Llandovery (Silurian) reefs in North Greenland have yielded extensive collections of remarkably well preserved trilobites dominated by scutelluids. These collections provide a unique opportunity to analyse the reefs and their trilobite faunas in a palaeoenvironmental context, and the study highlights the importance of pre-existing collections, to which further scientific value can be added. Counts of trilobite sclerites from thirteen reef localities are analysed using cluster and correspondence analyses to identify trilobite associations, and taphonomic signatures. Lithofacies analysis allows these to be understood in a detailed palaeoenvironmental context. Seven trilobite associations are identified: Scutelluid-Illaenimorph; Scutelluid; Scutelluid-Cheirurid; Scutelluid-Harpetid; Proetid; Encrinurid; and Illaenimorph. The number of associations highlights the variable distribution of taxa within the reefs, which is partly a reflection of intense hydrodynamic sorting in reef environments. Trilobites are predominantly associated with a cement-rich microbial lithofacies; it is apparent that microbial processes were the principal contributors to the production of carbonate mud, and played a significant role in the development of the Greenland reefs. The observed associations comprise time-averaged event beds in an otherwise low energy and slow sedimentation regime, and possibly result from storm activity remobilising sclerites, and depositing them within depressions and cavity fills within the reef surface. Correspondence analysis reveals two distinct taphonomic signatures within the reef, and transport distance is the primary influence on these. The Proetid Association is readily separated from the others by its taphonomic attributes, representing the least amount of transport, and suggesting the taxa involved took advantage of the more sheltered niches within the reef. This is the first study of reef-dwelling trilobites to take advantage of multivariate techniques to identify associations and describe their taphonomic attributes. In particular, the effectiveness of correspondence analysis in recognising taphonomic signatures is demonstrated. (c) 2011 Elsevier B.V. All rights reserved.