We address two aspects of Cretaceous plants and climate. Firstly, we briefly characterize Cretaceous global phytogeography and review some quantitative temperature estimates derived from plants. Secondly, by adopting a multivariate statistical approach to palaeophytogeographic mapping, we examine the effect of the rise in diversity and ecological radiation of angiosperms on the climate signal offered by non-angiosperms. In the Cretaceous, plant productivity was concentrated at middle and high latitudes. Polar cool temperate rain forests in coastal areas, where winter temperatures were ameliorated by the proximity of the ocean, were conifer-dominated and deciduous. In more continental high latitudes, winter temperatures probably fell well below freezing, thermally depressing metabolic rates and allowing some plants to retain their leaves year round. However, this `evergreenness' necessitated desiccation resistance in the form of reduced leaf size and thick cuticles. At mid latitudes, open canopy woodlands and forests were dominated by a mixture of microphyllous conifers, moderately xeromorphic ferns, cycadophytes, pteridosperms and sphenophytes. Late Cretaceous broadleaved coriaceous angiosperms were mostly subordinate shrubs and small trees. Large angiosperm trees were comparatively rare. Although the evidence is meagre, low latitude vegetation tended to be xeromorphic and only patchily forested. Tropical everwet vegetation was, if present at all, highly restricted and consequently not understood. Non-angiosperms exhibit a weak, poorly differentiated phytogeographic pattern linked most strongly to the evaporation:precipitation ratio. This relationship is reflected in foliar physiognomy, and was established long before the Cretaceous. As the advent of the angiosperms did not significantly alter this relationship, it may be possible to calibrate the non-angiosperms of the Mesozoic using the physiognomic techniques developed for woody dicots. Taxonomy independent of political and stratigraphic boundaries is essential in order to satisfactorily reconstruct palaeophytogeography and derivative climatic signals, as are non-selective reporting of floral composition and detailed descriptions of plant physiognomy. Without this basic information the full potential of the plant fossil record to reveal past climate and environmental change cannot be exploited.