Abstract: We estimate the global rate of biogenic silica production in the ocean to be between 200 and 280 × 1012 mol Si yr−1. The upper limit is derived from information on the primary productivity of the oceans, the relative contribution of diatoms to primary production and diatom Si/C ratios. The lower limit is derived independently using a multi-compartment model of nutrient transport and biogenic particle flux, and field data on the balance between silica production and dissolution in the upper ocean. Our upper limit is 30–50% lower than several previous estimates, due to new data indicating lower values for both the relative contribution of diatoms to primary productivity and their Si/C ratios. Globally, at least 50% of the silica produced by diatoms in the euphotic zone dissolves in the upper 100 m, resulting in an estimated export of 100–140 × 1012 mol Si yr−l to the deep ocean. Our estimates correspond to a global mean rate of biogenic silica production between 0.6 and 0.8 mol Si m−2 yr−1. Incubation experiments indicate that silica production rates exceed that mean by a factor of 3–12 in coastal areas and are 2–4 times less than the global average in the oligotrophic mid-ocean gyres. The mean silica production rate in waters overlying diatomaceous sediments (approximately 10–12% of the surface area of the oceans) is 0.7–1.2 mol Si m−2 yr−1. That rate is only slightly higher than the global average, indicating that the silica produced in those regions is only 10–25% of the global total. The estimated production of biogenic silica in surface waters of the mid-ocean gyres is approximately equal to that for all major areas of opal sediment accumulation combined. Regional comparison of silica production and accumulation rates suggests a strongly bimodal character in the efficiency of opal preservation in the sea. In waters overlying diatom-rich sediments 15–25% of the silica produced in the surface layer accumulates in the seabed, while virtually none of the silica produced in other areas is preserved. The global burial/production ratio of ˜ 3% is a composite of those two very different systems. The mechanisms leading to more efficient opal preservation in regions of silica accumulation are presently unknown, but they have no simple relationship to primary productivity. Regional differences in opal preservation appear to be controlled by factors such as low surface temperature, selective grazing and aggregate formation, which diminish the rate of silica dissolution in surface waters and/or accelerate its transport to the seafloor.