Abstract In the early 1830s Charles Lyell was convinced that much of western Europe had been submerged during the Pleistocene by cold seas strewn with icebergs; the relicts of whose loads of rock and mud occurred on land as boulder clay and erratic blocks. Swiss scientist Louis Agassiz disagreed, considering in 1837 that these were the products of deposition by a great ice sheet. Archibald Geikie realized by 1863 that Lyell was wrong. Mountain glaciers had carved the topography of Scotland and other parts of the UK, feeding an ice sheet that left glacial erratics behind when it melted away. He hoped, in vain, to change Lyell’s mind. Archibald Geikie’s mantle passed to his brother James, who compiled evidence from around the world to demonstrate the correctness of his brother’s thesis. It was published in 1874 just before Lyell died still arguing for the correctness of his iceberg theory, which gave us the word ‘drift’ for the unconsolidated deposits mantling the UK. Even so, by then Lyell had gone some way – no doubt partly influenced by the Geikies – to accepting that in certain instances glacial action had, indeed, moved large erratic blocks – locally even uphill, as in the Jura.

Abstract Iceberg discharges into the North Atlantic are important sources of fresh water, and the sediments they deposit can provide constraints on which sectors of different ice sheets were contributing icebergs. 40 Ar/ 39 Ar ages of sand-sized hornblende grains provide useful constraints on IRD (ice-rafted detritus) source areas. Heinrich events are intervals of anomalously high percentages of IRD in marine sediment cores of the North Atlantic IRD belt. In contrast to the others, Heinrich event 3 (H3) records a significantly lower flux of IRD. This study compares 40 Ar/ 39 Ar hornblende age distributions from the interval around and including H3 in giant gravity core EW9303-GGC31 from Orphan Knoll, in the southern part of the Labrador Sea, with piston core V28-82 in the eastern part of the North Atlantic IRD belt. Collectively, these results confirm that H3 represents a Hudson Strait IRD event, but that it was smaller than during H1, H2, H4 and H5, and therefore comprises only a small fraction of the detritus at the eastern North Atlantic location of V28-82. These results support a previously published interpretation of across-strait ice flow during H3 at Hudson Strait. Supplementary material: Appendix 1 is 40 Ar/ 39 Ar data from core EW9303-GGC31; Appendix 2 is grain counts across H3 from core V28-82; Appendix 3 is 40 Ar/ 39 Ar data from core V28-82; these are available at http://www.geolsoc.org.uk/SUP18631 .