The Brassicaceae (Cruciferae) is the largest family of the diverse order including the mustard oil glycoside-producing plant families. Morphological evidence (e.g., the presence of tetradynamous stamens, cruciform corolla, radially symmetric flowers, and characteristic fruit often with a false septum dividing it into two locules) and recent molecular phylogenetic studies support the monophyly of Brassicaceae. Nonetheless, relationships within the family remain problematic. Previous classification schemes for the family relied heavily on fruit characters to determine generic and tribal boundaries. Published preliminary molecular data show that fruit characters are highly homoplastic and that groups based on these features often do not reflect phylogeny. As a result, current generic circumscriptions in Brassicaceae make it difficult to identify putatively monophyletic terminal clades appropriate for sampling at the familial level. Recent molecular work at the generic and tribal levels suggests that the traditionally circumscribed tribe, Brassiceae, as well as a narrowly circumscribed Lepideae are monophyletic. We suggest that trichome morphology may be useful in delimiting monophyletic groups of genera in Brassicaceae. For example, the genera Alyssum, Arabidopsis, Arabis, Draba, and Lesquerella share branched trichomes and may form one putatively monophyletic group, while Armoracia, Barbarea, Cardamine, Nasturtium and Rorippa are either glabrous or have simple trichomes, forming another putatively monophyletic group. By integrating current molecular knowledge with detailed morphological investigations, we present a series of hypotheses of evolution in the family, testable using current molecular techniques. Furthermore, we discuss the importance of trichome morphology in predicting relationships in Brassicaceae.

Key words: Brassicaceae (Cruciferae), trichomes