Although a division of the Phanerozoic climatic modes of the Earth into "greenhouse" and "icehouse" phases is widely accepted, whether or not polar ice developed during the relatively warm Jurassic and Cretaceous Periods is still under debate. In particular, there is a range of isotopic and biotic evidence that favours the concept of discrete "cold snaps", marked particularly by migration of certain biota towards lower latitudes. Extension of the use of the palaeotemperature proxy TEX₈₆ back to the Middle Jurassic indicates that relatively warm sea-surface conditions (26–30 °C) existed from this interval (∼160 Ma) to the Early Cretaceous (∼115 Ma) in the Southern Ocean, with a general warming trend through the Late Jurassic followed by a general cooling trend through the Early Cretaceous. The lowest sea-surface temperatures are recorded from around the Callovian–Oxfordian boundary, an interval identified in Europe as relatively cool, but do not fall below 25 °C. The early Aptian Oceanic Anoxic Event, identified on the basis of published biostratigraphy, total organic carbon and carbon-isotope stratigraphy, records an interval with the lowest, albeit fluctuating Early Cretaceous palaeotemperatures (∼26 °C), recalling similar phenomena recorded from Europe and the tropical Pacific Ocean. Extant belemnite δ¹⁸O data, assuming an isotopic composition of waters inhabited by these fossils of −1‰ SMOW, give palaeotemperatures throughout the Upper Jurassic–Lower Cretaceous interval that are consistently lower by ∼14 °C than does TEX₈₆ and the molluscs likely record conditions below the thermocline. The long-term, warm climatic conditions indicated by the TEX₈₆ data would only be compatible with the existence of continental ice if appreciable areas of high altitude existed on Antarctica, and/or in other polar regions, during the Mesozoic Era.