Divergent Evolution in Antiherbivore Defences Within Species Complexes at a Single Amazonian Site
Total Page:16
File Type:pdf, Size:1020Kb
Journal of Ecology 2015, 103, 1107–1118 doi: 10.1111/1365-2745.12431 Divergent evolution in antiherbivore defences within species complexes at a single Amazonian site Marıa-Jose Endara1*, Alexander Weinhold1, James E. Cox2, Natasha L. Wiggins1, Phyllis D. Coley1,3 and Thomas A. Kursar1,3 1Department of Biology, The University of Utah, Salt Lake City, UT 84112-0840, USA; 2Health Sciences Center Core Research Facilities, School of Medicine, The University of Utah, Salt Lake City, UT 84112-0840, USA; and 3Smithsonian Tropical Research Institute, Balboa Box 0843-03092, Panama Summary 1. Classic theory in plant–insect interactions has linked herbivore pressure with diversification in plant species. We hypothesize that herbivores may exert divergent selection on defences, such that closely related plant species will be more different in defensive than in non-defensive traits. 2. We evaluated this hypothesis by investigating two clades of closely related plant species coexis- ting at a single site in the Peruvian Amazon: Inga capitata Desv. and Inga heterophylla Willd. spe- cies complexes. We compared how these lineages differ in the suite of chemical, biotic, phenological and developmental defences as compared to non-defensive traits that are related to habitat use and resource acquisition. We also collected insect herbivores feeding on the plants. 3. Our data show that sister lineages within both species complexes are more divergent in antiherbi- vore defences than in other non-defensive, functional traits. Moreover, the assemblages of herbivore communities are dissimilar between the populations of coexisting I. capitata lineages. 4. Synthesis. Our results are consistent with the idea that for the I. capitata and I. heterophylla spe- cies complexes, interactions with their natural enemies may have played a significant role in their phenotypic divergence and potentially in their diversification and coexistence. It also suggests that defensive traits are evolutionary labile. Key-words: herbivores, Inga, Peru, plant defences, plant–herbivore interactions, trait divergence, tropical rain forests The coevolutionary theory of plant–herbivore interactions Introduction suggests that the production of defences against insects has The arms race between plants and insect herbivores has been played a dominant role in host and enemy radiations (Ehrlich invoked as one of the main mechanisms driving trait diversifi- & Raven 1964). Specifically, this theory predicts a tight cor- cation and coevolution for both groups (Becerra 1997; relation between plant relatedness and plant defences. Thompson 1988, Becerra, Noge & Venable 2009; Futuyma & Although widely accepted, relatively few studies have tested Agrawal 2009; Kursar et al. 2009; Agrawal et. al. 2012). A this, and some even question the fundamental assumptions of fundamental prediction of this theory is that herbivores drive this theory. For example, Becerra (1997) found only a weak the evolution of plant antiherbivore defences faster than for relationship between the phylogenetic hypothesis and chemi- other traits (Thompson 2005; Kursar et al. 2009). Testing this cal similarity for the species of Bursera, common trees in the hypothesis requires demonstrating that sister species are more dry forests of Mexico. Likewise, Kursar et al. (2009) found a different in antiherbivore defences than in traits related to weak correlation between phylogenetic distances and chemical adaptations to other extrinsic factors, such as the abiotic envi- distances within the Neotropical tree genus, Inga. This lack of ronment. However, studies testing this idea are surprisingly phylogenetic signal in the expression of secondary metabolites few (e.g. Agrawal et al. 2009). Consequently, in this study, suggests divergent selection on antiherbivore defences, such we combine data on plant functional traits and insect herbi- that closely related species are not necessarily similar in vores to compare patterns of divergence in two groups of clo- defences. This should make it more difficult for herbivores to sely related species coexisting at a single site. track hosts over evolutionary time thereby reducing herbivore pressure on plants. Although the role of the physical environment on trait *Correspondence author: E-mail: [email protected] divergence has received considerable attention (Anacker & © 2015 The Authors. Journal of Ecology © 2015 British Ecological Society 1108 M.-J. Endara et al. Strauss 2014), the role of defensive traits in plant diversifica- 268 m.a.s.l. Los Amigos covers 453 ha of lowland Amazonian forest tion is not well understood (Futuyma & Agrawal 2009). Yet, and consists of a mosaic of terra firme and floodplain forests. Mean several studies have provided indirect evidence for the signifi- annual rainfall is between 2700 and 3000 mm, and the mean monthly ° cance of defensive traits in evolutionary diversification by temperature ranges from 21 to 26 C (Pitman 2007). showing a relationship between variation in these traits and species diversity (Farrell, Dussourd & Mitter 1991; Becerra STUDY SPECIES 1997; Agrawal & Fishbein 2008; Agrawal, Salminen & Fish- bein 2009; Agrawal et al. 2009; Kursar et al. 2009). A better Inga capitata comprises three phenotypically divergent ESUs: cap1, cap2 and cap3 (Kursar et al. 2009). In addition, they present different understanding of the relative importance of defensive traits in habitat preferences, with cap1 and cap3 showing a preference for terra phenotypic diversity and species divergence will require firme and cap2 for floodplains. The I. heterophylla species complex examining differences in defensive and non-defensive traits includes two phenotypically different lineages: het1 and het2 (Kursar simultaneously between recently diverged species or popula- et al. 2009), both on terra firme. For those ESUs found in terra firme, tions at an incipient state of divergence (Futuyma & Agrawal one ESU often is metres away from another and no intermediates 2009, Fine et al. 2013). were observed. The study plants were widely distributed within their Here, we examine the contribution of plant–insect interac- respective habitat types; aside from as noted above, inspection of the tions to divergence among species by determining variation in location data and our field observations showed no tendency for the functional traits and herbivore communities within two clades study species to be clumped or restricted to certain habitats (e.g. pref- of closely related species coexisting at a single site in the erence for treefall light gaps). Peruvian Amazon: Inga capitata Desv. and Inga heterophylla Willd. species complexes. The taxa in each complex are con- CENSUSES AND LEAF TRAITS sidered a single species based on the morphological traits of fi reproductive individuals (Pennington 1997). Our field obser- In the present study, antiherbivore defences are de ned as those plant vations of subtle differences within each clade in colour of traits that have been selected in response to herbivory. These include developmental defences (leaf expansion rate; Kursar & Coley 2003), the expanding leaves (see Appendix S1 in Supporting Infor- biotic defences (leaf-defending ants and the area of extrafloral nectar- mation), number of leaflets and stipule morphology have ies; Koptur 1984; Brenes-Arguedas, Coley & Kursar 2008), phenolog- motivated the present characterization of trait divergence ical defences (the timing and synchrony of young leaf production, within these clades. In fact, plastid DNA analyses distinguish Aide 1993; Kursar & Coley 2003) and chemical defences (phenolics each member as a different evolutionarily significant unit and non-protein amino acids; Coley et al. 2005). This set of defence (ESU) and as sister taxa (Kursar et al. 2009, Fig. S2). Based traits was measured only on expanding leaves because more that 80% on these analyses, I. capitata comprises three ESUs and of the damage accrued during a leaf’s lifetime happens during the I. heterophylla two ESUs. Four of the ESUs co-occur in terra short period (1–3 weeks) of leaf expansion (Coley & Aide 1991; Kur- firme habitats, often within metres of each other, whereas one sar & Coley 2003; Brenes-Arguedas et al. 2006). Therefore, young I. capitata ESU species (cap2) occurs primarily in the nearby leaf defences are under strong natural selection by herbivores. floodplains. Traits under selection from the physical environment are consid- In order to achieve a comprehensive analysis, we collected ered here as non-defence traits. These traits were measured only on mature leaves. These include leaf mass per area (LMA), leaf nitrogen data on many different defensive traits including chemical, content, area per leaflet, number of leaflets per leaf, and the presence biotic, phenological and developmental defences, as well as or absence of wings. These include some of the key ecophysiological on insect herbivores. We also collected data on non-defensive attributes that correlate with photosynthetic capacity and transpiration, traits that are related to habitat use and resource acquisition. with habitat type such as light availability and with resources such as Our previous studies on the genus Inga suggest that defences soil nutrient content (Cornelissen et al. 2003; Wright et al. 2004; Fuj- evolve rapidly (Kursar et al. 2009). Specifically, we expect ita, van Bodegom & Witte 2013). Although, in principle, the LMA the ESUs within an Inga lineage to be more similar with and leaf nitrogen of mature leaves can affect leaf palatability to herbi- respect to