Selected warm and cold water strains of the coccolithophorid Emiliania huxleyi and the closely related species Gephyrocapsa oceanica were cultured under controlled temperature conditions to assess genetic and physiological variability in the alkenone/alkenoate vs. temperature relationship. Differences in the strains' growth rates over the 6-30 degrees C experimental temperature range were small but consistent with their cold or warm water origins. E. huxleyi and G. oceanica had similar alkenone/alkenoate biochemistry, justifying the extension of alkenone stratigraphy to sediments predating the appearance of E. huxleyi. These species could not be distinguished by C-38/C-37 alkenone or alkenoate/alkenone ratios as previously suggested (Volkman et al., 1995; Sawada et al., 1996) but given samples from a range of temperatures may be distinguished by a plot of the C-38 ethyl vs. C-38 methyl unsaturation ratios (U-38Et(K) and U-38Me(K), respectively). Biochemical responses to temperature and the C-37 alkenone-based (U-37(K')) temperature calibrations differed significantly among the strains. The U-37(K') temperature calibration was nonlinear for five of the six strains examined. A reduction in slope of the calibration at temperatures < 12 degrees C and >21 degrees C suggests the cell's alkenone-based adaptation to temperature is limited at the extremes of its growth temperature range. The unsaturation ratios of the C-38 methyl and ethyl alkenones (U-38Me(K) and U-38Et(K)) varied similarly with temperature and were strongly intercorrelated. The experiments also documented an influence of cell physiological state on both alkenone and alkenoate composition and on alkenone unsaturation. Cells in late logarithmic and stationary growth had significantly increased abundance of alkenoates and C-38 ethyl alkenones relative to C-38 methyl alkenone abundance. In some strains the unsaturation ratios of both C-37 and C-38 alkenones also significantly decreased when cells entered the late log phase. Comparison of culture results with field data indicates that the average physiological state of alkenone-synthesizers in the open ocean differs from cultured cells growing under exponential growth and appears to be more similar to cells in late log or stationary growth phases. Differences in alkenone/alkenoate ratios between cultured cells and sediments underlying waters of a similar temperature most probably reflect a difference in cell physiology between cultured cells and oceanic populations and not greater diagenetic losses of alkenones relative to alkenoates, as previously suggested (Prahl et al., 1995). Our experiments confirm that biogeographical variations observed in the alkenone vs. temperature relationship in natural waters reflect, at least in part, differences in genetic makeup and physiological status of the local alkenone-synthesizing populations. Hence, alkenone-based paleo sea surface temperature estimates are subject to errors, albeit small, which arise from genetic differences between modern-day and paleo-populations. The reduction in slope of the U-37(K') temperature calibration for most strains at T > 24 degrees C indicate that linear U-37(K') temperature calibrations (e.g., Prahl et al., 1988) which are currently used to estimate paleo SST, and which are poorly constrained at higher temperatures, probably underestimate the magnitude of SST change for tropical and subtropical regions. Copyright (C) 1998 Elsevier Science Ltd.
