Much of Greenland experienced summers warmer than present in parts of the early to middle Holocene, during a precession-driven positive anomaly in summer insolation. However, the magnitude of that warmth remains poorly known, and its timing and spatial pattern are uncertain. Here we describe the first quantitative Holocene palaeotemperature reconstruction from central East Greenland based upon insect (chironomid) assemblages preserved in lake sediments. We postulate that landscapes like our study site, characterized by minimal soil and vegetation development through the Holocene and thus little influenced by some important secondary gradients, are especially well suited to the use of chironomids to reconstruct Holocene temperatures. The inferred timing of warmth at our study site near Scoresby Sund agrees well with other nearby evidence, including glacial geological reconstructions and temperatures inferred from precipitation isotopes at Renland ice cap, supporting the use of chironomids to reconstruct temperatures at this site. We infer highest temperatures from c. 10 to 5.5 ka, followed by gradual cooling after 5.5 ka and progressively colder and less productive conditions after 3.5 ka. Models based upon two independent training sets yield similar inferred temperature trends, and suggest an average summer temperature anomaly from c. 10 to 5.5 ka of 3 to 4°C relative to the preindustrial last millennium. The estimated overall rate of Neoglacial cooling averaged over the period from 5.5 to 0.5 ka was 0.6 to 0.8°C per thousand years, more than twice the rate previously estimated for the Arctic as a whole. Given strong apparent spatial variability in Holocene climate around the Arctic, and the utility of palaeoclimate data for improving climate and ice-sheet models, it should be a priority to further quantify past temperature changes around the margins of the Greenland Ice Sheet, where few quantitative reconstructions exist and future warming will affect global sea level.Yarrow Axford (axford@northwestern.edu),