Stingless bees (Hymenoptera, Meliponini) are a large and diverse group including 59 extant groups, representing the main pollinators of Amazon forests. Among those, Trigona is one of the largest endemic genera of Neotropical Meliponini. In this work, we updated the molecular phylogeny of Trigona proposed by Rasmussen and Camargo (2008), including data from 59 new specimens collected in 2020 in the forests of northern Peru, through a multigene phylogenetic approach combining sequences from four gene fragments (16S, ArgK, EF-1a, opsin). Our results confirmed the monophyly of Trigona and of all proposed subgenera, except Aphaneura. In addition, most Trigona species-groups resulted monophyletic but the ‘spinipes’ and ‘pallens’ groups appeared paraphyletic and polyphyletic, respectively. Moreover, the cohesion of the “fulviventris” species group was hindered by the inclusion of T. williana (previously included in the “pallens” group) within this clade. Finally, we provided further evidence for a subdivision into two (geographically) distinct clades within T. guianae in northern Peruvian Amazon, which highlighted the importance of Neotropical biogeographical barriers in Meliponini divergence and evolution. Finally, to avoid misidentifications of Trigona specimens, the need for a robust taxonomic revision based on a cladistic approach of the whole genus is discussed.

Apoidea, Neotropical biogeography, Peruvian Amazon, Taxonomy

Stingless bees (Hymenoptera, Meliponini) are major pollinators in tropical forests (Roubik 1989) with about 623 species belonging to 59 extant and fossil groups (considered as genera, subgenera or synonymized depending on the classification) (Engel et al. 2023). Meliponini are distributed throughout the tropical and subtropical areas of the Afrotropical, Australasian, Indo-Malayan and Neotropical Regions, exhibiting the highest diversity in the New World Amazonian rainforests (Michener 1978; Roubik 1993; Michener 2007). Currently, 26 extant genera are considered endemic to the New World (Engel et al. 2023). Among these, Trigona Jurine, 1807 is exclusive to the Neotropics and is one of the largest genera of stingless bees (Michener 2007; Rasmussen and Cameron 2007). Recent molecular phylogenetic data confirmed the monophyly of the New World species of Trigona, a genus with 32 currently considered valid species (Camargo et al. 2013; Costa et al. 2003; Rasmussen and Cameron 2007; Rasmussen and Camargo 2008). Nine species-groups have been recognized based on morphological, ecological and distributional data, and largely supported by genetic analyses (Rasmussen and Camargo 2008). More recently, seven of these species-groups were elevated to subgenera of Trigona (Engel 2021) [i.e., ‘cilipes’ as Aphaneuropsis Engel, 2021; ‘fulviventris’ as Koilotrigona Engel, 2021; ‘crassipes’ as Necrotrigona Engel, 2021; ‘pallens’ as Aphaneura Gray, 1832; ‘dimidiata’ as Dichrotrigona Engel, 2021; ‘fuscipennis’ as Ktinotrofia Engel, 2021; ‘recursa’ as Nostotrigona Engel, 2021], with the remaining two groups, ‘amalthea’ and ‘spinipes’ forming the subgenus Trigona s. s.r. Jurine, 1807.

This group of bees, characterized by small to large workers (5.5–11 mm), shows a variety of defense behaviors and nesting habits (i.e. nests are built on branches of plants or walls, in anthills or underground; Costa et al. 2004), as well as different foraging ecologies, from pollen and nectar gatherers (Fig. 1) to obligated necrophages (i.e., Trigona crassipes Fabricius, 1793, T. hypogea Silvestri, 1902 and T. necrophaga Camargo & Roubik, 1991; Roubik 1982; Camargo and Roubik 1991).

Figure 1. 

Trigona cf. chanchamayoensis Schwarz, 1948 sucking nectar from a flower (Photo M. Marconi).

About 22 species of Trigona have been reported in Peru (Rasmussen and Gonzalez 2009; Camargo et al. 2013; Sánchez Sandoval et al. 2015; Castillo-Carrillo et al. 2016; Rasmussen and Delgado 2020), but the overall number is likely underestimated because many forested areas of the country remain unexplored.

Recently, several Trigona specimens dwelling in humid and seasonally dry forests of northern Peru (in San Martin and Piura regions) were identified through an integrative taxonomy approach, i.e., considering both morphology and COI barcoding (Marconi et al. 2022). As expected, the COI-based reconstructed phylogeny was mostly unresolved at deep nodes. In addition, the newly collected Peruvian specimens ascribed to T. fulviventris Guerin-Meneville, 1845 and T. guianae Cockerell, 1912 were split into four distinct clades, two for each species (named provisionally as ‘A’ and ‘B’ clades in both cases). The same phylogenetic analysis also detected two lineages that were unrelated to other identified species, which were provisionally attributed to T. sp. 1 and T. sp. 2, respectively (Marconi et al. 2022).

In this work we conducted a multigene phylogenetic analysis of the Neotropical genus Trigona by integrating novel molecular data of four genes obtained from northern Peruvian specimens (Marconi et al. 2022) with a previously published dataset (Rasmussen and Camargo 2008). By updating the current phylogeny, we aimed to clarify the taxonomic issues emerged in our previous work (Marconi et al. 2022) and further validate the currently recognized species-groups within Trigona (Rasmussen and Camargo 2008) and the recently proposed subgenera (Engel 2021).

We obtained 58 sequences of 16S (Genbank Acc. n° OR353456–OR353513), 26 of ArgK, 41 of EF-1a and 26 of opsin (Genbank Acc. n° OR393480–OR393571) from a total of 59 northern Peruvian Trigona specimens collected in 2020 (Marconi et al. 2022). The combined dataset, including previously generated sequences of Trigona and outgroup species (Rasmussen and Camargo 2008), consisted of a total of 88 individuals (including 5 outgroups) with 2329 aligned positions composed by the four gene fragments: 485 base pairs (bp) of 16S, 592 bp of ArgK, 729 bp of EF-1a and 522 bp of opsin gene. ML and BY tree topologies largely overlapped (hence, BY topology only is shown; Fig. 2). The combined ML and BY analysis confirmed the monophyly of the genus Trigona (Fig. 2: PP = 1.00/UFB = 100) and the presence of two main distinct clades, one (PP = 1.00/UFB = 98) including members of the ‘amalthea’ + ‘spinipes’ (= Trigona s. s.r.), ‘fuscipennis’, (= Ktinotrofia) ‘recursa’ (= Nostotrigona) and ‘crassipes’ (= Necrotrigona) species groups (or subgenera) (PP = 1.00/UFB = 100), the other including members of the ‘cilipes’ (= Aphaneuropsis), ‘pallens’ (= Aphaneura) and ‘fulviventris’ (= Koilotrigona) species groups (or subgenera) (Rasmussen and Camargo 2008; Engel 2021) (Fig. 2). Five out of the 8 Trigona species groups were recovered as monophyletic: ‘amalthea’ (PP = 1.00/UFB = 99), ‘fuscipennis’ (PP = 1.00/UFB = 100), ‘recursa’ (also including T. sp. 1; PP = 1.00/UFB = 98), ‘crassipes’ (PP = 1.00/UFB = 98), ‘cilipes’ (PP = 1.00/UFB = 100). The ‘spinipes’ group appeared paraphyletic since it was split into two distinct, though poorly supported clades (Fig. 2). One included T. spinipes, T. hyalinata, T. corvina and T. amazonensis (‘spinipes’ 1; PP = 0.80/UFB = 94), whereas the other grouped T. nigerrima and T. dallatorreana (= T. sp. 2) (‘spinipes’ 2; PP = 0.93/UFB = 71). However, all ‘spinipes’ members clustered with those of ‘amalthea’, thus supporting the monophyly (PP = 1.00/UFB = 100) of the subgenus T. (Trigona s. s.r.) sensu Engel 2021. T. williana did not cluster within the ‘pallens’ group, but was genetically closer to members of the ‘fulviventris’ group (= Koilotrigona). However, its placement within the ‘fulviventris’ group or Koilotrigona subgenus remains doubtful since it received a low Bayesian support (PP = 0.52; Fig. 2). Hence, the ‘pallens’ group and the subgenus Aphaneura Gray 1832 (Engel 2021) are tenable only if T. williana is excluded and placed in a different group/subgenus, still to be defined. Finally, as already reported (Marconi et al. 2022), Peruvian specimens of T. guianae were subdivided into two well-supported and distinct clades (A and B) (Fig. 2), whereas those ascribed to T. fulviventris were included in the same clade (A+B; see Marconi et al. 2022).

Figure 2. 

Trigona Bayesian phylogenetic tree topology estimated from combined sequence data from four gene fragments (16S, ArgK, EF-1a, opsin). Posterior probability and ultra-fast bootstrap values (BY - PP/ML - UFB) are shown at deepest nodes only. Color marks are assigned to tips leading to the 59 northern Peruvian specimens belonging to Trigona species, whose taxonomic identification and geographic origin are reported in detail in table 1 of Marconi et al. 2022.

We here built upon the molecular phylogeny of the Neotropical genus Trigona (Rasmussen and Camargo 2008) by adding genetic data from newly collected specimens in northern Peruvian forests (Marconi et al. 2022).

We confirmed the monophyly of the Neotropical genus Trigona and of all proposed subgenera, except for Aphaneura Gray, 1832 (Engel 2021). In addition, most Trigona species-groups were found to be monophyletic (Fig. 2). However, as already observed (Rasmussen and Camargo 2008), the ‘spinipes’ and ‘pallens’ species groups were paraphyletic and polyphyletic, respectively (Fig. 2). Our results support combining members of the ‘amalthea’ and ‘spinipes’ groups into the proposed subgenus T. (Trigona s. s.r.). However, the closely related T. (Trigona) dallatorreana (= T. sp. 2; Marconi et al. 2022) and T. (Trigona) nigerrima should be ascribed to a different species-group (provisionally named ‘spinipes’ 2 in Fig. 2). As previously mentioned, the ‘pallens’ group (= Aphaneura) as usually recognized is polyphyletic due to the large genetic distance of T. williana from all other members of this group/subgenus. In fact, T. williana is similar only in coloration to members of the ‘pallens’ group and differs in the shape of metasoma and metatibiae (F.F. De Oliveira, pers. comm.). The placement of T. williana within the ‘fulviventris’ group is also doubtful as it differs in many morphological and biological features from other members of the group. Its true placement will require further investigation. In general, since some taxonomic issues affect the ‘pallens’ group (e.g., the types of both T. muzoensis Schwarz, 1948 and T. ferricauda Cockerell, 1917 should be re-examined to exclude possible synonymies), we cannot rule out that our specimens, formerly recognized through the integrative taxonomic approach as T. muzoensis (Marconi et al. 2022), could be instead ascribed to T. chanchamayoensis Schwarz, 1948 - occurring in Peru east of Andes (type locality: San Ramon, Valle de Chanchamayo, Peru) - or to T. pallens Fabricius, 1798. In fact, the specimens morphologically identified by Rasmussen and Camargo (2008) as T. chanchamayoensis and T. pallens (i.e., chanc 016 and pall 061; Fig. 2) are placed in two distinct clades including two separate groups of individuals previously identified as T. muzoensis (Marconi et al. 2022), respectively (Fig. 2). When these northern Peruvian specimens were identified in BoldSystems (www.boldsystems.org) (Marconi et al. 2022), they received ID scores ranging 96.43 (POA) - 98.46% (e.g., JP007) for T. muzoensis, but did not match with the single T. chanchamayoensis available in BoldSystems (from Brazil), nor with T. pallens, totally lacking COI data. Unfortunately, taxonomic keys to promptly distinguish morphologically all members of the ‘pallens’ group are also lacking. Similarly, doubts could be raised to our previous attribution of northern Peruvian specimens to T. cf. hypogea or T. cf. fuscipennis (Marconi et al. 2022), because these show a close (although scarcely supported) phylogenetic relatedness to two species identified by Rasmussen and Camargo (2008), i.e., T. crassipes (crassi 060) and T. albipennis (albi 168), respectively (Fig. 2). However, for these two species as well, data are lacking in BoldSystems, nor valuable keys of distinctive morphological characters are available for ‘crassipes’ and ‘fuscipennis’ groups. In general, the absence of published dichotomous keys based on reliable diagnostic morphological characters and cladistic approaches integrating extensive genetic (COI or other marker) datasets aimed to define species boundaries, still hinder the correct identification of stingless bee species (see also, Marconi et al. 2022). These data deficiencies are likely to generate conflicts in Trigona identification (as, in this case, with those of Rasmussen and Camargo 2008) and favor the description of new species without truthfully considering their morphological and genetic internal cohesion, as well as their distinction from other (sibling) taxa.

We also confirmed a genetic subdivision within T. guianae into two putatively distinct taxonomic and/or geographic units, possibly originated by limited gene flow due to biogeographic barriers in the Neotropics (Marconi et al. 2022). Indeed, comparative analysis of metatarsi of T. guianae (Clade A) and T. guianae (Clade B) revealed morphological differences at the retrodorsal margin and distal angle (unpublished data). Further data will allow establishing if T. guianae (Clade B) could be ascribed to a novel species endemic to Pabloyacu, or to one of the approximately 28 novel species awaiting description (Rasmussen and Camargo 2008). On the other hand, the combined molecular dataset did not support the split into two distinct entities in T. fulviventris, as previously suggested based on COI marker only (Marconi et al. 2022). However, a recent morphological analysis showed that T. cf. fulviventris (Clade A) has a narrow subtriangular metatibia, whereas T. cf. fulviventris (Clade B) (MA6) has a broad, “drop-like” shape (unpublished data). Additional specimens will be examined, both genetically and morphologically, to clarify such issues.

Concerning the two previously unidentified Trigona species (Marconi et al. 2022), as reported above we confirm that T. sp. 2 is T. dallatorreana, whereas T. sp. 1 seems to be related to T. recursa, although its taxonomic relationships need further examination.

Novel genetic/genomic data from populations sampled across the entire geographic ranges of all of the Trigona species groups will shed light on the phylogenetic relationships among members of this large genus of Neotropical stingless bees. Further morphological work is also needed to produce and/or refine taxonomic keys and accurately revise the taxonomy of this speciose genus. Such an effort would not only resolve some taxonomic issues within this large genus of stingless bees, but also enhance our understanding of the role of Neotropical biogeographic barriers in the evolution of this main group of pollinators of the Amazon forests.

The authors thank the community of Mangamanguilla, Paolo Villegas Ogoña, Miguel Tapullima and Hitler Panduro Salas for facilitating sampling activities in Northern Peruvian forests; Noemi Centrone and Alessandro Modesti for helping the laboratory work; Leydi Paz Alvarez to assist the field activities; the NGO Estudios Amazónicos to support all research on Meliponini in Peru. This work was supported by the Cooperation Project of Sapienza 2022 (Prot. n° 0001626) “Verso una Meliponicoltura Consapevole e Tutela delle Api senza Pungiglione nella Foresta Amazzonica del Perù”; by the Department of Excellence of the University of Roma Tre; Proyectos de Investigación Básica 2023-01 (PROCIENCIA n° 82163) “Modelo transdisciplinar para la comprensión de la diversidad clave de las abejas peruanas sin aguijón (Hymenoptera: Apidae: Meliponini) con fines de conservación y el desarrollo de una meliponicultura competitiva en la Amazonia”; the Project “Monitoraggio e protezione degli impollinatori del PN Circeo” (University of Roma Tre and PN Circeo).