Here we investigate the potential of Sr/Ca ratios in the marine gastropod Conus ermineus for reconstructing seawater temperatures. We present annually resolved records of Sr/Ca and δ18O for four shells collected alive from the Flower Garden Banks National Marine Sanctuary in the Gulf of Mexico. Our results show that variations in Sr/Ca and δ18O covary with the in situ seasonal temperature cycle. Sr/Ca and temperature are positively correlated, in contrast with the inverse relationship found in inorganically precipitated aragonite. The seasonal Sr/Ca variability is superimposed on a long-term trend of increasing Sr/Ca with age. Both the seasonal and long-term ontogenetic changes in Sr/Ca are associated with variations in growth rate, defined here as the shell linear extension rate (LER); the seasonal variability in LER is superimposed on a long-term decrease with ontogeny. Thus the covariance of Sr/Ca ratios with temperature and LER suggests that Sr incorporation is likely driven by temperature influence on growth rate, rather than by thermodynamic effects. Unlike the seasonal variability, the ontogenetic effect is characterized by inverse covariation between Sr/Ca and LER, suggesting that Sr/Ca variability is not controlled by growth rate alone, but probably by two different biomineralization mechanisms, one related to temperature and the other related to age. We use the seasonal Sr/Ca signal of four shells to construct a temperature calibration. To minimize the ontogenetic effects, we separate the records into juvenile and adult growth stages and calculate the Sr/Ca-temperature (T) relationships: Juvenile: Sr/Ca (mmol mol-1) = 0.042 (¿0.008) * T (¿C) + 0.24 (¿0.21) (R2 = 0.66) Adult: Sr/Ca(mmol mol-1) = 0.072 (¿0.014) * T (¿C) - 0.05 (¿0.34) (R2 = 0.68) Applying the calibration to a single specimen provides mean annual temperature estimates within ¿1¿C of the in situ temperature record but resolves the seasonal variability only within ¿3.5¿C. The large error in the seasonal estimates reflects the high variability among specimens. To reduce the uncertainty on seasonal temperatures, we propose combining records from multiple shells to generate an average temperature record. The potential of this approach needs, however, to be validated in other locations.