| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Botany, University of Texas, Austin, Texas 78713
Received for publication February 27, 1997. Accepted for publication August 14, 1998.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Asteraceae • chloroplast DNA • Ecliptinae • Heliantheae • Neotropical • restriction endonuclease • subtribal classification
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
3500 species classified in 302 genera (Robinson, 1981
; Karis, 1993
; Karis and Ryding, 1994
). The tribe includes many well-known plants such as the commercial sunflower cultivars of several Helianthus species, and many ornamentals including members of Cosmos, Coreopsis, Dahlia, and Tagetes. Its distribution is essentially New World with a few genera extending into the Old World tropical and subtropical regions. The tribe is predominantly North American with a high number of species concentrated in the highlands of Mexico and the southwestern United States, and comparatively smaller numbers in the Andes and southwestern Brazil. The highest generic diversity is in Mexico.
The taxonomic circumscription of the Heliantheae has been modified through time to include all genera with mostly trinerved leaves, yellow corollas, and radially or laterally flattened cypselae. This general taxonomic concept of the tribe, established by Cassini (1821)
, has been slightly modified to emphasize the presence or absence of receptacular bracts or paleae to separate two major groups within the Heliantheae sensu lato (Bentham, 1873
; Hoffmann, 1890
). The Helenieae contain all epaleaceous genera with mostly scaly pappi and tan or rosaceous anthers (Karis and Ryding, 1994
). The Heliantheae sensu stricto are paleaceous and most have dark-colored or black anthers. At the first major symposium on the phylogeny and chemistry of the Asteraceae in Reading, England, Turner and Powell (1977a)
treated tribe Helenieae as an artificial taxon and suggested its genera should be placed elsewhere. In the same symposium, Stuessy (1977)
provided a novel classification scheme, which included within his Heliantheae some helenioid taxa such as the Gaillardiinae and Bahiinae. Robinson (1981)
provided the most lucid and comprehensive account to date of the taxonomy and generic limits of tribe Heliantheae. He viewed tribe Heliantheae in its broadest sense by including all members of tribes Heliantheae sensu stricto and Helenieae in a classification scheme containing 35 subtribes. His circumscriptions of these subtribes were based mostly on a combination of floral and cypsela microcharacters anchored by other taxonomically distinctive features that have been traditionally used in the delimitation of these groups.
Recent molecular studies have led us to reevaluate Robinson's treatment with the potential recognition of a modified tribe Helenieae. Phylogenetic comparisons of restriction site data from chloroplast DNA (cpDNA) and sequence data from the chloroplast gene ndhF have revealed tribe Eupatorieae as derived from Heliantheae (Jansen and Kim, 1996
; other references therein). Because of limited sampling and/or weak phylogenetic signal the immediate sister taxa of the Eupatorieae within Heliantheae have not been identified. Regardless of such shortcomings, these studies all coincide in demonstrating that Heliantheae as circumscribed by Robinson (1981)
is not monophyletic and that the Eupatorieae is nested roughly between the traditional divisions of the Helenieae and Heliantheae. These findings lead us to look back at the polyphyletic Helenieae and its lineages as candidates for tribal recognition and the recognition of a more exclusive Heliantheae to contain only paleaceous taxa with dark anthers. Recent cladistic studies by Karis (1993
, 1996
) and Karis and Ryding (1994)
incorporating, or having as reference, information from these molecular studies have supported the recognition of tribe Helenieae and called for the reevaluation of the classification of the tribe. Their classification scheme for tribe Heliantheae, highlighted by the authors as an improvement over Robinson's treatment because of the use of rigorous cladistic methodologies, has not convinced us as such, and we have opted in this paper to use Robinson's subtribal circumscriptions for discussion and comparison purposes.
Our goal in this study was to evaluate phylogenetic relationships of the Ecliptinae. This is the largest subtribe recognized by Robinson, including 650 species of montane and lowland Neotropical species and some temperate North American species, e.g., in Silphium, Wyethia, and Balsamorhiza. This study grew out of our effort to elucidate the phylogenetic relationships and sister taxa of Verbesina; results from that study have been reported elsewhere (Panero and Jansen, 1997
). The results reported here reveal that subtribe Ecliptinae sensu Robinson is polyphyletic and that there is ample support for recognition of subtribes Engelmaniinae and Zinniinae and a more exclusive subtribe Verbesininae. These studies indicate that subtribe Ecliptinae should be circumscribed to include only most genera exhibiting a combination of variously winged cypselae, foliaceous phyllaries, and cypselae with some sort of setose or awned pappi exemplified by the genera Lasianthaea, Perymenium, and Wedelia.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
. Purification of DNA was accomplished using cesium chloride/ethidium bromide gradients (Sambrook, Fritsch, and Maniatis, 1989
). Restriction endonuclease digestions, agarose gel electrophoresis, bidirectional transfers of DNA fragments from agarose gels to Zetabind (CUNO Inc., Meriden, Connecticut) nylon filters, labeling of recombinant plasmids by nick-translation, filter hybridizations, and autoradiography were performed as described in Palmer (1986)
, Jansen and Palmer (1987)
, and Palmer et al. (1988)
. Sixteen six-base-pair restriction endonucleases were used, AvaI, BamHI, BanI, BanII, BclI, BglII, BstNI, BstXI, ClaI, DraI, EcoRI, EcoRV, HincII, HindIII, NsiI, and XbaI. Filter hybridizations used 22 cloned SacI fragments of lettuce cpDNA (Jansen and Palmer, 1987
) to construct restriction maps for all enzymes involved. Most maps are computerized in a Macintosh platform and are available from the first author. Mapping was performed as outlined in Jansen et al. (1991)
and Jansen, Michaels, and Palmer (1991)
using cpDNA maps of Helianthus constructed by Jansen et al. (1991)
and K.-J. Kim (unpublished data) as a reference. The data matrix contains 89 180 cells with 14 of them coded as missing. Complete maps are not available for enzymes DraI and EcoRI; sites in the small single-copy region were not coded because of uncertainty in assessing site order and/or homology among the genera studied. Differences in plastome DNA size exemplified by fragment size polymorphisms were not used in the analysis because of the difficulty associated in determining homology.
Phylogenetic analyses of the data employed Wagner parsimony (Farris, 1970
) using a Macintosh Quadra 840AV with 64 MB of RAM and PAUP version 3.1 (Swofford, 1993
). Tree Bisection Reconnection (TBR) and MULPARS options were used to search for the most parsimonious trees. The genera Hymenoxys and Coreopsis were used as outgroups (Jansen and Kim, 1996
). A strict consensus tree was constructed from the maximally parsimonious topologies using the PAUP CONTREE option. The amount of support for monophyletic groups was assessed using 100 bootstrap replicates (Felsenstein, 1985
). The MULPARS option was not used for the bootstrap analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
18 kb was detected in the genus Delilia. This inversion runs approximately from SacI fragments 4.6, 10.6, and 7.7 (clones 11, 12, and 13 of the lettuce chloroplast DNA library). Several of the missing data points in the matrix are associated with this inversion because of the difficulty in establishing homology between restriction sites for Delilia and the rest of the Heliantheae sampled. This situation was most prevalent with sites near to the ends of the inversion. Significant size variation in plastome DNA was observed in all taxa sampled. A loss of 300 bp was recorded for some taxa of the Ecliptinae in the area of fragments 4.3 and 3.4 SacI-HindIII, (clones 11a and 11b of the lettuce chloroplast DNA library). Incrementally smaller losses were recorded for most other genera sampled. This size variation was made evident by the use of restriction endonuclease BanI, which has sites flanking the variable region. Differences in size variation were not used in the phylogenetic analysis because they were not discrete but a gradient of sizes among the taxa sampled.
Phylogenetic analysis
The data matrix contains 910 variable restriction sites among the 96 taxa examined, 468 of which were phylogenetically informative. Wagner parsimony produced 30 517 equally parsimonious trees with a length of 1730 steps, a Consistency Index (CI) of 0.364 (excluding autapomorphies), and a Retention Index (RI) of 0.776. Bootstrap values and character support for each branch are shown on one of the most parsimonious trees (Fig. 1).
|
, is polyphyletic; members of the subtribe are distributed among four different lineages.
|
The next lineage contains the North American genera Balsamorhiza, Borrichia, Chrysogonum, Engelmannia, Silphium, Vigethia, Wyethia, and the Mesoamerican monospecific genus Rojasianthe, all members of the Ecliptinae. The clade is supported by three site changes and a bootstrap value under 50. The position of Rojasianthe as the basal lineage of this clade is not supported in the strict consensus tree. The other seven genera are included in a very strongly supported clade with 12 site changes and a bootstrap value of 99. The northern Mexican genus Vigethia is in a basal position. Silphium is the sister taxon to Borrichia and Engelmannia; this relationship is not supported in the strict consensus tree. The western United States genera Balsamorhiza and Wyethia are nested together with the eastern United States monospecific genus Chrysogonum as their sister taxon; these relationships are supported in the strict consensus tree. The last six genera are grouped in a clade that has strong support with eight site changes and a bootstrap value of 94.
All genera of subtribe Ecliptinae that have conical receptacles, the genera Encelia and Helianthella, and the genus Zaluzania, the only member of the Zaluzaniinae examined, are clustered in a weakly supported clade. Encelia and Helianthella form a monophyletic group and are sister to a clade containing two main branches. Zaluzania is basal to the North American genera having marcescent ray corollas, namely Heliopsis, Philactis, and Sanvitalia. The clade containing the last three taxa is strongly supported with 20 site changes and a bootstrap value of 98. The genera Acmella, Salmea, and Spilanthes, all characterized by flat cypselae with ciliate margins, are nested in a clade supported by ten site changes and a bootstrap value of 97. Members of the Galinsogiinae, Espeletiinae, Melampodiinae, and Milleriinae are sister to these genera (Figs. 1, 2).
Genera with functionally staminate disc flowers are distributed in two different clades. The five genera of subtribe Espeletiinae are grouped together (Figs. 1, 2). The monophyly of subtribe Espeletiinae is strongly supported with five site changes and a bootstrap value of 98. Rumfordia and Smallanthus form a monophyletic group and together they are sister to the Espeletiinae, and this clade is supported by three site changes but collapses in the strict consensus tree. Subtribe Galinsoginae is collateral to these taxa; the subtribe as circumscribed by Robinson (1981)
is monophyletic. In the strict consensus tree, however, the node supporting the two major lineages within the Galinsoginae collapses. Melampodium, the other taxon with functionally staminate disc flowers, is sister to Montanoa; collectively they are the sister taxa to the genus Guardiola, the only member of subtribe Guardiolinae. The lineage containing these three genera is in a position basal to the clade containing the fourth group of mostly Mesoamerican and South American genera of the Ecliptinae plus the genus Clibadium of the Clibadiinae. In the strict consensus tree this position is not supported.
The next clade is supported by ten site changes and a bootstrap value of 97 and is characterized by the sequential splitting of taxa (Fig. 1). This clade contains the genus Eclipta and most lowland, Mesoamerican genera of subtribe Ecliptinae. Within this clade, two interior nodes are strongly supported by a large number of site changes and bootstrap values of 99 and 100, respectively. The basal lineage contains the South American genera Monactis, Idiopappus, and Kingianthus.
The next lineage to split contains all genera having cypselae with a pappus of two awns produced as a continuation of the edges of the cypselae plus the morphologically distinct genus Delilia. This clade has good support with eight site changes and a bootstrap value of 87. The genera Otopappus and Oblivia are grouped together in a strongly supported clade with 13 site changes and a bootstrap value of 100. The genus Otopappus is not monophyletic. This clade is sister to an equally well-supported clade containing Lasianthaea, Synedrella, Damnxanthodium, and Delilia.
A strong clade, supported by 17 site changes and a bootstrap value of 100, contains most genera of Ecliptinae with winged cypselae and pappi of multiple caducous or persistent awns. The genus Eclipta is basal to this clade. The pantropical weedy genus Sphagneticola is sister to the Galapagos Island endemic, monospecific genus Macraea; this clade is supported by 13 site changes and a bootstrap value of 100. The genus Lundellianthus is monophyletic and sister to Perymenium; this relationship is not supported in the strict consensus tree. The genus Clibadium of subtribe Clibadiinae is nested within this clade. The terminal clade supported by 17 site changes and a bootstrap value of 100 includes the genera Oyedaea, Steiractinia, Wedelia, Elaphandra, and Zexmenia. The climbing, shrubby genus Zexmenia is basal to this group of taxa. The genera Oyedaea, Steiractinia, and Wedelia are paraphyletic.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
included in subtribe Ecliptinae mostly herbaceous genera with opposite leaves and mostly fertile ray flowers. With the exception of Enhydra, his concept of the subtribe coincides with the terminal clade of our tree, which also includes the type genus, Eclipta (Figs. 1, 2). The other genera sister to Eclipta in our studies were placed by Stuessy, because of their sterile ray flowers, in his subtribe Verbesininae. Robinson (1981)
enlarged the Ecliptinae to include most genera in Stuessy's subtribes Engelmaniinae, Ecliptinae, Verbesininae, and Zinniinae. Robinson's expanded Ecliptinae was not merely a fusion of Stuessy's subtribal concepts but the result of careful comparisons and analyses of novel floral microcharacters. Differences in expressions of these characters allowed Robinson to remove two genera that were regarded by him as anomalous in the Ecliptinae (i.e., Enhydra, Garcilassa). Cladistic studies of tribe Heliantheae by Karis (1993
, 1996
) and Karis and Ryding (1994)
produced different phylogenetic hypotheses concerning the generic relationships within the tribe. In their classification scheme they used the name Verbesininae rather than Ecliptinae because they considered Eclipta to be unassignable to any of their subtribes.
Results from our studies show that the Ecliptinae as circumscribed by Robinson (1981)
is polyphyletic (Fig. 2). This result in itself is not surprising given that the subtribe contains the most heterogeneous group of genera in the Heliantheae. According to him the unifying theme among these genera is a series of morphological features that in combination are distinctive in tribe Heliantheae. Most members of the Ecliptinae have black, nonstriate cypselae, weak development of colored resin patterns in the ducts of the corollas, and strongly blackened anthers. These morphological features, which are collectively treated as tendencies, because some genera variously lack some of them, are strongly unified as a taxonomic delimiter by the presence in most Ecliptinae of fibers along the veins of the disc corollas, a unique feature in Heliantheae. Even though these characteristics are useful in delimiting phenetic groups, their utility in revealing natural lineages is disputed if compared to DNA studies (Schilling and Panero, 1996
; Panero and Jansen, 1997
). A more detailed discussion of each of the major lineages recognized in this study as depicted in Fig. 1 is provided below, following a phylogenetic order from most basal to most recently derived.
The genus Enhydra
The subtribe Enhydrinae was established by Robinson (1981)
to contain the aquatic, pantropical genus Enhydra. Robinson (1981)
concluded that Enhydra has a unique combination of morphological characteristics present in several groups, remotely related among each other, such as the Neurolaeninae, Ecliptinae, and Montanoa. In our study, Enhydra is a basal lineage to the rest of the Heliantheae sensu stricto. In some trees, however, Enhydra and Montanoa are sister. This situation is the result of sharing some restriction sites unique to each other. Enhydra shares with Montanoa white limbs of ray corollas, a similar cypsela morphology, opposite phyllotaxy, and a base chromosome number x = 19. Enhydra appears to be an isolated element within the Heliantheae.
Neurolaeninae clade
Robinson (1981)
circumscribed the subtribe Neurolaeninae to include genera with prismatic cypselae, radially arranged pappus elements, reddish resin in ducts along veins of disc corollas, and unstriated cypselae, among other characteristics. We sampled four members of the subtribe: Calea, Greenmaniella, Neurolaena, and Tetrachyron. The Neurolaeninae as circumscribed by Robinson is supported by our results except for inclusion of Tetrachyron in the subtribe. The Neurolaeninae are not sister to the Galinsoginae, where, historically, many of its genera and species were placed.
Podachaenium, Squamopappus, Verbesina, and Tetrachyron clade
Robinson (1981)
placed three of these genera in subtribe Ecliptinae and Tetrachyron in subtribe Neurolaeninae. He believed Tetrachyron to be somewhat anomalous in the Neurolaeninae because of different floral and cypsela microcharacters. Wussow and Urbatsch (1979)
reevaluated Tetrachyron to include five species. They considered Tetrachyron to be closely related to Calea and excluded from Tetrachyron three other taxa now referred as podachaeniums. Podachaenium was known from one species, the ruderal P. eminens, until Robinson (1978d)
placed C. skutchii in Podachaenium. This transfer was not accepted by Jansen, Harriman, and Urbatsch (1982)
who described a new genus, Squamopappus, to accommodate C. skutchii. In the same paper they reevaluate Podachaenium to include two species. The most obvious difference between Squamopappus and Podachaenium is in ray flower corolla color, yellow in Squamopappus and white in Podachaenium. Except for this character, Squamopappus is essentially identical in most morphological characteristics to the lanceolate-leaved species of Podachaenium. None of these studies suggested a close relationship of these genera to Verbesina.
Traditionally, Verbesina has been variously allied to genera having laterally flattened, pappose, winged cypselae. Robinson and Greenman (1899)
, basing their conclusions on cypsela characteristics, believed Verbesina to be closely related to Otopappus. Stuessy (1977)
allied Verbesina to Monactis Kunth, Podachaenium Benth., and Podanthus Lag. among others, because of their similar phyllotaxy, disc morphology, and pistillate ray flowers. Robinson (1981)
, basing his conclusions mostly on floral microcharacters, suggested a close relationship of Verbesina simultaneously to Acmella Rich. ex Pers. and Salmea DC. on the one hand and to Encelia Adans., Enceliopsis A. Nelson, Flourensia DC., and Geraea Torr. and A. Gray on the other. Strother (1991)
, basing his conclusions on cypsela and corolla morphology, included Verbesina in his informal Group II (with Jefea Strother, Lundellianthus H. Rob., Oblivia Strother, Otopappus, Perymeniopsis H. Rob., Synedrella Gaertn., and Tuxtla Villaseñor and Strother
). The latest classification of the Heliantheae, by Karis and Ryding (1994)
based on a cladistic analysis, allied Verbesina to Flourensia, Encelia, Espeletia Mutis ex Humb. and Bonpl., and Zaluzania Pers. This diverse array of opinions concerning the phylogenetic relationships of Verbesina is probably the result of the tremendous morphological variation exhibited by the 300 species of the genus. For a more detailed discussion of the relationships of these genera and the infrageneric taxonomy of Verbesina the reader is directed to Panero and Jansen (1997)
.
Our results support the removal of Tetrachyron, Squamopappus, Podachaenium, and Verbesina from subtribe Ecliptinae as subtribe Verbesininae containing only these four genera.
Temperate North American Ecliptinae plus Rojasianthe clade
The Engelmanniinae were circumscribed by Stuessy (1977)
to include all genera with mostly alternate, dissected leaves, mostly functionally staminate disc flowers, and cypselae complexes (basally fused complex of one phyllary, one fertile pistillate ray flower, two functionally staminate disc flowers, and two to four pales). Robinson (1981)
included the Engelmanninae within his concept of the Ecliptinae. Once transferred to the Ecliptinae, he thought this group of North American genera to be related to tropical Ecliptinae because they all possessed a similar fiber sheath pattern along the veins of the disc corollas. Our results show that Engelmannia, Silphium, and Chrysogonum are derived from a common ancestor and that other Heliantheae genera of temperate North American distribution should be included in a revised concept of subtribe Engelmanninae. The genera Balsamorhiza, Borrichia, Vigethia, and Wyethia in spite of having hermaphrodite disc flowers and lacking the cypselae complex share their evolutionary history with that of members of subtribe Engelmanniinae.
The most surprising result is the phylogenetic position of the Mesoamerican monospecific genus Rojasianthe. Rojasianthe has been traditionally allied to Montanoa because of its accrescent pales, opposite leaves, and neuter ray flowers with white corollas. It also shares with Montanoa the same base chromosome number of x = 19. As discussed below under Montanoineae, Rojasianthe is quite different from Montanoa in many key morphological features and as our results show Rojasianthe is not closely related to Montanoa. Rojasianthe is apparently closely related to the mostly North American genera of Engelmanniinae. This relationship has never been suggested in the literature; the only obvious gross morphological similarity between Rojasianthe and these genera is the broadly ampliated phyllaries and the same base chromosome numbers (as in Vigethia, Balsamorhiza, and Wyethia). Given this result, we hypothesize a Mesoamerican origin for the lineage with the monospecific genera Rojasianthe, Vigethia, and Dugesia as palaeoendemic relicts isolated in the mountains of southern and eastern Mexico and successful radiation of the lineage in the temperate regions of the eastern and western United States. Similar conclusions can be drawn from phylogenetic studies based on sequence data of the Internal Transcribed Spacer region (ITS) of the nuclear ribosomal DNA (nrDNA), which support the basal position of Rojasianthe and a clade composition similar to the one revealed here by study of cpDNA (Clevinger and Panero, 1997
).
Conical receptacle clade
Most genera of Ecliptinae with conical receptacles are closely related. With the exception of Zaluzania, all these genera were classified by Robinson (1981)
in his subtribe Ecliptinae. Zaluzania is basal to a lineage containing Heliopsis, Philactis, and Sanvitalia. This result lends strong support for the recognition of subtribe Zinniinae as defined by Stuessy (1977)
. We agree with Robinson that the genus Borrichia should be removed from this alliance. Borrichia lacks marcescent ray corollas, a distinctive feature of all Zinniinae. As mentioned above, Borrichia is a member of the lineage containing members of the Engelmanniinae.
The genera Acmella, Salmea, and Spilanthes are sister to the Zinniinae and like that group they have strongly conical receptacles. The three genera share laterally flattened cypselae and somewhat navicular pales. Their corollas are white or yellow and, like the Zinniinae, all taxa have opposite leaves. Sampling of genera identified in the literature as related is necessary to understand phylogenetic relationships in the group and to recognize it formally as a distinct subtribe.
Basal to the above two clades is the clade containing Encelia and Helianthella. These two genera have alternate leaves and laterally flattened cypselae with conspicuously ciliate margins. In several species, the cypsela shoulders are very well developed, producing a distinctive morphology, which is also observed in the closely related genus Flourensia. These genera have been identified in previous molecular studies as sister to subtribe Helianthinae (Schilling and Jansen, 1989
; Kim, Turner, and Jansen, 1990
). Our results support the recognition of a new subtribe to either include these genera or to include them in a more heterogeneous subtribe Helianthinae.
Galinsoginae
Subtribe Galinsoginae as circumscribed by Robinson (1981)
included 16 genera of mostly herbaceous species characterized by having extensive development of resin ducts in the corollas, usually white, trilobed ray corollas, prismatic cypselae, and base chromosome numbers of x = 8 and 9. The subtribe contains 125 species distributed primarily in the highlands of Mexico and Central America, though some species are found in the Andes of South America.
Elucidation of phylogenetic relationships among genera of the Galinsoginae has been problematical because of convergence and reduction of key characters often used in Asteraceae classification. For example, Alepidocline and Schistocarpha, because of their pappus of capillary bristles and small, rudimentary pales (if any), were traditionally placed in tribe Senecioneae. In other cases, circumscription of some genera has been difficult because of reduction in morphological characteristics useful in classification. These situations led, inevitably, to a very tangled taxonomy and the formulation of artificial classification schemes. Stuessy (1977)
circumscribed the Galinsoginae to contain a very heterogeneous assemblage of genera that today we consider to have closer phylogenetic affinities with subtribes Ecliptinae and Neurolaeninae. Our studies support the Galinsoginae sensu Robinson (1981)
, but the support is weak because the clade collapses in the strict consensus tree. The two main Galinsoginae clades coincide with chromosome number information and pappus characteristics that have been traditionally used in the recognition of two main lineages in the Galinsoginae.
The x = 9 clade contains Bebbia, Cymophora, Jaegeria, Tetragonotheca, and Tridax. The mostly Mexican genus Jaegeria is basal to this lineage, but the position is weak as exemplified by the low bootstrap value. Jaegeria is quite heterogeneous: some species are aquatic herbs with white to purple ligules and the other species (centered about J. hirta) are mostly ruderal with small, yellow ray corollas. Most species of Jaegeria can be recognized by the translucent folds of the innermost phyllaries that tightly wrap the cypselae (Torres, 1968
). The southern United States genus Tetragonotheca with its pappus of minute squamellae is basal to the clade containing genera with plumose pappus (Bebbia, Cymophora, and Tridax). This result agrees with traditional classificatory assumptions based on morphological and chromosomal data. Most species of Cymophora have been transferred in and out of Tridax and much misunderstanding has surrounded the position of the Venezuelan species C. venezuelensis because of the unusual morphology of its outermost flowers (Canne, 1977
; Turner and Powell, 1977b
; Keil, Luckow, and Pinkava, 1987
). Our study clearly shows that C. venezuelensis is a distinctive taxon and not congeneric with Tridax. Furthermore, molecular studies of the Galinsoginae based on sequence data of the Internal Transcribed Spacer region (ITS) (J. Panero and A. Plovanich, unpublished data) including samples of the western Mexican species C. accedens, C. pringlei, and C. venezuelensis, confirm that Cymophora is monophyletic and sister to species of Tridax with subequal phyllaries (i.e., T. coronopifolia and allies). Bebbia is sister to Cymophora and Tridax.
The monophyly of the x = 8 group is very strongly supported. Two lineages were revealed by the study. The first clade contains Alepidocline, Alloispermum, Aphanactis, Oteiza, and Schistocarpha. All these genera share some sort of capillary bristle pappus that can either be caducous or persistent. The mostly Andean genus Aphanactis is basal to the clade. Alepidocline and Oteiza split sequentially. These two genera share many morphological features, but the most important and diagnostic characteristic is their pappus of easily caducous unequal bristles that contrast with those of Alloispermum and Schistocarpha in which the bristles are more or less persistent and roughly equal in length. Oteiza differs from Alepidocline in its paleaceous receptacles and suffruticose habit. Our results show that Alloispermum is paraphyletic and that Schistocarpha is apparently derived from it. Alloispermum was recently resurrected by Robinson (1978b)
to include species with pappose ray and disc cypselae. These studies and sequence data from the ITS region (J. Panero and A. Plovanich, unpublished data) support this decision but also clearly show need for a refinement of the generic concept of Alloispermum by a thorough study of the morphology and generic limits of Galinsoga, Sabazia, and Alloispermum and the role that hybridization may have played in the speciation and evolution of these genera. Sequence data of the ITS region do not support inclusion of Schistocarpha as sister or derived from Alloispermum but rather as sister to Alepidocline and Oteiza.
The South American species of Alloispermum sampled are placed in different clades of the Galinsoga–Sabazia lineage and are not related to the Mexican species of the genus. The involucre of the South American species is herbaceous, not scarious as in Mexican Alloispermum. Given the inherent difficulties in circumscribing genera in the Galinsoginae, the possible role of hybridization and polyploidy (both species sampled are polyploids; see Strother and Panero, 1994
) in the speciation of the South American Galinsoginae, and character convergence, it is not surprising that the species were allied to generic concepts with which they share no common evolutionary history. A simple transfer of these species to Sabazia or Galinsoga may be, at first impression, a solution to the problem, but the related genera are also poorly understood and consequently in need of a thorough study of their evolutionary history. This result is also supported by sequence data of the ITS region. It is clear from these studies that hybridization has played a major role in the speciation of Galinsoga and other genera of the subtribe and that this hybridization has contributed to the poor understanding of species and generic limits in the subtribe.
Espeletiinae
Subtribe Espeletiinae was established by Cuatrecasas (1976)
to include Espeletia and six other genera, all characterized by having the innermost disc flowers functionally staminate, base chromosome number of x = 19, phyllaries partly enclosing ray cypselae, ray flowers in several series, and vaginate to tubular leaf bases. In addition, all genera are characterized by the aggregation of leaves distally on growing shoots. In the genera adapted to Andean forest and shrubbery the leaves are soon deciduous and only the tips of growing shoots bear living leaves. Carramboa, Libanothamnus, and Tamania have this life form and tend to occupy Andean forest/shrubbery. At increasingly higher elevations, the aggregation of leaves is more marked, and a rosette is observed with the dead leaves covering the growing stem or trunk. Cuatrecasas (1976)
termed this growth habit as caulirosula. Espeletia, Coespeletia, Espeletiopsis, and Ruilopezia are all characterized by forming caulirosula.
We sampled species from five genera (Coespeletia, Carramboa, Espeletia, Libanothamnus, and Ruilopezia). The cpDNA tree confirms the monophyly of the Espeletiinae and supports its traditional placement along with other genera with functionally staminate disc flowers such as Rumfordia and Smallanthus. More recently there have been some discussions as to the close relationship of the Espeletiinae and some members of the Ecliptinae such as Wedelia and related genera (Karis, 1993
; Karis and Ryding, 1994
). This relationship is not supported by the cpDNA phylogeny. Furthermore, our results reveal a paraphyletic genus Espeletia. Espeletia pycnophylla from Ecuador is basal to all Espeletiinae examined. This result, if confirmed, supports an early split in the Espeletiinae lineage leading to Páramo-adapted, rosette-forming forms such as E. pycnophylla in Ecuador separate from those of the northern Andes of Venezuela and possibly northern Colombia.
Our results support a hypothesis of gradual diversification of the Espeletiinae lineage through adaptation and diversification to progressively higher elevations in the Andes. Arborescent genera such as Carramboa, Libanothamnus, and, to some extent, Ruilopezia, endemic to the Andean shrubbery vegetation belt just below the Páramo habitat, are basal to the lineages containing the rosette-forming genera such as Espeletia and Coespeletia, which are endemic to Páramo areas. Geological data indicate that Páramo regions in the Andes are of recent origin because Andean orogenesis has only recently produced mountains of sufficient elevation capable of harboring Páramo habitat (Simpson and Todzia, 1990
, and other references therein).
Smallanthus–Milleriinae
Smallanthus and Rumfordia were placed by Robinson (1981)
in his Melampodiinae and Milleriinae, respectively. Our studies show that the two genera are closely related and are not particularly close to Melampodium. This result supports a reevaluation of the generic relationships of these two subtribes. Rumfordia differs from Smallanthus in having bisexual disc flowers and strongly trifid ray corolla limb apices.
Smallanthus, because of its functionally staminate disc flowers and globose cypselae, was mentioned by Robinson (1981)
as possibly sister to Espeletiinae. This relationship is supported by our studies. Robinson highlighted that Smallanthus differs from Espeletiinae in having opposite leaves and different cypsela wall characteristics. Field observations of Andean Smallanthus species show additional, albeit tenuous similarities between the genus and the shrubby Espeletiinae (i.e., Carramboa, Libanothamnus). Panero observed in Venezuela that in Carramboa and to some extent in Libanothamnus the heads nod after anthesis. This unusual characteristic was also observed in the Andean arborescent species Smallanthus microcephalus. In addition, Smallanthus (at least South American smallanthuses) and the Espeletiinae show a tendency to have resinous exudates that have a mild camphor-like fragrance.
Sequence data from the ITS region (J. Panero and A. Plovanich, unpublished data) support the close relationship of Smallanthus and Rumfordia and collectively to other Milleriinae such as Axiniphyllum, Milleria, Sigesbeckia, and Trigonospermum. None of these genera appear to be closely related to Melampodium.
Guardiolinae
The first revision of Guardiola was by B. L. Robinson (1899)
, who recognized nine species and thought the genus to be related to Melampodium. Robinson (1978e
, 1981
) in his Guardiolinae stressed the unique set of morphological features characterizing Guardiola, its only member. He believed Guardiola to be related to Melampodium with which it shares similar cypsela characteristics. In our study, Guardiola is sister to the clade containing Montanoa and Melampodium. This relationship is not supported in the strict consensus tree.
Montanoinae
The genus Montanoa contains 25 species distributed mostly in Mexico and Central America with five species in northern South America (Funk, 1982
). Robinson (1981)
when erecting subtribe Montanoinae failed to identify a sister taxon for this attractive genus. As explained by Robinson, the combination of epappose cypselae and neuter ray flowers with white corollas of Montanoa is observed elsewhere only in subtribes Helianthinae and Ecliptinae. At the same time, Robinson observed that the cypselae of Montanoa were similar to those found in his subtribes Milleriinae and Melampodinae.
Funk (1982)
also failed to identify an outgroup for the genus. Funk reviewed the genera sharing with Montanoa either a base chromosome number of x = 19 or accrescent pales. She concluded that none of the genera identified by her as having these characteristics nor the ones proposed in the literature as related to Montanoa are closely related to Montanoa.
The accrescent pales of Montanoa are somewhat similar to those of Rojasianthe, the only other genus in the tribe having pales that grow significantly after anthesis. The first impression one obtains from observing Rojasianthe in the field is one of a very robust Montanoa, the only obvious difference between them being its open capitulescences, dark disc corollas, several series of suborbicular, imbricate phyllaries, and nodding heads, a set of characteristics essentially absent from Montanoa. It shares with Montanoa the same chromosome number, neuter ray flowers with white corollas, and opposite phyllotaxy. Rojasianthe differs from Montanoa in some notable features, the most striking being characteristics of the cypselae and disc flowers. The disc cypselae of Rojasianthe are red when immature and have a pappus of several yellow bristles, characteristics not observed in Montanoa. The disc corolla is strongly urceolate and bicolored, with the tube and base of throat dull white and the distal half and lobes black. This corolla morphology is not seen in Montanoa and the black or slate gray coloration of the corollas is only seen in M. pteropoda.
The most recent classification scheme for tribe Heliantheae (Karis, 1993
; Karis and Ryding, 1994
) also failed to identify a sister genus for Montanoa. They placed the genus in their Heliantheae-unassigned-to-a-subtribe group basal to a clade containing the Rudbeckiinae.
Our studies support a relationship of Montanoa with the Melampodiinae. The clade containing these two genera has a weak bootstrap value but does not collapse in the strict consensus tree. It is clear that additional studies including all genera of the Heliantheae sensu stricto should be undertaken to identify satisfactorily the phylogenetic relationships of Montanoa. Montanoa should be maintained in subtribe Montanoinae, exemplifying its isolated position in the Heliantheae. Even though we feel the question as to what is related to Montanoa has still not been satisfactorily answered, this study at least has given us a direction for future research.
Clibadiinae
Subtribe Clibadiinae was circumscribed by Robinson (1978e
, 1981
) to include the genera Clibadium, Lantanopsis, and Reincourtia. He considered the subtribe to be characterized by its strongly cymose inflorescences, the penicillate tufts of hairs on the tips of the disc corolla lobes, and the dense pubescence of the cypselae of most species. He treated the Clibadiinae as intermediate between his Ecliptinae and Ambrosiinae. The subtribe Clibadiinae shares with subtribe Ecliptinae black anthers and nonstriate cypsela walls. It shares with the Ambrosiinae epappose cypselae and reduced corolla limbs.
In our studies, Clibadium is revealed in the center of a clade containing the winged-cypselae members of subtribe Ecliptinae centered about Wedelia (Ecliptinae sensu stricto, see below). This result supports the inclusion of Clibadium in subtribe Ecliptinae and not with members of the Ambrosiinae.
Ecliptinae
As discussed above, Robinson (1981)
circumscribed a polyphyletic Ecliptinae to include most genera having dark anthers and fiber sheaths along the veins of the disc corollas. Our results support the recognition of a core subtribe Ecliptinae from here on referred as Ecliptinae sensu stricto to include the genera Clibadium, Damnxanthodium, Delilia, Eclipta, Elaphandra, Idiopappus, Kingianthus, Lasianthaea, Lipochaeta, Lundellianthus, Macraea, Melanthera, Monactis, Oblivia, Otopappus, Oyedaea, Perymeniopsis, Perymenium, Rensonia, Sphagneticola, Steiractinia, Synedrella, Wamalchitamia, Wedelia, Wollastonia, Tilesia, and Zexmenia. Within this clade we have identified some strongly supported major groups.
The most important result concerning the relationships of the Ecliptinae sensu stricto has been the placement of the clade containing Otopappus and Lasianthaea as a distinctive lineage separate from that containing the winged-cypselae genera with pappi of scarious crowns with or without bristles as exemplified by Perymenium, Wedelia, etc., and from the Verbesina lineage. We believe that it is only by the use of molecular data in combination with careful morphological studies that a classification that will depict the true evolutionary history of the groups involved can be achieved.
The Andean genera Kingianthus, Idiopappus, and Monactis are included in the basal clade of the Ecliptinae sensu stricto. This clade is supported by eight site changes and a bootstrap value of 100. These three genera are endemic to the central Andes of Ecuador and Peru. They are characterized by their prismatic cypselae and alternate phyllotaxy (opposite in monospecific genus Idiopappus), a distinctive characteristic in Ecliptinae sensu stricto.
All genera with laterally flattened cypselae and with pappi of awns or scales produced as a continuation of the cypsela wall wing are grouped in a well-supported clade. The recently described genus Oblivia, previously known as Otopappus mikanioides, the only South American member of the genus, is nested within Otopappus. This result supports the placement of the genus under synonymy in Otopappus. However, this may also be indicative of a heterogeneous Otopappus and consequently of the possibility of two distinct lineages otherwise classified together by convergence in cypselae characteristics. Villaseñor and Strother (1989)
after a thorough examination of the morphology of the genus have come to the conclusion that Otopappus is not monophyletic.
The morphologically distinctive genus Delilia is basal to the clade containing Damnxanthodium, Lasianthaea, and Synedrella. The phylogenetic relationships of Delilia have always been enigmatic because of its highly reduced inflorescence. Stuessy (1977)
placed the genus in his Milleriinae; Robinson believed the genus to be closely related to other herbaceous, ruderal Ecliptinae. The functionally staminate actinomorphic disc flower, fertile ray ligulate flower, and the broad herbaceous phyllaries of Delilia are characteristics that the genus shares with Stuessy's Engelmanninae. This suggestion was supported informally by Strother (1991)
in his study of the wedelioid taxa of North America. Karis (1993)
could not assign the genus with confidence to any of his subtribes but stressed the potentially close relationship of the genus to members of the Caribbean subtribe Pinillosiinae. Our results clearly reveal the genus as a member of the Ecliptinae sensu stricto and closely related to Damnxanthodium, Synedrella, and Lasianthaea.
Damnxanthodium is basal to the Lasianthaea and Synedrella clade. As discussed by Strother (1987)
, Damnxanthodium is anomalous in Lasianthaea because it lacks a pappus, its cypselae body is slightly quadratic, and it does not have winged cypselae. Its chromosome number of n = 12 contrasts with that of its initial placement in Perymenium (x = 15) and from its most recent position in Lasianthaea (x = 10, 11). Our results and the morphological and chromosomal differences outlined above support the distinctiveness of Damnxanthodium and are in disagreement with the suggestion of Karis and Ryding (1994)
that the genus is simply an apomorphic segregate of Lasianthaea. Instead, our studies suggest that the ruderal, tropical lowland genus Synedrella is sister to the mostly Mexican genus Lasianthaea. Synedrella, like Calyptocarpus, has sessile heads with yellow ligules and strongly winged cypselae with well-developed pappi of two awns. Synedrella shares with Lasianthaea a slightly similar cypselae morphology. Lasianthaea differs from Synedrella in its shrubby habit and tendency for scarious phyllaries. Future studies in this group are needed to elucidate the phylogenetic relationships of these genera. It is plausible that their closest relatives are the South American genera Dimerostemma and Angelphytum, which like Lasianthaea and relatives, have similar cypselae and involucre characteristics. Amphitropical disjunctions of Heliantheae genera are not unusual (e.g., Flourensia, Viguiera, Stachycephalum, Synedrellopsis, etc.) and may be another source of confusion in assessing phylogenetic relationships.
The genus Eclipta is basal to a strongly supported clade containing most genera with variously winged cypselae and pappi of combinations of persistent or caducous bristles arising from the bases or apices of shallowly to conspicuously constricted crowns. For ease of discussion, from here on this clade will be referred to as the wedelioid group and their distinctive characteristics as wedelioid. Most of these genera are dwellers of the Neotropical moist and rain forests. Strother (1991)
revised the North American members of this group. Although his classification of the genera and species of this group improved on the past, the wedelioid complex continues to be the least understood and most problematic in the Heliantheae sensu stricto. Chromosome numbers are of limited value in classification because they vary from x = 11 to x = 17.
The genera Macraea, Perymeniopsis, Rensonia, Sphagneticola, Wamalchitamia, and Tilesia split sequentially with Eclipta as basal. The clades containing these genera are weakly supported; they either collapse in the strict consensus tree or have weak support from bootstrap analyses forming a polychotomy at the base of the wedelioid clade (Fig. 1). The relationships supported are nevertheless very important from phylogenetic and taxonomic standpoints. These taxa are found in several areas of Mesoamerica and northern South America. The Galapagos Island endemic genus Macraea is strongly nested in the clade containing the pantropical, mostly coastal genus Sphagneticola. Macraea is a shrub adapted to dry areas of the islands; Sphagneticola includes hygrophyte perennial herbs of coastal or perpetually inundated tropical areas. Chromosomal and morphological studies support a relationship between Macraea and Wedelia sensu lato (Harling, 1962
; Eliasson, 1984
). The most plausible explanation for the origin of Macraea is long-distance dispersal from a tropical America progenitor because most of the Galapagos Island floristic elements trace their origins to that area (Wiggins and Porter, 1971
).
Another exciting result from this study is the sister relationship of Perymeniopsis and Rensonia. Perymeniopsis was considered until recently to be the northernmost member of the genus Oyedaea (Robinson, 1978c
). Like Oyedaea, Perymeniopsis has a pappus that combines a lacerate crown with awns or bristles borne at the base of the crown. Robinson considered this taxon to be anomalous within the mostly northern South American genus Oyedaea and recognized it as distinct from any other ecliptinous genera by segregating it as a monospecific genus, Perymeniopsis. Perymeniopsis is endemic to the northern and central areas of the Sierra Madre Oriental, Mexico; Rensonia is a dweller of the rain forests of northern and southeastern Chiapas and Central America. These two genera are each monospecific and probably palaeoendemics arising through vicariance. The overall morphology of Rensonia is suggestive of Perymeniopsis, especially its capitulescence structure, leaf morphology and phyllotaxy, and pappus morphology. This relationship has never been mentioned in the literature because Rensonia has always been allied to the Melampodiinae because of its functionally staminate disc flowers. Our studies support recognition of the genus Perymeniopsis as distinct from Perymenium and Oyedaea.
Wamalchitamia was recently described by Strother (1991)
to accommodate taxa with orange or yellow corollas, foliaceous phyllaries, and long four- or five-angled prismatic cypselae. The genus contains five species scattered mostly in tropical deciduous forests from Oaxaca and Chiapas, Mexico, through Central America to central Costa Rica. Based on our studies, Wamalchitamia is not closely related to Zexmenia sensu Strother (1991)
nor to Wedelia or any of the other Zexmenia segregates that collectively Strother suggested to be a monophyletic group.
Lundellianthus was established by Robinson (1978a)
with a single species Lundellianthus petenensis H. Robinson (= L. guatemalensis (Donn. Sm.) Strother and emended by Strother (1989)
to include eight species previously classified under Lasianthaea, Otopappus, Oyedaea, and Zexmenia. Robinson (1978a)
characterized the genus by a series of floral microcharacter characteristics including the lack of colored resin in the corollas, fibers present in the corolla forming sheaths, black anthers, and cypsela walls not having striations. Strother (1989)
based his circumscription of the genus on characteristics of the heads and cypselae (namely basally connate pales, disc cypselae that are quadrate in transection) and six-angled young stems. The genus, with the exception of Lundellianthus jaliscensis from southwestern Jalisco, is Mesoamerican in distribution.
Our results support recognition of Lundellianthus sensu Strother (1989)
. The genus is part of the Ecliptinae and is sister to Perymenium, although this position is weak because it is only supported by one site change and has a bootstrap value below 50. A sister relationship between Lundellianthus and Perymenium has never been reported in the literature. Perymenium has an overall morphology that is suggestive of Lundellianthus, especially its straggly habit, membranaceous leaf blades, and clustered-cymose capitulescences, however it differs greatly in many key morphological features of the cypselae. Even though the cypselae of Perymenium are three- to four-angled like those of Lundellianthus, they lack wings. In addition, most species of Perymenium have pappi of small bristles borne from a rostrate crown. The cypsela and pappus morphology exhibited by Perymenium is found elsewhere in Melanthera, a genus of uncertain affinities that, according to our studies, is basal to the Wedelia–Zexmenia clade.
Lipochaeta and Wollastonia are sister in a strongly supported clade. Gardner (1979)
considered Lipochaeta to be closely related to Wedelia and endemic to the Hawaiian islands. The genus Lipochaeta has basic chromosome numbers of x = 15 and x = 26. Some authors speculate that the x = 15 group is allied to Wollastonia biflora (n = 15); the x = 26 group has its origin in a hybridization event between either Wollastonia biflora or a Lipochaeta species and some unspecified wedelioid stock (Rabakonandrianina and Carr, 1981
). In some of the trees generated by our analyses the genus Melanthera is basal to the clade containing Lipochaeta and Wollastonia. The taxonomic and phylogenetic relationships of Melanthera are not clear; a modern revisionary study of Melanthera is needed. Melanthera shares the same base chromosome number of some lipochaetas of x = 15. Similarly, the taxonomic limits of the genus Wollastonia are not clearly defined. According to Strother (1991)
some species of Wedelia endemic to various Pacific islands may be congeneric with Wollastonia.
Zexmenia sensu Strother (1991)
contains two species of straggly shrubs or climbers of the moist or tropical deciduous forests of Mexico and Central America. The genus was circumscribed by Strother to include only the type species and an additional very closely related species. The rest of the species of the genus were placed elsewhere either in other established genera or in newly erected genera. Zexmenia has the overall morphology of Perymeniopsis, Lundellianthus, or a climbing Wedelia. According to Strother the relationships of Zexmenia are uncertain and may be with Oyedaea, Lundellianthus, and Otopappus. Our results reveal Zexmenia as basal to a very strongly supported clade whose genera can be recognized formally as a taxon within subtribe Ecliptinae. The genera Elaphandra, Oyedaea, Steiractinia, and Wedelia are included in this clade. As advanced by Strother (1991)
, Elaphandra may be paraphyletic and in need of a revision to include all of its member species, that can be found among the South American members of Aspilia and Wedelia. Wedelia as presently circumscribed is not monophyletic; the Mexican montane species as exemplified by W. purpurea may be phyletically different from the Andean or Caribbean species. The clade containing Oyedaea and Steiractinia is strongly supported by seven site changes and a bootstrap value of 100. Our results suggest that Steiractinia and Oyedaea are not monophyletic as presently circumscribed and that probably there is a need to combine the species of both genera into a more inclusive concept of Oyedaea.
Taxonomic implications
Homoplasy in the form of convergence and unequal rates of divergence accompanied by reversals of key diagnostic features have, as in many other groups, plagued the classifiers of tribe Heliantheae. As outlined above, many classification schemes have been put forward that have, in most cases, refined our understanding of relationships among genera of tribe Heliantheae. The advent of molecular techniques has allowed us to recognize the role of character convergence in classification and has facilitated implementation of classifications that reflect the evolutionary history of the groups involved. The phylogeny here presented is a plastome phylogeny (Doyle, 1992
, 1996
) and not a species phylogeny. The reader is advised to keep in mind that the taxonomic conclusions suggested below can only be implemented formally as a result of a thorough understanding of the morphology of the groups involved and the use of other independent molecular markers.
Results from our studies clearly show the need to reevaluate the taxonomic limits of subtribe Ecliptinae. Because of the limited sampling of members of the subtribe and other outgroups and our poor understanding of the morphology of the groups involved, the results obtained only allow us to provide recommendations for further study. We recommend changes in circumscriptions of subtribes Engelmanniinae, Verbesiniinae, and Zinniinae. It is clear that the characters that have supported the past circumscriptions of subtribe Engelmanninae are labile and that the subtribe should be viewed as an entity that has the propensity of having functionally staminate disc flowers, foliaceous phyllaries, and a cypselae-paleae complex. The case of subtribe Zinniinae is completely opposed to that of subtribe Engelmanniinae because all genera traditionally included in the Zinniinae show marcescent ray corollas, a characteristic unique in the Heliantheae sensu stricto and that has been used successfully to define this lineage of mostly Mexican taxa. Other Ecliptinae with conical receptacles such as Acmella and relatives need to be included in other subtribes or a new subtribe erected to accommodate them. Incidental information from this and other molecular studies (Schilling and Panero, 1996
) suggest that the lineage containing Encelia, Flourensia, and Helianthella may be sister to subtribe Helianthinae. From the present information it is not clear whether collectively these genera share a common ancestor or split sequentially as basal lineages in the evolution of the Helianthinae. If the latter is the case, a more inclusive subtribe Helianthinae that includes these genera can be circumscribed without troublesome loss of taxonomic cohesiveness. The basal lineages of the Helianthinae are the Sonoran desert and Baja Californian members of Viguiera sect. Bahiopsis (Schilling and Panero, 1996
, and other references therein). These taxa share with Encelia and Flourensia a similar chromosome number, overall gross morphology, and geographical distribution.
One of the most important results of this study is the support it provides for the recognition of a recircumscribed subtribe Verbesininae that includes Podachaenium, Squamopappus, Tetrachyron and Verbesina. By identifying the sister taxa of Verbesina and its phylogenetic position in tribe Helianthinae our task of elucidating the phylogenetic relationships of the Ecliptinae sensu stricto has been facilitated enormously.
Our future studies will concentrate on the implementation of a phylogeny based on sequence data of a chloroplast gene for all genera of Heliantheae sensu lato. We believe this approach will allow us to have a better understanding of the evolution of the group and to clarify the classification of tribe Heliantheae.
|
|
|
FOOTNOTES |
|---|
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Canne, J. M. 1977 A new combination in Cymophora (Compositae: Heliantheae: Galinsoginae). Madroño 24: 190–191.
Cassini, H. 1821 Hélianthées. In G. Cuvier [ed.], Dictionnaire des Sciences Naturelles, 2d ed., vol. 20. 354–385. Le Normant, Paris.
Clevinger, J. A., and J. L. Panero. 1997 Phylogenetic relationships in Engelmanniinae (Asteraceae: Heliantheae) based on ITS sequence data. American Journal Botany (Supplement) 84: 183.
Cuatrecasas, J. 1976 A new subtribe in the Heliantheae (Compositae): Espeletiinae. Phytologia 35: 43–61.
Doyle, J. J. 1992 Gene trees and species trees: molecular systematics as one-character taxonomy. Systematic Botany 17: 144–163.[CrossRef][ISI]
———. 1996 Homoplasy connections and disconnections: genes and species, molecules and morphology. In M. J. Sanderson and L. Hufford [eds.]. Homoplasy: the recurrence of similarity in evolution, 37–66. Academic Press, San Diego, CA.
———, and J. L. Doyle. 1987 A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15.
Eliasson, U. H. 1984 Chromosome number of Macraea laricifolia Hooker fil. (Compositae) and its bearing on the taxonomic affinity of the genus. Botanical Journal of the Linnean Society 88: 253–256.
Farris, J. S. 1970 Methods for computing Wagner Trees. Systematic Zoology 19: 83–92.[CrossRef]
Felsenstein, J. 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791.[CrossRef][ISI]
Funk, V. A. 1982 The Systematics of Montanoa (Asteraceae: Heliantheae). Memoirs of The New York Botanical Garden 36: 1–133.
Gardner, R. C. 1979 Revision of Lipochaeta (Compositae: Heliantheae) of the Hawaiian islands. Rhodora 81: 291–343.[ISI]
Harling, G. 1962 On some Compositae endemic to the Galápagos Islands. Acta Horti Bergiani 20: 63–120.
Hoffmann, O. 1890 Compositae. In A. Engler and K. Prantl [eds.], Die Natürlichen Pflanzenfamilien, vol. 4, no. 5, 87–391. Verlag von Wilhelm Engelmann, Leipzig.
Jansen, R. K., and K.-J. Kim. 1996 Implications of chloroplast DNA data for the classification and phylogeny of the Asteraceae. In D. J. N. Hind and H. J. Beentje [eds.], Compositae: systematics. Proceedings of the International Compositae Conference, Kew, 1994 vol 1, 317–339. Royal Botanic Gardens, Kew.
———, N. A. Harriman, and L. E. Urbatsch. 1982 Squamopappus gen. nov. and redefinition of Podachaenium (Compositae: Heliantheae). Systematic Botany 7: 476–483.[CrossRef][ISI]
———, H. J. Michaels, R. S. Wallace, K.-J. Kim, S. C. Keeley, L. E. Watson, and J. D. Palmer. 1991 Chloroplast DNA variation in the Asteraceae: phylogenetic and evolutionary implications. In P. S. Soltis, D. E. Soltis, and J. J. Doyle [eds.]., Molecular systematics of plants, 252–279. Chapman and Hall, New York, NY.
———, ———, and J. D. Palmer. 1991 Phylogeny and character evolution in the Asteraceae based on chloroplast DNA restriction site mapping. Systematic Botany 16: 98–115.[CrossRef][ISI]
———, and J. Palmer. 1987 Chloroplast DNA from lettuce and Barnadesia (Asteraceae): structure, gene localization, and characterization of a large inversion. Current Genetics 11: 553–564.
Karis, P. O. 1993 The Heliantheae sensu lato (Asteraceae), clades and classific
