Introduction As the contents of this volume, the international symposium that preceded it and, indeed, the long list of scholarly publications that have been generated over many decades illustrate, Tara looms large. The hill spans millennia of human activity on this island and still remains at the centre of current political and ideological debate. It is not surprising that several contributions in this volume approach Tara from the macro-scale, attempting to grasp the ‘bigger picture’ of the hill through a certain period or periods. This contribution starts at the other end of this spectrum, however. In attempting to address the first of four overarching research questions in this volume—What additional information can be extracted from the known archaeology of Tara?—it will use a number of small pieces of pottery from the Mound of the Hostages passage tomb to cast light on a poorly understood part of the Neolithic, the apparent ‘transition’ between the Middle and Late Neolithic at the turn of the third millennium BC, and to challenge some implicit assumptions about the ‘decline’ of passage tomb-building communities and the simultaneous ‘rise’ of Grooved Ware-using groups in Ireland. The Late Neolithic at Tara In amongst the abundant evidence for funerary activity in the Middle Neolithic and the Early Bronze Age at the Mound of the Hostages at Tara lie several small yet stubborn traces of a Late Neolithic presence: a sherd of Grooved Ware in a pit at the entrance to the tomb, short sections of palisades to the north and east of the mound, and a number of early third-millennium BC skulls from the tomb passage (O’Sullivan 2005). On a site that has seen very little open-plan excavation, it is hard to tell whether such activity is focused exclusively on the Mound of the Hostages and its immediate vicinity or whether further, more substantial structures remain to be uncovered across the hill (see Fenwick and Newman 2002). In any event, these remains hint at a continuing focus on the Mound of the Hostages at Tara into the early centuries of the third millennium BC. This in itself is nothing new; Late Neolithic monuments and cultural material have been recorded in the vicinity of a number of Middle Neolithic passage tombs, such as those within the Bend of the Boyne (Sweetman 1985; 1987; Roche and Eogan 2001) and at Ballynahatty (Hartwell 1998). The context of the material from the Mound of the Hostages, however, combined with good stratigraphic detail and recent programmes of radiocarbon dating and Bayesian modelling (O’Sullivan 2005; Bayliss and O’Sullivan, this volume), allow us to examine more closely the nature of the ‘transition’ between the Middle and Late Neolithic in Ireland and the relative levels of continuity and change that accompanied it. This paper takes as its starting point the contents of a pit, pit B, located at the south-eastern edge of the mound approximately 4m from the entrance to the tomb (O’Sullivan 2005, 42, figs 15 and 19). Measuring 1m by 0.45m by 0.86m deep, the pit contained several sherds of pottery, including a Grooved Ware rim sherd (in two pieces) and fragments of two decorated Carinated Bowl vessels. Pit B was overlain by a cremation deposit, burial 1, associated with a relatively large and elaborate stone setting (ibid., fig. 23). Spatially, burial 1 belongs to a group of seventeen cremation deposits, human where identifiable, that were uncovered at old ground level along the perimeter of the mound. Fourteen of these lay around the mound’s southern perimeter. Indeed, all but one of the cremation deposits were located to the south-western side of the axis 408 24. Tara in pieces—change and continuity at the turn of the third millennium BC JESSICA SMYTH