sign in sign up
<<
Home , Evolution of spiders

Molecular and 127 (2018) 146–155

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier.com/locate/ympev

Early Cretaceous greenhouse pumped higher taxa diversification in T ⁎ Lili Shaoa,b, Shuqiang Lia, a Key Laboratory of Zoological and Evolution, Institute of Zoology, Chinese Academy of Sciences, Beijing 100101, China b University of Chinese Academy of Sciences, Beijing 100049, China

ARTICLE INFO ABSTRACT

Keywords: The Cretaceous experienced one of the most remarkable greenhouse periods in geological history. During this Araneae time, ecosystem reorganization significantly impacted the diversification of many groups of organisms. The rise Phylogenomics of angiosperms marked a major biome turnover. Notwithstanding, relatively little remains known about how the Cretaceous global ecosystem impacted the evolution of spiders, which constitute one of the most abundant Diversification groups of predators. Herein, we evaluate the transcriptomes of 91 taxa representing more than half of the spider Angiosperm families. We add 23 newly sequenced taxa to the existing database to obtain a robust phylogenomic assessment. Phylogenetic reconstructions using different datasets and methods obtain novel placements of some groups, especially in the and the group having a retrolateral tibial apophysis (RTA). Molecular analyses indicate an expansion of the RTA at the Early Cretaceous with a hunting predatory strategy shift. Fossil analyses show a 7-fold increase of diversification rate at the same period, but this likely owes to the first oc- currence of spiders in deposit. Additional analyses of fossil abundance show an accumulation of spider lineages in the Early Cretaceous. We speculate that the establishment of a warm greenhouse climate pumped the diversification of spiders, in particular among webless forms tracking the abundance of insect prey. Our study offers a new pathway for future investigations of spider phylogeny and diversification.

1. Introduction precipitated mammalian interordinal diversification (O'Leary et al., 2013), and may have provided opportunities for modern birds to di- The Cretaceous period (145–66 Million years ago (Ma)) offers one of versify into new adaptive-zones (Prum et al., 2015). Therefore, change the best examples of a greenhouse climate on Earth because high in the Cretaceous ecosystem appears to have strongly influenced spatial average global temperature lasted until about 70 Ma (Wang et al., and temporal population dynamics. 2014). The progressive breakup of Pangaea as a result of intense tec- Spiders ( Araneae) are an important element of the food chain. tonic activity led to a near total reorganization of global ecosystem The group consists of more than 46,000 described (World Spider including the of dominant groups and subsequent diversifi- Catalog 2017) that occur almost everywhere from Arctic islands to dry cation of novel taxa (Coiffard et al., 2012; Lloyd et al., 2008). Of par- desert regions. Their evolution has witnessed mass , radical ticular interest, the Cretaceous Terrestrial Revolution (KTR; also re- alterations in terrestrial floras, continental rearrangements, and ferred to as the angiosperm revolution) saw an explosive radiation of changes in key environmental parameters. Throughout this time, spi- angiosperms and the replacement of ferns and gymnosperms between ders have developed a suite of morphological characteristics and re- 125 and 80 Ma (Lloyd et al., 2008). markably diverse modes of (Selden and Penney, 2010). In parti- Angiosperms were the fundamental players in terrestrial evolution. cular, the evolution of the orb web has been long considered a “key As primary producers, their domination provided new ecological and innovation” in spiders (Bond and Opell, 1998). The ancient origin of the evolutionary opportunities for insects such as herbivorous orb web and web change or loss associates with spider diversification (Zhang et al., 2018), lepidopterans (Wahlberg et al., 2013), and ants (Blackledge et al., 2009; Dimitrov et al., 2016; Garrison et al., 2016). (Moreau et al., 2006). Among vertebrates, crown squamates (lizards Nevertheless, it remains unclear how environmental changes, especially and snakes) appeared in the and underwent a Cretaceous ra- during the Cretaceous, influenced spider diversification. diation (Bronzati et al., 2015). Mammals and birds experienced a strong Molecular phylogenetic analyses have mostly driven our under- increase in diversification, mainly after the end-Cretaceous crisis. standing of the historical pattern of spider diversification. Multi- Nevertheless, the KTR increased ecospace diversity, possibly analyses have incrementally covered parts of the spider tree of life and

⁎ Corresponding author at: 1 Beichen West Road, Chaoyang District, Beijing 100101, China. E-mail address: [email protected] (S. Li). https://doi.org/10.1016/j.ympev.2018.05.026 Received 26 July 2017; Received in revised form 14 May 2018; Accepted 19 May 2018 Available online 24 May 2018 1055-7903/ © 2018 Elsevier Inc. All rights reserved. L. Shao, S. Li Molecular Phylogenetics and Evolution 127 (2018) 146–155 resolved the evolution of some groups such as (Bond shown to compromise phylogenomic analysis. Therefore, we re- et al., 2012), (Wood et al., 2012), orb-weavers grouped the initial 979 OGs (dataset A). Each OG was ranked by (Blackledge et al., 2009; Dimitrov et al., 2016), sparassids (Moradmand evolutionary rate following Telford et al. (2014) as follows: using an et al., 2014), and nearly all spider groups (Wheeler et al., 2016). Phy- alignment for each gene, a maximum-likelihood (ML) tree was cal- logenomic studies rejected the prevailing paradigm for orb web evo- culated using RAxML v8.0.0 (Stamatakis, 2014) and the total length lution and corroborated several unexpected results, such as the position of that tree (in estimated substitutions per position across all bran- of Leptonetidae (Bond et al., 2014; Fernández et al., 2014; Garrison ches) was divided by the number of taxa on the tree to give an es- et al., 2016). Still, much of the phylogenetic framework for spiders timate of the rate of evolution for each gene. Half of the slowest- remains uncertain. Molecular phylogenies cannot directly place extinct evolving genes were concatenated to form dataset B. Each OG was lineages, and yet the fossil record is our primary window onto the di- also ranked by missing data ratio: half of the OGs with lowest missing versification of ancient life. Fossil spiders supply a handful of calibra- data ratio (lower than 0.1535) were then concatenated to form da- tion points in molecular studies and only a few quantitative analyses of taset C. BaCoCa (Kück and Struck, 2014)andtwometrics(theχ2-test exist (Selden and Penney, 2010). Thus, phylogenies and the fossil of heterogeneity and relative composition frequency variability; record provide two complementary perspectives that can resolve tem- RCFV) (Zhong et al., 2011) were used to identify the OGs with amino poral variation in . acid compositional heterogeneity. OGs with p-values below 0.01 Herein, we reconstruct the phylogenetic history of spiders using were removed because the composition significantly deviated from previously acquired and de novo transcriptomes. We perform diver- homogeneity. Then, 490 OGs with the lower RCFV value (lower than gence dating and diversification rate shift analyses to hypothesize the 0.1053) formed dataset D. The full dataset was also optimized using tempo and mode of diversification at the level of family. We re-examine MARE (Meyer et al., 2011)tofind highly informative subsets of OGs the global fossil record to assess relative abundance of different groups. (dataset E). Detailed information for all the five datasets was sum- Finally, we reconstruct the ancestral predatory strategies among dif- marized in Table S3. ferent groups and suggest some of the possible drivers of spider di- versification during the Early Cretaceous greenhouse. 2.3. Phylogenetics analyses

2. Materials and methods OGs for each dataset were concatenated using FASconCAT v1.0 after alignment (Kück and Meusemann, 2010). A meta-partition ana- 2.1. Sampling, extraction, and assembly lysis using the PartitionFinderProtein v1.1.1 (Lanfear et al., 2012) was conducted for all five datasets to select the optimal meta-partitions and New transcriptomic data were generated for 23 specimens re- the best-fit amino acid substitution models. Maximum-likelihood presenting major groups of spiders. Previously available transcriptomic topologies were inferred with RAxML v8.0.0 (Stamatakis, 2014) using and genomic data were collected for 68 spider samples plus 14 out- partitions as indicated by ParitionFinderProtein, associated best-fit group taxa (Tables S1 and S2). All specimens used in this study were substitution models, and the GAMMA parameter to model rate-het- legally collected. Samples were flash frozen in liquid nitrogen and erogeneity. Nodal support was measured with 1,000 fast bootstrap mRNA was extracted using the TRIzol total RNA extraction method pseudoreplicates. Relatively small datasets B, C and D were also ana- (Life Technologies). Purification of mRNA, library preparation, se- lyzed under the site-heterogeneous mixture model, CAT- quencing, and quality control were done by Novogene Bioinformatics GTR + GAMMA, in PhyloBayesMPI v1.5a (Lartillot et al., 2013) using Technology Co. Ltd. Sequencing libraries were generated using the resources available from the CIPRES Science Gateway (Miller et al., ® ® NEBNext UltraTM RNA Library Prep Kit for Illumina (NEB, USA) fol- 2010). A coalescent-based method as implemented in ASTRAL (Accu- lowing manufacturer's recommendations and index codes were added rate Species Tree Reconstruction ALgorithm; astral v4.10.2; Mirarab to attribute sequences to each sample. Briefly, mRNA was purified from et al., 2014) was used to infer a species tree from a series of unrooted total RNA using poly-T oligo-attached magnetic beads. Fragmentation gene networks. was carried out using divalent cations under elevated temperature in NEBNext First Strand Synthesis Reaction Buffer (5X). Then first and 2.4. Divergence time estimation second strand cDNA were successively synthesized. For detailed methods of library preparation and sequencing refer to the supple- Selected meta-partitions from dataset B were used to estimate the mentary methods. Raw data (raw reads) in the fastq format were pro- time tree for spiders with each partition having at least partial non- cessed by removing reads containing adapter, reads containing ploy-N, ambiguous sequences for all 105 taxa and a minimum amino acid and low quality. Next, clean reads were assembled using default para- alignment length of 500 sites. Fourteen calibration points (5 for out- meters in Trinity v2.0.5 (Grabherr et al., 2011). Previously available group calibration nodes and 9 for ingroup calibration nodes) were sequence data were acquired from the NCBI database, quality-checked, chosen conservatively to calibrate the inferred phylogeny. and trimmed using the FastQC (http://www.bioinformatics.babraham. Although discussion persists on the nature of the Ediacaran fauna ac.uk/projects/fastqc/) and FASTXToolkit (http://hannonlab.cshl.edu/ (including putative ; Erwin and Valentine, 2010), 580 Ma fastx_toolkit/index.html) before assembly with Trinity. constituted a justifiable maximum hard bound for the origin of the major , which we used as the maximum rootHeight 2.2. Orthology determination and data filtering age. To constrain the node of Copepoda + Branchiopoda, we used a secondary calibration: a normal distribution with mean 390 Ma and Putative orthologs were determined for each species with 95% confidence interval of 513–267 Ma (Misof et al., 2014). The oldest HaMStR v13.2.3 (Ebersberger et al., 2009) using the Arthropoda core known fossil insect, Rhyniella praecursor, was used to constrain the ortholog set. The resulting orthologous groups (OGs) were further minimum age of 411.5 Ma for the node consisting of Protura + Col- processed by filtering them with fewer than half of all sample species lembola (Scourfield, 1940). The fossil Triassolestodes asiaticus re- while discarding amino acid sequences with length shorter than 75 presented the order crown Odonata, and we set the crown Odonata sites. The remaining OGs were then aligned with MAFFT v7.222 node with a minimum age of 237 Ma (Kohli et al., 2016). The earliest (Katoh and Standley, 2016) and processed with Aliscore (Kück et al., true insect fossil, Rhyniognatha hirsti, was used to constrain the 2010)andALICUT(http://zfmk.de/web/ZFMK_Mitarbeiter/ minimum age of Dicondylia at 411.5 Ma (Engel and Grimaldi, 2004). KckPatrick/Software/AliCUT/Download/index.en.html). Missing Although the unequivocal land plant megafossils date back into the data, evolutionary rate, and compositional heterogeneity have been mid-Silurian (about 425 Ma) (e.g., Hart et al., 2013), the earliest

147 L. Shao, S. Li Molecular Phylogenetics and Evolution 127 (2018) 146–155 terrestrial food webs probably existed somewhat earlier, most likely and Paratropidae). For further parameters, we used default or re- dating back to the mid- or earliest late-Ordovician (about 450 Ma) commended settings in Beast. Sampling each 1,000 tree, different mil- (Labandeira, 2005). Therefore, we set the maximum age of the three lion iterations were performed to achieve convergence depending on above nodes to 450 Ma with a conservative uniform prior. The fossil the meta-partition. All parameters (node estimates included) achieved Eophalangium sheari from the Rhynie chert has been referred to a crown- an effective sample size (ESS) > 200 after combining all runs of a meta- group harvestman and is one of the oldest known crown-group cheli- partition with adjusting the burn-in of each run using Tracer v1.6 cerates (Dunlop et al., 2004). Accordingly, we used an exponential prior (Rambaut et al., 2014). From each tree file (one tree file per meta- with dates closer to the younger bound 411 Ma and the maximum age partition), we trimmed the burn-in trees for the single run. Next, we was set to 500 Ma following Sharma and Giribet (2014). sampled 10,000 trees from the post-burn-in trees. Therefore, we took A fossil Palaeothele montceauensis from the Upper of 10,000 trees from 79 partitions, i.e. 790,000 trees. All collected trees Montceau-les-Mines, France, dated at around 299 Ma. This single fossil were combined with LogCombiner in the BEAST package. Consensus was assigned to (Selden, 1996). Therefore, we set the trees across all meta-partitions were generated finally. minimum age of 299 Ma to the crown spiders with a uniform prior. This fossil also provided evidence that opisthotheles (the sister group of 2.5. Diversification analyses mesotheles) were present in the Carboniferous period (Selden, 1996). Therefore, the maximum age of opisthotheles can be set to 299 Ma in- To visually examine the temporal pattern of lineage diversification, cluding both mygalomorph and araneomorph clades. A mygalomorph lineage through time (LTT) plots were constructed in the R package ape fossil, Rosamygale grauvogeli (Selden and Gall, 1992) from the middle v3.4 (Paradis, 2012). A reduced dataset (one individual per family) was was used to constrain the minimum age of 242 Ma for the used for estimating diversification rates on the family level. Incomplete crown of the mygalomorph clade and the maximum age was set to sampling was accounted for by applying clade-specific sampling 299 Ma with a conservative uniform prior. Extant araneomorphs were fractions on the family level in each of the principal spider clades as calibrated with the oldest known fossil Triassaraneus andersonorum defined in Garrison et al. (2016). We obtained an overall proportion of (Selden et al., 1999) setting a minimum age at 228 Ma and the max- 0.54 (61 out of 113 families). imum age was also set to 299 Ma with a conservative uniform prior. We Bayesian analysis of macroevolutionary mixture (BAMM v2.2.0, chose lognormal priors for the following calibration points because we Rabosky et al., 2014) was used to estimate and extinction considered a large uniform prior distribution to be too conservative and rates through time and to detect differences in diversification rates the increasing richness of fossils after the Triassic made us set the crown across lineages. Four independent BAMM runs were performed for node or stem age near the fossil age. The fossil Eoplectreurys gertschi 50,000,000 generations while sampling every 1,000 generation and (Selden and Huang, 2010) has been included in the extant family priors were chosen using the function “setBAMMpriors” in the R Plectreuridae. It served to represent the oldest and only Synspermiata package BAMMtools v2.0 (Rabosky et al., 2014). Convergence of the known during Jurassic. Therefore, we set the crown age of the Syn- runs was assessed, and the CODA library was used to check the effective spermiata clade to no earlier than Jurassic (lognormal prior, µ = 2.5, sample sizes of log-likelihoods and the number of shift events present in d = 0.5, minimum age 170 Ma). Some records of antrodiaetid myga- each sample (Becker et al., 1988). All parameters had effective sample lomorphs (Aliatypus and Antrodiaetus) have been dated to the Aptian sizes > 1000. A burn-in of 0.1 was applied and the posterior distribu- (Eskov and Zonshtein, 1990). Therefore, we set the crown age of this tion was used to compute the mean global rates of diversification clade in this period (lognormal prior, µ =2, d = 0.5, minimum age through time with the function “mean phylorate”. A diversification 110 Ma). Extinct Jurassic Mongolarachnidae (Selden et al., 2013) and rate-plot was obtained with the function “plotRateThroughTime”. Zhizhu daohugouensis (∼165 Ma) (Selden et al., 2015), which placed in Further, we quantified the percentage of taxa sampled per family and the stem-Deinopoidea, suggested a greater diversity of cribellate orbi- incorporated these data into analysis as a comparative diversification cularians during the Middle Jurassic. Additionally, large molecular analysis with the BAMM method. analyses for Araneae and our analyses here suggested that Orbiculariae The TreePar package (Stadler, 2013a) was used to assess speciation was not monophyletic and superfamily Deinopoidea was closely related and extinction rates through time under a birth- model (without to the RTAclade (Bond et al., 2014; Fernández et al., 2014). Because the mass extinction), and to specifically detect potential rapid changes in divergence date for the common ancestors of Orbiculariae may deeper diversification rates that might have owed to environmental factors. to the Upper Triassic, we set the crown age of Orbiculariae to between The “bd.shifts.optim” function was employed to estimate discrete the Upper Triassic and Lower Jurassic (lognormal prior, µ = 3.1, changes in speciation and extinction rates. TreePar analyses were run d = 0.6, minimum age 176 Ma). Among the Dysderoidea, the oldest with the following settings: start = 0, end = crown age (339 Ma) esti- fossil, Microsegestria poinari (Segestriidae) has been found in Lebanese mated by dating analyses, grid = 1 Ma, and posdiv = FALSE to allow amber (135–125 Ma) from the Lower Cretaceous (Wunderlich and the diversification rate to be negative (i.e., allowing for periods of de- Milki, 2004). Therefore, the crown age of Dysderoidea was set to this clining diversity). We obtained maximum-likelihood rate estimates for time (lognormal prior, µ =2, d = 0.5, minimum age 125 Ma). The zero, one, two and three rate shifts. The likelihood ratio test was used to oldest fossil of the Oecobiidae, Lebanoecobius schleei, was also in Leba- select the best model. nese amber (Wunderlich, 2004). Accordingly, we used the same date Fossil occurrence data were downloaded from the for the node Hersiliidae + Oecobiidae (lognormal prior, µ =2,d = 0.5, Database (PBDB; http://paleobiodb.org/) without revising its fossil minimum age 125 Ma). Although the oldest fossil of the family Liny- identifications. The analyses of spider diversification were carried out phiidae was in Lebanese amber (Penney and Selden, 2002), no Cre- using a hierarchical Bayesian model implemented in PyRate to infer the taceous records exist for its sister family, Pimoidae. Therefore, we set temporal dynamics of origination and extinction at the family level the stem of Linyphiidae + Pimoidae to a relatively young age during (Silvestro et al., 2014) as this was the best compromise between sam- the early stage of Cretaceous (lognormal prior, µ =2, d = 0.5, pling limitation and taxonomic uncertainty. Given the exceptional minimum age 125 Ma) (Table S4). preservation and sampling of the Burmese and Baltic , we also For each meta-partition, we used the best scoring substitution model conducted a PyRate analysis on only non-amber fossils to minimize the identified with PartitionFinderProtein. In Beast v1.8.1 (Drummond effect of amber preservation beginning in the Early Cretaceous. The et al., 2012), we chose an uncorrelated lognormal relaxed clock, and PyRate approach was followed, as in a previous study that analyzed the the Yule Process. The topology was fixed according to phylogenetic evolutionary history of vascular plants (Silvestro et al., 2015), which reference (mainly the analysis of all gene dataset A) and families with focused on variation in origination and extinction at the global scale unstable placement among different methods were removed (Eresidae and over large temporal ranges.

148 L. Shao, S. Li Molecular Phylogenetics and Evolution 127 (2018) 146–155

The distributions of fossil collections during different periods were Tetrablemmidae formed the sister-group to Pholcidae and Diguetidae. made by Fossilworks (http://fossilworks.org/). PyRate was run for Ochyroceratidae was the sister to the classical Scytodoidea. Deinopidae 5,000,000 MCMC generations on each of the 10 randomly replicated and Oecobioidea located close to the RTA clade. Sparassidae formed the datasets. After excluding the first 20% of the samples as the burn-in sister taxa to the clade ((((Desidae, Amphinectidae), Agelenidae), period, the posterior estimates of the origination and extinction rates (Dictynidae, Hahniidae)), Amaurobiidae). across all replicates were combined and used to generate rates-through- Our divergence time estimates suggested that crown-group spiders time plots. The magnitude of targeted rate changes and the rates of two shared their last common ancestor around 338.96 Ma (95% HPD adjacent intervals were considered significantly different when the [386.69 Ma; 299.00 Ma]). The most recent common ancestor (MRCA) mean of one laid outside the 95% highest posterior density (HPD) of the of diverged around 307.19 (346.10–274.15) Ma. The other, and vice versa. The marginal posterior distributions of origina- split within occurred in the early Permian around tion and extinction rates through the largest extinction events docu- 282.1 (299.00–255.21) Ma. The basal split in spiders producing the mented in geological history were examined. modern clade Orbiculariae + RTA happened around 197.18 (220.54–182.28) Ma and diversified during the Cretaceous. A chrono- 2.6. Ancestral state reconstruction gram based on selected meta-partitions from dataset B was shown in Figs. 2 and S9. To investigate the evolution of web patterns, ancestral state re- constructions were performed using the R packages ape (Paradis, 2012) 3.2. Diversification dynamics and phytools (Revell, 2012). The terminal state of each family was la- beled by the major type of web in this family (sensu Blackledge et al., BAMM suggested a global decrease in the diversification rate 2009) and spider lifestyles (Jocqué and Dippenaar-choeman, 2006). through time at the family level since the Carboniferous. The net di- Likelihood models for discrete characters have been based equal rates versification rates ranged from 0.036 events/Myr/lineage in the (ER), symmetrical (SYM), and all rates being different (ARD). We fitted Carboniferous to 0.006 events/Myr/lineage in the Cenozoic (Fig. 3b). A these models to our data and selected the one that resulted in the significantly supported rate increase occurred during the Early Cre- highest likelihood. The best performing model was then used to re- taceous. It mainly corresponded with the rise of the RTA clade. Another construct web evolution using a stochastic character mapping approach slight rate increase happened near the divergence of Opisthothelae. The (SIMMAP) as implemented in phytools. The stochastic character map- “per family” sampling strategy obtained an increasing diversification ping is a Bayesian approximation to ancestral state reconstruction rate (Fig. S12) that may have resulted from the small sampling ratio per (Bollback, 2006). We preferred SIMMAP to other likelihood approaches family, but it was reasonable of the overall trend of rate increase, to ancestral state reconstruction of discrete traits because it allowed especially since the Cretaceous. TreePar analyses supported a diversi- changes to occur along branches and assessed the uncertainty in char- fication model with varying rates. The models with zero and one rate acter history. Due to the dispute on the origin of orb web and the un- shifts were rejected in favor of a model with two rate shifts (P < 0.05). stable placement of Oecobioidea among our ML analyses and Phylo- A model with two rate shifts was not rejected in favor of a model with Bayes analyses, we reconstructed the ancestral state of webs for the three rate shifts (P = 0.62). The two shifts were detected at 95 and main nodes with or without the mainly free-living Hersiliidae, respec- 39 Ma (Table S5). The more recent rate shift may have been an artifact tively. because of relatively less complete sampling. After removing fossils that could not be classified to family, 2092 3. Results and 206 occurrences were assembled with or without amber fossil in- cluded (Tables S6 and S7), respectively. Spiders have extremely sparse 3.1. Datasets, phylogenomics and divergence analyses fossil records before the Early Cretaceous. No fossil spider is known from the Late Permian (Lopingian) to Early Triassic, and a single oc- Our sequencing of an average of 6.9 gigabases (Gb) of cDNA for 23 currence only exists in the Early Permian (Cisuralian), Middle Permian transcriptomes obtained 979 OGs after initial processing plus another (Guadalupian), and Early Jurassic. No time interval in the Permian, four datasets to account for possible biases and to assess the robustness Triassic, or Early/Middle Jurassic contains more than a single spider of the results (Table S3). All OGs in dataset A were used to construct family. Only the Pennsylvanian has three families and the Late Jurassic unrooted gene-trees to infer a species tree. has five. Therefore, PyRate may have greatly underestimated the un- Our database represented the largest assessment of spider phylo- certainty on the timing of speciation and extinctions events and the genomic analyses to date and refined existing knowledge of the re- error bars of zero on diversification. Therefore, we focused on the time lationships within major groups of spiders due to the addition of tran- intervals after the Late Jurassic. scriptomic data from 12 spider families not previously included in Upon including amber fossils, a notable increase in net diversifica- phylogenomic analyses. Our results were largely congruent with earlier tion rate occurred at the Jurassic–Cretaceous boundary (7-fold) works (Bond et al., 2014; Fernández et al., 2014; Garrison et al., 2016) (Figs. 3b and S10). This high rate continued to the and in recovering all major backbone lineages (e.g., Mygalomorphae, Ara- then decreased mainly due to a significant decrease in the speciation neomorphae, and RTA clade) (Fig. 1). Our ML analyses of dataset A rate at the family level (12-fold) following the Albian–Cenomanian recovered most clades with maximum support values (bootstrap sup- boundary. On the family level, extinction rates decreased about 9-fold port value = 100%). The other datasets resolved a few slightly different at the Jurassic–Cretaceous boundary and just slightly increased (4-fold) relationships (Fig. 1 and Figs. S1–S4). PhyloBayes analyses obtained after the Albian–Cenomanian boundary. The significant decrease of relatively stable relationship and most nodes had posterior probabilities speciation rate at the early Late Cretaceous could have partially owed to of 1 (Figs. S5–S7). The coalescent-based species tree analysis obtained a sparse Late Cretaceous spider records, especially after the Cenomanian well-supported tree (most nodes with posterior probability support stage. No notable diversification rate shift occurred on the family level values higher than 0.95, Fig. S8). With few exceptions, the species tree at the Cretaceous–Palaeogene boundary (Figs. 3b and S10). was compatible with the likelihood-based dataset A and PhyloBayes After excluding amber fossils, the overall diversification dynamics results. changed little. However, no notable increase in net diversification rate Our analyses obtained novel placements of some spider families occurred at the Jurassic–Cretaceous boundary. The two significant in- mainly due to the new representatives. For example, the Theraphosidae creases in fossils included in our analysis (Early Cretaceous and and Ctenidae were polyphyletic. Telemidae was the sister-group to all Paleogene) almost certainly reflected the exceptional preservation and other Synspermiata except Dysderoidea and Caponiidae. sampling of the Burmese and Baltic ambers, respectively. Baltic amber

149 L. Shao, S. Li Molecular Phylogenetics and Evolution 127 (2018) 146–155

Fig. 1. Reconstructed Maximum-Likelihood topology of spider relationships inferred with dataset A. Bootstrap support values below 100% are indicated.

alone included > 65% of all fossil spider occurrences reported. 3.3. Ancestral state reconstruction Therefore, the presence or absence of amber fossil spiders strongly in- fluenced the analysis. The maximum likelihood ancestral state reconstruction of web-type exclusive of Hersiliidae showed that the common ancestor likely used a

150 L. Shao, S. Li Molecular Phylogenetics and Evolution 127 (2018) 146–155

Fig. 2. Evolutionary history of spiders. Divergence times were estimated from 790,000 trees sampled from trees separately generated for 79 partitions that included all taxa and the 95% HPD age of main crown nodes are indicated; sectors of pies at main crown nodes are proportional to the probabilities of each state at that node (detailed information in Fig. S11); the bolded stem shows the best shift configuration with the highest posterior probability from BAMM analysis; palaeogeographical reconstruction with spider fossil collections (red dots represent fossil locations and maps modified from fossilworks http://fossilworks.org/).

151 L. Shao, S. Li Molecular Phylogenetics and Evolution 127 (2018) 146–155

Fig. 3. Diversification dynamics of spiders. (a) Proportional lineages-through time (LTT) plot based on the median values for spiders and distribution of clade ages for crown group spi- ders. (b) Global pattern of spider family di- versification rate based on the fossil record (corresponding to the left coordinate), the ∼7- fold rate change around 145 Ma is indicated with an increase arrow; global pattern of spider family diversification rate based on BAMM analyses (red line, corresponding to the right coordinate). (c): Global warming events after Scotese (2016). Stratigraphical abbreviations: C = Carboniferous, P = Permian, Tr = Triassic, J = Jurassic, K = Cretaceous, Pg = Paleogene, and N = Neogene.

subterranean burrow, perhaps one sealed by a trapdoor. The araneo- analyses including the Hersiliidae differed at some nodes. The araneo- morph ancestors probably foraged from a stereotypical aerial sheet and morph ancestors probably foraged without a web and the probability of the ancestor of orbicularians likely used a orb web. RTA ancestors foraging without a web increased in orbicularians and RTA ancestors probably largely abandoned webs (Fig. S11a). However, the result of (Fig. S11b).

152 L. Shao, S. Li Molecular Phylogenetics and Evolution 127 (2018) 146–155

4. Discussion indicators and the Fast Ocean-Atmosphere model analyses have shown that the expansion of temperate and tropical biomes culminated during 4.1. Pattern of temporal diversification the Cenomanian occurred at the expense of deserts (Chaboureau et al., 2014). At the time of the Early Cretaceous greenhouse, fossil spider Different model evaluations serve to assess distinct patterns of sites are numerous and scattered over the globe. This likely reflects the spider diversification. BAMM detects rate shifts within and among expansion of temperate and tropical zones (Fig. 2). Our diversification clades while TreePar assesses global patterns. Therefore, our diversifi- analyses also indicate that spiders diversified during this period. cation analyses complement one another. BAMM analyses indicate a Warming climate appears to promote speciation. Salamanders under- shift of diversification rates at the node for Opisthothelae and another went rapid episodes of diversification and dispersal that coincided with notable increase during the Early Cretaceous, which corresponds with major global warming events during the Late Cretaceous and again the origin of the RTA clade (Figs. 2 and 3b). TreePar analyses indicate during the Paleocene–Eocene thermal optimum (Vieites et al., 2007). higher diversification rates of lineages before the early Late Cretaceous, Chin et al. (2014) indicated that the diversification of all major lineages but no other rate shift (Table S5). The initial plateau in the TreePar in the modern Prunus may have been triggered by the global analyses before the Late Cretaceous could owe to the software assuming warming period of the early Eocene. In addition, Fernández et al. uniform sampling and specifying an equal probability for each node to (2018) showed that higher temperatures are associated with higher be sampled regardless of node age (Stadler, 2013b). Specifically, the diversification rates. Therefore, the warm climatic condition appears to analysis can drastically underestimate extinction rates when a diversi- have played a major role in shaping the extreme diversity of spiders fied sample is treated as random or uniform; this leads to biased esti- during the Early Cretaceous. mates for diversification rates (Höhna et al., 2011). Nevertheless, the LTT plot shows a remarkable accumulation of lineages in the Early 4.3. Ancestral state reconstruction of predatory strategies Cretaceous. Most extant spider families appear to have split from their common ancestor before ∼95 Ma, which might lead to the subsequent Sampling taxa more densely appears to be a reliable way to improve downturn of diversification rate (Fig. 3a). ancestral character state estimates (Salisbury and Kim, 2001). Our Although the PyRate analysis shows a notable increase in diversi- study adds samples of 12 families not included before and includes most fication rate of spiders in the Early Cretaceous, this probably owes to typical types of predatory strategies. Therefore, our adding sampling the first occurrences of spider-bearing amber deposits in the fossil re- increases the credibility of the ancestral state reconstructions. For ex- cord (Figs. 3b and S10). Still, fossil data serve to assess relative abun- ample, by adding of Zodariidae and Sparassidae, which now places at dance of different groups. Before the Jurassic, fossil spiders involve only the base of RTA clade, the analyses obtain further support for the ab- the Mesothelae and Mygalomorphae. Fossil occurrences of the main sence of a web in the ancestor of the RTA clade, and with a higher spider groups Mygalomorphae, Haplogynae, Palpimanoidea, Orbicu- percentage (Garrison et al., 2016). However, phylogenetic uncertainty lariae, and the RTA clade began to appear after the beginning of the can bias the evolutionary transitions estimated from ancestral state Early Cretaceous (Fig. 4). Most Cretaceous fossil spiders assign to extant reconstruction (Duchêne and Lanfear, 2015). In agreement with pre- families and analyses predict the presence of many other extant spider vious work, the cribellate orb-weavers and a mainly free-living group families at the same point in time. These data suggest that the primary (Hersiliidae) closely associate with the RTA clade. This refutes the hy- interfamily occurred during the Cretaceous or earlier. pothesis of the orb web having single origin (Bond et al., 2014; Therefore, corresponding diversification analyses based on molecular Fernández et al., 2014, 2018; Garrison et al., 2016). However, the and fossil data provide insights into the temporal variation in biodi- placement of Oecobioidea remains unstable in our phylogenetic ana- versity. The results point towards a diversification of spider lineages in lyses and, thus, the webless state may have emerged earlier (Fig. S11) the Early Cretaceous greenhouse, at least at the family level. than thought. Therefore, dense sampling, particularly of basal and webless groups, helps locate the phylogenetic position of character state 4.2. Early Cretaceous climate change and spider diversification shift. Our analyses suggest that RTA ancestors probably largely aban- Major climate change followed the breakup of Pangaea from the doned webs to become cursorial hunters during the Early Cretaceous. A mid-Triassic to the Jurassic and Cretaceous (Fig. 3c). Lithological slight overestimation of divergence time for the major webless lineages

Fig. 4. Ratios of different groups of fossil spiders in every geological epoch.

153 L. Shao, S. Li Molecular Phylogenetics and Evolution 127 (2018) 146–155

(RTA clade) exists when compared with the earliest record of fossil of mass extinctions on Crocodyliformes evolutionary history. R. Soc. Open Sci. 2, taxa. The median is still younger than Fernández et al. (2018) proposed 140385. Chaboureau, A.C., Sepulchre, P., Donnadieu, Y., Franc, A., 2014. Tectonic-driven climate and the age interval is narrower than that of Garrison et al. (2016). change and the diversification of angiosperms. Proc. Natl. Acad. Sci. USA 111, Notwithstanding, most divergences of webless spiders appear to have 14066–14070. occurred during the Early Cretaceous. Further, the various groups of Chin, S.W., Shaw, J., Haberle, R., Wen, J., Potter, D., 2014. Diversification of almonds, peaches, plums and cherries – molecular systematics and biogeographic history of amber-fossil spiders in the Early Cretaceous indicate the existence of Prunus (Rosaceae). Mol. Phylogenet. Evol. 76, 34–48. many cursorial spiders (Fig. 4, e.g. Oonopidae). The transformation of Coiffard, C., Gomez, B., Daviero-Gomez, V., Dilcher, D.L., 2012. Rise to dominance of their foraging strategy during the Early Cretaceous appears to have angiosperm pioneers in European Cretaceous environments. Proc. Natl. Acad. Sci. – been successful given that today these are speciose families. USA 109, 20955 20959. Dimitrov, D., Benavides, L.R., Arnedo, M.A., Giribet, G., Griswold, C.E., Scharff, N., The webless predation strategy correlates with spider diversity as a Hormiga, G., 2016. Rounding up the usual suspects: a standard target-gene approach notable increase of diversification rate occurs in the RTA clade (Fig. 2). for resolving the interfamilial phylogenetic relationships of ecribellate orb-weaving fi Cursorial spiders encounter mainly less mobile prey while web-weavers spiders with a new family-rank classi cation (Araneae, ). 33, 221–250. basically wait for prey to come, which tend to be highly mobile species Drummond, A.J., Suchard, M.A., Xie, D., Rambaut, A., 2012. Bayesian phylogenetics with (Michalko and Pekãr, 2016). Miller et al. (2014) showed that active BEAUti and the BEAST 1.7. Mol. Biol. Evol. 29, 1969–1973. predators ranged farther distances across broader domains and killed Duchêne, S., Lanfear, R., 2015. Phylogenetic uncertainty can bias the number of evolu- tionary transitions estimated from ancestral state reconstruction methods. J. Exp. more preys than sit-and-wait spiders. Many insect groups appear to Zool., Part B 324 (6), 517. have diversified extensively during the KTR, such as ants (Moreau et al., Dunlop, J.A., Anderson, L.I., Kerp, H., Hass, H., 2004. A harvestman (Arachnida: 2006), butterflies and moths (Wahlberg et al., 2013) and the most Opiliones) from the Early Rhynie cherts, Aberdeenshire, Scotland. Trans. R. Soc. Edinburgh: Earth Sci. 94, 341–354. speciose beetles (Zhang et al., 2018). Therefore, the prey abundance Ebersberger, I., Strauss, S., von Haeseler, A., 2009. HaMStR: Profile hidden markov model could help explain the concurrent increase in diversification and the based search for orthologs in ESTs. BMC Evol. Biol. 9, 157. webless hunting strategy in the RTA clade. Engel, M.S., Grimaldi, D.A., 2004. New light shed on the oldest insect. Nature 427, 627–630. A robust hypothesis for the phylogeny of spiders permits evolu- Erwin, D.H., Valentine, J.W., 2010. The Cambrian Explosion: The Construction of Animal tionary investigation of their life history, ecology, and in Biodiversity. Roberts and Company, Greenwood Village, CO. terms of patterns of distribution and diversification. Our study provides Eskov, K.Y., Zonshtein, S., 1990. First Mesozoic mygalomorph spiders from the Lower a well-supported phylogeny of spiders that clarifies the relationships of Cretaceous of Siberia and Mongolia, with notes on the system and evolution of the infraorder Mygalomorphae (: Araneae). Neues Jb. Geol. Paläontol. Abh. all major lineages. We also show novel placement of some families with 178, 325–368. high support using different phylogenetic methods. The evolutionary Fernández, R., Hormiga, G., Giribet, G., 2014. Phylogenomic analysis of spiders reveals – time frame and assessment, and analysis of the fossil records imply that nonmonophyly of orb weavers. Curr. Biol. 24, 1772 1777. fi Fernández, R., Kallal, R.J., Dimitrov, D., Ballesteros, J.A., Arnedo, M.A., Giribet, G., although the global pattern is complex, spider lineage diversi cation Hormiga, G., 2018. Phylogenomics, diversification dynamics, and comparative occurred primarily during the Early Cretaceous. We speculate that the transcriptomics across the spider tree of life. Curr. Biol. 28 (9), 1489. warm climate of this time was important in shaping the extreme di- Grabherr, M.G., Haas, B.J., Yassour, M., Levin, J.Z., Thompson, D.A., Amit, I., Adiconis, X., Fan, L., Raychowdhury, R., Zeng, Q., et al., 2011. Full-length transcriptome as- versity of spiders, especially for the webless clade following the di- sembly from RNA-Seq data without a reference genome. Nat. Biotechnol. 29, versification of insect preys. 644–652. Hart, J.P., Xiong, C., Wang, D., Wang, Q., Benton, M.J., Xue, J., Meng, M., Zhao, Q., Zhang, J., 2013. Diversity dynamics of Silurian-Early Carboniferous land plants in Acknowledgements South China. PLoS ONE 8, e75706. Höhna, S., Stadler, T., Ronquist, F., Britton, T., 2011. Inferring speciation and extinction The manuscript benefited greatly from comments by Miquel A. rates under different sampling schemes. Mol. Biol. Evol. 28, 2577–2589. Jocqué, R., Dippenaar-choeman, A.S., 2006. Spider Families of the World. Royal Museum Arnedo and three anonymous reviewers. We thank Zhonge Hou and Zhe for Central Africa, Tervuren. Zhao for their valuable comments on an earlier version of this article. Katoh, K., Standley, D.M., 2016. A simple method to control over-alignment in the MAFFT We thank Matthew Clapham for the detailed comments and suggestions multiple sequence alignment program. Bioinformatics 32, 1933–1942. fl on the analysis of fossil data. We thank Yuri Marusik for the species Kohli, M.K., Ware, J.L., Bechly, G., 2016. How to date a dragon y: fossil calibrations for odonates. Palaeontol. Electron. 19, 1–14. identification used in our study. We thank Jinzhong Fu, Bin Lu, and Kück, P., Meusemann, K., 2010. FASconCAT: convenient handling of data matrices. Mol. Weizhao Yang for their technical assistance in data process. This study Phylogenet. Evol. 56, 1115–1118. was supported by the National Natural Science Foundation of China Kück, P., Meusemann, K., Dambach, J., Thormann, B., von Reumont, B.M., Wägele, J.W., Misof, B., 2010. Parametric and non-parametric masking of randomness in sequence (NSFC-31530067, 31471960 to S. L.). alignments can be improved and leads to better resolved trees. Front. Zool. 7, 10. Kück, P., Struck, T.H., 2014. BaCoCa – a heuristic software tool for the parallel assessment Appendix A. Supplementary material of sequence biases in hundreds of gene and taxon partitions. Mol. Phylogenet. Evol. 70, 94–98. Labandeira, C.C., 2005. Invasion of the continents: cyanobacterial crusts to tree-in- Supplementary data associated with this article can be found in the habiting arthropods. Trends Ecol. Evol. 20, 253–262. online version, at https://doi.org/10.1016/j.ympev.2018.05.026. Lanfear, R., Calcott, B., Ho, S.Y., Guindon, S., 2012. PartitionFinder: combined selection of partitioning schemes and substitution models for phylogenetic analyses. Mol. Biol. Evol. 29, 1695–1701. References Lartillot, N., Rodrigue, N., Stubbs, D., Richer, J., 2013. PhyloBayes MPI phylogenetic reconstruction with infinite mixtures of profiles in a parallel environment. Syst. Biol. 62, 611–615. Becker, R.A., Chambers, J.M., Wilks, A.R., 1988. The New S Language. Wadsworth & Lloyd, G.T., Davis, K.E., Pisani, D., Tarver, J.E., Ruta, M., Sakamoto, M., Hone, D.W., Brooks/Cole, Pacific Grove, CA. Jennings, R., Benton, M.J., 2008. and the Cretaceous terrestrial revolution. Blackledge, T.A., Scharff, N., Coddington, J.A., Szüts, T., Wenzel, J.W., Hayashi, C.Y., Proc. R. Soc. B 275, 2483–2490. Agnarsson, I., 2009. Reconstructing web evolution and spider diversification in the Meyer, B., Meusemann, K., Misof, B., 2011. MARE: matrix reduction – a tool to select molecular era. Proc. Natl. Acad. Sci. USA 106, 5229–5234. optimized data subsets from supermatrices for phylogenetic inference, Bonn Bollback, J.P., 2006. SIMMAP: stochastic character mapping of discrete traits on phylo- (Germany): Zentrum für Molekulare Biodiversitätsforschung (zmb) am ZFMK. genies. BMC Bioinf. 7, 88. Version 01.2-rc. Available at < http://mare.zfmk.de >. Bond, J.E., Garrison, N.L., Hamilton, C.A., Godwin, R.L., Hedin, M., Agnarsson, I., 2014. Michalko, R., Pekãr, S., 2016. Different hunting strategies of generalist predators result in Phylogenomics resolves a spider backbone phylogeny and rejects a prevailing para- functional differences. Oecologia 181 (4), 1187–1197. digm for orb web evolution. Curr. Biol. 24, 1765–1771. Miller, J.R.B., Ament, J.M., Schmitz, O.J., 2014. Fear on the move: predator hunting Bond, J.E., Hendrixson, B.E., Hamilton, C.A., Hedin, M., 2012. A reconsideration of the mode predicts variation in prey mortality and plasticity in prey spatial response. J. classification of the spider infraorder Mygalomorphae (Arachnida: Araneae) based on Anim. Ecol. 83 (1), 214–222. three nuclear genes and morphology. PLoS One 7, e38753. Miller, M.A., Pfeiffer, W., Schwartz, T., 2010. Creating the CIPRES Science Gateway for Bond, J.E., Opell, B.D., 1998. Testing and key innovation hypotheses inference of large phylogenetic trees. In: Proceedings of the Gateway Computing in spiders. Evolution 52, 403–414. Environments Workshop (GCE), 14 Nov 2010, New Orleans, LA, pp. 1–8. Bronzati, M., Montefeltro, F.C., Langer, M.C., 2015. Diversification events and the effects Mirarab, S., Reaz, R., Bayzid, M.S., Zimmermann, T., Swenson, M.S., Warnow, T., 2014.

154 L. Shao, S. Li Molecular Phylogenetics and Evolution 127 (2018) 146–155

ASTRAL: genome-cale coalescent-based species tree estimation. Bioinformatics 30, Deinopoidea) from China. J. Syst. Palaeontol. 14, 1–26. i541–i548. Sharma, P.P., Giribet, G., 2014. A revised dated phylogeny of the order Misof, B., Liu, S., Meusemann, K., Peters, R.S., Donath, A., Mayer, C., Frandsen, P.B., Opiliones. Front. Genet. 5. Ware, J., Flouri, T., Beutel, R.G., et al., 2014. Phylogenomics resolves the timing and Silvestro, D., Cascales-Minana, B., Bacon, C.D., Antonelli, A., 2015. Revisiting the origin pattern of insect evolution. Science 346, 763–767. and diversification of vascular plants through a comprehensive Bayesian analysis of Moradmand, M., Schonhofer, A.L., Jager, P., 2014. Molecular phylogeny of the spider the fossil record. New Phytol. 207, 425–436. family Sparassidae with focus on the genus Eusparassus and notes on the RTA-clade Silvestro, D., Schnitzler, J., Liow, L.H., Antonelli, A., Salamin, N., 2014. Bayesian esti- and 'Laterigradae'. Mol. Phylogenet. Evol. 74, 48–65. mation of speciation and extinction from incomplete fossil occurrence data. Syst. Moreau, C.S., Bell, C.D., Vila, R., Archibald, S.B., Pierce, N.E., 2006. Phylogeny of the Biol. 63, 349–367. ants: diversification in the age of angiosperms. Science 312, 101–104. Stadler, T., 2013a. TreePar: Estimating Birth and Death Rates Based on Phylogenies. O’Leary, M.A., Bloch, J.I., Flynn, J.J., Gaudin, T.J., Giallombardo, A., Giannini, N.P., Comprehensive R Archive Network, Vienna (Austria) Available from: < http:// Goldberg, S.L., Kraatz, B.P., Luo, Z., Meng, J., et al., 2013. The placental mammal CRAN.R–project.org/package=TreePar > . ancestor and the post-K-Pg radiation of placentals. Science 339, 662–667. Stadler, T., 2013b. How can we improve accuracy of macroevolutionary rate estimates? Paradis, E., 2012. Analysis of Phylogenetics and Evolution with R. Springer Science & Syst. Biol. 62, 321–329. Business Media, New York. Stamatakis, A., 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis Penney, D., Selden, P.A., 2002. The oldest linyphiid spider, in lower Cretaceous Lebanese of large phylogenies. Bioinformatics 30, 1312–1313. amber (Araneae, Linyphiidae, Linyphiinae). J. Arachnol. 30, 487–493. Telford, M.J., Lowe, C.J., Cameron, C.B., Ortega-Martinez, O., Aronowicz, J., Oliveri, P., Prum, R.O., Berv, J.S., Dornburg, A., Field, D.J., Townsend, J.P., Lemmon, E.M., Lemmon, Copley, R.R., 2014. Phylogenomic analysis of echinoderm class relationships supports A.R., 2015. A comprehensive phylogeny of birds (Aves) using targeted next-genera- Asterozoa. Proc. R. Soc. B 281, 20140479. tion DNA sequencing. Nature 526, 569–573. Vieites, D.R., Min, M.S., Wake, D.B., 2007. Rapid diversification and dispersal during Rabosky, D.L., Donnellan, S.C., Grundler, M., Lovette, I.J., 2014. Analysis and visuali- periods of global warming by plethodontid salamanders. Proc. Natl. Acad. Sci. USA zation of complex macroevolutionary dynamics: an example from Australian scincid 104, 19903–19907. lizards. Syst. Biol. 63, 610–627. Wahlberg, N., Wheat, C.W., Pena, C., 2013. Timing and patterns in the taxonomic di- Rambaut, A., Suchard, M.A., Xie, D., Drummond, A.J., 2014. Tracer v1.6, Available versification of Lepidoptera (butterflies and moths). PLoS One 8, e80875.

from < http://tree.bio.ed.ac.uk/software/tracer >. Wang, Y., Huang, C., Sun, B., Quan, C., Wu, J., Lin, Z., 2014. Paleo-CO2 variation trends Revell, L.J., 2012. phytools: an R package for phylogenetic comparative biology (and and the Cretaceous greenhouse climate. Earth-Sci. Rev. 129, 136–147. other things). Methods Ecol. Evol. 3, 217–223. Wheeler, W.C., Coddington, J.A., Crowley, L.M., Dimitrov, D., Goloboff, P.A., Griswold, Salisbury, B.A., Kim, J., 2001. Ancestral state estimation and taxon sampling density. C.E., Hormiga, G., Prendini, L., Ramírez, M.J., Sierwald, P., et al., 2016. The spider Syst. Biol. 50 (4), 557. tree of life: phylogeny of Araneae based on target-gene analyses from an extensive Scotese, C.B., 2016. A New Global Temperature Curve for the . GSA Annual taxon sampling. Cladistics 33, 574–616. Meeting in Denver, Colorado, USA. Wood, H.M., Griswold, C.E., Gillespie, R.G., 2012. Phylogenetic placement of pelican Scourfield, D.J., 1940. The oldest known fossil insect (Rhyniella praecursor Hirst & spiders (Archaeidae, Araneae), with insight into evolution of the “neck” and pre- Maulik)-further details from additional specimens. Proc. Linn. Soc. 152, 113–131. datory behaviours of the superfamily Palpimanoidea. Cladistics 28, 598–626. Selden, P.A., 1996. First fossil mesothele spider, from the Carboniferous of France. Rev. Wunderlich, J., 2004. On selected higher and lower taxa of fossil and extant spiders of the Suisse Zool. 2, 585–596. superfamily Oecobioidea, with a provisional cladogram (Araneae: Hersiliidae and Selden, P.A., Huang, D., 2010. The oldest haplogyne spider (Araneae: Plectreuridae), from Oecobiidae). In: Wunderlich, J. (Ed.), Beiträge zur Araneologie, pp. 809–848. the Middle Jurassic of China. Naturwissenschaften 97, 449–459. Wunderlich, J., Milki, R., 2004. Description of the extinct new subfamily Selden, P.A., Shih, C.-K., Ren, D., 2013. A giant spider from the Jurassic of China reveals Microsegestriinae (Araneae: Segestriidae) in Cretaceous Lebanese amber. Beiträge greater diversity of the orbicularian stem group. Naturwissenschaften 100, zur Araneologie 3b:1867–1873. 1171–1181. Zhang, S.Q., Che, L.H., Li, Y., Dan, L., Pang, H., Ślipiński, A., Zhang, P., 2018. Selden, P.A., Anderson, J.M., Anderson, H.M., Fraser, N.C., 1999. Fossil araneomorph Evolutionary history of coleoptera revealed by extensive sampling of genes and spiders from the Triassic of South Africa and Virginia. J. Arachnol. 27, 401–414. species. Nat. Commun. 9 (1), 205. Selden, P.A., Gall, J.-C., 1992. A Triassic mygalomorph spider from the northern Vosges, Zhong, M., Hansen, B., Nesnidal, M., Golombek, A., Halanych, K.M., Struck, T.H., 2011. France. Palaeontology 35, 211–235. Detecting the symplesiomorphy trap: a multigene phylogenetic analysis of terebelli- Selden, P.A., Penney, D., 2010. Fossil spiders. Biol. Rev. 85, 171–206. form annelids. BMC Evol. Biol. 11, 369. Selden, P.A., Ren, D., Shih, C.-K., 2015. Mesozoic cribellate spiders (Araneae:

155