A new hypothesis of the relationships between arachnomorph arthropods, and the origin of chelicerates, is presented based on a cladistic analysis of 34 taxa and 54 characters. The present study provides a detailed discussion of primary hypotheses of homology and includes a more complete range of terminal taxa than previous analyses. The analysis provides the first convincing synapomorphies for the Arachnomorpha, and suggests that the marrellomorphs are not arachnomorphs. The assignment of Cambrian ‘great appendage’ (or megacheiran) arthropods to the Arachnomorpha is confirmed, and potential synapomorphies uniting them with chelicerates are discussed and tested. Principal amongst these are the loss of the first cephalic appendages (the antennae), loss of the exopods of the second cephalic appendages and modification of the endopods of these appendages into spinose grasping organs. The Arachnomorpha consists of two major clades: (1) a ‘chelicerate-allied’ clade, including chelicerates, megacheirans, Emeraldella, Sidneyia, cheloniellids and aglaspidids, in which chelicerates and a paraphyletic group of megacherian arthropods form the sister group to the remaining taxa; and; (2) a ‘trilobite-allied’ clade, including trilobites, xandarellids, helmetiids, tegopeltids and naraoiids, the relationships of which are more fully resolved than in previous studies.

A fourth specimen of the enigmatic Lower Cambrian arthropod Paleomerus hamiltoni is described from the Mickwitzia Sandstone of Sweden. The nature of the telson of P. hamiltoni was previously unknown and it has been assumed that the abdomen consisted of 12 segments, the posterior-most being elongate and tapering. This new material constitutes a mould of the entire animal and a cast of the posterior-most abdominal segment and telson, proving that the abdomen consists of 11 segments and a short, tapering telson, which was previously misinterpreted as the twelfth abdominal segment. The presence of 11 abdominal segments in P. hamiltoni eliminates significant diff erences between it and Strabops thacheri, supporting assignment to Strabopidae and rendering the family Paleomeridae redundant.

The Silurian arthropod Bembicosoma pomphicus Laurie, 1899 is re-studied in relation to other Palaeozoic chelicerate taxa. All three known specimens of Bembicosoma originate from the Silurian (late Llandovery) Eurypterid Bed of the Gutterford Burn Flagstones, Reservoir Formation, Pentland Hills, Scotland. Bembicosoma is removed from its previous tentative assignment to Eurypterida and re-assigned to Xiphosura. A morphological reconstruction of this taxon is presented for the first time. This work continues a reappraisal of the systematics of a number of taxa that belong within the synziphosurines, a loose grouping of early Palaeozoic chelicerate arthropods aligned with the Xiphosura (horseshoe crabs), but which have been mistakenly identified as eurypterids in the past. Bembicosoma is significant as it is one of the earliest known synziphosurines.

The type material of Chasmataspis laurencii Caster & Brooks, 1956 (Chelicerata: Chasmataspidida) from the Middle Ordovician (Tremadoc to Caradoc) of Sevier County, Tennessee, USA, is redescribed, and comparisons are drawn with recently discovered Devonian chasmataspids from Scotland, Germany and Russia. The depositional setting of the C. laurencii fossils is reinterpreted as an ash fall into shallow marine/tidal sediments. Chasmataspis laurencii confirms the presence of 13 opisthosomal segments in Chasmataspidida, a character of unresolved polarity. The phylogenetic position ofC. laurencii is difficult to resolve. A monophyletic Chasmataspidida has one convincing autapomorphy in the nine-segmented postabdomen, but C. laurencii shares a number of characters with xiphosurans (i.e. cardiac lobe, pre-abdomen with axial region, biramous and chelate limbs), while the Devonian taxa more closely resemble eurypterids (i.e. through pediform limbs and a genital appendage). Earlier interpretations of the respiratory system in C. laurencii appear unconvincing in the light of new evidence from Devonian forms with opisthosomal opercula. Resting impressions of a Chasmataspis-like animal from the Upper Cambrian Hickory Sandstone of Texas, USA, also appear to show evidence of opercula. These Texan fossils could represent the oldest record of Euchelicerata. Two chasmataspid families are recognised, and C. laurencii is placed in the monotypic Chasmataspididae Caster & Brooks, 1956, redefined here on the fused pre-abdominal buckler with an axial region and the narrow, elongate post-abdomen ending in a long, lanceolate telson.

A new chasmataspid (Chelicerata: Chasmataspidida) is described from the Early Silurian (Late Llandovery–Early Wenlock) of Lesmahagow, Scotland, as Loganamaraspis dunlopi gen. et sp. nov. It is distinguished from related forms by the low tapering ratio of the postabdomen, pediform sixth prosomal appendage and a heart-shaped metastoma. Additionally, a genital operculum with a genital appendage is preserved. The recognisable morphology of L. dunlopi gen. et sp. nov. bridges some of the gap between the Ordovician Chasmataspididae and the Devonian Diploaspididae. Traces of the gut are reported for the first time from a chasmataspid. Chasmataspids are regarded as sister group to the eurypterids: an anterior opercular plate anterior to the genital appendage in L. dunlopi gen. et sp. nov. is regarded as homologous to the anterior opercular plate of the three-segmented genital operculae of Dolichopterus and Stylonurina. This is considered a plesiomorphic character within Eurypterida, whilst the two-segmented genital operculum (with deltoid plates) of Eurypterina is considered apomorphic.

The record of Ordovician Eurypterida from New York State, USA, is shown to be largely false. Twenty-nine species in 17 genera are here recognised as pseudofossils, reducing by more than 75% the total number of named Ordovician eurypterid taxa. Consequently, 10 families now have their first occurrence either later in the Ordovician or in the Early Silurian. The implications for eurypterid palaeoecology, diversity and evolution are not as straightforward as would be expected from such a drastic taxonomic revision. All Ordovician eurypterids are now known to occur in shallow-water, near-shore shales or fine-grained carbonates. Diversity measures indicate that the end-Ordovician extinction event appears to have had less effect on eurypterids than previously known, and their turnover is level in the Ordovician.

Study of abundant phalangiotarbid (Arachnida: Phalangiotarbida) material – provisionally assigned here to Bornatarbus mayasii (Haupt in Nindel 1955) – from the Upper Carboniferous of Writhlington, UK, has revealed new information about some previously equivocal characters. The present authors report a trifurcate apotele, possible spiracles on sternite 5, and confirm the presence of 10 opisthosomal tergites plus a dorsal anal operculum. The affinities of phalangiotarbids are obscure, with most authors favouring affinities with Opiliones (harvestmen) and/or Acari (mites and ticks). Phalangiotarbida is scored for characters used in previous studies of arachnid relationships. A cladistic analysis based on 63 characters using 13 terminal arachnid taxa (plus a hypothetical outgroup), resolves Phalangiotarbida as sister group to (Palpigradi + Tetrapulmonata): the taxon Megoperculata sensu Shultz (1990). Even under cladistic analysis, the position of the Phalangiotarbida remains hard to resolve, but a prosomal sternite with distinct sclerites potentially groups them with the Megoperculata.

The first fossil opilioacariform mite (Acari: Opilioacariformes), Paracarus pristinus sp. n., is described from Baltic amber (Tertiary: Eocene). This well-preserved inclusion has six eyes and, thus, can be excluded from the derived, four-eyed genera. Based on characters such as the extremely long leg 4, it is assigned to the extant central Asian genus Paracarus Chamberlin & Mulaik, 1942. As with many Baltic amber inclusions, we lack an exact provenance for the type specimen, but the Baltic region is further north than the present day range of Recent opilioacariform species and suggests that the range of this group was once more extensive within Eurasia. The first solifuge (Solifugae) from Baltic amber is described as Palaeoblossia groehni gen. et sp. n. Based on its size and general appearance, it can be assigned to Daesiidae, and again represents an arachnid group which previously had a much wider geographical range. If adult, it may represent the smallest solifuge known.

The currently accepted cladogram of spider phylogeny and palaeontological data are used to evaluate spider family richness through geological time. A significantly more diverse spider fossil record is predicted than observed. The predicted rate of spider family diversification is considered more accurate because of its close similarity at 0 Ma to the number of extant families. Predicted spider family palaeodiversity is compared with insect family palaeodiversity to investigate whether spiders track insects through geological time. At the family level, the insects, and observed and predicted spider fossil records show an exponential increase over time, the pattern typical of a radiating group. No significant differences are observed in the rates of change in the slopes, and hence rate of diversification of spiders and insects over time. This suggests that spiders probably co-radiated alongside the insects, with the major radiations of both groups occurring at least 100 Ma before the origin of angiosperms.