In a recent contribution to Geological Journal Williams (2002) proposed a series of major modifications to the current understanding of the Early Ordovician tectonic evolution of the Irish Caledonides. The revisions are based mostly on a reassessment of the Rosroe and Derrylea Formations of the South Mayo Trough, which are used to suggest that the Laurentian passive margin did not come into collision with an oceanic island arc until much later than the Arenig age currently favoured (e.g. Dewey and Ryan 1990;Dewey and Mange 1999;Soper et al. 1999). Although this revised scenario does not explain the metamorphism, plutonism and exhumation of the Dalradian, we wish here to clarify several facts made by Williams that bear heavily on his final conclusion.
Williams asserts that tuffs of the Rosroe Formation 'have a geochemical signature typical of mafic-ultramafic rocks, komatiites/subalkaline basalts typical of mid-ocean ridge volcanism' (p. 138) and that these tuffs were 'derived from an adjacent oceanic island arc complex' (p. 139). Trace element analyses of the Rosroe and Derrylea tuffs (Clift and Ryan 1994;Draut and Clift 2001) show that both have clear relative Nb depletions in multi-element spider diagrams, a characteristic of subduction volcanism. The Rosroe tuffs display a strong relative light rare earth element (LREE) enrichment, with La/Sm ratios ranging from 3 to 6, supporting a strongly enriched continental, not oceanic, origin for these units. These data, which were not cited by Williams (2002), clearly negate any primitive, oceanic island arc or MORB-type origin for Rosroe volcanism. Also, Nd isotope analysis of these tuffs demonstrates that they do not show the unradiogenic character typical of oceanic arc rocks, but require significant reworking of continental material during petrogenesis (Draut and Clift 2001). The transition from oceanic to continental Nd isotope character between the older Bencorragh Formation ("Nd ¼ þ7.2; 485 Ma) and the Rosroe Formation ("Nd ¼ À10.6; 470-467 Ma) is not compatible with the tectonic model of Williams, and is most simply accounted for by a collision of an oceanic arc with the passive margin of Laurentia ("Nd ¼ À17.6 in Connemara).
New trace element and Pb isotopic data from the Rosroe Formation (Clift et al., in press) confirm the conclusion of Dewey and Mange (1999) that the source must be the Laurentian continent and not an island arc terrane. Also, Clift et al. (in press) and Dewey have collected 372 new palaeocurrent measurements that demonstrated a dominant flow from NE to SW, including the Derrylea Formation. These indicate a dominant axial flow, with major input from the north, i.e. Laurentia, as proposed by Dewey and Mange (1999). In this context, the finer grained facies of the Derrylea contrasted to the Rosroe can be understood as reflecting preferential subsidence and coarser grained sedimentation on the south limb of the South Mayo Trough.
Williams concluded that southward-directed subduction either never occurred, or had ceased by the time of Rosroe deposition in favour of northward-directed subduction. Because there is no clear reason why subduction