Permalink
Cannot retrieve contributors at this time
Fetching contributors…
Cannot retrieve contributors at this time
| <?xml version="1.0" encoding="UTF-8"?><!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.1d1 20130915//EN" "JATS-archivearticle1.dtd"><article article-type="research-article" dtd-version="1.1d1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><front><journal-meta><journal-id journal-id-type="nlm-ta">elife</journal-id><journal-id journal-id-type="hwp">eLife</journal-id><journal-id journal-id-type="publisher-id">eLife</journal-id><journal-title-group><journal-title>eLife</journal-title></journal-title-group><issn publication-format="electronic">2050-084X</issn><publisher><publisher-name>eLife Sciences Publications, Ltd</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="publisher-id">04490</article-id><article-id pub-id-type="doi">10.7554/eLife.04490</article-id><article-categories><subj-group subj-group-type="display-channel"><subject>Research article</subject></subj-group><subj-group subj-group-type="heading"><subject>Ecology</subject></subj-group><subj-group subj-group-type="heading"><subject>Plant biology</subject></subj-group></article-categories><title-group><article-title>Plant defense phenotypes determine the consequences of volatile emission for individuals and neighbors</article-title></title-group><contrib-group><contrib contrib-type="author" corresp="yes" id="author-1399"><name><surname>Schuman</surname><given-names>Meredith C</given-names></name><contrib-id contrib-id-type="orcid">http://orcid.org/0000-0003-3159-3534</contrib-id><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="corresp" rid="cor1">*</xref><xref ref-type="other" rid="par-1"/><xref ref-type="other" rid="par-2"/><xref ref-type="fn" rid="con1"/><xref ref-type="fn" rid="conf2"/></contrib><contrib contrib-type="author" id="author-3416"><name><surname>Allmann</surname><given-names>Silke</given-names></name><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="aff" rid="aff2">2</xref><xref ref-type="other" rid="par-1"/><xref ref-type="other" rid="par-3"/><xref ref-type="other" rid="par-4"/><xref ref-type="fn" rid="con3"/><xref ref-type="fn" rid="conf2"/></contrib><contrib contrib-type="author" id="author-1013"><name><surname>Baldwin</surname><given-names>Ian T</given-names></name><contrib-id contrib-id-type="orcid">http://orcid.org/0000-0001-5371-2974</contrib-id><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="other" rid="par-1"/><xref ref-type="other" rid="par-2"/><xref ref-type="fn" rid="con2"/><xref ref-type="fn" rid="conf1"/></contrib><aff id="aff1"><label>1</label><institution content-type="dept">Department of Molecular Ecology</institution>, <institution>Max Planck Institute for Chemical Ecology</institution>, <addr-line><named-content content-type="city">Jena</named-content></addr-line>, <country>Germany</country></aff><aff id="aff2"><label>2</label><institution content-type="dept">Department of Plant Physiology, Swammerdam Institute of Life Sciences</institution>, <institution>University of Amsterdam</institution>, <addr-line><named-content content-type="city">Amsterdam</named-content></addr-line>, <country>Netherlands</country></aff></contrib-group><contrib-group content-type="section"><contrib contrib-type="editor" id="author-1094"><name><surname>Greenberg</surname><given-names>Jean T</given-names></name><role>Reviewing editor</role><aff><institution>University of Chicago</institution>, <country>United States</country></aff></contrib></contrib-group><author-notes><corresp id="cor1"><label>*</label>For correspondence: <email>mschuman@ice.mpg.de</email></corresp></author-notes><pub-date publication-format="electronic" date-type="pub"><day>15</day><month>04</month><year>2015</year></pub-date><pub-date pub-type="collection"><year>2015</year></pub-date><volume>4</volume><elocation-id>e04490</elocation-id><history><date date-type="received"><day>24</day><month>08</month><year>2014</year></date><date date-type="accepted"><day>11</day><month>03</month><year>2015</year></date></history><permissions><copyright-statement>© 2015, Schuman et al</copyright-statement><copyright-year>2015</copyright-year><copyright-holder>Schuman et al</copyright-holder><license xlink:href="http://creativecommons.org/licenses/by/4.0/"><license-p>This article is distributed under the terms of the <ext-link ext-link-type="uri" xlink:href="http://creativecommons.org/licenses/by/4.0/">Creative Commons Attribution License</ext-link>, which permits unrestricted use and redistribution provided that the original author and source are credited.</license-p></license></permissions><self-uri content-type="pdf" xlink:href="elife04490.pdf"/><abstract><object-id pub-id-type="doi">10.7554/eLife.04490.001</object-id><p>Plants are at the trophic base of terrestrial ecosystems, and the diversity of plant species in an ecosystem is a principle determinant of community structure. This may arise from diverse functional traits among species. In fact, genetic diversity within species can have similarly large effects. However, studies of intraspecific genetic diversity have used genotypes varying in several complex traits, obscuring the specific phenotypic variation responsible for community-level effects. Using lines of the wild tobacco <italic>Nicotiana attenuata</italic> genetically altered in specific well-characterized defense traits and planted into experimental populations in their native habitat, we investigated community-level effects of trait diversity in populations of otherwise isogenic plants. We conclude that the frequency of defense traits in a population can determine the outcomes of these traits for individuals. Furthermore, our results suggest that some ecosystem-level services afforded by genetically diverse plant populations could be recaptured in intensive monocultures engineered to be functionally diverse.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.001">http://dx.doi.org/10.7554/eLife.04490.001</ext-link></p></abstract><abstract abstract-type="executive-summary"><object-id pub-id-type="doi">10.7554/eLife.04490.002</object-id><title>eLife digest</title><p>Plants are at the base of many food webs. This means that the different traits and characteristics of the plant species in an ecosystem can have a large impact on the animals and other organisms that live there.</p><p>Individuals within the same plant species often differ in multiple genes. This ‘genetic diversity’ can affect the populations of other organisms in their ecosystem, for example, by altering which species are present, and the number of individuals. However, previous studies that investigated plant genetic diversity in ecosystems used plants that varied in multiple, usually unknown, genetic traits, which made it difficult to identify specific genetic traits in plants that can influence the whole ecosystem.</p><p>One way that plants affect their ecosystems involves how they defend themselves against the herbivores that try to eat them. For example, plants can use sharp spines or harmful chemicals to deter herbivores, or attract predators that will attack the herbivores.</p><p>Here, Schuman et al. carried out a 2-year field study using wild tobacco plants that had been genetically altered to employ different defensive strategies. This was achieved by altering the expression of three genes in the plants in specific combinations. Two of the genes, called <italic>LOX2</italic> and <italic>LOX3</italic>, are required to make most of the chemicals that tobacco plants use to defend themselves against herbivores. The third gene, called <italic>TPS10</italic>, which comes from the crop plant maize, gives plants the ability to release a fragrance that attracts natural predators of their herbivores.</p><p>Except for these specific alterations, the plants were otherwise genetically identical. These plants were then grown either alone, or in groups of five plants, which reflects the normal size of groups of wild tobacco growing in its natural environment. The groups contained mixtures of plants with different gene alterations.</p><p>Schuman et al. found that the expression levels of these genes in individual plants could determine how well the whole group fared in several different measures of plant health and defense. For example, plants lacking <italic>LOX2</italic> and <italic>LOX3</italic> usually appeared to be less healthy and they were less likely to survive long enough to reproduce because they were less able to defend themselves against herbivores. However, if these plants were grown in a group with one plant that expressed <italic>TPS10</italic>, all of the plants in the group looked healthier and were more likely to survive long enough to reproduce.</p><p>Many crops are grown in large fields containing individual plants that are largely genetically identical, which makes them more vulnerable to disease. Schuman et al.'s findings suggest that some of the community-level protective effects provided by diversity in wild populations of plants could be introduced into crop fields by altering the expression of a few specific genes in some individuals.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.002">http://dx.doi.org/10.7554/eLife.04490.002</ext-link></p></abstract><kwd-group kwd-group-type="author-keywords"><title>Author keywords</title><kwd><italic>Nicotiana attenuata</italic></kwd><kwd><italic>Manduca sexta</italic></kwd><kwd>Geocoris spp.</kwd><kwd><italic>Trichobaris mucorea</italic></kwd><kwd>direct and indirect defense</kwd><kwd>diversity</kwd></kwd-group><kwd-group kwd-group-type="research-organism"><title>Research organism</title><kwd>other</kwd></kwd-group><funding-group><award-group id="par-1"><funding-source><institution-wrap><institution-id institution-id-type="FundRef">http://dx.doi.org/10.13039/501100004189</institution-id><institution>Max-Planck-Gesellschaft</institution></institution-wrap></funding-source><principal-award-recipient><name><surname>Schuman</surname><given-names>Meredith C</given-names></name><name><surname>Allmann</surname><given-names>Silke</given-names></name><name><surname>Baldwin</surname><given-names>Ian T</given-names></name></principal-award-recipient></award-group><award-group id="par-2"><funding-source><institution-wrap><institution-id institution-id-type="FundRef">http://dx.doi.org/10.13039/501100000781</institution-id><institution>European Research Council (ERC)</institution></institution-wrap></funding-source><award-id>Advanced Grant to ITB 293926</award-id><principal-award-recipient><name><surname>Schuman</surname><given-names>Meredith C</given-names></name><name><surname>Baldwin</surname><given-names>Ian T</given-names></name></principal-award-recipient></award-group><award-group id="par-3"><funding-source><institution-wrap><institution-id institution-id-type="FundRef">http://dx.doi.org/10.13039/501100003246</institution-id><institution>Nederlandse Organisatie voor Wetenschappelijk Onderzoek</institution></institution-wrap></funding-source><award-id>NWO-ALW 821.02.027</award-id><principal-award-recipient><name><surname>Allmann</surname><given-names>Silke</given-names></name></principal-award-recipient></award-group><award-group id="par-4"><funding-source><institution-wrap><institution>European Union</institution></institution-wrap></funding-source><award-id>FP7-PEOPLE-2011-IEF-302388</award-id><principal-award-recipient><name><surname>Allmann</surname><given-names>Silke</given-names></name></principal-award-recipient></award-group><funding-statement>The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.</funding-statement></funding-group><custom-meta-group><custom-meta><meta-name>elife-xml-version</meta-name><meta-value>2.3</meta-value></custom-meta><custom-meta specific-use="meta-only"><meta-name>Author impact statement</meta-name><meta-value>Both the frequency of sesquiterpene-emitting individuals and the defense capacity of individual plants determine the consequences of sesquiterpene volatile emission for individuals and their neighbors in populations of the wild tobacco <italic>Nicotiana attenuata</italic>.</meta-value></custom-meta></custom-meta-group></article-meta></front><body><sec sec-type="intro" id="s1"><title>Introduction</title><p>In <italic>The Origin of Species</italic>, Darwin wrote that ‘varieties are species in the process of formation’ (<xref ref-type="bibr" rid="bib19">Darwin, 1876</xref>). We now know that Darwin's ‘varieties’ are polymorphic characters resulting from natural genetic variation within species. The geneticist <xref ref-type="bibr" rid="bib23">Dobzhansky (1973)</xref> echoed Darwin's view: ‘It is evident that the differences among proteins at the levels of species, genus, family, order, class, and phylum are compounded of elements that vary also among individuals within a species’. It has also long been known that intraspecific genetic variation in plants can alter ecological community structure; for example, in 1960, De Wit reported that the reproductive output in mixtures of the grass species <italic>Anthoxanthum odoratum</italic> and <italic>Phleum pratens</italic> depended on associations of particular genotypes (<xref ref-type="bibr" rid="bib21">De Wit, 1960</xref>; <xref ref-type="bibr" rid="bib2">Allard and Adams, 1969</xref>). However, efforts to measure the broader consequences for communities are more recent (<xref ref-type="bibr" rid="bib6">Antonovics, 1992</xref>; reviewed in <xref ref-type="bibr" rid="bib37">Haloin and Strauss, 2008</xref>) and include the recently described field of community genetics.</p><p>Community genetics aims to quantify how genetic variation in specific traits affects the structure of ecological communities (reviewed in <xref ref-type="bibr" rid="bib83">Whitham et al., 2006</xref>; <xref ref-type="bibr" rid="bib37">Haloin and Strauss, 2008</xref>; <xref ref-type="bibr" rid="bib84">Whitham et al., 2012</xref>). Genes most likely to have community effects are those with a large effect in the foundation species that structure communities (<xref ref-type="bibr" rid="bib83">Whitham et al., 2006</xref>), and plants are foundation species in almost all communities (<xref ref-type="bibr" rid="bib37">Haloin and Strauss, 2008</xref>). Community genetics is a basic level of investigation—meaning it is close to the basic unit acted on by evolution, the gene—within the larger effort to meet what <xref ref-type="bibr" rid="bib37">Haloin and Strauss (2008)</xref> refer to as ‘one of the supreme challenges of biology’: understanding how the feedback loops between community ecology and evolution shape biodiversity.</p><p>The maintenance of biodiversity and the ecosystem services it supports is a matter of increasing concern as the human population grows, and agricultural resource use increases (<xref ref-type="bibr" rid="bib33">Green et al., 2005</xref>). In a meta-analysis of 446 biodiversity measures taken between 1954 and 2004, of which 312 manipulated plant species biodiversity, species-level biodiversity was shown to positively affect several ecosystem properties including stability, regulation of biodiversity, nutrient cycling, erosion control, and productivity at the primary, secondary, and tertiary levels (<xref ref-type="bibr" rid="bib9">Balvanera et al., 2006</xref>). Functional differences among species are often thought to be the source of effects resulting from species-level biodiversity (<xref ref-type="bibr" rid="bib26">Ebeling et al., 2014</xref>). In fact, a growing number of studies has shown that genetic diversity within species can have large effects on ecological community composition (e.g., <xref ref-type="bibr" rid="bib18">Crutsinger et al., 2006</xref>; <xref ref-type="bibr" rid="bib43">Johnson et al., 2006</xref>; <xref ref-type="bibr" rid="bib17">Cook-Patton et al., 2011</xref>). <xref ref-type="bibr" rid="bib84">Whitham et al. (2012)</xref> refer to the genetically based tendency for particular genotypes to support different ecological communities as ‘community specificity’. Some, or even most, biodiversity effects may be explained through the community specificity of particular genetic traits.</p><p>Four postulates of community genetics provide the roadmap for understanding the community effects of genes (<xref ref-type="bibr" rid="bib85">Wymore et al., 2011</xref>; <xref ref-type="bibr" rid="bib84">Whitham et al., 2012</xref>): (1) the focal species must have a significant effect on its community or ecosystem, (2) the trait under investigation must be genetically based and heritable, (3) conspecifics differing genetically in this trait must have quantifiably different effects on community and ecosystem processes, and (4) a predictable effect on those community and ecosystem processes must follow when the gene of interest is manipulated. As Koch's postulates first allowed medical researchers to identify disease-causing agents, these postulates allow community geneticists to identify genes responsible for structuring specific aspects of ecological communities. So far, although postulates one through three have been satisfied in several ecological systems, postulate four—the critical step in attributing community functions to specific genes—has been tested in only a handful of systems (<xref ref-type="bibr" rid="bib83">Whitham et al., 2006</xref>, <xref ref-type="bibr" rid="bib84">2012</xref>). <xref ref-type="bibr" rid="bib84">Whitham et al. (2012)</xref> suggest that improved genetic tools will soon permit the widespread testing of postulate four, as has been done for nicotine in the wild tobacco <italic>Nicotiana attenuata</italic>: an ecological model plant with a transformation system and a well-characterized ecological community (<xref ref-type="bibr" rid="bib55">Krügel et al., 2002</xref>; <xref ref-type="bibr" rid="bib78">Steppuhn et al., 2004</xref>; <xref ref-type="bibr" rid="bib53">Kessler et al., 2008</xref>; reviewed in; <xref ref-type="bibr" rid="bib73">Schuman and Baldwin, 2012</xref>).</p><p>In this study, we built on deep molecular ecological knowledge of <italic>N. attenuata</italic> to demonstrate that all four postulates for genes controlling a suite of ecologically characterized defense traits (<xref ref-type="bibr" rid="bib50">Kessler and Baldwin, 2001</xref>; <xref ref-type="bibr" rid="bib52">Kessler et al., 2004</xref>; <xref ref-type="bibr" rid="bib4">Allmann and Baldwin, 2010</xref>; <xref ref-type="bibr" rid="bib73">Schuman et al., 2012</xref>). Specifically, we manipulated direct defenses: traits which directly reduce plant fitness loss to herbivores, and indirect defenses: traits which attract enemies of herbivores which remove or impair herbivores, and thereby indirectly reduce plant fitness loss to herbivores. <italic>N. attenuata</italic>'s direct and indirect defenses, their biosynthesis and ecological functions have been characterized in detail over two decades of research (reviewed in <xref ref-type="bibr" rid="bib73">Schuman and Baldwin, 2012</xref>). The direct defenses comprise a suite of chemical traits, most of which are induced by herbivory via jasmonate signaling, and genetically silencing these defenses causes significant changes in herbivore communities (<xref ref-type="bibr" rid="bib52">Kessler et al., 2004</xref>; <xref ref-type="bibr" rid="bib30">Gaquerel et al., 2010</xref>; <xref ref-type="bibr" rid="bib47">Kallenbach et al., 2012</xref>). <italic>N. attenuata</italic> also produces a complex blend of herbivore-induced plant volatiles (HIPVs) which can be parsed into terpenoids, GLVs and other fatty acid derivatives, and aromatics, and are regulated by elicitors in herbivore oral secretions (OS) (<xref ref-type="bibr" rid="bib31">Gaquerel et al., 2009</xref>; <xref ref-type="bibr" rid="bib74">Schuman et al., 2009</xref>). Of these, the jasmonate-regulated sesquiterpene (<italic>E</italic>)-α-bergamotene (TAB) and the GLVs have been shown to reduce herbivore loads by attracting predators, which for the GLVs has also been shown to result in greater plant fitness (<xref ref-type="bibr" rid="bib35">Halitschke et al., 2000</xref>; <xref ref-type="bibr" rid="bib50">Kessler and Baldwin, 2001</xref>; <xref ref-type="bibr" rid="bib36">Halitschke et al., 2008</xref>; <xref ref-type="bibr" rid="bib74">Schuman et al., 2009</xref>; <xref ref-type="bibr" rid="bib4">Allmann and Baldwin, 2010</xref>; <xref ref-type="bibr" rid="bib73">Schuman et al., 2012</xref>). The majority of <italic>N. attenuata</italic>'s direct and indirect chemical defenses can be abolished by silencing the expression of two biosynthetic genes: LIPOXYGENASE 2 (<italic>LOX2</italic>), which provides substrate for GLV biosynthesis; and <italic>LOX3</italic>, essential for jasmonate biosynthesis, and thus also for the expression of most direct defense traits and the emission of TAB (<xref ref-type="bibr" rid="bib52">Kessler et al., 2004</xref>; <xref ref-type="bibr" rid="bib5">Allmann et al., 2010</xref>).</p><p>We specifically manipulated the expression of genes controlling direct and indirect defense traits in a factorial design by combining the silencing of <italic>LOX2</italic> and <italic>LOX3</italic> with the overexpression of <italic>Zea mays TERPENE SYNTHASE 10</italic> (<italic>TPS10</italic>), which produces TAB and (<italic>E</italic>)-β-farnesene (TBF) as its main products (Köllner et al., 2009; <xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>). This allowed us to independently manipulate total direct and indirect defenses, and the specific volatiles, TAB and TBF. We grew these transgenic, otherwise isogenic genotypes during two consecutive field seasons, both in replicated populations of different compositions, and as individuals.</p><p>Studies of genetic diversity and community specificity most commonly measure the diversity and composition of invertebrate communities on plants, but also invertebrate abundance, and occasionally functional measures such as predation rates (reviewed in <xref ref-type="bibr" rid="bib37">Haloin and Strauss, 2008</xref>). We focused on correlates of plant fitness: plant size, apparent health, flowering, and mortality before reproductive maturity; and on functional measures of the effectiveness of plant defense: total canopy damage from herbivores, herbivore abundance, and predation services received by plants in populations. The functional measures were chosen to reflect our observation of plants' ecological communities in both years. Nearly, all of these measures of plant fitness and interactions with herbivores and their predators were significantly affected not only by the genotype of individual plants, but by neighboring genotypes in populations.</p></sec><sec sec-type="results" id="s2"><title>Results</title><p>Data from the study was deposited in Dryad (10.5061/dryad.qj007, <xref ref-type="bibr" rid="bib90">Schuman et al., 2015</xref>).</p><sec id="s2-1"><title>Genetically engineered phenotypes are maintained in experimental populations</title><p>Transgenic lines manipulated in the emission of TAB, TBF, and GLVs, and the production of total jasmonate-mediated defenses (<xref ref-type="fig" rid="fig1">Figure 1</xref> and associated source data files) were screened and characterized as described in <xref ref-type="app" rid="app1">Appendix 1</xref>. To investigate the community-level effects of variation in TAB and TBF in well- and poorly defended plants, we used four genotypes: WT plants, which have jasmonate-mediated induced direct defense compounds and herbivore-induced volatile emission comprising primarily sesquiterpenes and GLVs; <italic>TPS10</italic> plants, which are like WT but have both enhanced induced, as well as constitutive emission of the herbivore-induced sesquiterpenes TAB and TBF (<xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>); <italic>lox2/3</italic> plants, which are deficient in both jasmonate-mediated direct defense compounds and herbivore-induced volatiles, and <italic>lox2/3</italic>x<italic>TPS10</italic> plants, which are like <italic>lox2/3</italic>, but with constitutive emission of TAB and TBF.<fig id="fig1" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.003</object-id><label>Figure 1.</label><caption><title>The <italic>lox2/3</italic> and <italic>TPS10</italic> transgenic constructs, alone and in combination (<italic>lox2/3xTPS10</italic>), independently alter herbivore-induced plant defenses and sesquiterpene HIPVs.</title><p>(<bold>A</bold>) The endogenous lipoxygenase genes <italic>LOX2</italic> and <italic>LOX3</italic> and the <italic>Z. mays</italic> sesquiterpene synthase gene <italic>TPS10</italic> (<italic>TPS10</italic>) were manipulated in transgenic lines of <italic>N. attenuata</italic> in order to uncouple the production of two HIPVs—the sesquiterpenes (<italic>E</italic>)-α-bergamotene (TAB) and (<italic>E</italic>)-β-farnesene (TBF)—from jasmonate-mediated direct defenses and other volatiles. LOX2 and LOX3 provide fatty acid hydroperoxides from 18:2 and 18:3 fatty acids (18:2/18:3-OOH) for the synthesis of green leaf volatiles (GLVs) or jasmonic acid (JA) and the jasmonate-derived hormones, respectively. GLVs are released in response to herbivore damage, and jasmonates regulate the production of most anti-herbivore defenses, including trypsin protease inhibitors (TPIs), and other HIPVs. LOX2 and LOX3 were transgenically silenced using an inverted repeat (ir) construct specifically silencing both (<italic>lox2/3</italic>). Jasmonate-regulated volatiles in <italic>N. attenuata</italic> are mostly sesquiterpenes, of which the most prevalent is TAB, and include trace amounts of the biosynthetically related sesquiterpene TBF. The <italic>Z. mays</italic> sesquiterpene synthase <italic>TPS10</italic> (Köllner et al., 2009), which produces TAB and TBF as its main products, was ectopically over-expressed in <italic>N. attenuata</italic> plants under control of a 35S promoter (<italic>TPS10</italic>) to uncouple the emission of these volatiles from endogenous jasmonate signaling. (<bold>B</bold> and <bold>C</bold>) Plants with <italic>lox2/3</italic> and <italic>TPS10</italic> constructs, and a cross with both constructs having the transgenes in a hemizygous state (<italic>lox2/3xTPS10</italic>) had the expected phenotypes both in glasshouse and field experiments. For each compound or group of compounds, the mean WT value was set to 1 (indicated by dashed lines), and all values were divided by the WT mean resulting in mean fold changes ±SEM. A complete description of <italic>TPS10</italic> plants including the demonstration of a single transgene insertion and WT levels of non-target metabolites is provided in <xref ref-type="bibr" rid="bib71">Schuman et al. (2014)</xref>, and additional information for <italic>lox2/3</italic>, <italic>TPS10</italic>, and <italic>lox2/3xTPS10</italic> plants (single transgene insertion for <italic>lox2/3</italic>, accumulation of target gene transcripts) is given in <xref ref-type="app" rid="app1">Appendix 1</xref>. (<bold>B</bold>) Glasshouse-grown plants were treated with wounding and <italic>M</italic>. <italic>sexta</italic> OS, and JA, GLVs, TAB and TBF were analyzed at the time of peak accumulation; n = 4. For <italic>lox2/3</italic> and <italic>lox2/3xTPS10</italic>, GLVs were not quantifiable (NQ) by GC–MS due to inconsistently detected or no detected signals. ***p < 0.001 in Tukey HSD tests following a significant one-way ANOVA (p < 0.001) of WT, <italic>lox2/3</italic>, and <italic>lox2/3xTPS10</italic>; # plants with the <italic>TPS10</italic> construct emit significantly more TAB + TBF (Holm-Bonferroni-corrected p < 0.01 in a Wilcoxon rank sum test of WT and <italic>lox2/3</italic> vs <italic>lox2/3xTPS10</italic> and <italic>TPS10</italic>); ns, not significant. TBF was not detectable in <italic>lox2/3</italic> or WT samples. For absolute amounts of JA and other phytohormones, GLVs, TAB and TBF measured in glasshouse samples, see <xref ref-type="supplementary-material" rid="SD1-data SD2-data">Figure 1—source data 1, 2</xref>. (<bold>C</bold>) TPI activity, total detectable GLVs, and total TAB and TBF measured in frozen leaf samples harvested from field-grown plants at the end of experimental season two (June 28th), after plants had accumulated damage from naturally occurring herbivores. TPIs (n = 11–24) were slightly elevated in field samples of <italic>TPS10</italic> plants, but TPI activity did not differ from WT plants in glasshouse samples from two independent <italic>TPS10</italic> lines including the line used for this study (<xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>). Total GLVs, TAB, and TBF were quantified by GC–MS in hexane extracts from the same tissue samples used to measure TPI activity, n = 9–22 (a few samples had insufficient tissue for the analysis). Neither TAB nor TBF could be detected in <italic>lox2/3</italic> samples (NQ); TBF was also not detectable in WT samples. *Corrected p < 0.05, ***corrected p < 0.001 in pairwise Wilcoxon rank sum tests following significant (corrected p < 0.05) Kruskal–Wallis tests across all genotypes for each category; p<italic>-</italic>values were corrected for multiple testing using the Holm-Bonferroni method; ns, not significant. For absolute amounts of TPIs, GLVs, TAB and TBF, and non-target volatiles measured in field-collected tissue samples, see <xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref>; emission of GLVs, TAB, TBF, and non-target volatiles from field-grown plants is given in <xref ref-type="app" rid="app2">Appendix 2</xref>.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.003">http://dx.doi.org/10.7554/eLife.04490.003</ext-link></p><p><supplementary-material id="SD1-data"><object-id pub-id-type="doi">10.7554/eLife.04490.004</object-id><label>Figure 1—source data 1.</label><caption><title>Absolute values for JA and GLVs shown as fold-changes in <xref ref-type="fig" rid="fig1">Figure 1B</xref> in <italic>lox2/3</italic> and <italic>lox2/3xTPS10</italic> plants, as well as JA-Ile, individual GLVs, and leaf areas for headspace collection (mean ± SEM); n = 4.</title><p>Data are from a single glasshouse experiment. Values for <italic>TPS10</italic> (line 10–3) are reported in <xref ref-type="bibr" rid="bib71">Schuman et al. (2014)</xref>.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.004">http://dx.doi.org/10.7554/eLife.04490.004</ext-link></p></caption><media xlink:href="elife04490s001.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p><p><supplementary-material id="SD2-data"><object-id pub-id-type="doi">10.7554/eLife.04490.005</object-id><label>Figure 1—source data 2.</label><caption><title>Absolute values for TAB and TBF shown as fold-changes in <xref ref-type="fig" rid="fig1">Figure 1B</xref>, and leaf areas for headspace collection (mean ± SEM); n = 4.</title><p>Data are from a single glasshouse experiment.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.005">http://dx.doi.org/10.7554/eLife.04490.005</ext-link></p></caption><media xlink:href="elife04490s002.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p><p><supplementary-material id="SD3-data"><object-id pub-id-type="doi">10.7554/eLife.04490.006</object-id><label>Figure 1—source data 3.</label><caption><title>Absolute values for analytes shown as fold-changes in <xref ref-type="fig" rid="fig1">Figure 1C</xref> (mean ± SEM).</title><p>Data are from tissue samples of several pooled leaves per plant from field-grown plants, taken at the end of experimental season two (June 28th).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.006">http://dx.doi.org/10.7554/eLife.04490.006</ext-link></p></caption><media xlink:href="elife04490s003.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xlink:href="elife04490f001"/></fig></p><p>These genotypes were grown as individuals and in five-plant populations in a field plot in <italic>N. attenuata</italic>'s native habitat (<xref ref-type="fig" rid="fig2">Figure 2A</xref>, <xref ref-type="supplementary-material" rid="SD7-data">Figure 2—figure supplements 1, 2</xref>). A population size of five plants is realistic for natural populations of <italic>N. attenuata</italic> (<xref ref-type="fig" rid="fig2s3">Figure 2—figure supplement 3</xref>), particularly before 1995, after which an invasive brome grass fueled a dramatic increase in fire sizes and consequently <italic>N. attenuata</italic> population sizes. This choice of population size also allowed us to clearly differentiate between edge and center positions (<xref ref-type="fig" rid="fig2">Figure 2B</xref>). Populations were either monocultures, or mixed cultures containing a single <italic>TPS10</italic>-expressing plant at the center and plants of the same defense- and HIPV level at the edges (WT around <italic>TPS10</italic>, and <italic>lox2/3</italic> around <italic>lox2/3</italic>x<italic>TPS10</italic>, <xref ref-type="fig" rid="fig2">Figure 2B</xref>).<fig-group><fig id="fig2" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.007</object-id><label>Figure 2.</label><caption><title>Field experiments designed to measure genotype-by-population effects of TPS10 volatiles in well-defended (WT, <italic>TPS10</italic>) or poorly defended (<italic>lox2/3</italic>, <italic>lox2/3xTPS10</italic>) plants.</title><p>Two similar experiments were conducted in two consecutive field seasons. (<bold>A</bold>) Each genotype was planted as individuals and in five-plant populations (see <xref ref-type="fig" rid="fig2s1">Figure 2—figure supplement 1</xref>). Each individual or population was planted ca. 1.5 m from the nearest neighboring individual/population, measured perimeter-to-perimeter. Populations consisted of five plants arranged as shown in a 0.4 × 0.4 m square. (<bold>B</bold>) Individual and population types: monocultures and, additionally, mixed cultures were planted in which a <italic>TPS10</italic> plant (<italic>TPS10</italic> or <italic>lox2/3xTPS10</italic>) was surrounded by four plants of the same level of jasmonate-mediated defense and GLVs (WT or <italic>lox2/3</italic>). Replicates (n = 12 in season one, 15 in season two of each individual or population type) were arranged in a blocked design (see <xref ref-type="fig" rid="fig2s2">Figure 2—figure supplement 2</xref>): blocks consisting of one replicate of each type (all six randomly-ordered populations followed by all four randomly-ordered individuals) were staggered such that consecutive blocks were not aligned, and thus populations and individuals were also interspersed. Random order was modified only when necessary to ensure that no two replicate individuals/populations were placed next to each other vertically, horizontally, or diagonally. This planting design reflects typical distributions in native populations of <italic>Nicotiana attenuata</italic> (see <xref ref-type="fig" rid="fig2s3">Figure 2—figure supplement 3</xref>), particularly before 1995, after which an invasive brome grass fueled a dramatic increase in fire sizes and consequently <italic>N. attenuata</italic> population sizes.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.007">http://dx.doi.org/10.7554/eLife.04490.007</ext-link></p></caption><graphic xlink:href="elife04490f002"/></fig><fig id="fig2s1" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04490.008</object-id><label>Figure 2—figure supplement 1.</label><caption><title>Photographs of the field experiment in season one.</title><p>(<bold>A</bold>) Photograph of a closed-headspace volatile trapping showing ca. 25% of the experiment in May of season one. (<bold>B</bold>) Photograph showing a similar portion of the experiment in June of season one.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.008">http://dx.doi.org/10.7554/eLife.04490.008</ext-link></p></caption><graphic xlink:href="elife04490fs001"/></fig><fig id="fig2s2" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04490.009</object-id><label>Figure 2—figure supplement 2.</label><caption><title>Layout in experimental season two, and color codes.</title><p>R, row of earth between two irrigation lines; not to scale.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.009">http://dx.doi.org/10.7554/eLife.04490.009</ext-link></p></caption><graphic xlink:href="elife04490fs002"/></fig><fig id="fig2s3" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04490.010</object-id><label>Figure 2—figure supplement 3.</label><caption><title>Example of plant distribution in a native <italic>N. attenuata</italic> population, photographed in 2004.</title><p>Reprinted with permission from Danny Kessler, Copyright 2004. All rights reserved.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.010">http://dx.doi.org/10.7554/eLife.04490.010</ext-link></p></caption><graphic xlink:href="elife04490fs003"/></fig></fig-group></p><p>There was no effect of population type (individual, monoculture, mixed culture) or plant position (individual, edge, center) on genetically engineered levels of GLVs, jasmonate-mediated defense (trypsin protease inhibitor [TPI] activity served as an indicator), TAB and TBF (<xref ref-type="fig" rid="fig1">Figure 1C</xref>, <xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref> and details in <xref ref-type="app" rid="app2">Appendix 2</xref>), with one exception: trace amounts of TAB were detected in open headspace trappings around young <italic>lox2/3</italic> plants at the edges of <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mixed cultures, but not around <italic>lox2/3</italic> individuals or <italic>lox2/3</italic> plants at the edges of monocultures (<xref ref-type="fig" rid="fig3">Figure 3</xref>).<fig id="fig3" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.011</object-id><label>Figure 3.</label><caption><title>The TPS10 product TAB can be detected in the open headspace around <italic>lox2/3</italic> plants in mixed cultures containing a <italic>lox2/3xTPS10</italic> plant.</title><p>(<bold>A</bold>) An ‘open headspace’ trapping was conducted during season two (May 16th) with rosette-stage individuals and edge plants as shown, using activated charcoal filters shielded from ozone and UV by MnO<sub>2</sub>-coated copper ozone scrubbers (black strips in Teflon tubes pointed at plant) and aluminum foil, respectively. The headspace was sampled during the day, when sesquiterpenes are most abundant (see <xref ref-type="app" rid="app2">Appendix 2</xref>). Eluents from all four filters were combined and analyzed using highly sensitive GCxGC-ToF analysis; n = 4 plants. (<bold>B</bold>) The <italic>TPS10</italic> product <italic>trans</italic>-α-bergamotene (TAB) could be detected in the open headspace of two of four <italic>lox2/3</italic> plants at the edges of mixed cultures containing a <italic>lox2/3xTPS10</italic> plant, but not in <italic>lox2/3</italic> individuals or <italic>lox2/3</italic> plants in monocultures. Peak areas were normalized to the internal standard (IS) peak. Arrows indicate the plant that was sampled.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.011">http://dx.doi.org/10.7554/eLife.04490.011</ext-link></p></caption><graphic xlink:href="elife04490f003"/></fig></p></sec><sec id="s2-2"><title><italic>LOX2</italic> and <italic>LOX3</italic> expression reduce foliar herbivore abundance and damage, while <italic>TPS10</italic> expression further reduces herbivore abundance</title><p>Prior to the field experiment, we tested whether <italic>TPS10</italic> expression affects defense against a specialist herbivore in a WT or <italic>LOX2/3</italic>-deficient background via a no-choice assay with larvae of the naturally co-occurring solanaceous specialist <italic>Manduca</italic> <italic>sexta</italic> in the glasshouse. <italic>M. sexta</italic> larvae grew five times as large on <italic>lox2/3</italic> or <italic>lox2/3</italic>x<italic>TPS10</italic> plants as they did on WT or <italic>TPS10</italic> plants (<xref ref-type="fig" rid="fig4">Figure 4</xref>; Wilcoxon rank sum tests, Holm-Bonferroni-corrected p<italic>-</italic>values < 0.001). In contrast, larvae grew to similar sizes on plants that differed only in the <italic>TPS10</italic> overexpression construct (corrected p-values > 0.2). The progression of <italic>M. sexta</italic> larval instars was also faster on <italic>LOX2/3</italic>-deficient plants: all larvae feeding on <italic>lox2/3</italic> or <italic>lox2/3</italic>x<italic>TPS10</italic> were in the third instar by day 8, while 11/17 larvae feeding on WT and 12/15 larvae feeding on <italic>TPS10</italic> were still in the second instar (corrected p < 0.001 in Fisher's tests for <italic>lox2/3</italic> v. WT, <italic>lox2/3</italic> v. <italic>TPS10</italic>, <italic>lox2/3</italic>x<italic>TPS10</italic> v. WT, and <italic>lox2/3</italic>x<italic>TPS10</italic> v. <italic>TPS10</italic>, corrected p = 0.4440 for WT v. <italic>TPS10</italic>, and corrected p = 1 for <italic>lox2/3</italic> v. <italic>lox2/3</italic>x<italic>TPS10</italic>).<fig id="fig4" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.012</object-id><label>Figure 4.</label><caption><title><italic>Manduca sexta</italic> larvae grow larger on <italic>LOX2/3</italic>-deficient plants, regardless of <italic>TPS10</italic> expression, in a glasshouse experiment.</title><p>Data are shown as mean ± SEM; n = 15–19 larvae on day 8 and 11–16 larvae on day 12. One <italic>M. sexta</italic> neonate per plant (starting n = 25 larvae) was placed immediately after hatching on the youngest rosette leaf of an elongated plant. Larvae were weighed at the third and fourth instars (of 5 total), corresponding to days 8 and 12, after which larvae become mobile between plants. <sup>a,b/A,B</sup> Different letters indicate significant differences (corrected p < 0.001) in larval mass on different plant genotypes within each day, in Wilcoxon rank sum tests following significant (corrected p < 0.001) Kruskal–Wallis tests for each day (WT vs <italic>lox2/3</italic> day 8, W<sub>17,19</sub> = 310, corrected p <italic><</italic> 0.001, day 12, W<sub>15,15</sub> = 217, corrected p <italic><</italic> 0.001; <italic>TPS10</italic> vs <italic>lox2/3</italic>x<italic>TPS10</italic>, day 8, W<sub>15,20</sub> = 290, corrected p <italic><</italic> 0.001, day 12, W<sub>11,14</sub> = 143, corrected p <italic><</italic> 0.001; WT vs <italic>TPS10</italic> day 8, W<sub>17,15</sub> = 95, corrected p = 0.227, day 12, W<sub>15,11</sub> = 56, corrected p <italic>=</italic> 0.361; <italic>lox2/3</italic> vs <italic>lox2/3</italic>x<italic>TPS10</italic> day 8, W<sub>19,20</sub> = 249, corrected p = 0.201, day 12, W<sub>16,14</sub> = 134, corrected p = 0.377). p<italic>-</italic>values were corrected for multiple testing using the Holm-Bonferroni method.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.012">http://dx.doi.org/10.7554/eLife.04490.012</ext-link></p></caption><graphic xlink:href="elife04490f004"/></fig></p><p>In the field, total canopy damage from foliar herbivores was assessed on June 9th in season one, and June 2nd in season two. Canopy damage in both seasons was due to both generalist and specialist herbivores from chewing and piercing-sucking feeding guilds. The generalists included noctuid larvae, <italic>Trimerotropis</italic> spp. grasshoppers, <italic>Oecanthus</italic> spp. tree crickets, and leaf miner spp.; the specialists included the sphingid larvae <italic>M. sexta</italic> and <italic>Manduca quinquemaculata</italic>, flea beetles <italic>Epitrix hirtipennis</italic> and <italic>E. subcrintia</italic>, and the mirid <italic>Tupiocoris notatus</italic>. Canopy damage from herbivores was more than twice as great in season two (9–24% total canopy) as in season one (2–9% total canopy).</p><p>In both seasons, <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants accumulated twice as much canopy damage as WT and <italic>TPS10</italic> plants, regardless of plant position or population type (<xref ref-type="fig" rid="fig5">Figure 5</xref> and associated source data files): there were no significant differences during either season between plants or populations differing only in <italic>TPS10</italic> expression, or plants of the same genotype in mono- vs mixed cultures with single <italic>TPS10</italic>-expressing plants; details of statistical tests are given in <xref ref-type="app" rid="app3">Appendix 3</xref>.<fig id="fig5" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.013</object-id><label>Figure 5.</label><caption><title>Only <italic>LOX2/3</italic> and not <italic>TPS10</italic>, plant position, or population type determined total foliar damage from herbivores over two consecutive field seasons.</title><p>Note differences in y-axis scale. Data were collected on June 9th in season one (left panels), and on June 2nd in season two (right panels) from plants at all different positions in the experiment: individual plants (<bold>A</bold>–<bold>B</bold>), and plants at the edges (<bold>C</bold>–<bold>D</bold>) or at the centers (<bold>E</bold>–<bold>F</bold>) of populations. Black bars denote WT, grey bars <italic>TPS10</italic>, red bars <italic>lox2/3</italic>, and pink bars <italic>lox2/3xTPS10</italic> plants. Total canopy damage reflected the trends in canopy damage caused by individual herbivores (<xref ref-type="supplementary-material" rid="SD4-data SD5-data">Figure 5—source data 1, 2</xref>). Total canopy damage in season one was low (left panels), ranging from 5–10% on <italic>lox2/3</italic> and <italic>lox2/3xTPS10</italic> plants, and only 2–5% on WT and <italic>TPS10</italic> plants. Damage levels were about twice as high in season two (right panels), but showed the same relative pattern, with <italic>LOX2/3</italic>-deficient plans having more damage. Damage levels within a season were similar for plants in different positions (p > 0.07, see <xref ref-type="app" rid="app3">Appendix 3</xref>). <sup>a,b</sup> Different letters indicate significant differences (p < 0.05) between <italic>LOX2/3</italic>-expressing and <italic>LOX2/3</italic>-deficient plants (WT, <italic>TPS10</italic> v. <italic>lox2/3</italic>, <italic>lox2/3xTPS10</italic>) in minimal ANOVA or linear mixed-effects models on arcsin-transformed data. For individuals and center plants n = 7–13, and for edge plants n = 16–48 (up to 4 per population; the blocking effect was accounted for by a random factor in statistical analysis); exact replicate numbers are given in <xref ref-type="supplementary-material" rid="SD4-data SD5-data">Figure 5—source data 1, 2</xref>. There were no significant differences in damage between plants differing only in <italic>TPS10</italic> expression (p > 0.4), plants in mono- vs mixed cultures (p > 0.5), or plants of the same genotype at different positions (individual, edge, or center, p > 0.07). Statistical models are given in <xref ref-type="app" rid="app3">Appendix 3</xref>.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.013">http://dx.doi.org/10.7554/eLife.04490.013</ext-link></p><p><supplementary-material id="SD4-data"><object-id pub-id-type="doi">10.7554/eLife.04490.014</object-id><label>Figure 5—source data 1.</label><caption><title>Total canopy damage from individual herbivores or groups of herbivores in experimental season one, corresponding to data shown in <xref ref-type="fig" rid="fig5">Figure 5A</xref> (mean ± SEM).</title><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.014">http://dx.doi.org/10.7554/eLife.04490.014</ext-link></p></caption><media xlink:href="elife04490s004.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p><p><supplementary-material id="SD5-data"><object-id pub-id-type="doi">10.7554/eLife.04490.015</object-id><label>Figure 5—source data 2.</label><caption><title>Total canopy damage from individual herbivores or groups of herbivores in experimental season two, corresponding to data shown in <xref ref-type="fig" rid="fig5">Figure 5B</xref> (mean ± SEM).</title><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.015">http://dx.doi.org/10.7554/eLife.04490.015</ext-link></p></caption><media xlink:href="elife04490s005.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xlink:href="elife04490f005"/></fig></p><p>Damage observed could be attributed to specific herbivores according to typical feeding patterns and the observation of feeding individuals. Individual herbivores generally followed the trend of total herbivore damage, causing greater damage on <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants (<xref ref-type="supplementary-material" rid="SD4-data SD5-data">Figure 5—source data 1, 2</xref>). The largest and most significant differences (p<italic>-</italic>values < 0.05) were between plants with vs without the <italic>lox2/3</italic> silencing construct (<italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> v. WT and <italic>TPS10</italic>, statistical comparisons in <xref ref-type="supplementary-material" rid="SD4-data SD5-data">Figure 5—source data 1, 2</xref>). All significant differences were between plants or populations with vs without the <italic>lox2/3</italic> silencing construct, and never between plants or populations differing only in the <italic>TPS10</italic> overexpression construct, nor between plants of the same genotype in different population types (<xref ref-type="fig" rid="fig5">Figure 5</xref> and associated source data files). It should be noted that in season one, multiple lepidopteran species were present on plants and their damage could not be clearly distinguished; since these included generalist species as well as <italic>Manduca</italic> spp., lepidopteran damage in season one also could not be categorized as generalist or specialist damage (<xref ref-type="supplementary-material" rid="SD4-data">Figure 5—source data 1</xref>). In season two, lepidopteran damage not resulting from experimental infestation for predation assays (not counted toward total canopy damage) was due to noctuid species (<xref ref-type="supplementary-material" rid="SD5-data">Figure 5—source data 2</xref>).</p><p>The abundance of arthropods on plant shoots was quantified for a subset of plants on June 26th in season two, at which time <italic>T. notatus</italic> adults and nymphs and <italic>Epitrix</italic> spp. adults were the only foliar herbivores observed. Interestingly, total herbivore abundance at the end of the season did not entirely reflect foliar herbivore damage assessed on June 9th (<xref ref-type="fig" rid="fig5">Figure 5B</xref>), although a visual assessment of the plants indicated that relative herbivore damage levels still reflected the June 9th data: herbivores were less abundant on plants expressing <italic>TPS10</italic>, independently of <italic>lox2/3</italic> expression (<xref ref-type="fig" rid="fig6">Figure 6A</xref> and associated source data files). Herbivore abundance on WT plants differed depending on plants' location: in a comparison of abundance on individuals vs plants at the edges or centers of monocultures, abundance was highest on individuals and lowest on center plants. In contrast, herbivore abundance on <italic>TPS10</italic>-expressing plants did not depend on plant location (<xref ref-type="fig" rid="fig6">Figure 6B</xref>, <xref ref-type="app" rid="app3">Appendix 3</xref>). Foliar herbivore abundance on WT and <italic>lox2/3</italic> edge plants was not affected by the presence of a <italic>TPS10</italic>-expressing neighbor: neither population type (monoculture or mixed culture), nor the interaction of genotype with population type was significant. However for WT and <italic>lox2/3</italic> edge plants, herbivore abundance did vary as a function of genotype, with abundance indeed being greater on <italic>lox2/3</italic> edge plants (<xref ref-type="fig" rid="fig6">Figure 6C</xref>, <xref ref-type="app" rid="app3">Appendix 3</xref>), consistent with greater damage to these plants (<xref ref-type="fig" rid="fig5">Figure 5D</xref>).<fig id="fig6" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.016</object-id><label>Figure 6.</label><caption><title>Foliar herbivore abundance is determined by <italic>TPS10</italic> expression, <italic>lox2/3</italic> expression, and plant position.</title><p>Data were collected on June 26th in season two, when herbivores and predators were more abundant (see <xref ref-type="fig" rid="fig5">Figure 5</xref>). Black denotes WT, grey <italic>TPS10</italic>, red <italic>lox2/3</italic>, and pink <italic>lox2/3xTPS10</italic> plants. Counts for individual herbivores and predators are given in <xref ref-type="supplementary-material" rid="SD6-data SD7-data">Figure 6—source data 1, 2</xref>. (<bold>A</bold>) Herbivore counts on focal plants of each genotype revealed fewer herbivores present on plants expressing <italic>TPS10</italic> (n given in <xref ref-type="supplementary-material" rid="SD6-data">Figure 6—source data 1</xref>). <sup>a,b</sup> Different letters indicate significant differences (corrected p < 0.05) between genotypes in Tukey contrasts following significant differences in a generalized linear mixed-effects model (see <xref ref-type="app" rid="app3">Appendix 3</xref>). (<bold>B</bold>) For WT and <italic>TPS10</italic> individuals or plants in monocultures (n given in <xref ref-type="supplementary-material" rid="SD7-data">Figure 6—source data 2</xref>), there was a significant effect (p < 0.05) of plant genotype (z = −3.662, p = 0.0003) and an interaction of genotype with position (<italic>TPS10</italic> by center vs individual position, z = 2.186, p = 0.0288, generalized linear mixed-effects model in <xref ref-type="app" rid="app3">Appendix 3</xref>). (<bold>C</bold>) The presence of a <italic>TPS10</italic>-expressing neighbor in mixed populations did not significantly affect herbivore abundance on WT or <italic>lox2/3</italic> edge plants, but there was a significant difference between WT and <italic>lox2/3</italic> genotypes at the edges of populations (n given in <xref ref-type="supplementary-material" rid="SD7-data">Figure 6—source data 2</xref>, z = 2.358, p = 0.0183, generalized linear model in <xref ref-type="app" rid="app3">Appendix 3</xref>).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.016">http://dx.doi.org/10.7554/eLife.04490.016</ext-link></p><p><supplementary-material id="SD6-data"><object-id pub-id-type="doi">10.7554/eLife.04490.017</object-id><label>Figure 6—source data 1.</label><caption><title>Total and median numbers of herbivores and predators counted on focal plants of different genotypes on June 26th in season two.</title><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.017">http://dx.doi.org/10.7554/eLife.04490.017</ext-link></p></caption><media xlink:href="elife04490s006.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p><p><supplementary-material id="SD7-data"><object-id pub-id-type="doi">10.7554/eLife.04490.018</object-id><label>Figure 6—source data 2.</label><caption><title>Total and median numbers of herbivores counted on focal plants of different genotypes on June 26th in season two, broken down into different population types and plant locations for WT and <italic>TPS10</italic> plants.</title><p>Replicate numbers of <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants were not sufficient for this analysis (n ≤ 3 per population type and position).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.018">http://dx.doi.org/10.7554/eLife.04490.018</ext-link></p></caption><media xlink:href="elife04490s007.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xlink:href="elife04490f006"/></fig></p><p>Thus <italic>LOX2</italic> and <italic>LOX3</italic> expression increased plants' susceptibility to foliar herbivores, measured as herbivore growth or percentage of plant canopy damaged (<xref ref-type="fig" rid="fig4 fig5">Figures 4, 5</xref>), but the relationship between foliar herbivore abundance and <italic>LOX2</italic>/<italic>LOX3</italic> expression was more complex and dependent on plants' position in populations (<xref ref-type="fig" rid="fig6">Figure 6</xref>). In contrast, herbivore abundance, but not damage or performance, was consistently reduced on <italic>TPS10</italic>-expressing plants (<xref ref-type="fig" rid="fig4 fig5 fig6">Figures 4–6</xref>).</p></sec><sec id="s2-3"><title><italic>LOX2/3</italic> deficiency accelerates flowering under low herbivory, while <italic>TPS10</italic> expression reduces flower production under high herbivory</title><p>We monitored plant growth and size, to control for these non-defense traits which could confound the analysis of differences among genotypes, plant positions, and population types, and their interactions with insect communities. Rosette diameter and plant stem height were measured at three times during the main period of vegetative growth in season one (May 1st–2nd, 8th–11th, and 16th–17th), and after plants had transitioned from vegetative growth to reproduction in season two (June 4th–6th), at which time stem diameter was also measured. There were no significant differences in rosette diameter or stem diameter (p-values > 0.08 in ANOVAs or linear mixed-effects models with LOX2/3 deficiency and TPS10 expression, plant position, and population type as factors: see statistical approach in <xref ref-type="app" rid="app3">Appendix 3</xref>). There were few differences in final measurements of plant height. <italic>LOX2/3</italic>-deficient plants were slightly taller at the final size measurement on May 17th of season one: the stem height of WT plants was 28.5 ± 1.1 cm, <italic>TPS10</italic> plants 27.2 ± 1.5 cm, <italic>lox2/3</italic> plants 35.4 ± 1.2 cm, and <italic>lox2/3</italic>x<italic>TPS10</italic> plants 32.2 ± 1.5 cm (mean ± SEM). In season two, final measurements of plant height on June 6th revealed only a minor difference for central plants expressing TPS10, discussed below.</p><p>We also monitored early flower production by plants as a measure of growth and reproduction rates. To assess the rate of flowering, flowers were counted once shortly after plants first began to flower, and again 1 week later; in season two, plant size and flower production was additionally assessed once immediately prior to quantification of herbivore damage. In season one, in which herbivores damaged <10% of total canopy area (<xref ref-type="fig" rid="fig5">Figure 5</xref>), <italic>LOX2/3</italic>-deficient plants produced more flowers earlier than WT and <italic>TPS10</italic> plants. Shortly after the first plants began to flower in season one (May 11th), more <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> than WT and <italic>TPS10</italic> plants were flowering (30% v. 14%), and this was still true 1 week later, after most plants had flowers (95% v. 79%; G-tests of independence, corrected p < 0.001 on May 11th and 17th; <xref ref-type="fig" rid="fig7">Figure 7A</xref>). In pairwise comparisons among individual genotypes, there were no significant differences in G-tests of independence between WT and <italic>TPS10</italic> (corrected p<italic>-</italic>values > 0.2) or <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> (corrected p<italic>-</italic>values > 0.7). However, more <italic>lox2/3</italic> than WT plants were flowering at both timepoints (May 11th corrected p < 0.001, May 17th corrected p = 0.022), and more <italic>lox2/3</italic>x<italic>TPS10</italic> than <italic>TPS10</italic> plants were flowering on May 17th (corrected p < 0.001). Both <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants also produced more flowers per plant: on May 17th <italic>lox2/3</italic> had 6 (median = mean) flowers per plant while WT had 4, and <italic>lox2/3</italic>x<italic>TPS10</italic> had 4/5 (median/mean) flowers per plant while TPS10 had 2/3.<fig id="fig7" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.019</object-id><label>Figure 7.</label><caption><title><italic>LOX2/3</italic> deficiency accelerated flowering under low herbivory.</title><p>Flower production was monitored twice per season: once shortly after plants began to flower and again 1 week later, when most or all plants were flowering. Bars indicate either total percentage of plants of each genotype which were flowering (left panels), or the mean number of flowers per plant (right panels). WT and <italic>TPS10</italic> (black and grey bars) are shown opposite <italic>lox2/3</italic> and <italic>lox2/3xTPS10</italic> (red and pink bars). Ratios in bars indicate numbers of flowering/total plants monitored; a few plants were not included in counts which had recently lost their main stem to herbivores and had not yet re-grown. (<bold>A</bold>) In season one, when plants suffered less damage from herbivores (<xref ref-type="fig" rid="fig5">Figure 5</xref>), plants with the <italic>lox2/3</italic> silencing construct produced more flowers earlier: asterisks indicate pairwise differences between genotypes differing only in the <italic>lox2/3</italic> silencing construct. ***Corrected p < 0.001, **corrected p < 0.01, *corrected p < 0.05, or no significant difference (ns) in G-tests (percentage flowering) or in Wilcoxon rank sum tests (flower number, WT vs <italic>lox2/3</italic>: May 11th, W<sub>120,120</sub> = 8582, p = 0.0002; May 17th, W<sub>120,119</sub> = 5043, p < 0.0001; <italic>TPS10</italic> vs <italic>lox2/3</italic>x<italic>TPS10</italic>: May 11th, W<sub>84,84</sub> = 3801, p = 0.2063; May 17th, W<sub>84,84</sub> = 2520, p = 0.0027). (<bold>B</bold>) In season two, during which plants received more damage from herbivores (<xref ref-type="fig" rid="fig5">Figure 5</xref>), flowering was monitored at later dates: the first timepoint in (<bold>B</bold>) is comparable to the second timepoint in (<bold>A</bold>) in terms of the proportion of plants flowering and the average number of flowers per plant (note difference in scale). In season two, <italic>LOX2/3</italic> deficiency rather decreased early flower numbers in the comparison of WT vs <italic>lox2/3</italic> on May 29th, and did not increase flower numbers at either measurement (WT vs <italic>lox2/3</italic>: May 29th, W<sub>103,95</sub> = 3860.5, p = 0.0091; June 6th, W<sub>97,69</sub> = 2940.5, p = 0.3018; <italic>TPS10</italic> vs <italic>lox2/3</italic>x<italic>TPS10</italic>: May 29th, W<sub>87,74</sub> = 3646.5, p = 0.1417; June 6th, W<sub>83,63</sub> = 2695, p = 0.7518). For a complete analysis of the effects of <italic>LOX2/3</italic> deficiency and <italic>TPS10</italic> expression, plant position, and population type on flowering and plant size, see source data file for <xref ref-type="fig" rid="fig8">Figure 8</xref>.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.019">http://dx.doi.org/10.7554/eLife.04490.019</ext-link></p></caption><graphic xlink:href="elife04490f007"/></fig></p><p>In season two, when herbivore damage ranged from 10–24% of total canopy (<xref ref-type="fig" rid="fig5">Figure 5</xref>), <italic>LOX2/3</italic>-deficient plants did not flower earlier (<xref ref-type="fig" rid="fig7">Figure 7B</xref>). Plants were in an earlier stage of flowering on May 29th in season two, than on May 17th in season one: both the proportion of flowering plants (72% May 29th season two, 87% May 17th season one) and the mean number of flowers per plant (4 on May 29th season two, 5 on May 17th season one) were lower for the May 29th observation. The differences between plants with or without the <italic>lox2/3</italic> silencing construct, which was evident on May 17th in season one, were not observed on May 29th in season two: the proportion of flowering plants was similar for all genotypes (68% for <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> v. 75% for WT and <italic>TPS10</italic>, G-test of independence, corrected p = 1), as was the average number of flowers per plant: 1/3 (median/mean) for <italic>lox2/3</italic>, 3/5 for WT, 2/4 for <italic>lox2/3</italic>x<italic>TPS10</italic>, and 2/3 for <italic>TPS10</italic>. 1 week later, 100% of plants were flowering, and the median number of flowers per plant remained similar regardless of <italic>LOX2/3</italic> deficiency (<xref ref-type="fig" rid="fig7">Figure 7B</xref>).</p><p>However, in season two, there were subtler effects on flower production which could be attributed to plant position, <italic>LOX2/3</italic> deficiency, and <italic>TPS10</italic> expression (<xref ref-type="fig" rid="fig8">Figure 8</xref>). The plant height and flowering data collected immediately prior to the assessment of herbivore damage in season two were analyzed as described in <xref ref-type="app" rid="app3">Appendix 3</xref> for effects of LOX2/3 deficiency, TPS10 expression, plant position, and population type (statistical analysis in the source data file for <xref ref-type="fig" rid="fig8">Figure 8</xref>). There were almost no differences in plant height (<xref ref-type="fig" rid="fig8">Figure 8</xref>, left panels), indicating that plant size differed little; the only significant difference was a small negative effect of <italic>TPS10</italic> expression for plants in the centers of populations, most pronounced in mixed cultures (<xref ref-type="fig" rid="fig8">Figure 8E</xref>). In contrast, flower production of individual plants was strongly decreased by <italic>TPS10</italic> expression and only slightly decreased by <italic>LOX2/3</italic> deficiency (<xref ref-type="fig" rid="fig8">Figure 8B</xref>), but differed little for plants in populations (<xref ref-type="fig" rid="fig8">Figure 8</xref>, right panels), with the exception that the shorter central plants also produced slightly fewer flowers (<xref ref-type="fig" rid="fig8">Figure 8F</xref>). Interestingly, for individual plants, there was an interactive effect of <italic>TPS10</italic> expression and <italic>LOX2/3</italic> deficiency resulting in similar flower production for <italic>lox2/3</italic>x<italic>TPS10</italic> plants and WT plants, although both <italic>lox2/3</italic> individuals and <italic>TPS10</italic> individuals produced significantly fewer flowers than WT (<xref ref-type="fig" rid="fig8">Figure 8B</xref>, statistical analysis in the source data file for <xref ref-type="fig" rid="fig8">Figure 8</xref>).<fig id="fig8" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.020</object-id><label>Figure 8.</label><caption><title><italic>TPS10</italic> expression reduced flower production under high herbivory.</title><p>Plant size as measured by stem length (left panels), and reproduction as measured by flower production (right panels); n given in <xref ref-type="supplementary-material" rid="SD8-data">Figure 8—source data 1</xref>. Data were collected on June 6th in season two from plants at all different positions in the experiment: individual plants (<bold>A</bold>–<bold>B</bold>), and plants at the edges (<bold>C</bold>–<bold>D</bold>) or at the centers (<bold>E</bold>–<bold>F</bold>) of populations. Black bars denote WT, grey bars <italic>TPS10</italic>, red bars <italic>lox2/3</italic>, and pink bars <italic>lox2/3xTPS10</italic> plants. Plant height differed little, with only a small negative effect of <italic>TPS10</italic> expression for center plants (<bold>E</bold>), which was more pronounced in mixed- than monocultures, indicating a slight competitive disadvantage for these plants. Individual plants differed much more than plants in populations in terms of flower production, with <italic>TPS10</italic>-expressing plants at a disadvantage (<bold>B</bold>); this effect was greatly reduced in populations.<sup>a,b</sup> Different letters indicate significant differences (corrected p < 0.05) in Tukey post-hoc tests on plant genotype following significant effects in a generalized linear model, or generalized linear mixed-effects model (<bold>C</bold>–<bold>D</bold>) with genotype as a factor; these p-values were corrected using the Holm-Bonferroni correction for multiple testing as the same data were also tested for effects of <italic>LOX2/3</italic> deficiency and <italic>TPS10</italic> expression (see <xref ref-type="supplementary-material" rid="SD8-data">Figure 8—source data 1</xref>). There was no difference in flower production for edge plants, but a small negative effect of <italic>TPS10</italic> expression on flower production in center plants (<bold>F</bold>) corresponding to the slight reduction in height of these plants (<bold>E</bold>). Statistical models are given in <xref ref-type="supplementary-material" rid="SD8-data">Figure 8—source data 1</xref>.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.020">http://dx.doi.org/10.7554/eLife.04490.020</ext-link></p><p><supplementary-material id="SD8-data"><object-id pub-id-type="doi">10.7554/eLife.04490.021</object-id><label>Figure 8—source data 1.</label><caption><title>The same statistical approach described in <xref ref-type="app" rid="app3">Appendix 3</xref> for <xref ref-type="fig" rid="fig5">Figure 5</xref> was used for plant size and reproduction data from May 17th in season 1, and June 6th in season two (the most complete measurements, closest to the date of herbivore damage screens, from each season).</title><p>It should be kept in mind that plants in the season one analysis were at an earlier stage of growth and flowering, for example, had about 10% as many flowers as plants in the season 2 analysis; for an analysis of flowering at similar stages, see <xref ref-type="fig" rid="fig7">Figure 7</xref>. Significant p-values and factors are given in bold, while p-values and factors marked in grey and in bold are marginal, but are required in the minimal model because their removal resulted in poorer (higher) AIC values. Other non-significant factors remained in models because they were the least insignificant factor (if only one factor remained), or because they are dependent on other factors which are significant or required.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.021">http://dx.doi.org/10.7554/eLife.04490.021</ext-link></p></caption><media xlink:href="elife04490s008.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xlink:href="elife04490f008"/></fig></p><p>Overall, these data indicate that accelerated reproduction of <italic>LOX2/3</italic>-deficient plants in season one was due to a reduced net cost for these plants of <italic>LOX2/3</italic>-mediated defenses under low herbivore damage, which disappeared under higher herbivore damage in season two (<xref ref-type="fig" rid="fig7 fig8">Figures 7, 8</xref> and associated source data file). In contrast, <italic>TPS10</italic> expression had little effect on plant growth and flowering in season one (<xref ref-type="fig" rid="fig7">Figure 7</xref>, source data file for <xref ref-type="fig" rid="fig8">Figure 8</xref>) and the effects of TPS10 expression on flowering in season two depended strongly on plant position (individual, edge or center, <xref ref-type="fig" rid="fig8">Figure 8</xref> and associated source data file). This indicates that <italic>TPS10</italic> expression did not have large, direct costs for plants in the field, but that the reproductive output of <italic>TPS10</italic> plants was sensitive to the presence and identity of neighbor plants. Especially under high herbivory in season two, there were few differences in plant size and these did not correspond to differences in herbivore damage (<xref ref-type="fig" rid="fig5 fig8">Figures 5, 8</xref>).</p></sec><sec id="s2-4"><title><italic>LOX2/3</italic> deficiency increases plant mortality, but mortality is mitigated by single <italic>TPS10</italic>-expressing plants in <italic>LOX2/3</italic>-deficient populations</title><p>We quantified plant mortality as an unambiguous measure of plant health and fitness (<xref ref-type="fig" rid="fig9">Figure 9</xref>). In our experiments, mortality can be interpreted as a measure of total reproductive potential and thus Darwinian fitness, because we observed mortality of plants before, or during, the peak of their flower production, and ended our experiments when plants would normally produce mature seed. (Production of viable seed is not generally allowed for field-released transgenic plants, and we did not allow our plants to produce seed.)<fig-group><fig id="fig9" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.022</object-id><label>Figure 9.</label><caption><title>One <italic>TPS10</italic>-expressing plant per population counteracts a mortality increase in <italic>LOX2/3</italic>-deficient populations.</title><p>Bars indicate total percentage mortality for all plants in each population type, and ratios in bars indicate numbers of dead/total plants (n = 75). WT and <italic>TPS10</italic> populations (left, black, and grey bars) are compared to the equivalent <italic>lox2/3</italic> and <italic>lox2/3xTPS10</italic> populations (right, red, and pink bars) differing only in the <italic>lox2/3</italic> silencing construct. <italic>LOX2/3</italic>-deficient plants had higher mortality in monocultures, regardless of whether they also expressed <italic>TPS10</italic> (**WT mono v. <italic>lox2/3</italic> mono, corrected p < 0.01; *<italic>TPS10</italic> mono v. <italic>lox2/3xTPS10</italic> mono, corrected p < 0.05 in G-tests; p<italic>-</italic>values were corrected for multiple testing using the Holm-Bonferroni method). However, the mortality of plants in mixed <italic>lox2/3</italic> + <italic>lox2/3xTPS10</italic> populations is similar to that of the plants in mixed WT + <italic>TPS10</italic> populations (ns, not significant). There was no significant difference in mortality among mono- and mixed cultures of WT and <italic>TPS10</italic> plants (corrected p > 0.3), or of <italic>lox2/3</italic> and <italic>lox2/3xTPS10</italic> plants, although there was a marginal difference between <italic>lox2/3</italic> monocultures and <italic>lox2/3</italic> + <italic>lox2/3xTPS10</italic> mixed cultures (corrected p = 0.091). In addition, a health index was assigned to each plant (discussed in text); this index was correlated with several plant growth and reproduction parameters (see <xref ref-type="fig" rid="fig9s1">Figure 9—figure supplement 1</xref>). Data were collected on June 29th at the end of experimental season two; in season one, all plants experienced much lower herbivory (<xref ref-type="fig" rid="fig5">Figure 5</xref>) and negligible mortality (0–5%).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.022">http://dx.doi.org/10.7554/eLife.04490.022</ext-link></p></caption><graphic xlink:href="elife04490f009"/></fig><fig id="fig9s1" position="float" specific-use="child-fig"><object-id pub-id-type="doi">10.7554/eLife.04490.023</object-id><label>Figure 9—figure supplement 1.</label><caption><title>The health index assigned to plants is strongly and significantly correlated to several measures of plant size and reproduction.</title><p>All measures of plant size and reproduction made on June 6th in season two are plotted against simultaneously assigned health indices >1 (a health index of 1 refers to dead plants). Spearman's ρ and the corresponding Holm-Bonferroni corrected p<italic>-</italic>value are shown for each correlation.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.023">http://dx.doi.org/10.7554/eLife.04490.023</ext-link></p></caption><graphic xlink:href="elife04490fs004"/></fig></fig-group></p><p>In season one, mean plant mortality was only 1%, or a total of 6 plants, by the end of experiments in mid-June. However, in season two, greater damage from herbivores (<xref ref-type="fig" rid="fig5">Figure 5</xref>) was associated with much higher mean plant mortality: 46%, or 236 plants by the end of experiments on June 29th. Overall, <italic>LOX2/3</italic>-deficient plants had higher mortality: 55% for <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> vs 38% for WT and <italic>TPS10</italic> plants (G-test of independence, corrected p = 0.002). Thus, <italic>lox2/3</italic> plants in monocultures had significantly higher mortality than WT plants in monocultures (corrected p = 0.003), and <italic>lox2/3</italic>x<italic>TPS10</italic> plants in monocultures had significantly higher mortality than <italic>TPS10</italic> plants in monocultures (corrected p = 0.022) (<xref ref-type="fig" rid="fig9">Figure 9</xref>). In contrast, mortality rates for plants in mixed cultures were similar regardless of <italic>LOX2/3</italic> deficiency (corrected p = 0.870). This was associated with marginally decreased mortality for plants in <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mixed cultures (<xref ref-type="fig" rid="fig9">Figure 9</xref>, <italic>lox2/3</italic> mono v. <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mix, corrected p = 0.092). Interestingly, mortality in WT + TPS10 mixed cultures was slightly higher than in monocultures (45% v. 37% for WT monocultures and 31% for TPS monocultures), although these differences were also not significant (corrected p-values > 0.090).</p><p>In summary, because plants in <italic>lox2/3</italic>x<italic>TPS10</italic> monocultures had higher mortality than those in <italic>TPS10</italic> monocultures, we conclude that expression of <italic>TPS10</italic> by <italic>LOX2/3</italic>-deficient plants did not compensate for the absence of <italic>LOX2/3</italic>-mediated defenses; yet surprisingly, single <italic>lox2/3</italic>x<italic>TPS10</italic> plants in <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mixed populations reduced the mortality of the entire population to the same level as in WT + <italic>TPS10</italic> mixed populations.</p></sec><sec id="s2-5"><title><italic>LOX2/3</italic>-deficient plants have reduced apparent health, unless they have <italic>TPS10</italic>-expressing neighbors</title><p>Plant health, in terms of turgor, expanded leaf area and senescence, varied noticeably under the higher levels of herbivore damage in season two (<xref ref-type="fig" rid="fig5">Figure 5</xref>), and so plants were given an apparent health rating on a scale of one (dead) or two (low) to five (high). This scale was sufficiently objective that several observers were able to independently assign the same health index number to randomly-chosen plants, after training on fewer than 10 other plants. Ratings >1 on this health scale (living plants) were strongly correlated to several growth and reproduction traits measured (Spearman's ρ values ≥0.55, corrected p<italic>-</italic>values < 0.001): number of buds, stem diameter, number of side branches, rosette diameter, stem length, and number of flowers (<xref ref-type="fig" rid="fig9s1">Figure 9—figure supplement 1</xref>).</p><p><italic>LOX2/3</italic> deficiency was associated with significantly poorer apparent health on this scale: excluding dead plants, a median health index of 3 for <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> vs 3.5 for WT and <italic>TPS10</italic> (Wilcoxon rank sum test, W<sub>153,186</sub> = 19,057.5, corrected p < 0.001). In pairwise tests between genotypes for surviving plants, WT appeared significantly healthier than <italic>lox2/3</italic>, but not <italic>lox2/3</italic>x<italic>TPS10</italic> (WT v. <italic>lox2/3</italic>, W<sub>102,82</sub> = 2722, corrected p < 0.001; WT v. <italic>lox2/3</italic>x<italic>TPS10</italic>, W<sub>102,71</sub> = 2687, corrected p = 0.059); whereas <italic>TPS10</italic> appeared significantly healthier than both <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> (<italic>TPS10</italic> v. <italic>lox2/3</italic>, W<sub>84,82</sub> = 1996.5, corrected p < 0.001; <italic>TPS10</italic> v. <italic>lox2/3</italic>x<italic>TPS10</italic>, W<sub>84,71</sub> = 1995, corrected p = 0.006). There was no difference in the apparent health of WT vs <italic>TPS10</italic> or <italic>lox2/3</italic> vs <italic>lox2/3</italic>x<italic>TPS10</italic> (corrected p<italic>-</italic>values = 1, <italic>TPS10</italic> v. WT, W<sub>84,102</sub> = 4619; <italic>lox2/3</italic> v. <italic>lox2/3</italic>x<italic>TPS10</italic>, W<sub>82,71</sub> = 2604.5).</p><p>Separate examination of surviving individual, center, and edge plants revealed that <italic>LOX2/3</italic> deficiency was only associated with poorer apparent health in edge plants (<italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> v. WT and <italic>TPS10</italic>: individuals, W<sub>21,22</sub> = 230.5, corrected p = 1; edge plants, W<sub>102,129</sub> = 9544, corrected p < 0.001; center plants, W<sub>30,35</sub> = 598.5, corrected p = 1). There were significant differences in the apparent health of edge plants in different population types: WT edge plants appeared significantly healthier than <italic>lox2/3</italic> plants at the edges of monocultures, but not healthier than <italic>lox2/3</italic> plants at the edges of <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mixed cultures (WT mono v. <italic>lox2/3</italic> mono, W<sub>42,29</sub> = 929, corrected p = 0.003; WT + <italic>TPS10</italic> mix v. <italic>lox2/3</italic> mono, W<sub>40,29</sub> = 891.5, corrected p = 0.003; WT mono v. <italic>lox2/3+lox2/3</italic>x<italic>TPS10</italic> mix, W<sub>42,34</sub> = 953, corrected p = 0.164; WT + <italic>TPS10</italic> mix v. <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mix, W<sub>40,34</sub> = 892, corrected p = 0.249). However, <italic>TPS10</italic> plants at the edges of monocultures appeared significantly healthier than <italic>lox2/3</italic> or <italic>lox2/3</italic>x<italic>TPS10</italic> edge plants in mixed- or monocultures (<italic>TPS10</italic> mono v. <italic>lox2/3</italic> mono, W<sub>47,29</sub> = 1129.5, corrected p < 0.001; <italic>TPS10</italic> mono v. <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mix, W<sub>47,34</sub> = 1209.5, corrected p = 0.001; <italic>TPS10</italic> mono v. <italic>lox2/3</italic>x<italic>TPS10</italic> mono, W<sub>47,39</sub> = 1389.5, corrected p < 0.001). No other comparisons between plants of individual population types were significant (corrected p<italic>-</italic>values > 0.08) and there was no significant association between apparent health and plant position (Kruskal–Wallis test across individual, edge, and center plants, χ<sup>2</sup><sub>2</sub> = 2.38, corrected p = 1) or plants of the same genotype in mono- vs mixed cultures (Wilcoxon rank sum tests, corrected p<italic>-</italic>values = 1).</p><p>In summary, as for foliar herbivore damage, growth, flowering, and mortality, there was no significant difference in the apparent health of plants differing only in <italic>TPS10</italic> expression. However, <italic>TPS10</italic> expression in <italic>lox2/3</italic>x<italic>TPS10</italic> plants was associated with an apparent health similar to WT plants, while <italic>lox2/3</italic> plants had a significantly lower apparent health than WT plants. Furthermore, the differences in apparent health were stronger between <italic>TPS10</italic> plants and <italic>LOX2/3</italic>-deficient plants than between WT plants and <italic>LOX2/3</italic>-deficient plants, indicating a weak, positive effect of <italic>TPS10</italic> expression on apparent plant health. Interestingly, <italic>lox2/3</italic> plants at the edges of monocultures appeared less healthy than WT edge plants, but <italic>lox2/3</italic> plants at the edges of <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mixed cultures did not, indicating a positive neighbor effect of <italic>TPS10</italic> expression on the health of plants in <italic>LOX2/3</italic>-deficient populations.</p></sec><sec id="s2-6"><title><italic>TPS10</italic>-expressing plants increase the infestation rates of their neighbors by the stem-boring weevil <italic>Trichobaris mucorea</italic></title><p>At the end of season two, we found that 35% of all surviving plants were infested with <italic>T. mucorea</italic> larvae. An adult <italic>T. mucorea</italic> is pictured in <xref ref-type="fig" rid="fig10">Figure 10A</xref>. Interestingly, <italic>TPS10</italic> expression was associated with slightly higher infestation rates: 29% of <italic>lox2/3</italic> and 34% of WT plants vs 39% of <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants. Looking at infestation rates by population type, we discovered a small difference between WT (34% infested) and <italic>TPS10</italic> plants (39% infested) which masked a much larger effect: WT plants in WT + <italic>TPS10</italic> mixed populations (55% infested) had an infestation rate more than twice that of WT plants in monocultures (19% infested, corrected p = 0.015, <xref ref-type="fig" rid="fig10">Figure 10B</xref>). We observed the same tendency in <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> populations: 33% of <italic>lox2/3</italic> plants in mixed populations with <italic>lox2/3</italic>x<italic>TPS10</italic> plants were infested, vs only 16% of <italic>lox2/3</italic> plants in monocultures; interestingly, <italic>lox2/3</italic>x<italic>TPS10</italic> plants tended to have higher infestation rates than <italic>lox2/3</italic> plants in populations (<xref ref-type="fig" rid="fig10">Figure 10C</xref>). Likely due to higher plant mortality in <italic>lox2/3</italic> populations (<xref ref-type="fig" rid="fig9">Figure 9</xref>), replicate numbers were too low for these effects to be significant (corrected p-values = 1 for the comparison of infestation rates of <italic>lox2/3</italic> plants in mono- vs mixed culture, and of <italic>lox2/3</italic> vs <italic>lox2/3</italic>x<italic>TPS10</italic> monocultures). Likely also due to low replicate numbers (n = 7–10), there were no significant differences in infestation among individuals (corrected p<italic>-</italic>values = 1 in pairwise tests), but 30% of <italic>TPS10</italic>, 38% of WT, 50% of <italic>lox2/3</italic>, and 57% of <italic>lox2/3</italic>x<italic>TPS10</italic> individuals were infested. Based on these trends, had there been a significant effect for individuals, it would have been due to <italic>LOX2/3</italic> deficiency and not <italic>TPS10</italic> expression.<fig id="fig10" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.024</object-id><label>Figure 10.</label><caption><title><italic>TPS10</italic> plants increase the infestation rates of their WT neighbors with the stem-boring weevil <italic>T. mucorea.</italic></title><p>(<bold>A</bold>) Photograph of a <italic>T. mucorea</italic> adult emerging from the stem which it infested as a larva reprinted with permission from Anke Steppuhn, Copyright 2005. All rights reserved. Adults mate and oviposit on young <italic>N. attenuata</italic> plants in March and April, and upon hatching, larvae burrow into the growing stem of the plant chosen by their mother (<xref ref-type="bibr" rid="bib22">Diezel et al., 2011</xref>). Infestation was scored as the number of plants with hollow, frass-filled stems (from which larvae were usually also recovered). (<bold>B</bold>) <italic>T. mucorea</italic> infestation more than doubled for WT plants, but did not change for <italic>TPS10</italic> plants in mixed cultures vs monocultures. Bars indicate total percentage of plants infested; ratios in bars indicate numbers of infested/total plants. Data were collected at the end of experimental season two (June 29th). *Corrected p < 0.05 in a Fisher's exact test, n = 31–47 (all surviving replicates). p<italic>-</italic>values were corrected for multiple testing using the Holm-Bonferroni method. (<bold>C</bold>) The presence of <italic>lox2/3xTPS10</italic> plants also tended to increase <italic>T. mucorea</italic> infestation in <italic>lox2/3</italic> + <italic>lox2/3xTPS10</italic> populations but, likely due to lower replicate numbers caused by higher mortality for plants in <italic>lox2/3</italic> and <italic>lox2/3xTPS10</italic> monocultures (<xref ref-type="fig" rid="fig9">Figure 9</xref>), differences are not significant (corrected p<italic>-</italic>values = 1).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.024">http://dx.doi.org/10.7554/eLife.04490.024</ext-link></p></caption><graphic xlink:href="elife04490f010"/></fig></p><p>Thus, although <italic>TPS10</italic> expression did not significantly alter plants' rates of infestation by <italic>T. mucorea</italic>, it did significantly increase the infestation of neighboring plants.</p></sec><sec id="s2-7"><title><italic>TPS10</italic> expression and <italic>LOX2/3</italic> deficiency interact to alter the distribution of <italic>Geocoris</italic> spp<italic>.</italic> activity in populations</title><p>Generalist <italic>Geocoris</italic> spp<italic>.</italic> predators co-occur with <italic>N. attenuata</italic> plants, and prey on their herbivores in response to specific HIPVs including GLVs and TAB, in an example of indirect defense (<xref ref-type="bibr" rid="bib50">Kessler and Baldwin, 2001</xref>; <xref ref-type="bibr" rid="bib4">Allmann and Baldwin, 2010</xref>; <xref ref-type="bibr" rid="bib73">Schuman et al., 2012</xref>, <xref ref-type="bibr" rid="bib75">2013</xref>). Near the end of experiments in season two, <italic>Geocoris pallens</italic> were abundant and <italic>Geocoris punctipes</italic> were also present at the field site. We considered that plants in populations with <italic>TPS10</italic>-expressing neighbors might either benefit from enhanced attraction of <italic>Geocoris</italic> spp<italic>.</italic>, or suffer in competition for the predation services of <italic>Geocoris</italic> spp<italic>.</italic> We also reasoned that the lack of all other HIPVs in <italic>LOX2/3</italic>-deficient plants could alter the strength and direction of such a relationship.</p><p>To test the overarching hypothesis that <italic>TPS10</italic>-expressing plants in populations alter the distribution of predation by <italic>Geocoris</italic> spp<italic>.</italic>, we conducted predation activity assays from June 21st to June 27th on edge and center plants that were similar in size and apparent health. We treated LOX2/3-expressing (WT, TPS10) and <italic>LOX2/3</italic>-deficient (<italic>lox2/3</italic>, <italic>lox2/3</italic>x<italic>TPS10</italic>) populations as two separate experimental groups because it was not possible to match plants across these two groups for apparent health or replicate numbers (both were lower for <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> populations). Selected <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> populations had a mode of 3 remaining plants, range 2–5; and selected WT and <italic>TPS10</italic> populations had a mode of 5 remaining plants, range 3–5; within these two groups, there was no significant difference between mono- and mixed cultures in number of plants remaining (G-tests, corrected p > 0.1). Plants were baited with <italic>M. sexta</italic> eggs and larvae (<xref ref-type="fig" rid="fig11">Figure 11A</xref>), and we counted the numbers predated by <italic>Geocoris</italic> spp<italic>.</italic> over 24 h (source data file for <xref ref-type="fig" rid="fig11">Figure 11</xref>, <xref ref-type="bibr" rid="bib50">Kessler and Baldwin, 2001</xref>).<fig id="fig11" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.025</object-id><label>Figure 11.</label><caption><title><italic>TPS10</italic> and <italic>LOX2/3</italic> interact to alter the predation of <italic>M. sexta</italic> by <italic>Geocoris</italic> spp from edge vs center plants.</title><p>(<bold>A</bold>) Photograph of a <italic>M. sexta</italic> egg and first-instar larva, and <italic>G. pallens</italic> adult on an <italic>N. attenuata</italic> leaf reprinted with permission from Danny Kessler, Copyright 2006. All rights reserved. As shown here, one first-instar <italic>M</italic>. <italic>sexta</italic> larva and one egg were placed on a lower stem leaf in a standardized position on healthy, size-matched plants (one leaf/plant), and their predation was monitored for 24 h. Data are total numbers from five consecutive trials conducted between June 21st and June 27th at the end of experimental season two, when <italic>Geocoris</italic> spp. and herbivores were more abundant (see <xref ref-type="fig" rid="fig5">Figure 5</xref>). (<bold>B</bold> and <bold>C</bold>) Predation rates of <italic>M</italic>. <italic>sexta</italic> are higher on <italic>lox2/3xTPS10</italic> than <italic>lox2/3</italic> plants at the edges of monocultures (<bold>B</bold>), but lower on <italic>TPS10</italic> than WT plants at the centers of monocultures (<bold>C</bold>). Bars indicate the total percentage of <italic>M</italic>. <italic>sexta</italic> eggs and larvae predated per population; numbers of predated/total eggs and larvae are given inside bars. The same trends were observed in both egg and larva predation; separate numbers of larvae and eggs predated are given in <xref ref-type="supplementary-material" rid="SD9-data">Figure 11—source data 1</xref>. Plants of the WT background and plants with the <italic>lox2/3</italic> silencing construct were separately matched for parallel experiments and thus analyzed separately. (<bold>B</bold>) Predation of herbivores from edge plants was similar (not significant, ns) regardless of the number of <italic>TPS10</italic> plants in WT + <italic>TPS10</italic> populations, but increased with increasing numbers of <italic>lox2/3xTPS10</italic> plants in <italic>lox2/3</italic> + <italic>lox2/3xTPS10</italic> populations. For plants at the center of populations, the pattern was reversed: (<bold>C</bold>) predation rates decreased with increasing numbers of <italic>TPS10</italic> plants in WT + <italic>TPS10</italic> populations, and were similar (ns) regardless of the number of <italic>lox2/3xTPS10</italic> plants in <italic>lox2/3</italic> + <italic>lox2/3xTPS10</italic> populations. <sup>a,b</sup> Different letters indicate significant differences (corrected p < 0.05) in Fisher's exact tests; p<italic>-</italic>values were corrected for multiple testing using the Holm-Bonferroni method.</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.025">http://dx.doi.org/10.7554/eLife.04490.025</ext-link></p><p><supplementary-material id="SD9-data"><object-id pub-id-type="doi">10.7554/eLife.04490.026</object-id><label>Figure 11—source data 1.</label><caption><title>Total numbers of <italic>M. sexta</italic> larvae and eggs predated by <italic>Geocoris</italic> spp. in experimental season two (corresponding to data shown in <xref ref-type="fig" rid="fig10">Figure 10</xref>).</title><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.026">http://dx.doi.org/10.7554/eLife.04490.026</ext-link></p></caption><media xlink:href="elife04490s009.xlsx" mimetype="application" mime-subtype="xlsx"/></supplementary-material></p></caption><graphic xlink:href="elife04490f011"/></fig></p><p><italic>TPS10</italic>-expressing plants significantly altered predation rates at the edges and centers of populations, and whether the effect was at the edge or the center depended on <italic>LOX2/3</italic> expression (<xref ref-type="fig" rid="fig11">Figure 11B,C</xref>). Specifically, <italic>TPS10</italic> expression increased predation from edge plants in <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> populations (Fisher's exact tests, corrected p<italic>-</italic>values < 0.05) but not in WT and <italic>TPS10</italic> populations (21–34%, corrected p<italic>-</italic>values > 0.3) (<xref ref-type="fig" rid="fig11">Figure 11B</xref>). At the edges of monocultures, <italic>lox2/3</italic>x<italic>TPS10</italic> plants received 43% predation whereas <italic>lox2/3</italic> plants received only 16% (corrected p = 0.049), and <italic>lox2/3</italic> plants at the edges of <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mixed cultures were intermediate with 23% (<italic>lox2/3</italic> mono v. <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mixed, corrected p = 0.723, <italic>lox2/3</italic>x<italic>TPS10</italic> mono v. <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> mixed, corrected p = 0.457). On the other hand, <italic>TPS10</italic> expression decreased, rather than increased predation from center plants in WT + <italic>TPS10</italic> populations (from 46% to 12%, corrected p < 0.04) and had no effect on center plants in <italic>LOX2/3</italic>-deficient populations (ranging from 13% to 23%, corrected p<italic>-</italic>values = 1, <xref ref-type="fig" rid="fig11">Figure 11C</xref>). WT plants received 46% predation at the center of monocultures, while <italic>TPS10</italic> plants received 38% predation at the center of mixed cultures and only 12% at the center of monocultures (WT mono v. WT + <italic>TPS10</italic> mixed, corrected p = 0.618; WT mono v. <italic>TPS10</italic> mono, corrected p = 0.012; WT + <italic>TPS10</italic> mixed v. <italic>TPS10</italic> mono, corrected p = 0.031).</p><p>Thus in populations of <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants deficient in all HIPVs except for the engineered TPS10 volatiles, <italic>TPS10</italic> expression by edge plants increased predator activity on those plants. In contrast, <italic>TPS10</italic> expression by center plants failed to attract more predator activity to the centers of populations. In populations of WT and <italic>TPS10</italic> plants with intact WT HIPVs, <italic>TPS10</italic> expression did not increase predator activity at all; however, <italic>TPS10</italic> expression by edge plants seemed to reduce the penetration of predators to center plants. Thus, depending the richness of the plants' HIPV blends, <italic>TPS10</italic> expression had either positive direct effects on predator activity for emitting plants, or negative indirect effects for their neighbors.</p></sec></sec><sec sec-type="discussion" id="s3"><title>Discussion</title><p>In this study, we sought to fulfil the four postulates of community genetics (<xref ref-type="bibr" rid="bib85">Wymore et al., 2011</xref>; <xref ref-type="bibr" rid="bib84">Whitham et al., 2012</xref>) to uncover the community effects of genes controlling direct and indirect defenses in the wild tobacco <italic>N. attenuata</italic>. <italic>N. attenuata</italic> has already been shown to fulfill postulate one: significantly affecting an ecological community of other plants and animals, and native microbes (reviewed in <xref ref-type="bibr" rid="bib73">Schuman and Baldwin, 2012</xref>; see also <xref ref-type="bibr" rid="bib56">Long et al., 2010</xref>; <xref ref-type="bibr" rid="bib59">Meldau et al., 2012</xref>; <xref ref-type="bibr" rid="bib58">Machado et al., 2013</xref>; <xref ref-type="bibr" rid="bib68">Santhanam et al., 2014</xref>). Postulate two, demonstration of genetically based traits, is ensured by our method of altering the expression of specific genes: <italic>LOX2</italic>, <italic>LOX3</italic>, and <italic>TPS10</italic> (<xref ref-type="fig" rid="fig1">Figure 1</xref> and associated source data files, <xref ref-type="app" rid="app1 app2">Appendices 1, 2</xref>). Thus, we created transgenic lines from a single genotype of <italic>N. attenuata</italic> and used them to investigate both individual- and population-level effects of two HIPVs, (<italic>E</italic>)-α-bergamotene (TAB) and (<italic>E</italic>)-β-farnesene (TBF), produced by the sesquiterpene synthase TPS10, in plants which either had wild-type direct defenses and HIPVs, or reduced production of direct defenses and other HIPVs. We used these plants to evaluate postulate three: different effects on community processes of conspecifics differing in <italic>LOX2/3</italic> and <italic>TPS10</italic> expression, and postulate four: predictable effects on those community processes when the genes are manipulated.</p><p>Specifically, we found that <italic>LOX2/3</italic>-mediated defense had a strong and direct effect on plant resistance to foliar herbivores and plant mortality (<xref ref-type="fig" rid="fig4 fig5">Figures 4, 5</xref> and associated source data files, <xref ref-type="fig" rid="fig9">Figure 9</xref>), as predicted (<xref ref-type="bibr" rid="bib52">Kessler et al., 2004</xref>). Interestingly, the effect of <italic>LOX2/3</italic> expression on herbivore abundance was less clear: <italic>LOX2/3</italic> deficiency was associated with significantly increased herbivore abundance for edge plants in populations, but not overall (<xref ref-type="fig" rid="fig6">Figure 6</xref>). Furthermore, although <italic>TPS10</italic> expression did not significantly affect foliar herbivore damage or performance, it did consistently reduce foliar herbivore abundance (<xref ref-type="fig" rid="fig6">Figure 6</xref>). It should be noted that the herbivore damage data in <xref ref-type="fig" rid="fig5">Figure 5B</xref> were collected just over 2 weeks earlier than the herbivore abundance data in <xref ref-type="fig" rid="fig6">Figure 6</xref>; however, the relative amounts of damage on plants at the time that herbivore abundance data were collected still reflected the patterns shown in <xref ref-type="fig" rid="fig5">Figure 5</xref>, with LOX2/3-deficient genotypes having visibly more damage and WT and <italic>TPS10</italic> plants not differing visibly in herbivore damage levels.</p><p>More surprisingly, we found population-level effects of <italic>TPS10</italic> volatiles which were as great as the individual-level effects of total <italic>LOX2/3</italic>-mediated defenses, and dependent on <italic>LOX2/3</italic> expression. Single <italic>TPS10</italic>-expressing plants reduced the mortality of plants in <italic>LOX2/3</italic>-deficient populations (<xref ref-type="fig" rid="fig9">Figure 9</xref>), corresponding to an increase in apparent health of those plants. Additionally, <italic>TPS10</italic> expression significantly increased the infestation rate of neighbors by the stem-boring weevil <italic>T. mucorea</italic> (<xref ref-type="fig" rid="fig10">Figure 10</xref>), and <italic>TPS10</italic> expression in a population strongly altered the distribution of predation services among edge and center plants, depending on <italic>LOX2/3</italic>-mediated defense (<xref ref-type="fig" rid="fig11">Figure 11</xref> and associated source data file).</p><p>It should be noted that our experimental design did not support the analysis of population-level effects resulting from changing <italic>LOX2</italic> and <italic>LOX3</italic> expression in single members of a population. Had we included these treatment groups as well (e.g., populations with a central <italic>lox2/3</italic> plant surrounded by WT, and with a central <italic>lox2/3</italic>x<italic>TPS10</italic> plant surrounded by <italic>TPS10</italic>), we predict that we would also have observed significant neighbor effects of LOX2/3 expression in populations. However, given the magnitude of direct effects of <italic>LOX2/3</italic> expression, it would be surprising if the indirect population-level effects from manipulating <italic>LOX2/3</italic> expression in single plants were larger than the direct effects—as was generally observed from the manipulation of <italic>TPS10</italic> expression. Our experiment was designed to determine whether TPS10 products have direct and neighbor effects in the presence, or absence, of total HIPVs and jasmonate-mediated defense.</p><p>We found no effects on non-target metabolites, (<xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref>, <xref ref-type="app" rid="app2">Appendix 2</xref>), and only minor effects on plant size and growth which were all associated with the alleviation of the costs of <italic>LOX2/3</italic>-mediated defense for single plants, benefits which disappeared under high herbivory loads (<xref ref-type="fig" rid="fig7 fig8">Figures 7, 8</xref> and associated source data file, <xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>).</p><sec id="s3-1"><title>Plant productivity depends on herbivore damage rates, LOX2/3 and TPS10 expression</title><p>It is well known that jasmonate-mediated defenses can be costly in the absence of herbivores (reviewed in e.g., <xref ref-type="bibr" rid="bib77">Steppuhn and Baldwin, 2008</xref>), and under low levels of damage from herbivores in season one (<xref ref-type="fig" rid="fig5">Figure 5A</xref>), growth and early flower production was greater in <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> than in WT and <italic>TPS10</italic> plants (<xref ref-type="fig" rid="fig7 fig8">Figures 7, 8</xref> and associated source data file). These data indicate that our measurements were sufficiently sensitive to detect developmental costs of defense metabolite production. Attempts to engineer the emission of terpene volatiles have sometimes resulted in plants with stunted growth or symptoms of autotoxicity (<xref ref-type="bibr" rid="bib67">Robert et al., 2013</xref>; reviewed in <xref ref-type="bibr" rid="bib25">Dudareva and Pichersky, 2008</xref>; <xref ref-type="bibr" rid="bib24">Dudareva et al., 2013</xref>). However, we did not detect any consistent costs of <italic>TPS10</italic> overexpression: <italic>TPS10</italic> plants were generally not smaller than WT (<xref ref-type="fig" rid="fig8">Figure 8</xref> and associated source data file). The same line of TPS10 also grew similarly to WT in a greenhouse competition study (line 10–3 in <xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>).</p><p>However, there were effects of <italic>TPS10</italic> expression on flowering which differed depending on plants' position in different populations. In particular, under high herbivory in season two, <italic>TPS10</italic> expression strongly decreased flower production for well-defended individual plants, but promoted flower production in <italic>LOX2/3</italic>-deficient individuals (<xref ref-type="fig" rid="fig8">Figure 8B</xref>). In contrast, flower production was not affected by <italic>TPS10</italic> expression for edge plants in populations, and was slightly reduced for <italic>TPS10</italic>-expressing plants in the centers of populations (<xref ref-type="fig" rid="fig8">Figure 8D,F</xref>). For central plants, in contrast to individuals, the reduction in flower production due to <italic>TPS10</italic> expression was similar both for <italic>LOX2/3</italic>-expressing and <italic>LOX2/3</italic>-deficient plants. Overall, these data indicate that <italic>TPS10</italic> expression may have altered the relative attractiveness or apparency of <italic>TPS10</italic> plants, with results depending strongly on plants' neighbors. The potential mechanisms in terms of herbivore abundance and predation rates, and consequences in terms of health and mortality, are discussed further below.</p><p>Higher levels of damage from foliar herbivores in season two (<xref ref-type="fig" rid="fig5">Figure 5B</xref>) were likely the main factor which shifted <italic>LOX2/3</italic>-mediated defenses from costly, in terms of early flower production (<xref ref-type="fig" rid="fig7">Figure 7</xref>), to beneficial, in terms of reduced plant mortality before and during the reproductive phase, and increased apparent plant health (<xref ref-type="fig" rid="fig9">Figure 9</xref> and <xref ref-type="fig" rid="fig9s1">Figure 9—figure supplement 1</xref>). Overall, there were few differences in growth and size among genotypes (<xref ref-type="fig" rid="fig8">Figure 8</xref> and associated source data file, and results reported in text), making it reasonable to attribute changes in herbivore damage, mortality, and other interactions to the manipulated chemical traits, rather than plant apparency or attractiveness based on size and development.</p></sec><sec id="s3-2"><title><italic>LOX2/3</italic> expression has large direct effects on plants, while <italic>TPS10</italic> expression has large indirect effects</title><p><italic>LOX2/3</italic> deficiency significantly increased damage from foliar herbivores and growth of the specialist <italic>M. sexta</italic>, whereas <italic>TPS10</italic> expression had no effect on either interaction (<xref ref-type="fig" rid="fig6 fig7">Figures 6, 7</xref>). There was, furthermore, no effect of plant population type (mixed or monoculture) on damage from foliar herbivores in field experiments (<xref ref-type="fig" rid="fig5">Figure 5</xref> and associated source data files). Together, these data indicate that feeding of <italic>M. sexta</italic> larvae and other naturally occurring foliar herbivores are not directly affected by <italic>TPS10</italic> volatiles. However, <italic>TPS10</italic> volatiles may have altered the attraction of herbivores to plants: TPS10-expressing plants had consistently smaller populations of foliar herbivores by the end of season two, despite the fact that these plants tended to look healthier, and independently of LOX2/3 expression (<xref ref-type="fig" rid="fig6">Figure 6</xref>). In contrast, ovipositing <italic>T. mucorea</italic> mothers may be attracted to <italic>TPS10</italic> volatiles (<xref ref-type="fig" rid="fig10">Figure 10</xref>). Interestingly, <italic>M. sexta</italic> moths laid fewer eggs on plants for which the headspace had been supplemented with TAB (<xref ref-type="bibr" rid="bib50">Kessler and Baldwin, 2001</xref>). A direct effect on colonizing herbivores and ovipositing females may result in indirect effects in terms of associational resistance or associational susceptibility for neighboring plants (<xref ref-type="bibr" rid="bib50">Kessler and Baldwin, 2001</xref>; <xref ref-type="bibr" rid="bib10">Barbosa et al., 2009</xref>). Although such indirect effects were not detected in foliar herbivore abundance on neighboring WT and <italic>lox2/3</italic> plants (<xref ref-type="fig" rid="fig6">Figure 6C</xref>), <italic>TPS10</italic> expression did significantly affect neighbors' mortality (<xref ref-type="fig" rid="fig9">Figure 9</xref>), apparent health, infestation by <italic>T. mucorea</italic> larvae (<xref ref-type="fig" rid="fig10">Figure 10</xref>), and predation services (<xref ref-type="fig" rid="fig11">Figure 11</xref>).</p><p>The most straightforward indirect effects of <italic>TPS10</italic> expression were evident in open headspace trappings (<xref ref-type="fig" rid="fig3">Figure 3</xref>): the presence of a <italic>lox2/3</italic>x<italic>TPS10</italic> plant in the center of a population was associated with TAB in the headspace of neighboring <italic>lox2/3</italic> plants, and TAB was not found in the open headspaces of <italic>lox2/3</italic> individuals or monocultures. The most parsimonious explanations are diffusion of volatiles into neighbors' headspace, or adherence and re-release of volatiles from the surface of neighbor plants, as has been shown in birch (<italic>Betula</italic> spp.) by <xref ref-type="bibr" rid="bib40">Himanen et al. (2010)</xref>. Because no TAB was detectable in extracts from the leaves of <italic>lox2/3</italic> plants even when plants were much larger and more likely to have direct contact with <italic>TPS10</italic> neighbors, it is unlikely that the TAB detected in open trappings of <italic>lox2/3</italic> headspaces came from the stimulation of TAB biosynthesis in <italic>lox2/3</italic> (<xref ref-type="fig" rid="fig1">Figure 1C</xref>, <xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref>).</p><p>Indirect effects of <italic>TPS10</italic> volatiles on neighboring plants may be attributed to the volatile compounds themselves, because we found no evidence that <italic>TPS10</italic> volatiles affected the defenses of emitting plants, or of neighbor plants in populations (<xref ref-type="app" rid="app1 app2">Appendices 1, 2</xref>, <xref ref-type="fig" rid="fig1">Figure 1</xref> and associated source data files); and <italic>TPS10</italic> plants also did not alter the defense response of neighbors in a glasshouse competition experiment (<xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>). In contrast, due to the large suite of defenses regulated by <italic>LOX2</italic> and <italic>LOX3</italic> products, it would have been more difficult to determine the proximate cause of neighbor effects resulting from manipulation of <italic>LOX2</italic> and <italic>LOX3</italic> expression. Although several components of plant HIPV blends have been shown to prime the direct and indirect defenses of neighbors in different systems (reviewed in <xref ref-type="bibr" rid="bib8">Baldwin et al., 2006</xref>; <xref ref-type="bibr" rid="bib7">Arimura et al., 2010</xref>; see <xref ref-type="bibr" rid="bib38">Heil and Kost, 2006</xref>), GLVs may often be the active cues (<xref ref-type="bibr" rid="bib27">Engelberth et al., 2004</xref>; <xref ref-type="bibr" rid="bib51">Kessler et al., 2006</xref>; <xref ref-type="bibr" rid="bib29">Frost et al., 2008</xref>). We did not mix GLV-deficient with GLV-emitting plants in our experimental design, but we can conclude that <italic>TPS10</italic> volatiles had no detectable direct effects on neighbor plants in our set-up.</p></sec><sec id="s3-3"><title>Indirect effects of <italic>TPS10</italic> HIPVs may benefit or harm emitters, depending on the emission and <italic>LOX2/3</italic>-mediated defense levels of their neighbors</title><p>We observed multiple significant indirect effects of <italic>TPS10</italic> expression in populations, as mentioned above; and two of these were strongest in mixed cultures containing only one <italic>TPS10</italic>-expressing plant. The mortality of <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants in monocultures was significantly higher than that of WT and <italic>TPS10</italic> plants in monocultures, but the mortality of <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> plants in mixed cultures was not higher than that of WT + <italic>TPS10</italic> plants in mixed cultures (<xref ref-type="fig" rid="fig9">Figure 9</xref>). Thus, a single <italic>TPS10</italic> emitter in populations of otherwise defenseless plants reduced overall mortality to levels comparable to well-defended populations containing single <italic>TPS10</italic> emitters. This apparent protective effect of <italic>TPS10</italic> volatiles was weaker for monocultures of <italic>lox2/3</italic>x<italic>TPS10</italic> plants. This may be because <italic>TPS10</italic> volatile emission makes plants more apparent to herbivores like <italic>T. mucorea</italic> as well as beneficial insects like <italic>Geocoris</italic> spp<italic>.</italic> (<xref ref-type="fig" rid="fig9 fig10">Figures 9, 10</xref>), and if plants are otherwise poorly defended, greater apparency may be dangerous (<xref ref-type="bibr" rid="bib28">Feeny, 1976</xref>). Interestingly, although there were no statistically significant differences in mortality among populations of WT and <italic>TPS10</italic> plants, mortality was lowest in <italic>TPS10</italic> monocultures; and WT + <italic>TPS10</italic> mixed cultures had mortality rates 1.5-fold that of <italic>TPS10</italic> monocultures (<xref ref-type="fig" rid="fig9">Figure 9</xref>). This indicates that well-defended <italic>TPS10</italic> plants may have benefited directly from emitting <italic>TPS10</italic> volatiles, and may have additionally increased the mortality of non-emitting neighbors. If so, <italic>TPS10</italic> volatiles may increase plants' competitive ability by increasing neighbor mortality.</p><p>Indeed, <italic>TPS10</italic> plants had a dramatic effect on the infestation of their WT neighbors by <italic>T</italic>. <italic>mucorea</italic> weevils (<xref ref-type="fig" rid="fig10">Figure 10</xref>). <italic>TPS10</italic> volatiles seemed only to affect the infestation rates of plant populations; for individual plants, jasmonate-mediated defense is likely more important (<xref ref-type="bibr" rid="bib22">Diezel et al., 2011</xref>, and see ‘Results’). Infestation patterns in populations (<xref ref-type="fig" rid="fig10">Figure 10</xref>) strongly indicate that ovipositing <italic>T</italic>. <italic>mucorea</italic> adults seek populations of plants based on their volatile emission and then oviposit on all plants in the population. Infestation rates did tend to be higher for <italic>lox2/3</italic>x<italic>TPS10</italic> plants than for <italic>lox2/3</italic> plants in populations (<xref ref-type="fig" rid="fig10">Figure 10C</xref>), indicating that the absence of other volatiles such as GLVs may reduce plants' apparency to ovipositing <italic>T. mucorea</italic>.</p><p>It is interesting that the <italic>TPS10</italic>-mediated increase in <italic>T</italic>. <italic>mucorea</italic> infestation was greater for WT plants in mixed WT + <italic>TPS10</italic> populations than for <italic>TPS10</italic> plants in monocultures (<xref ref-type="fig" rid="fig10">Figure 10</xref>): though <italic>TPS10</italic> monocultures tended to have higher infestation rates than WT monocultures, this difference was not statistically significant, whereas the difference between WT plants in monocultures and WT plants in mixed cultures was both larger and significant. It is tempting to speculate that <italic>T</italic>. <italic>mucorea</italic> may be attracted to lower relative abundance of <italic>TPS10</italic> volatiles, and deterred by higher relative abundance of these volatiles in the total mixture of the plant headspace. If true, this could provide a mechanism by which <italic>TPS10</italic> emitting plants could inflict an indirect cost of emission on non-emitting neighbors, which would be always slightly higher than their own cost of emission. Furthermore, if attraction or deterrence depended on the ratio of <italic>TPS10</italic> volatiles to other plant volatiles, the effects might differ for <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants (<xref ref-type="bibr" rid="bib13">Bruce et al., 2005</xref>).</p><p>The infliction of such an indirect cost on neighbors is reminiscent of extortionist strategies in the iterated prisoner's dilemma of evolutionary game theory: extortionist strategies lock a player's payoff to its opponent's payoff plus some positive number, thereby ensuring that the extortionist always benefits compared to its opponent (<xref ref-type="bibr" rid="bib64">Press and Dyson, 2012</xref>). Interestingly, such strategies have been shown in simulations not to be evolutionarily stable, but to catalyze the evolution of cooperation in populations (<xref ref-type="bibr" rid="bib39">Hilbe et al., 2013</xref>). Of course it is possible that the manipulation of <italic>LOX2</italic> and <italic>LOX3</italic> expression within populations would also reveal such extortionist-like effects, for example if <italic>T. mucorea</italic> preferred to oviposit in <italic>TPS10</italic> plants over <italic>lox2/3</italic>x<italic>TPS10</italic> plants in the same population, which would support the hypothesis that <italic>T. mucorea</italic> preference reflects the ratio between <italic>TPS10</italic> volatiles and other HIPVs.</p></sec><sec id="s3-4"><title>Effects of <italic>TPS10</italic> volatiles on indirect defense depend on <italic>LOX2/3</italic>-mediated direct and indirect defense</title><p>Predation rates of <italic>M. sexta</italic> herbivores from edge and center plants in populations (<xref ref-type="fig" rid="fig11">Figure 11</xref>) indicate that populations' total HIPV emissions can alter the roles of specific volatiles for indirect defense. In <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> populations, deficient in all HIPVs except those provided by <italic>TPS10</italic> expression, <italic>lox2/3</italic>x<italic>TPS10</italic> edge plants of monocultures experienced higher predation rates, whereas predation rates were similar for center <italic>lox2/3</italic>x<italic>TPS10</italic> and <italic>lox2/3</italic> plants in mono- and mixed cultures. For <italic>TPS10</italic> and WT plants, effects on predation rates were exactly the opposite. There was significantly less predation of <italic>M. sexta</italic> from <italic>TPS10</italic> center plants in monocultures vs mixed cultures, but <italic>TPS10</italic> edge plants had predation rates similar to WT. Together, these results indicate that the effect of <italic>TPS10</italic> volatiles on indirect defense was positive for <italic>lox2/3</italic>x<italic>TPS10</italic> plants, but negative for <italic>TPS10</italic> plants, and only significant for plants in monocultures. In mixed cultures, <italic>TPS10</italic> volatiles had no effect on the predation of <italic>M. sexta</italic> herbivores from emitting plants or their neighbors. It should be noted that due to higher mortality in <italic>lox2/3</italic> + <italic>lox2/3</italic>x<italic>TPS10</italic> populations, center plants were sometimes more exposed in these populations as a result of edge plants dying. However, the fact that <italic>TPS10</italic> volatiles benefitted only edge plants in these populations indicate that <italic>Geocoris</italic> spp<italic>.</italic> predators nevertheless distinguished between edge and center plants.</p><p>We conclude that <italic>TPS10</italic> expression had different consequences for plants' indirect defense, depending on whether one or all plants in a population expressed <italic>TPS10</italic>, and whether those plants also emitted other HIPVs. If only one plant in the population expressed <italic>TPS10</italic>, or if other HIPVs were also emitted in the population, <italic>TPS10</italic> expression did not enhance indirect defense and was sometimes detrimental to emitters. The attraction of most, or all, predators and parasitoids to HIPVs depends on learned associations between HIPVs and prey: naïve predators are just as likely to respond as they are not to respond (<xref ref-type="bibr" rid="bib3">Allison and Hare, 2009</xref>). Because <italic>LOX2/3</italic>-deficient plants were heavily attacked by herbivores (<xref ref-type="fig" rid="fig5">Figure 5</xref>), predators could associate <italic>TPS10</italic> volatiles emitted by <italic>lox2/3</italic>x<italic>TPS10</italic> plants with herbivore damage (though damage did not necessarily reflect abundance—see <xref ref-type="fig" rid="fig6">Figure 6</xref>). Because <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants had reduced emissions of other HIPVs (<xref ref-type="fig" rid="fig1">Figure 1</xref> and associated source data files, <xref ref-type="app" rid="app2">Appendix 2</xref>), there were few other volatiles which could provide the basis for learned associations between HIPVs and prey. On the other hand, TPS10 volatiles emitted constitutively by less heavily attacked <italic>TPS10</italic> plants (<xref ref-type="fig" rid="fig5">Figure 5</xref>) would be less often associated with feeding herbivores than the endogenous HIPVs released only when herbivores attack, and all the more so if <italic>TPS10</italic>-expressing plants supported fewer foliar herbivores (<xref ref-type="fig" rid="fig6">Figure 6</xref>). Perhaps because ‘honest’ endogenous HIPVs were also emitted by WT plants, <italic>TPS10</italic> plants had no advantage over WT in attracting predators.</p><p>The concept of honest vs deceptive, or ‘cry wolf’ HIPV emission has been investigated in the context of emission proportionate, or disproportionate to numbers of attacking herbivores (<xref ref-type="bibr" rid="bib76">Shiojiri et al., 2010</xref>). But what happens when plants ‘cry wolf’ by constitutively emitting volatiles which, in nature, are only associated with herbivore damage? Plants engineered to constitutively emit particular HIPVs can attract more parasitoids when parasitoids are trained to associate engineered volatiles with their prey (<xref ref-type="bibr" rid="bib49">Kappers et al., 2005</xref>; <xref ref-type="bibr" rid="bib70">Schnee et al., 2006</xref>), or when all emitting plants are also experimentally infested (<xref ref-type="bibr" rid="bib66">Rasmann et al., 2005</xref>; <xref ref-type="bibr" rid="bib20">Degenhardt et al., 2009</xref>). Such studies demonstrate that genetic engineering of HIPVs can enhance indirect defense, but they do not demonstrate that constitutive emission is an effective strategy in the complex environments of agriculture, or nature (<xref ref-type="bibr" rid="bib54">Kos et al., 2009</xref>; <xref ref-type="bibr" rid="bib73">Schuman et al., 2012</xref>; <xref ref-type="bibr" rid="bib86">Xiao et al., 2012</xref>). From an arthropod's perspective, constitutive HIPV emission is oxymoronic. If there is no effort to artificially associate constitutively emitted volatiles with prey, predators, and parasitoids may not learn to respond, or worse, learn to actively ignore the volatile emissions. Whether achieved via genetic methods or synthetic release stations (<xref ref-type="bibr" rid="bib48">Kaplan, 2012</xref>), constitutive emission is unlikely to enhance indirect defense if prey densities are low, or if plants emit other ‘honest’ HIPVs.</p><p>Moreover, assigning an indirect defense function to HIPVs based on their ability to attract predators or parasitoids to single plants may be misleading. In real populations, the effectiveness of HIPVs likely depends on the density of HIPV emission and the location of emitting plants (<xref ref-type="bibr" rid="bib86">Xiao et al., 2012</xref>). There may be competition among plants in populations either to attract, or to arrest predators and parasitoids of herbivores. Specifically, whether plants benefit from emitting more or less than their neighbors likely depends on the basal level of HIPV emission in a population (low or high), and the location of plants in a population (edge or center). Perhaps, wild-type plants are capable of adjusting their own volatile emission in response to the emission of their neighbors in order to adapt (<xref ref-type="bibr" rid="bib16">Choh et al., 2004</xref>; <xref ref-type="bibr" rid="bib8">Baldwin et al., 2006</xref>; <xref ref-type="bibr" rid="bib87">Yan and Wang, 2006</xref>). In this context, it would be interesting to know whether central <italic>lox2/3</italic> plants surrounded by WT edge plants would experience a similar predation disadvantage as central <italic>TPS10</italic> plants surrounded by WT edge plants, and whether central WT plants surrounded by <italic>lox2/3</italic> edge plants would have an advantage over their neighbors in attracting predators.</p></sec><sec id="s3-5"><title>Conclusion</title><p>Now that careful and extensive research has established the importance of biodiversity to ecosystem function, the next step is to understand how the two are linked in order to craft informed socioeconomic and environmental policy (<xref ref-type="bibr" rid="bib60">Midgley, 2012</xref>). The community genetics perspective is essential in order to understand biodiversity effects at the level of individual heritable traits (<xref ref-type="bibr" rid="bib83">Whitham et al., 2006</xref>). This is especially true in light of recent research showing that intraspecific, genetic biodiversity is at least as important as interspecific biodiversity for the structure of ecological communities (<xref ref-type="bibr" rid="bib18">Crutsinger et al., 2006</xref>; <xref ref-type="bibr" rid="bib43">Johnson et al., 2006</xref>; <xref ref-type="bibr" rid="bib17">Cook-Patton et al., 2011</xref>; <xref ref-type="bibr" rid="bib61">Moreira and Mooney, 2013</xref>) and the recognition that interspecific biodiversity is an emergent property of genetic biodiversity (<xref ref-type="bibr" rid="bib44">Johnson and Stinchcombe, 2007</xref>; <xref ref-type="bibr" rid="bib37">Haloin and Strauss, 2008</xref>).</p><p>The work presented here demonstrates that the population-level effects of an indirect defense gene may be greater than its effects on individual plants, demonstrating that the population perspective is vital to our understanding of gene function. Furthermore, our results indicate that populations comprising individuals differing in only a single genetic trait can functionally imitate more diverse communities. Engineering functional diversity may become an important strategy in a world with over 10 billion humans (<xref ref-type="bibr" rid="bib81a">United Nations, Department of Economic and Social Affairs, Population Division, 2013</xref>), where every patch of arable land may be planted with custom-engineered high-yield crop monocultures.</p></sec></sec><sec sec-type="materials|methods" id="s4"><title>Materials and methods</title><sec id="s4-1"><title>WT and transgenic plant lines and glasshouse growth conditions</title><p>Seed germination, glasshouse growth conditions, and the <italic>Agrobacterium tumefaciens</italic> (strain LBA 4404)–mediated transformation procedure have been described previously (<xref ref-type="bibr" rid="bib55">Krügel et al., 2002</xref>; <xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>). Seeds of the 31st generation of the inbred ‘UT’ line of <italic>N. attenuata</italic> (Torr. ex S Watson) described by <xref ref-type="bibr" rid="bib55">Krügel et al. (2002)</xref> were used for the wild-type (WT) plants in all experiments. We used homozygous transformed lines of the second transformed generation (T<sub>2</sub>), each with a single transgene insertion, to either ectopically overexpress <italic>Zea mays TERPENE SYNTHASE 10</italic> (<italic>TPS10</italic>) under the control of a 35S promoter in the pSOL9 plasmid (<italic>TPS10</italic> line A-09-389-6/10-3, described in <xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>), or silence <italic>N. attenuata LIPOXYGENASE 2</italic> and <italic>LIPOXYGENASE 3</italic> (<italic>LOX2/3</italic>) via an inverted repeat construct specifically targeting both genes in the pSOL8 plasmid (line A-07-707-2, sequences targeting <italic>NaLOX2</italic> and <italic>NaLOX3</italic> described in <xref ref-type="bibr" rid="bib5">Allmann et al., 2010</xref>) (<xref ref-type="fig" rid="fig1">Figure 1</xref>). Vector construction and the pSOL8 and pSOL9 plasmids have been described previously (<xref ref-type="bibr" rid="bib32">Gase et al., 2011</xref>). To combine the ectopic overexpression of <italic>TPS10</italic> with the silencing of <italic>LOX2/3</italic>, a hemizygous cross was created between the <italic>lox2/3</italic> and <italic>TPS10</italic> T2 homozygous lines (<italic>lox2/3</italic>x<italic>TPS10</italic>, <xref ref-type="fig" rid="fig1">Figure 1</xref>). Further details of transgenic line screening and characterization are given in <xref ref-type="app" rid="app3">Appendix 3</xref>.</p></sec><sec id="s4-2"><title>Field plantations and growth conditions</title><p>For field experiments, seedlings were transferred to 50 mm peat pellets (Jiffy) 15 days after germination and gradually hardened to the environmental conditions of high sunlight and low relative humidity over 10 days. Small, adapted, size-matched rosette-stage plants were transplanted into a field plot in a native habitat in Utah located at latitude 37.146, longitude 114.020. Field plantations were conducted under APHIS permission numbers 06-242-3r-a3, 10-349-102r, and 11-350-101r. Plants were watered thoroughly once at planting and as needed over the first 2 weeks until roots were established; all plants received the same watering regime in each year. WT, <italic>TPS10</italic>, <italic>lox2/3</italic>, and <italic>lox2/3</italic>x<italic>TPS10</italic> plants were arranged as individuals and populations according to the descriptions in <xref ref-type="fig" rid="fig2">Figure 2</xref> and <xref ref-type="fig" rid="fig2s2">Figure 2—figure supplement 2</xref> (n = 12 in season one, n = 15 in season two of each individual or population type).</p><p>A distance of 0.5 m is sufficient for <italic>N. attenuata</italic>'s herbivores and their predators to distinguish odor from individual plants (<xref ref-type="bibr" rid="bib50">Kessler and Baldwin, 2001</xref>; <xref ref-type="bibr" rid="bib73">Schuman et al., 2012</xref>): so that insects could perceive plants in a population as a single unit, they were planted in 0.4 × 0.4 m squares, while the distance between single individuals/populations was ca. 1.5 m. Individual and population types were marked with small (ca. 5 cm long, ⌀ 0.5 cm) bamboo sticks having different numbers of horizontal markings, which were checked during experiments only when establishing treatment groups, or confirming current location within the experiment. During experiments and data collection, plants were referred to only by their number or position (e.g., n1 population 2 upper left, where upper left was defined from a standardized observer position). Plant genotypes were never identified during data collection to avoid bias, and experimental treatments and data collection were spatially and temporally randomized corresponding to the randomized arrangement of individuals and populations in the field plot (<xref ref-type="fig" rid="fig2s2">Figure 2—figure supplement 2</xref>).</p></sec><sec id="s4-3"><title>Tissue samples from glasshouse- and field-grown plants</title><p>Prior to measuring gene transcripts and induced defenses or HIPVs, glasshouse-grown plants, which were not damaged by herbivores, were treated with wounding and <italic>M. sexta</italic> oral secretions (W + OS) as a standardized method to mimic herbivore feeding as described previously (<xref ref-type="bibr" rid="bib73">Schuman et al., 2012</xref>). The first fully-expanded leaf (node +1) was used for quantifying gene transcript (<italic>LOX2</italic>, <italic>LOX3</italic>, <italic>TPS10</italic>) and jasmonate (JA, JA-Ile) accumulation on a set of bolting plants. The adjacent older leaf (node +2) was used to measure headspace volatiles on the same plants (data in <xref ref-type="app" rid="app2">Appendix 2</xref>, and see ‘Quantification of volatiles in the headspace of glasshouse-grown plants’). A separate set of plants was used following the same procedure for WT and <italic>TPS10</italic> jasmonates and GLVs, and the results are reported in <xref ref-type="bibr" rid="bib71">Schuman et al. (2014)</xref> and ratios to WT are shown in <xref ref-type="fig" rid="fig1">Figure 1B</xref>. Control plants were left untreated. Treated leaves, and leaves at the same position on control plants were cut at the petiole and wrapped in a double layer of aluminum foil, then immediately flash-frozen in liquid nitrogen, and kept at −80°C until processing and analysis.</p><p>For field-grown plants, five similar, mature, non-senescent stem leaves were harvested after the end of predation assays, on June 28th of experimental season two. Two of these leaves had been damaged by the <italic>M. sexta</italic> first-instar larvae used for predation assays, and three were systemic leaves with similar levels of damage from other naturally-occurring herbivores. Leaves were cut at the petiole, pooled in a single sample per plant and wrapped in a double layer of aluminum foil, then immediately frozen on dry ice insulated with ice packs frozen at −20°C. Samples were stored at −20°C until transport to the MPICE on dry ice, where they were kept at −80°C until processing and analysis.</p><p>All sample processing was carried out over liquid nitrogen until the addition of the extraction solvents. Prior to analysis, entire leaves were crushed with a mortar and pestle and each transferred to a 2 mL microcentrifuge tube for storage. For specific measurements, aliquots were weighed into microcentrifuge tubes containing two steel balls and finely ground in a GenoGrinder (SPEX Certi Prep, Metuchen, NJ) prior to extraction.</p></sec><sec id="s4-4"><title>Quantification of JA and JA-Ile in leaf tissue</title><p>JA and the active jasmonate hormone JA-Ile were extracted from 100 mg of leaf tissue (from n = 4 plants) using ethyl acetate spiked with internal standards (IS) and quantified by LC-MS/MS (Varian, Palo Alto, CA) as described by <xref ref-type="bibr" rid="bib62">Oh et al. (2012)</xref>, with the modification that 100 ng rather than 200 ng of [<sup>2</sup>H<sub>2</sub>]JA, and 20 ng rather than 40 ng of JA-[<sup>13</sup>C<sub>6</sub>]Ile were used as internal standards (IS). Individual jasmonates were quantified in ng by comparison to the corresponding IS peak area and normalized per g leaf tissue fresh mass (FM). The quantification of TPI in field tissue samples as an indicator of jasmonate signaling is described in <xref ref-type="app" rid="app3">Appendix 3</xref>.</p></sec><sec id="s4-5"><title>Quantification of volatiles in the headspace of glasshouse-grown plants</title><p>For GLVs, the +2 leaf (n = 4 plants) was enclosed immediately after W + OS elicitation in two 50 mL PET cups (Huhtamaki, Finland) lined on the edges with foam to protect leaves and with an activated charcoal filter attached to one side for incoming air, and secured with miniature claw-style hair clips as described previously (<xref ref-type="bibr" rid="bib74">Schuman et al., 2009</xref>). For <italic>TPS10</italic> products in the headspace of glasshouse-grown plants (n = 4), leaves were enclosed 24 h after W + OS treatment and the headspace was collected from 24–32 h during the peak emission of sesquiterpenes in <italic>N. attenuata</italic> (<xref ref-type="bibr" rid="bib35">Halitschke et al., 2000</xref>). Headspace VOCs were collected on 20 mg of Poropak Q (<xref ref-type="bibr" rid="bib80">Tholl et al., 2006</xref>; Sigma–Aldrich, St. Louis, MO) in self-packed filters (bodies and materials from ARS Inc., Gainesville, FL) by drawing ambient air through these clip cages at 300 mL min<sup>−1</sup> using a manifold with screw-close valves set to provide equal outflow, via pushing air at 2 to 3 bar through a Venturi aspirator as described previously (<xref ref-type="bibr" rid="bib62">Oh et al., 2012</xref>). Background VOCs present in ambient air were collected using empty trapping containers and background signals were later subtracted if necessary from raw intensities of plant samples prior to further processing.</p><p>After trapping, Porapak Q filters were stored at −20°C until extraction by addition of 320 ng of tetralin as an internal standard (IS), and elution of volatiles with 250 mL of dichloromethane (Sigma–Aldrich). Filters were eluted into a GC vial containing a 250 μL glass insert. Samples were analyzed on one of two different GC–MS instruments from Varian with columns from Phenomenex (Torrance, CA; 30 m × 0.25 mm i.d., 0.25 μm film thickness). 1 μL of each sample was injected in splitless mode, and then the injectors were returned to a 1:70 split ratio from 2 min after injection through the end of each run. He carrier gas was used with a column flow of 1 mL min<sup>−1</sup>. GLV samples were analyzed by a CP-3800 GC Saturn 2000 ion trap MS with a polar ZB-wax column and a CP-8200 autoinjector; the GC and MS were programmed as previously described for this instrument (<xref ref-type="bibr" rid="bib73">Schuman et al., 2012</xref>), and compounds were separated by a temperature ramp of 5°C min<sup>−1</sup> between 40°C and 185°C. <italic>TPS10</italic> products and other VOCs were analyzed on a CP-3800 GC coupled to a Saturn 4000 ion-trap mass spectrometer with a nonpolar ZB5 column and a CP-8400 autoinjector; The GC and MS were programmed as previously described for this instrument (<xref ref-type="bibr" rid="bib62">Oh et al., 2012</xref>), and compounds were separated by a temperature ramp of 5°C min<sup>−1</sup> between 40°C and 180°C.</p><p>Individual volatile compound peaks were quantified using the combined peak area of two specific and abundant ion traces per compound using MS Work Station Data Analysis software (Varian) and normalized by the 104 + 132 ion trace peak area from tetralin in each sample. The identification of compounds was conducted by comparing GC retention times and mass spectra to those of standards and mass spectra databases: Wiley version 6 and NIST (National Institute of Standards and Technology) spectral libraries.</p><p>The area of trapped leaves was quantified for comparison by scanning and calculating areas in pixels using SigmaScan (Systat Software Inc., San Jose, CA), and subsequently converting pixels to cm<sup>2</sup> using a size standard which was scanned with leaves.</p></sec><sec id="s4-6"><title>Quantification of plant volatiles in hexane extracts of field tissue samples</title><p>Plant volatiles were quantified in tissue from field-grown plants (n = 9–22). Frozen tissue (100 mg) was spiked with 800 ng tetralin as an internal standard (Sigma–Aldrich) and extracted in 300 µL hexane during an overnight rotating incubation at RT. Tissue was allowed to settle and 100 µL of water- and tissue-free hexane was transferred to a GC vial containing a 250 µL microinsert. Samples were analyzed by a Varian C<italic>P-</italic>3800 GC-Saturn 4000 ion trap MS connected to a ZB5 column (30 m × 0.25 mm i.d., 0.25 μm film thickness; Phenomenex) as described in ‘Quantification of volatiles in the headspace of glasshouse-grown plants’. Analyte quantities were expressed as percent tetralin IS per 100 mg FM.</p></sec><sec id="s4-7"><title>Quantification of volatiles in the open headspace of field-grown plants</title><p>For the data in <xref ref-type="fig" rid="fig3">Figure 3</xref>, on May 16th of season two, four activated charcoal filters were placed around each plant (n = 4) as shown in <xref ref-type="fig" rid="fig3">Figure 3A</xref> and the headspace was not enclosed. The open headspace was sampled for 7 h during the day at which time emission of most volatiles was previously found to be greatest (<xref ref-type="app" rid="app2">Appendix 2</xref>). After elution (see <xref ref-type="app" rid="app3">Appendix 3</xref>), eluents from all four filters were combined and concentrated to a volume of 15–20 µL under a gentle stream of nitrogen gas. Pooled and concentrated samples were analyzed by Agilent 6890N GC equipped with an Agilent 7683 auto-injector coupled with a LECO (St. Joseph, MI) Pegasus III time-of-flight MS with a 4D thermal modulator upgrade. Injected samples were separated first on a nonpolar column (C1 RTX-5MS, 20 m × 250 µm i.d. × 0.5 µm; Restek, Bellefonte, PA) and every 6 s (modulation time) transferred to a midpolar column (DB-17, 0.890 m × 100 µm i.d. × 0.1 µm; Agilent Technologies, Santa Clara, CA) for the second separation. Chromatography and analysis conditions as well as deconvolution, alignment, and integration of VOC analyte peaks are described in <xref ref-type="bibr" rid="bib31">Gaquerel et al. (2009)</xref>. During peak table alignment using the comparison feature imbedded in ChromaToF software (LECO), mass spectra alignment was accepted at a similarity threshold of 500/1000. Individual volatile compound peaks were normalized by the peak area of the tetralin IS in each sample. Analyte quantities were expressed as percent tetralin IS per plant.</p></sec><sec id="s4-8"><title>Measurement of <italic>M. sexta</italic> growth</title><p>As for OS collection, we used <italic>M. sexta</italic> neonates from an in-house colony at the MPICE. One <italic>M. sexta</italic> neonate per plant was placed on the youngest rosette leaf of elongated plants (n = 25 larvae). Larval movement was unrestricted within a plant, but plants were spaced on the table so that larvae could not move directly between plants. Larval mass was determined on days 8 and 12, corresponding on average to the third and fourth larval instars. Plant genotypes were randomized spatially on a single glasshouse table and temporally in placement and weighing of larvae. Due to mortality and to larval movement off of plants, 11–16 larvae remained within each treatment group by day 12.</p></sec><sec id="s4-9"><title>Assessment of foliar herbivore damage</title><p>Total canopy damage due to herbivores occurring naturally on the field plot was quantified on June 9th in season one, and on June 2nd in season two. Damage was calculated as described in <xref ref-type="bibr" rid="bib73">Schuman et al. (2012)</xref> by identifying damage from specific herbivores according to their characteristic feeding patterns, counting the number of leaves per plant (small leaves were counted as 1/5 to 1/2 of a leaf based on leaf area and large leaves were counted as 1 leaf), estimating the total percentage of leaf area damage due to each herbivore, and dividing the total leaf area damage from each herbivore by the total number of leaves. Leaf area damage was estimated in categories of 1%, 5%, 10%, 15%, and so on, in steps of 5%. All such damage estimates were made by MCS. Plants with a health index <2.5 (see ‘Ranking of apparent plant health’) were excluded from herbivore damage calculations as it was difficult to accurately assess herbivore damage to these plants.</p></sec><sec id="s4-10"><title>Herbivore and predator counts</title><p>In the morning of June 26th in season two, a subset of focal plants with a health index ≥2 was chosen (see ‘Ranking of apparent plant health’) and numbers of herbivorous and predacious arthropods present on the shoots of these plants were counted (total n's in <xref ref-type="supplementary-material" rid="SD1-data">Figure 6—source data 1, 2</xref>). The counting was done by a team of two people in parallel moving down the plot in the same direction, so as to complete counts as quickly as possible and to avoid disturbing plants prior to counting. Plants were randomly assigned to each counter. Counting proceeded by looking at each branch of the plant in a systematic order from the top to the bottom of the plant, observing both the adaxial and abaxial sides of leaves as well as full areas of stems. In this way especially mobile arthropods which tend to feed near the top of plants were either counted before they fled (<italic>Epitrix</italic> spp.), or could be counted once they had relocated to the base of the plant (<italic>T. notatus</italic>).</p></sec><sec id="s4-11"><title>Measurement of plant size and flower production</title><p>Plant size (rosette diameter, stem length, and branching) was measured at three times during the main period of vegetative growth and beginning of reproduction in season one (May 1st–2nd, 8th–11th, and 16th–17th), and after plants had switched from vegetative growth to reproduction in season two (June 4th–6th), at which time stem diameter was also measured. Rosette diameter was measured as the maximum diameter found by gently laying a ruler over the rosette; stem length was measured from the base of the stem to the tip of the apical inflorescence by placing a ruler beside the stem resting at the base of the rosette; stem diameter was measured by placing a calipers at the base of the stem immediately above the rosette; and all side branches 5 cm or longer were counted.</p><p>Additionally, plant reproductive output was monitored by counting the number of flowers (counted once the corolla protruded visibly from the calyx) before removal at two time points after plants began to flower: on May 8th–11th and 16th–17th in season one, and on May 29th and June 4th–6th in season two. Additionally, buds >2 mm long were counted in season two. After these counts, floral meristems were removed as necessary to prevent the distribution of ripe seeds, as is required for field-released transgenic plants.</p></sec><sec id="s4-12"><title>Quantification of plant mortality and infestation by <italic>Trichobaris mucorea</italic></title><p>Plant mortality was quantified as the number of plants which died between the end of planting, and the end of the experiment when all plants and remains were removed from the field plot. During removal of plants on June 29th in season two, we split stems open lengthwise to check for <italic>T. mucorea</italic> frass and larvae (<xref ref-type="bibr" rid="bib22">Diezel et al., 2011</xref>). Plants were counted as infested if the main stem was hollow and filled with <italic>T. mucorea</italic> frass, or if a larva was found in such a hollow, frass-filled stem or axial branch. These counts were done by tallying totals without reference to specific plant ID's, and thus could only be analyzed using unreplicated statistical tests.</p></sec><sec id="s4-13"><title>Ranking of apparent plant health</title><p>In season two, plant health was also ranked by MCS on a scale of 1–5 in increments of 0.5 (1, 1.5, 2, 2.5, …, 5) using the index in <xref ref-type="fig" rid="fig7">Figure 7C</xref> of <xref ref-type="bibr" rid="bib73">Schuman et al. (2012)</xref>: 5, healthy, <5% of shoot senescent, no chlorosis or signs of water stress; 4, 5–10% of shoot senescent and may have visible chlorosis; 3, 10–15% of shoot senescent, some water stress, up to 50% of canopy chlorotic; 2, >15% of shoot senescent, heavy water stress or visibly sick; 1, dead. This scale was sufficiently objective that several observers not directly involved in the experiment arrived independently at the same ranks, to a precision of ±0.5, to describe the health indices of randomly-chosen plants, after training on fewer than 10 example plants.</p></sec><sec id="s4-14"><title>Assay of predation by <italic>Geocoris</italic> spp<italic>.</italic></title><p>Five consecutive predation assays were conducted between June 21st and June 27th at the end of experimental season two, when <italic>Geocoris</italic> spp<italic>.</italic> and herbivores were abundant. <italic>M. sexta</italic> eggs and larvae (total n's given in the source data file for <xref ref-type="fig" rid="fig11">Figure 11</xref>) were used as bait for <italic>Geocoris</italic> spp<italic>.</italic> predators as described by <xref ref-type="bibr" rid="bib50">Kessler and Baldwin, 2001</xref> and <xref ref-type="bibr" rid="bib79">Steppuhn et al. (2008)</xref>. Native <italic>Manduca</italic> spp. did not provide sufficient synchronously oviposited eggs for trials, and so we used <italic>M. sexta</italic> eggs kindly provided by C Miles from her laboratory colony at Binghamton University, Vestal, New York. Size-matched plants with health indices ≥2 in populations were used. We treated LOX2/3-expressing (WT, TPS10) and <italic>LOX2/3</italic>-deficient (<italic>lox2/3</italic>, <italic>lox2/3</italic>x<italic>TPS10</italic>) populations as two separate experimental groups because it was not possible to match plants across these two groups for apparent health or replicate numbers (both were lower for <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> populations). During the five predation assays it was occasionally necessary to select new plants, preventing a repeated measures analysis of the predation data; however, newly selected plants were matched as described. At the end of predation assays, tissue samples were taken from a subset of these plants for the extraction of volatiles and determination of TPI activity (<xref ref-type="fig" rid="fig1">Figure 1</xref>, <xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref> and preceding methods).</p><p>For each assay, one first-instar <italic>M. sexta</italic> larva and one egg were placed on one lower stem leaf in a standardized position on each plant, and the predation of eggs and larva was monitored after 24 h. No other <italic>Manduca</italic> spp. eggs or larvae were present on the plants used during the assays. Eggs were affixed to the underside of the leaf with a drop of α-cellulose glue (KVS, Germany), which does not damage plants or affect volatile emission (<xref ref-type="bibr" rid="bib50">Kessler and Baldwin, 2001</xref>). Feeding by the <italic>M. sexta</italic> larva ensured the locally elicited emission of HIPVs. Larvae were considered to be predated when either the larva was missing, but clearly identifiable, fresh <italic>M. sexta</italic> feeding damage was present, or when the predated larval carcass was found on the same plant; eggs were considered predated when the eggshell was empty but intact, except for a small hole which characterizes the typical damage caused by <italic>Geocoris</italic> spp<italic>.</italic> feeding as described in <xref ref-type="bibr" rid="bib50">Kessler and Baldwin (2001)</xref> and depicted in Figure 3 of <xref ref-type="bibr" rid="bib73">Schuman et al. (2012)</xref>. Eggs occasionally collapse during <italic>Geocoris</italic> spp<italic>.</italic> predation, but collapsed eggs were not counted unless the eggs were at least 50% empty with a visible feeding hole.</p></sec><sec id="s4-15"><title>Statistical analyses</title><p>Summary statistics were calculated in Microsoft Excel, as were G-tests, Fisher's exact tests, and tests of Spearman's correlation coefficients, which were calculated using Microsoft Excel spreadsheets from JH Macdonald's online Handbook of Biological Statistics (<xref ref-type="bibr" rid="bib57">Macdonald, 2009</xref>). All other statistical tests were calculated in R version 2.15.2, using the RStudio interface version 0.96.316 or version 0.97.449 (<xref ref-type="bibr" rid="bib65">R Core Team, 2012</xref>). In some cases, data which were clearly not different upon visual inspection and which had been found in similar data sets not to differ significantly were not statistically tested (e.g., rosette diameter in some subsets of data from season one).</p><p>When possible, Welch t-tests were used for pairwise comparisons and ANOVAs, linear mixed-effects models, generalized linear models, or generalized linear mixed-effects models were used for multiple comparisons as appropriate; we checked treatment groups graphically (quantile–quantile plots, residual v. fitted plots) and statistically (Shapiro–Wilk test, Bartlett test) for compliance with model assumptions, including normality and homoscedasticity. Data containing too many zero values or not meeting model assumptions were analyzed using ranking tests (Wilcoxon rank sum for pairwise comparisons, Kruskal–Wallis for multiple groups). R package nlme was used for linear mixed-effects models (<xref ref-type="bibr" rid="bib63">Pinheiro et al., 2014</xref>), and packages lme4 (<xref ref-type="bibr" rid="bib11">Bates et al., 2014a, b</xref>) and multcomp (<xref ref-type="bibr" rid="bib42">Hothorn et al., 2008</xref>) were used for generalized linear mixed-effects models.</p><p>The optimal fixed structure for ANOVAs, linear mixed-effects models, generalized linear models, and generalized linear mixed-effects models was found by stepwise model simplification and factor level reduction followed by the comparison of the models via the Akaike information criterion (AIC, <xref ref-type="bibr" rid="bib1">Akaike, 1973</xref>). Specific examples of model simplification for individual analyses are given in <xref ref-type="app" rid="app3">Appendix 3</xref> and <xref ref-type="supplementary-material" rid="SD8-data">Figure 8—source data 1</xref>.</p><p>Holm-Bonferroni p-value corrections were calculated in Microsoft Excel for families of tests on the same data. Like the Bonferroni correction, the Holm-Bonferroni correction controls the familywise error rate and is simple to calculate, but the Holm-Bonferroni method is more powerful (<xref ref-type="bibr" rid="bib41">Holm, 1979</xref>). Briefly, all p<italic>-</italic>values in a family of tests conducted on the same data set and subsets were listed from smallest to largest. The smallest p<italic>-</italic>value was multiplied by the total number of tests (n) conducted on that data. If the resulting corrected p<italic>-</italic>value was <0.05, then the next-smallest p<italic>-</italic>value was multiplied by n − 1. If the resulting corrected p<italic>-</italic>value was <0.05, then the third smallest p<italic>-</italic>value was multiplied by n − 2, and so on. At the first correction resulting in a p<italic>-</italic>value ≥0.05, that p<italic>-</italic>value and all larger p<italic>-</italic>values were considered non-significant. The corrected p<italic>-</italic>values reported in the text are the products of the correction procedure, for example, (smallest p-value) × n or (second smallest p-value) × (n − 1).</p></sec></sec></body><back><ack id="ack"><title>Acknowledgements</title><p>The authors gratefully acknowledge JO Miller for assistance with the processing of open headspace samples and tissue hexane extracts from field experiments; J Koenig, R Ludwig, B Naumann, N Petersen, I Schmidt, and D Schweizer for assistance with glasshouse experiments which required several careful hands and precise timing, and help processing the resulting samples; D Veit for constructing the volatile trapping set-up used in the glasshouse and the 125-sample plate used in the radial diffusion assay of TPI activity; D Kessler for photos of <italic>N. attenuata</italic>, <italic>G</italic>. <italic>pallens</italic> and <italic>M. sexta</italic> and A Steppuhn for the photo of <italic>T</italic>. <italic>mucorea</italic>; C miles for providing <italic>M. sexta</italic> eggs and larvae for use in field experiments; and fellow field researchers K Barthel, C Diezel, T Dinh, M Erb, I Galis, J Gershenzon, P Gilardoni, M Heinrich, M Kallenbach, D Kessler, D Marciniak, S Meldau, Y Oh, E Rothe, S Schuck, M Stanton, A Steppke, V Tamhane, L Ullmann-Zeunert, Al Weinhold, Ar Weinhold, and M Woldemariam for insightful scientific discussions and generous assistance when extra hands were needed. We thank APHIS for constructive regulatory oversight and Brigham Young University for the use of Lytle Preserve.</p></ack><sec sec-type="additional-information" id="s5"><title>Additional information</title><fn-group content-type="competing-interest"><title>Competing interests</title><fn fn-type="conflict" id="conf1"><p>ITB: Senior editor, <italic>eLife.</italic></p></fn><fn fn-type="conflict" id="conf2"><p>The other authors declare that no competing interests exist.</p></fn></fn-group><fn-group content-type="author-contribution"><title>Author contributions</title><fn fn-type="con" id="con1"><p>MCS, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article</p></fn><fn fn-type="con" id="con2"><p>ITB, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article</p></fn><fn fn-type="con" id="con3"><p>SA, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article, Contributed unpublished essential data or reagents</p></fn></fn-group></sec><sec sec-type="supplementary-material" id="s6"><title>Additional files</title><sec sec-type="datasets" id="s6-1"><title>Major dataset</title><p>The following dataset was generated:</p><p><related-object content-type="generated-dataset" source-id="http://dx.doi.org/10.5061/dryad.qj007" source-id-type="uri" id="dataro1"><collab collab-type="author">Schuman MC</collab>, <collab collab-type="author">Allmann SA</collab>, <collab collab-type="author">Baldwin IT</collab>, <year>2015</year><x>, </x><source>Data from: Plant defense phenotypes determine the consequences of volatile emission for individuals and neighbors</source><x>, </x><ext-link ext-link-type="uri" xlink:href="http://dx.doi.org/10.5061/dryad.qj007">http://dx.doi.org/10.5061/dryad.qj007</ext-link><x>, </x><comment>Available at Dryad Digital Repository under a CC0 Public Domain Dedication.</comment></related-object></p></sec></sec><ref-list><title>References</title><ref id="bib1"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Akaike</surname><given-names>H</given-names></name></person-group><year>1973</year><article-title>Information theory and an extension of the maximum likelihood principle</article-title><person-group person-group-type="editor"><name><surname>Kotz</surname><given-names>S</given-names></name><name><surname>Johnson</surname><given-names>NL</given-names></name></person-group><source>Breakthroughs in statistics volume 1. foundations and basic theory</source><publisher-loc>New York, NY, USA</publisher-loc><publisher-name>Springer-Verlag</publisher-name><fpage>610</fpage><lpage>624</lpage></element-citation></ref><ref id="bib2"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Allard</surname><given-names>ARW</given-names></name><name><surname>Adams</surname><given-names>J</given-names></name></person-group><year>1969</year><article-title>Population studies in predominantly self-pollinating species. XIII. Intergenotypic competition and population structure in barley and wheat</article-title><source>The American Naturalist</source><volume>103</volume><fpage>621</fpage><lpage>645</lpage><comment><ext-link ext-link-type="uri" xlink:href="http://www.jstor.org/stable/2459037">http://www.jstor.org/stable/2459037</ext-link></comment><pub-id pub-id-type="doi">10.1086/282630</pub-id></element-citation></ref><ref id="bib3"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Allison</surname><given-names>JD</given-names></name><name><surname>Hare</surname><given-names>DJ</given-names></name></person-group><year>2009</year><article-title>Learned and naïve natural enemy responses and the interpretation of volatile organic compounds as cues or signals</article-title><source>The New Phytologist</source><volume>184</volume><fpage>768</fpage><lpage>782</lpage><pub-id pub-id-type="doi">10.1111/j.1469-8137.2009.03046.x</pub-id></element-citation></ref><ref id="bib4"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Allmann</surname><given-names>S</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2010</year><article-title>Insects betray themselves in nature to predators by rapid isomerization of green leaf volatiles</article-title><source>Science</source><volume>329</volume><fpage>1075</fpage><lpage>1078</lpage><pub-id pub-id-type="doi">10.1126/science.1191634</pub-id></element-citation></ref><ref id="bib5"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Allmann</surname><given-names>S</given-names></name><name><surname>Halitschke</surname><given-names>R</given-names></name><name><surname>Schuurink</surname><given-names>RC</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2010</year><article-title>Oxylipin channelling in <italic>Nicotiana attenuata</italic>: lipoxygenase 2 supplies substrates for green leaf volatile production</article-title><source>Plant, Cell & Environment</source><volume>33</volume><fpage>2028</fpage><lpage>2040</lpage><pub-id pub-id-type="doi">10.1111/j.1365-3040.2010.02203.x</pub-id></element-citation></ref><ref id="bib6"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Antonovics</surname><given-names>J</given-names></name></person-group><year>1992</year><article-title>Toward community genetics</article-title><person-group person-group-type="editor"><name><surname>Fritz</surname><given-names>RS</given-names></name><name><surname>Simms</surname><given-names>EL</given-names></name></person-group><source>Plant resistance to herbivores and pathogens</source><publisher-loc>Chicago, IL, USA</publisher-loc><publisher-name>University of Chicago Press</publisher-name><fpage>426</fpage><lpage>449</lpage></element-citation></ref><ref id="bib7"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Arimura</surname><given-names>G</given-names></name><name><surname>Shiojiri</surname><given-names>K</given-names></name><name><surname>Karban</surname><given-names>R</given-names></name></person-group><year>2010</year><article-title>Acquired immunity to herbivory and allelopathy caused by airborne plant emissions</article-title><source>Phytochemistry</source><volume>71</volume><fpage>1642</fpage><lpage>1649</lpage><pub-id pub-id-type="doi">10.1016/j.phytochem.2010.06.021</pub-id></element-citation></ref><ref id="bib8"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Baldwin</surname><given-names>IT</given-names></name><name><surname>Halitschke</surname><given-names>R</given-names></name><name><surname>Paschold</surname><given-names>A</given-names></name><name><surname>von Dahl</surname><given-names>CC</given-names></name><name><surname>Preston</surname><given-names>KA</given-names></name></person-group><year>2006</year><article-title>Volatile signaling in plant-plant interactions: ‘talking trees’ in the genomics era</article-title><source>Science</source><volume>311</volume><fpage>812</fpage><lpage>815</lpage><pub-id pub-id-type="doi">10.1126/science.1118446</pub-id></element-citation></ref><ref id="bib9"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Balvanera</surname><given-names>P</given-names></name><name><surname>Pfisterer</surname><given-names>AB</given-names></name><name><surname>Buchmann</surname><given-names>N</given-names></name><name><surname>He</surname><given-names>J-S</given-names></name><name><surname>Nakashizuka</surname><given-names>T</given-names></name><name><surname>Raffaelli</surname><given-names>D</given-names></name><name><surname>Schmid</surname><given-names>B</given-names></name></person-group><year>2006</year><article-title>Quantifying the evidence for biodiversity effects on ecosystem functioning and services</article-title><source>Ecology Letters</source><volume>9</volume><fpage>1146</fpage><lpage>1156</lpage><pub-id pub-id-type="doi">10.1111/j.1461-0248.2006.00963.x</pub-id></element-citation></ref><ref id="bib10"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Barbosa</surname><given-names>P</given-names></name><name><surname>Hines</surname><given-names>J</given-names></name><name><surname>Kaplan</surname><given-names>I</given-names></name><name><surname>Martinson</surname><given-names>H</given-names></name><name><surname>Szczepaniec</surname><given-names>A</given-names></name><name><surname>Szendrei</surname><given-names>Z</given-names></name></person-group><year>2009</year><article-title>Associational resistance and associational susceptibility: having right or wrong neighbors</article-title><source>Annual Review of Ecology, Evolution, and Systematics</source><volume>40</volume><fpage>1</fpage><lpage>20</lpage><pub-id pub-id-type="doi">10.1146/annurev.ecolsys.110308.120242</pub-id></element-citation></ref><ref id="bib11"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Bates</surname><given-names>D</given-names></name><name><surname>Maechler</surname><given-names>M</given-names></name><name><surname>Bolker</surname><given-names>BM</given-names></name><name><surname>Walker</surname><given-names>S</given-names></name></person-group><year>2014a</year><source>lme4: linear mixed-effects models using Eigen and S4. R package version 1.1-7</source><comment><ext-link ext-link-type="uri" xlink:href="http://CRAN.R-project.org/package=lme4">http://CRAN.R-project.org/package=lme4</ext-link></comment></element-citation></ref><ref id="bib12"><element-citation publication-type="other"><person-group person-group-type="author"><name><surname>Bates</surname><given-names>D</given-names></name><name><surname>Maechler</surname><given-names>M</given-names></name><name><surname>Bolker</surname><given-names>BM</given-names></name><name><surname>Walker</surname><given-names>S</given-names></name></person-group><year>2014b</year><article-title>lme4: linear mixed-effects models using Eigen and S4</article-title><source>ArXiv</source><comment><ext-link ext-link-type="uri" xlink:href="http://arxiv.org/abs/1406.5823">http://arxiv.org/abs/1406.5823</ext-link></comment></element-citation></ref><ref id="bib13"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Bruce</surname><given-names>TJ</given-names></name><name><surname>Wadhams</surname><given-names>LJ</given-names></name><name><surname>Woodcock</surname><given-names>CM</given-names></name></person-group><year>2005</year><article-title>Insect host location: a volatile situation</article-title><source>Trends in Plant Science</source><volume>10</volume><fpage>269</fpage><lpage>274</lpage><pub-id pub-id-type="doi">10.1016/j.tplants.2005.04.003</pub-id></element-citation></ref><ref id="bib14"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Bubner</surname><given-names>B</given-names></name><name><surname>Gase</surname><given-names>K</given-names></name><name><surname>Berger</surname><given-names>B</given-names></name><name><surname>Link</surname><given-names>D</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2006</year><article-title>Occurrence of tetraploidy in <italic>Nicotiana attenuata</italic> plants after <italic>Agrobacterium</italic>-mediated transformation is genotype specific but independent of polysomaty of explant tissue</article-title><source>Plant Cell Reports</source><volume>25</volume><fpage>668</fpage><lpage>675</lpage><pub-id pub-id-type="doi">10.1007/s00299-005-0111-4</pub-id></element-citation></ref><ref id="bib15"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Calogirou</surname><given-names>A</given-names></name><name><surname>Larsen</surname><given-names>BR</given-names></name><name><surname>Brussol</surname><given-names>C</given-names></name><name><surname>Duane</surname><given-names>M</given-names></name><name><surname>Kotzias</surname><given-names>D</given-names></name></person-group><year>1996</year><article-title>Decomposition of terpenes by ozone during sampling on Tenax</article-title><source>Analytical Chemistry</source><volume>68</volume><fpage>1499</fpage><lpage>1506</lpage><pub-id pub-id-type="doi">10.1021/ac950803i</pub-id></element-citation></ref><ref id="bib16"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Choh</surname><given-names>Y</given-names></name><name><surname>Shimoda</surname><given-names>T</given-names></name><name><surname>Ozawa</surname><given-names>R</given-names></name><name><surname>Dicke</surname><given-names>M</given-names></name><name><surname>Takabayashi</surname><given-names>J</given-names></name></person-group><year>2004</year><article-title>Exposure of lima bean leaves to volatiles from herbivore-induced conspecific plants results in emission of carnivore attractants: active or passive process?</article-title><source>Journal of Chemical Ecology</source><volume>30</volume><fpage>1305</fpage><lpage>1317</lpage><pub-id pub-id-type="doi">10.1023/B:JOEC.0000037741.13402.19</pub-id></element-citation></ref><ref id="bib17"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Cook-Patton</surname><given-names>SC</given-names></name><name><surname>McArt</surname><given-names>SH</given-names></name><name><surname>Parachnowitsch</surname><given-names>AL</given-names></name><name><surname>Thaler</surname><given-names>JS</given-names></name><name><surname>Agrawal</surname><given-names>AA</given-names></name></person-group><year>2011</year><article-title>A direct comparison of the consequences of plant genotypic and species diversity on communities and ecosystem function</article-title><source>Ecology</source><volume>92</volume><fpage>915</fpage><lpage>923</lpage><pub-id pub-id-type="doi">10.1890/10-0999.1</pub-id></element-citation></ref><ref id="bib18"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Crutsinger</surname><given-names>GM</given-names></name><name><surname>Collins</surname><given-names>MD</given-names></name><name><surname>Fordyce</surname><given-names>JA</given-names></name><name><surname>Gombert</surname><given-names>Z</given-names></name><name><surname>Nice</surname><given-names>CC</given-names></name><name><surname>Sanders</surname><given-names>NJ</given-names></name></person-group><year>2006</year><article-title>Plant genotypic diversity predicts community structure and governs an ecosystem process</article-title><source>Science</source><volume>647</volume><fpage>966</fpage><lpage>968</lpage><pub-id pub-id-type="doi">10.1126/science.1128326</pub-id></element-citation></ref><ref id="bib19"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Darwin</surname><given-names>C</given-names></name></person-group><year>1876</year><source>The origin of species by means of natural selection; or, the preservation of favoured races in the struggle for life 6th, corr</source><publisher-loc>London</publisher-loc><publisher-name>John Murray</publisher-name><fpage>86</fpage><comment>URL: <ext-link ext-link-type="uri" xlink:href="http://darwin-online.org.uk/contents.html#origin">http://darwin-online.org.uk/contents.html#origin</ext-link></comment></element-citation></ref><ref id="bib20"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Degenhardt</surname><given-names>J</given-names></name><name><surname>Hiltpold</surname><given-names>I</given-names></name><name><surname>Köllner</surname><given-names>TG</given-names></name><name><surname>Frey</surname><given-names>M</given-names></name><name><surname>Gierl</surname><given-names>A</given-names></name><name><surname>Gershenzon</surname><given-names>J</given-names></name><name><surname>Hibbard</surname><given-names>BE</given-names></name><name><surname>Ellersieck</surname><given-names>MR</given-names></name><name><surname>Turlings</surname><given-names>TC</given-names></name></person-group><year>2009</year><article-title>Restoring a maize root signal that attracts insect-killing nematodes to control a major pest</article-title><source>Proceedings of the National Academy of Sciences of USA</source><volume>106</volume><fpage>13213</fpage><lpage>13218</lpage><pub-id pub-id-type="doi">10.1073/pnas.0906365106</pub-id></element-citation></ref><ref id="bib21"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>De Wit</surname><given-names>CT</given-names></name></person-group><year>1960</year><source>On competition. centrum voor Landbaupublikaties</source><publisher-loc>Wageningen, The Netherlands</publisher-loc><publisher-name>Landbouwdocumentatie</publisher-name><fpage>82</fpage></element-citation></ref><ref id="bib22"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Diezel</surname><given-names>C</given-names></name><name><surname>Kessler</surname><given-names>D</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2011</year><article-title>Pithy protection: <italic>Nicotiana attenuata</italic>'s jasmonic acid-mediated defenses are required to resist stem-boring weevil larvae</article-title><source>Plant Physiology</source><volume>155</volume><fpage>1936</fpage><lpage>1946</lpage><pub-id pub-id-type="doi">10.1104/pp.110.170936</pub-id></element-citation></ref><ref id="bib23"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Dobzhansky</surname><given-names>T</given-names></name></person-group><year>1973</year><article-title>Nothing in biology makes sense except in the light of evolution</article-title><source>Plant Physiology</source><volume>35</volume><fpage>125</fpage><lpage>129</lpage><pub-id pub-id-type="doi">10.2307/4444260</pub-id></element-citation></ref><ref id="bib24"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Dudareva</surname><given-names>N</given-names></name><name><surname>Klempien</surname><given-names>A</given-names></name><name><surname>Muhlemann</surname><given-names>JK</given-names></name><name><surname>Kaplan</surname><given-names>I</given-names></name></person-group><year>2013</year><article-title>Biosynthesis, function and metabolic engineering of plant volatile organic compounds</article-title><source>The New Phytologist</source><volume>198</volume><fpage>16</fpage><lpage>32</lpage><pub-id pub-id-type="doi">10.1111/nph.12145</pub-id></element-citation></ref><ref id="bib25"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Dudareva</surname><given-names>N</given-names></name><name><surname>Pichersky</surname><given-names>E</given-names></name></person-group><year>2008</year><article-title>Metabolic engineering of plant volatiles</article-title><source>Current Opinion in Biotechnology</source><volume>19</volume><fpage>181</fpage><lpage>189</lpage><pub-id pub-id-type="doi">10.1016/j.copbio.2008.02.011</pub-id></element-citation></ref><ref id="bib26"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Ebeling</surname><given-names>A</given-names></name><name><surname>Pompe</surname><given-names>S</given-names></name><name><surname>Baade</surname><given-names>J</given-names></name><name><surname>Eisenhauer</surname><given-names>N</given-names></name><name><surname>Hillebrand</surname><given-names>H</given-names></name><name><surname>Proulx</surname><given-names>R</given-names></name><name><surname>Roscher</surname><given-names>C</given-names></name><name><surname>Schmid</surname><given-names>B</given-names></name><name><surname>Wirth</surname><given-names>C</given-names></name><name><surname>Weisser</surname><given-names>WW</given-names></name></person-group><year>2014</year><article-title>A trait-based experimental approach to understand the mechanisms underlying biodiversity–ecosystem functioning relationships</article-title><source>Basic and Applied Ecology</source><volume>15</volume><fpage>229</fpage><lpage>240</lpage><pub-id pub-id-type="doi">10.1016/j.baae.2014.02.003</pub-id></element-citation></ref><ref id="bib27"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Engelberth</surname><given-names>J</given-names></name><name><surname>Alborn</surname><given-names>HT</given-names></name><name><surname>Schmelz</surname><given-names>EA</given-names></name><name><surname>Tumlinson</surname><given-names>JH</given-names></name></person-group><year>2004</year><article-title>Airborne signals prime plants against insect herbivore attack</article-title><source>Proceedings of the National Academy of Sciences of USA</source><volume>101</volume><fpage>1781</fpage><lpage>1785</lpage><pub-id pub-id-type="doi">10.1073/pnas.0308037100</pub-id></element-citation></ref><ref id="bib28"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Feeny</surname><given-names>P</given-names></name></person-group><year>1976</year><article-title>Plant apparency and chemical defense</article-title><person-group person-group-type="editor"><name><surname>Wallace</surname><given-names>JW</given-names></name><name><surname>Mansell</surname><given-names>RL</given-names></name></person-group><source>Recent advances in phytochemistry</source><publisher-loc>New York, NY, USA</publisher-loc><publisher-name>Plenum Press</publisher-name><fpage>1</fpage><lpage>40</lpage></element-citation></ref><ref id="bib29"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Frost</surname><given-names>CJ</given-names></name><name><surname>Mescher</surname><given-names>MC</given-names></name><name><surname>Dervinis</surname><given-names>C</given-names></name><name><surname>Davis</surname><given-names>JM</given-names></name><name><surname>Carlson</surname><given-names>JE</given-names></name><name><surname>De Moraes</surname><given-names>CM</given-names></name></person-group><year>2008</year><article-title>Priming defense genes and metabolites in hybrid poplar by the green leaf volatile <italic>cis</italic>-3-hexenyl acetate</article-title><source>The New Phytologist</source><volume>180</volume><fpage>722</fpage><lpage>734</lpage><pub-id pub-id-type="doi">10.1111/j.1469-8137.2008.02599.x</pub-id></element-citation></ref><ref id="bib30"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gaquerel</surname><given-names>E</given-names></name><name><surname>Heiling</surname><given-names>S</given-names></name><name><surname>Schoettner</surname><given-names>M</given-names></name><name><surname>Zurek</surname><given-names>G</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2010</year><article-title>Development and validation of a liquid chromatography-electrospray ionization-time-of-flight mass spectrometry method for induced changes in <italic>Nicotiana attenuata</italic> leaves during simulated herbivory</article-title><source>Journal of Agricultural and Food Chemistry</source><volume>58</volume><fpage>9418</fpage><lpage>9427</lpage><pub-id pub-id-type="doi">10.1021/jf1017737</pub-id></element-citation></ref><ref id="bib31"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gaquerel</surname><given-names>E</given-names></name><name><surname>Weinhold</surname><given-names>A</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2009</year><article-title>Molecular interactions between the specialist herbivore <italic>Manduca sexta</italic> (Lepidoptera, Sphigidae) and its natural host <italic>Nicotiana attenuata</italic>. VIII. An unbiased GCxGC-ToFMS analysis of the plant's elicited volatile emissions</article-title><source>Plant Physiology</source><volume>149</volume><fpage>1408</fpage><lpage>1423</lpage><pub-id pub-id-type="doi">10.1104/pp.108.130799</pub-id></element-citation></ref><ref id="bib32"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gase</surname><given-names>K</given-names></name><name><surname>Weinhold</surname><given-names>A</given-names></name><name><surname>Bozorov</surname><given-names>T</given-names></name><name><surname>Schuck</surname><given-names>S</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2011</year><article-title>Efficient screening of transgenic plant lines for ecological research</article-title><source>Molecular Ecology Resources</source><volume>11</volume><fpage>890</fpage><lpage>902</lpage><pub-id pub-id-type="doi">10.1111/j.1755-0998.2011.03017.x</pub-id></element-citation></ref><ref id="bib33"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Green</surname><given-names>RE</given-names></name><name><surname>Cornell</surname><given-names>SJ</given-names></name><name><surname>Scharlemann</surname><given-names>JPW</given-names></name><name><surname>Balmford</surname><given-names>A</given-names></name></person-group><year>2005</year><article-title>Farming and the fate of wild nature</article-title><source>Science</source><volume>307</volume><fpage>550</fpage><lpage>555</lpage><pub-id pub-id-type="doi">10.1126/science.1106049</pub-id></element-citation></ref><ref id="bib34"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Halitschke</surname><given-names>R</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2003</year><article-title>Antisense LOX expression increases herbivore performance by decreasing defense responses and inhibiting growth-related transcriptional reorganization in <italic>Nicotiana attenuata</italic></article-title><source>The Plant Journal</source><volume>36</volume><fpage>794</fpage><lpage>807</lpage><pub-id pub-id-type="doi">10.1046/j.1365-313X.2003.01921.x</pub-id></element-citation></ref><ref id="bib35"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Halitschke</surname><given-names>R</given-names></name><name><surname>Keßler</surname><given-names>A</given-names></name><name><surname>Kahl</surname><given-names>J</given-names></name><name><surname>Lorenz</surname><given-names>A</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2000</year><article-title>Ecophysiological comparison of direct and indirect defenses in <italic>Nicotiana attenuata</italic></article-title><source>Oecologia</source><volume>124</volume><fpage>408</fpage><lpage>417</lpage><pub-id pub-id-type="doi">10.1007/s004420000389</pub-id></element-citation></ref><ref id="bib36"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Halitschke</surname><given-names>R</given-names></name><name><surname>Stenberg</surname><given-names>JA</given-names></name><name><surname>Keßler</surname><given-names>D</given-names></name><name><surname>Keßler</surname><given-names>A</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2008</year><article-title>Shared signals – ‘alarm calls’ from plants increase apparency to herbivores and their enemies in nature</article-title><source>Ecology Letters</source><volume>11</volume><fpage>24</fpage><lpage>34</lpage><pub-id pub-id-type="doi">10.1111/j.1461-0248.2007.01123.x</pub-id></element-citation></ref><ref id="bib37"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Haloin</surname><given-names>JR</given-names></name><name><surname>Strauss</surname><given-names>SY</given-names></name></person-group><year>2008</year><article-title>Interplay between ecological communities and evolution: review of feedbacks from microevolutionary to macroevolutionary scales</article-title><source>Annals of the New York Academy of Sciences</source><volume>1133</volume><fpage>87</fpage><lpage>125</lpage><pub-id pub-id-type="doi">10.1196/annals.1438.003</pub-id></element-citation></ref><ref id="bib38"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Heil</surname><given-names>M</given-names></name><name><surname>Kost</surname><given-names>C</given-names></name></person-group><year>2006</year><article-title>Priming of indirect defences</article-title><source>Ecology Letters</source><volume>9</volume><fpage>813</fpage><lpage>817</lpage><pub-id pub-id-type="doi">10.1111/j.1461-0248.2006.00932.x</pub-id></element-citation></ref><ref id="bib39"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Hilbe</surname><given-names>C</given-names></name><name><surname>Nowak</surname><given-names>MA</given-names></name><name><surname>Sigmund</surname><given-names>K</given-names></name></person-group><year>2013</year><article-title>Evolution of extortion in iterated prisoner's dilemma games</article-title><source>Proceedings of the National Academy of Sciences of USA</source><volume>110</volume><fpage>6913</fpage><lpage>6918</lpage><pub-id pub-id-type="doi">10.1073/pnas.1214834110</pub-id></element-citation></ref><ref id="bib40"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Himanen</surname><given-names>SJ</given-names></name><name><surname>Blande</surname><given-names>JD</given-names></name><name><surname>Klemola</surname><given-names>T</given-names></name><name><surname>Pulkkinen</surname><given-names>J</given-names></name><name><surname>Heijari</surname><given-names>J</given-names></name><name><surname>Holopainen</surname><given-names>JK</given-names></name></person-group><year>2010</year><article-title>Birch (<italic>Betula</italic> spp.) leaves adsorb and re-release volatiles specific to neighbouring plants–a mechanism for associational herbivore resistance?</article-title><source>The New Phytologist</source><volume>186</volume><fpage>722</fpage><lpage>732</lpage><pub-id pub-id-type="doi">10.1111/j.1469-8137.2010.03220.x</pub-id></element-citation></ref><ref id="bib41"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Holm</surname><given-names>S</given-names></name></person-group><year>1979</year><article-title>A simple sequentially rejective multiple test procedure</article-title><source>Scandinavian Journal of Statistics</source><volume>6</volume><fpage>65</fpage><lpage>70</lpage><pub-id pub-id-type="doi">10.2307/4615733</pub-id></element-citation></ref><ref id="bib42"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Hothorn</surname><given-names>T</given-names></name><name><surname>Bretz</surname><given-names>F</given-names></name><name><surname>Westfall</surname><given-names>P</given-names></name></person-group><year>2008</year><article-title>Simultaneous inference in general parametric models</article-title><source>Biometrical Journal Biometrische Zeitschrift</source><volume>50</volume><fpage>346</fpage><lpage>363</lpage><pub-id pub-id-type="doi">10.1002/bimj.200810425</pub-id></element-citation></ref><ref id="bib43"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Johnson</surname><given-names>MT</given-names></name><name><surname>Lajeunesse</surname><given-names>MJ</given-names></name><name><surname>Agrawal</surname><given-names>AA</given-names></name></person-group><year>2006</year><article-title>Additive and interactive effects of plant genotypic diversity on arthropod communities and plant fitness</article-title><source>Ecology Letters</source><volume>9</volume><fpage>24</fpage><lpage>34</lpage><pub-id pub-id-type="doi">10.1111/j.1461-0248.2005.00833.x</pub-id></element-citation></ref><ref id="bib44"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Johnson</surname><given-names>MT</given-names></name><name><surname>Stinchcombe</surname><given-names>JR</given-names></name></person-group><year>2007</year><article-title>An emerging synthesis between community ecology and evolutionary biology</article-title><source>Trends in Ecology & Evolution</source><volume>22</volume><fpage>250</fpage><lpage>257</lpage><pub-id pub-id-type="doi">10.1016/j.tree.2007.01.014</pub-id></element-citation></ref><ref id="bib45"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Jongsma</surname><given-names>MA</given-names></name><name><surname>Bakker</surname><given-names>PL</given-names></name><name><surname>Stiekema</surname><given-names>WJ</given-names></name></person-group><year>1993</year><article-title>Quantitative determination of serine proteinase inhibitor activity using a radial diffusion assay</article-title><source>Analytical Biochemistry</source><volume>212</volume><fpage>79</fpage><lpage>84</lpage><pub-id pub-id-type="doi">10.1006/abio.1993.1294</pub-id></element-citation></ref><ref id="bib46"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kallenbach</surname><given-names>M</given-names></name><name><surname>Alagna</surname><given-names>F</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name><name><surname>Bonaventure</surname><given-names>G</given-names></name></person-group><year>2010</year><article-title><italic>Nicotiana attenuata</italic> SIPK, WIPK, NPR1, and fatty acid-amino acid conjugates participate in the induction of jasmonic acid biosynthesis by affecting early enzymatic steps in the pathway</article-title><source>Plant Physiology</source><volume>152</volume><fpage>96</fpage><lpage>106</lpage><pub-id pub-id-type="doi">10.1104/pp.109.149013</pub-id></element-citation></ref><ref id="bib47"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kallenbach</surname><given-names>M</given-names></name><name><surname>Bonaventure</surname><given-names>G</given-names></name><name><surname>Gilardoni</surname><given-names>PA</given-names></name><name><surname>Wissgott</surname><given-names>A</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2012</year><article-title><italic>Empoasca</italic> leafhoppers attack wild tobacco plants in a jasmonate-dependent manner and identify jasmonate mutants in natural populations</article-title><source>Proceedings of the National Academy of Sciences of USA</source><volume>109</volume><fpage>E1548</fpage><lpage>E1557</lpage><pub-id pub-id-type="doi">10.1073/pnas.1200363109</pub-id></element-citation></ref><ref id="bib48"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kaplan</surname><given-names>I</given-names></name></person-group><year>2012</year><article-title>Attracting carnivorous arthropods with plant volatiles: the future of biocontrol or playing with fire?</article-title><source>Biological Control</source><volume>60</volume><fpage>77</fpage><lpage>89</lpage><pub-id pub-id-type="doi">10.1016/j.biocontrol.2011.10.017</pub-id></element-citation></ref><ref id="bib49"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kappers</surname><given-names>IF</given-names></name><name><surname>Aharoni</surname><given-names>A</given-names></name><name><surname>Van Herpen</surname><given-names>TW</given-names></name><name><surname>Luckerhoff</surname><given-names>LL</given-names></name><name><surname>Dicke</surname><given-names>M</given-names></name><name><surname>Bouwmeester</surname><given-names>HJ</given-names></name></person-group><year>2005</year><article-title>Genetic engineering of terpenoid metabolism attracts bodyguards to <italic>Arabidopsis</italic></article-title><source>Science</source><volume>309</volume><fpage>2070</fpage><lpage>2072</lpage><pub-id pub-id-type="doi">10.1126/science.1116232</pub-id></element-citation></ref><ref id="bib50"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kessler</surname><given-names>A</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2001</year><article-title>Defensive function of herbivore-induced plant volatile emissions in nature</article-title><source>Science</source><volume>291</volume><fpage>2141</fpage><lpage>2144</lpage><pub-id pub-id-type="doi">10.1126/science.291.5511.2141</pub-id></element-citation></ref><ref id="bib51"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kessler</surname><given-names>A</given-names></name><name><surname>Halitschke</surname><given-names>R</given-names></name><name><surname>Diezel</surname><given-names>C</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2006</year><article-title>Priming of plant defense responses in nature by airborne signaling between <italic>Artemisia tridentata</italic> and <italic>Nicotiana attenuata</italic></article-title><source>Oecologia</source><volume>148</volume><fpage>280</fpage><lpage>292</lpage><pub-id pub-id-type="doi">10.1007/s00442-006-0365-8</pub-id></element-citation></ref><ref id="bib52"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kessler</surname><given-names>A</given-names></name><name><surname>Halitschke</surname><given-names>R</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2004</year><article-title>Silencing the jasmonate cascade: induced plant defenses and insect populations</article-title><source>Science</source><volume>305</volume><fpage>665</fpage><lpage>668</lpage><pub-id pub-id-type="doi">10.1126/science.1096931</pub-id></element-citation></ref><ref id="bib53"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kessler</surname><given-names>D</given-names></name><name><surname>Gase</surname><given-names>K</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2008</year><article-title>Field experiments with transformed plants reveal the sense of floral scents</article-title><source>Science</source><volume>321</volume><fpage>1200</fpage><lpage>1202</lpage><pub-id pub-id-type="doi">10.1126/science.1160072</pub-id></element-citation></ref><ref id="bib54"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kos</surname><given-names>M</given-names></name><name><surname>Van Loon</surname><given-names>J</given-names></name><name><surname>Dicke</surname><given-names>M</given-names></name><name><surname>Vet</surname><given-names>LE</given-names></name></person-group><year>2009</year><article-title>Transgenic plants as vital components of integrated pest management</article-title><source>Trends in Biotechnology</source><volume>27</volume><fpage>621</fpage><lpage>627</lpage><pub-id pub-id-type="doi">10.1016/j.tibtech.2009.08.002</pub-id></element-citation></ref><ref id="bib88"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Köllner</surname><given-names>TG</given-names></name><name><surname>Gershenzon</surname><given-names>J</given-names></name><name><surname>Degenhardt</surname><given-names>J</given-names></name></person-group><year>2009</year><article-title>Molecular and biochemical evolution of maize terpene synthase 10, an enzyme of indirect defense</article-title><source>Phytochemistry</source><volume>70</volume><fpage>1139</fpage><lpage>1145</lpage><pub-id pub-id-type="doi">10.1016/j.phytochem.2009.06.011</pub-id></element-citation></ref><ref id="bib55"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Krügel</surname><given-names>T</given-names></name><name><surname>Lim</surname><given-names>M</given-names></name><name><surname>Gase</surname><given-names>K</given-names></name><name><surname>Halitschke</surname><given-names>R</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2002</year><article-title><italic>Agrobacterium</italic>-mediated transformation of <italic>Nicotiana attenuata</italic>, a model ecological expression system</article-title><source>Chemoecology</source><volume>12</volume><fpage>177</fpage><lpage>183</lpage><pub-id pub-id-type="doi">10.1007/PL00012666</pub-id></element-citation></ref><ref id="bib56"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Long</surname><given-names>HH</given-names></name><name><surname>Sonntag</surname><given-names>DG</given-names></name><name><surname>Schmidt</surname><given-names>DD</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2010</year><article-title>The structure of the culturable root bacterial endophyte community of <italic>Nicotiana attenuata</italic> is organized by soil composition and host plant ethylene production and perception</article-title><source>New Phytologist</source><volume>185</volume><fpage>554</fpage><lpage>567</lpage><pub-id pub-id-type="doi">10.1111/j.1469-8137.2009.03079.x</pub-id></element-citation></ref><ref id="bib57"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Macdonald</surname><given-names>JH</given-names></name></person-group><year>2009</year><source>Handbook of biological statistics</source><edition>2nd edition</edition><publisher-loc>Baltimore, Maryland</publisher-loc><publisher-name>Sparky House Publishing</publisher-name><comment>USA. URL: <ext-link ext-link-type="uri" xlink:href="http://udel.edu/%7Emcdonald/statintro.html">http://udel.edu/∼mcdonald/statintro.html</ext-link></comment></element-citation></ref><ref id="bib58"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Machado</surname><given-names>RA</given-names></name><name><surname>Ferrieri</surname><given-names>AP</given-names></name><name><surname>Robert</surname><given-names>CAM</given-names></name><name><surname>Glauser</surname><given-names>G</given-names></name><name><surname>Kallenbach</surname><given-names>M</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name><name><surname>Erb</surname><given-names>M</given-names></name></person-group><year>2013</year><article-title>Leaf-herbivore attack reduces carbon reserves and regrowth from the roots via jasmonate and auxin signaling</article-title><source>New Phytologist</source><volume>200</volume><fpage>1234</fpage><lpage>1246</lpage><pub-id pub-id-type="doi">10.1111/nph.12438</pub-id></element-citation></ref><ref id="bib59"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Meldau</surname><given-names>DG</given-names></name><name><surname>Long</surname><given-names>HH</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2012</year><article-title>A native plant growth promoting bacterium, <italic>Bacillus</italic> sp. B55, rescues growth performance of an ethylene insensitive plant genotype in nature</article-title><source>Frontiers in Plant Science</source><volume>3</volume><fpage>112</fpage><pub-id pub-id-type="doi">10.3389/fpls.2012.00112</pub-id></element-citation></ref><ref id="bib60"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Midgley</surname><given-names>GF</given-names></name></person-group><year>2012</year><article-title>Ecology. Biodiversity and ecosystem function</article-title><source>Science</source><volume>335</volume><fpage>36</fpage><lpage>38</lpage><pub-id pub-id-type="doi">10.1126/science.1217245</pub-id></element-citation></ref><ref id="bib61"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Moreira</surname><given-names>X</given-names></name><name><surname>Mooney</surname><given-names>KA</given-names></name></person-group><year>2013</year><article-title>Influence of plant genetic diversity on interactions between higher trophic levels</article-title><source>Biology Letters</source><volume>9</volume><fpage>21030133</fpage><pub-id pub-id-type="doi">10.1098/rsbl.2013.0133</pub-id></element-citation></ref><ref id="bib62"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Oh</surname><given-names>Y</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name><name><surname>Gális</surname><given-names>I</given-names></name></person-group><year>2012</year><article-title>NaJAZh regulates a subset of defense responses against herbivores and spontaneous leaf necrosis in <italic>Nicotiana attenuata</italic> plants</article-title><source>Plant Physiology</source><volume>159</volume><fpage>769</fpage><lpage>788</lpage><pub-id pub-id-type="doi">10.1104/pp.112.193771</pub-id></element-citation></ref><ref id="bib63"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Pinheiro</surname><given-names>J</given-names></name><name><surname>Bates</surname><given-names>D</given-names></name><name><surname>DebRoy</surname><given-names>S</given-names></name><name><surname>Sarkar</surname><given-names>D</given-names></name><name><surname>Core Team</surname><given-names>R</given-names></name></person-group><year>2014</year><source>Nlme: linear and nonlinear mixed effects models. R package version 3.1-118</source><comment><ext-link ext-link-type="uri" xlink:href="http://CRAN.R-project.org/package=nlme">http://CRAN.R-project.org/package=nlme</ext-link></comment></element-citation></ref><ref id="bib64"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Press</surname><given-names>WH</given-names></name><name><surname>Dyson</surname><given-names>FJ</given-names></name></person-group><year>2012</year><article-title>Iterated prisoner's dilemma contains strategies that dominate any evolutionary opponent</article-title><source>Proceedings of the National Academy of Sciences of USA</source><volume>109</volume><fpage>10409</fpage><lpage>10413</lpage><pub-id pub-id-type="doi">10.1073/pnas.1206569109</pub-id></element-citation></ref><ref id="bib65"><element-citation publication-type="book"><person-group person-group-type="author"><collab>R Core Team</collab></person-group><year>2012</year><source>R: a Language and environment for statistical computing</source><comment>Available at: <ext-link ext-link-type="uri" xlink:href="http://www.r-project.org">http://www.r-project.org</ext-link></comment></element-citation></ref><ref id="bib66"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Rasmann</surname><given-names>S</given-names></name><name><surname>Köllner</surname><given-names>TG</given-names></name><name><surname>Degenhardt</surname><given-names>J</given-names></name><name><surname>Hiltpold</surname><given-names>I</given-names></name><name><surname>Toepfer</surname><given-names>S</given-names></name><name><surname>Kuhlmann</surname><given-names>U</given-names></name><name><surname>Gershenzon</surname><given-names>J</given-names></name><name><surname>Turlings</surname><given-names>TC</given-names></name></person-group><year>2005</year><article-title>Recruitment of entomopathogenic nematodes by insect-damaged maize roots</article-title><source>Nature</source><volume>434</volume><fpage>732</fpage><lpage>737</lpage><pub-id pub-id-type="doi">10.1038/nature03451</pub-id></element-citation></ref><ref id="bib67"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Robert</surname><given-names>CA</given-names></name><name><surname>Erb</surname><given-names>M</given-names></name><name><surname>Hiltpold</surname><given-names>I</given-names></name><name><surname>Hibbard</surname><given-names>BE</given-names></name><name><surname>Gaillard</surname><given-names>MD</given-names></name><name><surname>Bilat</surname><given-names>J</given-names></name><name><surname>Degenhardt</surname><given-names>J</given-names></name><name><surname>Cambet-Petit-Jean</surname><given-names>X</given-names></name><name><surname>Turlings</surname><given-names>TC</given-names></name><name><surname>Zwahlen</surname><given-names>C</given-names></name></person-group><year>2013</year><article-title>Genetically engineered maize plants reveal distinct costs and benefits of constitutive volatile emissions in the field</article-title><source>Plant Biotechnology Journal</source><volume>11</volume><fpage>1</fpage><lpage>12</lpage><pub-id pub-id-type="doi">10.1111/pbi.12053</pub-id></element-citation></ref><ref id="bib68"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Santhanam</surname><given-names>R</given-names></name><name><surname>Groten</surname><given-names>K</given-names></name><name><surname>Meldau</surname><given-names>DG</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2014</year><article-title>Analysis of plant-bacteria interactions in their native habitat: bacterial communities associated with wild tobacco are independent of endogenous jasmonic acid levels and developmental stages</article-title><source>PLOS ONE</source><volume>9</volume><fpage>e94710</fpage><pub-id pub-id-type="doi">10.1371/journal.pone.0094710</pub-id></element-citation></ref><ref id="bib70"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Schnee</surname><given-names>C</given-names></name><name><surname>Köllner</surname><given-names>TG</given-names></name><name><surname>Held</surname><given-names>M</given-names></name><name><surname>Turlings</surname><given-names>TCJ</given-names></name><name><surname>Gershenzon</surname><given-names>J</given-names></name><name><surname>Degenhardt</surname><given-names>J</given-names></name></person-group><year>2006</year><article-title>The products of a single maize sesquiterpene synthase form a volatile defense signal that attracts natural enemies of maize herbivores</article-title><source>Proceedings of the National Academy of Sciences of USA</source><volume>103</volume><fpage>1129</fpage><lpage>1134</lpage><pub-id pub-id-type="doi">10.1073/pnas.0508027103</pub-id></element-citation></ref><ref id="bib71"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Schuman</surname><given-names>MC</given-names></name><name><surname>Palmer-Young</surname><given-names>EP</given-names></name><name><surname>Schmidt</surname><given-names>A</given-names></name><name><surname>Gershenzon</surname><given-names>J</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2014</year><article-title>Ectopic TPS10 expression enhances sesquiterpene emission in <italic>Nicotiana attenuata</italic> without altering defense or development of transgenic plants or neighbors</article-title><source>Plant Physiology</source><volume>166</volume><fpage>779</fpage><lpage>797</lpage><pub-id pub-id-type="doi">10.1104/pp.114.247130</pub-id></element-citation></ref><ref id="bib90"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Schuman</surname><given-names>MC</given-names></name><name><surname>Allmann</surname><given-names>SA</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2015</year><article-title>Data from: Plant defense phenotypes determine the consequences of volatile emission for individuals and neighbors</article-title><source>Dryad Digital Repository.</source><pub-id pub-id-type="doi">10.5061/dryad.qj007</pub-id></element-citation></ref><ref id="bib72"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Schuman</surname><given-names>MC</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2012</year><article-title>Asking the ecosystem if herbivory-inducible plant volatiles (HIPVs) have defensive functions</article-title><person-group person-group-type="editor"><name><surname>Iason</surname><given-names>GR</given-names></name><name><surname>Dicke</surname><given-names>M</given-names></name><name><surname>Hartley</surname><given-names>SE</given-names></name></person-group><source>The ecology of plant secondary Metaboites</source><edition>1st edition</edition><publisher-loc>Cambridge, UK</publisher-loc><publisher-name>Cambridge University Press</publisher-name><fpage>287</fpage><lpage>307</lpage></element-citation></ref><ref id="bib73"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Schuman</surname><given-names>MC</given-names></name><name><surname>Barthel</surname><given-names>K</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2012</year><article-title>Herbivory-induced volatiles function as defenses increasing fitness of the native plant <italic>Nicotiana attenuata</italic> in nature</article-title><source>eLife</source><volume>1</volume><fpage>e00007</fpage><pub-id pub-id-type="doi">10.7554/eLife.00007</pub-id></element-citation></ref><ref id="bib74"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Schuman</surname><given-names>MC</given-names></name><name><surname>Heinzel</surname><given-names>N</given-names></name><name><surname>Gaquerel</surname><given-names>E</given-names></name><name><surname>Svatos</surname><given-names>A</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2009</year><article-title>Polymorphism in jasmonate signaling partially accounts for the variety of volatiles produced by <italic>Nicotiana attenuata</italic> plants in a native population</article-title><source>The New Phytologist</source><volume>183</volume><fpage>1134</fpage><lpage>1148</lpage><pub-id pub-id-type="doi">10.1111/j.1469-8137.2009.02894.x</pub-id></element-citation></ref><ref id="bib75"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Schuman</surname><given-names>MC</given-names></name><name><surname>Kessler</surname><given-names>D</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2013</year><article-title>Ecological observations of native <italic>Geocoris pallens</italic> and <italic>G. punctipes</italic> populations in the Great Basin Desert of southwestern Utah</article-title><source>Psyche</source><volume>2013</volume><fpage>465108</fpage><pub-id pub-id-type="doi">10.1155/2013/465108</pub-id></element-citation></ref><ref id="bib76"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Shiojiri</surname><given-names>K</given-names></name><name><surname>Ozawa</surname><given-names>R</given-names></name><name><surname>Kugimiya</surname><given-names>S</given-names></name><name><surname>Uefune</surname><given-names>M</given-names></name><name><surname>Van Wijk</surname><given-names>M</given-names></name><name><surname>Sabelis</surname><given-names>MW</given-names></name><name><surname>Takabayashi</surname><given-names>J</given-names></name></person-group><year>2010</year><article-title>Herbivore-specific, density-dependent induction of plant volatiles: honest or ‘cry wolf’ signals?</article-title><source>PLOS ONE</source><volume>5</volume><fpage>e12161</fpage><pub-id pub-id-type="doi">10.1371/journal.pone.0012161</pub-id></element-citation></ref><ref id="bib77"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Steppuhn</surname><given-names>A</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2008</year><article-title>Induced defenses and the cost-benefit paradigm</article-title><person-group person-group-type="editor"><name><surname>Schaller</surname><given-names>A</given-names></name></person-group><source>Induced plant resistance to herbivory</source><publisher-name>Springer Science+Business Media B. V</publisher-name><fpage>61</fpage><lpage>83</lpage></element-citation></ref><ref id="bib78"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Steppuhn</surname><given-names>A</given-names></name><name><surname>Gase</surname><given-names>K</given-names></name><name><surname>Krock</surname><given-names>B</given-names></name><name><surname>Halitschke</surname><given-names>R</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2004</year><article-title>Nicotine's defensive function in nature</article-title><source>PLOS Biology</source><volume>2</volume><fpage>e217</fpage><pub-id pub-id-type="doi">10.1371/journal.pbio.0020217</pub-id></element-citation></ref><ref id="bib79"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Steppuhn</surname><given-names>A</given-names></name><name><surname>Schuman</surname><given-names>MC</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2008</year><article-title>Silencing jasmonate signalling and jasmonate-mediated defences reveals different survival strategies between two <italic>Nicotiana attenuata</italic> accessions</article-title><source>Molecular Ecology</source><volume>17</volume><fpage>3717</fpage><lpage>3732</lpage><pub-id pub-id-type="doi">10.1111/j.1365-294X.2008.03862.x</pub-id></element-citation></ref><ref id="bib80"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Tholl</surname><given-names>D</given-names></name><name><surname>Boland</surname><given-names>W</given-names></name><name><surname>Hansel</surname><given-names>A</given-names></name><name><surname>Loreto</surname><given-names>F</given-names></name><name><surname>Röse</surname><given-names>US</given-names></name><name><surname>Schnitzler</surname><given-names>JP</given-names></name></person-group><year>2006</year><article-title>Practical approaches to plant volatile analysis</article-title><source>The Plant Journal</source><volume>45</volume><fpage>540</fpage><lpage>560</lpage><pub-id pub-id-type="doi">10.1111/j.1365-313X.2005.02612.x</pub-id></element-citation></ref><ref id="bib81a"><element-citation publication-type="book"><person-group person-group-type="author"><collab>United Nations, Department of Economic and Social Affairs, Population Division</collab></person-group><year>2013</year><source>World population prospects: the 2012 revision, key findings and advance tables. Working paper no. ESA/P/WP.227</source><comment><ext-link ext-link-type="uri" xlink:href="http://esa.un.org/unpd/wpp/Documentation/publications.htm">http://esa.un.org/unpd/wpp/Documentation/publications.htm</ext-link></comment></element-citation></ref><ref id="bib82"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>van Dam</surname><given-names>NM</given-names></name><name><surname>Horn</surname><given-names>M</given-names></name><name><surname>Mares</surname><given-names>M</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2001</year><article-title>Ontogeny constrains systemic protease inhibitor response in Nicotiana attenuata</article-title><source>Journal of Chemical Ecology</source><volume>27</volume><fpage>547</fpage><lpage>568</lpage><pub-id pub-id-type="doi">10.1023/A:1010341022761</pub-id></element-citation></ref><ref id="bib83"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Whitham</surname><given-names>TG</given-names></name><name><surname>Bailey</surname><given-names>JK</given-names></name><name><surname>Schweitzer</surname><given-names>JA</given-names></name><name><surname>Shuster</surname><given-names>SM</given-names></name><name><surname>Bangert</surname><given-names>RK</given-names></name><name><surname>Leroy</surname><given-names>CJ</given-names></name><name><surname>Lonsdorf</surname><given-names>EV</given-names></name><name><surname>Allan</surname><given-names>GJ</given-names></name><name><surname>Difazio</surname><given-names>SP</given-names></name><name><surname>Potts</surname><given-names>BM</given-names></name><name><surname>Fischer</surname><given-names>DG</given-names></name><name><surname>Gehring</surname><given-names>CA</given-names></name><name><surname>Lindroth</surname><given-names>RL</given-names></name><name><surname>Marks</surname><given-names>JC</given-names></name><name><surname>Hart</surname><given-names>SC</given-names></name><name><surname>Wimp</surname><given-names>GM</given-names></name><name><surname>Wooley</surname><given-names>SC</given-names></name></person-group><year>2006</year><article-title>A framework for community and ecosystem genetics: from genes to ecosystems</article-title><source>Nature Reviews Genetics</source><volume>7</volume><fpage>510</fpage><lpage>523</lpage><pub-id pub-id-type="doi">10.1038/nrg1877</pub-id></element-citation></ref><ref id="bib84"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Whitham</surname><given-names>TG</given-names></name><name><surname>Gehring</surname><given-names>CA</given-names></name><name><surname>Lamit</surname><given-names>LJ</given-names></name><name><surname>Wojtowicz</surname><given-names>T</given-names></name><name><surname>Evans</surname><given-names>LM</given-names></name><name><surname>Keith</surname><given-names>AR</given-names></name><name><surname>Smith</surname><given-names>DS</given-names></name></person-group><year>2012</year><article-title>Community specificity: life and afterlife effects of genes</article-title><source>Trends in Plant Science</source><volume>17</volume><fpage>271</fpage><lpage>281</lpage><pub-id pub-id-type="doi">10.1016/j.tplants.2012.01.005</pub-id></element-citation></ref><ref id="bib89"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Wu</surname><given-names>J</given-names></name><name><surname>Hettenhausen</surname><given-names>C</given-names></name><name><surname>Meldau</surname><given-names>S</given-names></name><name><surname>Baldwin</surname><given-names>IT</given-names></name></person-group><year>2007</year><article-title>Herbivory rapidly activates MAPK signaling in attacked and unattacked leaf regions but not between leaves of Nicotiana attenuata</article-title><source>The Plant Cell </source><volume>19</volume><fpage>1096</fpage><lpage>1122</lpage><pub-id pub-id-type="doi">10.1105/tpc.106.049353</pub-id></element-citation></ref><ref id="bib85"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Wymore</surname><given-names>AS</given-names></name><name><surname>Keeley</surname><given-names>AT</given-names></name><name><surname>Yturralde</surname><given-names>KM</given-names></name><name><surname>Schroer</surname><given-names>ML</given-names></name><name><surname>Propper</surname><given-names>CR</given-names></name><name><surname>Whitham</surname><given-names>TG</given-names></name></person-group><year>2011</year><article-title>Genes to ecosystems: exploring the frontiers of ecology with one of the smallest biological units</article-title><source>The New Phytologist</source><volume>191</volume><fpage>19</fpage><lpage>36</lpage><pub-id pub-id-type="doi">10.1111/j.1469-8137.2011.03730.x</pub-id></element-citation></ref><ref id="bib86"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Xiao</surname><given-names>Y</given-names></name><name><surname>Wang</surname><given-names>Q</given-names></name><name><surname>Erb</surname><given-names>M</given-names></name><name><surname>Turlings</surname><given-names>TCJ</given-names></name><name><surname>Ge</surname><given-names>L</given-names></name><name><surname>Hu</surname><given-names>L</given-names></name><name><surname>Li</surname><given-names>J</given-names></name><name><surname>Han</surname><given-names>X</given-names></name><name><surname>Zhang</surname><given-names>T</given-names></name><name><surname>Lu</surname><given-names>J</given-names></name><name><surname>Zhang</surname><given-names>G</given-names></name><name><surname>Lou</surname><given-names>Y</given-names></name></person-group><year>2012</year><article-title>Specific herbivore-induced volatiles defend plants and determine insect community composition in the field</article-title><source>Ecology Letters</source><volume>15</volume><fpage>1130</fpage><lpage>1139</lpage><pub-id pub-id-type="doi">10.1111/j.1461-0248.2012.01835.x</pub-id></element-citation></ref><ref id="bib87"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Yan</surname><given-names>ZG</given-names></name><name><surname>Wang</surname><given-names>CZ</given-names></name></person-group><year>2006</year><article-title>Wound-induced green leaf volatiles cause the release of acetylated derivatives and a terpenoid in maize</article-title><source>Phytochemistry</source><volume>67</volume><fpage>34</fpage><lpage>42</lpage><pub-id pub-id-type="doi">10.1016/j.phytochem.2005.10.005</pub-id></element-citation></ref></ref-list><app-group><app id="app1"><title>Appendix 1</title><boxed-text><sec sec-type="appendix" id="s7"><title>Herbivore-induced plant defenses and HIPVs are precisely manipulated in lox2/3, TPS10, and lox2/3xTPS10 plants.</title><p>We confirmed a single insertion of the inverted repeat <italic>LOX2/3</italic> construct in several <italic>lox2/3</italic> lines (<xref ref-type="fig" rid="fig12">Appendix Figure 1A</xref>) and chose lines with single insertions to screen for reduced jasmonic acid (JA) and jasmonoyl isoleucine (JA-Ile) (<xref ref-type="fig" rid="fig12">Appendix Figure 1B,C</xref>).<fig id="fig12" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.027</object-id><label>Appendix Figure 1.</label><caption><title>Transgene insertions and reduced jasmonate accumulation in lines of <italic>lox2/3.</italic></title><p>(<bold>A</bold>) Southern blot of genomic DNA from multiple lines of <italic>lox2/3</italic> digested with the restriction enzymes XbaI or EcoRI. Line 707 is highlighted: this line was used for experiments. A single transgene insertion in <italic>TPS10</italic> (line 10-3) is shown in <xref ref-type="bibr" rid="bib71">Schuman et al. (2014)</xref>. All lines of <italic>lox2/3</italic> screened had strong reductions in their induced (<bold>B</bold>) JA and (<bold>C</bold>) JA-Ile accumulation (mean ± SEM, n = 6). Line 707-2 was used for further experiments. <sup>a,b</sup> Different letters indicate significant differences (corrected p < 0.0001) in Tukey HSD tests following significant (p < 0.0001) 1-way ANOVAs on natural log-transformed data (JA: F<sub>4,25</sub> = 54.15; JA-Ile: F<sub>4,25</sub> = 18.00); ns, not significant in 1-way ANOVAs (p > 0.4).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.027">http://dx.doi.org/10.7554/eLife.04490.027</ext-link></p></caption><graphic xlink:href="elife04490f012"/></fig></p><p>Based on these data, we chose <italic>lox2/3</italic> line 707-2 for further experiments overexpressing <italic>TPS10</italic> plants (line 10-3) also harbor a single transgene insertion (<xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>), and both lines were also confirmed to be diploid by flow cytometry. All transgenic lines showed the intended reduction or increase in target gene transcripts: <italic>LOX2</italic> and <italic>LOX3</italic> transcripts accumulated to <7% of WT levels in <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic>, while <italic>TPS10</italic> transcripts were easily quantifiable in <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic>, but undetectable in WT or <italic>lox2/3</italic> (p-values < 0.01 for pairwise differences in Tukey HSD tests following p-values < 0.001 in 1-way ANOVAs on natural log-transformed data, <xref ref-type="fig" rid="fig13">Appendix Figure 2</xref>). Transcript abundance of <italic>LOX2</italic> and <italic>LOX3</italic> was similar in WT and <italic>TPS10</italic> (p = 1 in Tukey HSD tests) and in <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> (p = 0.471 for <italic>LOX2</italic> and p = 1 for <italic>LOX3</italic>); <italic>TPS10</italic> transcript abundance was similar in <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> (p = 0.157).<fig id="fig13" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.028</object-id><label>Appendix Figure 2.</label><caption><title>Target gene transcript levels in <italic>lox2/3</italic>, <italic>TPS10</italic>, and <italic>lox2/3</italic>x<italic>TPS10</italic> plants.</title><p>Transcript accumulation shows that <italic>lox2/3</italic> plants, <italic>TPS10</italic> plants, and hemizygous crosses (<italic>lox2/3xTPS10</italic>) have the expected silencing or accumulation of the target genes <italic>LOX2</italic>, <italic>LOX3</italic>, and <italic>TPS10</italic> (mean ± SEM, n = 4). <sup>a,b</sup> Different letters indicate significant differences (corrected p < 0.01) in Tukey HSD tests following a significant (p < 0.001) 1-way ANOVA on natural log-transformed data (<italic>LOX2</italic>: F<sub>3,12</sub> = 49.49, p < 0.001; <italic>LOX3</italic>: F<sub>3,12</sub> = 13.73, p < 0.001; <italic>TPS10</italic>: F<sub>3,12</sub> = 845.6, p < 0.001).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.028">http://dx.doi.org/10.7554/eLife.04490.028</ext-link></p></caption><graphic xlink:href="elife04490f013"/></fig></p><p>As a consequence of reduced <italic>LOX3</italic> transcripts (<xref ref-type="bibr" rid="bib5">Allmann et al., 2010</xref>), <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants accumulated ≤30% of WT JA levels in a glasshouse experiment in which leaves were treated with wounding plus distilled water (W + W) or wounding plus the addition of <italic>Manduca sexta</italic> oral secretions diluted in distilled water (W + OS) to mimic herbivore attack in a precisely timed manner (<xref ref-type="fig" rid="fig1">Figure 1B</xref>, see also overview of phenotypes in <xref ref-type="fig" rid="fig1">Figure 1A</xref>). Induced levels of the active jasmonate, JA-Ile, were also reduced to <50% of WT. Absolute values for JA and JA-Ile, and significant pairwise differences in Tukey HSD tests (p<italic>-</italic>values < 0.05) following two-way ANOVAs on natural log-transformed data (p-values < 0.002), are shown in <xref ref-type="supplementary-material" rid="SD1-data">Figure 1—source data 1</xref>. Likewise, reduced <italic>LOX2</italic> transcripts resulted in diminished GLV emissions from <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants after W + W or W + OS treatment to <2% of WT levels (<xref ref-type="fig" rid="fig1">Figure 1B</xref>, <xref ref-type="supplementary-material" rid="SD1-data">Figure 1—source data 1</xref>). Significant pairwise differences in Wilcoxon rank sum tests (Holm-Bonferroni-corrected p<italic>-</italic>values < 0.03) following Kruskal–Wallis tests (corrected p-values ≤ 0.020) are shown in <xref ref-type="supplementary-material" rid="SD1-data">Figure 1—source data 1</xref>. In contrast, GLV emission and JA and JA-Ile accumulation were similar in WT and <italic>TPS10</italic> plants (<xref ref-type="fig" rid="fig1">Figure 1</xref>, <xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>), but both <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants emitted elevated amounts of TAB and TBF, more than 10-fold WT levels as measured in a second glasshouse experiment (<xref ref-type="fig" rid="fig1">Figure 1B</xref>, <xref ref-type="supplementary-material" rid="SD2-data">Figure 1—source data 2</xref>). Genotypes with the <italic>TPS10</italic> overexpression construct vs those without differed significantly in their emission (Wilcoxon rank sum tests, corrected p-values ≤ 0.009, shown in <xref ref-type="supplementary-material" rid="SD1-data">Figure 1—source data 2</xref>).</p><p>As in the glasshouse experiments, <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants in the field had strongly reduced jasmonate-mediated defense and GLV emission, to <4% of WT levels; and <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants emitted elevated levels of TAB and TBF, >18× WT levels (<xref ref-type="fig" rid="fig1">Figure 1C</xref>; data reported in the following subsection). The reductions in GLVs, JA, and JA-Ile described here for <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> are similar to what has been previously reported for homozygous transformed lines silenced either in <italic>LOX2</italic> (GLVs) or <italic>LOX3</italic> (JA, JA-Ile) by separate, specific RNAi constructs (<xref ref-type="bibr" rid="bib5">Allmann et al., 2010</xref>).</p><p>In summary, JA, JA-Ile, jasmonate-mediated defense, and GLVs were significantly reduced in <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants to <4% of WT levels, and <italic>lox2/3</italic>x<italic>TPS10</italic> and <italic>TPS10</italic> plants produced more than 10 times as much TAB and TBF as WT or <italic>lox2/3</italic> plants; these phenotypes were robust in plants grown under controlled glasshouse conditions and in the field (<xref ref-type="fig" rid="fig1">Figure 1</xref> and supplements, <xref ref-type="app" rid="app2">Appendix 2</xref>). This set of transgenic plants (<xref ref-type="fig" rid="fig1">Figure 1A</xref>) allowed us to manipulate TAB and TBF independently of defense signaling and all other HIPVs, creating plants with enhanced emission in a well-defended (WT, <italic>TPS10</italic>) or poorly defended (<italic>lox2/3</italic>, <italic>lox2/3</italic>x<italic>TPS10</italic>) background.</p></sec></boxed-text></app><app id="app2"><title>Appendix 2</title><boxed-text><sec sec-type="appendix" id="s8"><title>Phenotypic measurements of transgenic lines in experimental field-grown populations.</title><p>Trypsin protease inhibitor (TPI) activity, a well-known jasmonate-mediated defense (<xref ref-type="bibr" rid="bib82">van Dam et al., 2001</xref>; <xref ref-type="bibr" rid="bib34">Halitschke and Baldwin, 2003</xref>), is more a stable trait than are JA levels in frozen tissue and thus was used as an indicator of jasmonate-mediated defense levels in field-harvested samples. TPI activity was reduced to <4% of WT activity in <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants grown in the field. A Kruskal–Wallis test among genotypes indicated a highly significant difference (χ<sup>2</sup><sub>3</sub> = 47.87, corrected p < 0.001), as did Wilcoxon rank sum tests between WT and <italic>lox2/3</italic>, and WT and <italic>lox2/3</italic>x<italic>TPS10</italic> (W<sub>24,12</sub> = 0, W<sub>24,11</sub> = 0 and corrected p<italic>-</italic>values < 0.001, <xref ref-type="fig" rid="fig1">Figure 1C</xref>, <xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref>). There was no significant difference in TPI activity between WT and <italic>TPS10</italic> (W<sub>24,21</sub> = 357, corrected p = 0.065), as was previously shown in glasshouse experiments (<xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>).</p><p>GLV pools extracted from leaf tissue samples of field-grown <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> plants were <3% of those found in leaf tissue from field-grown WT plants (corrected p<italic>-</italic>values < 0.001 in Wilcoxon rank sum tests), whereas <italic>TPS10</italic> and WT did not differ (corrected p<italic>-</italic>values = 1 in Wilcoxon rank sum tests, <xref ref-type="fig" rid="fig1">Figure 1C</xref>). (<italic>E</italic>)-Hexen-2-al and (<italic>Z</italic>)-hexen-3-ol were the only GLVs found in extracts (see also <xref ref-type="bibr" rid="bib73">Schuman et al., 2012</xref>), and both differed significantly by genotype. Significant differences in pairwise comparisons between <italic>lox2/3</italic>, <italic>lox2/3</italic>x<italic>TPS10</italic>, <italic>TPS10</italic> and WT in Wilcoxon rank sum tests (corrected p < 0.001) following Kruskal–Wallis tests (corrected p-values < 0.001) are shown in <xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref>. There was no significant effect of plant type or population type on extracted GLVs (corrected p<italic>-</italic>values = 1 in Kruskal–Wallis tests across individuals, edge and center plants, or across individuals, mono-, and mixed cultures). Emission of (<italic>Z</italic>)-hexen-3-ol, the only GLV detected from plants having natural background levels of herbivore attack in season one, was also reduced in <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> to <16% of WT emission (<xref ref-type="table" rid="tbl1">Appendix Table 1</xref>). Daytime emission of (<italic>Z</italic>)-hexen-3-ol was low and differed only marginally in a Kruskal–Wallis test among genotypes (χ<sup>2</sup><sub>3</sub> = 9.27, corrected p = 0.052), but nighttime emission differed significantly in a Wilcoxon rank sum test between plants with and without the <italic>lox2/3</italic> silencing construct (WT and <italic>TPS10</italic> v. <italic>lox2/3</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic>, W<sub>15,15</sub> = 171.5, corrected p = 0.035). No significant effect of population type on emission was observed (Wilcoxon rank sum tests at day or night of individuals vs center plants in monocultures, corrected p<italic>-</italic>values > 0.07).</p><table-wrap id="tbl1" position="float"><object-id pub-id-type="doi">10.7554/eLife.04490.029</object-id><label>Appendix Table 1.</label><caption><p>Volatiles (percent IS plant<sup>−1</sup>, mean ± SEM) trapped in the headspace around single plants in experimental season one (June 8th–9th).</p><p><bold>DOI:</bold> <ext-link ext-link-type="doi" xlink:href="10.7554/eLife.04490.029">http://dx.doi.org/10.7554/eLife.04490.029</ext-link></p></caption><table frame="hsides" rules="groups"><thead><tr><th colspan="3" rowspan="2"> </th><th colspan="8" align="center">GLVs (percent IS plant<sup>-1</sup>)</th><th colspan="17" align="center">TPS10 products (percent IS plant<sup>-1</sup>)</th><th colspan="10" align="center">Non-target volatiles (percent IS plant<sup>-1</sup>)</th></tr><tr><th colspan="8" align="center">(<italic>Z</italic>)-Hexen-3-ol</th><th colspan="10" align="center">TAB</th><th colspan="7" align="center">TBF</th><th colspan="6" align="center">α-Duprezianene</th><th colspan="4" align="center">Germacrene A</th></tr></thead><tbody><tr><td><bold>Genotype</bold></td><td><bold>Day n</bold></td><td><bold>Night n</bold></td><td colspan="3" style="background:silver"><bold>Day</bold></td><td colspan="5" style="background:#8C8C8C"><bold>Night</bold></td><td colspan="4" style="background:silver"><bold>Day</bold></td><td colspan="6" style="background:#8C8C8C"><bold>Night</bold></td><td colspan="4" style="background:silver"><bold>Day</bold></td><td colspan="3" style="background:#8C8C8C"><bold>Night</bold></td><td colspan="3" style="background:silver"><bold>Day</bold></td><td colspan="3" style="background:#8C8C8C"><bold>Night</bold></td><td colspan="3" style="background:silver"><bold>Day</bold></td><td style="background:#8C8C8C"><bold>Night</bold></td></tr><tr><td><bold>WT</bold></td><td>8</td><td>8</td><td style="background:silver">0.99%</td><td style="background:silver">±</td><td style="background:silver">0.99%</td><td style="background:#8C8C8C">7.96%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">4.25%</td><td rowspan="2" style="background:#8C8C8C"><named-content content-type="author-callout-style1">}</named-content></td><td rowspan="2" style="background:#8C8C8C"><bold><named-content content-type="author-callout-style1">a</named-content></bold><sub><bold><named-content content-type="author-callout-style1">lox</named-content></bold></sub></td><td style="background:silver">0.37%</td><td style="background:silver">±</td><td style="background:silver">0.29%</td><td style="background:silver"><bold><named-content content-type="author-callout-style1">a</named-content></bold></td><td colspan="3" style="background:#8C8C8C">—</td><td style="background:#8C8C8C"> </td><td rowspan="3" style="background:#8C8C8C"><named-content content-type="author-callout-style1">}</named-content></td><td rowspan="2" style="background:#8C8C8C"><bold><named-content content-type="author-callout-style1">a</named-content></bold><sub><bold><named-content content-type="author-callout-style1">TPS</named-content></bold></sub></td><td colspan="3" style="background:silver">—</td><td style="background:silver"><bold><named-content content-type="author-callout-style1">a</named-content></bold></td><td colspan="3" style="background:#8C8C8C">—</td><td style="background:silver">2.13%</td><td style="background:silver">±</td><td style="background:silver">0.85%</td><td style="background:#8C8C8C">0.96%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.42%</td><td style="background:silver">1.68%</td><td style="background:silver">±</td><td style="background:silver">0.98%</td><td style="background:#8C8C8C">—</td></tr><tr><td><bold><italic>TPS10</italic></bold></td><td>7</td><td>7</td><td style="background:silver">2.37%</td><td style="background:silver">±</td><td style="background:silver">1.55%</td><td style="background:#8C8C8C">2.84%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.63%</td><td style="background:silver">18.97%</td><td style="background:silver">±</td><td style="background:silver">6.03%</td><td style="background:silver"><bold><named-content content-type="author-callout-style2">b</named-content></bold></td><td style="background:#8C8C8C">0.94%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.53%</td><td rowspan="3" style="background:#8C8C8C"><named-content content-type="author-callout-style2">}</named-content></td><td style="background:silver">9.34%</td><td style="background:silver">±</td><td style="background:silver">3.44%</td><td style="background:silver"><bold><named-content content-type="author-callout-style2">b</named-content></bold></td><td style="background:#8C8C8C">0.20%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.20%</td><td style="background:silver">9.47%</td><td style="background:silver">±</td><td style="background:silver">5.26%</td><td style="background:#8C8C8C">0.94%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.26%</td><td style="background:silver">2.78%</td><td style="background:silver">±</td><td style="background:silver">1.52%</td><td style="background:#8C8C8C">—</td></tr><tr><td><bold><italic>lox2/3</italic></bold></td><td>7</td><td>8</td><td style="background:silver">0.13%</td><td style="background:silver">±</td><td style="background:silver">0.13%</td><td style="background:#8C8C8C">1.06%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.64%</td><td rowspan="2" style="background:#8C8C8C"><named-content content-type="author-callout-style2">}</named-content></td><td rowspan="2" style="background:#8C8C8C"><bold><named-content content-type="author-callout-style2">b</named-content></bold><sub><bold><named-content content-type="author-callout-style2">lox</named-content></bold></sub></td><td colspan="3" style="background:silver">—</td><td style="background:silver"><bold><named-content content-type="author-callout-style1">a</named-content></bold></td><td style="background:#8C8C8C">0.15%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.15%</td><td rowspan="2" style="background:#8C8C8C"><bold><named-content content-type="author-callout-style2">b</named-content></bold><sub><bold><named-content content-type="author-callout-style2">TPS</named-content></bold></sub></td><td colspan="3" style="background:silver">—</td><td style="background:silver"><bold><named-content content-type="author-callout-style1">a</named-content></bold></td><td colspan="3" style="background:#8C8C8C">—</td><td style="background:silver">2.75%</td><td style="background:silver">±</td><td style="background:silver">1.12%</td><td style="background:#8C8C8C">0.60%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.19%</td><td style="background:silver">1.68%</td><td style="background:silver">±</td><td style="background:silver">1.49%</td><td style="background:#8C8C8C">—</td></tr><tr><td><bold><italic>lox2/3</italic>x<italic>TPS10</italic></bold></td><td>7</td><td>7</td><td style="background:silver">0.07%</td><td style="background:silver">±</td><td style="background:silver">0.07%</td><td style="background:#8C8C8C">1.24%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.84%</td><td style="background:silver">7.39%</td><td style="background:silver">±</td><td style="background:silver">2.56%</td><td style="background:silver"><bold><named-content content-type="author-callout-style2">b</named-content></bold></td><td style="background:#8C8C8C">2.08%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.84%</td><td style="background:#8C8C8C"> </td><td style="background:silver">4.47%</td><td style="background:silver">±</td><td style="background:silver">1.70%</td><td style="background:silver"><bold><named-content content-type="author-callout-style2">b</named-content></bold></td><td style="background:#8C8C8C">0.40%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.40%</td><td style="background:silver">3.02%</td><td style="background:silver">±</td><td style="background:silver">1.42%</td><td style="background:#8C8C8C">0.73%</td><td style="background:#8C8C8C">±</td><td style="background:#8C8C8C">0.31%</td><td style="background:silver">0.66%</td><td style="background:silver">±</td><td style="background:silver">0.37%</td><td style="background:#8C8C8C">—</td></tr></tbody></table><table-wrap-foot><fn><p><named-content content-type="author-callout-style1">a</named-content>, <named-content content-type="author-callout-style2">b</named-content> Different letters indicate significant differences between lines for a vertical category (corrected <italic>P</italic><0.05) in Wilcoxon rank sum tests following significant Kruskal-Wallis tests across all genotypes (see Results text); <italic>P</italic>-values were corrected for multiple testing using the Holm-Bonferroni method. Community type was also tested and found not significant. Where there are no letters, there are no significant pairwise differences within a category.</p></fn><fn><p><named-content content-type="author-callout-style1">}</named-content> <named-content content-type="author-callout-style1">a</named-content><sub><named-content content-type="author-callout-style1">lox</named-content></sub>, <named-content content-type="author-callout-style2">}</named-content> <named-content content-type="author-callout-style2">b</named-content><sub><named-content content-type="author-callout-style2">lox</named-content></sub> / <named-content content-type="author-callout-style1">}</named-content> <named-content content-type="author-callout-style1">a</named-content><sub><named-content content-type="author-callout-style1">TPS</named-content></sub>, <named-content content-type="author-callout-style2">}</named-content> <named-content content-type="author-callout-style2">b</named-content><sub><named-content content-type="author-callout-style2">TPS</named-content></sub> Indicate significant differences (corrected P<0.05) between lines with and without either the <italic>lox</italic> construct (WT and <italic>TPS</italic> v. <italic>lox</italic> and <italic>loxTPS</italic>) or the TPS construct (WT and <italic>lox</italic> v. <italic>TPS</italic> and <italic>loxTPS</italic>) in Wilcoxon rank sum tests for a category; P-values were corrected for multiple testing using the Holm-Bonferroni method. There are no significant pairwise differences in these categories.</p></fn><fn><p>IS, internal standard; GLVs, green leaf volatiles; TAB, (<italic>E</italic>)-α-bergamotene; TBF, (<italic>E</italic>)-β-farnesene; —, not detected.</p></fn></table-wrap-foot></table-wrap><p>More than 18 times as much TAB and TBF were measured in <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> tissue extracts than in WT tissue extracts from field-grown plants. Significant differences (corrected p<italic>-</italic>values < 0.04) in Wilcoxon rank sum tests between individual genotypes following Kruskal–Wallis tests (corrected p-values < 0.001) are shown in <xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref>. There was no significant effect of plant type or population type on extracted TAB and TBF (corrected p<italic>-</italic>values > 0.8 in Kruskal–Wallis tests across individuals, edge and center plants, or across individuals, mono-, and mixed cultures). Emission of TAB and TBF from plants with natural background levels of herbivore attack in season one was also increased in <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic>, to more than 20-fold WT levels. During the day, when emission of TAB and TBF was highest, Kruskal–Wallis tests indicated highly significant differences among genotypes (TAB, χ<sup>2</sup><sub>3</sub> = 20.90, corrected p = 0.008; TBF, χ<sup>2</sup><sub>3</sub> = 19.63, corrected p = 0.001); significant pairwise differences (corrected p < 0.05) in Wilcoxon rank sum tests are shown in <xref ref-type="table" rid="tbl1">Appendix Table 1</xref>. At night, when TAB and TBF emission levels were 10- to 20-fold lower overall, genotypes with and without the <italic>TPS10</italic> overexpression construct still differed significantly in TAB emission in a Wilcoxon rank sum test (WT and <italic>lox2/3</italic> v. <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic>, W<sub>15,15</sub> = 52.5, corrected p = 0.022). No significant effect of population type on emission was observed (Wilcoxon rank sum tests at day or night of individuals vs center plants in monocultures, corrected p<italic>-</italic>values > 0.3).</p><p>In summary, all measures of GLV emission, jasmonate-mediated defense, and TAB and TBF emission revealed that the engineered phenotypes were stable and, except for open headspace measurements from <italic>lox2/3</italic> plants in mixed cultures, unaffected by population type in field experiments (<xref ref-type="fig" rid="fig1">Figure 1C</xref>, <xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref>, <xref ref-type="table" rid="tbl1">Appendix Table 1</xref>). Furthermore, non-target volatiles (phenylpropanoids, monoterpenes, and the sesquiterpenes germacrene A and α-duprezianene) were not affected by transformation or population type (<xref ref-type="supplementary-material" rid="SD3-data">Figure 1—source data 3</xref>, <xref ref-type="table" rid="tbl1">Appendix Table 1</xref>). Thus, there was no evidence for direct effects of TPS10 volatiles on neighbor plants, with the exception of headspace contamination (<xref ref-type="fig" rid="fig3">Figure 3</xref>) which could be due either to diffusion or adsorption and re-release of TPS10 volatiles.</p></sec></boxed-text></app><app id="app3"><title>Appendix 3</title><boxed-text><sec sec-type="appendix" id="s9"><title/><p>Screening of transgenic plant lines, quantification of TPI activity in field tissue samples, and quantification of volatiles in the headspace of field-grown plants (data shown in <xref ref-type="app" rid="app1 app2">Appendices 1, 2</xref>); and statistical models used to test the hypotheses that <italic>LOX2/3</italic> deficiency, <italic>TPS10</italic> expression, plant position, and population type were associated with differences in foliar herbivore damage (data shown in <xref ref-type="fig" rid="fig5">Figure 5</xref>) and foliar herbivore abundance (data shown in <xref ref-type="fig" rid="fig6">Figure 6</xref>).</p><sec id="s9-1"><title>Screening of transgenic plant lines</title><p>Transgenic lines were screened based on the procedure described by <xref ref-type="bibr" rid="bib32">Gase et al. (2011)</xref>. Briefly, plants likely to contain a single transgene insertion were selected based on the segregation of hygromycin resistance in the T1 generation (the <italic>HYGROMYCIN PHOSPHOTRANSFERASE II</italic> [<italic>HPTII</italic>] gene from pCAMBIA-1301 is contained in the pSOL vector). High-quality genomic DNA was extracted using a modified cetyltrimethylammonium bromide (CTAB) method (<xref ref-type="bibr" rid="bib14">Bubner et al., 2006</xref>) from non-senescent leaf tissue collected from transformed plants (see ‘Tissue samples from glasshouse- and field-grown plants’). Single transgene insertions were confirmed by Southern blotting of genomic DNA digested with EcoRI and XbaI using a <sup>32</sup><italic>P-</italic>labeled probe binding to <italic>HPTII</italic> (<italic>lox2/3</italic> shown in <xref ref-type="fig" rid="fig12">Appendix Figure 1</xref>, <italic>TPS10</italic> in <xref ref-type="bibr" rid="bib71">Schuman et al., 2014</xref>). Flow cytometric analysis (described by <xref ref-type="bibr" rid="bib14">Bubner et al., 2006</xref>) confirmed that all transgenic lines were diploid. Homozygosity of T2 plants was determined by screening for resistance to hygromycin.</p><p>The accumulation or reduction of target gene transcripts was quantified in leaf samples (n = 4) harvested 1 h after treatment with <italic>M. sexta</italic> oral secretions (W + OS, see ‘Tissue samples from glasshouse- and field-grown plants’ in ‘Materials and methods’). Total RNA was extracted from 100 mg tissue using TRI reagent (Invitrogen, Life Technologies, Carlsbad, CA) according to the manufacturer's instructions. Quality was checked on a 1% agarose gel and concentration was determined by absorbance at 260 nm with a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). Synthesis of cDNA from 0.5 μg of total RNA per sample and SYBR Green qPCR analyses were conducted as in Wu et al. (2007) using reagents from Fermentas, Waltham, MA, a Mastercycler (Eppendorf, Germany) and an Mx3005P qPCR system (Stratagene La Jolla, CA) and qPCR Core Kit for SYBR Green I (Eurogentec, Belgium). Transcripts were quantified using external standard curves for each gene, and <italic>N. attenuata ELONGATION FACTOR 1A</italic> (<italic>Elf1A</italic>) transcript abundance in each sample was used to normalize total cDNA concentration variations. Samples of RNA used to make cDNA were pooled to the same dilution as in cDNA samples and run alongside cDNA in all qPCRs to control for gDNA contamination; no contamination was detected. The primers for each gene have been described previously: <italic>LOX2</italic>, <xref ref-type="bibr" rid="bib5">Allmann et al. (2010)</xref>; <italic>LOX3</italic>, <xref ref-type="bibr" rid="bib46">Kallenbach et al. (2010)</xref>; and <italic>TPS10</italic> and <italic>Ef1A</italic>, <xref ref-type="bibr" rid="bib71">Schuman et al. (2014)</xref>.</p><p>Analysis of JA and JA-Ile is described under ‘Quantification of JA and JA-Ile in leaf tissue’ in ‘Materials and methods’.</p></sec><sec id="s9-2"><title>Quantification of TPI activity in field tissue samples</title><p>TPI activity (nmol mg protein<sup>−1</sup>) was quantified in 100 mg of frozen tissue from field samples (n = 11–24) using a radial diffusion assay as previously described (<xref ref-type="bibr" rid="bib82">van Dam et al., 2001</xref>) based on the protocol from <xref ref-type="bibr" rid="bib45">Jongsma et al. (1993)</xref> and using materials from Sigma–Aldrich and a radial plate sufficiently large to simultaneously accommodate all samples and two rows of standards, one at the start and one at the end of the plate. For the statistical analysis approach, see ‘Statistical analyses’ in ‘Materials and methods’.</p></sec><sec id="s9-3"><title>Quantification of volatiles in the headspace of field-grown plants</title><p>For volatile trapping in field experiments, we used single-use activated charcoal filters (ORBO-32 standard; Supelco, Sigma-Aldrich) shielded with aluminum foil from UV radiation and connected to self-made ozone scrubbers at the air inlet that contained eight-ply of 65 mm-diameter MnO2-coated copper gauze (OBE Corporation, Fredericksburg, TX) to prevent the oxidation of volatiles by ozone (<xref ref-type="bibr" rid="bib15">Calogirou et al., 1996</xref>). Ambient air was pulled through the charcoal filter by a vacuum pump (DAA-V114-GB; Gast, Benton Harbor, MI) powered by a car battery. After trapping, charcoal filters were sealed with provided caps and stored at −20°C until transportation to the MPICE, elution, and analysis. Filters were spiked with tetralin internal standard (IS) and eluted as described under ‘Quantification of volatiles in the headspace of glasshouse-grown plants’.</p><p>For the volatile data in <xref ref-type="table" rid="tbl1">Appendix Table 1</xref>, heat- and light-resistant transparent plastic cones with open tops (frost protection cones, company) were placed over individual plants and plants at the centers of monocultures (n = 7–8 plants per genotype) on June 8th in season one, and the charcoal filter was placed at the base of the plant such that air flowed in at the top of the cone, through the headspace, and out through the filter. Airflow was started at 8:45 am. Volatiles were trapped for a total of 22 h and filters were exchanged after 12 h, resulting in two samples from each plant for the periods from 9 am to 9 pm (day) and 9 pm to 7 am (stopped before daylight shone on plants, night). Samples were analyzed by a Varian C<italic>P-</italic>3800 GC-Saturn 4000 ion trap MS connected to a ZB5 column (30 m × 0.25 mm i.d., 0.25 μm film thickness; Phenomenex) as described in ‘Quantification of volatiles in the headspace of glasshouse-grown plants’. Analyte quantities were expressed as percent tetralin IS per plant. For the statistical analysis approach, see ‘Statistical analyses’ in ‘Materials and methods’.</p></sec><sec id="s9-4"><title>Statistical models</title><sec id="s9-4-1"><title>Statistical models used for data in <xref ref-type="fig" rid="fig5">Figure 5</xref></title><p>We tested the hypotheses that <italic>LOX2/3</italic> deficiency, <italic>TPS10</italic> expression, plant position, and population type were associated with differences in foliar herbivore damage. ANOVAs or linear mixed-effects models were performed on arcsine-transformed data. Models were fit by maximum likelihood and model reduction was performed by removing least significant terms and comparing the Akaike information criterion (AIC) value of more complex vs simplified models (<xref ref-type="bibr" rid="bib1">Akaike, 1973</xref>) until only significant terms were left. Linear mixed-effects models were re-fit by restricted maximum likelihood (REML). Genotypes were coded as follows: 1WT, 2TPS10, 3lox23, 4lox23xTPS10; <italic>LOX2/3</italic> deficiency was coded as 1WT, 2 lox; <italic>TPS10</italic> expression was coded as 1WT, 2TPS; plant positions were coded as 1Individual, 2Edge, 3Center; and monocultures vs mixed cultures were coded as 1Mono, 2Mix. All minimal models to explain differences in herbivore damage retained <italic>LOX2/3</italic> deficiency as the only significant factor, although there was a marginal interaction between <italic>LOX2/3</italic> deficiency and plant position in season one which resulted in a poorer AIC value when removed.</p><p>For individual plants, the following initial ANOVA model was used:</p><p>model1<-aov(Arcsin_total∼lox*TPS,ind),</p><p>and the final model was</p><p>model3<-aov(Arcsin_total∼lox,ind)</p><p>with the following results:<table-wrap id="tblu1" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th/><th>Df</th><th>Sum sq</th><th>Mean sq</th><th>F</th><th>p</th></tr></thead><tbody><tr><td colspan="6">Season 1</td></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>1</td><td>0.1008</td><td>0.1008</td><td>11.47</td><td><bold>0.0015</bold></td></tr><tr><td> Residuals</td><td>44</td><td>0.3864</td><td>0.0088</td><td/><td/></tr><tr><td colspan="6">Season 2</td></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>1</td><td>0.2321</td><td>0.2321</td><td>14.64</td><td><bold>0.0005</bold></td></tr><tr><td> Residuals</td><td>39</td><td>0.6183</td><td>0.0159</td><td/><td/></tr></tbody></table></table-wrap></p><p>For edge plants, a linear mixed-effects model was used to test the effects of <italic>LOX2/3</italic> deficiency and <italic>TPS10</italic> expression, so that the analysis of multiple plants per population could be included as a random blocking factor (lme function in R package nlme, <xref ref-type="bibr" rid="bib63">Pinheiro et al., 2014</xref>). The initial model was</p><p>model1<-lme(Arcsin_total∼lox*TPS,random=∼1|PopID,edge,method=“ML”)</p><p>and the final model was</p><p>model3REML<-lme(Arcsin_total∼lox,random=∼1|PopID,edge)</p><p>with the following results (fixed effects):<table-wrap id="tblu2" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th/><th>Value</th><th>Std. error</th><th>DF</th><th>t</th><th>p</th><th>(Intr)</th></tr></thead><tbody><tr><td colspan="7">Season 1</td></tr><tr><td> (Intercept)</td><td>0.1929</td><td>0.0102</td><td>200</td><td>18.86</td><td><bold><0.0001</bold></td><td/></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>0.0895</td><td>0.0145</td><td>70</td><td>6.152</td><td><bold><0.0001</bold></td><td>−0.704</td></tr><tr><td colspan="7">Season 2</td></tr><tr><td> (Intercept)</td><td>0.31</td><td>0.02</td><td>119</td><td>20.40</td><td><bold><0.0001</bold></td><td/></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>0.13</td><td>0.02</td><td>74</td><td>5.501</td><td><bold><0.0001</bold></td><td>−0.641</td></tr></tbody></table></table-wrap></p><p>Furthermore, WT and <italic>lox2/3</italic> edge plants were tested to determine whether levels of herbivore damage differed with the presence or absence of a <italic>TPS10</italic>-expressing plant in the population, starting with the following model:</p><p>model1<-lme(Arcsin_total∼lox*MonoMix,random=∼1|PopID,edgeWTlox,method=“ML”)</p><p>and the final model was</p><p>model6REML<-lme(Arcsin_total∼lox,random=∼1|PopID,edgeWTlox),</p><p>with the following results (fixed effects):<table-wrap id="tblu3" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th/><th>Value</th><th>Std. error</th><th>DF</th><th>t</th><th>p</th><th>(Intr)</th></tr></thead><tbody><tr><td colspan="7">Season 1</td></tr><tr><td> (Intercept)</td><td>0.1947</td><td>0.0125</td><td>130</td><td>15.60</td><td><bold><0.0001</bold></td><td/></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>0.0945</td><td>0.0178</td><td>46</td><td>5.309</td><td><bold><0.0001</bold></td><td>−0.702</td></tr><tr><td colspan="7">Season 2</td></tr><tr><td> (Intercept)</td><td>0.3132</td><td>0.0194</td><td>71</td><td>16.16</td><td><bold><0.0001</bold></td><td/></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>0.1454</td><td>0.0307</td><td>46</td><td>4.732</td><td><bold><0.0001</bold></td><td>−0.631</td></tr></tbody></table></table-wrap></p><p>For center plants, the following initial ANOVA model was used to test the effects of <italic>LOX2/3</italic> deficiency and <italic>TPS10</italic> expression:</p><p>model1<-aov(Arcsin_total∼lox*TPS,center)</p><p>and the final model was</p><p>model3<-aov(Arcsin_total∼lox,center)</p><p>with the following results:<table-wrap id="tblu4" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th/><th>Df</th><th>Sum sq</th><th>Mean sq</th><th>F</th><th>p</th></tr></thead><tbody><tr><td colspan="6">Season 1</td></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>1</td><td>0.1632</td><td>0.16325</td><td>23.57</td><td><bold>8.28E-06</bold></td></tr><tr><td> Residuals</td><td>63</td><td>0.4363</td><td>0.00692</td><td/><td/></tr><tr><td colspan="6">Season 2</td></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>1</td><td>0.2129</td><td>0.2129</td><td>6.545</td><td><bold>0.0131</bold></td></tr><tr><td> Residuals</td><td>60</td><td>1.952</td><td>0.0325</td><td/><td/></tr></tbody></table></table-wrap></p><p>Furthermore, <italic>TPS10</italic> and <italic>lox2/3</italic>x<italic>TPS10</italic> center plants were tested to determine whether levels of herbivore damage differed in mixed- vs monocultures, with the initial model</p><p>model1<-aov(Arcsin_total∼lox*MonoMix,centerTPSloxTPS)</p><p>and the final model</p><p>model3<-aov(Arcsin_total∼lox,centerTPSloxTPS)</p><p>with the following results:<table-wrap id="tblu5" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th/><th>Df</th><th>Sum sq</th><th>Mean sq</th><th>F</th><th>p</th></tr></thead><tbody><tr><td colspan="6">Season 1</td></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>1</td><td>0.0663</td><td>0.0663</td><td>8.726</td><td><bold>0.0052</bold></td></tr><tr><td> Residuals</td><td>40</td><td>0.3038</td><td>0.0076</td><td/><td/></tr><tr><td colspan="6">Season 2</td></tr><tr><td> <italic>LOX2/3</italic> deficiency</td><td>1</td><td>0.1679</td><td>0.1679</td><td>4.47</td><td><bold>0.0403</bold></td></tr><tr><td> Residuals</td><td>43</td><td>1.6147</td><td>0.0376</td><td/><td/></tr></tbody></table></table-wrap></p><p>Finally, to determine the effect of plant position, an ANOVA was conducted on individuals and edge and center plants from monocultures; values used for edge plants were the arithmetic mean of all edge plants in a population to reduce pseudoreplication, but a linear mixed-effects model was still used to correct for the fact that center and edge plants were from the same populations. The initial model was</p><p>model1<-lme(Mean_arcsin_total∼lox*Position*TPS,random=∼1|PopID,ind_mono,method=“ML”)</p><p>and the final model in season one was</p><p>model6REML<-lme(Mean_arcsin_total∼lox+Position,random=∼1|PopID,ind_mono)</p><p>with the following results:</p><p>LOX2/3 deficiency<table-wrap id="tblu6" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th>Season 1</th><th>Value</th><th>Std. error</th><th>DF</th><th>t</th><th>p</th><th>(Intr)</th><th>X2lx2l</th><th>Position2Edge</th></tr></thead><tbody><tr><td>(Intercept)</td><td>0.19326</td><td>0.01414</td><td>92</td><td>13.66</td><td><bold><0.0001</bold></td><td/><td/><td/></tr><tr><td><italic>LOX2/3</italic> deficiency</td><td>0.09517</td><td>0.01592</td><td>92</td><td>5.978</td><td><bold><0.0001</bold></td><td>−0.538</td><td/><td/></tr><tr><td>Position2Edge</td><td>−0.00666</td><td>0.01668</td><td>42</td><td>−0.3990</td><td>0.6919</td><td>−0.591</td><td>−0.021</td><td/></tr><tr><td>Position3Center</td><td>−0.03057</td><td>0.01689</td><td>42</td><td>−1.8097</td><td>0.0775</td><td>−0.588</td><td>−0.013</td><td>0.819</td></tr></tbody></table></table-wrap></p><p>The final model in season two was</p><p>model7REML<-lme(Mean_Arcsin_Total∼<italic>LOX2/3</italic> expression,random=∼1|PopID,ind_mono)<table-wrap id="tblu7" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th>Season 2</th><th>Value</th><th>Std. error</th><th>DF</th><th>t</th><th>p</th><th>(Intr)</th></tr></thead><tbody><tr><td>(Intercept)</td><td>0.3222</td><td>0.0153</td><td>121</td><td>21.12</td><td><bold><0.0001</bold></td><td/></tr><tr><td><italic>LOX2/3</italic> deficiency</td><td>0.1274</td><td>0.0227</td><td>121</td><td>5.619</td><td><bold><0.0001</bold></td><td>−0.673</td></tr></tbody></table></table-wrap></p><p>Statistical models used for data in <xref ref-type="fig" rid="fig6">Figure 6</xref>.</p><p>We tested the hypotheses that <italic>LOX2/3</italic> and <italic>TPS</italic> expression, plant position, and population type were associated with differences in foliar herbivore abundance. Generalized linear mixed-effects models were used to analyze data presented in <xref ref-type="fig" rid="fig6">Figure 6A,B</xref> so that the analysis of multiple plants per population could be included as a random blocking factor (glmer function in R in package lme4, <xref ref-type="bibr" rid="bib11">Bates et al., 2014a</xref>; <xref ref-type="bibr" rid="bib12">Bates et al., in review</xref>), while a generalized linear model (glm) was used for data in <xref ref-type="fig" rid="fig6">Figure 6C</xref> (in which case only one plant per population was analyzed). Models were fit by maximum likelihood (Laplace Approximation) using a poisson error distribution. Genotypes were coded as follows: 1WT, 2TPS10, 3lox23, 4lox23xTPS10; plant positions were coded as 1Individual, 2Edge, 3Center; and monocultures vs mixed cultures were coded as 1Mono, 2Mix.</p><p>For data shown in <xref ref-type="fig" rid="fig6">Figure 6A</xref>, the following model was used:</p><p>model<-glmer(Herbivores∼Genotype+(1|PopID),counts_all,family=poisson)</p><p>with the following results:<table-wrap id="tblu8" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th>Fixed effects</th><th>Estimate</th><th>Std. error</th><th>z</th><th>p</th></tr></thead><tbody><tr><td>(Intercept)</td><td>1.7</td><td>0.2361</td><td>7.219</td><td><bold>5.25E-13</bold></td></tr><tr><td>Genotype<italic>TPS10</italic></td><td>−0.8589</td><td>0.2116</td><td>−0.059</td><td><bold>9.30E-05</bold></td></tr><tr><td>Genotype<italic>lox2/3</italic></td><td>−0.612</td><td>0.3097</td><td>−1.983</td><td>0.07</td></tr><tr><td>Genotype<italic>lox2/3</italic>x<italic>TPS10</italic></td><td>−1.877</td><td>0.097</td><td>−0.582</td><td><bold>6.00E-06</bold></td></tr></tbody></table><table frame="hsides" rules="groups"><thead><tr><th>Correlation of fixed effects</th><th>(Intr)</th><th>GenotypeTPS10</th><th>Genotypelox2/3</th></tr></thead><tbody><tr><td>Genotype<italic>TPS10</italic></td><td>−0.36</td><td/><td/></tr><tr><td>Genotype<italic>lox2/3</italic></td><td>−0.369</td><td>0.27</td><td/></tr><tr><td>Genotype<italic>lox2/3</italic>x<italic>TPS10</italic></td><td>−0.325</td><td>0.22</td><td>0.531</td></tr></tbody></table></table-wrap></p><p>Tukey post-hoc contrasts were calculated using the multcomp package in R (<xref ref-type="bibr" rid="bib42">Hothorn et al., 2008</xref>) as</p><p>modelposthoc <- glht(model1, linfct = mcp(Genotype = “Tukey”))</p><p>with the following results:<table-wrap id="tblu9" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th>Linear hypotheses</th><th>Estimate</th><th>Std. error</th><th>z</th><th>p</th></tr></thead><tbody><tr><td><italic>TPS10</italic>–WT = 0</td><td>−0.8589</td><td>0.2116</td><td>−4.059</td><td><bold>0.0003</bold></td></tr><tr><td><italic>Lox2/3</italic>–WT = 0</td><td>−0.6142</td><td>0.3097</td><td>−1.983</td><td>0.1839</td></tr><tr><td><italic>Lox2/3</italic> x<italic>TPS10</italic>–WT = 0</td><td>−1.877</td><td>0.4097</td><td>−4.582</td><td><bold><1e-04</bold></td></tr><tr><td><italic>lox2/3</italic>–<italic>TPS10</italic> = 0</td><td>0.2447</td><td>0.3236</td><td>0.756</td><td>0.8675</td></tr><tr><td><italic>lox2/3</italic> x<italic>TPS10</italic>–<italic>TPS10</italic> = 0</td><td>−1.019</td><td>0.4168</td><td>−2.444</td><td>0.0644</td></tr><tr><td><italic>lox2/3</italic> x<italic>TPS10</italic>–<italic>LOX2/3</italic> = 0</td><td>−1.263</td><td>0.3592</td><td>−3.517</td><td><bold>0.0023</bold></td></tr></tbody></table></table-wrap></p><p>For data shown in <xref ref-type="fig" rid="fig6">Figure 6B</xref>, the following model was used:</p><p>model<-glmer(Herbivores∼Genotype*Position+(1|PopID), counts_WT_<italic>TPS10</italic>_Mono_Ind, family=poisson)</p><p>with the following results:<table-wrap id="tblu10" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th>Fixed effects</th><th>Estimate</th><th>Std. error</th><th>z</th><th>p</th></tr></thead><tbody><tr><td>(Intercept)</td><td>1.961</td><td>1.063</td><td>1.844</td><td><bold>0.0652</bold></td></tr><tr><td>Genotype<italic>TPS10</italic></td><td>−1.122</td><td>0.3064</td><td>−3.662</td><td><bold>0.0003</bold></td></tr><tr><td>PositionEdge</td><td>−0.8305</td><td>1.166</td><td>−0.712</td><td>0.4762</td></tr><tr><td>PositionCenter</td><td>−1.363</td><td>1.193</td><td>−1.143</td><td>0.2532</td></tr><tr><td>Genotype<italic>TPS10</italic>:PositionEdge</td><td>0.2737</td><td>0.7789</td><td>0.351</td><td>0.7253</td></tr><tr><td>Genotype<italic>TPS10</italic>:PositionCenter</td><td>1.748</td><td>0.7995</td><td>2.186</td><td><bold>0.0288</bold></td></tr></tbody></table><table frame="hsides" rules="groups"><thead><tr><th>Correlation of fixed effects</th><th>(Intr)</th><th>GenotypTPS10</th><th>PositionEdge</th><th>PositionCenter</th><th>G<italic>TPS10</italic>:PE</th></tr></thead><tbody><tr><td>Genotyp<italic>TPS10</italic></td><td>−0.071</td><td/><td/><td/><td/></tr><tr><td>PositionEdge</td><td>−0.912</td><td>0.065</td><td/><td/><td/></tr><tr><td>PositionCenter</td><td>−0.891</td><td>0.063</td><td>0.955</td><td/><td/></tr><tr><td>G<italic>TPS10</italic>:PE</td><td>0.028</td><td>−0.393</td><td>−0.269</td><td>−0.228</td><td/></tr><tr><td>G<italic>TPS10</italic>:PC</td><td>0.027</td><td>−0.383</td><td>−0.231</td><td>−0.323</td><td>0.797</td></tr></tbody></table></table-wrap></p><p>For data shown in 6C, the following initial model was used:</p><p>model1<-glm(Herbivores∼Genotype*MixMono,counts_WT_lox23_edge,family=poisson),</p><p>and the model was reduced to</p><p>model3<-glm(Herbivores∼Genotype,counts_WT_lox23_edge,family=poisson),</p><p>wit the following results:<table-wrap id="tblu11" position="anchor"><table frame="hsides" rules="groups"><thead><tr><th>Fixed effects</th><th>Estimate</th><th>Std. error</th><th>z</th><th>p</th></tr></thead><tbody><tr><td>(Intercept)</td><td>1.421</td><td>0.1313</td><td>10.83</td><td><bold><2e-16</bold></td></tr><tr><td>Genotype<italic>lox2/3</italic></td><td>0.4504</td><td>0.191</td><td>2.358</td><td><bold>0.0183</bold></td></tr></tbody></table></table-wrap></p></sec></sec></sec></boxed-text></app></app-group></back><sub-article article-type="article-commentary" id="SA1"><front-stub><article-id pub-id-type="doi">10.7554/eLife.04490.030</article-id><title-group><article-title>Decision letter</article-title></title-group><contrib-group content-type="section"><contrib contrib-type="editor"><name><surname>Greenberg</surname><given-names>Jean T</given-names></name><role>Reviewing editor</role><aff><institution>University of Chicago</institution>, <country>United States</country></aff></contrib></contrib-group></front-stub><body><boxed-text><p>eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see <ext-link ext-link-type="uri" xlink:href="http://elifesciences.org/review-process">review process</ext-link>). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.</p></boxed-text><p>Thank you for sending your work entitled “Plant defense phenotypes determine the consequences of volatile emission for individuals and neighbors” for consideration at <italic>eLife</italic>. Your article has been favorably evaluated by Diethard Tautz (Senior editor), a Reviewing editor, and three reviewers.</p><p>The Reviewing editor and the reviewers discussed their comments before we reached this decision, and the Reviewing editor has assembled the following comments to help you prepare a revised submission.</p><p>This manuscript makes an important contribution to understanding the role of two volatiles. The reviewers identified a few issues that need improvement.</p><p>1) You report that the role of <italic>lox</italic> mediated direct defence and its associated volatiles is determining damage on the plot level, not affecting focal or neighbouring plants. Whereas the <italic>TSP10</italic> plants do affect abundance of stem borer beetles and predator presence on individual plant level. This conclusion is sound based on the data presented, but it is also biased by the experimental design. You did not explore the role of <italic>LOX</italic> derived volatiles or direct defences on the individual plant level experimentally. You should acknowledge and discuss the fact that the effect of <italic>LOX</italic> mediated effects has not been tested fully. The Discussion should include how the missing treatment might affect your conclusions or further inferences about the role of particular volatile emissions.</p><p>2) Although from the plants perspective measurement of damage is a useful predictor of the costs of herbivory, the measurement may not accurately present the actual effect on herbivore abundance. Damage is confounding: herbivore abundance, their feeding intensity (that individuals may adjust to plant quality), and predator pressure on herbivores. Actual herbivore counts such as provided for the stem borer beetles would more strongly support the conclusions on <italic>LOX</italic> and <italic>TSP</italic> mediated community processes. Please discuss this point or provide the counts if you have them.</p><p>3) You rely very heavily on pairwise comparisons between treatments, usually with non-parametric tests or G tests for differences in proportions. The results would be much clearer to the readers, and avoid many of the concerns over multiple comparisons, if the authors used generalized linear models that better reflect the experimental design (see below for specific suggestion). For instance, the strongest results for their “community genetics” argument comes from the pattern of reduced mortality of <italic>lox2/3</italic> vs. WT plants when in the presence of a <italic>TPS10</italic> plant (<xref ref-type="fig" rid="fig9">Figure 9</xref>), but you do not test this directly. This pattern should be tested as an interaction between the <italic>lox</italic> and <italic>TPS</italic> treatments. You interpret this pattern strongly, but all of the tests performed show marginal or non-significant differences (a marginal decline in mortality for <italic>lox2/3</italic> plants and a non-significant increase in mortality for WT plants due to the presence of a <italic>TPS10</italic> plant). A similar model could be used for stem borer infestation and predation rates. In addition to providing more powerful tests, it may help the reader if all of these analyses are presented in a parallel way.</p><p>Specific suggestion: your statistical analysis could be improved for the analysis of the field experiment data (mortality, prevalence of stem borers, and prevalence of predation).</p><p>All of the outcome variables are binomial (1 or 0 at the level of an individual plant). It is recommended that you model this outcome with a generalized linear mixed model, using a binomial error distribution. This could be done with the glmer function in R or the GLIMMIX procedure in SAS, among other options. The model would include two crossed factors: the <italic>LOX</italic> genotype of the individual plant (2 levels, wild type or mutant) crossed with the <italic>TPS10</italic> treatment of the population (5 levels, individual WT, individual <italic>TPS10</italic>, monoculture WT, monoculuture <italic>TPS10</italic>, mixed), and plot as a random effect to account for the non-independence of plants from the same plot. The test of whether the presence of a <italic>TPS10</italic> plant alters mortality on non-<italic>TPS10</italic> neighbors differently based on the <italic>LOX</italic> genotype of those neighbors would come from the interaction term in this model, with the <italic>TPS10</italic> producing plant individuals excluded (and thus the <italic>TPS10</italic> monoculture treatment dropped). Or alternatively, you could drop the individual plant treatments and have a three way factorial model of <italic>LOX</italic> genotype*population treatment (mono WT, mono <italic>TPS10</italic>, mixed)*position (Center vs. Edge). You could then perform an a priori contrast to compare the interaction specifically between the (WT monoculture vs. mixture)*(<italic>LOX</italic> genotype) for edge plants to test the significance of the pattern in <xref ref-type="fig" rid="fig9">Figure 9</xref>.</p><p>4) Please consider that a streamlining of the presentation would make the manuscript easier to read and understand. A great deal of space is devoted to establishing that the gene silencing/overexpression worked and produced reliable phenotypes in controlled conditions. These methods and results could be moved to an appendix. This would allow the authors to focus on the results from the field experiment, where the interesting finding is not just that manipulating plant defenses and volatiles affects herbivore damage and plant fitness, but that the genotype of one individual can affect the outcome on neighboring individuals.</p><p>[Editors' note: further revisions were requested prior to acceptance, as described below.]</p><p>Thank you for resubmitting your work entitled “Plant defense phenotypes determine the consequences of volatile emission for individuals and neighbors” for further consideration at <italic>eLife</italic>. Your revised article has been favorably evaluated by Diethard Tautz (Senior editor), a member of the Board of Reviewing Editors, and two of the original reviewers. The manuscript has been improved but there are some remaining issues that need to be addressed before acceptance, as outlined below.</p><p>The reviewers found your manuscript to be improved. However there remains a statistical issue that requires resolution. This was framed by one of the reviewers as follows:</p><p>”My other concern with the original article was that the authors used a great deal of pairwise comparisons rather than building statistical models that better reflected their experimental design and could test directly for the interactive effects at the heart of their arguments. The authors made some attempt to respond to this suggestion, providing revamped analyses for plant size, flower number, and herbivore damage. However, I was disappointed to see that they did not collect their data in a way that would allow this more powerful analyses of the plant mortality data (which is the clearest case of “extended phenotypes” where the genotype of the center plant affected the fitness of the edge plants), or the data on stem borer abundance and predation rates. For the mortality data especially, this is surprising to me: why would the authors not have noted the live/dead status of plants individually, rather than just taking a tally? Do they mean that within a given population of five plants, they tallied the total number dead (1-5), or do they really mean that they cannot ascribe their mortality data even to the plot level (the base unit of replication for their experiment). If it is the former, they could still use a binomial generalized linear model, as long as they kept the data from the center and edge plants separately. If it is the latter, this is concerning to me because this analysis does not reflect the psuedoreplication induced by having five plants per plot. While a G test is “unreplicated” in that you need only enter a single frequency, it still assumes that the individuals comprising those frequencies are independent observations (i.e. a comparison between 28/75 vs. 50/75 is more powerful than 6/15 vs. 10/15). And I am truly puzzled as to how the authors can have data on flower production at an individual plant scale, but not have mortality data on that scale. Surely if they know how many flowers any given plant produced, they must know if it was alive or not. Am I missing something obvious here?”</p><p>In light of these comments, please clarify how the data on plant mortality and stem borer infection was collected. If the mortality analysis is unreplicated as described in the cover letter, please justify why it is still appropriate to use the statistical test that you used.</p></body></sub-article><sub-article article-type="reply" id="SA2"><front-stub><article-id pub-id-type="doi">10.7554/eLife.04490.031</article-id><title-group><article-title>Author response</article-title></title-group></front-stub><body><p><italic>1) You report that the role of lox mediated direct defence and its associated volatiles is determining damage on the plot level, not affecting focal or neighbouring plants. Whereas the</italic> TSP10 <italic>plants do affect abundance of stem borer beetles and predator presence on individual plant level. This conclusion is sound based on the data presented, but it is also biased by the experimental design. You did not explore the role of</italic> LOX <italic>derived volatiles or direct defences on the individual plant level experimentally. You should acknowledge and discuss the fact that the effect of</italic> LOX <italic>mediated effects has not been tested fully. The Discussion should include how the missing treatment might affect your conclusions or further inferences about the role of particular volatile emissions</italic>.</p><p>We agree that we did not always explicitly consider the limits of our experimental design when discussing the data, and to correct this, we have added the following text to the Discussion:</p><p>a) “It should be noted that our experimental design did not support the analysis of population-level effects […] in the presence, or absence, of total HIPVs and jasmonate-mediated defense.”</p><p>b) We changed the subsection title from “<italic>LOX2/3</italic> expression has large direct effects on plants, while the largest effects of <italic>TPS10</italic> expression are indirect” to “<italic>LOX2/3</italic> expression has large direct effects on plants, while <italic>TPS10</italic> expression has large indirect effects”.</p><p>c) We have also added the following sentence: “In contrast, due to the large suite of defenses regulated by <italic>LOX2</italic> and <italic>LOX3</italic> products, it would have been more difficult to determine the proximate cause of neighbor effects resulting from manipulation of <italic>LOX2</italic> and <italic>LOX3</italic> expression.”</p><p>d) We had already stated in the Discussion: “Although several components of plant HIPV blends have been shown to prime the direct and indirect defenses of neighbors in different systems (reviewed in <xref ref-type="bibr" rid="bib8">Baldwin et al<italic>.</italic> 2006</xref>; <xref ref-type="bibr" rid="bib7">Arimura et al<italic>.</italic> 2010</xref>; and see Heil<italic>,</italic> Kost 2006), GLVs may often be the active cues (<xref ref-type="bibr" rid="bib27">Engelberth et al<italic>.</italic> 2004</xref>; <xref ref-type="bibr" rid="bib51">Kessler et al<italic>.</italic> 2006</xref>; <xref ref-type="bibr" rid="bib29">Frost et al<italic>.</italic> 2008</xref>). We did not mix GLV-deficient with GLV-emitting plants in our experimental design, but we can conclude that <italic>TPS10</italic> volatiles had no detectable direct effects on neighbor plants in our set-up.” (This was in the first submission and has not changed.)</p><p>e) “Of course it is possible that the manipulation of <italic>LOX2</italic> and <italic>LOX3</italic> expression within populations would also reveal such extortionist-like effects, for example if <italic>T. mucorea</italic> preferred to oviposit in <italic>TPS10</italic> plants over <italic>LOX2/3</italic>x<italic>TPS10</italic> plants in the same population, which would support the hypothesis that <italic>T.</italic> mucorea preference reflects the ratio between <italic>TPS10</italic> volatiles and other HIPVs.”</p><p>f) And finally, we added: “In this context, it would be interesting to know whether central <italic>LOX2/3</italic> plants surrounded by WT edge plants would experience a similar predation disadvantage as central <italic>TPS10</italic> plants surrounded by WT edge plants, and whether central WT plants surrounded by <italic>LOX2/3</italic> edge plants would have an advantage over their neighbors in attracting predators.”</p><p><italic>2) Although from the plants perspective measurement of damage is a useful predictor of the costs of herbivory, the measurement may not accurately present the actual effect on herbivore abundance. Damage is confounding: herbivore abundance, their feeding intensity (that individuals may adjust to plant quality), and predator pressure on herbivores. Actual herbivore counts such as provided for the stem borer beetles would more strongly support the conclusions on</italic> LOX <italic>and</italic> TSP <italic>mediated community processes. Please discuss this point or provide the counts if you have them</italic>.</p><p>We thank the reviewers for this astute suggestion. We do indeed have one time point of herbivore count data and had previously not included it because the data are from only a subset of plants and we considered that they might not be as robust as the damage data. However, we re-visited these data and have now included them as a new <xref ref-type="fig" rid="fig6">Figure 6</xref> (the original figures have been re-numbered accordingly). Indeed, the herbivore abundance data differs from the damage data. This must be interpreted with caution as the abundance data were recorded on June 26<sup>th</sup> while the damage data were recorded on June 9<sup>th</sup>, but we state these dates clearly in the text and note, in the Results and in the Discussion, that observation of the plants indicated that there were no changes in trends of herbivore damage between these dates (i.e., relative damage levels remained consistent throughout the end of the experiment). These data are also introduced in the Introduction, reviewed in the Results and in the Discussion, in several commented locations, and a new paragraph has been included in the Methods under the subsection headed “Herbivore and predator counts”.</p><p><italic>3) You rely very heavily on pairwise comparisons between treatments, usually with non-parametric tests or G tests for differences in proportions. The results would be much clearer to the readers, and avoid many of the concerns over multiple comparisons, if the authors used generalized linear models that better reflect the experimental design (see below for specific suggestion). For instance, the strongest results for their “community genetics” argument comes from the pattern of reduced mortality of</italic> lox2/3 <italic>vs. WT plants when in the presence of a</italic> TPS10 <italic>plant (</italic><xref ref-type="fig" rid="fig9"><italic>Figure 9</italic></xref><italic>), but you do not test this directly. This pattern should be tested as an interaction between the lox and</italic> TPS <italic>treatments. You interpret this pattern strongly, but all of the tests performed show marginal or non-significant differences (a marginal decline in mortality for</italic> lox2/3 <italic>plants and a non-significant increase in mortality for WT plants due to the presence of a</italic> TPS10 <italic>plant). A similar model could be used for stem borer infestation and predation rates. In addition to providing more powerful tests, it may help the reader if all of these analyses are presented in a parallel way</italic>.</p><p><italic>Specific suggestion.: your statistical analysis could be improved for the analysis of the field experiment data (mortality, prevalence of stem borers, and prevalence of predation).</italic></p><p><italic>All of the outcome variables are binomial (1 or 0 at the level of an individual plant). It is recommended that you model this outcome with a generalized linear mixed model, using a binomial error distribution. This could be done with the glmer function in R or the GLIMMIX procedure in SAS, among other options. The model would include two crossed factors: the</italic> LOX <italic>genotype of the individual plant (2 levels, wild type or mutant) crossed with the</italic> TPS10 <italic>treatment of the population (5 levels, individual WT, individual</italic> TPS10<italic>, monoculture WT, monoculuture</italic> TPS10<italic>, mixed), and plot as a random effect to account for the non-independence of plants from the same plot. The test of whether the presence of a</italic> TPS10 <italic>plant alters mortality on non-</italic>TPS10 <italic>neighbors differently based on the LOX genotype of those neighbors would come from the interaction term in this model, with the</italic> TPS10 <italic>producing plant individuals excluded (and thus the</italic> TPS10 <italic>monoculture treatment dropped). Or alternatively, you could drop the individual plant treatments and have a three way factorial model of</italic> LOX <italic>genotype*population treatment (mono WT, mono</italic> TPS10<italic>, mixed)*position (Center vs. Edge). You could then perform an a priori contrast to compare the interaction specifically between the (WT monoculture vs. mixture)*(</italic>LOX <italic>genotype) for edge plants to test the significance of the pattern in</italic> <xref ref-type="fig" rid="fig9"><italic>Figure 9</italic></xref>.</p><p>We have revisited these raw data and regrettably, none of these data sets were collected in a manner which we think permits application of the modeling techniques suggested. Specifically, the <italic>Trichobaris mucorea</italic> infestation data and mortality data were simply collected as tallies per individual or population type and genotype at the end of the season, i.e., not replicated. Thus an unreplicated analysis such as the G-test, according to our understanding, must be used. The predation assays were not always conducted on exactly the same group of plants, as over the course of the 5 repetitions some plants were damaged by wind or severed by cutworms and new plants had to be assigned. Thus it would be difficult to appropriately apportion error to a random or repeated-measures term in a model, and we feel it is most appropriate for these data as well to treat them as total frequencies and conduct an unreplicated statistical analysis. We have stated these limitations clearly in the Methods and Results sections describing the relevant experiments, as marked by comments in the revised text.</p><p>However, we greatly appreciate the reviewers’ points and detailed suggestions. We thus reviewed all of the data in the manuscript and determined that other data sets would be more appropriately analyzed with such an approach. We therefore used an approach similar to that suggested by the reviewers to analyze the newly included foliar herbivore abundance data, and we furthermore re-analyzed our foliar herbivore damage, plant size, and flower production data. These changed analyses are presented in the new version of <xref ref-type="fig" rid="fig5">Figure 5</xref> (foliar herbivore damage data), the new <xref ref-type="fig" rid="fig6">Figure 6</xref> and source data files (foliar herbivore abundance data), and a new <xref ref-type="fig" rid="fig8">Figure 8</xref> and source data file (re-analyzed size and flower production data). The general statistical approaches are described in the revised statistical analysis section of the Materials and methods, and in addition, we have provided detailed descriptions of the analyses as part of the new <xref ref-type="app" rid="app3">Appendix 3</xref>.</p><p>In particular, the re-analysis of plant size and reproduction data allowed us to simplify the presentation of the flowering data in <xref ref-type="fig" rid="fig7">Figure 7</xref> to focus only on <italic>LOX2/3</italic> deficiency effects, while revealing subtler interactive effects of <italic>TPS10</italic> expression in the separate effects (new analysis presented in the new <xref ref-type="fig" rid="fig8">Figure 8</xref> and <xref ref-type="supplementary-material" rid="SD8-data">Figure 8–Source data 1</xref>), which fit well with the other data sets, and which were overlooked in our previous analysis.</p><p>Overall, the revised statistical analyses have revealed more population-level and interactive effects of <italic>TPS10</italic> expression which we feel facilitate our discussion of the data, strengthen our conclusions, and overall provide a more comprehensible and coherent picture for the reader.</p><p><italic>4) Please consider that a streamlining of the presentation would make the manuscript easier to read and understand. A great deal of space is devoted to establishing that the gene silencing/overexpression worked and produced reliable phenotypes in controlled conditions. These methods and results could be moved to an appendix. This would allow the authors to focus on the results from the field experiment, where the interesting finding is not just that manipulating plant defenses and volatiles affects herbivore damage and plant fitness, but that the genotype of one individual can affect the outcome on neighboring individuals</italic>.</p><p>We thank the reviewers for this suggestion to streamline the text. The <italic>eLife</italic> staff suggested that we create three appendices, each with one coherent topic or set of contents. Thus, the results regarding screening of transgenic lines are now in <xref ref-type="app" rid="app1">Appendix 1</xref>, results demonstrating field-stable phenotypes are in <xref ref-type="app" rid="app2">Appendix 2</xref>, and the corresponding methods are in <xref ref-type="app" rid="app3">Appendix 3</xref>. We feel this has greatly improved the readability and organization of the manuscript.</p><p>[Editors' note: further revisions were requested prior to acceptance, as described below.]</p><p><italic>The reviewers found your manuscript to be improved. However there remains a statistical issue that requires resolution. This was framed by one of the reviewers as follows:</italic></p><p><italic>“My other concern with the original article was that the authors used a great deal of pairwise comparisons rather than building statistical models that better reflected their experimental design and could test directly for the interactive effects at the heart of their arguments. The authors made some attempt to respond to this suggestion, providing revamped analyses for plant size, flower number, and herbivore damage. However, I was disappointed to see that they did not collect their data in a way that would allow this more powerful analyses of the plant mortality data (which is the clearest case of “extended phenotypes” where the genotype of the center plant affected the fitness of the edge plants), or the data on stem borer abundance and predation rates. For the mortality data especially, this is surprising to me: why would the authors not have noted the live/dead status of plants individually, rather than just taking a tally? Do they mean that within a given population of five plants, they tallied the total number dead (1-5), or do they really mean that they cannot ascribe their mortality data even to the plot level (the base unit of replication for their experiment). If it is the former, they could still use a binomial generalized linear model, as long as they kept the data from the center and edge plants separately. If it is the latter, this is concerning to me because this analysis does not reflect the psuedoreplication induced by having five plants per plot. While a G test is “unreplicated” in that you need only enter a single frequency, it still assumes that the individuals comprising those frequencies are independent observations (i.e. a comparison between 28/75 vs. 50/75 is more powerful than 6/15 vs. 10/15). And I am truly puzzled as to how the authors can have data on flower production at an individual plant scale, but not have mortality data on that scale. Surely if they know how many flowers any given plant produced, they must know if it was alive or not. Am I missing something obvious here?”</italic></p><p><italic>In light of these comments, please clarify how the data on plant mortality and stem borer infection was collected. If the mortality analysis is unreplicated as described in the cover letter, please justify why it is still appropriate to use the statistical test that you used</italic>.</p><p>This is an important concern and a considered comment, and we want to give a correspondingly careful and reasoned response, which is below. In summary, the data presented in <xref ref-type="fig" rid="fig9">Figure 9</xref> is the most biologically meaningful plant mortality data, which is unfortunately from an unreplicated data set; however the analysis presented in <xref ref-type="fig" rid="fig9">Figure 9</xref> is appropriate for these data and our question, and does not violate the assumptions of the G-test regarding independence between the groups compared. Likewise, the assumption of independence is not violated for the analysis in <xref ref-type="fig" rid="fig11">Figure 11</xref>. However the assumption of independence of the Fisher’s test was violated in one instance for the analysis presented in <xref ref-type="fig" rid="fig10">Figure 10</xref>. We have corrected our analysis for <xref ref-type="fig" rid="fig10">Figure 10</xref> and associated parts of the Results text, but the corrected analysis produces exactly the same conclusions.</p><p>Question 1: Should we replace our unreplicated plant mortality data set (presented in <xref ref-type="fig" rid="fig9">Figure 9</xref>) with a replicated data set, collected concurrently with flowering data?</p><p>Of course we did note plant mortality while collecting the data on flowering. However the flowering data was collected on May 29<sup>th</sup> and June 6<sup>th</sup>, while we wished to analyze mortality by the end of the season on June 29<sup>th</sup>. All other data sets collected between June 6<sup>th</sup> and June 29<sup>th</sup> (herbivore abundance, predation rates) were collected on subsets of plants and plant mortality was not noted again until June 29<sup>th</sup>, at which point all surviving plants were tallied in categories while checking stems for infestation by <italic>Trichobaris mucorea</italic>. Thus total mortality for the season was calculated on June 29<sup>th</sup> as the number of plants originally planted minus the number surviving in each category. Because these total mortality numbers from June 29<sup>th</sup> most accurately reflect the likelihood that plants would not have survived to produce ripe seed, we feel it is most meaningful to present the June 29<sup>th</sup> data.</p><p>One could argue that, although the final mortality numbers from June 29<sup>th</sup> are principally more meaningful, nevertheless if mortality did not change much between June 6<sup>th</sup> and June 29<sup>th</sup>, it would be preferable to present a replicated rather than an unreplicated data set. However there was a 1.5-fold increase in overall mortality, from 29% to 46%, between June 6<sup>th</sup> and June 29<sup>th</sup>. Although the mortality data from June 6<sup>th</sup> show similar trends as June 29<sup>th</sup>, with <italic>LOX2/3</italic> deficiency tending to increase mortality and <italic>TPS10</italic> expression tending to lower mortality, the differences on June 6<sup>th</sup> are not significant even when analyzed using a GLM with a binomial distribution, because the total mortality is lower. Thus the June 6<sup>th</sup> dataset has lower statistical power and, more importantly, is truly less biologically meaningful.</p><p>Answer to question 1: No. It is best to present the unreplicated June 29<sup>th</sup> data.</p><p>Question 2: Is the use of G-tests to analyze the data in <xref ref-type="fig" rid="fig9">Figure 9</xref> appropriate? What about the use of G-tests and Fisher’s tests in other data sets (<xref ref-type="fig" rid="fig10 fig11">Figures 10 and 11</xref>)?</p><p>In analyzing total mortality on June 29<sup>th</sup>, we wished to know whether mortality rate was dependent on, or independent of, population type. We thus used G-tests of independence and conducted post-hoc P-value corrections for multiple comparisons as necessary, which we understand to be the most appropriate choice of tests for our unreplicated data and our question. Although our data came from experimental populations (plots), we did not violate the assumption of independence incorporated into the G-test for the data shown in <xref ref-type="fig" rid="fig9">Figure 9</xref>. This is because, in <xref ref-type="fig" rid="fig9">Figure 9</xref>, we only compare data from separate population types (separate plots), and we never compare data from plants at different positions in the same population type. If we had compared e.g. center to edge plants, there would be dependencies within the data set because some of the plants in group “edge” would have come from the same populations as plants in group “center”, and thus we would not be comparing independent observations. However by comparing different population types, we ensured that no plants in group A shared a population with any plants in group B, and therefore there are no dependencies between groups: they are independent. The same reasoning holds for the comparisons in <xref ref-type="fig" rid="fig11">Figure 11</xref>, and for the comparisons of mortality between plants differing only in <italic>LOX2/3</italic> deficiency or from different population types made in the Results section text (in the subsection headed “<italic>LOX2/3</italic> deficiency increases plant mortality, but mortality is mitigated by single <italic>TPS10</italic>-expressing plants in <italic>LOX2/3-</italic>deficient populations”).</p><p>For <xref ref-type="fig" rid="fig10">Figure 10</xref>, we compare both plants from separate population types and plants from the same population type. The comparison of plants from the same population type violates the independence assumption of the Fisher’s test and we thank the Reviewer for bringing this to our attention. We have removed the incorrect comparisons, changed the representation of the statistical comparisons in <xref ref-type="fig" rid="fig10">Figure 10</xref> accordingly, and corrected the relevant parts of the figure caption and of the manuscript (subsection headed “<italic>TPS10</italic>-expressing plants increase the infestation rates of their neighbors by the stem-boring weevil <italic>Trichobaris mucorea</italic>”). In this part of the manuscript text, we also removed the Fisher’s tests of infestation by genotype: when comparing genotypes differing only in <italic>TPS10</italic> expression overall, rather than different population types, we also violated the assumption of independence because some populations contain both <italic>TPS10</italic>-expressing and non-expressing individuals. As a result of removing some of the pairwise comparisons, our post-hoc correction factor decreased and the <italic>P</italic>-value we report for the comparison of WT plants in mixed versus monocultures decreased from 0.025 to 0.015. Because the critical comparisons in the analysis of the <italic>Trichobaris mucorea</italic> infestation data are those made between, and not within populations, the correction of our statistics does not change the conclusions we draw from these data.</p><p>Answer to Question 2: Yes, the G-tests and Fisher’s tests are appropriate and generally the requirement of independent observations is not violated. However, one comparison shown in <xref ref-type="fig" rid="fig10">Figure 10</xref> and a few comparisons in the associated results text violated the independence assumption of the Fisher’s test. We have removed these comparisons. The corrected analysis does not change our conclusions at all.</p></body></sub-article></article> |