Abstract At the present time the dating of deep sea sediment horizons of upper Pleistocene age is based primarily on the use of radioactive daughter isotopes originating by decay of uranium dissolved in sea water. This is a fact often overlooked when cores ‘dated’ by microfossil assemblages or oxygen isotope cycles are used to reconstruct past events. Also not well understood is the present state of relative imprecision of this dating technique compared to other geochronological tools. Considering the importance of excess radioactive daughters (XRD's) in the interpretation of deep sea sediments, it is surprising that the data base and the theoretical framework have been so inadequate. A review of the literature shows that: (1) only a few cores have been systematically analyzed for 230Th and/or 231Pa; (2) the reliability and interpretation of even some of these studies have been questioned; and (3) despite a proliferation of microfossil and stable isotopic studies, no systematic XRD investigations have been attempted in this decade (except for indirect methods such as gamma ray or alpha track analyses, which measure neither uranium nor protactinum directly). Furthermore, the only models of XRD accumulation in sediments which have been seriously used as a basis for age determination and sediment accumulation rates have assumed either that the activity of an isotope in sediments has been a constant term, or that the ratio of activities of two isotopes has remained constant. Both of these assumptions are known to be only rough approximations, especially in the case of fluctuating sediment accumulation rates. A systematic analysis of the possible modes of covariation of accumulating sediment and accumulating XRD's in deep sea cores leads to a number of new conceptions: (1) A model which assumes a constant accumulation rate (flux) of XRD, regardless of changing sediment supply, deserves more attention than it has thus far received. (2) Even more useful may be ‘hybrid’ models, which assume that both constant activity and constant flux terms are involved in the activity of accumulating sediments. (3) Most of these models can be used to estimate the ‘transitory’ sediment accumulation rate for any given core horizon; the only model which cannot be used in this way is the standard constant activity model. (4) The ratio of two XRD's, i.e., 231Pa to 230Th, can be applied to these models in exactly the same way as activity data for any one isotope, as if the ratio were itself the activity of an imaginary isotope. This means that we are not limited to the assumption of constant ratio, which is patently unrealistic for many cores; rather the ratio data can be applied to the constant flux and hybrid models as well. As a test of these conclusions, the use of the various models, especially the ‘M’ hybrid, have been applied to several published cores in each of two deep sea basins where numerous data points and/or 231Pa/ 230Th are available: the Tasman Sea and the Caribbean Sea. In both regions, the curves of sediment accumulation rate versus age of core horizon appear to be consistent enough to warrant further application of these approaches. There does appear to be a need, however, for the collection of more 230Th and 231Pa data in cores, more carefully analyzed, and with the application of better accumulation models in mind.