ABSTRACT
Introduction
As many frustrated taxonomists working on Lilianae (Liliaceae sensu lato) and other monocotyledons have acknowledged (e.g. Cronquist, 1981), gross floral and inflorescence characters are often of limited systematic value above the genus level in these groups. Although a few families are well defined by floral characteristics (e.g. three stamens in Iridaceae; apical anther dehiscence in 'Fecophilaeaceae), many Lilianae have relatively unmodified flowers with six tepals and six stamens, and either a superior or inferior ovary. On the other hand, some micromorphological characters, such as microsporogenesis type and tapetum type, have proved to be significant at the family and order levels (e.g. Rudall et al., 1997; Furness and Rudall, 1998). The main problem with these characters is that they are cryptic: they are relatively inaccessible since they require specialist techniques and an understanding of the structures involved. There is always a need for good comparative and developmental data before scoring data matrices, but for micromorphological characters data are sometimes scored from the literature by non-specialists. It is also essential to avoid using 'types' of structures based on taxon names, which sometimes mask more than one character.
