Paleontological data provide essential insights into the processes shaping the spatial distribution of present-day biodiversity. Here, we combine biogeographic data with the fossil record to investigate the roles of parallelism (similar diversities reached via changes from similar starting points), convergence (similar diversities reached from different starting points), and divergence in shaping the present-day latitudinal diversity gradients of marine bivalves along the two North American coasts. Although both faunas show the expected overall poleward decline in species richness, the trends differ between the coasts, and the discrepancies are not explained simply by present-day temperature differences. Instead, the fossil record indicates that both coasts have declined in overall diversity over the past 3 My, but the western Atlantic fauna suffered more severe Pliocene−Pleistocene extinction than did the eastern Pacific. Tropical western Atlantic diversity remains lower than the eastern Pacific, but warm temperate western Atlantic diversity recovered to exceed that of the temperate eastern Pacific, either through immigration or in situ origination. At the clade level, bivalve families shared by the two coasts followed a variety of paths toward today's diversities. The drivers of these lineage-level differences remain unclear, but species with broad geographic ranges during the Pliocene were more likely than geographically restricted species to persist in the temperate zone, suggesting that past differences in geographic range sizes among clades may underlie between-coast contrasts. More detailed comparative work on regional extinction intensities and selectivities, and subsequent recoveries (by in situ speciation or immigration), is needed to better understand present-day diversity patterns and model future changes.B iodiversity is spatially structured at many scales. Biogeographic realms and provinces, the partitioning of species among environments within biogeographic units, and genetic population structure within species are all manifestations of the ecological and evolutionary processes that generate and maintain diversity in terrestrial and marine systems. For most terrestrial and marine organisms, the first-order global diversity pattern is the latitudinal diversity gradient (LDG): the increase in the number of species and higher taxa from the poles to the tropics (1). This gradient, as seen today, has been shaped by a combination of origination, extinction, and past geographic shifts of taxa; all of these variables might differ among clades, functional groups, and land masses or ocean basins (2-7), even when the overall LDG is similar. Thus, a mechanistic understanding of the present-day LDG in general, and along specific equator−pole transects, for distinctive regional faunas, and for individual clades, cannot be divorced from its history.The need for historical data is underscored by the differences and similarities among related clades that broadly conform to the LDG. As in many evolutionary question...