Chapter42 Coreopsideae Daniel J. Crawford, Mes! n Tadesse, Mark E. Mort, "ebecca T. Kimball and Christopher P. "andle HISTORICAL OVERVIEW AND PHYLOGENY In a cladistic analysis of morphological features of Heliantheae by Karis (1993), Coreopsidinae were reported Morphological data to be an ingroup within Heliantheae s.l. The group was A synthesis and analysis of the systematic information on represented in the analysis by Isostigma, Chrysanthellum, tribe Heliantheae was provided by Stuessy (1977a) with Cosmos, and Coreopsis. In a subsequent paper (Karis and indications of “three main evolutionary lines” within "yding 1994), the treatment of Coreopsidinae was the the tribe. He recognized ! fteen subtribes and, of these, same as the one provided above except for the follow- Coreopsidinae along with Fitchiinae, are considered ing: Diodontium, which was placed in synonymy with as constituting the third and smallest natural grouping Glossocardia by "obinson (1981), was reinstated following within the tribe. Coreopsidinae, including 31 genera, the work of Veldkamp and Kre# er (1991), who also rele- were divided into seven informal groups. Turner and gated Glossogyne and Guerreroia as synonyms of Glossocardia, Powell (1977), in the same work, proposed the new tribe but raised Glossogyne sect. Trionicinia to generic rank; Coreopsideae Turner & Powell but did not describe it. Eryngiophyllum was placed as a synonym of Chrysanthellum Their basis for the new tribe appears to be ! nding a suit- following the work of Turner (1988); Fitchia, which was able place for subtribe Jaumeinae. They suggested that the placed in Fitchiinae by "obinson (1981), was returned previously recognized genera of Jaumeinae ( Jaumea and to Coreopsidinae; Guardiola was left as an unassigned Venegasia) could be related to Coreopsidinae or to some Heliantheae; Guizotia and Staurochlamys were placed in members of Senecioneae. Melampodiinae; Jaumea was put in Flaveriinae; Microlecane In his revision of the tribal and subtribal limits of was kept as a synonym of Bidens following the work of Heliantheae, based on morphology, anatomy and chro- Mes! n Tadesse (1984); Sphagneticola was assigned to mosome studies, "obinson (1981) kept Coreopsideae as Verbesininae; and Venegasia was placed in Chaenactidinae. a synonym of Coreopsidinae. Bremer (1987), in his stud- Stuessy (1988) transferred two species of Oparanthus and ies of tribal interrelationships of Asteraceae, provided one species of Petrobium to Bidens. Shannon and Wagner Coreopsideae as a subset of Heliantheae. He retained many (1997) reinstated Oparanthus and recognized four species of the genera assigned to Coreopsidinae by Stuessy (1977a) in the genus. The genera recognized in this chapter are except for Guizotia (placed in Milleriinae), Guardiola (in a the same ones recognized by Karis and "yding (1994) with new tribe Guardiolinae; "obinson 1978), Jaumea (placed the exceptions that Megalodonta is not segregated from in Jaumeinae), Selleophytum (as a synonym of Coreopsis), Bidens and Selleophytum is segregated from Coreopsis. Staurochlamys (placed in Neurolaeninae), and Venegasia (in The ! rst cladistic analysis focusing on Coreopsideae Chaenactidinae). was by "yding and Bremer (1992). Parsimony analysis 714 Crawford, Mesfi n Tadesse, Mort, Kimball and Randle recovered three major groups, which were treated as genera Bidens and Coreopsis. "eservations have continually subtribes Coreopsidinae, Petrobiinae, and their new been expressed for decades not only about distinguish- Chrysanthellinae. The monospeci! c genera Dicranocarpus ing the two genera from each other, but also regarding and Goldmanella were eventually excluded from their the monophyly of each genus (Wild 1967; Agnew 1974; analyses because they occurred in various positions on Mes! n Tadesse 1984b, 1986, 1993). One of the shortcom- the cladograms, and inclusion of the genera produced ings of the phylogenetic analysis of "yding and Bremer so many shortest trees that computer memory was ex- (1992), which they readily acknowledged, is that both hausted. Of special interest, the largest genus Bidens occu- Bidens and Coreopsis were each accepted as “good” genera pied various positions in the shortest trees, including oc- in their analyses even though neither is likely monophyl- currence in the two di# erent subtribes Coreopsidinae and etic. It was necessary for "yding and Bremer (1992) to ac- Petrobiinae. Karis and "yding (1994) essentially followed cept both genera because it was beyond the scope of their the treatment of "yding and Bremer (1992) but recog- study to examine these large complex genera in depth. It nized the entire group as subtribe Coreopsidinae with the is evident that until there is better resolution of relation- three subtribes of "yding and Bremer (1992) treated as ships within and among elements of Bidens and Coreopsis, informal groups. as well as clari! cation of their relationships to other gen- era, it will not be possible to reach a proper understanding Tribes/subtribes based on molecular data of phylogenetic relationships within Coreopsideae. Jansen et al. (1991) were probably the ! rst to apply com- Kim et al. (1999) used ITS sequences to provide the parative cpDNA data to the evaluation of phylogenetic ! rst molecular phylogenetic study of Bidens and Coreopsis. relationships within Asteraceae as a whole. Coreopsideae The two shortcomings of the study were limited taxo- were represented by Coreopsis and Dahlia, and cpDNA nomic sampling in Bidens and inclusion of only repre- data supported the segregation of Coreopsideae from the sentatives of the two genera as the ingroup. Despite the core of Heliantheae. Based on complete sequences of the shortcomings, the results of Kim et al. (1999) indicated rbcL gene for 25 species of Asteraceae, Kim et al. (1992) strongly that neither Bidens nor Coreopsis was monophyl- did not ! nd strong support for relationships between the etic. Ganders et al. (2000) used ITS sequences to examine tribes but stated that Tageteae, Coreopsideae, Heliantheae relationships in Bidens, with emphasis on ascertaining the and Eupatorieae are close, and maintained Coreopsideae continental relatives of Hawaiian and Marquesan mem- as a tribe. Using chloroplast ndhF sequences from Dahlia, bers. Their results produced groups of Bidens similar to Coreopsis and Cosmos of Coreopsidinae, Kim and Jansen those detected by Kim et al. (1999), but since no other (1995) showed that this group is embedded within a clade of genera were included in the ingroup, the monophyly of Heliantheae s.l., (i.e., including Helenieae, Coreopsideae, Bidens was not tested. Eupatorieae, and Tageteae). Bayer and Starr (1998), using Kimball and Crawford (2004) conducted a molecular two non-coding chloroplast sequences (trnL intron, and phylogenetic study of Coreopsideae using ITS sequences trnL/trnF intergenic spacer), showed the same relationship from 20 of 24 genera (Table 42.1). Taxon sampling in between Tageteae, Heliantheae and Eupatorieae as Kim Bidens and Coreopsis included representatives of clades re- and Jansen (1995). Heliantheae were represented only by covered by Kim et al. (1999) and Ganders et al. (2000). Helianthus in their analysis. Panero and Funk (2002) used The tree presented in Fig. 42.1 was constructed with a combined dataset of chloroplast sequences totaling over maximum likelihood (ML) analyses of ITS sequences and 13,000 bp to produce a phylogeny-based subfamilial clas- includes exemplar taxa for clades present in the analysis of si! cation for Asteraceae with Coreopsideae treated as a Kimball and Crawford (2004). While prior studies used tribe. In the supertree (= metatree) of Funk et al. (2005), maximum parsimony for tree construction, Mort et al. Coreopsideae were retained at the tribal level. (2008) and Mort et al. (unpub.) show high congruence between maximum likelihood and maximum parsimony Genera based on molecular data analyses, and only the likelihood tree is shown. All ITS The following observations regarding resolution of rela- sequences are available in GenBank and have been previ- tionships within Coreopsideae can be generalized from ously published. Plastid sequences are available for some the morphological studies conducted through the early taxa (Mort et al., 2008) but will be mentioned only when 1990s. All studies recognized as monophyletic those they provide additional insights into relationships or re- taxa with C4 photosynthesis (the Chrysanthellum group). sults incongruent with the ITS tree. The only monophyl- Goldmanella was placed within Coreopsideae, but was etic group not collapsed in the tree is Bidens-2, 3, and the recognized as a somewhat discordant element in the tribe reason for this will be discussed. (Stuessy 1977a; "obinson 1981). Without doubt, how- The ! rst split in the ingroup is between the small ever, the biggest impediments to understanding rela- South American genus Ericentrodea (Table 42.1; "obinson tionships within Coreopsideae have been the two largest 1993) and the remainder of the sampled taxa. Two of Table 42.1. Currently recognized genera of Coreopsideae, giving number of species sampled for construction of Fig. 42.1 (followed by approximate total number of species sequenced for ITS in parentheses and total number of species in genus, known chromosome numbers, and assessment of monophyly of each genus with reference to most comprehensive study. Genus Taxa Chromosome number (n) Monophyletic; reference 1. Bidens L. 11(25)/340 10, 11, 12, 16, 17, 18, 22, 23, 24, 34, 36, 38, 40, 48, 72, ca. 73 N; Kimball and Crawford 2004 2. Chrysanthellum Rich. 1(1)/13 8, 12 ?; – 3. Coreocarpus Benth. 4(7)/7 9, 11, 12 N; Kimball et al. 2003 4. Coreopsis L. 16(55)/86 6, 7, 8, 9, 10, 12, 13, 14, 24, 26, 28, 32, 39, 56 N; Kimball and Crawford 2004 5. Cosmos Cav. 3(3)/36 11, 12, 17, 22, 23, 24, 33, 36 Y; Kimball and Crawford 2004 6. Cyathomone S.F. Blake 0/1 – N/A; – 7. Dahlia Cav. 3(33)/35 16, 17, 18, 32 Y; Saar et al.; 2003; Kimball and Crawford 2004 8. Dicranocarpus A. Gray 1/1 10 N/A; – Chapter 42:Coreopsideae 9. Diodontium F. Muell. 0/1 – N/A; – 10. Ericentrodea S.F. Blake 2(2)/6 – Y; Kimball and Crawford 2004 11.