Abstract. Campylandra Baker was published in 1875. Subsequently, various authors have either united it with Tupistra Ker Gawler (sometimes recognizing it as a subgenus or section of Tupistra) or maintained it as a separate genus. The present authors regard Campylandra as a separate genus on account of its having filaments only proximally (vs. almost wholly in Tupistra) adnate to the perianth tube, anthers as high as or higher (vs. lower) than the stigma, a shorter style, and a 3-lobed (vs. peltate to mushroom-shaped) stigma; this separation is confirmed by molecular phylogenies. In preparation for the accounts of genera in the Convallariaceae for Flora of China, Volume 24, it is therefore necessary to transfer into Campylandra 12 species hitherto treated under Tupistra, namely: T. annulata H. Li & J. L. Huang, T. chinensis Baker, T. delavayi Franchet, T. emeiensis Z. Y. Zhu, T. ensifolia F. T. Wang & Tang, T. fimbriata Handel-Mazzetti, T. jinshanensis Z. L. Yang & X. G. Luo, T. liangshanensis Z. Y. Zhu, T. lichuanensis Y. K. Yang, J. K. Wu & D. T. Peng, T. longipedunculata F. T. Wang & S. Yun Liang, T. verruculosa Q. H. Chen, and T. yunnanensis F. T. Wang & S. Yun Liang. It is also necessary to transfer two species originally described under a related genus, Rohdea Roth, but later transferred to Tupistra: T. tui (F. T. Wang & Tang) F. T. Wang & S. Yun Liang (R. tui F. T. Wang & Tang) and T. urotepala (Handel-Mazzetti) F. T. Wang & Tang (R. urotepala Handel-Mazzetti). In addition to these transfers, Campylandra ensifolia is lectotypified.

The genus Campylandra was originally described by Baker (1875: 582) with the single species C. aurantiaca Baker. Baker distinguished Campylandra from Tupistra Ker Gawler on the basis of its sessile leaves, presence of staminal filaments, and a smaller stigma. On the other hand, Bentham and Hooker (1883) regarded Campylandra as no more than a variant of Tupistra with somewhat more developed filaments and a more slender style, and reduced Campylandra to synonymy under Tupistra. Later, Engler (1887, 1897), Clarke (1890), Krause (1930), Dandy (1933), Chun (1934), Hutchinson (1934), Handel-Mazzetti (1936), Hara et al. (1978), and Conran and Tamura (1998) all recognized Campylandra as an independent genus. They evaluated it with varying circumscriptions, but as having, for example, sessile leaves, bracts longer than the flowers, tepals frilled at the margin, anthers positioned higher than or as high as the stigma, presence of filaments and/or a small, trilobed stigma. However, Hooker (1892), Baillon (1893), Franchet (1896), Wang and Liang (1976, 1978), and Huang and Li (1990) all followed Bentham and Hooker (1883) in regarding Campylandra as congeneric with Tupistra. Among them, Franchet (1896) and Huang and Li (1990) recognized Campylandra at the ranks of section and subgenus, respectively.

In the forthcoming Flora of China, Volume 24, Liang and Tamura (in press) will treat Campylandra and Tupistra as separate genera and stress the morphological differences as follows:

Campylandra: filaments proximally adnate to perianth tube; anthers positioned as high as or higher than stigma; style usually 0--1 mm, rarely to 3.5 mm (C. fimbriata); stigma small, 3-lobed.

Tupistra: filaments almost wholly adnate to perianth tube; anthers positioned lower than stigma; style 4--12 mm, cylindric; stigma peltate to fungilliform, 2--7 mm in diameter, fleshy.

According to Huang and Li (1990) and Huang and Hong (1997), Campylandra has unimodal karyotypes, and ellipsoid or subspherical, monosulcate pollen grains with reticulate or perforate exine, whereas Tupistra has bimodal karyotypes (except in T. nutans Wallich ex Lindley: unimodal), and spherical, inaperturate pollen grains with rugulose, verrucose, or gemmate exine. Thus the separation of Campylandra and Tupistra is more or less supported by karyological and pollen morphological data, although a comprehensive karyological study is needed to confirm this. The separation of the two genera has also been supported by molecular phylogenies based on both trnK (including matK) and rbcL genes (Yamashita & Tamura, in press). At the present stage, we consider that separation is better than inclusion of Campylandra within Tupistra. However, the detailed relationship of species and consequent generic circumscription of Campylandra, Gonioscypha Baker, Rohdea, Tricalistra Ridley, and Tupistra should be reexamined in the future. In the present paper, we propose the following 14 new combinations in Campylandra for the treatment in Flora of China, Volume 24. These species were hitherto treated under Tupistra, but agree with Campylandra as defined here. Two of them were originally described under a related genus, Rohdea Roth, but were transferred to Tupistra in Wang and Liang (1978). The key to genera, generic descriptions, key to species, and specific descriptions of both Campylandra and Tupistra will be provided in the Flora.

Wang and Tang cited two syntypes: the flowering specimen cited above (as "type") and the following specimen (as "fruit type"): China. Yunnan: "Shang-pa," 2000 m, 21 Oct. 1934 (fr), H. T. Tsai 58892 (PE). The latter specimen was preserved in the herbarium of the Fan Memorial Institute of Biology, Peiping (Beijing), which has since been transferred to KUN, NAS, and PE. The specimen at Kew has two inflorescences and is here chosen as the lectotype because flowering material is more useful diagnostically than fruiting material.

The holotype specimen was preserved in the herbarium of the Fan Memorial Institute of Biology, Peiping (Beijing), which has since been transferred to KUN, NAS, and PE.

Acknowledgment. We thank Anthony R. Brach (MO c/o Harvard) for help in locating literature.

Literature Cited