An assessment is made of the current state of research into Jurassic sea-level changes based on new insights into depositional patterns in relation to these changes and new stratigraphic information from several widely separated regions across the world. The bearing of this information on alternative models, respectively, involving sea-level rise interrupted by stillstands or falls, is then evaluated. The gross pattern appears to be one of more or less gradual rise of sea level through the period, interrupted by episodes of comparative stillstand rather than eustatic fall and several episodes of significant regression are shown to be the result of regional tectonics. Major episodes of eustatic rise took place in the early Hettangian, early Sinemurian, early Pliensbachian, early Toarcian, early and late Bajocian. middle Callovian and late Oxfordian to Kimmeridgian. A significant episode of rapid and very extensive regression, possibly global, took place at the end of the Triassic, but other regressive episodes, in the late Aalenian, Bathonian, late Oxfordian-Kimmeridgian and Tithonian, are clearly only regional in extent. There is no evidence for glacioeustasy and most if not all of thr regional or global changes recognised can readily be related to plate tectonics. The events at the Triassic-Jurassic boundary seem to be the result of mantle plume activity centred on the Central Atlantic region. (C) 2001 Elsevier Science B.V. All rights reserved.