An analysis of the vegetation of four islands of the Krakatau group (Rakata Besar, Rakata Kecil, Sertung and Anak Krakatau), Indonesia, was conducted based on our two expeditions of 1982 in centennial commemoration of the great explosion in 1883. Pioneer communities on lava flow, ash fields, beaches, cliffs and on a scoriaceous slope were quite different from each other. Forest communities were grouped into nine types using Morisita's index of similarity, Cλ(W). Two new forest types which have never been described from Krakatau were identified: Timonius compressicaulis forest and Dysoxylum caulostachyum forest. Successional relations among these community types and the Neonauclea calycina forest were discussed from the viewpoint of seed morphology, dispersal, and the development of juvenile generations in the three types of forest. Dysoxylum forest is the most advanced forest community on the Krakatau Islands. The climax forest is supposed to be a tropical monsoon forest similar to that of the volcanic island Panaitan. A successional scheme is proposed. The invasion times of three forest dominants on the three old islands of Krakatau are compared and it is concluded that the difference in vegetation between these islands, the two Rakatas and Sertung, is determined by the earlier invasion of species which could then extend their cover and build up a forest thus preventing later invaders to establish.
We are greatly indebted to Dr Fred Hehuwat, Director of the National Institute of Geology and Mining in Bandung, Mr Peter E. Hehanussa, Chief manager for soil and water resource development in the National Institute of Geology and Mining, Dr Soenartono Adisoemarto, Head of the Museum Zoologicum Bogoriense, Dr Kusuwata Kartawinata, Head of the Herbarium Bogoriense, National Institute of Biology, for their kind cooper-ation and use of their facilities. We also thank the staff of the Indonesian Institute of Sciences for their invaluable help in the field and Dr P. J. Grubb who kindly commented on the manu script.
The identification of flowering plants was carried out by T. Partomihardjo, who compared samples with specimens in the Herbarium Bogoriense. The identification was counterchecked by Dr M. Hotta, Associate Professor at the Yoshida College, Kyoto University. Dr C. G. G. J. van Steenis taught us a correct name of a species of Bignoniaceae. Pteridophytes were identified by Dr S. Mitsuda, Dept. of Botany, Faculty of Science, Kyoto University, and bryophytes by Dr T. Seki, Dept. of Botany, Faculty of Science, Hiroshima University.