South Georgia. Report 23, Department of Quaternary Research, Stockholm University, Stockholm. N. Van der Putten C. Verbruggen Geography Department, Ghent University, Krijgslaan 281 (S8/B2), 9000 Ghent, Belgium E-mail addresses: nathalie.vanderputten@ugent.be (N. Van der Putten), cyriel.verbruggen@ugent.be (C. Verbruggen) N. Van der Putten IMEP, CNRS UMR 6116, Universite´Paul Ce´zanne, Europoˆle Me´diterrane´en de l’Arbois, Baˆtiment Villemin, Avenue Philibert BP 80, F-13545 Aix-en-Provence Cedex 04, France doi:10.1016/j.quascirev.2007.07.004 Reply Reply to Comment by Van der Putten and Verbruggen We are pleased that Van der Putten and Verbruggen share our interest in the glacial history of South Georgia and we respond to their three sets of comments in turn below. 1. The extent of glaciation and the chronology of deglaciation of Stromness Bay/Husvik Harbour area The record produced by Rosqvist et al. (1999) from Tonsberg Lake 1 implied that glaciation was more restricted at 18.6 ka BP than any previous work had suggested: if the lake was accumulating organic and other sediment sub-aerially then the glaciers must have retreated. This struck us as an interesting and important result and so we set out to test it using geomorphological evidence from 15 sites, ranged over 60 km of the north-east coastline of the island, and using cosmogenic surface exposure dating on moraines from a sub-set of these sites. We explicitly discussed the possibility that the dates of Rosqvist et al. (1999) might be wrong (Bentley et al., 2007, p. 672), but rejected this as unlikely: there are seven 14 C dates older than 12,000 14 C yr BP in the basal part of the core, they are all in stratigraphical order, and yield a reasonable age-depth model. We do not agree with Wasell (1993) or Van der Putten and Verbruggen (2005) that Tonsberg Lake 1 somehow became deglaciated early and remained ice-free for the following 10 kyr while other sites on the Tonsberg Peninsula were covered by glacier ice. There is no geomorphological evidence of a glacier margin in this configuration, and we cannot envisage how this would have worked glaciologically for such a sustained period of time. We believe that the simpler explanation is that the range of basal dates from the other lakes (approximately 8000–9500 14 C yr BP (9000–11,000 cal yr BP) (Van der Putten and Verbruggen, 2005)) simply reflect varying times for onset of organic sedimentation or that coring studies did not retrieve the full basal sequences. In other words, the basal dates from lakes are not close minima for deglaciation prior to 18.6 ka. Diverse dates for the onset of organic sedimentation in closely adjacent lakes are a common feature of many deglaciated terrains across the world. 2. The Holocene glaciation history Van der Putten and Verbruggen (2007) suggest we should have compared our Holocene results to their data. We cannot comment on the unpublished data that they cite but can explain why we did not compare our record to Van der Putten et al. (2004). We were very wary of over- interpreting the Mid-Holocene advance, dated by us at 3.671.1 ka, because of the large error on this age. We were explicit about this reticence and that correlations to other proxy data may not be robust (see cautionary comments on pp. 648, 649 (Table 1), 672, 673, 674, 675 (Conclusions) of Bentley et al., 2007). For this reason, we did not cast around widely for proxy records to compare against—our interest was explicitly focussed on explaining why an interglacial advance apparently reached a similar extent to a Late Glacial one. The precise timing of the interglacial advance is not particularly important to this discussion. 3. The presence of a raised beach in Husdal Van der Putten and Verbruggen (2007) query our geomorphological mapping of raised beaches in Husdal (see Fig. 10d of Bentley et al., 2007). As we have no access to the unpublished study that they cite as evidence for their freshwater deposits we can only re-iterate what the geomorphology indicates to us. There is a flight of shingle and gravel ridges, extending horizontally for tens or hundred of metres and descending from 10 m a.s.l. to the present shoreline. The ridges are parallel to the coast but transverse to the Husdal river, which truncates and incises the ends of some ridges. The individual gravel ridges have similar dimensions and constituent sediment to the contemporary shingle beach at the mouth of Hudsal, and ARTICLE IN PRESS Reply / Quaternary Science Reviews 26 (2007) 2684–2691 2690