From Southeast Asia to global: phylogeny, biogeography and character evolution of the thread‐legged bug tribe Leistarchini (Hemiptera: Reduviidae: Emesinae)

Abstract

The thread‐legged bug tribe Leistarchini (Hemiptera: Reduviidae: Emesinae) is a cosmopolitan and diverse group characterized by a highly disproportionate spatial distribution across zoogeographic regions. Due to a historical lack of phylogenetic focus, the internal relationships and evolutionary history of the tribe remain poorly understood. In this study, we provide the first robust phylogenetic framework for Leistarchini by integrating molecular data from mitochondrial genomes and nuclear rDNA (88 taxa, 19 937 bp) with universal single‐copy orthologs (24 taxa, 667 loci). Our results support the monophyly of Leistarchini and identify five major clades, including a newly described genus Calliemesa gen. n. Our findings further reveal that the five most species‐rich genera (Nesita, Orthunga, Pleias, Ploiaria and Tinna) are either paraphyletic or polyphyletic as currently circumscribed. Molecular dating and biogeographic reconstructions suggest a Southeast Asian origin for the Leistarchini crown group during the late Palaeocene (ca. 57 Ma). Early diversification appears to have been driven by Paleogene geological and climatic shifts in Southeast Asia, while multiple intercontinental dispersals since the middle Eocene into the Afrotropics, Madagascar and the New World shaped the current global distribution. Ancestral state reconstructions indicate that the Leistarchini ancestor possessed a well‐developed posterior pronotal lobe and a three‐segmented protarsus. Subsequent evolutionary trajectories involved four independent transitions toward a shortened posterior pronotal lobe, and four separate reductions in protarsal segmentation.

Introduction

Assassin bugs (Hemiptera: Heteroptera: Reduviidae) represent one of the most successful lineages of predatory insects, characterized by a remarkable diversity of morphological, behavioural and ecological adaptations (Schuh and Weirauch, 2020). While most Reduviidae are general predators used as biological control agents against agricultural and forest pests (Grundy, 2007), some lineages, such as the hematophagous Triatominae, have transitioned to vertebrate blood‐feeding and serve as vectors for Chagas disease (Jurberg and Galvão, 2006; Galvão and Justi, 2015). Within the predatory groups, many lineages exhibit high degrees of prey specialization (Forthman and Weirauch, 2012; Bulbert et al., 2014; Gordon and Weirauch, 2016). A prominent example is the thread‐legged bug subfamily Emesinae, which is distinguished by its slender, delicate body, raptorial forelegs and frequent association with spiderwebs (Wygodzinsky, 1966; Standring et al., 2023). Several emesine species are obligate predators of web‐building spiders, utilizing specialized morphological and behavioural innovations to hunt in these complex environments (Wignall and Taylor, 2010, 2011; Soley and Taylor, 2012; Soley, 2016). Currently, Emesinae comprises seven tribes, over 130 genera and 1200 species, making it the second most species‐rich subfamily in Reduviidae (Maldonado‐Capriles, 1990; Standring et al., 2023; Castro‐Huertas and Melo, 2025).

The thread‐legged bug tribe Leistarchini is a diverse lineage of approximately 300 species described among 33 extant genera (Wygodzinsky, 1966; Maldonado‐Capriles, 1990, 1992). Members of the tribe have colonized nearly all non‐polar landmasses, occupying habitats from temperate grasslands to tropical rainforests and caves, with some species being observed dwelling and foraging on and around spiderwebs (Wygodzinsky, 1966; Ishikawa and Miyamoto, 2012; Chen et al., 2025). Nearly half of the tribe's described species belong to the cosmopolitan genus Ploiaria Scopoli, 1786, the most speciose genus in Emesinae (Ishikawa et al., 2008; Chen et al., 2023). In contrast, other genera contain fewer than 30 species and often exhibit restricted distributions (Wygodzinsky, 1966; Maldonado‐Capriles, 1990). Notably, Madagascar and its neighbouring islands represent a hotspot for Leistarchini diversity, harbouring 12 of the 15 genera recorded in the region, with high levels of endemism (Villiers, 1979; Chen et al., 2022). Conversely, the New World fauna lacks endemic genera, with Ploiaria being the sole representative (McAtee and Malloch, 1925; Wygodzinsky, 1966). This disproportionate global distribution suggests a complex biogeographic history; however, the temporal origin and ancestral distribution areas of Leistarchini remain unknown.

Originally proposed by Stål (1862), Leistarchini has been consistently recognized in subsequent classification schemes (Stål, 1874; Van Duzee, 1916, 1917; Wygodzinsky, 1966; Maldonado‐Capriles, 1990; Standring et al., 2023). While recent studies support the tribes's monophyly (Standring et al., 2023; Masonick et al., 2024), Leistarchini has never been subjected to a comprehensive phylogenetic analysis across its entire range. In Wygodzinsky's (1966) assessment of phylogenetic relationships, Leistarchini was placed between Collartidini and the lineages now recognized as Emesini, considered an early‐branching clade possessing numerous plesiomorphic features. He also proposed relationships among several genera based on morphological traits, which constitute testable hypotheses. Castro‐Huertas et al. (2021) published a morphology‐based phylogenetic analysis of Emesinae, focusing on the monophyly of Metapterini (now a junior synonym of Emesini), but included only two representatives of Leistarchini. Standring et al. (2023) inferred the phylogeny of Emesinae using molecular data from 37 Leistarchini species. Their analyses were primarily based on partial 18S and 28S rDNA sequences (3 865 bp), with backbone support provided by high‐throughput sequencing data (381 loci) from 12 emesine species. Leistarchini was recovered as monophyletic with high support, with Visayanocorini as its sister taxon. Furthermore, Ploiaria was shown to be polyphyletic, with Armstrongula Wygodzinsky, 1950, Gnomocoris McAtee & Malloch, 1926, and Leistarches Dohrn, 1860 deeply nested within it. Despite these findings, relationships among many genera remain unresolved, and an improved phylogenetic framework based on extensive sampling is necessary to clarify generic limits and provide context for the tribe's evolution.

A total 12 of the 33 valid genera of Leistarchini were described as monotypic, while another six comprise only two species, reflecting significant morphological diversity. A remarkable modification within the tribe is the reduction of the posterior lobe of the pronotum in winged individuals. In many genera, including Ploiaria, the posterior lobe is shortened into a collar‐like structure covering only the extreme anterior part of the mesonotum (Wygodzinsky, 1966; Villiers, 1979). An isolated instance is found in Lhostella Villiers, 1948, where the posterior lobe covers anterior one‐third to one‐half of the mesonotum. Wygodzinsky (1966) hypothesized that this reduction enables a flexible joint between the prothorax and mesothorax, enhancing mobility of the anterior part of the body. The morphology of the forelegs also varies significantly; the protarsus is often markedly elongated and strongly sclerotized, with tarsomeres reduced to two or entirely fused. Based primarily on the number of protarsal segments, Villiers (1948, 1949) distinguished the tribes Orthungini and Ploeariini, which correspond to Leistarchini in other schemes. This view was criticized by Wygodzinsky (1966) and rejected by most subsequent taxonomists, though it persisted in the work of Dispons (1970) and Villiers (1971, 1975, 1979). The ancestral state reconstruction of Standring et al. (2023) concluded that the non‐segmented protarsus was independently derived once within Leistarchini. However, the macroevolutionary patterns of these traits remain inadequately studied. Investigating the number and direction of these transitions within a well‐supported phylogenetic framework will provide critical insights into the classification and evolution of this group.

In this study, we present the most comprehensive phylogenetic analysis of Leistarchini to date, utilizing an expanded taxon sampling of 68 taxa representing ~65% of extant genera. We leveraged next‐generation sequencing (NGS) to acquire molecular markers from both ethanol‐preserved and dried museum specimens, the latter of which are critical for sampling rare taxa (Gillett et al., 2014; Timmermans et al., 2016; Lopes et al., 2024). We integrated mitochondrial genomes and nuclear rDNA with a subset of 24 taxa analysed via universal single‐copy orthologs (USCOs) to robustly resolve the tribe's backbone. Our primary objectives are to (i) generate a robust phylogenetic framework for Leistarchini based on extensive genus‐level sampling; (ii) perform divergence time estimation and ancestral area reconstruction to hypothesize the events that shaped the tribe's current global distribution; and (iii) explore the evolutionary history of key morphological traits, specifically the shape of the posterior pronotal lobe and the segmentation of the protarsus, to provide insights into the classification and character evolution of the tribe.

Material and methods

Taxon sampling

We sampled 88 terminal species for phylogenetic analysis including 68 species of Leistarchini. Nine species from other five tribes of Emesinae, as well as 11 non‐emesine Reduviidae were selected as outgroups. Neocentrocnemis stali (Centrocnemidinae) and Phymata pennsylvanica (Phymatinae) were used to root the tree as the Phymatine Complex is well recognized as the earliest diverging clade of Reduviidae (Hwang and Weirauch, 2012; Masonick et al., 2024). The taxon sampling for Emesinae was based on the taxonomic arrangements proposed by Standring et al. (2023) and Castro‐Huertas and Melo (2025) to include all recognized tribes, except for the monotypic Saiciretini. The 68 ingroup taxa represent 21 described genera of Leistarchini, covering approximately 65% of extant genera in the tribe. Molecular data of 53 species were newly sequenced in this study. Voucher information of the sampled specimens is shown in Table S1.

DNA extraction and sequencing

A total of 38 of the 53 newly‐sequenced samples originated from frozen (−20 °C) ethanol‐preserved specimens deposited in the Entomological Museum of China Agricultural University, Beijing, China (CAU). A further 15 dry‐pinned samples were selected from the historical collections of four institutions that were older than 25 years, with Lhostella sp. CONG being the oldest, collected in 1959 (Table S1). Total genomic DNA was extracted non‐destructively from the whole specimen (for dry‐pinned samples) or the right foreleg of the specimen (for ethanol‐preserved samples) using the Qiagen DNeasy Blood and Tissue Kit following the manufacture's protocol. To improve the quality of DNA extraction, each sample was soaked in tissue lysis buffer for about 20 h instead of the 1–3 recommended hours in the protocol. DNA concentrations were quantified using NanoDrop (Thermo Fisher Scientific, Waltham, USA) for quality verification before sequencing. Genomic DNA was sent to Berry Genomics (Beijing, China) for library preparation and sequencing. Sequencing libraries were generated with Illumina Truseq Nano DNA Sample Preparation Kit. All libraries were sequenced using the Illumina NovaSeq 6000 platform with 150 bp paired‐end reads and an insert size of approximately 350 bp, generating an estimated output of 10 G sequence data per sample. The raw reads were trimmed of adapters using Trimmomatic (Bolger et al., 2014) and short and low‐quality reads (>15 bp Ns or > 75 bp bases with a quality score ≤ 3) were removed using Prinseq 0.20.4 (Schmieder and Edwards, 2011).

The remaining data were downloaded from the National Center for Biotechnology Information (NCBI) database. To maximize the completeness of group coverage, a small number of representative species containing only partial molecular markers were also included. Among the 88 species sampled in our analyses, 24 have low‐coverage genomic or transcriptomic data and 70 have full‐length or near full‐length mitochondrial genome data. In addition, 81 and 86 species have full‐length or near full‐length 18S and 28S rDNA sequences, respectively. GenBank accession numbers of the newly‐sequenced and published data used in our analyses were listed in Table S2.

Molecular data processing

To acquire the mitogenome and nuclear rDNA (18S and 28S) sequences, the clean reads were assembled de novo using IDBA‐UD (Peng et al., 2012), with minimum and maximum k values of 41 and 141 bp, respectively. The mitogenome sequences were initially annotated by MitoZ 2.4 (Meng et al., 2019). The resultant gene boundaries were checked in Geneious Prime 2023 (Kearse et al., 2012, https://www.geneious.com) by alignment with homologous genes of the published mitogenomes (GenBank accession numbers PV626514 and PV626532), 18S (GenBank accession number PV647631) and 28S (GenBank accession number PV647665) of other Emesinae species. Fifteen mitochondrial genes (13 protein‐coding genes [PCGs], 12S and 16S) and two nuclear rDNAs (18S and 28S) were used in phylogenetic analysis. Each gene was aligned using MAFFT 7.450 (Katoh and Standley, 2013) with the L‐INS‐i algorithm. The alignment of the 13 mitochondrial PCGs was based on amino acid sequences translated with the invertebrate mitochondrial genetic codon table. All alignments were trimmed with trimAl 1.4.1 (Capella‐Gutiérrez et al., 2009) and concatenated with FASConCAT‐g 1.04 (Kück and Longo, 2014). Two datasets were generated for phylogenetic analysis: the PCGRNA matrix, which includes all three codon positions of the PCGs, and the four rDNAs (19 937 bp); the PCG12RNA matrix, which includes the first and second codon positions of the PCGs and the four rDNAs (16 438 bp).

The rapid‐assembly pipeline PLWS 1.0.6 (Zhang et al., 2019) was used to assemble the low‐coverage genomic and transcriptomic data. BBTools 38.29 (Bushnell, 2014) was used for quality control: duplicates were removed with clumpify.sh, quality trimming was done with bbduk.sh and normalization was done with bbnorm.sh. Redundant contigs were removed using Redundans 0.13c (Pryszcz and Gabaldón, 2016). We performed contig assembly using Minia 3.2.0 (Chikhi and Rizk, 2013). BESST 2.2.8 (Sahlin et al., 2014) was used for scaffolding, and GapCloser 1.12 (Luo et al., 2012) to fill gaps. We utilized BUSCO 3.0.2 (Waterhouse et al., 2018) to evaluate Benchmarking Universal Single‐Copy Orthologs completeness of assembly and extract USCOs from the assembled data, with the hemiptera_odb10 dataset (n = 2510) used as the reference gene set. The parameters of the mean USCO length in the control file “length_cutoff” were modified 2× to recover more “complete” USCOs, such that more fragmented USCOs could be treated as “complete” USCO sequences by the BUSCO pipeline. Loci were aligned in MAFFT 7.45 with the L‐INS‐i method and were trimmed in ClipKIT 1.1.5 (Steenwyk et al., 2020) with the “smart‐gap” model. We used TAPER (Zhang et al., 2021) to exclude errors within the alignments. To remove possible external contamination, we used Treeshrink 1.3.8b (Mai and Mirarab, 2018), a long‐branch detection algorithm, to detect the outlier long branches and remove their corresponding sequences. Finally, we generated four datasets with different taxon occupancies in PhyKit 1.11.1 (Steenwyk et al., 2021): the USCO75NT matrix, which comprises nucleotide sequences of 667 loci, with 75% taxon‐occupancy; the USCO75AA matrix, which comprises amino acid sequences of 667 loci, with 75% taxon‐occupancy; the USCO90NT matrix, which comprises nucleotide sequences of 142 loci, with 90% taxon‐occupancy; and the USCO90AA matrix, which comprises amino acid sequences of 142 loci, with 90% taxon‐occupancy. To improve backbone support, we combined the 13 mitochondrial PCGs and four rDNAs with 24 low‐coverage genomic or transcriptomic datasets to generate the most comprehensive matrix (the combined matrix; 159 loci, 259 164 bp). Detailed information of the matrices for phylogenetic analyses can be found in Table S3.

Phylogenetic analysis

For the combined matrix, phylogenetic relationships were reconstructed using the Maximum‐likelihood (ML) method in IQ‐TREE 2.1.2 (Minh et al., 2020). The best partitioning schemes and substitution models were selected using ModelFinder (Kalyaanamoorthy et al., 2017) based on the “greedy” algorithm and the Bayesian information criterion (BIC). The node support of the ML branching was evaluated with 1000 SH‐aLRT (Guindon et al., 2010) and 1000 UFBoot2 (Hoang et al., 2018) replicates. We also analysed the combined matrix using Maximum‐parsimony (MP) method in TNT 1.5 (Goloboff and Catalano, 2016). MP tree search was carried out using the New Technology Search employing the following search options: random drift, ratchet, sectorial search, tree fusing and find minimum tree length 100 times. Node support was assessed with 1000 bootstrap replicates. For the PCGRNA and PCG12RNA matrices, phylogenies were inferred using Bayesian inference (BI) and ML methods. BI analyses were conducted using PhyloBayes‐MPI 1.8 (Lartillot et al., 2013) with the site‐heterogeneous mixture CAT‐GTR model. Two independent Markov Chain Monte Carlo (MCMC) runs of 100 000 generations each were executed. Convergence was evaluated with the “bpcomp” and “tracecomp” procedure in the PhyloBayes package with a burn‐in of initial 25% by the recommended criterion of maximum discrepancy <0.1. A consensus tree was generated simultaneously by pooling the remaining MCMC trees of both runs, with Bayesian posterior probabilities (PP) on each node. ML partitioned analyses were conducted in IQ‐TREE 2.1.2, with node support evaluated with 1000 SH‐aLRT and 1000 UFBoot2 replicates.

Phylogenetic analyses for the USCO matrices were performed via concatenated and coalescent approaches. For the concatenated method, we conducted ML partitioned analyses in IQ‐TREE 2.1.2. Node support was computed with 1000 SH‐aLRT and 1000 UFBoot2 replicates. To address the effect of heterotachy on phylogenetic inference (Kolaczkowski and Thornton, 2004), we also used the General Heterogeneous evolution On a Single Topology (GHOST) model (Crotty et al., 2019) in IQ‐TREE. Multi‐species coalescent (MSC) supertree analyses addressing incomplete lineage sorting (ILS) were performed in ASTRAL 1.1.5 (Zhang et al., 2018) with default parameters, using individual gene trees generated in IQ‐TREE from the USCO matrices. We applied the weighted method in ASTRAL (Zhang and Mirarab, 2022) to improve the utility of tree reconciliation. The resulting trees were visualized and edited in Figtree 1.4.4 (Rambaut, 2018) and iTOL 6.8.1 (Letunic and Bork, 2021).

Divergence time estimation

MCMCTree in the Phylogenetic Analysis by Maximum Likelihood (PAML) 4.9j package (Yang, 2007) was used to estimate divergence times of Emesinae with a relaxed molecular clock based on topology corresponding to the combined matrix. The root age of the tree was set to less than 179.6 million years ago (Ma) according to the upper boundary of the split between the Phymatine Complex and the Trichobothrial Clade (formerly “Higher Reduviidae”) estimated by Hwang and Weirauch (2012). We used Acutiangulus densus Zhang, Liu & Yao, 2024 from the Cretaceous Burmese amber that is estimated to be 94.3–99.7 million years old (Zhang et al., 2024) to calibrate Neocentrocnemis plus Phymata. The Eocene compression fossil Aphelicophontes danjuddi Swanson, Heads & Taylor, 2021 (48.5–53.5 Ma, Swanson et al., 2021) and the Miocene Mexican amber fossil Empicoris electricus Thomas, 1992 (15.9–23.0 Ma, Thomas, 1992) were selected to calibrate Harpactorinae and Empicoris, respectively. Emesinae was calibrated by an undescribed species from the Cretaceous New Jersey amber (66.0–72.1 Ma, Michener and Grimaldi, 1988). Ectrichodiinae was constrained to 72.87–95.55 Ma following the result from Hwang and Weirauch (2012) as a secondary calibration. The upper and lower bound values of each fossil were designated based on the original reference and the Paleobiology Database (https://paleobiodb.org/). According to the assessed overall mutation rate obtained utilizing baseml (in PAML) under the GTR + G substitution model (model = 7, alpha = 0.5), the Dirichlet‐gamma prior to the overall substitution rate was set as rgene_gamma = 1, 20. The MCMCTree program was run for 1000,000 MCMC steps and sampled every 50 after a 10 000 burn‐in. The robustness of the MCMCTree results was checked by comparing the consistency of the two independent runs, with all parameters ≥200 for the effective sample sizes. The resulting chronogram was visualized and edited in TVBOT 2.5.3 (Xie et al., 2023).

Ancestral area reconstruction

An ancestral area reconstruction analysis was performed to understand the biogeographic history of Leistarchini. The 68 ingroup taxa were assigned to seven distribution ranges: A = Afrotropical, B = Madagascan, C = Oriental, D = Australasian, E = Palaearctic, F = Nearctic, G = Neotropical. Ancestral areas were reconstructed using Reconstruct Ancestral State in Phylogenies (RASP) 4.3 (Yu et al., 2020) with the BioGeoBEARS package (Matzke, 2014). The consensus tree derived from the combined matrix was used as input with outgroups excluded. The “+J” parameter (a model accounting for founder‐event speciation) was retained in our analysis as recommended in Klaus and Matzke (2020) and Matzke (2022). Six alternative biogeographical models (DEC, DEC + J, DIVALIKE, DIVALIKE+J, BAYAREALIKE and BAYAREALIKE+J) in the BioGeoBEARS were compared based on the AICc_wt value. To reduce the acquisition of multiple areas at nodes near the root, the maximum number of areas for each node was set to two. Species distribution data are based on the examined specimens in the present study and on literature.

Ancestral character state reconstruction

To better understand the evolutionary history of Leistarchini, we performed ancestral state reconstruction (ASR) for two representative morphological traits: the shape of the posterior lobe of the pronotum, and the segmentation of the protarsus. In winged individuals of Leistarchini, the shape of the posterior pronotal lobe can be distinguished into two types: a well‐developed posterior pronotal lobe, which covers the entire mesonotum, and a collar‐like posterior pronotal lobe that is strongly shortened, covering only the extreme anterior part of the mesonotum. Additionally, a moderately‐shortened posterior lobe is observed in Lhostella, which covers the anterior one‐third to one‐half of the mesonotum. This condition differs from the typical collar‐like posterior lobe and is therefore coded separately. The shape of the posterior pronotal lobe is considered consistent within a given genus and has historically served as an important diagnostic character for genera of Leistarchini. Similarly, the number of protarsal segmentation has been widely used as a genus‐level diagnostic character in Leistarchini. The most common configuration features three tarsomeres, though two‐ or non‐segmented protarsus has been recorded in certain genera. Villiers (1948, 1979) and Dispons (1970) respectively used segmented and non‐segmented protarsi to define their tribes Leistarchini and Orthungini. The analyses were conducted using Mesquite 3.70 (Maddison and Maddison, 2021) under ML approach with Mk1 model. The ML topology based on the combined matrix was used as input.

Results

Phylogenetic analysis

ModelFinder merged the original 159 partitions of the combined matrix into 30 partitions, with the best‐fitting model for each partition listed in Table S4. The best partitioning schemes and substitution models for the PCGRNA and PCG12RNA datasets are listed in Tables S5 and S6. The topologies of the combined dataset integrating USCOs, mitochondrial PCGs and four rDNAs (Fig. 1, Figs S1 and S2) are largely identical to those derived from the mitochondrial PCGs and four rDNAs only (Figs S3–S6) and USCOs only (Fig. 2, Figs S7–S18) datasets. All analyses recovered a monophyletic Emesinae. We employed the ML topology from the combined dataset for biogeographic and morphological character evolution analyses. This topology exhibits increased support across the backbone and is broadly congruent with a previous phylogenetic study (Standring et al., 2023).

Fig. 1.

Phylogenetic inference of Leistarchini. Topology derived from the maximum‐likelihood (ML) analysis of the combined matrix. Colour‐coded circles at nodes indicate ML support values as specified in the legend in the bottom‐left corner. An asterisk denotes that the clade is also present in the maximum‐parsimony (MP) analysis, with the colours indicate MP bootstrap support values as specified in the legend in the bottom‐left corner. Taxa displayed on the right: (a) Pleias fashengi Chen, Li & Cai from China (© Zhuo Chen); (b) Monicacoris atratus Chen, Li & Cai from Madagascar (© Hu Li); (c) Orthunga sp. from Madagascar (© Hu Li); (d) Calliemesa hainana (Hsiao) comb. n. from China (© Zhuo Chen); (e) Tinna picta Wygodzinsky from Namibia (© Ye Liu); (f) Ploiaria sp. from China (© Weiwei Zhang).

Fig. 2.

Phylogenetic inference of Leistarchini. Topology derived from the maximum‐likelihood (ML) analysis of the USCO75NT matrix. Colour‐coded square matrices at nodes indicate support values as specified in the legend in the bottom‐left corner. Unlabelled nodes indicate full support (UFBoot = 100, SH‐aLRT = 100, wASTRAL support = 1) in all analyses.

All analyses recovered the monophyly of Leistarchini with high support (PP = 1, SH‐aLRT = 88.2–100, UFBoot = 99–100, MP support = 97, wASTRAL support = 1). Collartidini was highly supported (PP = 1, SH‐aLRT = 99.6–100, UFBoot = 100, MP support = 93) as sister group to Leistarchini in analyses based on the combined dataset as well as the PCGRNA and PCG12RNA datasets. The position of Visayanocorini, previously considered closely related to Leistarchini, varied across analyses. In the combined analyses, Visayanocorini was recovered as sister taxon to Saicini, whereas this tribe was supported as sister taxon to the remaining Emesinae in the PCGRNA and PCG12RNA analyses. The phylogenetic relationships within Leistarchini were largely consistent across analyses, revealing five major clades (Clades A–E, Figs 1 and 2). Our results provide support for the placement of several genera previously excluded from phylogenetic analyses and reveal a novel lineage. In addition, we uncover new details concerning relationships within some species‐rich genera such as Orthunga Dohrn, 1859, Pleias Kirkaldy, 1901 and Ploiaria.

Gomesius Distant, 1903 is a small genus of six described species distributed in South and Southeast Asia. This genus formed an isolated lineage (Clade A) in all analyses and was recovered as sister taxon to the remaining Leistarchini in most analyses. Clade B (referred to as the Pleias clade) comprises Guithera Distant, 1906 (Oriental), Lhostella (Afrotropical), Pleias (Old World tropical) and Proguithera Wygodzinsky, 1966 (Oriental and Palaearctic). In most analyses, the Pleias clade was recovered as sister taxon to the remaining Leistarchini except Gomesius, but in ML and GHOST analyses for the USCO90AA dataset it formed a sister relationship with Clade C. Within the Pleias clade, the sister‐group relationship of Guithera and Proguithera was fully supported (PP = 1, SH‐aLRT = 100, UFBoot = 100, MP support = 100), while Pleias was paraphyletic with respect to Lhostella.

The eight Madagascan genera included in this study (Ambrinemesa Villiers, 1971, Bettyella Wygodzinsky, 1966, Millotina Villiers, 1953, Monicacoris Putshkov, 1985, Nesita Bergroth, 1906, Paranesita Wygodzinsky, 1966, Tinnunga Wygodzinsky, 1966 and Voloina Villiers, 1971) and the genus Orthunga (Afrotropical and Madagascan) constituted the Clade C, here referred to as the Orthunga clade. The phylogenetic relationships among these genera are largely consistent across analyses, including a fully supported sister relationship between Bettyella and Paranesita, and between Monicacoris and Tinnunga. Orthunga was rendered paraphyletic by Nesita. The Afrotropical genus Paraluteva Villiers, 1961 was also included in the Orthunga clade in most analyses, which was either recovered as the earliest diverging lineage of the entire clade (in PCGRNA analyses and ML analysis for the PCG12RNA dataset), or as sister taxon to the subclade comprising the genera Ambrinemesa, Bettyella, Millotina, Paranesita and Voloina (in the combined analyses).

The next clade recovered in our analyses (Clade D) comprised Ploiaria hainana Hsiao, 1965 and six morphologically‐similar undescribed species. We found, however, that this group of species does not belong to Ploiaria but instead forms sister branch to the clade comprising Ploiaria and other related genera (Clade E). This result is robust to all analytical treatments and necessitates a new genus name. Further morphological examination revealed a series of character states that distinguish them from the currently defined Ploiaria. The detailed description of the new genus Calliemesa gen. n. is provided in the Taxonomy section.

The hyperdiverse genus Ploiaria, along with five other genera (Gnomocoris, Leistarches, Orianocoris Villiers, 1964, Phryxobotrys McAtee and Malloch, 1926 and Tinna Dohrn, 1859), constituted the large Clade E (referred to as the Ploiaria clade) which is characterized by rampant paraphyly. The Old World genus Tinna clustered with Afrotropical species of Ploiaria and was highly supported (PP = 0.98, SH‐aLRT = 99.8–100, UFBoot = 97–100, MP support = 92, wASTRAL support = 1) as the earliest diverging lineage of the Ploiaria clade. While the Oriental genera Gnomocoris and Phryxobotrys were recovered as monophyletic, they were deeply nested within Ploiaria as sister taxon to some species currently classified as the latter. Leistarches is represented in this study by its type species L. serripes Dohrn, 1860, which formed a subclade with two Ploiaria species from the Australasian Region. The monotypic genus Orianocoris was also nested within Ploiaria in our analyses, but the exact position of this genus will benefit from additional testing. It was recovered as sister taxon to a subclade containing Ploiaria species from the Australasian and Oriental regions in the combined analyses, but showed close relationships to the New World species of Ploiaria in the PCGRNA and PCG12RNA analyses.

Divergence time estimation

Age estimates (median age values and 95% Highest Posterior Density [HPD]) from the MCMC molecular dating analysis are shown in Fig. 3 and reported in Table S7. The divergence of Leistarchini from other Emesinae tribes happened ca. 61.94 Ma (95% HPD 55.76–68.5 Ma), with its crown‐group age estimated at 57.04 Ma (95% HPD 49.53–64.86 Ma). The divergence of five major clades occurred during the late Palaeocene and middle Eocene: the Pleias clade was estimated to have diverged from the most recent common ancestor of the other three clades ca. 52.86 Ma (95% HPD 45.43–59.72 Ma); the Orthunga clade diverged from Calliemesa gen. n. + the Ploiaria clade ca. 49.85 Ma (95% HPD 43.58–56.54 Ma); the spilt of Calliemesa gen. n. and the Ploiaria clade was estimated ca. 46.07 Ma (95% HPD 37.76–54.72 Ma). Within the Pleias clade, Guithera + Proguithera was inferred to have diverged from “Pleias” + Lhostella ca. 45.48 Ma (95% HPD 38.27–53.10 Ma). The Orthunga clade was estimated to have a crown age of 44.5 Ma (95% HPD 38.55–51.05 Ma), with Paraluteva diverged from other Malagasy lineages ca. 40.93 Ma (95% HPD 32.83–48.69 Ma). The diversification of Calliemesa gen. n. was estimated to have taken place ca. 36.22 Ma (95% HPD 24.91–47.42 Ma). Within the Ploiaria clade, the subclade encompassing Tinna and Afrotropical Ploiaria diverged from the remaining taxa ca. 41.81 Ma (95% HPD 33.61–49.91 Ma); the New World Ploiaria diverged from the related Old World taxa ca. 29.65 Ma (95% HPD 18.59–41.07 Ma); the Mauritius Orianocoris was estimated to have diverged from the related taxa ca. 32.04 Ma (95% HPD 21.51–41.04 Ma); Leistarches and Australasian Ploiaria diverged from the related Oriental taxa ca. 27.28 Ma (95% HPD 12.62–29.97 Ma); two additional colonization events into the Australasian Region were estimated to have occurred ca. 18.42 Ma (95% HPD 7.46–27.56 Ma) and 18.14 Ma (95% HPD 8.21–27.72 Ma).

Fig. 3.

Chronogram with estimated divergence times among Leistarchini species based on a MCMCTree analysis, with ancestral areas reconstructed from DIVALIKE+J in BioGeoBEARS. Median age estimates (Ma) are shown along nodes; blue bars represent 95% highest posterior density (HPD) intervals of age estimates; purple stars indicate placement of fossil or secondary calibration points of clades; pie charts indicate relative likelihoods of each reconstructed distribution, with geographical distribution color coded according to the legend shown in the figure. Ple, Pleistocene; Pli, Pliocene.

Ancestral area reconstruction

The biogeographical model selection analysis suggested that DEC + J is the best scoring model for our data (Table S8). The ancestral geographic ranges inferred from the DEC + J analysis are summarized in Table S9, with the major results illustrated on the maximum clade credibility chronogram (Fig. 3). Our analysis recovered a tropical Asian distribution for the ancestral node of Leistarchini. Most earlier diversifications of the subordinate clades of Leistarchini were inferred to have taken place in the Oriental Region, followed by multiple intercontinental dispersals to other continents. The ancestral area of “Pleias” + Lhostella was inferred to include the Afrotropical and Oriental regions. If we take the biogeographic results at face value, at least two dispersals to the Afrotropical Region from the Oriental Region have taken place within this lineage. The ancestor of the Orthunga clade was reconstructed to have dispersed to Madagascar from the Oriental Region between 49.85 and 44.5 Ma, where it diversified in situ. Subsequently, the ancestor of Paraluteva dispersed to the Afrotropical Region from Madagascar. The biogeographic analysis indicated multiple colonization events from the Oriental Region to other areas within the Ploiaria clade. The ancestor of the subclade containing Tinna and Afrotropical Ploiaria was inferred to have dispersed from Asia to Africa between 41.81 and 34.33 Ma. The ancestor of Orianocoris dispersed from Asia to Mauritius during the early Oligocene. Colonization of the Australasian Region occurred at least three times, with one major dispersal event by the subclade represented by Leistarches and two minor incursions to the northern Australasian Region. Current evidence suggests that the Neotropical Region appears to have been colonized only once by Ploiaria between 29.65 and 25.41 Ma.

Ancestral state reconstruction on morphological characters

We reconstructed the evolutionary history of the shape of the posterior pronotal lobe and the segmentation of the protarsus in Leistarchini (Fig. 4). Our reconstruction indicated that the ancestral form of the posterior lobe of the pronotum in Leistarchini was well‐developed, entirely covering the mesonotum (Fig. 4a). Three independent transitions to a collar‐like posterior lobe (Fig. 4c) occurred at the ancestral nodes of Gomesius, Monicacoris + Tinnunga and the Ploiaria clade, respectively. In addition, an independent shift to a moderately‐shortened posterior lobe (Fig. 4b) occurred in Lhostella. The reconstruction of the segmentation of the protarsus indicated that the ancestor of Leistarchini possessed three tarsomeres (Fig. 4d,e). Reduction in protarsal segmentation occurred four times within unrelated lineages of Leistarchini, with transitions to a non‐segmented protarsus (Fig. 4g) at the ancestral nodes of Gomesius, the subclade containing Monicacoris, Nesita, Orthunga and Tinnunga, and Phryxobotrys. The ancestor of Guithera + Proguithera more likely possessed a two‐segmented protarsus (Fig. 4f), a trait retained in Proguithera but further reduced into a non‐segmented protarsus in Guithera.

Fig. 4.

Summary of ancestral state reconstructions for morphological traits in Leistarchini. Topology derived from the ML analysis based on the combined matrix. Pie charts along nodes indicate relative likelihoods of ancestral states reconstructed in Mesquite 3.70. Scoring of each character for taxa is indicated in the circles after the taxon names. (a) Pleias furcosa (Ribes), anterior part of body, dorsal view (scale bar 3 mm); (b) Lhostella leleupi Villiers, anterior part of body, dorsal view (scale bar 1 mm); (c) Gomesius hesione (Kirkaldy), anterior part of body, dorsal view (scale bar 2.5 mm); (d) Ploiaria hamulosa Chen, Li & Cai, protarsus, lateral view (scale bar 0.5 mm); (e) Pleias fashengi, protarsus, lateral view (scale bar 1 mm); (f) Proguithera caspersi Chen, Li & Cai, protarsus, lateral view (scale bar 0.5 mm); (g) Gomesius hesione, protarsus, lateral view (scale bar 2.5 mm).

Taxonomy

Genus Calliemesa gen. n.

(Figs 5 and 6)

Fig. 5.

Habitus and morphological details of Calliemesa hainana comb. n. (a) Macropterous male, dorsal view; (b) apterous female, dorsal view; (c) anterior part of body of macropterous male, dorsal view; (d) anterior part of body of macropterous male, lateral view; (e) anterior part of body of apterous female, dorsal view; (f) anterior part of body of apterous female, lateral view; (g) mesosternum of macropterous male, ventral view; (h) foreleg of macropterous male, ventral view; (i) hemelytron of macropterous male, dorsal view. Scale bar of A, B = 5 mm; of C–F, I = 1 mm; of G, H = 0.5 mm.

Fig. 6.

Genitalia of Calliemesa hainana comb. n. (a) Abdominal segment VIII of male, ventral view; (b) abdominal segment VIII of male, lateral view; (c) pygophore, dorsal view; (d) pygophore, lateral view; (e) paramere, lateral view; (f) paramere, dorsal view; (g) phallus, dorsal view; (h) phallus, lateral view; (i) phallus, ventral view; (j) genital segments of female, dorsal view; (k) genital segments of female, lateral view; (l) genital segments of female, ventral view. Scale bar of A–I = 1 mm; of J–L = 1.5 mm.

http://zoobank.org/urn:lsid:zoobank.org:act:341F77E9‐74F8‐42F8‐829D‐CE65A2FC4441

Type species. Ploiaria hainana Hsiao, 1965.

Included species. Calliemesa hainana (Hsiao, 1965) comb. n. The other six new species will be described in a separate paper.

Diagnosis. Recognized within Leistarchini by the following combination of character states: posterior lobe of pronotum well‐developed in macropterous form, covering mesonotum (Fig. 5c,d); posterior margin of prosternum rounded in macropterous form (Fig. 5g), emarginated in apterous form; profemur armed ventrally with two series of spine‐like setae arising from short tubercles, and accessory series composed of short spine‐like setae (Fig. 5h); anteroventral series widely interrupted at base (Fig. 5h); protarsus half of length of protibia, three‐segmented, with tarsomere I twice as long as tarsomere III (Fig. 5h). Macropterous and apterous forms are known. A detailed description of the new genus is provided in File S1.

Etymology. The generic name is derived from the Latin calli‐ (meaning beautiful) and Emesa (the nominate genus name of the subfamily Emesinae). Gender feminine.

Distribution. Oriental Region.

Remarks. The type species of Calliemesa gen. n. was originally placed in Ploiaria. The new genus differs from Ploiaria in the well‐developed posterior lobe of the pronotum in winged individuals (Fig. 5a,c,d) (vs collar‐like in Ploiaria) and the presence of the accessory series on the profemur (Fig. 5h) (vs absent in Ploiaria). Calliemesa gen. n. is morphologically similar to the Oriental genus Bagaudina Wygodzinsky, 1966, but can be readily distinguished from the latter by the following character states: posterior lobe of pronotum not expanded posteriorly, leaving scutellum exposed (Fig. 5c,d) (vs strongly expanded posteriorly, covering scutellum in Bagaudina); accessory series present on profemur (Fig. 5h) (vs absent in Bagaudina); anteroventral series of profemur widely interrupted at base (Fig. 5h) (vs not interrupted in Bagaudina); protarsus half of length of protibia (Fig. 5h) (vs much less than half of length in Bagaudina); tarsomere I of protarsus twice as long as tarsomere III (Fig. 5h) (vs tarsomeres of protarsus subequal in length in Bagaudina); abdominal segment VIII in male with nearly straight posterolateral margins (Fig. 6a,b) (vs distinctly curved in Bagaudina).

Discussion

Phylogeny and generic classification of Leistarchini

Our phylogenetic analysis provides strong support for the monophyly of Leistarchini. Since its initial proposal by Stål (1862), Leistarchini has been widely recognized as a tribe within Emesinae across various classification schemes (Wygodzinsky, 1966; Standring et al., 2023). However, a consensus regarding the phylogenetic position of Leistarchini within the subfamily remains elusive. Wygodzinsky (1966) tentatively placed Leistarchini between Collartidini and Emesini. Previous morphological phylogenetic studies focusing on various emesine lineages (e.g. Smith et al., 2019; Castro‐Huertas et al., 2021; Castro‐Huertas and Melo, 2025) naturally provided limited information regarding the phylogenetic position of Leistarchini, as this tribe was not the primary focus of their sampling strategies. While the recent molecular study by Standring et al. (2023) supported a sister‐group relationship between Leistarchini and Visayanocorini, our results recover Collartidini as the sister taxon to Leistarchini with high support, with Visayanocorini diverging distantly. Given that Collartidini and Visayanocorini are currently absent from the USCOs datasets, their precise relationship with Leistarchini warrants further testing.

Our reconstruction of the internal relationships within Leistarchini reveals five major clades, three of which were also recovered by Standring et al. (2023). Of the two newly identified clades, one is represented by Gomesius, a genus previously omitted from phylogenetic analyses, while the other constitutes a novel lineage that cannot be accommodated within any existing genus. This newly developed phylogenetic framework allows for a more comprehensive understanding of genus‐level relationships within the tribe.

Members of Gomesius are large‐bodied species recognizable by a collar‐like posterior pronotal lobe, a long penicillate process situated subbasally on the profemur and a non‐segmented protarsus (Wygodzinsky, 1966; Ghate and Mathew, 2018). No prior hypotheses have been proposed regarding the phylogenetic affinites of this genus. Our analyses indicate that Gomesius represents an isolated, early‐diverging clade within Leistarchini.

According to our results, Guithera, Lhostella, Pleias and Proguithera constitute the Pleias clade, corroborating the speculation by Chen et al. (2025). These genera are united by several morphological traits: a more or less convex dorsal head surface, with a shallow transverse interocular sulcus; a broadly rounded posterior prosternal margin; an accessory series on the profemur; a strongly sclerotized protarsus exceeding half the length of the protibia (where the first tarsomere, if segmented, is longer than the others); and similar hemelytral venation. The Afrotropical genera Bagaudella Miller, 1952 and Pseudobagauda Wygodzinsky, 1966, alongside the Oriental Lutevula Breddin, 1909, exhibit similar morphology and are likely members of this clade. Our results support a close relationship between Guithera and Proguithera. Although Guithera, Lutevula and Proguithera are considered closely related, their taxonomic status remains controversial (Wygodzinsky, 1966; Rédei, 2004; Chen et al., 2021). They are distinguished from other members of the Pleias clade by a narrowed anteocular region, paired transverse tubercles on the posterior pronotal lobe, a reduced number of spine‐like setae on the profemur and fused protarsal segments. Furthermore, we find Lhostella nested within Pleias. Although originally described within Pleias (Lhoste, 1939), Wygodzinsky (1966) treated Lhostella as a separate genus based on its markedly shortened posterior pronotal lobe. Increased sampling of Lhostella is required to fully resolve its relationship with Pleias.

The Orthunga clade comprises Orthunga and eight Malagasy genera, a grouping also supported by Standring et al. (2023). Interestingly, Orthunga is rendered paraphyletic by Nesita. Wygodzinsky (1966) suggested these genera might share a recent common ancestor. Nesita mirrors Orthunga in most respects, differing only in the absence of spiniferous processes on the ventral surface of the head. While these processes have traditionally been used to distinguish genera, our findings suggest their taxonomic significance at the generic level is limited, a pattern also observed in Tinna and Ploiaria. Notably, Paraluteva is recovered as sister to or nested within the Orthunga clade in most of our results, contrasting with Standring et al. (2023), where it was sister to the Ploiaria clade. This placement requires further investigation with expanded data.

We propose the new genus Calliemesa gen. n. for Ploiaria hainana and six similar undescribed species. This genus is recovered as the sister group to the Ploiaria clade. Unlike the currently defined Ploiaria, Calliemesa gen. n. possesses a well‐developed posterior pronotal lobe and an entire posterior prosternal margin in macropterous individuals, along with an accessory series on the profemur. While it resembles the Oriental genus Bagaudina, it differs significantly in thoracic and foreleg characters. Future inclusion of Bagaudina will be necessary to test their relationship.

Ploiaria is currently the largest and most morphologically diverse genus in Leistarchini. Previous taxonomists noted that many species, despite striking differences in habitus, could not be reliably partitioned into separate genera (e.g. McAtee and Malloch, 1925, 1926; Wygodzinsky, 1966; Ishikawa et al., 2008; Rédei, 2008). Molecular data (Standring et al., 2023) confirmed the polyphyly of Ploiaria, and our results reveal that the history of this taxonomic behemoth is exceedingly complex. At least six smaller genera are nested within Ploiaria across multiple lineages, signalling the need for a total systematic overhaul. For instance, Tinna is non‐monophyletic and forms a subclade with certain Afrotropical Ploiaria. Gnomocoris, originally a subgenus of Ploiaria (McAtee and Malloch, 1926) before its elevation by Wygodzinsky (1966), is closely related to Ploiaria species from the Malay Archipelago, suggesting it may need to be synonymized or returned to subgeneric status. Similarly, the monotypic Orianocoris is nested within Ploiaria in all analyses. Although its type species, O. spinosus Villiers, 1964, possesses unique spine‐like processes on the thorax and abdomen, its pronotal and foreleg morphology align with the currently defined Ploiaria, suggesting that Orianocoris may be considered a junior synonym of Ploiaria. Furthermore, the tribe's type genus, Leistarches, is deeply embedded within Ploiaria, consistent with the findings of Standring et al. (2023). While Leistarches can be distinguished by its rounded posterior prosternal margin and profemoral setal arrangement, these characters warrant reassessment given the internal complexity of Ploiaria. Finally, the peculiar genus Phryxobotrys is also deeply nested within Ploiaria, further highlighting the morphological complexity of the Ploiaria clade. Our current datasets provide a representative sampling of Ploiaria, which will serve as a foundation for future studies to disentangle the phylogeny and classification of the Ploiaria complex.

Historical biogeography of Leistarchini

By generating a time‐calibrated phylogenetic framework, we investigated the origin and historical biogeography of Leistarchini. Extant members of the tribe exhibit a cosmopolitan distribution, yet diversity varies significantly across zoogeographic regions. Madagascar and its neighbouring islands represent a pinnacle of generic richness and endemism, harbouring 80% of recorded genera and an entirely unique species‐level fauna (Villiers, 1971, 1979; Maldonado‐Capriles, 1990; Chen et al., 2022). In contrast, the New World is relatively species‐poor, represented by approximately 40 reported species of Ploiaria (Wygodzinsky, 1966; Maldonado‐Capriles, 1990). Furthermore, several endemic species are distributed across varies Pacific islands (Wygodzinsky and Usinger, 1960; Maldonado‐Capriles, 1990). Despite these striking patterns, formal biogeographic analyses of the tribe have been previously absent.

Our ancestral range reconstruction suggests that Leistarchini began diversifying in Southeast Asia approximately 57 Ma (95% HPD 49.53–64.86 Ma). Many early divergence events are inferred to have occurred within the Oriental Region during the early to middle Eocene (40–53 Ma). During this period, Southeast Asia underwent dramatic geological and climate transitions, including the breakup and insularization of the Sundaland, the collision of the Indian and Eurasian plates and significant global temperature fluctuations (Hall, 2009, 2013; Lohman et al., 2011; Westerhold et al., 2020). These events likely increased regional ecological complexity, mediating early diversification within the tribe. Once Leistarchini became sufficiently diverse and widespread throughout Southeast Asia, the probability of successful dispersal into extra‐limital regions likely increased.

A pivotal transcontinental dispersal event occurred when the ancestor of the Orthunga clade colonized Madagascar from the Oriental Region during the Eocene (44.5–50 Ma), subsequently triggering a major radiation on the island. Although Madagascar has been an isolated landmass throughout the Cenozoic (<65 Ma), much of its extant fauna arrived via later colonization (Rabinowitz et al., 1983; Vences, 2004; Yoder and Nowak, 2006). Several hypotheses address dispersal from the Oriental Region to Madagascar, often highlighting the role of the Indian Plate (Yoder and Nowak, 2006; Ali and Aitchison, 2008). While the Indian Plate moved progressively northward, it may have remained connected to Madagascar via the Seychelles‐Mascarene Plateau. This connection was potentially extended in the Paleogene by the development of volcanic island chains (Ali and Aitchison, 2008; Zhou et al., 2012). Warren et al. (2010) also noted that sea‐level fluctuations could have exposed a chain of islands between India and Madagascar, serving as stepping stones that reduced the open‐ocean distance. Climatic factors, such as winter monsoon winds and ocean currents, may have further promoted long‐distance dispersal from Asia to Madagascar (Goswami and Rajagopal, 2003; New et al., 2005; Cheke and Hume, 2008). Similar dispersal scenarios have been proposed for praying mantises (Svensen and Whiting, 2009), millipede assassin bugs (Forthman and Weirauch, 2016) and giant stick insects (Glaw et al., 2019).

Additional long‐distance dispersal events from Asia to Africa occurred between the middle Eocene and early Oligocene. The continued contraction of the Tethys Ocean during the Cenozoic may have provided terrestrial routes for these movements (Zhang et al., 2014). Alternatively, Asian ancestors may have reached Africa via the boreotropical forests that flourished during the Eocene and Oligocene (Wolfe, 1975). Subsequent climatic cooling in the late Eocene and Oligocene likely led to the fragmentation of these habitats (Griffin, 2002; Westerhold et al., 2020), promoting vicariance between African and Asian populations. Our results also support an “out‐of‐Madagascar” dispersal to the Afrotropical Region for Paraluteva approximately 18–41 Ma. A proposed land bridge, the Davie Ridge, may have existed between Africa and Madagascar during the middle Eocene to early Miocene (ca. 45–26 Ma) (McCall, 1997). Although this land bridge remains debated, others suggest that isolated islands in the Mozambique Channel provided the necessary stepping stones for dispersal from Madagascar to the African continent (Rogers et al., 2000; de Wit, 2003).

The current global distribution of Leistarchini was primarily shaped by multiple intercontinental dispersals within the Ploiaria clade. Colonization from the Oriental to the Australasian Region is inferred to have occurred during the late Oligocene and early Miocene. The collision of the Asian, Australian and Pacific plates around 25 Ma resulted in the formation of Wallacea (Hall, 2013) and the orogeny of New Guinea (Toussaint et al., 2014), allowing faunal exchange across shallow straits and ephemeral land bridges (Voris, 2000). Our findings indicate that the Ploiaria clade colonized the Neotropical Region only once, during the Oligocene (ca. 25–30 Ma). The ancestor of these lineages likely reached the New World via a high‐latitude land connection, such as the Bering Land Bridge, linking East Asia with western North America before migrating southward. The Bering Land Bridge facilitated biotic exchange from the early Cenozoic until the late Pliocene (ca. 3.5 Ma). Many insect groups are believed to have utilized this route, including ambush bugs (Masonick et al., 2017), butterflies (Tseng et al., 2022) and cave crickets (Kim et al., 2024). While the estimated timing of the Ploiaria expansion coincides with the availability of the Bering Land Bridge, direct evidence is still lacking. Further sampling of Nearctic and eastern Palaearctic taxa is required to test this hypothesis. Simultaneously, transoceanic dispersal should not be discounted; the presence of Ploiaria on remote Pacific islands like Fiji and the Marquesas Islands (Wygodzinsky and Usinger, 1960; Wygodzinsky, 1966; Maldonado‐Capriles, 1990) demonstrates the genus's capacity for long‐distance overwater dispersal.

Complex pattern of character evolution in Leistarchini

The current phylogenetic framework for Leistarchini reveals a complex evolutionary history for two key morphological characters: the shape of the posterior pronotal lobe and the segmentation of the protarsus. These traits have historically been used to delineate genera (e.g. Wygodzinsky, 1966; Maldonado‐Capriles, 1992; Rédei, 2004) and have been regarded as having signigicant phylogenetic weight (e.g. Villiers, 1948, 1979; Dispons, 1970; Standring et al., 2023).

In Emesinae, as in other Reduviidae subfamilies, the pronotum is divided into anterior and posterior lobes by a transverse constriction. In winged individuals, the posterior lobe is typically trapezoidal, well‐developed and completely covers the mesonotum. However, in several lineages of Collartidini, Emesini and Leistarchini, the posterior lobe is markedly shortened, leaving much of the mesonotum exposed (Wygodzinsky, 1966; Rédei and Tsai, 2010; Standring et al., 2023). While this reduction is unique to these emesine tribes within Reduviidae, it is considered a significant genus‐level taxonomic character. Of the 28 Leistarchini genera with known winged forms, 11 possess a shortened posterior lobe. Our reconstruction identifies a well‐developed posterior lobe as the ancestral state for Leistarchini, followed by four independent transitions to a shortened lobe across unrelated lineages. Notably, a moderate shortening of the posterior lobe occurred in Lhostella, a genus nested within Pleias in all our analyses, despite the latter possessing a fully developed posterior lobe. This finding necessitates a reassessment of the taxonomic status and generic limits of both Lhostella and Pleias.

Wygodzinsky (1966) proposed that a shortened posterior pronotal lobe enhances the mobility of the anterior body, providing a predatory advantage. This reduction may also offer greater structural space and freedom of movement for the mesothorax, potentially optimizing flight capabilities (Grimaldi and Engel, 2005; Chapman, 2013). Instances of a highly developed mesothorax have been observed in certain Leistarchini, such as Mafulemesa Wygodzinsky, 1966 and Phryxobotrys. Future research into the natural history and functional morphology of these insects is essential to further elucidate the evolutionary mechanisms underlying pronotal shape.

Variations in protarsal segmentation within Leistarchini have been documented since early taxonomic works (e.g. Stål, 1874; Distant, 1903). Among extant genera, 22 possess a three‐segmented protarsus, 10 possess a non‐segmented (fused) protarsus and only Proguithera exhibits a two‐segmented condition. Villiers (1948, 1949, 1979) and Dispons (1970) argued that these differences defined higher‐level taxa, leading them to separate Orthungini (non‐segmented) from the remainder of the tribe (two‐ or three‐segmented). Wygodzinsky (1966) viewed protarsal segmentation as a plastic trait useful only for generic definitions. While Standring et al. (2023) suggested a single independent derivation of the non‐segmented condition, our results, while supporting a three‐segmented ancestral form, identify four independent reductions in segmentation. Three of these resulted in non‐segmented protarsi, whereas the ancestor of Guithera + Proguithera is inferred to have had a two‐segmented protarsus, implying that the fused state in Guithera represents a subsequent independent event.

Given that a non‐segmented protarsus has emerged in several unrelated lineages, the classification schemes of Villiers (1948, 1949, 1979) and Dispons (1970) must be rejected. The fusion of tarsomeres is generally correlated with the elongation and sclerotization of the protarsus and the shortening of the protibia, likely serving as an adaptation for increased predatory efficiency. However, the specific mechanisms driving these changes remain unclear due to a lack of detailed observations of Leistarchini predatory behaviour. Comprehensive studies on their basic natural history will be vital to understanding the evolutionary processes governing the protarsus.

Conclusions

In this study, we generated the first time‐calibrated phylogenetic framework for Leistarchini by integrating improved taxon sampling (covering ~65% extant genera of the tribe) with expanded molecular datasets (including USCOs, mitogenomes and nuclear rDNA), providing a new perspective on the evolution of this cosmopolitan group of thread‐legged bugs. Our results corroborate the monophyly of Leistarchini and identify five major clades, including a newly described genus Calliemesa gen. n., while revealing that the five most species‐rich genera (Nesita, Orthunga, Pleias, Ploiaria and Tinna) are paraphyletic or polyphyletic, highlighting the urgent need for systematic revisions. Biogeographic reconstructions suggest a Southeast Asian origin during the late Palaeocene, with early diversification driven by regional geological and climatic shifts, followed by multiple intercontinental dispersals since the middle Eocene that shaped the tribe's current global distribution. These events include the colonization of Madagascar via the Seychelles‐Mascarene Plateau and volcanic island chains, dispersals to Africa through boreotropical forests, and the expansion of the Ploiaria clade into the Australasian Region and the New World. Furthermore, our ancestral state reconstructions revealed complex, homoplastic evolutionary patterns for the shape of the posterior pronotal lobe and the segmentation of the protarsus, demonstrating that their traditional use in higher‐level classification warrants critical reassessment. This phylogenetic framework provides a robust foundation for future research, encouraging the collection of basic biological data across a broader range of species to enable in‐depth investigations into the trait evolution and biogeographic processes that allowed this specialized group of assassin bugs to achieve global success.

Conflict of interest

The authors declare that they have no conflict of interests.

Supporting information

Fig. S1. Phylogenetic inference of Leistarchini. Topology derived from the maximum‐likelihood (ML) analysis of the combined matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S2. Phylogenetic inference of Leistarchini. Topology derived from the maximum‐parsimony (MP) analysis of the combined matrix. Support values on nodes indicate MP bootstrap support values.

Fig. S3. Phylogenetic inference of Leistarchini. Topology derived from the ML analysis of the PCGRNA matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S4. Phylogenetic inference of Leistarchini. Topology derived from the Bayesian inference (BI) analysis of the PCGRNA matrix. Support values on nodes indicate PP.

Fig. S5. Phylogenetic inference of Leistarchini. Topology derived from the ML analysis of the PCG12RNA matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S6. Phylogenetic inference of Leistarchini. Topology derived from the BI analysis of the PCG12RNA matrix. Support values on nodes indicate PP.

Fig. S7. Phylogenetic inference of Leistarchini. Topology derived from the ML analysis of the USCO75NT matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S8. Phylogenetic inference of Leistarchini. Topology derived from the GHOST analysis of the USCO75NT matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S9. Phylogenetic inference of Leistarchini. Topology derived from the wASTRAL analysis of the USCO75NT matrix. Support values on nodes indicate ASTRAL support.

Fig. S10. Phylogenetic inference of Leistarchini. Topology derived from the ML analysis of the USCO75AA matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S11. Phylogenetic inference of Leistarchini. Topology derived from the GHOST analysis of the USCO75AA matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S12. Phylogenetic inference of Leistarchini. Topology derived from the wASTRAL analysis of the USCO75AA matrix. Support values on nodes indicate ASTRAL support.

Fig. S13. Phylogenetic inference of Leistarchini. Topology derived from the ML analysis of the USCO90NT matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S14. Phylogenetic inference of Leistarchini. Topology derived from the GHOST analysis of the USCO90NT matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S15. Phylogenetic inference of Leistarchini. Topology derived from the wASTRAL analysis of the USCO90NT matrix. Support values on nodes indicate ASTRAL support.

Fig. S16. Phylogenetic inference of Leistarchini. Topology derived from the ML analysis of the USCO90AA matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S17. Phylogenetic inference of Leistarchini. Topology derived from the GHOST analysis of the USCO90AA matrix. Support values on nodes indicate SH‐aLRT (left) and UFBoot2 (right).

Fig. S18. Phylogenetic inference of Leistarchini. Topology derived from the wASTRAL analysis of the USCO90AA matrix. Support values on nodes indicate ASTRAL support.

Table S1. Molecular sampling information of the present study.

Table S2. Molecular data used in the present study. “*” means sequences newly generated in this study; “—” means missing data.

Table S3. Detailed information on the datasets for phylogenetic analyses.

Table S4. The best‐fitting partitioning scheme for the combined matrix.

Table S5. The best‐fitting partitioning scheme for the PCGRNA matrix.

Table S6. The best‐fitting partitioning scheme for the PCG12RNA matrix.

Table S7. Estimated divergence times (in millions of years) for nodes in Fig. 3.

Table S8. Statistical results of BioGeoBEARS analysis of six biogeographical models.

Table S9. Summary of results from DEC + J analysis.

File S1. Taxonomic treatment of Calliemesa gen. n.

Data availability statement

The molecular data newly generated in this study have been deposited in GenBank. The original documents for phylogenetic analysis, divergence time estimation, ancestral area reconstruction and ancestral character state reconstruction of this study are deposited in the online database Zenodo (doi:10.5281/zenodo.17809081).