Extended Data Fig. 3: The luminescence characteristics of the quartz and feldspar sampled from caves in the Guangxi region of southern China.

The resulting data from the application of three luminescence dating techniques; pIR-IRSL SA feldspars, pIR-IRSL SG feldspars and OSL SG quartz to typical samples from three different caves have been displayed. a, d, g, pIR-IRSL data for a single-aliquot from sample HEJ1 from Hejiang cave (a G. blacki bearing cave): a, a shine down curve including both the natural and regenerative dose; d, the dose response of the sample aliquot with a D_e of 631 ± 18 Gy, a recycling ratio of 1.04 ± 0.03 and a D0 of 1136 ± 139 Gy; g, the corresponding radial plot for this sample with an uncorrected MAM of 615 ± 11 Gy (n = 12 aliquots). b, e, h, pIR-IRSL single-grain data for sample GONG-16 from Gongjishan cave (a non-G. blacki bearing cave): b, single-grain shine down including both the natural and regenerative dose; e, the dose response of the sample aliquot with a D_e of 801 ± 51 Gy, a recycling ratio of 1.06 ± 0.05 and a D0 of 787 ± 124 Gy; h, the corresponding radial plot for this sample with an uncorrected MAM of 388 ± 23 Gy (n = 32 grains). c, f, i, OSL quartz single-grain data for sample CLPB-5 from Lower Pubu cave (a non-Giganto bearing cave): c, a single-grain shine down curve including both the natural and regenerative dose; f,  the dose response of the sample aliquot with a D_e of 108 ± 14 Gy, a recycling ratio of 1.03 ± 0.17 and a D0 of 378 ± 188 Gy; i, the corresponding radial plot for this sample with a MAM of 64 ± 7 Gy (n = 45 grains). j-m, Preheat, dose recovery, bleaching and fading tests for the pIR signal at the four preheat and pIR-IRSL stimulation temperature combinations (outlined in the supp) using sample CSH1 (n = 8). j, A preheat plateau test with the 270 °C pIR-IRSL stimulation and preheat temperature of 300 °C producing the least scatter between aliquots (n = 12). k, A dose recovery test with a surrogate dose of 175 Gy, with the 270/300 °C combination recovering the dose within error margins, also shown are the residual values (in Gy) after a 4 hours bleach using a solar simulator (n = 8). l, The difference in g values as a measure of fading at the 4 different combinations. m, An example of the difference in fading between the IR₅₀ measurement and pIR₂₇₀ measurement (n = 8). n, o, procedural tests for the OSL signal conducted on small aliquots: n, preheat plateau test using 2 aliquots per temperature point (n = 8), 220 °C was the chosen temperature; o, a dose recovery test using a surrogate dose of 100 Gy (n = 10) and the chosen preheat of 220 °C on 10 small aliquots – all aliquots returned the surrogate dose within standard errors. All data points from j-o are mean values with s.d. at 1σ uncertainties.