MIAMI UNIVERSITY The Graduate School CERTIFICATE FOR APPROVING THE DISSERTATION We hereby approve the Dissertation of Melanie A. Link-Pérez Candidate for the Degree: Doctor of Philosophy ____________________________________________ Co-Director (Dr. R. James Hickey) ___________________________________________ Co-Director (Dr. Linda E. Watson) ____________________________________________ Reader (Dr. Michael A. Vincent) ____________________________________________ Reader (Dr. Elisabeth E. Schussler) ____________________________________________ Graduate School Representative (Dr. Mark R. Boardman) ABSTRACT REVISION AND MOLECULAR SYSTEMATICS OF THE NEOTROPICAL FERN GENUS ADIANTOPSIS (PTERIDACEAE) by: Melanie A. Link-Pérez Of the five major clades in the ecologically diverse Pteridaceae, the cheilanthoid ferns, comprised of approximately 20 genera and 400 species, are most obviously characterized by adaptations to xeric habitats. The cheilanthoids are known to be replete with paraphyletic genera due to the difficulties of recognizing monophyletic groups due to homoplasious adaptive morphology. The objectives of this dissertation were to complete a revision and explore the systematics of the Central and South American species of Adiantopsis Fée, a small neotropical genus in the hemionitid subclade within the cheilanthoids. Classically, Adiantopsis has been characterized by the combination of echinate spores, golden or golden-red paired carinae on the upper side of leaf axes and sometimes on stipe axes, and well-differentiated pseudoindusia. Laminar morphology is one of the more notable aspects of Adiantopsis because its species display palmate, pedate, and pinnate architectures. Using both traditional morphological analyses and molecular phylogenetics, this dissertation investigates i) the monophyly of Adiantopsis and its generic boundaries, ii) the enumeration of its species, iii) the history of character evolution in Adiantopsis, particularly the origin and evolution of the diverse laminar morphologies, and iv) uses a molecular phylogeny as a framework to investigate its biogeographic history. Phylogenetic analysis of combined plastid rbcL and atpA DNA sequences revealed that Adiantopsis is not monophyletic as traditionally circumscribed; with an expansion of generic boundaries and the transfer to Adiantopsis of three Cheilanthes species, Adiantopsis becomes monophyletic with strong support. Morphological characters to circumscribe the expanded Adiantopsis include: large, reddish, pluricellular hairs or carinae on axes; distinct, scarious pseudoindusia that cover one to occasionally two sori; and compound leaves with small, asymmetrical ultimate segments, at least some of which are stalked. Previously unrecognized diversity was revealed, particularly among palmate taxa: three new species are described. Pinnate architecture appears to be pleisiomorphic in Adiantopsis and the palmate architecture arose just once. Biogeographic analyses suggest an origin for the genus in South America, possibly the Cerrado and associated dry areas of Brazil, with a minimum of three migrations into the Caribbean. REVISION AND MOLECULAR SYSTEMATICS OF THE NEOTROPICAL FERN GENUS ADIANTOPSIS (PTERIDACEAE) A DISSERTATION Submitted to the Faculty of Miami University in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Botany by Melanie A. Link-Pérez Miami University Oxford, OH 2010 Dissertation Co-Directors: Dr. R. James Hickey & Dr. Linda E. Watson © Melanie Link-Pérez 2010 ii Contents Chapter 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Chapter 2. Revision of Adiantopsis radiata (L.) Fée (Pteridaceae) with descriptions of new palmate taxa. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12 Chapter 3. Toward a redefinition of Adiantopsis Fée (Pteridaceae): Systematics, diversification, and biogeography. . . . . . . . . . . . . . . . . . . . 57 Synopsis of Adiantopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Chapter 4. Summary and synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94 iii List of Tables Chapter 2. Table 1: Diagnostic features that can be used in distinguishing palmate members of Adiantopsis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44 Table 2: Guard cell lengths of Central and South American palmate taxa of Adiantopsis.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Table 3: Spore lengths of Central and South American palmate taxa of Adiantopsis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46 Chapter 3. Table 1: Summary of nucleotide character data. . . . . . . . . . . . . . . . . . . . . . . . . .85 iv List of Figures Chapter 2. Figure 1: Distribution of palmate Adiantopsis species. . . . . . . . . . . . . . . . . . . . .47 Figure 2: Silhouettes of pinna apices from palmate species of Adiantopsis shown to same scale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48 Figure 3: Scanning electron micrographs of spores from palmate Adiantopsis species shown to same scale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 4: Adiantopsis crinoidea. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Figure 5: Adiantopsis dactylifera. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Figure 6: Adiantopsis radiata. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53 Figure 7: Adiantopsis ternata. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Figure 8: Adiantopsis timida. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 Figure 9: Adiantopsis trifurcata. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Chapter 3. Figure 1: Three types of lamina architecture in Adiantopsis—palmate, pinnate, and pedate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .86 Figure 2: Phylogeny resulting from Bayesian analysis of combined dataset (rbcL and atpA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Figure 3: Geographic distribution of Adiantopsis species mapped onto the Bayesian 50% majority rule consensus cladogram from the analyses summarized in Fig. 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .88 Figure 4: Phytogeographic regions of Adiantopsis species mapped onto the Bayesian 50% majority rule consensus cladogram from the analyses summarized in Fig. 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .89 Figure 5: Morphological characters traditionally used to circumscribe Adiantopsis—spore ornamentation and paired adaxial carinae—mapped onto the Bayesian 50% majority rule consensus cladogram from analyses summarized in Fig. 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 v Figure 6: Lamina architecture of Adiantopsis mapped onto the Bayesian 50% majority rule consensus cladogram from the analyses summarized in Fig. 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91 Figure 7: Phylogram of Adiantopsis from one of 5040 equally parsimonious trees from the maximum parsimony analysis of the combined dataset. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Figure 8: Estimated species diversity of Adiantopsis in various phytogeographic regions of Central and South America. . . . . . . . . . . . . . . . . . . .93 vi Acknowledgements A dissertation is completed only after a very long journey, and though parts of this journey are made in solitude in the wee hours of the night, much of it is made possible and bearable—sometimes, even enjoyable—by a multitude of people along the way. First and foremost, I wish to thank my wife, Angela Link-Pérez, for all her support and encouragement during every step of this undertaking. Whether it was lending a hand in the lab labeling PCR tubes, entering herbarium specimens into my database, bringing me a meal during long hours of collecting data, or luring me away from work to enjoy some much-needed relaxation, she was always present and willing to help out in any way that I needed. I could not have achieved my goals without her; indeed, I would not even have dared to dream them. I want to thank the members of my dissertation committee: Drs. Jim Hickey, Linda Watson, Mike Vincent, Beth Schussler, and Mark Boardman. I want to especially thank Linda and Jim for being outstanding dissertation advisors and rigorous reviewers of my writing and my thinking throughout this process. Linda has been a superb mentor, ready with encouragement at just the times I most needed it. Jim was constantly prodding, goading, cajoling, or tormenting me to hasten my progress down the path of taxonomic, pteridological, and philosophical enlightenment—in that peculiarly nurturing way that is uniquely his own: “Muchas gracias, Jefe!” Many thanks to my lab-mates who made work in the lab so much fun and joined me in various travel-related adventures: Mirabai McCarthy, Susannah Fulton, and Susan Sprunt. Mirabai and Susannah will always be close to my heart in that place typically reserved for sisters. Thanks to Eric Tepe, Aaron Kennedy, and Pieter Pelser for introducing me to the joys of the molecular lab and for the stimulating discussions about phylogenetics and troubleshooting PCRs. My research and scholarly development was also furthered in various ways by Rich Moore, Chris Wood, Matt Duley, Denghui Xing, and Lara Strittmatter. I am deeply indebted to Barbara Wilson and Vickie Sandlin for the myriad ways that they supported me during my graduate studies, from keeping track of grant funding, helping me secure necessary signatures, being ready with a smile, or just being available—the department just wouldn’t be the same without them. Finally, I want to thank my children, Dane Zurwell and Megan Martin. It was probably challenging for them at times to have to share their mother with a research project, but they did so generously and with good humor. I hope someday they will understand that my love for them strengthened me and motivated me every step of the way. vii “The first step in wisdom is to call things by their right name.” ~Chinese proverb 1 1 Introduction The objectives of this dissertation were to complete a revision of the Central and South American species of the neotropical fern genus Adiantopsis and to reconstruct its evolutionary history. Both traditional morphological analyses and molecular phylogenetics were used to resolve the generic boundaries and phylogenetic position of Adiantopsis and to identify and enumerate its species. Adiantopsis Fée (Pteridaceae) The Pteridaceae is remarkable among fern families in that its members occupy a broad range of ecological niches including aquatic, terrestrial, epiphytic, and rupestral. Schuettpelz et al. (2007) hypothesize that the initial diversification within Pteridaceae was tied to evolutionary innovations that allowed its members to colonize diverse habitats. Of the five major clades currently recognized within the Pteridaceae, the cheilanthoid lineage, about 20 genera and 400 species, is characterized most obviously by adaptations to xeric habitats, a rare condition in other lineages. The high degree of adaptive morphology to such habitats has made the taxonomic circumscription of cheilanthoids problematic (Gastony and Rollo, 1995; Gastony and Rollo, 1998), and has led to a proliferation of polyphyletic genera, notably the large genera Cheilanthes Sw. and Pellaea Link (Prado et al., 2007; Schuettpelz et al., 2007). To resolve the classification of the cheilanthoids into natural groups, it is necessary to revise generic circumscriptions and possibly recognize additional genera—an essential task for understanding the evolutionary history of this large group of ferns, and to make sense of the biological and ecological drivers of their diversification. The present work partially 2 addresses this need by resolving Adiantopsis, a genus which occupies a derived position within the cheilanthoid lineage (Schuettpelz et al., 2007). Adiantopsis takes its name from the superficial resemblance it bears to some species of Adiantum L., the “Maidenhair Ferns”, particularly in regards to its dark, shiny stipes and leaf axes, marginal sori, and the shape of the pinnules. Characteristic features that distinguish Adiantopsis from other cheilanthoid genera include short, golden or golden-red, paired ridges, carinae, on the upper side of leaf axes and sometimes on stipe axes, distinct sori, and well-differentiated pseudoindusia. Echinate spores have also been used to characterize the genus (Tryon and Tryon, 1982), but doing so excludes cristatespored taxa that are best included within Adiantopsis. Laminar morphology is one of the more notable aspects of Adiantopsis because its species display palmate, pedate, and pinnate architectures. The palmate architecture is arguably the most striking of the three and is best known from the widespread A. radiata (L.) Fée, which is found throughout the Caribbean and Central America, and southward into northeastern Argentina. Adiantopsis is predominantly a genus of tropical America (Tryon et al., 1990), although the ranges of some species extend into the northern portions of Uruguay (Zuloaga et al., 2008). Some taxonomists have postulated a relationship between neotropical Adiantopsis and certain African and Madagascan cheilanthoids (Christensen, 1932; Moran and Smith, 2001); however, persuasive evidence, either molecular or morphological, to support the inclusion of these paleotropical taxa in Adiantopsis is lacking. Tryon and Tryon (1982) provided the most recent review of Adiantopsis encompassing its entire geographic range. They employed conservative generic and species concepts and recognized only seven species, all neotropical in distribution. Their generic circumscription is established by the combination of echinate spores, adaxially bicarinate laminar axes (and sometimes stipes), and asymmetrical ultimate divisions. They excluded the South American A. dichotoma (Cav.) Moore and A. regularis Moore based on their possession of cristate rather than echinate spores, and they subsumed the Caribbean A. asplenioides Maxon and A. rupicola Maxon into A. paupercula (Kunze) Fée and A. reesii (Jenman) C. Chr., respectively. 3 A recent revision of only the Caribbean taxa recognizes nine distinct species in that region alone (Barker and Hickey, 2006), including the two Cuban endemics, A. asplenioides and A. rupicola, previously excluded by Tryon and Tryon (1982). This revision also described three new species: A. parvisegmenta M.S. Barker & Hickey, A. vincentii M.S. Barker & Hickey, and A. pentagona M.S. Barker & Hickey. When the present dissertation research began, approximately 13 species of Adiantopsis were generally recognized between the two treatments provided by Tryon and Tryon (1982) and Barker and Hickey (2006) and with the discovery of a sterile hybrid with pedate laminar morphology (Hickey et al., 2003). Various researchers working on regional treatments occasionally included other, lesser-known species within Adiantopsis: for example, A. seemannii (Hook.) Maxon of Mexico (Mickel and Smith, 2004) and A. perfasciculata Sehnem of Brazil (Figueiredo and Salino, 2005). The taxonomic circumscription of cheilanthoids has been a vexing problem because of difficulties interpreting patterns in the xeric-adaptive morphology that is common within its members (Gastony and Rollo, 1995; Gastony and Rollo, 1998). As stated succinctly by Gastony and Rollo (1998), “we do not have good generic characters in cheilanthoids because we do not have good genera in cheilanthoids.” This dissertation builds upon the Caribbean revision (Barker and Hickey, 2006) by including the Central and South American taxa. The morphological and molecular analyses remedy part of the cheilanthoid problem by resolving the generic boundary of Adiantopsis, which Weatherby bemoaned, in 1943, “is none too clear” (Weatherby, 1943), and by resolving its evolutionary relationships with other members of the cheilanthoid clade, particularly the large and paraphyletic Cheilanthes, with which Adiantopsis had been hypothesized to be closely allied (Tryon et al., 1990). A Note about Species Concept It is appropriate to include here a discussion about species concept, since species concepts provide the philosophical underpinning of any attempt to organize biological diversity. The goal of using a species concept is to partition diversity into its primary units—species—in a clear and consistent manner. “The problem,” according to Stuessy, 4 “is that the determination of species is one of the most important activities of the taxonomist but also one of the most difficult (2009).” More than 20 different species concepts have been proposed (Mayden 1997), and the variations on them are almost “endless” (Stuessy, 2009). Some species concepts currently employed by practicing plant taxonomists include morphological, biological, genetic (based on genetic distance), paleontological, phylogenetic (or cladistic), and evolutionary (Stuessy, 2009). Each concept has its limitations, and some work better than others on different groups even of closely related organisms (Levin, 2000; Niklas, 1997; Stuessy, 2009). Different species concepts can lead to different ways of partitioning diversity; however, when groups of populations differ in many ways— ecologically, genetically, and morphologically—it is likely that the same species interpretations will result, regardless of which concept is used (Levin, 2000). Levin (2000) comments, “The choice of a species concept has to do, in part, with the perspective that gives one satisfaction.” For this dissertation, the concept of species that gives “satisfaction” is the evolutionary species concept (Simpson, 1961; Wiley, 1978) as articulated by Wiley (1978): “a species is a single lineage of ancestor descendant populations which maintains its identity from other such lineages and has its own evolutionary tendencies and historical fate.” The evolutionary species concept includes the following critical aspects: importance of history (in the form of ancestry), contemporary isolation from other species (however that isolation is accomplished—for example, reproductive, genetic, phenological, or ecological), and accommodation of lineages that are sexual or asexual (Haufler, 2008). Often, we do not have enough evidence to accurately delimit species using the evolutionary species concept, so we make inferences based on a combination of comparative data: morphology, ecology, biogeography, and DNA sequence data (Haufler, 2008). To make the evolutionary species concept operational, we can use various other species definitions as “provisional hypotheses” of evolutionary species (Haufler, 2008; Hickey et al., 1989). Provisional morphological hypotheses are particularly useful in this respect. According to the morphological species concept, morphological variation within a group 5 can be used to circumscribe taxa when the variation is discontinuous (Stace, 1989). These discontinuities, or phenetic gaps, help us delimit taxa because we assume that the morphological distinctions reflect underlying genetic and reproductive distinctions of a similar degree (Stuessy, 2009). Speciation is usually accompanied by phenotypic variation or differentiation (Niklas, 1997), and reproductive (i.e., genetic) isolation can lead to populations of individuals that are morphologically more similar to each other than to other populations (Stuessy, 2009). Indeed, morphological discontinuities between groups can only persist if there exists some barrier to reproduction (i.e., gene flow) (Levin, 2000; Niklas, 1997; Stuessy, 2009), and if selection does not lead to canalization of characters due to ecological factors. In this dissertation, morphological studies were used to produce initial hypotheses of species boundaries. When possible, molecular data were also brought to bear, in order to more fully realize the historical context for species delimitation. Overall, an assortment of comparative data, including morphology, ecology, geography, and DNA sequences, were called upon to delimit taxa that fulfilled the criteria of evolutionary species to my satisfaction. “We are free,” according to Levin (2000), “to create and amend species interpretations until we have a mentally satisfying organization. If these interpretations are satisfying, they will have value; and that is about as much as we can hope for.” Research Questions and Hypotheses The research was guided by the following general questions. 1) How many species are in Adiantopsis? What are they? 2) What is the generic boundary of Adiantopsis? Is it monophyletic? What morphological characters circumscribe it? 3) What is the history of character evolution in Adiantopsis, particularly the origin and evolution of the diverse laminar morphologies (pinnate, pedate, and palmate)? 4) What is the biogeographic history of Adiantopsis, and how many (if any) colonization events were responsible for the present day diversity of Adiantopsis in the Caribbean? 6 5) Has evolutionary innovations that were adaptive to different habitats played a role in the diversification of the genus? My approach in addressing the research questions was to articulate explicit hypotheses, each associated with the general questions above, that could be tested with morphological and/or molecular data. These hypotheses, their predictions, and the approach taken to test them are outlined below. 1) Adiantopsis consists of a single species. Consistent with the essential nature of all taxonomic studies, the proposed research began with the null hypothesis that the group under investigation contains only one taxon. Taxonomic work is an iterative process, wherein taxa can be delimited from related taxa when the variation among them is discontinuous (Stace, 1989). In the present research, the hypothesis that there is only one taxon was continually tested and re-evaluated as study of herbarium specimens revealed sequential discontinuities in the morphological variation present. The hypothesis was serially rejected as the accumulated morphological data failed to support it; the number of hypothesized species was increased incrementally, as needed, and was challenged by additional data and iterative examination of herbarium material. The species hypotheses resulting from morphological studies were tested with molecular data whenever possible. 2a) Adiantopsis is monophyletic. 2b) Adiantopsis is a distinct group within the cheilanthoids, meriting generic status. These closely linked hypotheses (2a and 2b) were tested with molecular data. All available species of Adiantopsis were included, along with additional taxa hypothesized to belong to Adiantopsis based on morphological data. DNA sequences obtained in this study were added to a matrix of comparable sequence data for cheilanthoid taxa available on GenBank. These data were analyzed to resolve evolutionary relationships within Adiantopsis and to explore its relationship to Cheilanthes, Doryopteris, and other cheilanthoid genera. 7 3) Pinnate and palmate laminar architecture within Adiantopsis reflect separate evolutionary lineages within the genus. The three types of laminar architecture present in Adiantopsis provide a superb opportunity to explore the origin and evolution of a great range of morphological variation within one small lineage. In a survey of nearly nine thousand extant fern species, Tryon (1964) found that more than 85 % of species and genera have pinnate leaf architecture. Therefore, he hypothesized that the pinnate form is primitive and that other architectures, such as the palmate one found in less than 1 % of species, are derived. Most species within Adiantopsis are pinnate. The first clue to the nature and origin of the pedate architecture in Adiantopsis came following the discovery of a sterile hybrid with pedate morphology in South America, A. ×australopedata Hickey, M.S. Barker & Ponce (Hickey et al., 2003). Hickey et al. (2003) postulated that the progenitors of the sterile hybrid were the palmate A. radiata and either A. perfasciculata or A. occulta Sehnem, a hypothesis supported both by the morphologies and geographic distributions of the putative parents and the hybrid. Not long after the discovery of A. ×australopedata, Barker and Hickey (2006) hypothesized that the pedate morphology exhibited by the Caribbean A. pedata (Hook.) T. Moore and A. pentagona (both proposed to be tetraploids based on guard cell size, spore length, and arcus cell height) arose as a result of hybridization between the palmate A. radiata and two different pinnate taxa endemic to the Caribbean. Molecular studies helped reveal parentage within the pedate Adiantopsis, and optimizing laminar architecture onto the phylogeny provided insight into the evolution of variation within this character. 4) Adiantopsis endemics in the Caribbean are the result of a single migration, followed by diversification. This hypothesis would be supported if endemic taxa were members of a single clade. However, if endemic taxa occur in more than one clade, parsimony 8 would indicate that there have been multiple colonization events of Adiantopsis into the Caribbean, since outgroup taxa are South American. 5) Adiantopsis diversification has been tied to radiation into different phytogeographic regions. Although Adiantopsis is presumably derived from a xeric ancestor, since adaptation to xeric habitats characterizes the cheilanthoids in general, some species are associated with more mesic habitats. It was necessary to test whether the exploitation of mesic habitats and diverse phytogeographic regions is the result of a single radiation or is convergent among Adiantopsis taxa. This hypothesis was tested by optimizing phytogeographic regions onto the molecular phylogeny. If taxa in more recently derived lineages occupy more diverse habitats than those in early diverging lineages, this would provide some evidence that adaptations to different habitats has played a role in diversification. Outline of Chapters Chapter 2 is an investigation of the palmate taxa in Adiantopsis; what was once thought to be a single, widespread species, A. radiata, was found to be six geographically and morphologically distinct species. Chapter 3 is a molecular phylogenetic study of Adiantopsis in which the generic limits of Adiantopsis are examined, morphological characters are used to circumscribe the genus, and biogeographic patterns and morphological diversity are interpreted in light of the phylogenetic framework. Chapter 4 presents a summary and synthesis of the conclusions from the dissertation and places the research in a broader context of understanding the value of revisionary and monographic work for advancing our knowledge of diversity and patterns brought about by evolutionary processes. 9 LITERATURE CITED Barker, M. S., and R. J. Hickey. 2006. A taxonomic revision of Caribbean Adiantopsis (Pteridaceae). Annals of the Missouri Botanical Garden 93:371-401. Christensen, C. 1932. The Pteridophyta of Madagascar. Dansk Bot. Ark. 7:1-203, i-xv, + 80 plates. Figueiredo, J. B., and A. Salino. 2005. Pteridophytes from four "Reservas particulares do Patrimonio Natural" (RPPNs) in the South of the metropolitan region of Belo Horizonet, Minas Gerais, Brazil. Lundiana 6:83-94. Gastony, G. J., and D. R. Rollo. 1995. Phylogeny and generic circumscriptions of cheilanthoid ferns (Pteridaceae: Cheilanthoideae) inferred from rbcL nucleotide sequences. American Fern Journal 85:341-360. Gastony, G. J., and D. R. Rollo. 1998. Cheilanthoid ferns (Pteridaceae: Cheilanthoideae) in the southwestern United States and adjacent Mexico: A molecular phylogenetic reassessment of generic lines. Aliso 17:131-144. Haufler, C. H. 2008. Species and speciation. Pages 303-331 in Biology and evolution of ferns and lycophytes (T. A. Ranker, and C. H. Haufler, eds.). Cambridge University Press, Cambridge, UK ; New York :. Hickey, R. J., M. S. Barker, and M. Ponce. 2003. An Adiantopsis hybrid from Northeastern Argentina and vicinity. American Fern Journal 93:42-44. Hickey, R. J., W. C. Taylor, and N. T. Luebke. 1989. The species concept in Pteridophyta with special reference to Isoetes. American Fern Journal 79:78-89. Levin, D. A. 2000. The origin, expansion, and demise of plant species Oxford University Press, New York. Mickel, J., and A. R. Smith. 2004. The pteridophytes of Mexico. New York Botanical Garden Press, Bronx, NY. Moran, R. C., and A. R. Smith. 2001. Phytogeographic relationships between neotropical and African-Madagascan pteridophytes. Brittonia 53:304-351. Niklas, K. J. 1997. The evolutionary biology of plants University of Chicago Press, Chicago. 10 Prado, J., C. D. N. Rodrigues, A. Salatino, and M. L. F. Salatino. 2007. Phylogenetic relationships among Pteridaceae, including Brazilian species, inferred from rbcL sequences. Taxon 56:355-368. Schuettpelz, E., H. Schneider, L. Huiet, M. D. Windham, and K. M. Pryer. 2007. A molecular phylogeny of the fern family Pteridaceae: Assessing overall relationships and the affinities of previously unsampled genera. Molecular Phylogenetics and Evolution 44:1172-1185. Simpson, G. G. 1961. Principles of animal taxonomy. Columbia University Press, New York. Stace, C. A. 1989. Plant Taxonomy and Biosystematics, 2 edition. Cambridge University Press, New York. Stuessy, T. F. 2009. Plant taxonomy: the systematic evaluation of comparative data, 2nd ed edition. Columbia University Press, New York. Tryon, R. M. 1964. Evolution in the leaf of living ferns. The Torrey Botanical Club 21:73-85. Tryon, R. M., and A. F. Tryon. 1982. Ferns and allied plants with special reference to tropical America. Springer-Verlag, New York. Tryon, R. M., A. F. Tryon, and K. U. Kramer. 1990. Pteridaceae. Pages 230-256 in The families and genera of vascular plants, vol. I, Pteridophytes and gymnosperms (K. U. Kramer, and P. S. Green, eds.). Springer-Verlag, New York. Weatherby, C. A. 1943. The type species of Cheilanthes. American Fern Journal 33:6769. Wiley, E. O. 1978. The evolutionary species concept reconsidered. Systematic Zoology 27:17-26. Zuloaga, F. O., O. Morrone, and M. J. Belgrano (eds) 2008. Catalogo de las plantas vasculares del Cono Sur : (Argentina, Sur de Brasil, Chile, Paraguay y Uruguay). Missouri Botanical Garden Press, St. Louis, Mo., U.S.A. 11 2 Revision of Adiantopsis radiata (L.) Fée (Pteridaceae) with descriptions of new palmate taxa1 Abstract—Adiantopsis radiata has long been regarded as the only palmate species in the fern genus Adiantopsis (Pteridaceae). Here, three new species with palmate architecture are described: A. dactylifera, A. timida, and A. crinoidea. Additionally, a new combination is formalized for A. trifurcata, a typically ternate species originally assigned to Cheilanthes and frequently misidentified in herbarium collections as A. monticola, a pinnate species endemic to Goiás and Tocantins, Brazil. Adiantopsis ternata is affirmed as a distinct and valid species that should not be placed in synonymy with A. radiata. These six palmate species differ from each other in the form of their pseudoindusia, the attachment of ultimate segments to the costae, the form of adaxial carinae, the shape of pinnae apices, as well as spore size and ornamentation. All six palmate species are illustrated, and a distribution map and key to the palmate Adiantopsis species are provided. Keywords—Adiantopsis, cheilanthoids, diversity, ferns, neotropical, taxonomy. The Pteridaceae is remarkable among fern families in that its members occupy a broad range of ecological niches. It appears that the early diversification within Pteridaceae was tied to evolutionary innovations that allowed members to specialize to 1 This chapter is formatted for submission to Systematic Botany. 12 diverse habitats (Schuettpelz and Pryer, 2007). Extant species occupy aquatic, terrestrial, epiphytic, and xeric-adapted rupestral habitats. Of the five major clades in Pteridaceae, the cheilanthoid ferns, about 20 genera and 400 species, are most obviously characterized by adaptations to xeric habitats such as complex leaves with small ultimate segments, abscission zones in the petioles, and frequently scales, pubescence, or waxy indument (Tryon and Tryon, 1973). Within the cheilanthoids, similar morphological adaptations to xeric habitats have confounded their taxonomic circumscription. As a result, an abundance of polyphyletic genera have been recognized, notably Cheilanthes Sw., Doryopteris J. Sm., and Pellaea Link (Gastony and Rollo, 1995; Gastony and Rollo, 1998; Prado et al., 2007). Adiantopsis Fée is a small neotropical genus of approximately forty species (Link-Pérez, in preparation) within the hemionitidoid subclade of the cheilanthoids (Rothfels et al., 2008; Windham et al., 2009). Adiantopsis lacks the scaly or waxy aspect present in many cheilanthoids but does possess small ultimate segments that typically have abscission zones in their pinnule stalks. The genus takes its name from its superficial resemblance to certain species of Adiantum, the “Maidenhair Fern,” particularly in the appearance of its dark, shiny stipes, the shape of the pinnules, and its marginal sori. Laminar morphology is one of the most notable aspects of Adiantopsis because its species display three distinct architectures: palmate, pedate, and pinnate. The palmate architecture is arguably the most striking and is best known from the widespread A. radiata (L.) Fée, which is found throughout the Caribbean and Central America, and southward into northeastern Argentina. Adiantopsis radiata has long been regarded as the only palmate species in the genus (Tryon and Tryon, 1982); however, study of herbarium specimens as part of revisionary work on Adiantopsis revealed a surprisingly large amount of variation within the classical circumscription of A. radiata. The aims of this study were to examine the pattern of morphological variation in A. radiata and revise its taxonomy accordingly. 13 MATERIALS AND METHODS Morphology—Approximately 600 palmate specimens were examined for this revision. Morphological data were obtained from herbarium specimens with the aid of a dissecting microscope. More than 70 discrete and continuous characters were scored for each taxon and the data used to construct species descriptions (see Appendix). Two to three ultimate segments per taxon were rehydrated, cleared of natural pigment, and stained with safranin and fast green in order to observe venation patterns and epidermal features, such as laminar hairs, pavement cells, and guard cells. Spores from herbarium sheets were collected on the tip of a dissecting needle (Wagner et al., 1986) and mounted in Hoyer’s Mounting Medium (Anderson, 1954; Barrington et al., 1986) on glass slides and measured after 24 hours. Spore diameter was measured from the apex of one angle of the tetrahedral spore across its middle to the opposite side, thus bisecting the spore. Measurements included the perispore but not the echinate ornamentation. Spores of all taxa were mounted with double-sided carbon adhesive tabs onto aluminum stubs and examined with scanning electron microscopy (SEM) to examine spore ornamentation and shape. Spore measurements and guard cell lengths were used to generate hypotheses about ploidy level (Barrington et al., 1986). Morphological features that distinguish the palmate taxa were documented using camera lucida and photography. RESULTS Examination of specimens revealed that what was once considered a single taxon—Adiantopsis radiata—is, in fact, composed of at least six distinct species: A. radiata, A. ternata Prantl, Cheilanthes trifurcata Baker (=Adiantopsis), and three new species, A. crinoidea, A. dactylifera, and A. timida. These six species are quite different from each other morphologically, and the principal diagnostic characters are presented in Table 1 and described below. Only one species, A. radiata, is widespread, whereas the others are more or less geographically localized (Fig. 1). Pinnae characters—Adiantopsis radiata, A. dactylifera, and A. timida typically have 5–7 pinnae radiating from the stipe apex, although some fronds of A. radiata and A. timida have as few as 3 or as many as 9 (Tab. 1). The number of pinnae is generally 3 for both A. ternata and A. trifurcata, making them appear ternate, although some fronds 14 possess 4 or 5 pinnae, revealing their palmate nature. Adiantopsis crinoidea has the most variable number of pinnae, typically 5–15. The shape of pinna apices is distinctive for each species (Fig. 2, all shown to same scale). In A. radiata, pinnae end in a basally pinnatifid or lobate apical segment, this distally very elongate with an acute apex (Fig. 2A). Adiantopsis dactylifera pinnae have the distal pinnules gradually reduced toward the pinnatifid apex, the apical segment distally elongate and basally lobate with a rounded apex (Fig. 2B). The pinnae of A. timida end in an asymmetrical mitten-shaped apical segment with an acute, obtuse, or rounded apex (Fig. 2C). In A. trifurcata, the pinnae gradually diminish in width to a very small pinnatifid apex, with the apical segment trullate to mitten-shaped with an acute, sometimes spathulate, apex (Fig. 2D). Adiantopsis ternata has pinnae that end in a distinct and generally symmetrical apical segment, this basally lobate and with a rounded apex (Fig. 2E). The pinnae of A. crinoidea gradually diminish in width to a very small conform apex ending in an asymmetrical, narrow, mitten-shaped apical segment with an acute apex (Fig. 2F). Pinnule characters—Adiantopsis radiata has pinnules that are much larger than those in the other palmate species (Tab. 1). In A. radiata, pinnules typically range from 7.0–13.5 mm long. The next longest pinnules are found in A. dactylifera, ranging from 3.0–8.0 mm long. The other species have much smaller pinnules, with lengths less than 7.0 mm, with the smallest pinnules found in A. crinoidea (2.0–5 mm long, and 1.0–2.0 mm wide). Guard cell lengths of the six species fall into four groupings that have mean lengths that are significantly different from each other (Tab. 2). The shortest guard cells are in A. dactylifera, with a mean length of 39.52 µm. The second group consists of A. radiata and A. ternata, with mean lengths of 46.44 µm and 46.00 µm, respectively. The third group consists of A. crinoidea and A. timida, with mean lengths of 51.50 µm and 52.76 µm, respectively. The longest guard cells are found in A. trifurcata, with a mean length of 55.52 µm. The most informative pinnule characters, however, are the position of the stalk within the base of the pinnules and the shape and margins of the pseudoindusia (Tab. 1). In A. radiata, the stalk enters the pinnule in a basal to sub-basal position (Fig. 6D); this means that the stalk is located close to the basiscopic edge of the pinnule. In A. 15 trifurcata, the stalk enters the pinnule in a sub-medial to basal position (Fig. 9D, 9E). In A. crinoidea, the stalk enters the pinnule in a medial position (Fig. 4D, 4E), whereas in A. dactylifera, A. ternata, and A. timida, the stalk enters the pinnule in a sub-medial to medial position (Figs. 5E, 7E, and 8D, respectively). The pseudoindusia of A. radiata are lunate with entire margins (Fig. 6C, 6D). In A. ternata and A. trifurcata, the pseudoindusia are triangular, with lacinate margins in A. ternata (Fig. 7E, 7F) and erose to entire margins in A. trifurcata (Fig. 9D, 9E). Adiantopsis dactylifera has lunate to cuspidate pseudoindusia with entire to slightly erose margins (Fig. 5D, 5E), and A. timida has lunate to sub-cuspidate pseudoindusia with entire margins (Fig. 8D). The pseudoindusia of A. crinoidea are quadrangular to lunate, sometimes imbricate, with erose to very erose margins (Fig. 4D, 4E). Carinae characters—One of the characters traditionally used to circumscribe Adiantopsis is the presence of short, golden or golden-red, paired ridges, called carinae, on the adaxial side of leaf axes and sometimes on stipe axes. The form of the adaxial carinae shows some taxonomically useful variation. In A. radiata and A. timida, the cells of the carinae are tightly organized, small, and are oriented parallel to the costa axis, with the apical edge having a smooth margin (Fig. 6E and Fig. 8C, respectively). Conversely, in A. dactylifera, the apical cells of the carinae are elongate, distinct and divaricate (loosely resembling “fingers” and, thus, giving the species its name) (Fig. 5C, 5F). Differences in the orientation of cells along the apical edge of carinae are subtler in the other species: parallel to the costa axis with the margin frequently undulate or erose in A. crinoidea (Fig. 4C); of various sizes, with most parallel to the costa axis, and some elongate and diverging from parallel by a small angle, making the margin undulate in A. ternata (Fig. 7D); and elongate and diverging about 30º from parallel to costa axis in A. trifurcata (Fig. 9C). Spore characters—Spores of the six palmate species differ remarkably (Fig. 3, Tab. 3). Spore ornamentation is echinate in both A. radiata and A. dactylifera (Fig. 3A and 3B, respectively), is arachnoid-echinulate in both A. timida and A. ternata (Fig. 3C and 3E, respectively), and is echinulate in A. trifurcata and A. crinoidea (Fig. 3D and 3F, respectively). Spore length is highly variable between species, with mean spore lengths for all species being significantly different (Tab. 3). Adiantopsis radiata, A. dactylifera, 16 and A. timida have rather small average spore sizes, between approximately 31 and 35 µm in length. In contrast, A. trifurcata and A. crinoidea have larger spores, with average spore lengths between 47 and 51 µm. Adiantopsis ternata is somewhat intermediate with an average spore size of 40 µm in length. DISCUSSION Our research revealed that what was once considered a single taxon—Adiantopsis radiata—is composed of at least six distinct species, which differ from each other morphologically (summarized in Tab. 1 and in text, above) and in their geographic distributions (Fig. 1). Only A. radiata is widespread. These six species differ from each other in the form of the pseudoindusia, the position of pinnule stalks within the pinnules, the form of the adaxial carinae, the shape of pinnae apices (Fig. 2), frond dimorphism, and spore size and ornamentation (Tab. 3, Fig. 3). Three species are described below for the first time: A. dactylifera, A. timida, and A. crinoidea. Early workers provided names for two generally-ternate species of Adiantopsis—A. ternata Prantl and Cheilanthes trifurcata Baker—but most often these have been dismissed and placed into synonymy: A. ternata with A. radiata, and C. trifurcata with A. monticola. Although A. ternata and C. trifurcata generally have three pinnae, some fronds possess up to five pinnae, indicating that they are, indeed, palmate rather than simply ternate species. Adiantopsis ternata was described from a specimen collected by Humboldt in the Orinoco region of Venezuela; the species was set apart from A. radiata by Prantl in 1883 on the basis of A. ternata possessing three main lamina branches (pinnae). However, Tryon and Tryon (1982) considered A. ternata to be juvenile or depauperate specimens of A. radiata. On the contrary, A. ternata is quite distinct from A. radiata: for example, the pseudoindusia are triangular with lacinate margins in A. ternata but lunate with entire margins in A. radiata, spores are arachnoid-echinulate in A. ternata but echinate in A. radiata, and the pinnules are generally less than 7.0 mm long in A. ternata and more than 7.0 mm long in A. radiata. Cheilanthes trifurcata (new combination of A. trifurcata presented below) has most often been misidentified as A. monticola (Gardner) T. Moore, a small, pinnate taxon endemic to Goiás and Tocantins, Brazil. There are some superficial resemblances, 17 including the small size of the pinnae and pinnules, as well as habitat. However, the two taxa are quite distinct morphologically (and at the molecular level; unpublished data). For example, A. trifurcata has triangular pseudoindusia with erose margins whereas A. monticola has lunate pseudoindusia with entire margins. The triangular pseudoindusia, echinulate spores, and pinnules that are less than 6.0 mm long clearly distinguish A. trifurcata from A. radiata (characters listed above in the comparison of A. radiata with A. ternata). Data from guard cell lengths (Tab. 2) and spore lengths (Tab. 3) suggest that at least two palmate species, A. crinoidea and A. trifurcata, are polyploids, based on the larger sizes of their guard cells and spores in relation to the sizes observed in A. radiata, a known diploid (Walker, 1973). Spore length has been previously used as a proxy for estimating ploidy level in ferns (Barrington et al., 1986). For a sphere, a 1.26 increase in diameter accompanies a doubling of volume (Barrington et al., 1986); thus, for tetrahedral spores as in Adiantopsis, the spore length of a known diploid can be multiplied by 1.26 to estimate the spore length that might be observed with a doubling of chromosome number. Adiantopsis radiata has a known base chromosome number of n=30 (Walker, 1973) and can serve as a representative sexual diploid for the genus; a doubling of spore volume would be accomplished by an increase in spore length from a mean of 32.91 µm to 41.47 µm. Based upon this rationale, taxa with mean spore lengths equal to or greater than 41.47 µm are presumed to be tetraploid (A. trifurcata and A. crinoidea). The large size of A. crinoidea spores may indicate an even higher ploidy for this taxon (Fig. 3, Tab. 3), since spore size in A. crinoidea (51.41 µm) just falls short of the estimate for an octoploid (52.3 µm). Spore size in A. timida is in the range expected for a diploid (Tab. 3), but guard cell size in this species is significantly larger than those found in the known diploid A. radiata; in fact, guard cell size in A. timida is similar to that found in A. crinoidea, the species with the largest spores and the most compelling evidence to suggest that it is a polyploid species (discussed above). In light of the conflicting clues provided by spore and guard cell length data as to the possible ploidy of A. timida, we refrain from making any inference regarding its ploidy. All inferences of ploidy based on spore size and guard cell size must be taken with some caution, however, because both traits can be affected by numerous other selective pressures. For example, 18 in addition to changes in ploidy level, spore-size can be variable due to adaptation for dispersal (larger to stay closer to parent plant in insular habitats), increase for nutritional reasons, and environmental parameters (reviewed by Barrington et. al. 1986). Guard cell size, while also correlated with ploidy level, can be complicated by environmentally induced variability (e.g., smaller stomates in arid habitats) (Barrington et al., 1986). The six palmate taxa presented and illustrated below represent six morphologically distinct species. Molecular data for all available palmate taxa support this circumscription (Link-Pérez et al., in preparation). Our data also suggest the existence of several additional, undescribed palmate taxa in the Guyana, Suriname, and French Guiana region; however, description of those species is postponed until additional material can be studied. KEY TO THE PALMATE SPECIES OF ADIANTOPSIS 1. Lamina primarily ternate, pseudoindusia triangular . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 2. Pinnae fastigiate, pseudoindusia +/- entire. . . . . . . . . . . . . . . . . . . . . . . . A. trifurcata 2. Pinnae spreading, pseudoindusia laciniate. . . . . . . . . . . . . . . . . . . . . . . . . . .A. ternata 1. Lamina palmate with 5 or more pinnae, pseudoindusia lunate to quadrangular. . . . . . .3 3. Carinae digitate, composed of separate and distinct cells arranged perpendicular to the axes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .A. dactylifera 3. Carinae continuous, the cells contiguous and forming a continuous sheet, most cells parallel to the axes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 4. Fronds dimorphic, pinnules generally less than 7.0 mm long, pinnule stalks medial to sub-medial. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 5. Pseudoindusia lunate with entire margins, carinae adaxial on stipe beginning at midpoint or proximal third. . . . . . . . . . . . . . . . . . A. timida 5. Pseudoindusia quadrangular with erose margins, carinae absent on most stipes or adaxial on distal end only. . . . . . . . . . . . . .A. crinoidea 4. Fronds monomorphic, pinnules generally more than 7.0 mm long, pinnule stalks sub-basal to basal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. radiata 19 TAXONOMIC TREATMENT Adiantopsis crinoidea Link-Pérez & Hickey, ined.—TYPE: VENEZUELA. Territorio Federal Amazonas: Atures. Carretera Pto. Ayacucho—El Burro: km 22 al lado oriental: Laja de Galipero, 5°48´N, 67°30´W, 7 Apr 1992, Gröger & Barcroft 265 (holotype: MO!; isotype VEN!). Rhizome decumbent to erect; scales acicular, bicolorous, the central band shiny, dark brown, occupying 1/2 to 2/3 of the width, margins golden. Fronds erect, dimorphic, 14.5–47.0 cm long, shorter fronds sterile. Stipes medium brown, castaneous, to atropurpureous, lustrous, about 58–77% of frond length, 8.5–35.0 cm long; carinae mostly absent or adaxial distally, adaxial throughout on juvenile fronds, reddish-gold, to 0.1 mm tall; stipe scales similar to rhizome scales, bicolorous, sparse, limited to very proximal portion of the stipe and at the stipe apex where they are brown; hairs absent. Laminae palmate, circular, geniculate, 6.0–15.0 cm long, 7.0–18.0 cm wide; coriaceous, drying yellowish-green to dark tannish-brown, whitened epidermal cells randomly dispersed adaxially and abaxially; hydathodes marginal; scales absent; three-celled hairs diffuse abaxially, rare to absent adaxially, septate-capitate, basal cell elongate and colorless, middle cell short, red to orange, apical cell bulbous, amber, reddish-orange, or colorless; stoma anomocytic, guard cells 40.5–( x =51.5)–62.1 µm long. Costae persistent, brown to atrocastaneous; carinae adaxial on all costae, translucent reddishbrown to reddish-gold, apical cells neither distinct nor divergent, margins frequently undulate or erose, to 0.15 mm tall; scales and hairs absent. Pinnae 5–15, spreading radially from stipe apex, narrowly oblong, gradually diminishing in width to a very small conform apex ending in an asymmetrical, narrow, mitten-shaped apical segment with an acute apex; central pinna 5.0–15.0 cm long, 0.4–1.1 cm wide, bearing 23–49 pinnule pairs; second pinnae slightly shorter than central pinna; basal pinnae 1.4–4.8 cm long, 0.4–0.5 cm wide, bearing 5–20 pinnule pairs; basal flabellate divisions attached between pinnae at stipe apex, typically fertile when frond is fertile. Pinnules transverse, shortstalked, articulate, narrowly ovate, 2.0–5.0 mm long, 1.0–2.0 mm wide, length:width ratio 2.3–2.9, acroscopic auricles to 1.0 mm long; base acute basiscopically, obtuse to 20 truncate acroscopically; margins entire; apices acute to round; stalks medial, persistent, to 0.5 mm long, frequently with large remnants of senescent lamina tissue attached after pinnule abscission. Veins free, anadromous, occult, ending submarginally in adaxial hydathodes, basiscopic veins sometimes diverging then anastomosing. Sori marginal, discrete, occasionally two per pseudoindusia, (1–)3–11 per pinnule. Pseudoindusia distinct, quadrangular to lunate, scarious, chaffy and dry-looking, brownish-tan to yellowish-white, occasionally overlapping or imbricate, 1.0–1.6 times wider than long, 0.38–0.6 mm long, 0.48–0.77 mm wide, margin erose to very erose. Sporangia globose, short stalked. Spores pale yellow, tetrahedral, 47.3–( x =51.4) –55.4 µm long, echinulate with echinae formed of single rodlets extending above dissected bases of fused rodlets (visible with scanning electron microscopy), laesurae clearly visible with brightfield examination. Chromosome number unknown. Figure 4. A–F. Additional Specimens Examined—VENEZUELA. Estado Bolivar: Cedeño, rocky igneous forested slopes and bluffs of Cerro Medano, 22.5 km SW of Caicara del Orinoco, just SW of Sacuima, 07°36´N, 66°15´W, 100–200 m, 2 Sep 1985, Steyermark et al. 131214 (MO, VEN); Municipio Raul Leoni, bosque y arbustales sobre lateritas bauxíticas, cabeceras Río Túriba y Caño La Miel, 45 km al este de Pijiguaos, 6°34´N, 66°23´W, 800 m, Jun 1989, Fernandez and Delgado 5860 (MO). Territorio Federal Amazonas: Departamento Atures, 23 km NE of Puerto Ayacucho along the road to El Burro, laja and savanna at its base, 5°51´N, 67°29´W, 80–150 m, 17 Apr 1978, Davidse and Huber 15283 (GH, MO, NY); Vegetacion sobre laja granitica, ca. 3.5 km al N de Puerto Ayacucho, al margen este de la carretera Pt. Auacucho el Burro, 80 m, 14 Oct 1988, Romero 1708 (VEN). Distribution and Habitat—This terrestrial species is known only from the departments of Amazonas and Bolivar in Venezuela. It has been found among igneous rocks in forests or shrub islands, or on forested slopes, 80-800 msl. Comments— Adiantopsis crinoidea is readily differentiated from the other palmate species. Its most distinguishing features are its strong dimorphism, the chaffy quadrangular pseudoindusia with very erose, jagged margins, and the substantial pinnule stalks that are very distinct from the lamina tissue. Where these stalks enter the lamina, a 21 portion of the stalk seems to be flattened and raised above the lamina surface presenting a “shield-like” appearance with a texture very similar to the axes. Adiantopsis crinoidea is also unusual in frequently having more than 7 pinnae, whereas A. timida, A. dactylifera, and A. radiata typically have 7 or fewer. In addition to the aforementioned characters, it can be distinguished easily from A. dactylifera by lacking digitate cellular projections along the crest of the carinae (a condition unique to A. dactylifera), and from A. radiata by the differing position of the pinnule stalks within the pinnules, medial in A. crinoidea but basal or sub-basal in A. radiata. Adiantopsis dactylifera Link-Pérez & Hickey, ined.—TYPE: PERU. Cajamarca: San Ignacio Province. Chirinos, Localidad Las Juntas, Bosque primario, 5°22´S, 78°46´W, 550–650 m, 11 Mar 1998, Campos 4834 (holotype: NY!; isotype MO!). Rhizome ascending to erect; scales acicular, bicolorous, the central band shiny, dark brown to black, occupying 2/3 or more of the width, margins brown to reddishbrown, with conform concolorous tan to brown, lanceolate scales interspersed. Fronds strict, monomorphic, 7.0–41.0 cm long. Stipes ebeneous to atropurpureous, lustrous, terete, about 55–65% of frond length, 2.5–24.0 cm long; carinae absent; scales brown, lanceolate, sparse along the lower quarter of the stipe and at the stipe apex; catenate hairs, reddish-gold, rare. Laminae palmate (radially bipinnate), circular, geniculate, 7.3–18.0 cm long, 6.0–21.0 cm wide; carnose to sub-spongiose, drying olive-green to dark brownish-black, whitened epidermal cells randomly dispersed adaxially; hydathodes marginal; scales absent; three-celled hairs diffuse abaxially, rare adaxially, septate, basal cell elongate and colorless, middle cell short, white (orange to red), apical cell bulbous, yellow, white, or colorless; stoma anomocytic, guard cells 32.4–( x =39.52)–48.6 µm long. Costae persistent, ebeneous to atropurpureous, lustrous; carinae adaxial on all costae, translucent reddish-brown, to 0.27 mm tall, apical cells distinct and divaricate, elongate, to 0.12 mm long; scales absent; catenate hairs sparse, at abaxial costule-costa junctures. Pinnae 5–7, spreading radially from stipe apex, linear to narrowly fusiform, 22 distal pinnules gradually reduced toward the pinnatifid apex, the apical segment distally elongate and basally lobate with a rounded apex; central pinna 5.0–16.0 cm long, 0.9–1.5 cm wide, bearing 33–50+ pinnule pairs; second pinnae slightly shorter, 4.5–15.7 cm long, 0.9–1.4 cm wide, bearing 24–50 pinnule pairs; basal pinnae 2.4–12.8 cm long, 0.7–1.1 cm wide, bearing 8–43 pinnule pairs; basal flabellate divisions attached between pinnae at stipe apex, symmetrical, typically fertile when frond is fertile. Pinnules slightly ascending to slightly descending, sessile to short-stalked, articulate, narrowly oblong to oblong, drying revolute to conduplicate, 3.0–8.0 mm long, 1.1–2.3 mm wide, length:width ratio 2.7–3.6; acroscopic auricles to 1.0 mm long; base obtuse (rounded) to acute basiscopically, truncate (to acute) acroscopically; margins entire, flat to slightly revolute; apices round; stalks medial to sub-medial, persistent, peg-like to aculeate after pinnule abscission, to 0.77 mm long. Veins free, anadromous, occult to obscure, ending submarginally in adaxial hydathodes. Sori marginal, discrete, occasionally confluent on very fertile fronds, (1–) 8–20 per pinnule. Pseudoindusia distinct, lunate to cuspidate, scarious, yellowish-tan to grayish-green or olive, often black maculate, 1.1–1.5 times wider than long, 0.31–0.48 mm long, 0.43–0.65 mm wide, margin entire to slightly erose. Sporangia subglobose, long stalked. Spores yellow to golden at maturity, tetrahedralglobose, 27.7–( x =30.8) –33.8 µm long, echinate with dissected bases, echinae robust, 2– 3 µm long, interspersed with fine rodlets up to 1 µm long jutting up between echinae on perispore (visible with scanning electron microscopy), laesura obscured by echinae. Chromosome number unknown. Figure 5. A–G. Additional Specimens Examined—PERU. Cajamarca: Jaén, carretera between Shumba and Las Juntas, seasonally dry forest, with several species of columnar cacti, 5°42´S, 78°48´W, 750 m, 25 Mar 2001, van der Werff et al. 16380 (MO, NY); Jaén, 780 m, Feb 1954, Coronado and Cerrón 44 (MICH). Amazonas: Bagua District, Muyo Bosque primario, selectivamente intervenido, helecho terrestre, soros verdes, 5°18´S, 78°20´W, 600 m, 4 Feb 1999, Vásquez et al. 25968 (MO). Distribution and Habitat—This terrestrial species is known only from the departments of Cajamarca and Amazonas in Peru. It has been collected in seasonally dry forests and woods from 550-780 msl. 23 Comments—The most distinguishing feature of Adiantopsis dactylifera is the presence of the digitate cellular projections along the crest of the carinae, a condition that is unique to this taxon. Examination of spores using brightfield microscopy reveals large, moderately well dispersed echinae in A. dactylifera and A. radiata. These echinae are 2– 3 µm long in A. dactylifera and 2.7–4.05 µm long in A. radiata. In A. ternata and A. timida the ornamentation is arachnoid-echinulate, and in A. trifurcata and A. crinoidea the ornamentation is echinulate; in all four of these latter taxa, the echinae—less than 1.35 µm long—are considerably smaller than those of A. dactylifera and are difficult to discern using brightfield microscopy. Adiantopsis radiata differs in lacking digitate carinae and having larger pinnules. In addition, the position of the stalk in the ultimate segments (pinnules) is medial to sub-medial in A. dactylifera but sub-basal in A. radiata. In addition, the pinnules of A. radiata rarely become revolute or conduplicate upon drying. Of unknown significance is the observation that spores on herbarium sheets are often seen clinging to the carinae, entangled in the divaricate and distinct cells of the carinae. Adiantopsis radiata (L.) Fée, Gen. Felic. [Mém. Foug. 5]: 145. 1852.—TYPE: Sloane “Adiant. 3 radiatum” (lectotype LINN 1252.1, chosen by Lellinger, Mem. New York Bot. Gard. 23: 3, 1972). Adiantum radiatum L., Sp. Pl. 2: 1094. 1753. Cheilanthes radiata (L.) J.Sm., J. Bot. (Hooker) 4:159. 1841. Hypolepis radiata (L.) Hook., Sp. Fil. 2: 72. 1852 Rhizome erect; scales acicular, bicolorous, the central band shiny, dark brown to black, occupying 1/3 to 2/3 of the width, margins golden to reddish-gold, with conform concolorous tan to brown scales interspersed. Fronds strict, monomorphic, (7.0–)19.1– 24 63.0 cm long. Stipes atropurpureous to atrocastaneous, lustrous, typically 50–72% of frond length, 3.3–47.0 cm long; carinae absent on most fronds or rarely adaxial at distal end only, adaxially throughout on juvenile fronds, golden, to 0.1 mm tall; scales tan or golden, concolorous, sparse, these generally restricted to stipe base and apex; catenate hairs rare. Laminae palmate, circular, geniculate, 5.0–23.6 cm long, 4.8–24.3 cm wide; chartaceous (to spongiose), drying green, olivaceous, to blackish-brown; hydathodes marginal; scales absent; three-celled hairs diffuse abaxially, rare to absent adaxially, septate, basal cell elongate and colorless, middle cell short, amber, apical cell bulbous, white, pale-yellow, or colorless (occasionally amber or red); stoma anomocytic, guard cells 36.45–( x =46.44)–54.0 µm long. Costae persistent, atropurpureous to atrocastaneous, lustrous; carinae adaxial on all costae, golden, to 0.2–0.4 mm tall, apical cells parallel to the costa axis; scales absent; catenate hairs sparse, at abaxial costulecosta junctures, amber. Pinnae (3–)5–7(–9), spreading radially from stipe apex, narrowly elliptic to fusiform, ending in a basally pinnatifid or lobate apical segment, this distally very elongate with an acute apex; central pinna 3.7–17.0 cm long, (1.25–)1.6–2.45 cm wide, bearing (9–)19–40 pinnule pairs; basal pinnae to 1.9–9.6 cm long, 0.9–2.1 cm wide, bearing 6–27 pinnule pairs; basal flabellate divisions attached between pinnae at stipe apex, typically fertile when frond is fertile. Pinnules transverse to slightly ascending, short-stalked, articulate, narrowly oblong to oblong, 7.0–13.5 mm long, 2.0–3.5 mm wide, length:width ratio 3.0–( x =3.3)–4.0, acroscopic auricles to 1.0 mm long; base excavate basiscopically, truncate acroscopically; margins entire to crenulate; apices round to acute; stalks sub-basal to basal, slightly reflexed abaxially, persistent, peg-like after pinnule abscission, to (0.4–)0.7–1.0 mm long, frequently with remnants of senescent lamina tissue attached. Veins free, anadromous, obscure to occult (sometime prominent due to color), ending submarginally in adaxial hydathodes. Sori marginal, discrete, (1– )13–21 per pinnule. Pseudoindusia distinct, lunate, scarious, pale-yellow, white, or colorless, sometimes approaching lamina color, usually 1.5–3.0 times wider than long, 0.18–0.42 mm long, 0.36–0.9 mm wide, margin entire. Sporangia subglobose, long stalked. Spores yellow to golden at maturity, tetrahedral-globose, 27.0–( x =32.91)–37.8 µm long, echinate with dissected bases, echinae robust, 2.7–4.05 µm long, visible with brightfield microscopy, surface between echinae a complex network of rodlets (visible 25 with scanning electron microscopy), laesurae obscured by echinae. Chromosome number N=30 (Walker, 1973). Figure 6. A–F. Representative Specimens Examined—ARGENTINA. Corrientes: Dep. Ituzaingó, Estancia "Puerto Valle," densely wooded, moist ravine by the R. Paraná, 24 Oct 1954, Pedersen 2950 (GH, NY, US); Dep. Santo Tomé, Virasoro, habita en lugares húmedos y sombrios, 13 Nov 1944, Barrola 1196 (GH, NY). Misiones: Dpto. General Belgrano, Parque Provincial Urugua-i, Sendero Islas Malvinas, norte de Ruta 19, along trail in interior of jungle, mostly closed canopy, infrequent fern of understory, 2 Jun 2006, Link-Pérez & Cabral 183 (CTES, MU). BELIZE. Cayo: Vicinity of La Flor at Rio de la Flor, 6 mi. south of Grano de Oro, limestone out-crops abundant, terrestrial, in forest along trail, 3 Jun 1973, Croat 23775 (F, MO); Mountain Pine Ridge, Douglas Silva Village, along road to Rio Frio Cave, moist bank along small stream, 17 Jun 1993, Vincent et al. 6098 (MU). BOLIVIA. Beni: Ballivian Province, Km 35 on Yucumo-Rurrenabaque road, Agric.-Tech. High School at Rio Colorado, wet forest, 14°15´S, 67°5´W, 235 m, 11 Jul 1990, Fay and Fay 2666 (MO). Chuquisaca: Provincia Hernando Siles, outside town of Monteagudo, in Area Protegida Municipal (the water treatment plant for the town of Monteagudo), along a red dirt road through the protected "Bosque Montano Humido" (humid/semi-humid montane forest), 19°45.7´S, 63°54.8´W, 1199 m, 16 Aug 2006, LinkPérez et al. 327 (HSB, MU). Santa Cruz: Provincia Vallegrande, road from the town of Vallegrade to Masicuri (approximately 1-2 km before reaching Masicuri), on slopes above stream-on slopes of a gulley, partially closed canopy of about 10m, disturbed forest by river, 18°48.5´S, 63°49´W, 826 m, 21 Aug 2006, Link-Pérez et al. 355 (MU, USZ). La Paz: Prov. Murillo, 45.5 km below the dam at Lago Zongo, Zongo Valley, vicinity of Cahua hydroelectric plant, disturbed, moist forest, 16°3´S, 68°1´W, 1200-1400 m, 23 Dec 1984, Solomon 12916 (MO). BRAZIL. Bahia: Colonia de Gongogy, 9 Oct 1918, Curran 559 (F, MO). Distrito Federal: Gallery forest ca. 25 km. S.W. of Brasília, 1000 m, 20 Feb 1966, Irwin et al. 13070 (F, MO, SP). Espirito Santo: Santa Maria do Jetibá, Pedra do Garrafão, afloramentos rochosos, com pequenas matas úmidas nos vales, 20°10´S, 40°55´W, 1490 26 m, 18 Jul 2007, Labiak and Fontana 4181 (NY). Goiás: Serra Geral do Paraná, Riacho, ca. 3km S. of São João de Aliança, 850 m, 14 Mar 1971, Irwin et al. 31736 (F, SP, UB). Maranhão: Municipio Sta. Luzia, Fazenda AGRIPEC (Varig Airlines), 7km. W of Buriticupu on BR222, 11 km. along road from entrance to Rio Pindare where Carajas railroad runs parallel to river, 4°3´S, 46°24´W, 3 Apr 1983, Taylor et al. E1122 (MO). Mato Grosso: Município de Tangará da Serra, estrada para Tapirapuã, ca 2 km de Tapirapuã, mata, terrestre, 14°54´S, 57°45´W, 27 Dec 1994, Windisch et al. 7766 (SP). Minas Gerais: Munic. Poços De Caldas, Complexo da Serra da Mantiqueira, Serra dos Poços, Cahoeira da Antas (Usina Hidroelétrica das Antas), ca. 4 km da Cidade, mata de encosta seca semi-decidua com mata ciliar e campos, 21°47´S, 46°34´W, 1200-1250 m, 15 Jun 1995, Pietrobom da Silva 1849 (MO). Pará: Municipio Conceição Do Araguaia, range of low hills ca. 20 km. west of Redenção, near Côrrego São João and Troncamento Santa Teresa, delicate fern in forest near small stream, 8°3´S, 50°10´W, 350-620 m, 12 Feb 1980, Plowman et al. 8746 (F, MICH). Paraná: Tibagy, in dense woods, 5 Jun 1934, Reiss 68 (F, MICH). Rio de Janeiro: Paraty, extada Paraty-Cunha, 700 m, 3 Feb 2000, Amorim et al. 3264 (SP). Rondônia: Vicinity of Santa Barbara, 15 km. east of Km. 117, forrest on terra firme, laterite soil, terrestrial growing in small clumps, 16 Aug 1968, Prance and Ramos 7002 (F). Santa Catarina: Mun. Florianópolis, forest, Morro Itacorubí, Ilha de Santa Catarina, 27°34´S, 48°29´W, 150 m, 12 Mar 1952, Smith and Reitz 6164 (MO). São Paulo: Munic. Teodoro Sampáio, Região do Pontal do Paranapanema, Parque Estadual Morro Do Diabo, ca. 30 km da Sede pela trilha, Córrego 7 de Setembro ou Rio da Onça ou Angelim, mata latifoliada mesófila tropical seca semidecídua, terrestre com solo úmido e sombreado, margem do rio, muito abundante, 22°32´S, 52°11´W, 250-350 m, 16 Jan 1995, da Silva 1558 (MO). COLOMBIA. Cauca: Mun. Rosas, al SW. de la población, Vereda Mangavieja, Quebrada afluente de la Quebrada San Gil, 1200 m, 1 Jul 1975, Acosta-Arteaga 962 (AAU). El Magdalena: Growing on damp shaded sandstone banks in forest near Codazzi, 350 m, 22 Oct 1943, Haught 3770 (MICH). COSTA RICA. Alajuela: Vicinity of Los Chiles, Rio Frio, low tropical rainforest, with palms prominent, marshy, deep shade, 11°2´N, 84°44´W, 30-40 m, Aug 1949, Holm and Iltis 731 (F, MO). Guanacaste: Parque Nacional Guanacaste Cantón de Liberia, 27 Estación Cacao, Faldas de Cerro pedregal, frente a la estación, bosque medio húmedo transición a bosque seco, 10°55´N, 85°29´W, 700 m, 5 Jun 1990, Hammel 17786 (F, MO). CUBA. Oriente: Deciduous woods near base of Loma Menqura, ~680 m, 1-3 Feb 1910, Shafer 3777 (F); Monte Verde, Jan-Jul 1859, Wright 963 (F, MO). Unknown province: 1889, Eggers 4813 (F). GUATEMALA. Escuintla: Finca Volcan, Alta Verapaz, in soil, virgin forest, 2000 m, 22 Feb 1939, Wilson 238 (F). Huehuetenango: Canyon tributary to Río Trapichillo, between Democracia and canyon of Chamushú, terrestrial, 1000-1100 m, 24 Aug 1942, Steyermark 51241 (F). HONDURAS. El Cayo: Vaca, on dry log on hillside, 11 Apr 1938, Gentle 2475 (F, MICH). Olancho: Along Río Olancho, on road between Gualaco and San Bonito Oriental, 7.4 mi NE of San Estéban, rocky area along road and in rocky stream, on rocks, 15°20´N, 85°42´W, 540 m, 7 Feb 1987, Croat and Hannon 64353 (MO). JAMAICA. Clarendon: Peckham Woods, wooded limestone hillside, in soil, c. 2500 ft, 10 Mar 1954, Proctor 8445 (MO); Summit area of Crofts Mountain, 2500-2650 ft, in dense moist woodland sloping into a limestone sinkhole, 4 Oct 1968, Proctor 29256 (F). Trelawny: Vicinity of Troy, 600-660 m, 30 Jun 1904, Maxon 2954 (MICH). MEXICO. Chiapas: Municipio of Las Margaritas, tropical rain forest, low ridges at the confluence of the Rio Ixcán with the Rio Lacantum (Rio Jataté) on the Guatemala border, 300 m, 14 Mar 1973, Breedlove and McClintock 34125 (MICH, MO). Oaxaca: Mpio. Sta. María Chimalapa, Cañada del Río Negro, en la embocadura del Arroyo Huaponu, ca. 15 km al S. de Sta. María, 16°56´N, 94°39´W, 300 m, 15 Oct 1985, Hernández and González 1670 (MO). Veracruz: Municipio de Xalapa, Ejido Mártires de Chicago, matorral en cañada, 1300 m, 29 Jun 1974, Ventura 10238 (MICH). NICARAGUA. Matagalpa: Moist, dense rain forest about 6-10 km northeast of Matagalpa, road to El Tuma, terrestrial, 1000 m, 14-16 Jan 1963, Williams et al. 23863 (F). PANAMA. Veraguas: 18 km W of Las Minas, N slope of Cerro Alto Higo (Known locally as el Montoso), 2400-3000 m, 6 Aug 1978, Hammel 4274 (MO). 28 PERU. San Martín: Along road between Tarapoto and Moyabamba, valley of Río Mayo, midway between Km markers 562 and 563, ca. 57 km W of Tarapoto, 15 km W of Tabalosos, disturbed forest, 6°15´S, 76°41´W, 600 m, 11 Apr 1984, Croat 58126 (MO). VENEZUELA. Bolivar: Edo. Zulia, entre Las Tres Marías y el río Chiquito, trayecto de unos 8 km al sureste de Las Tres Marías, en bosque macrotérmico e higrófilo con areas ocasionales con elementos deciduos, en bosque, 10°25´N, 70°55´W, 14-16 Feb 1980, Bunting and Stoddart 9004 (NY). Tachira: Distrito Córdoba, Cerro La Camiri, just south of the town of Río Negro, steep forested sandstone cliffs with open landslide areas, 7°36´N, 72°10´W, 430-530 m, 6 Nov 1982, Davidse and Gonzalez 21589 (MO, NY). WEST INDIES. Dominica: Cliffs in woodlands on the North River watershed, on the western slopes of Morne Brule, Portsmouth, 29 Aug 1938, Hodge 72 (MO); St. Paul, mesophytic woods of Oreopanax, Myrcia, Vitex & c., rocky slopes of Morne Cola Anglais, ~1800 ft, 18 Jun 1965, Webster 13425 (MICH, MO). Martinique: 1868, Sidier sn (F). Montserrat: Roadside embankment north of Cudjor lerad, 8 Feb 1907, Shafer 453 (F). Guadeloupe: Falaises de las riviere des Peres, Blanchau sn (F). Distribution and Habitat—Adiantopsis radiata is the most widespread species in the genus and is found throughout the Caribbean and Central America, southward into Northeastern Argentina. It is primarily a terrestrial forest species growing in shaded ravines of streams and rivers. It has also been collected among rocks or on calcareous or sandstone slopes, from 100–2000 msl. Comments—The most distinguishing features of Adiantopsis radiata are the apical segments of the pinnae that have very long, contracted apices (Fig. 2 and Fig. 6F), the sub-basal to basal position of the stalk in the pinnules, and the scarious, lunate pseudoindusia. It is unlikely that A. radiata would be confused with either of the two ternate species, A. ternata or A. trifurcata, due to A. radiata having lunate pseudoindusia rather than triangular pseudoindusia, in addition to the differences in the number of pinnae: generally 7 or more in A. radiata but usually just 3 in A. ternata and A. trifurcata. Adiantopsis radiata differs from A. dactylifera in lacking digitate carinae. The pinnules of A. radiata are also much larger than those in A. dactylifera and the other palmate taxa. In addition to pinnule size, A. radiata can be distinguished from A. crinoidea in the shape of the pseudoindusia, lunate with entire margins in A. radiata but quadrangular with 29 erose margins in A. crinoidea. Adiantopsis radiata has pinnule stalks entering the ultimate segments in a sub-basal to basal position, whereas in A. timida, A. crinoidea, and A. dactylifera the stalks are in a medial to sub-medial position. The echinae of A. radiata spores are very evident with brightfield microscopy, whereas those of A. crinoidea and A. timida are too delicate and short to be easily visualized with brightfield examination. Adiantopsis ternata Prantl, Gartenfl. 32: 101. 1883.—TYPE: VENEZUELA. Orinoco-gebiet, Humboldt s.n. (holotype B digital loan!, B photo NY!, B fragment NY!). Rhizome erect to slightly ascending; scales acicular, bicolorous, the central band shiny, black, occupying 1/2 to 2/3 of the width, margins golden, lacking intermixed concolorous scales. Fronds erect, monomorphic, 4.5–29.0 cm long. Stipes ebeneous to atropurpureous, matte to sub-lustrous, terete, about 40–55% of frond length, 2.0–18 cm long; carinae absent; scales golden, concolorous, sparse at the stipe base, rare and biseriate at the stipe apex; catenate hairs rare. Laminae ternate to palmate, trullate to obtrullate on ternate fronds and pentagonal (with very narrow base) on palmate fronds, not to weakly geniculate, 3.0–13.5 cm long, 2.5–11.0 cm wide; spongiose, drying olivaceous to cinnamomeous; hydathodes marginal, green, occasionally dark margined; scales absent; three-celled hairs diffuse abaxially, rare to sparse adaxially, septate, basal cell elongate and colorless, middle cell short, orange to red, apical cell bulbous, white, yellow, or colorless; stoma anomocytic, guard cells 37.8–( x =46.0)–56.0 µm long. Costae persistent, ebeneous to atropurpureous, lustrous; carinae adaxial on all costae, golden, to 0.12–0.17 mm tall, apical cells neither distinct nor divergent; biseriate scales and catenate hairs sparse at abaxial costule-costa junctures, 3-celled capitate hairs scattered adaxially. Pinnae 3(–5), spreading radially from stipe apex in same plane as stipe, linear to lanceolate, apices conform, ending in a distinct and generally symmetrical apical segment, this basally lobate and with a rounded apex; central pinna 2.7–13.0 cm 30 long, 1.0–1.4 cm wide, bearing 17–35 pinnule pairs; second pinnae (the basal pinnae on most fronds) half to 2/3 length of central pinna, 4.5–7.4 cm long, 0.9–1.2 cm wide, bearing (1–)12–26 pinnule pairs; basal pinnae (third pinna pair, when present) no more than 1/4 length of second pinnae, 0.9–1.8 cm long, 0.6–0.9 cm wide, bearing 1–5 pinnule pairs; basal flabellate divisions attached between pinnae at stipe apex, more or less symmetrical, typically fertile when frond is fertile. Pinnules transverse to slightly ascending, short-stalked, articulate, oblong to ovate, somewhat recurved, drying revolute to rarely conduplicate, 4.2–7.0 mm long, 1.9–3.3 mm wide, length:width ratio 1.9– ( x =2.3)–2.8; acroscopic auricles minute, to 0.5 mm long; base obtuse basiscopically, truncate (acute) acroscopically; margins entire, somewhat revolute; apices round to crenate; stalks sub-medial to medial, persistent, irregularly conical after pinnule abscission, to 0.29 mm long. Veins free, anadromous, occasionally tending toward catadromous, obscure to occult (prominent in young tissue due to color), basiscopic veins sometimes bifurcate, ending in same sorus. Sori marginal, discrete, occasionally confluent on very fertile fronds, (1–)2–16 per pinnule. Pseudoindusia distinct, triangular, membranaceous, yellowish-tan to yellowish-green or white, often similar to lamina color, usually 1.1–1.6 times longer than wide, 0.53–0.86 mm long, 0.46–0.77 mm wide, margin lacinate. Sporangia subglobose, short stalked. Spores yellow to golden at maturity, tetrahedral-globose, 36.5–( x =39.96)–44.6 µm long, arachnoid-echinulate with dissected bases, echinae dense, up to 2 µm long, visible on some spores with brightfield microscopy, ornamentation arachnoid/stellate (with scanning electron microscopy), laesurae obscured by echinae. Chromosome number unknown. Figure 7. A–G. Additional Specimens Examined—COLOMBIA. Tolima: Honda, rock bluff, 200–250 m, 7 Jan 1918, Pennell 3685 (GH, K, NY, US); Piedras, common here on shaded banks pyroclastic rocks, 400 m, 29 Oct 1938, Haught 2402 (GH, UC, US); Límite entre Honda y Mariquita, crece en rocas o grandes piedras asoleadas, 500 m, 7 Jun 1960, Uribe 3466 (US); Cerca de Piedras en planitu calido ad pedeum montis guindio moro granateusis, en supibus arenaceis, 730 m, 1 Mar 1876, André 1980 (NY); An der Angostara an Rio Paëz, 800 m, 20 Feb 1883, Lehmann 2649 (K, US). 31 Distribution and Habitat—Adiantopsis ternata is known only from the department of Tolima in Colombia and the type locality in the Orinoco region of Venezuela. It has been found growing on rocky bluffs or on boulders, and among pyroclastic rocks or sandy soil on shaded banks, from 200–800 msl. Comments—The most distinguishing features of Adiantopsis ternata are the ternate fronds and triangular, lacinate pseudoindusia. It is most likely to be confused with A. trifurcata, which also typically has three main branches and triangular pseudoindusia. The two species differ in the pinnae being divergent in A. ternata but fasciculate in A. trifurcata, the margins of the pseudoindusia being very lacinate in A. ternata but slightly erose in A. trifurcata, and the ratio of the stipe to frond, with the stipe being more or less equal to the lamina length in A. ternata as opposed to typically much shorter than the lamina in A. trifurcata. The two taxa also differ markedly in pinnule size (A. ternata pinnules are larger). Adiantopsis timida Link-Pérez & Hickey, ined.—TYPE: BRAZIL. Rondônia: Rio Pacáas Novos, proximo a La Cachoeira, serra que fica em frente ao acampamento. Mata de terra firme, solo arenoso, 25 Mar 1978, Santos et al. 249 (holotype: NY!; isotypes MICH!, F!). Rhizome erect to decumbent; scales acicular, bicolorous, the central band shiny, blackish-brown, occupying 1/3 (to 1/2) of the width, margins brown to tan, with conform concolorous brown scales interspersed. Fronds erect, dimorphic, 7.0–40.3 cm long, shorter fronds sterile. Stipes atropurpureous, lustrous, about 62–70% of frond length in fertile fronds, approximately 50% of frond length in sterile fronds, 3.5–28.3 cm long; carinae adaxial, beginning at midpoint of stipe, occasionally in the proximal 1/3, golden, to 0.4 mm tall; scales golden-brown, sparse, only at base of stipe; catenate hairs sparse, more numerous at apex between pinnae attachments, amber. Laminae palmate, circular, geniculate, 3.5–16.0 cm long, 4.0–21.0 cm wide; spongiose, drying dark brownish-black, hydathodes marginal; scales absent; three-celled hairs sparse abaxially, rare adaxially, 32 septate, basal cell elongate and colorless, middle cell short, apical cell bulbous, amber to yellow; stoma anomocytic, guard cells 43.2–( x =52.8)–62.1 µm long. Costae persistent, atropurpureous, lustrous to matte; carinae adaxial on all costae, golden, to 0.3 mm tall, apical cells parallel to the costa axis; scales absent; catenate hairs few, at abaxial costulecosta junctures, amber. Pinnae (3–)5–7(–9), spreading radially from stipe apex, linear to somewhat fusiform, apices conform, ending in an asymmetrical mitten-shaped apical segment with an (acute) obtuse to rounded apex; central pinna (3–)4.5–15.0 cm long, 0.9–1.3(–1.6) cm wide, bearing (9–)15–38 pinnule pairs; basal pinnae 1.5–6.0 cm long, 0.5–1.1 cm wide, bearing 2–17 pinnule pairs; basal flabellate divisions attached between pinnae at stipe apex, typically fertile when frond is fertile. Pinnules transverse, shortstalked and geniculate, articulate, oblong, 3.9–7.0 mm long, 1.6–2.5 mm wide, length:width ratio 2.3–( x =2.7)–3.2, acroscopic auricles variable in size; base cuneate basiscopically, truncate to occasionally acute acroscopically; margins sinuate; apices round to round-sinuate; stalks medial to sub-medial, strongly reflexed abaxially, persistent, peg-like to aculeate after pinnule abscission, to 0.7 mm long. Veins free, anadromous, occult, ending submarginally in adaxial hydathodes. Sori marginal, discrete, (1–)6–14 per pinnule. Pseudoindusia distinct, lunate (sub-cuspidate), scarious to chartaceous, sometimes lustrous, pale brownish-gray distally, gradually approaching lamina color proximally, often black maculate, 1.4–1.9 times wider than long, 0.24–0.48 mm long, 0.42–0.72 mm wide, margin entire. Sporangia subglobose, long stalked, walls lustrous. Spores yellow to golden at maturity, tetrahedral-globose, 29.03–( x =34.98)– 41.85 µm long, arachnoid-echinulate with dissected bases, echinae densely distributed, delicate, to 1.35 µm long, appearing like pinpricks in proximal view with brightfield microscopy, ornamentation arachnoid in appearance with scanning electron microscopy, laesurae obscured by echinae. Chromosome number unknown. Figure 8. A–F. Representative Specimens Examined—BRAZIL. Rondônia: Município de Ariquesmes, Mineração Mibrasa, Setor Alto Candeias, km 128, Sudoeste de Ariquesmes, 10°35´S, 63°35´W, 18 May 1982, Teixeira et al. 592 (F, NY); Porto Velho to Cuiabá highway, vicinity of Santa Barbara, 15 km east of km 117, forest on terra firme, rocky laterite soil, growing on rocks, 12 Aug 1968, Prance and Ramos 6886 (F, NY). 33 COLOMBIA. Vaupés: Frequent in woodland, 250 m, 4 Aug 1959, Maguire et al. 44104 (MICH, NY, UB). Unknown department: Intendéncia Meta, Serránia de Macarena, alrededores de Río Guayabero, between central camp, La Macarena, and outcrop, La Meseta, on rocks in shade of trees, in quebrada (ravine), 1090 m, 11 Mar 1959, Barclay and Juajibioy 7122 (MO). GUYANA. Upper Takutu-Upper Essequibo Region: Kassikaityu R., 3-4 km N of landing at terminus of trail from Kuyuwini R., dense forest on brown sand, with Pouteria, Bertholettia & Oenocarpus, 1°50´N, 59°05´W, 240 m, 20 May 1997, Clarke 4745 (NY). VENEZUELA. Territorio Federal Amazonas: Departamento Río Negro, 1.5 km south of Cerro de La Neblina Base Camp which is on Río Mawarinuma, on ridge of hills, 0°50´N, 66°10´W, 190-340 m, 6 Mar 1984, Liesner and Funk 16429 (MO, NY); Departamento Atabapo, alto Río Orinoco, 15 km al SE de La Esmeralda, bosques medios en lomerio, estrato inferior dominado por Phenakospermum guianense Eichl., 3°08´N, 65°27´W, 180 m, 23 Feb 1990, Aymard and Delgado 8172 (MO, PORT); Atabapo, Salto Yureba, Cerro Yureba, lower Ventuari, 4°03´N, 66°01´W, 350 m, 15 Mar 1985, Liesner 18697 (MO); Departamento Atures, W side of valley of Rio Coro-Coro, 8km NNW of settlement of Yutaje, forested slope, 5°41´N, 66°08´W, 500–1000 m, 26 Feb 1987, Liesner and Holst 21412 (MO); Atures, forested areas and igneous outcrops along Río Coromoto, at Tobogán de la Selva, 35 km southeast of Puerto Ayacucho, vertical rock faces in forest shade, 5°22´N, 67°33´W, 150 m, 14 May 1980, Steyermark et al. 122466 (MO); Atures, fundo Santa Maria a 70 km al N de Pto Ayacucho via a Caicara–del Orinoco, 6°15´N, 67°40´W, 100 m, 14 Aug 1985, Melgueiro and Sanchez 317 (VEN); Atures, forest along Río Coromoto, above Tobogán de la Selva, 35 km SE or Puerto Ayacucho, 5°27´N, 67°33´W, 80 m, 7 Sep 1985, Steyermark et al. 131531 (MO); Atures, Laguna El Sillón y Caño Mariguaca, 78 km NE Puerto Ayacucho, 5°49´N, 66°50´W, 400 m, Oct 1989, Fernandez et al. 6433 (MO, PORT); Rio Orinoco just below mouth of Rio Atabapo, locally frequent on laja, 125 m, 3 Aug 1959, Wurdack, and Adderley 43734 (NY); Salto Coromoto (Balneario El Tobogán de la Selva) y aquas arriba del mismo, a +/- 11 kms del cruce desde la carretera Puerto Ayacucho-Samariapo, 120–150 m, 24 Mar 1979, Trujillo and Pulido 15166 (NY); Cerro de la Neblina, Loma de las Pinas (Pineapple Ridge), 1.5 km. S of Neblina Base Camp, 00°49´N, 66°09´W, 150–215 m, 27 34 Jan 1985, Nee 30569 (MO, NY, VEN). Estado Bolivar: Municipio Cedeño, Zona Minera El Guaniamo, bosque ombrófilo macrotérmico, helecho rupicola, 06°27´N, 65°54´W, 300 m, 24 May 1993, Diaz 1841 (GUYN, NY, PORT); Cedeño, vicinity of Panare village of Corozal, 6 km from Maniapure toward Caicara, semideciduous forest on granitic slopes and savanna, 06°55´N, 66°30´W, 90–400 m, 12 Oct 1985, Boom and Grillo 6345 (NY); Cedeño, vicinity of Panare village of Corozal, 6 km from Maniapure toward Caicara, semideciduous forest on granitic slopes and savanna, 06°55´N, 66°30´W, 400 m, 15 Apr 1986, Boom and Grillo 6508 (MO); Municipio Raul Leoni, bosque y arbustales sobre lateritas bauxíticas, cabeceras Río Túriba y Caño La Miel, 45 km al Este de Pijiguaos, 06°34´N, 66°23´W, 800 m, Jun 1989, Fernandez and Delgado 5692 (MO, PORT). Distribution and Habitat—This terrestrial species is widely distributed in a broad arc from Guyana westward through the departments of Bolivar and Amazonas in Venezuela, to Vaupes in Colombia, and southward into Rondônia in Brazil. It is found in forests or along forest edges, commonly among rocks and boulders, sometimes weathered volcanic rocks or igneous outcrops, and is often associated with slopes or riverbanks, from 80–1000 msl, usually below 400 msl. Comments—Among the most distinguishing features of Adiantopsis timida are the small mitten-shaped pinna apices and the reflexed pinnule stalks that enter the pinnule in a medial position. Often, the pinnules are revolute upon drying. This species is most likely to be confused with A. radiata but the pinnae apices in A. radiata end in basally pinnatifid or lobate ultimate segments that are distally elongate with an acute apex; in addition, A. timida differs in having much smaller pinnules than A. radiata, and their attachment to the costa is in a medial to sub-medial position in the pinnule in A. timida as opposed to sub-basal in A. radiata. Another useful character is the disposition of the stipe carinae: in A. timida they are typically present in the upper 1/2 or 2/3 of the stipe, whereas they are generally absent or restricted to the upper 1/10 of the stipe in A. radiata, at least in mature fronds. 35 Adiantopsis trifurcata (Baker) Link-Pérez & Hickey, comb. nov. Cheilanthes trifurcata Baker, Kew Bull. 144. 1901.—TYPE: BRAZIL. Goyaz, Glaziou 22637 (K!, K fragment NY!, B fragment NY!). Rhizome erect; scales acicular, bicolorous, the central band shiny, dark brown to ebeneous, occupying 1/2 (rarely to 2/3) of the width, margins pale golden or reddishgold, shiny, with intermingled conform, concolorous golden scales. Fronds caespitose, monomorphic, (3.0–)5.0–20.4 cm long (ca. 1/2 that size for Brazilian specimens). Stipes castaneous, atrocastaneous, to atropurpureous, lustrous, typically about 21–38% of frond length, as little as 10% in leaves with a single pinna, 0.7–7.0 cm long; carinae adaxial, beginning at base of stipe on younger fronds, frequently absent on older fronds, golden or reddish-gold, to 0.12 mm tall; stipe scales tan or golden, sparse, these generally restricted to stipe base and apex; septate-capitate hairs sparse. Laminae ternate to rarely palmate, trullate or trident-shaped, not to weakly geniculate, 4.5–14.0 cm long, 2.5–5.0 cm wide, young fronds often once-pinnate and linear; spongiose, drying grayish-green, olive, or brown, hydathodes marginal; scales absent; three-celled hairs diffuse abaxially, rare adaxially, septate, basal cell elongate and colorless, middle cell short, orange to red, apical cell narrow, colorless, amber to yellow; stoma anomocytic, guard cells 43.2– ( x =55.5)–70.2 µm long, guard cells of Bolivian specimens 43.2–( x =51.6)–61.4 µm long, guard cells of Brazilian specimens 48.6–( x =59.4)–70.2 µm long. Costae persistent, castaneous, atrocastaneous to atropurpureous, lustrous; carinae adaxial on all costae, reddish-gold to golden, shiny, to 0.2 mm tall, apical cells elongate and diverging at about 30° relative to the costa axis; scales amber, at costule-costa junctures; catenate hairs few, at abaxial costule-costa junctures, amber. Pinnae (1)3(–5), spreading radially from stipe apex, fasciculate, linear, gradually diminishing in width to a very small pinnatifid apex, apical segment trullate to mitten-shaped with an acute apex (sometimes spathulate); central pinna 5.0–14.0 cm long, 0.8–1.2 cm wide, bearing 17–48 pinnule pairs; second pinnae (the basal pinnae on most fronds) 0.7–5.7(–9.0) cm long, 0.3–0.8 cm wide, bearing 2–19(–36 in Bolivian specimens) pinnule pairs; basal flabellate divisions attached between pinnae at stipe apex, more or less symmetrical, typically fertile when frond is fertile. Pinnules transverse, short-stalked, articulate, oblong, 3.5–6.0 mm long, 1.6–2.5 36 mm wide, length:width ratio 1.9–( x =2.4)–2.7, acroscopic auricles to 0.6 mm long; base excavate to cuneate basiscopically, truncate to obtuse (acute) acroscopically; margins entire, somewhat “pebbly” in texture; apices round to lobulate; stalks basal to sub-medial, persistent, peg-like to aculeate after pinnule abscission, to 0.72 mm long, frequently with miniscule remnants of senescent lamina tissue attached. Veins free, anadromous, occult, ending submarginally in adaxial hydathodes. Sori marginal, discrete, (1–)2–11 per pinnule. Pseudoindusia distinct, triangular, scarious to membranaceous, yellowish-green or olivaceous, often black maculate, 0.9–( x =1.1)–1.5 times longer than wide, 0.43–0.84 mm long, 0.43–0.72 mm wide, margin erose to entire. Sporangia subglobose, distinctly short stalked. Spores yellow to golden at maturity, tetrahedral-globose, 39.15– ( x =47.32)–62.10 µm long, echinulate with dissected bases, echinae delicate, to 1.35 µm long, appearing like pinpricks in proximal view with brightfield microscopy so spores appear smooth, ornamentation revealed by scanning electron microscopy to be comprised of delicate rodlets that join together with single narrow rodlet extending outward to form echinae, laesurae visible. Chromosome number unknown. Figure 9. A–G. Additional Specimens Examined—BOLIVIA. Beni: Ballivian Province, 25 km from Yucumo on Yucumo-Quiquibey road, in the Pilón Lajas, growing on face of eastfacing cliff, 15°17´S, 67°4´W, 950 m, 18 Jul 1990, Fay and Fay 2764 (MO). Cochabamba: Prov. Carrasco, Parque Nacional Carrasco, al S del Campamento Petrolero Ichoa, bosque siempreverde, en pared rocosa de 6 m, 17°23´S, 64°24´W, 630 m, 20 Sep 1997, Acebey 691 (UC). La Paz: Prov. Abel Iturralde, Parque Nacional Madidi, campamento de guardaparques Sadiri, camino Sadiri–San José de Uchupiamonas, en la unión de los dos rios que forman el rio Yariapu, 14°10´S, 67°55´W, 620 m, 7 Aug 2004, Jimenez and Huaylla 2685 (NY); Lumupasa, 550 m, 12 Dec 1901, Williams 1328 (GH, NY, US). Santa Cruz: Prov. del Sara, Cerro del Amboró, Rio Yacuma, 1300 m, 21 Oct 1916, Steinbach 2990 (SI); Velasco, Parque Nacional Noel Kempff M., Campamento Las Gamas, bosque de colina, humedo con dosel superior de 30-40 m de altura, abundantes epifitas, suelo rocoso, rupestre, en caida de agua, poco comun, 14°48´S, 60°23´W, 900 m, 5 Apr 1993, Arroyo et al. 228 (USZ); Velasco, Campamento Las Torres, margen del Río Iténez (Guaporé), frontera con Mato Grosso, lado noreste del Serrania Huanchaca, 24 km 37 sur de Flor de Oro, ~50 km norte del Río Verde, bosque ralo de pendiente (3-10 m de altura), con fajas de roca abierto, temporalmente con flujo de agua, suelos arenosos, 13°39´S, 60°48´W, 200–400 m, 25 May 1991, Peña et al. 274 (USZ). Unknown department: Cerro Amboró, Ost. Cordillero, 1200 m, Oct 1907 [or 1909, unsure], Herzog 211 (UC, US); Subtrop. Region, Caranavi Tipuani, 1400 m, 13 Dec 1922, Buchtein 7029 (UC). BRAZIL. Goiás: Pirenopolis, Fazenda Bonsucesso, Centro oeste, na mata mesofitica, provincia do cerrado, 31 May 1997, Novelino 1354 (UB); Serra dos Pireneus, auf felsen, Dec 1892, Ule 3219 (HBG). Mato Grosso: Mun. Cuiabá, mata ciliar do alto da chapada dos Guimarães, perto da cach, Véu de Noiva, do Rio Coxipozinho, 15°30´S, 55°45´W, 21 Oct 1985, Pirani 1314 (K, NY); Chapada dos Guimarães, Gorge of Veu de Noiva, forest in rocky valley, growing on rocks, 17 Oct 1973, Prance et al. 19084 (NY); Chapada dos Guimarães, Veu de Noiva, nos blocos de pedra junto a corrego encachoeirado, interior do Canion, sombra, 16 Nov 1975, Hatschbach 37627 (UC); Serra Ricardo Franco, "Paraiso do Coatá”, próximo ao limite superior da mata, nas pedras da cascata, mas fora da água, 15°S, 60°W, 29 Jul 1974, Windisch 663 (GH); Chapada do Guimarães, Veu das Noivas, 500 m, 16 Feb 1988, Salino 388 (GH). Distribution and Habitat—Adiantopsis trifurcata is a lithophyte that occurs in Goiás and Mato Grosso in Brazil, and in Santa Cruz, Cochabamba, Beni, and La Paz in Bolivia. It is found growing on rocky walls and cliffs, or on boulders in forests, from 200–1400 msl. Comments—The most distinguishing feature of Adiantopsis trifurcata is the ternate fronds with fasciculate pinnae that resemble tridents (3-pronged spears) of classical mythology. The basal pinnae are typically half the length of the central pinna and the lamina 1 1/4–4 times longer than the stipe. This species is most likely to be confused with A. ternata, which also typically has three pinnae and triangular pseudoindusia; however, the pinnae in A. trifurcata are fasciculate whereas the pinnae in A. ternata are divergent, and the pseudoindusia margins in A. trifurcata are erose whereas those in A. ternata are lacinate. Another feature that distinguishes A. trifurcata from A. ternata is the stipe to frond ratio: 25–40% in A. trifurcata, but approximately 40–55% in A. ternata. 38 Adiantopsis trifurcata has been collected in the Andean foothills of western Bolivia and on the Precambric shield of eastern Bolivia along the border with Brazil, as well as in Goiás and Mato Grosso, Brazil. Most collections of this species are misidentified as A. monticola (Gardner) T. Moore, a small, pinnate species described from collections made in Goiás. There are some superficial resemblances, including the small size of the pinnae and pinnules, as well as habitat—the overall semblance is acknowledged by the handwritten note on the type specimens of A. trifurcata (apparently written by the collector of the type specimens, Glaziou) that the plant was a new species in Cheilanthes or Adiantopsis “near Monticolo Gardn.” However, the two taxa are quite distinct morphologically (and at the molecular level; unpublished data). Adiantopsis trifurcata has triangular pseudoindusia with erose margins whereas the pinnate A. monticola has lunate pseudoindusia with entire margins. In addition, the pinnule stalks in A. trifurcata are rather immersed within the lamina tissue and create a large abaxial bulge within the pinnule where the costa enters the lamina tissue; the pinnule stalks in A. monticola are very well-developed and quite distinct from the lamina tissue, with the stalks appearing “shield-like” on the surface of the lamina and with a texture very much like the carinae. There are several slight morphological differences between collections of Adiantopsis trifurcata from Bolivia and from Brazil, mostly in terms of size and apparent “robustness” rather than qualitative distinctions. In general, plants from Brazil are smaller and more delicate in appearance than those from Bolivia. Conversely, the guard cells of Brazilian collections are significantly larger than those of Bolivian specimens. Although future workers may choose to segregate these into different taxa, we see no compelling reason to do so based on our current understanding of the species. ACKNOWLEDGMENTS. The authors extend thanks to the curators of the following herbaria for loans or for providing access to material during visits: AAU, B, F, GH, GUYN, HB, HBG, HUT, K, MICH, MO, MU, NY, PORT, SI, SP, UB, UC, UPCB, US, USZ, and VEN. This research was supported by grants from W.S. Turrell Herbarium 39 Fund (Miami University) and Academic Challenge Research Grants (Department of Botany, Miami University). 40 LITERATURE CITED Anderson, L. E. 1954. Hoyer's solution as a rapid permanent mounting medium for bryophytes. Bryologist 57:242-244. Barrington, D. S., C. A. Paris, and T. A. Ranker. 1986. Systematic inferences from spore and stomate size in the ferns. American Fern Journal 76:149-159. Gastony, G. J., and D. R. Rollo. 1995. Phylogeny and generic circumscriptions of cheilanthoid ferns (Pteridaceae: Cheilanthoideae) inferred from rbcL nucleotide sequences. American Fern Journal 85:341-360. Gastony, G. J., and D. R. Rollo. 1998. Cheilanthoid ferns (Pteridaceae: Cheilanthoideae) in the southwestern United States and adjacent Mexico: A molecular phylogenetic reassessment of generic lines. Aliso 17:131-144. Prado, J., C. D. N. Rodrigues, A. Salatino, and M. L. F. Salatino. 2007. Phylogenetic relationships among Pteridaceae, including Brazilian species, inferred from rbcL sequences. Taxon 56:355-368. Rothfels, C. J., M. D. Windham, A. L. Grusz, G. J. Gastony, and K. M. Pryer. 2008. Toward a monophyletic Notholaena (Pteridaceae): resolving patterns of evolutionary convergence in xeric-adapted ferns. Taxon 57:712-724. Schuettpelz, E., and K. M. Pryer. 2007. Fern phylogeny inferred from 400 leptosporangiate species and three plastid genes. Taxon 56:1037-1050. Tryon, R. M., and A. F. Tryon. 1973. Geography, spores, and evolutionary relations in the cheilanthoid ferns. Pages 145-153 in The phylogeny and classification of ferns (A. C. Jermy, J. A. Crabbe, and B. A. Thomas, eds.). Academic Press, New York. Tryon, R. M., and A. F. Tryon. 1982. Ferns and allied plants with special reference to tropical America. Springer-Verlag, New York. Wagner, W. H., Jr., F. S. Wagner, and W. C. Taylor. 1986. Detecting abortive spores in herbarium specimens of sterile hybrids. American Fern Journal 76:129-140. Walker, T. G. 1973. Evidence from cytology in the classification of ferns. Pages 91-110 in The phylogeny and classification of ferns (A. C. Jermy, ed.) Academic Press. Windham, M. D., L. Huiet, E. Schuettpelz, A. L. Grusz, C. Rothfels, J. Beck, G. Yatskievych, and K. M. Pryer. 2009. Using plastid and nuclear DNA sequences to 41 redraw generic boundaries and demystify species complexes in cheilanthoid ferns. American Fern Journal 99:128-132. 42 APPENDIX 1. List of quantitative and qualitative morphological characters used to delimit species and construct species descriptions for palmate members of Adiantopsis. Measurement units of quantitative characters are presented in parentheses. 1. Rhizome aspect [erect, ascending, decumbent]. 2. Rhizome scales location. 3. Rhizome scales description. 4. Rhizome scales: dark central stripe is what proportion of total width of bicolorous scale. 5. Stipe length (cm). 6. Stipe color. 7. Stipe, location of carinae. 8. Stipe, color of carinae. 9. Stipe, maximum width of carinae (mm). 10. Stipe indument. 11. Stipe sulcus, continuous or discontinuous into costae. 12. Costa color. 13. Costa, location of carinae. 14. Costa, color of carinae. 15. Costa, maximum width of carinae (mm). 16. Costa indument. 17. Frond, overall shape. 18. Frond, aspect where lamina joins stipe [geniculate, weakly geniculate, in same plane]. 19. Frond length, from rhizome to apex of central pinna (cm). 20. Ratio of stipe:frond. 21. Dimorphism. 22. Lamina dissection. 23. Pinnae shape, overall. 24. Lamina length (cm). 25. Number of pinnae [main branches]. 26. Basal flabellate divisions, fertile or not. 27. Lamina trichomes, adaxial surface. 28. Lamina trichomes, abaxial surface. 29. Abaxial laminar trichomes, color of apical cell. 30. Appearance of veins, adaxial [obscure, occult, prominent]. 31. Appearance of veins, abaxial [obscure, occult, prominent]. 32. Lamina texture. 33. Central pinna length (cm). 34. Central pinna width at widest point (cm). Central pinna, length:width ratio. 35. Central pinna, position of widest point [middle, throughout, basal third]. 36. Central pinna, number of pinnule pairs. 37. Second pinna [next to central pinna], length (cm). 38. Second pinna, width at widest point (cm). Second pinna, length:width ratio. 39. Second pinna, number of pinnule pairs. 40. Basal pinna [farthest from central pinna], length (cm). 41. Basal pinna descriptor [3rd, 4th, 5th, 6th, nth pinna—as counted from the central pinna, which is “pinna 1”]. 42. Basal pinna, width at widest point (cm). 43. Basal pinna, length:width ratio. 44. Basal pinna, number of pinnule pairs. 45. Apical portion of pinnae, overall shape. 46. Pinnules [also known as “ultimate divisions” or “ultimate segments”], aspect of costule/pinnule junction. 47. Pinnules, position of segment attachment [bottom corner of segment, medial, or semimedial]. 48. Pinnule, articulation, description. 49. Pinnule, articulation, length of stalk left after pinnule abscission. 50. Pinnule, overall shape. 51. Pinnule, shape of basal portion, acroscopic side. 52. Pinnule, shape of basal portion, basiscopic side. 53. Pinnule, shape of apical portion. 54. Pinnule, margin. 55. Fertile pinnules [measurements taken from typical representatives from mid-portion of (typically) the central pinna of a fertile frond; five measurements taken from each specimen examined for collection of morphological data], length (cm). 56. Fertile pinnules, width [not including auricle, if present] (cm). 57. Fertile pinnules, length:width ratio. 58. Fertile pinnule, width with auricle (cm). 59. Fertile pinnules, number of sori. 60. Fertile pinnules, approximate area (cm2). 61. Sori [discrete or continuous]. 62. Pseudoindusia, shape. 63. Pseudoindusia, margins. 64. Pseudoindusia, texture. 65. Pseudoindusia, color. 66. Pseudoindusia, width (mm). 67. Pseudoindusia, length (mm). 68. Pseudoindusia, length:width ratio. 69. Sporangia, shape. 70. Annulus cells, number per sporangium. 71. Annulus cell height (µm). 72. Spore, ornamentation. 73. Spore, length (µm). 74. Spore, color. 75. Spore, shape. 76. Guard cell length (µm). 77. Carinae, orientation of cells. 43 TABLE 1. Diagnostic features that can be used in distinguishing palmate members of Adiantopsis. Taxa Typical number of pinnae Position of stalk in pinnule Typical size of pinnules (excluding auricle, if present) Pseudoindusia shape (margin) Spore ornamentation A. crinoidea 5–15 medial 2.0–5.0 mm long; 1.0–2.0 mm wide quadrangular to lunate (erose to very erose) echinulate A. dactylifera 5–7 medial to sub-medial 3.0–8.0 mm long; 1.1–2.3 mm wide lunate to cuspidate (entire to slightly erose) echinate with rodlets in between A. radiata (3–)5–7(–9) sub-basal to basal 7.0–13.5 mm long; 2.0–3.5 mm wide lunate (entire) echinate A. ternata 3(–5) sub-medial to medial 4.2–7.0 mm long; 1.9–3.3 mm wide triangular (lacinate) arachnoidechinulate A. timida (3–)5–7(–9) medial to sub-medial 3.9–7.0 mm long; 1.6–2.5 mm wide lunate to subcuspidate (entire) arachnoidechinulate A. trifurcata (1)3(–5) basal to sub-medial 3.5–6.0 mm long; 1.6–2.5 mm wide triangular (erose to entire) echinulate 44 TABLE 2. Guard cell lengths of palmate species of Adiantopsis. Taxa with means that are significantly different from each other (P < 0.001) are indicated by different letters. Sample size Mean (µm) Standard deviation (µm) Min (µm) Max (µm) A. crinoidea a 100 51.50 4.00 40.50 62.10 A. dactyliferab 100 39.52 3.00 32.40 48.60 A. radiatac 75 46.44 3.29 36.45 54.0 A. ternatac 100 46.00 3.99 37.80 56.03 A. timidaa 100 52.76 3.73 43.20 62.10 A. trifurcatad 150 55.52 6.42 43.20 70.20 Taxon 45 TABLE 3. Spore lengths of palmate species of Adiantopsis. Taxa with means that are significantly different from each other (P < 0.05) are indicated by different letters. Sample size Mean (µm) Standard deviation (µm) Min (µm) Max (µm) A. crinoidea a 25 51.41 2.20 47.25 55.35 A. dactyliferab 22 30.84 1.75 27.68 33.75 A. radiatac 63 32.91 2.39 27.00 37.80 A. ternatad 25 39.96 2.27 36.45 44.55 A. timidae 133 34.98 2.52 29.03 41.85 A. trifurcataf 50 47.32 4.92 39.15 62.10 Taxon 46 FIG. 1. Distribution of the palmate species of Adiantopsis. 47 FIG. 2. Silhouettes of pinna apices from palmate species of Adiantopsis shown to same scale. A. Adiantopsis radiata (Ventura 10238, MICH). B. Adiantopsis dactylifera (Vásquez et al. 25968, MO). C. Adiantopsis timida (Fernandez and Delgado 5692 (MO). D. Adiantopsis trifurcata (Buchtein 7029, UC). E. Adiantopsis ternata (André 1980, NY). F. Adiantopsis crinoidea (Davidse and Huber 15283, NY). 48 FIG. 3. Scanning electron micrographs of spores from palmate Adiantopsis species shown to same scale. A. Adiantopsis radiata showing echinate ornamentation (Link-Pérez and Cabral 183, MU). B. Adiantopsis dactylifera showing echinate ornamentation 49 interspersed with rodlets jutting up between echinae on perispore (Campos 4834, isotype MO). C. Adiantopsis timida showing arachnoid-echinulate ornamentation (Fernandez et al. 6433, MO). D. Adiantopsis trifurcata showing echinulate ornamentation (Hatschbach 37627, UC). E. Adiantopsis ternata showing arachnoid-echinulate ornamentation (Haught 2402, US). F. Adiantopsis crinoidea showing echinulate ornamentation (Steyermark et al. 131214, MO). 50 FIG. 4. Adiantopsis crinoidea. A. Habit showing dimorphic fronds. B. Spore with echinulate ornamentation. C. Detail of costa showing adaxial carina (right). D. Pinnule showing chaffy-looking, imbricate pseudoindusia. E. Line drawing of a cleared pinnule showing venation, medial placement of stalk, and quadrangular to lunate pseudoindusia. F. Pinna apex. A Gröger and Barcroft 265 (holotype MO); B Steyermark et al. 131214 (MO); C,F Davidse and Huber 15283 (NY); D–E Fernandez and Delgado 5860 (MO). 51 FIG. 5. Adiantopsis dactylifera. A. Habit. B. Spore with echinate ornamentation with dissected bases interspersed with rodlets jutting up between echinae on perispore. C. Detail of costa showing digitate apical cells of adaxial carina (left). D. Pinnule. E. Line drawing of a cleared pinnule showing venation, medial to sub-medial placement of stalk, and pseudoindusia. F. Detail of stipe apex from abaxial (white arrow) showing pinnae spreading radially; note the digitate apical cells of carinae. G. Pinna apex. A Campos 4834 (holotype NY); B Campos 4834 (isotype MO); C van der Werff et al. 16380 (NY); D,F Coronado and Cerrón 44 (MICH); E,G Vásquez et al. 25968 (MO). 52 FIG. 6. Adiantopsis radiata. A. Habit. B. Spore with echinate ornamentation. C. Detail of pinnule showing lunate pseudoindusia with entire margins. D. Line drawing of a cleared pinnule showing venation, basal placement of stalk, and lunate pseudoindusia. E. Detail of costa showing adaxial carina (right). F. Pinna apex. A,E Hernandez and Gonzalez 1670 (MO); B Link-Pérez and Cabral 183 (MU); C Link-Pérez et al. 355 (MU); D Link-Pérez et al. 327 (MU); F Ventura 10238 (MICH). 53 FIG. 7. Adiantopsis ternata. A–B. Habit showing mostly ternate fronds. C. Spore with arachnoid-echinulate ornamentation. D. Detail of costa showing adaxial carina (right). E. Line drawing of a cleared pinnule showing venation, sub-medial to medial placement of stalk, and triangular pseudoindusia. F. Detail of pinnule (line drawing and photo) showing triangular pseudoindusia with lacinate margins. G. Pinna apex. A–C, F Haught 2402 (US); D–E Pennell 3685 (US); G André 1980 (NY). 54 FIG. 8. Adiantopsis timida. A. Habit showing dimorphic fronds. B. Spore with arachnoid-echinulate ornamentation. C. Detail of costa showing adaxial carina (right). D. Line drawing of a cleared pinnule showing venation, sub-medial placement of stalk, and lunate pseudoindusia. E. Detail of two pinnae showing revolute pinnules and reflexed pinnule stalks. F. Pinna apex. A Aymard and Delgado 8172 (MO); B Fernandez et al. 6433 (MO); C–E Boom and Grillo 6508 (NY); F Fernandez and Delgado 5692 (MO). 55 FIG. 9. Adiantopsis trifurcata. A. Habit showing mostly ternate fronds. B. Spore with echinulate ornamentation. C. Detail of costa showing adaxial carina (right). D. Line drawing of a cleared pinnule showing venation, sub-medial placement of stalk, and generally triangular pseudoindusia, some with erose margins. E. Pinnule showing triangular pseudoindusia with entire margins. F. Pinna apex. A Williams 1328 (US); B,D Hatschbach 37627 (UC); C Steinbach 2990 (SI); E Herzog 211(UC); F Buchtein 7029 (UC). 56 3 Toward a redefinition of Adiantopsis Fée (Pteridaceae): Systematics, diversification, and biogeography2 We present here the first multi-gene phylogeny of taxa assigned to the fern genus Adiantopsis, one of about twenty genera in the large group of cheilanthoid ferns. Our goals were to clarify natural relationships, examine generic boundaries, and interpret character evolution within a phylogenetic framework. Phylogenetic analysis of combined plastid rbcL and atpA DNA sequences has led to an expansion of generic boundaries in Adiantopsis, a recircumscription that is clearly defensible with morphological data. Biogeographic analyses suggest an origin for the genus in South America, possibly the Cerrado and associated dry areas of Brazil, with a minimum of three migrations into the Caribbean. KEYWORDS: Adiantopsis, atpA, cheilanthoids, ferns, molecular phylogenetics, rbcL. INTRODUCTION Of the five major clades in the ecologically diverse Pteridaceae, the cheilanthoid ferns, comprised of approximately 20 genera and 400 species, are most obviously characterized by adaptations to xeric habitats. The high degree of adaptive morphology to such habitats has made the taxonomic circumscription of cheilanthoids problematic 2 This chapter is formatted for submission to Taxon. 57 (Gastony and Rollo, 1995, 1998; Prado et al., 2007) and has led to a proliferation of polyphyletic genera, including Doryopteris J. Sm., Pellaea Link, and the very large Cheilanthes Sw. It is necessary to revise generic circumscriptions to resolve a natural classification of the cheilanthoids, an essential task in understanding the evolutionary history of this large group of ferns and the biological and ecological drivers of its diversification. The present work contributes to the reorganization of the cheilanthoids by resolving the generic boundaries of Adiantopsis Fée, a small genus in the hemionitid subclade within the cheilanthoids (Rothfels et al., 2008; Windham et al., 2009). Classically, Adiantopsis has been characterized by the combination of echinate spores, golden or golden-red, paired ridges, carinae, on the upper side of leaf axes and sometimes on stipe axes, and well-differentiated pseudoindusia—a combination of characters that sets Adiantopsis apart from other cheilanthoid genera. Species with cristate rather than echinate spores, such as A. dichotoma (Cav.) Moore and A. regularis Moore, are generally placed in Cheilanthes. Adiantopsis is predominantly a genus of tropical America (Tryon et al., 1990), although the ranges of some species extend into the northern portions of Uruguay (Zuloaga et al., 2008). Some taxonomists have postulated a relationship between neotropical Adiantopsis and certain African and Madagascan cheilanthoids (Christensen, 1932; Moran and Smith, 2001); however, persuasive evidence, either molecular or morphological, to support the inclusion of these paleotropical taxa in Adiantopsis is lacking. Tryon and Tryon (1982) produced the most recent review of Adiantopsis encompassing its entire geographic range. They employed conservative generic and species concepts and recognized only seven species, all neotropical in distribution. Their generic circumscription is established by the combination of echinate spores, adaxially bicarinate laminar axes (and stipes in some species), and asymmetrical ultimate divisions. They excluded the South American A. dichotoma and A. regularis based on their possession of cristate rather than echinate spores, and they subsumed the Caribbean A. asplenioides Maxon and A. rupicola Maxon into A. paupercula (Kunze) Fée and A. reesii (Jenman) C. Chr., respectively. A recent revision of the Caribbean taxa recognizes nine 58 distinct species in that region alone (Barker and Hickey, 2006), including the Cuban endemics A. asplenioides and A. rupicola previously excluded by Tryon and Tryon (1982). This Caribbean revision also described three new species: A. parvisegmenta M.S. Barker & Hickey, A. vincentii M.S. Barker & Hickey, and A. pentagona M.S. Barker & Hickey. One of the most fascinating aspects of Adiantopsis is leaf morphology because its species display palmate, pedate, and pinnate architectures (Fig. 1). The palmate architecture is arguably the most striking of the three and is well illustrated by the widespread A. radiata (L.) Fée, which is found throughout the Caribbean and Central America, and extends southward into northeastern Argentina. Adiantopsis presents a superb opportunity to explore the relationship between leaf morphology and phylogeny within one small genus. Between the two treatments provided by Tryon and Tryon (1982) and Barker and Hickey (2006), and with the discovery of a sterile hybrid with pedate laminar morphology (Hickey et al., 2003), approximately 13 taxa of Adiantopsis have been generally recognized, although various researchers working on regional treatments occasionally have accepted other, lesser-known species: for example, A. seemannii (Hook.) Maxon in Mexico (Mickel and Smith, 2004) and A. perfasciculata Sehnem in Brazil (Figueiredo and Salino, 2005). The overall aims of this study are to resolve the generic boundary and the phylogenetic position of Adiantopsis within the cheilanthoids, and to enumerate its taxa and resolve their relationships to each other. A phylogeny, based on DNA sequences from two plastid genes—rbcL and atpA—is used as the framework to explore the history of character evolution in the genus, particularly the origin and evolution of the diverse laminar morphologies (pinnate, pedate, and palmate), and to investigate its biogeographic origins, specifically the high level of species diversity in the Caribbean. MATERIALS AND METHODS Marker selection. — The study used sequences from two plastid loci: rbcL and atpA. These genes were selected because both have been used successfully to reconstruct phylogenies in other fern lineages (Gastony and Rollo, 1996; Gastony and Rollo, 1998; 59 Nakazato and Gastony, 2003; Pryer et al., 2004; Schuettpelz et al., 2006; Schuettpelz and Pryer, 2007; Tsutsumi and Kato, 2005; Wolf, 1995; Wolf et al., 1994). The rbcL gene evolves relatively slowly and is good for inferring phylogeny at the generic and familial level in ferns (Hasebe et al., 1995), as well as at lower taxonomic levels (Gastony and Rollo, 1998; Haufler et al., 1996; Wolf et al., 1994). The atpA gene has recently emerged as an exceptionally useful marker in fern phylogenies due to its greater number of variable characters relative to other commonly used markers (Schuettpelz et al., 2006; Schuettpelz and Pryer, 2007). Taxon sampling. — The study included all available Adiantopsis species, which included 11 of the 13 generally recognized taxa and 5 taxa that are lesser-known or commonly treated as Cheilanthes, as well as 16 taxa hypothesized to belong in the genus based on morphology (including 14 previously undescribed species). Recent molecular analyses have placed Adiantopsis in the hemionitidoid clade of the cheilanthoids (Rothfels et al., 2008), so additional taxa were selected to provide a broad sampling from within the hemionitidoid genera, which includes Pentagramma Yatskievych, Windham, & E. Wollenweber, Aleuritopteris Feé, Aspidotis Nuttall, Hemionitis L., and Doryopteris. Voucher specimen data are provided in the Appendix. One to three individuals from each taxon were sampled, except for A. radiata and A. chlorophylla (Sw.) Fée, which were sampled more heavily due to their widespread distribution and the possibility that they represented multiple taxa or are not monophyletic. Before tissue was selected for DNA extraction, samples were examined under a dissecting scope to ensure that they were free of potential contaminants such as epiphyllous bryophytes. DNA extraction, amplification, and sequencing. — Total genomic DNA was extracted from silica gel-dried material (Chase and Hills, 1991) and herbarium specimens (Savolainen et al., 1995) using the DNeasy Plant Mini Kit (Qiagen, Valencia, California, USA) following the manufacturer’s protocol, including the optional 5 minute centrifugation during protein precipitation. PCR amplification was performed in 15 µL reactions using either 0.6 units Qiagen Taq DNA Polymerase (5 u/µL) with 1.5 µL 10X CoralLoad PCR Buffer, or 0.4 units Promega GoTaq® DNA polymerase (5u/µL) with 3 µL 5X Green GoTaq® Flexi Buffer (Promega Corporation, Madison, WI, USA). PCRs also contained 0.4 µM each 60 of forward and reverse primers, 200 µM each dNTP, 1.6–2.5 mM MgCl2, 0.3 µL of 0.1µg/µL BSA, 1 µL of total genomic DNA (full concentration, or diluted to 0.1x, 0.02x, or 0.01x), and molecular grade water to volume. Primers for rbcL are from Pryer & al. (2004; 2001) and Gastony & Rollo (1995), or were modified from them to be specific for Adiantopsis. Primers for atpA are from Schuettpelz & al. (2006) or were modified from them. A few primers were designed de novo. PCR protocol for both markers is as described by Pryer & al. (2001) except that the elongation step was extended from 90 seconds to 2 minutes to accommodate the length of the template. PCR products were electrophoresed on 1% agarose gels in sodium borate buffer and visualized with ethidium bromide or SafeView Nucleic Acid Stain (NBS Biologicals Ltd., Cambridgeshire, UK) to verify successful amplification. Excess nucleotides and primers were removed using Wizard SV Genomic DNA Purification System (Promega, Madison, WI) according to the manufacturer’s protocol. Purified PCR product was labeled with BigDye v3.1 (Applied Biosystems, Foster City, CA) sequencing chemistry. After fluorescent-labeling, fragments were purified using sodium acetate-EDTA ethanol precipitation. Automated sequencing of purified fluorescent-labeled fragments was performed on an ABI Prism 3130xl or 3730 DNA Analyzer (Applied Biosystems) at the Miami University Center for Bioinformatics and Functional Genomics. DNA sequence data were assembled and manually edited in Sequencher 4.8 (Gene Codes Corporation, Ann Arbor, Michigan, USA), and aligned in Se-Al v2.0a11 (Rambaut, 2002). Ambiguously aligned areas, confined to 3’ end of atpA, were excluded prior to phylogenetic analyses. Indels (“–”) were treated as missing data. Some additional sequences were acquired from GenBank (Benson et al., 2005). Phylogenetic analysis. — The DNA data matrix was used to construct a maximum parsimony phylogeny with PAUP* 4.0 (Swofford, 2000). The rbcL and atpA datasets were analyzed individually, as well as combined, using heuristic searches with 10,000 random addition sequences and TBR branch swapping. Support was estimated with 1,000 bootstrap replicates (Felsenstein, 2004), with ten random addition sequences per replicate using the heuristic search criterion. 61 Bayesian analyses were conducted with the software program Mr. Bayes v3.1.2 (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003). Models of nucleotide substitution for each gene were selected using ModelTest (Posada, 2006; Posada and Crandall, 1998) under the Akaike Information Criterion (AIC; Posada and Buckley, 2004). The GTR+I+G model was selected for rbcL, and the TrN+I+G model was selected for atpA. Two independent runs (six chains each) of 2 million generations of Markov Chain Monte Carlo simulation were used to estimate the posterior probability distribution of trees; trees were sampled every 100 generations. Log likelihood scores were plotted against generation to determine stationarity (48,000 generations), and all trees before this point (480) were discarded as burn-in. A 50% majority rule consensus tree was constructed with the remaining trees. Morphological and biogeographic character mapping. — Character states and geographic distributions were obtained from herbarium specimens and published sources. Spores were examined with scanning electron microscopy to determine ornamentation and shape. Selected morphological traits such as leaf architecture, spore ornamentation, and carinae presence or absence were optimized onto the resulting phylogenetic tree using MacClade 4.0 (Maddison and Maddison, 2000) in order to understand direction of character evolution within Adiantopsis. Taxa were assigned to phytogeographic regions modified from those of Gentry (1982), who subdivided the neotropics into nine regions based on distributional and ecological data compiled from reviews of 8,117 monographed flowering plant species; Gentry’s phytogeographic regions include Mexico and Central America, West Indies (“Caribbean” in the present research), Northern Colombia, Northern Andes, Southern Andes (not applicable), Amazonia, Guyana Highlands, Cerrado and associated dry areas, and Coastal Brazil. Geographic distributions and phytogeographic data were optimized onto the phylogeny to test hypotheses regarding geographic origin and distribution, and the influence of ecology on diversification. RESULTS DNA sequences and alignments. — The data matrix contained 44 taxa, with 3169 characters in the concatenated data set (Table 1). The study produced 90 newly obtained sequences representing 32 taxa, of which 31 are from the ingroup (Appendix). 62 Phylogeny. — Maximum parsimony analyses on both single-gene data sets resolved a clade containing all included Adiantopsis species and a handful of taxa currently treated as Cheilanthes (data not shown). The topologies of the single-gene majority rule trees were largely congruent, except for the placement of Adiantopsis monticola (Gardner) T. Moore and A. paupercula (Kunze) Fée; however, those conflicting placements had weak or no support in the rbcL tree (data not shown). All further analyses, therefore, used a concatenated data set including both plastid markers. Maximum parsimony and Bayesian analyses of the combined data resulted in a well-supported clade (bootstrap=96, posterior probability=1.00) with Cheilanthes flexuosa Kunze and Adiantopsis regularis Moore as early diverging lineages (Fig. 2). Cheilanthes tweediana Hook. and A. dichotoma (Cav.) Moore are resolved as embedded within Adiantopsis and are the next early diverging lineages, following C. flexuosa and A. regularis. The rest of the taxa are in a well-supported clade (bootstrap=98, posterior probability=1.00) with A. paupercula sister to two major clades (designated “Clade 1” and Clade 2” in Figure 2) that show only moderate to low support. “Clade 1” (no bootstrap support, posterior probability=0.54) has the pinnate Brazilian species A. monticola sister to a well-supported clade (posterior probability=0.99) of palmate species, in which the pinnate A. lindigii (Mett.) Prantl is embedded; see discussion below. “Clade 2” (no bootstrap support, posterior probability=0.62) contains three wellsupported sub-clades: one subclade (bootstrap=84, posterior probability=1.00) contains pinnate species from Colombia and Central America, including A. propinqua (Mett.) Prantl and A. seemannii, respectively; another subclade (bootstrap=87, posterior probability=0.99) contains pinnate taxa with a strictly South American distribution and includes A. chlorophylla; the third subclade (bootstrap=89, posterior probability=1.00) includes pinnate and pedate taxa including an undescribed Brazilian taxon sister to a well-supported clade (bootstrap=94, posterior probability=1.00) containing 6 species endemic to the Caribbean. The analysis also supports the recognition of 12 previously undescribed, but morphologically distinct, taxa, two of which have already been given provisional names (A. dactylifera and A. timida, in preparation). Biogeography. — The most parsimonious reconstruction of geographic distributions suggests an origin of Adiantopsis in South America, with multiple 63 migrations into the Caribbean (Fig. 3). The earliest migration into the Caribbean revealed by the phylogeny is the one associated with A. paupercula, which is found on the islands of Jamaica, Cuba, and Puerto Rico. A second migration appears to be responsible for a clade of Caribbean endemics that includes A. pedata, A. pentagona, A. rupicola, A. reesii, A. parvisegmenta, and A. vincentii. Only one species—A. radiata—is truly widespread, extending from the Caribbean, Mexico and Central America, to Paraguay and northern Argentina. Optimizing phytogeographic regions onto the Bayesian phylogeny indicates that the genus likely originated in the Cerrado, an extensive tropical woodland-savanna and associated dry areas of Brazil, with several subclades radiating into other areas (Fig. 4). Character evolution. — The most parsimonious reconstructions of spore ornamentation and presence/absence of paired adaxial carinae reveals that these two characters, the combination of which has traditionally been used to circumscribe Adiantopsis, do not circumscribe the same clade (Fig. 5). Carinae appear earlier, demarking a larger clade which includes the cristate-spored A. dichotoma and Cheilanthes tweediana. Optimizing laminar architecture onto the molecular phylogeny reveals a pattern that strongly indicates that pinnate architecture is pleisiomorphic, and that the palmate architecture is a synapomorphy for a clade of eight palmate species (Fig. 6). The presence of the pinnate species Adiantopsis lindigii within this clade is unexpected (explored in the Discussion under “Remaining questions”). The pedate architecture occurs in two distinct clades and, based on these data, has multiple independent origins. Branch lengths. — The phylogram in Figure 7, one of the 5040 equally most parsimonious trees resulting from the maximum parsimony analyses, indicates that the Caribbean taxa found in the “reesii” subclade, which contains Adiantopsis reesii, A. pedata (Hook.) T. Moore, A. pentagona, A. rupicola, A. parvisegmenta, and A. vincentii, are poorly resolved with short branch lengths due to very little DNA sequence divergence in the two genes used in this study. Poor resolution also occurs in the subclade containing A. chlorophylla and three undescribed taxa denoted as Species novum 7, 8, and 9 (Fig. 7). In contrast, branch lengths for most other subclades are relatively longer, 64 indicating greater accumulation of genetic change and providing better resolution for those parts of the phylogeny. DISCUSSION Toward a redefinition of Adiantopsis. — The taxonomic circumscription of cheilanthoids has long been a vexing problem because of their apparently high degree of convergence for xeric adaptations (Gastony and Rollo, 1995, 1998). These difficulties interpreting homoplasious morphology have led to an abundance of polyphyletic genera (see discussions in Prado et al., 2007; Schuettpelz et al., 2007). As stated succinctly by Gastony and Rollo (1998), “we do not have good generic characters in cheilanthoids because we do not have good genera in cheilanthoids.” The morphological and molecular analyses presented in the current paper have attempted to remedy part of the cheilanthoid problem by resolving the generic boundaries of Adiantopsis and by identifying a suite of morphological characters that can delimit it. Our results have revealed previously unrecognized diversity within Adiantopsis and support an expanded circumscription of the genus. The classical circumscription of Adiantopsis is based on the combination of golden or golden-red paired carinae along the axes and echinate spores; however, this circumscription is not satisfactory because these two characters circumscribe different clades. While a circumscription based upon possession of echinate spores includes the type of Adiantopsis, A. paupercula (Kunze) Feé, it excludes two carinate and cristate-spored taxa, A. dichotoma and Cheilanthes tweediana, that appear in a relatively basal position within a greater Adiantopsis clade resolved in our analyses (Fig. 2). If we were to maintain the more strict morphological definition of Adiantopsis based on possession of echinate spores, it would be necessary to erect three new genera, two of them monotypic, in order to effect a monophyletic classification: one for Cheilanthes flexuosa (the type of Cheilanthes, C. micropteris Sw., is in a distant clade (Fig. 2; also see Rothfels et al., 2008; Schuettpelz et al., 2007)), one for A. regularis, and a third for C. tweediana and A. dichotoma. Instead, we propose expanding the generic limits of Adiantopsis to include the non-carinate Cheilanthes flexuosa and A. regularis, as well as the carinate but cristatespored C. tweediana and A. dichotoma. This larger clade can be circumscribed by the 65 following morphological characters: large, reddish, pluricellular hairs or carinae on axes; distinct, scarious pseudoindusia that cover one to occasionally two discrete sori; and compound leaves with small, asymmetrical ultimate segments, at least some of which are stalked. With generic limits extended and the transfer into Adiantopsis of several species (C. flexuosa, C. tweediana, and C. trifurcata) currently placed in the polyphyletic Cheilanthes, Adiantopsis becomes monophyletic and clearly defensible. The expanded Adiantopsis proposed here is distinct from the other genera in the hemionitidoid subclade of the cheilanthoids: the compound leaves are essentially glabrous and lack any type of farina, which is present in most Notholaena and Pentagramma species; its sori are discrete and marginal, protected by distinct, scarious, flap-like pseudoindusia, whereas the sori are without indusia and occur along the veins in Hemionitis and Pentagramma or are protected by the recurved margins in Pellaea and Doryopteris (Kirkpatrick, 2007; Mickel and Smith, 2004; Rothfels et al., 2008; Tryon and Tryon, 1982). The laminae of Adiantopsis are pinnate to quadripinnate with small, asymmetrical ultimate segments, whereas the laminae of Hemionitis and Doryopteris are usually undivided, with shallow to deep lobes in Hemionitis, and are cordate, sagittate, or hastate, and deeply pinnatifid or bipinnatifid in Doryopteris (Mickel and Smith, 2004; Rothfels et al., 2008; Tryon and Tryon, 1982). The laminae of Aspidotis are adaxially lustrous and often striate (Mickel and Smith, 2004). Tryon and Tryon (1973) suggested that cristate (and rugulose) spores were pleisiomorphic in cheilanthoids, and our analysis is consistent with that hypothesis. Early diverging lineages in Adiantopsis, as circumscribed here, have cristate spores, with apomorphic echinate spores arising once and expressed in a majority of Adiantopsis species. Biogeography. — Based on our interpretation of the phylogeny we hypothesize a South American origin for Adiantopsis (Fig. 3), specifically within the Cerrado and associated dry areas (Fig. 4), which are home to several early diverging lineages. The subsequent radiation from the dry forest or wooded savanna vegetation of the Cerrado (Gentry, 1982) into distinct phytogeographic regions (Fig. 4), such as the Guianas, Northern Venezuela/Colombia and Mexico/Central America, or the Caribbean, rather faithfully follows the “generalized South American-Caribbean track” proposed by Rosen 66 (Fig. 8; 1975). In that study, Rosen (1975) reviewed Caribbean biogeography by examining the individual distributions of species or monophyletic groups of species of various organisms (primarily animals), looking for commonality in the distribution patterns. These “coincident distributions” he called “generalized tracks,” since they reflect the migration or dispersal of plants and animals from one region to another (Rosen, 1975). The present-day diversity of Adiantopsis in the Caribbean—9 species—is the result of a minimum of three (probably four) colonization events. Two distinct lineages of Caribbean endemics are designated with green branches in Figure 3: A. paupercula and a clade containing A. reesii (Jenman) C. Chr. (the “reesii clade”). Additionally, the widespread taxon A. radiata also occurs in the Caribbean, indicating a third (and possibly on-going; see below) colonization event. DNA from herbarium specimens of A. asplenioides Maxon, endemic to Cuba, could not be successfully amplified, precluding this taxon’s inclusion in the molecular phylogeny; however, based on its possession of carinae and cristate spores, we hypothesize that A. asplenioides would likely be positioned near the “tweediana/dichotoma” clade (Fig. 5). Hence, it appears that there have been no fewer than four migrations of Adiantopsis into the Caribbean. Biogeographic studies of ferns on oceanic islands often reveal that fern diversity is the result of multiple colonization events that are followed by limited divergence (Driscoll and Barrington, 2007; Geiger and Ranker, 2005; Schneider et al., 2004). Short branches in the “reesii clade” (Fig. 7) indicate little sequence divergence among these six species, suggesting either a relatively recent radiation within the Caribbean or, perhaps, ecological constraints to be slow-evolving. The palmate clade (Fig. 6) appears to be associated with a rather successful adaptive radiation, with species occurring in more diverse phytogeographic regions than any other subclade (Fig. 4). Morphological investigations of herbarium specimens revealed a surprisingly large amount of variation within the classical circumscription of A. radiata, the only broadly recognized palmate species in Adiantopsis. As currently interpreted, the complex consists of A. radiata plus several geographically localized, peripheral taxa. These taxa differ from A. radiata in the form of the indusium, the attachment of the ultimate segments to the costae, and the form of the adaxial carinae (see 67 Chapter 2). Branch lengths in the palmate clade of the phylogram (Fig. 7) appear consistent with a gradual and steady accumulation of palmate taxa, probably in connection with dispersal into new habitats followed by diversification. This is in contrast to a scenario involving a hypothetical widespread ancestral taxon giving rise to the many diverse palmate species via allopatric differentiation of populations following a vicariant event. The widespread A. radiata, arguably the most familiar and charismatic member of the genus, occupies a derived position in the palmate clade and seems to have evolved to be an especially good disperser and/or colonizer. In comparison to the other palmate taxa, the echinae on A. radiata spores are much longer and more robust, with exaggerated dissected bases below the echinae; the significance of these differences in spore ornamentation are unclear, but we speculate that well-developed echinae with dissected bases may impart increased loft and dispersability. Relationship between lamina morphologies and phylogeny. — The three types of laminar architecture present in Adiantopsis (Fig. 1) provide a superb opportunity to explore the origin and evolution of a great range of morphological variation within one small lineage (Barker and Hickey, 2006). In a survey of nearly 9000 extant fern species, Tryon (1964) found that more than 85 % of species and genera have pinnate leaf architecture; therefore, he hypothesized that the pinnate form is plesiomorphic in the ferns and that other architectures, such as the palmate one which is found in fewer than 1% of species, are derived from the pinnate form. Our data support that the pinnate architecture is, indeed, pleisiomorphic in Adiantopsis and that the palmate morphology evolved only once. Three species exhibit pedate lamina morphology, and their positions in the phylogeny indicate that the pedate architecture has had three independent origins (Fig. 6). Adiantopsis pedata, described from Jamaican material in 1852, was the only pedate taxon known until a morphologically similar specimen was collected in Argentina (Hickey et al., 2003). The Argentinian specimen, A. xaustralopedata Hickey, M.S. Barker & Ponce, was found to be a sterile hybrid, with hypothesized progenitors of A. radiata and A. perfasciculata Sehnem (Hickey et al., 2003). This discovery stimulated a closer look at the pedate taxon in the Caribbean and led to the recognition of two pedate species there: A. pedata and A. pentagona M.S. Barker & Hickey (Barker and Hickey, 2006). Barker 68 and Hickey (2006) hypothesized that the pedate morphology exhibited by A. pedata and A. pentagona arose as a result of hybridization between the palmate A. radiata and two different pinnate taxa. The hybrid origin for these pedate taxa is supported by similarities they share with A. radiata in details of spore ornamentation and laminar development, while similarities in ultimate division shape and mode of number of arcus cells found in the sporangium are some characteristics Barker and Hickey (2006) used to hypothesize the identity of the pinnate parents. Both A. pedata and A. pentagona were proposed to be tetraploids based on guard cell size, spore length, and arcus cell height. Our DNA sequence data places the Caribbean pedate taxa in the clade with their hypothesized pinnate parents; therefore, since plastids are maternally inherited in the cheilanthoids (Gastony and Yatskievych, 1992), it appears that palmate A. radiata is the paternal parent (Fig. 6). Considering that A. radiata is sympatric with so many other species in Adiantopsis, it is not surprising to find that A. radiata is implicated in a web of reticulate evolution, contributing as the sperm donor for no fewer than three taxa of putative hybrid origin. Remaining questions. — One surprising and fairly inexplicable result of this study is the placement of the pinnate Adiantopsis lindigii within the palmate clade and sister to the ternate A. ternata Prantl (Fig. 6). These two species share approximately four nucleotide substitutions and a Colombian distribution (Fig. 4), but are otherwise dissimilar morphologically. This placement may be the result of a character reversal or an artifact of the molecular analyses (although both individual markers were congruent with this surprising placement). Another possibility is that Adiantopsis lindigii is the product of a hybrid origin. Support for this is its spore length (mean=43.63 µm; unpublished data). In many cases, spore length can be used as a proxy for estimating ploidy level in ferns (Barrington et al., 1986). For a sphere, a 1.26 increase in diameter accompanies a doubling of volume; likewise, for tetrahedral spores, such as in Adiantopsis, the spore length of a known diploid can be multiplied by 1.26 to estimate the spore length that might be observed with doubling of chromosomes (Barrington et al., 1986). Adiantopsis radiata has a known base chromosome number of n=30 (Walker, 1973) and can serve as a representative sexual diploid for the genus; a doubling of spore volume would be accomplished by an 69 increase in spore length from a mean of 32.91 µm (unpublished data) to 41.47 µm. Based on this reasoning, taxa with mean spore lengths equal to or greater than 41.47 µm are presumed to be tetraploid (A. lindigii, A. minutula, A. monticola, A. trifurcata; unpublished data). Published spore lengths for putative Caribbean tetraploids A. pedata and A. pentagona (45.5 and 45.77 um, respectively; Barker and Hickey, 2006) indicate that these taxa also fit within the spore length range expected for tetraploid members of the genus. It is important to point out that in addition to changes in ploidy level, sporesize can vary due to numerous other factors: adaptation for dispersal (for example, larger to stay closer to parent plant in insular habitats), increase for nutritional reasons, and environmental parameters (reviewed by Barrington et al., 1986). Guard cell size is also correlated with ploidy level, but can be complicated by environmentally induced variability, such as smaller stomates in arid habitats (Barrington et al., 1986). Although it seems likely that A. lindigii is an allotetraploid, examination of its gross morphology has rendered no clues as to possible parentage. We might suspect that one parent is a palmate taxon (not A. radiata, in this hypothetical case) and that it is the maternal contributor, since A. lindigii is in the palmate clade of our phylogeny (Fig. 6). Unfortunately, the evolutionary history of this rare taxon may remain enigmatic indefinitely. There are several previously undescribed taxa, both palmate and pinnate, resolved in the molecular phylogeny (Fig. 2; species descriptions in progress). Among these are several accessions constituting the polytomy below A. chlorophylla (Fig. 2). An interesting feature shared by the members of this polytomy is the presence of ambiguous character states (peaks under peaks) in the electropherograms of the rbcL sequence data. There are seven such nucleotide sites shared between them, suggesting that there may be heterogeneous or polymorphic copies of the gene in those taxa. Other possibilities could include PCR error or heteroplasmy, although the latter seems implausible since there were no polymorphisms observed in the other plastid gene, atpA (if our marker were nuclear, rather than plastid, we could consider the possibility of hybridization). Morphological study of these pinnate taxa and the widespread and poorly understood pinnate A. chlorophylla reveals that canalization of morphology among taxa and extensive phenotypic plasticity within taxa has complicated species delimitation using morphology alone. Unfortunately, the small amount of sequence divergence captured by 70 our two plastid markers (Fig. 7) coupled with the polymorphic rbcL data for several of these taxa, has not permitted us to tease apart the distinct taxa that contribute to this species complex; additional study of herbarium specimens and possibly the addition of a third molecular marker will be required before we fully understand them. Summary and conclusions. — Our research revealed previously unrecognized diversity within Adiantopsis, and indicated that the morphological characteristics traditionally used to distinguish Adiantopsis are non-congruent with each other and lead to a circumscription that is overly conservative and excludes several basal members in the clade. Additionally, several species currently referred to Cheilanthes are embedded within the Adiantopsis clade, placements also supported by morphological similarities. The expansion of generic limits to include non-carinate and cristate-spored species in Adiantopsis, and the transfer into Adiantopsis of three taxa currently referred to Cheilanthes, provides for a clearly defensible and monophyletic circumscription. This expanded circumscription is well-supported by the molecular data and a suite of morphological characters that clearly separate Adiantopsis from other genera in the hemionitidoid subclade of the cheilanthoid ferns. Our interpretation of the biogeography suggests a South American origin for Adiantopsis, specifically in Cerrado and associated dry areas, followed by dispersal and diversification in various phytogeographic regions, including multiple migrations into the Caribbean. We currently recognize 28 species (3 newly described; see Chapter 2) in Adiantopsis, not counting the ten unnamed species for which we have molecular data; therefore, the number of species in Adiantopsis approaches forty. A synopsis is presented below. 71 SYNOPSIS OF ADIANTOPSIS Adiantopsis Fée, Gen. Filic. [Mém. Foug. 5] 145. 1852 — Type: Adiantum pauperculum Kunze [ ≡ Adiantopsis paupercula (Kunze) Fée]. Terrestrial, rupicolous, or xerophytic ferns. Rhizomes erect, ascending, or decumbent; scales bicolorous, the central band black, dark brown, or atrocastaneous, margins golden, reddish-golden, tan, or brown. Fronds erect, strict, or caespitose, monomorphic to hemidimorphic. Stipes atropurpureous, atrocastaneous, castaneous, or medium brown, always, occasionally, rarely, or never with paired adaxial carinae or large reddish pluricellular hairs; carinae, when present, golden or reddish-golden. Laminae pinnate, palmate, or pedate, apices difform or conform; laminar tissue spongiose, chartaceous, coriaceous, adaxially glabrous with marginal hydathodes, abaxially with diffuse septate hairs, three cells long, middle cell short, apical cell often bulbous. Laminar axes persistent or marcescent, atropurpureous, atrocastaneous, castaneous, or ebeneous, adaxially bicarinate or with reddish pluricellular hairs. Pinnules anadromous, slightly ascending, transverse, or slightly descending, deciduous, articulate, oblong, narrowly oblong, ovate, elliptic, lanceolate, trullate, or rhombiform, bases cuneate, obtuse, acute, or truncate, acroscopic auricles present in some species, margins entire, sinuate, crenulate, or deeply lobed, apices round, lobulate, crenate, acute, or acuminate; stalks persistent. Veins free, anadromous, occult, obscure, or prominent. Sori marginal, discrete, occasionally confluent on very fertile fronds. Pseudoindusia distinct, lunate, cuspidate, quadrangular or triangular, scarious, chartaceous, or membranaceous, margins entire, slightly erose, erose, or lacinate. Sporangia subglobose to globose, long stalked to sessile. Spores 64 or 32 per sporangia, tetrahedral-globose, tetrahedral, echinate, echinulate, or arachnoid-echinulate with echinae bases complete or dissected, or cristate, laesura visible or obscured by ornamentation The number of species in Adiantopsis approaches forty: twenty-eight accepted species (three of those newly-described; in preparation) and molecular data to support an additional ten taxa. Adiantopsis has a strictly neotropical distribution, with species occurring in the Caribbean, Mexico and Central America, and throughout much of South America, as far south as the northern regions of Uruguay. 72 Chromosome numbers have been counted for just one member of the genus, A. radiata (2n=60, see Walker, 1973). Adiantopsis alata Prantl, Gartenfl. 32: 99. t.1115. 1883. Brazil. Adiantopsis asplenioides Maxon, Amer. Fern J. 22: 14. 1932. Cuba. Adiantopsis xaustralopedata Hickey, M.S. Barker & Ponce, Amer. Fern. J. 93(1): 44. 2003. Argentina, Brazil, Paraguay. Adiantopsis chlorophylla (Sw.) Fée, Gen. Filic. [Mém. Foug. 5]: 145. 1852 ≡ Cheilanthes chlorophylla Sw., Kongl. Vetensk. Acad. Handl. 76. 1817. Argentina, Bolivia, Brazil, Paraguay, Uruguay. Adiantopsis crinoidea Link- Pérez & Hickey ined. (in preparation; Chapter 2). Venezuela. Adiantopsis dactylifera Link- Pérez & Hickey ined. (in preparation; Chapter 2). Peru. Adiantopsis dichotoma (Cav.) T. Moore, Index Fil. (T. Moore) 17. 1857 ≡ Cheilanthes dichotoma Sw., Syn. Fil.: 129, 335. t.3. f.7. 1806. Argentina, Brazil, Paraguay, Uruguay. Adiantopsis flexuosa (Kunze) Link-Pérez & Hickey, comb. nov. ≡ Cheilanthes flexuosa Kunze, Linnaea 22: 578. 1849. Brazil. Adiantopsis lindigii (Mett.) Prantl, Gartenfl. 32. 101. 1883 ≡ Cheilanthes lindigii Mett., Ann. Sci. Nat. Bot., sér. 5, 2: 218. 1864. Colombia. Adiantopsis minutula Sehnem, Fl. Ilustr. Catarin. 1 (Pteridac.), 73. 1972 Brazil. Adiantopsis monticola (Gardner) T. Moore, Index Fil. (T. Moore) xxxvii. 1857 ≡ Cheilanthes monticola Gardner, in Hook., Ic. pl. t. 487. 1842. Brazil. Adiantopsis occulta Sehnem, Pesquisas 3: 508. 1959 ≡ Adiantopsis chlorophylla (Sw.) Fée f. paludosa Rosenst., Hedwigia 46: 84. 1906. 73 Brazil. Adiantopsis parvisegmenta M.S. Barker & Hickey, Ann. Misouri Bot. Gard. 93(3): 388389; fig. 6C, D. 2006. Cuba. Adiantopsis paupercula (Kunze) Fée, Gen. Filic. [Mém. Foug. 5]:145. 1852 ≡ Adiantum pauperculum Kunze, Farrnkräuter 2(13):65. t. 127. 1850 ≡ Cheilanthes paupercula (Kunze) Mett., Fil. Hort. Bot. Lips. 52. 1856 ≡ Hypolepis paupercula (Kunze) Hook., Sp. Fil. 2:73. t 88-C. 1856. Cuba, Jamaica, Puerto Rico. Adiantopsis pedata (Hook.) T. Moore, Index Fil. (T. Moore) 18. 1857 ≡ Hypolepis pedata Hook., Sp. Fil. 2:73, t. 92-A. 1852 ≡ Cheilanthes pedata (Hook.) A. Braun, Index Sem. 1857. Dominican Republic, Jamaica. Adiantopsis pentagona M.S. Barker & Hickey, Ann. Missouri Bot. Gard. 93(3): 392394; fig. 8A, B. 2006. Cuba. Adiantopsis perfasciculata Sehnem, Pesquisas 5, Bot. 13: 21, t4, 5. 1961 Brazil. Adiantopsis propinqua (Mett.) Prantl, Gartenfl. 32. 101. 1883 ≡ Cheilanthes propinqua Mett., Ann. sc. nat. V. 2. 218. 1864. Colombia. Adiantopsis radiata (L.) Fée, Gen. Filic. [Mém. Foug. 5]: 145. 1852 ≡ Adiantum radiatum L., Sp. Pl. 2: 1094. 1753 ≡ Cheilanthes radiata (L.) J.Sm., in J. Bot. (Hooker) 4:159. 1841 ≡ Hypolepis radiata (L.) Hook., Sp. Fil. 2: 72. 1852. Argentina, Belize, Bolivia, Brazil, Colombia, Costa Rica, Cuba, Guatemala, Guyana, Honduras, Jamaica, Lesser Antilles, Mexico, Nicaragua, Panama, Paraguay, Peru, Trinidad, Venezuela, West Indies. Adiantopsis reesii (Jenm.) C. Chr., Ind. Fil. 22. 1905 ≡ Cheilanthes reesii Jenm., J. Bot. Brit. For. 24:267. 1886. Hispaniola, Jamaica. Adiantopsis regularis (Kunze) T. Moore, Index Fil. (T. Moore) 252. 1861 ≡ Adiantum regulare Kunze, Farnkr. Erde 2: 66. 1850 ≡ Cheilanthes regularis Mett., Cheil. 41 n. 56. 1859. Brazil, Paraguay. Adiantopsis rupicola Maxon, Contr. U.S. Natl. Herb. 10:485. 1908. Cuba. 74 Adiantopsis seemannii (Hook.) Maxon, Proc. Biol. Soc. Wash. 52: 113. 1939 ≡ Cheilanthes seemannii Hk., Sp. Fil. 2: 85, t. 97A. 1852. Mexico. Adiantopsis ternata Prantl, Gartenfl. 32: 101. 1883. Colombia, Venezuela. Adiantopsis timida Link- Pérez & Hickey ined. (in preparation; Chapter 2). Brazil, Colombia, Guyana, Venezuela. Adiantopsis trifurcata (Baker) Link-Pérez & Hickey, comb. nov. ≡ Cheilanthes trifurcata Baker, Kew Bull. 144. 1901. Bolivia, Brazil. Adiantopsis tweediana (Hook.) Link-Pérez & Hickey, comb. nov. ≡ Cheilanthes tweediana Hook., Sp. Fil. 84, t. 96B. 1852. Argentina, Bolivia, Brazil, Paraguay, Uruguay. Adiantopsis vincentii M.S. Barker & Hickey, Ann. Missouri Bot. Gard. 93(3): 398-400; fig. 10. 2006. Cuba. Taxa of uncertain placement Adiantopsis cheilanthoides R.M. Senna, Iheringia, Bot. 59(1): 108. 2004. Brazil. [unable to examine type material] Adiantopsis luetzelburgii Rosenst., Repert. Spec. Nov. Regni Veg. 20: 91. 1924. Excluded species Adiantopsis californica (Hook.) T. Moore, Index Fil. xxxvii. 1857 ≡ Hypolepis californica Hook., Sp. Fil. 2: 71, pl. 88a. 1851 ≡ Cheilanthes californica (Hook.) Mett., Abh. Senckenberg. Naturf. Ges. 3: 88. 1859 ≡ Aspidotis californica (Hook.) Nutt. ex Copel., Gen. Fil. (Copeland) 68. 1947. Adiantopsis capensis (Thbg.) Fée, Gen. Filic. [Mém. Foug. 5]:145. 1850-1852 ≡ Adiantum capense Thbg., Prod. Fl. Cap. 173. 1800 ≡ Cheilanthes capensis (Thbg.) Sw., Syn. Fil. 128. 1806. 75 Adiantopsis fordii (Baker) C. Chr., Index Filic. fasc. 1: 22. 1905 ≡ Cheilanthes fordii Baker, J. Bot. 17(202): 304. 1879. Adiantopsis incisa Moore, Index Fil. (T. Moore) 243. 1861. Adiantopsis linearis Bonap., N. Pt. 10. 185. 1920 ≡ possibly Cheilanthes, but this taxon will need a new combination (as discussed in Moran and Smith, 2001) since the name Cheilanthes linearis is already taken for another taxon (Cheilanthes linearis Moore ≡ Pellaea quadripinnata, but the original combination needs to remain available for when the taxon is placed in Cheilanthes). We make no recommendations as to final placement of Adiantopsis linearis Bonap. except to exclude it from Adiantopsis. Adiantopsis madagascariensis (Baker) C. Chr., Index Filic. 22. 1905 ≡ Cheilanthes madagascariensis Baker, J. Linn. Soc., Bot. 16: 198. 1877. Adiantopsis microphylla (Kl.) Prantl, nom. nud., Salom. Nom. 15. 1883. Adiantopsis pteroides (Sw.) T. Moore, Index Fil. (T. Moore) xxxvii. 1857 ≡ Cheilanthes pteroides Sw., Syn. Fil. 128. 1806. Adiantopsis schimperi (Kunze) T. Moore, Index Fil. (T. Moore) xxxvii. 1857 ≡ Cheilanthes schimperi Kunze, Farrnkräuter 1: 52 t. 26. 1840 ≡ Aspidotis schimperi (Kunze) Pic. Serm., Webbia 7: 326. 1950 ≡ Hypolepis schimperi (Kunze) Hook., Sp. Fil. 2: 70. 1852. 76 ACKNOWLEDGEMENTS Many thanks are extended to the curators of the following herbaria for loans or for providing access to material during visits: AAU, B, F, GH, GUYN, HB, HBG, HUT, K, MICH, MO, MU, NY, PORT, SI, SP, UB, UC, UPCB, US, USZ, and VEN. Alan Smith, George Yatskievych, Carl Rothfels, and Michael Kessler are thanked for providing plant material for sequencing. This research was supported by grants from W.S. Turrell Herbarium Fund (Miami University) and Academic Challenge Research Grants (Department of Botany, Miami University). 77 LITERATURE CITED Barker, M.S., & Hickey, R.J. 2006. A taxonomic revision of Caribbean Adiantopsis (Pteridaceae). Annals of the Missouri Botanical Garden 93:371-401. Barrington, D.S., Paris, C.A., & Ranker, T.A. 1986. Systematic inferences from spore and stomate size in the ferns. American Fern Journal 76:149-159. Benson, D.A., Karsch-Mizrachi, I., Lipman, D.J., Ostell, J., & Wheeler, D.L. 2005. GenBank. Nucleic Acids Research 33:D34-D38. Chase, M.W., & Hills, H.H. 1991. Silica gel: an ideal material for field preservation of leaf samples for DNA studies. Taxon 40:215-220. Christensen, C. 1932. The Pteridophyta of Madagascar. Dansk Bot. Ark. 7:1-203, i-xv, + 280 plates. Driscoll, H.E., & Barrington, D.S. 2007. Origin of Hawaiian Polystichum (Dryopteridaceae) in the context of a world phylogeny. American Journal of Botany 94:1413-1424. Felsenstein, J. 2004. Inferring phylogenies. Sinauer Associates, Sunderland, MA. Figueiredo, J.B., & Salino, A. 2005. Pteridophytes from four "Reservas particulares do Patrimonio Natural" (RPPNs) in the South of the metropolitan region of Belo Horizonet, Minas Gerais, Brazil. Lundiana 6:83-94. Gastony, G.J., & Rollo, D. 1996. Congruence of phylogenetic reconstructions in cheilanthoid ferns based on nucleotide sequences of maternally inherited rbcL and biparentally inherited nuclear ribosomal ITS. American Journal of Botany 83:125. Gastony, G.J., & Rollo, D.R. 1995. Phylogeny and generic circumscriptions of cheilanthoid ferns (Pteridaceae: Cheilanthoideae) inferred from rbcL nucleotide sequences. American Fern Journal 85:341-360. Gastony, G.J., & Rollo, D.R. 1998. Cheilanthoid ferns (Pteridaceae: Cheilanthoideae) in the southwestern United States and adjacent Mexico: A molecular phylogenetic reassessment of generic lines. Aliso 17:131-144. Gastony, G.J., & Yatskievych, G. 1992. Maternal inheritance of the chloroplast and mitochondrial genomes in Cheilanthoid ferns. American Journal of Botany 79:716-722. 78 Geiger, J.M.O., & Ranker, T.A. 2005. Molecular phylogenetics and historical biogeography of Hawaiian Dryopteris (Dryopteridaceae). Molecular Phylogenetics and Evolution 34:392-407. Gentry, A.H. 1982. Neotropical floristic diversity: phytogeographical connections between Central and South America, Pleistocene climatic fluctuations, or an accident of the Andean orogeny? Annals of the Missouri Botanical Garden 69:557-593. Hasebe, M., Wolf, P.G., Pryer, K.M., Ueda, K., Ito, M., Sano, R., Gastony, G.J., Yokoyama, J., Manhart, J.R., Murakami, N., Crane, E.H., Haufler, C.H., & Hauk, W.D. 1995. Fern phylogeny based on rbcL nucleotide sequences. American Fern Journal 85:134-181. Haufler, C.H., Ranker, T.A., & Li, J. 1996. Using rbcL sequences to test hypotheses of phylogenetic relationship in the fern families Polypodiaceae and Grammitidaceae. American Journal of Botany 83:126. Hickey, R.J., Barker, M.S., & Ponce, M. 2003. An Adiantopsis hybrid from Northeastern Argentina and vicinity. American Fern Journal 93:42-44. Huelsenbeck, J.P., & Ronquist, F. 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics (Oxford) 17:754-755. Kirkpatrick, R.E.B. 2007. Investigating the monophyly of Pellaea (Pteridaceae) in the context of a phylogenetic analysis of cheilanthoid ferns. Systematic Botany 32:504-518. Maddison, W., & Maddison, D. 2000. MacClade Version 4.0: analysis of phylogeny and character evolution. Sinauer, Sunderland, MA. Mickel, J., & Smith, A.R. 2004. The pteridophytes of Mexico. New York Botanical Garden Press, Bronx, NY. Moran, R.C., & Smith, A.R. 2001. Phytogeographic relationships between neotropical and African-Madagascan pteridophytes. Brittonia 53:304-351. Nakazato, T., & Gastony, G.J. 2003. Molecular phylogenetics of Anogramma species and related genera (Pteridaceae: Taenitidoideae). Systematic Botany 28:490-502. 79 Posada, D. 2006. ModelTest Server: a web-based tool for the statistical selection of models of nucleotide substitution online. Nucleic Acids Research 34:W700W703. Posada, D., & Buckley, T.R. 2004. Model selection and model averaging in phylogenetics: Advantages of akaike information criterion and Bayesian approaches over likelihood ratio tests. Systematic Biology 53:793-808. Posada, D., & Crandall, K.A. 1998. ModelTest: Testing the model of DNA substitution. Bioinformatics 14:817-818. Prado, J., Rodrigues, C.D.N., Salatino, A., & Salatino, M.L.F. 2007. Phylogenetic relationships among Pteridaceae, including Brazilian species, inferred from rbcL sequences. Taxon 56:355-368. Pryer, K.M., Schuettpelz, E., Wolf, P.G., Schneider, H., Smith, A.R., & Cranfill, R. 2004. Phylogeny and evolution of ferns (monilophytes) with a focus on the early leptosporangiate divergences. American Journal of Botany 91:1582-1598. Pryer, K.M., Smith, A.R., Hunt, J.S., & Dubuisson, J.-Y. 2001. rbcL data reveal two monophyletic groups of filmy ferns (Filicopsida: Hymenophyllaceae). American Journal of Botany 88:1118-1130. Rambaut, A. 2002. Se-Al Sequence Alignment Editor, Version 2.0aa11. Available at http://evolve.zoo.ox.ac.uk/software/SeAl/main.html. In. Ronquist, F., & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572-1574. Rosen, D.E. 1975. A vicariance model of Caribbean biogeography. Systematic Zoology 24:431-464. Rothfels, C.J., Windham, M.D., Grusz, A.L., Gastony, G.J., & Pryer, K.M. 2008. Toward a monophyletic Notholaena (Pteridaceae): resolving patterns of evolutionary convergence in xeric-adapted ferns. Taxon 57:712-724. Savolainen, V., Cuenoud, P., Spichiger, R., Martinez, M.D.P., Crevecoeur, M., & Manen, J.-F. 1995. The use of herbarium specimens in DNA phylogenetics: Evaluation and improvement. Plant Systematics and Evolution 197:87-98. Schneider, H., Russell, S.J., Cox, C.J., Bakker, F., Henderson, S., Rumsey, F., Barrett, J., Gibby, M., & Vogel, J.C. 2004. Chloroplast phylogeny of asplenioid 80 ferns based on rbcL and trnL-F spacer sequences (Polypodiidae, aspleniaceae) and its implications for biogeography. Systematic Botany 29:260-274. Schuettpelz, E., Korall, P., & Pryer, K.M. 2006. Plastid atpA data provide improved support for deep relationships among ferns. Taxon 55:897-906. Schuettpelz, E., & Pryer, K.M. 2007. Fern phylogeny inferred from 400 leptosporangiate species and three plastid genes. Taxon 56:1037-1050. Schuettpelz, E., Schneider, H., Huiet, L., Windham, M.D., & Pryer, K.M. 2007. A molecular phylogeny of the fern family Pteridaceae: Assessing overall relationships and the affinities of previously unsampled genera. Molecular Phylogenetics and Evolution 44:1172-1185. Swofford, D.L. 2000. PAUP*4.0 - Phylogenetic Analysis Using Parsimony (*and Other Methods). Sinauer Associates, Sunderland, MA. Tryon, R.M. 1964. Evolution in the leaf of living ferns. The Torrey Botanical Club 21:73-85. Tryon, R.M., & Tryon, A.F. 1973. Geography, spores, and evolutionary relations in the cheilanthoid ferns. Pp. 145-153 in: Jermy, A.C., Crabbe, J.A., & Thomas, B.A., (eds), The phylogeny and classification of ferns. Academic Press, New York. p 145-153. Tryon, R.M., & Tryon, A.F. 1982. Ferns and allied plants with special reference to tropical America. Springer-Verlag, New York. Tryon, R.M., Tryon, A.F., & Kramer, K.U. 1990. Pteridaceae. Pp. 230-256 in: Kramer, K.U., & Green, P.S., (eds), The families and genera of vascular plants, vol. I, Pteridophytes and gymnosperms. Springer-Verlag, New York. p 230-256. Tsutsumi, C., & Kato, M. 2005. Molecular phylogenetic study on Davalliaceae. Fern Gazette 17:147-162. Walker, T.G. 1973. Evidence from cytology in the classification of ferns. Pp. 91-110 in: Jermy, A.C., (ed), The phylogeny and classification of ferns. Academic Press. p 91-110. Windham, M.D., Huiet, L., Schuettpelz, E., Grusz, A.L., Rothfels, C., Beck, J., Yatskievych, G., & Pryer, K.M. 2009. Using plastid and nuclear DNA 81 sequences to redraw generic boundaries and demystify species complexes in cheilanthoid ferns. American Fern Journal 99:128-132. Wolf, P.G. 1995. Phylogenetic analyses of rbcL and nuclear ribosomal RNA gene sequences in Dennstaedtiaceae. American Fern Journal 85:306-327. Wolf, P.G., Soltis, P.S., & Soltis, D.E. 1994. Phylogenetic relationships of dennstaedtioid ferns: Evidence from rbcL sequences. Molecular Phylogenetics and Evolution 3:383-392. Zuloaga, F.O., Morrone, O., & Belgrano, M.J. (eds.). 2008. Catalogo de las plantas vasculares del Cono Sur : (Argentina, Sur de Brasil, Chile, Paraguay y Uruguay). Missouri Botanical Garden Press, St. Louis, Mo., U.S.A. 82 Appendix. Taxon; voucher specimen (herbarium); collection locality; GenBank accession number (and citations, for previously published data) for rbcL; atpA (in that order). “–” indicates missing data. An “*” after a specimen indicates that the accession was removed from the combined dataset after confirming that the DNA sequence was identical to other representatives of the same taxon; this was done to simplify analysis and to improve the appearance of the trees. Sequences generated in this study are indicated by “TBS” (“to be submitted”). Adiantopsis chlorophylla (Sw.) Fée; Prado & Yano 1047 (SP); Brazil; EF473684 (Prado et al., 2007); –. Adiantopsis dactylifera Link-Perez & Hickey ined.; van der Werff & Gray 16380 (MO); Peru; TBS; TBS. Adiantopsis dichotoma Sw.; Romero & al. 3819 (MU); Argentina; TBS; TBS. Adiantopsis dichotoma Sw.; Deginani & al. 1359 (MO); Argentina; TBS; TBS.* Adiantopsis lindigii (Mett.) Prantl; Lengenheim 3568 (UC); Colombia; TBS; TBS. Adiantopsis monticola (Gardner) T. Moore; Cordeiro & al. 2665 (UC); Brazil; TBS; TBS. Adiantopsis parvisegmenta M.S. Barker & Hickey; Jack 6877 (NY); Cuba; TBS; TBS. Adiantopsis paupercula (Kunze) Fée; Hespenheide 1191 (US); Jamaica; TBS; TBS. Adiantopsis pedata (Hook.) T. Moore; Harris 10878 (NY); Jamaica; TBS; TBS. Adiantopsis pentagona M.S. Barker & Hickey; Proctor 3229 (NY); Cuba; TBS; TBS. Adiantopsis propinqua (Mett.) Prantl; Killip & Smith 16360 (NY); Colombia; –; TBS. Adiantopsis radiata (L.) Fée; Christenhusz 4033 (TUR); Guadaloupe; EF452131 (Schuettpelz et al., 2007); EU268718 (Rothfels et al., 2008).* Adiantopsis radiata (L.) Fée; Cremers 5049 (NY); French Guiana; TBS; TBS.* Adiantopsis radiata (L.) Fée; Link-Perez et al. 183 (MU); Argentina; TBS; TBS. Adiantopsis radiata (L.) Fée; Link-Perez et al. 327 (MU); Bolivia; TBS; TBS.* Adiantopsis radiata (L.) Fée; Link-Perez et al. 355 (MU); Bolivia; TBS; TBS.* Adiantopsis radiata (L.) Fée; Prado & Yano 1046 (SP); Brazil; EF473685 (Prado et al., 2007); –.* Adiantopsis reesii (Jenman) C. Chr.; Ekman 7965 (NY); Haiti; TBS; TBS. Adiantopsis regularis Moore; Silva & Barbosa 3001 (UC); Brazil; TBS; TBS. Adiantopsis rupicola Maxon; Britton & al. 7497 (NY); Cuba; TBS; TBS. Adiantopsis rupicola Maxon; Morton 9791 (US); Cuba; TBS; TBS.* Adiantopsis seemannii (Hook.) Maxon; Reina & al. 2008-726 (MO); Mexico; TBS; TBS. Adiantopsis ternata Prantl; Haught 2402 (US); Colombia; TBS; TBS. Adiantopsis timida Link-Perez & Hickey ined.; Wurdack & Adderley 43734 (NY); Venezuela; TBS; TBS. Adiantopsis timida Link-Perez & Hickey ined.; Santos & al. 249 (NY); Brazil; TBS; TBS.* Adiantopsis vincentii M.S. Barker & Hickey; Morton 10389 (US); Cuba; TBS; TBS.* Adiantopsis vincentii M.S. Barker & Hickey; Morton 10544 (US); Cuba; TBS; TBS. Adiantopsis xaustralopedata Hickey, M.S. Barker & Ponce; Zuloaga & al. 8873 (MU); Argentina; TBS; TBS. Adiantopsis xaustralopedata Hickey, M.S. Barker & Ponce; Hickey & al. 01-63 (MU); Argentina; TBS; TBS.* Aleuritopteris argentea (S.G. Gmelin) Fée; Yatskievych 01-23 (MO); China; EF452137 (Schuettpelz et al., 2007); EU268719 (Rothfels et al., 2008). Aspidotis densa (Brack.) Lellinger; Pryer & al. 472 (DUKE); Oregon, U.S.A.; EU268773 (Rothfels et al., 2008); EU268723 (Rothfels et al., 2008). Cheilanthes [=Adiantopsis] flexuosa Kunze; Mello-Silva & al. 2566 (SP); Brazil; TBS; TBS.* Cheilanthes [=Adiantopsis] flexuosa Kunze; Pirani 5059 (SP); Brazil; TBS; TBS. Cheilanthes [=Adiantopsis] flexuosa Kunze; Thomas & al. 9680 (NY); Brazil; TBS; TBS.* Cheilanthes micropteris Sw.; Deginani & al. 1363 (MO); Argentina; EF452145 (Schuettpelz et al., 2007); –. Cheilanthes multifida ssp lacerata N.C. Anthony 83 & Schelpe; Gereau & al. 6409 (MO); Tanzania; TBS; TBS. Cheilanthes nitidula Hook.; Schneider s.n. (GOET); –; EF452146 (Schuettpelz et al., 2007); EF452085 (Schuettpelz et al., 2007). Cheilanthes [=Adiantopsis] trifurcata Baker; Jimenez & Huaylla 2685 (NY); Bolivia; TBS; TBS. Cheilanthes [=Adiantopsis] trifurcata Baker; Pirani 1314 (NY); Brazil; TBS; TBS.* Cheilanthes [=Adiantopsis] tweediana Hook.; Link-Perez & Meza Torres 213 (MU); Argentina; TBS; TBS. Cheilanthes [=Adiantopsis] tweediana Hook.; Zardini 6640 (MO); Paraguay; TBS; –.* Doryopteris concolor (Langsd. & Fisch.) Kuhn; Shiyong Dong 178 (PE); China; AY266414 (Zhang et al., 2007); –. Doryopteris sagittifolia (Raddi) J. Sm.; Schuettpelz 562 & Schneider (GOET); cult., origin unknown; EF452151 (Schuettpelz et al., 2007); EU268742 (Rothfels et al., 2008). Hemionitis palmata L.; Schuettpelz 297 (DUKE); cult., origin unknown; AY357708 (Ranker & Geiger, unpub.); EU268743 (Rothfels et al., 2008). Notholaena aschenborniana Klotzsch; Schuettpelz & al. 476 (DUKE); Arizona, U.S.A.; EF452159 (Schuettpelz et al., 2007); EU268745 (Rothfels et al., 2008). Notholaena standley Maxon; Schuettpelz & al. 435 (DUKE); Arizona, U.S.A.; EU268802 (Rothfels et al., 2008); EU268757 (Rothfels et al., 2008). Pellaea viridis Sw.; Janssen 2701 (P); Ile de la Reunion, France; EF452147 (Schuettpelz et al., 2007); EU2687767 (Rothfels et al., 2008). Pentagramma triangularis (Kaulf.) Yatsk., Windham & Wollenw. subsp. maxonii (Weath.) Yatsk., Windham & Wollenw.; Schuettpelz & al. 445 (DUKE); Arizona, U.S.A.; EF452165 (Schuettpelz et al., 2007); EU268768 (Rothfels et al., 2008). Species novum 1 (OTU68); Henkel & Chin 5721 (NY); British Guiana; TBS; TBS. Species novum 2 (OTU67); Henkel & Chin 983 (NY); British Guiana; TBS; TBS. Species novum 3 (OTU62); Mori & Smith 25095 (NY); French Guiana; TBS; TBS. Species novum 3 (OTU66); Granville & al. 15476 (NY); French Guiana; TBS; TBS.* Species novum 4 (OTU97); Haught 3901 (US); Colombia; TBS; TBS. Species novum 5 (OTU96); Lobo & al. 14 (MU); Costa Rica; TBS; TBS. Species novum 6 (OTU92); Rivera & al. 1732 (NY); Mexico; TBS; TBS. Species novum 7 (OTU5); Link-Perez & al. 193 (MU); Argentina; TBS; TBS. Species novum 8 (OTU7); Link-Perez & al. 356 (MU); Bolivia; TBS; TBS. Species novum 8 (OTU26); Link-Perez & al. 357 (MU); Bolivia; TBS; TBS.* Species novum 9 (OTU30); Meza Torres & Link-Perez 337 (MU); Argentina; TBS; TBS. Species novum 10 (OTU54); Irwin & al. 13128 (NY); Brazil; TBS; TBS. 84 Table 1. Summary of nucleotide character data. rbcL Mean sequence lengtha 1319 Alignment length 1325 Total included 1324 atpA 1708 1844 1776 1428 200 6.8% Total 3027 3169 3100 2497 351 6.4% a Included characters ParsimonyConstant informativeb 1069 151 Missing datac 5.7% Completely missing sequences (one rbcL; three atpA) and partially complete sequences missing more than 10% of data (four rbcL; six atpA) were not included in calculation. b Parsimony-informative characters were calculated in PAUP. c Missing data column displays the summed percentages of “?”s and “-“s in the matrix; includes completely missing sequences (one rbcL; three atpA, two of which are outside of Adiantopsis). 85 Fig. 1. Three types of lamina architecture in Adiantopsis—palmate (A), pinnate (B), and pedate (C). A. Adiantopsis radiata (Hernandez and Gonzalez 1670, MO). B. Adiantopsis paupercula (Maxon 4239, US). C. Adiantopsis xaustralopedata (Hickey et al. 01-63, MU). 86 placeholder taxa representing rest of hemionitidoids Adiantopsis clade 1 clade 2 Fig. 2. Phylogeny resulting from Bayesian analysis of combined dataset (rbcL and atpA). Numbers above branches are posterior probabilities. Numbers below branches are maximum parsimony bootstrap support values (≥50%). Tree is rooted with two species of Notholaena (representing the notholaenoid clade that is sister to the hemionitidoids); other outgroups are representatives from across the hemionitidoids. 87 Fig. 3. Geographic distribution of Adiantopsis species optimized onto the Bayesian 50% majority rule consensus cladogram from the analyses summarized in Fig. 2. 88 Fig. 4. Phytogeographic regions of Adiantopsis species optimized onto the Bayesian 50% majority rule consensus cladogram from the analyses summarized in Fig. 2. 89 Fig. 5. Morphological characters traditionally used to circumscribe Adiantopsis— spore ornamentation (A) and paired adaxial carinae (B)—optimized onto the Bayesian 50% majority rule consensus cladogram from analyses summarized in Fig. 2. Note that echinate spores and carinae fall out at slightly different places on the phylogeny, and both characters exclude early diverging lineages of the genus. See text for discussion about morphological characters that can be used for the entire genus as circumscribed in the current paper. 90 Fig. 6. Lamina architecture of Adiantopsis optimized onto the Bayesian 50% majority rule consensus cladogram from the analyses summarized in Fig. 2. Adiantopsis radiata is implicated by morphology in three instances of reticulate evolution (arrows), as the paternal donor for the sterile hybrid A. xaustralopedata and the putative allotetraploids A. pedata and A. pentagona. 91 “reesii” clade Fig. 7. Phylogram of Adiantopsis from one of 5040 equally parsimonious trees from the maximum parsimony analysis of the combined dataset. Note: In some MP trees, Species novum 7 and Species novem 9 have 0, 1, 2, or 3 nucleotide substitutions that differ; therefore, we are maintaining them as distinct species (supported by morphological distinctions) even though they show up in this phylogram as a polytomy. 92 A B 9 species Mexico & Central America Caribbean 7 species 3 species Northern Venezuela ­Colombia 4 species Guianas 3 species 2 species Northern Andes Amazonia 2 species Cerrado & associated dry areas Southern Andes 12 species Coast 5 species Fig. 8. (A) Estimated species diversity of Adiantopsis in various phytogeographic regions of Central and South America. Regions are modified and redrawn from Gentry (1982). ‘Coast’ and ‘Amazonia’ were left out of the phytogeographic regions optimized on to the tree in Fig. 4 since species in those areas were also present in the ‘Cerrado’, with which they were merged to simplify data presentation; some species are found in more than one region. (B) South American-Caribbean ‘generalized tracks’ of transcontinental and transoceanic dispersal of organisms (both plant and animal) proposed by Rosen (1975) to explain present-day distribution patterns of organisms (redrawn from Rosen, 1975); patterns observed in Adiantopsis are consistent with Rosen’s generalized tracks. 93 4 Summary and synthesis This dissertation presents a revision of the Central and South American species of the neotropical fern genus Adiantopsis and presents the first molecular phylogeny to focus on the genus. Hitherto unknown diversity was uncovered in Adiantopsis and a greater understanding of both its biogeography and the origin and evolution of its diverse laminar architectures was achieved. The dissertation has also contributed to the ongoing efforts within the seed-free vascular plants research community (Gastony, 1994; Gastony and Rollo, 1996; Gastony and Rollo, 1995; Gastony and Rollo, 1998; Kirkpatrick, 2007; Prado et al., 2007; Rothfels et al., 2008; Schuettpelz et al., 2007; Windham et al., 2009; Yatskievych et al., 1991; Zhang et al., 2007) to untangle the cheilanthoid clade of the Pteridaceae, a group long known to be replete with paraphyletic genera due to the difficulties of recognizing monophyletic groups when confronted with homoplasious adaptive morphology. Below is a summary of the major findings from this dissertation, paralleling the presentation of research questions and hypotheses from Chapter 1. How many species are in Adiantopsis? What are they? This question was addressed by a test of the null hypothesis that Adiantopsis consists of a single species. This hypothesis was serially rejected as the accumulated morphological and/or molecular data failed to support it. At the onset of the dissertation research, it was generally accepted that Adiantopsis was a small genus of around 13 species. The research revealed previously unrecognized diversity within Adiantopsis. The number of species that we presently recognize approaches forty: twenty-eight accepted species (three of those newly-described; Chapter 2 and in preparation) and an 94 additional ten taxa yet to be described (supported by molecular data, as documented in Chapter 3).3 In all, we have identified 10 palmate species, 3 pedate taxa, and 25 pinnate species—a tripling of the number of recognized species, and of these a third (13 species) are new to science. What is the generic boundary of Adiantopsis? Is it monophyletic? What morphological characters circumscribe the genus? This question was addressed by a test of the hypotheses that Adiantopsis is monophyletic and is a distinct group within the cheilanthoids, meriting generic status. The dissertation research revealed that Adiantopsis as traditionally circumscribed is not monophyletic. Classically, Adiantopsis has been characterized by the combination of echinate spores, golden or golden-red paired carinae (ridges) on the upper side of axes, and distinct pseudoindusia. Species with cristate spores (A. dichotoma (Cav.) Moore and A. regularis Moore) were generally placed in Cheilanthes. Our molecular data indicates that the morphological characteristics often used to distinguish Adiantopsis lead to a circumscription that is overly conservative and excludes several early diverging lineages in the clade that have cristate spores and are either non-carinate or carinate. Additionally, several Cheilanthes species with carinae and echinate spores (C. trifurcata Baker and C. tweediana Hook.) are embedded within the Adiantopsis clade. Adiantopsis will be monophyletic with the inclusion of three Cheilanthes species, for which three new combinations will be made as the result of this dissertation research. The inclusion of non-carinate and cristate-spored species in Adiantopsis provides for a clearly defensible and monophyletic circumscription that receives strong support with both Bayesian and 3 The distinctness of these taxa is supported by molecular data, geographic distributions, and morphological differences. Properly describing and validly naming these ten additional species will require extensive examination of additional herbarium material that will occupy a considerable amount of time following the conclusion of the dissertation research. Three of these species are associated with Adiantopsis chlorophylla, three are palmate species of the Guyana region, two are Central American species associated with the Mexican A. seemannii, one is a Colombian species associated with the Colombian A. propinqua, and one is a Brazilian species that is sister to the Caribbean “reesii” clade. 95 Maximum Parsimony analyses of the molecular data for the two plastid genes used in our study. Morphological characters that can be used to circumscribe Adiantopsis include: large, reddish, pluricellular hairs or carinae on axes; distinct, scarious pseudoindusia that cover one to occasionally two sori; and compound leaves with small, asymmetrical ultimate segments, at least some of which are stalked. Adiantopsis is distinct from the other genera in the hemionitidoid subclade of the cheilanthoids: the compound leaves are essentially glabrous and lack any type of farina, which is present in most Notholaena and Pentagramma species; its sori are discreet and marginal, protected by distinct, scarious, flap-like pseudoindusia, whereas the sori are without indusia and occur along the veins in Hemionitis and Pentagramma or are protected by the recurved margins in Pellaea and Doryopteris (Kirkpatrick, 2007; Mickel and Smith, 2004; Rothfels et al., 2008; Tryon and Tryon, 1982). The laminae of Adiantopsis are pinnate to quadripinnate with small, asymmetrical ultimate segments, whereas the laminae of Hemionitis and Doryopteris are usually undivided, with shallow to deep lobes in Hemionitis, and are cordate, sagittate, or hastate, and deeply pinnatifid or bipinnatifid in Doryopteris (Mickel and Smith, 2004; Rothfels et al., 2008; Tryon and Tryon, 1982). The laminae of Aspidotis are adaxially lustrous and often striate (Mickel and Smith, 2004). What is the history of character evolution in Adiantopsis, particularly the origin and evolution of the diverse laminar morphologies (pinnate, pedate, and palmate)? This question was addressed by a test of the hypothesis that pinnate and palmate laminar architectures within Adiantopsis reflect separate evolutionary lineages within the genus. Optimizing laminar architecture on the molecular phylogeny indicated that the pinnate architecture is pleisiomorphic for Adiantopsis and that the palmate architecture arose just once. The presence of a single, pinnate species, A. lindigii (Mett.) Prantl, within the palmate clade is currently inexplicable to our satisfaction; possible reasons are explored in Chapter 3, but the issue of this surprising placement of the extremely rare A. lindigii may remain unresolved indefinitely. 96 Optimizing other morphological features onto the phylogeny revealed that cristate spores are pleisiomorphic, with a single origin of echinate spores. The variability of the echinate perispores is high between species, ranging from echinulate, to arachnoidechinulate, to echinate; however, there is no apparent phylogenetic structure to this variation in echinate ornamentation (pers. obs.). Carinae arose once in the genus. What is the biogeographic history of Adiantopsis, and how many (if any) colonization events were responsible for the present day diversity of Adiantopsis in the Caribbean? This question was addressed by a test of the hypothesis that Adiantopsis endemics in the Caribbean are the result of a single migration, followed by diversification. This hypothesis was soundly rejected. Optimizing geographic distributions on the molecular phylogeny suggested a South American origin for Adiantopsis and indicated two distinct lineages of Caribbean endemics: A. paupercula (Kunze) Fée and a clade containing A. reesii (Jenman) C. Chr. and five other endemic Adiantopsis species. The widespread taxon A. radiata (L.) Fée also occurs in the Caribbean, indicating a third colonization event. The present-day diversity of Adiantopsis in the Caribbean—9 species—is the result of a minimum of three (probably four) colonization events. A fourth colonization event is postulated to account for the presence of the Cuban endemic, A. asplenioides Maxon. DNA from herbarium specimens of A. asplenioides could not be successfully amplified, precluding this species’ inclusion in the molecular phylogeny. However, based upon its possession of carinae and cristate spores, we hypothesize that A. asplenioides would likely be positioned near the “tweediana/dichotoma” clade (Chapter 3, Fig. 5). Hence, it appears that there have been no fewer than four migrations of Adiantopsis into the Caribbean. Biogeographic studies of ferns on oceanic islands often reveal that island fern diversity is the result of multiple colonization events that are followed by limited divergence (Driscoll and Barrington, 2007; Geiger and Ranker, 2005; Schneider et al., 2004). Short branches in the “reesii clade” (Chapter 3, Fig. 7) indicate little sequence divergence among those six species, suggesting either a relatively recent radiation in the Caribbean or, perhaps, ecological constraints to be slow-evolving. 97 Has adaptations to different habitats played a role in diversification within the genus? This question was addressed by a test of the hypothesis that Adiantopsis diversification has been tied to radiation into different phytogeographic regions. Adiantopsis has a strictly neotropical distribution, with species occurring in the Caribbean, Mexico and Central America, and throughout much of South America, as far south as the northern regions of Uruguay. Optimizing phytogeographic regions on the molecular phylogeny indicated that the Cerrado and associated dry areas are home to several early diverging lineages in Adiantopsis. The genus appears to have undergone a subsequent radiation into distinct phytogeographic regions (Chapter 3, Fig. 4), thereby suggesting that evolutionary innovation has played a role in its diversification, particularly within the palmate clade whose members are found in the highest number of different regions. The relatively recent colonization in the Caribbean that gave rise to the “reesi” clade represents an apparent rapid radiation, giving rise to six endemic species that have diverged very little at the molecular level, at least for the two plastid markers used in this study. Neotropical biodiversity & the value of revisionary studies This dissertation on Adiantopsis can be viewed as one “story” about evolution and diversification in the neotropics, that portion of the Americas between the Tropic of Cancer and the Tropic of Capricorn that includes southern Mexico, the Caribbean, Central America, and most of South America. The neotropics are physiographically complex with three continental plates—North American, South American, and Caribbean—contributing to the landmass (Clapperton, 1993). They are also geologically dynamic, shaped by volcanism and plate tectonics, with recent geological phenomena including the uplift of the northern Andes from about 5 mya and the bridging of the Isthmus of Panama about 3.5 mya (Gentry, 1982). Perhaps most fascinating, however, is that the neotropics are one of the most diverse bioregions in the world (Gentry, 1982). The neotropics contain more biodiversity than other tropical areas combined, and approximately 37% of all flowering plant species 98 are found there (Thomas, 1999). While the significant biodiversity of the neotropics is unquestioned, the evolutionary origin of the present day diversity has been and continues to be largely debated (Bush, 1994; Colinvaux et al., 2001; Rull, 2006). One hypothesis attributes the high species richness to gradual accumulation of species over time in stable equatorial climates (Colinvaux and De Oliveira, 2001; Stebbins, 1974), but an alternative hypothesis attributes it to recent and rapid radiation, possibly in response to climate fluctuations during the Pleistocene (last ~2 million years) when distributions of rainforest species may have contracted to refugial pockets during periods of cooler and/or drier climates brought on by cyclical glacial events, leading to speciation through allopatric differentiation of populations in separate refugia (Haffer and Prance, 2001; Prance, 1982; Prance, 1996; Van Der Hammen, 1991). Molecular phylogenies have much to bring to the debate on whether the high biodiversity in the neotropics came about gradually— which we could expect would be evidenced by numerous nucleotide substitutions among species and a well-resolved phylogeny—or whether present-day diversity arose in recent and rapid bursts of diversification—which we expect would be evidenced by short branches and a poorly-resolved phylogeny (Richardson et al., 2001). A meta-analysis of data from DNA molecular dating of >1400 neotropical species shows that species have been added in continual fashion since the late Eocene-early Oligocene (~39 million years before present), with no bursts of diversification (Rull, 2007; Rull, 2008), although other studies offer support for recent radiations (e.g., Kay et al., 2005; Richardson et al., 2001; Scherson et al., 2008). It is estimated that about 25% of the flowering plants in the neotropics are still unknown to science (Thomas, 1999). It is likely that the fern flora is similarly under estimated. In comparison to the relative completeness of fern floras for temperate regions, there are significant gaps in our knowledge of the fern flora in the neotropics (Thomas, 1999), particularly in the biodiversity hotspots of Colombia and Brazil (Smith, 2005). Although numerous fern floras and checklists have been published in recent years, and more are presently in preparation—for example, a fern flora for Bolivia— progress is slow, and individuals with the expertise required to produce them are in short supply (Smith, 2005). Furthermore, floras are dependent on monographic or revisionary studies, and few of these works are currently being written (Smith, 2005). This is an 99 unfortunate trend, considering that monographic and revisionary studies are critical to achieving a complete inventory of plant diversity in the neotropics, where so many species—seed plants and seed-free plants alike—remain unknown to science (Thomas, 1999). The type of information provided by monographic and revisionary studies— including descriptions of species, means of distinguishing between them, distributions, and habitat specifications—are essential for making appropriate conservation strategies (Grimes, 1998; Mori, 1992). Furthermore, the nature of these studies provides the researcher with the ability to discern species limits from natural variation present within species and permits clarification of the taxonomy of a group, including reducing some superfluous names to synonymy (Thomas, 1999). When molecular data can also be brought to bear, even cryptic species and those for which convergent evolution or phenotypic plasticity has complicated species delimitation based upon morphology alone can be adequately resolved. It seems that a full appreciation of the diversity of the neotropics as well as an understanding of the mechanisms that gave rise to its high biodiversity is such a big undertaking that accomplishing it is only possible by examining small pieces of that diversity—as is done, for example, in monographic or revisionary work—and bringing what is learned to the ever-growing body of knowledge about the neotropics and the organisms that live there. This dissertation research contributes another small chapter to the growing volume of knowledge on neotropical biodiversity, and with it comes another glimpse into the patterns and processes that contribute to the complexity and richness of the neotropics. Adiantopsis turned out to be—like so many neotropical genera—more diverse than originally suspected. Its evolutionary history includes evidence of: adaptive radiation into new habitats and distinct phytogeographic regions; long distance dispersal across barriers, including several migrations into the Caribbean; reticulate evolution, demonstrated in the pedate species; polyploidy, inferred in several species based upon spore and/or guard cell length; interesting patterns of morphogenesis, demonstrated in the three types of laminar architecture; and both a gradual accumulation of species, best demonstrated in the well-resolved palmate clade, and at least two recent and rapid radiations, indicated by the poorly-resolved “reesii” clade in the Caribbean and a handful 100 of species associated with A. chlorophylla (Sw.) Fée. A suitable question might be: if all this is true for Adiantopsis, how much more so for the neotropics in general? What we don’t know about the biodiversity in the neotropics is almost unimaginable. With so many species still unknown to science (Thomas, 1999), and with the current threats to biodiversity, including fragmentation of habitats, deforestation, and the climate crisis, we may lose the opportunity to learn so many “stories” like the one for Adiantopsis presented here. It is vitally important to discover and record as many of these stories as we can. 101 LITERATURE CITED Bush, M. B. 1994. Amazonian speciation: a necessarily complex model. Journal of Biogeography 21:5-17. Clapperton, C. M. 1993. Quaternary geology and geomorphology of South America. Elservier, Amsterdam. Colinvaux, P. A., and P. E. De Oliveira. 2001. Amazon plant diversity and climate through the Cenozoic. Palaeogeography, Palaeoclimatology, Palaeoecology 166:51-63. Colinvaux, P. A., G. Irion, M. E. Rasanen, M. B. Bush, and J. A. S. N. de Mello. 2001. A paradigm to be discarded: Geological and paleoecological data falsify the HAFFER and PRANCE refuge hypothesis of Amazonian speciation. Amazoniana 16:609-646. Driscoll, H. E., and D. S. Barrington. 2007. Origin of Hawaiian Polystichum (Dryopteridaceae) in the context of a world phylogeny. American Journal of Botany 94:1413-1424. Gastony, G. J. 1994. Phylogenetic relationships among genera of cheilanthoid ferns (Pteridaceae: Cheilanthoideae) inferred from gene sequence data. American Journal of Botany 81:120. Gastony, G. J., and D. Rollo. 1996. Congruence of phylogenetic reconstructions in cheilanthoid ferns based on nucleotide sequences of maternally inherited rbcL and biparentally inherited nuclear ribosomal ITS. American Journal of Botany 83:125. Gastony, G. J., and D. R. Rollo. 1995. Phylogeny and generic circumscriptions of cheilanthoid ferns (Pteridaceae: Cheilanthoideae) inferred from rbcL nucleotide sequences. American Fern Journal 85:341-360. Gastony, G. J., and D. R. Rollo. 1998. Cheilanthoid ferns (Pteridaceae: Cheilanthoideae) in the southwestern United States and adjacent Mexico: A molecular phylogenetic reassessment of generic lines. Aliso 17:131-144. Geiger, J. M. O., and T. A. Ranker. 2005. Molecular phylogenetics and historical biogeography of Hawaiian Dryopteris (Dryopteridaceae). Molecular Phylogenetics and Evolution 34:392-407. 102 Gentry, A. H. 1982. Neotropical floristic diversity: phytogeographical connections between Central and South America, Pleistocene climatic fluctuations, or an accident of the Andean orogeny? Annals of the Missouri Botanical Garden 69:557-593. Grimes, J. W. 1998. Before the floras-monographs. Australian Systematic Botany 11:243-249. Haffer, J., and G. T. Prance. 2001. Climatic forcing of evolution in Amazonia during the Cenozoic: On the refuge theory of biotic differentiation. Amazoniana 16:579-607. Kay, K. M., P. A. Reeves, R. G. Olmstead, and D. W. Schemske. 2005. Rapid speciation and the evolution of hummingbird pollination in neotropical Costus subgenus Costus (Costaceae): evidence from nrDNA ITS and ETS sequences. American Journal of Botany 92:1899-1910. Kirkpatrick, R. E. B. 2007. Investigating the monophyly of Pellaea (Pteridaceae) in the context of a phylogenetic analysis of cheilanthoid ferns. Systematic Botany 32:504-518. Mickel, J., and A. R. Smith. 2004. The pteridophytes of Mexico. New York Botanical Garden Press, Bronx, NY. Mori, S. A. 1992. Neotropical floristics and inventory: who will do the work? Brittonia 44:372-375. Prado, J., C. D. N. Rodrigues, A. Salatino, and M. L. F. Salatino. 2007. Phylogenetic relationships among Pteridaceae, including Brazilian species, inferred from rbcL sequences. Taxon 56:355-368. Prance, G. T. 1982. A review of the phytogeographic evidences for Pleistocene climate changes in the neotropics. Annals of the Missouri Botanical Garden 69:594-624. Prance, G. T. 1996. Islands in Amazonia. Philosophical Transactions of the Royal Society of London B Biological Sciences 351:823-833. Richardson, J. E., R. T. Pennington, T. D. Pennington, and P. M. Hollingsworth. 2001. Rapid diversification of a species-rich genus of neotropical rain forest trees. Science (Washington D C) 293:2242-2245. 103 Rothfels, C. J., M. D. Windham, A. L. Grusz, G. J. Gastony, and K. M. Pryer. 2008. Toward a monophyletic Notholaena (Pteridaceae): resolving patterns of evolutionary convergence in xeric-adapted ferns. Taxon 57:712-724. Rull, V. 2006. Quaternary speciation in the Neotropics. Molecular Ecology 15:42574259. Rull, V. 2007. On the origin of present Neotropical biodiversity: a preliminary metaanalysis about speciation timing using molecular phylogenies. Orsis 22:105-119. Rull, V. 2008. Speciation timing and neotropical biodiversity: the Tertiary-Quaternary debate in the light of molecular phylogenetic evidence. Molecular Ecology 17:2722-2729. Scherson, R. A., R. Vidal, and M. J. Sanderson. 2008. Phylogeny, biogeography, and rates of diversification of New World Astragalus (Leguminosae) with an emphasis on South American radiations. American Journal of Botany 95:10301039. Schneider, H., S. J. Russell, C. J. Cox, F. Bakker, S. Henderson, F. Rumsey, J. Barrett, M. Gibby, and J. C. Vogel. 2004. Chloroplast phylogeny of asplenioid ferns based on rbcL and trnL-F spacer sequences (Polypodiidae, aspleniaceae) and its implications for biogeography. Systematic Botany 29:260-274. Schuettpelz, E., H. Schneider, L. Huiet, M. D. Windham, and K. M. Pryer. 2007. A molecular phylogeny of the fern family Pteridaceae: Assessing overall relationships and the affinities of previously unsampled genera. Molecular Phylogenetics and Evolution 44:1172-1185. Smith, A. R. 2005. Floristics in the 21st century: Balancing user-needs and phylogenetic information. Fern Gazette 17:105-137. Stebbins, G. L. 1974. Flowering plants: evolution above the species level. Harvard University Press, Cambridge, Mass. Thomas, W. W. 1999. Conservation and monographic research on the flora of Tropical America. Biodiversity and Conservation 8:1007-1015. Tryon, R. M., and A. F. Tryon. 1982. Ferns and allied plants with special reference to tropical America. Springer-Verlag, New York. 104 Van Der Hammen, T. 1991. Palaeoecological background: Neotropics. Climatic Change 19:37-48. Windham, M. D., L. Huiet, E. Schuettpelz, A. L. Grusz, C. Rothfels, J. Beck, G. Yatskievych, and K. M. Pryer. 2009. Using plastid and nuclear DNA sequences to redraw generic boundaries and demystify species complexes in cheilanthoid ferns. American Fern Journal 99:128-132. Yatskievych, G., G. J. Gastony, and M. D. Windham. 1991. Generic concepts in Cheilanthoid ferns. American Journal of Botany 78:150. Zhang, G., X. Zhang, Z. Chen, H. Liu, and W. Yang. 2007. First insights in the phylogeny of Asian cheilanthoid ferns based on sequences of two chloroplast markers. Taxon 56:369-378. 105