Bot. Mag. Tokyo 94: 249-259, 1981 A Revision of the Genus Cyathorachis in Japan* HARUFUMI NISHIDA Laboratory of PhylogeneticBotany, Faculty of Science, Chiba University, Yayoi-cho1-33, Chiba 260 Cyathorachis, a genus of Cretaceous tree fern rachises is revised. In addition to the type specimens of the sole species, C. fujiiana Ogura, two specimens collected near the type locality were examined histologically. The histological data being used for the taxonomical comparison were obtained from anatomical studies on the rachises of extant tree ferns. As a result of the study three species are recognized, two of which are new. As a result of this study, it appears that species differentiatioh of the tree ferns in the late Cretaceous age of Hokkaido was rather high, as it is now in tropical and subtropical regions. Key words: Cretaceous - - Cyathorachis- - Histology -- Tree ferns. Cyathoraehis was established in 1927 by Dr. Yudzuru Ogura who described C. fujiiana from the Upper Cretaceous of Hokkaido. It was the sole species of the genus hitherto known. In ]968 and 1972 Dr] Makoto Nishida collected two petrified specimens of tree-fern raehises from the Upper Cretaceous of Oh-Yubari, near the type locality of C. fujiiana. He identified them tentatively as C. fujiiana, though some of their anatomical features did not accord well with the original description of C. fujiiana. Recently I had the opportunity to re-examine the type specimens of C. fujiiana to compare them with M. Nishida's specimens. Ogura cited three specimens, A, B, and C in his original description. A and C were collected at Yubari and B at Ikushumbetsu. These specimens are deposited in the University Museum of the University of Tokyo together with other specimens D and E, which were collected at Yubari and were identi- fied as C. fujiiana but not cited by Ogura. All specimens mentioned above were care- fully examined and compured using histological data obtained from anatomical observations on the raehises of extant tree ferns (H. Nishida, unpublished). It was revealed by this study that the type specimens of C. fujiiana should be separated into two species. That is, specimen B represents one species and specimens A, C, D and E another. Thus C. fujiiana should be emended and a new species established for the other specimen. One of M. Nishida's specimens shows histological features resembling emended C. fujiiana. The other specimen does not belong to any of the above species and represents a third species of the genus. * Contributions from the Laboratory of Phylogenetic Botany, Chiba University, No. 74.