The Indo-Pacific legume genus Serianthes was recently placed in the Archidendron clade (sensu Koenen et al. 2020), a subclade of the mimosoid clade in subfamily Caesalpinioideae, which also includes Acacia, Archidendron, Archidendropsis, Falcataria, Pararchidendron, Paraserianthes and Wallaceodendron. Serianthes comprises ca. 18 species, five subspecies and two varieties that are characterised by bipinnately compound leaves with alternate sessile leaflets, branched axillary corymbiform panicles and woody indehiscent pods. Generic relationships, as well as species relationships within genera in the Archidendron clade, remain uncertain. While the sister relationship between Serianthes and the genus Falcataria is strongly supported by molecular data, the distinction between Serianthes and the monotypic genus Wallaceodendron has been questioned, based on their similar flower and fruit morphologies. We combined three gene-enriched hybrid capture DNA sequence datasets (generated from the 964 mimobaits v1 probe set, the expanded 997 mimobaits v2 probe set and the GoFlag angiosperm 408 probe set) and used their overlapping markers (77 loci of the target exonic and flanking regions) to test the monophyly of Serianthes and to investigate generic relationships within the Archidendron clade using 55 ingoid plus two outgroup taxa. We show that Serianthes is monophyletic, confirm the Serianthes + Falcataria sister relationship to Wallaceodendron and recognise this combined clade as the Serianthes clade within the Archidendron clade. We also evaluated the use of overlapping loci across datasets in combination with concordance analyses to test generic relationships and further investigate previously unresolved relationships across the wider ingoid clade. Concordance analysis revealed limited gene tree conflicts near the tips of the Archidendron clade, but increased discordance at the base of the clade, which could be attributed to rapid lineage divergence (radiation) and/or incomplete lineage sorting.

Archidendron clade, Fabaceae, mimosoid clade, monophyly, phylogenomics, targeted enrichment sequencing

In the recent re-classification of legume subfamilies (LPWG 2017), the former subfamily Mimosoideae that is nested within the re-circumscribed Caesalpinioideae, was informally recognised as the mimosoid clade. Within the mimosoid clade, phylogenetic analyses (e.g. Luckow et al. 2003; Bruneau et al. 2013; LPWG 2017; Koenen et al. 2020) consistently show that none of the tribes in the traditional tribal classification of Bentham (1844) are monophyletic. Recent phylogenomic analyses provided greater resolution across the mimosoid phylogeny (Koenen et al. 2020) and Caesalpinioideae as a whole (Ringelberg et al. 2022), establishing the basis for the recognition of a number of informally-named clades, including the large pantropical ingoid clade (Koenen et al. 2020) that contains all genera of tribe Ingeae plus Acacia Mill. and all its segregates, except Vachellia Wight & Arn. Morphologically, this clade is characterised by flowers with > (10–)30 stamens that are often fused into a tube (Fig. 1; Brown et al. 2008; Koenen et al. 2020).

Figure 1. 

Phylogeny of the mimosoid clade modified from Koenen et al. (2020), based on the mimosoid 964 nuclear dataset with the Archidendron clade highlighted in red. Clade names follow Koenen et al. (2020) with branches collapsed and represented by green triangles. The Stryphnodendron and Mimosa clades, taxa from which were used to root trees in this study, are highlighted in orange and purple, respectively and indicated with an asterisk.

Koenen et al. (2020) found that the Indomalayan/Australasian Archidendron clade falls within the ingoid clade (Fig. 1). The Indo-Pacific genus Serianthes Benth., which is the focus of this study, is included in the Archidendron clade, together with seven other genera (Koenen et al. 2020; Table 1, Fig. 2): Acacia s.s., Archidendron F. Muell., Archidendropsis I.C. Nielsen, Falcataria (I.C. Nielsen) Barneby & J.W. Grimes, Pararchidendron I.C. Nielsen, Paraserianthes I.C. Nielsen and Wallaceodendron Koord. The Archidendron clade is restricted to the Indomalayan and Australasian realms, with highest species diversity and endemism in Malesia, Papua New Guinea, New Caledonia and Australia (Table 1).

Figure 2. 

Morphology and relationships of the genera of the Archidendron clade, based on relationships recovered in our ASTRAL analysis. The colour scheme follows that in Fig. 7. Images are used with permission from Flickr: Acacia rostellifera (PC: Russell Cumming, HS: K000779891), Archidendron grandiflorum (PC: fl, le: Russell Cumming, HS: K000724398), Archidendron lucyi (PC: fr: Russell Cumming), Archidendropsis paivana subsp. balansae I.C. Nielsen (PC: fl: Benoît Henry), Archidendropsis streptocarpa (Fournier) I.C. Nielsen (PC: fr, le: Benoît Henry, HS: K000822329), Falcataria falcata (Photo credits [PC]: flower [fl]: JB Friday), Falcataria toona (PC: fruit [fr], leaf [le]: Russell Cumming, herbarium sheet [HS]: NY0149795, Serianthes nelsonii (PC: Else Demeulenaere, HS: US00689615), Pararchidendron pruinosum (PC: Russell Cumming, HS: K000759556), Paraserianthes lophantha (PC: fl: Eric Hunt; fr: Russell Cumming, le: Forest Starr and Kim Starr, HS: OBI126697), Wallaceodendron celebicum (PC: Plantaholic Sheila, HS: LSU00096994).

Nielsen (1992) and Nielsen et al. (1983, 1984a, 1984b) solidified the classification of the genera in the Archidendron clade and this generic system is still largely followed today. However, apart from Acacia s.s. (Brown et al. 2008), the monophyly of most of the ingoid genera in the Archidendron clade has not been tested with modern phylogenetic and phylogenomic analyses until recently (Brown et al. 2022; Ringelberg et al. 2022). Recent efforts to resolve phylogenetic relationships within the species-rich Archidendron clade have been hampered by a paucity of molecular data or incomplete taxon sampling in previous studies (Brown et al. 2008; Brown et al. 2011; Koenen et al. 2020). These uncertainties are compounded by nomenclatural instability (Barneby and Grimes 1996; Brown et al. 2008), lack of fertile herbarium specimens and morphological homoplasy (Fosberg 1960; Nielsen et al. 1984a; Koenen et al. 2020), as well as extensive geographic ranges for some species spanning the Indo-Pacific and Australia (Strijk et al. 2020). In the age of museomics and collection-based phylogenomics, the ability to sequence DNA from historical museum specimens (Zedane et al. 2016; Moreno-Aguilar et al. 2020; Renner et al. 2021) provides new opportunities to analyse phylogenetic relationships within the species-rich Archidendron clade by expanding taxon sampling geographically and including expert-identified specimens. Targeted enrichment sequencing (e.g. Hyb-Seq) can generate phylogenomic data by extracting DNA from small amounts of leaf tissue from archived herbarium specimens to build phylogenies with greatly enhanced gene and taxon representation (Bossert and Danforth 2018; Johnson et al. 2019; Escudero et al. 2020; Bateman et al. 2021; Eriksson et al. 2021).

Serianthes is a genus of tropical trees and shrubs distributed in the Indo-Pacific (Southeast Asia, the Pacific Islands and Australia). The genus was described by Bentham (1844) and has been revised by Fosberg (1960) and Kanis (1979, only the Malesian species). The most recent revision of Serianthes (Nielsen et al. 1984b) recognised 18 species, five subspecies and four varieties. The infrageneric classification of Nielsen et al. (1984a) recognised two subgenera, based on the basic unit of the inflorescence, subgenus Minahassae Fosberg with racemosely arranged pedunculate spikes and subgenus Serianthes with racemosely arranged pedunculate racemes, umbels or glomerules, while pod dehiscence and pod valve morphology were used to define sections within subgenus Serianthes. Although the monophyly of Serianthes has not been questioned, certain Albizia and Acacia taxa have been transferred to Serianthes in taxonomic revisions (Fosberg 1960).

Most Serianthes species are island endemics confined to small archipelagos in the Indo-Pacific Ocean. These endemic species face varying degrees of extinction threat caused by habitat loss and spread of invasive species. The IUCN Red List of Threatened Species lists 12 species of Serianthes, with three designated as critically endangered (IUCN 2021). In addition, Serianthes nelsonii Merr., endemic to the Mariana Islands, Guam and Rota, is listed as critically endangered by the U.S. Endangered Species Act (U.S. Fish and Wildlife Service 1987); only a single mature tree remains in Guam (Indigenous name [IN] for S. nelsonii on Guam: Håyun Lågu) and fewer than 50 individuals on Rota (IN: Tronkon Guåfi). As traditional uses and endemic languages are intrinsically connected to these endemic species, the islands’ biocultural diversity is also vulnerable to extinction. Indigenous island communities traditionally use Serianthes trees for building canoes, boats and meeting houses, as ethnomedicines, in agriculture and in handicrafts (Demeulenaere et al. 2021).

Nielsen et al. (1983, 1984a, 1984b) discussed the generic limits of Serianthes and the other Malesian, Australian and Pacific Ingeae, based on comparative morphology. Nielsen et al. (1983) considered Serianthes to be closely related to Falcataria (as Paraserianthes falcataria) and Wallaceodendron, based on their wood anatomy and postulated that they were more closely related to the group of Paraserianthes s.s., Archidendropsis and Pararchidendron than to Archidendron. The eophylls of Falcataria and Serianthes are bipinnate, while all other genera in the Archidendron clade have once-pinnately compound eophylls (Nielsen et al. 1983). In 1996, Barneby and Grimes established Falcataria as a new genus, based on Nielsen’s Paraserianthes section Falcataria, which included three species. This treatment was validated by the phylogenetic study of Brown et al. (2011), which concluded that Paraserianthes was paraphyletic and provided strong evidence for a well-supported Falcataria clade (incl. Falcataria falcata (L.) Greuter & R. Rankin, Falcataria pullenii (Verdc.) Gill. K. Br. and Falcataria toona (F.M. Bailey) Gill. K. Br., D.J. Murphy & Ladiges), distinct from Paraserianthes lophantha (Willd.) I.C. Nielsen of Nielsen’s Paraserianthes section Paraserianthes.

A recent phylogenomic study of the mimosoid clade included seven of the eight genera of the Archidendron clade (Koenen et al. 2020) and was the first study to include one of the 18 species of Serianthes, but it did not sample the monotypic Wallaceodendron. Here, we used data from targeted sequence capture to evaluate the monophyly of Serianthes by combining a large dataset for mimosoid legumes (Koenen et al. 2020) with a separate phylogenomic dataset for Serianthes and genera of the Archidendron clade.

Our study provides the first molecular evidence that Serianthes, as delineated by Nielsen et al. (1984a), is monophyletic (node 10) (Fig. 3; Tables 3, 4). Diagnostic features of Serianthes include bipinnately compound leaves with alternate sessile leaflet insertion, branched axillary corymbiform panicles and woody indehiscent pods (Fosberg 1960; Nielsen et al. 1984b), as opposed to bipinnately compound leaves with opposite leaflets and dehiscent pods in Wallaceodendron and Falcataria. The spiciform racemes of Wallaceodendron are solitary, while they are compound in Falcataria. Nielsen et al. (1984a) also commented on differences in pollen morphology between Serianthes and other genera in the Archidendron clade (Table 5), whereby the tectum of Wallaceodendron and Serianthes (except for subgenus Serianthes sect. Minahassae) is perforated by non-isometric channels, as compared to isodiametric channels in the other genera of the Archidendron clade (Nielsen et al. 1984a). Further research is needed to evaluate the taxonomic significance of pollen exine stratification across the Archidendron clade as a whole.

The close relationship amongst Serianthes, Falcataria, and Wallaceodendron as suggested by Nielsen et al. (1983), based on morphology, is corroborated by our phylogenomic analysis and this group is here referred to as the Serianthes clade (Fig. 3). The centre of diversity of the Serianthes clade is the Malesian and Papuasian region. Of this clade, Serianthes is the only genus with Pacific Island representatives, while Falcataria is the only genus occurring in Australia. Serianthes is the most widespread, most likely because of its indehiscent pods, which are dispersed via ocean currents (Demeulenaere and Ickert-Bond 2022).

The monophyly of Serianthes and the relationships within the Serianthes clade (nodes 8, 9 and 10; Fig. 3; Tables 3, 4) received full support, suggesting that the alignments were informative and provided a clear signal for these relationships. Nielsen et al. (1983, 1984a) postulated that Paraserianthes falcataria (now Falcataria falcata) is closely related to Serianthes, observing that the bracts of the two are large and concave and have barely distinguishable wood anatomy (Nielsen et al. 1983). Serianthes and Falcataria also share opposite and bipinnate seedling leaves, while mature leaves of Serianthes, in contrast to Falcataria, have alternate leaflet insertion (Table 5). This phylogenomic study provides the first evidence of two deeply-divergent and robustly-supported subclades within Serianthes, one comprising S. germanii, S. calycina, S. melanesica and S. nelsonii and the second S. minahassae, S. hooglandii, S. vitiensis and S. kanehirae. The placements of other Serianthes species within these subclades and how they correspond to the classification of subgenera and sections from Nielsen et al. (1984a) will require more complete taxon sampling.

Serianthes and Falcataria are sister genera in our phylogenomic study (Fig. 3), corroborating the results of Ringelberg et al. (2022), but not Brown et al. (2022). Both genera have alternate leaves, while Wallaceodendron has leaves that are spirally arranged (Sosef et al. 1998). Wallaceodendron was recovered as sister to Serianthes + Falcataria in our study (Fig. 3). Fosberg (1960) treated these three genera as distinct, noting that, while the flowers and the fruits of Wallaceodendron and Serianthes are very similar, Serianthes has flowers arranged in panicles, rather than racemes in Wallaceodendron, the pods of Serianthes are indehiscent, compared to the dehiscent pods of Wallaceodendron (tardily dehiscent) and Falcataria, and Wallaceodendron and Falcataria have strictly opposite leaflets as opposed to alternate leaflets in Serianthes (Fosberg 1960; Kanis 1979; Nielsen et al. 1983). This combination of inflorescence, leaf, and fruit dehiscence differences supports recognition of three distinct genera.

Our phylogeny suggests that Pararchidendron is nested within Archidendron, rendering Archidendron paraphyletic (Fig. 8) as also found by Brown et al. (2022). Many nodes along the backbone of the Archidendron clade remain poorly resolved (Fig. 8). The sister relationship of Paraserianthes and Acacia s.s. agrees with Brown et al. (2022) and Ringelberg et al. (2022). A recent phylogeny of legumes as a whole found full support for the sister relationship between the monophyletic Acacia s.s. and a clade containing Falcataria, Pararchidendron and Archidendron (Zhao et al. (2021), but this study did not include Paraserianthes. The position of Archidendropsis within the Archidendron clade remains uncertain, but the genus is not supported as monophyletic in our analyses (Fig. 8) – see Brown et al. (2022). Increased taxon sampling with phylogenomic data is needed to resolve the relationships of Archidendron, Archidendropsis and Pararchidendron.

Figure 8. 

Comparison of relationships of the Archidendron clade recovered by different authors. Colour schemes follow those in Fig. 2. The branches that are fully supported (either by PP = 1.00 or BS = 100%) are indicated by blue stars and discordant placements of genera are indicated by dashed lines. The following abbreviations are used: Archidendron grandi. = Archidendron grandiflorum, Archidendron luc. = Archidendron lucidum, Archidendron quoc. = Archidendron quocense, Archidendropsis g. = Archidendropsis granulosa and Archidendropsis x. = Archidendropsis xanthoxylon.

Conflicting topologies amongst sites and genes occurred where nodes showed low sCF and gCF values (nodes 4, 5 and 8 in Fig. 3; Table 4), which are indicative of discordant signals between gene trees. This was also shown by the short internode distances (expressed in coalescent units) at these branches in our phylogeny (Fig. 3). High gDF_P values coincided with short branches and likely indicate rapid lineage divergence (evolutionary radiation) and/or ILS (nodes 3, 5, 6, 7 and 8; Fig. 3; Table 4). This is consistent with the large putative hard polytomy in the ingoid clade discovered by Koenen et al. (2020), which likely represents a rapid radiation of a set of six or seven lineages. The Archidendron clade (node 4 in Fig. 3; Tables 3, 4) is one of the lineages derived from that putative hard polytomy along part of the backbone of the ingoid clade. The gene tree analysis provided high node support (PP = 1, BS = 100) and a high sCF value of 69.59% supporting the obtained tree topology at this node. The observed low discordance factor values (between 0 and 14.72%; Minh et al. 2020a; Thomas et al. 2021), however, indicated notable conflicts between gene concordance and discordance factors. The concordance analysis provided high gDF_P values of 78.67%, indicating that the gene trees lacked a clear signal (Minh et al. 2020a; Thomas et al. 2021). The fact that high PP and BS values coincided with low gCF values illustrates that classical node support measures, such as PP and BS, do not capture all aspects of variation in large phylogenomic datasets (Brower 2006, 2018; Thomas et al. 2021).

Sequence capture (Grover et al. 2012) provides a cost-effective way to generate hundreds of informative markers for plant phylogenomics that can be used across taxonomic scales (Zimmer and Wen 2015), including recent radiations of species and in intraspecific phylogeography (Nicholls et al. 2015). There is growing interest in combining data from different probe sets and, particularly, the merger of data from universal probe sets with data from clade-specific probes (e.g. Hendriks et al. 2021). Our study shows that the merger of data from different probe sets can yield enough overlapping loci to resolve intergeneric relationships. Our ingoid dataset increased resolution in the ingoid and Archidendron clades and generated a well-supported phylogeny, representing the evolution of unlinked markers across the genome. In many cases, the concordance analysis provided a new perspective on bootstrap values, local posterior probability support levels and polytomy tests, which may be inflated in large, concatenated alignments (Minh et al. 2020a; Thomas et al. 2021). Our analyses provide robust evidence for: (1) the monophyly of Serianthes and two main lineages within the genus; (2) the Serianthes clade, which sets the stage for future biogeographic analysis of this clade and highlights the close sister relationship between Wallaceodendron and Serianthes + Falcataria; (3) rapid radiations across the backbones of the ingoid and Archidendron clades, which may be difficult to resolve without extensive genomic data; the concordance analysis clarified the interpretation of phylogenetic relationships; in particular, we found limited gene conflicts near the tips of the Archidendron clade, but an increase in discordance at the base of the clade; and (4) the utility of the polytomy test to further evaluate if gene tree discordance affects node support values. Continued sampling and sequencing of Serianthes species and other genera in the Archidendron clade are necessary to fully evaluate the generic delimitation and relationships within the Archidendron clade.

Mesuláng, Si Yu`os Ma`åse` to the Republic of Palau and Guåhan (Guam) for their hospitality. We thank all the people who assisted with acquiring permits for this project and for permission to conduct research on each Island. Special thanks to the Belau National Museum for their assistance with organising the field trips, to Sholeh Hanser and Naito Soaladoab for accompanying us with wonderful stories and enthusiasm and to Dr. Ann Kitalong for providing herbarium space to dry our specimens. We thank Gillian Brown, Daniel Murphy and Warren Cardinal-McTeague for their constructive criticism of our manuscript and the editors of the Advances in Legume Systematics 14, Colin Hughes, Luciano de Queiroz and Gwilym Lewis, for coordinating this Special Issue. We thank U.S. Fish and Wildlife for financial support under grant F16AP00679. The GoFlag probe set was developed with support from the United States National Science Foundation (DEB-1541506). The following permits were obtained to conduct this research: Guam (FWS/RI/AES/Recovery/GNWR-8 and FWS/RI/AES/Recovery/TE_64600C_0 (U.S. Fish and Wildlife Permits); Biological Research Permit from the Department of the Air Force Headquarters, 36^th Wing (PACAF), Andersen Air Force Base, Guam; Scientific Research Licenses 02590-17 and 03881-18), Palau (Memorandum of Understanding between the University of Guam, Center for Island Sustainability and the Ministry of Natural Resources, Environment and Tourism (MNRET), Government of Palau (2018)). Tom Schils is indebted to the University of Guam for supporting studies to document and conserve the natural heritage of Guam and the larger Micronesian Region. Part of this research was supported by the U.S. National Science Foundation (NSF; nsf.gov) under grant number OIA-1946352 awarded to the University of Guam. Any opinions, findings and conclusions or recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of NSF. The funders had no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript. This work was further supported by Swiss National Science Foundation through an Early Postdoc Mobility fellowship (grant number P2ZHP3_199693) to Erik Koenen and grants 310003A_156140 and 31003A_182453/1 (to Colin Hughes) which supported Jens Ringelberg.