THE PHYLOGENY ANDCLASSIFICATION OF THEZINGIBERALES'W. John Kress2ABSTRACT In the Zingiberales, a primarily tropical order of monocotyledons, most phylogenists currently recognize eightfamilies: Musaceae, Strelitziaceae, Lowiaceae, Heliconiaceae, Zingiberaceae, Costaceae, Cannaceae, and Marantaceae.Some taxonomists still prefer the earlier classifications that included Strelitziaceae, Lowiaceae, and Heliconiaceae inMusaceae s.l., and Costaceae as a part of Zingiberaceae s.l. Attempts to reconstruct the phylogenetic history of theorder have been made by Lane, Tomlinson, and Dahlgren & Rasmussen. An original analysis of the evolutionaryrelationships of the eight families of the Zingiberales based on the principles of phylogenetic systematics is presentedhere. The most parsimonious topology is (Musaceae (Strelitziaceae (Lowiaceae (Heliconiaceae ((Zingiberaceae, Cos-taceae) (Cannaceae, Marantaceae)). The cladogram rejects the recognition of Musaceae s.l. as an evolutionary group.A new phylogenetic classification based on the cladogram is proposed that recognizes eight families, two superfamilies,and five suborders within the Zingiberales. "The Scitamineae [Zingiberales] is a very nat-ural order of monocotyledons, somewhat compa-rable in its homogeneity to grasses, palms or or-chids... , long recognized by taxonomists as anatural entity" (Tomlinson, 1962). "The Order Zingiberales is well characterizedand sharply defined; its limits have occasioned nocontroversy .... One could wish that the familiesin ail orders were as well marked and sharplydefined as those in the Zingiberales" (Cronquist,1981). "The Zingiberiflorae, whether treated as a sep-arate superorder, as here, or an order in a morewidely circumscribed unit, is one of the most in-disputably natural suprafamilial groups" (Dahl-gren, Clifford & Yeo, 1985). The Zingiberales, or Scitamineae, are a groupof monocotyledons whose members are almost en-tirely restricted to tropical regions. As indicatedby the above quotations, the order is widely ac-cepted by most taxonomists and phylogenists to bea distinctly circumscribed "natural" or monophy-letic lineage of plants. No morphological charactersare in conflict with the acceptance of the Zingi-berales as a monophyletic group. In a brief diagnosis of the morphological andanatomical characters that distinguish the order,Tomlinson (1962) included the following: rhizo-matous herbs; leaves with open, sometimes ligulatesheaths; lamina entire with lateral veins divergingfrom a common midrib, one-half of blade com-pletely rolled around the other during development;hairs commonly unicellular; guard cells each withtwo narrow lateral subsidiary cells parallel to thepore; hypodermis of colorless cells below each sur-face of lamina; air canais in leaf axis segmentedby transverse diaphragms containing stellate cells;leaf axis with a single main arc of large vascularbundles and subsidiary systems of smaller bundles;silica cells or stegmata associated with vascularbundles in ail parts except roots; inflorescence ter-minal or lateral, commonly racemose with con-spicuous bracts; flowers (Fig. 1) zygomorphic, peri-anth consisting of separate calyx and corolla; fertilestamens usually five or one; one to five stamensusually represented by staminodes; ovary inferior,three-locular, with one to many ovules in eachlocule; seeds with abundant endosperm, often ar-illate. In reviewing these characters and subsequentlystudied features, Dahlgren and coworkers (1985)listed six apomorphies for the Zingiberales (their ' I thank D. Stone, B. Kirchoff, L. Andersson, and H. Kennedy for insightful comments on the Zingiberales, E.Christenson, T. Walters, D. Stone, and L. Eibest for editorial comments, P. Eckel for help with the Latin diagnoses,and B. Culbertson for her lucid illustrations. Cheryl Roesel was especially helpful in every facet of preparation of themanuscript. This work was supported by the United States National Science Foundation grants BSR-8306939 to D.E. Stone and W. J. Kress, and BSR-8706524 to W. J. Kress. Professor P. Barry Tomlinson's enthusiasm for large monocots provided the original inspiration for my studies ofthe Zingiberales. I dedicate this paper to him.2Department of Botany NHB-166, Smithsonian Institution, Washington, D.C. 20560, U.S.A.ANN. MISSOURI BOT. GARD. 77: 698-721. 1990.