ABSTRACT
The relationships of the agnostid trilobites have been controversial. They have been claimed as highly derived, neotenously-derived trilobites on the one hand, or as arthropods close to the stem group of crustaceans on the other. On the former view, their distinctive autapomorphies are regarded as specialisations related to their mode of life; on the latter, emphasis is placed on the differences from polymerid trilobites as indicating a separate derivation. We consider it necessary to include Eodiscina in any assessment of agnostid relationships. Previously suggested synapomorphies uniting crustaceans and agnostids to the exclusion of trilobites are reviewed, and the majority is rejected. A number of new synapomorphies uniting agnostids with all or some eodiscinids are described. These characters are tested in the context of a cladistic analysis of 79 eodiscinid and 3 agnostid taxa. We conclude that agnostids are deeply nested within the eodiscinids. The eodiscinid family Weymouthiidae is strongly supported as sister taxon to the Agnostina. Several supposed agnostid autapomorphies are found to be synapomorphies shared with one or more eodiscinids. At least as far as exoskeletal characters are concerned, the supposed distinctiveness of agnostids has been over-emphasised. To assert non-trilobite relationships of agnostids would require implausible levels of homoplasy.
