The taxonomic limits of Synstemon are established, and its relationship to Dontostemon is discussed. Synstemon lulianlianus, a new species from Gansu, is described and illustrated. Previously described taxa in Synstemon are reduced to synonymy of S. petrovii. The genus Synstemon was established by Botschantsev (1959) to include one species and two varieties. It was compared with Cymatocarpus O. E. Schulz and Arabidopsis Heynhold and was said to differ from these in having ciliate petal claws, connate and pubescent median filaments, and margined seeds. Synstemon is unrelated to Cymatocarpus or Arabidopsis, and it also differs from both by having flattened and strongly crisped instead of the rigid and straight trichomes, strongly differentiated instead of undifferentiated petal claws, and apiculate instead of obtuse anthers. Cymatocarpus also has yellow flowers, long setose trichomes, and submoniliform fruits, whereas Synstemon has lavender flowers, soft trichomes, and torulose fruits. In our opinion, Synstemon is unrelated to these genera. Botschantsev (1959) distinguished Synstemon from Dontostemon Andrzejowski ex C. A. Meyer because of the former's ciliate petal claws, veinless septa, and mucilaginous seeds, but a closer examination reveals that the difference in the fruit septum (veined vs. veinless) is unreliable, and most of the 10 species of Dontostemon have veinless septa. Synstemon differs in having pilose petal claws, entire stigmas, mucilaginous seeds, median stamens united only at the very base, terete replum, median nectaries present and confluent with the laterals, and flattened, strongly crisped trichomes, as well as in lacking the multicellular glands seen in some species of Dontostemon. By contrast, Dontostemon has glabrous filaments and petal claws, 2-lobed stigmas, nonmucilaginous seeds, median filaments united near to the apex, flattened replum, no median nectaries, straight or slightly twisted terete trichomes, and (in four species) well-developed multicellular glands. The two genera share many characters, and it appears that Synstemon is more closely related to Dontostemon than to any other Asian genus of Brassicaceae (Cruciferae). Because of fear that Synstemon might be considered a later homonym of Synostemon F. Miller, which is a synonym of Sauropus Blume of the Euphorbiaceae, Botschantsev (1980) proposed Synstemonanthus as a replacement of Synstemon. However, the pronunciations of the two are quite different, and we agree with Brummitt (1992) in recognizing Synstemon instead of Synstemonanthus. The presence in Synstemon of pilose petal claws and filaments is unique among all of the Old World Brassicaceae. This feature clearly supports the maintenance of the genus as independent of Dontostemon, as well as those Eurasian genera with connate median filaments. Pubescent petal claws and/or filaments are known among several genera in the New World, including the North American Stanleya Nuttall, Warea Nuttall, Rollinsia Al-Shehbaz, and a species of Dryopetalon A. Gray (AlShehbaz, 1982, 1985; Payson, 1923; Rollins, 1939, 1993), as well as the South American Chilocardamum 0. E. Schulz, Pterygiosperma O. E. Schulz, Sarcodraba Gilg & Muschler, Sisymbrium subscandens Spegazzini, and Trichotolinum O. E. Schulz (Boelcke & Romanczuk, 1984; Schulz, 1924, 1936). Synstemon remained monotypic until An (1981) added S. linearifolius An (as linearifolia) and S. petrovii var. xinglongnicus An (as xinglongnica). These two new taxa, along with Botschantsev's NOVON 10: 99-103. 2000. This content downloaded from 157.55.39.183 on Thu, 26 May 2016 05:36:53 UTC All use subject to http://about.jstor.org/terms
Read full abstract