Abstract The Milankovitch theory of climate change is widely accepted, but the registration of the climate changes in the stratigraphic record and their use in building high-resolution astronomically tuned timescales has been disputed due to the complex and fragmentary nature of the stratigraphic record. However, results of time series analysis and consistency with independent magnetobiostratigraphic and/or radio-isotopic age models show that Milankovitch cycles are recorded not only in deep marine and lacustrine successions, but also in ice cores and speleothems, and in eolian and fluvial successions. Integrated stratigraphic studies further provide evidence for continuous sedimentation at Milankovitch time scales (10 4 years up to 10 6 years). This combined approach also shows that strict application of statistical confidence limits in spectral analysis to verify astronomical forcing in climate proxy records is not fully justified and may lead to false negatives. This is in contrast to recent claims that failure to apply strict statistical standards can lead to false positives in the search for periodic signals. Finally, and contrary to the argument that changes in insolation are too small to effect significant climate change, seasonal insolation variations resulting from orbital extremes can be significant (20% and more) and, as shown by climate modelling, generate large climate changes that can be expected to leave a marked imprint in the stratigraphic record. The tuning of long and continuous cyclic successions now underlies the standard geological time scale for much of the Cenozoic and also for extended intervals of the Mesozoic. Such successions have to be taken into account to fully comprehend the (cyclic) nature of the stratigraphic record.

Published radioisotopic (K/Ar, 40 Ar/ 39 Ar, and Rb/Sr) and astronomical ages for the Eocene-Oligocene boundary are essentially consistent at ca. 33.8 ± 0.1 Ma, but the 40 Ar/ 39 Ar ages have been calculated relative to an outdated age of 27.83–27.84 Ma for the Fish Canyon Tuff sanidine dating standard. Application of a revised age of 28.02 Ma, or the new astronomically calibrated age of 28.201 Ma, leads to significant discrepancies, while others are eliminated. In particular, the astronomically tuned ages of ca. 33.79 Ma at Ocean Drilling Program (ODP) Site 1218 and of 33.90–33.95 Ma at Massignano–Monte Cagnero are now in good agreement with recalculated (alternative) 40 Ar/ 39 Ar sanidine ages for the boundary as derived from the volcanic ignimbrite complex in New Mexico and for the Persistent White Layer (PWL) ash bed in North America, which is supposed to closely correspond to the boundary. This mutual consistency suggests that the tuning is correct at the scale of the 400 k.y. eccentricity cycle. Evidently, additional single-crystal 40 Ar/ 39 Ar sanidine dates from the tuffs in North America and independent checks on the astronomical tuning and the intercalibration between the astronomical and 40 Ar/ 39 Ar dating methods are needed to definitively solve the problem of the numerical age of the Eocene-Oligocene boundary. It is anticipated that such analyses and tests will be carried in the coming years as part of the international Earthtime initiative and associated projects to significantly improve the geological time scale. Clearly, an accurate and precise dating of the Eocene-Oligocene boundary is crucial if we are to unravel the underlying cause of the major climate transition associated with it.