Phylogenomics of peacock spiders and their kin (Salticidae: Maratus), with implications for the evolution of male courtship displays
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Biological Journal of the Linnean Society, 2021, XX, 1–24. With 6 figures. Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 Phylogenomics of peacock spiders and their kin (Salticidae: Maratus), with implications for the evolution of male courtship displays MADELINE B. GIRARD1, DAMIAN O. ELIAS1,*, , GUILHERME AZEVEDO2, KE BI3, MICHAEL M. KASUMOVIC4, , JULIANNE M. WALDOCK5, ERICA BREE ROSENBLUM1 and MARSHAL HEDIN2 1 Department of Environmental Science, Policy and Management, University of California, Berkeley, Berkeley, CA 94720-3114, USA 2 Department of Biology, San Diego State University, San Diego, CA 92182-4614, USA 3 Museum of Vertebrate Zoology, University of California, Berkeley, Berkeley, CA 94720-3160, USA 4 Ecology & Evolution Research Centre, School of Biological, Earth & Environmental Sciences, UNSW, Sydney, 2052, NSW, Australia 5 Collections and Research, Western Australian Museum, 49 Kew Street, Welshpool, 6106, Western Australia, Australia Received 7 July 2020; revised 16 September 2020; accepted for publication 22 September 2020 Understanding diversity has been a pursuit in evolutionary biology since its inception. A challenge arises when sexual selection has played a role in diversification. Questions of what constitutes a ‘species’, homoplasy vs. synapomorphy, and whether sexually selected traits show phylogenetic signal have hampered work on many systems. Peacock spiders are famous for sexually selected male courtship dances and peacock-like abdominal ornamentation. This lineage of jumping spiders currently includes over 90 species classified into two genera, Maratus and Saratus. Most Maratus species have been placed into groups based on secondary sexual characters, but evolutionary relationships remain unresolved. Here we assess relationships in peacock spiders using phylogenomic data (ultraconserved elements and RAD-sequencing). Analyses reveal that Maratus and the related genus Saitis are paraphyletic. Many, but not all, morphological groups within a ‘core Maratus’ clade are recovered as genetic clades but we find evidence for undocumented speciation. Based on original observations of male courtship, our comparative analyses suggest that courtship behaviour and peacock-like abdominal ornamentation have evolved sequentially, with some traits inherited from ancestors and others evolving repeatedly and independently from ‘simple’ forms. Our results have important implications for the taxonomy of these spiders, and provide a much-needed evolutionary framework for comparative studies of the evolution of sexual signal characters. ADDITIONAL KEYWORDS: Australia – character evolution – multimodal signals – rapid evolution – sexual selection – taxonomy – ultraconserved elements. INTRODUCTION emerged from many of these studies is that sexually selected traits sometimes show low phylogenetic signal Sexual selection has driven the evolution of a (Kusmierski et al., 1997; Gleason & Ritchie, 1998; spectacular diversity of colours, sounds, smells, shapes Ohmer et al., 2009; Puniamoorthy et al., 2009; Hosner and forms across the animal kingdom, from darters & Moyle, 2012; Hebets et al., 2013; Owens et al., 2020), (Hulse et al., 2020) to bower birds (Frith & Frith, 2004) often in the traits that are the most obvious to human to mantis shrimps (Porter et al., 2010) to treehoppers (Rodriguez et al., 2004). One common theme that has observers (i.e. bower bird plumage and bowers, hind wings in swallowtails, tyrant flycatcher plumage, frog calls, etc.). It thus follows that for comparative *Corresponding author. E-mail: doelias@berkeley.edu studies, particularly those focused on the evolution © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24 1
2 M. B. GIRARD ET AL. of potentially homoplastic sexually selected traits, 2020). Sexual dimorphism is profound in this group, Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 species hypotheses and phylogenetic relationships with cryptic adult females contrasting with highly should be based on independent evidence (e.g. decorated adult males. Male Maratus are now famous phylogenomic data). for their elaborate visual and vibrational courtship Peacock spiders of the genus Maratus Karsch displays (Girard et al., 2011; Otto & Hill, 2011; Girard 1878 comprise a diverse clade of jumping spiders & Endler, 2014; Girard et al., 2018; Otto & Hill, 2019a; (Fig. 1; Supporting Information, Fig. S1) distributed Wilts et al., 2020), facilitated by conspicuous behaviour predominately in eastern and western Australia, with in which they raise their often colourful abdomens over 90 described species (Otto & Hill, 2019a; World vertically and along with elongated, brush-adorned, Spider Catalog, 2020) and active on-going species third legs perform a vigorous courtship display. Some discovery (e.g. Schubert & Whyte, 2019; Schubert, male peacock spiders have cuticular flaps that wrap Figure 1. Peacock spiders in courtship posture and variation in fan morphology. A, Maratus sarahae – elliptical fan. B, M. calcitrans – no fan. C, M. mungaich – elliptical fan. D, M. pardus – elliptical fan. E, M. amabilis – elliptical fan. F, M. vespa –posterior lobed fan. G, Saratus hesperus – no fan. H, M. flavus – no fan. I, M. rainbowi – round fan. J, M. plumosus – posterior lobed fan. K, M. harrisi – lobed fan. L, M. anomalus – no fan. M, M. jactatus – round fan. N, M. volans – elliptical fan. O, M. linnaei – no fan. P, M. clupeatus – elliptical fan. Q, M. cinerus – no fan. R, M. personatus – no fan. S, M. elephans – elliptical fan. T, M. madelineae – elliptical fan. U, M. proszynskii – no fan. V, M. australis – lobed fan. W, M. spicatus – no fan. X, M. vespertilio – lobed fan. Y, M. pavonis – round fan. Z, M. speciosus – no fan. Latitude and longitude data for all specimens can be found in Supporting Information, Table S1. © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
PEACOCK SPIDER PHYLOGENOMICS 3 beneath the abdomen (Dunn, 1947; Waldock, 1993; The genus Maratus included fewer than ten Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 Hill, 2009), unfurled during the male courtship display, described species in 2008 (Waldock, 2008). The revealing remarkable patterns of scale pigmentation description of new Maratus species has exploded over and structural colours (Stavenga et al., 2016; Hsiung recent years, as interest in these spiders has increased, et al., 2017, 2019; Wilts et al., 2020; Fig. 1; Fig. S1; Table and macrophotography and video of live specimens S1). The extensive radiation of peacock spider courtship has facilitated the discovery of complex male courtship ornaments and displays is perhaps comparable to ornaments and behaviours (Otto & Hill, 2019a; that of the better-known birds of paradise (Cooper & Schubert, 2020). Most peacock spider species have Forshaw, 1977; Nunn & Cracraft, 1996; Ligon et al., been placed into complexes of related species (species 2018), but these spiders have remained obscure groups) based on male morphology and/or display until recently because of their small size. Despite characters (Otto & Hill, 2019a). As of early 2020, tremendous sexual signal diversity seen in this group, 16 species groups have been recognized (Table 1). the evolutionary patterns and processes of ornament Relationships within and among groups have not been evolution remain mostly unstudied due to a lack of independently tested using other types of evidence, formal phylogenetic study and resolution. and resolution within these groups is mostly lacking. Previous molecular systematics research has placed Some species are not currently placed into species Maratus within a clade of mostly Australasian salticids groups (Table 1) and it remains unclear whether these in the Saitis Clade (Fig. 2A; Zhang & Maddison, 2013), taxa are truly phylogenetically isolated, or are nested within the tribe Euophryini (Maddison, 2015; Zhang & cryptically within described groups. Also, some species Maddison, 2015). Zhang & Maddison (2015) considered appear to share morphological features of multiple the multiple genera in the Saitis Clade (Maratus, groups (Schubert, 2020), blurring species group limits. Saratus Otto & Hill 2017a, Hypoblemum Peckham & Similar to arguments above for generic-level diagnoses Peckham 1885, Jotus Koch 1881, Saitis Simon 1876, and because species groups are defined mostly on male etc.) as so closely related as to be possibly congeneric. courtship morphologies and behaviours (but see Baehr Conversely, other authors have maintained the above & Whyte, 2016), current phylogenetic hypotheses do taxa as separate genera (Otto & Hill, 2012a, 2019a; not present an independent framework to understand Otto & Hill, 2017a; Prószyński et al., 2018; Baehr potential homoplasy in these same character systems et al., 2019). A hypothesized sub-lineage in this clade (Maddison & Hedin, 2003). includes the ‘Maratus group’ (Otto & Hill, 2012a, Of the now more than 90 accepted species in Maratus, 2019a; Otto et al., 2019), including Maratus, Saratus, no species hypotheses have been independently tested ‘Lycidas’ (now mostly synonymized with Maratus) and using molecular evidence. There are many species Hypoblemum. Otto et al. (2019) suggested that the two known only from single (type) locations, and these could species of Hypoblemum are sister to peacock spiders, represent divergent populations of otherwise known including Maratus and Saratus (Fig. 2B). Saratus is species (i.e. geographical variation). Some ‘varieties’ monotypic, with male and female genitalia that differ have been described within wide-ranging species [e.g. from Maratus, but with somatic morphology, male M. pavonis (Dunn 1947); Hill & Otto, 2011; Otto & coloration and male display behaviours otherwise Hill, 2012a; Baehr & Whyte, 2016], possibly indicating similar to the latter (Otto & Hill, 2017a). undetected speciation. Some of these wide-ranging Except for the Sanger-sequencing-based studies species have conspicuously disjunct distributions of Zhang & Maddison (2013, 2015), relationships with populations along the eastern seaboard and in within the Saitis Clade have never been tested via south-western Australia, but populations are absent phylogenetic analysis, although authors have inferred from the more xeric temperate arid zone [e.g. eastern relationships based on character argumentation vs. western M. pavonis, M. vespertilio (Simon 1901)]. (Fig. 2B). A unique aspect of generic-level diagnoses Finally, researchers have discussed possible evidence in this clade is that many are based mostly or entirely for gene flow across species boundaries (Otto & Hill, on male courtship morphology and behaviour (Otto & 2014b, 2016a; Schubert, 2020), again blurring species Hill, 2012a, 2017a; Prószyński et al., 2018; Baehr et al., limits, and testable using genetic data. 2019; Otto et al., 2019), rather than the traditional We collected ultraconserved element (UCE) and (for salticid systematics) male and/or female genitalic RAD-sequencing (RAD-seq) phylogenomic data to morphology (see Zhang & Maddison, 2015). It address relationships in the Saitis Clade, emphasizing remains to be seen whether independent evidence Maratus relationships. Our sample includes type (e.g. molecular evidence) supports taxa based on male species for core genera [Jotus auripes L. Koch 1881, courtship characters, or rather, if such characters Saitis barbipes (Simon 1868), Hypoblemum griseum might be subject to higher levels of homoplasy, calling (Keyserling 1882), ‘Lycidas’ anomalus (=Maratus into question current generic-level diagnostic traits anomalus) (Karsch 1878), Saratus hesperus Otto & (see also Zhang & Maddison, 2015). Hill 2017a, and Maratus amabilis (Karsch 1878)]. For © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
4 M. B. GIRARD ET AL. Maileus cf. fuscus (Malaysia)_JXZ098 A) Sanger sequence data - Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 Saitis barbipes (SPAIN)_JXZ147 Zhang & Maddison 2013, fig 1 cf Saitis sp. (WA)_JXZ091 Saitis mutans Maratus rainbowi (WA)_JXZ160 Maratus sp. (soAU)_JXZ159 Lycidas cf. karschi (NSW)_JXZ157 Hypoblemmum scutulatum Hypoblemum cf. albovittatum (soAU)_JXZ164 Hypoblemmum scutulatum? Hypoblemmum sp. (NSW)_JXZ085 Hypoblemmum scutulatum Lycidas cf. griseus (NSW)_JXZ092 Hypoblemmum griseum? Lycidas cf. vittatus (QL)_JXZ146 XZ1 uncertain cf. Jotus sp. (QL)_JXZ155 Prostheclina sp. (NSW)__JXZ095 Jotus auripes (QL)_JXZ090 Servaea vestita B) Compilation - multiple sources, see legend Saitis Saitis (5 species, S. barbipes type) image ima iim m maage male leg III long, w/ “bottle brush” slender unmodifed leg I flexibile pedicel MF elevation of abdomen Hypoblemum scutulatum Hypoblemum griseum heavy outer embolar ring image iima im m maage “Maratus Hypoblemum group” ?? Maratus M abdomen with image iima im m maage dorsal plate or scute Barraina (monotypic, B. anfracta type) M palp, Saratus hesperus F epigynum unique Maileus (monotypic, M. fuscus type) Maratus Jotus (8 species, Jotus J. auripes type) carapace with lateral stripes image ima iim m ma mag age abdomen with lateral bands male leg I setal brush combined conductor + embolus of male palp metatarsus leg I fringe in males Prostheclina (7 species, P. pallida type) Figure 2. Overview of the phylogenetic structure of the Saitis clade. A, results of Zhang & Maddison (2013: fig. 1), based on maximum-likelihood analysis of four concatenated genes. Annotation of uncertain taxon names reflects UCE by-catch phylogenetic results (see Supporting Information, Fig. S3). B, hypothesized phylogenetic structure of the Saitis Clade from multiple sources (Richardson & Żabka, 2007; Otto & Hill, 2012a, 2016a, 2017a; Prószyński et al., 2018; Otto et al., 2019; Baehr et al., 2019), with partial list of supporting characters for genera and generic groupings. © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
Table 1. Maratus species groups, following Otto and Hill (2019a), and results presented here Groups (no. of species) Otto & Hill (2019a) Otto & Hill (2019a) species (taxa Taxa sampled in this study RAD results characters; see also Baehr & described after 2019 denoted as such) Whyte (2016) fimbriatus group (4) Wheel-rim-shaped embolus; M. fimbriatus Otto & Hill 2016a, M. licunxini M. speculifer, M. ‘flame’, M. ‘carmel’ legs I extended in courtship Baehr & Whyte 2016, M. speculifer (Simon 1909), M. volpei Schubert 2020 chrysomelas group (3) Apex of the embolus with con- M. chrysomelas (Simon 1909), M. kiwirrkurra M. chrysomelas, M. nigromaculatus vergent inner and outer Baehr & Whyte 2016, M. nigromaculatus edge (Keyserling 1883) spicatus group (4) Conical proximal tegulum, M. nimbus Otto & Hill 2017b, M. purcellae M. nimbus, M. purcellae, M. robinsoni, spicatus group intermixed etc.; probably related to Otto & Hill 2013a, M. robinsoni Otto & Hill M. spicatus with chrysomelas group chrysomelas group 2012a, M. spicatus Otto & Hill 2012a pavonis group (8) Target on fan, sometimes ob- M. leo Otto & Hill 2014b, M. literatus Otto & M. cf. leo, M. leo, M. literatus, scured; zigzag ridges on em- Hill 2014b, M. maritimus Otto & Hill 2014b, M. maritimus, M. pavonis, bolic disc M. montanus Otto & Hill 2014b, M. pavonis M. rainbowi, M. watagansi (Dunn 1947), M. rainbowi (Roewer 1951), M. sylvestris Otto & Hill 2019c, M. watagansi Otto & Hill 2013b anomalus group (11) Bifurcated apex of the outer M. albus Otto & Hill 2016b, M. anomalus M. albus, M. anomalus, M. aurantius, Includes M. sceletus from ring of embolus (Karsch 1878), M. aurantius Otto & Hill M. cinereus, M. michaelorum, calcitrans group 2017a, M. cinereus Otto & Hill 2017a, M. neptunus, M. cf. neptunus, M. inaquosus Schubert 2020, M. julianneae M. vultus Baehr & Whyte 2016, M. kochi (Żabka 1987), M. lentus Otto & Hill 2017a, M. michaelorum Baehr & Whyte 2016, M. neptunus Otto & Hill 2017a, M. vultus Otto & Hill 2016b vespertilio group (2) M. sapphirus Otto & Hill 2017b, M. vespertilio M. vespertilio (Simon 1901) velutinus group (2) Squamous setae on abdomen M. proszynskii Waldock 2015, M. velutinus M. proszynskii, M. velutinus Otto & Hill 2012a harrisi group (3) Lobate flaps on fan M. constellatus Schubert 2020, M. harrisi Otto M. harrisi © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24 & Hill 2011, M. lobatus Otto & Hill 2016b calcitrans group (7) Males with asymmetric dis- M. calcitrans Otto & Hill 2012b, M. digitatus M. calcitrans, M. digitatus, M. eliasi, M. plumosus, M. cf. plumosus play, most inflating spin- Otto & Hill 2012b, M. eliasi Baehr & Whyte M. jactatus, M. ottoi, M. plumosus, near tasmanicus group, nerets 2016, M. jactatus Otto & Hill 2015a, M. ottoi M. cf. plumosus, M. sceletus M. sceletus in anomalus Baehr & Whyte 2016, M. plumosus Otto & group Hill 2013b, M. sceletus Otto & Hill 2015a volans group (3) Large fringed fans with dis- M. elephans Otto & Hill 2015b, M. pardus M. elephans, M. pardus, M. volans M. pardus in expanded tinctive figures Otto & Hill 2014c, M. volans (O. Pickard- mungaich (*from Cambridge 1874) south-west Australia) PEACOCK SPIDER PHYLOGENOMICS 5 Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021
6 Table 1. Continued Groups (no. of species) Otto & Hill (2019a) Otto & Hill (2019a) species (taxa Taxa sampled in this study RAD results characters; see also Baehr & described after 2019 denoted as such) Whyte (2016) tasmanicus group (3) Triangular fan with lobes, M. australis Otto & Hill 2016b, M occasus M. australis, M. tasmanicus M. australis in expanded black spots Schubert 2019b, M. tasmanicus Otto & Hill mungaich (*from 2013b south-west Australia) mungaich group (10) Wide fans with red scales; M. avibus Otto & Hill 2014a, M. bubo Otto M. avibus, M. caeruleus, M. madelineae, Not monophyletic, part of ex- south-west Australia & Hill 2016b, M. caeruleus Waldock 2013, M. mungaich, M. sarahae panded mungaich M. B. GIRARD ET AL. M. gemmifer Otto & Hill 2017c, M. hortorum Waldock 2014, M. karrie Waldock 2013, M. madelineae Waldock 2014, M. melindae Waldock 2013, M. mungaich Waldock 1995, M. sarahae Waldock 2013 flavus group (4) south-west Australia endemic, M. boranup Otto & Hill 2018a, M. felinus M. flavus Part of expanded mungaich related to linnaei and vespa Schubert 2019, M. flavus Otto & Hill 2018a, groups? M. suae Schubert 2020 linnaei group (4) Rotating abdomen during M. cuspis Otto & Hill 2019c, M. electricus Otto M. linnaei Part of expanded mungaich courtship; south-west Aus- & Hill 2017c, M. laurenae Schubert 2020, tralia; related to vespa M. linnaei Waldock 2008 group? personatus group (2) south-west Australia M. banyowla Otto & Hill 2019b, M. personatus M. personatus Part of expanded mungaich Otto & Hill 2015c vespa group (10) Lobate flaps with lateral orna- M. aquilus Schubert 2019, M. azureus Schu- M. vespa Part of expanded mungaich ments; south-west Australia, bert 2020, M. combustus Schubert 2019, related to linnaei group? M. cristatus Otto & Hill 2017c, M. icarus Otto & Hill 2019c, M. noggerup Schu- bert 2020, M. tortus Otto & Hill 2018b, M. unicup Otto & Hill 2018b, M. vespa Otto & Hill 2016b Taxa not placed into spe- M. amabilis (Karsch 1878), M. clupeatus Otto M. amabilis, M. clupeatus, M. speciosus M. clupeatus in ex- cies groups (6) & Hill 2014d, M. sagittus Schubert & Whyte panded mugaich (*from 2019, M. speciosus (O. Pickard-Cambridge, south-west Australia) 1874), M. tessellatus (south-west Australia) Otto & Hill 2016b, M. trigonus (south-west Australia) Otto & Hill 2017c © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24 Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021
PEACOCK SPIDER PHYLOGENOMICS 7 peacock spiders (Maratus and Saratus), we generated (Supporting Information, Table S1). We also included Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 data for a representative sample of species, including the more divergent genus Servaea Simon 1888 to root species from all 16 currently recognized species phylogenies (data shared by W. Maddison), following groups (Table 1). Our phylogenomic results reveal the results of Zhang & Maddison (2013, 2015). that both Saitis and Maratus, as currently recognized, We lacked samples of Barraina Richardson 2013, are paraphyletic taxa. Most described species are Maileus Peckham & Peckham 1907 and Prostheclina recovered as exclusive genetic groupings, but we find Keyserling 1882, all hypothesized members of the evidence for undetected speciation and we present Saitis Clade (Zhang & Maddison (2013, 2015). We used data for three probable new species. Phylogenetic the MYbaits Spider v.1 kit (Arbor Biosciences, Ann comparative analysis of male visual and vibratory Arbor, MI, USA; Kulkarni et al., 2020) to capture UCE signals reveals a combination of synapomorphic and loci, using methods of library preparation as in recent homoplastic signal. Overall, our research establishes a publications (Starrett et al., 2017; Hedin et al., 2019). phylogenetic framework for future studies of peacock Sequencing was conducted on an Illumina HiSeq400 at spider morphology, behaviour and evolution. the U.C. Davis Genome Center with 150-bp paired-end reads. Sequence reads were trimmed and assembled using TRINITY v.2.1.1. (Grabherr et al., 2011) using default settings (trimmomatic = full_cleanup, MATERIALS AND METHODS kmer = 25), then processed using Phyluce (Faircloth, For a majority of specimens, genomic DNA was 2016). Assembled contigs were matched to probes extracted from legs or whole spiders using a QIAamp using standard minimum coverage and minimum DNA mini kit (Qiagen, Valencia, CA, USA). Saitis identity values (80_80). UCE loci were aligned with barbipes samples used in UCE experiments were MAFFT (Katoh & Standley, 2013) and trimmed with extracted using a Qiagen DNeasy blood and tissue kit. Gblocks (Castresana, 2000; Talavera & Castresana, 2007) using strict settings (--b1 0.5 --b2 0.85 --b3 4 --b4 8). Resulting Phyluce alignments with at least Taxonomic correction 70% sample occupancy were imported into Geneious The name Maratus rainbowi (Roewer, 1951) is used 11.0.4 (Biomatters, Auckland, New Zealand), where all in this study, in accordance with the World Spider individual alignments were visually inspected. Catalogue (2020), replacing the use of Maratus Maximum-likelihood (ML) phylogenetic analyses splendens common in other catalogues (Otto & Hill, were conducted using IQ-TREE v.2.0-rc2 (Nguyen 2019a). The name Attus splendens was assigned et al., 2015). Initial partitions corresponded to to a North American salticid species by Peckham individual loci, and ModelFinder (Kalyaanamoorthy & Peckham (1883), which is now considered to be a et al., 2017) was then used to find best-fit models junior synonym of Habronattus decorus (Blackwall, and merge partitions (-s -p -m TESTMERGE 1846). Rainbow (1896) also established a species -rcluster 10); the relaxed hierarchical clustering that he called Attus splendens for a new species from algorithm (Lanfear et al., 2014) was used to reduce Australia, overlooking the prior usage by Peckham & computational burden. Support was assessed via 1000 Peckham (1883). Attus splendens Rainbow, 1896 was ultrafast bootstrap replicates (Hoang et al., 2018). transferred to Saitis by Simon (1901) and to Maratus An SVDQuartets analysis (Chifman & Kubatko, by Żabka (1991). In the meantime, Roewer (1951) 2014; Chifman & Kubatko, 2015) was conducted on recognized that Attus splendens Peckham & Peckham, a concatenated matrix using PAUP* 4.0a (Swofford, 1883 and Attus splendens Rainbow, 1896 were primary 1988), implementing the multispecies coalescent tree homonyms, and provided the replacement name model with exhaustive quartets sampling and 1000 Saitis rainbowi Roewer, 1951. With the transfer of all bootstrap replicates. Australian species of Saitis to Maratus, this species is Using IQ-TREE 2 we calculated gene (gCF) and correctly known as Maratus rainbowi (Roewer, 1951). site concordance (sCF) factors. For every branch of a This usage is in accordance with the International reference tree, gCF can be defined as the percentage of Code of Zoological Nomenclature (1999). ‘decisive’ gene trees containing that branch, while sCF can be defined as the percentage of decisive sites (in an alignment) supporting a branch (Minh et al., 2020). UCE data The latter support metric is particularly useful when Specimens of Jotus auripes, Saitis barbipes, Saitis individual gene trees are uncertain. Because ML and mutans Otto & Hill 2012a, Saitis virgatus Otto & Hill SVDQuartets UCE tree topologies agreed (see below), 2012a, Hypoblemum griseum, H. scutulatum L. Koch we used the partitioned ML tree as a reference for CF 1881, Saratus hesperus, and a subsample of seven calculations, with individual gene trees calculated as Maratus species were included in UCE experiments default in IQ-TREE 2 (-S command). © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
8 M. B. GIRARD ET AL. UCE by-catch taxa (rather than all taxa), this enabled us to extract Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 We assembled ribosomal 28S and mitochondrial more loci per clade. 16S_ND1 ‘by-catch’ using standard BLAST searches in Geneious 11.0.4 (Biomatters). Specifically, we RAD analysis – all taxa used published sequences to query UCE TRINITY assemblies using blastn (maximum e-value of 1e-5). A custom script was used to process RAD data (available We then stitched together UCE by-catch data with at https://github.com/CGRL-QB3-UCBerkeley/RAD), published sequences from Zhang & Maddison (2013, calling various external programs. This pipeline has 2015), aligned matrices manually, and conducted been used in several other recent publications (Krohn ML analyses using IQ-TREE 2 (-s, -B 1000, -bnni). et al., 2018; Maas et al., 2018; Klonoski et al., 2019). Our primary interest here was to potentially resolve Exact duplicates were removed using Super Deduper uncertain taxon identities from Zhang & Maddison (https://github.com/dstreett/Super-Deduper). Raw (2013) (e.g. see Fig. 2A), while also including Maileus reads were filtered using Cutadapt (Martin 2011) and Prostheclina, not sampled in other matrices. and Trimmomatic (Bolger et al., 2014) to trim adapter contaminations and low-quality reads. The resulting cleaned forward reads of each individual were first clustered using Cd-hit (Li & Godzik, 2006; Fu et al., RAD data collection 2012), and only clusters with at least two reads Specimens of Jotus auripes, Saitis mutans, S. supported were kept. To remove potential paralogues, virgatus, Hypoblemum griseum, H. scutulatum and Blastn (Altschul et al., 1990) was used to compare Saratus hesperus were included as ‘outgroup taxa’ clustered loci against themselves, and remove any (Supporting Information, Table S1). Lacking Servaea locus that matched to a locus other than itself. The RAD data, we used UCE topology results to root RAD resulting RAD loci from each individual were then trees. Specimens representing 48 described Maratus combined for all individuals and the resulting marker species were included, based on 2013–2015 collections sets served as a master reference. We then used Blastn from eastern and western Australia (ACT, NSW, QLD, (Altschul et al., 1990) to compare markers from each SA, TAS, VIC, WA; Table 1; Appendix S1). One to three individual to the master reference and only kept individuals per species were sampled per location, and those that had unique hits. These uniquely matched for more broadly distributed species we attempted to markers from each individual served as a reference sample populations spanning known geographical for that individual. Cleaned sequence reads from each ranges. Our taxon sample included five novel forms individual were aligned to its own reference using (Maratus cf. neptunus, M. ‘carmel’, M. ‘flame’, M. cf. Novoalign (http://www.novocraft.com) and reads that leo, M. cf. plumosus), representing potentially new mapped uniquely to the reference were kept. Picard species. (http://picard.sourceforge.net) was used to add read We generated RAD libraries using the protocol of groups and GATK (McKenna et al., 2010) to perform Ali et al. (2016) without completing the targeted bait realignment around indels. SAMtools/BCFtools and capture step and using Pippin Prep (Sage Science, ‘vcfutils.pl vcf2fq’ implemented in SAMTools (Li et al., Beverley, MA, USA) instead of beads to size-select 2009) were used to generate individual consensus fragments between 250 and 600 bp. Each of two sequences by calling genotypes and incorporating 96-sample libraries was sequenced on an Illumina ambiguous sites. We kept a consensus base only when HiSeq 4000 lane at the U.C. Davis Genome Center its depth was at least 3× or above and retained loci with 150-bp paired-end reads. Raw fastq reads were that contained no more than 80% missing data. We de-multiplexed allowing one barcode mismatch. also masked sites within 5-bp windows around indels. De-multiplexed reads were removed if the expected We converted the resulting consensus fastq sequence cut site (also one mismatch allowed) was not found at file to fasta format using Seqtk (https://github.com/ the beginning of the 5′-end of sequences. lh3/seqtk). Using MAFFT (Katoh & Standley, 2013), Two separate sets of analytical pipelines and final filtered loci from each individual were aligned by downstream analyses were conducted on RAD data: comparing against the master reference. Ambiguously (1) a custom pipeline was used to extract a set of loci aligned regions were then trimmed using Trimal conserved across all taxa – this locus set was used to (Capella-Gutierrez et al., 2009). To avoid excess reconstruct a phylogeny for Maratus and ‘outgroups’; missing data, we removed alignments if more than (2) based on the ‘all taxa’ RAD results, the ipyrad 90% missing data were present in at least 30% of pipeline (Eaton et al., 2017) was used to assemble samples. locus sets for well-supported species groups (or clades Using a subsample of the entire available sample comprising multiple species groups). Because ipyrad (152 specimens), we performed several all-taxa analyses were focused on more closely related sets of phylogenomic analyses. Species trees were inferred © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
PEACOCK SPIDER PHYLOGENOMICS 9 using the multispecies coalescent model implemented S2. Character scorings for Saitis barbipes were Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 in SVDQuartets. Originally developed for unlinked extracted from other sources (Hill, 2009; https:// single nucleotide polymorphism (SNP) data, where www.youtube.com/watch?v=wQT2bHTOdwo); some each site has an independent genealogy drawn from scorings for Hypoblemum scutulatum were taken from a coalescent model, the method has also been shown the courtship description of Otto et al. (2019). The to perform well for linked SNP data (Chifman & characters presented here are not intended to capture Kubatko, 2014), as used here. SVDQuartets was the full diversity or complexity of courtship characters used to construct a ‘lineage’ tree, relating individual seen within Maratus, but rather represent a set of sequences. We also used a taxon partition to assign scoreable traits with hypothesized deeper homology individual samples to species (partition = species); for across all genera included. M. pavonis, which includes multiple well-supported sub-lineages (see Results below), each sub-lineage here was treated as a ‘species’. For both lineage and Character evolution partitioned analyses we evaluated all possible quartets For tree-based character evolution analyses we used (evalq = all, quartet assembly algorithm = QFM), pruned UCE and all-taxa RAD matrices, removing and conducted a non-parametric bootstrap analysis, duplicate specimens/populations for individual resampling with replacement both loci and sites species, while retaining the most data-rich samples within loci (bootstrap = multilocus, loci = combined, (summarized using AMAS; Borowiec, 2016). We then nreps = 100). IQ-TREE was also used to reconstruct conducted partitioned and unpartitioned ML analyses an ML tree from a concatenated matrix, with a single of these matrices, as outlined above. Mesquite v.3.6 best-fit model automatically chosen by ModelFinder, (build 917) (Maddison & Maddison, 2018) was used to and support assessed via 1000 ultrafast bootstrap reconstruct ancestral states for courtship phenotypes, replicates (-s, -B 1000, -bnni). scored as alternative discrete states (Supporting Information, Table S3). ML character reconstructions were conducted using the one-parameter Markov RAD analysis – species groups k-state model (Lewis, 2001). Based on RAD phylogenomic results from all- taxa analyses (see below), additional analyses were conducted for members of four well- RESULTS supported clades (FC = fimbriatus + chrysomelas, pavonis, anomalus, MTVC = expanded Voucher specimen data and relevant summary values mungaich + tasmanicus + volans + calcitrans). Here, are presented in Supporting Information, Table S1. forward read (R1) RAD data were de novo assembled Raw UCE and RAD data have been submitted to the using ipyrad v.0.7.30 (Eaton & Overcast, 2016), with Short Read Archive (BioProjects PRJNA667490 and the following settings adjusted from default: max_ PRJNA665271), and all data matrices and resulting Indels_locus = 4, clust_threshold = 0.85 (within and tree files referenced below are available at Dryad across), min_samples_locus = 5; see Eaton et al., 2017). (https://doi.org/10.5061/dryad.9p8cz8wdp). We used a phylogenetic invariants analysis (Lake, 1987; Chifman & Kubatko, 2015) on linked SNPs to infer quartet trees and a species tree using tetrad UCE data v.0.7.30, part of the ipyrad.analysis toolkit, sampling The final primary 70% occupancy UCE matrix all quartets with 100 bootstrap replicates. included 472 loci, with a combined alignment length of 282 674 bp and 39 224 parsimony-informative sites. ML and SVDquartets tree topologies were identical, Courtship characters and both were well supported, with almost all nodes Adult male visual and vibrational courtship signals receiving maximum support (Fig. 3; Supporting were characterized using video and laser vibrometer Information, Fig. S2). Both gene (gCF) and site recordings (see Girard et al., 2011 for a detailed concordance (sCF) factors were relatively high, except description of methods). From these recordings we for a node uniting Saitis mutans with ‘Maratus scored nine characters for all taxa, based on original group’ genera. observations: (1) lateral fan flaps, (2) overall fan shape, The genus Saitis is not recovered as monophyletic (3) raises fan, (4) third leg use, (5) white brushes, (6) (Fig. 3). Saitis virgatus is sister to Jotus auripes, elongated spinneret display, (7) vibratory display, (8) possibly misplaced in Saitis. Saitis mutans and Saitis vigorous tapping and (9) pre-mount display. Detailed barbipes are also not recovered together; instead they character descriptions and scorings can be found are found to be successive sister species to ‘Maratus in Supporting Information, Appendix S1 and Table group’ taxa (Hypoblemum, Maratus, Saratus; Fig. 2B). © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
10 M. B. GIRARD ET AL. Saitis_virgatus_160 98.8/98.4 Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 Saitis_virgatus_159 79.2/59.4 Jotus_auripes_162 98.5/97.2 Jotus_auripes_161 Saitis_barbipes_SDSUG3313 99.7/98.5 Saitis_barbipes_SDSUG3314 gCF/sCF Saitis_mutans_178 ML/SVD bootstraps 99.4/98.7 (100 if not shown) Saitis_mutans_180 73.6/57.5 Hypoblemum_scutulatum_166 93.9/84.5 Hypoblemum_griseum_d529 Hypoblemum 86.4/86.8 Hypoblemum_griseum_168 Clade 32.6/35 53.8/50.5 55.1/43 Hypoblemum_griseum_167 95/96 M_chrysomelas_24c 98.3/89.3 Maratus_flame_18c 58.1/52.8 Maratus_rainbowi_84 83.4/81.5 Maratus_western_pavonis_44 Saratus_52c 89.6/94.9 67/67.1 Saratus_51c ‘core Maratus’ 45.4/45.8 0.02 Saratus_145 53.8/53.9 Maratus_volans_58c 65.4/74.6 Maratus_cf_plumosus_41c 26.9/41.2 Maratus_madelineae_120 85/80 Servaea_inanca_JXZ093 Figure 3. UCE maximum-likelihood phylogeny, with concordance factor (CF) values. All bootstrap values = 100, unless otherwise indicated. Bootstrap values from SVDQuartets analysis are also shown for nodes with values 7000 bp); these were trimmed to match the Sanger to this entire clade, also recovered in RAD analyses data. Based on phylogenetic placement in gene trees, (see below), as the Hypoblemum clade. A second we tentatively resolve the identity of five samples primary branch includes other Maratus species, with from Zhang & Maddison (2013, 2015) (see Fig. 2A; Saratus nested within this larger clade. We hereafter Supporting Information, Fig. S2). Two Hypoblemum refer to this clade as ‘core Maratus’ (also recovered in samples have discordant 28S vs. 16S_ND1 placements; RAD analyses, see below). we leave these identified to genus only. Prostheclina © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
PEACOCK SPIDER PHYLOGENOMICS 11 appears allied to Jotus (e.g. Fig. 2B), with Maileus we recovered 513 loci (alignment length = 59 997 bp, Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 surprisingly nested, although some nodes are poorly 11 290 SNPs, 7196 parsimony-informative sites) for supported for these individual gene trees (Supporting the sample including Jotus, Saitis, Hypoblemum, Information, Fig. S3). Saratus and Maratus. Following UCE results (see above), we rooted phylogenies using Jotus auripes + Saitis virgatus. With RAD analysis – all taxa this root placement, neither Saitis nor Maratus were Thirty-four samples were removed from downstream recovered as monophyletic in unpartitioned ML (Fig. 4; analysis because of low raw read counts. Of retained Supporting Information, Fig. S3) or SVDQuartets samples, 152 were included in all-taxa phylogenomic (Supporting Information, Figs S4, S5) analyses. One analysis. Using the custom pipeline, the number of primary branch includes members of the fimbriatus, filtered reads for these samples ranged from 101 077 chrysomelas and spicatus morphological species groups to 3623 248, with an average coverage of individual as defined by Otto & Hill (2019a; Table 1). As for UCE RAD loci of 12× (range from 3.0 to 61.1; Supporting data, members of this lineage occur on a conspicuously Information, Table S1). We recovered between 9400 long branch. Members of the chrysomelas group are and 27 893 loci per sample, but the number of shared intermixed with species of the spicatus group; because loci across the entire taxon sample was lower, probably the former species were described first, we hereafter due to the deep divergences among genera. In total refer to this entire larger clade as the chrysomelas Saitis_mutans_178 Saitis_mutans_180 Hypoblemum_grisuem_167 Hypoblemum_griseum_168 Hypoblemum_scutulatum_166 carmel_SpringHIllSA_12C carmel_SpringHIllSA_93C flame_LakeGairdnerSA_18C FC Hypoblemum Clade flame_LakeGairdnerSA_95C speculifer_TriggWA_121 fimbriatus chrysomelas_BordertownSA_17C chrysomelas_BordertownSA_22C chrysomelas_CarnavonQLD_24C chrysomelas_TregoleQLD_19C spicatus_KoondoolaWA_116 spicatus_KoondoolaWA_71 spicatus_KoondoolaWA_53 nigromaculatus_BoondallQLD_65C nigromaculatus_BoondallQLD_66C nimbus_LakeEyreSA_13C nimbus_LakeEyreSA_16C nimbus_LakeEyreSA_87C chrysomelas purcellae_CanberraACT_179 purcellae_CanberraACT_38C purcellae_CanberraACT_9C robinsoni_NewcastleNSW_131 robinsoni_NewcastleNSW_170 rainbowi_LaneCoveNSW_132 robinsoni_NewcastleNSW_172 rainbowi_LaneCoveNSW_84 maritimus_EsperanceWA_146 western_pavonis_HeardsmanWA_141 western_pavonis_TakalarupWA_88 western_pavonis_WalpoleWA_44 western_pavonis_HeardsmanWA_90 cf_leo_WaterfallSA_73C leo_DarlingtonSA_25C leo_DarlingtonSA_7C leo_DarlingtonSA_85C literatus_BooroopkiVIC_71C pavonis cf_leo_WaterfallSA_74C literatus_NhilSwampVIC_72C cf_leo_WaterfallSA_75C SE_pavonis_MaquarieTAS_37C SE_pavonis_OrfordTAS_35C SE_pavonis_OrfordTAS_36C SE_pavonis_MelbourneVIC_139 SE_pavonis_MelbourneVIC_129 ACT_pavonis_NamadgiACT_32C ACT_pavonis_NamadgiACT_33C ACT_pavonis_NamadgiACT_34C watagansi_WatagansNSW_169 albus_FlindersSA_21C albus_FlindersSA_2C albus_FlindersSA_8C albus_FlindersSA_90C albus_PtAdelaideSA_29C vultus_StrathdownieVIC_11C anomalus vultus_StrathdownieVIC_15C anomalus_DunmoreQLD_81 aurantius_OrangeNSW_62C neptunus_ButterwickNSW_173 aurantius_OrangeNSW_63C ‘core Maratus’ aurantius_OrangeNSW_76C cfneptunus_CassilisNSW_47C neptunus_ButterwickNSW_171 neptunus_ButterwickNSW_174 cinereus_StanthropeQLD_55C cinereus_StanthropeQLD_56C cinereus_StanthropeQLD_84C michaelorum_CarnavonQLD_1C michaelorum_CarnavonQLD_4C sceletus_IslaGorgeQLD_61C sceletus_WondulQLD_137 Saratus anomalus_NewcastleNSW_67 Saratus_CanberraACT_145 sceletus_WondulQLD_75 Saratus_NamadgiACT_51C Saratus_NamadgiACT_52C amabilis_CowanNSW_30C amabilis_KaputarNSW_123 amabilis_KaputarNSW_68 amabilis_MtBarneyQLD_27C amabilis amabilis_MtBarneyQLD_88C vespertillio_NgarkatSA_43C vespertillio_LittleDesVIC_46C vespertilio = bp > 95 vespertillio_MajuraACT_45C vespertillio_MajuraACT_44C speciosus_TriggWA_64 speciosus speciosus_TriggWA_86 = bp > 90 proszynskii_GoulburnNSW_3C proszynskii_NamadgiACT_143 proszynskii_MtWilliamTAS_89C velutinus proszynskii_NamadgiACT_103 velutinus_KuringgaiNSW_133 harrisi_WadabilligaNSW_14C harrisi_WadabilligaNSW_28C harrisi 87 avibus_CapeAridWA_92C caeruleus_MiddleIslWA_10C caeruleus_MiddleIslWA_6C pardus_CapeLeGrandeWA_157 clupeatus_GnangaraWA_126 clupeatus_GnangaraWA_96 flavus_YalgorupWA__124 linnaei_TwoPeoplesWA_112 expanded 65 linnaei_TwoPeoplesWA_87 linnaei_TwoPeoplesWA_63 madelineae_DardanupWA_120 mungaich madelineae_DardanupWA_57 madelineae_DardanupWA_69 sarahae_BluffsKnollWA_114 sarahae_BluffsKnollWA_85 mungaich_MtDaleWA_128 mungaich_MtDaleWA_52 vespa_lakeJasperWA_70C personatus_CapeRicheWA_154 australis_EsperanceWA_153 tasmanicus_MelbourneVIC_144 tasmanicus_MtWilliamTAS_48C tasmanicus tasmanicus_MtWilliamTAS_50C tasmanicus_PiccaninnieSA_49C (grade) tasmanicus_PiccaninnieSA_82C cf_plumosus_GleneigVIC_39C cf_plumosus_GleneigVIC_40C cf_plumosus_GleneigVIC_41C cf_plumosus_HeathertonVIC_101 cf_plumosus_HeathertonVIC_102 plumosus_KuringgaiNSW_108 plumosus_GoldburnRiverNSW_130 volans_BlackdownQLD_58C volans_BlackdownQLD_96C volans_LaneCoveQLD_94 volans_MtBarneyQLD_54C elephans_BowlingAlleyNSW_151 volans 0.0050 MTVC elephans_BowlingAlleyNSW_74 elephans_DungowanNSW_5C digitatus_KaputarNSW_125 digitatus_KaputarNSW_142 eliasi_NugaNugaQLD_80C eliasi_TregoleQLD_81C eliasi_NugaNugaQLD_97C ottoi_VenmanQLD_59C ottoi_VenmanQLD_67C calcitrans ottoi_VenmanQLD_68C calcitrans_BlackMtnACT_72 jactatus_WondulQLD_118 jactatus_WondulQLD_46 jactatus_WondulQLD_69C Jotus_auripes_161 Jotus_auripes_162C Saitis_virgatus_159 Saitis_virgatus_160 100 Figure 4. RAD all-taxa maximum-likelihood phylogeny. Bootstrap values below 90 are not shown for most nodes. Placement of Maratus sceletus, M. pardus and M. plumosus to the original groups of Otto & Hill (2019a) is designated with arrows. © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
12 M. B. GIRARD ET AL. group (redefined). Many species in the above three 2014b, M. neptunus Otto & Hill 2017a, M. aurantius Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 groups were previously placed in the genus ‘Lycidas’ Otto & Hill 2017a]. The only clear case of strongly (but not including the type species of ‘Lycidas’). supported non-monophyly involves the geographically Consistent with UCE data, fimbriatus + chrysomelas widespread M. pavonis. In both tetrad and ML results, (FC) clades are recovered as sister to Hypoblemum, M. pavonis specimens from Western Australia are more and together comprise the Hypoblemum clade. We closely related to Western Australia M. maritimus Otto discuss below possible morphological and behavioural & Hill 2014b, with eastern M. pavonis populations synapomorphies that unite members of this clade. phylogenetically closer to geographically adjacent A second primary Maratus branch includes all other M. literatus and M. leo (Figs 4, 5). This result supports Maratus taxa (‘core Maratus’), with Saratus samples the contention that M. pavonis currently includes nested within this larger clade (Fig. 4; Supporting several distinct species, as supported by both genitalic Information, Figs S4, S5). The ‘core Maratus’ clade (Baehr & Whyte, 2016) and male ornamentation includes the anomalus species group, including both characters (Hill & Otto, 2011; Otto & Hill, 2019a). the type species of ‘Lycidas’, and M. amabilis, the type Introgression with geographical neighbours might also species of the genus Maratus. Many clades within be driving this pattern, but if this were true we might ‘core Maratus’ are mostly concordant with the species expect more lineage non-monophyly (e.g. Western group divisions of Otto & Hill (2019a), as based on Australia M. pavonis intermixed with Western morphological/behavioural similarities. Exceptions Australia M. maritimus, etc.). include a monophyletic anomalus group, with the Multiple analyses support the hypothesis that three inclusion of M. sceletus Otto & Hill 2015a (from the novel forms included here represent undescribed calcitrans group). Maratus plumosus Otto & Hill species (M. ‘carmel’, M. ‘flame’, M. cf. plumosus). 2013b, a hypothesized (but divergent) member of the Two forms (M. cf. neptunus, M. cf. leo) might best be calcitrans group, also falls outside of this group. The interpreted as geographical variants (Figs 4, 5), but tasmanicus group is not recovered as monophyletic, larger sample sizes and formal integrative species but instead represents a grade sister to the ‘expanded delimitation analyses are needed to rigorously test mungaich’ (EM) clade. The large, poorly resolved EM these species hypotheses. clade includes species from several groups as follows: M. clupeatus Otto & Hill 2014d (not a straggler), M. pardus Otto & Hill 2014c (from the volans group), Character evolution M. australis Otto & Hill 2016b (from the tasmanicus We lacked some character data for Saitis barbipes group), and single sampled species from the flavus, and Hypoblemum (Supporting Information, Table linnaei, personatus and vespa groups. The EM clade S2), causing ambiguity in certain ancestral character is further discussed below, but this grouping is reconstructions. For the sparse UCE taxon sample that biogeographically compelling, as all included taxa are importantly includes Saitis barbipes, we reconstruct endemic to south-west Australia. the common ancestor of the entire Saitis Clade as males using third legs during courtship (character 4, Fig. S6), and with vibrations produced during a pre- RAD analysis – species groups mount display (character 8, Fig. S7; see Girard et al., Tetrad results and locus statistics for the FC, pavonis, 2011), with reversals observed for both characters. anomalus and MTVC clades are shown in Figure 5. The common ancestor of the Maratus group is Interspecific relationships in tetrad analyses are unambiguously reconstructed as with males raising generally better supported than in ML or SVDQuartets their abdomen during courtship (character 1, Fig. S6), analysis, perhaps reflecting the greater number of consistent with the hypothesis of Otto & Hill (2012a; loci recovered for each group. An exception is the EM Fig. 2B). Ancestral state reconstructions for other clade – although tetrad analyses included more SNPs characters are shown in Figures S6 and S7. and parsimony-informative sites than in all-taxa Ancestral state reconstructions for the denser RAD ML analyses, relationships are conspicuously poorly taxon sample that included the full Maratus sample supported in both (Figs 4, 5). (but lacked Saitis barbipes) are shown in Supporting Almost all a priori species are supported as Information, Figures S8–S15. These reconstructions monophyletic on tetrad trees (Fig. 5). One caveat of suggest the evolution of a strong white brush on the this claim is that several species with multiple samples male third leg evolving at the base of ‘core Maratus’ only include specimens from the same geographical (character 5, Fig. S11). Two characters suggest that location. Five species not recovered as monophyletic the pre-mount display and pre-mount vibrations may on ML trees are recovered as exclusive groupings on be unique in the Hypoblemum clade (characters 8 and tetrad trees [M. chrysomelas (Simon 1909), M. leo 9, Figs S14, S15), predicting that missing data for Otto & Hill 2014b + M. cf. leo, M. literatus Otto & Hill Hypoblemum species will conform to the FC clade. The © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
PEACOCK SPIDER PHYLOGENOMICS 13 84 flame_SA_95C FC Clade Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 100 flame_SA_26C flame_SA_18C 5381 loci, 100 carmel_SA_93C 63102 SNPs 83 89 carmel_SA_23C 17622 PIs 100 carmel_SA_12C speculifer_WA_121 chrysomelas_SA_20C 43 chrysomelas_QLD_24C 46 chrysomelas_QLD_19C 47 chrysomelas_SA_22C 90 68 chrysomelas_SA_17C spicatus_WA_71 95 spicatus_WA_116 40 74 spicatus_WA_53 nigromaculatus_QLD_64C 96 97 nigromaculatus_QLD_65C 90 nigromaculatus_QLD_66C nimbus_SA_16C 100 nimbus_SA_87C 100 43 nimbus_SA_13C purcellae_ACT_9C 100 purcellae_ACT_38C 71 albus_SA_2C 100 purcellae_ACT_179 albus_SA_8C anomalus 59 robinsoni_NSW_172 76 albus_SA_21C 100 robinsoni_NSW_131 100 67 52 albus_SA_90C 11609 loci, robinsoni_NSW_170 albus_SA_29C 100 115608 SNPs vultus_VIC_11C 100 33606 PIs vultus_VIC_15C anomalus_NSW_67 100 anomalus_QLD_81 cfneptunus_NSW_47C 47 neptunus_NSW_173 98 neptunus_NSW_174 100 45 neptunus_NSW_171 pavonis 100 aurantius_NSW_76C 85 aurantius_NSW_63C 9055 loci, 47 aurantius_NSW_62C 100 103477 SNPs cinereus_QLD_84C rainbowi_NSW_132 26175 PIs 83 cinereus_QLD_56C 100 rainbowi_NSW_84 50 72 cinereus_QLD_55C western_pavonis_WA_141 100 michaelorum_QLD_1C 68 western_pavonis_WA_90 92 65 michaelorum_QLD_4C western_pavonis_WA_107 sceletus_QLD_75 100 51 western_pavonis_WA_88 sceletus_QLD_137 99 88 western_pavonis_WA_44 100 sceletus_QLD_61C maritimus_WA_146 SE_pavonis_VIC_139 90 SE_pavonis_VIC_129 93 65 SE_pavonis_TAS_37C 46 avibus_WA_92C 79 SE_pavonis_TAS_35C SE_pavonis_TAS_36C 88 caeruleus_WA_10C 98 caeruleus_WA_6C 100 leo_SA_85C leo_SA_25C sarahae_WA_114 96 100 sarahae_WA_85 84 leo_SA_7C 53 vespa_WA_70C 55 cf_leo_SA_75C 22 linnaei_WA_112 36 cf_leo_SA_73C 94 linnaei_WA_63 63 45 50 cf_leo_SA_74C 52 36 linnaei_WA_87 ACT_pavonis_ACT_32C madelineae_WA_69 62 madelineae_WA_57 55 ACT_pavonis_ACT_33C 71 53 34 17 madelineae_WA_120 51 ACT_pavonis_ACT_34C EM flavus_WA__124 literatus_VIC_71C 78 34 mungaich_WA_52 57 83 mungaich_WA_128 literatus_VIC_72C clupeatus_WA_126 watagansi_NSW_169 57 79 clupeatus_WA_96 personatus_WA_154 australis_WA_153 44 pardus_WA_157 cf_plumosus_VIC_41C 68 46 cf_plumosus_VIC_39C 48 cf_plumosus_VIC_101 93 62 cf_plumosus_VIC_102 cf_plumosus_VIC_40C 77 tasmanicus_VIC_144 85 tasmanicus_TAS_48C 95 tasmanicus_TAS_50C 91 83 26 93 tasmanicus_SA_82C tasmanicus_SA_49C plumosus_NSW_108 49 plumosus_NSW_130 elephans_NSW_5C 99 elephans_NSW_74 99 elephans_NSW_151 97 volans_QLD_94 MTVC 55 volans_QLD_54C 100 volans_QLD_96C 13264 loci, 100 volans_QLD_58C jactatus_QLD_118 152942 SNPs 100 jactatus_QLD_69C 40981 PIs 82 80 jactatus_QLD_46 calcitrans_BlackMtnACT_72 eliasi_QLD_81C 91 100 eliasi_QLD_97C 94 eliasi_QLD_80C 100 digitatus_NSW_142 94 100 digitatus_NSW_125 ottoi_QLD_59C 99 ottoi_QLD_67C 38 ottoi_QLD_68C Figure 5. RAD tetrad results for the FC, pavonis, anomalus and MTVC clades. © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
14 M. B. GIRARD ET AL. ‘core Maratus’ Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blaa165/6126965 by University of California, Berkeley user on 05 February 2021 MTVC expanded mungaich elephans (2) avibus (2) pavonis anomalus volans (2) caeruleus (2) vultus (1) watagansi Saratus rainbowi (1) albus flavus calcitrans plumosus (4) amabilis (2) Hypoblemum Clade clupeatus (2) anomalus maritimus jactatus (1) cf plumosus (4) chrysomelas speciosus madelineae (2) western pavonis (1) robinsoni H griseum sceletus calcitrans tasmanicus (2) purcellae vespertillio (3) ACT pavonis (1)) H scutulatum linnaei cinereus outgroups nimbus ottoi Saitis mutans michaelorum SE pavonis (1) vespa (4) australis (3) fimbriatus chrysomelas nigromaculatus velutinus digitatus (3) speculifer proszynskii sarahae (2) Saitis virgatus cf neptunus neptunus + personatus literatus (1) flame eliasi (5) 5 mungaich (2) Jotus auripes aurantius leo + cf leo pardus (2) harrisi (3) spicatus carmel 1 1 1 1 1 1 1 2 3 3 1 3 5 2 2 4 4 2 3 2 2 2 2 2 2 2 4 Character 3: overall fan shape 0=no fan 1=round 2=elliptical 3=lobed 4=lobed(post) 5=diamond Figure 6. Maximum-likelihood ancestral character reconstruction for fan shape, on the skeletal RAD taxon sample phylogeny. Character states are noted on the phylogeny, and are associated with taxon names; all taxa with character state 0 (=no fan) are left unlabelled. evolution of fan shape diversity across Maratus group the abdomen during courtship are ancestral to the members is illustrated in Figure 6. ‘core Maratus’ clade. Other traits, particularly fan shape, show evidence of multiple independent gains and losses throughout what is currently recognized as Maratus. DISCUSSION Our phylogenomic results reveal that both Saitis and Maratus, as currently recognized, are paraphyletic Saitis Clade taxa. In Maratus, some species are more closely Zhang & Maddison (2015) define the Saitis Clade as related to Hypoblemum taxa. A ‘core Maratus’ including multiple genera (Saitis, Maratus, Saratus, clade includes several well-supported sub-clades Jotus, Hypoblemum, Prostheclina, Maileus, Barraina that mostly correspond to morphological species and possibly Margaromma), united in sharing several groups, although some hypothesized groupings were male genitalic features, including a lamella on the not supported. Saratus is nested within the ‘core tegular shoulder distinctive for euophryine spiders. Maratus’ clade. Based on this phylogenomic tree Zhang & Maddison (2015) recognized the striking topology, some traits such as using third legs and variation in male courtship ornamentation in the © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–24
You can also read