Chronostratigraphy—the temporal ordering of geologic strata—is at the heart of earth history. In reviewing chronostratigraphic theory, from d'Orbigny to Hedberg, we see that concepts and practices have evolved under pressure of constant improvements in time control and data quality. Such stimuli are naturally felt more strongly in the youngest part of the stratigraphic record, where dating is most precise and the biostratigraphic data are most abundant and well-preserved. It is not surprising, therefore, to find that chronostratigraphy is conceptually more advanced in the Cenozoic Erathem than in the Mesozoic and Paleozoic Erathems, where biostratigraphy and chronostratigraphy are widely confused, evidently out of necessity. Taking the five inter-series boundaries in the Cenozoic Erathem as case histories, we review their current status in relation to the lower boundary of the lowest `standard' stage above the boundary, in light of the hierarchical principle recommended in the International Stratigraphic Guide ( Hedberg, 1976; Salvador, 1994). In each instance, what should be a simple nested relationship turns out to be complex and difficult, and in the cases of the proposed global stratotype section and point (GSSPs) for the Eocene/Oligocene Series boundary and the Paleocene/Eocene Series boundary there are discrepancies so large with regard to the base of the Rupelian and Ypresian standard stages, respectively, that some essential relaxation of the hierarchic rule seems to be required if the historic value of these long-known stages is not to be violated. Since stage boundaries should not be defined a posteriori to fit series boundaries, we propose here to reconsider the role of the Standard Stage as the obligate elemental subdivision in a rigidly hierarchical chronostratigraphic scale and suggest that series and subseries should be formalized and their boundaries defined by GSSPs. We do not suggest that stages be abandoned. On the contrary, we propose to integrate them in a two-step hierarchical chronostratigraphic framework, so as to use them in complementary fashion with series and subseries. However, the boundary concept is more appropriate for stage than the GSSP. We recognize the need for precise and globally valid chronostratigraphic correlations based on GSSPs in attempts to understand the earth system, while at the same time, we emphasize the need for conceptual continuity with regard to the central place that the stage has played in chronostratigraphy over several decades.