J Syst Evol ›› 1994, Vol. 32 ›› Issue (3): 283-289.

• Research Articles •

### On Some Problems of Methodology of Bamboo Classification with Special Reference to the Circum-scription of Dendrocalamus

Li De-zhu

• Published:1994-05-18

Abstract: The taxonomy of Chinese bamboos has grown so rapidly since the end of 1970s that more and more new bamboos are discovered and many new names are therefore published. It might lead to the accomplishment of the subfamily Bambusoideae in the 《Flora Reipublicae Popularis Sinicae》 (FRPS) after the effort of several generations of Chinese taxonomists for over half a century. Meanwhile there remain some critical conflicts of the methodology and treatments in the classifications of Chinese bamboos among taxonomists on the mainland as well as between those in China and abroad. This includes the taxonomy of alpine bamboos, that of Arundinaria and its allies, and the problem of caespitose bamboos. In approaching the taxonomy of Dendrocalamus, reminded again by W. T. Lin´s two recent papers, some methods used and treatments proposed by Lin and others are discussed. Based on the selection of characters, theory and practice of numerical taxonomy, and principles of cladistics, the circumscription of Dendrocalamus is discussed. The present author believes that it is a monophyletic genus. In other words, the generic concept of Dendrocalamus accepted by Gamble, Chia & H. L. Fung follows a natural group. The attribution of Sellulocalamus W. T. Lin (1989) and Patellocalamus W. T. Lin (1989), based on the species treated in Dendrocalamus by most authors, and their relationships with the latter, are examined according to Clayton´s (1983) operational rules, including selection of characters, taxonomic distance between cluster centroides, clarity of moating, and predictive value. The result of numerical taxonomy showed similarities between the members within Dendrocalamus, while it could not be interpreted as phylogeny as Lin expected. Multidisciplinary data are needed for phylogenetic reconstruction, whereas the guess of phylogeny on the base of similarities of one or two vegetative characters can not be reliable. Taxonomy is a branch of science, and should be testable and repeat able. And a taxonomic treatment should be referred to all members of the group, including those in foreign countries. Finally, the generic names, Sellulocalamus and Patellocalamus, and three specific names under them, are merged into Dendro-calamus as synonyms.