Species delimitation in Taxus is controversial because morphological differences are mostly vegetative; nevertheless, a study of leaf characters was made of more than 1,000 specimens throughout the entire range of the genus. Evidence will be presented to show ecogeographical patterns of character traits that seem best explained by a long history of evolution. Diversity appeared greatest in E Himalayas, least in America. Some disjunct differences suggest a former continuous distribution in Laurasia, e.g.--epidermal cells angular in x-section in specimens from E Himalayas (e.g., T. wallichiana) and W America (e.g., T. brevifolia), while other differences seem ecogeographical within a continent; e.g., a reddish discoloration in specimens from subtropical Asia. Of evolutionary significance is clinal variation in numbers of stomata rows and marginal cells; for instance, specimens from SW China were found with up to 21 stomata rows/band, wherein W America, this number ranged from 8-9 in El Salvador to southern Mexico, (5-)6-8 in Nuevo Leon to Florida, and from 4-7 in California to Alaska; within the latter range, Californian specimens often had higher counts. A reverse cline, however, was observed in NE America (T. canadensis); where 5-6 stomata rows/band were usually seen, 7-8(-9) were occasionally found in specimens from Newfoundland. Taxus perhaps originated in central Asia during the Jurassic, spread to W America via Pacific during the Cretaceous, and to E America via Atlantic during the Tertiary. While some differences in character states might be interpreted as having evolved since the Jurassic, hybridization is also evident throughout most of the range of the genus. Nevertheless, character traits that have become established over wide geographical areas would seem to have taxonomic validity for species recognition.

Key words: conifers, gymnosperms, systematics, Taxus