Arachnids are an important group of terrestrial Arthropods, including the familiar Ticks, Mites, Spiders, and Scorpions, together with Pseudoscorpions, Camel Spiders, Vinegaroons, Whip Spiders, and a few other groups. Arachnids are important predatory Arthropods across almost every conceivable terrestrial habitat. While Ticks are ectoparasites that affect Humans and livestock, Spiders are ecologically the most successful Arachnids and as predators consume vast quantities of Insects. Thus, understanding when Arachnids colonised land and diversified is of interest from a macroevolutionary and macroecological perspective. Arachnids are chelicerates, together with the marine Horseshoe Crabs (Xiphosura) and Sea Spiders (Pycnogonida). They are the most speciose clade in Chelicerata, with more than 112 000 described extant species. Together with Hexapods and Myriapods, Arachnids represent one of three distinct and ancient events of Arthropod terrestrialisation (terrestrial Isopods are a younger addition to the continental Arthropod biota). While Arachnids have traditionally been considered monophyletic and terrestrial (apart from secondarily marine Mites) this picture has been challenged at different times. Scorpions were long thought to be the sister group of all other extant Arachnids or most closely allied to the aquatic 'Sea Scorpions', the Eurypterids. Early fossil Scorpions have been interpreted as aquatic, and in some cases even as marine. These phylogenetic hypotheses and interpretations of fossil ecology have been seen as requiring independent events of terrestrialisation. Some of these views have been overturned by strong molecular and morphological evidence for Scorpions being nested within the Arachnida as the sister group of the other Arachnids with book lungs, the Tetrapulmonata. Indeed, detailed correspondences in book lung morphology between Scorpions and tetrapulmonates support their homology. The supposed aquatic mode of life of various fossil Scorpions has also been questioned on both morphological and geological grounds, Another challenge to a single terrestrialisation event in arachnids came from analyses of phylogenomic datasets, which have often recovered the marine Xiphosura to be nested within Arachnida. This remains a contentious issue, as other phylogenomic analyses have yielded trees in which Arachnida is monophyletic.
Most Chelicerate lineages are predatory components of a diverse range of ecosystems, and the rock record attests to their presence in both earlier Palaeozoic marine settings and through into the Mesozoic and Cainozoic, which witnessed a prolific diversification of Spiders and other terrestrial Arachnids. The terrestrial rock record prior to the Silurian is very sparse and has presented some apparent discordances when investigating Myriapod, Hexapod, and Plant divergence times. While the body fossil record of terrestrial Plants and Arthropods does not extend much further back than the Silurian (roughly 443–419 million years ago), molecular clock estimates go back to the Ordovician (485–443 million years ago) and Cambrian (538–485 million years ago). However, likely Plant spores with desiccation-resistant adaptations extend back to the middle Cambrian. The fossil record presents no unequivocal evidence for crown-group Arachnids before the Silurian. The oldest crown-group Arachnids are of Silurian age (stem-group Scorpiones in the Llandovery), followed by the extinct Trigonotarbida in the late Silurian (early Přídolí), Acariformes and Opiliones in the Early Devonian (Pragian), and Pseudoscorpiones in the Middle Devonian (Givetian). Several other Arachnid orders first appear in the Carboniferous, including Araneae, Uropygi, Amblypygi, and Ricinulei. In contrast to a picture of scattered branches of the Arachnid crown-group first appearing in the Siluro–Devonian, older representatives of the Arachnid lineage are stem-group Arachnida, and are marine shoreline or brackish water/estuarine forms rather than being terrestrial. These include Chasmataspida and Eurypterida, the earliest members of which date to the Miaolingian Series of the Cambrian (Drumian Stage) and the Late Ordovician (Sandbian), respectively. Xiphosura-like Chelicerates have a good fossil record, showing considerable morphological stasis, with marine stem-group representatives of Xiphosura such as Lunataspis being documented from the Late Ordovician, about 445 million years ago, and a species from the Early Ordovician (Tremadocian) of Morocco extends the lineage’s history even deeper. With such a deep history revealed by the fossil record, any inferred phylogenetic position for Xiphosura within terrestrial Arachnids would imply that the marine ecology of this lineage should be a secondary acquisition. Although such a scenario is palaeontologically unlikely, molecular studies have often recovered Horseshoe Crabs in highly derived clades of Arachnids, such as sister groups to Opiliones or Palpigradi, or Scorpiones and Araneae.
As with other terrestrial groups, molecular dating has recovered old dates for the origin and main diversification of Arachnids. As part of wider campaigns investigating Arthropods using just a few Arachnid representatives, several previous syudies have recovered dates for the origin of Arachnida with credibility intervals bracketed between the Cambrian, in the first two studies, and Ordovician in the latter. One recent study reported a Chelicerate molecular phylogeny in which when they constrained Arachnida to be monophyletic inferred an Ediacaran origin for the group. Consequently, there are significant geochronological discrepancies, particularly for terrestrial lineages, between the molecular clock-based studies and the younger dates suggested by the first appearances of fossils. As fossils do not inform on the age of origin of clades, but rather provide minimum ages of divergence, clock-based methods are required to approach an accurate evolutionary timescale.
In a paper published in the journal Frontiers in Genetics on 11 March 2020, Jesus Lozano-Fernandez of the School of Biological Sciences and School of Earth Sciences at the University of Bristol, Alastair Tanner, also of the School of Biological Sciences at the University of Bristol, Mark Puttick of the Department of Biology and Biochemistry at the University of Bath, Jakob Vinther, again of the School of Biological Sciences and School of Earth Sciences at the University of Bristol, Gregory Edgecombe of the Department of Earth Sciences at the Natural History Museum, and Davide Pisani, once again of the School of Biological Sciences and School of Earth Sciences at the University of Bristol, present the results of a study in which they estimate the divergence time of Arachnids, using a method which previously recovered both Arachnida and Acari as monophyletic groups. To calibrate the molecular clock, Lozano-Fernandez et al. used a carefully selected and expanded set of 27 fossil constraints across the tree.
The molecular supermatrix used by Lozano-Fernandez et al. is composed of 89 species, 75 of them being Chelicerates, with 14 other Panarthropod species as outgroups. This matrix is a concatenation of 233 highly conserved and slow-evolving genes retrieved from transcriptomic data (45,939 amino acid positions and 78,1% complete). To evaluate the robustness of the results to an alternative topology, Lozano-Fernandez et al. also performed a divergencedate analysis in which Arachnida was non-monophyletic, with Xiphosura nested inside the Arachnids. The phylogenetic trees were inferred using PhyloBayes MPI v.4.1 under the site-heterogeneous CAT–GTR C 0 model of amino acid substitution. Convergence was assessed by running two independent Markov chains and using the bpcomp and tracecomp tools from PhyloBayes to monitor the maximum discrepancy in clade support (maxdiff), the effective sample size (effsize), and the relative difference in posterior mean estimates (rel_diff) for several key parameters and summary statistics of the model. We ran the analysis for 10 000 cycles and discarded as 'burn-in' the first 3000 generations.
Divergence time estimation was performed using PhyloBayes 3.3f (serial version). Lozano-Fernandez et al. compared the fit of alternative, autocorrelated and uncorrelated, relaxed molecular clock models generating ten different random splits replicates and performing cross-validation analyses. The tree was rooted on the Onychophora–Euarthropoda split. A set of 27 fossil calibrations and 1 node constrained by a maximum age was used. Lozano-Fernandez et al.imposed a soft maximum of 559 million years ago for the Onychophoran–Euarthropod split based on trace fossils in the White Sea/South Australian Ediacaran. This uses a radiometric date of 558 million years for strata at which body fossils such as Kimberella, a putative total-group Bilaterian Metazoan, occur. Metazoan trace fossils in the White Sea/South Australian Ediacaran indicate suitable preservation for Arthropod traces, were they present. Lozano-Fernandez et al. regard this to be a conservative soft maximum, as there is no body or trace fossil evidence for Arthropods in the Ediacaran. A minimum for the divergence of Onychophorans and Arthropods is set by the earliest Rusophycus traces (total-group Arthropoda), dated to a minimum of 528.8 million years ago. To allow the analysis to explore younger ages and prevent having posterior ages being much older than the fossil record, Lozano-Fernandez et al. also set maximum constraints on a few of the deepest calibrations within Euarthropoda. Lozano-Fernandez et al. infer that crown-group Mandibulata and Chelicerata do not predate the oldest fossil evidence for Arthropods (Rusophycus) and set soft maxima for each of this pair of sister taxa at the base of the Cambrian (538.8 million yars ago). Within Arachnida, Lozano-Fernandez et al. infer Acari and Arachnopulmonata do not predate the oldest body fossils of crown-group Chelicerata, using Wisangocaris barbarahardyae and its date of 509 million years old as a soft maximum. The amino acid substitution model used to estimate branch lengths was the CAT–GTR C 0 model.All analyses were conducted using soft bounds with 5% of the probability mass outside the calibration interval. A birth–death model was used to define prior node ages. Analyses were run under the priors to evaluate the effective joint priors induced by our choice of calibrations and root maxima. Convergence was considered achieved with tracecomp statistics dropping below 1 for all relative difference scores, and all effective sample sizes being above 50, for all chain parameters. The time-scaled phylogenies were plotted using the package MCMCtreeR, which allows the display of full posterior distributions on nodes and the inclusion of the geological timescale. We included as supplementary data the chronograms, the guiding trees and the calibration file used in PhyloBayes, the subset of sampled timescaled trees used to generate the posterior distributions shown on the figures.
Lozano-Fernandez et al. estimated rates of molecular evolution within Chelicerata using two different methods. For the first method, they modeled the rates of molecular evolution on a fixed tree topology constrained to the timetree relationships. On this tree they estimated relative branch lengths under the C60 model C 0 in IQTree. Lozano-Fernandez et al. then divided these relative branch lengths by the timetree lengths to provide an estimate of absolute molecular rates through time. For the second method, they inferred ancestral estimates of the amino acid sequence on the fixed timetree, again using the C60 model C 0 in IQTree. Lozano-Fernandez et al. divided the sum of gross amino acid changes between ancestral and descendant nodes by absolute time to obtain per-branch rates of change.
Lozano-Fernandez et al. estimated speciation and extinction rates on the fixed timetree by using a Bayesian episodic diversification rate model in RevBayes 1.0.10. This model estimates piece-wise rates of speciation and extinction on a phylogeny through time. Within each bin, rates of speciation are equal but can differ between bins. The initial episodic speciation and extinction rate was sampled from a log-uniform distribution U(-10,10). Moving backward in time for each distinct time bin, the model samples speciation and extinction rate from a normal distribution with the mean inherited from the value of the previous bin so rates are autocorrelated. Each normal distribution has a standard deviation inferred from an exponential hyper-prior of mean 1. In this manner, the model follows a Brownian motion pattern of rate change through time. To incorporate incomplete sampling in the model, we provided estimates of the known extant species numbers to complement the diversity shown in the tree using empirical taxon sampling by providing estimate diversity represented by each tip on our incomplete time tree. This empirical taxon-sampling approach is believed to produce less biased estimates of speciation and extinction parameters compared to diversified taxon sampling. Lozano-Fernandez et al. used values of the described extant species: Pycnogonida (1346); Xiphosura (4); Ricinulei (77); Opiliones (6571); Solifugae (1116); Acariformes (42233); Parasitiformes (12385); Pseudoscorpiones (3574); Scorpiones (2109); Uropygi (119); Amblypygi (172); and Araneae (44863). As Lozano-Fernandez et al. tested for the presence of early high rates compared to later times rather than differences in geological time units. Lozano-Fernandez et al. assumed there were 10 equally sized time intervals which can potentially possess distinct speciation and extinction rates.
In the topology tree used for the molecular clock analyses by Lozano-Fernandez et al. Chelicerata and Euchelicerata are monophyletic, with the Horseshoe Crabs retrieved as the sister group of monophyletic Arachnida. Bayesian cross-validation indicates that the autocorrelated CIR model most optimally fits the data. Accordingly, divergence time estimation was performed using the Autocorrelated CIR model.
The age of the Euarthropoda root, given the taxonomic sample, is recovered near the end of the Ediacaran, 546 million years ago, with the 95% highest posterior density lying between 551 and 536 million years ago. Chelicerata are inferred to originate at 535 million years ago (with highest posterior density 540–527 million years ago), similar to the age retrieved for Mandibulata 535 million years ago (539–526 million years ago). The origin of Myriapoda comprises ages centered on the early Cambrian 516 million years ago (524–505 million years ago) and precedes that of Pancrustacea at 486 million years ago (501–471 million years ago). Hexapods are inferred to be much younger in age than Myriapods, ranging through the Late Ordovician to Early Devonian, 422 million years ago (448–400 million years ago).
Arachnid terrestrialisation is inferred to date to the Cambrian to Ordovician, crown-group Arachnida having a mean at 485 million years ago (494–475 million years ago). Therefore, Lozano-Fernandez et al.'s results support a Cambrian or Early Ordovician origin of two of the three main terrestrial Arthropod lineages (Myriapods and Arachnids). The upper limit is consistent with fossil evidence for stem-group Arachnida (Chasmataspidid trackways) but is substantially older than any crown-group fossils. Within Arachnida, rapid cladogenesis then occurred during the 20 million years that followed their origin, with several crown-group supra-ordinal clades becoming established in this time interval. By around 450 million years ago, all 10 stem groups leading to extant orders of Chelicerates included in the analysis (out of 12 in total, Palpigrada and Schizomida are unsampled) were already established. Further cladogenesis is inferred to have involved a more gradual tempo of evolution, in particular for Arachnopulmonata and Acari, which originated at 470 million years ago but greatly expanded after the start of the Mesozoic (252 Ma to 66 million years ago). Lozano-Fernandez et al.'s dating suggests that the oldest crown-group Arachnid orders are Opiliones and Parasitiformes, with Silurian and Devonian origins, respectively. Crown-group Scorpions have a Devonian to Carboniferous origin, with the sampled extant lineages splitting more recently. For Araneae, the crown-group age is centered on the Devonian–Carboniferous boundary, with most extant Mygalomorph and Araneomorph lineages diversifying after the Jurassic.
In general, these Palaeozoic age estimates for deep nodes within the most intensely sampled Arachnid orders are similar to those inferred in other recent molecular dating analyses. For example, Lozano-Fernandez et al.'s estimates for crown-group Araneae is consistent with the Late Devonian or Early Carboniferous dates retrieved in other transcriptome-based analyses; likewise, a Carboniferous mean age for crown-group Opistothele Spiders is found in each of these studies. Lozano-Fernandez et al.'s estimates encompassing a Late Ordovician median age for crown-group Opiliones corresponds to that estimated using tip dating, whereas node calibration in that study recovered a Silurian median. In the case of Scorpiones, a Late Devonian to Carboniferous origin of the crown group is closely comparable to the date for the same node, but older than the strictly Carboniferous ages estimated by a previous study. However, in all cases mentioned the credibility intervals substantially overlap, indicating that these independent studies found results that, despite some differences, are not significantly different and corroborate each other. One exception is from a recent phylotranscriptomic study of Pseudoscorpiones, which retrieved an Ordovician to Carboniferous origin for the group, significantly older than the Permian ages retrieved by Lozano-Fernandez et al. This may reflect the much more complete taxonomic coverage of Pseudoscorpion diversity in that study.
Electrochelifer balticus, a Pseudoscorpion from Eocene Baltic Amber. Scale bar is 1 mm. Harms & Dunlop (2017).
To assess whether our joint prior assumptions were driving their posterior estimates, Lozano-Fernandez et al. also ran the analysis under the priors (i.e. they performed analyses without data) and found that the joint priors allowed a wide possible distribution of ages, for the most part encompassing but not enforcing the posteriors. Lozano-Fernandez et al. also performed a molecular clock analysis from a different matrix that resulted in a topology in which Xiphosura was nested within Arachnida, specifically as the sister group of Arachnopulmonata plus Pseudoscorpiones. Overall, the result it is in general agreement with the main analysis, with most significant discrepancies concerning the age of Pycnogonida, which encompasses Silurian to Devonian ages, whereas in the main analysis are centered on the Carboniferous.
Lozano-Fernandez et al. conducted estimations of molecular evolution and diversification rates based on their Chelicerate timetree in an attempt to clarify whether the explosive cladogenesis at the onset of the Arachnid radiation early in the Phanerozoic was matched by an increase in either of these rates. The analyses of rates of molecular evolution along the branches show a very high rate early in Chelicerate history, including at the origin of Euchelicerata. These rates remain high during the early radiation of the Arachnida until the end of the Cambrian. Molecular rate estimations using branch lengths or ancestral sequences under a non-clock model gave nearly identical results. Using the episodic model of speciation and extinction rates through time, Lozano-Fernandez et al. found no evidence of high rates of speciation during the Cambrian, the period that presents the highest rates of molecular evolution. Instead, there is evidence for higher rates of cladogenesis later, bracketed between the Permian and Early Cretaceous, but especially high in the Permian and Triassic.
Molecular clocks allow the reconstruction of evolutionary timescales, but the reliability of these timescales depends on a variety of assumptions, which includes fossil data that have robust stratigraphic and phylogenetic justification, the use of a robust phylogenetic framework for the extant taxa, and the use of well-fitting models of both amino acid substitution and change in rate of molecular evolution. In this context, fossil calibrations then provide minimum ages for the origin of crown groups. Lozano-Fernandez et al. report divergence times on a well-sampled phylogeny, using the best-fitting molecular substitution and relaxed molecular clock models Fernandez et al. therefore contend that their findings provide the currently most robust insights into early Chelicerate evolution. In Fernandez et al.'s analysis, the ancestral Pycnogonid divergence from Euchelicerata is inferred to have happened early in the Cambrian. This does not greatly predate the oldest unequivocal total-group Pycnogonid, Cambropycnogon klausmuelleri, from the late Cambrian Orsten Konservat-Lagerstätte. The from the late Cambrian Orsten Konservat-Lagerstätte. The Pycnogonid–Euchelicerate divergence date suggests cryptic evolution of the Euchelicerate stem group in the early Cambrian. Chelicerate, and, indeed, Arthropod, body fossils are lacking in the earliest Cambrian, the Fortunian, the Arthropod fossil record in the first 20 million years of the Cambrian being limited to trace fossils. Subsequently, it is estimated that Xiphosurids diverged from Arachnids in the late Cambrian, followed soon after by the radiation of crown-group Arachnida. While revising this paper, a new study on Spider fossil calibrations came out and suggests that a few of the shallower calibrations used within Araneae treated as crown-groups may instead be stem-groups, so Fernandez et al. add this caveat when interpreting the age of Spiders.
There remain major geochronological discrepancies between the inferred molecular and fossil age of the various terrestrial Arthropod groups. While these discrepancies may be thought to question the accuracy of molecular clocks, the differences need to account for pervasive biases in the terrestrial sedimentary rock record. It has been noted that in Euramerica (from which much of the data on early terrestrial arthropods and early Plant megafossils are derived), terrestrial sediments are rare before the late Silurian, and first become widespread in the Early Devonian. This temporal bias in the rock record almost certainly affects the fossil records of terrestrial organisms, and likely accounts for a major component of the discordance between molecular and fossil dates. The common recovery of Horseshoe Crabs as ingroup Arachnids is perhaps unsurprising, given the short molecular branch lengths among these nodes in the tree, which also suggest short divergence times. With a reasonably good Xiphosurid and Pycnogonid fossil record, the molecular clock is well constrained among Euchelicerates. This is evidenced by the short credibility intervals among deep Arachnid nodes.
From an ecological context, it has been suggested that appreciably complex terrestrial ecosystems may have existed as far back as 1 billion years ago, with molecular dating suggesting that crown-group land plants were already present by the middle Cambrian. If it is indeed the case that Myriapods and Arachnids were on land so early, we speculate that the Animals may have been early grazers on littoral bacterial mats, or predated on other amphibious or terrestrial organisms. These ecologies represent habitats highly unfavorable to fossilisation, such as high-energy environments characterized by erosion rather than deposition. It is unsurprising that paleontological insight is thus limited, and inference of the
molecular kind as used here becomes more important as an investigative tool.
Fernandez et al. estimate that Arachnids colonised the land near the Cambrian–Ordovician boundary, and diversified soon after. Rates of molecular evolution were high at the onset of Arachnida, coinciding with rapid cladogenesis. High rates are concentrated on the branches leading to the major clades within Arachnida, representing major morphological and ecological partitions within the group. In unusually large and ancient clades, such as Chelicerates, it is expected to find high rates of molecular evolution in their early lineages and Fernandez et al. retrieved results in agreement with that expectation. In order to avoid biases related to that fact, they imposed on the molecular clock analyses several maxima on the deepest nodes to account for possible overestimations of divergence times. Arachnids are predominantly predators, which must reflect the presence of an already diverse ecosystem, which the slightly older divergence times for Myriapods and Embryophytes established in the middle Cambrian. Fernandez et al. therefore suspect that Arachnids primitively represent carnivorous Arthropods rather than having adapted to this mode of life convergently several times. The carnivorous Centipede (Chilopoda) crown group is separated from other Myriapods by a long branch estimated to be much younger in age than the detritivorous and/or fungal-feeding Progoneate Myriapods, allowing Arachnids to be potentially the first carnivorous Animals on land with early Myriapod lineages as a likely source of prey. Hexapod divergence estimates are generally younger, suggesting a colonisation of land no earlier than the Ordovician.
Mesoproctus rowlandi, a Whip Scorpion from the Cretaceous Santana Formation of Brazil. Scale bar is 1 cm. Dunlop (1998).
There is a clear contrast in evolutionary tempo after the explosive radiation of the Cambrian and Ordovician. More gradual cladogenesis characterises later Phanerozoic macroevolutionary dynamics of Chelicerates, as is seen in the origins of ordinal clades. Fernandez et al. 's diversification studies reveal an increase in speciation rates bracketed between the Permian and the Early Cretaceous, in the origin of most sub-ordinal clades, with no evidence of higher speciation rates coinciding with the early rapid Arachnid cladogenesis. A heightened diversification of Spiders during the Cretaceous has previously been detected, suggested to result from the rise of Angiosperms, stimulated by a warmer climate that led to the proliferation of Spiders’ main prey, Insects. Interestingly, Fernandez et al. did not observe an early burst of diversification at the origin of Chelicerates followed by a slowdown toward the present, a statistical bias usually found in large clades that survive to the present, the so-called 'push of the past'. Instead, it seems that speciation rates are decoupled from the rates of molecular change. The common origin of Arachnids giving rise to a plethora of adaptations, together with high molecular rates on the short internodes at the origin of the group suggests an ancient adaptive radiation shortly after colonising the land, but our diversification analyses have not detected higher speciation rates at that time, one of the key features signaling an adaptive radiation. Fernandez et al. acknowledge that the taxon sampling may not be the most adequate to infer speciation rates, as it was originally designed to maximize diversity, particularly at the deepest nodes to resolve the splits at the ordinal level.
Fernandez et al.'s analysis corroborates euchelicerates having radiated in the Cambrian and Arachnids having diversified rapidly in the latest Cambrian–Early Ordovician. While this radiation was rather fast, Fernandez et al. found no evidence that the speciation rates that underpinned it were explosive. The late Cambrian to Early Ordovician emergence of Arachnid stem groups onto land was soon followed by a rapid radiation near that same geological boundary, cladogenesis coinciding with high rates of molecular evolution during that time. A later phase of diversification within Arachnida is detected between the Permian and Early Cretaceous, during which the living arachnid orders exhibit heightened rates of speciation.
See also...
Follow Sciency Thoughts on Facebook