Pit membrane remnants in perforation plates of Hydrangeales with comments on pit membrane remnant occurrence, physiological significance and phylogenetic distribution in dicotyledons

Perforation plates from ten species of seven genera of Hydrangeales sensu Thorne were studied using scanning electron microscopy (SEM). The presence of pit membranes in perforations ranges from abundant, as in Carpenteria and Hydrangea, to minimal, as in Deutzia, Escallonia and Philadelphus. Abnormally great pit membrane presence may result from the presence of secondary compounds that inhibit lysis, as in Quintinia serrata; such interference with the natural lysis process may or may not be evident from coarseness and irregularity of pit membrane surface and of threads composing the pit membrane remnants. The presence of pit membrane remnants in perforation plates is hypothesized to be a symplesiomorphy, found in a fraction of dicotyledons with scalariform perforation plates (but still in an appreciable number of species). Pit membrane remnant presence may represent incomplete lysis of primary wall material (cellulose microfibrils) in species that occupy highly mesic habitats, where such impedance in the conductive stream does not have an appreciable negative selective value. This physiological interpretation of pit membrane remnants in perforations is enhanced by the phylogenetic distribution as well as the strongly mesic ecological preferences of species that exemplify this phenomenon in dicotyledons at large. Families with pit membrane presence in perforations are scattered throughout phylogenetic trees, but they occur most often in basal branches of major clades (superorders) or as basal branches of orders within the major clades. Further study will doubtless reveal other families and genera in which this phenomenon occurs, although it is readily detected only with SEM. Phylogenetic stages in the disappearance of pit membrane remnants from perforation plates are described, ranging from intact pit membranes except for presence of pores of various sizes, to presence of membrane remnants only at lateral ends of perforations and in one or two perforations (arguably pits) at the transition between a perforation plate and subadjacent lateral wall pitting. Developmental study of the mechanism and timing of lysis of pit membranes in perforations, and assessment of the role of the conductive stream in their removal, are needed to enhance present understanding of vessel evolution. © 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146 , 41–51.