ee ta A 02 of Ca 4002 Ribe Article history: and has over 50 described species. Melipona, like Apis, possesses the remarkable ability to use represen- tational communication to indicate the location of foraging patches. Although Melipona has been the sub- (von Frisch, 1967; Michener, 1974; Dyer, 2002; Nieh, 2004). All Apis use a form of referential communication known as the waggle dance, whereby returning foragers inform colony members about newly discovered resource sites (von Frisch, 1967; Seeley, 1995; Dyer, 2002). The waggle dance communicates distance and direc- tion (von Frisch, 1967; Gould, 1976; Michelsen et al., 1992; Esch et al., 2001; Dyer, 2002; Sherman and Visscher, 2002). The commu- ences in the ability to communicate different spatial dimensions among species of Melipona, which correlate well with spatial distri- bution of ?oral resources in their current environment (Nieh et al., 2003). Whether and how this information is actually utilized by nest mates is still a subject of intense investigation, as was the case for decades in Apis. Although the genus has been the focus of behavioral, genetic, ecological, and pollination studies (Roubik, 2006), only partial phy- logenetic analyses have been carried out to date (Rego, 1990; Costa et al., 2003; Fernandes-Salom?o et al., 2005; Rasmussen and Cam- eron, 2010). The stingless bee genus Melipona is clustered within * Corresponding author. Present address: University of California, Berkeley, 137 Mulford Hall #3114, Berkeley, CA 94720-3114, USA. Molecular Phylogenetics and Evolution xxx (2010) xxx?xxx Contents lists availab ne .e l ARTICLE IN PRESSE-mail address: sramirez@post.harvard.edu (S.R. Ram?rez).1. Introduction The stingless bee genus Melipona contains at least 50 species of medium-sized (8?15 mm), robust, and often hirsute bees inhabit- ing forests of tropical America, from Mexico to Argentina (Schwarz, 1932; Michener, 2007). Most species of Melipona inhabit lowland wet forests, with the greatest species diversity concentrated in the Amazon Basin (Moure and Kerr, 1950). These bees are highly eusocial, which means they exhibit reproductive division of labor, cooperative brood care, and overlap of generations (Wilson, 1971). Similar to honey bees (Apis), Melipona are remarkable for in- sects, in their ability to recruit nest mates to speci?c foraging sites nication mechanisms of Melipona are less studied, but experimen- tal evidence indicates functional referential communication in some species (Esch, 1967; Aguilar and Brice?o, 2002; Nieh, 2004), but not in others (Hrncir et al., 2006). Upon returning to the nest, successful M. panamica and M. seminigra foragers may perform short piloting ?ights outside of the nest in the direction of the resource (Nieh, 1998; Nieh and Roubik, 1998), while inside the nest, they produce sound pulses while distributing food sam- ples to potential recruits (Esch, 1967; Nieh, 2004). The average duration of sound pulses correlates with, and thus potentially en- codes, distance to food sources relative to the location of the nest (Esch, 1967; Nieh and Roubik, 1998). Additionally, there are differ-Accepted 20 April 2010 Available online xxxx Keywords: Relaxed molecular clock Penalized likelihood Eusociality Stingless bees Honey bees Melipona Apis Referential communication1055-7903/$ - see front matter  2010 Elsevier Inc. A doi:10.1016/j.ympev.2010.04.026 Please cite this article in press as: Ram?rez, S.R., Evol. (2010), doi:10.1016/j.ympev.2010.04.026ject of numerous behavioral, ecological, and genetic studies, the evolutionary history of this genus remains largely unexplored. Here, we implement a multigene phylogenetic approach based on nuclear, mitochondrial, and ribosomal loci, coupled with molecular clock methods, to elucidate the phylogenetic relationships and antiquity of subgenera and species of Melipona. Our phylogenetic analysis resolves the relationship among subgenera and tends to agree with morphology-based classi?cation hypotheses. Our molecular clock analysis indicates that the genus Melipona shared a most recent common ancestor at least 14?17 million years (My) ago. These results provide the groundwork for future comparative anal- yses aimed at understanding the evolution of complex communication mechanisms in eusocial Apidae.  2010 Elsevier Inc. All rights reserved.Received 30 March 2009 Revised 2 April 2010 Stingless bees (Meliponini) constitute a diverse group of highly eusocial insects that occur throughout tropical regions around the world. The meliponine genus Melipona is restricted to the New World tropicsA molecular phylogeny of the stingless b Santiago R. Ram?rez a,*, James C. Nieh b, Tiago B. Quen Naomi E. Pierce a aMuseum of Comparative Zoology, Harvard University, 26 Oxford Street, Cambridge, M bDivision of Biological Sciences, Section of Ecology, Behavior, and Evolution, University c Smithsonian Tropical Research Institute, MRC 0580-12, Unit 9100 Box 0948, DPO AA 3 d Faculdade de Filoso?a, Ci?ncias e Letras de Ribeir?o Preto, Universidade de S?o Paulo, a r t i c l e i n f o a b s t r a c t Molecular Phyloge journal homepage: wwwll rights reserved. et al. A molecular phylogeny ofgenus Melipona (Hymenoptera: Apidae) l a, David W. Roubik c,d, Vera L. Imperatriz-Fonseca d, 138, USA lifornia, San Diego, 0116, La Jolla, CA 92093, USA -9998, USA ir?o Preto, 140400-901 S?o Paulo, Brazil le at ScienceDirect tics and Evolution sevier .com/ locate /ympevthe stingless bee genus Melipona (Hymenoptera: Apidae). Mol. Phylogenet. nuclear EF1-a (1.2 kb), ArgK (0.7 kb), and Pol-II (0.8 kb). DNA tions assumed unordered. We performed 100 random addition se- 2000), this fossil exhibits synapomorphic characters that unambig- uously place it within crown Meliponini. Thus, we used its age as a neti ARTICLE IN PRESSquences using the TBR algorithm, and estimated node support via non-parametric bootstrapping (100 replicates). A Maximum Likeli- hood (ML) phylogenetic analysis was performed in the software package GARLI (Zwickl, 2006) with model parameters estimated over the speci?ed number of runs. Bootstrap support values were estimated in GARLI with 100 heuristic tree searches using the same parameters as those implemented during tree searches. Addition- ally, Bayesian analyses were implemented in the software package MrBayes v3.1.1 (Ronquist and Huelsenbeck, 2003). Bayesian treewas extracted from individual bee specimens from either leg or thoracic muscle tissue using Qiagen DNA Extraction Kits (Qiagen Inc., Valencia, California). Polymerase Chain Reactions (PCRs) were carried on a Bio-Rad DNA Engine Dyad Peltier thermal cycler (Bio- Rad Laboratories Inc., Hercules, California) in 25 lL reactions with 2.5 mmol/L MgCl2, 2.5 mmol/L PCR buffer, and Taq polymerase (Qiagen Inc., Valencia, California) using various primer pairs (Dan- forth et al., 2004; Supplementary Table 1). We puri?ed PCR prod- ucts by incubating samples at 37 C for 35 min using Escherichia coli Exonuclease I enzyme (New England Biolabs, Hanover, Mary- land) and subsequently raising the temperature to 80 C for 20 min. Puri?ed products were cycle-sequenced using BigDye? Terminator v3.1 Ready Reaction Cycle Sequencing Kit (Applied Bio- systems Inc., Foster City, California). Samples were directly se- quenced on an Applied Biosystems Inc., 3100 Genetic Analyzer (Applied Biosystems Inc., Foster City, CA). Both forward and reverse strands were sequenced for each of the ?ve markers; complemen- tary strands were assembled using the software Sequencher? v4.2 (Gene Codes Corp., Ann Arbor, MI). All major lineages within Melipona and Apis were sampled for this study, including four subgenera, 35 species, and 51 individuals of approximately 50 described species of Melipona representing all main species groups, and three subgenera, six species, and 10 indi- viduals of Apis. Additionally, we sampled 30 taxa within the corbi- culate bees, including bumble bees, stingless bees, and orchid bees, and two outgroups (Epicharis and Centris). We include a total of 88 terminals. GenBank accession numbers are provided in Supple- mentary Table 3. 2.2. Phylogenetic analyses A single DNA matrix containing ?ve loci was assembled using MacClade v4.06 (Maddison and Maddison, 2003). Parsimony anal- yses were implemented in the software package Paup v4.0b (Swofford, 2003) with all characters weighted equally and transi-the Neotropical Meliponini (Rasmussen and Cameron, 2010), and its monophyly is well supported (Rego, 1990; Costa et al., 2003; Fernandes-Salom?o et al., 2005). A recent global phylogenetic anal- ysis of the entire tribe Meliponini supported a Miocene (24 My) origin for Melipona, but only 20 of the 50 described species were sampled and the internal relationships were not well resolved (Rasmussen and Cameron, 2010). Here, we present the ?rst com- prehensive species-level phylogenetic analysis of Melipona coupled with a molecular clock analysis. 2. Materials and methods 2.1. DNA sequencing and taxonomic sampling We sequenced 4.5 kb of DNA from ?ve different fragments including mitochondrial CO1 (1.2 kb), ribosomal 16S (0.6 kb), 2 S.R. Ram?rez et al. /Molecular Phylogesearches were made assuming both single (GTR+C+I) and multiple models of sequence evolution for each locus (see Supplementary Table 2). In addition, we ran a tree search where models of se- Please cite this article in press as: Ram?rez, S.R., et al. A molecular phylogeny of Evol. (2010), doi:10.1016/j.ympev.2010.04.026minimum age calibration for all Meliponini. The placement of E. moronei within extant (crown) Euglossa is justi?ed by the presence of elongated mouthparts, labrum shape, and pubescence (Engel, 1999b). The phylogenetic position of A. lithohermaea within extant Apis is justi?ed by the enlarged body size, elongated metabasitar- sus, wing venation, and infuscated wing membrane (Engel, 2006). The placement of P. dominicana within extant Neotropical Meliponini is justi?ed by the short trapezoidal clypeus, triangular shape of forewing medial cell, and shape of tibiae and basitarsi (Camargo et al., 2000). The concordance among calibration points was assessed with the cross-validation method (Near et al., 2005; Supplementary Fig. 3). Since our phylogenetic sampling included divergent extant lineages within Apis (Raf?udin and Crozier, 2007), Euglossa (Ram?rez et al., in press) and Neotropical Melipo-quence evolution were partitioned by codon positions, with parameters estimated separately for ?rst, second, and third codon positions of nuclear coding genes. Markov chain Monte Carlo (MCMC) searches were run for 10,000,000 generations, sampling every 1000 generations for a total of 10,000 trees; model parame- ters were estimated during the run. Three parallel runs were car- ried, and for each run one unheated and three incrementally heated chains were used. We checked for convergence within tree searches by plotting tree likelihood values against the number of generations, and among searches by comparing resulting topolo- gies. Bayesian posterior probabilities were estimated as the pro- portion of trees containing each node over the trees sampled during runs. The trees corresponding to the ?rst 1000 generations were discarded (??burn-in?). 2.3. Divergence time estimation Divergence times were estimated using a fully resolved topol- ogy obtained by applying a 50% Majority-Rule (MR) consensus to all the trees obtained from a Bayesian phylogenetic search; the remaining polytomies (six) were resolved randomly using the R software package APE v2.3. Using a Likelihood Ratio Test (LRT) we estimated this tree had a signi?cantly lower score value (nL 39222.08) when a molecular clock was enforced than when the assumption was relaxed (nL 38987.29). We calculated branch lengths on the 50% MR consensus tree via maximum likelihood in the software package Paup, optimized under the model of se- quence evolution GTR+C+I (molecular clock not enforced). Node divergence times were estimated with Penalized Likelihood (PL) using the Truncated-Newton algorithm in the software package r8s v1.71 (Sanderson, 1997). Mean ages ? SD were calculated using non-parametric bootstrapping. We used two sets of calibration ages, corresponding to the youngest and oldest estimates of the ages of the fossils used as node age constraints. A total of ?ve different ages were used to cal- ibrate our molecular clock trees (indicated by letters in Fig. 1): A, maximum root age (80?100 My, based on oldest stem bee fossil (Poinar and Danforth, 2006) and molecular clock analysis done by Hines (2008)); B, Cretotrigona prisca (65?70 My, Michener and Grimaldi, 1988; Engel, 2000) used as a minimum age calibration; C, Euglossa moronei (15?20 My, Engel, 1999b) used as a minimum age calibration; D Apis lithohermaea (14?16 My, Engel, 2006) used as a minimum age calibration; and E, Proplebeia dominicana (15? 20 My, Wille and Chandler, 1964; Camargo et al., 2000) used as a minimum age calibration. Although the age of C. prisca has been the subject of controversy (Michener and Grimaldi, 1988; Engel, cs and Evolution xxx (2010) xxx?xxxnini (Rasmussen and Cameron, 2010) we used fossil ages as mini- mum age constraints, even though in some cases lineage sampling was incomplete (e.g. Euglossa). the stingless bee genus Melipona (Hymenoptera: Apidae). Mol. Phylogenet. neti ARTICLE IN PRESSS.R. Ram?rez et al. /Molecular Phyloge3. Results and discussion 3.1. Phylogenetic relationships Our maximum parsimony, maximum likelihood, and Bayesian phylogenetic analyses, based on ?ve loci, resolved relationships 91 88 96 100 100 100 67 73 100 100 61 62 100 100 100 Parsimony bootstrap 100 100 76 94 100 100 94 100 99 92 * 100 100 Likelihood bootstrap * * B A Fig. 1. Strict consensus cladogram of 32 equally short maximum parsimony trees showin asterisks denote unavailable values. Five different ages were used to calibrate a molecula Hines (2008) and Poinar and Danforth (2006)); B, minimum age constraint, Cretotrigon constraint, Euglossa moronei (15?20 My, Engel, 1999b); D, minimum age constraint, Apis dominicana (15?20 My, Wille and Chandler, 1964; Camargo et al., 2000). Please cite this article in press as: Ram?rez, S.R., et al. A molecular phylogeny of Evol. (2010), doi:10.1016/j.ympev.2010.04.02675 76 M. panamica MP1 M. costaricensis MP14 Melipona melanopleura MP12 M. sp. (male) MP127 M. solani MP86 cs and Evolution xxx (2010) xxx?xxx 3within and between Melipona, Apis, and related clades of corbicu- late bees. We obtained well-resolved and supported phylogenetic trees (Figs. 1 and 2) that are congruent with each other under dif- ferent optimization schemes and model parameters (Figs. 1, 2 and Supplementary Figs. 1, 2). Parsimony analyses yielded 32 shortest trees (TL = 6344), with a strict consensus almost identical to a max- 58 100 81 76 87 97 86 95 100 100 100 59 57 93 87 100 100 99 97 52 100 50 98 100 99 73 97 53 95 100 100 100 100 100 67 95 80 99 93 97 92 51 51 70 69 * 100 100 94 95 91 100 93 100 71 92 100 100 100 100 67 100 76 96 99 100 100 100 68 76 100 100 * * * * * * * Nannotrigona perilampoides Euglossa asarophora M. quadrifasciata anth. MP23 Euglossa villosa M. crinita MP92 Trigona sp. M. aff. costaricaensis MP21 Geotrigona kraussi M. manda?aia MP81 M. quinquefasciata MP80 Bombus vagans A. dorsata Friesella schrotkyii A. cerana M. captiosa MP43 Eulaema peruviana M. favosa favosa MP19 M. compressipes MP37 M. sp (eburnea group) MP85 M. quadrifasciata anth. MP76 M. grandis MP128 M. grandis MP97 M. amazonica MP94 Cephalotrigona zexmeniae Apis cerana M. interrupta MP44 M. crinita MP32 A. cerana Euglossa decorata M. sp. (eburnea group) MP93 Nogueirapis mirandula Meliponula bocandei A. florea M. beecheii MP18 M. triplaridis MP8.1 A. koschevnikovi M. seminigra atrofulva MP38 Scaptotrigona barrocoloradensis Exaerete azteca Plebeia franki M. illustris MP46a Frieseomelitta silvestrii Bombus transversalis M. aff. costaricensis MP33 M. grandis MP22 Eufriesea caerulescens Cephalotrigona sp. Centris sp. M. sp (M de Dios) MP70 Scaptotrigona barrocoloradensis M. sp. (male) MP96 M. flavolineata MP49 M. sp. (queen) MP91 M. scutellaris MP78 M. asilvai MP82 Scaptotrigona polysticta M. lateralis MP45 Tetragonisca angustula M. rufiventris MP24 M. melanoventer MP34 M. bicolor MP83 Epicharis sp. A. andreniformis Aglae caerulea M. nebulosa MP48 M. marginata MP79 M. fuscopilosa MP35 M. n. sp. MP98 M. sp. (crinita group) MP90 M. ogliviei MP41 M. fuliginosa MP42 A. mellifera Euglossa mixta M. micheneri MP39 A. koschevnikovi Meliwillea bivea M. fulva MP15 M. fallax MP13 Lestrimelitta sp. M. illota MP95 M. rufiventris rufiventris MP77 M. interrupta MP20 M. flavolineata MP72 Apini Euglossini Bom bini outgroups M eliponini E D C g both maximum likelihood and maximum parsimony parametric bootstrap values; r clock (indicated by letters): A, maximum root age (80?100 My, estimated based on a prisca (65?70 My, Michener and Grimaldi, 1988; Engel, 2000); C, minimum age lithohermaea (14?16 My, Engel, 2006); and E, minimum age constraint, Proplebeia the stingless bee genus Melipona (Hymenoptera: Apidae). Mol. Phylogenet. neti ARTICLE IN PRESS4 S.R. Ram?rez et al. /Molecular Phylogeimum likelihood analysis (Fig. 1). Bayesian phylogenetic analyses also yielded well-supported trees that varied little whether using single (GTR+C+I), or partitioned (locus-speci?c and codon-speci?c) models of sequence evolution (Supplementary Figs. 1 and 2). All phylogenetic methods returned (i) Melipona as sister to the other Fig. 2. Relaxed-clock chronogram of Melipona and related groups; the tree correspon Divergence times were calculated (in millions of years, My) using the software package r8 single model of sequence evolution was ?tted to all loci (GTR+C+I) to estimate branch Please cite this article in press as: Ram?rez, S.R., et al. A molecular phylogeny of Evol. (2010), doi:10.1016/j.ympev.2010.04.026cs and Evolution xxx (2010) xxx?xxxNeotropical meliponine taxa included in our study (14 genera); (ii) Melipona + Neotropical Meliponini as sister to the African stingless bee genus Meliponula; and (iii) all stingless bees (Melipo- nini) + bumble bees (Bombini) as a monophyletic clade. These re- sults concord with previous molecular studies (Cameron and ds to the maximum a posteriori topology resulting from a Bayesian tree search. s v1.71. Node bars represent the 95% con?dence intervals of the age of each node. A lengths. Statistics for numbered nodes are indicated in Table 1. the stingless bee genus Melipona (Hymenoptera: Apidae). Mol. Phylogenet. quent molecular clock analyses, we used these two alternative tree topologies. The results from our PL analysis suggest that whereas extant Apis shared a recent common ancestor during the Oligocene, 29 ? 2 to 34 ? 2 My ago, Melipona shared a most recent common ancestor during the Miocene, 14 ? 1 to 17 ? 1 My ago, depending on whether we use the oldest or youngest ages of the fossil calibra- tions (Table 1). These time estimates varied little when using either of the two alternative models of sequence evolution (Table 1) or the two alternative topologies (data not shown). Melipona is one of the two largest (species rich) genera in the highly eusocial stingless bees (Plebeia is the other genus). Melipona is also exclusively Neotropical, and its origin is not known (Ras- mussen and Cameron, 2010). The few endemic species on islands in both the Lesser Antilles and Paci?c Panama (Camargo and Pedro, 2007) indicate relicts of past mainland connections during Mio- cene times (Roubik and Camargo, unpublished data). Our molecu- lar clock analysis coincides with this scenario. Our fossil-calibrated molecular clock provides an age estimate for the origin of Apis, Melipona, and the main clades of corbiculate Apidae. This provides a temporal framework with which to esti- mate the antiquity of referential communication. Although the genus Apis has an extensive fossil record, with the oldest fossil dat- ing to the Oligocene (25 My old) (Engel, 1998, 1999a, 2006; Engel et al., 2009), all honey bee fossils known to date (with the excep- tion of A. lithohermaea) are stem relatives of extant Apis (Engel, netics and Evolution xxx (2010) xxx?xxx 5 ARTICLE IN PRESSMardulyn, 2001; Rasmussen and Cameron, 2007, 2010; Kawakita et al., 2008). In a recent study by Rasmussen and Cameron (2010), Melipona was recovered as sister to most Neotropical gen- era, except Celetrigona, Dolichotrigona, Trigonisca, and Leurotrigona. Although our study did not include these taxa, our results are con- gruent with their hypothesis about the placement of Melipona. The placement of both honey bees (Apini) and orchid bees (Euglossini) were inconsistent in our study, depending on methodology (Sup- plementary Figs. 1 and 2). The parsimony and codon-partitioned Bayesian analyses supported the topology of (Euglossini (Apini, (Bombini, Meliponini))), whereas the Likelihood, Bayesian (opti- mized with both single and gene-partitioned models) supported the topology of ((Euglossini, Apini), (Bombini, Meliponini)). Although our analyses recovered the clade Apini + Euglossini as monophyletic only in some analyses, we obtained strong support for the clade Bombini + Meliponini, a grouping that has been con- troversial (Kawakita et al., 2008). Bombini is primarily boreal and temperate in distribution and Meliponini is restricted to tropical latitudes. Overall, our results differ from earlier morphology-based hypothesis about the relationships of corbiculate bees (Michener, 1944; Engel, 2001; Schultz et al., 2001) and, perhaps not surpris- ingly, support hypotheses based on molecular data (Cameron and Mardulyn, 2001; Kawakita et al., 2008). Our phylogenetic analyses indicate that three of the four sub- genera recognized by morphology within Melipona (Melikerria, Melipona s. str., and Michmelia) are monophyletic, but one (Eome- lipona) is polyphyletic (Fig. 2). All Melipona species currently unas- signed to a speci?c subgenus (Incertae sedis)?except M. fuliginosa (Camargo and Pedro, 2007, 2008)?form a monophyletic clade, sis- ter to the subgenus Michmelia (Fig. 2). In our MP analysis, M. amazonica was sister to the rest of species in the genus, but this placement was supported by a low bootstrap value (67). On the other hand, our Bayesian analysis supported the placement of M. amazonica as sister to M. marginata and M. bicolor, which was sup- ported by a high Bayesian posterior probability (98). M. amazonica constitutes a problematic taxon and additional gene fragments may be required to resolve its placement. Additionally, some of the internal branches within Melipona were not resolved or were not well supported. We note that multiple branches within Melip- ona are relatively short, particularly in the subgenus Michmelia and thus a rapid lineage diversi?cation may explain the observed low support values. The results from our phylogenetic analysis may guide future taxonomic studies, particularly on the delineation of subgenera and assignment of unplaced species. Our phylogenetic analyses agree with the proposed and widely accepted hypothesis of the internal relationships of honey bees: (Micrapis, (Megapis, Apis s. str.)) (Arias and Sheppard, 1996, 2005; Oldroyd and Wongsiri, 2006; Raf?udin and Crozier, 2007) and con- cur with molecular phylogenetic studies of the corbiculate bee tribes based on molecular data (Cameron and Mardulyn, 2001; Thompson and Oldroyd, 2004; Kawakita et al., 2008). 3.2. Molecular clock analysis We performed molecular clock analyses calibrated with ?ve dif- ferent fossil ages using Penalized Likelihood (PL). Because previous phylogenetic analyses (and our own results) have produced uncer- tainty in the placement of Apini and Euglossini, we used two alter- native tree topologies that resulted from our Bayesian analyses (Supplementary Fig. 1) that were produced by applying both single and partitioned models of sequence evolution. The main difference between these two alternative topologies was in the relative posi- tion of Apis and Euglossini, where Apis was recovered sister to S.R. Ram?rez et al. /Molecular PhylogeEuglossini (single model of sequence evolution), and Euglossini was recovered sister to the remaining corbiculate bees (gene and codon partitions). To account for this uncertainty in our subse- Please cite this article in press as: Ram?rez, S.R., et al. A molecular phylogeny of Evol. (2010), doi:10.1016/j.ympev.2010.04.0262006). By implementing molecular clocks, we show that extant Apis likely shared a most recent ancestor during the Eocene?Oligo- cene (29?33 My ago), whereas Melipona appears to have shared a common ancestor more recently, during the Miocene (14?17 My ago). Our age estimates for the most recent common ancestor of Melipona differ from those obtained by Rasmussen and Cameron (2010), which suggested that living members of genus shared an ancestor 25 My ago. In their analysis, Rasmussen and Cameron (2010) speci?ed a maximum age for the root node (Meliponini) of 125 My based on the oldest fossil angiosperms (most bees, ex- cept roughly 20 meliponine species (Lestrimelitta, Cleptotrigona, and the Trigona hypogea group) depend on ?owering plants for feeding. We used a different age estimate (80?100 My) for the Table 1 Mean age estimates (in millions of years, My) of major clades of corbiculate bees calculated via Penalized Likelihood (PL) optimized with a single model of sequence evolution (GTR+C+I) and a partitioned (locus-speci?c) model of sequence evolution. Tree branch lengths were calculated with maximum likelihood under the substitution model GTR+C+I using a 50% majority-rule consensus tree obtained in the Bayesian tree searches; SD were calculated via non-parametric bootstrapping. Divergence times for nodes that collapsed in the 50% majority-rule consensus are denoted by ??NA?. Node Younger calibrations Older calibrations GTR+C+I (l ? SD) Locus- speci?c (l ? SD) GTR+C+I (l ? SD) Locus- speci?c (l ? SD) Apis 29.92 ? 1.51 30.57 ? 1.57 33.18 ? 1.67 33.84 ? 1.72 Melipona 15.43 ? 0.89 14.56 ? 0.82 17.21 ? 0.98 16.21 ? 0.91 Euglossini 20.25 ? 1.27 20.89 ? 1.34 22.49 ? 1.33 23.16 ? 1.45 Neotropical Meliponini 32.92 ? 1.82 33.29 ? 1.79 36.73 ? 1.91 37.05 ? 1.91 1 6.34 ? 0.48 6.32 ? 0.44 7.07 ? 0.52 7.03 ? 0.48 2 8.49 ? 0.65 8.73 ? 0.65 9.47 ? 0.71 9.72 ? 0.72 3 11.66 ? 0.73 12.23 ? 0.78 13.00 ? 0.80 13.62 ? 0.86 4 8.29 ? 0.82 8.22 ? 0.88 9.25 ? 0.90 9.15 ? 0.98 5 NA 10.60 ? 0.83 NA 11.80 ? 0.92 6 32.04 ? 2.51 31.69 ? 2.44 35.37 ? 2.77 34.91 ? 2.69 7 17.27 ? 1.42 17.51 ? 1.49 19.16 ? 1.58 19.40 ? 1.65 8 22.32 ? 1.48 22.68 ? 1.54 24.76 ? 1.65 25.12 ? 1.72 9 18.56 ? 1.80 18.89 ? 1.79 20.59 ? 1.99 20.93 ? 1.98 10 77.62 ? 1.14 75.89 ? 0.98 85.24 ? 1.37 83.14 ? 1.17 11 83.87 ? 1.36 81.76 ? 1.23 92.76 ? 1.62 90.20 ? 1.47 the stingless bee genus Melipona (Hymenoptera: Apidae). Mol. Phylogenet. in Apis. This study should guide future comparative analyses of ref- neti ARTICLE IN PRESSerential communication and of other aspects of life history evolu- tion in Melipona and other eusocial bees. Acknowledgments We thank Paulo Nogueira-Neto and Jay Evans for providing cru- cial samples used in this study, Jennifer Davis and Jessica Girard for assistance in the laboratory, and Jo?o M.F. Camargo and Claus Ras- mussen for their assistance in the identi?cation of bee specimens. Beth Pringle provided useful comments on the manuscript. This project was supported by grants from ORBS (Opportunities for Re- search in the Behavioral Sciences Program), the National Science Foundation (NSF-DDIG, DEB 0608409, NSF-IBN 0316697, NSF DEB 0447242, and NSF-IBN 0545856) and grants from the Putnam Expedition Fund to J.C.N., S.R.R., and N.E.P. Samples were collected in Brazil under the Permit No. 057/2003 (IBAMA) and Peru under Permit No. 005141 (AG-INRENA). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ympev.2010.04.026. References Aguilar, I., Brice?o, D., 2002. Sounds in M. costaricensis (Apidae: Meliponini): effect of sugar concentration and nectar source distance. Apidologie 33, 375?388. Arias, M.C., Sheppard, W.S., 1996. Molecular phylogenetics of honey bee species (Apis mellifera L.) inferred from mitochondrial DNA sequence. Mol. Phylogenet. Evol. 5, 557?566. Arias, M.C., Sheppard, W.S., 2005. Phylogenetic relationships of honey bees (Hymenoptera: Apinae: Apini) inferred from nuclear and mitochondrial DNA sequence data. Mol. Phylogenet. Evol. 37, 25?35. Camargo, J.M.F., Pedro, S.R.M., 2007. Meliponini Lepeletier, 1836. In: Moure, J.S., Urban, D., Melo, G.A.R. (Eds.), Catalogue of Bees (Hymenoptera: Apoidea) in the Neotropical Region. (accessed 30.10.08). Camargo, J.M.F., Pedro, S.R.M., 2008. Revis?o das esp?cies de Melipona do grupo fuliginosa (Hymenoptera: Apoidea: Apidae: Meliponini). Rev. Bras. Entomol. 52, 411?427. Camargo, J.M.F., Grimaldi, D., Pedro, S.R.M., 2000. The extinct fauna of stingless bees (Hymenoptera: Apidae: Meliponini) in Dominican amber: two new species and redescription of the male of Proplebeia dominicana (Wille and Chandler). Am. Mus. Novit. 3293, 1?24.divergence between corbiculate bees and the outgroup (Centridi- ni). We suspect that the difference between both studies stems from applying different ages to the root node. The discovery of new fossils may shed new light on the time of origin of apid lineages. Because no morphological characters have been associated with the use of recruitment communication in bees, we cannot infer whether stem fossil Apis, or any other extinct stingless bee lineages (Engel, 2001), exhibited this form of communication behavior. However, since all extant members of Apis use recruitment com- munication, it is likely that the most recent common ancestor of extant Apis had a form of recruitment communication similar to that exhibited by modern species. If true, this would suggest that recruitment communication in honey bees has been stable since the Eocene?Oligocene. On the other hand, our results suggest that the genus Melipona shared a most recent common ancestor more recently, during the Miocene. Although detailed behavioral obser- vations are available for less than 10 of the 50 Melipona species in our phylogeny (Nieh, 2004), available data suggest that communi- cation abilities in Melipona are more variable (Nieh, 2004). Thus, our study suggests that the traits associated with communication behavior are perhaps younger and more ?exible in Melipona than 6 S.R. Ram?rez et al. /Molecular PhylogeCameron, S.A., Mardulyn, P., 2001. Multiple molecular data sets suggest independent origins of highly eusocial behavior in bees (Hymenoptera: Apinae). Syst. Biol. 50, 194?214. Please cite this article in press as: Ram?rez, S.R., et al. A molecular phylogeny of Evol. (2010), doi:10.1016/j.ympev.2010.04.026Costa, M.A., Del Lama, M.A., Melo, G.A.R., Sheppard, W.S., 2003. Molecular phylogeny of the stingless bees (Apidae: Apinae: Meliponini) inferred from mitochondrial 16S rDNA sequences. Apidologie 34, 73?84. Danforth, B.N., Fang, J., Sipes, S., Brady, S.G., Almeida, E., 2004. Phylogeny and Molecular Systematics of Bees (Hymenoptera: Apoidea). Cornell University, Ithaca, NY. Available from: . Dyer, F.C., 2002. The biology of the dance language. Annu. Rev. Entomol. 47, 917? 949. Engel, M.S., 1998. Fossil honey bees and evolution in the genus Apis (Hymenoptera: Apidae). J. Hymenopt. Res. 29, 265?281. Engel, M.S., 1999a. The taxonomy of recent and fossil honey bees (Hymenoptera: Apidae: Apis). J. Hymenopt. Res. 8, 165?196. Engel, M.S., 1999b. The ?rst fossil Euglossa and phylogeny of the orchid Bees (Hymenoptera: Apidae: Euglossini). Am. Mus. Novit. 3272, 1?14. Engel, M.S., 2000. A new interpretation of the oldest fossil bee (Hymenoptera: Apidae). Am. Mus. Novit. 3296, 1?11. Engel, M.S., 2001. Monophyly and extensive extinction of advanced eusocial bees: insights from an unexpected Eocene diversity. Proc. Natl. Acad. Sci. USA 98, 1661?1664. Engel, M.S., 2006. A giant honey bee from the middle Miocene of Japan (Hymenoptera: Apidae). Am. Mus. Novit. 3504, 1?12. Engel, M.S., Hinojosa-D?az, I.A., Rasnitsyn, A.P., 2009. A honey bee from the Miocene of Nevada and the biogeography of Apis (Hymenoptera: Apidae: Apini). Proc. Cal. Acad. Sci. 6, 23?38. Esch, H., 1967. Die Bedeutung der Lauterzeugung f?r die Verst?ndigung der stachellosen Bienen. Z. Vergl. Physiol. 56, 408?411. Esch, H.E., Zhang, S., Srinivasan, M.V., Tautz, J., 2001. Honeybee dances communicate distances measured by optic ?ow. Nature 411, 58.1?58.3. Fernandes-Salom?o, T.M., Rocha, R.B., Campos, L.A.O., Ara?jo, E., 2005. The ?rst internal transcribed spacer (ITS-1) of Melipona species (Hymenoptera: Apidae: Meliponini): characterization and phylogenetic analysis. Insectes Soc. 52, 11?18. Gould, J.L., 1976. The dance-language controversy. Q. Rev. Biol. 51, 211?244. Hines, H.M., 2008. Historical biogeography, divergence times, and diversi?cation patterns of bumble bees (Hymenoptera: Apidae: Bombus). Syst. Biol. 57, 58?75. Hrncir, M., Barth, F.G., Tautz, J., 2006. Vibratory and airborne-sound communication in bees (Hymenoptera). In: Drosopoulos, S., Claridge, M.F. (Eds.), Insect Sound and Communication: Physiology, Behaviour, Ecology and Evolution. Taylor & Francis Group, Boca Raton, FL, pp. 421?436. Kawakita, A., Ascher, J.S., Sota, T., Kato, M., Roubik, D.W., 2008. Phylogenetic analysis of the corbiculate bee tribes based on 12 nuclear protein-coding genes (Hymenoptera: Apoidea: Apidae). Apidologie 39, 163?175. Maddison, D.R., Maddison, W.P., 2003. MacClade v4.06 (Computer Program). Sinauer Associates, Sunderland, MA. Michelsen, A., Andersen, B.B., Storm, J., Kirchner, W.H., Lindauer, M., 1992. How honeybees perceive communication dances, studied by means of a mechanical model. Behav. Ecol. Sociobiol. 30, 143?150. Michener, C.D., 1944. Comparative external morphology, phylogeny, and a classi?cation of the bees. Bull. Am. Mus. Nat. Hist. 82, 151?326. Michener, C.D., 1974. The Social Behavior of the Bees: A Comparative Study. Harvard University Press, Cambridge, MA. Michener, C.D., 2007. The Bees of the World. Johns Hopkins University Press, Baltimore, MD. Michener, C.D., Grimaldi, D., 1988. The oldest fossil bee: apoid history, evolutionary stasis, and the antiquity of social behavior. Proc. Natl. Acad. Sci. USA 85, 6424? 6426. Moure, J.S., Kerr, W.E., 1950. Sugest?es para a modi?ca??o da sistem?tica do g?nero Melipona (Hymen.-Apoidea). Dusenia 1, 105?129. Near, T.J., Meylan, P.A., Shaffer, H.B., 2005. Assessing concordance of fossil calibration points in molecular clock studies: an example using turtles. Am. Nat. 165, 137?143. Nieh, J.C., 1998. The food recruitment dance of the stingless bee, Melipona panamica. Behav. Ecol. Sociobiol. 43, 133?145. Nieh, J.C., 2004. Recruitment communication in stingless bees (Hymenoptera: Apidae: Meliponini). Apidologie 35, 159?182. Nieh, J.C., Roubik, D.W., 1998. Potential mechanisms for the communication of height and distance by a stingless bee, Melipona panamica. Behav. Ecol. Sociobiol. 43, 387?399. Nieh, J.C., Contrera, F.A.L., Ram?rez, S., Imperatriz-Fonseca, V.L., 2003. Variation in the ability to communicate three-dimensional resource location by stingless bees from different habitats. Anim. Behav. 66, 1129?1139. Oldroyd, B.P., Wongsiri, S., 2006. Asian Honey Bees: Biology, Conservation and Human Interactions. Harvard University Press, Cambridge, MA. Poinar Jr., G.O., Danforth, B.N., 2006. A fossil bee from Early Cretaceous Burmese amber. Science 314, 614. Raf?udin, R., Crozier, R.H., 2007. Phylogenetic analysis of honey bee behavioral evolution. Mol. Phylogenet. Evol. 43, 543?552. Ram?rez, S.R., Roubik, D.W., Skov, C., Pierce, N.E., 2010. Phylogeny, biogeography and diversi?cation of the orchid bees (Hymenoptera: Euglossini). Biol. J. Linn. Soc., doi:10.1111/j.1095-8312.2010.01440.x. Rasmussen, C., Cameron, S.A., 2007. A molecular phylogeny of the Old World stingless bees (Hymenoptera: Apidae: Meliponini) and the non-monophyly of the large genus. Trigona. Syst. Entomol. 32, 26?39. cs and Evolution xxx (2010) xxx?xxxRasmussen, C., Cameron, S.A., 2010. Global stingless bee phylogeny supports ancient divergence, vicariance, and long distance dispersal. Biol. J. Linn. Soc. 99, 206?232. the stingless bee genus Melipona (Hymenoptera: Apidae). Mol. Phylogenet. Rego, M.M.C., 1990. Revis?o do genero Melipona Illiger, 1806: Genitalia e esternos pre-genitais de machos (Meliponinae: Apidae: Hymenoptera). Ribeir?o Preto, Brazil. M.Sc. Thesis, Universidade de S?o Paulo. Ronquist, F., Huelsenbeck, J.P., 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572?1574. Roubik, D.W., 2006. Stingless bee nesting biology. Apidologie 37, 124?143. Sanderson, M.J., 1997. A nonparametric approach to estimating divergence times in the absence of rate constancy. Mol. Biol. Evol. 14, 1218?1231. Schultz, T.R., Engel, M.S., Ascher, J.S., 2001. Evidence for the origin of eusociality in the corbiculate bees (Hymenoptera: Apidae). J. Kans. Entomol. Soc. 74, 10?16. Schwarz, H.F., 1932. The genus Melipona: the type genus of the Meliponidae or stingless bees. Bull. Am. Mus. Nat. Hist. 63, 231?460. Seeley, T.D., 1995. The Wisdom of the Hive: The Social Physiology of Honey Bee Colonies. Harvard University Press, Cambridge, MA. Sherman, G., Visscher, P.K., 2002. Honey bee colonies achieve ?tness through dancing. Nature 419, 920?922. Swofford, D.L., 2003. Paup v4.0b (Computer Program). Sinauer Associates, Sunderland, MA. Thompson, G.J., Oldroyd, B.P., 2004. Evaluating alternative hypotheses for the origin of eusociality in corbiculate bees. Mol. Phylogenet. Evol. 33, 452?456. von Frisch, K., 1967. The Dance Language and Orientation of Bees. The Belknap Press of Harvard University Press, Cambridge, MA. Wille, A., Chandler, L.C., 1964. A new stingless bee from the Tertiary of the Dominican Republic (Hymenoptera: Meliponini). Rev. Biol. Trop. 12, 187?195. Wilson, E.O., 1971. The Insect Societies. Harvard University Press, Cambridge, MA. Zwickl, D.J., 2006. GARLI, Genetic Algorithm for Rapid Likelihood Inference 0.94 (Software Program). Available from: . S.R. Ram?rez et al. /Molecular Phylogenetics and Evolution xxx (2010) xxx?xxx 7 ARTICLE IN PRESSPlease cite this article in press as: Ram?rez, S.R., et al. A molecular phylogeny of Evol. (2010), doi:10.1016/j.ympev.2010.04.026the stingless bee genus Melipona (Hymenoptera: Apidae). Mol. Phylogenet.