Cardosoa, a new genus of the subtribe Anisopappinae (Athroismeae, Asteraceae)

Ortiz, S. 2010. Cardosoa, a new genus of the subtribe Aniso pappinae (Athroismeae, Asteraceae). Anales Jard. Bot. Madrid.


Introduction
In 1995 Ortiz & Paiva described the species Anisopappus athanasioides (Asteraceae) from somewhat immature material collected by H. Cardoso in Chianga (Huambo), on the Angola Plateau. Eldenäs & Anderberg (1996: 174) considered the specimen from which it was described to be "significantly different from all other species of Anisopappus". We subsequently found more complete material of this taxon and discovered that, besides its distinct capitula and inflorescence, it has bilabiate (i.e. with one adaxial and three abaxial lobes) marginal florets when mature, very different to those of Anisopappus Hook & Arn., which always are true ray florets. In view of these and other very marked morphological differences with respect to Anisopappus and other related members of the Astera ceae, we describe a new genus for this species.

Material and methods
Morphological characters were studied in herbarium specimens of the 18 species of Anisopappus (including A. athanasioides) from BM, COI, K, LISC, LISE, LISU, LUA, LUAI, P and PRE.
Macromorphological characters were studied with the aid of a light microscope. For microscopic examination of floral parts, these were first boiled in water with a surfactant, and subsequently mounted in Hoyer's solution (Anderson, 1954). Cypsela sections were cut by hand with razor blades. Micromorphological and anatomical characters were studied with a compound light microscope.

S. Ortiz
The generic name honors Helder Cardoso, the first collector of this plant, and João Cardoso, who has accompanied us and provided invaluable help during our collecting trips in Angola.

Discussion
As a result of a molecular phylogenetic analysis using the cpDNA marker ndhF, Eldenäs & al. (1999) suggested that Anisopappus should be assigned to Heliantheae s.l. instead of Inuleae as proposed by seve ral authors in the last decade (Anderberg, 1991;Anderberg, 1994;Eldenäs & Anderberg, 1996;Ortiz & al., 1996;Eldenäs & al., 1998). In that analysis Anisopappus forms a trichotomy with the Athroisma-Blepharispermum group and with the remaining genera of Heliantheae s.l. Panero & Funk (2002), on the basis of a phylogenetic analysis of the entire Asteraceae family using a large number of chloroplast DNA markers, suggested that the group Athroisma-Blepharispermum, already found by Kim & Jansen (1995) in their analysis of the Asteraceae using the marker ndhF, should be considered a tribe, which they denominate Athroismeae (including the genera Athroisma DC., Blepharispermum Wight ex DC., and Leucoblepharis Arn.), phylogenetically situated between a) a clade including the tribes Inuleae and Plucheeae -the latter currently included in Inuleae by its author (Anderberg & Eldenäs, 2007;Anderberg, 2009) -and b) a large monophyletic group including the rest of subfamily Asteroideae (Heliantheae s.l.). A similar position for this tribe is also supported by the phylogenetic analyses of Panero & Funk (2008). These authors had already noted the possibility of including Anisopappus in this tribe. Also recently Panero (2005) described the new subtribe Anisopappinae, with the sole genus Anisopappus, within the tribe Athroismeae. In 2007 Panero also included in Anisopappinae the genus Welwitschiella O. Hoffm. although more recently Anderberg (2009) removed this genus from the Athroismeae based on his preliminary molecular phylogenetic analyses which nested it within the tribe Astereae as was confirmed by Brouillet & al. (2009). Cardosoa is very probably a member of that subtribe taking into account its morphological affinities to Anisopappus and we are currently carrying out a molecular phylogenetic analyses using the markers ITS and ndhF to check the relationships of Cardosoa.
Differences in inflorescence morphology can be readily used to separate Cardosoa from its putative sister taxon Anisopappus. Cardosoa has racemiform cymes (Fig. 1a) whereas the genus Anisopappus, has corymbose or subcorymbose cymes (Fig. 1d), sometimes with capitula on long peduncles that appear The involucre of Cardosoa is cyathiform (Fig. 1a), whereas in Anisopappus it is campanulate to hemispheric (Fig. 1d). Only A. smutsii Hutch. and A. pseudopinnatifidus S. Ortiz & Paiva have an involucre similar to that of Cardosoa.
The marginal florets of Cardosoa are bilabiate (one adaxial and three abaxial lobes) (Fig. 1b) or very rarely pseudobilabiate (one adaxial and four abaxial lobes), whereas the marginal florets of Anisopappus are always true ray florets (Fig. 1d). In addition, the marginal florets of Anisopappus are much larger than those of Cardosoa (abaxial limb 4-13 mm long in Anisopappus versus 1.5-2 mm long in Cardosoa), though in some species (such as A. pinnatifidus (Klatt) O. Hoffm., A. pseudopinnatifidus and A. chinensis subsp. scrophulariifolius (Baker) S. Ortiz, Paiva & Rodr. Oubiña) the ray florets may be absent. Cardosoa also differs from Anisopappus in the nerves of the apical lobes of the abaxial lip, which remain sepa-S. Ortiz rate to the base of the limb (Fig. 2a), whereas in Anisopappus these nerves normally join close to the sinuses (Fig. 2b).
The disc florets of Cardosoa are zygomorphic, with the depth of the sinuses varying among lobes (Figs. 1c,2c); in Anisopappus, by contrast, the disc florets are actinomorphic. In Cardosoa, the lobes are conspicuously more acuminate, with the nerve clearly extending to the apices (Fig. 2a); we have not observed this character in any of the species of Anisopappus.
The anthers of Cardosoa have a conspicuously apiculate apical appendage whereas in Anisopappus the apical appendage is acute, not apiculate.
The cypselae are not fully mature in the Cardosoa material available, but appear to differ markedly from those of Anisopappus, to judge from the number of nerves and incipient ribs: in Cardosoa the number of nerves per cypsela seems always to be less than or equal to 5, and the number of ribs less than 10, whereas in Anisopappus the number of nerves is characteristically 5-10(20?), and the number of ribs 10-20. The single exception to this is A. latifolius (S. Moore) B.L. Burtt, a distinctive species found to be sister to the other species of Anisopappus in the morphological cladistic analyses of Eldenäs & Anderberg (1996).  glish translation, to João Cardoso for his invaluable help during the collecting trips in Angola, and to the keepers of the herbaria mentioned for the loan of studied material.