Study of the accretionary biomineralised hardparts of organisms (sclerochronology) can make a useful contribution to palaeoclimatology. Ontogenetic sequences of isotopic data (δ¹⁸O and Δ₄₇ values) from the shells of marine molluscs are a source of information on seasonal sea-surface temperatures that can be used for detailed and rigorous evaluation of the outputs of numerical climate models. In situations where there is significant seasonality, and where shell preservation is adequate, accurate information about winter and summer surface temperature can be obtained from shallow-water benthic forms (bivalves and gastropods), in particular the early ontogeny of fast-growing species. Accurate information about winter surface temperature can also be obtained from individuals that lived at mid-shelf depths (20–40 m), but summer seafloor values from these need upward adjustment to derive a plausible surface temperature if thermal stratification of the water column occurs in this season. Ontogenetic δ¹⁸O profiles from planktonic pteropod gastropods are a potential source of insight into seasonal surface temperatures in the ocean basins; these organisms merit investigation for provision of information to complement shelf data.

Temperature profiles constructed from shell δ¹⁸O require an estimate of the δ¹⁸O value of ambient seawater, which can be derived by back-calculation from the Δ₄₇-temperature supplied by the same shell material. Alternatively, through appropriate sampling and data processing, seasonal temperatures can be obtained directly from Δ₄₇ profiles. Climate parameters are defined in terms of the mean state over a period of 30 consecutive years, a statistic (e.g., for seasonal temperatures) which can be derived from the long isotopic temperature records obtainable from bivalve species that live for many tens or hundreds of years. Efforts should be made to acquire such records, especially averaged data from crossdated shells, to specify climate parameters for precise times in the past. Information for precise times would be of particular value for icehouse intervals like the late Cenozoic, characterised by high frequency (high amplitude) climate fluctuation. Short records from non-crossdated shells can nevertheless provide useful insights into climate, particularly if a large dataset is obtained, supplying a reliable picture of the mean state and range of variation in climate parameters over the interval represented by the shells.