The Disposition of Trichopteris (Cyatheaceae)

Abstract

The delimitation of genera and families has been a persistent problem in fern taxonomy, and the Cyatheaceae sensu stricto is no exception. Christensen (190506) adopted clearly artificial genera (Cyathea, Hemitelia, and Alsophila) based on complete (totally surrounding the sorus), partial, and absent indusia. He included Lophosoria and Metaxya in Alsophila. The latter two satellite genera are distantly related to the major genera of the family, and nowadays are often placed in one or two families of their own. Holttum (1963) proposed a single genus Cyathea for the Flora Malesiana region with two very distinct subgenera, Sphaeropteris and Cyathea. Holttum has maintained (1981, p. 466) that the only subdivision of the genus clearly definable is that between subgenus Sphaeropteris and the rest. This indicates that Alsophila and Cnemidaria are less distinct from Cyathea than all three are from Sphaeropteris, which is confirmed by the lack of hybrids with Sphaeropteris. In studying the species of the Flora Malesiana region, Holttum came to the fundamental conclusion, among many, that indusium type is not an important generic character, for within a few species it varies widely. Tryon (1970) divided the Cyatheaceae sensu stricto on the basis of scale characters, indusium presence or absence, and venation. He adopted the genera Sphaeropteris (scales conform), Alsophila and Nephelea (scales non-conform and setate), Trichopteris (scales non-conform and non-setate, laminae free-veined, and sori exindusiate), Cyathea (scales non-conform and non-setate, laminae freeveined, and sori indusiate), and Cnemidaria (scales non-conform and non-setate and laminae net-veined. According to Holttum and Edwards (1983, p. 179), this classification has assorted closely related species into Cyathea, Sphaeropteris, and Trichopteris. It is apparent that genera based on these characters are not natural. Working from Tryon's (1970) generic concepts, I have found it possible to define readily recognizable and coherent genera in the Cyatheaceae sensu stricto by including in Cyathea the genus Trichopteris and the New World species considered to be Sphaeropteris, except for the S. horrida group. I accept the genera Sphaeropteris, Alsophila (including Nephelea), Cyathea (including Trichopteris), and Cnemidaria. Occasional hybrids occur within Alsophila and Cyathea and between Cnemidaria and Gyathea. This is evidence of a greater degree of relationship than with Sphaeropteris, but in my opinion should not alone be the basis for adopting an inclusive Cyathea (either excluding or even including Sphaeropteris), for the characteristics of Alsophila and Cnemidaria are sufficiently different from those of Cyathea to distinguish the genera readily, and intergeneric hybrids in ferns