Five sources of data force extensive revision of ideas about the nature and evolution of monocot xylem: scanning electron microscopy (SEM) studies of thick sections; availability of molecular phylogenies covering a relatively large number of families and genera; information on ecology and habitat; data concerning habit; and observations from xylem physiology. These five new sources of data, absent from the studies of Cheadle, plus added information from light microscopy, lead to a fresh understanding of how xylem has evolved in monocots. Tracheary elements hitherto recorded as vessel elements with scalariform end walls prove in a number of instances, to retain pit membranes (often porous or reticulate) in the end walls. There is not an inexorable progression from "primitive" to "specialized" xylem in monocots; apparent accelerations or reversions are also possible. The latter include such changes as the result of production of narrower vessel elements; or production of less metaxylem, which is probably heterochronic in nature (an extreme form of juvenilism). Tracheary elements intermediate between vessel elements and tracheids must be recognized for what they are, and not forced into mutually exclusive categories. Original data on tracheids and various types of vessel elements is related here to ecology and habit of groups such as Asteliaceae, Boryaceae, Cyclanthaceae, Orchidaceae, Pandanaceae, Taccaceae, Typhaceae, dracaenoid Asparagaceae, and Zingiberales. Data from palm xylem shows a nearly unique syndrome of features that can be explained with the aid of information from physiology and ecology. Vessellessness of stems and leaves characterizes a large number of monocot species; the physiological and ecological significance of these is highlighted. An understanding of how non-palm arborescent monocots combine an all-tracheid stem xylem with addition of bundles and vegetative modifications is attempted. The effect of the disjunction between xylems of adventitious roots and stems, providing a physiologically demonstrated valve ("rectifier") effect is discussed. "Ecological iteration" has occurred in some monocot lineages, so that early-departing branches in some cases may have more "specialized" xylem because of entry into xeric habitats, whereas nearby crown groups, which may have retained "primitive" xylem, probably represent long occupation of mesic habitats. Cheadle's use of xylem for "negations" of phyletic pathways can no longer be accepted. Symplesiomorphic mesomorphic xylem patterns do characterize many of the earlier-departing branches in the monocots as a whole, however. Cheadle's idea that monocots and non-monocot angiosperms attained vessels independently is improbable in the light of molecular trees for angiosperms. Vessels in roots seem an adaptation to major swings in moisture availability to adventitious roots as compared to taproots. The commonness of all-tracheid plans in stems and leaves in earlier-departing monocot clades is a feature that requires further clarification but is primarily related to the xylem disjunction that adventitious roots have. Secondary vessellessness or something very close to it can be hypothesized for Campynemataceae, Philesiaceae, Taccaceae, and some Orchidaceae. Eleven salient shifts in our conceptual views of monocot xylem are proposed and conclude the paper. Monocot xylem is not a collection of historical information, but a rigorously parsimonious system related to contemporary habits and habitats.
