This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright
Transcript
This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies areencouraged to visit:
DNA-based taxonomy of larval stages reveals huge unknown species diversityin neotropical seed weevils (genus Conotrachelus): relevance to evolutionary ecology
Sara Pinzón-Navarro a,b, Héctor Barrios c, Cesc Múrria a, Christopher H.C. Lyal a, Alfried P. Vogler a,b,*
a Department of Entomology, Natural History Museum, London, UKb Division of Biology, Imperial College London, Silwood Park Campus, Ascot, UKc Programa Centroamericano de Maestría en Entomología, Vicerrectoría de Investigación y Postgrado, Universidad de Panamá, Panamá
a r t i c l e i n f o
Article history:Received 28 October 2009Revised 12 February 2010Accepted 20 February 2010Available online 25 February 2010
High diversity in tropical phytophagous insects may be linked to narrow host specificity and host shifts,but tests are complicated by incomplete taxonomy and difficulties in food source identification. Speci-mens of the highly diverse New World genus Conotrachelus (Coleoptera: Curculionoidea) were rearedfrom >17,500 fruits (seeds) at six Central American rain forests. Interception traps were used for compar-ison with assemblages flying in the forest. Mitochondrial cox1 and the nuclear 28S genes were sequencedfor 483 larval and adult specimens. A Yule-Coalescent technique was used to group cox1 sequences intoputative species (17 from traps, 48 from rearing). Cox1 sequences of 24 species from museum collectionsprovided matches for three species from traps and no match for the reared species. Inga (Fabaceae) wasthe predominant host among 15 other genera and 67% of the weevils were monophagous. A three genetree (cox1, rrnL, 28S) recovered four well-supported clades feeding on Inga confirmed by phylogeneticcommunity analyzes that showed phylogenetic conservation of host plant utilization. This suggests thathost shifts are not directly involved in speciation, while the broad taxonomic host range and the evolu-tionary repeated shifts still contribute to the high species richness in Conotrachelus. The DNA-basedapproach combining species delimitation and phylogenetic analysis elucidated the evolutionary diversi-fication of this lineage, despite insufficient taxonomic knowledge. Conotrachelus is an example of thediverse tropical groups that require DNA-based taxonomy to study their evolutionary ecology.
� 2010 Elsevier Inc. All rights reserved.
1. Introduction
The factors that determine host specificity and promote theevolution of species diversity in herbivores and host plants remaininsufficiently understood (Basset, 1996; Novotny and Basset,2005). The diversity of host plants and the resulting diversity ofniche space may drive divergent adaptation and speciation in theherbivores, i.e., diversification is promoted by host shifts (Berl-ocher and Feder, 2002; Marvaldi et al., 2002; Janz et al., 2006; Dyeret al., 2007; McKenna et al., 2009). Alternatively, passive co-diver-sification with host plant lineages may provide mechanisms of spe-ciation, in which case host choice is conservative (Ehrlich andRaven, 1964; Mitter et al., 1991; Becerra, 2007). Phylogenetic anal-yses of herbivores and their host plants can provide the evolution-ary framework for testing questions about niche divergence orconservatism, and discriminate among various evolutionary sce-narios, also including the effects of community interactions (Webbet al., 2002; Emerson and Gillespie, 2008; Cavender-Bares et al.,
2009). Studying the role of herbivore–host interactions in diversi-fication of insect lineages requires, first, that the taxa and theirphylogenetic relationships are known and, second, that host re-cords can be established with high reliability. However, both taxo-nomic and ecological information is missing for many hyperdiversegroups of insects, precluding quantitative approaches to questionsabout insect–host plant co-diversification in most lineages.
Seed-feeding is a special case of herbivory of great evolutionaryrelevance due to the direct effect of seed predation on fitness of thehost (Janzen, 1980). Most studies of herbivore–host plant interac-tions in seed feeders have used well studied groups from the tem-perate regions (Hughes and Vogler, 2004; Morse and Farrell, 2005).Studies of tropical groups have been limited to a few taxonomicallyrestricted lineages based on time-consuming rearing studies fromlarge samples of seeds collected from potential host plants (Janzen,1980; Kergoat et al., 2007). Rearing has the advantage of linkingthe life history of all stages, but frequently it is prone to failure un-der laboratory conditions, undermining a quantitative analysis ofthe community. Alternative methods connecting larval specimensto a taxonomic identification (usually based on the morphologyof the adults) would greatly improve our power to detect the pat-terns of insect–plant interactions. DNA-based approaches using
1055-7903/$ - see front matter � 2010 Elsevier Inc. All rights reserved.doi:10.1016/j.ympev.2010.02.022
* Corresponding author. Address: Department of Entomology, Natural HistoryMuseum, London SW7 5BD, UK. Fax: +44 207 942 5229.
coalescence-based methods for grouping of sequences into evolu-tionarily significant groups (Pons et al., 2006; Abdo and Golding,2007) may now be applied to any life stage and circumvent the dif-ficulties with morphological species identification (e.g., Ahrenset al., 2007; Levkanicova and Bocak, 2009). This provides a new ap-proach to studies of herbivore–host plant interactions and co-evolution.
The analysis of herbivores at the larval stages also raises thepossibility that the species encountered in larval assemblages ob-tained directly from seeds do not match those described by tradi-tional taxonomy, which are usually obtained as adults. It is wellknown that different collecting and trapping methods yield differ-ent species (Basset et al., 2007; Missa et al., 2009) and samplinglarval stages may add previously unobserved species to commu-nity surveys. Hence, as sampling tropical insects commonly leadsto a significant proportion of undescribed species (Ødegaard,2003, 2006), sampling larval stages of tropical insects are expectedto do so even more. DNA-based clusters may be linked to knownspecies if they include adults, but no such link is possible in clus-ters based exclusively on larvae, even if they do represent knownspecies. However, morphological characters used to define widergroups of species may still enable newly discovered species or lar-vae-only clusters to be connected to the existing system at higherhierarchical levels such as species groups and genera, provided themorphological characters define monophyletic groups in the DNA-based phylogenetic trees. Hence, in a DNA-based system the avail-ability of adults for close relatives of larval-only groups may permitthe placement of these groups within the existing taxonomicsystem.
DNA-based approaches to the study of tropical insect communi-ties remain in their infancy, except for a few cases showing greatpromise in linking larval and adult assemblages (Ahrens et al.,2007) and geographically distant sites (Condon et al., 2008). Herewe apply this approach to a group of extremely diverse seed wee-vils in the genus Conotrachelus. This genus, comprising at least1200 species, is one of the most diverse in the family Curculioni-dae, distributed across North, Central and South America. Conotra-chelus are very species rich in tropical areas, with 135 describedspecies reported for Panama and Costa Rica (O’Brien and Wibmer,1984). Most species of Conotrachelus feed on seeds and fruits, whilesome are gall-forming in stems or inquilinous in other galls. Sev-eral species are important pests, including the plum weevil Cono-trachelus nenuphar that affects plum orchards throughout NorthAmerica (Jenkins et al., 2006; Leskey and Wright, 2007) and theguava weevil Conotrachelus psidii causing serious losses in SouthAmerica (Bondar, 1946; Bailez et al., 2003; Silva et al., 2007). Cono-trachelus in the Neotropics has been associated with numerousplant groups, ranging from Asterids and Eurosids to Magnoliids(Pinzón-Navarro et al., 2010) and seeds of Inga pods (Fabaceae)which have been identified as food source for many species (Va-lente and Gorayeb, 1994).
The current higher-level taxonomy of Conotrachelus is based ona small set of morphological characters, such as the shape andlength of the rostrum, pectoral canal, position of the coxae andthe overall body shape. These characters have been used to distin-guish between 4 and 6 major groups that each are further subdi-vided into up to a dozen subgroups. The most comprehensivemonographic treatment of the genus is that by Fiedler (1940)who compiled basic taxonomic and distributional information forthe South American species, including roughly half of the speciesnamed today. A study of the 37 species known from North Americawas published at approximately the same time but not cross-refer-enced (Schoof, 1942). Prior to these studies, the number and extentof species groups recognized in the older literature varied but in allcases was a phenetic grouping rather than an assessment of phy-logeny. As the starting point for a revised analysis of this huge
and important genus we here apply a DNA-based approach to sur-vey local samples of the genus reared from fruits and seeds in abotanically well studied forest at Barro Colorado Island, Panama(Croat, 1978) and additional Central American rainforests. TheDNA-based analysis provided a new perspective on the immensespecies richness in Conotrachelus and its evolutionary interactionswith the host plants.
2. Materials and methods
2.1. Taxon sampling
Fruits were collected from five sites in Panama: Barro ColoradoMonument (BCI), Bocas del Toro (BOC), Fortuna Dam Area (FOR),Parque Nacional Altos de Campana (CAM) and San Lorenzo Pro-tected Area (SHE) (Appendix 1). All fruiting trees encounteredalong forest trails over a total period of 6 months during the fruit-ing seasons of 2006 and 2007 were sampled. Fruiting trees from is-lets and peninsulas across BCI were sampled by boat. Furthersamples were obtained from La Selva Field Station (LSE), CostaRica, during 2007. Mature fruits were collected from tree branchesand forest floor; if fruits were green the tree was marked and re-sampled when fruits were mature. The total collecting effort in-cluded 17,532 fruits sampled from 215 trees. Fruits were morpho-logically identified to species-level (see Acknowledgments) andsome cases confirmed by DNA-aided identification based on chlo-roplast sequences obtained from the beetle specimens (Pinzón-Navarro et al., 2010). Fruits were sub-sampled; half of each fruitsample (per tree species) was opened in search of larvae feedingon seeds, which if encountered, were stored for DNA extraction,and the remaining fruits were used for rearing. Weevil larvae wererecognized among other beetle larvae present in the fruits by thepresence of two-segmented labial palps (distinguishing them fromBruchinae) and the absence of legs and pigmentation on the tho-racic plates (distinguishing them from Nitidulidae) (Stehr, 1991).
Rearing of seed feeding insects was done in plastic boxes, ven-tilated through a fine mesh in the lid and monitored every 2 daysfor emergence of adults or larvae. Emerging larvae were trans-ferred to plastic boxes with a layer of soil (covered with a finemesh). Time to pupation and emergence was recorded, which var-ied greatly among species from 2 weeks to several months. All spe-cies for which rearing was successful undertook pupation in thesoil, except for Conotrachelus pumilio feeding on Coccoloba manzan-illensis, which pupated within the fruit and emerged as adult. Addi-tional specimens were obtained from flight intercept traps(Appendix 2) placed at BCI near fruiting trees from which seedswere collected. Traps were set in the understory (1 m) and canopy(30–35 m) to obtain emerging adults or ovipositing females,respectively.
Total genomic DNA was extracted non-destructively fromabdominal tissue of larvae or from the head and thorax of adultsusing a Promega 96-well plate kit. Mitochondrial cytochrome oxi-dase I (cox1) was amplified with primers C1-J-2183 (Jerry) and TL2-N-3014 (Pat) (Simon et al., 1994) and the 16S rRNA gene (rrnL)(433–437 bp) with LR-N-13398 (16Sar) (Simon et al., 1994) andLR-J-12961 (16Sb2) (Cognato and Vogler, 2001). The nuclear largesubunit rRNA gene (28S rRNA) was amplified with primers 28SFFand 28SDD (Monaghan et al., 2007). Loaned museum (pinned)Conotrachelus specimens (see Acknowledgments) were selectedto include Panamanian species collected in the studied area, whichwere sequenced for cox1. Where primers Pat and Jerry failed to am-plify the full-length fragment, newly designed internal primerswere used to amplify smaller fragments (Appendix 3) under thesame PCR conditions (Simon et al., 1994). PCR fragments werepurified using Millipore Multiscreen 96-well plates (Millipore,
282 S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293
Author's personal copy
Billerica, MA, USA). Sequencing was done from both strands withthe same primers used for PCR amplification, using BigDye 2.1and an ABI PRISM3730 automated sequencer (Applied Biosystems).Sequence chromatograms were assembled and edited withSequencher 4.6 (Gene Codes Corp., Ann Arbor, MI, USA). All se-quences included in the analysis have been submitted to the EMBLNucleotide Sequence Database (see Appendix 4 for accessionnumbers).
2.2. Phylogenetic analysis and species delimitation
Genetic clusters representing putative species were establishedusing the generalized mixed Yule-coalescent (GMYC) model (Ponset al., 2006; Fontaneto et al., 2007). This technique detects the tran-sition from between-species to within-population branching pat-terns delimiting ‘independently evolving’ mtDNA clusters.Starting trees for this analysis were obtained from all full-lengthcox1 sequences, after identical haplotypes were collapsed, Maxi-mum Likelihood analysis was performed with RAxML 7.0.4 (Sta-matakis, 2006) under two data partitions (1st plus 2nd vs 3rdcodon positions) and a GTR + I + C substitution model. The result-ing topology was made ultrametric under a molecular clock modelin PAUP� 4.0b10 (Swofford, 2002) using the model parameters esti-mated from the ML search. Model choice and the specific techniquefor establishing ultrametricity may impact the outcome of theGMYC model. While this may change the precise branch length esti-mates, the shift in branching rates that defines the threshold fordelimiting the clusters is unlikely to be affected in the current dataset because most GMYC groups are composed of very tight branch-ing clusters that are very distant from others (see below). The GMYCmethod generally has been shown to be resilient to variation inparameters. For example, a model that permits variable thresholdvalues applied to each cluster individually (rather than a uniformthreshold applied to the entire tree as in the simple model appliedhere) did not significantly improve the likelihood of the model in acomplex sample of sequences from insects of Madagascar (Mona-ghan et al., 2009). Likewise, the tree topologies used here were reli-able, at least with respect to the nodes defining GMYC groups whichin most cases showed 100% recovery in 1000 bootstrap replicatesusing the fast bootstrap procedure in RAxML (Stamatakis et al.,2008). Denser sampling of the tree (from additional species andgeographic sites) may reduce the strength of the inference, but clus-tering has been shown to remain strong even if population sam-pling becomes largely complete (Pons et al., 2006). The GMYCanalysis was conducted with the R package ‘splits’ (SPecies LImitsby Threshold Statistics) (http://r-forge.r-project.org/projects/splits)with the ‘single threshold’ option (Pons et al., 2006).
Each GMYC group was labeled with a preliminary species num-ber, which can be updated when descriptions (species names) areavailable. The voucher number of the Natural History Museum(‘‘BMNH” followed by six digits) was included in the GenBank sub-mission with each sequence. After DNA extraction, specimens wereprepared for morphological identification and voucher specimenswere handed to the collection managers at the Natural History Mu-seum, London, while additional specimens will be stored at theUniversidad de Panamá.
One individual per GMYC group was selected for in-depth phy-logenetic analysis based on three genes (cox1, rrnL and 28S). No in-dels were observed in cox1 while there was only limited lengthvariation for rrnL and 28S, which were aligned with the L-INS-Imethod in Mafft 6.240 (Katoh et al., 2005; Katoh and Toh, 2008).Combined data matrices were used for Bayesian Inference underdifferent substitution models for up to six separate partitions. Thepreferred partitioning scheme was selected using the Bayes factor,requiring at least a 10 lnL increase in the harmonic mean per addi-tional free parameter before a more complicated model is accepted
(Miller et al., 2009). Each partitioning scheme was analyzed withMrBayes 3.1.2, selecting the most appropriate model for each parti-tion in MrModeltest 2.2 based on the Akaike Information Criterion(Nylander, 2004). Bayesian Inference involved two parallel runsusing one cold and two incrementally heated Markov chains(k = 0.1) sampling every 1000 steps. The number of generations un-til convergence varied between partitioning schemes (10–30 mil-lion), which were assessed with the AWTY software (Wilgenbuschet al., 2004). Trees were summarized using the all-compatible con-sensus command with burn-in calculated based on diagnosticparameters in Tracer 1.4.1 (Drummond and Rambaut, 2007).
2.3. Phylogenetic niche conservatism
Phylogenetic community analysis of host use was performedusing the species-level tree of Conotrachelus onto which host plantswere mapped. We tested patterns of phylogenetic communitystructure of oligophagous and polyphagous species to elucidatethe influence of host preference and trait conservatism (i.e., phylo-genetic clustering) or competition (i.e., phylogenetic overdisper-sion) during species divergence and assembly of communities(Emerson and Gillespie, 2008; Cavender-Bares et al., 2009). Thiswas carried out with Phylocom 4.0.1 (Webb et al., 2008) calculatingthe Net Relatedness Index (NRI) and Nearest Taxon Index (NTI) ofspecies of Conotrachelus feeding on the same host plant species.These indices associate observed (between weevil species presenton a particular plant species) and simulated (between random sam-ple of weevils species) Mean Phylogenetic Distance (MPD) and theMean Nearest phylogenetic Taxon Distance (MNTD). This null dis-tribution from re-sampled datasets was created using Model 0 inPhylocom in which terminals are shuffled on the phylogenetic tree,thereby randomizing the phylogenetic relationships. The number ofpermutations was set to 9999. A low negative NRI and NTI indexindicates that species using a single host are phylogenetically over-dispersed (from phylogenetic diverse origins, suggesting that hostshifts are common and may proceed to phylogenetically distanthosts) and a high positive index indicates that species are phyloge-netically clustered (conservation of host specificity, diversificationof herbivores remains within closely related host associations).For each host, we considered as significantly clustered or overdi-spersed the community of herbivores if the observed phylogeneticdistance was above or below 2.5% of the null distribution distances,respectively (a = 0.05). Wilcoxon tests were used to test whetherNRI and NTI values differed from zero.
2.4. Morphological fruit preference
Host choice may be attributed to the ability of the insect to ex-ploit a specific part of the plant. To analyze if oligophagous speciesare constrained by plant morphology we conducted clustering anal-ysis using pairwise dissimilarities (Euclidean distances) betweenfruits using several morphological fruits traits: fleshy/non-fleshy,number of seeds, seed size and seed shape (round/flat), and lifeform of host plant (Table 1). These traits were estimated as anaverage of all the fruits collected and corroborated by existingdatabases (http://striweb.si.edu/esp/tesp/plant_species_a.htm).Analyzes were carried out using the Cluster library implementedin the R package (http://www.r-project.org).
3. Results
3.1. Larval and adult stage associations and species delimitation
A total of 887 cox1 sequences of 700–765 bp were obtainedfrom reared larval and adult specimens. When subjected to phylo-
S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293 283
Author's personal copy
genetic analysis a highly supported clade was found that includedall adult specimens morphologically identified as Conotrachelus.This clade comprised 483 sequences, 77 from adults and 406 fromlarvae, which were attributed to the genus Conotrachelus (all otherswere not analyzed further for this study). This set of Conotrachelussequences was filtered keeping 121 unique haplotypes, whichwere used for an ML tree search and subsequent calculation ofbranch length under a molecular clock. The resulting tree was sub-jected to the GMYC analysis, which grouped these haplotypes in 48putative species (Fig. 1). Re-including the duplicated haplotypes,14 GMYC groups were composed of both adults and larvae,whereas 2 and 32 entities were composed exclusively of eitheradults or immature stages, respectively. Most GMYC groups werecomposed of more than five sequences, with a maximum of 68specimens, while four singletons were encountered (Conotrachelussp.43, C. sp.49, C. sp.29 and C. sp.36; Table 2).
Sequencing of the nuclear 28S marker was successful for 394 ofthe 483 Conotrachelus individuals. Sequence length ranged from600 to 671 bp and variation was very low showing only 11 differ-ent haplotypes in total, which differed from each other by 1–10 bpsand showed single insertions/deletions of 2 base pairs. The 28Smarker was invariant within any of the cox1-defined GMYC groups.However, the 28S haplotypes were not congruent with the cox1-based phylogenetic tree. The most common 28S haplotype (#5)was found in 15 GMYC groups, but these were separated into fiveclades scattered throughout the cox1 tree (Table 3). Likewise, an-other haplotype (#6) was present in four species that were widelyseparated based on the cox1 analysis. The remaining 28S haplo-types showed greater, but not complete congruence with thecox1 tree, or haplotypes were unique to individual GMYC groups(Table 3).
3.2. Assigning reared species to existing taxonomy and free-caughtsamples
The reared adult specimens were attributed to a total of 16 spe-cies that could be separated on morphological differences, each ofthem being represented by between 1 and 18 individuals. The mor-phological delimitations based on several external traits includingthose employed in higher-level morphological group recognition(e.g. Fiedler, 1940) (Table 4), matched precisely the extent of themtDNA-based GMYC groups confirming these species entities. Fiveadult morphospecies were assigned to existing Linnaean namesallowing association with the corresponding GMYC groups (Ta-ble 2), while the remaining 11 adult morphospecies could not beassigned to existing names despite extensive studies of the litera-ture and comparisons with the collections in Panamá and at theNatural History Museum in London; they are likely to constituteundescribed species.
The collection of reared specimens was further compared to col-lections made during the same period at BCI from flight intercepttraps. In total, 17 morphospecies (28 specimens) of Conotracheluswere sequenced from these traps, of which only one, Conotrachelusinexplicatus, could be identified morphologically. Moreover, two ofthe specimens collected in traps matched two undescribed speciesreared from seeds (Conotrachelus sp.8 and C. sp.38) while theremaining species from the rearing effort were not obtained inthe intercept trap samples.
DNA amplification of the museum specimens of Conotrachelusfrom Panamanian forests was used as a potential way to link thelarval GMYC groups to known species. This resulted in successfulamplification of cox1 sequences from 24 species. Full-lengthfragments of 765 bp could be assembled from shorter overlapping
Table 1Fruit morphological traits employed in clustering analysis. Life form: tree = 1, midstory = 2, liana = 3, understorey = 4 and shrub = 5.
Host plant Fleshy/non-fleshy Average of seeds/fruit Average seed size (cm) Seed form round/flat Life form
284 S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293
Author's personal copy
Fig. 1. Ultrametric tree of unique cox1 haplotypes of Conotrachelus, showing the threshold used by the GMYC model to delimit species. Species in black labels contain adultspecimens and species in gray larvae.
S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293 285
Author's personal copy
Table 2Summary of Conotrachelus species reared from seeds with associated host plants, collected during the fruiting seasons of 2006–2007. Sites in Panama: Barro Colorado Island (BCI),Bocas del Toro (BOC), Fortuna National Reserve (FOR), Campana National Park (CAM) and San Lorenzo National Park (SHE) and La Selva Biological Station (LSE) in Costa Rica.Ni = total number of COI sequences obtained, Nh = number of COI haplotypes per species, number of larvae/adults sequenced. Pupation substrate observed in weevils species (P)F: inside fruit, S: in soil and U: unknown.
Species Ni Host group Sites Nh P Larvae/adults
Eurosid IChrysobalanaceae
Conotrachelus sp.42 2 Maranthes panamensis SHE 1 U 2/0Clusiaceae
Conotrachelus mixtus 3 Chrysochlamys eclipes BCI 1 S 0/3Clusiaceae/Apocynaceae
Conotrachelus punctiventris 5 Calophyllum longifolium/Lacmellea panamensis BCI 1 S 1/4Fabaceae–Mimosoideae
Conotrachelus rufescens 18 Inga thibaudiana BCI 1 S 6/12Conotrachelus sinuatocostatus 19 Inga leiocalycina/I. capitata FOR 4 S 17/2Conotrachelus sp.2 68 Inga laurina, I. ruiziana, I. sapindoides, I. thibaudiana/I. sapidnoides BCI/LSE 4 S 53/15Conotrachelus sp.4 21 Inga laurina, I. ruiziana BCI 3 S 19/2Conotrachelus sp.5 40 Inga capitata, I. leiocalycina, I. marginata FOR 4 S 34/6Conotrachelus sp.11 2 Inga spectabilis/I. jinicuil BOC/LSE 1 S 2/0Conotrachelus sp.14 6 Inga jinicuil, I. marginata FOR 5 S 6/0Conotrachelus sp.15 10 Inga laurina/I. marginataI/I. alba,I. jinicuil/I. spectabilis BCI/FOR/LSE/BOC 6 U 10/0Conotrachelus sp.18 21 Inga alba, I. jinicuil, I. edulis, I. oesterdiana LSE 4 U 21/0Conotrachelus sp.19 11 Inga laurina/I. cocleensis, I. sp. BCI/SHE 2 U 11/0Conotrachelus sp.24 8 Inga jinicuil/I. marginata LSE/HOR 4 U 8/0Conotrachelus sp.32 2 Inga alba, jinicuil LSE 1 U 2/0Conotrachelus sp.34 2 Inga jinicuil LSE 1 U 2/0Conotrachelus sp.43 1 Inga capitata FOR 1 U 1/0Conotrachelus sp.50 2 Inga jinicuil/I. nobilis FOR/BOC 1 U 2/0Conotrachelus sp.36 1 Inga punctata FOR 1 U 1/0Conotrachelus sp.37 3 Inga jinicuil BCI/BOC 3 U 3/0Conotrachelus sp.47 14 Inga laurina, I. ruiziana, I. sapindoides BCI 3 U 14/0Conotrachelus sp.48 21 Inga spectabilis FOR 5 S 21/0
FlacourtiaceaeConotrachelus sp.13 13 Lacistema aggregatum BCI 3 S 6/7
MoraceaeConotrachelus sp.16 4 Trophis racemosa BCI 2 S 4/0
CaryophyllalesPolygonaceae
Conotrachelus pumilio 18 Coccoloba manzanillensis BCI 5 F 0/18Conotrachelus sp.6 2 Coccoloba manzanillensis BCI 1 S 2/0Conotrachelus sp.10 17 Coccoloba manzanillensis BCI 4 S 16/1Conotrachelus sp.23 2 Coccoloba parimensis BCI 2 U 2/0
Eurosid IIAnacardiaceae
Conotrachelus sp.8 9 Spondias mombin, S. radlkoferi BCI 3 S 6/3Conotrachelus sp.21 3 Anacardium excelsum BCI 1 S 3/0
MeliaceaeConotrachelus sp.17 2 Guarea grandiflora LSE 1 U 2/0Conotrachelus sp.26 8 Guarea grandiflora LSE 1 U 8/0
SapindaceaeConotrachelus sp.20 3 Cupania sylvatica BCI 3 S 3/0Conotrachelus sp.27 9 Paullinia ingifolia LSE 4 U 9/0Conotrachelus sp.29 1 Paullinia ingifolia LSE 1 U 1/0Conotrachelus sp.49 1 Cupania rufescens BCI 1 S 1/0
RosidMyrtaceae
Conotrachelus sp.9 20 Eugenia nesiotica BCI 4 S 19/1Conotrachelus sp.22 4 Eugenia venezuelensis BCI 4 U 4/0Conotrachelus sp.44 3 Eugenia galalonensis BCI 1 S 3/0
Myrtaceae/SapindaceaeConotrachelus sp.45 50 Eugenia nesiotica, Cupania rufescens, C. sylvatica BCI 13 S 49/1
AsteridsApocynaceae
Conotrachelus sp.12 2 Lacmellea panamensis BCI 1 S 1/1Sapotaceae
Conotrachelus sp.1 1 Pouteria sp. LSE 1 U 0/1Conotrachelus sp.28 14 Pouteria calistophylla LSE 1 U 14/0Conotrachelus sp.31 3 Pouteria calistophylla LSE 1 U 3/0
MagnoliidsMyristicaceae
Conotrachelus sp.35 3 Virola surinamensis BCI 2 U 3/0Conotrachelus sp.38 1 Celastraceae CAM 1 U 4/0Conotrachelus sp.39 5 Unknown CAM 2 U 5/0Conotrachelus sp.40 1 Quercus lancifolia CAM 1 U 1/0Conotrachelus sp.41 1 Unknown CAM 1 U 1/0
286 S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293
Author's personal copy
segments in five species, while others were shorter (Appendix 5).Including these sequences in the full dataset in the GMYC speciesdelimitation procedure matched three specimens from flight inter-cept traps as they joined clusters with museum specimens identi-fied as Conotrachelus brevisetis, Conotrachelus verticalis and C. sp.2,the latter being an undescribed species (Fig. 2). None of these se-quences matched those of larvae or adults reared directly fromthe seeds. In summary, the DNA sequencing exercise resulted inthe recognition of 83 species that were unique to either seed col-lecting, flight intercept collecting or museum samples, while threespecies were obtained both in museum samples and flight inter-cept traps and two species were obtained by rearing and flightintercept traps.
In addition, the hope of linking the unknown species to a largerspecies groups via morphological characters was not fulfilled. Sixexternal characters employed previously to describe speciesgroups within Conotrachelus (Fiedler, 1940) were scored for all spe-cies represented by adults but these were highly homoplasticwhen mapped on the phylogenetic trees from DNA (Table 4). Thisputs into question the validity of the taxonomic species groups andsubgroups as defined in the existing literature, which are largelybased on these characters. Where species names were available,the position of the species on the DNA-based tree were comparedto the groupings proposed by Fiedler (1940), but congruence withthe existing groups was weak (Table 4). The only case of mono-phyly was that of C. inexplicatus and Conotrachelus chevrolati, as-signed to Fiedler’s Group III, Subgroup III, among a total of 12species that could be identified and were also present in Fiedler’s(1940) treatment (Table 5). It should be noted, however, that thesespecies were represented by museum specimens, with limited se-quence information, thus their placement in the DNA-based treemay be uncertain.
3.3. Host associations and specificity
Out of 48 species for which we had feeding records from rear-ing, 14 species were associated with the genus Inga (Fabaceae),
four species with Coccoloba spp. (Polygonaceae), four species withdifferent Eugenia spp. (Myrtaceae), three species with Cupania(Sapindaceae) and three species with Pouteria spp. (Sapotaceae).The remaining Conotrachelus species were each associated withdifferent plant species (Fig. 3). Conotrachelus species showed vari-ous degrees of specificity; 32 species were monophagous (67%), 14species oligophagous (29%), feeding on at least two species of con-generic plants, and two (4%) species were polyphagous feeding onseveral non-congeneric plants. One of the polyphagous species,Conotrachelus sp.45, fed on Eugenia nesiotica, Cupania rufescensand Cupania sylvatica, with 7, 15 and 28 specimens, respectively,and composed of a total of 13 different haplotypes. These were re-stricted either to C. sylvatica (eight haplotypes), C. rufescens (twohaplotypes), E. nesiotica (two haplotypes) or shared between C.sylvatica and C. rufescens (one haplotype). The other polyphagousspecies was Conotrachelus punctiventris, composed of a single hap-lotype only and which fed on Lacmellea panamensis (two speci-mens) and Calophyllum longifolium (three specimens) (Fig. 3).
Host associations were analyzed in a phylogenetic frameworkusing the tree based on three gene fragments (cox1, rrnL and28S) and one representative for each GMYC group. Bayesian infer-ence was conducted under six different partitioning schemes. Thepreferred Bayesian tree was obtained applying separate substitu-tion models to four different partitions (cox1 1st and 2nd, 3rd co-don position, 28S, rrnL; P4 in Table 6), although otherpartitioning schemes produced very similar trees and mainly dif-fered in the nodes with low support. The selected tree (Fig. 3)was used to map host choice of the 48 reared Conotrachelus species.The analysis revealed a complex distribution of 36 host species and17 host genera on this tree, but showed some clear structuringaccording to host plant associations. All except one species feedingon the genus Inga were members of four well-supported clades(PP = 1.0) entirely confined to this host (Clades 1, 5, 6 and 8;Fig. 3). Four species exclusively feeding on Coccoloba manzanillensisand C. parimensis also formed a clade (Clade 4; PP = 1.00), whilethree species feeding on Pouteria calistophylla were monophyleticexcept for the inclusion of two species feeding on Eugenia (Clade3; Fig. 3). In addition, two pairs of sister species were found to feedon the same hosts: Conotrachelus sp.20 and C. sp.49 feeding on Cup-ania, and Conotrachelus sp.9 and C. sp.44 feeding on Eugenia. Inboth cases the hosts were also used by one distantly related speciesof Conotrachelus.
The analysis of host breadth in oligophagous and polyphagousspecies with Phylocom showed that Conotrachelus species foundon the same host plant were significantly more closely related toeach other than expected by chance (NRI = 1.03 ± 1.28 andNTI = 1.08 ± 1.17; Wilcoxon signed rank test, p < 0.05). Both theNTI values that measure the clustering of the terminal tips andthe NRI measurements were significant (two-tailed test on rankat the distribution of permuted null communities for p = 0.05).Hence, despite numerous host shifts (Fig. 4), host-usage in Conotra-chelus in oligophagous and polyphagous species is a phylogeneti-cally conserved trait.
Table 3Haplotypes of 28S obtained for the species of Conotrachelus with the correspondingnumber of GMYC species and number of separate clades on the cox1 tree in Fig. 3.
28S haplotype GMYC species Cox1 clades
1 2 22 2 23 2 14 15 15 56 4 37 3 38 5 19 2 2
10 111 1
Table 4Morphological characters of Conotrachelus used to delimit species groups, their presence in the individuals scored, and their degree of consistency with the mtDNA based tree of.
Characters Character states No individuals Consistency index Retention index Steps
State (1) State (2) State (3)
Fore coxae (1) Joint and (2) separate 30 16 0.06 0.20 35Mesosternum (1) Plain, (2) excavated, and (3) barely excavated 33 11 2 0.09 0.21 34Metasternum (1) Plain and (2) excavated 37 9 0.06 0.20 35Rostrum length (1) Short, (2) long, and (3) intermediate 35 1 10 0.06 0.23 34Rostrum width (1) Wide, (2) cylindrical, and (3) intermediate 30 15 1 0.08 0.15 37Body type (1) Diaconitus and (2) dentimanus 33 13 0.06 0.20 35
S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293 287
Author's personal copy
3.4. Host choice constrained by fruit morphology
Based on various quantitative measures of fruit morphology,nearly all species of Inga grouped together, except for Inga specta-bilis and Inga edulis with >20 seeds per pod. This trait coincidedwith the principal separation at the base of the cluster analysis(1 in Fig. 4). These two species of Inga were used by four speciesof Conotrachelus, which also fed on a range of other Inga, withthe exception of C. sp.48 that was only found on I. spectabilis. Like-wise, within the much larger cluster with <20 seeds, Inga feedingspecies frequently had associations with multiple host species,and the host ranges did not group in any apparent ways accordingto the morphological clustering (2 in Fig. 4). All remaining hostspecies clustered together showing no clear substructure accordingto their phylogenetic relationships, except at the genus level. Phy-logenetic relationships of the corresponding herbivore species alsodid not show any clear structure with regard to the morphologicalclustering of their hosts. However, the two polyphagous species ofConotrachelus fed on host species with similar seed characteristics.
C. punctiventris fed on L. panamensis and C. longifolium, which differonly in seed number (2 vs 1) and size (1 vs 2 cm). Conotrachelussp.45 fed on two species of Cupania and E. nesiotica that sharethe same number of seeds and range of seed size and cluster inthe morphological analysis (Fig. 4), thus showing the affinity of thisspecies to particular morphological traits.
4. Discussion
4.1. Towards a DNA-based taxonomy of Conotrachelus
‘‘How many species of organisms are there on Earth? We don’tknow, not even to the nearest order of magnitude” (Wilson, 1985).This quote by Edward O. Wilson is as true today as it was 25 yearsago. However the use of DNA in the study of biodiversity has re-cently emerged as a promising tool to speed up the explorationof the remaining unknown species diversity. Mitochondrial cox1is being used extensively in DNA barcoding studies (Hebert et al.,2003), and the use of DNA might have the greatest impact in thetaxonomically poorly known, highly diverse groups like tropical in-sects (Moritz et al., 2000; Pons et al., 2006; Monaghan et al., 2009).Samples with a high proportion of undescribed species and lackingupdated taxonomic revisions (for morphological identification)have been deemed inaccessible for biological research (Wheeler,2004), but can now be studied using DNA-delimited evolutionaryunits. The use of DNA also opens new doors in ecological commu-nity studies, e.g., by linking various life stages (Ahrens et al., 2007;Levkanicova and Bocak, 2009), identifying host species from the in-sects’s ingested material (Matheson et al., 2007; Jurado-Riveraet al., 2009; Pinzón-Navarro et al., 2010) and analyzing the
Fig. 2. Bayesian inference from cox1, rrnL and 28S combined data matrices for 83 Conotrachelus species including specimens from seed rearing, museum collections and flightintercept traps. Only one representative per GMYC group is included. Species that were recovered by two collecting methods are marked in bold letters.
Table 5Conotrachelus sp. and their assignment to species groups by Fiedler (1940).
Group Subgroup Conotrachelus sp.
II IV albofrontalis, brevisetis, verticalisII X pumilioIII I pectoralisIII III ibis, inexplicatus, chevrolati, scaberIII IV sinuatocostatus, planifronsIV V deplanatus
288 S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293
Author's personal copy
Fig. 3. Bayesian inference of Conotrachelus seed feeders from combined analysis of three genes, with host plants and degree of specificity (red: monophagous, black:oligophagous and gray: polyphagous) mapped on the tree.
S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293 289
Author's personal copy
phylogenetic structure of local species assemblages (Webb et al.,2002; Emerson and Gillespie, 2008; Cavender-Bares et al., 2009).
Here, the use of DNA revealed a high proportion of undescribedspecies, as expected from previous studies of tropical forest insectsthat also discovered numerous unknown species (Ødegaard, 2003)and a seemingly immense number of weevil species (Oberprieleret al., 2007). Based on our analysis, the sampling technique had agreat impact on the species encountered, with only two species incommon between the reared (48 species) and flight intercept trapcaptures (17 species). In contrast, within the cohort of reared indi-viduals, the rate of multiple captures of a species was much higher.Presence at two or more sites (in Panama and Costa Rica) was estab-
lished for seven species of Conotrachelus, while most of the speciescollected in BCI were present in various surrounding islands andother localities in Panama. The lack of overlap in the species compo-sition between freshly-collected and museum loaned species mightbe explained by the fact that most of the 135 species known fromPanama and Costa Rica (O’Brien and Wibmer, 1984) were not col-lected, suggesting that most of the reared species are unknown,as all previous taxonomic work has been based on free-livingadults. Cryptic lifestyles, long emergence times or low mobility ofadults may prevent many species from being encountered in traps.
The validity of the GMYC clusters to reflect Linnaean specieswas strongly corroborated by the fact that virtually all clusters
Table 6Bayes factor comparisons for selecting the best partitioning scheme. Five alternative partitioning schemes (P0–P5) were compared as described in the text. Numbers in bracketsindicate the total number of free parameters required for each partitioning scheme. Below the diagonals: ln(Bayes factor) (estimated as the difference of the natural logarithms ofthe harmonic mean of the likelihoods between two partition schemes). Above the diagonal: ln(Bayes factor)/Dp (where Dp = difference in total number of free parametersbetween two partition schemes).
Fig. 4. Cluster dendrogram based on fruit morphological traits with their corresponding species of Conotrachelus feeding on them. Phylogenetic structure of the ‘community’structure of Conotrachelus present on each plant species is classified either as phylogenetic clustering (e), phylogenetic overdispersion (�) or random patterns (�). Twopolyphagous species are marked in bold.
290 S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293
Author's personal copy
were congruent with entities that would be considered as separatemorphospecies by taxonomic experts. This study of Conotrachelustherefore adds to the growing list of insect lineages (e.g., Hebertand Gregory, 2005; Monaghan et al., 2005) showing this clean sep-aration of mtDNA lineages in step with the morphological varia-tion. While this finding strongly supports the GMYC approach,the nuclear 28S marker showed a surprisingly complicated patternnot fully congruent with the mtDNA-based groups. The overall le-vel of variation was very low, with only 11 haplotypes encounteredin 48 species (mtDNA-based GMYC groups), leaving many speciesundistinguishable by this marker, unlike other groups where sepa-rate species differ by at least one base pair (Markmann and Tautz,2005; Monaghan et al., 2005, 2009; Ahrens et al., 2007). In addi-tion, while no variation was observed within any of the GMYCgroups, most of the 28S haplotypes also extend to additional GMYCgroups (Table 3). Processes of concerted evolution affecting this lo-cus might cause incongruence of the gene and species trees,although the degree to which this occurs here is disconcertingand leaves some uncertainty about the validity of the mtDNA-based GMYC groups.
Finally, six key morphological characters were scored for allspecies represented by adults but these were revealed to be highlyhomoplastic when mapped on the phylogenetic trees from DNA(Table 4), while expected relationships from Fiedler’s (1940)groups were not recovered in the current sequence analysis. Hence,the current morphological classification, at least at the speciesgroup level may be misled by high levels of convergence, due tothe evolutionary plasticity of functional traits such as body shapeand rostrum length and width. In contrast, the DNA-based tree iscorroborated by the host choice, which was generally conservative.For example, four Inga feeding clades and four further clades of twoto four species feeding on Pouteria, Eugenia, Cupania and Coccolobanow constitute a provisional framework for grouping of additionalspecies that may be added to the tree in future. Although theseclades are composed of a small number of species and thereforeof limited value for a classification of the genus as a whole, theyprovide a more secure grouping than the current morphologicallydefined groups, which relied on highly homoplastic traits.
4.2. Evolution of species diversity and host choice
The great species richness of Conotrachelus raises the questionabout the role of the host plants in promoting diversity. Plant–her-bivore interactions can be regarded as a micro-community, amena-ble to phylogenetic analysis of the assemblage of co-occurringspecies (Webb et al., 2002; Vamosi et al., 2009; Cavender-Bareset al., 2009). Interactions with the hosts and the evolution of novelplant defense mechanisms are a driver of adaptive radiation(Anderson, 1995; Kursar and Coley, 2003; Sequeira and Farrell,2001; Kergoat et al., 2007), while various events of host rangeexpansion and subsequent specialization of niche space may alsofavor the evolution of host races and ultimately speciation drivenby host shifts (Janz et al., 2001, 2006; Nosil, 2002). The analysesof the combined data showed a significant trend towards phyloge-netic conservation of the host plants, while host shifts and re-peated evolutionary transitions to particular host plants werealso common events. In addition, host associations of a weevil spe-cies frequently extended to multiple plant species, although thesewere generally phylogenetically conservative, as evident from thepositive NRI and NTI scores. Taken together, this argues against aspecific role of host switches as the cause of species diversification,but the unexpectedly broad taxonomic range of hosts utilized byvarious lineages of the genus and the repeated shifts to the majorhost lineages, in particular the genus Inga, may contribute to theradiation.
Host-associated diversification may be coupled with vicarianceevents as an alternative mode of speciation. Thus, the observedpatterns of host associations may be affected by the spatial distri-bution of host plants and other environmental conditions thatdetermine the presence of a host. However, we found that closelyrelated species of Conotrachelus associated with a single host planthad a tendency to co-occur at a single site rather than being foundat multiple sites as expected under allopatric speciation. For exam-ple, the three species reared from Coccoloba (Clade 4) were encoun-tered only on BCI, while three Pouteria-feeding species (Clade 2)were exclusive to LSE and two Cupania-feeding species (Clade 7)again were exclusive to BCI (Table 2). The Inga feeding communi-ties were more complex because most species had multiple hostsdue to turnover of both Conotrachelus and Inga sp. among the sixgeographic sites sampled. Replication of host availability at multi-ple sites was very limited, possibly due to differences in fruitingseasons; only Inga sapindoides (found in BCI and LSE) and Inga jini-cuil (in BCI, BOC and LSE) were present at more than a single site(Table 2). Hence, the widespread GMYC groups inevitably altertheir hosts among sites. The findings suggest that for many speciesof Conotrachelus the precise Inga host is not fixed and feeding pref-erence depends on the temporal and local availability of the host,rather than narrow host specificity. Additional studies will be nec-essary to establish levels of beta diversity among host plants andamong geographic areas to analyze the patterns of host specializa-tion for monophagous vs oligophagous species in greater detail.Moreover, host shifts and race formation are likely to contributeto species formation by occasional shifts to distantly related hosts.Supporting this is the nearly perfect host segregation of haplotypesin the polyphagous Conotrachelus sp.45 indicating that hostswitches may be more relevant to host race formation and specia-tion when phylogenetically distant host plants are utilized.
While we focused here on the phylogeny of the herbivores,more explicit questions could also be asked about the phylogeneticrelationships of the plants, in particular those used by oligopha-gous and polyphagous herbivore species. We found that most her-bivores are specialists on a single species or a set of closely relatedspecies. This is in agreement with, for example, a study of insectherbivores in a New Guinean rainforest (Weiblen et al., 2006),which reported a high proportion of oligo- and polyphagous herbi-vores to feed on closely related plants, by using an explicit phylog-eny of plants. Our study used the genus level to provide a measureof phylogenetic proximity of the host given the good taxonomicknowledge of the plant species and the high fidelity of host useat this level. In contrast to the existing studies, we used the phylog-eny of herbivores itself to establish the kind of host switches in-volved in host use. In this way we focus on a slightly differentquestion, asking if the host choice in the herbivore lineage experi-ences any constraints by the plant phylogeny. This seems moreappropriate than previous studies, to answer the question abouthow the host choice might affect the diversification of insect lin-eages because the analysis specifically focuses on the step size ofhost switches in the herbivore. Thus, phylogenetic reconstructionof host choice and tests of conservation of host shifts in individuallineages may be more informative to assess the evolutionary pro-cesses underlying the diversification of host–herbivore systems.
The analysis of associations with morphological host traitsshowed that the morphology-based clusters were mostly deter-mined by the broad separation of Inga and other plants (Fig. 4).These phenetically defined groups were utilized by different spe-cies of Conotrachelus, specifically separating the Inga-feedinggroups from all others. Within the Inga-feeding group there wasno clear pattern of host-usage, e.g., seed size or number, nor wasthere a correlation of host morphology and herbivore phylogeny.In the group not feeding on Inga, only two species showed correla-tion with seed morphology, in particular the polyphagous C. sp.45
S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293 291
Author's personal copy
feeding on the morphologically similar, but distantly related gen-era Eugenia and Cupania. Overall, the study was not fully conclusiveregarding the relevance of seed morphology for host selection. Theobservations perhaps are reminiscent of other seed-feeding wee-vils like Curculio feeding on acorns of Quercus and other Fabaceae,which show a clear correlation to seed size and shape (Bonal andMuñoz, 2009) but host preferences are not correlated with phylo-genetic proximity (Hughes and Vogler, 2004; Bonal and Muñoz,2009). Unlike Curculio, which is distributed mostly in the temper-ate zone and only feeds on single-seeded fruits, Conotrachelus has agreater range of host species available in tropical forests and uti-lizes a greater diversity of fruit types, including seed in pods thatprovide fewer constraints on larval development. This broad re-source spectrum may have contributed to the great speciesrichness.
5. Conclusions
Approximately 1200 species of Conotrachelus have been de-scribed but, based on our finding of many unknown species in arelatively well studied area of the Neotropics, the true speciesnumber in this lineage must be much higher, as previous studieshave reported. However, even a relatively small number of individ-uals sampled here provide a reliable framework for the evolution-ary ecology of host use. Whereas host specificity can never becomprehensive given the great plant diversity and spatial and tem-poral turnover in the tropical forest, a combination of field obser-vations and phylogenetic analysis (at the level of species andhigher-levels) can provide the most important features of theinteractions and their evolutionary history.
Acknowledgments
We thank Salomon Aguilar, Osvaldo Calderon and Didimo Ureñ-a in Panamá and Orlando Vargas, Enrique Salicetti and NelsonZamora in Costa Rica for helping with plant identification. DouglasChesters provided scripts for collapsing haplotypes. Thanks toJohannes Bergsten for his help in previous versions. Thanks toPrograma Centroamericano de Maestría en Entomología of the Uni-versidad de Panamá for the loan of identified Conotrachelus speci-mens. This work was funded by the studentships from IFARHU-SENACYT-Panamá, CONACYT-México (196749) and the EuropeanUnion-Alban (E05D058103PA) to SPN.
Appendix A. Supplementary material
Supplementary material associated with this article can befound, in the online version, at doi:10.1016/j.ympev.2010.02.022.
References
Abdo, Z., Golding, G.B., 2007. A step toward barcoding life: a model-based, decision-theoretic method to assign genes to preexisting species groups. SystematicBiology 56, 44–56.
Ahrens, D., Monaghan, M.T., Vogler, A.P., 2007. DNA-based taxonomy for associatingadults and larvae in multi-species assemblages of chafers (Coleoptera:Scarabaeidae). Molecular Phylogenetics and Evolution 44, 436–449.
Anderson, R.S., 1995. An evolutionary perspective on diversity in Curculionoidea.Memoirs of the Entomological Society of Washington 14, 103–118.
Basset, Y., 1996. Local communities of arboreal herbivores in Papua New Guinea:predictors of insect variables. Ecology 77, 1906–1919.
Basset, Y., Corbara, B., Barrios, H., Cuenoud, P., Leponce, M., Aberlenc, H.-P., Bail, J.,Bito, D., Bridle, J.R., Castano-Meneses, G., Cizek, L., Cornejo, A., Curletti, G.,Delabie, J.H.C., Dejean, A., Didham, R.K., Dufrene, M., Fagan, L.L., Floren, A.,Frame, D.M., Halle, F., Hardy, O.J., Hernandez, A., Kitching, R.L., Lewinsohn, T.M.,Lewis, O.T., Manumbor, M., Medianero, E., Missa, O., Mitchell, A.W., Mogia, M.,Novotny, V., Odegaard, F., De Oliveira, E.G., Orivel, J., Ozanne, C.M.P., Pascal, O.,
Pinzon, S., Rapp, M., Ribeiro, S.P., Roisin, Y., Roslin, T., Roubik, D.W., Samaniego,M., Schmidl, J., Sorensen, L.L., Tishechkin, A., Van Osselaer, C., Winchester, N.N.,2007. IBISCA-Panama, a large-scale study of arthropod beta-diversity andvertical stratification in a lowland rainforest: rationale, study sites and fieldprotocols. Bulletin de l’Institut Royal des Sciences Naturelles de BelgiqueEntomologie 77, 39–69.
Becerra, J.X., 2007. The impact of herbivore–plant coevolution on plant communitystructure. Proceedings of the National Academy of Sciences of the United Statesof America 104, 7483–7488.
Berlocher, S.H., Feder, J.L., 2002. Sympatric speciation in phytophagousinsects: moving beyond controversy? Annual Review of Entomology 47, 773–815.
Bonal, R., Muñoz, A., 2009. Seed weevils living on the edge: pressures and conflictsover body size in the endoparasitic Curculio larvae. Ecological Entomology 34,304–309.
Bondar, G., 1946. Notas entomologicas da Baía. XVII. Revista de Entomologia, Rio deJaneiro 17, 78–113.
Cavender-Bares, Kozak, K.H., Fine, P.V.A., Kembel, S.W., 2009. The merging ofcommunity ecology and phylogenetic biology. Ecology Letters 12, 697–715.
Cognato, A.I., Vogler, A.P., 2001. Exploring data interaction and nucleotidealignment in a multiple gene analysis of Ips (Coleoptera: Scolytinae).Systematic Biology 50, 758–780.
Condon, M.A., Scheffer, S.J., Lewis, M.L., Swensen, S.M., 2008. Hidden neotropicaldiversity: greater than the sum of its parts. Science 320, 928–931.
Croat, T.B., 1978. Flora of Barro Colorado Island. Stanford University Press, Stanford,California USA, ix + 943 pp.
Ehrlich, P.R., Raven, P.H., 1964. Butterflies and plants: a study in coevolution.Evolution 18, 586–608.
Emerson, B., Gillespie, R.G., 2008. Phylogenetic analysis of community assembly andstructure over space and time. Trends in Ecology and Evolution 23, 619–630.
Fiedler, K., 1940. Monograph of the South American Weevils of the GenusConotrachelus. The British Museum, London.
Fontaneto, D., Herniou, E.A., Boschetti, C., Caprioli, M., Melone, G., Ricci, C.,Barraclough, T.G., 2007. Independently evolving species in asexual bdelloidrotifers. PLoS Biology 5, 914–921.
Hebert, P.D.N., Gregory, T.R., 2005. The promise of DNA barcoding for taxonomy.Systematic Biology 54, 852–859.
Hebert, P.D.N., Ratnasingham, S., Dewaard, J.R., 2003. Barcoding animal life:cytochrome c oxidase subunit 1 divergences among closely related species.Proceedings of the Royal Society B, 270 (Suppl.), S96–S99.
Hughes, J., Vogler, A.P., 2004. Molecular phylogeny of Curculio (Curculionidae)inferred from two mitochondrial and two nuclear data sets with differentnumbers of taxa. Molecular Phylogenetics and Evolution 32, 601–615.
Janz, N., Nyblom, K., Nylin, S., 2001. Evolutionary dynamics of host–plantspecialization: a case study of the tribe Nymphalini. Evolution 55, 783–796.
Janzen, D.H., 1980. Specificity of seed-attacking beetles in a Costa Rican deciduousforest. Journal of Ecology 68, 929–952.
Jenkins, D., Cottrell, T., Horton, D., Hodges, A., Hodges, G., 2006. Hosts of plumcurculio, Conotrachelus nenuphar (Coleoptera: Curculionidae), in CentralGeorgia. Environmental Entomology 35, 48–55.
Jurado-Rivera, J.A., Vogler, A.P., Reid, C.A.M., Petitpierre, E., Gómez-Zurita, J., 2009.DNA barcoding insect–hostplant associations. Proceedings of the Royal SocietySeries B 276, 639–648.
Katoh, K., Kuma, K., Toh, H., Miyata, T., 2005. MAFFT version 5: improvementin accuracy of multiple sequence alignment. Nucleic Acids Research 33, 511–518.
Katoh, K., Toh, H., 2008. Recent developments in the MAFFT multiple sequencealignment program. Briefings in Bioinformatics 9, 286–298.
Kergoat, G.J., Silvain, J.F., Delobel, A., Tuda, M., Anton, K.W., 2007. Defining the limitsof taxonomic conservatism in host–plant use for phytophagous insects:molecular systematics and evolution of host–plant associations in the seed-beetle genus Bruchus Linnaeus (Coleoptera: Chrysomelidae: Bruchinae).Molecular Phylogenetics and Evolution 43, 251–269.
Kursar, T.A., Coley, P.D. (Eds.), 2003. Convergence in defense syndromes of youngleaves in tropical rainforests. 4th International Legume Conference, Pergamon-Elsevier Science Ltd., Canberra, Australia, pp. 929–949.
Leskey, T.C., Wright, S.E., 2007. Host preference of the plum curculio. EntomologiaExperimentalis et Applicata 123, 217–227.
Levkanicova, Z., Bocak, L., 2009. Identification of net-winged beetle larvae(Coleoptera: Lycidae) using three mtDNA fragments: a comparison of theirutility. Systematic Entomology 34, 210–221.
Markmann, M., Tautz, D., 2005. Reverse taxonomy: an approach towardsdetermining the diversity of meiobenthic organisms based on ribosomal RNAsignature sequences. Philosophical Transactions of the Royal Society B,Biological Sciences 360, 1917–1924.
292 S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293
Author's personal copy
Matheson, C.D., Muller, G.C., Junnila, A., Vernon, K., Hausmann, A., Miller, M.A.,Greenblatt, C., Schlein, Y., 2007. A PCR method for detection of plant meals fromthe guts of insects. Organisms Diversity and Evolution 7, 294–303.
McKenna, D.D., Sequeira, A.S., Marvaldi, A.E., Farrell, B.D., 2009. Temporal lags andoverlap in the diversification of weevils and flowering plants. Proceedings of theNational Academy of Sciences of the United States of America 106, 7083–7088.
Miller, K.B., Bergsten, J., Whiting, M.F., 2009. Phylogeny and classification of thetribe Hydaticini (Coleoptera: Dytiscidae): partition choice for Bayesian analysiswith multiple nuclear and mitochondrial protein-coding genes. ZoologicaScripta 38, 591–615.
Missa, O., Basset, Y., Alonso, A., Miller, S., Curletti, G., Meyer, M., Eardley, C., Mansell,M., Wagner, T., 2009. Monitoring arthropods in a tropical landscape: relativeeffects of sampling methods and habitat types on trap catches. Journal of InsectConservation 13, 103–118.
Mitter, C., Farrell, B., Futuyma, D.J., 1991. Phylogenetic studies of insect–plantinteractions: insights into the genesis of diversity. Trends in Ecology &Evolution 6, 290–293.
Monaghan, M.T., Balke, M., Gregory, T.R., Vogler, A.P., 2005. DNA-based speciesdelineation in tropical beetles using mitochondrial and nuclear markers.Philosophical Transactions of the Royal Society B-Biological Sciences 360,1925–1933.
Monaghan, M.T., Inward, D.J.G., Hunt, T., Vogler, A.P., 2007. A molecularphylogenetic analysis of the Scarabaeinae (dung beetles). MolecularPhylogenetics and Evolution 45, 674–692.
Monaghan, M.T., Wild, R., Elliot, M., Fujisawa, T., Balke, M., Inward, D.J.G., Lees, D.C.,Ranaivosolo, R., Eggleton, P., Barraclough, T.G., Vogler, A.P., 2009. Acceleratedspecies inventory on Madagascar using coalescent-based models of speciesdelineation. Systematic Biology 58 (3), 298–311.
Moritz, C., Patton, J.L., Schneider, C.J., Smith, T.B., 2000. Diversification of rainforestfaunas: an integrated molecular approach. Annual Review of Ecology andSystematics 31, 533–563.
Morse, G.E., Farrell, B.D., 2005. Ecological and evolutionary diversification of theseed beetle genus Stator (Coleoptera: Chrysomelidae: Bruchinae). Evolution 59,1315–1333.
Nosil, P., 2002. Transition rates between specialization and generalization inphytophagous insects. Evolution 56, 1701–1706.
Novotny, V., Basset, Y., 2005. Review – Host specificity of insect herbivores intropical forests. Proceedings of the Royal Society B, Biological Sciences 272,1083–1090.
Nylander, J.A.A., 2004. MrModeltest v2. Program Distributed by the Author. UppsalaUniversity, Evolutionary Biology Centre.
O’Brien, C.W., Wibmer, G.J., 1984. Annotated checklist of the weevils (Curculionidaesensu lato) of North America, Central America, and the West Indies (Coleoptera:Curculionoidea). Southwestern Entomologist 9 (Suppl. 1), 286–307.
Ødegaard, F., 2003. Taxonomic composition and host specificity of phytophagousbeetles in a dry forest in Panama. Arthropods of Tropical Forests 220, 236.
Ødegaard, F., 2006. Host specificity, alpha- and beta-diversity of phytophagousbeetles in two tropical forests in Panama. Biodiversity and Conservation 15, 83–105.
Pinzón-Navarro, S., Jurado-Rivera, J.A., Gomez-Zurita, J., Lyal, C.H.C., Vogler, A.P.,2010. DNA profiling of host–herbivore interactions in tropical forests. EcologicalEntomology 35 (1), 18–32.
Pons, J., Barraclough, T.G., Gomez-Zurita, J., Cardoso, A., Duran, D.P., Hazell, S.,Kamoun, S., Sumlin, W.D., Vogler, A.P., 2006. Sequence-based speciesdelimitation for the DNA taxonomy of undescribed insects. SystematicBiology 55, 595–609.
Schoof, H.F., 1942. The genus Conotrachelus Dejean (Coleoptera, Curculionidae) inthe North Central United States. Illinois Biological Monograph 19, 1–170.
Sequeira, A.S., Farrell, B.D., 2001. Evolutionary origins of Gondwanan interactions:How old are Araucaria beetle herbivores? Biological Journal of the LinneanSociety 74, 459–474.
Silva, G., Bailez, O.E., Viana-Bailez, A.M., 2007. Dimorfismo sexual do gorgulho-da-goiaba Conotrachelus psidii Marshall (Coleoptera: Curculionidae). NeotropicalEntomology 36, 520–524.
Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H., Flook, P., 1994. Evolution,weighting, and phylogenetic utility of mitochondrial gene sequences and acompilation of conserved polymerase chain reaction primers. Annals of theEntomological Society of America 87, 651–701.
Stamatakis, A., 2006. RAxML-VI-HPC: maximum likelihood-based phylogeneticanalyses with thousands of taxa and mixed models. Bioinformatics 22, 2688–2690.
Stamatakis, A., Hoover, P., Rougemont, J., 2008. A rapid bootstrap algorithm for theRAxML web servers. Systematic Biology 57, 758–771.
Stehr, F.W., 1991. Immature Insects, vol. 2. p. 975.Swofford, D.L., 2002. PAUP�: Phylogenetic Analysis using Parsimony. Version 4.0b.
Sinauer Associates, Sunderland, MA.Valente, R.M., Gorayeb, I.S., 1994. Biologia e descrição dos imaturos de Conotrachelus
imbecilus Fiedler (Coleoptera: Curculionidade: Molytinae) em frutos de Ingaheterophylla. Boletim do Museu Paraense Emílio Goeldi, série Zoológica 2, 253–271.
Vamosi, S.M., Heard, S.B., Vamosi, J.C., Webb, C.O., 2009. Emerging patterns in thecomparative analysis of phylogenetic community structure. Molecular Ecology18, 572–592.
Webb, C.O., Ackerly, D.D., Kembel, S.W., 2008. Phylocom: Software for the Analysisof Community Phylogenetic Structure and Trait Evolution: Version 3.41.<http://www.phylodiversity.net/phylocom/>.
Webb, C.O., Ackerly, D.D., McPeek, M.A., Donoghue, M.J., 2002. Phylogenies andcommunity ecology. Annual Reviews of Ecology and Systematics 33, 475–505.
Weiblen, G.D., Webb, C.O., Novotny, V., Basset, Y., Miller, S.E., 2006. Phylogeneticdispersion of host use in a tropical insect herbivore community. Ecology 87,S62–S75.
Wheeler, Q.D., 2004. Taxonomic triage and the poverty of phylogeny. PhilosophicalTransactions of the Royal Society of London Series B-Biological Sciences 359,571–583.
Wilgenbusch J.C., Warren D.L., Swofford D.L., 2004. AWTY: A System for GraphicalExploration of MCMC Convergence in Bayesian Phylogenetic Inference. <http://ceb.csit.fsu.edu/awty>.
Wilson, E.O., 1985. The biological diversity crisis. Bioscience 35, 700–706.
S. Pinzón-Navarro et al. / Molecular Phylogenetics and Evolution 56 (2010) 281–293 293
