List of luminescence age-estimates. Samples prefixed by R were dated by the Nordic Laboratory for Luminescence Dating, Riso, Denmark using the coarse-grained (sand) method. For sample treatment see Aitken (1985). Dose rates (in Gy kal) were calculated for each sample after measuring beta activity in the laboratory. Measurements were carried out on both feldspar (F) and quartz (Q) due to differences in internal uranium and thoriurn contribution for the two minerals (a value of 0.1 Gy/ka for quartz and 0.18 Gy/ka for feldspar has been used) (Mejdahl, unpublished). For feldspar, equivalent dose, Dp, has been obtained using the plateau TL method (Mejdahl, 1988), total bleach additive-dose (both IRSL and OSL) and single aliquot regenerated and added dose (SARA) also using both IRSL and OSL (Mejdahl and Botterlensen, 1994). For quartz Dp was obtained using the single aliquot regenerated dose (SARI method (Murray and Roberts, 1998; Murray and Mejdahl, 1998) and partly SARA. The results from measurements on quartz are considered more reliable because quartz does not show anomalous fading (A. Murray, pers. comm.). The 'Age F/Q' column gives a comparison of the F and Q results. The subscripts given after the average values of On are the number of independent estimates included in the average. Datings from unit C, labelled PEC, were made by Quaternary TL Surveys, Nottingham, UK, by measuring luminescence of the fine-grained fraction (Debenham, 1985). These results can be considered maximum ages as the analysis does not account for remnant luminescence, which may yield higher apparent ages, but the datings were reported as minimum ages because instability of the TL signal limits the range of method to 110 ka (N. Debenham, pers. comm.)