Morphological, cytological, and molecular data support reduction of Boleum Desvaux and Euzomodendron Cosson to synonymy of Vella L. The new combination Vella bourgaeana is proposed. A key to species of Vella is presented. Boleum and Euzomodendron have generally been recognized as monotypic Spanish genera, and B. asperum (Persoon) Desvaux was originally described as Vella aspera Persoon. on other hand, includes five species distributed in Spain, Morocco, and Algeria. In his comprehensive revision of tribe Brassiceae, Schulz (1923) placed Boleum, Carrichtera DC., Psychine Desfontaines, Rytidocarpus Cosson (as Distomocarpus O. E. Schulz), Schouwia DC., and Succowia Medikus in subtribe Vellinae, and placed Euzomodendron, Savignya DC., and Oudneya R. Brown in subtribe Savignyinae. Schulz (1919, 1936) separated these subtribes primarily on basis of presence vs. absence of median nectar glands and seed wing: Vellinae have median glands and lack seed wing, whereas Savignyinae lack median nectar glands and have seed wing. G6mez-Campo (1978), however, indicated presence of vestigial wings in Boleum and Vella and suggested that Euzomodendron should be placed with these two genera in one group. He also suggested (G6mez-Campo, 1980) that Savignyinae be united with Vellinae. Except for shape of valvular segment and presence of a well-developed seed wing, Euzomodendron resembles Vella and Boleum in almost all other morphological characters, especially woody habit, connation of inner filaments in pairs, and presence of short pedicels, saccate lateral sepals, long petal claws, dark-veined petal blades, seedless flattened beaks, strongly 3or 5-veined valves, and acutely notched cotyledons (Table 1). In fact, G6mez-Campo (1978) indicated that the robust valves [of Euzomodendron] with narrowed base resemble that of Vella, and G6mez-Campo and Tortosa (1974) showed that cotyledonary notch is indistinguishable in three genera. However, not only did recent floristic accounts (e.g., G6m z-Campo, 1993; L6pez-Gonzilez, 1993; Marcos-Samaniego, 1993; Heywood, 1993) maintain three genera but also placed Euzomodendron remotely from Boleum and Vella. These three genera have been maintained historically because of taxonomic value placed on fruit length, fruit dehiscence, and seed wing. In our opinion, overwhelming similarities in all other characters have been either ignored or given insufficient weight. The connation of inner filaments into pairs in Euzomodendron, Boleum, and Vella is a remarkable synapomorphy not found elsewhere in tribe Brassiceae. Several large genera (e.g., Draba L., Arabis L., and Erysimum L.), and even a smaller one such as Lobularia Desvaux (4 spp.; Borgen, 1987), include taxa with winged or wingless seeds, whereas Draba, Rorippa Scopoli, and Leavenworthia Torrey, to name just a few, include species with both globose and narrowly oblong to linear fruits (for many examples, see Rollins, 1993). Therefore, fruit shape and presence of seed wing often do not warrant delimitation of genera, especially when used alone. The similarities between Boleum and Vella are even stronger than between these and Euzomodendron. The most reliable characters used to separate first two are presence in Boleum of short-pedicellate, indehiscent, sessile fruits, and in Vella of slightly longer-pedicelled, dehiscent, gynophorate or sessile fruits (Table 1). These characters do not justify maintenance of these genera. In fact, dehiscent and indehiscent fruits can be found within Tetracme Bunge, Coronopus Zinn, Sterigmostemum M. Bieberstein, and Ornithocarpa Rose (or even on same plant, as in Diptychocarpus strictus (DC.) Trautvetter, Cardamine chenNovoN 8: 321-325. 1998. This content downloaded from 207.46.13.92 on Wed, 30 Nov 2016 05:02:58 UTC All use subject to http://about.jstor.org/terms