A suite of 146 new accelerator-mass spectrometer (AMS) radiocarbon ages provides the first reliable chronology for late Quaternary sediments in Lake Baikal. In this large, highly oligotrophic lake, biogenic and authigenic carbonate are absent, and plant macrofossils are extremely rare. Total organic carbon is therefore the primary material available for dating. Several problems are associated with the TOC ages. One is the mixture of carbon sources in TOC, not all of which are syndepositional in age. This problem manifests itself in apparent ages for the sediment surface that are greater than zero. However, because most of the organic carbon in Lake Baikal sediments is algal (autochthonous) in origin, this effect is limited to about 1000+500 years, which can be corrected, at least for young deposits. The other major problem with dating Lake Baikal sediments is the very low carbon contents of glacial-age deposits, which makes them extremely susceptible to contamination with modern carbon. This problem can be minimized by careful sampling and handling procedures. The ages show almost an order of magnitude difference in sediment-accumulation rates among different sedimentary environments in Lake Baikal, from about 0.04 mm/year on isolated banks such as Academician Ridge, to nearly 0.3 mm/year in the turbidite depositional areas beneath the deep basin floors, such as the Central Basin. The new AMS ages clearly indicate that the dramatic increase in diatom productivity in the lake, as evidenced by increases in biogenic silica and organic carbon, began about 13 ka, in contrast to previous estimates of 7 ka for the age of this transition. Holocene net sedimentation rates may be less than, equal to, or greater than those in the late Pleistocene, depending on the site. This variability reflects the balance between variable terrigenous sedimentation and increased biogenic sedimentation during interglaciations. The ages reported here, and the temporal and spatial variation in sedimentation rates that they imply, provide opportunities for paleoenvironmental reconstructions at different time scales and resolutions.