Abstract Sea surface temperatures (SSTs) derived from the alkenone U37κ′) record of Quaternary sediments may be subject to bias if algae with different temperature sensitivities have contributed to the sedimentary alkenone record. The alkenone-derived SST records are usually based on a U37κ′-temperature relationship which was measured in culture experiments using the coccolithophorid Emiliania huxleyi (F.G. Prahl, L.A. Muehlhausen and D.L. Zahnle, 1988. Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions. Geochim. Cosmochim. Acta 52, 2303–2310). To assess possible effects of past species changes on the U37κ′-temperature signal, we have analysed long-chain alkenones and coccolithophorids in a late Quaternary sediment core from the Walvis Ridge and compared the results to SST estimates extracted from the δ18O record of the planktonic foraminifer Globigerinoides ruber. Alkenones and isotopes were determined over the entire 400-kyr core record while the coccolithophorid study was confined to the last 200 kyr when the most pronounced changes in alkenone content occurred. Throughout oxygen-isotope stages 6 and 5, species of the genus Gephyrocapsa were the predominating coccolithophorids. E. huxleyi began to increase systematically in relative abundance since the stage 5 4 transition, became dominant over Gephyrocapsa spp. during stage 3 and reached the highest abundances in the Holocene. Carbon-normalized alkenone concentrations are inversely related to the relative abundances of E. huxleyi, and directly related to that of Gephyrocapsa spp., suggesting that species of this genus were the principal alkenone contributors to the sediments. Nevertheless, SST values obtained from the U37κ′-temperature relationship for E. huxleyi compare favourably to the isotope-derived temperatures. The recently reported U37κ′-temperature relationship for a single strain of Gephyrocapsa oceanica (J.K. Volkman, S.M. Barrett, S.I. Blackburn and E.L. Sikes, 1995. Alkenones in Gephyrocapsa oceanica: Implications for studies of paleoclimate. Geochim. Cosmochim. Acta 59, 513–520) produces unrealistically high SST values indicating that the temperature response of the examined strain is not typical for the genus Gephyrocapsa. This is supported by the C37:C38 alkenone ratios of the sediments which are comparable to average ratios reported for E. huxleyi, but significantly higher than for the G. oceanica strain. Most notably, the general accordance of the alkenone characteristics between sediments and E. huxleyi persists through stages 8 to 5 and even in times that predate the first appearance of this species (268 ka; H.R. Thierstein, K.R. Geitzenauer and B. Molfino, 1977. Global synchroneity of late Quaternary coccolith datum levels: Validation by oxygen isotopes. Geology 5, 400–404). Our results suggest that U37κ′-temperature relationships based on E. huxleyi produce reasonable paleo-SST estimates even for late Quaternary periods when this species was scarce or absent because other alkenone-synthesizing algae, e.g. of the genus Gephyrocapsa, responded similarly to temperature changes.