38. Davalliaceae
骨碎补科 gu sui bu ke
Authors: Prof. Fuwu Xing, Wang Faguo & Hans P. Nooteboom
Plants small to moderate-sized, epiphytic, epilithic, or rarely terrestrial. Rhizome long creeping, dorsiventral, densely covered with peltate or basifixed scales; fronds usually in 2 rows. Fronds remote; stipe articulate at base to short or long phyllopodia; lamina simple, imparipinnate, or pinnatifid to 4-pinnate-pinnatifid, firmly herbaceous to leathery, glabrous or sometimes covered with scales or hairs; veins free, usually forked. False veins present in several species. Sporangia typically separately borne in a small discrete sorus terminal on veins, submarginal or sometimes medial [in Humata undulata (Alderwerelt) M. Kato & Tsutsumi connate and elongate along lamina margins]; indusium opening toward margin, attached at base and sometimes at sides, orbicular, reniform, or elongate toward margin. Sporangia long stalked, 3-seriate, annulus longitudinal, consisting of 12-16 thickened cells; spores monolete, elliptic or narrowly elliptic, translucent, usually without perispore.
Five genera (one according to Nooteboom) and ca. 35 species: mostly in tropical and subtropical Asia, a few species extending to Africa, one species in NW Africa, SW Europe, and Macaronesia; four genera and 17 species (three endemic) in China.
One of us (Nooteboom) considers that only one genus, Davallia, can be recognized in the Davalliaceae. Paradavallodes multidentata is nested in Araiostegia + Davallodes, which is a polymorphic basal clade. Among species examined, A. hymenophylloides (Blume) Copeland is the type of Araiostegia, and P. multidentata is the type of Paradavallodes. Therefore, the earliest published-Davallodes (Copeland) Copeland-should be used as the generic name. However, from morphology and DNA there are no arguments to recognize this genus. The Humata clade looks quite distinct as an entity, although it contains Davallia corniculata, renamed to H. corniculata by Kato, and in all other systems this is a Davallia. Tsutsumi, Zhang and Kato did not analyze D. wagneriana Copeland and included it in Davallia, but morphologically this is the closest relative of D. corniculata. It should belong to Humata; or better, Humata is part of Davallia (Tsutsumi, Zhang & Kato, Syst. Bot. 33: 44-48. 2008).
However, two of us (Xing and Wang) consider that morphological and molecular phylogenetic studies do not support there being only one genus (e.g., Sen, Sen & Holttum, Kew Bull. 27: 217-243. 1972; Kato and Tsutsumi, Acta Phytotax. Geobot. 59: 1-14. 2008; and Tsutsumi, Zhang & Kato, Syst. Bot. 33: 44-48. 2008). A recent study about the leaf epidermis and spore morphology of Davalliaceae (He Chunmei, A taxonomic revision of Davalliaceae in China. Graduate University of Chinese Academy of Sciences, Guangzhou. 2012) also supports Araiostegia, Humata, and Paradavallodes as natural genera.
Polyploidy, through hybridization and polyploidization, has been assessed in the Humata group, and these polyploids appeared to be apomictic. This could very well also be the case in other groups. With only cpDNA, the phylogeny of a group like this cannot be assessed. Moreover, in Davalliaceae, incongruence between CpDNA and nrDNA is found on the generic level but is not yet explained (C. W. Chen et al., Insights into Evolutionary History of the Humata repens Complex in Taiwan. 5th Symp. Asian Pteridol. Fern Show: Progr. & Abstr. 42. 2010; C. W. Chen & Nooteboom, ined.).
Ching Ren-chang, Fu Shu-hsia, Wang Chu-hao & Shing Gung-hsia. 1959. Davalliaceae (excluding Arthropteris, Gymnogrammitis, Leucostegia, and Nephrolepis). In: Ching Ren-chang, ed., Fl. Reipubl. Popularis Sin. 2: 280-319, 374-378; Wu Shiewhung. 1999. Davalliaceae (excluding Leucostegia). In: Wu Shiewhung, ed., Fl. Reipubl. Popularis Sin. 6(1): 161-197.
