Stackhousiaceae

Abstract: In traditional classifications Celastrales comprised families of usually woody plants with simple leaves, haplostemonous flowers with a disk and apotropous ovules which, in view of this character combination, formed an utterly heterogeneous assemblage. In the classification of Engler and Gilg (1912), for instance, families as divergent as Buxaceae, Rhamnaceae, Aquifoliaceae and Balsaminaceae were dumped into their Sapindales (= Celastrales). In later classifications, most authors gave up this broad circumscription but the Celastrales of Takhtajan (1987), to give just one example, comprised 12 families of which nine, according to our present knowledge, would have to be excluded from this order. In the absence of convincing morphological evidence, only sequ-ence-based phylogenetic studies have led to the recognition of a monophyletic order Celastrales which represents a clade of eurosids I that is sister to Oxalidales + Malpighiales, these three being sister to all remaining eurosids I (Fagales, Rosales, Zygophyllales, Curcurbitales, Fabales; Savolainen, ay et al. 2000; Soltis et al. 2000; APG II 2003). The Celastrales clade comprises only three families, Lepidobotryaceae, Celastraceae s.l. and Parnassi-aceae. The rbcL analysis by Savolainen, Fay et al. (2000) provided strong support for a position of Lepidobotryaceae as sister to the other two families. Parnassiaceae, comprising Parnassiaand Lepuropetalum, have often been included in Saxifragaceae. There exists, however, strong morphological (see Simmons on Parnassiaceae, this volume) and molecular evidence for their exclusion from Saxifragaceae. Various analyses of plastid and nuclear genes (Savolainen, Chase et al. 2000; Soltis et al. 1997, 2000) have resolved Parnassiaceae as sister to Celastraceae. In a multi-gene analysis of Celastraceae (Simmons et al. 2001b), Parnassia and Lepuropetalum have been resolved as members of an early branching but weakly supported lineage of that family, in which they are sister to Perrottetia and Mortonia. The latter two genera, as well as the early-derived Quet-zalia, are somewhat anomalous among Celas-traceae in lacking an aril in favour of a sarcotesta, and partly in possessing scalariform vessel perforations. For the time being it seems therefore justified to retain family status for Parnassiaceae, as suggested by Simmons in his contribution to this volume. As a result of Simmons’ (2001a, 2001b) analysis, Celastraceae are now re-circumscribed to comprise the genera Brexia, Canotia, Plagiopteron, Siphonodon, Stackhousiaceae and Hippocrateaceae, all of which at one time or another had been related to Celastraceae, and all of which have now been shown to be nested within that family.