Annals of the Missouri Botanical Garden ot VG olume 95 umber Volume 95, Number 1 April 2008 Annals of the Missouri Botanical Garden The Annals, published quarterly, contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, independent reviewers. Authors should write the Managing Editor for information concerning arrangements for publishing in the Annals. Instructions to Authors are printed in the back of the last issue of each volume and are also available online at www.mbgpress.org. Editorial Committee Victoria C. Hollowell Scientific Editor, Missouri Botanical Garden Beth Parada Managing Editor, Missouri Botanical Garden Allison M. Brock Associate Editor, Missouri Botanical Garden Barbara Mack Editorial Assistant, Missouri Botanical Garden Patricia M. Eckel Latin Editor, Missouri Botanical Garden Ihsan A. Al-Shehbaz Missouri Botanical Garden Gerrit Davidse Missouri Botanical Garden Roy E. Gereau Missouri Botanical Garden Peter Goldblatt Missouri Botanical Garden Gordon McPherson Missouri Botanical Garden Charlotte Taylor Missouri Botanical Garden Henk van der Werff Missouri Botanical Garden For subscription information contact ÁNNALS or THE Missouri BOTANICAL GARDEN, Yo Allen Mar- keting & Management, P.O. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2008 1s $170 per volume U.S., $180 Canada & Mexico, $205 all other countries. Four issues per volume. The journal Novon is included in the subscription price of the Annals. annals@mobot.org (editorial queries) http: //www.mbgpress.org THe ANNALS or THE Missouri BOTANICAL GARDEN (ISSN 0026-6493) is published quarterly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. Periodicals post- age paid at St. Louis, MO and additional mail- ing offices. POSTMASTER: Send address changes to ANNALS OF THE MissoURL BOTANICAL GARDEN, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. The Annal. bstracted indexed in AGRICOLA (through 1994), APT Online, BIOSISO, CAB Ab- stract/Global Health databases, ingenta, ISIO databases, JSTOR, Research Alert®, and Sci Search®. The full-text of ANNALS OF THE MISSOURI BOTANICAL GARDEN is available online though BioOne?M (http:// www.bioone.org). O Missouri Botanical Garden Press 2008 The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and their environment, in order to preserve and enrich life. " — 6&9 This naner meets the reauirements of ANSI/NISO 739 48-1992 (Permanence of Paper). Volume 95 umber | DOS messour E" APR 11 GARDEN LIBR " of | the Annals Missouri ,(; Dotanical Garden PALEOBOTANY IN THE POST- GENOMICS ERA: INTRODUCTION” ? William L. Crepet? and Maria A. Gandolfo' Today, addressed in the aftermath of the genomics revolution significant biological questions are being o O O in ways that suggest a remarkable understanding of life is now within our grasp. Post-genomics techniques are ja powerful, intriguing, and full of promise. In addition to the tools grouped under bioinformatics, we are now seeing the emergence of ancillary but complementary and informative technologies that are increasingly important, including breakthroughs in optical imaging such as confocal microscopy and various microarray methodologies. Such new techniques are helping us understand the cascade of information provided by genomics and are already yielding important insights into plant biology. These include advances in under- standing cell structure and function, control of gene expression, and intercellular communication that will inform our understanding of plant development and thus, ultimately, homologies. Genomics data are also vital to understanding the relationship between the organism and environment, and a better understand- ing of this relationship combined with better under- standing of homologies will provide new insights into plant evolution, especially in the context of another posl-genomics—era advance—comparative studies that allow estimations of phylogenetic relationships among extant taxa based on DNA sequence data using various algorithms. Phylogenetic context is a powerful tool for understanding many aspects of the history of life including those involving paleoecology and past dis- tributions. Phylogenies based on gene sequence data are also increasingly being used to date key evo- lutionary events through various molecular clock- based models. Dating key events in the past is critical to understanding the diversity and nature of life today. Thus. in the post-genomics era, many tools of molecular biology have implications for understanding aspects of plant life that were once illuminated exclusively by fossil evidence or inferred from com- parative studies of extant taxa based on morphology and anatomy. Now, in the excitement associated with the proliferation of genomics data, there is such reliance on various molecular genetic-based approaches and so much hope assigned to future applications of such methodologies that fossil data are in some danger of either being ignored or, because many fossil identi- "This and the nine a s "s : ial follow are the proceedings of the symposium “Pal: Mond in the Post-genomics Era, was held at Botany 200 August 2006 was supporte res in stud by e m obotanical Sectie 2 W e wish lo . at California State l niversity—Chico, vhich o, California, U.S.A. The symposium of the Botanical Society of ee ca. hank Victoria Hollowell, Beth peas and Allison Brock of the Missouri Botanical Garden Press for their editorial ON in He ee of these symposium proceedings. = IR epartment of Biology, Cornell University, wic dion ILedu. PP Ithaca, New York 14853, U.S.A. Author for correspondence: Bailey Hortorium, Department of Plant Biology, Cornell University, Ithaca, New York 14853. U.S.A. be 10.3417/2007180 ANN. Missouri Bor. GARD. 95: 1-2. PUBLISHED ON 11 APRIL 2008. fications in the literature have been inadequately justified, misapplied when they are used at the critical interface between genomics and paleontology. Thus, there is the possibility that the full value of fossil data in the context of genomics and post-genomics studies may not be fully realized at the very moment when these data might have the most significance. This symposium issue is dedicated to emphasizing the value of the fossil record by discussing issues likely to be important in the post-genomics era. There are too many areas where fossils have importance to include here, but certain issues and subjects are covered in some depth. In the first paper, Crepet further considers the value of the fossil record in general and specifically with respect to understanding angiosperm history. Gandolfo et al. discuss the importance of fossil selection in calibrating analyses using various liming models. Nixon considers the conflict between timing suggested by the fossil record versus timing inferred from the application of molecular clock—based dating using a specific example. Such conflicts are often attributed to flaws (i.e., incompleteness) in the fossil record. Burnham considers the accuracy of the fossil record directly and addresses some widely held notions about lag time while pointing out imperfections based not on the record itself, but on the level at which it has been studied. modern data is an fossil and Integration o important. recurring element in the post-genomics era, and in this symposium this issue is approached in a variety of ways. Hermsen and. Hendricks consider difficulties and benefits associated with integrating from modern taxa with molecular data morphological/structural data from fossil and extant sequence organisms in phylogenetic analyses aimed at eluci- dating character evolution. Endress reveals a creative and very interesting application of comparative floral organogenesis in extant taxa that has the potential to amplify the information retrievable from the fossil record, while Rothwell et al. illustrate interesting Annals of the Missouri Botanical Garden potential in the interface between the fossil record and molecular biology. They provide fossil evidence relevant to the regulation of cambial activity by auxin through time and in various clades. In so doing, they add yet another dimension to paleobotanical studies. Biogeography and paleoecology are areas of [e] e) | J | ter, particular and traditional importance in paleontology and ones eritical to understanding the evolution of life. There is no neontological substitute for the O information made available through such investiga- tions, but modern data and analytical techniques are vital to informing these studies. Tiffney takes on biogeography in te a P, the context of post-genomics-era methodologies using phylogenetics, applications of "physiological uniformitarianism.” and key fossil data in concert to evaluate certain phylogeographic hypotheses and to identify new ones. DiMichele and Gastaldo provide a comprehensive treatment of the interface between neoecological techniques and paleoecology. They illustrate how reciprocity between advances in paleobotany/paleoecological methods and [om modern ecological theory/methods holds great. poten- tial for a better understanding of the changes in Hutchinson's “ecological theater” through time and thus promises to impact our understanding of his “evolutionary play.” There are several themes emerging from this collection of papers. It is clear that new methodologies are eritical to increasing the value of fossil evidence and, reciprocally, that fossil evidence is important in exploiting many of these advances. Furthermore, it is likely that fossil data will be increasingly important as new informative techniques emerge that we cannot now imagine, just as we could not predict the value of fossils in the context of today's techniques 30 years is also very clear that the value of the fossil ago. | record, now so informative in so many areas, is limited by the accuracy of fossil descriptions, stratigraphic settings, and identifications. In the end, it is one of the most basic aspects of paleobotany that counts the most. THE FOSSIL RECORD OF ANGIOSPERMS: REQUIEM OR RENAISSANCE?! ABSTRACT the increasingly complete and informative fossil record of William L. Crepet? 1e fossil record of angiosperms has more potential than ever for contributing to the resolution of major questions in the evolution of the flowering plants due to the better understanding of the significance of leaf and pollen records and | »ecause of owers. Nonetheless, the record has fallen short of its potential (and of its potentially synergistic value) because, although it is better understood than ever, there are still problems in identifying fossils" affinities that have not been fully resolved and that have major implications angiosperm evolution with either molecular clock-based models or minimum-age node mapping. This issue significance with respect to early angiosperm radiation, where more carel i i unrealized with respect to determining liming in i is of particular ul studies of existing specimens seem to have value for increasing our comprehension of floral evolution and homology, potentially in the context of strides in understanding MADS-box genes. Subjective methods with typological overtones are still often used in identifying fossils, even though phylogenetic context is available. Identification using phylogenetic context, among other things, does not obscure relative character changes within monophyletic groups. does not lend itself to facultative interpretation of affinities to suit outcomes of various models, and, thus, does not impede our understanding of angiosperm evolutionary history. Nonetheless, a reasonably good fossil record of angiosperms is emerging from 1 evaluated, reveals an interesting and possibly informative patte is the rapid radiation of angiosperm taxa that are now unusually diverse around two particular times in geologica Early Tertiary. Possible reasons foi Turonian and ey words: Cretaceous, diversification. Hymenoptera, typological. YI S ie combined efforts of many laboratories and, when carefully rn of flowering plant evolution. One of its most striking aspects — history: the aese intervals of rapid radiation among angiosperms will be discussed. insect. pollination, MADS box, molecular clock, Tertiary, As the post-genomics era unfolds, the role of the fossil record in addressing broad questions related to plant evolution has come under scrutiny. While there are certain aspects of fossil data that are unique and cannot be replaced by comparative studies of extant taxa, the primacy of fossil evidence in what have been key areas exclusive to the domain of paleontology has been challenged. For example, the implications of o the principal areas of its importance, has often come into - fossil evidence in timing evolution, one of conflict with predictions derived from molecular clock-based models that sometimes produce dramat- ically contrasting results. (Wikström et al., 2001: Crepet et al., 2004; Graur & Martin, 2004: Sanderson record’s flaws are often invoked directly or by implication to explain the discrepancies in favor of the model’s results (Burnham, 2008), or there may be a tendency to look too hard for fossils that might corroborate the predictions of such models with the danger that standards of identification might be relaxed in so doing (Crepet et al., 2004). Another manifestation of conflicts between the results of molecular-based models and the fossil record involves dramatic examples where the taxonomic affinities assigned to fossil taxa have been reinterpreted because of conflicts with predictions based on models, despite the uncertainties associated with the models themselves and without supporting analyses that might favor the suggested change in affinities (Nixon, 2008). These instances reveal trends that could diminish the impact of fossil evidence in the face of alternative methodologies before the potential value of fossil evidence is fully realized and before the accuracy of such alternative means of timing has been fully evaluated. With the possibility of controversy in such a critical area as timing, the full potential of fossil evidence in understanding evolutionary history cannot be appreciated because the potential informa- tive value of plant fossil data depends on correlative studies that, in turn, depend on timing (i.e., models explaining evolutionary events can be evaluated for This research was supported in part by National Science Foundation grant DEB-0108369 to WLC and Kevin Nixon. The author thanks Maria Gandolfo, Karl Niklas. and Kevin Nixon for helpful comments and corrections and is indebted to Victoria C. Hollowell of the Missouri Botanical Garden for her careful editing and excellent suggestions. The author acknowledges David L. Dilcher for illustrations in Figures 1 and 8, and Michael Rothman for his reconstructions in Figures 9-11. He also thanks Wolf B. Frommer, Plant Biology, The Carnegie Institution at Stanford for valuable discussion. * Department of Plant Biology, Cornell University, Ithaca, New York 14853, U.S.A. wlelGcornell.edu. doi: 10.3417/2007065 ANN. Missouri Bor. GARD. 95: 3-33. PUBLISHED ON 11 APRIL 2008. Annals of the Missouri Botanical Garden S consistency with the fossil record) or on paleobiogeo- graphic/paleoecological context that, again, requires an This collection of symposium contributions aimed accurate time frame. paper is one in a ex- amining the fossil record of plants to illustrate how it might continue to be of value in the face of dramatic new developments in molecular genetics and their applications (e.g.. Downie & P: almer, 1992: 1999: l.. 2000), examples of new applications of fossil data in Parkinson and et al., Chaw et ¢ at providing = context. Due to obvious space limitations. it is by no means a comprehensive treatment of all aspects of the integration of fossil and modern data, but is intended to underscore the importance of fossils and their un- realized potential at a time when excitement generated by alternative, powerful techniques may obscure the value of fossil data and ultimately diminish its con- questions of interest and tributions to addressing importance. This would eventually have the ironic result of limiting the impact of important breakthroughs in molecular genetics by removing a valuable set of tools from interpreting and using such approaches. Certainly, the fossil record does have uncontested and exclusive domain in its potential for revealing taxonomic history in an ecological or a paleobiogeo- graphic context. Thus. the fossil record empowers an studies of evolution/development entire class of where, in addition to timing and relationships. ecological context is also important for understanding Jablonski, 2005). molecular genetics have the potential to illuminate key clarifying genetic control mechanisms associated with evolution (e.g.. While studies of events in evolution by identifying genes and timing in development (in addition to the taxonomic distribution of these genes). critical aspects of the relationship between evolution and development would be invisible to studies that focused exclusively on contemporary species (e.g.. correlations between environment or key environmental components [s.l.] and character innovations [e.g.. Crepet, 2000; Gri- maldi E Engel. 2005: Jablonski, 2005: Labandeira. 2006]. Thus. separate from its historical importance in allaining our present unc erstanding of the history and evolution of life, the fossil record has unique properties that are important to a deeper understand- he nature of the ing of evolution depending on particular fossil record and how it has been inter- preted and applied. Tur Fossi RECORD VALUE OF FOSSIL EVIDENCE The fossil record provides information relevant to understanding plant evolution and diversity in a ) Determining the pattern of taxa appearances through number of important areas including the following: ( time—this allows tests for consistency with various evolutionary ecological hypotheses via comparisons with hypothesized related parallel events. (2) Deter- mining minimum times of appearance of characters or taxa—especially important in the context of deposi- tional environment. (3) Identifying morphoclines— transitional states that form a sequence linking two disparate homologous conditions where. transitional types may be difficult to hypothesize (verifiable in phylogenetic context). (4) Identifying mosaics or links among taxa, such as extinct taxa that include combinations of characteristics now found in different taxonomic groups and, therefore, are possibly useful in linking (i.e., missing links through the combination) or in checking existing phylogenies for (i.e... if groups consistency the combination of characters verified in a fossil taxon cannot occur on a phylogenetic tree, then that tree is incorrect). These taxa. as informative as they might be, are invisible in baffled their them without including cases where typologists, by unique nature, attempt to identify them in phylogenetic context. This approach has obscured the observation of evolution within lineages and opens the door for controversy about affinities by making determinations necessarily. to at least some degree, subjective. (5) Observing distribution changes through time. allowing for a check on paleobiogeo- eraphy patterns and their implications for explaining 1974). (6) Exam- modern distributions (Raven & Axelrod, Reconstructing paleoecologv/climatology. (7 ining the evolutionary implications of depositional (9) Observing adaptive radiations and mass extinctions environments. (8) Identifying extinct diversity. e.g.. turnover rates). As implied above, the plant fossil record, which includes only morphology and structure with very occasionally preserved. chemistry, might be consid- ered one of two independent sets of data relevant to plant history, evolution, and relationships. The other is based on extant taxa and includes comparative ala. morphology. structure. and molecular genetics « With respect to analyzing the significance of these separate sets of data, complementarities now exist in various combinations of fossil and modern data through phylogenetic analyses. in assessments of congruence checks between fossil mosaics and phylogenies generated without fossils. by comparisons between apparent character transformations based on fossil intermediates and those implied by phylogenetic context or even by underlying molecular genetics, and by general comparisons between patterns based on comparative analysis of extant taxa with their implied sequences of evolution and patterns observed in fossil Volume 95, Number 1 2008 Crepet 5 The Fossil Record of Angiosperms history. Analyses that include some synthesis between these data sets aimed at understanding evolution have changed progressively with the development of new methodologies in phylogenetic analysis (e.g.. Nixon 1999, 2003) and There are, however, enduring first principles upon H advances in molecular genetics. which effective interpretations of fossil data depend. The identification of the fossils (Gandolfo et al.. most fundamental of these is accuracy in 2008). Without such accuracy, none of the potential of fossil taxa can be realized. Timing. a critical example. — ollows from the pattern of appearance of taxa that are This enormous value beyond simply determining minimum accurately identified. pattern in time has times of appearance of taxa or characters or as references for molecular clock-based models (Crepet. 1996; Magallón et al., 1999: Crepet et al.. 2004: 2006). Friis et al., IDENTIFICATION OF FOSSIL TAXA With respect to maximizing the potential informa- tive value of the fossil record in any of the areas mentioned above (or in any of its usual applications), is the seemingly simple and straightforward process E taxon identification that has caused a great deal of controversy in interpreting the fossil record of plants. particularly, but not exclusively, angiosperms, and has also prevented the full realization of the potential informative value of the plant fossil record. For vascular plants, the fossil record is further compli- cated by the fact that entire organisms in reproductive state are seldom preserved with component organs in Often, particular taxa are known as fossils. connection. some e! = organic only parts « This phenomenon (i.e., that of missing morphological/structural charac- ters) has significant implications for the identification that inferences and insights that may of fossil taxa are the foundation for other be derived from fossil evidence in the areas noted above (Nixon & Davis, 1991; Nixon, 1996). The potential associated with a pattern of evolution of taxa set in chronology and ecological context cannot be realized unless fossil identifications are precisely accurate or precisely as accurate as possible because affinities based on subsets of characters are, after all, only best estimates. Reliable identification of fossils is more important than ever in light of widely accepted phylogenies based on molecular data that imply a sequence in timing (relative appearances) due to their hierarchical nature (i.e., creating a possible subliminal tendency to look for siis might be expected in the fossil record f rigorous as mentioned above). e! Facultative interpretations « fossils’ affinities, easier to do with less methodologies for fossil taxon identification and inspired, whether subliminally or not, by phylogenies based on molecular sequence data, would be harmful to the scientific method because such interpretations might effectively eliminate the independence of fossil evidence as a corroborative instrument by linking expectations to outcomes in the realm of identities. Thus, the significance of the fossil record would be diminished, and the test for congruence between two independent sources of data, which should ideally strengthen our understanding of evolutionary history. would be compromised. Proof against this eventuality would be a careful, open methodology for taxonomic affinity determination—a methodology that, ideally. might take into consideration and compensate for challenges in describing and classifying fossil taxa whose character associations reflect evolutionary change (that is, those different from their closest modern relatives). In other words, a methodology that would allow for precise placement of extinct taxa. PHYLOGENETICS AND IDENTIFICATION Cladisties or phylogenetics has made it possible to identify fossils through placement in phylogenetic 1989: Nixon. 1996). may be identified context (e.g.. Thus, precisely and with objectivity, even though they would be difficult to tables of characters in cases where the fossil taxon's Donoghue et al., affinities of extinct. taxa determine through comparisons of characters do not exactly correspond to the characters found in any possible modern relatives. In addition, the evolutionary significance of such fossils can be evaluated as changes in character-complements within monophyletic groups, which can be trackec because relationships among these fossil taxa are revealed through phylogenetic context. Phylogenetic context also focuses investigators on taxon-defining synapomorphies. IDENTIFICATION METHODOLOGIES: HISTORICAL PERSPECTIVE Previous to and leading up to the advent of phylogenetic methodology, there has been an evolu- tion of identifying methodologies throughout the 20th century that is perhaps most easily illustrated in that evolution of identifying methodologies began with angiosperm paleobotany. One can posit the the criterion of general similarity in what were considered taxon-defining characters. and, in the absence of phylogenetic methodology. proceeded to what has amounted to an essentially typological method of identification. involving gross character comparisons via tables. The latter approach, initiated the need for a in response to more objective identification methodology, has been used historically Annals of the Missouri Botanical Garden in conjunction with increasingly rigorous standards for | While not ideal. it is still despite the that the weaknesses of such an approach are now obvious describing | fossils. sometimes employed fact (by depending on congruence of character associa- tions for taxonomic placement, the approach cannol deal effectively with extinct taxa characterized by unique combinations of characters without subjective interpretation). In essence, this method could result in un — Yoehorning fossil taxa into particular extant. group- ings defined by strict character associations (typolo- This process by denying or minimally, by obscuring, one obscures the gy). methodology evolutionary of its fundamental aspects—character association changes. This tabular character-matching method then, by necessity, also involves subjective. assess- ments of affinities in cases (most depending on the taxonomic level and age of the fossil taxon in question) where the match is imperfect. Thus, these methods have limited the impact of fossil evidence in plants, particularly in angiosperms, given that entire suites of characters are usually not analysis. context provides an In contrast, phylogenetic alternative methodology for identifying extinct taxa that do not precisely conform to the character complexes now found in their closest modern relatives 1992; Crepet & Nixon, 1998b). In addition to representing an optimal means (e.g.. Herendeen & Crane, of placing fossil taxa, precision identification through phylogenetic analysis is also the best way lo cope with inevitable differences of opinion that characterize the scientific community by reducing the subjectivity inherent in other methodologies. These are inevitable in the face of the limited numbers of characters available in fossils and, frustratingly, often involve fossils of unusual potential signif- icance. Nonetheless, such disputes, while relatively common and potentially distracting, are importan he scientific method and, if conducted e correlates of within the boundaries of phylogenetic methodology. are resolvable within the limits of the relevant fossil evidence. Indeed, phylogenetic context 1s the dif- regress to only reasonable method for resolving such ferences because alternatives represent a methods abandoned years ago by systematists because of the well-known flaws in fundamentally subjective approaches (e.g., angiosperm systematics in 1960). and. to reiterate, determination of affinities by comparisons of characters through tables is inherently subjective because there is no logical way to deal with incongruities that are the norm and not the exception when dealing with fossils (at any but the often less informative or relevant, higher, and taxonomic levels). av ail: ible for level of Two Case STUDIES: CRETACEOUS MoNocor FLOWERS AND AN EARLY CRETACEOUS ANGIOSPERM There are many examples of differences of opinion on the affinities of important fossils. Two, in particular, are appropriate in the context of this paper: (1) reports of certain Cretaceous monocot flowers and (2) reports of a unique Early Cretaceous angiosperm. These illustrate the differences between what might be characterized as subjectively based disagreements on affinities and ones founded on conflicting results of alternative phyloge- netic analyses. The monocot example is a case of a challenge to an assignment of affinities involving a number of Cretaceous. taxa, exemplified by Mabelia Gandolfo Nixon & Crepet (Gandolfo et al., 2002; Figs. 2A, B. 10A), that have been placed with modern Triurida- ceae on the basis of phylogenetic analysis. Although the triuridaceous affinities have been effectively universally accepted by monocot specialists (e.g.. Chase, 2004) and experts in modern Triuridaceae Rudall, 2003) on the placement of these fossils. Friis et al. (2X [ema basis of phylogenetic JOGO) question such affinities. Their challenge is based on differenc- 1 pollen. morphology between the one modern es 1 taxon of Triuridaceae for which pollen morphology is known and the fossil taxa, and on the presence of a staminal column in one of the fossil taxa. They consider the staminal column found in some of these ANITA clade and fossil taxa as a more generally dicol 2000) [n fact, the character (Friis et al., not one found modern Triuridaceae. character assem- blage of the fossil triuridaceous taxon Mabelia reflects remarkable congruence with those of taxa found in the 2002). However, its pollen is monosulcate and, thus, more modern family Triuridaceae (Gandolfo et al., generalized. than the monosulcate-derived pollen found in the only modern taxon of Triuridaceae (out of 34) for which pollen has been described (Furness et al.. 2002). Furthermore, staminal columns are actually well known in the modern family Triuridaceae (Maas € Rubsamen, 1986: fig. 14h: Maas, 1988: figs. 3, 10a). Suggestions of alternative affinities with more generalized ANITA dicots have not been supported by an evaluation/criticism of the original matrix of Gandolfo et al. (2002) on which triuridacean affinities have been assigned. Thus, this difference in interpre- (2006) does not include. nor is it based on. a contravening phylogenetic analysis that tation by Friis et al. would support the assertion that the fossil has more generalized non-triurid affinities (Friis et al., 2000). In the absence of an alternative analysis based on accurate characterization of modern and fossil characters, assertions of alternative affinities become subjective and an extension of the character-matching Volume 95, Number 1 2008 Crepet The Fossil Record of Angiosperms r phenetic approach used widely before phyloge- netics was available. The difference in interpretation seems based on only the alleged incongruence in characters between the fossil and modern taxa. These alleged instances of incongruence are misplaced given that at least one of the characters actually does occur in modern triurids and that the pollen differences, considering their nature and the severely limited extent to which pollen morphology is known in modern members of Triuridaceae (i.e., similar pollen could ultimately be found to exist among the 33 taxa of living triurids for which pollen morphology is now unknown), are in any case consistent with generally accepted trends in pollen evolution as well as with the © well-documented tendency for pollen morphology t ag floral morphology as documented numerous Cretaceous taxa (see Friis et al., 1994; Zhou al., 2001; 2003; Crepet et al., 2004). Just as original identifications of fossils via this temporally Hermsen et al., method of character matching have been historically difficult because they do not accommodate evolution- ary change within (monophyletic) lineages, criticisms of this type are similarly based on methodologies that =— do not allow for the possibility of differing rates o evolution among characters found in particular clades. So, in this case, according to the criteria suggested by Friis et al. (2006), pollen would have had to evolve at the same rate as the other floral characters in order for the fossils to be defined as triurids (also assuming that no monosulcate pollen types will be found in any of the 33 species of extant triurids for which pollen is now unknown). In this regard, it is striking that some orchids might not be defined as orchids under such a methodology because they have monads or tetrads instead of polyads (Pacini & Hesse, 2002: A good example of a disagreement on affinities taking place within the framework of phylogenetic analysis is the Lower Cretaceous fossil taxon Archae- Dilcher, implications as to early angiosperm floral morphology fructus Sun, Zheng & Zhou, a fossil with and perhaps to the origin of flowers (Figs. 1A, B, 8A). It is an unusual angiosperm, having naked, unisexual that organized in flowers, with affinities either to all of 1998, 2002) or, with a different interpretation of the reproductive organs as to the Nym- phaeales (Friis et al., 2003). This difference of opinion reproductive structures are apparently not the angiosperms (Sun et al., secondarily simple due to reduction, within the context of 2002 and The challenge to the original has taken place entirely phylogenetic analysis (compare Sun et al., 2003). placement of Archaefructus at the base of all other Friis et al., angiosperms was based on a reanalysis of the Sun et al. (2002) matrix following the addition of another existing nymphaeoid taxon Cabomba Aubl. (Friis et al.. 2003). The methodology was explicit and can be — objectively evaluated. It happens that there was an error in coding leaf characters in Cabombaceae in the Friis et al. (2003) paper (Cabombaceae were coded as having dissected leaves when, actually, they have both dissected submerged leaves and entire-margined floating leaves; Nixon, 2006). When corrected and the modified matrix reanalyzed, there is no support for the within Nymphaeales inclusion of Archaefructus — Nixon, 2006). While this particular challenge to the affinities of a fossil taxon may have been based, at least partially, on a coding error, the point is that the explicit nature of phylogenetic context allowed for an evaluation of the original identification by peers and also permitted an evaluation of the criticism. This is a good example of a healthy difference of opinion that is supported by critical evaluation, just as the original identification was supported by such evaluation. Nonetheless. the issue of affinities of Archaefructus is not fully resolved. Limited numbers of characters are known for the fossil taxon, keeping its actual [2 affinities controversial. Recent assertions that some of the reproductive branches of similar, presumably closely related taxa bear both carpels and stamens (an interpretation that would make the reproductive branches inflorescences) complicate the issue (Ji et al., 2004). However, this interpretation is not con- vincingly supported by the illustrations purporting to reveal such bisexual fertile axes, which seem instead to reflect several layers of plant organs compressed together. This is an important issue that can only be resolved through further empirical study of the fossils in question followed by critical analysis. It focuses on — the importance of first principles in paleobotanica studies (1.e., that the fossil taxon's characters should be explica understood and clearly demonstrable based on the study of the fossil). BROADER RELEVANCE TO ANGIOSPERMS The issues above play out in important ways that are current with respect to a set of evolutionary mysteries that hold a commanding position among those concerning plant history. The circumstances surrounding the radiation of angiosperms comprise some of the most addressed and notably unresolved questions in plant evolution. There have numerous articles on the many facets of angiosperm evolution, and many of these have contributed to our understanding of the questions associated with angiosperm history. These questions include: (1) ancestry—still famously unresolved with underlying homologies (to be revealed by molecular genetics studies y yet-to-be discovered fossil evidence) nr being key to its resolution, and its correlate, the origin Annals of the Missouri Botanical Garden Figure 1. A, B. and SZ0916). —A. Note that stamens subtend carpels on th subtende d : bracts. there are no bracts or obvious floral enve PNE axes Ar a, Dileher & Crane (University of Florida collection numbers UFI5 frui Tw "a Cenomanian of Kansas. U.S.A, > Creek. ska. —E. flowers from Rose Creek. Nebra: Note the ri s symmetry ME the d ua nt nectary ring and partitions separating the carpels (UFI permission of I . Dilcher. Florida Natural History Museum. of angiosperm-defining characters in ecological con- text: (2) within angiosperm relationships—long prob- lematic due to convergence and homology problems and the domain of molecular now systematics (a possible escape route from convergence issues bul perhaps one not vet fully resolved): (3) the dramatic rise of angiosperms to numerical species dominance and possible reasons for it: and (4) the history of angiosperm distribution patterns. There is a great deal the e area of of enthusiasm about prospects for homology resolution in molecular genetics (e.g.. Theissen et al. 2000. ete.), especially given that morphological ee may never have actually Mature carpels are helically arranged on elongate axes " F Note stamens and carpels with separate styles Archaefructus fertile branches (Nanjing Institute of Geology and Paleontology collection numbers SZO917 * branch on the left; and that while the fertile branches are tamens along the fertile land UF15703-2300): Compressed dicot 703-4026). —F. A-F arts subte nding the « carpe ‘ls or 106- 308 and stigmas (UF T5 5113-3429). with existed, account for which might the difficulty i identifying angiosperm outgroups. These questions have been addressed using new New methods continue to provide hope that they will be applications of fossil evidence. analytical successfully resolved. In posing the question as to how the potential of fossil evidence may be realized with respect to angiosperm history, it should again be kept in mind that reliable and appropriate fossil evidence is required, i.e., the identifications and ages of fossils must be accurately determined, and. likewise. the nature of the fossil evidence must be appropriate to the questions being addressed (e.g., questions regard- Volume 95, Number 1 2008 Crepet The Fossil Record of Angiosperms ing flowers for evolution of pollination). In addition. methodologies must appropriately compensate for weaknesses in the fossil record (an area itself that is subject to question) in, for example, timing (minimum lom age node mapping or molecular clock-based models or in phylogenetic placement of the fossils. With respect to the questions above, it is interesting En to review changing approaches to the ^ angiosperm fossils because the fossil record has fallen — short in some areas. One of the issues affecting our understanding of angiosperm evolution in historical context has been the unreliability of the fossil record. Studies of angiosperm fossils have been plagued by poor identification methodologies and limited by the incomplete nature of the fossil record. Nonetheless, there has been a historical trend toward increasingly rigorous methods for angiosperm fossil identifications resulting in what has now become a much more informative fossil record. With respect to angiosperm fossil identification, an important revolution of sorts was already underway in the late 1960s. New criteria were being applied in the identification of angiosperm leaves by Hickey (1971), Dilcher (1974), and Hickey and Wolfe (1975), while Doyle (1969), building on the careful stratigraphic palynology of Brenner (1963) and employing careful pollen character analysis developed with Walker (Walker € Doyle, 1975), LeThomas (1981), and others, began a comprehensive look at angiosperm pollen radiation (see also Muller. 1981). Thus, the groundwork was being laid for a better — understanding of angiosperm history, an area o E considerable confusion prior to that time because o rampant misinterpretations of angiosperm fossils that often erred in the direction of assigning modern es affinities to fossils where such assignments were unwarranted. And, as noted above, methodologies evolved to include careful comparisons of character complexes that were more reliable than previous methods but that have aforementioned limitations in the face of better approaches afforded by phylogenetic — context. With respect to the incomplete nature of the fossil record, phylogeneties and minimum age node mapping provide a means of filling in gaps as do various molecular clock-based models for determining timing. THe VALUE or FLOWERS Given that the background of extensive leaf and pollen records and their increasing value as criteria ey for fossil identifications have become more rigorous, the advent and continuing investigations of fossil flowers have added significant critical mass to the angiosperm fossil record. Furthermore, flowers add several important dimensions to the record. They embody many taxon-defining characters and reveal study of aspects of past angiosperm reproductive biology, an aspect of angiosperms that has often been invoked to explain their relative success (Grant, 1949; Raven, 977: Regal, 1977; Burger, 1981; Crepet, 1984: Crepet & Friis, 1987; Friis & Crepet, 1987; Eriksson & Bremer, 1992). A great deal of hope has been — associated with the prospect of the recovery of a reasonably accurate fossil record of flowers as a remarkable flower record has been unfolding. While historically sporadic studies of fossil flowers have been a characteristic of angiosperm paleobotany (e.g.. 1886). describe fossil flowers began with investigations of Conwentz, sustained efforts to find and compressed Tertiary flowers (e.g., Crepet et al., 1974, 1975; Crepet & Dilcher, 1977; Crepet & Daghlian. 1980: Crepet & Taylor. 1985, 1986; Crepet & Nixon. 1989a. b). followed in short order by investigations of lower Middle and Late Cretaceous fossils, represented by a variety of preservational modes (Dilcher et al., 1976: Tiffney. 1977; Basinger & Dilcher, 1984; Crane & Dilcher, 1984; Dilcher & Crane, 1984; Crepet & Friis, 1987: Friis & Crepet, 1987; Crepet € Her- endeen, 1992; Crepet et al., 2004; Friis et al., 2006). As a result, a pattern of evolution of flowers has now — — emerged that can be reviewed with the fossil records of other plant organs as well as fossil insects that are potential pollination vectors. This collective “biotic” approach can provide insights into the evolution of reproductive biology in the angiosperms. In addition, this record of flowers, in the contexts of the carefully vetted leaf, fruit, and pollen records, provides : comprehensive record of angiosperms that can be compared to the hypothetical sequences of taxon appearances implied by phylogenies based on extant laxa and to hypotheses of timing in angiosperm history derived from various molecular clock—based models (e.g.. Crepet et al., 2004). In addition, the flower record provides insights into the evolution of pollination mechanisms in angiosperm history and further illuminates relative rates of evolution among different plant organs, thus increasing the value of the fossil records of dispersed non-floral organs. This kind of information also provides insights into angiosperm reproductive strategies during key periods during their diversification. TIMING ANGIOSPERM HISTORY An important aspect of interpreting the significance of the history of the angiosperm fossil record and one that was once an uncritically accepted correlate of that record is the timing implied by it. Insights into questions that may be addressed by fossil evidence depend on correlation consistency with various ecological/evolutionary hypotheses, distributional Annals of the Missouri Botanical Garden changes (Tiffney, 2008), and character evolution and its significance, and these depend on accurate timing. ru While timing is implied by the fossil record, the fossil record is far from perfect, and assessing the accuracy of timing implied by the fossil record in the context of molecular clock—based models has become an yel controversial issue (see al., 2004: number of interesting, Important, Crepet et al., 2004; 2008). determining timing with fossil data. Nixon. Sanderson € There are a approaches to First are direct inferences from raw fossil data (i.e., data accepted from literature), second are inferences from the carefully screened (for reliability) fossil record, and finally, inferences based on various models, notably clock- each of these approaches minimum age node mapping or molecular based models. Of course, has strengths and weaknesses (e.g. uncritically O 2 s evaluated literature can be misleading, critically evaluated literature may leave gaps in taxon histories, and models fill in gaps, but methodologies are in transition and their reliability controversial), but whatever the methodology employed, a more accurate and complete fossil record would and will enhance our understanding of timing. Thus, even with respect to establishing accurate timing, and in the face of a radically improved fossil record of angiosperms, the major questions surrounding angiosperm history are only partially resolved. Tue Fossi. RECORD oF ANGIOSPERMS In light of the discussion above, a review of the carefully vetted fossil record of angiosperms (1.e., fossils found to have synapomorphie characters defining the taxa to which they have been assigned — to have had their affinities determined directly in the context of phylogenetic analyses) that integrates new discoveries does reveal some interesting aspects of their relevance tlo history that have major unresolved issues in angiosperm evolution. There has been a recent analysis of the fossil angiosperms based on the criteria of synapomorphy or phylogenetic context (Crepet et al., 2004) and also a more recent review of the fossil record of angiosperms (Friis et al., 2006) that takes these criteria into account to some degree in its summary of angiosperm history. Friis et al. (2006) also raise a number of issues worthy of discussion. and these will be or have been considered throughout this paper (e.g.. monocot fossil. record, » following discussion of the angiosperm == ete). Th record (taxon appearances and character appearances) is based principally on flowers but is congruent with other fossil organs as noted in the relevant discus- sions. This discussion incorporates results of earlier fossil record of angiosperms. reviews of the early There is emphasis on Turonian (89—93.5 million years the of angiosperms as = ago [Ma]) and younger angiosperm fossils i context of the earlier fossil recor summarized and discussed in previous papers (Friis el al.. 1986, 1997: Crepet et al., 2004; Friis et al., 2006). Key Early Cretaceous taxa, however, are discussed al length in the contexts of the overall fossil record of flowering plants and of the Turonian reports (when these are deemed to be more reliable evidence than the earlier reports of the same taxa) and how they have been investigated. IMPORTANCE OF THE Fossil FLOWER RECORD OF THE "TURONIAN As with new floral evidence from all Cretaceous and Tertiary intervals, Turonian data have supplied a wealth of information of angiosperm history and, given generally accepted patterns of angiosperm diversifi- calion, represent the angiosperms at an important moment in their history. These fossils are remarkable for their quality of preservation (e.g.. Staedler et al., 2007) and document surprising diversity in angio- sperms at a relatively early time (depending on the taxon). In addition, they reveal that certain characters These Turonian fossils and taxa had yet to evolve. include numerous initial appearances, changing our view of timing in angiosperm history in significant ways, especially with regard to the rosid-asterid radiation and also, more broadly, with respect to adaptations for insect pollination. EARLIEST EVIDENCE OF TAXA AND CHARACTERS REPRESENTED BY TIMING: TURONIAN FossiLs The following taxa appear first in Turonian deposits (consistent with the Angiosperm Phylogeny Group [APG] Il system except as noted): Nymphaeaceae s. Monocot megafossils, putative Magnoliales, Car- yophyllales, Hamamelidaceae, Altingiaceae, Hydran- Brassicales Malphigiales (Clusiaceae), Myrtales, Melastomataceae, and Ericales (Ericanae) (Crepet et al., 2004). lleaceae, Fagales, Rosaceae, geaceae, The Fossi EVIDENCE NYMPHAEACEAE The Nymphaeaceae comprise an interesting family placed at the base of angiosperms by molecular on leaves systematics analyses and reported based and other flowers from other sediments including sediments earlier than Turonian deposits (Friis et al., 2001, 2000). unsettled in the sense that there is some level The fossil record of Nymphaeaceae is Volume 95, Number 1 2008 Crepet The Fossil Record of Angiosperms controversy around each of the reports of nymphaea- ceous flowers and inflorescences. The possibility that an unusual nymphaeaceous taxon is represented by Archaefructus (Fig. 1A, B) has been discussed above. and it is too controversial to include this taxon in the fossil record of Nymphaeaceae. More plausible reports of Nymphaeaceae are based on mesofossil evidence that includes a putative flower of Nymphaeaceae from 100 Ma sediments of the Early Cretaceous (Friis et al., 2001). str. does not withstand rigorous phylogenetic analysis However, its assignment to Nymphaeaceae s. using the same matrix employed by Friis et al. (2001: 1999) and including only those characters available in the fossil as opposed to adapted from Les et al., those inferred characters figured by Friis et al. (2001: Gandolfo et al., 2004). Upon phylogenetic analysis using the Les et al. (1999) matrix originally employed by Friis et al. (2001), results indicate that affinities are equally compatible with Illiciaceae as well as with other angiosperm families, suggesting that Microric- toria. Gandolfo, Nixon & Crepet (Figs. 2C, D, 10B), (2004), is the only unequivocally strictly nymphaeaceous fossil from the Early phaeaceous fossil taxon described by Friis et al. described by Gandolfo et al. Cretaceous. Nonetheless, the putative nym- (2001) has significant implications that are relevant to the age of Nymphaeaceae because of its equal compatibility with llliciaceae, which suggests the node that includes oo Nymphaeales and Illiciales can be dated at 11: | which is consistent with phylogenetic dum aes et al., 2004; Friis et al., 2006). Thus, the fossil nymphaeaceous taxon Micro- victoria, while the earliest confirmed record, probably does not reflect the actual age of the family, which is projected at least 113 Ma by minimum age node 2004). Another perspective on the age mapping (Cre et et al., pping i of Nymphaeaceae that challenges both reports is derived from applica- — tion of molecular clock-based models (Yoo et a 2005). The challenge is based not on the morphology of fossils, but on the fact that they dramatically alter the outcomes of model-based angiosperm dating if they are included as calibration points (Yoo et al., 2005). The conflict with model-based analyses brings up issues involving the accuracy of the applications of at least some of these models and is discussed in detail in Nixon (2008). MONOCOTY LEDONS Flowers of Turonian-age Triuridaceae have already been discussed above. They exist in the context of an interesting, relatively sporadic, and somewhat contro- versial fossil record. Based on minimum age node mapping, Crepet et al. (2004) projected an origin for monocots of no later than 98 Ma, consistent with some model-based projections (Magallón & Sanderson. 2002). Despite the projections of an Early Cretaceous origin, the accuracy of the Lower Cretaceous fossil record of monocots (Doyle, 1973) has recently been 1998, 2000; et al., 2004), but a recent report of distinctive pollen challenged (Gandolfo et al., Crepet pm provide valid Lower Cretaceous (Friis et With respect to traditionally cited monocot fossils of Early (2000) make the case that the common Early Cretaceous of Araceae may evidence of the monocots al., Cretaceous age, Gandolfo et al. angiosperm fossils, leaves, and pollen are not sufficiently divergent from basal ANITA clade dicots in characters that have been preserved to allow their unequivocal identification as monocot remains. With respect to dispersed palynomorphs, it is not clear that monosulcate pollen with reticulate micromorphology and a gradient in lumina size is restricted to monocots. or even to angiosperms (Zavada, 1984). from the Cretaceous (Turonian) are represented by what are generally agreed to Monocots Je flowers representing Triuridaceae (see discussion above; Figs. 2A, B, 10A), a family of saprophytes with very simple and presumably highly reduced vegetative habit. Inasmuch as they nest with modern triurids in phylogenetic analyses (Gandolfo et al.. 2002), these fossils suggest that the saprophytic habit had evolved within monocots by the Turonian and, given the presumably derived nature of the saprophytic habit in monocots, are indicative of an earlier origin of monocots consistent with timing suggested by minimum age node mapping and some molecular clock—based models (Bremer, 2000). Mono- — cots are not well represented in progressively younger Cretaceous deposits. Zingiberaceae fruits and leaves are found in Campanian deposits, and Musaceae are known from petrified fruits in the Deccan Inter- trappean series (Hickey & Peterson, 1978; Friis et al., 2006). Palms appear in the Campanian, and pollen similar to modern grass pollen a appears in the Maastrichtian (Crepet et al., 2004: 2006), originated earlier than in- —. riis et al., suggesting that grasses dicated by the first megafossil remains of leaves and florets/pollen in the Lower Eocene (Fig. 7A, B; Crepet & Feldman, 1991). Furthermore, Prasad et al. (2005) have reported fascinating grass and palm remains from the Deccan Intertrappean series, con- sistent with reports of Cretaceous palm pollen (Harley, 1996; Harley & Zavada, 2000). T ossil record is better in Tertiary deposits, but some — 1e monocot — no fossil record (e.g.. 1995; major groups have little or Daghian, 1981; 2000). orchids; Herendeen & Crane, Gandolfo et al., 12 Annals of the Missouri Botanical Garden Figure 2. Charcoalified flowers from the Turonian of New ed V. Cornell University Paleobotanical collection pumber | CUPC re aos Flowers of Triuridaceae showing six tepals and three stamens with connective extensions CUI (CUPC1260). —C, D. Microvictoria (Nymphaeaceae) illustrating the morphology and arrangement of floral parts EN conform with be of Ens m m Lindl. in the Nymphaeaceae (CUPC1475). oF in Cre O & Nixon (Magnoliales) illustrating the tepal-bearing floral cup and numerous spirally arranged carpels (CUPC1175). — . Detrusandra Crepet & Nixon (Magnoliales) showing floral cup with incurved stamens and st ele and spiri a arranged carpels within CUPC1 188). Volume 95, Number 1 2008 Crepet The Fossil Record of Angiosperms MAGNOLIALES Fossil flowers of apparent magnolialean affinities are a relatively common component of the Sayreville Turonian angiosperm flora. Interestingly, there are a greater number of magnolialean taxa represented by flower fossils in these sediments than by dispersed pollen (Brenner, unpublished palynological data). This observation raises questions about relative rates as of character evolution in Magnoliales, suggesting that pollen character evolution lags floral morphology. a phenomenon also observed in other taxa from these E Turonian sediments. Thus, the pollen record by itself may not reveal the full extent of Turonian or earlier magnolialean diversity because dispersed grains representing different taxa may have too many common characteristics to be distinguished. This phenomenon also raises questions about the repre- sentative nature of the dispersed angiosperm pollen flora in general in the context of overall patterns of angiosperm evolution. Among putative magnolialean taxa from Sayreville, two, representing extinct lineages, have been de- scribed and have a common characteristic: a cupular receptacle that is also shared by numerous unde- scribed fossils from the same sediments (Crepet & Nixon, 1998a, unpublished data; Fig. 2E-C). remain the only two representatives of Magnoliales to These and, — be identified in phylogenetic context (Fig. 2E-C while clearly extinct taxa, suggest a minimum age for the Magnoliales of 98 Ma using minimum age node mapping consistent with the compression fossils reported by Crane and Dilcher from Kansas in the USA: verified phylogenetic context (Dilcher & Crane, 1984; Crane & Dilcher, 1984; Figs. 1C, D, 8B, C). Very interesting of entire compressed plants from However, these have not been new discoveries Brazilian Aptian—Albian deposits have been com- Fy pared with Magnoliales. but assessments of their affinities have been based on comparisons of char- acters arranged in tables, a methodology discussed above that suffers from being, essentially, a phenetic approach to identification with inevitable subjective elements (Mohr & Eklund, 2003). The possible weak- nesses of this mode of identification are particular- ly evident in the instance of these Aptian-Albian magnolialean flowers, where comparisons extended to purported affinities with extant Eupomatiaceae are based on the presence of glandular tufts on what are characterized as staminodes in the context of suile of an additional generalized characters. The affinities of these fossils seem in need of further review, given that modern Eupomatiaceae have a distinctive cupular floral receptacle (in contrast to these fossils) and a number of distinctive features that are either not preserved in these fossils or have not been elucidated by the techniques employed by the authors (Mohr & Eklund, 2003). Thus. the characters that are reported in tables and illustrated there do not seem to convincingly restrict these fossils to the modern taxon Eupomatiaceae. angiosperms | including Generalized Cretaceous these putative Magnoliales are a particularly good example of a critical group of available fossils that should be carefully analyzed in phylogenetic context because it is among this general class of fossils (i.e.. Karly Cretaceous angiosperms with conduplicate carpels) that valuable insights might be gained into due to their early angiosperm radiation. However, many common features, full understanding of the pattern of evolution of these taxa and its implica- likely to. be assessments of affinities that do not involve. phy- that taxonomic lions is not gained from uncritical logenetic analyses and tend toward forcing extinet taxa into. extant categories, thus potentially obscuring information that might other- wise reveal evolutionarily significant patterns. This is an area that needs attention, but the fossil rec- that the than the conservative estimate provided by Crepet et al. (2004). ord suggests Magnoliales will be older CALYCANTHACEAE The earliest taxon assignable to the Calycanthaceae based on identification in phylogenetic context is Jerseyanthus Crepet, Nixon € Gandolfo (Figs. 3A. B. 9A, B: Crepet et al. 2004). This taxon is distin- guished by unusual staminodes or B; Staedler et al., stamen connective extensions that were discovered in 2007) and remarkable branched- the fossil before reports of less elaborate but similar connective extensions in modern taxa (Staedler et al.. 2007). However, there is an earlier report of a fossil member of the Calycanthaceae, Virginianthus Friis, Ek 1994), but this fossil 1 caly- and, Pedersen & Crane (Friis et al.. taxon is more generalized than modern canthoid taxa and, among other things, has mono- sulcate pollen consistent with a trend in pollen morphology evolutionarily lagging behind floral mor- phology that has been observed in other taxa discussed in this article. Thus, the presence of more generalized pollen does not necessarily remove this taxon from Calycanthaceae, but. upon phylogenetic analysis including all of its characters, Virginianthus appears in an unresolved position. basal to the remaining Laurales or as sister to the entire Laurales while Jerseyanthus excluding the Calycanthaceae, nests with the monophyletic Calycanthaceae (Crepet et al., 2004). Annals of the Missouri Botanical Garden Charcoalified flowers from the Figure 3. staminodes (A) and elaborate connective extensions (B) Caryophyllales showing we Tl-p reserved ovules within the « campylotropous nature. — inflorescence showing stamens within a tepal cup (F, G). (CUPCIIO staminodes CARYOPHYLLALES The from etaceous deposits (Nixon & Crepet, Turonian of de aw Jersey. A. B. CUPCIA8 y ovary (C UPC m RS ). pantoporate pollen grain. from one of the Carvophyllales have only been reported once 1993). These Turonian fossil flowers may represent this order Jerseyanthus (Calycanthaceae) showing stalker . Flower sharing basie characters with modern single well-preserved ovule showing its anthers, F-H. | A floret with stamens removed to illustrate the position of the hamamelidaceous staminate pollen and 3C-E, 9C, D), but there are differences, notably partitions based on very distinctive pantoporate generally similar floral morphology (Figs. retained in the ovary, but these could be the result of the fossils having been preserved at a relatively early Volume 95, Number 1 2008 Crepet 15 The Fossil Record of Angiosperms developmental stage, consistent with the early devel- opmental stages of the well-preserved campylotropous 1993b). Although they possess significant synapomorphies that define ovules (Fig 3D; Nixon & Crepet, the caryophylls, these fossils have not been fully analyzed in phylogenetic context and these results are preliminary. CORE EUDICOTS, SAXIFRAGALES, HAMAMELIDACEAE Exceptionally preserved. staminate and pistillate inflorescences and detached stamens have been discovered from Turonian deposits at the Crossman site in New Jersey, U.S.A. (ca. 88.5-90.4 Ma) in the 1992; floral Raritan Formation of New Jersey (Crepet et al., Fig. 3F-H). The fossils have a combination of and pollen characters found various genera of modern entomophilous and anemophilous Hamameli- daceae and anemophilous Platanus L. (Platanaceae). including a sepal cup, staminal tube, and apparently nectariferous staminodes (these characters are con- sistent with insect pollination). This unnamed fossil (Fig. 3F-H), Hamamelidaceae in phylogenetic taxon while grouping with modern = analysis, is al example of a mosaic of familial-level characters relative to modern taxa and might have been hard to the character table comparison method would result in place outside of phylogenetic context (i.e., subjective assignment). Interestingly, these fossils have apparent staminodal nectaries (based on position and structure) with morphology intermediate between the fossils" functional stamens and modern hamame- lidaceous petals (Fig. 3H). They are positioned within the fossil flowers where petals are found in modern related genera (Fig. 3H), suggesting the possibility that petals originated from such staminodes in the hamamelid lineage. The appearance of the character — complex embodied in these flowers during the late Middle Cretaceous may signal the early stages of the relationship between specialized pollinators, such as bees, and the hamamelid—rosid—asterid lineage of angiosperms, arguably one of the most important 1991; Crepet, 1996) and one reflected in the contemporary events in angiosperm radiation (Crepet et al., record of other angiosperms (Figs. 12, 13) CORE EUDICOTS, SAXIFRAGALES, ALTINGIACEAE Zhou et al. (2001 inflorescences that have perigynous florets with a described fossil — capitate hypanthium, two-carpellate gynoecia, unisexual flow- ers, phyllomes, numerous ovules per carpel, a three- layered carpel wall, and tricolpate reticulate pollen (Fig. 4A, B). have scalariform perforation plates with oblique end Vessels in the inflorescence pedicel walls and scalariform and opposite/alternate intervas- cular pitting (Zhou et al., 2001). fossil As in the case of the — hamamelidaceous inflorescences described above, this fossil taxon includes a mixture of characters relative to the sets found in modern taxa within Hamamelidaceae, but cladistic analysis sug- gests that these fossils represent a basal member of the Altingiaceae that lacks the derived pollen found in extant Altingiaceae and retains the more plesio- morphic tricolpate pollen found in most of the Hamamelidaceae. This phenomenon, i.e., generalized pollen relative to morphology of the flower and inflorescences, is observed in other fossil angiosperm taxa from these Turonian sediments (including the lteaceae; Hermsen et al., 2003). Also, as in the case of the hamamelidaceous taxon above, certain characters of the fossil including phyllomes with stomata, short and straight styles, and small perprolate pollen grains also indicate the possibility of insect pollination. CORE EUDICOTS, SAXIFRAGALES, ITEACEAE Fusainized three-dimensionally preserved flowers from the Turonian Raritan Formation of New Jersey (Divisestylus brevistamineus Hermsen, Gandolfo, Nixon & Crepet and D. longistamineus Hermsen, Gandolfo, Nixon & Crepet) provide another example of fossil taxa whose character complements are not ound in any single extant family (Figs. 4C—E, =t 10C, D). In cases of this nature, tabular comparisons might lead to subjective assignments of affinities through some sort of intuitive character weighting. whereas inclusion of their characters in phylogenetic analysis provides an objective and transparent method for assessing their relationships. These are the kinds = f ana that also reveal trends in character a o yses evolution within monophyletic groups. These fossil species D. revistamineus and lon gistamineus) share characters with both Saxifragaceae and Iteaceae T. n the order Saxifragales. Similarities include a pentamerous perianth, calyx fused below into a hypanthium with free sepal lobes above, haplostemo- nous androecium with stamens situated opposite the calyx lobes, inferior ovary, bicarpellate gynoecium, numerous ovules on axile placentas, conspicuous intrastaminal nectary ring, and capsulate fruits that open apically. Distinctively, carpels and stigmas are fused, while the styles are free, indicating closer affinities with extant Iteaceae, whereas other charac- ters, such as basifixed anthers in D. brevistamineus, tricolpate and striate pollen grains, and anomocytic stomata, indicate closer affinities to Saxifragaceae. Cladistic analyses using molecular and morphological characters, and morphological characters only (Herm- 2003), demonstrate that these sen et al., however, Annals of the Missouri Botanical Garden flowers from the Figure 4. Charcoalified \llingiaceae) with bicarpellate florets Divisesrylus ane elongate stamens (CUPCI3 be —VF. Myrtalean flower is an elongate floral tube. Myrtaceae, hexa ovary (CUPCI 50 nerous flower ` fossils share a more recent common ancestor with lteaceae than Saxifragaceae, thereby making Divises- tylus Hermsen, Gandolfo, Nixon & Crepet the oldest fossil taxon known with clear affinities lteaceae These fossil taxa are distinguished by having tricolporate pollen. (Fig. 4E), a more generalized pollen aperturial configuration than the dicolporate pollen found in modern species of Iteaceae. Turonian of phyllomes, and numerous olesi arpel (A, CUPCIOGA: B. cae). Elongate distally fused style Fusion at the stigmas (CUPC two incurved styles with contorte d corolla lobe 2S, phyllom es alte 'rnaling with stamens, and he "'Xanmerotis. Jersey. —A- D. Vicroaltingia Zhou, Crepet & Nixon CUPC1063). C-E. es are distinctive characters of this taxon. —C. Sepals and 1359). —E. Tricolporate pollen on an iin Y (CUPCI359). . and coiled stamen filaments (CUPCIS10). —G. H. epigynous ROSIDS, MYRTALES Myrtaceae are not definitively represented by fossil evidence until approximately 65 Ma (Crepet et 2004): however, Turonian fossils with a hypanthium, elongate floral tubes, and coiled stamen filaments may be examples of the Myrtaceae (Fig. 4F). These fossils require further study and. despite their apparent Volume 95, Number 1 Crepet 17 2008 The Fossil Record of Angiosperms Figure 5. —A. Astronia So M. ual 9413) in modern Myrtaceae. —B. Fossil f i an inflorescenc 'e preserved in amber from the Turonian of New Jersey, C-I. Charcoali = ied flowers from the Turonian of New Jers —C. Part of a robust in Pod bearing bisexual f Mee florets Nee to those illustrated as seen in Fig. 5E. — = Staminate fagalean floret. —E. Bisexual fagalean floret. —F. Fruit of Juglandaceae. G-I. Paleoclusia. —G. Flower with cune om stigmas and fascicled stamens. —H. Pistil with broken ovary wall revealing anatropous ovules. —I. Mature seeds within the ovary possession of certain synapomorphies, phylogenetic firmed, may resolve disputes about a possible placement to confirm apparent affinities. Gondwanan origin of melastomes by revealing Well-preserved flowers having a set of char- that they originated early enough so that complex acters typical of less-specialized Melastomata- long-distance dispersal models are not needed to ceae (e.g. as in Astronium Jacq; Fig. 5A) are explain their modern-day distribution (Raven & also found in Turonian deposits (Figs. 4G, H. 5A, Axelrod, 1974; Renner et al., 2001: Morley & Dick, 1OE, F). These fossils, if their affinities are con- — 2003). EUROSIDS I, FAGALES Flowers, inflorescences. and pollen (Figs. 5C-E, LLA. B) and another taxon represented by flowers in amber (Fig. 5B) suggest Fagales were extant by the (Nixon & 1994). mesofossils have six tepals in two whorls and are Turonian Crepet, Charcoalified either staminate (Figs. 5D, 11A, B). or bisexual with a well-developed ovary (Figs. 5E. 11B). They are linked by similar floral structure and identical pollen. There are six stamens in each type of flower. and in staminate florets, stamens consist of long filaments and dorsifixed anthers bearing tricolporate that pollen. with exine micromorphology similar te of modern castaneoids and generalized Saxifragales. Stamen filaments are flanked by globose structures bearing trichomes that may be part of a nectary dise or represent separate nectaries (Fig. LIA). Ovaries are tricarpellate trigonous and slightly winged as in 11B). There are three recurved styles with adaxial grooves fruits of certain modern Fagaceae (Figs. 5E, (Fig. 11B). Characters of flowers in conjunction with compact, apparently fleshy bisexual flower-bearing inflorescences (Fig. 5C) suggest a morphocline lead- ing to the more compact inflorescences observed in more recent Fagales (Nixon & Crepet, 1994: Sims et al.. 1998) and. ultimately, to cupules characteris- ic of modern Fagales and are indicative of fagalean affinities that have been preliminarily confirmed by phylogenetic analyses (Gandolfo et al., in prep. The floral characters also suggest insect pollination because the combination of well-developed tepals, neclaries, and pollen of a size incompatible with wind dispersal (Whitehead, 1969) is indicative. of insect pollination. Insect pollination in ancestral Fagales would also be compatible with apparent insect pollination in early Platanaceae as suggested by Friis et al. (1988) and also with floral characters in early Juglandales that also are consistent. with insect pollination, suggesting that extant wind- pollinated hamamelids are derived with respect to pollination biology. EUROSIDS I, JUGLANDALES Excellently preserved fossils of Juglandales from Santonian Coniacean deposits have provided the firsl important insights into the affinities of the Norma- polles-bearing taxa and the history of the Juglandales (e.g., Friis, 1983; Sims et al., 1999; Schónenberger el al.. 2001) and have provided further insights into — pollination biology of early Juglandales. Several inflorescence fragments with associated Normapolles pollen and nuts (Nixon, in prep. Fig. 5F) suggest Juglandales were present by the Turonian. Annals of the Missouri Botanical Garden EUROSIDS I. MALPIGHIALES, CLUSIACEAE Remarkably well-preserved flowers are five-merous with fascicled stamens and cuneiform stigmas and are > almost identical to flowers of modern Clusiaceae 1 every respect down to pollen morphology and resin canals in the floral tissues (Fig. 56-1). These flowers group with modern. Clusiaceae in phylogenetic analysis (Crepet & Nixon. 1998b). EUROSIDS Hl, BRASSICALES (CAPPARALES) A distinctive Turonian taxon is represented by flowers having a combination of characters now found in families in the order Capparales (now included in Brassicales; Angiosperm Phylogeny Group. 2003). This taxon, Dressiantha Gandolfo, Nixon & Crepet (Gandolfo et al., 1998b), has flowers with a well- developed gynophore, a distinctive sepal arrangement, an unequal petal size (bilaterality), monothecal anthers, and a bicarpellate gynoecium, characters n extant families in the order Brassicales found (Angiosperm Phylogeny Group. 2003: Fig. 6D—F). This taxon represents the oldest known fossil record for the Brassicales and has been important in dating in the history of the related extant. taxon Arabidopsts (DC. i Craene Heynh. Interestingly, in a recent paper, de Haston (2006) place the different morphology matrix from the one used by Gandolfo et al. (1998b), the (1991a, b) matrix. The de Craene and Haston matrix and fossil in a who used Rodman was larger and included characters, such as wood anatomical features, not found in the Dressiantha fossil. It may be that the resultant addition of missing values for Dressiantha in their matrix might account he anomalous change in the position of this for fossil taxon from Brassicales to Sapindales. The change is anomalous because, as discussed by de Craene and Haston (2006), there is no conflict between morphological characters found in Brassi- cales and those displayed by the fossil flowers. This is A phylogenetic context in discussing affinities of fossil a good example of the value of the medium taxa. Different outcomes can be objectively analyzed and explained; new questions can be posed in a level of discourse that is both appropriate and potentially illuminating. ASTERIDS, ERICALES OR ERICANAE, THEACEAE Flowers five-merous with three carpels, fascicled stamens, and tricolporate pollen. represent Theales, are diverse in the Turonian deposits of northern New Jersey (Figs. 6A. 11€, D; Crepet et al., 1996), and are also represented in somewhat younger (Campanian) deposits in Georgia (Keller et al., 1990). Volume 95, Number 1 2008 Crepet 19 The Fossil Record of Angiosperms Figure 6. glands. —B. —C. A ricalean flower ÚS stamen tubules. D— irte corolla in bud. — the geniculate filament. ASTERIDS, ERICALES OR ERICANAE, ERICACEAE interesting. and LIE, F) and shortly thereafter in Santonian—Campanian deposits 1985; & Friis, They are remarkably diverse in the Turonian Fossils of Ericanae are first in Turonian deposits (Figs. 6B, C, of Sweden (Friis, Schönenberger 2001). and demonstrate a complex of characters that is now associated with highly specific, usually apid pollinators. Among these, a more general and basal Ericalean taxon has elongate sepals, nectary disc, carpels with separate stigmas, a sympetalous corolla, and two whorls of stamens with stamen awns (Nixon & Crepet, 1993a). related, A more interesting associ- ation of closely more derived ericalean taxa is linked by a basic five-merous floral plan and a set of shared but variable characters (Figs. 6C, l1E, F). These include abaxial sepal glands, sepal Charcoalified flowers from the Turonian of New Jersey. —A. A Th Paleoenkianthus Nixon & Crepet (Ericac eae) flower bud m n to show = corolla, pistil, and stamens within. Dresstantha (Brassicales). —D. Flower . An open flower ba showing the pistil m i p stamens. —F. appear iealean flower with remarkable abaxial sepal E g the bilaterally single stamen illustrating indumentum, haplostemony, dorsifixed anthers, basie pollen morphology, flattened stamen fila- ments, staminodal nectaries, and superior ovaries with five-lobed stigmas having distinctive circular indentations at the junctures of the lobes. However, there are distinct variants representing at least three taxa and probably more. One taxon has anthers without appendages or tubules, elongate stamen filaments, monadinous pollen grains, a non-globose stigma, and fruits that are loculicidal capsules. This taxon is the best known among this complex because flowers are occasionally preserved in inflorescences (Fig. 11E). Flowers are subtended by two bracteoles (Fig. 115). with mar- n the axil of foliar bracts Leaves associated with these fossils are folded, ginal glands similar to those found on the margins this taxon of the sepals of flowers (Fig. 11E, F). representing o Annals of the Missouri Botanical Garden Ii II — ——— A second taxon differs from the first ericad of this five-merous complex in having differently distributed sepal glands, shorter stamen filaments (anthers do not extend beyond the stigma height). stamen filaments — with distinctive stellate trichomes that are also found — on the pistil, anthers that are notched and have tubules (Fig. 6C). and pollen that is apparently in loose polyads connected by short viscin threads. Both fossil taxa group within Ericaceae upon phylogenetic analysis (Crepet et al.. 2004). OVERVIEW OF THE ANGIOSPERM Fossil RECORD IN THE ApriAN—TURONIAN INTERVAL There is a striking pattern in the fossil records of 2 | both floral characters within this interval that has a number of angiosperm taxa and their defining interesting and possibly meaningful implications. Early Cretaceous taxa have a relatively sparse but significant record that has already revealed the timing of the occurrences of major events in angiosperm history including the possible origin of the angio- sperms, the evolutionary assembly of the bisexual flower, the origin of the floral envelope. and the origins of several major angiosperm lineages, includ- ing the eudicots, accompanied by a corresponding Early Cretaceous angiosperm fossils include representatives diversification of pollen types. In addition. of what appear to be extinct taxa basal to existing lineages such as Nymphaeaceae. Existing evidence Early NITA clade + Chloranthaceae + the early tricolpate indicates that Cretaceous. taxa included the lineages: platanoids and Buxaceae (Drinnan et al., 1991: 1994: 20006). The modern-day diversity of taxa that first appear in Pedersen et al.. 1991. Friis et al.. the Early Cretaceous is relatively low, with little indication from known fossils that they might have once been more speciose (Magallón et al. 1999 Crepet et al., 2004; Friis et al., 2006). With respect to flowers of either structurally appropriate for coleopteran pollinators or function. fossil this time were suggestive of generalized kinds of insect. pollination Diptera. Lepidoptera (Micropterigidae). and. conceivably, even that may have involved Coleoptera, early hymenopteran lineages transitional to Apidae (Thein, 1980: Grimaldi, 1999; Grimaldi € Engel, 2005). The uniformly small floral fossils of this age suggest that their pollinators at least included very small insects (Friis et al., 2006). The possibility that some of these taxa may have been wind pollinated has been raised by Dilcher (2000). The lower Late Cretaceous time frame (more or less including the Ceno- the middle of the Cretaceous). manian, Turonian. and Coniacean ages. is important because angiosperms became dominant in species — numbers (inferred from modern-day diversity in clades that appear by that time) during this interval (Fig. 12) and because there was a dramatic coinci- dence of co-occurring factors that may be significant y explaining some measure of the success of the flowering plants. These include: (1) The striking peak 39.000 species today included in the dicot taxa known to have in diversity of angiosperm taxa. There are ca. existed by that time. Although the early record of the monocols is too fragmentary to give an accurate idea Middle certainly of actual monocot diversity during this Cretaceous interval, they were almost present in greater numbers than indicated by the known fossil floral record, and, thus, total angiosperm species diversity was likely to have been even greater than reliable evidence now indicates. (2) The earliest (Cenomanian) appearance of the first cyclic eudicot flower that had both petals and sepals in addition to definite numbers floral parts (Figs. 1E. F. 8D; Basinger & Dilcher, 1984). configurations are considered wide- Also note that eyelic floral parts in other spread if not fundamental to most angiosperm taxa by 2003). (3) The variety of some authors (e.g. Zanis et al., dramatic Turonian appearances of a wide floral morphologies, some now specifically associated with more derived pollinators including hymenopter- ans (Fig. 13). The evolution of some of the latter floral types (e.g. those having a fused corolla or bilateral symmetry, ete.) would logically have depended on the prior occurrence of the relatively simple cyclic flower because these more specialized configurations would been attainable in a single from not have step ancestors with a more generalized floral plan that included at least some multiple spirally arranged 2001). observed in the fossil record (i.e.. floral parts (Endress. Given the timing the rapid appear- ance of derived floral configurations closely following the appearance of the cyclic flower with regular numbers of parts), various floral specializations for more derived pollinators must have evolved rapidly the Cenomanian—Turonian interval, indicating possi- bly rapid adaptation to radiating hymenopteran or other highly derived pollinators. This possibility is consistent with the first appearances in the Turonian of taxa that are now specifically and closely associated with particular and derived Apidae (e.g.. the Clusia- ceae, Paleoclusia Crepet & Nixon, with Meliponini). In addition, this fossil Clusiaceae indicates that even at this relatively early stage of their history, at least some of these fossil taxa apparently had the same his reward structure as their modern relatives (in instance, resins; Crepet & Nixon, 1998b). (4) Timing Middle Cretaceous burst of angiosperm species diversity is of the early fossil record of bees. The observed followed relatively closely by the first fossil evidence Volume 95, Number 1 2008 Crepet The Fossil Record of Angiosperms of bees. The earliest bees are preserved in Upper Albian Burmese amber (Poinar & Danforth, 2006). and another fossil bee that shares many characters with modern Meliponinae is preserved in apparently Upper Cretaceous amber of uncertain age (Grimaldi. 1999: Engel, 2000), but the possibility that this fossil from New Jersey is of Turonian or even earlier in age is supported by the presence of the fossil Paleoclusia noted above from Turonian deposits of New Jersey. U.S.A., with apparent reward system that today ties the Meliponi- (Crepet & Nixon, 1998b). similar morphology and the same nae to the Clusiaceae ( Post-TURONIAN CRETACEOUS DEPOSITS The post-Turonian pre-Maastrichtian fossil record of angiosperms is not as complete as that of the record in the Cenomanian-Turonian interval and, so far. has of detail of subsequent Late Cretaceous angiosperm radiation. As failed to provide a corresponding level it is presently understood, the fossil record documents — a limited (but significant) number of innovations between the Turonian and Maastrichtian. This interval is characterized by the radiation of rosid orders Fagales and Juglandales that appear first in the Turonian, represented by flowers, partial inflores- l., 1998). Inferences from other carefully analyzed plant organs, cences, and fruits (Fig. 5F; Sims et a — while perhaps risky without more detailed knowledge of the relative rates of evolution among different plant organs, do suggest that monocot diversification pro- ceeded throughout the Upper Cretaceous with evi- dence of Zingiberales (fruits and leaves) and palms (fruits, pollen, and leaves) in this interval (Hickey & Peterson, 1978; Daghlian, 1981; Herendeen & Crane, 1995; 2000). remains are generally not preserved in great detail — Gandolfo et al., Maastrictian floral but are significant because the record of monocots suggests that they had radiated considerably by this time, consistent with the occurrence of Triuridaceae in the Turonian and other Lower Cretaceous remains that have been attributed to monocots. Finally, as noted above, pollen with morphological and structural features compatible with the distinctive pollen of -— grasses 1s known from Maastrichtian deposits (Linder, 1986: Crepet & Feldman, 1991), and grass-like leaf remains have been isolated from | Maastrichtian dinosaur coprolites (Prasad et al. 2005). further suggesting that monocots had radiated relatively invisibly during much of the Cretaceous. Interestingly, there is not a great deal of evidence in the pre- Maastrichtian—Late Cretaceous of taxa with the kinds of fruits that are adapted for animal seed dispersal in groups that now are characterized by such dispersal mechanisms (e.g., some Juglandaceae and Fagales). EARLY TERTIARY-—PALEOGENE There is a burst in Tertiary angiosperm diversifica- tion characterized by the radiation of grasses, radiation f Fagaceae and Juglandaceae, appearance and radiation of Euphorbiaceae (Crepet & Daghlian, 1982, unpublished data), legumes (mimosoid, caesal- Fig. 7C-E; Herendeen & 2005), and many asterids piniod, and papilionoid, 992: \sterid I clade (although the fossil record of Crane, ] Lavin et al., n the / reliably identified asterids is relatively sparse: Crepet el al., 2004: Friis et al., 2006). This interval represents a period of modernization for many taxa and is also characterized by the radiation of many pollinating insects including the Macrolepidoptera. The Tertiary 1999: ,rimaldi & Engel, 2005). Perhaps the most dramatic record of bees is relatively good (Grimaldi, ~ innovations in the Tertiary include the brush flower/ inflorescence (mimosoid legumes), the papilionoid legume flower with its pronounced zygomorphy, and the prolonged corolla tube (Fig. 7C—F). In addition, the Tertiary is characterized by the radiation of additional taxa (e.g.. asterids and grasses), which include modern relatives with herbaceous habits. DISCUSSION um There are a number of central questions surround- ing angiosperm history in the context of updated fossil evidence. ANGIOSPERM ORIGIN AND EARLY RADIATION Additional clarification. of the structures of early fossil angiosperms and analysis of their relationships an in phylogenetic context promise to illuminate early angiosperm history and perhaps to reveal certain important homologies, and already pose certain questions. For example, were early angiosperms apetalous? Are sepals de novo organs, are they modified stamens, or do they have multiple origins Did coalescence of reproductive structures borne on and homologues? the flower originate by elongate axes such as those of Archaefructus, and do a number of Early Cretaceous angiosperm flowers represent a morphocline linking Archaefructus-like ancestors with condensed flowers, or is this model too simplistic and not supported in phylogenetic context? Such clarifications, if forthcoming and if conducted within a phylogenetic framework, may complement genetic studies of MADS-box genes and transcription factors known to influence floral development in ways that might reveal homologies between angiosperm reproductive structures and those of non-angiosperms Theissen et al., 2000 Annals of the Missouri Botanical Garden © Figure 7. florets. —B. Grass rhizome with leaf sheath Protomimosoidea flower. —E. Tertiary compression fossils from = Paleocene-Eocene of Tennessee, U. ». Protomimosoidea Crepet & Tas RR C repet & Herendeen, the earliest ee flower in the fossil record. — rass spikelet with two FM ‘avlor ve with dene legume. > w flower (Asterid 1?) illustrating "i the funnelform corolla is well developed at this time. ANGIOSPERM DOMINANCE IN SPECIES NUMBERS Given the scale of angiosperm preeminence i modern plant species diversity and its obvious ecological significance, the matter of angiosperm predominance and success has received a great deal of attention and continues to be an speculation, and controversy Regal, Doyle, 1978; Burger. 1981: Stebbins. 1981; Niklas et al., 1983: Crepet. 1984. 1996; Tiffney. 1984, 1986: Eriksson & Bremer. 1992; Waser, 1998; 1999: Magallón et al.. 1999; Verdú, 2002; 2004; 2004; Friis et al., There is considerable interest. (Raven, 1977: Grimaldi. Feild et al.. Fenster et al., 2005; 20006). doubt that there are separate and Grimaldi € Engel, little possibly complementary reasons for the present level = of angiosperm species diversity (Raven, 1977: Crepet, 1984; Tiffney, 1984; Eriksson & 1992; Tiffney € Mazer, 1995; Verdú. 2002: Fenster et al., 2004). There have been appraisals of the correlations Bremer, area of between diverse angiosperm taxa and derived polli- 1977), ments of the potential significance of animal seed ney, 1984), possible effects of synergy among multiple factors (e.g. Raven, 1977: 1977; Crepet, 1984; Tiffney. 1984: Eriksson & Bremer. 1992). In addition, consideration has been given to factors involved nators (including vertebrates: Raven. assess- lamer) dispersers (Tif and examinations of the Regal, minimizing extinction, as well as to those involved in 1995). various hypotheses linking promoting speciation (Tiffney & Mazer, There have been challenges to angiosperm success to various factors including pollinator specificity and efficiency following the model first proposed by Verne Grant (1949; e.g. Midgely € Bond, 1991; Waser, 1998: Fenster et al., 2004). While there are FO correlations between hyperdiverse taxa and insect iecur (e.g.. orchids 968, 1982), there are s ; or circumstances thal and euglossine bees; Dress obvious alternative may have been responsible for high species diversity Volume 95, Number 1 2008 Crepet 23 The Fossil Record of Angiosperms 4 cm D: L. Dilc igure 8. Rec Hei lions with permission o Archaeanthus flower. within particular monophyletic groups. Sorting out the relative influences remains very much at issue, and, as molecular genetics studies continue to reveal more about genetic control mechanisms and within-species Bomblies & expect the reproductive barriers (hybrid necrosis: Weigel, 2007), we discovery that other factors have had important roles might reasonably in promoting species diversification in angiosperms. Currently, our understanding of the fossil record of angiosperms is much improved and more relia- 5 mm ^. Florida Natural ET Museum. —A. Archaefructus. —B. che . Archaeanthus fruit. —D. Eudicot Mode from Rose Cree ble, and we have a better idea of the pattern of appearances of angiosperm families, floral charac- ters, and possible insect pollinators. Likewise, we have this knowledge in the context of a more accu- rate understanding of overall angiosperm relation- ships (Soltis et al., 2000), such that some questions concerning angiosperm success can be addressed from a fresh perspective. However, in considering the phenomenon of angiosperm diversification through geologic time, it is important to frame the major ques- Annals of the tions before proceeding to examine how recent, more complete, and reliable evidence affects our possible understanding of them. With respect to timing, one can ask two questions about angiosperm preeminence in species numbers: (1) When do angiosperms become dominant in species number (measured by first appearances of modern diverse groups), and what are the circumstances that might have been relevant based on timing? (2) How did even greater angiosperm diversity arise following the initial attainment of angiosperm dominance in numbers of plant species? It is interesting to examine fossil evidence now available with respect to the questions above and to consider whether it rules out any possibilities, such as involvement in preeminent angiosperm diver- the insect sity. due to timing considerations. As noted ii 24 Missouri Botanical Garden C 1 mm Figure 9. Reconstructions of Turonian fossil flowers with permission of Michael Rothman. —A, B. Jerseyanthus (Calycanthaceae). —C. D. ?Caryophyllales. receding summary of the angiosperm fossil record. conservative estimates of timing in first appearances of angiosperm families suggest that the angiosperms had become dominant by the Turonian (ca. 90 Ma). In limes now Fig. 12), fact, families appearing by Turonian encompass approximately 39,000 species while all gymnosperm and vascular non-seed plant taxa combined today only include about 22,000 species. The dramatic coincidence of characters associated with advanced modes of insect pollination by the Turonian (Fig. 13), the similarity of certain Turonian angiosperm taxa to those now specifically associated with such pollinators, and the proximally temporal fossil evidence of the earliest. bees are consistent with insect. involvement in the Cenoma- nian—Turonian proliferation of angiosperm species according to mechanisms discussed by Grant (1949), Figure 10. illustrating a foss Divisestylus longis Volume 95, Number 1 2008 repet The Fossil Record of Angiosperms 25 sa — LA SES AS A v r 277. Jd e a vw " Lie = oO Y y $ UN QM A mi 375 um 315 um Re constructions of Turonian fossil flowers with permission of Michael Rothman. —A. Mabelia (Triuridacae) sil species with a staminal column, a character also found in modern Triuridaceae. —B. Microvictoria. —C. tamineus. —D. Divisestylus brevistamineus. —E, F. Reconstructed fossil Melastomataceae. 26 Annals of the Missouri Botanical Garden Figure 11. Reconstructions of Turonian fossil flowers with permission of Michael Rothman. A, B. Fagalean flowers. —A. Staminate floret. —B. Bisexual floret (this taxon is andromonoecious). —C. D. A generalized thealean fossil illustrating the pistil with three styles and numerous ovules and the fascicled stamens of uneven length. —E. Ericalean inflorescence with remarkable glands on the margins of the large bracts subtending paired florets. which are themselves enclosed by a pair of bracteoles. each with a terminal gland. —F. Ericalean flower reconstruction. Volume 95, Number 1 2008 Crepet The Fossil Record of Angiosperms 50000 4 200000 — 100000 45000 4 | 90000 80000 40000 4 = /0000 L 60000 = 35000 = -50000 >, Ez D D - 40000 5 > 2 3 300004 T UU Y 0 N [^ —- 20000 o D O ® 25000- = 10000 Y o 09) > = 9000 Q > © 200007 - 8000 E = Is =] P — 7000 E & 150007 - 6000 O " —5000 ? 10000 4 — 4000 — 3000 5000 - — 2000 — 1000 Cretaceous Figure 12. the relative he lea ol the peaks (black dots) reflec families that first occur in the fossil record at the corresponding times indicated on the x-axis. Thus, 38 laceous and Paleogene angiosperm species diversity through time. IMlu[ Liu | LTU Aptian | Albian LEE nur Eocene' Oligo Paleogene As has been the convention elsewhere. ‘t the numbers of species (y-axis scale on left) now found in angiosperm 34 species occur in modern families that first appear in the Turonian (see text discussion). Cumulative species diversity through time is illustrated with diamonds, according to the y-axi s scale on the right side of the figure. Thus, while 38,334 species are now included in families that first appear in Turonian de ‘posits, 39,755 species are now included in all angiosperm families that appeared by the Turonian. Crepet (1984, 1996). Grimaldi (1999), and field between pollinators (or classes of them) and speciation Eriksson and Bremer (1992), others. However. based on recent and related studies, the relationship is as vet imperfectly understood and in need of further study before a cause-and-effect relationship can be 1994; Waser, 1998 2 also illustrates that a dramatic : Fenster et postulated al., 2004). increase in number of angiosperm species occurs from e.g.. Grant, Figure | the end of the Cretaceous to the Middle Eocene. There is a drop in the rate of appearances of new families during the Upper Cretaceous (but against a back- ground of steadily increasing species diversity; Fig. 12), which could be the result of sampling error because there are few, if any, reported fossil localities with preserved diverse fossil flowers between the Coniacean and the Maastrichtian. However, best estimates of timing major groups based on molecular 2005) or on minimum 2004) suggest this apparent absence of many of the major diverse groups clock models (e.g.. Lavin et al., age node mapping (Crepet et al., from the Upper Cretaceous implied by fossil evidence is real (Leguminosae, Asterids I, Euphorbiaceae, etc.). What involved in the proliferation of angiosperm species in the Early Tertiary, involved in the initial establishment of angiosperm the question arises: were the factors from those and were they different species dominance? 28 Annals of the Missouri Botanical Garden oBBE 0) Floral Characters BEESSs | 2) N LLE | EBENEN [ Fund T Barr Aptian Albian Cen T C Sa Camp ul MU L U Paleocene Eocene Time of Occurrence Figure 13. Times of occurrence of select floral characters. These are identified on the f we by numbers in colored boxes corresponding to the character summary below and grouped into color-coded categories (c.g. . arrangement of floral parts— green) to illustrate patterns in the evolution of D cular categories of floral characters. Times of appearance are noted in the summary below and illustrated on the x-axis of the figure. The y-axis illustrates numbers of characters occurring at the times indicated on the x-axis. Floral ree eptacle n ). l. Cupular (Aptian). 2. Hypanthium (Turonian). Arrangement of floral parts (green). 3. Paired (Barremian/Aptian). 4. Spiral i ar ls 5. Cyclic (Aptian). Floral envelope (red). 7. Apetaly (Barremian/Aptian). 8. Regular merosity (Cenomanian). Subtending bracts bracteoles (Aptian). 10. Numerous tepals (Albian). 11. Sepal tube ( Taran, (Cenomanian). 6. Co: s ‘ed into “flowers” 2. Corolla asymmetry m 13. Clawed petals (Turonian). 14. Corolla tube (Turonian). 15. Zygomorphy (Turonian). 15: Prolonged corolla tube (Lower Eocene) Stamens (yellow). 17. Fascicled (Turonian). 18. Geniculate (Turonian). 19. With tubules (Puronian). 20. Dorsifixed (Turonian). 21. Inverted Deen, 22. Staminal tube (Puronian). 23. Staminodal nectary (Puronian). Pollen (orange). 24 Viscin threads (Puronian). 25. Polvads (Turonian). 26. Pantoporate pollen (Turonian). Gynoecium (pink). 27 )ocarpy (Barr/ \ptian). 28. Inferior Pp 29, Multiple carpels and elongate style (Turonian). 30. Multiple carpels and single style (Turonian). 31. Carpels fused (Albian). 32. Carpels whorled rii 33. Ovary wall nectary (Turonian). Sexuality (blue). 34. Bisexual (perfect) (Aptian). Reproductive condition (brown). 35. Monoecy (Turonian). 36. Dioecy (Turontan). 57 siue) (Puronian). Reward structure (silver). 38. Nectar Rabe. 39. tary disk (Turonian). 40. Resin (Turonian). 41. Pollen (Turonian). 42. Staminal food bodies (Puronian). To place the Early Tertiary radiation of angio- — over contemporary gymnosperms due to their short life sperms in the context of earlier events in angiosperm — cycles, self-incompatibility systems. biosynthetically history, early angiosperms originated and diversified diverse secondary compounds, probable insect polli- T the Lower Cretaceous, enjoying some advantages — nation, and possibly wider ecophysiological tolerances Volume 95, Number 1 Crepet 29 2008 The Fossil Record of Angiosperms (e.g.. Doyle, 1978: Verdú, 2002; Feild et al., 2004) origin of the herbaceous habit in certain asterids with and possibly because of hybrid necrosis (although its phylogenetic distribution in angiosperms has vet to be determined, making it difficult to assess its impact on & Weigel, 2007 Toward the Middle Cretaceous, Cenomanian deposits angiosperm radiation; Bomblies include the first evidence of the regular eudicot flower with distinct corolla and, in evolutionary context, what might be considered developmentally interchange- floral configuration might have constituted a baseline or able, cyclically arranged floral parts. This platform morphology that would facilitate the parsi- monious origin of more advanced floral morphologies. Because many families with such derived floral morphologies are now quite diverse, the attainment of such a platform morphology in certain lineages might have catalyzed the quantum leap in angiosperm that documented by angiosperms from Turonian deposits. diversification closely followed its origin as Turonian angiosperm flowers do include those with novel morphologies and characters now associated with bee pollinators (Figs. 5G, 13). These apparently adaptive characters (Fig. 13) also occur in the same geologic timeframe as early bee pollinators, during observed increases in gross angiosperm diversity, and along with first appearances of angiosperm families that now include over 32,000 species. This broad co- occurrence of related factors suggests that the major burst of evolution accomplished in angiosperms by the Turonian supports the possibility that pollinator involvement in numerical species dominance by flowering plants cannot be ruled out based on timing alone. Yet, based on recent studies discussed above, this finding does not determine whether this co- occurrence of factors is a cause of angiosperm success or an effect of it. Further, with respect to the second question posed above about causes of the later, but equally dramatic radiation of angiosperms in the Early Tertiary, one ask: critical to angiosperm diversification, why, with the could if flower-pollinator relationships were regular cyclic flower, bilaterality, and even the corolla tube established by Turonian times, was there an apparent lag between the Turonian and Lower Tertiary before the resumption of rapid radiation of families thal pollinator relationships associatec was characterized by even more specific a with sometimes even more highly adapted floral structures (Fig. 7C— F)? Here, coordinated character (s.l.) evolution may make sense out of the patterns observed in the angiosperm fossil record. For example, with respect to the asterids, present since Turonian times, why do dramatically diversify in the Tertiary and not Could the benefits of the co-occurrences of the apparent Tertiary they earlier? answer lie in the synergistic associated short life cycle, the floral tube (extant since the Turonian in the asterid lineage), and modes of pollination associated with it (Fig. 7F)? Could speci- ation-amplifying synergy be related to yet other co- occurring temporally coincident characters? Could enhancements of the speciation process postulated to be associated with specific insect pollinators com- bined with vertebrate seed and fruit dispersers and their possible catalysis of allopatry also account for subsequent Tertiary radiations (Crepet, 1984; Tiffney, 1986)? consistent We don't know. Yet these possibilities are with the separate angiosperm radiations observed in the fossil record. And, in the context of character evolution. and its timing in angiosperm history, is it unrealistic to assume that coordinated radiations between insect pollinators and angiosperms would occur only once, especially given that the Early Tertiary. radiation. included taxa with zygomorphic TE), flowers (Fig. 7C, D), elongate tubes (Fig. 7F), flowers (Fig. brush pro- nouncedly corolla and presume d herbaceous habits? better interpreted fossil record now provides details of Thus, a understood, more conservatively angiosperm history that preclude eliminating the importance of insect pollinators in angiosperm diversification and instead point to its potential significance. Yet, based on available evidence from neontological investigations of pollination ecology, one cannot assume that insect pollination was the single most important factor in angiosperm radiation wo really be its relative precise in assessing importance among the array of Coan! Dora uides of angiospermy in e context p paleon- tological data might be to carefully monitor the history and to assess the possible of individual families causes for speciation patterns within those families. This approach could ultimately give us more accurate insights into causes of angiosperm numerical species dominance and is within reach of the better fossil record of angiosperms that has come from stunning advances in fossil discovery combined with precise analytical techniques that now aid in accurately identifying fossil taxa. Literature Cited po Phylogeny Group. 2003. An update of the up classification for the orders I. Bot. J . Linn. Soc. giosperm phylogeny gro ma Samilis of flowering plants: APG 399—430. d — e o I F. & D. L. Dilcher. Ancient bisexual flowers. Science 224: 511— Bomblies, K. & D. W eigel. 2007. Hybrid Autoimmunity as a potential gene ~ barrier in plant i -39: Genet. 8: 382 nec rosis: species. Nat. Rev. Annals of the Missouri Botanical Garden Bremer, K. 2000. flowering plants. 1707-4711. G. J. 196: Karly Cretaceous lineages of monocot Proc. Natl. Acad. Sci. U.S.A. 97 Brenner, The spores and pollen of the Potomac Group of Martane Maryland Dept. ( Geol. Mines Water Resources Bull. —125. Burger, W. C. T Why are there so many kinds flowering plants? BioScience 31: 572, amak Burnham, R. J. 2008. fossil record? Hide and go seek: What does pe mean in the Ann. Missouri Bot. Gard. 51-71. Chase, M. W. 2004. Monocot pa Amer. J. Bot. 91: 1645—1655 Chaw, S-M., C. D. Palmer. 2000. three plant genomes: Monophyh of An overview. . Cheng. T. M. Vincent & J. Seed e MCA ny inferred from all Parkinson. xtant (er rms and origin of Gnetales from conifers. " roc. Natl. Acad. Sci U.S.A. 97: 4086—4091. Conwentz, H. 1886. Die Bernsteins. des Bernsteins, 2. Die Wilhelm Flora ad rmen des Engelmann, Leipzi Crane, P. R. & D. L. Dileher. 1984. Lesqueria: An early angiosperm fruiting a from the mid-Cretaceous. Ann. Missouri Bot. Gard. 7 : 384402. ——, K.R. o E. M. Fr . Drinnan. 1993. A. middle i an) platanoid from Early Cretaceous. (early to inflorescences associated with Sapindopsis leaves the Potomac Group of eastern North America. Syst. Bot. 344. Crepet, W. L. 1984. mechanisms: Patterns of evolution and implications vis Missouri. Bot. wo Ñ | Qo Advanced. (constant) insect aters vis angiosperm diversity. Ann. Gard. 007-030. 1990. floral characters: Upper Cretaceous (Turonian) taxa with tricol- Timing in the evolution of derived pate and eo als pollen. Rev. Palaeobot. UL 90: 339—300. 2 2000. Progress in understanding angiosperm history, success, and re Jationshipa: Darwin’s r] i. a xing pus nomenon.” Proc. Natl. Acad. Sci. : 129; | & a p D aghlian. 1980. Castaneoid inflorescences from the Middle Eocene of E messee and the diagnostic value of E n (at the subfamily level) in the Fagaceae. mer. J. Bot. 67: o & ———. 1982. Euphorbioid inflorescences e the Middle Koce ici eee Formation. Amer. J. Bot. 258-200. — — — & D. L. Dilcher. 1977. Investigations of angiosperms from the Eocene of North America: A mimosoid inflorescence. Amer. J. Bot. 64: 714-725. v C. D. Feldman. 1991. The earliest. remains of grasses in the fossil record. Amer. J. Bot. 78: 1010-1014. E. M. Friis. 1987. The insect evolution of i pollination in angiosperms. Pp. 181—201 in E. M. Friis, W oner € P. R. Crane (editors). Moo rms and Their Biological Consequences. Cam- bridge P ress, Cambridge. ^ PS . Herendeen. 1992, | Cane Howe TS Pn the icu D ;ocene of onn North America. Pp. 43— 55 in P. S. Herendeen & D. in Legume i5 stematies: e 4, a Dilcher (e ditors), ro es The Fossil Record. Roya Botanic ( pat ns, K. Kev ———— 4 . Nixon. 19894. Earliest megafossil evider of Fagaceae: Do ne tic one biogeographic implications. Amer. J. Bot. 76: 842-8 Origin of & ————. |989b. from the Oligocene and their phylogenetic: implications. imer. jh Bot. 76: 1493-1505. ———— 1998a. Extinct transitional Fagaceae Two new fossil flowers of magnoliid affinity from the Late Cretaceous of New Jersey. Amer. J. Bot. 85: 1273-1288. SN 1998b. Cretaceous (Turonian) « & ————. Fossil Clusiaceae from the Late Jersey and ru Amer. J. Bot. regarding the history of bee nollination. 1122-1 i — & D. W. Taylor. 1985. The dive rsification of the Leguminosae: Fi rst fossil evidence of the A Mimosoide ae and Palilionoide ae. Science 288: 1087-1089 1986. Primitive mimosoid flowers from the | ale 'ocene-Eocene and their systematic and evolution- ary implications. Amer. J. Bot. 73: 548-5 Dileher & F. W. Potter. 1974. Eocene angiosperm flowers. Science 185: 781-782. = O ~) 1 & Investigations of angio- A catkin with 813-823. e evidence for ocene of North ~ rica: . Amer. J. Bot. 1991. the evolution of biotic pollination. Philos. Trans., Ser. B. : 187-195 sperms from the F juglandac eous affinities n Nixon & E. M. Friis. DD ©: 2d. ). 1992. Oldest fossil flowers of Hamame didus eous affinity, fs ge Late Acad. . Freudenstein. Cretaceous of New Jersey. Proc. Natl. 809: 8969-8989 & M.A. Amer. J. Gandolfo. 1996. Turonian flowers Bot. 83: 148—149 Fo & The age of major angiosperm vete s based on of the Theales. evidence and Ww = glisse ny: mesofossil and maecrofossil evidence from Cretaceous deposits. Amer. J. Bot. 91: 1060-1082 Daghlian. €. P. 1981. A review of the fossil record monocotyledons. Bot. Rev. (Lancaster) 47: 517—555. De Craene, L. P. Ronse & E. 2006, The systematic relationships of E osinolate-producing plants and relatec Haston. families: A cladistic investigation based on morphologic Li and molecular characters. Bot. J. Linn Soc. 1514: 45 94 Dilcher, D. L. 1974. eS rm leaf remains. Bot. Rev. (Lancaster) 40: 1—157. To Approaches to the identification of 00. Toward a new synthesis: Major evolutionary tre "t in the _angiosper rm fossil record. Proc. Natl Acad. Sci. U.S.A. 97: 7030-7036 & P. R. une. 1984. Irchaeanthus: An early angiosperm from vi Genomanian of the a rn n rior of North America - Missouri Bot. Gard. 51-383 W. . B. Beeker & H. C. cn 1916. he produc live and vegetative morphology of a Cretaceous Science 191: 854-850. — pow ~ ar 18108 perm. Donoghue, M. J., J. A. Doyle, J. Gauthier, A. C. Kluge & T. Row 1989. The importance of fossils in phylogeny reconstruction. Ann. Rev. Ecol. Syst. 200: 431—460. Downie, S. R. & J. D. Palmer. 1992. Use of e eor DNA . 14- D OY le (e sito) rearrangements in reconstructing plant phylogeny. 35 in P. 5. Molecular Systematics of Plants. Chepmen and Hall, N Soltis. D. E. Soltis & J. York. Doyle, J. A. 1969 Atlantic C pa a lain and ils e evolutionary signifie ance. Arnold e rr. 50: 1-35. a Monocotyledons: Cre laceous anglospe rm polle n of [ the Their evolution and aaa alive Cd V. Fossil evidence on p evolution 5 f the monocotyledons. Quart. Rev. Biol. 399 Volume 95, Number 1 Crepet 31 2008 The Fossil Record of Angiosperms 978. The origin of angiosperms. Ann. Rev. Ecol. Gandolfo, M. A.. K. C. Nixon & W. L. e 1998. A new 2 Syst. 9: 365-392. fossil flower form the Turonian of New Jersey: Dre a Dressler, R. L. 1968. Pollination by euglossine bees. pie d gen. et sp. nov. (C A ae J. I Evolution 22: 202-210. 74 . 1982. Biology of the orchid bees (Euglossini). Ann. ———— 4 ————. 2000. Monoc apne eview Rev. Ecol. Syst. 13: 373-39¢ of their Early TE record. Pp. 44—52 a ilson Drinnan, A. N., P. R. Crane, E. M. Friis & K. R. Pedersen. & D. Morrison (editors), Proceedin of cn Second 199]. Angiosperm flowers and tricolpate pollen of International Conference on the ane Biology of buxaceous affinity from the Potomac Group (mid-Creta- the Monoe oiledons CSIRO, Sydney. ceous) of eastern. North America. Amer. J. Bot. 78: i, & 2002. Triuridaceae fossil flowers 153-176. from the Upper Cretaceous of New Jersey. Amer. J. Bot. Endress, P. K. 2001. Origins of nos n J. Exp. 89: 1940-1957 Zool. B. tpe Developm. Evol. 2 -115. — —— & —————. 2004. The oldest complete Engel. M. 900. A new bad ii M oldest fossil fossil flowers of Nymphaeaceae and implications for the bee xd chee ra: Apidae). Amer. Mus. Novit. 3296: are insect entapme nl eee e in -l arly angiosperms. Proc. Natl. Acad. Sci. U.S.A. 101: Eriksso ). B. Bremer. 1992. Pollination systems, 8056-6000. dispersal ds life forms and diversification rates in ————. ———— & —————. 2008. Selection of fossils angiosperm families. Evolution 46: 258-2060. salibration of PET ar da ating models. Ann. Missouri bs Feild, T. S., N. C. Arens, J. A. Doyle, T. E. Dawson & M. J. Gard. 95: 3442. Donoghue. 2004. Dark and disturbed: A new image of Grant, V. 1949, Pollinating systems as s isolating mechanisms early ae ecology Paleobiology 30: 82-107. in angiosperms. Evolution 3: 82— ge CB . Armbruster, P. Wilson. M. R. Dudash & — ————. 1994. Modes and origins w mechanical and J. Thom apson. 2004. m syndromes and floral aps d isolation in angiosperms. Proc. Natl. Acad. spec an Annual Rev. Ecol. Evol. Syst. 35: 375—403. Sci. U.S.A. 91: 3-10. Friis, E. M. 1983. Upper Cretaceous e nonian) "floral Er D. € W. Martin. 2004. Reading the entrails of structures can affinity containing Normapolles chickens: Molecular timescales of evolution and the pollen. Rev. Palaeobot. Palynol. 39: 161-188. illusion of ue Trends Genet. 20: 80-86. 1985. Actinocalyx gen. nov., sympetalous angio- Grimaldi, D. A. The co-radiations of aio insects sperm flowers from the Upper Es taceous of southern and v in the Cretaceous. Ann. Missouri Bot. Sweden. Rev. Palaeobot. Palynol. 45: 171-183. Gard. 313—400. W. L. Crepet. 1987. Time a appe arance of floral ———— € M. S. Engel. 2005. Evolution of the Insects. features. Pp. 145-179 in E. M. Friis, W. G. Chaloner & P. Cambridge Univ. Press, Cam bridge. R. Crane (editors), The Origin of a e Their erie M. M. 1996. I Pollen and the Fossil Record. Biological Consequences. Cambridge Univ. Press, Cam- Ph.D. Thesis. University of East London and The Royal bridge Si rur Ke Crane & K. R. Pedersen. 1986. Floral | ———— & M. 5. Zavada. 2000. Pollen of Monocotyledons: idence: for € retaceous Er angiosperms. Nature e m ‘ters for cladistic analysis. Pp. 194-213 in 320: 163-164. Wilson € D. Morrison (editors), Proceedings of the ———, ——— &— 1988. Re a aie o Conference on the Comparative rotación P ds ae. ion Skr. 31: 1-55. Biology of d Yd Pur CSIRO, Sydney — — —, H. Eklund, K. R. Pedersen za P. R. Crane. 1994. Herendeen, P. ane. 1992. Early cae x Virginianthus calycanthoides gen. et sp. nov.—4A caly- fruits from cd DE ne E southern England. | 08 canthaceous flower from the Potomac Group (Early in P. S. Herendeen & D. L. Dilcher (editors). del “es in Cretaceous) of eastern North America. Int. J. Pl. Sei. Legume Systematics: Part t, The Fossil Record. Royal 155: sia 35. Botanic Gardens, ew. . R. Crane € K. R. Pedersen. 1997. Fossil history ——— ——. 1995. The fossil history of the of inaenolid angiosperms. Pp. 121-156 in K. Iwatsuki € Monat ledons. Pp. 1221 in P. J. Rudall, P. J. Cribb, P: aven (editors), E in and Diversification of ). F. Cutler & C. J. Humphries (editors), Monocotyledons: jan. Pu Springer-Verlag, Tokyo. Systematics and o Roy: il Botanic Gardens, Kew. R. Pedersen € P. R. Crane. 2001. Fossil Hermsen, E., | . Gandolfo, K. C. Nixon & W. L. Crepet. AE of water lilies Ea s) in the Early 2003. Diisestplus gen. nov. (Aff. Iteaceae), a fossil Cretaceous. Nature 410: : ). saxifrage from the Late Cretaceous of New Jersey, U.S./ A. Doyle, P. K. Endress & Q. Leng. 2003. Amer. J. Bot. 90: 1373-1388. Archacfructus Angiospen rm d or specialized early Hickey, L. 1971. Classification. of us architecture of angiosperm? Trends Pl. Sci. 8: 369-373. ion leaves. Amer. J. Bot. 60: 17-33, , P. R. Crane 8 K. R. Pedersen. 2004. Araceae from ——— & J. A. Wolfe. 1975. The 2 of angiosperm the Early Cretaceous Portugal: Evidence on the E E s alive morphology. Ann. Missouri Bot. emergence DE. monoc a dons. Proc. Natl. Acad. Sci. Gard. us —509. U.S.A. 101: 16565-165 —— —— & R. E Peterson. 1978. A a fossil genus , K. R. Pedo rsen a a R. Crane. 2006. Cretaceous of llis ginger family from the Late Cretaceous to Early angiosperm flowers: Innovation and evolution in plant Eocene sediments of western interior North America. E ion. Palaeogeogr. Palaeoclimatol. Palaeoecol. Canad. J. Bot. 56: 1136-115 232: 251-293. da D. 2005. Evolutionary innovations in the fossil 'eco Furness i A. P. J. Rudall € A. Eastman. 2002. The intersection of ec P oe and € dai buin of tapetal and [und n characters lo the mac Noevalution: J. Exp. Zool. B. Mo ; systematics of Triuridaceae. Pl. Syst. Evol. 235: 209-213. 304B: 504—519 . Developm. Évo Annals of the Missouri Botanical Garden Ji. Q.. H. Li. L. M. Bowe. Y. Liu & D. W. Taylor. 2004. Early Cretaceous hanes M dur sp. nov. with bisexual flowers from Bei E » western Liaoning, Cina, Acta Geol. Sin. 78: 883—890. Keller, J. A.. P. H. He rS & P. R. Crane. 1996. Fossil flowers and fruits of the Actidiniaceae from the Campa- man (Late Cretaceous) of Georgia. Amer. J. Bot. 83: 3-541. bdbsndel ira, C. C. . The four Ks of plant-arthropod re A in n p time. Geol. Acta. 4: 400-438 S. He Lavin, M.. , 2005. E o eguminosae amplie ales a erendeen & M. F. We Dun. rates analysis of La rapid diversification of ie sages during the Tertiary. Syst. Biol. 5: 530- o ). Les. D. H.. E. L. Schneider. D. J. Padgett. P. S. Soltis, D. Soltis & M. uiis. s UE classification an) floral. evolution of water lilies (Nymphaeaceae: Nym- phaeales): A synthesis of non-molecular, rbeL, matK, and 18S rDNA data. Syst. Bot. 24: 28—46. A. 1981. Ultrastructural characters of the pollen grains of African Annonaceae and their significance for the phylogeny of primitive angiosperms. Pollen & Spores 7 949 ND == c 1986. The evolutionan w EA the Poales/ e 97-318 in LP. Pinto g G. Lozano (editors), Flora de Colombia. Univ iE Nacional y ) “Triuridac eae. Pp. e Colombia, Bogota. & T. Hübsamen. 1986. Triuridaceae. Flora Neo- trópica, ws 10. Hafner, New York Mi aisló S. € M. seed pl: Sanderson. 2002. Re lationship among ints mo rre ad from highly conserved genes ing conflicting pi oe netic signals among ancient linea ages. in J. : 1991- poe = le POS pattern, diversity, and Fudicots. Ann. Missouri Bot. Midgley, J. J. & W.J pollination and dispersal to the success of th ? Philos. Trans., Ser. B. 333: 209-215. Mohr, B. n R. & H. Ecklund. 2003. magnoliid T from the | ormation (Brazil). Rev. P Morley, R. J. & C. W. clocks, and the mer. J. Bot. —, P Herendeen. 1999, Phylogenetic Gard. 86: 297-372. Bond. 1991. How important is biotic anglo- Araripia florifera, a alaeobot. Dick 2003. Missing f fossils. molecular .] origin of the Melastomataceae. Ame Bot. 90: 1038-1045 Muller, J. 1981. Fossil pollen records of extant angiosperms. Bot. Rev. (Lancaster) 47: 1-141. Niklas, K. J., B. H. Tiffnex 8 A. H. Knoll. 1983. Patterns in vascular land d diversification. elt 303: 314—316 Nixon. K. C. 1996, Paleobotany in cladistics and cladisties in paleobotany: : Lists nment and uncertainty. Rev. Palaeo- jm Palyi : 90: 361-373. 1999. The parsimony method. for 4414. for ac m ratchet, a new rapid parsimony analysis. Cladistics 15: . 2003. Winclada. Software package analysis. Windows 95/98/NT version. Published by the author, tha a. New York. JO6. Early flowering plants: Evidence for and against p habit and pollination syndromes. P. 99 in stracts, 7th European Paleo hor and Prague. evidence, and molecular dating: Program and ; pp C No ‘rence, . 2008. Paleobotany, An cin from the Nymphaeales. Ann. Missouri Bot. Gard. 95: 43-50, & flowers of ericalean affinity. fossil 1993a. Late Cretaceous Amer. J. Bot. 80: 616-623. Crepet. diversification of 37: SÁ — 190; Cretaceous of New Jersey. Amer. V sympetalous « Sp ies Ww flower from el x Bot. 6O(Supy jl.) & 994. A atomi from ni Late Cretaceous of New Jersey. Amer. J. Dot. Pita 176-177. . 1991. values A c a anaha Cladistics 7: 233-241. . Hesse. 2002. in orchids, and mes consequences for germination. and Ann. Bot. (London) 89: 653-604 K ossil flower of possible fagalean Davis. digo ap laxa, missing 3-24 = acini; E. & Types of polle n dispersal units fe rtilization. Parkinson. C. E. K. L. Adams & J. D. Plots 1999, Multigene analyses identify the ban PUE. lineages of extant flowering E: mts. Curr. wx 185-1488. Pedersen, K. R., P. R. Crane. A. N. s in & E. M. Friis. 1991. Fruits S a Nat ol North America with pollen grains of the Clavatipollenites type. 211-990, Grana 30: , P. R. Crane & A. N. 994. Re — tive structures of an extinel ra from the Early a eastern North \me a a. Rev. Palaeobot. Palynol. 80: 291-303. . O. & B. N. Danforth. 2006. Science 314: 614. Drinnan. Cretaceous Poinar. A fossil from Early Cretaceous Burme se amber. Prasad, V.. C. X. id 2005. Dinosaur csi " the early evolution P jence 310: 71-1180. i Kim ee the Cretaceous Stromberg. ee & grasses and grazers 1977. rise to dominance of the 151—452. & D. l. Axe Irod. 197 |. and pi ast continental movements. Ann. Missouri Bot. Gard. a Raven. P. H. i angiosperms. Evolution 31: Angiosperm biogeography 07 1977. E rence Ol Ral. . J. Ecology and evolution of flowering plant 196: 622-629. & K. Meyer. 2001. The roles of Bot. 88: \ taxonomic pee s of Glucosinolate- Syst. Bot. 16: Lae Renner, S. S. G. Clausing Historical bioge rd of Melastomalaceae: and long distance dispersal. Amer. J. 199la. plants. Part ]. Rodman. J. producing phenetics. 5 (9) . 1991b, A taxonomic analysis of Glucosinolate- woducing plants. Part 2, cladistics. Syst. Bot. l6: | p à ol — Rudall, P. J. 2003. Monocot pseudanthia revisited: Floral structure of the mycoheterotrophic family PL Sei. o 5): 5307-5320. N. Wikstrém & K. 2004. el is on plant divergence times. Amer. J. Bot. 91: 1656-1665 Schónenberger. : x E. M. 2001. Fossil flowers. of ericalean affinity from the Late Cretaceous of southern Sweden. Amer. J. Bot. 88: 467—480. K. R. 1. Friis. 2001. from the Late 205-230. Triuridaceae. Int. Bader: M. J Thorne. Bremer. Normapolles ` . 2, € flowers E fagalean affinis: Cretaceous. of Portugal. Pl. Syst. Evol. 22€ Sims. H. J.. E S. Herendeen & P. R. Crane. 1998, New genus of fossil Fagaceae from the < paseo (Late Cretaceous) of central Georgia (USA). Int. J. PL . 159: 39 ———. Ro Lupia, R. A. Chr copies & P. R. Crane. 1999, Fossil r idt Nommápallat | pus n. from the Upper Cretaceous. of. southeastern North rica, Rey Palaeobot. Palynol. 106: 13 er Soltis, D. E. P. S. Soltis; M. W. Chase. M. E. Mort, D. Albach. M. Zani V. Sai n, W. H. e S. p. Ya M. F. Fay, M. Axtell, . Swensen, lL. M. Prince, W. J. Volume 95, Number 1 2008 Crepet 33 The Fossil Record of Angiosperms Kress, K. C. Nixon € J. S. Farris. 2000. 188 rDNA, P equences. Tw J. Linn. Soc. 133: 381-46 Staedler, . P. H. Weston & P. K. no: 2007. Floral rss and s architecture in Calycanthaceae (Laurales). Int. J. da . 168: 285-300. Stebbins, G. 5 198 E hy are ur so many species of mos ring Eo a ience 31: 573-577 E D. Jilcher, S. Zheng & Z Zhou. f the n T A Jur from northe r E hina. Science ! — ———, Q. Ji, . Dilche Tes K. Nixon € X. F. Wang. vA p» haef aa a new basal angiosperm familv. Science 296: 904. Thein, L. B. 1980. Patterns = ee in the primitive Apiapi rms. miohoples ica 12: 1-13 Theissen, G., A. Becker, A. - Rosa. Non & K.-U. Winter. 2000. box genes in june PI. Molec. Tiffney, B. H. 197 Martha's 36-137. —————. 1084. Seed size, dispersal syndromes, and the rise Angiosperm phylogeny inferred and atpB 1998. In search rassic angiosperm, Archaefructus, 05 A. Kanno, J. T. Kn A short que of M. ADS. Biol. 42: —-]149. . Upper Cretaceous angiosperm flowe r Massachusetts. Nature 2 from y; ineyard, of the P ert and hypothesis. Ann. Missouri Bot. Evidence 76 Gard. 51-570. . 1986. Fruit i. dec dispersal and the evolution of 13: 394—410. the Hamamelidae: Ann. Missouri Bot. Gard. 73: 394— )08. 20( Phylogeography, fossils, and Northern Hemisphere biogeography: The role of physiological uniformitarianism. Ann. Missouri Bot. Gard. 05, 135-143. 1995 —— & 5. J . Angiosperm growth habit. pau and diversification reconsidered. Evol. Ecol. 9: -117 Mazer. Ve i M. 2002. Age at mame and diversification in woody Evolutio 352-136 & J. A. Doyle. 1975 . The base 's of angiosperm Ann. Misco Bot. Gard. 62: T eus Walker, J. W. phylogeny: 664—723. pes M. 1998. Pollination, angiosperm speciation, and the ure of species boundaries. Oikos 82: 198-201. W hitehes ad, D. R. 1969. Wind pollination in the a : Evolutionary and environmental ae d -35. Pa Iynology. > ngio- Savolainen € M. Chase. 2001. ona dl the Era rms: albina the family tree. ^oc. Roy. Soc. London, B., Biol. Sei. 268: . Bell, S. Soltis € D. E. Soltis. 2005. pl. limes and histories] biogeography of Nym- phaeales. Syst. . 30: 693—704 Zanis, M. J., P. $ S Solt tis, Y. Qiu, E. 2003. Phologenche analyses and pe basal AEN Ann. Missouri Bot. Gard. 90: 129-150. e & D. E. Soltis. rianth evolution in Zavada, 984. Angiosperm origins and evolution based on NR fossil pollen ultrastructure. Bot. | 71: e 4—463. Zhou. Z.. W. L. M & K. C. Nixon. fossil dan e of the laa lidaceae: (Turonian) inflorescences and fruits of Altingioidea. Amer. J. Bot. 88: 753-766. Ann. Missouri 2001. The earliest Late Cretaceous SELECTION OF FOSSILS FOR Maria A. Gandolfo,? * Kevin C. Nixon,” and CALIBRATION OF MOLECULAR — William L. Creper DATING MODELS' ABSTRACT In the past decade, there has been a rise in interest in the plant fossil record. Fossils potentially provide information for assessing homology and evolutionary change (e.g., the popular missing link phenomenon), character evidence that affects phylogenetie conclusions and, thus, our unde nds of modern relationships, evidence of past distributions that can aid in E rstanding biogeographic hiona and estimates of minimum ages of the clades to which they belong. Recently, many molecular biologists have used fossils in their analyses as a way of providing a c alibration eh for evolutionary models used to approximate ages for the nodes of phylogenetic 1 rees. However, there has been little, if any, discussion of the criteria by which calibration fossils can be selected for these studies. When considering the use d a fossil as a calibration point, it is critical to take into account the quality of preservation, the method and details of identific ae liability of the taxonomic placement), and the accuracy of the published age. Here, we provide basic criteria for the use ssils to calibrate molecular evolutionary models. These approac Jes nol only provide better primary estimates for ages of c Ps but also provide more reliable sources for those molecular biologists wishing to clean up their male "c alae clocks. Key words: Age, calibration point, fossils, geology, identification, molecular clock, preservation, sampling. Although fossil remains have been the objects of During the past 10 years, there has been a tendency o H studies since Greek and Roman times, they are — to place fossils onto phylogenetic trees (with or without frequently ignored by researchers dedicated to simultaneous cladistic analysis) to calibrate models — a molecular and development studies of plants and estimating the age of the obtained nodes or clades (e.g., animals. Nevertheless, knowledge and interpretation Bremer, 2000; Bremer et al., 2004; Donoghue et al., of the fossil record are fundamental to understanding 2001; Renner & Meyer, 2001; Renner et al., 2001: the history of life in general, including evolutionary Wikström et al., 2001; van Tuinen & Dyke, 2004; Yoo — processes. Fortunately, we are witnessing an explo- — et al., 2005; and others). These methods, such as the sion in the utilization of the fossil record, largely nonparametric rate smoothing. (NPRS; Sanderson, as a consequence of interest in ages of clades, as — 1997), semiparametric rate smoothing by penalized part of major biogeographical studies, and as a tool likelihood (18s program; Sanderson, 2002), and Bayes- for assessing divergence times. Therefore, there has ian divergence time estimation (Huelsenbeck & recently been a rise in interest in data obtained from Ronquist, 2001), typically use fossils for calibration. fossils. Although the bulk of the fossil record remains Nevertheless, the fossils often are not included in the undiscovered, there is no doubt that fossils potentially analyses performed using these models, and the provide important information on diverse aspects of placement of fossils requires either the results of biology. Fossils are essential for understanding the previous morphological or combined cladistic analyses sequence of evolutionary changes (e.g., the popular or acceptance of fossil identifications by paleobota- missing link phenomenon) They make available nists. Fossils also have been used extensively for character evidence that affects phylogenetic conclu- — providing historical distributional data in biogeograph- sions and assessment of morphologic homology, and, — ical studies (Donoghue et al., 2001; Manos & Stanford, consequently, they influence our understanding of 2001; Morley & Dick, 2003; Renner et al, 2001; modern relationships among taxa. They also provide Wagstaff et al., 2002; Renner, 2004a, b, 2005; and support for past distributions that can aid in others), but, once more, fossils are usually not included ~ comprehending biogeographic histories. More impor- in the primary analyses and, instead, are simply tantly, they provide estimates of minimum ages of the mapped onto molecular phylogenetic trees based on clades to which they belong. assumptions about their relationships. 'This research was supported in part by National Science Foundation grant DEB-0345750 to MAG. The authors thank E. J. Hermsen for discussions and the photo of fossil Ribes. * L. H. Bailey Hortorium, Department of Plant Biology, Cornell University, Ithaca, New York 14853, U.S.A. ! Author for correspondence: mag4@cornelledu. doi: 10.341 7/2007064 ANN. Missouri Bor. Garp. 95: 34-42. PUBLISHED ON 11 APRIL 2008. Volume 95, Number 1 2008 Gandolfo 35 Fossils Du aes of Molecular Dating Models The tendency to map fossils onto phylogenetic trees without including them directly in analyses makes the original identifications and quality of the fossils used extremely important. Although it is refreshing to see that fossils are utilized by molecular biologists to date thal sometimes they do not consider several issues when trees, it is also disappointing to see selecting fossils to be used as calibration points. Most will agree that merely mapping fossils on trees is not optimal because such placements are not directly tested and are, therefore, subjective. More important- ly, any conclusion based on the position of phyloge- netically untested fossils on a tree will be debatable. For example, Doyle and Donoghue (1993: 151) mapped the “most likely positions of Barremian- Aptian fossils? without directly including them in an analysis onto one of the most parsimonious trees obtained in a previous analysis of only extant taxa. Among the fossils were the earliest (early Cretaceous) Liliacidites Acaciaephyllum Fontaine (2000) actually included the fossils within a cladistic analysis and found that putative monocots, represented by a Couper pollen grain and leaf. Later, Gandolfo et al. Acaciaephyllum | and Liliacidites were placed in various positions within various clades, but never in the paleoherbs clade in which Doyle and Donoghue placed them with the monocots. Thus, basing a most likely position for a fossil on a generalized identifi- cation without cladistic analysis, in this case, has not held up to later scrutiny. In this paper, we will address various problems associated with fossil selection for calibration of molecular trees and some that should solutions minimize errors in selection of fossils. These recom- mendations are not meant to replace direct inclusion of fossils in a simultaneous total evidence analysis, which should be undertaken whenever possible, e.g., Hermsen et al. (2003) and Crepet et al. (2005). THE PROBLEM For any researcher of paleontology, it is common knowledge that working with fossils tremendous challenge. This challenge has been made more complex by the need to either include fossils directly in modern cladistic analyses or provide strong pl: acement in the context of evidence of their morphological synapomorphies for the groups in This often must be accomplished without the help « analyses, question. cladistic © f pre-existing morphological resulting in an increasing number of paleobotanical studies in which morphological matri- ces for modern groups must be constructed in order to place fossils (e.g.. Nixon € Crepet, 1993; Crepet & Nixon, 1998a, b; 2003). Nixon (1996) Hermsen et al.. reviewed several issues concerning the relationship between traditional paleobotanical work and modern cladistics, including issues of homology, missing data, and reconstruction. Clearly, some of these issues are aggravated when fossils are mapped, without inclusion in an analysis, onto previously calculated phyloge- netic trees for calibration of molecular dating models. discussion focused on the fossil In a surprising divergence time between the ancestors of the birds o — and the mammals 310 million years ago (Ma) (1996) and based on (2004) listed three sources of error that can influence clade age proposed by Hedges et al. molecular dating, Graur and Martin estimates: (1) errors in the topology of the phyloge- netic tree, (2) errors in the taxonomic identification of the fossil material, and (3) errors in the chronological assignment of the geological strata. Combining these concerns about molecular dating per se with concerns about phylogenetic analysis of fossils (Nixon, 1996), we propose that there are several other factors that be addressed that impinge on each of the (1) fossil nature of fossil preservation, incompleteness of the need — above caveats: preservation (the biased fossils or fragmentary taxa, and reconstruction of fossil taxa); (2) methods of identification and taxonomic placement: (3) interpretation of the fossils and establishment of homologies; (4) sampling of taxa/ fossils; and (5) the age assigned to the fossils based on B — the age of the formations where they were collected. 1. FOSSIL PRESERVATION Biased nature of fossil preservation. Plant fossils can be preserved as compressions, impressions, permin- eralizations, charcoalifications, and various other less common ways (Taylor & Taylor, 1993; Stewart & Rothwell, 2001). Each type of preservation pro- — vides different kinds of data and varying qualities of information. the fossil In a permineralization, usually has preserved internal structure (anatomy) providing details of cell types and tissues. However, some aspects of overall morphology may be difficult to interpret. In compressions, on the other hand, typically only characters of the external morpho- logy and some cuticular features can be observed. Charcoalified (fusainized) fossils are an exception, since they are three-dimensional and mostly show a spectacular array of cell-by-cell preservation, pro- viding a wealth of information unusual in the fossil record (Friis et al., 1986, 1992, 1994, 1995; Friis, 1990; Crepet et al., 1992; Herendeen et al., 1993, 1994, 1999: Nixon & Crepet, 1993; Crepet & Nixon, 1994, 1998a, b; Gandolfo et al., 1998a, b. 2000, 2002, 2004; Hermsen et al., 2003). It is common knowledge that soft parts are more difficult to preserve than hard Annals of the Missouri Botanical Garden parts: there are more bones preserved than internal organs of vertebrates and more leaves and fruits than embryo sacs in flowering plants. Based on this, it is reasonable to assume that the majority of >a are incomplete and, therefore, preservational bias is high. Preservalional bias is also dependent on the environment in which fossilization occurs. Fossils are typically preserved better (with less damage and more detail of delicate structures) in environments of low energy (e.g.. deltas and lagoons) than in areas of high energy (e.g., high-energy rivers and ash flows). the fossils or Although plants typically have many parts Ini ompli teness of fragmentary Laxa. that are relatively resistant to immediate decay and tend to preserve well, in general very few complete fossils with attached stems, roots, leaves. and reproductive organs are available in the fossil record. Notable exceptions occur with some aquatics Archaefructus Sun, Dilcher, Zheng & Zhou, presumably due to in situ preservation in low-energy 2002). exacerbated by t the such as environments (see Sun et al., This tendene y lo fossil fragmentation is further tendency for some parts to be delicate and/or ephemeral on the living plants. Thus, we may have leaves, fruits, and wood from a single locality, but it is atypical to find a single fossil of the complete and stems), much organism (attached. fruits, leaves, less the softer reproductive parts (i.e. flowers). Because of this fragmentary nature of fossils. many characters, even for well-known fossil taxa, are unavailable or incomplete. In the final analysis, most plant fossils are typically represented by single categories of organs (leaves, pollen grains, roots, fragments of stems, reproductive structures, etc.). Isolated pollen grains, perhaps the most ubiquitous of plant fossils, provide a good example of the difficulties inherent in. paleobotanical identification. Common characters studied in pollen grains are size. aperture number and type, and exine sculpture and structure, but determining taxonomic position based on these characters is often very difficult and in general has relied on direct comparisons with living 1983, phylogenetic analyses. Isolated pollen remains diffi- taxa (Zavada, 1991) rather than comprehensive cult to include in large-scale phylogenetic analyses because the available characters that lend themselves to cladistic coding (e.g., aperture number and form) generally conform lo very broad groups of taxa, and the details that differentiate smaller clades are often uniform (e.g.. quantitative. and labile, or relatively fagaceous pollen). Thus, without additional organs. assuming the occurrence of a taxon based on fossil pollen grains alone can be misleading (e.g.. Liliaci- dites), such identifications must have less weight than identifications based on a broader range of features. The dispersed organs (e.g., pollen alone) is very important issue of identifications based on very few to address when choosing a fossil as calibration point, in view of the fact that typically only one organ is used for calibration and not the complete plant. common laxa. of fossil Another paleobotanists is to Reconstruction practice among assemble dispersed parts, usually from the same or nearby q localities, into full-organism or partial-organism reconstructions. A well-known case of reconstruction is the extinet genus Glossopteris Brongniart, based on an assemblage of variously preserved (but not directly attached) fossil organs including leaves, stems. roots. and a wide array of reproductive organs (e.g.. Gould & 1977). Another corystosperm group of Mesozoic seed ferns. Corysto- Delevoryas, interesting case is the sperms have been considered to have Rhexoxylon- type wood based on the association of putatively Gothan Archanglesky & stems in the same strata and the presence of similar Dicroidium ol yslospe THOU: Rhexoxylon Bancroft. emend. — types of secretory cavities in the two fossils. However. Dicrodium leaves now have been found organically attached to fossil stems with Dadoxylon Enlitcher- (Mever-Berthaud et al.. 1992, 1993), making the association with Rhexoxylon wood am- type wood biguous. These two facts may imply that corysto- sperms have more than one type of wood; alternatively, perhaps Rhexoxylon-type wood is not The reconstruction problem IS even more aggravated {Í for actually found in the corystosperm clade. angiosperms. The most common angiosperm organs preserved in the fossil record are leaves and, less commonly, wood, seeds, fruits, and flowers. Unless all the organs are found in organic connection, it is very difficult to produce reconstructions of a whole plant, and there is a tendency to attempt reconstructions based on minimal data. Sometimes. fossils are preserved in the same sediments at the same horizon and they share one or more characters (e.g., the same type of stomata or venation pattern), and, such cases, a reconstruction seems reasonable. This is the case with Archaeanthus Dilcher & Crane, which is based on five fossil taxa independently described by Dilcher and Crane (1984): Archaeanthus linnenbergeri (multifollicular fruits), Kalymmanthus walkeri (probably bud-seales), Archaepetala beekeri and A. obscura (putative. perianth parts). and Liriophyllum These associated but not in organic connection in narrow kansense (leaves). organ fossils were found horizons in different localities, and, furthermore, all these fossils shared the presence of amber-colored 1984). resin. bodies (Dileher & Crane, Nevertheless, Volume 95, Number 1 2008 Gandolfo et al. 37 Fossils for Calibration of Molecular Dating Models without organic connection, reconstructions are always hypotheses to be tested with more data. One approach to testing hypothetical reconstructions is to include the [2m disparate parts in a cladistic analysis to see if, indeed, they can be placed together in the context of parsimony based on the evidence present in each detached organ fossil. This approach was successfully used to test reconstructions of Cretaceous fern fossil organs by Gandolfo et al. (1997). Another type of fossil reconstruction goes beyond connecting existing dispersed fossil organs and instead utilizes extrapolation of missing parts, usually based on the presence of attachment scars. Such reconstructions are the most highly suspect and, when the identification of the fossil in question relies on such extrapolations, should probably not be used as calibration points for molecular dating. One such extrapolated reconstruction is a putative Nymphaeales/Nymphaeaceae flower based 9s. on a single specimen that lacks petals/tepals and stamens. A reconstruction of these organs was presented, indicating filamentous stamens with basi- fixed anthers and elongate perianth parts based solely on the nature and position of attachment scars around the floral rim (Friis et al., 2001) 2. IDENTIFICATION AND TAXONOMIC PLACEMENT OF FOSSILS Due to the uncertainty and ambiguity inherent in the fossil record, one of the most challenging aspects of paleontology is the identification and taxonomic + placement of fossils. For example, during the late 1800s and until at least the 1980s, remains of leaves, flowers, and fruits of the Tertiary age were mostly described based solely on direct comparisons with extant genera and species based on the premise that for every given fossil there is one extant species that is most similar (Wolfe, 1973). These fossil species were often considered to be direct ancestors, or sister taxa, ofextant species without explicit phylogenetic analysis and based on phenetic similarity alone. Unfortunately, these associations were too often based on vague character- istics, and the methodology, known as picture-matching, predominated in paleobotany until the 1980s. The use of this technique produced a great number of fossil a misidentifications, as well as a high proportion o incorrect taxonomic placements (for further discussion, see Dilcher, 1973; Hermsen & Gandolfo, 2004) The most problematic aspect of poor reconstruc- tions, fossil misidentifications, and taxonomic mis- placements results from the use of suspect fossil data by non-paleobotanists in studies that utilize fossil occurrences to estimate ages of clades. Such uses are not restricted to molecular dating but also include — Xogeographical studies in which the age of clades may be used to postulate historical patterns of vicariance and dispersal. Unfortunately, the fossils used in such studies are often not critically evaluated and the identifications and taxonomic affinities are taken for granted from the literature. An example of this problem is provided by Schultheis and Donoghue (2004), fossil assigned to the modern genus Ribes L. Axelrod 1996, UCMP 11156; Fig. 1A) as the oldest reliable fossil occur- f Ribes in North molecular phylogeny who used (Ribes species rence of America in a study of the Ribes. > fossil in and biogeography of of the a review of the group by a paleobotanist revealed a However, a subsequent careful study — ack of consistent characters with modern Ribes. and this calls into question the Another example is a fossil identified ee Hermsen, 2005) any conclu- sions based on misidentification (Fig. 1A). the putatively araceous monocot Arisaema herperia Knowlton 1926 (USNM 36885; Fig. 1B). Subsequent- ly, A. herperia was used to estimate the age of the North America, Africa based on a Bayesian chloroplast 2004). The A. herperia fossil and previous Tertiary floristic link between Asia, and East phylogeny (Renner et al., is based on a single specimen, Knowlton’s identification was performed using the picture- matching method of gross similarity. Hermsen and Gandolfo (2004 ac Q S reexamined A. herperia and found a a definitive the Araceae. Furthermore, in a previous paper not cited f sufficient characters to make taxonomic placement with Arisaema or family by Renner et al., A. herperia had been removed from Arisaema and reassigned to Liquidambar L. by Brown (1946). Although the taxonomic position p of this fossil is still in dispute (Renner et al., 2004; Hermsen & Gandolfo, 2004). it thal paleobotanists face in identifying and assigning fossils illustrates the difficulties and the problem of uncritical acceptance of fossil identifications from the older literature. Pollen grains are also frequently used as calibration points (e.g.. Bremer, 2000), and these present a challenge as well (see paragraph above). Different establishment the dis- interpretation of the fossils and (1996) assessment in of homologies. | Nixon cussed in detail issues of homology fossils; these will not be repeated here. Different researchers often interpret the same fossil structures differently, increases sometimes dramatically so, and this the uncertainty of homology. Imperfect preservation, lack of context of detached parts. missing features, and lack of precise detail (e.g.. anatomy) all contribute to more complex and difficult (1996) rovided several examples of these issues. including p | homology assessment in fossils. Nixon differing interpretations of the homology of the multi- ovulate cupules of Mesozoic seed ferns such as 38 Annals of the Missouri Botanical Garden “igure 1. Ribes sp. (UCMP 11156) housed at the l Califonia —.b. Type specimen of Arisaema heperia ede History, Smithsonian ME tion, Washington, D.C. Thomas as compared with angiosperm Undoubtedly, influenced Caytonia ovules. differences in interpretation mav be either by different views of background knowledge (e.g... a given scenario of o Le] e eo morphological change) or may be a form of verificationism—fitling our assessment of homology to reflect a given (often popular) view of phylogenetic history. Even when fossils are considered to be very well preserved, there are a limited number of characters one can investigate, and homology assessments are in general weaker than with the extant plants used for comparisons (Nixon, 1990). 3. FOSSIL SAMPLING Another aspect that has added confusion to the process of the selection of fossils for calibration is the The least spotty and incomplete fossil record sample. incompleteness of the record is apparent in niversity of California Museum of Paleontology. Berkeley. (USNM 36885) housed at the National Museum of Natural three dimensions: temporal, geographic, and taxo- nomic. Originally, paleontological collections were secondary particular research programs, and most paleobotany collections were maintained by scientists coming from different backgrounds. The majority were geologists gathering data for physical geological purposes or for commercial endeavors such as mines and oil prospecting. Therefore. the information recorded for specimens within these collections (exact locality, strata and type of environment in which deposition occurred, other fossil associations, etc.) was minimal. As with older herbarium labels. old fossil collections often lack precise location data and other informative facts, and sometimes the information ossil provided on the labels was incorrect. Today, collections are usually made with an emphasis on careful stratigraphic mapping and, more recently, global positioning system waypoints (GPS data) to allow fairly exact relocation of localities and quarries. Volume 95, Number 1 2008 Gandolfo et al. Fossils for Calibration of Molecular Dating Models Special care is put into documenting vertical and horizontal variations, and, usually, extensive and intensive sampling is undertaken at productive sites. This approach increases the probability of collecting more than a single specimen per taxon, and therefore, taxonomic assignment and identification are more accurale. This more thorough sampling also reduces the proportion of missing data when multiple documented specimens from the same site become available. t. GEOLOGY AND AGE Perhaps one of the most pervasive issues related to the utilization of fossils for calibration of molecular dating methods is the tendency to accept the original description, taxonomic assignment, and geologic (Hermsen & classifications, * dating without further investigation Gandolfo, 2004). As with classification of geological groups, formations, strata, plant horizons, etc., and their assigned (calculated) age are Thus, calibration of a molecular tree, not static. when considering a fossil for iL is necessary to thoroughly search for later modifications of both the taxonomic assignment and, just as importantly, changes in the assigned age. The early angiosperm Archaefructus is a good example of an original geologic age that was disputed and subsequently changed from Late Jurassic to Early Cretaceous based on additional 2002). Other issues are purely data (Sun et al., terminological. Many competing stratigraphic schemes were used until around 1880, when geologists started to address the problem of the necessity of a standard stratigraphic scale. Although it took them more than 100 years to produce the guidelines for defining global chrono- stratigraphic units, the International Commission on Stratigraphy ([ICS], a commission within the Interna- tional Union of Geological Sciences [IUGS]) produced the first Global Stratigraphie Chart (Cowie & Basset, 1989), which is revised frequently. Different geologic en methods often produce different estimated ages r the same deposits. It is now generally accepted are obtained by data (U-Th-Pb, Rb- it is perhaps more common to n the most accurate paleomagnetic and radiometric Sr, K-Ar). However, date by other means, such as stratigraphy, biostratig- ages raphy, palynology, and presence or absence of indicator taxa or assemblages of taxa in combination. Both (floristic) dating have caveats, including unconformity of strata stratigraphic and biostratigraphic and the potential for what are considered to be unique floral assemblages to persist or reappear. Therefore, depending on what method has been used to calculate the age of strata, the assigned age and method should be carefully considered when selecting calibration fossils. A Hunco paleoflora that occurs in Patagonia, Argentina. clear example of this is the Laguna del Initially, its age was considered Miocene (23.03— 5.33 Ma) based on comparisons with other paleofloras of South America (Berry, 1932), but, based on argon/ potassium dating, Archangelsky (1974) suggested a —55.8 Ma). Mazzoni et al. (1991) age al 51—43 Ma and, — Paleocene age (65.5 ater estimated the based on ignimbrite radiometric dating, most recently, Wilf et al. (2003), using paleomagnetic data, deter- mined the age to be early Eocene (52.8-51.7 Ma). Depending on which of these age ranges are used, fossils from this site might produce very disparate results if used for calibration of molecular trees. The problems of the most accurate date for a particular locality or stratum not only are dependent on dating methods but also are related to the instability of the geological scale, which has changed through time. In 1937, most researchers agreed that the Tertiary started 70 Ma, but today we accept 65.5 Ma as the basal age for this period. Overall, the assignment of the age of the beginning of the Tertiary has fluctuated between 70 and 63 Ma. In addition, the duration of the periods within the Tertiary has changed, with the Oligocene having a 16-million-year duration in 1937 (from 48-32 Ma) to a 17.4-million- year duration in 1982 (from 38—24.6 Ma), and today it has 10.87-million-year duration from 33.9-23.03 Ma 2004; I when (Gradstein et al., ‘able 1). Such differences, if not considered using fossils as calibration points, can also result in widely differing estimates of ages. This kind of variability is observed for all the eras, periods, and stages of (Gradstein et al., 2004. of the geological scale THE SOLUTIONS It is impossible to completely eliminate problems associated with selection of fossils as calibration points for molecular dating, but such problems as discussed above can be reduced to a minimum if certain steps are taken. This should reduce the errors introduced into such analyses and provide Undoubtedly, most of the suggested steps that require a paleobot- more repeatable results. some or perhaps anist or paleobotanical or taxonomic knowledge to undertake may be beyond the scope of many studies. We would hope that the same care and concern should go into the selection of fossils as that which goes into methods of the the final result of the selection of appropriate genes, phylogenetic analysis, and taxon sampling ii original studies. In many ways, molecular dating may be more dependent on the selection of a calibration fossil than on the particular 40 Annals of the Missouri Botanical Garden Table I. Comparison at the period level of selected Paleogene time scales from 1937-2004: modified from Gradstein et al. (2004). Age in million years (Ma). B, beginning of period: E. end of period: D. duration of period. Paleogene Paleocene (Ma) Eocene (Ma) Oligocene (Ma) B E D p p D B E D 193% 70 08 2 08 18 20 K 32 lo 1960 (0 00 10 00 10 20 10 25 I5 1972 05 253.5 11.5 503.5 11.5 16 11.5 22.5 l5 1978 65 59.9 115 03.5 37 16.5 37 24 13 1982 65 54.9 10.1 54.9 38 16.9 38 24.0 13.4 1985 00.5 o1. 0.0 TA 30.0 21.2 36.5 23.0 12.9 1987 00.5 54 12.5 54 30 le 30 20.2 10.8 1990 05 56.5 0.5 06.5 35.4 21.1 35.1 23.3 12.1 1995 05 54.5 10.5 54.5 32.4 20.0 33.7 23.8 9.9 2004 05.5 55.8 9.7 22.0 33.9 21.9 33.9 23.9 10.0 gene sequences utilized for the phylogeny, assuming that multiple sequences are selected. RECOMMENDATION 1: TAXON SELECTION Verification of the fossil identification: (a) If possible, verify the identity of the fossil with qualified paleobotanists. O = ten, suspect fossil taxa may be well known to the paleobotanical community, but there may be relatively little published on these subsequent to the original description. For those who are qualified, it is, of course, better to directly study the actual specimen(s) if possible. (b) Evaluate the original diagnosis provided by the author of the fossil genus/species. In many older papers, fossils lack a formal diagnosis, and such fossils are immediately suspect. in. terms of their identity. At a minimum, some characteristics that are unique to the group to which the fossil is assigned should be determined. If it is not possible to characterize the fossil with a diagnosis that clearly associates it with its presumed modern relatives and eliminates assigning it to multiple other taxa, it should probably be excluded immediately as a candidate for a calibration point. (c) Check for all taxonomic/nomenclatural changes subsequent to the original publication. Sometimes, this is sufficient to eliminate a fossil from consideration. (d) Check for synapomorphies that provide unequiv- ocal taxonomic assignment of the fossil. If the fossil lacks synapomorphies or at least an assemblage of characters that confirmed its taxonomic placement, it should be rejected and not used as a calibration point. (e) If the fossil taxon has not been previously in- cluded in a phylogenetic analysis, it should be. if at all possible. This can often be done by includ- ing the fossil taxon in an existing morphological analysis of modern taxa that are within the scope of the identification. RECOMMENDATION 2: FOSSIL SOURCE (a) Confirm the data associated with the collection of the fossils: formation, strata, and locality. Initially, these should be obtained from the original paper, but because concepts of formation and strata can change, these need to be verified. (b) Verify that the specimen(s) are still available. This may require contacting the presumed collection in which the specimen is housed. Any fossil taxon used as a calibration point should have available specimens and, presumably, a clearly designated type specimen, so that future studies can evaluate the quality of the assignment. RECOMMENDATION 3: AGE (a) Confirm the age of the associated formation, strata, and locality. Again, because concepts change and strata are sometimes remapped, it is important to determine if any changes have occurred since publication of the fossil. (b) Check in the geological literature for changes in the estimated age of the strata since the date of original publication of the fossil. (e) Use the youngest, most probable, most precise period (f the period is Eocene, use 33.9 Ma and not 55.8 Ma). (d) Provide the correct bibliographic citation for the age adopted and, in particular, the source of any revised e I ge estimates for the strata in which the fossil(s) occur. Volume 95, Number 1 2008 Gandolfo et al Fossils for Calibration of Molecular Dating odels (CONCLUSIONS The use of fossils as calibration points in molecular dating 1s an exciting and potentially useful endeavor. independent of which molecular dating methods are used or the theoretical justifications for particular approaches. The renewed focus on fossils in this context has exposed numerous issues that have been well known to paleobotanists, but may not be apparent to workers seeking to simply obtain a calibration point as easily as possible. Providing carefully identified and accurately aged fossils that can be utilized by molecular biologists in such studies is the responsi- bility of the paleobotanist. Unfortunately, the vast body of paleobotanical work is largely inappropriate for use to calibrate molecular trees without careful scrutiny and attention to issues related to traditional, often nonrepeatable, approaches to fossil study. The paleobotanist must be a primary partner in these endeavors, and, hopefully, paleobotanists will increas- that molecular ingly become collaborators in studies use molecular dating methods. At a minimum, botanists should carefully consider which fossils they are using, and paleobotanists should seek to provide fossils in a verifiable phylogenetic context based on combined molecular and morphological analyses. In these ways, both fields of endeavor will be improved. Literature Cited a tafoflora de la Ameghiniana 11: Archangelsky, S. 1974. Sobre la edad de | Laguna del Hunco. Provincia de Chubut. 413—417 Bremer, K. 2000. Early Cretaceous lineages of monocot flowering glanis. Proc. Na Acad. Sci. U.S.A 7 17074711. M. Friis € B. Bremer. 2004. Molecular nioena dating of asterids reta p shows early Cretaceous diversification. Syst. Bi 505 Brown, R. W. 1946. Alterations in some 2 ind living floras. J. Wash. Acad. Sci. 36: 344-355. Cowie, V. & M. C. Bassett (editors). 1989. Global timo Chart with geochronometric and magneto- stratigraphic calibri ation. Episodes 12(2), 1994. Magnolidea and the implications. /n P. suppl. Crepet, Nixon. F P ers of Turonian E. M. E ad ndress Friis GN E arly Evolution of Hs rs. Pl. Syst. Evol. SU "a 1998a. Cretaceous. ( Turon an Clusiaceae from the Late f w Jersey and implications re erg the history of bee SA ed Amer. J. Bot. 85: 1122-1133 ————, 1998b. dieran O de the Late Cretaceous of New Jersey. r. J. Bot. e DA ———— E, M. Fr 1992. Oldest fossil flowers of a lidac eous affinities from ee Late eous of New Jerse Acad. Sei. U.S.A. "s Pond 8989. Two new fossil flowers of & J. Freudenstein. . Proc. Natl. Gandolfo. 2005. An Jerseyanthus b ipic J —— —— & M. A. extinct sóla anthoid taxon, Late Cretac 1475-1485. Dilcher. D. L. 1973. A of the Bauen n North America. P s obot from the ceous 0 (ew Jersey. Amer. J. ot. 9 Eoc One A of anist 20: rchaeanthus: ey revision Crane. 1984. sails angiosper rm from the Cenomanian io the 7 n [o = = = © = ~ Q £o E a w 'a. Ann. Mi i 1. 2001. Phylogenetic patterns in dNosihieit Hemisphere plant geography. Int. Pl. Ser. 162: 541-852. Doyle, J. A. & M. J "LHP diversificatic >» M. 1990. Silv new sax gata rea from the Late Cretaceous of Biol. ' P. R. o e K. R. Pedersen. 1993. ves and Paleobiology 19: 141— Donoghue. Friis o suecicum gen. et sp. nov., a Sweden. 1986. Floral ev idenc «e for Cretaceous chloranthoid angiosperms. Nature 320: 163- 164. Pedersen & P. R. Crane. 1992. Esqueria gen. nov., fossil flowers with combretaceous features from the r. 41: 1—45. P. R. Crane. 1994. ov.—AÁ caly- late E re ae eous of ac Bio . Eklund, Keen a da o A gen sp. canthaceous flower from the potus np ee, Int. J. Pl Cretaceous) of eastern North America. L552 7 í 12-785. .. R. Pedersen & P. R. Crane. 1995. oe Saleen gen. et sp. nov., a new Early staceous plant with similarities to Circaeaster an extant ^ dde Amer. J. Bot. 82: 933—943 & 2001. Early Cretaceous. Nature 410: ———, ———— & ———— Fossil evidence of water lilies (Nymphacales) in the 357-360. un fo. M. A.. K. C. Nixon, W. L. Crepet & G. E. Ratcliffe. . A new fossil fern assignable to Gleicheniaceae from E. ens sediments of New Jersey. Amer. J. Bot. 84: 483—493. . Crepet. 1998a. Tylerianthus crossmanensis gen. et sp. nov. (aff. uid n from A Upper Cretaceous of New Jersey. Amer. J. Bot. 85: 370—380. ———, — . 1998b. A new fossil flower from the Turonian of New jene y: Dressianthia rera pis gen. et sp. nov. di dian) M . J. Bot. 85: 964— JOO. se D : review of the Early Cretaceous rec¢ . Pp. . Wilson & D. Morrison (editors), Proce ee ka 2s 2 Interna- tional Conference on the Comparative Biology of the Monocotyledons. CSIRO, Sydney. ——— & — —— ' 2002. Triuridaceae fossil flowers from the Upper Cretaceous of New Jersey. Amer. J. Bot 89: 1940-195 e 9 . The oldest e omple le fossil leer of Nymph aeaceae Bonn implications for the complex insect entrapment pollination mechanism in early angio- sperms. Proc. Natl. Acad. Sci. U.S.A. 101: 8056-8060. Gould, R. E. & T. Delevoryas. 1977. The biology of Glossopteris: Evidence from pe Bc seed- SERE and pollen-be aring organs. Alcheringa 1: 3 Gradstein, F., gg & A. Smith. Tn A E Time Scale. Cambridge Univ. Press, Cambridge Graur, D. € W. Martin. 2004. Reading ihe entrails of chickens: Molecular timescales of evolution and the 86. illusion o at precision. Trends Genet. 20: 80— Annals of the Missouri Botanical Garden Hedges, S. B., P. H. Parker, >. Sibley € S. Kumar. 1996. Continental. bre uh Sa nem yaa of Nature 381: 226-229, Herendeen, repet & K. Nixon. 1993. Chloranthus- like stamens from the a Cretaceous of Amer. J. Bot. e 865—861. & .19 994. Fossil flowers and o Pl birds and AE: S., W. "y y Jersey. CREAS ON of Lauraceae due a Upper E relaceous of New Jer Syst. Ev e —40. E Puebla, R. Lupia, P. EU m R. Crane & J. A preliminary conspectus of the Allon E a. F lora from the Late Cretaceous (Late Santonian) of central 2 Hermsen, E. J. å Grossulariaceae in the Cretaceous and ' y oft United States and Canada. Ph.D. Thesis, 5 Univer- sity, Ithac a, New Yor ————«&M.A.( Canddio. 2004. Fossil identification used in molecular dating studies should be justified. Amer. J. Bot. , ac- cessed 16 Oc tober 20€ ; & W. L. Crepet. 2003. (aff. Iteaceae), a fossil saxifrage from USA. J. Bot. 90 co al gen. nov. the Late Cretaceous of New Jersey, Amer. 1373-1388. Huelsenbeck, J. P. & F. Ronquist. 2001. MrBayes: Bayesian inference of xod ny. Bioinformaties 17: 754—755. Manos, P. M. Stanf 2001. The historical E da agaceae: Tracking the Tertiary history Northern ord. owe zap of temperate and subtropical fore sts of the Gemienhece, Int. J. PL Sei y pda -893. 1. M.. K. Kawashita, S. Harrison & E. Aragón. Edades liada Eocenas en el borde oce a seol. Argent. 46: Mazzoni, M 199] del E izo Norpatagónico. Revista Asoc. 150-15 Meyer- e N. Taylor & E. L. Taylor. 1992. Re- constructing t on fuite seed fern uoce Evidence from the : 341-344. Triassic of Antarctica. a 25: — 1993. Petrified stems bearing a ee mum from os ‘Triassic of Antarctica. d 36: 337 Morley, . W. Die an 2003. Missing fossils, mol ind cloc » Pan s origin of the Melastomataceae. Amer. J : 1638- 16: . 1996. Pale obotany in cladisties pale sohatamy: Enlightenment and uncertainty. Rev. bot. Palynol. 90: N & W. L. Crepet. sand cladisties in Palaeo- 361-373. 3. Late Gretaceous fossil flowers ~ J. Bot. 80: 616-623 iltiple Miocene Melastomalaceae Africa, and India. Philos. of ericalean Ped Renner, S. S. dispersal between Madagascar, 2 BS -— = Trans., Ser. B. 359: 1485-1494. —. 200 4b. Bayesian analysis of combined chloroplast loci, using multiple calibrations, support the recent arrival of Me lastomataceae 1 Africa and Madagascar. Amer. J Bot. es 1427-143: JOS. Relaxe d molecular eloc ks for dating historical plani: dits rsal events. Trends Pl. Sei. 11: 550-55 — ——— & K. Meyer. 2001. Melastomeae come full circ le: Bioge 'ographic n reconstruction and molecular clock dating. Evolution 55: 5-1324 G. C ec Meyer. 2001. Historic al la of Melastomataceae: The roles of Tertiary migration and long-distance dispersal. Amer. J. Bot. 88: 290 —1300 . L. Zhang & J. Murata. 2004. A chloroplast phylogeny of Arisaema (Araceae) o trates Tertiary floristic links betw America, and East Africa. Amer. J. Bot. € Sanderson, M. J. e A een dcs : 881 North —888. non-parametric aaa 'h to rale estimating divergence time in the absence of constancy. Molec. Biol. Evol. 14: 1218-123 ————, 2002. Estimating De rates A olecular —9 and divergence DÀ o likelihood approach. Molec a E vol 19: 101—109 Schultheis, L. M. & M. J. Donoghue. 2004. Molecular phylogeny and aie of Ribes (Grossulariaceae), vith an an on gooseberries (subg. Grossularia). Syst Bot. 77-90. Wak the Evolution A a N. € G. W. Rothwell. 2001. Plants. Cambridge Univ. Press. Paleobotany and a . G. - a Ji, a Dilcher, a Zheng, K. Nixon & E ang. a. Uer i laceae, a new basal angiosperm family. Science 296: 899-904 Taylor, T. N. & E. L. Taylor. 1993. The Evolution of Fossil Plants. Prentice Hall. Upper Saddle Biology and River, New Jersey van Tuinen, M. & pm * Dyke. 2004. Ca clocks using quet fossils 0: 74-86. LT s Jones & D. = ibration of galliform molecular and genetic partitions. Molec. Phylogen. a Wagstaff, S. J., M. ayly. Albach. 2002. Classification, origin, and diversification of the New Zealand hebes (Scrophulariaceae). Ann. Missouri eee. ur 89: e ue RA K. R. Johnson, J. icks, S. L. Wing & J. - bao a 2003. High au t 'rsity in Eocene South - : Evidence from Patagonia. Science 300: 122-1 wa N., V. Savolainen & M. W. Chase. 2001. Evolution of the angiosperms: Calibrating Ed d tree. Proc. Roy. Soc. London, Ser. B, Biol. Sci 8 2211-222 220. ies n A, 1973. Fossil forms of Amentiferae. Brittonia 25: Bell, P. S. Soltis & D. E. Soltis. 2005. and historical biogeography of Nym- 0: 693—704 33/ Yoo, Pa i D e limes phaeales. Syst. Bot. 30: 6 Zavada, M. S. 1983. Cane e. morphology of monocot pollen and evolutionary tr ls 1, 49: 3 ends of apertures and wall structures. Bot. Rev 31-379. 991. De Kenning ; character i e in pollen. Pp. 239-256 in S. Blackmore & S Polle n and Spores: Patterns of rena ation. Oxford Univ. Press, New York = arns (editors), PALEOBOTANY, EVIDENCE, AND Kevin C. Nixon! MOLECULAR DATING: AN EXAMPLE FROM THE NYMPHAEALES ABSTRACT recent years, most systematics studies have focused on phylogenetic analyses of molecular data sets. The latest trend has een to add molecular dating to these phylogenies utilizing methods such as nonparametric rate Wen: a and penalized likelihood (PL) and Soon these analyses using (often only one or very few) fossils. The suecess of such approaches is dependent on several assun iplions, ine ae a local clocklike behavior of evolution, i: accuracy of the phylogeny, the correct phylogenetic o of ossils, nd the consistency of particular fossils in extrapolating rates throughout a given phy oge enelic tree. An example ol Hes an analysis of the Nymphaeales is provided to illustrate inappropriate use of fossils in this context and faulty results based on inade quate and/or inappropriate analyses. Neither fossil identifications nor a particular method of molecular dating should be called into question based on the dispari of a single analysis. Indeed, fossil observations and molecular dating are often at odds due to failure of the data to meet minimum assumptions of a clocklike behavior and vui or inadequate sampling of extant taxa, molecular sequence data, and/or fossils ejection or acceptance of either the fossils or the molecular dates resulting from their use should be considered in light of direct analysis of the fossils and compare «d to a analyses using other fossils and/or other extant data sets. Rejection of fossils based on unexpected results is merely verificationism. Key words: Angiosperm, molecular clock, nonparametric rate smoothing, Nymphaeales, penalized likelihood. In recent years, the fields of plant and animal rates of particular genes are constant either globally systematics have progressed from a relatively subjec- (the ideal situation) or sufficiently within parts of the be amenable to models that attempt to tive and often authoritarian process of classification tree & (see Cronquist, 1981) to one based on monophyletic minimize rate change from one node to another. This groupings following the recommendations of Hennig idea, that evolution may occur at least locally in a (1966), although proponents of paraphyletic classifi- clocklike manner, can be traced back at least four cations persist (e.g., Brummit, 2006: Hörandl, 2007). decades (Zuckerkandl & Pauling, 1962, 1965), the The existence of comparative DNA sequence data for — although the recent discussion of this topic ii most major taxa now provides a basis for coordinating context of phylogenetic analysis has shifted somewhat, phylogenies of extant species and fossils via simulta- and several issues deemed important in the original neous molecular-morphological analyses that include dialogue (such as gene neutrality: see King & Jukes, both living and extinct terminal taxa (numerous 1969) are seldom mentioned in current discussions. In papers, including Sun et al., 2002). In turn, such many ways, the idea of a molecular clock underlies all combined analyses (or surrogates that attempt to place currently used model-based approaches such as can | maximum likelihood or Bayesian phylogenetic anal- == fossils on independently calculated phylogenies be used to estimate minimum ages of hypothetical — ysis, which rely on predictable patterns of molecular ancestral nodes (Crepet et al., 2004). A recent and change that can be designated in a model with increasingly popular use of this approach is to use relatively few parameters. Clocklike evolution is also fossil calibration points to estimate evolutionary rate the underlying justification for using neighbor-joining parameters for trees on the basis of molecular data trees as an approximation of phylogenetic pattern, and thus extend the dating of nodes beyond the based on a model of minimum evolution and a least- immediate attachment point of the fossils (Sanderson squares—based algorithm (Saitou & Nei, 1987). et al., 2004). These methods can also be used to It is not the intent of this paper to review or criticize calculate statistical confidence intervals for these various methods of node-dating using molecular trees estimated ages. In the broadest sense, such analyses calibrated with fossil data: indeed, some of the rely on what was originally referred to as the proponents of these methods already have provided molecular clock-—the assumption that evolutionary relatively strong criticisms and caveats (Sanderson el 'L. H. Bailey Hortorium, Department of Plant Biology, Cornell University, Ithaca, New York 14853. U.S.A. KC a lled doi: 10.341 7/2003 17063 ANN. Missourt Bor. Garp. 95: 43-50. PUBLISHED ON 11 APRIL 2008. 44 Annals of the Missouri Botanical Garden al. 2004). To a great extent, the acceptance or One major assumption of the original approach to rejection of molecular-dating approaches is based on faith in the general concept of the molecular clock and the accuracy and/or reliability of particular models that are incorporated. into such methods. In many cases, age estimates of important nodes based on molecular dating are far older (and only rarely. younger) than minimum age estimates based on existing fossil evidence alone (see an excellent review of these issues by Graur & Martin, 2004). Of interest here is not whether molecular dating is accurate but. instead, what the relationship between these two approaches is and how to resolve discrepancies in results. The two approaches can be summarized as follows: (1) Estimation of the minimum ages of clades is based on direct observations of fossils—their identification in a phylogenetic context —combined in which they occur and with dating of the strata subsequent estimation of minimum ages for clades based on the oldest fossil. record included by that 2004), may nol involve additional extrapolation of ages based clade (e.g., Crepet et al., This method may or on branch lengths. The phylogeny on which the estimations are based might be molecular, morpho- logical, or combined, but there must be some way to explicitly assign fossil taxa to positions in the tree. (2 Estimation of ages is based on molecular-generated branch lengths and topologies interpreted with a particular algorithm (and implicit model), such as penalized likelihood (PL) or nonparametric smoothing (NPRS: Sanderson, 1997), with one or more fossils identified and dated by the rate and calibrated above methods. A molecular tree must be available. and. in general, there has been little discussion of the fossil selection and how te a) exact requirements for determine exactly where it should be placed in the 2008 In order to discuss problems that arise when these Gandolfo et al.. tree (see two approaches yield different answers, 1 will use a particular example that involves fossils that I have studied directly to illustrate some of these issues. BACKGROUND: THE MOLECULAR CLOCK lt is generally accepted that the idea of the molecular clock was first introduced by Zuckerkand and Pauling (1962, 1905). 1960s, the idea of potential to use sueh clocks for constructing phyloge- Soon after. in the mid- to late molecular clocks and the estimating evolutionary rates, and/or the Because DNA sequencing was not feasible nelic trees, age of particular clades emerged. at that estimating most of the discussion of molecular clock tree time, construction and dating centered on (very scarce) protein. sequence data, particularly cytochrome C. molecular clocks was that neutral genes or sites, and what can be termed synonymous changes at these sites, were the most likely to have clocklike rates and thus were the best candidates for phylogenetic reconstruction. It is important to note that the concept NL of (selective) neutrality in this context excludes both sites that are variable (and selected for or against in (mem various ways), as well as, and perhaps most im- portantly, sites that are conserved (e.g.. stabilizing selection) and show little variation. A truly neutral site is one in which selection is absent, whether it varies much or little. At one point, the idea that most or all evolutionary changes are actually (or virtually) neutral also entered the debate. This neutral theory—the idea that most evolutionary change is non-Darwinian—is closely tied to the molecular clock, as can be seen in the following 1969: 796): non-Darwinian change equals the rate of selectively quotation (King & Jukes, “The rate of — neutral mutation and is independent of environmental fluctuations and of population size. For a given protein, the rate of such change should be nearly constant. Darwinian change, in contrast, is under the influence of changing environment, adaptive radia- —— tion, fluctuations in population size, and such factors as adjustment to major changes in the genetic background. Thus it might well be subject to bursts of rapid change in some species and relative stability in others." Following this reasoning, genes (or individual nucleotide bases) undergoing selective (Darwinian) evolution will be poor candidates for a molecular clock. Sites under selective pressure will have heterogeneous both vertically within clades and horizontally Such sites will not exhibit. clocklike rates, among clades. behavior in their rate of evolutionary change. Some recent papers have suggested that bases al codon third positions have more of a clocklike pattern of evolutionary rate than first- or second-position Mercer & Roth, 2003). This is intuitive— if substitutions in third positions are synonymous in bases Les terms of the proteins coded for, then they would be neutral in the sense of the above discussion and might be better candidates for modeling a molecular clock. The potential for third-position bases to provide informative variation has also been (1999) found that more groups were supported in one large molecular analysis phylogenetically debated, and Kállersjó et al. by (presumably neutral) third-position bases than by first and second positions. This may be due to rate issues (because they are neutral and evolve faster, variation in third-position bases resolves more — terminal groups). Át any rate. the issues of saturation and the Felsenstein zone (see Felsenstein, 1978) have Volume 95, Number 1 2008 ixon Paleobotany, Evidence, and Molecular Dating generally been invoked to suggest that third positions are less useful due to repeated substitutions. These issues are not meant to be addressed here but certainly are necessary to an understanding of how molecular dating might perform with different data sets and different analyses (e.g., parsimony vs. maximum likelihood). In a recent review of the subject, Benton and Ayala (2003) did not address the primary issue of whether evolutionary rates are clocklike or issues related to Darwinian versus non-Darwinian evolution. Instead, they framed the problem as a conflict between results from fossils providing dates that are younger and molecular clock methods providing dates that are much older. Anecdotally, they provide an example where, with additional data and analyses. the two disparate estimates come closer together, and they present this as a hopeful indication that we are on the road to reconciling the problems associated with molecular clock dating. It is important to note that although some of the NPRS most popular molecular-dating methods, e.g., 1997), on the surface may not seem to (Sanderson, require rate. constancy, they do assume local rate constancy as opposed to global rate constancy. If rates are completely unpredictable (i.e., poorly correlated with time) and/or change rapidly and frequently within and among clades, even methods such as NPRS will fail to provide results. Other issues related to the selection of suitable fossils to calibrate molecular dating are discussed elsewhere in this issue (Gandolfo et al., 2008). AN EXAMPLE: THE NYMPHAEALES In the past two decades, molecular analyses of angiosperms have placed various modern groups as concordant sister taxa to the bulk of the extant based on 1993; Nixon, 1999) to various permutations with Amborella Baill. as the sole sister angiosperms, ranging from Ceratophyllum L. rbcL alone (Chase et al., taxon of all remaining extant angiosperms (e.g., Soltis et al., 2000). In these latter trees, the nymphaealean clade, along with Illiciales and Austrobaileyales, is generally placed as a sister group of all remaining angiosperms excluding Amborella. It should be noted that. statements regarding these trees often suggest that primitive angiosperm, Amborella (or Ceratophyllum) is a basal or which of course is merely a misinterpretation of sister-group relationships (such context of fossils has 2002). that if the three-gene tree (Soltis et al., 2000) i terminology in the more meaning; Sun et al., It is easily understood correct, then Amborella and any extant angiosperm diverged from the common ancestor of angiosperms at the same time, and therefore the lineage culminating in Amborella, have had the same amount of time to diverge from the ancestral angiosperm as any orchid, mistletoe, grass, or Texas bluebell. Amborella is no more basal within the angiosperms than any other extant species of angiosperm. Because of their aquatic nature and presumably early divergence. (whether termed basal or not), the Nymphaeales has attracted considerable attention in the context of the putatively primitive angiosperms. Indeed, this interest has increased with the discovery that some of the earliest identifiable angiosperm fossils appear to be aquatic (Sun et al., 2002) A recent paper by Yoo et al. (2005) that focuses on Nymphaeales provides a useful example of the conflict between traditional paleobotanical evidence and the results of molecular-dating methods. Yoo et al. (2005) used four approaches (strict molecular clock, NPRS, PL, Bayesian) times for the modern erown group of Nymphaeales as in an attempt to calculate divergence well as the age of the angiosperms. They utilized a previously published morphological cladogram of 2005: fig. 1 ‘his tree was used to map molecular Nymphaeales (Yoo et al., . from Les et al., i changes using sequences for rbeL, matK, and 18S that were downloaded from GenBank and aligned using CLUSTAL X (Thompson et al., 1997) with the default options, the alignment then (Yoo et al., 2005). terminals, including four (Ginkgo L., Gnetum L., Amborella, and was manually This tree gymnosperm outgroups Mill.. Taxus L.). Austrobaileyales Larix the Illiciales or repre- sented by three terminals (//lictum L.. Schisandra Michx., White), commonly recognized genera of Nymphaeales (Victo- , Euryale ' sad Nuphar Sm. from the Aubl. Austrobaileya C. T. and the eight ria Lindl., Nymphaea Salisb., Ondinea Hartog, Barclaya W le Nymphaeaceae, and Brasenia and Cabomba Schreb. from the Cabombaceae). It is important to note that Yoo et al. (2005) did not calculate a new cladogram based on the data matrix that they generated from downloaded sequences, but, instead, the pre-aligned molecular sequences were superimposed on the morphological tree from Les et al. (1999). then used to analyze the molecular matrix using the Les et al. (1999) tree as the input tree. The program r8s (Sanderson, 2003) was p t is not clear why Yoo et al. (2005) did not attempt a new analysis of the molecular data, given that the original data matrices needed to be included in order to estimate molecular divergence times. They then analyzed the methods of "[B]ecause all tests among lineages matrix/tree combination with various molecular dating available in 18s: were of rate heterogeneity highly significant, we used three approaches that have been Annals of the Missouri Botanical Garden proposed for use with heterogeneous rates. NPRS (Sanderson, 1997, 1998), PL (Sanderson, 2002), and a Kishino et al., 2005: 695). ysis by Yoo Bayesian method (Thorne et al., 1998; 2001: Thorne & Kishino, 2002)" (Yoo et al., The results of the molecular dating ana (2005: divergence for 697) suggested a relatively recent Nym- phaeales, even though Cretaceous fossils had been et al. Tertiary the crown group of previously described for the family Nymphaeaceae: “Our divergence time estimates indicate that extant Nymphaeales diversified relatively recently, whereas the stem lineage to Nymphaeales is old, based on a Nymphaeales from the Early Friis et al., 2001) and Nymphaeaceae from the middle 2004).” other words, the age of diversification found by Yoo et al. (2005) for Nymphaeales based on molecular-dating methods is considerably younger than the published fossil attributed to Cretaceous (125-115 mya; fossil attributed In Cretaceous (~90 mya; Gandolfo et al., fossil evidence (in contrast to the more common situation, where fossils are typically younger than estimated ages based on molecular trees). Oddly, Yoo tal. (2005: 697) use the rare but similar contradic- tory results found in other studies to validate their own results: “These results for Nymphaeales indicating recent diversification in an ancient lineage agree with similar findings for the basal angiosperms Chlor- anthaceae (Zhang & 2003) MMlicium (Illiciaceae; A. Morris, unpublished data). The fossil Renner, and record indicates clearly that Chloranthaceae represent one of the oldest angiosperm lineages, with unequiv- ocal reproductive structures resembling those of from the Barremian-Aptian boundary, approximately 125 mya (see Friis el al.. 1994, 1999; Friis, 1997; 2003; Eklund et al., 2004, for recent interpretations and lists of earlier references). Hedyosmum Doyle et al., However, divergence time estimates based on molec- ular data indicate that the extant genera of Chlor- anthaceae are relatively young (i.e., 60-29 mya for Hedyosmum, 22-11 mya for Chloranthus, and 18- 9 mya for Ascarina; Zhang & Renner, 2003).” There seems little point in belaboring the faulty logic in citing agreement with other, need analyses in which fossil ages are older than the age of crown groups estimated by molecular models. Usually, when molecular dating methods repeatedly provide significantly older dates than fossil evidence, the argument of an incomplete fossil record is typically voiced. to explain. the discrepancy. In contrast, since in their study Yoo et al. (2005) found significantly younger diversification of the crown group for Nymphaeales than well-studied fossils, they were forced to decide that the fossil identifications were incorrect. Unfortunately, for this interpretation, in the case of the Laurales example, the Chlorantha- ceae have a long unequivocal fossil record with some that the Hedyosmum Sw. from the early Cretaceous (Eklund et al.. 2004). this Hedyosmum did not diversify fossils have same characters as modern However, molecular clock dating tells us to the rate model(s) The pattern here is apparent—if the dates are too old then cannot be so; according until much later. we just do not have enough fossil data: if the dates are too young, then the fossils are misidentified. In other words. no independent tests of the molecular dates are allowed, and fossils are the only independent tests. the dating are not called into question in cases of conflict, = Apparently, particular results of the molecular only the fossil evidence. THE A ANGIOSPERMS E OF Apparently in order to provide evidence that both nymphaealean fossils from the Cretaceous (Friis et al., 2001; Gandolfo et al., 2004) were erroneously placed, Yoo et al, (2005: 695) then compared calibration of the age of the entire angiosperm clade using these two fossils (in two independent PL analyses) in r&s. Given the previous putative overestimation of the age Nymphaeales. the results were rather shocking: the angiosperm clade diverged 1093 million years ago (Ma; more than one billion years ago) based on Vicrovictoria Gandolfo, Nixon & Crepet (Gandolfo et al.. 2004) and 1457.3 Ma ago; Yoo et al., 2005: table (2001) fossil. Again, Yoo et al. (2005) suggested that this provides evidence that the fossils are phyloge- As the age of (almost 1.5 billion years 2) based on the Friis et al. netically misplaced. an aside, note high for the (2001) fossil calibration (to ca. years of almost 1.5 billion years). Yoo et al. (2005: performance of PL (in the sense of providing w angiosperms based on the 300.000 precision Friis et al. use the poor 700) proceed to — lal might be considered wildly inaccurate estimates of the the crown group Nym- origin of angiosperms) as an indictment of placement of Microvictoria as a “There seem to be two possible the phaeaceae fossil: ossil mu] explanations for the disparity between record for angiosperms and the age inferred here using PL and Microvictoria, as placed within Nym- phaeaceae by Gandolfo et al. (2004) as the calibration point. Either methods of estimating divergence times from molecular data are highly dubious or the netic placement of Microvictoria in Gandolfo et al. (2004) may need to be reconsidered.” e are more than two possible explanations for the problematically old estimates calculated using PL for the origin of angiosperms, and the dichotomy of two possible explanations presented by Yoo el al. (2005) is a false one. First. the fact that in a particular Volume 95, Number 1 Nixon 47 2008 Paleobotany, Evidence, and Molecular Dating instance a method fails does not make it highly the Nymphaeales lineage, e.g., a rapid diversification dubious, and certainly this is not a reasonable in the Cretaceous, followed by much slower rates within the crown groups. Second, a single fossil from a particular clade might not be sufficient to provide accurate results using molecular-dating methods. These interpretations (not considered by Yoo et al.. 2005) do not impugn molecular dating per se or traditional paleobotanical observation (which, after all, must be the basis for calibration of trees used in molecular dating). The idea of highly heterogeneous rates, with rapid diversification followed by relative stasis, was popularized a few decades ago by Eldredge conclusion. This is simply a classic case of a straw man, clearly intended to steer the reader toward the second alternative explanation provided. Maximum likelihood, parsimony, and almost any conceivable phylogenetic method fail under certain circumstances e.g., long-branch attraction or repulsion, a.k.a. the Felsenstein zone or the Farris zone; Siddall, 1998), and this fact alone does not impugn either of these widely used methods. Indeed, with any method one expects failure a certain percentage of the time, particularly when the assumptions of the method are and Gould (1972) when they championed the concept not met. Certainly, there is a strong possibility that of punctuated equilibrium. Perhaps this has fallen out evolution is not acting in a sufficiently clocklike of favor because it is difficult to model and, if manner in the region of Nymphaeales relative to other demonstrated parts of the angiosperm tree, and/or the rates in this part of the tree are not representative of rates elsewhere. Such heterogeneity of rate could result in fossils. from these clades being poor choices for calibration of an entire seed plant cladogram, and this to be common, would cause severe problems with simple rate-smoothing methods, which would tend to over-equalize rates among ancestors, descendants, and disparate clades. In this context, a ased on current large-scale phylogenetic trees, it is likely that the original angiosperms were not fully is actually the best alternative explanation. In other aquatic (Crepet et al., 2004), and it is possible that words, it may very well be that rates can differ there was a period of rapid diversification in t dramatically between ancestor and descendant nodes, nymphaealean clade i and using a single fossil from an area of the tree that is — habitat, followed by subsequent relative stasis in a heterogeneous for rates and not typical of average stable aquatic habitat. rates elsewhere will likely result in estimates that are obviously wrong. Given that Yoo et al. (2005) had already established the high significance of rate heterogeneity among lineages, it seems odd that they did not provide this obvious possible interpretation of their results. — 1e n invasion of the aquatic Yoo et al. (2005: 700) provide a rather complicated explanation. of why Microvictoria might have been misplaced: “Alternatively, Microvictoria may be misplaced in the phylogenetic analysis of Gandolfo et al. (2004), perhaps due to homoplasy in the crucial morphological characters scored and included in that In addition to the third explanation above (assump- study. That i tions of PL were not actually met due to highly heterogeneous rates), there is also the possibility, for those who adhere to the primacy of molecular-dating methods, that the angiosperms are much older than previously thought (interestingly, in the majority of results of molecular dating, this is the finding—that the age of major groups is much older than previously thought—yet, in those cases, the argument is typically provided that we merely lack fossils of sufficient age; \ see Graur & Martin, 2004). It seems that when ages Yoo et al. (2005) impinges on the effect of fossils with are acceptably older than expected based on fossil ambiguous characters on local parts of a consensus evidence alone, molecular dating is working well, and, tree, which is a well-known phenomenon (Nixon & when they are unacceptably older than expected, the Wheeler, 1992; Nixon, 1996). Yoo et al. (2005: 699) fossil record (or identification) is bad. [om s, a now-extinct assemblage of early angiosperms may have possessed suites of traits not found in any extant groups." This speculation needs little discussion. There might be any number of clades that are extinct and. never discovered as fossils that any particular existing fossil might be related to. The same can be said for any extant taxon as well. There is always the possibility of something as yet unobserved to counter what careful observation and analysis show. => As a relative aside, another issue brought forth by apparently misinterpret deresolution due to multiple If one accepts the general validity of molecular- placements of a fossil with actual clade conflict and dating methods, as well as the identification of the misinterpret the consensus tree presented by Gandolfo nymphaealean fossils by two different paleobotanical et al. (2004): “However, Gandolfo et al.’s topology research groups (Friis et al., 2001; Gandolfo et al., disagrees with the morphological analysis of Les et al. 2004)—one of which was included in cladistic (1999), who found strong support for Victoria and analysis—the “too old” (Yoo et al, 2005: 699) Euryale, consistent with many previous inferences of angiosperm results might suggest two interpretations. relationship in Nymphaeaceae; this sister-group First, perhaps there were significant changes in rate in relationship is not evident in Gandolfo et al.’s tree.” Annals of the Missouri Botanical Garden Apparently, Yoo et al. (2005) misunderstand poly- tomies—an unresolved polytomy is not in topological disagreement with any dichotomy that it implies— n systematics theory and the basis for kinds of 1990, complete discussion of these issues). In fact, the trees well known { various creation and interpretation ¢ consensus trees (see Nixon & Carpenter, for a identical to those the are (1999) Microvictoria is pruned (removed after the analysis). found by Gandolfo et al. reported by Les et al. when fossil with high support for the Victoria-Euryale clade. The 1992) reduces fact that Microvictoria floats (Nixon & Wheeler, within this clade does not mean that it support for Victoria-Euryale. which always have exactly the same relationship to other extant. Nym- phaeales in all most parsimonious trees (Gandolfo et al., 2004). Microvictoria is ambiguously It merely means that given these data. placed, sometimes as a sometimes to both Victoria and sister to Victoria, Euryale. The consensus is thus unresolved in this area. If viewing the component most parsimonious trees, when only extant taxa are considered, Victoria and Euryale always form a monophyletic group of two taxa, as in Les et al. (1999) and as stated by Gandolfo et al. (2004). (2004), because they restricted their analysis to the original Les et al. (1999) matrix, did not include the feature of — Note also that Gandolfo et : prickles on the outside of the flower : that since synapomorphy of the Victoria-Euryale clade would exclude Microvictoria from that clade the floral cup of Microvictoria is covered externally with bracts, not prickles. Other molecular-dating analyses, using much larger data sets and different fossils to calibrate their analyses, have found results that are in conflict with Yoo et al. (2005). Wikstróm et al. (2001) did an NPRS analysis on a much larger and more diverse angiosperm data set (507 taxa vs. 16 taxa, also with fagaceous Cretaceous three genes), using putatively fossils to calibrate one of the 567-taxon three-gene trees (Soltis et al., 2000). The results presented by Wikstróm et al. (2001) suggest the Nymphaeales of 171-153 Ma, with divergence of the an age for extant crown group occurring 144-111 Ma (consistent Micro- Given these results, one with. and considerably older than the fossil victoria, aged at ca. 90 Ma). might suggest that perhaps there is a problem with the Yoo et al. (2005) paper, f Microvictoria. Wikstróm et al. (2001: 2211) also note the important extant. clades of angiosperms: identification not the -= [ej p . fossils considered to be members of derived angiosperm lineages are being documented from increasingly older geological deposits. Crepet and Nixon (1998), Turonian for example, documented Clusiaceae from yenomenon of discovering ever-older examples of (90-88 Myr) deposits of New Jersey... . The presence of such. derived. groups in. Cenomanian-Campanian deposits implies either that we have underestimated the rapid and explosive nature of the angiosperm diversification or that eladogenesis in basal angio- earlier than fossil- Note that, a discrepancy between fossils and molecular dating (a Wikstróm et al. (2001) reasonable alternative expla- the the sperms took place considerably based estimates have indicated.” faced with very common occurrence). provide two scientific, nations, and neither reject fossils nor molecular-dating methods. CONCLUSIONS evolution (a rare at best. In the (2005). significant This is consistent Homogeneous rates of sequence molecular clock) seem to be example provided by Yoo et al. heterogeneity of rates was detected. fact, seen in cladograms (whether derived by likelihood or with many other studies and, is often easily parsimony) when branch lengths are displayed. Methods such as NPRS and PL merely move the molecular clock assumption to within-clade calcula- — lions (by smoothing rates locally) and are susceptible to poor calibration—using one or a few fossils from clades that are not. representative of rates in other parts of the tree. If rates are very slow near the calibration fossil. one will come up with overestimates ol ages in other parts of the tree; if rates are very fast, one will come up with underestimates. The possibility that two different fossils from differen parts of a clade could provide very different estimates on the same rees is not at all surprising. Because of this. if one wishes to use molecular-dating methods, one should utilize as many different fossils as possible (assuming they are well identified and phylogenetically placed): fossils might be used together or in separate Bae to compare how disparate the results become. The fact that a particular fossil provides different rate estimations than other fossils no more impugns the identification of the fossil than it does the assumption ft clocklike that representative of rates in the tree as a whole. Because rates near the fossil node are fossils are based on direct observations and rates cannot be observed, one must favor fossil identifica- tion over rate assumptions based on a model. By smoothing vertically through clades, NPR and related methods explicitly apply a local molecular clock where possible. Rate smoothing, by isolating f the is biased against rapid diversifi- chunks o tree, cation followed by slow evolutionary rates—t.e., punctuated equilibrium, a popular concept based on The issue here is not whether Microvictoria is or is not a observation of actual fossils through the record. Volume 95, Number 1 2008 Nixon Paleobotany, Evidence, and Molecular Dating member of the crown group of Nymphaeaceae/ Nymphaeales. The issue goes to the core of how we (2005) identifications based on careful morphological work Pa ( do science. Yoo et al. rejected fossil and cladistic analyses and presented only two alternative explanations: that the method is wrong from [Yoo et a 700]. or the angiosperms are unreasonably ("methods of estimating divergence times molecular data are highly dubious” 2005: old. These K + rima facie false premises were then used - question the placement of the nymphaealean fossils, which. as noted above, were consistent. with larger, more robust analyses (Wikstróm et al.. 2001). o J , (2005) did alternative morphological analysis of Microvictoria but Because Yoo et al. not an we supp instead created a scenario that some unspecified lineage may have existed with the same characters as modern Nymphaeaceae, their conclusion that Micro- 2001) are doubtfully part of the Nymphaeaceae crown clade 'he Yoo et al. (2005) paper is useful for pointing out what not to do in terms of fossil victoria (and the fossil from Friis et al.. cannot be evaluated. 1 data and molecular dating. Because rates may (and do) change radically within lineages, any particular fossil may be a poor choice for calibration of an entire Calibration with tree. Microvictoria produces one answer: calibration with fagaceous fossils on a larger n taxon-sampled) tree produces an entirely different result. This is not surprising. The conclusion presented here is simple: it is dangerous, and verificationist, to accept or reject fossils. based on whether they fit a preconceived It is likely that with could be notion of the correct results. almost any data set, a fossil found to calibrate a molecular-dating model that would provide a wide range of dates for a particular node of interest. f the results become the primary consideration in deciding which fossils should be used, the scientific endeavor is abandoned and the study becomes an exercise in verification. Because of these constraints, the most casually studied fossil by a competent paleobotanist, if placed in phylogenetic context with explicit characters, should be favored in every case over what might be termed a secondary analysis—and rejecting direct observation based on an interpretation of history constructed on a model. Literature Cited Benton, M. J. € F. J. Ayala. 2003. Dating the tree of life. Science 300: 1698-1700. Brummitt, E bs aa Am La bony fish? Taxon 55: 268-269. Chase. . Soltis & R. G. Olmstead, et al. 1993 PN "nelies s seed plants sequences from the pl: m gene rbcL. Ann. Missouri Bot. Gard. 80: 528— 980. Nixon, K. C. . C. Nixon & M. A. ee 2004. Fossil evidence an "i phylogeny: The of 1 Crepet, W. L., jor angiosperm c lade s based on mesofossil ae mac ‘crofossil evidence from retaceous deposits. Amer. J. Bot. 91: 1666-1682. A. 1981. A C BM n Integrated S Sy stem of e ae of . Press, Nev 1 lerendeen. Flowering Plants. Columbia b Doyle, J. A., H. Eklund & P. e ot in Chlor 2003. Floral oi: a of a morpho- logical phylogenetic analysis. Int. J. Pl. Sci. 164(suppl.): 5365-5382 Eklund, H., J. A. Doyle & P. Herendeen. 2004. Morphological pyogeuetie LE of living and fossil .PLS : 107-151. N. & 5. J. Could alternative to phy C hloranthaceae. nt. Eldredge, 972. * unctuated sa An ‘letic adu lism. Pp. 82-115 in T. J. M Schopf (editor), Models in Paleobiology. Freeman Cooper. San Francisco. J. 1978. Cases patibility methods will be positively misleading. Syst. Zool. 27: pem 41 Di Felse nstein, in which parsimony or com- . Fossil history of magnoliid angiosperms. Pp. 1211: A in IK. Iwatsuki . Raven (editors), Evo- lution and Diversification of Land Plants. Springer, Tokyo. . R. Crane. 1994. Angiosperm el struc ud from the pu Cretaceous of Portugal. Pl. Syst. Evol. 9, Pedersen & I . Early PEN diver- The pm of re associated with angio- sperm reproductive structures in Early Cretaceous floras Ann. Missouri us Gard. 86: 259—296. & . 200 ———, ———— & ————— . Fossil evidence of water lilies Tpha 60 sification: from Portugal. in d n Cretaceous. Nature 410: 397-300. Gandolfo, M. Nixon & W. L. Crepet. 2004. Cretaceous How rs ak Nymphaeaceae and implications for complex insect ida is pollination mechanisms in early SUD IR . Natl. Acad. Sei. U.S.A. 101 8056- 806 —— & ——. 2008. Selection calibrátion of molecular M models. Ann. 12 Gard 95: 34— of fossils for Missouri Bot. Graur, D Martin. 2004. Reading the entrails of chickens Moléculas timescales o ex] ution and the illusion of precision. Trends Genet. 20: 80- 86. Hennig, W. 1966 Boone System . Translated by D. Dwight & R. Zangerl. Univ. « Illinois iren Urbana. Horandl E 2007. Neglecting evolution is bad taxonomy. —5. Källersjö, 1 a V. A. t & J. S. Farris. 1999, uw increases phylogenetic structure. Cladistics 15: King, J. L. & T. H. Jukes. 1969. Non-Darwinian m er 164: 788-79 O ,J.1 don & W. J. Bruno. 2001. Performance a ps vergence estimation method under a evolution. Molec. Biol. Evol. babii model of rate l —361. Les, E. L. Sc n D. J. Padgett, M. Zanis, D. E. a js P. S. Soltis. 1999 floral evolution of water lites (Nymphaeales): A synthesis 2. Phylogeny, classification, and of non-molecular, rbcL, matK, and 185 rDNA data. Syst. Bot. 24: 28-46. Mercer, J. M. & V. L. Roth. 2003. The effects of Cenozoic global change on squirrel phylogeny. Science 299: 1568-1572. 1996. Paleobotany in cladistics and cladistics in Dn min, E nlightenment and uncertainty. Rev. Palaeo- 61-373. . Palynol. 90: : Annals of the Missouri Botanical Garden 999. The parsimony ratchet, a new method for . Cladisties 15: 407—414. 1996. On consensus, collaps- ibility ul clade concordance. Cladisties 12: 305—321. & Q. D. Wheeler. 1992. Extinction Mee the origin of species. Pp. 119-143 in M. J. Novacek & Q. D. "d ler S Extinction and Phylogeny. Columbia Univ. Press, "T parsimony analy sis & ;'arpenter. ew Yo Saitou, N. & M. Nei. 1987. The neighbor-joining method: Á new me ethod for reconstructing phylogenetic trees. Molec. Biol. Evol. 4: 406-425 S OA M. J. 1997. A estimating. divergence m in the absence of rate constancy. Molec. Biol. Evol. 14: 1218-1232 1998 Estimating rate and nonparametric approac h to molecular lige nies: Beyond the dis icu clock? Pp. 242-204 in D. E. Soltis, P. S. Sol & J. J. Doyle (editors). Molecular Systematies of Be Il, DNA Sequencing. Kluwer, lime in E )2 Q = © . Estimating absolute rates of molecular a and divergence times: À pen: alized likelihood approach. Molec. Biol. Evol. 19: 101-109 2003. ré 03. r8s: Inferring absolute rates of molecular and divergence times in the absence of 301—302. Bremer. mm molecular cloc k. Bioinformatics 19: y ds Ik Thorne, N. & K. C. 2004. Molec e "M nce on plant divergence times. Amer. J Bot. 91: 1656-1665. Siddall, " E. 1998. Success of parsimony in the four-taxon case: Long-branch Kw by likelihood in the Farris zone. A ladistics 14: 209—22( Soltis, D. E., P. S. Soltis, M. -i Chase, M. E. Mort, D. C. Me t M. Zanis, V. e n, W. J. Hahn, s. i Hoot, M. F. Fay, M. Axtell, 5. M. Swensen, L. M. Prince, W. J. Kress, K. C. Nixon & J. S. Farris. 2000, Angiosperm inferred. from 185 rDNA, rbcL, and atpB phylogeny se quence es. Bot. J. Linn. a 133: 381461. Sun, G., Q. Ji, D. L. Dile ther, S. L. Zheng, K. C. Nixon & X. F. Wang. 2002. Arc NE SRM a new basal angiosperm nily. rw 296: 899—904. J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin & far Thompson, J. I G 1997. The EA Es ClustalX windows interface: Flexible strategies for multiple is e alignment aided by quality analysis tools. Nucl. A ee Res. 24: 4876-4882 Thorne, J. L. € H. Kishino. 2002. Divergence time and e rale estimation with multilocus data. Syst. Biol. 689—702. N . Painter. 1998. Estimating the rate of evolution of the rate of molecular evolution. Molec. Biol Evol. 1647 a Wikstró avolainen & M. Chase. 2001. E is E the angiosperms: C d the family tree. Proc. bd a du 1, Ser. B, Biol. Sei. 2 a 2211-2220. Yoo, M.-J., . Bell, p 9. P & D. D times and historical bioge cea of Nym- phaeales. Syst. Bot. 30: 693—704. Zhang. L.-B. & S. Renner. 2003. Chloranthaceae as resolv by ome sequences. Int. Pl. Sei —5392. e deepest splits in . Pl $ 64(5 suppl.): S Zuckerkandl, E. € L. evolution, and genetic Kasha & B. Academic Press, New York —— ——, 1965. convergence in ataie: Pp. B. 1962. Molecular and. 1M heterogeneity. Pp. 189-22 Pullman (editors), Horizons in hem Evolutionary ee and 97 . Bryson & H. J. Vogel (editors), Evolving Genes and Proteins. Academic —166 in V Press, New HIDE AND GO SEEK: Robyn J. Burnham" ? WHAT DOES PRESENCE MEAN IN THE FOSSIL RECORD? ABSTRACT Our efforts to reconstruct accurate, complete records of events in vegetation history and in plant e ew na d depend f on accuracy in dating sediments, interpretation of structures preserved, reconstruction of who ganisms or communities from the preserved material, and interpretation of the interaction between past abundance and { sala esenc lis ua Hp examines the interaction Di 'en past abundance of a A ah plant and the ecc ol retriev al of that species 1 the fossil record. By examining records of recolonization in volcanic areas, records of invasive Lee spread, s ccession in denied habitats, and historical an patterns, we can ae pe estimates of the likelihood of appearance in fhe potential fossil rec an of newly evolved and reasonably successful species. The lag in discovery, recognition, and publication of a fossil as an import nt representative of a critical clade is also e annie a and is nde as a more important constraint on the use of fossils in paa evolutionary and ecological hypotheses than the recolonization rate. The lag be ween redi and public ation is particularly relevant in areas of the modern world Vus fossil plant-bearing deposits are either rare or inaccessible. Greater awareness of the density and reliability of the plant record should allow eee hos ists and D CN to ies ii not only time intervals but also geographic regions where the fossil record can be cuc largely at face value. At the e time, more effort should be focused on intense collecting efforts and training in areas where fossil = o are a present, but poorly collected and evaluated. Key words: Invasive species, paleobotany, plant evolution, stratigraphic resolution, taphonomy. I 8 À ne The fossil record has a pivotal role in retelling the important in research on plants in the geological story of plant evolution. Research into the molecular record basis for development of plant organs and the However, a caveat can almost always be raised: the application of D amplification and sequencing fossil record is incomplete. Ecological and evolution- techniques to outline plant phylogenies interfaces ary processes often operate over short time intervals, with direct evidence from the fossil record. The fossil and, therefore, the incomplete fossil record is best record has the potential to place time constraints on used with, or even replaced by, other types of data. So models of lineage evolution, diversification, specific what should the interpretation be when, at arm’s a traits and combinations, as well as to length, the fossil record indicates that many of the clarify the longevity of both traits and characters significant and defining features of the angiosperms (Rothwell, 1999; Frohlich & Parker, 2000; Magallón ^ appeared virtually simultaneously (Darwin, 1903; & Sanderson, 2001; Pryer et al., 2001; Crepet et al., Crane et al., 1995: Magallón et al., 1999; Crepet, 2004; Davies et al., 2004; Soltis & Soltis, 2004; 2000; Anderson et al., 2005; Davis et al., 2005: Friis Anderson et al., 2005). et al., 2006)? Even when the sequence of phylogenetic These new and exciting applications require events is addressed at a finer scale, such as within increased rigor and specificity from our reading of angiosperm subgroups (Saxifragales, Malpighiales). the fossil plant record. Era-level dates, formation- the density of fossil data and molecular analyses level stratigraphic placements, and nonquantitative available point to the rapid evolution of currently descriptions and reconstructions are no longer diverse and pus lineages in the Albian and 2001; Davis et al., paleontological literature are no longer likely to — 2005). With these inus well-constrained studies, sufficient. Similarly, superficial readings of the ^ Cenomanian (Fishbein et a adequately test evolutionary hypotheses. Original confidence intervals on individual branching events provenance of the plants in a community, as well as — can range from seven to 28 million years. In some the current and past location of the geological crust on cases, fossil evidence applied to dating molecularly which they are deposited, have become much more — based phylogenies rests on a single occurrence of a ' L acknowledge helpful critique of the manusc ript and ideas from Shanan E. Peters; literature searches by Susu Yuan and Marko Melymuka were instrumental to data collection; Bonnie Miljour rendered the final drafts of the figures; financial support from the National Science Foundado ADVA NCE Project Crosby Funds is gratefully acknowledged. ? Mu useum of Binh and Department of Ecology and Evolutionary Biology. University of Michigan. Ann Arbor. Michigan 48109-10 S.A. rburnham@umich.edu. doi: 10. 3417/2007002 ANN. Missouni Bor. Garp. 95: 51-71. PubLisHeD on 11 APRIL 2008. Annals of the Mun. Botanical Garden single plant organ from few sedimentary a ag Wikstróm et al., 2001; 004; 2006). While we should take seriously the verified presence of : (e.g.. Bremer et al., Crayn et al., taxon as excellent data, regardless of its uniqueness, neither the absence of a lineage nor the gap between occurrences is typically so carefully weighed. This biased approach to the fossil record is analogous to a hypothetical example of a molecular biologist extracting DNA from a 1 X 1 em piece of tissue from a single mature leaf from a single forest in Thailand and characterizing the phylogenetic position of an entire family of angiosperms based on a sequence of 850 bp from that leaf. More data are potentially available, but for logistical reasons they just are not collected, processed, and applied. Treatment of the obvious gaps in the fossil record can take two routes. Either gaps are ignore xd, reporting only what we know now, or gaps are acknowledged and the weakness of the record highlighted as an area for future research. The second route does not need to declare defeat but, rather, to state the uncertainty associated with the lack of sequential, fossil occurrences. A parallel situation occurs with the consistent enthusiastic reports of increased crop yield from radically elevated CO» concentrations based on small enclosures. The conclusion is that increased CO» may be a positive result for humans. Simultaneously. however, open-grown crop plants in elevated CO» sites (Free Air CO, Enrichment [FACE] marginally higher than in control sites (Long et al., 2006). the field experiments fills a gap that otherwise would provide are only Knowledge of data from open for misplaced optimism under increased CO» for crop yields (Tubiello et al., 2007). In a similar fashion, this contribution is intended to focus attention on the gaps the of investigalions that will improve our understanding of in fossil record as a means encouraging the limits of those gaps. An outline of the events relevant to fossilization of plants is first presented, starting with the evolution of a new species and ending with the use of that fossil species in a phylogenetic or morpho-developmental context (Fig. 1). Each step involves some modification of the original abundanc e, morphology, and likelihood of the the potential to distort subsequent ERT The uncertainty introduced by each step in this chain of recovery of plant species. Each has events should theoretically be quantifiable and, thus, provide us with an estimate of potential error. Two of these steps are highlighted (bold print, Fig. 1) in the of their importance in interpretation of the quality of the remainder this contribution, relative to reported fossil record. Steps in the fossilization of fossil plants have been addressed particularly carefully by Spicer (1981), Scheihing and Pfefferkorn (1984), Ferguson (1985), Gastaldo et al. (1987), Gastaldo (1988, 1994), Burnham (1989, 1993). Greenwood (1991), Burnham et al. (1992), and van der Burgh (1994). In each of these studies, the steps in the fossilization process are thus — outlined for slightly different reasons, and slightly different interpretations emerge. A frequent casualty of the fossilization process is the qualitative and quantitative correspondence between the source forest and the plant debris incorporated into the fossil record. As toward thicker and more abundant leaves most often fossilization proceeds, selective bias esults. even if these are not the most common in the source forests. In addition, small-statured species (herbs, shrubs, relative to large-statured species like trees (Scheihing, 1980; 1997). distinctive and well-preserved morphological features lianas) are reduced in abundance, Burnham, Furthermore, species with are most readily identified by paleobotanists, and species with a preference for moist, soggy, lacustrine, or riparian habitats are favored during fossilization over those with high abundance in dry or mesic habitats (Greb et al, 2006). These are important aspects of interpreting the fossil record, and prior studies have suggested the number of. specimens needed for accurate interpretations, the environments most likely to preserve an accurate record of the paleovegetation, and the effect of differential degra- dation of plant species and plant organs. This set of steps, from the first ecological and evolutionary appearance of a plant group to its use by humans in describing evolutionary history, is a critical pathway in interpretation and fidelity of the fossil record. Each factor can vary with geographic region, climatic regime, type and continuity of sedimentation, plant habit and abundance, and distinctiveness of the species. Any blanket statements about the character of the fossil plant record and its fidelity are clearly distillations of the variability inherent in the fossil if the fossil plant record is to be record. In general, used most reliably, records should be based on large- yodied, abundant, perennial wetland species, whose — reproduction is frequent and includes both sexual and asexual reproductive organs. Even so, we should expect large gaps in the fossil record due to periodic loss of habitat, lack of sedimentation, and erosional activily. Two fossilization steps are addressed here (Fig. 1) to clarify the constraints and advantages of applying the plant. fossil record to evolutionary history. First addressed is abundance increase (increase in popu- lation density and extent). which asks the question: £ how long does it take from simple presence as : species to abundance of the species with biomass large enough to have a high likelihood of represen- Volume 95, Number 1 Burnham 53 Presence in the Fossil Record Evolutionary event Population establishment ON ABUNDANCE INCREASE Abundance near depositional settings = ED Incorporation into sediments (some parts) Lithification subsidence [ELA es EE n RIO RI ‘igure l. Steps original plant form, ue ation size anc [em tation in the fossil record? Then, several important steps are bypassed to address outerop exposure and exploration, both of which are important in controlling the quality of interpretations made from the plant fossil record. This second step poses the R Uplift, recycling, Long-term sediment preservation erosion OUTCROP EXPOSURE AND RESTORATION Accurate identification of | what remains — Accurate dating of sediments Accurate | phylogenetic placement = wolved in fossilization of a newly evolved plant species. Each step involves some modification of the extent, abundance, or relationship to associated organisms, question: are appropriate fossil-bearing deposits well represented and sufficiently studied in areas proposed as cradles of evolutionary events? A final note is added that opens the door on issues surrounding long- term sediment preservation. 54 Annals of the Missouri Botanical Garden Srer One: POPULATION EXPANSION AND MIGRATION. How ONG Does Ir Take FOR POPULATIONS TO EXPAND TO Fossil. RECORD? CONTRIBUTE TO THE Following the evolutionary establishment of a new plant species, it has been proposed that there is a substantial lag m before species have the potential to contribute to the fossil record, either because new species tend to be geographically restricted, or because they tend to be relatively rare in comparison to their longer-established relatives. This proposition, often linked to the idea ghost lineages and added to (justifiable) suspected lack of taxonomic identification, is frequently cited in the plant evolution literature. This leads to the conclusion that the lag time between the evolution of the first angiosperm and its appearance in of the fossil record can create an apparent explosion of angiosperm forms at its first appearance (Frohlich & Parker, 2000; Qiu et al., 2000). Interestingly, a gap in the record is rarely cited radical innovations of grasses, where a more stately progression for the similarly is generally easily described (Jacobs et al.. 9; acobs, 2004; Stromberg, 2004; 2005). Although the lag time is rarely explicitly stated, for the explanation to fit the fossil record, the ecologically controlled lag times would need to be on the order of Prasad et al., — one lo five million vears to cause much discernable diff y is of critical importance in using fossil plants to date families, erence in analysis. Currently, this potential lag time the appearance of higher-order clades (e.g.. orders) and calibration of phylogenies based on molecular sequence data. Although this argument has a great deal of potential validity, il using modern plant data from analogous evolutionary and it has never been seriously tested ecological scenarios. Here, four ecological scenarios are examined in which expansion to large population sizes is used as a test of the lag time explanation for apparent gaps the The argument proposed is that if we evaluate settings under which can use the time population sizes as a measure of the minimum in fossil record. plants colonize and expand, we required for colonization and expansion of large ag time expected in the geological record. It is explicitly acknowledged that data are presented from success ( capability for expansion, dispersal, and population | typical for the stem group of many plant groups). j stories (1.e., species with ongevily that may not be data presented are the minimum time newly Therefore, required for colonization and expansion of a evolved species. On the other hand, it could be argued that it is just those species with strong expansion capabilities that will ultimately form the stem group of a successful plant group. Either way. it is prudent to view these time frames as minimum estimates. SCENARIO I: POST-VOLCANIC COLONIZATION Volcanic environments present a setting in which clearly defined we can follow succession using a starting date. Numerous volcanic eruptions over the past 1000 years are sufficiently studied to provide descriptive data on vegetative colonization. Á massive eruption in Papua New Guinea (Long Island) is dated at 1645 and is estimated to be about the magnitude of the island volcanic eruption on well-known Krakatau. Some 350 years later, 273 species of seed plants, including 31 species of the genus Ficus L. alone, plus 32 pteridophytes are present (Harrison et al., 2001; Shanahan et al. 2001). Motmot, a caldera island established as a habitable island in 1968, had established Ficus seedlings within only three years of volcanic activity cessation (Shanahan et al., 2001). this level of recolonization is species-poor compared to the rich flora of nearby Papua New Guinea. The slow colonization has been attributed to the dry volcanic soils relative to the richer flora of Krakatau, which has a similar island setting. and shorter eruptive history. In comparison, the massive eruption in 1883 u destroyed vegetation (and habitations) on Krakatau left ehind fertile ash soils in a humid tropical environ- ment, both of which promoted rapid and diverse colonization. Only 100 years after the eruption, over 400 species had colonized (Whittaker et al., 1997), including some 100 species of climbing plants. Whit- aker and colleagues propose that the island’s flora is still disharmonic but “becoming less so” just 104 years following the devastation of the original vegetation (Whittaker et al., 1989; Bush et al., 1995; Thornton et al., 1996; Whittaker et al., 1997: 1671). Volcanic environments can be rich in nutrients but relatively slow to accept colonizing species because the mantle of tephra and ash blankets the relatively more nutrient-rich debris flows and pre-volcanic soils. Even 50, lal ^ jon Because of high dispersal capabilities, ruderal species are generally the quickest to colonize (del Moral & Bliss, 1993). On Mount Lassen in California, forest trees may not have begun colonization until as much as 30 years after the cessation of eruptive activities (Kroh et al., 2000). Although colonization in volcanic environments can be slow, growth rates can accelerate once successful establishment has taken place, with many tree species reaching maturity within 20 years. On Mount St. U.S.A.), interactions between dispersal and colonizing ability prevent any concrete Helens (Washington State, site amelioration, ability, predictions on reassembly of communities (del Moral t al, 2005). lo approach equivalence across all habitats surveyed by » However, species richness started 1 ^ 20 vears following the major eruption of 1980. In Volume 95, Number 1 2008 urnham Presence in the Fossil Record contrast, percentage cover was most rapidly increas- ing in safe sites and strongly ameliorated sites. Recovery of volcanic environments is clearly a slow process and is highly subject to the species attributes and environmental changes of the specific area. Vegetation recovery on Mount St. Helens 22 years after eruption indicates that stochastic processes are still in effect; deterministic processes have not yet taken over (del Moral & Lacher, 2005). With relatively sparse data on long-term events in volcanic environments, we can still propose that in the absence of repeated events, even 2000 years should be sufficient for species in plant communities to expand to fossilizable population sizes. In the Neogene geological record of the western United States, it is clear that frequent disturbances can reset Ne the ecological setting on long-term succession, as shown by Taggart and Cross (1990) for the Succor Creek and Trapper Creek floras. SCENARIO 2: INVASIVE SPECIES INTRODUCTIONS Another approach to estimating the rate at which new species can become abundant enough to contribute to the fossil record is to determine the rate at which introduced species spread across a new landscape. Knowing the date of the first documented arrival of a species (based on herbarium specimens, first cultivation dates, or vegetation surveys) provides a proxy for a new species origination. Literature on introduced and invasive species is reviewed here with the following question in mind: How long does it take to become a pest (common enough to become a fossil)? Published literature is rich with reports of the impact and importance of invasive species, so this review is not intended to be exhaustive, but rather a sampling of life forms and phylogenetic diversity covered by the literature. In addition, summary ideas on invasive plants are extracted, relevant to the arguments here. Although attributes of species can help predict their propensity to be invasive, it is generally agreed that predictions are statistically based and individual cases are normally unpredict- able (Kowarik, 1995; Rejmánek & Richardson, 1996; Williamson & Fitter, 1996; Kolar & Lodge, 2001; Hallet, 2006). A temporal framework for this discus- sion was provided by Usher (1988), who remarked that 1000 years, it is distinguish exotic from native plant species. after virtually impossible to Based on intentionally imported species, a rule has been proposed that one of 10 species is likely to become naturalized once imported. Of those, one of 10 is likely to become a pest (Kowarik, 1995; Williamson & Fitter, 1996). tant because crop plants are very likely to become Exceptions are impor- naturalized (71 of 75 surveyed; Williamson & Fitter, 1996) and of 235 species introduced for horticulture, 82% have been found naturalizing outside of cul- tivation (Reichard, 1997). Still, the important step to pest status is quite low for crop species, yielding a one in 100 chance of rising to pest status for most = imported crop species. Interestingly, this value is similar for Hawaiian plants (90 in 900 of naturalized invasive plants are threats to ecosystems; Loope, 1998), which have gained a great deal of attention because of their island setting. An abbreviated list of invasive plants for which dates of introduction and approximate dates of pest status are readily available is presented in Table 1. This compilation does not include crop plants or plants that are only invasive in island ecosystems. Sampled are a range of life forms (trees, herbs, lianas, shrubs, aquatic plants) and a range of phylogenetic Pami ie conifers, dicotyledonous angiosperms, level, introduced species rapidly rise to pest status. Át a Henry and Scott (1981) reported that 29% of the plant species in Illinois were community-floristic level, documented aliens, and 13% of all woody species were aliens as of their study date. These figures are in line with figures reported from other preserved areas of the United States (Pimental et al., 2000: 26% in Great Smoky Mountain National Park; Yost, 1991: 36% alien to the New York State flora). Table 1 suggests that if a species has a propensity to become abundant, it can rise to abundance in less than 500 years in almost all cases studied. There is a surprising amount of consensus on what combination of attributes most predispose introduced plant species t by a that “The rate of immigration of plants and arthropods via to become invasive. Despite the statemen — National Academy Committee (Committee, 2002 natural [italics mine] means appears so low as to cause little concern. about the arrival of detrimental species,” it is likely that on a scale of 1000 years, this natural immigration would be of note. It is this longer scale we are concerned with here. Similarly, while plant traits are important, the complementary view that the disassociation of plants from their coevolved parasites and herbivores is also an important aspect of current plant invasiveness (Lake & Leishman, 2004; Hallett, 2006). Newly evolved species may have few parasites early in their evolutionary history, which would only add velocity to the rate of expansion. I emphasize here that the ecological literature clearly treats lag times as real and they certainly are so on an ecological time scale, but the time lags mentioned in the ecological literature are almost always less than 2000 years (Sakai et al., 2001; Crooks, 2005). In a particularly Table 1. Examples of selected invasive species for which the following can be determined: place of origin, place of invasive record, time to recognized pest status, and literature source. Genus and species Family Place of origin Place of invasion (first appearance date) Time to pest status (years) Source(s) leer platanoides L. Schinus terebinthifolius Raddi Vincetoxicum rossicum Kleop.) Barbar Ambrosia icm LL. Solidago altissima L. Solidago canadensis L. Berberis thunbergii DC. Butomus umbellatus L. Lonicera japonica Thunb. Celastrus orbiculatus Thunb. — jn Bryonia alba Sphaeropteris cooperi (F. Muell.) R. M. Tryon [syn. Cyathea cooperi (F. Muell.) Domin.] Elaeagnus umbellata Thunb. ex Murray Sapium sebiferum (L.) Roxb. Aceraceae Anacardiaceae Apocynaceae Asteraceae Asteraceae Asteraceae Berberidaceae Butomaceae Caprifoliaceae Celastraceae Cucurbitaceae Cyatheaceae Eleagnaceae Euphorbiaceae continental Europe and the C 4zducasus South America (Brazil, Argentina, Paraguay) U ane and North. America North America central and southern apan mainland Europe, U.K.. Ireland. "mperate western Asia Asia southeastern Ásia central and eastern Europe Austra Japan. China. rea Kor subtropical China MA. NJ. DE (introduced ca. 1775) Central America, Bahamas, Caribbean islands, FL, Mediterranean Europe, North and South Africa. Southeast Asia. Australia. (1832 or 1642-1849) northeastern and midwestern Ámerica (1890) France (1865-1880: possibly 16817) Europe (17507) China (1935) to Shanghai J to NH. west ND and SD: introduced to Arnold Arboretum (ca. 1870) northeastern North America (1897) eastern U.S.A. (ca. 1860) eastern U.S.A. (1860) HI (1950s) IL (and throughout the northeastern U.S.A.) GA (1772) China, South and Pacific islands < 200 70 100 10 ca. 200 Webb & Kaunzinger, 1993 Morton, 1978; Ferriter, 1997; Williams et al., 2005 Sheeley € Raynal, 1996 no et al., 2005: Pederson et al., 2005 Weber. uu. 2001 Weber. 2001; Dong et al., 2006a, b Vm 1997, 1999 Tutin et al.. 1980: Brown € Eckert. 2005: Kliber & Eckert. 2005 Schierenbeck, 2004: Yates et al.. 2004: Yurkonis & Meiners, 2004: McNab & Loftis, 2002: Ellsworth et al., 2004: Leicht & Si 2006 Novak & Mack, 1995, 2000 Medeiros et al., 1992: Durand & Goldstein, 2001 ander, Ebinger, 1983: Yates et al.. 2004 Bruce et al., 1997: Barrilleaux & Grace. 2000: Siemann & Rogers, 2001 USPIBE) [eoiuejog unossi|N 9g au) Jo sjeuuy Table 1. Continued. Genus and species Family Place of origin Place of invasion (first appearance date) Time to pest status (vears) Source(s) Pueraria lobata Willd.) Ohwi Lygodium microphyllum (Cav.) R Syzygium jambos (L Alston Psidium guajava L. Pinus radiata D. Don Spartina alterniflora Loisel. Polygonum perfoliatum Rhamnus cathartica L. Paederia foetida L. Salvinia molesta D. S. itch 5e y die kraussiana (Kuntze) A. Bra Atlanthus altissima (Mill.) Swingle Fabaceae Lvgodiaceae Myrtaceae Myrtaceae Pinaceae Poaceae Polygonaceae Rhamnaceae Rubiaceae Salviniaceae Selaginellaceae Simaroubaceae Asia tropical Southeast Asia, Africa, Australia Southeast Asia tropical America North America, 4000 ha. on CA coast Gulf of Mexico, east coast of North America East Asia: Japan, Philippines Europe Japan southeast Brazil South Africa, Azores eastern and central China Philadelphia, FL, southeastern U.S.A. 876) southern FL (1958) Puerto Rico (1821) Fiji, Namibia, Galapagos (1870) Australia, South Africa, and New Zealand (1665), but large-scale planting in 19: 930 WA (Willapa Bay) (ca. 1905) OR then PA (1890) 1824) North America X FL, NC, TX (ca. 1867) n Natl. Park, Australia; SC (1995); X (1998); LA to NC (2000) Neotropics, southern U.S.A., Europe, New Zealand esp., HI, Norfolk Island, Waitakere (7) (1919, New Zealand) U.S.A. (1784); Ontario, Canada (1998) ¡om 120 ca. 100 pa < 80 ca. 100 130 max.. possibly alv 60 Miller € Edwards, 1982; Miller, 1985: er h, 2000; m et al : Lembke, Nauman A d. FE A & Ferriter, 1998; Dougherty, 2003; Lott et al., 2003: Volin et al., 2004 Brown et al., 2006 Partridge, 1979 Cronk & Fuller. 2001 Stiller & Denton, 1995; Davis, 2005 i — 9; Oliver, 1996; Kumar, en 1985: Heneghan et al.. 2004. Brown, 1992; Diamond, 1999; Flores, 2003 Cowie & Werner, 1993; Jacono & Pitman, 2001; MeFarland et al., 2004. Wilson, 1996; Roy et al., 2004 Meloche & Murphy, 2006 8005 | JequinN ‘G6 euinjoA p1098H |ISSO4 au} ui eouesaJg weyuing Annals of the Missouri Botanical Garden detailed review of lag times, Kowarik (1995) reported a time lag of about 147 years between introduction and invasive status among 182 species studied. Characteristics commonly cited as beneficial to plants establishing expanding populations across a new habitat or continent, and thus potentially leaving an immediate fossil. record, have been addressed frequently in the literature (Baker, 1974: Scott & Panetta, 1993; Pyšek et aL, 1995; Rejmánek & Richardson, 1996; Reichard, 1997; Reichard & Hamilton, 1997; Clark et al., 2001; Kolar € Lodge, 2001; Sakai et al., 2001; Lake & Leishman, 2004; Richardson € Rejmánek, 2004; Rejmánek et al., 2005). From these, we can list attributes most likely to be found in plant species with the best potential for an immediate fossil record following evolutionary ap- pearance as follows: vegetative reproduction, includ- ing forming thickets with high leaf area: large number — of propagules produced and high consistency in production; small seeds and/or propagules with strong adaptations for vertebrate dispersal: short juvenile period; family or genus includes other invasives; broad ecological tolerance (via phenotypic plasticity and/or broad latitudinal range). Species of plants most likely to bear these characteristics. are — ossil the best candidates for use in reconstructions and paleoclimatie and phylogenetic studies. SCENARIO 3: SUCCESSIONAL FOREST AND REGENERATION RATES A third strategy for estimating the rate of expansion oJ of new species is to examine records of regenerating forests in temperate and tropical landscapes. Forests that have been clear-cut for ranching, lumbering, or mining, or devastated by landslides or hurricane winds are ideal settings for estimating rates of regeneration of populations to abundances that can contribute to the fossil record. This scenario is treated here separately from that of volcanic settings largely because of a more rapid rate of regeneration due to plant root and stem stocks and the smaller geographic extent of destruction expected. The two scenarios overlap to some degree. Again, the literature is rich in case studies of forest regeneration. Only a few of the many studies from temperate, subtropical, and tropical climates are selected for estimation of regeneration rates. This summary is restricted also to the regeneration of woody species, which are the most likely modern analogs to contributors to the fossil record (Ferguson, 1985; Burnham, 1997). In Neotropical lowland sites (either maritime-in — lu- enced or wholly continental), the time of regeneration for heavily disturbed or damaged forests is on the order of 100—400 years (Guariguata & Ostertag, 2001) and even en disturbance is short or less intense (Table 2). — less w Several authors point to the importance of assessing forest recovery in terms of species composition, basal area, and forest function, rather than simply species richness (e.g., Guariguata & Ostertag, 2001). Accurate data on forest regeneration as much as 200 years after disturbance are very rare, so estimates are often made either by comparison among different-aged | forest disturbances or by extrapolation from single forests from much shorter time intervals. A direct assessment of forest regeneration could increase the average time of recovery well beyond 400 years, but it is unlikely that it would extend it as much as one order of magnitude t« 4000 years. Significantly, repeated and continual jm [! shown by Rico-Gray and Garcia-Franko (1992) in forests in Yucatán, Mexico. A review of the successional isturbance suppresses total forest regeneration as literature for the Neotropics published in 2001 was characteristically noncommittal on any maximum rate of forest regeneration but acknowledged that “the regen- erative power of neotropical forests vegetation is clearly high" (Guariguata & Ostertag, 2001: 185). In temperate forests, data on human clearing are more readily available and can be used as the forest destruction date. In these forests, similar results seem to apply: regeneration is on the order of 100—500 years (Table 2). Regeneration time is potentially. longer under continued anthropogenic disturbance. In a review of temperate. regeneration. under. large-scale (1998) indicate that distance for dispersal and random processes (lack of predictable competitive hierarchies) are likely to apply disturbances, Turner et al. p but, again, over time scales that are not explicitly stated. Most authors report studies from relatively short time intervals, so their reluctance to commit to full regeneration is understandable. The combination of approaches mentioned below under Scenario 4 (migra- tion rates) adds a dimension improving interpretations of these shorter-scale results. This brief summary of the literature on regeneration in forested ecosystems in a variety of climates indicates that, again, plant species can and do expand to large populations in remarkably short periods of time, geologically. It would be surprising to detect any gap in the fossil record longer than 1000 years due to a single destructive event. Models of forest succession based on observations in the temperate zone generally employ time periods substantially less than that Busing, 1995). o e.g.. SCENARIO 4: MIGRATION—EXPANSION RATES BASED ON FOSSIL POLLEN LINKED TO HISTORICAL RECORDS The fourth approach taken here is to apply evidence from pollen in sediments to estimate rates of plant Table 2. Sele ected studies indicating woody species” composition time to recovery in tropical and temperate forests (including one bog). ordered north to south. Site Climate Recovery time (yr.): woody species Type of disturbance Source(s) Sweden s a ars Canoe Area S.A. a song Area CT), U.S.A. Fernow po Vang Forest (WV), Great bns e (TN), U.S.A Great Smoky Mountains (NC/TN) Yucatán. Mexico Dominican Republic uerto Rico Luquillo, Puerto Rico Sarapiqui, Costa Rica La Selva, Costa Rica Tropical summary: Nigeria to Thailand to Venezuela Barro "REM Island, Panama Colombia, Venezuela, Upper Río Negro Gran Pajonal, Peru temperate ombrotrophic peat mire temperate moist forest temperate moist forest temperate mountainous rest — temperate moist forest temperate lowland dry tropical tropical maritime tropical maritime tropical maritime tropical lowland wet forest tropical lowland moist forest lowland secondary forests tropical lowland tropical continental montane tropical rainforest > 50 60 yr. for demographic transitions > 50 Structurally similar but unrestored after 100 vr ca. 200 50-200 much more than 26 (forest not recovered at publication) 0 40 (similar to > 80-year-old forest) ca. 150: dominants regenerated each large sizes < n (longest tree lifespan) average 80 (75—150) 50-100 ca. 80-200 35 yr. to ca. 300 metric tons; uncut control ca. 600 metric tons peat harvesting 50 yr. before abandonment pan reduced fire regimes; yr. of direct observation pasani logging, followed 20 vr. later by Castanea blight abandoned agricultural fields small natural gaps agave cultivation 60 yr. of agriculture abandoned pastures human clearing and hurricane isturbance 5-10 yr. of pasture before abandonment undisturbed forest human disturbance natural river meanders 2-3 yr. of intensive agriculture on poor soil light pasture, 0-3 yr. of crop cultivation Soro et al., 1999 Frelich & Reich, 1995 Fike & Niering, 1999 Schuler & Gillespie, 2000 lebsch & Busing, 1989 Runkle, 1985 Mizrahi et al., 1997 Martin et al., 2004 Aide et al., 1996, 2000 Foster et al., Guariguata et al., 1997 Lieberman & Lieberman, 1987 Brown & Lugo, 1990 Knight, 1975; Dewalt et al., 2000; Denslow & Guzman. 2000 Saldarriaga et al., 1988 Scott, 1977 8003 | JequinN ‘G6 euinjoA DJ098H |ISSOY SY] ui eouesaJd weyuing 60 Annals of the Missouri Botanical Garden species expansion from nascent o lo popu- lations large enough to be recovered in the fossil record. Pollen deposited lake, bond, estuary, or riverine sediments documents the presence of species in the catchment area. First appearances of plant invaders are used as a proxy for species evolution and employed in combination with historical exploration and land use records to give a minimum estimate for definite representation in the fossil record (at least in wetland areas). Similar to other approaches used here. pollen evidence is a best-case scenario resulting in the most rapid rates. The evidence presented overestimates he appearance in the fossil record of slow-moving taxa - or taxa whose pollen is poorly represented in sediments. In addition, pollen is likely to be the first plant organ potentially recognized in sediments but is species (but see recent 2006). Pollen evidence overestimates the rate at which an average species can only rarely diagnostic t advances in Finkelstein et al., spread and be recognized in the fossil record. Crawford example of pollen records including Native American ake (southern Ontario) is a well-known and European colonizers (McAndrews, 1988; Jackson, 1997). Records indicate that with the expansion of Native Americans using small-scale cultivation meth- ods, species that are known to favor disturbed habitats, such as Portulaca oleracea L. and Zea mays L., appeared and expanded. Evidence from a variety of sites (Easter Island; Lindsley Pond, Connecticut) was reviewed by Jackson (1997) and the obvious point is that the appearance of pollen of an invasive or introduced species in sediments is often interpreted as synchronous with the arrival of colonizing humans. This interpretation, while certainly not absolutely accurate, is widespread (Whitehead & Sheehan, 1985 — and underscores the speed with which introduced species spread on arrival. Jackson (1997) summarized the rate of spatial expansion being on the order of tens to thousands of kilometers within 5000 years for entire p (Endl.) Carriere into several Pinus L. ant communities. Synchronous invasions by Tsuga forests in North 1998), evidence of invasions that appear to have taken place ess than 1000 years. Woods and Davis (1989) document range expansions over periods of 500-2000 America are documented (Davis et al., — + over years for Fagus L. Quite appropriately, data for Fagus and Acer L. 2005). showing that rates may not have been as rapid genetic polymorphisms (McLachlan et al., some have envisioned. While the molecular estimates double the previously accepted migration rates of forest trees, even this doubling of migration rates suggests no major temporal gap between the appear- ance of new species and their recognition in the fossil record for the majority of the fossil record. were revisited using the distribution. of These four approaches to estimating the rate of species expansion following establishment support the — idea that it is unlikely that gaps in the fossil record can be explained by an intrinsic lag in population establishment and expansion. Plant species appear to be able to both migrate to almost any area inhabitable by the species increase in population size and on the earth. Data are based on rapidly migrating. expandable species, but there is no intrinsic. geolog- ical lag due to lack of colonizing or dispersal abilities. Time lags of as much as 2000 years may be expected for some species prior to establishment of large populations, contributing to the fossil record. Howev- er, this is not a long period of time relative to the acuity of the fossil record and is particularly short relative to the phylogenetic and morpho-developmen- likely, expansion 1s — questions applied correctly. It is not that the responsible for observed ta therefore, rate of species gaps in the fossil record itself or disharmonie results between molecularly based phylogenies and the sequence of appearance of taxa in the fossil record. Instead, observed or perceived gaps must be explained by other mecha- nisms, of which there are many. Srep Two: OUTCROP EXPOSURE AND EXPLORATION. CAN THE PorkNTIAL Fossi. RECORD PROVIDE ANSWERS? The sequence of events leading from species origination to recognition and application of plant fossils includes a step that receives little attention: the potential for discovery of fossil-bearing outerops and recognition of taxa represented (Fig. 1). This aspect of our understanding of earth's history has received remarkably little direct attention in the past, giving us little hard data with which to estimate the potential distributions. n stratigraphic ages or global error Because my interests have focused on fossil plants from northern South America (Burnham & Graham. 1999; Burnham & Johnson, relatively 2004). I cover this topic in some detail here for young sediments (Cretaceous to Pliocene) in northern South America. The topic deserves much more attention and could be profitably covered continent by continent by special- ists. Two questions are addressed using data derived from Cretaceous- through Pliocene-age sediments from literature and accumulated outcrop data: (1) Is the quality of the published fossil record sufficient for he timing of evolutionary events in constraining northern South America? (2) If we knew every plant ace 'essible fossil deposited in every oulcrop in northern South America, would that be sufficient for dating phylogenetic trees, mapping past ecological or geographic distributions, or documenting a complete sequence of evolutionary events? Volume 95, Number 1 8 Burnham 61 Presence in the Fossil Record I 50°W Late Cretaceous £^ 67 Figure 2. Known late Early and Late Cretaceous-age plant fossil localities in northern South America as of December 2006. Data for localities were extracted from the literature, with each site numbered in a database available from the author. Localities included from Aptian to Maastrichtian. THE QUALITY OF THE PUBLISHED PLANT Fossi. RECORD In 1999, Burnham and Graham published a series of five maps of northern South America, plotting all fossil plant localities known to us from the Late Cretaceous, Paleogene, Miocene, Pliocene, and Qua- ternary. Collectively, the data appeared rich (Burn- 1999: fig. 2), but the 243 fossil localities were scattered over more than 80 million years and over 13,000,000 km*; localities per time unit or per unit area was really ham & Graham, thus, the number of quite small, even compared to our limited under- standing of the distribution of modern biodiversity. Moreover, those localities (28%) represented Quaternary fraction (roughly 2%) of the 80 million years of more than a quarter of sediments, a tiny sediment surveyed. In 2006, a resurvey of the paleobotanical and geological literature, published Internet sites, and more complete bibliographic searches turned up a much larger number of sites for each time period evaluated, in addition to seven years of added publication activity (see Table 3). the Late Cretaceous, Paleogene (three subintervals included), Updated maps are published here for Miocene, and Pliocene (Figs. 2-5). Quaternary de- posits were not resurveyed in 2006. A total of 332 pre- Quaternary localities are now known (compared to 174 pre-Quaternary localities known in 1999). A compar- ison of these data sets is presented here as Table 3. Significant in both data sets and from inspection of both series of maps (Figs. 2-5) is the paucity of known plant localities from the Amazon Basin for more than 80 million years (Late Cretaceous through Pliocene: 1999; Figs. 2-5). The rocks of the Amazon Basin are complex, and, yet, a == see also Burnham € Graham. large proportion is mapped either as igneous and metamorphic rocks of Precambrian age, or as Pliocene- and Quaternary-age sediments blanketing an unknown base (SIG, 2002). Many of the known fossil plant localities are sediment cores, which typically can only provide palynological data. How- ever, in parts of the Brazilian Amazon, rocks of Cenozoic or Cretaceous age are mapped but poorly studied. These areas represent the sites of greatest changes in density of localities between the 1999 and 2006 surveys of localities, but the large extent of rock pd that could potentially contribute to the plant fossi record has been underexplored. Improvements on the Annals of the Missouri Botanical Garden Table 3. Comparison of locality numbers per geological period for two literature census dates. Because literature was combed more deeply for the Cretaceous localities in the 2006 survey, the two Cretaceous values differ not only because of the addition of Aptian and Albian localities in the 2006 census, but also because of intensified publication activity and better searching techniques (search engines, online libraries, and software) in 2006. my — million years. Geological period Approximate duration (my) localities presented in 1999 Localities presented im 2006 (Quaternary 1.8 69 Not censused Pliocene 3.5 35 38 Miocene 18 15 116 Oligocene 10 18 37 Eocene 22 27 44 Paleocene 10 6 23 35 (1999) 13 7 Late Cretaceous 55 (2006) Total localities Cenomanian to Maastrichtian 174 74. Aptian to Maastrichtian 332 E 243: including Quaternary known fossil locality maps Burnham and Graham published in 1999 give us hope that further explora- tions, with this more complete set of known localities. will help close the gaps where they yawn the widest. The number of fossil data points that contribute to dating recent land plant phylogenies based on molecular evidence is remarkably small and yet is drawn from global sources (Fishbein et al., 2001: Conti et al., 2002; Bremer et al., 2004; Davis et al., 2005; Crayn et al., 2006). South America has the potential to contribute substantially, as demonstrated by the spectacular fossils of early angiosperms and gnelaleans recovered from the Crato Formation in Brazil (Mohr & Eklund, 2003; Rydin et al., 2003) and the diverse angiosperms reported from the Paleocene and Focene from Colombia (Jaramillo, 2002; Jaramillo et al., 2006) as well as from the Miocene of Ecuador (Burnham, 1995; Burnham & Carranco, 2004; Burn- ham, unpublished). However, much of the Late Cretaceous and Cenozoic sedimentary rock that is known to exist in South America has not been adequately explored. to determine whether it. bears fossil plants, much less sustained detailed taxonomic studies. For example, a large expanse of Cretaceous sediment is evident from maps of the adjoining states of Goias, Tocantins, and Bahia, Brazil, but | have not vet found a single reference to fossil plants from this area. Similarly, the geologic map of Bolivia bears indications of appropriately aged sediments, and, yet, this country consistently shows few localities on each of the 2006 maps presented here. These specific examples underscore the importance of gaps in our knowledge of the potentially available fossil record, in contrast to gaps that may be created by lack ` preservation, Or deposition, lack olf fossiliferous sediments prior to discovery. Some gaps of erosion of might be filled by data that are still held as proprietary by commercial concerns (oil, gas, and other mining exploration). — The impact of gaps on applications of the fossil record to paleogeography, paleoclimate, phylogenetics, and even ecosystem change is extreme. For example, due to the limited subset of fossil calibration points for phylogenetic trees, the same data can be used to constrain timing for alternative scenarios of plant evo- lution. Drastically different scenarios can be equally plausible if calibration points are few, or are biased toward specific regions of the world. Similarly, data on paleoclimatic parameters based on fossil plants are limited by known localities that can be analyzed for such data. A single fossil assemblage from a small area can be assumed to be representative of a large geo- eraphic area (as is generally assumed), but it can also be atypical, representing a low moist area in a surrounding nondepositional, dry area (Burnham et al., 2001; Kowalski & Dilcher, 2003). Interpolation of paleoclimatic points between sites that are more close- e ly spaced is obviously preferred. Lack of intensive collection and documentation of fossils in certain geo- graphic areas will severely bias paleogeographic re- constructions of species distributions. In contrast to the promising results from the first step of the fossilization process presented here (Abundance Increase: Fig. 1). in some parts of the world and for some time periods, the fossil record should be read very cautiously. Presence of a taxon in the fossil record is a known anchor for evolutionary events, but such data cannot be used to weight particular areas or time periods if sufficient geological and paleobotanical exploration has not been carried out elsewhere (see, for example, the distribution of localities in fig. 2 of Gingerich, 2006). A recent Volume 95, Number 1 2008 Burnham 63 Presence in the Fossil Record 2 2501918 17 x I | qa P41 42.10 A 50?W O Oligocene 28 6,8, ‘29, 30, 35,37 2 E i o 5 O Paleocene 2 : 19,36 Vv: 11/3) 15 14 12 5 34 16 15,34 9 20 D 39 3755 4 8)5 5 29 9 1 2 2132) 0? e 3. Known Paleogene-age plant fossil localities in northern South America as of Dec :ember 2006. Data for localities were extracted from the literature, with each site numbered in a database available from the author. Localities are coded to show age assignment to Paleocene, Eocene, or Oligocene extensive review of the Malvoid pollen record (Krutzsch, 2004) proposed evolutionary and migration scenarios based on worldwide pollen records. While terms of published the review is complete in information, we have to wonder how much of the scenarios might be altered with just two additional records of Malvoid pollen from South America. The absence of evidence in unexplored areas da to be recognized as just that. CAN THE PorkNTIAL FossiL RECORD ANSWER Our QUESTIONS? The second question posed above is potentially unanswerable because of our lack of understanding of pa global depositional patterns over time: if we knew the identity and morphology of every fossil deposited in every accessible outerop, would that be sufficient for confirming the validity of dates for phylogenetic trees or for confirming evolutionary events? While we can more intensively sample areas of the earth that are poorly represented by fossils of particular taxonomic groups and ages (addressed in the preceding paragraphs), a more substantial issue lurks within the deposition and preservation of the sedimentary rock record itself. Ronov and colleagues 1979, 1980; Ronov, 1994) painstakingly measured the area of rock outcrop for epochs over the (Ronov et al., past 543 million years. These monumental efforts were hampered at the time by the same kinds of gaps in our knowledge of surficial geology in remote areas that we the work 1980) that within each geological have today. Nevertheless, demonstrated (Ronov et al., period from the Lower Permian to the present, the proportion of rock volume composed of coal-bearing rocks (the most likely to preserve plant fossils) is between 0.3% and 8.1% (a high of 11.2% is recorded in the Carboniferous period). If we add all other sedimentary rocks that could potentially incorporate plant remains (but probably do not, such as coarser river sediments) to those coal-bearing sediments, we get a maximum of 16%-30% of the rocks preserved from any geological period, which have the potential to bear identifiable plant remains. Of course this is highly variable, temporally. This result is sobering because this may mean that mapped geological exposures of any specified time period have, on average, less than a 30% chance of 64 Annals of the Missouri Botanical Garden " m i : 50?W 38, 41] 0411116. 103, 107, 108, 409, 115 : NS E d 63 113,114, Oe Miocene "El AGE aie 6 234039, 110, 11 y 0 uS C oa Ma oE q q a7. (20) 7 194853) REE os 0 S909. | 82 (8 @5 0*— 8) 83 Q9 58 @2) O | _@) Figure 4. Known Miocene-age plant fossil localities in northern South America as of December 2006. Data for localities were extracted from the literature, with each site numbered in a database available from the author. yielding rocks that potentially bear plant remains. What is of even more concern is a concept recently cogently explained by Peters (2005) that the potential for sedimentary deposition and preservation is strongly controlled by continental position and, in particular, sea-level change. Peters (2005, 2006) employed the COSUNA (Correlation of Stratigraphic Units of North America) charts, which are based on some of the most well-known terrestrial and marine sedimentary sequences in the world. He asked the question: Is the temporal continuity of sedimentary rock over time correlated with either first or lasi appearances of taxa (in his case, genera), thus indicating that the two are not mechanistically independent? He showed that, in fact, almost every aspect of temporal variation in the fossil record is reflected in similar varialion in the spatial and temporal distribution of sedimentary rocks. In his work, Peters (2006) also demonstrated that the total volume of terrestrial sedimentary units over North America increased dramatically over the last 90 | the million years. This increase in all sediments i recent geological past is overprinted on extreme variability in the proportion of sedimentary packages constituted by fine-grained sediments (ideal for fossil plant preservation) over time. The presence of multiple marginal depositional basins is correlated with a high potential for sedimentary accumulation. An occurs when plate tectonic configurations are such that the of sedimentary basins. One implication of this result is increase in sediment accumulation logically they accommodate formation large that we should expect to find that many plant fossils appear synchronously in the fossil record, not for evolutionary reasons, but as a result of tectonic forcing of the sedimentary record. Peters’ data would suggest that interpreting first occurrences as true originations is likely to be erroneous, given that the preservation potential is so much higher when tectonic plate configurations foster the accumulation of fine-grained sediments. One specific geographic location where this might have had a strong effect is in the eastern portion of Brazil, where well-preserved plant remains uniquely represent early angiosperms and gnetaleans. However, this was just the kind of tectonic: setting where deposition would be predicted. So were angiosperms and gnetaleans living there preferential- ly, or are they simply preserved preferentially there? Volume 95, Number 1 2008 Burnham 65 Presence in the Fossil Record 50°W ; | Pliocene | a @) Figure 5. Known Pliocene-age plant fossil localities in northern South America as of December 2006. Data for localities were extracted from the literature, with each site numbered in a database available from the author. While this evidence might be far from the minds of those involved in phylogenetic analysis and the By application of fossil plant data to constraining evolutionary scenarios, it is critical to recognize that the fossil record is both preserved and subsequently sampled in a highly nonrandom manner. Overriding considerations are the completeness of the record of any one time period, the depositional circumstances under which the original data were preserved, and the erosional circumstances of the intervening time period between deposition and exhumation. In conclusion, this contribution addresses two important steps in fossilization and recovery that apply to a wide variety of paleobiological and evolutionary questions. Four scenarios are provided under which it can be seen that plant populations grow rapidly from a very small number of individuals to. populations large enough to contribute substan- tively to the fossil record in a geologically short period of time (ca. 2000 years). This period of time may not be resolvable in the rock record and, as such, can be thought of as instantaneous. Migration rates of plants are generally rapid enough that there is no discern- able lag time that would be produced in a fossil record if a successful plant or plant group is to be anticipated to migrate over large land areas. These data apply only to the rock record prior to the Quaternary, in which resolution is not expected on these ecological scales. Most of the modern analogies I have used to address this question are based on successful immigrants or invasives, and some are expanding under circum- stances of lowered competition and predation. There- fore, the quantitative data here are recognized as representing the short end of the time-to-expansion spectrum. It is not possible to give estimates of expansion rates for a given fossil species if reproduc- tive intervals, time to reproduction, or even ability to vegetatively reproduce are unknown for that species. Instead, it is reasonable to assume that a plant group that ultimately does contribute substantively to plant diversity, density, or vegetational importance (defined as many species within the group, biomes dominated by members of the group, continent-wide coverage by members of the group) has done well at some time in its evolutionary history, and the time scale of — expansion of 10-1000 years is applicable. In contrast, the recovery of a high-quality fossil record depends on a number of factors that are not as 66 Annals of the Missouri Botanical Garden easily dismissed. We should be especially cautious in assuming the adequacy of the currently known and potentially retrievable fossil record in answering detailed questions about plant evolution. While the rock record is incomplete, the analysis does not need to stop there. If the fossil record can be quantified both globally and regionally relative to the probability of bearing and preserving fossil plants, we would be able to assign confidence limits on the presences and absences that are applied from the rock record. Approaches like this have been carried out on pure sedimentary gap analysis and for paleobiodiversity analyses (see papers by Sadler & Strauss, 1989, 1990; Holland, 2000; Crampton et al., 2006) and hold great promise for the geographic and stratigraphic absences of fossil plants. Literature Cited Aide, T. M., J. K. Zimmerman, M. Rosario & H. Marcano. 1996. Forest recovery in abandoned cattle pastures along an elevational Nn nt in northeastern Puerto Rico. Biotropica 28: : Rivera & H. Marcano- nosequence of ——, ————, E Pascarella, L. Vega. 2000. Forest tropical abandoned E s: une alions for restoration d Ecol. 8(4): 32 Bremer regeneration in a chron ecology. Ander & E. M. Friis. 2005. Dating od basal eudicots using rbcL sequences and reference points. Amer. J. Bot. 92 = multiple fossil 1737-1748 CIE evolution of weeds. Annual Rev. B oe e, 2000. Growth and invasive within the Barrilleaux ere of 2m sebiferum ara i coastal prai egion: The effects of soil and moisture regime inet Bot. 87: 1099-1106. Breme E. Friis € B. Bremer. 2004. Molecular phylogenetic dating of asterid flowering wq s early Cretaceous diversification. veia 153: 496 Brown, J. S. & C. G. Eckert. 2 aa ees increase in sexual us clonal 2s capacily during Edd al invasion in an aquatic D EE umbellatus (Buto- er. es Bot. 495—502. 'atena A J. Gurevitch. 2006. Effects of an invasive tree on community structure and diversity in a tropical forest in do Rico. Forest Ecol. Managem. 226: 5-152. Brown, L. E. 1992. Cayratia japonica (Vitaceae) and Paederta de (Rubiaceae) adventive in Texas. Phyto- logia 72 Brown, S. & A E pus 1990. Tropical secondary forests. —32. Int. J. pe Ecol. € : Bruce, K. A., G. N. Cameron, P. A. Harcombe & G. Jubinsky. 1997. O impact on native habitats, and management of a woody invader, the Chinese Tallow Tree, Sapium sebiferum (L .) Roxb. Nat. Areas J. 17: 255-260. Burnham, J. 1989. a between standing vegetation and leaf litter in a paratropical forest: n tions for n Rev. Palae rm p jl. 58: 5-32. 3. Reconstructing richness in the du. fossil 90; Palaios 8: 376-384. — A new species of winged fruit from the Miocene ES E js ecuatoriana (Leguminosae). . J. Bot. 82: 1599-1607 Stand ae and leaf litter species ió of a dry forest out in mon Rosa National Park, Fa Rica. pr 29: 384-395. — ——— & A. Graham. 199 e nes of Neotropical an tation: Ta evelopment and status. Ann. Missouri . Gar d. i 546— — & N. L. RAD 2004. Miocene winged fruits Ec Pda — Loxopterygium (Anacardiaceae) from the Andes py J. Bot. 91: R. Johnson. 2004. South American paleobotany and de origins of € NI noe Philos. Trans., p 1595-1610. A. Pitman, K. R. Johnson € P. Wilf. 2001. Habitat- ee error in estimating temperatures from leaf t. 88: Amer. J. — WB margins in a humid tropical forest. 1096-1102 . Wing & G. G. Parker iunities in leaf litter 5. . 1992. The reflection of deciduous forest comm : Implications for autoc ‘hthonous litter assemblages from the fossil record. 34—53. | Whittaker & T. Partomihardjo. 1995. onization m succession on Krakatau: Án analysis of the guild of ud plants. Biotropica 21:355 Busing, R. 1995. Disturbance and the population dynamics jm can tulipifera: pes with a spatial model of forest succession. J. Ec 5-53. Clark, . Lewis & 4 2001. extremes: mm spread with variation in dispersal 537-554. Paleo c 18 ush, M. Co w —372 ¡om pos Invasion by and EE Amer. Naturalist |: o C. C & R. T. Busing. 1989. Secondary succession, gap dynamics, and community structure in a _southe rn Appalachian cove forest. Ecology 70: 728— Committee on the Scientific Basis for Predicting the Invasive Potential of p e Plants and Plant Pests in the United Stat 002. P eh invasions of nonindige- nous P P plant pests. National Academy Press, Washington, D.C. , ac November 2007. 2 T. Eriksson, J. Schonenberger, K. J. Sytsma & D. . 2002. Early Tertiary out-of-India dispersal of nce from phylogeny and molecular 42. Conti, ed A. Bau one aceae: Evide dating. Evolution 56: 1931— i D. € P. Cowie, I. . Werner. 1993. Alien plant species vasive in Kakadu National Park, tropical northern Mal Biol. Conservation 63: 127-135. Crampton, J. S., A. G. Beu, R. A. Cooper, C. M. Jones, B. Marshall € P. A. Maxwell. 2006. Estimating the rock volume bias in paleobiodiversity studies. Science 301: Friis & K. R. Pedersen. 1995, The origin Nature 374: 358-300. nane, P. R., E. M and early ae ‘ation of angiosperms. 27-33 1. Rossetto € D. J. Maynard. 2006. Molecular ro-Miocene radiation Crayn, D. M prelatus Md dating reveals an Olig of dry-adapted shrubs (former Tre ee from rain- forest Iree progenitors (Elaeocarpaceae) in Australia. Amer. J. Bot. 93: 1328-13 Crepet, W. L. 2000. Progress in understanding angiosperm history, success, and relationships: Darwin's abominably “perplexing phenomenon." Proc. Natl. Acad. 97: 12939-12941. Burnham Volume 95, Number 1 67 2008 Presence in the Fossil Record K. C. Nixon € M. A. Gandolfo. 2004. Fossil eden and phylogeny: The age of major angiosperm clades based on mesofossil and macrofossil evidence from Cretaceous e Amer. J. Bot. 91: 1666-1682. Cronk, Q. C & J. L. Fuller. 2001. Plant Invaders: The Threat to Natural Po ms. Earthscan, Sterling, Virginia Crooks, J. A. 2005. Lag times and exotic species: The ecology and management of biological invasions in slow- motion. Ecoscience 12: 316-329, Darwin, C. 1903. Pp. 20-22, 239-240 in F. Darwin & A. C. Seward (editors), More Letters of oe Darwin, Vol. 2. D. Appleton & Company, New Davies, T. J., T. G. Barraclough, M. y. Chase, P. S. Soltis, D. E. Soltis & V. Savolainen. 2004. Darwin’s abominable 1999. Structure and dynamics of populations of Tipies barberry (Berberis puntera eee in deciduous forests of New ad Biol. ips Ellsworth, J. A., R. A. Vies & J. pu 2004. dling emergence, growth, and ee of Oriental bittersweet: Effects of seed input, seed bank, and forest r litter. Forest Ecol. Managem. 190: 255-264 "il D. K. 1985. The origin of leaf-assemblages—New light on an old problem. Rev. Palaeobot. Palynol. 46: 117-188 c >» Ferriter, A. (editor). 1997. Brazilian Pepper Management Plan for Florida. Brazilian Pepper Task Force Florida Exotic Pest Plant Council. untley & T. " Pp. 673-697 in B. Vegetation MEE Handbook of Vegetation Scienc series. Kluw Pus McFarland, D. Ea | S. Nelson. M. J. Grodowitz, R. M. Smar in: Den 2004. S. Mite m (Giant Salvinia) in the United States: A review of species Salvinia molesta | ecology and approaches to management, U.S. Army Corp of E oe rs Engineer Research Development Center SR- 04-2 MeLac dis J.S.. J. S. Clark & P. S. Manos. 2005. Molecular indicators of tree migration capacity under rapid climate change. des 86: 2088-: 0t MeNab, W. & | occurrence 2002. habitat features for oriental bitters mÓ 2: olii Probability of i weel in an oak forest in the southern ee mountains, USA. Forest Ecol. Managem. 155: 45-54 Volume 95, Number 1 Burnham 69 2008 Presence in the Fossil Record Medeiros, a L. Loope, T. Flynn, S. J. Anderson, L. W. Pryer, A., H. Se a ider, A. R. Smith, R. Cranfill. P. G. Cuddihy e K. 4 Wilson. EDU Notes on the status of an Wolf, J. 5. aie: & 5. D. SM 2001. Ho ned and ferns invasive Australian. tree fern (Cyathea cooperi) in are a monophyletic gip Hawaiian rain Ue Amer. Fern J. 82: 27-33. Meloche, C. € S. Murphy. 2006. Managing tree-of-heaven (o Ailanthus oliin. in parks and prote al Pa e [ies 13: 264-772. Miller, J. H. 1985. Testing herbicides for kudzu eradication on a Piedmont site. S. J. Appl. 128-132. Forest. € — lo K ected areas: A case study o ndeau Provinci (Ontario. Canada). edw war ds. udz Zu: Where did it come J from? And how can we stop it? S. J. Appl. Forest. 6: 16: Mitich, I: W. | is [Pueraria lobata (Willd.) Ohwi]. Weed Technol. 31-235. Mizrahi, A., J. pube & J. Jimenez-Osorio. 1997. Composition, structure, and management of secondary dry tropical forest vegetation in two abandoned henequen plantations in. Yucatan, Mexico. Forest Ecol. Managem. o: E —282. Mohr, B. A. R. € H. Eklund. 2003. Mii is a a ee angiosperm from the Lower Cretaceous Crato Formation (Brazil). Rev. Palaeobot. Palynol. 126: 279-292 Morton, J. F. peop 1978. e. animals Brazilian pepper—lts impact | on and the environment. Econ. . Austin. Nauman, C. E. 978. Spread of the exotic fern Lygodium ons m in Florida. Amer. Fern J. 68 65—00. Novak, S. J. & R. N. Mack. 1995. Allozyme diversity in the apomictic t vine Bryonia alba (Cucurbitaceae): Potential aoe nces of multiple introductions. Amer. J. Bot. 82: -116 02 . 2000. Clonal diversity within and among introduced populations of the d vine ome alba (Cucurbitaceae). Canad. J. Bot. 469— Oliver, J. 1996. M onum ile-a-minute weed | (Polyg: ca L sl an invasive vine in natural and disturbed a 6 4—2: sites. Castanea 61: 244-251. s rt, R. A. . L. Hamric & L. A. Donovan. 2000. Genetic variation in Pueraria lobata (Fabaceae), an introduced clonal invasive plant of southeastern United States. Amer. J. Bot. 87: 1240-12 Partridge, I. a 9. The guava a in Fiji. S. Pacific Bull. 29(2): 28 . Peteet, D. Kurdyla € T. Guilderson. 005. Me d a Ps arming, Little Ice Age, impact on the environment during the last ra in Pederson, a and European the lower Hudson Valley, New York, USA. Qua 93: 238-249. Pemberton, R. W. € A. P. Ferriter. 1998. Old World climbing fern (L n M EE a dangerous invasive weed in Florida. . Fern J. 88: 165-175. Peters, S. E. 2005 tionary history of marine animals. U.S.A. 102: 12326-12331 2006. ~ ‘rostratigraphy of North America. J. Geol. 114: 391 Pimental, Ge ne constraints on n an Proc. Natl. Lach, R. Zuniga S D. Morrison. 2000. and economie costs of U nited States. 5 A. E. Strömberg, H. Alimohammadian & A. Sahni. 2005. Dios coprolites and the early evolution of 310: 1178-1180. Environme n nonindigenous Specs in the BioSeic ence 50: 53— Prasad, Brasses and grazers. Science s the closest living relatives to seed | Nature 409: 618-622. Pyšek, P. Prach & P. Bs 1995. Relating invasion success to plant traits: An analysis of the Czech alien flora. Sek, K. Prach, M. y sow & M. Wi 2 lant siesta A pects slants. and Special Amsterdam. Qiu, Y La J. Le ur Quadroni, D. E. Soltis, P. S > . A. Zimmer, Z. Chen, V. Savolainen & ase. oy Phylogeny of esal Analyses of five genes from three genomes. Int. ĉi. 161: 3531 S21. Reed, F. 1979. Maryland Reichard, S tions on vata and local levels. Pp. 215-227 in J. Luken & J. A. Thieret (editors), Assessment and Manage- ment of Pa Invasions Springer-Verlag, New York. & 1997. Predicting invasions of woody an oe ‘ed into North America. Conservation Biol. 93—203. A 7 & D. M. Richardson. make some 1655-1601. a ademic, angiosperms: J. PI. S nd ae (L.) Greene in a 43: 219-22 997. Prevention is invasive plant introdue- 215 A. milt lon. 1996. What attributes plant species more invasive? Ecology 77: ———, — ——Á , S. I. Higgins, M. J. Pitcairn & E. O 2005. Boi: of invasive plants: State of the art. Pp. 161 in H. A. Mooney N Mc PA Waage editar. Invasive A New Synthesis. Island Press, Washing- - lack, J. A. = 1, Ric ae D. M. & M. Rejmánek. 2004. Conifers as invasive aliens: A global survey and predictive framework. Diversity a oo 10: 321-331. Rico-Gray, V. € J. Garcia-Franko. 1992, Vegetation and soil sed a of successional stag dec :iduous forest. J. Veg. Sei Robertson, D. J.. M. €. 1994. Colonization dynamics of four exotic plants in a nort Piedmont 107-11 E Ronov, A. B. 1994. sions on d continents: À quantitative approach based on es in tropical lowland 1 7-624. Robertson € T. Tague. ern natural area. Bull. Torrey Bot. Club 121: w ` Phanerozoic transgressions and re gres- areas flooded by the sea and areas of marine and ey dus deposition. Amer. J. Sci. 294: 777-801 Balukhoy Khain & A. N. a associations of the continents. Int. 21: 415-446. 'sky. 1979. Paleogene Geol. Rev ————, —, ——— & eslavinsky. 1980. Ovantitaio analysis of phiano a Sed- imentary Geol. 311—325. Rothwell, C. W. 1999. dus and ferns i in n the resolution of land plant phyogs ny. Bot. Roy, B., I. Popay, P. uo T. James & A. 2004. An ib usd Guide to the Common Weeds of Zealand. New Zealand Plant Protection Society, New Zealand. Runkle, J. R. 1985. temperate M 17-33 in S. A. Pickett € P. S. lite , The Uol of Natural Disturbance and Patch co Co Rahman. Lips Disturbance regimes in T. ress, Orlando. ohr & E. M. . 2003. Cratonia cotyledon gen. et sp. nov.: À unique Cretaceous seedling relate i a Welwitschia. Philos: Se 270(suppl.): 529 Friis Trans., Ser ~ B. 70 Annals of the Missouri Botanical Garden Sadler, P. M. & D. J. Strauss. 1989. Classical confidence intervals and Bayesian probability for ends of local taxon ranges. Math. G 1: 411-427. E Estimation. of completeness of stratigraphical Es p empirical data and theoret- J. Geo ol. in 147: 471-485 Tharp & C. Uhl. 1988. est succession in the ical models. Saldarriaga, J. G., Long-term a aoe a d Negro of Colombia and Venezuela. J. Ecol. 76: 938-9 Sakai, i < F. W. E J. S. Holt, D. M. Lodge, J. Molofsky, K. A. With, S. Baughman, R. J. Cabin, J. E. Cohen, N. C. Ellstrand, p. E. McCauley, P. O'Neil, L M. Parker, J. N. Thompson & S. C. Weller. 2001. The population biology of invasive species. Ann. Rev. Ecol. Syst. 32: 305-3: Scheihing, M. H. 1980. Reduction of wind velocity by the forest canopy and the rarity of non-arborescent plants in a Upper Carboniferous fossil record. Argum. Palaeobot. : 133-138. & H. W. Pfefferkorn. 1984. The taphonomy of plants in the Orinoco delta: A model for the incorporation of plant parts in clastic sediments of Late Carboniferous Age of Euramerica. Rev. echó Palynol. 41: 205-240. Schierenbeck, K. A. japonica) as an invasive species; context. Crit. Rev. Pl. Sci. 23: 391 n T.M. & A.R. Gillespie. 2000. Tanpoi patterns of 2004. Japanese honeysuckle (Lonicera pistons ecology, and ody cies diversity in a central Appalachian forest from 156 to 1997, J. Torrey Bot Soc. 127: 149-161 cott, G. 1977. The importance of old-field succession VUE increments to shifting cultivation. Great Plains d hu eed Geograph. J. 6: 318-327. Scott, J. K. € F. D. Panetta. 1993, Predicting the Australian weed status of southern. African. plants. J. Biogeogr. 20: 87-93. Shanahan, M., R. D. Harrison, R. Yamuna, W. Boen & I. W. B. a 2001. Colonization of an island voleano, Long Island, Papua New Guinea, and Tdi in its caldera lake. V. Colonization by figs (Ficus spp.), 1 an emergent island, thei ajr r dispersers and pollinators. J. Biogeogr. 28: F D. d na 1906. The distribution a tus 'eloxicum in eastern ioken: Buil. Tomes E "Club 123: 148-156 Siemann, E. s. 2001. Genetic d in wth of an invasive Pd species. Ecol. I 514—518 sic (Sistema de Informações Geográficas). 200. dos Geo- AD do Brasil (1:2.500.000) AS Tec- dou ursos aes Hp Brasil. 13/31 /start.htm?t I / |= homes. ac- cessed. l November 2007. D. Soltis, P. S. & E. Soltis. 2004. The a and diversification of angiosperms. Amer. J. Bot. 91: 1614— 1626. Soro, A., S. Sundberg & H. Rydin. 1999, Species diversity, niche metrics and species associations in harvested and undisturbed us J. Veg. Sci. 10: 549-560. Spicer : 1981. The sorting and dlog hnos plant material in a modern environment at Silwood Lake, Silwood Park, Berkshire, England. United States Cel Al Survey e Paper 1143 Stiller, J. W. & A. L. Denton. Spartina s mb] (Poaceae) in » illapa B ton: Random amplified polymorphic DNA analysis of an 3. )ne hundred vears of ay, Washing- invasive population. Molec. Ecol. 4: 355-36: deposition of Stromberg, C. A. E. 2004. Using phytolith assemblages to reconstruct the origin and spreac grass-dominated habitats in the great plains of North America during the — ate Eocene to early Miocene. Palaeogeogr. Palaeoclima- tol. Palaeoecol. 207: 239-275. ie R. E. & A. T. Cross. 1990. Plant successions and interruptions in Miocene volcanic deposi Pacific Northwest. Pp. 57-68 in M. G. Lockley (e ditor, pos ‘anism and Fossil Biotas. Ge lid Sociely of America Special Papers 244. Thornton, L BaS The role of animals in the OEN of ~ ; Compton & C. N. Wilson. 1996. ` the Krakatua Islands by fig trees (Ficus specie Biogeogr. 23: 577-592 Tubiello, F. N., J. S. Amthor, K. J. Boote. M. Donatelli, mi G. Fise ther, R. M. Gifford, M. Howden, J. RD & evated CO, "Food for a de T 2 S z o gs No ES Ss S Crop response to e and world food supply, A comment on Thought...” by Long et al, Science 918-1921, 2006. Eur. J. E 26: 215-223 Turner, M. . L. Baker, C. J. Peterson & R. K. Peet. 1998. Pach tors us ncing succession: Lessons from large, infrequent natural disturbances. Ecosystems 1: 511—523. Tutin, T. G., V. H. Heywood, N. A. Burgess, D. M. Moore, D. H. Valentine, S. M. Walters & D. A. Webb. 1980. Flora Europea. Cambridge Univ. Press, Cambridge Usher, M. B. 1988. Biological ias A search for generalizations. Biol. Conserv. 44: 119-135. van der Burgh, J. 1994. wood-, and leaf-floras, táphonomy a us LUN impli- cations. Rev. Palaeobot. Palynol. 8 )-129. Volin, J. C.. M. S. Lot, J. D. n : m Owen. 2004. Predicting rapid invasion of the Florida Everglades by Old Differences in fossil seed/fruit-, World C m E ern A ygodium microphyllum). Diversity & Distrib. 10: 4 446 Webb, S. L. & C.K. x r. 1993. $E 'al Invasion of (New Jersey) Fore Norway rU. ). Bull. Torey Bot. Club 120; 343-349, the Drew University reserve by Weber, E. 1998. The dynamics of plant invasions: A case study of three exotic goldenrod species (Solidago L.) in uar Ai Ta ogr. 25: 147-154. . Current and potential ranges of three esoo gole Tea (Solidago) in Europe. Conservation Biol. 122-128. Whitehead, D. R. & M. C. 1985. O vegetational changes in the Tombigbee River eee Eastern eee Ud Midl. Naturalist 113: ir ] Whittaker, R. J., M. B. Bush € K. Richards. 1989. ie iru es pee vegetation succession on the Krakatau 23 Sheehan. Islands, Indonesia. Ecol. Monogr. 59: 59— & T. 1997. The rebuilding, of an lated rain forest assemblage: How he flor Krakatau? Biodivers. & Partomihardjo. Chase. 2001. E volutión of the angiosperms Calibrating the family tree. Londe on, Ser. B, Biol. Sci. 268: 2211-2220. A. e holt, J. P. Cuda & C. R. Hughes. 2005. C hloroplast and microsatellite DNA diversities al the in eee verebinthifolius) 3043-3050. Savolainen Proc. Roy. Soc. troduction history of Brazilian peppertree in Florida. Molec. Ecol. 14: Williamson, M. « A. Fitter. 1996. The varying suecess of waders. ene 77: 1661-1666 . 1996. Alien ferns in awe Pacific Sei. 50: Wilson K. / 127—141 Volume 95, Number 1 2008 Burnham Presence in the Fossil Record Woods, K. & M. B. Davis. 1989. Paleoecology of range | Yates, E. D F. Levia Jr. & C. L. Williams. 2004. Recruitment of deed non-native invasive plants into a fragmented forest in southern Illinois. Forest Ecol. Managem. 190: 119-130 imits: Beech in northern Michigan and Wisconsin. Ecology 70: 606. Yost, S. E. S. Antenen & G. Hartvigsen. 1991. The vegetation of the Wave Hill Natural Area, Bronx, New York. Bull. Torrey Bot. Club 118: 312-325. Yurkonis, K. A. & S. J. Meiners. 2004. Invasion impacts local species turnover in a successional system. Ecol. Lett. > 7: 764—169 W(H)ITHER FOSSILS? STUDYING © Elizabeth J. Hermsen? and Jonathan R. MORPHOLOGICAL CHARACTER Hendricks EVOLUTION IN THE AGE OF MOLECULAR SEQUENCES ABSTRACT A major challenge in the post-genomics era will be to integrate molecular sequence dala from extant organisms with morphological data from fossil and extant taxa into a single, coherent picture of ‘phylogenetic relationships: only then will these phylogenetic hypotheses be effectively applied to the study of morphological character evolution. At least two analytical approaches to solving this problem have been utilized: (1) simultaneous analysis of molecular sequence and morphological data with fossil taxa included as terminals in the analysis, and (2) the molecular scaffold ada in which morphological data are analyze d over a molecular backbone (with constraints that force extant taxa into positions s suggestel by sequence data). The perceived obstacles to including fossil taxa directly in simultaneous analyses of m aan al and molecular sequence data with extant taxa include: (1) that fossil taxa are missing the molecular sequence portion of the character data: 2) that morphological characters might be misleading due to convergence; and (3) e character weighting. specifically how and us ther to weight characters in the duh ‘al partition relative to characters in the molecular sequence data partition. The molecular scaffold has been put forward as a potential solution to at least some of these problems. Using examples of simultaneous analyses from the literature, as pun as new analyses of previously published morphological and molecular sequence data matrices for extant and fossil Chiroptera (bats), we argue that the : imultaneous analysis approach is superior to the molecular scaffold approach, specifically addressing the problems to which the molecular sc affold has been suggested as solution. Finally, the application of phylogenetic hypotheses including fossil taxa (whatever their derivation) to the study of iaracter evolution is discussed, with special emphasis on scenarios in which fossil taxa are likely to be most morphological e enlightening: (1) in determining the sequence of character evolution: (2) in determining the timing of character e volution; and in 3)1 in making inferences about the presence or absence of characteristies in fossil taxa thal may not be direc tly observable in the fossil record. Key words: Character mapping, Chiroptera, convergence, echolocation, fossil, homoplasy, molecular scaffold, molecular sequence data, morphological character evolution, phylogeny, simultaneous analysis. total evidence. At one time, extinct taxa represented by fossils features (le... synapomorphies) were considered im- (hereafter, fossil taxa) were considered central to portant in interpreting relationships. With the advent understanding the evolution of organisms through time of the framework explicated by Hennig (1900) and the (see, for instance, Eldredge & Cracraft, 1980; Smith, development of analytical methodologies and pro- 1998). Phylogenetic hypotheses were developed by a — grams. for tree-building (e.g. Farris, 1970: Fitch. qualified expert or experts on the basis of comparative — 1971: Felsenstein, 1981), paleontology became less anatomy and morphology, to which fossils were central to understanding evolution through geologic considered to contribute primitive and intermediate — time (despite the early recognition of the logic and forms through which one could trace evolution from — utility of the cladistic methodology by some paleon- ancestor to descendant to the most recent members of — tologists [e.g.. Schaeffer et al., 1972]) because extant a group. Characters considered meaningful to the organisms could be grouped on the basis of shared without refer- development of evolutionary scenarios were entirely at derived. traits—or. synapomorphies the discretion of the investigator, and overall — ence to the fossil record. In fact, fossil taxa, for which — similarity as well as the appearance of advanced many data were often missing and whose interpreta- We thank William Crepet for inviting us to participate in this symposium volume and thank Paulyn Cartwright, Mark Holder, Victoria C. Hollowell, Bruce Lieberman, Linda Trueb, d Gramarye, and members of the Department of Ecology and Evolutionary Biology systematics od group at the University of. Kansas for helpful comments and editorial assistance that improved the quality of this manuscript. Emma Teeling a provided us with her aligned molecular sequence data set for Chiroptera. JRH's ió to this research were supported by National Science Foundation EAR 0518976. ° honing of Ecology and Evolutionary Biology, Haworth Hall, University of Kansas, 1200 Sunnyside Avenue. Lawrence, Kansas 66045-7534, U.S.A. chermsen@ku.ec “De partment of Geology, University of Kansas, Lindle pe enm doi: 10.3417/2006206 — ey ^ Hall. 1475 Jayhawk Blvd, Lawrence. Kansas 66045-7613, U.S.A. ANN. Missouri Bor. Garp. 95: 72-100. PuBLISHED ON 11 Arri 2008. Volume 95, Number 1 2008 Hermsen & Hendricks Morphological Character Evolution tion was potentially difficult, became viewed by some as an impediment to understanding phylogenetic (e.g.. 1981). Molecular systematies, which provides relationships among extant taxa Patterson, arge numbers of sequence characters, has altered our understanding of the relationships among and within many groups, often without reference to fossil taxa. However, fossil — - — axa provide unique types of information not available in extant organisms, and, because of this, the recognition of fossil taxa as an important component of phylogenetic studies has recently experienced a renaissance (Smith, 1998). Some of this may be due to the temporal information that fossil taxa can provide about the rate and timing of group diversification, principally in the application of temporal data associated with the occurrences. of fossils as calibration points in studies of the rate of evolution of molecular sequence characters (e.g. a 2004; Schneider et al., 2004). Fossils are also unique repositories of data on extinct both representation in the extant biota. Thus, fossil taxa — Peterson et a morphologies for groups with and without can provide insight into the sequence of evolution within morphological characters that are correlated in extant taxa, as well as access to suites of characters or variations within characters that would be entirely lost if not for knowledge of the vast extinct flora and fauna that once flourished on earth (Donoghue et al., 1989; Smith, 1998; Forey & Fortey, 2001). One of the biggest c — iallenges for paleontologists and systematists alike in the post-genomics era will be to figure out how best to incorporate paleontological data (primarily anatomical and morphological data, vereafter. simply referred to as morphological data) with molecular sequence data from extant organisms to take advantage of these unique aspects of fossil taxa (e.g.. see Peterson et al., 2007). The issues involved in this subject are complex, ranging from character delimitation and interpretation, the effect of missing data on analyses, and whether to combine data sets and analyze fossil taxa directly with extant taxa referred to as simultaneous analysis (Nixon & Carpenter, 1996) or combined analysis, or the total evidence (Kluge, 1989) or supermatrix approach (see review in de Queiroz & Gatesy, 2007)—or to use more indirect methods, such as trees based on molecular referred to as "molecular scaffolds" (Springer et al., 2001: 6242). Herein, we compare several methods for combining backbone constraints, sometimes morphological and molecular sequence data for a group (Chiroptera, bats) that has attributes (e.g.. a arge body of molecular sequence data conflicting with traditional groupings based on morphology, several well-preserved fossil representatives) emblem- atic of the current problems confronting the integra- tion of extant with fossil taxa in which molecular sequence data are involved. We will use these analyses as examples in a review of major issues surrounding tree-building and the interpretation of character evolution in joint fossil-extant taxon anal- yses that include a molecular sequence and morpho- logical component. In the first part of this discussion, we argue that, because many of the issues surrounding character mapping on a phylogeny are not unique to analyses in which fossil taxa are included, the underlying problem in studying fossil taxa in a phylogenetie context is to identify the most effective way to integrate our knowledge of morphological characters with our evolving knowledge of the tree of relationships among extant organisms as suggested by molecular sequence data. In the second part, we discuss the utility and complications of mapping characters and studying character evolution in a context in which fossil taxa are included as terminals in a phylogenetic analysis, emphasizing examples from simultaneous analyses. BACKGROUND The Cenozoic record of the placental mammal clade Chiroptera (bats) provides a good data set to explore how paleontological, morphological, and molecular sequence data interact in phylogenetic analyses, and how, in turn, these data types can inform hypotheses of the sequence and timing of morphological character evolution. A plethora of morphological and molecular sequence data has been collected about bats. and several analyses have integrated data sets in order to explore patterns of character evolution and biogeog- raphy within both extant and fossil members of the Chiroptera (Springer et al., 2001; Teeling et al.. 2005). Traditionally, bats have been grouped by morpholog- ical data into two clades, Microchiroptera (microbats). with laryngeal echolocation (a biological form o — ` sonar to hunt prey), and Megachiroptera (Pteropodidae; flying foxes and Old World fruit bats), lacking laryngeal echolocation (Simmons, 2005a). Recently, analyses of molecular sequence data have challenged this traditional view of bat evolution (see reviews in Simmons, 2005a; Jones & Teeling, 2006); these data suggest that Pteropodidae (flying foxes and Old World fruit bats) and an echolocating microbat group called Rhinolophoidea (horseshoe bats) are more closely related to one another than either is to the remaining microbats. The clade including Pteropodidae and Rhinolophoidea is known as Yinpterochiroptera, while the clade including other echolocating bats has been referred to as Yangochiroptera (Springer et al., 2001). While it was once thought that laryngeal echoloca- tion—which has a complex morphological basis (Arita 74 Annals of the Missouri Botanical Garden & Fenton, 1997)—evolved only once in bats, the new view of bat phylogeny based on molecular sequence data raises the possibility that laryngeal echolocation either evolved twice independently, or evolved once, but was lost in Pteropodidae (Springer et al., 2001; Jones & Teeling, 2006) To date, direct simultaneous analyses (Nixon & Carpenter, 1996) of combined morphological and molecular sequence data sets from extant families across the order Chiroptera (with or without fossil taxa) are lacking, despite the potential that these combined data may have for further clarifying the 20052) the evolution of echolocation. Here, we explore whether relationships of bats (Simmons, and performing such an analysis will support the new molecular view of bat phylogeny, as predicted by Simmons (2005a: 167). Prior research in this area has been undertaken only indirectly: Springer et al. (2001) and Teeling et al. (2005) used a molecular scaffold approach (e.g., used backbone constraints) to place fossil taxa within a phylogenetic context. MATERIALS AND METHODS COMBINED DATA MATRIX — The combined morphological and molecular se- quence data set analyzed here was constructed from two previously published data matrices. The NEXUS data file representing the morphological matrix published by Gunnell and Simmons (2005) was downloaded from the American Museum of Natural History FTP site linked directly from Nancy e T "n Hl mons nomepa ge ( h.amnh personnel/simmons. gig) This wiorebé osten ve features a total of 35 terminal taxa, six of which are fossil taxa Archaeonycteris Revilliod, Hasstanye- teris Smith & Storch, /caronycteris Jepsen, Palaeochi- ropteryx Revilliod, Tanzanycteris Gunnell et al., and an undescribed genus from the Eocene Green River Formation of Wyoming (Gunnell & Simmons, 2005)— and five of which are extant outgroup taxa (Cyno- cephalus Boddaert, flying lemur; Erinaceus L., hedge- hog; Felis L., cat; Sus L., pig; and Tupaia Raffles, tree shrew). There are 204 morphological characters, 94 of which are soft tissue and 110 of which are osteological characters; 165 have nonadditive (unordered) transfor- mations and 39 have additive (ordered) transformations. All extant morphological matrix are extant familial or subfamilial ingroup terminals (n = 24) in this — bat taxa. Fossil taxa are scored at the genus level. The aligned molecular sequence data matrix, the basis for the study by Teeling et al. (2005), was kindly provided to the authors (specifically, JRH) by Dr. Teeling on June 9, 2006. The complete molecular 13,792 sequence characters representing "nuclear sequence sequence data matrix includes aligned data from portions of 17 nuclear genes" (Teeling et al., 2005: 581) from 30 bat genera (representing all families of Chiroptera; sequence data for some ingroup terminal genera are composites from multiple infrageneric species) and four outgroup terminals represented by composite sequence data gathered from two or more genera each. Details of this matrix, including GenBank accession numbers, were provided by Teeling et al. (2005; supplementary table S6). Three options for reconciling overlapping taxa between the two matrices presented themselves at the beginning of this study: (1) culling taxa from the molecular sequence data set, leaving one generic-level terminal to combine with the corresponding family or subfamily terminal in the morphological data set; (2) fusing terminals in the molecular sequence data set so that all molecular sequence variability for each family or subfamily was encompassed in one corresponding terminal in the morphological matrix; or (3) duplicating morphological terminals in order that each terminal represented by a molecular sequence data set also had a morphological data set, some of which would be identical for members of the same family or subfamily. Each option has potential pitfalls. The first would discard the most data, the second would result in an increase of polymorphisms in the molecular sequence data set, and the third would result in multiple terminals sharing the same set of morphological characters. For a more generalized discussion of the problem of terminal mismatch in combining data matrices, see Nixon and Carpenter (1996). We decided to use option three (duplication of morphological terminals), because at least some studies have suggested that phylogenetic accuracy increases with greater taxon sampling (e.g., Zwickl & Hillis, 2002), information: further, and we did not wish to discard we did not want to add to the ambiguity of the combined data set (which already includes many cells coded as missing) by fusing molecular sequence terminals. This option allowed us to keep all terminals represented in the molecular sequence data set except Perissodactyla (odd-toed hoofed mammals), for which we could find no reasonable combination with a terminal in the morphological data set (see further discussion below). The number of terminals represented by molecular data that morphological data in the combined data set range sequence share the same duplicated from zero (17 taxa, including three of the outgroups) to two (two groups of two taxa), three (one group of three taxa). or four (two groups of four taxa). For genera with identical morphological data sets, the morphological data obviously supply no information on intrafamilial Volume 95, Number 1 2008 Hermsen & Hendricks 75 Morphological Character Evolution relationships; in the simultaneous and molecular scaffold analyses, these are completely structured by the molecular sequence data. One potentially problematic aspect of the morpho- logical data set is that, where polymorphisms existed in the earlier Simmons and Geisler (1998) morpho- logical matrix, of which the Gunnell and Simmons (2005) matrix is a modification, Gunnell and Simmons (2005) replaced them either with an inferred ancestral state (IAS) for the family or subfamily (IAS coding) or with the most common state in the family or subfamily, or used ambiguity coding for that character (see Geisler, 2002 (1998) listed the number and percent of polymor- Simmons & Simmons and Geisler phisms within each terminal in the previous version of this matrix, thus giving some indication of how many cells may have been converted from polymorphic to single state for each terminal in Gunnell and Simmons (2005). This means that, in cases in which polymor- phisms may occur intrafamilially or intrasubfamilially, the inferred plesiomorphic state within the higher- level terminal may be substituted for a polymorphism (Simmons & Geisler, 2002), and this state may or may not occur in the genus matched with the higher-level terminal. Hence, the analyses could be improved in the future by coding the states present in individual genera, rather than in families or subfamilies. As a corollary, not all characters coded for each higher- level terminal may have been observed in each genus included in the molecular sequence matrix, so some extrapolation of character states within genera may be occurring (see Nixon & Carpenter [1996] concerning — m extrapolation). According to Simmons (2005c: 527 extant bats are classified into 18 families, another six families are known from fossils, and *biologists have long agreed that these groups represent distinct evolutionary lineages,” although “there has been no consensus concerning relationships among them." Because the monophyly of the families within Chiroptera is apparently not in question and is further supported by the molecular sequence data set employed here (Teeling et al., 2005), we do not anticipate error caused by incorrectly assigning some genera to the wrong family. According to Simmons and Geisler (1998), monophyly of all bat taxa (families and subfamilies) included as terminals in the analysis is well established, except perhaps for Vespertilioninae; only one genus, Rhogeessa H. Allen, is assigned to Vespertilioninae in this study. Combining these two matrices required renaming the terminal extant bat taxa in the morphological matrix (suprageneric) with the generic terminal names in the molecular sequence data matrix. The classifi- cation of Simmons (2005b) was used to match each extant bat genus with the suprageneric taxon to which it belongs .(see Table 1). composite terminal Felis/Panthera Oken in the mo- For the outgroups, the ecular sequence data set was matched to the morphological data set for Felis and named Felis; the composite terminal Condylura llliger/Talpa L./ Scalopus Desmarest (moles) was matched with Erina- ceus (hedgehog), and these were renamed Eulipo- typhla, a monophyletic clade composed of some former members of the Insectivora, including hedge- 2001). Finally, the composite molecular sequence terminal hogs, shrews, and moles (e.g., Murphy et al., Tragelaphus Blainville/Bos L. (bovines) were com- bined with the morphological data for Sus (pigs) to form the terminal Cetartiodactyla, a clade supported by molecular sequence data (e.g., Montgelard et al., 1997; Murphy et al., 2001; Boisserie et al., 2005) that also includes additional ruminants, whales, and hippopotami. The Perissodactyla outgroup, composed of sequence data from Equus L. Ceratotherium Gray (rhinoceroses), combined with a morphological terminal, as no perissodactyls are outgroups in the morphological matrix. Because it lacks morphological data, the Perissodactyla outgroup can then act as a “wild card” & Wheeler, 1992: 134; see discussion below), interacting with fossil bat taxa that group between the ixon oa outgroup taxa and extant ingroup bats, since Perisso- dactyla and the fossil bat taxa are coded for mutually exclusive data sets. Perissodactyla was thus removed from the combined matrix. Four extant bat subfamilies and two outgroup terminals represented in the Gunnell and Simmons (2005) matrix that are not represented at all in the Teeling et al. (2005) data set were allowed to remain in the combined matrix, since they were all coded for the morphological characters. These taxa were thus similar to (but more complete than) the fossil taxa. The total combined morpholog- ical and molecular sequence data set (hereafter referred to as the combined data set) included 45 terminal taxa (five outgroups) and 13,996 characters (39 additive). All characters were weighted equally. CLADISTIC ANALYSES Terminals were duplicated and renamed in Word- Pad (Microsoft Corporation, Redmond, Washington), matrix dimensions were modified (where necessary), and the files were opened in WinClada (Nixon, 1999— ) Matrices were combined in WinClada, with in Table 1. The combined data set was saved in .ss format before it terminals matched as detailed was opened in the software program TNT (Goloboff et al., 2003a), where it was resaved in TNT format. Tree searches were performed under the parsimony crite- ron using TNT (Goloboff et al., 2003a). For each 76 Annals of the Missouri Botanical Garden Table 1. Morphological (family and subfamily) data sel matched to each molecular sequence (genus) terminal. taxa follow Simmons (2005b). correspond to the GenBank molecular sequence accession information in supplementary table S6 of Teeling et al. (2005); the Assignments of genera to higher Abbreviations in the molecular sequence data column first letter corresponds to the genus name, the next two letters to th € molecular source species name, and the number in the bracket refers to composite terminal numbers table S6 of Teeling et al. (2005). Four extant terminal taxa possess morphological data (from Gunnell € Simmons, 2005) but lack corresponding molecular sequence data; these include Tomopeatinae Miller (Molossidae) and Miniopterinae Dobson, Murininae Mil pa er, and Kerivoulinae Miller (Vespertilionidae). Taxonomic level (family or sub Extant bat genus family) from which morphological data were coded! Molecular sequence data source” Antrozous H. Allen Antrozoidae Apa[19] Craseonycteris Hill Craseonycteridae Cih[30] Emballonura Temminck Emballonuridae Ea 11] Taphozous E. Geoffroy Emballonuridae Inu[12] Rhynchonycteris Peters Emballonuridae Hna[13] Furipterus Bonaparte Furipteridae Fho[26] Hipposideros Gray Hipposideridae Heo[6] Megaderma V. Geoffroy Megadermatidae Mly[7] Macroderma Miller Megadermatidae Mgi[8] Tadarida Wafinesque Molossinae Tbr[28] Eumops Miller Molossinae Fau[29] Pteronotus Gray Mormoopidae Ppa[23] Myotis Kaup Myotina Mda[21] Mystacina Gray Mystacinidae Mtu[25] Wyzopoda Milne-Edwards & A. Grandidier Myzopodidae Mau[22] Vatalus Gray Natalidad Nst[27] Voctilio | Noctilionidae Noctal[ 18] Nycteris G. Cuvier & E. Geoffroy Nycteridae v1 [9] Tonatia Gray Phyllostomidae Tsif 14] Irribeus Leach Phyllostomidae Aja] 15] Desmodus Wied-Neuwied Phyllostomidae Dro[ 16] Inoura Gray Phyllostomidae Agel[ 17] Pteropus E nde ben Pteropodidae Peif 1] Cynopterus F. Cuvier Pteropodidae Cbr[2] Rousettus Gray Pteropodidae Hla[3] Nyctimene Borkhausen Pteropodidae Nal[4] Rhinolophus Lacépede Rhinolophidae Rer[5] Rhinopoma Y. Geoffroy PRA Ag Hha[10] Thyroptera Spix Thy Pu[24] Rhogeessa H. Allen al Ktu[20] ' From Gunnell e Simmons, 2005. ? From Teeling et al., 2005. analysis, the collapsing rule (determining which nodes will be collapsed from dichotomous to polytomous in most parsimonious trees [MPTs]) was set to rule 3. which only collapses nodes with no character support 0). search was emploved using the following parameters: (max length — For each analysis, a heuristic starting Wagner trees were calculated using a random seed of 0; 1000 search replications were performed with tree bisection-reconnection (Swolford & Olsen, 1990). trees per search replication. branch-swapping saving up to 10 shortest Trees were collapsed with a after the search (in other words. branches maximum length of O were collapsed into polytomies rather than being ensemble 1969) were maximum consiste number displayed ncy index of steps calculated by WinClada as dichotomies). The (Cl; Kluge & Farris, and ensemble retention index (RI; Farris, 1989) calculated based on the total minimum and for each matrix as For analysis partition was analyzed la. the modified morphological data without constraints. This analysis was performed to confirm the results of the original Gunnell and Simmons (2005) analysis with a matrix including the cloned terminals. For analysis Ib. bat genera represented in the molecular sequence data matrix (Table 1) were constrained to conform Volume 95, Number 1 2008 Hermsen & Hendricks 77 Morphological Character Evolution with the Yinpterochiroptera- Yangochiroptera group- ings as shown in Teeling et al. (2005). Only two positive constraints (one for each clade) were used. Fossil taxa and supergeneric extant ingroup taxa were designated as floaters (unconstrained). This analysis was performed to determine how many additional steps would have to occur in the morphological data in order for them to support the major dichotomy in the Chiroptera suggested by the molecular sequence data set (Teeling et al., 2005). For analysis 2a, the modified Teeling et al. (2005) matrix was analyzed without constraints. For analysis 2b, the outgroup Perissodactyla was deactivated and (2005) analyzed. The purpose of this analysis was to insure that the Yinpterochiroptera and Yangochiroptera were still recovered as monophyletic clades with Perisso- the modified Teeling et al. data set was dactyla removed from the molecular sequence data For analysis 2c, the molecular sequence data matrix minus Perissodactyla was analyzed with taxa traditionally assigned to Megachiroptera and Micro- chiroptera constrained to belong to those groups (two positive constraints). The purpose of this analysis was to determine how many additional steps must occur in the molecular sequence data in order for them to support the major dichotomy in Chiroptera suggested by the morphological data (Gunnell & Simmons, 2005) and traditional classifications. For analysis 3a, the combined data matrix was analyzed with only those taxa for which both data partitions were coded (other taxa were deactivated). The purpose of this analysis was to determine whether taxa with large amounts of missing data were significantly affecting the results of the simultaneous analysis with fossil and extant taxa. For analysis 3b, the combined data set with all taxa exc ept Perisso- dactyla was analyzed. For analysis 3c, the combined data set was analyzed with Yinpterochirpotera— Yangochirptera constraints, as in analysis 1b For analysis 4, all nodes from the molecular tree topology found in analysis 2b were constrained, and only taxa lacking molecular sequence data were allowed to float. This a lates the methods of Teeling et al. (2005) and others (see below). nalysis emu CHARACTER MAPPING The binary (presence/absence) laryngeal echoloca- tion character was mapped onto the trees resulting from analysis 3b using Fitch optimization (Fitch, 1971). RESULTS Table 2 summarizes the results for each analysis, including number of MPTs, length of MPTs, and CI and RI. relevant (e.g.. under constraints). For analysis la (morphological Topological results are discussed below as when analyses were not performed data with no constraints), the strict consensus of all MPTs was concordant with the results shown in Gunnell and Simmons (2005: fig. 1). The of the MPTs for analysis 2a (the molecular sequence data set without constraints) has consensus two the same overall structure as that shown in Teeling et al. (2005: fig. 1), which illustrated the maximum likelihood (ML) tree calculated under the GTR + T +1 model of molecular sequence evolution. In both trees, Yinpterochiroptera and Yangochiroptera are mono- phyletic sister groups composed of the same taxa. The Pteropodidae (flying foxes and Old World fruit bats) and Rhinolophoidea (horseshoe bats) form monophy- letic clades sister to one another in Yinpterochir- optera, although the internal arrangement of Ptero- podidae is different in our parsimony and the ML trees of Teeling et al. (2005). Emballonuridae, Phyllosto- midae, Vespertilionidae, and Molossidae form mono- phyletic groups within Yangochiroptera (other fami- lies are represented by only one terminal), although the internal structure of Yangochiroptera is both different from and less resolved in the strict consensus of the parsimony trees found here as compared to the ML tree (see Teeling et al., 2005: fig. 1). The results of analysis 2b (the molecular sequence data set with no constraints, Perissodactyla deactivated) also sup- port the Yinpterochiroptera- Yangochiroptera. group- Support values—standard bootstrap (Felsen- stein, 1985), Horovitz, 1999b), and symmetrical resampling (Golob- off et al., 2003b) with traditional search—calculated for this pruned data set that decrease the support (as ex- ings. Poisson bootstrap (Farris et al. in suggest removing Perissodactyla does pressed as absolute frequency) for the Yinpterochir- optera clade, although the degree to which support was affected was dependent on the resampling procedure used. Support values for Yinpterochiroptera ranged from 75 (5000 replicates of standard bootstrapping) to 89 (5000 replicates of symmetrical resampling at 3396 change probability). as compared to 96 as reported by Teeling et al. (2005) when Perissodactyla was included (1000 replicates of standard bootstrapping in PAUP 4.10b10 [Swofford, 2003 J. The results of 3a (the unconstrained combined data set both partitions. only) support the traditional groupings analysis including taxa with data Microchiroptera and monophyletic a monophyletic Megachiroptera) among extant bats. Similarly. the results of analysis 3b (the combined data set without constraints, Perissodactyla deactivated) do not support paraphyly of Microchiroptera with respect to Mega- chiroptera. Two fossil bat taxa, /caronycteris and the Table 2. Results of phylogenetic analyses. Descriptions of analyses are as follows (see text for further details): analysis la, morphological data set; 1b, morphological data set with Yinpterochiroptera—Y angochiroptera constraints; 2a, molecular sequence data set, all outgroups: 2b, molecular sequence data set, Perissodactyla deactivated; 2c, molecular sequence data set with Megachiroptera-Microchiroptera constraints, Perissodactyla deactivated; 3a combined data set, extant taxa only; 3b, combined data set, extant and fossil taxa; 3c, combined data set, Yinpterochiroptera- Yangochiroptera constraints; and 4, molecular scaffold analysis, all nodes constrained. MPTs = most parsimonious trees. Analysis la 1b 2a 2b 2c 3a 3b 3c 4 Taxa (outgroup, ingroup, 45 (5, 40, 6) 45 (5, 40,6) 34(4, 30,0) | 33 (3, 30, 0) 33 (3; 30, 0) 39 (5, 34, 0) 45 (5. 40, 6) 45 (5, 40, 6) 45 (5, 40, 6) fossil) Characters (total and no. 204 (202) 204 (202) 13792 (3792) 13792 (3533) 13792 (3533) 13996 (3735) 13996 (3735) | 13996 (3735) NA informative) Ts 4 30 2 1 2 3 16 1 3 Morphological characters Length 791 827 NA NA NA 750, 763, 771 875 (8). 888 (8) 901 923 CI 0.363 0.347 NA NA NA 0.357, 0.360, 0.323 (8). 0.328 0.319 0.311 0.367 (8) RI 0.732 0.713 NA NA NA 0.659, 0.664, 0.681 (8), 0.688 0.674 0.662 0.673 Molecular characters Length NA NA 15782 15238 15249 15258, 15271. | 15258 (8). 15271 15248 15238 15250 (8) CI NA NA 0.531 0.539 0.538 0.537, 0.538 (2) | 0.537 (8), 0.538 0.538 0.539 (8) RI NA NA 0.485 0.49] 0.490 0.489 (2), 0.488 (8). 0.489 0.490 0.491 0.490 (8) All characters Length NA NA NA NA NA 16021 16146 16149 16161 NA NA NA NA NA 0.529 0.526 0.526 0.526 RI NA NA NA NA NA 0.506 0.512 0.512 0.511 uepJer) jeoiuejog unossilN aul jo s¡euuy Volume 95, Number 1 2008 Hermsen & Hendricks Morphological Character Evolution 79 Outgroups New Green River Bat f Icaronycteris t Cynopterus Nyctimene Pteropus Rousettus Archaeonycteris t - Tanzanycteris t Hassianycteris t Palaeochiropteryx t Megachiroptera —- — Rhinolophus Hipposideros Rhinopoma Craseonycteris SEE Megaderma Macroderma Nycteris Taphozous Emballonura Rhynchonycteris Mystacina Myzopoda — Thyroptera E Noctilio Kerivoulinae ure l. Strict consensus of 16 MPTs (16,146 steps; CI = 0.526; RI = 0.512) resulting from the analysis ( 3b) of the combined morphological and molecular sequence data set, including all extant and fossil taxa, without constraints. Fossil taxa are indicated by a dagger (T). Annals of the Missouri Botanical Garden Green River Bat, group outside of all extant bats. The other four fossil bat taxa and extant Microchiroptera Fig. 1). The clade that includes Rhinolophidae, Megadermatidae, Craseonyc- form a monophyletic group teridae, and Rhinopomatidae is partially collapsed due in part to interaction with the fossil taxa (Fig. 1). The strict consensus of three MPTs found during analysis 4 (the molecular scaffold. with all nodes constrained, Perissodactyla deactivated) is shown in Figure 2. Tanzanycteris is resolved sister to Rhinolo- phidae. In one tree, Hassianycteris is sister to all Yangochiroptera and in the other trees is outside of the clade including all extant bats. All other fossil taxa are on the stem lineage of extant bats in all DISCUSSION The most basic problem of studying character evolution in a phylogenetic context is a problem of methodology: how will the trees be built? Different methodologies have been employed to integrate data from fossil taxa (primarily composed of morphological often (now including, or data) and extant taxa composed entirely of molecular sequence data) in analyses that incorporate information from multiple sources. One of these is simply to suggest the position of a fossil taxon on a tree of extant taxa by reference to morphological synapomorphies mapped on a tree (e.g., Rowe, 1988, for select fossil taxa with more than 12% jer et al., 2003). Two others, which take a more analytical approach, are the molecular — missing data; Bouc scaffold (also known as molecular backbone con- straints, molecular constraints, etc.) and simultaneous analysis (also known as total evidence or combined analysis, or the supermatrix approach). The first « these, the molecular scaffold, gives greater prece- dence (at least to some degree, depending on the constraints used) to the topology suggested by the molecular sequence data. In this type of analysis. extant taxa are analyzed using molecular sequence data, all or some of the relationships among these taxa on the resultant tree(s) are constrained, and then a morphological matrix including fossil taxa is analyzed under the constraints (e.g, Springer et al., 2001: Sánchez-Villagra et al. 2003; Roca et al., 2004: supplementary data; Asher et al., 2005a: Teeling et al., 2005). In a total evidence approach (Kluge. 1989) combined or simultaneous analysis (Nixon & r a Carpenter, 1996), all character data are combined into a single supermatrix (see de Queiroz & Gatesy, 2007), and extant and fossil taxa are analyzed together. In the greater influence on the resultant. tree topologies. atter approach, morphological data can have a and inclusion of morphological data with molecular sequence data has sometimes been shown to signif- icantly alter the topologies recovered relative to those found when sequence data are analyzed alone (see discussion below). Generally, simultaneous analyses have been performed under equal-weights parsimony with sequence data aligned prior to analysis, although some authors (e.g.. Giribet et al., 2002: Asher et al., 2003, 2004: Wheeler et al., 2004; Arango & Wheeler. 2007) have chosen to implement direct optimization o — E sequence data (Wheeler, 1996, 2003) during phylog- eny reconstruction, in which different costs can be assigned to morphological and various types of molecular transformations. Inclusion. of fossil taxa in phylogenetic. analyses increases taxon sampling and does so in a very unique way. Fossil taxa represent lineages sampled. through time and. as such, can be repositories of unique morphologies that may not be represented in today's biota. Thus. direct inclusion of fossil taxa in phylogenetic analyses, rather than overlaying a morphological analysis onto a molecular scaffold, can alter tree topologies, sometimes in ways that yield different results than simply combining data partitions E for extant taxa alone. In fact. the addition of fossil taxa representing extinct diversity has the potential to alter he interpretation of relationships among extant taxa any time homoplasy occurs in the data set on which a phylogenetic hypothesis is based (Nixon & Wheeler, 1992). Thus, effectively, fossil taxa that possess unique combinations of characters almost always have the potential to alter the hypothesis of phylogenetic relationships when included in an analysis and, in certain situations (e.g., cases in which large accumu- lations of apomorphies distinguish extant taxa: see 1989), significantly affect the perceived relationships among Donoghue et al., might be expected to extant laxa. Perhaps the earliest illustration of this property in a combined analysis of morphological and molecular sequence data was presented by Eernisse and Kluge (1993). who studied amniote relationships. They performed analyses including morphological (Gau- thier et al., 1988) and molecular sequence (Hedges el al.. 1990) data sets in various combinations with and without fossil taxa. In some pairs of analyses (e.g.. combined 188 rRNA sequences plus morphology). inclusion of fossil taxa was critical; without fossil taxa, birds and mammals formed a monophyletic group to the exclusion of crocodiles, turtles, and lepidosaurs (e.g.. snakes, lizards, tuataras); with fossil taxa, birds and crocodiles were sister taxa in a monophyletic group with turtles and lepidosaurs to the exclusion of mammals. When all characters were considered, the results were consistent with a monophyletic bird- extant replile clade, although the positions of turtles Volume 95, Number 1 Hermsen & Hendricks 81 2008 Morphological Character Evolution Outgroups == New Green River Bat T == /caronycteris T m— Archaeonycteris T pe Crans T Megachiroptera — = Palaeochiropteryx T Rousettus Pteropus Cynopterus — Nyctimene Tanzanycteris 1 Rhinolophus Hipposideros Rhinopoma AR Craseonycteris Megaderma Macroderma Nycteris Taphozous Emballonura pese Rhynchonycteris Natalus Tomopeatinae Tadarida Eumops Antrozous Rhogeessa Miniopterinae Myotis Murininae Kerivoulinae Figure 2. Strict consensus of three MPT’ resulting from analysis (4) of morphological data over a molecular scaffold under full constraints (all taxa with molecular sequence data constrained). Taxa lacking the molecular sequence : ala partition were analyzed without constraints. Tree statistics are given in Table 2. Fossil taxa are indicated by a dagger (F 82 Annals of the Missouri Botanical Garden and lepidosaurs were reversed among extant-only and fossil-extant analyses. Another demonstration of the difference that inclusion of fossil taxa can make was provided by Wheeler et al. (2004), including fossil taxa in a supermatrix of morphological and molecular sequence data for arthropods resulted in some differences in relationships among extant uch showed that Z groups relative to an analysis with extant taxa alone. For example, the extant-only analyses always re- .g., crabs, shrimps, and barnacles) and Hexapoda (insects) as sister groups, solved Crustacea (e whereas the fossil-extant analyses sometimes instead resolv- ed Hexapoda and Myriapoda (e.g., lipedes) as sister groups, depending on the cost centipedes and mil ratios used to optimize the molecular data. A recent study of euphyllophytes by Rothwell and Nixon (2006) compared simultaneous analyses of sequence 2001) with morphological data taxa by data (Pryer et al., (Pryer et al., 2001, with addition of fossil | & Nixon, 2006) when fossil taxa were in- One of the notable differences — Rothwe cluded and excluded. in the resultant parsimony topologies was that lyco- phytes were a stem taxon of the euphyllophytes in 2001), whereas fossil the extant-only analyses (Pryer et al., they to the lignophytes taxa were included and the extant outgroups (bryo- phytes) were replaced with a fossil outgroup taxon presumably more closely related to the lignophytes (Rothwell & Nixon, 2006). Few studies to date have directly compared the were sister when results of fossil-extant molecular scaffold and simul- taneous analyses, specifically considering the effect each approach may have on the relative placements of fossil and extant taxa. Asher et al. (2005a) compared affold analyses of fossil the fossil lipotyphlan simultaneous and molecular scaf = and extant placental mammals with goal exploring the affinities of the (mammalian insectivores) genus Centetodon Marsh. that Centetodon (insectivores) in a more Simultaneous analyses indicated grouped with Eulipotyphla derived position than Solenodon Brandt (solenodons, insectivorous mammals endemic to the Caribbean). The molecular scaffold analysis also indicated that Centetodon belongs within Eulioptypha but did not decisively resolve the position of Centetodon relative — to Solenodon. There were many substantive differ- ences in the inferred interrelationships of mammalian orders and molecular scaffold topologies. (2007) compared simultaneous and molecular scaffold analyses of fossil family between the supermatrix Manos et al. and extant members of the angiosperm walnut family). Both analyses recov- ~ Juglandaceae ered two clades (englehardoids within the family, and the strict consensus trees The and juglandoids) resulting from each analysis were similar. positions of the fossil taxa were also similar between analyses, the biggest difference being that Paleo- oreomunnea stoneana Dilcher, Potter & Crepet (a fruit taxon; Dilcher et al., 1976) grouped with the juglandoids in the simultaneous but not the molecular scaffold analysis. Magallón (2007) compared simulta- neous and molecular scaffold analyses of fossil and extant taxa in the angiosperm family Hamamelidaceae of the whereas (witch hazel family). The strict consensus simultaneous analysis was poorly resolved, the strict consensus of the molecular scaffold analysis had greater resolution. In both the fossil taxon Archamamelis Endress & Friis (a floral taxon: Endress & Friis, 1991) to Hamamelis L. (witch hazel), although the relationships of the other fossil taxa were more ambiguous. The our study show a clear difference in the analyses, was resolved sister results of ea positions of fossil taxa between simultaneous anc molecular scaffold analyses, with fossil taxa being divided into two stem group and four erown group taxa in the unconstrained combined analysis (3b, Fig. 1) and five stem group taxa and one erown group taxon or four stem group and two crown group taxa in the molecular scaffold analysis (4, Fig. 2). The difference in the basal dichotomy among extant bats between the two analyses—caused by addition of morphological data to the pruned molecular sequence data set in the simultaneous analysis, as demonstrated when the combined data set is analyzed with all taxa lacking the is likely molecular partition removed (analysis 3a) affecting the inferred positions of the fossil taxa. Comparison of the results of simultaneous and molecular scaffold analyses and simultaneous analy- ses with and without fossil taxa clearly demonstrates, even though examples are relatively few, that choice of methodology can affect the optimal topologies and, thus, that the type of analysis performed and/or the direct of The rationale for using a total evidence or simultaneous analysis approach to analyzing data (not necessarily including fossil taxa) was made by Kluge (1989), later by Nixon and Carpenter (1996), and more recently by de Queiroz and Gatesy (2007); perhaps the most persuasive argument for such an approach is that all inclusion fossil taxa does matter. putatively phylogenetically informative data should be used to construct. phylogenetic hypotheses. Recent arguments against simultaneous analysis of morpho- with molecular sequence data, against use of that and/or ) that the analyzing = logica select morphological characters (those incongruent with a molecular scaffold), the molecular scaffold approach, include: (1 acc epted tec :hniq ue for — most generally morphological data is parsimony, whereas molecular sequence data may be better analyzed using other 2001; Asher et al., methods (e.g., Springer et al., Volume 95, Number 1 2008 Hermsen & Hendricks 83 Morphological Character Evolution 2005); (2) that molecular sequence data are not available for most fossil taxa, or the missing data 2001; Manos et al., (3) that morphological data are subject to argument (e.g., Springer et al., 2007); Bomoplésy due to convergence (e.g., Springer et al., 2007; Eick et al, 2005); and (4) that simultaneous analyses “fail to address the weighting problem posed by including molecular and morpho- logical data in the same data matrix” (Springer et al., 2004: 436). The first point is certainly debatable with regard to molecular sequence data sets (e.g., Frost et al., 2001) morphological data (Lewis, = and is becoming moot with regard to 2001; Springer et al., 2004). For example, recent analyses of morphological (e.g., Müller & Reisz, 2006) or combined morphological and molecular sequence 2004; sometimes including fossil taxa (Lee, 2005; Xiang et al., 2005; Asher & Hofreiter, 2006; Müller & Reisz, 2006), have used Bayesian methods, although this little both phylogeny building and the study of character The latter three arguments may be considered aspects of the also noted by data (Glenner et al., Nylander et al., 2004 9 — approach is relatively explored for evolution using morphological data sets. topic of character evolution especially pertinent to including fossil taxa in simultaneous analyses with extant taxa. Below, we address the arguments against including fossil taxa in simultaneous analyses of morphological and molecular sequence data (primar- ily consisting of arguments against using morpholog- ical data in phylogenetic analyses) and go on to discuss some of the specific benefits accrued. and obstacles encountered to the study of character evo- lution when including morphological data from fossil taxa in phylogenetic analyses. MISSING DATA—A JUSTIFIABLE REASON FOR EXCLUDING FOSSIL TAXA FROM THE PROCESS OF PHYLOGENY RECONSTRUCTION? One criticism that has been levied against the inclusion of fossil taxa in simultaneous analyses is that fossil taxa may be missing significant amounts of data. For example, Springer et al. (2001: 6242) in part rejected the total evidence approach because “mo- lecular data are usually unattainable for fossils." Perhaps the most serious methodological consequence of including fossil taxa with significant amounts of missing data into an analysis is a weakening of the parsimony criterion, which is strongest when present- ed with maximum evidence: “the tree that is best corroborated is the tree that best explains (e.g. homology) all character distributions among all taxa” (Nixon, 1996: 369 One way in which this weakened test of parsimony may manifest itself is through the wild card taxon phenomenon, in which a taxon (or taxa) with large amounts of missing data may group at numerous different positions on the shortest discovered tree topologies due to its limited distribution of character states (Nixon & Wheeler, 1992). Nixon and Wheeler (1992) noted that the inclusion of wild card taxa in a result in a significant increase in the MPTs consensus of all MPTs. matrix may number of and deresolution of the strict The best solution to such a problem, when encountered following phylogenetic analysis, may be to remove such taxa, provided that the lack of problematic taxon can reasonably be attributed to resolution in the position of the missing data and not at least in part to character (2002). in recognition of three different types of wild card taxa: conflict. Kearney fact, has suggested [om data wild cards, whose instability is ~ 1) missing entirely caused by missing data; (2) mixed wild cards, whose instability is due both to missing data and character incongruence; and (3) conflict wild cards, taxa whose instability is entirely due to character conflict. Missing data wild cards can be identified by taxonomic equivalent analysis (Wilkinson, 1995), in which fragmentary taxa that are identical with more complete taxa in the characters for which they are coded are removed if an initial analysis shows them to be wild cards, thus eliminating redundancy from the analysis and possibly increasing the resolution of the resultant topologies (see examples in Kearney, 2002). more significant barrier to including fossil taxa is the unpredictable effect(s) that missing data can have on a parsimony analysis when character incongruence is encountered, and/or when mixed or conflict wild cards are present. Due to the weakened test of character congruence and the tendency of parsimony to underestimate tree length when large amounts of un data are present, Nixon (1996: 370) suggested *we should be suspicious when the addition of fossils with large numbers of missing data results in significantly different topologies than when they are excluded." When large amounts of extinction have occurred within a clade, however, significant rear- rangements may be expected when fossil taxa are sampled (see examples from Eernisse [1993], Wheeler et al. [2004], and Rothwell and Nixon [2006] discussed above), especially when these taxa represent much of a group's diversity. This conundrum may be insoluble, since more complete character data will often be unavailable for fossil taxa. Worse, in combined morphological-molecular sequence data sets, fossil taxa will be coded for a very small proportion of characters, as they will likely be missing the entire molecular sequence partition. In the present study, for example, /caronycteris is coded (with one or states) for only about 2% of all more character Annals of the Missouri Botanical Garden parsimony-informative characters in the combined matrix, and the Green River fossil bat taxon for about 1.396; Tanzanycteris, whose position within or outside of crown group Microchirptera varies, is coded for only about 0.696 of parsimony-informative characters, the smallest proportion. among all fossil. bat taxa included in the matrix. Although threshold values for excluding taxa on the basis of the proportion of missing data have been used prof 8 in some studies (e.g., Rowe, 1988; Benton, 1990; 1998), these have been rejected both logically ( Kearney & Clark, 2003) and on the basis of simulation and empirical studies. Using simulations, Wiens (2003a) showed that it not the amount of missing data that determines whether Grande & Bemis, e.g., IS the position of a taxon will be unambiguously and correctly resolved but rather whether critical charac- that This conclusion principle of ters are coded for laxon. conforms to a basic phylogenetic systematics: it is only critical character state trans- formations—synapomorphies—that provide grouping information (Hennig, 1966). Thus, the missing data em becomes one of too few characters rather Pede prob than too much missing data, since greater sampling of characters increases the chance that critical char- acters needed for precise and accurate. resolution of a taxon's position will be included in the analy- i 2003a). In TIME study, 309) predicted, placed in an anal phological data alone, they should be accurately placed in the combined analyses as well, regardless of their relative level of incompleteness when the sis (Wiens, Eee Wiens (2003b: can be accurately the fossil taxa ysis of the mor- molecular data are added.” Wiens’ prediction has been, for the most part, borne out by studies employing pseudofossil analyses. In a pseudofossil analysis, one or more extant taxa are coded for only a subset of characters in the combined data matrix, with all molecular sequence characters and often a portion of the morphological character matrix coded as missing so that the pseudofossil(s) simulate the behavior of a fossil taxon (or taxa) during simultaneous analysis of a real data set (real fossil taxa are excluded). Jordan and Hill (1999), Jordan and Macphail (2003), and Asher and Hofreiter (2006) used an approach in which one taxon at a time was treated as a pseudofossil in combined morphological and molecular sequence data matrices in order to evaluate the of terminals in simultaneous analyses of real data sets. In each case, to the behavior these incomplete pseudofossils were found to place reliably general area of the tree suggested by analyses of the taxa with the full complement of characters, although the results of analyses with pseudofossils were not identical necessarily to analyses in which the full matrix was analyzed. Manos et al. (2007) took different approach and created pseudofossils by duplicating one extant taxon at a time and randomly eliminating all but 25%, 50%, 75% of the duplicate's morphological characters or by eliminating all but its organ-specific (i.e., floral, fruit) characters and performing simultaneous analy- Although the randomly generated pseudofossils to their parent species in the or vegetative, were seldom sister “in the correct ~ resultant topologies, they did place local and neither of the two large clades (engelhardioids or juglandoids) was disrupted” (Ma- nos et al., 2007: 422); “[rlemoval of suites of organ- specific characters did not show appreciably different results” (Manos et al., 2007: 422). contrast, Springer et al. (2007)—in a pseudofossil study of a n which clade, In combined matrix of placental mammals, 1 ordinal or superordinal groups were treated as pseudofossils by eliminating all but the osteological data partition for each taxon in that group—often found profound rearrangements in topology relative to a tree based on molecular sequence data alone. Although Springer el (2007) attributed these rearrangements to the inadequacy of the morpholog- they did not establish whether taxon removal (i.e., taxon sampling) affected the topologies favored by the molecular sequence data partition al. ical data set, alone. Thus, the underlying reasons for these rearrangements may have been more complex. Studies such as those cited above suggest that, as previously noted by Kearney (2002: 370), ds PE effects of incomplete taxa and concomitant. missing character data are not general, but matrix-specific, and depend on the precise distribution of question marks and characters states across taxa"; also see Novacek (19922). Therefore, taxa with large amounts of missing data should not be considered a priori unsuitable for inclusion in phylogenetic analyses. including simultaneous analyses of morphological and molecular sequence data. The biggest problem that missing data present, then, may be the weakened application of parsimony (however, see Kearney & Clark, 2003) case when incongruent which may have undesirable effects on the results. such as deresolution of the strict consensus tree or support for misleading topologies. Novacek (1992b: 75) perhaps best summarized the problem: “The kinds not just f account for the potential his is especially the data are concentrated in fossil taxa, of characters preserved, the degree o character representation, influence of characters preserved are the combination of primitive and derived states that force relationships in a particular direction, then the included taxa—even an added taxon. If among the few when poorly represented—will play a significant role Volume 95, Number 1 2008 Hermsen & Hendricks 85 Morphological Character Evolution in the outcome. Of course, the possibility that these incompletely preserved taxa force the wrong outcome, which would be apparent had the taxa been better represented, cannot be eliminated." Unfortunately, at this time, there is no definitive way to differentiate between the correct and the incorrect topology in such a situation. The molecular scaffold approach may appear to circumvent this problem by building a topology that includes only extant taxa (for which all or most molecular sequence characters can be scored) to serve as a basis for constraints on the analysis of the morphological data, thereby isolating the ambiguity caused by missing data into unconstrained taxa in the morphological partition. Although excluding fossil taxa from the scaffold-building step thus inoculates the molecular partition from the computational and philosophical problems that can be caused by in- cluding large numbers of ambiguous cells in a data matrix, the underlying missing data problem is not truly eliminated. This is because excluding fossil taxa from direct phylogenetic analysis also reduces taxon sampling, which in general can negatively impact the accuracy of results (e.g., Zwickl € Hillis, 2002). CONVERGENCE According to Givnish and Sytsma (1997: 56), there are four mechanisms that can result in homoplasy: "evolutionary convergence, recurrence, transference, and character misclassification [bold emphasis re- Evolutionary convergence or parallelism (cf. 2005 morphological moved ]." Wiens et al., 2003; Desutter-Grandcolas et al., result A of form may when similar d to similar selective pressures are discovered. ere appears to be a perceived notion among some workers that convergence of morphological form is so pervasive and misleading that it is justifiable to either (1) simply exclude morphological data from the construction of phylogenetic hypotheses or (2) regard morphological data as suspect when topological conflict with molecular sequence data occurs (Hedges & Sibley, 1994; Hedges & Maxson, 1996; Givnish & Sytsma, 1997; Eick et al., 2005). There is no denying that convergence of form is a widespread feature of evolution and that it has occurred at many scales and in many taxa through geologic time (some examples include ichthyosaurs and dolphins, rudist bivalves and reef-building corals, some placental and marsu- pial mammals, the mangrove habit in plants, etc.). As by Wiens et al. (2003: 501), *Convergence is a critical issue in systematics because it can potentially [om notec mislead phylogeny reconstruction methods, for exam- ple. causing analyses to group distantly related organisms that share similar habitats." Of particular concern are cases in which conver- gence has potentially led to correlated patterns of evolution in suites of functionally related morpholog- ical characters included in an analysis (see, for instance, Wiens et al., 2003). In such cases, however, precise character definition (i.e., hypotheses of homology) and careful morphological study have reduced and will continue to reduce the occurrence of this problem (that is, the fourth class of homoplasy & Sytsma [1997]: misclassification). One approach is atomization and given by Givnish character coding of such morphological complexes, which can highlight differences in apparently convergent mor- phologies. Bruneau (1997), for example, contrasted the character complex of pseudo-tubular corollas, convergent among hummingbird-pollinated species of the legume Erythrina L. (a legume), with the atomized corolla characters detailing specific aspects of size and morphology. Consideration of pseudo-tubular corollas as a single presence-absence character would have led to a relatively uninformative hypothesis of homoplasy due to multiple origins, while breaking the complex pseudo-tubular corolla character into several atomized petal characters allowed for the detection of convergence in a single aspect of the corolla morphology (see also Luckow & Bruneau, 1997). Nixon (1996: 368) pointed out that it is especially important that homology assessments be approached carefully with fossil taxa because “poorly preserved fossils may have a higher likelihood of being misunderstood and therefore incorrectly scored for — those characters that are not missing [boldface in original].” Thus, in some situations in which struc- tures are poorly understood, perhaps fossil taxa should be coded using different homology assessments and the results of analyses compared or the fossil taxa should be excluded from analysis altogether. Despite this last caveat, the problem of morphological data being seriously compromised by evolutionary conver- gence may be overstated. There convergence of morphological characters may have have been demonstrated cases in which produced misleading results. An interesting example was provided by a study from Gatesy et al. (2003) on the relationship of erocodylians to the extant gavial (Gavialis gangeticus (Gmelin)), in which results of analyses from a morphological data set (modified from Brochu, 1997), molecular sequence data sets, and a combined data set for fossil and extant crocodilians were compared. Analysis of the morphological data alone with and without fossil taxa suggested that the extant gavial was the most basally diverging lineage of the extant crocodylian groups; in contrast, molecular sequence data suggested that the gavial was in a derived position in the tree, sister to the false gavial 86 Annals of the Missouri Botanical Garden Tomistoma schlegeli Müller. When the data sets were combined, the strict consensus of the resultant MPTs was congruent with the molecular results for extant taxa, suggesting that the extant gavial is an atavistic laxon. (2003) could have chalked up these results to the unsuitability of the morpho- Although Gatesy et al. logical data, beset by convergence, to discover the proper position of the extant gavial among crocodyl- ians, they instead tested the morphological data for secondary signals. Secondary signals (Nixon € Carpenter, 1996) or hidden support (Gatesy et al., 1999) are patterns present in a data partition that are masked when that alone—reflecting a primary partition is analyzed signal—but which may manifest themselves as character support for clades conflicting with the primary signal for the partition when analyzed combination with other data partitions. When data sets that independently produce competing. results e., have different primary signals) are combined and analyzed together, secondary phylogenetic signals common to two or more data sets may even lead to novel hypotheses of phylogeny not expressed in any see Barrett et al., 1991; Nixon & 1999). In other words, "[t]he peculiarities of — individual partition 1906; Gatesy et al., during simultaneous analysis, Carpenter, each data set are cancelled. out by the unique pe C uli arities of f the othe US and the rc maining common 1999: 301). al. (2003) tested for hidden support for the derived signal emerges" (Gatesy et al., Gatesy et extant gavial hypothesis in their morphological character data set using partitioned hidden branch support (PHBS; Gatesy et al., 1999). PHBS is the difference between the branch support values for a eiven data partition and a given node in combined and 999); a score for a particular data set indicales a partitioned analyses (Gatesy et al., "positive BS secondary phylogenetic signal for the relationship of that (Gatesy et al., emerges in simultaneous analysis” 2003: 407). Using PHBS, Gatesy et al. (2003) found hidden morphological support for interest nodes linking the extant gavial to the extant false gavial and the Crocodilynae (crocodiles) to the gavial and false gavial in the combined morphological and molecular sequence topology for extant taxa. This support was increased by the addition of fossil taxa. In a more recent expanded analysis of crocodylians that increased taxon and character. sampling, (2004: 347) found “11 that emerged in the supermatrix analysis” that implied by any combination of trees supported by included Gatesy et al. eroups “were not separate analyses of the 17 data sets in the super- matrix.” thereby revealing the secondary signals that manifested themselves during simultaneous analysis. — Similarly, secondary signals must be at work in the simultaneous analysis of Chiroptera presented here, because the two major clades of bats resolved during simultaneous analysis are concordant with the primary morphological but not the primary molecular se- signal (Fig. 1), suggesting that secondary the support the traditional Megachiropteran and Micro- quence signals in molecular sequence partition may D ? chiropteran groupings. The discovery of these secondary signals is only possible through simultaneous analysis and is an the important advantage of this approach over molecular scaffold approach. For instance, in the al. (2005a) on the extinct study from Asher e insectivorous mammal Centetodon, simultaneous anal- yses of molecular sequence and morphological data indicated that Centetodon was more derived than the Caribbean endemic insectivore Solenodon, whereas a molecular scaffold analysis failed to decisively resolve the position of Centetodon relative to Solenodon. Asher et al. (2005a: 919) considered this ironic since Holarctic lipotyphlans [mammalian insectivores] results from “a basal position for Solenodon within the molecular signal responsible for the molecular scaffold, and not from the influence of the morpho- alone, morphological data logical data. Analyzed EE] weakly support Erinaceus as basal” within the clade. Thus, the molecular scaffold actually obscured both the molecular sequence signal as well as the signal common to both data sets when analyzed together. Furthermore, the molecular scaffold approach may tend to force more conflict into the data set that is analyzed on the scaffold, depending on the constraints used. This creates a self-fulfilling prophecy whereby A homoplasy—sometimes explained ad hoc as evidence amounts ^ the morphological data contain larger of convergence (see also Luckow & Bruneau, 1997)— relative to the results of a morphological analysis alone, and sometimes relative to the results of a simultaneous analysis. The bat data set presented here is an interesting example of this phenomenon. The morphological data gained 132 steps relative to the MPTs for the morphological partition alone, and 35 or 48 steps relative to the MPTs for the simultaneous analysis when analyzed under a strict molecular scaffold (compare analyses la and 3b to analysis 4, Table 2). The overall CI and RI for the combined analysis (3b), the combined analysis with Yinpter- ochiroptera-Y angochiroptera constraints (3c), and the strict molecular scaffold (4) are nearly identical for (Table 2). distribution of homoplasy between partitions is shifted the combined data set However, the from the molecular sequence to the morphological characters as the data are put under constraints that are progressively more favorable to the primary Volume 95, Number 1 2008 Hermsen & Hendricks 87 Morphological Character Evolution molecular sequence signal. The CI and RI become progressively lower (e.g., more homoplasious) for the morphological data and higher for the molecular 2). The change in the CI and RI values for the morphological partition is also greater, that, although difference in the number of steps for each partition sequence data (Table reflecting the fact the absolute is similar between the most and least parsimonious analyses for that partition (35 or 48 steps for morphology, 20 or 33 steps for molecular sequence), proportionally more steps are forced into the morpho- logical partition in less parsimonious topologies for that partition than for the molecular sequence data set, because the molecular sequence data set is much larger (Table 2). Eick et al. (2005) recently took assumptions about convergence in morphological characters to an extreme and analyzed fossil-extant bat relationships on a molecular scaffold using only morphological characters that were nonhomoplasious when mapped on a topology suggested by molecular sequence data because “the exclusion of a significant amount of homoplasious characters can potentially alter the (Eick et al., 2005: 1872) and "the high level of parallel evolution when using (Eick et al., they did not Cc 'onclusions reached" morphological characters is problematic" 2005: 1879). homoplasious molecular sequence characters from Interestingly, remove their analyses, although homoplasy was present in the molecular sequence data (the ensemble RI was less than one for each data set when analyzed using parsimony). Molecular sequence data are not neces- Lee (1997) noted that biologists understand functional sarily exempt from evolutionary convergence. morphology well and are adept at recognizing and articulating the adaptive significance of some aspects of the morphologies of organisms. Conversely, Lee (1997) questioned whether the young discipline of functional molecular biology is yet as capable of recognizing adaptive convergence in sequence data. In fact, recent work has demonstrated the occurrence of evolutionary convergence at the molecular level (e.g., Bull et al., 1997; Zhang & Kumar, 1997; Cuevas et al., 2002; Zakon, 2002; Protas et al., 2006). While we agree with some authors (e.g., Hedges & Maxson, 1996) that molecular sequence data may be less prone to similarity caused by evolutionary convergence or parallelism, molecular sequence data are also subject to homoplasy from recurrence (random mutation leading to noisy homoplasy; see Wenzel & Siddall, 1999) and transference (horizontal gene transfer), as well as character misclassification (e.g., due to the amount of homoplasy in sequence data may be as misalignment). Because of these four factors, great as or greater than that in morphological data 1989; Baker et al., 1998). Thus, it should be borne in mind that sequence (e.g., Sanderson & Donoghue, data do not have special immunity from homoplasy— as powerfully evidenced by "the fact that different genes yield different phylogenies" (Lee, 1997: 394)— and, as in morphological data, this homoplasy is — potentially misleac Only through the addition of more taxa and more characters Ing. additional tests of a phylogenetic hypothesis can be the ultimate arbiter of the robustness of a suggested set of relationships. As perhaps best summarized by O'Leary et al. (2003: 861), "Without insights into some yet undiscovered law of nature, there is no particular reason to think that a functional, developmental, or ecological explanation for homo- plasy is a better explanation of covariation than is synapomorphy. Simply proposing such generalities does not condemn characters to being phylogeneti- cally uninformative.” THE WEIGHTING PROBLEM It has been argued that in a simultaneous analysis of morphological and molecular sequence data, the large amounts of molecular sequence data will simply — overwhelm (or swamp) the signal of the morphological data set. For example, Alvarez et al. (2000: 184) argued that the morphological and molecular se- quence data sets they used to reconstruct the phylogeny of a family of sponges "are unequally sized (95 parsimony-informative molecular characters vs. 16 parsimony-informative morphological characters) so that the molecular signal, which is different, will swamp the morphological signal. Therefore, a simul- taneous analysis (e.g. one including both types of data) was considered inappropriate.” Some might argue that because morphological transformations might be presupposed to involve more complex underlying mechanisms than transformations among acid perhaps morphological data nucleic pases, partitions should be upweighted with respect to jd molecular sequence data partitions. This assumption may be analogous to that employed in some analyses “Within paradigm of probabilistic assumptions, it is often of molecular sequence data sets: argued that transitions are more probable (i.e., observed more frequently) and should therefore deserve a lower cost...than transversions. However, no consensus has 39 reached for the appropriate cost ratio, and an arbitrary choice is required” (Frost 2001: 354 schemes may be the problem alluded to by Springer et al. (2004: 436) difficulty with combined analyses is that they fail to et al., ) The arbitrariness in weighting "One potential when they wrote, address the weighting problem posed by including Annals Meet. Bos Garden molecular and morphological data in the same data matrix." Empirical studies of combined morphological and molecular sequence data sets do not bear out the assumption that large numbers of molecular sequence data will necessarily overwhelm the signal present in less numerous nonmolecular sequence data when all characters in a combined matrix are weighted equally (e.g.. Omland, 1994; Mattern & 2004: Wahlberg et al., 2005). This observation ties into the McLennan, idea of secondary signals discussed above: when the data are combined, a common signal present in both = data sets may be expressed that is not evident i either data set (or one of the data partitions) when — analyzed alone (see similar argument made by Barrett et al., 1991). Obviously, the bat data set analyzed here exemplifies this: although the molecular sequence data set is much larger than the morphological data set, even when only parsimony-informative characters are considered (202 morphological, 3533 molecular sequence, or a ratio of parsimony-informative molec- ular sequence to morphological. characters of over 17:1), ings that are congruent with the morphological but not the combined data set produces major group- the molecular result, clearly showing that the morphological signal is not completely swamped by the signal from the molecular sequence data set but rather contributes to the structures of the most parsimonious topologies. A study done by Baker et al. (1998) specifically and examined the distribution of morphological molecular sequence character support for nodes in trees generated by equal-weights parsimony applied to combined data sets. These authors compared the partitioned Bremer support (PBS; Baker € Desalle. 1997; Lambkin et al., 2002) values from 10 previously published combined morphological-molecular se- quence data sels (five additional combined data sets in their study featured non-DNA sequence molecular data) with ratios of parsimony-informative morpholog- ical to molecular sequence characters ranging from 0.08-0.33; results of the incongruence length differ- ence (ILD) test (Farris et al., 1995a, b) for these data sets indicated that in five of 10 cases, there was significant incongruence between morphological and molecular sequence data partitions. When PBS values were summed across nodes for each of the studies, the morphological data provided proportionally slightly more support (relative to the number of parsimony- informative characters in each partition) to the topology resulting from simultaneous analysis than molecular sequence data for nine of the 10 data sets. Notably, 10 instances, less homoplasious than the molecular sequence data sets on the topology(ies) resulting from the morphological data were also, in nine of simultaneous analyses as measured using the CL of parsimony-informative characters for each partition: Baker et al. (1998) suggested that this may account for the greater impact of. the morphological partitions relative to their sizes. Similarly, Lee (2005) combined morphological and molecular sequence data in analyses including extant and fossil squamates (e.g. lizards, snakes, amphisbaenians, and extinct relativ es PBS The (under parsimony values. results of the and examined combined analyses and Bayesian criteria) were mostly congruent with the result of the morphological analysis alone (under the parsimony Furthermore, “PBS parsimony analysis] revealed that the morphological (although weak) is still than the " (Lee, 2005: 229), even though there were more parsimony-informative molecular sequence criterion). values [from the signal stronger molecular signal characters than morphological characters. (See also Wortley € Scotland, 2006, using different methods.) Baker et al. ( for a more recent study 1998) have suggested, in fact, that the larger data partition in a combined matrix should have more influence on the results of a phylogenetic analysis—for the result to be otherwise would cast doubt on the quality of the larger data set. Another argument against the necessity of weight- ing—character independence—has been effectively addressed by a number of authors (e.g. Kluge & 1969: Kluge, 1989: Nixon & Carpenter, 1996). Nixon and Carpenter (1996), who made a thorough Farris, argument for employing equal weights in a parsimony analysis a priori, addressed the issue of weighting data partitions using an appeal to character independence. Characters chosen for inclusion in a cladistic analysis are hypothesized to be independent of one another they are logically independent, or assumed to evolve 1983; Kluge & Wolf, 1993; 1996). If characters are assumed independently (Farris, Nixon & Carpenter, to be independent from the outset, as they are in a parsimony analysis, then there is no more justification lo apply weights among different data partitions than lo apply differential weights to characters within a data partition. Thus, the perceived greater complexity of morphological characters need not be used justification for upweighting them relative to molec- ular sequence characters. Some have employed differential costs for morpho- molecular transformations, however, logical and without resorting to completely arbitrary weighting schemes. In the context of simultaneous parsimony analyses including both morphological and molecular POY which differential sequence data, this has been done using (Wheeler et al.. 1996-2003). in costs can be assigned to transitions, transversions, insertion and deletion events, and morphological Volume 95, Number 1 2008 Hermsen & Hendricks 89 Morphological Character Evolution transformations, and direct optimization is employed 1996, 2003). In empirical is applied to simultaneous during analysis (Wheeler, POY analyses of morphological and molecular sequence n which studies data with fossil and extant taxa, measures such as [LD metrics and a topological index have been used to choose the optimal weighting scheme(s) from those applied to the data (Giribet et al., 2002; Asher et al., 2003, 2004). Finally, the molecular scaffold approach does not overcome the weighting problem and, in fact, appears to resurrect a widely recognized weighting problem that arises when topology is used as a proxy for character matrices: although an explicit weighting scheme is not imposed, an implicit weighting scheme Barrett et al., 199 scaffold implies differential weights for the characters is employed (e.g., 1). The molecular used ti construct. the molecular scaffold topology relative to the characters in the matrix analyzed under the scaffold constraints, and these weights are node specific. The morphological characters are effectively weighted to zero for extant taxa where nodes are constrained (e.g., the nodes cannot be overturned by morphological characters), whereas the molecular sequence characters, where applicable, are effectively Thus, far weighted to zero at unconstrained nodes. from circumventing the weighting problem, the molecular scaffold generates a unique weighting problem of its own. Whether or scaffold approach under the rubric of a weighting not one wishes to view the molecular e! the implicit assumptions built into the that constrained, the molecular sequence data are consid- scheme, approach are contradictory. At nodes are ered reliable, whereas at unconstrained nodes they are not. Thus, in a strict molecular scaffold—in which all nodes suggested by the molecular sequence data are 1999: Sánchez- Villagra et al., .. 2004; Asher et al., 2005a; Lee, 2005; 2005)—the sequence data are assumed to provide a constrained (e.g., O'Leary & Geisler, 2003; ; Roca et al 2005; Teeling et al., Xiang et al., reliable estimate of phylogeny at all nodes where they are informative, whereas the morphological data only contribute to the structure of the tree where molecular In a semi-strict sequence data cannot be applied. em molecular seaffold—in which only some nodes data are Lee, 2005; Manos et al., recovered. by the molecular sequence 2001: 2007: only Springer et al., 2006: Magallón, sequence data are constrained. (e.g say & Cozzuol, 2007)—1he contribute to the tree topology at nodes that meet a allowed to given criterion (for instance, have certain support values: cut-off values for nodes in semi-strict molecular scaffolds have ranged from 70% [Lee, 2005] to 90% [Springer et al., 2001] in previous studies), whereas the morphological data contribute structure only at nodes that do not meet the threshold criterion or where sequence data do not apply. Thus, an implicit and contradictory judgment about the value and reliability of each data partition for phylogeny reconstruction is built into the molecular scaffold approach—molecular sequence data are generally favored, whereas morphological data have value only when evidence from sequence data is poor or nonexistent. CHARACTER MAPPING AND THE STUDY OF MORPHOLOGICAL CHARACTER EVOLUTION Regardless of their effect on tree topology or the method used to include them in an analysis, fossil taxa are undoubtedly very significant in the application of phylogenetic hypotheses, specifically | interpreting morphological evolution within and across monophy- letic groups. In fact, it seems safe to suggest that the study of character evolution across major organismal groups such as the seed plants and amniotes must incorporate fossil taxa to be relevant. This is because so much of the diversity that links extant end points— or even stands on its own as a major component of the evolutionary history of these groups—is missing in analyses that incorporate only extant taxa. In such cases, the hierarchical nature of evolution coupled with the removal of diversity through extinction may have the power to obscure relationships and morpho- logical character evolution in a way that can only be addressed through reference to the fossil record (e.g., Wheeler, 1992; Cobbett et al., 2007). Before discussing the potential benefits and difficulties of studying character evolution using a phylogenetic hypothesis that includes fossil taxa, it is necessary to address an old argument that continues to persist in the literature: that it is circular to map a character on a phylogeny when that same character is incor porated into the matrix on which the phylogeny is based. Thus, the argument goes, if one wishes to explore morphological character evolution, molecular sequence data should be used to construct a phylogeny, and morphological data should be mapped on the phylogeny after the fact, wherein they can be used to trace the evolution of morphological traits (Armbruster, 1992; Hedges & Maxson, 1996, 1997); or, more generally, characters to be mapped should be — removed from the matrix when building a phylogeny so that that phylogeny is independent of the mapped 2001). If inclusion of morphological characters in matrices and subsequent characters (e.g., Springer et al., mapping of those same traits on the resultant trees — were circular, this would pose a grave problem for paleontologists. How best to study morphological Annals of the Missouri Botanical Garden character evolution in a phylogenetic context when morphological characters are usually the only source of information about fossil taxa that can be harnessed to build phylogenies including those taxa? The most fundamental assumption of the indepen- dence or circularity argument appears to be that the study of character evolution should be independent of tree building, or that the character(s) under consid- eration in proposing a specific evolutionary hypothesis must be independent of the construction of the tree used to test that evolutionary hypothesis (Coddington, 1988; Brooks € MeLennan, 1991). However, there is no logical reason that a phylogeny should be independent of the characters to be mapped on it. If the characters to be mapped are assumed to be among the class of characters that are suitable for phylogeny reconstruction, they should not be independent of phylogeny, and there is no circularity in mapping ogy or topologies on which they have been optimized during the tree search (e.g., Deleporte, 1993; Kluge & Wolf, 1993). This is because all characters coded for a terminal and hypothesized to be the result of the evolutionary history those same characters onto the phylogenetic topo coexist within the same taxon are of that taxon and informative with respect to phylogeny. Thus, independence of a phylogeny from a character to be mapped on that phylogeny is not really achieved by removing a phylogenetically informative character from the analysis: although the removed character is independent of the mechanics of building the phylog- eny, it is not independent of the true phylogeny or evolutionary history of the taxa under consideration in the analysis, which we presume to exist independently the in order of our ability to correctly recover it. In fact, unintended consequence of character removal to avoid circularity may be to make phylogenetic hypotheses themselves less robust, because evidence pertinent to evolutionary history is being ignored (see Kluge, 1989; Nixon & Carpenter, 1996), which can impact the tree topologies recoverec. That argument addressed, inclusion of fossil. taxa among the terminals in a. phylogenetic MC has great potential for elucidating the evoluti y histories of individual aspects of whole organisms. mem ations include: (1) determination of the sequence of character transformation; (2) insight into the timing of evolution- ary events; and (3) inference of unknown aspects of the morphology (e.g., due to incomplete preservation) or behavior of extinct organisms and especially testing of evolutionary hypotheses. SEQUENCE OF CHARACTER EVOLUTION Inclusion of fossil taxa as terminals on trees can help to bridge gaps in our understanding of morphological character evolution as viewed solely from the present. Fossil taxa provide a broader sampling not only of morphologically but also of temporally diverse taxa, and thus are actual data points representing samples stages in the 2007). As such, their inclusion as terminals in phylogenetic taken from across time at different evolutionary history of a group (Cobbett et al., analyses may alter the perceived sequence of character state transformations in a group regardless of whether their inclusion significantly alters tree topology. A simple situation in which a fossil taxon or fossil taxa can impact understanding of character evolution is when a fossil taxon or fossil taxa are intercalated among the synapomorphies along an internode that would be unbroken if only extant taxa were consid- ered. Intercalation of such taxa may indicate the sequential evolutionary order of synapomorphies that otherwise cluster at the same node when only extant in a simultaneous axa are considered. For instance, - analysis of morphological and molecular sequence a data for extant and fossil taxa within the angiosperm order Saxifragales by Hermsen et al. (2006), inclusion of the Cretaceous flower and fruit taxon Microaltingta Zhou, Crepet & Nixon indicated that the evolution of unisexuality and loss of the corolla preceded the evolution of winged seeds and decurrent stigmas in the stem group leading to the clade including extant Cercidiphyllum Siebold € Zucc. (katsura) and Altin- family). Thus, inclusion of glaceae (sweet Microaltingia broke up a series of synapomorphies gum = c that would otherwise have occurred together at t node subtending the extant clade, assuming the same topology and character optimizations if Microaltingia were removed (Hermsen et al. 2006). Horovitz (1999a) compared the results of simultaneous analy- ses of morphological (primarily skeletal) data and molecular sequence data from platyrrhines (New World monkeys) which the relative relationships among the extant taxa were the same with and without inclusion of the fossil taxa. In at least three specific broke "a stepwise instances, inclusion of the fossil taxa up clusters of synapomorphies, introducing appearance of different characters along the phylog- eny, that would seem to appear in larger clusters of synapomorphies in the phylogeny composed of Recent (Horovitz, 1999a: 24). One subfamily Callitrichinae (marmosets laxa only” instance involved the and tamarins), which was supported by eight synapo- morphies in the extant-only analysis; these character transformations were spread among four separate included. Horovitz nodes when fossil taxa were (1999a: 26) suggested that the breakup of the synapomorphy cluster by the fossil taxa "shed some light on the process that was presumably a conse- D » quence of reduction in body size." Volume 95, Number 1 2008 Hermsen 4 Hendricks 91 Morphological Character Evolution Inclusion. of fossil taxa may also change the perceived utility of characters to provide grouping information. Character states that are autapomorphic (parsimony-uninformative) when only extant taxa are considered may turn out to be synapomorphies when 1989). Horovitz (1999a) provided a good example of this in fossil taxa are included (Donoghue et al., her comparison of topologies recovered from simulta- neous analyses for extant platyrrhines with and without fossil taxa. In that example, Callimico goeldii Thomas (Goeldi's marmoset) is offset by four autapo- morphies in the tree including only extant taxa; two of these autapomorphies became synapomorphies for Callimico and two fossil taxa (Patasola magdalenae y & Meldrum and Carlocebus carmenensis Fleagle) os the fossil taxa were included in the analysis 19993). mations that appear to be unambiguous synapomor- (Horovitz, Alternatively, character transfor- phies when only extant taxa are considered may become pure gr more diversity is sampled . The Gatesy et al. (2003) with and (Donoghue et al., simultaneous a of crocodylians without fossil taxa provide examples of this. As they noted, “A more complete sampling of taxa [inclusion of fossil taxa] uncovered additional homoplasy, revealed uncertainties in character optimizations, and ultimately overturned hypotheses of homology that were based solely on the extant biota” (Gatesy et al., 2003: 412). As a specific example, among extant taxa alone, presence of "rectangular dorsal midline osteoderms" was an unequivocal synapomorphy for Tomistoma schlegelii (false gavial) and Gavialis gangeticus (gavial), whereas “[t]he combined evidence topology [including fossil taxa] implied that square/ equant osteoderms instead were independently de- rived from rectangular osteoderms within Alligator- 2003: 412). Fossil taxa may also help to establish the polarity oidea and Crocodylinae" (Gatesy et al., within characters present in a clade but inapplicable in the most closely related extant clades; this may be especially true for groups offset by many characters not represented in their nearest living relatives (Donoghue et pi 1989). Illustrative of this phenom- enon in the context of simultaneous analysis o morphological with molecular sequence data is that of the relationship of the gavial to the remainder of 2003, above). Simultaneous analysis of morphological and extant crocodilians (Gatesy et al., introduced molecular sequence data including both extant and fossil crocodilians resolved some fossil taxa along the stem lineage leading to extant crocodilians, their presence clarifying and, in some cases, overturning the polarity of character states within characters for the ingroup when compared to the analysis of data from extant taxa alone. For instance, when only extant crocodilians were included in the simultaneous analysis of morphological with molecular sequence data, character state polarity could not be established in the occlusal pattern of dentary teeth in crocodilians because the character was inapplicable in the outgroup (Aves, birds); however, when fossil taxa representing more closely related outgroups were included, a clear polarity was established, suggesting Alligatoroidea (alligators) have the plesiomorphic condition. This occurred because the nearest outgroup (Aves, coded as inapplicable for 4296 of the morphological birds) for the extant erocodilians alone was characters present within the extant erocodilians, and thus the fossil taxa proved more relevant in analyzing character evolution within the least inclusive clade 2003). also help to establish a single including all extant crocodilians (Gatesy et al., Fossil taxa may preferred. optimization sequence of character states that have multiple equally parsimonious optimizations when only extant taxa are considered, reducing om ambiguity byincreasing taxon sampling density. For an example from crocodylians of how a suite of characters was differentially optimized (in terms of gains and losses) when fossil taxa were included or excluded from phylogenetic analysis, see Gatesy et al. 200: Finally, fossil taxa may add entirely novel charac- — ters or variations within characters to an analysis, providing information that we would have been completely ignorant of were they not included. For instance, inclusion of fossil penguins in an analysis of morphological and molecular sequence data for all Gs extant penguins by Clarke et al. (2007) demonstrated is that the ancestral beak morphology for penguins "[a|n elongate, powerfully constructed beak unknown (Clarke et al., 2007: 11550). in extant penguins” TIMING OF CHARACTER EVOLUTION Another, perhaps presently underexploited, aspect of character mapping in the context of a phylogeny is bracket the timing of character transformations. While determining a maximum limit on the time of appearance of a given synapomorphy by reference to as it is wrapped up the fossil record is controversial with determining ancestry (Hermsen & Hendricks, 2007) and is impeded by the fact that the true time of the first appearance of a fossil taxon (and thus the characters it bears) cannot be confirmed simply fossil record— through a literal reading of the e determining the minimum age of a synapomorphy is straightforward. It has long been recognized that the minimum age of a clade can be estimated by reference the most recent 1966): to the oldest of the descendants of common ancestor of that clade (e.g.. Hennig. Annals of the Missouri Botanical Garden this technique for estimating the minimum age of a clade has been exploited using various permutations of the same basic idea, such as comparison of the age its sister 1993), 2004: The minimum of a clade to that of 1992, mapping (Crepet et al., 20006). ancestor of a taxon or ghost lineage analysis (Norell, or minimum age node also see Hermsen € age of the most — Hendricks, recent. common clade is also the minimum age of the synapomorphies that define that clade, unless one or more of the synapomorphies is demonstrably older. Thus, mapping of characters onto a phylogeny and consideration of the minimum ages of the nodes on a cladogram can allow one to place minimum ages on synapomorphies as well. The bat phylogeny does not provide a particularly compelling example of this, since some of the characters of most interest that are not preserved in the fossil taxa (e.g.. laryngeal echolocation) cannot be mapped unambig- uously on the resultant trees, or too many topologies exist to make accurate age interpretations; however, in some of the trees from the unconstrained simultaneous analysis (analysis 3b, Fig. 1) and from the molecular scaffold analysis (analysis 4, Fig. 2), at least one Focene fossil bat taxon is nested within a clade of bats laryngeal echolocation. characterized in part by Therefore, minimum age node mapping performed 8 pping concurrently with character mapping on these trees EB suggests that laryngeal echolocation had evolved by the Eocene, an inference that has also been made analysis of the fossils 1998). through structural (e.g.. Simmons & Geisler, Hermsen et al. (2006) provided another example of minimum age node mapping overlain on character mapping in their simultaneous analysis of the — Saxifragales (introduced above i angiosperm clade S the section “Sequence of Character Evolution”). In that example, which occurs only in two of eight MPTs, minimum and maximum ages for the evolution of dorsifixed anthers in the clade including the extant families Pterostemonaceae and Iteaceae (Virginia willow family) were inferred on the basis o traits present in the fossil flower and fruit taxon Divisestylus Hermsen, Gandolfo, Nixon & Crepet (two species of which were included as terminals in the analysis) as well as fossil pollen that was not directly included as a terminal in the analysis. The synapo- morphy of dorsifixed anthers was mapped on the branch immediately descendant to the node where the species of Divisestylus attached. Because one species, Nixon Q Crepet, had been documented to have had basifixed D. brevistamineus Hermsen, Gandolfo, anthers (Hermsen et al., 2003) —suggesting that the dorsifixed condition had not yet evolved—it was used to dar a maximum age of about 90 million years » (Ma) on the aj pearance of dorsifixed anthers. timing of na £ Dispersed diporate pollen, a synapomorphy for Iteaceae (Choristylis Harv. and ftea L.), mapped at the node descendant to the node supported by dorsifixed anthers, provided a minimum age of about 50 Ma, the approximate time of first appearance of this pollen in the fossil record (Moss et al., 2005). Thus, it was suggested that a transition from basifixed to. dorsifixed anthers occurred somewhere between about 90 and 50 Ma. Hendricks (2007) methodology for applying minimum and maximum Hermsen | and formalized a — ages (relative or numerical) on the appearance of synapomorphies that are mapped on a cladogram that includes fossil taxa as terminals. While methods for inferring the timing of first appearance of synapomor- phies with reference to fossil taxa included directly in — cladistic analyses have not often been explicitly employed by overlaying minimum age node mapping and character mapping on a cladogram, the timing of first appearance of a trait has been inferred using inference of the existence similar logic. For instance, ee o given trait at a given time might be made when optimization suggests that tralt was present in a fossil taxon in which it cannot be directly observed see discussion below on inference, ambiguity, and hypothesis testing), thus providing a minimum age for the appearance of that synapomorphy. INFERENCE, AMBIGUITY, AND HYPOTHESIS TESTING One great. promise of phylogeneties as applied to the fossil record is the ability to test structural, functional, or behavioral hypotheses by inferring the presence or absence of particular structures, behav- iors, or functional features in fossil taxa through optimization of these attributes on phylogenetic trees; this is especially important when these attributes are unlikely to be preserved or are unpreservable in the fossil record. As pointed out by Bryant and Russell (1992) and Witmer (1995), a phylogenetic bracket (to use Witmer's term) is necessary in order to unambig- uously infer the presence or absence of traits in a fossil taxon in which these traits cannot be observed: in other words, taxa that are unambiguously known to ack taxon in — or possess a feature of interest must flank the unknown T order to which the trait is unambiguously infer that trait’s absence or presence in that taxon (e.g., to unambiguously optimize the character state transformation at the node where the fossil taxon attaches to the tree; see also discussion in O'Leary, 2001). O'Leary (2001), for instance, was able to infer whether basal Cetacea (whales) had hair O'Leary, 2001) using a simultaneous analysis of morphological and (extant cetaceans do not have hair: molecular sequence data for cetaceans and artiodac- Volume 95, Number 1 2008 Hermsen & Hendricks 93 Morphological Character Evolution tyls (even-toed hoofed mammals). Mapping of the a character for hair (with presence/absence states) onto indicated that the Hussain & Arif and Pakicetus Gingrich € Russell (extinct basa O two of 33 MPTs unambiguously fossil taxa ia Thewissen, whales) were hairless, since they were bracketed by extant taxa, the hippopotamids and extant cetaceans, r (O'Leary, 2001: figs. 6b, 6c). The extinct mele group Mesonychia was reconstructed which lack ha as either hairless or hairy depending on where the group was resolved in the MPTs Phylogenetics may be limited as a tool for inferring behavioral or functional (or even unpreserved struc- O'Leary (2001) provides several examples of this in her total tural) features of extinct taxa, however. evidence analysis of cetaceans and artiodactyls. Structural. evidence has been used to hypothesize that, for terrestrial. quadrupedal locomotion and of processing instance, basal whales were capable of sound under water (see citations in O'Leary, 2001). However, O'Leary (2001) was unable to unambigu- ously infer whether basal whales were capable of — these behaviors through optimization of behaviora characters on phylogenetic trees. She suggested that direct fossil evidence—when available—can provide the ultimate corroboration (or refutation) of a functional or behavioral hypothesis, and, where direct indirect evidence (for evidence is unavailable, instance, interpretation of an organisms habitat) could be utilized to address some questions that cannot be addressed successfully through character optimization. For instance, the hypothesis that early whales were quadrupedal was suggested by observa- tion of the breadth of the sacro-iliac joint in the extinct whale Ambulocetus and also inference about Pakicetus through character optimization on phylogenetic trees O’Leary, 2001: fig. locomotion was, however, ambiguously optimized for the breadth of the sacro-iliac joint in ma 6). The behavior of quadrupedal these taxa; thus, the hypothesis of terrestrial quad- rapedal locomotion in early ales cannot be corroborated through character optimization. In this case, direct fossil evidence could theoretically provide a solution: “if footprints of a pakicetid or ambulocetid (O'Leary, 2001: 501), hypothesis of quadrupedal locomotion in these early were found" the behavioral = whales could be corroborated. In the case o adaptation to hearing under water—where presence of a pachyostotic bulla in the ears of extinct basal whales suggests they were capable of hearing underwater—optimization is ambiguous and direct fossil evidence about hearing capabilities does not ikely character correlation argument only or an inference (and cannot) exist. Thus, “we are left with from design or environment based on the character- — istics of organisms that happen to be alive” (O'Leary, 2001: 501 For bats, extrapolation from structural features has been used to suggest that the extinct taxa Archaeo- nycteris, Hassianycteris, Palaeochiropteryx, and Tan- zanycteris were capable of laryngeal echolocation (e.g.. Simmons & Geisler, 1998; Gunnell et al., 2003). While mapping of the echolocation character on the topologies based on unconstrained combined data is equivocal for presence or absence of this behavior in these taxa in most instances (except Tanzanycteris, which can be inferred as capable of the behavior in eight of 16 trees: Fig. 3), hypothesizing that they were capable of laryngeal echolocation on the basis of structural features does not add extra steps to the MPTs. preferred approach to mapping the echolocation Such an interpretation instead imposes a character such that its presence would support the node immediately descendant to the node at which Megachiroptera attaches to the tree, although pres- ence of this behavior is not directly testable using the fossil record. Certainly, the examples from bats and whales expose one of the great weaknesses of phylogenetics for addressing aspects of evolution that go unrecorded in the fossil record: *While questions about stem taxa to major clades are often some of the most interesting, they can expose areas where it is functional and behavioral very difficult to test inferences in fossils with cladistic character data” (O’Leary, 2001: that can serve as phylogenetic brackets that allow for 501) due to the lack of extant taxa ey unambiguous optimization of intangible features. When structural (or other) evidence contravenes the optimization of a behavioral character on a phylogeny, the problem of functional and behavioral inference becomes even more complex. Simmons and Geisler 1998), for evidence suggests that /caronycteris was capable of have argued that structural ca instance, laryngeal echolocation (see similar discussion of echolocation in /caronycteris by Novacek [1987], Gunnell & Simmons [2005], and Simmons [2005a]). although mapping of the laryngeal echolocation character on the more recent most parsimonious (Gunnell & 2005) and the tree resulting from the topologies from | morphological data Simmons, combined analysis here (Fig. 3) suggest that it was not. When structural features and character mapping disagree, one must choose which is preferable, or whether the evidence is equivocal. In the bat example, therefore, one must decide whether /caronycteris should be coded as missing (?) for the presence of the matrix of Gunnell and Simmons, 2005), or eis: it should be laryngeal echolocation (as it -. assumed to be an echolocator a priori on the basis of other evidence. The latter extrapolatory approach 94 Annals of the Missouri Botanical Garden > O Outgroups New Green River Bat T Icaronycteris T Megachiroptera Archaeonycteris t Hassianycteris T Tanzanycteris T > Palaeochiropteryx t | Microchiroptera Outgroups New Green River Bat t Icaronycteris T Megachiroptera Archaeonycteris 1 Palaeochiropteryx t Tanzanycteris t | > Hassianycteris T | Microchiroptera Outgroups New Green River Bat T Icaronycteris T Megachiroptera Archaeonycteris T Palaeochiropteryx t | > Hassianycteris T Microchiroptera 1 Tanzanycteris t Hipposideros Rhinolophus Microchiroptera 2 Outgroups New Green River Bat T Icaronycteris T Megachiroptera Archaeonycteris T Palaeochiropteryx T Hassianycteris T Rhinolophoidea | Tanzanycteris T » Microchiroptera | Figure 3. Four scenarios (each representing four MPTs) of character evolution based on 16 MPTs re iu from analysis of combined morphological and molecular sequence arrows indicate a reversal of the positions of taxa in two of four MPTs. data sets without constraints (ana = ysis 3b; see Fig. 1). In each case, Fossil taxa are indicated by daggers us y "ue ually opti- mal positions for a transition from absence to presence of laryngeal echolocation are indicated by hash marks. Each scenario supports one origin iion of laryn 18e: al echoloc ation wil th no losse es. Mapping of this character does not O affirm the hypothesis that trchaeonycteris. Palaeochiropteryx, and Hassianycteris were able to echolocate. Tanzanycteris is Volume 95, Number 1 2008 Hermsen € Hendricks 95 Morphological Cm Evolution (Bryant & Russell, 1992), of course, overrides the potential utility of phylogenetics to help one infer the presence or absence of behavioral characteristics that cannot be directly observed; furthermore, the phylo- genetic topology and/or interpretation of character evolution may be impacted by choosing a character state on the basis of inferred rather than observed characteristics or behaviors. Mapping of co-varying structural characters that indicate the presence of a particular behavior may be one way to circumvent this 2001). However, when this reasoning is employed, it renders — apparent conundrum (e.g., Springer et al., questionable the coding of both structural and behavioral characters in the matrix on which the analysis is based—as they are in the Chiropteran but not the Cetacean data set—since doing so would violate the assumption of character independence. Another promising avenue for hypothesis testing is the potential utility of phylogenies to provide a framework on which the sequence and timing of this i mn character evolution can be juxtaposed; important, for instance, for corroborating or refuting proposed correlations between the evolution of structural features and extrinsic selective forces that occurred during geologic time. Hypothetical instances of this are given by Maddison and Maddison (2000) and Hermsen and Hendricks (2007). Both examples rely on the existence of a clade of extant taxa that is characterized by a unique adaptation that has been hypothesized to have arisen due to a selective force (e.g., appearance of a predator, climatic shifts, etc.). The extant clade is then analyzed with fossil taxa, which are found to nest within it. These taxa are older than the selective force that was hypothesized to have favored the adaptation that is a synapomorphy for the clade, so the cause-effect hypothesis is rejected. As in determining minimum and maximum possible ages for the appearance of synapomorphies, the utility of this approach is biased—it is much easier to reject a cause-effect relationship between an extrinsic fac- tor and the appearance of a synapomorphy if the minimum age of that synapomorphy is older than the selective pressure that supposedly favored its estab- Clarke et al. (2 ample of this in penguins, although it is not directly lishment. 007) provided a good ex- linked to one specific character. Molecular divergence dating has suggested that the origin of the Sphenisci- dae (the clade including all extant penguins) occurred in the Paleogene (~40 Ma), possibly “concomitant with the initiation of the circum-Antarctic current, initial onset of Cenozoic global cooling, or at the proposed extinction of giant penguins” (Clarke et al., 2007: 11549). Clarke et al. (2007) cast doubt on these correlations through simultaneous analysis of fossil and extant penguins, which suggested instead that Speniscidae arose in the Neogene, based on the stratigraphic occurrences of the fossil penguins included in their analysis. They noted that accommo- dating a Paleogene-Spheniscidae radiation given their inferred phylogeny would require a 164.1-334.2 million year ghost lineage (estimated using MSM* [Manhattan Stratigraphic Measure], Pol & Norell, 20 e extensive penguin fossil record known from between 40 Ma and 8 Ma includes no fossils that can be assigned to Spheniscidae. Despite this, they noted that their evidence is suggestive, not conclu- sive, since “stratigraphic data can only falsify a divergence date when a fossil discovery is older than (Clarke et al., 2007: 11549). estimated" CONCLUSIONS The challenge for icu and paleontologists i in coming years will using molecular sequence data with our growing knowledge of the fossil record. Molecular sequence data have certainly helped to clarify relationships among extant taxa and will likely continue to modify and improve our understanding of the tree of life. Although fossil taxa almost always lack sequence data, they represent extinct morphologies that may be informative to the overall history of life on earth, as well as evolution within particular extant and extinct groups of organisms. Several perceived obstacles to identifying the phylogenetic positions of fossil relative to extant taxa have been addressed above. Some of these purported problems are closely related to fossil taxa in particular (e.g., missing data), while others are related to morphological data in general (e.g., convergence and character weighting). These issues have been over- emphasized as obstacles to effectively including fossil taxa in phylogenetic analyses, despite significant empirical advances in our understanding of how well the simultaneous analysis (or supermatrix) approach Drs Md inferred as possessing laryngeal echolocation in eight of 16 trees (C, D). includes the extant genera Rhinopoma E. Geoffroy, Craseonycteris Hill, Megaderma E. Geoffro n scenario D, Rhinolophoidea includes Rhinolophus Lacépède, Microc a l contains all other ingroup microbat taxa. In In scenario C, Microchiroptera 2 oy, and Macroderma Miller and e Gray, Rhinopoma, Craseonycteris, Megaderma and Macroderma, and Microchiroptera represents the remainder of the microbat 96 Annals of the Missouri Botanical Garden often works for elucidating the phylogenetic context of biota. A arguments that fossil taxa should not be included fossil taxa relative lo the extant priori simultaneous analyses are insupportable—the suil- ability of fossil taxa for inelusion in these data sets needs to be tested empirically in the context of each data matrix and their behavior evaluated on a case-by- case basis (for instance, using pseudofossil analyses). The molecular scaffold approach has several weak- nesses as a method for integrating morphological and molecular sequence data in order to include fossil taxa in phylogenetic hypotheses. Perhaps the largest of these is that the methodology obscures secondary signals that may manifest themselves during simulta- neous analysis, that it minimizes the impact of fossil semi-strict and, when a axa on tree topology. molecular scaffold is employed, it provides a contradictory perspective on the reliability of molec- ular sequence and morphological data. A final paleobotany in particular, since the subject of this note should be added in reference to symposium is paleobotany in the post-genomics era. Most examples of simultaneous (and scaffold) analyses including fossil taxa in this paper come from studies of vertebrates (in addition to studies cited above, see, for instance, Brochu, 1997; Shaffer et al. 1997: O'Leary, 1999; Gao & Shubin, 2001; O'Leary et al., 2004; Asher et al, 2005b; Démére et al. 2005; Geisler & Uhen, 2005: 2006: Ksepka 2006). This is because vertebrate paleontolo- gists and zoologists have been leaders in Horovitz et al.. et al., the field, exploring methods to integrate fossil and extant taxa in greater numbers than those who study plants (in addition to the studies cited above, see Sun et al., 2002; 2003; Gandolfo et al., 2004: Crepet et al.. 2005) or inverte- cited above, se 1995; Arango H Hermsen et a 2005; Xiang et a brates (in addition to the studies Littlewood & Smith, 1995; Smith et al., — & Wheeler, 2007; Cardinal € Packer, 2007). Some of this may de the result of the large number of osteological characters thal can be scored for vertebrates, perhaps making them more amenable to such analyses than other groups. Paleontologists and biologists who study vertebrates are also leading the way in the study of character evolution on phylogenies including extant and fossil taxa (see examples discussed above). While botanists have certainly kept pace in the arena of collecting sequence data and building phylogenies of extant organisms from those data, more work needs to be done to integrate. our evolving view of the relationships of extant groups to knowledge of morphology. Only provide one another with our development, and the fossil record. integration — of these disparate data types wil us with a holistic view of plant evolution. ADDENDUM Following the c rcli of the analyses document- ed in this paper, the “Green River bat” was formally published as Onychonycterts ph Simmons. Sey- Habersetzer & Gunnell, 21 formal description of this bat, Sante et: Along with the 2008) published new morphological and molecular sent ld mour, phylogenetic analyses. These analyses were based on an updated morphological matrix and a molecular scaffold derived from a more recent molecular analysis than the Teeling et al. (2005) data used in our study. Notably. structural evidence from the newly published taxon is interpreted as confirming our inference that Onychonycteris was incapable of echolocation. Literature. Cited J. N. A. Hooper & er. ri tic relationships of the Alvarez, B., M. D. Crisp, F. \ 2000. I family Axinellidae (Porifera: De M ede using uum : 169- (ue ny of the R. W. M. Van Soest. Y heeler. 2007. sea E (Arthropoda, Pycnogonida) based on direct > logical a molecular data. Zool. Script: \rango. C & W.C. m of six loci and morphology. Cladistics 23: IRS- ae H.” E M. B. Fe 1997. the solus and evolution of bats. Trends Ecol. anton. Flight and ec B ‘ation in 12: Phylogeny a n. A: of S. W. 1992. plant- rs interactions. baste T. BioScience As a t. J. & M. Hofreiter, 2006, Tenrec Pisteen ny um the MR extraction of nuclear DNA. Syst. Biol. 181-194. . M. J. Novacek & J. H. Geisler. 2003. Relationships of endemic African mammals and their fossil relatives ased on PA «al x molecular evidence. J. Mammalian E vol. 31-19 . Horovi E a hez- Villagra. 2004. is of cee mammal rela- 3 10-250. First N co => => vol, 33: 24025 . J. Emry € M. McKenna. 2005a. New mates mes (Mammalia. L ipotyphla) ^ s importance of ios M [de gcn al of = = (missing) D! — nees in systema c paleontology. J. 92: Vertebrate Paleontol. 911-923. J. Meng. J. n. Wine M. €. McKenna. G. W. Housié r, D. Dashzev . Novacek. 2005b. Stem Lagomorpha and he a a Glires. Science 307: 1001-1094. Baker, R. character iade Pe Ps 1997. Multiple phylogeny of Hawaiian H. & R. Salle. sources of drosophilids. Syst. Biol. 54-673. Yu & H. bie : relative ae of molecular and ee ‘al characters in dec 2 Assessing the simultaneous analysis trees. Mc Phylogen. Evol. x: 274; 6. Barrett, M.. M. J. e A E. consensus. Syst. Zool. : 4.9: Benton. M. 1990. Pd rhynehosaur (Reptilia, d ues the Middle ‘Triassic of England. E Trans.. 328: 213—300. Boisserie, J.-R., F. Lihoreau & a iin 2005. The position of Hippopot: ded within Cetartiodactyla. Natl. Acad. Sci. U.S.A. 102: 1537-1541. Sober. 1991. Againsl 3 iynchosaurus. a species of p roc Volume 95, Number 1 2008 Boucher, L. D., S. R. Manchester & W. S. Judd. 2003. An extinct genus of Salicaceae based on twigs with attached flowers. fruits, and f foliage from the Eocene Formation of Utah and Colorado, USA. Amer. J. Bot. 9 1389-1399, CO Brochu, C. A. 1997. Morphology, fossils, divergence timing, and the phe netic relationships of Gavialis. Syst. Biol. 46: 479-5 Brooks, D. R. P D. A. McLennan. 1991, Behavior: A Biology. i Phylogeny, Ecology, and esearch Program in Comparative d of Chicago Press, Chicago. Bruneau, A. 1997. Evolution and homology of bird D ud tion inició in Erythrina (Leguminosae). Amer. J. Bot 84: 54-7 A Bryant, H. N, & A. P. Russell. 1992, The role of phylogenetic analysis in the inference of unpreserved attributes of Philos. Ser. B. 337: 405—418. M. R. Badgett, H. A. Wie voa J. P. Huelsenbeck, A . A. Gulati, C. Ho & I. J. Molineux. 1997. enion convergent evolution in a virus. Genetics 147: 1497-15 Cardinal, S. 8 L. extinct e Trans., ae ^" er. 2007. ' corbiculate Apinae Phylogenetic analysis of based on morphology of the sting apparatus (Hymenoptera: Apidae). Cladisties 23: 99-118 Jarke. J. A., D. T. Ksepka, M. Stucchi, M. Urbina. N. Giannini, S. Bertelli, Y. Narváez € C. A. Bovd. e Paleogene equatorial penguins challenge le proposed relationship between biogeography, diversity, and Ce zoic Natl. Acad. Sei. U.S.A 545 Gabhett eno- An change. Proc. 104 50. E Wilkinson & A. Willis. 2007. in d as hard as living taxa in parsimony analyses of morphology. Syst. Biol. 56: 160. J. A. 1988. hypotheses. Cladistics 4: Crepet, . Nixon 3 7 F ossils Coddington, : ilii tests of adaptational A. Gandolfo. 2004. p The age angiosperm clades based on mesoloea = macrofossil evidence from 91: 1666-1682. Fossil evidence ed i logeny: e of major Cretaceous de pos Amer. J. Bot. ———. 2005. An extinct enn taxon, + Jesevants salycanthuudis from the : ate ceous e ^w Jersey. Amer. J. Bot. 92: o 485. Cuevas, J., S. F. Elena & A. Moya. 2002. ms basis of dien convergence in epe rimental eiua of RNA vir ; s 162: : 542. de Queiroz, A n & T o un - un ~ ` A 7 )07. The Ps rmatrix approach : Trends Ecol. Evol. 34—41. Characters, n. 's and evolutionary scenarios. Cladistics 9: 427—432. Berta & M. R. McGowe an. 2005. and evolutionary history of fossil 2 modern balaenopteroid Mysticetes. J. Mammalian Evol. 2: 99-143. Desutter-Grandcolas, L., F. Robillard & J. 10 and pa = lism: Is a new life ahead of old concepts? Cladistics 21: 51-61. Dilcher. D. L.. F. W. Potter Jr. € W. L. Crepet. 1976. Investigations of angiosperms from the Eocene i 2d rica: Juglandaceous winged fruits. Amer. J. . 63: 544. Te M howe. 1989. reco nstuehon: Eernisse, D. J. € versus total Lo sy: Dele "OMA tests of Démére, T. A. taxonomic Legendre, Murienne. 2005. A. G. Kluge & T. of fossils in phylogeny 31—460. le Gauthier, Rev. 2 Sys . G. Kluge. ev ip nee, e Disspamó congruence and amniote phylogeny inferred from fossils, moleculen, and morphology. Molec. Biol. Evol. 10: 1170-119 reen River Hermsen & Hendricks Morphological Character Evolution 97 D. S. Jacobs & C. hylogenetic Eick, G. ». A. Matthee. 2005. A nuclear DNA the evolution of echolocation and historic al d of extant. bats (C hiroy ra). Molec. Biol. Evol. 22: 1869-1886. 1. & J. Cracraft. 1980. evolutionary process. on perspective T s netic E and Eldredge, N th Univ. Pre New Columbia ‘ork. Endress, P. K. & E. M. 1991. Archamamelis, hamame- lidale ean flowers from the Upper Cretaceous of Sweden. Pl. Syst. . 175: 101-114. Farris, Ll s. es » Methods for computing Wagner trees, Syst. Zool. 19: 83-92, 1983. Pp. 1-36 in Advances in C od T3 New York. 989. The re ele ntion index and rescaled consistency Cladisties 5: d 419. Küllersjó, A. . Klug e & C. Bult. 1995a. Testing sente 'ance of incongruence. "t ladistics 10: 315-319. S Friis. The a basis of phyloge netic analysis. N. Aatnick € V. / “un Vol. (editors), -Columbia T Univ. Press, index. ————, ———— & ——.. 1995b. C rais a ines tés for incongruence. Sy st. Biol. 44: 570-57 . 1981. trees from DNA a nc- es: À maximum likelihood approach. J. Molec. Evol. —376 308 Felsenstein, , Evolutionary 1985. Confidence limits on phylogenies: approac A using the bootstrap. Evolution 39: 783-791. Fitch, W. M. 1971. Toward defining the course of evolution: a Shane for a specified tree topology. Syst. Zool. 20: 406-416 An Forey L & Fortey. 2001. Fossils in the reconstruction of ton Pp. 515-519 in D. E. G. Bri P. oii a. Palaeobiology IL. Blk Scie ence, alden, o UN es T. Grant & T. A. Phy eos of the liz Inc., Frost, Titus. 2001. ard genus sud o Tropidurinae ): Direct optimization, des rip- and sensitivity analysis of congruence between molec e res and morphology. Molec. Phylogen. Evol. 21: 352-371. T ropid uridae: tive efficiency, Nixc Crepet. 2004. Creces Hows TS ( P. LM aceae anda implic alions for complex insect entrapment pollination mec sp an in early Angiosperms. Proc. Natl. Acad. Sci. 101: Gao, K.- . Shubin. 2001. Late Jurassic salamanders from ioe Tn ie hina. N: ipd 110: 574-577. J., P. O'Grady & R. H. Ba jd 1999. Corroboration data sets in Bale analysis: Hidden sup P. for legs netic rel: Me among higher level art dactyl t; Cladisties 15 ps 313. . G. Amato, M. Nass i DeSalle & C. Hay 2003. Combined support for Ms d o crocodylians. Syst. Biol. — ——, R. H. Baker & C. arguments s the Gatesy, among axa. yashi. taxic atavism in —422. pul consistencies in 200% e th: Hayashi. neue 2E SUpormatriees ersus supertrees of Crocodylia. Syst. Biol. 53: 342—355. CDM J.. A. G. Kluge P T. Rowe 1988. Amniote phy- logeny and the importance of fossils. Cladistics 4: 105-2 J. H. & M. D. Uhen. 2005. relationships of extinct Cetartiodac tvls: Res taneous analyses of 1 Geisler, Phylogenetic sults of simul- nolecular, vorpal al, and strati- graphic pr J. Mammalian á vo 15-160. E ). Edgecombe. Y NS Tit ler & C. Babbitt. 2002 p con ny and systematic position of Opiliones: A combined alysis of chelicerate rel emo using morphological and Tos pe data. Cladistics 18 Sytsma. 1997, Homoplasy in molecular Givnish. T. J. & K.J. VS. morphological data: The likelihood of correct phil genetic analysis. Pp. 55-101 in T. J. Givnish & K. E Sytsma (editors). Molecular Bsoldtin He Adaptive iara Cambridge l ress, New York Glenn J. Hansen, | Sørensen, F. Ronquist. J. P. Hue i s k & E. Wille a y. 2004. Bayesian inference of * metazoan phylogeny: A combined molecular and ee approac h. Curr. - Biol. 14: 1644-1649. Goloboff. Farris & . Nixon. 2003a. TNT: Tree REN Using New s Published by the \rge ntina. M. Kiillersjó. B. Oxelman, M. J. . Tucumán. Goloboff. p A. J. S. Fan OA Ramirez & C. Szumik. 2003). improve ments to resampling measures of support. Cladistics 19: 324—332. Grande, L. & W. E. Bemis. 1908. A comprehensive phylogenetic study of. amiid Te (misa) based on comparative skeletal anatomy. An empirical search for interconnec $5 «d patterns of natural history. Soc. Vertebrate 1-09 rice em. 4: 190. Gunnell. : N. B. Simmons. 2005 e E nce and the origin of bats. J. Mammali an Evol. 200-246. —————, B. F. Jacobs, P. Cmn E T: mad E. Kowals ki, C. P. Msuya. F. A. Mizambwa, T. Harrison, J. Habersetzer & G. Stroch. 2003. Oldest place Pu mammal \frica: ‘arly evolution of sophisticated echo- from Sub-Saharan Eocene microbat from Tanza- nia—evidence for « Palacontologia Electronica 5(2). 1994. morphology in avian evolution: The case P a "pe we ani- birds. Proc. Natl. A A. 91: 9861-9805. & 1996. Re: KA morphology in amniote phylogeny. Molec. Phylogen. Evol. 6: 312-319 ———— & — 1997 location. Hedges, S. B. & €. G. Sibley. Molecules vs. form” & L. R. Maxson. cules and Re: Molecules. morphology, and phylogeny: A jols to Hedges and Maxson. Molec. Phylogen. Evol. 7: 391—395. = K D TA & L. R. Mason: 1990. Tetrapod RNA p "s ny inferred. from E and 285 ribosomal review the evidence for amniote A 607-033. Phylogenetic Sean Univ. of Illinois sequences and a relationships. Molec. Biol. Hennig. W. 1966. Press, Urbana. J. & J. R. Hendricks. Sene 21: 403-405 -——— A method for constraming E age of origination of derive He characters. Cladis —179. . M. A. K. C. Nixon & W. a kde go Dioses gen. nov. (aff. Iteaceae), a fossil saxifrage from USA. Amer. J. Bot. 90: The Hermsen, 2006. Spotlight: hierare i df time. EN — & andolfo, the Late Cretaceous of New Jersey. 27: 20 73-1388. — 4 K. €. Nixon & W. L. Crepet. 2006. The impact of extineL taxa on unde ellas the early evolution of Angiosperm clades n P incorporating fossil reproductive structures "m ixifragales. Pl. Syst. Evol. 260: 141—169. Horovitz. I. 1999a. A phylogenetic study of living and fossil plat "o s. Amer. Mus. Novit. 3209: 1—40. - 29b. A report on “One Day Symposium on Nume e C ladistie s." Cladi: p 15 Hr a M. R. Sánchez Villagra, T. Ma & 0. A. Aguilera. 2006. The fossil record of o9 dins kae Mones 1980 (Mammalia, Rodentia) M Urumaco (Late Miocene, Venezuela), with pu of its phylogenetic relationships. J. Syst. Palacontol. 4: 293—: 06 C. Teeling. 2006. Qs Evol. 21: 149-156. Jones. a & E. echolocation in bats. Trends Ecol. evolution of Annals of the Missouri Botanical Garden . Hill. 1999. The RN affinities based on Jordan, G. J. & R. of Nothofagus A e AO leal combined molecular and morphological data. Int. J. PL [ossi S 160: 1177-1188. — & M. K. Macphail. 2003. V Middle-Late Eocene inflorescence of Caryophyllaceae from Tasmania, Austra- lia. Amer. J. Bot. 90: 761 Kay. R. F. & M. A. Cozzuol. 2006. New p monke NS from the Solimões dieu (late Miocene. Acre State, Brazil). J. Human Evol. Kearney. M. 2002. Sod n aken assumptions and conclusions. —108. : 61 13-080. e taxa. missing data, and ed Biol. St: -381. &, A Clark. 2003. Problems due ^ missing data in phyloge netic analyses ine i fossils: A critical review. tebrate Paleontol. 23: 263-274. Th e G. 1989. A concern for evidence and phylogenetic Espai sis of re eee among Epicrates (Boidae. Se rpe ntes). Syst. Zool. 1-2 & Farris. 1960. Quantitative ee ties and the evolution of anurans. Syst. Zool. 18: 1-32 —— & A. J. Wolf. 1995. C lei What's in a word? Cladistics 9: 183—199, Ksepka. D. T. 5. Bertelli & N. P. Giannini. 2000. The phylogeny of the living and fossil Sphenisciformes (penguins). Cladistics 22: 412441. = Lambkin. C. L.. M. S. Y. Lee, S. L. Winterton & 2002. pre Bremer support. and aule 30-444. Rë Molecules. trees. Cladistics 18: Lee, M. S. Y. pts re d poA morphology, and phylogeny: A to Hedges and Maxson. Molec. Phvlogen. Evol. 7: 391—395. — 2005. Mole ular evidence and marine snake origins. Biol. Lett. ]: 227-230. Lewis. P. 0. 2001. V likelihood approach p from d rete morphological character data. Syst. Biol. 50: 913 Littlewoo: D. T 1 3 Smith. 1995. momke al and molecular phylogeny for sea urchins " de "us Trans., Ser. B. 347: stimating V combined idea: Echinodermata). Philos 31-234. Log nd M. & A. Bruneau. 1997. Circularity and indepen- ls of ecologic al hypothe ses. dence in phylogenetic les Cladisties la 145-151. Maddison, D. R. & W. P. Maddison. 2000. MacClade 4. ersion 4.02. Sinauer Associates. Ine., Sunderland, Massachusetts. Magallón. S. 2007. From fossils to molecules: Phy logeny and the core eudiet flor: d em un in Hamamelidoideac (Hamame Tidac ‘eae, Saxifragales). Syst. Bot. 32: 317—341 Manos, P. S p 5. ae * E. rus S. R. S.-H. Oh. ( . Dileher & D. E. p hylogeny i extant a e sil Juglanc T eae inferred from \ Stone, 2 the integration of molecular and morphological data sets. Syst. Biol. 56: 412-430, Mattern. M. Y. & D. A. MeLennan. 2004. Total evidence phylogeny of Gasterostek lae: C Qu. n Sd C vu mor- phological and 7 havioral data. Jadis Lies Montgelard, C., M. Catze e no Douzery. 1997. Phylogenetic. re [mde hips of Artiodactyls and Cetaceans as deduced from the comparison of cytochrome b and 125 dec. Biol. Evol. 14: rRNA iloc hondrial sequences. Me Greenwood € S. B. Archibald. 2005. nity dynamies of the Moss. Ps T DR.: Regional and local v Eocene Okanagan Highlands (British C columbia x astInEión Earth Sci. 42: vege talion comm State) from palynology. Canad. J. Volume 95, Number 1 2008 Hermsen € Hendricks 99 Morphological Character Evolution Müller. J. & R. R. Reisz. 2006. Eureptiles: Comparing parsimony and Bayesian approac h- The phylogeny of early es in n the investigation of a basal fossil clade. Syst. Biol. J., E. Eizirik, S. O'Brien, O. Madsen. M. Seally, D. I Christophe, E. Recline, O. A. Ryder, M. J. Stanhope, W. W. de Jong & M. 2001. Resolution of the early end mammal radiation using . Science 294: 2348-235 | =] m = y - Springer. Bayesian e logenetics Nixon, K. C. paleobotany: E rlightenment and uncertainty. Rev. Palaco- bot. Palynol. 90: -31: 1999- 2 E inClada: Ithaca, w York. & ( 96. Paleobotany in eladisties and cladistics in Published by the author. .D. Whe eler. 1992. Extinction and the origin of species. = 119-143 in M. J. Novacek € Q. D. Wheeler (editors). Extinction pe Phylogeny. Columbia Univ. Press. New York. — —— & J. M. 4 ire nter. 1996. On simultaneous analysis. C hei s 12: 221—241. Norell, V . 1992, Taxic origin and temporal diversity: The effect dd. Pp. 89-118 in M. J. Donoghue & Q. D. Wheeler (edito Ee Extinction and Phylogeny. Columbia Univ. Press, New 993. Tree hased approac hes to understanding history: s nts on ranks, rules, and the quality of the fossil rec us jer, J. Sei. 29: : 407417 Novacek, M. J. Auditory pana and a ffinitie s of the Eocene ni Icaronycteris and Palio de (Micro- us. Novit. 2877: |— ng data. and the e o ra, incertae sedis). Amer. Fossils. topologies. jüghr j A T1000 of eutherian m: E Syst. Biol. bl: Pe S 2b. Fossils as critical data for phylogeny. Pp. 46— i Novacek & Q. D. and a Columbia Univ. a ona J. A., F. Ronquist, J. P. ves- vee 200 nt W data. Syst. Biol. O'Leary, M. 1999 extinct and extant taxa and the Cetac ceo —330. 88 in M. ‘heeler (editors), Extinction Press, New York. Huelsenbeck & J. I bus sian phylogenic : 47-67. analysis o "arsimony ande of total evidence from question (Mammalia, Ungulata). Cladistics 15 2001. The ; position di e elaceans: Further combined ds analysis, phylogenetic comparisons with the strati record and a character graphic discussion of optimization. Amer. —— & J. H. Geisler. . The position of Cetacea within Mammalia: AR tic m of Dr pe data Tom extinct and extant taxa. Syst. Biol. —490 . Gatesy & M. aen Are lle FT ntal dates re sally at odds with the molecular De wid e s al Novacek. e mas nce for whale phylogeny (re) reexamined. Syst. Bi 52: 853-804. Allard, M. J. Novacek, J. Meng & J. M. Buildins the laa sector of the tree of life, === Galesy. 2004. combining ditis rent data and times for placental mammals. Pp. 1 a discussion of divergence 90-516 in J. Cracraft & Assembling the Tree of Life. New Y ork. Omland. K. E. 3s 394. Chai a morphological E ny for dabbling iol. 43: 369— pep (edite racter congruence between molecular and : ducks (Anas). Syst. Patterson, C. 1981. relationships. Ann. Rev. ae in dete rmining Ecol. Syst. Significance ol evolutionary 19 e Donoghue. 2007. 75-809. Peters . J.. R. E. Summons & P. C. J. Mole: c i ar paleobiology. Palacontology 50: 7 ————, J. B. Lyons, a Si m wak, C. & M. A. ie Pele times with a molec Ep e a k. Proc. Natl. 101: 6536-6541. I. Takacs. M. J. Wargo bons melazoan rl i. U.S./ Acad. S Pol. D. & M. A. Norell. 2001. Comments on he Manhattan stratigraphic measure. Cladistics 17: 285 Protas, M. E., C. Hersey, D. Kochanek, Y. e In Wilkens, W. R. Jeffery. L. I. Zon, R. Borows sky & L T abin. 2006. Genetic a analysis of cavefish reveals molec "e convergence 07 in the e vilius of albinism. Nat. Genet. B Pryer, K. Schneider, A. R. Smith, R. Cranfill P. G. Wolf, ar m & 5. D. Sipes. 2001. Horsetails and ferns are a neral tic a and the closest living relatives to seed plants. Ne ture | 109: ws 622 toca, A. L., G. Bar- Gal, Fiz zirik, . Helgen, R. Maria. M.S. Springer, 5 ail 2x n & W. i ae 2004. Mesozoic origin for West Indian i insectivores. Nature 429: 649—051. Rothwell, Y ] Nixon. 2006. How does the inclusion of fossil data change our conc d about the oe history of e Ms ere 's? Int. J. Pl. Sci. 167: diagnosis, and origin of a 64. . Horovitz. 2003. The anatomy of the ie s largest extinct rodent. Science 301: 1708-1710. Sanderson, M. J. & M. J. Donoghue. 1989. ERISTUS of VaL OR 1 levels of me lasy. Evolution 43: 1781— Se hae ffer. B.. M. K. Hecht & N. E d 1972. Phylogeny and lel. Pp. 31-46 in T. Dobe hansky, M. K. Hecht & E nina Biology. Vol. 6. lens Century-Crofts, New Definition, NE D us ne em ii Sánchez- Villagra. M. R. ee. a Steere (editors). Schneider. H., E. Schuettpelz, K. M. . R. Cranfill, Magallón P R. Lupia. 2004. Ferns pH rsified in i: shac a S a x d rms. | Natur » 428: 553-557 Shaffer, . P. Meylan & M. L. McKnight. 1997. Tests of turtle ae ny: Molecular, sis ae s and paleon- tological appyoae thes. Syst. Biol. 46: 235-2 2005a. C Krople ra. Pp. EP in K. D. Archibald (editors). The Mammals: pode and Relationships of the Major Extant Clades. The Johns Hopkins Unis gs s. Baltimore. 2005b. Order ie Ph 312- 520 in D. E. — Wils ¡lena i D. M. Reeder (editors), Mammal Species of the Work T ‘axonomic and Geog ede Reference, 3rd ed., Vol. “The Johns Hopkins Univ. Press, Baltimore. Simmons, Ww ose & Rise of Placental — 5c. E volution—An Eocene E bang for bats. Science e 307: 5 527-528. & J. H. Geisler. 1998. Icaronycteris, chiropteryx to extant bat lineages, with comments on the evolution of Microchiroptera. Bull. Amer. Mus. & Phylogenetic aa of |Z Archaeonycteris, Hassianycteris, and Palaeo- echolocation and foraging strategies in RI Nat. Hist. 235: 1-182. ——— . 2002. Sensitivity analysis of different methods of coding taxonomic polymorphism: An example from higher-le - bat phylogeny. Cladistics 18: 571-584. ——, K. L. Seymour, J. Habersetzer & G. F. Gunnell. 2008. Ponutive Early Eocene bat from Wyoming and the evolution of flight and echolocation. Nature 451: 818-822. Smith, A. B. 1998. What does palaeontology contribute to systematics in a molecular world? Molec. Phylogen. Evol. 9: 437-4 H. . D. T. J. Littlewood & G. )atterns i evolution: . 1995. Comparing aa “ai e stages and Philos. arval : ribosomal I of post-Palaeozoie ec hinoids : Tra 18 , Ser. B. 349: 11-18 100 Annals of the Missouri Botanical Garden g. O. Madsen, M. J. Stanhope & dd d fossil and molecular data Acad. Sei; U.S.A. a r, M. ©. C. Teel W. W. de a 2001. reconstruet bat echolocation. Proc. Natl. 98: 6241-6246. AA J. Stanhope, O. Madsen & W. W. de Jong. 2004. Molec me s consolidate the placental mi m tree. Trends Evol. 19: 430—438. —————, A. Burk-Herrick, R. Meredith, E. Peeling, S. J. & W. J. => = Ecol. Eizirik, E. Murphy. 2007. The adequacy of morphology for reconstructing the early history of placental mammals. Syst. Biol. 56: 673-068 Sun, G., Q. Ji, D. L. Dileher, S. Zheng, K. C. Nixon & X. Wang. 2002. Archaefructaceae, a new basal angiosperm family nce 296: 899—9( Swofford, a 2003. PAUP? lena Analysis Using uini, (and Other Methods), Vers. 4.0. Sinauer Associates, Inc., Sunderland, Massachusetts. N 1990. Phylogeny reconstruction. D. M. Hillis € C. Moritz "e Sunder- —— & G. J. Pp. 411-501 in (editors), Molecular Systematics. Sinauer ler Associates, land, Massachusetts. Teeling. E. C., M. S. Springer, Madsen, P. Bates O'Brien € W. J. a DR s i molec al ar ¿llosa ny are bats illuminates biogeography and the fossil record Science 307: 580-584. Wahlberg. N., M. F. Braby, A. V. Z. Brower, R. de Jc -M. Lee, S. Nylin, N. E. Pierce, F. A. H. Sperling, R Vila, A. D. Warren & E. Zakharov. 2005. Synergistic effects of combining morphological and molecular data in resolving the phylogeny of bu bun and skippers. Proc. Roy. Soc. o Sci. Ser. B. 27 577-1580. Wenzel, J. S M. E. Siddall. o do Cladisties 15: 51- M le e le T, ES . 1992. Extine tion, d and molecular phylogenetics. Pp. 205-215 I. J. Donoghue & Q. D Wheeler. (editors), binden p Phylogeny. Columbia Univ. Press, New York 1996. Optimization alignment: The end of multi- No E. sequence alignment in phylogenetics? Cladistics 2003. Implied alignment: A ipd ee based de sequence alig gnment me ss and its use in cladogram search. Cladistics 19: 201—208. . D. Gladstein & J. De T aet. 1996-2003. POY. phylogeny reconstruction via op Ed of DNA and other data version 3.0 (Current version 3.0.11). American Museum of Natural History, New G. Giribet & G. D. systemalics, the NUR E: study 2004. of genomic, Edgec aie piis anaton cal. and paleontological information. Pp. 281-205 in 1 a rac raft & M. J. Donoghue (editors), Assembling the Tree Oxford U Intv. Press, New Yor Wiens, " J. 20023a. liasa data, 3 phylogenetic accuracy. Syst. incomple te tan and NE t6 ANO 528-5: Y t . 2003b. Incomplete taxa, inc omplete alpibus and phylogenetic accuracy: A there a missing data problem? J. Verte gus Paleontol. 23: 207—310. 1 ea & D. M. Hillis. 2003. When are phy m ne ue nd s misled by convergence zd A case study 1 Biol. 52: 501—514. W Ikinson, » 1995 ». Coping with abundant missing entries ‘ave salamanders. Syst. in phylogenetic inference using parsimony. Syst. | iol. 44: 501—5 i Witmer, aaa of reconstructing soft tissi 33 in . Verte brate Pale ontology. Cambri idge Univ. Press, New $ Wortley, A. H. & R. V 2006. The effe of combining molecular and morphologie ‘al data in publishe d ,, M. 1995. The extant E netic bracket and x s in fossils. Pp. 19- Thomason (e: al Morphology. in Scotland. phylogenetic pud 's. Syst. ies 55: 677—685. Xiang, Q.-Y. J.. 5. R. Mane hester, D. T. Thomas, W. Zhang & C. Fan. 2005. Phy logeny, B. ography, and molecular Tracking Tertiary plant migration. Evolution 59: 1685-1700. es H. H. 2002. C 1. Brain Behavior Evol. 59: 250-2( Zhang, J. & 5. Kumar. 1997. parallel evolution at the amino acid sequence level. Molec Biol. Evol. 14; 527-5: Zwickl, D. J. & D. M. Hillis. * Roc greatly reduces phylogenetic error. Syst. Biol. 51: 588-599. dating of cornelian cherries (Cornus, Cornaceae): onverge je e Magus on the molecular | of convergent and 2002. Here en taxon poe THE WHOLE AND THE PARTS: RELATIONSHIPS BETWEEN FLORAL ARCHITECTURE AND FLORAL ORGAN SHAPE, AND THEIR REPERCUSSIONS ON THE INTERPRETATION OF FRAGMENTARY FLORAL FOSSILS' Peter K. Endress? ABSTRACT Floral architecture and floral organ shape are interrelated to some extent as can be seen in the diversity of extant angiosperm groups. I e shap a reconstruction of the architecture of a flower. This i is paniy a review and par of view on organ-archi ral topic itecture interrelationships. Sev shape, exemplified by euneate organs, especially stamens; (2) condition for valvate anther dehiscence: (3) e of fragmentary fossil material, for the w points : (1) autonomous and cise s single organs, may therefore give hints È rtly ode: original materi s are illustrated with examples and shape of reduced organs that have decreased in size and lost their original function; (4) long hairs as filling material of irregular s spaces; (5) architectural conditions for the presence of roo ovules; ) structural differences between exposed and covered organ parts in bud; and (7) sepal aestivation and petal elaboration. Key words: Autonomous shape, cuneate, floral architecture, imprinted shape, reduced organs. There are relationships between the whole and the parts in floral structure. The ensemble of the floral organs constitutes the floral architecture, and, vice versa, the architecture of the whole flower or even inflorescence may have repercussions on the shape of individual organs. Under certain conditions, organ shape is conspicuously affected by the floral archi- tecture. Thus, the shape of isolated organs may tell something about the architecture of flowers. Floral fossils are often not complete flowers, but are instead fragments and sometimes single organs. The question then arises on how the whole flower may have looked. Can certain traits of a flower be inferred from single organs, even if the systematic affinity is unknown? This is possible to some extent. The influence of the floral architecture on the floral parts may be direct, shaped individually by neigh- boring organs during development. It may also be more indirect, shaped historically by evolutionary constraints on the best fit of the components of the floral architecture (Endress, 1975, 1994a). The first is seen in pressure marks on organs made by adjacent organs. The second is seen in harmoniously fitting floral) bud. Both are related to eco- p parts within a nomic use of space and protective function in the bud. In flowers at anthesis, the floral organs are also more or less highly synorganized with each other, which is related to the floral function in reproductive biology (Endress, 1994a, 2006). The of floral fossils are important aspects in the integration of reconstruction. and interpretation fossils into phylogenetic and evolutionary research. Information extracted from floral fossils is continu- ously being improved by technical advances in prep- aration and reconstruction (Schénenberger, 2005; De 2006; , 2007; et al., 2007) and by the ela of fossils in Vore et al., Friis et al. von Balthazar morphological or combined molecular and morpho- logical character matrices of extant plants for the determination of their phylogenetic position (Crepet & Nixon, 1998; Gandolfo et al., 1998; Sun et al., 2002; Friis et al., 2003a; Hermsen et al., 2003, 2006: Crepet et al., 2004). flowers helps in the interpretation of related floral Conversely, information from extant ' I am grateful to William L. Crepet for the opportunity to participate in the symposium “Paleobotany in the Post-genomics Era” held at the Botany 2006 meeting in Chico, Californi 2 Bernie material of several plants used in this et ation, and K. thank Rosemarie o for microtome sectioning, Ur the Wb uc Falso thank Victoria C. m h for support with the S Hollowel i for carefully reading t nked for support in the field and for pickled mic ana section of Hydrocharis morsus-ranae. | and Piel Hyland is tha ónig for a Bernhard for support with ne manusc rip Ji sugge stions * Institute of Systematic Botany, University of Zurich, Ze ae 107, 8008 Zurich, ia pendre m stbot.uzh.ch. ras 10.3417/2006190 ANN. Missourt Bor. Garp. 95: 101—120. PunuisuED on 11 APRIL 2008. 102 Annals of the Missouri Botanical Garden 1990: 2003: fossils (Crane et al., 1989; Friis € Endress, Frits, 1994, 2006; Doyle et al., 2003; Eklund et al., 2004). This light phenomenon that is informative in the understanding Endress & Hermsen et al., shed on a neglected study tries to of floral architecture and organ shape, for paleobo- tanical reconstruction, and would also be of interest from a molecular developmental and evolutionary perspective. It is partly a review, bul also presents original material. MATERIAL AND METHODS Species and specimens used for original illustra- lions are listed in Appendix 1. Material used for light microscopy (LM) or scanning electron microscopy (SEM) was fixed in formalin/ acetic acid/aleohol (FAA) or 70% ethanol. For serial microtome sections, specimens were either embed- ded in paraplast and the section series stained with safranin and Astra blue or embedded in Technovite Hereus Kulzer, Wehr- (2-hydroethyl methacrylate; heim, Germany) and stained with ruthenium red 5-10 Um and toluidine blue. Sections were mostly treated thick. For SEM studies, specimens were with 2% osmium tetroxide, dehydrated in ethanol and acetone series, critical point—dried, and sputter- coated with gold. AUTONOMOUS SHAPE AND IMPRINTED SHAPE (SUPERIMPOSED By CONTIGUOUS NEIGHBORS) The influence of mechanical pressure on the shape of plant organs has rarely been studied, perhaps contrast & Hillen, 2003). although the influence of mechanical stimulation on 2005). H distinguish between autonomous m imprinted shape. o animals (Benjamin plants is conspicuous (Braam. is useful to Autonomous shape is the shape of an organ de- veloping without physical influence from contiguous organs. Imprinted shape is the shape of an organ altered by the pressure of contiguous organs (Endress, 1975, 1994a, 2006). this effect. In the extreme case (level 1), —- There are various intensities of irregular imprints of adjacent contiguous organs are present on the affected organ. This goes so far that the disassembled organs can be put together like the pieces of a jigsaw puzzle (Fig. 1A. B). If two such organs are of the same kind (e.g.. two stamens), the mutual effect is about equal in both. If they are of different kinds. one of them is more deformed than the other. As an example, in a floral bud of an Annonaceae, the deforming influence of the stamens on the carpels is more pronounced than that of the carpels on the stamens for most of their length. This ts shown on transverse. sections by broad. indentations within carpel flanks but lesser indentations in stamen flanks (Fig. LA, left side). stamens of the bud are shaped by the three inner In addition, the peripheral is. relatively of the B). Thus. in these two examples it appears that the outer verianth organs; the peripheral line smooth and shows three corners at the site nargins of the three inner perianth organs (Fig. 1 older) organs influence the inner (younger) organs more than the other way round. In a less extreme case (level 2). but the organs fit tightly together in a harmonious way. there is no or less direct local deformation, As an example, in a floral bud of Gillbeea adenopetala F. Muell. whorls have filaments of different lengths, and thus Cunoniaceae). the stamens of the two the anthers are orderly stacked in two superposed rows (Fig. 2A). However. there are also cases in which, in contrast to an orderly stacking, the floral organs are not orderly stuffed in bud. e.g.. in flowers with a high number of stamens; such flowers may then have somewhat distorted anthers in bud, but the shape may be more or less restored when the flowers are open and the pressure is released, e.g., Barringtonia calyptrata R. Br. ex Benth., Lecythidaceae (Fig. 2B, C) (Endress. 19944). l appears affected by imprinted shapes are more common in that carpels | general, il stamens ( basal angiosperms and basal eudicots than in core eudicots or monocots. However, there are exceptions of such stamens in core eudicots, especially in flowers in which the perianth and androecium are not highly that perianth and androecium organs. An example is the synorganized and have irregular numbers of elusive genus Sphenostemon Baill., which was once Gibbs) with Trimeniaceae 1925), most angiosperm families. In complete contrast, today classified (as /denburgia (Austrobaileyales) (Perkins. one of the basal- Sphenostemon is in a family of its own (Sphenostemo- of the euasterids H Savolainen et 2003). lis of the | part because of i uncertain. position in 000; APG, erroneous placement Was because naceae) al., earlier gene ral Uu appearance of its flowers and i irregularly cuneate stamens ño an anther that is clearly differentiated from the filament. The stamens are densely crowded bud and have an imprinted shape by mutual pressure (Fig. 2D-G) (see also Endress, 2002). However, the reverse may also occur: imprinted shapes may play a role in highly regular and synorganized flowers. In Asclepiadoideae (Apoeyna- ceae), the two carpels fuse postgenitally in their uppermost part and this united. region attains a pentangular shape, molded by the five adjacent 1990). evnoecium looks like a anthers (e.2., Endress, Thus, the apex of the anthetic pentamerous slruc- Volume 95, Number 1 2008 Endress 103 Floral Architecture and Organ Shape Figure Autonomous and imprinted shape in flowers of Annonaceae. — Cananga odorata Hook. f. & Thomson. transverse section of floral center with inner stamens and carpels (with ventral slit and dorsal vascular bundle indicated). All organs are contiguous and arranged like the pieces of a jigsaw ; puzzle by mutually imprinted shapes. The section is slightly oblique (lower on the right side than on the left). On the left side, the stamens shape the carpels, and on the right side the carpels shape the stamens. B-D. Cyathocalyx martabanicus Hook. f. & Thomson, transverse sections. —B. Androecium and unicarpellate gynoecium. The stamens shape each other and the gynoecium, and the periphery of the androecium is shaped by the three inner (removed) tepals (arrows). —C. yynoecium shaped by stamens (arrows). —I ame gynoecium with autonomous shape at the base where the inner stamens are united into a ring and the imprinting influence of single stamens is lack C. D = 0.5 mm. acking. Magnification bars: A, B = 1 mm: C. | ture, although it is dimerous. Mutual adjustment of adjacent organ surfaces is necessary for postgenital fusion of the five stamens with the upper part of the eynoecium in these highly synorganized flowers. The developmental processes involved in this adjustment are largely unknown (see also Endress, 2006). However, such imprinted shapes are not commonly used for subsequent postgenital fusion. A similar quadrangular shape of a dimerous gynoecium, im- printed by four stamens, is present in Triaenophora Solereder (Plantaginaceae) (Wang & Wang, 2005). In some Lauraceae, the monomerous gynoecium attains a triangular shape in early development, molded by the trimerous preceding whorls. This has led to the erroneous interpretation of a trimerous nature of the lauraceous gynoecium (Singh & Singh, 1985), al- — though monomery is evident from comparative devel- opmental studies (Endress, 1972). Shape of a floral primordium may also. be influenced by adjacent parts of the inflorescence. Instead of being circular (Couroupita Aubl., Lecythi- daceae; Fig. 3A), it can be transversely extended, especially if the inflorescence is elongate and the flowers have a reduced perianth (e.g.. Euptelea Siebold & Zucc., Eupteleaceae, Fig. 3B, and Endress, 1986; Ren et al., 2007; Styloceras Kunth ex A. Juss., Buxaceae, von Balthazar & Endress, 2002). It can also be asymmetrical, if the inflorescence is monochasial 104 Annals on oad Garden Figure Spacing tn buds. . Gillbeea cn F; dx ll. (C unoniaceae). Regular spacing of the two whorls of five stamens em the floral bud. covere 2 by five ate sepals (s). B, C. Barringtonia calyptrata (Lecythidaceae), androecium, perianth removed. Irregular spacing (stuffing) " the numerous stamens in bud, anthers with imprinted aur by Paca de SDE ON —B. Bud view from the side. —C. Bud view from above. D-G. Sphenostemon lobosporus (F. Muell.) 1 n (Sphenostemon . —D. Flowering shoot. —E. Perianth part (p) and four stamens (st) from bud, irregularly contiguous. Lp. Stamen from bad (from ventral), showing irregular, imprinted shape. —G. Transverse section of bud with tight arrangement and imprinted shape of anthers. Magnification bars: A = 0.5 mm: B. C = 1 mm: E-G = = 0.2 (e.g., Heliconiaceae, Strelitziaceae, Kirchoff, 2003). flowers (Doust, 2001). Not only can floral primordia be The difference in shape of floral primordia may be asymmetric. but inflorescence primordia can be as especially striking if inflorescences have both a well, if they are under space constraints or under the lateral and a terminal flower, such as in Drimys J. influence of a polarity induced by the entire plant or a R. Forst. & G. Forst. (Winteraceae), in which the larger unit of the plant, such as in Loteae (Legumi- different shape of the floral primordium has reper- nosae) with a superposed dorsiventrality (Sokoloff et cussions on floral phyllotaxis of lateral versus terminal al., 2007). In Fuptelea (see above), in addition to the Figure 3. Shape of floral primordia. 3. Young inflorescence lips from above. —A. peta guianensis Aubl. (Lecythidaceae), floral ieee and young flowers widely spaced, round, with autonomous shape. —B. Euptelea polyandra Siebold € Zuce. (Eupteleace floral primordia and young flowers crowded, flattened, with imprinted shape: the abaxial transverse line is caused by pressure of the subtending bract (arrows). Magnification bars: A = 1 mm: B = 0.5 mm. Volume 95, Number 1 2008 Endress Floral Architecture and Organ Shape 105 general transversely extended shape, direct pressure by the subtending bract leaves its mark on the abaxial the young flower (Fig. 3B). side o ARCHITECTURAL CONDITIONS FOR THE PRESENCE. ( CUNEATE OR SHIELD-SHAPED ORGANS An example for imprinted shape are Lower Cretaceous isolated fossil stamens, many of which show two striking features: (1) cuneate (wedge- shaped) anthers and a massive apex. and (2) anther laterally hinged valves and not by simple longitudinal slits (Friis et al., 1991, 20064). Such cuneate stamens (anthers) also occur in some extant plants (Figs. 4, 5) (Endress, 1975; Endress & Hufford, 1989; Hufford & Endress, 1989). We may ask: Are cuneate stamens typical for certain clades, or dehiscence by is this shape due to certain architectural constraints independent of systematic relationship? Both ques- lions can be answered affirmatively to some extent. The first aspect, cuneate shape. which is related to imprinted shape, will be addressed in this section: the second aspect, valvate dehiscence, will be addressed in the next section. Isolated cuneate (wedge-shaped) anthers with PS massive sterile apex and short filament indicate with some probability that the anthers were contiguous and densely arranged in bud, that the anther apex had a protective function and was not covered in bud by other organs, and that stamen arrangement was in a spherical, hemispherical. fashion. or cylindrical Therefore, the surface of the entire architecture is largest at the periphery, which results in the cuneate form of the stamens. There are two possibilities for such an architecture: (1) inflorescences with dense flower arrangement and the perianth lacking or small and not protective, such as in extant Hedyosmum Sw. (basal angiosperms) (Fig. 4A-C). Platanaceae eudicots) (Fig. 4D—F), (basal core eu- G-I). 4J-L). and Typhaceae (Fig. 4M-0) (both commelinid monocots), and (2) single flowers with numerous stamens. such as in Nym- both basal angiosperms (Endress, 1975 basal Altingiaceae dicots) (Fig. Sparganiaceae (Fig. Annonaceae (Fig. 5) and Nuphar Sm. in phaeaceae, 1987 a ). In Platanaceae, the female inflorescences bus similar architecture. as the male ones. and the carpels have a cuneate ovary (Crane. 1989). In some Annonaceae (although they have a well-developed protective perianth). the stamens become exposed in late bud, and thus their broad apices are then the protective parts for the thecae and the ovaries. In both cases, the massive anther apices may also be a protection against pollinators with biting mouth parts or other ne o insects. Thus, for isolated fossil stamens with such a shape. even if they may not be determined as to their larger lacking), alliance (if pollen is partial reconstruction of the architecture of their flowers or inflorescences may be possible. Cuneate carpels arranged in a spherical gynoecium occur in Kadsura marmorata (Hend. € Andr. Hend.) A. C. Sm. 1947), and cuneate gynoecia (without perianths and subtending in Schisandraceae (Smith, bracts) arranged in a spherical inflorescence occur in Pandanus S. Parkinson (pers. obs.). There are not only cuneate f kinds of protective function and similar dense arrangement into cone- like bracts that subtend flowers or partial inflorescences (cymes) are often cuneate or loral organs but also other cuneate organs with a architectures. In extant plants, have a shield-like portion at the periphery of the bud of dense inflorescences (e.g Piperaceae [Fig. 6A-C], Mimosoideae of Legumino- sae [Fig. 6D-F], and Betulaceae | Fig. 6G-I]. Endress, 1975; Balanophoraceae, Kuijt, 1969; Moraceae, Berg, 1990). This is also true for sporangiophores of various “9 gymnosperms and groups other than seed plants (e.g .. Equisetum L. [Fig. 6P-R]. Endress, 1975; Taxaceae [Fig. 6M-O]. Endress. 1975; Mundry & Mundry, 2001; cycads [Fig. 6]-L]. Endress, 1975; Mundry « Stützel, 2003). Cuneate organs are also known from a number of various fossil plants. Subtending bracts (or sporangiophores) are cuneate in some lvcophytes (Mazocarpon Benson; Schopf, 1941; Pigg. 1983), sphenophytes (Palaeostachya Weiss: Baxter, 1955), eycads (Androstrobus Schimper, Beania Carruthers, Harris, 1941, 1964; Delemaya Klavins, Taylor, Krings & Taylor, Klavins et a (Ohsawa, 1994; Stockey, 1994). Ovules are cuneate in 1948: Crane. 1985) and inter- seminal scales in Bennettitales (Harris, 1969; Crepet. 1974; Sharma, 1982: Pedersen et al.. 1989: Nishida, 1994; Rothwell € Stockey, 2002; Stockey & Rothwell, 2003). The formation of such massive peripheral, contiguous protective parts is commonly provided not , 2003). and many conifers Pentoxylales (Sahni, by a small-celled, eytoplasm-rich marginal meristem but mainly by cell enlargement (Endress, 1975). The well-preserved Bennettitales described by Crepet and Delevoryas (1972) also indicate that cell enlargement alone was Instrumental i 1 the thickening of the peripheral, protective part of the ovular integument in female reproductive structures. CONDITIONS FOR VALVATE DEHISCENCE IN ANTHERS AND SIMPLE LONGITUDINAL The Cretaceous stamens (Friis et al.. second striking feature of fossil Lower 2006a: fig. 21). addition to the cuneate shape (see above), is that a number of them have a valvate dehiscence pattern with two laterally hinged valves in each of the two 106 Annals of the Missouri Botanical Garden > ¿AS s: Won y Figure 4. D, G. J, M. Stamen (anther). — — C. F. E. L. O. Entire inflorescence. A-C. Hedyosmum mexicanum C. Cordem. (Chloranthaceae). D-F. 7 (Platanaceae). G-L Liquidambar styraciflua V. (Altingiaceae). J-L. Sparganium erectum V. (Sparganiaceae). | —O. Typha minima Eunck in Hoppe (Typhaceae). Tanniferous tissue black in B, E. N. (From Endress. 1975, with permission: ). therefore, valvate dehiscence is more efficient (Fig. 7D—F). Valvate stamens with this architecture (thick connective) are concentrated in some clades and not common in angiosperms as a whole. Although these clades are in disparate places in the angiosperm tree, they are more or less restricted to basal angiosperms Such among basalmost angiosperms in Nymphaeales (Vu- phar; Hufford, 1996); among other basal angiosperms in Laurales (Sinocalycanthus (Cheng & S. Y. Chang) Cheng & S. Y. Chang and some fossil Calycanthaceae: and basal eudicots. anthers were recorded 107 108 Annals of the Missouri Botanical Garden DO 29,949 9.9 2, p! p f Figure 6. cylindrical inflorescences or shoots with pr a —B, E, e Subtending bre and P. A, B. Peperomia inna (L E Kunth (Piperaceae). (L.) Gillis & Stearn Kee el (Zamiaceae). M-O A ata l. ( permission; «http: Jw ww.schweizerbart.de- Sporangiophore a P Friis et al.. 1994; Crepet et al., 2005; Staedler et al., 2007; Atherospermataceae, Lauraceae. Hernandiaceae, Monimioideae-Monimiaceae; Endress & Hufford, 1989). Magnoliales (Annonaceae [Fig. 8A]. Magnolia- ceae, Eupomatiaceae, Himantandraceae, Degeneria- ceae; Endress € Hufford, 1989), and Piperales (Piper augustum Rudge; Endress, 1994b): among basal eudicots in Ranunculales (Eupteleaceae; Endress 1986; Ranunculaceae p.p.; Endress & Hufford, 1989; Weber, 1993), Proteales (Platanaceae |Fig. 8B]; Huf- ford & Endress. 1989), and Trochodendraceae (Endress, 1986; Hufford & Endress, 1989; Chen et al., 2007); and among basal core eudicots in Saxifragales (Altingiaceae, Hamamelidaceae [Fig. 8C]; Endress, 1989; Hufford € Endress, 1989). Valvate anthers appear to be absent from monocots and higher core eudicots. LABILITY IN NUMBER AND SHAPE OF REDUCED ORGANS THAT Have DECREASED IN SIZE THEIR ORIGINAL FUNCTION AND LOST Reduced organs that have decreased in size (1.€., have become shorter and narrower) and lost their Cuneate subtending bracts of flowers or p inflorescences or p p. C. Alnus glutinosa (L.) Gaertn. (Betulaceae). J-L. jj. P-R. Equisetum arvense L. sporangiophores, and hemispherical o . D. G. J. M. P. Inflorescences or shoots with roni O, R. Longitudinal sections of (parts G. ]. M. Leucaena latisiliqua — of) ^ P. caperata Yunck. (Piperaceae). D-F. Zamia verschaffeltii Miq., (Equisetaceae). (From Endress. 1975, with fe SER original function may become labile in number and shape because functional constraints on shape are lacking. A reduction in size and loss of function of an organ type may be concomitant with a decrease ii irregular increase in number, but the reverse—an number—is also possible. What are the conditions for maintaining a constant number of organs in whorls? [n eudicots, sepals are basically the protective They cial-imbricate (with overlapping flanks) or. less often, organs in floral buds. are commonly quincun- valvate (with margins of adjacent sepals tightly appressed to each other) Especially in the latter case, all sepals of a flower have the same shape and their number is stable, as they have to build a precise Heisteria — envelope for the young floral organs (e.g., Jacq.. Olacaceae) (Fig. 9A). However, here and there in the phylogenetic tree, the protective function has been evolutionarily transferred from the sepals to another organ category (e.g., petals) or, less often, floral subtending bracts or prophylls. If petals have become protective, they commonly become larger than the sepals early in development and often have a The sepals have then valvate aestivation in bud. Volume 95, Number 1 Endress 109 2008 Floral Architecture and Organ Shape NZ JS . Pas 7 Figure 7. Ahon with opening by longitudinal slits (A-C) and laterally hinged valves (D-F). —A, p. Pie section of O befe (thick, interrupted line indicates thickness of anther: thin, ee ed lines i ay areas indic 'ate pollen sacs). —B, E. Anther before dehiscence, from transverse side, showing theca with ae lines as pem indicate ends of dehiscence lines). —C, F. Anther after dehiscence, from transversal side (interrupted line indicates septum between the two pollen sacs of a theca; asterisks indicate ends of dehiscence lines; arrows indicate directions of opening of theca) commonly become small and sometimes irregular in shape of the organs. An example is Mimosa L. shape and number. In some cases, the number has (Leguminosae), in which several subgroups have a irregularly increased. This irregularity is due to a lack valvate corolla (Fig. 9C) and a highly reduced calyx. of a functional constraint on a precise number and which is irregularly denticulate, with more teeth than Figure 8. Anthers with dehiscence by two laterally hinged valves for each theca. —A. Artabotrys hexapetalus (L. f.) Bhandari (Annonaceae). Platanus orientalis L. (Platanaceae). —C. Folherzilla major Lodd. (Hamamelidaceae). Magnification bars: A-C = 0 110 Annals of the Missouri Botanical Garden Figure 9. Well-developed versus reduced perianth organs. well de "ve pne i B with valvate aestivation. B, C. ae M B = 0.1 mm. would be expected from the merism of the flower (Fig. 9B) (see also Tucker, 1984; Barneby, 1991). Even more conspicuous examples are Asteraceae and Valerianaceae, in which the originally pentamerous calyx is dissolved into more than five pappus bristles e.g., Semple, 2006) Thunbergia Retz. and related genera (Thunbergioi- deae, Acanthaceae) (Schónenberger & Endress, 1998; 1999) are organs. As in Schonenberger, examples for bracts as protective floral most Lamiales, the basic sepal number in Acanthaceae is five, and the sepals are protective in bud. However, in Thunber- vioideae, the mechanical protective function for the prophylls, OA). In flowers has been transferred to two large bud (Fig. E which are postgenitally united contrast, the sepals are not protective at any stage of Thunbergia, reduced calyx. Figure. 10. flower bud. —C. € 0.25 mm; C = 5 mm Mimosa spegazzinit Pirotta (Leguminosae). — etals well developed, with valvate aestivation. p = petals, s = sepals, a —A. T. grandiflora Roxb., sal yx diffe rentiation in five species of Thunbergia. . Heisteria parvifolia Sm. ne aceae), sepals and pe ‘tals epals reduce d. with open = ition bars: A, C = 0.5 mm; floral development (except indirectly for bearing extranuptial nectaries, which attract ants that defend the flowers against herbivores) (Fig. 10B). Thus, the sepals are much reduced in size and sometimes completely lost. However, if present, they often have irregularly increased in number from originally five to 15 or more in some species; they form short scales of slightly variable size. In sphingid-pollinated species (e.g.. T. guerkeana lindau) with increased flower length, these ca. 15 narrow organs may be secondarily elongate in terms of evolution (Fig. 10C) (Sehönen- berger, 1999). Subtene another possibility for ing bracts are surrogate protective organs if the sepals are reduced. In Balanophoraceae, which has massive protective flowers of Helosis Rich. still have bracts. male open flower from the side. —B. T. alata Sims, young ) = petal, pr = prophyll, s = sepal. Magnification bar: B Volume 95, Number 1 2008 Endress Floral Architecture and Organ Shape 111 Figure 11 sepal initiation; in B, one of t dels of one of the two Carpinus rin Niue a ), female flowers, with reduced calyx. - j lowers r Two flower buds in a dic hasium at —A, B. emoved. —C. Two flowers in a dichasium, at anthes prophyll, sb — subtending RR poene ads = sepal primordia. Magnification bars: A, B = 0.2 mm: C = 1 mm. protective, valvate sepals of a fixed number, whereas those of the closely related Corynaea Hook. f. have nonprotective, reduced sepals with irregular shape (Kuijt, 1969). A. Mey. and Sycopsis Oliv. (Hamamelida- and increased, unstable number — Parrotia C. ceae) also have reduced, nonprotective sepals with irregular shape and increased, unstable number, in contrast to other Hamamelidaceae with larger, protective. sepals with a fixed number (Endress, 1989, 1990). A (Betulaceae), and shape ast example, female flowers M Carpinus L. should be treated in more detail. In contrast to the male flowers of Alnus Mill. and Betula L. with four and two protective sepals, according to tetramerous and dimerous flowers, respectively, in Carpinus (Abbe, 1935), the sepals are not protective; ea they are short, of irregular shape, and increased in number. However, although in C. betulus L. sepal number and shape are irregular (Eichler, 1878), there are commonly four somewhat larger sepals in the median and transverse planes, which is still reminis- (Abbe, 1935; for the young basic 1967). flowers are the subtending bract of each flower pair, cent of tetramerous pattern Endress, Protective organs — which represents a dichasium without a central flower, plus the subtending bract and the two prophylls of each single flower (Fig. 11C). The sepals are much delayed in development, and at no time do they enclose the inner parts of the flower. Sepal delay is so pronounced that they only appear after the two carpels have been initiated. They first appear as two irregular rims, each in the median plane of the bicarpellate gynoecium (Fig. 11A, B). Each sepal (or calyx tooth) ends in a colleter in early development—a secretory tip that probably has a chemical, and not mechanical, protective function for the young flower (Figs. 12, 13). Later in development, pronounced longitudinal ribs develop in the calyx teeth extending along the surface of the inferior ovary and are most conspicuous in fruit (Fig. 13). situated, which probably differentiate so strongly due In these ribs, strong vascular bundles are to the early secretory function of the colleters, which they serve. Such pronounced ribs are also present in fossil Betulaceae-like flowers (Schónenberger et al., 2001; Friis et al., 2003b, 2006b). In Carpinus betulus, structural lability is expressed at the of the (Fig. 12A—H) between individuals (Fig. 13A—F). Within an individ- level individual and Annals of the Missouri Botanical Garden Figure ] ¿arpinus betulus (Betulaceae). —A-H. Female flowers from a single tree showing intra-individual variability of sepal ecu with corresponding se pal formula of the visible side indicated at upper right. —I-K. Floral diagrams with corre dos sepal formula of one side of the flower indicated. Magnification bars: A-H = 0.5 mm. ual, between the commonly four larger sepals in the median and transverse planes there may be zero, one, or two commonly smaller sepals. Within a flower, all these numbers may be present without a recognizable pattern (Fig. 12A-K). Inter-individual variation can be seen by comparing developmental series of flowers y | : | In the second 13A-F) the sepals are shorter, more numerous, and, i from two different trees (Fig. tree, older stages, more curved so that in the fruit they form a more convoluted rim (Fig. 13F) than in the first tree (Fig. 130). Thus, in both Thunbergia (Acanthaceae) and Carpinus (Betulaceae), the sepals appear to have convergently lost their mechanical protective function with size reduction and concomitant loss of a fixed size, shape, and number, but perhaps e gained a secretory protective. function. Therefore, small sepals of irregular size associated with secretory structures in fossil flowers may indicate that they did not have a mechanical protective function and that their number may be variable. Increase in number concomitant with reduction in size and loss of function is not only present in sepals. It may also occur in stamens. An example is the large genus Bauhinia L. (Leguminosae), in which the basic number of 10 stamens in two whorls as common in the family may be reduced to five, three. or even a single In B. g (anterior) stamens of the outer . HA, D). reduced to staminodes 14C, D) 19944) two upper staminodes of the outer whorl are sull inet [n the number. In the expected alpinii N. E. Br., only the three lower stamen. androecial whorl are The small other seven without anthers (Fig. (Endress, . The than those of the inner whorl. inner whorl, the organs are increased in position of each single organ, there are two or three (Fig. 14B, D). In Dillenia L. numerous stamens tiny. collateral. organs (Dilleniaceae). the flowers have Volume 95, Number 1 200 113 Floral Architecture and Organ Shape ary (date: s of collection indicated), showing inter- individual variability of se ) —D-F. Another tree. Magnification bars: A-C. F = 0.5 mm: that are centrifugally initiated. The last initiated organs of the androecium are staminodes. They are narrower than the stamens, and two collateral organs occupy a position in which a single organ would be expected (Endress, 1997: fig. 8B) Long Hairs as FILLING MATERIAL OF (IRREGULAR) SPACES The presence of long hairs concentrated in certain areas of flowers is sometimes an indication of irregular spaces in flower buds, as hairs may be used as flexible filling material, which may have protective functions against herbivores or physical factors (e.g.. drought or frost). Such hairs occur preferably at or around carpel bases. There they may also play a role in fruit development. Examples are seen in Hamamelidaceae and Cunoniaceae (Matthews & Endress, 2002), which 10, or stamens. In these flowers, long hairs are at and around (Fig. 15A-D). have only two carpels but have five, more the ovary In some Monimiaceae, the irpinus betulus (Betulaceae). A-F. Female flowers from two different trees. in the same deve lopmental stages ALC. S; A-H p formation. — ame tree as Figure 12 m. D, E = 0.2 carpels are free in the bottom of a floral cup. and hairs develop in the interspaces between the carpels (Fig. 15E) (Endress. 1980). (Araliaceae), the hairs are filling material for the open the that increased number o (Endress, 2006). In between floral organs or (Berg, 1990). Another instance n Munroidendron Sherf space in floral center originates. by whorl (Fig. protective carpels in a Moraceae, flowers are also common is hairs on anthers, as seen in male flowers of Carpinus (and other Coryleae. Betulaceae). Here, sepals are not only reduced as in female flowers but completely absent. The inner space of the bud is provided by the subtending floral bract and the bracts of the adjacent flowers, and has an irregular shape so that the anthers cannot be tightly packed in a strict pattern. Thus, the space is not spherical, as often in flowers, but broad and flat. To fill features: this space, these flowers have two unusual (1) the two thecae of each anther are 114 Annals of the Missouri Botanical Garden Figure 14. reduc ed organs. — low d galpinii N. E. . Flower with "dh fertile stamens y). zx staminodes inc me in number of inner whorl (x organs of outer whorl, white: organs of inner akoni- -D. number of strongly rec letters corresponding to those in Fig. 14B). Magnification bar: separated from each other by a filamentous connec- tive, which allows flexible arrangement of the thecae (Fig. 15H); and (2) again, long hairs fill empty spaces. These hairs occur as tufts on top of each theca (Fig. 15G, H). Such hairs on anthers are present in a Be ew unrelated groups of wind-pollinated plants with a reduced perianth (e.g., some Anacardiaceae, Faga- ceae, Juglandaceae) (Elias, 1972; Endress & Stumpf, 1991: Bachelier & Endress, 2007). interesting to know whether the presence of hairs on It would also be these anthers is just a consequence of sepal reduction and a concomitant irregular space in bud, or whether it has acquired an additional ecological function in the context of pollen dispersal by wind. ARCHITECTURAL CONDITIONS FOR THE PRESENCE OF ORTHOTROPOUS OVULES IN. ANGIOSPERMS In contrast to other seed plants, angiosperms EN commonly have anatropous ovules. Ovules with this shape have the micropyle topographically adjacent to the placenta and, thus, can take up pollen tubes directly from the placenta. Nevertheless, orthotropous ovules occur sporadically or regularly in some larger (arrow 3r. (Leguminosae), reduction of androecium and concomitant inerease in number of Stamen and staminode of outer whorl (X, Y) and ie k: ws). mE lagram of — indroeci lum common in many Leguminosae ( Dian am of androecium of Bauhinia galpinii, showing increase in uced staminodes " inner whorl (black and ane organs of outer whorl; white: organs of inner whorl; 151 B= mm. groups. They conspicuously tend to occur under certain conditions of the ovary architecture, three of which are listed here. Commonly, the inner architec- ture of the ovary is such that the pathway of pollen tubes from the stylar canal to the micropyle of the ovules is alleviated by the orthotropous shape. ( single basal (or almost basal) ovule, [evi ) The most simple condition is the presence of < which positions its micropyle directly to the lower end of the stylar canal, thus enabling direct passage of a pollen tube pen Fig. 16A, D, C). This condition has evolved in many different angiosperm families, such as in Piperaceae 1998), Didymelaceae (basal eudicots; von Balthazar et al., 2003), lales), Juglandaceae and Myricaceae (core eudicols, (basal angiosperms; Igersheim € Endress, Polygonaceae p.p. (core eudicots, Caryophyl- Urticaceae (core eudicots, rosids, 1937), cols; Dahlgren, 1939), and Araceae (basal monocots). Fagales), rosids, Rosales; Eckardt, Zosteraceae (basal mono- (2) If the locule is filled with secretion, ovules of any number can be orthotropous as the pollen tubes may directly grow through the secretion (e.g., Buzgo, 1994) (Fig. 16B, E, H). This condition is known from a number of plants that grow in wet or moist habitats, Volume 95, Number 1 Endress 115 8 Floral Architecture and Organ Shape Figure 15. Long hairs as filling material of empty spaces in HOM buds and flowers. —A. Hamamelis d L; (Hamamelidaceae). —B. Trichocladus grandiflorus Oliv. (H sae). —C. Acsmithia davidsonii (F. Muell.) Hoogland (Cunoniaceae). —D. Geissois biagiana F. Muell. (Cunoniaceae). —E. Wilkie ea sp. nee (Monimiaceae -F. Munroidendron racemosum (C. N. Forbes) Sherff (Araliaceae) G, H. Carpinus betulus, male (Betulaceae). Par af male inflorescence, flower-subtending bracts partly removed. -—H. Stamen, with the two thecae flexible by lament sntous connective and crowned by tuft of long hairs. Magnification bars: A, B, F = 0.5 mm; C = 0.3 mm; D = 0.4 mm; E, G = 1 mm; H = 2 mm. such as in Barclaya Wall. of Nymphaeaceae (basal- These three architectural conditions for the pres- most angiosperms; Schneider, 1978; Igersheim & ence of orthotropous ovules were briefly discussed in Endress, 1998), Acoraceae (basal monocots; Rudall & Endress (1994a). There are also other instances of Furness, 1997; Buzgo & Endress, 2000; Igersheim et — orthotropous ovules, in which the ovary architecture al. 2001) Pistia L. and other Araceae (basal does not so evidently favor the orthotropous condition monocots; Buzgo, 1994), and some Hydrocharitaceae as in the three types mentioned. (basal monocots; Igersheim et al., 2001). (3) If the placentae are parietal, the micropyle of — Structural DIFFERENCES BETWEEN EXPOSED AND orthotropous ovules may be directed not to their own Covereb ORGAN Parts IN BUD placenta but to the neighboring placenta, where they take up pollen tubes (Fig. 16C, F, I). Examples are Organ surfaces that are exposed in bud may behave Casearia Jacq. (core eudicots, rosids, Malpighiales) differently from those that are covered in bud. This and Mayacaceae (monocots, commelinids). is especially conspicuous in imbricate aestivation, 116 Annals of the Missouri Botanical Garden igure l6. Orthotropous ovules and ovary architecture in oun "ms n micropyle directed to stylar canal. —B. I. ules in a locule with parietal placentae. n micropyles directed to neighboring place nta. A-C. \ single ovule al the base of a loc ule, —A,D, ( Several ovules in a spacious loc Me fille d with secretion. —C, F, I. Several ematie figures of ovaries ies ovules. D-I. Microtome sec i of ovaries with ovules. —D. Piper augustum (Pipera : E Hydr -charis morsus- ranae L. (Hydrocharitaceae). — . Mayaca sp. indet. (Mayacaceae). G. Polygonum firme De (Polvgonaceae). —H. Pistia stratiotes L. (Araceae). = C asearia silvana Schltr. (Salicaceae) (arrows point to micropyle). Magnification bars: b E. H = 0.1 mm: F, | 0.2 mm: G = 0.5 mm. (A, B, C from Endress, 1994a. Reprinted with permission from Cambridge University Press.) where the same organ has exposed and covered areas. For instance. if there is an indument of hairs, the hairs may be restricted to or more strongly developed on the Nelson. Fremontodendron exposed areas (Stellaria media (L.) Vill., 1954; Coville, von Larreat., Balthazar et al., 2006: Staples, 2007; Convolvulus tricolor Another Chiranthodendron Rivea Choisy, pers. obs.). difference is that the exposed flanks are whereas the covered flanks are (L.) Roth.. detailed study on this positional effect is that by Warner et al. (2005, Nymphaeaceae. green hyaline (Ipomoea purpurea pers. obs.). The most 1 press) for Nuphar and other Often in imbricate aestivation the margins are thin (two cell layers). The thin part of the covered flanks is often broader than that of the exposed flanks within a flower (e.g.. sepals in Caryophyllaceae, Rohweder, 1970; Corynocarpus J. R. Forst. & G. Forst., Matthews & Endress. 2004: Brexia Noronha ex Thouars; petals )05a: and sepals and petals in /xerba A. Cunn., Matthews € Endress, 2005b). differentiation of the sepal flank is especially obvious 1 Siphonodon Griffith, Matthews & Endress, 2( This influence of the position on the pentamerous flowers with quincuncial aestivation. The third-formed sepal of the five has a covered and an exposed flank, and the covered flank is regularly 1970). A well-known example are the sepals of rose flowers, in thinner than the exposed flank (Rohweder. which the exposed margins have pinna-like append- ages, whereas the covered margins lack them: Troll Volume 95, Number 1 2008 Endress Floral Architecture and Organ Shape 117 (1957) rose sepals, which reportedly goes back to Albertus Magnus (13th century). In the organs have thick, valvate tepal, sepal, or petal aestivation, abruptly ending margins, they are triangular, and the curvature of the margins reflects the shape of the bud—the more curved the margins are, the bud. Among floral fossils, the monocot Mabelia Gandol Nixon & Crepet has such valvate 2002). the shorter and rounder as o. tepals (Gandolfo et al., SEPAL AESTIVATION AND PINNATE PETALS Petals that are elaborate and have several lobes along the margins (pinnate shape) tend to occur in flowers with valvate calyx aestivation (Endress $ Matthews, 2006). that in valvate aestivation the margins of two adjacent A reason for this correlation may be sepals are contiguous in a mirror symmetrical fashion and do not change position relative to each other during development. Thus, the space for the compli- cated petal to develop does not change its shape, and the petal lobes are unobstructed in their development. Such pinnate petals are also known from floral fossils (Carpenter & Buchanan, 1993). In these fossils, sepal aestivation is unknown, but it can be predicted that it was probably valvate not only because they appear to belong to Cunoniaceae, which often have valvate sepals (also in cases with simple or no petals), but also because of the presence of pinnate petals. Flowers with such lobed or pinnate petals in combination with a valvate calyx are known from many core eudicot families, especially in rosids (some Myrtaceae, Onagraceae, Rhynchocalycaceae, Anisophylleaceae, Cunoniaceae, Elaeocarpaceae, Celastraceae, Rhizo- Matthews et al., 2001: Matthews € Endress, 2002, 2004; Schónenberger & Conti, 2003; Endress & Matthews, 2006). phoraceae; CONCLUSIONS The mechanisms of the imprinting of organ shapes by contiguous neighboring organs are unexplored. From the diversity of cases, it appears that there are different degrees of imprinted shapes. There are directly imprinted shapes (such as in the stamens of Sphenostemon), and shapes that just fit in the general architecture and may not be directly shaped during the individual development but were installed during evolution (such as in the floral primordia of Euptelea). The stamens of exhibit. different levels of imprinting. A more general level may form Annonaceae may the cuneate shape, and a more individual level may form the individual deformations that are different in cites a rhyming Latin riddle on this behavior of each single stamen. lt may not always be easy to recognize to what extent one or the other is involved. The observations discussed in this paper should entice further and more detailed studies of interrela- tionships between the whole and the parts of structural systems in plants. 1 hope they will also help in the reconstruction. of fragmentary fossil. material. There are certainly more influences of the whole architec- ture on the shape of single parts than those shown here, which represent conspicuous examples. Literature Cited E E. C. 1935. Studies in the phylogeny of the Betulaceae .F bra and ee ‘ence anatomy and morphology. Bot. ne : 1-67 APG utm Phylogeny Group). 2003. An update of the Angiosperm Phylogeny Group classification for the orders and f OE of flowering plants: APG II. Bot. J. Linn. Soc. 141: —436. pu A B. & P. K. Endress. 2007. Inflorescence, cupule, p np and flowers of P and comparison with Pistacia (Anac Tom eae). Int. J. Plant Sci. 168: 1237-1253 Barneby, R. C. 1991, cine: censitae: A description of = genus Hein World. s New Yo = ar, R. [9] o us dep eas 2m the rk | 65: I= 955. ps ‘hya ui a new species s cone from the Amer. J. Bot. a 342-351. Benjamin, M. € B. Hillen. 2003. Mechanical influences on cells, tissues wi organs: Mechanical morphogenesis. Eur. J. ue Berg, C. € d Differe ntiation of flowers and inflores- cences af Urticales in relation to their protection against breeding insects and to pollination. Sommerfeltia 11: —34. Gard. 6 American Carboniferous. UN J. 2005. In touch: n responses to mechanical stimuli. w ge 165: —389. Buzgo, M. Ml. rence | development of Pistia stratiotes TU Bot. Jahrb. Syst. 115: 557-570 — & FP. Endress. 2000. Floral structure and development of Keo raceae and its systematic relationships : 23-41. with basal pu qua Int. J. PL Sci. Carpenter, R. J. & A. M. Buchanan. 1993. 3. Oligoc ene leaves, fruit and oss a the Cunoniaceae from Cethana, Tasmania. Austral. Syst. Bot. 6: 09. Chen, L., en, P. K. Endress, X. H. Tian & X. H. Zhang. 2007. F local development of Tetracentron sinense 1 A ho- dendraceae) and its systematic significance. Pl. Syst. Evol. 264: 183-193. Crane, P. R. 1985. Phylogenetic analysis of seed plants amd the n of angiosperms. Ann. Missouri Bot. Gard. 116—793. 1989. Palaeobotanical evidence ion of nonmagnoliid dicotyledons. 162: Cum 191. on PL Boi early st Evol. . M. Friis & K. R. Pedersen. 1989. Re eq ^ struc es and function in Cretaceous Chloranthaceae. Syst. 165:-2 26. Boo W. 1974. Evol. O Investigations of North American g of P Me 148: —16 Investigations ie 'adeoids: The reproductive biology P kis 'ontographica, Abt. B, Palüophytol. & T. Delevoryas. 197 s h J. Bot. iz. American cyc poe Pul ovule ontogeny. Amer. 59: 2€ 118 Annals of the Missouri Botanical Garden € K. C. Nixon. 1998. Two new fossil flowers of magnoliid uM from the Late Cretaceous of New Jersey Amer. J. Bot. 85: 1273-1288. M. Gandolfo —, —— 904. Fossil evidence and ulis logeny: The age of major P M rm clades based on mesofossil and macrofossil evidence from Cretaceous 3 1666— 1682. . 2005. An extinct Ue anthoid deposits. Amer. al Bot. 91: taxon, een aei from the La reta- Amer. J. Bot. 92: 1475- 1939. Endosperm- und Embryobildung yel usted marina. Bot. . 4: 607-615. De Vore, M. L., P. ick, K. B. Pigg & R. A. 2000. Utility of high poro X-ray computed 1 lomogra- phy (HRXCT) using London Clay fruits and seeds. 1848-1851. Doust, A. N. 2001. The developmental basis of floral variation in Drimys winteri (Winteraceae). Int. J. Pl. Sci. 6: 7-717 p cc ec ous of N Dahlgren, Ketchar n exam] ple Bot. for paleobotanical studies: Amer. J. ( d. E Doyle, J. A. H. Floral jae e phylogenetic analysis. Int. J. PI. » 2003. Implications of a Sci. 164 Herendeen. Eklund $ S evolution in Chloranthaceae: (5. Suppl.): 5305-5382. E ede T. a Untersuchungen Entwicklungsgesc des pse ae ren Gynoeceums. N.F. 5(26): 1-112. Kichler, A. i 1878. Leipzig. Eklund, H., J. A. Doyle € P. 5. 2004. e o ri of ae and fossil Chlo PI. Sci. 165: 107- Elias S. d 22. venera of Ju Arnold Systematische iiber Morphologie, Bedeutung Acta Leop. systematische Nova hichte und Blüthendiagramme 2. Engelmann, Here ndeen. ranth aceae. Int. J. a pn gla Me astern United States. J. nes Endres ^ K. 1967. Studie verw Mos haftlichen Beziehungen. zwischen den Hama- Bot. Jahrb. Syst. 87: the 53: 26-51. Ws 'eae über die m relidaceen und Betulaceen. 31-525. . 1972. Zur vergleichenden Entwicklungsmorpholo- gie, Embryologie und Systematik bei Laurales. Bot. Jahrb. Syst. 92: 331-428 975. Nachbarliche Formbeziehungen mit Hüll- funktion i im infloresze mz- und Blütenbereich. Bot. Jahrb. 96: 1—44. . 19 -— — Ontoge ny, 2 function and e ol extreme floral construction in Monimiaceae. Pl | Evol. 134: 79-120 198 36. Floral structure, systematics and T logeny in Troe hode a . Missouri Bot. Gard. 297-324. . The € hloranthac reae: Reproductive struc ay s -220. and geet position. Bot. Jahrb. Syst. 109: . 1989. As Hamame ae 'eae and the Lower Hamamelididae. ames 19 ects of evolutionary differentiation of the PI Syst. y JO. Un rns s of floral construe tion in ontogeny and phslogen ny. Biol. J. Linn. Soc. 39: 153-175. — Oda. Tropic a Hose srs. Cambridge Univ. Press. 94b. Shapes stamens a Magnoliidae. Bot. 1997 Dia and Prato Cambridge. s, sizes and Un trends in Jahrb. Syst. 115: 420-460 Relationships between floral DART architecture. and ea mode in Dillenia (Dillenia- . Syst. Evol. -118 2002. Morphy hf angiosperm systematics in Bot. Rev. 08: 5 ceae). d mole cular era. Biology of Angiosperm floral au um D MR d framework. Adv. Bot. Re : 1-61. & E. M. Friis. 1994. ditor Ma tre nds in study of early flower e Pel PI. Syst. Evol. Suppl. 8: 1-6. 2006. R quur auper tant, am the ‘ir bearing on ~ re e 2 structure C fossil and « Introduction. Pl. Sy 10: — un 1989. The i dius rsity of stamen structures P dehiscence patterns among Magnoliidae. Bot. J. Linn. Soc. 100: 45-85. —— QU M. L. Matthe ws. 2006. Elaborate petals and staminodes in eudicots: Structure, function, evolution. Organisms Dive sity Evol. 6: 257-293. — & S. Stumpf. 1991. The diversity of stamen structures in “lower” Rosidae. Bot. J. Linn. Soc. 107: 217-29: 3. Friis. E. M. € P. K. Endress. 1990. Origin and evolution of flowers. Adv. Bot. Res. 17 1 angiosperm fÍ ] p Pedersen. | ; Bar diversity and | in a pollen. of Cretaceous angio Pp. 197-224 in S. Blackmore & S l Pollen and die s: Patterns of E Clare aden — | irane & K. Stamen sperms. 3arnes (editors Ixford. de Kone, Virg anthoi Press, a 1994. A calycantha- Pedersen & P. R. Crane. sp. nov —A ^5 gen. ets eous flower from the Potomac group (E arly Cretaceous) of 785. eastern North se rica. Int. . Sci. 155: 772-1 — ——, J. A. Doy yle, P. 2003a. Archaefruc SACHEN rm precursor or specialized early ; 8: 36 K. Trines & Q. Leng. angiospe rm? Trends PI. Sci. 369-373. R. Pedersen & J. Schónenberger. 2003b. new cele: producing plant genus Cretaceous of Portugal. 23. "uU gg of T affinity = the S te t. J. PL . 164(5, & ———— R. n Suppl): 201-52 2000: . Cretaceous um and Nes in nt angiosperm flowers: reproduction. Palaeogeogr. Palaeoclimatol. ens "c ol. 292: 251-292. ————, —— & J. Se hónenberger. 2000b. sem plants: A m nt Pp a ol e Cretaceous rosid divereuieation. Pl. Syst. Eve a 14 n — Ci rane, T edersen. S ‘Be "ngtson, P. C. J. "rep . W. Grimm p M. Stampanoni. 2007. Phase- contrast x-ray mie EUR Sa inks Cr erace eous seeds ith Gnetales and Bennettitales. Nature 450: : 552. Gandolfo, M. A., K. C. Nixon & W. L. Crepet. 19 n A new Jersey: d J. Bot. fossil flower Don the Turonian of New ee gen. et sp. nov. (Capparales). Amer. 97: 964— ———,———— & 2002. Triuridaceae fossil flowers = the po Cretaceous of New Jersey. Amer. J. Bot. ): 1940— Tun ES M. ru iaa E the Jurassic rocks i es hire. Philos. Trans. . B. xm 15-98. The Yorkshire ii : Flora Caytoniales, rae pes PEA rms. British v um. (Natural Mr lor The Yorkshire Jurassic Flora HH. ins, D Museum (Natural History). Londo A. Gandolfo, K. C. Nixon & W. i. e y Bennetti- Hermsen, E. J., M. 2003. Divisestylus gen. nov. (aff. Iteaceae), a ossil saxifrage from the Late Cretaceous of New Jersey, USA. Amer. J. Bot. 90: 1373-1388. Nixon € W. L. Crepet. 2006. The impact of xtinct taxa on unde das the early evolution of angio- sperm clades: An example inc p e reproductive yst. Evol. 260: 141—169. structures of Saxifragales. Pl. S Volume 95, Number 1 Endress 119 2008 Floral Architecture and Organ Shape Hufford, L. 1996. The origin and early e of Ren, Y., H. . Zhao & . Endress. 2007. Floral angiosperm stamens. Pp. 58-91 in W. G. D'Arcy € R. ja d nesis in n fup a | re Ranunculales). C. Keating (editors), The Anther: Form, A unction, and Ann. Bot. 100: al Cambridge Univ. Press, Cambric Rohweder, 0. eo C om rmen-Studien 4. Morphologie & Endress. 1989. The etis of anther und Anatomie der B men Friichte a i amen bei structures mi dehiscence on among Hamamelidi- Alsinoideen und P s.lat. (Can y ae). dae. Bot. Ann. Soc Bot. Jahrb. Syst. 90: 201-271. Igersheim, a & P. Prides. 1998, Gynoecium dive "sity Rothwell, G. W. & I sioe Kex 2002. Anatomically and systematics of ER paleoherbs. Bot. J. Linn. Soc. 289—370. 2001. diversity and systematics in basal monocots. Bot. J. I Buzgo € P. K. Endress. Gynoecium Soc. 9: jð: Kirch off. y A primordium shape, and flower orientation. Int. J. Pl. Sci. 64: 50: -517 Klavins, S. D. n M. Krings & T. N. Taylor. 2003. Gymnosperms CN KE middle Triassic of Antarctica: The J Shape matters: Hofmeister’s rule, first T pres served cycad pollen cone. Int. J. Pl. Sci. 164: 1007-1020, Kuijt, A 1969. The Biology of Parasitic Flowering Plants. Univ. of California Be Berkeley Matthews, M. L. & . Endress. 2002. Comparative floral structure and systematics in Oxalidales (Oxalidaceae, Connaraceae, Cephalotaceae, Brunelliaceae, Cunonia- ceae, BA Tremandraceae). Bot. J. Linn Soc. 140: : a & — ieee teal floral structure and systematics in Cue a Coriaria- eae, Datiscaceae, Tetrame = 'eae, Begoniaceae, Anisophylleaceae). Bot. J. rynocarpaceae, Cucurbi- taceae, Linn. Soc. 145: 129-185. & ———. 2005a. Comparative floral structure and Celastrales o B A Bot. J. Linn. Soc. 149: & systematics in . Parnassiaceae, 129-194. 005b. Gace floral structure and systematics in ais ales (Crossosomataceae, Sta- Geissolomata- J. Linn. Soe. chyuraceae, Staphylea ane ae, ceae, Ixerbaceae, Bun es 'eae). Bot. 147: 1-46. . 2001. A s of Anisophylleaceae and -, J. Schónenberger € E. M. Friis comparison of floral structure Cunoniaceae and the aad a their systematic position. Ann. Bo Mundry, L& M. Mundry. 2001. Male cones —An a of Wettstein's pseudanthium concept. PI. Biol. 3: 405—416 A Suitzel. 2003. Morphogenesis of male sporangiophores of Zamia md D.W. Stev. Pl. Biol. 5: 297-310. 1954. Gesetzmiissigkeite in Taxaceae s.l. Nelson, E. im Blütenbereic h—ihre Bedeutung idis das Pr = er Gestaltwandlung blem der ernex-Montreux, Switzerland. Morphol and the evolution of PL. Res. : 479-492, Anatomy E relationships of petrified Taxodiaceae, an Sciadopityaceae. J. PL f Pedersen, K. R., P. R. Crane € E. M. Friis. 1989. The morphology and phylogenetic Harris (Bennettitale 's). Rey Perkins. J. 1925. significance of cree . Palaeobot. Palynol. 60: 7-24, Ubersic h t über die Canones Pn. der _Monimiac cae. Engelme elpzig . K. B. 1983 us) B The A and reproductive biology x1 the sillata 600—613 cone Mazocarpon. Bot. Gaz. 144: preserved Coue] (C ation of systematic charac E rs for the seed Bennettitales. we J. Bot. 89: 1447-1458. Rudall, P. J. & C. A. Furness. po . Systematics of Acorus: Ovule E ales Int. J. Pl. Sei. 158: 640—651. Sahni, B. 1 . The Pentoxyleae: A new g f Jurassic ou] gymnosperms ud the Rajmahal Hills of "dia. Bot. Gaz. 110: 47-80. Savolainen, V., M. F. Fay, D. C. , K. M. Cameron, S. Albach, A. Backlund, DN E D. m Whitten, K. J. o & M. 3 . 2000. Phylogeny of the eudicots: A nearly aa analysis based on rbcL gene sequences. Kew Bull. 30 complete 55: 251—309 Schnei E T, E. Te 1978. Morphological studies on the a IX. D seed of Barclaya longifolia Wall. Bot. . 139: 223-2: co Je pen F lori structure, development and diversity in Thunbergia (Acanthaceae). Bot. J. Linn. Soc. 130: 1-36. 2005. Rise from the ashes. Reconstruc tion € charcoal fossil flowers. 436— Trends Pl. Sei. 10: — é Conti, 2003. Molecular phylogeny uid foral evolution of Penaeaceae, Oliniaceae, Rhynchocalycaceae Amer. J. Bot. 90: 293-309. and Alzateaceae (Myrtales). & P. K. Endress. 1998, flowers in Mendoncia, Structure and development of Pseudocalyx and Thunbergia vule de and their systematic implications. Int. J. PI. Sci. 159: E 465. -. R. Pedersen & E. M. Friis. 2001. Normapolles ower a foe an , afinig from the Late Cretaceous of . 226: 205-230. en Contributions to Pennsylvanian paleo- m = gal. Se de a M botany: Mazocarpon b ives sp. nov. and sigillarian Rods Illinois Geological Survey, Report of Investigations 75: 1-40 Semple, J. C. 2000. Giedi: pappi in the goldenasters, subti Asteraceae: Astereae). Sida 22: 503-531. Sharma, B. D. 1982. Morphology of interseminal se pus in Williamsonia Carr. (Bennettitales). Ind. J. Earth Sci. 9: 1—5. Singh, V. € A. Singh. 1985. Flor: ipia empor mpa double, and simple Chrysopsidinae organogenesis i|! 61-67. bonn udo) 35: Smith, A. C. The families Hliciaceae and Sehisan- draceae. “Sane ntia 7: 1-224. Sokoloff, D. D.. C. V. Degtjareva, P. K. Endress, M. V. pon XH Samiguilin & C. M. Valiejo-Roman. 07. Inflorescence and early flower development in n a phylogenetic and taxonomic Ser. 168: 801-833. H. Weston & P. K. and floral phyllotaxis in aurales). Int. J. Pl. Sci. 168: 285-306 2007. A 2 of Rivea (Convolvulaceae). bd, a NN 213 19 Me "SOZOLE 2007. Floral Calycanthaceae Endress. EL mE a Lp = ¢ - s OR ae Pl. Res. 107: 493 and sysle matic re nais i J. 120 Anatomically preserved . W. Rothwell. 2003. W mde (V i = = Evidence for bennetti- talean reproduction in the Late Cretaceous of western North 2d rica. a J. PL Sci. 164: 251-202. . Dileher, S. Zheng, K. C. Nixon & X. 2 re n refructaceae, a new basal angiosperm pen ience 296: 899—904. . Praktische Einführung in die Pflanzenmor- Un PES phologie 2 Fise ne 984. ( e a N T n flowers. Pp. 351-395 . Dic Hon (e nor. Contemporary pex O in R. A. White & W. Proble ans in Plant ba Pu Academic Press, Orlando. von Balthazar, M. & P. K. Endress. 2002. Development of orescences and e rs in Buxaceae and ne to m of Ser. 163 mr! inf perianth 2 tation. Int. J. PL Sei. 103: 847- , G. E. Schatz & P. K. 2003. Female Ing and in e scencees E Didymelaceae. Pl. Syst. Evo 208. 199— Endress. W. S. Janka. 2006. Structure v evolution of ihe an Schónenberger, Alverson, H. --—À A. Baum. androecium in the Malvatheca clade alvaceae s. implications for Malvaceae and sad s. Pl. Syst. Evol. 260: 1 ea . Peder P. V Pinnae Pa blan gen. a new Lauraceae retaceous (E ay Lo y Mile Albian) of eastern North ad ica. Amer. J. Bot. 94: 2041-2053 Wang, L. € Y.-Z. Wang. 2005. vp qs : eronicaceae) and phylogenetic implica- lions. Pl. Syst. Evol. 250: Ap Crane & E. M. Friis. 2007. sp. nov flower from the Early Floral development of ~ 9-76 nes P. J. Rudall & M. W. Frohlich. 2005. ani ‘ntal control of sepalness and petalness in waterlilies. XVII. International Congress, Vienna. Ab- stracts Heo — —, Differentiation of perianth organs In ds ES s. Taxon (in Weber, 1993. Struktur. der Blüte von Me arvensis. Österreich 130: 1 pr (CSS Antheseverlauf und Bestäubung Verh. Zool.-Bot. Ges. APPENDIX 1. List of plant specimens, collectors, and collection numbers used for original illustrations. Col ection dates are provided in cases without a collection number. All vouchers deposited in Z Acsmithia davidsonti (F. Muell.) Hoogland, Cunoniaceae, ne 1919 1. K. Irvine 1212 Annals of the Missouri Botanical Garden Artabotrys hexapetalus (L. l.) Bhandari, Annonaceae, P. K. ndress 52 Barringtonia calyptrata R. Br. ex Benth., Lecythidaceae, P. K. Endress 4300 Cananga odorata Wook. f. «€ Thomson, Annonaceae, P. A Endress 1134 capa ey I... Betulaceae, P. K. Endress 925 (Figs. FH. . H), 3794 (Fig. 13D-F), 3795 (Figs. 12A-H, 13A-C e 1908 (Vig. 15H) ou silvana Schltr, Salicaceae, P. K. Endress 6019 Couroupita guianensis Aubl., Lecythidaceae, P. A. Endress 0305 Cyathocalyx martabanicus Hook. f. & Thomson. Annona- ceae, P. K. Endress 9372 Euptelea prot S Sie tools | & Zucc.. Eupteleaccae, PORE & Endress s.n.. summer | koihencilla major Lodd., Hamamelidaceae, P. K. Endress 717 Muell.. Cunoniaceae, P. K. Endress k biagiana F. $ - Gillbeea a F. Muell.. Cunoniaceae. P. A. 12 Endress 427. Hamame lis virginiana l.. Hamamelidaceae. P. A. Endress 23 IX 1988 biens parvifolia Sm., P. K. Endress 97-11 K. Urmi Olacaceae. Hydrocharis morsus-ranae |... Hydrocharitaceae, König s.n.. s.d. P. K. Endress 9504 P. K. Endress Mayaca sp. indet.. Mayacaceae, nore spegazzinii Pirotta, Leguminosae, 723 Munroidendron racemosum (C. N. Forbes) Sherff.. Aralia- ceae, P. K. Endress 90-11 Piper augustum Rudge, Piperaceae, P. A. Endress 7890 . Platanace 1C P ou eae. P. K. Endress Poly gonum nie mobs, va (e Sphenostemon pei (F. Muell.) S. Sm. Spheno- P. ess 9205 \canthaceae, P. A. Endress 7871 \canthaceae. slemonaceae, . Endr Thunbergia d Sims. Thunbergia he A Roxb. no voucher (photo of flow Trichoc ludus Rauh & Schlieber Wilkiea sp. ce erin Oliv.. Hamamelidaceae., Mia Br. M. Hyland 10127 A FOSSIL RECORD FOR GROWTH REGULATION: THE ROLE OF AUXIN IN WOOD EVOLUTION' Gar W. Rothwell,” Heather Sanders,? Sarah E. Wyatt,” and Simcha Lev-Yadun? ABSTRACT Appreciation for the role of m ny in plant evolution has been heightened by advances in studying eels using ulatory, p molecular techniques, with a gro diochemical mechanisms by iid th and or extending that understanding to t — — that developmentally diagnostic features can be identified in the fossil record, where The first paleontological evidence for the ian of cambial activi mediated regulatory pathways. 1e ontogeny and evolution of whole organisms through time. Recent stu ving number of specific structural features now understood in terms of the genetic, they are produced. Paleontological approaches to plant development provide a vehicle es have shown they Ere fingerprints for gen e polar axial flow of auxin consists of circular patterns of tracheary elements above buds and branch junctions in the id of the 375-million- year-old fossil progymnosperm Archaeopteris Dawson. That evidence strongly su orts homology of se nas vascular tissues in progymnosperms and seed plants, and monophylesis of the lignophyte 1 have now been identified arborescent i tes that belong to ids pendent lineages, co that a ale mechanism involving the reg anatomical patterns at the same positi of secondary vascular tissue production by the polar axial flow ophy tes, lyc a that the wood in lignophyt of secondary tissue production a auxin in each clade and Crane (1997): oe combining future clades of Paleozoic plants sensu Kenri flora and illuminate promising avenues substantially impact our unde ade of the role « Key words: Evolution, fossils, polar auxin flow, A ytes, and equisetophytes origi WEE inc onj unction with the parallel evolution of regulation .T al ee sois spiral vascular tissues, * (progymnosperms and seed plants). Similar od of -size in MS fossil equisetophytes and ulation of auxin characterizes those clades as well. These DE imply independent origins of s ei vascular tissue in three m paleontological studies with molecular and genetic studies to end ee te in vascular plant evolution wood development, xylem The rise of classical genetics during the 1920s and 1930s fostered a fundamental advancement in Dar- winian evolutionary theory, with the resulting modern synthesis placing biological evolution on a solid genetic footing (Gould, 1994). As a consequence, the focus of the study of evolution shifted from organisms to populations of organisms, and the definition of evolution was modified from changes in the structure of organisms through time to changes in relative gene frequencies through time (Gould, 1994). This modification of focus is appropriate because it is DNA (not individual organisms) that is modified through time to provide the continuity of biological Another of the synthesis was the separation of studies of evolutionary evolution. consequence modern patterns from studies of the evolutionary process. Molecular phylogenetics notwithstanding, evolution- ary patterns are established primarily by changes in organisms through time (a major role of paleontology; Kenrick & Crane, 1997), whereas studies of evolu- tionary process focus primarily on the population genetics and the population biology of living organ- isms (Howell, 1998 Recent advancements in developmental genetics and the rise of molecular developmental biology have, for the first time, allowed us to understand structural features in light of the regulatory mechanisms by which structure is patterned. For example, Reinhardt et al. (2003) have shown how auxin accumulation in the shoot apical meristem stimulates leaf outgrowth and thus controls phyllotaxy. Various works are beginning to reveal gene expression patterns that give rise to morphological variation in leaves os et al., 1996; Hareven et al., 1996; Chen et al., ; Kim et al., 2003; Gleissberg et al., 2005; b. et al., 2005). Because we are now able to ascribe particular structural features of plants to the influence of specific genes, ontogenetic studies from the fossil record have the potential to combine with knowledge of molecular developmental pathways and allow us to infer much more precisely the evolution of tissues, organs, and whole organisms through time. Although at first glance ' We thank William Crepet, Cornell University, for inviting the symposium presentation from which this paper de svelope ad. Seve oral f figures wes nm this paper are deri a nd Wilson N. Stew . This study was dd in part by the National S GWR) and a MORPH ae search Coordination Network Cross-Disciplinary Training Grant awarded t ? Department of Environmental and Plant Biology, Ohio University, Athens, Ohio 45701, U.S.A. ed from drawings originally prepared by Donald A. Eggert, Kathleen B. Science Foun E EF-0629819 to p lah; * Department of Biology, Faculty of Science and Science Education, University of Haifa-Oranim, Tivon 36006, Israel. 10.3417/2006208 doi: ANN. Missouni Bor. GARD. 95: 121-134. PUBLISHED ON 11 APRIL 2008. 122 Annals of the Missouri Botanical Garden paleontology seems to offer us scant potential for inferring plant development at the genetic and biochemical levels, a closer look exposes a sound theoretical and empirical foundation for the approach (Gould, 1977; Rothwell, 1987; Stein, 1993; Sanders el al., 2007) and illuminates an opportunity to reincorpo- rate evolutionary patterns from paleontological studies mainstream of evolutionary process. As into the compared to most animals, the potential for the excellent. preservation of sequences of developmental processes in fossil plants is common because the plant skeleton occurs at the cellular level (in the form of the cell wall) instead of at the organismal level (as in most animals). Consequently, virtually all plant cells are readily preserved in the paleontological record (Stewart & Rothwell, 1993; Taylor & 1993) available to studies of developmental biology. Taylor, and The fossil record of vascular plants provides an impressive array of evidence about the ontogeny of tissues, organs, and whole organisms, as well as for the evolutionary pathways by which organs and species of organisms have evolved (Stewart & Rothwell, 1993; Bateman € DiMichele, 2002; Kenrick, 2002). Tradi- tional ontogenetic studies of fossil plants emphasize topies such as the development of shoot systems 1964: Scheckler, 1978), growth architec- sporophytes (Eggert, 1961: 1993; Tomescu et al., 2006), primary and secondary meristems (Good & 1 1972; Cichan, (Rothwell, 1971), and microgametophyte development (Millay € Eggert, 1974; Taylor € Rothwell, 1982). Other paleobotanical studies model apical growth (Stein, 1993), meristematic activity deduced from vein patterns (Boyce (Delevoryas, ture of Trivell, l'avlor, 1985), ovule ontogeny infer leaf blade development in relation t & Knoll, 2002), hypothesize regulatory mechanisms for vascular differentiation 1993: Boyce & Knoll, 2002), and explore the evolution of meristematic activity (Stein, in relation to. structural changes brought about by experimental studies of gene activity in living plants 2007). These strategies for inferring ontogeny from fossils (Sanders et al., recently have been extended to the investigation of developmental regulatory mechanisms by which the fossilized structures were produced (Rothwell & Lev- Yadun, 2 es introduced 005). Together with the theoretical approach- above, that study illuminates the exciting potential for developmental regulatory mech- anisms to be recognized and characterized from the paleontological record. In the present study. we examine and provide an example of the broad potential for fossil plants to provide developmental regulatory data and to infer the evolution of tissues. organs, and whole organisms. Paleozoic fossil lycophytes and equisetoplhiytes has been examined to determine if specific evidence for The secondary xylem of the regulation of cambial activity by polar axial auxin flow circular tracheary elements) extends beyond the lignophytes to other euphyllophytes and (6.95 to lycophytes (Fig. 1). Circular patterns of tracheary elements are present in lignophytes, other euphyllo- phytes, and lycophytes but v vary considerably in the extent and owe of their development. These data are evaluated in relation to the known positions of deviations of the direction of the polar auxin flow in 1990), with the goal of interpreting the evolution of regulatory living vascular plants (Lev-Yadun & oni. mechanisms for secondary vascular tissues across the major fossil clades of vascular land plants. MATERIALS AND METHODS Anatomically preserved fossils of woody lycophytes carbonate concre- 2002) were sectioned in tangential view at nodes using the well- and equisetophyles preserved in tions known as coal balls (Rothwell, known cellulose acetate peel technique (Joy et al., 1950) to branches expose the patterns of tracheary elements above and wide leaf-trace rays. These include the Middle 310 million years old) lycophyte plant Paralycopodites brevifolius (Williamson) DiMichele from the Lewis Creek locality (Phillips, 1990). 295 million. vears old) lycophyte Pennsy lvanian (ca. in eastern. Kentucky the Upper Pennsylvanian (ca. plant Chaloneria cormosa Pigg & Rothwell from the Phillips, 1980), 310 million years Steubenville locality in eastern Ohio ( Middle Pennsylvanian (ca. of the from the and old) Arthropitys Goeppert species equiselophyte morphotaxon Lewis Creek locality. Peels were mounted on microscope slides under cover Kinder, vermany: distributed by Calibrated Instruments, Inc. slips in the mounting medium Eukitt. (QO. ~ Hawthorne. New York) and examined using transmit- ted light. with Microlumina (Leaf Systems, and/or reflected Images were captured Ine., Southborough, Massachusetts) and Photophase (Phase One Indus- tries, es Denmark) digital scanning cam- eras mounted on a copy stand or a Leitz Aristophot (Berlin, using Adobe Photoshop CS (www.adobe.com). Spec- bellows camera, and processed -~ sermany) imens, peels, and microscope slides are housed in the Ohio University Paleobotanical Herbarium. Microscope 3,43 7—13,448 slides bear acquisition numbers lé RESULTS PALEONTOLOGICAL EVIDENCE FOR THE DEVELOPMENTAL REGULATION OF WOOD—INITIAL WORK Evidence for polar auxin regulation of vascular cambial activity recently has been recognized in the Volume 95, Number 1 2008 Rothwell et al. Fossil o for Growth Regulation 123 Lycophytes Euphyllophytes Other Euphylloph Isoetalean Clade er Euphyllophytes wood (375 mya) Equisetophytes Lignophytes Seed Plants * wood (360 mya) wood (375 mya) 380 mya, minimum +/- 410 mya, minimum Figure 1. e aes d Euphyllophyte relationships reflect s 999: Pryer et al., 2001). Minimum ages ws ol des t eviden lycophytes from euphyllophytes occurred at least => or diverge nce of ma ajor « t 35 million years b * z Progymnosperms Simplified phylogenetic tree of vascular ed d currently recognized relationships among lycophyte Rothwell " ac de 's are placed at relevant nodes. Hz da narks on branches nsus of most recent competing hy pothe "ses nce for secondary growth in lycophytes, e ue. and lignophytes. Note that the divergence of efore evidence for the origin of secondary growth in either clade and that lignophytes also appear to have diverged from equisetophytes long before the appearance of secondary vascular tissues in either 1999: and Pryer et al., 2001.) 375-million-year-old tree trunk Callixylon whiteanum Arnold (Rothwell & Lev-Yadun, 2005). whiteanum is a morphospecies for woody stems of the ixylon a tree Archaeopteris Dawson (Trivett, angential sections of C. whiteanum wood show circular patterns of tracheids immediately above branches (Fig. 2A) that are comparable to patterns in living seed plants (Rothwell € Lev-Yadun, 2005). Because the circular patterns of living plants are pe produced by interruption of polar auxin flow a in the conifers (Sachs & Cohen, 1982), such patterns may be JOVe an obstruction cambium of woody dicots and regarded as a structural fingerprint for the regulation of vascular tissue development by polar axial auxin flow (Hejnowiez € Kurezynka, 1987; Lev-Yadun & . (Data primarily from Stewart & Rothwell, 1993; Taylor & Taylor 1993; Kenrick & Crane. 1997: Rothwell, Aloni, 1990) in Callixylon Zallesky and other fossils as well (Fig. 2B) ISOETALEAN (RHIZOMORPHIC) LYCOPHYTES Giant, much-branched, tree-sized relatives of the modern quillwart /soetes L. were dominant canopy trees in the equatorial wetlands of the late Paleozoic Era (360—300 million years ago [Ma]; Stewart & Rothwell, 1993). Such plants are commonly assigned to the Lepidodendrales of the isoetalean clade (Stewart & Rothwell. 1993) and are referred to as rhizomorphie because their much-branched rooting system (i.e., rhizomorph) is actually a shoot modified for rooting (Rothwell & Erwin, 1985). In a pioneering 124 Annals of the Missouri Botanical Garden Figure a. Brane ih stele (Br) ii ection surrounc ross se ed by wood (W) of stem, showing circular | SOR HOS ; M i ak MACH Na 1 VA Y Y d e MW Ses N i VEINS Secondary xylem of a Callixylon whiteanum Arnold stem in tangenti al section at the level of branch divergence. allerns of tracheids scd by uid a of polar auxin posi and pool ing of auxin above branch. —b. Enlargement of ond distal to pine arranged tracheids of the primary xvlem (PX) of the diverging branch stele. showing distorted tracheids and 1 circular patterns. 1961) documented a complex mode of primary and second- study of development in fossil plants, Eggert ary growth for the vascular tissues of lepidodendralean lycophytes (Fig. 3A). Eggert determined that the shoot system of lepidodendralean trees had a primary vascular system that increased in size up to the level of the (Fig. xylem (i.e., Fig. 3B) stem). with the apical regions of the branches having 3B). opmental pattern demonstrates that the wood is the first. branch and then decreased distally 3B). By contrast, the thickness of the secondary wood: zone with horizontal lines in decreases distally (i.e.; from the base of the no secondary. vascular tissues (Fig. This devel- product of secondary. growth. similar to that of living seed plants in some respects but different in others 1985). growth (Cichan, Therefore. primary growth and secondary are products of independent meristems in arborescent rhizomorphic lycophytes (Eggert, 1961; 1985). Up however, there has been no definitive evidence for the Cichan, lo the present, structure or developmental regulation of a lepidoden- dralean-type vascular cambium. (Williamson) | DiMichele. Paralycopodites Morey & Morey is a stratigraphically Paralvcopodites brevifolius long-ranging genus of arborescent rhizomorphic lepidodendralean lycophytes that is roughly similar to the plant reconstructed in Figure 3A. Paralycopodites trunk bisporangiate cones that terminate lateral branches has a long and much-branched crown, with D and a highly branched stigmarian rooting system cells forming (DiMichele, 1980 branching and numerous deciduous lateral branches in but with infrequent dichotomous the crown as compared to the plant in Fig. 3A). Branches of Paralycopodites that have steles in the range of 2—3 cm in diameter (Fig. 4A) display a large pith surrounded by a medulated protostele of primary tracheids with a smooth outer margin, and abundant secondary xylem that com- — prises a relatively dense wood composed of tracheids 4B). At the level where a branch base diverges from the stem stele, the the stele of tracheids and adaxially oriented. parenchy- and infrequent narrow rays (Fig. xylem of branch consists of a C-shaped ma (Fig. 4B) that becomes a radial stele at more distal levels (DiMichele. 1980). stems near the periphery of the Tangential sections f p ] hi 0 1 aratycopoctces wood (i.e.. along line c-c of Fig. 44) show regularly arranged, long, straight tracheids and interspersed rays (Fig. 4B, at left to the branch trace narrow and bottom). Imme- diately distal the axial files of tracheids are disrupted, many tracheids are distorted, r horizontal ly (Fig. 4B, at upper right). By contrast, tracheids immediately and others extend obliquely o below the branch form regular axial files that are similar in pattern to those in non-branching regions (Fig. 4B. at bottom right). Chaloneria cormosa Pigg & Rothwell. Compared to Paralycopodites and other lepidodendralean trees, Chaloneria cormosa is a relatively small and un- branched rhizomorphie lyeophyte of Upper Penn- sylvanian age that grew to about 2 m tall (Fig. 3C). Volume 95, Number 1 Rothwell et al. 125 2008 Fossil Record for Growth Regulation > v Mba a Y a Mm ig AJ fs RD bi ae | « P? i > Ef j E: e 3. Diagrams of fossilized woody lepidodendralean and isoetalean lycophytes from the Pennsylvanian of North America and Europe. —a. Le pidodendreleat i tree ca. 8-10 m tall, showing elongated trunk and branching crown with terminal cones and ie i assignable to ihe morphogenus Stigmaria Goeppert. Modified from Stewart € Rothwell, 1993. —b. Diasrarmatié ía of growth architecture of lepidodendralean trees. Note that the size of the poa xylem (black = primary xylem; white = pith) increases up to first branch and then decreases distally with each subsequent branch. Also note that the thickness of secondary xylem (horizontal lines) decreases from base of stem distally. Modified from Eggert, 1961. Used with permission from Cambridge University Press. —c. Unbranched plant of the isoetalean aui. cormosa Pigg & R Mens a 2 m tall, showing cormose base, vegetative basal region of stem, and fertile apex. Modified from Pigg and Rothwell, 19 This species produced vegetative leaves toward the occurs at and near the base of the Chaloneria Pigg & base, displayed alternating zones of microsporangiate Rothwell plant (Fig. 4C), but the secondary xylem and megasporangiate sporophylls distally, and had a diminishes in thickness rapidly and is absent from all corm-like base similar to the living quillwort, /soetes but the most basal regions of the stem (Pigg $ (Fig. 3C; Pigg & Rothwell, 1981). Secondary xylem Rothwell, 1981). 126 Annals of the Missouri Botanical Garden Figure 4 PIA NOAA DITA i [71711 er e AER A Qs LA. UL ME z : p wi NC AES a 2 zn eae == TI == tih NS 2 a ay A Cr + + SR SR SI E fr 0, N | IAS > E SSA DEAS US PUMP > SS ELE IUE EL PU UD Ie = ANOS > SS MEAE pe hi l i Ni ^ DA Vi il NY LA 1 TAA IND ON iM AT adt VERA l il h Q8 W R RY ) y ie M ni aaa aN y ANOS Ras ENDS F] ES A e a 72 hoi y, CET s g s | [č ^ d do » [Y AN NV mS: z S oy A NA Nea ti Y, P». y] t ET) 8 | PES NN Ew RS NS UN ` NA ci s xod m v3 zi S, 77 SÍ l> P. d t ZN a / * q A a ad PON, TA fit = aX A Nd eN me AN ^ SAG MEL ce SS — US ft y» TA Ly IN. , à PPS à ES sS RD ES DU SAS no i y E M D R ATA A Y EU y x E Y NS ELSA D i E oa X ye IN E d A SERE. UN UE NOV EEG ON doa CANO See Mie Mi T DE VIO HO MEL PER I TA N == CM Or a N DM TUR. di s D Y ; a ADA RE a SN rA ERA O ES E e. 2 D > 25 AS 23 p BEE LA Lt Oe Sa cd (ASS es Sf) RK s NN NS Tks erate SN a SS Je SEDE T m LA ORE Se > EU 5 RITOS X= VA SZ o ABO Niraj. f is SVs se i polar auxin flow in regulation of secondary vascular tissue production. —a. Paralycopodites brevifolius (Williamson) DiMic stem from the lower Middle Pennsylvanian of eastern Kentucky, showing radial view throug surface. The pith region is surrounded by a cylinder of primary xylem (PX; also seen in c which is enclosed in a cylinder of wood (W). Note diverging branch stele (BS) at right. Line c—c indicates plane of sec shown in part « vel of a branching at coal ba Anatomical sections through stems of fossilized woody rhizomorphie lycophytes showing evidence for axial hele lion —b. Tangential section through secondary xylem of stem, cut through line e—e of part a. Note that the Volume 95, Number 1 2008 Rothwell et al. Fossil Record for Growth Regulation 127 Tangential sections near the base of a Chaloneria stem show secondary xylem that is perforated by helically arranged leaf traces (Fig. 4C). each within a broad ray (Fig. 4D). At the base of the plant (Fig. 4C). root traces diverge from the rounded bottom of the (Fig. 4E, at *RT"). traces at the base of the plant indicates that leaves woody zone An absence were not produced in that region (Fig. 4C, at bottom: AE). of Chaloneria Tangential sections through the secondary xylem reveal radial files of tracheids that follow a sinuous course around the large leaf-trace rays (Fig. 4C, D). In some areas, particularly imme- diately distal to each ray, the tracheids are highly distorted, with some tracheids being branched and others forming circular patterns (Fig. 4D). Highly dis- torted patterns. of tracheids are also present at the branched tracheids and base of the plant. where swirls of tracheids are common (Fig. 4E, at arrows). ARBORESCENT WOODY EQUISETOPHYTES Arborescent equisetophytes of the Pennsylvanian coal swamps were essentially giant woody Equisetum L.-like plants (Fig. 5A; 1975) that. produced that is assignable to three morphogenera including Arthropitys (Taylor € Taylor, 1993). h Good, wood cross sections, Arthropitys stems have a large pith surrounded by a bundles (Fig. 5B). the primary xylem and vallecular canals oceur in the 5B. 6A). both of identical to comparable structures in living species of Equisetum (Fig. 6D). By Arth ropilys stems range ring of primary xylem Conspicuous carinal canals characterize cortex. (Figs. which are virtually comparison Equisetum, much larger in diameter. Arthropitys stems also differ from Equisetum by the production of abundant secondary xylem that forms wedges of radially aligned tracheids separated by wide parenchymatous rays (Figs. 5B, OA: Stewart & Rothwell, 1993 At the level of a node, tangential sections through the secondary xylem of Arthropitys stems show branch bases that are positioned slightly above and on alternating radii with leaf traces (Fig. 6B). The branch base shows a large hollow pith that is surrounded by a canals and wood, all of which are ring of carinal similar to those in the stem from which the branch has of leaf arisen (cf. Fig. OA, B). As seen in tangential sections, the secondary xylem of the stem typically consists of regular axial files of long and straight tracheids and interspersed narrow rays (Fig. OB, sides) except where leaf traces and especially branch junctions are positioned. Tracheids of the secondary xylem extend around leaf traces (Fig. 6B, at bottom arrows), often — showing a small area of tracheids that form a circular pattern (Fig. 6C, at bottom left). Evidence of regulation of vascular differentiation by the axial polar auxin flow is more pronounced the where wood extends around the much larger branch bases (Fig. 6B). Cells distal to and extending around to the sides of the branch stele form distinct swirls and circular patterns (Fig. 6C), and there are much larger numbers of ray cells than in other areas of Also. he secondary xylem. some tracheids branch (Fig. 6B. arrow at top). SYSTEMATIC SCOPE OF POLAR AUXIN REGULATION IN Woop DEVELOPMENT wood assigned to Fossil the Callixylon (Fig. ZA, B) is produced by the plant Archaeopteris genus (Beck, 1960), which is a member of the Progymno- spermopsida, a paraphyletic grade subtending seed clade that is known as lignophytes ; Crane, 1985; Rothwell € Serbet. 1994). discovery of a common regulatory pathway for the plants in the (Fig. The development of secondary vascular tissues i = pro- gymnosperms and seed plants strengthens the hy- pothesis that such tissues are homologous and provides additional support for the definition of the lignophyte clade (Fig. 1). Although large woody trees with substantial secondary vascular tissue production he are restricted to seed plants in the modern flora, fossil record includes highly branched, woody equi- selophyte and lycophyte trees that were prominent components of wetland forests during the Paleozoic Era (Taylor & I 1993). Data developed in the current study reveal cellular , aylor, patterns in the secondary xylem of both Paleozoic lycophytes (Paralycopodites and Chaloneria) and equisetophytes (Arthropitys) that are diagnostic for regulation of cambial growth and secondary xylem production by polar axial auxin flow among living tracheids distal to the branch stele (BS) are distorted, whereas those below the branch stele : ot. —ec.' base of Chaloneria cormosa Pigg € Rothwell from the Upper Pei with large leaf-trace rays (LTR) distal to a cormose plant base. of the plant. —d. Enlargement of wood in stem region of p Note that no leaf l'angential section al insylvanian « of Ohio, (sc. he lically arranged leaf traces traces or rays traverse the wood (W) at de hase part e (W) Ar showing dive 'rglng leaf trace (LT) surrounded by a wide . Note contorted and circular patterns of tracheids distal to the LTR. —e. E | nlargement of part e at base of plant showing diste d and circular patterns of tracheids (at arrows) in the wood (W) above a dive ‘rging rootlet trace (RT) at the base of the p 128 Annals of the Missouri Botanical Garden <> NX 2 ee NLIS E S > ig; a5 Se Few] A = IDA RT SE See = Wass TINS SASS ‘NE A equiseto- Figure 5. Diagrams of fossilized calamitalean tH 3 e] z phytes from the Pennsylvanian of North America anc . Calamitalean tree +8 m tall, divergi underground rhizome. Modified from Stewart and Rothwell, 993 i ing ca. e —b. ¿ross section of characteristic Equisetum-like anatomy and secondary xylem. Important features include hollow pith (P), carinal canals (C) surrounded by wood, and vallecular canals (VC) in cortex. Modified from Stewart and Rothwell, 1993. seed plants. Distorted files of axial secondary xylem cells, branched tracheids, and circular patterns. of neids are associated with leaf trace divergence — trac in Chaloneria (Fig. 4C, D), reflecting swirling polar auxin currents above disruptions in the vascular cambium that are every bit as pronounced as those that result from obstructions to auxin flow in the cambium of living and fossil lignophytes (Fig. 2). These patterns document that axial polar auxin flow was involved in the regulation of wood development in lycophytes as well as in lignophytes (Fig. 1; Rothwell & Lev-Yadun, 2005). Likewise, the occurrence of 1e — — well-developed diagnostic cellular patterns in secondary xylem of Arthropitys demonstrates. the levelopment by the regulation of secondary xylem axial polar auxin flow in extinct equisetophytes. Together, these data from lignophyte, equisetophyte, and lycophyte taxa reveal that polar auxin transport facilitates a common mode of developmental regula- tion for secondary vascular tissues across the vascular p auxin regulation of wood development is common to ant evolutionary tree (Fig. 1). That is to say, polar both major clades of living vascular plants, Lycophy- tina and Euphyllophytina (which includes both equisetophytes and lignophytes: Fig. 1). VARIATIONS IN. Fosstt EVIDENCE FOR REGULATION BY AXIAL POLAR AUXIN FLOW 1 anatomical evidence for the regulation of Althoug secondary vascular tissue production by axial polar auxin flow is present in all three major clades of vascular plants that have produced abundant wood through time (ie, lignophytes, equisetophytes. lyco- phytes; Fig. 1), there are significant variations in the degree of development and in the locations of evidence among representatives of the various clades. In all of the examples studied to date, evidence for regulation of vascular differentiation by deviation from the typical orientation of the axial polar auxin flow is preserved in the secondary xylem above the positions where there would have been obstructions to auxin flow in the cambium during tissue patterning (Figs. 2. 4, 6). Lignophytes (Fig. 2B). such as the extinct progymnosperm Archaeopteris (represented here by the morphotaxon Callixylon. Fig. 2). living species of both conifers and dicotyledonous flowering plants (Rothwell € Lev-Yadun, 2005: figs. 2, 3), and Fig. 6B) all show obvious regions of distorted tracheids, branched My ossil equisetophytes such as Arthropitys tracheids, and circular patterns of cells distal to branch junctions. The equisetophytes Arthropitys and Archaeocalamites Stur also show evidence of circular patterns with leaf (Fig. 6B; Smoot et al., 1982), demonstrating. that small obstructions in polar auxin regulation were associated trace divergence sufficient to produce the diagnostic fingerprint in this clade by the Mississippian Period (ca. 330 Ma). Files of axial tracheids of the secondary xvlem distal to branches in Paral ycopodites are also disrupted above the branches, but the extent of the zone is much smaller and the distortions are far less pronounced (Fig. 4B. at Volume 95, Number 1 8 Rothwell et al. 129 Fossil Record for Growth Regulation This subtle top) than in the other plant clades studied. ' evidence for regulation of secondary xylem differenti- ation by the axial polar auxin flow in Paralycopodites is most easily recognized by noting differences in the wood immediately above and below a branch. By comparison to the subtly distorted patterns distal to a — branch, files of cells below the same branch (cf. cells above and below the branch in Fig. 4B) of Paralyco- podites show no evidence of distortion at all. Were it not for the obvious disruptions of cell files with branched tracheids and circular patterns of cells present in the wood of other lycophytes (e.g., Chaloneria; Fig. 4D, E). it would be difficult to interpret regulation of secondary xylem development by the axial polar auxin flow in fossil isoetalean lycophytes with certainty. Chaloneria cormosa is unbranched and smaller than lepidodendraleans (cf. Fig. 3). Chaloneria also has a cormose base like living /soetes, rather than the elongated and much-branched stigmarian rooting system that characterizes Paralycopodites and other members of the Lepidodendrales (Fig. 3A). Neverthe- less, evidence for the regulation of secondary vascular tissue development by the axial polar auxin flow is well represented in Chaloneria (Fig. 4C-E). Disrupt- ed cellular patterns, branched tracheids, and swirling patterns of tracheids occur in association with the that (Fig. 4C, D). Because of the extremely short inter- wide rays accompany diverging leaf traces nodes and numerous helically arranged leaves of Chaloneria (Fig. 4C), the closely spaced wide leaf- trace rays must have produced a considerable obstruction to the axial polar auxin flow in the developing vascular cambium. That complex pattern of obstruction probably accounts for the occurrence of contorted and branched tracheids around several sides of many rays (Fig. 4D). One of the most distinctive and initially unexpected occurrences of an axial polar auxin flow fingerprint in fossils is found at the base of the Chaloneria plant, where particularly. well-developed circular patterns — are present (Fig. 4C, E). Such patterns were first recognized as being distinctive and bilateral in distribution before the role of the axial polar auxin flow in regulation of secondary xylem development 1979). those cireular patterns are present in a region that is was understood (Pigg & Rothwell, Because below all typical obstructions in polar auxin flow that would have been caused by the divergence of branch junctions and leaf traces, their occurrence must be attributable to another phenomenon. We know from wounding and partial girdling experiments in oak (Quercus ithaburensis Decne. and Q. calliprinos Webb) that circular vascular tissues are also formed above any large obstacle to the polar axial auxin flow in tree trunks (Lev-Yadun & Aloni, 1991). An understanding of the unique morphology of rhizomorphic lycophytes provides a plausible explanation for the formation of the circular tissues at the base of Chaloneria plants. Although some questions about the homologies of the rooting system of /soetes remain among authors who work primarily with living species (Yi & Kato, 2001). there now is overwhelming evidence that the bipolar growth and rooting systems of isoetalean lycophytes are unique among all vascular plants. i2 ;vidence from the developmental morphology of Isoetes (Karrfalt & Eggert, 1984) structure, and embryology of fossil species (Frankenburg & Eggert, 1969; illips, 1979; Karrfalt, 1980) supports the hypothesis that isoetalean lycophytes have no true roots that could be and from the development, homologous to the roots of Selaginella P. Beauv., other non-rhizomorphic lycophytes, or any other group of vascular plants (Rothwell & Erwin, 1985). Rather, the rhizomorph of /soetes and all of the isoetalean = lycophytes is homologous to a shoot system that has become highly modified for rooting the plant. This means that the structures commonly referred to as the roots of /soetes are actually leaves that have been modified for rooting (Frankenburg & Eggert, 1969). That interpretation is both explained by and further strengthened by the distinct evidence for disruption in polar auxin flow at the very base of the Chaloneria plant. If, indeed, Chaloneria and other rhizomorphic lycophytes have no true roots, as interpreted from the rt, 1969; Rothwell & Erwin, 1985), then basally flowing auxin would fossil record (Frankenberg & Egge have nowhere to go after reaching the cormose base of the plant. The auxin must have pooled and swirled in thus inducing the distinctive anatomical patterns in the the cambial zone at the base of the trunk, wood at the base of Chaloneria specimens (Fig. 4C, E; Pigg & Rothwell, 1979). Anatomical descriptions of the basal regions of other fossil rhizomorphic lycophytes with cormose bases are scarce. However, in a genus that is even more ancient than Chaloneria (i.e. the 360-million-year-old Lower Carboniferous lycopohyte Trabicaulis flabellilignis Meyer-Berthaud; Meyer-Berthaud, 1984), there is evidence for disrup- tions of tracheid patterning similar to those of Chaloneria both above leaf trace divergences (ib. 1984: fig. 8) and at the base of the Meyer-Berthaud, 1984: Planch 3, fig. 4). These "m reveal that regulation of secondary Meyer-Berthaud, plant (i.e., vascular tissue production by the axial polar auxin flow was probably widespread among rhizomorphic lycophytes with cormose bases. Moreover, there is now additional compelling evidence that the most distinc- live synapomorphy for isoetalean lycophytes, the rhizomorph, represents a shoot system that has been modified for rooting. 130 Annals of the Missouri Botanical Garden um P $ ^ rj ¿0% s 23 etn te Biv Sateen 5 ES OOO red, s ; "VA IAE B: : Uy, Wi iy; ARE S 7 2 AY Y SS y. Gi M S p AS SN SS ROS, Y v» $ 2s > " d =) t} de Ne? Ne Ñ N At "t NN A 5- C DL qf ge AE N QNS ANN > ANNAN ES MU ANSE TO a INES) \ Min TAA 1 NA WN) FN Waa Nie ein aai Sa Y Ñ bid A JI a ^ lj LU n gru 3 y on) W "UN M i 2» sl i A 1 WE j P AR is di AUT AA SW 4 Figure 6. Anatomical sections through stems of large, woody equisetophytes assignable to the Paleozoie Calamitales (ae o D T J | I Ñ D and living Equisetum L. (d). —a. Cross section through segment of Arthropitys Goeppert stem from Westphalian A sediments of g Eg : 8 : PI | England. with pith at right and cortex at left. Carinal canals (CC) in the primary xylem and vallecular canals (VC) in the cortex that are comparable to those of living Equisetum (d). Note the well-developed segments of secondary xylem that are separated BEY | B y X) | by wide pith rays. This wood (W) shows both long radial files of secondary tracheids and interspersed narrow parenchymatous rays. —b. Tangential section through wood of Arthropitys at the level of a node. Diverging branch base in cross section shows a Volume 95, Number 1 2008 Rothwell et al. 131 Fossil Record for Growth Regulation FossiL EVIDENCE FOR GROWTH REGULATION IN THE EVOLUTION OF PLANTS Methodology for recognizing specific developmen- tal regulatory pathways from diagnostic structural features of fossil plants has recently begun to emerge as an additional approach for improving the under- standing of evolution through geological time and for testing hypotheses of evolution and phylogenetic relationships among major clades of vascular plants (Rothwell & Lev-Yadun, 2005). This represents a new approach, the first application of which is based on our knowledge that there are disrupted files of that patterns immediately distal to branches and = contorted and branched cells form circular — arge leaf-trace rays in living seed plants. Because those patterns are known to represent an anatomical fingerprint for the axial polar auxin flow that regulates a ea patterns of secondary vascular tissue development (Sachs & Cohen, 1982; Hejnowicz & Kurczyñska, 1987; Lev-Yadun & Aloni, 1990), the discovery of circular patterns in the wood of a 375-million-year-old species of Archaeopteris has provided the first direct evidence that progymnosperms and seed plants share a common mechanism of growth regulation for secondary vascular tissues. Those data also reveal that auxin-mediated secondary vascular tissue devel- opment is a synapomorphy for lignophytes (i.e., a clade that includes seed plants and their progymno- spermous sister groups). IMPLICATIONS FOR THE EVOLUTION or Woop Across VASCULAR PLANTS These newly developed data for the fossil evidence for regulatory mechanisms of wood production now can be combined with an understanding of phyloge- netic pattern and first occurrences. of secondary vascular tissues in divergent clades to develop hypotheses for the evolution of secondary vascular issues across the vascular plant evolutionary tree. Although the overall topology of vascular plant remains a subject of controversy Nixon, 2006), there is wide agreement of the tree. Living plants and their closest relationships (Rothwell & about some of the deep internal nodes evolutionary extinct sister groups comprise two major clades, Lycophytina and Euphyllophytina sensu Kenrick and Crane (1997), can be identified as early as the some representatives of which Late Silurian. This places the minimum age for divergence of lycophyte ancestors from euphyllophyte ancestors at about 410 million years, which is at least 35 million years earlier than the first oceurrence of secondary vascular tissues in either clade SE Fig. 1). Therefore, secondary vascular tissues in lycophytes and euphyllophytes must have evolved independently. Evidence for separate origins of secondary vascular tissues in equisetophytes and lignophytes is not as strong because of the current uncertainty about relationships Rothwell € Nixon, the first evidence for secondary among euphyllophytes (Fig. 1; 2006). vascular phytes is on the order of five million years younger than the minimum age of divergence between these However, n either equisetophytes or ligno- tissues two clades (Fig. 1). Therefore, it is also likely that secondary vascular tissues evolved independently the ancestors of Equisetum and seed plants. We now recognize that the developmental mecha- nism (Le. involving polar axial auxin flow) for secondary vascular tissue production is common to lignophytes, equisetophytes, and lycophytes, but that the secondary vascular tissues evolved separately either in lycophytes and euphyllophytes or in all three wood production in the A The question remains whether Therefore, clades (Fig. 1). major clades of vascular plants is the product convergent evolution. the underlying regulatory mechanism (i.e., by the axial polar auxin flow) for secondary vascular tissues also evolved independently in the various clades or whether it was present in the genomes of all of the clades prior to the evolution of secondary vascular tissues. Among living plants, this question has been addressed in a molecular survey of embryophytes and their algal sister groups (Cooke et al., 2002). Although conjugation rates and the complexity of both increase toward the distal branches of the tree, that indoleacetic acid (IAA) metabolism is present in the auxin metabolism these authors record bryophytic sister groups of vascular plants as well as all of the lycophytes and euphyllophytes that have Within the tracheophytes, auxin been surveyed. hollow pith (P) surrounded by pith parenchyma and a ring of carinal canals (arrows at CC). Branch is situated just distal to, and f on a radius that alternates With, radii apon which lea tracheids (indicative of the role of the positioned distal to the branch, but are not produced below the branch. Arrow at top of photo identifies branching trac traces (arrows at L axial polar auxin flow in regulation of secondary vascular tissue produc S are ) are produced. Note that circular patterns o —c. Enlargement of disrupted wood pattern distal to branch in part b, showing swirls and cire E e rns dus cells. a Equisetum arvense L. stem in cross section, showing anatomical features (CC = carinal canal, P ith, VC = vallecular canal) that are mi identical to the primary tissues of Arthropitys (parts a and b). 132 Annals of the Missouri Botanical Garden metabolism is widely associated with primary vascular tissue production (Leyser & Berleth, 1999); it also may be associated with development of conducting tissues in bryophytes (Cooke et al., 2002). The basal/apical ratio of auxin flow is one to two orders of magnitude lower in the only lycophyte studied, Selaginella, than it is in the euphyllophytes 2002: table 4). | is characteristic of lycophytes as a whole, it would studied (Cooke et al.. If that difference account for the much less obvious anatomica — = ingerprint for the role of axial polar auxin flow in the regulation of secondary xylem formation above the branches of Paralycopodites than in similar positions of the fossil euphyllophytes studied (i.e.. Archaeopteris and Arthropitys). This ratio has yet to be determined for Equisetum. However, if the distinctive lingerprint Arthropitys (Fig. 6B) flow, we n` present in the wood of correlated with levels of polar auxin hypothesize that future studies will document polar auxin metabolism in equisetophytes that is compara- ble to other euphyllophytes. CONCLUSIONS Until the recent recognition of structural finger- I prints for axial polar auxin flow in fossil plants (Rothwell & 2005). possible to employ data from the paleontological t has not been Lev-Yadun. record to infer plant evolution from changes in developmental regulation. However, the discovery of developmentally diagnostic circular patterns in fossil secondary xylem reveals that secondary vascular tissue development was probably regulated by the axial polar auxin flow in all of the major clades of vascular plants that have produced woody trees through time. Because such patterns result from obstructions of the axial polar auxin flow in the vascular cambium, both the positions of the circular patterns and the extent of their development provide meaningful information about specifies of the auxin metabolism, the growth and development that has resulted from regulation by auxin, and the organog- raphy of the extinct plants. These data imply that the much lower basal/apical [AA transport ratios found in living Selaginella also characterized Paleozoic fossil lepidodendraleans and that the patterns preserved above branch junctions of Paleozoic fossil Calamitales are predictors of the auxin flow in living Equisetum that is similar to other modern euphyllophytes. Likewise, the extensive development of circular patterns at the base of Chaloneria and similar Paleozoic lycophvtes with a plant base similar to living /soetes strongly supports that highly the hypothesis rhizomorphic lycophytes are rooted by a modified shoot system and that regulation by the axial polar they lack true roots that are homologous to those of other vascular plants. Although the fossil record now provides strong anatomical evidence that a common mode of devel- opmental regulation underlies secondary vascular tissue production in all of the major clades of vascular plants that have produced extensive wood through time, the currently recognized pattern of phylogeny for tracheophytes reveals that secondary vascular tissues probably evolved independently in lycophytes. equi- setophytes, and lignophytes. The occurrence o regulation by polar auxin flow of primary growth across the entire spectrum of vascular plants (Sachs, 1991; Scarpella et al., this developmental regulatory mechanism were pre- 2006) indicates that genes for sent in the herbaceous ancestors of all woody plants, although they were not expressed in circular patterns before the emergence of secondary. vascular tissues (Rothwell € Lev-Yadun, 2005) and were available for exaptation to the new role of secondary vascular tissue development. Therefore, while the secondary xylem of lignophytes, equisetophytes, and lycophytes is not inherited. from a common ancestor that homologous (1.e.. that produced secondary xylem), the genes —— underlie the regulation of wood production by the axial polar auxin flow in the three clades probably are homologous (i.e., they probably are paralogues). As a occurred within all ancient result of the extensive extinction that has clades of vascular plants (Mishler, 2000), including the large woody trees of —— lycophytes and equisetophytes investigated in this study, there is a rapidly growing recognition that the in- depth understanding of evolution at both the + / pdt TITLIVCALE ilatory and organismal levels relies heavily on paleontological data. Information presented in this study reveals that such data are, indeed, available from fossil plants. Although the work presented here is restricted to increasing our understanding of growth of secondary vascular tissues in terms of regulatory mechanism evolution, the rapidly growing number of newly developed techniques for recognizing develop- mental regulatory mechanisms from fossils (Boyce & Knoll, 2002: Rothwell et al., 2006: 2007) promises to increase dramatically the breadth of Sanders et al., y such studies. In so doing, paleontology is becoming more intimately incorporated into the mainstream of plant evolutionary studies. Literature Cited M. € W. A. DiMichele. 2002. Generating and iltering major phenotypic novelties: NeoGoldschmidtian I 9-159 in Q. B. C. Bateman, R. — = saltation revisited. Pp. Cronk, J. Bateman & A. Hawkins (editors), Dev x ntal Geneties and Plant Evolution. Taylor & Francis, New York. Volume 95, Number 1 2008 Rothwell et al. Fossil Record for Growth Regulation 133 eck, C. B. 1960. The ipe of Archaeopteris and Callixylon. uuo 12: K. & A. H. Knoll. 2002. lila of developmen- tal potential E E^ multiple independent origins of leaves in Paleozoic vascular plants. Paleobiology 28: Chen, J. J., B. gene fusi J. Janssen, A. Williams & N. Sinha. 1997. A ion at a homeobox locus: Alterations in leaf a and implications for a olee evolution. PI. Cell 9: 289-13 e Chuck, G., C. Lincoln & S. Hake. lobed leaves 1996. KNATI with ec Des meristems when overexpressed ell 8: 1277-1289, induces in M M Pl. Cichan, A. 19 ment in n. Een rous plants. I. Lepidodendrales. Amer. Bot. 72: 1 ES P 17 Cooke, T. J. E A. E. Sztein & J. D. nue n. 2002. Polo ds in auxin action. Pl. Molec. Biol. 49: 319- 85. oe ular cambium and wood Mn Crane, R. 1985. Phylogenetic analysis of seed P PA the origin of angiosperms. Ann. Missouri Bot. Gard. 716 Dele F E 1964. Pitomorpholoss : 4: 14 DiMichele, W. 980. Paradies Morey & Morey, from the C saree of Euramerica— pue ne studies of fossil plants. -3 reassessment of E i of Bot. 67: nerie affinities and A de tfolium Williamson. Amer. J. 1466— Eggert, D. A. 1961. 6 itogeny of the arborescent Lyc eposica. Palacontgraphic a 108 M. & D. 69. E Carl mn rous -92. Frankenberg, J. ¿ge E rom North Ame Stigmaria ficoides, the underground portions of Le a eae. Palaeontogra- phica far 1-47. Gleissberg, S., E. P. Groot, Kólsch & m Hutter. 2005. to peltate leaf structure in Tropaeolum majus (Tropaeola- ?etrified Sigmar erica: Part Schmalz, M. Eichert, A. Se ie ntal events leadin ceae) are associated with expression domain changes of a dud gene. Dev. Genes Evol. 215: 313—319 Good, C. W. 1975. elate i Palacontographic a l5 v Pennsylvanian-age calamitalean cones, aring spores, n associated vegetative organs. 2-99, KP . Taylor. 1972. The ontogeny of mee articulates: The apex o Amer. J. Bot. 59: 917-6: Gould, S. I 1977. Ontogeny and phylogeny. Harvard Univ. se Press, C le Massachu 1994. Tempo and mode in the macroevolutionary rec gy tion of Darwinism. Proc. Natl. Acad. Sci. U.S.A. Til. Hare 5 T. Gutfinger, A. Parnis, Y. Eshed & E. I ise s 1996. The making of a compound leaf: Ge E M of leaf architecture in tomato. Cell 8 35-7 Harrison, C. J.. M. Moller, J. Langdale, Q. B. C. Cronk & A. Hudson. e: um role of morphological novelty in Streptocarpus. Pl. 430—443. Hejnowicz, Z. & E. KNOX genes in the evolution of cel . Kurezytiska. 1987. Occurrence of circular vessels sho ve axillary buds in stems of woody plants. ae Soc. Bot. 415-419. Howell, S. 1998. C x : niv. Press, i^ K. W., J. Willis & W. S. :ellulose t lec Ds in CE (Oxford) 20: 635-6: Poloniae 56: Tod cular Gene tics of Plant Development. . À rapid E Bot C mE Mei Karrfalt, E. E. 1980. Further Wu O of Isoetes roots and stigmarian appendages. Canad. J. Bot. 8-2322. D. A. Eggert. 1984. The development of the ma origin and early root-producing meristem of /soetes E omez. Bot. Gaz. 145: 372-377. ick, P. . Telome theory. Pp. 365-387 in Q Cronk, E. A Bateman vkins (editar). Developmental. Genetics and Plant Evolution. Taylor € Francis, New York. & P. R. Crane. 1997 Plants. The Origin and Early Diversification of I Land Smithsonian Institution Press, i Washington, D.C Kim, M., Ee S. McCormick, R. K. Kuzoff € N. M eee Reduced leaf complexity in tomato wiry "HAN and KNOX generating nm leaves. Development 130: 44.15. Lev-Yadun, S. € R. Aloni. branch june tions of trees: Circular patterns and functional 9-54. " Pham, A mutants sugges role for H genes in 405- 1990. Vascular differentiation in significance. Trees 1991. Natural induced dispersion of aggregate and Caras f Quercus ob. Ann. Bot. s in shoots o — ays and Q. loro: We ery Decne. 68 Leyser, 2n &' . Berleth. ction. Seminars Cell Developmental Biol. 10: 131-137 1999, A molecular D for auxin Me yer-Berthaud, B. 1984. Les axes de lye ud a structure servée du Carbonifere Basal (T Noire: eyrodendron gen. nov. Palaeontographica 190B: 1-36. Millay, M. A. & D. A. 974 dn: nt in the Paleozuit seed fÍ Amer. J. Bot. 61: anatomique cons de la Montagne Trabicaulis gen. nov. B: “ggert. 1974. Microgametophyte fern family Calli 1067-1075. E islo- phytaceae. Mishler, B. D. 2000. Deep phylogenetic relationships among “plants” E their implications for classification. Taxon 49: 66 m Phillips, 1 197 9. Reproduction of heterosporous arbores- lississipplan- Por of Palynol. 27: ). Stratigraphic and a occurrences of cent a in the ume à. Rev. Palaeobot. permineralized coal-swamp plants— of North America and Europe. Pp. 93-118 in D. L. Dilcher I Tavlor (editors), Biostratigraphy of Fossil Plants. Dowden, Hutchinson & Ross, Ine., Stroudsburg, Pennsyl- Pigs, K. B. & G. W. Rothwell. 1979. an Upper Pennsylvanian woody lycopsid. Amer. J. Bot. 66: 914—924 Stem-root transition of & —— —. 198 rous lycophytes from n Pennsylvanian of North America. Bot. Gaz. 144: 132-147. . Chaloneria gen. nov.; heterospo- I. Schneider, A. R. Smith, R. Cranfill, P. G. Wolf, J. S. Hunt & S. D. A 2001. Horsetails and ferns are a S lie group and the closest living relatives to > ints. Nature 409: 619-622 Bernat D., E. Pesce, P. Baltensperger, M. . J. Friml 2003. Regulation k oa by polar auxin al ort. Dow 426: 255-200. Rothwell, G. W. 1971 ui Rn T. Mandel K. Bennett, i Kuhlemeier. . Ontogeny of the fau Amer. J. Bot. 706 e role of deve lopme nt in a phy logen "ny: li ovule: 1]: 1987. Th 7 paleobotanie 'al perspective. Re —114, 1999. Fossils and ferns in the resolution of land '. Palaeobot. Palvnol. € 86-2 plant lora ny. Bot. Rev. 65: 134 Annals of th Missouri Botanical Garden . 2002. Coal Balls-Bemerkenswerte Zeugnisse di a- p her Pflanzen ae et anzengese mé hafen. Pp. 3 . Dernbach & W. D. Tidwell (editors), Ge bu nuc Pflanzen. Doe Verlag, Hergestellt; Ger- many. & D. M. Erwin. 1985. The rhizomorph apex of Paurodendron; implications. for homologie: s among the rooting organs of Lycopsida. Amer. J. Bot & S. Lev-Yadun. 2005. Evidence of a auxin o o; Amer. J. zm =i in ee 5 LEE -year-old fossil wood. 903 — 2006. How does the inclusion of fossil data s ange our conclusions about the history Ü DE Int. J. Pl. Sei. 167: —— & erbe 904. Lignophyte . Nixon. > 37-749. ER and the MS P. E & a numerical cladistic anal- 'S 9: 443—482. 1 . H. Sanders, S. Wyatt € S. Lev-Yadun. 2006. Paleontology provides a context for the a nlal mechanisms of evolution. Abstracts of Botany 2006, the Botanical Society of America 100th uda meeting. : he 0 State mu sity, Chico, California, Abstract No. 39, p. 21 Sachs, 1991. Pattern Fonna an sd Plant Tissues. Cambri e Cole n. 1982. Circular vessels and the control o of ves differentiation in plants. Differentiation 21: 22- e H. . Rothwell & 5. Wyatt. 2007. tological context for the deve eae ntal mechanisms of evolution. Int. J. Pl. Ser. 168: 119-128. A paleon- Scarpella, E., D. Marcos, J. Friml & T. Berleth. 2006. Control of leaf vascular patterning by polar auxin transport. Genes Developm. 20: 1015-1027. Scheckler, S. E. 1978. Ontogeny of Progymnosperms Il. d = a Devonian Archaeopteridales. Canad. J. Sol 6—3170 N pae Smoot, T. N. Taylor € B. S. Serin. 1982. Arabasacatomllos from the U n oss of Arkan- sas. Rev. Palaeobot. Palynol. —334. > /. E. 1993. Modeling ^ "lon of stelar archi- 4: 229-26: tecture in vase [E oe Int. J. PL . 15: )—263. Stewart, W. N. & C. Rothwell. 1993. p POM: and the Evolution of ae Be ed. Cambridge Univ. Press Cambridge. Taylor, T. N. & G. W. Rothwell. 1982. Studies of seed fern pollen: Development of the exine in Monoletes (Medullo- sales). Ame — —— & E. L. Taylor. 1993. The Biher and Evolution of Fossil Plants. Pr entice Hall, Engle pes Cliffs, New Jer Tomescu, A. M. F., G. W. Rothwell € Trivett. 2000. Kaplanopte Mo ien fam. nov., od dive ud in the in D radiation of filicalean ferns. Int. J. Pl. 167: 6154 Trivett, M. L. 1993. An architectural analysis of Archae- opteris, a fossil tree with pse yd qr and opportu- u ol. J. nistic adventitious growth. n. Soc. 11l: 301-3209. Yi, S.-Y. & M. Kato. 2001. Basal meristem and root itu eut nt in eg asiatica and Isoetes japonica. Int. J-P Sei- 225-1235. 62: PHYLOGEOGRAPHY, FOSSILS, AND NORTHERN HEMISPHERE BIOGEOGRAPHY: THE ROLE OF PHYSIOLOGICAL UNIFORMITARIANISM! Bruce H. Tiffney* ABSTRACT Biogeography is one of the most synthetic of biological undertakings; it requ in a geological, climatological, and ecological context, all of which shift through time. of cladistic and molecular techniques has diversified our grasp on phylogeny of past distribution patterns based on samples from available contribute to these cladistic approaches by providing a basis for estimates of rates of dive rgence. Howe a fossil of the taxon occur whe re predicte a) and by ec predicte d. environment?). The first te and time. sl ls S cond is based on the peii pi common das possess similar a iolog If c f the t hypotheses, together with new pale llene cal dat (b) the need to reevaluate others; and (c) t the Old and New Worlds into the Later Tertiary, contrary Key words: predicted physiological tolerances hytogeography. Te ran , uniformitarianism. D D y adding ANAL al ehec o ts to character Bering land bridge, boreal, fossil E North Atlantic 1 ind bridge, uires placing a substantiated Rh die model 5, the applicati . This has allowed the onn of dete eses living mate " and algorithms for their interpretation. Fossils In the past two deca socialions in time and by er, fossils also check these hy hee: by direct occurrence (does ological suitabilit y (is it reasonable that the taxon could occur in the Ma aii wits if difficult, re qug : the finding of a specific fossil in a specific place [ physiological unil corre ormitarianism—that fossil and modern taxa united by a then hypotheses of past distribution must accord with the xon in question. Ap ol ‘ation of physiological uniformitarianism to phylogeographic ] PI l ) geogra] the validity of many established phylogeographic hypotheses; > recognition ue a North Atlantic land bridge likely functioned as a link between is author's earlier pay Norihe rn Hemisphere, phylogeography, The establishing historical biogeographic hypotheses was classic approach to the use of fossils in essentially scientific but not especially satisfactory. One would survey the known fossil record, assemble the reputable records of the taxon in question, and then suggest a pattern of migration consistent. with The hypothesis was constrained by known deposits and their distribution in space and time. resulting existing identifications. This hypothesis could be predictive (it might suggest where new specimens — could be expected to be found in time and space), but the prediction were lacking. often necessary fossil deposits to test the Furthermore, the discovery of a new representative of a given clade in an unpredicted time or place often required the hypoth- esis to be completely reconceptualized. Thus, advane- es took new data into account, but not infrequently left a historical trail of contradictory hypotheses. It would lua] be far more satisfactory 1 | narrow the possibilities through iterative testing and the refutation of a priori predictions (cf. Platt, 1964). Hecent advances in molecular techniques and data analysis provide a complementary approach to pale- ontology that helps to constrain phytogeographic hypotheses to a greater molecular J D degree. Using analyses of genetic material from living plants, coupled with analysis of morphological characters, practitioners. invented the field of phylogeography (coined by Avise et al., 1987), overlaying phylogenetic interpretations on fossil and modern biogeographic patterns (e.g., Wen, 1999, 2001; Xiang et al., 2000; Donoghue et al., 2001; etc.). Furthermore, based on the gross assumptions of the molecular clock, phylogeographic hypotheses can also predict the time of phylogenetic splitting, placing a particular clade in a particular geographic locale at a particular geologic time, which creates a testable hypotheses based largely on non-fossil data. This is an important advance, as Donoghue et al. (2001: S41) note: ‘Geological and climatological processes that have impacted the biota of the Northern Hemisphere during the Tertiary are expected to yield little resolution when area cladograms are compared without taking the »hvlogeosraphic phytogeogra] timing of diversification into account.” Thus, approaches have the potential to offer alternative hypotheses to those generated. from I thank William L. ` and Colleg U 8. ne doi: 10. 3417/20061 90 ANN. Missouri Bor. ` of Creative Studies Tiffney@ccs.ucsb.edu or Tiffney ge ol.ucsb.edu. (GARD. 95: 135-143. Cre for his invitation to participate in this symposium. . University of California-Santa Barbara. Santa Barbara, PUBLISHED ON 11 APRIL 2008. 136 Annals of the Missouri Botanical Garden paleontological data alone. However, fossils retain an important role in this brave new world. First, the calibration of the clock for a particular clade is dependent on its antiquity as defined by the fossil that assumptions and is fraught with complications thal record. This is a process involves many have been the focus of several presentations in this issue (Burnham, 2008; Gandolfo et al., 2008; Nixon. 2008). Additionally, Shaw and Small’s (2005) caution on using molecular data from a single specimen, rather than exploring the potential for molecular variation within a species through multiple samples. deserves wider recognition. Second, fossils offer both a direct and an indirect method for testing phylogeo- graphic hypotheses. Considerations of these tests and, particularly, the ecological fit of past distributions into phylogeographic models form the core of this paper. MoDEs oF TESTING PHYLOGEOGRAPHIC HYPOTHESES WITH FOSSILS DIRECT TESTING—THE RIGHT FOSSIL IN THE RIGHT PLACE A phylogeographic hypothesis may suggest the presence of a particular clade at a particular geologic time in a particular geographic place. In general, the most important of these predictions. will involve migrational bottlenecks. Thus, in the present case, the focus is on intercontinental disjunctions of taxa between the New and Old World in the Northern Hemisphere, in which the two greatest barriers are the Pacific and Atlantic oceans. These, in turn, create two bottlenecks over which terrestrial organisms need to pass—the Bering land bridge (BLB) and the North Atlantic land bridge (NALB). existence, although limited by seasonal climate, while The former is still in the latter has been affected both by oceanic widening (plate tectonics) and global cooling (Tiffney « Manchester, : The divergence logically should correspond to the origin estimated time of the onset of genetic of the interruption of gene flow. This interruption could result either from a migration event across a barrier causing the establishment of a new, isolated population or from the vicariance of a pre-existing and perhaps long-established distribution. These alterna- tives cannot be distinguished from molecular data, although a pre-existing fossil record can support the atter. However, both hypotheses imply the presence of representatives of one or both clades involved in the ^ (om isjunction at the geographic locality that is inferred — to form the barrier-to-gene flow at the time of its interruption. If a phylogeographic hypothesis thus predicts the taxon in a particular place at a presence of a particular time, then paleontological data may be sought in the appropriately located and aged rock. For Donoghue et al. (2001) predicted that Thunb. and Diervilla Mill. (Caprifoliaceae) split ca. 5.2 + 0.5 million years ago (Ma) to estab- example, Weigela lish a present-day eastern Asian (Weigela)-eastern North America (Diervilla) distribution. The clade was — | wpothesized to have arisen in China or Japan and to have migrated over the BLB to North America. (2001) wrote, they were not aware of any fossils of Weigela or Diervilla in the White et al. (1999), in a fairly obscure publication, place pollen similar to that At the time Donoghue et al. However. Bering region. of Weigela and Diervilla in northwestern Canada and Alaska, appearing 15 Ma and disappearing ca. 6.5 Ma. While the fossils suggest an older first appearance in the area and a slightly older disap- a pearance than that suggested by phylogeography, the agreement is quite acceptable and the hypothesis is not refuted. However, it is to be noted that Weigela also occurs in the European Miocene and Pliocene (Mai € Walther, 1988: Mai, 2001) that the hypothesis of a North American-European link must and still be at least acknowledged. INDIRECT TESTING—PHYLOGENETIC UNIFORMITARIANISM Phylogeographic hypotheses place specific taxa in ylogeogra] yl | } specific places at specific times. When they do so, the the surrounding vegeta- general ecological context of tion should be in accord with the physiological tolerances of the living members of the clade in question. To use a trivial example, the Palmae are classically subtropical and tropical in distribution: a phylogeographic hypothesis that placed a palm in a temperate or boreal context would violate our understanding of the physiologic tolerances from modern examples. This interpretation is based on the assumption of physiological uniformitarianism. Succinctly, we identify taxa in the fossil record by their morphol- ogy. assigning them to living families. genera, o oecasionally species in the assumption that similar morphologies reflect. similar underlying genotypes. However, the taxa in question also possess unique that are again genetically shysiological characters pm 8 controlled. If we make the identification of Vitis L. in the fossil record, we assign it to the genus on the basis of morphology and assume that it represents a liane, as Vitis is overwhelmingly dominated by vines Mabberley, 1997). that it brings with it a set of physiological tolerances that g pm B Similarly, we assume allows it to grow in specific thermal and moisture regimes that fall within the tolerances of modern species within the genus. In the case of Vitis, this Volume 95, Number 1 2008 Tiffney 137 Phylogeography, Fossils, and Northern Hemisphere Biogeography results in a very broad range of ecological tolerance, as the genus extends from temperate to subtrop- ical areas in the Northern o (Tiffney & Barghoorn, 1976; Mabberley, 1997). An alternative possibility must be considered: that morphology has remained sufficiently stable over time to allow placement of a fossil in a modern taxon while physiology has evolved independently, dissociating ecological tolerances from morphological identifica- tions. If this were the case, flora and vegetation would — exhibit a similar dissociation through time, and no clear relationship would be observed between cohorts of taxa and specific climates. While the nature of biological systems is too variable to state that this has never happened (e.g., that a genus of subtropical angiosperms evolved a temperate species to cross through a cooler climate, which, in turn, gave rise to a subtropical species upon re-encountering warm cli- mates), the coherence of the history of vegetational and floral change through the Tertiary (e.g., Tiffney, T Mai, 1995; Graham, 1999: Tiffney & Manches- r, 2001), in concert with climatic change adduced s other sources (e.g., isotopic sources, Zachos et al.. 2001), suggests that physiology is generally stable and tracks morphology. To date, my discussion has implied that the phys- iological tolerances are those involving moisture and seasonal warmth. However, there are other aspects of physiology that potentially play an important role in determining clade survival in a given environ- ment. Chief among these is the annual distribution of light at the poles. In previous papers (Tiffney, 1985, 1994, 2000), I have queried whether extended periods of polar winter darkness might prove to be a barrier in the Northern Hemisphere to the migra- tion of evergreen taxa in periods of global warmth, as respiration. would carry on in the absence of photosynthesis. If so, evergreen lineages might be restricted to more southerly migration routes in order to balance respiration with a minimum level of photosynthesis. This assumption appeared reason- able, given (a) the frequency of deciduous taxa in the fossil record of the North Pole (Falcon-Lang 2004; Brentnall et al., 2005), (b) earlier experiments (Read & Francis, 1992), and (c) the importance of some deciduous taxa (e.g., Larix Mill.) — et al., in boreal realms in the present day. While this is a rational conclusion, in fact the current global climate is sufficiently cold that we do not have the opportunity to observe the response of thermophilic evergreen taxa to extended periods of darkness, as could have potentially occurred at times in the Paleogene of the Northern Hemisphere. Recent experiments suggest that the carbon loss involved in seasonal leaf shedding is far greater than the loss involved in dark-season respiration (Royer et al., 2003; Osborne et al., 2004) and that some feature other than carbon balance must drive the prevalence of deciduous taxa in the boreal realm. Brentnall et al. (2005) suggested that this feature reflected the frequency of fire-based disturbance. Such experi- ments are conducted over brief time spans; in order to survive and migrate, evergreen plants would have — o succeed over a multigenerational time frame, allowing reproduction, dispersal, seedling germina- and establishment through repeated cycles. Further both balance and/or fire seems in order, especially since tion, examination of the role of carbon another pattern of note is that many extant high- latitude austral taxa are evergreen. Other features may be involved in explaining the apparent barrier to evergreen taxa; these might be ecological (e.g.. Givnish, 2002) or involve phylogenetic constraints inherent in the boreal taxa involved. If we accept the reality of physiological uniformi- larianism, then its application to testing phylogeo- graphic hypotheses has one very useful aspect; it does not require the occurrence of the taxon under investigation in a particular place in the fossil record. s long as there is a fossil flora or fauna with sufficient detail to allow estimation of the paleoenvir- onment, or even a good geochemical estimate of paleotemperature, one can infer whether the particu- lar taxon in question would logically have survived in the predicted environment. However, this advantage is balanced by disadvan- tages. Foremost, it can only be applied to living taxa wherein the physiological tolerances of light, mois- ture, temperature, etc., are fully known. Furthermore, it cannot be effectively used on taxa with very broad tolerances. It would be difficult to support or falsify a historical hypothesis about a subtropical taxon if it were predicted that its occurrence fell within the scope of a warm-temperate vegetation. This aspect becomes more problematie in the earlier Tertiary where reduced climatic gradients allow the wide- spread occurrence of equable climes with interdigi- lated temperate and subtropical taxa. Nevertheless, the method still has power to refute hypotheses where the environmental tolerances of the taxon in question are substantially different from those of the predicted environment of occurrence. SoME Tests EmPLOYING PHYSIOLOGICAL UNIFORMITARIANISM From the foregoing, we can proceed to look at a number of taxa for which phylogeographic hypotheses have been made, placing them in particular geo- graphic situations and climates. 138 Annals of the Missouri Botanical Garden THE BERING BRIDGE The BLB is a frequently invoked route of exchange between the Old and New Worlds in the Tertiary (Tiffney, 1985; Tiffney & Manchester, 2001). Several southern Alaskan Eocene floras hosted evergreen thermophilic lineages (Wolfe, 1972, 1977), which may indicate that the bridge was open to evergreen lineages at this time (see discussion Tiffney «€ anchester, 2001: S7). Later Eocene through Oligo- = cene floras displayed an increasing dominance of warm- s to temperate dec e taxa (Wolfe, 1972, 1992). By the Early to dle Miocene, deb floras hosted moderate- to pide Msi e a taxa (e.g., Juglandaceae, Fagaceae, Liquidambar L., Nyssa L., Tilia L.; Leopold & Liu, 1994; White et al., 1997, 1999) Many of the older crossings hypothesized from molecular evidence fit well within the physiological tolerances of the taxa involved. Rhus L. was suggested to be present on the BLB at ca. 34 Ma (Yi et al., 2004), Menispermum L. at ca. 29 Ma (Lee et al., 1996), Campsis Lour. at ca. 24.5 Ma (Wen & Jansen, 1995), Gleditsia L. and Gymnocladus Lam. at ca. 20 Ma (Schnabel et al., 2003), and Liquidambar L. a (Ickert-Bond & Wen, 2006). s are all M EIS mesic taxa, and all but Rhus are 15- or before 15.6 M deciduous: the latter includes deciduous and ever- ereen taxa. (Late hypothesized crossings also fit within the predictions of physiological uniformitarianism. Nie et al. (2005) that Benth. (Rubiaceae) crossed the Bering Bridge ca. 5.5 Ma. Several more recent Miocene-Pliocene) Thus, predict Kelloggia Torr. ex lts distribution in montane western North America (viz. Burke Herbarium, 2006) and cool Asia fits this hypothesis. Nie et al. (2006a) predict that Symplo- carpus Salisb. ex Nutt. and Lysichiton Se ‘hott (Araceae) crossed the Bering Bridge ca. 4-7 Ma. This, again, appears consonant with the Late Miocene—Pliocene Beringian vegetation and ecology of the extant taxa. However, other hypotheses appear to pose problems for physiological uniformitarianism. Some are mar- Nie et al. (2006b: 1349) suggest attained its einally problematic. that Phryma L. Asia—eastern North America distribution "at least 3.68 + 2.25 to 5.23 35 through the Pleistocene. Phryma has a wide geo- ranging growing in rich and often moist eastern 1.37 mya," or in the very latest. Miocene graphic range in North America, from Manitoba to Florida, Fernald, 1970). Its geographic distribu- 2006b: fig. an equally temperate environment in southeastern woods (e.g., tion in eastern Asia (Nie et al., 4) suggests Asia north through Japan and Korea. None suggest an affinity for boreal forest conditions, which would have been present on the BLB and environs during the 1994; Phryma’s wide predicted time of crossing (Leopold & Liu, White et al., 1999). modern distribution, especially in North America, Nevertheless, offers the argument that it might be able slip through the BLB during a brief warm spell within this time frame. Other nature and are harder to imagine making the crossing taxa are more clearly warm temperate in in latest Miocene and Pliocene. For example, Xiang et al. (1998) proposed that Calycanthus L. (now occurring in China, western North America, and southeastern North America—a classic Tertiary o distribution, e.g., Li, 1952) crossed Beringia al | Ma. Wen et al. (1996) and Zhou et al. (2006) RR ds crossing a little earlier at ca. 6 Ma. Wen (1999, l., 1995) suggested that Magnolia tripetala (L.) L. split from M. officinalis var. biloba (Rehder & E . H. Wilson) Y. W. Law of China ca. 2-5.5 Ma. By 6 Ma, marginal exolics that had been present even two million years earlier (Reinink-Smith € Leopold, 2003, 2005) were largely absent from the coastal coniferous forests of Beringia (Wolfe, 1994). from central Alaska suggest a boreal forest (Leopold € Liu, 1994; White et al., 1999) with a few temperate taxa (e.g., Tilia, Corylus L., Juglans L.). In Similarly, after Qiu et a Post-6 Ma pollen records cool- light of its current locales of growth, Calycanthus would not be ecologically suited to these environ- ments. While M. tripetala does co-occur with Tilia, Corylus, and Juglans in its present eastern. North American distribution (Little, 2006), it has a more Addition- ally, the proposed crossing for Magnolia L. overlaps southern distribution than these other taxa. vith the opening of the Bering Strait (Gladenkov et al., 2002). The hat these phylogeographic hypotheses require re- E. ack of physiological consonance suggests — examination. For example, while Schnabel and Wendel (1998) placed Gleditsia in Beringia at 3.9— 7 Ma, again associated with an implausibly cool (2003) ‘onclusion and placed this crossing in a much more — flora, Schnabel et al. re-evaluated this 5 appropriate climatic setting at 15-20 Ma. While I focus on the bottleneck created by the BLB, in fact, the hypothesis of such a crossing involves more than just the transition across the BLB—it also invokes migration to the bridge and from it to the of distribution. Thus, \ the BLB by Magnolia Calycanthus also must account for the modern areas involving the crossing of tripetala or movement from the point of crossing to their present if the postulated time of molecular divergence measures a areas of survival in eastern North America, migrational event. To make this transition, these taxa need not only deal with the ecological setting of the Volume 95, Number 1 2008 Tiffney Phylogeography, Fossils, and Northern Hemisphere Biogeography vegetation of Beringia at circa 6 Ma, they must also cross the Rocky Mountains and the interior grasslands before becoming established in the mixed deciduous forests of eastern North America. Given the changing nature of North American vegetation through the Tertiary (Graham, 1999) and the ecological affinities of Magnolia and Calycanthus, this would be more easily accomplished in the Eocene through Middle Miocene rather than later in the Tertiary. that achieved their distribution at an earlier time, but that > 1 alternative hypothesis is these taxa the molecular divergence measures the time of their vicariance. This eliminates the objection of the need to migrate through other potentially inhospitable the Magnolia or Calycanthus co-associated with boreal =- environments, but still leaves incongruity o taxa in Beringia at the proposed time of interruption of gene flow. This leads me to wonder if, in cases of vicariance or migration, the actual molecular diver- gence lags behind the time of cessation of gene flow. THE NORTH ATLANTIC LAND BRIDGE In Alaska and northeastern Siberia, there is a relatively good fossil record through the Tertiary 1989; Wolfe, 1994; 2005). Therefore, it is possible to both track ongoing; Baranova et al., Fradkina et al., exchange throughout the Tertiary and test phylogeo- graphic hypotheses in this area against fairly numerous taxonomic and ecological data. This has allowed me to be rather black and white in ap- proaching the role of fossils in evaluating molecularly derived phylogeographic hypotheses and focus on support or refutation. Hypotheses and tests lie at the core of science at its best, but often the scientific process is far more labile, resulting not in tests and refutations, but in an awkward upward spiral of understanding as new data modify our interpretation of preceding data. | turn now to just such a case involving the NALB, where I believe that molecular phylogeographic interpretations coupled with new paleobotanical data suggest a changing interpretation of the history of this important exchange corridor. 1985, 1994, 2000), I interpreted that the NALB was an important route for the based on a range of fossil evidence. Its importance In earlier papers (Tiffney, biogeographic exchange in earlier Paleogene in the late Paleogene into the Miocene was difficult to assess by direct evidence, but plausible arguments for the migration of taxa have been made in the Oligo- cene (Hably et al., 2000) and Miocene (Liquidambar, Ickert-Bond & Wen, 2006). Miocene, | was dubious of its significance in light of missing biological (fossil record) and geological Our However, post Early data, both to the east and west of Greenland. paleontological and geological knowledge of the possible later Cenozoic links between western Green- land and eastern North America remains quite limited but is illuminated by more recent findings. The fascinating Pliocene flora from Kap København in northernmost Greenland (Funder et al. 1985: Bennike, 1990; Bennike & Bócher, 1990) suggests a tundra vegetation bordering on taiga. To the south and Pliocene floras (Vincent, 1990; Matthews & Ovenden, 1990) and insect faunas (Elias et al., 2006) on Prince Patrick, Meighan, Ellesmere, and Banks west, islands of Arctic Canada suggest a mixed hardwood/ conifer forest in the late Tertiary giving way to larch- dominated taiga and tundra by 2 Ma (Vincent, 1990) in some places, and pine-dominated taiga in others (Matthews & Ovenden, 1990), including Sciadopitys Siebold & Zucc. on Meighan Island. Precise interpre- the complicated by a stratigraphy that is still being resolved (Fyles, 1990). Climatic interpretations of (Elias & Matthews, 2002) suggest that the latest Miocene through Pliocene had a lower latitudinal temperature gradient than at present, allowing for the wider spread of the taiga in this period (Matthews et al., 2003). It would be valuable to trace these transitions back tation of timing of vegetational change is entomological fossils to the Miocene, but only the rich flora from the Middle Miocene Mary Sachs Gravel on Banks Island exists to provide a norm for mid-Tertiary floras of Arctic North Interestingly, this includes Glyptostrobus Endl., Metasequoia Hu & W. C. Cheng, Juglans, Liriodendron L., Phyllanthus L., and Actinidia Lindl., among other taxa (Matthews & 1990). No presently reported between the high Arctic and 35°— 40° eastern America. flora Ovenden, urther Miocene floras are atitude in eastern North America. The best continuous evidence for sub-boreal vegetational and climatic change throughout this time frame in the North Atlantic region comes from ca. 50°N latitude in Germany (Utescher et al., 2000). Turning to geography, we presently lack evidence establishing or refuting possible physical links between western Greenland and eastern Canada south of the northern tip of Greenland in the mid-Tertiary. A second line of evidence for estimating the timing of the opening of the Banks Strait might involve tracking changing biogeographic affinities of marine inverte- brates in these waters in response to the opening of the Bering Strait (e.g., Marincovich, 2000), although 1 am unaware that such data currently exist. Since publication of my previous papers, several phylogeographic appeared that support later Tertiary exchange via the NALB. For example, Whitcher and Wen (2001: 296) suggested Corylus sect. Corylus migrated from Europe to North hypotheses have 140 Annals of the Missouri Botanical Garden America “in the late Pliocene or Pleistocene.” Realizing both the purported physical barrier and real climatic barrier (taiga and tundra), they invoked long-distance dispersal, perhaps via a bird. This 8 | } | hypothesis requires a lot of the bird, given the distance, the prevailing westerlies, and the size of a r Milne (2004), using molecular an Corylus fruit. Similarly data, suggested that Rhododendron L. subgen. Hyme- (Blume) K. Pontica achieved a distribution between Europe and North nanthes Koch subsect. Tagg America ca. 6 Ma or less. Given existing geological and paleontological data, I was predisposed to assume that both timings were anomalously young. Quite recently, Denk et al. and Denk (2007) provided an updated evaluation of (2005) and Grimsson the Miocene floras of western Iceland, together with a discussion of their geologic setting. These authors placed Corylus on proto-Iceland between 12 and 8 Ma the —— in an ecologically suitable temperate flora (vs. suggested 3-2 Ma via molecular hypothesis) and the same paper placed Rhododendron aff. ponticum L. in a 12-10 Ma Icelandic flora (vs. 6 Ma as suggested by Milne [2004]. Their geological analysis focused on the links between Greenland, Iceland, and Europe. and made clear that terrestrial bridges or island stepping-stone options were present through. the period. diminishing toward 6 Ma. While their work the Greenland—North clearly indicates the feasibility of a does not address America barrier, it late Tertiary Europe—Greenland link. In light of these data, the history of both Corylus and Rhododendron aff. ponticum would be consonant a European—North Amer- Middle to Late followed by a vicariance event. But, as with the Bering with their having achieved ican distribution in the Miocene, example, the molecular data predict that the vicari- ance event occurred substantially later. Is there any evidence suggesting the possibility of a suitable environmental corridor in the Late Miocene through Plioc Glaciers capable of generating extensive ice-rafted ene? dropstones in the Norwegian-Greenland Sea existed in the Late Eocene (Eldrett et al.. 2007), indicating an early development of a cooling climate on Greenland. Ice-rafted debris became more frequent in the mid- the but there is evidence Miocene and gathered in strength through Pliocene into the Pleistocene, to suggest that there was a significant variation in the 1998). Ma taiga-tundra vegetation at the norl strength of this process (Thiede et al., Thus, the Ca. 2 of Greenland (Bennike, 1990) suggests the possibility — ern point of a deciduous forest over more southern portions of M Greenland in the latest Neogene. It is perhaps worth noting that a roughly 300,000-year warm period in the s associated with the re-invasion of mid-Pliocene Antarctica by Nothofagus Blume forests. (Hill. & Seriven, 1995; Haywood & Williams, 2005), suggest- ing the feasibility of rapid vegetational response to relatively short-term climatic fluctuations. Repenning 1990) provides an interesting discussion of the interplay of topography, climate, and vegetation in the ing short periods of warmth affecting eastern North While such events have not been clearly demonstrated in the ater f the Arctic, duration would make them difficult to detect, and they ale Pliocene and early Pleistocene, hypothesiz- America north to the eastern Canadian Arctic. Miocene or Pliocene of their. brief might have had an effect on the ability of deciduous taxa to migrate across Greenland as late as the Pliocene or earliest Pleistocene. In sum, the fossils and the molecules point toward the need to reassess the possibility of a later Tertiary The NALB may the later Neogene. Arctic exposure of North Atlantic biogeographic. link. geogra| sull However, functional ii =) have been we need more data, especially The sediments with such evidence could become a positive eastern. Canada and Greenland. side to current global warming. — One consistent observation that arises in comparing the paleontological and molecular data as presented rere is that the fossils and the phylogeographic predictions converge but do not agree on the dates of several Neogene migrations. The molecular data seem predict. younger dates than the fossil the uniformitarianism would suggest. to generally occurrences or application of physiological Are paleontologists simply missing data, particularly for short-term climatic fluctuations that open windows of opportunity the Tertiary? Or do disjunctions arise from the vicariance of pre-existing for migration in latest mosl (om distributions, as the fossils often suggest? If the latter mode dominates, the discrepancy in timing observed between the fossils and molecular predictions be- comes a pattern worthy of investigation. Practitioners of both disciplines have to confer to see if a source of consistent error can be identified, again as discussed by several participants in this symposium. DISCUSSION The theme of this symposium was paleobotany in a time of molecular genomics. This could easily be taken as an invitation for paleontologists to weigh the fossil record against the neontological perspectives and tools of molecular systematics. However, at the risk of oversimplification, the question is not the choice of fossils or molecules as has occasionally been propounded by discussants on both sides. Rather, the reality is fossils and molecules. Our two fields exist in a circumstance of reciprocal illumination. Fossils can Volume 95, Number 1 008 Tiffney 141 Phylogeography, Fossils, and Northern Hemisphere Biogeography test. predictions arising from molecular phylogenies, and molecular phylogenies can identify possible times and routes requiring further paleontological investi- gation. To achieve this, both sides must freely discuss their limitations and insights with practitioners of the other discipline. But there is a further insight from this symposium, one I feel has not been properly realized. The central question is not fossils in an age of molecular phylogenetics but rather the role of desktop computing in both disciplines. This symposium is a testimony to the growth in power of molecular techniques in roughly the past two decades. But these techniques are informative only because of our ability to analyze the immense amount of data they generate using desktop computing. I suggest that the same set of tools will lie at the core of the successful integration of molecular and fossil data to understand the evolution of the flora (or fauna) of the last ca. 60 Ma. I believe we will truly advance only when all paleontological data are easily available online so that molecular phylogeneticists can check temporal calibrations and the goodness of ecological fit of their hypotheses— and this is but one use of such data. To this end, if paleobotany is to be a full partner in evolutionary biology in the age of information, then all verified paleobotanical data need to be databased in an internationally available, online system. Then we can use this summed data to seek and test patterns in and fully i interdisciplinary power of phylogeography. evolution and distribution. realize the Literature Cited Avise, J. . Arnold, R. M. Ball Jr., E. dre EJ Neisgl; C. A. Reed & N. C. eee The Lamb, To Intraspecific mitoc niis bridge between population genetics and systematics. Ann. Rev. Ecol. ds 18: 489—5: a Baranova, Y. P., 0. M LV L. P. Zharikova, S Laukuin, A. LS Sergeenko . Fradkina. 1989. m ie ne and Neogene of ihe i pee USSR. Sci. USS th, Yakutsk (In Russian). Bennike, 5 1990. The e Kubenbum Formation: Stratig- raphy and palaeobotany of Peary | Land, North Greenland. Meddel. Grónland, Geosci. Acad. v == Siberian Branc a Plio-Pleistocene sequence in — ——— & J. Bócher. 1990. Forest-tundra neighboring the North ps f Pleistocene Kap Kobenhavn Formation, North Greenland. ic 43: 331—338. J. D. J. Beerling, C. P. Osbor W. M. o. Frances, P. J. Valdes & V. E. Wittig 2005. Climatic and ecologic al determinants of leaf Cretaceous * 95. Plant and insect remains from the Plio- NN in 227 “greenhou — =) Hide and go seek: What does oe Burnham, R. J. 2008 i Ann. Missouri Bot. Gar mean in the fossil record? 51-71 Burke Herbarium. 2006. Arct Jeean molluscan migrations. = a = O = ga zd two “A co g E, Oeri n. 1990, Late Tertiary plant macrofossils on localities in Py d North America: Á review of the e Archie Es A. Westgate, y NE OF ouc ili 2003. Strada py. Tossils and age of sediments at ! Chicken mine: New information on the Late Pliocene environment of east central Pei ies Hes. 60: 9— I. Phylogeny and PS of Rhodo- . Ovenden, the upper pit of the Lost PE tern. dendron a Pontica, a group with a Te i relict distribution. Molec. Phylogen. Ev Nie, Z-L. J. Wen, H. Sun & B. Ba Monophyly of Kelloggia Torrey ex Benth. ( eae between Amer. J. Bot. evolution of its intercontinental western North America and eastern Asia. 92: 642-652. ————, H. Sun, cus hy Lysichiton, and ‘tii of / rn Asia and America. r lec ae n. - Evol. e 55-16 ———, — Mende ley, R. G. e & J. “Wen. 2006h. ex of biogeogr: iphic ea e tween eastern North Amer in iryma 1343-1 356 6. and molecular . Ann. . 2006a. Intercontinental poe 1 . Li & J. subfamily iplocarpus, North —— Asia and (Phrymace eae). Amer. J. Bot. ne Nixon, K. C. 2008. dating: An e dec from the Nymphaeales eastern T Paleobotany, evidence. Missouri Bot. Gard 95: 43—50. Osborne, C. P., D. L. Royer & D. J. Beerling. 2004. di o role of leaf habit in extinct polar forests. Intl. Forest. Rev 6: 181-1 Platt, J. 4 9o |. Strong inference. Science 140: 347—353. Qiu, Y.- . R. Parks & M. W. Chase. 1995. bie e ic e the eastern Asia—eastern North America sje s section rd b a Magnolia Macael ceae). / J. Bot. 82: 1580-1508 Responses of some Southerr Read, He dos re tree s & J. Francis. 1992. species 0 a pne dark period and implications for high-latitude g and Lee pals celine: their Cretaceous pe loras. Palaeoecol. 99: 271—290. Pai Smith. L. M. & E. B. f warm-temperale Leopold. 2003. Miocene pollen rom the Homerian type section near Homer, Alaska: an in-depth study. Geological Society of America Abstracts With Programs 35: 546. ————& —. 2005. Warm climate in the Late Miocene of us south coast. of Alaska and the occurrence of 202. Podocarpaceae pollen. Palvuielory 29: 205- P Repenning. C. 1990. Of mice and ice in the Late Pliocene of e Americ a. Arctic 43: 314-323. Royer, |C. P. Osborne & D. J. Beerling. 2003. Carbon de um in a COs-rich ancient polar » PA environment. ip: 424: )-6 Schnabel, A. & . We n) p Cladistic e f Gleditsia (gumines based on nd nd ci EOM. ast e sequences, Amer, J. Bot. € 1753-17 e——— BE. Donal & ]. 1 ` Wendel. Ha Tots re onda n Glec ce de guminosae) based on ITS Amer. J. Bot. 90: 310—320. Shaw, J. € R. L. Small. n Chloroplast DNA phylogeny and phylogeography of the North American plums (Prunus — subgenus Prunus, section Prunocerasus, Rosaceae). Amer J. Bot. 92: 2011-2030. Thiede, J., A. Winkler, T. Wolf-Welling. O. Eldholm, A. M. Myhre, K.-H. Baumann, R. Henrich € R. Stein. i Late Cenozoic history of the polar Ne Atlantic: Results from ocean drilling. Quatern. Sci. Rev. 17: 185-208. Tiffhey. B. H. 1985. The Eocene | i isnt Land Bridge: Its importance in Tertiary and modern dp org of the Northern e re. J. Arnold Arbor. 66: 243-273. 1994. An estimate of the Early al paleocli- male E the e rn Arctic. Pp. 267-295 tn M. | C. Boulter & H. V. Fisher (editors), Cenozoic H lants and Climates of the n tic. NATO ASI ser. 1: C hange, Vol. 27. Springer, Berlin. ;lobal e ones 2000. Geographic a climatic influences on the Cretaceous and Te d history of > uramerican floristic 6. similarity. Acta Univ. Carol. Geol. 5- ——— & E. S. Barghoorn. 1976. "em and seeds of the Brandon Lignite. l. Vitaceae. Rev. Palaeobot. Palynol. 22: 169-1 a 1. Manchester. 2001. The use of geological and m evidence in evaluating. plant phylogeo- graphic o in the ps rn. Hemisphere Tertiary. Int. J. PL > 62(suppl.): 53-517 Utescher, \ Terrestrial e b Ashraf. 2000. evolution in northwest Germany over 30-449 Fully mp E : VE ORA the last 25 million years. Palaios 15: Tertiary nd Pleistocene deposits and history of Banks Isl ind, UE rn Canadian Arctic Archipelago. Arctic 43: 53. Wen, J. 1999. Evolution of eastert m rican PRO aM in . Ecol. a, ed American bioge graph es A few i. 162(suppl.): 5117-5122. a n. 1995. Te ‘al and molecular 1 Asian ies eastern North lowering plants. Ánn. 421-45 didum of eastern. Asian—eastern North ee issues, .J. PI. ——— va kl comparisons of Campsts is pr and Campsis radicans (Bignoniaceae). an easte Asian and eastern P . American vicariad oa pair. Pl. Syst. Evol. 196: ei oi "— E * Volume 95, Number 1 8 Tiffney 143 Phylogeography, Fossils, and Northern Hemisphere Biogeography —, —— E. A. Zimmer. 1996. relationships e DNA se quence divergence Asian and eastern N Nei & N. Takahata (editors), Current Topics in Molecu lar Evolution. Pennsylvania State Univ. and the Graduate School y e Studies, Hayama, Japan. Whitcher, I. N. & J. Wen. 2001. s and and biogeography Phylogenetic aster of Cor lus (Beralas eae): Inferences from IT'S sequences. Syst. Bol p 283-298. White, J. M., e c D. P. Adam, E. B. Leopold, C. Liu, H. Jetté € 5. Schweger. 1997. An 18 million year record of vegetation and climate change in northwestern Canada and Alaska: Tectonic and global climatic d pes . Palaeogeogr. Palaeoclimatol. Palaeoecol. 130: 293- 1999, Neogene and uate: ernary ae Se asda pol and paleoclimatology from sections in idjacent a hwest Territories and Alaska. Can. Bull. 1-30. 972. An o of Alaskan Tertiary 2 —233 in A. Graham (editor), Floristics and of Asia Eastern. North America. T, cee ds 1977. Paleogene floras from the Es of Alaska 1-10 region. U 5. Geol Surv. Prof. Pap. 997: and & Geol. Su. A. Wolfe, J. and floristic and vegetational changes America. . 1992. Climatic, near the Eocene/Oligocene boundary in North ; Pp. 421-436 in D. R. Prothero & W. A. Berggren (editors), Eocene-Oligocene Climatic and Biotic Evolution. Prince- ton Univ. Press, Princeton. . 1994. An analysis of Neogene climates in Md eringia. 0g P alcogeogr x alaeoc shapes Palaeoecol. 108: 207-216. E. Solti P. S. Soltis. 1998, The eastern Xiang, Q.-Y., Asian and eastern es western North America floristic disjunction: Congruent phylogenetic e in seven diverse genera. Mol. Phylogen. Evol. 78-190. 5. - CHR & dD. J. rn Asian—eastern North corrobo- C tal 2000. ae Tm ea America floristic disjunction: Molec ular rates e 'al estimates. Mol. Phylogen. Evol. ae clock r & J. Wen. 2004. Phylogenetic and Anacardiaceae) in Evol. A. J. Mille grep HOME RUN of Rhus ( Northern. Hemisphere. Mol. Phylogen. 76 — = T. 9. Zachos, J., M. Pagani, L. Sloar Thomas & K. Billups. 2001. Trends, ar dn and Euh in global climate Ma to present. vende 292: 186-693. n. 2006. Molecular phylogeny and intra- and adim Ep ography of Caly- canthaceae. Mol. Phylogen. Evol. 39: 1-15. PLANT PALEOECOLOGY IN William A. DiMichele? and Robert A. Gastaldo? DEEP TIME! ABSTRACT The paleoec ology of plants as a modern discipline, distinct from traditional floristics or cer has undergone an 1e past 20 years. In MET to baseline stu characterizing extinct plants and plant assemblages enormous expansion in t in terms of their growth habits, environmental preferences, and xui. of association, valeoccology has converged on neoecology and represents a means to extend our basic moderada of the world and to contribute to the theoretical framework of s writ large. Reconstruction of whole plants, including studies of physiology and developmental biology. and analyses of biomechanics have become mainstays of autecological studies. Assemblage studies now are informed by sophisticated phon models that have helped g guide sampling strategies and helped with the interpretation of statistical data. Linkages of assemblage patterns in space and time with sedimentology, geochemical proxies for atmos pheric composition and cli melt pru analyses, and increasingly refined geochronological and sequence nc d data have permitted paleoec n EE to examine rates and extents of vegetational response to environmental change and to time intervals. of quiescent climatic cond dons: Studies of plant-animal interaction, explicit consideration of phylogenetic information in assessing assemblage time-space dynamics, and examination of ecological structure in terms of developing metabolic scaling theory are all having = sct impact on paleoecological as well as neoecological studies. The growth of paleoecology shows no a sign iof diminishment—closer linkages with pene elon are needed. ords: Environmental eee paleobotany, paleoecology. taphonomy. The discipline of ecology encompasses the charac- can be the basis for a robust understanding of the terization of multispecies systems, the description of ways in which organisms respond and interact over their spatial and temporal dynamics, and the search short and long spans of time. Such responses may be for organizing principles and general models. Of either to environmental or biologic al crises of differing course, any explanatory models developed from magnitudes, or both studies of modern systems should apply consistently This paper is intended to review the broad field of to long-extinet ecological realms as well, if they are to plant paleoecology in the pre-Pleistocene, including claim generality. Thus. in the largest theoretical the fundamental building blocks of the discipline, as sense, ecology not only should, but must, include and well as considerations of the application of the basic integrate data from modern systems with those from data to problems of broader theoretical interest and the past. The richness of empirical neoecological — importance. We do not believe that any caveat or studies, which include more observable system-level apology is necessary for the various limitations dimensions than paleoecology (not the least of which presented by the fossil record. The matters at hand is the possibility of experimental manipulations), — are: what can be done with the record: to what degree nonetheless is limited temporally. Paleontology pro- can ils undoubted biases be understood, considered, vides access to biological phenomena on a wide and accounted for; how can one evaluate those data in spectrum of time scales. But, in contrast to neoecol- real time to understand the ecological dynamics, ogy, it has limited spatial resolution at finely re- preserved therein; and how can models based on solved time scales, is restricted to comparative extant systems be tested against that record? It is the data only, and includes many systems that are intent of this contribution to provide a way into the composed of organisms about which biological literature of this rapidly growing, increasingly diverse understanding may be limited. However, neobiology field, while trying to identify the major areas of and paleontology overlap in many ways and together research, mature and just developing. ! WD acknowledges support of the Evolution of Terrestrial Ec osystems Program of tlie National Museum of Natural History. We wish to thank Bill Crepet for the invitation to participate in the symposium in which this paper was ae presented. We thank Scott Wing, Caroline Strémberg, Peter Wilf, and C a Labandeira for helpful discussions. In addition, we wish to thank the numerous colleagues who responded to our request during the summer of 2006 for their t ear on the state and future directions in p paleoecology: we are deeply indebted to them for their thoughtfulness and candor. This is publication 102 of the Evolution of Terrestrial Ecosystems Program at the 3mithsonian Institution. ? Department of Paleobiology, bonds Smithsonian Institution, Washington, D.C. 20560, U.S.A. dimichel@si.edu. ‘Department of Geology, Colby College, 4000 Mayflower Hill, Waterville, Maine 04901. U.S.A. ragastal@colby.edu. dor: 10.3417/2007016 ANN. Missouni Bor. Garp. 95: 144—198. PUBLISHED ON 11 ApriL 2008. Volume 95, Number 1 2008 DiMichele & Gastaldo Plant Paleoecology in Deep Time 145 Mayor AREAS OF RESEARCH Paleoecology is a discipline that includes research a approaches that range from morphology and floristics, which might be considered the traditional strengths of paleobotany, to some that were embryonic 25 years ago, such as growth modeling, evaluation of physio- logical proxies, and linkages between biotic and abiotic studies (geochemistry, sedimentology. paleo- pedology). The immediate questions of interest in each of these areas vary considerably. The following discussion addresses what seem to be the major foci of current research. Basic systematic investigations remain the core of paleobotanical research, and the continuance and importance of those endeavors are implicit throughout this paper. Alpha systematics provides the basic language with which we describe multispecies systems, and phylogenetic analyses are an increasingly important part of ecological studies. AUTECOLOGY The reconstruction of past organisms, as much possible as whole plants, is probably the fundamental building block of paleobotanical research and, with a twist, of paleoecological studies. The key that takes this endeavor into ecology is the step from systematics as the end goal and the move into the ever-growing — focus on functional morphology, growth architecture, and whole-plant reproductive biology, which ulti- mately aims at understanding autecology in as much detail as possible. The real challenge in this research is to see the past on its own terms. In other words, the challenge is to imagine and allow for the possibility that extinct organisms may have functioned in ways not known today, and that such function will, or can, be revealed in the gross morphology of the organism. FLORISTICS AND VEGETATION STUDIES Description of fossil floras, along with basic morphology. is perhaps the longest-standing research area of paleobotany. Development of protocols to describe floras quantitatively, to sample them in ways that permit reconstruction of original vegetation, and to analyze them statistically, is a major area of study and a building block of larger studies of ecological dynamies in time and space. RESOLUTION Considerable effort has been devoted to taphono- my—the study of the fossil record and the processes that underlie its origin. Perhaps the most fundamental paleoecological question in taphonomy is that of spatio-temporal resolution. Before the record can be used to address questions in ecology. sensu lato, it is necessary to consider the limitations of the record in candid detail. DYNAMICS OF ECOSYSTEMS IN RESPONSE TO PERTURBATIONS Paleontology may be able to contribute uniquely to ecology through the examination of multispecies assemblage dynamics over time intervals not acces- sible to neobiological analysis. The most publicly visible of these kinds of studies are those that examine biological responses to major environmental pertur- bations, such as catastrophie environmental change (brought on by such things as an asteroid impact or massive volcanism), major climate change (global to regional temperature increases or decreases linked to precipitation patterns), or linkages between biotic e dynamics and changes in atmospheric composition. Presently, researchers are, or have been, focused on several time intervals of major extinction in Earth's history, and one can ask just how much can be generalized from any one of these. In other words, are there dependencies or contingencies unique to any that generalization from it alone? Is any one extinction system at any point in deep time limit interval unique, and thus, not of broader interest? Are there temporal trends suggesting the evolution of system-level responses through time? ECOSYSTEM ASSEMBLY OVER SHORT AND LONG TIME SPANS There has been a long-standing discussion in ecology about the existence of levels of organization above species populations, but can such systems even be circumscribed realistically? Is there the possibility of emergent properties in multispecies assemblages? Consider the early debate between adherents of the superorganism versus individualistic schools of Clem- ents (1916) and Gleason (1926). A vast exists about the concept of equilibrium in ecological (e.g.. MacArthur & Wilson, 1967), extending to the relationship between ecological — iterature organization stability and complexity (May, 1973), and systems ecology (Odom, 1983). Recent debate has centered on Hubbell's (2001) neutral theory of ecological organi- zalion, including such concepts as metacommunities (Holyoak et al., 2005) and assembly rules (Belyea & Lancaster, 1999; Weiher & Keddy, 1999). There also is increasing interest in the relationship between the historical-phylogenetic structure and the ecological architecture of assemblages of organisms (Shugart, 1997; Webb et al., 2002 Paleontological data can be construed to indicate — long periods of ecological persistence, even in the face of environmental fluctuations (e.g.. Brett et al., 1996, 146 Annals of the Missouri Botanical Garden 2007; DiMichele et al., 2004). And although not particularly couched in the terminology of assembly — rules, it does appear that ecological systems under- went long periods of evolutionarily driven assembly, controlled by processes not unlike those proposed for very general rules of ecological assembly in space (compare Valentine [1980] with Belyea € Lancaster [1999)). If there are very general rules, to what extent are they taxon or assemblage specific? More recently, intriguing work with great potential impact has been conducted on scaling relationships among plant-body sizes, metabolic rates, and ecological factors such as self-thinning of both populations and multispecific stands (e.g. Niklas € Enquist, 2001; Niklas et al., 2003), which may relate directly to the manner in which plants sort out resources within multispecies assemblages. Aspects of the first two topics have been investigated in various deep-time taxa (e.g., rhyniophytes and zosterophylls, cf. Roth-Nebel- sick et al., 2000; Roth-Nebelsick, 2001: angiosperms, cf. Roth-Nebelsick et al., 2001) as well as in universal generalities (e.g., Roth-Nebelsick et al., 1994a, b), with the possibility of applying these results to assemblage- level communities over the Phanerozoic. ARE THERE UNIQUE PATTERNS RECOGNIZABLE ONLY ON LONG TIME SCALES? Are there kinds/types/classes of relationships be- tween ecological and evolutionary dynamies that can be understood, or even simply recognized, only within the context of the geological-time rate spectrum (see below)? The recognition of such patterns will likely be an empirical outcome only of trends identified within the fossil record. Tur GEOLOGICAL-TIME RATE SPECTRUM There is no single entity that can be characterized as Geological Time. Rather, the geological record preserves a wide spectrum of resolvable time intervals. Thus, the problem is one of time resolution, A major element of time resolution in the geological record is lime averaging (see Behrensmeyer et al., 2000), which is more pronounced in the invertebrate and vertebrate fossil records than in that of plants. The stratigraphic record includes plant-bearing beds of various attributes, each of which represents accumulation of debris within one original environ- — ment. Each genetic assemblage (the stratigraphic interval over which the assemblage is preserved) also can be considered to be a time layer of different irreducible temporal thicknesses. These may range from those preserved instantaneously (10%-year de- posits), such as forest floors buried under volcanic . 1993), to those that are catehments of material accumulated over hundreds of ashfalls (e.g., Wing et a years, such as tidal estuarine settings and packrat midden deposits (e.g.. Mazon Creek, Pfefferkorn. 079; Pearson & Betancourt, 2002), to many thousands of years, such as deposits of resistant material floated into marine depositional environ- ments (London Clay, Reid € Chandler, 1933). The thickness of a rock unit is not related necessarily to the amount of time it encapsulates. Thus, the biological assemblage from a given deposit (bed or horizon) may have accumulated over some period of time that cannot be further subdivided, depending on the nature of the rock unit in which it is embedded. lt may be temporally irreducible at the time scale over which the deposit accumulated, which is its level of depositional (or taphonomic) time averaging. Such deposits can be combined into meta-assemblages, each of which can encompass a hierarchy of time units of variable length. Such units can be described as artificially or analytically time averaged (Behrens- meyer & Hook, 1992). Consequently, any deposition- ally or analytically time-averaged fossil deposit can be analogized to a photographie snapshot of the fossil record but taken with variable shutter speed. These shutter speeds range from ultrafast, such as in situ burial of a forest-floor litter (e.g., Gastaldo et a 2004b), to ultraslow, such as the diversity of Linnean families sampled globally and lumped into geological stage-level bins (Rees et al., 2002). In the latter case, .. we can imagine shadowy species moving in and out of the exposure, much like the ghosts of people in 19th- century long-exposure negatives. Thus, the geological record preserves, in natural and artificial packages, a broad spectrum of time resolution. These time packages can be combined in various ways and manipulated analytically and statistically, providing a very powerful tool rather than irresolvable obstruction. APPROACHES TO THE STUDY OF PALEOECOLOCY At its most fundamental level, plant paleoecology is built on studies of individual plants and reconstruc- tions of plant assemblages. In each case, there are various, often high degrees of uncertainty that must be overcome or accommodated for credible interpreta- tions to be made. Taphonomy plays a role at all levels of inference in structuring interpretations and must be considered explicitly as a key element in all paleoecological analyses. PALEOAUTECOLOGY Reconstruction of whole plants. Comparative mor- phology has long been the core of paleobotanical Volume 95, Number 1 2008 DiMichele & Gastaldo 147 Plant Paleoecology in Deep Time research. The ultimate goal of this research is to elaborate and understand the history of plant life on Earth. However, because of their modular construction and open, continuous manner of growth, whole plants are rarely preserved, instead tending to fall apart into a variety of organs. Putting these plants back together is a challenge. Not only is the objective to determine what isolated fossil organs go together but, where possible, to determine such things as the timing of reproduction relative to the plant’s growth and life history or to variations in its he reproduction physical environment and, more removed, proportions of biomass allocated to versus vegetative growth, to the extent this can be estimated. Perhaps the greatest biofantasy of whole-plant that to ascertain reconstruction is the the association in an acceptable scientific manner. It is misconc eption organ attachment is only way organ necessary to recognize that reconstructions of fossil plants are hypotheses, subject to test and revision. The more explicitly they are reconstructed and the more clearly the bases for reconstruction are made known, the more rigorously they can be evaluated by the First of all, attachment unique addition of new information. hardly provides certainty of association (e.g.. the stigmarian root systems of Paleozoic arborescent lycopsids, which are found in attachment or association with many different species of several genera of stems belonging in different families; Bateman et al., 1992). More important is the degree to which this perspective denies the basic probabilistic nature of the reconstruction process and the fundamentally inductive, statistical nature of science. Likelihood is the main means by which the reconstruction of extinct plants must proceed if we are than a handful of species ever to have confidently reassembled. The underlying means for more such reconstructions include recurrent association at sample sites, association at sites with only one kind of each organ, anatomical and morphological similarities among dispersed organs, and similarities in the morphology and anatomy of organ cuticles and epidermal cell patterns. The most important point may be that in paleoeco- logical analyses, it is necessary to work with a variably complete understanding of the biology of the organisms, and keep hypotheses congruent with what is known or what can be reasonably inferred from the data. Some recent examples of the application of these various approaches to whole-plant reconstruction are highlighted below. There are a relatively large number of plants, too many to cover exhaustively here, that have been completely or partially reconstructed. by various means, but these still represent just a tiny fraction of the plant species thought to have existed in the geological past. Reconstruction from organ attachment is the most difficult of all ki attention to fine detail as well as a little luck. Perhaps kinds to undertake and requires the greatest abundance of whole plants is found among the early land plants of the Late Silurian and Early to Middle Devonian. Because of their simple morphol- ogies, complete or nearly complete plants, including reproductive organs, often have been found and reconstructed. Such reconstructions have been effect- ed directly and most often from compression fossils, some of which also have anatomical details preserved via permineralization. A few examples include the enigmatic fungus Prototaxites Dawson (Hueber, 2001; oyce et al., 2007), the lycopsid Sawdonia ornata (Dawson) F. M. Hueber (Hueber, 1971), the trimer- ophyte Pertica quadrifaria Kasper & Andrews (Kasper & Andrews, 1972; Gensel & 1984), and numerous plants from the classic Rhynie Chert in Andrews, Scotland, which usually have been reconstructed by finding parts attached in series rather than as single whole plants (Paleobotanical Research Group, Uni- versity of Münster, 2007). One of the most puzzling nearly whole plants is Cooksonia W. H. Lang, often touted as the earliest vascular land plant but known 1992). Where these only from aerial parts (Edwards et al., were to which presumably were attached? Rothwell (1995) suggested that Cooksonia may have been a sporophyte incapable the prostrate axes organs of a free-living existence, instead growing attached to a photosynthetic gametophyte, much like modern moss sporophytes and, thus, far from completely known morphologically. Gerrienne et al. (2006) found a cluster of Early Devonian Cooksonia axes attached to a thalloid-like pad and offered three possible interpretations, among which was the possibility that the axes represented sporophytes attached to a gametophyte. This hypothesis has been given biome- chanical and physiological support by Boyce (2006), who demonstrated that most Cooksonia species are too narrow to have had sufficient photosynthetic tissues to sustain themselves independently. The earliest fully reconstructed seed plant, Elkinsia Rothwell, Scheckler & Gillespie (Rothwell & Serbet, 1992; Serbet & Rothwell, 1992), is known to a degree that the timing and extent of reproductive allocation can be evaluated (e.g.. Rothwell € Scheckler, 1985; Scheckler, 1986). Reconstructions are not limited to members of the understory; creeping ground cover and/or lianas are also are well represented (Manches- ter & Zavada, 1987; Krings & Kerp. 1999). The late Paleozoic gymnosperm Callistophyton (Rothwell, 1981) may be reconstructed plants of the entire geological record. Delevoryas one of the most fully 148 Annals of the Missouri Botanical Garden The detailed nature of Rothwell’s (1981) reconstrue- tion, based on permineralized anatomically preserved specimens, has permitted some revision based on findings from compression fossils (Galtier & Béthoux, 2002), illustrating the concept of reconstruction. as testable hypothesis. — Whole-plant preservation is most likely for smal For example, the only known herbaceous Aethophyllum Brongn. from the Triassic, was plants. conifer, reconstructed from a complete plant, a finding made perhaps more likely because of the small stature of the plant (Rothwell et al., 2000). Some other whole plants found in the fossil record lycopsids Clevelandodendron Chitaley & Pigg (1996), Chaloneria include the isoetalean 2o of Late Devonian and Pigg Rothwell (1983) and Hartsellea Gastaldo, Blanton-Hooks (2006) of Pennsylvanian age. all with pole-like growth forms and comose base. The former is age, Gibson & fertile zone. a single specimen terminating in : whereas Chaloneria was put together from a number of variably complete specimens, terminating either in a fertile zone or with intercalated fertile and sterile zones. Hartsellea. on the other hand, was preserved erect and in situ within a back-barrier marsh. Escapa and Cúneo (2005) reconstructed the early equiseto- phyte Peltotheca furcata Escapa & Cuneo from segments of plants that permitted linkage of repro- ductive and vegetative parts. Small aquatic plants occur in that preservation. Examples include Archaefructus liao- Dilcher, e & Zhou, an aquatic angiosperm (Sun el a 002a). pinnata Rothwell & Sader | (1994), a heterosperous taphonomic settings favor intact ningensis Sun, and Hydropteris aquatic fern, each proposed as the basis of new families of phylogenetic significance Larger shrubs and/or trees are extremely difficult to find preserved as whole plants. Rare discoveries of . Stidd & Phillips, 1968) can provide information about development and assist in the adult only from The discovery of whole trees does juveniles (e.g reconstruction of plants known isolated organs. occur occasionally, however. For example, Wnuk and fefferkorn (1984) discovered a forest of lyeopsid trees with medullosan understory blown down during a storm and buried in place. The medullosans con- formed to two growth habits. In one, trees were solitary, upright in growth stature, with trunks more than 20 em in diameter and less than 5 m in height. Fronds were closely spaced, about 10 per stem length, and were curved downward after death, suggesting development of a skirt, a morphological means of protection against epiphytes and vines. The other form was found in groups, suggestive of thickets in which the large fronds intertwined, allowing the trees to be mutually supporting. Stems were flexuose. meter of me up to 13 em in diameter and greater than 5 m in height, with widely spaced fronds that appear to have been shed regularly rather than remaining attached. Pfefferkorn et al. (1984) evaluated published recon- structions of medullosans, which show a wide range of variation on the flexuose versus upright habit and determined that most were considerably inaccurate. Unfortunately, while revealing a great deal about medullosan growth habits, the actual foliage of these plants was not preserved attached; hence, taxonomic identity could not be established. Partial reconstructions from attachment may some- times have greater ecological significance than full reconstructions. Consider, for example. the Mesozoic eycadeoids, reconstructed primarily from silicified found at trunks, in which mature strobili were developmental states suggesting monocarpic (or “pseudomonocarpic”) reproduction (Wieland, 1921). Similarly, the reconstruction of the false-trunk fern Tempskya Corda from the Cretaceous reveals a growth form unknown among ferns or seed plants today (Andrews & Kern, 1947). stem/wood, sepals. and stamens with included pollen Fortuitous connection of permitted the reconstruction of several plants by Manchester and colleagues (Manchester & Crane, 1983: Manchester et al.. 1986: Manchester, 1989): we consider these to be “partial” reconstructions because the attached parts are from branches of presumed the although representing vegetative an trees, and because suite of organs found in attachment, reproductive parts, is an incomplete set. Probabalistic from dispersed organs, means of assembling whole plants rather than through organ attachment, have produced a large number of reconstructions and given us considerable insight into the biologies of the plants. Paleozoic plants, in particular, are challenging because they generally have no close living relatives. Pennsylvanian arbo- rescent lycopsid reconstructions, for example, histor- ically illustrate egregious errors, particularly as related to the dynamic biology of the plants. Individually or in forest stands, these trees are almost — always reconstructed along the lines of angiosperms. Trees are given light-capturing crowns and frequently are shown tipped over with their root bases in the air, presumably from blowdowns. However, evaluation of the development of these plants reveals dramatic architectural changes from juvenile to reproductively mature adult, which has major implications for individual plant and landscape reconstructions. Stands of these plants probably looked little like their representation in dioramas and paintings. Fossil data suggest two major growth forms, with different reproductive biologies. In one type, which is the most widely reconstructed, individual trees spent nearly all Volume 95, Number 1 08 DiMichele & Gastaldo 149 Plant Paleoecology in Deep Time of their lives as unbranched poles, forming a crown only at the end of tree life, which was determinate (Andrews € Murdy, 1958; Eggert, 1960). Anatomical (DiMichele, 1979) and compression (Wnuk & Pfef- ferkorn, 1987) studies indicate that cones were produced on these plants among the branches in the crown during the terminal phases of growth, rendering them effectively monocarpic (Bateman & DiMichele, 1991; Bateman, 1994). That cones were borne only in the terminal-phase crown suggests that height and tree — crowns had little to do with light capture and much to do with dispersal of reproductive organs (DiMichele & Phillips, 1985; Phillips & DiMichele. 1992). The other growth architecture consisted of a determinate monopodial trunk bearing determinate deciduous lateral branches (DiMichele, 1980). Cones were borne among the lateral branches in some taxa or were borne directly on the trunk in others. In either case, the trees were polycarpic. The small, deciduous lateral branch systems were too small to account for much light capture in most species with this life strategy. These peculiar trees had tenacious root systems that spread far from the trunk at shallow depths. The longest recorded stigmarian axis approaches 13 m. Neither of the authors have ever seen a tipped-up tree in their many visits to mines where stumps of these trees are exposed (e.g., Gastaldo, 1986; DiMichele & DeMaris, 1987; Wnuk «€ Pfefferkorn, 1987; Di- Michele et al, 1996; Gastaldo et al., 2004a) nor have they read of any such tip-ups reported in the literature. Yet, uprooted lycopsids are staple elements of Carboniferous swamp reconstructions, implying that blowdowns and uprooting, for which there is no evidence, were parts of the disturbance mode in such systems. Finally, anatomical studies of arborescent lycopsids raise certain matters with regard to their physiology (Phillips & DiMichele, 1992). The lack of a clear phloem connection between root and shoot, the generally limited phloem throughout the plant's aerial shoot, the leaf-like rootlets borne on the stigmarian axes, and the long leaves on stems and cones, are all consistent with extremely localized use of photosyn- thate and perhaps even self-supporting root systems in carbon fixation. The purpose of this long = terms o discussion is to illustrate how much can be known and understood about fossil species from disconnected organs when intense work by many investigators is integrated into a series of reconstructions of consid- erable, though hardly absolute, certainty. more remarkable given that a whole lepidodendrid This is even lycopsid tree has never been found intact from top to bottom. Disconnected organs can be linked into nearly living whole plants by using common features of gross morphology. An example of this comes from the work In a series of papers X Krings and colleagues. (summarized in Krings et al, 2003), they used cuticular features and gross morphology to link disconnected organs and reconstruct liana and vine- like habits in several species of Pennsylvanian-age pteridosperms. They detail a variety of climbing mechanisms, from hooks to pads to tendrils, that have structural analogues in taxa of other clades. In addition, they note much higher relative abundances and diversities of these plants in Late Pennsylvanian environments compared to environmentally similar Karly and Middle Pennsylvanian tropical ecosystems. Clearly, there were major changes in the dynamics of these tropical forests, associated. with major species turnover and change from lycopsid to tree fern canopy that dance (possibly disturbance and patchiness of the 2004b). Strong patterns of organ association dominance, favored the increase in vine abun- forests; see Gastaldo et al., also allow whole plants to be reconstructed from disconnected organs, especially when an assemblage consists only of one of each organ type attributable to a particular higher-level group of plants. This is strengthened further by repeated occurrences of organ associations at multiple collecting sites. An example of this comes from the work of Kvacek et al. (2005), Cretaceous who proposed a reconstruction of a Late ginkgoalean consisting of the following dispersed organs: wood (Ginkgoxylon greutii Pons & Vozenin-Serra), short shoots (Pecinovicladus kvacekii Faleon-Lang), leaves (Eretmophyllum obtusum (Velen.) Kvacek), ovulifer- ous organs (Nehvizdyella bipartita Kvacek, Falcon- Cycadopites Wode- a Lang € Dasková), and pollen different collecting sites in consort with organs of house). found this same association at four other types, none of which were ginkgoalean in affinity. In addition, due to anatomical preservation of some remains, they were able to link the various organs sequentially in terms of their relative spatio- structural position on the plant. Finally, these plant remains occurred in a specific, recurrent lithofacies, indicative of peat deposits formed in salt marshes (making this the first described salt-tolerant mangrove ginkgophyte and indicating that the biology of this plant cannot be understood by analogy primarily with modern representatives, of which there is only one in the case of ginkgophytes). Other examples include Archaeanthus linnenbergeri Dilcher & Crane, an early angiosperm, reconstructed from partial attachment and associated organs, using anatomical similarities and the presence of distinctive resin bodies (Dilcher & Crane, 1984), and Pentoxylon Sahni. reconstructed from repeated associations of leaves, stems, and reproductive organs found in detailed sedimentolog- 150 Annals of the Missouri Botanical Garden M ical context (Howe & Cantrill, 2001). Examples of organ linkage into whole or partial plants can be found in many groups and, remembering the hypothetical nature of strongest and most likely means of developing such reconstructions, is probably one of the associations. The use of large samples from single sites also facilitates the plants from isolated and partially preserved organs. An example -Castillo on the early reassembly of whole of this is the work of Hernandez walchian conifer Thucydia mahoningensis Hernandez- Castillo, Rothwell & Mapes (Hernandez-Castillo el al., 2001, 2003). Hernandez-Castillo and colleagues were able to reconstruct this species, including growth habit, from a large population of fragmentary compression fossils, some with preserved anatomy, including leafy branches, stems, and reproductive organs, and one specimen with branches attached to a They concluded that early walchian generally stem segmenl. conifers were considerably smaller than envisioned, perhaps reaching only 2 m in height, with architecture much like juveniles of the araucarian conifer Araucaria heterophylla (Salisb.) Franco (Nor- folk Island wal- chians, although of larger size, have been suggested (2002a) based on much smaller Pine). Similar reconstructions of et al. by Ziegler populations of fossils. The larger-scale study informs and adds confidence to these other investigations and interpretations. Whole-plant concepts can also be created by direct comparison with living descendants or close relatives. This is especially true of deciduous angiosperms, where isolated leaves make up nearly all of their fossil record, and post-Paleozoic conifers, where many groups have living descendants. It is also true of ferns, where most of the modern families appeared in the Mesozoic (Rothwell, 1994; 2002; van Konijenburg-van Cittert, 2002a, b). Through a combination of morphologic and taphonomic/sedimen- Collinson, tologic analyses and comparison with modern species, a complex of early Tertiary elms has been character- ized and examined with regard to changing ecological preferences through time (Burnham, 1983, 1980). Taxa were found to have high degrees of ecological fidelity over millions of years. For conifers, pd the Eocene age mummified Metasequoia Hu & W. Cheng forests found north of the Arctic a Canada (Williams et al., 2003a. b). Here, stumps were found in association with mummified logs and thick beds of peat containing three-dimensionall y bul leaves in profusion, pe rmitting reconstruction of us can in dn preserved, disconnected, cones and plant. biology and landscape structure. These ye compared directly to living Metasequoia from which a wide range of inferences is possible, including physiology (LePage et al.. 2005). An example from Mesozoic ferns is that of Skog and Dilcher (1994) from Dakota These deposits the fern families Schizaeaceae, the Cretaceous Formation. include species of Matoniaceae, and Gleicheniaceae in wetland deposits, where they remained dominant even as angiosperms were increasing in other habitats. Today, members of these families form expansive tropical fern brakes in open areas or occur as undergrowth in more open forests. Direct evidence for this in the past is limited, but Skog and Dilcher (1994) infer this as a likely habit for the fossil taxa by analogy to situations. They also propose that some Dicksoniaceae based solely ea these modern may have formed dense thickets, on analogy to modern forms. In the absence of grasses, it has been suggested that ferns played a similar role, forming extensive areas of ground cover (LePage & Pfefferkorn, 2000). Reproductive biology. The study of reproductive biology is a central part of both modern and fossil plant ecology. The most common approach, based on sheer numbers of published papers, has been the study of reproductive isolated entities, generally placed within a particular evolutionary lineage or higher taxon but not tied to a particular whole-plant species (e.g., Taylor, 1982). Such studies have been compiled into analyses of patterns through most often with a focus on morphological evolutionary history 1965). Other have focused on functional morphology of organs as time, (e.g. Taylor, studies reproductive organs, with a particular focus on seeds and pollination biology (e.g.. Rothwell, 1972). Some of these studies have been experimental, using physical models, sometimes including mathematical modeling as a component (e.g., Niklas, 1981). A few analyses looked at the distribution and relative abund- organs of have ances of reproductive versus vegetative single species in attempts to estimate reproductive allocation and dispersal behaviors (e.g., Schwende- mann et al., 2007). The study of reproductive organs occupies too much of the literature to be detailed here. However, some compilations include Taylor (1978) on Paleozoic seed- fern pollen, Millay and Taylor (1979) and Rothwell and Eggert (1986) on Paleozoic seed-fern pollen organs, Tiffney (1986) and Sims (2000) on the distribution of seed sizes and dispersal syndromes through geologic time, Tiffney (1994, 1999) or Gee 2005) on seeds and fruits from Tertiary deposits, Crookall’s (1976) seeds, the massive compilation of seeds and fruits in the Eocene London Clay flora (Reid & Chandler, 1933: Collinson, 1984), and the compilation of late Paleozoic marattialean reproductive organs by Millay — summary of Late Carboniferous [evum Volume 95, Number 1 2008 DiMichele 8 Gastaldo Plant Paleoecology in Deep Time (1979). In organs can yield insights into likely pollination some instances, isolated reproductive — mechanisms, especially when comparisons can be made to extant relatives (Klavins et al., 2003) Pennsylvanian arborescent lycopsid reproduc- tion is understood in some detail. This is made possi- ble by a wide range of studies of the structure and function of microsporangiate and megasporangi- ate cones (e.g., Brack-Hanes, 1978; Phillips, 1979; Brack-Hanes & Thomas, 1983; Pigg, 1983), linkages between microsporangiate cones and dispersed spores (which also permits differentiation of species from peat vs. clastic substrates [Willard, 19892, b]), linkages between parent plants and reproductive 92. for summary), that tied together growth architecture and reproductive patterns (DiMichele & Phillips, 1985; Phillips & DiMichele, 1992). Quantitative studies of coal balls (permineralized peat) from Pennsylvanian-age coals (e.g., P 1977; Phillips & DiMichele, 1981) have comes coarse-grained studies of prop- agule distribution patterns; comparison of the dis- organs (see Bateman et al., 19 and morphological studies, as detailed above. illips et al., tribution of certain seed-like megasporangiate units with vegetative parts of the parent plants (Lepido- carpon Williamson vs. Lepidophloios Sternberg) dem- onstrates that the propagules are dispersed more widely than vegetative material, consistent with dis- tribution by flotation under aquatic swamp condi- tons and an invasive distribution strategy. Once more, this kind of analysis is made possible by a combination of detailed morphological analyses, whole-plant reconstructions based on likelihood of — association, and study of statistical patterns of plant distribution in time and space. The ecology of reproduction in ferns is facilitated by the common presence of spore-producing organs in intimate association with laminate, photosynthetic foliage. In addition, it is often possible to make direct comparisons between ancient and modern fern bi- ology (Page, 2002) because many extant fern lineages were in existence during the Mesozoic, and, addi- tionally, all ferns share certain reproductive and life-history traits. The late Paleozoic marattialean dominant elements of wetland ecosystems (Phillips et al., 1985) and are from both anatom- — erns, for example, were very well known morphologically, ical (coal-ball) studies and compression-impression fossil analyses. More cheaply constructed in terms of carbon allocation than any other group of Pennsyl- vanian welland plants (Baker € DiMichele, 1997), with tree habit made possible by a mantle of ad- filled (Ehret € these massive ventitious roots with airspaces Phillips, 1977), numbers of highly dispersed spores, permitting them plants produced to act as invasive weeds as well as forest canopy dominants (DiMichele & Phillips, 2002). Many of the basal extant families of filicalean ferns arose dur- ing the early Mesozoic (Tidwell & Ash, 1994; van Konijenburg-van Cittert, 2002a, b), their importance tapering off into the later Mesozoic and Cenozoic (Collinson, 2001; Wang, 2002). However, during the later Mesozoic, ferns remained ecologically dominant in many habitats and in many ecological roles, from colonizers of disturbed substrates to biomass domi- nants (e.g., Wing et al., 1993; van Konijenburg-van Cittert, 2002a, b). to angiosperms in most floras from this time inter- Cenozoic ferns are subordinate val, leading to a comparatively understudied record relative to ferns from older eras (see summary of Collinson, 2002), but are known to have occupied a range of habitats similar to extant ferns. In some instances, it takes years to put together a picture of the biology of a fossil plant or group of plants. Reproductive biology is generally the most elusive part of such a reconstruction. A case in point is the biology of the Bennettitales, an important and diverse group of plants in many Mesozoic ecosystems that included both outerossing and potentially breeding species. Members of the Cycadeoidaceae M. n- were considered to be monocarpic (Wieland, 1921) and likely animal, possibly beetle, pollinated (Crepet, n some species of the Williamsoniaceae, P No] -l however, anatomical evidence supports wind pollina- tion, seed dormancy, and outcrossing (Stockey & Rothwell, 2003). Modeling studies of reproductive dynamics have been applied mainly to early land plants, where modern analogues of the fossil morphologies are lacking. Niklas (1981, 1983a) built scaled-up models of Late Devonian and Mississippian seeds bombarded them with jumbo-sized, grains in wind tunnels. The seeds studied varied from model pollen those with completely unfused integuments and megasporangial modifications that facilitate pollen capture, described as hydrasperman reproduction (Rothwell, 1986), to those with closed integuments, some further enclosed in cupules. Such models allow direct testing of functional morphological scenarios. They also make specific predictions about how pollination biology of such ovules may have func- tioned, which can be tested against the fossil record (see comment and response from Rothwell & Taylor, 1982; Niklas, 1983b). Preliminary studies also have been carried out on pollen sedimentation rates, which have permitted testing of hypotheses about wind certain For versus animal pollination in groups. example, how do large medullosan pteridosperm pollen grains settle in liquids, such as pollen drops, such as those relative to saccate pollen types, 152 Annals of the Missouri Botanical Garden produced by early conifers (Schwendemann et al., Study of the tissues has Biomechanical attributes of fossil plants. biomechanical properties of plant permitted insights into growth and development that are difficult to attain by other means of reconstruction, Such approaches can be of great value even when only Speck, 1994). The grouping and distribution within the stem o parts of plants can be evaluated (e.g. anatomically preserved tissues can be analyzed and compared directly with modern analogues because 0 the physical nature of the system (Niklas, 1992). Study of the growth nian-age plant, Sphenophyllum oblongifolium (Germar & Kaulfuss) Unger by Galtier and Daviero (1999) illustrates a statistical approach to understanding architecture of the Pennsylva- plant growth form and its structural and ecological implications. By examining patterns of internode length, axis diameter, and leaf length, these authors were able to examine the growth ontogeny of the plant and infer scrambling, thicket-forming habit. An example of contrasting approaches with con- vergent conclusions is demonstrated for the relatively well-known Middle Devonian aneurophytalean pro- Beck, 1957), which was originally reconstructed as a small eymnosperm Tetraxylopteris schmidtu Beck tree, perhaps as much as 3 m in height. Speck and Rowe (2003) studied the plant's anatomy in addition to growth architecture and concluded that it was nol like ly to traditional but ive been self-supporting. In a more extremely detailed morphological analysis of a related species, T. reposana Hammond & Berry, the authors (2005) came to a conclusion growth similar regarding habit. Tetraxylopteris Beck is now considered to have been a small bushy thickets of Fertile branches were lense interlocking plant that formed branches. ikely borne on the more distal parts of the plant, where the sporangia would be most exposed. The combination of such = studies provides insight into this entire group of extinct plants and suggests that this kind of growth form may have been characteristic of the clade (Speck & Rowe, 2003). Similar approaches have been used to examine growth form of the Calamopityaceae, an extinet group of Mississippian pteridosperms. These were small plants with limited secondary xylem, probably scrambling ground cover and/or vines. Rowe et al. (1993) analyzed their stem architecture and deter- mined that internodal distances are relatively short in arge-diameter stems but become increasingly long in smaller-diameter stems. The decrease is so dramatic that the small stems appear to have been incapable of self-support. Hence, it is possible that the mature — — plant had an upright or corm-like basal portion with a trailing or semi-self-supporting upper part. In con- trast, Speck and Rowe (1994) conducted a biome- chanical analysis of the much larger Mississippian pteridosperm Pitus dayi Gordon and concluded that it was certainly self-supporting and abscised its frond- like leaves. The potential power of biomechanical studies is additive. In isolation, they may enhance understand- ing of the ecologies of individual plants. But, in aggregate, they allow those individual ecological insights to take on considerably greater significance. Most studies, biomechanical and architectural that look een aimed at understanding evolution- multi-taxon especially those broadly across clades, have ary phenomena (e.g.. Niklas, 1999). However, such bmi antal importance in | Niklas and Speck (2001), who investigated the ae of structural studies can also be of ecology. Consider the study o safety factors to wind shear in early land plants, revealing distinct trends both through time and with regard to plant height, largely independent of clade membership. The phylogenetic independence of the trends is consistent with biomechanical properties of common structural elements and their stereotypical spatial deployment within plant stems. This approach opens the door to the estimation of wind effects in the context of its effect on stand density, plant morphol- ogy. and the relationships between wind, plant architecture, and. reproductive strategies. PALEOSYNECOL PUTTING TOGETHER A COMMUNITY The reconstruction of past assemblages of plants is another of the central, and traditional, pursuits of terrestrial. plant. paleoecology. Until recently, such endeavors could be described broadly as floristics, the elaboration of a taxonomic list from one or more collecting sites in the same bed/sedimentological unit, or even from a time interval of variably long duration, with the underlying assumption being that they were growing in common, part of the same life assemblage. Often, visual or descriptive renderings of the inferred parent vegetation were made, generally representative of the species pool or even of the regional biome, although not explicitly stated or even recognized as such (e.g.. Davies, 1929: Becker. 1972). of sedimentological context to species lists greatly — The addition refines such studies, restricting the inferred. assem- blages to those taxa that grew in or proximate to the environment of deposition, thus making the list more representative of actual life communities (e.g., Hickey & Doyle, 1977; Scott, 1978; Wing, 1984; Spicer et al., 2002). The most representative reconstructions, both visual and statistical, are based on data, quantitative Volume 95, Number 1 DiMichele & Gastaldo 153 2008 Plant Paleoecology in Deep Time and presence-absence, from field samples. The — taphonomic investigations that decay rates not only correlation of sedimentary settings with quantitative floristic composition, at as finely resolved a taxonomic level as possible, provides the closest approximation f the original vegetation. We will focus on such approaches. It should be evident from the following that a thorough understanding of the sedimentologic, stratigraphic, and taphonomic frame- discussion work is essential prior to any attempt to make paleoecological inferences from plant-fossil assem- blages in space or time. a Taphonomic considerations. Vegetation and land- o scapes vary enormously in systematic composition, plant density, and vegetational architecture on continental spatial scales, influenced by climatic, topographic, and edaphic controls. Yet, plant remains strici subset of all 1994; Spicer, the are preserved only under a relatively geochemical 1992a, These conditions greatly sedimentological | and possible conditions (Gastaldo, 1988, 1989). likelihood. of reduce fossilization for many types of plant assemblages, depending on landscape position and climate. Additionally, there are a limited number of sites within any landscape where the potential exists to preserve plant parts because this can occur only in depositional regimes where: (1) dysoxia and/or anoxia prevails (i.e., at the sediment-water interface in a (2) micro-environmental geochemical ake system); gradients are strong (e.g., fluvial channel-bar troughs, Gastaldo et al., 1995); (3) resistant and diagnostic phytoclasts persist. unaltered long after removed all soft tissues (e.g., PEE Strómberg, 2002: palynomorphs, 1994); or (4) buried assemblages are maintained below the d water decay has Traverse, table preventing oxidation and decay over the long term (10%—10%-year time scale; Gastaldo & Demko, 2005). In the majority of instances, biomass is fated to be recycled within the carbon cycle by the living biota—this is the rule rather than the exception. The disarticulation of vegetative and reproductive with their tendency to degrade on the order of weeks (e.g., structures, along quickly, flowers and leaves), months (leaves), or years to several decades (wood, fruits, and seeds; Burnham, 1993b), presents some advantages and disadvantages when attempting to reconstruct anything, from a single plant to the original plant community or a landscape mosaic. It is especially rare for dispersal structures (e.g., seeds or spores) to be found attached to higher vascular plants. But, when found attached, i provides direct evidence for the operation of traumat- such structures are ic, as opposed to physiological processes, in the origin of the assemblage (Gastaldo, 1992a, 1994). It is well documented from neoecological and experimental differ within taxa of a single clade or between various clades, but also under different climatic conditions (Gastaldo & Staub, 1999) and even within microhab- itats under the same general climatic regime (Bray & Gorham, 1964). Additionally, the degree of similarity between a standing forest and the litter it produces may vary greatly and is not necessarily taxonomically jdn (including forest-floor litters; Burnham, , 1997; 1992), but depends on vise and growth strategy (Gastaldo, 19922). Yet, because decay rates of most plant parts are Burnham et al., extremely rapid when compared with hard parts of invertebrates or vertebrates, it is possible to identify many of the constraining mechanisms and processes responsible for preservation of an assemblage. In addition, means have been developed to estimate the taxonomic diversity of the source vegetation (e.g., Burnham, 1993a; Gastaldo et al., 2004a). Hence, nearly all plant assemblages can provide what amounts to a To estimate of the taxonomic composition of the source community and some degree of relative abundance of the source plants (this is more the case when leaves and other soft-tissue structures are preserved). It is true that certain resistant plant parts may be reworked, such as woody debris, charcoal, and palynomorphs, within particular depositional systems or particular intervals within stratigraphic sequences. Criteria have been developed, however, that allow for recognition of such recycled parts (e.g., wood-clast rounding, Gastaldo, 1994; change in palynomorph fluorescence, Traverse, 1988), the presence of which can thus be eliminated from consideration. when evaluating an accumulation of plant parts from an identified transported assemblage. There is an overrepresentation of wetland commu- nities in the plant fossil record because these are in depositional regimes with the Thus, pre-Neogene, closest proximity to such wetlands, particularly in will account for the most fossils in autochthonous and highest probability of preservation. the parautochthonous assemblages. In situ, or autochtho- nous, assemblages are deposits of high importance and significance from which we can gain our greatest insights into community composition and spatial These assemblages consist of plant parts buried within the organization within. the landscape mosaic. site where the source plants grew (Behrensmeyer & Hook, 1992). In transportation of plant parts into or out of the area although there may be palynological — general, there is little or no of burial, contribution from outside the immediate area, espe- cially if the source plants are wind pollinated. Hence the sediment-air interface, litters buried below particularly below the mean water table, represent 154 Annals of the Missouri Botanical Garden those specimens capable of being preserved in the fossil record. Not all taxa contribute biomass equally to a litter assemblage. The contribution of any taxon = may be a function of physiology, climate stress, or growth strategy, and the fidelity of the community within microhabitats of a landscape also may vary. For Burnham (1989: table IV, p. 16) founc within a paratropical floodplain that the percentage of [om example, 5 local species in the leaf litters, those within 100 m o the sample site, ranged from 45%-—63%. The time at which litter is buried can further limit the fidelity of an assemblage, considering that different taxa con- tribute different proportions of their vegetative and/or reproductive structures during any growth year (Bray & Gorham, 1964), and that those proportions may vary seasonally and annually. In particular depositional settings, such as within intermontane Tertiary sequences, care must be taken to differentiate the components actually preserved in situ from those that may have been transported downslope via various mechanisms, e.g., mudflows or lahars (Fritz, 1980a, b, 1981). Criteria have been developed to distinguish in situ from transported erect, root-laden trunks (Fritz, 1980b; Fritz & Harrison, 1985; Gastaldo, 1984, 1999). In coastal plains, such elements will be found in fluvial sediments and will have diagnostic features including root-encapsulated soil that may bear extrabasinal (e.g Liu & Gastaldo, 1992a; Gastaldo & Degges, 2007). Autochthonous burial can be instan- clasts taneous when entombment is the result of volcani- clastic ashfall (e.g., Burnham & Spicer, 1986; Wing et al., 1993), or over a period of days to months when increased sedimentation rate accompanies a change in local or regional base level (e.g., Ellis et al., 2003), which, in some instances, may be the result of coseismic subsidence (e.g., Gastaldo et al., 20044). Parautochthonous assemblages (Behrensmeyer et al., consist of disarticulated parts that have been transported some distance from the parental vegetation but not outside of the landscape setting in which they grew. Such assemblages most often are found within floodplain lakes (Rich, 1989; Gastaldo et al., 1989, 1998; Howe & Cantrill, 2001), but also in fluvial channel bedload and bar deposits depending on the fluvial dynamics and aerial extent of the vegetational unit across the landscape. In general, plant parts preserved in fluvio-lacustrine settings are derived from vegetation. directly adjacent to and fringing the body of water. But in such environments, the potential exists for the assemblage to include allochthonous elements that have been transported greater distances, depending on the geographical and altitudinal position of the lake (e.g., Spicer & Wolfe, 1987) and the proximity of tributaries debouching into 1987). it is imperative that an integrated geologic, principal river channels (e.g., Gastaldo et al., Hence taphonomic, and paleontologic approach be used to understand any assemblage prior to interpretation. Allochthonous assemblages warrant caution when attempting to address certain paleoecological ques- tions, although they may provide the only insight into the plants living in extrabasinal (Pfefferkorn, 1980) settings. In an intermontane lake deposit of Tertiary age (Taggart & Cross, 0), it proved difficult to separate the plant parts originating in the fringing vegetation surrounding the lake from those transport- ed into the lake from outside the area, particularly because all plant parts were transported as suspen- sion-load debris. Use of phytoclasts as taphonomic features (e.g., bedload vs. suspension-load transported n Gilbert-type delta—front deposits modification) provided a means by which such local versus more regional contribution could be assessed (Spicer & Wolf, 1987). samples from other proximal and distal lake sites from Using such data in conjunction. with the same geographic region, of the same approximate age, and formed under similar climatic conditions can permit assessment of the proportional contribution from different communities within the landscape to any given deposit. In assemblages preserved in offshore marine lowstand deposits (lowstand systems tract in the terminology of sequence stratigraphy), — shore-fringing and more distal vegetation will be mixed difficult to interpretation of the flora to regional spatial scales e.g., Mapes, 1985; Rothwell et al., 1996). Often, the regional components of such floras are unknown in and differentiate, thus limiting po other parts of the stratigraphic sequence. The sedimentary context of a fossil plant assemblage is a Sedimentary sequences, climate, and eustacy. crucial part of the equation that must be accounted for when using the record for paleoecological analyses. Often, depositional these data are restricted to reports of the environment (facies) and/or regional setting (correlative facies associations). But, without hu, an analysis of the sedimentary sequence al a finer scale, it is possible to overlook pertinent information relevant to understanding the genesis of a plant- rock (Galloway & or the interpretation of the fossil bearing deposit or Hobday, 1996) assemblage itself (Falcon-Lang € Bashforth, 2004; 2005; Falcon-Lang, 2006). Isotaphonomic assemblages (Behrensmeyer et al., sequence Gastaldo et al., 2000) must be distinguished and used when the — paleoecological questions concern changes in taxo- nomic dominance or richness patterns in space or facies the time. The identification of the sedimentary within a particular depositional regime, and Volume 95, Number 1 2008 DiMichele & Gastaldo 155 Plant Paleoecology in Deep Time taphonomic processes associated with that setting, provide the basis for comparison of communities. It is not sufficient to describe only the stratigraphic position or sediment matrix of the assemblage, because each of these criteria For demonstrated that not all Carboniferous roof-shale landscape. xample, oras preserved in coastal lowlands originate in the same way. Under specific depositional settings, these floras may be allochthonous, parautochthonous, or autochthonous, with the latter either preserving the ultimate peat-mire community (e.g., Gastaldo et al., 2004b; DiMichele et al., 2007) or a community that grew in a mineral substrate, unrelated to peat Gastaldo, 1987; Falcon-Lang, 2006). The disparity in absolute time represented within such formation (e.g., assemblages (e.g., coal-ball permineralized peats vs. buried-in-place clastic swamp vegetation) may be on the order of 10—100,000 years, depending on the combination of depositional and analytical time averaging. The same disparity holds true for floras preserved within the seatearth (paleosol) beneath a coal bed (Staub & Gastaldo, 2003). The relative completeness of depositional sequenc- es differs dramatically from continental to marine Continental the incomplete, highly dependent on tectonic regime, basinal dynamics, and climate (Shanley & McCabe, 1994), with resolution varying within and between depositional environments (e.g., Pelletier & Turcotte, 1997; 2003). Sediment transport from a source area is controlled by variations in long- basins. stratigraphies | are most Lowenstein et al., term climate patterns, particularly the timing and amount of yearly rainfall (Cecil & DuLong, 200 Similarly, rainfall patterns the distribution of plant communities. Thus, there must strongly control be a coincidence of plant preservation with periods of sediment transport, adequate physical accommodation space, and geochemical conditions that promote plant-part preservation rather than decay. Although it may seem counterintuitive at first, little sediment is generated and transported in most envi- ronments wherever wet, humid to perhumid (rainfall 10-12 months/year; 2003) climates prevail and precipitation exceeds evaporation and evapo-transpiration. these the landscape is covered by a multi-tiered vegetation. Cecil et al., Under conditions, Extensive rooting at various levels within the soil profile binds sediments and prevents large-scale soil erosion and sediment transport down gradient to depositional environments. Pedogenesis (especially formation of podozols to laterites) is the principal process operating across the landscape under such climatic conditions (Parrish et al., 1993; Sellwood & Price, 1993). If soils contain sufficient proportions of expandable (swelling) clays, stagnant paludal condi- tions may occur with resultant peat accumulation (Gastaldo & Staub, 1995). communities that grew under these climatic condi- ence, records of plant tions often include peat mires, with a lower probability for preservation of macrofloral remains in other parts of the landscape due to low sediment yield. At the other climate extreme, available sediment may be higher where evaporation greatly exceeds precipita- tion, but rainfall may be extremely limited and temporally concentrated (rainfall less than four months of the year in warm climates; Cecil & Dulong, 2003). sediment transport is restricted to times of high- Under such climatic conditions, however, intensity rainfall with high erosive power. Even with little vegetative cover to bind sediments, sediment transport is limited by the lack of water to do the work, which diminishes potential for organic matter burial and preservation (see the effect of water table below). Under such arid and semi-arid conditions, taxonomic diversity and plant density per unit area are relatively low. Plant settings, such as alkaline lakes, which may preserve material y be preserved in unique aerial debris from small deciduous taxa that sur- rounded the site (e.g., Beraldi-Campesi et al., 2006), or gypcrete and/or calcrete soils, in which subterra- nean rooting structures may be preserved. But, such occurrences are rare and may be associated with a rapid influx of sediment (e.g., volcaniclastic ash- fall or higher-than-average rainfall resulting in fine clastic transport and deposition) and changing geo- chemical conditions. Hence, the probability of pre- serving plant material growing under extremely dry conditions is very low to insignificant, although such plant assemblages have been inferred from parts of the stratigraphic record (Retallack & Dilcher, 1988) and seasonally dry to arid communities have been reconstructed (e.g., Francis, 1984; Cantrill et al., 1995; Falcon-Lang & Scott, 2000; Falcon-Lang & Bashforth, 2004). It would appear, then, that plant preservation in clastic rocks is promoted by equable climatic conditions, where precipitation equals or just exceeds evaporation and rainfall patterns are more seasonal in distribution (4-7 months/year), which would result in increased sediment load in streams and catchments (Cecil et al., 2003). The potential preservation, though, still requires accommodation space for sediment burial and geochemical conditions for plant-part that retard or prevent decomposition. Hence, it is more likely that a plant-fossil record will be generated tidal channel, and near- within lacustrine, fluvial. shore environments than in aerially exposed sites where pedogenesis and biomass recycling control the landscape (Gastaldo & Huc, 1992; Gastaldo, 2004). 156 Annals of the ned. Botanical Garden In the broadest generalization, two sedimentologi- cal processes control the stratigraphic record in continental and marine environments, aggradational and degradational (erosional) processes, both of which climate (Bull, 1991). As the aggradational sequences build up are controlled by L concepts imply, sediments in parts of a sedimentary basin, whereas degradational sequences remove sediments and re- distribute them elsewhere within that or an adjacent basin. There is an interplay between both aggradation and degradation at various scales (within beds, — architectural elements, depositional systems, etc.; Miall, 1990). Additionally, the record of each is influenced by autocyclic (e.g.. channel migration) and allocyclic (e.g.. tectonism, eustacy, climate) processes o at any point in time and space (Beerbower, 1961, 1964: Demko & Gastaldo, 1996). The highest probability for preservation of plant fossils exists within regional-scale al o oo sequences immediately above the contact with the previous regional-scale erosional event (Gastaldo, ¡om 2006). This point of transition between erosion anc ageradation reflects a change in base level and the generation of increased accommodation space. Ag- eradational processes build up continental sequences via the transport and deposition of siliciclastics from an erosional source area. Sediment accumulation may intermontane basins. within the occur in confined, source areas or further away, near local or regional base level. Pedogenesis may modify overbank sedi- ments spread between the interfluves (Retallack, 2001), assisted by t and soil-forming processes generally are — 1e presence of some type of vegetational cover, the composition of which is controlled by the 2003) and the invertebrate fauna that utilizes this resource (Ray- 2001). table will control the geochemical pore-water proper- prevailing climate (Ziegler et al., — mond et al., The level of the regional water ties at all depths, concomitantly impacting decay of buried organic matter. Preservation of buried plant detritus is most likely when the regional water table is persistently high over millennial or longer time scales, rather than when the water table fluctuates dramat- ically, permitting percolation of oxygenated rainwaters through soil and buried. sediments (Gastaldo & Demko, 2005). Base-level changes in transitional (coastal plain, deltaic, estuarine, and tidal flat) and nearshore marine (c tectonic or eustatic in nature, jon elta front, barrier island) settings may either be ora combination of both of these allocyclic mechanisms operating over differ- ent time scales. Debate continues as to how to tease apart tectonic from eustatic effects within the stratigraphic record as the primary mechanism channel-fill complexes, major controlling accommodation space for sediment burial (e.g.. Klein, 1992; Christie-Blick & Driscoll, 1995; Miller el al; 1997). But. aat the long-term record of aggradational and degradationa the fact remains t — processes in coastal and marine depositional regimes may be controlled by eustacy (for this discussion, the short-term tectonic influences on generating accom- modation and plant-part preservation will not be see Gastaldo et al., 2004a). At certain scales of inquiry, paleoecology must addressed: consider the regional sedimentological framework within which ancient landscapes are preserved. The paradigm of sequence stratigraphy (Posamentier & Allen, 1999) provides a model that permits the effects — of eustacy, climate, and tectonics to be integrated into a larger-scale framework. Eustatic cycles are based on the marine stratigraphic record and are asymmetrical over time, with rapid transgressive (onlap) sequences Miall, 1990). Sequences are bounded above and below by — — and prolonged regressive (offlap) sequences unconformities, which often are exposed subaerial en surfaces (identified by paleosols in terrestrial land- scapes). Each onlap or offlap sequence is composed of parasequences, which are defined (van Wagoner el al., 199 genetically related beds bounded by marine flooding 0) as a relatively conformable succession of surfaces and their correlative updip (inland) surfaces as paleosols). Note that although the marine record may consist of a number of stacked parasequences, the continental record of these sea level changes may be locked within a single (thick) paleosol. Parasequence sedimentation occurs when accommodation 1s gener- ated by a rapid rise or fall in sea level. Each change results in an asymmetrical, shallowing upward cycle somewhere within the basin but not everywhere coincidentally. Parasequences are stacked in various Transgressive patterns. across the basin over time. — systems tracts (TSTs) occur where successive para- sequences overstep each other in a landward direc- into tion, and coastlines move over the continent, more interior areas. When sea level has reached its the shoreline reverses its direction f transitional and nearshore sediments in a series of maximum depth, and progrades seaward, building up deposits « c parasequences within a highstand systems tract (HST). As sea level begins to fall, coastal progradation may continue until a time when the shoreline drops below the position of the previous shelf-slope break or one where the fall in relative sea level does change the position of the shoreline in the basin. The first situation results in the fluvial incisement (down- —, cutting) and erosion of former HST deposits, and transport and deposition of these mobilized sediments as lowstand systems tract. (LST) turbidites in deep marine settings. Volume 95, Number 1 2008 DiMichele & Gastaldo Plant Paleoecology in Deep Time 157 For the general purposes of this discussion on transitional environments, preservation of autochtho- nous and parautochthonous fossil plant assemblages occurs primarily when sea level is high, that is, within ageradational HST parasequences. It is in these settings. where high sedimentation rates and high regional water tables prevail (characterized by inceptisol formation and wetland vegetation). When sea level drops and LST deposition takes place, degradational processes operate within the then- exposed HST deposits. All buried assemblages are subject to effects brought on by a lowering of regional water table coincident with the newly established, lower (oceanic) base level. In summary, climate not only controls the distribu- tion of vegetation across the landscape but also the timing and nature of sedimentation, within which plant parts are buried and potentially preserved. For these reasons, the plant fossil record in more proximal and, therefore, somewhat higher-elevation siliciclas- tic continental environments is biased toward low diversity, seasonal to humid, wetland communities in which geochemical conditions exist that may re- tard decay of plant debris. When these settings are subjected to sudden base-level change (e.g., tectonic coseismic subsidence), new accommodation space is generated and filled rapidly. This sedimentary in- filling (aggradation) is accompanied by a rise in regional water table, removing some buried organic matter from the effects of oxygenated groundwater. If climate remains relatively humid during landscape stasis, groundwater level will remain high and buried organic matter can be preserved. But, if climate trends toward more seasonal and/or semi-arid to arid during landscape stasis, water tables will drop and decay of buried organics will occur. Entombed organics may be f or may be lost entirely - from buried sedimentary sequences sediments are not somewhat compacted, reduced to impressions where some compaction has taken place in fine-grained sediments. Early diagen- esis may promote mineralization of buried detritus, including pyritization, limonitization, and perminer- alization with calcite or silica. Plant assemblages in coastal plain and transitional settings are also biased toward wetland communities The reasons are similar to those provided above. Coastal growing under seasonal to humid climates. and transitional communities already are at or within ca. 2 m of mean sea (base) level and grow within a progradational regime where sediments, in general, accumulate. Even in these basinal settings, preser- vation potential is high only when accommodation space for sediments is generated either by localized tectonic or compactional subsidence and concomi- tant sea-level rise, or eustatic rise in association with transgression without ravinement (Liu € Gastaldo, 1992b). The latter situation commonly occurs within incised valleys where gradual sediment aggradation in a more proximal area, closer to the sediment source, allows for a thick sedimentary package in which plant 1998). Preserva- tional potential is highest in subsequent parase- fossils are preserved (Demko et al., quences during the phase of rapid aggradation, again, if the regional water table ascends as sediment accu- mulates and is maintained over the long term (fourth- to-third order cycle durations of 10'-10? years). The fact that there is a bias toward an overrepresenta- tion of wetland communities through geologic time, preserved within predictable settings by identifiable mechanisms, allows for their direct comparison under broadly isotaphonomic conditions. Al- though now readily recognized and widely studied, it Paleosols, biochemical, and geochemical proxies. is hard to believe that until 25 years ago, paleosols were barely recognized in pre-Quaternary rocks. Paleopedology focuses on the soils that formed in past landscapes. Paleosol classification schemes may be based on modern soil-forming processes (Mack, 1997; Retallack, 2001b) or pedogenic processes that 1993; ). The mosaic of soil types that result in identifiable characters (Mack et al., James et al., 1998 exists across any landscape reflects the variation in physical and chemical features resulting from variations in climate and vegetation (Kraus, 1999). Soils may form in as little as a few hundred years (wetland entisol/inceptisol) or over several millennia (oxisol/spodosol), and well-developed soils usually define major unconformities in terrestrial stratigraphic sequences 1994; Shanley « McCabe, Cecil et al., e.g., Driese et al., 1994; Joeckel, 1995; Olszewski, 1996; 2003; Demko et al., 2004). Evidence of colonization by plants and animals may include the presence of root traces or drab halos, where organic matter has been degraded and color differentiation enhanced by dehydration of ferric oxyhydrates during burial (Pfefferkorn & Fuchs, 1991; Retallack, 2001b), or ichnology (e.g., Hasiotis, 2003). Most often, though, organic matter is recycled through the soil system, leaving little or no physical evidence of the specific that sediment. Dispersed organies may remain within the kinds of plants biologically modified the soil profile, such as phytoliths (Strómberg, 2004), carpolit 1983), identifiable — is (Thomasson, and meso- or micro- be whereas others may be amorphous and unidenti- debris; some may systematically, fiable. Regardless, many plant parts retain original . 2004b, 2006) that can be used for paleoecological restoration. In biochemical information (Jahren et al. addition, such debris may act as the nucleating agent 158 Annals of the Missouri Botanical Garden 1990), conditions for soil nodules or concretions (e.g., Pye et al., if bacterial activity and geochemical permit. By their assemblages allow correlation between paleosols and nature, rooted-in-place, autochthonous a community under which they developed. . because in situ assemblages are a rarity, the source plants and vegetation under which the majority of paleosols formed remain unknown. In addition, long-exposed terrestrial surfaces can be subject to changing climatic and biotic agents resulting ir pan] compound soils affected by a number of potentially very different kinds of physical and biotic conditions. The question further arises as to possible relationships between specific soil types and specific kinds vegetation (arborescent vs. herbaceous; gymnosper- mous vs. spore-producing), and whether diversifica- Phanerozoic has Such questions are focused more on those soils that form tion of vegetation types over the resulted in the diversification of soil types. under seasonally dry conditions, such as vertisols, alfisols, and spodosols (Cecil & DuLong, 2003; Driese et al., 2005). In contrast, very wet and very dry soils, at the ends of the climatic spectrum, present plants with a unique set of physical and physico-chemical conditions controlled by persistently high water tables (e.g., histosols, laterites), or high evapo-transpiration Um, rates and very low rainfall (e.g., calcisols, gypsisols that affect vegetational structure independently of its phylogenetic composition. Note the great similarities among Late Mississippian, Pennsylvanian and Perm- ian, Late Cretaceous, and Neogene peat-accumulating histosols, despite the enormous differences in the phylogenetic composition of the vegetation under which they formed (it is not possible to compare the other end of the spectrum because of taphonomic constraints). There is evidence of change in soil structure with the advent of arborescent (tree and shrub) growth forms, accompanied by the increasing depth of root penetration (Driese et al., 2000). During the Early to Middle Devonian (Elick et al., 1998), soil thicknesses and types were controlled more by illuviation and (Driese 1992; Driese et al, 1992) than by organic degradation and sol translocation & Foreman, — have been 1992), pre- vegelalion respiration. Pre-vascular plant-soil types Retallack, bryophytic described (e.g., polsterland; sumably formed under a Strother, 2000). Many of th in which autochthonous assemblages are preserved e Middle Devonian soils — are poorly developed wetland paleosols (e.g., Allen € Gastaldo, 2006; Greb et al., 2006), although Retallack (1997) infers better-drained and -forested alfisols from Antarctica during this time (but see Dahms et al., 1998). A wider range of soil types appears within ^ © e continental sequences during the expansion of gallery forests in the Late Devonian (Algeo & Scheckler, 1998). The subterranean plant parts preserved in if any (Pfefferkorn & Fuchs, 1991), are generally root traces, clay-filled (Retallack, 1999), rhizoliths (e.g.. Schneider & 1996; Kraus & Hasiotis, 2006), or rhizotu- (Demko et al. 2004) Typically, the community structure responsible for these soils, such as voids Rossler, bules and rhizocretions origin of a paleosol is inferred by analogy with modern biomes (Retallack, 1992, 2001 b). Such more consistent with angiosperm-dominated land- Retallack et al., 2001) than with ecosystems dominated by extinct. gymno- | an approach is scapes of the Tertiary (e.g.. sperms and lower vascular plants. matter (OM) retains the carbon isotope (°C) signature of the contributing Paleosol bulk-organic plant matter (Jobbágy & Jackson, 2000). Consequent- ly, isotopes have been used as a proxy to recognize unique metabolic pathways (C3 or C4 grasslands; e.g., Fox & Koch, 2003, 2004; Pagani et al., 1999) and to characterize proportions ir their relative =) parent vegetation. Isotopes also have been used to assess 2002). Soil OM is derived from the decay of biomass produced under C3, paleoclimate (e.g., Boom et al., C4, or Crassulacean acid metabolism (CAM) photo- synthetic pathways. The C4 and C4 isotopic compo- sition of carbon varies from —25%o0 to —3296o and —14%o te evidence to — 1096o, respectively. Presently, there is no that Ca Miocene, although inferences have been made about CAM (Raven & Spicer, 1996). persists upward through the food chain, with incorpo- pathways existed prior to the This fractionation ration of the isotopic signal into tooth enamel (e.g., Fox & Koch, 2003, 2004). isotopic signal in soils may be altered by decay However, the carbon processes (e.g., Bowen & Beerling, 2004), leading to the use of taxon-specific biomarkers (e.g., Lockheart et al., 1997) that, in the absence of recoverable fossil- plant debris, may make it possible to differentiate the Smith & Whit Stable at oxygen isotopes recovered from contribution of ae taxa (Nguyen Tu et al., 2004; 04). calcite cements within soil carbonates (concretions, nodules, amorphous masses) also provide paleobaro- metric data for the time at which calcite precipitated (Cerling, 1984, 1991). been used as an indirect method to estimate the paleo- Pedogenic carbonates have atmospheric CO» concentrations and climate, despite complexities for which analytic compensation must be made (Ehleringer € Cerling, 2002; Tabor et al., 2007), and several models have been developed (Cerling, 1991; Yapp & Poths, 1996). In the cases where pedogenic carbonates reflect the incorporation of atmospheric CO», it is possible to Volume 95, Number 1 2008 DiMichele & Gastaldo 159 Plant Paleoecology in Deep Time reconstruct paleoclimates and, in turn, paleoprecipi- tation regimes (e.g., Tabor & Montañez, 2002). There are numerous Tertiary-age examples based on multi- ple proxies (pedogenic carbonates and tooth enamel; 1995, 2003), with fewer further back example, e.g., Koch et al., in deeper time. Triassic climate, for — generally has been interpreted as warm to semi-arid under monsoonal seasonality (e.g., Rees et al., 2002). Yet, Triassic macrofloral assemblages are interpreted to indicate the prevalence of warm, humid climates (Ash, 1999), probably reflecting the deposition. of plant-bearing beds during wetter intervals within incised valley systems (Demko et al., 1998). Recently, Prochnow et al. (2006) acquired isotopic data from Triassic pedogenic carbonates. They estimated atmo- spheric CO» concentrations from the micritic fraction using the equations of Cerling (1999), perature using both oxygen and carbon isotopic They inferred a trend in the Early Triassic and tem- values. of the southwestern U.S.A. from arid to semi-arid (aridisols) conditions in the upper Moenkopi Forma- tion to a more humid climate (alfisols) in the lower Chinle Formation, followed by a drying trend in the upper Chinle (semiarid aridisols and inceptisols). I this same lithologic sequence, there is an ol paucity of fossil plants, those that are concentrated within a pluvial interval, which is and no surprise, given the taphonomic constraints on plant preservation (see earlier discussion). Hence, in those parts of the stratigraphic record where aridity and/or high seasonality is inferred from isotopic evidence, the absence of well-preserved megafloral assemblages is not unexpected, given the geochemical conditions. The most comprehensive spatio-temporal insight into a fossilized terrestrial plant community comes from the integration of paleosol and (bio)geochemical data from rocks bearing autochthonous mega- and microfloral plant fossils (e.g., Gastaldo et al., 1998). It is most common, however, to encounter parautochtho- nous assemblages, which occur in association with, but not as part of, paleosols. Such assemblages may have little or nothing in temporal context with the paleosol(s) immediately beneath or above them within the section (Gastaldo et al., 1995b) advised particularly where there are incongruencies Caution is in paleoclimate indicators in successive beds, such as when megafloral assemblages occur in sequences with, Such paleosols form at that end of — put not actually within, calcisols or aridisols. ` the climate spectrum where preservation of labile plant tissues is unlikely. Hence, only rarely can the presence of a well-preserved plant-fossil assemblage in such a rock sequence be used to interpret a dry or arid community. Spatial patterns. Spatial relationships of taxa within the landscape are dictated by a variety of abiotic (soil) and topography, insolation, local water availability, and factors including edaphic conditions climate. Most landscapes are vegetational mosaics with taxa capable of growing anywhere conditions do not exceed the limits of their physiological tolerances Beleya & Lancaster, 1999; Jackson, 2000). It is recognized that incumbency, competition, the role of light breaks, and other factors operate over the short term (10'—10? years) to dictate the composition of any local community. But, in theory (e.g., Hubbell, 2001), all taxa within a species pool oa over the longer term, could be considered to be roughly equivalent and capable of growing anywhere within the region of — the species pool or metacommunity (Law & Leibold, 2005), given the proper abiotic conditions. Hence, the composition of a plant-fossil deposit will depend on whether plant parts were shed physiologically (resulting in a lower fidelity record) or traumatically (resulting in a higher fidelity record), which plant parts were available at the time the deposit was formed, the chemistry of the depositional environ- ment at the time, and the process-geometry of the sedimentary environment. As a result of these factors, it is unlikely that most taxa in the immediate source vegetation will be represented within any given assemblage of fossil plants. Add to this the depositional environments, which are often orders of fact spatial scales of magnitude greater than what is exposed at most outcrops. It thus becomes clear that collections from a single hole-in-the-ground locality, less than 1 m? to several square meters, capture only a fraction of the forms that existed. On the other hand, one must consider the regional metacommunity, broadly speaking, the regional spe- 2005), that will contain many more species than any local site. This allows a cies pool (Holyoak et al., solution to the limitations of small local samples, which is to make many small collections from across a sedimentary environment (e.g., Johnson & Ellis, 2002; Gastaldo et al., 2004b; DiMichele et al., 2007) o coeval stratigraphic horizon (e.g., Davies-Vollum à Wing, 1998). Statistically, such an approach will maximize the information (Bennington, 2003), particularly where an autochthonous assem- available blage is sampled. This approach can be confounded by difficulties in tracing short-term time horizons over large spatial areas—the more finely one resolves time at a single outcrop exposure, the more difficult it can be to trace that time line laterally, without a well- defined and sedimentologically understood marker (such as an ashfall tuff). In addition, different times in geological history appear to have had different 160 Annals of the Missouri Botanical Garden landscape diversities, i.e., different metacommunity/ species pool richnesses, even while retaining similar diversities on small, subhectare scales (Wing & DiMichele, 1995). This may involve a systematic time that Paleozoic less diverse than those of the Mesozoic bias, such species pools may be significantly and Cenozoic, leading to more uniformity among the best sampled sites along a gradient. Sampling strategies. It is apparent that many inter- acting factors affect the genesis and maintenance of a fossil plant assemblage. In order to ensure maxi- mum data recovery for paleoecological analyses, a appropriate sampling strategies must be employed. the assemblage, aerial extent of the exposure, both the dependent on the taphonomic signature of lateral face and horizontal and degree of matrix lithification. Different techniques. for example, will need to be employed for autoch- thonous well-lithified where standing vegetation and forest-floor litter are bedding surfaces, assemblages in sediments, preserved lithified mudstone (e.g., Taylor et al., 1992) versus those preserved in a non-lithified sedi- ment in a semi-mummified state (e.g., Mosbrugger et al, 1994; Williams et al.. 2003a, b). Similarly. lithified sediments preclude sieving and, thus, reduce the possibility of recovering small fruit and seed remains from parautochthonous assemblages, remains that would provide a more complete representation of the community than leaf or wood analyses alone (e.g., Gastaldo et al., 1998). What's the question? Most macrofossil plant assemblages targeted for paleoecological analyses ‘an be characterized broadly as parautochthonous. lt is in these kinds of assemblages that the widest range of sampling strategies has been used. It is intuitive that the methods of sampling and analysis should fit the nature of the question (Magurran, 1998). One of the most significant questions regards the matter of biodiversity. One can ask legitimately what generates reveal about ecosystem attributes and functions, and to what biodiversity, what does taxonomic richness extent do rare species matter in comparative ecology? These have become subjects of considerable discus- 2001: For example, Buzas and Culver sion in neoecology (e.g.. Loreau et al., Ives & 2007). (1999) demonstrated, using highly diverse assemblag- Carpenter, es of modern benthic foraminifera, the for example) in species richness can be accounted for that interregional differences (between arctic and tropics, mainly by the rare taxa that lie in the tails of species abundance curves, largely those known only once the fossil within huge numbers of samples (81% of left a record. Conversely, the top 10 dominant species in an difference), very few of which have area (species pool) were found to have approximately similar distributions and relative abundances in each biogeographic region. Thus, if one is interested ecological structure, it is likely that the dominant species in a local area, such as a hectare, will emerge quickly from small samples and most certainly if multiple samples are taken (e.g.. Burnham, 1993b; Wing & DiMichele, 1995). If comparisons are to be made across allochtho- nous assemblages, presence/absence data collection is most appropriate. If the objective 1s to determine richness (€ diversity) of a local assemblage. for example, the focus should be on large numbers of specimens from within a given depositional environ- ment, in order to of taxa; it is maximize recove rv not necessary to laboriously collect data on debris cover (e.g., biomass estimate based on digitized area: 2000). concerns the proportionality of taxa with the intent of the « Pryor € Gastaldo, Similarly, if the question reconstruc ting community structure, whic h is is juestion of most paleoecological studies, i generally appropriate lo use a quantitative sampl- ing method, the choice of [shi will depend on the taphonomy of the plant deposit, the nature o — the plant fossils, and the nature of the geologica exposure. In such studies, the objective most often is to understand vegetational structure, so that exhaus- tive sampling to detect a high proportion of the rare species is likely not necessary nor would the results ce justify the sampling effort required. different aches to quantitative sampling have been employed Quantitative sampling. | Several appro- in paleoecological studies of fossil plant. deposits. (1) In studies served in fluvial and lacustrine settings, it is common practice to count leaves individually (Wing. 1984: Johnson, 1992; Burnham et al., 1992: Burnham, 1994: Davies-Vollum € Wing, of angiosperm-dominated floras pre- 1997: 1998a), permitting direct. conver- Gemmil € Johnson. 1998; Wilf et al., sion of leaf counts to biomass estimates, if — desired (assuming size and leaf thickness to be approximately fe habit). (2) In of Pennsylvanian- the same across taxa of a given lif permineralized peats (coal balls) age coal beds, square-centimeter (Phillips et al., 1977) or square-millimeter (Pryor, 1996) grids have been used to estimate taxonomic and organ num- bers and biomass when overlaid on cut When bedding surfaces (3) point count (Scott, 19772, quadrat. methods (Gastaldo et al., surfaces. extensive are accessible, 1078. 1985) (4) 2004b: DiMichele (5) For spoils or museum collections from a single deposi- et al., 2007) have been employed. mine tional setting, rapid estimates of relative abundance can be made by the use of hand samples. treated (Pfefferkorn et al. DiMichele as quadrats 1975; Volume 95, Number 1 2008 DiMichele 4 Gastaldo Plant Paleoecology in Deep Time 161 et al., 1991). These methods produce similar but not always identical results when applied to the same data sets. Wing and DiMichele (1995), for example, compared quadrat (frequency) methods and percent- age representation (count) methods and found that the quadrat method systematically increases the abun- dance of rare taxa while reducing the abundance of common taxa. Proportionality of taxa within an autochthonous or parautochthonous assemblage also can be assessed using subjective and/or semi-quantitative measures. Simple rank-order abundance (ordinal scale measure) Is an attempt to assign a subjective representation to each taxon based on occurrence. Abundance classes most often are assigned on the basis of a visual estimate of the sample and ordered into abundant, common, and rare categories (Spicer, 1988). Kershaw (1973) noted that this scheme often underestimates small (fragmentary) taxa, and differentiation between — can vary among individual observers. Adjustments of the method that impose a semi- quantitative approach include the modified scale of Braun-Blanquet (1932; Podani, 2006) and the Domin scale, which are based on an estimated percentage of cover for each taxon. These methods appear to be more robust and can be accomplished more rapidly when used in compression-impression macrofossil paleoecological analyses than estimates of cover using 1988) Autochthonous assemblages, those where stems and point counts (Spicer, derived litter are preserved in situ, permit the most complete data collection. In such instances, it may be possible to evaluate the stature of individual plants, 1993; Gastaldo et al., 2006) and woody (Williams et al., 2003a), includi ing using both herbaceous (Wing et al., jon mathematical estimates based on general Niklas, 1 assemblages studies of 1994). may permit evaluation of the density of individuals plant growth allometry (e.g., — Perhaps most importantly, suc within the community, if the exposure allows, using both simple direct measurements and various means of calculating nearest-neighbor patterns (e.g., Wnuk & Pfefferkorn, 1984, 1987; Gastaldo, 1986; DiMichele & DeMaris, 1987; DiMichele & Nelson, 1989; Wing al., 1993; DiMichele et al., 1996; Cúneo et al., 2003; Williams et al., 2003a, b). Such analyses can provide insight into vegetational structure in addition to the proportional taxonomic composition, which is the currency. of most paleoecological analyses. In some cases, such as the Late Cretaceous flora found (1993), it could be shown that species diversity and relative abundance beneath an ash bed by Wing et al. were not correlated; angiosperms, although represent- — ed by the most species, did not account for the highest biomass in the assemblage, raising questions about studies that assume a rough equivalence of taxonomic richness and ecological dominance. Biomass contri- bution from individual taxa also can be calculated in autochthonous assemblages, resulting in productivity estimates that can be compared with Recent commu- nities (e.g., Williams et al., 2003b). Spicer (1988) noted at the time that megafossil-assemblage sampling had been a Spatial sampling. somewhat haphazard process. He proposed a more codified set of strategies to be applied to bedding (three- dimensional) samples. Several of these methods are a surface (two-dimensional) and volumetric effectively the same as neoecological sampling prac- tices. These techniques can be supplemented by bulk can provide additional data on the oft-overlooked 2004). The disarticulation of aerial plant parts is notorious for others, including sieving of sediment and meso- and micro-fossils (Eklund et al., causing overrepresentation of some taxa and under- representation of others, relative to their proportion- al abundance in the parent vegetation (e.g., Burnham 992; DiMichele & Phillips, 1994). This means that various measures of taxon abundance in et al., the fossil sample may be irrevocably biased, though correctable in some cases. The first step is to recognize that there rarely will be a 1:1 correlation between the quantitative composition of a parent vegetation and that of the derived fossil assem- 1992; oxbow blage, whether autochthonous (Burnham et al., Burnham, 1997) or parautochthonous (e.g., lakes; Gastaldo et al., 1989) For those quantitative studies that use quadrat- based methods, the size of the quadrat selected and the number of quadrats assessed are dependent on many factors, including logistical hurdles. The details of the sampling strategy will be dictated by the size and number of available outcrops, the degree of cementation and mineralization of the matrix, and the time and resources available to carry out the study — particularly if one is working under time constraints in a mine, or on another continent when the best localities are found at the end of a field season) meaning the choice is often less than ideal. In concept, the heterogeneity and scale of plant-part size have to be taken into account when selecting a suitable quadrat size. Spicer (1988) suggests a pilot study during which a pre-survey taxon/area curve (similar to a collection curve, see below) is developed to determine the minimum quadrat size for a specific assemblage. For example, it is most common for 0.5— 1.0-m? quadrats to be evaluated at the outcrop scale where autochthonous leaf litter is concentrated, whereas assemblages in which large prostrate trees are preserved may be resolved more quickly and effectively by larger areas. Burnham (1989) further 162 Annals of the ec Botanical Garden developed guidelines for the choice of sampling strategies (few large samples vs. many small samples) based on the heterogeneity of the source vegetation, which can be approximated by a few simple on- outcrop tests. Most recent studies of fossil. plant assemblages consider explicitly the spatial aspect of the sample. Sample replicates, represented by two or more side- by-side samples, permit intra-site sample variabi- lity to be assessed and compared with among-site DiMichele et al., 2007). AI- though such replicates are not samples from ex- sample variability (e.g.. actly the same site, they are as close as can be expected from fossil-field samples. Among-site com- parisons require collections from multiple sampl- ing sites (and replicates thereof) along a transect within a single sedimentary environment (e.g.. Wing & DiMichele, 1995; Wilf et al., 1998a; Ellis et al.. 2003; Gastaldo et al., 2004a; Willard et al., 2007). Transect leaves preserved in modern 1992; 1993b have demonstrated limits to leaf mixing in parauto- studies of — sediments (Burnham et al., Burnham, chthonous assemblages and, thus, suggest that most square-meter samples represent relatively local, at most half-hectare source areas, if the parent vege- tation was a forest. Data analysis. It is beyond the scope of this paper == to discuss in detail the methods that can be applied to the analysis of various kinds of paleoecological data (see DiMichele & Wing, 1988). Data are often overanalyzed. It is frequently possible to grasp and describe major patterns accurately simply by in- spection of the data. The techniques used to analyze patterns in modern vegetation usually can be applied equally well to fossil data. These techniques can be found in any number of general references, such as McCune and Grace (2002), ble in many different computer software packages and are widely availa- (including free shareware such as the R package, which includes manuals that offer various insights into the techniques and their peculiarities; R Project for Statistical Computing, 2007). Most for interpretable patterns in numerical data, often paleoecological analyses are searching with an explicit spatial component. Data. generally are stored in a matrix of samples by taxa. Sample similarities can be compared with a wide range of metrics, varying from those based on taxono- mic presence-absence (e.g., the Jaccard coefficient). rank-order abundance (e.g., rho), or Spearman's ratio-scale quantitative measures of sample compo- sition (e.g., Euclidean distance). These metrics are the basis m assessing sample similarities. although the choice of metric may be influenced by sample or Wilf size (Archer & Maples, 1988). variate 1987; Maples € Archer, The most commonly used exploratory multi- visualization of similarity method for the among samples is ordination, and the most com- monly used ordination methods at the present lime are nonmetric multidimensional scaling (NMDS) and detrended correspondence analysis (DCA) (e.g.. Miller et al., 2001; Holland, 2006; Willard et al., Data from other sources, such as the sedi- mentary environment, can be superimposed on these analyses. Classification methods, such as clustering, are used frequently in conjunction with exploratory Liang, 2004). regarding the existence of structure in the data, can be carried out with the use of such analysis of similarity (ANOSIM). Methodologically NMDS, ANOSIM examines variation in taxonomic compositional data within and between methods (e.g., Hypothesis testing, methods as similar to predefined samples. Discriminant function analysis is a hypothesis-testing classification method that employs a priori-defined that assign unknowns to one of the existing groups with a eroups to develop a classification. algorithm then can be used to specified degree of confidence. tecently, several workers (e.g., Wing & Harrington, 2001: fig. 9) have plotted litio first-axis sample scores vatterns of against time detect — temporal change in similarity among samples. This is a pow- erful visual technique that can quickly reveal vege- — ational responses to environmental changes in time. Spatial analyses of in situ vegetation may be reported in two (horizontal space) or three (horizontal space and vertical canopy height) dimensions | (e.g.. Wnuk & Pfefferkorn, 1987; Wing et al. 1993; Williams et al., 2003b; Gastaldo et al. 2004b). Additionally, the analysis may include various kinds ^ of nearest-neighbor statistics to determine if plant distributions are random, ordered, or clumped (e.g.. DiMichele et al., 1996; see Hayek & Buzas, 1997). A final comment on method involves sample standardization. Under many circumstances, partic- ularly when taxonomic richness is the objective of a study, it will be necessary to standardize sample size in order to reduce the biases of collecting intensity. This is generally accomplished with a method known as rarefaction (Raup, 1975). With this method, a distribution of samples of different sizes is created. by repeated resampling from the original quantitative sample. This is done for each sample in the analysis. Then a standard sample size is chosen for the entire sample suite, drawn from the re- sampling curves created for each of the individual samples (for examples of the use of this method in plant paleoecology, see Wing & DiMichele, 1995 2001). et al., Volume 95, Number 1 2 DiMichele 8 Gastaldo Plant Paleoecology in Deep Time 163 RECONSTRUCTING THE PAST Reconstruction. of the past, in either visual or descriptive form, of individual species or of entire landscapes, is one of the major ways in which paleontologists make their work accessible, both to specialists and nonspecialists. Throughout the follow- ing discussion, a significant concern is the extent to which proportions of species are represented accu- rately in assemblage reconstructions. Of even greater concern is the degree to which the biologies of the individual plants are captured and integrated into the landscape reconstruction. The implicit dynamics of these past systems are in large part implicit in the portrayal of the reconstructed plants. The main fault of restorations, especially for those that purport to represent the pre-angiosperm world, is the implicit assumption that the plants can be analogized to extant angiosperms with life histories, growth habits, repro- ductive biologies, and stand characteristics that might be found in a local temperate woodland, so familiar to Northern Hemisphere scientists. This led Pfefferkorn (1995: 389) to label the paleobotanical community, on Take, for the ever-present uprooted lepidodendrid average, as “temperate climate chauvinists.” Pennsylvanian wetland | reconstructions, suggesting blowdown and uprooting as major distur- bance elements in such assemblages. In fact, there is virtually no evidence ever reported for uprooting of these trees, or the tree-pit soil disturbances and tree gaps that accompany such events. Equally suspect are the many representations of dense, dark canopies in Carboniferous lepidodendrid forests, given that for many of the species the plants had no crowns until late in life. Such assumptions and analogies to modern vegetation are present to various degrees in recon- structions of the past from any time in geological history. Thus, it may be best to make clear statements about the assumptions regarding the plants and vegetation and, roughly, the confidence that can be placed in a particular reconstruction when offered for general consumption. Disturbance and succession. The ecological dyna- mies of past communities, like those of today, are strongly influenced by disturbance and subsequent patterns of succession. Perhaps the most commonly studied disturbance agent is fire. An excellent review of the distribution of fire through time and its rela- tionship to atmospheric oxygen levels is provided by Scott and Glasspool (2006). Wind disturbance, also common in the fossil record, is so at difficult to document. Wnuk and Plefferkorn (1987 documented a blowndown forest of Pennsylvanian age through access to large bedding surfaces exposed in a coal mine and excellent sedimentological con- trol. Flood disturbance has been documented for a variety of deposits throughout the fossil record (see citations below), the problem being the differentia- tion of flood disturbance from flood burial. Distur- bance in voleanogenic landscapes has been described by Taggart and Cross (1980), bs et al. (1982), and Cross and Taggart (1982) i posits and by Scott and Galtier T as disturbance iocene-age de- affected the evolution of early ferns, from the Mis- sissippian of Europe. Succession following disturbance is more difficult to document. Analyses of vegetational responses to fire need to be done at a finely resolved level, such as the study of Arens (1993), who analyzed a plant succession across a charcoal layer in Pennsylva- nian-age deposits by taking small samples across Pryor (1993, ii lyzed patterns of succession in a P the disturbance horizon. ana- peat deposit using statistical patterns af plant co- occurrence in coal ball macrofossils. Also studying haymond (1988) ana- lyzed succession during peat accumulation through DiMichele and Phillips (1988) correlate patterns of plant re- coal-ball permineralizations, patterns of successive root penetration. sponse to disturbances within coal beds, again us- ing macrofossils, by linking sedimentological fea- tures to recurrent assemblages. Many studies of coal palynology have revealed patterns of vegeta- tional responses both to disturbance and to direction- al changes in the physical environment of the peat body through time of accumulation, generally related to hydrology and nutrient peur (e.g., Smith, 1962; Grady & Eble, ; Willard, 1993; Eble, 1999a, b; Figueiral et al., n Family portraits: Lots of plants but not much accu- racy. Museums are populated by reconstructions of past worlds—dioramas and paintings and com- binations thereof. These reconstructions of the past now can be found in abundance and readily on Internet websites, in textbooks, and in various popu- lar venues. They generally are agglomerations of representative species from each of the main plant groups living at a particular time and in a par- ticular region (for example, the wet tropics, a broad dry steppe, or a periglacial area). Such reconstruc- tions often mix together plants from different sub- habitats, portray density unrealistically and, overall, can be terribly inaccurate as ecological depictions of the past. Add and in unlikely combinations, and these reconstruc- animals, at excessive densities tions become little more than family portraits, time- averaged gatherings of extinct organisms. 164 Annals of the Missouri Botanical Garden Family portraits have their role, in that few places in the past are sufficiently well known or, if well known, are insufficiently diverse to make for a captivating reconstruction, one that passes along the sense of strangeness of a lost world. In their diversity, family portraits offer the viewer an abstract of the past. Thus, fossil family portraiture has its place and will not be displaced by reconstructions that aim at Improvements could be made by greater accuracy. paying more attention to autecology in particular. Statistical reconstructions: Reconstructions based on the law of averages. More accurate representations of past vegetation have been based on data from actual paleoecological samples. Because most of such samples are from parautochthonous assemblages (see discussion above), the samples are representations he original flora, of subject to various kinds of taphonomic modifications. Thus, most paleoeco- logical reconstructions based on data are based on some form of likelihood estimate, when the data are considered within the context of the various aspects of taphonomy. Such reconstructions may lack the repre- sentational breadth of family portraits, depending on the amount of the landscape they represent. Statistical reconstructions often are based on the output of numerical analyses, particularly ordinations juantitative samples. Such or cluster analyses of c samples provide a sense of recurrent taxonomic associations and may reveal gradients, outliers, and the patterns of relative commonness or particular combinations of taxa. In some instances, no actual image is produced from such analyses and the Broadly statistical reconstructions may also be derived from reconstruction is descriptive. considered, detailed sedimentological analyses in which plants are related to depositional environments and thence back to distribution on a larger landscape. Examples of statistically reconstructed landscapes have grown in recent years. Some examples include schematic reconstructions based on various statistical summaries of data (e.g.. DiMichele et al., 2002; Liang, 2004; Ricardi-Branco & 2004), artistic reconstructions based on quantitative analyses (LePage & Pfefferkorn, 2000: Falcon-Lang et a 2001: 2002; Willard et al., 2007), and descriptive reconstructions based on a Rosler, detailed — Johnson & Raynolds, n variety of quantitative approaches and degrees of depth of statistical analyses of data (e.g.. Spicer & Hill, 1979; Phillips € DiMichele, 1981; Raymond & Phillips, 1983; Wing, 1984; DiMichele et al.. 1991: Pryor, 1996; Gemmill & Johnson, 1997; Wilf et al., 1998a; Figueiral et al., Johnson & Ellis, 2002; Spicer et al., 2 999: Howe & Cantrill, 2001 002: Hofmann n rarity of Martin-Closas & Galtier, 2005; 2005). Johnson (1999) discusses in detail the procedures & Zetter, 2005; Wehrmann et al., that went into the production of numerous recon- structed landscapes in the Prehistoric Journey exhibit These reconstructions are based on detailed field sampling of the Denver Museum of Nature and Science. of exposures and, thus, represent local views of much broader landscapes, not intended to be family portraits. Examples represented in this exhibit are (2002) and include the Pennsylvanian-age Hamilton. Quarry. of 1988), ate Cretaceous upper Hell Creek Formation illustrated in Johnson and Raynolds Kansas (Rothwell & Mapes. the biota from the of the northern Great Plains (Johnson et al., 2002), — and the exceptionally high-diversity early Paleocene Castle Rock flora of Colorado (Johnson & Ellis, 2002: Ellis et al., 2003) The Castle Rock flora demonstrates the fine line that separates a parautochthonous assemblage from an autochthonous assemblage. Preserved in a catastroph- ic flood, the Castle Rock flora is represented by a buried litter horizon with in situ tree trunks with roots in place in a subjacent paleosol. A later litter horizon with (possibly derived from dying trees—compare Burnham and Spicer [1986]) is preserved above the initial flood deposit (Ellis et al., 2003: fig. 4). The abundance patterns of buried by a second flood event. leaf positions of disarticulated individual plants demon- distribution in transect samples and relative strate minimal transport. Thus, this fossil deposit probably preserved regional spatial patterns and is similar to other stands of plants catastrophically buried by floods, where vegetation baffles sediment- laden water, inducing burial with minimal transport 1977; Allen & Gastaldo, 2006). (e.g., Andrews et al., — Direct reconstructions: Raising the de ad. Possibly the = most accurate reconstructions of past landscapes come from sites that were buried catastrophically by air-fall or waterborne volcanic ash, sediment carried by flood waters, often associated with low areas created by seismic activity, or by the encapsulation of vegetation by precipitation of minerals. The re- construction of such vegetation is done, essentially, by raising the flatte ned vegetation backup, hence the label as Burnham and Spicer (1986) described vegetation buried by and preserved by volcanic ashfall near El Chichón volcano in Mexico—an initial burial of the forest-floor litter, ground-cover plants, and bases of trees, with a second layer of leaves preserved above bed, the initial representing the leaves from the f the ashfall. A similar deposit was reported by Wing et al. (1993) standing forest, live at the time of Volume 95, Number 1 2008 DiMichele & Gastaldo Plant Paleoecology in Deep Time 165 from Big Cedar Ridge, Late Cretaceous of Wyoming. buried by air-fall volcanic ash. In situ preservation of the vegetation revealed that ferns were the dominant group, diverse at the whereas angiosperms were the most that groups is not major species level, demonstrating diversity within higher taxonomic necessarily an indicator of dominance of that group. measured as biomass. Other examples of in situ floras buried by ash have been described by Wagner (1989) from the late Pennsylvanian of Spain, and Pfefferkorn and Jun (2007) from the Early each case preserving plant stems and litter and Permian of China, in allowing spatial and whole-plant reconstructions. Catastrophic burial by flood-borne sediments often is associated with seismically induced changes in elevation, such as those accompanying the formation of Reelfoot Lake, Missouri, and the burial of the forest 1981). subsiding basins proximate lo sediment source areas, growing on that site (Penick, In rapidly forests may be buried autochthonously in multiple, 2004a), less active areas, seismic activity closely spaced layers (Gastaldo et al., and even in tectonically may be essential for the creation of accommodation space and the in-place burial of forests (DiMichele et al., 2007). Autochthonous vegetation buried by flood- borne sediment can be found throughout the fossil record of terrestrial. plants. Examples include the Devonian Trout Valley flora of Maine (Andrews et al., 1977; Carboniferous lowland, wetland tropical vegetation (Gastaldo, 1992b; DiMichele & DeMaris, 1987; DiMichele € Nelson, 1989; D . 2004b; Calder et al., latitude forests = the Permian (Taylor et al., Allen & Gastaldo, 2006), numerous cases of ¡Michele et al., ; 2006), high- 1992), Triassic (Cúneo et al., 2003) and Cretaceous (Falcon- 2001) cycadeoid stands (Wieland, 1916), and in situ fern thickets (Cantrill, 1996) of Cretaceous Gastaldo et al Lang et al., age, and younger vegetation, such as that of the Castle Rock flora mentioned above (Ellis et al., 2003). In all of these instances, To spatial aspects of the original vegetation are preserved. The most unusual way to preserve vegetation is by entombment in mineral matter precipitated around the standing plants. The best example of this kind of preservation is the Devonian Rhynie Chert, preserved in travertine deposits from mineral-laden hot springs (Rice et al., 2002). T plants and animals in this deposit is spectacular. In I I I he anatomical preservation of the addition, different assemblages of species, plant associations in particular sedimentary environments, and successional patterns have all been identified in what amounts to one of the most completely reconstructed fossil ecosystems. Autochthonously preserved plant deposits will have their greatest impact when many are known, espe- — cially at multiple stratigraphic horizons in close succession, and before and after major extinction boundaries. This is an area in need of much more attention, and in which the gradual accrual of data will eventually provide a richer array of data points for comparison. PALEOBIOGEOGRAPHY Paleobiogeography is an increasingly important component of paleoecology. Such studies of plants were among the biological patterns that lent strong support to the idea of continental drift, prior to the widespread acceptance of plate tectonies (see Chal- oner [1959] for review), and, later, helped refine study 1996; Berthelin et al., 2003). Vegetational studies also are of of continental positions (e.g., Ziegler et al., importance in the ground-truthing of climatic models (e.g.. Wing & Greenwood, 1993; Upchurch et al., 1999: Beerling & Woodward, 2001; Rees et al., 2002; 2007). geographic studies is, of Poulson et al., The resolution. of e course, much better in younger rocks, something paralleled by deu graphic reconstructions. Furthermore, although basic patterns were apparent in some of the earliest studies, the degree of resolution has improved significantly through time, with improved taxonomies and better paleogeographies. Paleozoic. The Paleozoic is characterized by a combination of major evolutionary changes in plants, major changes in the relative positions of continental landmasses, high levels of variation in atmospheric composition and global climate, and great variation in sampling intensity at different times and places (DiMichele & Hook, 1992). Consequently, from the Silurian to the end of the Permian there really was a series of worlds in which plant distributional patterns were changing, sometimes dramatically, through time. There also are time intervals of considerable un- certainty due to poor sampling, intercalated among density during which subtle can be detected (e.g.. Edwards, 1990; 2000). Through most of the later continental. landmasses were gradually times of rich sample patterns Laveine et al., Paleozoic, coalescing into the Pangean supercontinent, result- ing in many land-based connections permitting plant dispersal across wide areas. Thus, for the Silurian and Devonian and certainly for the later Carboni- has been traditional to ex ferous and Permian, 1 subdivide the world into large-scale floristic realms. Sampling evidence has consistently pointed to differentiation of three distinct regions: equatorial/ low latitude, and north and south temperate high latitude (e.g.. Chaloner & Lacey, 1973; Chaloner & 166 Annals of the Missouri Botanical Garden Meyen, 1973; Wnuk, 1996). Ziegler and colleagues 1981; 1985) added the element of climatie zonation, differentiating floristic provinces. Later, Ziegler (1990) and Rees et al. (1999, 2002) took an explicitly recognizing ancient biomes as climatically bounded (Ziegler et al., Rowley et al., — climate-based approach, paralleling vegetational distri- 1985). Ziegler et al. (2003) argue persuasively that atmospheric circula- vegetational units, bution on the modern earth (Walter, tion patterns are the major controls on vegetational distribution, with due regard for the effects of elevation and paleogeographic locations of moisture sources (e.g., Ziegler et al., 16 Summaries of Silurian and Devonian paleobiogeo- graphy, the most problematic of the Paleozoic, can be (1985, 2006), and Raymond and Metz Raymond = found in Raymond et al. (1987), Edwards (1990), (1995): these analyses suggest some largely latitudinal floristic differentiation with limited taxonomic overlap The fossil plant collecting sites from the Carboniferous between regions. much greater yields a more detailed understanding of phytoge- ography at that time. General summaries for the — Mississippian include van Der Zwan et al. (1985), Raymond (1985), Raymond et al. (1985), Scott and Galtier (1996), and Raymond (1997). lannuzzi and Pfefferkorn (2002 previously unrecognized warm-temperate floral belt in — recently presented a case for a the late Mississippian of the paleo-Southern Hemi- sphere, demonstrating the continuing need for field sampling and refinements of the systematic framework on which these floristic analyses rest For the Pennsylvanian and Permian periods, there are a relatively large number of analyses of plant distributions, including some on fine spatial scales. The classic papers are those of Chaloner and Lacey (1973) and Chaloner and Meyen (1973), which argue ifferentiation of floristic regions em for the progressive c throughout the later Paleozoic, from one to three to five. Thirty years later, Rees et al. (1999, 2002) and Ziegler et al. (2003) had subdivided the late Paleozoic world more finely and demonstrated. considerable similarity to patterns in the modern world. Fine-scale refinements of these broader patterns include those of Laveine and colleagues for the Pennsylvanian (La- 1989, 1993, 2000; Laveine, 1997). These authors tracked the migration patterns of pterido- veine et al., sperms across the tropies and suggested significant geographic rearrangements of the positions of micro- continental landmasses in the eastern Tethys. Simi- (2003) have demonstrated yet-to-be-understood physical connec- larly, for the Permian, LePage et al. tions permitting overlap of north temperate Angaran f ican equatorial regions, though principally extrabas- oral elements with those more typical of Euramer- abundance of inal equatorial elements. Broutin and colleagues 1995, 1998; Fluteau et al., 2001; 2004; Broutin & Berthelin, 2005) have demonstrated complex mixing of floras along the (Broutin et al., Berthelin et al., margin of the tropical regions, bringing together floras from western and eastern portions of the tropical belt with those from more southerly temperate regions. Cuneo (1996) and lannuzzi and Rósler (2000) have examined phytogeographie patterns in the Southern Hemisphere during the Permian. Pant (1996) summa- rized paleophytogeographic patterns in India during Li and Wu (1996 patterns in China, and Oshurkova (1996) compares examine such <= the late Paleozoic. north-temperate Angaran and equatorial Euramerican floristic trends. Mesozoic. During the Mesozoic, plants recovered from the major ecological disruption at the Permo- Triassic boundary, the Earth's climates were warm, at times from equator to pole, large animals dominated the herbivorous fauna for much of the time, and the angiosperms began to diversify and dominate some ecosystems. Consequently, as with the Paleozoic, ts many factors impinged on plant distribution and a simple summary of plant paleobiogeography during this time is not really possible. During the Early Triassic, low latitudes and low- elevational areas of the Earth appear to have been dramatically altered, probably in terms of their biogeochemical cycles, to such a degree that fossil floras are rare. There certainly were extinctions at or — near the Permo-Triassic boundary, the extent of which, a however, is poorly understood. Despite claims of devastation (e.g., Retallack et al., 1996), it is clear that most major plant clades survived and vegetation was well along in the process of recovering by the Middle Triassic (Looy et al., 1999; Kerp et al., 2006). Ziegler et al. (1993, 2003) summarize some of the physical and biotie changes that occurred during the Triassic. The Jurassic appears to have been a time of relatively long-term climatic uniformity over much of the globe, following a period of global warming at the 1999), By tying climatically sensitive sediments (especially Triassic-Jurassic boundary (McElwain et al., coals and evaporitie rocks) to the distribution. of plants, Rees et al. (2000) concluded that maximum plant productivity during the Jurassic was concentrat- ed in the midlatitudes, where the diversity of plant major groups also was highest. The tropics, in contrast, were dry and supported little if any rainforest vegetation; rather, such areas were populated by xeromorphic conifers and cycadophytes. Polar regions tended to be dominated by deciduous conifers and ginkgophytes. They further conclude that there were five major biomes that remained latitudinally stable Volume 95, Number 1 2008 DiMichele & Gastaldo 167 Plant Paleoecology in Deep Time throughout the period, moving across the continental landmasses as the positions of the continents changed. ant diversity followed the patterns of maximum climatic equability, being highest in the midlatitudes, dropping off toward the equator and poles (Rees et al., 2004); these authors also note a discrepancy between the distribution of plant and dinosaur fossils, which they attribute to taphonomic factors and argue that dinosaurs were most abundant in dry, savannah-like habitats rather than forests. Skog (2001) considered the biogeographic patterns of evolution of leptospor- angiate ferns during the Mesozoic and found that most have evolved in the with are known from and may midlatitudes or higher, associated moisture availability. ne Cretaceous has been studied extensive ly by E climate modelers. Some paleobotanical studies that have addressed Cretaceous climate, including plant- climate feedbacks, include those from Spicer et al. (1994, 1996), Upchurch et al. (1998, 1999), DeConto et al. (1999), and Beerling and Woodward (2001). Cretaceous plant productivity and biodiversity were broadly similar to those of the Jurassic in being where climates were most 1993). However, there may highest at midlatitudes, favorable (Spicer et al., have been high levels of terrestrial productivity in some areas of the tropics if continental and atmo- spheric circulation patterns favored high rainfall, given the climate-biome models of Upchurch et al. (1999). Certainly, peat accumulated at midlatitudes in 1987). Climate changes just prior to the end of the Cretaceous, western North America (Ziegler et al., particularly global warming, have been suggested. Cenozoic. By the Cenozoic, the amount of geographic and elevational coverage preserved in the fossil record is so much greater than that known or preserved in the Mesozoic or Paleozoie that patterns of plant distribution can be evaluated at a much higher level of detail than in deeper time. For example, Manchester (1999) provides a broad review of North American Tertiary floras, noting a long and well-documented record of floristic exchange among the major Northern — Hemisphere continents. during this time interval. Similarly, Jacobs et al. (1999) follow the history of during the Cenozoic, grass-dominated ecosystems lending support to a complex history prior to appearance of extensive grasslands in the Miocene. Retallack (2001c) has also reviewed the origin and spread of grass-dominated ecosystems and argues that the long-term co-evolution of grasses and grazers may have contributed significantly to cooling, drying, and climatic ae The most recent studies of the origin of g have used phytolith evidence E CS. 2002, 2004, 2005) and suggested that the evolutionary radiation of grasses and grazers was decoupled. Prasad et al. (2005), through the exami- nation of grass phytoliths in dinosaur coprolites, were that significant enough numbers during the Cretaceous to able to determine grasses were present in ve a food source for dinosaurs. Burnham and Graham — (1999) were able to examine the pattern of Neotrop- ical rainforest. species composition and turnover from the Miocene to the Recent by a combination of paleobotanical and neobotanical data. They show that plant migrations did not follow those of animals after the formation of the Panamanian isthmus three million years ago; rather, the southern forests have remained relatively free from major, ecologically dis- ruptive invasions. In addition to larger-scale biogeo- graphic studies, it is possible to detect elevational controls on vegetation in Tertiary-period deposits because of the preservation of basins at higher elevations, typically eroded in deeper time (e.g., Taggart et al., 1982; Wehr & Manchester, 1996). The point of these examples is to demonstrate the high degree of spatial and temporal resolution that can be achieved in the study of the biogeography of Cenozoic floras and major plant lineages, to a degree not com- parable in deeper time. In contrast to this generality, Collinson (2001) notes that the Cenozoic although often well known and with distributions exceeding those of their extant descendants, are in ferns, some cases more poorly known than ferns from the Mesozoic and Paleozoic, in large part because they are less abundant elements of Cenozoic vegetation than ferns from these earlier times. APPLIED PALEOECOLOGY The obvious applications for paleoecological data sets are in the assessment of spatial (within time units at the resolution of age or stratigraphic sequence) and temporal (across time at the resolution of epoch or period) patterns from which (1) communities and ecosystem structure can be evaluated (e.g., Taylor et al., 1992; Wing et al., 1993; Mosbrugger et al., 1994; Falcon-Lang € Cantrill, 2002; Gastaldo et al., 2004b; Falcon-Lang & Bashforth, 2004, 2005; Falcon-Lang, 2006; DiMichele et al., ples of stasis, turnover/replacement, and extinction 2007); (2) governing princi- and assembly laws can be tested (e.g., Knoll et al., 1984; Wolfe & Upchurch, 1986; Pfefferkorn et al., 2000; DiMichele et al., 2001b, 2002, 2004; Willard et al., 2007); and (3) phytogeographic realms can be Raymond, 1985; Raymond et al., 1993; 2000, 2002: Rees, 2002). Because plant communities are non- delimited (e.g., 985; Spicer et al., Rees et al., mobile, they reflect the unique climatic conditions Annals of the econ, Botanical Garden governing growth, reproduction, and dispersal across the landscape. Hence, the suite of plant organs, from which the community and/or ecosystem was recon- structed, is a proxy for the paleoclimatic conditions under which the parent plants lived. LEAVES AND LIVING RELATIVES AS PALEOCLIMATE PROXIES (1915, that leaf morphology could be used Nearly a century ago, Bailey and Sinnott a 1916) recognizec as a proxy for climatic conditions, and paleoclimate estimation protocols have long been in the refine- ment stage. Two principal approaches to climate estimation have arisen. (but see also Uhl € Mos- brugger, 1999) based on the morphology of angio- sperm fossil-leaf assemblages: leaf margin. analysis (LMA: Wolfe, 1979; Wing € Greenwood, 1993: Wilf, 1997) and climate leaf analysis multivariate pro- gram (CLAMP; Wolfe, 1993). LMA correlates mean annual temperature (MAT) with the proportion. of woody dicot species in a flora having entire. (non- toothed) leaf margins. CLAMP takes a multivariate (canonical correlation) approach. involving 31 morphological characters and 11 climate variables 2004, 2005). LMA is faster and simpler to use, but CLAMP offers the computation of a ing MAT, mean cold and warm month temperature, and mean (e.g., Spicer et al., ae number of different climate variables incluc annual precipitation (MAP), and offers the opportunity for additional statistical and. graphical manipulations (Green, 2006). For comparison of methods, see recent papers by Jacobs (2002) and Uhl et al. (2007). MAP has been somewhat more difficult to estimate. from leaves, and leaf-size analysis has been investigated as a simple and straightforward method to obtain. such estimates (Wilf et al., 1998b, 1999). both climate-estimation continue to test the validity, Proponents for models constraints, and controls on foliar physiognomic responses, often with disparate results when evaluating the same flora (e.g.. Wiemann 1998; Gregory-Wodzicki, 2000). With a bias toward preservation. within wetlands, Kowalski and et al., Dilcher (2003) found that living dicot communities growing in these North American settings have ¿ higher proportion of toothed leaves that resulted in an underestimation of inferred paleotemperatures rang- ing between 2.5 C and 10.0 C, regardless of method- ological approach. To redress the underestimate, they introduced wet-site regression equations and proposed that predictive equations be redeveloped using sites more analogous with those identified in deep time. Depending on the biome under consideration, both overestimates (e.g., Jacobs, d and underestimates (e.g., Greenwood, 2005) of een reported, with lero the result of low- limate parameters have — leaf diversity angiosperm floras. But, when high species diversity exists, such as in the paratropics and tropics, there may be a minimum number of taxa above which credible MAT estimates and somewhat less credible MAP estimates can be made. Burnham et al. (2005) [mm found a correlation between species richness and the accuracy of climate estimates in the Neotropics: estimates came within + 2°C of recorded tempera- tures when 50% of the highest rank-ordered taxa were used. More accurate estimates were achieved as the taxonomic number increased. When less than 25 taxa were used in an analysis, estimates were outside the error (Wilf, 1997). In their data. a bias toward serrated leaves existed in the data accepted + 2 C margin o set (comprised of the top 10 rank-order taxa), with this bias disappearing once the top 20 taxa were taken into (2005) con- cluded that habitat may not be relevant to the bias MAT found that MAP estimates for all their sites were too consideration. Hence, Burnham et al. associated with estimates. Additionally, they low, even when all taxa were used. The other approach to climate reconstruction ts the nearest. living relative (NLR: Mosbrugger. 1999) model, where the climatic requirements of the NLR of a fossil taxon is, or assemblage components are, estimate the paleoclimate under which the (1997) the coexistence approach used organism(s) lived. Mosbrugger and Utescher introduced a variation, (CoA). in which the climate requirements of a fossil assemblage are estimaled by using a compilation of the ranges of climatic requirements of each taxon's systematically NLR. Climatic variables (MAT, MAP, etc.) under which all taxa within the assemblage could have grown together are used to develop the estimates. Theoretically, this method can be applied to assem- blages where at least one NLR can be identified, but, obviously, resolution increases with an increase in the number of taxa in an assemblage with NLRs. Mosbrugger and Utescher (1997) demonstrate a strong congruence between climate parameters in both Recent and Neogene floras (88%-100% of coexist- ing taxa). Tests of the CoA against both LMA (Uhl et al.. 2003) and CLAMP (Utescher et al.. 2000: Liang 2003) demonstrate its applicability to Tertia- both approaches can produce sensible and consistent re- et al., ry assemblages. When compared with LMA, sults, depending on the flora (e.g.. Liang et al.. 2003), when the standard error of the leaf-physiognomy taken caution that reconstructions based on paleoclimate data is into account. (2003) leaf physiognomy are influenced by et al. factors nol re- lated to climate. such as sample size and taphonomy. When results are compared between CLAMP and CoA analyses, lower paleoclimate estimates were cal- — culated using CLAMP for European Mio-Pliocene Volume 95, Number 1 2008 DiMichele & Gastaldo Plant Paleoecology in Deep Time 169 1997; et al., 2000) and for a Miocene assemblage preserved assemblages (Mosbrugger & Utescher, in diatomaceous lake shales in Asia (Liang et al., 2003). A similar CLAMP underestimate is report- ed for the Shanwang site when compared with LMA Sun et al., 200; This begs the — as to whether or not such ~ protocols can be applied to the non-angiosperm world. Most non-angiosperm deciduous taxa have no NLRs. Thus, to date, climate estimates based on such taxa have met with only partial success. In the deepest time analysis to attempt LMA, Glasspool et al. (2004 evaluated — Permian-age Cathaysian gigantopterids, which have angiosperm-like broad leaves; the results of the analysis are limited, however, by lack of another independent proxy for temperature. Another taxon, -— Ginkgo L., is interpreted to have had similar eco- logical distribution since the Mesozoic in disturbed stream-side and levee environments (Royer et al., 2003). representatives (Dicranophyllum Grand'Eury) are still The ecological constraints of Late Paleozoic uncertain; the taxon has been found associated with Lower Permian Southern. Hemisphere coals (Guerra- Cazzulo-Klepzig. 2000). And, living Ginkgo is monospecific (G. biloba L.) and there Sommer & because are no unequivocal natural stands today (unlike the > recently discovered Wollemi Pine, Araucariaceae; Chambers et al., 1998), there is a high probability that Ginkgoaleans were adapted widely to a variety of 2002). On the other hand, of fossil Ginkgo are believed t habitats (see Rees et al.. stomatal trends o parallel those of the living taxon grown under varying atmospheric gas concentrations (Beerling & Royer, 2002a; see below). Due to the seeming obstacle of ecological homol- ogy, other workers have utilized permineralized wood assemblages to estimate paleoclimatic parameters 1990). Recently, though, Falcon-Lang (2005a) and Poole and van Bergen (2006) that the between climate and growth-ring parameters in (e.g., Chaloner € Creber, have demonstrated relationship modern trees, in part, invalidates the use of fossil woods as climate proxies. This is due to unconstrained variability in tree response to climate-forcing (i.e., growing-season length and growth conditions, season- ality, productivity) on a global scale, applicable to both angiospermous and coniferous woods (Falcon- Lang, 2005a, b). Proxy approaches to understanding ancient ecosys- tems and the environmental conditions of their for- mation at whole-system levels have been applied extensively to such settings as peat-forming environ- ments, where physical controls and taphonomy often are considered to be conservative and. independent of plant composition. There is an extensive literature Utescher on peat-to-coal models, including environmental inferences. Models, such as those of Diessel (1986) and coworkers, that would identify modern vegeta- tional equivalents, such as reed-dominated versus in Carboniferous take consideration the significant differences in the plants woody forest-dominated systems, forests based on coal macerals do not into from different times in geological history and, thus, risk inappropriate ecological conclusions. As pointed out by Collinson and Scott (1987) in a comparison of Carboniferous lycopsid-dominated systems with Cre- taceous to Recent taxodiaceous-dominated systems and DiMichele and Phillips (1994) in an examination of Carboniferous systems, the qualities of the plants cannot be ignored. STOMATAL DENSITIES AS CLIMATE INDICATORS The developmental response of leaves to elevated pCOz as reflected in leaf stomatal density has garnered great attention over the past decade (e.g.. McElwain & Chaloner, 1996). Application and limi- tations of the approach have been reviewed else- where (e.g., McElwain, 1998; Royer, 2001; & Royer, 2002a; Roth-Nebelsick. 2005), reader is directed to these contributions. An in- Beerling and the verse relationship exists between the pCO, and index (SI; 1999) such that stomatal indices (the ratio of stomata to stomatal Poole & Kürschner, epidermal cells) decrease as the concentration of at- mospheric CO» increases. This inverse response is found in both angiosperms and gymnosperms (e.g.. 2003) but is not limitless. It has been demonstrated in modern angiosperms that not Kouwenberg et al., only will different species within the same genus reach their own SI limit at different pCO. (e.g., 1997), but species same genus (Quercus) grown under identical climate Quercus L.; Kürschner, of the and atmospheric gas conditions exhibit statistically significant differences in both stomatal density and 2006). necessary when using this approach in deep time (Kürschner, 1997). There SI values (Cantor et al., Hence, caution is — aave been attempts to apply non-angiosperm stomatal indices in other parts of the pre-angiosperm record. For example, a number of researchers have used the Ginkgoales to examine geological trends in pCO». Retallack (2002) used extant Ginkgo biloba, in part, to calibrate responses of fossil Ginkgo species back into the Late Triassic, and extended this even further back into the Permian, using pteridosperm taxa with purported evolutionary affinities to the Ginkgoales. The assumption, of course, is that all these related species exhibited the same physiological and developmental responses (Beerling et al., 1998) at 170 Annals of the Missouri Botanical Garden all times and places. Chen et al. (2001) examined modern G. biloba and found significant differences in stomatal patterns relative to timing of leaf matu- ration, leaf development, placement on short versus long shoots, canopy position, and sexuality. They concluded that only mature leaves of G. biloba should be used in any analysis. They (Chen et al., 2001) then matched stomatal estimates of pCOs from four extinct species of Ginkgo, ranging in age from Early Jurassic to Early Cretaceous, to the pCO» curve of Berner (1998) and found congruence for only two. NLRs for Jurassic cycadaleans also have been used (e.g., McElwain et al., 1999). Using stable carbon-isotope data, Beerling (2002) extended the approach to Carboniferous arborescent lycop- sids using extant tropical Lycopodium cernuum L. for calibration purposes, although acknowledging closer affinity to extant /soetes L. Values derived from the stomatal indices of Carboniferous arborescent lycop- sids were consistent with the independent indicators of a drop in atmospheric pCO» concentration during the Late Paleozoic glaciation. In the absence of one or more NLRs, McElwain and Chaloner (1995) chose to use nearest living — equivalents (NLEs), defined by them as Recent taxa that grow in a comparable ecological setting and/ or are structurally similar, in their evaluation of Devonian (Juncus L. = Sawdonia Hueber and Aglaophyton D. S. Edwards; Psilotum Sw. = Aglao- phyton) and Carboniferous (Araucaria Juss. = Swil- lingtonia A. C. Scott & Chaloner and Lebachia Florin) plants. They found a marked contrast in all stomatal parameters between the extant and the extinct taxa. The Devonian fossil taxa exhibited significantly higher SI values than their NLEs, whereas values for the Carboniferous taxa plotted within the range of extant araucarians. These results were consistent with Berner’s (1991) GEOCARB model, confirming a massive atmospheric drop in CO» from the Early Devonian through the Permo-Carboniferous. Again, using the NLE approach, with ginkgoaleans and cycadaleans, McElwain et al. (1999) linked major species turnover at the Triassic—Jurassic boundary to a 3 C—4 C increase in global temperature. For deep-time taxa without either an NLR or NLE, physiological and developmental responses have been inferred using even more generalized assumptions about stomatal responses. For example, Hesselbo et al. (2003) used the stomatal densities (SD) instead of SI values in plants from a Middle Jurassic sequence of Yorkshire to develop a general pattern. Although SI is more reliable than SD (Poole € Kurschner, 1999; Royer, 2001; Beerling & Royer, 2002b preservational problems dictated use of the latter. — 5 Stomata patterns in pre-Mesozoic gymnosperms also have been used to develop proxies for pCO». Cleal et al. (1999) evaluated stomatal parameters in the pteridosperm Neuropteris ovata Hoffmann across the Moscovian-Kasimovian (Wespthalian-Stephanian) — »»undary. Their results suggest an increase in al- mospheric CO» in the very early Kasimovian, con- sistent with changes in vegetational composition and carbon sequestration (DiMichele & Phillips, 1996; Cleal & Thomas, 2005). More recently, Wang and Chen (2001) extended the application to the extinct arborescent lycopsids, using cuticles recovered from leaf cushions and distal sporophyll laminae. The shift in values across the Upper Permian of China, tied to a shift in lithofacies, is attributed to a change in the physiological response of these taxa from Cs to CAM photosynthetic pathways in response to an increasingly dry climate. INFERRING PHYSIOLOGICAL RESPONSES OF EXTINCT PLANTS AND THEIR IMPLICATIONS Paleobotanists are increasingly finding ways to estimate plant physiological functions and responses (2006) developed a model to evaluate photosynthetic capa- to environmental conditions. Konrad et al. bility as it relates to diffusion resistance of the epi- dermal cells based on the size and depth of stomatal pores and mesophyll parameters (e.g. intercellular air space, mesophyll-cell area, and membrane resis- — ances) They also added functions to evaluate the relationship between assimilation and temperature. The model permits use of stomatal morphologies and densities to calculate past CO». Similarly, Royer and Wilf (2006) have proposed and evaluated a gas- exchange hypothesis that explains the relationship between leaf-margin morphology and climate vari- ables. They tested the efficacy of the model, found — that it accurately predicted the response of leaves from colder climates, and found an increased effici- ency in transpiration and photosynthate production early in the growing season in taxa with toothed margins versus those with entire margins (this also may explain the higher proportion of toothed-leafed taxa in stressed wetlands). Additionally, plant re- sponse is enhanced at higher-temperate latitudes (38.8°N, colder temperate Pennsylvania; 35.7 N, war- mer temperate North Carolina) where the plants maximize carbon gain when temperature is limiting but moisture and nutrient availability are not. They have concluded that the increased physiological functioning associated with serrated leaf margins may provide a proportionally increasing selective advantage with decreasing temperature (or water stress). This is reflected in empirical correlations used for paleo-temperature estimation. Another recent DiMichele & Gastaldo Volume 95, Number 1 2008 Plant Paleoecology in Deep Time — hypothesis relating toothed-leaf margins to temperate climates is that of Feild et al. (2005), who suggest that teeth act as points of guttation, a process of likely physiological traits of evolutionarily basal angiosperms (Feild et al., 2000, 2001, 2003a, b). In the deeper past, Phillips and DiMichele (1992) relieving flooding of leaf tissues by root pressure on considered the consequences for arborescent isoeta- the vascular system. These authors propose that by relieving the negative effects of flooding, guttation permits root pressure to drive leaf developmental expansion. Cool, high-humidity spring weather, com- — ean lycopsids of CAM metabolic pathways, which have been found in modern isoetaleans (Keeley & Busch, 1984) and also inferred from carbon isotopic analysis (Raven & Spicer, 1996). Boyce and col- leagues (Boyce & Knoll, 2002; Zwieniecki et al., 2004; Boyce, 2005) examined the physiological at- tributes of leaves and their xylary support systems in modern seed plants and pteridophytes as indi- cators of possible constraints on the evolution of leaves in primitive gymnosperms and ferns. As part of this, Boyce et al. (2004) examined the physiology of tracheid lignification in a phylogenetically di- verse sample of extant vascular plants. They postu- lated trade-offs between hydraulic transport and bined with moist soils in many temperate regions, would be conditions in which guttation was favored. In addition, this model explains the overabundance of toothed-margin leaves in wet habitats (Burnham et al., 2001; Kowalski & Dilcher, 2003). In contrast to these organ-scale studies, Beerling and Osborne (2002), working at the level of the biome, considered such responses as carbon balance in high-latitude conifer- ous forests to elevated carbon dioxide levels. Studies of the leaf economies spectrum in extant plants (Wright et al., 2004) have revealed nonlinear correlations among leaf morphological and physio- logical characteristics arrayed along a single rate spectrum. At the extremes, what Wright et al. (2004) call quick-return leaves are characterized by high nutrient concentration, high photosynthetic and re- spiratory rates, short life spans, and low biomass per support as the evolutionary driving force of lignifica- tion patterns. Lignified tracheidal cells are found in plants with small leaves or leaves with needle- like construction. In those plants with larger lami- nate leaves (ferns and angiosperms), lignification of tracheidal cell walls is decreased, enhancing water transport, with support functions shifted to heavily unit area; at the slow-return end are leaves with low lignified fiber cells for support. nutrient concentrations, long life spans, low photo- synthetic and respiratory rates, and high investment in ROUND-TRUTHING CLIMATE MODELS biomass per unit area (usually in antiherbivore tissues). Correlations of these traits with climate were Inferences about climate based on plant physi- present, especially when constrained by biome or growth-form/functional group. Wilf et al. (2001) used these morphological traits to estimate | ognomy and ecological tolerances casts plants as natural, independent sources of information with which to ground-truth global climate models at con- tinental, hemispherical, and global scales. Such an eaf life spans and position within the economics spectrum for late Paleocene and Early to Middle Eocene dicotyle- approach was used by Wing (1991) in a critique. of donous angiosperm leaves, and found close correla- computer models (Sloan & Barron, 1990). The tions among herbivory and inferred plant defenses presence of large nonburrowing tortoises, palms, tree in response lo independently assessed changes in paleoclimate. Royer et al. (2005) used and extended these morphological-physiological correlations. Based western interior North America was on a preliminary study (Huff et al., 2003), Royer et incompatible physiologica al. (2005) quantified univariate correlations among ~ ferns, and dicot trees lacking strong seasonal growth rings in Late Cretaceous and Paleogene deposits of considered ly with sustained below- freezing winter temperatures and a high degree of seasonal temperature fluctuations predicted by the models. Wing (1991) extended the biological argu- ment to the entire Tertiary prior to ca. 34 Ma. He — leaf traits and. demonstrated. statistically significant patterns (number of teeth, perimeter ratio, shape factor, and margin percentage); this latter study also demonstrated several relationships between leaf physi- — found paleontological evidence to be compatible with ognomy and leaf economics. From these results, Royer the climate-simulation parameters in only one in- et al. (2005) developed a set of refined leaf-climate stance for the Paleocene and models using multiple linear regressions. The NLR approach to study of extinct plant physiology has been applied to various plant groups from different times in Earth's history. Perhaps later Eocene. More recently, Fricke and Wing (2004) have evaluated both LMA and stable isotopes, from phosphate in mam- malian tooth enamel and freshwater fish scales, to estimate the MAT in the Early Eocene across a North American latitudinal gradient from Ellesmere Island to Texas. They acknowledge that even in areas where ct the most explicit comparisons have been made by Feild and colleagues, who have combined studies of angiosperm phylogeny with physiology to examine — both megafloral and megafaunal assemblages occur in 172 Annals of the Missouri Botanical Garden the same stratigraphic section, these are rarely preserved. within the same bed (same chronostrati- graphie horizon; this also applies to the entire is difficult to MAT between areas due to age Phanerozoic record). Furthermore, it make direet statistical comparisons in different pale OLE ographic uncer- MAT short periods of time (Milankovic time scales). tainties because can change significantly ove Con- trary to the straight-line application of Recent corre- latives between these parameters, Fricke and Wing (2005) demonstrate that the oxygen isotope-MAT correlation during the early Eocene was significantly =) different from the climate proxies obtained througl LMA ("0 estimates were 0.2°C to 2.1°C They also found that the absolute values for higher). o'?0 differed between the Bighorn, Green River, and Pow- der River basins of Wyoming, implying that different hydrological regimes existed in these areas. Regard- less, this combined approach toward paleotemperature estimates captured a global-warming event at the Paleocene—Eocene boundary not determinable from paleobotanical data alone. Deeper time climate models are more problematic simply because both abiotic and biotic data are more remote, meaning sparser data, limited modern proxies with which to compare, poorer re- gional stratigraphic temporal correlations, and, in some cases, exacerbation of a basic lack of under- standing of the complexities of the Earth-climate system. Nonetheless, it is possible to combine geo- logical data and paleontological data in the search for common patterns, allowing climatic inferences based on physical data to be s to biological Montañez et al., 2007 2007). Late Paleozoic changes from a cool to warm Earth (Gastaldo et al., kind of Southern Hemisphere changes (e.g.. : Poulsen et al., 1996) serve as an exam- ple of this study. During the Permian. deglaciation occurred in a stepwise pattern, inferred from stable isotopes of marine (Veizer et al., 1999) and paleosol (Royer et al., 2004) interglacial pulses have been identified, each distin- carbonates of Euramerica. Three glacial and guished by a òO excursion, with glacial ice present during the late Gzhelian—Asselian (Carboniferous— Permian), across the Sakmarian—Artinskian boundary, and during the Kungurian—Roadian. Complete Gond- wanan deglaciation occurred during the Middle Per- although there are some suggestions 2005). carbon dioxide mian (Roadian), of later minor ice pulses (Fielding et al., Preliminary estimates of atmospheric concentrations suggest that Permian pCO: was up 10 times greater than during the Late Carbonifer- ous, in which a pCO» concentration of ~350 ppm parallels the present atmosphere (Montañez et al., 007 ^ complex spatio-temporal vegetational re- sources of pCO» is recorded in tropical paleolatitudes of the south- U.S.A. This wetland megafloras to those that grew under more sponse to these cycles and overall increasing western pattern. shows a trend from seasonal rainfall patterns, many elements of which are typically thought characteristic of the Late Permian and Mesozoic (DiMichele et al., 2001a). The overall megafloral compositional changes are similar in western North America and Central Europe, although = correlation problems prevent exact comparisons o timing (Scott, 1980; Kerp & Fichter, 1985; Broutin et al.. 1990; DiMichele & Aronson, 1992). Overall, this 'al- ( lc | ge ologic pale mntologic al combined approach dem- onstrates a close correlation of changes in the abundances of major groups of plants and inferred changes in atmospheric CO, soil Major patterns occur at each inflection point of the pCO» moisture, and regional temperature. changes in landscape record rather than at stage boundaries (Montanez et al., 2007), as suggested by global diversity compi- Rees et al., 2002). patterns provide insight into the combined effects of Continental-scale allons (e.g. changing global climate and paleog . Howev- er, finer sequence-scale analysis within the regional sedimentological, chronostratigraphic, and geochemi- nuances of terrestrial cal framework captures the ecological response to perturbation. BIODIVERSITY PATTERNS IN SPACE AND TIME Applications of synecological analyses have primar- ily been focused on diversity studies at the continental, hemispherical, and global scales. In general, the sys- tematic diversity of collection sites across an area of interest has been compiled into databases (e.g.. The Paleobiology Database, 2007), with all floras binned — within a stratigraphic interval (usually stage-level). Spatial and temporal changes can then be evaluated to identify patterns of diversification, range extensions (or contractions), and ecological assembly. The best resolution of paleobiogeographic E is in the extensive Tertiary. plant-fossil record, which there are many extant angiosperm families. The quality of this record has led to many recent studies of group origination patterns, diversification, and range changes relative to present biogeographic distribu- 2001). It is beyond the scope of this review to document all that has been tions (e.g., Tiffney & Manchester, published relative to angiosperm paleobotany in these respects. A few illustrative examples will suffice to demonstrate major research approaches. Recently, Corbett and Manchester (2004) evaluat- ed the stratigraphic and paleogeographic distribu- tion of Ailanthus Desf., a Nort commonly — ern. Hemisphere genus encountered as fossils in lake sediments Volume 95, Number 1 2008 DiMichele 8 Gastaldo Plant Paleoecology in Deep Time 173 associated with temperate evergreen taxa. The oldest megafloral remains are known from the Early Eocene of North America and Asia, with the genus appearing in Europe by the Middle Eocene and ultimately achieving a cireumboreal distribution. Presently, Ailanthus is restricted to the Indian subcontinent, Southeast Asia, the Indonesian archipelago, and northern Australia (violating Wallace's line). Corbett and Manchester (2004) conclude that Ailanthus either originated in the Northern Hemisphere and subsequently extended its range southward, or vice versa. Macrofloral evidence exists for the former but is suspect for the latter (there is a dearth of fossil evidence in these geographies). Any attempt to resolve this dilemma would require the use of a molecular phylogenetic analytical approach that may provide evidence for the ecological limits of the ancestral lineage. Such an approach has been taken by Xiang et al. (2005, 2006) for Cornus L. They combined DNA data with morphological traits of extant and extinct taxa. From an extensive set of analyses (PAUP [maximum likelihood], parsimony, and Bayesian phylogeny in- ference and divergence-time analyses), they conclude that Cornus underwent five sequential intercontinen- tal dispersals since origination in Europe. During the late Paleocene, its range expanded into North America via the North American land bridge, with separation and climate forcing isolation (C. sessilis Torr. western U.S.A.) and extinction (eastern and central North America). Range extension occurred into Africa at about the same time, with ex Durrand. these taxa retaining an evergreen habit in tropical montane-evergreen forests of eastern Africa. Europe- to-Asia transferral occurred twice, in the mid-Oligo- cene and mid-Miocene, arriving on this continent also as an evergreen lineage. Documentation of such phy- logeographical histories and assemblages becomes the basis for understanding ecosystems and their respons- es to perturbation (e.g., Collinson & Hooker, 2003). Extrapolation into deeper time where there are stratigraphic, paleogeographic, and paleoclimatic com- plications results in much coarser temporal and biogeographic resolution. Rees et al. (2004), for ex- ample, used climatically sensitive sediments, plant diversity, and dinosaurs to infer broad geographic patterns for the Late Jurassic (Kimmeridgian and Tithonian, 155.7-145.5 Ma). Using the data com- piled by Rees et al. (2000) and plotted within latitu- 10°, latitudes where dinosaur dinal bins of they note low plant diversities in low remains are well preserved (ascribed to a taphonomic bias and a small number of localities), increasing diversity to- ward the midlatitudes, and a pole-ward diversity decrease (where megafaunal elements are virtually unknown). High-midlatitude diversity forests were dominated by a mixture of conifers, cycadophytes, pteridosperms, ferns, and sphenophytes. This stands in marked contrast with low-latitude xeromorphic vegetation, characterized by microphyllous conifers and cycadophytes, and polar assemblages dominated by macrophyllous conifers and deciduous ginkgo- phytes. Although it was only possible to delimit coarse spatial oe and Se and tem- pora 3 Ma) re- ae con- ; Tithonian, (Kimmeridoian 4€ gian, 4.9 solution of these vegel oul Ls clude that “an immense biomass and diversity of [om dinosaurs was apparently supported by sparse vege- (Rees et al., 2004: 649) occupying a tropical savannah biome (Hees et al., 2000) A similar stage-based approach was taken by Rees et al. tation” (2002) for the transition from icehouse to greenhouse conditions in the Permian (Gastaldo et al., 1996). taking into account the northward movement of continents over this interval. Plant-assemblage diver- sities were compiled at the generic level using taxonomic taken from the literature for the Sakmarian (285-280 Ma) and Wordian (267-264 Ma; Rees et al., lists 1999). These were used in conjunction with dry and wet climatically sensitive lithologies — o interpret polar-to-equatorial climate | gradients within each stage. To limit spatial and temporal averaging, megafloral data were restricted to inter- locality distances of ca. 100 km and vertically stricted throughout the stratigraphies. The resultant 112 genera from 128 localities (799 occurrences); the Wordian was 1001 occurrences). From these data, nine biomes (tropical ^ re- Sakmarian data set consisted of comprised of 104 genera from 147 localities — ever-wet to tundra) were identified, and the morpho- logical adaptions of the plants within each biome were used to infer precipitation and temperature patterns within the established paleogeographical context. Biomes were compared against computer climate- generated biome models run at two levels of pCO» (4X, 8X) to assess the credibility of the interpreted distributions. (2002) found that biome simulations run at elevated pCO: matched the em- Rees et al. pirical data in the equatorial, midlatitudes, and northern high latitudes, although no simulation re- produced the temperate climates in high southern latitudes where alpha diversity is limited to fewer than 15 taxa (Rees, 2002). As Rees (2002) notes, each regional pattern is a reflection of sampling and the effects of regional-level geographic and climatic changes (wet to more seasonal) on different different. biomes. taxa growing in In addition, per- turbations, extinctions, and turnover al stage level play a significant role in the percep- tion of broad regional patterns when these assem- blages are amalgamated. 174 Annals of the Missouri Botanical Garden PHYLOGENETIC ANALYSES IN PALEOECOLOGY Accounting for the effects of history on ecological patterns and dynamics is rapidly growing in impor- tance. Attempts have been made both to accommodate and to remove the effects of historical patterns in ecology, by considering both phylogenetic relatedness and biogeographic history. The integration of phylo- genetic pattern in the analysis of ecological pattern and process leads inexorably toward further consid- eration of the fossil record. Fossils not only provide direct evidence of the combinations and distributions of extinct organisms, but also permit dating of nodes in cladograms (Crepet et al., 2004), which may prove where, and under important in determining when, — what environmental conditions clade ecological traits may have been established. Conversely, if clade ro membership strongly affects local ecological struc- ture, it may be difficult to examine patterns of con- em vergence among ecosystems from different times anc pai places, and thus, clade membership attempts to ac- count for phylogeny by recasting ecological descrip- tions in terms of functional groups or ecomorphs. INCLUDING PHYLOGENY In neoecology, the consideration of phylogeny has Be led to findings that contradict traditional views o plant ecological variation. One need not look far to find the opinion that all plants do the same thing, and, hence, much convergence in ecological patterns is to be expected across widely unrelated lineages, render- ing history unimportant. However, work such as that of Ackerly (1999), Webb (2000). or (2001), for example, or in a number of papers (e.g.. Cavender-Bares et al, 2006; Kembel & Hubbell. 2006; Silvertown et al., 2006) in a special issue of the journal Ecology devoted specifically to this subject (Webb et al., iid suggests that ecological distri- butions of species o Prinzing et al. taxa tend to be more — Ig sher clustered with i to spatial and habitat conditions than expected by chance. Thus. all plants are not doing the phylogenetic and other historical effects on biodiver- same thing and there may be strong sity and the ecological mediation of evolutionary patterns. Such integration of explicit historical con- siderations and ecological patterns and processes appears to be an area with tremendous potential to have wide, ecology (see summaries and comments by Webb et al. [2002, 2006], Ricklefs [2006], and Westoby [2006]). There are some paleoecological studies that have examined or considered ecological patterns in light of phylogenetic patterns. Examples at a relatively finely Nem reso profound effects on many aspects of ved taxonomic scale include Burnham’s study of early Cenozoic elms (1983, 1986), in which the major genera of the subfamily Ulmoideae were characterized morphometrically and. with respect to their climatic, particularly temperature, tolerance ranges. In each of three genera, species-level diversifications were found to occur during times of climatic instability in the late Eocene. However, new species had the same basic climatic tolerances as both the ancestral forms and those found in the living representatives today, suggesting strong phylogenetic niche conser- of another the clade level. In a study f the Ulmaceae in the subfami- vatism at Cenozoic member of ly Celtoideae, Cedrelospermum Saporta, Manchester (1989) found that the several known species of this genus were lypically colonizers of areas disturbed by volcanic ash deposition; some extant members of the Celtoideae share this peculiar ecology. Although e: he edaphic distribution of Cedrelospermum was nar- row, Manchester notes that it was widely distributed through a variety of climatic conditions. Wing and Hickey (1984) examined the systematies and pa- eoecology of phylogenetic relationships within the Juglandaceae, focusing on the genus Platycarya Siebold € Zucc. They found Tertiary Platycarya spe- cles be thicket-forming shrubs characteristic. of open, early suecessional habitats. This particular eco- logy is characteristic of species of this genus up to the present day. The use of nearest-living relatives lo infer the climatic tolerances of fossils, discussed above (Mos- brugger & Utescher, 1997; Mosbrugger, 1999), makes the implicit assumption that clades are conservative in their ecological tolerance ranges. By looking across a flora instead of using a single species, the coexistence approach to estimate such climatic tolerances of 1997) adds considerable support to the general assumption of extinct forms (Mosbrugger & Utescher, clade conservatism in niche breadth. From a still. broader, (1988a) examined the origin of Eocene and Oligocene floras in the Rocky Mountains and concluded that derived microthermal lineages, those typical of cooler. floristic perspective, Wing — more climatically seasonal climates, originated in high-elevation areas, whereas the lower elevations were dominated by ancestral megathermal (warm equable climate) forms. Once climatic cooling spread throughout North migrated into and dominated cooler climates that had western America, these lineages become widespread through the lowland regions. Similarly, DiMichele et al. (2001a) found Early Permian floras dominated by phylogenetically derived conifers and cycads, plants previously known only from Late Permian or Mesozoic deposits, associated with a tongue of sediment in the western Pangean equatorial belt deposited during cooler. strongly Volume 95, Number 1 2008 DiMichele & Gastaldo Plant Paleoecology in Deep Time seasonal climates, possibly coincident with a last-gasp pulse of high-latitude glaciation (Montañez et al., 2007). Ancestral forms typically grew in more equable seasonal settings. Both these examples lend support to models that link evolutionary innovation to the crossing of ecological thresholds (e.g., Wiens € Donoghue, 2004; Rickleffs, ) and the establishment of basic clade-level eco- logical traits during these initial radiations. At a more global spatial scale, the role of history in constraining hypotheses of evolutionary inferences was considered by Manchester and Tiffney (2001) who, again using elms, demonstrated that inclusion of extinct taxa and information on past geographic dis- tributions not only provided more complete phyloge- nies and added constraining age data, but permitted significant refinement of interpretations on place and conditions of clade origin and evaluation of age and area hypotheses. They conclude that the fullest picture comes from a combination of neobotanical and paleobotanical data within a phylogenetic context. At the deepest time level, DiMichele and Bateman (1996), Bateman et al. (1998), and DiMichele et al. (2001b) examined the origin of vascular-plant body plans, roughly equivalent to Linnean classes, during the Middle to Late Devonian and concluded that there was strong overlap between phylogenetic affinity and ecological distribution, with lycopsids occupying primarily wetland habitats, seed plants in terra firma habitats, sphenopsids in a narrow range of aggrada- tional settings, and ferns as opportunistic weeds in terra firma environments. Hotton et al. (2002) found evidence of this same basic clade-by-environment sub- division as far back as the Early Devonian, wherein the ancestral clades (zosterophylls and trimerophytes) of the main Late Devonian (and beyond) body-plan groups were already established along the wet-dry- substrate spectrum. Perhaps the largest scale and most general study to attempt to understand patterns of ecological conver- gence and divergence is that of Niklas (2006), who compared allometric patterns of biomass partitioning bryophytes, plants, explicitly examining the patterns within a in aquatic. macrophytes, and vascular phylogenetic framework. He found a similar pattern of resource allocation to functionally equivalent body parts that are otherwise developmentally distinct and non-homologous in these different groups. From this, he concluded that there may be a single scaling relationship describing biomass partitioning into stem, leaf, and root for all eukaryotic photoautotrophs. This finding, as Niklas (2006) noted, permits the use of simpler, more general model systems to develop analytical solutions to explain biomass partitioning patterns in higher plants. At this level, the inclusion of explicit phylogenetic consideration leads to the conclusion that at some level of analysis, such con- siderations can be ignored (see next section). similar to studies of allometry within a phyloge- netic framework are recent studies of physiology linking fossil and modern plants by Boyce and 2003, 2004; Boyce, 2005), discussed above. This work has permitted convergent colleagues (Boyce et al., patterns in morphology to be identified and linked to underlying physiological controls, revealing certain patterns of trade-off in the basic plant functions, sup- port versus hydraulics. These studies link closely with those of the leaf economics spectrum (Royer et al., 2005, also discussed above), in that plants with small, structurally complex and highly defended, long-lived leaves differ in terms of the lignification patterns in their hydraulic and support systems (having the more primitive condition) from plants with shorter-lived, more physiologically active foliage. EXCLUDING PHYLOGENY Just as there have been attempts to include phylogeny and examine the effects of relatedness on ecological patterns, there have been (much more long- standing) attempts to remove such effects. In neoecol- ogy, these attempts fall broadly under the study of guilds (Root, 1967; Simberloff € Dayan, 1991) or functional types (Smith et al., 1997). Such approaches implicitly assume that evolutionary relatedness will bias the distribution of plant taxa in space and by resource use paltern, thus impeding comparisons among ecosystems that might otherwise be found to converge (or not) in structure. This approach also aims to make systems comparable in such cases where true morphological convergence exists, but is masked by the use of taxa as the variables by which ecological assemblages are categorized. Functional-type analyses may reclassify plants via multivariate analysis based on their morphological traits, or place them into groups or guilds based on inferred functional roles within ecosystems (Citay & Noble, 1997; Westoby & Leish- man, 1997). This approach may be applied to entire ecosystems as well as to the invididual taxonomic components of those systems (Shugart, 1997). There have been a few paleobotanical attempts to remove taxon effects, an approach that has been used somewhat more extensively in studies of vertebrate paleontology (see Damuth et al., 1992). Wing (1988b; Wing et al., flowering plants, gymnosperms, and lower vascular 1992) placed 52 species of Eocene-age plants into 12 ecological categories based on foliar physiognomy or comparisons with NLRs, basically a guild approach. The subsequent assemblages then were analyzed statistically; structurally distinct assemblages 176 Annals of the Missouri Botanical Garden were found to correlate with different physical habitats. DiMichele and Phillips (1996) rescored Pennsylva- nian-age plants according to a suite of ecological traits such as reproductive allocation, body size and form, and dispersal capability. They found that the phyloge- netic signal could not be removed: Pennsylvanian wetlands were dominated by several widely distinct phylogenetic lineages, representing Linnean class and ordinal level groups. Thus, distinct body plans repre- sent different fundamental approaches to being a plant and result in strong partitioning of ecological resources by clade. A similar pattern was found in the Middle to Late Devonian radiation of vascular plant classes (body plans) by DiMichele et al. (2001b), in which phylogenetic patterns, ecomorphic (functional. type distributional patterns, and ecological distributional patterns were found to overlap strongly, meaning that historical-phylogenetic patterns could not be removed except to separate ground cover from canopy and understory plants. Thus, functional-type categoriza- tions of angiosperm-dominated systems appear to contain more convergences than those of the Paleozoic, where different kinds of ecosystems were dominated by phylogenetically widely disparate groups. This is an — area where more examination is needed and one thal should tell us much about evolution and the nature of ecological resource use and occupancy. INTEGRATIVE STUDIES EVOLUTION AND ECOLOGY [em The conjunction of evolution and ecology has long been sought after. Yet, the two have long been prac- ticed largely separately, despite such pithy charac- — terizations as Hutchinson's (1965) ecological theater and evolutionary play, or Dobzhansky's (1973) classic observation that nothing in biology makes sense ex- cept in the light of evolution. Certainly, explicit link- age is more common in neobiology than in paleontol- ogy, if for no other reason than the vast differences in the number of practitioners. Nonetheless, because paleontologists are faced relentlessly with evolution- ary and ecological history and their contingencies, there have been many attempts to bring them together at all scales from individual structural features to particular species to major groups. Perhaps the best known example in paleobotany of the quest for an ecological explanation of evolutionary pattern is the origin of the angiosperms, Darwin's abominable mystery. In fact, this group is no more — unheralded in its appearance than many of the major classical Linnean plant ordinal or class-level body- plan clades; there are just so many angiosperms around today to attract our attention. And. despite intensive study of seed-plant phylogeny over the past 20 years, a consistent, agreed-upon sequence of morphological transformations linking the angio- sperms to the other seed plants has not been reached (e.g.. Hilton & Bateman, 2006. for review). Thus, the angiosperms are representative, almost iconic, in fact, f an evolutionary problem—the origin of major morphological discontinuities. Explanations vary all over the conceptual map, and, still again, the subject is too complex for a comprehensive review here (see Feild & Arens, 2005). The implicitly gradualist hy- pothesis of a long-shrouded upland origin (Axelrod, 1952. 1972; Stebbins, 1965) visualized aridity as the driving selective force. A more sophisticated ver- sion of this model, making fewer assumptions about the tempo and mode of evolution and taking a more critical look at early angiosperm life histories, iden- tified early angiosperms as weedy opportunists con- fined to river corridors, where they may have moved into areas of reduced competition in lowland settings from a place of origin in drier, marginal, extrabasi- nal habitats, also characterized by low competition (Hickey & Doyle, 1977). Taking a similar view, that physiological stress was the driving force of angio- sperm morphology, but with an opposite viewpoint on where the selection occurred, Retallack and Dilcher (1981) put forth a coastal hypothesis, arguing that early angiosperms moved up from the coast rather than down from the uplands. In a modification of the upland weed model, early angiosperm evolution was hypothesized to have occurred directly within dis- turbed lowland habitats (Taylor & Hickey, 1992). Most recently, Feild and colleagues (Feild & Arens, 2004; Feild et al.. 2004) have attempted to integrate physiological analyses of phylogenetically identified basal angiosperms (e.g., Feild et al., 2003a. b) with the fossil record. They suggest that disturbance and browsing by large vertebrates, specifically dinosaurs, n forest-shaded understories suppressed slower- growing ferns and gymnosperms and permitted rapidly growing, more physiologically escalated, angiosperms (sensu Vermeij, 1987) to make significant gains in resource occupancy. Labeling their hypothesis dark mm and disturbed, they note that the concept of ke innovation in this instance is unlikely and that angiosperm dominance is a product of the overlap of a series of essentially unrelated factors. Considerable evidence has accumulated from Late Paleozoic deposits, suggesting that novel body plans originate in areas of reduced competition, marginal to basinal settings, in effect lending credence to earlier models of upland (in reality, extrabasinal) origin of novelties, but without the presumption of long periods of gradual, microevolutionary change (recall the geo- logical context, in this case the continental strati- Volume 95, Number 1 DiMichele 8 Gastaldo 177 2008 Plant Paleoecology in Deep Time graphic record discussion, above). Most of this evi- in locally variable ways. Furthermore, there are dence comes from the appearance of scrappy, allo- jon chthonous remains of plants from derived clades (such as conifers, peltasperms, cycads, and corystosperms) in association with basinal wetland floras dominat- ed by primitive lineages (such as lepidodendrid ly- copsids, medullosan pteridosperms, marattialean tree ferns, and calamitean sphenopsids). For example, conifers first appear in the Middle Pennsylvanian, but only as bits and pieces of transported material, usually in basins proximate to tectonically active up- land areas (Scott & Chaloner, 1983; Lyons & Darrah, 1989). Later floras rich in these xeromorphic plants oecur in rare beds, intercalated within sequences otherwise characterized by dominance of wetland plants, with little overlap in taxonomic composition between these two kinds of floras. Ultimately, these initially. extrabasinal, seasonally dry floras become the predominant vegetation in the basinal lowlands, correlating with directional changes in climate (Broutin et al., 1990; DiMichele & Aronson, 1992). rp - — 1e pattern repeats again as vet additional floras and vegetation types appear consisting of even more derived elements (DiMichele et al, 2001a, 2004; Kerp et al., 2006); in these examples, the derived plants often had been known only from much younger rocks, even of Mesozoic age, prior to their discovery in rare Paleozoic assemblages. At the more detailed level of specific organisms, studies of fossil-plant biomechanies, when combined with stratigraphy and phylogenetic analyses, permit f ecology. Examples, discussed earlier, include such o the evaluation of morphological trends in light « things as the response of plant-growth architecture to wind shear via the evolution of stem structural safety factors (Niklas & Speck, 2001), or the filling o morphospace, considered in light of plant functional morphology (Niklas, 1999, 2000). lur] VEGETATIONAL RECOVERY FROM MEGADISTURBANCES Studies of modern vegetational distribution along 1985), and responses of past vegetation to climatic change altitudinal and latitudinal gradients (Walter, during the Holocene and Pleistocene, demonstrate that plant distribution tracks climate and that plants respond to climatic changes rapidly (see reviews by Post, 2003; modern and past biomes may simply reflect. the Parmesan, 2006). The boundaries of position and boundaries of major atmospheric circu- lation cells (Ziegler et al., 2003). In addition, there is evidence from the deep past that megadisturbances, such as the massive Cretaceous—Tertiary boundary bolide impact, had their greatest effects by changing climatic conditions, even if just for the short term and theoretical reasons to suspect that rapid changes in vegetational composition and structure could repre- sent regime shifts, rapid changes from one stable compositional state to another once an environmental/ climatic threshold is crossed (Scheffer & Carpenter. 2003; Ives & Carpenter, 2007). The pre-Pleistocene record contains many well- studied examples of major changes in biotic compo- sition in response to catastrophic (effectively instan- taneous), rapid-directional, or prolonged-directional environmental change. The following are examples of some of the major events. The onset of the Carboniferous ice age in the Mississipplan is coincident with the development of widespread wetland floras and the demise of vegeta- ion dominated by the most primitive seed plants Pfefferkorn et al., 2000; Cleal & Thomas, 2005). Major reorganization of tropical wetland vegetation oa of Euramerica occurred at the middle-late Pennsyl- vanian (Moscovian-Kasimovian) boundary, after mil- lions of years of stability in the face of glacial- interglacial fluctuations (Pfefferkorn et al., 2000; DiMichele et al., 2002: Falcon-Lang, 2004), probably in response to rapid global warming (Phillips et al., 1974; Pfefferkom & Thomson, 1982; Phillips & Peppers, 1984; DiMichele € Phillips, 1996; Cleal & Thomas, 2005). This may have been caused or ex- acerbated by tectonically driven elevational changes in Central Europe. which would have liberated large amounts of buried carbon and reduced the areas of carbon burial in the tropics (Cleal & Thomas, 2005). Middle Pennsylvanian-type wetland vegetation per- sisted through the Permian in China (Wang & Chen, 2001), where it remained very wet due to proximity to moisture sources and atmospheric circulation patterns (Ziegler, 1990). Dominance of weedy vegetation and a period of chaotic dynamics follow the Euramerican regional extirpations (Peppers, 1979, 1996; Pfeffer- korn et al., 2008). Vegetational changes in the dominant biome in most of equatorial Pangea occurred across the Car- boniferous—Permian boundary (Kerp & Fichter, 1985; Broutin et al., 1990; DiMichele & Aronson, 1992) in response to long-term changes in temperature and rainfall patterns, again perhaps linked to changes in the extent of glacial ice and atmospheric CO» (Montafiez et al., 2007). Terrestrial and marine biotic changes near the Permian-Triassic boundary, in response to complex causation (Erwin, 1993), resulted in long pseudo- successional recovery lags and wholesale vegetational 1999), with evidence of persistence of many forms in extrabasinal areas (Kerp et al., 2006). restructuring (Looy et al., 178 Annals of the Missouri Botanical Garden Major vegetational restructuring and species turn- over occurred at high latitudes (and possibly globally) at the Triassic-Jurassic boundary (McElwain et al., 1999) apparently caused by a rapid 396—496 increase in atmospheric CO». The Cretaceous-Tertiary boundary had profound effects on terrestrial vegetation globally. Initially debated with regard to extent and rapidity of plant response (Hickey, 1981; Tschudy, 1984), it is now clearly documented to be rapid (Johnson, 1992) and global, but with variation dependent on local and regional climate and topography (Wilf & Johnson, 2004). Followed by what may have been a pseudo- succession similar to that of the early Triassic (Wolfe & Upchurch, 1986, 1987). like that former event, appears that lowlands may have been more homoge- -. — nous than was characteristic of extrabasinal areas, indicated by rare finds of distinct and diverse floras Johnson & Ellis, 2002). Significantly different inter- actions between plants and animals may have ensued, — e delaying the radiation of large mammals for millions of years (Wing & Tiffney, 1987). A short, extreme pulse of global warming at the aleo- Paleocene-Eocene boundary, the so-called — cene—Eocene thermal maximum (PETM), was driven by rapid global warming, which correlates with sharp changes in carbon isotopic composition of marine sediments (Zachos et al., 2005). Vegetational response was rapid and accompanied by large range shifts (Wilf, 2000; Wing et al., 2005). Finally, long-term cli- matic cooling occurred from the late Eocene through the early Oligocene (see papers in Terry € Evanoff, 2006) and was accompanied by vegetational changes that effectively tracked the changes in climate (e.g., Jaramillo et al., 2006). Although each of these changes is unique, com- prising different taxonomic compositions and different causes, either in kind or in degree, and occurring under different continental configurations and pre- vailing climates, there are unmistakable similarities. First, all confirm that vegetation responds rapidly to environmental changes, be those changes in extremes or seasonal distribution of temperature or rainfall. Long-term climatic changes demonstrate that plants, particularly if considered at the scale of biomes or track climate relatively large-scale ‘cies pools, closely and faithfully. Often, this is accompanied by intrabiomic changes in dominance-diversity while retaining much of the basic species-pool presence- absence composition. Rapid to catastrophic changes are almost always accompanied by dominance of opportunistic weedy vegetation during the initial response. In some instances, this vegetation gives rise to persistent dominant forms, though, in most cases, it is rapidly replaced by more K-selected growth forms. Pseudo-successions follow several of these events; however, there are indications from occurrences of rare intercalated or precocious floras that more complex and diverse vegetation continued to persist or exist refugially in extrabasinal areas. Thus, such pseudo-successions would be most in- dicative of long-lasting environmental disruption of t ment of most surviving lineages from surrounding — 1e lowland wetlands and inhibition of reestablish- areas, Finally, these events indicate that plants and vegetation can and do survive major environmental changes, but that recovery from these disruptions generally takes millions of years (e.g., plants typical of Mississippian seasonally dry habitats reappear in similar habitats during the late Pennsylvanian and 1991; Mamay, 1992]; plants of Mesozoic habitats appear precociously in the Paleozoic [DiMichele et al., 2001a; Kerp et al., 2006]; dominant conifers both occur precociously in the Permian [Mamay & Bateman, Pennsylvanian and survive the Permo—Triassic ex- tinction though returning after a long lag [Lyons & Darrah, 1989; Looy et al., 1999]. It must be reiterated that apparent stratigraphic lags in the occurrence of — axa are impacted by the regional and basinal pro- cesses operating on preservation potential of organic matter (see sequence stratigraphic discussion above; Gastaldo et al., 2005). PLANT-ANIMAL INTERACTIONS Our coverage of this expanding field of paleoecology will be abbreviated because most of the research focuses on animal paleobiology in which the plants are substrates for animal activities. There are important areas where the ecological attributes of the plants and animals are considered jointly, generally involving comparative phylogenetic patterns. Insect-mediated pollination biology of angiosperms (and potentially other groups of plants, though there is greater speculation there) has been an area where some debate has arisen due to potential conflicts between theory and data. Crepet (1979, 1985) sug- gested that the coevolution of angiosperms and faith- ful insect pollination during the Cretaceous was a powerful driving force both in the evolution of the flowering plants and of their pollinators. Many modern students of insect systematics and ecology have suggested that the expansion of the flowering plants substrate for the tremendous in fact created a radiation of the insects. Labandeira and Sepkoski (1993). in an analysis of the fossil history of insect diversity and timing of origination of functional feeding groups, found that insects radiated signifi- earlier than the during the cantly angiosperms, Triassic and Jurassic, and argued for the decoupling Volume 95, Number 1 DiMichele & Gastaldo 179 008 Plant Paleoecology in Deep Time f insect diversification from that of flowering plants. the later Carboniferous (Hotton et al., 1996; Sues & Grimaldi (1999) has challenged this interpretation of no relationship, considering conjointly the fossil record of the major groups of insect pollinators and phylogenies of these clades. He argues that there is strong coincidence in the timing of insect and ento- mophilous angiosperm radiations. This debate contin- ues (e.g.. Labandeira, 2002) Much of the literature on plant-arthropod interac- tions has focused on the trace-fossil record of arthropod activity. Most evidence from the earliest land plants suggests that arthropods were acting mainly as detritivores, with only limited evidence for attacks on living plants (Kevan et al, 1975; Labandeira, 1998). By the Carboniferous, there is considerably more evidence for arthropod predation on living plants. The majority of data, such as that 1977b; Labandeira et al.. 1997), suggests that detritivory remained the major from coprolites (Scott, way in which plant productivity entered the inverte- brate food chain (Scott € Taylor, 1983: Scott et al., 1985). There is some indication that insects were involved in medullosan pollination, based on coprolite data (Scott & Taylor, 1983). Labandeira (2006a reports that the earliest evidence of external foli- — age feeding occurs in the Late Mississippian, and that insect herbivory is concentrated particularly on medullosan pteridosperms in tropical floras, also noted by Scott et al. (1992), and on glossopterid seed plants in the Southern Hemisphere. Evidence of distinct forms of feeding, such as piercing and sucking, has also been documented in the Pennsyl- vanian (Labandeira & Phillips, 2002). During the Permian, in tropical regions, it appears that insect feeding was also concentrated preferentially on cer- tain clades (Beck & Labandeira, 1998; Labandeira & Allen, 2007). Labandeira (2006a) finds two further expansions of insect herbivory, one beginning in the early Mesozoic, and the other taking place in the mid- Early Cretaceous, coincident with the evolutionary radiation of flowering plants. Insects seem to have established modern food-web patterns by the end of the Carboniferous, with the full spectrum of insect functional feeding groups active by the end of the Triassic (Labandeira, 2006b) The record of vertebrate-plant interaction is more difficult to interpret. Not only might it involve herbiv- ory (Chin & Gill, 1996), which will be considerably more difficult to detect from feeding traces on plants than is the case for arthropods, and detritivory (Chin, 2001), but vertebrates, particularly those of large size, can considerably alter ecosystems through various kinds of disturbance (e.g., transformation of vegeta- tional structure by elephants; Wing & Buss, 1970). The earliest evidence of vertebrate herbivory is from Reisz, 1998). Interestingly, these early occurrences are not followed by a major radiation of vertebrate herbivores. Based on patterns of taxonomic diversity, there is a diversification lag paralleled by the evo- lution of herbivorous feeding strategies in insects — Labandeira, 2006a). Some of these early vertebrates had large bodies and small heads, suggesting that they may have been fermentative gut processors. The first vertebrate with indisputable evidence for oral pro- cessing of tough, high-fiber plant material is of Late Permian age from Russia (Rybezynski € Reisz, 2001). Prior to the Late Permian, vertebrate communities are dominated by what appear to be carnivorous forms. It is not until the Late Permian that trophic pyra- mids appear similar to those of modern communities (DiMichele & Hook, 1992; Sues & Reisz, 1998). Many Late Paleozoic seeds are large and have fleshy seed coats: Tiffney (2004) suggests that these fleshy seeds may have been consumed and dispersed by verte- brates, which may have confused them for carrion. However, he also concludes that there is, overall, relatively limited evidence for vertebrate dispersal of most Paleozoic seeds (but recall the disparity of preservational requirements between plant and verte- — rate remains). deep understanding of the effects of dinosaurs on Mesozoic plant communities has been elusive. It is clear that these large animals would have been major agents of mechanical disturbance and also consumed, probably in an ie dieron manner, a wide spectrum of plants (Coe et al., ; Tiffney, 1997; Chin & Kirkland, 1998; Chin, pd In addition, they may have been major agents in promoting the expansion of weedy angiosperms (Feild & Arens, 2005). Coevolu- with effects that can still be seen today on the chemical tion dinosaurs also may have had selective and morphological attributes of clades that reached their zenith in the Mesozoic, such as cycads. Many of these plants contain neurotoxins in their leaves, which may have deterred vertebrate herbivory, but not in the (edible) sarcotestas of their seeds, permitting seed herbivory and dispersal (Mustoe, 2007) Following the demise of the dinosaurs at the end of the Cretaceous, large areas of forest developed in the Paleocene, coincident with a delayed appearance of large-bodied mammalian vertebrates (Collinson & Hooker, 1991); Wing and Tiffney (1987) attributed this suppression of vertebrate body size increase to the vegetational structure. Tiffney (2004), in a review of the role of vertebrates in seed-plant dispersal through time, goes so far as to suggest that only with the [ol emise of the dinosaurs could the angiosperm-bird- mammal dynamic evolve, resulting in a fundamental change in ecological dynamics between the Mesozoic Annals of the 180 Missouri Botanical Garden and Cenozoic. We may presume that Tertiary ecosys- of populations and individuals that dictate and scale tems, especially those of the post-Paleocene, func- tioned much like those of the present with regard to the variety of plant-vertebrate interactions (see the reviews by Collinson & Hooker [1991], Wing [1998], and Tiffney [2004], which bring together much of the literature on vertebrate interaction with plants). ECOLOGICAL ASSEMBLY RULES IN SPACE AND TIME The concept of assembly rules in ecology has been debated with varying intensity and in various guises nearly throughout the history of the discipline. The very idea of ecological assembly implies determi- nism and predictability of outeome—it is the nature of this determinism and the particular nature of the expectations about outcome that lie at, or close to, the root. of this long-term controversy. From the time of the long-running Clementsian superorganism (Clements, 1916) versus Gleasonian species individ- 1926) to more modern incarnations of jon Gleason, debate over succession an — ualism climax communities, this discussion about the degree of interaction and coevolution in communities (e.g., Diamond, 1975; Conner € Simberloff, ized formulations (Weiher € Keddy, 1979), to recent more general- 1999), drawn em- the in- ference of rules of community assembly, pirically from time-space patterns, persists despite 1990). Gene- ral skepticism is drawn from things such as the re- general skepticism (e.g.. Foster et al., sponses of plants to the retreat of ice following the ast glaciation (e.g., Overpeck et al., 1992; Jackson, 2000), or studies of short- and long-term, apparently stochastic, changes in the composition of ecological 1986). Hubbell’s a general theory of ecological communities (Hubbell & Foster, (2001) formulation of dynamies, in which species are treated. neutrally as identical. elementary particles, changes the terms of the debate to a great degree, in that deterministic Fur- © processes may lead to a variety of outcomes. thermore, the explicit consideration of historical con- straints on modern ecological patterns, via inclusion of phylogenetics and fossil history (Webb et al., 2006), into the explanatory framework, impinges on the matter of assembly because past events may have impacts far into the future, a basie tenet of assembly- rules considerations. In deep-time paleontology., there has been little explicit consideration of the matter of ecological assembly rules, primarily because of the need for a sequence of fossil deposits over a specific time interval and covering a spatial extent that would of sequences and combination patterns to be rigorously ff species addition permit evaluation patterns evaluated statistically. However, it is the interactions upward into larger ecological patterns (e.g., Hubbell, 2001): surely, such interactions are what also underlie the dynamics of evolution. Thus, it should be expected of vegetation over long that. evolutionary. assembly periods of time should be under the same basic controls as assembly in space over short periods of time. The most general models for controls on community assembly are summarized by Belyea and Lancaster (1999) and Weiher and Keddy (1999). Using somewhat different jargon, these authors pre- sent basically the same hierarchical model, consist- ing of three components. The model also applies to evolutionary assembly over long spans of time as it does to spatial assembly on post-glacial time frames. The following factors control the composition of any site-specific assemblage. (1) Can a species get to the site? In other words, is it a member of the regional species pool, or *metacommunity" (sensu, Holyoak et al., 2005)? (2) Once on a site, can the species survive under the local physical conditions? Does it have the physiological and structural capabilities to exist under he soil edaphic conditions, the local aspect, and the light regime of the site? (3) If the first two criteria are — met, can the representatives of the species DUM QN representatives of other species already on the site, those that subsumes and may largely be a consequence of the effects of incumbency, which are related to population size, likelihood of expropriating resources, likelihood arrive later? To a certain degree, diis of colonizing unutilized or underutilized resources (e.g., Gilinsky € Bambach, 1987), and fecundity. The most explicit of a evolutionary assembly of ecological systems is that of Valentine (1980). This heuristic model describes an adaptive zone as a checkerboard, or a pattern of statement model for Each square represents an tesserae. through. time. ecological niche. The initial population to establish within this landscape finds little or no competition for resources and can survive even if only marginally able to live on the site. Given that most speciation will uce descendants similar to the ancestor, ecolog- [om pro ical space is gradually filled by new species around the resource space of the initial occupant. Coincident with small-scale evolution are rare larger-scale de- partures in ancestor-descendant similarity, most of which produce nonviable offspring, but some of which will produce forms that can survive if they can reach survive in with low or no Gradual filling of ecological resource and resource spaces competition. space then begins around these new resource nodes. Eventually, as ecological resource space fills, large- scale departures decline in frequency because their chances of survival decline. This is due to the effects of incumbency within the overall adaptive landscape Volume 95, Number 1 2008 DiMichele & Gastaldo Plant Paleoecology in Deep Time Walker and Valentine (1984) later criticized this model because of the supposed lack of niche saturation in ecospace (see Patzkowsky «€ Holland [2003] for a similar paleontological perspective), a matter widely addressed in the neoecological litera- ture. In addition, the static view of niche space which niche 1973), is as defined by species rather than as ee nor comports neither with the way in generally is conceived (Whittaker et al., that with the similar, expanded view, known as niche construction (Odling-Smee et al., 1996), which argues that organisms, by their actions and interactions, create and expand niche/resource space. Nonetheless, the Valentine (1980) model, though in need of modi- fication, is still a strong basic framework for uniting evolutionary models of ecological assembly with mainly spatial neoecological models. DiMichele et al. (2001b, 2005) have applied some aspects of these models to the examination of ecological assembly patterns in Paleozoic plant communities. An entirely different way of examining ecological assembly and the rules by which plant assemblages may be structured is emerging from the study of plant allometry. In short, beginning with metabolic scaling theory (West et al., 1997; 2004), a straightforward set of linkages has been developed all the way to spatial-density aspects of populations and mixed-species stands (Enquist et al., 2002, 2003, 2007; Niklas et al., 2003; Niklas, 2006). In the model from West et a 7), the metabolic rate of an organism scales to its body mass raised to the 3/4 power. These relationships have been found to apply rown et al., among plant size, mass, metabolic rate, etc., to the scaling of plant-body size within populations, and from that to entire multispecies plant communi- ties (Niklas & Enquist, 2001; Niklas et al., 2003). In its most fully elaborated form (Enquist et al., 2007), these scaling laws permit basic physiological attri- butes of organisms to be linked directly to the eco- logical structure of plant communities in a “taxon- free” manner, given that the basic rules appear to apply to all major groups of eukaryotic photoauto- trophs (Niklas, 2006). The strength of this approach, to quote Enquist et al. (2007) “is that it reduces much of the complexity of organisms and ecosystems to simple, but universally applicable, physical and chemical principles.” In certain ways, this approach meshes with Hubbell’s (2001) neutral theory, in that the particulars of species composition do not really matter to ecosystem a Where the fossil record permits (e.g., of the examples given in Enquist et al. [2007). Pus predictions about organism size and stand densities can be tested directly, often in systems for which there is no modern equiv- alent. Conversely, to the extent that these scaling generalities are proven to hold (also consider the leaf physiological rate spectrum, discussed above), they may be useful to allow estimation of attributes of ancient systems. The degree to which this approach becomes central to ecology remains to be seen. Literature Cited Ackerley, D. 1999. Comparative plant ecology and the role of p DUCERE e in comparative data. Evol. Ecol. 5-909 Mco T teleconnections in . E. Scheckler. the evolution of land plants, weathe ring proe esses, and marine 3. 353: 113-130. 1998. Terrestrial-marine Devonian: Links between the anoxic events. Philos. Trans., Ser. E Allen, J. P. & . Gastaldo. 2006. Sedimentology and nd of he Bally to Middle Devonian plant-bearing a of the Trout Valley Formation, Maine. Pp. 57-78 in W . A. DiMichele & S. Greb (editors), Wetlands Through Tine Geological Society of America Special Publications le ews, H. N. & E . M. Kern. 1947. The Idaho ar and associated adi plants. Ann. Missouri Bot. Gard. : 119-183. & W. H. Murdy. 1958. e and oed in arborescent lycopods. Amer. J. B : 002—560. . E. Kasper, W. H loses a E Gensel & Y 1977. Early dud flora of the Trout V is Formation of northern Maine. Rev. Palaeobot. Palynol. 23: 255-205. Chaloner. Archer, A. W. & C. G. Maples. 1987. Analysis of binary similarity a Effects of sample sizes upon dis- tributions. o 2: 609-617. lC, . Reconstructing Community Dynamics in a o Floodplain Forest. Ph.D. Dissertation, Harvard University, Boston. Ash, S. 1999. An Upper Triassic upland flora from north- central New Mexico, U.S.A. Rev. Palaeobot. Palynol. 105: 83-199 Axelrod, D. I. 1952. A theory of angiosperm evolution. Evolution 6: 29-60. . 1972. Edaphic aridity as a factor in angiosperm Naturalist 106: 311—320. Bailey, I. & E. Sinnott. 1915. A botanical index of Cre- taceous and Tertiary climates. Science 41: 831—834. ev nos Amer. 916. The climatic distribution of certain nius of wc 2 Amer. J. Bot. 3: 24—39. Baker, . & W. Michele. 1997. Biomass allocation in mu p An coal-swamp plants. Palaios 12: 1-132. Bateman, R. M. 199 the growth architecture of fossil rhizomorphic lycopsids: Scenarios constructed on cladistic foundations. Biol. Rex 69: 527-597 4. Evolutionary-developmental change in . DiMichele. 1991. las en. verbaceous segregate of epider context pees ary 1 lycopsid growth architecture. Syst. Bo 76: (Pennsylvanian) Phylogenetic changes i Willard. 1992. Experimental of ‘anatomically: -pre "served arborescent lycopsids from the Carboniferous of Euramerica: An essay on paleobotanical phylogentics. Ann. Missouri Bot. Gard. 79: 500-559. kj A cladistic analysis 182 Annals of th Missouri Botanical Garden , P. R. Crane, ce DiMichele, T rick, N. Berthelin, M., J. Broutin, H. Kerp, S. Crasquin-Soleau, J.-P. Rowe, T. Speck & 34 . Stein. 1998. a p ie Platel $ E Roger. 2003. a Oman Gharif mixed land plants: Phylogeny, a de les of paleoflora: A useful tool for testing Permian Pangea the o terrestrial radiation. Annual Rev. Ecol. Syst. 2€ Beck; A. L 1998. folivory on a yes rid- EF riparian flora from Labandeira. ‘arly Permian insect north-centra Palaeogeogr. Palaeoclimatol. Pa- laeoecol. je TM Beck, C. B. 7. Tetraxylopteris schmidtii gen. el sp. nov., a ps M die p from the Devonian of New York. P or. J. Bot. 50-367. of the u Palaeontogr. Becker, H. F. 1972. The Mevel R anch flora / River southwestern Montana. 141B: 1-61. Beerbower, J. R. 1961. Origin of cyclothems of the Dunkard group (upper Pennsylvanian-lower Permian) in dirus Bull. 7 upper ed vania, West Virginia, and Ohio. Geol. Soc. Amer. 1029-1050. 1964. Cyclothems and cyclic depositional mecha- D. F. nisms in alluvial plain sedimentation. /n Merriam (editor). uu on Cyclic Sedimentation. Kansas eol. Surv. Bull. 1 31-42. Beerling, D. J. 2X e. atmospheric CO» levels during the Permo- M A E aciation inferred from fos- sil is ds. Proc. Natl. Acad. Sci. U.S.A. 90: 12567- 1257 ———- e C. P. Osborne. 2002. Physiological ecology of are Said forests in a high CO) environment. Ann. Bot. 329-339. ja D. a ^ yer. aia a a CO» signal from fossil e omab, i "Ww fps 53: —397 ———— & ——— 2002). E P as indic 'ators of the as 2 jn d cycle. Ann. Rev. Earth Planet. 5 7-3: [om deest 2001. Vegetation and the Modelling the First 400 Million Cam om ne. | = > gı A Terrestrial Carbon a Years. Rc: Univ. McElwain & d "1998. Stomatal re sponses of the ‘living fossil’ Ginkgo biloba L. to changes in do CO» concentrations. J. Exp. Bot. 49: 1603-1607 Behrensmeyer, A. K. € R. W. Press, Hook. 1992, mental contexts i ML RE wu in the terrestrial Paleoenviron- ossil recort , W. Y DE 1e £s W ing (editors): Terrestrial mA ni 2 — in A. K. Be I nsmeyer, R. Puts; H. Es s& Jue Ecosystems Time. ———,S. M. Kidwell & R. A. Castaldo, 2000. Taphonomy and paleobiology. Paleobiology 26: -147 Belyea, L. E Lancaster. 1999, la pe le 's within a iru ec ology. Oikos 86: 402—4 1€ gton, J. B. 2003. comparative ROM sof Bennin Transcending vun hiness in the paleocommunities: A test case New Jersey. Palaios 18: a 2m the Upper Cretaceous o -33. anteno- García, C. 2006. Sedimentology and rin edo of an Eocene -Oligocen 1 LS. R. e os-Ferriz, E. Ce — Be ña Campsel, Arenas- "ug & L. Fe scm PR dac ustrine arid system. S. Mexico Sediment. Geol. 191: 227-254. Berner, "i 1991. Phanerozoic time. 998. > Amer. J. $ : 339—316. carbon eye s and CO» over Phanerozoic . B. 353 "The role of land plants. Philos. Trans., Ser. B. 353: lime: 15-8 A model for atmospheric CO, over ] . 291: 339-37 reconstructions. Palaeogeogr. T t) ol 196: 85 . C. Vozenin-Serra & J. Broutin. 2004. Phytogeo- graphic and climatic implications of Permian wood discoveries in Oman. Palae MOLIS 208B: * -112. Boom, A., | E LE Marchant, H. Hooghiemstra & J. S )2. COs- and temperature- cola altitudi- Cy-dominated grasslands allow reconstruction of palae Borar pCO». Palaeogeogr. Palaeoc pe Palaeoecol. 177: 151-168. Bowen, G. J. & soil organic ea an id ( ). J. Beerling. 2 An integrated model for 20»: Implications for paleosol a pCO: paleobarometry. Global Biogeochem. e : 1-12 a 2005. Patterns of segregation and convergence E emilia of fern and seed plant leaf morphologies. e Paleobioloss 31: 117-140. . The role of size for the potential phy na of "m T like vasc ular plant. relatives. Geol. Amer n 38(7 Knoll. 2002. Buchulton of developmen- tal "heben x the multiple independent origins o ca Abstr. A leaves in Paleozoic vascular plants. Paleobiology 28: 70-100. ——, G. D. Cody, M. L. Fogel, R. M. Hazen, C. M. O'D. Ale ils & A Pari 2003. Chemical evidence for cell wall (eni oanh and the evolution of tracheids in Sel. ] 691—702 Karly Devonian plants. Int. — — —, C. Hotton, M. L. Fogel, C. D. Cody, R. M. Hazen, A. 2007. a glant fungus. Hueber. Devonian landscape H. Knoll & F. M. heterogeneity recorded by Geology 35: 99—402. 40 —— M. A. Zwieniecki, cody, C. J. Jacobse W s A. H. Knoll & N. E S m Evolution o E xylem O and hydrogel O regulatio Proc. Natl. Acad. im U.S.A. 101: 58. Brack-Hanes, 5. D. | epilodendise ean cones. Bol. & B. A. Thomas. 1983. dostrobus Brongniart. Bot. J. Linn. Soc Braun-Blanquet, J. 1932. Plant Sociology: The 2 of Plant 978. On the ngage s of the . 139: 140-146 A re-examination of Lepi- 6: 125-13: Communities. Stechert-Hafner Service Agency y York. Bray, J. R. & : oer 1964. Litter Mi in pu "sls of the world. 1157 in J. p. Cragg (editor), Advances 'ademic p ress, New in sd P Vol. 2. Ae Brett, C. E., L. C. Ivany € K. M. Schopf. 1996. Coordinated Palaeo- stasis: An overview. Palaeogeogr. Palaeoclimatol. ecol. 127: 1-20. —, A. J. W. Hendy, A. J. Bartholomew, J. R. Bonelli & POL 2007. biotas to sea-level fluctuations aa ement and the process of MER tracking. Palaios 22: —244. Soult, J. &M.I floras during the McLaughlin. Danse of shallow marine A review of faunal re- 3erthelin. 2005. Dynamic of settlement of mixed ng Permian in the Peri-Tethyan domain: Paleogeographic and paleoclimatic signific vance. Bull. New Mexico Mus. Nat. Hist. Sci. 30: 24— ———, J. Doubinger, G. Farjanel, F. Freytet & H. Kerp. 1990. Ler renouvellement des flores au passage Carboni- fère Perm Approches stratigraphiques. biologiques, Ee E Compt. Rend. Acad. Sci. Paris 321: 1503-1509 Volume 95, Number 1 8 DiMichele & Gastaldo 183 Plant Paleoecology in Deep Time — J. Roger, ^ P: uc L. Angiolini, A. Bau Buc he er, J. Marco . Al Hasmi. 1995. The nes Pangea. pala implications of new paleontolog- ical discoveries Oman (Arabian Peninsula). Compt. Rend. Acad. Sci. Paris, Sér. lla 321: 1069-1086. , H. Aassoumi, M. El Wartiti, P. Freytet, H. Kerp, C. Quesada & N. Toutin-Morin. 1998. The Permian basins of Tiddas, Bou Achouch and Khenifra (central Morocco). Biostratigraphic and ad du Mo in a J.F. Gini. P, Allen, V. M. Sava 'sl. . Toward a wetabulie theory of ar Ecology 85: 1771-1789 Bull, W. B. 1991. Geomorphic responses to climatic change. Oxford Univ. Press, New York. Burnham, R. J. 1993. Diversification and stable distribution with respect to temperature in early ora Ulmoideae of western North America. Abstr. Bot. Soc. Amer. 68. . 1986. Foliar ir dA aa of the iary i M North Ulmoi- deae (imac eae) from the Early T America. Med. d 201B: 135 Relationships between ms vegetation and b ui. in a paratropical oo for paleobotany. Rev. Palaeobot. Palynol. 58: 5-32. . 1993a. enc richness in ifie plant fossil -— e 8: 376—384. 1993b. Time resolution in terrestrial macrofloras: Guidelines from o accumulations. Pp. 57-78 in S. Kidwell & A. V E (editors), Taphonomic Approaches to Time Resolution in Fossil Assemblages, Short ee in Paleontology N 4. Paleoecological and ‘floristic heterogeneity in the a pue record —and analysis based on the Eocene of Washington. U.S. Geol. Surv. Bull. 2085B: B1-B36. . 1997. Stand characteristics and leaf litter species composition of a dry forest hectare in Santa Rosa National Park, Costa Rica. Biotropica 29: 384— & am. 1999, The histo of ra Ann. Missour vegetation: New developments and status. —589. Bot. Gard. 86: 5 & R. puse 1986. Forest litter preserved by bon aoti at El Chichon, Mexico: A potentially urate record of the pre- eruption n. Palaios 1: 158-1 5] Wing & G. C. Parker. 1992, The reflection of decidida forest Com das hi iod litter: Implications the fossil record. for autochthonous litter Paleobiology 18: 30-49. . A. Pitman, K. R. Johnson & P. Wilf. 2001. Habit related error in estimating temperatures from leaf margins in humid tropical forest. Amer. J. Bot. 88: 1096- e R. Johnson & B. Ellis. 2005. Modern tropical M ae Does high biodiversity affect paleocli- terpretat tions? Palaios 20: 439-451. Buzas, M tulver. 1999, Understanding regional species diversity ions the log series distribution of occurrences. Diversity Distrib. 8: 187— e un J. M. R. Gibling, A. C. Scott, S. J. s & m 20006. A fossil lycopsid ues succession in m classic Joggins section of Nova Scotia: Paleoecology of a disturbance-prone Pennsylvanian wetland. Pp. 169-194 in S. J. Greb & W. A. DiMichele (editors), Wetlands Thr ough xm Geological Society of America Special Papers 3 matic in 2 Cantor, B. M., B. V. Aigler, D. W. Pace, S. B. Reid, C. Y. Thompson & h A. Castaldo, 2006. nue and interspecific variation in stomatal proxies for Quercus and Nyssa in the subtropical southeastern USA. Geol. Soc. Amer. Abstr. Prog. 38(7): 487. B , D. J. 1996. Fern thickets from the Cretaceous of Alexander Island, Antarctica containing Alamatus ei Douglas and Aculea acicularis sp. nov. Cretaceous Res. 169-182 A. N. Drinnan & J. A. Webb. 1995. Late Triassic plant fossils from the Prince Charles Mountains, East Antarctica . Antarc. Sci. 7: 51-62 Cavender-Bares, J., A. Keen & B. Miles. 2006. Phylogenetic structure of ! andis ine communities depends on taxo- nomic and spatial scale. Ecology 87(7)(suppl.): S109-S122. Cecil, C. B. & F. T. Dulong. 2003. m models for i ent supply in warm climates. Pp. 21-27 in C recil € N. T. Edgar ERR Climate Controls on ort SEPM = ial Pub. Vol. 77. , R. A. Harris, J. A. Cobb, H. G. Gluskoter & H. Nugroho. 2003. pcs on climate and sediment discharge in selected tropical rivers, Indonesia. Pp. 29-50 in C. B. Cecil & N. T. Edgar d oe Controls on Stratigraphy. SEPM Socal Pub. V Cerling, T. E. 1984. The stable o composition of modern is ed and its 2 UM to climate. Earth s . 71: 229-240. uo pu in fie atmosphere: Evidence from Pain and Mesozoic paleosols. Amer. J. Sci. 29 377-400. . 1999, Stable carbon isotopes in paleosol carbonates. 4 Thi 1 Simon-Coingon (editors), Palácoveatlistine- Palacosurfaces, and Related Continen- tal uri Blackwell, Cambridge. Chaloner, W. G. 1959. Continental ge Pp. 7-30 in M. L. Johnson, a dbenrombie E. Fogg (editors) New Biology, No. 29. Penguin Boake Balti imore. & G. Creber. 1990. Do Pe plants give a climatic m ee Soc. 147: 343 Lacey. 1973. ene peers of Late Paleozoic i. Pp. 271-289 in N. F. Hughes (editor Organisms and Continents icu d ime. Special Papers in Palaeontology No. 12. The PRO Society, ndon —— S. V. Meyen. 1973. Carboniferous and Permian floras of the northern continents. Pp. 169-186 in Hallam (editor, Atlas of Palaeobiogeography. Elsevier Scientific Publishers, Amsterdam. Chambers, I Drinnan € S. McLoughlin. 1998. Some morphologic al jedes of Wollemi pine (Wollemia nobilis: Araucariaceae) and their Br on to Creta- ceous plant fossils. Int. J. Pl. Sci. 159: m Chen, L.-Q., C.-S. Li, W. G. Chaloner, D. Pr. g, Q.-G. Sun M M. E. Collinson & P. L. Mitchell. 2001. yee the ee or the stomatal characters of extant and fossil oS o leaves : rd atmospheric CO» change. Amer. J. L. 88. 1309-1315. Chin K. 2007. 2i paleobiological implications of herbiv- ous dinosaur coprolites from the Upper Cretaceous Two Medicine Formato of Montana: Why eat wood? Palaios 22: 554-56 & B. E Gill. 1996. Dinosaurs, dung beetles, and conifers: Participants in a Cretaceous food web. Palaios 11: 280—285. & J. I. Kirkland. 1998. Probably herbivore coprolites from the Upper Jurassic Mygatt- ha oore Quarry, Western Colorado. Modern Geol. 23: 249-2 184 Annals of the Missouri Botanical Garden Chitaley, S. & I Pigg 1996. Clevelandodendron Cúneo, N. R. 1996. Permian Erona in Gondwana. ohioensis, gen. et sp. nov., a slender upright lycopsid from Palaeogeogr. Palaeocli matol. Palaeoecol. 104. the Late Devonian Clevel land Shale of Ohio. Amer. J. Bot , E. L. Taylor, T. N. m & M. ue pum. An in 83: 781-789. situ fossil forest from the upper Fremouw Formation Christie-Blick, N. W. sco s stratigraphy. Ann. Earth vg Sci. 2; Cleal, C. J. € B. 2005. Palaeozoic tropical rainforests and hei lr -— cb on e climates: Is the the key to the present? Neca a 3-31. R. M. James & E. Em 1999. Variation in hata density in the Late s gymnosperm )-185 & N. Sequence Rev. 451-478. Thomas. past frond t ovata. Palaios 14 Clements, F. R. 19 Development. of oe Carnegie Institution, Wash- . Plant Succession: An Analysis of the ington, D.C. Coe, M., D. L. Dik her. J. Farlow, D. M. Jarzen «€ D. . Dinosaurs and i land plants. Pp. 225-25 Sin . Chaloner & P. R. Crane (editors). The erms and Their Biological Consequenc- . M. Friis, W. Os of e es. Cambridge Univ. Press, Cambridge. ( rer M. E. 1984. Fossil Plants of the London Clay Flora ield Guide to ' os Fossils). Paleontological Association, 2001. Cainozoic ferns and their distribution. Brittonia 53: 173-235 2002. The ecology of Cainozoic ferns. Rev. Palae abot. Pab nol. 119: 51-68. & J. J. Hooker. 1991. between rpg and plam: eating mammals. | Ser. B. : 197 Fi —. 2 3. Framework for F & Fossil evidence of interactions Philos. Trans. Paleogene vegetation of Eur- TET feg] aunas. Pp. 41-83 in i a Wessels (editors), Distribution asia: J. W Reumer é and Migration of Texans Mammals in Eurasia. Deinsea Vol. 10 - & A. C. Scott. 1987. change through the geological record on a ls for coal- forming environments. Pp. 67 ott (editor). Coal and Coal-bearing as hecent "M “es. 2 E ical Society of London Special Publications Vol. & D. Simberloff. 1979. T Chance or competition? Implications of vegetational in A. — Conner, E. F. ie assembly of M Ecology 60: communities: 1132-1140. Sebe 9. fossil history of Ailanthus (Simaroubaceae). | Manchester. 2004. P ps a and JPL :S Crepet, W. L. 1974. Investigations of North American The Ta biology of Cycadeoidea. Palae ontographic a 148B: 144—169 1979. s A paleontological per- 29: 102 ms 1985. Advanced (constant) nsecl spective. BioScience 185 insect pollination my ‘eli and implications vis-a- . Missouri Bot. mec Toews Pattern o vis angiospe rm diversity. Gard. 607 (—€ Gandolfo. Fossil & M. A. 2004. The age of major angiosperm Nixon e ids nce and phylogeny: ; based on mesofossil and macrofossil evidence from Amer. J. Bot 91: 1666-1682. clades Cretaceous deposits. Crookall, R. Greal ed d section]. Mora, Geol. Surv. Great Britain, Palaeontol. 4(7) 1-100 Cross, A. T. & R. E. changes in 1982. are of short-term Some Taggart. sequential fossil plant assemblages: considerations tee d on a Miocene flora of the northwe "sl 34. United States. Ann. Missouri Bot. Gard. 69: 679- 1976. Fossil plants of the Carboniferous rocks of of Antarctica: es oak ntal setting and paleoclimate analysis. Palaeogeogr. Palaeoclimatol. Pa- Miss s 239-26 Dahms, D., V. Holliday & P. W. Sr 1998. Paleosols and iNOS forests. Science 279 Damuth, J. D., D. up J. A. ds . R. Potts, R. Stucky ). B. Weishampel. 1992. Tona is calegorization of animal a a s. Pp. 183- E 3 in A. K. Behrensmeyer, J. Damuth, DiMic Ne le. R. Potts, H.-D. £ bouem] p Through Time. Univ. Davies, D. 1929. Correlation and measures in East roman 217: 91-153. E Wing (edi MS of Chicago Press Sues & S. L. . Chicago. palaeontology of the coal B. Philos. Trans., Ser. E M K. 5. wem Wing. 1998. Sedimentological, taphonomic, and dum ic aspects of Eocene swamp deposits (W m UN NA Bighorn Basin. Wyoming). Palaios 13: 28—40. DeConto. R. | a W. W. Hay, ; 1999. Late Cretaceous climate and vegetation interactions: The cold continental Pp. 391—406 in E. Barrera & C. Thompson & J. Bergengren. interior paradox. Johnson (editors), Evolution of the Cretaceous Ocean/Climate System. Ge ologic al Society of America Special Papers 332. Demko, R. A. Gastaldo. 1996. Kustatie and aulocyclic influe ances on n of the Lower Pennsyl- vanian Mary po Coal zone, Warrior Basin, Alabama. Int. . Goal Geol. 31: 3-19. . F. n & J. T. I in inc uen valleys: Implications for interpre ting 119-112 Parrish. 1998. Plant taphonomy Le COC li- male from fossil plants. Geology 26: 1 Es] . ———, B. 2004. ET paleoclimatic and stratigraphic implic ations of paleosols and fluvial/overbank architecture in the Morrison Forma- Lion = 2 r Jurassic ). Western Interior, USA. Sediment. Currie & . Nicoll. Geol. E 5-135. Din ] M. 15." ‘he asse rs of species Communities. Pp. 342-44 ie | M. L. Cody € J. M. Diamond Ecology. and Evolution of de Harvard Univ. (editors), Press, Cambridge. Diessel, C. F. K. 1986. On the correlation between coal faci and ao environments. Advances in the Study of de Sydne : we 20th Symp. Univ. Newcastle, pp. 19-22 Dilcher, D. L. € Crane. 1984. angiosperm from the Cenomanian of western interior of 38: Archaeanthus: An early — a] North i Ann. Missouri Bot. Gard. 71: 351-368: DiMichele, W. A. 1979. Arborescent SyEOpods a Pennsyl- vanian age eae Lepidophloios. Pal graj 171E 57-77. 1980. Paralycopodites Morey & Morey, from the Carboniferous of reassessment of the and “Le Pap ndron^ generic affinities t brevifolium Williamson. Amer. J. Bot. 67: 1466-1476 Aronson. 1992. The Pe e Permian transition: A terrestrial analogue to the is. Evolution 46: 807-824. 1996. Plant paleoecology and € . vegetational onshore-offshore hypothes & R. M. Bateman. ‘ence: Two examples from the Paleozoic. anon ot. Palynol. 90: E 247 UP. J. DeMaris. 1987. Pennsylvanian-age Lepi Ui i forest: o ionary infer Res cture and dynamies of a Colonizers of a (No. 6) Coal of disturbed swamp habitat in the Herrin 146-157 llinois. Palaios 2: Volume 95, Number 1 2008 DiMichele 8 Gastaldo Plant Paleoecology in Deep Time 185 & R. W. Hook. 1992. Paleozoic terrestrial tems. Pp. 205-325 in A. P. J. Damuth, W. A. DiMichele, R. Potts, H.-D. Sue ors), Terrestrial Lo Through Univ. of Chicago Press, Chic & W. 1989. "Small -scale spatial het- erogeneity in Pennsylvanian-age vegetation from the roof- shale of the Springfield Coal. Palaios 4: 276-280. 1985. Arborescent lycopod reproduction p e ‘ology in a coal-swamp environ- ment of late Middle Pennsylvanian age (Herin apal, Illinois). Rev. pn obot. in jl. 44: & 988. Eo of the Middle Pennsylvanian-age : Hes coal swamp near a contempo- er system, the Walshville Paleochannel. Rev. Palaeobot. Palynol. 56: 151—176 . 1994. Paleobotanical and paleoecological f Time. —— — & J. Nelson. —— —— & T. L. Phillips. [om raneous riv constraints on models o peat formation in the Late Car Te Euramerica. Palaeoclimatol. Palaeogeog. : 39—90. Palaeoec :0 jl. & = & 996. Climate change, plant extinction and vegetational recovery during Middle-Late Pennsylvanian transition: The case of tropical peat-forming environments in North America. Pp. 201-221 in M. B. Hart (editor), Biotic ' from Mass e ction pen ee al Society of London Special Publications Vol. 2002. Rev - Palaeobot. Pulvaol; 11 & 5 Wing. 1988. Methods and Applications A Plant Paleoecology. Paleontological Society Special P Recovery -oÈ 5 S — pis A Paleozoic ferns. 43-159. lic dr: No. 3. . L. Phillips € G. E. McBrinn. 1991. Quantitative EE sis Mn floras a Pennsylvanian, 390—409 paleoec olog gyo of the Se ecor coal and roof-sha le Oklahoma). Palaios 6: ; Eble & D. S. Chaney. 1996. Dess forest above the Mahoning coal D Group, Upper Pe n in eastern Ohio, U.S.A. I J.C : 249 A drowned ar Gec l. 31 27 . H. Mamay, D. T 2001a. An Early P and Mesozoic gd from Paleontol. 75: 449-460. ,W.E Pied & R. M. Bateman. 2001b. Ecological sorting during the Paleozoic 2x on of oa ut ant classes. Pp. 285—335 in W. pe & D. J. Pun (edited, Evolutionary Ate Columbia Univ. Press, d York oe y, R. W. mian flora wi north-central Hook & J. W. ith Late Permian Texas. o & W. J. Nelson. 2002. Place vs. time onal persistence: A comparison of four tropical paleomires e the Illinois Basin at the hei jx of the Int. J. Coal Geol. 50: 43-72. T. D. TUM vegela Pennsylvanian ice age. A. K. Behrensmeyer, PACA C. C: Pandolfi, 5. L. Wing & R. Bobe. J. Long: term stasis in ecological assemblages: Evi- Labandeira 2004. dence from the fossil record. Ann. Rev. Ecol. Evol. Syst. 35: 28 a Hook, W. J. Nelson & D. S. Chaney. 2004. Middle Formation (Pease River Group, Leonardian-Guadalupian West Texas. J. Paleontol. 78: An a Permian flora from the Blaine Series) of King County, 765-782. Gastaldo & H. W. bnduenit: partitioning in the Late P Pfefferkorn. 2005. Plant Carboniferous and arly Permian and its implications for ecosystem assembly. Proc. California Acad. Sci. 36(suppl. 1): 32-49. , H. J. Falcon-Lang, W. J. Nelson, S. D. Elrick & P. R. Ames: 2007. Ecological gradients within a Pennsylva- nian mire forest. Geology 35: 415—418. lobor: T. 1973. ari T sense in biology except Biol. Teacher 35: 125-129. Foreman. 1992. Paleopedology and implications of paleosols, Juniata Formation, Southern Appalachians. J. Sediment. Res. 62: 71-83. I in the light of evolution. Am Driese, S. G. & J paleoclimatic Late Ordovician vertic E. Cotter & J. L. Foreman. 1992. pe stable isotope chemistry of Late Silurian vertic paleosols, P Formation, central Pennsylvania. A Sediment. Res. 62: 825-841. M. Elick. 2000. me paleosol record of increasing 2 velo and depth of rooting and changes in atmospheric pCO, in the Siluro-Devonian. . 47-61 in R. A. Gastaldo & W. A. Ex M Phase ozoic Terres nce Society Papers, Vol. 6 . C. Nordt, W. Lynn, C. A. Siles; C. I. Mora & L. P. Milans, 2005. Distinguishing climate in the soil record using chemical trends in a Vertisol climosequence from the Texas Prairie, and application to interpreting Neu o in the Appalachian basin. J. Sediment. 75: : 340 —35: Res. 7 Paleopedology trial Ecosystems, A Short Course. Coastal P nivasan, C a & F. W. Stapor. 1994. ries of rae Monteagle Limestone preceding development of a Lower Chesterian transgres- sive systems tract and sequence boundary, middle Tennessee and northern Alabama. Geol. Soc. Amer. Bull. 106: m Eble, C Greb, D. A. Williams & J. C. Hower. 19993. n e on the ur a petrography and geo- chemistry of the Western Kentucky No. 4 coal bed. Int. J. Co: 3 140. al Geol. .]. € Hower € W. M. Andrews Jr. Viae Compositional variations in the Fire Clay coal bed o eastern Kentucky: Geochemistry, petrog c. _. and paleoecology. K ntuc el bel Survey Report of Devonian an le analysis based on sils. Pp. 233-242 in W. S. errow & C. (editors), Palaeozoic E i. and orcos Geological Society Memoir Vol. 12. Geological Society, London. , K. L. Davies & L. Axe. strand in e early land plant Cooksonia. Nature 357: 683-685 1992. A vascular conducting Eggert, D. The ontogeny of ge arborescent D Palacontograpincs 108B: & T ced E Carboniferous 43-92, d . Stable isotopes. a in H. A. Mooney & J. G. Canadell Cas b ou of Global Environmental Change, Vol. 2 John Wiley & Sons, Chichester. Ehret, D. L. & T. L. Phillips. me Psaronius root systems— m and development. Palaeontographica 161B: 1 ON . J. Cantrill & J. E. Francis. 2004. A late due :eous mesofossil aen aie from Table Nunatak, Antarctica: Lycopods, ferns, and vegetative structures of mq and angiosperms. Cretaceous Res. 25: 211- Elick, J. M., S. G. Driese & C. I. Mora. 1998. Very large m a root traces from the Early t to 5 Middle Devonian: mplications for early terrestrial ecosystems and atmo- ce pCO». Geology 26: 143-146. 186 Annals of the Missouri Botanical Garden Ellis, B., K. R. Johnson & R. E. arly Paleocene px : Dunn. 2003. Evidence for an from Golorado: dn Mountain Geol. 38: 73— Enquist, B. J., J. Sanderson & M. D. = eiser. 2002. Lem macrose ue patterns in n Science 295: 1835-1830. VE Aller E D. xman, : UM 2003. Se NUM metabolism from 423: 639 Castle Rock tz = a) ——— . Economo, T. E. Huxr benda &I E. organisms lo poi. Nature 4 B. H. Tiffney & J. Niklas. 2007. sc alim: and i Pub. dynamics of plan and diversity: To Mea t size, form ward a synthesis of ec dn evolution and puente E J. PL Sei. 168: 7 órwin, D. H. Erwin, D. The Great a d Crisis. Columbia Univ. Press, a or Escapa, I. & R. Cúneo. 2005. A new equisetalean plant from the early Permian of Patagoia, Argentina. Rev. Palaeobo Palynol. 137: 1-14. Falcon-Lang, H. J. 2004. Pennsylvanian tropical rain forests Hop to glacial-interglacial rhythms. Geology 32: 689 aoi 5a. Global climate analysis of growth rings in —3 nd its implications for deep-time paleoclimate —444. 31: 434. 05b. Intra-tree variability in wood anatomy, and inr Pale E its in ena ations for fossil wood systematics and palaeocli- matic studies. e ontology 48: 171—183. . 2006. Latest Mid-Pennsylvanian tree-fern forests in retrograding ner plain deposits, Sydney Mines Forma- tion, J. Geol. 163: 81-93. R. Bashforth. 2004. Pennsylvanian uplands logy 32: Nova Scotia, Canada. X A. were forested by giant cordaitalean trees. Geo 417-420. 2005. Morphology, anatomy, and upland ecology of large cordaitalean trees from the Middle Pennsylvanian of Newfoundland. Rev. Palaeobot. Palynol. 135: 223-243. & D. J. Cantrill. 2002. Terrestrial paleoecology of the (Early , South Shetlands Islands, Antarctica: Á rec ‘ord of polar S s 17: 491—506. 2000. ipland ha of some Late cretaceous Aptian) Cerro Negro Formation ina e arc environment. Pa A. C. Zo mex d puc an trees from Nova Scotia and England. Palaeogeogr. Palaeoclimatol. Palaeoecol. 156: 22 ix . J. Cantrill & G. J. Nichols. 2001. Biodiversity and e ecology of a ae Cretaceous, x latitude floodplain, Ale Sander Antarctica. J. oc as 158: 709-7: Feild, S. Nc 2005. Form, envionments of the early angiosperms: Island, function. and Merging extant phyloge ny and ecophysiology with fossils. New Phytol. 166: 383—408. M. hans & N. M. Holbrook. 2000. W biela eae evolution: ue e cophy siological perspective. Ann. e. Bot. pun : 323-334. . Brodribb, po & N. M. Holbrook. 2001. Ace a of ~ bu. photosynthetic light use, and xylem hydrau to light Amborella trichopoda (Amborellac ud - Pl. Sci. 162: 999-1008. C. Aren . Dawson. 2003a. The e Emerging perspectives fror Plant Sci. 164 suppl: Zwieniecki ecology of angiosperms: extant basal lineages. Int. J. 5129-5142. P. J. Franks & T. L. Sage. 2003b. Ecophysiological adaptation in the shade basal angiosperm Austrobaileya J. 4: scande ens (Austr obaileyac eae). Int. . Sel. —324. s, J. A. Doyle, T. E. Dark and disturl bed: Dawson & M. —, N. me E A new image i -107 angiosperm ec ERN dn 30: l. . T. Frank, L. Birgenheier, S. Thomas, M. Rygel € A. Tones 2005. a Permian s rec A of eastern Abstr. Progr. Es A IB No Australia. Geol. Soc. 2005. —— d & W. J. e Uvdahedal la al teeth of Chloranthus japonicus (Chlor- T. L. Sage. anthaceae) prevent gultation- ind E of the mesophyll. Pl. Cell Environm. 28: 1179-1190 Figueiral, L, V. Mosbrugger, N. P. Rov Utescher & T. P vegetation of northwestern Germany: An analysis of wood , A. R. Ashraf, T. . Jones. 1999. The Mic ‘ene peal-forming remains and m with previous ee inter- pretations. Re obot. Palynol. Fluteau, F., J. b. se, Y Broutin & "n Berthelin. 2001. flora during the Permian. Climatic and paleogeographie constraints. Earth Planet. Sci. Lett. 193: 603-616 Foster, D. R., P. K. Sc m & S. T. A. Pickett. Insights from k ^oecology to community ecology. Ecol. Evol. 5: 119-122. Fox, D. L. & p. m Koch. 2003. Tertiary history of C4 biomass in the Great Plains, USA. ps ology : 09-812. & . 2004. variability in Neogene dus ol carbona Extension of Cathaysian 1990. Trends Carbon and gs isotopic : Constraints on the evolution of the Cy-gras E “of the Great Plains, USA. Palaeogeogr. laeoclimatol. Palaeoecol. nM ;. 1984. The seasonal environment of the Purbeck (Upper dd he: il fore 3: 285 -307. . Wing. 2004. Oxygen isotope and nd ó'?0-latitu- Francis, J. E. sts. Palaeogeogr. Palaeoclima- tol. Palaeoecol. < Fricke, H. & 5. paleobotanical estimates of temperature a gradients over Ja C during the early Eocene. Amer. J. Sci. 304: —635. Fritz, W. J. 1980a. e of the depositional Geology 8: environment of the Yellowstone “fossil forests.” 309-313. . 1980b. Stumps transported and deposited upright by Mount e e mud flows. Geology 8: 586-588. . Comment and reply on ‘Reinterpretation of the de depen environment of the Yellowstone fossil forests." Geology 9: 53-54. : r Harrison. 1985. Transported trees from the 1982 Mount St. Helens a su e ir use as paleo- current indicators. xor . Geol. Galloway, W. E. & D. " Hobda ay. o Terieenous ls de ban one systems: Applications to fossil fuel and E rk. groundwater resources. Springer-Verlag, Nev E > Galtier, J. & O. Béthoux. 2002. Morphology i growth habit of Dicksonites pluckenetii from the E pper Carboniferous of Graissessac (France). Geobios 35: 525-5: & V. Daviero. 1999 — gblumedhltum from the Upper Carbonifer- ous of Frar nt. J. Pl. Sci. 160: 1021-1033 Gastaldo, R. n 1984. A case against pelagochthony: The untenability of Carboniferous arborescent lycopod-domi- Structure and o of nated floating peat mats. Pp. 97-116 in. K. R. Walker (editor), The Evolution-Creation Controversy. E on Religion, Handbook. Paleontological Society Special Publications Philosophy, Science and Education: No. I ———. 10986. Implications on the E of autoch- sedimentary Palaeoecol. 53: thonous Carboniferous lycopods in clastic environments. Palaeogeog. Pslecocliiaatdl. 191- Volume 95, Number 1 2008 DiMichele 8 Gastaldo Plant Paleoecology in Deep Time ———. 1. Confirmation of Carboniferous clastic swamp communities. Nature 326: 869—871. Taphonomic considerations 2 plant evolu- tionary fs Palaeobotanist 41: 211-223. 1992 b. Regenerative growth in fossil bupsetsils (Calamites) following burial by alluvium. Hist. Biol. 6 203-220 . The genesis and sedimentation of phytoclasts l environments. Pp. 103-127 in Traverse (editor), Sedimentation of Organic Particles. Cambodee Univ. Press, Cambridge. with examples from coasta . 1999. Debates on autochthonous and alloc morum origin of coal: Empirical science versus the diluvi Pp. 135-167 in W. L. Manger ee The Evolution. Creation Controversy II: Perspectives on Science, Reli- gion, and Geological Education. Dn dec al pu Papers, Vol. : . 2004 4. The relationship m — and log WeiPelster of coarse ge sis. Palaios 19: 595 . Hydrological controls on the M plant- fossil record: A model for preservational modes. 7th European Palaeobotany-Palynology Conference, Abstracts and Progen National Museum, Prague. orientation in a Paleogene fluvial c basin, Germany: Implications for the use debris for paleocurrent analys 20€ C. W 007. Sedimentology and paleontology of a Carboniferous log jam. Int. J. Coal Geol. 69: 103-118. & T. Demko. 2005. Long-term ciis controls b continental plant-fossil record. Geol Abst Prog. 37(7): 11 & A. Y. Huc. 1992. Sediment facies, depositional environments, and distribution of phytoclasts in the Recent Mahakam River delta, Kalimantan, d is ee 591. . Amer. Indonesia. >. 1995. Peat or no peat; the Rajang and ms River 2e. compared. Geol. Soc. Amer. Abstr. Prog. 27(6): 30-31. & mechanism to explain the preservation of leaf litter lenses in coals derived from Palaeogeogr. Palaeoecol. raised mires. Palaeoclimatol. 49: pur P. Douglass & S. M. McCarroll. 1987. Origin, c hae ‘teristics and provenance of plant macrodetritus in a Holocene crevasse splay, Mobile delta, Alabama. Palaios 2: 229— 240. arce, C. Degges, R. J. Hunt, M. W. Peebles & D. M y T 1989. Biostratinomy of a H lake: A backswamp - mid-channel Palaeobot. e 58: 47-60. n & A. Y. P 1995a. The tidal character of futiel iens of the Recent Mahakam River delta, Kalimantan, oo Special. Publ. Int. Assoc. Sedi- mentol. F olo ocene oxbow Rev. transect. ^w. d & W. A. DiMichele. 1995b. c and sedimentologic characterization of *roof- shale” fois Pp. 341-352 in P. L yons, R. H. Wagner & E. Morey (editors), Historical Perspective of Early Twentieth Century Carboniferous Paleobotany in North America. as Society of America Memoir 185. . A. DiMichele & H. W. Pfefferkorn. 1996. Out of the ic ene into the greenhouse: A late Paleozoic analogue for modern global vegetational rs GSA Today 6: 1-7. gel, W. Püttmann, U. the Late Oligocene vegetation in central Europe. Palaeobot. Palynol. 101: 71—94. N. Ware. 2004a. In situ, erect forests are evidence for large-magnitude, coseismic base-level s within Poss vanam pes oth ems z the Black Warrior Basin, USA. EP. 2 in J. Pashin & R. A 3 , Coal- [o 'aring Tute Sequence Stratigraphy, Paleoclimate, and American Association of Petroleum Geologists DM in Geology Series, No. Ma , L. Stevanovic-Walls & W. 2 Bo S 2- F. Greb. 2004b. Community he depre: of Early 3 oa peat mires. Geolog 32: 693-696. —— ——, R. cn M. K. Bamford, C. C. Labandeira, f Na J. Sims. 2005. Ec trends macrofloral ==. across the rmia es boundary, Karoo Basin, South porn Palaios 20: 478-497. . A. Gibson & A. Blanton-Hooks. 2006. The Late Mississippian back-barrier mars y E in the Blac Warrior and Appalac ‘hian basins. Pp. 1 & W. A. DiMichele (editors), E Through ‘Tine. Geological Society of ims Special Papers see, C. T. 2005. The genesis of mass doned deposits (bedload a ree in the Tertiary of the Lower Rhine Basin, es Palaios 20: 463—478 Gemmil, R. Johnson. 1997. Paleoecology of a late ee 'ene “iin me gaflora from the northern Great Divide Basin. Palaios 39-448 i 1984. Plant Life in the 5 B uem shi, I. Milagres, E. C ptional land plant Cn paranensis, and a hypothesis. e on the life cycle o eutrachoephytes. Rev. Palaeobot. Palynol. Gilinsky, N. L. € K. Bambach. 1987. patterns of origination za extinction in higher taxa. Paleobiology 13: 427—44: Gitay, H. & I. R. Noble. 1997. What are functional types and how ie we e seek them? Pp. 3-19 in T. M. Smith, Woodward (editors) Plant Funcuonal C Roduienes. 2006. An exce specimen of the sal ac) i) = E E the earliest 142: 123-130. Asymmetrical Relevance to Ecosystem e es and Global Change. Cambr Univ. Pre Glasspool, I. J., J. Hilton, M. E. 2004. a physiognomy Eiesnteptetids and the potential to ud. Paleo oic climates using an extinct plant group. TEET D ol. 205: 69-110. Gleason, H. A. The individualistie concept of the plant e B Torrey Bot. Club 53: 7-26 Grady, W. C. & C. Eble. 1990. Relationships among x minerals, m. paleoecology in a column of HERR Coal (Uy pper Pennsylvanian) from north- Gaal West Virginia (USA). Int. J. Coal Geol. 15: 1-26. Greb, S. F., W. A. Bn. i R. A. Gastaldo. 2006. Evolution of wetland types and the importance of wetlands in earth history. Pp. 1—40 in S. Greb & W. DiMichele o W. a (editors), Wetlands fou i Time. eee Society of merica Spe puo pers Green, W. A. 2006. RE the CLAMP: An exploratory graphical approach to the E Leaf Analysis Multi- variate Program. Palaeontol. Electronica, Article 9.2.9A RE D. R. 2005. Leaf margin analysis: aints. Palaios 20: a 05. ory- EIE K. M. ea and climate, Bolivia: Impli mating paleocli 688, Taphonomic 2000. e between leaf 'ations for esti- mate from fossil floras. Paleobiology 26: 188 Annals of the Missouri Botanical Garden Grimaldi, D. 1999. The co-radiations of pollinating insects and angiosperms in the Cretaceous. Ann. Missouri Bot. Gard. 86: 373—400. Guerra-Sommer, M. & zzulo-Klepzig. 2000. Early Permian palaeofloras lita dn m Brazilian Gondwana: A paleoclimatic approach. Revista Brasil. Geoci. 30: 486—490 Hammond, S. E. € C. M. Berry. 2005. A new species Tetraxylopteris (Anerophytales) ps t Linn. Soc. 148: —303. T. 2003. Complex ye sil: of solitary and Venezuela. Bot. J. Hasiotis, S. social soil s Understanding their evolution and roles in te ial paleoecosystems. Palaeogeogr. Palaco- climatol. Palae: 'oecol. 192: 259-320. ayek, L. & M. B Surveying Natural Poy MN C ‘olumbia Univ. Hernandez-Casu G. R, G. W Rothwell & G. Mapes. 2001. Thuc s eae fam. nov. with a review and lia of Pa leozoic qudm conifers. Int. J. Pl. Sei. 162: ———, ——— & Stockey. 2003. Growth architecture of Thue ydia pohonin nsis, a model for primitive walchian conifer p e . Pl. Sei. 164: 443-452. S . S. Morgans-Bell, J. C. McElwain, P. M. Pda & C. E. Ross. 2003. Middle Jurassic = Rees, al A. perturbation in the V Carbon-cycle ' and accompanying 11 changes 1 in the terrestrial paleoenvironment. J. Geol. 111: 95 16. Hickey, L. J. gradual, not catastrophic change al the end o Cretaceous. Nature 292: 529-531. & J. A. Doyle. 1977. idence for angiospe rm MAE Bot. Rev. -] T J. GR: backbone of seed-plant phylogeny. J. Torrey Bot. Soc. 133: 119-168. Hofmann, 1981. Land plant evidence compatible: with Early Cretaceous fossil 13: 3-104 . Bateman. 2006. Pt sec rms are the | & R. Zetter. 2005. different land plant habitats of the Pannonian Pasin system (Neogene, Eastern Australia). Palaios 2 266-27 Holland, S. P 2006. Reevaluating the correspondence analysis and non-metric multidimensional Amer. Abstr. utility of detrended s caling for ecological ordination. Geol. Soc. .M. eibold. N. Mouquet, R. D. Holt & M. Ho opes. 2005. tacna : framework for une scale ei ecology. E 1—31 in M. cola E M. Leibold « Holt i ditors p l Spaliz Dynamic S E A cological Communities. of n 'ago Univ. 5. Bridge. An pics M. Hueber, D. H. Griffing & J. 2002. Rudy ierat] plant. environments: from the Emsian of Gaspé, 17 D. editors), Plants Invade the e Canada. Pp. 179-212 in P. Gensel & Edwards E dp dud and Environmental Perspectives. Columbia lv. ress, Ne W 'or ol N. TIL E. C. Olson & S. R. Beerbower. 1996. Amniote origins and the discovery of herbivory. Pp. 207— 264 in S. S. Sumida & K. M. Martin (editors); Amniote ira : ademic Press, New York. . J. Cantrill. 2001. E e from the Albian of Antarctica. Cretaceous Palaeoecology and taxonomy s. 22: 779-7 n We. i5. E 2001. The Unified Biodiversity and Biogeography. Biology 32. P Theory Neutral Monographs in Population Princeton. Princeton Univ. Press. 1€ Devonian of Reconstruction of X 1986. Biology. chance, and history and the structure of tropical rain forest tree communities. Pp. 314-320 in Diamond € T. J. Case (editors). Community Ecology. Harper and Row, New York. Hueber, F. M. 1971. a princeps. Taxon 20: 641—642. . Foster. Sawdonia ornata, a new name for . Rotted wood-alga-fungus: The n and life * homi Rey oo Palynol. 116: Huff, P. M., Wilf & E. J. Azumah. 2003. ep future for d oc D estimation from leaves? Preliminary results. Palaios 18: 266-2 Hutchinson. G. E. : 206-274 1965. The Ecological Theater and the Evolutionary Play. Yale Univ. R. & O. Rösler. 2000. Floristic migration in South Carboniferous: Phytogeographic and Press, New Haven. lannuzzi, merica during the hati consequences. Palaeogeogr. Palaeoclima- . Palaeoecol. 71-9: & Lo H. W. Ho oen 2002. floral belt in Gondwana (Late Visean, Early = A pre-glacial, warm- Pe adis Carboniferous). Palai 571-590. Ives, A. R. & S ecosystems, POIBIUS 317: Jacana, S. T. 2000. A Quaternary perspective on Late Cenozoic Pp. 287-308 in R. A. Gastaldo & Michele (editors), Phanerozoic Terrestrial. Eco- h. Carper nler. 2007 7. Stability and diversity of 50-C čcosystem reorganization into the te rre sirial i oecology. | systems. Pal ontological Society Papers, Vol. 6. Jacobs, 2002. Estimation of low-latitude paleoclimates using p» sil angiosperm leaves: Examples from e Hi 99-42 Miocene M ills, Kenva. d 28: : Js acobs. 1999, The origin o ). Kingston & L. grass- al ecosyslems. po Tee Bot. Gard. 86: 590—643. Jahren. A. H.. B. A. Methanogenesis in Eocene arctic s E LePage & S. P. Werts. 2004a. soils inferred from delta of tree fossil carbonates. Palacogeogr. Palaeoclimatol. ias 214: 347-358. tersen & 0. Seberg. 2004b. Plant DNA: A new sola for € arbon stable isotope analysis and a potential indicator. G E 32: 241—244. — ———, K. Kelm E Wendland, G. Petersen & O. Seberg. 206 o. Cobos stable ey composit ion of DNA isolated 34: 381-384. N James, W ack & H. C. Mon O M | V ial from an incipient paleosol. Geology W. C. C. H. M E 998. poene Paleosol classification modified}, po rn. Int. 51/52: e Jaramillo, C., M. J. Rueda & G. Mora. 2006. egi: plant diversity in A Neotropics. Science 31 893-1890. a The carbon and its relation to Ecol. Applic. 10: 423-436 1995, Tectonic and paleoclimatic significance Jackson. vertical — Jobbágy. D. G. & R. distribution of soil organic climate and vegetation. Joeckel, R. M. of a prominent Upper Pennsylvanian Virgilian Stepanian) weathering profile, lowa and Nebraska, USA. Palaeogeogr. Palaeoclimatol. Palaeoecol. 118: 159-179 Johnson. K. J. 1992. Leaf- extinctions at Dakota. : E fossil evidence for extensive floral (is Cretaceous- rliary E North . Cretaceous Res. 13: 91-117 The reconstruction i ancient des ‘apes: An Late Cretaceous Hell Creek F ee of North Dakota Acad. Sei. 53: 134-140. T m the North Pad Proc. Johnson, K. R. & B. Ellis. 2002. : tropical um est in Colorado |. i million years after the Cretaceous-Tertiary seer Science 296: 237 pie ~ ~ — — ». Raynolds. 2002. cient. Denver: Scenes from 300 sillon years of the Col qe Front Range. Denver Museum of Nature and Science, Denv Volume 95, Number 1 2008 DiMichele & Gastaldo Plant Paleoecology in Deep Time 189 ———, D. J. Nichols € J. H. Hartman. 2002. Hell Creek Formation A 2001 synthesis. Pp. 503-510 in J. H. man, K. R. Johnson & D. J. Nichols (editors), The Hell C au Formation and the Cretaceous-Tertiary boundary in the northern Plains: An integrated continental record of the end of the Cretaceous. Geological Society of America Special iud 361 Kasper, A. E., Jr. & H. N. Andrews. 1972. Pertica, a new genus of d onian ur from northern Maine. Amer. J ot. : 97-911. Keeley, Í. & G. Busch. 1984. Carbon assimilation characteristics P a E CAM plant /soetes howellii. PL a 20 Kembel, S. W. & E js E 2006. The Sg ona structure 7 a — 'al forest tree community. Ecolo 87 a ps 6-599. er . & J. ea 1985. Die Macrofloren des saarp- cius he ən Rotliegenden (?Obr- bue D Perm; SW- Mi tt. 59-280. & Pi Dd 2006. nan Neue elements in the Upper log 265-208 Deutse a Meinzer Geowiss. A. Hamad, B. Vö Typic ut prm sic Permian of the Doe dd Geology Kershaw, ntitative and Dynamic Ecology, 2nd. ed. ur yen London. Klavins, S. D., E. L. Taylor, M. Krings € T. N. Taylor. 2003. Gymnosperms from the Middle Triassic of rms tica: The Gondwa Plant first pue preserved cycad pollen cone. Int. Sci. 164: 1007-1020. Klein, G. deV. 1992. Climatic and tectonic sea-level gauge for Midcontinent Pennsylvanian cyclothems. Geology 20: 43: cl. A. H., K. J. Niklas, P. G. sel & B. H. Tiffney. 1984. Character Wivensitie 'ation i patiems of evolution in early vascular plants. O 10: : 7. Re P.EJ. C. a hos & D Dettman. 1995. Stable trati and E ee of the Paleogene Piom B yoni USA). Pal Palaeocli- matol. Palaeoec ol. 11 -89. yde, R p. Heppel, M. L. Fogel, S. L. Wing & J. C. Zalos 2003. Isotopic records of carbon cycle and climate change across the Pale ~ alaeogeogr. eocene-Eocene boundary from the Bighorn Basin, a Pp. 49-64 in S. Wing, P. D. Gingerich, B. Schmitz & E. Thomas (editors), Causes and Consequences of p rly Paleogene Warm Climates. Geological Society of America Special Papers 3 dl W., A. Roth-Nebelsick € M. Grein. 2006. Stomatal density and atmospheric COz—Mechanistic modeling approaches based on diffusional resistances and photo- is. 7th European idR Palynology Confer- ence, Abstracts and Program, Natio n, Prague. L. R., D. J. Bead ing, E E. M: Stomatal frequency synthes ae Wagner, adjustment of iur conifer species to historical changes in atmospheric CO. Amer. J. Bot. 90: 610—619. Kowalski, E. A. & D. A. Dilcher. 2003. Warmer paleotem- peratures for terrestrial ecosystems. Proc. Natl. Acad. Sci. U.S.A. 100: 167-170. Kraus, M. J. 1999. Paleosols in clastic sedimentary rocks: 7: 41-70. Their a S m Earth S Sci. Rev Significance of different . Hasiotis. 2006 modes of PT s bs to interpreting paleoenvir- onmental and paleohydrologic settings: Exam[ ples from Paleogene paleosols, Bighorn Basin, Wyoming, U.S.A. J. Sediment. Res. 76: 633-646. Krings, M. & H. Ker ecology of Blanz an praedentata (Gothan) nov. comb., a climbing neuropteroid seed fern from the Stephanian of central France. Int. J. Pl. Sci. xd 603-619. i . L. Taylor. 2003. How Paleozoic hanes off the ground: On scrambling and climbing Carboniferous—Early Permian pteridosperms. Bot. Ren 69: 204—244 Kürschner, W. M. 199 7. The anatomical diversity of rec and fossil leav s of the durmast oak Lieblein/Q. peudocasianea Goep their use as biosensors of To dE CO» lev Rev. Palaeobot. Palynol 96: 1-30. Kvaček, J., H X Falcon-Lang J. Dašková. 2005. A new Late Cretaceous ginkgoalean reproductive structure Neh- vizdyella gen. nov. from the Czech Republic and its whole- plant reconstruction. Amer. J. Bot. 92: 1958-1969. C. C. 1998 1999. Morphology, growth habit, and vines and Quercus ides ert) implications for els. Labandeira, Early history of arthropod and vascular plant associations. Ann. Rev. Earth Planet. Sci. 26: 329-377. . 2002. The history of associations between plants rera & O. Pellmyr Evolutionary and als Pp. 26-74 in C. Her Plant-Animal Interactions: An m E Blackwell Science, Oxford. (editors), 6a. The four phases of plant-arthropod associ- ations in t ep tim eol. Acta )9—438. 2006b. clades and their associations: Silurian to Triassic o and hexapod New d 7. Minimal insect herbivory for the Lower Permian Coprolite Bone Bed site of north- central Texa ISA, and late Paleozoic floras. Palaeogeogr. Palaeoclimatol. Palaeoecol. 247: 197-219. & T. L. Phillips. 2002. Stem DA and petiole | from PM s, USA: Im; compa rison to other tree ferns of Illino tions for the origin of the borer and E Ded fee ng groups and holom ia kolu insects. Palaeontogra- 64A: 1-84. phica J. MNA es Insect diversity in the fossil oe | ence 261: 310-345 Phillips € R. A. Noon 1997. Oribatid mites and > (s 'omposition a plant tissues in Paleozoic li c coal- swamp forests. Palaios 12: 319-353. Laveine, J.-P. 1997. Synthetic analysis of the Neuropterids. Their interest for the decipherment of Carboniferous ae Rev. Palaeobot. Palynol. 95: 155-189. , Y. Len e € 5. Zhang. 1993. T the n ^ some Eurasiatic Carboniferous plants 12th The importance of for i si a reconstructions. Compt. Rend., Inter. Cong. Carb. Perm. 2: 89-106. A. & Y. Lemoigne. 1989. Global paleobotany, as exemplified by some Upper EN pterido- sperms. Bull. Soc s 98: 115-12 & )00. aie PME and palacogeography on the dus of examples from the '. Paléobiol. Genève 19: 409—425 :. Belge Gé Carboniferous. Law, R. & Leibold. 2005. Assembly dynamics in netac ree Pp. 263-278 in M. Holyoak, M. A Leibold & D. Holt (editors communities: Spatial Dynamics = Ecological Commnnines: Univ. of Chicago Press, Chicag LePage, B. A. & H. W. Pfefferkorn. 2000. Did ground cover change over geologic time? Pp. 171—182 in R. A. Gastaldo & W. A. DiMichele (editors), Terrestrial Ecosystems. Paleontological Society Papers, Vol. € Phanerozoic 190 Annals of the Missouri Botanical Garden Beauchamp, H. W. Pfefferkom & J. Utting. 2003. Late Early Permian plant fossil from the Canadian High Arctic: A rare paleoenvironmental/climatic window in Palaeogeogr. Palaeoclimatol. Palaeoe- n Pange a. e 345—312. o«J- MW. a & H. Yang. 2005. The ge ed qa Topics in Geobiology, Vo : col. ae sy of Metaseqi uoia. Springer, Dorde ‘ht, The Netherlands. ai, X. & X. Wu. 1996. Late Paleozoic phytoge ographic provinces in China and its adjacent regions. Rev. 2 Pal: vba fer ge 90: 41-62. Liang, 2004. Palynology, palecorco ogy and Sie i climate “ol ilis Miocene Sha Province, eastern China. A. Bruch, M. B Li Q. a Sun & J. M. 5. Destine ilia climatic estimates from different pala eei methods: An ex- ample from the Middle Miocene Shanwang flora > ~ Palaeogeogr. Palaeoclimat. ME 198: 279—: Liu, Y. € R. A. Gastaldo. CM NUN and provenance of log-transported gravels in a As arboniferous — peep : Me nt. Patrol 62: 1072-1083 )92b. Characteristics m a Pennsylva- nian jb gar ue Sediment. Geol. 77: 197-214. Lockheart, M. J., P. F. van Bergen & R. P. Eve srshed. 1997, Variations in the stable carbon isotope compositions of individual pes from the leaves of modern angiosperms: Implications for the study of higher land bos derived matter. Organ. Geocher 26: sedimentary organic 137-153. ,ooy, C. A. Brugman, D. L. Dilcher & H. Visscher. 1999, The de Fue resurgence of equatorial forests after the Permian- Triassic ecologic crisis. Proc. Natl. Acad. Sci 96: 13857-13862. S Bengtsson, J. P. ton, D. Inchausti, m Biodiversity and ecosystem functioning: Current. knowl- e and future challe nges. Science 294: 8 s l. Lowe idc . Hein, Bobst, Jordan, T.- Ku & B Luo. 2003. ino assessment of nun rend s in climate-sensitive closed-basin lake sediments: Sal de Atacama, Chile. J. Sediment. Res. 73: 91- 104. Lyons, P. C. & W. C. Darrah. 1989, Earliest conifers in North America "er and/or paleoecological indicators? Palaios 4: ee Mac Arthur, e 5 O. Wilson. 1967. The Theory of Is e os ae Princeton Univ. Press, Princeton. Mack, 1997. Paleosols for Sedimentologists, 2nd ed. Ge Bo r Soci iety of America Short Course Notes. W. 93. Classification of paleosols. Geol. Soc. Amer. Bull. 105: 129-136 199 Magurran, A. E. 8. Ecological Diversity snd Its ES . 1. James & H. C. Mong Princeton Univ 992. Sphe pud. and 1 from the Amer Measurement. Mamay, 5 l Dinidir- like associatior Pennsylvanian) of New Mexico. 1092-1101. Telangiopsis in a Virgillian (Upper J. Bot. 79 — — . M. Bateman. 1991. Archaeocalamites lazerit, sp. : The range of Archaeocalamitace tended from the nov. Pecan Pennsylvanian to the nid. qum Permian. Amer. J. Bot. x 489—496. Manchester, S. & P. R. Crane. 1983. Attached leaves, inflorescences, and fruits of Fagopsis, and extinct genus of Fagaceous affinity from the Oligocene um Flora of Colorado, U.S.A. Amer. J. Bot. 70: 1147-116 Manchester, , D. L. Dilcher & W. D. Tidwell. 1986. Intercon- nected reproductive and vegetative remains of Populus (Salicaceae) from the Middle Eocene Green River Formation, northeastern Utah. / oda tive vegetative remains the extinct American-European genus Cedrelospermum u Imaceae) from the t ETIN of Utah and Colorado. Amer. J. Bot. 76: 256— 1999, Biogeographical relationships E ips E Tertiary floras. Ann. Missouri Bot. Gard. 72-522. & B. H. Tiffney. 2001. Integration of paleobotanical and neobotanical data in the assessment of phytogeo- graphic history of is E ce is clades. Int Plant Sci. 162(suppl.): 5 ¿ M. S. Zavada. v ' rou foliage with intact ense from the Eocene of Wyoming. Bot. Gaz. 148 c Es apes, G. 1985 ena in midcontinent North e Bot. Gaz. 146 -167 Maples . & A. W. A n 1988. Monte Carlo simulation of ioe i binomial ae coefficients (II): Effect sparse data. Palaios 3: 9 3. Martin-Closas, C. & J. on tier. 2005. Plant taphonomy anc paleobotany of Late Pennsylvanian intramontane wetlands — 2 [e the Graissessac-Lodéve Basin (Languedoc, France). Palaios 20: 249—265. May, R. M. 1973. Stability and Complexity in Model Ecosystems. Monographs in Population Biology No. 6. Princ Univ. Press, oou McCune . B. Grace. 2002. Analysis of Ecological Com- Merc erm San Design, G leneden Beach, Oregon. McElwain, J. C. 1998. Do fossil plants signal palaeoatmo- spheric COs concentration in the geological past? Philos. Trans., Ser. ha 353: 83-96. & W. G. Chaloner. 1995. Stomatal density and index of fossil elena track uud carbon dioxide in the Ann. Bot 389—395. & jc sh The fossil cuticle as a ske ~ record of environmental change. Palaios 11: 376—38¢ eerling € F. I. Woodward. 1999, Fossil s Triassic—Jurassic boundary. Paleozoic. and label warming al the Science 285: 1386-1390. Miall, A. D. 1990. Principles of Sedimentary Basin Analysis. ui a -Verlag, New York. Millay, M. A. 1979. Studies of Paleozoic marattialeans: Á Se of es of DP MS Palaeontographica 169B: —— ——— & T. N. Taylor. pem Paleozoic seed fern pollen Mea a Rev. 45: 301-375. Miller, S.M. Aa D. L. Meyer the a speci 69 & B. F. Dattilo. 2001. Tho use p faunal gradient analysis for d correlation and assessment of changes in sea-floor topo- up in 1 the type Cincinnatian. 603 3613. J: Geol se Mille G., J. V Browning, S. F. P P. J. Sugarman. 1997. ao evolution of the Ne W re coastal plain: a sediment supply. 3 in K. G. Miller . W. Snyder (editors), Proc. Ocean pp us Sci. ae Vo : Montañez, I. P., N. J. Tabor, D. Niemeier, W.A Michele, Frank, C. R. ooy & J. L. Isbell. pu COs- and vegetation instability during Late : 87-91. evel, tectonics, Paleozoic de une Scienc Mosbrugger, V. 1 The nearest s elative method. Pp. 261-265 in i > Jones & N. P. Rowe (editors), Fossil Plants and Spores: Modern Ta einn Geological Society, London. Volume 95, Number 1 2008 DiMichele 8 Gastaldo 191 Plant Paleoecology in Deep Time Utescher. 1997. The coexistence approach —A method bs quantitative reconstructions of Tertiary terres- trial palaeoclimate data using wy — Palaeogeogr. Palaeoclimatol. Palaeoecol. 134: 6 JC , G. Belz & A. E TE esonsiticuon of an in situ forest from the Lower Rhine Embayment, Germany—New methods in palaeovegetation analysis. Palaeogeogr. Palaeoclimatol. Palaeoecol. 110: 295-317. Mustoe, G. E. 2007. Coevolution of cycads and dinosaurs. Cycad Newslett. ou: 6-9 A vi raf. ] 994. Three- > peat nort E western Nguyen Tu, T. T., . Kuerschner, S. Schouten & P. F. van Bergen. 2004. I eaf carbon isotope composition of fossil and extant oaks grown under differing atmospheric CO, levels. jl. 212: 199-213. Palaeogeogr. ergs de Palaeoecc Niklas, K. J. 1981. Simulated wind pollination and airflow around ie " some early seed plants. Science 211: 215-211. 83a. The influence of Paleozoic ovule and cupule morphologies on wind pollination. Evolution 37: 968-986 3b. Early Be Lp wind pollination studies: A ole Taxon 32: 99— 92. Plant PM cal to Plant Form and Function. Press, o 1994. Plant Allometry: The Se aling of Form and Process Univ. of Chicago Press, Chica 1999. Evolutionary walks through a land plant morphospace JJ. Bot. 50: 39-52. 2000. Modeling plant form- psi ‘tion relationships: A ne prepa 26(suppl.): 289-304. 2006. A phyletic perspective on the allometry of slant eee itioning m EM equivalent organ-categories. New Phytol. : 27-40. & . Enquist. 2001. ships for foU ifie a biomass Poean rates and Natl. Acad. Sci. U.S An Engineering Ap- Univ. of Chicago patterns Invariant se s relation- body size. Proc. J.S 22-2927 ¿ T. Speck. is factors against wind-induced stem failure. Amer. J. Bot. 1278. 88: 1266— ah, Bvoluüerny MA in safety Midgley € R. H. Rand. 2003. Tree size requency P plant density, age and community disturbance. A Lett. 6: 40541 1. nd Smee, Laland & M. W. Feldman. iche a don Amer. Naturalist Ve 641—064 s H. T. 1983. Systems Ecology: An Introduction. John Wiley E "Mn New York. Olszewski, 19 quence stratigraphy of an Upper a ote cyclothem from North Am ica (Iola Limestone, Kansas and Missouri, USA). Facies 03 1996. Oshurkova, M. V. 1996. Paleoecological parallelism between the Angaran and Seg" ^an phytogeographic provinces. Rev. x p cus ): 99-111. Overpeck, J. ' Rs : T. Webb III. 1992. Mapping ie de a th American vegetation change of the last Ka: No-analogs and the future. Geology 20: 1071-1074. n K. H. Freeman € M. A. Arthur. 1999. Late Miocene bins logic CO» a and the expan- sion ko Ca grasses. Science 285: 876-879 Page, C. N. 2002. Ecological ec in fern ^ nd overview. Rev. Palaeobot. Palynol. 119: ES FT -— evolution: A Palae obotanical Research Group, University of Münster. 2007. Hhynie Chert and its Flora: d w.uni-muenster.de/ Ge MAUS EE yP t sed 14 December 20€ , ACCES Pant, D. D. 1996. The biogeography of the = ee floras of India. Rev. Palaeobot. Palynol. 90: Parmesan, C. 2006. Ecological and evolutionary responses to climate change. Ann. Rev. Ecol. Evol. Syst. 37: 637-669 Parrish, J. T., T. Demko Tanck. 1993. ee nlary paleoclimatic leas: Ahat hey are and «T Ser. B. 344: 21-25. what they tell us. Philos. Trans., Ser. Patzkowsky, Holland. 2003. Lack of ae saturation at the beginning of the Paleozoic plateau: The dominance s regional over local processes. Paleobiology 29: 545-560 Pearson, S. & J. um 2002. Understanding arid A environments using fossil rodent middens. J. Arid Environm. 50: 499-511. : D. L. Turcotte. 1997. Synthetic stratigraphy with a stoc hasi pon model of fluvial sedimentation. J. Sediment. Res. 67: 1060-1067. Penick, L a Jr: 198r The New Madrid Earthquakes dl: edition). Univ. of Missouri Press, Columbia. Peppers, R. A. 1979. Development of coal-forming floras during the early has of the bua in the Illinois Basin. Pp. 8-14 in J. E. Palm . Dutcher (editors), Depositional Ee Structural Histor of the Pennsylvanian System of the asin: Invited Papers. Ninth International Congress on Carboniferous Stratigraphy and Geology, Illinois State Geological Survey, Guidebook llinois Series 15a. 996. Palynological Correlation of Major Pennsyl- vanian (Middle and Upper Carboniferous) Chronostrati- raphic as in the Illinois and Other Coal Basins. Geological Societ Pn Memoir 188. Pfefferkorn, H. W. 1979. High a and stratigraphic age of the Mazon ien Flora. Pp. 129-142 in M. H. Nitecki (editor), Mazon Creek Fossils. i nic Press, New York. Ic A note on - term “upland flora.” Rev. ba Palynol 30: 157-158. . 1995. We are Miu climate chauvinists. Palaios 10: 389-391. & K. Fuchs. 1991. A field classification of fossil plant substrate interactions. Neues Jahrb. Geol. Palaeon- tol. 183: Lie 36. —— & V. Jun. 2007. Early p hou forming floras Mit as compressions from the Wuda District (Inner Mongolia, China). Int. J. Coal Geol. pA 90-102 & M. Thomson. 1982. patterns in Upper ÓN plant-fossil assemblages. Geology 10: 641-644. H Changes in oia Mos & H. Hass. 1975. Quantitative Charakt risierung ober- dx. Abdruckfioren N. Jahr. Geol. Palaeontol. Abhandl. 150: 2 W. H. Gillespie, D. A. em 5 M. . Scheihing. onstruction and architecture A medullosan pride rms Pts arial The Mosasaur 2: 1-8. R. A taldo € W. A. DiMichele. osy ystem ee during the Late Paleozoic med Pp. 63-78 in R. A. Gastaldo & W. A. DiMichele (editors), Phanerozoic Terrestrial Ecosystems. Paleonto- Mes Society Papers, Vol. 6. & T. L. Phillips. 2008. Pennsyl- vanian floras af the United States as a record of usd climate. In C. R. Fielding, T. D. Frank & J. Isbell (editors), Resolving the Late Paleozoic e Ice Age in Time and Space. cad Society of America Special Paper 441 (in press Phillips, T. L. 1979. Reproduction of heterosporous arbores- cent lycopods in ssippian-Pe RN an of ( he Euramerica. Rev. Palaeohot.| ELE 21: Annals of the Missouri Botanical Garden f Middle Hlinois— Paleoec FRIES of uthern . DiMichele. 1981. T van age coal swamps in Herrin Coal Member at Sahara Mine. Pp. 231 -285 ink. J. Niklas (editor), | Paleoecology and Evolution, V . Praeger Publi York. & 190; Paleobotany, shers, Comparative ecology and ol. 1 -c — life- history biology d oeste lycopods in Late C a erous swamps of Euramerica. Ann. Missouri Bot. Gard. 7 560-5 & R. A. Peppers. Pennsylvanian coal-swamp climatic control on coal occurrence. Int. 205-255. 1984. Changing patterns of vegetation and implic ations J. Coal Geol. , ——., M. J. Avcin & P. F. Laughnan. 1974. Fossil ts and s Patterns of change in Pennsylvanian coal iA Basin. Science 184: 1367-1369. —— Kunz & D. J. Mickish. 1977. permineralized peat (coal balls) Member of the Illinois Basin. Pp. 18—49 in P A. D. Cohen (editors), iem eg Studies of Peat and s rerica Microform Publication 7. T eer of zoal swamps of the I A.B Geological Society of An ——— Peppers & W. DiMichele. 1985. Stratigraphic and interregional chá nian-age coal-swamp Ma Ce Environmental inferenc- J. Geol. 5: 43-109. The morphology and reproductive biology 6 144: Bot. ages in Pennsylva- es. Int. Pigg, K. B. of the Rare cone Mazocarpon. 600-613. G. W. Rothwell. 1983. Chaloneria gen. heterosporous ve Bo from the Pennsylvanian of North America. Bot. 44: 132-147 Podani, J. 2006. id B anquet s las gaey and data analysis in i vA science. Sei. 17: 113- 7 W. Kürschner. 1999, s iei and 257-260 in T. ones & N. P. Rowe Plants Modern Techniques. Geological Society of I on. & PE De 2000. an. s ‘ters in wood as palaeoclimate proxies. Pl. : 175-195. az. nov. LÈ? = — eg. Poole, I. M. ws The practice. Pp. (editors), Fossil and ca ondon, Lonc van rgen. Physiognomie and Ecol. 1999. Siliciclastic nd Applications. SEPM e H. Allen. Sequence Straligrap ibi oncepts Concepts in vc pire and Paleontology 7 l. Sedimentary Geology, Tulsa E. 2003. Climate-vegetation dynamics in the fast lane. Trends fica Evol. 51-553. Poulsen, C. J., D. Pollard, I P. Montañez & D. Rowley. 2007. Late Paleozoic tropical climate response to Gondwanan deglaciation. Prasad, V., C. A. imberg. H. A E & Sahni. 2005. MAE ja a the early evolution of grasses and grazers. Science 310: 1177 0. Prinzing, A., W. D n 5. Klotz & A Brandl. 2001. The niche of higher plants: Evidence for phylogenetic niche . Proc. Roy. Soc. Lond., Ser. D. Biol. Sci. 268: Society for Post, id e 771-77: conservatism 2383-2389. Prochnow, S. J., L Nordt, S. C. Atchley & M. R. 2006. Multi- proxy paleosol evide nce for | Middle An T. Palaeogeogr. in eastern 232: 53 rical succession within the riassic climate trends Palacoclimatol. Palaeoecol. Pryor, J. S. 1993. Patterns of ecolog Upper Pennsylvanian ones coal of Ohio (USA). Evol. —72 Trends Plants 7: 5 ———. 1996. The Upper Pennsylvanian Duquesne Coal « Ohio (USA): Evidence for a dynamic peat du 119-146, swamp community. Int. J. Coal Geol. 29: _ A. Gastaldo. 2000. Palaeoecological analysis of two Barly Pennsylvanian mineral-substrate we lands. Palaios 15: 3-13. ye, K., J. A. D. Dickson, N. Schiavon, M. L. Coleman & M. Cox. 1990. Formation of siderite-Me-calcite-iron sulphide concretions in intertidal m rents, North Norfolk, England. Se ‘dime ntology 37: 325-343. R Project for Statistical C MEHR The. 2007. , accessed 14 December 2007. rsh and sandflat sedin Raup, D. M. 1975. Taxonomic a rsity estimation using rarefaction. Pale e E 342. Raven, A. & R. icer. 1996. The evolution of 5p crassulacean acid metabolism, Pp. 360-385 in K. Winter _ A.C. Smith (editors), Crassulacean Acid Metabolism: Biochemistry, 5 ophy ee and E volution. Ecological Studies Vol. 114. Springer-Verlag, Ber Raymond, . Gee s diversity and climatic amelioratio cones the Early Carboniferous nd Paleobiology 11: 293-309. 987. TA Arii of Early Devo- nian plan traits. Palaios 2: 113-13 988. The lace Um of a p ball deposit from the Vra Pennsylvanian of lowa dominated by Cordaita- lean gymnosperms. Rev. Palaeobot. Palynol. 53: 233-250 1997. Latitudinal patterns in the diversification of Climate and the floral (editor), Patterns ch-U.S. Carboniferous mid- Carboni erous land plants: break. Pp. 1-18 R. L. Pale e Proceedings of a Czec lina 2 UA illinois State Museum Scientific Leary in . 26. & C. M. Metz. 1995. Laurussian land- E diversity Ma during the Silurian and Devonian: xtinction, sampling d or both? Paleobiology 21: 74-91. ——— & L. Phillips. 1983. Evidence d an Upper e rous mangrove community. Pp. 30 in H. Teas (editor). Biology of Mangroves. Tasks fo and Ecology Mangrov Vege p Se ience 8. i W . dois Publishers. The ———— . Parker & 5 € able Es Devonian. ^! 129-167 in p. Tiffney (ex litor), Geological Factors and the Evolution of Plants. Yale Univ. Press. New Haven ————,———— & Parrish. 1985. Phytogeography and paleoc ‘limatology of the Early Carboniferous. Pp. 169-22: in B. Tiffnev (e i Geological Factors and the Evolution of d Yale Univ. Press, New H: P. Cutlip ^ M. Sweet. 2001. Rates - pee esses of The world before beetles, termites and flies. Pp. : 235-283 in W. Allmon & D. Bots p Evolutionary Paleoecology. Columbia Univ. Press, New P. G. Gensel & W. E. of ius Silurian macrofloras. Rev. 165-192. Rees, P. M. 2002. Land-plant o and 7-830 aven. terrestrial nutrient cycling in the Paleo Stein. 2006. Phytogeography Palacobah Palynol. | the end-Permian mass extinction. Geog y 30: . M. T. Gibbs. En PAR J. E. Kutzbach & P. J Em 1999, TR climates: Evaluating model predictions using global paleobotanical data. Geology 27: 1-894. A. M. Ziegler € P. J. Valdes. 2000. Jurassic chytoseonaphy p climates: 2 ur 2 mode comparisons. Pp. 318 in B.” - Huber. >. Macleod & S. L. Wing m Warm C Tene in A de History. ress, C e _—- yibbs, J. E. Kutzbach, P. Behling € Mia 0 Permian a D NM 2 Pu tale comparisons. J. Geol. 1-31: Cambridge Univ. Volume 95, Number 1 2008 DiMichele & Gastaldo Plant Paleoecology in Deep Time 193 . R. Noto, J. M. Parrish € J. T. Parrish. 2004. Late Jurassic climates, vegetation and ovis distributions. 53. J. Geol. 112: 64: Reid, E. M. & J. Chandler. 1933. The London Clay Flora. British M Museum (Natural History), London. we! t to call early plant formations on 997. Karly forest soils and their role in Devonian global change. Science 276: 583-585. . 1999. Permafrost palaeoclimate of Permian B o- sols in the Gerringong volcanic facies of New South Wales Austral. J. Earth Sci. 46: 11-22. 200 A 300- a -year record of atmospheric plant cuticles. Nature 411: carbon dioxide from fossil 287-2 001». Soils of the Past An Introduction to l Science, Oxford. lenozoic expansion n erasslands and lado soole: J. Geol. 109: 407—426 — Paese 2nd ed. Blackwe 200 . 2002. Carbon dioxide and ina over the past 300 pr ge Trans., Ser. A. 360: as . Dilcher. 1981. A s te sis for the ee ie rise to dominance of flowering plants. Pp. 27-77 in K. J. Niklas ipaam, and Evolution. M New Paleobotany, Paleoecology, . Rec onstructions T selected seed Missouri ee Gard. 75: 1010 57. Era Permian ferns. Ann. J. Veevers & R. Mone Triassic coal gap between Late obal early extinction and ian Triassic recovery of peat-forming plants. Bull. ;eol. . Amer. 108: 195-207. "a je dn F. & O. Rösler. 2004. The paleoflora of Figueira in the context of the neopaleozoic of the Paraná il. Terrae 1: 44— -H. O 8 L. i duds rson. 2002. Geological se vols wv ac Early Devonian Rhynie cherts, Aberdeen- I An early terrestrial hot spring system. J. *. London 159: 203-214. A review of the taphonomy of plant remains in lacustrine sediments. Rev. Palaeobot. Palynol. 58: 33-46. Ricklefs, R. E. 2006. Evolutionary diversification and the origin of the diversity-environment relationship. Ecology 87(7 fum 83-513 Roth, A., 1994a. ! : Neugebauer. Efficiency and evolution of water transport systems in eee & H. J. higher plants: A modelling approach. I. The earliest land plants. Philos. Trans., Ser. B. 345B: 137-15 . 1994b. Efficiency NE evolution of water transport systems in higher plants: A modelling ee :h. H. Stelar evolution. Philos. Trans., Ser. B. 345B 153- Roth-Ne ie k, A. 2001. Heat transfer of rhyniophytic plant axes . Palaeobot. Palynol. 116: 109-122. a2 p )». Reconstructing atmospheric. carbon dioxide with stomata: Possibilities and limitations of a botanical Funct. 19: 261-295. rimm, V. E H. Haas & H. Kerp. 2000. ome analysis of Rhynia and Asteroxylon: CO»-sensor. Trees Struct. Testing functional aspects of early land plant evolution. a 26: 405—418 Uhl, V. Mosbrugger & H. Kerp. 2001. Evolution A review. Ann. and fi unction of leaf venation architecture: 87: 553—566. ce G. W. 1972. Evidence of pollen tubes in Paleozoic pteridosperms. Science 175: 772—774. = 1981. The Callistophytales (Pteridospermopsida): lam sophisticated Paleozoic gymnosperms. n ig Palynol. 32: 103-121. E the a gymnosperms. Pp. epe in R. A. . A. Thomas (editors), Systematic and Taxonomic nee in Palaeobotany. Systematies Association Special Vol. No. 31. Oxford Univ. Press, Oxford. Saa — ———. 1994, Phylogenetic a de among ferns and gymnosperms: Án overview. J. Pl. . 107: 411-416 . 1995. The fossil history of Sine ed Implications for the phylogeny of land plants. Pp. 71-86 in P. C. Hoch & A. G. Stephenson (editors), Experimental and nt Biosystematics. Missouri Botanical Molecular Approaches to Pla Garden, a ouis. ms COH Te. A monograph 2d Polcon e | organs, Pe Roy. Soc. Edinburgh Earth Sci. 77: 47-19. : G. Mapes. 1988. Vegetation of a Paleozoic conifer community. Pp. 213-224 in G. Mapes & R. apes (editors), Regional Geology and Paleontology af Upper Paleozoic Hamilton Quarry Area in Southeastern Kansas. Kansas Ge r Surv. ". 6. & 5. E. Scheckler. 1985. Biology of ancestral gym- nosperms. ^ 84-134 in C. B. Beck (editor), Origin and Evo- ja Univ. Pr Guidebook Se LL lution of MT Colum ess, New York & R. et. 1992. Pollination biology of Elkinsia abis co de for the origin of gymnosperms. rier Forschungsinst. Senckenberg 147: 225-231. & R. A. Stockey. 1994. The role of Hydropteris pinnata gen. et sp. nov. in reconstructing the cladistics of heterosporous ferns. Amer. J. Bot. 81: 479—492. 1982. Early seed plant wind pollination studies: A commentary. Taxon 31: 308-309. G. Mapes & R. Mapes. 1996. Anatomically presene Vojnovskyalean sael plants in Upper Pennsyl- vanian d marine shales of North America. J. Paleontol. 70: 1067—1079. el Stamm Taylor. & G. Mapes. 2000. An d ruderals in herbaceous e conifer: i -alaeogeogr. Pa- the evolutio Mesozoic laeoclimatol. Palac a 56: 139-1 vegetation. Rowe, N. P k & J. Galtier. 1993. S ea analysis of a faci bir stem. Proc. Roy. Soc. London, Ser. B, Biol. 252: Mn 28 Rowley, D. B., A. Raymond, JL. ish, A. I s, C. R. Scotese € A. M. Ziegler. 1 m “Carboniferous eae graphic and ml reconstructions. Int. J. Coal Geol. 5: 7 Royer, D. L. 2001. Stomatal density and stomatal index as indicators Palaeobot. Palynol. 114 ¿ P. Wilf. 2006. Why d toothed leaves correlate with cold climates? Gas exchange at leaf margins provides of peleontmosphi eric concentration. Rev. new insights into a classic paleotemperature proxy. Int. J. Sci. 167: 11-18. . L. J. Hickey & S. L. Wing. 2003. Ecological UE A in the “livi ing fossil” Ginkgo. Paleobiology 89: 84— Si . I. P. Montañez, N. J. Tabor & D. J. Kealing S C 0, as a primary driver of Phanerozoic climate Ee d Today 14(3 2: 4—10. ud f, D. A. Kowalski & D. due and plant e unt to Wi . Janesko, Dilsher: 2005. uhi eR of « leaf size and shape: Potential proxies for ibo fossil record. Amer. J. Bot. 92: 1141-1151. 194 Annals of the Missouri Botanical Garden . & R. R. Reisz. 2001. e zi pus essing in a terrestrial herbivore. 411: 684-68 Schec Td S. E. 1986. Geology, floristics and paleoecology of ate Dern coal Me from Appalachian Laurentia (US A.). Ann. Soc. € Géol. Belg. 109: 209-2 Sche "m M.&S Carpenter. 2003. cación regime shifts in erens Linking theory to observation. Trends Rybczynski, Nature ' £o c Schneider, | . Roessler. . A Permian calcic paleosol containing duse] liths ne microvertebrate remains from the Erzgebirge 8 B Neues J Basin, Germany : Environment and ahrb. 202 taphonomy. Geol. cal. Abh. 202: 243-258. G. . Osborn & S. L That . 2007. Aerodynamics t saceale pollen and its Amer. J. — Schwendemann, A. B., G. Wa che 97 27: — el e microns for wind pollination. 1371-1381. Scott, A. C. 19774. A review of the ecology of Upper Carboniferous plant es with Strathclyde. Palae ontology uh 4747: .1977b. € new data from g plant material from the abades of Britain. Palscanlolany 20: 59-08. — ———. 1978. Sedimentological and ecological control of W asian B plant o from west Yorkshire. Proc. Yorkshire Geol. Soc. 41: 46 . 1980. The ecology of some Upper Palaeozoic floras. Pp. 87- 115 A. L, The E AAA dud the Land ache Panchen (editor), Origin of ‘Terrestrial Vertebrates. Academic Press, London. 1985. Techniques in Carboniferous m an "e Problems and perspectives. Compt. Rend., trat. Geol. 5: 35- 1983. B of Yorkshire. Proc. Inter. Cong. Carb. S The uan fossil conifer Roy. Soc. ———— G. Chaloner. from a Westphalian Lond., Ser. B, Biol. Sei. 220: 163-182 3 J. Galtier. 1985. Distribution and ecology of early ferns. Proc. Roy. Soc. Edimburg JOB: 141-149. & . 1996. A review of the problems in the —— ó stratigraphical, palaeoecological and. palaeobiogeographi- cal interpretation of Lower Carboniferous (Dinantian) floras from Western Europe. Rev. Palaeobot. Palynol. 90: 141-153 & LJ. 2006. The diversification. of Paleozoic fire systems and rp in unas xygen concentration. Proc. Natl. Acad. Sei. U.S.A. 103: 10861-10 865 Glasspool. 1983. Plant/animal interac- Upper Carboniferous. Bot. Rev. 49: Taylor. tions during the 259-307. — W. thaloner & S. Patterson. 1985. Evidence of E arthropod rra in the fossil record. Proc. Roy. Soc. Edinburgh 86B: 133-140. PR & W.C 1992. Interaction and conva lutian of dn. and arthropods during the AUR Philos. 3. 335B: 29-165. MIR B. W. € G. D. Price. indic ators of Mesa palaeoclimate. Philos. Trans, Ser. 341B: 225-2 — o and Mesozoic. Trans., Ser. | 1993. Sedimentary facies as Ser lior R. € C. W. Toled. 1992. Characterizing the most primitive seed ferns. I. A reconstruction of Elkinsta aa Int. da PI. Sci. 153: 602-621. Shanley W. & P. J. McCabe. 1994. Perspec tives on the sequence — of continental strata. Amer. Assoc. Petrol. ( Bull. 78: 544-508 seol. Farliest evidence for dayi: Early implications on emt habit. J. Pl. Res. 107: —— — € ——— 206 Spicer, R. A Society Special Publi ——.. 1989. correspondence ps. of phylogeography and ur signals. F ^p. Shugart, H. H. 1997. Plant and ecosystem functional types. Pp. 20-43 in T. M. Smith, H. H. Shugart & F. Il. Woodward (editors), Plant Functional Ecosystem Properties Types: Their Relevance to and Global Change. Cambridge Jniv. Press, Cambridge Silvertown, J., M. D odd. s onway. 2006. Phylogeny and the hierarchical organi- on of Dus diver sity. Ecology 87(7)(suppl.): 539-549. EE 8C 199]. Th ne guild concept and the structure a ers 'al communities. Ann. Rev. Ecol. Syst. 22: 115-143. Sims, H. J. 2000. Diversity, Turn and Seed Size Evolution in the Late Paleozoic Radiation of Seed Plants. Unpublished dissertation, Skog, J. E. 200 erns re P to extant Gowing, C. Lawson & — `. Dayan. over, University of Chicago, Chicago. E . Biogeography of Mesozoic leptosporangiate t ferns. Brittonia 53: 236—269. & D. L. Dilcher. 1994. Lower vascular ET E the Jakota ins in n and Nebraska, USA. Rev. Mio Palynol. € Sloan, L. C. & E. J. ek Ear wil Geology 18 Smith, A. V. H. 1962. The peals is on miospores -18. 1990. “Equable” elimates during : 489-492, paleoecology of Carboniferous ; T bios of bituminous or Geol. $ 1 J. V Sie coals. Pv 2346: Smith, F. 004, Modem calibration of um hdi isotope signatures for C3 aleograss- land reconstruction. Palaeogeogr. Palae M den Palaeo- ecol. 207: 271-304. Smith, T. M., H. H. Shugart € F. I 1997. Plant Functional Types: Their Relevance to Eco- system Properties and Global Change. Press, 7 ambridge. Woodward (e ditors). Cambridge Univ. Speck, T. 1994. A biomechanical method to distinguish between self-supporting and non-self-supporting fossil is Rev. Palaeobot. Palynol. 81: 65-82. v N. P. Rowe. E Bome ek al analysis of Pitys eed plant vegetative MD and its 443—400. ' and secondary growth processes in e A summary of f the methodology and new data from an early lignophyte. Philos. Trans., Ser B. 358B: 1473-1485. 3. Modelling primary . 1988. Quantitative Va of plant megafossil menlis . Pp. 29-51 in W JiMichele & S. L. Wing (editors), Methods b Applic ations of Plant Paleoecology. 3. ications No e formation and ua E of plant fossil mb s. Advances Bot. Res. 16: 96-191. & C. R. Hill. 1979. Panna ida: and quantitative dala Jurassic plant bed. Rev. Palaeobot. Palynol. 28: 273-299 ——— & J. A. Wolfe. 1987. in Trinity (Clair Engle California. Pa levbiolgy 13: 227— , P. M. Rees, J. L. Plant taphonomy of Late Holocene deposits i :) Lake, northern Chapman, " A. Jarzembowski & D.C avt 1993. Cretaceous pare ch and climate 3. : 277-2 signals. Philos. Trans., Ser. B. 341 & J : en 69-78 in J. R. Sellwood, R. A. Spicer & P. J. s Eu m and Their Modelling. C d and Hall, London. & A. G Cin 3. J. Hoskins, B. care E = v 1996. The Some insights. Pp. . Herman. of Asia: S in A. Sahni (editor), Cretaceous Stratigraphy and diera vironments. Geological Society of India Memoir Vol. 37. é Cretaceous 405—43: vegetation and climate Volume 95, Number 1 2008 DiMichele 8 Gastaldo Plant Paleoecology in Deep Time 195 . A. Ahlberg, A. B. Herman, S. P. Kelley, M. umor & P. M. Rees. 2002. Pa Pn bud ecology of the Middle Cretaceous Gre eae Asia. Palaeogeogr. RE Palaeo- —105. ecol. 184 de B Henan : K M. Kennedy, T Climate Leaf a ce Program (CLAMP) ad the cold month mean temperature. J. Geol. 112: 685-70! 05. The essensilivity of m ou Foliar to Boone loss of foliar physiognomic characters. Palaios 20: 429—438. Staub, R. . Gastaldo. 2003. Late s incised- valley fill and deltaic Era in the Raja ng River Delta . Nummedal, Imbert, H. E did Topical Deltas of ed t Asia—Sedimentology, Stra- tigraphy, and Petroleum Geology. SEPM Special Publica- tions 76. Society for Sedimentary Geology, Tulsa. Stebbins, G. L. 1965. The probable growth habit of the earliest flowering alanis Ann. Missouri Bot. Gard. 52: 457—468. Stidd, B. M. & T. L. Phillips. 190 8. Basal stem anatomy of a Amer. J. Bot. 55: 8 40. Stockey, A. e ll. 2003. Anatomically ipu a (Williamsoniaceae) Eviden Bennettitalean reproduction in the Late Cretaceous of western North America. Int. J. Pl. Sci. 164: 251-262. K. 2 : The origin and e nce for & W. A. DiMichele Rd eae Terrestrial Ecosystems. Paleontological Society Papers, Vol. 6 Strömberg, C. A. E. 2002. The origin and spread of grass- ( e - SCOSySIeUis in the late Tertiary of North America: Preliminary results concerning the evolution of i Palaeogeogr. Palaeoclimatol. Palaeoecol. 177: 9-7 Using jore assemblages to reconstruct the origin bs spread of grass-dominated habitats in the Great Plains of North America during the late Eocene to early Mioc ‘ene. Palaeogeogr. Palaeoclimatol. Palaeoecol. 207: 239— . 2005. Decoupled taxonomic ar and ecological ji of open-habitat grass e Cenozoic of North erica. Proc. Natl. Acad. Sci. USA. 102: 11980-11984. Sues, H.-D. & R. R. Reisz. 1998. Origins and Er. evolution of herbivory in tetrapods. Trends Ecol. E Sun, G., Q. Ji, D. L. Dilcher, S. Zheng, K. E Nixo urs 2002. Archaefructaceae, a new basal angiosperm mily. Science 296: 899-904. ollinson, C.-S. Li, Y. Wang & D. J. Beerling. 2002. Quanta reconstruction of palaeoclimate from the Middle Miocene Shanwang flora, eastern China. Palaeogeogr. Palaeoclimatol. Palaeoecol. 180: 315-329. Tabor, N. J. & I. P. Shifts in late Paleozoic ad spheric circulation over western equatorial Pangea: nsights from pedogenic mineral delta "O compositions. b 30: 1127 Es P. Montañez. 2002 r & D. Schwindt. 2007. 6C julie: of be E across the Permian—Triassic boundary in the Karoo Supergroup (South Africa) reflect a stinking sulfurous swamp, not perpe CO». geogr. Palaeoc limatol. Palaeoecol. 252: 370—381. Taggart, R. E. & A. T. n 1980. d change in the Miocene Súler © Palae- Creek flora of rios and | aho: À case n D. L. Dilcher & Taylor (editors), id of Fossil Plants. Dowden, Hutchinson, and Ross, Stroudsburg, Pennsylvania. study in paleosuccession. Pp. 1 Flora of . 2000. và Unique North American ón Univ. Carol. Geol. s upland vegetation: val perspectives. Acta Sate ai : periodic volcanism on a oce 1982. The effects of ne vegetation distribution in eastern Ore and southwestern Idaho. Proceedings of the Third North American Paleontological Convention, Montreal 2: 535- 40. Taylor, Hickey. 1992. Phylogenetic evidence for T herbae 'eous origin of the angiosperms. Pl. Syst Evol. 7-156. Taylor, ^ ae . Taylor & N. R. Cúneo. 1992. The present is not the key to the past: A polar forest from the Permian of Antarctica. Science 257: 1675-1677 d T. N. 1965. Paleozoic seed studies: A monograph of > American species of Pachytesta. Palaeontographica : 1-46. 1178 . 1978. The ultrastructure and reproductive signifi- cance of Monoletes (Pteridospermales) pollen. Canad. J. Bot. 56: 3105-3118. . Reproductive biology in early seed plants. _ Biosci ‘lence 32: 23-28. ry, ¿ E. Evanoff. 2006. The terrestrial Eocene- Olesen boundary revisited: A comparison of multi- proxy records of paleoenvironmental and paleoclimatic change. Geol. Soc. Amer. Abstr. Prog. 38(7): 201-203 The a Database. 2007. , 4 ee 2007. an: e oli acces 3. Carex graceii sp. n., Cyperocarpus p. os erocarpus Lerrestris sp. n., a sp. n. (C of e Amer. J. Bot Tidwell, . & S. on c. to oe Cretaceous eliasi s and Cyper- gee eae 2 from the Miocene y 435— 1 n review of selected ferns. J. Pl. Res. 107: 417—442. Tiffney, B. H. 1986. Evolution and seed dispersal raa according to the ies record. Pp. 273-305 in (editor), Seed Dispersal. 1 irray Academic Press, S dier: 9 Re o of the age of the Brandon Lignite ord USA) Pos on plant megafossils. Rev. bot. Palynol. 82: 299—315. Land o as food and habitat in the age of the dinosaurs. Pp. 352-370 in J. O. Farlow & M. K. Brett- Surman (editors), The Complete Dinosaur. Indiana Univ. ?ress, Bloomington. 1999 Palaeo Fossil fruit. and a flora from the early Eocene Fisher/Sullivan site. Pp. 139-159 in R. E. Weems G. J. Grimsley (editors), Early pun Vertebrates and Plants from the Fisher/Sullivan Site (Nanjemoy Formation) Stafford County, Virginia. Virginia Division Mineral Resources Publication Vol. 152. Virginia Department of Mines, Minerals, and Energy, Richmond. . 2004. Ve rtebrate i ied of seed plants through Ecol. Evol. Syst. 35: 1-29. R.M OVER 2001. The use of geological and paleontological evidence in evaluating plant dea graphic hypotheses in the northern os Tertiary Int. Sci. 162: S3-817 yas A. 1988. Paleopaly solo Allen and Unwin, London. (editor). 1994. Sedimentation of Organic Particles. ds e Abe Cambridge. Tschudy, R a Pillmore, C. J. Orth, J. S. Gilmore & J. Disruption of the terr estrial plant Cretaceous-Tertiary boundary, Interior, Science 225: 1030-1032. Uhl, D. & V. Moa 1999. Leaf venation density as a limat tal proxy: A critical review ien new data. ONE SEDAN e ue col. : 15-26. Western Annals oa Eod Garden — ———, S. Klotz Traiser, €. Thiel, T. Utescher, E. Kowalski & D. tures and lea . Dilcher. 2007. Cenozoic paleotempera- hysiognomy—4A European perspective. PORE. Palae 'oclim Wwe rer 'oecol. 248: 24-31. r T. Utescher. 2003. using leaf floras: Case margin. analysis. and Palynol. 126: , V. Mosbru igger, A. Bruch ] Reconstruction | studies for a anaon of leaf the coexistence approach. Rev. Palaeobot. 49—64. U i en h, G. R., Jr, B. I to-Bliesner & C. Scotese, 98. V ege tation- prd inleractions and their role in i warming during i latest Cretaceous. Philos Trans., Ser. B. 353B: 97-11 & ————. 1999. during = latest ‘Terrestrial vegetation and its effects on _ climate Cretaceous. Pp. 407-435 The Evolution ‘of E. Barrera € €. €. Johnson (editors). Cretaceous Oce e limate pele ms. Geologie al Soci ie P of America Special Papers : A. R Ashraf. 2000. Terrestrial climate evolution in northwest Germany over Utescher, T., Mosbrugger & A. R. the last 25 pee years. Palaios 15: 4304 Valentine, J. W. taxonomic Serer of diversity in higher I-45 de . Hubbard. Lower a rous in categories. T d O: : J.. M & R. LI van der Zwan, C. Boulter 1985. Climatie c duh during the | Euramerica, based on multivariate. statistic 'al analysis of xalynological data. Palaegeogr. Palaeoclimatol. Palaeo- : geog -20. van Konije nburg-v van Cittert, J. H. A. 2002a. Ecology of some Jurassic ferns Palynol. 199: | Eurasia. Rev. Palaeobot. 113-124. . 2002b. Ecology of some Late Tria to Early Cretaceous ferns in Eurasia. Rev. Palaeobot. pue 119: 113-124. van Wagoner, J. C., R. M. Mitchum, K. M. Campion & V. 1990. ae ‘iclastic sequence stratigraphy 1 Ra a American Association " Methods in well logs, and outcrops. cores, Petroleum Geologists Exploration Series, No. 7. Veizer, J., D. Ala. K. Azmy, P. Bruckschen, Buhl, Bruhn, G. A. E Pio. i: Min S: e Y: Mp. T. Jasper. C. vellek, € . Podlaha veg. o e Chem. Geol. I & H. Strauss. 1999. Phane TOZOIC seawater. ri G. J. 1987. E so nts of 59-88. = E TUMOR and ^» 'alation. Princeton . Press, Princetor Un Wagner, R. Sporangiostrobus fossilized by volcanic ash fall H. 1989. E e Stephanian forest o with in the RUE ertollano Basin, central Spain. Int. J. Coal Geol. 12: Walker, T & J. W. Valentine. of S species dive m and the number of empty Naturalist 124: 887—899 Vegetation of ded 1984. Equilibrium models niches. Amer. Walter, H. 1985. Systems of the sá and Ecological Geo-biosphere, 3rd ed. Springer-Verlag. New York. Wang, Y. 2002. Fern ecological implications from the Lower Jurassic in western Hubei, China. Rev. 119: 125-142. Wang, Z.-Q. & A. 2001. lycopsids and dieback of the forest vegetation in relation to Palaeobot. Palynol. -S. Chen. Traces of arborescent the terminal Permian mass e a tion in North China. Rev. Palynol. 117 7-244. 2000. E d the phylogenetic structure of Palaeobot. Webb, C. O. ecological communities: An example for rain forest trees. Amer. Naturalist 156: 145-155. Ackerly, A. McPeek & M. J. Donoghue. 2002. P nies pa community ecology. Ann. Rex Ecol. Syst. 33: 475-505. J. B. Losos & A. A. Agrawal (editors). 2006. a a into community ecology. Ecology 8T(7) (suppl): 51-52. Wehr. W. C. & S. R. Manchester. 1996. pn WA significance of Eocene flowers. fruits. pr eds fro TM Washington. Washington Geol. 2 :25-21. Wehrmann. A.. G. Hertweck, R Brocke,- " Jansen, P. cónigshof, G. Plodowski, E. Schindler. V. Wilde, A. Sel Blieck € S. Se a 2005. Paleoenvironment of an Early land-sea transition: A case study from the Old Red Continent (Mosel Valley, Germany). Palaios 20: Eme ‘el & P. K j Assembly rules as general constraints on aint cd Pp. 251-271 in E. Weiher & P. Perspectives, na. es, Retreats. Cambridge Univ. Press, eddy (edito rs), Ecological Assembly Rules: C e st. . J. H. Brown & B. J. E iu 1997. A general model for the origin of allometric scaling laws in biology. Science 270: 122-126. Westoby, M. 2006. Ad netic ecology at World scale, a new fusion Ecology between ecology and evolution. 7(7)(suppl.): 5163-8 )5. & M. a wishman. 1997. Categorizing plant species . M. Smith, H. H. Plant. Functional Hn i o and into functional ly pes. Pp. 104—121 in E. IL ar R G dobsd C A C RIDES Unis Whittaker, habitat, 2 ide Amer N Wieland, G. Carnegie Institute of Washington, 9 . Root. 1973. rg he. d 107: 321-3: Fossil Cycads, pre 2: Publication No. 34. and pseudomonocarpy in the Amer. . 8: 218-230 sy R. . Levin & -- 1916. American 1921. Monocarpy cadeoids. jot ~ — Wiemann, ] Manchester, D. L. Hinojosab & E, A 1998. Estimation of temper- ature and prec pem from morphological characters of J. Be 796-1802. . Wheeler. dicotyledonous leaves. Amer. 1.85: —18( Wiens, J. J. & M. pay. ecology and species richness. Trends Ecol. Evol. —644. W " a 1997. case for le al mar, gin analys J. Donoghue 2004. Historical oo When are leaves good thermomete a A new 73-390. . Paleobiology 23: ——.. 2000. Late Paleocene—e 2 Eocene c bis chang- es in southwestern Wyoming: Paleobotanical analysis. Geol. Soc. 2 Bull. 112: 292 a, & K. R. Johnson. 2004. Land plani extinction at the end of the Cretaceous: A quantitative analysis of the North Dakota na rec cord. alesis 30: 347-308. . Beard, K. S. I s-Vollum & J. W. Norejko. 10083. a of a Late a (Early. Clarkforkian) terrestrial ecosystem: Big Multi Quarry and associated p W eats Basin, southwestern Wyoming. Palaios 13 — Wi ing, D. R. Greenwood & C. L. Greenwood. 1998). "D sing du ie as pale epi indica- tors: An Eocene example. Geology 26: s : ———— & ; x": fossil leaves as pi Vica DE sei An Eocene D. le Re ‘ply. Geology 27: oC. D. Coley & A. D. Cutter. . Insect he ee plant defense, 25 early Cenozoic ipe change. Proc. Natl. Acad. Sci. U.S.A. 98: 6226 622 1- ». Labandei ‘ira, K. . Johnson, P. Volume 95, Number 1 2008 DiMichele & Gastaldo Plant Paleoecology in Deep Time 197 Willard, D. A. 1989a. Source plants for Carboniferous microspores: Lycospora from permineralized Lepidostrobus. Amer. a e 76: 820-827. 39b. Lycospora from Carboniferous Lepidostrobus U Amer. J. Bot. 76: 1429-1440. “Vegetational puta rns in the Springfield Coal (Middle Pernsuo anian, Illinois Basin): Comparison of miospore and coa ar led Pp. 139-152 in C. Ceci J. Cobb (editors), Modern and Ancient Coal- Forming Environments. Geological Society of America Special Papers 2 ——, T. L. Ph isa W. DiMichele. 2007. TE nian-age Calhoun coal bed and ue: for long-term dynamics of wetland ecosystems. Int. J. Coal Geol. 69: 21-54. Williams, C. H. Johnson, B. A. LePage, D. R. 1 K nd EN Reconstruction of Tertiary Nl forests: I. Test of a method for biomass determination bci on stem dimensions. Pale »obiology 29: 212-237. eda. 2003b. forests: H. Lesnikowska & id of the Late Pe = Ter tia biomass and modus vily o Arctic. RON. of Structure, plain. forests 238-21 Wing, L. p. & I. O. Buss. World e Society, S. L. 198 and cvclic Metasequoia cocene flood- in the Canadian Paleobiology 29: 1970. pM and Forests. Washington, Wing, Relation of da tation to geometry ‘ily x some fluvial carbonaceous deposits. Petrol. 54: : 88a. Eocene and Oligocene floras and vegetation of the Rocky Mountains. Ann. Missouri Bot. Gard. 74: 748-784 Se "diment. 5200. oO ES Taxon-free paleoecological analysis of Eocene _megalloras from Wyoming [abstract]. Amer. J. Bot. 75 — “Equable” climates during Earth history? Comnet e 18: 539—540. ——. 1 'ertiary vegetational history of North America as a context for mammalian evolution. Pp. 37-65 in C. L . Jacobs & K. Scott (editors), Psion of Tertiary Mammals of North America, Vo Carnivores, Ungulates, and Ungulate-like Memals; Cambridge Univ. Press, Cambridge. . DiMichele. 1995. Conflict between local and E ble: in pant diversity through geological time. Palaios 10: 551—564 R. Greenwood. Janis . Ek Terrestrial 1993. Fossils and fossil climate: The case for equable continental interiors in the Eocene. Philo s. Trans., Ser. B 43-252. . Harrington. 2001. warming at athe Paleocene/Eocene boundary and impli- = = Floral response to rapid cations for concurrent faunal change. Paleobiology 24: 539—562. Hickey. 1984. The Platycarya perplex and the evolution of the Juglandaceae. Amer. J. Bot. 71: 388411. ——— & B. H. Tiffney. 1987. The reciprocal interaction of angiosperm evolution oy tetrapod herbivory. Rev. Pala eobot. Pals 50: 179-210 ichele, T. L. Phillips: R. Taggart, B. H. Tif ho & 5, J.I | vs azer. dos RA characterization of fossil pam Pp. 139-180 in A. K. Behrensmeyer, J. Damuth, W. A. DiMichele, R. Potts, H.-D. Sues & S. L. Wing UN Terrestrial Ecosy stems Through Time: Evolutionary Paleoecology of estrial Plants and Animals. Univ. of Chicago Press, Chie ago. Gier low ski, W. C. l a J. Hickey & C. C. Swisher. 1993. Implications of an exc meee fossil flora for Late Cretaceous vegetation. Nature a suas 344. E , F. A. Smith, J. I. Bloch, Buyer e K. H. E reeman. (2005. i T eae 2 rapid po Warming at the Paleocene-Eocene boundary. Science 310: 993-906. Wnuk, C. ioe The development of floristic provinciality during the Middle and Late Paleozoic. Rev. Palaeobot. Palynol. 90: 540, W. PUMP paleoecology of Middle arr ‘ing 1984. The life habits and Pennsylvanian pteridosperms based on an in situ pedes from the Bernice Basin (Sullivan ed P U.S.A.). Rev a 4]: . Palaeobot. 7. A Pennsylvanian-age terrestrial Using plant foss the history and process of sediment accumulation. J. Sedi- ; Petrol. 51: 212-221. Wolfe, J. A. 1979. Temperature parameters of humid to mesic forests of « storm dams ils to characterize 'astern Asia and relation to forests of other regions of the northern. hemisphere and Australia. U.S. Geol. pu Prof. Pap. 1106: 1-37. 993. A method of obtaining climatic parameters from leat semble U.S. Geol. Surv. Bull. 2040: 1-71. . Upchurch. floral iin anges - Nature 324: & 198 6. Vegetation, climatic, and the Cretaceous—Tertiary 152. boundary. ——— ———— American climates and vegetation during He Cretaceous. Palaeo- geogr. Palaeoclimatol. Palaeoecol. 61: 33-77. Wright, I. J., P. B. Reich, M. EU. D. Ackerly, Z 1, F. Bongers, d. Cavender-Bares, T. Chapin, J. H. C Ciel MR M. lexas, E. Garnier, P. K. Groom, J. Gulias, K. Hikosaka, B. B. Lamont, T. Lee, W. Lee, C. Lusk, J. J. Midgley, M.-L. Navas, Ü. Niimemets, J. A. Poorter, P. e L. Prior, V. L Pyankov, C. Roumet, S. C. T ». Tjoelker, E. J. Veneklaas & I > ed 27 1987. North non-marine j = = economics spectrum. Nature ve 21-827. Xiang, (J.) Q.-Y., . er, D. Thomas, W. H. Zhang & C. Z. Fan. 2005. Phylogeny, dE a molecular ane of Cornelian Cherries (Cornus, Cornaceae)—Track- ing eu plant migration. Evolution 59: 1685-1700. Thomas, W.-H. Zhang, S. R. Manchester & Z. Es pos Species level phylogeny of the Dogwood Cornus (Cornaceae) molecular and morphologic ‘al evidence—Implication in taxonomy and ertiary PEN migration. Taxon 55: 9-30. Yapp, C. J. & H. Poths. 1996. Carbon isotopes in continental ved environments and variations in ancient atmo- Manches genus based on spheric CO» pr 'ssure. Earth Planet. Sci. 71-8: Zachos, J. Ca Rohl, S. A. Schellenberg, A. Shui S, 5 A Hodell, D. C. Ke lly, E. Thomas, M. Nicolo, I. Raffi, L Lourens, H. ar & Kroon. 2005. Rapid acidification of the ocean during the Paleocene-Eocene thermal maximum. ous : 308: 1 615. Ziegler M. "hytogeographie patterns and conti- nental e during the Permian period. Pp. 363— 9 in W. S. McKer row & Scotese (editors), Paleozoie Pulsa n and Biogeography. Geological Soc ety of L ondon Memoir 1; "A. . Bambach, J. T. Parrish, . Barrett, E. H. Parker, A. Rayona: i J. Sepkoski Jr. 981. Paleozoic biogeography and climatology. Pp. 231-266 in K. J. Niklas (editor), Paleobotany, Paleoecology Evolution, Vol. 2. Praeger Publishers, New York. and 198 Annals of the Missouri Botanical Garden , A. Raymond, T. C. Gie srlowski, M. A. Horrell, D. B. Ee & A. L. Lott 7. Coal, climate and terrestrial produc tivity: The Peren and Early Cretaceous compared. Pp. 25-50 in A. C. Scott y ditor), Coal and Coal-bearing Strata: Recent Advances. Geological Society of Lonc Special Public sation „J.N a on 1. Parrish, Jiping C. Gyllenhaal, D. B. E ch A. Bekker & M. L. E Mesozoic phytogeography | and s Trans., Ser. B. 341B: 297-305. H Rees, D. B. M. L. S 1996. \ Constraints from fossil ctonics and paleomag- netism. Pp. 371—400 in A. Yin b M. Harrison (editors), The Tectonic Evolution of Asia. Cambridge Univ. Press Cambridge climate. P. Bekker, assembly of Asia: Zwieniecki, . Hulver & D. B. Rowley. 1997. Permian world Tes and climate. Pp. 111-146 in I. P. Martini (editor), Late Glacial and Postglacial Environmental C uio Quaternary, Carbonife a ic. Oxford Univ. Press, » Oxford. . V. Naugolnykh. 2002. The Early Panis Mons of Prince Edward Island, Canada: Differ- entiating en nem loc al effects of climate change. ( J. Earth Sci. : 38. ———, G. Eshe d P. M. Rees, T. A. Rothfus, D. B. Rowley & D. nda: 2003. Tracing the a across land and sea: a to present. Lethai Au C. , Proterozo- . Rees & S van. : 227-254. Boyce & N. M. Holbrook. 2004. ene p space of single veined leaves: Role of issue hydraulic propere. in constraining leaf size and shape. Ann. Bot. 94: 507-513. Erratum Watanabe, K., T. Yahara, G. Hashimoto, Y. Nagatani, A. In Appendix | on p. 653, under Eupatorieae, the Soejima, T. Kawahara & M. Nakazawa. 2007. Chromosome second accession for Chromolaena pedunculosa numbers and karyotypes in Asteraceae. Ann. Missouri Bot. (Hook. 8: Arn.) R. M. King & H. Rob. was erroneously Gard. 94: 643-654. d a HU DC = dme listed as 2n = 2011 + 101. The correct accession is 2n = ]OII + 101. The aut — 10rs regret the error. doi: 10.3417/2008009 ANN. Missouni Bor. Garp. 95: 199. PUBLISHED ON 11 APRIL 2008. Volume 95, Number 1, pp. 1-200 of the ANNALS or THE MISSOURI BOTANICAL GARDEN was published on April 11, 2008. Missouri Botanical Garden Lib A 3 1753 00350 n www.mbgpress.org CONTENTS Paleobotany in the Post-genomics Era: Introduction William L. Crepet & Maria A. Gandolfo 1 The Fossil Record of Angiosperms: Requiem or Renaissance? ——— William L. Crepet 3 Selection of Fossils for Calibration of Molecular Dating Models Maria A. Gandolfo, Kevin C. Nixon & William L. Crepet 34 Paleobotany, Evidence, and Molecular Dating: An Example from the Nymphaeales |... : Kevin C. Nixon 43 Hide and Go Seek: What Does Presence Mean in the Fossil Record? |. Robyn J. Burnham — 51 W(h)ither Fossils? Studying Morphological Character Evolution in the Age of Molecular Sequence Elizabeth J. Hermsen & Jonathan R. Hendricks — 72 The Whole id the Parts: Relstianshipe Between Floral Architecture and Floral Organ Shape, and Their Repercussions on the Interpretation of Fragmentary Floral Fossils Peter K. Endress 101 A Fossil Record for Growth Regulation: The Role of Auxin in Wood Evolution — = Li. TAPA E Gar W. Rothwell, Heather Sanders, Sarah E. Wyatt & Simcha Lev-Yadun 121 Phylogeography, Fossils, and Northern Hemisphere Biogeography: The Role of Physiological Uniformitarianism Bruce H. Tiffney 135 Plant Paleoecology in Deep Time - William A. DiMichele & Robert A. Gastaldo 144 Erratum for Chromosome Numbers and Karyotypes in Ásteraceae Kuniaki Watanabe, Tetsukazu Yahara, Goro Hashimoto, Yoshimi Nagatani, Akiko Soejima, Takayuki Kawahara & Miyuki Nakazawa 199 Cover illustration. — Archaefructus reconstruction by David L. Dilcher and K. Simons, Flor- ida Museum of Natural History. Annals of the Missouri Dotanical Garden T VG olume 95 umber Annals of the Volume 95, Number 2 008 Missouri Botanical Garden June 2 The Annals, published quarterly, contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, independent reviewers. Authors should write the Managing Editor for information concerning arrangements for publishing in the Annals. Instructions to Authors are printed in the back of the last issue of each volume and are also available online at www.mbgpress.org. Editorial Committee Victoria C. Hollowell Scientific Editor, Missouri Botanical Garden Beth Parada Managing Editor, Missourt Botanical Garden Allison M. Brock Associate Editor, Missouri Botanical Garden Barbara Mack Editorial Assistant, Missouri Botanical Garden Patricia M. Eckel Latin Editor, Missouri Botanical Garden Ihsan A. Al-Shehbaz Missouri Botanical Garden Gerrit Davidse Missouri Botanical Garden hoy E. Gereau Missouri Botanical Garden Peter Goldblatt Missouri Botanical Garden Gordon McPherson Missouri Botanical Garden Charlotte Taylor Missouri Botanical Garden Henk van der Werff Missouri Botanical Garden For subscription information contact ANNALS or THE Missouri BOTANICAL GARDEN, 96 Allen Mar- keting & Management, P.O. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2008 is $170 per volume U.S., $180 Canada & Mexico, $205 all other countries. Four issues per volume. The journal Novon is included in the subscription price of the Annals. annals(? mobot.org (editorial queries) http://www.mbgpress.org The A ] i-r d Uy THe ANNALS OF THE Missouri BOTANICAL GARDEN (ISSN 0026-6493) is published quarterly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. Periodicals post- age paid at St. Louis, MO and additional mail- ing offices. POSTMASTER: Send address changes to ANNALS OF THE Missouri BOTANICAL GARDEN, o Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. indexed in AGRICOLA (through 1994), APT Online, BIOSISO, CAB Ab- stract/Global Health databases, ingenta, ISI® databases, JSTOR, Research Alert®, and Sci Search®. The full-text of ANNALS or THE Missouni BOTANICAL GARDEN is available online though BioOne™ (http:// www.bioone.org). ( Missouri Botanical Garden Press 2008 The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and their environment, in order to preserve and enrich life. © This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Volume 95 Annals Number 2 œ of the 2008 7 ISSOUTI we? y ws . s us Botanical NS arden ASSUMPTION 0 ANALYSIS: Daniel R. Brooks? and Marco G. P. van Veller? COMPARATIVE PHYLOGENETIC STUDIES IN THE AGE OF COMPLEXITY" ? ABSTRACT arwin's panoramic view of biology enc ompassed two metaphors: the phylogenetic tree, pointing to relatively linear (and divergent) complexity, and the tangled bank, pointing to reticulated (and convergent) complexity. The emergence o phylogene tic systematics half a century ago made it possible to investigate linear complexity in biology. Meno sl O, first 986, is not needed for cases of simple evolutionary patterns, but must be invoked when there are complex k is reticulated iis itionships. A corollary of Assumption 0, the duplication convention, matic ontology to be used in discovering reticulated evolutionary “= s in comparing trees (PACT), was developed sj »ecifically for use in analyses invoking Assumption 0. PACT can help discern complex evolutionary explanations for historical biogeographical, coevolutionary, phylogenetic, and tokogenetic processes Key wt Assumption 0, er cladogeny, coeva evolutionary complexity, horizontal transfer. hybridization, PACT, PM. analysis for comparing trees, phylogenetic systematics, phylogeography, reticulation, taxon pulse. tokogeny. ! This and the three articles that follow it are the proceedings of the 52nd Annual Systematics Rea UN of the Missouri Botanical Garden, “Reconstructing Complex Evolutionary Histories: Gene-Species Trees, Historical Bi ogeograpin and Coevolution." The symposium was held 7—9 Oc ober 2005, at the Missouri Botanical Garden in St. Tu Missouri, U.S.A. “This was the 50th Missouri ee ‘al Garden Annual Systematics Symposium to be pu d by a grant from fe Te Science Foundation d DEB-0515933). The meeting was organized and moderated by Richard Mayden of Saint Louis University; P. Mick Richardson was responsible for the smooth running of the symposium, with the welcome and able o of Alina Freire-Fierro, Mary McNamara, Sandy Lopez, Donna Rodgers, William Guy, and Zubin o Victoria . Hollowell, Beth Parada, and Allison Brock (Missouri Botanical Garden Press) were responsible for the publication of the sy MD um proceedings. "This contribution is based on a presentation made by DRB at the 2005 Annual Systematics Symposium of the Missouri Botanical Garden. We thank Rick Mayden for the invitation to iim ipate in 1 that symposium, as well as Peter Raven, Mick appreciation to Marco Roos, E. O. Wiley, and Richardson, and the other organizers of the symposium. We also Natural Sciences and Engineering res Rino Zandee, who first articulated Assumption 0. DRB eM finding d the | Research Council (NSERC) of Canada. ‘Department of Zoology, University of Toronto, 25 Harbord Street, Toronto, Ontario M55 365, Canada. dbrooks@zoo. uloronto.ca. ° Biosystematics Group, Wageningen University, Generaal Foulkesweg 37, 6703 BL Wageningen, The Netherlands. marco.vanveller@wur.nl. doi: 10.3417/2006017 ANN. Missouri Bor. Garp. 95: 201-223. PUBLISHED ON 18 JUNE 2008. 202 Annals o Missouri Botanical Garden The universe is structured by laws, and science is the search for methods of data analysis and theories roviding powerful general explanations, all in terms of I J £ I — those general laws. This ontology of simplicity has guided the development of western science for nearly 2500 years, embodied in the principle of parsimony (Latin Aristotle (350 B.C.E.) articulated the ontological view of the principle of — oe ^^ parcere, to spare). parsimony, the postulate that “nature operates in the "the is always preferable." shortest way possible” and more limited, if adequate, The principle of parsimony is also linked with the English philosopher and Franciscan monk William of Ockham (ca. 1285- 1349), who advocated the use of what is known as “Ockham's neccesitate” (“plurality should not be posited without razor”: “pluralitas non est ponenda sine necessity”) and “non sunt multiplicanda entia praeter necessitatem” (“entities should not be multiplied unnecessarily”). In this sense, the principle of simplic- ity obliges us to favor theories or hypotheses that make the fewest unwarranted, or ad hoc, assumptions about the data from which they are derived. This epistemo- = logical use of the principle does not necessari y imply that nature itself is parsimonious. Indeed, despite the best efforts of philosophers for more than 700 years, no link truth established. Nonetheless, most scientists conduct their between parsimony and has ever been research as if they believe that nature is parsimonious in some sense and, as a consequence, they place much more credence in simple than in complex theories. The cosmologist Stephen Hawking has dubbed the 21st century the century of complexity, leading a parade of physicists who began to think about ontological complexity in the latter decade of the 20th century. Biology should be well poised to take — advantage of this sea change in our understanding o the discovered an ontology of complexity called Darwin- Few fu that Darwinism is not a simple theory: the nature of science and universe, having ism almost 150 years ago. ly realize there y od and the nature of the conditions. The former are two factors: namelv, the nature of the seems lo be much more the important; for nearly eM variations sometimes arise unde FT. as far ds WC can judge, dissimilar conditions; and, on the other hand, dissimilar variations. arise. under conditions 5 which appear to be nearly uniform: —Darwin (1872: 32) — Most evolutionary biologists consider this passage no more than a general repudiation of Lamarckism. We believe, however, that it is one of the first modern articulations of a complex theory: Darwin proposed that evolution is an emergent property of asymmetrical interactions between two different causal agents, each with its own properties, producing outcomes that are not readily predictable from knowledge of the properties of either agent and, thus, historically contingent. Darwin thought that organisms were historically and developmentally cohesive wholes, and it was in the "nature of the organism" to produce offspring that were all highly similar (but not identical) to each other and to their parents and other ancestors. He also that without regard for environmental postulated reproduction produced variation conditions, and it was in the "nature of the organism" to produce these offspring in numbers far exceeding the resources When overproduction produced variety in available for their support. this inherent critical charac- ters, natural selection would preserve the versions that were functionally superior in that particular environ- mental context (adaptations). Whenever an environ- ment changed, those organisms that already had the adaptations necessary to survive would do so, whereas — — aose lacking appropriate adaptations would not. The production of organismal diversity thus required that organisms be at once autonomous from, and sensitive o, the environment, another example of complexity. Darwin visualized his viewpoint about the com- plexity of evolution with two metaphors, the phyloge- netic tree and the tangled bank. The phylogenetic tree — points to complexity arising from irreversible phe- nomena. By referring to species as “communities of 39 descent” and placing them in a single “tree of life, Darwin emphasized that the fundamental explanatory The contrast, points to complexity arising from biological principle is shared history. tangled bank, by associations that do not share the same temporal and spatial origins, producing reticulated (tangled) pat- terns of descent and association. By the end of the 19th century, most biologists had adopted the view that evolution had occurred, but many were uncomfortable with Darwinism’s lack of simple general laws, and the discovery of the “Laws of Inheritance” by Mendel did little to quell this unrest. Bowler (1983) documented the eclipse of Darwinism 1890-1940. time were neo-Lamarckism and during the period of The most influential theories during that orthogenesis, both theories of simplicity. As a theory of environmental determinism, neo-Lamarckism placed causality solely in the nature of the conditions. that the evolution, Orthogeneticists, by contrast, asserted environment played little or no role thereby placing causality solely in the nature of the organism. The eclipse of Darwinism is assumed to have ended with the emergence of neo-Darwinism in the 1930s. mechanism, natural selection. Buoyed by advances in Neo-Darwinism focused on Darwin’s key understanding the mechanisms of genetic inheritance Volume 95, Number 2 2008 Brooks & van Veller Assumption O Analysis and some elegant experiments, this became the dominant theory for evolutionary biologists during the second half of the 20th century. fully expected neo-Darwinism to Some founders encompass the panorama of phenomena encompassed in Darwin's original formulation: very part of the whole, wonderful history of in lif fe, SET ce modes and all the factors of evolution are some understanding is in our grasp, and we may trust our own powers to obtain more. —Simpson (1953: 393) this was not to be the case. Sadly, As noted by Gould (1983) using the rubric “the hardening of the synthesis,” during the latter third of the 20th century. neo-Darwinism maintained a narrow focus on popu- lation genetics, acknowledging other aspects of evolution but asserting that all of them were manifestations of population genetics and natural selection and, therefore, all appearances of complexity could be reduced to the sometimes computationally complex, yet ontologically simple, principles of population geneties. Like the neo-Lamarckians, neo- Darwinians focused their attention on the environ- mental conditions correlated with the survival of organisms, populations, and species. Thus, many explanations expounded by neo-Darwinians sound Lamarckian because the nature of the organism has been reduced to almost nothing. For many today, the nature of the organism is to be a blind watchmaker made up of selfish genes. Organisms exist, but they have no “nature” autonomous from their surroundings. This produces a theory that is parsimonious, vet lacking Darwin's panoramic view of biological diversity. During the past 25 years, a growing number of scientists and philosophers have called for modifica- tions of evolutionary theory (e.g., Eldredge, 1985, 1986, 1989; Brooks & Wiley, 1986, 1988; Maynard Smith & Szathmary, 1995; Brooks, 1998, 2000, 2001, 2002; Collier & Hooker, 1999; Brooks € McLennan, 2000 and references therein). Many of these calls seek to reintroduce Darwin's postulates about “the nature of the organism" and the “nature of the conditions” to develop analytical methods for integrating the tree of life and the tangled bank metaphors, and to establish fundamental guiding principles or null hypotheses for complex evolutionary phenomena occurring on differ- ent spatial and temporal scales. The phylogenetics revolution that began in the late 1960s 1966) entrenched beginning in the 1980s (for a review, see Brooks & and became (Hennig, McLennan, 2002) produced quantitative analytical methods for effectively documenting the complexity of the tree of life. Phylogeneticists discovered quickly that most phylogenetic tree complexity is manifested in a relatively small number of linearly organized historical correlations (which we consider to be the communities of descent) and, therefore, is (once again) computationally complex but ontologically simple. Methods developed for analyzing historical associa- tions described by tangled bank complexity assumed that such associations could also be adequately described and explained using phylogenetic tree— based reasoning (e.g., Brooks, 1985). Tangled bank complexity, however, produces reticulated historical relationships in addition to the relatively simple linear ones. We now know that methods of analysis based on phylogenetic tree-based reasoning produce internally Qo inconsistent results in direct proportion to the degree of tangled bank complexity in the data, an indication that those methods have oversimplified the data (van Veller et al., 1999, 2000, 2001, 2002, 2003; Dowling, 2002; Dowling et al., 2003). Wojcicki and Brooks (2004, 2005) developed an algorithm called phyloge- netic analysis for comparing trees (PACT) that permits phylogenetic systematic principles to be used in describing tangled bank type complexity. PACT is based on the assumption that tangled bank complexity is based on individual parts that each conform to relatively simple tree-like complexity which, when combined, may nonetheless produce of both relationships, including the possibility of reticulated complex patterns general and unique relationships. PACT uses three guiding principles: (1) Assumption O (Wiley, 1986, 1988a, b; Zandee $ Roos, 1987): in order to not oversimplify, you must analyze all input data without modification, and your final result must be logically consistent with all input data; (2) the duplication rule (Brooks € McLennan, 2002): methods of simplicity when the data result from Assumption 0 is violated with analyses by reticulated processes. Oversimplification can be avoided (i.e., Assumption O can be satisfied) in such — cases by duplicating entities with reticulated histo- ries; and (3) epistemological parsimony (Brooks « 2002): number of generalities supported by the data, entities McLennan, in order to find the maximum should not be duplicated beyond necessity. Therefore, we should make only enough duplications to satisfy Assumption 0. Underscoring its connection with the ontology of complexity, note that Assumption 0 more fundamental than parsimony. If evolutionary patterns were simple, Assumption 0 would never be violated, so we would never need to duplicate entities What seeking is the most general explanation possible and would never need parsimony. we are Annals of the Missouri Botanical Garden B C Figures | and 2. —l. pertaining to entities A, B, C, and D. Second from top = lop = | +} th hierarchical patterns aa rimpose d on each other. Y). —2. PACT r twe (Y = Y) —2. analys C. panem to entities A, B, D. " ue F. Se cond from from top = “boih hierarc de patterns superimpose "IN. Y( YN (AF) = YN AP: Y ( — 3 given the data (1.e., the ontology of complexity), not the ontology of simplicity) (see van Veller & Brooks. 2001; Brooks & MeLennan, 2002). What we have to euard against is the simplest explanation we can imagine (Le., the use of models and analytical methods that oversimplify the data in an effort to generalize. That is, simplicity is not always the most parsimonious depiction of the real world (van Veller & Brooks, 2001). PACT utilizes any data that can be represented as hierarchical strings, ranging from gene trees and other character state trees to species phylogenies. Wojcicki and Brooks (2004, 2005) recognized that nested (historic al) hierarchical patterns in nature are combi- nations of only different classes of SIX pattern modules. Analysis of a set of data is accomplished by selecting any one of the input data strings as the initial template. Next, the final analysis is built by sequentially aligning each additional data string with the template. Thereafter, combinations are made satisfy Assumption 0 using four combination rules. P ACT analysis for simple congruence. Top — the two hierare is al patterns aligi o Inc ongruent patterns. Toy j= > + YN (ED D> m NY NN y two hierarchical patterns based on different e but ed by common elements. Third from Bottom — PACT solution based on applying combination pen vo hierarchical patterns based on different data e top = the two hierare hical patte rns aligned by common elements. Third o — PACI ED); Y solution based on up e rule (B) + YN (C(DE) — YN (B(C The combination rules are illustrated in Figures 1—6 and explained briefly in the figure legends; for more details, see Wojeicki and Brooks (2004, 2005). Some workers seem to have assumed that PACT is an algorithm for the method called. Secondary BPA (Brooks & McLennan, 2002). A purported critique of PACT al. (2008) addresses shortcomings of secondary BPA overcome by PACT by Salvador Arias et that were specifically addressed by Wojcieki and Brooks (2004, 2005). Elucidating complex evolutionary patterns using PACT comes with a cost of sorts. Explaining the complexity of a set of observations can only be achieved by abandoning a priori expectations of mechanism. For example, is the presence of a species in area F in Figure 2 an indication of a unique event involving the production of a novelty in one of the data strings, or is it an indication of a unique event involving loss from the other? PACT produces the same result regardless of the interpretation, so no assertions or assumptions about processes are em- Volume 95, Number 2 2008 Brooks & van Veller Assumption O Analysis A D B C D E an id < gures A, B, C, D, and E. Seco Fig data bui pertaining lo entities top = male 2 (Y + N). Y (A) + YN (AD) = elements. d fron ] Hn Y N (B C DE))). Note that this example is precisely the same as the one shown in Figure 2, with the exception that in thi e find evidence of reticulated relationships. —4 ee al 4D. and hacia patterns = YNN). Y (A)N (BC hased on diff ferent data but p e to entities common elements. Third from top = on applying combination rule 3 (YN + interprel the data bedded in the algorithm. Similar reasoning can be applied to concepts such as lineage duplication, lineage sorting, and ancestral areas/hosts. All these concepts pertain to hypotheses of mechanism, and in order to maximize empirical robustness and avoid the possibility of circularity, all processes should be inferred following an analysis of data, not built into it. It is for this reason that PACT makes combinations beginning with the youngest members of the data strings being compared. This avoids the assumptions, unintentionally embedded in all previous P for all species associated today originated together at the S gm historical biogeography and coevolution, that same place and time; and (2) when there is ambiguity. we resolve the ambiguity beginning with the oldest (most basal) portions, amounting to an unacceptable assumption that the older the evolutionary event, the more accurate our information about it. PACT analysis is, therefore, only the beginning of a study, not an end in itself. Explaining the result of a PACT analysis requires reference to guiding princi- ples that are more comprehensive than those currently . PACT analysis for two incongruent patterns. Top nd from t both hierarchical patterns MU SR on each other. Bottom YN — Y Y alysis for tw > [es] e = two hierarchical patterns based on different top — Le two hierarchical pee rns aligned by common CT solution based on YN (ED): Y (B) + N(C(DE)) = N (AD): Y (E) + YN (ED) = is case vo congruent patterns. Top = two hierarchical patterns E. Second from top = the two hierarchic i: as rns aligne d by ~ I solution based superimpose d on eac h othe Zr. Bol tlom )+ Y (A) N (DE)) = YNN (A(BC)(DE)). (One should not over- in use. We illustrate this briefly for four different research programs in comparative evolutionary biology. RESEARCH PROGRAMS HISTORICAL BIOGEOGRAPHY Biogeography is the basis for understanding the communities, and the differential radiation of clades (Brooks & McLennan, 2002). In latter third of the 20th century, historical biogeography produced two simple origin of species, the assemblage of the and elegant theories that were based on an ontology of simplicity rather than complexity. — The first of these was the equilibrium theory of island biogeography (ETIB) (MacArthur & Wilson, 1963, 1967). dispersal from This theory is based on the view that ee , source" areas to "islands" (actual or metaphorical), mediated by island size and distance, produces linear log-normal species-area relationships. Noise in the system or the effects of contingency comprise in situ speciation and extinction. From this, 206 Annals of the Missouri Botanical Garden Figure 5. different data but pertaining to entities A, B, . and D. S elements. Third from top left = both hierarc i al patterns superim] Y-Y-Y) Note tha based on applying vue rule node. Bottom left = two based on applying Rond rule | one infers that data that conflict with the expected (the “law”) are the of historica contingencies, is therefore permissible t As a result, island biogeog- — pattern result ) and i remove or modify them. raphers are admonished to study small, young islands in order to minimize the potential for such historical contingencies that cloud our ability to see the true (and simple) pattern. simplicity was the maximum vicariance (Max Vic), vicariance biogeography or cladistic biogeography (Humphries & Parenti, 1999). Max Vic is based on the theory that in situ speciation and extinction produce The second theory of hypothesis also known as simple area cladograms in which each area occupies a unique position. PACT analysis for lineage ww ation (sympatric speciation). Second from top = the two hiera (CD) groups coming lm a common reda superimposed on each other. Y = Y) Noise in the system or the effects of C D C D A B C D C D ——»- A B C D — a Top = two hierarchical patterns based on rchical patterns aligned 2i common posed on each other. Third from top right = PACT solution ) elements coming Ju a common t this leaves duplicate (CD Bottom right — PACT solution contingency result from dispersal. From this, one infers that data that conflict with a single area cladogram in which each area appears once (the “law”) are the result of historical contingencies, and it is therefore permissible to remove or modify them. As a result, cladistic biogeographers developed Assump- modify (“reconcile”) tions ] and 2 to remove o incongruent data with a single simple area cladogram in which each area occupies a unique position. One persistent concern about both these paradigms has been this: if it is necessary to remove and modify data and to restrict one's scope of analysis, just how general and powerful are the explanations produced? comparison of the two paradigms reveals they are complementary A closer another interesting feature: Volume 95, Number 2 2008 Brooks & van Veller Assumption 0 Analysis Figure 6. PACT analysis of two partly congruent patterns. Top = two hierarchical patterns based on different data but O to entities A, B, C, and D. Second from top = the two hierarchical patterns aligned by common element . Third from top 7 = both hierarchical patterns nan :d on each other. Bottom = PACT solution based on applying uin rule 4 (Y(Y- de (Y(Y—. Y (A) (Y- (A- + Y (A-) = Y (A) (Y (A-. This exemplar is also known as the “problem of paralogy.’ theories, each one excluding the other's domain of Taxon pulse and taxon cycle models both assume that explanation. What is missing from ETIB are assess- ments of the geographic ou of peres even though this is what distinguishes a “source” from an “island.” What is missing in MaxVic are assessments of post- speciation movements, and yet this is how ancestral species become widespread enough to be affected by vicariance. If each of these theories describes something valid and each one excludes the other's explanatory domain, perhaps the problem lies in the adoption of ontological parsimony on the part of advocates of both theories of biogeography. At vicariance A "new" guiding principle: The taxon pulse. the paradigm emerged, Erwin (1979, 1981) proposed the nearly same time the maximum taxon pulse hypothesis as a model incorporating both dispersal and vicariance. Erwin’s model stemmed — from an idea proposed by Darlington (1943), later named the “taxon cycle" by Wilson (1959, 1961). species and their adaptations arise in "centers of diversification," and that distributional ranges of taxa periodically fluctuate around a more stable, continuously occupied center. This general biotic dispersal may be interrupted by the formation of barriers, producing episodes of vicariant speciation. Breakdown of those barriers produces new episodes of biotic expansion, setting the stage for yet more episodes of vicariance. Taxon cycles occur over relatively short periods of time (“ecological time” and involve species that disperse actively and colonize new areas during expansion episodes, then contract their ranges during periods of habitat contraction without producing new species. Taxon pulses, by contrast, occur over relatively long periods of time (“evolutionary time") and are characterized by [om dispersal along a broad front during expansion into suitable habitat when previous barriers break down. During this expansion phase, different species within Annals of the Missouri Botanical Garden a biota encounter additional geographic heterogeneity, including range contractions. Such heterogeneity may (1) stop the expansion of some species, resulting in species of restricted. distributions; (2) affect only the some species, producing rate of expansion for widespread species; or (3) act as barriers to dispersal of sufficient magnitude to produce new species as a result of peripheral isolates speciation. Geological evolution, operating on longer time scales than biological evolution, may also produce barriers, resulting in episodes of vicariant speciation affecting members of these same biotas. Despite the existence of an alternative to MaxVic and despite concerns that exemplar taxa were being selected to show a preponderance o (Simberloff et al., 1980; 1987), vicariance has become the default e a And yet, MaxVic has carefully vicariance Simberloff, for any observation of allopatry. always been deficient because it neglects the issue of how ancestral species of mary clades become widespread enough to be affected by vicariant events. f vicariance affects many members of ancestral biotas in the same way, it seems reasonable to assume that al the biota some point in the past the members of expanded their geographic ranges to such an extent that vicariance event. they could be affected by the subsequent Advocates of vicariance biogeogra- phy have acknowledged that this must happen. Wiley 1981) colonization of islands, might produce general distri- noted that circumstances, such as some rather than Endler (1982) suggested that such correlated dispersal patterns might be common. [n bution patterns based on dispersal vicariance, and however, practice, historical biogeographers have o o a simply assumed that such dispersal does not produce general patterns, so it is permissible to invoke dispersal only to explain departures from the general pattern, which is always explained as the result of vicariance (Wiley, 1986, 1988a, b; Brooks & McLennan, 1991, 2002: Humphries & Parenti, 1999). Taxon pulse-driven biotic diversification differs from vicariance-driven biotic diversification in three impor- First, biotic expansion, we expect to find general patterns tant ways. because diversification is driven by associated with dispersal, not just with vicariance. General patterns resulting from biotic expansion occur when barriers to dispersal, especially the large-scale break episodes of biotic expansion, even those involving large ones leading to vicariance, down. Second, areas, will inevitably lead to reticulated historical relationships among areas and biotas within areas of endemism comprising species of different ages derived from different sources. Third, the absence of particular clades in particular areas is more parsimoniously explained as a lack of participation in that particular = expansion episode by a particular clade, rather than dispersal with extinction. Taxon pulses are also historically contingent, meaning that at any given time, different clades comprising a complex biota may form a mosaic of area relationships. Recent studies have shown extensive reticulated area relationships even for data sets carefully chosen to emphasize vicariance (e.g., Brooks & McLennan, 2001: Green et al., 2002: MeLennan & Brooks. 2002 Spironello € Brooks, 2003: Bouchard et al., 2004: Brooks € Ferrao, 2005: Brooks € Folinsbee. 2005: Halas et al., 2005; Folinsbee & Brooks. 2007). 4 new research program for historical biogeographic Malas et al. (2005 research program based on an ontology of complexity analyses. Following a new for historical biogeographical analyses comprises two sleps: Step 1—Producing a general area cladogram. Convert all phylogenetic trees into taxon-area clado- grams by replacing the names of the species with the areas they inhabit, then combine them using PACT, producing a general area cladogram (GAC). Figure 7 is the GAC resulting from PACT analysis of elephants, hyenas, and hominoids since the Miocene (Brooks & Fo Folinsbee € Brooks. 2007). Step 2—Inferring biogeographic history from — insbee, 2005; the GAC. This is a two-part process. First, we musl distinguish general from unique nodes. Unique nodes are produced by an evolutionary event affecting only a single clade. At present, all nodes associated with evolutionary events affecting more than one clade are m termed general nodes. Clearly, a general node involving two of 100 clades likely is not the same as f 100 clades. At the 1owever, we do not have enough empirical data or any moment. one involving 75 « models to allow us be more fine-grained in our distinctions among general nodes, but this is clearly an area of interest for future research. Second, we Unul historical biogeographers believed that all general must interpret the general nodes. recently, nodes in an area cladogram should be ascribed Lieberman (2000, 2003a. noted that episodes of geodispersal (geographic areas vicariance. b). however. fusing rather than splitting) could produce general nodes resulting from dispersal. In a similar vein, the taxon pulse hypothesis asserts that episodes of general biotic expansion occurring between vicariance (or, more generally, isolation) events set the stage for Lieberman (2000, 2003a, b) proposed a protocol for distinguishing general nodes future isolation. events. in area cladograms due to isolation from those caused by geodispersal or biotic expansion. The protocol based on phylogenetic character optimization (Fitch optimization). using areas inhabited as the characters. Volume 95, Number 2 Brooks 4 van Veller 209 2008 Assumption O Analysis = = > > D = > 2 3 25 25 ul « 4 ul u au < u a uoa q w ur ww du L L uL 2 nu O kee EN O ww k d uL Se i. 3 2 i i i1 iu A Lu q < X u qo dm do Xd aaqa «cd a «c 3 od uy co c co au dd «oc coc oc ud A Fig PACT analysis of elephants, hyenas, and hominoids since the Mioc ene (Brooks & Folinsbee, 2005; Folinsbee & TEN M 2007 ). General area cladogram. AF = Africa; AM = Americas, AS = Asia, EU = Eu General nodes associated with vicariance or other forms of isolation exhibit decreasing numbers of areas occupied relative to the next oldest node (Fig. 8). whereas general nodes associated with biotic expan- breakdown of barrier) are associated with sion increasing numbers of areas occupied relative to the next oldest node (Fig. 9). Figure 10 summarizes the nodal analysis for the G AC shown in Figure 7. The alternation of isolation ~ , or vicariance) nodes with biotic expansion (BE) and within-area (W) differentiation nodes produces the characteristic signature of a taxon pulse radiation. The relatively large number of general nodes (22 out of 28, or 7996 that all participated in the same general taxon pulse scenario, three clades suggests summarized in Figure 11. Complexity studies in biogeography: Integrating historical and ecological biogeography. That larger islands have greater species richness than smaller islands, and that “islands” need not be oceanic because species richness increases with any increased sample of area, have long been recognized. ETIB (MacArthur & Wilson, 1963, 1967) predicts a linear log-norma function relationship between species richness and the size of an island resulting from a dynamic balance between immigration, that is, colonization from a source area, and extinction. The extinction rate is assumed to increase with the number of species present on any island, so that small areas with higher species richness than expected have a higher extinction rate than immigration rate and are not yet in equilibrium. MacArthur and Wilson (1963 acknowledged that “local speciation” (in situ specia- tion) would confound the species-area relationship, but suggested that for most cases it was probably safe to omit in situ speciation from the model as its effect on the species-area relation 1s “probably significant only in the oldest, largest, and most isolated islands” (1963: 380). Recent discussions, however, have suggested that such historical phenomena require closer investigation (e.g., Heaney, 2000; Whittaker, 2000) Halas et al. = (2005) performed the first direct analysis of the phylogenetic impact on the species- area using the extensive data set presented by Marshall and Liebherr (2000), repre- clades of relationship senting 33 insects, vertebrates, an flowering plants occurring throughout Mexico and parts of Central America. Nodal analysis of their GAC permitted them to classify each species in each area as a colonizer or a resident. Correlating species richness and area size for the 33 clades and nine areas in Mexico and Central America (for details, see Halas et al., 2005) produced a low correlation coefficient for the species-area curve (r^ = 0,47), due primarily to two relatively small areas (the Transmexican Volcanic Belt and the Madre del Sur) unusually numbers of species. Sierra containing large Correlating extinction events and species richness for this data set produced a high correlation coefficient (r^ = 0.75). indicating strong support for this prediction of the 210 Annals of the Missouri Botanical Garden A A B E [o 8 BCD ABCD A F Figures 8 and —8. Lieberman's nodal analysis 1. Thick branches indicate general nodes; thin brane 1es represent elements of the GAC produced by only one clade. Progressive decrease in the number of areas associated with nodal values indic: = that all general nodes are ea nodes. J. Lieb Thic harem by only one clac | rule. erman’ S e analysis 2 This corresponds to either maximum vicariance or Hennig's k branches indicate general nodes; thin branches represent elements of the GAC . Increase in the number of areas associated with nodal values from node D to node CDE in o idicates an episode of biotic expansion or general dispersal; decrease in the number of areas associated with nodal values from CDE to CI ) indicates an episode of isolation, corresponding to either vicariance r Hennig’s progression rule. Alternation of biotic expansion and isolation nodes is characteristic of taxon pulse radiations. ETIB. Following Macarthur and Wilson’s lead, Losos and Schluter (2000) reported that, for anoline lizards on large Caribbean islands, inferred extinction rates were low and in situ speciation was a more important source of species richness than colonization. They predicted that effects similar to those they observed on the largest islands should be found on continental Halas et (2005) predictions: in situ speciation correlates better with islands. al. corroborated those area size than does colonization, and inferred extinction rates are low. The protocol produces inferences of only 19 extinction events involving 16 of the clades, beca direct 33 ause parsimonious inferences of extinction can be made only for episodes of vicariance. If extinction rates are the same following biotic expansion events, which account for the number of inferred 19 to 48, change the correlations with area size. 60% of the general nodes, extinctions increases from but does not These data provided additional insight into the evolutionary relationship between colonization and in situ speciation. There is a very poor correlation 2 -— 1 situ. speciation (r^. = between colonization and i 0.02), independent of each other. We suggest that the reason ~ indicating that these phenomena are relatively historical effects on large islands confound the species-area relationship is the subsequent dispersal of some species produced in situ to other islands, so sources become islands and islands become sources on evolutionary time scales. All nine areas studied by Halas et al. (2005) have served as both sources and islands at different times and to different degrees. Halas (2005) episodes of vicariance and nine of biotic expansion et al. identified six fundamental as the main determinants of the biogeographic patterns observed in the data set. The Transmexican ic. Belt species by colonization, for all nine biotic expansion Volcanic has acted as an “island,” receiving events, and the Sierra Madre del Sur has acted as a dispersal source for three of the episodes of biotic expansion. At the same time, both areas were involved in five of the six episodes of vicariance. This explains why these relatively small areas are disproportionately species rich, without any evidence of an accompany- ing high extinction rate. — The analysis by Halas et al. (2005) indicates that MacArthur and Wilson (1963, 1967) were correct —in situ speciation confounds the species-area relation- — ship. This occurs when the same areas function as islands and as sources at different times. How can this happen? The answer is niche diversification associ- ated with speciation, producing island species that can colonize the source area(s). Niche diversification most often occurs at a much slower rate than Volume 95, Number 2 Brooks & van Veller 211 2008 Assumption O Analysis z z q « YN € N nn o a nn a nnn q q q II a < qa Il «x q qoc a > o a > 2 2 z 2 2 2 2 z 2 22252 ul q 4 ul ui u «a ul x uon O Iw] [re] > m ma Figure 14. efficiently than species 3, needing only two duplica- tions to achieve 13 inter-population connections as opposed to four duplications to achieve 14 inter- population connections. PHYLOGENETICS The Hennigian assumption that phylogeny can be represented accurately by an internested branching diagram is refuted each time a new species is formed by the hybridization of two (or more) other species. This produces a reticulated phylogenetic history that cannot be represented by a purely internested PACT and phylogeography 1. Three gene trees for various populations. branching diagram that is the result of analyses with jem standard phylogenetic methods. Most zoologists consider hybrid speciation to be rare in animals, but botanists have recognized for a long time that many plant species have been formed in this manner (Vriesendorp & Bakker, 2005 and he endosymbiont — examples therein). Furthermore. t theory of the origin of eukaryotes (e.g.. Schwartz & Dayhoff, 1978; Margulis, 1981; Sitte, 1993) postulates that certain major transitions in evolution (sensu Maynard Smith € Szathmary, 1995) are the result of hybridization. At least some prokaryotes have an interesting ability that refers to the capability of Annals of the Missouri Botanical Garden A D B C A B B C E EF Figure 15. producing reticulated phylogenetic relationships. In times of stress, these prokaryotes not only shut down their metabolic activities but may actually jettison the portion of their genomes associated with metabolic activities. Furthermore, they are capable of regaining genetic machinery for metabolic activities, but not always from a close relative. Partial hybridization is a common process for prokaryotes (Woese, 2000, 2002). In bacteria, genes for resistance to antibiotics are usually situated on These chromosomes and can be transferred by plasmids. lasmids behave as accessory | I j bacterial bacterial conjugation. Their plasmids can be ex- (om changed via a tubular connection between two bacterial cells, and resistance to antibiotics can become widespread to relative as well as nonrelative species. Another way of transferring genetic material [om between two species of bacteria can be accomplishec by transduction by a bacterial virus (bacteriophages or phage). This partial hybridization by exchange of genetic material, however, is not only restricted to bacteria. In case of bacterial transformation (by, e.g., Agrobacterium tumefaciens [Smith € Townsend, 1907] Conn [1942] or A. rhizogenes [Riker et al., 1930 [1942]), genetic material may even be transferred from Conn == a prokaryote to a higher plant species. A “new” guiding principle. | Phylogeneticists were aware early on that species of hybrid origin might pose 1979; Funk, 1981), and this was the subject of a symposium at the meeting of the Wi Hennig | Society (Humphries, 1983; Nelson, 1983; Wagner, 1983; Wanntorp, 1983). Funk (1985), however, was the first to propose that analvzing correlated homoplasies in special problems (Bremer & Wanntorp. — second analyses of clades containing species of hybrid origin G G F H L I t | Jd K K L PACT and phylogeography 2. PACT analysis of three gene trees shown in Figure 14. could be used to detect the hybrid species. She noted that when there are multiple equally parsimonious trees for a clade suspected of containing species of hybrid origins, those trees often represent a small number of distinct topologies. Furthermore, inspection of the trees indicated that most of the homoplasy and, thus, most of the ambiguity was contributed by a small number of species. Removal of those species from the analysis iL dis ‘ed a stable topology and reduced homoplasy. Funk's insights, combined with PACT, form the basis under our] or an approach to phylogenetic analysis Assumption 0. A new research program for phylogenetic analy- ses. Within 1 program, phylogenetic analyses of entities the realm of an ontologically complex — researc show cladogenetic as well as reticulated relationships, and the following two steps are applied. Step 1—PACT analysis of all available data. Convert all transformation series into character-state trees, then combine them using PACT (Figs. 16, 17), ignoring the plesiomorphic state. This produces a phylogenetic tree in which all traits interpreted as homoplasy by standard Hennigian analysis will be — interpreted as homologies resulting from horizonta transfer/hybridization. Step different, yet intertwined, issues involved in this step. 2—Explaining reticulations. There are two The first is differentiating reticulated homology from true homoplasy. The second is differentiating hori- zontal transfer from hybrid speciation. The key to deciding the first issue is having 2 phylogenies based on as many characters anc character types as possible—comprehensive total- evidence studies. We would then expect, as suggested by Funk (1985), that cases of hybrid speciation would Volume 95, Number 2 Brooks & van Veller 219 2008 Assumption O Analysis B Z A Z B Z C B A Z C B A Z C B Z A Z 16 C A B Z C B A Z C B Z A Z 17 Figures 16 and 17. —16. PACT and a of hybrid species. Top four are binary m n ee trees indicating distribution of apomorphic states. Bottom three ar 2 +3) and 4. —17. PACT combining the gene trees. Parental species are sister species appear in a PACT analysis as duplications supported by many characters (because hybrid speciation involves combinations of whole genomes), whereas cases of true homoplasy would appear as duplications supported by single characters (because homoplasy involves single characters). For such cases, relying on epistemological parsimony would have an ontological justification. Consider this thought experiment: One species (Z) forms a sister relationship with each of two other species (A and B) in a PACT analysis. If you have one case of (BZ) and one of (AZ), there are three possible explanations: (1) hybrid speciation (Assumption O default); (2) one case of horizontal transfer (A is the true sister species of Z, with horizontal transfer to B or vice versa); or (3) one case of true homoplasy (independent origin). Using epistemological parsimony alone, there is no way to choose, based on the evidence at hand PACT results of (from left) combining I and analysis of hybrid species. = 2 3. then (1 + T result for l and 2, then (1 + 2) and diagrams are gene trees; bottom fab is PAC Next, suppose you have five cases of (BZ) and five cases of (AZ). Now the explanations become (1) one case of hybrid speciation; (2) five cases of horizontal transfer or one to four cases of horizontal transfer of linked and (3) five evolution or one to four traits: cases of independent cases of “co-adapted trait complex" or "adaptive syndrome" evolution. In this case, epistemological parsimony would lead us t the choose hybrid speciation unambiguously as preferred explanation. That means we hypothesize that all shared similarities among the species are homologous. This approach is not guaranteed to produce correct results, of X the heritage of each parental species is expressed in the hybrid product(s). We expect that the more distantly related the parents are, the less likely we will be able course. It works when enough to recover the parental species, but also the less likely 220 Annals of the Missouri Botanical Garden that hybrid speciation will occur. In addition, we often expect hybrid species lo segregate phylogenetically more with one parent than another (see McDade, 1990, 1992; Mindell, 1992, 1993; Zrzavy & Skala, 1993; Skala & Zrzavy, 1994). For ambiguity in. phylogenetic cases in which results from a [ev] analysis small number of species and when there is abundant evidence of correlated homoplasy, Funk's approach gives us an excellent first approximation as to the possibility that the ambiguity is due to those species being hybrids, justifying the use of Assumption 0 analysis. However, when larger numbers of species are involved, the explanations for character ambiguity may become very complex. For example, recent studies of prokaryote phyloge- netics have raised the possibility that this propensity on the part of prokaryotes has produced some wildly reticulated) phylogenetic relationships. Some have even expressed the opinion that this means we should change our view of phylogeny, from a tree to a reticulated network (Doolittle, 2000). Maynard Smith Szathmary (1995), machinery and however, pointed out that genetic associated with storing and transmitting Information shows no similar evidence of horizontal transfer. Thus, the true phylogeny may still be tree-like in homoplasy in metabolic genes may be present due pattern, although substantial to horizontal transfer, Regardless of the eventual outcome of this debate, M we do not have to develop fundamentally different approaches to assessing the evidence about relation- that the the more we will have pied ships. It simply means more horizonta transfer there has been, ) — resort to species duplications those transfers. Much more work, including substan- tial modeling effort, is needed to help determine when it is appropriate to use Assumption O in phylogenetic analysis and when it is not. We will When Assumption O and PACT are applied in this context, now address the second issue. all reticulations are taken as evidence of introgression, but introgression can produce multiple outcomes, of which speciation is only one. Differentiating the various forms and outcomes of introgression requires a robust and unified ontological species concept d evolutionary species concept; Simpson, 1944, 195 Wiley, 1 view of 978) as well as a broad-based Pre the diversity of kinds of species. As we suggested earlier, the PACT analysis is the beginning of an explanation, not an explanation. in itself. Inference of species boundaries requires additional tests (such as nested clade analysis) to highlight the independent lineages and experiments to determine the strength of cohesion of those lineages (for a discussion and protocol, see Brooks & McLennan, 2002). uen — find the source of Finally, we come to the issue of how to incorporate reticulated phylogenetic events into our classifica- tions. There are two choices. Do we classify a tree with reticulations, or do we try to classify a reticulated network? If we choose the former, hybrid species will appear in two different clades, reflecting their phylogenetic origins. This disturbs traditional classi- fications in which each taxon has a unique place. If we choose to try to classify a network, each species of hybrid origin would be classified separately from either parental species, so the resulting classification would maintain the tradition of each taxon having a unique placement, but could not provide information about the phylogenetic relationships of the hybrid species. If we wish our classifications to reflect what we think we know about evolution, it seems that we will have to opt for the first alternative. CONCLUSIONS: BACK TO THE FUTURE Evolution is historically. contingent. and complex. Darwinism provides an ontology of complexity. Histor- ical contingency requires phylogenetic information to his > document and explain patterns. Representing complexity so that we have some understanding of what needs to be explained requires a method that allows reticulated relationships to be found. A new algorithm, PACT (Wojcicki € Brooks, 2004, 2005), developed for that purpose. Applying this algorithm in has been historical. biogeographical, coevolutionary, phylogeo- graphical, and phylogenetic. studies invokes Assump- tion 0; Thereby, the resulting patterns of related areas, species. all data must be analyzed without modification. and populations contain duplications of entities to such an extent that these patterns are logically consistent with all input data. Further inference of these duplications asks for an ontology of complexity in which multiple evolutionary processes resulting in both divergent and reticulated patterns nest comfortably. Methods for phylogenetic analysis that only deal with divergent patterns and do not allow duplication of entities disallow complex evolutionary processes for explaining the convergent (reticulated) patterns and thereby oversimplify the historical relationships among the entities of analysis. Furthermore, several studies mentioned in this paper show that several older and newer concepts like the taxon pulse (Erwin, 1979), ecological fitting (Janzen, 1985b), phylogenetic 1985), and the oscillation hypothesis (Janz & Nylin, 2007) allow for explanation analysis of hybrids (Funk, of reticulated patterns. In order to perform comparative analyses based upon an ontology of complexity, we do not have to begin again because Darwinism is based on such an We do, have to move forward, ontology. however, Volume 95, Number 2 2008 Brooks 4 van Veller Assumption O Analysis beyond the oversimplification of evolution associated with much of 20th-century evolutionary biology. Basic principles of phylogenetic systematics, coupled with Assumption 0, can accommodate analyses of complex evolutionary associations at various levels of organi- zation. Once we begin doing this, we will recapture earlier ideas and integrate them into a renewed appreciation for the Darwinian panorama. Literature Cited Adamson, M. . N. Caira. 1994. Evolutionary factors influer le nature of parasite specificity. Parasitol. NL 109: 585-595. vise, J. C. 2000. Phylogeography: The History and Formation of Species. Harvard Univ. Press, Cambridg ge. Bouchard, P., D. R. Brooks & D. K. Yeates. 2004. Mosaic Australian tropics arthropods: 2 19 in croevolut wW o eat by taxon pulses. Pp. 425-46 C. Moritz . Bermingham (editors), Rainforests: Past, Present, Future; Univ. of Chicago Press, Chicage Bowler, P. J. 1983. The Eclipse of Darwinism. us opens Univ. Press, Baltimore. H.-E. Wannt torp. 1979. . Syst. Zool. 28: 6 Historical ecology: A new approach to T. Bremer, Hierarchy and retic lation i in systemat 024-627. Brooks, . 1985 ae ie on : ecological associations. Ann. Missouri. Bot. Gard. 72: —680. 998 e m es of evolution and selection acess in G. van de Vijver, S. N. Salthe & M. Delpos 2 E dicens Systems: Biological and Epistemological Perspectives on Selection and Self- Organization. Kluwer Academic Publishers, Dordrecht. D O. The nature of the ME Life takes on a life at its own. Proc. New York Acad. Sci. 901: 257-265. . 2001. Evolution in the ino s Rediscov- s Evol. Energy e. WW ru ary.utoronto 002. Taking evolu Üonaty transitions. seriously. Semiotics Evol. Ene 2: 6-24, Available at: , er 17 March 2008 ergy Developm. ). Wiley. 1986. Evolution as Entropy: Toward a United ae of Biology, Ist ed. Univ. of Chicago Press, Chica 88. Evolution as Entropy: Toward a Unified Theory ot Biology, 2nd ed. Univ. of Chicago Press, Chica . MeLennan. 1991. Phylogeny, Ecology and Behavior: A m earch Program in Comparative Biology. niv. of Chicago Press, Chicago. Š 1993. Parascript: Parasites and the Smithsonian Institution Press, ge of Evolution. " eun. D & —— —— ` 2000. The nature of the organism and the emergence of selection processes and biological signals. Pp. 185-218 in E. dion (editor), Semiotics, Evolution, i. Shaker Verlag, Aachen, Germany. $ A comparison of a discovery-based aan an event- A d method of historical biogeography. J. -161. Biogeogr. 28: i . 2002. The Nature of Diversity: An e Voyage of Discovery. Univ. of Chicago Press, Chica 2003. Extending phy eee studies of coevolution: Sec das. Brooks parsimony analysis, para- sites, and the Great Apes. Cladistics 19: 104—119. & rrao. 2005. E lib dtr Emerg gin 18 infect tious The historical aa dis sease re evolutionary accidents wailing to e n. J. Biogeogr. 32: 1291-1299. K. “olinsbee. 2005. Paleobiogeography: da the ebb and flow of evolutionary diversifi- cation. Pp. 15-43 in B. S. Lieberman & A. RM Rode (editors), Paleontological Society Papers, Vol. 1 & E. P. Hoberg. 2006. Systematics E. emerging E management to solution infectious diseases: Parasitol, 92: 426 & 2007a . How will global climate change a parasite- host assemblare? Trends Parasitol. 23: 571-574. 8 —— ———. 2007b. D sloshing bucket: The evolutionary Pa 5 Pme infectious diseases. Evol. Outreach Educ. I: 1920. Darwin’ s necessary misfit and the E Brues, C. with spec jal iE :e to lepidopterous larvae. Amer. Midl. Naturalist 54: 332. 1924. mS specificity of food-plants in the evolution of phytophagous insects. Amer. Midl. Naturalist 58: 127-144. Collier, J. & C. xlexly p dynamical systems. Open Syst. Info. Dynamics 6: 302. The selection of food-plants Yu insects, Hooker. 9. Com organised 241- Cracraft, J. 1985. Biological O and its causes. Ann. Missouri. in Gard. 72: 194—822. sai ci P. J. 1943. nds: n the E of isolated faunas, and on atrophy of wings. . Ecol. Monogr. 37-61. Darwin, C. 1872. The Origin of Species, 6th ed. John Murray, London E Y. 4 2000. Uprooting the tree of life. Sci. Amer. Caral ce of mountains and islar Ce 2002. Testing the accuracy of TreeMap and ses of coevolutionary patterns 16 pa Ne ud Brooks pu analy using artificial pA I ass 18: ,M.G. P. v ‘eller ; >, Hoberg € D. R. Brooks. Á priori er a Mew methods in a e x 2003. evo d studies of host-parasite associations. Cladis- 5: E P. z & P. H. Raven. 1964. pes and plants: A study in LN Evolution 18: 608. Eldredge. N. 19 chies and ie Evolutionary 5. Unfinished Sy hs "sis: Wurm ‘al Hierar- Thought. Oxford Univ. Press, id York. 1986. S economies and evolution. Ann. Rev. Ecol. Syst. 17: 351-369. . 1989. uM ad Dynam es and pu Peaks. MeGraw-Hill, New York. Endler, J. 1982. from ec o factors in biogeography. 441—452. Erwin, T. L. 1979. Thoughts on the evolutionary history of ground beetles: Hypotheses generated from comparative faunal analyses of lowland forest sites in pu rd and tropical regions. Pp. 539—592 in T. D. R (editors), Carabid Evolution, Natural History, The Hague. =] 1 ics: Species, Nic Problems in distinguishing Ds ^ Amer. Zool. /hitehead Junk, and Classification: W. vicariance, and dispersal: An Taxon pulses S, beetles. 196 in G. Nelson & D. E Vicariance Biogeography—4A | Critique. Press, New York. evolutionary synthesis illustrated by Pp. 159- ) Columbia Univ. Annals of the Missouri Botanical Garden Fahrenholz. H. 1913. E ctoparasite n unde abstammungslehre. Zool. Anz. (Leipzig) 41: 371—374. Folinsbee & D. R. Brooks. 2007. Early hominoid Pulse Ss of dispersal and differentiation. J —391 biogeography: I: E T . Special concerns in estimating plant 3-66 in V. A. Funk € D. R. Brooks Cladisties. Vol. L New York New York. P logenetic patterns and hybridization. Ann. Missouri Bat. rd. 72: 681-715 Gould, S. J. 1 pas Pp. 75-16 in M. un in Botanic al Garde D. 98 hardening ui the modern synthesis. Grene (editor), Dimensions of a Themes and Counterthemes in Twentieth-Ce ' Evolu- tionary Theory. Cambridge Univ. Press, oia Green, M. D.. M. C. P. van Veller & D. R. Brooks. 2002. viia asymmetry and s congruence istics 112-124, Halas, D., D. Zamparo & D. CA protocol for dri biotic diversification un laxon eina J. Biogeogr. 249—200. . R. 2000. Dynamic disequilibrium: Assessing modes of o A long-term, pen scale perspective on the equilibrium model of island b ao Global Ecol. Biogeogr. 9: 5974 ». Phylogenetic Systematics. Urbe of Ilinois ress, ae E. P. & A. 2000. biodiversity: Understanding faunal structure and er "og- raphy in the marine realm. Bull. Scand. Soc. Hobere, \dams. Phylogeny, history and Parasitol. 19-37. ———, A. Jones, R. L. Rausch, K. S. Eom & S. L. Gardner. 2000. A shy lose nelic nd n sis for species of the genus Taenia (Eucestoda: Taeniidae). J. Parasitol. 86: 89-098. © N. Alkire, A. de Queiroz & A. Jones. 2001. Out of oss Origins of the Taenia tapeworms in oo Proc. jl. Sci. 268: 787 Roy. Soc. London, Ser. B. Bio Humphries, € J 1983. Primary data Cii analysis. p. 89-103 in N. l Platnick & Funk pU s), \dvances in Cladistics H. Columbia Univ. e ress, New & L. R. Pare nti. 1999. Cladistic oda Interpreting Patterns of Plant and Animal Distributions, 2nd ed. Oxford Univ. P New York Janz, N. & S. Nylin. 2007. The oscillation hypot thesis ol host plant-range and speciation, Pp. 203-215 in K Tess, (editor), Specialization, wie and Radiation: Evolutionary Biology of Herbivorous Insects. Univ. of BE rkeley. California Press, —, —— Wahlberg. 2006. Diversity be "gels diversity: Host e D and the ia sification of plant- RES BMC Evol. Biol. Janzen, D. H. 1908. Host Pn as aad in evolutionary and conte ied lime. Amer. Midl 02: 592—595 197: Host Fc e “ind contemporary time. ralist 107: 9 Naturalist I plants as islands: Compe in Amer. Midl. Natu- 780-7 1973b. ona nts on host- “spec ific ‘ity of tropical p. 201- DEUS and Ecology. herbivores zie ils relevance to ui cies richness. P 211 V. 2 in . Heywood (editor Academic Press, New Yor 198 "i When is ib coevolution? Evolution 34: aL Patte ' ns da herbivory in a tropical deciduous 271-282. seeds by Futuyma & M. Associates, vertebrate Slatkin (editors), Sunderland, 5 of 232- 262 in D. J. Coevolution. Sinauer guts. Pp. 23 Massachu- setts . Coevolution as a process: What p of NOT ha Coevolution of Par isitic Mammals. Y Wiley and Sons, New . 1985 ina and pan do lave in common. Pp. 8 3-99 in im (editor), Arthropods O and » On oe cal fittin “Oikos 45; 308—310. ———— & P. S. M . 1982. Neotropical anac oo T fruits the ( D qat. ate. Science 215: 19-27. Kellogg. V. L. 1896. New 1e Mallophaga, l. pe ith special » a collection from maritime birds of the Bay of MES Proc. Calif. Acad. Sci. 6: 31-168. . Distribution and species-forming of ectopar- pie: Naturalist 47; 129-158. ». Permanent splits as erence to TAPA rey, asiles. es Kornet, D. J. formal rec MON tion of n internodal species concept. J. Theor. Biol. 164: 40 Lieberman, HB. 2000. Pale da Plenum/Kluwer Acader _ New York. 20 3a. Unifying theory Evol. Biol. 33: 1-25. 2003b. Paleobiogeography: The uor ance of fossils Ann, Rev. Ecol. Syst. 51—69. 2000. Analysis i ane D Nature 408: € speciation events: A and methodology i 1 bioge ography. E to bore na .Sc in Losos, ~% ies-area ron Mac e R. H. . Wilson. 1963. Ps y equi theory of a zoogec er vy. Evolution 17: 373-387, p L——— d 207. The Theory of Island Biogeogra- w Jersey. New Jerse Cell Evolution. W. Princeton, - & phy. Princeton Un Margulis, L. 1981 [mm CSS ‘Saal esis i San Francisco. Marshall, T & J. K. Liebherr. 2000. Cladistic biogeog- raphy of ua Tid transition zone. J. Biogeogr. 27: —210. Maynard Smith, J. € E. Szallmary. 1995. The Major Transitions in Evolution. W. H. Freeman, San Francisco. McDade, L. A. 19 Patterns of 90. Hybrids and phylogenetic systematics I. character expression in hybrids implications for cladistic analysis. Evolution 44: . 1992. Hybrids and phylogenetic systematics I. The impact of hybrids on cladistic analysis. Evolution 46: 1329-1: E McLennan, .. A. & D. R. Brooks. 2002. Complex histories of pes iation and dispersal: An example using some Nus des birds. J. Biogeogr. 29: 1055-1006. Minde ll, P. 1992. Phylogenetic consequet of symbi- 5 oses: E Wide va and Eubacteria are not D EN laxa. oe “ms s 27: 53-62. 1993. Merger of taxa and definition of monophyly. Biosy stems 31: 130-133. Mode, C. J. 1958. A mathe matic ‘al model for the co-e volution of E parasites and their hosts. Evolution 12: 15 ius G. 1983. peces in cladograms. Pp. 105—111 in Platnick € V. A. Funk (editors), Advances in Al e ss: Proceedings of the Second Meeting of the Willi Hennig Society. Columbia Univ. Nylin, S. & N. Janz. 2007. example of plasticity Ecol. DOT: 10. Press, New York Butterfly host plant range: An as a promoter of Aa ciation? Evol. 1007/s10682- 007-9205- Page, R. D. M. a 2002. Tangled Trees. Univ. of Chicago Press, Chicage Posada, D. K. A. it 2001. Intraspecific gene genealogies: Trees grafting into networks. Trends Ecol. Evol. 16: 37-45. Riddle, B. R. & D. J. Hafner. 2004. The past and Me roles of phyloge Mond in d : bL Ad ye ^p. 93-110 in M. V. iolino & L. R. ey (editors), Frontiers in EU cus New pm in e Geography of Nature. Sinauer Associates, Sunderland, Massachusetts. Volume 95, Number 2 2008 Brooks & van Veller Assumption O Analysis 223 Salvador Arias, J., 1. J. Garzon-Orduna, F. Lopez-Osorio, s Parada-Vargas & afael Miranda-Esquivel. 200 What is PACT sedi? Cladistics 24: 1-12. Schwartz, & M. O. Dayhoff. 1978. Origins of prokaryotes, NM onda and chloroplasts. Science 199: 395—403. re D. S. 1987. Calculating probabilities that ladograms no A method of biogeographical inference. —59. Syst. ae a ER E. D. MeCoy & E. F. Connor. 1980. one no statistical tests oí cladistic biogeo 0- 0—63 in G. Nelson & D. Rosen (editors), Vicariance Biogeography: Columbia Univ. Press, New York. Simpson, G. G. 1944. Tempo and Mode in Evolution. cue Univ. Press, New York 951. The species con cep Evolution 5: 285-2 en The io or Features of Evolution. Bud Univ. Press, New Yo Sites, J. W., Jr. & J. C. Massa 2003. Delimiting species: À renaissance issue in systematic biology. Evol. 18: 462-470. & There a A (sque. Trends Ecol. 4. Operational criteria for delimiting ecies. Ann. Rev. Ecol. Syst. 35: 199-227. Sine, P. 1993. Sumbiopenetic evolution of ye cells and MU plastids. Eur. J. Protistol. 29: 43. Skala, Z. & J. Zrzavy. 1994. Ba retic m and rad Discussion of methodological concepts. Cladis- tics 10: 305-313 Spironello, Brooks. 2003. Dispersal and diversification in the evolution of Inseliellium, an archi- 30: 1563-1573. s, F. N. Renaud € S. Morand. 1998. The evolution of life. chee traits in parasitic anc free-living p. nce A new perspective. Oecologia 115: 370—378 van n Veller, M. G. P. & D. R. Brooks. 2001. When simplicity simonious: A priori and a a methods in rnc biogeography J. Biogeogr. 28: 1-11 Zandee & D. J. Korn ds valid common et. 1999, Two equinenients atterns under different assumptions in vicariance biogeography. Cladistics 15: 393-406. for el € M. Zandee. 2000. Methods in vicariance lo Assessment of the implementa- tions of assumptions zero, 1 and 2. Cladisties 16: 319— 3 . Zandee & D. J. Kornet. T inclusive sets assumptions zero, 2001. Measures for of ograms under l and 2 with different methods for vicariance biogeography. Cladistics 17: 248-259. ). J. Kornet & M. Zandee. 2002. A posteriori and a er methodologies for testing hypotheses of causal cesses s & M. Zandee. 2003. Cladistie i Fest TRUE The art and the science of discovery. J. Biogeogr. 30: 319—329. Verschaffelt, E. 1910. The cause ao selection o food in some herbivorous insects. Proc. Acad. Sci Amsterdam 13: 536—542. Vriesendorp, B. & F. T. Bakker. 2005. Reconstructing patterns of reticulate evolution in angiosperms: What c we do? Taxon 54: 593-604. Wagner, W. H., Jr. 1983. Reticulistics: The recognition of bo bilds and s ae in cladistics and classification. Pp. 63-79 in Platnick € V. A. Funk (editors), Advances in Cladistics I Columbia Univ. Press, New York Wanntorp, H. 983. Reticulated cladograms and the identification of hybrid taxa. Pp. 81-88 in N. I. Platnick & V. A. Funk (editors), Advances in Cladistics II. Columbia Univ. Press, New Yor Whittaker, R. J. 2000. Sole succession and complexity in island biogeography: Are v we asking the right questions? Global Ecol. Biogeogr. 9: 75-85. Wiley, E. O. 1978. reconsidered. Syst. Zo ] s$ The a ol. 27: 17-26. Phylogenetics: The Theory and Practice of Pluloscnete i eae Wiley-Interscience, New York. . 1986. Methods in vicariance biogeography. Pp. 283— 306 in P. Hovenkamp (editor), o and Evolution: A Matter of Diversity. Utrecht . Press, Utrecht. . 1988 a a sis gut vicariance biogeog- "s Syst. "Zool. 37: 2 1988b. Vicariance ala Ann. Rev. Ecol. st. 19: 513-542, W D E. O. 1959. Adaptive s zn dispersal in a tropical and fauna. Evolution 13: 122- —. 196]. The nature of jos taxon cycle in the Melanesian ant fauna. Amer. Midl. Naturalist 95: 169-193. . R. 2000. Au the universal phylogenetic tree. Prec e Acad. Sel. . 97: 8392-8396. On the ee di eis Proc. Natl. Acad. Sci. iru 2» 8742-8741. Wojcicki, M. € D. R. Brooks. 2004. Escaping the matrix: A new algorithm for phylogenetic comparative studies of co- evolution. Cladistics 20: 341—361. & species concept . 2005. PACT: A simple and s algorithm for generating area cladograms. J. Biogeogr. 32: —(14. Dandee, M. & M. C. Roos. 1987. Component-compatibility in historical Diageon, Cladistics 3: 305—332. Zrzav Z. Skala. hybrids and c Jadistic 's. o 3l Holobionts, 30. 121-1 RESOLVING SPECIES L. Lacey Knowles? and Yat-Hei Chan? PHYLOGENIES OF RECENT EVOLUTIONARY RADIATIONS' ABSTRACT For recently derived species and when the time se parating speciation events is short, the phylogenetic distribution of taxa in a gene tree may not accurately reflect the actual species relationships. The eo tradition of relving on gene tree and species tree synonymy is not reliable under such historical scenarios. Nevertheless, recent studies have oa d that accurate estimates of species relationships are possible when the method of d eue inference considers not only the stochastic processes of nucleotide E e but also the random loss of gene lineages by genetic drifi—even when there is wide ee incomplete lineage sorting. This simulation study examines how the broader phylogenetic context, that is, the species tree topology and branch RUM influences the ability to recover species relationships when taxa have undergone a recent and pur radiation, As expected, the time since species divergence and the time between speciation events influences whether prog netic relationships are accurately estimated. However, the influence of the timing of divergence on the ability lo recover species relationships accurately differed. depending on the relative en of the taxa in the species tre Differences in the ability to recover these re rni. )S across multiple simulated s s trees highlight the potential effects of taxon sampling on phyloge netic inference at, or near, the species boundary and pine hek rates of speciation. By focusing attention on the species tree, rather than on ihe individual gene trees as the basis for interpretations about species re lationships, these results also represent a fundamental shift from the phylogenetic paradigm. ey words: Coalescence, divergence time, genealogical discord, incomplete lineage sorting, radiation, speciation, statistical phylogeography. Within a traditional molecular phylogenetic frame-— nized—the genealogical histories of individual loci work, species relationships are inferred directly from may not faithfully reflect the phylogenetic history the topology of the estimated gene trees (Felsenstein, because of retention and sorting of ancestral polymor- 2004). Under this paradigm, phylogenetic resolution phism (Avise et al, 1983: Pamilo € Nei, 1988; (at the most basic level) becomes a question of how Takahata, 1989: Maddison, 1997; eee. 2002, many unlinked loci are necessary to provide corrob- 2003). To avoid misleading conclusions about species orative evidence of species relationships (e.g.. Miya- relationships, arising from either discord between a moto & Fitch, 1995; Poe & Chubb, 2004; Jennings & gene tree and the species tree (Maddison, 1997) Edwards, 2005; Rieseberg et al., 2006) or number of widespread incomplete lineage sorting that obscures nucleotide sites required to estimate a gene tree (or any obvious pattern of historical relatedness (e.g. topology from combined data across multiple loci) Jennings € Edwards, 2005; Maddison € Knowles, (e.g.. Walsh et al. 1999: Rokas et al. 2003) 2006; Knowles & Carstens, 2007a). methods of However, when the time separating species diver- phylogenetic inference need to take into account the gence is short, genetic drift is unlikely to have time to process of gene-lineage sorting (e.g... Maddison $ bring loci to fixation before subsequent speciation Knowles, 2006; Liu & Pearl, 2007). These methods events (Tateno et al., 1982; Tajima, 1983; Pamilo € shift the focus away from the idiosyncrasies of Nei, 1988). Consequently, the estimated gene tree has — individual gene trees as the basis for inferring limited utility because the most recent common — phylogenetic relationships, to estimating the history ancestor of individuals is not likely to occur within of species divergence (ie. the species tree) directly e.g., Carstens & Knowles, 2007; Edwards et al., 2007). The problems that incomplete lineage sorting (i.e.. One of the conditions where the stochastic loss of a species lineage. the failure of gene lineages to coalesce into a common gene lineages by genetic drift and incomplete gene- ancestor before subsequent species divergence) poses — lineage sorting will make gene trees an unreliable for estimating species relationships are widely recog- basis for inferring taxonomic relationships involves "Thanks to Rick Mayden, and the organizers of the symposium, for inviting me to participate in the symposium. The research was funded by the following awards to LLK: a National Science Foundation grant (DEB-0447224) and the Elizabeth NC Caroline Crosby Fund, National Science Foundation ADVANCE Project, University of Michig “Department of Ecology and E a Biology, 1109 Geddes Ave., Museum of 7 Giclee ey, "Room 1089, University of Michigan, Ann Arbor, Michigan 48109-1079, U.S.A. Author for correspondence: knowlesl(?umich.edu. doi: 10.3417/2006102 ANN. Missouni Bor. Garp. 95: 224-231. PUBLISHED ON 18 JUNE 2008. Volume 95, Number 2 2008 Knowles & Chan 225 Evolutionary Radiation and Phylogenetic Inference the rapid diversification of species. During evolution- ary radiations (e.g., Shedlock et al., 2004; Nikaido et al., 2006). speciation events occur in rapid succes- sion. Consequently, reconstructed gene trees may not accurately reflect the history of speciation among taxa because of the persistence of ancestral polymorphism across subsequent species divergence events. In fact, the expected lack of congruence among independent genes (Hasegawa et al., 1985; Slowinski, 2001) has been used to infer that species have radiated (e.g.. 1999: Poe & Chubb, 2004). Rather than commenting on the discordant genea- Jackman et al.. logical histories as evidence for a radiation, here we examine whether a signal of species relationships can be extracted despite the genealogical discord that accompanies rapid speciation. We consider the effects "both the timing of species divergence and the internode distance (i.e., the timing of the preceding speciation event) on recovering phylogenetic relation- ships when there is widespread incomplete lineage sorting. The ability to estimate the phylogenetic relationship of a pair of sister taxa is explored over a natural range of branch lengths and topologies, rather than focusing on a specific species tree (e.g., Takahata, 1989; Rosenberg, 2002). thereby providing a broad historical context. for examining how the history of species divergence during a recent radiation will affect our ability to recover species relationships. METHODS PROCEDURAL RATIONALE The information for estimating phylogenetic rela- tionships is extracted from the pattern of gene-lineage coalescence (as described in Maddison & Knowles, 2006). as opposed to synonymizing the gene trees with the species tree. We focus on historical scenarios where the hazards of incomplete lineage sorting are expected to predominate—namely, recently diverged species. Because the probability that a particular gene under the coalescent process (Takahata & Nei, 1985; Rosenberg, 2003; Degnan & Salter, 2005), a species or population tree can be inferred (in principle) using a full-probabilistic model (e.g.. 1997; Degnan & Salter, 2005). In practice, Maddison and Knowles (2006) showed that the species history can be accurately inferred, even if the actual probabilities of tree for a population model can be calculated Maddison, incomplete lineage sorting are not quantified under a stochastic model. Therefore, an approach that incor- that results in the of gene lineages below the divergence porates the genetic process coalescence of the species (ie. deep coalescence) during the phylogenetic inference procedure is applied here. COMPUTER SIMULATIONS natural spectrum of Species trees that include a topologies and branch lengths were simulated using Mesquite version 1.1 (Maddison & Maddison, 2004). Species trees were simulated for six species to have a total time depth (summed length of branches from any terminal down to the root) of 100,000 generations according to a Yule model. With an N, of 100,000, a total time depth of 100,000 generations (i.e., a total tree depth of 1 D leads to considerable incomplete lineage sorting (e.g., Fig. 1). One thousand species trees were Saini, 100 of which were selected at random as species trees in which a particular taxa pair was sister (species E and F were randomly chosen as the target sister taxa), out of the 109 species trees with E and F as each others closest relative, and species A was the outgroup. To account for the stochasticity of the genetic process of gene lineage coalescence, the ability to recover the sister relationship of species E and F was examined over 100 replicate data sets for each species tree, in which either one locus or three loci were sampled for each of the 10 individuals in each species. Sene genealogies for each replicate were simulated by a neutral coalescence (Kingman, 1982; Hudson, 1990 through Mesquite’s Neutral Coalescence module, which uses an exponential approximation to avoid fully explicit modeling of individuals. A biologically rea- sonable effective population size (N,) of 100,000 was used for all simulations. Population size of the ancestral species lineage was also set to the common size of 100,000; this represents a reasonable null model of speciation. Otherwise, varying the effective population sort of size of the ancestor would imply some demographic event (e.g., a bottleneck or expansion if its size was set to smaller or larger values, respectively, compared to the descendent taxa). A species tree was then estimated for each replicate from the simulated gene trees by minimizing the number of deep coalescences between the gene trees the species tree (for details, see Maddison & vade nd 2006). This method is based on searching for the species trees that minimize the implied number of deep coalescences in the contained gene trees (Maddison, 1997). The number of deep coalescences was counted assuming the reconstructed gene trees taxon addition were unrooted and using an “as is” followed by subtree pruning regrafting sequence, branch swapping, saving only a single tree at any stage (MAXTREES = STATISTICAL ANALYSES AND SUMMARY OF RESULTS To evaluate the ability to recover the sister relationship of E and F, each replicate was examined mm 0000 IA i AA AAA ree PY Y ABAAAMAMA In minim mimi AAAA AR AS "PY Species: ABCDEF Individuals: 0 € x A €» Figure 1. spec ies studied here (on the left are three representative white circle Example of the degree of incomplete e sorting apparent in a gene tree of any single sampled locus and the degree of discord among sampled loci for the recent radiation of ne trees); the effects of the species divergence time (1), internode distance (i), and nodal position (shown in the s) of the sister taxa relative to the other species on recovering species lobos accurately are considered here. Gene trees were simulated with 10 Sas a (i.e., individuals) per species by neutral whe coalescence within a simulated species tree (shown on the right) with a total depth of species tree from root to tips = 1 N, where N = 100 920 > zi nH 3 o» cw -— 9 DW _. os 9 D 2. cC m C) £5 — o (9) = Volume 95, Number 2 2008 Knowles & Chan 22 Evolutionary Radiation and Phylogenetic Inference and F To examine and the number of times species E were estimated as sister laxa was recorded. what factors influence the accuracy of estimated ationships, (1) the depth of the species divergence. — re (2) the time of the preceding speciation event (i.e., the length of the internode defining the sister taxa E and F), as well as (3) the position of the E and F species split relative to the other four taxa in the original species tree were noted (Fig. 1). Analyses of variance (ANOVAs) were used to test for a relationship between divergence time and internode distance (Fig. 1) on the ability to recover the sister relationships of species E and F. An analysis of covariance (ANCOVA) was used to test whether this relationship differed depending on the relative position of the target taxa in the species tree (i.e., nodal position). To separate the effect of nodal position from the influence of differences in the timing of divergence on the accurate recovery of E and F as sister taxa, this analysis was carried out on the residuals from the regression of phylogenetic recovery and divergence time. RESULTS AND. DISCUSSION For the shallow species histories considered here, there is a very low probability of reciprocal monophyly (Hudson & Coyne, 2002). incomplete lineage sorting and gene tree discord predominate (see also Maddison & Knowles, 2006: Knowles & Carstens, 2007b). With a total tree depth of 1 N., which corresponds to 100,000 years assuming one generation per year and an effective population size of 100,000 for each species, the timing of species divergence and interval between speciation events is of the species and very short (see Fig. 1). For example, the average depth of the divergence of species E and F ranged from 0.01 0.16 N, and averaged 0.1 Ne, or 10,200 whereas the average internode distance, or approximately years assuming one generation per year, the time of the preceding speciation event, ranged from 0.002 to 0.03 2300 years (Fig. 2). The rate of speciation in the simulations (i.e., six species originating with the total tree depth of N. with an average of about 100,000 years) corresponds to the origin of a new Therefore, the the affecting phylogenetic accuracy when species have species about every 17,000 years. simulations serve as a guide to conditions undergone an evolutionary radiation (e.g., Mendelson & Shaw, 2005). The range of species divergence times represented in the simulation study is also common to species-level studies (Arbogast et al., 2002), espe- cially those involving Pleistocene divergences (e.g., Knowles, 2000, 2001: Masta & Maddison, 2002; Hewitt, 2004), making the study broadly relevant to the general difficulties with reconstructing species relationships at or near the species boundaries. While species the discordance between gene trees and trees (Edwards € Beerli, 2000: Knowles & 2002; Hey «€ Machado, 2003) interpreting the genealogical patterns observed in any Maddison, makes single gene tree problematic (see Fig. 1), as shown by (2006), relationships of species can nevertheless be accurately Maddison and Knowles the phylogenetic inferred despite widespread incomplete lineage sorting. However, study design is critical to such inferences (Takahata, 1989; Rosenberg, 2002). that the accuracy of phylogenetic. estimates. differed This study shows depending on the number of loci considered, where accuracy reflects the proportion of replicates in which species relationships were estimated correctly for each species tree. À higher percentage of correct. species relationship was recovered with three loci compared to standard error vs. > one locus (average of 59 KS 48.2 + 1.37 standard error, respectively). Both species divergence time and the internode distance explained a significant amount of the variance in phylogenetic accuracy across the species 3). However. much stronger influence on phylogenetic trees (Fig. internode distance had a accuracy compared to species divergence time. The ability to my estimate the sister relationships of species E and F — e.g., Fig. 2) decreases dramatically with a short time interval between the divergence time of species E and F and the preceding speciation event (1.e.. small internode distances). This matches theoretical expec- tations that demonstrate that not only the timing of species divergence, but also the interval between speciation events, strongly influences the consistency probability between species and gene trees (Takahata, 1989; Rosenberg, 2002). (1.e.. with recent than 0.1 N). the and F Even very species divergence less recovered mes relatedness of species E was accurately using genealogical information from a single locus (Fig. 2). However, if the timing of the E and F species divergence was similar to that of the internode preceding speciation event (ie. short distance), there is a low probability of recovering he sister relationships of E and F, even with three loci (Fig. 3). ANCOVA showed that when controlling for differences in divergence time (Table 1), internode length still explained a significant amount of the variance in phylogenetic accuracy across the different species trees when both a single locus and three loci were used to estimate the species tree (r^ = 0.80, P < 0.0001, 0.68, P < 0.0001, for the whole model for one locus and three loci, respectively). and 7” The timing of divergence is not the only factor that affects phylogenetic. accuracy—the relative position of taxa in the species tree (i.e., nodal position) also 228 Annals of the Missouri Botanical Garden 1 Locus ol 80 f > El o 60 QUA 3 am P n oo O i (D (Y E o, Recovery Figure focal taxa relative to the other species (i.e.. nodal position) is impacted whether the sister relationships of E and F were accurately recovered (Table 1). Moreover, position the effect of internode distance on phylogenetic accuracy (i.e., the tion term in Table 1 the ability t reconstruct the and F nodal also influenced \ interac- ). These effects on sister status of E no doubt Phylogenetic accuracy across the range of species ¢ 100 simulated species trees when species re eationships are estimated f. —> Nodal Position 2 —a- Nodal Position 3 | —— Nodal Position 4 —- Nodal Position 2 —o- Nodal Position 3 —— Nodal Position 4 - 0.20 E: 0.15 © @ NS Ed S „Ò S livergence times and internode lengths represented in the rom one and three loei, i marked by the different shapes. — respectively: the position o = ‘the reflect the probability that gene lineages from species » and F are likely to reach as deep as two, three, or more of the preceding species branchings (see also Shedlock et al., 2004). In addition to the obvious implication. this observation has for the effect of species sampling on phylogenetic accuracy, it also has Volume 95, Number 2 2008 Knowles & Chan 229 Evolutionary Radiation and Phylogenetic Inference LE o o œ o % Correct Recovery iyo) o . 10000 15000 Internode Length (c) 1 locus 100 + % Correct Recovery 30000 50000 0 10000 20000 40000 Species Divergence Time Figure 3. correct species a were recovered target taxa (7 ).80, time < 0.0001 for one locus [a] an internode length has been isolated % Correct Recovery 10000 15000 Internode Length 0 5000 % Correct Recovery 50000 40000 20000 Species Divergence Time 0 10000 30000 Phylogenetic accuracy (percent of the 100 replicate data sets for each of the 100 species trees in which the ) is rq influenced by both the length of the internode leading to the 68, P < 0.0001 for three loci [b]) and the species diverg gence (r? = 0.36, E < 0.0007 for one > loc us [c] and P = E 24, P < 0.001 for three loci [d]; the effect of divergence lime and ANCOVA based on either the residuals from a regression of percent correct recovery on either internode length, or divergence time, with node position as a covariate). important consequences for the ability to recover species relationships during evolutionary radiations. Because the species tree influences the distribution of gene genealogies (see also Degnan & Salter, 2005), the ability to recover the sister relationship of a particular pair of taxa depends on their placement relative to other species, in this case, the placement of species E and F relative to the other four species (Fig. 1). Further of this effect will provide important clues into how the ability to investigation reconstruct these relationships will change over time with the sorting of ancestral polymorphism. Table 1. times, the analyses were based on the resic divergence). However, this relationship depends on the shape of the species tree (i.e. The ability to recover the species relationships is a unfortunately expected to diminish with the sorting of ancestral polymorphism when using an approach such as minimizing the number of deep coalescents, as suggested by the decrease in accuracy with increasing e ^ species divergence time (Fig. 3). since any signa — apparent in the pattern of incomplete lineage sorting will be lost with time. While this suggests limited application of this approach for inferring relationships during evolutionary radiations that have occurred in the more distant past, it could be used as a tool to evaluate whether or not such relationships are likely Phylogenetic accuracy increases with increasing internode length (where, to control for differences in divergence uals from a regression of accurate phylogenetic recovery on the timing of species the relative position of the target taxa in the species tree) as is evident from the significant effect of node position and significant interaction term in the ANCOVA, Degrees of freedom Sum of squares F-ratio Prob > F One locus Internode distance ] Node position 2 Internode distance X node position 2 Three loci Internode distance 1 Node position 2 Internode distance X node position 2 7168.88 207.30 < 0.0001 : 1.5 0.0136 516.41 8.33 0.0005 13369,94. 136.38 < 0.0001 1512.06 7.7) 0.0008 2388.90 12.18 < 0.0001 Annals o Missouri Botanical Garden to be estimated accurately from DNA sequences. This general issue of which particular evolutionary scenar- ios will defy accurate phylogenetic estimation has been largely overlooked. despite its obvious implica- tions for historical inference. (see also Knowles & Maddison. 2002). CONCLUSION Despite the relatively modest sampling (i.e, one or three loci sequenced in 10 individuals per species) and challenging conditions (i.e.. the radiation of six species over a lime spanning just | N generations) considered here, species relationships were nonethe- less accurately estimated for many of the individual species histories examined (Fig. 2). As with estimates 2006), increased sampling of loci will no doubt also increase of population genetic parameters (Felsenstein, the accuracy of estimated species relationships when the process of gene lineage coalescence is incorpo- rated into the phylogenetic approach (Liu & Pearl, 2005: Maddison & Knowles, 2006; Carstens & Knowles. 2007). Moreover, these findings have important implications for how taxon sampling may influence the ability to recover species relationships and point to further investigation into how phyloge- netic accuracy may shift as the time since speciation increases and when taxa have undergone a rapid radiation. Literature Cited P. Beerli & J. B. divergence Arbogast, B., S. V. Edwards, J. Slowins m 2002. molecular data on phylogenetic and population ae Rev. Ecol. Syst. 33: a a 5. W. Daniel, C. F. Aquadro & R. DNA during the spec lation process in Peromyscus. Molec. Biol. Evol. Carstens, B. 2007. plivlogeny Bom gene-tree probabilities despite incomplete Wakeley, Estimating times from genetic time scales. Avise, J. C. A. ik “108: Mitochondrial ese Ww X C T « Y A é . Knowles. Estimating species lineage sorting: Sp from Melanoplus grasshop- pers. Syst. e : 1-12 Degnan. J. H. E: Salter. under the seent process. Evolution 59: 2 N Beerli 2000. boul ition divergence, 2005. Gene tree oe Edwards. S. & P. Jin d Gene divergence, and the variance in coalescence time in phylogeography studies. Evolution 54: 1839- i. I. n Liu & D. trees w ET concatenation. Proc. Natl. Acad. 104: 0941. Ho. n 2004. Inferring s NER ale Sunde rland, earl. 2007. High-re PrN iu ies \ Sinauer Associ- Massachus —., 2006. Do we need more thea more sequences, or mk Molec. Biol. Evol. 23: 691—700. Hasegawa, M.. H. Yano. 1985. Dating of the eee a splitting 2E a Ee c pen clock " mitochondrial DNA. J Accuracy of pan scent likelihood estimates: ` loci? a. & T. Molec. Evol. : 160-174. Hewitt, G. M. 2004. Genetic consequences of climatic ose SD in the Quaternary. Philos. Trans., Ser. B. 359: 183-19 Hey, J. & C A. Machado. pn hope for a difficult and science, Nat. Rev. Genet, 4: 535-543 Hudson, R. 1990. Gene gene ren m the process. Oxford Surv. E a iol. 7: 1-44. S 2003. of structured The study divided coalescent & J. A. Coyne. 2002. voire consequences of the genealogical species concept. Evolution 56: 1557- 1565. Jackman, T. R. 1 K. de Queiroz & J. Losos. 1999, He cana E and tempo i early 254—285. oe in Anoli is lizards. Syst. Biol. 48: í m nnings, W. B. & $ y grass finches Mee inferred from thirty V. Edwards. 2005. Speciational history gene trees, | volution 59: 203: Kingman, J. F. C The scd ‘ent. Stochastic Process Applic. 13: 235-248. Knowles, L. L. 2000. montane grasshoppers from the Tests of Pleistocene speciation in y islands of western North America (genus Melanoplus). ln 54: 1337- 1348. elaciations is Did the Pleistocene divergence? promote Tests of explic : Ne models in montane 69 gri dps rs. Molec. Ecol. 1-70 . 2004. The TE oning pee of statistical phylogeog- EN Biol. 17: 1-10. ain 2002. Statistical phylogeography. Ec ei : 202; ee & B. cally ony it model af po] ism divergence. Evolution 61: 477—493. e raphy. J. & Molec. pn . 2007a. Estimating a. geographi- ep t 2007 b. monophyletic gene trees. Syst. Delimiting species without Biol. 56: 887-895 Liu, L. € D. K. Pearl. 2007. Species trees from m trees: Reconstructing Bayesian posterior distributions of : pa cies phy lagé "ny using estimated gene tree ¢ distributions. . Biol. 56: 504—514. E. W. P. 1997. Gene trees in species trees. Syst. Biol. 46: 52: pe ———— & D. addison. 2004. Mesquite: A modular system for elton analysis. Version 1.01. Prolisbiities of monophyly, paraphyly, and roca de in a coalescent model. Evolution 61: 225-247. Shedlock, / , K. Takahashi € N. Okada. 2004. SINEs of Ia ois lineages with retroposons. Trends Ecol. Evol. 19: 545—553. Slowinski, J. B. 2001. Molecular polytomies. Molec. . Evol. 19: 114—120. Phylo- Tajima, F. 1983. Evolutionary O = is sequenc- es in abd popu s 105: 4 Takahata, N. 1989. Gene ge are ee in 3 EE popula- tions—Consistency probability between gene and popula- es. Genetics 122: 957-966. . Nei. 1985. Gene genealogy and variance of Genetics — ations. Gen tion tree oteo paatin] nucleotide differences. 5-344. ei & F. Tajima. 1982. Accuracy gon from molecular data. I. Distantly Kidd, T. Moum & V. L. Friesen. 1999. Polytomies amd we ana of phylogenetic inference. Evolution 53: 932- LIFE HISTORY PATTERNS AND Lynne R. Parent? BIOGEOGRAPHY: AN INTERPRETATION OF DIADROM Y IN FISHES’ ABSTRACT Diadromy, broadly defined here as the regular move d between freshwater and marine habitats at some time during their lives, characterizes numerous fish and invertebrate taxa. Explanations for the evolution of diadromy have focused on ecological requirements of individual taxa, rarely reflecting a comparative, phylogenetic component. When incorporated into phylogenetic studies, center of origin hypotheses have been used to infer dispersal routes. The occurrence and distribution of diadromy cho fish (aquatic non-tetrapod vertebrate) phylogeny are used here to interpret the evolution of this life history pattern and demonstrate the 1 elationship between life history and ecology in cladistic biogeography. Cladistic biogeography has been mischaracterized as rejecting ecology. On the e poses cladistic biogeography has been explicit in interpreting ecology or life history pate rns within the per [M work of Nem patterns. Today, in inferred ancient life history patterns, such as diadromy, we see remnants of previously s istum patterns, suc h as antitropicality or bipolarity, that spanned both marine a freshwater pants: Biogeographic 1 regions that span ocean basins and incorporate ocean margins better explain the rela Ip g vy. Its evolution, and its distribution than do biogeographic regions ‘entered on continents Key words: iua d distributions, biogeography, diadromy, eels, fishes, global biogeographic regions, life history, m. The extremely slow growth of the larvae of t Thomas Wemvss Fulton 855-1929) was a European eel is... an qus to the d iu — Scottish fisheries biologist best remembered today journey [to spawning grounds in the Sargasso Sea]. It is hoe Vis dasi The Sa aa " the Sea (Ful ` scarcely possible to understand this unique phase in the or his classic fhe Sovereignty of the Sea (Fulton, life cycle of the European eel on the hypothesis that the 1911), a treatise on the development of the notion o dii ia were formerly the same as now territorial waters. Fulton’s life began in 1855, a few exist. But if Wegener's theory [of continental drift] be years before Charles Darwin’s publication of On the iid: the explanation is simple. As the coasts slowly r^s > Origin of Species in 1859, and ended in 1929, just as receded from one another the larval life of what became 2 f I . . e. J iiie the European spec ies was more and more prolonged by Alfred L. Wegener published the fourth edition of Die natural selection in correspondence with the greater Entstehung der Kontinente und Ozeane, the proposal of distance to be il a theory of continental drift. In his letter to Nature, a —Fulton (1923: 360) f sia n quoted above, Fulton described a hypothesis for the evolution of catadromy in the European eel, Anguilla Animals that migrate between marine and freshwa- anguilla (L.), which incorporated modern concepts of er al some time during their life are diadromous, a — panbiogeography. First, the earth and its biota evolved ~ term coined by American ichthyologist George S. together, a tenet of botanist Léon Croizat (1958). Myers (1949a). Several categories of diadromy were Second, global biogeographic regions should be defined or redefined (Myers, 1949a, b: McDowall, centered on ocean basins, not continents (Croizal, 1988) collectively referred to as diadromy herein. Some from the first two, migration patterns evolve in concert anadromy, catadromy, and amphidromy— 1958; Craw et al., 1999: figs. 6-13). Third, following invertebrate taxa, such as atyid crustaceans and with geology, and these changes lead to lineage neritid gastropods, are also diadromous (e.g.. Myers differentiation, here changes in length of larval et al., 2000). period, and ultimately to what we recognize as E thank Richard L. Mayden, Peter H. Raven, and P. Mick Richardson for the invitation to present this paper in the 52nd Annual Systematies Symposium of the Missouri Botanical Garden, “Reconstrueting Complex Evolutionary Histories: Gene- S] ; and colleagues at the S 5 SE Mp esie in n ‘ular Kevin Conway, for discussion. John R. Grehan and James €. Tyler provided references. Malte C. Ebach, Anth ; Gill, David C. Smith, and Victor G. Springer read and commented on a be aft of the manuscript and DUE references Division of Fishes, rg nt z Vertebrate Zoology, Smithsonian Institution, P.O. M 37012, National Museum of Natural History. Room WG-12, MRC 159, W. ae D.C. 20013-7012, U.S.A. parentilOsi.edu. doi: 10.34.17/200605 | ANN. Missouri Bor. Garp. 95: 232—247. PUBLISHED ON 18 June 2008. Volume 95, Number 2 2008 Parenti Life History Patterns and Biogeography ee speciation. When Fulton referred to “... what became the European species...,” he postulated that there had been a widespread northern Atlantic lineage that, along with geological evolution of the Atlantic Ocean, differentiated to become the European eel in the North Atlantic rostrata (Lesueur), in the western North the American eel, A. Atlantic rostrata are hypothesized 2001a, 2005; see below). Other early 20th-century biologists adopted earth eastern and Anguilla anguilla and A. sister lineages (Lin et al., — listory to explain the evolution of migration patterns. Notable among ornithologists was Albert Wolfson of Northwestern University, Illinois, who, in 1948, wrote lead article in Science, “Bird Migration and the Concept of Continental Drift,” which correlated the increasing distances birds traveled with the separation of land masses and their latitudinal shift. Nearly four decades later, Wolfson (1986) vividly recounted the negative and positive reaction to his correlation. of bird migration and continental drift. Many modern explanations for the evolution of migration and diadromy conflict with those of Fulton and Wolfson because, for the most part, they reject the coupled evolution of life and earth, at global and local scales, and and rely exclusively on ecological evolutionary models. For example “...the relative productivity or growth advantage of sea and freshwater — habitats appears to be the key to [diadromy's evolution. The productivity differential can probably explain why fish migrate across the sea—freshwater boundary, predict their direction of movement and account for where in the world diadromous species occur" (Gross, 1987: 21). Earth history was separated from biological history, in part as a return to center of origin explanations for geographic distributions as outlined by Matthew (1915) in his influential, often (see Nelson & Ladiges, 2001). Equally influential was Myers, who reprinted. “Climate and Evolution" classified fishes in ecological categories for the purposes of biogeography. A decade before he defined diadromy, Myers (1938) proposed divisions of fresh- water fish taxa based on their ability to tolerate saltwater: freshwater fishes, including primary divi- sion freshwater fishes (no salt tolerance), secondary division freshwater fishes (some salt tolerance), and peripheral or migratory fishes (diadromous) versus marine fishes (salt tolerant). These categories have been incorporated into fish biogeography and are still used po today (e.g., MeDowall, 1988; Helfman et al., ; Berra, 2001). md eee fishes was modified by paleontol- (The category of primary ogist Colin Patterson [1975], who recognized archae- olimnic taxa, those inferred to have originated and always lived in freshwaters, and telolimnic, those living in freshwater now, but not necessarily through- out their history. These two categories have been used almost exclusively by paleoichthyologists [e.g., Hilton, 2003]. Following Myers's ecological classification, the distribution of primary division freshwater fishes has been interpreted with respect to earth history (viz. 1957), division freshwater fishes, peripheral fishes, and marine fishes has largely not (e.g., Briggs, 1974, 1995; but see Springer, 1982; Parenti, 1991; Mooi & Gill, 2002; Nelson, 2004). Accepting Myers's ecological categories precludes Darlington, whereas the distribution of sec- ondary the comparison of distribution patterns among dia- dromous, freshwater, and marine fishes: “It is plainly evident that a fish which can swim through sea water from one river mouth to another is not of much use in (Myers, 1938: believed by and studies of terrestrial 343). And so it contemporaries (e.g., Herre, 1940) and subsequent zoogeography" was Myers generations of biogeographers. Adhering to Myers's philosophy, biogeography of marine and freshwater fishes on a global scale remains largely independent. Textbooks and courses have traditionally drawn the line, contrasting Zoogeography of the Sea (Ekman, 1974) with 1990, 992, 1995), for example. More general texts, such 1953) and Marine Zoogeography (Briggs, Zoogeography of Fresh Waters (Bánárescu, as Zoogeography: The Geographical Distribution of 1957), r Biogeography, an Ecological and Animals (Darlington, (Briggs, 1995), o Evolutionary Approach (Cox & Moore, 2005), likewise Global Biogeography have divided their attention between marine and freshwater taxa. The arguments for dispersal (e.g., Gross, 1987; MeDowall, 1970, 1988, 2001. 2002; Berra et al. 1996; Keith, 2003), 1974; Croizat et al., 1974), or some combination of the two (e.g., Choudhury & Dick, 1998) as explanations for vicariance (e.g.. Rosen, the distribution of diadromous taxa, especially the antitropical salmoniform fishes, have been debated extensively. Likewise, broad geographic ranges have been interpreted as explicit evidence of e.g., McDowall, 2001, 2002), of Rosen, 1974), of either dispersal or vicariance 1984). distribution. (above) will not convince a dispersal —— vicariance (e.g., (Leis, Fulton’s Schrank dispersalist to embrace vicariance or a vicariance explanation for Anguilla biogeographer to adopt dispersalist explanations, as each represents a particular view of the relationship between the organism and the environment in forming biogeographic distributions. adistic biogeography (viz. Nelson & Platnick, 1981; Humphries & Parenti, 1986, 1999; Crisci et al., 2003) puts primary emphasis on the phylogenetic relationships among organisms, not their physiological or ecological requirements, to discover distribution Annals of the Missouri Botanical Garden Cladistic biogeographers do not ignore id but. “life history patterns. ecology (viz. Wiens & Donoghue, rather, are explicit in interpreting ecology o patterns within the broader framework af phylogenetic 2005). Detailed and commentaries on the systematics, biology, patterns (e.g., Sparks & Smith, reviews and distribution of diadromous fishes, largely within the dispersalist framework, have been provided by New Ze al me le s Robert M. McDowall (e.g.. 1970, 1987, 1988, 1992, 1993, 19972, b, 2001, 2002, 2003). | ne argued mene that the distribution of each genus of diadromous gobies of the subfamily Sicydiinae (family Gobiidae) may be described by the ocean basins in which it lives and, further, allopatry level: the Sicydium Valenciennes does not may be recognized at the generic eastern Atlantic overlap with its sister genus Sicyopterus Gill i Pacifie and in the Indian (Parent Ocean and western and central Pacific , 1991; 1998). My p purpose here is to reinterpret the phylogeny and both marine, within a cladistic biogeographic framework to ask if, diadromous fishes Parenti & Thomas, global distribution of fishes. freshwater and and if so, how, the biology and ecology of are related to their distribution patterns and to competing models of earth history. DIADROMY AND CLASSIFICATION OF FISHES Table 1 lists the major groups of living aquatic vertebrates that we call fishes, the estimated number ol species in each taxon, the number of species that live almost exclusively in freshwater, and the total number of species that enter freshwater during some phase of their lives (modified from Nelson, 2006: 4— 5). Diadromous species are included in the last column, which also enumerates those species that do not undergo regular habitat migrations but are euryhaline, such as eyprinodontiform killifishes (see, 1973). An diadromous (McDowall. e.g., Rosen, estimated 225 species are 1992: Nelson. 2006). Not al obligately so (McDowall. diadromous species are 2001), precluding a precise enumeration of such taxa. A large concentrated among inferred basal lineages: lampreys proportion of diadromous species is (Petromyzontiformes). sturgeons (Acipenseriformes), eels (Elopomorpha), herring (Clupeomorpha). and salmon (Protacanthopterygii), as noted by McDowall (1988. 1993) and others, and many species in these lineages are euryhaline. Many, but not all, diadromous fishes are also antitropical or bipolar. distributed the northern boreal and/or the southern austral zone. and absent from the tropics, again as noted by McDowall (1988, 1993) and others. Salmoniforms and 1974: fig. 15). lampreys (Petromyzontiformes; Berra, 2001: 6, 8, osmeriforms (Protacanthopterygii; Rosen, Nelson, 1986) bipolar distributions, whereas sturgeons (Acipenser- Choudhury & Dick, 1998; 2001: 42) and sticklebacks (Gasterosteidae; Berra. 2001: 351) are 12) and some clupeomorphs have idae: Berra, boreal, for example. EPICONTINENTAL SEAS? THE SETTING FOR BoNY Fisu EVOLUTION Living actinoplerygian (ray-finned) fish clades are — the bichirs (Polypteriformes), sturgeons and paddle- fishes (Acipenseriformes), gars (Lepisosteiformes), bowfins (Amiiformes), and the bony fishes (Teleostei; see Nelson, 2006). Relationships among these line- ages have been debated by both morphologists and molecular biologists (e.g., Nelson, 1969a; Patterson, 1973; Wiley & Schultze, 1984; Inoue et al., 2003) and, for the purposes of this discussion, may be summarized by the cladogram of Figure 1. Polypteriformes (e.g.. Britz, 2004) live in African freshwaters. Fossil bichirs of Africa and South America are of Middle Cretaceous age (Nelson, 2006). 1991) com- prise the anadromous and freshwater eireumboreal Poly- odontidae. There are two extant species of paddlefish, Acipenseriformes (Grande & Bemis, sturgeons, Acipenseridae, and the paddlefishes. one in freshwater in North America and the other ii freshwater and estuarine habitats in the Yangtze River and associated areas of China. Fossil paddlefishes date to the Lower Cretaceous from China (Berra. 2001). Sturgeons are an old lineage, known from at least the Middle Jurassic (175 Choudhury & Dick, 1998). Lepisosteiformes includes one living family, Lepi- million years ago |Ma]; sosteidae (Wiley, 1976), which comprises two extant genera with seven species living in fresh, brackish, North and Central gars are marine waters in America and Cuba. Fossil more broadly distributed throughout North and South America, Europe, Africa, and India (Wiley, 1976: fig. 69); they date from at least the Early Cretaceous (Janvier, 2007 1998, 1999) are Amia calva L., Amiiformes (Grande & Bemis, represented by one living species, distributed in freshwater habitats in North America. It is the single, living representative of the Halecomorphi, known from marine and fossil taxa dating to the Jurassic. The ancestral habitat for bony fishes is inferred from the phylogeny and distribution living and fossil basal actinopterygian taxa to be shallow freshwater margins, (Fig. 1). A widespread ancestral range was illustrated for a epicontinental seas and their including river. deltas reconstructed — subgroup of halecomorph fishes by Grande and Bemis (1999: f present-day geography contrasted. with that of the ig. 6), who outlined the distribution on a map -= Volume 95, Number 2 2008 Parenti Life History Patterns and Biogeography ble 1. Classification of living aquatic vertebrates (non-tetrapods) or fishes, estimated number of species in each taxon, Table number of those that live almost exclusively in freshwater, and total number of those that enter freshwater al some time during their life (from Nelson, 2006: 4—5; Stiassny et al., 2004). The last column includes all of the species in the second column, plus those that are migratory (diadromous or euryhaline). No. of No. of Total no. of species Taxon species freshwater species that enter freshwater Myxiniformes 70 0 0 Petromyzontiformes 38 29 38 Elasmobranchii 937 24 10 Chimaeriformes 33 0 0 Osteichthyes Sarcopterygii Coelacanthiformes 2 0 0 Ceratodontiformes 6 6 6 Actinopterygii Polypteriformes 16 l6 16 Acipenseriformes 27 14 27 Lepisosteiformes 7 6 7 Amiiformes 1 l l Teleostei Osteoglossomorpha 220 220 220 Elopomorpha 857 6 33 Clupeomorpha 304 79 85 Ostariophysi 7980 7847 7858 ee 356 127 152 Esociformes 10 10 10 Stomil i: m 391 0 0 ims mes 12 0 0 Aulopiformes 236 0 0 Myctophiformes 246 0 0 Lampridiformes 21 0 0 Polymixiiformes 10 0 0 Paracanthopterygii 1340 2] 24 Stephanoberyciformes 15 0 0 Zeiformes 32 0 0 Beryciformes 144 0 0 Gasterosteiformes 278 21 13 Synbranchiformes 9 96 99 Percifort 10,033 2040 2335 Mugiliformes 72 l 7 Atherinomorpha 1552 1304. 1352 'orpaeniformes 1477 60 62 Pleuronectiformes 618 10 20 Tetraodontiformes 357 14 22 ‘ S 27,977 11,952 12,457 Early Cretaceous, 118 Ma. The present-day, disjunct transatlantic distributions are continuous when drawn across Cretaceous epicontinental seaways. An area cladogram, or areagram, of halecomorphs indicating habitat, marine or freshwater (Fig. 2A), was by Grande and Bemis (1999: fig. 7) to infer a marine ancestry for the group. This hypothesis results from optimizing habitat or, at the least, inferring that the habitat of the basal taxon is the ancestral habitat. [t rests on the assumptions that the habitat has Qu. usec changed (e.g., marine to freshwater), even when there may be no evidence for such a transformation, and that one or the other taxon of a sister group pair may be identified as basal (see Parenti, 2006; Santos, 2007; also see below). Without these assumptions, the ancestral habitat may be reconstructed as epiconti- nental seas, spanning marine and freshwater habitats (Fig. 2B). Africa also represent a transatlantic distribution, that Likewise, bichirs in South America and is, a remnant of the broader distribution. pattern. Annals of t 236 Missouri a Garden Polypteriformes Epicontinental Seas and Acipenseriformes Freshwater margins Lepisosteiformes Amiiformes Teleostei Figure l. Bemis (1999). Paddlefishes reflect transpacific relationships, as do other North American taxa (Grande, 1994a, b). A lineage with marine and freshwater representa- tives (Fig. 34) may be interpreted to have a marine and freshwater ancestral distribution (Fig. 3B). In a likelihood approach to reconstructing ancestral ranges of lineages, Ree et al. (2005: fig. 6c) also reconstruct a = widespread ancestral range equivalent to that o Figure 3B when dispersal is considered unlikely. that rocess is specified to explain the 1999; 2007). variable ancestors" are analogous to - Is, when no a priori ER. pattern (see also Ebach, Santos, “Ecologically “geographically widespread” ancestors (see Hardy, 2006: 13). (The method of Ree et al. [2005] does not correspond to other cladistic biogeographic methods, in part because it defines an area as a discrete geographic unit, not the organie areas of endemism that are defined by the ranges of the organisms that inhabit them [see Harold & Mooi, 1994.) Our notion of freshwater and marine fishes comes, in large part, from the existence of speciose taxa that such as the Ostar- nearly all. freshwater, characins, minnows, and relatives) or are today iophysi (catfish, Osteoglossomorpha (bony tongues), or from deep-sea marine orders, such as the Myxiniformes (hagfishes). Chimaeriformes (chimaeras), Stomiiformes (dragon- fishes), Aulopiformes (lizardfishes). (lanternfishes), and so on (Table 1). Osteoglossomor- Myctophiformes Hypothesized relationships among living actinopterygian clades, following Inoue et al. The ancestral distribution is inferred to be throughout epicontinenta (2003) and Grande and | seas and their freshwater margins pha is the sister group to all other teleosts. all extant of which live exclusively in freshwater (Nelson, 1969b, 2006; Table 1). It is not historically freshwater lineage: Paleocene-Eocene fos- the Spec ies — marine sils, such as genus TBrychaetus Agassiz, are hypothesized to be nested the osteoglossomorph in the family Osteoglossidae (Hilton. cladogram 2003). The oldest osteoglossomorph fossils are Middle Mesozoic, and Bánárescu (1995: 1162) acknowledged their long history and widespread ancestral distribu- tion by proposing that Osteoglossomorpha is Pangean. Osteoglossomorpha, therefore, has both marine and freshwater representatives and is a freshwater group today because of widespread extinction of marine taxa. If the majority of species in a taxon are marine and just a handful are freshwater, it is often assumed that the group originated in marine habitats and several laxa dispersed to freshwater (e.g., Myers 1938; Berra, 2001: 2002). The reverse is also assumed. Such an assumption is not an Tsukamoto et al., analysis of biogeographic data, but an implicit and untestable center of origin hypothesis (viz. Croizat et al., 1974), Both Nelson (1973) and Patterson ( er habitat of origin of 1975) optimized habitat on areagrams to infer osteoglossomorphs and came to different conclusions. (1975: 162) that view of the history of For example, Patterson asserted “if the conventional Volume 95, Number 2 Parenti 237 8 Life History Patterns and Biogeography FW M/FW M M/FW FW M(FW) M(FW) M FW M/FW M M/FW FW M(FW) M(FW) M M/FW | M/FW | M/FW | M/FW | MEW | M/FW | M/FW Figure 2. —A. Areagram of halecomorph habitats from Grande and Bemis (1999), who inferred a marine ancestry for the group. M/FW means distribution in marine and freshwater habitats; M (FW) means distribution mostly in marine habitats with some representatives in freshwater. —B. Reconstruction of nodes to infer a widespre: ad marine/freshwater ancestral habitat, interpreted here as epicontinental seas | their freshwater margins. 238 Annals of the Missouri Botanical Garden B Figur re 3 I ly pothetical areagram represe nting taxa in marine (M) and freshwater M/FW FW M/FW M/FW M/FW M/FW M/FW M/FW . Optimization of (FW) habitats wid to infer a marine center of origin. —B. Reconstruction of nodes to infer a widespread marine/fr ms ancestral habit Gondwana is accepted, marine origin of the Osteo- Nelson (1973: 9), who elossomorpha is indicated. . the read Patterson in manuscript, countered that totality of evidence concerning the relationships and both fossil and — distribution of osteoglossomorphs, indicates that FBrychaetus is secondarily Recent, "that is, that osteoglossomorphs originated in marine,’ freshwaters. Even though Nelson later dismissed such arguments (e.g., Nelson, 1974; Nelson € Ladiges, 2001), they persist (see Parenti, 2006). Optimization of habitat on an areagram has been applied broadly to interpret the ecological and distributional history of lineages (see, e.g., Brooks € McLennan, 1991; McLennan, 1994). Optimization of habitat is often justified by the principle of parsimony, yet, in practice, can require Volume 95, Number 2 8 Parenti 239 Life History Patterns and Biogeography multiple hypotheses of switching between habitats (e.g., Baker, 1978). The distributional history of gobioid fishes, for example, was hypothesized by Thacker and Hardman (2005: 869): * ancestor, . gobioids arose in freshwater, from a marine then returned. to marine habitats once or many times.” Whether diadromy or euryhalinity is a primitive or derived life history pattern for fishes has been debated seemingly endlessly (Tchernavin, 1939; Denison, 1956: Patterson, 1975; Griffith, 1987; McDowall. 1988, 1993, 1997a; Johnson & Patterson, 1996; Bemis & Kynard, 1997; Waters et al., n Janvier, 2007). throughout epicontinental seas, including along their 2000; see especially The distribution of fishes review margins, as proposed for basal actinopterygians, is in accord with numerous explanations for the evolution of solely freshwater or marine taxa from widespread diadromous or euryhaline ancestors. Suppression of the marine or of the freshwater life history phase of a migratory lineage could lead to evolution of solely freshwater or marine fishes, respectively, as speculated by Patterson (1975: 168—169). In addition to evolution of development, the roles of extinction and earth history have been recognized: isolation or stranding through changes in sea level or other vicariant events can result in a former euryhaline species being restricted. t freshwater, for example (e.g.. Choudhury & Dick, 1998; 2000; Heads, 2005a), a phenomenon termed (Craw et al., 1999 Widespread extinction of the earth’s biota since the Waters et al., “ecological stranding” Upper Mesozoic has left just remnants of these formerly widespread distributions, “the ring on the bathtub” te paraphrase Heads (1990: 225). Isolation in one habitat or the other by stranding does not require or imply invasion of that habitat (see also Heads, 2005a). Some taxa persist in freshwater, others in marine habitats, and some in both. EaRnTH Hisrory MODELS Plate tectonics is the model used most often to interpret. global distribution patterns with respect to earth history (e.g., Rosen, 1974; Patterson, 1975; Springer, 1982; Choudhury € Dick, 1998; Grande & Bemis, 1999). The model specifies an Atlantic Ocean expanding since the Mesozoic and supports vicariant explanations for transatlantic taxa, yet leaves bioge- ographers scrambling for similar explanations, such as migration of allochthonous terranes, for the distribu- taxa (McCarthy, 2003: 1556). A vicariance explanation. of transpacific taxa finds geological support in the still-maligned theory of Expanding Earth, which specifies that the age of the Pacific Atlantic tion of transpacific o the same as that of the 1979, 1983, Ocean is roughly and Indian oceans (Shields, 1991, 1996; McCarthy, 2003, 2005). Biogeographic sister areas across the Pacific have complementary. matching geological outlines (McCarthy, 2003: fig. 3). That there are competing models of earth history endorses the view that biogeographic patterns should be discovered, then interpreted with respect to these models, rather than the models a priori specifying biogeographic patterns (Ebach & Humphries, 2002). Minimal divergence times between lineages have been estimated using ages of fossils or a molecular clock and used to test hypotheses of dispersal versus vicariance (e.g., Rosenblatt € Waples, 1986; San- 2001; Burridge, 2002, to cite just three Minimal estimates from sequence data martín et al., examples). have been interpreted as absolute, thus calling into — question earth history—Plate Tectonics or Expanding Earth—interpretations of distribution patterns (e.g. de Queiroz, 2005). This interpretation has countered by Heads (2005b), Parenti. (2006), others, continuing a long debate within biogeography that techniques (see especially Nelson, 2004). Teleost as been and was revived with the advent of molecular well as basal actinopterygian lineages are old enough to be interpreted within earth history models. Fossils of living teleost lineages date from about 150 Ma in ade are estimated to be at least of Triassic age 2007; see also Benton & Donoghue, 2007). Differentiation is the the Upper Jurassic, and teleosts as a (Janvier. also ancient: tetraodontiforms, nested within apomorphie percomorph clade, are represented by fossils from the Upper Cretaceous of Slovenia, Italy, and Lebanon, the oldest of which dates from 95 Ma (Tyler € Sorbini, 1996). DISTRIBUTION PATTERNS OF DIADROMOUS FISHES Historical biogeographic patterns may be complex, and are, by definition, hierarchical. They are characterized by repeated elements that have become the basis of broadly recognized global distribution patterns (Croizat, 1958, 1964) in a classification (Fig. 4): Atlantic, (Pacific, Antitropical), or by fragments or composites of these areas. One stated aim of historical biogeog- that may be expressed Indian Ocean, raphy is to develop a standard language or nomen- clature for global distribution patterns (e.g., Morrone, 2002) including their fragments. Fragmentation may be caused by extinction, which has left remnants of formerly complete distribution patterns, such as a lone freshwater representative of a taxon once living broadly in marine and freshwater habitats. These remnants have been used to interpret, and misinter- pret, the distributional history of taxa (see Heads, 2005a, b; Nelson & Platnick, 1981; Nelson & Ladiges, 1991, 1996; also see below). Annals of the Missouri Botanical Garden Atlantic Indian Pacific Austral Figure 4. Areagram or classification of Croizal's (1958 Nelson and Ladiges (2001: fig. 8b) The distribution of antitropical or bipolar taxa has been interpreted with regard to evolution of the Pacific Basin (e.g., Humphries & Parenti, 1986, 1999; Nelson, 1986). In a recent discussion of the evolution O of diadromy among fishes of the southern oceans, McDowall (2002: 208) acknowledged but did not discuss bipolarity. He divided fishes of New Zealand into two kinds: those with close relatives throughout the southern oceans, such as lampreys and galaxioids, both groups bipolar, and those that he inferred, a priori, to be part of another global pattern: “Anguillid eels and gobiid [fishes] have nothing in common with the southern cool temperate fish fauna regarding historical origins, distributions and relationships. probably They are not Gondwanan and and vel, anguillids gobiids express a series of relationships and distribu- tions between the faunas of eastern Australia and New Zealand that are essentially the same as those seen in (McDowall, 2002: 212-213). Sharing a pattern implies sharing a history: I examine the southern families” (italics added) ' distribution of eels and gobies to ask if they have anything in common with the southern cool fish fauna in other words, whether they conform to the same global as well as local patterns. PHYLOGENY AND DISTRIBUTION: EELS Kels of the genus Anguilla are catadromous (Myers. 19192): adults live in freshwater streams and migrate to marine spawning grounds (Tesch, 1977; Smith, 19893, Boreal 3) and Craw et al/s (1999) global biogeographic patterns. following b). Phylogenetic relationships among the 19 extant taxa ~ Fig. 5) were hypothesized using molecular data by two research groups (Lin et al., 2001a, 2001). : and these two sets of analyses were contrasted 9; Aoyama et al., and compared with morphology (Lin et al., 2005). An areagram of Anguilla taxa, with names of Figure 5 replaced by a brief description of distribu- taxon tional limits, is given in Figure 6. Another mo ecular phylogeny based on whole mitochondrial genome sequences was published by Minegishi et al. (2005). we Anguilla species are associated with all continents, except Antarctica, and do not live in the eastern Pacific or South Atlantic (Berra, 2001). Distribution of anguillid eels has largely been interpreted. within a 1968) and, recently, to conflicting ends (Heads, 2005a: 700): Lin et al. (2001a, b) put the center of origin of eels in the southwest Pacific and proposed dispersal eastward dispersalist paradigm (e.g., Harden Jones, across the Pacific and further across the Central American Isthmus to enter the Atlantic Ocean. whereas Aoyama et al. (2001), don and Tsuka- and Tsukamoto et moto (1997), . (2002) hypothe- sized dispersal from a center of origin near Borneo, westward through an Atlantic Ocean. rejected by ancient Tethys Sea to the 3oth of these explanations were (2005: 141), «the present geographic distribu- Minegishi et al. who concluded that tion [of eels of the genus Anguilla| could be attributed — o. for example, multiple dispersal events, multidi- rectional dispersion, or past extinctions, ele.” Volume 95, Number 2 2008 Parenti 241 Life History Patterns and Biogeography A. marmorata Quoy 8 Gaimard A. interioris Whitley > . malgumora Kaup > . bengalensis labiata Peters > . bengalensis bengalensis (Gray) obscura Gunther > . bicolor bicolor McClelland > D . bicolor pacifica Schmidt D . celebesensis Kaup A. megastoma Kaup . reinhardtii Steindachner A. japonica Temminck & Schlegel A. australis australis Richardson D> . australis schmidtii Phillips TT D . dieffenbachii Gray > anguilla (L.) rostrata (Lesueur) > > mossambica (Peters) Figure 5 arbitrary. Anguilla malgumora, as used by Lin et al., is a synonym of A. borneensis, the species name used b A. borneensis Popta Cladogram of relationships of eels of the genus Anguilla, following Lin et al. (2001a). Br a lengths are ! Aoyama et al. (2001) and MEE gishi zb al al i pelis of species names and authors follows Eschme 'yer's online cate dies (http://www. cala g Panbiogeographers have long endorsed the view that the distribution and evolution of eels of the genus Anguilla may be interpreted with respect to earth history, specifically evolution of a Tethys Sea biota 1958), and hypotheses (see Heads, 2005a). Cladistic biogeogra- (Croizat, dismiss ad hoc dispersal phers endorse the panbiogeographic view and add relationships among areas, as interpreted from log/), accessed on 19 October 2007 phylogenetic hypotheses, as a framework for further refinement of distributional history patterns. Rather than optimizing areas, the information in the areagram of Anguilla taxa (Fig. 6) may be summarized by a variety of methods, such as reduced area-cladograms or paralogy-free subtrees (Nelson & Ladiges. 1996). or by comparing it to the cladistic summary of Croizat's (1958) global biogeographic patterns (Fig. 4). Notable 242 Annals of the Missouri Botanical Garden Indo-Pacific E New Guinea Borneo, Sulawesi, Philippines fare E Africa, Reunion, Mauritius CON p N Indian Ocean, East Indies | W New Guinea, Queensland, Society I. Indo-West Pacific | Indo-West Pacific W Pacific Pacific E Australia, Tasmania, NG, NC, LHI 3i 3 NW Pacific SW Pacific New Caledonia New Zealand 4 [cum E North Atlantic ERN W North Atlantic W Indian Ocean, NC Borneo, Sulawesi, Philippines Figure 6. Areagram of Anguilla laxa, with a names of Figure 5 replaced by brief description of distributional limits. NG = New Guinea, NC = New Caledor AM = Lord Howe Island. Shaded ; area l: western margin of Pacific Basin; shaded area 2: Indian Ocean: shaded area 3: anb iouis al shaded area 4: antitropical. in the areagram of Anguilla (Fig. 6) are laxa and these are, in turn, sister to a group of Pacific taxa. distributed along the western margin of the Pacific Another clade is also antitropical: the North Atlantic Basin (shaded group 1). the Indian Ocean (shaded sister species (boreal) are closely related to a group of group 2), antitropical distributions, and antitropical southwest Pacific, New Zealand, and New Caledonian taxa sister to Pacific taxa. Eels demonstrate one of the taxa (austral; shaded group 4). Freshwater eels of the three-area relationships in the global biogeographic genus Anguilla have not been characterized as pattern (Fig. 4): Pacific (boreal, austral). Northwest antitropical or bipolar, as antitropicality of fishes is Pacific taxa (boreal) are closely related to those from seen largely as a phenomenon of strictly marine taxa eastern. Australia; Tasmania. New Guinea. and New (e.g.. Nelson, 2006: 12). These remnants of antitropi- Caledonia (austral, with additional elements from the cality would likely not be identified without a western margin of the Pacific Basin: shaded group 3). hypothesis of phylogenetic relationships or without Volume 95, Number 2 2008 Parenti 243 Life History Patterns and Biogeography the aim of identifying general patterns, rather than hypothesizing individual distributional histories. Are anguillid eels ancient, like other antitropical taxa? Perhaps. A molecular clock estimate of 20 Ma a for the beginning of Anguilla speciation was rejectec by Minegishi et al. (2005: 141) as an underestimate because of fossil evidence: anguilliforms date from the Cretaceous (113-119 Ma) and the genus Anguilla from the Focene (50-55 Ma) (Patterson. Upper 1993), both minimum estimates of the age of these lineages. PHYLOGENY AND DISTRIBUTION: GOBIES a Gobioid fishes comprise over 2000 species distrib- uted broadly in pantropical and temperate freshwater, estuarine, and marine aquatic habitats (Nelson, 2000). Because the vast majority of gobioids live in marine waters, they have been characterized as a marine 1938; Berra, 2001). subgroups, such as Gobiomorphus Gill, and the gobiid group (e.g., Myers, Several subfamily Sicydiinae are diadromous (e.g., Parenti & Maciolek, 1993; Keith, 2003). Phylogenetic analysis of sicydiine gobies and their relatives at the generic level (e.g., Parenti, 1991; Parenti & Thomas, 1998) has been offered as evidence that the repeated, recognizable biogeographic areas for gobioids are ocean basins, not continents (see also Springer, 1982; 2000) Hypotheses of relationships among gobioids (e.g.. Gill in Lundberg et al., Thacker & Hardman, 2005: fig. 1) are preliminary, yet, when combined with distribution, display remnants of elobal patterns (Fig. 4): Rhyacichthyidae, the sister group to all other gobioids, lives in the tropical western Pacific (Berra, 2001: 458 sister to the remaining gobioids, exhibits a remnant ) whereas Odontobutidae, boreal distribution, living in freshwaters of the northwest Pacific (Berra, 2001: 460). The southeast Australian and New Zealand Gobiomorphus, a goby genus that McDowall (2002: *, 213) characterized as “probably not Gondwanan," is interpreted as having a remnant of an austral distribution; it is sister to a large gobioid lineage distributed throughout the pantropics. The fossil record of gobiids is scant; they date from at least the Middle Tertiary (Patterson, 1993). Endemism of gobioids on islands or island groups is high, notably throughout the Indo-Pacific (Maugé et al.. 1992; Parenti € Maciolek, 1993, 1996; Larson, 2001; Donaldson & Myers, 2002). Endemism through- out the marine realm is underappreciated in large part because marine species are defined traditionally as Gill & Kemp, 2002, fishes; Meyer et al, 2005, gastropods). widespread (e.g.. Endemism may be under- recognized at both lower (species) and higher (genera) taxonomic levels. The lack of genetic structure among populations of the five freshwater or amphidromous gobiid fish living streams (e.g., Chubb et al., 1998) prompted McDowall 2003) to conclude that they must have colonized species throughout Hawaiian po Hawaii by long-distance, chance dispersal. Alterna- tive Wadi) are plausible. Four of the five species are endemic | Hawaiian Islands, as McDowsll a 705) acknowledged. Teditional island biogeography tells us that what inhabits an island or island group is what is able to physiolog- ically and physically stand a long ocean voyage from he mainland (MacArthur & Wilson, 1967), although islands, as well as continents, are inhabited today by o the remnants of once more diverse faunas. Islands and coastal zones support life that is able to withstand the wide swings in salinity, temperature, and so on (see also Heads, 2005a), and chief among that life is the diadromous or euryhaline taxon. CONCLUSIONS Biogeography of diadromous fishes is not about whether they are capable of surviving in the seas. They are, history. [t is not about whether diadromous fishes have at least during some phase of their life — freshwater ancestors because all fish both. fishes is like that of all other marine or lineages have Biogeography of diadromous taxa: a search for a pattern. of. relationships among endemic areas that conforms to or differs from general distributions of taxa worldwide. Eels and gobies, phylogenetically disparate and distinct diadromous taxa, reflect global can acknowledge this — biogeographic patterns. One « congruence, this conformity to a pattern, or one can that distribution of diadromous taxa based on their life ignore it. Predictions may be made about history are not necessarily corroborated by their phylogeny or their distribution patterns. The answer to “Why migrate?” is straightforward when earth history is considered: migration evolves as the earth 1948; Croizat, 1958; Heads, 2005a). Migration. of indemana species is “the result...not the cause” of their evolution (Wolfson, 1948: 30; see also Grehan, 2006). When interpreting the evolution of diadromy in fis evolves (see also Wolfson, — hes, or of any migration pattern, the relationship between phylogeny and distribution must be considered along with ecological patterns. Ecology, phylogeny, and distribution are inseparable (see also Heads, 1985). The ability to live in both freshwater and marine habitats is interpreted as an ancestral life history pattern for fishes. The evolution of diadromy is tied to the evolution of ocean basins. Optimizing habitat on the nodes of an areagram to invoke an ancestral habitat, here marine or freshwater, is equivalent to Annals of the Missouri Botanical Garden invoking a center of origin for areas. A repeated Qa s general pattern among taxa that are not closely relatec and/or that differ in ecological requirements is most parsimoniously explained by the same events. There should be no separate global historical biogeography for marine and freshwater fishes. Proposing separate explanations for each prevents discovery of general patterns. Today, we see but a remnant of the past distribution of life on earth. Global biogeographic patterns are the organizing framework for interpreting. distributional and ecological history of both marine and freshwater taxa. Recognizing that anguillid eels have remnants of antitropical patterns, for example, is a first step toward accepting that their distribution may be classified along with that of other taxa, rather than a priori requiring a unique explanation. Literature. Cited J. & 1997, . Tsukamoto. freshwater eels. Naturwisser nschaften 84: M. & K Aoyama, Evolution of the 1721. Nishida Tsukamoto. 2001. Molecular loe Bind evolution E the freshwater eel, genus Anguilla. Molec. Phylogen. Evol. 20: 450—459. Baker, R. 1978. The Evolutionary Ecology of Animal 'ol Migration. oo & Stoughton, London Bánárescu, P. ol Fre "t Waters, Vol. General and Dispersal of Freshwater e Aula- dé rlag, Weisba 992. esa of JO. Zoogeography Fresh Waters, Vol. 2: a and Dispersal of Freshwater Animals in North America and Eurasia. Aula-Verlag, Weisbaden. 1995 Waters, Vol. : sal of Fre Pe din Animals in nes ‘a. Pacific Areas, Meam America. Aula- Bemis, W. E. & Kynard. 1997. introduction to acipenseriform biogeography and life history. Environm. Biol. h 67-183. Benton, M. J. & P. C. P. 2001. iau is evidence to date the tree of life. Molec. Biol. Evol. a d of Fres as and Dispe Verlag, Weisbaden. Sturgeon rivers: An Fi Isnes 48: Donoghue. + Ber ‘rra, . M. 2001. Press, New York P E. rowley, W. Ivantsoff & P. A, 1906. Galaxias. mac Ny An d of its biogeog- vater Res. 47: 845-849 (4. Marine Saas ee M Freshwater Fish Distribution. Academic Fuerst. ceGraw-Hill, New . 1995. Global Biogeography. Elsevier, New York. Britz, R. 2004. Polypterus teugelsi, a new spec ies of bichir from the Upper iross River system Cameroon (Actinopterygii: 2 ladistia: Polypte ridae), Ichthyol. Explo Freshwaters 15: 209-224. Brooks, D. R. & D. A. McLennan. pA Phyloge ny, Ecology, Univ. of Chic = and Behavior. cago Press, Chicago. Burridge, C. P 2002. Antitropicality of Pacific fishes: Molecular insights. Ton Biol. Fishes. 65: 151- Choudhury, A. & T . 1998. The ography of sturgeons (ie li historical n Acipenseridae) synthesis of phylogenetics, palacontology and palae ogeo- graphv. J. Biogeogr. 25: 623-640 — — & W. E. Chubb, A. L., R. M. Zink & J. M. of mtDNA variation in. Haw phylogeographie consequences e fone J. Hered- ity 89: 8-16. Cox. C. Fitzsimons. 1998. Patterns freshwater fishes: the b. & P. D. Moore. 2005. Biogeography, an Ecological and Evolutionary Approach, 7th ed. Blackwell Publishing. Oxford. Craw, R. C., J. R. Grehan € M. J. Heads. 1999, ; Panbioge- ography : Trac iur the History of Life. Oxford Univ. Press, Oxforc Crisci, J. V., L. Katinas € P. Posadas. 2003. Historical Biogeography, an Introduction. Harvard Univ. Press, Cambridge. Croizat, L. 1958. Panbiogeography. Published by the author. Caracas. ace, Time, Form: Published by the author, The Biological Synthesis. Caracas . G. Nelson & osen. 197 t. Centers of origin and related concepts. Syst. Zool. 23: 265-287. Darlington, P. J., Jr. 1957. Zooge n The Ge ographio cal Distribution of i RD 1859. . John Wiley € Sons. New Yor Darwin, C. Jn the Origin of Species by Means E Natural deuda or the Preservation of Favoured Races in the BE ve or Life. John Murray, London. Denison, R. . A review > pic he a at of the earliest erte ine "s. j^ le AU Geol. 9]- de Quei iroz, A. 2005. The resurrection 2 oceanic SpA sal in historical biogeography. Trends Ecol. Evol. 20: 68-73. Donaldson. T. J. & R. F. Myers. 2002. Meis UE Patteri Biol. s of species richness an Fishes. 65: 139-149 . Environm. Ebach, M. C. 1999 Paralogy and the centre of origin concept. C Cladistic 's 15: 387-391. . Humphries. 2002. n p a . Biogeogr. 29: 444. and i art of discovery. Ekman, S. 1953. Zooge la of the Sea. E and The Sovereignty of the Sea. An Historical Account of the Claims of England to the Dominion of the British Seas, and of the Evolution of the Territorial Waters: With Special Reference to the Rights of Fishing and the Naval Salute. William Blackwood & Sons. Edinburgh, 1923. The life-cycle of Wegener’s hypothesis. Nature 111(27 Gill, A. C. € J. M. i -fish species the eel in relation. to 785): 359-300. Kemp. 2002. Widespread Indo-Pacific : A challenge for taxonomists, biogeog- raphers, ec lowes and fishery and conservation managers. Biol. 165-174. a 1994a. Repeating patterns in nature, predictabil- impact in science. Pp. 61-84 in L. Grande € O. Rieppel (editors), Interpreting the Hierarchy of Nature, from Systematic Pi Environm. Fishes. 65: Grande, | ily, sade atterns lo Coevolutionary Process a . Academic Press, London. 94b. Studies of paleoenvironments and historical loge pean in the fossil Butte and Laney members of the sreen River formation. Contrib. Geol.. Wyoming 30: 5-32. Univ. Bemis. 1991. Osteology and phylogenetic relationships of fossil and recent paddlefishes (1 "olyodo on- seniores: J. Verte xd = = tidae) with comments on the inte eure of Acipen- Paleontol. m. l) 908. A eder nsive phylogenetic study of amiid "ub TRCERE based on comparative Paleontol., Spec. Mem. 4 skeletal eae J. Verte (suppl. to vol. Volume 95, Number 2 2008 Parenti 245 Life History Patterns and Biogeography & —— — — Historical biogeography and historic al paleoecology of Amiidae and other halecomorph fishes. Pp. 413—424 in G. Arratia & H.-P. (editors), Mesozoic Fishes 2—Systematics and Record. by g Dr. Friedrich Pfeil, Munich. Grehan, J. | 2006. Non-Darwinian evolutionary mecha- nisms. Pp. E 747-1750 in H. J. Birx (editor), I Anthropology, Vol. 4. Schultze Fossil encyclopedia Sage Publications, Thousand i ks, California. Griffith, R. W. 1987. Freshwater or marine origin of the vertebrates? Comp. Biochem. Physiol. A 87: 523-531 Gross, M. R. 1987. The evolution of diatom in fishes. Pp. 14-25 in M. J. Dadswell, R. J. Klauda, C. M. Moffitt, Saunders, R. A. Rulifson & J. E. ee (editors), = Common Strategies of Anadromous and Catadromous “ishes. American Fisheries Society Symposium |, Amer- can Fis Marylan sheries | Society, Bethesda, Wade Jones, 1968. Fish Migration. Pawar Arnold Ltd., London Hardy, C. R. 2006. oe and possible solutions. Divers. Reconstructing ancestral ecologies: Distrib. 12: . & R. D. Mooi. 1994. Areas of ie "eod Send and criteria for recognition. Syst. Biol. 261—260. Heads, M. 1985. On the nature of ancestors. Syst. Zool. 205-215. 1990. Mesozoic terranes and the deconstruction of bioge oet A new model of Australasian biogeography. 223-225. — iogeogr. 17: — . 2005a. Towards a s P robivaehgiaphy of the seas. Biol. J. Linn. men 84: 675 A, . Dating m. on molecular Jp: gre A critique el a biogeography. Cladistics À Helfman, G. S., B. B. Collette & D. E. Facey. 1997. The diversity of fishes. Blackwell Science, Malden, Massachu- setts. Herre, A. W. 1940. Distribution of fresh- p fishes in the Indo-Pac Sci. Monthly [August] 51: 165-168. Hilton, E. J. 2003. Comparative osteology n phylogenetic systematics of fossil and living bony-tongue fishes (Actinopterygii, Teleostei, Osteoglossomorpha). Zool. J. Linn. Soc. 137: 1-100 Humphries, & L. Parenti. 1986. Cladistic Biogeography. Oxford University Press Monographs on Biogeography. Clarendon Press, Oxford. & 99, Clad istic Biogeography, 2nd ed. Pida Da rd. . Tsukamoto & M. Nishida. 2003. A mitogenomic per- h." Molec Oxford Univ. J. G., M. Miv Basal actinopte rygian pd Inoue, spective on the phylogeny of the “ancient fis lo Phylogen. Evol. 26: 110-120. Janvier, P. 2007. euh primitive fishes and fishes from deep p. 1-51 in D. J. McKenzie, A. P. Farrell € (editors), Primitive Fishes. Fish Phy sila: . Academic Press, New York. & atterson. 1996. Re are of lower euteleostean f fishe s. Pp. 251-332 in M. L. J. Stias N ). Johnson (editors), mE of Fishes. do 'ademic Press, New York Keith, P. )3. nt Indo-Pacific and ( 847. o x o. e D. Parenti & G. D iology and ecology of amphidromous i of zaribbean regions. Fish Biol. Larson, H. E. es A revision of the gobiid fish genus Mugilogobius (Teleostei: Gobioidei), and its systematic placement. Rec. West. Aust. Mus. (Suppl. No. 62). Leis, J. M: 1984. Larval K a Fd the East Pacific Barrier. M 'éanogr. Trop. Lin, Y.-S., Y.-P. Poh & C. a (uas on freshwater eels inferred from mitoc TN genes. Molec. Phylogen. Evol. 20: 252-261. . 2001 b. e of the ancient nme tic phylogeny of freshwater eels. J. Taiwan Fisheries one 51-173. . A phylogeny of Which is the bn rsal eels: Reanalys E , C.-S. Tzeng & J.-K. Hwang. 2005. perra nt of ld al characteristics in freshwater eels (genu Anguilla, Anguillidae) shows congruence ih ane ub a estimates. Zool. Scripta 34: 34. win G.. M. Kottelat, G. R. Smith, A L. J. Stiassny & C. Gill. 2000. So many fishes, so little time: An overview of recent ichthyological discovery in continental 87: O. Wilson. 1967. Princeton Univ. Th e Press, = D Q : za zu c E Theory of a _Bioge ography. Princeton, Nev hen D D. 1915. Climate and evolution. Ann. New York Acad. Sci. 24: 171—416. G. Marquet & P. Laboute. 1992. Les Sicydiinae nésie Francaise. Maus, À., (Gobiidae) des eaux douces de la Poly: i de trois espéces nouvelles. Cybium 16: 213- 231. McC arthy, D. 2003. The trans-Pacific zipper effect: Disjunct ink the — ster taxa and matching ge eologic al outlines that Pacifi ic oo J. 545-1561 5. Biogeographical and A ue evidenc 'e E a Biogeogr. 30: 5-15 m—. er, ct tely-enclosed Pacific Basin the ‘retaceous. J. Biogeogr. 32: 216 1 cA. Mc Pen n. M. 1970. T 2 OM of New Zealand. Bull. ie comp Zool. Es occurrence p "dias of diadromy E d ;. Pp. 1-13 in M. J. Dadswell, R. J. Klauda, C. . Moffitt, n. L. Saunders, R. A. Rulifson € J. E. Cooper Common Strategies of Anadromous and Cata- din = ie : dromous Fishes. ], American Fisheries Society, American Fisheries Society Symposium Bethesda, Maryland. 1988. Diadromy in Fishes: iur Between Freshwater. and Marine Environments. Timber Press, Portland. 992. Diadromy: dc and definition of termi- nology. Nal 1992: 248-251. A recent marine ancestry for diadromous iol E paren yes, but mostly no! Environm. Biol. 37: 329-335. 997a. The evolution of diadromy in fishes (revisited) and its place in phylogenetic analysis. Rev. Fish Biol. Fish. 43—46: 199 Tb. a there "e a thing as amphidromy? Micronesie: :a 30: . 200 av ced diversity and divergence: Impli- cations nés speciation processes in fishes. Fish Fisheries 2: 21 ; 20 02. Accumulating evidence for a dispersal biogeography of southern cool temperate freshwater fishes. J. Biogeogr. 29: 207-219. 2003. Hawaiian biogeography and the islands’ eua fish fauna. J. 0. McLennan, D. A. 1994. / Biogeogr. 30: 703-71 phylogenetic approach to the evolution of fish behaviour. Rev. Fish Biol. Fish. 4: 430—460 Meyer J Mino & G. Paulay. 2005. Fine scale endemism on coral re rchipelagic ro m in turbinid gastropods. E dde 59: 113- Annals of the Missouri Botanical Garden Minegishi, Y., J. Aoyama, J. G. Inoue, M. Miya, M. Nishida & K. Tsukamoto. 2005. Molecular evolution of the freshwater eels genus Anguilla based on the whole mitochondrial genome sequences. Evol. 34: 134-146. Mooi, R. D. & A. C. Gill. 2002. Historical biogeography of fishes. Pp. 43-68 in P. J. B. Hart € J. D. Re (editors), Handbook of Fish and Fisheries Biology, Vol. 1 Fish Biology. Blackwell Publishing, Oxford. Morrone, J. J. 2002. Bioge dnd: ‘al regions under track and cladistic scrutiny. J. Biogeogr. 29: 149-15: Myers, G. S. 1938 zoogeography. 92(1937): 339-36 —.. 1949a. relation to zoogeographical problems. Bijdr. Dierkunde 28: 322 3l5- lolec. Phylogen. ynolds Fresh: aa fishes and West Indian Board Regents Smithsonian Inst. Salt tolerance of fresh-water fish groups in 1949b. lied lerms ior migratory fishes. Copeia wd .P Myers, M. J., UV. H. a a Neritid and thiarid pu from French pee streams: How — Use 0 “anadromous, O and a p OST Meyer & reproduction (sexual, parthenogenetic) and dispersal (active, passive) affect population structure. Biol. 44: 535-545. Nelson, G. J. 1969a. Gill arches and the phylogeny of fishes with notes on the classific ation nok vertebrates. Bull. Mus. pee Hist. 141: 475 969b. Infraorbital prd ang their be BUE on the Amer. nd and gragap y of fishes : 1-37. gl 1 Amer. Mus. Novit. . 1973. pons on Fa structure and relationships of Amer. certain Cre E eous and Eocene teleostean fishes. : 1-231. Mus. Novit. l 974. pono biogeography: An alternative 3: 555—550 eme lizanon: Syst Zool. : 1986. f of challenge: The rape Pp. J6 in A. E Aldrich Pun Publication of Earth Sciences History Vol. 4 ———— qe 12 27 (editors), Plate Tectonics and d Ai 2004. Cladistics: Its arrested development. 147 in D. M. W Es & P. L. in Systematics. Systematics Association Special Vol. 67. CRE Pen Boca jar Florida. . Platnick. 1981. phy: € e '" and Vicariance. New York. Forey (editors), Milestones Systematics and Bioge ogra- Columbia Univ. Pres —— —— & P. Y. Ladiges. 1991. Three-area statements: Stan p d p ctm for biogeographic analysis. Syst. Zool. ‘70-485. ——— & — ». Paralogy in cladistic biogeography and analysis of Peat ree subtrees. Amer. Mus. Novit. 3167: 1-58. ————— ——— 2001 pay and = New York school revisited. Austral. J. Bot. - 89—40 Gondw ana, v Le arianc e E d Ne ud J.S. ` 2000. Fishes of the World. Sons, New York Parenti, L. R. freshwater fishes. Austral. Syst. Bot. 4: Ath ed. John Wiley & 13714 . 2006. Common cause m historical ores Pp. 61-82 in M. €. Ebach R. Tangney (edito World. Fifth Biennial United King- Biogeography in C qe Conference of dom. CRC ` the Systematics Association, Press, Boca Raton, Floric A. Maciolek. Palau, 1993. New sic icydiine gobies from a snl Micronesia, with comments on systematies of the ded Sic ) (Teleostei: Gobii- —972 dae). Bull. Mar. Sci. Freshwater uture of 1091. Ocean basins and the biogeography of 1996. Sicyopterus rapa, new species of sicydiine goby ( Teleostei: Gobiidae) from Rapa, French Polynesia. ae Mar. Sci. 5 007, & K. Paris: 1998. ' Pharynge al jaw morphology and rios in sicydiine gobies (Teleostei: Gobiidae) and allies. J. Morphol. 237: 257-274, Patterson, c. 1973. Interrelationships of holosteans. Pp. 233— 305 in P. H. Interrelationships of Fishes. Greenwood, R. S. Miles € C. Patterson Academic Press, distribution of Mesozoic freshwater Mus. Natl. Hist. Nat., Sér. A., Zool _ . The fishes. Mém. 56-174. 993. Osteichthyes: Teleostei. Benton (editor), The Fossil Record London. Ree, R. H.. B. R. Moore, C. O. Webb & M. J. 200 2005. A like Tur frame ne for inferring the evolution of Pp. 621-656 in M. J. 2. Chapman & Hall, Donoghue. geographic range on phylogenetic trees. Evolution 59(11): d l Rosen, Cypr scm a. Familie . 1973. Suborder Cyprinodontoidei. Superfamily s Cyprinodontidae, Poeciliidae, Anablepidae. Fishes of x Western North Atlantic. Mem. ears ponm d. Mar. Res. 1(6): 229-26: 1974. The cub ny and mcm of salmoni- form Ries ind the relationships of Lepidogalaxtas salamandroides. Bull. - Mus. Nat. Hist. 153: 265-326. Rosenblatt, R. H. 'aples. 1986. c o of allopatric populations of shore fish species Amer. genetic from the eastern and central Pac ifie ocean: Dis vicariance? Copeia 1986(2): 275-284. Enghoff & F. 2001. Patterns of animal dispersal, vicariance and diversification in the Biol J. Linn. Soc. 73: 345-390. . D. 2007. On basal clades and ancestral areas. J. Biogeogr 34: 1470-1471. Shields, a P the spersal or Sanmartín, 1.. H. Ronquist. Holarctic. 919. Evidence for initial opening of the Pacific Jurassic. Palaeogeogr. Palaeoclimatol. Palaeoecol. 26: 181-220. 3. Trans-Pacif Er tie links that suggest earth 2 Pp. 199-205 in S. W. Carey (editor), Expanding Marth M tes U niversity of Tasmania, Hobart. 19 Pacific O and rapid earth expansion. J. Biogeogr. 18: Ties Plate ni or an expanding earth? J. 08. Geo . India 47: 399—4 Smith, G. 1989a. Order Anguilliformes: Family Anguilli- dae. Fishes of Atlantic. Mem. Sears the Western North Hes. 9(1): 17 89b. Family Anguillidae: Leptocephali. Fishes of the Te estern i Atlantic. Mem. Sears Found. Mar. Res. 9(2): 898— Sparks, E P Smith. 2005. dispersal ability, and nonevidence: “Condi Life Rafts” — 2547. e uis Freshwater. fishes, to the rescue. Syst. Biol. 54: 158-1€ Springer, V. G. 1982. Bact fic plat special S MN to shorefishes. Smithsonian Contr. Zool. 307: 1 el , with O. Wiley, G. D. Johnson & M. Carvalho. Pp. 410—420 in J. Crac E. & Ae the Tree of Life Stiassny, M. 4 2004.. CR NUM fishes. J. uri Get Oxford Univ. Press, New York Tchernavin, V. i 3o. "The origin of salmon—lIs its ancestry marine or freshwater? Salmon Trout Mag. 95: 120-140. Tesch, F. W. 1977. The Kel— Anguillid Eels. Chapman and Hall, Biology and Management of London. Volume 95, Number 2 2008 Parenti 247 Life History Patterns and Biogeography Thacker, C. E. € phylog of basa Rhyacichthyidae, us Xenisthmidae, Eleotridae eri Per- ciformes: Gobioidei) | MU Phylogen. Evol. 37: 858-871. Tsukamoto, K., F Migration, speciation, and the evolution " Nene in anguillid eels. Canad. J. Fish. Aquat. Sci. 59: 1989-1998. ; i. 1996. New superfamily and three new families of ndo fishes from the Upper Cretaceous: The earliest and Hardman. 2005. | grbiid fishes: J. Aoyama & M. J. Miller. oe Tyler, J. C. & I most morphologically primitive plectognaths. Smithsonian Contrib. 82: 1-59, Waters, J. M A. López & G. P. Wallis. 2000. Molecular phylogenetics and biogeography of galaxiid fishes ( teichthyes: Galaxiidae): Dispersal, vicariance, and the position of Lepidogalaxias salmandroides. Syst. Biol. 49: 777-795 Paleobiol. Molecular Wegener, A. 1929. Die ens a der Kontinente und Ozeane. Frie iedric h Vieweg & Sohn, Braunschweig. Wiens, J. J. & M. J. onda. a. Historical Pu ogra- phy, ecology and species richness. Trends Ecol. Evol. 639- Wiley, E. O. 1976. The phylogeny and biogeography of fossil es recent gars. (Actinopterygii: Denise) n ee Nal. Misc. Publ. 64: 1-111. -P. Schultze. 1984. Family Lepisoste ridae (gars) as living Pd Pp. 160-165 in N. Eldredge & S. M. Stanley (editors). Living Fossils. Springer- -Verlag, New Ya Wolfson, A. 1948. Bird migration and the contine he drift. Science 108: 23-30. 986. Bird migration and the concept of continental drift. = 182 Levition € M. L. Aldrich (e Wh Plate Tectonics p Biogeography. Spec ial Issue Earth Science History, Vol. 4. Univ. concept o -186 in A. E. CONGRUENCE AND CONFLICT Jeanne M. Serb? and M. Christopher Barnhart’ BETWEEN MOLECULAR AND REPRODUCTIVE CHARACTERS WHEN ASSESSING BIOLOGICAL DIVERSITY IN THE WESTERN FANSHELL CYPROGENIA ABERTI (BIVALVIA, UNIONIDAE)' ABSTRACT Organisms with complex life histories and unusual modes of genome inheritance can present challenges for phylogenetic reconstruction and accurate assessment of biological diversity. This is B wticularly true for freshwater bivalves in the family Unionidae he cause: (1) they have complex life cycles that i include a parasitic larva and obligate fish host; (2) they possess both found in riverine habitats with complex hydrogeological histories. E xamination of mitochondrial DNA (mtDNA) ) sequen nees, lutinate morphology, and host fish compatibility of the western fanshell Cyprogenia aberti (Conrad, 1850) revealed significant character variation across ils range. Although variation was correlated among the different data sets and supports discrete groups, these groups did not always correspond to geographically isolated populations. Two discrete mtDNA clades exist sympatrically within most C. aberti populat tions, and these same clades are also diagnosed by at least one morphological character, egg color, The surprisingly high genetic distance (14.6196—20.1966) between the members of these sympatric clades suggests heritance infidelity of the two different mitochondrial genomes. This hypothesis was tested and falsified. More general patterns in geography were observed in host fish compatibility. Populations of C. aberti from the major river systems differed in their ability to utilize fish species as hosts. These differences in reproductive traits, which are presumably genetically based, suggest that these. populations are not ecologically exchangeable with one another and represent biological diversity not previously recognized within Cyprogenia Agassiz, 1852. Key words: Central Highlands. coevolution, Cyprogenia aberti, double uniparental inheritance, mitotype. Unionidae. =, Comparison among associated historical patterns is association can provide insight into the biology of a powerful tool in evolutionary biology. Such patterns understudied organisms, direct research by focusing include organisms and geographic areas in vicariance attention on meaningful patterns, reveal hidden biogeography (Rosen, 1978). the co-speciation of biological diversity, and help identify the processes hosts and parasites (Hafner € Nadler, 1988), and and mechanisms involved in speciation. congruence between gene trees and species trees Although evolutionary associations may provide (Goodman et al., 1979). Areas. organisms. and genes mechanisms for speciation, lineages (organism vs. are analogous because each has a history that can be gene vs. area vs. host) may or may not show congruent reconstructed via phylogenetic methods. Likewise, the patterns. Pattern differences may reflect a unique affiliations between these patterns are analogous. for history at one level that did not influence lineage each pair has one entity that is connected to, and may diversification at another. An example of such track, the other because of causal interaction through incongruence would be the failure of a parasite to evolutionary history (Page, 1993: Page & Charleston, track its host after a host-switching event (Page. 1998). Comparing lineage patterns in an evolutionary 1994). However. when data from different evolution- 'We would like to thank Richard Mavden, Mick Richardson, and Peter Raven for organizing the 52nd Annual Systemalics Symposium of the Missouri Botanical Garden. “Reconstructing Complex Evolutionary Histories: Gene-species Trees, Historical Biogeography, and Coevolution.” We ds thank John Harris, Betty C E Chris ig tok Nathan Eckert, Brian e meyer, Susan Oetker, Bill Posey, and Kevin Roe for assistance with this project. Funding was generously provided by the S. Fish and Wildlife Service, Kansas De partment of Wildlife and Parks. Quac ita National Forest Service, the American E of Natural History’s Theodore Roosevelt Memorial Fund (JMS), and the University of California President's Postdoctoral Fellowship (JMS) “Department of Ecology, Evolution, and Organismal Biology, lowa State University, Ames, lowa 50011, U.S.A. serb@ iastate.edu. t Department of Biology, Missouri State University. Springfield. Missouri 65897, U.S.A. chrisbarnhartOmissouristate.edu. doi: 10.3417/2006103 ANN. Missourt Bor. Garp. 95: 248-261. PupLISHED on 18 June 2008. Volume 95, Number 2 2008 Serb & Barnhart Biological Diversity in Cyprogenia aberti ary associations converge on a single pattern, it is a strong indication that the pattern represents the true species tree (Page & Charleston, 1998 North American freshwater mussels in the family Unionidae offer an unusual opportunity to examine simultaneously, because the multiple associations greatest diversity occurs in areas with complex hydrogeological histories, their life cycle includes an obligate parasitic stage, and they possess an unusual first unionid mode of mitochondrial inheritance. The two factors appear to play a role in 1984) assessing biological diversity (Roe & Lydeard, King et al., 1999; 2001). In mitochondrial genes are often used in phylogenetic major when 1998; contrast. speciation (Kat, and may be useful Roe et al., reconstruction because most models of mitochondrial DNA (m and purifying selection as the major processes t tDNA) evolution consider neutral evolution — lat shape sequence divergence. Thus, mtDNA variation should track species history. However, alterations in mitochondrial inheritance. could affect phylogenetic inferences from these genes, making them incongruent with other lines of evidence (Hoeh et al., 1996, 1997) The life eycle of North American e includes a larval stage that is obligately parasitic on a vertebrate host (for a general description of unionid life history, 1921; McMahon, 1991). The parasitic larvae, called glochidia, are brooded inside the female see Coker et al., mussel in a modified portion of the gill. Once mature, the glochidia must attach. to a host to complete development. Generally, the host is a fish, and the elochidia encyst in the gill or fin epithelium of the appropriate host species. Unionids are typically able to utilize only a limited number of species as hosts. Glochidia that attach to a compatible host are encysted by migrating epithelial cells and subsequently trans- 1932; Glochidia that attach to form into the juvenile stage (Arey, Rogers- Lowery € Dimock, 2006). incompatible (non-host) species either fail to be encysted or are subsequently sloughed before transfor- mation is complete (Arey, 1932; Meyers et al., 1980; O'Connell & Neves, 1999: Rogers & Dimock, 2003). Relationships between mussels and host species may influence speciation rates and/or provide infor- r mation for phylogenetic reconstruction. The degree of host specificity varies among unionid taxa (Hoggarth, 1992). Some host-parasite relationships are highly restrictive, so that a mussel species will only transform on a specific (sympatric) population within 2001: Eckert, associations a host species range (Rogers et al., 2003). In between mussel parasite and fish host should be these instances, lineage highly correlated. Females of some host-specific mussel species (e.g., 1820), Lampsilis cardium (Rafinesque, Villosa iris (Lea, 1829)) attract particular hosts by producing lures that mimic prey items of intended host (see Kat, 1984; Lures can be part of the female’s mantle tissue or may Barnhart et al., 2008 and references within). take the form of conglutinates, aggregations of larvae (glochidia) and unfertilized eggs that are released from the female. When a fish attacks the conglutinate, it comes in direct contact with the glochidia, which At the time, the lure limits the diversity of fish that are likely increases the probability of infection. same to come into contact with the glochidia of the species and, therefore, affects the range of potential hosts. The evolution of lures may thus affect the evolution of host For example, change in display timing Kat (1984) and Graf (1997) suggest that this change in fish specificity. may alter the level of attractiveness to a host. hosts might result in the formation of new unionid species via sympatric means. Another evolutionary association in unionids occurs between the organism and its geographic range (area). Similar to freshwater fishes and crayfishes, unionids exhibit distinct faunal assemblages and high endem- 1987; unionid diversity is ism among river basins (Neves & Widla 1997). The found in ancient rivers with dynamic hydrogeological 1995; Neves et al., 1997). Many of these rivers have undergone reorga- Vaughn, greatest histories (Lydeard & Mayden, nization during the cyclic glacial advances and retreats of the Pliocene and Pleistocene (e.g.. Donn et al., 1962; Melhorn & Kempton, 1991) or experi- enced episodic tilt-block tectonics (e.g., Cox, 1994) during the Pleistocene and Early Holocene. These geologic events have profound evolutionary effects by restricting gene flow, causing allopatric speciation through vicariance, or creating opportunities for rapid radiations and secondary contact zones between once- isolated populations through stream-capture events (Roe et al, 2001; Kozak et al., 2000). understanding the historical geology of the drainages Thus, in which the study organisms occur is important when interpreting phylogenetic reconstruction. Phylogenetic patterns should reflect the hydrogeological history of the drainage basins when the organism is restricted to and has limited dispersal aqualic environments capabilities. There will be some exceptions to this prediction. In a dynamic environment with multiple, rapid cladogenetic events, area relationships and gene lineages may be in conflict with one another due to incomplete lineage sorting of ancestral polymor- phisms. Incongruence between the unionid and area history also may occur when mussel species use hosts with high vagility. Finally, complicate phylogenetic analysis and interpretation unusual modes of genomic inheritance can where the gene tree may not reflect the species tree 250 Annals of the Missouri Botanical Garden O A 11B1p1a, N 9€ Ps Y KS; Y Y O Clade A O Clade B ae C iih Sp g Figure 1. clades recovered in the phylogenetic analyses, clade A and e Strawberry (5), Current (6), Black (7), White (8), Quac abbreviations: Arkansas (AR), Kansas S Missouri (MO) The Unionidae and two other bivalve families (Mytilidae and two different mitochondrial genomes (mitotypes) that are associated with gender (F for female and M for ma Veneridae) possess zaan e) (Zouros et al.. 1992, 1994; Hoeh et al., 1996, 1997; Liu et al., 1996; Passamonti & Scali, 2001). These sexual mitotype lineages remain separate via doubly uniparental inheritance (DUI) (Skibinski et al.. 1994; Zouros el al., 1994), where the maternal mitochondrial genome (F-mitotype) is inherited by both male and female progeny from the mother, while the paternal mitochon- drial genome (M-mitotype) is only transmitted from father to male offspring and housed in the male gonads. The resulting heteroplasmy is not necessarily limited to males. In both — aboratory crosses and natural popula- tions, low levels of DUI infidelity have been docu- mented in Mytilus L. species (family Mytilidae) (Zouros et al., 1994; Stewart et al.. 1998). This event is a role reversal of sexual mitotypes M') feminization of the M-mitotype (F") (as describe Hoeh et al., 1997). functional mitotype (e.g.. 1995: Garrido-Ramos et al.. by either masculinization of the F-mitotype by The result is that two lineages of the " and E ") co-exist in a population. These events can confound the interpreta- tion of phylogenetic analyses when non-orthologous genes (F vs. F") are inadvertently compared (see Hoeh et al., 1996, 1997), so that the resulting gene tree would nol represent the species tree. ade B. respectively site on the Verdigris River that was used in a host ide p e study by phylogenetic POEM (but see phylogenetic analysis in Serb, 2006 and Serb (2006). Collection sites are os by river name and number: Fall ( hita (9), ). and Oklahoma (OK). Distribution map of sampled P nia aberti localities. Squares and circles represent the two major genetic de A he (see Fig. 3). A Eckert (2003). but was not alude in this ). Specific Ps z information i is e d in Appendix | . Verdigris (2). St. Francis (3). Buffalo c jd Caddo ur. States are star represents 1 ‘ollection Saline a abe led y ith CYPROGENIA ABERTI, A CASE STUDY The western fanshell Cyprogenia aberti (Conrad, 1850) is the only congener of the federally endangered 1820) (U.S. Fish & Wildlife Service, 1990). These two species occur in the Central Highlands of North the Interior Highlands west of the Mississippi River and C. Mississippi River. C. stegaria (Rafinesque, America, with C. aberti found in stegaria in the Eastern Highlands east of the The Central Highlands are drained by high-gradient, clear streams and are isolated by low-gradient habitats that are typically silt’ loaded. The two Highland provinces contain a diverse, well- studied aquatic taxa, Mayden, 1994), and fauna and many endemic (Wiley & Mayden, 1985; 1987, 1988), amphibians (Routman et al.. crayfishes (Crandall & Templeton, 1999). Cyprogenia aberti is endemic to rivers draining the Ozark Plateau and the Ouachita Highlands of the Interior Highlands. These include upland portions of the Arkansas, Ouachita, White, and St. including fishes Francis river systems in Arkansas, Kansas, Oklahoma, and Missouri (Fig. 1). Many C. geography belween rivers and within river systems. aberti populations are isolated by For example, populations in streams of the upper Arkansas, Black, and White rivers are all separated the lower-elevation from one another by . low-gradient portions of these systems. Volume 95, Number 2 2008 Serb & Barnhart Biological Diversity in Cyprogenia aberti 251 Figure 2. differs among Cyprogenia aberti females. —A— Variation in conglutinate morphology. Color of the unfertilized eggs and the structure of conglutinate ure vite conglutinates were produced by a female collected in the Verdigris River, Kansas. Similar conglutinate morphology is produced by Fall River females. —C—D. Red and brown conglutinates were 8 a | à g produced by two sympatric females collected in the St. Francis River, Missouri. Geography would appear to be the best predictor of how populations of Cyprogenia aberti would be related em to one another due to their disjunct distribution. However, recent genetic investigations of C. aberti using mtDNA sequence data indicate that geography only partly explains the phylogenetic. pattern. (Serb. 2006). Although the phylogenetie clades correspond to drainage, this pattern was replicated within C. aberti, where two genetically distinct sympatric lineages were identified within the Ouachita River (Serb. 2006). If the mitochondrial gene tree accurately describes the system as non-sister clades species tree, these data suggest that there are as many as five distinct evolutionary lineages within C. aberti (Serb, 2006). To test the of this biodiversity compared molecular, accuracy assessment, we geographic, and reproductive characters. Reproductive traits employed in attracting a fish host, such as lure morphology, presumably have a genetic basis and may provide additional data for determining the species tree. In Cyprogenia, conglu- tinate coloration varies among species and popula- tions. For example, C. stegaria typically produces red & Neves, 2002). \mong populations of C. aberti, red conglutinates worm-like conglutinates (Jones have been documented from the St. Francis River (Chamberlain, 1934), while white conglutinates were observed in the upper Arkansas River (Barnhart, 1997) (Fig. 2A, B). Most recently, Eckert (2003) identified populations in the Ouachita and St. Francis rivers that contain females that produce either red or brown conglutinates (Fig. 2C, D). Conglutinate color is the result of pigmentation in the sterile (structural eggs (Barnhart, 1997; Eckert, 2003). The red and brown forms are pale when immature but darken during development (Barnhart, pers. obs.). mr 1e goal of this study is to better understand the types of data needed to accurately assess biological diversity in organisms with complex evolutionary histories. We explore how three evolutionary associ- 252 Annals of the Missouri Botanical Garden ations of a North American unionid, Cyprogenia aberti, relate to one another: reproductive traits associated with host preference, geographic distribu- tion, and mitochondrial gene variation. Specifically, we compare the geographic distribution and repro- ductive traits with a mitochondrial gene tree derived from previous work (Serb, 2006). MATERIALS AND METHODS Thirty-one individual specimens of Cyprogenia aberti representing 11. different localities were collected across its extant range in Arkansas, Kansas. and Missouri. Individuals are identified with a unique museum number and river location. The following river and geographic notations are used. “White” refers to the upper White River above the confluence of the Black River and its tributaries (Strawberry, Current, and Buffalo rivers). The upper St. Francis River is included in the White designation, because the two rivers have similar fish faunas attributed to interchanges between headwaters (Cross et al., 1986). “Arkansas” refers to the upper Arkansas River and its tributaries in Kansas (including the Fall and Verdigris rivers). “Ouachita” refers to the Ouachita. Caddo. and Saline rivers in the Ouachita Highlands. Rivers of the ‘the Ozark Plateau and Ouachita Highlands. Figure 1 is a map of Interior Highlands include all drainages o rivers and localities in this study. DNA SEQUENCING Genomic DNA was extracted from the Cyprogenia aberti specimens and six outgroup taxa also within the Unionidae: C. stegaria (eastern fanshell), Dromus dromas l. Lea (dromedary pearly mussel), Lampsilis ornata Conrad (southern pocketbook), L. siliquoidea Barnes (fatmucket), Obliquaria reflexa Rafinesque (three-horn wartyback), Ptychobranchus fasciolaris Rafinesque (kidneyshell) (Appendix 1). Mantle tissue along the shell was used in standard phenol/ chloroform extraction with ethanol precipitation (Roe & Lydeard, 1998) to reduce the probability of non- orthologous mitotype contamination. from the male gonad. DNA amplification and sequencing of two mito- chondrial gene portions were performed as described in Serb (2006). A 600-bp region of the first subunit of cytochrome oxidase e gene (COI) was amplified and sequenced using primers 1001490 and HCO2198 (Folmer et al., 1994) or a modified HCO2198 primer (Graf & Ó Foighil, 2000). A 700-bp region of the 5' end of the first subunit of the NADH dehydrogenase gene (NDI) was amplified and sequenced using primers Leu-uurF and NIJ-12073 (Serb et al., 2003). Purified polymerase chain reaction (PCR) products were used as a template for cycle sequencing reactions with the ABI. Prism Big Dye Terminator kit (Applied. Biosystems, Foster City, California). Sequencing reactions were visualized on an ABI 3100 automated sequencer. A male mitotype from Cyprogenia aberti was isolated and sequenced for phylogenetic analysis. Testis tissue was dissected from an ethanol-preserved C. aberti male from the Black River, Missouri (see Appendix 1). Total RNA was isolated from gonad tissue using TRIZOL (Invitrogen, Carlsbad, Califor- nia). followed by phenol/chloroform extraction and ethanol precipitation. RNA was used as a template for complementary DNA synthesis in reverse transcrip- tase-PCR with a poly-T primer. Complementary DNA was quantified and used in subsequent PCR to amplify the male mitotype copy of COL. A 550-bp region of the male copy of COL was amplified and sequenced using primers UNIOMCOIF (Curole, 2005) and HCO2198, by means of the following thermal cycling conditions for 34 cycles: denaturing at 94°C for 40 seconds, annealing at 45 C for 60 seconds, and extension at 72 € for 90 seconds. CONGLUTINATE PHENOTYPES Six to eight gravid females of Cyprogenia aberti were collected from the Fall (Arkansas system), St. Francis, and Ouachita rivers. These females were brought back to the laboratory, where conglutinate release was induced by raising water temperature from 5 C to 20 C over a period of seven hours. Maturity of glochidia and conglutinate color were determined, which allowed for individuals to be designated as known phenotypes. All animals were preserved. in 9596 ethanol. DNA sequencing was performed as — described above for correlation with conglutinate phenotype. PHYLOGENETIC ANALYSES All double-strand DNA sequences. were visually aligned unambiguously and were converted into amino acid sequence to check the accuracy of the nucleotide sequence in the data editor BioEdit (Hall. 1999). For each protein-coding gene, ar evolutionary model was determined using the log likelihood ratio test (LRT) as implemented in Modeltest version 3.0 (Posada & Crandall, 1998). Pairwise sequence divergences were calculated between all taxa using the appropriate evolutionary model. Mean sequence divergence was calculated within between and phylogenetically defined clades. Volume 95, Number 2 2008 Serb & Barnhart Biological Diversity in Cyprogenia aberti 253 Aligned sequences were analyzed phylogenetically under two optimality criteria: maximum parsimony (MP) using PAUP* version 4.0b10 (Swofford, 2002) and Bayesian using the program MrBayes version 3.0 (Huelsenbeck & Ronquist, 2001). All MP analyses emploved 100 random-order a heuristic search of addition of taxa and tree bisection-reconnection TBR) branch swapping. Only minimum-length trees were retained and zero-length branches were col- : the resulting phylogenetic hypotheses was assessed by lapsed. Support for the individual nodes of bootstrap values using the FAST step-wise addition option (2000 replicates) in PAUP*, and decay index TreeRot Sorenson, 1999). Outgroup taxa within the values (Bremer. 1994) were calculated with version 2c Unionidae were chosen based on a molecular phylogeny of the tribe Lampsilini using COI] and 16S ribosomal RNA (rRN 1999). Trees were rooted using the species Lampsilis A) sequence data (Roe, ornata. MP analyses comparing M- and F-mitotypes were based on the COI gene alignment. Trees were rooted using the M-mitotype copy of CO1 from L. teres Rafinesque (GenBank accession number AF406794). A Bayesian analysis of the concatenated mitochon- drial genes was conducted using six data partitions: first, third positions Iwo protein-coding genes. Four Markov chain Monte Carlo second, and codon for (MCMC) simulations were run simultaneously using a 3 x 10% estimate the topology and posterior probability for 100 discarded random starting tree for generations to node support. Trees were sampled every generations, and the number of trees (burn-in) was determined by plotting the log likeli- hood scores of all saved trees versus generation time. Only trees with likelihood scores after stationarity was achieved were retained for inclusion in the consensus tree. Node support was determined by the frequency at which a particular clade occurred within all trees retained after the burn-in trees were discardec RESULTS Conglutinates from Fall River (l, cf. Fig. 1) Cyprogenia aberti females were all white in color, similar to previous observations of specimens from the upper Arkansas system (Fig. 2A, B; Barnhart, 1997; Eckert, 2003). Females from the St. Francis (3, Fig. 1) and Ouachita (9, Fig. brown conglutinates (Fig. 2C, D). 1) rivers produced either red or Sequence alignment of the COI and NDI gene portions yielded 1126 bp, containing 402 polymor- 261 of informative under MP. phic sites, which were phylogenetically The MP analyses conducted on these concatenated sequences resulted in 7476 equally most parsimonious trees (MPT) of 734 steps (Fig. 3). None of these trees supported a monophyletic Cyprogenia aberti due to the placement of two American Eastern Highland species, C. stegaria and Dromus dromas, within the C. aberti clades. In the Bayesian after 650,000 generations, and 8500 trees were discarded analysis, stationarity was achieved as burn-in. The 50% majority-rule consensus tree was the 21,437 (ln L = 4895.566) and was similar to the MP topology (Fig. 4). In both phylogenetic analyses, Cyprogenia form two major clades (A and B). Clades within the large A and B clades can be defined generally by river system based on remaining trees membership of the C. aberti individuals (designated as Ouachita, White, or Arkansas) creating parallel geographic structure. Both clade A and clade B contain Ouachita (Ouachita [9], Saline [10], Caddo [11]. sites, cf. Fig. 1) and White (St. Francis [3], Buffalo [4], Strawberry [5], Current [6], Black [7], White [8]. Fig. 1) clades. Sympatric individuals from these two drainages segregate by sites, cf. om conglutinate color: clade A producers belong to clade B (Fig. red conglutinate producers have membership, while brown conglutinate 2). Clade A also lite conglutinate producers from the Fall River site (1, cf. Fig. 1) and the Verdigris River (see Serb, 2006) in the Arkansas Basin. Only Bayesian topologies. includes w MP and In the MP topology, the red one node differs between the conglutinate—producing Ouachita clade (Ouachita, Caddo, Saline exemplars) is the sister group to the White/Cyprogenia stegaria + Arkansas clade (2 decay index; less than 5096 bootstrap support), while the Bayesian topology places the red conglutinate— producing Ouachita clade as the sister group to the White/C. stegaria clade (83 posterior probability). This topology places all red conglutinate producers together (Ouachita, White, and C. stegaria sites) and the white conglutinate producers in the Arkansas » (Fall River site, Fig. 4). Sequence divergence was calculated for each river > — River as the basal group drainage and phylogenetically defined groups within a single river (Table 1). Because COI and NDI gene portions were determined by Modeltest to have two different sequence evolution models, we calculated sequence divergence for each gene separately. Evolutionary models for COI and ND1 were variations (GTR) model: the Tamura-Nei model (TrN-I) with an adjustment for among-site rate heterogeneity (x = 0.2428) for COL, and the transversion model (TVM-I) with a parameter 0.5211) for sequence divergence within a of the general time reversible for the proportion of invariant sites (I = D ;enerally. phylogenetically defined clade was low. less than 1%. These groups corresponded to conglutinate color (red, brown, or white) within Cyprogenia. In contrast, 254 Annals of the Missouri Botanical Garden 60/1 C. stegaria 1499 Black 2724 C. stegaria 1500 E Black 1648 51/1 mei St. Fran R6 E. t. Fran R8 white Clade A — T white | gem Fall 1647 Arkansas Saline 76 Ouac 2378 I———— Cadd 1838 Cadd 2384 Cadd 2383 Ouac R6 red Ouac R7 62/1 Current 1537 PAL Buffalo 1650 Straw 1652 Straw 2726 99/9 | 62/1 brown St Fran B1 H rown St. Fran B5 St. Fran B2 Clade B DIOWI. Se Fian Ba 100/24 St. Fran B4 White1654 Ouac 2377 Cadd 1837 Cadd 2381 51/2 a Ouac 2374 O 98/13 64/1 L2% Guac B1 a Ouac B3 51⁄1 White/ C. stegaria 92/1 Ouachita red 99/18 White 68/2 Otiachita Ptychobranchus fasciolaris Obliquaria reflexa Lampsilis siliquoidea Lampsilis ornata Figure 3. Strict consensus of 7476 trees recovered in maximum-parsimony analysis of the mtDNA sequence data set (tree length [TL] = 734; consistency index [CI] = 0.683; retention index [RI] = 0.895). Numbers above branches represent bootstrap support greater than 50% (2000 replicate Ph Meus E decay index values. The two major Cyprogenia clades are labeled as A and B. ener clades are labeled by river drainage along terminal branches. Individuals with known conglutinates color are labeled (red. white. or brown) on termin: e branches. Species currently recognized as C. aberti are listed by University of Alabama fiue Collection museum number and river locality. River abbreviations are: Cadd = Caddo: St. Fran = St. Francis; Straw = Strawberry: Oua Juachita. Other species are listed by species name in italics: C. stegaria specimens include museum accession bs ars (see Appendix 1). when means were calculated for drainages that COI = 14.6%: NDI = 17.8%). These values were the Ouachita riverine system included mitochondrially distinct, red and brown slightly higher i conglutinate producers, these values increased dra- (averages, CO] = 15.3%: NDI = 20.2%). Again. matically. Within a single location in the St. Francis these values are estimated from phylogenetically River that contained red/brown conglutinate variation, — distinct genotypes (in A and B clades) found among individuals were more than 14% divergent (averages, sympatric individuals in two different river drainages 7 o $ | o Volume 95, Number 2 2008 Serb & Barnhart 255 Biological Diversity in Cyprogenia aberti 100 RED 100 red c stegaria 1499 Black 2724 = fea" © stegaria 1500 Black 1648 pee St. Fran R6 L__fed st Fran R8 Saline 76 53 Ouac 2378 p LES Cadd 2383 100 Cadd 1838 Cadd 2384 97 red Ouac R6 Toe a Ouac R5 fed uac R7 100 NE - Fall 1647 LAS Fall 1646 BROWN 98 Current 1537 ~ me Buffalo 1650 Straw 1652 100 Straw 2726 ——À4 . 100 100 Drown St. Fran B1 — rom St. Fran B5 brown St. Fran B2 D St. Fran B3 brown St. Fran B4 White 1654 Ouac 2377 Cadd 1837 T 100 Hn 2381 uac 2374 Ouac 2375 100 22% ouac Bt — rom Ouac B3 prown Ouac B2 red Dromus dromas Ptychobranchus fasciolaris Es 89 Obliquaria reflexa m Lampsilis omata LL —— Lampsilis siliquoidea Figure 4. Bayesian 50% veu -rule consensus topology of the mtDNA sequence data set. Posterior probabilities based on 21,437 post burn-in trees (-In L = 4895.566). Bold branches identify taxon rel: Modus dm differ from the MP analysis (Fig. 3). The clade containing ws als with red conglutinates is shaded in dark grey (node with circle). The brown conglutinate-producing clade is in light grey (node with square). Cyprogenia aberti from the Fall River (Arkansas drainage) produce white conglutinate lures. Cyprogenia stegaria and Dromus dromas are both red conglutinate producers (see references in text). Taxon labels follow Figure 3. Annals of the Missouri Botanical Garden Mean Arkansas). within conglutinate color lineages Table 1. sequence divergence within drainage lineages (e.g.. (Os red or brown), or between conglutinate color lineages (c.g. red vs. brown) as identified through phylogenetic 3). Specific determined for (TrN+T) analysis (see Fig. models of sequence evolution were each mitochondrial gene portion: Tamura-Nei corrected. distances. for the first subunit of cytochrome oxidase e gene (COT): transversion (TVM+I) corrected distances for the first subunit of the NADH dehydrogenase gene (NDI COI NDI TIN+T TVM+I White/C. stegaria 0.47% 0.54% Arkansas 0.08% 0.82% St. Francis Re 0.19% 0.35% Brown 0.66% 0.21% Red vs. brown 14.61% 17.76% Ouachita Red 0.61% 0.47% Brown 0.15% 0.09% Red vs. brown 15.33% 20.19% (Ouachita and St. Francis, within the White system) and not monophyletic entities. the M- and F-mitotype copies of the COL gene yielded 529 bp, containing Sequence alignment. of 253 polymorphic sites, 167 of which were phyloge- netically MP. The MPTs of 456 steps. None o MP topologies placed the Cyprogenia aberti M- mitotype within the F-mitotype Cyprogenia clade (100% bootstrap Fig. 5). the dissimilarity between the amino acid sequence of the informative. under MP analyses resulted in four equally the peo support; In addition, M- and F-mitotypes suggests that they are non- orthologous gene copies (data not shown; available from first author). DISCUSSION CONGRUENCE BETWEEN MITOCHONDRIAL PHYLOGENY AND REPRODUCTIVE PRATT: This study attempted to test whether a mitochon- drial gene tree with a replicated geographic pattern is an accurate assessment of the species tree in an organism with a complex life history. We compared the mitochondrial gene topology to life history trait patterns in. Cyprogenia aberti. First, we examined reproductive trait potentially key to the host-parasite association, conglutinate color. Variation in congluti- nate lure color corresponded to the different mito- chondrial lineages of C. aberti. All Cyprogenia ps e worm-like conglutinates that are either white, red, or brown (Fig. 2). Each color morph appears to oie clades defined by the mitochondrial analyses, thus supporting the mitochondrial gene phylogeny with an independent data set. Clade A has mostly red conglutinate producers, with the exception of white conglutinate producers of the upper Arkansas system (Fall River site). clade B consists of brown-producing females, with the excep- In contrast to clade tion of Dromus dromas. Conglutinate lures produced by the closely allied D. dromas are morphologically On they i and flatter than Cyprogenia conglutinates and, unlike C. of clade B. in color (Ortmann, 1912: 2004). Other reproductive characters may be important in distinct. average, are shorter, wider, aberti ae are rec when mature Jones & Neves, estimating biological diversity and understanding One aberti speciation mechanisms of Cyprogenia aberti. Although all C. populations use small, benthic-dwelling trait is host fish preference. species in the darter genera Etheostoma Rafinesque and Percina Halderman, i (Barnhart, host use varies among river basins 1997: Eckert, 2003). Eckert (2003) found variation in host preference among four C. (in 1) rivers of the aberti Qs Spring Kansas) and Verdigris (2, Fig. Ouachita River (9, Fig. 1). and St. Francis River (3, Fig. 1). All C. aberti populations examined had high transforma- Arkansas system, tion success (£e of glochidia that complete develop- ment) on P. caprodes Rafinesque, a large darter species that is widely distributed, locally D 992). In addition, each C. aberti population used at least one other host of more limited. distribution. but not. always abundant, across range of C. aberti — Robison & Buchanan. For example. P. phoxocephala Nelson, a host for Verdigris River C. aberti, co-occurs with C. aberti only in the upper Arkansas drainage (Page, 1983). mussel populations and host species showed poor Significantly, allopatric = pairings of transformation success. However, differences in host compatibility between red and brown conglutinate— producing individuals were not evident (Barnhart, 1997; Eckert, 2003; Barnhart & Eckert, unpubl.). Difference in host compatibility among river drainages provides additional evidence for the distinction among genetically defined allopatric lineages of C. aberti and not the sympatric lineages in the Ouachita and White rivers. Change in host utilization may be an important factor in diversification of C. aberti lineages (see below). Second, we looked for direct evidence of gene tree/ species tree conflict by comparing two different mitochondrial lineages in Cyprogenia aberti. A copy of COL from the C. aberti M-mitotype was included in phylogenetic analyses to test the hypothesis that the Volume 95, Number 2 2008 Serb £ Barnhart 257 Biological Diversity in Cyprogenia aberti 90 Vi 99 C. stegaria 1499 ————— C. sfegaria1500 Black 1648 Black 2724 St. Fran R6 LL St. Fran R8 Fall 1647 Fall 1646 Saline 76 L—— —— Ouac 2378 Cadd 1838 A Cadd 2384 — Cadd 2383 Ouac R6 Ouac R7 Current 1537 Buffalo 1650 Straw 1652 Straw 2726 a A a St. Fran B1 St. Fran B5 St. Fran B2 St. Fran B3 St. Fran B4 White 1654 Ouac 2377 Cadd 1837 Cadd 2381 Ouac 2374 100 Ouac 2375 61 Ouac B1 Ouac B3 Ouac B2 Dromus dromas 53 65 Lampsilis ornata Lampsilis siliquoidea | Ptychobranchus fasciolaris Obliquaria reflexa Figure 5. 456; CI — 0.724; RI — 71). Numbers above MP stric E consensus topology comparing the male (M-) MPT = — > branches re present bootstrap support greater than 50% (2000 Em All sof COI (M and female (F-) mitotype copies individuals are F- Es s ien explicitly labeled as M-mitotypes. M-mitotypes are shaded in gray and have a bold *M' Cyprogenia aberti Black M and Lampsilis teres M. Taxon labels follow Figure 3 genetically disparate but sympatric individuals are the result of a role reversal of sexual mitotypes. In the MP analysis, the C. aberti M-mitotype was never included in the Cyprogenia clade (Fig. 5), suggesting that none of the C. aberti copies of the COI gene were male in origin. Since the C. aberti M-mitotype was not similar to any female-derived gene copies, it is most likely that the genetie variation and replicated geographic pattern found in C. aberti is not due to a role-reversal event. Instead, the observed genetic diversity found in C. aberti mitochondrial sequences arose within the Cyprogenia F-mitotype lineage. These data support recent evidence that DUI has been operating at a high level of fidelity in Unionoidea for over 100 million years (Curole € Kocher, 2002; Hoeh et al., 2002; Walker et al., 2006). Annals of the Missouri Botanical Garden EXPLANATIONS FOR DIVERSITY PATTERNS Examination of the mitochondrial sequence data in conjunction with observed phenotypic diversity within populations of Cyprogenia aberti suggesls existence of multiple evolutionary entities (a species complex). In this section; we present several possible processes that may have produced this observed diversity. One mechanism is that the two lineages (A and B) formed in allopatry, and the river systems where they co- occur represent a secondary contact zone (clades A 3, 4). vicariance events and subsequent range expansions and B in Figs. This hypothesis depends on both that may have been driven by abiotie changes during the Pliocene or Pleistocene. As glacial advances during these epochs altered the ecology and geology of 1987 these changes impacted distribution patterns of the ee the Central Highlands region (see Mayden., biota. Fluctuations in climate, water levels. and reorganization of drainage patterns within the High- = lands could isolate and displace populations of aquatic fauna, while creating opportunity for re- colonization and range expansion at a later time. The post-glacial secondary contact hypothesis has been used to explain similar genetic patterns in other molluscan taxa, including slugs (Pinceel et al., 2005) and terrestrial gastropods (Davidson, 2000). Sympatric speciation is another process that can generale genetically diverse sympatric lineages. It has been demonstrated that shifts in reproductive timing can result in sympatric speciation (allochrony) in invertebrate laxa, specifically those that have complex mating behaviors or periodicity, or those that utilize hosts (Wood & Guttman, 1982; Feder et al., 1993: Harrison & Bogdanowicz, 1995; Simon et al.. 2000). Cyprogenia aberti possess a complex life cycle that provides many opportunities for allochronic speciation, where a shift in the timing of early reproductive events (i.e., gamete release, brooding, or larval release) could affect the success of their obligate parasitic larva. For example, larvae from animals that reproduce later in the reproductive season may be exposed to a different suite of potential fish hosts (Graf, 1997). If allochrony is a speciation mechanism in C. aberti, it does not appear to be a recent event, as evident from the non-sister clade pattern of the Ouachita groups. Instead, allochrony may have occurred within the Ouachita system, followed by range expansion of the two new Additional life history, geological, and perhaps nuclear gene sequenc- clock data determine which speciation mechanism created the lineages into other Highland drainages. es and molecular will be needed to diversity in C. aberti. n the Regardless of how the sympatric lineages Ouachita and White river systems were formed, the question remains of how the color polymorphism could be maintained evolutionarily. Either the maintenance of color polymorphism within one species or the coexistence of sympatric e could result from frequency-dependent selection. In this model. host fish might avoid conglutinates after an initial encounter because of discomfort associated with glochidia attachment. Mussels producing a rare might, therefore, experience ¿ conglutinate color selective advantage because fish would be less likely that selective advantage of a given color would increase to have previous experience with color. The with decreasing frequency in the population, protect- ing the polymorphism. It is necessary to establish several phenomena in order for this theory to gain credence, including the aversion of fishes to similar- colored conglutinates after an initial exposure. This theory also does not explain the very large genetic differences observed in the mitochondrial sequence data. CONCLUSIONS Gene trees are a critical first step when assessing biological diversity, but gene trees may be suspect if organisms have complex traits, such as unusual modes of genetic inheritance. We advocate that a compara- live approach is the best method to assess biological diversity in these under-studied taxa. In general, we found congruence among the different data sets for Cyprogenia aberti. The congruence of conglutinate color and the mitochondrial gene tree appears to support separate nuclear lineages in C. aberti without direct evidence from a nuclear marker. In inverte- brates, color morphs associated with mitochondrial lineages are nol uncommon, and these data have been used to identify species even when color forms are otherwise morphologically and ecologically indistin- 2004). followed, up to five species could be described in the guishable (Tarjuelo et al., If this precedent is C. aberti complex. Within the Ouachita River, two pue species exist: the red conglutinate produc- ers (cf. Fig. 3. clade A) and the brown conglutinate Fig. 3, clade B). A White River. producers and brown conglutinate producers co- Francis River (Fig. 3). biological surveys of rivers in the White River system producers (cf. similar division occurs in the with red conglutinate occurring in the St. Future will be needed to determine these two species’ ranges. astly, the upper Arkansas River contains a single species that produces white conglutinates. In contrast to conglutinate color, the data on host fish indicate that there are no differences host utilization among sympatric red and brown congluti- nate forms within the Ouachita and White systems. Volume 95, Number 2 2008 Serb & Barnhart 259 Biological Diversity in Cyprogenia aberti These data do not support, but also do not falsify, the hypothesis of sympatric species. Both genetic differ- ences and host compatibility differences among geographic populations are consistent. with limited gene flow and the possibility of several species within the Cyprogenia aberti group. These geographically clades are not ecologically separated C. aberti exchangeable. This conclusion is based on genetic divergence among lineages, difference in reproductive 9s. traits, such as conglutinate morphology. and prelim- inary data on host fish preference. Finally, the — UNDE among different data sets suggests that C. aberti is a species complex that contains cryptic will be described in a forthcoming manuscript. aa diversity. This diversity formally Literature Cited Arey, L. B. 1932. A Pa study of glochidial immunity. J. Morphol. 53: 367-379 Re] production. and fish host of the 50). Kansas De paren! of Wildlife and Parks, Pratt, Kansas . Ha aag & W. N. Roston. 2008. mw Lions to TN Ducato. in the J. N. Benthol. Soc. 27(2): 71-95. Bremer, K. 1994 Cladistics 10: 295- Chamberlain. T. K. es " rkansas fanshell, Cyprogenia aberti (Conrad). Biol. Bull. Barnhart, M. C esl i (Conrad 18 Unionoida. Amer. d support and tree stability. ode glochidial conglutinates of the Coker, R. E., A. F. Shira, Clark & . Howard. 1921. Natural lla e ps T ee mussels. eran ur. Fish. 37: 77-18 Cox, R. T. 1994. An rapid tec VPN to identify areas of possible Qui ternary alysis of drainage-basin symmetry as a ult-block tectonics: and centers of "m of the nea sired rocer- N w | icambarus (Decapoda: Cambaridae). Evolution. 53: Cross, F, B., a Mayden & J. D. Stewart. 1986. Fishes in the western n Basin. (Missouri, Red Rivers). Pp. 363-412 in C. H. Hocutt & E. O. Wiley (editors), The Zooge 'ography of North American Fre la T Fishes. John Wiley & Sons, New York. 2005. Universal primers for the specific Curole. J. P. 2005. amplification of the male a of the Unionoidea (Bivalvia). Conservation Ger 3-735. & T Kocher. E 02. Anci ient sex-specific extension of the cytochrome c oxidase I gene in bivalves and the POI 3 EOS “uniparental inheritance. Molec. Biol. Evol. 19 13 Davidson, E. east-west distribution of divergent mitoe honda haplotypes in British iuge of the land snail, Cepaea nemoralis (Pulmonata). Biol. J. Linn. Soc. 70: 697—706 Farrand & M. Ewing. 1962. Pleistocene 214 Reproduc Du Requirement Differences among I Cyprogenia aberti (Conrad. 1850). MS Thesis, Southwest Missouri State University, Springfield. solated Popul ations of Feder, J. L., T. A. Hunt € G. L. Bush. 1993. Geographic patterns of genetic differentiation between host-associated populations of Rhagoletis pomella. Entomol. Exp. Appl. 69: 117-1: 35. Folmer, O., . Hoeh, M. B. Black € R. L. 1994 DNA o for amplific ation of mitochondrial cytochrome C oxidase subunit I p Aes zoan inverte- brates. Molec “Mer. Bi ol. Biotech. : A. A., ", Stewart, B = elas d&E . The dis Dubon of male-transmitted and Vrijenhoek. Goodman, M., J. G pies iak, G. W. Moore, A. E. Romero- Herrera & G. Matsuda. 19 nd Fitting the gene lineage into its species lineage, a parsimony strategy illustrated. by cladograms constructed from globin sequences. Syst. Zool. 28: 132-168. Graf, D. L. 1997. Sympatric speciation of freshwater mussels (Bivalvia: Unionidae): A model. Amer. Malacological Bull 14: 35—40 & D. Ó Foighil. 2000. The shwater ^ arly mussels (Bivalvia: l. Stud. 66: 157-170. evolution of brooding haracters among the fre i i Unionoide E e ae America. J. M Hafner, M. A. Nadler. o Phylogenetic trees Toy he coe ae of parasites and their hosts. Nature 506-250. PNE . 1999, BioEdit: A user-friendly biological sequence eis nt editor and pus d pur n y indows 95/98/ NT. Nucl. / ipe i Symp. Ser. 41: Harrison, R. NÀ sides ny ne American Mitochondr ial field cric .M. pur opes of North Perspectives on the evolution of life cycles, bue assoc a d 2 Biol. 8: 209—232. Hoeh. T. Ste , B. W. Sutherland € E. Zour e Mele origins ag ander associated mitoc ‘hondrial \ lineages in biv alves (Mollusc a: Bivalvia). Evolution A pee 2286. songs. E vedra, B. W. Saa Philopenetle ev ide lace eal — ——__ Sutherland & Zouros. 1997. for role-reversals af gender-associatec | Nar rn DNA in Mytilus (Bival- via: Mytilidae). a Biol. ol. 14: 9: pss av mitoc hond ial genome doubls uniparental mode of inheritance in freshwater mussels (Bivalvia: VL a Evolution 56: 2252-2261 Hoggarth, M. A. 199 examination of the glochidia-host re lationship re pedi in ate literature for North American Bivalvia). Malacology species of Unionacea (Mollusca: Te Data Net 3: 1-30. Huelsenbeck, J. P. Ronquist. 2001. us ES: Bayesian inference of elote Bioinformatics 17: 155. Jones, J. W. 2004. Life history and propagation of the endangered dromedary pearly mussel Unionidae). J. N. Amer. romus dromas) (Bivalvia: Benthol. Soc. 23: 515—525 R. J. Neves. 2002. Life history and ecology of the fanshell (Cyprogenia stegaria) (Bivalvia: Unionidae). J. N. Amer. Benthol. Soc. 21: 7 : Kat, P. W. 1984. Parasitism and the Unionacea (Bivalvia). Biol. Rev. 59: 189-207. King, T. L., M. :kles, B. Gjetvaj & W. R. Hoeh. 1999, Intraspecific RE hy Lasmigona subviridis (Bivalvia: Unionidae): Conservation implications of range 5—578. Molec. Ecol. endangered discontinuity. Annals of the Missouri Botanical Garden R. A. Blaine & A. North Ame Phylogeography and 2006. Gene lineages palaeodrainage basins: Kozak, K. H.. Larson. and eastern ‘rican salamanders of the Molec. Ecol. 15: speciation in Eurycea bislineata species complex. 191-207. Liu. H.-P.. J. B. Mitton & S.-K. Wu. 1996. mitochondrial DNA differentiation far exceeds maternal mitochondrial DNA and allozyme Paternal differentiation in the fre a sr mussels, Anodonta grandis grandis. Evolution 50: 952-957. & R. L. Mayden. 1995, A southeast United Lydeard, diverse and ecosystem of the : 900—980: s endangered. aquatic States. : conservation Biol. Mavden, n bioge m of Noth ie rican central highland fishes. Pp. 141-151 in dec hes and historical Johnson de Quaternary PUMA of p Kansas Geological Sun Lawrence. Kansas. ———. 1988. Vicariance biogeography, parsimony, and e evolution i in | North American freshwater fishes. Syst. Biol. 7: 329-3! Me M thon, R 199 l. Bivalves. Pp. 315-399 in J. H. Thorp ido (editors), American Freshwater Invertebrates. Academic Press, San Diego. Melhorn, W. N. € J. P. ke o 1991. The Teays system: A | Melhorn & J. P. Kempton (editors), Geology and B of the Bedrock Valley System. (n cial Paper No. xS Meyers, T. R., R. E 1980. in A. G p hinaus ol de fishes IV. Humoral and tissue summary. Pp. 125-128 in Teays- Mahomet Ge ologie al Soc iely of America lillermann & C. Fustish. responses of coho and chinook salmon to experimental infection with Margitifera. J. Parasitol. 66: 274—281 Neves, R. J. & J. €. Widlak. 1987. (Bivalvia: Unionidae) ir Malacol. Bull. 5: freshwater mussels juvenile a headwater stream in Virginia. Amer. ES —— A. E. Bogan, J. D. Williams, S. A. Ahlstedt & P. W. Hartfield. 1997. Status of aquatic States: A diversity. Pp. 43-85 in G. W. mollusks in the southeastern downward (editors), Aquatie Fauna in Perspective. Southeastern Aquatic Lenz Design and Communication, Decatur, Georgia. Neves 90 i O'Connell, T. € R. J. immunological responses by a host fish (Ambloplites E I wo non-host fishes (Cyprinus carpio and LI "upestris) ) and t Carassius auratus) to glochidia a | freshwater mussel (Villosa iris). J. Freshwater Ecol. -7 Ortmann, A. E. the najades. Ann. Carnegie Mus. 8: 222-365 Page, L. M. 1983. Handbook of Dnus T.F.H. Pi Neptune City, New Jersey. 912. Notes bed us D sand genera of Page, R. D. 1995. Genes, d Md areas: The problem of rue lineages. Syst. Biol. 17-84 — ———. 1994. Parallel phylogenies: O LaS history of a assemblages. Cladistics 10: bd & M. Charleston. 1998. P hyloge ny and historical associations. Trees within trees: Trends Ecol. Evol. Passamonti, M. & V. Scali. 2001. DNA het in the ae pu (Mollusca Bivalvia). Curr. Genet, 39: 117-12 Gender-associated mitochondrial eroplasmy venerid clam Ecology and Classification of Habitat ecology of Backeljau. 2005. Extreme lerrestrial slug (Gastropoda, Pinceel, J.. K. mtDNA diver Arionidae): Jordaens & T. gences dn a Accelerated: evolution, ae Biol. 18: Pulmonata, divergence and secondary contact. J. Evol. 1264-1280. Posada. D. & K. A. Crandall. model of DNA sittin M Roberts, D. M. €. Ba 1998. Modeltest: Testing the Bioinformatics 14: 817-818. J99. Effects o id COs on transformation of glochidia = rnhart. Le So rature, MH. ar of the flat floater mussel, Anodonta suborbiculata. oni Benthol. Soc. 18: 477-487. Robison, H. W. T. M. Arkansas. Univ. of Arkansas Press, Fayette Roe, K. J. 1999. ee ar Systematics Freshwater Mussel Tribe | lenta 1992. Fishes o ; and E iren on of the Lampsilini (Bivalvia: Unionidae). Univer SN of Tuscaloosa, Ph.D. Dissertation, Alabama, Alabama. & . Lydeard. 1998. freshwater specs genus Potamilus (Bivalvia: Unionidae). Malac po 39: MW Molecular systematics of 1 )5-205. . Hartfield € C. Lydeard. 2001. Phylogeo- "e analysis of the threatened and endangered super- conglutinate "producing. mussels of ne. genus s Lampsilis (Bivalvia: pup e). Molec. Ecol. 10: . Dimock. 2003. "e uire ae re of bluegill suní os Lepomis macrochirus to gloc hidis a pes ae of Utterbackia imbecillis multiple infections. J. Rogers, C. L. the freshwater mussel (Bivalvia: Unionidae) after Parasitol. 89 1-50. Rogers. S. O.. e history Watson & R. J. Neves. 2001. E and population biology of the ad tan sere ll psi sma ei d (Bivalvia: Unionidae). J. N. Amer. Benthol. Soc. 20: 5 —504. C. & E - Dimoc k. 2006. E neapsulation of f fre -— Rogers-Lowery, attached ectoparasitic glochidia larvae sshwater mussels by epithe hial 2 on fins of naive and resistant host n Biol. Bull. 3 Rosen, E. "s icariant patterns and historical 5 38 i nia in M eae Syst. Zool. 27: 159- Routman, E., R. Wu € A. R. Templeton. 1994. Parsimony. molecular evolution. imd biogeography: The case of the North American giant salamander. Evolution 48: 1799— 1309, Serb, J. M. 2006. Discovery of genetically distinct sympatric lineages in the freshwater mussel Vi aberti (Bivalvia: Unionidae). J. Moll. Stud. 72: 425—42- —— ——, J. E. Buhay € C. Lydeard. 2003. lema systematics of the North freshwater bivalve (Unionidae: | sequences, Ambleminae) based on Molec. Phylogenet. Evol genus Quadrula ~ hondrial ND : 1-11. E C.. J. Tang. S. Dalwadi. G. Staley. J. Deniega & T. R. Unnasch. 2000 Cenc: evidence for assortative mating between 13-year cicadas and sympatric “17-year cicadas | e cycles" provides support for allmattis spe 1326-1330. Skibinski, D. O. F., C. Gallacher € €. M. Be ynon. 1994, Sex- limited mitochondrial DNA transmission in ll marine ) mussel Mytilus edulis. Genetics 138: 801— orenson, M. D. 1999. TreeRot. Boston University. Boston. Stewart, | C. Saavedra. R. R. Stanwood. A. O. Ball & E. Zouros. 1995. Male and mitochondrial DNA lineages in a blue musse yh (Mytilus edulis) species group. Molec. Biol. Evol. 12: 735-747. Benito: D. L. un PAU a Parsimony (and other Sunderland, Massachusetts lemale Phylogenetie Analysis Using mend Sinauer Associates, Volume 95, Number 2 2008 Serb € Barnhart Biological Diversity in Cyprogenia aberti 261 K. A. Crandall. Phylogeography and s Tarjuelo, D. Posada, . Pascual & X. e morphs in the colonial ascidian Pseudodistoma crucigaster. Molec. Ecol. 13: 3125-3136. Turon. | Service. 1990. Endangered and threatened wildlife and plants: Designation of the freshwater mussel, the fanshell as an endangered species. Pp. 2 a Federal Register. Vaughn, C . Regional patterns of mussel species modal in enn American rivers. 107-115 Walker, J. M., J. P. Curole, D. E, Wade, E. G. s A. E. Bogan, G. T. Watters /. R. Hoeh. 200€ distribution and polsenei utility of g de r-associated mitochondrial (Bivalvia). | Malac ologia - X E CORTE iphy 20: a, . Taxonomic genome U ond 265— 292. .. Mayden. 1985. Species sand speciation in phylogenetic sei s, with examples from ps North uie fauna. Ann. Missouri Bot. Gard. 72: 596-635. Wood, basis i re um isolation in the sympatric Enchenopa binotata us (Homoptera: Membracidae). Evolution 36: 233-242. . Guttman. 1982. Ecological and be a K. R. Freeman, A. O. Ball & G. H. Pogson. 1992. Direct evidence for extensiv inheritance. in. the 412—414. | A. O. Ball, C. 1994. An unusual type of mitochondrial DNA inheritance in the blue mussel Mytilus. Proc. Natl. Acad. Sci. U.S.A. 91: 7463-7407. Zouros, E., e paternal mitochondrial DNA marine mussel Mytilus. Nature 359: Saavedra & K. R. Freeman. APPENDIX 1. Voucher information for specimens sequenced in this study. GenBank accession numbers for the first subunit of cytochrome oxidase c (COL 1) and the first subunit of the NADH pare mG "ses. B dd S, respectively, are in V ouche Pos are deposited at ia liste i = museum acces and GenBank accession aba for other spec cies are found in Serb (2006). speciation of colour ls nia aberti: — Fall River (1). U.S.A. Kansas: he ilson Co., University of Al: ae t nionid C Ys ection [UAUC] 1646 p 35. DQ494711), UAUC 1647 (DQ494728, DQ4947€ St. Francis River (3). U.S.A. Missouri: Wayne UTM 15 722800E 4123900N (Patterson quad), Bl (DQ6 060 6: DQ663617), B2 (DQ667061, DQ663618), B3 (DQ667 DQ663619), B4 (DQ667063, DQ663620). B 5 (DQ667064. DQ663621), R6 (DQ667065, DQ663622), R8 (DO667066, DQ663623). Buffalo River (4). U.S 1650 (DQ494730, DO4947 n Strawberry River (5). U. 1652. (DQ494744., (bQ4 947 15, DQ494720). rrent River (6). U.S.A. Arkansas: Clay Co. 1537 (DO404727. DQ494701). Black River (7). U.S.A. Arkansa Co.. UAUC 1648 (DQ4947 29. DQ494709), (DQ494742, DQ494717). Missouri: Butler Co., 732729E 4071085N, js e (DQ063624). Vhite River (8). Arkansas: Woodruff/Jackson/ White Co., UAUC vs E 36, DQ494712). ~ hita River (9). U.S.A. | UTM 15 434359E 3828410N (Mount Ida (DQ667054. DQ663611), B2 (DQ667055, DQ663612), B3 (DQ667056, DQ663613), R5 (DQ667057. DQ663614). RO (DQ667058, DQ66361 5), me (DQ667059, DQ663616). Saline River (10). l Arkansas: Saline Co., UAUC 76 pane 33, AY 1587: o Caddo River ). U.S Arkansas: E 1837 (DQ494739, DANI. UAUC » 8 DOMUS 43, 10494718). Arkansas: Clark Co., UAU E (DQ494741, DQ494716), UAUC 2383 (DQ494749, Do 194724), UAUC 2384 (DQ49447, DQ494722 G yproge nia stegaria: . Arkansas: Newton Co.. UAUC Arkansas: U A. Sharp Co., DQ4947 19), AUC 2726 , UAUC iS! Randol phil awrence Arkansas: Montgomery Co., quad), Bl k Co., UAUC 1500 (DQ494726. Clinch River. U.S.A. Tennessee: Hancox 1499 (AY654992, Bs UAUC DQ4947006) THE IMPACT OF PETER RAVEN ON EVOLUTIONARY AND BIODIVERSITY ISSUES IN THE 20TH AND 215T CENTURIES: INTRODUCTION” * Peter C. Hoch? 'he 53rd Missouri. Botanica Annual Systematics Symposium of the Garden, 2000. had as its theme “The Impact of Peter Raven on Evolutionary and Biodiversity Issues in the 20th and 21st Centuries.” one individual, a departure from previous Garden symposia based on single topics, was occasioned by the celebration in 2006 of the Peter of the Missouri haven's arrival as Director Botanical Garden. Given that two thirds of the Garden's annual symposia have taken place during his tenure, this seemed a particularly appropriate opportunity to. celebrate. Peter Raven's intellectual contributions to the broader scientific community and, indeed, to the world at large. A committee of Raven's colleagues (Olga Martha Montiel, Barbara Schaal) and former students (Paul Berry, Peter Hoch, and Warren L. decide what issues to cover and who would speak. The Wagner) convened to selection process was difficult. Raven's intellectual contributions and the constraint of only eight speaking slots. Faced with his massive publication record (more than 650 entries at the end of 2005), made major contributions that continue to be important the committee chose topics to which Raven areas of research, hence the selection of population biology/gene flow. biogeography. coevolution. pollina- — tion biology, and cytogenetics. Other areas to which Raven has made immense contributions are less easily characterized, but we determined that the broad area of conservation/biodiversity certainly should be included. Finally. in thinking not only about the exceptional | held on 13-14 October This focus on the work and impact of 35th anniversary of given the breadth of intellectual leadership he has provided to the Garden for 35 years. to his own students, and to legions of other students influenced by his three top-selling college textbooks, but also about the many institutions he has Academy of . the Na- tional Science Board, among others, the inmiversités at served with distinetion—the National Sciences. the National Geographic Society which he has held positions, and the many scientific societies to which he belongs or has been elected—we decided we needed a talk that would address Raven's in the broadest contributions to “building capacity” sense. Fortunately, to round out the final slate of eight talks, P which provided a more personal perspective on the very aul Ehrlich agreed to give the evening address. a fruitful collaboration he shared with Raven over the course of more than 40 years. Due to various extenuating circumstances, not least the unique historical perspective adopted by most of the speakers. only two manuscripts resulted from this symposium, those by Diane Campbell (pollination i weck What follows here is a brief synopsis of the remaining talks. biology) and J. Chris Pires and Kate L. Hert (cytogenetics), which follow this introduction. in the order of their presentation at the symposium. Barbara Schaal (Washington University) presented "Differentiation of Populations: Gene Flow Redux." revisiting the seminal paper published in Science i 1969 (Ehrlich & Raven. 1969). Populations.” This paper challenged the prevailing "Differentiation. of orthodoxy that species were evolutionary units held together by gene flow, and opened a broad and productive field of research into the empirical 'This and the two articles that ved it are the proceedings of the 53rd Annual Systematics Symposium of the Missouri “The Impact of Pe E 13-14 October ie Botanical Garden. sy poso Was This was the 5 Science F n un (grant DEB-0515 Hoch of the Misso Wagner (Smithsonian Institution). I uri Botanical Garden, with the , pee and able assistance of Mary r Raven on Evolutionary and Biodive t the Missouri Botanical Garden in St. Louis. t Missouri Botanical Carde n Annual Systematics Symposium to 933). The meeting was organized by Barbara Scha assistance of Paul Berry (then at rsity Issues in the 20th and 21st Centur Missouri. U.S.A. be supported by a grant from the National al of W ashington University and Peter University of Wisconsin) and Warren Mick Richardson was responsible for the smooth running of the symposium, with the McNamara, Sandy Lopez. Donna Rodgers, William Guy. and Zubin Chandran. Victoria . Hollowell, Beth Parada, and Allison Brock (Missouri Botanical Garden Press) were responsible for the publication of the symposium proceedings. This symposium, together with 36 previous ones. again benefited conside rably from the contributions. [e SUP} port, and encouragement of Peter Ray 'Miss souri Botanic al Garden doi: 10.3417/200805 95: 262—203 ANN. Missourt Bor. GARD. . P.O. Box on St. Louis, Missouri 63166-0299, U.S.A. peter.hoch@mobot.org. PUBLISHED ON 18 June 2008. Volume 95, Number 2 2008 Hoch Introduction 263 measure of gene flow, which remains active and vibrant to the present. Through the use of several model plant systems, Schaal explored the processes that enhance and retard population differentiation. Stephen J. O'Brien (Laboratory of Genomic Diver- sity, National Cancer Institute) presented “Genomic Archaeology —Solving the Mysteries of History.” which explored the topic of diversity from the perspective of comparative genomics in mammals. and Re recent advances in conservation genet- s. Using examples from his remarkable genomic a diverse mammals as cats anc 2003). approaches to the elucidation of biological diversity. Mae on such elephants (O’Brien, O'Brien discussed new with applications for sampling methods and the analysis of extinction, in these topics echoing many publications by Raven (e.g., Phillips & Raven, 1997; Raven & McNeely, 1998) Ulrich Mueller (University of Texas at Austin) presented “Coevolutionary Principles of Insect Fun- giculture: Lessons for Human Agriculture?,” which demonstrated the rich intellectual fruit borne from the wlich Evolution. — original planting of the idea of coevolution by E and Raven (1964) in their article in “Butterflies and Plants: A Study Using the remarkable adaptations found in fungicul- 2006), Mueller demonstrated the sophistication of recent coevolu- =. Coevolution.” turist attine ants (Gerardo et al., tionary analyses and presented the ant-fungus-mi- crobe interactions as an intriguing analogy to human agriculture, with important potential applications for improving the latter. Jun Wen (National Museum of Natural History, “Evolution of Smithsonian Institution) presented ajor Patterns of Plant Disjunctions,” reviewing and updating many of the hypotheses proposed by Raven e.g., Raven, 1963), most nota his collaborative work with the late Daniel Axelrod , Raven & Axelrod, 1974, 1978). Wen enumer- the major intercontinental disjunction patterns, — over many years (e bly in with many examples from recent work, noting the insights gained from modern phylogenetic analyses and the nascent field of lineage dating based on fossil and molecular data. Cristián Samper (National Museum of Natural History, Smithsonian Institution) presented “A Ra- ven's Garden: Building Capacity for Plant Research and Conservation," in which he discussed the need for building capacity in all countries and how that relates destruction of biological diversity to the rapid A a E California flora. Üniv. Calif. & J. A. worldwide. He then explored some of the impacts that Peter Raven has had in this area, developing flora projects in many biodiverse regions, placing staff in the countries of origin, developing tools to enhance the flow of information about species, and developing long-term partnerships for conservation. Paul Ehrlich (Stanford Symposium with his evening address, University) closed the “Saving the World,” an informal and humorous recollection of his long and fruitful collaboration with Raven, and how some of those ideas have developed into new fields of study. With a playful allusion to the title of one of Raven’s (1988: “We're Killing Our World”) singular contributions on conservation and the human impact on the global ecosystem, Ehrlich shared his deep passion for the study of biology and the need for (Ehrlich, 2000), and his personal insights and humor regarding personal responsibility in conservation his relationship with Peter Raven made for a fitting close to a day of celebration. Literature Cited Ehrlich; P. R. 2000. Human Natures: Genes, C ultures, and the Human Prospect. Islan & . Raven. 1964. Butte thie 's and plants: / in coe volution Evolution 18: 586— ress, Washington, D.C vere — & ——. 5 Differentiation of populations. Science 165: 1228-1232 Gerardo, N. M., U. ; R. Currie. 2006. Complex host- NR cóevolution ir the Apterostigma | £ ymbiosis. BMC C Ev ol. Biol. 6: ungus-grow ing ant-microbe O'Brien, : Tears of the Cheetah: The Genetic ur a Du Vos Ancestors. St. Martin's Press, Ne Phillips. O. L. & P. H. Raven. 1997. A strate g y for sampling 65 in A. C. Aso cal Biodiversity and ers Univesity of neotropical forests. Pp. 141-1 n (editor), alifornia, io Angeles. Raven, P. H. 1963. Amphitropical relationships in the floras of North de South America. Quart. Rev. Biol. 38: 151— 177. 1988. We're vstem in Crisis Killing Our World, The Global The MacArthur Foundation Occ sional al Paper Mac echo Pow Chicago. D xelrod. 1 . Angiospert D O and past 2 M movements. Ann. Missouri Bot. Gard. ol: 1978. Origin and EE of the Publ. Bot. 72 134. McNeely. 1998. ip extinction: lts scope and. meaning for us. Pp. 13-32 in L. Guruswamy & J. A. McNeely (editors), eds of Global B Converging Strategies. Duke Univ. Carolina. Biodiversity: North Press, Durham, POLLINATOR SHIFTS AND THE ORIGIN AND LOSS OF PLANT SPECIES' Diane R. Campbell? ABSTRACT Pollinators have long been implicated i rgetics of animals into our Min standing of diversification of floral traits. More rece ntlv. efforts biologists to consider the implications of a no and losses due to human activities. This n plant s spec lation, how Peter Raven's earlier work was instrumental integrating 1 major pollinators influence the ey shifts volutionary y Raven and others in the area of conservation have inspired pollination s paper uses ihe shift between hummingbird and hawkmoth pollination as a model for exploring impacts of pollinator shifts on plant populations. Recent studies M ichx. " ave quantifie d the degree of reproductive isolation due to such pollinators in several genera. Data from /pomopsis urther allow us to consider whether recent changes in pollinator regimes have demographic consequences for plant de A majority of piant populations may currently suffer from pollen limitations on seed production, but few data exist or the der emogt raphic consequ ces ol pool reproduction. In /pomopsis, reduced seed production due to pollen limitation can impact the number of inc aaa surviving to reproduce in the next generation. Some diues e of I. tenuituba (Rydb.) V. E. Grant are estimate hawkmoth pollination. In the absence of an increase in hawkmoths, selection for wider corolla tubes and other could, in prin a 1 by natural selectior d to have finite rates of increase less than unity, which can be expla may leave the population vulnerable to loce uned in part by current low levels of floral traits ciple, attract enough hummingbird pollination to result in a growing population, but models show that such extinction. We need more studies of the ou demographic consequences of changes in pollinator regimes. Suc hh studies should consider how evolutionary changes influence the risk of extinction. Key words: Aquilegia, coevolution, extinction, pollinator shift. speciation. hawkmoth. hummingbird. /pomopsis. Nicotiana. pollen limitation. Changes in pollinator regimes are thought to be of critical importance in speciation and diversification of the flowering plants. Such changes include qualitative shifts from one type of pollinator to another, quantitative shifts in the relative proportions of major types, and overall reductions in visits by animal pollinators that select for wind pollination or self- fertilization. In the past few decades, biologists have made considerable progress in testing the role of pollinator shifts in fundamental aspects of the origin of species: diversification of floral traits and develop- ment of reproductive barriers between incipient species. They have also begun to evaluate the role played by recent changes in pollinator regime for the potential loss of species. One major goal of this paper is to explore the roles played by pollinator changes in both hummingbirds and hawkmoths as a case study and the origin and the loss of species, using paying attention to the historical roots of these ideas, particularly those resulting from the work of Peter Raven and his colleagues. A second major goal is to integrate the two themes by considering how we can incorporate. evolutionary. studies into examining the impact of pollinator losses. Where data are otherwise scarce in the current literature, I will focus on studies Michx. including of Ipomopsts (Polemoniaceae). ongoing work. PETER RAVEN AND POLLINATION BIOLOG) Pollination is central to so much of angiosperm evolution that it perhaps should not surprise us that so many diverse aspects of Raven’s work have made a Although n pollination focused mainly on mark on pollination biology. Raven's experimental work 1 specialized los systems in the Onagraceae, ue molut his widest impact on this field can be traced to three papers that introduced new conceptual approaches during the 1960s and 1970s. Prior to that time, much of pollination biology emphasized de- scriplive analysis of associations between flower characters and types of pollinators (Faegri & van Pul. 1966). One of the pollination biology occurred with a fuller incorpora- der biggest changes in l thank Peter Hoch and Barbara Schaal for the invitation to ay ipate in the 5 Missouri Botanical Garden. Comments by Mary Price, Robe contributed to revision of this manuscript. The author's work is Med by grant DEB-0542876 from the I oundation. “Department of Ecology and Evolutionary Biology. ue. edu, doi: 10.3417/20070006 ANN. Missouri Bor. University of. California, 53rd Annual Systematics Symposium of the Sehlising, Niekolas Waser, and an anonymous reviewer National Science Irvine, California 92697, U.S.A. dreampbe@ Garb. 95: 264—274. PUBLISHED ON 18 JUNE 2008. Volume 95, Number 2 2008 Campbell 26 Effects of Pollinator Shifts on Plant Populations tion of the animal's perspective, a change precipitat- ed, in part, by publication of Ehrlich and Raven's (1964) landmark paper on coevolution and Heinrich and Raven's (1972) fundamental paper on foraging energetics of pollinators. In addition, the idea that gene flow is too restricted to be responsible for holding together a species (Ehrlich & Raven, 1969) sparked a whole industry focused on estimating movement of pollen (e.g., Levin € Kerster, 1974). These three papers by Raven helped to launch a renaissance in pollination biology, as biologists increasingly recognized plant-pollinator systems as models for understanding major issues in ecological and evolutionary processes. As the ideas of gene flow through pollen, and of coevolution, are considered elsewhere in the symposium, I will focus on what follows on other influences, especially the perspective brought by considering animal foraging. a J te] D D POLLINATOR SHIFTS, FLORAL DIVERSIFICATION, AND ORIGIN OF PLANT SPECIES HUMMINGBIRD VERSUS INSECT POLLINATION Some of the most compelling evidence for the role of pollinators in speciation comes from studies of related plant species that are pollinated either by hummingbirds or by insects. This emphasis is perhaps features natural, given the longstanding interest in ten found in = 0 hummingbird-pollinated plants: red, tubular flowers that are relatively broad (compared, for example, to moth-pollinated flowers) and produce [em large quantities of relatively dilute nectar (comparec 1978). Raven (1972) considered the question of why so many to bee-pollinated flowers; Bolten & Feinsinger, plants visited by hummingbirds have red flowers. He argued that birds learn to recognize red as a signal for a large calorie nectar reward even though they have no — intrinsic preference for red, and that red is the signal because it is less conspicuous to most insects. This topic has continued to fascinate pollination ecologists and remains not fully resolved. Recent studies have altered the picture by demonstrating that bees can perceive red flowers, although the color is more difficult for them to distinguish against a green background of leafy vegetation (Chittka & Waser, 1997): presumably renders red flowers less efficient, but not 2001). The orimary significance of reward energetics present in J o o this results in increased handling time and invisible, for bee foraging (Spaethe et al., Raven's original argument for flower choice has been amply documented and reinforced. This emphasis on rewards was fleshed out in the highly influential paper by Heinrich and Raven (1972). The idea, simply put, was that the caloric reward of a flower could be used to predict which pollinator species would visit it, the number of flowers visited per plant, and other aspects of foraging biology that could influence important features of plant reproduction such as the outerossing rate. Along with the concept of coevolution between insects and plants elaborated by Ehrlich and Raven 1964), this paper stimulated a new generation to work on pollination from both the plant and animal points of view. In particular, it emphasized the idea that one could predict aspects of pollination by understanding (Waddington, 1983), an idea that also owed much to the interjection of foraging theory (Pyke, 1984). Foraging energetics ultimately offered new perspectives on the adaptive ie. foraging energetics of the animals significance of many plant traits, such as number of flowers (Wyatt, 1982), distribution across flowers (Best & Bierzychudek, 1982), & Price, 1983), packaging of pollen (Harder & Wilson, 1994). per inflorescence nectar nectar guides (Waser and pus QUANTIFYING POLLINATOR-MEDIATED SELECTION AND REPRODUCTIVE ISOLATION These ideas about foraging energetics also revitalized work on speciation resulting from pollinator shifts. Speciation involves divergence of notype and development of reproductive isolation, and both could be produced by pollinator shifts. Different types of have pollinators (for example, bumblebees vs. birds) different energetic. requirements, morphologies, and sensory systems that may lead them to visit plants with Such | livergent selection mediated different suites of floral traits. behavioral differences would y [om by pollinators, potent ellie leading to differentiation of the floral phenotype, and could also generate repro- ductive isolation (see below). While this viewpoint was clearly articulated early on (Grant, 1949), it took much longer for biologists to document selection mediated by pollinators. In an introduction to an influential volume on pollination biology published two decades ago (Real, 983), necessary to document that genetically variable f Raven remarked that it will ultimately be oral traits actually contribute to reproductive success. This feat has since been accomplished for several species (e.g., Raphanus raphanistrum L. by Stanton et al., 1986; Polemonium viscosum Nutt. by Galen, 1996); however, most studies have focused on selection within a particular plant species. Only recently has the question of divergent selection and reproductive isolation by divergent pollinators been addressed with the same rigor. However, recent studies of genera with both hummingbird and hawkmoth pollination (Aquilegia L., Ipomopsis, and Nicotiana L.) offer repeated examples in which the extent of reproductive isolation mediated by pollinators has been quantified. Annals of the Missouri Botanical Garden To see how reproductive isolation has been quantified, it is important to consider in more detail With pollinator-mediated divergent selection, hybrids be- how pollinator behavior can generate it. tween the two plant species could have low pollination and, therefore, low fitness, a form of postzygotic — reproductive isolation. Furthermore, if pollinators focused on a single species because of its more suitable reward or morphology, pollen movement between species would be restricted, causing pre- zygotic reproductive isolation as well. This contribu- tion of behavior to reproductive isolation is referred to as ethological isolation. (Rlamotogica) and can be quantified as: heterospecific plant transitions Rlethological =] conspecific plant transitions where transitions refer to movements made by pollinators in a setup with equal choices available. For cases of species visited predominantly by hummingbirds and hawkmoths, estimates for etholog- ical isolation range from 0.50-1.00, a considerable range (Campbell & Aldridge, 2006) that begs further consideration. A second form of reproductive isolation — (mechanical isolation) can also be produced by ix pollinators when they move pollen more efficiently between conspecific flowers because of differences in aspects of floral morphology such as the positions of reproductive organs. CASE STUDIES WITH HUMMINGBIRDS AND HAWKMOTHS Two species of Aquilegia illustrate the case of strong ethological isolation and show the contribution of at least one specific floral trait (Fulton & Hodges. 1999). Aquilegia formosa Fisch. has pendent, red and yellow flowers with short nectar spurs, whereas A. pubescens Coville has upright, pale flowers with long spurs. Although visited primarily by hummingbirds and hawkmoths, respectively, the two species produce the southern Sierra Nevada of hybrid zones in California, a situation originally studied by Chase and Raven (1975). In an array of A. pubescens in which half of the flowers were tied to make the flowers pendent and, thus, given the trait of the other parental species, hawkmoths visited upright flowers more than o P 10 times as ¢ 1999). produced strong preference for A. pubescens; however, as pendent flowers (Fulton & Hodges, dissipation of this one trait other floral traits such as flower color are likely also involved, as the preference was still not as extreme as observed in arrays of the two parental species (Fulton 1999). clearly involved in strengthening the current repro- & Hodges, Although flower orientation is ductive isolation, it is not clear whether this trait was the particularly without information on how each trail also involved in original speciation event, influences the energetics of nectar extraction by hawkmoths. Hodges and his collaborators are now studying the genetic basis of the floral traits that maintain reproductive isolation (Hodges et al., 2002). Nicotiana alata Link € Otto and N. forgetiana Hort. ex Hemsley in Brazil provide a similar example in which prezygolic reproductive isolation is normally 2004). In experimental plots strong (Ippolito et al., with only plants of the two species, hummingbirds visited N. forgetiana exclusively, and hawkmoths strongly preferred N. alata, in agreement with their pollinator syndromes. However, the presence of some hybrids led to high interspecific pollen. transfer, producing a “hybrid bridge" between the species (Leebens-Mack & Milligan, 1998). When contained the two species along with F, individuals, both that movements could have produced interspecific gene the finding plots hummingbirds visited species, so their flow, and backcrossing between hybrids and parental species was also possible. This supports the suggestion of Arnold (1997) that even if F; hybrids rarely arise due to prezygotic ethological isolation, their presence can accelerate further gene flow between species. RELATIVE POLLINATORS AND HABITATS IMPORTANCE OF SELECTION MEDIA’ ED BY Besides pollinator differences, an alternative form of selection that could drive plant speciation is differ- — ences in habitat such as soil types, other physical attributes of the environment, or biotic interactions that are unrelated to pollinators (Waser & Campbell, 2004). Adaptation to different habitats could generate postzygotic reproductive isolation if hybrids are unfit as a result of genomic incompatibilities or specifically unfit in the Schluter, that habitat differences might actually be parental environments (Hatfield & 1999). Chase and Raven (1975) SE more important than pollinator differences in maintaining the species difference between Aquilegia. In principle, these two forms of selection could be distinguished by comparing the extent to which hybrids suffer from low pollination success versus low survival or other aspects of fitness. Few studies have examined pollination success of plant hybrids, although visitation to hybrids was studied in the Nicotiana arrays mentioned above ito et al., 2004). In /pomopsis (Polemoniaceae), the importance of (Ippo both pollinator-mediated and habitat-mediated selec- tion has been investigated. As in the two cases the conspicuously in floral mentioned above, two species involved differ trails. Ipomopsis ageregata Volume 95, Number 2 2008 Campbell 267 Effects of Pollinator Shifts on Plant Populations (Pursh) high nectar production and often exserted anthers and V. E. Grant has broad-tubed red flowers with stigma, whereas /. tenuituba has narrow-tubed, long, white or pale pink flowers with lower production and strongly inserted reproductive parts. l: 1993), and Grant and Grant (1965) suggested they provide an example of recent or incipient speciation These two species are close relatives (Wolf et a via pollinator-mediated divergent selection in which /. aggregata is pollinated by hummingbirds and /. tenuituba by hawkmoths. However, at some contact sites, the two species are visited by both pollinator types when present, and, as a result, they often form extensive hybrid zones in nature (Grant & Wilken, 1988; Wu & Campbell, 2005). One such hybrid zone occurs in Poverty Gulch in Gunnison County, Colorado. Hybrids between the two species of /pomopsis taken from that area are easily made, with seeds resulting from just as many heterospecific crosses as from conspecific crosses (Campbell & Waser, 2001). To determine how hybrids compare with the parental species in pollination success versus other aspects of fitness, Campbell and Waser (2007) planted more than 3000 seeds from conspecific and heterospecific crosses (Fy reciprocal —— hybrids) into the two parental sites and a site where natural hybrids are common and followed the progeny for 12 years until nearly all of these Eie perennial plants had completed their life cycle. Differences in survival and reproduction both nr substantial contributions to the fitness. differences. Overall, survival of Fı hybrids was equivalent to the average of the parental species; however, in the two parental sites, the direction of the cross was critical. Hybrids with 7. tenuituba as the mother survived well only in the natural hybrid site, suggesting that habitat differences are important. One factor involved is water availability. The hybrid site is warmer and drier than the parental sites, and the hybrids have higher photosynthetic water-use efficiency than the parental 2005). critical. In experimental arrays of potted plants with species (Campbell et al., Pollination is also equal numbers of each of the parental species and hybrids, plants of /. aggregata received and donated the most pollen to stigmas of other plants (as overall 2002). Hybrid pollination success was intermediate between estimated by dyes), thereby enjoying the highest pollination success (Campbell et al., that of the two parental species in one experiment and equal to /. aggregata in another. In a separate study, hybrids produced and sired intermediate numbers of seeds to those of the parental species (Meléndez- 1998). So, suffer low pollination success in situations where only Ackerman & Campbell, hybrids do not hummingbirds are present. nectar These primarily the visitation preference of hummingbirds patterns in pollination success reflect for Ipomopsis aggregata and hybrids rather than /. 1997). In this area, two species of hummingbirds, broad-tailed hummingbirds tenuituba (Campbell et al., and rufous hummingbirds, make most of the flower visits. Their preference also results in partial ethological isolation, with about 796 as much pollen transferred between species as transferred conspeci- 2002). Mechanical isolation although to a fically (Campbell et al., well, 1998) The visitation preference of hummingbirds reflects contributes as esser extent (Campbell et al., the wider corolla tubes, red color, and greater nectar production of /pomopsis aggregata (Campbell et al., 1997; Meléndez-Ackerman € Campbell, 1998). In particular, the preference of hummingbirds for Z. aggregata generates directional selection for wide corolla tubes. One species of hawkmoth (Hyles lineata) is also a visitor at this study site but only in rare years (two years out of the last 20). In years when they are present, they disproportionately visit plants with narrower tubes. The combination of birds and moths in the natural hybrid zone then generates disruptive — selection on corolla width, in line with the expectation that pollinator-mediated divergent selection maintains species differences (Campbell et al., 1997). Campbell and Waser (2007) used life table analysis to show that differences in age-specific survival and both the patterns of lifetime fitness seen in the plants in the fecundity made large contributions to reciprocal transplant study described above. The low fitness of [pomopsis tenuituba at I. aggregata sites could be explained by low pollination, whereas the low fitness of Fy hybrids with 7. tenuituba mothers at both pollinator-mediated and habitat-mediated selection — hat site reflected poor survival. In. summary, contribute to the fitness differences seen in this hybrid zone and to maintenance of species differences. They may also have contributed to the original speciation 1993). Another way to approach the problem of the role between these sister species (Wolf et al., played by pollinator-mediated selection is to compare cases where species hybridize with cases where they the lack of hybridization can be do not, to see if explained by lack of shared pollinators. Such a comparison is most informative if it involves the same species pair in different localities. /pomopsis aggre- gata and I. tenuituba show considerable geographical At Grizzly Ridge on the north rim of the Black Canyon of variation in the extent of hybridization. m Gunnison, Colorado, there are few if any morpholog- ical hybrids. That site thus provides an extreme contrast with the Poverty Gulch site, where hybrid- ization is extensive. This difference can be explained 268 Annals of the Missouri Botanical Garden by differences in behavior of pollinators between the two sites. At Grizzly Ridge, the same species of hummingbirds and hawkmoths found at Poverty Guleh are completely species-specific in foraging. Experi- ments with potted plants in four combinations, representing all combinations of site of origin of the plants and site of observation, showed that the difference in ethological isolation was primarily the result of hawkmoths changing their behavior between sites (Aldridge € Campbell, 2007). n summary, the extent of reproductive isolation due to behavior of pollinators has now been documented in at least a few cases, with Aquilegia, Nicotiana, and fpomopsis serving as examples of species pairs involving hummingbird and hawkmoth pollination. More work will be needed, however, to quantify the general importance of these differences in pollinators versus differences in habitats, to the two facets of plant speciation: divergence in phenotype and development of reproductive isolation. ANTHROPOGENIC CHANGES IN POLLINATOR REGIMES AND POTENTIAL FOR SPECIES Loss POTENTIAL CONSEQUENCES Plants and their pollinators, like other parts of the natural world, face increasing peril from human activities. One example is the recently documented parallel decline in insect pollinators and insect- pollinated plants in Britain and The Netherlands (Biesmeijer et al., 2006). Understanding and mitigating these perils is an important challenge for pollination ecologists and an important arena for future work. Peter Raven’s extensive work in conservation biology has helped to fuel wide concern over threats such as habitat loss and fragmentation, global warming, and invasive species (Raven, 2002). While a recent report from the Nalional Academy of Sciences (National Research Council, 2007) points out the considerable gaps in our knowledge of effects on pollination, concern over the potential for harm is rising. Potential consequences of pollinator changes extend not only to native plant populations but also to agriculture, as 60% of the world's crops depend on honeybee pollination or pollination by other bees (Roubik, 1995). Pollinator changes include shifts in species com- position of pollinators and reductions or even losses of pollinators. These changes can potentially have demographie and/or evolutionary. consequences for plant populations. Demographics can be impacted if (1) the change increases pollen limitation of seed production, and (2) the lowered seed production alters population growth. If the rate of increase for the population drops below replacement, the result could even be local extinction, especially if seed production drops even farther in small populations (as in the Allee effect seen in Banksia goodii R. Br.; Lamont et al., 1993). Even if demographic changes are not seen, evolution could be altered if the new pollinator regime selects for different floral or reproductive traits. POLLEN LIMITATION Most effort to date has focused on direct. demo- eraphic changes, rather than evolutionary ones. The first. requirement for a demographic change has considerable evidence. In the 1970s, it was proposed that seed production would rarely be pollen limited 1979; Willson, 1979). papers were based on application of current sexual Charnov, These influential a selection theory to plants and contributed, along with the conceptual papers by Raven mentioned above, to a renaissance in pollination ecology. Among other things, the notion of sexual selection spawned a large number of studies of pollen limitation (Burd, 1994). Pollen limitation can be demonstrated by an increase in reproduction. when pollen loads on stigmas are supplemented by hand pollination, ideally of all flowers on a plant to rule out compensatory changes due to re-allocation of resources between flowers. A recent meta-analysis of studies in 306 species across 80 plant families found that fruit produetion is limited by receipt of low levels of pollen in the majority of cases (Knight et al., 2005), contrary to the original prediction from sexual selection theory. Furthermore, thal habitat l as lower reproductive success (Aguilar et al., 2006). populations have recently suffered — fragmentation have lower levels of pollination, as we What we don’t know in these cases is how critical seed production is to the growth of the population, the second requirement for a demographic change. A population that increases its seed production may nevertheless fail to thrive if a density-dependent process, such as competition of seeds for a limited number of suitable germination sites, negates the higher seed production. Several studies have com- bined the extent of pollen limitation with population 2005) attempt to look at the demographic consequences; projection matrices (Knight et al.. in a first however, in all of these cases, absence of density- dependence was assumed. EXPERIMENTAL TESTS FOR DEMOGRAPHIC EFFECTS IN /POMOPSIS My colleagues and I have begun to examine the possibility of demographic effects of low pollination in Price et al., 2008). In populations of Z. aggregata near Gothic, Ipomopsis aggregata and Í. tenuituba S Colorado, hummingbirds are responsible for aboul Volume 95, Number 2 2008 Campbell Effects of Pollinator Shifts on Plant Populations Price el 94% of all flower visits made by pollinators al., 2005). These populations are often pollen limited. with supplemental hand pollination of all flowers on a plant approximately doubling total seed production 1991; 1993). simulated this increase experimentally by adding the (Campbell, Campbell & Halama, estimated number of seeds produced by five naturally = pollinated plants or five plants with supplemental and pollination (750 or 1500 seeds) to areas of 8 m”. Seeds were added again the next year to simulate persistent pollen limitation, and each treatment was Doubling the seed input t replicated six times. simulate full pollination roughly doubled the number of seedlings that emerged in each 8-m* plot (means = 132 vs. 252 seedlings) and the number of individuals that survived to flower (7 vs. 17 flowering plants; Price et al., 2008). Over this range of densities correspond- ing to natural versus full pollination, per-capita success was independent of density even though density dependence was evident over a wider 10-fold range of seed inputs. To estimate the demographic effect of full pollina- A) for a population using the observed values of age-specific tion, we calculated the finite rate of increase survival and age at flowering, while estimating age- specific. fecundity as the observed surviving individuals that reproduced, multiplied by the average number of seeds made by either fully or naturally pollinated plants. A population would have — ].14 (positive growth) under a scenario of full pollination versus A = 0.94 (negative growth) under a scenario of continued natural pollination (Price et al., 2008). In this case, low seed production due to low pollen transfer does have a demographic impact, and natural populations may currently be in some peril from low pollination. POTENTIAL DEMOGRAPHIC EFFECTS OF POLLINATOR SHIFTS Applying this approach to /pomopsis tenuituba allows us to consider the potential demographic consequences of a shift from hummingbird to hawkmoth pollination. The reciprocal transplant studies at Poverty Gulch discussed above have indicated that /. tenuituba may be in decline even in its native habitat. In 1994, we planted seeds from 10 full-sib families into a natural population of this The 197 produced five plants that survived to reproduce (or species. seeds we planted eventually that were still alive at the time our estimate was made in 2005), themselves (details given in Campbell & Waser, and produced an estimated 172 seeds 2007), giving a pooled net reproductive rate of R, = 0.87 and a pooled finite rate of increase of A = 0.98. These values for R, and A are based on small samples, proportion of but if taken at face value would lead /. tenuituba to be considered Vulnerable (VU) according to the IUCN List (IUCN, 2001). In addition, obtained in an experiment where seeds were planted Red they were just under the soil surface, which increases germina- tion relative to seeds that are simply scattered naturally (M. Price, N. M Campbell, A. Brody, unpublished data). would occur Waser, D. In part, the low success of /pomopsis tenuituba can be attributed to low rates of hawkmoth visitation. Demographic studies indicate that the finite rate of increase for /pomopsis is elastic lo increases in fecundity (Campbell € Waser, 2007); increase or decrease in fecundity during the first big percentage vear of flowering (age five or six) can have a substantial effect on the finite rate of increase. Indirect evidence that the level of hawkmoth pollination influences fecundity comes from compar- — ison of fruit sets in years with and without substantia numbers of hawkmoths. In the time span that we have studied this abundant in 1993. Fruit set (% flowers that set fruit) that compared with 25% in 199: hvbrid zone, hawkmoths were most tenuituba populations was 90%, and 26% in 2002, both years without hawkmoths (Campbell et al., 1997, in 1992, fashion from the high-elevation /. year in /. unpublished. data). Furthermore, fruit. set increased in clinal tenuituba populations to the lower-elevation 7. ag- gregata populations where hummingbirds were more abundant, while no such eline was observed in 1993 (Campbell et al., 1997). These data suggest that high fruit set is obtained only in years when hawkmoths are present. During the past 20 years, my colleagues and I have — observed hawkmoths in substantial numbers in the surrounding area in only two years. It is possible that hawkmoth frequency was historically higher (Grant & Grant, 1965). higher The current rarity of hawkmoths in the mountains might conceivably be a recent anthropogenic change, perhaps from widespread use of insecticides in the desert areas, which harbor larger populations of these animals, and/or from other causes, although there are no data on population trends of moth species in the United States (National Research Council, 2007). Regardless of the reason for low abundance, it is useful to ask: would a higher abundance of hawkmoths maintain viable populations This addressed by using a demographic transition matrix of Ipomopsis tenuituba? question can be with elements corresponding to age-specific survival and fecundity, while also considering what happens that corresponds to frequency of years with hawkmoths. when fecundity is multiplied by a factor Assuming that fruit and seed set are linearly related to the frequency of years with hawkmoths, and fruit set is Annals of the Missouri Botanical Garden k times as high when hawkmoths are present: R, — (1 — m)B + kmB where m = frequency of years with moths and B = R, when only birds and no hawkmoths are present. In our case, k = 0.90/0.25 = 3.6. Noting that one out of LO years in which plants bloomed in our experiment had hawkmoths (m = 0.1), and in that Mp y — 0.87 seeds per seed planted, solving gives B = 0.69. Thus. we would expect R, = 0.69 if E were never present and R, = 2.49 if hawkmoths were always present. lt would take a minimum of two years out of l0 with hawkmoths to raise R, above |. This simplistic calculation ignores evolutionary change (see below) but suggests that these /. tenuituba populations are in some danger unless visitation by hawkmoths and/or hummingbirds increases. It is worth noting that this species combines a number of traits that could make it particularly vulnerable: self- incompatibility, a monocarpic life cycle with no seed bank, and some dependence on hawkmoths that are North ` butterflies and moths are extremely variable from year to year. In America, many species ol endangered in various 2007). While effects of such low pollinator numbers on plant regarded as threatened or regions (National Research Council, populations are rarely studied, one exceptionally clear-cut case of endangerment due to low hawkmoth pollination is provided by the tree Oxyanthus pyriformis (Hochst.) Skeels in South Africa (Johnson et al., 2004). We need many more such studies that combine pollination and demography in order to assess the severity of what some have referred to as a pollination 1998). It is perhaps easier to list what we do know about the problem than what we crisis (Kearns et al., don't know. However, from the plant side it is clear that little is known about demographic effects and about what types of plants will suffer the most. It is important to identify such situations in order to know where to focus conservation efforts, and this will likely be an area of much future work. Aguilar et al. (2006) found greater effect of habitat fragmentation. on pollination for. self-incompatible species, but the importance of most other traits has yet to. be demonstrated. AN INTEGRATIVE EVOLUTIONARY PERSPECTIVE ON THE CONSERVATION OF PLANT-POLLINATOR INTERACTIONS PUTTING EVOLUTION INTO IMPACT ASSESSMENT To date, pollinator have ignored the influence of evolutionary change on demographic processes. A few studies have considered the possibility that reduction in pollinators most assessments of the impact of loss of would alter selfing frequency and inbreeding depres- 2000): changes sion (e.g., Lazaro & Traveset, however, most other genetic or evolutionary have been ~- ignored. It is now abundantly clear that some natural populations experience selection strong enough (End- ler, 1986) to produce evolutionary. change on short time scales (Reznick € Ghalambor, 2001). Adaptation of populations to new selective regimes could, i principle, happen quickly enough to prevent extinc- and the potential for this process has been (Gomulkiewicz & Holt, and the tion, demonstrated theoretically 1995). Still, potential for such evolutionary empirical data are lacking, changes has hardly been considered in any particular case, especially outside of the case of global climate change (Rice & Emery, 2003). We need to consider the possibility of such evolutionary change in the context of loss of pollinators. AN ILLUSTRATION BASED ON /POMOPSIS TENUITUBA In this section, | will outline how we might consider n response Lo pollinator the possibility of evolution loss using the case of /pomopsis tenuituba. Large eo variation in fitness among the 10 full-sib families we planted suggests a high opportunity for selection on traits that influence reproduction (Campbell & Waser, 2007). If hawkmoths disappeared entirely from these populations, could these plants evolve characters making them more attractive to hummingbirds quickly enough that the risk of extinction would be amelio- rated? Selection on traits such as corolla width is strong, raising the possibility that an increase in corolla width might increase hummingbird visitation sufficiently to raise reproduction above the replace- ment rate. Of course, this scenario would require that hummingbird abundances not decline below their present level, and there are some suggestions of recent declines in the hummingbird pollinators as well (Nabhan, 2004). evolutionary change in traits to increase hummingbird Even if it were possible for an visitation, this would lead to greater similarity between the two species, and, thus diversity in £ phenotypic sense would not necessarily be main- Most of question are available, but sample sizes are, in some tained. the data necessary to model the cases, rather small, so that the quantitative answers will be extremely rough. This example should be viewed mainly as a guide to the approach. The first step is to estimate how fast a character. such as corolla width, would change evolutionarily due to selection. According to a simple quantitative width between generations ean be predicted by the strength genetic framework, the change in corolla of phenotypic selection on the trait (measured by the Volume 95, Number 2 2008 ampbell 271 Effects of Pollinator Shifts on Plant Populations selection differential S) multiplied by heritability of the trait. A more precise prediction can be made by taking into account genetic correlations. between characters (Lande & Arnold, 1983), but for simplicity, those correlations will be ignored here. This proce- dure is equivalent to assuming that the selection by pollinators is only on corolla width and no other Campbell (1996) measured the heritability of corolla width genetically correlated characters. natural conditions in Ipomopsis aggregata as 0.30, 30% explained by the average eff: indicating that the variation could be — ects of alleles passed on by parents. Combining that figure with the phenotypic selection exerted by hummingbird visitation when presented with arrays of both /pomopsis species and their hybrids (selection differential = 0.064; Camp- bell et al., 0.02 mm per generation, about 1/50 of the difference in this trait between the two species. This calculation assumes heritability would be identical in 7. tenuituba (a simplistic assumption; see Roff & Mosseau, 1999) and ignores any selection arising from greater efficiency of hummingbirds at dispersing pollen from 1998). Furthermore, i assumes directional rather than stabilizing selection wider flowers (Campbell et al., on the character, in contrast with Gomulkiewicz and — Holt (1995), who envisioned a shift from one optima joe value to another under the new selective regime. Incorporating stabilizing selection instead can also be done using simple quantitative genetics models, but that approach was rejected here because the observed phenotypic selection was primarily directional (Camp- bell et al., 19 The second step is to allow fecundity to increase in the next generation to the extent predicted by how the new phenotype (in this case a corolla that is 0.02 mm wider) would increase pollinator visitation. These data are available from studies of hummingbirds visiting E lowers in experimental arrays of the two species and hybrids (Campbell et al., 1997). The net reproductive rate would increase between generations as follows: R, = Bl(v + bz)/v] = dl + =, where B = net reproductive rate in the original generation following loss of hawkmoths, v = current hummingbird visit rate for /pomopsis tenuituba, b = slope representing the increase in visit rate per change in corolla width, z = the change in corolla width per generation, and ¢ = time in generations. The part in square brackets expresses the proportional increase in visit rate relative to the original rate by that strength. of selection and the response to selection hummingbirds. Notice what matters is the (z) relative to the original visit rate. Campbell et al. under 1997) predicts an evolutionary change of do X 20 g 18 A Pd uu we = AB P" o 2 14 pt 5 A = a2 Pd © P d a Z m BLZ dh $ 0.6 - ] > 0 10 20 30 40 50 N/No Generations Figure 1. rate (R,) and (B) pop on size (N) value (No), fecundity due to response to selection on a floral trait. Strength of selection (b) = 0.18 (solid lines) or b = 0.33 (dashed lines). Estimated changes in the (A) net reproductive relative to its initial according to a model that allows for increased — un) (1997) found an increase of 0.18 visits per millimeter flower per hour for /. 1997: fig. 3). from a mean of 0.17 visits per tenuituba plants (Campbell et al., Substituting, 0.17 0.18(0.02) R, = 0.69|1 + —— ———t = 0.69|1 + 0.0211]. This approach assumes that fecundity is linearly related to visit rate. This assumption is unlikely to hold exactly, as seed production is expected to level 2004), but may be a reasonable first approximation in this system off at high pollen receipt (Ashman et al., since even /. aggregata, with its much higher visit rate, suffers pollen limitation. The final step is to iterate the population size using a this dynamic model for R,. According to this simple model, it would take 22 generations for corolla width (198 years for this plant with its generation time of nine years) to increase sufficiently in response to selection for the net reproductive rate to rise above 1 (Fig. 1A). Using this change in R,, in the meantime, the population size would decline to 2% of its initial value (Fig. 1B, solid Even more line). using a 272 Annals of the Missouri Botanical Garden optimistic. value for how visit rate would increase (b = 0.33 based on a measure of selection taking into account correlated effects of other characters; Camp- bell el al., 1997), the population size would decline to 11% of its initial value before starting to recover after 13 generations (Fig. 1). Given that the current size of the population is probably in the hundreds, nol thousands, these caleulations suggest the population would likely be vulnerable to extinction through demographic or environmental stochasticity before it could evolve a sufficient increase in pollination level, despite the strong selection on corolla width. All of the above considerations ignore the potential effects of hybridization. Hybridization can impact as population in one of two ways. [f hybrids are relatively unfit, hybridization can lower the net reproductive rate of the population. Second, even if hybrids are fit, they may show signs of introgression, changing the population through genetic assimilation, as has occurred for example in Helianthus L. (Carney et al., 2000) and Spartina Schreb. (Anttila et al., 2000). the case of /pomopsis hybridization, hybrids with tenuituba as the mother and /. aggregata as the father had zero fitness in the home site for Z. tenuituba. sí likely decline of the population rather hybridization would contribute further t than to genetic assimilation. | stress that these predictions concerning the effects taken at face value of loss of hawkmoths, even if despite the uncertainty in estimates of parameters, would apply only to the potential of local extinction from the region near Poverty Gulch, not global extinction of the species. Pollinator Aider on vary greatly in space, and hawkmoth numbers, in partic- ular, seem to be highly variable. For MEE. in the columbine Aquilegia caerulea James, hawkmoths accounted for 68% of flower visitation in an Arizona population, while being absent entirely from another population in Colorado near our study site (Brunet & Sweet, 2006) tenuituba studied at Grizzly Ridge, hawkmoth visita- In other populations of /pomopsis tion was 50% higher than at Poverty Gulch, which might be sufficient to maintain viable populations (Aldridge & Campbell, 2007). CONCLUSIONS AND FUTURE DIRECTIONS Since the publication of important conceptual papers by Raven in the 1970s, pollination ecologists have made much progress in understanding the evolution of floral diversity. In a number of cases. the influence of pollinators on microevolution of floral traits has been documented, and even quantified, through estimates of selection and heritabilities (Galen, 1996; Ashman & Majetic, 2006). Much recent been directed at the and floral and progress in this area will undou attention has origin maintenance of differences between species, tedly continue as new genetic tools and methods of analysis allow a finer look at the genetic basis of the traits (Bradshaw et al., & Conner, 1998) and their fitness consequences (Morgan 2001). We understanding how changes in pollination regimes can have come a long way in lead to diversification and origin of species. Much less, however, is known about how recent changes in pollination regimes impact the FU a of plant populations and might thereby lead to loss of plant species. Given the mounting evidence for Tecline of some pollinators, there is urgent need to design experiments that evaluate whether lower pollination will result in smaller populations of wild plants. particularly for those species that are already threatened. Such studies should consider not only direct demographic: effects of lower pollination and seed production, but also the possibility of indirect demographic effects that occur through rapid evolu- tionary changes in levels of trait values or in inbreeding. It is time to heed Raven's (2002) call for action and bring the growing knowledge of ecology and evolution of pollination to bear on evaluating the impact of anthropogenic changes in populations of animal pollinators. Literature Cited . Galetto & M. A. 2006. Plant ne live susce ED to habitat fragmentation: Aguilar, R., L. Ashwor Aizen. Review and synthesis through a mela- analysis. Ecol. Letters 9: 968 ISO Aldridge. G. € D. R. Campbell. 2007. eran in pollinator | | preference between two ide conta p cd 110. . King. C. e D. R. Ayres & D. R. 00. ip al hybrid E ar Spartina in s that vary in hybridization. Ba: Molec. Ecol. 765-770. > 1997. Press, Natural bo and Evolution. Oxford Univ. Ashman, T. L. & C. floral Heredity Oxford. J. en 2006. Genetic constraints on icu A review and evaluation of patte 3: . A. Steets, P. ai a Dudash, M. ( gage S, Mitchell, M. T. Morgan ^ W. G. Wilson. 2004. Pollen limitation of plant reproduction: j: PM al M. Burd, and evolutionary « causes and consequences., Ke do ogy 85: Best, L. S. & floxglove (Digitalis purpurea): A 70-79 P. Bierzychudek. 1982. Pollinator foraging on ‘st of model. of a new Evolution 36: 70— Biesmeijer, J. x. S. P. M. 2 muller, Edwards. T. Peeters. A. P. Schaffers, S. G. Potts, R. ee C. D Settele & W. E. Kunin. 2006. Parallel declines in pollinators and insect- Roberts. M. Reemer, R. Thomas. J. pollinated plants in Britain and the Netherlands. Science 354 313: 351- Volume 95, Number 2 2008 Campbe 27 Effects of Pollinator Shifts on Plant Populations Bolten, A. B. . Feinsinger. 1978. Why bird flowers secrete dilute nectar? Biotropica 10: 309. Bradshaw, H. D. Jr., K. I do pro 307 G. Otto, B. E. Frewen, J. K. 1998. Quantitative differences in floral morphology between two species of e (Mimulus). Genetics 149: 367-382 . R. Sweet. 2006. Impact of insect po dus m m display size on outcrossing rate. Evolution McKay & V. Schemske. trait loci affecting E Brunet, J. & I linator Burd, " 1994. ‘Roca s principle and plant reproduction: The role of pollen limitation in fruit and seed set. Bot. Rev 60: 83-1: Campbell, D. R. 1991. one nts of a in ial aggregata. Amer. Midl. Naturalist 137: 713-7 T E nde of f floral trails in a he dile ied Effects of floral traits on sequential plant: Fie Jd measurements 0 ` heritabilities and genetic correlations. 2-1453. NE 2 N Evolution 50: 1 1993. limitations to lifetime seed production in a natural plant Halama. Resource and pollen — ie IN. poy pulanon. Ecology 74: 1043-1051. N. M 2001. fitness of plant hybrids in the wild. = Waser. Genotype-by-environment interaction and the dio 55: 669-676 Alc a 2006. F loral 3 of hybrid zones. Pp. e in L. larder & S. C. H. Barrett (editors), Ecology and faa of Flowers. Oxford Univ. Press, Oxford. & N. M. Waser. 2007. Ipomopsis hybrid zone: € Mp ion fitne: Midl. Evolutionary dynamics of an s with lifetime ss data. ARE ris 19: 298-310. —— — lelé im z-Ackerman. 1997, odio m -me EN A selection in a plant hybrid zone: LE visitation pure on three spatial scales. Amer fps st 149 315. E ollen transfer by & P. G. Wolf. 1998. atual hybrids and pare ae species in an /pompsis hybrid 02— l. . Evolution 52: a 161 & G. Ye derson. 2002. Predicting patterns of Son "m: br hybridization from pollinator aa ipu Naturalist 159: 438—450. salen & C. A. Wu. 2005. Ecophysiology of first and d generation hybrids in a natural plant hybrid Decologia 144: 214-225 Carney, S. E., K. A. Gardner & L. H. be Bae, ior. zone. 2000. Ri eseberg. r history of a hybrid anges over the fifty E ¿volutionary ch: population of sunflowers (Helianthus). E volution 54: 462— 474. Charnov, E. L. 1979. one "ous he rmaj Suri and se Res selection. 2484 Proc. Natl. ad. U.S.A. 76: 2480— Chase, V. C. € P. H. 1975. Evolutionary anc ecological relationships between Aquilegia formosa and A. pubescens anneal eae), two perennial plants. Evolu- tion 29: 474— a E & 3 Waser. w Why ; red flowers are nol sible to bees. Israel J. . 45: 169-183 E hrlie d. P. R. & P. H. Raven. i B utte a "s and plants: A study in coevolution. Evolution 18: 58 Raven. & ————. 1909. Ae ae 2 opinas Science 165: 1228-1232. Endler, J. A. 1986. Natural Selection in the Wild. Princeton Univ. Press, A00 Pergamon Press, Toro: Faeg K 4 van The e of Pollination E sole Fulton, M. € S. A. Hodges. 1999. Floral isolation between Aquilegia formosa and aa pep 'oc. Roy. Soc London, Ser. B, Biol. Sci. 266: 2247-225 2 Galen, C. 1996. bumblebee pollination in an pu wildflower, Polemoni- pur : 120-125. Gomulkiewicz, R. € R. D. Holt. y natural selection peer extinction? Evolution 49: 201-207. Grant, V. in angiospe rms. Rates of floral evolution: Adaptation to S um viscosusm. 1995. When does evolution 1949. Pollination systems as s isolating mechanisms ver 3: 82 . 1965. dst Pollination in the C M Univ. Press, rk. D. H. Wilken. a (s ral between pomopsis aggregate e I. tenuituba (Polem- G az dus 213-22 Phlox Family. hybridization Hd E o Cc ei ^ lI . Wilson. pd M. evolution and male reproductive success—Optim ules for pollen dispersal by animal- MS :d plants. Evol. Ecol. 8: 542-559 & D. giulaii 1999. Ecological speciation in ebacks: Environment-dependent hybrid fitness. Evo- lution 53: 866-873. p. Heinrich, . & q al dispensing sche Hatfield, T. Raven. 1972. Energetics and pollination ecology. Science 176: 597—602. IB. d M. Fulton & J. Y. Yang. 2002. its influencing reproductive isolation Hodges, S. A Genetics of floral t between Aquilegia, formosa and Aquilegia pubescens. Ame Midl. Naturalist 159: 851—560. Ippolito, A., G. W. P & Holtsford. 2004. icollana alata, N. forgetiana, 534—2644. I lllinatót preferences for | and their F-1 hybrids. Evolution 58: IUCN. 2001. IUCN Red List C ategories and Criteria Version 3.1. Prepared by the IUCN S J we ipa uua . Neal, C. I. Peter & T. : Edwards. 2004. Fruiting e and a recruitment in remnant pala dns of ie hawkmoth- Sende tree Oxyanthus Biol. cies Survival Commission. => juo S. D. Conserva- pyriformis subsp. pyriformis (Rubiaceae tion 120: 31—39 Kearns, C. A., D. W. Inouye Endangered The pollinator interactions. Annual Rev. Ecol. Syst. 112 aight. T. J. A. Steets, J. C. V , S. J. Mazer, M. Burd, D. 3 Camel M. R. Dudash, m 0. Johnston R. J. Mitchell & T. L. Ashman. 2005. Pollen limitation of plant reproduction: is and process. Annual Rev. Ecol. Syst. 36: 467-49 Lamont, B. B., P. G. L. Klinkhamer & E. T. 1993. Population fragmentation may reduce zero in Banksia goodie—A demonstration of the Allee effect. Oecologia 94: 446—450. Lande, R. & S. J. Arnold. 1983. on comelaren characters. E e 37: & N. Waser. conservation of 1998. mutualisms: plant- 8 F. Witkowski. fertility to The a of selection 210-1226. » Re ane tive success of the Pollen E . Traveset. 200 endan ur shrub Buxus Disco Dues dene ak painot, and o anc : | “vol. 261 a —128. ers Mack, J. ¢ G. Milligan. 1998. Pollination in hybridizing a (Fabaceae) populations. Amer. Bot. 85: 500-507 Levin, D. A. & H. ants. Evol, b de. 5 E 2. significance of an Ipomopsis hybrid zone. Evolution 52: 12 biology J ERU 1974. Gene flow in seed 9-220. R. Campbell. 1998. Adaptive flower color and inter-trait corre d deus it 93-1303. 274 Annals of the Missouri Botanical Garden Morgan, M. T. & J. K. Conner. 2001. Using genetic markers to directly estimate male selection gradients. Evolution 55: 72-5 Nabhan, G. P. 2004. Conserving Migratory Pollinators and " Corridors in Western North a Press, America. Univ. of pod il. 2007. Status of Pollinators in North America. National Academies Press, Washington, D.C. Price, M. V., N. M. Waser, R. E. Irwin, D. R. Campbell & A. K. Brody. 2005 x Temporal and spatial variation in pollination of a montane herb: A seven-year da jc d 86: p B ). R. Campbell, N. M. Waser & A. K. Brody. pede ging the generation gap National R in bas sien eee fec pu and M demography. Ecology (in press). Pyke, G. H. 1984. al foraging theory: Á critical review. ns Eh Ecol pum 5: 523—575. P. H. . Why are bird-visited flowers predomi- nantly red? ene 26: 674. 200: - Raven, . Science, sustaimability, and the human ‘OS T Science 297: 954—958 Re: aL slo v. SM Biology. Academic Press, Orlando. panic GC K 2001. The population ecology eu temporary adaptatic Wha Ghalambor. empirical studies BUE about d 1e conditions thai promote ip ee evolution. pus tica 83-198. > Emery 2003. Managing microevolution: Restoration in mu face of global change. Frontiers Ecol. 1: 78. Rice, K. J. & 469— Roff, D. A. alter Ail tic een ture . À. Mousseau. 1999. Does natural selection An evaluation of quantitative genetic variation among —— of Allonemobius socius and E J. Evol. Biol. 12: 361-369 Roubik, D. W. 1995. Tropics. Food I Pollination of Cultivated Plants in the Agriculture Organization, United Nations, ome. Spaethe, J., J. Chittka. 2001. Tautz € L. Visual constraints in foraging bumblebees: Flower size and color affect seare a time and flight behavior. Proc. Natl. Acad. Sei. U.S.A. 9 3898-3903 Stanton, M. L., A. Snow € 5. N. Handel. 1986. Floral Aura liveness to oe increases male Science 232: 1625-16 Wang K. D. 1983. o behavior of pollinators. 213-239 in L. Real (editor), po ade mic Press, Orlando. Waser, . & M. V. Price. 1983. Pollinator behaviour and i M ‘tion for flower colour in Delphinium nelsonii. Nature 3 : : 422—424. & Pollination Biology. ME N . R. Campbell. 2004. Ec ologic e speciation in flowering n eckmann, M. Doebeli, J. A. J. Tautz hee . Press, Cambridge. per ^p. 264-277 etz & D. Cambridge Univ Adaptive Spec ation. Willson, Naturalist 113: 777— Wolf, P. G., P. S. Soltis & p. . Soltis. 1993. Phylogenetic loma of chloroplast pu restriction site variation in the Ipomopsis aggregata complex and related. species p se le ection in plants. Amer. Midl. (Polemoniac xod Syst. Bot. 18: 652-602. Wu, C. A. & D. R. Campbell. 2005. C LK gee and nuclear markers reveal contrasting. patterns of spatial genetic structure in a natural /pomopsis hybrid zone. "Mole c. Ecol. 14: 781-76 Wyatt, R. 1982. flower number, arrangement, and phenology affect pollination Amer. J. Bot. 69: 585-594. Inflorescence a rehitect ure: I low and fruit-set. A RENAISSANCE OF CYTOGENETICS: STUDIES IN POLYPLOIDY AND CHROMOSOMAL EVOLUTION! J. Chris Pires? and Kate L. Hertweck? ABSTRACT Systematics and ev d ions are being bolstered by a renaissance in cytogenetics and comparative genomics as illustrated by reviewing Pete three key research of chromosome numbers in the Annonales and revealed paleopolyploid events, which appear Raven's characterization of chromosome evolutio n's integration of cytogenetics an areas. The first area is the evolution of chromosomes during the origin of the other basal A inspired modern genomic comparisons that r to have occurred early an in various genera of ros id phylogenetics and by presenting apa ites to his work i s. Raven's Nice have t diversification. Second, it of a angl Osp er ften during flowering plant 9s, agraceae is updated in lig contemporary Onagraceae ME The possible construction of ancestral karyotypes in the Onagraceae is feasible using tec ae ies that have been successful in analyses of genomic blocks in the Poaceae and Brassic caceae. Third, Raven’s work on catastrophic Ane identified the importance of chromosomal rearrangements in the evolution of Clarkia Pursh and the abil ity of new species to inhabit different environments. Current work in Brassica L. speciation events can arise from relatively few chromosomal rearrangements. Á fusion of sys opening new a evel characters into phylogenetic analyses li Key words: Brassica, Brassicaceae, chrom phylogenomics, Poaceae, polyploidy. has shown that phe netypic changes contributing to ematics and cytogenetics 1s ireas of research, with phylogenomics allowing ancestral genome reconstruction, mo. incorporation of genome- s, and new theories about evolution on a genomic scale. Clarkia, molecular cytogenetics, Onagraceae, phylogenetics, In 1986, conceptual difficulty with the field of biosystematics Peter Raven commented, “a genuine is that it was always conceived as a way of improving plant classification, rather than a way of learning more about plant evolution" (Raven, 1986: 26). Indeed, the history of biosystematics itself is problematic. The origin of phylogenetics in the mid-20th century caused drastic changes in how systematics is viewed philosophically. Before the advent of phylogenetics in the mid-20th century, biosystematists had concentrat- ed on closely related species and within-species variation, with methodologically and theoretically flexible views of breeding and chromosomes, viewing evolutionary relationships as webs. The arrival of tree thinking switched the focus of systematics primarily toward resolving phylogenetic relationships of more distantly related species and higher taxa, with a more monistic interpretation of traits and homology (Sytsma & Pires, 2001). In 2006, Peter Raven remarked personal communication, “the two biggest advances in my lifetime were (1) the conceptual distinction between synapomorphy and pleisiomorphy and (2) the development of technology to allow us to look at then flavonoids, then and chromosomes, isozymes, + inally DNA.” The most important advancements over ~ he past several decades were improved methods of building trees and visualizing phylogenetic relation- ships. While the importance of phylogenetics need not be questioned, we can ask whether its growth comes at the expense of other disciplines, such as evolutionary development and cytogenetics. The past few decades field can emerge, have shown how a such as evolutionary development (evo-devo) suggesting the potential for resurgence of other fields (Love, 2003; Breidbach & Ghiselin, 2007; Muller, 2007; Pighucci, 007). Here, cytogenetics and comparative genomics is occurring i) we argue that a similar renaissance in 'This paper was presented at the 53rd Annual Systematics va of the Missouri Botanical Garden, Peter Raven on Evolutionary and Biodiversity s in the 20th , ting. JCP ac el s funding E the and KLH is supported by a Univ Sel s Baum, Eric Schranz, Martin Lysak, and i iie Me hell-Olds for discu 'ussion We Peter Hoch and Barbara Schaal, for a elle ant Genome Program (DBI 0501 ae aa DBI a Graduate Fellowship. We thank Jon Lamb, Davi on Brassicaceae, and members of the Pires Lab and would like to thank Peter Ken c OBS neticisls ev oe Raven, Sytsma, and other where “The Inpact a of * We thank the symposium orga Mese Science Foundation of Missou 21st Centuries. anizer ri Life external reviewers for their thoughtful comments on the manuscript. members of our academic ancestry, and, finally, classic ? Division of Biologic als Sc iences, 371 B Life Science Center, 1201 Rollins Road, University of Missouri, Columbia, Missouri 69211-7310, U.S 10. 3417/20071 76 dol: ANN. Missouni Bor. GARD. 95: 275-281. . Author for correspondence: piresjc€missouri.edu. PUBLISHED ON 18 JUNE 2008. Annals of the Missouri Botanical Garden with the advent of modern cytogenetic techniques and methods for analyzing genomes. We illustrate this by Peter and phylogenetics and by presenting reviewing Raven’s integration of cytogenetics updates to his pivotal work in three key research areas. First, how can the chromosome numbers of basal angiosperms contribute any information to chromosome evolution pad across angiosperms? Raven's investigation of chromo- some numbers in the Annonales supports subsequent conclusions about the basal chromosome number of angiosperms, and these analyses inspired genomic comparisons that revealed paleopolyploid events. how have chromosomes evolved in Onagra- the Second, ceae? Raver's characterization of types of chromosomes found in various genera of Onagraceae reveals interesting patterns of chromosomal evolution when mapped onto the contemporary Onagraceae phylogeny. Recent advances in similar work allow the construction of ancestral karyotypes in grasses and analyses of genomic blocks in Brassicaceae. Finally, how have chromosomal rearrangements led to rapid Clarkia Work catastrophic speciation, which occurs when a minority speciation in the genus Pursh? n cytotype becomes adaptively advantageous following dramatic habitat change, allowed Raven to identify the importance of chromosomal rearrangements in Clarkia speciation, and new examples from the same genus highlight the abilities of these species to inhabit Brassica L. different environments. Current work has consequently shown that relatively few chromo- somal rearrangements can lead to phenotypic changes that may contribute to speciation events. ANGIOSPERMS BASAL CHROMOSOME NUMBER IN THE — From a phylogenetic viewpoint, the term basal is used to denote organisms that are placed at or near the a phylogenetic tree, and is always used in When discussing deep phylogenies of angiosperms, authors base of reference to a certain group of organisms. often refer to groups of organisms whose branches ase of the tree, such as Amborella — originate near the Baill. and the Magnoliales, as basal angiosperms. In fact, these plants are merely descendants of the true ancestral angiosperms, which are presumably now extinct (Crisp & Cook, 2005). Indeed, all extant flowering plant lineages are equally distant Amborella and from a common ancestor. For the purposes of this paper and in concordance with the literature being reviewed, a lineages of plants that diverged from a common angiosperm ancestor prior to the monocot- eudicot split will be referred to as basal angiosperms, are with the understanding that collectively they merely extant approximations of the true ancestors of flowering plants. One of Raven's earliest contributions to cytology involved the basal chromosome number in the angiosperms (Raven & Kyhos, 1965), and these conclusions have been updated by more recent 2003). The advent of genome sequencing has consequently chromosome counts (Soltis et al., allowed comparisons between species for more thorough analyses of chromosomal evolution. It was Raven's thought that chromosome counts from the Annonales, or the woody Ranales, might shed light on angiosperm base chromosome number. His 1965 paper published the chromosome counts for Degeneria |. W. Bailey € A. C. Sm. (n = 12) and Drimys J. R. Forst. & G. Forst. (43 bivalents). (1965: 247) to state, appears very likely (1956) is the angiosperms.” These counts led Raven and Kyhos = “with our present information, that n = 7, as suggested by original base chromosome This low base Darlington number of conclusion seemed fairly reasonable; a relatively chromosome because it 247). and number is selectively advantageous reduces recombination (Raven € Kyhos, 1965: However, haven cautioned that aneuploidy polyploidy confound the issue, and phylogenetic considerations should be applied to interpret the base chromosome number of angiosperms. Moreover, “there are still six families of annonalean affinities i: which not a single chromosome count is available" (Raven & 1965: 247) chromosome counting did not stop with this paper, Kyhos, . Raven's interest in comprehensive however; he was instrumental to the founding of the Index to Plant Chromosome Numbers (IPCN Series), which originated at the University of California, Berkeley, with Marion Cave’s efforts in 1956, but moved to the Missouri Botanical Garden in 1978. Subsequent research in major patterns of cytology and phylogeny in the angiosperms has confirmed the relevance of Raven’s conclusions regarding basal chromosome numbers in angiosperms. Major devel- opments facilitating current advancement include more accurate phylogeny of angiosperms and more chromosome counts from taxa of early diverging lineages. Subsequent researchers have continued 1 — ill in gaps in chromosome numbers to provide a more complete picture of chromosome evolution in angio- 1968; Walker, 1972; Additionally, are sperms (Ehrendorfer et al., Ehrendorfer € Lambrou, 2000). gaining a better understanding of relationships among we the basal current phylogenies place Amborella and Nuphar Sm. as sister groups to the rest 1 | angiosperms; of the flowering plants. Paradoxically, many basal angiosperms are now known to have high haploid chromosome numbers (1.e., Amborella with n = 13 [Oginuma et al., 2000], Mlicium L. with n = 11 [Fedorov, 1969], Magnoliaceae with n = 19 [Fedoroy, 1969]. Despite these observations, current thought Volume 95, Number 2 2008 Pires & Hertweck Polyploidy and Chromosomal Evolution -—] indicates the original base chromosome number is low (x = 6-9) (Raven, 1975; Soltis et al., 2003, 2005). Current methods of characterizing genome evolu- tion go beyond mapping chromosomes onto phyloge- nies by exploring the prevalence of polyploidy and other genome duplications in angiosperms. A variety of research methods, including isozymes and phylo- genomics, suggest that genome duplications have arisen in many angiosperm families, with repeated cycles of polyploidization and diploidization. Classic estimates of genome evolution in polyploidy indicate 30%-50% of extant angiosperms are polyploids, and recent studies suggest that most extant angiosperms share at least one polyploid ancestor (Cui et al., 2006). =— Isozymes first detected duplicated loci in Magnolia- ceae and other basal lineages, providing the first suggestion of paleopolyploid events; isozyme evidence for duplications in eudicots confirmed earlier hypoth- eses of polyploidy in angiosperms (Soltis. 2005). A variety of inventive methods of genome analysis have been applied to detect the history of whole- genome duplications in angiosperms. Bowers (2003) fused phylogenetics and genomics to evaluate ancient genome duplications within Arabidopsis (DC.) Heynh. and found three nested polyploid events. The ages of duplicated genes suggest an alpha duplication event after the Brassicaceae-Malvaceae split, an older beta duplication event after the monocot-eudicot split, and perhaps a more ancient gamma duplication event that is difficult to Comparisons of sequenced place. genomes reveal duplicated genes, which provide evidence of ancient polyploidy in Arabidopsis, Oryza L., and Populus L. (Blanc et al., 2003; Paterson et al., 2004; Yu et al., 2005). Genome Project investigators examined basal angio- that In a parallel effort, Floral sperm lineages for older polyploidy events — predated the monocot-eudicot divergence. Ks analyses revealed independent ancient polyploidy events in several early diverging angiosperm lineages. Much slower rates of sequence evolution occur in basal angiosperms, especially Liriodendron L. (Magnolia- ceae), allowing analyses to determine that at least one ancient polyploidy event occurred in the common ancestor of all angiosperms or in the common ancestor of all angiosperms other than Amborella (Cui et al., 2006) Raven's early work counting chromosomes in the Annonales allowed him to make hypotheses about the base chromosome number for all angiosperms. This type of work laid the foundation for more thorough analyses of chromosome number with the discovery of other basal angiosperm lineages, which, in turn, have advanced more in-depth investigations of chromosom- al evolution and ancestral karyotype inferences via phylogenomic comparisons. CHROMOSOME EVOLUTION IN ONAGRACEAE [om Although Peter Raven's ideas about basal angio- sperm. chromosome number were insightful, he is perhaps more famous for his work on chromosome evolution within the Onagraceae, the evening prim- rose family. His initial characterization of Onagraceae chromosomes eventually led to hypotheses about the phylogenetic patterns in the family, and current work on ancestral karyotype reconstruction in the Poaceae and Brassicaceae reaffirms the importance of ad- dressing chromosomal organization in a phylogenetic framework. on milosis in the and 46 described An early collaborative work Onagraceae covered all tribes, 15 genera, species. This comparative work three modally distinct groups of chromosomes that vary in size and heterochromatic content. Members of the first group of chromosomes are presumably ancestral and medium sized, contract evenly along length during meiosis, and are not visible in interphase; they are found in Fuchsia L., Circaea L., and Lopezia Cav. The second group is intermediate in age; these chromo- somes are morphologically dot-like and heteropycnot- ic throughout meiosis, and are not visible in interphase; they are found in Ludwigia L., Epilobium L., and Boisduvalia Spach. Chromosomes in the third group are derived, medium sized, and heteropycnotic near the centromere, and have long, lightly staining JC. and euchromatic tails; they are found in Hauya DC. Onagreae. These groupings were used to support Fuchsia and Ludwigia as less-derived genera of the 1962). Raven's family (Kurabayashi et al., When combined with classification of Onagraceae chromosomes, the current. Onagraceae phylogeny (Levin et al., 2003) suggests subtle patterns The previously men- 1962) do not map simply onto the Onagraceae phylogeny: about chromosome evolution. — tioned chromosome types (Kurabayashi et al., small, dot-like chromosomes have evolved indepen- dently multiple times. Later karyomorphological studies in the Onagraceae described four types of 1988). There were two major differences between the old and new chromo- chromosomes (Tanaka et al., some groupings. In the new chromosome classification scheme, the genera of section Onagreae were split into two groups, and both the first and second groups in the ii This chromosome classification scheme suggests variation original classification were reorganized. new in the origin of chromosome morphology and indicates that cytology of genera reflects phylogenetic relation- ships. For example, the independent evolution. of small chromosomes in Onagraceae can be explained by more thorough analyses of chromosome structure. Although Epilobium and Ludwigia both have small 278 Annals of the Missouri Botanical Garden chromosomes, Epilobium chromosomes seem to be a specialization of more diffuse (type 1) chromosomes (P. Raven, Recent pers. comm.). comparative. genomic analyses of the Poaceae and Brassicaceae are illustrative of potential future practices in family-wide plant systematic studies. Comparisons among grass genomes indicate that there is colinearity among the grass genomes with rice as the central reference genome (Devos et. al., 2002). using synteny to find candidate genes in non-model The benefits of this comparative approach (e.g.. taxa) make it a desirable analytical tool to apply to other taxonomic groups with copious genetic informa- tion available. For example, is it possible to compare colinearity among crucifer genomes with Arabidopsis Although A. functional reference thaliana (L.) Heynh. as a reference? thaliana will continue to be the point, its small, derived genome (n = 5) complicates larger-scale structural genomic comparisons to other members of the Brassicaceae. Given that complete genome sequences for at least four crucifers (A. lyrata a) OF Kane € Al-Shehbaz [n = 8], Capsella Medik. [n £ 0]. and Thellungiella ©. E. Schulz [n = 7]) will be available within a few years. ya = 8|, Brassica rapa L. |n = itis unlikely that A. thaliana alone will continue to be used for structural comparisons. It may be possible for genetic maps and cytogenetic data from several Brassicaceae genomes to be combined to create a generalized reference set of ancestral genomic blocks. The A. thaliana genome can be split into at least 21 conserved blocks, and comparative analyses reveal A. thaliana and B. napus L. share large, colinear blocks that have undergone numerous rearrangements (Par- kin et al., isons allow extrapolations about the evolution of more 2005). These types of structural. compar- derived states. In fact, the karyotype of A. thaliana is derived from a tentative n = 8 ancestral karyotype: genetic maps of closely related species, A. lyrata and C. rubella Reut., A. thaliana (n = 5) is rearranged and derived (Lysak 2000). Brassica genomic blocks with other crucifer genomes the karyotype, which will better facilitate comparisons are collinear (n = 8) and show how et al., Investigating the A. thaliana and will allow for reconstruction of an ancestral N = across the Brassicaceae (Schranz et a Peter Raven’s analysis of chromosomal forms ii pao] Onagraceae provided invaluable information to re- searchers who could later analyze chromosomal evolution in the family. Current research in Poaceae and Brassicaceae has built on similar types of analyses by deseribing how genomie blocks can show conserved and derived patterns of chromosomal evolution across closely related species. These studies indicate a major role for cytogenetic data in phylogenomic comparisons. At the family level, analyses of chromosomal blocks and ancestral karyo- type reconstruction will most likely be developed first i Solanaceae, Poaceae, Brassicaceae, Leguminosae, and Asteraceae. However, as genomic resources and comparative analyses improve, Onagraceae, and any other family of choice, will not be far behind. One can imagine the possibility of generating ancestral genome browsers for deeper-nested nodes (e.g., Brassicales, rosids, eudicots, angiosperms) that would allow for a myriad of phylogenetic inferences and comparisons across clades of flowering plants. RAPID SPECIATION IN THE GENUS CLARKIA AND CHROMOSOMAL REARRANGEMENTS Peter Raven's research interests not only spanned comparisons of orders and families of flowering plants but cytogenetics within genera and species. Specifically, ytog g ) also involved integrating phylogenetics and Raven’s theories about catastrophic selection and pty edaphic endemism as a factor in speciation later informed phylogenies about Clarkia and modes of adaptation to habitats. Modern research incorporating cytogenetics and phenotype change in Brassica confirms the ability of species to adapt quickly to new environments and to become | reproductively isolated. Early Lewis (1962) theorized that the removal of most of a population would allow a deviant, work by or better adapted, genome to establish a new marginal population. This concept of catastrophic. speciation was later expounded upon by Raven (1964: 338) when he suggested that “catastrophic selection and conse- quent genolype reorganization may be a potent force m mg creation of new edaphically restricted endem- les.” He hypothesis in Clarkia rubicunda (Lindl.) F. H. noted a pattern possibly supporting this Lewis Lewis & P. H. & M. E. Lewis and C. franciscana F. H. Raven, but found further support in distinctive endemies like Microcycas (Miq.) ^. DC. (Mason, 1946) and narrow endemics like the Sireptanthus glandulosus Hook. complex (Kruckeberg. 1957). Recent phylogenetic and cytogenetic analyses of Clarkia reveal fascinating patterns of speciation and — iromosomal evolution. Although very closely related with little sequence divergence, Clarkia biloba A. Nelson € J. F. Macbr. (n = 8) and C. lingulata V. H. Lewis € M. E. Lewis (n = 9) are effectively reproductively isolated due to hybrid sterility; their differences. in base chromosome numbers result in oor chromosome pairing (Sytsma & Smith, 1992). In | B AM contrast, Clarkia sect. Rhodanthos subsect. Rho- danthos presents a more complicated example. Several researchers have utilized. different. types of data to determine the mode of speciation in this group. Volume 95, Number 2 2008 Pires & Hertweck Polyploidy and Chromosomal Evolution Subsequent models of speciation project modes of speciation at different points in time (Lewis & Raven, 1958; Gottlieb, 1973; Sytsma et al., 1990; Gottlieb & 1992). Clarkia, therefore, interesting and complex system in which to examine Edwards, provides an chromosomal speciation, as it involves several modes shifts in genome downsizing, and polyploid of chromosomal evolution: base chromo- some number, speciation, em While work on Clarkia has revealed the formation and establishment of taxa through catastrophic speciation, newer genomic methods are beginning to sort out mechanisms through which this process can occur. Work in diploid Helianthus L. hybrids (Riese- berg et al., 2003) and allopolyploid Brassica (Pires et al., 2004; Gaeta et al., 2007) mechanisms that can generate rapid variation, which suggests molecular may support the ecological theory of catastrophic speciation. A modern application of Peter Raven's work on speciation in Clarkta is found in analyses of show resynthesized Brassica allopolyploids, which rapid chromosomal and phenotypic change. In brief, chromosomal rearrangements caused by homoeolog- ous recombination can alter the dosage of homoeol- ogous alleles. This variable dosage of alleles can result in novel variation for flowering time, and genotypes with rapidly diverging flowering times can become isolated in the absence of cross pollination of early and late flowering groups. These studies of resynthesized B. napus polyploids elucidate potential mechanisms of speciation because rapid changes in genome structure alter gene expression and result in es variation of important life history traits e flowering time (Pires et al., 2004; Gaeta et al., 200 7. Initial Peter Raven on selection in Clarkia suggested that dramatic changes work by catastrophic in phenotype could arise from chromosomal changes and become fixed in a population; later work Clarkia reaffirmed this pattern throughout the genus. Current research in taxa like Helianthus and Brassica also reveals that chromosomal rearrangements can cause phenotypic change that may eventually lead to speciation. TOWARD A CYTOGENETIC RENAISSANCE IN SYSTEMATICS left research that have been continually Peter Raven his mark on several areas of reexamined and expanded. His ideas regarding the basal chromosome C number in angiosperms were eventually confirmed by updated phylogenies and corresponding chromosome counts of newly identified basal angiosperm lineages. In turn, these studies have led to genome-wide investigations that have detected signatures of ancient polyploidy events occurring throughout the diversifi- cation of the angiosperms. Raven's characterization of chromosome types in Onagraceae has revealed complex evolutionary patterns in the family: these types of analyses have been extended in grasses and Brassicaceae to show how chromosomal blocks can be compared between organisms. Raven's work on catastrophic selection in Clarkia provided the theo- relical framework for further research on the molec- ular mechanisms of chromosomal rearrangements and rapid speciation. The revolution in evolutionary development was made possible by the formation of new, accessible methods of experimentation and analysis. Similarly, the cytogenetic renaissance will depend upon new methods that facilitate combining cytogenetic and phylogenetic analyses. The key to ushering forward the cytogenetic renaissance will arise from conceptu- — ally linking cytogenetics and phylogenetics by utilizing new eytogenetic tools, improving sequencing technologies, and cultivating phylogenomics (includ- ing comparative chromosomics) as a research priority (Dobigny et al., 2004; Graphodatsky, 2007). First, new cytogenetic tools and applications will make cytogenetic research more accessible. Repeti- tive DNA sequences, like retroelements, are very useful for easily labeling chromosomes from parental species in a hybrid. Fluorescent in situ hybridization (FISH) probes for retroelements, as well as many other types of probes, can be quickly obtained from a variely of library-building techniques. One relatively inexpensive and efficient protocol for building small insert libraries (about 1000 bp) produced probes that effectively labeled retroelements from the genera Zea L. and Tripsacum L. (Lamb & Birchler, 2006). New techniques in FISH are al — owing easy visualization of gene clusters or even single genes (Kato et al., 2000); regions as small as 3 kb can be effectively probed with these methods. These new developments are making cytogenetics accessible and useful for anyone interested in chromosomal evolution. Second, new sequencing technology will alter the availability of data for phylogenomic analyses. Genomic sequences for select species are allowing that could potente be used for phylogenetic analyses )4; Gilchrist et al., 2006: Bouck & Vision, 2007). Complete genome sequences can be more rapid development of nuclear gene loci mall et al., ound in nearly every plant group (gymnosperms, moss, fern, fungi, algae), allowing all plant biologists to take advantage of this information. The use of genomic data for systematics will initially derive from mining complete genome sequences from closely but the advent of related crop and model species, low-cost genome sequencing technology and phyloge- Annals of the Missouri Botanical Garden netically informed genomic comparisons (e.g.. Rokas et al., 2003; Sjolander, 2004; Delsuc et al.. 2005; Koonin, 2005: dis et al., 2005; Jeffroy et al., 2006: Jansen et al., 2007; Philippe & Blanchette, 2007) will provide resources for many species in question. Finally, research area will actively integrate cytogenetics and the cultivation of phylogenomics as a genomics with phylogenetics. Phylogenomic analyses will ultimately establish not only evolutionary relation- ships among organisms but also relationships among all genes within an organism. A phylome, the complete collection of all gene phylogenies in a genome, has been developed for humans (Huerta-Cepas et al., 2007) and in the future will be developed for plants. Currently available resources such as bacterial artificial chro- mosome (BAC) libraries and expressed sequenced tag (EST) libraries can both be used by systematists to assess gene copy number and integrate cytogenetic data and nuclear genes with phylogenies. For example, Pires et al. (2006) utilized genome size and chromo- some number to explore genome evolution in Aspar- — agales and proposed that far more information could be unearthed using BACs or ESTs from crop plants like asparagus and onion to investigate non-model taxa like hyacinths, agave, amaryllids, irises, and orchids. The proliferation of genomic resources coupled to a fusion of systematic and cytogenetic methodologies promises to open up whole new areas of research. Phylogenomics ancestral reconstruction, the will allow for genome incorporation of genome-level characters into phyloge- netic analyses, and new theories about evolution on a genomic scale. Literature Cited Blanc, G., K. Hokamp € K. H. Wolfe. 2003. A recent polyploidy super Impose «d Or 1 olde 35 large ¿SE ind duplic alions 144. in the pr em genome. Genome Res. 13: 137 Bouck, A. & T. Vis 2007. The molecular e ets s ; guide to expresse ^ se Pu nce tags. Molec. Ecol. E 24. Bowers, J. E., B. Rong € hi 2002. genome phylogenetic analysis of e and duplication events. Nature 422: 433—438. Breidbach, O. & M. T. deve Ea nt: Past, 125: 157-171. M. D. & L. G. Cook. Chapman, eae acne O n n by 2007. future. Evolution. and Biosci. Ghiselin. Theory present, and Crisp, 2005. Do early branc hing lineages signify ance a traits? ‘Trends Ecol. Evol. 122-12 Cui, L., P. K. Wall, J. H. I eebe ns-Mack, B. G. Lindsay, D. E. Soltis, J. J Ee p. - J. E. o K. Arumuganathan, A. Barakat Albert, H. Ma & W. dePamphilis. 2006. W ide pude genome Vp throughout the history of 1 flowering plants. Genome Res. 16: 7 38-749 ] b po .D. Abe "en 1956. Chromosome Botany. The University Delsuc, F., H. Brinkmann € H. Philippe. 2005. Phyloge- nomies and the reconstruction iem the tree of life. Nat. Rey Genet. 6: 361-375. Devos. K. M.. J. K. M. Brown & J. L. Bennetzen. 2002. Genome size re «due lion through ille 'gitimale rec combination counteracts genome lrabidopsis. Genome Res. 12: 1075-1079 Dobieny, G., J. F. 2004. Cytogenetics and cl ied lies. Sys y expans lon in R oe & v. Bpod, . Biol. 53: 470-484. Lambrou. 2005 UE of Ducroz, T. Ehren Takhtajania, other Y dorf Gr; F. « IVI. Winte raceae, and Canellac Gacy P hy lo- genetic implications. Ann. Missouri Bot. Gard. 87: 107—413. — E. Kr 1908 anglo- E. Habeler € W. Sauer. primitive idl, Chromosome numbers and. evolution Taxon 17: 337—353. 1969. Acade P of Sciences, Leningrad. Gaeta, R. dT Osborn, 2007. ra NOMIC c Piin ‘s dn res ae Chromosome Numbers of Flowering ) rp Leon & a © Pires, F. Iniguez-Luy, E. ized lucus i and o ire ffe ct on gene expression e phenotype. 3417 Pl. Cell 19: 3403— E. ` e W. H e” C. C. Ying, 5. P. Otto, J. Zhuang, D E heung, B. Hamberger, F. Mrmr. T Kalynyak, L. Johnson, J. Bohlmann, B. E. Ellis, A . Cronk. 2006. Use of Ecotilling as an ry tool to variation i1 in Ecol. 1 efficien ey genetic wild ations of Pun Mns Molec. 1367-1376 Gottlieb. L. " in Clarkia franciscana, C. Evolution 27: 205-214. & S. W. Edwards. 1973. Enzyme differentiation and phylogeny rubicunda, and C. amoena. 1992, test of a newly disc An € le 2 troy shore lc the genetic independence o ered popula- tion of re franciscana. Madroño 39: 1-7 Graphodatsky, S. 2007. Comparative chromosomics. Tolec. Biol. " 301-375. cook ue J.. H. Dopazo, J. Dopazo & T. Gabaldon. 2007. The human phylome. Genome Biol. 8: RI Jo. . Cai, L Raubeson, H. Leebens-Mack, K. F. singer-Bellian, R. C. ; : Chumley, S. B. Lee, R. . R. McNeal, J. V. Kuehl & J 2007. Analysis of 81 genes from 64 plastid resolves Jansen, R. K. Deparia i Boore. genomes relationships in angiosperms and evolutionary patterns. Proe. Natl. 19369-19374. Delsuc € H. Philippe. 2006. beginning of incongruence? Trends identifies genome-scale Acad. Sei. U.S.A. 104: Jeffroy, O., H. Phylogenomics: The Genet. 22: 225-231. Kato, A., P. S. Albert, J. M. Vega € J. A. Birehler. 2006. in situ hybridization signal detec- Brinkmann, F. ” Sensitive fluorescence in in maize using directly labeled probes produced by nick high concentration DNA poo rase translation. Biotechnol. Histochem. 81: 7 Koonin, E. V. 20095. s e alogs. and evolutionary genomics. Ann. Rev. Genet. 39: 300-338 d H. 195 Var lation p fertility of hybrids between NITE d populations of the se rpe sey Mun cles Sireptanthus g landulos Hook. Evolution 11: 211. Kurabayashi, M. Lame & P. H. Raven. do: \ Amer. J. comparative sind ul mitosis in the Onagraceae. Bot. 49: 1003-11 E 6. Lamb. J. C. & J. A. Birchler. 2006. painting: C en al visualization of retroelement expan- Retroelement genome sions in ~ genera Zea and Tripsacum. Genetics 173: 1007-10 Volume 95, Number 2 2008 Pires & Hertweck Polyploidy and Chromosomal Evolution Levin, R. A., Pires, ^. L. Wagner, P. C. Hoch, M. Nepokroeff, J. € is mmer & K. J. Sytsma. 2003. Family- level eo of Onagraceae based on chloroplast rbcL and ndhF data. Amer. J. Bot. 90: 107-115. 2. Catastrophic selection as a factor in speciation. Evolution 16: 257-271 ——— & P. H. Raven. 1958. Rapid evolution in Clarkia. Kogani E 319-336. Lov 03. Evolutionary i ue innovation, and e es iP Be nary Biol. P arat 9: Lysak, Es a Pecinka, R. Schmidt, K. McBree & I. 3 ds 2006. Mos s of chromosome ole T reduction in S thaliana and rs d B species. Proc. Natl. Acad. Sci. U.S.A. 5224— Mason, H. L. en e CE factor in narrow n E d influences. Madroño 8: 209-2 Muller . 2007. Ev Rum Nat. Rev. Conan Oginuma, K., T. Jaffre € H T ind developmental biology. rassicaceae ee the es Tobe. 2000. > karyotype analysis of somatic chromosomes i in a trichopoda V d J. Pl. Res : 281-283. Parkin, I. A. P., S. M. d A. G. Sharpe, I Trick, T. C. on & D. J. Lydiate. 2005. structure the Brassica napus . Lukens, V Segme he et 0 genome based on comparative analysis with Arabidopsis thaliana. Genetics 171: 765-781. Paterson, s € B. A. Chapman. Ancient vac parere diverge cereals, and its « consequences for co c. Natl. Acad. Sci. U.S.A. 101: 9903-99( H. & M. Blanchette. 2007. TER She of phylogenomics conference. BMC Evol. Biol. S 2004. ance of the mparative genomics. — Philippe. "the first (suppl. 1): ———, F. H. Parr & N. ano GE 2005. a ae | Rev. Ecol. Syst. 541—502. Pigliueci, M. 2007. Do we ven an Bone evolutionary synthesis? Ev olution 61: 2743-2749. Pires, J. C.. . Zhao, M. E. Schranz, E. J. Leon, P N. A & T. C. divergence J on, P. A. Osborn. 2004. Flowering rearrangements in re Bugs Brass be (Brassicaceae). d 675-6 088. — — ——, |. J. Maureira. T. J. Givnish, K. J. Sytsma, O. Seberg G. Petersen. J. I. Da W. Stevenson, P. J. Rudall, M. F. Fay & M. W. Chase. 2006. Phylogeny, genome size, and e evolution of Asparagales. Al 5-302. Raven, P. H. 1964. Catastrophic selection and edaphic a Evolution 18: 336-338. . 197 Quijada, L. time nd genomic syn- Biol. J. TM avis, D. iso 22: 75. The bases of Ear We phylogeny: Cytology. 64 Ann. Missouri Bot. Gard. 62: ———. 1986. Modern aspects of the Biologie ia Pe cies in lanis Pp. 11-30 in K. Iwatsuki, P. H & W. Bock (e nk Modern Aspects of Species. Univ. of Tokyo Press, Tokyc — —— & D. y. Kyhos. 1965. New evidence concerning the original ba chromosome number of asic angiosperms. E mon 19: e 248. : S uci D. M. T. Nakaza UR E a an ia c. 2003. Major ecological transitions in wild sun b TS oe e d by AEE Science 2 1211-1216. a Williams, N. King € S. B. Carroll. 2003. Genome-sc es approaches to resolving incongruence in Rosenthal, Z. Lai, K. d A. E. Schwarz- molecular es Nature 425: 798-804. Schranz, M. E. A. Lysak € T. Mitchell-Olds. 2006. The ABC? s of c ompar rative genomics in the e 'aceae: dep blocks of crucifer genomes. Trends Pl. 535- Sjolander, K. 2004. e a nomic inference of protein molecular function: we ces and challenges. Bioinfor- s 20: 17 p 19. 55 It. Small. R. I . Cronn & J. F. Wendel. 2004. Use of nuc i 'ar genes ra phy loge ny reconstruction in plants. Austral. Sy x: Bot. 17: 145-170. Soltis, D. E., P. S. Soltis, M. D. Bennett & I. J. Leitch. 2003. Evolugon " genome size in the angiosperms. Amer. J. Bot. 90: 1596-1603. ———— . andress & M. W. Chase. 2005. Ehslósény and bu of the Angiosperms. Sinauer Associates, Inc., Sunderland, Massachusetts. Soltis, P. S. 2005. Ancient and recent polyploidy in angiosperms. New v Phytol . 166: Sytsma, K. J. & J. F. Smith. 1992. Mole “cular systematics of rl eae: Exam les from Clarkia and F uschia. In D. E. Soltis. S. Solus € J. J s). Molecular Syste EY s P a Doyle Chapman and I & J. C. s. 2001. Plan syste matics in he. next 50 years: Re- mapping the new frontier. Taxon 50: 713-732. ————, —— & L D. Gottlieb. 1990. Phylogenetics in Clarkia (Onagraceae) ; Res E ee site mapping of chloro- plast DNA. Syst. Bot. 15 Tanaka, R., K. Oginuma & s. P 1988. ASA ocal studies on 26 s ecies in ter generc 1 52: 1675-1696. numbers, =) ae. E Kromosomo II-51- W. 2. P ines ek phylogeny, elos mos of qM and their bearing on the one nl all basic omame number of angiosperms. Annonaceae oH —= N DUE ET Soy “4 (s qn S. Li. H. The | genomes of 2m saliva: PLoS Biol. & J. Zhou, et al. 2005. ^ history of duplications. TAXONOMIC REVISION OF A. Michele Funston? ROLDANA (ASTERACEAE: SENECIONEAE), A GENUS OF THE SOUTHWESTERN U.S.A., MEXICO, AND CENTRAL AMERICA! ABSTRACT A revision of the genus Roldana La Llave (Asteraceae: Senecioneae) is given, inc ae a key to species and complete nomenclature with synonymies, description, distribution, and n 'ussion sections for 48 species; eight varieties are recognized and ll new combinations are made, including one species transfer to the genus Pucalip H. Rob. € Brettell phytogeography section for the genus, which is distribute paran the highlands of Mexico and Central America, with one species entering the southwestern-most U.S.A., is also included. New combinations are e stablished for R. e (Bertol.) Funston, R. alieni (B. L. Rob. & Se x Funston, R. hartwegit var. carlomasonii (B. L. Turner & T. M. Barkley) Funston, R. hartwegii var. durangensis (H. Rob. € Brettell) Funston, R. hartwegii var. Tie de (H. Rob.) Funston, R. kerberi var. calzadana (B. L. Turner) Funston, x kerberi var. manantlanensis (R. R. Kowal) Funston, R. petasitis var. cristobalensis (Greenm.) Funston, A. petasitis var. oaxacana (Hemsl.) Funston, R. petasitis var. sartorii (Sch. Bip. ex Hem? Funston, as wel as for the excluded taxon P. er Ru (He e ) Funston & Villaseñor. Neotypes are designated for Cineraria platanifolia Schrank, R. Ja dn (Klatt) H. Rob. & Brettell. R. lobata La Llay Ni petasitis (Sims) H. Rob. € Brettell, Senecio acerifolius K. Koch, 5. canicidus Sessé & Mew. and 5. prainianus A. Berger. Lectotypes are an E lor Cacalia nutans Sessé & Mc 1. peltata Sessé & Moc., esta R. v nervia a n.) H. Rob. € Brettell, R. aschenborniana (S Sha er) Ho Ri " x ed R. gilgii (baa Rob. & Brettell, R. d Hem 1 ) H. Rob. & Brettell, R. Rob. & Brettell, R. kerberi jae: un & Brette ae inglasset (Greenm.) H. Rob. € Brettell, R. lanie idis (ea enm. : H. Rob. & x tell, R. petasitis var. a R. petasitis var. sartorit, S. Ares Greenm., 5. api var. areolatus Greenm., S. chrismarii Greenm., S. ghi ~ T var. pauciflorus J. M. Coult., 5. -o Loes. var. glabrior Hemsl., 5. MONS Greenm., 5. mange us Greenm., 5. lobatus Sessé € Moc.. s rias Hemsl., D Sessé & Moc., and S. schumannianus S. Schauer. = > e = sac) 3 Q US Sa Ee = a = = O ey wore Asteraceae, Central America, Compositae, Mesoamerica, Mexico, Roldana, Senecioneae, taxonomy. The genus Roldana La Llave (Asteraceae: Sene- arises from a hairy, tuberous caudex with numerous cioneae) is distributed throughout the highlands of fleshy roots; leaves typically palmate or, if pinnate, Mexico and Central America. It is placed in the 1 — ien. typically longer than 10 cm and/or variously subtribe Tussilagininae because it has stigmatic hairy. In this synopsis, 48 species and eight varieties surfaces united across the inner face of the style are recognized and 11 new combinations are made, branches, upper part of stamen filaments (anther including one species transfer to the genus Psaca- collar) cylindrical, chromosome numbers n = 30, and liopsis. principal phyllaries often with the midrib notably The genus has two centers of diversity; one is along thickened at the base. The other tussilaginoid genera the Trans-Mexican Volcanic Belt and Sierra Madre with which it is associated, based on geographic del Sur of south-central Mexico and the other is along distribution, are Digitacalia Pippen, Psacalium Cass., — the Sierra Madre de Chiapas and Cordillera Centro- Robinsonecio T. M. Barkley € Janovec, Pippenalia americana of Chiapas and Central America. The McVaugh, Rn H. Rob. € Brettell, Villase- species of Roldana vary in growth form from small noria B. L. Clark, Pittocaulon H. Rob. € Brettell, perennial herbs that seldom exceed 1 m in height to Telanthophora H. Rob. € Brettell, Nelsonianthus H. freely branching trees that may reach 12 m in height. Rob. € Brettell, and Barkleyanthus H. Rob. € The architecture of the plants is insufficiently well Brettell. It is separated from these genera because it understood. to apply the rather precise descriptive ‘If I were to properly recognize all those who helped with this revision, it would surely take as many pages; in memoriam, | thank my mother Sandra Scharbach Funston and my O professor Theodore Mitchell Moon y for their guidance of this work. Runde ial support was provided by the Division of Biology. Kansas State University, anc dı pene Botanical Garden. issouri Botanical Garden, P.O. Box 299, St. DEN Missouri 63166-0299, U.S.A n@att.net. a 10.341 7/2003 151 ANN. Missouri Bor. Garb. 95: 282-337. PUBLISHED ON 18 June 2008. Volume 95, Number 2 2008 Funston 283 Taxonomic Revision of Roldana terminology of contemporary ecology, e.g., Hallé et al. (1978), although the utility of such a classification scheme is acknowledged. The genus is predominantly made up of perennial suffruticose herbs that may grow to shrubs. There are also four species of acaulescent perennial herbs and nine species that may grow to trees; stems are solitary or in clumps. Plants are typically distributed in the understory of pine-oak forests at elevations between 1000 and 4000 m. Herbaceous plants are also found along the forest [em edge in patches of disturbed grasslands. All species may be found in disturbed areas and roadsides. The main characters used to separate the species are habit, type and density of pubescence, number of phyllaries, and, to some extent, the shape of the leaf blade and its form of attachment. In 1901, J. M. Greenman wrote a revision of Senecio L. s.l. North and Central America. In his revision, he separated the genus into 22 sections. from Since that time, there has been considerable debate whether the genus should remain together or be split into segregate genera. Currently, Greenman's concept of Senecio is split into two subtribes, with 11 genera from Mexico and Central America placed in the subtribe Tussilagininae. While Robinson and Brettell (1973a, b, these genera that included keys, none was given a comp c, 1974) made a comprehensive review of = ete revision. MATERIALS AND METHODS Information and materials were derived from field experiences in Costa Rica in January 1994, and Mexico in March 1996, from consultations in CR, IEB, MEXU, MO, and from herbarium ea that were lent to KSC from A, ASU, CA , F, GH KANU, MEXU, MICH, MO, NMC, nid OS; SD, TEX, UC, UMO, and XAL—herbarium abbreviations follow Holmgren et al. (1990). Herbaria referred to, but from which no loans were requested, are B, C, C, K, P, and US. Gibson (1969: 97) notes in his dissertation, “Correspondence with Dr. G. Wagenitz indicates that all of the types of the genus Senecio at the Berlin Herbarium were destroyed in Species are presented in alphabetical order. Complete lists of exsiccatae and distribution maps are available from the author. Microcharacters as reviewed by King and Robinson (1970), Jeffrey (1987, 1992), and Nordenstam (1978) have proven useful in the Senecioneae. Microchar- acter studies of Senecio include Drury and Watson (1965), Jeffrey et al. (1977), Jeffrey (1979), Vincent et al. (1992), and Wetter (1983). For my dissertation (Funston, 1999), investigated for consistency within the genus. Not all microcharacters in Roldana were species could be coded; however, those that were While microcharacters did confirm the placement of species showed a significant degree of similarity. in the subtribe Tussilagininae, they were not sufficient —— to include or exclude a species at the generic level. A (1987, 1992) analytical descriptors is available from the author. character state matrix of Roldana for Jeffrey's A numerical analysis performed on the species Roldana hartwegii (Benth.) H. Rob. & Brettell is also available from matrix. of 64 collections and five character states was analyzed using NTSYS-PC v. 1.8 (Rohlf, 1993). The data matrix was standardized using a simple matching coefficient. the author. A data The standardized matrix was then clustered using the unweighted pair-group method, arithmetic average. l'he product moment correlation was used to obtain a cophenetic correlation of r = 0.8 TAXONOMIC HISTORY The genus Roldana stems from sect. Palmatinervii Hoffman (1894), a portion of the section Fruticosi Greenm. (1901), and two genera, Pericalia Cass. (1924) and Digitacalia (1968), from the old “cacalioid” complex of Mexico and Central America. Senecto The derivation of the genus is representative of the tortured struggle to segregate natural groups from the ubiquitous and heterogeneous Senecio s.l. Hoffman (1894) defined palmately veined senecios of In Engler and Prantl, Central America as forming the section Palmatinervit. He did not list any f the T. species. Greenman composed a monograph of North which was published in his doctoral dissertation and Central American species of Senecio, (1901) and through a series of papers in the Annals of the Missouri Botanical Garden (1914, 1915). In his dissertation, he divided the genus into 22 sections that included the sections Palmatinervii and Fruti- cosi. A revision of the sections Palmatinervit and Fruticosi was never completed. These were included that is, with in the subgenus Fusenecio, style LI branches truncate, rounded-obtuse, or occasionally terminated by a penicillate tuft of hairs: stems erect or ascending; stems without a foreshortened axis: oil tubes not richly developed in the cortex. The Fruticosi consisted of 13 species of pinnately veined shrubs or small trees. The Palmatinervii included 33 species and two varieties of palmateh veined herbs, shrubs, or small trees. His concept of the genus remained the same for his treatment of Senecio in Standley (1926), tions and deletions of the species although there were several addi- that he ascribed to the two sections. 284 Annals of the she Botanical Garden Tangled within the attempt to separate sections 5 and/or genera from Senecio s.l. is the concept of cacalioid and senecionoid complexes, first defined by Cassini (1827). His cacalioid complex included those plants with white- or cream-colored disk florets, eradiate capitula, and elongate corolla lobes, and his senecionoid complex had yellow corollas, typically radiate capitula, and short corolla lobes. These complexes have met with varying degrees of accep- ance as natural groupings. ec Over time, several philosophical trends developed. In 1837, de post-Linnaean authors, revising the cacalioid complex Candolle summarized Cassini and other pain to include species occurring in Asia, eastern. North America, and Mexico that differed consistently. from Senecio in a number of respects, treating it as the single widespread genus Cacalía L. Botanists who have accepted this cireumscription are Gray (1883), Hoffman (1894), d (1950), Blake (1960). and Gleason and Cronquist (1963). Many authors did the Cacalia, preferring to relegate all of the species to Senecio: Schultz (1845), Bentham (1€ cuc Greenman (1901). Ferna not accept de Candollean 373a, b). Hemsley (1881), MeVaugh (1972. 1984). Rzedowski (1978), and Barkley (1985a. b). Concurrently, in the 20th century, some authors accepted the concept of splitting Senecio s.l. into cacalioid and senecionoid complexes, and then within these complexes creating segregate genera. These authors include Rydberg (1924), Cuatrecasas (1960, 1981), Pippen (1968). Robinson and Brettell (1973a. b. 974). and Nordenstam (1978). In 1984, Jeffrey and Chen were the first to recognize the two complexes as subtribes: the senecionoid complex as the subtribe Senecioninae and the cacalioid complex as the subtribe Tussilagi- ninae. Bremer (1994. two subtribes primarily on three characters. Species in used this concept, defining the — the Senecioninae have balusterform filament collars, style branches with separate stigmatic areas, and : basic chromosome number of x = 10 or 5 the stvle branches with a continuous stigmatic area, and a >. Species in pa Tussilagininae have cylindrical filament collars, basic chromosome number of x = 30. As for Roldana. in 1924 Rydberg genera distributed in North and Central of the circumscribing these genera, | revised three America oul cacalioid and senecionoid complexes. In ie pointed out that the characters by which Cacalia was being separated from Rydberg noted that some disk many are rayless, and that a few species of Cacalia “The trouble has arisen because botanists have included in Senecio were inconsistent. “true” senecios have cream-colored corollas, have short corolla lobes. As he so aptly stated: Cacalia several different types which can easily be distinguished. from each other and. from Senecio if considered separate but not if taken as an aggregate" (Rydberg, 1924: 369). It is on this foundation that he the three genera Psacalium, Odontotrichum Zucc., and Pericalia. For the purposes of this paper, revised only the genus Pericalia is relevant. Pericalia is circumscribed as having four species (P. sessilifolia Hook. & Arn.) Rydb., P. suffulta (Greenm.) Rydb., P ovatifolia (Sch. Bip.) Rydb., Rob.) Rydb.) with the following suite of characters: short corolla lobes, the style branches of Senecio, and and P. michoacana (B. L. corollas white or nearly so (although, in fact, the corollas may be orange in P. suffulta). Pippen defined cacalioid and senecionoid charac- teristics (1968: 371) and circumscribed species based on the combination of these and other characters. Following. Rydberg, he expressed the opinion that: “The distinctive elements of North America and Asia formerly placed in Cacalia would be better placed in several narrowly circumscribed segregate genera.” He considered the section Palmatinervii as cacalioid and and 5. cooperi Greenm. and recognized four segregate genera o O > o related to Psacalium, Senecio angulifolius DC., from the Mexican cacalioid complex: Digitacalia, Without explanation, he only included P. sessilifolia and P. Odontotrichum, Pericalia, and Psacalium. michoacana in the genus Pericalia, yet he was aware of Rydberg’s paper and cited it. The first monograph of the section Palmatinervii was completed by Gibson in 1969. While never published, it is a reservoir of information. Gibson included 27 species and five varieties in the section. He was the first to. pull several species from the section Fruticost and the genus Pericalia under the fabric of Palmatinervii. While all of the species were placed in the genus Senecio, he considered most species of Palmatinervii as senecionoid in character, with other species alternately categorized as senecio- noid, cacalioid, or as occupying a confusing interme- diate group. — irst to the They considered it to be a Brettell (1974) were the recognize the section as a distinet genus under Robinson and resurrected name Roldana. cacalioid entity and grouped it with other genera that are concentrated in Mexico and have single fused stigmatic surfaces on the style branches and unen- larged bases on the filament collars. Giving the genus a broad interpretation, they included the entire section Palmatinervii, part of section Fruticosi, the us Poesia. andoneproblemnate species. Senec genus ericatia, and one pro tematic spec les, Senecio Klatt, scribed 46 species and one variety within the genus. heteroideus from Digitacalia. They circum- Barkley (19854) retained the concept of Senecio s.l. the senecionoid and cacalioid assemblages, respectively. divisible on basis of microcharacters into It was his belief that until revisionary work the Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana species level was completed, further generic consid- His are fundamentally in accord with the above eration would not prove meaningful. “working groups” authors. In 1996, Barkley, Clark, and Funston outlined the segregate genera of Senecio s.l. and Cacalia s.l. in Mexico and Central America, placing Roldana in the subtribe Tussilagininae. The circumscription of these genera is in agreement with the views of King and Hobinson (1970), wherein consistent morphological characters at the generic level were outlined. These characters have been used worldwide to help illuminate the natural relationships within the tribe Senecioneae. Furthermore, it has become apparent that the groupings that authors since Greenman have used are essentially equivalent to genera. It is against this backdrop that the current revision of Roldana is written. Here, the genus is placed in the subtribe species and eight varieties. It closely follows Robinson and Brettell (1974) in that it includes species from section Palmatinervii, the genus Pericalia, four species ussilagininae and defined as having 48 from section Fruticosi, and Senecio heteroideus, as well as 19 taxa culled from botanical literature. MORPHOLOGY HABIT AND ARCHITECTURE The species of Roldana fall into three general eroups of life forms, even though they are somewhat imprecise and some species are intermediate. The first eroup is comprised of sparingly branched or acaules- cent herbaceous perennials that are mostly less than 1 m in height. The second group includes strict, erect, single-stemmed perennials that may grow to 4 m in height and are herbaceous or weakly ligneous toward the base. These plants are often solitary or loosely clustered in few-stemmed groups. Plants of the third eroup are shrubby or tree-like with branching stems that are either solitary or in crowded clusters. At least the lower stems are clearly ligneous, and growth rings are present in the small trees. Root structure in the genus consists of a woody, typically lanate caudex the with spreading rhizomes. In perennial herbs, caudex is often bulbous. Field observations in Mexico revealed that vegetative reproduction occurs either by new growth from the caudex, by rhizomatous growth from the caudex, or by adventitious vegetative growth from stems that have fallen. In the latter case, where nodes from fallen stems come in contact with the eround, roots and then shoots develop, expanding the colony of plants. The important characters relative to the stem are shape, color, and vestiture; stem shape is typically terete, although Roldana acutangula (Bertol.) Funston has angulate stems and R. hartwegii has both types of stem shape. The bark on all of the tree species is — smooth with small fissures and brownish grey in color, and new growth is red or light brown. The stems of the other habit forms are either red or greenish cream. Vestiture is an important diagnostic feature used in — = ae key, varying from glabrous to a tawny lanate tomentum. Species fall into four general categories. The first category includes those plants that are glabrous, or essentially so, and are suffruticose herbs from the Trans-Mexican Volcanic Belt and northward. The second category contains those species with a light to dense covering of stipitate-glandular hairs and multicelled hairs that are found throughout Roldana’s distribution. The third type of pubescence has thin white trichomes that are arranged in either a light to dense arachnoid pubescence or a tomentum of short curly or long tawny hairs. Species with this type of pubescence are predominantly found along the Trans- Mexican Volcanic Belt and southward. Lastly, R. petasitis (Sims) H. Rob. & Brettell has a unique pubescence of tawny canescent hairs on the under- — surface of the leaves, while the pubescence of the capitulescence is stipitate-glandular. LEAVES The majority of species have cauline leaves that are mam evenly distributed on the upper half of the stem, the l Many perennial herbs have cauline leaves that are clustered about the middle of the stem, and the low-growing lowermost being deciduous. of the glabrous — herbs have basally clustered leaves. Leaf blade venation is pinnate to pinnipalmate, palmate, or triple-nerved. There are nine species with The palmately veined or triple-nerved, 1.e., there are three pinnately veined leaves. rest are distinctly main veins arising from the base of the leaf with secondary veins arising from the central main vein. The petiole is typically marginally attached to the blade. Roldana subpeltata (Sch. Bip. H. Rob. & Brettell, R. cordovensis (Hemsl.) H. Rob. & Brettell, R. heterogama (Benth. . Rob. & Brettell, and some- times R. angulifolia (DC.) H. Rob. € eccentrically peltate leaves. Roldana sessilifolia Hook. & Arn.) H. Rob. & Brettell may have sessile cauline leaves. — rettell have a Blade shape is typically ovate, rotund, lanceolate, or elliptic in outline. The blade margin is typically palmately lobed with some secondary lobing, the sinus depth extending one quarter to one half of the way to the midrib. Several species have blades with serrate All margins. The smallest principal cauline leaves in the margins. species have callous denticles on the 286 Annals of the Missouri Botanical Garden genus belong to Roldana reticulata (DC.) H. Rob. & Brettell and are 4-8 1-8 em. Roldana greenmanii H. Rob. & Brettell and R. gilgú (Greenm.) H. Rob. € Brettell have the largest leaves, up to ca. 30 X 30 cm. Leaf pubescence varies from glabrous to a persistent anate tomentum. CAPITULESCENCE The capitulescence in Roldana is terminal, except in R. glinophylla Gibson ex H. Rob. € Brettell, where Most compound paniculate cyme, but there are 10 species es} it is axillary. species have a pyramidal Jaagi that are flat-topped and described as corymbiform. The capitulescences of R. mexicana (MeVaugh) H. Rob. & Brettell and R. glinophylla have an umbel-like appearance. Plants with a simple panicle and | to 5 capitula per branch are the small perennial herbs. In an attempt to quantify the density of capitula within the capitulescence, the character number of capitula per branch was recorded. Although it does not aid in Un the separation of capitulescence types as well a anticipated, it was kept in the deseriptions as it is of some diagnostic use. Ultimate peduncle length is a useful diagnostic character. The typical peduncle length is 2-10 mm. Those plants with distinctly long ultimate peduncles are Roldana ehrenbergiana (Klat) H. Rob. € Brettell and R. sessilifolia, which have peduncles exceeding 20 cm, and R. guadalajarensis (B. L. Rob.) H. Rob. & Brettell, R. heteroidea (Klatt) H. Rob. & Brettell. R. michoacana (B. L. Rob.) H. Rob. & Brettell, and R. suffulta. (Greenm.) H. Rob. € Brettell, which have peduncles ranging from 10-15 em. Roldana lobata La (Greenm.) H. Rob. & Brettell have sessile capitula that are described as Llave and A. robinsoniana being arranged in glomerules. Pubescence varies from glabrous to lanate tomentose. The character primary bracts of the capitulescence refer to those bracts that arise at the level of the secondary peduncles as opposed to the bracts of the (1974) were the first to rachis. Robinson and Brettel use the description of these bracts as a diagnostic The filiform, and obovate. The typical forms that these character. bracts have three forms: linear, bracts take are linear or filiform and are 3—5 mm in length. The species Roldana angulifolia, R. cordo- vensis, R. greenmani, R. grimesii (B. L. Turner) C. Jeffrey, R. petasitis, and, on occasion, R. sessilifolia have large obovate bracts that are 5-15 mm in length. With regard to the rachis bracts, R. ereenmanti. R. 8 S petasitis, and R. cordovensis may have uniquely large obovate bracts. The term bracteole refers to those structures that are found on the ultimate peduncles of the capitulescence. These are typically linear in shape Roldana ebracteolate. Only R. angulifolia, R. grimesii, and R. and 1-3 mm in length. subpeltata is suffulta have obovate bracteoles. CAPITULA Roldana has radiate and eradiate capitula. They are homochromous; corolla color is yellow, yellow-orange, or cream-white. Capitula are turbinate, campanulate, or funnelform in shape. The average capitula size is ca. 10 X 3 mm. Capitulum height is measured from — the point where the capitulum attaches to the peduncle to the top of the mature florets, excluding ray ligules. Width is measured across the base of the capitulum where the phyllaries attach to the recep- tacle. All herbarium specimens. Calyculate bracts arise from measurements were taken from dried the base of the capitula. Typically they are ] to 3 in number, linear in shape, and 1-3 mm in length. |i KR. subpeltata. Obovate calyculate bracts are found on =~ Calyculate bracts are completely lacking i v. angulifolia, R. grimesii, R. reticulata, and R. suffulta. The phyllaries are of uniform length but are arranged in two weakly defined rows, the margins of = adjacent phyllaries overlapping. The phyllaries lying toward the inside of the row have hyaline margins. The base of the phyllary is thickened dorsiventrally, a condition representative of the subtribe and best observed on fresh specimens. The one exception in the genus is R. mixtecana Panero & Villaseñor, which has phyllaries in five graduated series. When dry, they brown and aave a greenish appearance purple strialions may be present. The pubescence of the phyllaries typically matches that of the peduncles, although at times it may not, and this difference can aid in distinguishing the species. The number of phyllaries present has traditionally been used as a diagnostic character. Typically, species fall into one of three categories: 5 phyllaries, 6 to 8 phyllaries, and 10 to 14 phyllaries. However, there are exceptions: R. hartwegii has collections with the full range of phyllary number and, while R. scandens Poveda & Kappelle and R. schaffneri (Sch. Bip. ex Klatt) H. Rob. € Brettell typically have 5 phyllaries, many collections have 6-phyllaried capitula as well. The receptacle surface of the capitula is alveolate and ridged on all species of Roldana. FLORETS The number of ray florets per capitula is typically 3 eds o 5. There are three types of ray florets: ligulate, inconspicuously ligulate, and tubular. Ligulate ray ha lorets have conspicuous ligules that are typically 5-7 X 2-3 mm. Roldana ehrenbergiana has the largest ray Volume 95, Number 2 2008 Funston Taxonomic Revision of Holdana ligules, measuring 10-30 X 5-10 mm. Inconspicuous ray florets have ligules that measure only ca. 2 X 1 mm. Tubular ray florets do not have a ligule, and the tube itself is typically 2-3 mm long. The presence/ absence of rays in some species is a variable character. Species that have pubescent ray florets are R. kerberi (Greenm.) H. Rob. & Brettell, R. R. hartwegii, R. hirsuticaulis (Greenm.) ob. grimesii, Funston, R. lobata, and R. subcymosa The number of disk florets per capitulum is typically less s 30. Roldana ehrenbergiana, R. sessilifolia, and R. suffulta have large capitula with 90+ disk florets. The shape of the corolla is typically funnelform, although a few species are distinctly campanulate. The corolla is divided into tube, throat, and lobes (see Pippen, 1968). All species have 5 acute lobes. Species in Roldana have the senecionoid type of lobe length, that is, it is shorter than or equal to the throat, with one exception, R. eriophylla (Greenm.) H. Rob. & Brettell, which has lobes longer than the throat. Species with a pubescent disk corolla are R. kerberi, R. grimesii, R. hartwegii, R. hirsuticaulis, R. lobata, R. subcymosa, and R. subpeltata. 1 have found the presence or absence of ray and disk corolla pubescence to be a reliable character. ACHENES The achenes within Roldana are typically 1-2 mm long and cylindrical in shape, and have 5 or 10 vertical ribs. Nine species have pubescent achenes, and R. barba-johannis (DC.) H. Rob. & Brettell and R. cordovensis have achenes with distinct resin glands. The pappus is of numerous capillary bristles and is in several series. It is typically ca. 1 mm shorter than the total length of the disk corolla. CHROMOSOMES of Roldana have had For those that have been Relatively few species chromosome counts made. examined, all are reported to be n — 30 (Turner et al., 1962; Ornduff, 1963; Turner & King, 1964; Turner & Flyr, 1966). PHYTOGEOGRAPHY There are two centers of diversity in Roldana: one is along the Trans-Mexican Volcanic Belt (TVB, Eje Neovolcánico) and the Sierra Madre del Sur, the second is along the Sierra Madre de Chiapas and Guatemala. The 48 species can be grouped into six putative clades, subjectively based on a combination of morphological characters and, to a lesser extent, The of Roldana are found distribution. species predominantly in the pine-oak forests of Mexico and Guatemala, and it is through their phytogeography that clues to Roldana’s dispersal are derived. Roldana’s distribution spans eight countries, from the extreme southern parts of Arizona an ew Mexico in the United States to the southernmost end of the Cordillera de Talamanca in Panama. Composed of weedy species, it grows in disturbed areas and in the understory of pine-oak forests throughout the highlands of Mexico and Central America. It typically grows at elevations between 2000 and 4000 m although it has been collected at elevations as low as 900 m and as high as 4100 m. Mexico is comprised of 15 physiographic regions; Roldana occurs within eight of these regions. Appendix 1 provides a species list for each region. Fifty-six Bem of the species in Roldana occur along the with four regional endemics. The Sierra Madre del Sur also contains 5696 of the species and has six regional endemics. The Sierra Madre Oriental has 3896 of the species and three endemics, the majority being restricted to the extreme southern end of the mountain chain, which runs into the eastern end of the TVB. The Sierra Madre Occidental has 23% of the species, including three endemics, with the species evenly dispersed along the mountain chain. Another significant area of diversity for Roldana hes the Isthmus of Tehuantepec in Mexico southward to the Cordillera de Talamanca in Panama, where there are 14 species equaling 29% of the genus. While the species in this region have a distinct aspect to them, there is little doubt that they should be included within Roldana. Roldana acutangula, R. gilgii, R. greenmanit, and R. scandens (endemic to Costa Rica) do not occur above the Isthmus of Tehuantepec. All of the other Central American species are concentrated in the Guatemalan highlands, yet may reach as far north as the Sierra Madre Oriental and as far south as the aforementioned Cordillera de Talamanca in Panama (Appendix 2). In contrast, R. aschenborniana (S. Schauer) H. Rob. € Brettell and R. barba-johannis are the only two species with distributions that are concentrated in the TVB and Sierra Madre del Sur that extend south of the Isthmus of Tehuantepec. The patterns of concen- tration in the regions of Mexico and Central America are similar to those found in Senecio s.l. by Barkley (1990) and that of the generic distribution in Mexico of the Senecioneae by Turner and Nesom (1993). To further elucidate the distributional patterns within the genus, I grouped the species of Roldana below into six putative clades primarily based on similar habit and morphology (Appendix 3). The Central American complex is predominantly made up of woody shrubs and small trees that have a stipitate- that includes | multicelled glandular pubescence Annals of the e Botanical Garden trichomes. E species are distributed in the highlands of Guatemala and Central America, with lwo sab exceptions. Roldana hintonit H. Rob. & Brettell occur in the — Turner) Funston TVB and the Sierra Madre Occidental. respectively, but were placed here due to their strong „a DNA) showing the relationships among these three species and A. mezquitalana (B. L. resemblance to R. schaffneri. Further data (e.g would give clues to the overall dispersal history of the pg genus. The R. hartwegii complex is composed of herbs and subshrubs with an arachnoid floccose pubescence on the peduncles and a light tomentum on the undersurface of the leaves. They are typically found along the TVB and northward along both the Sierra Madre Oriental. The R. lobata complex is made up predominantly of woody shrubs Occidental and and small trees that have a persistent. tomentum, which may be tawny on the new growth, and a dense lanate tomentum on the undersurfaces of the leaves. Turner) B. L. Turner) B. L. Roldana gonzaleziae (B. L. Turner and R. sundbergii (B. L. Turner are perennial m herbs, but were The concentrated. about the placed here due to their tawny pubescence. R. lobata complex is centrally TVB and Sierra Madre del Sur but also extends northward along the Sierra Madre Oriental and Occidental as well as southward into America. Central The species A. petasitis with its three varieties was placed in a complex unto itself due o its broad distribution and mixed pubescence of woolly hairs on the undersurface of the leaves and hairs on the and stipitate-glandular peduncles phyllaries. pang Its core distribution is in the highlands of Guatemala. At its northernmost extension, much of its unique pubescence is lost, and R. pelasilis appears to hybridize with R. angulifolia in the southern ranges of the Sierra Madre del Sur. is made up of suffruticose shrubs that have a stipitate- like the R. hartwegii complex in that it is distributed throughout the TVB and Madre Oriental. Lastly, the R. pericalia complex includes all of those The R. cordovensis complex glandular pubescence. It is the Sierra northward along species formerly placed in the genus Pericalia by Rydberg (1924), suffruticose herbs to subshrubs which included eradiate glabrous . | have extended that definition to include radiate species and plants that are essentially glabrous. This group, like the R. lobata complex. is centered in the TVB with some northern and southern extensions. Examination of the clades reveals that there are two areas of diversity in the genus. The Central American and Roldana petasitis complexes are concentrated about the Guatemalan highlands and rarely reach above the Isthmus of Tehuantepec. The majority of four clades are concentrated TVB and Sierra Madre species in the other — throughout the del Sur with northern extensions along both the Sierra Madre Occidental and Oriental. What is the explanation for this distribution? Were the progenitors of Roldana from North America o South that descendant of the senecios of Eurasia, America? One theory is Roldana is a which were dispersed to North America via the arcto-tertiary geoflora of classical phytogeography ( Barkley. pers. comm.) Another possibility is that they dispersed from South America across the Panamanian land bridge. The difficulty with the latter theory is that the tropical flora of Mexico are typically found below a. 1000 m (3500 ft, Miranda € Sharp, 1950). Miranda and Sharp (1950) also note that the transition zone between 1000 m and 1200 m (3300-4000 ft.) is the Neotropical floris dividing line between the arcto-tertiary and Graham (1973, 1975, 1976) and Gentry (1982) give further evidence that, while the lowland tropical plants of Central America lc regions. and southern Mexico have dispersed northward from South America, this does not include plants above the elevation. of the tropical flora. Roldana makes its 1500 m, these forests have traditionally been allied. with the home in the pine-oak forests above ca. and flora of North America. Even in the regions of Central America where Roldana is found, the habitat has more affinities with the Mexican flora than that of South America (Steyermark, 1950; Graham, 1993). In particular, Steyermark remarks that the pine-oak forests of the Cuchumatanes and Sierra de las Minas Moun- Therefore. I of Guatemala resemble the southern Rocky New Mexico. concur with Barkley that the most likely ains, especially those of avenue of dispersal is from North America. The geo-histories of Mexico and Central America give clues to the distribution patterns of senecios into Mexico. Barkley progenitors diversified into these regions during the (1990) believes that Senecio s.l. or its Late Tertiary. From the Cretaceous. through the Quaternary, these regions underwent extensive tec- tonie activity (Coney, 1982). The Sierra Madre Occidental is believed to have arisen during the Cretaceous and Paleocene. The TVB was developed during the end of the Tertiary, with its principal uplift ms rom the Miocene to the Quaternary. By the end of the Tertiary, southern Mexico had access to temperate biotas of North America and tropical biotas to the south (Graham, 1993). These invasions were precip- itated by warm-cool and wet-dry climatic fluctuations that varied independently of each other. It is believed that the areas of temperate biotas increased during the Middle Miocene, tropical cool climates of the Late Eocene, the increased during the warm climates of Middle and the Pleistocene, while biotas the Paleocene, Early and Eocene, Oligocene, and Early Volume 95, Number 2 2008 Funston 289 Taxonomic Revision of Roldana ‘he ebb and Miocene (Toledo, 1982; Graham, 1993). I flow of these biotas created cycles of refugia and primary communities that surely affected speciation rates through vicariance. McDonald (1993) provides indirect geological and that acent regions paleological evidence that suggests montane vegetation in northeastern Mexico and adj descended ca. 1000 to 1300 m below their present-day positions during the peak of the Wisconsin glacial 70,000-12,000 years before present). would have provided migration corridors for montane the This lowering poem elements of Mexico during colder periods. As montane vegetation returned to higher elevations, it would cut off these routes, isolating populations and His that montane forests would have occurred all creating refugia. theory holds continuous along the — Sierra Madre Occidental and would have been intermit- tent along the TVB and Sierra Madre del Sur. These events recurred, creating short-term isolated populations followed by the re-integration of populations. Perhaps the most cogent explanation of the factors affecting the dispersal and speciation events within montane flowering plants is given by Takhtajan (1968). populations in which genetic Montane regions form pockets of disjunct J Genetic that drift occurs. the isolation of forms are + drift enhances statistically unlikely to be expressed in larger populations due to natural selection. The accumula- tion of small neutral mutations in these isolates is ereater than that of larger populations. Polymorphisms or new adaptations in intraspecific forms may arise, which may ultimately play a role in interspecific differences. Roldana hartwegii, with its variation in phyllary number and achene pubescence, may well be an example of this phenomenon. Occasional inter- breeding between such populations increases their variability and, hence, their plasticity. Hybridization produces more material suitable for rapid adaptive evolution, thus ereating evolutionary jumps without evidence of precursor organisms. The endemics found in the TVB, Sierra Madre del Sur, and both the Sierra Madre Occidental and Oriental may be examples of this type of speciation event. An important piece that is missing from the puzzle is the reproductive biology of the genus. Unfortunate- ly, the breeding mechanisms of Roldana are a matter of complete speculation. Barkley (1990) believes that North outcrossers with they function similarly to the aureoids of America, in that they are normally generalized pollinators and that their cytological structure allows easy hybridization and introgression. Because the chromosome number is n = 30 in a tribe in which the base number is thought to be x = 10, the ancestor of the genus (and probably all of the tussilaginoids) was hexaploid, although the descen- dants are now functioning as diploids. It should also be noted that many species are prodigious vegetative reproducers. All signs point to the fact that Roldana is a genus that is comfortable in disturbed habitats. l ancient iis regard, the role of man cannot be ignored. Since the populace has inhabited the pine-oak forests of the times, greatest part of the Mexican TVB and Sierra Madre del Sur. Irrigated agriculture has been carried on there for some 6000 years (Barkley, 1990). Bye (1993) discusses the interplay between mankind and the diversification of weeds in the Mexican flora. For millennia, the people of the Tehuacán Valley have exploited the biotic community creating disturbed habitats that the prolif eration of Ww eedy spec les. for medicine and food, throughout the region are favorable to — the species ly In conclusion, Xf Roldana proba have breeding mechanisms that allow for hybridiza- tion and introgression when populations come in contact with each other. The fluctuating climate and high level of tectonic activity of Mexico and Central the Early provided an ever-changing montane environment. America since at least Tertiary have This varying habitat caused a recurring. cycle of refugia and primary communities n the pine-oak forests with a concomitant cycle of isolation of breeding populations in Roldana followed by a re-integration of those populations. The net effect of these cycles, and recent human habitation, 1s a morphologically variable genus resulting from the natural selection and evolutionary processes of genetic drift. As with all genera. details concerning the phytogeography and evolution await a well- resolved phylogeny of the species. TAXONOMIC TREATMENT Roldana La Llave, Nov. Veg. Descr. 2: 10. 1825. TYPE: Roldana lobata La Llave. Senecio sect. Palmatinervii Hoffm. in A. Engler & K. Prantl. 1894. Die Naturlichen Pflanzenfamilien. Leipzig. 4: 64 Perennial herbs. suffruticose shrubs, and small trees, 0.2-8.0 m tall, woody, often lanate caudex, arising from a fibrous-rooted, growing singly in colonies; stems erect to ascending, p not foreshortened below capitulescence, glabrous to having a tawny lanate tomentum. Leaves simple. alternate, venation pe to pinnate; petiolate to rarely peltate or sessile, cauline, variously clustered to — evenly distributed on upper half of stem; ovate, orbicular, lanceolate, elliptic, or pinnatifid, typically 5- to 13-lobed, lobes acute or rounded, sinus depth midrib but typically about. 1/2 of the way to the ranging from subentire to dissected, abaxial surface Annals of the Missouri Botanical Garden glabrate to densely lanate tomentose and canescent, adaxial surface typically glabrescent, margins vari- ously callous denticulate. Capitulescence typically a compound paniculate eyme, rarely a corymbiform yme, or a simple panicle; ultimate peduncles typically 6-10 mm, but reaching up to 22 cm in some species; rachis bracts typically minutely foliaceous, or obovate; primary bracts and bracteoles may be filiform, linear, or obovate. Capitula radiate or eradiate; homochromous, corolla cream, yellow, or orange (may appear white when dry); capitula 7— 10(25) X ca. 3(10) mm; filiform, linear, or obovate; phyllaries 5 t calyculate bracts ca. 5, 14(25), typically in | apparent series, margins overlapping, or graduated in 5 series, typically glabrous or stipitate- pubescent, but may be arachnoid or tomentose, ca. 4) X 1-3 mm. Receptacle surface alveolate and Ray to. 5(8). inconspicuous (ligule measuring less than florets absent. or 3 corollas ridged. ligulate, 3 mm), or reduced to only a tube, glabrous or pubescent, ligule 5—7(30) X ca. 3(10) mm; venation Disk florets 6 to 30(904-), campanulate, 5-lobed, glabrous typically 4, unbranched. corolla funnelform or or pubescent, typically 8-10 mm, lobes of the corolla as long as or shorter than the throat. Stigmatic surface entire, style branch apex conical without apical tufts. Stylopodium free, style base cylindrical or expanded. Filament collar cylindrical. Anther base rounded short-tailed, anther appendage about twice as long as KEY TO THE SPECIES OF ROLDANA IN THE SOUTHWESTERN UNITED S TATES, MEXICO, wide ¢ Endothecial tissue semi- © r slightly longer. polarized. Achenes glabrous or pubescent, cylindrical, 1-2 mm, ribs 5 or 10, resin glands typically absent; > B y! ) pappus of numerous white capillary bristles, in several series, ca. 7—9 mm. Pappus hairs tapered (rarely elavate), 2 or 3 apical cells divided (rarely appressed). Carpopodium medium or large. Ovary wall crystals mixed, predominantly druses quadrates. Achene glabrous, or pubescent with short clavate or long thin hairs. Distribution and ecology. Extreme southern Ar- izona and New Mexico in the United States, Mexico, Guatemala, Honduras, El Salvador, Nicaragua, Costa Rica, and 1000—4100 m. Montane cloud forests and pine-oak forests. Distrib- Panama at elevations of uted in understory, margins of open grassland, waste places, and along roadsides. Associated with Abies Mill., Alnus Mill, Arbutus L., Baccharis lL., Chi- ranthodendron Larreat., Clethra L., Colubrina Rich. ex Brogn., Cornus L., Cupressus Dahlia Cav., Distylium Siebold & Zuce., Eryngium L., Ficus L., Fraxinus L., Fuchsia L., Ipomoea L., Liquidambar L., Meliosma Blume, Mimosa L., Ostrya Scop., Penstemon Schmidel, Physalis L., Phytolacca L., Platanus L., Podocarpus Labill., Populus L., Potentil- la 13 Salvia L., Solanum L., and Symphoricarpus Kunth. Flowering Pinus L., Prunus L., Quercus L., Senecio, time is year round, typically November through April. AND CENTRAL ÁMERICA la. Leaf blade venation pinnate (to pinnipalmate as in R. barba-johannis and R. neogibsonii). 2a. Phyllaries 5 to 8. 3a. Leaf blade 's el liptic-linear, margins entire to d serrate; phyllaries 5 or6 ......... 43. R. schaffneri 3b. Leaf blades pinnate ly 7- to 11-lobed; phyll aries 7 or 8. 4a. Achenes pubescent; leaf blade ca. 7-lobed n sinus de ph more than 1/2 of the way to p midrib, base decurrent on E petiole foin a wing, abaxial surface glabrescent ...... 20. R. herac n 4b. Achenes glabrous; leaf blade obs d with sinus Ds typically no more than 1/3 if the we to the midrib, WR not on on petiole, abaxial surface densely pubescent m ed iuri P "rrr 9, R. lineolata 2b. Phyllaries 9 to 14. 5a. Plants perennial herbs; capitula with 50 to 60 a florets: nro ta aa a 9. R. ehrenbergiana 5b. Plants suffruticose herbs to shrubs; capitula with 9 to 20 disk flor 6a. Peduncles and abaxial leaf blade surface persistently tomentose. Ta. Peduncles and abaxial leaf blade surface velvety tomentose: leaf blades oblong TET eee gages ode a REI LEE D EM EE M ». R. UM 7b. e les and abaxial leaf blade surface tomentose with white woolly hairs; leaf blac broadly lanceolate, oval, or ovate ooo 7. R. barba ni Ob. Peduncles and abaxial leaf blade surface glabrescent, stipitate-pubescent, or arachnoid pubescent. Plants essentially glabrous; leaf blades linear, 7-10X longer than wide ............ A E EE E E at EE AEA R. guadalajarensis 8b. pubescent; leaf blades elliptic, 3X longer than wide. Capitulescence pubescent with appressed arachnoid hairs; leaf margins entire to subentire; known only from Temascaltepec, México... .......0.... 23. R. hintonii Ob. Capitulescence densely e puse ent; leaf margins serrulate; known only fro the states of Durango and Jalisco ooo... ooo... 33. R. iia Ib. Lea f blade venation tog to triple-nervec 10a. Phyllaries 25 to 35, in five LOb. Phyllaries 5 5 to nn in one apparent series. graduated series .... 35. R. mixtecana Volume 95, Number 2 Funston 291 2008 Taxonomic Revision of Roldana lla. Ta 5 to 8. essentially glabro 13a. Leaf blades dde Sine ly lobed to dissected; achenes glabrous ...... 22. R. heteroidea 13b. = blades ovate to orbicular, margins serrate to lobed with sinus E Eb less than a of the way to the midrib; achenes pubescent. 14a. Leaf blades with nie marginally attached; corollas glabrous ...... 32. R. mexicana Ab. Leaf blades eccentrically peltate; corollas pubescent .....o.o.o.o.o.o.. 45. R. subpeltata 12b. Plants variously pubescent. 15a. Plants acaulescent perennial herbs, typically 20-70 cm tall. 16a. Achenes glabrous; corollas pubescent; found in Durango and Jalisco .. 14. A. bue s Achene pureren pies gionan e in Nuevo León ....... 7. R. sundbergü 15b. t perennial herbs, shrubs, or trees, typically greater TM m high. Plants with a primary Br nce type i stipitate d and/or multicelled hairs: - e sduncles always variously stipitate-pubescent, abaxial leaf blade surface glabrous to dens ' and persistently pubescent with stipitate-glandular ae multicelled hairs or variously tomentose (as in R. petasitis). 1 ine leaves variously hastate to having 3 divaricate lobes, secondarily lobed go E to merely denticulate. Capitula radiate; ray ligule 10-13 X 2—4 mm; from Oaxaca .. 5. R. anisophylla 19b. Capitula eradiate or radiate with ray ligule inconspicuous. 20a. Capitula eradiate; from Oaxaca o... .ooo coco... 19. R. hederifolia 20b. Capitula eradiate or rays 3 to 5, ray ligule ca. 3 X 0.5 mm: dn México and Michoacán ........ llle ene . R. aliena 18b. Cauline leaves variously palmatifid, Pie ue ovale, or oblong: 5- to 7(4)- m 2la. Leaf blades orbic ular to ovate, typically 7(+)-lobed. 22a. Leaves eccentrically peltate; c id leaves palmately (5 to)7- to 17- lobed, lobes acute llle 2]. R. heterogama 22b. Leaf petioles marginally attached: cauline leaves ovate to orbicular, - to 13-lobed. 23a. Leaf lobes 9 to 13, deeply cut forming rectangular segments topped with 3 acute lobes: blades abaxially pube scent with shor and long multice led hairs, densely so on veins .. 15. R. greenmanii 23b. Leaf lobes ca. iia allow; blades abaxially ae scent lo densely lanate tomentose . .. eee 7. R. petasitis 21b. Leaf blade "s variously palmatifid. e ovate to oblong, 5- to 7(13)- be d 24a. Capitulescence a simple cymose panicle: leaves clustered at the base CA e O «2462s ex xr eu E idees s 31. R. metepeca 24b. Capitulescence a compound paniculate cyme; leaves cauline, evenly distributed on stem, lowermost deciduous 25a. Mg uh leaves distinctly sec andara lobed; calyeulate bracts ca. 3, obovate, 5-15 X ca. 2 mm: phyllaries 9-13 X 1.3— AN A ee ee ee ae L. R. angulifolta 25b. Cauline leaves without secondary lobing or minutely so; calyculate bracts 0 to 3, linear to filiform, 1-2 mm: phyllarie s 7-9 X ca. | mm. 26a. Cauline leaves eccentrically peltate ..... 8. R. cordovensis 26b. Cauline leaf petioles marginally attached. 27a. Cauline leaves ovate to orbicular, shallowly ca. 7- lobed; adaxially puberulent, abaxially pubes scent to de nsely lanate tomentose ......o... . R. petasitis 27b. Cauline leaves palmatifid, 5- to 7-lobed: us lly glabrescent, abaxially glabrescent to stipitate- glandular pubescent. 28a. Capitula 12-15 X 3-5 mm: from northern Mexico (Chihuahua. Durango, Sinaloa). . SE Aeneas ons io eee ee |. R. gentryi 28b. Capitula 6-11 X 2-3 mm: from southern exico (C s Oaxaca, Veracruz) and Cenal America 29a. Primary boss linear, 24 mm long: ultimate ae ca. 6 mm long; ) bracteoles O to 3, linear to scale-like, 1-2 mm long ........ 25. R. jurgensenü 20b. Primary bracts linear, 5—15 mm long or obovate 10—30 X 5-15 mm: ultimate peduncles 8-20 mm long: bracteoles O to 3, filiform, up to 3 mm long 3T de beer 292 Annals of the Missouri Botanical Garden 17b. Plants bun a primary pubescence type of arachnoid or woolly hairs; peduncles and xial leaf blade surfaces variously glabrate, arachnoid, or tomentose pubescent. 30a. Pe n les and abaxial leaf blade surface variously tomentose pubescent. 3la. Leaves seasonally deciduous, new leaves emanating from stem apex, creating a foreshortened stem — ~ o K. de 31b. Leaves cauline about evenly distributed on upper half of stem, lowermos de ciduous, stem not foreshortened. Plants with a light tomentum on peduncles and abaxial leaf surface, becoming glabrate with age. 33a. Shrubs and small trees, 3—7 m tall .......... 2. R. albonervia 33b. Single-stemmed pere a he 2c 1-2 m tall . . 18. R. hartwegii 32b. Plants with a dense persistent, at times pem tomentum on pe peduncles and abaxial leaf surface, never glabrate. 34a. Leaf blades palmately veined, rotund to cordate-reniform, crenate to shallowly 9(+)-lobed; some margins e: long ile E: denticles up to 3 mm long ............... R. lanicaulis 34b. Leaf blades triple-nerved to palmately veined, ovate to c Dr entire to shallowly ca. 5- lobed: margins never with elonga denticlesS uuu BE ee ode eee eee nS 41. R. ica A 30b. Peduneles and Fes le af blade surface arachnoid floccose to d “ent. 35a. Leaf blades distinctly palmately nerved, orbicular o... 2. R. scandens 35b. Leaf blades triple-nerved, variously palmatifid or ovate. Leaf blades palmatifid: plants of montane cloud forests in Chiapas, Mexico, and Guatemala; phyllaries 8, glabrous, 6—7 X 1-2 mm -—À— —— É———— R. acutangula 36b. Leaf blades cue ovate; plants of pine-oak forests in the uthwestern U.S. nd Me xico; phyllaries (5 toJ8 to 13, oe to arachnoid floccose. ra 5 X ca. | mm. 37a. Plants shrubs iud mall trees, 3-7 m tall ..... 2. R. albonervia 37b. Plants single-stemmed suffruticose herbs, 1-4 m tall. 38a. Plants 1.5—4.7 m tall, principal cauline leaf blades omis X (2-)10-25 em. petioles (2-)10-20 cm A EE hens agen ap & be oe © oo 26. R. vi 38b. Plants 1-3 m tall, princ ue cauline leaf blades 5-14 5-16 em, petioles 3-7 em long 39a. Plants I Ea ee era pubescent with hirsute hairs: rare, uds [rom one collection in Hidalgo VE Gr Ex ELEM up a nies Begg 39. R. reglensis 30b. Plants Re to arachnoid floccose ae COMMON zac eee RA a RA HO ek 8. R. hartwegü IIb. P hyllaries 9 to 14. 40a. Plants essentially glabrous. la. Capitula radiate. 2a. «cal blades palmatifid, 5-lobed, distine ind star-shaped, margins entire .. 13. R. glinophylla 42b. Leaf blades ovate, margins variously serrate aooaaeoa aaao aaa aaa. 40. R. reticulata Alb. Capitula eradiate 43a. Calyculate pum ls oval-elliptie; achenes pubescent o... 46. R. suffulta 43b. Calyculate bracts linear: achenes glabrous. dda. Capitula with 36 to 58 florets: leaf blades never sessile, subpalmatifid, 3- to 5(7)- R Obed A set eee eee uP Asse ee eae 34. R. michoacana 44b. ( e with 10+ florets: leaf blades may be sessile. E to ovate- ipte; 5--to 9-lobed 4 corria da pga Gea rte 44. R. sessilifolia LOb. i ants variously Med "scent. Plants ac aule 'scent to subcaulescent perennial herbs, typic cally 20-70 em tall. x 4-7 loa. Petioles 1.5—7 cm long; leaf blades 3.5—7 —i em, ovate, with ca. 6 shallow acute lobes: plants glabrous to floecose arac duod A a . Hacotepecana 46b. Pe ile ss 6-12 em long; leaf blades 6-11 X 8-13 em, pao - lo 7 lobed. lobes acute, secondarily lobed, sinus depth about 1/2 of the way to the midrib; plants pubescent with a mixture of stipitate-glandular m i multicelled hairs ...... TTD IRR 38. A R. platanifolia lob. Plants caulescent suffruticose perennial herbs, shrubs, or trees, typically greater than 1 m tall. 47a T ants with a primary pubescence type of stipitate-glandular and multicelled hairs: peduncles and abusi blade surface slightly to dense ly pubescent (sometimes so dense as ^ appear tomentose in A. gilgii and R. grime ‘i. Ba. Phyllaries ca. 1 mm wide. a. Leaf blades eccentr ‘ally peltate; capitula eradiate ..... 21. R. heterogama 19b. Leaf blades NS attached to petiole; capitula radiate .. 27. R. langlassei Funston 293 Taxonomic Revision of Roldana Volume 95, Number 2 2008 48b. aa greater than 2 mm wide. 50a. itulescence a simple cymose panicle; leaves clustered at the base or mic mm o a ce rad aa . metepeca 50l l | le; | ly di abi | 50b. Capitulescence a compound cymose gle eaves evenly distributed on upper half of stem, lowermost deciduou 5la. Leaf blades triple-nerved, palmatifid, 5 - to 7(13)-lobed, sinus depth 2 to less than 1/2 of the way to the midrib; petioles glabrate, marginally attached to blade or eccentrically peltate ... 4. R. angulifolia 51b. Leaf blades palmately veined (to weakly triple -nerved), rotund to reniform. 10(+)-lobed, sinus depth less than 1/4 of the way to the midrib; epa densely pubescent, marginally attached to blade. 52a. Calyculate bracts to inear, 2-3 mm long, densely pubescent with stipitate-g alandi and long multicelled hairs, margins pubescent wih short multicelled hairs, not ciliate A E ae EEEE SURE UR 2. R. gilgit 52b. Calyculate bracts 3 to 6, obovate to elliptic, 10-14 mm long, pubescent with spitale -glandular hairs, margins ciliate with R. grimesit 5 47b. Plants with a primary pubescence type of arachnoid or woolly hairs; peduncles and abaxial leaf blade surface glabrate to tomentose. 3a. Pedunc les and abaxial leaf blade surface variously tomentose pubescent. Pubescence of capitula restricted to tomentose tufts at base 40. R. reticulata 54b. Pubescence of capitula not restricted to base 55a. Corollas eus scent. 56a. Leaf blades ovate to cordate, with ca. 10 shallow lobes to crenale; petioles and stems lanate tomentose 24. R, hirsuticaulis 56b. Leaf blades ovate to sag subpalmatifid, 5- to 7-lobed, sin depth 1/4 or less of the way to the midrib, lobes acute, or ea rotund to ue with subentire margins: Jue stems floccose tomentose .... oo... o... .. ). R. lobata 55b. Corolla E )us. seal blades ovate, oblong, or rotund, mostly 5- to 9-lobed, sinus depth less than 1/4 of the way to the er or margins sinuate; peduncles floccose arachnoid . e e N 57b. Leaf des cordate-orbicular, margins crenate to nine 9(+)-lobed, sinus depth less than 1/8 of ile way lo us midr x peduncles lanate tomentose ......o.oo.o.oo.o.. 28. R. lanicaulis 53b. du and abaxial leaf blade surface variously arachnoid pubescent. 58a. f blades oval and subentire to ovate and weakly (3 to)5- to 7- s d, b surface arachnoid to 2 tomentose puis scent... 30. R. lobata 58b. Leaf blades ovate to oval, (5 to)7- to 15-lobed, abaxial surface ac to weakly arachnoid pubescen nt. 59a. Plants 1.5—4.7 m tall; oup (2—)10—20 em long; ied ipal cauline leaf blades (5—)10—35 0-25 cm; foune along the Sierra Madre de Manantlán and Madre del Sur ........ . kerberi 59b. A 1—3 m tall; petioles n long: princ Kel cauline leaf bl ade X 8-16 cm; found ap as ihe Sierra Madre Occidental and tal 18. / s Sierra Madre Orien €. hartwegii [2 l. Roldana acutangula (Bertol.) Funston, comb. distributed on upper half of stem, lowermost decidu- nov. Basionym: Cineraria acutangula Bertol., Novi ous; petioles 8-18 cm. glabrate to sparsely pubescent; Comment. Acad. Sci. 1840. Senecio acutangulus (Bertol.) Hemsl., Biol. Bot. 2: 235. 1881. Roldana acutan- gula (Hemsl.) H. Rob. & Brettell, Phyto 415. 1974, nom. illeg. TYPE: Inst. Bononiensis 4: 435. blades 12-24 X 6-16 em, papyraceous, triple-nerved, palmatifid, 5- to 9-lobed, sinus depth 1/2 of the way to lobes acute, Cent.-Amer., less than 1/2 of the way to the midrib, — irregularly serrate, margins with multicellular hairs: ogia 27: Guatemala. Sacate- abaxially glabrate with pubescence of short and long 1837, J. Velasquez s.n. (holotype, BOLO not seen, microfiche IDC 748. 46 III 7!; isotype, BOLO fragm. US not seen). péquez: slopes of Volcán de Agua, 1-4 m tall; new growth glabrate to arachnoid floccose; stems 4- to Robust suffruticose herbs or shrubs, 6-angulate, red or brown. Leaves cauline, evenly glandular hairs on veins, adaxially glabrous to sparsely pubescent with short glandular hairs. Capitulescence compound paniculate cyme (sometimes with a flat- topped corymbiform aspect), 20 to 50 capitula per branch, peduncles arachnoid pubescent to glabrate, primary bracts linear, 1-3 mm, ultimate peduncles 2— 7 mm, bracteoles 1 or 2. scale-like to linear, 1-2 mm. Capitula radiate, 8-10 X 2-3 mm, funnelform, corollas 294 Annals of the Missouri Botanical Garden yellow: calyculate bracts 0 to 5, linear, 1-2 mm: phyllaries 8. glabrous, greenish, 6-7 X ca. | mm. Ray florets 3 to 5, corolla ligulate, glabrous, tube ca. 5 mm ligule 5-7 X Disk florets 5 to 10, , funnelform, glabrous, 7-8 mm, tube: —3 mm. corolla 3A mm, throat ca. 2 mm, lobes ca. 2 mm. Achenes sparsely pubescent, ribs 10, pappus bristles 5-6 mm. cylindrical, ca. 2 mm, resin. glands absent; Distribution and phenology. Southern Mexico and Huehuete- San Sololá, Suchitepéquez. Totonicapán: Mexico: Chiapas at elevations of 1400-3300(3800) m. Montane cloud forests. Flowering December through March. Guatemala (Guatemala: Chimaltenango, nango. Quezaltenango. Sacatepéquez, Marcos, Comments. — Roldana acutangula (Hemsl.) H. Rob. & Brettell is a International Code of Botanical Nomenclature (Greu- ter et al., 2000: Art. 33.6, Ex. 11). make their combination, Robinson and Brettell (1974) nomen illegitimum under the In attempting to did not correctly reference the basionym author, and as Hemsley (1881) provided an explicit reference to Bertolini (1840), Art. 33.6(a) (Greuter et al., 2000) does not apply and the combination under Roldana was nol validly published by H. Robinson and Brettell. Bertolini (1840) listed the specimen “in In the protologue for Cineraria acutangula, Vulcano " Annalisa Managlia at BOLO (pers. June 2003) confirms the existence of the J. d'acqua. Perenn? comm., Velasquez s.n. specimen and further notes that some fragments of it are housed at US. The angulate stems of Roldana acutangula are unique. The only other species in which this occurs, = on occasion, is R. hartwegii. Morphologically, acutangula is most similar to R. jurgensenit (Hemsl.) H. Hob. & Brettell. pubescence, stem shape. leaf base, phyllary pubes- They differ in types of cence, and achene pubescence. Roldana acutangula has an arachnoid floccose pubescence of long thin white hairs, whereas KR. jurgensenii has short-stipitate and long multicelled brown hairs that vary from sparse to dense. For the remaining diagnostic characters, R. leaf base cordate, acutangula has angulate stems, phyllary pubescence glabrous, and pubescent achenes: A. jurgensenii has terete stems. leaf base truncate or weakly cordate, phyllary pubescence stipitate-glandular, and glabrous achenes. UATEMALA. Chi- Chicha 1500 m, 5 J; [uehuetenango: Sierra Representative spec imens examined. G maltenango: a Chimaltenango, Dec. 1933, A. F. Skutch id ds de m Cuchumatanes. rd. beyond La Prader: 2m 32, 3300 m, 31 Dec. 1940, P. C. Standley es m om. E of San Mateo Ixtatán, Sierra de los Cue :humatane "ss pus m, 31 July 1942, J. A. Steyermark 49886 (F. MO). Quezalte- nango: Km 172 on hwy. to entrance Quezaltenango. Huehuetenango & Totonicapán, 2860 m. 10 Jan. 1974 Ll R. 30209 (V, MICH); nil, 2500 m, 4 Mar. 1939, P. équez: Ai olcán EM 'atenango, 2500 m, Hunnewell 14922 (GH). Fuentes Georginas & Standley 67334 s 15 Feb. 2 D rd be lw. San Marcos: mins. betw. San & Serchill, 2900 m, 30 Jan. 1941, P. C. Standley 85378 (F). Sololá: betw. María Tecun ^5 Los Encuentros, Ss 2800 m, 29 Nov. 1969, Molina R. 24990 (F) Suchitepéquez: SW lower slopes of Volcán E nil, vic. of PUER Asturias, NE of Pueblo Nuevo, 1250 m, 1 Feb. 1940, . Steyermark 35335 (F). Totonie apai vin CAI betw ANA lenango & Chimaltenango, i ‘tw jet. in rd. to Quezaltenango & Mee ca. SSE of jet. to Que zog Énango, 14: 52'N, 9 bed MA E - m, 23 Jan. 1987, T. vat 63497 (KS T Encuentros y. María l'ecun, 2800 m, 21 Nov. f Molina R. bis (F, MO, NY). MEXICO. Chiapas: in the paraje ol Ollas, mpio. Chamula, 2600 m, 20 Dec. 1964. D. E. func 8012 (DS, MICH, MO, NY). 2. Roldana albonervia (Greenm.) H. Rob. & Brettell, Phytologia 27: 415. 1974. Senecio albonervius Greenm., Ann. Missouri Bot. Gard. l: 275. 1914. TYPE: Mexico. México: Valley of Apr. 1831, C. J. W. Schiede s.n. (lectotype, designated here, GH! ea — Shrubs to small trees, 3—7 m tall; old growth woody, brown, terete, sparsely lenticeled; new growth pur- plish red, terete, arachnoid floccose to tomentose, elabrate with age; the contrast between old and new growth may give the stem a foreshortened appearance Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 5-10 cm, gla- 10-17 x 8 14 em, blades triple-nerved, broadly cordate to ovate, rate to arachnoid floccose: blades 5- to 9-lobed, sinus depth less than 1/4 of the way to the midrib, lobes acute, margins glabrous: pubescent on both surfaces, mature leaves glabrate with hirsute hairs on veins, may be tomentose on abaxial surface. Capitulescence pyramidal compound paniculate cyme, 20 to 60 capitula per branch, peduncles arachnoid floccose to glabrate, primary bracts linear, 2—5 mm, ultimate peduncles 2-6 mm, bracteoles ca. 2, linear, 1-2 mm. 10-15 turbinate, corollas bright yellow; Capitula radiate, ‘a. 3 mm, ^ calyculate bracts 3 to 5, linear, 1-2 mm; phyllaries 8, glabrate, 5-8 X ca. | mm. Ray florets 3 to 5, corolla ligulate, glabrous, tube 4-5 mm, ligule 5-6 X » 12, corol at s of corolla lobes; up to 2.5 mm. Disk florets a narrow, glabrous (sparsely pubescent . 8-11 mm, tube 3 | mm. —5 mm, throat — 3—4 mm, lobes ca. Achenes glabrous, cylin- ribs 5 or 10, resin glands absent: drical, ca. 2 mm, pappus bristles 6-8 mm. Distribution and phenology. Central Mexico (Hidalgo, Jalisco. México. Michoacán. Morelos, Pue- bla, Veracruz) at elevations of 1500—3200(4100) n Found in pine-oak-fir forests, open woods, and margins of open grassland. Flowering February through April. Volume 95, Number 2 2008 Funston 295 Taxonomic Revision of Roldana Comments. Greenman (1914) cites B and GH as the type material for Roldana albonervia; the Berlin specimen was destroyed in 1943. Roldana albonervia is a common species along the roadsides of central Mexico. It is very similar to R. aschenborniana; however, there are several characters both are shrubs, R. albonervia has the propensity to become small trees that can be diagnostic: while reaching 7 m tall; the peduncles of R. albonervia have a floccose pubescence versus a persistent pubescence in R. aschenborniana; and the phyllary number in R. albonervia is 8 versus 13 in R. aschenborniana. Representative specimens examined. MEXICO. Hidalgo: 23 km N of rd. Pachuca Zacualtipán, mpio. Mineral del Chic :0, 2630 m, 24 Mar. 1978, García P. e AS, F, MICH, plateau. center 4-5 km NNE of Cerro San Miguel & 4 n: S of Rincón de Manantlán, NE of crossroads 3-35'N Jalisco: = ae on plateau, Sierra de Manantlán ce antral, 19°; 104° 10- d de 2900 m, 13 Jan. 198 % R. Kowal 2844 MIC H, IS). Méxicc sop un P a a tl, ca. 1, 2: 198 4, T. M. Barkley 3350 m, Mar.— Michoacán: 3002 (KS ` NY) poe 2430— 1903, C. P. jode Mr VER, t W of (Mil Cumbre, cumbre on rd. to Morelia, 10°37'47'N, een 25 Mar. 1996, T. M. Barkley : Sierra de Tres Marías, 3050 m, 15 Apr. È T2 8903 (F, GH, MICH, MO). P ca. P km W of Huejotzingo on Fed. hwy. 190 Libre, ca. 4 km E of Puebla— Mexico state line, 2800 m, 22 Mar. 1984, T. M. Barkley 2998 NY Mineral del Monte, C. Ehrenberg 324 228 racruz: (fragm. at GH). o Roldana aliena (B. L. Rob. & Seaton) Funston, comb. nov. Basionym: Senecio alienus B. L. Ro & Seaton, Proc. Amer. Acad. Arts 28: 110. 1893. TYPE: Mexico. Michoacán: Mtns. near Pátzcuaro, 22 Dec. 1891, C. C. Pringle 5056 (holotype, GH!). Ann. Missouri Bot. Gard. 1: 278 oc.|, Chrismar s.n. (lectotype, designated here, B fragm. & tracing GH! photos F!, GH!, MICH!, MO!, NY!, B tracing MO). Senecio chrismarii Greenm., c 1914. Syn. nov. TYPE: Mexico. [s. | Simple or sparingly branched suffruticose herbs, 0.5-2 m tall; plants sparsely to densely pubescent with stipitate-glandular and multicelled hairs; stems terete, striate, red or brown. Leaves cauline, evenly distributed on upper half of stem, lowermost decid- uous; petioles 5-11.5 cm, glabrate to sparsely pubes- 6-10.5 Xx marginally cent: blades 2-11 em, eccentrically peltate or had to petiole, triple- nerved, hastate to subhastate, 3 primary lobes, weakly secondarily lobed to denticulate, margins sparsely pubescent; abaxially glabrate with glandular hairs on veins, adaxially glabrous to sparsely pubescent with short glandular hairs. Capitulescence simple to compound paniculate cyme, 3 to LO ol per branch, peduncles densely stipitate-glandular pubes- cent and sparsely so with long multicelled hairs, primary bracts linear, 5-10 mm, ultimate peduncles 10-30 mm, 3 mm. Capitula radiate or eradiate, linear or filiform, 1— 10215. 5€ 2- e alyc ulate bracts bracteoles | to 3, 5 mm, funnelform, corollas pl filiform or linear, 1-2 mm; elandular pubescent, ca. absent or 3 to 5, ligule inconspicuous, glabrous, tube ca. 2 mm, ligule ca. 3 X 0.5 mm. Disk florets 14 to tube ca. 23, corolla funnelform, glabrous, ca. 9 mm, 4 mm, throat ca. 4 mm, lobes ca. | mm. Achenes glabrous, cylindrical, ca. 2 mm, ribs 5 or 10, resin glands absent; pappus bristles ca. 8 mm. Central Mexico (Mé- 2000—2500 m. Flowering December Distribution and phenology. xico, Michoacán) at elevations of Found in pine-oak forests. through February. Comments. In the protologue for Senecio chris- marii, Greenman (1914) listed the type material as being a specimen in B and a tracing and fragments in 4: 279), in the protologue of S. kerberi, make it known that he took GH. Comments made by Greenman (191 fragments and tracings of the Berlin specimens; therefore, the GH tracing and fragments are deemed an isotype at the time of publication. Roldana aliena is part of a complex of plants that has been difficult to circumscribe due to the small number of collections available. The complex in- cludes R. aliena, R. anisophylla, and R. hederifolia. The relevant characters used to distinguish these species are marginal versus eccentrically peltate blade attachment, the presence or absence of ray florets, and the distribution of each species. Roldana aliena was described by B. L. Robinson and Seaton (1893) as having eccentrically peltate, triangular-ovate leaves with 3 to 5 angulate lobes, eradiate capitula, and with the type collection from Michoacán. | have extended this description to — include radiate capitula with inconspicuously ligulate rays based on three collections from México that I believe belong to the species concept of R. aliena. The lectotype designation for Senecio chrismarii was made according to the International Code of Botanical Art. 8.3, Ex. 5). The type collection is from an unknown location. Senecio Nomenclature (Greuter et al., 2000 chrismarii was described by Greenman (1914) as having ¡om leaf blades marginally attached to the petiole anc triangular-ovate with 3 to 5 hastate lobes, and eradiate capitula. Greenman separates this species from Rol- dana aliena based on its having deeply cordate leaves and the absence of peltation. In my concept of the genus, this form is not enough to distinguish the species, so it is placed in synonymy. XICO. México: 1955, Carlson Representative specimens examined. Mk Avendero, Valle de Bravo, 2000 m, 8 Jan. Annals of the Missouri Botanical Garden 2895 (F); Nanc hititla. Temascállepec, 23 Feb. 1933, NY): Zacualpan, 2100 Michoacán: 35 rd. mi. Eo 36.5 mi. E of Morelia, MICI KSC, 1965, E. MIC H); 2500 m, 20 Feb. S. Gibson 1033 (KSC, Rob. & Brettell. Senecio angulifolius 4. Roldana angulifolia (DC.) H. Phytologia 27: 415. 1974. DC., Prodr. 6: 431. 1837 [1838]. TYPE: Mexico. México: Toluca, flor. Apr., G. Andrieux 297 (lectotype, designated by MeVaugh, 1984: 809, G-DC not seen, microfiche IDC 800. 1142 IE 6!, G-DC photos F!, MICH!, MO!). Cacalia berlandieri DC., Prodr. 6: 328. 1837 [1838]. Senecio berlandieri (DC.) Sch. Bip., Flora 28: 498. 1845, non Senecio. berlandieri (DC.) Hemsl., 1881, nom. illeg Senecio desertorum Hemsl.. Biol. Gent.-Amer., Bot. 2 239. 1881, xicanae vallis montanis ad Viej jo desierto, nom. superfl. TYPE: Mexico. Querétaro: “in E -DC not s . mic on he -DC photos F!, MICH. Wochenschr. Vereines Beford. Staaten 4: 237. Senecio angulifolius DC. ws K. Koch, sartenbaues Konigl. TYP Senecto ace rifolit Preuss. E: same as that of (neotype, designated here, Senecio adenolepis Greenm., Publ. Field C n Mus.. E: Mex Bot. Ser. 2: 281. 1907. Syn. nov. TYP xico. ig los: Sierra de Tepoztlán, 7500 ft., E Feb. 1907, » Pringle 13909 (holotype, GH). Senecio edo var. ingens e enm., Ann. Missouri Bot. Gard. |: 276. 1914. TYPE: Mexico. México: Mt. Ixtaecthuatl, above re r line, Mar.—July 1903, C. Purpus 193 (holotype. MO!: US not seen). = Isoly pe. Sparingly branched shrublets to tree-like shrubs, 2-7 m tall, typically found growing as solitary stems connected by root stocks; new growth variously pubescent with stipitate-glandular and long multi- celled hairs; stems terete, maculate, reddish brown. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 4-15 em, typi- cally densely pubescent; blades 12-18 X 16-25 cm, marginally attached to petiole or sometimes eccentri- 7(13)- lobed, sinus depth 1/2 to less than 1/2 of the way to cally peltate, triple-nerved, palmatifid, 5- to the midrib, lobes acute, secondarily lobed: surfaces densely pubescent to glabrate with hairs persisting on veins and blade Capitulescence a loose 10 to 20 capitula per branch, peduncles densely stipitate-glandular pubes- margin. compound paniculate cyme, cent, primary bracts typically sessile, variously 15-30 30 mm, bracteoles ca. | X ca. 8 mm, ultimate peduncles 10 )x 2- elliptic, ped pe variously ovate, ca. | 4 mm. Capitula radiate or eradiate, 10-15 X 4- 6 mm, campanulate, corollas yellow; calvculate bracts ca. 3, variously oblong, 5-15 X ca. 2 mm, may have ciliate margins of long multicelled hairs; phyllaries 8 ll, densely stipitate-glandular pubescent, 9-13 X 3-2 mm. Ray florets absent. or to 7, corolla reduced to a tube. 5X 1-2 mm: 22(40). 4 mm, inconspicuously ligulate or glabrous, tube 4-5 mm, ligule 3- when Disk florets 12 to 10 mm, reduced, tube 2-3 mm. corolla narrow, glabrous, ca tube ca. throat ca. 5 mm, lobes ca. 1.5 mm. Achenes glabrous, cylindrical, ca. 3 mm, ribs 10, resin glands absent; pappus bristles 8-9 mm. Distribution and phenology. Mexico (Distrito Fe- deral, Durango, Guerrero, Guanajuato, Hidalgo, Jalisco, México, Michoacán, Morelos, Nuevo León. Oaxaca, Puebla, Querétaro, San Luis Potosí. Tlaxcala, 1200-3400 m. pine-oak or fir forests and in mixed shrub-grassland at Veracruz) at elevations of Found in higher elevations. Flowering February through April. Comments. Regarding the lectotypification of Roldana angulifolia, de the specimens E. M. Mairet s.n. and G. Andrieux 297, Candolle's protologue lists but he did not designate a holotype. The destroyed in type of Senecio acerifolius K. Koch was the Berlin herbarium fire of 1943: Mexico. |s. loc. & s. coll.]. Herb. Hort. Reg. Bero- linensis 6153 (B tracings GH!, on the GH tracing, in hand, E O!). Greenman notes “appears to be of type no specimen of the the neotypified Do making it homotypie with Roldana material from Berlin.” Because type material is believed extant, name is angulifolia. The epithet can be confused with its multiple heterotypic homonyms, which include S. acerifolius Hemsl., 1881, S. acerifolius C. Winkl., 1893, and S. acerifolius Klatt, 1894. Under Cacalia berlandieri, Senecio berlandieri Bot. 2: 236. 1881 Gynoxts cent.-Ámer., d Hemsl., Biol. a heterotypic nom. illeg. based on herlandieri DC., 1837. Hemsley made the superfluous nomen novum desertorum believing that he had blocked the transfer of C. berlandieri to Senecio, when in actuality the combination had already been made in E 845. Roldana angulifolia is most similar to R. cordo- vensis but is distinct in that it has obovate calyculate bracts and leaf blades typically marginally attached to has linear to filiform petiole. Roldana cordovensis calyculate bracts and eccentrically peltate leaf blade attachment. The obovate calyculate and primary bracts exhibited by œR. angulifolia are unique diagnostic features. The combination of these two bract types is found only one other species, R. grimesit, from which it is easily distinguished. Roldana angulifolia is one of the more beautiful and widespread species in the genus; therefore, it is that with it. In his original description, de nol surprising there are several Synonyms associated Candolle (1837) listed Cineraria angulata Mairet ex DC., nom. nud., in synonymy. Greenman created the Volume 95, Number 2 2008 Funston 297 Taxonomic Revision of Roldana variety Senecio angulifolius var. ingens based on its having fewer and larger capitula than the typical variety, stating that the capitula of the variety is 1.5— 2 cm high and 40- to 50-flowered. While there are specimens with the above characters, there is no correlation of these characters with distribution. growth being due lo an possibility of this larger excellent growing environment cannot be excluded. Based on this possibility and an examination of the holotype, the variety is placed in synonymy. Cacalia berlandieri was placed in synonymy by McVaugh (1984) and Gibson (1969); upon examining photo- graphs of the holotype, it is also placed in synonymy here. Senecio acerifolius K. Koch was placed in synonymy by Greenman (1926); upon examining the type material for this species it is again relegated to synonymy. Senecio adenolepis was placed in synonymy under S. chapalensis by Gibson (1969). My examina- tion of the holotype places the — entity within KR. angulifolia. has obovate bracts and leaf blades marginally attached to petiole. These characters are [om not diagnostic of 5. chapalensis; therefore, it is placed in synonymy here. MEXICO. Distrito t W of La Venta. Representative aes examined, Federal: on hwy. ' Mexico City, just 27 Dec. 1970, Al 17068 (MO). Durango: Puerto Buenos Aires, 53 km Salto, along hwy. Durango— Mazatlán, 2650 m, J. ae P. 902 (MO). Guerrero: Cerro Teotepec mpio. ae about 40 km N of Coyuca de Benitez 3350 m, 5 De 196: RE 2922 (M NY) uM ca. 3 km E nta Rosa, 2650 m, 11 Nov 970, R. McVaugh 242 200 us H). Hidalgo: 6 km N de + es 3000 m, 31 Dec. 1963, J. Rzedowski 18227 (DS, > hills of e 2 H) mpio. o 2900 m. 17 p 1975. Ventura A. 7 (CAS, MO). Jalisco: 25-30 km SE of Autlan, on iss rds. E rd.-eross called Pa Cumbre," betw. El Chante € Cuzalapa, 35'N, 10: 15'W, 2600 m. 20 Mar. 1965, R. Mc cup E 36 (MIC b México: 3.1 mi. W of summit of divide on toll rd. to Puebla. 3 Jan. 1972, Dunn 16779 (MO). Michoacán: on ies “an site of Mex—Mich border, Km 71 W of pas 'a, 1926'26"N, 100 11'38"W, 2585 1 Mar. 1996, . Barkley p (KSC, MO). Morelos: mins. Tres Mapas. is m, 20 Feb. 1932, Froderstrom 253 (F). Nuevo León: mpio. Zaragoza [E of Puerot Pinos on NE side of Cerro Pena Antonio Pena Ne Vae 2550 m, 24 48'N, 99°51'W, 25 Aug. 1989, G. Nesom 7125 N slope of Sierra San auum ca. 30 km NE 1971 m a, 2900 m, 12 Nov. (0, A. Cronquist 10907 (NY). Puebla: Mt. Popocatepetl, W-facing slope, 4100 m, 9 Feb. 1965, E. S a za (F, K AT lon, 1 km : ied Reed ski 407 ye AS, MICH). Sar : 5 rd. do C ue del rm 1200 m. å “SC, MICH m s ec Facenda de San Diego de Pinar, Mt. Malinche, : [e] Oct. 1938, Balls 5653 (F. GH, MICH). Veracruz: 1-2 kn above Ese poe. on i. NM slopes o of Cofre de Perote, mpio. Perote, 1¢ ik . 3300 m, 21 oo 1984, Taylor 166 (KSC, NY, s 5. Roldana anisophylla (Klatt) Funston, Novon 11: 305. 2001. Senecio anisophyllus Klatt, Leo- poldina 24: 124. 1888. TYPE: Mexico. Pelado, F. M. Liebmann 160 (holotype, C not seen, C drawing GH! C tracing MO"). Oaxaca: Roldana cronquistii H. Rob. & Brettell, Phytologia 27: wi 1974. Senecio cronquistii (H. Rob. € Brettell) B Turner & T. M. Barkley, Phytologia 67: 392. a: in wet fores en t 1 mi. or less S of pi tepec, 65 mi. N of Oaxaca, 9300 i. 11 p 1962, ^n psc 9648 (holotype, US not seen, US photos F!, GH! MO!; isotypes, GHI, KSC!, MICH!, NY!). Simple to sparingly branched suffruticose herbs or 0.5-2.0 m tall: densely pubescent with stipitate-glandular and mul- ticelled subshrubs, new growth sparsely to hairs, terete, striate, red or brown. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 4-8(—15) cm, glabrate to sparsely pubescent; blades 10-18 X 5—10(-15) em. triple-nerved, hastate, margins weakly secondarily lobed to denticulate, sparsely pubescent; abaxially elabrate with glandular hairs on veins, adaxially glabrous to sparsely pubescent with short glandular -~ hairs. Capitulescence simple paniculate cyme, ca. 10 capitula per branch, peduncles densely stipitate- glandular with sparse pubescense of long multicelled ultimate 2 turbi- hairs, primary bracts linear, 5-15 mm, 2. linear, ca. 4 mm. to 3, bracteoles ca. 15 x calyculate bracts O peduncles 10-25 mm, 5 mm. Capitula radiate, 10— nate, corollas yellow: linear, 2-5 mm; phyllaries 8, stipitate-glandular, 8-10 X ca. 2 mm. Ray florets 4 to 5, corolla ligulate, glabrous, tube 5-7 mm, ligule 10-13 X 2-4 mm. Disk florets 20 to 23, corolla weakly campanulate, glabrous, ca. 10 mm, 2 mm. tube ca. 4 mm, throat ca. 4 mm, lobes ca. Achenes glabrous, cylindrical, ca. 2 mm, ribs 5 or 10, resin glands absent; pappus bristles ca. 9 mm. Distribution and phenology. Southern Mexico (Oaxaca) at elevations of 2800-3000 m. pine-oak forests. Flowering August through December. Found in Comments. These are beautiful plants with sharp- ly hastate, deep green leaves and large, bright yellow i Brettell (1974) described the species Roldana cronquistit, in part, based on their rays. Robinson and agreement with Gibson’s (1969) circumscription of R. hederifolia (Hemsl.) H. Brettell reduced R. chrismarii. where he and R. anisophylla to synonymy. However, R. anisophylla is hob. aliena, Senecio a distinct species, and, on examining the protologue and type material of all species, R. cronquistu IS placed in synonymy under R. anisophylla. MEXICO. to Valle Representative specimens examined. 43 km N of Ixtlan-de Juarez jet, Oaxaca: Nacional. on rd. Annals of the Missouri Botanical Garden 2870 m | Sep. 1983, D. F. Me 00015 (CAS): carretera Ciudad de Oaxaca a Tuxtepec, cerca del Cerro Pelon, mpio. Yolox, 2850 m. 18 Jan. 20 30. Cháx zaro B. 5809 d. 40 € . 45 km NE of 17°35 10" N, 90 30'45"W, 2 320 m, 28 1781 (MO); Km 130 Oaxaca- MacDougall s.n. ( 0 tla, 17 » 31'20"W., 2650 m. 20 Dec. 1993, A. Rincón G. pu (MO). . Roldana aschenborniana (5. Schauer) H. Rob. & Brettell, Phytologia 27: 415. 1974. aschenbornianus S. Schauer, Linnaea 20: 698. 1847. TYPE: Valle Toluca, Aschenborn [716] (lectotype, designated here, B fragm. & tracing GH). Senecio Mexico. [stale unknown]: 5 j zalticus L. O. Williams. "v se 31: 446. 1975. . Williams) H. Rob., EU Der Quezi il- Sur. nov. Rec eg (e e 32: 331. 19 m go: mtns. SE of S on old rd. to San Juan 2550-2850 m. 21 Jan. i" P: Standley 84280 (holotype, Fl; isotype, MON. Phytologia 87: 235. 2005 eg hos nov. TYPE: Mexico. Oaxac 100 m del por la carre ole ra de e a D de Flores magon (Mex 182)" ca. 2320 m, 13 Fe . Mun- Estrada & Mendoza. 1947 (hololype, TE TEX digital TEX!; MIXI Roldana mazatecana B. L. Turner, Puerto de la “Aprox. image Isolvpe. nol secr i Shrublets or shrubs, 1-3 m tall; new growth reddish tomentose elabrate with woolly hairs: bark distributed on upper half of stem, lowermost decid- variously stems terete, brown. Leaves cauline, evenly uous; petioles 2-5 em, densely pubescent; blades 5— 12 X 3.5-10 em, triple-nerved, ovate, oval, or rotund, mostly 5- to 9(to 13)-lobed, sinus depth less than 1/4 of the way to the midrib to subentire, margins weakly serrate to denticulate; abaxially pubescent to lanate tomentose and canescent, adaxially glabrate, arach- noid hairs on veins. Capttulescence pyramidal com- pound paniculate cyme, 20 to 50 capitula per branch, peduncles lanate tomentose to sparsely floccose tomentose, primary bracts linear, 2-5 mm, ultimate peduncles 5-10 mm, bracteoles 0 to 3, linear to scale- like, turbinate, — mm, 3 to 5. 4—6(—7) Ray florets 6 to 8, corolla ligulate, ligule 4-6 X ca. 2 mm. Disk —8 mm, 0.5—] mm. 1-2 mm. Capitula. radiate, 6-10 corollas vellow: calyculate bracts linear, 1-2 mm; phyllaries ca. 13, glabrous, ca. ] mum. 3-5 mm. florets 12 to l6. tube 2— elabrous, tube corolla narrow, glabrous, 5 mm. throat 1.5-2 mm. lobes glabrous, cylindrical, | mm, ribs 10, 7 Achenes resin glands absent: pappus bristles 4-7 mm. Distribution and phenology. Eastern. Mexico and Guatemala (Guatemala: Chimaltenango. Huehuete- nango, Quezaltenango, Sacatepéquez; Mexico: Hidal- vo, Nuevo León, Oaxaca, Puebla, Querétaro. San Luis Potosí, Tamaulipas, Veracruz) at elevations of. 1000— 3200 m. forests. Flowering November through April. Found on gulf slopes in pine-oak and fir Schauer (1847) lists Aschenborn s. in the protologue for Senecio 718 specimen at GH that is a fragment and drawing of the Gibson (1969: 54) notes that J. M. Greenman on Liebmann 153 (GH) > Comments. from “valle Toluccana” aschenborntanus, and there is an Aschenborn presumed destroyed B type material. states “compares well with the type number Aschen- 718 Mus. Bot. Berlin.” further remarks that a drawing of Liebmann 153 (GH- that Aschenborn 718 is the type. There is a possible type herb. Gibson born’s no. Klatt) has a notation, perhaps by the artist, specimen al Kew: in Hemsley’s treatment of Senecio, he (1881: from Toluca. However, the specimen was not readily April 2007). hirsu- 236) reported seeing Aschenborn s.n. (K) found by Nicholas Hind (pers. comm., Roldana aschenborniana most resembles R. ticaulis and R. albonervia. For further elaboration, see the discussion sections under those species. Williams (1975) stating that il described Senecio quezalticus is closely related to Roldana barba- Johannis but differs in leaf shape ud pubescence from that species. While the deseription of 5. quezalticus certainly separates it from R. barba-johannts, it does not separate it from R. aschenborniana. He also notes that the into 5. da din were annotated. by Gibson as belonging two specimens he would put | R. aschenborniana. Upon examining the holotype and paralypes of S. quezalticus, it is placed in synonymy. Roldana mazatecana is placed in synonymy here as there is nothing to distinguish it from my concept of R. aschenborniana. However, that shallow. il may represent a variety leaf and a correlates rounded blades with ca. i acute lobes southern distribution versus a more ovate leaf blade with to 9 acute lobes and a northern distribution, de the typical variety. Representative specimens examined. 6 U ATE M; ALA. € hi- maltenango: Cerro de 2550 m. 26 Dec 3 2000 m. 6 Huehuetango: rd. Mar. 1933, . to San Pedro Soloma, Chichavae, MICH). of San Mate 0 Ixtatán, d of San Mateo Ixtatán, | edlove 8649 (V). 3900 m. i dh E | Santa María, iid Pow 2400— . 6 Mar. 1938, P. C. Standley 07593 (F). quez: slopes of Volcán de td a. above Mid María de Jesús. 2250-3000 m. 11 Feb. ? C. Standley 65185 (F): lower N slopes of Volcán e as [ia above Palojunoj. 2000 m. 1941. P. C. Standley 83436 (F. MO). MEXICO, > hwy. 85, p 236 San dais 2300 m. 30 Dec. 1954, Tono 3 2796 (V); 3 km al W de Zacualtipan, 2000 m. lO Jan. 1971. Diaz M. 270 (DS, MICH. MO). Nuevo León: Sierra Madre Oriental; Puerto Blanco to Sacatepé- I5 a aray, ca. 15 mi. Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana SW of Galeana, 23 July 1934, NY). Oaxaca: rd. Mueller 1211 (A, E, MICH Teotitlan del Camino to Huatla de Jimenez, 17.8-18.3 mi. E of Teotitlan, 2300 22 Feb. 1979, T. B. Croat 48210 (MO). Puebla: 6 km A NE de Zac un sobre la carretera a Cuetzalan, 1500 m, 2 Feb. 1975, J. I e Mie ASU). Derat barranc a de Amealco, 2350 m, 21 Feb. 1977, Arguelles 730 (C San Luis Potosí: e Hwy. 80, 16.2 mi. E of f El T from .5 mi. E of El Platanito, 17 km (by air) NE of ~ del Maiz ward = tiguo, Morelos, 22 a 99°28'W., 980 m. 22 Mar. 19 nr: 2809 MIC H. WIS); chiefly in the region of San nu Potosí, 2500 m, 20 F a 1905, Palmer 539 (F, GH, MO, NY). Tamaulipas: on e e rd. ca. 13 mi W of Ciudad Victoria, near summits of Sierra Madri 1000m, 9 M 194 McVaugh 9859 (MICH, NY) Dulces Nombres, Nuevo León, just E of border into Tamaulipas, 1690 m, 18 x 19. " e 2800 (G a MO). Veracruz: Choubla, Mar. 1910, Daidinas 521 (MO): hills above & NE of Xico Viejo, 8 km n of Xico, mpio. um 19°28'N, 97704'W, 1800 m. 5 Feb. 1984. M. Nee 29360 (KSC, NY). 7. Roldana barba-johannis (DC. H. Hob. & 415. 1974. Prodr. 6: 4230. 1837 México: Brettell, Senecio ipis -johannis [1838]. TYPE: E. M. Mairet designated by Phytologia 27: DC., Mexico. & G. McVaugh, microfiche IDC 800. MICH!, MO)). around Toluca, Andrieux 290 (lectotype. 1984: 812, G-DC not 1142 I 1!, G-DC seen, photos IDA Hartw. 18. PE: Mexico. [state unknown]: * 1839, as * “mexican O Senecio grahamii Benth., Pl. grahami.” TY nee s.n. (holotype, K not seen: isotype, GH!). Senecio pullus Klatt, Abh. Naturf. Ges. Halle 15: 333. 1882, non Senecio pullus Klatt, 1896, nom. illeg. TYPE: Mexico. [state unknown]: Cordillera Guchiloque, J. L. Berlandier 1177 (holotype, G-DC 534c not seen: isotype, G-DC fragm. € drawing GHN). Senecio donnell-smithii J. M. Coult., Bot. Gaz. 16: LOL. 1891. Syn. nov. Roldana donnell-smithi (J. M. & Brettell, Phytologia 27: 418. 1974. TYPE: Guate a Sacatepéquez: Volcán de Agua, 11,000 ft., . John Donnell Smith 2362 (holotype, K not seen: aa FL MOD. Branching woody shrubs or small trees, 1-6 m tall, new growth variously lanate tomentose; stems arising from a woody caudex, terete, reddish brown. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 4-8(215) cm, tomen- — tose to densely pubescent; blades 10-18 xX 5-10(-15 cm, pinnate to pinnipalmate, variously lanceolate, ovate, or oval, margins subentire, repand-sinuate or occasionally weakly 5- to 9-lobed, variously dentic- ulate; abaxially densely lanate tomentose and canes- cent, rarely weakly tomentose, adaxially glabrous to pubescent along veins. Capitulescence pyramidal compound paniculate cyme, 20 to 40 capitula per branch, peduncles persistently tomentose, primary bracts linear, 3-5 mm, ultimate peduncles 2-10 mm, bracteoles 0 to 2, linear, 1-2 mm. Capitula radiate or 7-10 X ca. calyculate eradiate, 3 mm, campanulate, corollas yellow; bracts ca. 5, linear, 1-3 mm; phyllaries 10 to 13, tomentose, arachnoid to glabrate with age, + purple coloration on phyllary tips, 4-5 X P es; l mm. Ray florets absent or 1 to 8, corolla ligule prominent or inconspicuous, glabrous, tube 2-3 mm, ligule 47 X ca. 2 mm, or when inconspicuous florets l to 3, ligule ca. 2 X 0.5 mm. Disk florets (10 to)13 to 18(to 21), corolla funnelform to moderately campan- ulate, glabrous, 6-9 mm, tube 2.5—4 mm, throat 2— 4 mm, lobes 0.5—2 mm. Achenes glabrous, cylindri- cal, 1-2 mm, ribs 5, resin glands present; pappus bristles 6-9 mm. Distribution and phenology. Central Mexico ex- tending southward into Guatemala along the Sierra Cuchumatanes and Honduras (Guatemala: Chimalte- nango, Huehuetenango, El Progreso, Quezaltenango, Sololá: Chiapas, Distrito Federal. Durango, Guerrero, Hidal- Morelos, Tlaxcala, Veracruz) at elevations of Sacatepéquez, Honduras: Lempira; Mexico: go, Jalisco, México, Michoacán, Oaxaca. Puebla, Sinaloa, 2400—4100 m. Found on steep humid mountainsides, understory of Also found on rich, dark volcanic soils, rocky ravines, and barrancas, often in dense shrubby secondary pine-oak forests, and in deep shade. erazed forests. Associated with Quercus, Pinus, Alnus Abies. Senecio, Solanum. Poten- tilla. Chi- ranthodendron. Flowering December through April. Penstemon, Physalis, Eryngium, Liquidambar, Cupressus, and Comments. In the protologue for Senecio barba- johannis, de Candolle (1837) made the notation “Mairet!” | agree with McVaugh’s (1984) selection of the E. M. Mairet & G. Andrieux 290 (G-DC) specimen for the lectotype. Roldana barba-johannis is closely associated. with R. lanicaulis, R. robinsoniana, and R. lobata due to its overall pubescence and general habit. When present. the lanceolate leaf shape is distinct among these four Standley, Greenman remarks (1926: 1627) that “the sap of the species. In his treatment of Senecio for thick stems is sometimes resorted to by travelers as a substitute for drinking water. Senecio grahamii and S. pullus were (1926); type material, they are placed in synonymy here as placed in T synonymy by Greenman after examining | well. Coulter recognized the species S. donnell-smithii based on those plants from Chiapas and Guatemala having a tendency to have involucral bracts that are As Gibson (1969) the | have found that they are not glabrescent and purple-tipped. noted, these characters are not restricted to southern distributions. consistent within the southern plants themselves. ;»ecimens from Guatemala have been found with lanate phyllaries that are not purple-tipped. Gibson Annals of the Missouri Botanical Garden proposed that 5. donnell-smithti be made a variety based on the southern distribution, essentially glabrous and purple-tipped phyllaries, leaves with markedly sinuate margins, and leaves pinnately to subpalmately veined versus those that are mostly pinnately veined. Plants from the south do have a tendency to have purple-tipped phyllaries; however, their pubescence shows the same variation as those plants from the north. The leaf blade of plants from the south does have a tendency to be more sinuate and whereas leaf blades from the north have a The distinction between pinnately to subpalmately veined ovale, tendency to be lanceolate and entire to subentire. leaves and mostly pinnately veined leaves is dubious at best. Therefore, the trend within the southern plants for leaf shape and purple-tipped phyllaries is not enough to separate the plants into a species or a variety. Representative no e *xamined. GUATEMALA. Chi- malte 'nango: "dr Je ; of AACE Calderas, 2800 m, 3 Jan. 1 930, Pu. S MO). e rd. t of San Mateo Ixtatán, mpio. San Mateo Ixtatan, 32 D. E. Breedlove 8049 (F). El Progreso: 3000 m, 21 Jan. 1942, J. Canton La 3200 m, Feb. 1965, Calera & summit of Volcán Siglo. A. Steyermark 43053 (MO). Quezaltenango: Esperanza about 6 km from San Juan Ostuncalco, 12-13 Jan. 1966, A. Molina R. 16598 (V) i of Los Monzo, mtns. above San Juan Ostuncaleo, 2800 m, 21 Jan. 41, P. C. Standley 84189 (E, MO). Sac o : Volcán de Aqua, Apr. 1890, Smith 2362 (F, MO). Sololá: Volcán Polimán, side facing Volcán Atitlán to summit, 3000 m, 13 June 1942, J. A. Steyermark 47558 (F, MO). HONDURAS. Lempira: Camino de El A be = Los Cedros Parque Celaque, 14°32'N, 2'W, . P. House 1199 (MO). ME ed C igs Bail al of Matsab, mpio. m japa, 2500 r . Breedlove 9242 (DS, F). Distrito Fede a ca "i mi. near crest of continental ae area, 25 mi. S of jet. Universidad & Churubusco. 2440 m. 28 Dec. 1970. Dunn 17246 (MO). Durango: along Mexico hwy. 40, 10.6 mi. E of El Palmito, ca. 57 mi. E of Cone a mpio. Pueblo Nuevo, 2500 m, 30 Dec. 1962, D. E. E 4264 (DS). Guerrero: mpio. Minas, 2200 m, 9 ns 1937, G. B. Hinton 10170 (GH, NY). 4.5 km E de mpio. Epazoyucan, 2850 m, 29 Dec. 1964, ( WIS). Jalisco: top of Sierra de Manantlán plateau, from N end of Sierra de Manantlán Central to Hidalgo: along lumber rd. \rroyo de Neverias at W end. of Sierra de Manantlán Oriental, 24 km E of Cerro La Cumbre, 17.5 km S of El Chante, 19 33'N, 104 11-13/W. 6 y 1080, A. He fltis 2547 (MICH, WIS). México: ca. 0.5 mi. W of summit on rd. from Toluca to Temascáltepec, 2800 m, 27 Dec. 1970, Di 17151 (MO): on S slope of Sierra de Ajusco, ca. 3—4 k e on rd. to Zempoalo, $ 2800 m, 14 July 1960, H. H. Ilis 224 (MICH, WIS). Mic vio d & T 3 km W of ( And pa NNW of . 3200 m. 20 oe if ine 4154 (F, MIC m Oaxaca: lumber rd. alone dee of Sierra, departing 2 N of Ixtlan, 2800 m. Ll Jan. 1970, 5414 (MICH, Anderson MO). Puebla: W-facing slope, 4100 m, 9 Feb. 1965, E. S Gibson 997 (F. KSC, MICH). Sinaloa: 2 air mi. NW of E | Palmito, N of Hwy. 40, very near Durango state line, mpio. Concordia i Occidental. ca. 105 51N Sanders 4926 » Pinar, Mt. Malinche. 3100 m. a2 Dec. 193t pn 5632 (F. € GH, MICI jel. d j»pec-Escola-Jacal rd., 14 qe W of Esc ola. ca. 19°07'N, 97 14W, 2900 m, 15 Nov. 1981, M. Nee 23109 (F, WIS) 8. Roldana cordovensis (Hemsl.) H. Rob. & Brettell, Phytologia 27: 417. 1974. Senecio cordovensis Hemsl., Biol. Cent.-Amer., Bot. 2: Mexico. Veracruz: Fôret de la 1800, E. Bourgeau 2020 (holotype, K not seen; isotypes, F!, P not seen, P photos MICH!, MO!. I 238. 1881. TYPE: Trinidad Cordova, 7 Mar. > drawing GH!). Bot. 2: 243. Orizaba. Biol. Cent.-Amer., 1881. Syn. nov. TYPE: Mexico. Botteri Es (lectotype, designated here, K not seen: GH! G not seen, G phot os FL MICH! MOD. ies peltata Sessé & Moc., Pl. Nov. Hisp. 132. 1887, non Cacalia peltata Kunth, 1818; 1895. TYPE: Senecio macrobotrys Hemsl., Veracruz: isotypes, m. illeg., syn. nov. nec Cacalia pelata | Rob. & Greenm., s. Se: Y Mexico. signated he re. Senecio cupiens S. Watson, Proc. Amer. Acad. Arts 25: E . 1890. Syn. nov. Roldana E prd (5. Watson) Rob. & Brettell, 6. 1974. TYPE: A xico. Jalisco: C as a de Sio 'aje, montañas ca. 1892, C. G. Pringle 2419 NY! 2 P hytologia 2T: (hol os G Ht: ole FL MICH MO! Senecio brachyanthus € ;reenm., Ann. Missouri Bot. rd. | 211. 1914. Syn. nov. TYPE: Mexico. Br pode e & Petatlán 1540- 2155 m, E. W. Nelson 2137 (lectotype, designated here, GHI; isotype, US not Ei Senecio chapalensis var. rur Greenm., Ann. Missouri Bot. Gard. 1: 278. 1914. . nov. TYPE: Mexico. bluffs s a wel canyon above 15 Feb. 1899, C. C. un ed GH!; Morelos: on a Cuernavaca, 1980 m, isotypes, Fl, (lectotype, ae here, NY!) MO! Shrublets to tree-like shrubs, 1-6 m tall; new growth pubescent with stipitate-glandular and multi- celled Leaves cauline, evenly distributed on upper half of hairs to glabrate; stems terete, green, red. stem, lowermost deciduous; petioles 3-12 cm, gla- 2) X triple-nerved, weakly pal- brate to pubescent; blades 6-17(22: 7-15 (222) em, eccentrically peltate, matifid, base truncate, 5- to 7(to 15)-lobed, sinus depth about 1/4 of the way to the deltoid, margins weakly denticulate: abaxially weakly pubescent on veins and margins, rarely with a light tomentum of multicelled hairs, adaxially glabrous to sparsely pubescent with short glandular hairs. Capi- tulescence pyramidal compound paniculate cyme, 10 to 30(to 50) capitula per branch, peduncles glabrate to densely pubescent with short stipitate-glandular hairs, Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana primary bracts sessile, broadly oval, 2-9 X 17 cm, 3-12 mm. 2-3 mm. X 3—4 mm, funnelform, corollas yellow ultimate peduncles bracteoles 2 to 3, filiform or linear, Capitula radiate or eradiate, 8-15 to whitish; calyculate bracts ca. 3, linear to filiform, 1—5(-10) mm; phyllaries (5 to)8, stipitate-glandular 1-10 X ca. 5, corolla ligulate or reduced to a tube, glabrous, pubescent, ca. 1 mm. Ray florets absent or tube 3—5 mm, ligule ca. 7 X 2 mm, or when reduced tube 2-3 mm. Disk florets 8 to 16, corolla funnelform, glabrous, 8-11 mm, tube 3—4 mm, throat 4-5 mm, lobes 1-2 mm. Achenes glabrous, cylindrical, 1— 2.5 mm, ribs 10, resin glands present; pappus bristles 7—9 mm. Distribution and phenology. Central and southern Mexico (Aguascalientes, Durango, Guerrero, Jalisco. México, Michoacán, Morelos, Nayarit, Oaxaca, San Luis Potosí, Veracruz, Zacatecas) at elevations of 1500-2800(—3500) m. including shaded bluffs of Found in pine-oak and fir forests, wet canyons. Flowering November through February. Comments. In the protologue for Senecio macro- botrys, Hemsley (1881) listed the syntypes Bilimek 560 (K not seen, F!, F fragm. F!) and Botteri 1103 (K, chosen here as lectotype), both from the same locality. In the protologue for Cacalia peltata Sessé & Moc.. Sessé and Mociño (1887) did not list any specimens: the lectotype is chosen from original Sessé & Mocino material at F. Greenman (1914) listed E. W. Nelson 2137 (GH, US) in the protologue as the type material for Senecio brachyanthus, and Greenman (1914) listed G; G. Pringle 8010 (GH, MO) in the protologue as the type material for S. chapalensis var. areolatus. In both cases, the GH specimen was chosen as the lectotype. Roldana cordovensis is a wide-ranging species. The and leaf blades is morphology of its ray florets variable. The ray florets may be completely absent from the capitula, reduced to tubes, or ligulate. While there appears to be a trend toward the reduction of ray florets in the southern range of the distribution, none of these forms can be correlated with a specific distribution. The shape of the leaf blade is highly variable; at times the blade has distinct acute lobes, and at others it has indistinct rounded lobes. There has also been some uncertainty as to whether the blade attachment is restricted to being eccentrically peltate, to the exclusion of having blades marginally attached to petiole. In my concept of this species, 1t 1s so restricted, Roldana cordovensis was described by Hemsley (1881) as having eccentrically peltate blade attach- ment and ray corollas either with or without a ligule. He also described Senecio macrobotrys as essentially an eradiate R. cordovensis. However, as Gibson (1969) noted, the isotype has reduced rays present. Senecio chapalensis was described by Watson in 1890 as having eccentrically peltate leaves and radiate capitula. After comparing the type material available and numerous collections annotated under all three — epithets, I have concluded that they are all within the fabric of R. cordovensis. Greenman (1901) proposed 5. chapalensis var. areolatus to call attention to those plants with areolae on the undersurfaces of the leaf blade and with smaller rays. In comparing the type material of the variety with other collections of the species, there is no such difference: due to the high variability of the ray florets in this species, the variety is placed in synonymy. Senecio brachyanthus is a form of R. cordovensis from Guerrero with only multicelled trichomes as opposed to the more common condition multi- of a combination of stipitate-glandular and celled trichomes. From my observations of the genus as a whole, these two types of trichomes are found in concert with varying densities, and the separation of a species based on this character alone is not tenable. Cacalia peltata Sessé & Moc. was desc ribed in 1887, and, after viewing a photograph of the type. it is placed in synonymy here. The 9055 and G. B. Hinton 11320 both have notes by the collectors that t collections Mexia — 1ey were white-flowered. Greenman (1926) noted that the rays of this species are whitish in the dried state. Due to and “senecionoid” charac- the mixing of “cacalioid” ters in this genus, it is not surprising to see a variation in corolla color accompany the variation in ray floret morphology. MEXICO. Aguasca- i de pu de en 10 5 lientes: ca. mi. SE ol lvillo, 3 hours bi horse from Rancho "s Los Adobes, Den m, d nie 195 . McVaugh 217 (MICH). Durango: 48 km of Huei El Alto, Jalisco on rd. to Canoas. ee Y zquital, 2530 m. 21 Oct. 1983, D. E. Breedlove 5912 AS). Guerrero: Mina, second ridge N of dee pe. m, l Jan. Mexia 9055 (CAS, ne GH, MO, NY): ca. 2km NE of rr El Gallo, ca. 17°28'N, 100 13 W, 2650 m, 27 Jan. 1965, J. Rzedowski 132 (DS, MIC H). Jalisco: ca. 31 rd. mi. W d Ayutla, ca. 70 mi. NW of Autlán, 2200 m, 3 Nov. 1962, A. kiap 9797 (GH, KSC, MICH. MO, NY). México: cerca de Amecameca. 2500 m, 25 Feb. 1968. J. ene di 25440 (DS. MICH, WIS). Michoacán: Rincón, . 1911, Arsene 6860 (MO); top of Bosques de ; of bona (ca. 20 km NE of 3500 m, 19 July 1960, H. ax above Cuernavaca, T Pringle 8010 (F, GH, MIC = EEN a > S P > ; ayaril: ine 8-9 km E of a hin, 12 E from entrance of is ri pic-Ixtlan, 21°27'N, 104745" W., 1340 m, 11 Nov. I9 G. Flores- D 3263 (MO). x 7 km NE of ru Larga, rd. betw. Piedra Larga Miahuatlan, distr. of Juquila, 1260 m, 22 Nov. 1982, Martínez S. 2750 (KSC, NY). San Luis Potosí: 22 km W of Santa Catarina on hwy. 86 at Km 49, 22 5'N, 100 40'W. 302 Annals of the Missouri Botanical Garden 0 km by rd. NW of LW, Zacatecas: 2200 m, 29 Sep. 1965, Roe 2185 er Veraeruz: 5.3 km W of Escola on rd. to Jacal, \ 19-10'N, 97 WIS). 4 ( S rU M. Nee rad (F, mplo. 1981, Coscomatepec, 2300 m, 12 Jan. ca. 25 km SW of Tlaltenango. ) km W of the jet. Jalpa— Juchipila, 2550 m, 24 Jan. t R. MeVaugh 25941 (MICH). hob. & 9. Roldana ehrenbergiana (Klatt) H. é Senecio Brettell, Phytologia 27: 1974. ehrenbergianus Klat, Leopoldina 24: 125. 1888, TYPE: Sessé & Moc. not seen). same as that for Senecio canicidus (neotype, designated here, MEXU Senecio canicidus Sessé & Moc., Fl. Mexic. (ed 2) 185. 1894. Syn. nov. TYPE: Mexico. Puebla: Sessé & Mociño s.n. (neolype, designated here, MEXU not seen, MENI photo MICH 17551). Senecio ~ Klatt, Ann. K. K. a Hofmus. 9: 303. 1894. nom. illeg.. svn. non e TYPE: Moxico. Is. seen, W Koch, 1861. OS W nol iS: ri I. Barkley, Mexico. Brittonia 30: 73. 1978. México: in moist swale on S of Mexico City, ca. 20 km NW 30 June 1974 Cronquist CASL GHI, KSC! Senecio semperamatae T. N Syn. nov. TYPE: slope of volcanic ridge S 1300 m, NY => of Cuautla, ca. 11130 (holotype. isotypes, QO!) Small single-stemmed or clustered perennial herbs. 0.5—] m tall, from a tuberous thickened fibrous rooted caudex, densely brown villous-lanate; plants glabrate l caudex patches of arachnoid pubescence : to weakly pilose, axils and under capitula, base of stem and densely white-villous; stems terete, striated, maculate, stems reddish. mature Leaves cauline, evenly distrib- uted on stem: petioles weak, 2-4 cm, sparsely pubescent; aie 50-10 X 4-7 em, pinnately veined and lobed, 5 principal lobes, sinus depth more than 1/2 of the way to the midrib to incised, lobes weakly to lobed, weakly cuneale and unequal, margin pube scent with strongly secondarily leaf base truncate to multicelled hairs; abaxially glabrate with multicelled hairs on veins to pilose, adaxially glabrate to sparsely pilose with multicelled hairs along veins. Capitules- cence variable from a simple panicle to a flat-topped capitula per branch, corymbiform cyme, l to 2 peduncles glabrate to pilose, primary bracts linear, 3-10 mm, ultimate peduncles 30-200 mm, bracteoles O to 2, filiform, up to 10 mm. Capitula. radiate, 10— 15222) X 5-10 mm, campanulate, corollas yellow; filiform, ea. 5, 3-5 mm; phyllaries 8-15 Ray and disk ae may have resin glands present. Ray florets 8, ulate bracts Í ca. 13, glabrate, 2.5—4 mm. corolla ligulate, glabrous, tube ea. 3-5 mm, ligule 10— 30 X 5-10 mm. Disk florets 50 to 60, corolla funnelform, glabrous. 7-13 mm. tube 2.5-5 mm, throat 3-5 mm, lobes 1-3 mm. Achenes glabrous, ribs 10, resin. glands absent: 2—3 mm, pappus bristles 5-10 mm. cylindrical, Distribution and phenology. Central Mexico (Mé- xico. Morelos, Puebla) at elevations of 1300-1600 m. Found in subtropical deciduous forests with Prunus, Ipomoea, Ficus, and Mimosa. Flowering May through June Berlin for ehrenbergianus was destroyed in 1943: Mexico. Puebla: C. Ehrenberg s.n. (B photo MICH!, B drawing GH!). No other type material was found at HAL by Det. 2004); from the senior Comments. The holotype at Senecio Uwe Braun (pers. comm., therefore, the neolype is chosen synonym S. canicidus Sessé & Moc. In the protologue of S. lobatus, Sessé and Mociño (1894) did not list any specimens; the lectotype was chosen from original Sessé and Mociño material at Roldana ehrenbergiana is a distinctive perennial herb with large capitula and deeply lobed pinnate leaf blades. Barkley (1978) created the species Senecio semperamatae based on its having somewhat larger capitula and involucre 12-15 mm high versus 6— 10 mm high. Due to the range of variation of involucre size that is present, one cannot base a separate species on this character alone. Afler examining the protologue and type material for both S. acerifolius Klatt and S. canicidus Sessé € Moc., they are placed in synonymy here. MEXICO. México: Representative specimens e EXE : 1 | Morelos: near México, 1893, Ranirez s.n. (GH). Ejido Palpan, Km 20 Miac atlan-T Mon mpio. Miacatlan, 1500 m, 5 June 1977, Garci 338 i pad Puebla: Valsequillo, Arnica del Pams, 8 May 106, Boege 74 (CAS): Hierba del Perro, hierba de la Pue vá 8 May 1961, Boege 561 (GH); vic. » San Luis vr Puebla, near Oaxaca, | June 1908, 519 (CAS, F, GH, MO NY). A. Purpus 2 Rob. «€ 10. Roldana eriophylla (Greenm.) H. Brettell, Phytologia 27(6): 418. 1074. Senecio ertophyllus Greenm.. Publ. Field Columbian Mus., Bot. Ser. 2: 282. 1907. TYPE: Mexico. 20 May 1906, C. G. GHI, hills near Tula, (holotype, CH Oaxaca: Pringle 13804 Hh). Isotypes, Simple or sparingly branched shrubs or small trees, —3 m tall; stems solid, new growth lanate, becoming glabrate; terete, striated. Leaves seasonally decid- uous, new leaves emanating from stem apex: petioles 3.05 cm, lanate to densely pubescent; blades 9-12 X 6-8 cm, palmate- to triple-nerved, ovate to oval, 7 to 11-lo midrib, lobes deltoid, acute or rounded at apex, leaf ved, sinus depth 1/8 or less of the way to de base truncate to cuneate; abaxially variously lanate and canescent, adaxially glabrate to densely pubes- Volume 95, Number 2 2008 Funston 303 Taxonomic Revision of Roldana cent canescent. Capitulescence pyramidal compound panicle with terminal cymes, 5 to 10 capitula per branch, peduncles lanate, primary bracts linear, 3— 5 mm, ultimate pedicels 3-8 mm, bracteoles ca. 3. Ww linear, ca. 2 mm. Capitula eradiate, 8-9 X ca. 3 mm, v funnelform, corollas yellow; calyculate bracts ca. 3, 2-3 mm; distally glabrous, purple coloration along midrib, 6-7 X 1-2 mm. Disk florets 10 to 12, corolla slightly P phyllaries 8, proximally lanate, linear, campanulate, glabrous, corolla 7-9 mm, tube 2- 4 mm, throat 1-3 mm, lobes 2—4 mm, lobes longer than throat. Achenes glabrous, cylindrical, 1.5- 2.5 mm, weakly 5-ribbed, resin glands absent; pappus bristles 6-7 mm. Southern Mexico 1000-2300 m. Found in pine-oak forests, growing with Quercus. Euphorbia L., Hauya DC., Cedrela P. Trichilia P. and Heliocarpus Flowering from May through October. Distribution and phenology. (Chiapas, Oaxaca) at elevations of Diospyros L., Browne, Browne, Comments. Roldana eriophylla has several fea- tures that make it unique: its corolla lobes are longer than the throat and the tomentose leaves are seasonally deciduous with new leaves emanating from the stem apex, creating a foreshortened stem. Both Bonnie Clark and José Luis Villaseñor recommend retaining it in Roldana, rather than transferring it to Pittocaulon or Telanthophora (pers. comm.. July 2002) dabo cd specimens examined. MEXICO. € hiapas: of Ciapa de : a Mauer above El mpio Breedlove 50639 (C alia 9000 (CAS . ; June 1906, Conzatti 1366 (GH): 5 km SE of — s & 9 km E of Torres 2431 (TEX). Corzo. D. E. c - P z A t e > a S: m = z3 Mam = Totolapan, carr. Oaxaca-Tehuantepec 11. Roldana gentryi H. Rob. & Brettell, Phytolo- gia 27: 418. 1974. Senecio gentryi (H. Rob. & M B. L. Turner & T. M. B . 1989. TYPE: arkley, Phytologia Mexico. Sinaloa: above La Surotato, pine-oak zone, 5000— 17-24 Mar. 1945, H. S. Gentry 7178 (holotype, US not seen, US photos F!, GH!, MO!, NY!; isotypes, F!, GH!, NY!). Senecio a popan B. L. Tur ; Phytologia 74: 383. 1993. 'n. nov. Roldana popa (B. L. Turner) B. 1 Lo Tuner, Phyiologia 87: 248. 2005 [2006]. TYPE: Mexico. Sonora: epos: Sierra a Río Mayo drainage, “U 935, H pper Sonoran; dark canyon,” 9 Mar. | Gentry 1411 (holotype, CAS! F!, MO 1089265! RIZ not seen; isotypes, Al, !). o 1094405 often tree-like, 2—3 m tall; new growth Shrubs, glabrate to densely pubescent with stipitate-glandular and multicelled hairs; stems terete, light green. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 6-13 cm, sparsely pubescent; blades 9-15 X 9-18 cm, palmate- to triple-nerved, shape various: cordate-deltoid with acute lobes; ovate, 5 to 7 shallow acute lobes; or palmatifid, 5 to 7 acute lobes, sinus depth about 1/4 of the way to the midrib; margins glabrous, weakly callous denticulate; abaxially glabrate with stipitate- glandular hairs along veins, adaxially glabrous. Capitulescence pyramidal compound paniculate cyme, jarsely t 5 to 25 capitula per branch, peduncles s - densely pubescent with stipitate-glandular and long i 1.5—4 X 0.5-2 mm, or linear, 2-5 mm, ultimate peduncles multicelled hairs, primary bracts sessile, oval, 5-17 mm, bracteoles 2 to 4, linear to scale-like, ca. 2 mm. Capitula radiate, 12-15 X 3-5 mm, campan- ulate, corollas yellow; calyculate bracts 0 to 3, linear to scale-like, 1-2 mm; phyllaries 8, densely stipitate pubescent, purple-tinted, 7-9 X 1-2 mm. Ray florets 4 to 6, corolla ligulate, tube 5-6 mm, ligule 6-10 X 5-2 mm. Disk florets 6 to 12, corolla funnelform, 9-10 mm, throat 4-5 mm, Achenes glabrous, eylindrical, ca glabrous, tube 2.5—4 mm, lobes 1-1.5 mm. 3 mm, ribs 10, resin glands absent: pappus bristles ca. 3 mm. Distribution and phenology. Northern Mexico (Chihuahua, Durango, Jalisco, Sinaloa, Sonora) at 900-)1500-2700 m. Found in pine-oak forests. Flowering November through March. elevations of Comments. Roldana gentryi most closely resem- bles R. cordovensis, but it differs by having marginally attached leaf blades and distribution restricted to northern Mexico. MEXICO. Chihua- thorn fore sl & barranc ‘an 35 Representative specimens examined. hua: E of La Bufa in ecotone betw. Batopilas, PS La 900 m. = Mar. 1979, R. Sanders De an ). dd 12 Mar . 1991, Cház aro B. 6579 "m I5). 8 ug 3 mi. NN Los pps rd. to $ Mocorito, Sierra S Feb. 1969, D. E. Breedlove 16529 (CAS Tacuic a 1500 m, 18 Feb. 1940, H. 5. Gentry 5059 (GH, MICH, MO "NY. Sonor sd Chiribo, Río Mayo, Upper Sonora, dank canyon, 9 Mar. 1935, H. S. Gentry 1411 (A, CAS, F, MO) 12. Roldana gilgii (Greenm.) H. Phytologia 27: 419. 1974. Senecio gilgit Greenm., Publ. Field Columbian Mus., Bot. : 282. 1907. TYPE: Mexico. Chiapas: near deis 6500-8000 ft., 8 Feb. 1896, E. W. Nelson 3772 (lectotype, designated here, US not seen, US digital image US!: isotypes, GH!, MO!). Rob. € Brettell, Annals of the Missouri Botanical Garden Suffruticose robust herbs. shrublets or shrubs. I— Em tall: new growth densely pubescent with stipitate- elandular and long multicelled hairs; stems terete, greenish brown. Leaves cauline, evenly distributed on upper half stem, lowermost deciduous; petioles 8-22 cm, densely pubescent: blades 10-30 10- 30 em, palmately nerved, rotund to reniform, 12(+)- lobed, sinus depth less than 1/4 of the way to the midrib, lobes acute, weakly secondarily lobed to irregularly serrate, margins pubescent with multi- celled hairs; abaxially moderately to densely pubes- cent with stipitate-glandular and multicelled hairs, adaxially pubescent with stipitate-glandular hairs. Capitulescence pyramidal compound paniculate cyme, 5 lo 25 capitula. per branch, peduncles densely pubescent with a mixture of stipitate-glandular and long multicelled hairs, primary bracts linear, ca. 5 mm, ultimate peduncles 5-30 mm, bracteoles 2 to 6. li 12-17 X ca. 5 mm. campanulate, corollas yellow; calyculate bracts 2—4 mm. ~ linear, Capitula radiate, to 3, linear, 2-3 mm: phyllaries 10 to 13, densely slipitate pubescent, 9-13 X ca. 2 mm. Ray florets 8 to 10, e m ligulate, glabrous, tube ca. 5 mm, ligule 3— 9 Disk florets ca. 30. elabrous. 9-10 mm, lobes ca. 2 mm ribs 10, pappus bristles 7-9 mm. —3 mm. corolla weakly le tube ca. 5 mm, throat 2-3 mm. Achenes glabrous, cylindrical, 3—4 mm, resin glands absent; Distribution and phenology. Southern Mexico and Guatemala (Guatemala: Chimaltenango, Guatemala, Huehuetenango, Quezaltenango, Quiché, San Marcos, Sololá, Suchitepéquez, Totonicapán; Mexico: Chiapas) 1600—3200 m. rainforests. Flowering January through March. at elevations of Found in montane Velson in the protologue; however, Comments. Greenman (1907) cited K. 3773 (GH, Us, B) appears to be a typographic in hand, the £. W. from GH and US as the ty] this error, as Greenman annotated, Velson 3772 specimens The Berlin Gibson selected the > collection. specimen is presume ed iun 'stroyed. ( lectotype, in hand, in 196 The most a ue of Roldana gilgit — are the large rotund to reniform multilobed leaf blades and long brown multicelled trichomes. GUATEMALA. Chi- Representative specimens examined. maltenango: rd. to Panajachel betw. Los Idolos & Chocoyosm, 2400 m, 12-23 Jan. 1966, Molina R 16189 (F). Guatemala: slopes of Volcán i Pacaya, betw. San Francisco Sales & the base of the active cone, Dee. 1940, P. C. nA 8004.3 on Hu about. laguna bs Ocubila, E of Huehuetenango. 1900 m, 7 Jan. 1941, P. C. Standley pes (F). Qu of Volcán sd 2 at & above Aguas Feb. 1936 Standley 65207 canyon de Madre Mins. 2000 m, 20 ichuetenango: DO zallenango: E 2050 m, 17 Quiché: ue os (MO). betw. Sierra Los Encuentros & a nN 1900 m. 22 Dec. 1972. Williams 4108€ (E, MICH). San Mareos: betw. Todos Santos & Finca El diee ds v Jo middle slopes of Volcán Tajumulco, 2300 m. 1 Mar. 1940, J. A. Steyermark 37038 (V). Sololá: ravines near Nahuala. Sierra Madre Mtns.. 2500 m. 17 Dec. 1962, Williams 23187 (F. NY A Suc iu 'péquez: Volcán Atitlán, S side, 2200 m, 14 x» 935. A. F. Skutch 2151 (V, 10. NY). Totonicapán ri qus San Francis sco El Alto & Momoste 'nango. 2000 m, 19 Jan. 1941, P. - Standley Spr (F). MEXICO Chiapas on i a side of Volcán Tacaná above | alguin; 2200 m, 5 Mar. 2D. E Bradio 24411 (DS, MICH, MO). 13. Roldana glinophylla Gibson ex H. Rob. & Brettell, Phytologia 27: 419. 1974. Senecio acerifolius Hemsl., Biol. Cent.-Amer., Bot. 2: 235. 1881, nom. illeg., non Senecio acerifolius K. Koch, 1861. Senecio glinophyllus (H. Rob. & Brettell) MeVaugh, Fl. Novo-Galiciana 12: 823. 1984. TYPE: "Oaxaca," Mexico. Michoacán or Oaxaca: Ghiesbreght s.n. (holotype. K not seen). Simple or sparingly branched. perennial herbs, ca. — m tall, plants glabrous: stems terete, finely striated, vellowish green. Leaves cauline, clustered at the midstem, reduced upward; petioles 3—4 em. glabrous: blades 6-10 X 606-10 em, papyraceous, palmately nerved, palmatifid, 5-lobed, sinus depth 1/2 to 1/4 of the way to the midrib, lobes acute, leaf base truncate, margins without callous denticles: surfaces glabrous. Capitulescence ascending simple to compound. cor- ymbiform cyme, sometimes giving an umbellate appearance, 8 capitula per branch, peduncles a glabrous, primary bracts minutely foliaceous with 3 divaricate lobes, the lowermost secondarily lobed or linear, ca. 3 mm, ultimate peduncles 2-25 mm, bracteoles O to 3, linear, 4-7 mm. Capitula. radiate, 10-15 X a campanulate, corollas yellow: calyculate bracts 5 to 8, linear, 3-8 mm; phyllaries 13. glabrous, 10-11 X ca. 2 mm. Ray florets 5 to 8, corolla ligulate. glabrous, tube ca. 5 mm, ligule ca. 5 xX 2.5 mm. a florets ca. 30, corolla glabrous, ca. tube ca. funnelform, mm, 3.5 mm, throat ca. 3 mm, lobes ca. L2 mm. Achenes glabrous, cylindrical, ca. 2 mm, ribs 10, resin glands absent: pappus bristles 6—1 mm. Distribution and phenology. Mexico (Michoacán) 1000-1500 m. oresls. Collected in flower in September and October. al elevations of Found in pine-oak == Comments. Hemsley (1881) lists Ghiesbreght s.n. (K) from southern Mexico, Oaxaca in the protologue. McVaugh (1984: 823) the Ghiesbreght s.n. (K) specimen, and that it is the type notes that he saw for both Senecio acerifolius Hemsl. and $. glinophyl- Ile that duplicates erroneously lus. also notes many of Ghiesbreght's were labeled. from Oaxaca Volume 95, Number 2 2008 Funston 305 Taxonomic Revision of Roldana when they were in fact from Michoacán or from other r parts of Mexico. There is another set of possible type material at GH, a note by Greenman, 19 Sep. 1904. in “Ghiesbreght 371, Hemsl. verified at Kew with the original.” Roldana perennial herbs found along the Belt. corymbiform capitulescence make it unique. hand, reads Senecio acerifolius glinophylla is one of the glabrous Trans-Mexican Volcanic The star-shaped leaf blade and MEXICO. Michoa- temperee, Ghiesbreght 371 (GH, distr. B. Hinton Represe ntative specimens examined. cán: near terre MICH. NY) Pras 12220 (GH, MICI Y). Iruapan, Coalcomán, G. — 14. Roldana gonzaleziae (B. L. Turner) B. L. Turner, Phytologia 80: 278. 1996. Senecio gonzal- Phytologia 57: 377. Roldana gonzaleziae (B. L. Novon 11: 304. 2001, as isonym. TYPE: Mexico. Durango: mpio. de Mez- quital, 3 km S de Sta. Ma. de Ocotan, 17 Oct. 1984. M. Gonzalez & S. Acevedo 1558 (holotype. TEX not TEX digital image TEX!). 1985, as Turner) “gonzalezi,” Turner, ^ eziae B. L. ‘gonzalezae. Funston, seen, Perennial herbs, acaulescent, to 60 em tall, arising from a fibrous rooted caudex, typically lanate; plants glabrate to sparsely tomentose; stems terete, striated, reddish. Leaves clustered at base; petioles 2-4 cm, pubescent to pilose; blades 5-7 X 5-9 cm, subpal- mately nerved to triple-nerved, oval, weakly 8- to 13- lobed to crenate, margins pubescent; abaxially pilose to lanate pubescence, adaxially sparsely pubescent with glandular hairs. Capitulescence flat-topped cor- ymbiform cyme, atop a 9.5-55 cm naked scape, ca. 15 capitula in the entire capitulescence, scape floccose tomentose, primary bracts linear, up to 5 mm, ultimate bracteoles O to 3, peduncles 0-10 mm, linear, l- 3mm long. Capitula radiate, 7-9 X 2-3 mm, turbinate, corollas yellow; calyculate bracts ca. 3, linear, 1-2 mm; phyllaries 8, glabrous to arachnoid floccose, greenish, 5-6 X ca. 1 mm. Ray florets 5 to 8, corolla ligulate, glabrous, tube ca. 3 mm, ligule 4-6 X 2-3 mm. Disk florets 9 to 16, corolla funnelform, glabrous or minutely pubescent, 5—6.5 5 mm, tube 2.5— 3 mm, throat 1-2 mm, lobes 1.5-2 mm. Achenes glabrous, cylindrical, ca. 1.5 mm, ribs 10, resin glands absent; pappus bristles 4-6 mm. Distribution and phenology. North-central Mexico (Durango, Jalisco) at elevations of ca. 2700 m. Found in dry pine-oak forests. Collected in flower September and October. Comments. Roldana gonzaleziae is one of the few small perennial herbs in the genus. It closely resembles R. sundbergii but is separated from it by f form C O having shorter scapes and a different capitulescence. They are separated geographically: R. sundbergii has only been collected in the Sierra Madre Oriental of Nuevo León, whereas R. gonzaleziae is found further south in the Trans-Mexican Volcanic Belt. Duran- Representative specimens a MEXICO. | i S . 1558 ( Aon image TEX). Jalisco: 74 dm WNW of Huequilla El be Jalisco Canoas, mpio. ~ 2700 m, 22 Oct. 1983, D. E. Pr 59162 (CAS Roldana greenmanii H. Rob. & Brettell, Phytologia 27: 419. 1974. Senecio greenmanii H. Rob. & Brettell) L. O. Williams, Phytologia 31: 441. 1975. TYPE: Guatemala. Santa Rosa: — Pueblo Nuevo, Tetalhuleu, 800 m. “Mano de leon," edible, Mar. 1923, R. Stricker 359 (holotype, US not seen, US photos F!, GH!, MO!" NY !). Robust. suffruticose herbs, shrubs or small weak trees, up to 8 m tall; new growth densely pubescent with stipitate-glandular and long multicelled hairs: Leaves evenly distributed on upper half of stem, lowermost stems terete, striate, green or red. cauline, deciduous; petioles 10-15 cm, pubescent; blades 7— 30 X 10-25 em, palmately or triple-nerved, orbicular, 9- to 13-lobed, sinus depth about 1/2 of the way to the midrib, lobes deeply cut producing almost rectangular segments topped with 3 acute lobes, upward on stem blades becoming subpalmatifid, margins always denticulate; abaxially pubescent with short and long multicelled hairs, densely so on veins, adaxially sparsely pubescent with short glandular hairs. Capi- tulescence a loose paniculate cyme, 7 to 20 capitula per branch, peduncles pubescent with stipitate- elandular and long multicelled hairs, primary bracts obovate to oval, up to 10 X 5 mm, ultimate i uei les 15-30 mm. Capitula radiate or eradiate, bracteoles 2 to 3, —3 mm. 11-15 X 4-5 mm, calyculate bracts absent eylindrical, corollas yellow: or ca. 3, linear, ca. 2 mm; phyllaries ca. 8, sparsely stipitate pubescent, greenish purple-tipped, 9-11 X 1-2 mm. Ray inconspicuously ligulate, glabrous, tube ca. 4 mm, 3 X 0.5 mm. Disk florets 18 to 21, corolla florets absent or ca. 4, corolla ligule ca. narrow, glabrous, ca. 10 mm, tube ca. 5 mm, throat ca. 4 mm, lobes ca. l mm. Achenes glabrous, 4 mm, absent: ribs 10, pappus bristles 7-9 mm. cylindrical, ca resin. glands Distribution and phenology. | Southern Mexico and Guatemala (Guatemala: Alta Verapaz, Quezaltenango, jan arcos, Santa Rosa; Mexico: Chiapas) at elevations of 1800-2 forests. Flowering February through April. 2300 m. Found in montane cloud 306 Annals of the Missouri Botanical Garden Comments. Roldana greenmanii is a very distinct The producing almost species unlikely to be confused with any other. rotund leaves are deeply cut, rectangular segments that are topped with three acute lobes. with delicate The capitulescence is open peduncles that have relatively few capitula. Repre sentative specimens examined. Verapaz: vie. of Coban. 1300 m. 90921 (F). anta Mana ^ GUATEMALA. i Mar.—Apr. 1941, P. Quezaltenango: along Río e Se sús, 1600 m, 25 Jan. 1941, P. C y Mare ‘OS: Barran “a Er minencia, rd. = c 19 - e. barranca betw. Fi inca La Lucha & Buena Vista, 2600 m, 6 Feb. 1941, P. C. Standley 86357 (F). Sa la Rosa: Pueblo Nuevo, Te m M u, BOO m, Mane de leon.” Mar. 1923, R. Stricker 359 (US inis FK. GH, MO, NY) MEXICO. Chiapas: third ridge along enn rd. from las Margaritas lo C am Alegre, mpio. La Inde pe E neta, 2300 m. 18 Feb. 19 . E. Breedlove 33710 (DS, F. MICH, MO, NY). 16. Roldana grimesii (D. L. Turner) C. Jeffrey. Kew Bull. 47: 55. 1992. Senecio grimesii B. L. Turner, Brittonia 40: 81. 1988. TYPE: Mexico. Hidalgo: 45 mi. N of Zimapán along hwy. 85, 14 Mar. 1983, B. L. Turner 15088 (holotype, TEX not seen, TEX digital image TEX!: isotype, XAL not seen). 1958. Syn. Senecio ips "sut B. L. Turner, acies e ): a nov. Roldana ros (B. Turner) C. Jeffrey, Kew Bull. 47: 55. 1992. TYPE: ve ico. Venen Saltó del Gato, 4 SM al este de Xalapa, 19 35'N, 96 33'W, ca. 1500 i os 2: VEU "aducifolio. secundaria, suelo rojizo 5 Apr. 1972, W. Marquez R. 24 Ge TEX not seen, TEX digital image TEX!; isotypes, KSC MEXU not seen, TEX! XAL not seen). Senecio nesomiorum B. L. Turner, Phytologia 66: 461. 1980. Syn. nov. Roldana ne 'esomiorum (B. L. Turner) € Jeffrey, Kew Bull. 47 . 1992, TYPE: Mexico. Nuevo León: mpio. Doctor ae. ca. 35 mi. NE of Doctor \rroyo NE of San Antonio, Pena Nevada, N side of Cerro Peña Nevada, rd. ca. 2 mi. N of “El Puerto,’ 23 45'N, 90 52'W, 2650 m, 15 Sep. 1988, G. Nesom 0712 e TEX not seen, TEX digital image TEX: isotype, MEXU not seen). Shrublets or shrubs, to 1.5 m tall new growth greenish brown villous. covered with a mixture of short and long (up to 2 mm) glandular multicelled hairs; stems terete, bark dark brown. Leaves cauline, evenly distributed on upper half of stem, lowermost pubescent with short 18 X 22 em, 9- to 13-lobed, about 1/8 of the way to the midrib, lobes acute, weakly deciduous; petioles 7-15 em, and long multicelled hairs: blades ca. triple-nerved, ovate-orbicular, sinus secondarily lobed to denticulate: abaxially hirsute to velutinous hirsute, margins pubescent, adaxially elabrate to weakly hirsute. Capitulescence compound paniculate cyme, 3 to 10 capitula per branch, peduncles villous, glandular, primary bracts broadly 10 mm, oblong, ca. 20 ciliate margins. of long multicelled hairs, hairs 1.5-2 mm, ultimate peduncles 30-35 mm, bracteoles 20 mm, Capitula radiate or eradiate, 12-16 X elliptic-lanceolate, up to ciliate margins of long multicelled hairs. 5-10 mm, campanulate, corollas yellow; calyculate bracts 3 to 6 ml or oblanceolate, 10-14 mm, 2 apparent series, glabrate, ciliate margins of long multicelled hairs: phyllaries 9 to 13, stipitate-glandular pubescent, )J11 Ray absent or (5 to)9 to 13, corolla ligulate or reduced to a margins ciliate, 1 a. 2 mm. florets tube, pubescent, tube 4-8 mm, ligule 9-10 X ca. 3 mm. Disk florets 25 to 65, corolla campanulate, pubescent, ca. 10 mm, tube ca. 5 mm, throat. ca. lobes Ca. | 4 mm, mm. Achenes glabrous or pubes- cent, cylindrical, ca. 2 mm, ribs 10, resin. glands absent: pappus bristles 8—10 mm. Western Mexico (Hi- dalgo, Nuevo León, Puebla, Veracruz) at elevations of D Distribution and phenology. 1500-2700 m. Found in pine-oak forests, scattered in an oak woodland in relatively level terrain. Flowering February through April. Comments. — Roldana grimesii is densely pubescent with both stipitate-glandular and long multicelled hairs. » brown, The pubescence is dark-purple somewhat like the pubescence found on R. metepeca Jeffrey. T lescence and the calyculate bracts are oblong-ovate -— Turner) Phe bracts of the capitu- with ciliate margins. The only other species in which this occurs, on occasion, is R. angulifolia. Roldana grimesii was described from Hidalgo as having eradiate capitula and glabrous achenes. Turner (1989) described R. marquesii from Veracruz and separated it from œ. grimesii based on its having radiate capitula. and pubescent achenes. He further separated R. nesomiorum based on its having radiate capitula and glabrous achenes, from León. examination of the Nuevo ased on my collections at hand, the above correlation of charac- ters does not exist. Achene pubescence is variable and may be related to maturity of the capitula and, there is no while rays are either absent, present distributional relationship to this phenomenon. All three of these entities are alike with respect to their and habital characteristics; other morphological therefore, they are synonymized under the oldest name. Representative specimens examined. MEXICO. Hidalgo: Pachuca & Tampico, 27.2 mi. S of 0.9 mi. N of Santa María, Feh. 1987. : mpio. Doctor / Arroyo, NE of San 2650 m. 15 Sep. 1988, C. Vesom 671 2 (TE X digital image TEX). rod pe o, 2 Apr. 1967, [TA 536 (CAS). Veraeruz: on hwy. 131, 13 mi. along hwy. 105 betw. S of Quebradora, 98 34 W, 1315 Volume 95, Number 2 200 Funston Taxonomic Revision of Roldana from jet. with hwy. 141, 1500 m, 29 Apr. 1986, LaFrankie 078 ( MO) 17. Roldana guadalajarensis (B. L. Rob.) H. Rob. & Brettell, Phytologia 27: 420. 1974. Senecio Rob., Proc. Amer. Acad. TYPE: Mexico. Jalisco: cerca guadalajarensis B. L. Arts 26: 166. 1891. Lugares fertiles de la Barranca, Guadalajara, 10 Sep. 1890, C. G. Pringle 3280 GH!, GH fragm. & e (holotype, GH!; isotypes, Fl, drawing GH!, MO!, NY!). tall, . 2.5m plants essentially [om Single-stemmed perennial herbs arising from a woody caudex, glabrous, stipitate-glandular hairs on leaf blades and occasionally with fine hairs on peduncles: terete, with rough vertical striations, reddish brown. stems Leaves cauline, clustered at midstem, reduced upward; petioles 1-3 cm, glabrous; blades 12-25 X 2-6 cm (making blade 7-10X longer than wide), pinnately veined, linear-oblong, margins serrate to only callous hairs sometimes denticulate, — stipitate-glandular present; abaxially glabrous to having a covering o — stipitate-glandular hairs on veins, pale, adaxially essentially glabrous. Capitulescence flat-topped com- pound corymbiform cyme, 5 to 30 capitula per branch, peduncles essentially glabrous, occasionally with tufts of fine hairs subtending the capitula, primary bracts linear up to 5 mm, ultimate peduncles 40—150 mm, bracteoles linear, 1-2 mm. Capitula radiate, 8-10 X 2—4 mm, campanulate, corollas bright yellow to golden; calyculate bracts ca. 5, linear, 2-3 mm: phyllaries ca. 12, glabrous, purple-tinted, 5-9 X ca. Ray florets ca. 5, corolla ligulate, glabrous, tube 3-5 mm, ligule 6-8 X 2-3 mm. Disk florets 16 to 19, corolla funnelform, glabrous, 6.5-9 mm, tube 2.54 mm, throat pu L5 Achenes pubescent with fine hairs, cylindrical, ca. 2 mm, ribs 10, resin glands absent; pappus bristles 4— | mm. — ^ 2.5-3.5 mm, lobes mm. 6 mm. Distribution and phenology. Western and central Nayarit) al — Mexico (Guanajuato, Jalisco, Michoacán, elevations of 1000-2200 m. Oak-pine zone, on grazed and cut mountainsides, may be found on open woodlands, waste places, and grasslands. Associated with Quercus, Pinus, Salvia, and Cercocarpus Kunth. Flowering July through October. Comments. | Roldana guadalajarensis is one of the few species with pinnately veined leaves in the genus. The leaves are further distinguished by being typically 7-10X longer than wide. This species is also one of the few possessing a flat-topped corymbiform cyme for its capitulescence type Representative specimens examined. MEXICO. Guana- juato: De Moro Leon, 10 July 1900, Duges 1005 (GH). Jalisco: 32 km Colima, 2-3 km SW of El Terreo, 19725'50"N, a b 7 OW. 2100 m, 22 Sep. 1987, Cuevas 2480 (WIS). Michoacán: grazed woodland N of Los Reyes, 1200 m, 28 Oct. 1962, R. McVaugh 21962 ieri Nayarit: along unused rd. from Mesa del Navar to Santa Teresa, mpio. El Nayar, 1800 m, 12 Aug. 1980, D. E. 20 45420 (CAS). 18. Roldana hartwegii (Benth.) H. Hob. & Brettell, Phytologia 27: 420. 1974. Senecio hartwegii Benth., Pl. Hartw. 18. 1839. TYPE: Mexico. Jalisco: 1839, Hart- weg 124 (holotype, K not seen, K photos Fl, MICH; isotypes, GH!, NY!). In sylvis Bolaños, Senecio seemannii Sch. Bip. in Seem., Bot. Voy. Herald 311. 1856. TYPE: Mexico. [state unknown; according to Seemann's travel agenda, this either western Durango, adjace collection came from nt Sinaloa, or N ierra Madre, B. Seemann 2010 (holotype, GH!; Vo eR K not seen, P not seen, K photos F!, MICH!). Cac alia tepicana M. E. Jones, Contr. W. Bot. 15: 156. 1929. exico. Nayarit: Tepic, 10 Feb. 1927, Marcus 2. Jones ; 23357 (holotype, POM not seen; isotype, US!). Roldana e H. E & Brettell, Phytologia 2 n. 422. K necio e B. L. Tumer & 1989, non Sere a Mex Barranca below Sar E. Station, 7000 fi., n G. Hil 13568 (holotype, US not seen, US sliotos P. GH!, MO!, NY! isotypes, CAS!, GH!, MICH!, MOL, NMC). T. M. Pues Divi 67: < pennellii Greenm., 1923. TY Pes a aris- . 1-2 m tall, ing from a woody caudex with Er woolly buds Single-stemmed perennial herbs and/or with branching rhizomes; new growth arach- noid pubescent; stems terete or angulate, purplish red or green, striate or maculate. Leaves cauline, evenly distributed on upper half of stem, lowermost decid- uous; petioles 3-7 cm, glabrous to sparsely arachnoid pubescent; blades 8-15 8-17 cm, palmate- to triple-nerved, oval to ovate, 9- to 15-lobed, sinus depth about 1/4 of the way to the midrib, lobes acute, weakly secondarily lobed to irregularly serrate- denticulate or rarely orbicular and subentire; abaxi- ally glabrous to sparsely arachnoid pubescent on veins, rarely densely arachnoid pubescent, adaxially glabrous. Capitulescence compound paniculate (rarely 10 to 30(to 50) capitula per y arachnoid pubescent, corymbiform) cyme, — branch, peduncles sparse primary bracts linear, 1-3 mm, ultimate peduncles 3-10 mm, bracteoles 0 to 3, linear, 1-2 mm. Capitula radiate or eradiate, 6-10 X ca. 4 mm, Pe ARE corollas yellow; 1- 2 mm; phyllaries (5 to)8 to 13, oid lo sparsely arachnoid pubescent, greenish purple-tipped, 4-7 X 1-2 mm. Ray florets absent or (3 to)5 to 8, corolla ligulate, pubescence of glandular hairs on tube, tube ca. 4(—5) mm, | —6) X 2-3 mm. Disk florets 6 to 12, corolla funnelform, pubescence of glandular calyculate bracts 1 to 3, linear, ==, igule ca. 4( Annals of the Missouri Botanical Garden hairs on tube, ca. 8 mm, tube ca. 4 mm, throat ca. 3 mm, lobes ca. | mm. Achenes glabrous or pubes- cent, cylindrical, 1—2 mm, ribs 10, resin glands absent; pappus bristles 4—6 mm. Distribution and phenology. Southwestern United States, Mexico Durango, — northern and west-central Mexico, (Mexico: Jalisco, southern Chiapas, Chihuahua, Nayarit, United Mexico) at elevations of Coahuila, Oaxaca, New Found in Nuevo Leon, States: Arizona 1200-2700 m. pine-oak forests and thickets near streams. Flowering Sinaloa, Sonora: and August through November. Comments. In the protologue for Senecio hart- wegii, Bentham (1839) does not list any specimens: (1881: treatment of however, Hemsley 241) lists three specimens from Kew in his Senecio, including Seemann 2010, Gregg 622, and Hartweg 124. It is presumed that Hemsley would be examining Ben- tham’s type material: based on this and the K photographs, the lectotype is chosen here. Roldana hartwegii is a common species along the northern mountain. chains and western end of the Belt of morphology, it is consistent and easily recognizable. the variation found in the number of Trans-Mexican Volcanic Mexico. In gross there has been of the However. disagreement over significance phy Ilarie s, pubescence of the achenes, and morphol- ogy of the stems. Upon performing a numerical the Materials and Methods section), | analysis of complex on 64 collections (see have concluded that there is a recognizable trend in the correlation of Plants with 10 to 13 pł pubescent achenes are distributed along the Pacific these characters. wllaries and Trans- Mexican Volcanic Belt; plants with ca. 8 phyllaries slopes of the Sierra Madre Occidental. and Qu istributed on the inner slopes of the Sierra Madre Occidental and Trans-Mexican Volcanic and pubescent or glabrous achenes are « Belt, and on the northern end of the Sierra Madre Oriental; lastly, plants with 5 phyllaries and pubescent or glabrous achenes are seemingly restricted to southern Durango. Concolor- ous stems occur in steppe and desert climates. while maculate stems occur in tropical and midlatitude rainy climates. Angulate stems are associated with tropical rainy climates. while terete stems are associated with the midlatitude rainy climates and steppe and desert climates. Over the years, several authors have recognized this variation by circumseribing new species, but in gross morphology and general habit, all of the collections examined are deemed to be the same species. Roldana hartwegii Bentham in 1839 as arachnoid tomentose, was originally described by having stems floccose to leaves glabrous adaxially and - to 4- raved, 8 to 10 disks, and achenes us: etl ; | B white pubescent abaxially, 6 to 8 phyllaries, examined the protologues and type material for Senecio seemannii and Cacalia tepicana, they are easily placed in synonymy here, as was done by McVaugh (1984). Robinson and Brettell (1974) described KR. pennellii as having eight phyllaries, glabrous achenes, concolorous stems, and distribution in the interior of Mexico. In general, these characters occur within R. hartwegii as circumscribed here; achene pubescence in this region is quite variable, and therefore does not warrant taxonomic recognition. KEY TO THE VARIETIES OF ROLDANA HARTWEGI la. Phyllaries 5; plants e southern Durango .. aE abd 8b. R. hartwegii var. durangensis lb. Phyllaries 6 to 8 or or s 10 to 13. 2a i llaries 10 1 ladre Oriental 3: plants from the Sierra Es 2b. Phyllaries 6 to A Phyllaries ca. 5 X | . R. hartwegii var. carlomasonii mm: abaxial leaf pubescence glabrate: plants found in the southwestern United States. north and west-central Me ON TEN Ja. mu var. 3b. Phyllaries ca. 7 X M mm; pubescence pen plants found in hartwegii abaxial leaf Oaxaca and Chiapas .......... 18d. R. hartwegii var. subeymosa 18a. Roldana hartwegii var. hartwegii Distinguishing characters. Stems typically terete, occasionally angulate, concolorous or maculate leaves abaxially glabrate; phyllaries 6 to 8, 4—5.5 ca. | mm; achenes pubescent or glabrous. Distribution. Primarily along the inner slopes of the Sierra Madre Occidental and northern end of the Sierra Madre Oriental: Ja- lisco, Nayarit, Nuevo León, Sinaloa, Sonora, Zacatecas. Chihuahua, Coahuila, Durango, Representative specimens examined. MEXICO. Chihua- hua: Tarahumara sect. of Sierra Madre Occidental, ca. 5 km NE of Memelichi. € ca. 20km SE of Cascada de Basaseachie, 2200 m, 25 Aug. 1980, A. Cronquist 11761 . Coahuila: cs Tn S oahuila, higher elevations in the Sermi Jardin, 2800 m, . 1966, Flyr 1182 (MO); middle & upper reaches d» Canon ui ds Hacienda, S of Rancho Cerro de la Madera, N slope of Sierra de la Madera, 27 3—4' N. '"W. 2000 m. 21 Sep. 1972, Chiang 9452a (TEX). Durango: steep slopes at base of Espinosa del Diablo, 4 km NW of T e id. betw. Mazatlán & s 2500 m, 28 Oct. 1973, D. E. Breedlove 35753 (CAS, MO): Madre m c aa . 10 mi. W of El Salto, 1962, / Cee 9587 (G, KSC, MICH, MO, mpro. Bulsnds. 25 km W of Bolanos, camino 9823. Lott 2064 ( Sierra 200 m, 2 Oct. ^ No) ~ — mmn Huic holes, 2530 m. KSC, l, 10440'W, 6 Nov. 1950. S 590 ME XU. MO). Nuevo León: Sierra Madre Oriental, —. Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana La Encautada, 23°55'40"N, 99750'50"W, mpio. Ee L Zaragosa, 2250 m, 28 Aug. 1975, Robert s.n. (MICH). a on small dirt logging rd. near Loberas Microwave l 2000 m, 18 Oct. 1983, D. E. A 58848 ). Sonora: region of the Río de Bavispe, NE Sonora, E i TE del Pilar, 2200 m, 13 Oct. 1941, White 4818 (M Jalpa, sobre la carretera Pena 30 km del entronque con la carretera Jalpa-Juchipila, 2550 m, 21 Oct. 1973, J. Rzedowski 939 (MICH, MO, NY, OS). tion, Mpio. Concordia, ¿AS C 18b. Roldana hartwegii durangensis (H. Rob. & Brettel Roldana pennellii var. durangensis H. Rob. & Brettell, Phytologia 27: 423. 1974. Senecio octo- bracteatus var. durangensis (H. Rob. & Brettell) B. L. Turner & T. M. Barkley, Phytologia 67: 392. 1989. TYPE: Mexico. Durango: El Saltó, Sierra Madre Occidental, rocky, canyon, 2500-2530 m, herb, flowers yellow, 31 Aug. 1934, F. W. Pennell 18500 (holotype, US not seen; isotypes, GH!, NY!) var. — ) Funston, comb. nov. Basionym: andesitic, pineland « Distinguishing characters. Stems terete (rarely angulate), concolorous (rarely maculate): leaves abaxially glabrate; phyllaries 5, 4—5.5 X ca. l mm: achenes pubescent or glabrous. Distribution. Southern Durango, Mexico. Comments. Robinson and Brettell (1974) de- scribed variety durangensis from the interior of a southern Durango as having five phyllaries anc elabrous achenes. In the eight collections | examined, five had glabrous achenes and three were pubescent. Regarding distribution, all eight were distributed near El Saltó, many to the west and seemingly on the slopes of the Sierra Madre Occidental. Due to the restricted distribution of five phyllaried plants in southern Durango, the taxon is being maintained as a variety, although within the species Roldana hartwegii. MEXICO. ae a i T rd., 1970, D Repre sentative specimens examined. em an- go: 54 mi. N of Estatión Coyotes of Cuachichilas, 3000 m, 5 N 18788 (CAS); Puerto Buenos na 8 a E of La Ciudad, 53 km W of El Saltó, rd. Durango-Mazatlán, 2650 m, 8 Nov. 1978, García 898 (KSC, MO): 24 mi. E of El Saltó, ca. 3 mi. W of Los Navios, 23 Oct. 1980, Spellenberg 1541 (NMC): 3 mi. W of El Salto, 11 Aug. 1956, Waterfall 12671 (MICH). 5. uud ^ o pl 18c. Roldana hartwegii var. carlomasonii (B. L. Turner & T. M. Basionym: Senecio carlomasonit B. L. M. Barkley, Phytologia 67: 390. carlomasonii (B. L. Turner € T. Barkley) Funston, comb. nov. Turner & T. 1989. Roldana M. Barkley) C. Jeffrey, Kew Bull. 47: 54. 1992. TYPE: Mexico. Sonora: “along the road betw. Yapachic € Yecora, 19 mi. W of border with Chihuahua, dry, rocky hillside, with oaks,” 24 Sep. 1984, S. Sundberg CH). tecas: ca. 38 km W of 2839 (holotype, TEX not seen, TEX digital image TEX!; isotypes, CAS!, MEXU not seen). Distinguishing characters. Stems terete late, or angu- typically maculate, occasionally concolorous; leaves abaxially glabrate; phyllaries 10 to 13, 4—5. ca. | mm; achenes pubescent. Distribution. Primarily along the Pacific slopes of — the Sierra Madre Occidental: Chihuahua, Jalisco, Nayarit, Sinaloa, Sonora. Comments. Turner and Barkley (1989) described Senecio carlomasonii 10 to 13 phyllaries, concolorous or the species based on the following: pubescent achenes, either maculate stem color, and a distribution from Nayarit northward to southern Arizona. Their concept of Roldana hartwegii had a distribution restricted to southern Durango, Nayarit, and adjacent Jalisco. My treatment of the species distribution emphasizes an east-west trend rather than a north-south one, placing the variety carlomasonii on the Pacific slopes of the Sierra Madre Occidental. MEXICO. ón; specimens examined. prd YN hua: Pinas Altos; S wall of Verde Arroyo, 28 16/3 108°18' 00"W. 2100 m, 21 Oct. 1945, ip 50 a Jalisco: Volcán Tequila, rd. to microwave station, ca. 20°47'N, 103°50'W, 1600 m, 25 Oct. 1970, ae 15963 (DS, MICH). Nayarit: along rte. 28, near Villa Caranza, ca. 5 mi. from jet. of hwy. 15 & hwy. 28, along hwy. from Al De ). T. B. Croat 45137 (KSC. MO). Sinaloa: along a small stream 0.5 mi. N of Los Dot mplo. Badiraguato, 1600 m, 1 Nov. 1969, D. E. Breedlove 16765 (CAS, CH): hillside 18 mi. NE of town 6.5 mi of El Cajon, 26 N, 108" W, 1400 m, 26 Nov. 1975, Lehto 19526 (MICH, NY, OS). Sonora: Sierra de / Mandos, 1100 m, 6 Nov. ek H. S. Gentry 4902 ME GH, , NY); along the rd. Yapachic & Yecora, 19 mi. W a the eg with ; R 24 Sep. 1984, $ Sundberg 2839 (CAS 18d. Roldana hartwegii var. subeymosa (H. Rob.) Funston, comb. nov. Dasionyn 2 subcymosa H. Rob., Phytologia 32: Senecio subcymosus (H. Rob.) B. L. Aces ES T. M. Barkley, Phytologia 67: 392. 1989. TYPE: Mexico. Oaxaca: in a moist ravine in pine-oak- alder zone, Sierra Madre del Sur, ca. 125 km 8 of Oaxaca, on rd. to Puerto Angel, ca. 2400 m, 7 1970, A. Cronquist & J. J. Fay 10888 Eus US not seen; isotypes, CAS!, F!, GH!, KANU!, MICH!, NY!). 199]. Phytologia tonii B. 7 ue T. ii dana tonii 249. 2005 Sn San Cristóbal iP las € 15 Nov. 1986, . Senecio Phytologia 71: 304. Turner) B. L. TYPE: asas, Santa Cruz en . Mendez Ton 9481 e Turner, Mexico. Ba mpio. m e lipe, TEX). Distinguishing characters. Leaf blades abaxially lightly tomentose to persistently arachnoid pubescent 310 Annals of the Missouri Botanical Garden on veins and veinlets, adaxially glabrescent. Capitu- lescence compound paniculate cyme, 30 to 50 capitula branch. Capitula radiate, ca. 9 X 3 mm; 8, glabrous to arachnoid, ca. 7 2 mm. florets ca. 3, corolla ligulate, tube pubescent or glabrous, tube ca. 5 mm, ligule ca. 6 x 2mm. Disk florets ca. Achenes pubescent. 7, pubescent or glabrous. Mexico (Chiapas, Oaxaca) at elevations of 2400 m. Found in Distribution and phenology. Southern pine-oak forests. Flowering in November. Comments. Roldana hartwegii var. subcymosa is the southern-most extension of R. hartwegii. Due to its denser pubescence, larger phyllaries, the occasional glabrosity of the corollas, and the geographic separation between the southern boundary of the typical variety and variety subcymosa's distribution, il is being given varietal rank. Based on my examination of the protologue and type material, Senecio tonii is placed in synonymy here. Ps XICO. Chiapas: 1 Cristóbal de las owns e 23034 (DS, t near summit Cerro Z« of $ Saw Ci istóbal de las Casas, 2800 m, 17 Nov. 1981, D. E. Breedlove 5 55611 (CAS); below Zinacantan Center along the ) ft. 5 Dec. 1966, R. M. Laughlin ? 2929 Oaxaca: Sierra Madre del Sur, ca. 125 km S of . 7 Nov. 1970, A. rH, KANU, MICH, NY). Representative specimens na on Cerro Hite epec (Muk’ta i m ue OR xaca, ot naa & Fay 10888 (CAS, F, ( Rob. & Senecio Bot. 2: 19. Roldana hederifolia (Hemsl.) H. Brettell, Phytologia 27: 420. 1974. hederifolius Hemsl., Biol. Cent.-Amer., 241. 1881, “hederaefolius.” Cacalia hederi- folia (Hemsl.) A. Gray ex Urbina, Cat. Pl. Mexic. 188. 1897. TYPE: Mexico. Oaxaca: Pilado € on Canetzeq, NE from Oajaca, Feb. 1845, 379 (holotype, K not isotypes, G not seen, K fragm. € drawing GH! G photos F!, MICH! from Monte Jurgensen seen: Perennial suffruticose (scandent?) herbs, 1-3 m tall; new growth glabrate to sparsely — stipitate- glandular pubescent(?); stems terete, yellowish. Leaves cauline, evenly distributed on hn r half of. stem; blades -l0 X 7.5-10 cm, o» hastate, 3 dus lobes, lowermost petioles ca. 10 cm; — — trip e-nerved, su weakly secondarily lobed or subcordate with 5 weak — lobes(?). Capitulescence a lax compound. corymbose panicle, 3 to 10 capitula per branch, primary bracts lanceolate. Capitula eradiate, narrowly campanulate: phyllaries ca. 8, slightly pubescent: disk florets ca. 20, corolla pappus about equal in length to corolla. (Description funnelform, glabrous; achenes glabrous, compiled from the literature.) Mexico (Oaxaca) at elevations of ca. Distribution. 2000 m. Found in pine-oak forests. Comments. — Roldana hederifolius as described by Hemsley (1881) has leaf blades marginally attached to the petiole and discoid capitula, with the type collection from the state of Oaxaca. It may ultimately fall into synonymy under K. aliena. 20. Roldana heracleifolia (Hemsl.) H. Rob. € Brettell, Phytologia 27: 420. 1974. Senecio heracleifolius Hemsl., Biol. Cent.-Amer., Bot. 2: 241. 1881. TYPE: Guadalupe valley, 7 Oct. Mexico. [east-central Mexico], 1869, Bilimek 558 (lectotype, designated here. K not seen: isotypes, NY!, P not seen Large single-stemmed perennial herbs or sub- 8 8 | 1-3 m tall, caudex, new growth glabrate with tufts of arachnoid shrubs, arising from a lanate woody hairs: stems terete, striate, maculate, brown to red. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous on shrubs or clustered. al midstem on herbs; petioles 5-10 em, glabrate with tufts of arachnoid hairs at axils: blades 15-30 X 15- 20 cm, typically 7-lobed, sinus depth 3/4 of the way to the pinnately veined and lobed to dissected, midrib or incised (larger leaves may be condensed so as to appear ovate and deeply palmately lobed, yet the venation is still pinnate), lobes acute to obtuse, weakly lobed merely callous denticulate, margin weakly pubescent, secondarily to irregularly serrate or leaf base weakly auriculate with a winged petiole; adaxially glabrate, sparsely to densely pubescent on Capitulescence ascending 10 to 30 capitula per veins, abaxially glabrous. compound corymbiform cyme, Ram branch, peduncles glabrous with arachnoid tufts at 5 mm, ultimate — s, primary bracts linear, up to 3-10 mm, 3 mm. Capitula radiate, 8-12 X axi peduncles bracteoles 2 to linear, l— 2—3 mm, narrowly campanulate, corollas bright yellow: calyculate bracts ca. 3, linear, 1-2 mm; phyllaries ca. 8, glabrous to slightly pubescent, sometimes purple-tinted, 6-8 popa 1.2 mm. Ray florets 5 to 7, corolla ligulate, glabrous, 2-3 mm. Disk glabrous, ca. Wn ca. long. ligule 5-7 florets 12 to 16, 3 mm, 5 mm corolla funnelform, 8mm, tube ca. throat ca. 3 mm, lobes ca. 2 mm. Achenes pubescent with short hairs along the ribs, cylindrical, 2-3 mm, ribs 10, resin glands absent; pappus bristles 5-7 mm. Distribution and phenology. Central Mexico (Aguascalientes, Guanajuato, Jalisco, México. Mi- choacán, Querétaro, Zacatecas) at elevations of 1600— 2700 m. In oak forests and open disturbed woodlands, roadsides, pastures. Flowering September through November. Volume 95, Number 2 Funston 311 2008 Taxonomic Revision of Roldana Comments. Hemsley (1881) lists two specimens terete, dark brown. Leaves cauline, about evenly from K in the protologue: S Mexico, Guadalupe. valley Bilimek 558; The Bilimek 558 (K) specimen is As McVaugh (1984: 826) notes, the exact location from which the type was of Mexico, and, without locality. Mackenzie s.n. chosen as the lectotype. collected is an enigma, as there is no modern record of its occurrence in or about Guadalupe. Roldana des R. however, it grows slightly taller, typically has more resem — heracleifolia lineolata; disk florets per capitula, and has pubescent achenes. While both have pinnately lobed leaf blades, R. heracleifolia's sinus depth is more than 1/2 of the way to the midrib and its blades are abaxially glabrate with veins pubescent, whereas R. lineolata's sinus depth is typically only 1/3 or less of the way to the midrib and its blades are abaxially tomentose. AUREIS specimens examined, MEXICO. Ag lientes: Mes pn Alto, 2400 m, 23 Sep. 198( Gutierrez HL 'renadas Palo / ). R. ie 8 (MICH, NY). : Tlalpujahua, 200: m, | Oct Pd sm Michoacán: at a small s a. E of More = A m, 9 Oct. 1965, A. o dore (C. AS, G H, , MICH, NY, WIS). Qu Bon VR de Allende, ca. Km 8, 2150 m, (CAS). Zacatecas: "a 18 de la anueva-El Plateado, mpio. Villanueva, 2000 m, 1978, García 866 (CAS, DAV, F, MICH, MO). od rd. along SLP & 1 PER mall 5 Nov. — 21. Roldana heterogama (Benth.) H. Rob. € Brettell, 420. 1974. Cacalia heterogama Benth. in Orst.. Vidensk. Meddel. Dansk Naturhist. Foren. Kjobenhavn 1852(5—7): 107. 1853. Senecio heterogamus (Benth.) Hemsl., 242. 18 TE Phytologia 27: Biol. Cent.-Amer., Bot. Costa Rica. Cartago Province: Volcán de Irazu, [1845]. A. S. Orsted 192 (lectotype, designated here, microfiche IDC 8858 not seen, 190 III 4—5!, K fragm. & C photos F!, GH! MO!, K not seen; isotypes, C 2204. drawing GH not seen, NY!) Field ae nov. Senecio heterogamus var. kellermanii Greenm., Publ. Columbian Mus., Bot. Ser. 2: 282. 1907. TY Sacatepéquez: Volcán de Aqua. 9100 ft, 15 Feb. 1905, W. A. Kellerman 4706 iolotype, F!) Cacalia aclamdha S. F. Blake, Contrib. Gray Herb. 52: 58. 1917. Syn. nov. TYPE: Guatemala. Sacate ba zi ol án de Aqua above Santa María de Jesús ; . D. Holway 570 (holotype, GH not seen, GH plicio MICH!, GH drawing NY”. Suffruticose herbs, shrublets or shrubs, 1-5 m tall; new growth glabrous to densely pubescent with stipitate-glandular and long multicelled hairs; stems lago: ca. 19-20 km NW of C distributed on upper half of stem, lowermost decid- uous; petioles 8-16 cm, pubescent with short and long blades 8-20(30-35) x 8-20(30- em, principal leaves eccentrically peltate, triple- multicelled hairs: 35 nerved, ovate-orbicular, about 1/4. of secondarily lobed, — T- to 11-lobed, sinus depth the way to the midrib, lobes acute, margins callous denticulate, pubes- with hairs, densely so on veins with long multicelled brown cent; abaxially pubescent stipitate-glandular hairs, adaxially sparsely pubescent with short glandular hairs. Capitulescence a loose compound paniculate cyme, O to 20 pubescent with a mixture of stipitate-glandular and capitula per branch, peduncles long multicelled hairs, primary bracts obovate to der io 10-30 X 3-10 mm, ultimate peduncles 10-30 mm. bracteoles I to 3, linear, 3-6 mm. Capitula n (5-)9-15 X 3-5 mm, narrowly eal uas corollas yellow: calyculate bracts ca. 3, linear 2-6 mm: phyllaries 8 to 13, stipitate-glandular petis 8-13 X ca. 1 mm. Disk florets (17 to)20 to 40, corolla campanulate, glabrous, ca. greenish purple-tinted, 8 mm. tube ca. 3 mm, throat ca. 2 mm, lobes ca. 2 mm. Achenes glabrous, cylindrical, ca. 1 mm, ribs 10, resin glands absent: pappus bristles ca. 7 mm. Distribution and phenology. Southern Mexico to northern Panama (Costa Rica: Puntarenas, San Huehuetenango, Quezaltenango, San Marcos. Sololá, Chiapas, Oaxaca: Panama: Bocas del Toro, Chiriquí) Cartago, José; Guatemala: Chimaltenango, Guatemala, Quiché, Sacatepé- quez, Totonicapán; Mexico: at elevations of 3000-4100 m. In pine-oak forests. mostly on upper volcanic slopes. Flowering December through April. Bentham (Örsted. 8000 ft.” but did not listany specimens. Hemsley (1881: 242) listed A. S. Orsted 192 (K) when he made Senecio. It is Comments. ln the protologue. 1853) listed type material from “Irasu. his combination in assumed that he was examining Bentham’s type material. Furthermore, Gibson (1969: 100) states that he saw a drawing and fragment of A. S. Orsted 192 (GH) and that it was made from K of the lectotype. Roldana heterogama is a common Central Ameri- can species and has eccentrically peltate leaves. The variety kellermanii was based on plants having larger leaf blades, longer petioles, and the involucre villous- reddish brown hairs. These hirsute with red or characters are not inconsistent with the species as a whole; therefore, the variety is placed in synonymy. Upon examining the protologue and type material for Cacalia calotricha, it is also placed in synonymy. COSTA RICA. Car- Representative specimens examined. E Asuncion (Cerro de ¿erro 312 Annals of the Missouri Botanical Garden ) on C.R. #2, 19.6 a beyond El E mpalme, 2500 m. :. 1973, (iss 1091 (F). P Perez Zeledon, Cordillera 83 O 2600-2800 m, R. + Muerte 23 De . 197 "0. 58356 (F). Huehuetenango: rd. to . S of San Juan Ixcoy. 3000 m, 4 Feb. 1965, D. £. Breedlove 2475 (F, MICH): Cerro E p San Ildefonso . 15 Aug. 1942, J. A. Steyermark 50556 (F, MO). Quezaltenango: Fuentes Georgin: des :án de Zunil, 2850 m, 4 Mar. 1939, P. C. Standles 67453 MO): SE of Palestina. on old rd. to San Juan Ostuncalco. Jan. 1941. P. C. Standley 64269 (V). Quiché: s, 2200 m, Apr. 1892, Smith e (F, GH, z: slopes of Volcán de Agua, above : Jesús. 2500 m, 11 Feb. 1939, P. C. Sta ndley San Marcos: betw. San va & summit of Volcán Tajumulco, 4200 m, 13 Feb. 1940, J. / 35567 (F, MO): rd. betw. San Se M. m Km 21 & Km 8, 8- 18 mi. NW of San Marcos, 3200 m, 15 Feb. 1940, J. Steyermark 55616 (F). Sololá: Volcán Tolima (side facing J. A . Steyermark Volcán Atitlán to F 2900 m. n 1942, Steyermark 47582 (F, M( onicapin 1 sleep slopes of Marta Tecun, 3300 m. vd 23 Jan. Ml R. 16400 (NY). MEXICO. Chiapas: on the E ys ol f the summit a Volcán Tacaná, mpio. of Union Juarez, 3600 m, 3 Mar. 197 D. E. Breedlove 24295 (MO): SW of Me: Xx1can ER 190 near Rancho Nuevo, ca. 9 mi. SE of S mpio. San Cristóbal de las Cas Breedlove 9227 (DS, MICH, NY). y betw. Oaxaca € Pochutla, 7.3 mi. N of 2750 2] Jan. e . T. B. Croat 46166 (MO). : AN, AMA. Bocas del Toro: le de headwaters of the Río C dubie 6 ae km NW of the peak of Cerro Ec handion on the Costa 5! Talamanca, Rican—Panamanian inte Eom al border 82 50'W, 2500 m. 2 Mar. 198: idse 25154 (M o» M just below summit, El us 3000 m. 20 Mar. 1977, D'Arey 11028 (MO). 22. Roldana heteroidea (Klatt) H. Rob. «€ Brettell, Phytologia 27: 420. 1974. Senecio heteroideus Klatt, Leopoldina 24: 125. 1888. Digitacalia ds (Klatt) Pippen, Contr. U.S. Natl. Herb. 1968. TYPE: Mexico. Oaxaca: Andres et S. Maiguel, F. M. Liebmann 178 (holotype, C not seen; isotype, GH C drawing GH! C tracing MO!) inler S nol seen, Amer. J. Sei. xico. Oaxaca: on 10 Oct. GH not Cacalia iig cr B. L. Rob. € Greenm., Art 23,50: . 1895. TYPE: Mex duy de o 's of ps Sie rra de San Felipe, 7000 ft., 1894, C. G. ingle 5828 US not seen). (holotype, seen: isotype, Large perennial herbs, 1-2 m tall; plants essen- tially glabrous; stems terete. striate. cream to red. Leaves cauline, clustered at the midstem: petioles 8- 12 em, glabrous; blades 7-10 X 9-13 cm. palmately nerved, palmatifid, 5-lobed, sinus depth about 3/4 of the way to the midrib, lobes acute, secondarily lobed , WwW slope of to merely denticulate; abaxially glabrate with short glandular hairs along veins and margins, adaxially glabrous. Capitulescence simple paniculate cyme, 2 to 5 capitula per branch, peduncles glabrous, primary bracts oblanceolate, 359-130 mm, Capitula eradiate (? or 22.5 mm, ultimate peduncles linear, ca. 4 mm. radiate), 12-18 X 5-8 mm, campanulate, corollas pale yellow: bracteoles O to 3, calyculate bracts a. 3, linear, ^ 5-7 mm: phyllaries ca. 8, glabrous, Disk glabrous, ca. | mm. lobes 2- ca. 2 mm, 10 ribs, resin glands absent; pappus bristles 7-8 mm. greenish purple-tinted, 12-15 X ca. 4 mm. lorets 30 to 40, 10 mm. tube ca corolla funnelform, 7mm, throat ca. 3 mm. Achenes glabrous, cylindrical, Distribution and phenology. 2400-2700 m. Collected in flower October through Decem- Mexico (Oaxaca) only al elevations of a e Found in pine-oak forests. Jer. Comments. Roldana heteroidea is a unique spe- cies with a very restricted distribution. It has only been collected along the ledges of the Sierra de San Felipe in Oaxaca. However, a collection by 7. F. Daniel 366 ASU; 120, Cadereyta) could be a variant of Roldana heteroidea. MICH, betw. Jalpan and — Queretaro, along hwy. Its distinctive features are the ca. 15-mm-long ray ligules and bright green deeply 5-lobed to dissected palmate leaves. Its taxonomic determination depends on further investigation. Representative pcia examined. No XICO. Oaxaca: Cerro de San Felipe, 7500 ft., 23 Sep. 1895, C. C i San > e es G "m 1894, C. L. onzatti 715 3500 fi., 2 . NY): Cerro de pa 287 (F, MO, B NIU 6170 (V. San Felipe. p ft. Oct. NY) 23. Roldana hintonii H. Rob. & Brettell, Phyto- logia 27: 420. 1974. Senecio hintonii (H. Rob. & Brettell) Pruski & T. M. Barkley, 238. 1989. TYPE: Mexico. México: pec. Comunidad, upper oak belt, Mar. 1936, G. B. Hinton 8967 (holotype. US not photos F!, GH!, MOL isotypes, Fl, GH! MO!, NY!, UC, Phytologia 67: Temascálte- seen, US MICH!, Suffruticose herbs to subshrubs. 1—2.5 m_ tall arising from a woody caudex; new growth appressed arachnoid pubescent; stems terete. light brown. Leaves cauline, evenly distributed on upper half of stem. deciduous; pubescent: 8-16 X apex acute, base cuneate, margins entire to subentire, lowermost petioles 14 cm, blades 2.5-5 cm, pinnately veined, elliptic. — minutely callous denticulate; surfaces glabrate. ap- pressed arachnoid hairs on veins, may be densely so abaxially. Capitulescence pyramidal compound panic- Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana ulate cyme, 10 to 25 capitula per branch, peduncles pubescent with appressed arachnoid hairs, primary bracts linear, ca. 2 mm, ultimate peduncles 9-11 mm, bracteoles O to 4, Capitula radiate, 8-12 X 2-3 mm, funnelform, corollas yellow: calyculate bracts ca. 3, scale-like, to 1 mm: phylla- ries, ca. 13, glabrous, greenish purple-tipped. 4-6 X 1-2 mm. Ray florets 6 to 8, corolla ligulate. glabrous, tube 3-5 mm, ligule 4-6 X ca. 2 mm. Disk florets ca. scale-like, to ] mm. 12, corolla funnelform, glabrous, 7-10 mm, tube 2- 4 mm, throat ca. 4 mm, lobes 1—2 mm. Achenes glabrous, cylindrical, 2-3 mm, ribs 10, resin glands absent; pappus bristles 6-8 mm. Distribution and phenology. Central Mexico. Known only from the vicinity of Temascaltepec, Mexico, at elevations of 2100—3000 m. Found in pine-oak and fir forests. Flowering February through March. Comments. | Roldana hintonii is a rarely collected species due to its restricted distribution. At glance, it may appear to be a misplaced R. schaffneri. However, R. hintonii has 13 phyllaries versus 5 or 6 in R. schaffneri, and it has more numerous and larger ray first and disk florets. Representative specimens examined. MEXIC . Mé xico: as Cruces, mplo. bs mascáltepec, 2900 B Hinton 3264 (F, GH); omini Hinton a a Hes (G H, MIC ; , UC); » s 1936, . Hinton 8966 i GH, 3» Temascáltepee on e to Toluca, 19° 06' 58"N, m SEED 22 km 99^56' a n . 24 Mar. 1996, T. M. Barkley et al. 4030 (KSC, MO). 24. Roldana hirsuticaulis (Greenm.) Funston, Novon ll: . 2001. Senecio hirsuticaulis Greenm., Publ. Field Columbian Mus., Bot. Ser. : 1907. TYPE: Mexico. San Luis Potosí: en route from San Luis Potosí to Tampico, Dec. 1878 to Feb. 1879, Dr. E. Palmer 1114 (holotype, GH!; isotype, NY!). Suffruticose herbs to shrubs, 2 m tall; new growth tawny, lanate tomentose; stems terete, reddish brown. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 5-6 em, lanate tomentose; blades 11-14 9.5-13.5 nerved, ovate to cordate, 9(+)-lobed, sinus depth less than 1/8 of the way to the midrib or merely dentate, lanate tomentose and canescent, cm, triple- abaxially densely adaxially pubescent along veins with long thin hairs. Capitulescence — flat-topped compound paniculate cyme, 20 to 40 capitula per branch, peduncle lanate ca. 5 mm, ultimate bracteoles ca. 2, l- X 2 mm, turbinate, 3; 1- tomentose, primary bracts linear, peduncles 2-10 mm, linear, 2mm. Capitula radiate, ca. corollas yellow; calyculate bracts ca. inear, 2 mm; phyllaries ca. 13, glabrous, greenish purple- tinted, 3-5 X ca. Ray florets ca. 3, inconspicuously ligulate, pubescent, tube ca. 4 mm, | x pubescent, ] mm. corolla Disk florets ca. 7, tube ca. 3 mm, ligule ca. | mm. corolla funnelform, ca. 7 mm, 2 mm. Achenes glabrous, 10, throat. ca. en mm, lobes ca. eylindrical, 2-3 mm, ribs resin glands absent; pappus 7 ca. 5 mm. Distribution and phenology. Eastern Mexico (Nuevo León, San Luis Potosí, Tamaulipas) at elevations of 2000-2500 m. Found in pine-oak forests. Flowering December through February. Roldana and Comments. The differences between hirsuticaulis and R. aschenborniana are minor, much consideration was given to placing R. hirsuti- caulis in synonymy. The species is recognized due to the persistent tomentum and often tawny pubescence found on the stem, peduncles, and underside of the leaf. Roldana aschenborniana typically is less densely pubescent and never tawny. Roldana hirsuticaulis seems to be restricted in distribution to the northern extreme of R. aschenborniana’s range. Gibson (1969) believed the type material for R. hirsuticaulis to be a hybrid between R. aschenborniana and R. lanicaulis. The relationships between these species are in need of more detailed analysis. Representative specimens examined. MEXICO. Nuevo León: trail from Potrero Redondo to Laguna a mpio. Villa ma l6 Aug. 939, Muller 2748 (GH, N UC). San Lui chiefly in the region of San a a 22°N ree "2000- 2500 ı m, 1878, Parry & Palmer 539 G ae en ped from San Luis Potosí SS Tampio o, Dec. 1878 to Feb. Palmer 114 GH, . Tamaulipas: 5 mi. W, zu vd Harrell M 59 9 (MO); W of Gomez Farias on rd. to Rancho o del Cielo, McCarten 2605 (MICH); N of Frank Harrison’s, Rancho del Cielo, in Sierra de Guatemala above Gomez Farias, Sharp 52267 (NY). 25. Roldana jurgensenii (Hemsl.) H. Rob. & Brettell, Phytologia 27: 421. 1974. Senecio jurgensenii Hemsl., Biol. Cent.-Amer., Bot. 2: 242. 1881. TYPE: Mexico. Oaxaca: San Pedro Nolasco, NE by E from Oaxaca, Feb. 1845, Jurgensen 309 (holotype, K not seen; isotypes, G not seen, K fragm. & drawing GH!, G photos F!, GH!, MICH!, MO). Roldana 2s dlovei i Rob. & Brettell, Phytologia 27: 416. 1974. TYPE: Mexico. Chiapas: SW of Mexican Hwy. pm near Ranc i Nuevo, ca. 9 mi. SE of San eus de las Casas, steep slope near crest of ridge 9 ts. Mar. 1965, D. E. Breedlove 9228 (holotype. Us not seen, US photos F!, GH!, MOL; isotype, NY!). shrubs, 1—4.5 m fruticose herbs, shrublets. or Bey Su tall; new growth densely covered with a combination of short stipitate-glandular and long multicelled hairs. 314 Annals of the Missouri Botanical Garden glabrate with age; stems terete, maculate, color — ler states were also found to overlap to such an extent varying from green leaves with purple stems to that the designation of separate taxa was nol tenable. reddish brown leaves and stems, numerous lenticels either present or completely absent. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 5-14.5 cm, glabrous; blades 10— 8-14.5 em, palmatifid, 5- to 7- triple-nerved, oval-ovate to sub- lobed, sinus depth less than 1/4. of the way to the midrib, lobes acute, weakly secondarily lobed to merely denticulate; surfaces glabrous to sparsely pubescent with short glandular hairs. Ca- pitulescence pyramidal compound paniculate cyme, 50 to 70 capitula per branch, peduncles densely pubescent with stipitate-glandular and long multi- celled hairs, primary bracts linear, 2—4 mm, ultimate peduncles ca. 6 mm long, bracteoles 0 to 3, linear to 1-2 mm. 3 mm, turbinate, corollas yellow; scale-like, Capitula radiate, 6-11 X 2- calyculate bracts 0 o 3, linear, ca. 2 mm; phyllaries ca. 8, stipitate- ES 25 pubescent to glabrate, purple-tinted, ca. 5 Ray florets ca. 5, glabrous, tube (0.5)6-7 mm, ligule (4.5)6—7 X (1)2— Disk florets ca. 7 to 13, corolla funnelform, 6-8 mm, tube (1-2)4-5 mm, 3 mm, lobes ca. (1)2 mm. Achenes glabrous, cylin- 1.2 mm. corolla ligulate, 3 mm. glabrous, throat ca. — drical, 1-3 mm, ribs 10, resin glands absent; pappus bristles 5-6 mm. Distribution and phenology. Southern Mexico through Central America into Honduras (El Salvador: Santa Ana; nango, El Progreso, Guatemala: Chimaltenango, Huehuete- Quiché, Marcos, Sololá, Suchitepéquez, Totonicapán; Hon- Lempira, Santa Barbara; Mexico: at elevations of 1000-3000 m. montane cloud forests, open meadows, mixed broad- Quezaltenango, San duras: Chiapas, Oaxaca) Found i leaf and cypress forests, and moist hillside thickets. sold in Totonicapán markets for The leaves wrapping February. are tamales. Flowering December through Comments. Roldana jurgensenii is one of the more common species in Central America. Robinson and Brettell (1974) establis on the basis that it has more truncate leaf bases, less pagas ed the species R. breedlovei closely palmate leaf veins, throats of the corollas much shorter, and stems smooth (vs. raised lenticels). From my examination of the available collections, | believe that the species described by Robinson and Brettell is at one end of a morphological continuum. In distribution, the complex has small the southern capitula, dark brown leaves, and numerous raised ight-colored lenticels on the stems and leaf blades. In the north, the capitula are larger with longer ray ligules and leaf blades are light green abaxially, dark These charac- green adaxially, and without lenticels. leaf morphology, leaf shape and With venation in the genus are notoriously variable, and regard to it holds true in this species as well Representative specimens examined. EL SALVADOR. summit of Los Santa Ana: edge of cloud forest near E sesmiles, 14°21'N, 89 pa W, 2580 m, 10 Mar. 1942, J. M. ucker 987a. (F, UC). GUATEMALA. Chimaltenango: C ¿erro Chichoy near Chichoy, where dept. of ( O Quiche & Sololá ca. 14 48'N, 91 W, 2800—3200 m, 26-27 Jar an. 1949, L. O. Williams 15324 i yam, region of Sant: 38, P. C. Stanley 60995 (F, MO). Huehuetenango: rd. to Huehuetenango, 7 mi. S of San Juan Ixcoy, mpio. San Juan Ixcoy, Sierra de los Cue 'humatanes, 9750 m, 4 Feb. 1965, D. E. Breedlove e DS, F ; Sierra Cuchumatanes TA Paqu Juan 3000- 3350 m, 8 Jan. . A. R a (F, MO, NY). El Progreso: uppermost portion of slopes, betw. Calera & 21 Jan. 1942, J. A Quezaltenango: pe summit of Volcán Siglo, 2000-3 ovr 43100 (F, GH, La Esperanza, ca. ^ km from San Juan Ostuncalco, 12-23 Jan. 1966, A. Molina R. 16596 (F, Quiché: Chiul, 8000 ft., ns 1892, Heyde et Lux. 2276 (F, GH). San Marcos: wet mtn. forest near Aldea Fraternidad. betw. San Rafael Pie de la Cuesta & Palo Gordo, W-facing 1800-2400 m, 10-18 Dec. 300 m, NY Canton 3200 m. slope of the Sierra Madre Mins., 1963, L. O. Williams et al. 25818 (F). Sololá: Volcán Santa Clara, S-facing slopes to Roo 2100—3000 m, 5 June 1942, J. A D UE 46973 (F). Sue pe Volcán Atitlán, S de 7800 ft., Ti 35. A. F. Skutch 2130 (A, O, NY). Totonicapán: ka n of ( mtns. pe Totonicapán, on rd. biu Jolom, Jan. 1941, P. C. Standley 34415 F MO). HONDU RAS. Lempira: Siquatepeque, Campo Narangjo, 14°33'N, 88^ bubus 2455 Jan. 1992, P. House 1183 (MO). Santa Bárbara: upper rocky slopes € summit of Cerro MEXICO. Chiapas on the NE slope of Zontehuitz near ne nmit, mpio. C UE rs IE: x Dec. 1964, D. E. Breedlove 7801 (D, MICH, MO); rd. from Motozintla 2 Mendoza to Siltepec e El ida 14.1 mi. NW of Motozintla, n ue on da slopes facing the Atlantic, principal forest, 11 Feb. 9. T. B. Croat 47298 (KSC, MO). Oaxaca: 65. pus mi. N of > Es of hwys. 175 & 190, 17 Dec. 1977, V. Funk 2731 (OS). 26. Roldana kerberi (Greenm.) H. Rob. & Brettell, Phytologia 27: 421. 1974. Senecio kerberi Greenm., Ann. Missouri. Bot. Gard. |: 279. 14. TYPE: Mexico. Colima: Mesa del Arrero, “trompetero,” 21 Nov. 1880, E. Kerber 94 (lectotype, designated here, B fragm. & tracing at GH!, B photos F!, GH!, MICH!, MO!, NY, B tracing MO!). Senecio Ts McVaugh, Contr. Univ. Michigan Herb. t 473. fee aa a usua SE ro H. Rob. ^ e 27: 418. TYPE: Mexico. Cac oma, ca. 4 md alisco: Sierra. de precipitous mountainsides, San Miguel de la Sierra, 2000 m, 3 Nov. 1962, R Me nel 2201 10lotype, MICH!; isotype, NY!). ( Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana Suffruticose herbs, 1.54.7 m tall; new growth lightly to densely arachnoid floccose or tomentose to rarely tawny; stems terete, striate, maculate, reddish brown. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles (2-)10-20 cm, glabrate with arachnoid hairs at axils: blades ca. (5-) 10-35 x (2-)10-25 cm, palmate- to triple-nerved, leaf shape various with 7 to 13 lobes: oblong-ovate to oval, fan-shaped with distal lobe sinus depth ca. 1/2 of the way to the midrib, apex acute to rounded, proximal lobes sinus depth ca. 1/4 of the way to the midrib, lobes [om acute; or broadly ovate-elliptic, sinus depth ca. 1/4 of the way to the midrib throughout, lobes broadly to narrowly acute; margins weakly denticulate, weakly secondarily lobed to variously serrate; abaxially glabrate, sparsely to densely arachnoid pubescent to having a tomentum, adaxially glabrous to slightly pubescent on veins and margins. Capitulescence pyra- midal or flat-topped compound paniculate cyme, 25 to 100 capitula per branch, peduncles glabrate to arachnoid floccose, primary bracts linear, 2-10 mm, ultimate peduncles 2-10 mm, bracteoles O to 3, linear to weakly filiform, ca. 1 mm. Capitula radiate, 5-9 X 2—4 mm, turbinate, corollas yellow to golden; calycu- late bracts ca. 3, linear, ca. 2 mm: phyllaries 8 to 13, 4-6 X 1-2 mm. Ray florets 3 to 6, corolla ligulate, glabrous or pubescent, tube 2-3 mm, ligule 3— glabrous, 6 X ca. 2 mm. Disk florets 5 to 11, corolla funnelform, glabrous or pubescent, 5-7 mm, tube 2-3 mm, throat ca. 2 mm, lobes 1-2 mm. Achenes glabrous, cylindri- ribs 10, cal, 1-2 mm, resin. glands absent; pappus bristles 3-6 mm. Distribution and phenology. Mexico (Colima, Jalisco. Oaxaca) at elevations of 1500-2600 m. Found in pine-oak and fir forests. Flowering October through March Comments. The lectotype designation for Senecio kerberi was made according to the International Code of Botanical Nomenclature (Greuter et al., 2000: Art. 8.3, . 5). In the protologue for 5. kerberi, man (1914) listed the type material as being a specimen in B and a tracing and fragments in GH. The Berlin material was destroyed in 1943. Greenman (1914: 279) states that he was permitted to make tracings and take fragments for the Gray Herbarium from the specimens he was studying at the Royal Botanical Museum of — Berlin. It is apparent that the GH tracing done by Greenman is of the B specimen that is depicted in all of the photographs, as well as the MO tracing. Therefore, the GH fragment and tracing is determined to be of the x type and, thus, eligible to be the lectotype. Roldana kerberi superficially resembles R. hart- plant with typically 280) notes in the wegii; however, it is a stouter larger leaves. Greenman (1914: protologue that it is Benth. from other species of sect. Palmatinervii to which it “related to Senecio hartwegi and S. reglensis Greenm., but from these and belongs it is readily distinguished by the somewhat dn more or less fan-shaped and bluntly lobed eave It has been surmised (Gibson, 1969; oe & ee 1974; MeVaugh, 1984; Kowal, 1991; Turner, 2005) that the type is a specimen of the more recently circumscribed S. galicianus McVaugh ca — 972). He described S. galicianus, citing specimens that had been annotated by Gibson as “S. roldana var. calvescens,” a variety never published, commenting that S. of the associated with S. roldana (R. lobata). In Flora Novo- Galiciana (1984: 829), MeVaugh also states that he believes S. galicianus is of closer affinity to S. roldana galicianus has few characteristics (R. lobata) than to S. hartwegii. While this may be true, I do not completely dismiss an affinity to R. hartwegii. For 77 years, the only recognized gathering of Roldana kerberi was the type material. In 1991, two gatherings were collected from the general area of the type: R. R. Kowal & M. A. Wetter 3025, as reported by Kowal (1991: 131) and A. C. Sanders 10314 (MO!). Kowal also discusses the relationships between R. kerberi, Senecio galicianus, and R. lobata. He found that while R. kerberi is ecologically similar in habit with R. lobata, it is morphologically more similar to 5. galicianus, differing from the latter in leaf morphology and pubescence type, the leaf blade being much more leathery with a tomentum formed by rough, septate hairs and having a rust-colored pubescence. Kowal provisionally accepted R. kerberi as distinct from 5. galicianus. While I had originally intended to do the same in this manuscript, upon finding the Sanders gathering, S. galicianus is relegated to synonymy. McVaugh (1972) understood this possibility as he stated in the protologue: “It is possible S. galicianus is a synonym of S. kerberi Greenm.” The three species Roldana hartwegii, R. kerberi, and R. lobata variously overlap along the Sierra de Manantlán, Sierra Madre del Sur, and Sierra de Miahuatlán. The varieties recognized here are illus- trations of the possible ecomorphologic variation and hybridization that may occur among any of the species of Roldana The interested reader is directed to the A. hartwegii that have a broad distribution. section of this paper, as well as Greenman (1914), Kowal (1991), McVaugh (1972), and Turner (1996, 2005) for further discussion on the interrelationships of these species and their habitats. KEY TO THE VARIETIES OF ROLDANA KERBERI distal lobes m Leaf blade oblong-ovate to oval, elongate, apex acute to rounded, sinus depth 316 Annals of the Missouri Botanical Garden 1/2 of the way to the e proximal lobes acute precipitous mins. ca. 40 km W of Ayutla, 3-6 mi. SE of El with sinus aepth ca. 1/4 of the we ay to the midrib; Carmen, 2100 m, 29 "nb 1960, R. MeVaugh 21560 (MICH). phyllaries 10 to 13; bin glabrous ...... A te eee see ee Oe kerberi var. kerberi 26b. Roldana kerberi var. calzadana (B. L. Ib. Leaf blade broadly ovate-elliptic, lobes variously Purner) Funston, comb. nov. Basionym: Roldana acute, sinus depth ca. 1/4 of the way to the midrib: phy pos 830 ie pubescent. lobes broadly acute; capitulescence ee by upper foliage; plants arach- noid to densely tomentose, from Sierra Madre del Sur .. 26b. R. kerberi var. calzadana 2b. Leaf lobes nasi acute; capitulescence elabrate-arachnoid to from Sien E le sra JR. ether var. manantlanensis 26a. Roldana kerberi var. kerberi Distinguishing characters. Suffruticose herbs, 1.5—4 m tall: new growth weakly arachnoid floccose. Leaf petioles (2-)10-15 em, glabrate with arachnoid hairs at axils; blades ca. (5-)10-15 X (2-)10-17 em, oblong-ovate to oval, distal lobes sinus depth 1/2 of the way to the midrib, apex acute to rounded, proximal lobes acute: abaxially glabrate to sparsely arachnoid pubescent, adaxially glabrate. Capitulescence pyrami- dal compound paniculate cyme, ultimate peduncles glabrate to arachnoid floccose. Capitula radiate, 6-9 X 24 mm: phyllaries 10 to 13. Ray florets 3 to 6. Disk florets 6 to 11, mm. glabrous. glabrous, 5-7 slopes at the Belt in the Sierra de Manantlán at elevations of 1200-2200 m. Distribution. Found on Pacific western end of the Trans-Mexican Volcanic Comments. The type specimen of Roldana kerberi consists of an inflorescence and what appears to be a cauline leaf. foliaceous bract as opposed lo a Nevertheless, cauline leaves have traditionally been 5-10 X 2-8 cm tomentose abaxially. It is suggested described as and subarachnoid- here that the relatively smaller leaf size is an artifact of the type specimen and should not be taken as a strict descriptive characteristic. Regarding plant pubes- cence, as is often found in the genus, plants growing along the Pacific slopes of Mexico have a tendency toward persistent and denser pubescence in the southern ranges. while the northern ranges new growth may be densely pubescent but has a greater tendency to become glabrate with age. Representative specimens examined. MEXICO. C gee Rancho El Jabali, 20 km (air) NNW of Colima in the l © Volcán de ( s 19 26.3'N. 103 42.5' : 245 Jan. 1991, A. C. Sanders 10314 (MO). durs 0: NW flank of the Nevada En Colima, ca. 23 rd. W of C oa bene 2000 m, 31 Dec. 1962. 4. us 9767 (GH, . MICH. MO. NY); 12-15 mi. SSE of ae e to Corralitos, 4-10 mi. above (SE of) Ahuacapan, 1500 m, 22 Nov. 1989, R. McVaugh & Koetz 978 (MICH) Autlán. on Turner, Phytologia 80: 276. 1996. Mexico. Oaxaca: mpio. San Martin Peras, calzadana B. L. TYPE: carretera Coicoyan de las Flores, Santiago 17N, 98 1]1W, “200 m de la desviacion a San Martin Peras,” ca. 2535 m, 16 Feb. 1995, J. 1 Calzada 19738 (holotype, TEX nol seen, TEX digital image TEX!; isotype, MEXU nol seen). Juxtlahuaca, 17 Distinguishing characters. Suffruticose herbs, ca. 2 m tall: new growth tawny-puberulent. Leaf ae S 16-20 cm broadly ovate-elliptie, long, blades ca. 30-35 16-20 cm, sinus depth about 1/4 S the way to the midrib, lobes broadly acute: abaxially glabrate to arachnoid-tomentose, adaxially glabrate with hairs along major veins. Capitulescence. flat- topped congested cymes, arachnoid-tomentose. Capit- ma radiate, 5-6 mm high: phyllaries 8. Ray florets ca. ligule 4-6 X ci Disk 7-8 mm. (Description 3, tubes puberulent, u 2 mm. a florets ca. 5, pubescent. compiled from the literature.) Southern Mexico (Oaxaca) along the 2500-2000 m. Distribution. Sierra Madre del Sur at elevations of Comments. ‘Turner (1996) separates Roldana cal- its having larger deltoid) each irregularly serrate, and having a zadana from R. manantlanensis by with marginal lobes, leaves more broadly acute (broadly tomentose versus arachnoid pubescence. Turner also congested suggests a difference in capitulescence: flat-topped cymes as opposed to pyramidal cymes. While the somewhat disjunct distribution between gatherings and the possible difference in capitules- cence form is not enough for me to recognize the species, it is being given varietal rank. The difference 230) discusses that the area is also ecologically hospitable to R. lobata: not be at all surprising to find these two species in distribution is not great, and Turner (2005 would therefore, as discussed above. i growing together. 26c. Roldana kerberi var. manantlanensis (R. Funston, comb. Basionym: R. Kowal nov. — N galicianus var. manantlanensis R. Brittonia 43: 109. 1991. antlanensis (R. R. Kowal) B. L. logia 80: 278. 1996. TYPE: Mexico. 23 km SSE of Autlán, 17 km SSW of El Chante. on or below NW-SE crest Occidental, on lumber rd. f 4 km NW Senecio Kowal, Roldana man- Turner, Phyto- Jalisco: of Sierra de Manantlán from Cerro La Cumbre km NE « — toward las Joayas, ca. Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana Cerro El Almeal, at spring above rd. and along brook below. 19 33-34'N, 104°14.5-15.5'W, 2090 m, 2 Jan. 1980, R. R. Kowal 2776 (holotype, WIS!: isotype, MICH!). Distinguishing characters. Suffruticose herbs, 2- 4.7 m tall: new growth weakly arachnoid floccose. Leaf petioles 16-20 cm, blades 30-35 X 16-25 cm. broadly ovate-elliptic, sinus depth about 1/4 of the way lo the midrib, lobes acute to narrowly so; abaxially glabrate to arachnoid-tomentose, adaxially glabrate with hairs along major veins. Capitulescence flat-topped congested cymes, arachnoid tomentose. Capitula radiate, 5-6 mm high; phyllaries 8. Ray florets ^ a. 3, tubes puberulent, ligule 4-6 X ca. 2 mm. Disk florets ca. 5, pubescent, 7-8 mm. Distribution. Found along the Sierra de Ma- nantlán Occidental of Mexico (Colima, Jalisco). Elevation ca. 2000 m. . Comments. Kowal (1991) used discriminant anal- ysis to support recognizing variety manantlanensis. The deltoid) lobes of characters distinguishing it are acute the leaf blade, broader phyllaries, fewer phyllaries and florets in the capitula, and larger ray and disk florets. His analysis is based on only seven and four collections of varieties galicianus and manantlanensis, respective- ly, so more data are needed to more firmly substantiate the recognition of this variety. MEXICO. Jaliseo: . en el ranc sho I n carretera para rancho 10 Dec. 1982, J. 1. Cerro la Representative specimens examined. mpio. Cuautitlán de la Joya, pine Pr ca. Calzda & Nieves H. 9461 TA in add le betw. Cumbre | and odd a Imeal. called Cloud Camp. 19733/40"N. 104°14'40"W, 2150 m, 12 Oct. 1982, H. H Iltis 28869 (KSC, MICH, WIS). 27. Roldana langlassei (Greenm.) H. Rob. $ 421. 1974. Columbian TYPE: Mexico. Sierra Madre, Brettell, Phytologia 27: Senecio langlassei | Greenm., Mus., Bot. Ser. 2: 283. 1907. "Michoacán et Guerrero: secu de 3—4 m, fleurs jaunes, sol granitique, 1600 m," 21 1899, E. Langlassé 1005 (lectotype, designated here, GH!; isotype, US not arbris- Sep. seen). Suffruticose herbs to shrubs, ca. 3 m tall; new growth pubescent with glandular hairs; stems terete, maculate, greenish yellow. Leaves cauline, evenly distributed on upper half of stem, lowermost decid- uous; petioles 11.5-26 cm, sparsely pubescent; blades 12.5- 11- to 13-lobed, sinus depth about 1/4 of 25 X 12.5-35 cm, triple-nerved, ovate- orbicular, the way to the midrib, lobes acute, slightly secondarily lobed; abaxially glabrate to pubescent with stipitate- hairs, adaxially gla- glandular persistent on veins, brous to sparsely pubescent with short glandular hairs. Capitulescence pyramidal compound paniculate cyme, 50 to 80 capitula per branch, peduncles densely pubescent with stipitate-glandular hairs, primary E bracts linear, up to 5 mm, ultimate peduncles ca 10 mm, bracteoles O to 4, 2-14 radiate, 1: linear. 2-3 mm, funnelform, bracts 1 to 3, phyllaries ca. 10, pubescent with stipitate-glandular ca. 2 mm. Capitula corollas yellow; calyculate linear, ] mm: hairs, mildly purple-tinted, ca. 6 X 1 mm. Ray florets ca. 5, corolla ligulate, glabrous, tube ca. 5 mm, ligule ca. 7 X 2 mm. Disk florets ca. 12, corolla funnelform, E glabrous, ca. 10 mm, tube ca. 5 mm, throat ca. 4 mm, lobes ca. | mm. Achenes glabrous, cylindrical. 1— 2 mm. ribs 10, resin glands absent: pappus bristles ca. 8 mm. Distribution and phenology. Mexico (Guerrero) at elevations of 1600—2700 m. Pacific slopes of pine-oak forests with Pinus, Abies, Clethra, and Quercus. Flowering December through February. Comments. Greenman (1907) cited Langlassé 1005 (B. GH) in the protologue: the Berlin specimen was destroyed in 1943. Roldana collected species. langlassei is an elusive and poorly Morphologically, it seems closer to the southern R. petasitis than with its geographically closer relatives. MEXICO. rero: Ca. a E de Campamento El suroccidentales del Cerro Teotey 17 28'N, 100 13 Je 2660 m. 27-29 Jan. 1965. R. McVaugh 132 (DS, MICH); 27 km al NE del Paraiso. camino a Puerto del Gallo, mpio. wv de Alvarez, 1720 m. 29 Mar. 1985, Marinez S. 3804 (KSC). Representative specimens examined. Guer- 2 km N ; lc | Gallo, estribaciones Rob. & Senecio 28. Roldana Brettell, Phytologia 27: : lanicaulis Greenm.. Publ. Field Columbian Bot. Ser. 2: 283. 1907. TYPE: Chiapas: near Pinabete. 1800-2400 m. 1896, E. Nelson 3771 (lectotype. here, GH!; isotypes, MO!, US not seen). (Greenm.) H. 1974. lanicaulis Mexico. 8 Feb. designated — = Simple or sparingly branched herbs or subshrubs, 1-3 m tall; tose, new growth densely white lanate tomen- often tawny; stems terete, light brown. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 7-12 cm, tomentose; 15-25 X 12-22 cm, 9(4)-lobed, sinus depth less than 1/8 of blades palmately nerved, reniform, the way to the midrib, lobes acute, margins densely denticulate, occasionally denticles up to 3 mm long: abaxially densely lanate tomentose and canescent, adaxially sparsely pubescent with glandular hairs. Capitulescence round topped to ps ramidal compound 318 Annals of the Missouri Botanical Garden 60 peduncles floccose tomentose, primary bracts linear, 0—3(—5) bracteoles 0 to 2, linear, to 1 mm. Capitula. radiate, 5-8 X 2-3 mm, turbinate, bracts ca. 3, linear or filiform, 1— 31 mm: phyllaries 7 to 1-2 mm. Ray florets 5 to 8, corolla ligulate, glabrous, tube 3—4 mm, ligule 2-4 X 1-2 mm. Disk florets 7 to 13, funnelform to moderately campanulate, glabrous, 4— 1-2.5 mm, l- cylindrical, 0.5-1 mm, paniculate cyme, 40 to up to 2 mm, ultimate peduncles mm, corollas yellow; calyculate 13, glabrous, purple-tinted, 3.5-5 X corolla 7 mm, tube 2—3 mm, throat lobes 1.5 mm. Achenes glabrous, ribs 5 or 10, resin glands absent; pappus bristles 4— 7 mm. Restricted to the Sierra Madre Oriental and south ward into Guatemala Distribution and phenology. (Guatemala: Huehuetenango: Mexico: Chiapas, Oa- 1000-2500 m. Montane cloud forests with Quercus. Pinus, Abies. Acer L., and Sauravia Spreng. Flowering November through March. xaca, Veracruz) at elevations of Styrax L., Comments. In the protologue for Senecio lanicau- lis, Greenman (1907) cited two syntypes: Guatemala. Dept. Quiché, Chiul, 2400 m, Apr. 1892, Heyde & Lux 3377 (GH not seen); and E. W. Nelson 3771 (GH. US Roldana southern Mexico and Guatemala lanicaulis is a distinct. species from i. It is distinguished by its tawny pubescence and reniform leaf blades that have numerous elongate denticles along the margins. examined. GUATEM rest al Cruz de oe de los ( Representative ALA. . betw. specimens Huehuetenango: wet cloud f San Mataeo Ixtatán € Nuca, 2800 m, 31 July 1942, J. j [1256275, 125627. ie MEXICO. Chiapas: on ridge above Silte 2 on e rd. to Huixtla, mpio. Siltepec, 2200 m, | Feb. 1982, ` Breedlove 98235 (C 1800— near Pinabele, 2400 m P b. W. Nelson 3771 (GH, MO). Oaxaca: 4.3 " 7 (m ssviacion a Zacatepec, 2400 m, 23 Apr. n Torres 2677 (MO, NY). racruz; pasando el map lel Rosario ak a Buena Vista, 2400 m, 28 Dec 986, Cházaro 4307 (WIS) 29. Roldana lineolata (DC.) H. Rob. € Brettell. Phytologia 27: 421. 1974. Senecio lineolatus DC.. Prodr. 6: 427. 1837 [1838]. TYPE: Mexico. |state unknown]: in Mexico Cordillera de Guchilapa, J. L. Berlandier 1227 (lectotype, designated by McVaugh, 1984: 836, G-DC not seen, microfiche IDC 800. 1141 1 4!, G-DC photos F!. MICH! MO!). Nov. non Senecio sinuatus Gilib., Senecio sinuatus Kunth, 818, Gen. Sp. (folio ed.) 4: 14.1. 1798. Roldana sinuata , Phytologia 87: 245. 2005 [2006], superfl. TYPE: Mexico. Guanajuato: betw. uasa justo & Santa Rosa, Humboldt lleg., nom. Turr nom iles E capitula per branch, de oo s.n. (holotype, GH!, 1D. P not seen, B photos F!, ¡pea tuberous thickened fibrous-rooted Single-stemmed tall, caudex: perennial herbs or subshrubs, 3m from ¢ ge new growth weakly arachnoid floccose to tomentose; stems terete or slightly ribbed, maculate, green to dark red. Leaves cauline, evenly distributed half of clustered at midstem when herbaceous; petioles 1— on upper stem, lowermost deciduous or 4 em, pubescent; blades 10-30 X 6—19 em, pinnately veined and lobed, 7 to 11 lobes, sinus depth to 1/3 of the way to the midrib, lobes acute to rounded, may be weakly secondarily lobed, leaf base cuneate; abaxially densely pubescent to lanate tomentose and canescent, glabrous. adaxially Capitulescence ascending com- pound corymbiform cyme, 15 to 40 capitula per branch, peduncles pubescent to densely arachnoid floccose, primary bracts linear, up to 5 mm, ultimate = 2mm. Capitula radiate, 10-17 X ca. 3 mm, funnel- a peduncles 5-20 mm, bracteoles 0 to 3, linear, form, corollas yellow; calyculate bracts 2 to 3, filiform, 1-7 mm; phyllaries 7 to 8, glabrous. sometimes purple-tinted, 8-9 X ca. | mm. Ray florets ca. 5, corolla ligulate, glabrous, tube 4-6 mm, ligule 6-10 ca. 2 mm. Disk florets 9 to 13, corolla funnelform. ca. 10 mm, throat 4-5 mm. = 2 mm, ribs 5, resin glands absent: pappus bristles 6— glabrous, tube 3—4 mm. lobes 1-2 mm. Achenes glabrous. cvlindrical. o ` Distribution and phenology. Central and southern (Distrito Jalisco, Mexico Guerrero, Hidalgo, Morelos, Oa- xaca, Puebla, Querétaro, Veracruz) at elevations of 2000-3500 m. Found in pine-oak forests with Pinus. Abies, Arbutus, Federal, Guanajuato, México, Michoacán, Quercus, Prunus, Senecio, Phytolacca, Baccharis, Symphoricarpus, and Alnus. Flowering October through February (March). Comments. In the protologue for Senecio line- olatus, de Candolle (1837) listed two syntypes from J. L. Berlandier 1226 (G-DC, lectotype) and J. L. Berlandier 1227 (G-DC not seen. IDC 800. 1141 I 5). sinuatus, Greenman notes, in hand. on the B specimen the the same locality, microfiche Regarding S. “doubtless a fragment of the type.” Unfortunately, Berlin specimen itself was destroyed in 1943 Roldana lineolata is a common species in Mexico. To the novice, it may be confused with R. heraclei- Jolia. See the discussion section under that species for an outline of diagnostice characters. Representative SEEEN examined. MEXICO. Distrito Feders > La Campana, Ajusco, 3000 m, Dec. 1964, J NE "19310 (D S, MICH, TNR mtn. slopes ca. 3 km E of S 2700 m. 11 Nov. 1970. R. MeVaugh 24204 v z- (MICH). Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana Guerrero: steep slope near summit of Cerro Teotepec, 3000 m, 12 Nov. 1973, D. E. Breedlove 36082 (CAS, MICH, MO). Hidalgo: along es 130 betw. Poza R City, near reservoir, 23 km W of Huauchinango, 23 Oct 1986, T. M. Barkley 3917 (KSC, NY, WIS). Jalisco: plata NNE of Cerro San Miguel € 4 km S of ur ón de P E 19734'N, 104^12'W, 2300 m, 10 Jan. 1980, R owal 2838 (MICH, WIS). México: Llano ae hasta B cima, 3000 m, 20 Nov. 1965, Boege 26 (CAS); c km N of summit betw. Toluca & Mexico City, on the PR i 3100 m, 20 Oct. 1970, A. Cronquist 10824 (CAS, . MICH, NY); 19 km E of Texcoco, 1 km S of rd. o '0— span n Tepetlaoxtoc, 2800 m, 28 Sep. 1976, García P. 116 S. MO). Michoacán: Mt. Patamban, ur m, 29 Jan. ia P W. Nelson 6578 bs H, MO); rd. to jos Azufres, 19 km by rd. from Mex 5, 6km W of a iudad Bd D N, 100737' W, 3400 m, 19 1975, Steingraeber 21 (WIS). re on hwy. 95 betw. Mexico City & Cuernavaca, ca. 1 km N of Tres Marías, 17 ih 1986, T. d 3901 (KSC, WIS; Tres Marías, 2900 m, 7 Nov. 1903, C. G. Pringle 11569 (CAS, F, GH, MICH, NMC). d xaca: n to La Cumbre, 17 mi. lumber rd. to lumber camp, 9 mi. along crest San Felipe range, 2000 m, 17 Jan. 1965, Carlson 4012 (F, MICH, NY). Puebla: above & 8 of hwy. Mex. 190 betw. Puebla & Mexico City, 4 km E of Río Frio recreation area & Mexico-Puebla state line, 2800 m, 7 Jan. 1981, M. Nee 19592 (NY). Querétaro: Puerto de Agua Fria, ca. 10 km 5 of Pinal de Pinal de Amoles, 2800 m, 5 Sep. 1985, 8 km E of Los Altos 912 W center 4—5 km from Oaxaca, E on Amoles, mpio. Fernandez 3065 (NY). Veracruz: Veracruz & 12 km W of Ayahualulco, 19%24'N, 97 2900 m, 7 Nov. 1981, M. Nee 22869 (F, WIS). 30. Roldana lobata La Llave, Nov. Veg. Descr., Fasc. 2: 10. 1825. Senecio roldana DC., Prodr. 6: 431. 1837 [1838], non Senecio lobatus Pers., 1807; nec Senecio lobatus Sessé & Moc., 1894. TYPE: Mexico. [s. loc., 1831, Alaman s.n. designated here, G-DC not seen, microfiche IDC 800. 1142 I 3-4!, G-DC photos F!, MO!). (neotype, Senecio rade S. Schauer, Linnaea 20: 698. 1847. TY o. Hidalgo: ca. Zimapán, Aschenborn 691 (lec UE aa here, B fragm. GH!). ., Pl. Nov. Hisp. 140. 1887 non EM rotundifolius Stokes, Y Senecio ion Sessé & Moc ie 90]. nom. illeg., 812; nec Senecio Cia pd 1813; nor a rotundifolius Hook. f., YPE: Mexico. > unknown]: Xochitlán, ne P Ys 2820 », designated here, MA not seen, MA photo Amer. Acad. Arts 26: Mexico. Jalisco: 1889, C. G Senecio jaliscanus S. Watson, Proc. 143. 1891, as “jaliscana” ' Chapala Mtns., near Guadalajara, 9 Dec. Pringle 2931 (holotype, GH!). Roldana igs gi ate B. L. Turner, e 31: 230. 2005 [2006]. TYPE: Mexico. Oaxaca re ee 6-7 km from El Manzana along rd. to Infiernillo, 17°12'N, 9804'W, pine-oak forests, loc sally d 13 Feb. 1996, J. I. Calzada 20776 (holotype, EX not seen, TEX digital image TEX!). Single-stemmed perennial herbs or unbranched shrubs, 1-4 m tall, arising from a woody caudex; glabrate below, upwards new growth arachnoid pubescent to floccose tomentose, at times canescent: stems terete, red with patches of cream-green. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 7-13 cm, glabrescent, arachnoid pubescent to floccose tomentose; blades 5-25 X 10-15(-20) cm, triple-nerved, blade shape various: oval-subpalmatifid, 5- to 7-lobed, sinus depth 1/4 of the way to the midrib, lobes acute, first set of proximal lobes longer and broader than others, giving blade a rhomboid aspect; or ovate, (3 to)5- to 7-lobed, sinus depth less than 1/8 of the way to the midrib, lobes acute; or occasionally rotund-oval, margins subentire; all margins sparsely callous denticulate, abaxially arachnoid pubescent to lightly tomentose and canescent, adaxially glabrous. Capitulescence elongate pyramidal compound paniculate cyme, 10 to 30 capitula per branch, peduncles arac hnoid pubescent to densely floccose tomentose, bracts linear, 2-5 mm, ultimate peduncles (0—)5— 10 mm, bracteoles O to 2, linear, de 2 mm. Capitula 7-13 X 3-5 corollas yellow to light orange; calyculate bracts ca. eradiate or radiate, mm, turbinate, 5, linear or filiform, 2-3 mm; phyllaries 10 to 13(to 14), arachnoid pubescent to tomentose, purple-tinted, 3.5-6.5 X 1-2 corolla ligulate, pubescent, hairs typically restricted to the top of the tube, tube 2-5 mm, ligule (2.5-)4-8 X 2-3 mm. Disk florets (11 to)l5 to 22, funnelform to moderately campanulate, pubescence mm. Ray florets absent or 3 to 6, typically restricted to the top of the tube, (? rarely tube 2-3.5 mm, throat 2- Achenes glabrous, cylin- glabrous), 5-8.5 mm, 4.5 mm, lobes 0.5-2 mm. drical, 1-3 mm, ribs 5 or 10, resin glands absent; pappus bristles 5-8 mm. Distribution and phenology. Central and south- western Mexico, Guatemala (Guatemala: Jalapa: exico: Distrito Federal, Guanajuato, Guerrero, Hidalgo, Jalisco, Méxic Oa- xaca) at elevations of 1300-2700 m. Ravines, steep , Michoacan, Morelos, mountainsides, in tropical subdeciduous forest with Quercus, Fraxinus, Prunus, and Pinus. Flowering October through February. Comments. The type for Roldana lobata came from an unknown locality in Mexico and grown in the Royal Botanical Garden, Mexico City. Specifically, La Llave (La Ll 1825) reports in the protologue “Januario floret in horto mexicano.” The La Llave collections are cited in Taxonomic Litera- ture, 2nd edition (TL-2), i ; type collection of R. lobata was not found in G by ave & Lexarza, as deposited at G, but the — Laurent Gautier (pers. comm., Sept. 2004). A type collection of this plant is thus presumed nonexistent. en de Candolle (1837) made the nomen novum Senecio roldana, he cited two gatherings in addition to 320 Annals of the Missouri Botanical Garden R. lobata. He cited the gatherings with their herbarium names, which were the replaced synonym, nomina nuda and, hence, not validly published. These invalid names are included here, as they provide an additional reference when viewing the collections. The gatherings cited by de Candolle are both labeled from Mexico without locality: Cineraria angulata Alaman, nom. nud., 1831, Alaman s.n. (G-DC not seen, microfiche IDC 800. 1142 1 34!; G-DC photos F!, MOD): and Cineraria lobata Mairet, 1833, E. M. Mairet s.n. (K not seen, 800. 1142 I 2N. material is nom. nud., microfiche IDC because no original R. lobata, Alaman s.n. (G-DC) is chosen here as the neotype. Therefore, available for the specimen Regarding Senecio schumannianus, Schauer (1847) lists Aschenborne 691 in the protologue but does not designate a herbarium. No information is provided on L-2. Nevertheless, fragment from Berlin at GH, which is believed to be of the the B destroyed in 1943, the GH fragment, while mislabe Sebastian Schauer in T there is a type material; because specimen was ed “Aschenborn 641,” is chosen here as the lectotype. No collection was cited in the protologue (1887) for Senecio rotundifolius Sessé & Moc.; the lectotype was chosen from original Sessé and Mociño material Roldana lobata is a distinct and common species. lts sometimes rhomboid-like leaf shape and dense floecose to. tomentose arachnoid pubescence are unique. McVaugh (1984: Nueva Galicia, 832) notes that material from Jalisco, is strietly discoid, a feature used to circumscribe Senecio jaliscanus. However, MeVaugh agrees that this distinction is not enough to separate the two species. He also states that the radiate form is known from central Michoacán eastward, | can neither confirm nor deny this observation. Gibson (1969: variety calvescens based on its being more coarse and 149) contemplated the with stems more herbaceous, having somewhat smaller heads, and a lack of floccose tomentum. Specimens he cited with these characters have been placed ii Roldana kerberi or R. petasitis var. oaxacana. Roldana juxtlahuacana differs from the type of R. While noleworthy, it it does not separate it from my species lobata by having less pubescence. this is concept of R. lobata. In addition, there are plants to the north with less pubescence as well (México: García 206). While Turner (2005) reports the type specimen to have glabrous corolla lobes, which would be a possible distinguishing character, the material 1 have seen that fits the description and distribution of R. ee has weakly pubescent corolla lobes. Therefore, the s until further specimens are examined. species is placed in synonymy here GUATEMALA. Ja- lapa: mtn. rd. betw. Jalapa € Paraiso, 1400-1700 m, P. € Standley 77351 (MO). MEXICO. Distrito Federal: 2600 m, C. G. a 11571 (CAS, GH, MICH). nm. W Chilpancingo, rd. to Omiltemi, Oct. a h. 21907 Representative. specimens. examined. — z => WeVaugh cerca de 2500 m, 27 Dec. MICH. "Hidalgo: ca. Zimapan, drawing at GH). Jalisco: N of Sierra de ine of El Chante, 19 36'N. 104. 13'W, 1500 m, 14 s R. Kowal 2853 (MICH, WIS). | km N betw. Tramo-Tlalmanalco-Amecameca, 1976. García 206 ee F. NY): dist. Temascáltepec 1500 m, 4 Dec. 1932, B. Hinton 2859 (GH. MO. NY). Mie hoa 'án: S-facing o s of mtns. betw. Río del e & f Morelia, 2200 m, 9 Nov. 1961. deca along free rd. 25381 Aschenborn 691 ly zm. & e 3 b ms 980. R. N of S that 2300 m. 12 = cc 3 Ls pie Cumbres, yp 14231 (CAS. MICH). betw. pr “a & Pochutla, 7: . 2800 m. iv Jan. 1979, T. B. Croat 46166 . Putla, 18 km N of Putla de ( Det. 1980. 0. Tellez pm s dum lexico. City to Cuernav: Tres , 16 Jan. 1966, D. E Oaxaca: along hwy. 7 ruerrero, 21 (B... L. Kew Bull. 47: 55. 1992, as “metepecus.” Senecio metepecus B. L. Turner, Phytologia 66: 465. 1989. TYPE: Mexico. Hidalgo: mpio. de Doria, 18 km NNE of | Metepec, * terracería que va de Metepec a Tenango de Doria.” 2200 m. 31 Oct. 1977. J. García P. 514 (holotype. TEX not seen, TEX digital image TEX! isotypes, CHAPA not seen, KSC!). 3l. Roldana metepeca Turner) Jeffrey. Tenango de "sobre camino de tall. spreading from creeping rhizomes; new growth stipi- Suffruticose herbs and subshrubs, 14 m tate-glandular pubescent with purplish elongate crin- kled trichomes: stems terete, purple. Leaves cauline, clustered at base and midstem; petioles 7.5-10.5 cm, stipitate-glandular pubescent: blades 4.5-12 X 9- 15 em, triple-nerved, subpalmatifid to ovate. 5 to 7 lobes, lobes acute, weakly secondarily lobed to denticulate: abaxially densely stipitate-glandular with long multicelled hairs along veins, adaxially sparsely pubescent with short glandular hairs. Capitulescence simple paniculate eyme, 2 to 5 capitula per branch, peduncles densely stipitate-glandular pubescent, pri- X 3 mm, bracteoles O to mary bracts linear-elliptic, ca. ultimate peduncles 24-40 mm. linear to weakly obovate, 2-5 X 2 mm. Capitula eradiate. ca. 16 X 4 mm, turbinate, corollas yellow; calyculate bracts 3, linear, 5 mm; phyllaries 8 to 13, eee ae glandular pubescent, purple-tinted, ca. 12 X 3 mm. Disk florets 20 to 30, corolla narrow, eode ca. 10 mm, tube ca. 4 mm, throat ca. 4.5 mm, lobes ca. .0 mm. Achenes glabrous, cylindrical, ca. 3 mm, ribs 5. resin glands absent: pappus bristles ca. 10 mm. Distribution and phenology. Central Mexico (Hi- dalgo, Puebla, Oaxaca) at elevations of 2000-2200 m. Volume 95, Number 2 Funston 321 2008 Taxonomic Revision of Roldana Found in pine-alder forests. Collected in flower versus the numerous florets (90+) and obovate August and October. calyculate bracts found in R. suffulta. Comments. | Roldana metepeca is a poorly collect- Representative iis examined. MEXICO. Guer- ed species. It has eradiate capitula and a densely stipitate-glandular pubescence. While other species in the genus have red stems and may have light purple trichomes, the deep violet purple stems and trichomes in this species are unique. Representative specimens examined. 8-11 km SW of Tenango de Doria, E. Breedlove 59570 (CAS). Oaxaca: Maria Tiltepec. 10 km SW of Totontepec, 2360 m, 20 Jan. 1984, R. Torres C. 4608 0 Infiernillo, Dec. 1954, E. Lyonnet MEXICO. Hidalgo: 1830 m, 30 Oct. 1983, D. zi (MO). Puebla: El 541200067 (MO). Rob. & 32. Roldana mexicana (McVaugh) H. Brettell, Phytologia 27: 421. 1974. Senecio mexicanus McVaugh, Contr. Univ. Michigan Herb. 9: 473. 1972. TYPE: Mexico. Michoacán: hills of Pátzcuaro, 11 Nov. 1890, C. G. Pringle 3332 (holotype, MICH!; isotypes, Fl. GH!, MICH!, MO!, NY!, UC”). 12.5 m tall, plants essentially Single-stemmed perennial herbs, arising from a rhizomatous caudex: glabrous; stems terete, striated, cream to dark red in color. Leaves cauline, clustered at midstem, reduced upwards; petioles 7.5-28 cm, glabrous; blades 8-28 X 8-28 cm, triple-nerved, suborbicular to subovate, sinus about 1/4 of to the T- to 13-lobed, the way midrib. lobes acute or rounded, lo and proximal veins of larger weakly secondarily = ved to merely denticulate; surfaces glabrous, margin leaves with short elandular hairs. Capitulescence flat-topped to rounded compound paniculate cyme (umbel-like), 10 to 20 capitula per branch, peduncles glabrous, primary bracts linear, up to 5 mm, ultimate peduncles 3— 15 mm, bracteoles 3 to 5, scale-like to linear, 2- 3 mm. Capitula eradiate, 8-15 X 3-5 mm, funnel- form, corollas orange to orange-yellow; calyculate bracts ca. 5, scale-like, 1-2 mm; phyllaries ca. 8, glabrous, greenish purple-tinted, 7-13 X 1-2 mm. Disk florets 9 to 12, corolla funnelform, glabrous, 8— 10 mm, 2 mm. Achenes pubescent, cylindrical, 1-2 mm, ribs TJ o tube 3-5 mm, throat. ca. 3 mm, lobes ca. 10, resin glands absent; pappus bristles 7-9 mm. Distribution and phenology. Central Mexico (Guerrero, Jalisco, México, Michoacan, Oaxaca) at elevations of 1500-2500 m. Found in pine-oak forests with Pinus, Quercus, and Alnus. Flowering October through December. Comments. | Roldana mexicana is a distinct spe- cies that resembles R. suffulta; however, they are easily separable by the fewer florets (9 to 12) and scale-like calyculate bracts found in R. mexicana 36 km NE of on ridge Bre uae rero: Pueblo El Gallo, rd. to Filo del ze NE of Teotopec, 2200 m, 13 Nov. 1973, D. E 36123 m CAS, MICH); along hwy. 95) & A c de Alvavez, ae mi. W of turn-off onto rd. Chichihualo 2 m, 14 Jan. j 561: (MO). Jalisco: Cerro de Santa Fé, o. pae Zapotlá- 1600 m, 21 Oct. 1972, Díaz L. 3571 (F, MICH, NY, ). Mé a, dist. RAO spec, 26 Nov. 1935, G. B. Hintor ; MICH, NY). Michoacán: of Morelia, 2100 m, 15 Sep. nejo, 5) ~ 2 a 2 ps pan] Lx o erro Azul, vic. 2697 (GH, MO); Pátzcuaro, Holway 3182 (GH); oak zone 5 mi. N of Pátzcuaro, 2900 m, 25 Oct. 1962, R. McVaug xh 21929 (MICH). Oaxaca: 2 Coicoyán, Sierra Su Juxtlahuaca, Siki Nami al de Coicoyán, 17%16 98 17'W, 2100 m, 5 Nov. oo A. de Avila 471 (MO). 33. Roldana mezquitalana (B. L. Turner) Fun- ston, Novon 11: 305. 2001. Senecio mezquitala- nus B. L. Turner, Phytologia 71: 56. 1991. Senecio gesnerüfolius B. L. Turner, Phytologia 62: 75. 1987, nom. illeg., non Senecio gesner- ifolius Cuatrec., 1950. Roldana gesneriifolia C. Jeffrey, Kew Bull. 47: 54. 1992, nom. superfl. TYPE: mplo. 26.5 km de La Guajolota por el Mexico. Durango: Mezquital, camino ¢ Platanitos, ca. 2610 m, 15 Mar. 1985, M. González 1677 (holotype, TEX not seen, TEX digital image TEX!). Senecio diode B. L. Turn ED s e 367. 1993. yn. nov. Roldana floresi siorum (B. er) B 3): 222. a E Totatiche, Ranchi On urner, ' Phytologia 8T > Temastian ca. 20( A. Flores M. & M. Flores M. 2688 (holotype, seen, TEX digital image TE X!; isotype, IGE not seen). shrubs, ca. Suffruticose herbs or 2 m tall: new growth densely white tomentose; stems terete, dark red. Leaves cauline, evenly distributed on upper half petioles 2-2.5 cm; 34 cm, pinnately veined, elliptical of stem, A deciduous; blades 10-15 to MN apex acute, base cuneate, margins irregularly serrulate; surfaces glabrate, pubescent on veins. Capitulescence terminal somewhat rounded corymbiform cyme, ca. 18 capitula per branch, peduncles densely stipitate-glandular | pubescent, primary bracts linear, ultimate peduncles | (3— 5)6 mm, bracteoles 2 to 3, linear, ca. | mm. Capitula 7-8 X ca. yellow; calyculate bracts 6 to 8, linear, 1-2 mm; phyllaries 8 or 11 to 13, glabrous, (4—5)7—8 mm high. Ray florets 5 to 6, corolla ligulate. glabrous, tube ca. 5 mm, ligule (2-4)6-7 X radiate, 7 mm, campanulate, corollas 2.5—3 mm. Disk florets ca. 18, corolla funnelform, 8 glabrous, ca. 8 mm, tube ca. 3.5 mm, throat ca. 3.5 mm, lobes ca. 1 mm. Achenes 322 Annals of the Missouri Botanical Garden glabrous, cylindrical, ca. 2 mm, ribs 5, resin glands absent; pappus bristles 5-7 mm. (Description com- piled from the literature.) Distribution and phenology. Mexico 00-2600 m. pine-oak forests. Collected in flower in March. (Durango, at elevations of 20 Found in = Jalisco Comments. Roldana mezquitalana, like R. hinto- schaffneri, former two species have phyllary numbers of 8 or 11 nit, superficially looks like but the to 13 and occur in west-central to northern Mexico, whereas R. shaffneri has 5 to 6 phyllaries and prolific in southern Mexico and Central America. Roldana mezquitalana is separated from R. hintonii by a denser pubescence, serrulate leaf margins, geogra- phy, and, in minor ways, by characters related to the capitulescence and capitula. | have only seen an illustration of this species, and it is possible that it is conspecific with R. hintonii. Roldana floresiorum is described as being identical to R. mezquitalana except for differences in capitu- lescence morphology and in having smaller capitula. In my overall concept of the genus, species based on such differences, when they occur within the same range, are not recognized; therefore, R. floresiorum is placed in synonymy and the description of R. mezquitalana is broadened to include plants with ultimate peduncles 3-5 mm; phyllaries 8, 4-5 mm high; and ligules 2-5 mm. 34. Roldana michoacana (B. L. Rob.) H. Rob. & Brettell, Phytologia 27: 421. 1974. Cacalia michoacana B. L. Rob., Proc. Amer. Acad. Arts 43: 46. 1907. Pericalia michoacana (B. L. Rob. Rydb., Bull. Torrey Bot. Club 51: 377. 1924. Senecio michoacanus (B. L. Rob.) B. L. Turner & T. M. Barkley, Phytologia 67: 392. 1989. TYPE: Mexico. Michoacán: == on a pine-covered crater 5500 ft. 31 Oct. 1905, C. G. Pringle 10117 (holotype, GH not seen, GH photo MICH!; isotypes, F!, NY! cone, Uruapan, Cacalia a 5. F. Blake, : Wash. Acad. Sci. 280. 19 TYPE: xico. Jalisco: in dense growth dern stream on steep m trail to La Sabala 2s und Sebastian, Sierra Madre Mtns., 1500 m, 10 Feb. 1929, Ynes Mexia 1656 (holotype, US 1318107 US photo MICH! isotypes, F!, MICH!, NY!) Sy n. NOV, not seen, Small single-stemmed perennial herbs, 0.5-1.5 m tall, arising from a bulbous caudex with woolly buds: plants essentially glabrous, may have scattered pubescence of short stipitate-glandular hairs; stems terete, striated, greenish to red in color. Leaves cauline, clustered at midstem; petioles 5-12 cm, glabrous; blades 4-9 4-13 cm, triple-nerved, subpalmatifid, 3- to 5(7)-lobed, sinus depth about 1/4 of the way to the midrib, lobes acute, weakly secondarily lobed to denticulate; surfaces glabrous, margins and proximal veins may be weakly pubescent on larger leaves. Capitulescence simple to compound paniculate cyme, l to 4 capitula per branch, peduncles glabrous, primary bracts linear, ca. 3 mm, ultimate peduncles 10-100 mm, bracteoles 0 to 3, linear, 2-5 mm. Capitula eradiate, 10-20 X 3-6 mm, campanulate, corollas yellow-cream or white; calycu- late bracts 5 to 10, filiform, 2-10 mm; phyllaries ca. 13, glabrous, purple-tinted, 10-12 X 1-2 mm. Disk florets 36 to 58, campanulate, glabrous, throat 4-6 mm, cylindrical, ca. 2 mm, ribs 10, resin glands absent: corolla funnelform to slightly 10-13 mm, tube lobes 2-3 mm. 4-5 mm, Achenes glabrous, pappus bristles 8-11 mm Distribution and phenology. | Central Mexico (Ja- México, 1500— 2500 m. Pacific slopes, pine-oak and fir forests with isco, Michoacán) at elevations of Pinus, Quercus, Ostrya, Cornus, Meliosma, and Podocarpus. Flowering November through January. Comments. Roldana michoacana is a somewhat smaller version of R. suffulta. While these two species are difficult to separate, at times they are distinct. Roldana michoacana has filiform calyculate bracts, glabrous achenes, a simple to compound paniculate cyme, and typically smaller capitula. Roldana suffulta has obovate calyculate bracts, pubescent achenes, a compound corymbiform cyme, and typically larger capitula. MEXICO. epi betw. s onica & N to Aserradero Agua Blanca, ca. °30'W, 2275 m, 13 Nov. 1952, R. McVaugh ae MICH, es steep mtns. 11—12 mi. 5 of Talpa de Allende, headwaters of W branch of Río de Talpa, 1450 m. 24 Nov. 1960, Pippen 63 ieri WIS). México: 8 km NE of Temascaltpec, 2070 m, 25 Nov. 1983, D. E. Breedlove s s NL dist. Pu 5 De -~ B. Hinton 12718 (GH, MICH us carretera Morelia, por Mil Cumbres, ap Ciudad Hidalgo, 2450 m, 11 Oct. 1983, Martínez S. 4717 (KSC). Representative specimens examined. = & Villaseñor, Mexico. 35. Roldana mixtecana Panero Brittonia 48: 83. 1996. TYPE: Dist. largo de la nueva brecha que lleva a campos de fresas al W de Cañada Lobos, 1895 m, 17%19'53.7", 98*08'12"W, 24 Oct. 1994, J. L. E. Manrique & J. Il. Calzada 5409 (holotype, MEXU not seen; isotypes, MSC not seen, TEX not seen, TEX digital image TEX!). Oaxaca: Juxtlahuaca, Cañada del Río Ratón a lo Panero, Perennial herbs, 0.5-1 m tall; plants sparsely to moderately pubescent, arachnoid, burgundy: stems 3 to 5, only 1 to 3 producing flowers, terete. Leaves cauline, evenly distributed along stem, reduced Volume 95, Number 2 2008 Funston 323 Taxonomic Revision of Roldana 36 XxX 1.3- 6.5 cm, subcordate-deltoid, 5 shallow lobes to merely 5 or 3 simple corymbiform cyme, 3 upwards; petiole 0.3-2.7 cm; blade 1. dentate; surfaces glabrescent. Capitulescence to 24 capitulescence, peduncles 2.5-6 cm, sparsely pubes- capitula per cent, green. Capitula radiate, 1.2-1.5 X 0.7-1 cm, campanulate, corollas bright yellow: ee 25-10 series in 5 graduated somewhat pus ellas of first to third series 2-7 X 1.2-1.7 mm, appressed, chartaceous, ovate to lane ies glabrous, stramineous-green rimmed with purple; phyllaries of fourth to fifth series ca. 10 X 1.7-2 mm, appressed, chartaceous, ovate to oblong, glabrous, stramineous. Ray florets 5 to 7, corolla ligulate, glabrous, tube ca. 4 mm, ligule 9-10 X 3—4 mm. Disk florets 20 to 27, corolla campanulate, glabrous, ca. 8 mm, tube ca. 3.5 mm, throat 2-2.5 glabrous, columnar, ca. 2.5 mm; pappus brist 2 mm, lobes mm. Achenes — es 5.5— 6.5 mm. (Description compiled from the literature.) Distribution. Mexico (Oaxaca) elevation ca. 2000 m. Found on well-drained, shady, rocky slopes of pine-oak forests. Comments. Roldana mixtecana is morphologically an outlier in the genus due to its multiseriate phyllaries, but in all other aspects it has the gestalt of Roldana. The authors compare it to R. michoacana and R. My impression was that it was the rarely collected R. — sessilifolia in the protologue. initia hederifolia; however, the multiseriate phyllaries argue against this. Having seen only a digital image of the type specimen, I will refrain from further comment and recognize the species as circumscribed. 36. Roldana neogibsonii (B. L. Turner) B. L. Phytologia 80: 278. 19906. Senecio Turner, Brittonia 37: 119. neogibsonit (B. L. Turner) 304. 2001, isonym. TYPE: Mexico. Veracruz: mpio. Huayacocotla, orilla del camino entre Helechales y Ocotes, 20 39'N 98^26' W, 1750 m, 2 Oct. 1980, J. García P. 177 (holotype, XAL not seen; isotype, TEX not seen, TEX digital image TEX!). Turner, neogibsonii B. L. 1985. Roldana Funston, Novon ll: : Shrubs or trees, ca. 3 m tall; new growth densely Leaves half of stem, velvety tomentose; stems terete, grey-green. cauline, evenly distributed on upper lowermost deciduous; petioles 0.8—1.5 cm, tomentose; blades 10-15 -2.3 cm, pinnate to pinni- palmate, oblong, apex acute, base cuneate, margins subentire, weakly callous denticulate, in-rolled; abaxially glabrate to light velvety tomentose, adaxially glabrous, dark green. Capitulescence pyramidal com- pound paniculate cyme, 10 to 20 capitula per branch, bracts linear, ca. peduncles velutinous, primary | mm, ultimate peduncles 1-4 mm, bracteoles O to 2. scale-like, ca. 1 mm. Capitula radiate, ca. 2.5 mm, turbinate, corollas yellow; calyculate bracts 3 to 6, velutinous, linear, ca. 1 mm; phyllaries 11 to 14, glabrous to slightly pubescent, purple-tinted, 4-5 X ca. 1 mm. Ray florets ca. 8, corolla ligulate, glabrous, tube ca. 4.5 mm, ligule ca. 4.5 X 2 mm. Disk florets 9 to 11, corolla narrow, glabrous, ca. 7 mm, tube ca. 3 mm, throat ca. 3 mm, lobes ca. 1 mm. Achenes glabrous, columnar, ca. 2 mm, ribs 5 or 10, resin glands absent; pappus bristles 5-6 mm. Eastern Mexico (Hi- Veracruz) at elevations of 1800- Distribution and phenology. dalgo, Puebla, 2000 m. Found in pine-oak forests. Flowering No- vember through December. Comments. Roldana neogibsonii is a unique and easily recognized species. It has a persistent velvety tomentum on its stems, on the undersides of the leaf dlades, and in the capitulescence. Its leaf blades are up to 5X longer than wide with subentire in-rolled margins; the callous denticles may require a hand lens to see. Representative specimens examined. MEXICO. Hidalgo: Zacualtipán, 1800 m, Jan. 1940, Martínez s.n. (F); Zacualti- 1800 m, 15 Dec. 1939, O s.n. (GH). Puebla: mins. W of Huauchinango, 1 Nov. 1943, Lundell 12633 MICH); Puente E of Svid ee uila, rd. to Zacatepec, oa io. Huauchinango, 20°5'N, 98 E W. 2000 m, 26 Feb. 1987, Tenorio L. 12688 (KSC). Veracruz: mpio. Huayac al, orilla del camino es PE hales y Oc “oles, 20°39'N, 98°26'W, 1750 n 980, J. García P. 17 TEX Rose image TEX. - pán, m H Madera, — y = 37. Roldana petasitis (Sims) H. Rob. & Brettell, Phytologia 27: 423. 1 Sims, Bot. Mag. 37: t. 1536. 1813. Senecio petasitis (Sims) DC., Prod. 6: 431. 1837 [1838]. YPE: same as that of Senecio lobatus Sessé & 74. Cineraria petasitis Moe. (neotype, designated here, F!). Cineraria eui Schrank, Pl. Rar. Hort. 95. 1819 [1821]. TYPE: same as that of pie mu Sessé & M. (ne va designated here, F!) Fl. (ed. 2 illeg., non Senecio lobatus Pers., Mexico. [s. loc.], Sessé & Mociño ar E d. designated here, ". n photo MICH 17 Senecio petasioides Greenm. inJ. D. Smith, E ux 37: 419. 1904. Syn. nov. Roldana — ee (Greenm.) H. Rob., 1975. TYPE: Guatemala. Santa B Hs 2. Heyde et Lux. , NY!, US not apis lobatus Sessé & M Mexic. a S 185. 1894, 7. TYPE nom. Phytologia 32: 331. Rosa: Cenaguilla, 4000 " 4522 (holotype, GH!; isotypes, F!, MO! n). Senecio prainianus A. Berger, Gard. Chron. 3: 82. 1911. Syn. nov. TYPE: same as that of Senecio lobatus Sessé & Moc. (neotype, designated here, F!). 3 m tall; growth densely pubescent with stipitate-glandular and Suffruticose herbs or shrublets, 0.5— new 324 Annals of the Missouri Botanical Garden long multicelled hairs on stems and lanate tomentose D on abaxial leaf surfaces; stems terete, maculate. dark red to greenish brown. Leaves cauline, evenly — distributed on upper half of stem, lowermost decid- uous; petioles 2-14 cm, variously pubescent with glandular hairs; blades 5-23 X 5-23 cm, thick or at times membranous textured, triple-nerved, shape of two types: variously orbicular to subovate, 7- to 9- lobed, sinus depth about 1/8 of the way to the midrib or less, lobes acute to rounded: or subpalmatifid, 5- to i-lobed, sinus depth about 1/2 of the way to the midrib, lobes acute: abaxially variously glabrescent, e pubescent, or tomentose, adaxially glabrous to pubescent with short glandular hairs. Capitulescence 10 to 60 ensely stipitate-glandular pubes- 10- mim. compound paniculate eyme, capitula per branch, peduncles « cent, primary bracts linear, 5-15 mm or obovate, 8-20 Capitula 30 X 5-15 mm, ultimate peduncles filiform, up to 3 mm. 6-16 X corol las vellow : bracteoles 0 to 3, eradiate or radiate, 2—4 mm, campanulate or turbinate, calyc ws bracts 0 to 3. linear, 1-2 mm; phyllaries ca. 8, densely stipitate- glandular pubescent, 3-9. X s mm. Ray florets absent or 3 to 8, corolla ligulate or reduced to a tube, glabrous, tube 4-6 mm, ligule 2-10 X 1-4 mm. Disk florets 5 to 15, corolla campanulate, glabrous, 6— llo mm, tube 3-5 mm, throat 3-5 mm, lobes ca. 2 mm. Achenes glabrous, cylindrical, ca. 2 mm, ribs 10, resin glands absent; pappus bristles 6-8 mm. Guatemala, (El Salvador: Santa Ana, Distribution and phenology. Mexico, El Salvador, Ahuachapán, Chalatenango, San Salvador, San Honduras, and Nicaragua Vicente, Sonsonate: Guatemala: Alta Verapaz. Baja Verapaz. Chimaltenango, El Progreso. Guate- J | B 8 mala, Huehuetenango, Jalapa, Jutiapa, Quezalte- nango, San Marcos, Sololá, Totonicapán, Zacapa: Honduras: Comayagua, Copán, El Paraíso. Intibucá. yag | La Paz, Mexico: Chiapas, Hidalgo, México, Oaxaca, Puebla, Tamauli- Lempira, Morazán, Ocotepeque: pas, Veracruz: Nicaragua: Estelí, Madriz) at elevations of 1000-2500 m. Found in montane cloud forests and pine-oak forests with Pinus, Quercus, Liquidambar, Ulmus La December through April. Weinmannia L.. and Styrax. Flowering A single naturalized roadside population in Victo- ria, Australia, is reported in Flora yd Victoria (ed. D. B. Forman & N. G. Walsh, 4: 968, f. 198d. 1999). Comments. The typification of Roldana petasilis is confounded by the fact that it is a pretty plant: several names were described from horticulturally tended plants that were never collected for specimens. The names without type material are Cineraria petasitis Sims, C. platanifolia Schrank, and Senecio prainianus A. Berger The details concerning the search for type material are listed below. In the protologue of S. Sessé and Mociño (1894) did not list any specimens: the lectotype was chosen from original lobatus, Sessé and Mociño material at F. The same specimen is designated here as the neotype of C. petasitis, C. platanifolia, and S. prainianus, thereby making the four names homotypic. Cineraria petasitis Sims: based on a plant grown in England by A. B. I material found at G by Laurent Gautier Oct. 2004). No material found at BM Hinds (pers. comm., Nov. 2004). Cineraria platanifolia Schrank: no material found Oct. 2004). No collected by ambert, blooming in Dec. 1812. No pers. comm., or K by Peta t K by Nicholas Hind (pers. comm., that Schrank at M by Franz Schuhwerk (pers. comm., Oct. 2004). Senecio prainianus A. Berger: ard found was definitely based on cultivated plants "Mexico, in montibus supra Vera Cruz; frutex a Hortulum cl. Horto No herbarium gather- l. A. Purpus lectus et in C. Sprenger floret in mensibus, Febr.-April, 1910." p introductus, primum Mortolensi ing is noted. Roldana species. petasitis is a beautiful and common has been widely collected and grown in cultivation. It typically has a lanate tomentum on the abaxial leaf surface; the tomentum often with a silky distribution in appearance. At its northern-most central Mexico, the tomentum is lost from the abaxial surface and the leaves are membranous in texture. In Oaxaca and Veracruz, some populations have devel- which also lacks a oped a more palmatifid leaf. tomentum and has smaller capitula. Plants in Oaxaca and Chiapas. Mexico, and Guatemala sometimes have smaller eradiate capitula and abaxial leaf surfaces only at most lightly tomentose. In the seribed based on the above pattern. However, when past, several species have been circum- viewing this complex of plants as a whole, the gross morphology makes it apparent that they are all of the species. the morphological same species. As one variability from north to south is recognized at the level of varieties. The long list of synonyms for Roldana petasitis is Cineraria platanifolia probably due to its ubiquity. and Senecio lobatus have traditionally been placed in — synonymy of 5. petasitis, and | concur with their Senecio prainianus was part of a collected by A. The apparently grown from seed by Berger at a garden in placement here. group of plants Purpus in the barrancos above Veracruz. specimens were England. He was unable to identify his plant but stated that it resembled R. petasitis. His protologue also reads as a very fine description of R. petasitis, and so the name is placed in synonymy. Senecio petasioides Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana was described by Greenman in Smith (1904) as being different from R. petasitis due to its involucral bracts being glandular-hirsute and the heads being shorter. In comparing the type material of S. petasioides with other collections of R. petasitis, I have found that the pubescence of S. petasioides is not distinct and that the capitula size falls within the range of the species. — Robinson and Brettell (1974) created the species A. chiapensis, stating that it differs from R. cristobalensis in the presence of rays. While this is correct, it does not separate it from R. petasitis. From examining the type material, I can find no distinction between R. chiapensis and R. petasitis. KEY TO THE VARIETIES OF ROLDANA PETASITIS la. Capitula eradiate .. 37b. R. petasitis var. cristobalensis lb. Capitula radiate, ray floret may be reduced to a tube. 2a. Leaf blade palmatifid, 5- to 7-lobed .... EET 37c. R. petasitis var. oaxacana 2b. Leaf blade orbicular 7 lobed. 3a. Abaxial leaf to ovate and 7(+)- surface glabrescent to ightly tomentose... see 0 ++ 37d. R. ns var. sartoril 3b. Abaxial leaf C surface tomento to ensely lanate tome ntose 1. R. petasitis var. petasitis 37a. Roldana petasitis var. petasitis Distinguishing characters. Leaf blade variously E orbicular to subovate, 7- to 9-lobed, sinus depth about ~ 1/8 of the way to the midrib or less, lobes acute to rounded; abaxially tomentose to densely lanate tomentose, adaxially pubescent with on glandular 10-16 ‘a. 8, densely stipitate-glandular pubescent, 7-9 X rairs. Capitula radiate, x 2-4 mm, phyllaries 12 mm. Ray florets 5 to 8, corolla ligulate, glabrous, tube ca. 6 mm, ligule 7-10 X 2—4 mm. Disk florets 10 to 15, 8-11 mm, tube 3-5 mm, throat ca. 5 mm, lobes ca. 2 mm. Pappus bristles 6-8 mm. Distribution and phenology. Mexico, Guatemala, El Salvador, Honduras, and Nicaragua (El Salvador: Ahuachapán, Chalatenango, San Salvador, Santa Ana, Guatemala: San Vicente, Sonsonate; Alta Verapaz, Baja Verapaz, Chimaltenango, El Progreso, Guate- mala, Huehuetenango, Jalapa, Jutiapa. Quezalte- nango. Sacatepéquez. San Marcos, Santa, Rosa, Sololá, Totonicapán, Zacapa: Honduras: Comayagua, Copán, El Paraíso, Estelí, Intibucá, La Paz, Lempira, Morazán, Ocotepeque; Mexico: Oaxaca, Veracruz; Nicaragua: Jinotega, Madriz) at elevations of 1000— 2500 m. Found in montane cloud forests and pine-oak with Pinus. Weinmannta., forests Quercus, Liquidambar, Ulmus, and Styrax. Flowering December through April. SALVADOR. Ahuachapán: W slopes rra ie e ca, vic. of Apaneca, 400—1600 m, 24 Jan. 1947, Standley 2991 (F). Chalatenango: near summil E Los Be smiles, 14 21"N, 890 W. ca. 2580 m. 10 Mar. 1942. J. M. Tucker 987b (F, MICH). San Salvador: Volcán de San Salvador, rd. Finca Florencia to the S Jan. 1946, Miramundo. 1890-2400 Representative specimens examined. EL from rim of the crater, 1680-1860 m, 30 Carlson 379 (F). Santa Ana: e iue» Los dou. NE of ee . 25 Feb. . Carlson 947 (Y). Cerro C. Standley 21553 (GH, NY). Sonsonate: San Juan e m, 27 Feb. 1907, 1997 (F). GU ATEMAL A. Alta Verapaz: Pastizales cenagosos a la margen del lío Frio, entre Tactic y Santa Cruz, 1460 m, 14 May 1963, A. Molina R. 12231 (NY). Baja Verapaz: dry | of Santa Rosa. 30 Mar. T P s ug ru ep Panag 'a de rra, SE Patzum, 2100 m, 31 Dec. 1938, P. C. ae vi (F). El Progreso: hills we tw. Finca Piamonte & slopes SE of Finca Bru 2400—2500 m, 4 Feb. 1942, Steyermark 13429 (F). Guatemala: on hills rd. 1650-1950 m. 16 Jan. 1939, P. C. 62997 (F). Huehuetenango: rd. to Standley 81281 (V). ms mtns. along the rd. betw. Jalapa 00-1700 m. 14 Nov. 1940, P. C. Standley 77292 (F). Jutiapa: vales ‘an Suchitan. NW of 2 íon Mita. 500-2050 m. 18 Nov. 1939, J. A. Steyermark 31895 (F). Que ne uh s of Volcán de Zunil, Aguas Amargas, 2 285( 17 Feb. 1939, P. C. Standley 65377 (F region of N A N slope of Volcán de Zunil, 2300-2500 m, 3 Feb. 1941, P. C. Standley 85734 (F). Sac NE oak- pine woods along upper reac ches of Río Sitio Nuevo betw. Santa Rosalia & s water fall. 1200-1500 m. 9 Jan. 1942, J. A. mu 42232 (F). S Pittier ud usté ue & San Haimundo, tandley mtns. e of Aguacatan, on the € 1940, P. Huehuetenango, ca. 50 m, 27 Dec. & ae E n Marcos: mtns. along National Route 1, ca. 2 mi. E of 7 e 2 July 1960, R. M. King 3158 (MICH). Santa Rosa: Santa Rosa, Cenaguilla, 4000 ft.. Feb. 1892, e et Lux 4522 (F, GH. MO, NY) Sololá: mixed forest area in Sierra ar Si Nahuala, 2800 m, 2 Dec. 1972, L. O. MICH). Totonicapán: along the rd. to Santa Cruz del Quiche, ca. 14 km generally NE of ER ca. 7800 fi., ~“ Jan . 1977, R. M. King 7280 (MO [2576149, 2576297], uere trail betw. Santa Rosalía de a « Vegas, Ys Dn J. A. Steyermark 42959 (MO). HONDURAS. omayagua: trail from Firea prae to Cerro ps near Cannes 18 Feb. 1955, M. C. Carlson 3165 (V). E km SO de Santa Rosa de j opan, 1200 m, 29 M. / Molina R. 11679 (F, NY). leading from Guinope to y ca. 1440 m 1947, P. C. Standley 2111 (F). Estelí: Faldas Tisey, 12°58'N, 86 22'W. 1400-1460 m. 6 Me 1982, P. P. Moreno 15768 (MO). Intibuca: vic. of La Esperanza 7 Intibuca, 1500- d 3] Jan.-12 Feb. 1950, P. Standley 25337 (F). La Paz: cloud forest area betw de larías & Montaña Verde Cordillera. Do umbo. 1900 m, 6 69. A. Molina R. 24053 (DS, F, NY). Le N e nto Naranjo, 11 km S arque Nacional de Celaque, 14°33'N, 88°40'W, 2470 m, 31 Jan. 1992, Mejia n. 3 (MO). Morazan: al NE de Valle Encantado, drainage of the Río Yeguare, AN 1700 m, 23 Mar. 19606. A. i Osona about Belén Gualcho, 1500-2000 m, 2-15 Apr. 1977, Alberto up 57 7 (MO). MEXICO. Oaxaca: Cerro del Machete, Pachutla, . 1941, 1200 m. Reko 62006 (F). Azumiate, pu a de San Miguel El Paraiso: Cumbre on "i hwy. Mar. 196 Siquatepeque. Pa T Veracruz: Tlagguio- 326 Annals of the Missouri Botanical Garden tiopa, entre Tlacuiotiopa y Vagueria, 19 Jan. 1989, 2800 m, Cházaro B. 0827 PS : NICARAGUA. Jinotega: i l èst area, Cordillera Central de Nicaragua E of Jones ga, Ji 400-1600 m. 20 Feb. 1963, ( Williams 24735 (F, NY). J : Cerro El Fra 86 16'W. 1 [REA m, 1984, P. P. Moreno 23521 (MO) Madri 10 Mar. 37b. Roldana petasitis var. cristobalensis (Greenm.) Funston, comb. nov. Basionym: Senecio Greenm. in Loes. Bull. Herb. Boissier ser. 2. 6: 867. 1906. Roldana cristoba- Greenm.) H. Rob. & Brettell, Phytologia 27: 1974. TYPE: in fruticeto in valle cristobalensis pm lensis 417. centrali, Mexico. Chiapas, in distr. x Río Prospero in Hacienda Tierra colorada sita, E. Seler 2106 (lectotype, designated here, GH!). Distinguishing characters. Leaf blade orbicular to subovate, 7- to 9-lobed or occasionally subpalmatifid, 5 to 7 lobes, sinus depth about 1/2 of the way to the midrib, lobes acute; abaxially leaf surface variously pubescent adaxially glabrate. Capitula to lightly tomentose, discoid, ca. 6 X 2 mm, turbinate, phyllaries ca. 8, densely stipitate-glandular pubescent, 3-6 X ca. 1 mm. Disk florets 5 to 10, ca. 6 mm, tube ca. 3 mm, throat ca. 2 mm, lobes ca. 2 mm. Pappus bristles ca. 6 mm. Distribution and phenology. Southern Mexico and Alta nango, Jalapa, Quezaltenango; Mexico: Chiapas) at Guatemala (Guatemala: Verapaz, Huehuete- elevations of 1000-1600 m. Found in montane cloud forests. Flowering November through February. Comments. Greenman in Loesener (1906) desig- nated the holotype Æ. Seler 2106 (B photos F!, GH!, MICH!, MO!, NY!) in the protologue. He also listed a Mexico. Chiapas, betw. San Cristóbal & 4 Dec. 1895, E. W. The Berlin specimens of paratype: Teopisca, 6700-8500 ft., 3469 (GH!, US!, B). collections were destroyed in the 1943 herbarium fire. Nelson — oth Roldana petasitis var. cristobalensis has many charac- ters that overlap between the typical variety and variety oaxacana. Its leaf shape is usually that of variety petasitis, but on occasion it has the subpalmatifid form of variety oaxacana. The abaxial leaf pubescence is usually lightly to densely pubescent, but it can have a light tomentum. The capitula are the smaller turbinate capitula, although eradiate, of variety oaxacana. Consideration. was given to maintaining. Roldana cristobalensis and/or R. oaxacana as species, but when taking into account the gross morphology and distribution of all the taxa involved and then the their recognition as intermingling of characters, varieties was deemed appropriate. Gl se A. Alta | f Coban, Villans 40725 (F. Representative specimens examined. Río Chio, ca. Verapaz: hills along — » 1969, L. O. 1300-1400 m, Jan.—Fe above Malacatancito, 30219 (F, MICH). ICH, NY). 1900 m, IO Jan. 1974. A I E lenango: Molina R. Jalapa: vic. of Sole ni Montaña Miramundo, betw. Jalapa € Mataquescuintla. 2000-2500 m, 4 Dec. 1939 4. Steyer- mark 32641 (F). Quezaltenango: Fuentes Ge 'oginas, western slope of Volcán de Zunil, ca. 2850 m, 4 Mar. 1939, P. C. Standley 67498 (F). MEXICO. Chiapas: SW of i xican hwy. 190 near Rancho Nuevo ca. 9 mi. SE of San ‘ristobal de | Cristobal de las Casas, mpio. S S, 9000 ft.. 7 Nov. 19 E. Breedlove 14151 (DS, MICH. NY): Clinica Rl Buena, 2 km NW of eblo Nuevo 5400 ft.. 23-24 Jan. 1965. P. H. Raver Solistahuacan, (E. MICH); shrubby slope in a of Yash’ ed mpio. Na japa. 6000 fi., 21 Feb. A. S. Ton 2092 (DS, F, CH, NY). 37c. Roldana petasitis var. oaxacana (Hemsl.) Funston, comb. nov. Basionym: Senecio oaxacanus Biol. Cent.-Amer., E 2: 244. 188l. Roldana oaxacana (Hemsl.) H. Rob. & Brettell, Phytologia 27: 422. 1974. TY P. Mexico. Oaxaca: 9900 ft., 1840, H. 2009 (holotype, K not seen; isotypes, F not seen, G not Hemsl.. Cordillera, Galeotti seen, GH!, NY not seen, US not seen, G photos F!, MICH!, MO!, GH photo MO!). Senecio hederoides Greenm. in Loes., Bull. Herb. Boissier. . 2. 6: 868. 1906. Syn. nov. Roldana hederoides Rob. & Brettell, Phytologia 27: 4: (Greenm.) H. 1974. TYPE: Mexico. RC m 3045 m, 20 Oct. 1894. E. designated here, GHI; e near Reyes, 2X Nelson 1002 ae S!). Gard. | Senecio da es Greenm., . Missouri. Bot. l 9]. mtns. of 1914. TYPE: Mexico a ll ‘a, 7800 ft, LO Dec. 1894, Rev. Lucius C. Smith 368 (lectotype, designated dnd GH!; isotype, 7H fragm. € photo 1). gods vacua da H. Rob. & Brettell, Phytologia 27: 416 1974. Syn. nov TYPE: Mexico. Chiapas: Mt. Peta: 31 e Coe Matuda 5-34 lose US not seen. US photos Fl. G HI MO!: isotype. MICH!). peu gushing characters. Leaf blade subpalmati- fid, 5 to the midrib, 7 lobes, sinus depth about 1/2 of the way to lobes acute; abaxially pubescent with stipitate-glandular and multicelled hairs, adaxially glabrate. Capitula radiate or disciform, ca. 6 X 2 mm, turbinate, phyllaries ca. 8, densely stipitate-glandular pubescent, 3-0 X ca. | mm. Ray florets ca. 3, corolla ligulate or reduced to a tube, glabrous, tube ca. 4 mm, ligule ca. 3 X 1 mm or when reduced tube ca. 2 mm. Disk florets 5 to 10, ca. 6 mm, tube ca. ca. 2 mm, lobes ca. 2 mm. Pappus bristles ca. 6 mm 3 mm, throat long. Distribution and phenology. Southern Mexico (Chiapas, Oaxaca, Veracruz) al elevations of 1400— = 2000 m. Found in pine-oak forests. Flowering October through January. In the protologue for Senecio oaxaca- (1881) cites the Galeotti 2009 (K) Comments. nus, Hemsley Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana specimen as the holotype. The GH specimen has a note from Greenman, in hand, that it is of the type collection. The G specimen has a Galeotti 2009 label and was determined by Greenman, in hand; however. it should be noted that the G photograph is seemingly of a different plant than the GH specimen, although undoubtedly a collection of Roldana oaxacana. Regarding the typification of Senecio hederoides, Greenman in Loesener (1906) designated the holo- type: xico. Oaxaca, distr. Nochiotlan, Cerro del Puebla Huautlilla, 15 Dec. 1895, Seler 1571 (B photos F!, GH!, MICH!, MO!, NY!); however, the only known specimen was destroyed in the 1943 Berlin herbarium fire. The paratype E. W. Nelson 1002 (GH) is chosen here as the lectotype. The label on the US specimen reads “E.W. Nelson 1802,” a typographic error. In the protologue for Senecio hypomalacus, Green- man (1914) cites Rev. Lucious C. Smith 368 (GH, photo & fragm. MO) as the type material. The lectotype is chosen here. Roldana petasitis var. oaxacana has a palmatifid form of leaf blade that distinguishes it from the typical variety. It also has smaller, turbinate capitula. Gibson (1969) recognized Senecio hederoides as a that it differs oaxacanus in that the leaf blade lacks secondary distinct species, saying from 5. lobing and prominent callous denticles. Greenman (1901) separates S. hypomalacus from S. oaxacanus by its having more distinctly lobed leaves that are thicker — in texture and with a soft tomentum beneath. In comparing the type materials, I do not find the above combinations of characters distinguishable from the variation present in the collections for this taxon. Representative Eden examined. MEXICO. Chiapas: mpio. Cintalapa, SE of Cerro Baul bordering Oaxaca, 16 km NW of Rizo ds x slong lo ogging rd. to Colonia Figaron, 1600 m, 8 Jan. 19 P. 31449 (MO). Oaxaca: on hwy. 175 N of du ca. 21 km N of jet. with hwy. 190, on first. high ridge in mixed forest, 19 Oct. 1956, T. M. PEU 3904 (KSC, WIS); ranch of Boone Hallberg, ca. 5 km W of Ixtlan de s 17 18'N, 9 ?^26'W. 2000 m, 20 Nov. 1982, Benz 689 (KSC, WIS); 14 mi. NE of Tlacolula, 3100 m, 28 Oct. 1965, A. Cronquist 10432 (CAS, F, GH, . KSC, MICH, NY); hwy. 175, 12 mi. E of jet. of hwy. Oaxaca, | mi. W of La Cumbre, 9 Jan. 1983, Spellenberg 6910 (NMC); S of El Limon el cual esta, 11.1 km al SW of jet. Tehuantepec and 1983, feme e 4289 (NY). Veracruz: Vic. Puente Acabaloya, ca. 5 km NW of Xico on trail to Xico Viejo, mpio. Xico (— Jico), d m, 5 Feb. 1984, M. Nee 29399 (KSC, NY). Buenos Aires, 9 Dec. 37d. Roldana petasitis var. sartorii (Sch. Bip. ex — Funston, comb. nov. Basionym: Senecio sartorii Sch. Bip. ex Hemsl., Biol. Cent.-Amer., Bot. 2: 247. 1881. Roldana sartorii (Sch. Bip. ex Hemsl.) H. Rob. & Brettell, Phytologia 27: 423. 1974. TYPE: Cordillera of Hemsl. Mexico. Veracruz: Vera Cruz, 3000 ft., Galeotti 2225 (lectotype, designated here, K not seen; isotype, GH!). Distinguishing characters. Leaf blade variously orbicular to subovate, 7- to 9-lobed, sinus depth about 1/8 of the way to the midrib or less, lobes acute to rounded; usually with a membranous texture to blade; abaxial surface glabrescent to lightly tomentose, adaxially glabrous. Capitula radiate, 8-12 X ca. 2 mm. Ray florets 5 to 8, ligule 6-8 X ca. 2 mm. Disk florets 10 to 15, 7-10 mm, tube ca. 4 mm, throat ca. 3 mm, lobes ca. 2 mm. Pappus bristles 6-8 mm. Distribution Mexico (Hidalgo, México, Oaxaca, Puebla, Tamaulipas, Veracruz) at of 1200-1800 m. forests. Flowering in January and February. Hemsley (1881) listed three syntypes from southern Mexico: Cordillera of Vera Cruz, 3000 ft., Galeotti 2225; Mirador, Linden 1164; and, probably Mirador. F. M. Liebmann 156— all deposited at Kew. Having only seen the Galeotti and phenology. elevations Found in pine-oak Comments. n the protologue, gathering, it was chosen for lectotypification. Roldana sartoril the northern extreme of the species R. petasitis. Much petasitis var. represents consideration was given to placing the taxon in synonymy; however, because its distribution is disjunct across the Isthmus of Tehuantepec and the difference in pubescence so extreme, it was given varietal rank. Representative specimens examined. MEXICO. Hidalgo: 5 km S of po weh 1800 m, 25 Mar. 1981, Hernandez M. 5611 Rop Taine ‘uatlan near rd. from Zacualtipán to Mahelonpa 2000 m, 7 Nov. 1946, Moore Jr. 1885 (GH). México: Toluca, 3300 m, 18 Feb. 19 Hunnewell 11851 (GH). "Da aca: along hwy. 175 betw axaca & Pochutla 77.8 mi. S of Miahuatlan, 20.3 mi. S of DIU 9.9 mi. N of turnoff to Pluma Hidalgo, 1480 m, 20 Jan. 9. T. B. Croat 46075 (KSC, MO). Puebla: 6 km above a to Matitla, 12 km f 1200 m, 30 Apr. 1983, García P. 1752 Villa Juarez, 1000 m, 23 Feb. 1961, KANU). E 4 mi. W ft. vic. of Gomez Farias, 4 =a 19°13'N 381 (MO). Veracruz: lk of Pails, 96°46'W, 600 m, 13 Mar. 1985, Castillo C. 4267 (NY): 2.5 mi. N of La irt rd., ca. 17 mi. NW of Jalapa, 7 Dec. 1974, Stuessy 3667 (MICH, OS). 38. Roldana platanifolia (Benth.) H. Rob. & Brettell, Phytologia 27: 423. 1974. Senecio platanifolius Benth., Pl. Hartw. 43. 1840. TYPE: Mexico. 1840, Hartweg 331 (holotype, K not seen; isotypes, G not seen, NY!, G photos F!, MICH!) Michoacán: in pinetis, Chico, Perennial herbs, acaulescent to subcaulescent, 0.4— ] m tall, arising from a slender woody rhizome (not bulbous); plants densely pubescent with a mixture of 328 Annals of the Missouri Botanical Garden stipitate-glandular and long multicelled hairs, stems cream with patches of purple. Leaves clustered. at base; petioles 6-12 em, densely pubescent: blades 6— Il X 8-13 cm, triple-nerved, palmatifid, 5- to 7- lobed, sinus depth about 1/2 of the way to the midrib, lobes acute secondarily lobed, callous denticulate; surfaces glabrate, sparsely to densely pubescent along veins. Capitulescence a terminal simple paniculate 20-40 cm, or a terminal compound paniculate cyme arising from an cyme borne atop a naked scape. obvious stem, 6 to 20 capitula per capitulescence. scape and peduncles stipitate-glandular pubescent, primary bracts linear, 4-10 mm, ultimate peduncles 38-20 mm, 2- 5 mm. Capitula radiate, 12-15 X bracteoles O to 3, linear to filiform, 4-6 mm, campan- ulate, corollas yellow; calyculate bracts 0 to 3 13, glabrate, ciliate hairs on mar- ca. 2 mm. hay florets ca. 8 glabrous. ligule 9-14. X 3-5 mm. Disk florets 30 to 40, corolla funnelform. glabrous. . linear or filiform, ca. 2 mm: phyllaries ca. densely stipitate-glandular to gins, greenish 8-13 s corolla ligulate, tube 5-8 mm. 7-11 mm, tube 4-5 mm. throat ca. 5 mm. lobes ca. 1 mm. Achenes glabrous, cylindrical, 2- 3 mm, ribs 10, resin glands absent; pappus bristles 9— 11 mm. Distribution and phenology. Central Mexico (Dis- Michoacán, Morelos, Nuevo León, Oaxaca, Puebla) at elevations of 2700— 4100 m. Found in pine-oak and fir forests. Flowering trito Federal, Hidalgo, México, October through February. Comments. the widely distributed of the smaller herbs in the genus. lts palmately lobed leaf blade is very similar to R. angulifolia. ls Roldana platanifolia is most somewhat coloration is a unique creamy green with purple hues throughout. The dense pubescence of mixed stipitate-glandular and long multicelled hairs also distinguishes this species. Representative specimens examined. MEXICO. Federal: Parc que Nac we Mese de los Leones, 3000 m. 28 No irrios 19 (KANU. MICH, UMO, WIS). Hidalgo: o Parque a ional El Chico on rd. from Real del i: to E 1C P :0, 3000 m, 27 Oct. 1949, Moore Jr. 5426 (MIC do N slope of Volcán Popocatépetl, 3300 m. dde 198 p" Cronquist H7 e H. KS CNY rd. 2—4 km fom dec km N« 134, 19 10' 3YN. M. Barkley Distrito ): on forestry f Temescalte ee to turn off 99 52 s isl Mar. V ic ies "án: rocky for hwy. 1996, T. E M. DeJong 1057 (MICH). Morelos: Parque Nacional de Zempoala 19 km E : Tres Marías, 2960 m, 10 Feb. d Garcia P. s.n. (MICH). Nuevo León: Zaragoza. a ¿erro del Viejo, 15 mi. W Dulces Lane 19 Aug. 1948, D 3011 (MO). Oaxaca: Uxpanapa Region, along grave 7 rd. from Esmeralda to Rio Verde, 1.1 mi. S of Esmeralda, 17 m a 94 ASW, 18 Jan. 1987, T. B. Croat 65784 (MO Popocatepetl, 4100 m, 9 Feb. 1965. £. Puebla: a s. p 996 (KSC 39. Roldana Brettell, reglensis (Greenm.) H. Rob. € 27: 423. 1974. Senecio reglensis n .. Publ. Field Columbian Mus.. Bot. Ser. 2: 283. 1907. TYPE: Mexico. Veracruz: Regla, Sept-Oct. (s.d). €. Ehrenberg 454 (holotype, B fragm. & tracing GH!, B photos Fl, Phytologia GH!, MICH!, MO!, B tracing MO!) Single-stemmed perennial herbs, 1—2.5 m tall; plants glabrate to weakly pubescent with hirsute hairs: stems terete. Leaves cauline, evenly distributed on stem; blades 5-7 X 5-7 cm, blades palmatifid. T-lobed, sinus depth about 1/2 of the way to the midrib, lobes acute, weakly secondarily lobed to denticulate; abaxially hirsute-pubescent, adaxially glabrous. Capitulescence a rounded compound panie- ulate cyme, 10 to 20 capitula per branch: calyculate bracts ca. 5, 1-2 mm. M Mn radiate, 9-12 X 3 mm, . Disk florets ca. 20. Achenes glabrous. (Description patas from linear, campanulate. Ray florets ca. the literature.) Distribution. Mexico (Veracruz) at elevations ca. 1200 m. Greenman (1907) lists the type locality as “Regla, Vera Cruz.” However, that city is in Hidalgo, not far from the Veracruz border. making the collection locality somewhat unclear. Comments. Greenman (1907) designated the type i The GH Berlin Comments C. Ehrenberg 454 (GH) in the protologue. specimen is a tracing of a that 1943. made by Greenman (1914: 279) in the protologue of Senecio kerberi make it known that the GH fragments came from the Berlin type collections. fragment and specimen was destroyed in Also, it is apparent that the GH tracing done by Greenman is of the B specimen that is depicted in all of the photographs, as well as the MO tracing. Therefore, the GH fragment and tracing are determined here to be the holotype. Roldana reglensis is seemingly known only from a fragment, tracings, and photographs. It superficially resembles R. mexicana but has radiate capitula. It is provisionally accepted. 40. Roldana reticulata (DC.) Phytologia 27: 423. 197 E Senecio reticulatus DC., Prodr. 6: 431. 1837 [1838], non Senecio reticulatus C. Clarke, 1876. TYPE: Mexico. [state unknown]: Villalpando, Mendez 18 type, G-DC not seen, microfiche IDC 800. 1142 Il 8!, G-DC photos F!, MICH!, MOL; isotypes. G-DC not seen, G-DC microfiche IDC 800. 1142 HE 1—3(pp)!). . Rob. & Brettell, holo- p 43. Heal 1940. del TYPE: Monte. Senecio dictyophyllus Benth., Pl. Hartw. Me XICO Hidalgo: Guajolote near Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana Hartweg 327 e K not seen, K photos F!, CH; isotype, NY! = — Single-stemmed perennial herbs, 0.5-1 m tall, much branched above, arising from a woody caudex: plants essentially glabrous; stems terete, striated, cream to red. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 0.5— 1:9 cm, 4—8 cm, nerved, ovate, 5- to 12-lobed, sinus depth about 1/8 of the secondarily x= glabrous; blades 4-8 triple- way to the midrib, lobes acute, may be lobed, giving margins an irregularly serrate appearance to merely subentire and denticu- late, surfaces glabrous, margins may be sparsely pubescent with short glandular hairs. Capitulescence flat-topped or rounded compound corymbiform cyme, 3 to 20 capitula per branch, peduncles glabrous, some plants with tufts of arachnoid pubescence at base of 2.0 cm, Y 9; 10 mm, — W capitula, primary bracts linear-oblong, to ultimate peduncles 5-25 mm, bracteoles O t linear, 1-2 mm. Capitula radiate, 8-15 X 5- campanulate, corollas yellow to yellow-orange; ca- lyculate bracts 3 to 6, linear to obovate, 2-13 X I- 5 mm; phyllaries 9 to 13, glabrous, greenish, some purple-tipped, 7-12 X 1-2 mm. Ray florets 5 to 8, corolla ligulate, glabrous, tube 4-6 mm, ligule 8-15 X ca. 3 mm. Disk florets 19 to 37, corolla funnelform, throat. 3—5 mm, 2 3 mm, ribs 10, resin glands absent; pappus bristles 5— glabrous, 7-11 mm, tube 3—4 mm, lobes 1-2 mm. Achenes glabrous, cylindrical, 9 mm. Central Mexico (Dis- trito Federal, Guanajuato, Guerrero, Hidalgo, Jalisco, México, Michoacán, Morelos, Puebla, Querétaro) at 3600 m. in pine-oak forests. Flowering Bonon through. December. Distribution and phenology. elevations of Found Comments. The most distinguishing feature of Roldana reticulata is its rather small ovate leaf — blades, which are attached to the stem by a short petiole and have a short internodal distance. When pressed, these characters combine to appress the leaf blades adaxially to the stem essentially covering the stem from view. It is also one of the few species in the genus with a corymbiform cyme. Upon examining the protologue and type material for Senecio dictyophyllus, it is placed in synonymy. MEXICO. Distrito epalcate, about Puerto de Las Cruces, . 19 Nov. 1965, J. Rzedowski 21728 (DS, F, GH, MICH, UC, Santa Rosa, 2600 m, 1 Nov. 1970, R M Guerrero: Teotepec, distr 1937, G. B. Hinton 11136 (GH, JC). Hidalgo: parte alta del Cerro Xihuingo, mpio. Tepeapuleo, 3000 m, 7 Oct. 1973, J. Rzedowski 31478 MICH, UMC UMO). Jalisco: Volcán Tequila, rd. to microwave Representative. specimens examined. Federal: Cerro El T — station, ca. 20°47'N, 103750'W, 2800 m, 24 Oct. 1970, Webster 15914 (DAV, MICH, MO). México: just S of Pan- Am hwy. at Pass of Monte Río Frio, 5 km WNW of Río Frio : p Grande, 19 22'N, 98" 43' W. 3100 m. a UR 1960. Iltis 1090 n an m, 27 Oct. Tancítaro, o 1165 (F, MO). Morelos: EA de 932, Lyonnet 797 Es bove Huejotzingo. 4000 n KSC). Querétaro: oe Colon, ] da SW . 13 Nov. 1971, J ME = Cerapoda, Ixtaccíhuatl a o X. 139( of la cumbre us Zamorano, 3100 n Rzedowski 417 (CAS, MICH). 41. Roldana robinsoniana (Greenm.) H. Rob. & Brettell, Phytologia 27: 423. 1974. robinsonianus Greenm. in Sarg., Trees & Shrubs l Senecio 902. TYPE: Mexico. Oaxaca: betw. Nopala & Mixistepec, 5 Mar. 1895, E. W. Nelson 2439 (holotype, GH!; isotype, MO). Simple or sparingly branched herbs or subshrubs, 1-3 m tall; lanate tomentum; stems terete, light brown. Leaves evenly distributed on upper half of stem, lowermost deciduous; petioles 5-10 cm, blades 10-15 X 10-15 em, palmate- to triple-nerved, ovate to cordate, 5(+)-lobed, sinus depth less than 1/8 of the way to the midrib or merely denticulate; abaxially densely pubescent to lanate tomentose and new growth covered with a dense white cauline, tomentose; canescent, adaxially glabrous to sparsely pubescent with short glandular hairs. Capitulescence compound paniculate cyme terminating in glomerules or short pedicellate cymules, 30 to 60 capitula per branch, peduncles lanate, primary bracts linear, ca. 4 mm, ultimate peduncles 0-2(-4) mm, bracteoles 0 to 3, filiform, ca. 1 mm. Capitula radiate, 6-8 X 2-3 mm, turbinate, corollas yellow; calyculate bracts ca. 5, ili 8, glabrous to 35 X ca. ] mm. Ray florets ca. 5, corolla ligulate, glabrous, tube 2-4 mm, ligule 2-4 X ca. 1 mm. Disk florets 7 to 13, corolla funnelform to moderately campanulate, glabrous, 4—7 mm, lobes 1-2 mm. Achenes glabrous, cylindrical, 0.5— filiform, 1-2 mm; phyllaries ca. arachnoid pubescent, purple-tinted, tube 2-3 mm, throat 1-2 mm, — 1 mm, ribs 5, resin glands absent: pappus bristles 4— 6 mm. Distribution and phenology. Southwestern Mexico (Jalisco, Michoacán, Oaxaca) at elevations of 1200— 1900 m. Pacific slopes, tropical deciduous and pine-oak forests with Podocarpus, Distylium, Ostrya, — ower Quercus, and Pinus. Flowering December through March. Comments. Roldana robinsoniana is a poorly collected species from southwestern Mexico. It most closely resembles R. lobata in that it has an elongate panicle that may terminate in glomerules and has a persistent tomentose pubescence. Annals of the Missouri Botanical Garden Representative specimens examined. MEXICO. Jalisco: 30-35 km SE of Autlán, seaward-facing slopes, kn be e summit, La Cumbre, near lumber rd. betw. El Chante à above the abandoned site Durazi CW, R. Me d 23226 (MICH). Mic Bos = & Cuzal: cy 19 m N, cán: Un 901 (GH). Oaxaca: 17 km NE of Pedia Larga, rd. Piedra Larga-Miahu: FAN Martínez S. 2733 (KSC) ca. 80 km SSW of Sola de Vega, seaward side of pass 25 s m S of the Río Verde . McVaugh 22373 above 5. Gabriel Mixtepec & ca. 30 k crossing al s a atengo, mpio. Juquila, R (DS, MICH, Roldana scandens Poveda € Kappelle, Bre- nesia 37: 157-160. 1992 [1993]. TYPE: Costa Rica. San Santos, Cordillera de Talamanca, por el camino 42. Prov. José: Reserva Forestal Los hacia San Gerardo de Dota cerca del caserió de Jaboncillos de Dota (Distrito de Copey), 2900 m, 26 ene. 1992 (fl.), Kappelle 5843 (holotype, CR nol seen: isotypes, ASD not seen, COL not seen. F not seen, MEXU not seen, MO!, NY not seen. U not seen, US not seen). Shrubs, to 5 m tall; new growth glabrate to arachnoid floccose; stems terete, dark red. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous: petiole 4—9 em, blades 4—9 5-14 cm, palmately nerved, orbicular, 9(+)-lobed, sinus depth glabrate; 1/8 of the way to the midrib or less, lobes acute, weakly secondarily lobed to denticulate; surfaces glabrous at maturity, plush arachnoid when immature. Capitules- cence pyramidal compound paniculate cyme, 30 to 70 peduncles weakly arachnoid 2—4 mm, ultimate capitula per branch, pubescent, primary bracts linear, peduncles 4-15 bracteoles 1 to 2, linear, ca. l mm. Capitula radiate, 8-10 X 2-3 mm, turbinate, mm, corollas yellow; calyculate bracts 3 to 5, linear, ca. | mm: phyllaries 5(to 6), glabrous, purple-tinted, 4-7 Ray florets 2 to es glabrous, tube ca. 5 mm, ligule ca. 1-2 mm. corolla ligulate, 5 X 1-2 mm. Disk iA funnelform, dou: 6—8 mm, 2—3 mm. 1-2 mm, 5 ribs, resin glands absent; pappus bristles ca. 7 florets ca. 5. tube ca. 3 mm, throat 1-2 mm, lobes Achenes glabrous, cylindrical, mm. Distribution and phenology. Costa Rica (Puntare- nas, San José) at elevations of 1800-3300 m. Found in open areas of oak forests. Collected in flower in September. Comments. | Roldana scandens is endemic to Costa Rica. It is best characterized by its palmately nerved, orbicular leaf blades. I affiliate it with R. schaffneri based on similarities in phyllary number (5 or 6), capitulescence, and size of overall pubescence, capitula. COSTA RICA. Pun- slopes of Cerro 9701'30"N., Representative specimens examined. tarenas: Cordiller Talamanca, Echandi, oak forest with Chusqa understory, ra de upper 82 49'00"W, 23 Aug. 1983, Davidse et al. 23996 (MO). San sé: high mtn. oak forest, on d & ridges, along the trail via Los Ange n above (N of) the 930" N, (F); a de | of Cerros re ci, trail along 34 lom Canaan to Chirripó Talari, 3100-3200 m, W. Cerro ind 7433 ridge 197 70. upper slopes, weste ridge with iio fes rstory, ca 'N, 83 40'W 3160 m, 15 Sep. 1983, G. Davids 24716 har slope in woods above cave, along Canaan—Valle de de Conegos (Chirripó ssi. 3150— $300 m, 10 Dec. 1966, A. 3. Weston 3688 (UC). 43. Roldana schaffneri (Sch. Bip. ex Klatt) H. Rob. & Brettell, Phytologia 27: 423. 1974. Senecio schaffneri Sch. Bip. ex Klatt, Leopold- ina 24: 126. 1888. TYPE: Mexico. Mirador, F. M. 151 C-Liebmann not seen). Veracruz: Liebmann (holotype, Senecio grandifolius Loes. var. glabrior Hemsl.. Biol. Cent.- Amer., Bot. 2: 240. 1881. TYPE: Mexico. Orizaba Valle Cordova, E. Bourgeau 2207 (lectotype, designated here, K not . Veracruz: seen; isolype c! Senecio idas Regel var. panier J| M. Coult.. Bot. Gaz 1891. Senecio santarosae Greenm., Publ. : 281. 1907, non Senecio pauciflorus Pursh, FI. limes Sept. 2 929, 181: TYPE: Guatemala. Dept. Guatemala: E 1310 m, Mar. 1890, Smith 2359 (lectotype, here, US-Donnell Smith not seen; isotypes, F!, GH not OQ). Field poe bian Mus., Bot. Ser. designated seen, | Senecio iL L. O. Williams, og la 31: 443. 1975 Syn. nc YPE: Nic di ^ Re Tj P MR: e m~ 500 m, Williams & T. P. Williams 24679 (olope, Fl; pa EAP not seen). Suffruticose shrubs to small trees, 2—4 m tall; new growth arachnoid to tomentose; stems terete, dark red. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous; petioles 2-7 cm, glabrate with arachnoid tufts at axils; blades 12-50 X 5-9 em, pinnate, elliptic to ovate, apex acute, base cuneate, margins subentire to serrate, or sinuately lobed: abaxially glabrate, weakly arachnoid pubescent on some veins, adaxially glabrous. Capitulescence pyra- midal compound paniculate cyme, 30 to 50 capitula per branch, peduncles floccose arachnoid, primary bracts linear, ca. 2 mm; ultimate peduncles glabrate. | mm. 2-5 mm; bracteoles absent or scale-like, to Capitula radiate, 7-10 X ca. 2 mm, funnelform, corollas yellow; calyculate bracts ca. 3, linear, l- 2mm; phyllaries 5 to 6, glabrate to floccose arachnoid, purple-tinted, 4-7 X 1-2 mm. Ray florets 3 to 5, corolla ligulate, glabrous, tube ca. 4 mm, ligule 3-5 X ca. 2 Disk d glabrous, 6-8 mm, tube 3—4 mm, throat mm. florets ca. 6, corolla 2-3 mm, lobes 1-2 mm. Achenes glabrous, cylindri- an ca. 2 mm, ribs 10, resin glands absent; pappus bristles 4-6 mm. Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana Mexico, and Nicaragua Distribution and phenology. | Southern Guatemala, El Salvador, Honduras, (El Salvador: La Libertad, San Salvador, Guatemala: Alta Verapaz, Baja Verapaz, Chimalte- nango, El Jalapa, Quezaltenango, Sacatepéquez, San Marcos, Santa Rosa, Sololá, Suchitepéquez, duras: Copán, Intibucá, La Paz, Ocotepeque: Mexico: Santa Ana; Progreso, Guatemala, Huehuetenango. Zacapa: Hon- Chiapas, Guerrero, Oaxaca, Veracruz: Nicaragua: Estelí, Matagalpa) at elevations of 1300-1800 m. Found in pine-oak forests. Flowering April through February. Comments. In the protologue for Senecio schaf]- neri, Klatt (1888) lists the paratype: Orizaba, Thomas 1864 (C-Klatt not seen). glabrior, In the protologue for grandifolius var. (1881) several syntypes housed at Kew: S Mexico, region of Orizaba, Botter, 608, 819, 1102; S Mexico, valley of Cordova, Bourgeau 2207, 2335; Costa Rica. Volcan de Irazu, Oersted s.n. Hemsley listed and In the protologue for Senecio ghiesbreghtii var. pauciflorus, Coulter (1891) lists two syntypes from aca Chucaneb, Dept. Alta Verapaz, 6000 ft.. Apr. 1880, J. D. Smith 1606; and the chosen lectotype J. D. EC 2359. While a herbarium abbreviation does not follow any of the specimens in the article. Coulter (1891: 95) states that he was asked to determine specimens for John Donnell Smith and that the “final and critical study of species” was done at Harvard University. While I have seen neither the US- Donnell Smith nor GH specimens, Greenman (1914) cites the gathering and these two specimens when he made the nomen novum $. santarosae, although he did not designate it as type material. Roldana schaffneri is a unique species in the genus. It has pinnately veined leaf blades, the corolla lobes can be as long as the throat, and the pappus consistently exceeds phyllary height, giving the involucre a foreshortened appearance. These charac- teristics lead one to suspect an affinity with the genus Digitacalia. Within schaffneri i superficially similar to R. hintonii and R. mezquita- lana. The latter two species SN ca. 13 phyllaries R. schaffneri. The numerous synonyms are a product of the species” un the genus, R. versus the 5 to 6 phyllaries of wide variation of the leaf margin. The margins vary from subentire to serrate to having distinct sinuate lobes. Otherwise, Roldana schaffneri. However, a correlation between leaf these. synonymized species all resemble margin type, distribution, and other morphological characters may be worth further investigation. cladis specimens examined. EL SALVADOR. La Libert ‘| Boqueron, Volcán San Salvador, 1887 m, 24 Feb. 1968, A. Molina R. 21663 (F). San Salvador: Volcán de $ San Pe from Finca Las Brumas, 1680-2010 m, ‘eb. 1946, M. C. Carlson 465 (Y). Santa Ana: Hacienda a Miguel near e tapan, chiefly on pine-forested slopes betw. Río San Miguel & summit of Cerro E l em 600-1380 m, 22 Feb. 1946. Carlson GUATEMALA. Alta rc along Quiche hwy. ca. 12 m W of San Cristóbal, . 1100 m, 24 Mar. 1941, P. C. Standley 89728 (F. MO). Ba aja ae min. slopes N of divide N of Santa Rosa, 1650 m. 30 Mar. 19 à Fe ou Standley en (F). Chimaltenan rd. be Chimaltenango & Martín Jilotepeque, i po l o. m, 22 De 1940, P. dia 80924 (F). El Progreso: hills N af Finca Seid betw, Finca Piamonte & summit of Volcán Santa Luis 00— 3333 m, 5 Feb. I: 4. Minis 43601 E. GH). Guate ae slopes of Volcán de Pacaya. betw. San Francisco Sales € the base of the active cone, 1800- 1940, P. C. Standley 80522 (F). mins. W of Aguacatan, on the rd. to 1950 m. 27 Dec. 1940, P. C. Standley Miramundo & summit of Montana 2300 m. 20 Dec. Huehuetenango: Huehuetenango, ca. 81207 (F). Ja Miramundo, betw. Jalapa and s scuintla, 6 mi. S of 500 m. 5 I 1939, J. 32699 (F). Quezaltenango: nm old rd. be i" Pirineos & Patzulin, 1200-1400 m. 9 Feb. 194 "ps y 80725 (F I Sachitepéquez: Volcan a S side. 8100 14 Jan. 2138 (A, F, NY). San HA sinus as 6 mi. SW of town of Tajumulco, NW slopes of Volcán ee 2300-28( 00 m, 26 Feb. 1940, J. A. Steyermark 366 d Santa Rosa: Santa ye or La edi 3000 m. Pe. 1893, Heyde et Lux. 4520 (F, GH). Sololá: Volcán Santa e S-facing slopes 2100-3000 m. 5 June 1942. J A. aoro 17039 (F). is E. p quez: V , S side, 8100 ft., t Jan. 19 . F. Skutch 2138 (A. , NY). s. trail betw. inla on us de Marmol & $ 19 Jan. 1942, J. A. 42915 ( F) . HOND Uh An: ie forest, Monte ). M. Hernandez . La xad ilapa: betw. 2000-25 Miramundo, . Steyermark Finca P. E 935, A. F. Skutch Lo a Ste) ermark o m, k fue Molina R. 62 : betw. Las Marías & M nie iia p Cordillera aG juajiquiro, “1900 m, 6 Mar. 1969, A. Molina R. 24049 (NY). Ocotepeque: along Yoroconte river betw. El Moral € Sinuapa. 1300 m. 10 Mar. 1969, A. Molina R. d (DS, F). MEXICO. Chiapas uud E e Tres Picos montane rainforest al the Cerro Bola n. a ae rd. Coloni donnes Mexicanos, mpio. Villa Corzo, | ie d m, 4 May 1972, D. E. Breedlove oe (DS, MICH). Guerrero: 5 km SE of El Carrizal de los Bravos, rd. Filo de a aballo-C hichil M ‘0, 2500 m. 21 Feb. 1983, Martinez S. 3293 (NY). Oaxaca: Juan Juquila, dist. Mixe, 1400 m, x Apr. 1083. His | (NY). Veracruz: Cerro de San Martín, Mapa 24.0, 5. J. L Calzada 217 (F, GH, mor arene winding rd. from Naolinco to Misantla, just above Paz de Enriquez, 19°50'N, 96 i» W, KSC, WIS). NICARAGUA. Estelí: . 14 Jan. 1981. P. Moreno Nicaragua, m, 19 Fe ios m, 2 Apr. ¿€ ordille ra Central de 1500 n le! k quarry,” dead forest area, 1963, Williams 24679 (F Rob. & C 'acalia 44. Roldana sessilifolia (Hook. & Arn.) H. Brettell, Phytologia 27: 423. 1974. sessilifolia Hook. & Arn., Bot. Beechey 436. 1841. Senecio beecheyanus Sch. Bip.. 28: 499. 1845, non Senecio sessilifolius Sch. Bip.. Flora 28: 50. 184: 5. nom. illeg. superfl. Senecio Annals of the Missouri Botanical Garden sessilifolius (Hook. & Arn.) Hemsl., Biol. Cent.- Amer., Bot. 2: 247. 1881, nom. illeg. Pericalia sessilifolia (Hook. & Arn.) Rydb., Bull. Torrey Bot. Club 51: 376. 1924. TYPE: Mexico. Nayarit: betw. San Blas & Tepic, ca. lat. N 22^, Dec. 1837, Sinclair s.n. (holotype, E-GL not seen: isotype. K not seen, K photos F!, MICH?). Flora 28: 498. 1845. ee gine Kunth, Nov. Gen. Sp. ( ed.) 4 * 2l . 1818, nom. illeg.. non Cacalia a ; sa 35 L ws p 782 Jn nec Pis necio dae » 181 [1 Diae 0 yllu ke C ent.- oe T, 2: 23 7 " jen oa S th. Bip.) a wl Torre y His C E TYPE: Mexico. México: 2: 377 T m, F me Sep. [1803]. Mole £ plan s.n. sere E not seen, microfiche ID iin ovatifolius Sch. Bip., Syn. nov. (folio dau os F!, MIC MOT, microfie E e Pn . Wes I 2t i type, K not see Cacalia nutans Sessé & Moc., | Nov. Hisp. 132. 1887. Syn. TYPE: Mexico. México: Topelpa, near Mexico City. Sessé & Mociño 2822 (lectotype. designated here. 35 MA not seen, microfiche IDC BT13. 235 I 1). Single-stemmed perennial herbs, 0.5-2 m tall, arising from a bulbous caudex with woolly buds: plants glabrous; stems terete, striated, greenish to red in color. Leaves cauline, clustered at midstem to base: petioles 0-18 em, glabrous; blades 8-18 X 9-20 cm. reticulate, leaves sessile or triple-nerved, may be petiolate, ovate, subcordate, 5- to 9-lobed. sinus about 1/8 of the way to the midrib, lobes acute, secondarily lobed to denticulate: surfaces glabrous, margins and veins rarely with multicelled hairs present. Capitules- cence a simple to compound paniculate cyme, 2 to 6 capitula per branch, peduncles glabrous, primary bracts linear, up to 20 mm or on larger plants obovate, 20 mm, bracte- 10-25 corollas yellow-eream or linear, 2-10 mm; 11-14 x 2- up to 40 mm, ultimate peduncles 5-2 oles O to 3, linear, 2-5 mm. Capitula eradiate. X 5-10 mm, campanulate, white; calyculate bracts 5 to 10, phyllaries ca. 13, glabrous, purple-tinted, 3mm. Disk florets 40 to 70, corolla funnelform to slightly campanulate, glabrous, 10-12 mm, tube 5- 7 mm, throat 2-3 mm, lobes 2-5 mm. Achenes ca. 2 mm, ribs 10. absent; pappus bristles 8—11 mm. glabrous, cylindrical, resin glands Distribution and phenology. Central Mexico (Aguascalientes, Chihuahua. Distrito Federal. Du- rango, Guanajuato, Hidalgo, Jalisco, México. Michoa- cán, Nayarit, Querétaro, San Luis Potosí) at elevations of 1000-2900 m. Acacia Mill., Clethra, Piper L.., and Dahlia. Flowering August through February. Found in pine-oak forests with Abies, Salvia, Fuchsia, Comments. Regarding Cacalia nutans, Sessé and Mociño did not designate a type in the protologue. According to TL-2, the original Sessé and Mociño herbarium is at MA. In 1963, Pippen, in hand, chose Sessé & Mociño 2822 from the material at MA as the ed ea There are three other collections inclu in the microfiche material, IDC BT 13. 235 11-4: & Mociño 2821, 2829, 4118. Type material of Senecio ovatifolius was destroyed in the 1943 Berlin lectotype. Sessé and herbarium fire. Roldana sessilifolia, as its name implies, may have leaf blades; however, this is not the typical The distinguishes the species: eradiate capitula, terminal sessile condition. following combination of characters few-headed capitulescence with bracts linear, 2—4 cm long, and glabrous achenes. MEXICO. Represe ntative specimens examined. Aguasc ta- ow 959 lientes: : f Aguascalientes, 2 : bluffs of ana Pringle 7983 (F. GH). 3 km N x t Soldado along the rd. to Otinapa, 21658 (MICH, NY). a W of Dolores Hidalgo, 2100 m, 9 Nov. 1970, 24095 (MICH). Hidalgo: elo Mule o on rd. from Metepec to Agua Blanca, 2400 m, 7 Sep. 948, Moore Jr. 4889 (A. MICH). e» isco: 28 km WNW of E rers E : Ln Jalisco, on rd. canoas, 2530 21 Oct. 1983, č Breedlove 59068 Du ; km SSW of Rincón S of El Chante, 19 36'N, 104 13'W, dem. H H. Hitis 28807 | KSC, WIS): pls 1-5 km N NNE of Cerro San Miguel & I km of ivatiside na center al (NY) ). Michoacán: re L be i JO m, 23 Oci . 1960. f Río Jesús María. ca. " km E of village Jesús María. W slope Sierra de Huichol. 1000 m. 20 Sep. 1960, Feddema 1345 (MICH). San Luis Potosí: 27 mi. E of San Luis Potosí on the Río Verde hwy.. 19 Sep. 1960. Johnston 5642 (MICH). 45. Roldana subpeltata (Sch. Bip.) H. Rob. € Brettell, Phytologia 27: 424. 1974. Senecio subpeltatus Sch. Bip. in Seem., Bot. Voy. Herald 211. 1856. TYPE: Mexico. [state unknown]: Sierra Madre, B. Seemann 1984 (holotype. GH!: isotypes, K not seen, P not seen, P photos F!. MICH!) Sparingly branched perennial herbs, 1.5-2.5 m tall. arising from a rhizomatous caudex with fleshy innovations (not lanate on specimens examined); plants glabrous; stems terete, finely maculate, finely striate, cream with patches of purple. Leaves cauline, evenly distributed on upper half of stem, lowermost 21 cm, blades 18— deciduous: petioles 15— glabrous: Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana 22 X 22.5-24 cm, eccentrically peltate, texture. wax paper-like, triple-nerved, ovate to orbicular, 5- to 7- to the — obed, sinus depth less than 1/4 of the way midrib to merely denticulate, lobes acuminate: surfaces glabrous, abaxially paler than above. Capi- tulescence simple to compound paniculate cyme. axillary peduncles delicate, 3 to 12 capitula per branch, peduncles glabrous, primary bracts absent ultimate peduncles 22—40 mm, scale-like, ca. 0.5 mm. Capitula eradiate, 15-20 X 3— bracteoles O to 5 mm, funnelform, corollas cream to white; calyculate bracts absent; phyllaries ca. 8, glabrous, purple-tinted at tips or throughout, ca. 15 X 1-2 mm. Disk florets ca. 12, corolla funnelform, pubescent with sparse small-clubbed multicellular hairs, ca. 14 mm, tube ca. 6 mm, throat ca. 7 mm, lobes ca. | mm. Achenes pubescent with sparse small hairs. (difficult to see without magnification), cylindrical, ca. 2 mm, ribs 10. resin glands absent; pappus bristles ca. 7-8 mm. Northwestern Mexico 1800-2300 m. Flowering Distribution and phenology. Sinaloa) at elevations of Durango, Found in moist ravines of pine forests. February through April. Comments. McVaugh (1984: 841) discussed the locality from which the type collection was made. “probably came from [western Durango near the boundary of Sinaloa], or from somewhere between Durango and Tepic.” Roldana subpeltata is one of the few species with eccentrically Other features are the delicate lax peduncles of peltate leaves. distinguishing the capitulescence, the distinctly funnelform and eradiate UTA and the wax paper-like texture of the leaf blades MEXICO. Du- wy. from Durango to Mazatlán, 50-54 mi. ic. of Rancho Los Angeles, near Sinaloa 1951, R. MeVaugh 11587 (MICH): Mazatlán—Durango hwy., 3-15 km toward El Saltó from the Sinaloa al at El Palmito, 2000 13 Apr. 1965, R. 2 (DS, MICH, MO); mpio. LE 1-46 mi. N of state line € 1.6 3-4 mi. N of El . 10550" AE 23 36'N. Se ln 4950 (KSC). Sinaloa 2000 n 1 12 Mar. Representative specimens examined. rango: a WSW of eee 2100 m, 25 Mar. "T P along = y. mi. ` Palmito. da erra "us 2200 m. 27 Mar. 1984, : just W of El Palmito, 1980, Lehto 24312 [one Rancho el Liebre, (ASU (Greenm.) H. Rob. & 424. 1974. Carola Greenm., Proc. Amer. Acad. 310. wa Pericalia iiis o Rydb., ' Bot. Club 51 . 1924. Senecio j (Greenm) McV "i on Univ. Mi- Herb. 9: 477. 1972. TYPE: Mexico. Morelos: wet barranca above Cuernavaca, 1 Nov. 46. Roldana suffulta Phytologia 27: Arts . 1896, C. G. Pringle 6626 (holotype, GH not seen, GH photo MICH!; isotypes, F!, MO!, NY!). Large single-stemmed perennial herbs, 1-3 m tall, rhizomatous caudex with woolly EI spreading from a buds; plants essentially glabrous, multicelled hairs may be dense on stem base: stems terete, striated, ereenish to red. Leaves cauline, evenly distributed on upper half of stem, lowermost deciduous: petioles 8— 20 em, glabrous: blades 10-25(35) X 10-25(35) cm, ovate, 7- to 13-lobed, sinus depth about 1/8 of the way to the midrib, lobes acute, weakly secondarily lobed to denticulate; surfaces glabrous, margins and proximal veins with multicellular hairs present. Capitulescence flat-topped to rounded compound corymbiform cyme, 2 to 6 capitula per branch, primary bracts obovate, up to 20 X LO mm, ultimate pe duncles 10-110 mm, bracteoles O to 3, obovate, 2— 1-10 mm. 10-27 X 5- campanulate, corollas bright orange (vellow- peduncles glabrous, Capitula eradiate, 10 mm, orange to yellow); calyculate bracts ca. 5, oval- elliptic, enclosing the capitula, (3-8)10-25 X (2-5)8— 15 mm; phyllaries ca. 13, glabrous, purple-tinted, 8— 17 X 2-3 mm. Disk florets (29-47)90+, corolla funnelform, glabrous, (7.5-)12-16 mm, tube (3—)8— 10 mm, throat 3—4 mm, lobes ca. 2 mm. Achenes pubescent, cylindrical, 2-3 mm, ribs 10, resin glands absent; pappus bristles 11-14 mm. Mexico Mi- 900— Tropical deciduous and pine-oak phenology. Central Hidalgo, at elevations of Distribution and Guerrero, Jalisco, México, choacan, Morelos, 2400(-2900) forests. Flowering October through December. (Durango, Nayarit Roldana suffulta’s large obovate ca- lyculate bracts that envelop the capitula are emblem- Comments. atic of the species. Some of the specimens examined have smaller, fewer-flowered capitula and reduced obovate calyculate bracts. This form may be a hybrid between R. suffulta and R. mexicana. examined. MEXICO. Du- 20, =. 9 mi. E AS) Ade specimens rango: 41 mi. WSW of Durango along hwy. of Las res 27 Sep. 1984, Guerrero: 15 km NE Puerto de TO hw iue Filo de Caballo. mpio. Tlacotepec, 3100 m, 23 Nov. 1983, 635 (KSC). Hidalgo: mpio. UR 5-6 k of Ismolintla, 1900 m. R. Herbández M. 5910 (MO). Jalisco: 11 rd. r p of Autlán & 2 rd. mi. SW of the pass, 1500 m, 1 . A. Cronquist 9774 (GH, MICH, MO, NY): trail to ı W La Cumbre, 5 as an N . 1l a Toluca to a alle de Bravo, e. 15 be . Morelia € Cd. Hidalgo, 19°40'N, 100 57'W, 2200 m, 2 Nov. 1970, Webster 16158 (MICH, UC). Morelos: Annals of the Missouri Botanical Garden min. canyon Pg Cuernavaca, 11 Nov. 1902, 2000 m, C. G. a ee ae M. MICH, MO, NMC, NY). Nayarit: along es from hwy. 15 N of i MICH): 8 8.0 mi. 1 O Nov. 1977, LaDuke 452 (MO hwy. Nov. 1968, > Boutin 2041 ( lo cen 1000 m. sundbergii (B. L. Turner) B. L. ) 278. 1990. Brittonia 37: 117. Roldana 47. = — Turner, Phytologia € Senecio sundbergu B. L. Turner, 1985. Roldana sundbergii Funston, Novon 11: 305. 2001, isonym. TYPE: Mexico. Nuevo León: mpio. Galeana, ca. 25 km S of Iturbide, along main rd. to Agua pigs 0.6 km W of Ejido la Purisima, ca. 2100 m, 27 Oct. 1982, S. Sundberg 1921 (holotype. TEX not seen, TEX digital image TEX! isotypes, KSC!, NY). Perennial herbs, acaulescent, 20-50 em tall. aris- ing from a fibrous rooted caudex; plants tomentose, tawny al base. floccose to elabrate above; stems terete, striated, reddish. Leaves clustered at the base; petioles —8 em, blades 6-20 X 20 em, triple-nerved, ovate to oval, 6- to 13-lobed, floccose to tawny lanate; crenate or weakly secondarily lobed, sinus depth less 1/4 of the way abaxially than to the midrib, margins callous denticulate; lanate tomentose and canes- cent, adaxially hirsute with multicelled hairs. Capi- tulescence terminal paniculate eyme, surmounting a 4— 30 cm naked scape, 3 to 6 capitula per branch (up to 50 E elabrate to lanate, primary bracts linear, up loo capitula in the overall capitulescence), scape ninm, ultimate peduncles 2-10 mm, bracteoles O to 2, filiform 1-3 mm. Capitula. radiate, 5-9 X 2-3 mm, O corollas yellow; calyculate bracts ca. 2, filiform, 2-3 mm: phyllaries ca. 8, glabrous, greenish, 356 X ca. | mm. Ray florets 3 to 5, corolla ligulate, glabrous, tube 44.5 mm, ligule 6-10 2-3 mm. Disk florets 10 to 25, corolla funnelform, glabrous, 5— 7 mm, lobes 2— 2.8 mm. Achenes glabrous or sparsely hispid distally, ribs 10, pappus bristles 5-6 mm. tube 2.5-3 mm, throat ca. 1.5 mm, cylindrical, ca. 2 mm, resin. glands absent; Distribution and phenology. Mexico. Restricted to the Sierra Madre Oriental of Nuevo León at elevations of 800-2100 m. Found in pine-oak forests with Pinus. Quercus, Platanus, Populus, and Colubrina. Collected in flower in October. Comments. — Roldana sundbergii is perhaps the smallest. perennial herb in the genus. It is very similar to R. gonzaleziae, but differs in pubescence, leaf shape, and scape size. For a complete review, see the Comments section under R. gonzaleziae. = Representative specimens examined. MEXICO, Nuevo Cola de Caballo, 100 10'W, 1982. J. A. Ville 1204 (KSC); areas Caballo. bosque mesofilo de montaña, León: areas 25 23'N, 27 Oct. cercanas a Cola de real V- 100 10 W, 24 23'N, 24 Oct. 1987, J A. Villarreal 5546 (MO) 48. Roldana tlacotepecana Funston, Novon ll: 305. 2001. TYPE: Mexico. mplo. Tlacotepec, 19.5 km al NE de Puerto del Gallo, camino Atoyac-Filo de Caballo, 2000 m, 23 Nov. 1983, E. 5648 (holotype, KSC!). Guerrero: Martinez 5. 20-60 cm tal plants glabrate to Perennial herbs, acaulescent, , aris- ing from a rhizomatous caudex; floccose arachnoid, long multicelled hairs present; stems terete, red. Leaves clustered at base; petioles 0-1 cm, glabrate to weakly arachnoid; blades 3.5-7 X 4-7 cm, triple-nerved, ovate to orbicular, ca. 5- lobed, sinus depth about 1/8 of the way to the midrib, lobes acute or rounded, margin callous denticulate; abaxially lanate tomentose and canescent, adaxially glabrous. Capitulescence simple paniculate cyme, borne atop a naked scape, ca. 10 em, | to 2 capitula per branch, pubescence patches of strigose hairs, dense beneath capitula, primary bracts linear, ca. 10 mm, ultimate peduncles when present ca. 4.5 em, bracteoles absent. Capitula radiate, ca. 16 X 7 mm, campanulate, corollas yellow; calyculate bracts ca. 3, linear, 2-4 mm; phyllaries ca. 14, glabrous, ca. 13 X 2 mm. Ray florets ca. 8, corolla ligulate, glabrous, tube ca. 8 mm, ligule ca. 12 X 4 mm. Disk florets ca. 35, corolla funnelform, glabrous, ca. 11 mm, tube ca. O mm, throat 4 mm, lobes ca. | mm. Achenes glabrous, columnar, ca. 3 mm, ribs 10, resin glands absent; pappus bristles 7-11 mm. Distribution and phenology. Mexico (Guerrero) at elevations of 2600-3500 m. Found in pine-oak and fir forests. Flowering November throu sh Januar £ 5 8 y small Roldana herbaceous perennial. It has the pubescence and leaf Comments. tlacotepecana is a characteristics of R. sundbergii and R. gonzaleziae but with larger capitula. It is named after the municipality from which the only known specimens were collected. ik nod specimens examined. MEXIC M Guer- rero: Cerro Teotepec, i Tlacotepec, ca. 4 of Coyuca de n nitez, 3200 m, 12 pod 1963, o 2918 (MICH); Cerro Teotepec, mpio. Tlac 7°29' | 100 12'W, 3400 m, 29 Jan. 1965, J. a 207 (MIC T ExcLUDED TAXA — Funston € Villa- comb. Basionym: Senecio pinetorum Hemsl., Biol. Bot. 2: 245. 1881 Roldana pinetorum (Hemsl.) H. Rob. & al: Psacaliopsis pinetorum (Hemsl. senor, nov. Cent.-Amer., Phytologia 27: 423. 1974. TYPE: Mexico. Oaxaca: Cordillera, 5000-7000 ft., Nov.—Apr. 1840, Galeotti 2019 (lectotype, designated here, Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana 335 K not seen; isotypes, G not seen, K fragm. & drawing GH!, G photos F!, GH!, MICH!, MO!). Distribution and ecology. Mexico, Honduras, and El Salvador (El Salvador: Santa Ana; Honduras: Lempira; Mexico: Guerrero, Oaxaca) at elevations of 2600-3500 m. Flowering November through April. Found in pine-oak and fir forests. Comments. In the protologue for Senecio pine- torum, Hemsley (1881) listed two syntypes, housed at Kew, both from southern Mexico, Galeotti 2019 and, without locality, Sallé s.n. Having only seen the Galeotti gathering, | have chosen it for lectotypifica- tion. The species Senecio pinetorum is transferred to the genus Psacaliopsis based on the following characters: blade centrally peltate leaf attachment, orbicular shape, and perennial subscapose habit. Literature Cited Barkley, T. M. 1978. Three new species (Asteraceae) from Mexico. Brittonia 30: 69-75 985a. Infrageneric groups in Senecio s.l. Cacalia s.l. (Asteraceae: Senecioneae) in m a a. Brittonia 37(2): 211-218. ue Generic boundaries in Senecioneae. Taxon of Senecio and Mexico and 34: a. A geo-historical perspective on the distribu- tion aid variation in Senecio s.l. (Asteraceae: Senecioneae) . America. Pl. Syst. Evol., Suppl. 4: 113- in Mexico and ( 119, , B. L. Clark & / A. M. Funston. 1996. The segregate genera of Sener sensu lato Cacalia sensu lato neae) in Mexico and Central Amer- —620 in D. J. N. Hind € H. J. Beentje Compositae: Proceedings of the International Compositae Conference, Kew. 1994. (D. J. N Hind, Editor-in-Chief). Vol. 1. Royal Botanic Gardens, (Asteraceae: Senecio ica. Pp. 613 (editors), Systematics. a 1839. ew. ed G: >, Hartw. iii—iv, 5-393. G. Pamplin, ondon | 73 163-533, 536-537 in G. er (editors), Genera Plantarum. B & Co. and Williams & Norgate, : du dia Pp. Bentham b e Hoo Vol. 2, part 1. ini 873b. Notes on the classification, history, and e distribution of the Compositae. J. Linn. Soc. Bot. 13: 335-577. Bertolini, A. 1840. Florula Guatimalensis. Novi. Comment. Acad. Sci. Inst. Bononiensis 4: 403—443. 1960. Compositae. Pp. 829-971 in T. H. Kearney € R. H. Peebles (editors), Arizona Flora, ed. 2. Univ. of California Press, Berkeley. Bremer, K. 1994. Asteraceae: Cladistics and Classification. Timber Press, Portland, Oregon. Bye, R. 1993. The role of humans in the diversification of plants in "| Pp. 707-731 in T. P. Ramamoorthy, R. Bye, A. Lot & J. Fa (editora s Diversity of Mexico, Origins and Distribution. Oxford Univ. Press, New York. Candolle, A. P. de. 1837 [1838]. Prodromus giai Naturalis Regni Vegetabilis, Vol. 6. Treuttelet & Würtz, Paris. Cassini, H. 1827. Cacalia in Dict. Sci. Nat. (Paris) 48: 459. eprinted in: Cassini, collected from the Dictionnaire des Sciences Naturelles from 1816 to 1830. R. M. dl & H. W. Dawson (editors). Oriole Editions, New Y :oney, P. J. 1982 Plate tectonic M oU of Middle America nn. Missouri Bot. Gard. 69: 432—443 Coil n VÀ M. 91. New or oe xis from Guatemala. Bot. Ca 16: 95-102. oe A us on Andean Compositae IV. 12: 182-195. 981 75. Cassini on Compositae, — p = > constraints on and the Caubbesn Cuatrecasas, J. Brittonia — ] . Studies in Neotropical Senecioneae Transfers to genus Pentacalia of north Andean species. Phytologia 49: 241-260. Drury, D. G. & L. Watson. 1965. Anatomy and the taxonomic significance E gross vegetative morphology in Senecio. New Phytol. 64: 307-314. Fernald, M. 1950. Gray's Manual of Botany, ed. 8. American Book d New Yor 999. Genus Roldana Kansas Funston, A. A Revision of the (Asteraceae: Posse Ph.D. Dissertation, State University, Manhattan, Kansas. H. . Neotropical floristic diversity: Phyto- connections between of the x dean Orogeny? Ann. Missouri Bot. Gard. 69: 551—59 Gibson, E. i 1969. A Revision of the Section Palmatinervii of the Genus Senecio (Compositae) and Its Allies. Ph.D. Dissertation, Kansas State University, Manhattan, Kansas. Gleason, A. € Cronquist. 1963. Manual of Vascular Plants. Van Nostrand, Princeton, New Jersey. Graham, ; 973. History of clemente in the northern Latin American biota. Pp. 301— 312 in A. Graham (editor), Vegetation and Vegetational History of Northern Latin Ámerica. nr Company, Amsterdam. 1975. La the arborescent te mperate Elsevier Scientific ste te Cenozoic evolution of o ng land vereion in Veracruz, Mexico. Evolution 29: 5. . 1976. Studies in i a paleobotany. IL The Miocene communities o z, Mexico. Ann. Missouri Bot. eee 63: 787-842. 993. eee factors and biological id el in Meo Pp. 109-128 in T. P. Ramamoorthy, R. Bye, A. t & J. Fa (editors), Biological ane of Mexico, Oris: a Distribution. Oxford Univ. New York. G A 83 Contributions to North here botany I. E Press, lompositae with revisions of certain da and critical notes. Proc. Amer. / : 1-96. ma M. I9 Monographie der c MuR hen Arten Allgemeines und Mone 1902. Bot. Jahrb. Syst. 32: 1-33. New or s Dn from Mexico, Central America and the West Indies. Publ. Field Columbian Mus., Bot. Ser. 2. 6: > 7-287 14. Descriptions of pons American Senecioneae. Ann. Missouri Bot. Gard. 1: -281 19 er Gattung Senecio. I. Teil. Leipzig. Reprinted in ic and Central part TL North Anena species of Ann. Missouri Bot. Gard. 2: 5 . 1926. Senecio. Pp. 1621-1636 in P. C. Standley a Trees and Shrubs of Mexico. Contr. U.S. Natl. Herb. Vol. 23. g3 wo genus Senecio, 79 Annals of the Missouri Botanical Garden - McNeill, F. R. Barrie, M. apium V. 5. Filgueiras, D. H. Nicolson, Pe Gy Greuter, W., Demoulin, TS E. Skog, Trehi ane, N. J. Turland & D. d (editors). 2000. nano! Code of eee Nomen- clature (Saint a Code). Regnum M g, 138. Hallé, F., R. A. Oldeman & I Tropical Trees m Forests: "pa Verlag, New York Hemsley, W. B. 1881. Biologia Centrali-Americana Bot 2: 1— 576. R. H. Porter, London. Hoffman, O. 1894. Compositae. Pp. 87-3 vA. Engler € K. Prantl (editors), Die Naturlichen o Vol. 4(5). N Holmgren & C arnett - 1990. Index 1978. An ee Analysis. Tomlinson. 979. Generic and s (C ompositae) II. Evaluation of some recent studies. Kew n 19-58. sectional limits in Senecio Bull. 37. Developing descriptors for systematic analy- ses Wt E necioneae (Compositae). Bot. Jahrb. Syst. 108(2/3): 201-211. . 1992. The tribe Senecioneae (Compositae) in the Mascarene Islands with an annotated world check-list of the genera in the tribe. Notes on Compositae VI. Kew Bull. 109. 47: 49— & Y. L. Chen. 1984. Taxonomic studies on the e Senecioneae (C d or Eastern Asia. Kew Bull. 39: 205-446. , P. Holliday, de and sec M. Wilmot-Dear & S. W. Jones. 197 T C paa I: —. tional ME in Senecio Progress pow pour Bull. 32(1): 47-67. $ M. & H. Robinson. 1970. The new synantherology. Taxon 19(1): 5 l. Klatt, F. W. . Beiträge zur Kenntniss der Compositen. Leopoldina ae 124— 128. Kowal, R. R. 1991. A new variety of Senecio (Asteraceae: Senecioneae) from the Sierra Madre de Manantlan, Jalisco, Mexico, with notes on the S. roldana complex. Brittonia 43(2): 102-115. a Llave, P. & J. M. Lexarza. 1825. Nov. Veg. Descr., 2: 10. de Fasc. Loesener, T. 1906. Plantae Selerianae. Bull. Herb. Boissier ser. 2. 6: ue x McDonald, J. 903, Phytogeography and history of the alpine-su lpn flora of northeastern Mexico. Pp. 681— 703 in T. Lot € J. Fa (editors), Biolo gical ; of Mexico, Origins and Distribution. Ordi Univ. Press, New York. MeVaugh, R. 1972. Compositarum Mexicanarum Pugills. Contr. Michigan Herb. 9: 359—484. . 1984. Flora Nov — amamoorthy, h. Bye, A. Univ. vo-Galiciana, Vol. 12: Compositae. Univ. of pur Press, Ann Arbor. 1950. Characteristics of the vegetation in certain 1e be regions of eastern Mexico. 3): 313— Miranda, F. & A. J. Sharp. = ; ae Ta studies in Senecioneae. Opera Bot. 44: 3-77. Ornduff, R., P. H. s D. W. Kyhos & A. R. Kruckeberg. 1963. a ru in Compositae III. Senecio- e J. Bo 131-139. 853. C Menem au p (Bestem- eiser af G. Bentham, Ved A. S. Orsted). Vidensk. Meddel. Dansk Naturhist. ie Kjobenhaven 1852(5-7): 65-121 Pippen, R. W. 1968 Senecio (Compositae). 447. neae. Örsted, meiser og Em Mexican cacalioid genera to bs U.S. Natl. Herb. — Contr. Robinson, B. L. € H. 1893. Additions to the phaenogamic flora of va 0, eee by C. G. Pringle. —115 . Seaton. Proc. Amer. Acad. Arts 28: pet un 1975, fangs in is Senecioneae (Asteraceae); VIL. sid. Phyt 23: 331-332 ——— & R. D. Brettell. 1973a. Studies in the Senecioneae (Asteraceae); |. A new genus, Pittocaulon. Phytologia 26: 451-453. I 1973b. Studies in the Senecioneae (Asteraceae); IL A new genus, Nelsonianthus. Phytologia 7: 535: & : in the Senecioneae (Asteraceae); II. 254-264. Phytologia 27: 1973c. Studies The genus Psacalium. 1974. Studies the Senecioneae (Asteraceae); V. T nera Psacaliopsis, aa "n and hom Phytologia 27: 402—439. hohlf, F. J. 1993. NTSYS-PC. Numerical o. and ere a P siens version 1.80. Applied Biostatistics, New York. Rydberg, P. A. - Some s senecionoid genera; I. Bull. Torrey Bot. Club zu 369— Rzedowski, 1978. oe tación de Mexico. Editorial Limusa, Mexico (D.F) Schauer, S. 1847. ERT el descriptiones generum novorum specier HDI plantarum in terris Mexicanis cresc entium. Linnaea 20: 698-699. Schultz, Ueber einige mit Senecio zu verbindende Gattungen. Flora 28: 497—4€ Sessé, M. & J. M. Mociño. 1887 [1890]. Pl. Nov. ud 1x11, 1-9: Imprenta de Ignacio Escalante, Mex & 1894. Fl. Mexic., ed. à. "i -213. Oficina ografica de La Secretari e ia de Pons. Mexico. E J. D. 1904. Undescribed plants from Guatemala and other Central American republies, XXVI. Bot. Gaz. 37(6): 417—423. Standley, P. C. 1926. Senecio. Trees p shrubs of Mexico. Contr. U.S. Natl. Herb. 23: 1621-16 dps J. A. 1950. Flora of le Ecology 31 308 3): es de 1969. Ecology and genetics in flowering plant idera Pp. 20-33 in C. Jeffrey (translation). Flowering ants: Origin o Dispersal. Oliver and Boyd, Edinburgh. m V. M. I tropic k Mexi Pleistocene changes of vegetation in Prance (editor), Biological Diversification in the Tropics. Columbia Univ. » York. ico. um 93-111 in G. T. ress, Ne B. L. 1988. Palmatinervii (Asteraceae) from eastern Mexico. Brittonia 40(1): 81-84. Turner, Two new species of Senecio section hety New Senecio northeastern Mexico. 6): 467—469. new species of R hiin Oaxaca, Mexico Phytologia 80( 2005 [20 —————, 1989. rom Phytologia 66 . 1996. E from pe mecioneae Et y ri SN i recession of be Mex 'an species of Roldana (Asteraceae: Senecioneae). Phytologia 87(3): 204-249 . King. Compositae. VIL. Mexican and Central American species. . Naturalist 9(1): 27-39, & D. Flyr. 1966. Compositae. X. North American species. 53(1): e & T. M. Barkley. combinations in Senecio sect. Palmatinervii (Asteraceae). Phytologia 67(5): 390-393. 1964. Chromosome numbers in the numbers in the Amer. J. Bot. Chromosome 1989. New species, names and w ño Volume 95, Number 2 2008 Funston Taxonomic Revision of Roldana & ,. Nesom. endange sed or threatened status of Me Pp. 559-576 in T. P. emai R. By (editors), Biological Div of Mexico; Distribution. Oxford Univ. Press, New Yor . M. Powell & M. King. 1962. Clwompeeine numbers in ilia Compositae. VI. Additional Mexican and Guatema- "s. dU es 64(759): 251-273. szek & F. M. Getliffe. 1992. Elucidative studies on 5 generic concept of Senecio (Asteraceae). Bot. J. Linn. Soc. 108: 55-81. W jos S. 1890. e to American botany. . Acad. Arts 25: 124—163. SR M. 983. Micromorphologica characters and generic World Senecioneae (Asteraceae). E 35: 1- Williams, L. O. 1975. Tropical American plants, XVIII. Phytologia 31 ^ 435—441. 1993. e ied and 1 Asterace n Origins and Proc. — eo delimitation of sor ew V -22 a. APPENDIX 1. Distribution of species in physiographic regions of Mexico and Central America. Species marked with an asterisk (*) are endemic to that region. Sierra Madre A (SMO, Roldana barba-johannis, R. Rr EAE R. ger a R. hartwegii, R. prada R. sessilifolia, R. subpeltata*, gonzaleziae, R. ia, R. mezquitalana*, suffulta Sierras y Llanuras del Norte (SLN, 1 sp.) R. iris. Sierra Madre Oriental (SMR, 18 s R. ones R. ap Hp R. chen, R. ~ johannis, R. grimesii, R. hartwegii, cleifolia, hirsuticaulis, R. lineolata, R. lobata, ru metepeca, neogibsonii*, R. petasitis, piano, R. reticulata, R. reglensis*, R. sessilifolia, n o Mesa del Centro (MC, 8 s R. angulifolia, R. iE R. cordovensis, heracleifolia, R. hirsuticaulis, R. lineolata, R. reticulata, R. € rans-Mexican Volcanic Belt (TVB, 27 s 5 albonervia, R. aliena*, R. angulifolia, R. deno ER ana, R. barba-johannis, R. cordovensis, R. ehrenbergiana* glinophylla*, R. gonzaleziae sit, R. guadalajarensts R. hartwegii, R. heracleifolia, R. hintonii*, R. kerberi lanicaulis, R. lineak wa, lobata, R. mexicana, R. R. michoacana*, R. petasitis, R. platanifolia, R. reticulata, R. robinsoniana, R. Shaffnen, R. Ses R. suffulta Sierra Madre del Sur (SMS, 27 s angulifolia, R. anisophylla*, R. E R. m -johannis, R. cordovensis, R. eriophylla, R. galiciana, R. hartwegii, R. do uad. R. heterogama, R. e oidea*, R. Jurgensenii, erberi, R. langlassei*, R. lanicaulis, R. lineolata, R. aa R. metepeca, R. mexica R. mixte- ana*, )etastLis platanifolia, R. re KM R. "d R. nier R. suffulta, R. ia da Janura Costera del Golfo del Sur (LCG, 1 sp.) I ae Sierras de Chiapas y Guatemala (SC & GU, 8 spp.) R. acutangula, R. barba-johannis, R. a R. hartwegii, R. heterogama, R. jurgensenii, R. petasitis, R. Cordillera Centroamericana in Mexico (CCM, 8 spp- R. acutangula, R. eriophylla, R jurgensenii, R. lanicaulis, R. petasitis, R. schaffneri . gilgii, R. heterogama, R. Central America (CA, R. acutangula, R. aschenborniana, R. barba-johannis, R. gilgii, g R. heterogama, R. jurgensenii, R. lanicaulis, R. petasitis, R. scandens*, R. schaffneri reenmanil, APPENDIX 2. Species distribution in Central America. Guatemala ro b pon R. aschenborniana, R. barba-johan- nis, R. g i, R. greenmanii, R. heterogama, R. jurgensenii, R. scada R. As asitis, R. schaffneri Sa ^ i iin R. petasitis, R. schaffneri Hondur R. jurgensenit, R. petasitis, R. schaffnerii Nicaragua R. petasitis, R. schaffneri Costa Rica R. heterogama, R. scandens Panama R. heterogama APPENDIX 3. Hypothesized species groups. Physiographic areas from north to south: SMO, Sierra Madre Occidental; SLN, Sierras y Llanuras del Norte; SMR, Sierra Madre Oriental; MC, Mesa del Centro; TVB, b Mexican Volcanic Belt; S ead Madre del Sur; LCG, Llanura Costera del E del Sur; SC & GU, Sierras ra Chiapas y d . Cordillera Centroamericana in Mexico; antral Modes a. Zentral American Complex Roldana acutangula (SC € GU, CCM, CA), R. gilgii (CCM, CA), R AE. (SC & GU, CA), R. heterogama (SMS, E. P GU, z CA), R. hintonii (TVB (SMS, SC & GU, LA), R. mezquitalana (SMO), s (CA). uir (TVB, SMS, LCG, SC & GU, C A) R. cordovensis Complex R. aliena s. MC, TVB, SMS). R. angulifolia (SMR, MC, TVB, SMS), R. anisophylla SMS). R. cordovensis (SMO, MC, TVB, ey R. een (TVB), R. gentryi (SMO), R. grimesii (SMR, T v. hederifolia (S R. ae MS), R. petaca (SMS), R. platanifolia ( MR, T VB, SMS) R. hartwegit Complex . hartwegii (SMO, SMR, TVB, SMS, SC € GU), R. heracleifolia (SMO, SMR, MC, TVB), R. kerberi (TVB, SMS), R. langlassei (SMS), R. lineolata (SMR, MC, TVB, SMS), R. tlacotepeca (SMS) R , R. jurgensenu — iit . lobata Complex R. "pus (SMR. ' TVB, CA), R. barba n (SMO, S GU, CA), a eriophylla pi CCM), : TVB), i. hirsuticaulis (SMR, MC), R. lanicaulis (TVB, R. lobata y TVB, S! . neogibsonit (SMR), A ina (TVB, SMS), R. andagi (SMR) ericalia Complex linophylla (TVB), R VB), R. aschenborniana (SMR, MC, . TVB, SMS, SC & R gonzaleziae (SMO, SMS, 35 0 "a EN a (SMO, T ui (SMS), R. mexicana (TVB, JR. mio TV reglensis (SMR), F ene an TVB, SMS), R. Cu e (SMO, SLN, E . TVB), R. dd (SMO), R. di uA (SMO, TV p. Rm s Complex R. us (SMR. TVB, SMS, SC & GU, CCM, CA) REVISIÓN DEL GÉNERO Paola Peralta,” María E. Múlgura de Romero,” JUNELLIA (VERBENACEAE)" 2 Silvia S. Denham,” y Silvia M. Botta’ RESUMEN Se enta una revisión del género ie Anas nom. cons. El género a 39 especies y seis variedades distribuidas en América del Sur, desde Perú y Bolivia hasta la Argentina y Chile. Se establece un reordenamiento de las categorías infragenéricas a nivel de secciones; se cambia de subgénero Junellia subgén. Thryothamnus (Phil.) Botta secc. Joni Bolta, dándose la siguiente nueva aa ion: Junellia subgén. Junellia sece. Junelliopsis (Botta) P. Peralta Múlgura; y se funda una nueva sección para las especies excluidas de d subgén. EE sece., nea Junellia subgén. Thryothamnus secc. Dentium P. Pe b & Múlgura. Las sigule ecclones y 's se sinonimizan: Junellia subgén. Thryothamnus secc. Ju Pu Botta, Verbena L. secc. Junellia Goldene Trone. se > Minutifoliae Tronc., Verbena secc. Junellia ser. Thymifoliae Tro: ferbena secc. Verbenaca Walp. ser. Erinaceae Walp.. Verbena secc. Verbena 'ü ser. Seriphioide ae Walp. v Verbena ser. Paus n pu Se neotipifican: ee Juniperina Lag., V. ligustrina Lag.. V. minima Meyen, V. thymifolia Lag., V. tridactylites Lag. y V. tridens Lag. Se lectotipifican: Doo. alla Miers, V. asparagoides Gillies & Hook. ex I ook., V. bisulcata Pu e V E PIS Kuntze, V. digitata Phil., V. lorentzii Niederl. ex Hieron oen Briq.. V. pygmaea R. E. V. selaginoides Kunth ex Walp., V. spathulata Gillies & Hook. ex Hook. var. inis dd Schauer, V. spathulata var. parii iflora Schauer y V. uniflora Phil. var. glabriuscula Kuntze. Se transfieren a la sinonimia los siguientes taxones: Junellia echegarayi "m on.) Moldenke var. cordifolia Moldenke, J. echegarayi var. pulerulon Moldenke [= J. echegarayi], J. spathulata (Gillies & Hook. ex Hook.) Moldenke var. grandiflora (Schauer) Botta . spathulata var. glauca (Gillies & Hook. ex Hook.) Bottal, J. tripartita Moldenke, J. longidentata Moldenke var. glandulosa pug |= J. aspera (Gillies & Hook. ex Hook.) Moldenke var. aspera], V. Nor Sucdwi th E J. caespitosa (Gillies 0 look.) Moldenke], V. dolichothyrsa Sandwith |= J. connatibracteata (Kuntze) Moldenke], V. colchaguensis Phil. E a a (Phil.) Moldenke] y V. scoparia Gillies & Hook. ex Hook. var. puberula Trone. |= J. scoparia (Gillies & Hook. ex Hook.) Botta]. Se realizan dos nuevas combinaciones a nivel subespecifico: J. toninii (Kuntze) Moldenke var. mulinoides (Speg.) P. Peralta & Múlgura, J. aspera var. ie Aa (Molden e) P. Peralta & Múlgura. Se incluye una clave para identificar las esp ecies y variedades, se actualiza la sinonimia de cada una de ellas, que se describen e ilustran y se rinda su distribución geográfica. ABSTRACT A systematic treatment of the ae Junellia Moldenke, nom. cons. is presented. The genus comprises 39 species and six rieties growing in South America from Peru and Bolivia to Argentina and Chile. Infrageneric names are re-ordered at the a level; Junellia Bub: UE (Phil.) Botta sect. Junelliopsis Botta is moved to another asa and the re vesulling new combination is made: Junellia subg. Junellia sect. Junelliopsis (Botta) P. Peralta & Málgura: and a new name hryothamnus sect. e is establis} ied: bins SHE. E sec S d | se ar Junelliopsis Botta, Verbena L. sect. Junellia (Moldenke) Tronc. ser. inuti liae Tr ronc. Tetera sect. Tunellia ser. Thymifoliae Trone., Verbena sect. Verbenaca Walp. ser. Erinaceae Walp., lla se Verben naca ser. Seriphioideae Walp., and Verbena ser. a Briq. The following taxa are neotypified: V. juniperina 2 Lag. . ligustrina Lag., V. minima Meyen, V. thymifolia Lag., V. tridactylites Lag., and V. a Lag. Lectotypes for the Ws species are given: Diostea filifolia Miers, V. e Gillies S Hook. ex Ho ok., . bisulcata Hayek, V. connatibracteata Kuntze, V. digitata Phil., . lorentzii Niederl. ex Hieron., f . E. Fr, V. Gillies & Hook. ex Hook. var. aa Schauer, V. Sau tar o var. parviflora Schauer, and V. uniflora Phil. var. — o x a © x z E 2 EN > = S Te Mel > Y = 3 = 2 c& 2 a nj 4 Pd Q ` ~ R -. E 3 5 O = a c 2 a“ = << > Q = = pud = = e E Ex S glabriuscula Kuntze. The follow laxa are synonymized: nra echegarayi (Hieron.) Moldenke var. cordifolia Moldenke. ma echegarayi var. pulerulenta. Mo denke [= J. echegarayi|. J. a (Gillies € Hook. ex Hook.) Moldenke var. grandiflora (Schauer) Botta [= J. spathulata var. glauca (Gillies & Hook. ex Hook.) Botta], J. tripartita Moldenke, J. longidentata Moldenke var. glandulosa Botta |= J. aspera iue s & Hook. ex Hook.) Moldenke var. aspera], V. comberi Sandwith [= J. caespitosa a : Hook. ex Hook.) Moldenke], V. dolichothyrsa Sandwith |= J. connatibracteata (Kuntze) Moldenkeļ, V. colohaguensis Phil. [= J. Ivana ii ) Moldenke], and V. scoparia Gillies & Hook. ex Hook. var. puberula Tronc. |= J. scoparia (Gillies ex Hook.) Botta]. 15 new combinations of subspecifie names ar established: J. toninti Cane a var. pa (Speg.) P. Peralta & Múlgura and J. aspera var. imm — "Se agradece al CONICET, por la financiación durante el desarrollo del presente trabajo, a través del PIP 2514/01, como asimismo a los curadores de E diferentes herbarios que nos uds información y facilitaron el préstamo de material. Asimismo agi radecemos al Dr. O. Morrone, por la lectura crítica del manuscrito, a la Dra. S. Martínez por su ayuda para esolver problemas nomenc da y al personal del IBODA, por la asistencia brindada. Los manuscritos dejados por S. Botta (1942- aves) sirvieron de base para la realización de este trabajo. ? Los editores e ecen a Diana Gunter su colaboración en P redacción de este manuscrito. a de Botánica Darwinion, La bardén 2 200. CC 22 (B1642H YD), San Isidro, Buenos Aires, Argentina. Autor para la corre legs ncia: pperalta@darwin.edu.ar : 10.341 7/2004167 ANN. Missouri Bor. Garb. 95: 338-390. PuBLIsHED oN 18 June 2008. Volume 95, Number 2 08 Peralta et al. Revisión del Género Junellia (Verbenaceae) and geographical distribution. Key words: Junellia, Verbenaceae, Verbeneae. (Moldenke) P. Peralta € Múlgura. A key to identify these species is given, as well as synonymy, descriptions, illustrations, Junellia Moldenke es un género americano repre- sentado por 39 especies distribuidas desde Perü y Bolivia hasta la Argentina y Chile En 1989, Botta resume la historia taxonómica del a — género y explica los caracteres usados para | delimitación del mismo dentro de la tribu Verbeneae. Propone además una clasificación | infragenérica basada en caracteres morfológicos y anatómicos de órganos vegetativos y reproductivos. Las especies del género Junellia no han sido revisadas en su totalidad y sólo han sido tratadas parcialmente en estudios florísticos regionales (Botta, 1984, 1987, 1993, 1999). Dicha autora estaba realizando una revisión integral del género 1994. En contribución se continúa con esta tarea y se revisan las 39 especies del género Junellia y sus variedades, sobre la base del estudio de los caracteres morfoló- cuando falleció en esta gicos vegetativos y reproductivos. Botta (1989: 382) hace referencia a una propuesta de conservación del nombre Junellia, pero ésta no había sido aán presentada. Finalmente, Botta et al. (1995) proponen formalmente la conservación del nombre Junellia en contra de Monopyrena Speg. y a Phil. La misma fue recomendada por : 867), aceptada en el XVII Congreso publicada en Brummit Internacional de Botánica de Missouri y el Código Internacional de Nomenclatura Botánica (Greuter et al., 2000). Al aceptarse el nombre Junellia como nombre conservado y Monopyrena y Thryotham- nus como nombres eee es, quedó establecida como especie tipo del género M. serpyllifolia Speg. |= J. serpyllifolia (Speg.) Moldenke| y por lo tanto se hace necesario reordenar los nombres de las divi- siones infragenéricas de acuerdo con el Código Internacional de Nomenclatura Botánica. MATERIALES Y MÉTODOS Los ejemplares examinados provienen del material depositado en los siguientes herbarios: BÀ, BAB, CONC, CTES, F, G, H, K, LIL, LP, MERL, MO, NY. P, SI, TEX (Holmgren et al., 1990). Se proveen las listas de las especies aceptadas del género Junellia y de los ejemplares examinados en los Apéndices 1l y 2. Lectotipificaciones. Las especies descritas por Lagasca, en base a las colecciones de L. Née, Malaspina, 1993; como — integrante de la Expedición ueron destruidos (Fernández Casas & Gamarra, com. Blanco), por eso para las especies de este - pers. autor se hizo necesario designar neotipos para las mismas. Acevedo de Vargas (1951). crítico de las Verbenáceas de Chile, lectotipificando, dados realizó un estudio m los casos necesarios. los nombres por Philippi. Para esto, la autora, utilizó los términos “Typus” e “Isotypus”, se considera que los ejemplares señalados por ella como “Typus”, deben ser tratados como lectotipos. CARACTERES MORFOLÓGICOS Hábito. Las formas biológicas que se observan en el género Junellia son nanofanerófitos y caméfitos. Dentro del primer grupo pueden encontrarse: arbustos subáfilos, con hojas usualmente caedizas y general- mente sin desarrollo de braquiblastos (J. spathulata (Gillies & Hook. ex Hook.) Moldenke). o arbustos con hojas de los macroblastos y braquiblastos de desa- rrollo normal o transformadas en espinas (J. aspa- (Gillies & Hook. ex Hook.) Moldenke, J. aspera (Gillies & Hook. ex Hook.) Moldenke, J. seriphioides (Gillies & Hook. ex Hook.) Moldenke). vegetales ragoides ser pulvinados: — os caméfitos pueden lefiosos o sublefiosos, bajos, con ramas muy apretadas formando cojines hemisféricos más o menos densos; son espinosos, densos como en J. erinacea (Gillies & Hook. ex Hook.) Moldenke y J. (Phil.) Moldenke, o inermes, generalmente con hojas carnosas, laxos como en J. thymifolia (Lag.) Moldenke. J. tridactylites (Lag.) Moldenke o en placas (con renuevos apenas sobre el nivel del suelo formando placas leñosas de apenas estos cojines en algunos casos ulicina bien unos centímetros de altura como en J. aretioides (R. E. Fr) Moldenke, J. micrantha (Phil.) Moldenke, J. minima (Meyen) Moldenke, y J. uniflora (Phil.) Moldenke). Indumento. Los tricomas observados son de dos tipos: glandulares y no glandulares. Los primeros son unicelulares o multicelulares formados por una cabeza globosa (2-8-celular), una célula del cuello y la célula del pie más o menos desarrollada dando origen a pelos glandulares largos y cortos. Los pelos no glandulares un cuerpo presentan un pie o base multicelular y apical unicelular, generalmente con cutícula gruesa. — El tipo de pilosidad: hispido, estrigoso. escabroso, etc., y la combinación de los distintos tipos de pelos es constante en las especies estudiadas. Por otra 340 Annals of the Missouri Botanical Garden parte, dentro de una misma especie, la densidad del indumento puede ser muy variable. Hojas. Las hojas de los macroblastos son opues- brevemente 2—3- herbácea. las, O a veces tetráslicas, sésiles o pecioladas. con lámina entera, 2-3-lobada, partida o 2-3-secta, de consistencia cartácea o crasa, a veces con ápice punzante o bien pueden presentarse reducidas (Junellia spathulata) o transformada en espinas (J. seriphioides). Las hojas de los braquiblastos en general son dísticas o imbricadas. sésiles, carnosas o no y a veces también transformadas en espinas (J. ulicina). La venación es pinnada, con vena primaria única de recorrido recto. En el caso de especies con lámina desarrollada se presentan cinco o seis venas de 2” orden, que pueden presentar vena marginal (J. longidentata Moldenke). En especies con hojas de los macroblastos tripartidas y espinescentes, la vena primaria origina dos ramas laterales. que ingresan a cada lóbulo, las secundarias son delgadas y puede presentarse vena marginal (J. tridens (Lag.) Moldenke). Inflorescencias. Las inflorescencias son dibotrios heterotéticos o monobotrios, formados por florescencias en espigas contraídas, cilíndricas o corimbiformes, plurifloras o paucifloras, en varias especies reducidas a 2-3 flores (Martínez et al.. 1996). Flor. Las flores son hermafroditas, levemente zigomorfas, con pedicelo muy breve. Cáliz tubular, generalmente 5-dentado, dientes pungentes (Junel- lia erinacea, J. mulinoides (Speg.) Moldenke) o no Moldenke, J. uniflora), en general subiguales, los abaxiales más desarrolla- (J. connatibracteata (Kuntze) dos que los adaxiales. La morfología de los dientes calicinales, por lo común, es constante para cada especie. La corola es hipocraterimorfa, conspicua. blanca, blanco-crema, amarillenta, rosa a rojiza o lila, generalmente con tubo bien desarrollado, exteriormente glabro, pubescente o en algunos Casos glanduloso (J. ulicina). interiormente con una banda longitudinal y abaxial de pelos agudos, adpresos y retrorsos; fauce glabra, a veces con escasos pelos, semejantes a los del interior del tubo corolino, se observaron pelos moniliformes, en J. tetragonocalyx Moldenke. didínamos, insertos en la mitad superior del tubo (Trone.) Los estambres son cuatro, corolino. con filamentos breves; el conectivo puede o no superar la longitud de las tecas. algunas veces se presentan apéndices conectivales desarrollados. El gineceo presenta el estilo filiforme, más de tres veces el largo del ovario, la base del estilo es ensanchada, cubriendo el ápice de las clusas (subgén. Junellia) o no desarrollada (subgén. Thryothamnus). Polen. radialmente simétricos, isopolares, elípticos en vista Los granos de polen son ecuatorial, y triangulares en vista polar, tricolpados sin ornamen- taciones notables (Raj, 1983). Fruto. Clusas subcilíndricas y de sección sub- Irígona, a veces con expansiones laterales y laminares del pericarpio que forman dos alas notables (Junellia sección Verticiflorae sensu Botta). de base angostada y sin desarrollo de un repliegue interno y basal de la cara comisural. El ápice puede ser rostrado como en J. bisulcata (Hayek) Moldenke. Número cromosómico. De acuerdo con Botta y Brandham (1993) el número cromosómico puede ser x =90x DISTRIBUCIÓN Y HABITAT El género Junellia se extiende en forma continua desde Perú y Bolivia hasta el sur de la Argentina y Chile, abarcando aproximadamente desde los 65 a los 75. de longitud oeste y desde los 16° a los 51^ de latitud sur, creciendo desde el nivel del mar. en a Patagonia argentina, y hasta los 4600 m en la 2). El encuentra en la Patagonia argentina. región andina (Figs. 1, mayor número de especies se Algunas especies son exclusivas de Chile (J. bryoides (Phil.) Moldenke, J. cinerascens (Schauer) Botta. J. lavandulaefolia (Phil.) Moldenke. J. (Walp.) Moldenke) o Perú (J. arequipensis (Botta) Botta). Fitogeográficamente, se observa una mayor diver- selaginoides sidad de taxones en el Dominio Andino-patagónico (Cabrera € Willink, 1973). i Junellia son elementos importantes en las comuni- 1985): ejemplo. J. seriphioides es frecuente en la provincia Ciertas especies de dades regionales (Boelcke et al., así, por Patagónica y en la provincia Puneña (Cabrera. 1957, 1978; Roig, 1970); Morello (1958: 117) la menciona. bajo Verbena L. y junto con J. aspera y J. juniperina (Lag.) Moldenke, como una de las especies impor- tantes en las estepas de arbustos bajos en los faldeos de la provincia del Monte. Otras especies son frecuentes en algunos distritos de la provincia Patagónica: J. tridens es muy abundante en la provincia de Santa Cruz, subdistrito santacrucence del Distrito 1956a. 1983). J. ligustrina (Lag.) Moldenke es frecuente en el Distrito Central (Soriano. [em del Golfo de San Jorge y en la parte septentrional del Distrito Central y J. connatibracteata y J. succulenti- Jolia (Kuntze) Moldenke aparecen en el Distrito Occidental (Cabrera, 1978). Algunos otras especies tienen un área más restringida, por ejemplo J. congesta. (Vronc.) Moldenke y J. cedroides (Sandwith) Moldenke, endémicas de la provincia Patagónica Volume 95, Number 2 Peralta et al. 341 Revisión del Género Junellia (Verbenaceae) T VAL T VAT 4 EN N mes ( ; ( f eE i P. A A EN ` 1 ` y : H ‘ 2 i 2——]|'iss po 20 bo. q0 y 1 y 1 1 te ssy ! Mo 2! l 2 a - y ` — MES -4b >n] h- AA pu 4 M3 | Y e MN TE N D N N K i + ‘5 r~ MM E SN jh o HHE pe Y SEC 4 otter rf » d PAS z 4 t h As pS [| e 1 J A. S rd | eee : y O; ig a un 2 [| at ei PCS i O TN y? e s ! / 3 t Í ` ¿qa a t y y LI [| L P mH a H O i 'H a i | r 40 r 40-f se) [ i ES I Mi ' i } D j H , , LI 4 4 4 LI , > , , > a DE T pt v S S v x 7 $ t O S 1 3 ES n ¢ O a} | L "Y. c "A c r r ES ^ B B B ` e M gi E SA 3 i ^ - ms - o o; UN t * hii J Bisulcala 60 gy. caespitosa A .arequipensis 3 asper 32 4 i @ J. cedroides ; . QJ. aretioides P» v noe ata i vine campesis Ay SLE mvar digtat Jeconnatibracieata : i .e i 4 J. asparagoides DJ. .azorelloides m J. bryoid 2J. congesta ar. integerrima QJ rayi T EIS 7 ts. T ws i ; E s h í ; ( i M y 1 ` \ » e, 9 ! t- A los í e (« © H © { ^N Or, ` z^ ` - Y 7 10 ? d r 20 » i 20 e. p e EAS 20 y ges 1 1 P 1 H i 1 E ¡0 a A ! e A eee 1 + wr Hz y wr on a O we A | as —> mI A ` -——— OA +- ^ poe ; te Ae -4 . ; dr nius N M 7 "|^ i e 5 Eh ND H M LA ilte , ie = ; ME 9 AN y 9e £ oe? E 7 ANS "i eua i S a ^ d S l * Prod P rd d m e N ^ N ow Po e EN y [ s H + ! ` O i x +, { iN ro ; ¿$ y P y for i Pg ie 40-f | 40-f | i 40 B - 1] 1 T jn A Jj } 5 > ‘ lj quta A S 7 ' Nb , o E nOs . i tu Em I = = o tT 2 E P 2 LA 1 o , I di 1. dp - E S E , E : F4 E Au UO A D "A E ‘AY F 4 4 . hera P, : asi P N e ~ £ E : SUL! a e o a? bad o ] D 60 RJ 60 TOF IJ. erinacea: QJ. idvandulaefolia Ou. miran 1 J. odonellii J. selaginoides OJ. hystrix J. ligustrina , : J. patagonica eJ.m J. pseudojuncea . patag O var. selaginoides P J. juniperina B var. ligustrina € J. scoparia avar. illapelina 1i var. lorentzii p Distribución de: Junellia arequipensis, J. aretioides, J asparagoides, J. aspera, J. azorelloides. —B. cinerascens, J. congesta. — digitata, J. connatibracteata, J. ecliopanenet J. minima, J. odonellit, J. Figura 1. aes J. bryoides, . J. erinacea, J. H —F. J. patagonica, J. scoparia, J. selaginoides. strix, J. juniperina, J. caespitosa, J. cëdroidi es, hy J. lavandulaefolia, J. ligustrina. —E. J. micrantha, Annals of the Missouri Botanical Garden wi eT 1 d 1 LI I 1 ( 4 pe ( je 9 Di , ^ , ¡00 Aa Yu z D aa a e s D inl 40- r 40- n , , 4 4 "i "S$: A e Ut 3 ¿ ^ A : £ , p pr t t Y Y A B e X e .. 60 Eu Wee y E hulata i OJ. spissa e J. seriphioides A J. silvestrii a "spat W J. succulentifolia OJ. tetragonocalyx oe» > Nr >> ` Neos.) Py / 40 3 ; : Ma L i500 i ' (j E if Id 5 2 ` 2 8 i 8 — "d — , £ A J, ! - F E g Z = T-- 60 & "^ 60 E toni E J. thymifolia l " O var. toninii O J. tridactylites I J. tridens 3J. ulicina I var. mulinoides @ J. triturcata A J. uniflora - Junellia seriphioides. —B. J. silvestrii, i spathulata. —C. J. spissa, J. succule ao J. ridens, J. trifurcata, J. ulicina, J. uniflora Figura 2. Distribución de: — tetragonocalyx, —D. J. toninii. Jo thymifolia, J. tridactylites. —F argentina, la primera con un área restringida al Tratamiento TAXONÓMICO departamento Zapala, provincia de Neuquén, mientras Dentro de la familia Verbenaceae, Junellia se ubica a tribu Verbeneae Schauer (Atkins, 2004: 459), que la segunda se encuentra en Zapala y en el el departamento Pilcaniveu, provincia de Río Negro Volume 95, Number 2 2008 Peralta et al. 343 Revisión del Género Junellia (Verbenaceae) caracterizada por: fruto seco esquizocárpico, separado a la madurez en cuatro clusas; anteras con tecas paralelas, o apenas divergentes, a menudo con conectivo dilatado; estilo 2-lobado, lóbulo anterior estigmatífero. CLAVE PARA RECONOCER LOS GÉNEROS DE LA TRIBU VERBENEAE la. Estambres fértiles 2; cáliz d x en el O eerte or REX eee igs robotana ug lb. Estambres fértiles 4, cáliz no acrescente en el fr rbustos con hojas bien Pee a o subáfilos, espinescente o no, o caméfitos pulinades o en placas. : láliz con largos pelos higroscópicos que cubren casi por completo a li corola; exocarpo liso . Urbania Phil. 3b. Cáliz sin pun pelos dicsisconicoss exocarpo generalmente reticulado . Erde Junellia Moldenke 2b. Hierbas anuales o perennes, con hojas bien desarrolladas, no espinesc entes. Sin orescencia en dibotrio o pleiobo- trio heterotético; espigas cilíndricas laxas densas en la floración .. Verbena L. mente contraídas en la floració O Glandularia I F. Gmel. Junellia Moldenke, Lilloa 5: 392. 1940, nom. cons. TIPO: Monopyrena serpyllifolia Speg. [= Junellia serpyllifolia (Speg.) Moldenke]. Monopyrena Speg., Rev. Agron. La Plata 3: 559. 1897, nom. rej. CLAVE DE SUBGÉNEROS, ESPECIES Y VARIEDADES DE JUNELLIA total o parcialmente el ápic 2a. He c caméfitos. Ww = - I a secc. Guedesi Arbustos bajos o caméfitos; conectivo no superando la longitud de las tecas; base del estil sanchad e las clusas: monobotrios pauci- o plurifloros (Junellia subgén. Junellia). ojas generalmente vea nunca transformadas en espinas, sin ápice punzante; arbustos inermes o Menden Phil., Anales l Univ. Chile 90: 618. mai 1895, TI hryothamnus junciformis Phil. Verbena sect. T Junellia (Moldenke) Tronc., Darwiniana 18: 312. 1974. TIPO: Verbena erinacea Gillies & Hook. ex Hook. [= Joelia erinacea (Gillies & Hook. ex Hook.) aa Arbustos bajos o caméfitos, inermes o no, tallos de sección poligonal o cilíndrica, con desarrollo o no de braquiblastos. Hojas homomorfas o dimorfas, opues- tas, a veces imbricadas, sésiles, lámina desarrollada o no, enteras, lobuladas, partidas o sectadas, pungentes o no. Monobotrios o dibotrios, bifloros hasta pluri- floros, raquis alargado o no en la fructificación. Flores levemente zigomorfas; cáliz tubuloso, 4- o 5-dentado, diente adaxial breve persistente, no contortos en la fructificación; corola hipocraterimorfa (4—)5-lobada, interior del tubo corolino con pelos retrorsos, fauce glabra o con pelos semejantes o no a los del interior del tubo; estambres 4, didínamos, insertos en la mitad superior del tubo corolino, generalmente incluidos tecas paralelas, conectivo mayor o menor que as tecas, con o sin apéndice glandular; ovario general mente glabro (Junellia tridens con ápice piloso), 2- carpelar, 4-locular, 4-ovular, óvulos basales; estilo filiforme, más de 3 veces el largo del ovario, con la ase ensanchada o no, estigma bilobado, lóbulo anterior estigmatífero. Clusas 4, subtrígonas, margen alado o no, ápice truncado o no, base angostada, sin repliegue interno de la cara comisural, blanca y generalmente papilosa y cara dorsal castaña, general- mente reticulada. y cubriendo Par superior de estambres con apéndices conectivales; corola glabra exteriormente; hojas generalmente trilobadas a tripartidas; caméfitos; monobotrios pauci- o plurifloros ( Junellia secc. Junellia) CLE 1. J. micrantha (Phil. Moldenke Par superior de estambres sin apéndices conectivales; corola con pubescencia variable exteriormente; 10Jas enteras o trilobadas a tripartidas; arbustos bajos o caméfitos; monobotrios pauci- o plurifloros Inflorescencias con 1 a 4 flores; caméfitos. Hojas enteras. Cáli Ms glabro; hojas generalmente tetrásticas, 6b. rsutas glabras o m ciliadas en de eee ek ea . silvestrii (Speg.) Moldenke Cali con pubescencia variada sobre la superficie; hojas no tetrásticas, pubescentes Ta. Cáliz pubérulo en la zona apical; clusas con las caras comisurales papilosas O E EN 6. J. minima (Meyen) Moldenke Tb. Cáliz pubescente en toda su superficie; clusas gen nearer con las comisurales lisas 2... 0... ee ee ee 13. J. uniflora (Phil.) ! Molde nke 5b. Hojas tripartidas o trilobadas 8a ojas tripartidas, iébules s con ápice agudo y punzante .. 12. J. trifurcata (Phil.) Moldenke 8b. Hojas trilobadas, algunas enteras, lóbulos con ápice redondeado no Tid uA ad a ido Steine t aes 2. J. aretioides (R. r.) Molde nke Ab. Inflorescencias con más de 4 flores; caméfitos o arbustos bajos. da. Caméfitos hasta 5 cm alt.; hojas enteras o n l Hojas seneralmente on enteras, de margen revoluto. Annals of the Missouri Botanical Garden lla. Hojas lineares, pubérulas; cáliz y tubo corolino hispídulo, éste apenas más larg jue-el Calle ache a rl 3. J. azorelloides (Speg.) Molde nke lb. Hojas ovadas, pilos; is; cáliz cano-velloso, tubo c iiu velloso, éste el doble de largo que el cáliz 22... ooo... . patagonica (Speg.) Moldenke 10b. Hojas tripartidas o trilobadas 12a. Hojas glabras, lóbulos enteros; brácteas enteras, oblongas: corola apenas mayor que el cáliz, exteriormente pubescente ....... 4. J. congesta (Tronc.) Moldenke 12b. Hojas con pubescencia variable. 13a. Hojas híspidas, lóbulos divididos: corola dos veces la longitud del cáliz. exteriormente glabra. 14a. Brácteas tripartidas ..... 5a. J. digitata var. digitata (Phil.) Moldenke 14b. Brácteas enteras, anchamente ovadas | Libis dog ERROR ee E Puce d china QAM UU EU E EE y J. € ata. var. integerrima. (Botta) Botta 13b. Hojas con pubescencia adpresa; lóbu l os Eo ros; corola menos de dos veces la longitud del cáliz, exteriormente con pelos aplic e relr A pad e LI. J. tridac n (Lag) Moldenke Ob. Arbustos de desarrollo variable; hojas enteras y algunas trilobadas 15e lojas con pelos cortos simples y glandulares; tallos con SUN encia escabrosa, con pelos glandulares ooo... oo. 9. J. succulentifolia (Kuntze) Moldenke 15h. Ii con pelos largos bl: eeu no annie eo con pelos adpreso- retrorsos, sin pelos glandulares .........o.o.o.o.o.. smile (Lag.) Moldenke 2b. Hojas homomorfas, todas e o dimorfas, P» aie los macroblastos us s 0 al menos con ápice punzante, las de los braquiblastos, tetrásticas, carnosas; caméfitos pulvinados o arbustos espinosos (Junellia secc. Junelliopsis l6a. Hojas ences espiniformes, enteras; caméfitos pulvinados; clusas brevemente rostrad a e Gut EA 18. J. ertnacea (Gillies & Hook. ex Hook) Molde nke 16b. M dimorfas; arbustos o caméfitos pulvinados: clusas generalmente truncadas u obtusas, a vece ostradas 17 la. Hojas de los macroblastos la mayoría enteras 18a. Inflorescencias bifloras (raro 3—4- Hors en los braquiblastos. og Arbustos de 30-80 cm alt.; clusas brevemente aladas o... O NON 21. J.. Me (Gillies & Hook. ex Hook.) Moldenke 19b. Caméfitos hasta 30 em alt: clusas no alada 20a. Brácteas híspidas en la cara BURN cáliz 3.5—4 mm: gineceo de 3.5- pa EU MERE EE DLP 22. J. spissa (Sandwith) Moldenke 20b. Brac ‘teas pubescentes en ambas caras: cáliz de 4.5-5 mm; gineceo de 6.5- Md 16. J. caespitosa (Gillies & Hook. ex Hook.) Moldenke 18b. ds 3—4- a plurifloras, en los macroblastos o en braquiblastos alargados: arbustos bajos. 2la. Brácteas connadas, ovadas, pubescentes; cáliz híspido sobre los nervios; hojas recurvas, a veces algunas hojas más o Menos rectas o... ooo E ed (Kuntze) Moldenke 21b. Brácteas no connadas, ovadas o angostamente ov: Pues agudas hasta acuminadas, pilosas o glabras; cáliz pubé m a pubescente o glabro; drum rectas. 22a. Brácteas pilosas; cáliz con pubescencia diversa: pubérulos hasta pubescen- MOS RP RETE EM 23a. J. toninii (Kuntze) Moldenke var. toninii 22b. Brácteas glabras; cáliz P sólo ciliado en los márgenes ........ LLL. = OPER 23b. J. toninii var. mulinoides (Speg.) P. Peralta & Múlgura 17b. Hojas de los macroblastos tripartidas o trisectas, punzantes, a veces también con hojas enteras, 3 el mismo e jemplar. 22 ~ usas rostradas; hojas de los braquiblastos tripartidas; brácteas enteras: tallos con pube scencia adpreso-retrorsa: one bajos. 24a. Cáliz con dientes triangulares, menores a l mm, pubescencia hispida: hojas de macroblastos generalmente patentes ... 14a. J. bisulcata (Hayek) Moldenke var. bisulcata 24b. Cáliz con dientes lineares atenuados, de i. D : mm, pubescencia hirsuta; hojas de los macroblastos no patentes .........., ». J. bisulcata var. campestris (Griseb.) Botta 23b. C TONS no rostradas; hojas de "braquibl astos enteras o bi- o tripartidas; brácteas enteras o tripartidas; 1 tallo con LE encia híspida, pelos patentes o glabrescentes; arbustos bajos 0 caméfitos pulvinac 25a. Brácteas TOME idas; hojas de los braquiblastos tripartidas. 26a. Monobotrios de 3 a 10 flores: brácteas de 4.5—7 mm: dientes d | cáliz de 0.5 mm ong., con ápice da TICKET 20. J. odonellii Moldenke 26b. Monobotrios de 2 o 3 flores; brácteas de 11.5-12 mm: dientes del cáliz de 1.5 mm long., con ápice acuminado .......... 4. J. ulicina (Phil. ) Moldenke 25b. Brácteas enteras; hojas de los braquiblastos enteras o paridas 27a. Arbustos con ramas erguidas, hasta de | m alt; hojas de los braquiblastos generalmente enteras o bilobadas o bi- o ee punzantes: tallos con Volume 95, Number 2 Peralta et al. 345 Revisión del Género Junellia (Verbenaceae) pubescencia patente... e 19. J. juniperina (Lag.) Moldenke Arbustos postrados o caméfitos; hojas de los braquiblastos todas enteras, no l J ] N > nis tallos con pubescencia patente o glabrescentes. 28a. Brácteas agudas o ac a iguales o menores que la longitud del tubo salicino; Bae is, brácteas y cáliz escasamente pilosas o glabras; tallos T. glabrescentes. 29a. Brácteas pilosas; cáliz con pubescencia dcin pubérulos hasta pubescentes: zucca Oo eee a : . toninit var. toninti 29b. Brácteas glabras; cáliz glabro, sólo ciliado en p márgenes . 23b. J. toninii var. made a mitad de la longitud del 28b. Brácteas ovadas y de ápice agudo, menor que tubo calicino; in brácteas y cáliz pilosos en toda su superficie: con pelos patentes 2... 15. J. bryoides (Phil 24 P MT Arbustos, no formando caméfitos; conectivo RCM te superando la longitud de las tecas (excepto iube da base del estilo no ensanchada y más o menos hundida entre js cuatro clusas (Junellia Bn Thryothamnus). 30a. Corola de fauce glabra; hojas enteras, lineares (muy rara vez bi- o tetralobadas) no espiniformes, papiráceas o crasas; espigas contraídas; estambres con conectivo menor que las tecas: clusas castañas (Junellia secc. Thryothamnus). 3la. Arbustos con hojas subopuestas, a veces alternas, y gene E aaa bien desarrollados. 32a. Cáliz con dientes subulados, en general mayores de 1 long., con pubescencia híspida y abundancia de pelos glandulares largos; hojas de 5-15 17 7 mm. long. ice Rees TL 27a. J. selaginoides (Kunth ex Walp.) Moldenke var. selaginoides 32b. Cáliz con dientes triangulares y agudos, menores de 1 mm long., con pubescencia hispida o aro etu es e URL E pe 27b. J. selaginoides var. illapelina (Phil.) Botta escasos pelos glandulares largos: hojas hasta de 5 31b. Arbustos con hojas opuestas, generalmente sin braquiblastos desarrollados. Brácteas lineares, ápice agudo, apenas menores que la longitud total del cáliz .......... 26. J. a (Gay) Moldenke A 33b. Brácteas no lineare 34a. Brácteas angostamente elípticas, ápice acuminado, no Dar TUTTO A O 5. J. lavandulaefolia (Phil.) Moldenke 34b. Brácteas ovadas, generalmente ápice rígido, punzante. 35a. Brácteas todas similares, menores o o iguales a la mitad de la longitud del cáliz, de 0.5— 0.6 mm lat, margen no membranáceo ... seh dogm dock the auus pecus “(Gillies & Hook. ex Hook.) Moldenke var. spathulata 35b. Brácteas diferentes, s leise mayores a la mitad de la longitud del cáliz: las 3—4 mm lat., margen membranáceo o no, las medias y superiores más angostas, de 2— 3.5 mm lat., margen no membranáceo .. 2... 0. RD PED 28b. , spathulata var. oo (Gillies & Hook. ex Hook.) Botta 30b. Corola de fauce pubescente o glabra; hojas enteras o ee inermes o punzantes; monobotrios en espigas ntraídas o alargadas; estambres con conectivo mayor que 36a, Hojas de los macroblastos i aa ae las de los A uo con lámina reducida o no; monobotrios en espigas contraídas hasta 3 em long.; corola con fauce pubescente (Junellia secc. Dentium). 37a. Hojas de los is enleras; cáliz 4- Pu 4-cos tado; fauce de la corola con pelos A REM EU Chess Pil Md x (Trone.) Moldenke 37b. Hojas de los macroblastos tripartidas; cáliz 5-dentado, 5- cta fauce de la corola con pelos claviform pane AN ak RU a ie See 33. J. tridens (Lag.) Moldenke 38b. Gineceo TE en el ápi 3 Cáliz ne con dientes triangulares de ápice agudo . 2... lisse 30. J. cedroides (Sandwith) Moldenke 38a. Gineceo piloso en e 39b. Cáliz de dientes subulados, punzante 40a. Hojas de los braquiblastos enteras con lámina desarrollada, ovado- a lineal- oblonga, a veces con un diminuto mucrón ... sse E J. asparagoides (Gillies & Hook. ex P Moldenke 40b. Hojas de los braquiblastos tripartidas, con lámina reduc ida, carnosa ...... ————— giant Tete aia) cae ate ts Se EE ee eee ee aE . hystrix NT iil.) Moldenke 36b. Hojas de los macroblastos no espiniformes, las de los He con ; [amine desarrollada; monobotrio o dibotrios en espigas mayores de 3 em long.; corola con fauce glabra (excepto en scoparia y J. cinerascens) (Junellia secc. Vertic iflora). 4la. Plantas ge »neralmente sin desarrollo de braquiblastos; fauce de la corola pubescente: hojas de 2— n de largo, membranáceas o subcrasas. 42a. Tallos de sección tetragonal; apéndices conectivales presentes; base del estilo profundamente hundida entre las d plantas grisáceas que no se ennegrecen SeCHISO lel Mb on eu E drogues xm ac O. J. cinerascens (Schauer) Botta 346 Annals of the Missouri Botanical Garden 42b. Tallos de sección poligonal: apéndices conectivales ausentes; base del estilo apenas iundido entre las clusas; plantas que se ennegrecen al secarse ooo E tater ig eld yk, oat Gk da at . J. scoparia (Gillies & Hook. ex Hook.) Botta 41b. Plantas con diferenciación en braquiblastos y macroblastos; fauce de la corola glabra; hojas de los macroblastos de 5-25 mm de long., cartáceas o coriác 43a. Clusas con el pericarpo apenas expandido ba mente formando un ala insconspicua; dientes del cáliz triangulares, menos de 1 mm de long.; tubo corolino 2—3 veces la longitud del cáliz EA EE E A EEA ERS 4. J. arequipensis aua Botta 43b. Clusas con el pericarpo expandido calas nte, formando un ala conspicua. I I pe p 44a. Hojas coriáceas y enteras; dientes del cáliz triangulares hasta de 1 mm long., Tac ia del cáliz adpresa o hispida: corola más de 2 veces la longitud del cáliz 45a. Cáliz con pubescencia adpresa; inflorescencia de (3—4.5-9 em long. ..... V LP 38a. J. ligustrina var. ligustrina (Lag.) Moldenke 45b. Cáliz con pubescencia I: ná de 1.5-3(-60) em long. ...... A a b. J. ligustrina var. lorentzii (Niederl. ex Hieron.) Moldenke 44b. Hojas membranáceas, enteras o trisectas en los macroblastos. 46a. Dientes del cáliz E de 1-2.5 mm long. 47a. Pubescencia del cáliz híspido-glandulosa o escabroso-glandulosa; tubo de corola de 6.5-10.5 mm, apenas más largo que el cáliz; ramas jóvenes de sección circ ular hojas del macroblasto generalmente trisectas . 2... TEES: a. J. aspera (Gillies & Hook. ex Hook.) Moldenke var. pam 47b. Pubescencia del cáliz incano-ve ‘Hosa, r no glandulosa, más abundante sobre las costillas: tubo de la corola de 8—9.5 mm, generalmente 2 veces el p del cáliz; ramas n enes de sección tetrágona; hojas del aid generalmente enteras... UNTER. 35b. D aspera var. longidentata (Moldenke) Múlgura & P. Peralta 46b. Dientes del cáliz triangulares, subiguales, de ápice rígido, los abaxiales de 1 mm ong.; pubescencia del cáliz homogene a a aon corola menos de 2 eces la longitud del cáliz VA . J. echegarayi (Hieron.) Moldenke Junellia Moldenke subgén. Junellia. TIPO: Mono- erinacea Gillies & Hook. ex Hook. como especie tipo pyrena serpyllifolia Speg. |= Junellia serpyllifo- del subgénero. Al realizar la propuesta de conserva- lia (Speg.) Moldenke]. ción del nombre (Botta et al., 1995) la especie tipo del género pasó a ser Junellia serpyllifolia. La misma Verbena secc. Verbenaca Walp., p.p., Repert. Bot. Syst. 4: 14. había sido considerada por Botta (1989: 390) dentro 1845. ey 5 i Ae de la sección Guedesia Botta, ser. Minutifoliae Arbustos bajos, muy ramificados, con láminas de — (Tronc.) Botta. Con posterioridad Botta y Brandham las hojas bien desarrolladas, a veces espiniformes o (1993: 150) no la asignan a ninguna de las divisiones caméfitos; conectivo no superando la longitud de las — infragenéricas propuestas hasta ese momento, por lecas; base del estilo ensanchada y cubriendo el ápice tener un número cromosómico diferente a las restantes de las clusas. especies de la sección Guedesia. Asimismo los f VN . autores, sugerian la necesidad de la reclasificación Observaciones. La sección Verbenaca Walp. com- : M es y . de la especie en una nueva serie o sección, criterio prendía muchas de las especies que actualmente se . ; : p l que se ha adoptado en este tratamiento. consideran pertenecientes a los géneros Glandularia y F ». | , | 2. ; : . x ista sección queda conformada por esta única Verbena. Schauer (1847: 536) funda la sección : l pr j , : : oe especie por sus características morfológicas y cromo- Verbenaca, que resulta ser un homónimo posterior. Edd DAA sómicas. Las restantes especies de la serie Minutifo- Junellia Moldenke subgén. Junellia secc. Junellia. liae Tronc., permanecen en la sección Guedesia. TIPO: Monopyrena serpyllifolia Speg. |= Junellia Ey Pte SONG pee ad Ane l. Junellia micrantha (Phil. Moldenke, Lilloa 5: 399. 1940. Basónimo: Verbena micrantha Phil., 'IP serpyllifolia (Speg.) Moldenke]. Verbena secc. Junellia ser. Minutifoliae Trone., p.p., Anales Univ. Chile 90: 915. 1895. TIPO: Chile o arwiniana 18: 314. 1974, syn. nov. Junellia sece. II Argentina. "Andes provincianus centralis", Guedesia Botta, ser. i ai (Tronc.) Botta, p.p., 72, P. Ortega s.n. (holotipo, BM!; isotipo, ( - E $ « 3 Darwiniana 29: 389. 1989. BM foto SI!). Figura 3A-F. Gamentas; Higgs Fibimortiortes, eterna 7 ns menmes. Verbena ramulosa Phil., Anales Univ. Chile 90: 610. 1895. Monobotrios plurifloros. Apéndices conectivales de- TIPO: Chile. “Prope locum Angostura de sarrollados. Número cromosómico x — 9, x = 10. provincia O'Higgins lecta est”, | nov. 1886, J. Monreal Marin s.n. (holotipo, SGO!, SGO foto SI!; isotipo, Per Observaciones. Botta (1989: 386), siguiendo el Monopyrena serpyllifolia Speg., Revista Fac. Agron. l criterio de Troncoso (1974: 312) eligió Verbena Nac. La Plata 3: 559. 1897. Verbena pim Volume 95, Number 2 Peralta et al. 347 Revisión del Género Junellia (Verbenaceae) Figura 3. A-E. did ird —A. Aspecto general, con hojas enteras. —B p —C. C i —D. Antera. A-D de Bot 6. —E. Pg to general, con hojas pu ws (de i ibrera 3397). F, G. Junellia ae —F Aspecto general. —G. E E via. F, G de Cabrera 31136. H, I. Junellia SU —H. Aspecto general. I. Inflorescencia. H, I de Volponi 24. 348 Annals of the Missouri Botanical Garden (Speg.) Speg., Anales Soc. Ci. Argent. 53: 245. 1902. Junellia serpyllifolia (Speg.) Moldenke, Lilloa 5: 401. 1940. TIPO: Argentina. s Jorge. feb. 1896, C. Imeghino LP 10405 (holotipo, LP: isotipo, SID. Verbena wile zekii Briq., Annuaire Conserv. Jard. Bot. Genève 18 beam iud (Briq. ) ues nke, > Disk ib. Verb. Avice . TIPO: ] ae Atuel, 7 (holotipo, LAU no visto; isolipos, K!, G foto FM 24701 SI!) ada s Speg., Anales Soc. Ci. Argent. 53: 245. 1902. June s an (Speg.) Moldenke. Lilloa 5: 394. 1940 Argentina. Chubut: Corcovado, aest. 1900, N. jm sn, i [doti pd. LP isotipo, SE). I0 em alt: bescencia patente, entrenudos de 0.3-3 mm. Hojas e 14 trilobadas, con lóbulos espatulados o enteras. Camefitos hasta ejes hispídulos, pu- homomorfas. « 5-2 mm, no imbricadas, elípticas anchas hasta orbiculadas, ápice redondeado u obtuso, base cuneada, densamente estrigosas. en ambas superficies o menos densa la superficie adaxial, pelos adpresos. Monobotrios contraídos, plurifloros, de 20 a 30 [lores sésiles, blancas, celestes, liláceas o rojizas. en espigas cortas, hasta 2 em: brácteas de 1.5-2.5 X 1.5-2 mm, ovadas o elípticas, cóncavas, cara adaxial elabra, cara abaxial estrigosa. Cáliz de 2.5-3.5 mm, cilíndrico, dientes cortos, triangulares, generalmente todos iguales, superficie externa estrigosa: corola de 5-6.5 mm, glabra en la superficie externa, fauce glabra: par superior de estambres generalmente con apéndices conectivales exertos, conectivo no supe- rando la longitud de las tecas; gineceo de 44.5 mm, —2.5 X ca. | mm. glabras, ápice obtuso cubierto por la base del base del estilo ensanchada. Clusas de 2 estilo, cara dorsal reticulada, castaña, caras comisura- les papilosas, blancas. Cariología = 9;pn Botta & dla am, 1993: ad To ^ 145 IB. C). Distribución Habita el sur de Mendoza hasta Santa Cruz. geografica. en Argentina, desde crece desde la región de la costa hasta la zona cordillerana. hasta 3500 m, en la 15). provincia fitogeografica Patagó- nica (Fig. Fenología. Florece y fructifica entre octubre y febrero. Botta, Junellia serpyllifolia. Iconografía. 1999: 177, fie. 128. sub Observaciones. Esta especie se caracteriza por la presencia de apéndices glandulares conectivales y por la corola apenas sobrepasando el cáliz. En algunos ejemplares de la región oeste de Neuquén, no se ha observado la presencia de glándulas conectivales: sin embargo los restantes caracteres de la morfología foliar y floral coine Laen "n con los de la especie (Sehajovskoy s.n. ; Roig 13056 2562: nicum 715; Unui 10. > Her patago- ARGENTINA, Chubut: Me uerial adriana, — 5 de Pto. Madryn, ruta 1, Correa 10171 (SI). Mendo a Dole Malargiie, O lo del Ancho, 30 km W de Los Molles, Fabris 8514 (SD. Neuquén: Dpto. Catan Lil, ruta 40, entre La Negra y Catan Lil. Gentili 817 ). Rio Negro: Dpto. Norquinco, El Portezuelo, a 25 km de des E 26600 (SI). Santa Cruz: Dpto. . ruta 5 10 km al N de los Manantiales, Boelcke 16229 (SI). Junellia Moldenke subgén. Junellia secc. Guedesia Botta, Darwiniana 29: 388. 1989. TIPO: Verbena uniflora Phil. |= Junellia uniflora (Phil.) Moldenke Verbena ser. Pauciflorae Briq., ud. Bot. Geneve Joli ser. . 197 syn. nov. Annales Conserv. A(2) 17. 1899 Pauciflorae (Briq.) Trone., Junellia Moldenke Junellia “asa ser. Pauciflorae (Briq.) Botta, Darwini- ana 29, 389. 1989, TIPO: Verbena uniflora Phil. |= Junellia dnd Phil.) Moldenke Verbena secc. Junellia ser. 18: 314. 1974. | erbena secc. Darwiniana subgén. SEC, Darwiniana Junellia Minutifoliae Tronc., Junellia Moldenke ser. Minutifoliae (Trone.) Botta, p.p., 389. 1989, TIPO: Verbena minutifolia tellia minutifoliae (P hil. ) Moldenke Junellia ser electa: Cali & 314, 1974. TIPO: Verbena thymifolia y. [= Junellia thymifolia (Lag.) Moldenke]. subgén. Guedes Darwiniana 29: Phil. [= Jur Verbena sece. Tronc: Darwiniana Arbustos bajos o cojines. homomorfas, Hojas enteras o divididas, imbricadas o no. inermes o con ápice punzante pero no transformadas en espinas, a veces subcarnosas. Monobotrios plurifloros o pauci- floros. Apéndices conectivales no desarrollados. Número cromosómico x = 9. Observación. Aquí aceptamos la sección Guede- sia, sin las series propuestas por Botta (1989), porque los caracteres enunciados por la autora no resultan suficientes d avalar la subdivisión. Por otra parte Botta (1989: 389) consideró a J. odonellii. en esta sección: en e presente contribución. se ubica a la misma en la sección Junelliopsis. 2. Junellia o (R. Fr.) Moldenke, Lilloa : 393. 1940. Basónimo: Verbena aretioides R. E. ES in Acta Regie Soc. Sci. Upsal, ser. 4, 1: 109. 1905, non Verbena aretioides Hayek, 1908. TIPO: Bolivia. “Salitre pr. Yavi (In Argentina) in 1000 m, 6 ene. 1902, R. E. 1053 (holotipo, S no visto, S SI!; isolipos, GH no visto, NY! SE). montis siccis apricis? Fries foto Verbena o R. E. Fi 108. 1905 . Nova a Regie Soc. Sci. Upsal, ) 5 Mono pygmaea (R. Fr.) Distrib. Verb. Avicenn.: 77. Jujuy, E ve Known Con 1912. TIPO: Argentina. AE 4100 m. . F. Kurtz 1515 (estopa, aquí designado, S!; isotipos. NY!, SI). in petrosis, 27 ene. 1901 Caméfito en placas de 1.5-2 em de alto, con raíces gruesas; muy ramificado, ramas de 3-5 em. Hojas « e 23 X ca. homomorfas. « | mm. densamente Volume 95, Number 2 2008 Peralta et al. Revisión del Género Junellia (Verbenaceae) imbricadas, tetrasticas, subcarnosas, sésiles, en general 3-lobadas, lóbulos de 1 mm. oblongos. ápice obtuso, a veces enteras, incano-pubescentes en ambas caras, cilioladas en el margen. Monobotrios pauci- floros, - 2 a 3 flores; brácteas anchamente ovadas, de 2-2.5 en la cara bod con pelos adpresos en la adaxial, 5-2 mm, ápice obtuso o subagudo, vellosas ciliadas en el margen. Cáliz de 2.5-3 mm, velloso en la mitad superior, dientes de 0.5—1 mm, triangulares: corola blanca, de 4-5 mm, tubo algo ensanchado en el ápice, externamente glabro; par superior de estambres sin apéndices conectivales, conectivo no superando la 2.5-3.5 mm, Clusas ca. 2 mm, longitud de las tecas; gineceo de base del estilo apenas ensanchada. elabras, subreniformes, ápice truncado y cubierto por la base del estilo, base angostada, cara dorsal de color pardo obscura, caras comisurales blancas, papilosas. Nombre vulgar. “Yaretilla” (Botta, 1993: 76). Distribución geográfica. Bolivia, Chile y Argen- tina en las provincias de Jujuy y Salta. Crece en suelos pedregosos de la provincia Altoandina, desde los 3000 y 4500 m (Fig. 1A). Fenología. Se ha coleccionado en flor desde octubre hasta abril. 1 (sub Verbena Iconografía. Botta, 1984: 348, fig. pygmaea). Botta, 1999: 75, fig. 34. Observaciones. Junellia aretioides se diferencia de J. trifurcata (Phil.) Moldenke, porque esta última presenta los lóbulos foliares vellosos en la superficie y ápice agudo. En la descripción original de Verbena pygmaea se 347) cita depositado citan dos ejemplares de Kurtz. Botta (1984: Kurtz 11513, como isotipo de la especie, en SL En realidad la autora realizó una lectotipifica- ción, sin mencionarlo, por lo que en esta oportunidad se hace efectiva la misma. examinado, ARGENTINA. Jujuy: ipan, Botta 356 He) Salta: Dpto. Los Andes, San Antonio los Cobres, subida a Alto Chorrillos, Cabrera 8270 (SI). BOLIVIA. Potosí Sud López, cerro Liberman ee (SD. Tarija: Punta Patanca, Fiebrig 2616 (BM, K, LIL, SD). CHILE. I Región: Tarapacá, Tranque de Caritaya, pue 25595 (SI). adicional Material Dpto. Tumbaya, Abra « Tapaquillcah, 3. Junellia azorelloides (Speg.) Moldenke, Known Geogr. Distrib. Verb. Avicenn.: 77. 1942. Basónimo: Verbena azorelloides Speg., Revista Fac. Agron. Univ. Nac. La Plata 3: 562. 1897. TIPO: Argentina. Santa Cruz, Dpto. Deseado: “In altiplanitie secus Golfo San Jorge”, feb. 1896, C. Ameghino s.n. (holotipo, LP 1881!, LP foto SI! isotipo, SI). Caméfito pulvinado, de 2-3 cm de alto, inerme, formando cojines subhemisféricos o chatos, muy ramil ficados, compactos: ramas de —3 cm, con entrenudos muy abreviados. Hojas homomorfas, de 2-3 X ca. 0.5 mm, lineares, enteras, sésiles, crasas, no espinosas, ápice obtuso y soldadas en la base. pubérulas en ambas superficies, con pelos más largos en la parte basal del margen. Monobotrios en espigas contraídas, plurifloras de 4 a 5 flores; brácteas de 2.8— 3.5 X 1.5-2.5 mm, ápice agudo, con el nervio medio ésta villosa, cara prominente en la cara abaxial v abaxial subglabra. Cáliz tubuloso, de 6.5-7 mm, hispídulo, ápice 5-dentado, dientes breves: corola de 1.5-12 mm, hispídula: par superior de estambres sin apéndices blanco-rosada, exteriormente rojiza o conectivales, conectivo no superando la longitud de as tecas; gineceo 8 mm, base del estilo ensanchada. Clusas 4 mm, glabras, ápice cubierto por la base del estilo. Distribución geográfica. Crece en Argentina en las provincias de Mendoza. Chubut y Santa Cruz, Dominio correspondiente fitoge 08 ráfic amente en el Alto-andino, provincia Patagónica, desde el nivel del mar hasta los 2000 m (Fig. Fenología. Se ha coleccionado en flor desde noviembre a enero. Botta, 1999: Iconografía. 67, fig. 124, 169. Observaciones. Entre las especies de la serie, esta se caracteriza por sus pequeñas hojas lineares, pubérulas. ARGENTINA. Chubut: Vallerini 2082 (5 1). Sosneado, el Material adicional examinado. Dpto. Ameghino, 21 km 5 de Garayalde, Mendoza: Dpto. San Rafael, El Lagiglia 8370 (SI). Santa Cruz: Dpto. Deseado, Deseado, Ameghino 78 (BA) Angulo, (Trone.) Moldenke, Phytolo- gia 19: 435. 1970. Basónimo: Verbena congesta Darwiniana 14: 631. 1968. TIPO: Argentina. Neuquén, Dpto. Zapala: In mountains in the Zapala district, 1925-1927, E. Opazo s.n., Herb. A. F. 1290 (bolsa po. K!, K foto SI). 4. Junellia congesta 435 Tronc., Comber Caméfito pulvinado denso, de 3 cm alt., ramas breves, densamente foliosas. Hojas homomorfas, de 3.54 mm, imbricadas, profundamente tripartidas, connadas en la base, carnosas, glabras, lóbulos linear espatulados, obtusos, enteros a veces los laterales algo divididos. Monobotrios en espigas contraídas, pluri- floras de 10 a 12 oblongas, subcarnosas, ciliadas en el borde. flores. densos: brácteas de 5-6 mm. Cáliz de 5.56 mm, piloso en la mitad inferior y sobre los nervios, con pelos glandulares breves. dientes suba- 350 Annals of the Missouri Botanical Garden gudos; corola de 6.5-8 mm, pubescente en la mitad superior externa, internamente con densos pelos retrorsos, fauce glabra; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo con estilo ensanchado. Fruto no visto. Distribución geográfica. Se conoce hasta el pre- sente sólo el material tipo de la especie (Fig. 1B). Iconografía. Troncoso, 1968: 631, 632, fig. 1 sub V. congesta; Botta, 1999: 170, 171, fig. 125. Observaciones. Especie afín a Junellia digitata de la que se diferencia por la pubescencia y la división de las hojas, las forma y división de las brácteas y tamaño de la corola. 5. Junellia digitata (Phil.) Moldenke. 5a. Junellia digitata (Phil. Moldenke var. Distrib. Verb. digi- tata, Known Avicenn.: 77. 1942. Basónimo: Verbena digitata Phil., Anales Mus. Nac. Bot. 8: 59. 1891. TIPO: Argentina. Catamarca, Antofa- 1885, F. Philippi s.n. (lectotipo, aquí designado, SGO 42522!). Figura 6F. Geogr. Santiago de Chile, gasta de la Sierra: Colorados, ene. Anales Univ. Chile 90: 615. Chile: “In Araucania habitat”, La T s. m a ee SGO 547831, isotipo, SI!). T spegazzinii Moldenke, Phytologia 7: 258. mbre ree vana Verbena P E Speg., Comun. B. Aires. 1(4): 137. Verbena liem Poe pp., Froriep's Ser. 1, 23: 292. 1829. | Tronc., Darwiniana 18: : Argentina, Salta, Cachi: Cerro de Ca achi, ene. 1897, C. Spegazzint n. (holotipo, LP 10399!; isotipo, SI!). Junellia punctulata Hieron. & Moldenke ex Moldenke, hytologia 3: 112. 1949. TIPO: Argentina. | Famatina: . Cuesta del Tocino, 2 feb. 1979 Hieronin & C. O 086 d MV M no visto; isotipos, CORD!, SI, 7 SID. Verbena rida : x Jot. V w a Pi 1895. Cueva, ene SGO foto SI; Dp © z g Mus. 899, nom. illeg., non s Not. Natur-Heilk. (de alada (Moldenke) 314, 1974 à Rioja, Sierra de Famatin E. W. = G foto Anales Mus. Nac. Santiago de Chile, l. Junellia tridactyla (Phil.) Moldenke, B 2: ar 1946. TI Copacoya, 19 feb. 1885, designado por Acevedo de 54801!; isotipo, SI!). O: Chile. Antofagasta: F. Si s.n. (lectotipo, Vargas [1951: 68], 5GO Caméfito en placa, estolonífero, formando pequeños de 3-5 cm alt.; con corteza ferrugínea que se desprende fácilmente y manchones aislados, tallos postrados entrenudos | mm. Hojas homomorfas, de 3—5(-7) X 5-7 mm, subcrasas, profundamente 3-partidas (a veces 2-partidas) a veces con la parte basal angostada, plana y limbo con lóbulos oblongos de ápice redondeado, generalmente dividido, con ambas caras — ifspido-glandulosas, las de los macroblastos opuestas as de los braquiblastos imbricadas; pecíolo de 2 mm o menor. Monobotrios en espigas contraídas, pluri- floras de 5 a 8 flores, densas; brácteas de 3-5.5 mm, 3-partidas hasta 3-sectas. Cáliz de 3-5 mm, velloso, con pelos glandulares, en la región apical: dientes breves, triangulares, subiguales; corola de 7.5-9 mm, rojiza, glabra externamente y pilosa en el interior; fauce glabra o con escasos pelos; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 4— Clusas de 2.5- 3.5(—4.5) mm, glabras, ápice obtuso y cubierto por la 6 mm, base del estilo ensanchada. base del estilo, base angostada, cara dorsal reticulada, LI de color castaño oscuro y las comisurales blancas, papilosas. Distribución geográfica. Norte de Chile y noroeste de Argentina, desde Salta hasta San Juan, en la provincia fitogeográfica Altoandina, crece en general en suelos pedregosos, entre los 3500 y 4700 m (Fig. 1C). Fenología. Florece y fructifica en verano. Iconografía. Botta, 1984: 350, fig. 8, a—h. Observaciones. Junellia digitata es afín a J. congesta; esta ültima se caracteriza por sus hojas glabras, con lóbulos enteros. Se diferencia de J. tridactylites, porque presenta la corola con pubescen- cia adpresa retrorsa, en la superficie exterior. En la descripción original de Verbena digitata, Philippi (1891) cita un ejemplar de Guanaqueros. Consultado el herbario de SGO, y otros herbarios, no se localizó ningún ejemplar de esa localidad, correspondiente a esta especie. Por lo tanto se procedió a elegir un lectotipo entre los materiales estudiados por el autor, colectados por F. Philippi en su viaje a Tarapacá y depositados en SGO, lugar de trabajo del autor, eligiendo el material citado por ser el más completo. En el caso de Verbena tridactyla, Acevedo de Vargas (1951: 68), la autora señaló SGO 54801, como “typus”, criterio que se comparte. Material adicional examinado. ARGENTINA. Cata- marca: Dpto. Antofagasta de la Sierra, Las Punillas, cerca de Falda de Ciénaga, 4300 m s.m., 19 feb. 1980, Cabrera Abra Chorrillos, 16 feb. ioja: Dpto. Vinchina, El , 3 feb. 1947, Hunziker J. E Pastos Grandes, | La Julia, Corrida de Cori, 4500 m s.m.. 17 . Jujuy: Dpto. a 1980, COTÉ 3l 769 (SI). H. 2146 sn Saltas L OS camino a Mu feb. 1945, Cohn 6788 i San Juan: Dpto. Iglesias, Reserva de in P as Carachas al Cajoncito . 3500 m s.m., .29 feb. 1984, Kiesling 4629 sn 'ucumán: an to. Taff, Cumbres Cale E cerro Negrito, 4200 m s.m., 3 ene. 1953, a 867 (SI). CHILE. Región: iem Manda. 15 abr. 1985, s. col., SGO Volume 95, Number 2 2008 Peralta et al. 351 Revisión del Género Junellia (Verbenaceae) 68385 (SGO). Il Región: Antofagasta, inter Vegas de Diablo & Antofagasta, ene. 1885, Philippi s.n. (SI); Desertum Atacama, feb. 1888, Philippi s.n. (K). 5b. Junellia digitata var. integerrima (Botta) Botta, Hickenia 2: 128. 1995. Basónimo: Verbena digitata var. integerrima Botta, Darwiniana 25(1—4): 350. 1984. TIPO: Argentina. Jujuy, Susques: lomás al pie del Cerro Tuzgle, 2 feb. 1944, A. L. Cabrera 8401 (holotipo, LP!; isotipo, SI!). Figura 6C. Difiere de la variedad típica por tener brácteas enteras, de igual o menor longitud que el cáliz. Distribución geográfica. En Argentina, en la provincia de Jujuy y en Chile en las regiones I y Il. oe suelos pedregosos aproximadamente entre los 450 m, fitogeográficas Altoandina y Prepuneña (Fig. 1C). en el límite de las provincias Fenología. Florece y fructifica en verano. Material adicional examinado. ARGENTINA. Jujuy: pto. Susques, Planic cie al S del Cerro Tuzgle, Werner 123 (SI). CHILE. I Región: Iquique, Collaguasi, Quebrada La Represa, Teiller 3082 (CONC). II Región: El Loa Ojo de San Pedro, Pisano 1792 (CONC). 6. Junellia minima (Meyen) Moldenke, Prelim. Alph. List Inval. Avicen.: 47. 1940. Basónimo: Verbena minima Meyen, Reise Erde 1: 451. 1834. TIPO: “Pérou, sur le gran plateau", Weddell, H. A. 1857-1861. Chlor. And. 2: 154-155, pl. 62B (neotipo, aquí designado). Incor. Sci. Names Verben. s die nail Moldenke, Phytologia 1: 483. 1941. entina. Jujuy: Tilcara ys above San munde a 000 ft., (holotipo, US no visto, US foto SI!; isotipo, K!, K foto SI!). Junellia minima var. strigosa Moldenke, Phytologia 5: 340. . TIPO: Argentina. Mendoza: Las Heras, Cruz de Paramillo, 3 ene. 1950, Melis y Paci 279 (holotipo, NY no visto; isotipo, SI!). Caméfitos en placas, de 1-5 cm de alt., con raíces 1-1.5 mm lineares o angostamente ovadas, por lo general leñosas. Hojas homomorfas, de 4—5 X enteras, a veces las inferiores 2—3-fidas, sésiles, densamente imbricadas, connadas en la base, sub- cóncavas, subcarenadas, rígidas, generalmente mu- cronadas, a veces espinescente en el ápice y algo incurvo, cara abaxial con el nervio medio prominente, cara adaxial densamente pubessente hacia el ápice, margen ciliado. Monobotrios de 2 a 3 flores; brácteas de 2.5-3.5 X 1.5-2.5 mm, ovadas hasta anchamente obovadas, cóncavas, de ápice agudo casi obtuso, pubescente en su cara abaxial, margen ciliado. Cáliz de 2.5-3 mm, pubérulo en su mitad superior, dientes subobtusos, ca. 0.5 mm; subiguales, triangulares, corola de 3.5—4.5(—5.5) mm, blanca o lilacina, glabra en el exterior, tubo corolino recto; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 2.5- Clusas de 2- 2.5 mm, glabras, ápice obtuso y cubierto por la base 3 mm, base del estilo ensanchada. del estilo, base angostada, cara dorsal reticulada, pardo-obscuro, caras comisurales papilosas, blancas. Cariología. 2n = 18 (Junellia minima, Botta y Brandham, 1993: 147, fig. 2D, “Kohtta” (Lilloa 5: 398. 1940); “césped de la puna” (de Cárdenas 3740) “yareta”, “yaretilla” (Botta, 1993: 74); “hareta” (Moldenke & Moldenke, 1949: 7) Nombre vulgar. Distribución geográfica. Perú, Chile, Bolivia y Argentina. Crece entre los 2800 y 4600 m, en suelos pedregosos de la Provincia Altoandina (Fig. 1 Fenología. Florece y fructifica de noviembre a febrero. Iconografía. Weddell, 1857-1861: pl. 62 (sub 'erbena minima); Ancibor, 1980: 164, fig. 1; 182, lam. 2; 184, fig. SA—E (sub V. minima). Observaciones. Especie muy afín a Junellia uni- flora de la que se diferencia por la morfología de la hoja, distribución de la pubescencia en el cáliz y las superficies comisurales de las clusas. No se han registrado hasta el presente de ejemplares que corroboren la presencia de esta especie en Chile, ya que Tacna, localidad donde se colectó el ejemplar Wederman 1120, actualmente pert a Perú. Por otra parte el ejemplar de Budin 7441 (LL 32890, SI), publicado por Moldenke (1944: 371) como J. hayekii Moldenke, se considera aquí J. minima El holotipo de Perú, Tacora, abril 1831, Meyen s.n., depositado en B, fue destruido. Debido a la ausencia de isotipos, fue necesaria la neotipificación; se designó la lámina de Weddell por tener claramente representados los caracteres que caracterizan la especie (Greuter et al., 2000: Art. 8.1 Con respecto a Verbena nubigena Poepp. (Froriep's Not. Natur-Heilk., Ser. 1, 23: 292. 1829), el autor no indica ejemplar de referencia, se ha considerado que podría ser un sinónimo de Junellia digitata, teniendo en cuenta la descripción origina Verbena straguloides (Moldenke) Tronc. (Darwini- ana 18: 313 porque no hay referencia a la cita bibliográfica donde 974) no era validamente publicado, se describió la especie. Material adicional examinado. ARGENTINA. Jujuy: Dpto. Cochinoca, Abra de la Cruz, camino a Casa Colorada, Dell'Arciprete s.n., BACF 2324 (BACK, SI). La Rioja: Dpto. Gral. Sarmiento, Reserva Laguna Brava, Pefias Negras, 352 Annals of the Missouri Botanical Garden campamento de la empresa minera El Dorado, Biurrun 4555 (SD). Mendoza . Las . Paramillos Carette s.n. MERL 2504 (ME FRI. NY). San Juan: Dpto. De | Observatorio El Leoncito al Portezuelo del Kiesling 4074 (SI). Dpto. Tafí, Castillon 3205 (LIL). BOLIVIA. La Alto, borde de la ciudad, Beck 3980 (SD. Oruro: Avaroa, De Challapata hacia Potosí, Ventilla, 14 km Macha, Beck 14110. (SD. Potosí: Frías, Tre s Palcas (atocha s Cardenas 37 740 NO 26 P y. CHILE. Chucuy de 17, Calingasta, O Tontal, Tucumán: Muñoz, Las Ánimas, Paz: Murillo, La Paz, El / antes del cruce hacia = =~ <` = = y | Cailloma. sobre Chivay, Gordillei Juliaca, Arequipa: (F). aa Torata. Pun Stafford 1130 (V). Cordillera Volos :án Tacara, 7472 (F) Tacna: Tacna, Ancara, Werder- mann 1120 (SD. Lilloa 5: Junellia patagonica (Speg.) Moldenke, 399. 1940. Basónimo: Verbena patagonica Speg.. Anales Soc. Ci. 15: 114. 1883. TIPO: Argentina, Santa Cruz: Río Santa Cruz, Cerro de los 1882, C. Spegazzini s.n. (holotipo, LP*; isotipo, SI). Argent. Caracoles, Patagonia Austral, ene. Verbena morenonts Kuntze, TIPO: Argentina. * 176 (holotipo, NY! Revis. Gen. Pl. 3(2): 256. 1898. "Patagonia, s.f”, F. Pai & Tonini Caméfitos pulvinados hasta 5 em alt., densísimos, a veces erectos, verde-claros. Hojas homomorfas, de 1-2 X 1-2 mm en la base, ovadas. enteras, revolutas, carnosas. con pubescencia adpresa-retrorsa y un mechón impor- tante de pelos canos en la base interior. Monobotrios plurifloros hasta de 10 flores, flores perfumadas, brácteas de 3-3.5 mm, ovadas, ápice agudo, tomentosas. Cáliz de 5-6 mm, tomentoso, adpreso-antrorso, a veces adpreso- retrorso, 5 dientes agudos, 2 más cortos que los otros; 12-12.5 violáceo y colores intermedios, perfumada, pubescencia corola de mm, blanca, rosada, amarilla, rojo pilosa a sericea, adpreso-retrorsa exteriormente: par superior. de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas, gineceo de 7,58 mm, base del estilo ensanchada. Clusas 3.5— 4 mm, ápice cubierto por la base del estilo, cara dorsal reticulada. Cariología. = 9 (Junellia patagonica, Botta & Brandham, co 147, fig. 2A). Distribución geográfica. Especie abundante en la provincia fitogeográfica Patagónica. Crece alrededor de los 1900 m (Fig. 1 — `) Fenología. Florece en primavera. Iconografía. Botta, 1999: 176, 180, fig. 133. Observaciones. Esta especie se caracteriza por la forma de las hojas y el mechón de pelos canos en la base de las mismas, además por formar cojines inermes. con. flores. intensamente perfumadas. En algunos ejemplares por ejemplo Donat 221 (BM, K, MO, SI), Liberman 4300 (MO), Vallerini 3430 (SI), Ruíz Leal 25711 (SD, se ha observado la presencia de hojas trilobadas, siendo los restantes caracteres propios de la especie. Phil. ex Ball (J. Linn. Soc. 27: 497. 1891) es un homónimo posterior. La descripción Verbena thymoides del mismo es muy incompleta y además no se ha podido localizar el material de Andrews: sin embargo Darwin 225, de ha determinado como Junellia en K, existe un ejempar de Porto Desiderato, que se patagonica, y podría ser el referido en la descripción original: “A plant which appears identical with this was collected Port Desire by Darwin”. Por esta razón podría tratarse de un sinónimo de J. patagonica. ARGENTINA. Chubut: Cabrera 33118 1000 m al Material adicional examinado. rg SI). Neuquén: Dpto. Añelo, sa. s Languineo, cerca de Fa. Quichana, Auc à. Mahuida. de la antena repetidora, Rossow 1615 (BAB, SI. Río pe Dpto. Valeheta, Somoncurá. lag. ida Correa 250 (BAB). Santa Cruz: Dpto. an de la Deseado, a 17 ) al a Ea. Romberg, una 10264 (BAB, SI Lilloa 5: 8. Junellia silvestrii (Speg.) Moldenke, peg 401. 1940. Verbena silvestrii Speg., Anales Soc. Ci. Argent. 53: 243. 1902. TIPO: Argentina. Santa Cruz: “secus río Santa Cruz” D feb. 1900, F. Silvestri s.n. (lectotipo, designado por Botta (1989: 389), LP! ISOLIPO, SID. Basónimo: Camefitos pulvinados de 3—4 em alt., inermes, muy duros, ramosisimos, glabros. Hojas homomorfas, de 2— 3(-5) X 1 slicas, carnosas, glabras o ciliadas en el margen y er mm, enteras, ovadas, imbricadas, tetrá- la base. Monobotrios bifloros, brácteas de 2-2.5 mm. ovadas, inflexas en el ápice, glabras, a veces ciliadas. Cáliz de 4 mm long., tubuloso, 5-costado, 5-dentado, dientes reducidos, subtriangulares. glabro: corola de 0-8 mm, tubo recto piloso o a veces glabro. limbo patente, 5-lobado: par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 6 mm, la base del estilo ensanchada. Clusas de 3 mm, elabras. ápice cubierto por la base del estilo, cara dorsal apenas relieulada, caras comisurales blancas con papilas cortas. Distribución geográfica. Argentina y Chile. Crece en la provincia fitogeográfica Patagónica. distritos del Golfo de San Jorge y Patagónico occidental y provincia Subantártica, distrito Magallanico (Fig. 2B). Fenología. Florece en primavera y verano. Botta, 1999: 183, fig. 135. Iconografía. Observaciones. Entre las especies de la serie, se caracteriza por sus hojas glabras, brillantes y dispuestas en forma imbricada. Volume 95, Number 2 2008 Peralta et al. 353 Revisión del Género Junellia (Verbenaceae) Material adicional examinado. ARGENTINA. Chubut: (5 Florentino Ameghino, Begoña, Soriano 2065 ($ ice Santa Cruz: Dpto. Deseado, camino Jaramillo— bosque petrificado, Crespo 1759 (BAB, SI). CHILE. XI Magallanes, Sierra Cazador, Magens 19090 (SI). 9. Junellia succulentifolia (Kuntze) Moldenke, Lilloa 5(2) 402. 1940. Basónimo: Verbena succulentifolia Kuntze, Revis. Gen. PL 3(2): 257. TIPO: Argentina. Patagonia, s.f., F. Moreno y Tonini 335 (holotipo, NY no visto; isotipo, LP!, Bt foto FM 17449 SI!). Verbena uu c Mi Turrill, Curtis Bot. Mag. 167(1): tab. 98. 1950. Junellia Sd oN (Turrill) d Por 3: 310. 1950. TIPO: Argentina. euquén, Huiliches: Malleo, 7 ene. e H. F a 957 (listino, n K foto SI). Arbustos bajos de 15 em alt., inermes, globosos, escabrosos con pubescencia glandular corta y pelos simples en tallos, hojas y flores, entrenudos de 5- 30 mm. Hojas homomorfas, de 4 X 1.5 mm, sésiles, opuestas, linear-oblongas, obtusas, brevemente loba- das en la base, a veces hojas trilobadas y enteras, carnosas, margen revoluto. Monobotrios en espigas contraídas plurifloras de 10 a 15 flores, 5-30 XxX 4-15 mm; ápice agudo, densas, alargadas en la madurez, de brácteas de 6 mm, elípticas, pubes- centes. Flores blanco-azuladas a violáceas, intensa- mente perfumadas. Cáliz 9 mm, con pelos cortos, rígidos; corola de 12 mm, tubo recto, exteriormente con pelos cortos retrorsos, fauce glabra; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 6.5— 7 mm, base del estilo ensanchada. Clusas de 4 X 1- 2 mm, glabras, ápice cubierto por la base del estilo, cara dorsal castaña, estriada, caras comisurales blancas, papilosas, margen lateral engrosado, for- mando una pequeña ala. Cariología. 2n = 18 (Junellia succulentifolia, Botta & Brandham, 1993: 147, fig. 2B). (de Offer- Nombre vulgar. “Yerba de la perdiz” man s.n., Sl). Distribución geográfica. Habita en la provincia fitogeográfica Patagónica, desde La Pampa a Chubut; 2208 5 i ; nivel del hasta los 1110 m crece desde el mar (Fig. 2C). Fenología. Florece en primavera y verano. Iconografía. Botta, 1999: 182, 187, fig. 139. (o e ^ [e] Observaciones. Se caracteriza por ser un arbusto bajo, con hojas carnosas, con pelos glandulares, lo que da a la planta un aspecto viscoso. Es afín a Junellia thymifolia, de la que se diferencia por la pubescencia de las hojas. Material adicional examinado. ARGENTINA. Chubut: Dpto. Ameghino, lkm antes del dique F. a 10196 (SI). La ~ el y Puelches, Cano 532 (S Florentino Ameg desde Dol: )pto. Curáco, entre Tigel Cal Bois n: Dpto. Catan Lil, Sierra de Catan Lil, ad 1 (SI). Río Negro: Dpto. Avellaneda, 70 km al NW de os a nucos, Ruíz Leal 26345 (S — c 10. Junellia thymifolia (Lag.) Moldenke, gia 2(2): 51. 1941. Basónimo: Verbena thymifolia Gen. Sp. PL 18. 1816. TIPO: Argentina. Santa Cruz, Lago Buenos Aires: ruta 40 y Valle 1 feb. 1949, A. Soriano 3406 (neotipo, aquí designado, SI!). Phytolo- Lag., del río Deseado, Arbustos bajos, erectos, tallos con pubescencia adpreso-retrorsa, en general con desarrollo de braqui- blastos, poco o nada ramificados. Hojas homomorfas, de 2—4.5 recurvos y con un falso ojal en el dorso, híspidas y con mm, enteras, oblongas, carnosas, márgenes algunos pelos largos blanquecinos. Monobotrios, en espigas contraídas, de 10 o 11 flores en el ápice de cáliz de 7- macroblastos; brácteas de 4-4.5 mm. 8 mm, dientes de 1-1.2 mm, pequeños, subiguales, pubescencia densa, adpresa-retrorsa; corola de 15 mm, lila o azul, pubescencia híspida, fauce glabra; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 7.5-8 mm, base del estilo ensanchada. Clusas de 4 mm, glabras, cilíndricas, ápice obtuso y cubierto por la base del estilo, cara dorsal reticulada especialmente en el ápice, caras comisurales papilosas. Distribución geográfica. Especie abundante en Neuquén, Chubut y Santa Cruz, y en Chile en la XII Región, en la provincia fitogeográfica Patagónica donde se encuentra en zonas áridas y en médanos (Fig. 2E). Fenología. Florece y fructifica de octubre a enero. Iconografía. Botta, 1999: 184, 187, fig. 138. Observaciones. | Esta especie se caracteriza por la presencia de pelos blanquecinos largos y la ausencia de pelos glandulares, carácter que la diferencia de Junellia succulentifolia. Como neotipo se eligió un material proveniente del mismo lugar que la colección original, citado por Botta (1989: 390) Sprengel (1825: dato que se repite en el Index Kewensis (Hooker & Jackson, 1977: 1179). La expedición de Malaspina, Europa llegaron a 148) cita la especie para Paraguay nunca pasó por Paraguay: desde Uruguay y desde allí continuaron hacia el sur. llegando a Puerto Deseado, por lo que la cita. para En el caso de Verbena Paraguay es incorrecta. thymifolia el nombre fue usado por Philippi (1895: Annals of the Missouri Botanical Garden 613) con duda (7), coincide con los materiales observados citando la y amplía la descripción, que especie para Lago Santa Cruz (actualmente Argen- tina). ARGENTINA. o Dpto. Biedma, Península de Valdéz, Frazier s.n. (BA); Dpto. Escalante, a del Casullo, brea o. > e Indios, a 51 km SW ruta 24, Ea. La Marfil, idis 212 (SI); Dpto. Río Senguerr, 130 km al E de , Soriano 3211 (S entre Material adicional examinado. ; Dpto. Sarmiento, Cabrera 33201 1/2 (S Soriano 504, Hermoso, (5). Neuquén: Dpto. Huiliches, a de lo Andes, Cerro Santa Julia, i ee 79/V II (SI). Santa Cruz: Dpto. Deseado, ruta 281, 29 km al W de Puerto De a Botta 472 (SI); Dpto. e Buenos Aires, ruta 40, 33 al N del a con ruta 43, Botta 534 (SI); Dpto. Rio Chico, ruta 28, : is Horquetas, valle del río Chico Boelcke 12787 (SI); Dpto. Magallanes. $ à A (SI. CHILE. XII Región: Magallanes, CONC). n de Gob. ( Gregores a "24 0°40'W an Julián, Blake 424 Gunckel 70300 11. Junellia tridactylites (Lag.) Moldenke, Lilloa 5(2): 402. 1940. Basónimo: Verbena tridactylites 19. 1816. TIPO: Argentina. Santa Cruz: Lago Argentino, ruta Nac. 40, 7 km Lag., Gen. Sp. Pl. = al N del empalme con el camino a Calafate, 1993, 5. M. Botta 517 designado, SI!). ene. (neotipo, aquí Verbena struthionum Speg., Anales Soc. Ci. Argent. 15: 116 Junellia struthionum (Speg.) eae a Known n Geogr. Distrib. Verb. Avicenn.: TIP( " entina. Santa Cruz: Río Santa e p 1122 pd. LP!, LP foto SI!; m SI! Anales Univ. Chile 90: 694. 1895. ombre remplazado: Verbena a naea 29: 21. 1857, , Nov. Gen. Sp. 2: 27 8. Junellia minutifolia (Phil.) Moldenke, Lilloa 5(2): Me 1940. TIPO: Chile. Cordillera, ene. 1856, Germain s.n., S 47 (holotipo, SGO!, SGO foto isotipos, F no visto, SI!). 43. 33. = 05. Verbena prichardii Rendle, J. Bo gaz V s minutifolia P hil., . Lin- nom. illeg., non Verbena i 21 2. Co E Nombre », J. Bot. 42 370. 1904, nom. illeg., non Verbena a Es, 153. l eemplazado: Verbena ete Renc Sp. Argentina: South Dansónie: 0, a X Mt. Buenos Aires, 1901, richarc Junellia rosulata Moldenke f. rosulata, Phytologia 2: 135. 1946. TIPO: Argentina. Chubut: Teeka, 25 dic. 1945 T. er 9586 (holotipo, NY no visto; isotipos, LILL Junellia roudata Moldenke f. alba Moldenke, Phytologia 2: . TIPO: Argentina. Chubut: Ea. “La Mimosa”, 2 T. Meyer 9587 (holotipo, NY no visto: isotipo, LIL! 29 od . Caméfitos pulvinados, postrados o erectos de 15— 20 cm alt., foliares en cojín o carpeta, laxos, inermes; ramas con entrenudos de 5-6 mm, pubescencia híspida. Hojas homomorfas, de 3-5 mm, trilobadas a Monobotrios en a 15 flores; tripartidas, carnosas, pubescentes. espigas contraídas, plurifloras de 10 brácteas de 4 X 2 mm, ovadas. pubescentes, de bordes recurvos. Cáliz de 7—7.5 mm, dientes de 1.5 mm, desiguales, pubescencia generalmente densa, adpreso-retrorsa; corolas de 10-12 mm, perfumadas, blancas, viólaceas, rojizas v/o amarillas, pubescencia densa adpreso-retrorsa, fauce glabra; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 9— 11 mm, base del estilo ensanchada. Clusas de 4.6 X 1.1 mm, glabras, ápice obtuso y cubierto por la base del estilo, cara dorsal reticulada. 18 (sub Junellia atf. eae 1993: 144, fig Cariología. 2n = lia, Botta & Brandham, Distribución geográfica. Argentina, desde el sur de Mendoza hasta Santa Cruz, y Chile. En la provincia fitogeográfica Patagónica y áreas limítrofes del Monte, crece entre los 2300 y 2350 m (Fig. 2E). Fenología. Florece en diciembre. Iconografía. Botta, 1999: 174, 177, fig. 129. Observaciones. El color de las flores es muy variable: se han observado en la espiga flores centrales blancas y marginales violáceas. Es probable que la coloración las flores varie en la relación con la madurez de las mismas. Esta especie se caracteriza por la pubescencia del cáliz, abundante y adpreso- retrorsa, al igual que la parte exterior de la corola. Es afín a Junellia thymifolia, de la que se diferencia principalmente por la pubescencia en las flores. Se designa como neotipo el ejemplar de Botta por ser el más completo y tener como lugar de colección uno de los puntos donde arribó la Expedición Malaspina (ver comentario sub Junellia thymifolia). Lagasca en su descripción original, cita como hábitat de Verbena tridactylites las Islas Malvinas, lo cual, a pesar que Née coleccionó allí, la especie no proviene de ese lugar ya que el tipo de vegetación de las islas es totalmente diferente. Además, la distribu- ción más austral de la Tribu Verbenae es la provincia de Santa Cruz. or Verbena philippiana Kuntze (Revis. Gen. Pl. 3(2): 257. 1898), para V. microphylla Phil., el autor designa un nuevo nombre cuando ya existía un nombre reemplazante para esta especie. Por Verbena patagonica Moldenke (Phytologia 1: 171. 1935), non V. patagonica Speg., Moldenke usa este nombre, siendo éste un homónimo posterior y además lo designa como nombre reemplazante para V. bonariensis Rendle, que ya tenía un nombre asignado. Material Ape examinado. ARGENTINA. Chubut: o. Futaleufú, Portezuelo del C hoique, fun Leal 21454 (SD. Neuquén: Dpto. Chos Malal, Extremo NW de la Pampa Ferraina, Boelcke 11377 (BAA, — = > Ww ‘XO ~ = c Volume 95, Number 2 2008 Peralta et al. Revisión del Género Junellia (Verbenaceae) fie! tc e S- / MI n A Z UE EEN ID NS MW Of Witty Y 5 yA 3 Cath i y Z NYA KI CX N i Y s. eS 2 MZN EN y NES a A OR VEES eres SEA pee As N AP —- OS — m (NOT my Figura 4. A-D. Junellia toninii var. toninii. —A. Aspecto general. —B. Detalle de inflorescencia. —C. Detalle de la pubescencia en el cáliz. —D. Detalle de un nudo. A-D de Cabrera 33013. E, F. Junellia toninii var. mulinoides. —E. Flor. — F. Detalle de un nudo. E, F de Ruíz Leal 26359. —G. Junellia trifurcata. Detalle de inflorescencia (de SGO 54802). BAB, SI). Río Negro: Dpto. Norquinco, cerca de Río Chico, hacia Norquinco, Cabrera 33067 (SI). Santa Cruz: Dpto. ; I). 12. Junellia trifurcata (Phil. Moldenke, Known Geogr. Distrib. Verb. Avicenn: 77. 1942 Basónimo: Verbena trifurcata Phil., Linnaea 29: 21. 1857. TIPO: Chile. Cordillera de Santiago, feb. 1857, Germain s.n., SGO 54802 (lectotipo, designado por Botta (1989: 389), SGO!, 5GO foto SI!). Figura 4G. Arbusto hasta de 10 em de alt., cespitoso y muy ramificado. Hojas homomorfas, de 3-5.5 X 1.5- mm, imbricadas, sésiles, no carnosas, tripartidas, lóbulos lineares con margen revoluto, agudos y punzantes, villosos en ambas superficies; monobotrios en espigas bifloras; brácteas ca. ovadas, enteras, villosas. Cáliz de 6-8.5 mm, villoso, dientes triangulares ca. 2-3 mm; corola de 10-11.5 mm, exteriormente pubescente, fauce glabra; par superior de estambres sin apéndices conectivales, conectivo no o mm, superando la longitud de las tecas; gineceo de 6 mm, base del estilo ensanchada. Clusas de 3 mm, glabras, 356 Annals of the Missouri Botanical Garden ápice cubierto por la base del estilo, cara dorsal reticulada, caras comisurales papilosas. la cordillera Distribución geográfica. Chile, en cercana a Santiago, en la provincia fitogeográfica Chilena Central (Fig. 2F). Observaciones. Especie afin a Junellia aretioides; esta última presenta hojas trilobadas, lóbulos redon- deados. Material adicional examinado. CHILE. Región Metro- politana: Cordillera cercana a Santiago, 1850, Germain s.n., SGO 42483 (SGO, SI): e Yeso, Dec. 2 (K, foto FM 17455 1873, Reed I Lilloa 5: Junellia uniflora (Phil. Moldenke, 102. 1040. Basónimo: dle uniflora Phil., Linnaea 29: 20. 1857. TH Chile. Santiago, die. 1856. C. Ga) 1060 msn SGO 5480( isotipos, P! SIB). Figura 3E, G Ii \ erben na uniflora var. glabriuscula Kuntze, Revis 258. 1898. Junellia a var. ae ) Moldenke, Lilloa 5: 403. 1940. TEPC Arge P: F. More eno & Tonini E ed aquí designado, NYL; isotipo, SI). sen. e a J tagi onia, Camefitos en placas densísimas, de 3—4 em alt. Hojas homomorfas, de 2.5-5 X 1.3 mm, carnosas, imbricadas, enteras. lineares, obtusas, inermes o con el ápice punzante pero no transformadas en espinas, hirsutas en ambas caras, ciliadas, nervio medio prominente en la cara adaxial. Monobotrios de la 2 3 X 2 mm, ciliadas, ápice agudo. flores: brácteas ca. anchamente ovadas, Cáliz de 5 mm, con 5 dientes triangulares, pubescencia hirsuta; corola de 8.5- 9 mm, glabra exteriormente; par superior de estam- bres sin apéndices conectivales, conectivo no supe- rando la longitud de las tecas; gineceo de 6-7 mm, base del estilo ensanchada. Clusas de 3.5 X 1 mm, glabras, cilíndricas. ápice obtuso y cubierto por la base del estilo, cara dorsal reticulada, caras comisurales lisas o con escasas papilas pequeñas en los márgenes. Nomore vulgar. “Jazmín del campo” (de Kiesling ) Distribución geográfica. Crece en Argentina en las provincias de San Juan, Mendoza, Neuquén, en Chile en las Regiones Metropolitanas, HE IV. V y VII, Altoandina, la provincia fitogeografica entre los 2500 y 4000 m (Fig. 2F). el Fenología. Se ha colectado en flor en el mes de enero. Sanzin, 1919: 109, fig. 13 (sub 1963-1972: 26, fig. 1979: 8, 9, fig. 1, 2 199, fig. 144. Iconografía. Verbena uniflora); Bocher et al., Bocher, 1999; 13 (sub Verbena uniflora): (sub Verbena uniflora); Botta, Observaciones. Especie afín a Junellia minima; se diferencia por la pubescencia del cáliz y la superficie de las caras comisurales en las clusas. En la descripción original de Junellia uniflora var. glabriuscula, se citan dos ejemplares, de los cuales se eligió F. Moreno y Tonini 294 por ser el más completo. Material adicional examinado, ARGENTINA. Men- doza: Dpto. San Car «m * sobre camino a Laguna Diamante, Boelcke one (SI). Neuquén: Dpto. Chos Malal, Cajón de ) Boelcke 1 Calingasta, paso desde Cuesta i Rio Blanco i la el Valle de Los Patos Norte, Kiesling 7774 (St). CHIL oe Me seas arta Prov. Grandjot codd 0050 Ou lle, Quebrada /^ Región: Los Andes. ol Inca, (CONC). VIT Región: Cordillera de Linares, Philippi s.n., S1 3447 (SD. O o del Cruce, faldeo S del macizo del Domuyo, 283 (SI). San Juan: Dpto. Ac eua, Cordillera à a D IH. Región: (B = Cordille uva de Jiles 2957 (SI). Arroyo 81273 . IV Region: Junellia Moldenke liopsis (Botta) P. Peralta € Múlgura, comb. nov. subgén. Junellia secc. Junel- Junellia subgén. Thryothamnus sece. Junelliopsis Darwiniana 29: 392. 1989. TIPO Phil. |= Junellia Botta, p.p., Verbena bryoides bryoides (Phil.) Moldenke]. Verbena sece. Bol. Syst. Erinacea (Walp.) Tronc., . Reper dun llia Verbenaca ser. Erinaceae X s rb AI n Z SOT; 1974. DONA "m alp.) Botta, TIP Junellia erinacea (Gillies & Darwiniana 29: 386. 1989. Verbena erinacea Gillies & Hook. ex Hook. [= Hook. ex Hook.) Mol- - denke Verbena sece. Verbenaca ser. Syst. 4: 14. E Walp.. p. Repert. Bot. Syst. 4: 18 syn. nov. Verbena subgén. Junellia ser. voii n ap emend. Tronc., pa Darwiniana 18: 312. 1974. Junellia subgén. ] j Seriphioideae (Walp. ) Botta, Darwiniana 29: 386. 1989. TIPO: ) Gillies & Hook. ex Hook. |= (Gillies € Hook. ex Hook.) Tu ser. 2 seriphioides Moldenke]. Arbustos bajos o cojines. Hojas homomorfas, todas espiniformes u hojas dimorfas, las de los macroblastos opuestas, enteras o partidas, espiniformes o al menos punzantes, las de los braquiblastos tetrásticas, carnosas, enteras o partidas. Monobotrios pauci- o plurifloros. Anteras sin apéndices conectivales. Nú- mero eromosómico n = 10. Observaciones. Se pudo analizar material más Junellia subgénero Thryothamnus sección Junelliopsis Botta 1989: 393) y morfológicos propios del subgénero Junellia, por lo completo de bryoides, especie lipo del se comprobó que presenta caracteres cual se transfiere la sección Junelliopsis a este subgénero. Se hace necesario entonces, realizar una Volume 95, Number 2 Peralta et al. 357 2008 Revisión del Género Junellia (Verbenaceae) AL rd SS E ARA LEE E OO a =o ra D DAE ct : Xe II PET T LACE = E P EE oe E GT = Se mat MM WV n a | MEME Figura 5. A, B. Junellia bisulcata var. bisulcata. —A. Inflorescencia. —B. Cáliz abierto. A, B de Cabrera 31700. C. D. Junellia Dun var. campestris. Eu Inflorescencia. —D. Cáliz abierto y fruto. C, D de ii diia 7751. E-l. Junellia bryoides. —E. Aspecto general. . Detalle de rama e inflorescencia. —G. H. Anteras. —l. Fruto. E-1 iis Rhamer s.n.. : 2472. nueva combinación a nivel de sección para los Patanca, 3700 m, 1903-1904, K. A. G. Fiebrig taxones agrupados por Botta en la sección Junellia 2473 (lectotipo. nO K!: isotipos, BM!, series Erinaceae y Seriphioideae. Aquí se considera K!, L no visto, P!, SIL US no visto, W no visto, BY que estas series no presentan caracteres suficientes foto FM 17404 E Figura 5A, B. para el mantenimiento de las mismas. Arbustos bajos, postrados, muy ramosos, de ramas Hayek) Moldenke. decumbentes, estriadas, de 10—40 cm. l4a. Junellia bisulcata (Hayek) Moldenke var. erguidas, pubescencia hispida, pelos adpreso-retror- bisulcata, Known Geogr. Distrib. Verb. Avicenn.: — sos. Hojas dimorfas, las de los macroblastos de 3-6 77. 1942. Basónimo: Verbena bisulcata Hayek, | (—7.5) X 1.5-5 mm, trisectas, lóbulos desiguales, Bot. Jahrb. 42: 164. 1909. TIPO: Bolivia. Puna — angostamente oblongos, el central de mayor longitud, 14. Junellia bisuleata — a veces 358 Annals of the Missouri Botanical Garden rígidos, espinescentes, ápice mucronado, el nervio central prominente en el envés y margen ciliado entero, revoluto, pubescencia escasa, generalmente patentes; hojas de los DraqiubiBstos 2.5-3 X 2.5- 2.8 mm, raramente enteras, ápice obtuso hasta subagudo, por glabras subcrasas, 3- o 2-lobadas, más lo general brevemente mucronadas, si el braquiblasto se desarrolla aparecen hojas trisectas. Monobotrios en espigas contraídas de 1—1.5(—2.5) em, plurifloras de 6 a 12 (o 18) flores; brácteas de 3-5 mm, enteras, ovadas, en general de ápice agudo, a veces cuspidado, cilioladas, iguales o más breves que el cáliz. Cáliz de 5-7 mm, con pelos breves, patentes, esparcidos en su superficie, más densos en su mitad superior, dientes del cáliz triangulares, menor a 1 mm; corola de 10-13 mm, lila, rosada o púrpura, glabra exteriormente, velloso en el interior, fauce pubescente; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 6.5-8 mm, base del estilo ensanchada. Clusas de 4.5-5.5 mm, glabras, subcilíndricas, rostradas, cubierto por la base del estilo, cara dorsal reticulada, lisa en el ápice, castaño claro, caras comisurales blancas, papilosas. Distribución geográfica. Sur de Bolivia y noroeste de Argentina desde las provincias de Jujuy a La Rioja. Fitogeográficamente crece en el dominio Andino- 2040 y patagónico, provincia los 3300 m (Fig. IB). Puneña, entre Fenología. Florece y fructifica de noviembre a marzo. Iconografía. Botta, 1984: 344, fig. 5. Botta, 1993: 73, fig. 32 Observaciones. Especie afín a Junellia juniperina (Lag.) Moldenke que presenta las clusas truncadas y los tallos con pubescencia híspida. El holotipo de esta especie estaba depositado en B, y actualmente está destruido, razón por la cual se eligió un lectotipo entre los isotipos encontrados. Material marca ARGENTINA. Cata- ujuy: Dpto. Humahuaca, : ‘hapu Rodeo, nacientes ; del A rroyo La Cueva, Novara 8770 SI). La Rioja: Dpto. Chilecito, Cuesta de Miranda, Halloy D (SI). Salta: Dpto. Cachi, Cachipampa, Cabrera 30747 SI). Tucumán: Dpto. Tafí, subida al Abra del Infiernillo B Amaicha, Kiesling 5855 (SI. BOLIVIA Tarija: Ehrich 48 (SI); Avilez, cerca ds Cobre, adicional examinado, orgensen 1733 ($ , Iscayachi, Mér Baian 602 (SD. 14b. Junellia bisulcata var. campestris (Griseb.) Botta, Hickenia 2: 127. 1995. Basónimo: Verbena Abh. Königl. juniperina var. campestris Griseb., Ges. Wiss. Göttingen 19: 193. 1874. Junellia Juniperina var. campestris (Griseb.) Moldenke, Phytologia 44: 215. 1979. TIPO: Argentina. Córdoba: Chañar, N de la prov. de Córdoba, 29 nov. 1871, P. Lorentz 7 (holotipo, GOET no visto; isotipos, CORD!, K!, SI!, K foto SI!). Figura 5C, D. Difiere de la var. bisulcata, principalmente por las hojas en general no patentes y los dientes del cáliz atenuados, los anteriores ca. 1.5 mm long. Distribución geográfica. Argentina, en las provin- clas de Córdoba y Santiago del Estero, en la provincia fitogeográfica Chaqueña, distrito Chaqueño-serrano Fig. 1B) Fenología. Florece y fructifica en verano. Material adicional examinado, ARG ENTINA. Córdo- 2 b to. Sobremonte . Burkart 2 (SI). Santiago de 1 Es tero: Dua. Ojo de AE ruta 9, re con Caoba Burkart 20110 (SI). Francisco de Cha — 15. Junellia bryoides (Phil) Moldenke, Lilloa 5: 394. 1940. Basónimo: Verbena bryoides Phil., FL Atac.: 40. 1860. TIPO: Chile. Desierto de Ataca- ma: Alto de Varas, ene. 1854, R. A. Philippi s.n., p.p. SGO 54075 (lectotipo, designado por Ace- vedo de Vargas (1951: 53), SGO!). Figura SE-I. Arbustos erectos 1 m de alt; tallos con pelos simples, patentes, macroblastos más o menos erectos. Hojas dimorfas, sésiles, las de los macroblastos 3- partidas o 3-sectas, lóbulos de 1.5-2.5 mm, oblongos, carnosos, ápice agudo, rígido, escabrosos en toda su superficie, pelos algo más largos en el margen, segmento medio de mayor longitud que los laterales: hojas de los braquiblastos de 1.5-2 mm, enteras, ovadas, carnosas, ápice rígido; pubescencia igual a las anteriores. Monobotrios plurifloros; brácteas de 3- 3.5 mm, enteras, ovadas, ápice agudo, rígidas; con los pelos más largos que los de las hojas, muy aislados en la superficie, más abundantes en el margen. Cáliz de 3.04 mm, pubescente exteriormente; dientes desi- guales, subulados, rígidos, el mayor 1 mm; corola de 10-10.5 mm, glabra en el exterior; fauce glabra o con — escasisimos pelos semejantes a los del interior del tubo corolino; par superior de estambres sin apéndices conectivales, conectivo no superando la de 7- ensanchada. Clusas 2 mm, longitud de del glabras, ápice truncado y las tecas; gineceo mm, base estilo cubierto por la base del estilo, cara dorsal pardo obscura, caras comisurales papilosas. Nombre vulgar. “Cola de zorro", “pata de perdiz” 8 | I (Acevedo de Vargas, 1951: 53). Distribución geográfica. Chile, desierto de Ata- cama. Crece en el dominio Andino-patagónico, Provincia del Desierto, alrededor de los 3500 m (Fig. 1B). Fenología. | Florece y fructifica en verano. Volume 95, Number 2 2008 Peralta et al. Revisión pa Género Junellia (Verbenaceae) llia respit JS jb Am IA NS OIZ & F Figura 6. A-E. J Corola. —E. Ant j —B. Deta as. A-E de ae E —F. Ju nelli ee var. us Flor y brácteas (de Cabrera 31856). ). j ( hb DEA ANS V IU —C. Cáliz abierto. E e de rama e inflorescencia. ~ Junellia noun var. integerrima. Flor y bráctea (de Cabrera 8041 Iconografía. Philippi, 1860: tab. V, fig. C, sub Verbena bryoides. Observaciones. Junellia bryoides y J. hystrix son muy similares en su aspecto vegelalivo, porte y desarrollo de la planta; esta ültima presenta la base del estilo incluido entre las clusas. Esta similitud entre las especies mencionadas, motivó la sinonimia propuesta por Moldenke (1940: 39 Acevedo de Vargas (1951: 52) las reconoce como entidades independientes (bajo el género Verbena) basándose en las características de la pubescencia de las hojas y de la corola, haciendo constar que en el ejemplar SGO 54675 hay mezcla de las dos especies y elige al mismo como lectotipo de V. bryoides. Botta (1989: 393) funda subgénero Thryothamnus sección Junelliopsis, y designa como especie tipo a Junellia bryoides sin mencionar a J. hystrix. Material adicional examinado. CHILE. I Región: Tara- paca, Jaiña, Rahmer s.n., SGO 42472 (SG e s. loc., rie s.n. (BM) 16. Junellia caespitosa (Gillies & Hook. ex Hook.) Moldenke, Lilloa 5: 394. 1940. Basónimo: Verbena caespitosa Gillies & Hook. ex Hook., Bot. Misc. 1: 165. 1829. TIPO: Argentina. Mendoza: “Uspallata summit of the Uspallata range moun- tain”, J. Gillies s.n., Herb. Hook. (lectotipo, designado por Botta (1989: 388), K!, K foto SI). Figura GA—E i comberi Sandwith, Kew Bull. 1927: 185. 1927, syn. "eU. T ian (Sandwith) Moldenke, Lilloa 5: entina. Ye ed río Coyunco, 12 dic. Tus P P e n 10 [oip K!, K foto SI!; isotipos, BM!, TEX no visto). Caméfitos pulvinados, en cojines hemisféricos densos o caméfitos en placas, macroblastos rastreros, bien desarrollados o no. Hojas dimorfas, la de los macroblastos de 5-15 X 0.5-2 mm, enteras, lineares, espiniformes; hojas de los braquiblastos de 1.5-2.5 X 1-1.5 mm, imbricadas, ambas con escasos pelos cortos, antrorsos en la cara abaxial y pelos largos adpresos densos en la cara adaxial. Monobotrios bifloros; brácteas 2.5-3(-5) X (1.5-)2-2.5 mm, ovadas o anchamente ovadas, ápice agudo, margen ciliado, pubescencia adpresa en ambas caras. Cáliz 4.5—5(-6) mm, dientes desiguales, l mm, triangulares, no punzantes, con pelos cortos enteras, ovadas, carnosas, suaves y adpresos, sobre las costillas y los dientes, a 360 Annals of the Missouri Botanical Garden veces erectos en la parte basal del cáliz: corola 10— 10.5 mm. lila claro. glabra en el exterior, pubescente en el interior, fauce glabra; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas: gineceo de 6.5-7 mm, base del estilo ensanchada. Clusas de 2.5-3(—4) mm, glabras. ápice truncado, cubierto por la base del estilo. base caras comisurales angostada, cara dorsal reticulada, papilosas. Nombre vulgar. “Llareta” (Hooker, 1829: 165). Distribución geográfica. Chile, regiones HI y IV y en Argentina en San Juan, Mendoza y Neuquén: en el extremo sur de la provincia fitogeográfica Puneña del dominio Andino-patagónico, al oeste de la provincia del Monte, dominio Chaqueño, y en la provincia Patagónica. dominio Alto-andino, entre los 1400 y 3700 m (Fig. 1B). Fenología. Florece y fructifica en verano. Iconografía. — Sanzin, 1919: 108, fig. 11; Botta. 1999: 167, fig. 123 (sub J. comberi). Observaciones. Esta especie es afin a Junellia toninit de la que se diferencia por ser un arbusto bajo con hojas enteras o partidas, monobotrios plurifloros y cáliz glabro o con pubescencia hispida. ARGENTINA. Men- 414 Neuq Dpto. C < M: dal, 21 km a de Chos Vallerini 3478 (SI). San en Dpto. AÑ » Kiesling 0558 (Sl). CHILE Laguna Grande, Material adicional nw doza: Dpto. Las Heras, del vg te Rotman Calingasta, >» HI Región: Johnston 5937 aia cordillera de Huanta, Gay 310 Dpto. Y allenar, (K). IV Región: P Junellia denke, Mol- Verbena Gen. PL 32) Mendoza: 1650 m, 1992, O. Kuntze s.n. (lectotipo, aquí designado, NY!) connalibracteata (Kuntze) 394. 1940. Kuntze. TIPO: arroyo de Lilloa 5: Basónimo: Revis. i " " CUCUCHTLTULUEELUT CEC EC CELLULE er vat D99. 1893. Cruz. Ww Argentina. Paso los. Papagayos, ene. Verbena eet ia Sandwith, Kew Bull. 1927: 186. 19 |. Junellia dolichothyrsa (Sandwith) vn a Lilloa. 5: 395. 1940. TH Neuquén, d Zapala, Cerro oleha and Coyunco, 900 m, l4 Oct. 1925, H. F. Comber 79 (holotipo, K! K foto SIL; Ee BM!, TEX no visto - Pd — Argentina. —0.60 m de alt. Macroblas- Hojas dimorfas, I=1 ovado-angostas, de Arbustos bajos de 0.15 tos erectos, hispídulos. las de los macroblastos de 5-10 X .5 mm, sésiles. connadas en la base, ovadas a sección Iriqueta, ápice agudo, rígido punzante, en general recurvas, con el nervio medio prominente en la cara abaxial, con pelos cortos; adpresos y de densidad variable en ambas caras; hojas de los macroblastos de 3.504 mm, ovadas, carnosas, ápice agudo rígido, algo plegadas, pelos adpresos en la cara abaxial, margen ciliado. Monobotrios contraídos, plurifloros; brácteas de 3.54 mm, anchamente ovadas, connadas en la base, de ápice agudo, a veces largamente atenuado. con pubescencia adpresa. Cáliz de 5.5-7 mm. dientes triangulares con ápice atenuado y rígido, el adaxial y rk muy breve; pubescencia de densidad variable. adpresa en las regiones intercostales, a veces algo híspida sobre los nervios: ciliado en el ápice, entre los 10-13 mm, rosada, glabra en el dientes; corola de blanco crema hasta exterior, fauce. glabra y tubo pubescente en el interior; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 6-6.5 mm. cilín- drico, base del estilo ensanchada. Clusas de 4 X .o mm, glabras, cilíndricas, ápice truncado. cubierto La elal línd pice 1 | bierl por la base del estilo, cara dorsal castaña, rugosa. caras comisurales blancas, papilosas, papilas digita- liformes. (Ruíz Leal, 1972: 240) Vombre vulgar. “Tomillo” Distribución geográfica. Argentina en las provin- cias de Mendoza. La Pampa. suroeste de Buenos Aires, Neuquén, Río Negro y Chubut: fitogeográfica- habita en la mente provincia Patagónica en los distritos de la Central, a 900 m (Fig. 1C). Payunia, Occidental y Fenología. Florece de octubre a febrero. plc bs Sanzin, 1919: 123, fig. 27; Ruíz Leal, 1972: 239, lám. 74, fig. 275; Botta, 1999: 171, fig. 126. Observaciones. La especie se caracteriza por sus brácteas connadas y hojas recurvadas, que la diferencian de las restantes de la sección. En la descripción original se citan dos ejemplares de los cuales, se designa como lectotipo al ejemplar Kuntze s.n.. por ser el más completo El mayor o menor grado del desarrollo de los braquiblastos, así como la variación en la densidad de la pubescencia, otorgan aspectos muy variados a esta Monticelli (1938 estos caracteres, consideró tres formas ecológicas, sin especie. 358) teniendo en cuenta descripción en latín, ni cita de ejemplares, por lo tanto son nombres inválidos: (1) Verbena connatibrac- 1938); (2) glomerata Monticelli 1938): (3) Verbena connatibracteata f. 1938) \Vunellie connatibracteata (Kuntze) Moldenke f. rosulata (Mon- tecelli) Moldenke (Phytologia 2: 135. 1946) |. teata T. normalis Monticelli (Lilloa 3: 358. Verbena connatibracteata T. (Lilloa 3: 358. rosulata Monticelli (Lilloa 3: 358 En Flora Patagónica (Botta. 1999: 175). se consideró a Junellia dolichothyrsa, como sinónimo de J. mulinoides. El aspecto general, las hojas enteras Volume 95, Number 2 2008 Peralta et al. 361 Revisión del Género Junellia (Verbenaceae) y las brácteas connadas han sido los caracteres considerados para sinonimizar J. dolichothyrsa con J. connatibracteata. ARGENTINA. Min pita examinado. Patagonia SI 50/3, unt (paratipo NY, isoparatipos LP, SI). oe ru Partido Villarino, Fernández y Zalba s.n., Herb. Villamil 7159 (BBB). Chubut: Dpto. Paso de los c ruta 25, a 42 km S de Paso de Indios, S. Arroyo 66 D La Pampa: Dpto. Limay Mahuida, Puesto Garro, Steibel 86 (SI). Mendoza: Dpto. Luján de Cuyo, 100 km S of Me ndoza, Bocher & Hjerting 672 (Sl). Neuquén: Dpto. Catan Lil. tkm de Las Cortaderas hacia Charahuilla, Correa 7901 xen entre Viedma y San Antonio Deste, Correa 2457 (Sl). Li Río Negro: Dpto. A. Alsina, ruta 3, Km ex Hook.) 18. Junellia erinacea (Gillies & Hook. Moldenke. Lilloa 5: 396. 1940. Basónimo: Verbena erinacea Gillies & Hook. ex Hook., Bot. Misc. 1: 164. 1829. TIPO: Argentina. Mendoza: Portezuelo de Uspallata, “a low shrub with flowers growing near community along with G. C. h.", 29 ene. 1824, J. Gillies s.n. (lectotipo, designado por Botta (1989: 386), K!, K foto SI: isotipo, BM). Caméfitos pulvinados, cojines hemisféricos, espi- 7-12 cm. 10-20 X 1-1.5 mm, nosos, no compactos. Ramas de Hojas lineares, homomorfas, espiniformes, connadas en la base, margen piloso, híspidas en la cara abaxial y con escasos pelos casi adpresos en la adaxial, opuestas en macroblastos e imbricadas en braquiblastos. Monobotrios paucifloros, brácteas de 4-6 X 3- ovado-anchas, generalmente de 2 o 3 flores, 4 mm, connadas en la base, ápice subulado, rígido, con el margen piloso y la superficie 00.5 híspida y glandulosa. Cáliz de 5 5 mm, hispido, elanduloso, dientes subulados, espinescentes, e mayor de 1.5 mm, el adaxial muy breve; corola de 10.5-11.5 mm, lila o blanca; en el exterior glabro y piloso en el interior; fauce glabra o con escasos pelos y piloso en el interior; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 8-8.5 mm, base del DD [e estilo ensanchada. Clusas de 4.5-5 mm, glabras, ápice brevemente rostrado y cubierto por la base del — estilo, base angostada, cara dorsal reticulada, caras comisurales lisas. Distribución geográfica. Argentina en las provin- cias de La Rioja, San Juan, Mendoza y Río Negro; fitogeográficamente crece en el límite de la provincia del Monte, Patagónica, provincia Alto-andina, y la los 2500 y 3000 m, en suelos rocosos, generalmente sueltos (Fig. 1D). con la en general entre Fue (1967: 26) pero dado que no hay registros de herbario que avalen citada para San Luis, por Minoprio et al. esta distribución, no se ha considerado que la especie este presente en la provincia de San Luis. Fenología. Florece y fructifica en verano. Hooker, 1829: fig. 108. fig. 12 (sub Verbena erinacea), Botta, 48: Sanzin, 1919: 1989: 385 Iconografía. oO dg. 5. Observaciones. | Especie afin a Junellia caespitosa de la que se diferencia por tener las hojas homomorfas espiniformes y las clusas brevemente rostradas. Material adicional examinado. ARGEN T s A. ME nds ladrid, entre Lampaya v : TN ute ren, 10 (AUS. MERL). e Somuncurá, base del 2002. Bártoli s. BAA 24181 (SI). Cea, El Mendoza: 111 a Observatorio, 1 Juz p 6498 19. Junellia juniperina (Lag.) Moldenke, Lilloa 5: 396. 1940. Basónimo: Verbena juniperina Lag., Gen. Sp. Pl. 19. 1816. TIPO: Mendoza: San Carlos, camino a Laguna Diaman- te, 2000 m, 2 feb. 1950, A. 4057 (neotipo, aquí designado, SI!). Argentina. jasad Soriano Verbena juniperina var. grisea Johnst., Contr. de 81: 98. 1928 dur due. var. grisea ( . Johnst.) Molde RN Lilloa 5 . 1940. TIPO: Perú. E ‘quipa: ravines, 5 slope of s fed N of Arequipa, 3355 m, M . 1920, Hinckley 76 (holotipo, GH no visto: satis. NY 1 US no visto). Gray Arbustos erguidos o postrados. Macroblastos ergui- dos, a veces decumbentes, tallos con pubescencia generalmente híspida, variable en densidad. Hojas dimorfas, las de los macroblastos, de 4-5 X 2.5- 3 mm, profundamente 3-sectas, connadas, segmentos de 5-6.5 X | mm, angostamente elípticos o lineares, rígidos, pungentes, subiguales, nervio central promi- nente en la cara interior y margenes engrosados: hojas de los braquiblastos enteras, 2-lobadas o 2-3-sectas, más pequeñas que las anteriores, pubescencia con pelos simples y glandulares. Monobotrios « en espigas alargadas, de 2-3 em, plurifloras, de ores; brácteas de 5-7.5 X 1.5-2 mm, enteras, ovales, pungentes, pubescentes y con pelos largos y rígidos en el margen. Cáliz de 6.5-8 mm, dientes de 0.5-1 mm, subiguales, pungentes, pelos simples y glandulares: corola lila de 10-12.5 mm. glabra en el exterior, fauce elabra; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 6.5-8 mm, base del estilo ensanchada. Clusas de 3.5-4 mm, glabras, ápice truncado, a veces brevemente rostrado cubierto por la base del estilo, cara dorsal reticulada, castaño rojizo, caras comisurales casi planas. papilosas. blanquecinas. Cariología. n= al., 1988: 543). 10 (Junellia juniperina. Poggio et Annals of th Missouri Els Garden jon “Monte de la yegua” (Ruíz Leal, (de la Peña & Pensiero, Nombres vulgares. 1972: 168), “diablo fuerte” 2004: 68), “heliotropo de la sierra" (Sanzin, 1919: 113), "cola de león" (del ejemplar Meyer 8269, LIL). Usos. de la sangre. Según Roig (2001) es medicinal, depurativa de Perú, sur de Sur Chile Distribución | geográfica. Bolivia, centro de y noroeste de Argentina; en la provincia fitogeográfica Altoandina norte y y Puneña; crece entre los 1900 y 3600 m (Fig. 1D). Iconografía. Sanzin, 1919: 114, fig. 17; Ruíz Leal, 1972: 169, lám. 52, fig. 189; Botta, 1984: 340, fig. 4 (todos sub Verbena juniperina). Observaciones. Especie similar a Junellia bisul- cata de la que se diferencia por el ápice de las clusas rostrado y la pubescencia de los tallos adpresa. En este caso para neotipificar el taxón se eligió un nuevo ejemplar que reúne las características de la especie. Verbena juniperina Gillies & Hook. ex Hook. (Bot. Misc. 1. posterior y comprobó dicha especie es un sinónimo taxonómico. El ejemplar Kurtz 6788, citado por Moldenke (1940: 397) para la provincia de Córdoba, corresponde a Además Moldenke 829) es ilegítimo por homónimo de la observación del material tipo se Junellia bisulcata var. campestris. (1940: 397) expresa: "originally distributed as Verbena Juniperina. var. charoides Hieron.”; este nombre no se ha registrado en ninguna publicación, por lo tanto se debe considerar como nom. nud. Por áltimo, Moldenke (1980: 186) cita la especie para Chubut, sin mencionar ningún ejemplar; probablemente se trate de J. ulicina. adicional examinado. ARGENTINA. Cata- algala, along roadside betw. Andalgalá 31 (SD. Material & , Cantino 4. za Rioja: Dpto. Se ito, Quebrada Los Manzanos, Okada 2719 (SI). to. Las Heras, Caracoles de Villavicencio, mns 395 (SI). Salta: Dpto. Cafayate, de Salta a Cafayate, samerro s.n., SI 26244 (SI). San Juan: m Sarmiento, quebrada de 1 a Flecha y que *brad a el Río de los Sombreros, Kiesling 4160 (SI). Tucumán: o. Tafí, Cuesta del Infiernillo, Km 84, Burkart 22071 (S D. BOLIVIA. Tarija Aviles, Yunchara, Meyer 21561 (LIL). CHILE. I Región Arica lo tre, X-112 e can ariscal Nieto, ES 7 (U Til Hua elic 30268 (L Tacna: Tarata, camino a Caro, La Torre 1837 20. Junellia odonellii Moldenke, Phytologia aves 466. 1948. TIPO: Argentina. Santa Cruz, Gü Aike: Río Gallegos, s.f., C. A. O'Donell Pn (holotipo, NY! isotipos, P no visto, SI!, P foto Caméfitos pulvinados, tallos glabros o escasamente híspidos, macroblastos de 6-12 cm, erectos a decum- bentes. Hojas dimorfas, las de los macroblastos de 4— 7 mm, opuestas, subsésiles, profundamente 3-partidas, lóbulos lineales, con ápice agudo, punzante, los laterales algo más breves, ciliados en el margen, con pubescencia antrorsa; hojas de los braquiblastos tripartidas, carnosas, glabras, de menor tamaño que las de los macroblastos. Monobotrios de 3 a 10 flores en el ápice de braquiblastos más desarrollados; brácteas 3-partidas, similares a las hojas del macroblasto. Cáliz de 5-6 mm, híspida, rala; dientes triangulares, desiguales, breves, de ápice atenuado, 12-14 mm, blanca, pubescente-glandulosa en el exterior; tubo pubescente, en el interior, fauce glabra; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 8.5- 9 mm, base del estilo ensanchada. Clusas de 4 X 0.8 mm, glabras, ápice truncado y cubierto por la base =m de 4.5—7 mm, con pubescencia corola de rígido; del estilo, base angostada, cara dorsal reticulada. Distribución geográfica. Se la encuentra desde Neuquén a Santa Cruz, en Argentina y en el sur de Chile region XII, formando carpetas leñosas entre rocas, de la provincia fitogeográfica Patagónica (Fig. 1E). en cu mbres cerros; ern Fenología. Florece y fructifica en primavera y verano. Iconografía. Botta, 1999: 175, 180, fig. 131. odonellii (Moldenke) 974), esta combinación Observaciones. Verbena Trone. (Darwiniana 18: 31: realizada por Troncoso, es ilegítima por omitir el lugar de publicación del nombre original de acuerdo con el Art. 33.3 del Código Internacional de Nomenclatura Botánica (Greuter et al., 2000 Especie afín a Junellia ulicina de la que se diferencia por el mayor tamaño y el ápice acuminado de los dientes del cáliz, los monobotrios 2- o 3-floros y el mayor tamafio de la bráctea. ARGENTINA. Chubut: astillo, punto panorámico, js, Ea. Fortín Material adicional e Dpto. Escalante, Pam Cabrera 33253 (SI). Neu Chacabuco, co. e Bele 7 d Dpto. Güer Aike, ruta e Güer A peranza 12 km al W de Rio s. Botta 507 (SI). € Región: Magallanes, Estancia La Portada, e 1368 (CONC). Junellia seriphioides (Gillies € Hook. ex Hook.) Moldenke, Lilloa 5: 400. 1940. Basó- nimo: Verbena seriphivides Gillies € Hook. ex Hook., Bot. Misc. 1: 164. 1829. TIPO: Argentina. Mendoza: betw. Río Tunuyán & Hío Mendoza, 27-Nov. 5, 2]. also in S side of latter river, Oct. Volume 95, Number 2 2008 Peralta et al. Revisión del Género Junellia (Verbenaceae) 363 1824, J. Gillies s.n. (lectotipo, designado por Botta (1984: 334), K!, K foto SI!; isotipo, BM!, BM foto SI!). Figura 3H, I. Verbena echinata Phil., Anales Univ. Chile 36: 191. 1870. TI rgentina. Mendoza: Andes mendocinum, 1868— 1869, R Philippi s.n., SGO 42479 (lectotipo, designado por Acevedo de Vargas (1951: 65), SGO!; isotipos, K!. Pl). Bol. Acad. Ci. 1880. TIPO: mp si Cruz: a del Rio Santa a 874, C. 152 (holotipo, P no Verbena “Sriqueti F. Candollea 5: 402. Nombre o ee PE strix Briq., / Cosme Jard. Bot. Genéve 4: 15. 1900, non non V. hystrix Phil., Anales us ren 90: 610. 1 895. TIPO: Argentina. Mendoza: Pampa de San Rafael, s.f., E. Wilczek 45 (holotipo, G!, G foto FM 24687 SI! isotipos, K!, SI!, K foto SI!). Junellia seriphioides var. tomentosa Moldenke, Known Geogr. istrib. Verb. "ie 7. 1942. TIPO: Argentina. Mendoza: a, ac Dek 1 896. col. Bodenbender. herb. F. Kurtz 10012 (holotipo, NY no visto; isotipos, D no visto, SI!). Verbena seriphioides var. lanigera Hieron., órdo ils d 1934. Tinella s var. - glabra Moldenke, Phytologia 2: 409. 1948 : Argentina. Chubut: Puerto Madryn, 24 oct. oe P. O'Donell 3240 (holotipo, S no visto, S foto SI!; isotipos, NY no visto, SI!). Arbustos de 0.5—0.8 m alt. Macroblastos erectos, ramificados, a veces con raíces adventicias en la base, los más jóvenes pubescentes, al envejecer exfoliables en lonjas o láminas alargadas, braquiblastos de 2— 2.5 mm, muy ramificados. Hojas dimorfas, las de los macroblastos de 5-7 X 1—1.5 mm, enteras, rara vez 3- sectas (en este caso los lóbulos laterales apenas desarrollados), espiniformes, connadas en la base, cuando jóvenes pubescentes en ambas caras, hojas de los braquiblastos de 1 X l mm, enteras, carnosas, tetrásticas, imbricadas, lámina ovada, la mitad superior engrosada, bordes incurvos formando un surco tapizado por pelos suaves, adpresos y antrorsos, margen ciliolado. Monobotrios paucifloros, de 2 a : flores en el ápice de los braquiblastos; brácteas de 2.5 mm, ovadas, levemente pubescentes en la cara adaxial. Cáliz de 3.5—4 mm, externamente velloso (a veces glabro), y con pelos glandulares breves, hacia el ápice la pubescencia adpresa; corola de 10-12 mm, blanca o lila-pálido, glabra exteriormente; tubo recto, velloso en el interior; fauce glabra o con escasos pelos; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 4-5 mm, base del estilo ensanchada. Clusas de 3.5 mm, glabras, brevemente aladas, ápice truncado y cubierto por la base del estilo, cara dorsal reticulada, pardas, caras comisu- rales papilosas, blanquecinas. Cariología. n 10 Schnack & Covas, 1947: 5). (Junellia seriphioides, "T omillo macho” Leal, 1972 “rosita” (Cabrera, 1957: 398), (Moldenke & Moldenke, 1949: (Mufioz Pizarro, 1966: 150). Usos. En Salta refrescante cerebral para cuyo efecto se lavan la ”, del ejemplar BA 27/891. Nombre vulgar. (Paan, 1919: 87; Ruíz 7), “rosa de cordillera” “los nativos lo usan como cabeza con dicha infusión Distribución geográfica. Bolivia, Chile y Argen- tina, desde el noroeste hasta Patagonia. Crece en suelos arenosos y pedregosos en las provincias Puneña y Patagónica del Dominio Andino-patagónico hasta los 4200 m y en la provincia del Monte del Dominio Chaqueño (Fig. 2A). Fenología. Florece y fructifica en primavera y verano. Iconografía. Sanzin, 1919: 110, fig. 14 (sub Ver- bena seriphioides); Cabrera, 1957: fig. 1; Morello, 1958: 118, fig. 50; Roig, 1970: 143, lám. 70 (sub V. ser- iphioides); Ruíz Leal, 1972: 241, lám. 75, fig. 282 (sub V. seriphioides); Botta, 1984: 333, fig. 1 (sub V. seriphioides); Botta, 1993: 68, fig. 29; Botta, 1999: 183, fig. 134. Observaciones. Esta especie se caracteriza por sus ramas rígidas, sus hojas punzantes y sus inflorescen- cias paucifloras. Es muy afín a Junellia spissa, de la que se diferencia por la presencia de clusas no aladas. ARGENTINA. Cata- marca: Dpto. Antofagasta de la Sierra, Antofagasta de la Sierra, Ulibarri 747 (SI). Chubut: Dpto. Ameghino, 9 km N de Uzcudum, Ruta 31, Correa 10212 (BAB, SI). Jujuy: Dpto. Cochinoca, Mina Aguilar, camino a la escuela del Río grande de la Poma, Ancibor 31 (BAB, SI). Material adicional examinado. La Pampa: Dpto. Caleu Caleu, barrancas and flat upland above comisaría de Río a. Bartlett 19935 (Sl). La Rioja: Dpto. Gral. Sarmiento, quebrada del Río Peñón inziker 2128 (SD. Mendoza: Dpto. San Carlos, Bun La Jaula, cerros vecinos a la Agrupación de Gendarmería Nacional, Ambro- Herb. Ruíz Leal 34616 (MERL). Neuquén: Dpto. ia Fortunato 90 (BAB). Dpto. Avellaneda, La Japonesa (Gob. E Krapovickas 22393 (SI). Salta: Dpto. La Poma, ruta 40, ntonio de los Cobres, 5 km antes del vae con Jujuy. Cialdella 45 (SD). San Juan: Dpto. Angaco, Sierra de Pie de Palo, al Este de Mogote de los Corralitos, Kiesling 4774 (SI). San Luis: Dpto. La Capital, Estancia San Martín del Alto Negro, Anderson 1664 (SI). BOL IVIA. Oru Salinas de G. Mendoza hacia el W, vía Iswaya, Beck 11795 (SI). Pot tosi: Telamaya, Cárdenas 48 (SI). CHILE. I Región: Iquique, Quebrada de Chusmiza, Villagran 9072 (CONC Región: Antofagasta, Calama, Caspana, Olivares 9621 (SI). setti s.n., Confluencia, egro: 5 km de San ro: L. Cabrera, de — — == 22. Junellia spissa (Sandwith) Moldenke, Lilloa 5: 401. 1940. Basónimo: Verbena spissa Sandwith, Kew Bull. 1927: 187. 1927. TIPO: Argentina. Neuquén, Zapala: Zapala, co. Lotena, 14 Oct. 1925, H. F. Comber 78 (holotipo, K!, K foto SI!). 364 Annals of the Missouri Botanical Garden 30 em alt. ramificados desde la base. sinuosos, muy apretados: Caméfitos hasta Macroblastos muy braquiblastos de 2-2.5 mm, arrosetados. Hojas di- morfas. las de los macroblastos de 4-7 mm, lineares, espinosas, híspidas en ambas caras, las de los braquiblastos de | mm, tetrásicas, sésiles, angosta- mente ovadas, agudas, carnosas, con pelos cortos a antrorsos en ambas caras. Monobotrios de 2 flores en el ápice de los braquiblastos; brácteas ca. 2 X 3 mm, [EN anchamente ovadas, ápice agudo y rígido, connadas en la base, híspidas en la cara exterior, ciliadas en el margen. Cáliz de 3.54 mm, híspido, dientes ca. 0.5 mm, triangulares, no pungentes, agudos; corola de 8.5-9 mm, lila o blanca, tubo recto, glabra exte- riormente, pubescente interiormente; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 3.5- 4.5 mm, base del estilo ensanchada. Clusas de 3— 3.2 mm, glabras, ápice truncado y cubierto por la base caras comisurales del estilo, cara dorsal reticulada, papilosas, no aladas. Distribución geográfica. Ha sido coleccionada desde Neuquén y Chubut, en suelos aridísimos, ei la provincia filogeografica Patagónica (Fig. 20). | O to} t [o] Fenología. | Florece entre noviembre y febrero. Iconografía. Botta, 1999: 187, fig. 137. Observaciones. | Especie muy afín a Junellia con- natibracteata de la que se diferencia por el número de flores en la espiguilla, siempre más de cinco flores. A veces parece un ejemplar muy achaparrado de J. seriphioides de la que se distingue por tener el fruto alado. ARGENTINA. Chubut: Dpto. Florentino Ameghino, Ameghino, Ruiz Leal 25670 (S1). Neuquén: Dpto. Picunches, bajada del Aorio, Vallerini 3388 (SI); Dpto. Zapala, Ruta 22, Material adicional examinado. camino Neuquén=Zapala, bosque petri- ficado. Botta s.n. (Sl 20074). 23. Junellia toninii (Kuntze) Moldenke. 23a. Junellia toninii (Kuntze) Moldenke var. toninii, Phytologia 21: 253. 1971. Basónimo: Verbena toninil Kuntze, Revis. Gen. PL. 3(2): 258. 1898. TIPO: 1882-1884. T. Moreno & Tonini s.n. (holotipo. NY!: isotipos. LP! SIB. Figura 4A-D. Argentina. Patagonia, Arbustos bajos, intrincados hasta 40 em de alt., « camefitos pulvinados hemisféricos ca. 15 em alt., muy espinosos. Ramas erectas, densamente foliosas, de 10-20 cm. ( 'spiniforme s, gene ‘ralmente 7 los macroblastos 4—6(—-10) xX l- lóbulos lineares, Hojas dimorfas, las de enteras, 1.5 mm, lineares, a veces trisectas. lóbulo medio ca. 5.5 X 1.5 mm. lóbulos laterales de 5X ] mm, o ambas formas en la misma rama. todas sésiles, agudas, ciliadas: hojas de los braquiblastos 1.5-2 X 0.5- das, l mm, enteras, lineares, sésiles, cilia- carnosas 0 mem! yranosas. ápice no punzante Monobotrios en espigas de 3, 5 hasta 10 flores, en el ápice de macroblastos; brácteas de 3-6 X 1.5 mm, ovadas o angostamente ovadas, agudas o acuminadas, pilosas. Cáliz de 6.5-7 mm, pubérulo a pubescente, dientes de 1.5 mm, el adaxial muy breve, ciliado en los bordes; corola de 10—16.5 mm, blanca, azulada o lila. glabra exteriormente, pubescente interiormente en la mitad basal, fauce glabra: par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las del 4 mm, glabras, ápice truncado y cubierto por la base del comisurales blancas, papilosas. tecas: gineceo de 4.5 mm. base estilo ensanchada. Clusas de estilo, cara dorsal reticulado-estriada, caras Distribución geográfica. Desde sur de Mendoza. hasta Santa Cruz, desde el nivel del mar hasta la serranías de la precordillera, en la provincia fitogeo- gráfica Patagónica (Fig. 2D) Fenología. | Florece y fructifica en verano. Iconografía. Botta, 1999: 191, fig. 140. Observaciones. Se diferencia de la variedad mu- linoides por la pubescencia del cáliz. Además es afín a Junellia connatibracteata de la que se diferencia por as brácteas connadas y las hojas de los macroblastos recurvas. ARGENTINA. Chubut: 31 km S de Garayalde, Material adicional Maie Dpto. Escalante, ruta Nac. 464 (SI). Mendoza: Dpto. VENE camino a Valenciana, Cabrera 33169 (SI). Neuquén: Dpto. Catán Lil, 15 km de Las Cortaderas hacia Charahuilla, Corea 7211 (SI). Río Negro: Dpto. Norquinco, Choique, Cabrera 33043 (SI). 281, 52 km al W de Santa Cruz: Dpto. Deseado, ruta Prov. Puerto Deseado, Botta 178 (Sl) 23b. Junellia toninii var. mulinoides (Speg.) P. . Basónimo: Verbena Anales Soc. Ci. Argent. 53—54: 1902. Junellia mulinoides (Speg.) Moldenke, Distrib. Verb. “In aridis '. mayo 1899, Lahille s.n. (lectotipo, desig- Botta (1989: 388). LP). Peralta & Múlgura, comb. nov mulinoides Speg., 24T. known Geogr. Avicenn.: 77. 1942 TIPO: Porfirio’ Argentina. maritimis Caleta nado por IE F. Figura Se diferencia de la variedad típica, por presentar el cáliz glabro. apenas ciliado en los bordes y las brácteas elabras. o Distribución geográfica. Se la encuentra en terre- nos áridos y laderas pedregosas desde Neuquén a Santa Cruz, en la provincia fitogeográfica Patagónica (Fig. 2D). DM Volume 95, Number 2 2008 Peralta et al. Revisión del Género Junellia (Verbenaceae) Iconografía. Botta, 1999: 177, fig. 130 sub Junellia mulinoides. Fenología. Florece y fructifica en verano. Observaciones. El lectotipo de Verbena mulinoides fue designado por Botta (1989: 388), como se expresa más arriba, pero en los manuscritos posteriores de la autora aparece el ejemplar de Illín, con anotaciones que consideran al mismo como lectotipo. Probable- mente esto llevó a confusión en la citación del tipo en la Flora Patagónica (Botta, 1999: 175; “Chubut. dpto. Futaleufú, N. Hlín s.n., Mar. 1900, Holotypus LP 10941"). Por otra parte, en la etiqueta original del citado ejemplar, dice: Chubut: Carreuleufú, Patag.. N. Illin, Mar. 1900, LPS 10401. Acerca de los sinónimos de esta variedad v - Junellia connatibracteata. Material adicional examinado. ARGENTINA. Chubut: Dpto. Biedma, 8 km S de Madryn, ruta 1, Correa 10160 (S1). Neuquén: Dpto. Añe slo, Sierra Auca CTUM Fabris 870 (SI). Río Negro: Dpto. San Antonio, Camino a Las Grutas, desvío al mar, Botta 406 (SI). Santa Cruz: Dpto. Corpen 1 70 km S Gobernador Gregores, camino a La Julia, Boelcke 16255 (SI). 24. Junellia ulicina (Phil. Moldenke, Lilloa 5: 402. 1940. Basónimo: Verbena ulicina Phil.. Anales Univ. Chile 90: 611. 1895. TIPO: Chile. Colchagua: Valle Hermoso, ene. 1872, s. col., SGO 54682 (lectotipo, designado por Botta (1989: 386), SGO! SGO foto SI!). Verbena ameghinoi Speg., Revista Fac. Agron. Univ. Nac. La Plata 3: 61. 1897. Junellia ameghinoi (Speg.) Mol- denke, Lilloa 5: 392. 1940. TIPO: Argentina. Santa Cruz: golfo de San Jorge, feb. 1896, C. Ameghino s.n. (holotipo, LP! LP foto SI). Verbena ra cM Annuaire Conserv. Jard. Bot. Genève 4: 900. Junellia intrincata (Brig Moldenke, a E 416. 1942. TIPO: Argentina. Mendoza, San Rafael: Río Tordillos, mina de las Choicas, feb. 1897, E. Wilezec 48 (holotipo, LAU no visto; isotipos, K!, P!, SIL K foto SII, G foto FM 24689 SIN: — Caméfitos pulvinados, pubescentes, ramas glabras o escasamente híspidas. Hojas dimorfas, las de los macroblastos tripartidas, espiniformes, de 5-10 mm (las basales menores que las superiores), lóbulo medio 2 mm mayor que los laterales, escabrosas; hojas de los braquiblastos no espiniformes, tripartidas, de 2 mm, carnosas, lóbulo medio mayor, superficie adaxial escabrosa. Monobotrios en espigas, de 2 a 3 flores, brácteas tripartidas, de 11.5-12 mm, similares a las hojas de los macroblastos. Cáliz de 6-7 mm, híspido- elanduloso, dientes ca. 1.5 mm, con ápice acuminado, rígido; corola de 12-15 mm, tubo recto, exteriormente elanduloso; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de as lecas; gineceo de 12 mm, base del estilo ensanchada. Clusas de 4.5 mm, subcilíndricas, ápice obtuso y cubierto por la base del estilo. cara dorsal reticulada en la mitad superior, caras comisurales suavemente papilosas. Distribución geográfica. Chile y Argentina, al sur del paralelo 32%, en las provincias de Mendoza, Neuquén, Chubut, Santa Cruz, desde el nivel del mar hasta la zona cordillerana a 1750 m (Fig. 2F). Fenología. | Florece en verano. Iconografía. Botta, 1999: 199, fig. 143. in a Junellia odonellit: = Observaciones. Es muy a ambas especies son cojines, espiniformes, presen- tando las brácteas florales tripartidas. Junellia odonellii tiene monobotrios de 3- a 10-floros, brácteas menores y dientes del cáliz de 0.5 mm, atenuados. Material adicional examinado. ARGENTINA. Chubut: Dpto. Escalante, ruta Nac. 3, 76 km al S de Garayalde, Botta 165 (SI). Me P Dpto. Malargüe, Valle del Salado, Alto de Los c aracoles, camino del Sol, Km 13-14, A. A. Lagiglia 7 : É rquín, Cerro Huaile, ladera SE, Biganzoli 1198 (SI). Sania Gi Dpto. Magallanes, Tres Cerros, Ea. La Lomita, pie de > cerro, cerca de 3 km al E, ruta 3. Boelcke 15368 (SI). CHILE. VI Región: Valle Hermoso, s. col., SGO 54682 (SGO, ai ae do = co 25 = > e - E 2 a = Junellia subgén. Thryothamnus Botta, Darwiniana 29: 390. 1989. TIPO: Thryothamnus junciformis Phil. Arbustos de desarrollo variable, con hojas normales o reducidas; monobotrios o dibotrios plurifloros en espigas contraídas, cilíndricas; fauce de la corola glabra o pubescente; conectivo superando o no la longitud de las tecas, con o sin apéndices conecti- vales; base del estilo no ensanchada y más o menos hundida entre las clusas. Número cromosómico x = Junellia subgén. Thryothamnus secc. Thryotham- nus Botta, Darwiniana 29: 391. 1989 Verbena secc. Verbenaca ser. Junciformis Briq., p.p. in Engler & Prantl, Nat. Pflanzenfam. 4(3A): 147. 1895. Verbena secc. Junellia ser. UNUS (Briq.) Tronc., Darwiniana 18: 31: Arbustos con ramas de sección poligonal o circular; hojas desarrolladas, enteras, no espiniformes; mono- botrios o dibotrios plurifloros, en espigas contraídas; fauce de la corola glabra, estambres con conectivo de menor longitud que las tecas y sin apéndices conectivales. En este caso las especies consideradas en esta sección coinciden con el tratamiento dado por Botta (1989: 391). 366 Annals of the Missouri Botanical Garden 25. Junellia lavandulaefolia (Phil) Moldenke, 1910. TIPO: [Chile.] Coquimbo: Arqueros, in Lilloa 5: 397. 1940. Basónimo: Verbena lavan- collibus, dic. 1836, C. Gay 1081, SGO 54710 Chile 43: 521. 1872, designado por . SGO!, SGO foto dulaefolia Phil., Anales Univ. 873. TIPO: Chile. Las Damas, Río Teno, s. col., SGO 54719 ( pr Acevedo de Vargas (1951: SI; isotipo, SIT). Figura 7 521. var. Phil., Anales Univ. Chile 43: 1873, sy Verbena hands Phil colchaguensis (Phil.) Reiche, Fl. Chile 5: 281. 1910. Are CURA PME an var. Ne (Phil.) Moldenke, 235. 1946. TIPO: Chile. Valle de pa n 1872, s. SGO 54708 (holotipo, O! SGO!). | e rbe na cole haguensis | syn. nov. -hytologia 2: col., Arbustos glabros, ramas erectas, estriadas de sección circular, entrenudos de 2-3 cm, ee sin desarrollo de braquiblastos. Hojas de 5-15 1.5 mm, generalmente todas enteras, a veces aus hojas trifidas, lineares, sésiles, margen revoluto, base persistente, pubescencia ensanchada y a veces escabrosa escasa, pelos adpresos. Monobotrios en espigas, plurifloras, alargados a la madurez ca. 2 cm: 3-5 X 2 mm, angostamente elípticas, acuminadas, márgenes mem- brácteas de enteras, todas similares, branosos y ciliados, cara abaxial escabrosa. Cáliz de 5-6 mm, dientes triangulares de 1 mm, híspidos en el »]2 mm, exteriormente con exterior; corola de pubescencia adpresa, fauce glabra; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 5 mm, Fruto no visto. base del estilo no ensanchada. Distribución geográfica. Chile en las regiones IV, V, VI, provincia Región Metropolitana, en Central. (Fig. a y VII y la fitogeográfica Chilena Se ha coleccionado a 1900 m. Fenología. Florece en verano. Junellia pseudojuncea por su porte, diferenciándose principal- Observaciones. Especie muy similar a mente por los entrenudos mayores y hojas apenas desarrolladas y brácteas lineares. à variedad se ha sinonimizado porque se pudo observar la presencia de hojas enteras y algunas divididas en un mismo ejemplar de herbario. .HILE. Largo > de Las Condes, Material adicional examinado. Región Metro- cordillera de Santiag go, Valle llo s.n. (CONC). VI Región del Flaco, Montero 9627 (CONC). Región: Curico. Teno. Zöllner 5183 (CX Valle superior Río 26. Junellia pseudojuncea (Gay) Moldenke, Lil- loa 8(2): 418. 1942. Basónimo: Verbena pseudojun- Fl. Chile 5: 19. 1876. Verbena spathulata Reiche, Fl. Chile 5: 280. cea Gay, var. pseudojuncea (Gay) (lectotipo, designado por Acevedo de Vargas (1951: 60), SGO!). Figura 7A-D. Anales Univ. Chile 90: 618. (Phil.) Reiche, Fl. Chile 5: 281. 1910. TIPO: Ch s e jare Torca, in dpto. Ovalle”, 188€ uk . Gei 42533 designado por (ub a can (1951: Thryothamnus p Phil., 1895 na Rp (lectotipo, )! 60), SGO!). Arbustos de 1—1.5 m de alt., tallos aparentemente áfilos, de sección circular, entrenudos hasta de 10 em, generalmente sin desarrollo de braquiblastos. 5-10 1-2 mm, lineares, enteras, a veces tripartidas, glabras. Mono- Hojas de reducidas, sésiles, espigas contraídas, 3—4 cm, X ] mm, lineares, agudas, botrios o dibotrios plurifloros, alargadas en la madurez hasta flores — perfumadas; brácteas de 5 pubescentes en el dorso, todas similares entre sí, iguales o apenas menores que la longitud del cáliz. 4—4.5 mm, cilíndrico, híspido de 8-11 mm, Cáliz de dientes de | mm, triangulares; corola violáceo clara, escasos pelos en la parte superior externa del tubo y en los lóbulos, fauce glabra, pelos retrorsos en ‘| interior del tubo; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 5.5-6 mm, base del estilo no ensanchada. Clusas de 5-5.5 mm, glabras, cara dorsal reticulada, caras comisurales papilosas. Distribución geográfica. Crece en Chile en las entre los 2000 y 2500 m, campos pedregosos. Fitogeográficamente habita en las Regiones HI y IV, en provincias del Desierto, distritos Coquimbano, de los Cardonales y Central (Fig. 1E). Botta, 1989: 387, fig. 6. Iconografía. Observaciones. Este taxón es similar a Junellia spathulata; se diferencia principalmente en as brácteas florales ovales. Se asemeja también a J. muy lavandulaefolia, ver observaciones en esa especie. El ejemplar C. Gay 1081 había sido nombrado por Acevedo de Vargas (1951: 60), como *Cotypus"; dado que la autora no hace referencia a otro ejemplar como tipo de la especie, se considera al mismo como lectotipo. Material CHILE. HI Región: Huasco, Km Arancio d (CONC). gión: i. ee de trica Los . Bocatoma, 5 (SI). Mol Ricardi 755 (CONC 27788); río Flamenca, pon 2905 a Coen 9 río del | cho, 27. Junellia selaginoides (Kunth ex Walp.) Mol- denke 27a. Junellia selaginoides (Kunth ex Walp.) Moldenke var. selaginoides, Prelim. Alph. List Inval. Incor. Sci. Names Verben. Avisen.: 48. 1940. Basónimo: Verbena selaginoides Kunth ex Volume 95, Number 2 2008 Peralta et al. 367 Revisión p Género Junellia (Verbenaceae) Figura 7. e Ricardi 753, CONC 27788; D Javandulaefalia o :encia (de Grandjot . a F-I. Junellia aspera var. a m Aspecto general. —G. Corola abierta. Ant tomado de Botta, 1 —H. Anteras. —I. Fruto. F-I de Boelcke 1567. 3. J-L. Junellia tridens. — A-D. me pseudojuncea. —A. Aspecto general. —B. Ar a an —C. Corola abierta. —D. Antera. A-D 989: 387, fig. 6, E. Junellia n el permiso de F. Zuloaga. to general. —K, K'. Hojas. —L. eras. J-L de Pew 3388; L tomado de Botta, 1999: 191, fig. 141, con el permiso m n Zuloaga. Walp., Repert. Bot. Syst. 4: 15. 1845. TIPO: Chile. Coquimbo: 1839, C. Gay s.n. (lectotipo, G!, G foto FM 7856 SI?) aquí designado, Figura 8A-D. Verbena e Turez., Bull. Soc. Imp. Naturalistes u 36(2): 194. 1863. TIPO: Chile. Coquimbo: 7. Bridges 1353 (holotipo, K!, K foto SI!; isotipos, BM!, SI). Arbustos de 0.5—0.8 m de alt.; de 0.5-2.5 em, ramas erguidas, entrenudos estriados, de sección poligonal, con pubescencia de pelos suaves, en general adpresos, retrorsos, con braquiblastos desa- de 5-15(-17) xX 1- 1.5 mm, subopuestas, sésiles, lineares, subcarnosas, rrollados. Hojas homomorfas, margen revoluto, cara adaxial con pubescencia escabrosa, cara abaxial con pubescencia híspido- glandulosa. Monobotrios plurifloros en espigas contra- alargadas a la madurez hasta de 1—3.5 em; brácteas de 7-9 X 1.5- el margen apenas revoluto, ídas, 2.5 mm, ovado-angostas, con ciliado, pubescencia Annals of the Missouri Botanical Garden ABESTARD Figura 8. A-D. Junellia selaginoides var. selaginoides. —A. Aspecto general. —B. Flo A BESTARD r. —C. Corola abierta. —D. Anteras. A-D de CONC 30923. E-H. Junellia selaginoides var. illapelina. —E. Aspecto general. —F. Flor. —G. Corola 20. abierta. —H. Fruto. E-H de Goodspeed 16512 semejante a la de las hojas, con pelos glandulares más abundantes. Flores perfumadas. Cáliz de 5-5.5 mm, con dientes subulados, en general mayores de 1 mm, hispido sobre los dientes y costillas, más adpresos en el resto de la superficie, con abundantes pelos elandulares; corola de 7.5-9.5 mm, pubescente ex- teriormente con escasos pelos glandulares, fauce glabra; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las tecas; gineceo de 5.5-7 mm, base del estilo no ensanchada y más o menos hundida entre las clusas. Clusas de de 2.5-3 mm, ápice obtuso, cara dorsal reticulada, pardo-obscura, caras comisurales papilosas. Distribución geográfica. Chile, en las Regiones Il y IV. crece en la zona litoral entre los 25 S y 305 y entre los 50 y 150(500) m; en la provincia fitogeográfica del Desierto, distrito Coquimbano (Fig. Las restantes especies de la sec- la falta Observaciones. ción se diferencian por de desarrollo « braquiblastos. Volume 95, Number 2 2008 Peralta et al. Revisión del Género Junellia (Verbenaceae) 369 Figura 9, Junellia inflorescencia i = ijo 1713). — mn var. spathulata. As] 0. Junellia spathulata var. Detalle de la inflorescencia En la descripción original, Walpers describe la especie refiriendo el nombre a un binomio usado por Kunth en mss. Para la lectotipificación se eligió un i ejemplar que lleva una etiqueta manuscrita por Kunth y que coincide con la descripción original. En el K, figura la localidad Concepción, lo que probablemente ejemplar Bridges 1353 depositado en sea un error en la etiqueta (Botta mss.). e CHILE. H Región: Antofagasta, Cerro l, Grandjot 4456 a (SI). IV Región: C be aie. Hacie ns “Talinay. Garaventa 4459 (SI). Material adicional 27b. Junellia selaginoides var. illapelina (Phil.) Botta, Hickenia 2: 128. 1995. Basónimo: Ver- bena illapelina Phil., Anales Univ. Chile 90: 612. 1895. Verbena selaginoides var. illapelina (Phil.) Reiche, Fl. Chile 5: 288. 1910. Junellia illapelina (Phil) Moldenke, Phytologia 2: 51. 1941. Chile. Cordillera de Hlapel, La Polcura, ene. 1888, s. col., SGO 54727 (lectotipo, designado por Botta (1989: 391), SGO! SGO foto SI!; isotipo, PP). Figura 8E-H. Difiere de la variedad típica por el tamaño de las hojas: de 1.5-5 mm en los macroblastos y de 1.5-2 mm en los braquiblastos; el cáliz con dientes triangulares y agudos de 0.6-1 mm y escasos pelos glandulares. —A. p c (OE E ; "ew 2 dE a LEO (AO A j G A N AMIA SUS C ESOS. PE Af i An BS NRA AI tdi SAS Ae T E PUE Y pe b NAHE, s AGAN VAR - E: WM EE jj Ed 27 AAA: Lee nts CEE PERT ae RE Ad. EL a>. MS Ve RUE E Em P NY m SEO AA NES 2 Er =. TUS SE uem m NEST A i AGS ore um E Y na me PR Aine: TI ot er ZO A SRM S N Loe $ UNA ES de la alle Ruíz Leal 20835). Cabrera 33382). —B. Det (de ecto general (de — Chile, en la Región IIL y IV, crece en cerros y quebradas entre los 600 y Distribución geográfica. 1500 m; en la provincia fitogeográfica del Desierto, distrito Coquimbano (Fig. 11). examinado. CHILE. HI Región: Los Cristales, Quebrada Las Salinas. INC). IV Region Coquimbo. Linari. d adicional iayco. O. Mineral, Marc 1723 (CC >», Jiles 1919 (SI). 28. Junellia spathulata (Gillies & Hook. ex Hook.) Moldenke. 28a. Junellia spathulata (Gillies & Hook. ex Hook.) Moldenke var. spathulata, Lilloa 5: 401. 1940. Basónimo: Verbena spathulata Gillies & Hook. ex Hook.. Bot. Mise. 1: 162. 1829. TIPO: Argentina. Mendoza: en valle de las Leñas Amarillas, in E ascent to El Planchon, s.f., J. Gillies s.n. (lectotipo, designado por Botta (1989: 392), K!, K foto SH; isotipo, SI!). Figura 9A, B. as Hook. ex Hook. var. parviflora Prodr. 11: 543. 1847. Wu snp Miers, Pomi Linn. Soc. London Bot. 105. TIPO: Chi Andes of Chile, T. BE 1220 a aquí designado, K!, K foto SI!: isotipo, BM!) Verbena spathulata Gillie thauer in DC., alt.; tallos erectos, glabros, estriados, aparentemente áfilos, sin Arbustos de 70 cm sección circular, Annals of the Missouri Botanical Garden Hojas de 2-20 X 2- 4 mm, enteras, lineares, sésiles, margen algo revoluto, desarrollo de braquiblastos. carnosas, generalmente glabras, a veces estrigosas en ambas superficies. Monobotrios plurifloros en espigas contraídas densas, alargadas en la madurez hasta brácteas de 2-3 X 0.5- ovadas, ápice agudo y rígido, siempre más cortas que 4 cm, 0.6 mm, angostamente el cáliz y todas similares, margen no membranáceo. Cáliz de 5-6.5 mm, glabro en la base, híspido hacia el ápice, con dientes desiguales, agudos, pubescentes exteriormente; corola lila, de 9 mm, exteriormente con pelos retrorsos, tubo recto, glabro, fauce glabra, limbo patente; par superior de estambres sin apéndices conectivales, conectivo no superando la longitud de las base del estilo no ensanchada. Clusas de 4 X 1.2 mm, glabras, con alas tecas; gineceo de 6.5 mm, laterales pequeñas, cara dorsal reticulada, caras comisurales papilosas. Distribución geográfica. Crece en la región an- dina de Argentina, sur de Mendoza y Neuquén, y en Chile, en las regiones Metropolitana, Vl, y VII. Habita entre los 1900 y 2200 m, en la provincia fitogeo- gráfica — Altoandina, distrito — Altoandino-cuyano (Fig. 2B). Fenología. Florece en verano especialmente en los meses de enero y febrero. Iconografía. Botta, 1999: 183, fig. 136. Observaciones. Tanto la variedad típica como la variedad glauca, se diferencian principalmente de Junellia lavandulaefolia y de J. pseudojuncea, por la morfología de las brácteas. El análisis de EP Md de Verbena spathulata var. parvifolia, en el herbario K, fue realizado por Botta, quien en sus manuscritos dejó indicado el citado ejemplar como lectotipo, ya que las caracterís- ticas del mismo se corresponden con la descripción original. Por otra parte el mismo ejemplar fue citado como tipo de Diostea stenophyla en la descripción original. ARGENTINA. Men- 60 km al N de 33492 (SI). Neuquén: Dpto. Norquinco, Trolope, 121 (SD. CHILE. egión olitana: Santiago, Refugio Lo Valdes, Valle Maipú, Garaventa 6286 (CONC). VI Región: Vegas del » a la mina, Ricardi 3246 (SI). VIL Región: Volcán Peteroa, Werderman 577 (BM, SI). Material adic tonal examinado. Río Grande, Las Sanso Flaco, caminc Cord. Curicó, — 28b. Junellia spathulata var. glauca (Gillies & Hook. ex Hook.) Botta, Hickenia 2: 128. 1995. Basónimo: Verbena glauca Gillies € Hook. ex Hook., Bot. Misc. 1: 163. 1829. Junellia glauca (Gillies & Hook. ex Hook.) Moldenke, Lilloa : 396. 1940. TIPO: Argentina. Mendoza: near the River Diamante, J. Gillies s.n. (holotipo, K!, K foto SI!). Figura 9C. Verbena spathulata var. aes igs Se hauer en DC., Prodr. 11: 543. 1847, syn. nov. Verbena ochreata TRN 3. 19( Annuaire Conserv. Jard. 2 Genève 4: 8. JO. Junellia spathulata var. pe edo is hauer) Botta, Hicker 28. 1995. ule. Gay sn. (lec pcd aquí moms e ee SI”. de e Miers, Trans. Linn. Soc. London Bot. 27: 869. TIPO: Chile. Cordillera de Santiago AM s.n. (lectotipo, aquí designado, K!; isotipos, BM!, P!). Verbena glauca Gillies € Hook. ex Hook. var. n Niederl. en Lorentz & Negro Bk ndina (Niederl.) Moldenke, Lilloa 5: frecuente en el TIPO: capa Nenq uén: * embocadura Viederlein s.n. (holotipo, CORD!). Se diferencia de la variedad típica por las brácteas basales de 3-6 X 3—4 mm, ovadas u ovado-oblongas, margen membranáceo o no, ciliado, ápice agudo a punzante; brácteas superiores de 4—6 X 2-3.5 mm, mayores que la mitad de la longitud del cáliz, ovadas o angostamente ovadas, híspidas hacia la base, ápice mucronado, rígido u obtuso margen ciliado. Cáliz de 4.5—6 mm. Cartología. 2n = 20 (sub Junellia glauca, Covas & Schnack, 1947: 225, 229, fig. 30). n = 10 (Poggio et al., 1988: fig. LA). 2n = 20 (Botta & Brandham, 1993: 146, fig. 1A). Precordillera del S y W regiones Distribución. geográfica. de Mendoza y Neuquén, centro de Chile, Metropolitana, IV, V fitogeográficas del Monte y Altoandina (Fig. 2B). Vll; en las provincias enero. tab. XII, Fenología. Florece entre noviembre y Iconografía. Lorentz y Niederlein, 1881: Il, fig. 6 (sub Verbena glauca var. cisandina). Observaciones. La variedad Verbena spathulata var. grandiflora, fue lectotipificada en base a las observaciones realizadas por Botta, quien en sus manuscritos dejó indicado el citado ejemplar como lectotipo y se pudieron corroborar sus observaciones. En el caso de Diostea filifolia, el material citado en la descripción original, fue también analizado por Botta, quien en sus manuscritos dejó asentado la elección [em como lectotipo y que dicho material se corresponde a Junellia spathulata var. grandiflora. aterial adicional examinado. ARGENTINA. Men- alargüe, Los Ranquiles, Martínez Crovetto E E mo E H [ o [a] Sa S po Bo ae o p ~ Us NN SI). CHILE. Región Metropolitana: Santiago, Los Condes, Gillies s.n.. Volume 95, Number 2 2008 Peralta et al. 371 Revisión d Género Junellia (Verbenaceae) LP 8470, 8469 (LP) IV Región: Coquimbo, Illapel, i la Vega escondida, Goodspeed 16594 (SI). V ión: Los Andes, Los Maitenes, Río Colorado, Zóllner 6528 (CONC). VIL Región: Curicó, Laguna del Planchon, Zóllner 5991 (CONC). Junellia subgén. Thryothamnus secc. Dentium P. Peralta € Múlgura, sect. nov. Junellia subgén. Thryothamnus secc. Junelliopsis Botta, syn. nov. Darwiniana 29: 392. 1989, p.p., tipo. TIPO: Verbena tridens Lag. [= tridens (Lag.) Moldenke]. excluyendo el Junellia Arbustos altos, ramas de sección poligonal o subcircular. Hojas dimorfas: las de los macroblastos, enteras o tripartidas, espiniformes; las de los braquiblastos con lámina desarrollada, enteras, car- osas, no espiniformes. Monobotrios en espigas cilíndricas, alargadas o contraídas. Corola con fauce densamente pubescente, pelos diferentes a los del interior del tubo corolino; estambres con conectivo generalmente de mayor longitud que las tecas y sin apéndices conectivales. Observaciones. Las especies de esta sección in- cluyen todas las especies de Junellia subgén. Thrytham- nus secc. Junelliopsis sensu Botta, excluido el tipo, J. bryoides que se transfirió al subgénero Junellia. Junellia tridens se designa como nueva especie tipo por presentar los caracteres típicos de la sección. 29. Junellia asparagoides (Gillies & Hook. ex Hook.) Moldenke, Lilloa 5: 395. 1940. Basónimo: Verbena asparagoides Gillies & Hook. ex Hook., Bot. Misc. 1: 165. 1829. TIPO: “near the upper.. Argentina. Mendoza: .from Portezuelo to the valley of Uspallata”, s.f., Gillies s.n. (lectotipo, aquí designado, K!, K foto SI!). Verbena diversifolia Kuntze, Revis. Gen. Pl. 3(2): 255. 1898. Junellia diversifolia (Kuntze) Moldenke, Lilloa 5: 395. 40. TIPO: Argentina. Mendoza: Paso Cruz, 2000 m, 5 (holotipo, NY!). = s.f., Kuntze 5 Arbustos de 1-1.5 m de alto, con ramas erguidas, flexibles, pubescentes hacia el ápice; entrenudos de 0.5-1 cm. Hojas dimorfas, las de los macroblastos de (2-)3-4(-8) mm, central de mayor longitud que los laterales; hojas de tripartidas, espiniformes, lóbulo 1-1.5(-3) mm, enteras, los braquiblastos 9(-13) subcrasas, obovadas a linear-oblongas, ápice obtuso a base veces subobtuso, con un mucrón breve y cuneada, superficie escabrosa. Monobotrios en espi- gas plurifloras, de 3 cm; brácteas de 2.5-8(-9) mm, subconnadas, ápice agudo a ovadas o lineares, punzante, a veces acuminadas, margen ciliado. Cáliz tubuloso, 5-dentado, con tubo de 3—4.5 mm, marca- damente 5-costado; dientes de 0.5—3 mm, subiguales, rígidos y algo recurvos en el ápice, subulados, pubescencia de pelos cortos, esparcidos, blanquecinos, especialmente entre las costillas; corola lila, de 8.5— 12.5 mm, con pelos suaves y cortos, más o menos erectos, adpresos, desde la mitad superior hasta la base de los lóbulos; fauce con pelos claviformes; par superior de estambres sin apéndices conectivales, conectivo de igual longitud que las tecas: gineceo de 56.6 mm, base de Clusas de 2.8-3 mm, ápice obtuso, angostándose hacia estilo no ensanchada, ápice glabro. — a base, cara dorsal convexa, castaño oscura, estriada, caras comisurales papilosas, blanquecinas. Cariología. 2n = 20 (Junellia asparagoides, Covas € Schnack, 1946: 156, 159, fig. 2G). Nombre vulgar. “Barba de viejo” (Ruíz Leal, 1972: 38). Distribución geográfica. Habita en Argentina en — as provincias de Catamarca, Jujuy, La Rioja, Salta, las regiones IH y IV. Crece entre los 2000 y 2800 m Mendoza, San Juan y Tucumán y en Chile en en la provincia Prepuneña y asciende por las quebradas llegando hasta la Puna (Fig. LA). Fenología. Florece en verano desde diciembre hasta febrero. Iconografía. | Sanzin, 1919: 111, fig. 15 (flor sub Verbena asparagoides); Ruíz Leal, 1972: 37, lam. 9, fig. 27 (sub Verbena asparagoides); Botta, 1984: 337, fig. 2 (sub Verbena asparagoides): Botta, 1993: 7( fig. 30. Observaciones. La especie se caracteriza por las hojas de los macroblastos espiniformes tripartidas y los dientes del cáliz subulados y rígidos. En Junellia cedroides, los dientes son triangulares, aunque ambas especies presentan hojas dimorfas. Los ejemplares de Gillies, depositados en K, fueron estudiados por Botta, quien eligió el ejemplar citado como lectotipo, porque tiene los caracteres propios de la especie y la etiqueta se corresponde en parte con la descripción original: “Hab. Uspallata apud Cerro del Portezuelo: alt. 10,000 ped". In summum fere jugum Material — avidi ARGENTINA. Cata- marca: Dpto. Belén, Cuesta de Randolfo, Cabrera 32449 (SI). Jujuy: Des ee Quebrada de Huasamayo, Cabrera 31563 (SI). La Rioja: ae Famatina, Sierra de Famatina, entre Los Corrales y Los Berros, a 3.5 km de la primera, ee 7091 (Sl). Mendoza: Dpto. Las He ras, e de Jspallata, J. H. Hunziker 3071 (SI). Salta: Dpto. Cachi. desde ruta 40, camino dud rno hacia Las Paylas, Cialdella 264 (SI). San Juan: Angaco, Sierra de Pie de Palo, camino a Mogote ie hs Corralitos, Kiesling 3108 (SI). Tucuman: Dpto. e La Cienaguita, Cumbres Calchaquíes, X). CHILE. III Copiapó. Que- rancio 657 (CONC). Región: Ovalle. Cordillera C "s alle, 2898 (SI): € n Baños del Toro. Estero de Huanca. Región: IV Werdermann 236 (S 372 Annals of the Missouri Botanical Garden 30. Junellia cedroides (Sandwith) Moldenke. Lilloa 394. 1940. Verbena cedroides Sand- with, Kew Bull. 1927: 184. 1927. TIPO: Argentina. Neuquén, Zapala: Zapala, 27 nov. 1925, H. F. Comber 189 (holotipo, K!, K foto SI!; isotipos, BM no visto, LP! TEX no visto). Arbustos espinosos de 0.15-1 m de alt; ramas jóvenes casi glabras. Hojas dimorfas, las de los macroblastos de 4-5 X 1.8—2.2 mm. 3-partidas, lóbulos angostame nte elipticos, apice agudo, punzante, con- nadas entre sí en las ramas jóvenes; hojas de los braquiblastos de 2.5-3 20.7 mm, espatuladas, subcarnosas, ápice ph y base decurrente, casi elabras. Monobotrios en espigas plurifloras: brácteas de 3-4 mm, ovadas, ápice obtuso, base decurrente, ciliadas en el margen. Cáliz de 6-6.5 mm, campanu- lado, dientes triangulares, ápice agudo, punzante. pubescencia de pelos cortos, antrorsos, ralos; corola -ll mm, lila, externamente con pubescencia blanca, pelos muy cortos y suaves, adpresos, antrorsos, densos, menos en la base del tubo y ápice de los lóbulos, internamente pubescente, fauce con pelos claviformes: par superior de estambres sin apéndices conectivales. conectivo superando la longitud de las tecas: gineceo de 7.5-8 mm, obovoide, ápice glabro, base del estilo no ensanchada. Clusas de 44.2 mm, glabras, ápice obtuso y base angostada, cara dorsal estriada casi negra, caras comisurales papilosas, blancas. Distribución geográfica. Crece en las provincias de Neuquén y Río Negro, en provincia fitogeográfica Patagónica, alrededor de los 1200 m (Fig. 1B). Fenología. Florece en diciembre. Iconografía. Botta, 1999: 167, figs. 122, 169. Observaciones. Especie que se caracteriza por tener los dientes del cáliz triangulares y de ápice Junellia Es afín también a J. tridens, de la cual agudo y punzante, que la diferencia de asparagoides. se diferencia por tener el ovario glabro. Material adicional examinado. ARGENTINA. Neu- quén: Dpto. Zapala, de Zapala a Primeros Pinos, Cabrera 32850 (SI). Río Negro: Dpto. Pilcaniyeu, hd ~ a, 64 (8 rula 40, 28 dic. 1963, Schajovskoy s.n., SI 2 31. Junellia hystrix (Phil) Moldenke, Known Geogr. Distrib. Verb. Avicenn.: 77. 1942. Basó- nimo: Verbena hystrix Phil., Anales Univ. Chile 90: 610. 1895. TIPO: Chile. Atacama: Acerillos, 1844, SGO 42514 (lectotipo, designado por Acevedo de Vargas (1951: 55), SGO! SGO foto SI). Fig- ura LOA-D. desierto de Atacama, nov. Villanueva s.n., Arbusto de 0.50—1 m alt. tallos jóvenes densamente vellosos, con pelos adpresos, glabrescente a la madurez. dimorfas. trisectas, las de los macroblastos Hojas opuestas con lóbulos de 3-3.5 X ] mm, ovados, ápice — atenuado, punzante, nervio medio prominente en la cara abaxial, pelos glandulares en ambas caras, cara adaxial villosa en la base, margen ciliado; hojas de los 5 braquiblastos imbricadas, carnosas, con lóbulos de X 0.5 mm, ovado-angostos, ápice agudo, rígido, nervio medio no prominente, pubescencia similar a las hojas de los macroblastos. Monobotrios en espigas plurifloras, de —1.5 em, acrescentes en la fruetificación hasta 4-6 em: brácteas de 3.5-8 X 1-1.5 mm, enteras, ovadas, ápice generalmente punzante. Cáliz de 7-8 mm. con costillas prominentes, ciliado en el margen superior, superficie landulares, patentes, dientes 1-12 mm, con pelos s 'siguales y de densidad variable en la cara con pelos glandulares y no g subulados. punge antes hasta 3 mm: rol | externa del tubo, fauce con pelos claviformes; par superior de estambres sin apéndices conectivales, conectivo superando la longitud de las lecas; gineceo de 8 mm, ápice glabro, base del estilo no ensanchada, levemente hundida entre las clusas. Clusas de 3- 3.5 mm, glabras, ápice obtuso, cara dorsal negruzca reticulada, caras comisurales blancas y papilosas. Nombre “Yote” (Acevedo de Vargas. 1951: vulgar. 55). | norte de Distribución geográfica. Crece en Chile, en la Región H y HI y en San Juan (Argentina): en el dominio Andino-patagónico, provincia fitogeo- eráfica del Desierto, hasta los 3500 m (Fig. 1D). Fenología. Se ha coleccionado en flor en los meses del verano. Observaciones. Junellia hystrix es afín a J. asparagoides, por las hojas de los macroblastos rígidas y punzantes y los rígidos dientes del cáliz: J. asparagoides presenta un porte mayor y hojas enteras en los braquiblastos. Asimismo esta especie fue considerada sinónimo de Junellia bryoides (ver observación en este taxón). ARGENTINA. San Morrillos (Reserva Privada Fortunato 5602 (St). CHILE. H Material | adicional examinado. Juan: Dpto. Calingasta. Los Fundación Vida Silvestre), egión: Antofagasta, Cordillera Domeyko, Salar Punta egra, Chong s.n., Herbario Gunkel 49305 (CONC). HI Región Macama, Chañaral: camino de Pedernales, Km 14, Ricardi 1589 (4 de po» de Peder- nales, Behn s.n., CONC 23739 (CONC, 32. Junellia tetragonocalyx (Tronc.) Moldenke, Phytologia 3: 167. 1949. Basónimo: Verbena tetragonocalyx Tronc., Darwiniana 8: 481. 1949. TIPO: Argentina. Chubut, Paso de Indios: 23 km al E de El Sombrero, 6 ene. 1948, Soriano 2839 (holotipo, SIl; isotipo. K no visto. K foto SI!). Volume 95, Number 2 2008 Peralta et al. Revisión del Género Junellia (Verbenaceae) A. BESTARD. Figura 10 A-D. Junellia hystrix. —A. Aspecto general (de Zöllner 4397). —B. Detalle de una rama. —C. Flor y bráctea. B, C de Behn 23739. —D. Fruto (de CONC 31655). —E. Junellia ligustrina var. ligustrina. inflorescencia (de M. H. Nat. San Rafael 3615). —V. Junellia ligustrina var. lorentzii. Inflorescencia (de Dactuk 25). 0.8-1.2 m alt., macroblastos con entrenudos de 4—10 mm, braquiblas- Arbusto de oscuro, subeglabro: tos de 3-4 mm. Hojas dimorfas, las de los macroblastos de 2-25 X ]-1.5 mm, enteras, sésiles, connadas, agudas, en general punzantes al envejecer, cara abaxial con nervio medio prominente, recurvas; hojas de los braquiblastos enteras de 1.5 .5-] mm, carnosas, sésiles, angostamente oblongas, obtusas, a veces brevemente apiculadas. Monobotrios en espigas pluri- floras; brácteas de 2.5-5 2-3.5 mm, anchamente ovadas, obtusas y apiculadas, margen ciliado. Cáliz de 6-8 mm, con cuatro costillas, dientes subulados recurvos, ensanchado hacia el ápice, pubescente en la mitad superior; corola de 11-12 mm, blanco-amari- llenta, pubescente en la mitad superior, fauce con pelos moniliformes: estambres con conectivo de mayor lon- gitud que las tecas; par superior de estambres sin apéndices conectivales, conectivo superando la lon- gitud de las tecas; gineceo de 9-10 mm, ápice glabro, base del estilo no ensanchada, hundida entre las clusas. Clusas de 4.5-5 mm, cilindricas, ápice obtuso, cara dorsal oscura, caras comisurales blancas, papilosas. Distribución geográfica. Se la encuentra en Río Negro y Chubut; en la provincia fitogeográfica Patagónica (Fig. 2C). Fenología. Florece y fructifica en verano. Iconografía. Troncoso, 1949: 481, 482, figs. 1. 2 (sub Verbena tetragonocalyx); Botta, 1999: 182, 191, fig. 142. Observaciones. Especie muy similar a Junellia tridens de la que se diferencia por tener hojas partidas en los macroblastos, el cáliz 5-dentado, 5-costado, clusas pubescentes en el ápice. examinado. ARGENTINA. Chubut: P] umás, almacén El Guanaco, Soriano cm de Material adicional Dpto. Mártires, Las 418 (SI). Río Negro: Dpto. San Antonio, cerca de 2 Gendarmería Nacional en Sierra Grande, camino a Boca- mina, Crespo 2471 (BAB). res, Las | V 33. Junellia tridens (Lag. Moldenke, Lilloa 5: 402. 1940. Basónimo: Verbena tridens Lag., Gen. Sp. PI. 19. 1816. TIPO: Argentina. Chubut, Florentino Ameghino: a 7 km del cruce de Ruta Provincial 28 con Ruta Nacional 3, 24 feb. 1990, N. Correa 10227 (neotipo, aquí designado, SI!). Figura 7J—L. Verbena carroo Speg., Anales Soc. Ci. Argent. 15: 112. 1883. TIPO: Argentina. Santa Cruz: 17 ene. 1882, C. Spegazzini 1125 (holotipo, LP!; isotipos, K!, SH, K foto SI!). I!) Annals of the Missouri Botanical Garden |-1.5 em alt, negruzcos, erectos. as de los macroblastos de 4—4.5 X Arbustos de Hojas dimorfas, | 3-3.5 mm, sésiles, trilobadas a tripartidas, lóbulos agudos, punzantes con margen revoluto; cara adaxial con pelos adpresos, antrorsos, la abaxial hispídula; hojas de los braquiblastos ca. 1.5 X 1.3 mm, sésiles, ovadas, enteras. carnosas, dísticas, agudas, con margen revoluto, cara adaxial lisa, hispídula, abaxial surcada longitudinalmente en la parte media, con pelos breves, adpresos. Monobotrios de 15-20 mm, en espigas plurifloras, sobre braquiblastos; brácteas de 2-3 mm, trilobadas, semejante a las hojas de los macroblastos. Cáliz de 5—5.5 mm, densamente hispí- dulo, dientes agudos, rígidos, los adaxiales muy breves, los abaxiales más desarrollados, de 0.5- l mm: corola de 9.5-10 mm, lila a blanquecina, con pubescencia adpresa-retrorsa en el exterior de la mitad del tubo y de los lóbulos, fauce con pelos claviformes: lóbulos obtusos, emarginados; par supe- rior de estambres sin apéndices conectivales, co- nectivo superando la longitud de las tecas; gineceo de 6—6.5 mm, entre las elusas, ápice del ovario pubescente. Clusas base del estilo no ensanchada, hundida de 4-4.5 mm, ápice obtuso, pubescente, cara dorsal reticulada, casi negra, caras comisurales blancas, papilosas. Cariología. 2n = 20 (sub Verbena tridens, Dollenz, 1976: 165, fig. 1, 166). de la Peña & p Nombre vulgar. “Mata negra” Pensiero, 2004: 367). Distribución geográfica. Crece en las provincias de Chubut, Río Negro y Santa Cruz y en Chile en | Región XI; en la provincia fitogeográfica Patagónica, en el distrito occidental (Fig. 2F). Suele ser elemento a dominante de la estepa, se la encuentra en terrazas, mesetas y en el fondo de cañadones o valles fluviales cerca de mallines, forma matas hemisféricas. Fenología. | Florece entre diciembre y febrero. Iconografía. — Macloskie, 1905: pl. 23 (sub Verbena carroo Speg.); Soriano, 1956b: 368, fig. 8C; Botta, 1999: 191, fig. 141. Observaciones. droides de la que se diferencia fundamentalmente por a pilosidad del ápice del gineceo. Fue necesario designar un neotipo, porque el material que estudió Lagasca fue destruido. Se eligió más el material de Correa, por ser uno de los completos. Material adicional examinado. ARGENTINA. Chubut Dpto. Ameghino, Ruta Nac. 3, 10 km al S de Garayalde Botta 454 (SI). Río Negro: Dpto. Avellaneda, 18 km de Choele Choel | Especie muy afin a Junellia ce- acia Río Colorado. Correa 10519 (BAB) Santa Cruz: Dpto. Corpen Aike, Ruta 3, 32 km al S del Río Chico, Botta 488 (SI); Dpto. Lago Argentino, Soriano 3368 (SD. CHILE. XH Región: Magallanes, Kimiri Aike, Pedersen 14430 (SI). Junellia subgén. Thryothamnus secc. Verticiflora (Schauer) Botta, Darwiniana 29: 392. 1989. Verbena secc. Verbenaca ser. Verticiflorae Schauer in DC., Prodr. 11: 544. 1847. Verbena secc. Junellia ser. Verticiflorae (Schauer) Tronc., Darwiniana 18: 313. 1974. TIPO: Verbena aspera Gillies € Hook. ex Hook. |= Junellia aspera (Gillies € Hook. Hook.) Moldenke var. aspera |. €x Arbustos con ramas generalmente de sección poligonal o menos comúnmente tetragonal. Hojas con láminas variables, no espiniformes. Monobotrios o dibotrios en espigas, alargadas o contraídas. Corola con fauce glabra o con pelos semejantes a los del interior del tubo corolino; estambres con conectivo de mayor longitud que las tecas, con o sin apéndices conectivales desarrollados. Junellia arequipensis (Botta) Botta, Darwini- ana 29: 392. 1989. Basónimo: Verbena arequi- pensis Botta, Darwiniana 28: 237. 1987 [1988]. TIPO: Perú. Arequipa: encima de baños de Jesús, 23 abr. 1961, 2600-2700 m, A. Ferreyra 14261 (holotipo, SI!). Y = e Arbustos ramosos, de 1.20—1.80 m de alto; ramas jóvenes de sección poligonal, hispídulas. Hojas dimorfas, las de los macroblastos de 0.8-2 X 0.3— 0.7 cm, opuestas, a veces alternas, enteras, coriáceas, subsésiles, elípticas, ápice agudo y base decurrente; cara adaxial estrigosa y abaxial híspida, especial- mente sobre el nervio medio prominente; hojas de los braquiblastos iguales a las de los macroblastos pero menores, ocasionalmente, con lámina obovada, ápice obtuso. Monobotrios plurifloros, en espigas alargadas, de 10-14(-16) em. cilíndricas: brácteas de 2.54. X 0.5 mm, angostamente ovadas a lineares, pubes- centes, ciliadas en el margen. Cáliz de 4-5 mm, 5- dentado, dientes triangulares, desiguales de 0.4— 0.5 mm, híspido sobre las costillas y estrigoso entre as mismas; corola de 13-15 mm, pardo rojiza y lóbulos amarillos, pubescente en la parte superior del ase de los lóbulos, el resto de la superficie tubo y ł glabra: interior del tubo pubescente a la altura de la inserción de los estambres; fauce glabra; par superior de estambres sin apéndices conectivales, conectivo superando la longitud de las tecas; gineceo de 11— 11.5 mm, del no ensanchada hundido entre las clusas. Clusas de 2.5-3 mm, pericarpo con alas a penas desarrolladas, cara dorsal reticulado- base estilo estriada, caras comisurales papilosas, blancas. Volume 95, Number 2 2008 Peralta et al. Revisión us Género Junellia (Verbenaceae) ^ c Nombre vulgar. “Romerillo”, Arenas 84 [SI] y Hinckley 54 [SI]). capo colorado” (de Distribución geográfica. Coleccionada en Perú y Chile, entre los 15°S y 20°S, en laderas rocosas entre los 2200 y 4300 m (Fig. 1A). La región corresponde al dominio Andino-patagónico, provincias Altoandina y del Desierto, distrito de los Cardonales. Fenología. Florece desde octubre hasta abril. Iconografía. Botta, 1987: 238, fig. 1 (sub Verbena arequipensis). Observaciones. Junellia arequipensis se caracte- riza por sus flores, con tubo recurvo, de color pardo rojizo y muy perfumadas. Es afín a J. aspera diferenciandose por los dientes del cáliz más largos y los frutos marcadamente alados. Material nu e CHILE. I Región: Arica, 0 (SI); Tarapacá. Cord. Co. LM Werdermann 1111 ts PERÜ. Arequipa: Arequip Estanquillo Ahogado, Chachani, Arenas 84 (SI). Cajatambo, ima, cerca de Churin, Ferreyra 18706 (SI, USM). Tacna: Tarata, Tisaco, Cano 8312 (USM). 35. Junellia aspera (Gillies & Hook. ex Hook.) Moldenke. 35a. Junellia aspera (Gillies & Hook. ex Hook.) Moldenke var. aspera, Lilloa 5: 393. 1940. Basónimo: Verbena aspera Gillies & Hook. ex Hook., Bot. Mise. 1: 163. 1829. TIPO: Argen- tina. Mendoza: Paramillo de Uspallata “near Los Hornillos, eastern descent from Paramillo de Uspallata”, nov. 1822, J. Gillies s.n. (lectotipo, designado por Botta (1989: 392), K!; isotipo, BM!, BM foto SI!). Figura 7F-I Verbena ourostachya Briq., Annuaire Conserv. Jard. Bot. e 4(2): 20. 1900. TIPO: nig e Mendora San Rafael: Plantes des environs de St. aël & de la Vallée du Rio Mendoza, jan. & fév. T T Wilczek 51 (lectotipo, ucc aquí, G!, G foto FM 24695 SI!; isotipos, F!, NY!). sisi tripartita Moldenke, Phytologia 3: 36. 1948, syn. TI Argentina. Mendoza, Santa Rosa: Las Catia 4 abr. 1942, A. Ruiz Leal 7966 (holotipo, NY no visto; isotipos, BA!, MERL).. ee a a var. glandulosa Botta, Hickenia 2: 128. syn. nov. TIPO bog ined entre Jagüe y Sierra Punta Negra, )0— m, 23 ene. 1949, A. Krapovickas 5480 ee SI!; isotipo, BAB!). : Argentina. La Rioja, Arbustos ramosos de 0.80-1.20 m de alto. Ramas jóvenes de sección circular, divaricadas, glanduloso- pubescentes, al envejecer glabrescentes y espines- centes; corteza pardo-oscura, con estrías longitudi- nales. Hojas dimorfas, sésiles, las de los macroblastos a veces alternas, raramente verticiladas, generalmente lámina trisecta, lóbulo medio de 4.5-15 X 2-3.5(-7) mm, notablemente mayor que los laterales, elípticos a obovados, ápice agudo a obtuso, con la cara adaxial más densamente escabroso-glandulosa que la abaxial, ésta con el nervio medio prominente; hojas de los braquiblastos de (6—)8-14 X 1.2-1.3(-2) mm, enteras, elípticas a obovadas, con igual pubescencia que la de los macroblastos. Monobotrios plurifloros en espigas contraídas, de más de 5-10 cm a la madurez; brácteas de 3—5 mm, enteras, angostamente elípticas a lineares, ápice subulado, margen ciliado. Cáliz de 6—10 mm, híspido-glanduloso o escabroso-glanduloso, dientes subulados, el desiguales, de 1-1.8 mm, adaxial muy breve; corola de 6.5-10.5 mm, blanca, glabra en el exterior y escasamente glandulosa- pubescente en el interior, a la altura de la inserción par superior de de los estambres, fauce glabra; estambres sin apéndices conectivales, conectivo superando la longitud de las tecas: gineceo de a. Clusas de 5-7 X 2.5-3 mm, elipsoidales, glabras, con el pericarpio 8 mm, base del estilo no ensanchad prolongado lateralmente en forma de dos alas membranáceas, conspicuas, cara dorsal pardo-oscura, lisa, caras comisurales blancas, papilosas. Cariología. 2n = 60 (Junellia aspera, Covas & Schnack, 1946: 156, 159, fig. 2H). Nombre pn "Capo colorado" (Moldenke & Moldenke, 1949: Argentina, desde Cata- Habita en el Monte, en zonas Distribución. geográfica. marca hasta Río Negro (Fig. 1A). dominio Chaqueño, provincia del áridas entre los 270 y 2200 m. Fenología. Florece desde noviembre hasta abril. Iconografía. Sanzin, 1919: 119, fig. 23 (sub Verbena aspera); Roig, 1970: 146, fig. 72 (sub Verbena aspera); Botta, 1999: 164, fig. 120. Observaciones. La variedad típica presenta el cáliz glanduloso, tubo de la corola apenas más largo que el cáliz y las hojas generalmente trisectas. Los ejemplares provenientes de San Juan, Dpto. Iglesia presentan hojas hasta de 6 mm lat, sin embargo dado que el resto de caracteres coinciden con Junellia aspera se los incluye dentro de esta especie. De acuerdo con Botta, 1987: 242, la especie ha sido excluída de la flora peruana. Se eligió como lectotipo de Verbena ourostachya el ejemplar E. Wilezek 51, depositado en G, por estar manuscrito el nombre de la especie y la referencia “sp. nov.”, por el autor de la misma. ARGENTINA. Cata- ma ;apayán, Ambato, Quebrada y Cuesta de la Sébila, ruta 60, entre Chumbicha y el desvío A. T. Hunziker 18390 (SI). La Pampa: Dpto. Material adicional examinado. Sierra de hacia Pomán, Annals of the Missouri Botanical Garden Curacó, a 25 km de Puelches, Laguna Amarga y salares de Chadileufú, Rutz Leal 26046 (SI). La Rioja: Dpto. y lipe Varela, Parque de Talampaya, A. T. Hunziker 11659 (SI). Mendoza: Dpto. o El Carrizal, Múlgura nid (SI). Neuquén: Dpto. Anelo, meseta al N de Mina Esc 'ondida . lgles Sed 1. H (Auca M a), NAR 8662 (SD). San Juan: ruta 150, 18 km al W de Las Flores, Valle de | Hunziker 11622 (SI). San Luis: Dpto. Belgrano, Sierra de Las Quijadas, quebrada del Alambre, al W de San Antonio, ruta nac. 147, A. T. € 10359 (SI). Río Negro: Dpto. Avellaneda, 35 km al | © Choele Choel, Cabrera 18605 (SI). 35b. Junellia aspera var. longidentata (Mol- denke) Mulgura € P. Peralta, comb. nov. Basonimo: Junellia longidentata Moldenke, Known Geogr. Distrib. Verb. Avicenn.: 77. 1942. Verbena aspera Gillies & Hook. ex Hook. var. longidentata (Moldenke) Botta, Dar- winiana 25: 338. 1984, TIPO: Argentina. Tucu- man: Amaicha, 2090 m, 29 die. 2 Castillón 2460 (holotipo, NY; tsotipo, LIL 317531). Ramas jóvenes sección tetrágona, hispídulas fle- generalmente homomorfas, enteras, de ).3-0.8 em, en los macroblastos, xibles. 8 1.3-2.5 X oneness opuestas, pecioladas, y en los braquiblastos, fasciculadas, lámina membranácea, elíptica u obo- vada, ápice generalmente agudo, a veces mucronado, margen entero, pubescencia de pelos cortos y rígidos, dispuestos espaciadamente en ambas caras. Raquis, Tubo calicinal de brácteas y cáliz, incano-vellosos. 1.5-5 mm, dientes subiguales ca. 1.5-2 mm, ápice subulado. Corola de 9-10.5 mm, tubo de 6.5-8 mm, recurvo, glabro. Cariología. 2n = 40 (Junellia longidentata, 1088: 542, 544, fig. 1C). Poggio et al., Distribución geográfica. Noroeste de Argentina, desde Jujuy hasta La Rioja (Fig. zona alta de la provincia del Monte, llegando hasta los Habita en la 3000 m en la Prepuna. Fenología. Florece de diciembre a marzo. fig. 3 (sub Verbena 31 (sub Botta, 1984: 339, aspera. var. longidentata); Botta, 1993: 71. Iconografía. fig. Junellia longidentata). Observaciones. Especie afín a Junellia arequipen- sts por el porte; se diferencia por presentar sección del tallo poligonal, cáliz estrigoso, con dientes triangula- es, corola pardo-amarillenta, 2 a 3 veces más larga que el cáliz, y las alas del pericarpo pequeñas. ARGENTINA, nar marca: Dpto. Andalgalá, Andalgalá, Jörgensen 1614 Dpto. Belén, Hualfin, Sayago 600 (SD). La Rioja: Dpto. al y el Jagúe, aproximadamente ¿ Material adicional examinado, fev] Sarmiento, entre el peñón 11 km del primero, que ‘brada del río El Peñón. Biurrun 4578 > rri Dpto. Tilcara, Quebrada de Huasamayo, Botta SI). Salta: Dpto. Cachi, recta de itin, Kiesling 6411 e oe s. loc., Morello 1287 " 36. Junellia cinerascens (Schauer) Botta, Dar- winiana 29: 392. 1989. Basónimo: Verbena cinerascens Schauer in DC. Prodr. 11: 545. 1847. Diostea ipn (Schauer) Moldenke, Revista Sudamer. Bot. 5: 1. 1937. TIPO: Chile. IV Región, Caan 1829, C. Gay sn (lectotipo, designado por Botta (1989: 392), G foto FM 7855 SI! [ejemplar a la derecha de la foto]: isotipos, F!, KD. Figura 11A-4 Arbustos cinéreos, erguidos, 2 m alt; ramas espinescentes, rígidas, sección tetragonal, cano- lomentosas, generalmente sin desarrollo de braqui- blastos. Hojas homomorfas de 4-8(215) X 1—5 mm, subsésiles, lámina ovada o eltptica-angosta, ápice agudo, base decurrente, pelos patentes y rígidos en ambas caras. Monobotrios plurifloros, espigas de 4 cm, alargadas; brácteas de 2-2.5 mm, ovadas, pelos adpresos en su superficie. Cáliz de 4.5—4.8 mm, con pelos blancos rígidos y adpresos, dientes de 1 mm, los adaxiales mayores, los restantes brevísimos: corola de 12-14 mm, blanca hasta rosada, superficie externa pubescente en la mitad superior del tubo y base de los lóbulos: superficie interna del tubo y fauce pu- bescente; par superior de estambres con apéndices conectivales exertos y claviformes, conectivo supe- ando la longitud de las tecas; gineceo 7.5-8 mm, hundida profunda- base del estilo no ensanchada. Clusas ca. 1.2 mm, mente entre las elusas. ápice obtuso, cara dorsal parda, caras comisurales blancas papilosas. Distribución geográfica. Hasta el momento sólo se ha coleecionado en Chile en las regiones IV y V, entre los 30 S y 33 S (Fig. 1B). Habita en faldeos y quebra- das entre los 1000 y 1500 m, en la provincia fitogeo- gráfica Chilena Central, dominio Andino-patagónico. y fructifica en verano. Fenología. Florece £ Observaciones. Esta especie se caracteriza por el estilo profundamente hundido en el ovario; este carácter se hace más pronunciado al desarrollarse el fruto. Es afín a Junellia scoparia que presenta sección del tallo poligonal, ausencia de apéndices conecti- vales y el estilo apenas inserto entre las elusas. Poepp.. es un nombre Citharexylon alpinum manuscrito en herbario, no publicado validamente. Fue citado por Walpers (1845: 78) como sinónimo de Gillies & Hook. ex Hook. y teriormente por Schauer (1847: 545) como sinónimo de V. Moldenke (1960: 309) lo consideró cinerascens Verbena scoparia pos- cinerascens Schauer. en la sinonimia de Diostea Volume 95, Number 2 Peralta et al. 2008 Revisión E Género Junellia (Verbenaceae) 77 Figura 11. AG. 2 cinerascens. —À. longitudinal de ovario. —E. 4t -] ions 326 Aspecto general. —B. Flor. —C. Anteras. —D. Esquema del corte ulo cara adaxial. —F. Fruto cara abaxial. —G. Corte tauia de fruto. A-G de Marticorena ne —J. pe recto general. —K. Inflorescencia. —L. Fruto vista adaxial. —M. Fruto vista abaxial. J-M de (Schauer) Moldenke y hace referencia a las fotos del n: Retorca, Los Andes por Río Colorado, 2 dic. 1929, . HE o T " citado ejemplar de Poeppig (FM 7855). D uA E Material adicional examinado. CHILE. IV Región: 37. Junellia echegarayi (Hieron.) Moldenke, Li- Limarí, Estancia Cabrería, faldeos, Jiles 1535 (CONC). V lloa 5: 395. 1940. Basónimo: Verbena echegarayi Annals of the Missouri Botanical Garden Hieron., Bol. Acad. Nac. Ci. 4: 66. 1881. TIPO: Argentina. San Juan: Leoncito, ene. 1876, S. Echegaray s.n. (holotipo, CORD!). Figura 11J-M Junellia echegarayi var. corola Moldenke, Phytologia 2: 465. 1948, e doza, Las , K. Grandjot is, s.n., Hb. Ruíz Leal 4714 elut: ME um Junellia echegarayi var. puke enia Moldenke, Phytologia 2 2: 466. 2 syn. nov. 19 37, K. undo s.n. He e Ruíz e 4713 a MERL!). Arbusto de 0.40-1.20 m de alt.; sección poligonal o circular, pubérulos, al envejecer tallos jóvenes de glabrescentes y espinescentes; corteza pardo ama- rillenta, estriada longitudinalmente. Hojas dimorfas, enteras, las de los macroblastos opuestas, de 6-8(212) 1.5—4 mm, subpecioladas, recurvas, ovadas, ápice base obtusa o cordiforme, algo agudo, margen revoluto, cara adaxial algo escabrosa y abaxial híspida; las de los braquiblastos sésiles, imbricadas, de 4—12 X 1.5—3 mm, espatuladas, ápice obtuso, base atenuada, margen algo revoluto y pubescencia similar a las hojas de los macroblastos. Monobotrios en 5(—6) em, ciladas; brácteas de 3—5 X 2-3 mm, ovadas a veces espigas plurifloras, de 2.5— flores subverti- rombiformes, ápice agudo o acuminado, pilosas en el margen. Cáliz de 5.5-6 mm, hispídulo; dientes muy breves, triangulares, ápice rígido, subiguales, el 9-11 blanca, glabra en el exterior o con pocos pelos abaxial de 0.5 mm; corola de 5 mm, lila o esparcidos, pubescente en el interior, fauce glabra, ápice obtuso; par superior de estambres sin apéndices conectivales, conectivo superando la longitud de las tecas; gineceo de 6 mm, base del estilo no ensan- chada. Clusas de 44.1 pericarpio expandido lateralmente formando dos alas X 2-2.2 mm, glabras, con el conspicuas, ápice obtuso, base angostada, cara dorsal lisa, pardo oscuro, caras comisurales papilosas, blancas. Distribución geográfica. Argentina en San Juan y Mendoza; citada para Perú por Macbride (1960: 620), en base al ejemplar Weberbauer 1398, no se ha tenido acceso a este ejemplar, ni tampoco a otros materiales, que avalen efectivamente la presencia de la especie en Perú (Fig. 1C). provincia fitogeográfica del Monte y la Prepuna, entre Crece en el límite entre. la los 1900 y 3000 m, en planicies secas y arenosas. Fenología. Florece y fructifica en verano. Iconografía. Sanzin, 1919: 120, fig. 24 (sub Verbena echegarayi). Observaciones. Especie muy afín a Junellia as- pera, de la que se diferencia principalmente por los dientes del cáliz subulados y de mayor tamaño y la pubescencia del cáliz glandulosa. Las variedades fundadas por Moldenke se con- sideran sinónimos de la especie porque al observar mayor cantidad de ejemplares no se encontraron diferencias suficientes como para mantenerlas. Material examinado. ARGENTINA. Men- doza: Dpto. Las Heras, Paramillo de Us 23294 (SD); Dpto. Tupungato, Tupungato, Ruiz Leal 2913 (SI). San Juan: Dpto. Iglesia, camino al Valle del Cura, ies del Romo, Kiesling um Sp. adicional »allata, i Leal => E 38. Junellia ligustrina (Lag.) Moldenke. 38a. Junellia ligustrina (Lag.) Moldenke var. ligustrina, Phytologia 2(11): 466. 1948. Basó- Gen. Sp. PL 1816. Lippia ligustrina (Lag.) Britton, Trans. York Acad. Sci. 9: 181. 1890. Aloysia ligustrina (Lag.) Small, Fl. South. U.S. 1013. 1903. TIPO: Argentina. Santa Cruz, Deseado: a 107 km al S del límite con Chubut, ruta Nac. 3, 25 feb. 1990, M. N. Correa et al. 10256 (neotipo, aquí designado, SI!). Figura 10E. nimo: Verbena ligustrina Lag., New Arbustos erguidos, de 1—1.5 m alt. Ramas rígidas de sección tetragonal. Hojas homomorfas, opuestas, a ‘ € veces subalternas, de 1—3 X 0.2-1.2 cm, enteras, coriáceas, elípticas u oblongas, ápice obtuso, mucro- nado, base obtusa o atenuada, escabrosas. Monobo- trios en espigas alargadas, plurifloras de (3—)4.5— 9 cm, flores perfumadas; brácteas de 4 mm, angosta- mente ovadas, pubescentes. Cáliz de 5-6 mm, dientes triangulares, subiguales, de 1 mm, con pubescencia adpresa; corola amarillenta con fauce rojiza, glabra, de 10-16 mm, tubo recurvo, glabro en el exterior, pubescente en el interior; par superior de estambres sin apéndices conectivales, conectivo superando la longitud de las tecas; gineceo de 10-14. mm, estilo Clusas de 4—4.5 mm, glabras, pericarpio prolongado lateralmente formando con la base no ensanchada. dos alas conspicuas; cara dorsal pardo-oscura, caras comisurales blancas papilosas. Cariología. 2n = 20 (sub Verbena ligustrina, Covas & Hunziker, 1954: 251); (Junellia ligustrina, Botta & Brandham, 1993: 146, fig. 1D). Nombre vulgar. “Verbena” (Soriano, 1956a: 326). Distribución geográfica. Habita en Argentina, en Neuquén, Chubut y Santa Cruz; crece la provincia fitogeográfica Patagónica desde el nivel del mar, hasta los 2000 m formando parte de las estepas arbustivas áridas y semiáridas (Fig. 1D). Fue citada para Perú; no se ha observado material que lo avale (Pool, 1993: 1171 Fenología. Florece en verano. Volume 95, Number 2 2008 Peralta et al. 379 Revisión id Género Junellia (Verbenaceae) Iconografía. Botta, 1999: 171, figs. 127, 173. Observaciones. Comparte con Junellia aspera y J. echegarayi los frutos alados lateralmente. Se diferen- cia de ambas porque presentan las hojas membraná- ceas; J. aspera además presenta las hojas de los macroblastos tripartidas. Como se expresó en otros casos los ejemplares estudiados por Lagasca han sido destruidos, por lo tanto se eligió un ejemplar que coincide con los caracteres de la especie. Lippia ligustrina (Lag.) Kuntze (Revis. Gen. Pl. 3(2): 252 homónimo posterior. 1898) es un nombre ilegítimo por ser un Material adicional examinado. ARGENTINA: Chubut: Dpto. Biedma, 8 km S de Puerto Madryn, ruta prov. 1, Correa 10161 (SD. pug Dpto. Catan Lil, La Negra, Schajov- skoy s.n., SI 25628 (SI). Santa Cruz: Dpto. Deseado, Valle del río Deseado, ruta 277, Soriano 5051 (SI). 38b. Junellia ligustrina var. lorentzii (Niederl. ex Hieron.) Moldenke, Phytologia 47: 222. 1980. Basónimo: Verbena lorentzii Niederl. ex Hieron., Bol. Acad. . Ci. Córdoba 3(4): 370. 1880. Junellia lorena Pe ex Hieron.) Moldenke, Lilloa 5: 397. 1940. TIPO: Argentina. Río Negro, set.-nov., 1874, E^ Berg 100 (lectotipo, aquí designado, CORD!). Figura 10F. Annuaire Conserv. Jard. Bot. Geneve 4(2): 19. 1900. TIPO: PME e Cerroyo Manga, entrée de la vallée de l'Atuel, Wilczek 49 (hol E " isotipos, MO no visto, SI!, » foto FM 24688 SI!). Verbena | alatocarpa p Darwiniana 8: 485. 1949. Nombre reemplazado: Junellia chubutensis Moldenke, Phytologia 2: 46: nom. illeg., non Junellia chubutensis ee) Moldenke, Lilloa 5: 394. 1940. Ju get es cds (Tronc.) ld Phytologia 3: 167. . TIPO: Argentina. Chubut: Puerto Madryn, 24 oct. au. ( 0 "Donell ia NY!; isotipos, LIL!, SI! Verbena inconcinna Bri Difiere de la variedad tipica por las hojas en general más pequeñas, las inflorescencias menores, 1.5—3(—6) cm y el cáliz con pubescencia híspida o hispídula. Cariología. 2n = 20 (sub Verbena inconcinna, Rahn, 1960: 123). Distribución | geográfica. Crece en Argentina desde el sur de Mendoza y Buenos Aires hasta Santa Cruz, en la provincia fitogeográfica Patagónica, desde el nivel del mar hasta los 2000 m (Fig. 1D). Fenología. Florece desde octubre hasta febrero. Iconografía. Lorentz y Niederlein, 1881: tab. XII, I, figs. 2-6 (sub Verbena lorentzii Niederl.). Observaciones. Se eligió como lectotipo de Verbe- na lorentzii, el ejemplar de C. Berg, ya que la publicación de Hieronymus, se basa en materiales de dicho coleccionista y los originales se encuentran depositados en CORD. Existe una foto: B foto FM 17428, que podría corresponder a la descripción del homónimo posterior Verbena lorentzii Niederl. (Lo- rentz € Niederlein, 1881: 264, tab. XII, I, figs. 2-6). ARGENTINA. Buenos Aires: Dpt rmen de Patagones, Médano Comody, Miccio Peralta 30 (S D. Chubut: Dpto. Biedma, Puerto Madryn, Soriano 2717 (SI). Pampa: Dpto. Chical Co, Aguas de Torres, Adler s.n. (SI); Dpto. Lihuel Calel, Sierra Lihuel Calel, Cabrera 19437 (SI). Mendoza: Dpto. Malargiie, Sierras de Chachahuén, Cerro de Ureta, Prina 1454 (SI). Neuquén: Dpto. Catán Lil, Ruta 40, entre La Negra y Catán Lil, Gentili 831 (SI). Río Negro: Dpto. El Cuy, a 70 km del licia, Correa 4750 (SI). Santa Cruz: Dpto. Lago Argentino, ruta 40, 16 km N de Paso Biggeri, sobre río La Leona, Boelcke 12654 (SI). Material adicional examinado. Cerro Po 39. Junellia scoparia (Gillies & Hook. ex Hook.) Botta, Darwiniana 29: 392. 1989. Basó- nimo: Verbena scoparia Gillies & Hook. ex Hook., Bot. Misc. 1 9. Diostea scoparia (Gillies & Hook. ex EM Miers, Trans. Linn. Soc. London 27: 104. 1869. TIPO: Argentina. Mendoza: Las Heras, shrub valley nr. Villavi- cencio, J. Gillies s.n. (lectotipo, designado por Botta (1989: 392), K!, K foto SI!). Figura 12A-G Anales Univ. Chile : 402. 1862. Lippia scirpea Phil., Di s.n., SGO 4 A s designado por Acevedo de Vargas (1951: 42), SGO!; isotipo, SI!). "e. 2d Phil, Anales Univ. Chile 90: 623. 1895. hile. Valparaíso: Cam e Quillota, Aug. a 1884, SGO 54829 doo designado por Acevedo de Vargas (1951: dea SGO!). Verbena Mi gels Gillies & Hook. ex Hook. var. puberula Tronc., Dansk. Bot. Ark. 220) 109. 1963, syn. nov. Diostea scoparia var. puberula (Tronc.) Moldenke, — Phytologia 19: 1970. a se (ue puberula (Tronc.) Botta, Hickenia 2 "TIPO: Argentina. e. San Raluel: Cue p los terneros, 30 km SW de San Rafael, 34°S, 1200 m, 22 nov. 1955, T. p 1143 (holotipo, SI!). Arbustos erguidos, de 0.90—1.20 m de alt., subáfilos; ramas opuestas, rígidas, de sección poligonal, espini- formes, sin desarrollo de braquiblastos. Hojas de 5—6 X 3 mm, opuestas, sésiles, enteras, caedizas, lámina ovada o linear, ápice obtuso y base decurrente. Monobotrios o dibotrios plurifloros, en espigas de 2.5—5 cm, alargadas, raquis de las florescencias glabro o pubérulo; brácteas de 1.5-3 mm, ovadas o excepcionalmente redondeadas, subuladas, generalmente ciliadas en el margen, super- ficie adaxial glabra o pubérula; flores perfumadas. Cáliz de 4—5.5 mm, glabro o pubérulo, dientes desiguales, triangulares los abaxiales de 0.5 mm, los adaxiales 380 Annals of the Missouri Botanical Garden x t: A NE "E M y A Ye NA y iy P y Figura. 12. Anteras. —F. Junellia scoparia. —A. ecto general. Fruto cara adaxial. —G. B rulo cara abaxial. más breves; corola blanco-violácea, de 9-13 mm. glabra exteriormente; tubo recurvo, fauce pubes- cente; par superior de estambres sin apéndices conectivales, conectivo superando la longitud de las tecas; gineceo de 7 mm, base del estilo no ensanchada. Clusas de 44.2 mm, ápice obtuso, cara dorsal estriada- reticulada, pardo-oscura. caras comisurales blancas, papilosas. Cartología. 2n & Schnack, 1946: = 20 (sub Diostea scoparia, Covas 155). Vombre vulgar. “Clavillo del campo”, “escobilla del campo” (Acevedo de Vargas, 1951: 44); (Ruíz Leal, 1972: 228) “solupe negro” Distribución geográfica. Chile, regiones IV, V y Región Metropolitana y Argentina desde Catamarca hasta Río Negro, en las provincias fitogeográficas del D | O O —B. A=G de Cabrera 4301. == LS Sn | I A MEN uU id SS Sa Inflorescencia. —€. Cáliz abierto y fruto. —D. Flor. . Monte y Patagónica, en laderas arenosas no pedrego- de meseta, entre los 1800 y 3000 m, en general en la zona nd formando man- sas, en bordes chones más o menos densos (Fig. Fenología. Florece generalmente en verano tem- prano, habiéndose encontrado en flor hasta el mes de abril. 1829, Bot. Mise. I: tab. 47 (sub Verbena scoparia); Sanzin, 1919: 123, fig. 26 (sub Verbena scoparia); Acevedo de Vargas. 1954: 43. fig. Iconografía. Hooker, LA (sub Lippta aphylla), C (sub Lippia scirpea), D (sub Verbena scoparia), E (sub Verbena scoparia var. 1957: 169, fig. 2, Vs > tallo (sub Verbena scoparia); Ruíz Leal, 1972: 227, lám. 70, 265 (sub Verbena scoparia); Botta, 1999: 178, 180, fig. 132 (sub Junellia scoparia). aphylla): Troncoso, corte anatómico de Volume 95, Number 2 2008 Peralta et al. Revisión del Género Junellia (Verbenaceae) Observaciones. Los ejemplares de esta especie se caracterizan por sus tallos aparentemente áfilos y que Junellia cinerascens que presenta tallos de sección tetragonal y ennegrecen al secarse. Se diferencia de base del estilo profundamente hundido entre las clusas. De acuerdo a la observación del material es muy díficil poder establecer los límites entre las diferentes jon variedades, ya que por un lado la pubescencia es muy variable en densidad y por otra sólo se encuentra algún ejemplar con brácteas más subovadas o suborbiculares, como para justificar la revalidación de la variedad definida por Acevedo de Vargas. En el caso de Lippia aphylla, Acevedo de Vargas (1951: 44) escribe: “Valparaíso: Campana de Quillota. Aug. Borchers 1884 (Clastotypus et typus de Lippia aphylla: SGO 42424 et 54829)". misma autora (1954: 42) explica que el ejemplar SGO Posteriormente. la 54829, está “en perfecto estado y que permite distinguir bien sus caracteres”. En el material estudiado, vuelve a citarlos, pero invirtiendo el orden de los números de los ejemplares ("Typus y Clastotypus SGO 42424. y 54829”). De acuerdo con nuestras observaciones efectivamente el ejemplar SGO 54829, es el más completo, por lo cual se considera efectiva la lectotipificación de Acevedo. Hook. Vargas (Acevedo de Verbena scoparia Gillies & ex Hook. var. aphylla (Phil.) Vargas, 1954: 45) es un nombre ilegitimo, de acuerdo al Art. 33.3 del Código Internacional de Nomenclatura Botánica (Greuter et al., 2000). Verbena scoparia Tausch (Flora 19(2): 390. 1836) es un nombre ilegítimo por ser un homónimo posterior. Acevedo de Moldenke describió Diostea scoparia var. subulata Moldenke (Phytologia 44: 123. 1979). El material citado por Moldenke, pertenece a Neosparton aphyllum (Gillies & Hook. ex Hook.) Kuntze; por esta razón este taxón no se considera en la sinonimia de Junellia scoparia. examinado. ARGENTINA. Cata- La Plaza, — 1403 (BA, K, Santa Florentina, . SI 19233 (SI). Ruíz Leal Material adicional marca: Dpto. A SI). La Rioja: camino a la Mina de Oro, endoza: Dpto. Las Heras, Quebrada del 25904 (SI). Neuquén: Dpto. Aluminé, entre piedra Gaucha y Cabrera 32952 (SI). Río N de Pi E Toro, egro: Dpto. Picaniyeu. a 3684. (SI). an: Dpto. Ang: de Pie de I har camino al Mogote de los Corralitos, Kiesling 4810 (SI). HLE. Región: Coquimbo, Illapel, Cerro C MR an d 16692 (SI). V Región: Juncalillo, Ricardi 2910 (Sl); inson 42 (K). Región Metropolitana: de Santiago, Cerro Cortadera, Werdermann Aluminé, camino a Bariloche, Nicora Aconcagua, llano de Valparaíso, Cerro La Campana, Hutch- Santiago, Cordillera 188 (K, S TAXONES DUDOSOS Moldenke. 396. 1940. e reemplazado: Verbena aretioides Hayek, Junellia hayekii Lilloa 5: Nom — Bot. Jahrb. 42: 163. 1908, hom. illeg. TIPO: Perú. Iter a Tacora ad Pomarape, 4200-4400 m s.m., oct. 1876, Stübel 100b (holotipo, BY, B foto FM 17402 SI). existía el material citado por lara conocer si Hayek, se consultó a los siguientes herbarios: B. BR, C, CAS, F, G, GB, GH, GOET, H, K, LE, NY, PH, S, TEX, US, W, WRSL, WU, y en ninguno de ellos existe el material. El siguiente ejemplar: PERÚ. Entre Puno Arequipa, mar. 1943, Sanderman 3855 (K). que fue citado bajo este nombre, se considera aquí, que es Junellia minima, especie citada para la misma región geográfica. Gen. Sp. Pl. 18. 1816. Verbena polyenemoides Lag., TIPO: “in dni Como se expresó anteriormente los materiales de agasca fueron destruidos, en este caso la descripción es poco clara y además en la “planicie bonaria”, no crece Junellia. Verbena spartioides Turez.. Bull. Soc. Imp. Natura- Moscou 36(3) 195. 1863. TIPO: Chile. Mac Rae 43 (no localizado). listes Cumbre, No fue posible localizar el ejemplar tipo. por lo que se considera a esta especie como un taxón dudoso. De acuerdo con la descripción original por el tallo tetrágono, cinéreo y hojas pequeñas, podría tratarse de un nombre sinónimo de Junellia cinerascens. Literatura Citada Vargas, R. 1951. Indice Específico de as, nuevas o críticas del Herbario del Mus. Nac. Hist. Nat. Acevedo Verbenáceas Chilen Museo Nac ional. Bol. Chile 25: 35-72 . 1954. Consideraciones sobre Verbena scoparia Gill. el Hook. Revista Chilena Hist. Nat. 54(4): 41—45 Ancibor, E. 1980. puna le Jujuy. IL Soc. Argent. Bot. 19: 157-202. Atkins, S. 2004. (e con) The Santiago de Estudio anatómico de la vegetación de la Anatomía de las plantas en cojín. Bol. 449-408 in K. Kubitzki Plants, Verbenaceae. Pp. Families and Genera of Vascular Vol. 7. Springer-Verlag, Berlin. Bócher, T. W. 1979. Xeromorphic leaf types. Biol. Skr. 22: 1-73 , J. P. Herting € K. Rahn. 1963-1972. Botanical studies in the Atuel dues area, LONE Province Dansk Bot. E Moore & = s (editores). 1985. Seid Austral. CONICET, E Boelcke, O.. D. M. Transecta Botánica de Buenos Aires Dotta, s M. 108 84. Las especies del género Verbena > cc. ics del NW Argentino. Darwiniana 25: 331-35 Verbena (seco. rm 987 [1988]. Una nueva especie de m Darwiniana 28: 237-243. 1989. Estudios en el género sudamericano Junellta trata- (Verbenaceae—Verbenoideae) 1. De pup ación miento infragenérico. Darwiniana 29: 371— Annals of the Missouri Botanical Garden . 1993. Verbenaceae, Junellia. Pp. 63-76 in A. L. potes Flora de la Provincia de Jujuy c . Verbenaceae. Pp. (editor), A d ia Vol. t del a i pianos Air E. Brandham. 1993. The taxonomic signifi- ance mes number in Junellia (Verbenaceae). 3-150. Kew es ay i la de Romero € S. Martínez. 1995, P "e conserve e name Junellia Moldenke. Taxon 44: 6394 Brummit, e m 199 the Report of Committee for 863-872 n 4T. n 4T: t Cabrera, A. L. 1957. La vegetación de ja Puna Argentina. TEM iuret née 11(4): 317412. 1978. L : Ni de relaci "ones con la ve ue ión puneña. Pp. 329-343 en E Troll & L. Lauer (editors), Geoecolo- gical A eds between the Sr Temperate Zone and the Tropical Mountains. d us schaftliche Forschung, XI. is Steiner Verlag aden. Willink. 1973, Nake de América Latina. Serie m Biología, monografía 13. Organización de Estados EE Washington, p. C. Covas, € . Schnack. 1946. Número de cromosomas en Antófitas ^s la Región de Cuyo (Repüblica Argentina). Revista Argent. Agron. 13: 153-166. Patagonia y sus altoandina y los & . 1947. Estudios cariológicos en Antófitas. Revista Argent. Agron. 14: 224-231. & J. H. Hunziker. Antófitas. Revista Invest. Agric. de la Peña, M. R. € F. Pensiero. 2004. Plantas Argentinas. Catálogo de Nombres Comunes. L.( . Buenos Aires. uds 7 O. 1976. ia ‘08 de Verbena tridens ccharis patagonica Hook. & Arn. y Adesmia bo- ilh Hock. f. Anales Inst. Patagonia (Chile) 7: 163—167. Fernández Cas J. & 1993. 1954. Estudios cariológicos en 49-253. Gamarra. Lagascanum. Fomueria 36: 67-108. Herbarium Greuter, Neill, R. Barrie, H. M. Burdet, V. a T. E B. 4 ja rcd P. C. Silva, J. Skog, P. Trehane, N. J. Turland & D. + Hawksworth. md Inte Sonal I Code of naan al N Louis Code). I m Veg. 138. Holmgren, P. K., . Holmgren & L. C. Barnett. 1990. Index en ih ed. Regnum Veg. 120. Hooker, J Jackson. 1977. Index Kewensis ens bonnm LIL, 3rd ed., O. eltz, Koenigstein. Hooker, W. J. 18 reprint. ip. On the species of m genus cn and a. Bot. Misc. 1: 159-175, t Niederlein. [ Expedició ión idi Río >67. ly atle ia P .& Pa Negr uo Botánica 2: 26 Macbride, J. F. 1960. Verbenaceae. Pp. 609-721 en Flora of Per Id Mus. Nat. Hist., Bot. Ser. 13, Vol. 5(2). ] 1905 . Parte V, Flora Patagónica sect. 3: "ug Um —Compositae. Pp. 595-810, pl. 21-27 en W. B t, Reports of the Princeton Univer eren 1896-1899, Vol. 8 (5, Princeton, New Jers Martínez, S., 5. Baltar dr M. E. Múlgura. 1996. a de las inllorese 'encias en Verbenaceae, Verbenoidea I: Tri Verbeneae. 1-17. Minoprio, J. sity Expeditions to 3). Princeton University, Darwiniana 34: H. Naves é Abdon. 1967. Factores ecológicos que ipn la trichiniasis silvestre en oeste de San Lu de Mendoza. Anales Soc. Argent. 183: 19— el Ci. a) oe Cientifica del INTA, Buenos Aires, . N. Corre 8(6). Colección Científica Moldenke, H. N. 1940. Contributions to the tropical South America. I. Verbenaceae. 432. flora of extra- Lilloa 5: 369— c eee rs to the flora Ea extra-Lropical Sowth e VI. »enaceae. Lilloa 10: 336-348. . Ma terials toward a De of the genus Diosteñ. Phytologia 7 I-3]9 = e T > sixth. summary of the Verbenaceae, Avicenniaceae, Stilbaceae, Chloranthaceae, Symphorema- ceae, Nyctanthaceae and Eriocaulaceae of the world as to valid taxa, E distribution and synonymy. Phyto- logia us s 2. — —— & A. L Moldenke. 1949. The rec O and of of Stilbaceae, and ebd eae ar- ranged according to genera and species. Supplement 1. Anales Inst. Biol. Univ. Nac. Auton. l. 20: 1-15 Monticel li, vernac WEN names me mbers the erbenac "ede, Avicenniaceae, Mexico, Bio El V. 1938. Sas "ones fitogeográficas de La . Lilloa 3: 251-382. Morello, 3 1958. E Opera Lilloana 2: Muñoz a provincia fitogeográfica del Monte. 5. c Pizarro, C. Sinopsis de la flora chilena. "iones P T bem de Chile, rw d de Chile. Pili F a . Atacam, 1-62, pl. 1—6 E a En Catalogus Praevius Plan- e ad Tara Anales Mus. Nac. pacá aderico pl Santiago de Chile, Bot. 8: Plantas nuevas Chilenas de las familias que cor ae al Il V de la obra de Gay. Anales Univ. Chile 607—625. Poggio, . Greizerstein & S. Botta. cromosómicos en Verbenaceae. Bol. Soc. 542—545. — 1988. Estudios Argent. Bot. 25: Pool, A. 1993. Verbenaceae. En L. Brako € J. L. Zaruechi (editors), a of the Flowering Plants and wt sperms of Peru. Monogr. Syst. Bot. Missouri Bot. Gard. 4 1164—1184 Hahn, K. 1960. Chromosome American angiosperms. Bot. Tidsskr. 56: 117— Raj, B. 198 contribution to the a E! of Verbenaceae. Rev. Palaeobot. " Palynol. 39: 343—4 Roig, F. A. 1970 [1971]. Flora y vegetación de E reserva LU de Nancufian. Deserta 1: 21-239. l Medicinal Mendocina: Las Plantas Medicinalss y Aromáticas, de la Provincia de Mendoza Universidad Nacional de numbers n some South Cuyo, Mendoza, Ruíz Leal, A. 1972 [1975]. Deserta 3: 1-296. Sanzin, R. 1919. Las o Contribución a la Flora de MN ge :. Ci. Argent. 88: 9! Schauer, J. C Ve cane 'eae. Pp. 522— 700 i in A. L. de Candolle a Prodromus, Vol. 11 Schnack, B. & G. Covas Flora Popular Mendocina. 947. Fauda Cariológicos en Antófitas. Poena 1-10 Soriano, A. . Los Dau: s hi la Provincia Patagónica. en Invest. Agric. 10 -348. 1956b. Aspectos les | y n) de la patagónica relacionados con su estado y capacidad de recuperación. Revista Invest. ic. 10: 372. 349 vegetación Agric. . 1983. Desert and semidesert of Patagonia. Pp. 423— 460 N. E. West (editor), Temperate Deserts and Semideserts. Elsevier, Amsterdam. Volume 95, Number 2 2008 Peralta et al. Revisión py Género Junellia (Verbenaceae) CE C. 1825. i sadi En C. Linneo, Syst. Veget., ed. 16, 2: 747—165, Góttin one oso, N. S. 1949. Des ane dos Verbenas patagónicas. niana 8: 481-487. 1957. El oe ro “Neosparton” Darwiniana 11: 163-19 68. Notas Ens nuevas o Jua dd Darwiniana 14: 6 . Los géneros de Verbenáceas de Sudamérica (Verbenaceae). Verbenáceas Sudamericanas 638. extratro du Darwiniana 18: 295-412 "E G. G. 1845. Verbenaceae. Repert. Bot. Syst. 4: 134. LI Weddell, H. A. 1857-1861. Chlor. Andina 2: [1]-361, pl. 43-90, Par APÉNDICE 1. Lista de las especies aceptadas del género Junellia. 34. J. arequipensis (Botta) Botta 2. J. aretioides (R. E. Fr.) Moldenke 29. J. asparagoides (Gillies & Hook. ex Hook.) Moldenke 35a. J. aspera (Gillies € Hook. ex Hook.) Moldenke var. aspera 35b. J. aspera var. longidentata (Moldenke) Múlgura & P. Peralta 3. J. azorelloides (Speg.) Moldenke 14a. J. bisulcata (Hayek) Moldenke var. bisulcata 14b. J. bisulcata var. campestris (Griseb.) Botta 15. J. bryoides (Phil.) Moldenke 16. J. caespitosa (Gillies & Hook. ex Hook.) Moldenke 30. J. cedroides (Sandwith) Moldenke 36. J. cinerascens (Schauer) Botta 4. J. congesta (Tronc.) Moldenke 17. J. connatibracteata (Kuntze) Moldenke 5a. J. digitata (Phil.) Moldenke var. digitata 5b. J. digitata var. integerrima (Botta) Botta 31. J. echegarayi (Hieron.) Moldenke 18. J. erinacea (Gillies & Hook. ex Hook.) Moldenke 31. J. hystrix (Phil.) Moldenke 19. J. juniperina (Lag.) Moldenke 25. J. lavandulaefolia (Phil.) Moldenke 38a. J. ligustrina (Lag.) Moldenke var. ligustri: 38b. J. ligustrina var. lorentzii (Niederl. Moldenke l. J. micrantha (Phil.) Moldenke 6. J. minima (Meyen) Moldenke 20. J. odonellii Moldenke 7. J. patagonica (Speg.) Moldenke 26. J. pseudojuncea (Gay) Moldenke 39. J. scoparia (Gillies & Hook. ex Hook.) Botta 27a. J. selaginoides (Kunth ex Walp.) Moldenke var. selaginoide es 27b. a ex ei ) a inoides var. illapelina dne. Botta ies & Hook. ex Hook.) Moldenke 8a. J. spathulata (Gillies & Hook. ex Hook.) Moldenke var. pala 28b. J. spathulata var. glauca (Gillies & Hook. ex Hook.) B 22. J. spissa (Sandwith) Moldenke 9. J. succulentifolia (Kuntze) Moldenke 32. J. tetragonocalyx (Tronc.) Moldenke 10. J. thymifolia (Lag.) Moldenke ls 23b. J. toninii Mülgura ii var. nulinaddes (Speg) P Peralta & tridactylites (Lag.) Moldenke tridens (Lag.) Moldenke trifurcata (Phil.) Moldenke ulicina (Phil.) Moldenke . uniflora (Phil.) Moldenke APÉNDICE 2. Indice de colecciones examinadas. Cada espéci- men es € stage apo el nomoi del primer colector en el caso en sionistas partic ipen de la colección. En el caso de no poseer número a poteg ción se e mga eli nureto nde herbario. Be indie entre parentesis el número de orden de la especie a que corresponde (ver Apéndice 1). Abadie 748(9), 986(20b), 1152(11); Adler s.n.(35a), s.n.(17); Ager 580(33), 581(18); Aguirre 455(21); Albornoz 11(21); Allende 222(39); Ambrosetti 1591(35a), 1257(18), , BA 27/887(21) 3133(28b) Ameghino 19(10) ), 190(33), s.n.(20a), s.n.(1), s.n.(24), 1135(24), s.n. LP 1000 — LP 10008(11), s.n. LP 10009(11), s.n. LP 12990(7), s 12995(7), s.n. LP 13000(24), s.n. LP 13001(24); Fo 114(17), 117(20b); Anchorena s.n.(9); Ancibor 1(6), 31(21). 48(2), 51(2), 102(7), s.n. BAA 445 RI s.n. BAB 4458(20a), s.n. BAB 4480(23p.p.), s.n. BAB 4480(8p.p.), s.n. BAB 4630(7), s.n. BAB 5788(24), s.n. BAB n s.n. BAB 90130(17), s.n. BAB 90141(16), s.n. BAB 90212(20a), s.n. BAB 90222(28b), s.n. BAB 90250(38b), BAB 90263(30), s.n. BAB 90289(17); Anderson 1664(21); Andrada s.n.(37); o 8(28b); s o, SGO 78425(33); d 4(21); Apochian 58(37), 183(37), 200(18), 195(21), 248(37); Arancio a 657(29), 662(13), 10139(31), 10147(31), 10292(31), 10355(31), 10655(31), 94183(26); Araque s. 1084(19), 1130(19); Arenas 84(34), s.n. BACF 954(29), s.n. BACP 1723(6); Arriaga s.n. BA 75895(21), BA 75881(21); Arroyo 16(20b), 66(17), 67(10), 101(20a), 114(10), 116(20a). 130(33), 131(20a), 14501) 182(20a), 194(7), i 216(20a), 256(7), 283(38b) 329(31) 3. 337(1), 341(20a), 355(10), 376(38a), 417(31), 418(20a), 515(5b) 503(6) : ; 1306(31), 2589(6), 81273(13), 81603(13), 81296(39), s.n. S T TBPA 2178(8), s.n. T s.n. TBPA 2455(20), s.n. TBPA 2 ere ), s.n. TBPA 2760. s.n. TBPA 2583(3); Asensi s.n.(7); Asp 125(9), s.n. BAB 64416456(9); Asplund 5882(21); (e de la e s.n.(33); Ayarde 350(5a) BA 34771(7); BAA 11625(17), BAA 12577(21), BAA 16518(3), BAA 16518(3), BAA 17794(35a), BAA 17795(17), BAA 18527(20b), BAA 18546(20), BAA 18548(16), BAA BAA 18650(20a), BAA 18817(11), , BAA 24780(20b), BAA 24781(1), 65(35a), BAB 9882(35a), 90259(38b); Baer s.n. LIL 31617(19); Baeza 466(31), 649(31 254(14a), 1180(14a); Ball s.n.(5a), 5973(29), 6000(6); Barkley 19 Ar 743(35a), 19 Ar 758(6), 19 Ar 852(19), 9232(6); Bartlett 19361(35a), 19363(35a), 19426(19), 19429(35a), 19434(21), 19448(35a), 19474(21), 19518(17), 19935(21), 20482(14b); Bártoli 9/03(9), 21(20a), 25(20a), 25/02—2(20b), 26/03(20a), 47/03(32), 50/02-2(20a), 55/03(9), 58/02-2(38b), 52/02 BAA 24787(20a), 65/02- ivl > > ivl 2(10), 69/02-2(32), 114/02-2(38b); Barros s.n.(27b). 1377(27a), s.n(36), 2233(27b) 2233(28b); Bastión 602(14a); Baumann 61(5a); Bayer 2019(2); Beaufilo Annals of the Missouri Botanical Garden Z C n wy j1 2372 ac ino 68 4(3t 3b), 690(17), 696(32). pos X11). wn »15(20a). 547010). 549(10). 5500). cy 55((11). S59 11), 56101). 562017), s.n. SI 26674(22) SI 26673(20b), s.n. SI 26676(17): s 2033€ Bridarolli 2207(9). 4321(21): Bridges s.n.(39 s.n(26b), 4582884). 12202890). 1353(27a): Bruch LP 832519) Brun 703(32). 88821). p 39(20b). 23791 10). 2421011). 25890 1). 2774(1 1): Buchtien s.n (6), s.n.(39), uL [v 7455(35b). 7487(29). s.n. St 1496(6). s.n. BA 31/91 H(0); Burkart | 2149(21), 2152(35a). 7455(35b). | 8392(21). 341539), 148550), 1586421), 15927(35a). 15950(21). 19220(17). 201 1LO(14b). 20784(14b). 22065( 14a), 22071(19), 26513(9), 29642(14b). s.n. SI 11728(29), SI 11950(29), s.n. SI 14247017), SI 14248(28b). s.n. 14253284). SI 14254(28a). SI 14255(28a). SI 14256(28a). s.m. SI | poe SI M SI 14260(38b). SI 14261(16), SI 14263(Q8b). SI 14204(29), SI 14205.(28a). SI | aa ne SI 1426€ m. SI 14275(28b). SI 14279(24), S.H. i 14278(38b), SI 142850(21). SI 14851(38b), SI 14992(17). sn. SI 2207121): Burkart. S. 1268(20a): bs "ster. 12(11), 23(33),. 450 b),. 10301), 113(33), 144(20), 344411). 652701). 6136(20a), 6437(17), 60439(11), 6526(17). 0529(38b). 0532(33). 6533(20b). 65347) 6535(20a). 11882(20b). 11898(11), 36530(38a). s.n. BAB 119 21(33). s.n. SI 3446(33) . E. 516(27a): Cabrera 18(17), 22(33), 30(21). 38(38b). : 5 " da), 121(38b), 129(2), 981(21). I913(21), 3137(19). 10(38b), 438221) 4802(38b). 4833(10). 4845(17), p 1847(1 t 7677(29), 791321), 7932(35h), 8216(21). 8244(21). 8270(2). 8401(5b), 8788(5a). 8952(5a). 9023(21). 9024(35b). xt (6). 9277(6). 9346(2). 9491(29), 11015(20b). 11022(20b), 11049(20b). 11105(28b). 11108Y28b), 1101506). 1104921), 11071007). 11169204). 1117007. 1331849), 1334 10(35)), 13983(6). E413 7(DIda). 15015(29), Io718(Ida), 1677421) 17520029). 17536(6), 17580(2). 17594(2). 170604(6). 18527 (21). 1859217), 15605(35a). 18652(17), 18676(17). 18080(38a). 18711 (28b). 187! ene 19435(17), L943 7(38b), 19466(28b), 19743(9).. 20693(6). eiii E 60 ta), f- 119(29), 21508(6), 21805(35b), 2181721). 22034(21). 22410(35b), 22440(21), 22804(28b), 22001 X16). 2 1386013). mm 520(35a). 24522(18). 24616(39), 24650(29). 24680(19). 24736(19). 25857(11), 26378(014a), 26378 p.p.(29), 2703713). 27214(29), 27417(29), 27141(19), 27214(29), 30048(13), 30129(18), 30479(21), 30564(14a). 30600(35b), 3074 7(14a). 30754(35b), 30764(35b). 30801044), 311: 30(13), 31276(13). 3130337). 3150329). 31618(29), 31698(6), 31713Q1), 3285 sol: 30), 32804(17). 32867(9), 32881(24), Ds 32904(7). 232971(0). 33009(17). 3301 1(9), 33012(17). 33013(20a). 33028(20a). 233030(17). 33034(38b). 33043 20a), 3305807), 33070(9), 3307 7(1 1), 3311800). 33125 3313400). 33135(20a). 33151(33). 3315701 D. 331580 1). 3310691011) 83195 (10), 331960 | ), 233249(11).. 33251(10). 33252(24), 33250(11). 33261 5(10).. 33264(20a), peers (20a). 33277 (20a). 3319320) . 33213(33), 3324: 33252(7), 33253(20) , 33257(1) | 33284(17), 33 286 (323). 33289(38a) 33308(11). 33315(21), 33328(28b), 33382(284). 33391(24). 3339701). 33420(24). 334414). 33460(298hb). 323409(20a). 33474(28b), 33478(10), 33492(28a). 34149(33), 3415401), 341550 1D). 34157(28b): Cáceres 250019): Cajal 121) 21016) 1256543). 136006): Calcagninto s.n... BAB 13496(38b): Calderón 1122(19). 1069(39): Canevari s.n. BA 6951120) Cano s.n.(38b). 112(38b). 330017). 500(38b). 532(0). 1099(38b). 1227(35a). w E o Volume 95, Number 2 2008 Peralta et al. Revisión del Género Junellia (Verbenaceae) 385 4679(35a), 8312(34), 11872019); C E 4 31019): Capare n 55(19); Cardenas 285(6). 312(21) 740(6). 4912(2 Cardoso s.n. BA 23695(20a), s.n. BA s s Carette 5(28b), 62(16), 150(35a), 325(28b) 329(38b) 330(38b). 33129). 332(29). 334(19) 335(17). 336(106) 337(E 1), 3380). 3045(21). s.n.(33). s BA 23670(39). s.n À i (19 LP 86591(33), s.n. MERL 2564(6), s.n. MERL 3732(35a), ipo SI 3342016), s.n. SI 3392 "o s.n. SI 3409( ID. SI 27715(39); Castagno 34(28b): 80/282), 160(24), 387(20a). POLA 3541(9). 20160(17). . BA 6180(1), s.n. BA 6181(8), s.n. BA 7023(5a), s.n. BA 61 75(7 ) s.n. BA 6176(33), s.n. 31 77(33). s.n. BA 6178(20a). BA 7885 (2), 7886( ] 1 s.n. Castellanos s.n.(21). das 29), B A 2021 1(38a). A 25/2939(21). s.n. BA 202272550) BA eel BA pure ) BA ue 2 E BA 27/88921). - s.n. BA 2 E s.n. BA 28322 . s.n. BA 16962(5a). BA BA 51175(14b), sn. LIL 2(2 368501200), s.n. BA ant 560/19), s.n. BA ( 46963(35b). ( 108120), 2. 3 19(19), 244 "1 | 9). pr 3199(19), 3321(21). . 1421309), 16393(19), s.n. LIL . L1L31588(5a LIL 31595(5a), LIL 31 dae bu. ; 13032(14a). LIL oS bo u astro s.n. (16); Ci 12(5a), 14421), BA 25/1 515 21), p. A 26/1 72021); Catedra de Geobotánica 5377(29), 5581(21); Ceballos 84(6); 269660(11). s.n. MERL 30189(33). s.n. 34970(16). s.n.(1): Cekalov le: ; Charpin 20799(21), 25504)(28b): 25121). 411(35b). 11400(2)(S. E 3947(21 Cel s.n. 850). BE Citadini CUR s.n. 455(21); no Clemens s.n. ; Cocucci, A. 2702(29). 194(14a). a Claren 8281; Herb. Parques 2207(19), 2282(28a). 2287(24), 2 " 1133). 2512(33); Col- lantes s.n.(27a); C ma v 25(38b), 78(sp. 79/07). 112(1) 133(17). 138(16). 189(30), 957(9), 1290(4); Cordo 522/ 78(19), sn. SI 28455(19); Correa 332(20), 1715(20a) 2457(17), 2461(20b), 2485(38a). 2504(20b). 2514(38b). 2529(33). 2583(10). 2617(11), 2029(7), 2040 1/2(7) 2649(10). ed Ja), 2746(38b). 2796(8). 2811(20). 2957(11), X11), 806101), 3179(17), | 3185(38b) 3230(38a). vu 20). 3303(10), 3337(10), 3379(7), 3392(7 3393(20a). 3395(20a) 3407(11). 3411010). 3417(38a 3436(20a) 3441(7). 3568(11). 3573(10) 3618(38b 3002(20a), | 3665(7) 3073(9) 3682(17) 3699(20a), 31742(17). 3866(1 7001) 872(20a), 3932(38 3948(1) 2(: 1205(9) 1697(21 ), 17 l po 1750(38b), | A771(38b 1821(38b), 4884(20b), 5024(17). 6190(38b), 6239(20a 6243(11), 6250(7) spat one 6303(20a), 6397(38b 2(38a). 6470(24), 6473(11), i aay 6520(11), 6525(33), 6617(11), 6674(1), 6695(20b). 073300), 07306(20a). 6786(20b). 7045(20a). 7059(38b) 7186(20b). 7266(32). 7311602) 7327(20b), 7331(20b) 7333(20b), 7339(9). 7352(9), 7387(9), 7854(38b), 7860(17), 7868(17), 7879(17), 7885(20a), 7889(20a), 7901(17), 7901 1/2(30) 7907(9) 7911(20a), 7913(28b), 7956(17), 7965(17), 8541(381 8543(21). 8563(38a). 8637(17). 863921). 8645(21) 8662(35a), 8716(35a), 8731(21), 8748(20a) 8761(21) 8760(21). 8781(20a), 8799(20a). 8801(21) 8803(21) 8817(10), 8848(20b), 8851(10) 9095(1D) 9116(17) 9157(9), 9159(20a), 9397(9), 9417(20b), 9447(39), 9448(9) 9455(9). 9459(17). 9530(35a). 9592(17). 9596(21). 9602(21). 96106(35a), eer 38b), 9638(32), 9654(20b). 90665(9). 9760(28b) 981911), | 9855(11), 9980(284) 1008 (28b). ig 10159(20b), 10160(21). 10161(38a). 10162(21), 10163(20a), 10165(20b), 10169(20a), 10171(1), 10178(38b), 10196(9), 10212(20a), 10213(38b), 10214(20b), 10216(20b), 10219(20b), 10220(1). 10223b(20). 10227(33), 10229(20a), 10232(24), 10236(7), 10239(24), 10239(3), 10243(38b). 10240(10), 10249(1), 10256(38a). 10258(20a). 10258/b(10). 10260(10). 10261 (10). 10262 b(10), NUBE 10265(11),. 10268(10). 10221(20b), 10223(20a). 10269(33). 10269b(1 1), 13(20a). 10290(11). 10296(7). 10299(10), 10303(10), io dn 0). 10284(7). 10309(10), 10311(10), 10312011), 10317(38b).. 10332(1). 10355(10), 10358(20a), 10360(1 1), ), 10383(11), 10399(11), 10453(9) 10486(17). 10491(28b), 10492(28b), 10503(28b), 10507(17). 10513(9), 10514(21). 1051721). 12140(33), 19519(21), ' 94(3). 311(20), 610(35b); 674(35a), 907(19), 97121), 210421). 2105(19. 2647(38b). 3125(35a), 3146(19), 3415(28a), 15068(37), 19990(37); Crespo 1632(11), 1638(11), 1639(11), 1652(10). 1653(10), 1712(38a), 1713(10), 1714(33), 1718(10), S9 1760010), 10357(20a), 103 74(11). 10588(11), 10476(9), 105041). 10519(33). 20227(33); 10361011), 10366(1 1), Correl 17617). 1768011), 1909(28b), 2471(32). s.n. BA 47386(21) Cristóbal 36(29), 145(35b); Cuezzo - 14a). 866(14a). 1963(19), 1991(19), 2008(16), 1963019). 1991(19) 2574(19), 20M200(29), 20M7400(29). 20M24065(29); Cuming 282(28a); Cunquero 1054(21). Daciuk 25(38a), 30 p.p.(21), XXX111-26(38b). XXXIII-27(20b). 794(1), XXXIII-20(38b), XXXIII-28(20b), XXX- iMI-30(20a); Dalmaso 37(39). 250(3): D'Antoni 44(21): Darwin s.n(38a) s.n.(11) Dauben 11001); Dawson 13389) 1053(28b), 1060(21). 1079(39) 1210(2: 1242(28b) 1366(9), — 1400(21) o : ) Pu 1656(28a), 2 Azkue s.n. BAB 91360(21), s.n. BA 2077(9); De la Barrera 27(2 De Marco 118(10), 1396 38a), 195(38b). 2 Debandi De inani 450(14a); 1157821). s.n. 1156617): Del Vitto 27710(28a); Dell'Arciprete s.n. BACP 2303(21), s.n. BACK 2324(6); Descole 1419(19), 1728(17), 2128(17), 2427(9), 3525(19; Di lorio 76(35b); Diaz 1861—62(39); Diem ol): 2443(33), 3207(9), 3580(28b), 3592(9) Diers ; Dillon 3336(19); Dimitri 4314(28b). s.n. Parques 5086(28h »), s.n. Herb. Parques 7208(9): Dinelli s.n.(35b), s.n. LIL 12890)(19; Dollenz 1368(20) Dominguez 29(20), 389(33); Donat 56(38a), 108(33). 11101). 124(11), 138(24), 1870), 201(10), 220(38a) 221(7) 281011) 14308), 143(20a); Duek 2314(12) Duqu: (46(38a); Dusén 5314(20b), 5333(38a). 5399(38a). 5444(33). 6023(20b). 6080(1), 6236(7). s.n. SI 3431(8), s.n.(8). s.n. LIL 31570(33) Eca 4(5aj) Egeli 2240(27b) 2566(27b). 2728(31). 2825(34); Ehrich ae 1), 341(29); Elkin s.n.(5a), 19(21): Ellemberg | 4302(2 4716(6) Elliot 3727b): Escobar 386 Annals of the Missouri Botanical Garden s.n.(20a), 60(20a), X-149(34); 62(10); Eskuche 168(9), 736—4(38b), 915(9), 938(20b). 1291(17), 1292(28b), 1293(7), 1294(20a), 1870—12(20a), 1878-79(7), 1898-3(33), 1903-13(10), 1912-1(1); Espina 18817(33); Eyerdam | 10073(39), 22121(34), 23445(17), uq um 23804 p.p.(24); Ezcurra 37(9), 173(11), 183(28b), 314(33). Fabris 817(21), 858(38b), 876(20b), 877(16), 897(38b), 1280(28b) | 1309(9, 1331(14a), Escomel 68(19); Eskobar <= = Faggi 147(20b), BAA 7490 5(21). s.n. BA 77756(29) Ferrando 12(20); Ferrando s.n. SI 3837(11); Ferreyra 14261(34), 18706(34); Feruglio 63(7), s.n. BA 23733(33), s.n. BA 30/1886(33), s.n. BA oe sn. BA 30/1888 p.p.(1), BA 30/1888 p-p.(11), s.n. BA ee aay s.n. BA 30/1890(24), s BA 34756(20). s.n. BA 3 PDN, . BA 3475911), s BA 34760(7), s.n. BA 3 s 61) s BA 347717); Ficbrig 2616(2); Fiedler 47(25), 48(19), 19(28b), 50(28a), 91(13), 101(24), 156(24); Figueroa Romero 1512(19); Fischer 24(21), 25(38b), 89(35a), 183017); Fisher CONC ; Flossdorf 50(5a), s.n. SI 28454(5a); , 4365(9), — Fortunato = [91] BA nem n s.n. LP 544(21); Frey de Jones 38(9); F 3(2). Gallardo 67 j 7), per 1222(33), s.n. BA 23593(20a), s s.n. BA 591 Garaventa 5409(28b), 3 . 5383(28b), 5414(28b), Ere 6582(39); García 214(19), 490(21), 553(29), CS: 8(19); Garino s.n. BAB 91769(17), s.n. BAB 01790(28a), s.n. Bab 91771(39), sn. BAB 91814(19); — 1150 D, 21); Gamerro 4459(27a), 5794(33), Garrido 18(38b), 119(20a), 452(20a), 447(20a), 473(7), 88(20a) 515(11), 584(11), 596(7), 686(20a), s.n.(21); Gaudichau 68(27a); Gautier enn Gay 310(16 980(13), s.n. US 1706333(36), s.n.(19), s.n.(26), s.n.(39), s.n. 293(27a ); Gentili 335(11), 409(33), 500(39), penes . 565(9), 574(20a), 567(7), 659(17), 680(9 ) t 8170 866(9), 887(7), 980(7), 1091(9), 1092( 64(35a), s.n.(35b), s.n.(9), 126(39), s.n. BA 23711(28b); Germain s.n.(28b), Gert 61(11); Gerth Giareli s.n. SI 3357(20a); BACF m Gillies s.n.(19), s.n. a (16), sn (35a), s.n., p.p.(37), s.n.(18), s.n.(39), s.n.(21); Girola 11780(21), s.n.(35a); Gómez 2492(21), 5840(13); González 26(14a), 725(38b), 749(17), 763(0 ) 922(28b); González, J. A. 174(21), 205(21); Gonzales ouk 13(9); Nada EE Goodspeed 16101(27a), ; 2 2397601), 23080(33). 24355(11), 24378(20a), d 30388(19) P Grandjot s.n.(25), s.n.(39 s ^ 61(27b), 730(27a ), 2639(13), - dona TETA 2038(33), 2167(33), 22031 ih ee 2275(11), 2348(1) = co JU 2 — pg = Y co ¡2 A — w o = ; Guaglianone 155l 3253(28b) Guerrero 54(17), 15801). 3810 1): 464(33); te Paci 15656(13); Guiñazú 169(21), 188(33), s.n. BA 31/1717(1 1); Guiraldes 24(21), 30(28b); Gunkel s.n. T 70300(1( ). Haber 109(5a; Haedo 06973(19; Haene 490(21 1159(21), 1621(29), 1737(29), 1771(21), 1785(6), 2002(16). - 2004(37), 2021(21), 2031(18), 2092(16); Hager 57(20a), 96(17), 135(9), 159(7); Hajardo s.n. n His. Nat. San Rafael 3615(38b); Halloy e 55(14a); Hammel 5959(21); Harvey s.n.(27a Hatc RD s.n. 1896/99(33); Modi 287(13), s.n.(13), s.n. SI > BA 23602(38b). s.n. 48(0), s.n. Hayward : Helandér. s.n. BAB 60928(1 1): . Mun. ro 10259(11); Hernández 2(5a): Herrera 966( 18) Herting 416(28b); Hicken 62(16), 66(17), 503(1), s.n. SI 3403(38a), SI 3430(21); Hiel 390(6), 391(6); O & Niederlein s.n. BA 23668(5: 771(6), s.n.(19), 482(19), s.n. BA 54(34), 76(19); Hjerting 69(29), Hoffmann | 8925(34), | 8927(19), 8955(34); Hogberg 110(33), s.n. BAB9467(20); Hokenaker 808(28b); Huajardo s.n. Herb. Museo San Rafael No. 3294(28b); Hueck 272(21), 360(21), a); Hieronymus Humbert sn. SI 18156(13), SI 1818803); 21082(21) Hunziker, A. T. 3079(35a), 11659(35a) 13079(35a) 14225( 35a), 14402(35a), 16359(35a), 18390(35a), 22859 (19; Hunziker, J. H. 1868(19), 1962(29), 2128(21) 1460(5a), 2200(21) 3092(29, 3071(29), 3108(39) 3145(39), 4047(35a) 4095(21), 4160(5a), 4211(35a) 4239(35a), 4276(35a) 4897(18), 6372(28b) 6448(21) 7802(29), 7835(28a), 10427(6), 11110(35a), 11130(19) 1113567), 11142(37), 11235(37), 11262(39), 1 1017) 11366(13), 11391(35a), 11272(19), 11288119), 11289(19) 11622(35a), 12548(35b). 1313921), 13140(35a), 13170(35b, | 13176(21); Hurrell 781(29): Hutchinson 42(39), 1245(19), Reus 4). Ibar s.n. SGO 54788(33), SGO 42535(33); IllÁn 7(10), 31(17), 42(7), 53(1 n. 56(11), 69(17), 71(17), 75(17), 99(17), 122(20a). 179(20b), 240(1), 1132(11), 3764(17), 6504(1). 6505(20b), 6506(38b), 6508(11), 6509(7), 6510(20b). 6518(1), s.n. BA 23694(20a), s.n. BA 23702(20a), s.n. BA 23753(20a); Irisarri 32(38b), 263(38b), 267(11), 268(34), 273(3), 274(3), 481(1), 482(11), 485(7); Iter : . 145(33), 247(33), 275(33), Patagonicum 320(38b), nn 71501), ~ E ~ A o = s.n.(7), s.n. “BAB 651 107). James 439(33), 440(11); Jaffuel s.n.(28b); Jiles 406(27a), 1402(27a), 1509(27b), 1535(36), 1598(36), 1919(27b), 1920(27b), 1971(26), 2243(27a), 2437(27a) 2476(28b), 2857(27b), 2898(29) 905(26), 2927(160), 2957(13), 2927(6), 3193(27b) 3266(27b), 3375(26), 3499(36), 3506(27b), 3736(36), 3762(26), 3780(28b). 3854(27a), 4088(39), 4160(28b), 4278(29), 4387(26), 4393(39), 395(28b), 4450(39), 4451(39) 4551(39) 4567(25). 4727(26), 4821(39, 5047(27b), 5572(26), 5913(29) 3071(39), 6103(13) Pee 6406(25), | 0474(13), 6509(26); Johns 82-56(6), 82-34(6): Johnston 4732(15), 5380(27a), 5509(27a) 5618(27a), 5937(16), 6050(13), 6253(29); Jones 38(11); Jordan 56(6); Jorgensen 115(21), 1027(19), | 1403(39),, 10612(29), 1614(35b), 1732(21) (< = > 173(14a), s.n. BAB 23483(35a): Joseph 814(39) 4996(36); Juárez 2283(35b). Kalela 198(21), 340(35a), n 1876(1), 1895(11), s 1914(10), 2426(38a), 2440(20a); Kaiser 2 esling 9), 271(21), 342(29), 354(29). 384(29 ace 651(29), 2953(21), 3108(29), 3116(39), 3267(3 a 3227(13), 3288018), 3368(14b), 3433(29), 3513(14a). Volume 95, Number 2 2008 Peralta et al. 387 Revisión ud Género Junellia (Verbenaceae) 3522(14a), 3524(14a), 3642(29), 3654(29), 4030(35a). 4108(39) 4160(19, 4165(39, 4196(19) 11(18 4407(29), 4409(39), 4616(5a), 4629(5a), 40662(29 4671(29), 4674(6), 4774(21) 4810(39) 5302(14a 5246(21), 5855(14a), 6235(35a) 6325(29), 6411(35b 6498(18), 6503(21), 6558(16), 6683(39. 6735(13 6865(29), 6902(18), 6962(29), 7034(14a), 7077(29), 7260(24), 7343(21), 7365(16), 7419(37) 7453(29) 7545(16), 7710(29), 7774(13), 7872(35a), 7885(37) 7936(18), 7980(13), 8699(24), 8757(21), 8782b(21) 8821(13), 8834(5a), 9085(21) 9086(18), 9109(29) 9208(13), 9253(29), 9256(18), 9257(21), 9432(13); King 37(19), 135(19, 249(21), 326(29), 327(16), 341(18), 9391(17), 10355(21), s.n. LP 890400(21); Kofalt 1(8), 2(8). Kopta s.n. herb. Cat. Geobotan. 10073(13); Koslowsky 40(33), 41(7), 42(10), 152(33), 191(11), 194(20b), 326(11). 9386(20b), 10247(10), 10249(20b), 10259(20b), 12335(20b). 12245(7), 10247(10), 12304(10), 12334(10), 12336(10), 12337(33), 12338(10), 12354(20a), BA 65100(20a), s.n. BA 23657(10), BAB 2337(33), 89-85 BA 23667(7), s.n. BA 23732(33), s.n. SI 3843(10), s.n. BAB 12355(17), s.n. BAB 12354(20a); Krapovickas 3130(21), 3655(35a), 3658(21) 3703(28b), 3712(10) 3720(39), 3751(32) 3825(17) 3833(11), 3807(35a) 4338(33), 4339(10), 5072(35b) 5169(19, 5189(19), 5210(14a), 5216(31), 5225(29) 5233(19), 5315(39), 5318(29) 5356(5a) 5388(21) 5480(35a), 5563(18), 5568(16), 5705(19), 5706(19) 5882(35a), 5932(35a), 6130(19), 7751(14b), 14184(16) 14584(16), 14595(19), 14597(19), 20537(19), 21908(14a), 41(21), 21959(35b), 21960(14a), s.n.(16), 22393(21), 22394(17), 22488(21), 22520(38b); Kuefer s.n. MERI 32394(35a); Kühn 60(33); Kuhneman s.n. BA 37375(11), BA mte 1), s.n. BA 81544(33); Kuntze 5(29), 6788(14b), (19), s.n. LP 8468(28a); Kunze s.n.(17); Kurtz 32(39), 86121), 3320(14b). 3349(21), 5519(21), 5645(29), 5920(24), 7142(24), 326(14a), 9464(29), 9621(29), 9653(13 9781(21), 10012(21), 10962(19), 11492(14a), 11513(2), 11660(29), 13392(14a), LP 14402(35a), 15141(19), 15435(29); Kuschel s.n.(25) La Torre 1837(19); Lagigha 1194(16), 1208(38a), 1216(13), 1230(28a), 1280(28a) 131116), 1330(13 1374(28a), 1375(1), 3302(28b), 4831(17), 5098(28a), 6543(17), 6547(17), 6640(28a), 6809(24) 6814(28a 6869(28b), 6874(28a), 6878(24), 935(28a), 6895(28b 6062(28b), 7228(19) 7267(18), 7229(28b), 7736(17) 7759/24), 8446(28a), 8488(28a), 8465(1) 8370(3), 8669(29), 8756(29), 9173(28a), 9174(1), 9976(17); Lahille s.n.(20b); Lahitte s.n. Herb. Parques 7296(33), 7305(11), 7491(33), 7492(33); Lahitte R., s.n. BAB 51976(11), s.n. BAB 51985(1), s.n. BAB 51995(11); Landrum 8404(33); Lammers 7597(27a); Larguia s.n. BAB 8888(9); Larrain 652(34); Latorre rs Latour 1812(33), s.n. TBPA-fIT 109 1/2(8), s.n. TBPA-fIT 154 1/2(20), s.n. TBPA-fIT 182 1/2(20), s.n. TBPA-fIT 319 1/4(9), s.n. TBPA-fIT — 314 (20), s.n. TRPA-fTT 350(8), s.n. TBPA-fIT 350 3/4(8), s.n. TBPA- [IT 374(8), s.n. TBPA-fIT 823(20), s.n. TBPA-fIT 871(20), s.n. TBPA-fIT 913(20; Legname 3(19, 41(21) 14(35b) 216(35b), 156(29), 5162(21) 5234(14a), 5254c(35b) 5510(19); León 2254(38b), 2623(1), 2508(11), 2533(20a), 2534(20a), 2548(10), 2550(11), 2559(11), 2685(11), 3220(7), 3454(9), 4009(7), 4107(21), 4185(11), 4337(24); Lewenber- ger 3572(21), 3679(20a) 3720(33), 3760(1), 3799(39), 3907(28a), 3950(39), 3961(21), 4101(8), 4120(7); Liberman 202(2), 287(6); LIL 4984(29) 4697(35b), 5851(14a), 5877(29), 11958(29), 12973(29); Lillo 5522(6), 5623(6), 1114019), n Loos 32(29), 60(29); d s.n.(25); López s.n.(27a), 126(21); Lorentz s.n.(19), s 3394(14b); Lourteig 388(21), 672(39), 821(21), 829(28a), 851(21), 873(35a) 874(21); Loyola 13(31) 21(31); LP 402(39); Eos Fa s.n. Cat. Geobtánica 5360(29); Lutz Mix ae 19090(8); Mahu 5295(39); Maldonado 290/21). 683(28b) 726(17); Mandon 196(6), 526(6), P536(6), 796(6); Marq-mold 23(35a); Márquez s.n.(21), 1); Marti- corena 23(19), 129(27a), 203(6), 233(27a), 270(27a 306(27a), 325(2), 346(27b), 401(36), 427(31), 464(27b 580(39), 613(39), 623(28b), 658(25), 693(28b), 695(39 991(25), 1072(25) 1424(27a), 1549(27b) 1625(27a 1723(27b), 8359529), 83478(13), 83497(16), 83603(29 Martin(38b), 194(21), 232(21), 352(21), 425(21), 435(5a). 457(5a), 511(21); Martínez 455(33), 6051(17), 6196(17), 6197(17); Martínez Carretero 1533(19), 1608(14b), 1633(29), 2091(21), 80(1), 6193(7), 6226(28b), s.n. MERL 50977(21); Martínez Crovetto P-24(33), P-52(11), 6051(17), 6153(28b); Martínez Garrido 455(33), 678(8), 681(11), 701(38a); Matthei 271(19), 314(34), 323(34); Mazzuconi 1035(9 Medan 759(20a); Medrano 7561(38a), 7659(12), 77331 TIAM), 773133), 7764(38b), 7787(33); Meglioli 37(18); Melis 20 Mz 084(19), s.n. NYBG 3441(29), 88(21); Melis y Paci 279(6); Méndez 1153(8), 1154(33); Menéndez s.n. BA 54342(28b); Merck s.n. SI 3448(6); MERL 194(13), 216(21). 438(39), 668(35a), 797(28b), 1023(39), 1148(35a), 1169(21). 1208(38a), 1419(29), 2087(28b) 2103(37), 2247(21), 2378(38b), 2556(29), 2564(6), 25068(20b). 2868(37), 2913(37) 3002(19), 3250(35a), 3520(35a), 3543(35a), 3566(35a), 3631(38b), 3632(39), 3850(28a), 4124(19), 4397(37), 4713(37), 4788(19), 6253(35a), 6955(21), 7223(13), 7520 : 797/146( 38b). 7806(28a), c 3 3 > ry m EN aoa BQ aA a = SUN NS O 5 - EN © 32% w œ Z ~] A - 1619(21), 7701(28b), 7746(28b), 7774(28b), 7966(35a), 8167(21) 8220(37), 8456(35a), 8498(35a), 8803(35a), 8950(35a), 9617(35a), 10242(37), 11054(28b), 11256(35a), 11425(19), 11765(28b), 11813(35a), 11841(29), 11954(28b), 13060(17), 13097(21), 13404(17), 14592(13), 14857(21), 15436(5a), 15583(1), 15692(38b), 15752(1), 16473(29), 16953(35a), 17014(6). , 17455(35b), | 17987(17), 18034(17), 20136(18), 20640(21), , 20883(28a), 21074(17), < , 21435(11), , 21698(19), 22553(35a), 23: . 23636(29), 23642(18), 2: , 23851(17), 23858(38b), 23972(11), 23996a . 23996(10), 23997(38a), 24076(33), 24090 = , 24120(11), 24124(7, 24165(11), 24146(10), 24199(38b), 24209(10), 24212(20a), 24213(10), 24222(1), 24305(11), 24328(9), 24229(10), 24551(24), 24582(28a), 24583(28a), 24649(28a), 24652(38a), 24657(20a), 24673(1), 24691(38b), 25559(20a), 25574(7), 25586(20a), 25697(10). . 25581(7), 25586(20a), 25604(1) 25660(38b), 25876(1), 25882(1), 25904(39), 26000(9), 26046(35a), 26128(20b), 26131(38b), 26177(11), 26911(11), 26209(7), 26232(7), 26256(7), 26359(18), 26345(9), 26359(20b), 26412(10), 26421(20), 26438(20), 26466(20), 26480(10), 26488(11), 26489( 26567(11), 26580(20a), 26595(20a), 26651 (20a), 26732(17), Ee E = ~] A orn IAS a E] BY o —T PK] Y o — qa Es n WO ——-— 000-10! ou0oc SS o e A n 3 04 Ne} o © E ESRA-S 28 16901(13). 15710(28a), D Annals of the dm Botanical Garden NW N L049(21), 1367119), 1580017 1), 161021), 163221). 1079/21), ^ ys 0731(37), s.n. BA 61933(13), s.n. BA 61935(21), s.n. BA 61937(19); Pearce s.n. (21); Pedersen. 13248(21), 13256(9), 1331721). 13332(35a). 141530). 14430683). 14489(1): Pedreros 88(31): Peirano 3283621). s.n.(19). s.n. LIL 32842(5a): Pemberton 12(33); Pennell 4162(19), 1318734) P ped 167; Pereyra 5881(29); Pérez Moreau R. A. 318(0), s.an. BAA 30/178(13), s.n. BA 30-179(16), s.n. BA 12704(19); s.n. m 1270609). s.n. BA 23354(28a), s.n. BA 45439(7). s.n. BA 45449(9. sn. BA 48730(9). 55121(16), BA 55134(29), s.n. BA 61935(21). s.n. Ba 69967(33); Pérez Moreau R. L. 3113(28b), 3139(20a), 3152(17), 3192(17), 3219(30). 3221(1), 3244(9), 3263(28b), 3315(1), 358 Pu 3078(9). s.n.(1): Perrone s.n.(7). s.n. BA 586 7(1). s.n. SI 20332(9), SI iir 30). s.n., SI 20339(9), BA 584360 1). A 28800(1), s.n. BA 61618(30), s.n. BA 01623(11), BA 7045316), s.n. SI 2033724), s.n. SI 20330(20a); Petersen € Hjerting 807(5a); Petersen 85(6), s.n. BA 45887(7), s.n. 15889(9); Pfister s.n.(27b). s.n.(33). 5433), 12119(33); Philippi F.. s.n. 8GO 54801(5a); Philippi s.n.(5a). s.n. SGO 42479(21), SGO 42522(5a). s.n.. pp.(Lectotipo SGO 54675)(15), s.n. SGO 54075 pp.(31)(5GO). SGO 54673(31), SGO 68374(31). s.n. año 1904(27a ). s.n. SIE 3447(13) Piccinini 1255(38b). 1300(9). 1332(20b), 1347(20b), 1532(38b). 1574(20b), 1580(20b). 1581(20b). 1652(20b). 2372(32). 3928(14b); a Sermoli 7428(33); Pierrot 7485(21)(K); Pisani 11533): Pisano 182(27b), 188(27b), 1396(39), 1399(39), ( psi Y > 1792 5b) 1793(5b). 4180(33) Plotnik 3(38b), 27(17); Poeppig s.n.(36); P rflanzenflur & al. 4120(7); Pozner 128(21), 262(19), 266(20a); Pozzi 6438(20b); Prichard s/ n(ll); Prina 1 a ). 1200(20a), 1227(35a), 1228(21) 1283(38a). 1293(16). 1334(17). 1345(20b). 1351(20a). 15560(35a). 1397(21), 1454(38b), 1456(16), 1488(29), 1494(28b), — 1543(11), 1566(24). 1577(16). 1625(29), 1781(21), |.1784(20), 179021), 1879(1), 1868(28a), 1907(28b), 1935(20a). 2122(28b) 2251(1), 2390(38b), 2429(24); Propal 87(20), no. 154(20), 339(9) Pujalte 30(21). 34(13), 106(5a), 241(13). (21), 60(17). 1247). 228(11), 351(28a). 352(21). 6060(4a). 758821), 765721 ) 814 121 ), FE e N a. 1622(35b), 1641(21), 1649(35b). 165621). eooo ), v x o 29731). 391(34), 434(2), 4069(5b), 500(21). 7 mes " 13608(34), 1404(21), 1550(27a), 1575(31), 158831), 158931), 1766(13), 2157(27a). 2441(25). 2487(27a), 291039), 3245(25). 3246(28a). 893(28a), 3378(6), 4272(27a) 4541(27b). 4659(21). 4674(5b), 1763(19), 4802(2). 49060(27b). 40635(27b). 4960(21b). 5513(21). 1281527a), CONC 13886(28a), 25464(5a), 25595(2) Ringuelet 11221): Rizzo s.n. BAB 7866121): ps 3196(29): Rodríguez 284(19), 1218(29); Rodríguez, 1379(5a), 1984(19); Rodríguez Ríos 317031). Dm 3505 131). 3646( 31); a 811), € 1119), s.n. LIL 29388(5a); Roig 20(29), 365(1). 422(1), 6010019), 807(29), 2589(14b), 2914(29), 316318), 4648(19), 528619), M o: 6640(1). 7269(38b), 8190(35a). 8909(38b). 9023(35a). 9255(33). 9810(1). 9812(38a), 9860(20b). 99863(20b). Volume 95, Number 2 Peralta et al. 389 2008 Revisión del Género Junellia (Verbenaceae) 10212(8). 10870(29), 10888(39). 10950(37). 10961(18), a 61(28b) 64(19) Sesberg 132(24): SGO 42507(25). 12476(18), 13057(7), 13058(1), 13059(9) 13060(17). | 42480(24). 42483(12). 42526(27b), 54708(25). 54719(25) 13390(38b), 14060(38a), 14092(7), 14365(11), 14435(10). 54726(27b). 54802(12). 54682(24); Shepard 26(6): SI 14445(17), 14457(28b), 14692(33), 14537(33), 14580(10), — 2039(19), SI 3394(14b), SI 3408(1). 19178(14a); Silvestri 14669(11). 14683(38b), 14685(10), 14781(11), 16009(38b). 6470(11). 6471(20a). 6472(20a)J. 6473(38a). 6474(30). 16015(7). 16023(28b). 16411(28a) 16032(38b). 2399( s.n.(8): map. 8568(34); Skostisberg 624(38b). s.n. BA bis(10), 3219(30), 3221(1), 3219(30), 3221(1). s.n. TBPA- 23725(11), s.n. CONC 18195(28a). s.n. SGO 58993(33) fIT 2397(8), s.n. TBPA-fTT 2452(8), s.n. TBPA-(TT 2453(1). 11069(39). 1107 > 28b); Slanis 195(19), 219(14a); Sleumer s.n. TBPA-ITT 2478(8), s.n. TBPA-(TT 2841(20), s.n. TBP N 324(21), 381016). 464(19), 515(I4a), 547(38b), 589(24), (IT 286 7(5); Ramen n 2509(38b). 2660(33). s.n. B/ 598(28a), 644(24). 664(19), 772(20). 897(20), 906(33) 66335(33), s.n. BA 66363(38a), s.n. H 1061031(11). s s 1082(11), 1088(33), 1108(20a). 1109(20a). 1160011), 1061032( 3) Romanczuk 961(33), 969(10), LOTO( 20 1). 120924). 1457(11). 1469(38a). 1916(5a). 2341(35b) 1055(1): Romanella 2(21); Roquero s.n. Herb. Parques 2364(19), 2077(35b). 2078(35b). 20679(21). 3228 (2) 8932(33), s.n. Herb. Parques 8938(11); Rose 18834(0) 3243(14a). | 3316(2 3319(2), 3345(35b), | 3580(29) Rossi 358(28a), ed 360(28a), 368(17), 369(20a) 615029). s.n.(8); Smith 4058021): ae 182(21); Solomon 370(24); Rossow 33(20a) 152(1). 166(20a). 196(9) 166433). 12850(6); Soriano 550(1), 662(29). 1274(9) 212(20b). 265(9), SM 119(21). 942(28a). 1000(24) 1440(38b), 1500(1). 1870(20b). 1884(38b). 2003(11). 1545(17). 1547(21) 1568(20b) 1657(20b) 16156 2014(3). 2038(7). 2044(3). 2060(7). 20651 2101(11) 1663(16), 1682(38b). 1723(35a) 1736(21). 1738(28b) 211711), 2136(20a), 2142(10). 21023(7). 21064(10) 1739(17). 1743(21), 1785(28b) 2495(17), 24.98(21 2224(11), 2225(1), 2272(20a). 2433(38b), 2479(1). 2575(1). 2511(28b) 2528(7), 2686(28a), 2895(11), 3059(28b) 257610). 2597(33). 17(38b), En (20b) 2865(20a). 4491(28b) 15047) 515(28b), 4550(11). 4585(l) 3211(10) 3284(20) 3344(20) 385(11). 3388(33) 4.586(28a) js sn. LP 861507). s.n. LP. 8375(38b) 3402(20a). 3406(10). 3418(32) 4907 ) 3182(17) Rotman 345(35a) 374(37). 384(39). 395(19), 405(29) 3816(1). 3839(32), 387 1(11). 3875(1) ee 1). 3980(38b) 408(19), 410( e 414(16), 436(35a), 449(29); Rusby 3985(10), 3989(10). 4006(20b) 4035(35a) 4048(l7) 2670(6); Ruthsatz 119(21), 136 15(14: » i 238(6) 4057(19), 4072(28b). 321(10), 4574(20a), 4700(11) 273(2), s.n.(2), 314 p 367 b(35b). 373(29), 271(29) 4701011), 4733(10). 4764(11) 181(10), | 4781b(11) 6347(16), 6471(16), 10(13). s.n. p 9991(29), BAA —4798(38b) 4814(11), 5037(38b). 5043(10), | 5044(10) 9992(29). BAA M BAA 9994(14a). BAA 9995(6). 5051(38a). 5095(33). 5096(20a). 5199(38b). 5264(10) BAA 9996(2) BAA 9997(21), XXVI 14(14a). s.n.(6). s.n. BA 5265(10), 5266(20a). 5281(35a) 5282(20a) 5288(10) 30/116(31). 5316(9), 5338(10). 5367(10), 5373010), 5408(12), 5674(33): Saénz 701(35a); Saile(37); Garay s.n.(37); Samler s.n.(36); Sosa 32(16), 33(24); Soukup 43(6), s.n.(( B SpA s 5941(19); Sánchez 145(21), 184(20b), 241(10), 244(11). 246(7), 252(9) Spegazzini C. s.n.(19), s.n.(7), s.n.(33). 1122(11), 1134(8) 291(38a), 314(38a), 342(10), 344(33), 347(1), 381(20a) 6476(35a), 6477(35a). 13009(11). s.n. ane 11294(35a), 384(11), 386(10), 486(10), 488(11). 520(20), 584(20b) l 92732(39), BAB 22733(39. BAB 22734(39), 645(20a); Sanderman 3855(6)(K), 3891(34), 896(6) 23404(39), s.n. BAB 23152(35a), BAB 23451(35a). 3916(6); Sanso 121(28a), 135(28a); Sanzin 44(35a), 59(18) LP 10010(3), s.n.(l4a), s.n. BAB 33722(19); 9 Mh 120(39), 277(18), 554(21), 661(16), 733(29), s.n.(35a) P. L. s.n. BAB 29026(19): Spegazzini, 178(16). s.n. 3102(35a), 3103(21) 3106(16), 3113(28b), sn. MERL LP 10399, s.n. BAB 57952(35a), s.n. BAB 58190(18). s.n. 2561(18); Sarasola 1099(19); Saravia Toledo 14725(14a), BAB 61204(38b) Squo 88061(13) Stafford = 333(19), 14733(29), 1 n L ee E MT ek 723(19). 1136(6): Steibel 2486(17). 2520(21). 3192(17) 19(35a), 616(21), 2517(35a), 273 d = 281921), 3245017), 33167); Scala AID. 33507). sn. LP 8317(17), s.n. BA 23706 Schajovskoy a 33/1V (24). 35/1V (7). 36/1V (9), 43/ IV(9). 051-1V(7). 54(20a),54-1V(30), DT , 66-4 V(7). 79/ VII(10), 139/V Nn 15111V(9). 175(28a), s.n.(17), s.n.(38b), s.n(39), s.n. SI 25623(20a), s.n. SI 25624(39), s.n. SI 25625(1), s.n. SI 25628(38a), s.n. SI 25664(30), s.n. Herb. Parques 9315(28b); Schlegel 4827(19), 6205(25); Schinini j: Schikendantz 89(19); Schmitt 176(6): Schreiter ee A LL 857 51 03(19), « s.n. 5(27a); Schul 6168(21), 6210(28b) 02 i . 6668(19), 6675(19); Schwa 2079(21) 2097(17) Scolnik 1706(6). , 417(11), 423(10), 425(33), 428(1 1), 634(6): Seibert s.n. “TBPA- fIT P, 2180(33); Seijo 10560(19), s.n. 86245(35b): Schub X Nn ~ e = S 2157, sn. TBPA-TI 1623(28b), 1713(28a), 2068(9), 2162(21), 2256(9); Selander 88-a-28(21); Semper s.n.(21), 28(35a). 119(21), 407(35a) 149(18), 574(29), 631(39), 4347(35a). s.n.(35b), MERI 9982(18), 10006(18); Senn 4305(17); Serra 29(13), 30(13), 4718(21), 5270(17). 5279(17): Stipanicic sn. BA 44143(28b); Storni 29(39), s.n. BAB 69417(28b); Straw 2263(34); Stuckert 22289(20a): Stuessy 6811(11), 689711 1), od 10284(28a). 10355(21); Suárez 3(38b): Suero 11573631), 1017(27b), 1118621). 1154121), Teillier s.n.(28b), 1028a). 5230(28a). 4583(25), 25), 4585013), joco Pes 5460(16), 5462(13); Tessleff 5328(20b), )9(38b). 5414(11). 5444(33), Thunger 107 nm 149(39): ;.n.(9); Tovar 6779(19); Troiani 3650(21), 15003(20b). 15043(20b), 15068(20a), 151550). 15202(1), 15212(7) 155510). 15705(10), 15708631), 15712010), 7); Troncoso s.n. SE 6062(19). s.n. SI 20498(21), s.n. SI 2051307 y Tutin 1374(21); Tweedie 311(33). Ulibarri 74721), 1480(21). Valentin 42(20b), s.n. SI 3306(20b), s.n. SI 3445(33); Valla 33(35a) (BAA 17794), 39(21), 46(BAA 17807)(16), 55(20b), 65(28b). s.n. BAA 17796(17). s.n. BAB 17600(2 1). s.n. BAA 17826(28b): Vallerini 20001) 234(9), 239(1). 542(20a), 638(17). 703(32). 900(20b). 907(38b). 986(20a). ; league s. 149) i = gi 1m ] d Je SS REGE Ww So 1293(9), 1375 A 9). 1528( Vel 1533017), 552(20b). 2033(38b), 2040(20b) 2082(3). 2053(38a). 2070(11). 2115(11), 2124(10). 2126202). 2131(10)..— 2146(l). 390 Annals of the Missouri Botanical Garden 2149(20a), 2167(1), 2172(33), 2324(28b) 2814(38b), 2825(39), 3387(17), 3388(sp, 3430)(7), ena ] 3492(28b), 3500(38b), 3520(38b), 3522(17), 3641(11), 3666(11), 3693(20b), 3772(20a), 3787(38b), 3803(1), 3805(35"), 3813(10); Vals 100013); Var- gas 7927(34), 11348(6); Vattuone 199(2 407(6); Venturi 4163(29), 4282(29), 45: 3929). 6713(19), 6715(35b), 6718(19), 6943(5a), 6970(39), 8141(21), Z © ee 354 N © = Ww Ke) a s.n. LIL Eno. Vidal 300: M 9(11); s.n. ); Villagran s.n.(39), 1358(31), 2328(34), 2430(19), 2474(19), 4063(19), 9072(21), 9107(5b), 9238(5b), 9529(31); Villamil 2257(1), 2259(10), 2281(11), 2282(11); Villanueva s.n. (31); Vogel 9523L(29); Volponi s.n.(21), 21(21), 24(21), 148(29), 319(21); von Rentzell s.n. SI 19233(39); Vuilleu- mier de 302(6), 387(21) 41 enknecht s.n.(36), 112(39), 18421(36); Walter s.n. mon s.n.(17); Weberbauer 5779(34), 6895(6), 7 Weigend 97/771(34); Weisser 400(39); Weldt 1 dermann 231(13), 236(29), 488(39), 496(28b), 577(28a), 848 7a), 1111(34), 1120(6); Werner 123(5b), 336(29), 444(29), 921(21); West 3909(27b); Wilezeck 37(1), 43(21), 45(21), apu ee): 51(35a) 53(35a) (paratipo 2 isoparatipo ; Wingenroth 238(29), 319(39), 356(19); Worth VN Wulff A. 19917), 200(21), 204(35a), 371(28a). 859(19 = (9). = — Zardini 42(38b), 86(17), 148(29), s.n.(31), 871(31), 998(31), 999(27a), 1374(13), 1486(13), 10964(28b), 3075(28b), 1537(28b), 5991(28b), 4091(19), Eu 4402(6), 5183(25), 5533(26), 6482(33), 6528(28b), 06(27b), 9875(26), 10019(31) 12533(27b); Pee 5946(6), 6004(6), 6408(14a). 1931(29); Zöllner Zuloaga STAUFFERIA AND PILGERINA: TWO NEW ENDEMIC MONOTYPIC Valéry Maléco: ARBORESCENT GENERA OF SANTALACEAE FROM MADAGASCAR! Zachary S. Rogers,? Daniel L. Nickrent,” and 4 ABSTRACT Two new pe Nay species of Santalaceae, both endemic to Madagascar, are described in the new deis genera, Staufferia ers, Nickrent & Malécot and Pilgerina Z. S. Rogers, Nickrent & Maléc ot. Based on available dd and nobile p bn new species are part of a clade temas with Ty una Michx. of Asia and North jo. the Indo- opyrum Arn.. and the central and w ka Pellegr. & Normand. Staufferia is vea restern AIT morphologically rd Okoubaka by the smaller inflorescences (4 to 10 vs. 50 to 100 o: smaller (1.8-2 X 1.1-1.6 cm v m), obovoidal (vs. ellipsoidal) fruits; smaller (ca. 15- 20. mm thick), 5-sulc: ate (vs. smooth) exocarp; and thinner (ca. 0. 1.5 mm vs. 7-8 mm diam.), aues perianth; thinner (0.5-1 mm vs. 5 mm vs. 3 thick), smooth (vs. deeply striate or alveolate) mesocarp. Pilgerina differs from Screropyrurm by the smaller inflorescences Bt to 23 vs. 60 to 100 flowers); pedic ellate X I. (vs. sessile) flowers; smaller (1. obovoidal to pyriform) fruits; and thinner (ca. 0.5 mm vs. 1.5-3 n mesocarp. Both species are illustrated and assigned an IUCN preliminary cons Key words: IUCN Red List, molecular phylogeny, Pilgerina, Sula :m), broadly transversely ellipsoidal to subspheroidal (vs. um thie k), e or finely striate (vs. deeply striate or alveolate rvation status of Least Concern (LC). Staufferia, Santalales, taxonomy. Cavaco and Keraudren (1955), in their treatment of Santalaceae for the Flore de Madagascar et des Comores, recognized the shrubby genus Thesium L. as the only member of the family occurring in Mada- gascar. Two years later, Capuron (1957) noted that Olacaceae were represented on the island by five genera, Anacolosa Blume, Olax L., Phanerodiscus Cavaco, Ximenia L., and by a fifth undescribed genus only known from a single collection (Service Forestier 987), which differed markedly from these named taxa in having five alternisepalous longitudinal furrows on the external surface of the fruit, thereby giving the false impression of a dehiscent 5-locular capsule. In an ineffectively published technical report, Capuron (1968: 39-40) recognized the material as an unplaced of Santalaceae providing the arborescent taxon provisional name “Santalacée A.” Within that same document, he attributed three other collections (H. Humbert 28660, Service Forestier 22622, 23390) to a second taxon, recognized as “Santalacée B” (Capuron, 1968: 40-41). At that time, both taxa were only known from fruiting material, but specimens still showed well-preserved, persistent floral organs re- maining at the apex of the fruit. Capuron, so sufficiently confident that the two santalaceous taxa were unknown to science, provided provisional descriptions and illustrations for both in his report. Since that report, however, neither taxon has received much attention, but over the course of the past 15 years, additional flowering and fruiting specimens leaf that allow us to reexamine have been collected, as well as material preserved in silica gel, Capuron's work and to specifically address the phylogenetic position of his “Santalacée A” and "Santalacée B" using molecular and morphological means. MATERIALS AND METHODS PHYLOGENETIC ANALYSIS The specimens used for DNA extraction are listed in Table 1. Leaf samples were taken from either silica ' The authors thank Martin Callmander, Laurent Gautier, and Louis Nusbaumer for making newly collected material of Stauferia available; Johny Rabenantoandro, Rolland Ranaivojaona, Faly Randriatafika, and Fidisoa Ratovoson for assistance with Der, Romina Vidal-Russell, and Miguel A. Gar au for the Latin « fieldwork; Joshua the Pilgerina pyrene; Roy Ger cía for laboratory assistance; spia i MERDA for his Pro of liagnoses; and Brendan Lepschi, P an helpful review comments. The Botanical Research Foundation of Idaho provided financial Rd cane fieldwork for ZR. A National Science Foundation Grant (DEB 0108229) to DLN supported the pe ular bs of this study. ? Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0 299, U.S zac ud rrogersmobot org. S.A. * Department of Plant Biology, Southern Illinois University, Carbondale, Illinois a -65t * [Institut National d'Horticulture, UMR S doi: 10.3417/2006148 AGAH, 2 rue le Nótre, F-49045 Angers Cedex D Penes, ANN. Missourt Bor. Garb. 95: 391—404. PuBLISHED ON 18 JUNE 2008. Annals of the Missouri Botanical Garden matK accD GenBank accession numbers rbcL number DNA accession Country pers for the 11 taxa from the Santalaceae used in the phylogenetic analyses. The DNA accession numbers refer to the collection by Voucher I Voucher information and GenBank accession num Taxon Nickrent archived at SIUC. Table E, A oS. 2 a eS gel-dried leaves or fresh material. Comandra umbel- LAN ŻA S LAA L ; mE / DAARAAAARAAAS lata (L.) Nutt. and Buekleya distichophylla (Nutt) DDD DO DOS ooo lorr. were used as outgroups. A modified 2X 668655512356 siii vs cetyltrimethylammonium bromide (CTAB) method &i c6 Ev Gb X cB B GEN tO P was employed for all genomic DNA extractions ee ee E B Eee (Nickrent, 1994). The polymerase chain reaction Soo oO eS or (PCR) mixture (final concentration) contained 1X Promega buffer (5 mM KCl, 1 mM TrisHCl), 1.5 uM TND Ee HNO OE ie M p. RS Gb Ff O0 pop FSOD. [SO [- MgCl. 50 UM dNTPs, 0.4 UM forward and reverse DODDDDDRDADAD : en A T & & & & S S GG primers, 1 Unit Taq polymerase, 30-50 ng genomic SSS SS SS SS d DNA. Nuclear small-subunit (SSU) rDNA and three : chloroplast genes (rbcL, matK, and aceD) were NX noo QA Ka i : sac Mida O S Eo amplified. PER amplifications for SSU rDNA and DSORDDADIDDVDDOD E e COLI GOL OI | GI ux GLO rbcL generally followed the conditions outlined in 33358238 8288 A mo s cu e" | C xum a as Nickrent (1994). For amplification of matK and accD, a step-up procedure modified from Palumbi (1996) was used: 94 C for 5 min. followed by 5 cycles at o A ee rr 94. C for 30 sec., 46 € for 30 sec., 65 °C for 90 see. SSKRARGHGRASG EPA eget TG RIPE E a a followed by 25 exules at 94 C for 30 sec., 48 C for 30 sec., and 68 C for 90 sec. The following primers were typically used for PCR am ieri and I 4 I - = z sequencing (all 5' to 3%): for SSU rDNA, 12f (TCC S g% z bo" nr S 3.3 2 $ & uou uL DR Ea zo3 El s fes PV Dey Ret ge SE. = TAA GTT TCA GCC), and 1769r (CAC CTA CGG AAA: oS: = "(op PEOP rp ij TATOO TG | a dh E ps SE 1420r (TCG AAG TAT ATA CTT TAT TCG ); and for NF ue SES M pa A "A TOR VOD TT . js Dee Se = c accD, 1f (TCT ATG GAA AGA TGG YGG TT), 1Bf ND) — -= N Z . É . M - e EN Y - zs "S E = re (ATG GAA AAA [GG YGG ITY AA), and 1300r wz cs ERO LU Leu "De my mp wp Opp pp val ` ER 2e du e det (TGY TCA ATT ACT CTT TTA CC). Sequencing was = O ps o T^ uas E. US : f . "RD ISSS conducted using automated methods (ABI Prism 377 SE BUS SIE am : i doxes 9 Powec-x SSRs automated DNA sequencer; Applied Biosystems. 4 q vB E S o W wy ^R GS J Ss Bob estes Bm s baer Ci ala lio ot SE ZI. EE Foster Tas California) according lo pans rs SA SS is protocols. Sequences were deposited with GenBank under the accession numbers given in Table Z The above ingroup and outgroup sequences were = aligned manually using SeAl version 2.0 (Rambaut, = = 2004). and, for the three protein-coding genes, it E alignment was guided by assumed amino acid compo- D = si E p sition. The multigene alignment is available on the age = second authors web site: http://www. itiepl D - Zo Y ; $565 x siu.edu/Alignments/Alignments.html. Analyses t ve ZE Go BS Log zz separate gene partitions showed that all were generally o » dc m CON SS A o 2a ee on mH E congruent, hence the four genes were concatenated. E 23% E cj EN 2 El thereby. producing a matrix of 11 taxa by 5094 sites. (2578S 1. Ow amm ; i oe "m ` : a 1n re CUP Gaps were treated as missing data. The data matrix was ao Ajo. 2 sj EWM | o OS uc pl cp analyzed using maximum parsimony (MP) as imple- & Su. iS SEn ; " SM OA Se eats 8B 22 8 8 mented by PAUP* 4.0b 10 (Swofford. 2002). A branch- S22 € $9ov PS? i . : B.S Eus and-bound MP search was performed using 100 random E o oos 58 cR n . ; i ; Pees 9 > SEARS addition sequence replicates with tree-bisection-recon- SS IIS SERA R nection (TBR) branch-swapping, holding 10 trees at BSS ESPS ES Tzu n s 7 ] Ske es £2 2 2 § = each addition step. with all sites equally weighted. o8 3.8 s pe e 8 id ) x x00o0Xo0&5üud Nodal support was estimated using equal-weights MI Volume 95, Number 2 2008 Rogers et al. Staufferia and Pilgerina (Santalaceae) 393 Table 2 was nol obtained for Okoubaka aubrevillei. Cl- = Consiste Gene diversity statistics for all data partitions for the 11 taxa of Santalaceae. The nuclear SSU ney ue x minus uninformative siles ; rDNA sequence ¿ep = chloroplast: HI- = = rescaled consistence y index. homoplasy index minus uninformative sites; RI = retention index: No. Parsimony Data partition Sites trees length informative sites CI- HI- RI RC Nuclear SSU rDNA 1813 | 160 39 0.7414 0.2586 0.7826 0.7092 Chloroplast rbcL 1421 2 176 60 0.7204. 0.2796 0.8129 0.6929 Chloroplast matK 1180 | 377 121 0.1606 0.2394 0.8289 0.7300 Chloroplast aceD 1280 l 242 8: 0.8598 0.1402 0.9148 0.8581 Combined cp genes 3881 l 801 264 0.7665 0.2335 0.8408 0.7443 Combined four genes 5094. l 961 302 0.7633 0.2367 0.8346 0.74.17 bootstrap analysis (heuristic search using 1000 random sequence addition replicates). TAXONOMY Herbarium specimens were examined from G, MO, NEU, P, TAN, and TEF. Additional Okoubaka Pe & Normand and Seleropyrum Arn. was studied from several additional herbaria (B, B K, L, LAE). Okoubaka were predominately based on an examina- tion of dried specimens of both members of the genus, & Normand and O. michelsonii material of legr. Observations and measurements for O. aubrevillei Pellegr. J. Léonard & Troupin. Our morphological observations for the poorly known Okoubaka were supplemented by additional flowering and fruiting information taken (Louis & Villiers, from several pertinent literature sources Léonard, 1948; Léonard & Troupin, 1950; 1973a, b; Hallé, 1987). For Scleropyrum, the mea- surements and observations provided below are based Asiatic and S. mostly on herbarium material of the two 5. pentandrum (Dennst.) Mabb. which are the most morphologi- species, maingayi Hook. f., cally similar to the Malagasy taxa. Two other taxa, $. aurantiacum (K. Schum. & Lauterb.) Pilg. leptostachyum Pilg., both from New Guinea, probably do not belong to Scleropyrum (Malécot, pers. obs.), and thus were not included here. Additional morpholog- and S. ical information regarding the flowers and fruits of Seleropyrum was taken from the most recent, compre- hensive literature (Macklin & Parnell, 2000, 2002; Nianhe & Gilbert, 2003). The botanical illustrations and floral measurements of the Malagasy species were made from rehydrated material. Complete collection data for cited exsicca- tae, including photos of types and other representative specimens, are posted on the Tropicos database at http://www.tropicos.org. Coordinates and elevations of collecting localities were assigned post-facto, when necessary, using the “Gazetteer to Malagasy Botanical Collecting Localities” (Schatz & Lescot, 2007; http://www.mobot.org/MOBOT/Research/madagascar/ gazetteer/), and are enclosed by square brackets in the text. The distribution map was created using ArcGIS 9, and species distributions are superimposed over the zones of Madagascar 1974). Conservation five simplified bioclimatic (Schatz. 2000, following Cornet. status for each species is provisionally assigned based on the /UCN Red List Categories and Criteria Version 3.1 (IUCN, 2001). RESULTS PHYLOGENETIC ANALYSIS Gene diversity statistics from MP analyses of each of the individual gene partitions are shown in Table 2 The lowest number of parsimony informative sites (39) was obtained from the SSU rDNA partition and the highest (121) from matK. The (CI-, consistency index minus the uninforma- highest consistency index tive sites) was from the accD partition (86%) and the lowest came from rbcL (72%). Only matK and accD resulted in one most parsimonious tree, whereas SSU rDNA produced four, and rbcL yielded two equal- length trees. Combining all three chloroplast gene partitions resulted in one tree that was identical in topology to the one obtained with the four-gene data set that included nuclear SSU rDNA (Fig. 1). The tree expands the well-supported “clade b” previously described in Nickrent and Malécot (2001: Seleropyrum, 71) by the addition of Cervantesia Ruíz & Pav., Capuron's Santalacée A (= Staufferia Z. S. Rogers, Nickrent € Malécot), and Santalacée B (= Pilgerina Z. S. Rogers, Nickrent € Malécot). The nine ingroup taxa form two well-supported clades (Fig. 1, 100% bootstrap support [BS]. the first (the Cervantesia South American Cervantesia, clade) composed entirely of the genera Acanthosyris (Eichler) Griseb., and Jodina Hook. & Arn. ex Meisn., and the second (the Pyrularia clade) of the remaining genera, namely Pyrularia Michx., Pilgerina, Staufferia, Scleropyrum, and Okoubaka. All members of the Old Wor latter clade are d taxa with the exception of the basal genus Annals of the Missouri Botanical Garden 353 Figure 1. The single ae ies and 10 genera of Santalaceae used in this study. BS Okoubaka aubrevillei (AF) Scleropyrum pentandrum (AS, NG) | y Staufferia capuronii (MA) 3$ Pilgerina madagascariensis (MA) a Pyrularia pubera (AS, NA) Acanthosyris asipapote E: Acanthosyris falcata (SA) E 9 Cervantesia tomentosa (SA) e i Jodina rhombifolia (SA) ES Comandra umbellata (AS, EU, NA) IE Buckleya distichophylla (AS, NA) 3 MP tree obtained from the analysis of a combined SSU rDNA, rbcL, matK, accD data set for the 11 > given above branches and branch length (number of steps) below. "able 2 for tree statistics. The a 2a for each genus is enclosed in parentheses to the right of the , AS = Asi constituent species. AF = Africa SA = South America. Pyrularia, present in eastern Asia as well as eastern North Okoubaka Scleropyrum, although BS is only moderate (Fig. 1, 10% BS). Staufferia is sister to the Okoubaka— Scleropyrum clade (100% BS) with Pilgerina sister to those three genera (99% BS). America. appears as sister to "l'AXONOMY Staufferia Z. 5. Rogers, Nickrent & Malécot, gen. nov. TYPE: Staufferia capuronii Z. S. Rogers, Nickrent & Malécot Hoc genus ab Okoubaka Pellegr. & Normand inflorescen- 0- vs. 50- ad 100-flora), fructu ellipsoideo), tia staminata minore (4- ad 1 minore obovoideo (vs. perianthio in fructu persistente minore, pyrena minore (1.1-1.6 vs. plus quam 3 cm in diametro) obovoidea (vs. ellipsoidea), exocarpio 0.5— | (vs. 15-20) mm crasso longitudinaliter 5-sulcato (vs. laevi) atque mesocarpio laevi (vs. profunde striato vel alveolato) ca. 0.5 (vs. 3—4) mm crasso differt. Dioecious shrubs or trees; branches estipulate. Leaves alternate, distichous, margins entire; venation brochidodromous. Inflorescences axillary; staminate inflorescences thyrsoid, 4- to 10-flowered, bracteate; pistillate inflorescences not seen. Flowers unisexual, obovoidal (in bud), 5-merous, epigynous, actinomor- sessile; calyx pak petals 5, phic, free, valvate; staminate flowers with 5 oppositipetalous stamens (adnate near base of petal lobe and below the adjacent disk), introrse, dorsibasifixed; anthers tetrasporangi- = Europe, MA = Madagascar, NA = North America, NG = New Guinea, — ate (two pairs of divaricate locules), each loculus of a theca opening by a common, longitudinal slit; nectary disk star-shaped, fleshy, located between stamen and pistillode style; disk lobes 5, triangular, entire, apex rounded, each lobe alternating with a filament; pistillode style + orbicular in transverse. section; pistillate flowers not seen (based on persistent remains at the apex of the fruits): staminodes 5, minute, hidden under the apical lobes of fruit exocarp; 1-locular; ovary inferlor, placental column straight; ovules 1 to 3, apical, pendulous; style very short ith 3 to 5 lobes. Fruits drupaceous, single-seeded, obovoidal, cylindric or absent; stigma subsessile, with persistent floral parts at apex; persistent perianth 0.5-1 mm thick, with 5 fused indehiscent segments, each segment ca. 1.5 mm diam.; exocarp fleshy, alternating with a longitudinal densely pubescent furrow; pyrene (i.e., seed plus mesocarp) obovoidal, 1.1-1.6 cm diam., thick Seeds globose; endosperm copious. with a bony, smooth, ca. 0.5 mm mesocarp; endocarp papyraceous, very thin. Etymology. ‘The genus is named in honor of Hans Ulrich Stauffer (1929-1965), whose 10 “Santalales- Stauffer, 1957, 1961, 1969) made major contributions to our understanding 1965 prevented him from summarizing his morphological Studien" publications. (e.g., 4 Santalaceae. a Mis unexpected death in observations and taxonomic conclusions that would have eventually resulted in a complete monograph of Volume 95, Number 2 2008 Rogers et al. 395 Staufferia and Pilgerina (Santalaceae) the family. One of his most astute observations was that some members of the heterogeneous tribe Osyrideae, with its two superimposed loculae per theca, should be segregated, as was done posthu- mously (Stauffer, 1969) as Amphorogyneae. is significant that early in his career Stauffer (1957) recognized the affinity of Okoubaka (a large East African tree that was nebulously placed in either Octoknemaceae or Olacaceae) with Scleropyrum and Pyrularia, and that these genera plus Staufferia have now been shown to be closely related using molecular data. The description of Staufferia marks the first time a genus has been named in Stauffer's honor. Staufferia capuronii Z. S. Rogers, Nickrent & Malécot, TYPE: Madagascar. Antsi- ranana Prov.: Analabe forest, Fivondronana sp. nov. Vohémar, Firaisana Nosibe, Fokotany Anjiabe, forét littoral sur sable d'Anaborano, pres du Lac Sahaka, 13%04'42"S, 49%54'13"E, 25 m, 2 Nov. 2002 (© fl.), J. Rabenantoandro, R. Rabevohitra, G. McPherson, H. Ranarivelo & M. Sola 1100 (holotype, MO!; isotypes, G!, INH!, K!, P!, TEF!). Figure 2. Dioecious shrubs or trees to 20 m tall; buds and young vegetative organs densely to moderately trichomes sericeous, sparsely pubescent or glabrous on mature branches; simple, un- 0.3-0.5 mm, branched, adpressed or subadpressed; branches + zig-zag; bark longitudinally striate, cracked and exfoliating when dry, sometimes lenticellate. Leaves 2-7 X 1-3 cm, length:width ratio (2-)2.5-3:1; leaf blades elliptic or ovate, adaxially glabrous, abaxially moderately pubescent to glabrescent, chartaceous, apex acuminate or acute, acumen 3-8 mm when present, margins revolute, base short attenuate; midrib depressed adaxially, prominently raised abaxially, more pubescent than the blade; venation usually more strongly raised abaxially; secondary veins 4 to 6 pairs per side; fine venation very loosely reticulate, rarely obvious adaxially; petioles 2-5 mm, pubescent to less often glabrous, caniculate adaxially, rounded abaxi- ally, articulate at base. Staminate inflorescences 5— 7 mm, 4- to 10-flowered (with 2 or 3 subsessile flowers at the tip of each inflorescence axis), 3- to 5- branched, located near the base of young shoot with the main axis subtended by | or 2 scales or rarely by a leafy bract or developed leaf (at early stage, a young shoot with several inflorescences may be interpreted as a long inflorescence, but young leaves are located near the apex of the axis); inflorescence axes ca. 0.3 mm diam., densely pubescent; primary inflores- cence axis 1.5-2.5 mm; bracts linear, 0.8-1 X 0.2— 0.3 mm, densely pubescent abaxially, otherwise glabrous. Flowers sessile; staminate flowers: l- 1.5 mm, 1-2 mm diam., obovoidal (in bud), vellow or yellow-green, externally densely to moderately sericeous, trichomes in a pattern resembling a calyx; 0.8-1 0.7—1 mm, papilliate near apex, along upper 1/3 of margin, and calyx absent; petals ovate, along base adaxially, puberulent behind filament adaxially, those trichomes 0.1-0.2 mm, matted and wavy (trichomes weakly attached to the abaxial surface of the filament near the anther connective), pubescent abaxially, those 0.25- glabrous; moderately to sparsely 0.1-0.3 mm, 0.15-0.2 mm anthers ca. 0.3 X 0.4 mm, glabrous; nectary disk trichomes straight; filaments 0.3 mm, wide (at base), ca. | mm diam., fleshy, glabrous; disk lobe 0.3—0.4 X 0.4—0.5 mm; 0.3 mm sometimes with a few distinct pistillode style ca. diam., papilliate, lobes: pistillate flowers not seen. Fruits drupaceous, obovoi- dal, 1.8-2 em, (1.1—)1.4—1.6 cm diam., 1(2) developing per infructescence; fruiting 5-4(-6) mm, 0.8-2 mm thicker where meeting the fruit: exocarp fleshy, 0.5— light yellow, pedicels and diam., articulate | mm thick; mesocarp bony, ca. 0.5 mm thick; endocarp papyraceous, ca. 0.1 mm thick. Seeds globose, 1.1—1.3 cm diam.; embryo 5-6 mm, in- curved. Distribution and phenology. Staufferia capuronii is widespread but patchily distributed in the eastern half 25-1827 m (Fig. 3). Populations are distributed from the humid of Madagascar from elevation wet forest of the Montagne d'Ambre in the far north of the island to the sandy littoral forest of Mandena in the southeast (near Fort Dauphin, Toliara province). The species has been recorded in flower in October, Fruiting collections November, February, and May. were made in January, February, and September. Conservation status. The species occurs inside two protected areas (Mandena, Montagne d'Ambre). Some of the population located near Zahamena National Park may actually fall within park boundar- les. The extent of occurrence (EOO) of the species is 81,000 km”, and the area of occupancy (ADO) is 60,000 km? based on a 100-km? species is assigned a provisional IUCN conservation status of Least Concern (LC) (IUCN, 2001). erid cell size. The Etymology. ‘The epithet was chosen to honor René Capuron (1921—1971), as he was the first to collect this species and bring attention to this novel component of the Malagasy flora. Discussion. Several examined collections exhibit notable morphological differences from plants made at the type locality: M. Callmander et al. 317, a fruiting 140 km 500 m) elevation, has collection made about west of the type locality and at higher (ca. 396 Annals of the Missouri Botanical Garden 0.5 mm 1cm Figure 2. Staufferia pp ae —A. Flowering branch. pubescent shoot with caducous bracts . Axillary th from interior with petal (left), p ral vi view (mic ed from back Nap i5. and from front (bottom PRAE yrsoid infl orescences 1cm —B. M leaf showing venalion pattern, —C. Young 7. Opening flower, apical vi longitudinal section. —H. Fruit, lateral view. —I. Fruit, apical v —J. Detail of the persistent floral organs at the apex af the fruit. = Lengo section through apex i of fruit; st = signal sm = staminode, di = disk, ex = exocarp, me = mesocarp, e Yerated endocarp, and se = seed endosperm. —L. Staminode from Fe > apical lobe of " des sp interior (left) and a (right) surfaces. All drawn from Service o elabrescent buds and shoots and a nearly glabrous fruit surface (Fig. 4A, B); L. Nusbaumer et al. 1436, collected ca. 25 km southwest of the type locality and made on lateritic soil rather than sand, has more densely pubescent buds, shoots, and infructescence axes, and a denser indument over the surface of the immature fruits (Fig. 4C); J. Rabenantoandro et al. (C ( Capuron) 987 => = 161, an immature fruiting collection made ca. 400 km south of the other localities, has consistently smaller and more pubescent leaves (but still of the same shape). The leaf and pubescence variation we have observed in these specimens may be related to environmental factors, whereas the fruit differences could be due to different stages of maturity. At Rogers et al. Volume 95, Number 2 2008 Staufferia and Pilgerina (Santalaceae) 45°E 50°E 20°S 25°S Figure 3. Distribution of Pilgerina madagascariensis (shaded circle) and Staufferia capuronii (square) on Madagascar. 398 Annals of the Missouri Botanical Garden Figure 4. Fruit variation in Staufferia capuronil. same collection as A but dry (right), apic 'al portion (upper specimen (L. Nusbaumer et al. 1436, photo) present, sufficient material is not available to ade- quately address whether these three collections are sufficiently distinct to warrant taxonomic recognition. Paratypes. MADAGASCAR. Antsiranana Proy.: Am- Commune Rurale Monee chaine Galoka, Mont jelinta, 13°35'07"S ` Manjaribe 317 bilobe, iw. Anketrabe- 5 s rues (C. MOJ2]. P. TE E ): Ane labe forest, F DUC Mi a Firaisana Nosibe, | 13 04'45"8 40754'17"E, R. de a Ps eed 1485 (MO, I "TE F), azakamalala, > & S. Mathieu 1254 (BR, wa a TE F); Anjiabe, de ndn o 7 (TAN), Z. Rogers, F. Davidson & S. Christoph 1166 (MO. P. TAN); Pare Na Montagne d Ambre, ca. 4 km WSW of Gite Etape, 49 08'E], S. Malcomber & S. Rapanarivo 1201 (P); Bots prefecture Vohé ‘mar, Commune Rurale de Daraina, Massif de Bobankora, partie Nord, 13. 13'39"8, 49 45'20"E, Vus- baumer, S. Wohlhauser & P. Ranirison 1436 (G, MO, P, cF) hajanga Prov.: Ambohimirahavavy, Massif de l'Ambohimirahavavy, rebord S du plateau de Marofamamo, Bealanana, [14 12/5, 49°00'E], Service Forestier (Capuron) 987 (MO, P[2], TE F); Bemafo, Versant rd Ouest. du campement Be mafo, 14°13'48"S, 49°03'32"E, C. Rakotovao, S. Buerki, J. z & Torize 2349 (G MOL: Toamasina Prov.: e OU. à d iaon a Sud, ju hors du pare ia, à 800 m de la vies 17 739'43"S, 48^39'28"E, J. Renontondr S. Rake xtonandrasana & I. Rak 161 (G, K. MO[2], P, TEF). Toliara Prov.: Forêt littorale à E RF, Aom Nahampoana, Tort Dauphin, | 24 57'26"5, AT 01'34"E, F. Ratovoson, F. Ronda toandro & E. Ramisy 113 (MO, P, TEE — Zahame J. Fee Pilgerina Z. 5. Rogers, Nickrent € Malécot, gen. nov. "PE: Pilgerina madagascariensis Z. S. Rogers, Nickrent & Malécot. —A. From a fresh fruit (M. Callmander et a 317, photc ). . Immature pum from a "m left), basal portion (lower left). — . —B. From t Hoc genus a Scleropyro Arn. inflorescentia minore (8- a 23- vs. e ad 100-flor i floribus p ructu minore (1.2-1. 1.7-2.7 em vs. ca. 3 X m p transverse een usque bci (vs. obovoideo usque pyriformi), pyrena lato transverse ellipsoidea usque sub- sphaeroidea mesocarplo laevi vel minute striato (vs. protunc ^" striato vel alveolato) ca. 0. 5 (vs. 1.5-3) mm crasso differ Shrubs or Leaves alternate, probably distichous, glabrous, mar- trees; branches estipulate, glabrous. gins entire; venation brochidodromous. Inflorescences axillary or subterminal, racemose, 8- to 23-flowered, bracteate. Flowers hermaphroditic, transversely ellip- soid (in bud), dorsiventrally flattened, (4)5(6)-merous, epigynous, actinomorphic, pedicellate; calyx absent; petals (4)5(6), free, valvate; stamens (4)5(6), oppo- sitipetalous (adnate disk), tetrasporangiate (two pairs of divaricate locules), near base of petal lobe and below the introrse, dorsibasifixed: anthers each loculus of a theca opening by a common, nectary disk suborbicular, fleshy, disk (4)5(6), entire, rounded, each lobe alternating with a disk l-locular; placental column longitudinal slit; lobes — ocated between stamen and stigma; filament; gynoecium embedded in nectary tissue; ovary inferior, straight; ovules 1 to 3, apical, pendulous; style very short cylindric or absent; stigma subsessile, with 3 to 5 lobes. Fruits drupaceous, single-seeded, broadly transversely ellipsoidal to subspheroidal, glabrous, smooth, with persistent floral parts at apex; exocarp fleshy, thin; pyrene (i.e. seed plus mesocarp) broadly transversely ellipsoidal to subspheroidal, with a thin, bony, smooth or finely striate mesocarp; thin. Seeds endocarp papyraceous, very broadly transversely ellipsoidal; endosperm copious. Volume 95, Number 2 2008 Rogers et al. 399 Staufferia and Pilgerina (Santalaceae) Robert In the early Etymology. he generic name honors Knud Friedrich Pilger (1876-1953). 1900s, Pilger published treatments of Santalaceae for the floras of New Caledonia and New Guinea (Pilger, 1906, 1908, 1924). He later wrote the treatment for Santalaceae for Engler and Prantl’s Die Natürlichen Pflanzenfamilien (Pilger, 1935). Compared with the earlier work on the family in that series (Hieronymus, 1889), including six Pilger’s treatment expanded Santalaceae by genera (all placed in his tribe, Osyrideae), and his observations have greatly im- proved our understanding of the relationships between Santalaceae genera. Three genera have previously been named in Pilger's honor, Pilgerodendron Florin (Poaceae), an (Cupressaceae), Pilgerochloa Eig Pilgeria Schmidle (a cyanobacterium). Pilgerina madagascariensis Z. S. Rogers, Nickrent & Malécot, sp. nov. TYPE: Madagascar. Toliara Mandena Forestry Station, littoral forest on white sand, : 47 01'55"E, 10 m, 19 Jan. 2006 (fl.), Z. Rogers, R. Ranaivojaona, F. Randriatafika, J. Rabenan- toandro, C. Davidson & S. Christoph 890 ne MO; isotypes, B!, BM!, BR!, CAS!, , L!, LE!, MA!, MO[2]!, NSW!, NY!, P!, E PREL TAN!, US!, WAG!). Figure 5. Prov.: Shrubs or trees to 12 m tall; branches £ zig-zag, glabrous; young branches angular; mature branches terete; bark smooth, exfoliating in strips on older growth. Leaves 4-13.2 X 1.2—4 cm, length:width ratio 3—4.5:l, glabrous; leaf blades elliptic to lanceolate, rarely ovate, glabrous, chartaceous or less often chartaceous-coriaceous, apex acute or slightly acumi- nate, tip usually rounded, margin revolute, base long attenuate or cuneate-attenuate; midrib raised on both surfaces; venation slightly raised on both surfaces; secondary veins 6 to 10 pairs per side; fine venation irregularly reticulate, loosely arranged, often incon- spicuous; petioles 3-7 mm, glabrous, often weakly caniculate adaxially, rounded abaxially, articulate at 1-2 cm, 8- to 23-flowered; inflorescence axes ca. 1.5 mm diam. at base, sparsely base. Inflorescences puberulent to nearly glabrous, trichomes ca. 0.1 mm, erect or subadpressed; bracts triangular-ovate, 0.3— 0.6 X 0.4-0.5 mm, clasping each pedicel base, often caducous before anthesis, glabrous or only sparsely puberulent along margin adaxially, sparsely to moderately puberulent abaxially, trichomes generally than concentrated along margin and near base. Flowers less 0.1 mm, erect or subadpressed, more ca. | mm (excluding the pedicel), ca. 5 mm diam., transversely ellipsoid, dorsiventrally flattened, green; pedicels 3—4 mm, sparsely puberulent, trichomes ca. less often subadpressed; calyx 0.1 mm, erect or 1.8-2.5 X 1.5- spreading (becoming reflexed near apex), absent; petals triangular-ovate, 2 mm, pubescent behind filaments adaxially, those trichomes 0.5-1 mm, matted and wavy (trichomes weakly attached to the abaxial surface of the filament near the anther connective), sparsely puberulent or gla- E brescent abaxially, those trichomes 0.05-0.1 mm, erect or less often subadpressed, apex adaxially puberulent, margin puberulent; filaments 0.5-0.7 X 0.3—0.35 mm (at base), glabrous; anthers ca. 0.5 X disk 2.5-3 mm fleshy, glabrous, green, shiny exudate observed on 0.6 mm, glabrous; nectary diam., dry material; disk lobe ca. 0.3 X 1 mm; style very short cylindric to absent, ca. 0.1 X 0.3-0.4 mm; stigma depressed in the middle. Fruits drupaceous, broadly transversely ellipsoidal to subspheroidal, 1.2— 1.9 em (excluding pedicel), 1.7-2.7 cm diam., green, elabrous, only 1 or 2 developing per inflorescence, vase attenuate with pedicel; fruiting pedicels l- 10.5 mm, often elongating substantially, 1.5-2.5 mm thick; mesocarp bony, ca. 0.5 mm thick, smooth to finely diam., glabrous; exocarp fleshy, 0.5-2 mm striate; endocarp papyraceous, ca. 0.1 mm thick. Seeds broadly transversely ellipsoidal to subspher- oidal, 1.2-1.4 cm, 1.4-1.7 cm diam.; embryo 7- 8 mm, incurved. Distribution and phenology. | Pilgerina madagas- cariensis is a widespread but patchily distributed Malagasy endemic occurring between 0 and 1200 m elevation (Fig. 3). Populations are known from several small littoral forest fragments (Mandena, Sainte Luce) in the extreme southeast of the island on sand, in the drier forests (Ihosy, Isalo, Zombitse) of the west on sand and sandstone, in the humid forest of the central Alaotra), calcareous massif of Ankarana located in the north of plateau (Andilamena, Lac and on the the island. The species flowers from October through January and fruits from November through April. Vernacular name. Sakaimboalavo (J. Rabenan- toandro et al. 315A; Z. Rogers et al. 890, 976). Conservation status. Pilgerina madagascariensis has been collected inside five formally protected areas (Ankarana, Isalo, Mandena, Sainte Luce, Zombitse). The species has an EOO of 230,000 km? and an AOO of 70,000 km? based on a 100-km^ grid cell size. The species is assigned a preliminary IUCN conservation status of Least Concern (LC) (IUCN, 2001). aratypes. MADAGASCAR. Antsiranana Prov.: An- karana, ouest (Nord). plateau de l'Ankarana, environs sud M d [12754'S, 40%08'E], K, MO, P, TEF). de Mahamasina (Antan Service Forestier (Capuron) 23390 (G, Fianarantsoa Prov.: Isalo, o el vallées de l'Isalo, à l'Ouest de Ranohira, [22724' S, 45' . H. Humbert 28660 (P); Menarahaka, ouest duro du dune d — dans le 400 Annals of the Missouri Botanical Garden 1 mm Figure 5. Pilgerina madagascariensis. —A. Flowering branch. —B. Cleared leaf showing venation ja e. Racemose inflorescence. —D. Open flower, viewed from top. —E. Stamens from back (left) and front (right). . Bud in longitudinal section showing free-central pendulous placentation. —G. Fruit, lateral view. —H. Apex of m showing persistent perianth lobes, disk, & and style in 5- (left) and 4-merous (right) flowers. —I. Longitudinal section of fruit; ex = exocarp, me = ‘arp, se — seed endosperm. —J. Detail of embryos, lateral and side views. —K. Surface view of pyrene p mesocarp Me enc iie d seed). A-F drawn from J. Rabenantoandro et al. 3154 (MO), G-K n from H. Humbert 28660 (P). P bassin de la Menarahaka, pres du o des routes 36 (MO), N. Dumetz 1235 (MO, P, TAN); Mandena, Forêt 2732! d'Ihosy à lvohibe et lakora, [22 32'5. O'R]. Service marécageuse, M7 à Mandena, Ampasy. Taolagnaro. Forestier Ai dnb 22622 (MO, P, n Toamasina — 24 56'09"3, 47 OI'49"E, J. Rabenantoandro, R. E Prov. mbodisaina, forêt degradée, — 17 17'25"85,.— tra, L ud & E. Ramisy 315A (INH, MO[2]. I 18^ 40' SOE, S. Rakotonandrasana 627 (MO); Andilamena. TAN, TEF); S E» Luc e, N of Fort-Dauphin, NW of Sainte Fivondronona Andilamena, Commune Andilamena, Foko- Luce b 2445128, 47 10'25"E, Z. Rogers. R. tany Anony Ampamoha, forêt de Soavalivato route Vohi- Ranaivojaona, 3 Ramisy, C. Davidson & S. Christoph traivo-Andilamena, 17 121275, 48 33/08", R. Randria- 976 (B, BR, CAS, K. L. MO[2]. NSW, P[2]. PE, ue naivo, A. Ratodimanana. T. Razafindrabeaza, H. TEF, US, WAG). G. McPherson. N. Du R. Rajanatsoa, P: Rakotondramanz za & 0. Rabozanahary 825 Rabevohitra 14227 (B, MO, P, TAN, TEF), C js Ic Pherson. (INI .MO. I |) Alaotra, Chutes du Maningory, N. Dumetz & R. Rabevohitra 14843 (MO): : ea 18 km rive. gauche, [17 22'$, . de ABE], A. Homolle 531 (P). E of e on rd. to Ihosy, 22.525, 44 41'E, J. Miller Toliara Prov.: Mandena, Fort Dauphin, rt. de Saine & R. Keating 4552 (MO. P. TAN): E. Forêt de Luce, a de Mandena. forêt sèche littorale, 2457'S. Besaka- a 15-20 km E of Sakahara. of Tulear, 1700 Allorge & C. Bourgeois 850 (P), C. Bourgeois — on Rte. #7, 22 46'S, 44 A0'E, T. Croat 30716 (P). Volume 95, Number 2 Rogers et al. 401 g Staufferia and Pilgerina (Santalaceae) DISCUSSION Features of the inflorescence and fruit indicate that Staufferia and Pilgerina share an affinity with a group of eight small Santalaceae genera previously recog- nized by Stauffer (1957), namely Acanthosyris (five to seven species, Costa Rica to northern Argentina; Nee, 1996; Ulloa Ulloa & Jørgensen, 2002), Cervantesia (one to four species, Andean Colombia, Bolivia, Ecuador, and Peru; Pilger, 1935; Stauffer, 1961; Nee, 1996; Ulloa Ulloa € Jørgensen, 2002), Jodina (one species, Bolivia, Brazil, Uruguay, and Argentina: Stauffer, 1961; Nee, 1996), species, roughly India and China to Malesia and New Guinea; Macklin & Parnel, 2002; Nianhe & Gilbert, 2003), Pyrularia (two species, eastern United States and approximately India to China; Leopold & Muller, 1983; Nianhe & Gilbert, 2003), and Okoubaka (one to Scleropyrum (four lo six two species, Ivory Coast, Ghana, and Democratic Republic of Congo; Léonard & Troupin, 1950: Villiers, This grouping of genera, never classified at ceae, was previously considered by Pilger (1935) to be 1973a, b). a subfamilial level within the Santala- part of Osyrideae, which is a heterogeneous assem- blage of 21 genera (Der & Nickrent, 2008). All these genera share drupaceous (pseudodrupaceous) fruits with bony mesocarps (not endocarps as sometimes erroneously reported; see Bhatnagar & Sabharwal, 1969), and most possess fruits larger than other members of Santalaceae (greater than 3 cm diam.), although fruit size in Staufferia and Pilgerina is smaller (less than 2 cm long). The genera of our Cervantesia clade (Fig. ers with half-inferior ovaries, whereas four of 1) have hermaphroditic flow- the five genera of the Pyrularia clade (Fig. 1) are dioecious with inferior ovaries (the exception being Pilgerina with hermaphroditic flowers). Features that appear diagnostic for the Pyrularia clade (vs. Cervantesia clade) are the straight (vs. contorted) placental columns, reflexed (vs. straight) petals at anthesis, and dorsibasifixed (vs. dorsifixed) anthers. (i.e. diagnostic) Relatively few autapomorphic character states exist for any one genus in the group, and variation between genera generally involves quantitative and continuous character states. A cladistic analysis of morphological characters, such characters Table 3, unpublished data) that as those selected shown in recovers a tree (Nickrent, shares few features of the molecular tree (Fig. 1), thus suggesting that many of the morphological characters are homoplastic. Despite this, genera in the Pyrularia and Cervantesia clades are defined by unique combinations of character states. For example, Stauffer (1957) listed a number of diagnostic morphological features that distinguish Okoubaka from Scleropyrum, particularly inflorescence structure and fruit organization, which can be extended to differentiate all members of the Pyrularia clade, including our two newly named genera (Table 3). Staufferia differs most notably from Okoubaka, the most morphologically similar genus, by the smaller inflorescences (4 to 10 vs. 50 to 100 flowers); smaller (182 x 9x 5 obovoidal (vs. ellipsoidal) fruits: smaller ( vs. > 3 cm diam.), obovoidal (vs. ellipsoidal) pyrenes [1.1-]1.4-1.6 cm vs. ca. cm), I.1-1.6 em smaller (ca. (i.e., mesocarp plus the enclosed seed): 7-8 mm diam.) persistent perianth; thinner (0.5-1 mm vs. 15-20 mm (vs. smooth) exocarp: and by the thinner (ca. 0.5 mm vs. 34 mm thick), alveolate) mesocarp. Pilgerina is distinguished from 5 mm vs. thick), 5-sulcate smooth (vs. deeply striate or Scleropyrum by the smaller inflorescences (8 to 23 vs. 60 to 100 flowers); pedicellate (vs. sessile) flowers; 7 3 X 2 cm), broadly transversely ellipsoidal to subspheroidal (vs. smaller (1.2— —2.7 em vs. obovoidal to pyriform) fruits; broadly transversely ellipsoidal to subspheroidal (vs. subspheroidal to obovoidal) pyrenes; and by the thinner (ca. 0.5 mm finely striate (vs. vs. 1.5-3 mm thick), smooth o deeply striate or alveolate) mesocarp. An argument could be made to lump instead of split axa in with regard to placement of these two new j distinct novel genera within the Santalaceae. Clearly both are related to the three genera of the Pyrularia clade, but given the topology of the phylogenetic tree Fig. 1). retaining monophyly would require several pane Staufferia were the unfavorable taxonomic. transfers. If considered congeneric with Scleropyrum, two species of Okoubaka would also have to be included, resulting in an extremely morphologically heteroge- neous Scleropyrum. A more radical ay pproac ‘+h would be genera, into => o lump Pilgerina, or Pyrularia (the oldest name, 1803), but this would only exacerbate the problem of heterogeneity. As will be shown below, the genera of the Pyrularia clade are mutually as both by molecular characters, as other genera in Santalaceae, distinct, morphological and thus their classification at the generic rank is not incompatible with the existing classification. When using molecular data to delimit genera within families, workers usually examine phylogenetic trees, identify monophyletic and diagnosable clades, and then name these clades (genera) with the goal of minimizing disruption of existing nomenclature. This approach was recently used to justify recircumscrip- tion of genera that display polyphyly and age e.g., Kellermann et al., 2005; Pfeil € Crisp, 2005: 2005). In the latter study, the genus tribe a ptr Alejandro et al., Mussaenda l. (Rubiaceae) recently and relatives in Mussaendeae were reclassified based on able 3. Comparison of some selected morphological features among nine exemplar species of the Pyrularia and Cervantesia clades. Table based on a matrix of morphological characters (Nickrent. unpublished data Exocarp Mesocarp Anther Placental Fruit diam. Exocarp Fruit Pyrene thickness thickness Mesoca Taxon Floral bract Pedicel attachment column (cm) surface indument shape (mm) (mm) ornamentation Acanthosyris persistent absent or dorsifixed twisted > 3 smooth glabrous ellipsoidal l 3—4 — smooth asipapote subsessile Cervantesia caducous absent dorsifixed twisted <3 5-valved (basally glabrescent ellipsoidal- 0.3 0.2 smooth tomentosa and apically spheroidal dehiscent from fruit) Jodina rhombifolia persistent absent dorsifixed twisted « 3 5-valved (basally pubescent ellipsoidal- 0.5-1 0.1-0.2 smooth and apically spheroidal dehiscent from fruit) Okoubaka persistent present dorsibasifixed straight >3 smooth glabrous ellipsoidal 15-20 3—4 deeply striate or aubrevillei alveolate Pilgerina persistent present dorsibasifixed straight < 3 smooth glabrous broad 0.5-2 0.5 . smooth or finely madagascariensis transversally striate ellipsoidal to subspheroidal Pyrularia pubera caducous present dorsibasifixed straight < 3 smooth glabrous or obovoidal 0.5-1 0.5-1 smooth or sparsely verrucose pubescent at least near apex Scleropyrum persistent absent dorsibasifixed straight <3 smooth glabrous su p vg to 1-3 1.5-3 deeply striate or pentandrum obov alveolate Staufferia caducous present or dorsibasifixed straight <3 5-fused segments pubescent or obovoidal 0.5-1 0.5 smooth capuronit absent (not dehiscent) glabrescent cor əy} Jo s¡euuy USPJEL) [jeoiuejog UNOSSIN Volume 95, Number 2 2008 Rogers et al. bo and Pilgerina (Santalaceae) 403 molecular data. Combined trnT-F and ITS data were used to justify lumping Aphaenandra Mig. into Mussaenda, maintaining Pseudomussaenda Wernham as a distinct genus and renaming a clade (composed of both Landiopsis Capuron ex Bosser and Mussaenda p.p.) as a new genus Bremeria Razafim. & Alejandro. Unlike the Pyrularia and Cervantesia clades are monotypic or these analyses, however, genera of small, with each member recognized by distinct diagnostic inflorescence and fruit features (Table 3; see also Stauffer, 1957). Molecular data have also been used to show that the taxon “genus” is not equivalent within vertebrates (Johns & Avise, 1998) and euascomycetes (Lumbsch, 2002), and it is reasonable to extrapolate this finding to other groups, including angiosperms. The reasons for inequivalency were summarized by Lumbsch (2002), as were the effects this has on nomenclature, specifically the ongoing debate between those advo- cating the use of Linnaean ranks versus rankless Our approach in this paper is essentially a combination systems that utilize phylogenetic. principles. of these two philosophies. By naming two new monospecific genera, we demonstrate our recognition of this rank, while at the same time we adhere to the concept of monophyly. Moreover, because molecular data are available for these taxa, we are compelled to utilize distances to maintain intergeneric genetic internal consistency in our classification. ee. 95 Using our four-gene matrix, uncorrected “p distances were calculated using PAUP* (Swofford, 2002) for all pairs of Santalaceae taxa (other models of molecular evolution, including the general time reversible, gave comparable results). The average P value between the Pyrularia and Cervantesia clades was 0.039. Within the Pyrularia clade, intergeneric averaged 0.0165, the Cervantesia clade the average was 0.0125. The latter changed little whether or both Acanthosyris were included. Thus, there is greater differentiation between Old World distances whereas within one species of it appears that genera than between New World genera, at least for the genes used in this study. As a reflection. of the lower level of generic distinctiveness of Cervantesia and Acanthosyris, it can be pointed out that several species have been transferred from the former to the latter genus: C. colomibrana A. C. Sm. (Smith, 1937) to A. colombiana (A m.) Cuatrec. by Cuatrecasas (1950) and C. 1906) to A. glabrata (Stapf) Stauffer by Stauffer (1961). The average intergeneric poe Stapf (Stapf, distances within the Pyrularia clade ranged from 0.0136 (Staufferia) to 0.0188 (Okoubaka) with a mean across all genera of 0.0165. The distance for Pilgerina (0.0176) was greater than the mean distance values for both the Pyrularia clade and the Cervantesia clade (0.0125). This approach provides additional, internal- ly consistent. genetic data that justify recognizing these taxa as new monospecific genera, a result that is the level of angiosperm taxa on Madagascar. not surprising given endemism of Our phylogeny suggests that in the Pyrularia clade, the North American taxa are sister to an Afro-Asiatic clade, whereas the Malagasy taxa form a paraphyletic group. Using the age estimation of Malécot (2002), the divergence between Pyrularia and other taxa occurred during Middle Eocene, while the split of Okoubaka and Scleropyrum took place sometime in the Late Eocene. Given this, the speciation of the two Malagasy genera may have occurred between the Middle and Late Eocene, an age consistent with divergence dates derived in other groups (e.g., cichlid fish | Y oder et al., 1996], 2001], chameleons [Raxworthy et al., 2002], and Acridocarpus Guill. & Perr. in the Malpighiaceae [Davis et al., 2002)). lemurs [Vences et al., Literature Cited Alejandro, G. D., S. G. Razafimandimbison & S. Liede- Schumann. 2005. Polyphyly of Mussaenda udis from ITS and trnT-F data and its implication for ge neric limits in d. (Rubiaceae). Amer. J. Bot. 92: 544—557. 1969. rud and embryology of lodina rhombifolia Hook. & Arn. Beitr. Biol. Pflanzen 45: - Capuron, R. 1957. Essai d'Introduction a l'Etude de la Flore Bhatnagar, S. P. € G. Sabharwal. 165—419. Forestiére de Madagascar. Inspection Générale des Foréts, tananarivo. 968. Materiaux pour l'Étude de la Flore Forestiere de Madagascar: Olacaceae, Arbustives ou Technique - Forestier, Opiliaceae et Santalaceae Arborescentes Madagascar. Centre Tropical Sl tion de Madagascar, Cavaco, A. M. Keraudren. 1955. Santalaceés 58). Pp. 1-9 in H. Humbert (editor), Flore de Mada- gascar et des Comores (Plantes Vasculaires). Didot, Paris. Cornet, A. 1974. Essai de cartographie bioclimatique a Madagascar. Notic. Explic. No. 55. ORSTOM, Paris. Cuatrecasas, J. 1950. p in South American plants Il. Fieldiana, ae = Davis, C. C., “Bell. D Fritsch & S. Mathews. 2002. Phylogeny of ae PM arpus-Brac hylophon (Malpighiae eae): (family Firmin- Implications for Tertiary tropic al. floras and Afroasian el Evolution 56: 2395-2405. Der, J. P . L. Nickrent. 2008. A mole cular phylogeny of ape (Santalales). Syst. Bot. 33: 107-11 Hallé, N. 1987. Okoubaka Palles & Nonmand est vraiment un ge de Santalaceae. Bull. Mus. Natl. Hist. Nat., B, a iU 9: 355-30: Hieronymus, G. 1889. Sinala ‘eae. Pp. 7 ler & K. Pranil (editors), Die Natürlichen Pflanzen- milien, Ist ed., Vol. 3(1). W. Engelmann, Leipzig. IUCN 2001. IUCN Red List Categories and ( IUCN Species Survival Commission, Gland, Switzer- land and Cambridge, United Kingdom. 202-227 in A. Criteria Version 404 Annals of the Missouri Botanical Garden Johns, G. C. genetic dis SNP |. A vise. 1998. A comparative summary of sin ke vertebrates from the mitochondrial . Biol. Evol. 15: 1481-1490. . Ladiges. 2005. Phyloge- and ge neric limits of the tribe Pomaderreae cytochrome b gene. Kellermann, J., F. l netic analysis (Rhamnaceae) using internal transcribed spacer DNA sequences. Taxon 54: 619-631. Léonard, J. & G. Troupin. 1950. Observations sur le genre Okoubaka Pellgr. et Normand (Octoknemaceae). Bull. Jard. Bot. at ean 0: 11-14. Leopold, D. J. R. Muller. 1983. pubera Mic in the Castanea 48: Louis, J. & J. vien 1948. Okoubaka. Pp. 292-293 in Comité Exécutif de la Flore du Congo ae (editors). 1948. Flore du Congo Belge et du Ruanda-Urundi, Vol. 1 INEAC, Bruxelles. Lumbse h. H. T. 2002. How genera in euascomycetes? Perspect. Pl. Ecol. Evol. Syst. 5: 91-101. Macklin, J. A J. Parnell. 2000. An introduction to the Santal- aceae R. Br. of Mp Thai Forest Bull., Bot. 28: 112-122. ——— An account of the Santalaceae of Thai aes Bull., Bot. 30: 75-115 Thailand. V. 2002. Histoire, a el Pologne des Hosts of Pyrularia field and in culture. E? a 'eae 38-145. a are Malécot, Olacaceae R. Brown as s). Ph.D. T Pierre et Marie Curie-Paris VI, Paris. Nee, M. 1996. A new species of Acanthosyris (Santalaceae) from Bolivia and a key to the woody South Santalaceae. qe a 48: 574—519. Nianhe, X. & M. G. Gilbert. 2003. Scleropyrum. P. 210 in W. Zhengyi, P. E Raven € H. Deyuan (editors), Flora of Vol. 5. Science Press, Botanical Garden Press, St. Louis. Nickrent, D. L. 1994. From field to film: Rapid sequencing methods for field collected plant species ;. BioTechniques 16: 470—475. & hesis, Université American China, Beijing, and Missouri — V. Malécot. 2001. A molecular phylogeny of Santalales. Pp. 69-74 in A. Fer, P. Thalouarn, D. M. Joel, L. J. Musselman, C. Parker & J. A. C. Verkleij (editors), Pro- ceedings of the 7th International Parasitic Weed Symposium. Faculté des Sciences, Université de Nantes, Nantes. Palumbi, S. R. 1996. Nucleic acids II. The reaction. Pp. 205-248 in D. M. Hillis, C. Mable (editors), Molecular Systematics ates, Sunderland, Massachusetts. Pfeil, B. E. & M. D. Crisp. 2005. What to do with Hibiscus? A proposed nomenclatural resolution for a large and well known genus of Malvaceae and comments on paraphyly. Austral. Syst. Bot. 18: 49-60. Pilger. R. 1906. Santalaceae. Pp. 102-103 in. R. Schlechter (editor), Beiträge zur Kenntnis der llora von Neu- Kaledo- nien. Bot. Jahrb. ———. 10908. Salas olymerase chain Moritz & B. K . Smauer Ássoci- (editor), e zur Kenntnis der Flora von e. Bot. Jahrb. Syst., Vol. 40, Beiblatt no. 92. W. Engelmann, Leipzig. Beitrüg 1924. Die Santalaceae Neu-Guinea, mit beme lunes n über die gattung Exocarpus im allgemeinen. Bot. Jahrb. Svst. 59: 28. von 1935. Santalaceae. Pp A. Engler € K. Prantl (editors), Die Mide hs ecd 2nd ed., Vol. 16b. W. Engelmann, Leipzig. Rambaut, A. 2004. Se-Al Sequence EU ent Editor, fers . Department of Zoology, University of Raxworthy, C. J., M. R. J. Forstner & R. A. Nussbaum. 2002. Chi ameleon Jada by oceanic dispersal. Nature 415: 84-787. Se hate G. E. 2000. Endemism in the Malagasy tree flora. 1-9 i E R. Lourenço & S. Goodman ee Der pe Endemism in Madagasc ar, Mémoires de été de B oe MNHM. e Par ——— . Lescot. 2007. Collecting aliti Société de Do adis ales to Malagasy Botanical les Missouri: -Botanical Nd an web- site (http: //www.mobot M( IR g / ed 20 March 2008. 1937. A O species of Cenaniaa h.& P. Trop. s 5l: 12-14 Stapf, O. 19 à ill. Mie U. gazelleer/), acces Smith, A. C. 2 "y Kewenses, decades XL., piam 1906: 71-78. Stauffer, H. )51. Santalales-studien I. gatlung Okoubaka Pellegrin et Normand. Bot. Ges. 67: 422—427 . 1961. ls studien VII. Südamerikanische Santalaceae I. Acanthosyris, Cervantesia Vierteljahrsschr. Naturf. Ges. Zü n 106 1969. Santalales-studien X. a ae eine neue e tribus cee T im ‘eae. V o hr. I XLI. zur stellung der Ber. Schweiz. a N vn Jodina -—412. Naturf. Ges. - 2002. PAUP*: Phylogenetic analysis using 2s cand other methods), version 4.0 ssociates, pander an a Massachusetts. Ulloa Ulloa, C. € Jørgensen. 2002. Santalaceae. 1-118 in €. i & L. Andersson (editors), Flora . No. 69. Botanical Institute, sity, Göteborg. b10. Sinauer Göteborg Univer- nces, M., eyhof, R. Sonnenberg, J. Kosuch & M. Veith. 2001. 35 onc E fossils and molec e Cenozoic divergence of Cichlid biogeography of ishes and the adagascar. J. di ogr. 191-1099, Villiers, J.-F. 197: Okoubaka. Pp. 190-192 Aubréville & d Leroy (editors), Flore du Cameroun, Vol. e Muséum National d'Histoire Naturelle, 2 Parts. 973b. Okoubaka. Pp. 190-192 in A. Aubréville & -F. o (editors), Flore du Gabon, Vol. 20. Muséum National d'Histoire Naturelle, Yoder, A. D., M. Cartmill, M. Vilgalys. 1996. primates. Proc. N Paris. K. Smith & R. Ancient single origin for atl. Acad. U.S.A. 93: 5122-5126 Ruvolo, Malagasy Volume 95, Number 2, pp. 201-404. of the ANNALS OF THE MISSOURI BOTANICAL GARDEN Pl was published on June 18, 2008. MU www.mbgpress.org CONTENTS Reconstructing Complex Evolutionary Histories: Gene-Species Trees, Historical Biogeogra- phy, and Coevolution, the 52nd Annual Systematics Symposium of the Missouri Botanical Garden Assumption 0 Analysis: Comparative Phylogenetic Studies in the Age of Complexity __ Daniel R. Brooks & Marco G. P. van Veller Resolving Species Phylogenies of Recent Evolutionary Radiations : L. Lacey Knowles & Yat-Hei Chan Life History Patterns and Biogeography: An Interpretation of Diadromy in Fishes — 2 Lynne R. Parenti Congruence and Conflict Between Molecular and Reproductive Characters When Assessing Biological Diversity in the Western Fanshell Cyprogenia aberti (Bivalvia, Unionidae) Jeanne M. Serb & M. Christopher Barnhart The Impact of Peter Raven on Evolutionary and Biodiversity Issues in the 20th and 21st Centuries, the 53rd Annual Systematics Symposium of the Missouri Botanical Garden The Impact of Peter Raven on Evolutionary and Biodiversity Issues in the 20th and 21st Centuries: Introduction To Peter C. Hoch Pollinator Shifts and the Origin and Loss of Plant Species Diane R. Campbell A Renaissance of Cytogenetics: Studies in Polyploidy and Chromosomal Evolution... J. Chris Pires & Kate L. Hertweck Taxonomic Revision of Roldana (Asteraceae: Senecioneae), a Genus of the Southwestern U.S.A., Mexico, and Central America A. Michele Funston Revisión del Género Junellia (Verbenaceae) ME Paola Peralta, María E. Múlgura de Romero, Silvia S. Denham & Silvia M. Botta Staufferia and Pilgerina: Two New Endemic Monotypic Arborescent Genera of Santalaceae from Madagascar | Zachary S. Rogers, Daniel L. Nickrent & Valéry Malécot Cover illustration. Archaefructus reconstruction by David L. Dilcher and K. Simons, Florida Museum of Natural History. Annals of the Missouri Botanical Ga m Volume 95 umber Annals of the Missouri Botanical Garden Volume 95, Number 3 September 2008 The Annals, published quarterly, contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, independent reviewers. Authors should write the Managing Editor for information concerning arrangements for publishing in the Annals. Instructions to Authors are printed in the back of the last issue of each volume and are also available online at www.mbgpress.org. Editorial Committee Victoria C. Hollowell Scientific Editor, Missouri Botanical Garden Beth Parada Managing Editor, Missouri Botanical Garden Allison M. Brock Associate Editor, | Missouri Botanical Garden Tammy Charron Editorial Assistant, Missouri Botanical Garden Cirri Moran MBG Press Assistant, Missouri Botanical Garden Patricia M. Eckel Latin Editor, Missouri Botanical Garden Ihsan A. Al-Shehbaz Missouri Botanical Garden Gerrit Davidse Missouri Botanical Garden Roy E. Gereau Missouri Botanical Garden Peter Goldblatt Missouri Botanical Garden Gordon McPherson Missouri Botanical Garden Charlotte Taylor Missouri Botanical Cardin Henk van der Werff Missouri Botanical Garden For subscription information contact ANNALS or THE Missourt BOTANICAL GARDEN, Y Allen Mar- keting & Management, P.O. Box 1997, Lawrence, KS 66044-8897. Subscription price for 2008 1s $170 per volume U.S., $180 Canada & Mexico, $205 all other countries. Four issues per volume. The journal Novon is included in the subscription price of the Annals. annals@mobot.org (editorial queries) http://www. mbgpress. org THE ANNALS OF THE Missouri BOTANICAL GARDEN (ISSN 0026-6493) is published quarterly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. Periodicals post- age paid at St. Louis, MO and additional mail- ing offices. POSTMASTER: Send address changes to ANNALS OF THE Missourt BOTANICAL (GARDEN, Yo Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. The Annals are abstracted and/or indexed in AGRICOLA (through 1994), APT Online, BIOSIS®, CAB Ab- stract/Global Health databases, ingenta, ISIO databases, JSTOR, Research Alert®, and Sci Search®. The full-text of ANNALS or THE MISSOURI BOTANICAL GARDEN is available online though BioOne'M (http:// www.bioone.org). © Missouri Botanical Garden Press 2008 The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and their environment, in order to preserve and enrich life. This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). "Er olume 95 Annals umber 3 of the 008 Missouri Botanical Garden A REVISION OF THE SOLANUM Sandra Knapp? HAVANENSE SPECIES GROUP MISSOURI BOTANIC/ AND NEW TAXONOMIC ADDITIONS TO THE GEMINATA OCT 0 1 2008 CLADE (SOLANUM, SOLANACEAE)' GARDEN LIBRARY ABSTRACT R t f Solanum L. (Solanaceae) phylogeny using both plastid and nuclear DNA sequences h ı distin clade comprising the members of Solanum sect. pde ua (G. Don) Walp. (woody, non-spiny solanums with d Or jdm anc de trichomes and usually leaf-opposed inflorescences) also inc dee a small Caribbean group of s do known as section Diamonon (Raf.) A. Child, whose relationships have pre uus been obscure: S. acropterum Gris AE 5: Dnacatpum Dun al, S. havanense Jacq., nd S. troyanum Urb. Work on a DP dd of section Holophylla (G. Don) Walp. (s.1.) | | specie i dy excluded from section Geminata on morphological grounds belong to the more Yee circumscribed g as the Geminata clade. In addition, since the publication of the monograph of Solanum sect. Geminata, seven new spec ies ul these rare, forest understory shrubs have been described (S. chalmersii S. Knapp. S. E Granados-Tochoy & S. Knapp, $. monanthemon S. Knapp, S. naucinum a Knapp, S. pseudodaphnopsis L. A. Mentz & Stehmann, S. sa, DER PUN Granados- Tochoy & C. I. Orozco, and 5. sumacaspi S. Knapp). An additional new species from Ec n 5. elvasioides S. Knapp, is described here. | review the status and eireumseription of section Geminata in the New World and its relation to the Geminata clade, and provide jon riptions for all the species of the Caribbean S. havanense species group and all species previously excluded from section Geminata. Lectotypes are selected for S. acropterum, S. hookerianum A. Spreng., S. myrtifolium Lodd., S. me Vell., S. troyanum, S. argentinum Bitter & Lillo, and S. evonymoides Sendtn., and a neotype is selec ted for S. havanense. Ke l: ey words: Caribbean, IUCN Red List, Neotropics, section Diamonon, section Geminata, Solanaceae, Solanum. The Solanaceae is an economically important, in all habitats, from the driest deserts on the western cosmopolitan family with approximately 3000 species coasts of South America to the dense, wet, tropical in some 90 genera. The family has members occurring rainforest of the Amazon and Southeast Asia. Life I thank the National Science Foundation for pus this work through the ins Ps ue ud Inventory program, award DEB- bg "PBI i oda worldwide treatm ; J. Bennett, J. Mallet silbert, M. Nee, L. Bohs, d Barboza, D. , Neill, C. Ulloa, and Jorgensen for searching din Leiburam specimens latín du dew p ortions of the paper. [E Men insights 2 Solanum taxonomy; the curators of herbaria cite idi in the text for loan of herbarium specimens; ha Photogray eT Unit of the Natural History Museum, especially Kevin Webb, for preparing the photographic figures; and Lynn Bohs, Iris | eralta, and Victoria C. Hollowell for providing useful comments on the original manuscript. E aint of Botany, The Natural History Museum, Cromwell Road, London SW7 5BD, United Kingdom. GC uk doi: 10.3417/2006159 ANN. Missouni Bor. GARD. 95: 405—458. PUBLISHED ON 23 SEPTEMBER 2008. 406 Annals of the Missouri Botanical Garden forms in the family range from canopy trees to minute ephemeral herbs, with enormous variation in between. Members of the family include globally important food crops such as potatoes and tomatoes (Solanum tuberosum L. and S. lycopersicum L., respectively) and a number of widely used drug plants such as Nicotiana nọ, the source of tobacco, and Atropa L., the source of atropine. Approximately half of the species in the family are contained in five genera, the largest and most diverse of which is Solanum L. (Knapp et al., 2004). Solanum is one of the 10 most species-rich genera of flowering plants (Frodin, 2004). With approximately 1500 species (J. Bennett € S. Knapp, unpubl. occurring on all temperate and tropical. continents, the genus occupies an incredibly wide range of habitats and habits, paralleling that of the family. The history of Solanum classification has been reviewed in Knapp (2002a), but the last time the genus was monographed in d Candolle’s Prodromus T was in de its entirety (Dunal, 1852). Current work by participants of the “PBI Solanum” project (see ) will result in a modern mono- graphic treatment of the entire genus available online. Solanum can be divided into 13 well-supported Bohs, 2005: Weese & Bohs, 2007). Solanum sect. Geminata (G. Don) Walp. is one of the monophyletic clades argest sections of the extremely large plant genus Solanum. In Dunal’s (1852) treatment of the group, he included 37 species (the history of classification and species description in section Geminata is reviewed in Knapp, 2002a). Knapp (2002a) included 126 species in section Geminata, defined as woody plants with leaf-opposed inflorescences, white flowers, and green fruits al maturity. These species have been considered difficult and are very similar morphologically; they have “remarkably the herbarium sheet? (D'Arey, 1973: 740). and Morton (1944: 54) described a new taxon (S. orygale C. V. been described as uniform on Morton, a synonym of 5. arboreum Dunal) as “having no outstanding features, but does not agree with any other species," Phylogenetic studies of Solanum using chloroplast and nuclear DNA sequences (Bohs, 2005; Weese & Bohs, 2007) have shown that the members of section Geminata (sensu Knapp. 2002a) form one of the 13 major clades in the genus, if a small group of endemic Caribbean species (see below) and several Neotropical taxa previously excluded from the section are included. Bohs (2005) called this larger, more inclusive monophyletic group the Geminata clade. This clade is well supported (100% bootstrap support in a three-gene analysis, Weese € Bohs, 2007), and, with 144 species currently known (see Appendix 1), it is the largest major clade of non-spiny solanums. and S. Dunal, two Caribbean endemic species previously Solanum havanense Jacq. COnocarpum recognized as the monospecifie Solanum sect. Dia- monon (Raf) A. Child or of unknown relationships, respectively, have been included in the most recent molecular analyses (Weese & Bohs, 2007; J. C. Granados-Tochoy & L. Bohs, pers. comm.). These species are sister taxa al the base of the Geminata clade and are, in turn, apparently sister to Solanum sect. Geminata as recognized by Knapp (2002a). Solanum delitescens C. V. Morton, recognized as a member of Solanum sect. Geminata, but of uncertain relationships by Knapp (2002a), has also been shown to be firmly nested within the Geminata clade in molecular analyses (Weese & Bohs, 2007), confirming its relationships as suggested in the earlier mono- graph. I excluded a number of species from the original treatment of Solanum sect. Geminata (Knapp. 2002a) based on inflorescence morphology, which | considered to be similar to that found in species of Solanum sect. Holophylla (G. Don) Walp. (s. str., see Knapp, 1989). Phylogenetic analyses using DNA sequences (Bohs, 2005; Weese € Bohs, 2007) have also shown one of these species (S. argentinum Bitter & Lillo) to be nested well within the Geminata clade. Subsequent study of the rest of these taxa as part of © the preparation of a monograph of Solanum sect. Holophylla has shown that based on morphology, these previously excluded species are members of an enlarged and redefined Geminata clade. In addition, seven new species of the Geminata clade have been described since the publication of the monograph of Of the additional species included in the more broadly Solanum sect. Geminata (Knapp, 2002a). defined Geminata clade, only S. argentinum, S. havanense, and 5. conocarpum have been used in molecular phylogenetic analyses (Weese & Bohs, 2007; L. morphology, all of the other species included here Bohs, pers. comm.), but based on overall em belong in the larger, monophyletie Geminata clade. For this treatment, | am defining the Geminata clade as the larger group containing all the species treated in Knapp (20022), the set of species related to Solanum havanense, those taxa excluded from Sola- num sect. Geminata previously, and all the species described as members of the section since Knapp (2002a). The clade has few unambiguous morpholog- ical synapomorphies (Bohs, 2005), but its members are generally woody plants with simple leaves that are glabrous or have simple or branched (never stellate) trichomes, terminal or leaf-opposed inflorescences, and white or purple flowers. In my Flora Neotropica monograph (Knapp, 2002a), I divided the members of section Geminata treated into a set of putatively monophyletic species groups. Recent molecular work has shown that these groups are, for the most part, not monophyletic (J. E. Granados-Tochoy. pers. comm.), Volume 95, Number 3 Knapp Solanum havanense Species Group 2008 but sampling for molecular analyses has been relatively sparse within the group, and groupings obtained are very poorly supported. Although these species groups may not be monophyletic, they serve as initial, pragmatic groupings lo assess relationships and aid in identification. To that end, I place each of the species described here in one of the species groups from Knapp (2002a) in which it would belong (Table 1), similar or but in the descriptions I also elaborate on other sympatric species from the entire clade. If these species groups, or the monophyletic eroupings recovered from further analysis, were to be given sectional status (as has been done for the 2008), the name Solanum sect. Geminata would apply to the tomatoes and their relatives; see Peralta et al., grouping containing 5. nudum Dunal, the type species of the light of the phylogenetic data available for the group, | am not section. [n only preliminary formally recognizing any of these groups al sectional level, but continue to use a pragmatic set of species order to aid Whalen, 1989) in communication and further work with this, the largest groups (sensu of the clades of non-spiny solanums. Within the species of the Geminata clade as treated here, trichome morphology and inflorescence strue- ture are both extremely variable attributes. Variation occurs within species, populations, and even individ- of trichome branching, which the uals in the degree traditionally has been one of most important characters used in Solanum taxonomy (see Knapp, 2002a). Section Geminata, as previously treated (Knapp. 2002a), includes species with simple tri- chomes and species with branched trichomes, a conclusion strongly supported by recent molecular analyses (Bohs, 2005) for the As | have studied the members of the Geminata Geminata clade as a whole. clade in its broadest sense, it has become clear that, although useful for identification purposes, trichome morphology is not likely to be a particularly reliable synapomorphy for monophyletic groups. Although distributed primarily in the New World tropics, a he Geminata clade are = species of single species (Solanum spirale Roxb.) is found orest habitats in the Old World, from southern China to Australia. As for Solanum as a whole (see Knapp. 2002b), the species richness of the Geminata clade is cireum-Amazonian, although some species are found in the Amazon Basin. Highest numbers of species are found in Colombia and Peru (with 41 species each), 7 of the —- but the highest endemism is in Brazil, where 34 species occurring there are endemic, most of these to southeastern Brazil (see Table 2). In this paper, l provide descriptions for the additional species shown by molecular analyses to be members of the Geminata clade (see above), species previously excluded from the group by Knapp (2002a), and taxa in the group described since Knapp (20022) (see conservation Appendix 2). I also provide preliminary the to contribute to Target 2 (a assessments for each of species treated here, in order preliminary conservation assessment of all known plant species) of the Global Strategy for Plant (Secretariat of the Convention on 2002) Conservation Biological Diversity, THE SOLANUM HAVANENSE SPECIES GROUP Shrubs or trees; stems glabrous or minutely pubescent with uniseriate trichomes, these often white; new growth glabrous or minutely pubescent; sympodial units plurifoliate. Leaves simple, elliptic to narrowly elliptic, glabrous. Inflorescences terminal or internod- al. often borne on short shoots, simple or branched, usually glabrous or minutely pubescent with simple trichomes. Flowers white or purple, relatively large, to 6 cm diameter, expanding during anthesis, the corolla lobes planar at anthesis. Fruit ellipsoid, dark green or fruiting pedicels erect or blue-green lo purple; deflexed; seeds rounded or ovoid, dark brown. Distribution. All species are endemic to islands in the Caribbean, from coastal and montane forests. Members of what I am calling here the Solanum havanense species group (also see Knapp, 2002a) have long been of uncertain affinity within Solanum. Rafinesque (1836: 76) based his genus Diamonon, with the single species D. coriaceum (Hook.) Raf., on the possession of a "constantly and unequally four-partile” (zygomorphic) calyx shown in the plate of 5. cortaceum Hook. (Hooker. 1827). Although Hooker (1827: opposite plate 2708) stated that flowers were ^ text often with an irregularity” in the corolla lobes, this is not apparent from the plate in his publication. The calyx structure as depicted in Hooker's plate (see Fig. 4) shows a five- parted calyx with two of the lobes fused at the margin, a condition common in Solanum, where the calyx lobes do not split regularly at corolla egress (see description of 5. 1998), not recognizing the relationship of 5. havanense to the havanense below). Subsequent authors (Child, other species included here, considered section Dia- monon as a distinct, monospecific group at the sectional level (as Solanum sect. Diamonon) within Solanum. The inflorescences borne on short shoots and large purple flowers led Nee (in litt., October 1989) to consider 5. havanense a species of Lycianthes (Dunal) Bitter; closer examination of morphology coupled with molecular data (Bohs, 2005) clearly place it in the Geminata clade of Solanum. Two species of the group (S. havanense and S. conocarpum) have been included in the most recent Bohs, molecular analysis (J. C. Granados-Tochoy & L. 408 Annals dle oia Garden a T T. : Table 1. Continued. Fable 1. Species groups in the Geminata clade (Knapp ` [2002a] and species treated here) using the species group names e sstablishe d by Knapp (2002a); several taxa are treated Solanum microleprodes Bitter here as incertae. sedis. They are clearly members of the Solanum nutans Ruiz & Pav Geminata | clade but with no clear affinities to any of the Solanum xanthophaeum Bitter pragmatic species groups listed here. Species treated in this Solanum youngii S. Knapp paper are in boldface. Solanum amblophyllum species group Solanum amblophyllum Hook. Solanum havanense species group Solanum cee Zahlbr. Solanum acropterum Griseb. Solanum elvasioides S. Knapp Solanum conocarpum Dunal Solanum ae Bit Solanun havanense Jacq. Solanum e nie Dunal | troyanum Urb. Solanum tovarii S. Knapp Solon oblongifolium species group Solanum vacciniiflorum Standl. & L. O. Williams Solanum clivorum S. Kna Solanum validinervium Benitez € S. Knapp Solanum mi Bitter Solanum deflexiflorum species group Solanum oblongifolium Dunal Solanum deflexiflorum Bitter Solanum venosum Dunal Solanum imberbe Bitter Solanum pseudocapsicum species group Solanum incomptum Bitter Solanum paca Bitter & Lillo Solanum irregulare C. V. Morton Solanum delicatulum L. B. Sm. & Downs Solanum sieberi Nan Heurck & Müll. Arg. Solanum lia 42 Solanum arboreum voy group Solanum kleinii L. B. Sm. & Downs Solar amnicola 5. Solanum malacothrix S. Knapp Solanum m Van He "urek € Müll. Arg. Solanum pseudocapsicum L. Solanum arboreum Dunal Solanum spissifolium Sendin. Solanum falconense S. Knapp Solanum nudum species group Solanum goniocaulon S. Knapp Solanum acuminatum Ruiz & Pav. Solanum gratum Bitter Solanum aphyodendron S. Knapp Solanum humboldtianum Granados-Tochoy & S. Solanum caavurana Vell. Knapp Solanum an Roe m. & Schult. Solanum laevigatum Dunal Solanum cassioides L. B. Sm. & Downs Solanum lucens S. Knapp Solanum chalmersii S. Knapp Solanum plowmanii : Knapp Solanum cordioides S. Knapp Solanum ramonense C. V. Morton & Standl. Solanum daphnophyllum Bitter Solanum ripense RAE Solanum gertii S. Knapp Solanum roblense Bitter Solanum intermedium Sendtn. Solanum sagittantherum Granados-Tochoy & C. I. Solanum lasiopodium Dunal rozco Solanum monanthemon S. Knapp Solanum tanysepalum 8. Knapp Solanum nudum Dunal Solanum unifoliatum species group Solanum pabstii L. B. 5m. & Downs Solanum bahianum S. Knapp Solanum pseudoquina A. St.-Hil. Solanum bellum 8. Knapp Solanum reitzii L. B. 5m. & Downs Solanum cyclophyllum S. Knapp Solanum restingae S. Knapp Solanum longevirgatum Bitter Solanum santosii S. Knapp Solanum marantifolium Bitter Solanum spirale Roxb. Solanum triplinervium C. V. Morton Solanum ar Rusby Solanum unifoliatum S. Knapp Solanum tepuiense S. Knapp Solanum robustifrons species group Solanum trachytrichium Bitter Solanum abitaguense S. Knapp Solanum trichoneuron Lillo Solanum cucullatum S. Buapp Solanum warmingii Hieron. Solanum heleonastes 5. Knapp Solanum leucocarpon species group Solanum robustifrons Bitter Solanum alatirameum Bitter Solanum superbum S. Knapp Solanum corumbense S. Moore Solanum narcoticosmum vi group Solanum evonymoides Sendin. Solanum dasyneuron S. Knap Solanum leucocarpon Dunal Solanum foetens S. Knapp Solanum lindenii Rusby Solanum narcoticosmum Bitter Solanum oblongum Ruiz & Pav. Solanum nigricans species group Solanum nutans species group Solanum callianthum C. V. Morton Solanum malletii S. Knapp Solanum canoasense L. B. 5m. & Downs Volume 95, Number 3 2008 Knapp 409 Solanum havanense Species Group Table 1. Continued. Solanum cornifolium Dunal Solanum maturecalvans Bitter Solanum nigricans M. Martens & Galeo a ochrophyllum cda Heurck & Mu Arg. Solanum platycypellon S. Knapp Solanum quebradense S. Knapp Solanum arenarium species group Solanum arenarium Sendin. ssum L. B. Sm. & Downs Solanum gnaphalocarpon Vell. Solanum pseudodaphnopsis L. A. Mentz & Stehmann Solanum smithii S. Knapp Solanum tunariense Kuntze Solanum sessile species group Solanum chlamydogynum Bitter Solanum confertise riatum Bitte Solanum monadelphum Van Heurck & Müll. Arg. Solanum obovalifolium Pittier ex Benítez Sl wm oppositifolium Ruiz & Pav Solanum palmillae Standl. Solanum rovirosanum Donn. Sm. Solanum sessile Ruiz & Pav. Solanum triste Jacq. Solanum turgidum S. Knapp Solanum confine species g PONAM capillipes Britton roup Solanum confine Dunal Solanum cruciferum Bitter Solanum ida id 5. Knapp . V. Morton Solanum erosomarginatum S. Knapp Solanum dis simile C Solanum hypocalycosarcum Bitter Solanum leptopodum Van Heurck & Müll. Arg. Solanum leptorhachis Bitter Solanum morii S. Knapp Solan wm nematorhachis S. Knapp 'olanum bbb rd 5. Knapp Solum pastillum S. Knapp V. Morton um pertenue Standl. & C. orti stipulatum Vel Solanum ter al s. UAE Solanum tuerckheimii Gree aom & Standl. Solanum yanamonense 8. Knapp Solanum ie rae C. V. Solanum dolosum species group Solanum dolosum €. V. Morton ex S. Knapp Solanum gonyrhachis S. Knapp aan habrocaulon = Knapp 1 naucinum S. Knapp incertae V. Morton Knapp sedis pond delitescens C. Solanum lasiocladum S. Solanum mapiricum S. Knapp Solanum sumacaspi S. Knapp pers. comm.), and they are resolved as sister taxa at the base of the Geminata clade and are, in turn, apparently sister to Solanum sect. Knapp (20022). Preliminary Geminata as recognized by conservalion assessments are given here for the members of the Solanum havanense group in advance of a paper outlining the specific criteria for formal IUCN listing for all endemic Caribbean species of Solanum (Yoder et al., in prep.). KEY TO THE SPECIES OF THE SOLANUM HAVANENSE SPECIES GROUP Ib. Flowers purple; pedicels not winged. 2a. New growth completely glabrous; leaves drying dark; flowers 3.5-6 cm diam. cT S. troyanum 2b. New growth pubescent with white trichomes, these usually simple but sometimes dendrit- ic. 3a. Berry erect, dark purplish black, fleshy; anther pores minute, never splitting to slits ........ S. havanense 3b. Berry pendent, green, hard and woodv anther pores lengthenin to S slits with age ......... Y. conocarpum l. Solanum acropterum Griseb., Fl. Brit. W.L [Grisebach] 437. 1862. TYPE: Jamaica. s. loc., s.d., J. Waters s.n. (lectotype, designated here, K 000005250!). Figure 1 Small trees to 10 m; stems erect, completely glabrous; bark pale gray-brown when young, darker red on older stems; new growth glabrous; sympodial units plurifoliate, appearing difoliate, but the inflo- rescence one per stem on short shoots. Leaves (2)5.5— 14 X (1)3-5.5 cm, elliptic, glabrous on both surfaces, when dry densely white-spotted with crystal sand (oxalate inclusions), darker adaxially (fide Stearn 499), the margins entire, the apex acute, rounded at the very green the base acute to somewhat attenuate, tip; primary veins 5 to 8 pairs, drying vellow on older leaves, reddish orange on young leaves; petiole 1— 2.5 em, glabrous, markedly corrugated. Inflorescences leaf-opposed and on _ short leafy shoots, 2-6 cm, with 3 to 15 flowers, completely glabrous, the peduncle almost absent to branching 1 to 3 2 cm; pedicels 15-20 mm, strongly and conspicuous- ly 5-winged from the midrib of the calyx lobes, glabrous, nodding at anthesis, ca. 5 mm diam. at the apex, ca. 1 mm diam. at the base, the wings to 1.5 mm wide, articulated at the base; pedicel scars irregularly 1-2 mm slightly above spaced apart, the pedicels articulating the base leaving a peg 1-3 mm on the inflorescence; buds ellipsoid, the corolla ca. 1/2 of the way exserted from the calyx tube, included in the lobes. Flowers apparently all perfect; calyx tube 3— 410 Annals of the Missouri Botanical Garden Table 2. Geographical distribution of the Neotropical members of the Geminata clade (Knapp. 2002 and the species treated here) by country. Countries are listed in alphabetical order, followed by the number of endemic ‘s/total number of species. Country-endemic species are in boldface. Solanum spirale is the only exclusively Old World member of the group; its distribution is given in Knapp (20022). Country (endemics/total) Species of the Geminata clade Argentina (1/14) argentinum, caavurana, compressum, corumbense, delicatulum, delitescens, heleonastes. maturecalvans, platyeypellon, pseudocapsicum, pseudoquina, symmetricum, trachytrichium, trichoneur Belize (0/4) diphyllum, nudum, rovirosanum, tuerckheimii Bolivia (0/20) acuminatum, aphyodendron, argentinum, ehalmersii. confine, corumbense, daphnophyllum. gonyrhachis, heleonastes, leucocarpon, lindenii, mapiricum, maturecalvans, monanthemon, nudum, nutans, ochrophyllum, oppositifolium, platycypellon, pseudocapsicum, robustifrons, sessile, superbum, symmetricum, trichoneuron. tunariense Brazil (17/34) alatirameum, anisophyllum, arenarium, bahianum, caavurana, campaniforme, 'anoasense, cassioides, compressum, cordioides, corumbense, delicatulum, evonymoides, gertii, intermedium, kleinii, leptopodum, leucocarpon, nudum, oppositifolium, pabstii, pseudocapsicum, pseudodaphnopsis, pseudoquina, reitzii, restingae, robustifrons, santosii, sessile, spissifolium. stipulatum, symmetricum, trachytrichium, warmingii Chile (0/1) pseudocapsicum Colombia (9/41) aphyodendron, arboreum, barbulatum, callianthum, cornifolium, eyclophyllum. darienense, deflexiflorum, dissimile, dolosum, humboldtianum, imberbe, irregulare, laevigatum, lasiopodium, laurifrons, leptopodum, leptorhachis, leucocarpon, longevirgatum. lucens, malletii, marantifolium, microleprodes, nematorhachis, nudum, nutans, oblongifolium, eee a pseudocapsicum, psychotrioides, ripense, rovirosanum, sagittantherum, se sssile, sieberi, triplinervium, unifoliatum. vacc iniiflorum, validine vium, venosum Costa Rica (0/15) aphyodendron, arboreum, diphyllum, imberbe, incomptum, narcolicosmum, nudum, pastillum, pertenue, ramonense, roblense, rovirosanum, tuerckheimii, vaeciniiflorum, valerianum Cuba (0/2) havanense, nudum Dominica (0/2) nudum, triste Dominican Republie (0/1) nudum Ecuador (3/35) abitaguense, anisophyllum, barbulatum, bellum, confertiseriatum, confine, cornifolium, corumbense, cruciferum, cucullatum, eye i, deflexiflorum, dolosum, goniocaulon, he eptopodum, leptorhachis elvasioides, hypaleurotric 1. hy] ; leucocarpon, malletii. marantifolium, itur calvans, nudum. nutans, FM T oppositifolium, plowmanii, pseudocapsicum. robustifrons, sessile, smithii. triplinervium. venosum, youngil El Salvador (0/3) aphyodendron, diphyllum, nudum French Guiana (0/4) campaniforme, leucocarpon, morii, oppositifolium Grenada (0/1) nudum Guatemala (0/9) aphyodendron, dasyneuron, diphyllum, narcoticosmum, nigricans, nudum, pseudocapsicum, rovirosanum, tuerckheimii Guyana (0/3) campaniforme, leucocarpon, oppositifolium Haiti (0/1) nudum Honduras (0/4) aphyodendron, diphyllum, nigricans. nudum Jamaica (2/4) acropterum. havanense, nudum. troyanum E (0/2) sieberi, triste Mexico (2/11) aphyodendron, dasyneuron, diphyllum, malacothrix, narcoticosmum, nigricans, nudum, almillae, pseudocapsicum, rovirosanum, tuerekheimii Nicaragua (0/8) aphyodendron, diphyllum, nudum, pastillum, pertenue, rovirosanum, tuerckheimil, valerianum Panama (0/15) aphyodendron, arboreum, darienense, imberbe; incomptum, leucocarpon, microleprodes, narcolicosmum, nudum, pertenue, ramonense, rovirosanum, tuerckheimii, vacciniiflorum. valerianum Paraguay (0/10) argentinum, caavurana, compressum, corumbense, delicatulum, heleonastes, pseudocapsicum, pseudoquina, symmetricum, trachytrichium Volume 95, Number 3 Knapp 411 2008 Solanum havanense Species Group Table Continued. Country (endemics/total) Species of the Geminata clade Peru (11/41) barbulatum, bellum, clivorum, confertiseriatum, confine, cucullatum, daphnophyllum, goniocaulon, habrocaulon, leptorhachis, leucocarpon, lindenii, malletii, maturecalvans, monadel[ abitaguense, acuminatum, amblophyllum, amnicola, anisophyllum, aphyodendron, corumbense, cruciferum, aevigalum, je ptopodum, hu = - - . haucinum, nudum, oblongifolium, oblongum, ochrophyllum, oppositifolium, plowmanii, pseudocapsicum, robustifrons, sessile, smithii, sumacaspi, tovarii, xanthophaeum, yanamonense, youngii Puerto Rico (0/1) nudum cet (0/2 Prinidad and Tobago (0/6) Uruguay (0/1 Venezuela (12/32) N leucocarpon, morii arboreum, capillipes, pseudocapsicum aphyodendron, arboreum, erosomarginatum, falcone campaniforme, se, fi nudum, pseudocapsicum, sieberi, triste capillipes, chlamydogynum, cornifolium, dissimile, tens, gratum, imberbe, lasiocladum, leucocarpon, lucens, morii, nudum, oblongifolium, obovalifolium, ombrophilum, DAGA pseudocapsicum, quebradense, ripense, sessile, sieberi, tanysepalur tenuiflagellatum, tepuiense, triste, Virgin Islands (1/2) conocarpum, nudum turgidum, validinervium 3.5 cup-shaped, glabrous, the lobes 3-8 mm, narrowly triangular to lanceolate, winged abaxially, mm, the wing continuing onto the pedicel, glabrous, the apex rounded; corolla 2-2.5 cm diam., white, lobed 1/3 to 1/2 of the way to the base, the lobes ca. 1 em, deltate to narrowly deltate, campanulate at anthesis, minutely papillose with tangled white trichomes at the tips; filaments with the free portion 1.5-2 mm, the filament tube less than 0.5 mm, glabrous; anthers 4-5 the dorsal face thickened and darker slightly the poricidal at the tips, the pores lengthening to slits with 1-1.5 mm, when dry, rounded-sagittate at base, age; ovary conical, glabrous, the style 8-9 mm, straight, glabrous, the stigma minutely capitate, the surface papillate. Fruit a globose berry, becoming ellipsoid, ca. 1.2 X 1 cm (immature), green; fruiting 1.5 mm diam. at the base, X 2mm, pedicels ca. 2 em, ca. woody, pendent; seeds ca. rounded reniform (immature), dark brown, the surfaces mi- nutely reticulate-pitted. Chromosome number nol known. Distribution. Endemic to Jamaica, where it occurs in woods on limestone hills in the cloud forests of primarily the John Crow Mountains from 100—700 m. Discussion. Solanum acropterum is a distinctive species with its strongly winged, completely glabrous pedicels and ellipsoid berries. The inflorescences are borne one to a short shoot, making the flower-bearing portion of the plant look somewhat bushy. In contrast to the other members of the 5. havanense group, the flowers of 5. acropterum are white rather than purple. Mature fruits of 5. acropterum are not known, but — appear to be ellipsoid and are likely to be blue-green, purple, or black, as are the fruits of S. havanense. Solanum acropterum is known only from the John Crow Mountains and a few other forest habitats 1 ~ 1 central Jamaica; due to this restricted distribution on likely concern (Yoder et al., in prep.) and was listed in 1997 as Rare (Walter & Gillett, 1998). This species can be assigned a preliminary conservation status of Vulner- able (VU) (IUCN, 2001), but there is a reasonable basis to consider S. acropterum as Endangered (EN) a single island, it is be of conservation due to paucity of collections and its status as a single island endemic (Yoder, 2006 Grisebach (1862) cited two collections in his original description of Solanum acropterum, Waters both at K. The Waters sheet (K 000005250) has TM chosen as the lectotype because s.n. and Wilson s.n., it has both flowers and fruit (Fig. 1). Neither sheet has specific collecting locality information. mode (mens examin 1. JAMAICA. s. loc., 1844, urdie s.n. 2 Wilson s.n. (K). Hanover Parish: o: DES & vic., Britton 2294 (K, NY), Harris 10298 (BM, F, NY, US). Por i Parish: E slopes, John Crow Mins., Britton 4137 (NY). Saint Ann Parish: Union Hill, near Moneague, Britton & po 2807 (NY, US); Stirling Castle Fo ju i "serve, vic. Alumina aia a tramline Howard 15024 (A). i Thomz rish: E slopes of S d of John di row Mins., Harris & Britton iud à K . US), Harris 10752 (F, ', US); Big Level Distr., SE end of the John Crow Mins.. Pr roctor 11802 (US), pup 499 (A, BM [4], K [2], MO), Webster & Proctor 5546 (BM, ES d: , US). Trelawny Parish: Paynes Patent, WNW [Oak Step, 9 Mar. 1949, Lewis s.n. (US); W vi a 1490 (US). 2. Solanum conocarpum Dunal Poir., Encycl. Suppl. 3: 748. 1814. TYPE: U.S. Virgin Islands. St. John: Coral Bay, L. C. Richard s.n. (holotype, P!; isotype, F 976721 [fragm.]!). Figure 2A. Annals of the Missouri Botanical Garden Figure A K000005250 LECTOTYPE LECTOTYPE of Solanum acroptevam Ori seh. Det. Sandra Knapp BM 2005 i |. rrt. WT. 422, (862 Databased for the PBI Solanum project Det. Sandra Knapp BM 2006 ELC 7 s i is HAS 23:54 A v. — e « e eats 1. Lectotype specimen of Solanum acropterum (Waters s.n. [K]). Volume 95, Number 3 2008 413 Solanum havanense Species Group gu Acevedo). —B. s. CH flower phot, C. D. Adams). cm . Adams). Scale bars = 1 flowers (photo, ( Shrubs to 2 m; 10) brown, stems + stout, erect, pale grayish new densely white glabrescent; growth pubescent with simple uniseriate trichomes ca. 0.5 mm, soon glabrescent; bark white, exfoliating; sympodial units plurifoliate. Leaves 3-6 X 1.5- 2.5 cm, elliptic, glabrous on both surfaces, the base acute, the margins entire, the apex acute to primary veins 7 to 8 pairs, 0.5-0.8 cm, Inflorescences leaf-opposed, terminal on Son axillary 0.4-0.6 cm, densely pubescent with simple, uniseriate trichomes to 0.5 4 mm, acuminate: sometimes drying yellowish green: petiole glabrous. shoots, simple, with l to flowers, mm, the peduncle absent to 0.3 em; pedicels 2— stout, ca. | mm diam. at base and apex, erect at anthesis, glabrous, articulated at the base; pedicel scars closely spaced, somewhat overlapping; buds globose, the corolla ca. 1/2 to completely exserted. Flowers heterostylous; calyx tube 1-2 mm, broadly cup-shaped, glabrous, the lobes 1.5-2 mm, deltate, glabrous with an apical tuft of uniseriate trichomes; corolla 1.5-2 em diam., bright purple or violet, lobed ca. 1/4 to 1/3 of the way to the base, the lobes ca. 2 X | mm, deltate, broadly planar at anthesis, densely papillose-pubescent at the tips: filaments glabrous, F lowe S and fruit of me ambe STS of the Solanum havanense Spec ie S gr ro up 0.8 .—A. 5 conocarpum flower (photo, 5: havanense fruit (ha t . D. Adams). —D. troyanum — the free portion 1-15 mm, the filament tube less than 0.5 mm or absent; anthers 2-2.5 mm, poricidal at the tips, the pores with thickened edges, slightly lengthening to slits with age; ovary long-conical, the style 6-8 mm in long-styled flowers, less than 2 mm in short-styled flowers, glabrous, the stigma broadly capitate, the surface minutely papillose. Fruit a conical berry, 3-5 X 2-3.5 cm, hard and green al maturity, the pericarp woody, the apex sharply pointed; fruiting pedicel 2-3 em, ca. 5 mm diam. at the base, woody and pendent from the weight of the berry; seeds not seen. Chromosome number not known. Distribution. Endemic to the island of St. John in the U.S. Virein Islands (and perhaps also St. Thomas 8 I I 1996), in forest and Virgin Gorda, fide Acevedo-Rodríguez, coastal habitat, growing in seasonally dry, understory or in open areas from 0-500 m “Marron bacoba” (Dunal, 1852). Common name. Discussion. Solanum conocarpum is quite similar in vegetative morphology to S. havanense, and it inhabits the same sorts of coastal habitats. The two species do not appear to be sympatric, and 5. 414 Annals of the Missouri Botanical Garden havanense is found on Cuba, Jamaica, and the Cayman The flowers of S. conocarpum are more stellate in outline Islands but is not found in the Virgin Islands. than those of 5. havanense, and usually have a distinct white eye (see Fig. 2A, B). species are completely different: S. conocarpum has The fruits of the two very large, conical berries that are green al maturity and are borne on woody, pendent pedicels, while the berries of 5. havanense are smaller, ellipsoid, and dark purplish blue at maturity and are borne. on erect pedicels (Fig. 2C) Solanum conocarpum was proposed for listing as endangered under the U.S. Fish and Wildlife Endan- Act in 1998, was made that the listing was not warranted due to lack gered Species but in 2006 the decision of information about this species (). paucity of collections and the current population data, Based on the this species can be assigned a preliminary conservation 2006). The conocarpum from Puerto Rico cited in the 1997 IUCN Red List (Walter & Gillett, 1998, based on a record in The status of Critically Endangered (Yoder, record of 5. Nature Conservaney's 1996 National Heritage Database, not available online) is likely to be an error. | have seen no authentic material of S. conocarpum from Puerto Rico. Acevedo-Rodríguez (1996) knew of Solanum con- ocarpum only from a single plant on St. John. but suggested it might also occur. on Thomas and Virgin Gorda. | have seen no material from either island, although Acevedo-Rodríguez cites a sterile specimen (Little 23836) from Virgin Gorda that I have 2005) have found fewer than 200 individuals of S. conocarpum in not examined. Forbes et al. (pers. comm., the wild and have suggested that the species is self- incompatible, since plants do not set fruit when grown in isolation, Solanum conocarpum has suffered severe 2002; A. M. Stanford, pers. comm.), indicating that this species is loss of genetic diversity (Stanford et al., a high priority for listing at both a global and national level. U.S. VIRGIN ISLANDS. along trail from Lameshur to MO M a ns examined. St. John: Reef Bay Quarter Reef Bay ruins, FRA et al. 5437 (K, ). 3. Solanum havanense Jacq., Enum. 1760. TYPE: (neotype, designated here, tab. 35 1763!). Figures 2B, C, 3. Syst. PL 15. 1763 in Jacquin, Jamaica. Tab. 35, Jacquin, Solanum hookerianum A. Spreng., Tent. Suppl. 1828, replacement name for Solanum coriaceum E. 1827. Replaced synonym: Solanum coriaceum Hook.. Bot. Mag. 54: 2708. 1827. Diamonon coriaceum (Hook.) Raf., Fl. Tellur, 2: 76. 1836 |18 Suburb. Solanum UA bis ull. a L 1845. nom. illeg., supe. TY p E Plate 2708, based on a plant Mr e in England, the seeds said to be sent by R. m from Mexico (lectotype, designated here. plate 2708 11 Hooker, 18271). Solanum E es Bol. TYPE: Plate 143 London a. oddiges, 1828!). bos ra Vell., Cab. 25: pl. 1431. 1828. . based on a plant cultivated in I designated here, plate 1431 in . Flumin. 82. 1829, nom. illeg. TYPE: Icon. Tab. i razil. Rio de Janeiro: ted Um designated here, tab. 94 in Vellozo, 1827 [18: Solanum paint dip mn Prodr. e ) 18(1): 160. 1852. Pob dd (Dunal) Ti uba. Havana, F neg. . 330441. IDC mum havanense Jac Griseb.. Cat. PL Cub. 189. 1866. de la Sagra 294 Maong, T ~- mie a he 800-61.2075:1.2 bark dark new growth densely pubescent with Shrubs to 3 m; stems erect, glabrescent; or pale gray: minute appressed, simple or occasionally dendritic uniseriate trichomes to 0.5 mm, soon glabrescent; sympodial Leaves (1.5)2- 0.8)1.1—4.5 brous and shiny adaxially, sparsely pubescent abaxi- units plurifoliate. — em, elliptic, thick and coriaceous, gla- ally with minute simple and = dendritic uniseriate trichomes less than 0.5 mm along the main veins and with scattered longer trichomes to 1 mm, the base attenuate, the margins entire and revolute, the apex acute or sometimes rounded; primary veins 5 to 6 pairs, drying reddish orange; petioles 0.2-1 cm, pubescent with uniseriate trichomes like the stems. Inflorescences leaf-opposed, internodal or sometimes 0.4-1.3 cm, densely pubescent with simple uniseriate peduncle 0.2-0.5 em; pedicels 0.6-1 em, + fleshy, ca on short shoots, with 3 to flowers, simple, trichomes to 0.5 mm, the | mm diam. at base and apex, erect at anthesis, sparsely pubescent with like the inflorescence, 1-2 mm articulated at the base or with a small peg simple uniseriate trichomes articulated the base; pedicel scars ca. apart, remaining alter abscission; buds globose, the corolla included until just before anthesis. Flowers apparent- ly all perfect: calyx tube 1-2 mm, cup-shaped, the lobes 3-5 mm, lanceolate, tearing at anthesis along the sinuses, sparely pubescent with a tuft of simple 2-3 cm diam., violet, fading to pale violet or white, lobed ca. 1/4 of the way anthesis proceeds, the uniseriate trichomes at the apex; corolla to the base, the lobes splitting as lobes 0.5-1 em, deltate to ovale, planar at anthesis, densely white papillose at the tips; filaments with the free portion 1—1.5 mm, the filament tube less than 0.5 mm or absent; anthers 3.5— 4 X 1.5-2.5 lengthening; ovary glabrous, somewhat conical, the mm, poricidal at the tips, the pore not style 7.5-9 mm, glabrous, the stigma capitate, the surface minutely papillate. Fruit an ovoid to ellipsoid berry, 1—1.2 X 0.9-1.2 em, dark blue-purple or black at maturity, soft and fleshy, the pericarp thin, the apex Volume 95, Number 3 Knapp 415 2008 Solanum havanense Species Group A Fig . Neotype of ana havanense (Jacquin, 1763: tab. 35). Image reproduced with permission of Botany Library, The M History Muse 416 Annals of the Missouri Botanical Garden rounded; fruiting pedicels 1-1.2 em, woody and erect, ca. | mm diam. at the base: seeds ca. 7 X 3 mm, ovoid, reddish brown, the surfaces minutely pitted, the testal cell walls straight. Chromosome number not known. Distribution. In woods on limestone from 0— 100 m, on Cuba. Jamaica, and the Cayman Islands. — Common name. Dominican Republic: “roblina” (Jimenez 1907) Discussion. Solanum havanense is a striking plant, — with its large purple flowers and bright purplish blue to black fruits (Fig. 2B, C). H cultivation in Europe and appears to have been spread was brought into widely amongst botanic gardens (see synonyms), with some loss of information as to its original provenance. The flowers of S. havanense expand through anthesis, with the corolla lobes becoming both larger and more divided with time (see Fig. 2B; and shown clearly in Hooker, 1827). bright to pale purple, The corolla color also fades from and the lobe orientation becomes more planar as the flower ages (Fig. 2B). Solanum havanense is relatively common and Antilles, and the Cayman Islands (a single widespread in the Greater occurring on Cuba, Jamaica, collection, possibly cultivated), but apparently absent from Puerto Rico and Hispaniola; the single specimen (Jiménez 1907) is mosl probably cultivated. Specimen data indicate that it is from the Dominican Republic relatively common, although its habitat is the coastal risk It has been assigned a preliminary fringe, making it possibly at from | coastal development. [om conservation status of Near Threatened (NT) an standard monitoring is suggested (Yoder, 2006). No specimen of Solanum havanense attributable to Jacquin has been found, despite intensive searches in LINN. and W: Stafleu € Cowan, 1979) where such a sheet, if it We many herbaria (principally BM, ever existed, might have been preserved. No original 1760), The illustration of 5. havanense in Jacquin’s first illustrat- material was cited in the protologue (Jacquin, so any type selected must be a neotype. ed edition of the Selectarum Stirpium Americanarum Historia (Jacquin, 1763), a copper engraving taken os rom original drawings done by Jacquin in the field (see Wiltshear, 1913), is very accurate, and I have chosen it here as the neotype of the species (Fig. 3). The much more detailed description in Jacquin (1763) S. havanense, mentions the distinctive blue fruit of and the plate in the hand-painted color edition of — Selectarum Stirpium Historia Iconibus (Jacquin, 1780 clearly shows this. Similarly, specimens of neither 5. coriaceum nor 5. myrtifolium have been found. A specimen of 5. havanense labeled as “Solanum Bot. Mag.” at Kew (K 00005238) is nol clearly associated with the description and does coriaceum Hook. not have the distinctive calyx morphology depicted in the plate. These names too are lectotypified using the Vellozo's (1825) description of S. havanense may be a re-description of original illustrations (Figs. 4, 5). 8 8 Jacquin’s plant, but no reference to Jacquin is made in — the protologue or accompanying material. 1 have lectotypfied it using the illustrations published in 1831, as no herbarium specimens are extant from Vellozo's work (Carauta, 1973). Solanum hookerianum (Sprengel, 1828: 9) and S. byrsinum (Voigt, 1845: 511) were both proposed as replacement names for S. Hook.. with S. coriaceum Dunal, a spiny vine from northern South coriaceum which is homonymic America. , Additional specimens examined. CAYMAN ISLANDS rand Cayman: 3/4 mi. N of Half Moon Bay, Brunt 2200 iN. CUBA. s. loc., Anon. s.n. (MO); San Antonio, Hitchcoc s.n. (E); s. loc., Mérat de Vaumartoise s.n. (W); 1823, Poeppig s n. (W); s. loc., et (Ws s Vd s.n. (W): s. loc., Wright 302 . G-DC, S, W); Santa Clara, vic. of Sole n Rs cre i n h erosses trail to Guaos, Howard 5740 (GH, JBSD, US). La Habana: Cojimar, Alain 2394 (US), Alain, Bro. 2396 (GH): Sierra del Grillo. Madruga, Alain 5280 (GH); Havana, Anon. s.n. (W); La Playa de Marianao, Curtiss 716 (A, BM, F, G, K, MO. US): Loma Esperon, Ekman 10555 (F); Sierra de Anafe, León 7135 (GH): Tamon Mtes, Rugel 94 (BM). Matanzas: dd Rugel, F. 144 (BM s K). Pinar del Rio: base of erra de Jíbaro, El HER Viñales, Alain 4292 (GH, US); of Mariel, Britton & Earle 7609 (Us); Mariel Tinaja, Ekman 12876 (A, C); Loma Harenales f Sumidero, Schafer 3795 (A, US); Loma de Esperon, on ihe NW outskirts of Ms near border of Habana Ds Webster 3700 GH, US): Sierra de Anafe, Wilson 1 Cintura] Zenfuegos, Ciene : E K 2 Guabairo, Soledad, Cienfuegos. Jack US): E m s Woods, near the E shore of Cienfuegos Bay, a of Soledad, Webster et iid ol | a Villa Clara: Dolores (A [2). DOMINICAN Santiago, | Jiménez 1907 Manchester Parish: Gut. iver, Adams 6333 (BM): | mi. ESE of God's Well, N side of Round ine Stearn 654 (BM). Portland: Priestman's Bay, Bailey 7 J vic. of Port Antonio, mouth of Priestman’s River, Long Bay, Britton 4213 (NY); 104 (U S). near Rural Hill, Proctor 31565 (MO); Turtle Cove, near Bennett Point, about 2 mi. due E of Port Antonio, Stearn 7 (BM): 2. a Saint Andrew Parish: vic. of Kingston, Ferry Rivet Britton 401 (F, NY), Britton 2831 (NY); Ferry Pen, Campbell 6175 (F): hill behind the Quarry at the Ferry, Harris oe (BM, F. NY, US), Harris 10036 (K); Hectors River, 9 Mar. 1909, Harris & Britton 10718 (F); Ferry Pen, Herb. Bot. Dept. Jamaica 6175 on 3), Herb. Bot. Dept. Jamaica 6248 (GH); W side of Ferry Hill, 3/4 mi. up Old f Ferry Rd., Rive or E (BM); | 8076 (BN 1, G [2]. Saint Ann Parish: Fort Point, Se Asprey 2353 (NY); Queens Hwy., 13 Dec. 1970, Pleumon near mouth of Priestman's River, Yuncker 18423 Volume 95, Number 3 Knapp 417 2008 Solanum havanense Species Group N. 2708 Pub, by S,Curtís, Walworth, Jan, 1827, — SwanSo - e 4. Lectotype of Solanum coriaceum (Hooker, 1827: tab. 2708). Image reproduced with permission of Botany T The Natural History Museum. 418 Annals of the Missouri Botanical Garden í D V T e 5. Lectotype of Solanum myrtifolium The Nz Figur y, tural History Museum. Librar (Loddiges. 1828: pl. 1431). Image reproduced with permission of Botany Volume 95, Number 3 2008 Knapp Solanum havanense Species Group 2986 (GH); along Queen's Hwy., about 2 mi. E of Rio Bueno, Stearn 14 eM ME Saint Elizabeth Paris Pe pper Britton 1511 (N A: Kaiser mine area 5 of Gt er Howard & Proctor 13765 ( NY); Pepper. Miller 13 US) Fort harles, Feb. 1936, Sangster s.n. (BM); vic. of Merriman's Point, Stearn 865 (BM). Saint James Parish: Orange River valley, 2 mi. SE of Sign, Proctor 16625 (BM). Saint Thomas Parish: Greal Morass, Adams 11633 (MO): Holland Bay, Purdie 135 (K). Trelawny Parish: near n Bue along coast rd. to Danena Plowman 3241 (A, K, US); W hite Bay, 4 mi. due E of Falmouth, Proctor 23668 (BM); ca. 2 mi. W of Rio Bueno Harbour, Stearn 319 (BM). Westmore land Parish: Negril & vic., & Hollick 2058, 2063 (NY): Negril, Harris 10231 (BM, F, K, NY, US), Harris 10257 (K Y). 4. Solanum troyanum Urb., Antill. 5: 487. TYPE: Jamaica. Trelawny Parish: Mt. o, 29 Aug. 1905, W. Harris 9000 (lectotype. des enel here, BM 0008491631; isotypes, K 000005227!, NY 00111387!). Figures 2D, 6. Symb. Slender shrubs to small trees, to 4 m; stems + stout, erect, glabrous; bark very dark on new shoots, white on older stems, peeling and exfoliating in irregular patches; new growth completely glabrous, drying black; sympodial units plurifoliate. Leaves 4.5-16.5 fleshy, shiny adaxially, matte abaxially, completely X 1.56 cm, elliptic, thick and somewhat glabrous, drying dark, the base attenuate, the margins entire, slightly revolute, the apex acute; primary veins 6 to 9 pairs, drying reddish green; petioles 1-1.5 cm, glabrous. Inflorescences leaf-opposed on short termi- nal leafy shoots, 3—8 cm, simple or occasionally once- branched, with 3 to 10 flowers, glabrous, the peduncle .5-3 em; pedicels 2-2.5 cm, stout and fleshy, 1— .5 mm diam. at the base and the apex, + erect at anthesis, articulated just above the base; pedicel scars 5-8 mm apart, the pedicels flush with the rachis or leaving a peg 0.5-1 mm; buds globose to ellipsoid, strongly exserted from the calyx tube. Flowers apparently all perfect; calyx tube 3-4 mm, broadly cup-shaped, glabrous, the lobes 5-6 mm, deltate and on the tips and overlapping, minutely papillate corolla 3.5—6 cm diam., violet, intensely fragrant, lobed 1/4—1/2 of the way to the base, the lobes 0.5-2 cm, with age as corolla expands, margins; deltate and emarginate, enlarging planar to somewhat campanulate at anthesis, densely white papillate at the tips; filaments with the free portion 2.5—3 mm, the filament tube less than 0.5 mm, glabrous; anthers 6—7 2.5-3 mm, poricidal at the tips, the pores not lengthening to slits with age; ovary conical, glabrous, glabrous, the stigma the style ca. 15 mm, straight, minutely capitate, the surface minutely papillate. "ruit a globose berry, ca. 1-1.5 em diam. (immature), dark bluish green, the pericarp thin; fruiting pedicels 3—4 cm, 1.5-2 mm diam. at the base, woody, erect; seeds 4-5 X 2.5-3 mm, ovoid-reniform, pale tan, the surfaces minutely pitted, the testal cell margins sinuate. Chromosome number not known. Distribution. Endemic to Jamaica, in moist crev- ices and hilltops on limestone, from 500—1000 m. Discussion. Solanum troyanum is a spectacular plant with its large, purple flowers to 6 em diam. and shiny dark green leaves. Like those of 5. havanense, the flowers become larger through anthesis. Solanum similar to S. havanense but, in troyanum is very addition to having larger flowers, is glabrous on all vegetative parts, including the very young growth. Solanum havanense occurs along the coastal fringe on limestone, while S. troyanum is a mountain plant, growing in moist forests on slopes. Solanum troyanum is assigned a preliminary conservation status (IUCN, 2001) of Vulnerable (VU) based on its narrow geographical range (< 20.000 km?) and the length of time since the last specimen was collected (1976: Yoder, 2006), but its status as a single island endemic may warrant an assessment of Endangered (EN). Re-collection of 5. troyanum is a high priority, and given its large, showy flowers, it is a potential candidate for ex situ conservation and propagation as an ornamental. Three syntypes were cited by Urban in the original description (Urban, 1908): Harris 8530, 8681, and 9000. The gathering Harris 9000 from Mt. Diablo was chosen as the lectotype (BM, see Fig. 6) as it represents more complete flowering and fruiting material, and is distributed in three herbaria. Additional specimens examined. JAMAICA. Claren- of Quaco Rock, vic. Ritchies don Parish: vic. . Proctor 33808 (BM, MO). Manchester Parish: Banana Ground, Adams ) (BM, MO); Christiana, 1/2 mi. NW of Christiana, Howard & Proctor 14305 (A, BM); Coleyville, Gourie Forest, ca. | mi. due SW of Coleyville, Proctor 31768 (F, MO). d Parish: Douglas Castle Distr., Proctor 26410 (BM, GH). t Catherine Parish: vic. of Hollymount, Mt. Diablo, D. 2206 (NY. US [2]). Saint ish: Mulgrave, along rd. betw. Mulgrave & ie, Proctor e (I pies sh: Sweetwater, | mi. Trelawny Parish: ea 5.5 mi. NW of Troy, / 12827 (BM); Troy, Cockpit country, Britto near Troy, Harris 8530 (A, BM, ds Harris 8681 (NY |2]). Perkins 1375 (GH); View Hill, Wilson Valley Distr., 1.5 mi. N « of Wa sop, Plowman 3256 (A, K, US): Troy, Proctor 9911 (N Y); Island View Hill, Wilson Valley Island distr 5mi. N of Warsop, Proctor 21339 (BM); Miss Laura’s Hill, Wilson Valley distr., ca. 1 mi. N of W. arsop, Proctor 24845 (A, BI Y, US); Burnt Hill, overhanging Donkey Trail, 1/2 mi. N of Burnt Hill intersection, Read 1918 (BM, GH, US); Troy, iE E n, near Trov (ca e ~ y lez . 3 mi. NNW), Stearn 922 (BM [2]): sland Vie y Hill, N of ANON Webster 13669 (BM); p Hill. W hitefoord 1399 BM, F, MO). ~ 420 Annals of the Missouri Botanical Garden Figure 6. THE NATURAL HISTORY Vi wi T M000 FLORA JAMAICENSIS. PUBLIC GARDENS, JAMAICA, and V YORK JOTANICAL GARDEN um, dol, able DD ATE A4 1904 f Lecto TYPE or Solanum Won ¿ram Urbaz, Sugmb Ab. S: 454. Heg. Det. Sandra Knapp BM 2006 E LECTOTYPE » q "tud 1 Solanum troyanum Det. Sandra Knapp BM 2005 Lectotype of Solanum troyanum (Harris 9000 [BM |). Volume 95, Number 3 2008 Knapp 421 Solanum havanense Species Group ADDITIONS TO THE SOLANUM PSEUDOCAPSICUM SPECIES GROUP Small shrubs, often with woody tap roots; young stems and leaves glabrous or densely pubescent, the trichomes either simple and uniseriate or dendritic; sympodial units difoliate, geminate, the minor leaves similar in shape to the major leaves, but usually smaller. Leaves simple, lanceolate to elliptic, glabrous to densely pubescent, the trichomes like those of the stems. Inflorescences opposite the leaves, usually simple but in some species many times branched. Flowers white, the calyx lobes deltate to leafy and relatively large, the corolla lobes and sepals reflexed at anthesis. Fruit an orange or red berry, soft and juicy at maturity; fruiting pedicels erect, thickened and seeds flattened-reniform, pale yellowish woody; brown, the surfaces pitted. Distribution. Species of the group are widespread in dry habitats, from Mexico to Uruguay. 5. Solanum argentinum Bitter & Lillo, Repert. Nov. Regni Veg. 12: 547. 1913. TYPE: Argentina. Córdoba: prope Córdoba, T. Stuckert 15164 CORD!; isotypes, G [2]!. LIL not seen). Figure 7. Spec. (lectotype, designated here, Solanum ML ie Bitter & Lillo var. chroniotrichum epert. Spec. 913 Bitte Nov. Regni Veg. 12: 549. . co chroniotrichum (Bitter) C. V. Morton, Revis. Argentine Sp. Solanum 183. 1976. TYPE: Argentina. Jujuy: s. loc., i Stuckert 21337 (neotype, designated by Morton, 1976: 183, CORDI: isotype, G!). = Shrubs (clonal?) to 2 m; stems erect, ca. 0.5 cm diam., sparsely to densely pubescent with uniseriate dendritic trichomes 0.25-1 mm, longer trichomes found in more densely pubescent populations, pubes- cence extremely variable and correlated over the entire plant; bark red or dark brown, very shiny; new growth sparsely dendritic pubescent like the stems; sympodial units (2)3-foliate, s occasionally geminate. Leaves (2-)3.5-15 X (0.6- elliptic, adaxially i. to sparsely pubescent with .7 em, elliptic to narrowly uniseriate simple or dendritic trichomes 0.5-1 mm on the veins and lamina, abaxially glabrous to densely pubescent with uniseriate dendritic trichomes 0.5— l mm over entire surface, the base attenuate, the margins entire, the apex acute to acuminate; primary veins 10 to 12 pairs, often drying yellowish green; petioles 0.3-1.2 cm, pubescent. Inflorescences internodal or terminal, 2.5— elabrous to densely dendritic 10 cm, many times (to 6X) branched, with 50 to 200 flowers, sparsely to densely pubescent with dendritic trichomes, the pubescence less dense apically, the 0.5-0.6 cm, at the base and apex, peduncle 2—4 em; pedicels less than 0.5 mm diam. nodding at anthesis, sparsely dendritic pubescent; pedicel scars closely packed distally, to 3 mm apart toward the base of the inflorescence, plane with the rachis; buds globose, the corolla exserted approximately halfway. Flowers all perfect; calyx tube ca. 1 mm, conical, the calyx lobes 1-1.5 mm, narrowly triangular, sparsely dendritic pubescent abaxially; corolla 0.8-1 cm diam., white, lobed ca. 3/4 of the way to the base, the lobes 0.5-0.8 cm, anthesis, glabrous to densely papillate toward the apex deltate, planar to slightly reflexed at abaxially; filaments with the free portion 1-1.2 mm, the tube essentially absent, glabrous; anthers 2-2.5 X ca. 1 mm, poricidal at the tips, the pores lengthening to slits with age; ovary conical, glabrous, the style 4— 5 mm, straight, glabrous, the stigma capitate, the surface minutely papillate. Fruit a globose berry, 0.5— 0.6 cm diam., bright orange or yellowish orange at maturity, the pericarp thin, the pulp sticky; fruiting pedicels 0.7—0.9 cm, ca. 2 mm diam. at the base, swollen just below the berry, ca. 1 mm diam. at a apex, woody and erect; seeds 6 to 8 per berry, 3-3. 2-2.5 mm, flattened-reniform, pale yellowish tan, a surfaces minutely pitted. Chromosome number n = 12 (Moscone, 1992). Distribution. Bolivia, Argentina, and Paraguay, primarily a species of Chaco habitats, but in many secondary habitats in a wide range of elevations from 100-3000 m mmon names. Argentina. “hedionillo del mon- Com ” (Lorentz & Hieronymus s.n.); Salta: “cabia yuyu” n 77); Santiago del Estero: “afata” (Gramaja 16); (Stuckert 18156). 7€ Bolivia. (Var- “cabara Tucumán: “runa-yuyu” Santa Cruz: “amarguillo negro,” “yacurembiú” Presidente Hayes: gas 1052). Paraguay. yuyo” (Soria 564). Discussion. I excluded Solanum argentinum from section. Geminata based on its complex, branched and thought it was possibly more Holo- phylla s. str. such as S. aligerum Schltdl. (Knapp, 2002a). Analyses of DNA sequences (Bohs, 2005; Weese & Bohs, 2007) show that 5. argentinum 1s firmly embedded in the Geminata clade, and that it is inflorescence, closely related to members of Solanum sect. apparently close to 5. pseudocapsicum L. and its relatives. Closer analysis of inflorescence morphology revealed that 5. argentinum lacks the distinctive platform pedicel base morphology of 5. aligerum and relatives (see Knapp, 1989), and that the inflores- character, is ts highly branched cence, despite similar to those elsewhere in the Geminata clade. Solanum argentinum shares with other members of the S. pseudocapsicum species group (sensu Knapp. 2002a) brightly colored fruit with red or orange Annals of the Missouri Botanical Garden Systematic 5 Solanum argentinum t A ne HERBARIUM G. Iso c&ecro TYPE of Solanum 64, mb na n, Ba Hae +h Uo leyut. uec. Vov., Aper Ves. 12: $43. 1413 Det. Sandra Knapp BM 2006 ISOLECTOTYPE Bitter & Lill Solanum Det. Sandra Knapp BM 2006 X J TJ, Te “Oe E UTER Thor ANI Hai d l i tis Boin, Un, Ck TEx Che LORIE A eT 2 5 EN CC. NH ito, I A J CH n beep Nk, Ses X de Hopp Hans y On No à e “Col, Prov! Diny ; TF Pew ^ Delos uf chion nep og Pod 5 roy loja, üt, UM SAC) SL 2 —— Onnar DL, O 5 D. AOL Jeane Ce tt Ce 7t i Sia QD, Cite Mk 5 t a 4 E Le, E Ly eI Ki, c tten Y A, E H Cr Lo, A ARIUM GENAVENSE (G) R B Ko / e. WE y 016911 Ii tudies 1n + Sandra Knapp (BM) 198 Isolectotype of Solanum argentinum (Stuckert 15164 |G). Volume 95, Number 3 2008 Knapp 423 Solanum havanense Species Group pericarp, yellow, flattened seeds, and pedicels that are nodding in flower and erect in fruit. Individual plants of Solanum argentinum vary enormously in pubescence density, from nearly glabrous to densely pubescent over the entire abaxial leaf surface. This variability in pubescence density led to the description of the two taxa now synony- mized under S. argentinum. Rigonatto et al. (2005) examined trichome and leaf surface morphology along the range of pubescence densities and found no differences that would lead them to regard them as different taxa, even at the subspecific level. Pubes- cence density may also vary in response to environ- mental variables such as aridity or elevation. To test this possibility, I scored mature leaves (four leaves below the growing tip of a shoot) into three classes of pubescence density on 160 geo-referenced specimens with elevation recorded (see Appendix 3) from throughout the species range. Each duplicate of a given collection was scored separately, as some collections were obviously not from the same plant (i.e., they were very different in pubescence density). The scoring was a simple three-point scale: 1 = ,3 = data glabrous, 2 = tufts of trichomes in the vein axils pubescence on the entire lamina. The are displayed as box plots in Figure 8. No effect of latitude, longitude, or altitude was apparent, and the confidence limits of each pubescence class over- lapped for each environmental variable (see Fig. 8). of S. composed of large continuous patches of plants of Local populations argentinum are often identical pubescence density. My observations of populations in eastern Bolivia (e.g., Nee et al. 51722, 51723) suggest to me that the species may be clonal, or spreading through underground connections. This, however, has not yet been investigated. Solanum argentinum is a widespread and weedy species. lt can be assigned a preliminary conservation status (IUCN, 2001) of Least Concern (LC). Lillo (Bitter, 1913) cited 16 collections in their original description, most of which were listed Bitter and without herbaria. Those that were identified as being in a particular herbarium were either from Berlin, herbarium Hieronymous (subsequently held at Berlin) or herbarium Stuckert (now at G). Morton (1976) identified potential lectotypes in CORD, but did not lectotypify Solanum argentinum, as he had not seen material from Teodoro Stuckert’s herbarium in Geneva at the time he was writing his treatment of Argentine Stuckert 15164 as the the specimen in G is Solanum. I have chosen lectotype of S. argentinum, annotated in Bitters hand, and duplicates are in CORD and LIL (fide Morton). I have chosen the lectotype as the specimen in CORD so as to place the type of this species in its country of origin. ARGENTINA. Baer (G); Río Juramento, 1972, Morton s.n. (MO, US); FAQIA, e 21 (US); s. loc., Stuckert 8896 (G); Capital Federal, a Ortúzar, nb Vet, Weisz ma (US. Catamarca: De 'smonte, 19631 (US); Santa Rosa, Brizuela 166 (W); Dept. La Paz, n Brizuela 523 (MO): Dept. La Paz, El Aibal, Brizuela V); Dept. La Paz, El Lindero, Brizuela 893 (MO); bs La Paz, El Divisadero, Brizuela 1145 (C); Dept Carrizal, Brizuela 1334 (W); Dept. Paclín, La Pierotti s.n. (US); Dept. Santa Rosa, Lavalle, Quintero 63 (MO); Dept. Paclín, Superi, Risso 156 a Dept. Paclín, La Merced, Risso 239 Dept. Paclín, La Wee ed, Rio Paclin Nte, Risso 295 (NY). Chaco: Avia Terai, Aguilar 161 (W); Dept. Napalpí, Presidencia Roque Saenz Peña, Buratovich 446 (G); Malvárez 1335 (W); Dept. Resistencia, Cote Me 1401 (MO); Dept. 1 de Mayo, Colonia Benítez, pou (F); Dept. b vs Enrique Urién, Schulz 8170 (US); Doe Maipú, Ruta inas Km 75 al N de Saenz Peña, Schulz 15635 (F); UE Benítez, Schulz 15825 (Y), Schulz 17038 (F). € órdoba: Gute nberg, Ferrocarril Centra l Bartlett 19888 (US ; Dept es oe 16820 (G); Cruz ‘del Eje, Stuckert 17237 (G); Dept. Ischilin, Quilino, Villafañe 74 (US loc., Burkart 2585 (G), Pedersen 2585 Additional specimens examined. ca. Ta- a x À G Dept. Resistencia, Fortin Cardoso, . Corrientes: s. MO. US); Corrientes, — near Facultad de Agronomia, Plowman 2729 (F, US). Formosa: Río Pilcomayo, barrancas la altura del Campamento Central, Cordini 67 (US S) comun Ric. 83, i 2622 (MO, US); Dept. Pilagá 3 km de a Piin 7268 (US); Dept. Pilasás. E nillo. Morel 350 (W); Dept. Bermejo, Nuevo PIS Pi erotti 4012 "m NY, US); Dept. Bermejo, Laguna Yema, alrededores de Laguna Yema, Vanni et al. 4334 (NY). Ja alre de dores de la ciudad, Araque M. & dd ¿dr42 e Estate, Sé El Carmen, Peri 0, dur rin (US); Dept. El Ca | MO); Per TICO, Lillo 9837 (US); Dept. Valle Grande, ca. Grande, ca. 7.5 (NY). Salta: Joaquín V. González, Anta, Joaquín V. Gonzalez, Aari 226 (U S); vic. (El Dique), Coronel Moldes, Bartlett 19683 (US): Dept. Orán, Ballivián-Agua Linda, Borsini 744 (W); Dept. Orán, Puesto de Agua Tranquila, Borsini 6721 3 5 km W de Cafayat (W); Dept. Cafayate, ca. e, Rio Colorado, rel & Novara 20717 (G, US); Dept. La Viña. ruta 68, El Hongo, d & ND a (G [2]; Dept. m um Salant ha, bor 8, Charpin 25819 (G [2]); Dept. s para la Medicion de un / g seen 'ación, Meridiano 896 (F); Orán, alrededores, Cozzo 77 Capital, Km 34 al E de Salta, Descole 1494 (W); Cafayate, CMS Hayward 2076 (US), 19 Oct. 1948, Fries s.n. (MO); s. loc., Herb. Hoehne 38942 (US); Dept. Metán, Río Piedra, Huidobro 536 (US); Río Pasaje Juramento, Huidobro 617 (US); Dept. La Viña, Coronel Moldes, Hunziker 1064da (US); Dept. Rosario de vaca: 28 km E de Rosario de la F 28017 (MO); Dept. Rosario de Lerma, Cerrillos, Krapovickas et al. 28397 eee 16 km W de Gral. Ballivian, illa oyo afluente del Río Sehinini 30909 (F, MO); Dept. la Frontera, Frontera, ruta 34, Krapovickas et al. San Martin, Seco, Krapovickas & San Martin, Campo Duran, 424 Annals of the Missouri Botanical Garden o © S -25 5 G : -30 4 : o Longitude ° O C e ~ E . ~< 2000 4 o0 e ge! e 2 — i . < 1000 e e 0 T T T Pubescence class Figure 8. Environme m distribution of pubescence classes in Solanum argentinum. Box plots of the three classes of pubescence (1 = glabrous, 2 = tufts in vein axils. 3 = pubescence uniformly distributed on lamina) are plotted against latitude, longitude, and alütüde above sea level. The al box in each class represents the central 509 of the data (with the median shown as a line). The outer bars re m sent 90% of the nes with the remaining 1096 shown as outlying points. The figure shows that each of the three pubescence classes ov is strongly in all three geographic dimensions. i: each case, Kruskal-Wallis tests suggest that the me dlians do not differ (P > 0.05). = Volume 95, Number 3 2008 Knapp 425 Solanum havanense Species Group Río d Krapovickas & Schinini 39269 (G); Dept. Capital, 20 km E de Salta, ruta Salta a Gral. Güemes, Krapovickas & Cristóbal. pus (G); Dept. Anta, Gonzalez, camino a Puerta Blanca, Krapovickas & ( ME 46303 s Dept. Anta, lra. sección, El Dorado, a 1049 . Metán, Metán, da 155 (G); Dept. Anta, , ies 305 (Vy . Orán, Bande ae hada, 197 (W); I 3663 (F): Dept. Cerrillos, Cerrillos, Meyer 3652 2 (P: De pt. Anta. León Pozo, 30 km al W de Piquete Cabodo, Meyer 9900 (BM. F, NY); Dept. Orán, Orán, Meyer 5048 (F); Dept. Anta, Coronel Olleros, Novara 527 (G [2], MO); Dept. Capital, Salta FFCC a Quijano a la sia de calle eran Novara 535 (G); ruta 68 km 83 entre Alemania y el Guayacán, Novara 4572 (G) Dept. Cuachip pas, Alemania, alrededores del pueblo, Novara 4856 (G [2]; Dept. La Viña, Talapampa, en 93, Novara 5625 17 km NE de J. Ó i sobre el los alrededores de la hostería, ruta 68 Km (G [2]; Dept. Capital, Chachapoyas, Sierra de la Universidad Católica, Novara 5892 (G Viña, adyacencias del digue Puerta de Diaz, 4-6 kn al W de Coronel Moldes, Novara 7248 (G); Dept. Cerrillos, Finca Santa Margarita, Ruta 21, 4 km al N de San Agustin, Novara G [2]; Dept. Orán, Orán, O'Donell 3135 (BM); Dept. ur Santo, antes de Cabeza de Buey, O'Donell 4357 (MO): Dept. Orán, Orán, Pierotti 2 (F); Dept. Rosario de Lerma, Quebrada del Toro, Ruthsatz 364/2 (MO): Dept. i El Piquete, Ragonese 274 (US) Schulz 204 Y) Dept. Chicoana, Dique Cabra Corral, en las Serranas « Guachipas, Alemania, , MO, US). Santa Fé: Dept. 9 de Juli entre E] Bolic he del Turco y el Gato Colorado, Ragonese 3000 (US). Santiago del Estero: betw. Brea Pozo and Río Huayco Hondo, Ferrocarril Central de Argentina, Bartlett 19750 (US): a 20474 (US) Dept. € ; Dept. Silípica, El jar Mrs Talavera, Eu ept. Robles, Vilmes, capital, Capital, Bianc hi 12 (F, 16 (BM, US); Dept. iM Mal Paso, ay 5 (W); Dept. Re ble es, Fernandez, Krapovickas 2715 nes Dept. Pellegrini, El Tigre, Luno 24 (W); Dept. Robles, Turena, Maldonado 227 (F, US); Dept. Copo, Urutaú, Malvárez 653 (MO); Dept. Copo, Pampa de Los Guanocos, Malvárez 691 (W); Dept. Guasayán, El Cevilas, Mar. 1944, Pierotti s.n. (US); Dept. Pellegrini, Estancia El Remate, Venturi 5797 (BM, F, MO, US). Tucumán: la Molle, Castellanos 14825 (US); Dept. Leales, Los Puestos, Krapovickas & Cristóbal 27143 (MO); Dept. Capital, s. p Lille 3 3343 (G [2], US); rd. Tucumán to Racas, railroad ados in open country, Mexia 4341 (MO), Mexia 7834. (BM, MO, US); Burruyacú, Taruca RE Finca San Augine, pg 91A3 (US); D Alta, Las Cejas, O'Donell 4321 (NY); Dept. Enuncijadas, Ousset 112 Je Tapia, Rodriguez 249 | s. loc., Stuckert 8170 (G); Dept. Burruyacü, s. loc 18156 (C), pes 19721 (G); Venturi 2434 (US ept. Trancas, d 'u, Cerro de BOLIVIA. Chuquisaca: prov. Azero, 5 km N de Carandaytí, aida 31246 (MO); prov. Oropeza, Chataquila, Murguia 379 (NY) un Luis Calvo, proximo al lugar de el Palmar Murguia R. NY); prov. Luis Ca l Salvador, zona central, o 4323 (MO, NY). Gran Chaco: Cururenda JO m, Cárdenas 2525 (US Santa Cr TOV Cordillera, Santa Cruz, ca. 200 km hacia el S, Proyecto Abapó Izoz6g, cerca al Rio Grande, Beck 6459 (NY); prov. Cordillera, Aguarati Izozog, en las grandes dunas del Río Parapetí, 500 m al S de Aguarati, D 2017 7 (NY); prov Cordillera, Camiri, Cárdenas 754 (US); prov. Chiquitos, 7 km al E de la planta de gas de Río Grande, sobre la brecha del gasoducto, Fuentes & Novarro 1958 (MO); prov. Cordillera, Cuevo, adentro la sierra cerca la planchada Jardim & Roas-Hurtado 1550 (NY); del Pueldo, 11 km N del Pueblo, . Cordillera, 3 i NW from Camiri, Kuroiwa & Maeda 2332 2 (NY): prov. Cordillera, Alto Parapetí, orilla del Río Parapetí, Michel 2 (MO, NY); prov. : d E € NE sides of Camiri, eu Río Parapetí, /Vee 8 (NY): prov. Cordillera, 2-3 km NW of Boyuibe, along d railroad to Cuevo, Nee 35317 (MO, NY); prov. Cordillera, in Boyuibe, Nee 35330 (NY); Prov. Andrés Ibáñez, city of Santa Cruz, Nee 40420 (MO, NY); Prov. Andrés Ibáñez, Basilio, on rd. from Santa Cruz to O, NY, US); Cordillera, around hwy. and railroad bridges over Río Seco on N side of settlement of Río Seco, along new hwy. from Santa Cruz to Abapó, Nee 49083 (NY); prov. Cordillera, 6 km (by air) 5 of center of Camiri, along dirt rd. to Comunidad Yuti, Nee 51132 (G, NY); prov. Cordillera, 2 km SW of center of Camiri, on SW s side of Camiri, Nee 51149 (NY); prov. Cordillera, 7 km S of Salinas on hwy. from Ten (NY); prov. Cordillera, along rd. from Camiri to Ipatí, 18.5 km N of bridge over Río Parapetí on N side of Camiri, le 51320 (G [2], NY, ); prov. Cordillera, along rd. from Camiri to Ipatí, 9.5 km N of bridge over Río e of Camiri, Nee 51337 (NY); . Nee et al. 51722 (BM, G [2]. NY); prov. Cordillera, S side of bridge over Río Seco in town of Río Seco, Nee et al. 51723 (BM, NY); prov. Florida, along hwy. from Mairana to Mataral, 3 km NE of bridge at Los Negros, /Vee et al. 52101 (BM): prov. La Brecha, Bañados de Izozog, alrededores del hospital y 3 km camino hacia Río Parpetí, Vargas 1052 (MO, NY). Tarija: prov. Avilés, Sunari Pampa, cerca Colon N, Bastian 1347 (NY); prov. Arce, 31 km S of jet. of rd. to Entre Ríos, on rd. to Padeaya, Solomon 10579 (! (MO, US); ca. 25 km S of Tarija toward Padcaya, Wood 9529 (NY). PARAGUAY. Jardín Botánico, Rojas 14486a (W). Alto Paraná: Río / Itaipú Binacional 668 (MO). Boquerón: Jno, Pozo Hondo, a orillas de Río Pilcomayo, Degen & Mereles 3149 (FCQ, MO [2]; a 22 km sobre Picada Mistolar de la Picada 4ta. División de Infantería, Pedro P. Peña, Degen & Mereles 3176 (FCQ); zona del cana Paraguayo, en el Pilcomayo, Degen & Mereles 3179 (MO); Pozo Hondo, orillas del Río Pilcomayo, Mereles & Degen 5570 (FCO, MO); Tyto. Misión-Estancia Oasis, 6 km S del desvío a Pozo Hondo, Mereles & Degen 5616 (FCQ); entrada al Puesto de Buenos Aires, Mereles & Degen 5642 (FCQ, MO); Estancia María Esther, Cañada Madrid, Mereles £ Degen 5859 (FCQ); cercanías del Río Pilcomayo, Mereles & Degen 6114 (FCQ); Avalo Sanchez-Estancia Brusquetti, Soria 364 (FCQ); Retiro Aloncito (Estancia d Campo Mon- ñ anchez, Spichinger l. a de Chevron, ~ antigua '. Cordillera, Boyuibe, 29 (NY Camiri, Nee 44597 (BM, = = prov. Camiri to Cuevo, Nee / T a m = © L4 D a = e. > Cordillera, borde de la taño, 12 km al N del Retiro Avalos RS2185 (FCQ); Estación E a rime he us Pot, Vanni et a 7 (NY). Presidente Hay de Pozo Colorado versus Fortin Gen. Díaz, Bernardi. 20283 (MO, NY); Bruschetti Avalos Sánchez, Soria 364 (FCQ. MO). Estancia 6. Solanum delicatulum L. B. Sm. & Downs, Phytologia 10: 424. 1964. TYPE: Brazil. Santa Catarina: Itajaí, Cunhas, 10 m, 8 Feb. 1955, R. Klein 1151 (holotype, HBR not seen: isotype, US 00027002 [fragm. |! |. B. Sm. & Downs, Phytologia 10: 428. 1964. TYPE: Brazil. Santa Catarina: Papanduva, Lejeadinho, 800 m, 3 Jan. 1962, P. Reitz & R. Klein 11438 (holotype, US!; isotypes, HBR not seen, NY!). Solanum pavimenti L. Annals of the Missouri Botanical Garden Herbs to subshrubs to small shrubs 0.5-1.5 m; stems slender and erect, + densely pubescent with a mixture of simple and branched (dendritic) uniseriate the trichomes closely bark pale growth densely pubscent like the stems with a mixture trichomes ca. 0.5 mm, ap- pressed and ascending: yellowish; new of simple and branched trichomes, especially on the veins; sympodial units difoliate, usually geminate, but leaf distant, some nodes occasionally the pair occasionally trifoliate. Leaves adaxially glabrous to sparsely pubescent on the veins and lamina with a mixture of simple and dendritic trichomes ca. 0.5 mm, abaxially sparsely to densely pubescent on the veins, more sparsely pubescent on the lamina, the trichomes ‘a. 0.5 mm; major leaves the simple or furcate-dendritic, 3-7 X 1.4-3 cm, elliptic to ovate, the base acute, margins entire, the apex acute; primary veins 6 to 7 pairs, drying yellowish; petioles 0.41 cm, glabrous to sparsely pubescent; minor leaves 1.5 1-1.6 em, similar in shape to the major leaves but usually somewhat more orbicular, the base acute, the margins entire, the apex acute; petioles 0.2-0.4 em. Inflores- cences internodal or somewhat leaf-opposed, 0.2— 0.5 em, simple, with 2 to 3 flowers, sparsely = pubescent with a mixture of trichomes like those o the stems, the peduncle 0.1—0.2 cm; pedicels 0.5- 0.6 em. less than 0.5 mm diam. at the base, ca. | mm diam. at the apex, nodding at anthesis; pubescent witl simple and dendritic trichomes like those of the stems and inflorescences; pedicel scars closely spaced, overlapping at the tip of the inflorescence, plane with the corolla included in the the rachis; buds globose, calyx lobe s. Flowers all perfect; calyx tube 1-1.5 mm, the lobes 2-2.5 mm, long-triangular, ads best with simple uniseriate trichomes: 3/4. of lanar corolla 0.9-1 em diam.. white, lobed the way to the base, the lobes 0.7-0.9 em, deltate, l lo slightly reflexed at anthesis, elabrous or sparsely pesi pubescent with scattered. trichomes along the peta midveins and tips adaxially; filaments with the free than 0.5 poridical al portion ca. 0.5 mm, the tube less mm. glabrous; anthers 2-2.5 X ca. 1 mm, he tips, the pores lengthening to slits with age; ovary elabrous, the style 4-5 mm, the 2-lobed to bifid. minutely papillate. Fruit a globose berry. — conical, straight, gla- the surfaces 1-1.2 em brous, stigma diam., red-orange at maturity, the pericarp thin and brittle; fruiting pedicels 0.5-0.7 em, ca. 1 mm diam. at the base, woody and erect; seeds 3—4 —3 mm, flattened-reniform, yellowish tan, the margins thick- ened, the surfaces minutely pitted. Chromosome number not known. Distribution. Brazil (Santa Catarina and Sao Paulo), Argentina (Misiones and Corrientes). and eastern Paraguay (fide Gutiérrez et al.. 2006), forests o J and forest edges at ca. 800 m. Common name. “Canema mirim” (Smith & Downs, 1966) Discussion. Solanum delicatulum was excluded from section Geminata (it was thought to be a member of section Solanum or a related group), and its synonym 5. pavimenti included in the synonymy of 5. pseudocapsicum by Knapp (2002a). Subsequent mor- phological studies of this rare species by L. A. Mentz (Mentz & Oliveira, 2004) have shown that it is a member of the section, and a member of the S. pseudocapsicum group with bright red fruit, and pedicels nodding in flower and erect in fruit. Some authors (G. E. Barboza, comm.: Gutiérrez et al., 2006 listinet based on a suite of fruit and leaf characters that | delicatulum. Studies Solanum pers. delicatulum as regard S. pavimenti and S. think are part of the overall variation in 5. hal show (2006) relatively by Gutiérrez et al. delicatulum is widespread in Argentina and Brazil. and occurs. sporadically in eastern Paraguay, although | have seen no Paraguayan forests and Near or Vulnerable (VU), depending upon material. a plant of understory in along streams, and assessed as Threatened (NT) the circumscription of the species (as including S. here or excluding it, as 2000). pavimenti, as defined delimited by Gutiérrez et al., ARGENTINA. > Austral, segunda Additional SPERARE" eee Rio Ck i Dese ae margen del BR a Sao Paulo, f the Instituto de Botánica. 10 kn 2.2 km E of center of São Paulo, Praga da Sé. Kiten & E 5790 (NY): Cidade Jardim, Hoehne (SPF 10696) s.n. (NY): Cidade Jardim, Hoehne (SP 44929) s.n. (NY). São Paulo: ADDITIONS TO THE SOLANUM NUDUM SPECIES GROUP treelets, or small trees; Shrubs, leaves glabrous or pu young stems and escent, the trichomes simple or occasionally fureate or dendritic, often only on one side of the stem: sympodial units difoliate, usually geminate, usually anisophyllous. Leaves simple, elliptic to lanceolate, glabrous or pubescent, often with tufts of trichomes in the axils of the main veins the Inflorescences opposite the leaves, usually abaxially, trichomes simple or occasionally dendritic. simple, occasionally branched, glabrous or pubescent. Flowers white or greenish white, sometimes fleshy, the calyx lobes large and leafy to deltate, the corolla lobes Volume 95, Number 3 2008 427 Solanum havanense Species Group planar or slightly reflexed at anthesis; anthers usually Fruit a yellowish at maturity; flexed; seeds flattened-reniform, pale tan to vellow. quite stout. ereen, hard berry, becoming o J O fruiting pedicel woody, de- Distribution. Members of the species group are often widespread in secondary habitats through the Neotropics, but some species are more local and occupy primary forest habitats. EQ. 7. Solanum chalmersii S. Knapp, Brittonia 58: 330. 2006. TYPE: Bolivia. La Paz. Prov. Sud Yungas: 7.5 km (by rd.) from Huancané on rd. to San Isidro, moist montane forest, l6 21'S, 67 30'W, 2225 m, 10 May 2001, M. Nee, L. Bohs, 3. Knapp & J. M. Mendoza F. 51777 (holotype, LPB! isotypes, BM!, NY!, USZ!). Shrubs or small trees 2-6 m tall: young stems and leaves densely white pubescent with simple, uni- seriate trichomes ca. 1 mm, composed of 2 to 5 cells: remaining densely white pubescent, occasionally glabrate; bark of the older trunks pale yellowish white; sympodial units unifoli- elliptic t older stems stems and ate or difoliate and geminate. Leaves narrowly elliptic, chartaceous, evenly pubescent adaxially with simple uniseriate trichomes ca. | mm, densely pubescent abaxially with white uniseriate trichomes 1-1.5 mm, the trichomes denser on the veins; major leaves 9-17 X 3-6 cm, the base acute, the margins entire, the apex acute, rounded at the very tip; minor leaves 1.5-3 X 1-2 em, differing from the major ones only in size, but occasionally somewhat rounder in outline; primary veins 9 to 12 pairs, raised adaxially, prominent and yellowish green abaxially: petioles 0.3-0.7 minor leaves, densely pubescent. Inflorescences 1.5— cm in major leaves, ca. 0.5 em ii 5 em, opposite the leaves or occasionally somewhat internodal, unbranched, with 10 to 20 flowers, densely white pubescent with simple uniseriate trichomes 0.5— 1.5 mm, the peduncle 1-3 cm; pedicels 1—1.2 em, tapering from the abrupt base of the calyx tube to a slender base 0.5-0.8 mm diam., articulated. at the base; pedicel scars closely packed, often overlapping: buds when very young appearing globose, the corolla buds later soon exserted from the calyx lobes, the becoming obovoid just before anthesis. Flowers all — perfect; calyx tube 1—1.5 mm, the lobes ca. | mm, deltate to broadly triangular, abruptly con- stricted to an elongate tip ca. 0.5 mm long, densely pubescent with simple, uniseriate trichomes like those of the rest of the inflorescence; 1.52 em diam., stellate, lobed 2/3 to 3/4 of the way to the base, white or sometimes tinged purplish, the lobes ca. 1 corolla 0.5 mm, ovate-lanceolate, reflexed at anthesis, dense- ly and evenly pubescent abaxially with simple uniseriate trichomes ca. 0.5 mm, glabrous adaxially he tips of — except for the densely papillose margins, the lobes densely papillose and somewhat cucullate; filaments with the free portion 0.8—1 mm, the tube 1— 2 mm, glabrous, with small teeth arising between the free portions. of anthers 4-5 X l- 1.5 mm, oblong, slightly sagittate at base, the pores the filaments; teardrop-shaped, opening into longitudinal slits with age: glabrous; ovary glabrous; style 6-7 X ca. 0.05 mm, cylindrical, straight, glabrous, the stigma capitate. Fruit a globose ber rry 11.2 cm diam., green, glabrous fruiting 2.2 cm, the fruiting calyx not accrescent, the pedicels 0.5-1 mm diam. at the base, pendent, woody, lobes ca. 1 X 1 mm, brittle and somewhat patent; seeds 3-3.5 X 2-2.5 mm, flattened-reniform, pale yellow in dry material, the surfaces minutely pitted, the margins incrassate and darker yellow. Distribution. In the understory of montane forest in northern Bolivia, on eastern Andean slopes from 1900-2200 m. both in the forest understory and in disturbed areas along roads and streams, attaining higher population Plants of Solanum chalmersti grow densities in open areas. Discussion. Solanum chalmersii was mentioned as a probable new species in Knapp (20022), but at that time, material was not sufficient to distinguish it from other similar species in the large 5. nudum species group, and good fruiting material was not available in order to place it confidently. Solanum | chalmersü possesses the flattened-reniform, pale-colored seeds, simple uniseriate trichomes, and more or less closely spaced pedicel scars typical of the S. nudum species group. It is most similar to S. acuminatum Ruiz & Pav. of central Peru Bolivia, with which it is nearly sympatric, but differs in its uniform covering of simple uniseriate trichomes (usually confined. to the vein in 5. flowers with longer anthers with sagittate bases, and in its longer fruiting pedicels. In the 5. acuminatum occurs at higher elevations its axils acuminatum), Bolivia, where two species are sympatric, than S. chalmersii. Solanum chalmersii, despite its small overall range, is common where it occurs and forms quite dense thickets at the margins of roads and streams. It can be assessed (IUCN, 2001) as Vulnerable (VU) due to this restricted range (< 5000 km”) and occurrence in only a few locations, but it is of less concern due to its probably weedy nature. Additional specimens examined. BOLIVIA. La Paz: Prov. Sud Yungas: de Chulumani hacia el N unos 5 km hacia Irupana, entrando haci ia Apa Apa, Beck 24480 (NY); . Buchtien 315 ( primary 7.0-9.4 km NE Luteyn & Dorr 13722 (BM, LPB, NY): 5 Sirupaya vic. de Y: through (a Huancané, 52 km (by 428 Annals of the Missouri Botanical Garden rd.) from Huancané on rd. to San Isidro, mois! montane 773 (BM, LPB, NY, USZ); rd. betw Unduavi € Puente e around hotel € As del Cedo & down to the Río Unduavi, Nee et al. 51795 (BM, USZ); Yanacachi, 3.5 M hacia Chojlla (as Chajlla). Seidel & Richter 851 (LPB, : ong bios achi, Mina Chollja, camino a Kacapi, a 261 (LI [rom Wood n. 05 forest, Nee et al. ca. 6 km along rd. f (K). Huancané to San Isidro, 8. Solanum monanthemon 5. Knapp, Brittonia 58: 332. 2006. TYPE: Bolivia. Santa Cruz. Prov. Caballero: Parque Nacional Amboró, Cerro Bravo, próxima a las juntas del Río Alizar y Amparo (20 km al NW de San Juan de Potrero), 17°57'S, 64 54'W, 2000 m, 10-14 Apr. 1994, 7. Vargas C., D. Ayala & J. Quiroga 3138 (holotype, LPB!; isotypes, BM!, NY!, USZ!). Shrubs to 1.5 m tall; young stems glabrous and somewhat shiny; older stems glabrous; bark of older stems shiny reddish brown; sympodial units difoliate leaf often deciduous and and geminate, the minor nodes thus appearing unifoliate. Leaves ovale to elliptic or narrowly elliptic, chartaceous, glabrous and shiny adaxially, glabrous abaxially but with tufts of white simple uniseriate trichomes in the axils of the ca. 0.5 interlaced, arising from the lamina; major leaves 7— veins, the trichomes mm, very thin and 11 X 2-4 cm, the base acute to slightly attenuate onto the petiole, the margins entire, the apex acute to long- acuminate and rounded at the very tip; minor leaves y the midrib —2.1 em, differing from the major leaves on D = in size; primary veins 5 to 6 pairs, only raised adaxially, the veins reddish brown abaxially; leaves, 0.5-0.7 em 1 > petioles 0.5—1 cm in major minor leaves, glabrous. Inflorescences consisting of a single flower (occasionally the scar of a second flower apparently present). opposite the leaves or occasion- ally somewhat internodal, glabrous, the peduncle and rachis absent or the pedicel appearing jointed 0.6— = 0.9 cm from the base: pedicels 3-3.5 em (to 4 em if measured to the base of the inflorescence), deflexed, tapering from the base of the calyx tube to a slender base ca. 0.2 mm diam., articulated at the base; buds when very young appearing globose, the corolla included within the elongate calyx lobes, the buds remaining globose until anthesis. Flowers apparently calyx tube 1—1.5 mm, the lobes 3—5 X ca corolla 0.8-1 em 2/3 of the way to the base, all perfect; 0.5 mm, long-triangular, glabrous: diam., stellate, lobed ca. — white. the lobes ca. 0.5 X 0.3 mm at base. ovate- anthesis, glabrous the papillose tips; filaments with the free portion ca. lanceolate, probably reflexed abaxially and adaxially except for minutely 0.5 mm, the tube ca. 0.5 mm, glabrous; anthers 2.5-3 as l mm, oblong, slightly sagittate at base, connivent, yellow, the pores teardrop-shaped, opening into longitudinal slits with age: ovary glabrous; style A X ca. 0.1 mm, cylindrical, straight, glabrous, the stigma capitate. Fruit a globose berry, 1—1.2 em diam. glabrous; fruiting pedicel 4.5— (immature), green, 5 em, 0.5-1 mm diam. at the base, ca. 3 mm diam. al the apex, pendent, slender, the fruiting calyx slightly accrescent, the lobes 5-6 X ca. 0.5 mm; seeds not known from mature fruit. Distribution. Only known from two specimens collected in cloud forests in Parque Nacional Amboro and Parque Nacional Carrasco on the eastern slopes of the Andes in Bolivia at ca. 2000 m. where Solanum monanthemon grows have Podocarpa- The cloud forests ceae and Myrtaceae as canopy trees (fide label on Vargas et al. 3138). Discussion. Specimens of Solanum monanthemon have occasionally been identified as the more common Bolivian species 5. trichoneuron Lillo. It is superfi- cially similar to that species, but differs in its dark bark and single-flowered inflorescences, which look like a single flower on an articulate pedicel arising opposite the geminate leaves. The apparent articula- tion is actually the joint between the inflorescence the i inflorescences with minute second buds that appar- axis and pedicel, as evidenced by some ently abort early in inflorescence development. Many more collections of S. monanthemon are needed investigate this phenomenon fully. Mature fruits of 5. monanthemon are not known, so species group placement is tentative at present. Overall morphology tufts of hairs in the vein axils, short anthers) suggests 5. monanthemon belongs to the S. nudum species group, as accepted here, but this can be misleading e.g. S. smithii S. Knapp of the S. arenarium Sendtn. — species group also possesses these hair tufls; see Knapp, 20024). Solanum monanthemon can be distinguished from other species with tufts of hairs in the abaxial vein axils by its single-flowered inflorescences and its tree- like individuals habit. It also apparently grows as isolated dense cloud forests, unlike many of the other members of the S. nudum species group that are plants of Open areas along streams and roads both in Bolivia and throughout the Neotropics. Solanum monanthemon is known only from two collections and, thus, cannot be assessed with confidence using the IUCN criteria (IUCN, 2001), conservation concern. A but is likely to be of preliminary conservation assessment Endangered (EN) l area of occupancy) some hundreds conservative would be based on the few c :ollec ‘tions (sma of kilometers apart (indicating a larger extent of 5000 km*), but species occurrence of < further searches for this apparently rare are a priority. Its Volume 95, Number 3 2008 Knapp Solanum havanense Species Group occurrence in the Parques Nacionales Amboro and Carrasco may afford it some protection. Additional specimens examined. BOLIVIA. Cocha- bamba: Prov. Carrasco, Sehuencas, Río Fuerte cerca puente |El Puente?], Parque Nacional Carrasco, /bisch & Ibisch 93.1477 (LPB). 9. Solanum pabstii L. B. Sm. & Downs, Phytologia 10: 427. 1964, TYPE: Brazil. Santa Catarina: Lajes-Sáo Joaquim, bank of the Rio Lavatudo, 1050 m, 22 Oct. 1961, G. Pabst 6200 & E. Pereira 6373 (holotype, US 00027014: isotype. HBR not seen). Figure 9. Treelets or small trees 3—5 m; stems erect, glabrous: bark dark on young stems, yellowish on older stems; new growth glabrous or occasionally minutely papillate: sympodial units plurifoliate. Leaves 3.5-10.5 Xx 2.5 cm, narrowly elliptic to lanceolate, glabrous adaxially, t um the vein axils with resinous dots, glabrous and with distinct pit-like domatia in the vein axils abaxially, the base acute to attenuate, the margins entire, slightly revolute, the apex acute to acuminate; primary veins 6 to 8 pairs, impressed adaxially: petioles 0.2-1 em, glabrous. Inflorescences internodal to ap- parently terminal, 2-8 em, 3 to 4 times branched, with 3 to 50 flowers, glabrous, thin and brittle in dry specimens, the peduncle 1-2 em; pedicels 1.3-1.5 cm, ca. 0.1 mm diam. at the base, ca. mm diam. at the apex, glabrous, + nodding at anthesis, articulating at the base; pedicel scars irregularly spaced 1-10 mm apart, plane with the rachis; buds narrowly ellipsoid, the corolla strongly exserted from the calyx. Flowers all perfect; calyx tube 1.5-2 mm, conical, the lobes 1— 1.2 mm, deltate, apiculate, the margins membranous and paler, glabrous; corolla 1.3-1.9 cm diam., white, lobed ca. 2/3 of the way to the base, the lobes 0.8— 1.3 em, broadly deltate, campanulate at anthesis, the tips densely papillate; filaments with the free portion 1-2 mm, the tube absent, glabrous; anthers 2.5—4 X 1— 1.5 mm, poricidal at the tips. the pores lengthening to slits with age; ovary glabrous, conical, the style 5— 6 mm, straight, glabrous, the stigma clavate to slightly 2-lobed, the surface minutely papillate. Fruit a globose berry, 0.8-0.9 cm diam., pedicels ca. 2 em, not markedly enlarged at the apex; jon green when ripe; fruiting seeds not known. Chromosome number not known. of Santa Catarina, Paraná, and Rio Grande do Sul, in Atlantic Forest and forest margins, from 600-1300 m. Distribution. In the Brazilian states Common “Canema” 1966). name. (Smith € Downs, Discussion. Solanum pabstii is a very striking species, with its large, open, many-branched inflores- The inflo- rescence axes of 5. pabstii are extremely thin and filiform, cences of relatively large, white flowers. and are brittle on herbarium specimens. Specimens of S. pabstii have been misidentified as 5. pseudoquina A. St.-Hil., another species of the S. nudum species group occurring in southern Brazil. The two species share filiform inflorescence axes (although the inflorescences of S. pabstii are much more branched) and leaves that dry pale green on the herbarium sheet. They differ in flower morphology, with S. pabstii having all anthers of equal size and dehiscence of pores lengthening to slits with age, and S. pseudoquina having markedly unequal anthers that dehisce by tiny pores that never lengthen to slits. Solanum pseudoquina has tufts of uniseriate trichomes in the vein axils on the abaxial leaf surfaces, while 5. pabstii has small resinous patches on the new growth in the vein axils that develop into cavities with resinous margins in mature leaves (see Fig. 10). The avities are probably domatia (sensu O'Dowd & Willson, 1991) and are a distinctive feature of the species, found nowhere else in the Geminata clade, and to my knowledge nowhere else in Solanum. Not all mature leaves develop the cavities to their full extent, but at least some leaves on every specimen examined had them. The ecological function and development of these cavities will be an interesting study in plant-insect interaction, should they be found to house mites. Although the seeds of S. pabstii are not known at present, I suspect that they might be flattened- reniform with enlarged margins. Solanum pabstii conforms to other characters of the but phylogenetic analyses. S. nudum species group. as yet has not been included in Solanum pabstii is a species of disturbed habitats and of relatively wide distribution and, thus. can be assigned a preliminary conservation status of Vulner- able (VU) due ts relatively narrow global distribution (< 20,000 km?) using the IUCN criteria (IUCN, 200 > distribu- tion within the threatened interior Atlantic Forest of to 1). Further work on its specific Brazil may indicate that it may be of further concern. Additional specimens examined. BRAZIL. Paraná: Pal- mas, Hatschbach 15030 (F, NY, US): Hatsch- bach 22707 (F, MO, NY); Mun. Lapa, Segundo Faxinal, trevo para Monte Alegre, Ribas & Silva 190 (BM). Rio Grande do Clevelandia. Sul: Aparades da Serra, Pabst 6281 & Pereira 6454 (US): Sáo Salvador, Sehnem 2203 (NY, US); Sáo Salvador, Montenegro, Sehnem 3934 (V, NY, US); Estrada B. Ouro- Riozinho, Sobral et al. ICN 111138. (NY): Mun. Sao Francisco do Paula-RS 235, Wasum 270 (G). Santa Catarina: Mun. Lajes, Painel, Lajes, Klein 4567 (NY. US): Mun. Urubici, cachoeira do Avencal, 29 Mar. 2001, Hatschbach et al. 72547 ( eh z Mun. Urubici i, ie acas Gordas, entz & Sobral 188 (NY); | se, perto Sta. Cecilia, Pabst 6123 & 2 "reira a (K. US x Mun. in Doce. nalt Annals of the Missouri Botanical Garden Figure rags No. gir 0. Holotype of Solanum pabstii (Pabst 0200 & Pereira Databased for the | PBI Solanum project. Non m Nome vulg Loc. Brasil: Obs. Arvo: Col. Gat t Del 6373 |USp. HOLOTYPE HOLOTYPE of Solanum pata 4L. &»u YL boy (52) : Plololopa 10: 427. 1964 Det. Sandra Knapp BM systematic Studi QAI 2006 Solanum pabstii L.B. Smith & Downs Det. Sandra Knapp BM 2005 HERBARIUM BRADEANUM Ri Brasil o de Janeiro — a ta Catari ni ma pom do x S 6 SeJo2quinms+1050m e 5-lOn,fl.alvas 520045 Pereira 6373 TYPE Fam. Solanaceae Nome cient. Solanum pab sti Smith & Downs avaiido entre La- Data 22.10+61 Data Volume 95, Number 3 2008 Knapp 431 Solanum havanense Species Group undersurface of Fi Detail of domatia from leaf aa pe (from Hatschbach 22707, MO). Scale bar in millimeters. gure 10. Campos a euer 24 km E of Horizonte (Paraná), Smith & Klein 13489 (MO, NY, US); Bom Jardim da Serra, estrada Jardim-S. n Km 29, i n Dutilh 1549 (NY). — ADDITION TO THE SOLANUM LEUCOCARPON SPECIES GROUP Shrubs or small trees; young stems and leaves glabrous or pubescent with simple, uniseriate tri- chomes; sympodial units difoliate and geminate or plurifoliate, often anisophyllous. Leaves simple, elliptie to occasionally ovate, glabrous or variously pubescent with simple, uniseriate trichomes, these occasionally dendritic in some populations. Inflores- cences opposite the leaves or somewhat internodal, simple or branched, glabrous or pubescent like the stems and leaves. Flowers white, fleshy, and relatively large, the corolla lobes planar at anthesis. Fruit green or yellowish green at maturity: fruiting pedicels erect or slightly deflexed, usually enlarged just below the berry, woody; seeds flattened-reniform, the margins incrassate, paler, reddish brown to yellowish tan. Distribution. Species in the group are distributed in secondary habitats in South America, from Panama to Brazil. Mart., Brazil. 625 10. Solanum evonymoides Sendtn. in Flora 24, Beibl. 2(5): 87. 1841. TYPE: Nov. 1839, C. Martius Bahia: “Ilheos.,” a totype, designated here, NY!; isotypes BM!, . G-DC!, K!, MO!, P! [Morton neg. 8189]. Fi igure 11. Shrubs or small trees 1-5 m; stems erect, stout and hollow (shattering in dry specimens), glabrous and shiny, usually dark reddish; bark dark reddish brown; new growth glabrous or sparsely red-papillate; sympodial units difoliate, not geminate. Leaves (7—) 10-25 X adaxially, glabrous or with sparse simple uniseriate 1.5-)3.2-6 em, elliptic, glabrous and shiny trichomes 0.5-1 mm in the vein axils and along the midrib abaxially at the junctions of veins and lamina, the the acuminate; veins margins entire, the apex 9 to 13 yellowish; petioles 1.5-6 cm, glabrous. Inflorescences base attenuate, primary pairs, drying leaf-opposed apparently (but structurally not in axils) branch forks, (1.5-)3—4 cm, 3 to 4 times branched, with 10 to 40 flowers, glabrous, the peduncle absent and the inflorescence branching at the very base, or 0.1 to rarely 1 cm; pedicels 1.5- 2.2 em, ca. 1 mm diam. at the base, ca. 2 mm diam. at the apex, slender, nodding to + erect at anthesis, completely glabrous, shiny, articulating at the base: pedicel scars irregularly spaced 2-6 mm apart, plane with the rachis; buds fusiform with the tip strongly pointed, strongly exserted from the calyx. Flowers calyx tube 1-1.5 mm, broadly deltate, apparently all perfect; conical, the lobes 1.5-2 mm, glabrous but dense tuft of uniseriate trichomes on the corolla 2.1—2.6 cm diam., white, with a swollen lobe tip: sweet-scented (fide Kallunki et al. 706A), lobed ca. 3/4 of the way to the base, the lobes 1.6-2.2 cm. planar at anthesis, densely papillate on the margins on both surfaces, hooked and cucullate at the apex: filaments with the free portion 0.5—1 mm, the tube ca. 0.25 mm, anthers 3.5-5 1.5-2 mm, poricidal at the tips, the pores lengthening to slits glabrous; with age: ovary glabrous, conical, the style 6-9 mm, straight, glabrous, the stigma clavate or occasionally (Sáo Paulo populations) 2-lobed, the surfaces minute- ly papillate. Fruit a globose berry, 2-3 cm diam., green al maturity, the pulp sticky; fruiting pedicels 2.5-3 cm, ca. 3—4 mm diam. at ihe apex, woody and pendent; seeds 4-5 X 3.54. mm, flattened-reniform to + ovoid, the margins white, the 1.5 mm diam. at the base, not markedly thickened, pale yellow or surfaces minutely pitted, sometimes appearing hairy. Chromosome number not known. Distribution. Coastal Atlantic Forest of southeast- ern Brazil, in the states of Bahia, Espírito Santo, Rio de Janeiro, and São Paulo, from 0—550(—1800) m. "(Lima & pimientera brava” (Santos 73). Common names. Brazil. Bahia: “caigara’ Messias Santos 144), “ Annals MES ESTE Garden Discussion. Solanum evonymoides is a distinctive species, with its shiny, glabrous leaves and its open, highly branched inflorescences with large, white flowers. Label data indicate that the flowers are sweetly scented (Kalllunki et al. 706A). The cucullate, strongly hooked petal lobe apices are distinctive in 5. evonymotdes. Overall morphology suggests it is best placed in the 5. leucocarpon Dunal species group, with other species possessing large flowers, rather large open inflorescences, and flattened seeds. Within that group it is most similar to 5. alatirameum Bitter, also of southeastern Brazil, but can be distinguished from that species by its petiolate (rather than sessile) leaves and larger berries. Confusion over the identity of Solanum evonymoides has existed for some time, with specimens being annotated as a variety of taxa. Specimens of S. evonymoides have been annotated as 5. glomuliflorum Sendtn., a species of uncertain identity (most probably ` Athenaea Sendtn., see Knapp, 2002a). The later name S. evonymoides J. Rémy in Gay (1846), a species of coastal southern Chile, is a synonym of S. valdiviense Dunal (Knapp, in prep.). a member of Solanum sect. Holophylla s. str. (see Knapp, 1989) and the Dulcamaroid clade sensu Bohs (2005). Solanum cordioides S. Knapp has been annotated as 5. evonymoides in several herbaria (see Knapp. 2002a), but the two species differ in flower morphology (a smaller corolla 0.6-0.8 cm diam. and a species of Aureliana Sendtn. oi green in 5. cordioides vs. 2.1-2.6 em diam. and white with cucullate, hooked apices in S. evonymoides) and fruit size (1—1.2 cm diam. in S. cordioides vs. 2-3 em in S. eronymoides). Solanum evonymoides has with S. Mart. Sendtn. (ex Cyphomandra Sendtn.), with which it is been confused fec sycocarpum very similar morphologically. Solanum evonymoides differs in its uniformly reddish bark, the pedicels abscising flush with the inflorescence rachis and thus not leaving conspicuous pegs, and in its anthers without an enlarged connective (typical of members of the Cyphomandra clade, sensu Bohs, 1994, 2005; Weese & Bohs, 2007). Populations of S. evonymoides that have smaller, shinier leaves and are altogether smaller in all parts (flowers included) from Sáo Paulo and Rio de Janeiro states in Brazil are here treated as a geographical variant, but may warrant distinction at the species level (Custodio Filho 265; Hoehne 3042; Martinelli et al. 13274, 13360; Martins et al. 12361: Pirani et al. 3028A). Some of these plants with very coriaceous leaves and smaller flowers are from higher elevations (1000—1800 m) and are components of saxicolous vegetation (Martinelli et al. 13274). despite their marked habitat differences from other These plants, populations of 5. evonymoides, share the key charac- teristics of the species: glabrous foliage, difoliate but not geminate sympodia, branched inflorescences with white flowers, petals with cucullate, hooked apices, and green fruits at maturity. Solanum evonymoides is relatively widespread in the Atlantic Forest of coastal southern Brazil and appears to be relatively common where it does occur. However, the restricted nature of its habitat and the threats to this habitat make it necessary to monitor the conservation status of this species. Based on current herbarium specimen data, it can be assigned a status (IUCN, 2001) of Vulnerable (VU) (extent of occur- rence < 20,000 km? and occurring in approximately 10 locations, some of which are fragmented), with monitoring across the whole species range and field assessment of the southern populations both being priorities. The holotype of Solanum evonymoides was de- Berlin, so the NY lectotype (Fig. 11) as it is annotated with the exact collecting locality (Hheus) and date (1839). It has several oper stroyed in sheet is chosen as the destroyed B sheet (F neg. 2817); isolectotypes are widely distributed. flowers and is a good match for Additional specimens examined. BRAZIL. s. loc., Anon BN 625 (GH), Sellow 819 (BN) Bahia: E unapolis-Htamaraju, Almeida 16 (US); Mun. Urucuca, dist. à — Sellow s.n. (BM), Serra Grande. 7.3 km na pena Serra Grande Itacaré, conjunto F {mor NY); loc Blanchet bo (e Blanc ha Bees G 12): Mun. ani cruz Cabrália, Es A ológica de Vai Brito & Da Vinha 117 (F); bos Ina, Riberáo s Caveria, Serra ramal com entrada no ne Ll dar n Jos dae lado S, 6 km do S de entrada, Brito & pei 3986 (NY, US): Mun. ltamaraju, rod. ay eee xeira de F indu 101). k Chapadao, Callejas et al. 1618 (NY); Mun. llhéus, ón 18 Rod. Hhéus-Itacaré, Bom nds et al. 6672 (BM); Mun. Itacaré, Marambaia, ca. 6 km SW de Itacaré na estrada para BR-101 Carvalho et al. 4121 (K, NY); Mun. Una, Faze a 9 km do BR-101. dos Santos & Alves Ubaitaba, lbitapitanga; s Santos 1051 (US); Cruz Cabrália, Est. Ecol. Pau-Brasil, CEPLAC Euponino & Da Vinha 481 T. Mur Santa ( F Jose—Una, £5 [av] ruz C abrália, az. Guiberti, Lima 144 (NY); Mun. Belmonte, Estacáo Exp. Gregorio Bondar, Km 58 da rodavia Belmon vae sbi, Mattos Silva et al. 376 (F); Mun. Una, Km 27 da Rod. Sao Mattos Si et 2 12 298 Uña povoado s Comandatuba, 17 km ao S de Silva et al. (F, MO); Mun. Itamaraju, ca. p et al. 10737 (NY, US); Mun. Sta. Cruz de C -o alrededores da Est. Ecol. Pau-Brasil, ih et al. 108 E NY); Mun. Marau, Faz. a ‘ae 22 km a : de Ubaitaba, Mori 12747 (NY): Re ES ne e d E (US); Mun. Santa Cruz Cabrália, area da Est. Ec V Porto Seguro, H od. B 13067 s Seguro— ore aces Santos "OS, Slow 244 (B); , proximo a faz. Pie dade; e ltamaraju, T o = | ¡an e Y — m Ui UN > Z a -q ci = Mun. . Be Imonte, Baro km E of BR 101 on na, Ne "Thomas el b : m na Biol. Res. ca. 500 m N of Rio Ma Volume 95, Number 3 Knapp 433 2008 Solanum havanense Species Group LECTOTYPE vecro TYPE of Solanum LU run io foun, in Part. Flora Ze Bubl. US): EY. 1841. Det. Sandra Knapp BM 2006 * y S 4s ite PEA Bu Solanum evonymoides Sendt, 2 isotyne Determined by W.G. D'Arcy 1972 Missouri Botanical Garden Systematic Studies in Solanaceae Det. Sandra Knapp BM 2005 | A TERT | FS Databased for the porum n m DOE a I E 0013 dedu V. Sa - E £ E, AI E TRACT <> IMAGED Figure 11. Lectotype of Solanum evonymoides (Martius 625 [NY |). Annals of the Missouri Botanical Garden 11202 (BM). Porto Espírito Santo: Mun. Linhares, Res. Flor. CVRB/BA, Folli 1040 (NY): Bananal, Faz. Santa Angelica, gated dirt rd. to rt. Seguro, from church in Bananal on paved rd. from Bananal to Novo Brasil, Kallunki et al. 7064 (BM); Estrada X-I, Res. Linhares, Docemade. Sucre 8298 (F); Mun. pcs io da Barra, Res. Biol. a AMA). lkm de Pedro Canário (na Avenida Aracruz) e 6 km E do IBAMA, Pirani et s 3028-A (BM, Y) Minas Gerais: s. loc., 1839, Claussen s.n. (F); € 200 (C). Rio de sopolis, Glaziou 8856 (G); Mun. a Grande laxoeira do Campo, Claussen ec Serra dos Orgãos. Santa Maria de i dm na, Par. Est. Desengano, Martinelli 13274 (V), Pedra do Desengano, Martinelli 13360 (F). São Paulo: C a Res. Flor., Custodio Filho 265 (NY); A Hoehne 3042 (US); Cunha-Res.. 12361 (NY) 1919, Martins et al. ui to da Serra, 28 Jan. Est. de Cunha, ADDITION TO THE SOLANUM AMBLOPHYLLUM SPECIES GROUP Shrubs or small trees of high elevations, often growing in windswept edges of páramos: voung stems and leaves glabrous to densely pubescent, the trichomes simple and uniseriate; plurifoliate, or variously difoliate, geminate or not, if geminate, the minor leaves usually similar in size and shape to the major leaves. Leaves simple, elliptic to Inflorescences obovate, often thick and coriaceous. opposite the leaves, or occasionally lateral or appearing terminal, simple or several times branched. Flowers white, fleshy and rather large, the petals planar or campanulate at anthesis. Fruit green or yellowish, hard at maturity; fruiting pedicels deflexed or somewhat erect, woody and occasionally rugose; seeds flattened-reniform, pale tan to reddish brown. Distribution. Species of the group occur at high elevations in the Andes, from Colombia to southern Peru 11. Solanum elvasioides S. Knapp, sp. nov. TYPE: Achiras (Uritu- Universidad Nacional de Loja, 2800 m, 4 03'21"S5, 79°13'60"W, 20 May 2001, J. E. Madsen, A. Byg & C. Chimbo 8054 (holotype, LOJA!; isotypes, AAU! MOL QCA!, QCNE!). Figure 12. Ecuador. Loja: rd. Loja-Las singa), Km 9 from the Species Solanum barbulato Zahlbr. similis, sed caulibus et. foliis. ominino. glabris, sympodiis unifoliatis interdum difoliatis subgeminatis, venis numerosis subparallelis differt. Shrubs to small trees. 2-5 m; stems slender, brittle, gray on older stems; new growth glabrous or sparsely seemingly glabrous and grayish; bark pale papillate, the papillae white or reddish: sympodial units usually unifoliate on reproductive shoots, occasionally difoliate and almost geminate. Leaves 3.5-9 X 0.9-2.5 cm, narrowly elliptic to lanceolate, sympodial units elabrous on both surfaces, the base attenuate, the margins revolute in dry material, the apex acuminate or acute; primary veins 15 to 30 pairs, very closely spaced and subparallel, barely visible adaxially, not at all prominent abaxially, the midrib drying yellow- slightly ish; petioles 0.2-0.9 cm, winged from the decurrent leaf bases. Inflorescences leaf-opposed occasionally internodal, 0.1—1 em, simple, with 2 to 6 f flower owers, glabrous, the peduncle 0-0.2 cm; pedicels in not known, in bud glabrous; pedicel scars closely spaced and overlapping, plane with the rachis: buds (very voung) elliptic. Flowers not known. Fruit a elobose berry, 0.9-1 em diam., green at maturity, o J e J glabrous; fruiting pedicels 1-1.5 cm, ca. 0.5 mm al the base, ca. 15 mm at the apex, deflexed to Seeds 0.3-0.4 2-0.3 mm, flattened-reniform, reddish brown, the margins slight- — — somewhat deflexed. ly paler and thickened, the surfaces minutely pitted. Chromosome number unknown. Distribution. In páramo and shrubby vegetation from 2600-2800 m in the province of Loja, Ecuador. Etymology. Solanum elvasioides is named for its tightly parallel. venation that slightly f Elvasia DC. leaves with resemble those o (Ochnaceae). Discussion. Although Solanum elvasioides is only known from fruiting material, | am describing it here in order that flowering material might be sought. [Ht must be a very rare species, as the numerous collections made in the vicinity of Loja in the last decades by collectors from the University of. Aarhus and the Missouri Botanical Garden have only revealed two specimens (cited here). These collections are about 10 km apart on the western slope of the north- south ridge west of Loja (the continental divide), where the terrain drops steeply to the deep, dry Catamayo valley (D. Neill, Solanum elvasioides is superficially similar to both S. barbulatum Zahlbr. (of the S. amblophyllum Hook. smithii (of the S. both of — pers. comm. jJ. species group) and arenarium occur in the same but the tightly spaced, parallel leaf venation and total glabrousness species group), which general region of southern Ecuador, of S. elvasioides are distinctive. It is likely to belong to the 5. 2002a), flattened-reniform seeds. The 5. amblophyllum species group (sensu Knapp, with other taxa from high elevations with amblophyllum spe- cies group Knapp (2002a) that may be monophyletie (J. C. Granados-Tochoy, is one of the few species groups from pers. comm.). A preliminary conservation assessment of Solanum elvasioides would be Endangered (EN), based on its very small range (two currently known populations 10 km apart, so extent of occurrence < 100 km”: see Volume 95, 2008 Number 3 Knapp Solanum havanense Species Group Solanum ef vas or des S. Det. Sandra Knapp BM 2006 Figure 12. cA =. aa ISOTYPE FLORA OF MA dong Jens Elgaard Madse: with Anja Byg, Carlos Chiñibo 8054 Solanaceae Solanum or the PRI Solanum Project. "HERBA RIU UM 2A Province: LOJA MONA AARHUS i Rd. - : Nor p vs DE 2 AN AAU. ^ Reg At dA burned paramo. (79° 13' 60" W 04° 03. 21" Alt. 2800 m. May 20, 2001 Shrub, 1 m high. Fruits green while unripe Department of Systematic Botany, University of Aarhus, Denmark (AAU) in collaboration with Herbario Re falda Espinosa, Ecuador (LOJA) Isotype of Solanum elvasioides (Madsen et al. 8054 | AAU ]). Annals of the Missouri Botanical Garden IUCN, 2001) and the threat from fire spreading into its montane forest habitat from the drier forests in the heavily populated Catamayo valley (see Kessler [1992] for a southwestern Ecuador). discussion of threats to the forests of Paratype. ECUADOR. Loja: Cerro Villanaco, 2s rd. oja-Pedestal-La Toma. Km E 2670 m. 4 00'I3"5, 19 15'14"W, 10 Apr. 1994, P. M. Jorgensen, C. n & H. Vargas 44 (LOJA, MO, QCA not seen). ADDITIONS TO THE SOLANUM ARBOREUM SPECIES GROUP Shrubs or small trees; young stems and leaves elabrous to densely red-papillose, occasionally pu- bescent with simple, uniseriate trichomes; sympodial units difoliate and geminate, often strongly aniso- phyllous, occasionally unifoliate or plurifoliate. Leaves simple, narrowly elliptic to obovate, usually elabrous on both surfaces. Inflorescences opposite the occasionally internodal, simple. few to leaves or many-flowered. Flowers white, small, and with the lobes deflexed at anthesis or larger and fleshy with the lobes planar at anthesis. Fruit usually green and hard at maturity; fruiting pedicels usually erect and thickening at maturity. occasionally deflexed; seeds ovoid-reniform, pale tan to dark brown, often rather large. =] Distribution. Species of the group occur i primary forests from Central America to Bolivia, at a wide range of elevations. 12. Solanum humboldtianum Granados-Tochoy & S. Knapp, Syst. Bot. 32: 202. 2007. TYPE: Colombia. Cundinamarca: Mun. Bojacá, vereda San Antonio, finca El Triunfo, saliendo hacia la carretera que conduce de Mosquera a Tena 2000-2700 m, 29 Jan. 2005, J. C. Tochoy & W. Meier 860 (holotype, isotypes, BM!, FMB!, HUA!) Granados- COLI Small branched; stems scandent, 10 m, 4-angled, diffusely drying pale brownish or trees to more yellowish when young; new growth with the vegetative buds minutely pubescent with simple, glandular 2- celled trichomes, reddish in dry material, soon deciduous; bark of older stems exfoliating; sympodial units Ac or difoliate, not geminate. Leaves 5— 25 3-12 em, elliptic to obovate, chartaceous to S Rm ncs glabrous, markedly discolorous, adaxi- ally bright green when fresh, becoming olive green or pale tan in dry material, somewhat shiny, abaxially vellowish green when fresh, becoming golden in dry material; base acute, slightly asymmetric; margins entire, revolute; apex acute to acuminate or emargin- ale, somewhat asymmetric; primary veins 7 to 14 pairs, with the midrib impressed adaxially, keeled abaxially, drying reddish green; petiole 1-4 cm, channeled adaxially, corrugated transversely and somewhat corky, glabrous. Inflorescences leaf-op- posed or internodal, 1-11 em, usually simple, occa- sionally bifurcate, with 6 to 50 flowers, the peduncle 0.5-5 em, ca. 3 mm diam. at the base, stout; pedicels 0.5-1 em, ca. 1.5 mm diam. at the base. ca. 1 mm diam. at the apex, fleshy, nodding at anthesis, elabrous, articulated at the base: pedicel scars closely spaced 2 distinct rows, the rows and individual scars overlapping: buds globose, becoming ellipsoid, sparsely pubescent with simple trichomes to 0.5 mm, Flowers eh about halfway exserted from the calyx tube. — neterostylous; calyx tube 2-3 mm, campanulate, the lobes 1-1.5 mm, broadly deltate to truncate, pubes- cent with tufts of simple, uniseriate trichomes c: 0.5 mm at the apices, sometimes bifid and irregular: 2/3 of the way to the base, the lobes 0.7-1.1 em, adaxially corolla 1-2 cm diam., white, fleshy, lobed ca. glabrous, abaxially densely pubescent with minute uniseriate glandular trichomes, these reddish when dry, the petal midvein keeled adaxially, the lobes cucullate at the tips; filaments with the free portion ca. 0.1 mm, the tube less than 0.2 mm, sometimes with tiny apiculae ca. 0.5 mm on the margin between the filaments, glabrous; anthers 3—4 X 1.1-5 mm, poricidal at the tips, the pores lengthening to teardrop-shaped slits with age; ovary 1-2 mm, sub- pyramidal, glabrous, the apex somewhat tuberculate: short styles 2-3 mm, long slightly 2-lobed, Fruit a globose to style straight, glabrous, styles 4-5 mm, the stigma capitate, the surface minutely papillate. 1.5-2 X subelobose berry. .2-1.6 em, green and 8 ] 8 hard maturity, the pericarp coriaceous: fruiting pedicels 1.5-3 cm, distally enlarged, woody and deflexed, the calyx lobes in fruit expanding to twice their size in flower, tightly investing the base of the fruit, 34 X 2-3 mm, creamy yellow in fresh material, berry; seeds ca. 30 per flattened-reniform. yellowish tan when dry, the margins incrassate, the surfaces minutely pitted. Chromosome number not known. Distribution. Endemic to the western Andean slopes of Colombia in the departments of Cundina- marca and Santander, in premontane cloud forests from 2400-2700 m. Discussion. Solanum E is very sim- ilar morphologically to 5. goniocaulon S. Knapp (of the S. arboreum species group) from iem Ecuador and adjacent northern Peru and to 5. laurifrons Bitter of the S. Cordillera amblophyllum species group) from the Occidental of Colombia (see Knapp, 2002a). The three species share papillate new growth, Volume 95, Number 3 2008 Knapp 437 Solanum havanense Species Group angular fleshy stems, markedly discolorous leaves that dry yellowish abaxially, and cucullate petal apices. Solanum humboldtianum is also scandent or somewhat lianescent shrub, a habit not recorded in other similar species. Solanum goniocaulon is a member of the S. arboreum species group (sensu Knapp, 2002a), all with ovoid seeds. Within the S. arboreum species group, it may be that S. humbold- tianum is closely related, as well as morphologically similar, to S. goniocaulon. Solanum humboldtianum has been assessed as Endangered (EN) using IUCN Red List criteria (IUCN, 2001), based on its extent of occurrence of less than 5000 km", 100 km?, its locations) and small overall population size, and the area of occupancy less than severely fragmented (four current unprotected nature of most of the populations near a mim population center (Granados-Tochoy et al., 07). Solanum humboldtianum 1s being propagated in the Jardín Botánico de Bogotá José Celestino Mutis, with plants crossing easily and producing ample seed (J. C. Granados-Tochoy, unpubl.) making ex situ conservation a possibility. Additional specimens examined. COLOMBIA. Cundina- rca: Mun. Bojacá, vereda San Antonio, Granados- Tochoy et ja 389 (CO D. Granados-Tochoy & Giraldo-Cafias 844 vee near Sibaté, San Miguel, Hawkes pe García-Barriga 36 (K); Santa Fé de Bogotá, 1805, Humboldt & ue s.n. (P); Mun. Granada, hac. “El Soc > Idrobo 7 (COL, FMB); Mun. Bojacá, jan yos ie M á próximo a la carretera Mosquera-Tena, Lozano-C. & Torres- — vereda S R. 98 (COL); Mun. Cabrera, vereda Núñez, Morales et al. 497 (COL). Santander: Mun. El Encino, vereda Santa Helena, predio La Sierra, margen izquierdo aguas abajo del Río La Rusia, Santuario de Flora y Fauna Guanetá, Alto Río Fonce, Cadena et al. 78 (COL, UIS). 13. Solanum sagittantherum Granados-Tochoy & Orozco, Caldasia 28(1): 2. 2006. TYPE: Colombia. Cundinamarca: Mun. Sibaté, sector El Soche, vía Bogotá-Silvania, ca. 2900 m, 31 May 2005, J. C. Granados-Tochoy, D. Giraldo-Cañas & D. Canal 888 (holotype, COL!; isotypes, FMB not seen, HUA not seen). to 1.5 m, rhizomatous; slightly angular, minutely papillose with trichomes Subshrub young stems less than 0.05 mm, sparsely pubescent with simple uniseriate trichomes to 0.5 mm, glabrescent; new growth papillose, puberulent, the trichomes to 0.5 mm, leaf margins in bud ciliate distally; sympo- dial units difoliate, geminate, anisophyllous. Leaves elliptic to ovate, occasionally obovate, chartaceous, larger on vegetative shoots, adaxially dark green in fresh material, drying olive green to brownish, abaxially yellowish, drying golden-green, glabrous on both surfaces, abaxially minutely puberulent on the veins, the trichomes simple, uniseriate, to 0.5 mm, the veins pale green in fresh material, drying brownish green; major leaves 3-21 X 1-8 cm, the base acute to cuneate, occasionally oblique, the margins entire to slightly sinuate, revolute when dry, the apex acute to acuminate, inconspicuously ciliate at the tip; minor leaves differing from the major leaves only in size, 0.5-3 X 0.3-2 em; impressed to slightly prominent adaxially, prominent primary veins 4 to 12 pairs, abaxially; petioles 0.2-3 em, glabrous, channeled adaxially, with a few transverse stripes near the base. Inflorescences opposite the leaves, 0.5—0.9 cm, sim- ple. with to 8 flowers, papillose and sparsely pubescent like the young stems, the peduncle 0.05— 0.5 em; pedicels at anthesis 0.5—1 cm, weakly to strongly reflexed, somewhat enlarged at the apex, articulated at the base; pedicel scars closely packed, somewhat separated or overlapping; buds globose when young, later ellipsoid, one calyx lobe promi- nently veined and larger than the rest, the corolla strongly exserted from the calyx tube. Flowers tetramerous, isostylous; calyx white or greenish white in fresh material, cream in dry material, the tube ca. l mm, broadly conic, glabrous or pubescent on external surface with simple uniseriate trichomes ca. 0.5 mm, on internal surface minutely glandular- papillose, the papillae less than 0.1 mm, the lobes ]-1.5 X —3 mm, broadly deltate, glabrous or sparsely pubescent with simple uniseriate trichomes and glandular papillae like the tube, the apex of one of the lobes often enlarged to a fleshy projection ca. 0.2 mm, an extension of the midvein; corolla 1.5— 2 cm diam., fleshy, white, stellate, lobed 2/3 of the way to the base, the lobes 0.5-0.9 cm, elliptic, planar at anthesis, cucullate at the tips, the tips and margins minutely papillate, the papillae less than 0.1 mm; filaments with the free portion ca. 0.1 mm, the tube 1.5-2 mm, glabrous; anthers 4—5 X sagittate at the base, poricidal at the tips, the pores l mm, lengthening to slits with age; ovary ca. 1 X 1 mm, conical, glabrous; style 6-8 mm, straight or sigmoi- dally curved, white in live plants, the stigma ca. 1 mm diam., capitate, the surface minutely papillate. Fruit a globose berry, 1—1.5 X 1-1.5 cm, green at maturity, drying brownish green; fruiting pedicels 1.5-2 cm, strongly deflexed beneath the foliage, distally en- larged, the calyx lobes in fruit ca. 2 times the size at anthesis, tightly appressed to berry in fresh material, reflexed in dry material; seeds ca. 2-3 mm, flattened- reniform, pale green in fresh material, reddish brown the testa minutely pitted. Chromosome number not known. in dry material, the margins enlarged, Distribution. Only known from the type locality on the western slopes of the Cordillera Oriental of the 438 Annals of the Missouri Botanical Garden Colombian Andes, in pristine cloud forest from 2700— 2900 m. Discussion. Solanum sagittantherum is unusual in the Geminata clade in possessing tetramerous flowers, but other species originally described with tetramer- ous flowers have later been shown to have populations with pentamerous flowers as well (e.g., 5. cruciferum Bitter, see Knapp, 20022). Should more populations of S. sagittantherum be found, they may indeed prove to be pentamerous. Granados-Tochoy and Orozco (2006) were uncertain of the affinities of 5. sagittantherum, as it combines many of the characteristics of several of the pragmatic species groups delimited by Knapp (2002a). fruiting pedicel make it similar to members of the S. lis flattened seeds and apically enlarged leucocarpon group, while the rhizomatous habit is shared with both S. arboreum and S. robustifrons Bitter (of the S. arboreum and S. robustifrons species groups, respectively). The rhizomatous habit is unusual in the be related Geminata clade and may to the humid nature of the habitat in which S. sagittantherum occurs. The sagittate anther bases for which the species is named are also very unusual in the amblo- from 5. sagittantherum in distribution, in having plurifoliate, Geminata clade, but are also present in 5. phyllum of central Peru, which differs ralher than difoliate and geminate, sympodial units, and in its complex branched inflorescences, and in 5. chalmersii of Bolivia, which is densely pubescent. have tentatively placed S. sagittantherum in the S. arboreum species group based on its overall morphol- ogy, but molecular results may clarify its status. Solanum sagittantherum is known only from the type locality (four collections were cited in Granados-Tochoy and Orozco [2006], all from El Soche, but no population details were reported; I have not seen this material) and needs further investigation before a confident conser- valion status can be assigned. Given the fragmented nature of the cloud forest habitat in the Colombian Department of Cundinamarca (Granados-Tochoy et al., 2000), it can be given a preliminary assessment (IUCN, 2001) of — IS likely to be of conservation concern, but it Critically Endangered (CR) due to its very restricted distribution (< 100 km’) in a single locality. ADDITION. TO THE SOLANUM NIGRICANS SPECIES GROUP Shrubs or usually pubescent with usually arachnoid, but occa- small trees; young stems and leaves — sionally dendritic or simple trichomes; sympodia units unifoliate, difoliate or trifoliate, or difoliate and usually geminate. Leaves simple, elliptic to ovate, pubescent abaxially, occasionally glabrous, the tri- chomes like those of the young stems. Inflorescences opposite the leaves, 3- to 20-flowered. Flowers white, thick and fleshy, the corolla lobes planar at anthesis; ovary glabrous or pubescent. Fruit green and hard at maturity, glabrous or pubescent; fruiting pedicels erecl or somewhat deflexed, woody; seeds ovoid- reniform, usually dark reddish brown. Distribution. Species of the group are usually found in forests, in montane forests from Mexico to Honduras and the Andes, and in Araucaria and coastal forests in southeastern. Brazil. 14. Solanum canoasense L. B. Sm. € Downs, Fl. lustr. Catarin., Pt l Solanac. 104. 19660. Solanum cataractae l. B. Sm. € Downs, Phytologia 10: 427. 1964. TYPE: Brazil. Santa Catarina: Mun. Bom Retiro, Rio Canoas, Cm dos Padres, 1300-1400 m, 22 Nov. 1956, L. B. Smith & R. Klein 7843 (holotype, US 00026997); Isoly pes, HBR not seen, R not seen). Small elabrescent; 5-2 m; reddish: shrubs 0. stems slender, bark pubescent with uniseriate dendritic and/or arachnoid erect, new growth densely trichomes, these soon deciduous; sympodial units 3-9 1-2.1 cm, oblanceolate to narrowly lanceolate or oblanceolate, Caves o plurifoliate. lanceolate ti widest in the upper 1/3, both surfaces glabrous, the leaves crowded at branch apices, the base attenuate, the margins revolute, the apex acute; primary veins 7 to 9 pairs, drying paler than the lamina; petioles 0.4— 0.6 cm, sparsely dendritic pubescent adaxially. Inflo- rescences internodal, appearing terminal, 0.6-1 cm, simple, with 2 to 4 flowers, densely pubescent with trichomes, the 1.3-1.6 em, ca. at the apex, stout, + nodding at anthesis, sparsely branched arachnoid or dendritic 0.2-0.5 cm; l mm diam. uniseriate peduncle pedicels at the base, ca. 1.5 mm diam. pubescent like the inflorescence, articulated at the base; pedicel scars 1-2 mm apart; buds ellipsoid, the corolla strongly exserted from the calyx tube. Flowers all perfect (2); calyx tube (1-)2-2.5 mm, the lobes 2—4 mm, broadly conical, broadly and irregularly deltate, thick and fleshy, the margins thickened and paler when dry, sparsely pubescent with arachnoid or dendritic trichomes especially near the tips; corolla white, lobed ca. 2/3 of the way to the base, the lobes 1-2 cm, broadly deltate, + planar at 2—2.2 cm diam.. anthesis, sparsely lo densely pubescent abaxially especially in the sinuses and distally with branched uniseriate or papillate trichomes, these especially dense at the tips; filaments with the free portion 1— 5 mm, the tube ca. | mm, glabrous; anthers 5—6 X 1.5-2 mm, poricidal at the tips, the pores lengthening to slits with age; ovary conical, glabrous, the style 1— glabrous, the stigma clavate, the L.l em, straight, Volume 95, Number 3 2008 Knapp 439 Solanum havanense Species Group surface minutely papillate. Fruit not seen; seeds not seen. Chromosome number not known. Distribution. In Araucaria forests, cloud forests, and forest edges in the Bre Catarina and Paraná, from 1300-1400 m. azilian states of Santa Common name. “Canema mirim” (Smith & Downs, 1966) Discussion. Solanum | canoasense is distinctive among southeastern Brazilian members of the Gemi- nata clade in its large flowers and floccose, but soon glabrescent, new growth. The species is known from very few collections, and it is apparently very rare. Mentz and Oliveira (2004) put 5. synonymy with 5. cassioides L. B. 5m. & Downs, from canoasense | — 1 which it differs in not possessing the distinctive broad-based trichomes found in S. cassioides and its relative 5. trachytrichium Bitter, and in its plurifoli- ate, rather than difoliate, sympodial units. Solanum canoasense is a member of the S. nigricans M. Martens & Galeotti species group (sensu Knapp, 2002a) based on the possession of floccose trichomes on the new growth, but seed morphology is not known. The type of S. canoasense is nearly glabrous, like 5. bahianum S. Knapp of coastal Bahia, Brazil. Solanum canoasense differs from S. bahianum in its plurifoliate rather than unifoliate sympodia, and in its larger flowers that are somewhat campanulate at anthesis. The two species occupy very different habitats, S. bahianum the coastal restingas and 5. canoasense upland Araucaria forests. No fruiting specimens are known, and collections of fruiting material are a priority. canoasense are Conservation assessments of S. difficult, due to the few specimens available. The small area (few estricted distribution of the species in threatened habitat (Araucaria forests) suggest it is of concern and could be assigned a preliminary conservation status (IUCN, 2001) of Endangered (EN). occupancy collections) and Additional Speers examined. BRAZIL. a Mun. Pu Serra do Aracatuba, Kummrow 2410 (NY). Santa na: Mun. estrada Serra do Corvo Branco, Catari Urubici, Peis 1756 (NY). ADDITIONS TO THE SOLANUM ARENARIUM SPECIES GROUP Small shrubs; young stems and leaves sparsely to densely pubescent with dendritic trichomes, these often very densely branched; sympodial units unifo- liate or more commonly difoliate and geminate. Leaves simple, lanceolate to elliptic, occasionally shiny adaxially, pubescent abaxially with dendritic trichomes, these confined to the vein axils or densely covering the surface, the apex and base various. Inflorescences opposite the leaves or occasionally internodal, simple, densely pubescent like the stems pedicel scars and leaves, the trichomes dendritic; closely spaced or unevenly spaced and ca. 5 mm apart. Flowers white, usually somewhat fleshy, the corolla lobes planar at anthesis. Fruit green o greenish yellow and hard maturity, the pericarp occasionally thin and brittle in dry specimens, glabrous or densely pubescent with dendritic tri- chomes; fruiting pedicels woody, deflexed; seeds very large, ovoid-reniform, pale tan or yellow. Distribution. Species in the group are distributed in montane forests in the Andes in Ecuador, Peru, and Bolivia, and in a variety of habitats in southeastern Brazil. 15. Solanum compressum |. B. Sm. € Downs, Phytologia 10: 430. 1964. TYPE: Brazil. Santa Catarina: Sáo Joaquim, Fazenda de Laranja. Bom Jardim, 1400 m, 13 Dec. 1958, P. Reitz & R. Klein 7867 (holotype, US 000270011; isotypes, HBR not seen, NY!). Figure 13. to 15 cm DBH; 1 Loose, Trees to small treelets 5-10 m, stems. spreading, sparsely pubescent with golden uniseriate dendritic trichomes 0.5-1 mm, soon glabrescent; bark dark reddish brown; new growth densely golden dendritic pubescent, the trichomes ca. ] mm; sympodial units plurifoliate. Leaves 4—15 1.6-5 em, elliptic to narrowly elliptic, glabrous and slightly shiny adaxially, occasionally with a few scattered uniseriate simple trichomes on the lamina, abaxially very variably pubescent, from a few denditic trichomes 0.5-1 mm in the vein axils to sparsely pubescent over the entire lamina with dendritic trichomes 0.5—1 mm, the trichomes with a uniseriate stalk and very short branches, always denser in the vein axils, the base acute to attenuate, the margins entire, the apex acute; primary veins 9 to 10 pairs, the midrib keeled adaxially; petioles 1-2.5 em, glabrous or sparsely dendritic pubescent. Inflorescences ter- minal, much longer than the new growth, 4—9 cm, 3 to A(to 6) times branched, with 30 to 40 flowers, glabrous or sparsely pubescent with loose dendritic trichomes like those of the stems, the peduncle 2-6 em; pedicels 1.4—1.6 cm, diam. at the apex, filiform, + deflexed at anthesis, ca. 0.5 mm diam. at the base, ca. | mm glabrous to sparsely dendritic pubescent; pedicel scars irregularly spaced 0.5-1 mm apart, plane with the rachis; buds ellipsoid, pointed at the tips, strongly exserted from the calyx tube. Flowers apparently all mm, broadly conical, the perfect; calyx tube 1-1.5 lobes 1.5-2.5 mm, deltate, densely pubescent with dendritic trichomes ca. 0.5 corolla 1.5-2 cm diam., white, lobed ca. 3/4 of the way to the mm on the tips; 440 Annals of the Missouri Botanical Garden ISO TYPE of Solanum Lom p EG m LB. dati, + Tame 1 | "Phdoleya. 10: 420. 1944. | Det. Sandra Knapp BM — 2006 RT E CUTS TT [ | Solanum compressum L.B, Smith & Downs | Det. Sandra Knapp BM — 2005 | = 7 Plantas de SANTA CATARINA - BRASIL Família Solanaceae N. cientifico Solanum. compressum Smith & Downs Sin - Var. IBotypus.n4.Ap Nome vulgar Localidade | F&zonda..da..Laranjs,..Bom Jardim,..3,. Joaquim... Habitat Pinhal Altit. 1400... m Habit A e e Seas e In EA SZ Lon Flor (côr, odor, etc.)..... c8; estames amarelos Fruto (tamanho, odor, cór, etc.) € olecionador 7...Data_13.12.1958...... Determinador Le. Ba. Smi. Date . 1964 . Tu. <> IMAGED Observacóes (usos económ., abundár GARDEN CO 00139102 Figure 13. Isotype of Solanum compressum (Reitz & Klein 7867 [NY |). Volume 95, Number 3 2008 Knapp 441 Solanum havanense Species Group base, the lobes 1-1.5 cm, deltate, planar at anthesis, minutely dendritic pubescent on the tips and margins; filaments with the free portion ca. 0.5 mm, the tube ca. 0.5 mm or less, glabrous; anthers 3.5-4 X 1.5- 2 mm, poridical at the tips, the pores lengthening to slits with age, the abaxial surface papillate and rough: ovary conical, glabrous, the style 6—7 mm, straight. glabrous, the stigma clavate, 2-lobed, the surface minutely papillate. Fruit a globose berry, 1-1.5 em diam., pale green at maturity, the pericarp woody .o mm € 2.5-3 cm, when dry; fruiting pedicels ca. at the pendent, the calyx lobes enlarging in fruit to 3—4 mm; diam. at the base, ca. 3 mm diam. apex, seeds ca. 2 X 1.5 mm (immature?), ovoid-reniform, pale yellow, the surfaces minutely pitted, the margins not incrassate. Chromosome number not known. Distribution. In Brazil (Rio Grande do Sul, Catarina, and Paraná) and adjacent Argentina and Santa Paraguay, in open forests on the planalto, interior Atlantic Forest, 750- 1400 m. and Araucaria forests, from Common name. 1966) “Canema mirim” (Smith & Downs, Discussion. Solanum compressum is somewhat similar to 5. pabstii, also of southeastern Brazil, but the dendritic pubescence, larger flowers, and more robust inflorescences easily distinguish it from that species. The trichomes and ovoid seeds of S. ‘the S. arenarium species group (sensu Knapp, 2002a) and are partic- ularly like those of S. gnaphalocarpon Vell. from compressum place it as a member of Brazil. Solanum gnaphalocarpon has difoliate, gemi- nate sympodia that are anisophyllous, a simple inflorescence, and a distinctive densely long-pubes- cent fruit. The seed morphology of S. compressum needs further examination, as those I have examined are from what may be immature fruits. Solanum compressum is very variable in pubes- cence density throughout its range, with some plants being nearly glabrous and others densely pubescent on all parts. The trichomes are always dendritic, with a long uniseriate stalk and very short branches. The abaxial surface of the anthers of S. compressum is papillate in herbarium specimens and somewhat reminiscent of the anther margins of wild tomatoes (Carrizo García, 2003; Peralta et al., 2008). Whether this is merely an artifact of drying is not known, but it occurs in all flowering specimens I have examined. Solanum compressum is relatively widespread in the Atlantic Interior Forest of Brazil, Argentina, and Paraguay. It grows often in disturbed habitats at the edges of forests and roads. It can be assigned preliminary conservation status of Near Threatened (NT which it occurs (Atlantic — (IUCN, 2001), due to the fact that the habitat in Selva Paranaénse) is highly threatened; monitoring will be Interior Forest or necessary. ARGENTINA. Mis- Ruta Nac. 101, 8 km San Antonio, Salto Andrecito, 8 (MO [2]; Guaraní, Predio Guaraní, junto s, Tressens et al. 5910 (F US). . Guarapuava, Serra da Esperanga, Himalbank 7384 (F); Mun. Laranjeiras do Sul, Rio das Pedras, Hatschbach & Pereira 13071 (US): Vittorino, Hatschbach 22679 (BM, F. W); Mun. Gal. Carneiro, da Galinha, Hatschbach 30713 (BM, NY); Ttaperussú, Jonsson 10054 (F); Fazenda Reserva, ca. 60 km SW of Guarapuava, Lindeman & de Haas 4915 (K [2]. NY). Grande do Sul: Tres Arroios, S. Almeida, ES 153, Babe et al. 7468 (US); Mun. Sáo Marcos, Linha Feijó, Kegler 286 (NY, US); Mun. a ^u do Sul, Forqueta, Kerr 1301 (US); Mun. Caixas do . Vila Oliva, Kegler 5 (US) Sáo Francisco de | Mu no paradeiro da ix para Taquara, Mentz & Sobral 215 (NY); sentre o rio Tainhas e Prince 'esa do Campo, Pabst 6693 & Pereira (F, NY Zaixas, sas D x JE rid 52193 (US [2]; $ sato et al. bs (C, . Dois Ermáos, Sehen xd 505 (US); Linha cea] Nova Petrópolis, Sobral et al. 1 (F); Par. de Espigão Alto, Barrancáo, Stehm 728 m. i Barra do Ouro, estra al. 1524 (NY Additional P eui examined. iones: Dpto. Gral. Manuel Belgrano, de Bernardo hi Pos hacia Morrone et al. 138 Passo =< = = [s] ^ © i a Kappesberg, p. P nann da para Riozinho, Stehmann et Cambará, Stehmann ICN 111359 (NY): Faxinal. Cambará do Sul, Stehmann & Sobral ICN 111362 (NY); Mun. Sáo Francisco de 1174 (G, US), Botánico, W la: id Lar: Paula, estrada p Wasum m : US); ara Taquara, Caixas do Sul, 14584 (US): Rio I ts São nia Reitz & p AA (US); Potreiro Grande, Mafra, Reitz & Klein 2 (F). PARAGUAY. Itapua: Pirapó, Centro de Desarrollo hs Pérez 173 (MO) 16. Solanum pseudodaphnopsis L. A. Mentz & Stehmann, Novon 13: 97. 2003. TYPE: Brazil. Paraná: Paranaguá, Matinhos, 20 Sep. 1946, G. Hatschbach 382 (holotype, MBM not seen; isotypes, BM!, PACA not seen, US 00810597?, US 00810597!, W!). — Shrubs to 1.7 m; branches woody, cylindrical, dark green to almost dark brown, glabrate, apical branches light, tinged in green or gold; new growth sparsely to gut, B 8 8 8 I à [om densely covered with dendritic-echinoid trichomes; sympodial units unifoliate, less commonly difoliate and geminate. Leaves 5-12.5 X 1.5-5.5 cm, obovate, coriaceous and shiny, drying olive green to cinereous, glabrous or with a few dendritic-echinoid trichomes adaxially along the veins, glabrous or with a few dendritic-echinoid trichomes along the midrib abaxi- ally, the base attenuate, the margins slightly revolute in dry material, the apex obtuse, rounded or slightly acute: primary veins 4 to 7 pairs, drying yellow abaxially; petiole 0.3-0.5 em, sparsely pubescent 442 Annals of the Missouri Botanical Garden with dendritic-echinoid trichomes. Inflorescences — cassioides of the S. nudum species group, but differs opposite tne leaves or occasionally internodal, simple or rarely branched, 1—4 em, 12- to 19-flowered, the peduncle to 2 em, entire inflorescence pubescent with dendritic-echinoid trichomes; pedicels at anthesis slender, io 2 em, sparsely pubescent with dendritic- echinoid trichomes, articulated at the base; pedice scars ca. | mm apart, not overlapping, beginning 1/3 to 1/2 of the way from the base of the inflorescence; buds ellipsoid, the corolla soon exserted from the calyx tube. Flowers apparently all perfect: calyx tube conical, ca. 2 mm long, the lobes ca. 2 mm, ovale, sparsely dendritic-echinoid pubescent adaxially, gla- brous abaxially; corolla 1.5-2 cm diam., white, lobed ca. 3/4 of the way to the base, the lobes reflexed at anthesis, glabrous or with scattered dendritic-echinoid trichomes adaxially, the tips of the lobes with a few dendritic-echinoid trichomes, otherwise papillate; anthers 2-3 X ca. 0.5 mm, poricidal at the tips, the pores teardrop-shaped; free portion of the filaments less than 0.1 mm long, the filament tube ca. 0.5 mm; ovary glabrous; style to 7 mm; stigma a thickened area at the tip of the style, minutely papillose. Fruit a globose, giabrous berry, to 1 cm diam., green; fruiting pedicels woody, somewhat deflexed, to 2.5 cm long, ‘a. | mm diam. at the base, ca. 2.5 mm diam. at the apex: seeds ca. 3 X 2.5 mm, ovoid-reniform, brown, the surfaces minutely pitted, nearly smooth. Distribution. Restricted to the coastal sandy habitats of restinga vegetation in Santa Catarina and — ¡a arana in southern Brazil; growing in open places and in gaps along the borders of coastal rain or swamp forest at about sea level. Common names. Brazil. Paraná: “canema-mirim, caneminha, canema do brejo” (Hatschbach 382). Discussion. Solanum pseudodaphnopsis is easily recognized by its obovate, leathery, glabrous, or very sparsely pubescent leaves with delicate, dendritic- echinoid trichomes. The leaves dry a distinctive olive- green color. The trichome branches of 5. pseudodaph- NOPSLS are not as congested as the true echinoid in members of the Brevantherum trichomes found clade (e.g., S. erianthum D. Don or S. rugosum Dunal) or in the 5. nitidum Ruiz & Pav. species group of the Dulcamaroid clade (e.g., 5. storkit C. V. Morton & Standl.) but the branches are more tightly packed along the axis than other related species in the Geminata clade (sensu Bohs, 2005). Solanum pseu- dodaphnopsis is a member of the 5. arenarium species 5: arenarium, differing in its larger, obovate leaves and group and within that group is most similar tc much sparser, somewhat echinoid trichomes. Solanum pseudodaphnopsis also is morphologically similar to 5 from that species in its leaf shape, ovoid-reniform seeds, and simple inflorescences; within the 5. nudum species group, S. cassioides is related to S. trachy- trichium. Both species possess peculiar thick-based i and S. pseudodaphnopsis. Solanum arenarium is a species trichomes not present in arenarium of upland forests in southeastern Brazil, mainly on granitic hills, at elevations to ca. 300 m, while 5. pseudodaphnopsis is restricted to the restinga habitat at sea level. Solanum pseudodaphnopsis occurs in a very re- stricted and threatened habitat type, the coastal restinga of Brazil (see Mentz «€ Stehmann, 2003). It is known from very few (10 cited in the original description) collections along a narrow, ca. 100- — 50 km long, coastal strip of critically endangered habitat (Falkenberg, 1999). Based on these observa- lions, it can be assigned a preliminary conservation status of Endangered (EN) using IUCN criteria (IUCN, 2001), but may indeed turn out to be Critically Endangered (CR) when detailed population data become available. Additional specimens examined. BRAZIL. Parana: Guaratuba, Rio da 10198 (US) Mun. Paranaguá, Rio da Vila, Hatschbach 43175 (F). Praia. Hatschbach ADDITION TO THE SOLANUM DOLOSUM SPECIES GROUP Spindly shrubs or hemiepiphytes; young stems and leaves variously pubescent, strongly zigzag: sympodial units unifoliate. Leaves simple, linear to ovate, glabrous to variously pubescent, the apex long-acuminate, the margins of the apex with minute ciliate trichomes. Inflorescences opposite the leaves or internodal, strongly zigzag. Flowers white, usually less than 1 em diam., the lobes reflexed at anthesis. Fruit a globose green berry; fruiting pedicels deflexed; seeds ovoid-reniform. Distribution. Species of the group occur in montane Andean cloud forests from Colombia to Bolivia. 17. Solanum naucinum 5. Knapp, Ann. Missouri Bot. Gard. 92: 251. 2005. TYPE: Peru. Pasco: Prov. Oxapampa, 4-5 km N of Mallampampa, 2400 m, 10 02'S, 75 45' W, 22 Jan. 1984, Smith & Canne 5793 (holotype, USM!; isotype, MO! — Small shrub i densely golden pubescent, the trichomes 0.5-1 mm, | 50 em; young stems and leaves dendritic, uniseriate with single-celled branches; older stems glabrescent, thin and flexuous; bark of the older stems pale brown; sympodial units unifoli- ate, occasionally difoliate and geminate on non- Volume 95, Number 3 Knapp 443 2008 Solanum havanense Species Group reproductive nodes. Leaves 7-15 X 1.5-3.2 cm, and our knowledge is limited due to lack of narrowly elliptic to narrowly oblanceolate or lanceo- — collections. It has been given the status of Data late, glabrous adaxially, pubescent abaxially on the veins and lamina with golden uniseriate dendritic trichomes to 1.5 mm, the trichome branches unicel- lular, pubescence denser along the veins, the apex long-acuminate, blunt at the extreme tip, the margins entire, slightly revolute, the base acute, sometimes somewhat oblique; primary veins 8 to 11 pairs, these looping to form a distinct marginal vein; petiole 0.3— 0.6 cm, reproductive nodes smaller, but similar in shape to the densely pubescent; minor leaves on non- major leaves. Inflorescences opposite the leaves, simple, 5—8 cm, 8- to 10-flowered, densely pubescent with golden dendritic trichomes to 1 mm; pedicels at anthesis not seen, articulated at the base; pedice scars spaced 2—4 mm apart, beginning ca. | em from base of inflorescence; buds not seen. Flowers white (fide Smith & Canne 5793), not seen. Fruit a globose, diam., drying green (immature?) berry, ca. l cm mustard yellow on herbarium specimens; fruiting pedicels 1-1.2 cm, pendent, ca. 1 mm diam. at the base, ca. 2 mm diam. at the apex; calyx lobes in fruit ca. 3 mm, deltate, pubescent with golden dendritic 2 mm, not mature in fruit slightly trichomes; seeds ca. 3 examined, ovoid-reniform to flattened, the surfaces minutely pitted, the margins not incrassate. Distribution. Only known from the type specimen collected in the montane forests of the Cordillera de Yanachaga (site of Parque Nacional Yanachaga- Chemillén) in the Department of Pasco, Peru, at about 2400 m. Discussion. Solanum naucinum is a member of the Solanum dolosum C. V. Morton ex S. Knapp species group (sensu Knapp, 2002a), with unifoliate sympodia a on reproductive nodes, filamentous inflorescences, and long-acuminate leaves with ciliate margins near the apex. It differs from all other species in the group in its distinctive golden dendritic pubescence. Like the other taxa in the 5. dolosum species group, 5. naucinum 1s a spindly shrub, possibly also growing as a hemiepiphyte in its cloud forest habitat. The seeds in the fruits on the type specimen are immature, but without incrassate appear to be ovoid-reniform margins; however, mature fruits and flowers are needed to complete the morphological description of this rare species. Solanum naucinum is apparently an extremely rare plant; no new collections of this rare species have come to light since the collection of the type specimen in 1984. Solanum naucinum shares an elongate inflorescence with widely spaced pedicel scars with S. gonyrhachis S. Knapp of Bolivia. uncertain conservation status (IUCN, 2001) as it is known only from the type, Solanum naucinum is of Deficient (DD) (Knapp et al., 2007), but it is likely to be Endangered (EN) or Critically Endangered (CR) due to low population size (common in other members of the S. dolosum species group) and very restricted range (Knapp, 2005). SPECIES INCERTAE SEDIS 18. Solanum sumacaspi 5. Bot. Gard. 92: 248. 2005. Urubamba, re Knapp, Ann. Missouri TYPE: Urubamba to Cusco: Quilla- bamba, betw. Ollantaytambo & Abra Malaga, 1312155, 72^17'49" W.. ca. 3500 m, 12 Sep. 2002, M. Ackermann & D. Kollehn 289 (holotype, USM!; isotypes, BM!, BSB!, HUSA not seen, M!, NY). Small shrubs to 2 m tall; Peru. from young stems and leaves minutely and sparsely papillate, soon completely glabrous, pale greenish yellow and shiny, a few (at most ] or 2) simple uniseriate trichomes sometimes present on very new stems; older stems ridged, shiny: bark of the older stems and trunks pale greenish white, exfoliating; sympodial units difoliate, appar- ently geminate, the inflorescences borne at the branch üps and overtopping the leaves. Leaves simple. elliptic, completely glabrous adaxially and abaxially, shiny, thick and fleshy, drying olive green adaxially, pale bright yellowish green abaxially; major leaves 3— 1.3-5 cm, the apex acute to somewhat truncate, the base attenuate onto the petiole; primary veins 5 to 10 pairs, raised and yellowish green adaxially, prominent and bright yellow abaxially; petioles 0.1— l em, ridged, the base remaining as a prominent scar on the older stems; minor leaves not differing from the major ones in size or shape. Inflorescences opposite 1-2 em, 7- to 10-flowered, the leaves, simple, glabrous and shiny, the peduncle 0.5-1 em; pedicels at anthesis somewhat deflexed, 0.5-1 em long, tapering from the abrupt base of the calyx tube to a slender base 0.5-0.8 mm diam., green and shiny, articulated at the base; pedicel scars widely spaced ca. 2 mm apart, the pair of scars closer than the distance between the sear pairs, flowers borne from near the base of the inflorescence; buds when very young appearing globose, the corolla soon exserted from the lobes, the buds later calyx becoming ellipsoid just before anthesis. Flowers all apparently tube —2 mm, broadly and 2-2.5 mm, deltate to glabrous, the margins thickened and drying white, the perfect; calyx open conical, the lobes rounded, tip with a thickened mucro ca. 1 mm which may be swollen in fresh occasionally notched; 1—2.5 cm material, corolla white or cream-colored, 1. diam.. 444 Amnals era ried Garden lobed nearly to the base, the lobes planar or somewhat campanulate at anthesis, glabrous on both surfaces, the tips of the lobes minutely papillose and somewhat cucullate; filaments with the free portion ca. 1 mm long, the tube ca. 0.25 mm long to absent; anthers 5—6 X I-1.5(-2 teardrop-shaped, the tips thickened and paler; ovary — mm, poricidal at the tips, the pores glabrous; style straight, 1-1.2 mm; stigma capitate, green. Fruit and seeds not known. Distribution. In steppeland in southern Peru, forests near 2100—3400 m. with most collections from the vicinity of high secondary from Cusco. Discussion. Solanum sumacaspi was included i the of S. similarly glabrous species from delimitation daphnophyllum Bitter, a olivia in the S. nudum species group, with some reservations in the 2002a). Additional collections confirmed the distinctness of monograph of section Geminata (Knapp, S. sumacaspi, both morphologically and in habitat. Solanum daphnophyllum is a species of middle to high elevation (700-2800 m), dry to semi-humid forests on the eastern Andean slopes from central Bolivia to the Department of Puno in Peru, while S. sumacaspi is much disturbed forests ir > only found in higher elevations (2100-3500 m) in grassland (puna)/forest transition areas in the Urubamba River drainage. Morphologically, 5. sumacaspt differs from S. daph- nophyllum in its more elongate inflorescences with more widely spaced pedicel scars and in its pale green, exfoliating bark. Solanum daphnophyllum has inflorescences usually only ca. 1 em long and dark reddish brown shiny bark on all stems. The apparently geminate sympodial units also differentiate 5. suma- caspi from S. daphnophyllum, which has difoliate, but not geminate, units. Fruiting material is lacking for Solanum sumacaspt, preventing its firm allocation to one of the putative the Geminata, sensu Knapp, 2002a). Should the seeds be flattened-reniform, S. sumacaspt is likely to be a member of the $. rada species group (sensu Knapp, 2002a), along with 5. daphnophyllum. lt differs from other completely glabrous members of the 5. nudum species groups of Geminata clade (section species group in its unwinged pedicels, deltate calyx lobes, and high elevational distribution. — Solanum sumacaspi also morphologically resembles two species in the S. arboreum species group (sensu Knapp, 2002a) in its showy inflorescences of flowers all apparently opening at once: S. gratum Bitter of the Cordillera de la Costa in Venezuela and 5. plowmanit Knapp of the western Andean slopes in Peru and Ecuador. also are similar to 5. The shiny stems goniocaulon of the eastern Andean slopes of northern Peru and southern. Ecuador, also of the 5. arboreum species group. It can be differentiated from all those species by the combination of its yellowish exfoliating bark, shiny leaves that dry a very pale green, its non- anisophyllous leaves, and its high-elevation southern Peruvian distribution. Solanum | sumacaspi to be appears relatively common in the narrow Cusco area. lt can tentatively be assigned a conservative preliminary conservation status of Endangered (EN) based on IUCN criteria IUCN, 2001) of range (< 5000 km”), locations (< 5 to 6), species was treated as Data Deficient (DD number of the by Knapp and habitat disturbance: — et al. (2007). Further investigations at the population level, however, may reveal a more significant threat status. Label data indicate the species grows in disturbed forests, but this may be indicative of the highly altered state of the habitat in the region rather than indicative of the weedy status of 5. sumacaspt. Additional specimens examined. PERU. Cusco: Uru- bamba, a Huaytar e Calatayud et al. 966 (BM, MO M). m et al. 373 (BM, CUZ not seen, HUT not seen, A SM not sen) Urubamba, Yucay Huaran, Farfán & bnt 424 (MO, NY) Urubamba, Macchu | above town of Page ha, Peytor : Urubamba. Macchu Pice trail in Tres oe Blancas, 1 km Peyton & Peyton 1026, 1026b (MO); Valenzuela et al. 1214 (MO, NY). n 1. along Înca from is. Calca, Lares. Manto, REVISED KEY TO THE SPECIES OF THE GEMINATA CLADE DICHOTOMOUS KEY TO THE SPECIES OF THE GEMINATA CLADE In contrast to the keys in Knapp (2002a), here | include all the species of the Geminata clade in a single key, rather than breaking them up into the pragmatic species groups defined in that monograph. The key includes only species found in the New World; if you are in the Old World and encounter a species of the Geminata clade, it is Solanum spirale with red fruits on deflexed pedicels or the introduced diphyllum L. with orange fruits on erect pedicels (see Knapp, 2002a). I have tried to use vegetative characters as much as possible, largely because many of these plants are usually found sterile, and also because several species are known only from flower or fruit. The key should be used conjunction. with Table 2, where the geographic distribution of all the New World species of the Geminata clade is listed. Where two species differ radically in distribution but indicated. their key out with the same lead, | have distribution parentheses after the species name, e.g., couplet 5. Volume 95, Number 3 Knapp 2008 Solanum havanense Species Group Key TO THE New WORLD SPECIES OF THE GEMINATA CLADE, INCLUDING THE SOLANUM HAVANENSE SPECIES GROUP la. Mature leaves glabrous, with no well-developed trichomes more than a single cell long. 2a. Sympodial units plurifoliate, difoliate, or unifoliate, never geminate. 3a. Sympodial units plurifoliate or vas 4a. Inflorescence branched (> 2 5a. cs of river courses; ds saves linear to lanceolate ...... 22e. palmillae Standl. (Mexico); 5. monadelphum Van Heurk & Müll. Arg. (E slope of Andes) 5b. Not x of river courses; leaves elliptic to obovate. ja. Stems strongly winged — Pm S. alatirameum 6b. Stems nol winge bu orange, on erect ee ds Dolus e ede US e S. argentinum "an green, on deflexed 8a. Sympodial units difoliate or ilia te. 9a. Leaves drying black; flowers greenish; plants of SE Brazil. ... S. cordioides 9b. aves drying golden haus ially; flowers white; Andes ... S. laurifrons 8b. Sy padia units plurifoliate 10a. Leaves with resinous sic -like domatia in leaf axils abaxially .. 5. pabstii 10b. Leaves with no pit-like domatia ....... .. 0000008 iS. aaide 4b. Inflorescence pla or al most furcate. lla. Flowers purple. 12a. Fruit green, conical with a pointed tip; pericarp woody ........ iss. S. conocarpum 12b. Fruit blue or blue-green, ellipsoid; pericarp thin. New growth completely glabrous... S. troyanum New growth pubescent with appressed uniseriate trichomes ....... S. havanense llb. Nu whi le. — 14a. Le aves drying golden abaxially 5a. Shrubs with pale bark; C tordillera de Avila, Venezuela 15b. Lax shrubs with reddish brown bark; Andes of Colombia 14b. Leaves not drying golden abaxially. e densely arac hnoid pubescent, the trichomes soon deciduous; plants O ER S. canoasense lob. New o glabrous; plants of N South America and the Caribbean. edicels strongly winged; Jamaica ....... lille. S. acropterum 17». 1 'edicels not winged; South America. 18a. Bark of mature stems pale and shiny; plants of high elevations . S E E rd E Ge tie hee A de VERD E a 18b. Bark of mature stems reddish brown; plants of low elev "E in COUTSES 2 ig Ux verbe EUR NUES ORE UI eU aov spe 5. S imber Bitter Jb. Sympodial units unifoliate. 19a. Inflorescence axis filiform; fruiting pedicels elongate. 20a. Ste ms glabrous. nflorescence axis elongate, to 30 cm; leaf margins ruffled; W Andean slope pes "pur Radars Bitter 21b. Inflorescence congested, < 1 em; leaf margins plane; Andean Peru N id a cun diana ue ea S. habrocaulon S. Knapp 20b. Stems variously pubescent, the trichomes usually golden 22a. Inflorescence axis elongate, to 30 cm; petals planar a at anthesis eee eee S. leptorhachis 22b. Inflorescence axis shorter; petals reflexed at anthes l. 23a. Inflorescence l- to 5-flowerec 24a. Leaf bases acute ........... <<<... S. pertenue us & C. V. Morton 24b. Leaf bases oblique svc cece ie s . darienense e Knapp 23b. NC 'ence 5- to 15-flowered, more elongate. 5a. Stem trichomes 1-1.5 mm, usually curling ........... 4 S. capillipes Britton ae Stem trichomes shorter, erect. 26a. Calyx lobes minute, < 1 mm: plants of the northern Pera in enezuela and Colombia ............-.. S. dissimile C. V. Morton 26b. Calyx lobes deltate, longer; plants of E Andean slopes 2 Leaf margins ruffled; inflorescence axis not marked zig- WOO. sete Gace dn ee tee YES Re ea EE S. confine Dunal 1 seal margins plane; inflorescence markedly zigzag .. S. gonyrhachis 19b. lge ‘ence axis stout; fruiting Miro aa not very elongate. a. New growth variously pubescent, sometimes minutely so. Trichomes of new growth golden, erect D. Trichomes of new growth ERA arachnoid S. b s £e ES SS up dre Sura n oto iut S. bahianum (SE Brazil); S. cornifolium Dunal (N Andes) 28b. New growth completely glabrous. 446 Annals of the Missouri Botanical Garden 30a. Leaves markedly 3-veined from the base. 3la. Leaf venation with parallel secondary veins, looking like Melastomataceae a. - A SR O triplinervium » V. rton 31b. Leaf ve nation not like Melastomataceae ....o.o.o..o... S. m 5: e 30b. | weaves nol: ned from the base 32a. Leaf margins ruffled; all orders of venation prominent .. S. ombrophilum S. Knapp 32b. Leaf margins plane; venation finely parallel or obscure. 33a. Venation obscure; fruiting pedicels erect 5... llle. S. bellum S. Knapp 33b. Venation finely ld fruiting pedicels deflexed ......... 5. elvasioides 2b. Sympodial units difoliate, geminate. 4a. Leaves of a geminate pair not markedly anisophyllous; minor leaves not differing in shape. 35a. Inflorescence axis variously pubescent, the trichomes uniseriate. 36a. Trichomes of inflorescence axis and new growth floccose, arachnoid 5... llis. Lia ee qu e ee ree S. nigricans (Mexico to Honduras); 5. maturecalvans Bitter (Andes) 36b. Trichomes of inflorescence axis and new growth erect, usually golden. 37a. Fruit pubescent; leaves always petiolate .... S. oppositifolium Ruiz & Pav. (Amazonia): eux ieu cube A 5. turgidum S. Knapp (Peninsula de Paria, Venezuela) 37b. Fruit glabrous; leaves petiolate or more usually sessile ........ S. sessile Ruiz & Pav. 35b. Inflorescence axis glabrous or papillate. 38a. Leaves usually < 9 cm long, never repand. 39a. Leaves drying golden abaxially. 40a. Bark red and exfoliating; Mexico to Panama ......... S. tuerckheimii Greenm. 40b. Bark nol red and exfoliating; South America 2... gratum (Cordillera de Avila, Venezuela); S. plowmanii (Ecuador and Peru) eaves nol ligt golden abaxially. 39b. 1 Ila. Fruiting pedicels not erect, slightly or strongly deflexed; South America .. pt io did S. abd Vell. (SE Brazil); S. barbulatum (nd es) Alb. Fruiting pedicels erect; Central Ameri l2a. Calyx lobes long-triangular ........ S. ramonense C. V. Morton & Standl. 42b. Calyx lobes quadrate to deltate 2... nananana ollis. S. roblense Bitter 38b. Leaves usually > 10 cm BE to 35 em, often re cun 43a. lr illoréso 'ence branched (> 2X). 4a. Stems strongly winge m plants of SE Brazil ...oooooo.o.o.oo.o.o.. ar avon 44b. Stems terete; plants of N South America... .. obovalifolium s ex Bei . (Cordillera de la Costa, Venezuela); S. MR (E Andean slope, ido 43b. lloras 'scence simple, or al most furcate. 45a. Calyx lobes petaloid or lone: triangular. 46a yen drying pale green; plants of SE Brazil ....... 5. warmingii Hieron. 46b. Leaves drying brownish; plants of Bolivia to Argentina. ........... 15b. lobes deltate. Inflo orescence axis glabrous (in S. lindenii o o fruit < 2 em nii (Peru and Bolivia); PTT S. narcoticosmum — ue xico and Central America) 47b. Inflorescence axis papillate; fruit > 2 em diam. when mature. 48a. Stems winged: buds ene qe in calyx until just before anthesis . oie Se eee Eee hee od eee tee eee 5. Wo neve ver p 8b. Stems terete; buds early exserted from calyx .. S. cucullatum S. Knapp 34b. Leaves of a geminate pair markedly anisophyllous; minor leaves differing in shane and si 49a. Fruiting pedicels deflexed at maturity, often from weight of fruit. 50a. Fruiting pedicels stouter. not markedly longer than in flower; inflorescence axis congested, stoul, »la. Flowers 4-merous; anthers e ly connivent, sagittale oonu anuau 5. sagittantherum tb. Flow 5-merous; anthers tightly connivent, not sagillale 2.2... llus 5. leucocarpon 50b. F ruiting pe vedi els Ren quem FW in flower); inflorescence axis usually filiform or thin. 52a. Minor leaves stipule-like, very small and orbicular; bark reddish .......o..... D T E UT ERIT 5. leptopodum Van Heurck € Müll. Arg. 52b. Minor le saves various, not stipule-like: bark pale. dda. | s strongly winged; bark white and exfoliating; plants of SE Brazil .. S. stipulatum 53b. "i terete; bark pale, but not exfoliating; plants of montane regions, Central S. pastillum S. Ka (Central America); S. hypocalycosarcum Bitter (SW Ecuador) 49b. F Mp pedicels erect al maturity. 54a. uit red or orange (occasionally bright yellow) at maturity; pericarp thin; seeds flattened, Ss vellow. 554; Fruit > lem diam., red or orange-red: flowers > | em diam., the corolla lobes PINAL qua e Seg est an od ee ae a dog E don Shee Be 5. pseudocapsicum Volume 95, Number 3 Knapp 447 2008 Solanum havanense Species Group 55b. Fruit < 1 em diam., pale orange; flowers < 1 em diam., the corolla lobes iu "——————————ÉÉPÉÉEmO AN S. diphyllum 54b. E green; seeds ovoid, variously brown. orescence branched. scence elongate, the pedicel scars ice widely spaced; petals strong Ae GIlexedi ioco dren toy acid eas Did ass S. sieberi Van Heurek € Müll. Arg. 57b. Infloresec ‘ence congested, the pedicel scars ov news petals planar or slightly reflexed. 58a. n ruit glabrous; leaves sessile... S. sessile 58b. Fruit pubese ent; leaves petiolate 2... csse . S. opp md cee a S. turgidum (Peninsula de Paria, Venezuela) 56b. Inflorescence imple or flow 59a. Calyx lobes long- ide 60a. Flowers sessile ... eee 5. tanysepalum 5. Knapp Ob. Flowers not sessile. an inflorescence axis present Ola. Inflorescence axis elongate, the pedicel scars spaced; buds SUITED O ote . irregulare C. V. Morton 6lb. Inflorescence axis stout, congested, the NS scars overlapping: buds globose 62a. adds shrubs; new growth sometimes golden puberulent: leaves: sessile o co o ds S. arboreum 62b. Erect shrubs; new growth various; leaves petiolate ...... S. ramonense (Central e 5: gn et Van Heurck & ba TAE Müll. Arg. (E slopes of Andes, Amazonia) 59b. Calyx lobes deltate or rounded. 53a. Leaves with a well-differentiated petiole .............. S. ripense Dunal 63b. Leaves attenuate at base, the petiole not well-differentiated. 4a. Leaves shiny adaxially, the midrib keeled ....... S. lucens S. Knapp 64b. Leaves dull adaxially, the midrib not markedly keeled. 65a. Inflorescence axis with tiny golde n erect trichomes ...... Sapo tes Ren ce M Ae Fb ots E UU S. oppositifolium 65b. Inflorescence axis glabrous or pubescent, the ue not golden and erect. 66a. Small trees; buds elliptic .........o.o.o.ooo o... 5S. sieberi Rhizomatous shrublets; buds globose ........ S. arboreum e Ib. Mature leaves with at least some well-developed trichomes more than a single cell long 67a. Tric se variously branched. 8a ichomes of new growth and leaves floccose, arachnoid, the structure and branching not easy to FEAR > 69a. Sympodial units not gemina 70a. Pubescence dense on pm undersurface, lamina not visible O RPM EE M E "Er TIL daa 5. e M Van Heurck & Müll. Arg. 70b. Lamina clearly visible through pubescence : leaf undersurfac 71a. Trichomes in tufts in vein axils abaxia ACIE S. laevigatum Dunal 71b. Trichomes on lamina or veins, not in i ts a. Calyx lobes deltate; fruit sometimes pubescent; N South America .. S. cornifolium 72b. Calyx lobes absent, a mere rim; fruit glabrous; S Bolivia, CI EE" A E S. platycypellon 5. Knapp 69b. Sympodial units difoliate, geminate. 13a. Pubescence dense on leaf undersurface, lamina not visible ............. S. ochrophyllum 73b. Lamina clearly visible through pubescence of leaf undersurface. Trichomes in tufts in vein axils abaxially, reddish ..........o..o..o... S. laevigatum a Trichomes se iu on We or veins, or if in tufts, the trichomes grayish. 75a. Corolla 1.5-2.2 cm diam.; Andes serar erinit 20200000005 S. maturecalvans 75b. AL i .3 cm diam . nigric 68b. Trichomes of new growth and leaves dendritic or echinoid, with distinct stiff branches. 76a. Inflorescence branched, > 2X, complex. Mexico to Honduras ..... lees S. nigricans 77a. Fruit orange when ripe, sometimes n yellowish orange .....o.o.o o... o..-. S. argentinum 77b. Fruit green or greenish yellow when ripe 78a. e wingec 79a. ia oe geminate, petiolate... 0.0.0... 0.000000 0 oe 5. RD. num Bitter 79b. Leaves not geminate, sessile ....... eee delitescens 78b. Stems terel 80a mii units plurifoliat Tric homes loose, wea L with elongate branc Pubaaconce reddish, the branches of nc trichomes congestec d. "crc rc "I PEE S. venosum Dunal 448 Annals of the Missouri Botanical Garden 82b. Trichomes transparent, the branches of the tric Es loos "rrr, S. oblongifolium Dunal 81b. n homes uniseriate, erect, the branches shor ubescence gray, the trichomes some UT [loc cose; leaf ba neate and somewhat caer Andes . S. hy] xaleurotrichum. Bitter 83b. Pubesc ‘ence golden or reddish: leaf bases acute; SE Brazi Argentina, and Paraguay 2... sees S. EA 80b. Sympodial units difoliate. 84a. Fruit glabrous ..........ooo ooo ees S. clivorum S. Knapp 84b. : ruit pubescent, the trichomes minute and golden. corolla 0.8-1 em diam.; plants dioecious; coastal Ecuador and b] GEL orate pcg e Venues deett ruber rag: ra qe deem T ads S. confertiseriatum Bitter 85b. Corolla 1-1.5 em diam.; plants not dioecious; Venezuela .. S. turgidum 76b. Inflorescence simple or at most a vale, "ruit red or orange al maturity. 81a. Trichomes mixed Bun and dendritic; sy m t units difoliate or trifoliate .. S. delicatulum 87 " Tric homes all dendritic; sympodial units difoliate SAVES ea panera pr doe ed ttm ea zur S. spissifolium Sendtn. 88b. Le aves ellipti 89a. ed dense on both leaf surfaces: Me vt 1-2 mm, io O TET S. kleinii L. B. Sm. & Downs 89b. Pubescence denser on undersides; trichomes 0.25-0.5 > mm, ye los Drown ANO ils apsicum 86b. Fruit green at maturity 90a. Leaves sessile or nds dly decurrent on the petiole. Ola. Sympodial units not geminate; anthers obovoid; calyx lobes MARA S TEM Box 91b. spots units geminate; anthers ellipsoid; ¢ idus lobes deltate or angular, 92a. Fruit pubescent; corolla > 1 em diam.; minor leaves orbicul: AN Di mallet S S. Knapp 92b. F ruit glabrous; corolla < 1 em diam.; minor leaves elliptic ........ inthophaeum Bitter (Andes); S. pseudodaphnopsis (restingas of SE Brazil) SOS eaves dicuact n pies w h^ m ue units unifoliate. m e: i. scence stout, congested; leaves coriaceous, thick .. S. nutans Ruiz & Pav. 94b. In florescence filiform, the pedicel scars widely spacec d leaves membra- PREX 5. naucinum 93b. Sympa units difoliat 95 Adaxial leaf sue e em snsely pubescent with simple tric homes; leaves large and Sepa. x 1s sace oo lisina 5. superbum S. Knapp 95b. Adaxial leaf surface variously pubescent or glabrous, the trichomes not simple; leaves not large and repan 96a. dne leaf surface shiny and glabr 97 Tric homes of the abaxial leaf aire e found in tufts in the vein S. 98a. Seeds flattened, yellowish brown; flowers usually < E E S. nudum 98b. See d ovoid, dark brown: flowers > 1 em diam. . S. smithii 97b. Trichomes of abaxial leaf surface on lamina and veins, not in tufts. 99a, Trie homes of abaxial leaf surface mostly confined to the EE E EE S. microleprodes Bitter 99b. Tric ME of abaxial leaf surface on the lamina and veins, dendritic 100a. New oh resinous; Andes ....... is. 5. lindenii 100b. New growth not resinous; SE Brazil ... S. arenarium 96b. Adaxial leaf surface variously pubescent. Ola. Seeds flattened. 2a. Marginal leaf trichomes with multicellular bases .. S. cassioides 102b. No leaf trichomes with multicellular bases 103a. 103b. eaves bullate .......o..... S. youngu S. Knapp Leaves not bullate. = Ma. Leaves densely red papillate and s pud ‘ent, the trichomes weak and tra Mb S: elis Dunal = 104b. Leaves with uniseriate trichomes on yon E nsely papillate ....... 5. tovarii S. Knapp Volume 95, Number 3 Knapp 449 2008 Solanum havanense Species Group 101b. ne ovolt . Fruit k insely e E with densely branched trichomes; SE Brazil 5. gnaphalocarpon a e s of $ 105b. Fruit e E of the Andes 1( Joa. Leaf surfaces abaxially loan from dens pubescence ............-. S. tunariense e Kuntze 106b. Leaf surfaces clearly visible abaxially, trichomes loose ..... S. callianthum G, * Morton 67b. Dd. simple or at most furca a. Trichomes of leaf a ene 'e in tufts in vein axils or tucked along midrib, not on lamina or veins (if so very few trichomes). 108a. Inflorescences branched (> 2X). = 09a. Leaves thick and fleshy; plants of high elevations in central Peru ....... 5. amblophyllum 109b. Leaves membranous; plants of lower elevations, coastal regions. 110a. Petals strongly reflexed at anthesis; buds ellipsoid ................... 5.5 eberi 110b. Petals campanulate at anthesis; buds globose ...... 5. campaniforme Roem. Se le 108b. ra Ei 'es ie simple Dx at most furcate). lov fleshy, > 1.5 em diam.; lobes + planar at anthesis. “1124 e es petaloid. E Plants diying black... ceu dew oh one wet APER 5. caavurana Vell. 3b. Plants drying pale preen sacs eA Eust RUNE RR ES S. warmingii 112b. Cal lobes deltate, not petaloid. l Fruiting pedic el swollen apically. ST S. corumbense S. Moore y DASE attenuale face 9 93 3x esp IUS tS Ros wis id 115b. Leaf base acute to Ble oe ee ie Pb urs eee oe 5. leucocarpon 114b. Da pedicels not Biber swollen apic ally. Leaves not casal TOP PET S. validinervium Benítez & S. Knapp Yee Leaves gem 117a. Leaf hd es reddish, weak and collapsing, on lamina .. . ychot oe os Leaf trichomes white, uniseriate and erect... 5. KU ‘Ruiz & Pav 111b. Flowers < 1.5 cm diam.; lobes reflexed at ies 118a. Flower buds ellipsoid (elongate) ............ 5. yd (coastal N South America); S. deflexiflorum Bitter (Andes). 118b. e buds globos Sympodial units tolón NEA IEA DS eS S. mapiricum S. Knapp Te Sympodial l units pr usually geminate. 120a. er buds large; flowers 1.5-2 cm diam ae of leaf undersides toaline tepuis of T enezuela . o ls eat . leputense s. Knapp 121b. a of leaf undersides uniseriate; Sierra Ne vada de Santa S. lasiopodium Dunal 120b. Flower ded smaller; 5 22a. Flov lowers < 1.5 cm diam. wer buds flattened ........o.o.o ooo... ... S. barbulatum — — 122b. Flower buds globose. 23a. Stems winged ............. lessen. 5. santostt S. Knapp 4a. Bark « f twigs white or very pale yellow, even on young stems; veins of leaf undersurface pale. 125a. Stems and inflorescence axis with en uni- eriate trichomes ............. S: i S. Knapp 125b. Hin idi inflorescence axis glabre 126a. Anthers unequal (3 TE pers 2 short) the pores never elongating ....... . ps seudoquina 126b. Anthers all of inei: length, the pore lengthening to slits with age. 127a. Inflorescence axis ps cm; calyx lobes eltate ........ >: la cad S; T 127b. Inflorescence axis 3—8 cm; x lobes AS: - . trichoneuron qua 124b. Bark dark brown or a veins drying dark. 28a growth densely floccose pubescent, trichomes reddish llle S. laevigatum 128b. New growth glabrous or with scattered uniseriate trichomes, not floccose. 129a. Inflorescences of 1 to 2 flowers; calyx lobe long-triangular ........... 5S. di n TF = 450 Annals of the Missouri Botanical Garden 107b. 129b. Inflorescences with > 5 flowers; calyx lobes deltate to triangular. 130a. Inf loresce ‘ence subumbellate, all flow the ap cal part 24 symmetricum » Rusby 130b. fallen eee ‘ence with flowers along entire CUSCO TUE ERE 5. nudum Trichomes of leaf undersurface on veins and lamina, not confined to vein axils. 131a. Pubescence of leaf undersurface on lamina (occasionally also denser on veins 132a. Some trichomes of leaves and/or new growth with multicellular bases 133a. Leaves scabrous, like o to the touch; trichomes unicellular, ‘stiff | S. trachytrichium 133b. Leaves not scabrous; trichomes multicellular 134a. Leaves geminate; Andes ............ 00. cee ee eee eee S. psychotrioides 134b. Leaves not geminate; Central America and N South America... iss. EEEE He gah ie Pav RA A S. vacciniiflorum Standl. a O. Williams 132b. Trichomes uniseriate, without multicellular bases 135a. Sympodial units unifoliate. 136a. Leaves sessile, decurrent onto the winged stem .....o.o.o.o... 5. foetens S. Knapp 136b. Leaves pe e the stem lelele oo... o... S. quebradense S. Knapp 135b. Sympodial units difoliate, geminate. 137a. Leaves of a gemina 138b. Leaf margins ruffled: trichomes soft, yellowish; f 137b. Leaves of a geminate pair the same shape, but often d 39a. Inflorescence branc 140a. Anthers unequal (3 long and 2 te pal arkedly anisophyllous, hed. differing in size and RIS . Leaf margins 25) tric ichomes stiff, curled in one direction; fruit yellow S. fa lconense S Sonn ruit DECON Jc wk a aao 'rosomarginatum 5. Knapp iffering in size short), the pores never e M ue 5: 40b. Anthers ie the pores lengthe ning to slits with age 139b. ie 'scence l4la sim e or al most lurcate. a B. Sm. € Downs E confertiseriatum Anthers pM (3 long and 2 short), the pores never elongating . . date Aa unde Ga LX EU ur E ngs Mah UR UR ER CREE ROB s . reitzi 141b. poe: equal. Flowers > 1.5 em dian 143a. Calyx broadly urceolate, constricted at apex peliole «xs rp eR E EXE S. corumbense 143b. -Calyx-eonieal: «esee Scots, ao Tears 5. tovarit 142b. Flowers — 1.5 em diam. 144a. Flower buds ellipsoid .......o.... 5. incomptum Bitter 144b. Flower buds globose or obovoid. 131b. Pubescence of leaf undersurface confined to the veins, only occasionally lamina. IE 145b. 15a. Flower buds obovoid; plants 146b. Calyx lobes deltate . 5. malacothrix S. Knapp (Mexico); 147a. Sympodial units difoliate to plurifoliate. 148a. Inflorescence branchec a Flowers < len = o trichomes 149b. Flowers > | em 50a. Leaves ses 150b. Leaves peu 1 diam., s. confertisertatum (W Ecuador and Peru); 5. oppositifolium (Amazonia) diam., fleshy: fruit glabrous. sile tiolate 148b. yn a n simple or at most furc pie I5la. Leaf trichomes reddish, 152a. 5 uiting dicels e weak collapsing rect; trichomes all un Lua ak seriale S. intern a few of the Bolivian S. chalmersii iedium Sendtn. (Brazil) trichomes on the nol fleshy; fruit pubescent with minute golden rovirosanum Donn. Sm. ess S. psychotrioides Se 152b. Fruiting pedicels de flexed; some tric i with multicellular bases l i Leaf trichomes w hite, ur useriate and ¢ S. ia cl. 53a. Leaves of a geminale pair E e differing in size and shape .. S. i y D Le saves of a geminale pair not markedly anisophyllous, differing only i s. de 147b. Sympodial units MEE asyneuron S. bus Volume 95, Number 3 2008 Knapp 451 Solanum havanense Species Group 154a. . Pubescence of leaf nde de 's minute, uniseriale, 154b. Inflorescence not elongate and slender, 156 ge Inorese 'ence axis very elongate and slender. Pubescence of leaf md ellular k S. tenuiflagellatum S. E (Cordillera de la € du Venezuela) 18: luct rs 0.4—0.5 em diam.. Flowers 0.8-1 cm diam., white; leaf margins plane y d edd with eoi trichomes l- to 30 cm. the trichomes usually unicellular S leptorhachis the tric iematorhachis S. Knapj undersides of trichomes to ] mm long. mes (W » olombia): usually < 2 cm long. eaf apex long acuminate, the ultimate tip blunt greenish white; usi margins ruffled ..... yanamonense S. Knapp oes eas S. dolosum ip acute 1.5 mm long . S. rton & Standl. ad an 158b. Stem puberulent, the les very a it: 160a. F Dui. pe at ‘els 1-1.5X thei ES 3-flowered: plants of Central America 5. pertenue nce lowered plants of South Americ a. "ir length in flower: calyx lobes — 1 long; Andes of Venezuela and C olombia S. dissimile 160b. Fruiting pedicels 2-3 X their length in flower; eal hes ca. SYNOPTIC OBSERVATIONS ON THE SPECIES OF THE GEMINATA CLADE This synoptic set of observations is intended to simplify the task of identifying a member of this large and, at first glance, very uniform group. Sterile plants of members of the Geminata clade are often difficult to identify, and even plants with either flowers or fruit can be difficult. I have included many leaf and whole plant characters so that sterile plants in many cases can be identified to a choice of a few taxa. Once this step has been done, the user can read the descriptions and match distributions in this paper and Knapp (2002a) and consider other characters not used in this list to the plant in hand. All descriptions of members of the Geminata clade are also presented on the Solanaceae Source website (). In the following list, geograph- ical regions and then character states are followed by a list of species epithets in alphabetical order. Epithets in parentheses indicate that the state is relatively uncommon in that species or occurs in only some populations of that species. A question mark (7) following a species epithet indicates that the character is likely to occur, but has not been absolutely verified. Epithets for species described in this paper are in italics. Species that vary in a particular character, for example in leaf trichome distribution, will be listed for each state present in that species. Not all relevant characters are considered; for example, | have listed fruit color other than green, but not fruits green— common in most species in the group. In this way, diagnostic states can be easily seen. Many species of the group are rather narrow endemics, occurring in only one country or region. Table 2 lists species recorded for each country of the Neotropics; this list n long: E Andean slopes and Amazonia ....... S. confine must be used with care, as new collecting is constantly revealing range extensions in these rare forest plants. Plants of Central America and Mexico: aphyoden- dron, arboreum, darienense, dasyneuron, diphyllum, imberbe, incomptum, leucocarpon, malacothrix, micro- leprodes, narcoticosmum, nigricans, nudum, palmillae, pastillum, pertenue, pseudocapsicum, ramonense, roblense, rovirosanum, tuerckheimii, vacciniiflorum, valerianum Plants of southeastern Brazil: alatirameum, arenar- ium, bahianum, caavurana, campaniforme, canoa- sense, cassioides, compressum, evonymoides, gertii, gnaphalocarpon, intermedium, kleinii, pabstii, pseu- docapsicum, pseudoquina, reitzii, restingae, santosii, spissifolium, stipulatum, trachytrichium, warmingii Plants of Caribbean islands (including Trinidad and Tobago): acropterum, arboreum, capillipes, conocar- pum, diphyllum, havanense, nudum, pseudocapsicum, sieberi, triste, troyanum Shrubs of river courses, rheophytes (usually with very narrow leaves): amnicola, imberbe, monadel- phum, palmillae, spissifolium, stipulatum Plants growing in weedy thickets at roadsides: acuminatum, aphyodendron, caavurana, chalmersit, nudum, rovirosanum, sieberi, sumacaspi?, trichoneuron Pré- model with tiers of plagiotropic branches): Pagoda-like architecture (Chamberlain’s or vost’s dissim- capillipes, confine, cruciferum?, darienense, ile, erosomarginatum, | leptopodum, — leptorhachis, morii, nematorhachis, pertenue, stipulatum, tuerck- heimii, valerianum, yanamonense Stems markedly winged: abitaguense, chlamydogy- delitescens, foetens, humboldtia- num, goniocaulon, num, oblongifolium Bark of older stems exfoliating: incomptum, suma- caspi, tenuiflagellatum, tuerckheimii, vacciniiflorum Annals of the Missouri Botanical Garden Sympodial its unifoliate: bahianum, bellum, capillipes, eie cyclo lila darienense, dolosum elvasioides, foetens, gonyrhachis, habrocaulon, peo boldtianum, leptorhachis, longevirgatum, if hachis, nutans, oc yllum, ombrophilum, pertenue, stipulatum, triplinervium, uni- foliatum, valerianum, yanamonense ants completely glabrous, any trichomes not visible to the naked eye: acropterum, alatirameum, cordioides, diphyllum, elvasioides, evonymoides, menee (cultivated), sumacaspi, troyanum, warmingii ium, naucinum, nematornac New growth resinous: amblophyllum, lindenii Trichome yllum, um, m, laevigatum, (nutans), a robustifrons, sagittantherum, superbum, vaccinii- florum, venosum Leaf margins markedly undulate (ruffled): capil- lipes, confine, cruciferum, darienense, dissimile, erosomarginatum, leptopodum, leptorhachis, morii, nematorhachis, ombrophilum, M rus pertenue, stipulatum, valerianum, yanam Leaf pubescence ped (arachnoid trichomes): ba- hianum, canoasense, cornifolium, hypaleurotrichum, evigatum, maturecalvans nigricans, ochrophyllum, piben ello Leaves wes or markedly rugose: amblophyllum, maturecalvans, vacciniiflorum, you aves drying black: alatiram caavurana, campaniforme, cordioides, dasyneuron, foetens, (la- siopodium), narcoticosmum, (symmetricum), (trachy- trichium), troyanum -veined from the base: cyclophyllum, ngii eum, Le triplinervium Leaves with strongly parallel venation: elvasioides, triplinervium Leaves sessile: alatirameum, delitescens, malletii, pct — s of a geminate pair anisophyllous, markedly fesses in size and differing in shape: anisophyllum, arboreum, erosomarginatum, hypocalycosarcum, irre- gulare, leptopodum, leucocarpon, lucens, malletii, ombrophilum, sagittantherum, tuerckheimii af and stem trichomes echinoid or Christmas- tree-like: compressum, gnaphalocarpon, arrian chum, nutans, pseudodaphnopsis, xanthophaeum Trichomes dendritic or antler-like: arenarium, argentinum, callianthum, chlamydogynum, clivorum, compressum, confertiseriatum, delicatulum, eroso- marginatum, gnaphalocarpon, kleinii, leucocarpon, lindenii, naucinum, (nigricans), nutans, oblongifo- lium, oblongum, pseudocapsicum, pseudodaphnopsis, smithii, spissifolium, superbum, triste, tunariense, youngi Leaf trichomes simple, long and uniseriate: capil- lipes, chalmersii, (confertiseriatum), confine, bense, dasyn erosomarginatum, falcone oetens, poa irregulare, lasiocladum, he cothrix, monanthemon, nematorhachis, quebradense, pero tepuiense, tovarii, turgidum valerianum euron, foe virgatum, m trichomes with enlarged bases: cassioides, Tia NU (psychotrioides), vacciniiflorum Tufts of trichomes in the vein axils abaxially: acuminatum, aphyodendron, barbulatum, caavurana, campaniforme, compressum, cornifolium, laevigatum, lasiocladum, lasiopodium, leucocarpon, monanthe- mon, nudum, pseudoquina, psychotrioides, santosii, smithii, symmetricum, tepuiense, trichoneuron Pit-like domatia in vein axils abaxially: pabstii Pubescence uniform and dense on leaf undersides: arenarium, argentinum, callianthum, chlamydogynum, be ressum), icum omen. deli- inatum, gnaphalocarpon, m otrichum, ag eae longevirgatum, malacot malletii, microleprodes, nutans, (oblongifolium), d longum, ochrophyllum. docaps cens, erosoma: icum, reitzii, su- perbum, tovarii, ME valerianum, venosum, oa youngi ves shiny sail, uniformly pubescent on the TR abaxially: arenarium, lindenii, pseudodap: nopsis, tunariense a golden abaxially aa goniocaulon, gratum laurifrons, (plowmanii), tuerckheimii when dry (herbarium , humboldtianum, Inflorescence branched many times: (acuminatum), alatirameum, amblophyllum, chlamydo num, nfe cordio E orum, compressum, confertiseriatum, delitescens, evonymoides, hypaleurotrichum, (lasio dium), laurifrons, monadelphum, (malletii), EE folium, (ombrophilum), oppositifolium, pabstii, (reitzii), (robustifrons), rovirosanum, sessile, (smithii), (tenuiflagellatum), triste, (turgidum), venosum, ungii Set © Flowers sessile, no inflorescence axis present (if so, very small): canoasense, cruciferum, delicatulum, dolosum, elvasioides, monanthemon, pertenue, tany- sepalum Inflorescence filiform, longer than 4 cm: gonyrha- chis, leptorhachis, oo nematorhachis, stipu- atum, tenuiflage Corolla green: ae malletii, xanthophaeum, yanamonense Corolla — conocarpum, havanense, troyanum rolla campanulate: campan P ids PREPS. validinerv Corolla lobes strongly reflexed at old bellum, capillipes, chalmersii, cruciferum, deflexiflorum, di- phyllum, dolosum, erosomarginatum, lasiocladum?, canoasense, Volume 95, Number 3 2008 Knapp 453 Solanum havanense Species Group leptopodum, morii, ombrophilum, pertenue, sieberi, valerianum, yanamonense Corolla lobes tomentose (not merely papillose on the tips and margins) without: na pa aE num, Une compress cornifolium, haloca arcup rotri os klei- nii, (lasiocladum), las, microleprodes, nutans, (quebradense), superbum, (triste), venosum, (xantho- phaeum) a 8y delitescens, gnap lyx lobes petaloid or long-triangular: acropterum, caavurana, (arboreum), delicatulum, irregulare, mar- antifolium, pseudocapsicum, ramonense, spissifolium, warmingii em lobes orbicular (live plants): sumacas, pastillum, of ee lengths (2 long, 3 short): Diac i re Anthers eet connivent: alatirameum, corum- bense, evonymoides, leucocarpon, lindenii, (oblongum) Anthers obovoid in shape, the base narrow: delitescens the base: chalmersii, (havanense), sagittantheru Anther pores small, never ee to slits with nthers sagittate me age: havanense, pseudoquina, reitzii, troyanum Fruiting pedicels erect: amnicola, anisophyllum arboreum, bellum, cyclophyllum, diphyllum, kleinii, laevigatum, lucens, marantifolium, pseudocapsicum, roblense, robustifrons, sagit- tantherum, spissifolium, triplinervium, turgidum, uni- ramonense, ripense, foliatum es pubescent: Roin M callianthum, chla- confertiseri ifolium, cuculla- tum, mA cep mallet triste, turgidum Bista Berries pointed at the tip: conocarpum, nigricans, stipulatum Berries > 3cm at largest axis: abitaguense, p papa superbum erries red or orange at maturity: argentinum, discus kleinii, pseudocapsicum, spissifolium Berry purple or bluish: acropterum?, havanense, troyanum Literature Cited Acevedo-Rodriguez, P. 1996. Flora of “i dem US Virgin Islands. Mem. New York Bot. Gard. 7 Bitter, G. 1913. Solana nova vel status d dis Repert. Spec. Nov. dt eg. 12: 542-555. Bohs, L. 199 oe (Solanaceae). Fl. Neotrop. Monogr. 63: 1-17 . 2005. Ma = oa in sos based on ndhF coa 9 in R. C. Keating, V. C. Hollowell & T. B. Croat Lese, A Festschrift for William G. D'Arey: The € : a Taxonomist. Monogr. Syst. Bot. Missouri Bot. Gard. 1 2007. Phylogeny of the Cyphomandra clade of the genus 5 Misiin (Solanaceae) based on ITS sequence data. Taxon 56: Mug 1026 Carauta, J. P. P. 1973. The text of Vellozo's Flora apre and its Be as date of publication. Taxon 22: 281-284. Carrizo nes C. 2003. M o l da d sii mates y especies afines wmm Solanaceae). een 30: 27-39. pres A. bon boc in Solanum and related ipm (6). ric taxa for the gen Solan E iens a Feddes Repert. 109: 407—427. ird G. 1973 [1974]. pae IX. AR 170. Solanaceae. Woodson Jr. & R. W. cp a Flora of dine i An Missouri Bot. Gard. 60 Dunal, M. -F. 1852. Solanaceae. Pp. e Candolle edion, Prodromus Systematis Naturalis "v Vegetabili : Falkenberg, D. B. 1999. Aspectos da flora e da vegetação pom e iSo nd de Santa tenia sul do Brasil. Insula Forbes, ds x Par & T. Zimmerman. 2005. Outcrossing mechanisms in Solanum conocarpum (Solanaceae), a rare Miis Islands’ endemic. Poster geal at Botany 2005 meetings in i owbird, Utah, U Frodin, D no and ones of big plant genera. ane 5s 8 58-7 storia Fisica y mee de Chile. Botanica Thun lago. i No peii Vol. 1. p t, Paris and Sant Granados-Tochoy, J. C. & C. I. ma 2006. Una nueva especie de num cinia Geminata (Solanaceae) de Solan Colombia. Caldasia 28(1): 1 AUN 2007. Solanum Eneida (Solas) An regu new species from Colombia rediscovered 200 rs after its first 1862. vum of the British West Indian Lowell, Reeve ., London lala. A., G. E. Barboza » L A. Mentz: 2006. Solanum dinde Golanscsan): Nueva cita para Argentina y y su sinonomía. Darwiniana 44: 508-513. «ies n pl 1827. Solanum coriaceum. Bot. Mag. 54: tab. 2708. IUCN. 2001. IUCN Red ce dea and Criteria Version 3.1. Prepared by the IUCN Species Survival Commission. IUCN, United Gland, ar hie Cambridge, Kingdom Jacquin, N. I 1760. Enumeratio Systematica Plantarum. T. Haak, Leiden. . 1763. Selectarum Stirpium Americanarum Historia. Krauss, Vienna. 1780. Selectarum Stirpium Historia Iconibus Pictus. Pritzel, Vienna. PE M. 1992. The vegetation of south-west Ecuador. the Ecuadorian Dry Forest Project 1991. Biosphere Publications, Leeds. United Kingdom Knapp, S 1989. K revision of the Solanum nitidum pta (section Holophylla en pe Bull. s. (Na t. Hi st.), Bot. ge toii section ca (Solanaceae). Fl. Resin. Mon 84: 1405. b. Assessing patterns of Prog endemism in Neopia uplands. Bot. Rev. 68: 2 . 2005. Taxonomic additions ^ Silian section Geminata Solanaceae in Peru. Ann. Missouri Bot. Ga rd. 248-25. 454 Annals of the Missouri Botanical Garden L. Bohs, M. Nee & DN 2004. Solanaceae: A model for linking g genomics and biodiver- sity. Comp. Funct. Genomics 5: 285-291, , D. M. Spooner & B. León. 2007 [2006]. Solanaceae endémicas del Perú. Pp. 612-643 in Spooner. León, et al. (editors), Libro Rojo de las Plantas Endémicas del Perú. 612 Revista Peru. Biol. 13(2): 612-643. Loddiges, C. L. 1828. Solanum myrtifolium. Bot. Cab. 25: plate 14: Mentz, L. A. € P. A. de Oliveira. 2004. Solanum onlenaanae) na regiao sul do Brasil. Pesquisas, Bot. 54: . R. Stehmann. 2003. Solanum pseudodaphnopsts TEN eae), a new species from the critically endangered restinga vegetation in southern Brazil. Novon 13: 97-100. Morton, V. 1944. Some e UP species of Solan, Contr. U.S. Natl. Herb. 2 72. 976. A Revision of n. Wen ntine Species of lanes. Academia Nacional de Ciencias, Córdoba, Argentina. Moscone, E. A. 1992. Estudios de cromosomas qs en Solanaceae de Argentina. Darwiniana 31(14): 261—297. ee, M., L. Bohs & S. Knapp. 2006. New species and lona records in the Solanaceae of Boli 322-3: " 199]. Associations betwe en ia. Trends Ecol. Evol. 6: 179-1t ner & S. Knapp. 2008, l'axonomy P. tomatoes and their atives Willson. jr < (Solanum sections e o P ae Lycopersicon; Solanaceae). Syst. Bot. Mor 186. C s Ao y 37]. Flora Telluriana, part 2. . Dottori & T. Cosa. 2005. órganos ve setas en Solan argentinum (Solanaceae). Kurtziana 31: 21-2€ ole O. Anatomía de Secretariat of the Convention on Biological Diversity. E Global Strategy for Pl: Convention on Biological Diversity and Botanic Garde ns ant Conservation. Secretariat of the Conservation International, Quebec and London. Smith, L. B. & R. "i P. R. Reitz (editor), Flora s Catarinense, fase. SOLA, 1—321. t. J. Downs. 1966. Solanaceae. Y P Tentamen Supplementi ad Systematis D Annaeani editione m decim am sextam al Taxonomic 1979, Literature. Vol. e Le. de gnum Veg. 98. ro M., L. Mustafa, C. Richardson € S. Saunders. 2002. b. genetic. dive E. in a rare. Virgin. Islands’ endemic, Solanum. conocarpum deed pete ab 2002. © a Collections of Appendix 2, while other "iur s in the clade are only in Table 1, the key, and the synoptic observations. The numbering system here applies to Appendices | and 2 only. Solanum havanense species group Solanum pseudocapsicum species group 5. Solanum pios a & Lillo Solanum delicatulum L. B. S — nudum species m. & Downs s group Solanum chalm ersii S. Knapp i Solanum monanthemon S. Knapp . Solanum pa m. & Downs Solanun leucocarpon species group . Solanum evonymoides Sendtn. Solan amblophyllum species group . Solanum elvasioides S. Bnapp Stan arboreum species group volanum turban enanos Toe hoy & 5. Knapp 0 Granados-Tochoy € C. I. Orozco Wann s nigricans species group B. 5m. & Downs solus arenarium species group Solanum canoasense L. Sm. € Downs A. Mentz & Stehmann >. Solanum compressum L. B. Solanum pseudodaphnopsis L. stan dolosum species group . Solanum naucinum S. Knapp se cies incertae sedis . Solanum sumacaspi S. Knapp APPENDIX 2. Exsiccalae for species treated in detail in this paper id & ae D. 289 Pw B 2827 (4), 11230 (4); Ag 1 (5 den P 4(5 ); Alain, Pu 2394 (3). 23961 um deas 0 (3). Aine J. 16 (10); Alvarez, M. R. Amorim, oe 3 (10); Araque M., J. & 235 : 3). vedo, P. R. ls ; Ac es 9Ar482 (5): nen E F. 2. : Baer, G. n 110 (5); Bailey, L. H. 731 (3); Bartlett, H. H. 19631 (5), 19750 (5), 19683 (9. Kis o Qe (5), 20474 (5) Bastian, - 1347 (5); Be 1459 (5), 24480 (1): Bernardi, L. 20283 (5); Bianchi A 2 (5): Blanchet, DES - 30964 (10); Borsint, O. 7: 2 (5). i 21 (5); Bourd y. G. 2017 (5); Brito, H. 5. & Da Vinha M G. 117 (10); Brito. T. 3980 " 2 L. 401 (3), 591 ).2 294 (1), 2294 (1). . N. L. & Earle, 2 & Hollick, / A. 2058 (3), 2063 (3). 7009 (3); Britton, N. I 2807 Volume 95, Number 3 2008 Knapp Solanum havanense Species Group 455 (1); eg, A. 523 (5), 893 (5), 1145 (5), 1334 (5); Brizuela, J. 166 (5), 744 (5); Brunt, M. 2200 (3); Buchtien, O. 7472 (7); Buratovich, F. 446 (5); Burkart, A. 2 5); 2 A. & 7 (5); Butzke, A. et a 7: . 31427 (5); is J. J. 78 (12); Calatayud, G. et al. 966 (18); Callejas, R. et al. 1618 , 4754 (5); Carvalho, A. M. de et 10), 6672 (10); Costellonas. ? 14825 (5); s A. 25819 (5), 23928 (5 da bg 1, A. & Lazare, J. J. 23928 ra, L. 20717 (5); Claussen, P. s.n. (10), 200 (10); Combs. R 114 (3); Comisión para la Medición = un Arco de da 896 (5); Cordini, J. 1. 67 (5); Cozzo, 77 (5); rg a H. 716 (3); s Filho, i 265 (10). gen, ereles, F. 3149 3176 (5), 3179 (5); Descolea ? 1494 e ): s A E. ^B & Alves, M. C. 224 (10); Dos Santos, T. S. Eiten, G. & Clayton, p D. 5790 (6); Ekman, i * 10555 (3), 12876 (3); Euponino, A. & Da Vinha, S. G. 481 (10). Farfán, J. et al. 373 (18); Folli, D. i es ay Fries, R. E. s.n. (5); Fuentes, A. & Novarro, G. 1958 (5). dur A. F. M. 8856 (10); Gramaja, Pi 16 (5); Granados- Y, et al. 389 (12), 844 (12), 860 (12), 888 (13). i W. 8530 (4), 8681 (4), 9000 (4), 10036 (3), 10231 (3), 10237 (3), vos ) 10718 (3), 10752 (1); Harris, W. & Britton, N. L. 10723 (1); Hatschbach, G. 382 (16), 10198 (16), 15030 (9), "o (15), 22707 (9), 30713 (15); Hatschbach, G. & Pereira, E. 13071 (15); Hatschbach, G. et al. 72547 (9 Hawkes, J. G. 36 (12); Hayward, K. J. 2078 (5); Herb. Bot. Dept. Jamaica e ins 6248 (3); Herb. Hoehne 38942 (5); Hoehne, F. C. 10); Hoehne, W. s.n. (6), s.n. (6); Howard, R. A. 5740 (3), Mam (3), 15024 (1); Howard, R. A. & Proctor, G. R. 13765 (3), re Ds A A. M. R. 617 (5), 536 (5), 3045 (5); Humboldt, F H. A. & Bonpland, A. s.n. (12); m 4 (6 » Ibisch & Ibisch 93.1477 (8); ldrobo. J. M. 5437 (12); ltaipú Binacional 668 (5). Jack, J. G. 7317 (3); Jardim, A. de Rosas-Hurtado, N. 1550 (5); Jiménez, P J. 1907 (3); Jàrganedi P. 2622 (5); Jorgensen, 44 (11). unki, a A. et al. 706A (10); es A. 286 (15), 1301 15); Killeen, fi et al. 4209 (5); Klein, R. 4567 (9); tóbal, C. 17526 (5), 27143 (5), 27157 (5), 3 (5), Pe ? 6) Krapovickas, A. ee e 30909 a ed ^ 39269 (5); Krapovickas, A. et al. 2 ; Kummrow, R. 2410 (14); "sicud (5). León, Bro. 7135 (3); Eur C. B. s.n. (1); Lillo, M. 3343 (5), 9837 (5); Lima, J. : Messias Santos, M. 144 (10); dn J. C , J. H. de 4915 (15), 5302 (15); no-C., G. 98 (1: 2 a F. E. 1049 (5); Luno, M. Y. (7). s.n. (35; Madsen, J. E. (5); Malvárez, M. R. 305 653 (5), 691 (5); Un R. 155 (5), 1335 x , W. T. 1666 (3); is, C. F. P. von 625 (10); Mattos Sila, L A. 2d 376 (10), 1208 (10), 1378 (10); Maxon, W. R. 2206 (4); Mentz, " & Sobral, M. 188 (9), 215 (15); Merat P m rtoise, F V. s.n. (3); Mereles, F. & Degen, R. 5570 (5), 5616 (5), 5642 (5), 5859 (5), 6114 (5); Mexia, Y. 4341 (5), 7834 (5); Meyer, : 3652 (5), 3663 (5), 5048 (5), 13127 (5), 9900 (5); Michel, R. de 2 (5); Miller, G. S. 1358 (3), 1490 (1); Morales, G. 497 (12); Morel, I. 7268 (5), 7350 (5); Mori, S. A. 12747 (10); Mori, S. A. et al. 10737 (10), 10813 (10); Morrone, O. et al. 1388 i 5); Morton, C. V. s.n. (5); Murguia R., O. 379 (5), 536 5) Nee 35288 (5), 35317 (5), 35330 (5), 40420 (5), 44597 e, M. (5), 49083 (5), 51723 (5), 51723 (5), 51149 (5), 51171 (5), 51132 (5), 51320 (5), 51132 (5), 51320 (5), 51337 (5); Nee, . & Bohs, L. 50828 (5); Nee, M. et al. 50736 (5), 51722 (5), 51773 (7), 51777 (7), 51795 (7), 52101 (5); 91/A3 (5); Novara, L. J. 527 (5), 535 (5), 4572 5625 (5), 5892 (5), 7248 (5), 7755 (5). O Donell, C. A. 3135 (5), 4321 (5), 4357 (5); Orozco, C. I. 1062 (12); Ousset, ? 112 (5). Pabst, G. & Pereira, E. 6123 (9), 6200 (9), 6281 o e (15); Partridge, W. 61701 (6); dy ersen, T. M. 5 (5) . & Mariano, D. 4323 (5); Pérez, L. n (15); Perkins, J. R. 1375 (4); Pierotti, s. A. s.n. (5), 2 (5), 4012 (5); Pinheiro, R. S. 1490 (10); Pirani, J. R. et al. 3028-A (10); Plowman, T. C. 2729 (5), 3241 (3), 2986 (3), 3241 (3), . (4); ig, E. F. s.n. (3); Proctor, G. 4 (3), 7980 (4), 9911 (4), 16625 (3), 20480 (4), 21339 (4), 23668 (3), 24845 . 26410 (4), 31565 (3), 31768 (4), 33808 (4); yos G. R. 1802 (1); Purdie, W. s.n. (1), 135 (3). — 5), 4856 (5), T E RS * E ae (5); Rambo, B. 44621 (15), 1, R. W. 8 (4); Reitz, P. R. & Klein, R. 7867 (1: 5), 11438 (6 ), iS 15947 (15); Ribas, O. S. & Silva, J. M. 190 (9); Richard, L. C. s.n. (2); Risso, J. L. 156 (5), 295 (5); Risso, L. 239 (5); Rodriguez F. 21 (5); Rodriguez, F. M. 249 (5); Rojas, T. 14486a (5); Rossato, M. et al. 2484 (15); Rugel, F. 94 (3), 144 (3); Ruthsatz, B. 364/2 (5). P R. de la s.n. (3), 294 (3); Sangster, I. s.n. (3); Santos, F. C. 73 (10); Schafer, J. A. 13795 (3); Schulz, A. G. 2041 (5), 3639 (5 : 8170 (5), 15635 (5), 15825 (5), 17038 (5); Sehnem, A. 2203 (9), 3934 (9), 12505 (15); Seidel, R. & Richtre, E. 851 (1); Sellow, F. s.n. (10), 244 (10), 819 (10); Sessé, M. 219 (3), 535 (3); Siñani 261 (7); Smith, D. N. & Canne, J. 5793 (17); Smith, L. B. & Klein, R. M. 13489 (9); Sobral, M. et al. ICN 7991 (15), 111138 (9); Solomon, J. C. 10579 (5); Soria, N. 364 (5), 394 (5); Spichinger, R. RS2185 (5); Stearn, W. T. 140 (3), 319 (3), 499 (1), 537 (3), 634 (3), 865 (3), 821 (3), ; Stehmann, : R. 728 (15), 1524 (15), 1549 (9), 1756 R. & Sobral, M. ICN 111362 (15), ICN Sudke, . 8170 (5), 8896 (5), 15164 (5), 16820 e 17237 (5), 18156 (5), al. 9926 (10), 11202 ( ~ omas, W. W. et 10); Tressens, S. G. et al. 5910 (15 alenzuela, L. et al. (18); Vanni, R. et al. 2362 (5), 4334 (5); Varela, F. 1545 (5); ed C., I. G. & Tapia, E. 1052 (5); Vargas C., I. G. et al. 3138 (8); Ventus. S. 2434 (5), 2655 (5), 5797 (5), 7774 (5 ) us. (5), 9818 (5); Villafañe, T. 74 (5). Wasum, R. 270 (9), 1174 (15), 1422 (15), 9015 (15); Waters, J. s.n. (1); Webster, ^ L. 3700 (3), 13669 (4); Webster, G. L. & Pro 6 (1); Me bster, G. L. et al. 51 (3), Pa O C. 1399 (4); W bon P. UM , J. R. I. 9529 , 9955 (7); Vig C. SM 3); us [^ s.n. (3). Yuncker, G. 456 Annals of the Missouri Botanical Garden APPENDIX 3. Specimens used in the analysis of pubescence in Solanum argentinum. Collections are listed by senior collector only (for complete data, see Appendix 2), and duplicates in different herbaria are listed separately. ARG = i ntina; BOL = Bolivia; PGY = Paraguay. Sample sizes for latitude and longitude = 160; elevation = 70. Specimens for which an elevation was not given were not included in the analysis for altitude. Digital latitude and longitudes are rounded to two decimal places. Digital Digital Elevation Collector Number Location latitude longitude (m) Score Aguilar, R. M. 304 ARG 25.17 04.17 l Alvarez, M. R. 226 ARG 25.17 04.17 | Baer, G. A. 110 ARG — 26.00 —65.30 l Bartlett, H. H. 19631 ARG = 25.50 —064.30 l Bartlett, H. H 19750 ARG 28.25 — 603.92 l Bartlett, H. H 20474 ARG —24.78 — 04.15 1 Bartlett, H. H 19888 ARG — 18.38 — 62.92 2 Bartlett H 83 ARG — 19.27 — 62.47 3 Bartlett, H. H 20321 ARG —29.22 — 65.45 3 Bastian, E. 1347 BOL — 29.12 —65.35 1800 3 Beck, S. G. 6459 BOL —28.62 —65.80 1 Bernardi, L. 20283 PGY —24.82 —65.45 2 Bernardi, | 20283 PGY — 20.05 — 63.53 2 Bourdy, G 2017 BOI — 26.08 —05.97 l Brizuela, A 1145 ARG — 26.08 —65.97 l Brizuela, A 523 ARG —23.33 —04.25 1 Brizuela, A 893 ARG — 22.33 —02.02 1 Brizuela, J 144 ARG — 22.30 —02.53 ] Buratovich, I 146 ARG —22.30 —02.53 2 Burkart, A. 12017 ARG — 22.58 —063.33 1400 3 Cabrera, A. | 31020 ARG — 22.30 —62.53 3 Cabrera, A. | 31427 ARG — 26.08 — 65.97 3 árdenas, M 4754 BOL — 17,58 — 62.67 900 ] Charpin, A 20717 ARG —206.10 —04.32 1700 l Charpin, A 25819 ARG — 26.10 — 64,32 3 Charpin, A 25819 ARG — 26.08 00:91 3 Charpin, A. 5. 20717 ARG =25.25 — 65.42 1700 l Comisión paa la Medición 896 ARG —21.28 — 065.00 2 de un Arco de Meridiano Cordini, I. i 67 ARG =271:92 03.90 2 Cozzo, ? 77 ARG — 27.92 — 63.90 1 Degen, R 3149 PGA =05,17 64.17 1 Degen, R 3149 PGY — 26.82 — 65.22 1 Degen, R 3149 PGY — 206.82 —65.22 ] Degen, R 3176 PGY =259:70 063.07 1 Degen, R 3179 PGY 22.33 —02.62 l Fries, R. F s.n. ARG —22.30 —02.53 l Fuentes, A. 1958 BOL —25.17 =05.17 300 l Gramaja, Á 16 ARG —25.17 —065.17 l amo A 16 ARG —2595T —065.17 1 Haywe 2: 2078 ARG = Zod 09.17 ] s A.M. R. 536 ARG — 18.83 65.25 l 3 Hunziker, A. T 1064 ARG -20.03 63.52 1200 l Jardim 1550 BOL — 18.65 =63:25 1385 3 Killeen, I 4209 BOL — 29.87 03.73 700 3 Krapovickas, Á 27143 ARG — 18.67 —03.24 l Krapovickas, A. 27157 ARC — 18.07 —03.24 1 Krapovickas, A 2715 ARG — 206.88 — 64.73 1 Krapovickas, A 46303 ARG =28.17 —65.07 1 Krapovickas, Á 56552 ARG —25.62 —65.47 2 Krapovickas, A. 17526 ARG 28.20 —065.13 j Volume 95, Number 3 2008 Knapp Solanum havanense Species Group APPENDIX 3. Continued. Digita Digital Elevation Collector Number Location latitude longitude (m Score Krapovickas, A. 28017 ARG —28.17 —65.61 3 Krapovickas, A. 28397 ARG —28.17 —65.67 3 Krapovickas, A 30423 ARG —24.58 —65.5 3 Krapovickas, A 090 ARG —31.40 —64.18 350 3 Krapovick A 30909 ARG — 24.82 —65.45 350 3 Krapovickas, A. 31246 BOL —32.42 —63,25 3 Krapovickas, A. 39269 ARG — 19.58 —62.58 3 Kuroiwa, N. 2332 BOI —26.78 —65.25 3 illo, 3343 ARG — 24.28 —65.10 1 Lillo, M 3343 ARG — 24.28 —65.10 ] Lillo, M. 9837 ARG — 24.28 —65.10 3 Malvarez, M. R. 653 ARG — 24.28 —65.10 1 Malvárez, R. 1401 ARG —26.15 —65.30 3 alvárez, R 155 ARG — 26.50 —64.9 3 Mereles, F 5570 PGY — 26.50 — 04.02 1 Mereles, F 5570 PGY 25.60 05.03 1 Mereles, F 5642 PGY —25.60 — 65.63 2 Mereles, F. 5859 PGY — 25.60 — 65.63 2 Mexia, Y 4341 ARG — 30.20 —64.48 760 l Mexia, Y 7834 ARG —34.58 —58.48 762 l Mexia, Y 7834 ARG —21.68 —64.83 762 1 Mexia, Y 7834 ARG — 29.83 —63.50 762 1 Meyer, 1 3652 AR 22.63 —59.67 1260 3 Meyer, T 5048 ARG — 22.63 —59.67 360 3 Meyer, 1 9900 ARG — 26.90 —60.13 3 Meyer, 1 9900 ARG 23.22 — 64.10 3 Meye 9900 ARG —25.20 — 58.00 3 Morel, I 1268 ARG — 24.68 —64.25 3 Murgia, O 379 BOL 22.97 61.83 3400 1 Nee, M 35288 BOL 23.28 —61.47 820 l Nee, M 49083 BOL —]17.77 —03.18 550 1 Nee, M 50736 ARG =Z —63.18 425 1 Nee, M 51723 BO — 24.07 — 64.27 550 1 Nee, M 1723 BOL — 24.07 —64.27 550 l Nee, M 40420 BOL — 25.53 —65.57 420 2 Nee, M 40420 BOL —25.53 —65.57 420 2 Nee, M 3531 BOL — 25.60 —65.30 800 3 Nee, M 35317 BOL — 24.33 —59.50 800 3 Nee, M 35330 BO — 24.33 —59.50 810 3 Nee, M 44.597 BOL — 24.33 —59.50 500 3 ee, 44597 BOL —24.93 — 58.93 500 3 Nee, M 14597 BO —23.43 60.15 500 3 Nee, 14597 BOL 26.48 65.37 500 3 Nee, M 50828 ARG —25.25 —65.42 1300 3 Nee, M 51132 BOL —24.25 — 64.83 810 3 Nee, M 51132 BOL —21.67 — 64.83 810 3 Nee, M 51149 BOL 55.33 —65.17 850 3 Nee, M 51171 BOL —24.30 64.97 920 3 Nee, M 51320 BOL 25.50 68.50 900 3 Nee, M 51320 BOL — 26.20 04.78 900 3 Nee, M 51337 BOL — 26.20 — 64.7 850 3 Nee, M 51722 BOL — 25.30 — 64.90 520 3 Nee, M 51722 BOL 20.35 —63.47 520 3 Nee, M 51722 BOL — 20.30 —63.27 520 3 ee, M. 51722 BOL — 25,38 —65.10 3 Novara, L. J. 527 ARG 26.47 65.18 2 458 Annals of the Missouri Botanical Garden APPENDIX 3. Continued. Digital Digital Elevation Collector Number Location latitude longitude (m) Score Novara, L. J. 527 ARG 24.90 — 065.48 2 Novara, L. J 5625 ARG 24.75 — 65.25 2 Novara, L. J 5625 ARG — 22.90 — 03.98 2 Novara, L. J 7248 ARG — 22.90 — 03.98 1400 2 Novara, L. J. 4856 ARG —20.75 —63.15 3: H Novara, L. J. 4856 ARG —22.23 —63.70 3 Novara, L. J. 535 ARG —20.07 63.54 1180 3 Novara, L. J 5892 ARG —24.38 —65.10 1200 3 Novara, L. J 5892 ARG — 27.50 —59.60 1200 3 Novara, I 7755 ARG —25.98 — 64.95 1150 3 O'Donell, C. A 4321 ARG —24.90 —65.48 l O'Donell, C. A 3135 ARG =23.33 —64.25 j O'Donell, C. A 4357 ARG —24.78 — 64.23 3 Pierotti, S. A s.n ARG — 24.78 — 64.23 l Pierotti, S. A s.n. ARG —24.78 — 64.23 l Pierotti, S. A 4012 ARG —25 — 58.67 2 Pierotti, S 4012 ARG — 20.40 — 63.28 2 Pierotti, S. A 4012 ARG — 20.40 —63.28 2 Pierotti, S. A 2 ARC 20. 63.28 j Plowman, T. € 2729 ARG — 18.13 — 63.20 77 3 Plowman, T. ( 2729 ARG — 18.13 —63.20 77 j Quintero, J 63 ARG —18.13 —63.20 l Ragonese, ? 274 ARG — 18.13 —03.20 j Ragonese, ? 3000 ARG —23.56 —064.97 j Risso, J. I 156 ARG — 20.09 03.52 l Risso, J. L. 295 ARG — 20.09 —63.52 900 l Ruthsatz, B. 364 ARG — 20.05 —03.53 2120 l Schulz, A. G 17038 ARG — 20.28 —63.45 2 Schulz, A. G 15635 ARG — 19,90 —63.53 j Schulz, A. G 15825 ARG — 19.90 — 063.53 j Schulz, A. G 39 ARG — 19.94 — 63.52 3 Schulz, A. G 8170 ARG —18.14 —63.20 3 Solomon, J. € 10579 BOI —18.14 —63.20 3 Spichinger, R. 2185 PGY — 18.13 —63.19 100 2 Stuckert, T 15164 ARG —18.13 —63.19 l Stuckert, 1 15164 ARG —25.00 — 65.63 1 Stuckert, 1 16820 ARG — 25.60 — 65.63 l nni, R ARG —24.78 — 64.88 3 Vargas, I. G 1052 BOL —25.15 —55.38 400 l Venturi 2434 ARG 05.05 —55.38 550 l Venturi, S 5797 ARG — 23.93 62.55 500 l Venturi, S 5797 ARG 230.39 64.25 500 l Venturi, S 5797 ARG —24.78 — 65.02 500 l Venturi, S 5797 ARG — 23.33 — 64.25 500 l Venturi, S 7774 ARG —21.41 5 1000 l Venturi, S 1851 ARG 27.47 — 58.83 800 l Venturi, S 1857 ARG 24.22 07.07 800 l 9818 ARG — 28.38 —62.12 1300 | Venturi, S 9818 ARG — 26.83 —61.00 1300 l Venturi, S 9818 ARG 27.33 58.93 1300 l Venturi, S 2055 ARG —21.33 —58.93 800 2 Villafañe, T. 74 ARG 27.57 —60.53 l Weisz, P. 5761 ARG =21.82 — 64.07 l Wood, J. R. I. 9529 BOL —24.25 —61.25 1900 l PHYLOGENETIC RELATIONSHIPS Yang Liu,?* Yu Jia,?* Wei Wang," Zhi-Duan OF TWO ENDEMIC GENERA FROM EAST ASIA: TRICHOCOLEOPSIS AND NEOTRICHOCOLEA (HEPATICAE)! Chen,’ E. Christine Davis,* and Yin-Long Qiu? ABSTRACT Trichoc CR S. Okamura and Neotrichocolea S. Hatt. are two water saccate liverwort genera endemic to East Asia whose emali u < Nn =, c positions have been controvers sial. T nine gene 2 265; mitoc a nad5) of 2 margins, water sacs, coelocaules, and interior capsule wall c y and maximum likelihood methods: ( s (chloroplast small ae [SSU], Hee subunit [LSU], atpB, nu psbA, and md. nuclear of o ells with thin w To address their phylogenetic positions and re ee within liverworts, two ) four 4, 265, and es (psbA, rps4 analyses suggested the sister relationship between which is also d b lla characters, such as a lobed leaf with ciliate alls. Traditional classifications usually considered Trichocoleopsis and Neotrichocolea as members of L ne or Trichocoleaceae, whereas our results strongly suggested that the Trichocoleopsis-Neotrichocolea clade allied with Ptilidiaceae, which is supported by the similar ciliate and lobed leaves. However, considering many differences between them, such as in the > perianth, sporeling type, water sacs, and rhizoids, we support a family rank for the Neotrichocoleaceae, containing Trichocoleopsis and Neotrichocolea, as premi; suggested by 974 Inoue in 1 Key words: East Asia, endemic, multiple genes, Neotrichocolea, Neotrichocoleaceae, phylogeny, Trichocoleopsis. Trichocoleopsis S. Okamura and Neotrichocolea S. Hatt. are two liverwort genera from East Asia that are considered to be related based on the shared charac- teristic of bearing water sacs on their leaves. Trichoco- leopsis contains two species, T. sacculata (Mitt.) S. Chen ex M. X. Zhang, and the latter is only distinguished from the Okamura and T. tsinlingensis P. C. former by its entire or only 1- to 3-ciliate lobe margins (Zhang, 1982). Neotrichocolea is a monotypic genus with N. bisseti (Mitt.) S. Hatt. The two genera are restricted to moist forests in China, Japan, Korea, Burma, and the Russian Far East. As habitat destruction increases, all three species of both genera are seriously threatened and on the verge of extinction (Wu et al., 1997; Cao et al., 2006). An updated distribution map of these species is shown in Figure 1. This map is based on our own consultations of specimens as well as other published research (see Table 1). In 1891, Mitten published two species from Japan, Blepharozia sacculata Mitt. and Mastigophora bisseti Mitt; the former is the basionym of Trichocoleopsis sacculata and the latter is the basionym of Neotricho- colea bisseti. Thereafter, the two species were trans- ferred to Ptilidium Nees (Stephani, 1897; Evans, 1905) before Okamura (1911) and Hattori (1947) established the monotypic genera Trichocoleopsis and Neotrichoco- lea, respectively. In addition, Horikawa (1934) suggest- ed that N. bisseti be placed in the genus Trichocoleopsis. Since their establishment, the systematic positions of these two genera have been controversial. Tricho- coleopsis has been placed in Ptilidiaceae (Evans, 1939; Hattori, 1947; Mizutani & Hattori, Trichocoleaceae (Nakai, 1943; Miiller, 1951-1958; Schuster, 1957; Mizutani, 1972), or Lepidolaenaceae (Schuster, 1966); meanwhile, Neotrichocolea has been recognized as a member of Ptilidiaceae (Hattori, 1947; 'The cae thank Y. X. Xiong for abc the materials of Neotrichocolea bisseti; P. C. Wu reading an early draft of the manuscript; C and National Natural Science Foundatio *State Key Laboratory of Systematic g for direction in fieldwor assistance. This research was supported by Naona] Basic Research Progr n of China grants 30121003 na 30228004 to YQ. ECD is supported by National BG "ence nbus gun 0531730 e ipal Mae Jon Shaw volutionary Botany, AE of Botany, Chinese Acade and r L. He-Nygrén for nd in the herbarium; and Q. H. Wang for lab m of China (973 Program no. a to emy of Sciences, Beijing 100093, People’s Republie of China. liuyang@ibe as.ac.cn; yjia@ibcas.ac.cn; wangweil 127@ibcas. ac.cn; zhiduan@ibcas.ac.cn. na. * Graduate School of the Chinese Academ * Department of Biology, Duke University, Box 90338, Durham, North Carolina 27708, U.S.A. christine.davis@d ey sa of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, Mic ha 48109-1048, U.S.A. yiqin Com: (h.e ^ Author forc corre meson didt yjia@ibcas.ac.en. doi: 10.3417/200607 1 ANN. Missouni Bor. GARD. 95: 459—470. PUBLISHED ON 23 SEPTEMBER 2008. Annals of the 460 Missouri Botanical Garden 60E 100 140 I 1 Ve e 1 " i NG E | ¿E N, \ Russia . po Nm | em Le +440 LM “mn, Z ^r | oe | | M C o mms mo) A. P 2 “so U 7 n H y. ) Japan PA . UM Y ( ) ee | | e ¿0.. dA, «ci bad e A L7 & E A 20N LI o — 400 800km Í ( Figure 1. Distribution map of Trichocoleopsis and Neotrichocolea. Filled circle = T. sacculata, open circle = T. tsinlingensis, filled triangle = N. bisseti Table l . Mapped distributions in China and adjacent regions are supported by e 1974 [Japan]; Konstantinova et a (Gao & Cao, 1988, 2000; Hong, 1966 [Korea]: Inoue, Specimen vouchers for geographic distributions of Neotrichocolea and Trichocoleopsis in eastern Asia (Fig. 1). the specimen vouchers or inferred from the literature l., 1992 [Russia]: Koponen et al., 2004: Wu. 1992; Zhang. 1982). Asterisks mark the two specimen vouchers for the new chloroplast and nuclear Pe DNA sequences reported in this paper. IFSBH = Herbarium, Kunming Institute of Botany, Chinese Academy of Scienc Locality Herbarium Collection Literature Neotrichocolea bisseti (Mitt.) S. Hatt. Guang mingding, Mt. Huang, Anhui Province, China Chen et al. 6900 PE Qing liangtai. Mt. Huang, Anhui Province, China Chen et al. 7061 PE Qing liangtai, Mt. Huang, Anhui Province, China Chen et al. 7128 PE Shi xinfeng. Mt. Huang, Anhui Province, China Chen et al. 7161 PE Shi xinfeng. Mt. Huang. Anhui Province, China Chen et al. 7104 PI Shi xinfeng, Mt. Huang, Anhui Province, China Chen et al. 7171 PI Shi xinfeng, Mt. Huang. Anhui Province, China Chen et al. 7172 PI Xi haimen, Mt. Huang. Anhui Province, China Chen et al. 7038 PE Xi haimen, Mt. Huang, Anhui Province, China (alt. 1600 m) Chen et al. 7003 PE Lian huafeng, Mt. Huang, Anhui Province, China Chen et al. 6939 PE Qing liangtai-Er daowan, Mt. Huang, Anhui Province, China Chen et al. 7188 PE Se-Tse-ling. Mt. Huang. Anhui Province, China (alt. 1800 m) Ku, C. M. s.n. PE Bei har, Mt. Huang, Anhui Province, China (alt. 1600 m) Gao, C. H. 21260 SHM Bei hai, Mt. Huang, Anhui Province, China (alt. 1600 m) Hu, R. L. & Gao, C. H. 8679 HSNU Song yanan, Mt. Huang, Anhui Province. China (alt. 1300 m) Hu, R. L. & Gao, C. H. 8737 HSNU Xi haimen, Mt. Huang, Anhui Province. China Chen et al. 9571 HSNU Huang gang, Mt. Wuyi, Fujian Province, China Gao, C. H. 25077 SHM Yuan jiang, Yunnan Province, China rao & Cao, 2000 *ML. Fanjing. Guizhou Province, China He, L. 040522 PE Trichocoleopsis sacculata (Mitt.) S. Okamura Shi zilin, Mt. Huang, Anhui Province, China (alt. 1800 m) Chen et al. 6399 PE Shi zilin, Mt. Huang, Anhui Province, China Chen et al. 6364 PE Volume 95, Number 3 2008 Liu et al. 46 Phylogeny of Trichocoleopsis and Neotrichocolea Table 1. Continued. Herbarium/ Locality Collection Literature Shi zilin-Xihai, Mt. Huang, Anhui Province, China Chen et al. 6497 Xi haimen, Mt. Huang, Anhui Province, China then et al. 6997 PE Qing liangtai, Mt. Huang, Anhui Province, China d 1500 m) Chen et al. 7172 PF w iangtai, Mt. Huang, Anhui -o Chir Chen et al. 7179 PE zilin-Yun gusi, Mt. Huang, Anhui Province, C hina Chen et al. 7428 PE Feng huangsi, Mt. Jinfo, sha deeds China Chen et al. 1632 PE San guandian, Mt. Taibai, Shanxi Province, China (alt. 3000 m) Li X. J. 705 PE Ba xiantai, Mt. Taibai, Shanxi Province, China (alt. 4000 m) Li X. J. 718 PE *Oing liangtai, Mt. Huang, Anhui Province, China Liu Y. 4170 PE Qing liangtai, Mt. Huang, Anhui Province, China Li D. K. 02274 SHM Qing liangtai, Mt. Huang, Anhui Province, China Zheng W. F. 14307 HSNU Qing liangtai, Mt. Huang, Anhui Province, China (alt. 1600 m) Hu L. & 2. C. H. 8632 e Mt. E mei, Sichuan Province, China Gao Q. 19374, 19485 IFSB Mt. E mei, Sic huan Province, Chir ;ao C. H. don 15946 Gao A Cao, 1988 Yang yuping, Mt. Jinfo, e China (alt. 1400 m) Cao T. & Li Q. 41495 IFSBH Po niangyan, Daguan, Yunnan Province, China Yi liang, Yunnan Province, China Qu jing, Yunnan Province, China Mt. Wuyi, Fujian Province, China Bai shanzu Natural Reserve, Zhejiang Province, China (alt. 1010—1600 m) Mt. Jiu long, Zhejiang Province, China (alt. 1300 m) Mt. Xitianmu, Zhejiang Province, China (alt. 1300 m) Mt. Fanjing, Guizhou Province, Chi Mt. Ba dagong, Hunan Province, Chus Mt. Huping, Hunan Province, China Tai dong, Taiwan Province, China Trichocoleopsis tsinlingensis P. C. Chen ex M. X. Zhang Mt. Taibai, Shanxi Province, China Fang yangsi-Ping ansi, Mt. Taibai, Shanxi Province, China (alt. 2800-3000 m) De qin, Yunnan Province, China Da guan, Yunnan Province, China Mt. Wuyi, Fujian Province, C Mt. Wuyi, Fujian Province, C — nna — lina Bai shanzu Natural Reserve, Zhejiang Province, China (alt. 1600 m Li X. J. 4520 KUN Gao & Cao, 2000 - Gao & Cao, 2000 Gao C. H. 25099 Gao & Cao, 1 Zhu R. a 90402, 90413, 90554, HSNU 90564, 90839, 90883 Liu Z. E 3080 HSNU d Y. F. 00405 HSNU He L. s GACP Koponen et al., 2004 Koponen Le 48168 8 us 521 2004. Koponen et al., Wu, 1992 l Zhang, 1982 = Gao & Cao, 2000 CUN X. J. KU Li E K. 11604 SHM 2. K. & H. 11614, SHM po Zhu R. L HSNU Müller, 1951-1958; Schuster, 1966) or Trichocolea- ceae (Mizutani & Hattori, 1969). Schuster (1972) proposed a subfamily rank, Trichocoleopsidoideae, to include Trichocoleopsis under the family Lepidolae- naceae, and later another subfamily, Neotrichocoleo- 1979). (1983) followed these treatments, but because ideae, to include Neotrichocolea (Schuster, Grolle the name of subfamily Trichocoleopsidoideae is illegitimate in nomenclature, he used Trichocoleop- soideae instead. Furuki and Mizutani (1994) placed the two genera in Lepidolaenaceae and kept Tricho- coleopsis in subfamily Trichocoleopsoideae, but they moved Neotrichocolea to subfamily Lepidolaenoideae. Based on differences in capsule wall anatomy and other structures between the Trichocoleopsis—Neotricho- colea group and Lepidolaenaceae, Inoue (1974) pro- posed a new family, consisting of the two genera, called Neotrichocoleaceae, which was accepted by Schuster 1984). The recognized in the recent liverwort classification systems (Crandall-Stotler & Stotler, 2000; Frey & Stech, 2005; He-Nygrén et al., 2006), but the circumscription of the also P cS family Neotrichocoleaceae was family has been disputed. Frey and Stech (2005) and He-Nygrén et al. (2006) did not list the members of Neotrichocoleaceae, whereas Crandall-Stotler and Stot- ler (2000 Neotrichocolea, and Trichocoleopsis should be trans- = suggested that the family should only include ferred back to Lepidolaenaceae. Molecular data ranging from a single gene to multiple genes have been used to resolve the 462 Annals of the Missouri Botanical Garden relationships of liverworts (e.g., Capesius & Bopp, 1997; Lewis et al., 1997; Stech et al., 2000; Wheeler, 2000; Boisselier-Dubayle et al., 2002: Davis, 2004: Forrest & Crandall-Stotler, 2004, 2005; He-Nygrén e al., 2004, 20006; Crandall-Stotler et al. 2005; 2005; Forrest et al., 2006; Hentschel important — Heinrichs et al., 2006). progress in understanding the phylogenetic relation- These studies = el al., represent ships of liverworts at different levels. These studies also provide a useful instrument for resolving the systematic positions of previously unsampled taxa whose placements are controversial or uncertain when based solely on morphology. Among the above studies, Davis (2004), He-Nygrén et al. (2004, 2006), Heinrichs et al. (2005), Forrest et al. (2006), and (2006) sampled Neotrichocolea. All molecular analyses indicated that Ptilidium; however, the position of this Ptilidium—Neotrichocolea — Hentschel et al. Neotrichocolea was closely related to clade within the leafy liverworts was ambiguous. In em these molecular studies, leafy liverworts are divide “Leafy I” Pulidium— ‘Leafy P" or but no results have into two major clades, which were named and “Leafy Il" by Davis (2004). The Neotrichocolea clade has been resolved in * “Leafy II" by different studies, Until has not been included in any molecular study. received strong support. now, Trichocoleopsis In this study, we generated sequence data from four chloroplast genes (small subunit [SSU], large subunit [LSU]. atpB, and rbcL) for Trichocoleopsis and Neo- trichocolea, and two nuclear ribosomal (185 and 265), one mitochondrial (nad5), and two chloroplast (psbA and rps4) genes for Trichocoleopsis. Two different data sets were constructed with different taxon and character sampling schemes. Using these data, we aim to: (1) determine the position of. Trichocoleopsis within the leafy liverworts, and (2) clarify the relationship between Trichocoleopsis and Neotrichocolea. MATERIALS AND METHODS TAXON SAMPLING The majority of taxa sampled for this study were derived from the recent four-gene analysis reported by Davis (2004) and the six-gene analysis by Qiu et al. (2006). Yang (collection number — Trichocoleopsis sacculata was collected by Liu 4170) in 2005 from Mt. Huang, Anhui Province, China (Table 1), and Neo- trichocolea bissett by 040522) in 2004 from Mt. Fanjing, Guizhou Province, China (Table 1). The voucher specimens have been deposited in the herbarium of the Institute of Botany, the Chinese Academy of Sciences, Beijing (PE). This study obtained four new sequences for N. bisseti: He Lin (collection number chloroplast SSU ribosomal DNA (rDNA) (GenBank accession number EF460698), LSU rDNA (EF460696), atpb (EF460094), and rbcL (EF460702); and nine new chloroplast SSU rDNA sequences for T. sacculata: (EF460097), LSU rDNA (EF460695), atpB (EF460693), rbcL (EF460701), rpsá (EF460703), and psbA (EF460700), nuclear SSU (185) rDNA (EF460091) and LSU (205) rDNA (EF460692), and mitochondrial nado (EF460699). For the four-gene data set, 57 taxa representing 57 genera in 34 families of liverworts were included. The sampling came mostly from Davis (2004) and focused on the leafy liverworts. For the nine-gene data set, including 24 genera of 19 families, the strategy was to further stabilize the tree topology as sequence dala increased. Although taxon density of this data set was lower than that of the four-gene data set, all major lineages of leafy liverworts were sampled. According to revious studies (Davis, 2004; He-Nygrén et al., 2004, 2006: 2005; Forrest et al., 2006), six genera from simple thalloid liverworts were chosen as — Heinrichs et al., outgroups in all analyses. In this study, classification 2000; — — followed Crandall-Stotler and Stotler MOLECULAR AND PHYLOGENETIC ANALYSES DNA extraction, polymerase chain reaction (PCR) amplification, and sequencing protocols followed Davis (2004) and Liu et al. (2005). All alignments were made using CLUSTAL X 1.83 (Thompson et al., 1997) and then adjusted manually in BioEdit 5.0.9 (Hall, 1999). By comparing the alignments made by Davis (2004) and our (mainly in the nad5 and 20S) were excluded from own, ambiguous positions phylogenetic analyses. Because we were unable to sequence the 185 gene from Neotrichocolea, despite repeated attempts, this was coded as missing The deposited in TreeBASE (; sequence data. resulting matrices have been accession number SN3273). Phy out using maximum parsimony (MP) and maximum likelihood (ML) methods in PAUP* version 4.0b10 (Swofford, 2003) and PHYML version 2.4.3 (Guindon 2003), heuristic searches were conducted with 1000 repli- = ogenetic analyses for each matrix were carried = & Gascuel, respectively. For MP analyses, cates of each 100 random sequence addition, tree bisection-reconnection (TBR) branch swapping, Mul- unordered and Trees in effect. All characters were equally weighted. and gaps were coded ds missing data. To assess node support, bootstrap analyses (Felsenstein, 1985) were performed using 1000 heuristic search replicates as described above. Likelihood analysis was performed using a general time reversible (GTR) substitution model with invari- ant sites and additional among-site rate variation Volume 95, Number 3 2008 Liu et al. 46 Phylogeny of Trichocoleopsis and Neotrichocolea Table 2. Comparison of 57-taxon and 24-taxon MP tree statistics. CI, consistency index; RI, retention index. ota Variable Variable informative Tree characters characters sites Trees length CI RI 57 Taxa Nuclear only 1077 244 130 (12.1%) 510 686 0.484 0.650 Mitochondrial only 1125 479 241 (21.4%) 12,360 946 0.685 0.687 Chloroplast only 1764 819 660 (37.4%) 4 3756 0.319 0.521 4-Gene combined 3966 1497 1028 (25.9%) 4 5356 0.386 0.549 24 Taxa Nuclear only 2391 356 132 (5.5%) 9 672 0.641 0.516 Mitochondrial only 1052 318 130 (12.4%) 14,445 499 0.741 0.584 Chloroplast only 6619 1891 1277 (19.3%) 1 5359 0.477 0.422 9-Gene combined 10,062 2565 1539 (15.3%) 3 6586 0.509 0.430 modeled as a discrete gamma distribution (Yang, 1994). ML parameter values were then optimized, with a BIONJ tree as a starting point (Gascuel, 1997) with the appropriate parameters. Nodal robustness on the ML tree was estimated by the nonparametric bootstrap (1000 replicates). RESULTS FOUR-GENE ANALYSIS This data set included nuclear 26S (1077 sites), chloroplast psbA (1171 sites) and rps4 (593 sites), and mitochondrial nad5 (1125 sites) genes for 57 taxa. Of the total 1028 informative characters, 64.2% and 23.496 derived from the chloroplast and the mito- chondrial gene partitions, respectively. MP analysis resulted in four maximally parsimonious trees (MPTs) of length 5356. retention index (RI) values are consistently higher he consistency index (CI) and for mitochondrial genes than for the nuclear or chloroplast genes (Table 2). Topologies resolved by MP Leafy liverworts are recognized as monophyletic 9295"" ML analyses subdivided leafy liverworts into two major clades, bed I" (70%) and “Leafy II" (76%), but the “Leafy ” clade is not found in the strict consensus of four qe s. In all analyses, Trichocoleopsis and Neotricho- colea formed a clade (71%™'; 63%"”) sister to Ptilidium with strong support. The Ptilidium—(Tricho- and ML analyses are mostly congruent (Fig. 2). with strong support (97%™'; coleopsis—Neotrichocolea) clade was resolved in “Leafy Mu 7090"). II" and sister to the remaining taxa (76% NINE-GENE ANALYSIS This data set contained nuclear 268 (1061 sites) and 185 (1330 sites); rs SSU (892 sites), LSU (2162 sites), atpB (972 sites), rbcL (835 sites), rps4 (590 sites), and psbA (1168 sites); and mitochondrial nad5 (1052 sites) genes for 24 taxa. Of the total 1539 informative characters, 83.0% and 8.4% derive from the chloroplast and the mitochondrial gene partitions, respectively (Table 2). MP analysis recovered three PTs of length 6586. The tree was largely congruent with the MPTs (Fig. 3). Although taxon sampling was more limited compared with the four- gene data set, the results were similar to the four-gene analysis. Trichocoleopsis and Neotrichocolea formed a clade (819 *!; 84904”), sister to Ptilidium with strong support, but the Ptilidium-(Trichocoleopsis—-Neotricho- colea) clade was identified as sister to *Leafy P” in ML The Ptilidium—(Trichocoleopsis—Neotrichocolea) clade was analysis without significant support (50%). again resolved in “Leafy II" by MP analysis, also without significant support (55%). DISCUSSION TRICHOCOLEOPSIS AND NEOTRICHOCOLEA ARE NOT MEMBERS OF TRICHOCOLEACEAE AND LEPIDOLAENACEAE Some authors (e.g., Nakai, 1943; Müller, 1951- 1958; Schuster, 1957; Mizutani & Hattori, 1969; Mizutani, 1972) placed Trichocoleopsis and/or Neo- trichocolea in Trichocoleaceae, based mainly on the presence of a coelocaule. However, as pointed out by Schuster (1966, 1972), there are numerous characters suggesting the remoteness of the relationship between them (Table 3). For example, the former two gener have a coelocaule with simple paraphyllia (usually branched (succubous in Trichocoleaceae), in Trichocoleaceae), incubous leaves relatively reduced underleaves (similar to the lateral leaves in Tricho- coleaceae), freely developed rhizoids (usually lacking in Trichocoleaceae), water sacs (absent in Trichoco- leaceae), unelongated leaf cells, collenchyma, smooth cilia cells (striolate-papillose in Trichocoleaceae), and relatively large spores (40-55 um in Trichocoleopsis and Neotrichocolea but 13-20 um in Trichocolea- ceae). In fact, Asakawa et al. (1981b) hypothesized that Trichocoleopsis and Neotrichocolea had a distant 464 Annals of the Missouri Botanical Garden Acromastigum Bazzania Lepidoziaceae Lepidozia Telarane Chiloscyphus Geocalycaceae El Plagiochila Plagiochilaceae A gr ed Mm | e tedas Mastigophora Em 57 Lepi Lepicolea Td chochiesceds | Scapaniaceae | Jungermanniaceae | Cephaloziellaceae B Gymnomitriaceae Cephaloziaceae Leafy ll Adelanthaceae | Gymnomitriaceae Jungermanniaceae Antheliaceae C Calypogeiaceae Geocalycaceae | Balantiopsidaceae | Acrobolbaceae eS Schistochilac Neotr choculacen Lepidola e tilidia Jubulopsidaceae Lepidolaenaceae ebeliella Goebeliellaceae Ceratolejeunea 70 Lejeune Leafy | Cyclolejeunea Lejeuneaceae B teri Odontolejeunea A bul p Frullania Jubüjacpag 100 E Radula Radulaceae 100 Pallavicinia 5 E 0 Symphyogyna 88 Fo oe Outgroups Simple thalloid 100 Metzgeria 9 Riccardia Figure 2. ML tree inferred from the four-gene data set. Numbers above and below branches are ML a MP bootstrap s (> 50%). / — =) ^ D water Higher family names follow cal sacs, circumscription of the Leafy I and H liverworts pe clades A, relationship with Trichocolea Dumort., because their chemical components are quite different. In the present molecular analyses, Trichocoleaceae was resolved as a member of the strongly supported clade A in “Leafy IL" congruent with previous studies (e.g.. Davis, 2004; He-Nygrén et al., 2004, 2006; Heinrichs et al., 2005; Forrest et al., 2006). It is clear that the two endemic genera and Trichocoleaceae belong to two remote clades, respectively (Figs. 2, 3). Asterisks indicate the nodes not found in the strict consensus o X MPTs. Shading repre s those genera Crandall-Stotler m (2000). The beside the tree are based on Davis (2004). classification of and B, and C In modern classification of liverworts, the place- ment of Trichocoleopsis and Neotrichocolea in Lepi- dolaenaceae was also widely accepted (Schuster, 1972, 1979; Grolle, 1983: Furuki € Mizutani, 1994; al. 2006; Yamada & Iwatsuki, 2006). However, Lepidolaenaceae, together with Jubu- ade (1009 %; 96%"™") (Fig. 2), which was consistent with the results 2004; Cao et in this study, — opsidaceae, formed a well-supported c of previous molecular studies (Davis, He- Volume 95, Number 3 Liu et al. 4 2008 Phylogeny of Trichocoleopsis and Neotrichocolea 100 Pazzana | Lepidoziaceae 00 Lepidozia 94 Plagiochila Plagiochilaceae A 90 Trichocolea Trichocoleaceae 321, E Lepicolea Lepicoleaceae 100 Marsupella Gymnomitriaceae Leafy II ie 100 Calypogeia Calypogeiaceae C ut cl Cephaloziaceae 22 100 00 Odontoschisma B 100 Scapania Scapaniaceae Schistochila Schistochilaceae 100 81 Trichocoleopsis Lepidolaenaceae Pa DO [84 Neotrichocolea Neotrichocoleaceae 00 Ptilidium Ptilidiaceae ' a a Frullania Jubulaceae Leafy I DO} 92 Radula Radulaceae 37 [100 Lejeunea Lejeuneaceae = Lepidogyna Lepidolaenaceae 4 Pallavicinia 71 00 Symphyogyna a Fossombronia | outgroups Simple thalloid | 100 Riccardia 100 Metzgeria Figure 3. ML tree inferred from the nine-gene data set. iue above and below branches are ML and MP bootstrap percentages (= 50%). Asterisks indicate the nodes not found in the strict consensus of MPTs. Shading represents those genera having leaf water sacs. Higher family names follow the classifi ration of Crandall-Stotler and Stotler (2000). The circumscription of the Leafy I and II liverworts and clades A, B, and C beside the tree are based on Davis (2004). Nygrén et al., 2004, 2006; Heinrichs et al., 2005). In contrast, this structure originates from the ventral- the aligned matrix of the nad5 sequences, there are most lobe in Trichocoleopsis and Neotrichocolea. several gbique 1 insertions (“GCACGA” insertion at Additionally, Lepidolaenaceae is distributed in 740-745, "C. at 992-994, and “AAGA” or Gondwanian regions, whereas Trichocoleopsis and “AATA” at 1010-1013) shared by Lepidolaenaceae Neotrichocolea are endemic to eastern Asia. If the and Jubulopsidaceae, but they are not found in two genera were considered members of Lepidolae- Trichocoleopsis, Neotrichocolea, or other taxa, which ^ naceae, this disjunction would be so rare that Schuster provides additional evidence for the distant relation- (1983), under this assumption, once called Trichoco- ship between them. leopsis and Neotrichocolea “the most striking hepatic Initially, Trichocoleopsis and Neotrichocolea were endemics in eastern Asia." placed in Lepidolaenaceae (Schuster 1966, 1979), based mainly on the endosporic development of the — qyp RELATIONSHIP BETWEEN TRICHOCOLEOPSIS protonema and bearing water sacs. However, there are AND NEOTRICHOCOLEA remarkable differences between the two genera and Lepidolaenaceae. In Lepidolaenaceae, the interior Recent molecular studies all support that Neotri- cells of the capsule wall have l- or U-shaped chocolea is closely related to Ptilidium (Davis, 2004; secondary thickenings (Grolle, 1967), but they are — He-Nygrén et al., 2004, 2006; Heinrichs et al., 2005; extremely thin walled in Trichocoleopsis and Neotri- Forrest et al., 2006; Hentschel et al., 2006). Our chocolea. Additionally, water sacs occur both on results showed that Neotrichocolea had a closer lateral leaves and underleaves in Lepidolaenaceae, relationship with Trichocoleopsis than with Ptilidium, but only on lateral leaves in Trichocoleopsis and although the three genera formed a strongly supported Neotrichocolea. In Lepidolaenaceae, the water sac is monophyletic group (Figs. 2, 3). Inoue (1974) indi- derived from the middle lobe of the trifid leaf; by cated that Trichocoleopsis and Neotrichocolea were Table 3. (1974). N Comparison of morphological characters NA, not available in Trich ocoleopsis Neotrichocolea and other related genera as described in detail by Schuster (1966, 1972), Grolle (1967). and Inoue Character Trichocoleopsis Neotrichocolea Ptilidium Lepidolaena Dumort. Trichocolea Branching types Subfloral innovation Leaf insertion pattern Leaf and underleaf Water sacs Leaf cells and oil bodies Rhizoids Perianth Capsule Spore diameter Sporeling type l to 2 pinnately branched, Frullania type resent incubous lobed, margins ciliate, cilia cells smooth, underleaf smaller than lateral leaf irregularly polygonal, with small trigones, oil bodies granular botryoida present absent, having coelocaules apex rounded, epidermal ith secondary cies at corners, interior cells walled, e e 48-55 um endosporous, Trichocoleopsis 3 to 4 ene branched, Frullania present incubous lobed, both margin and dorsal surface of leaf ciliate, cilia cells smooth, underleaf smaller than lateral leaf present round rectangular, with weak trigones, oil bodies granular trvoida presen absent, having coelocaules apex rounded, epidermal cells with secondary thickenings at corners, inner cells thin walled, easily removed 40-48 um endosporous l to a pinnately branched, ıllania type absent transverse (upper end incubous) lobed, margins ciliate, cilia cells smooth, underleaf smaller than lateral leaf irregularly polygonal, with bulging trigones, oil bodies nearly homogeneous absent present apex beaked, epidermal cells with nodul ] inner cells thinner with ar thickenings, semicircular thickenings 25-27 um exosporous, Vardia type 2to4 scd branched, Frullania present incubous not lobed, margins without or with short cilia, cilia cells smooth, underleaf smaller than lateral leaf esent irregularly polygonal, with bulging trigones, oil bodies nearly homogeneous absent, having coelocaules apex beaked, T and inner cells sized, all have I- or i shape secondary thickenings 50—80 um endosporous, Frullania type l to 3 pinnately branched, Frullania type resent succubous strongly lobed, margins ciliate, cilia cells cuticle striolate-papillose, underleaf similar to lateral leaf absent rectangular, without obvious trigones, oil bodies nearly homogeneous absent absent, having coelocaules apex rounded, epidermal cells hyaline, without secondary thickenings, inner cells with distinct secondary thickenings 13-20 um exosporous, Nardia type USPIEL) jeoiuejog LNOSSIN 99r au} Jo s¡euuy Volume 95, Number 3 2008 Liu et al. 46 Phylogeny of Trichocoleopsis and Neotrichocolea closely allied to each other. Indeed, the two genera are similar in many morphological characters, such as the lack of a perianth, the presence of a paraphyllose, C ells 0 X the capsule wall being extremely thin walled and easily massive coelocaule, and the interior removed. Additionally, they all possess incubous leaves, with cilia on the lobed leaf margins, water sacs, and semblable oil bodies (Table 3). Moreover. a similar distribution pattern (Fig. 1) and habitat (Chen & Wu, 1965) are shared by the two genera. However, as described by Hattori (1947), differences to separate them at the generic level: cilia there are enough occur only on leaf margins in Trichocoleopsis but both on the margin and dorsal surface of leaves in Neotrichocolea; and water sacs are present on stem and branch leaves of Trichocoleopsis but only on branch leaves in Neotrichocolea. Moreover, the development and shape of water sacs are distinct in the two genera. Water sacs are widely observed in certain members of “Leafy |" (see shading in Figs. 2, 3) and were regarded as a device for facilitating water storage (Schuster, Frullania Raddi), horn-shaped (in Goebeliella Steph.), 1984). Their shapes can be cup-like (in or even specialized by having a trap-like structure in Colura (Dumort.) Dumort. and Pleurozia Dumort., which are considered t (Hess et al., dissection, saccate lobules are divided into two kinds: have a zoophagy function 2005). According to development in the “Lejeunea Lib. type” and “Frullania type” A Lejeunea-type water sac comes from an inrolling of the free margin of the lobule against the back side of the lobe and is formed from both lobe and lobule tissue. A Frullanta-type water sac, in contrast, involves the development of a secondary meristematic region in the midzone of the lobule. At an early stage, the lobule inrolls to the dorsal lobe. Afterward, continued divisions cause the upper part of the lobule to balloon (Crandall-Stotler & The opening of this kind of lobule or inflate down itself. 1980). faces away from the apex of the stem. upon Guerke, Trichocoleopsis and Neotrichocolea were described as 1972) and Frullania- Sacs, having Lejeunea-type (Shuster, 1905) Trichocoleopsis, the free margin of the lobule is deeply type (Evans, waler respectively. In inrolled, and it can be drawn open like a spring. This any Lejeuneaceae (such as Neurolejeunea (Spruce) Schiffn., found in -— type of deep inrolling is also Mastigolejeunea (Spruce) Schiffn., and Siphonolejeunea Herzog). As in other Lejeunea-type lobules, a Tric coleopsis lobule is inserted in a U line on the stem, so the whole leaf has a J-shaped insertion. In Neotricho- colea, the Frullania-type water sac is a slipper-like structure, opening parallel to the stem, but is not cup- like as in Frullania. lt is remarkable that genera with such different lobule development patterns are so closely related. Trichocoleopsis and Neotrichocolea do share unique lobule position: water sacs of both come from the ventral-most lobe of the trifid leaf. In contrast, [om other Lejeunea-type and Frullania-type water sacs are typically derived from the middle lobe of the trifid leaf. The third lobe can convert to a stylus, as in certain species of Frullania and Lepidolaenaceae, or be reduced to a slime papilla as in Lejeuneaceae and Jubula Dumort. (Schuster, 1972) AFFINITIES OF TRICHOCOLEOPSIS AND NEOTRICHOCOLEA In the pem nes dicia of the four- and nine- at E pa Naoirichocplea. In ns D there was an exclusive “CAT” insertion at the position of 593—595 shared by Neotrichocolea, Trichocoleopsis, and Ptilidium, which ment of nad5 sequences, further supported the close relationship between them. i special endosporic Trichocoleopsis was described as having a sporeling type (Trichocoleopsis 1956), exosporous Nardia Gray type of Ptilidium. However, type [Inoue, which is different from the Inoue (1963) described that in 7. sacculata, the early stage of development of the leafy shoot was essentially the same as that of Ptilidium, indicating a potential relationship between them. Additionally, a unique furanosesquiterpene, deoxopinguisone, and its deriv- atives were shared by Ptilidium and Trichocoleopsis 1981 a). Ptilidiaceae, once a portmanteau family of liver- (Asakawa et al., worts, was suggested to contain 15 genera by Evans (1939). Over subsequent years, most of these were from the family, until only Ptilidium remained (Schuster, 1979, 1984; Grolle, 1983; Crandall-Stotler & Stotler, 2000). Trichocoleop- s and Neotrichocolea were previously placed in Ptilidiaceae (Evans, 1939; Hattori, 1947; Schuster, 1966; Mizutani & Hattori, 1969). Some key differ- ences in morphology between the Trichocoleopsis— gradually removed uz ce Neotrichocolea group and Ptilidiaceae precipitated their removal (Table 3). For example, Ptilidium has a well-developed perianth, which was considered the most important difference between Ptilidium and the Trichocoleopsis—Neotrichocolea group, and also be- tween Trichocoleaceae and Lepidolaenaceae (Schus- ter, 1972; Inoue, 1974). Additionally, Trichocoleopsis and Neotrichocolea have water sacs on their leaves Ptilidium), subfloral rhizoids (absent in Ptilidium), (beaked in Ptilidium), different oil body types than Ptilidium (Table 3). (absent in innovations and rounded capsules and much larger spores, Phytogeographically, Trichocoleopsis and Neotricho- colea are endemic to East Asia, but Ptilidium has a 468 Annals of the Missouri Botanical Garden holarctic distribution pattern. Schuster (1983) pre- sumed that Trichocoleopsis, Neotrichocolea, and some other eastern Asian endemic liverwort genera, such as Vipponolejeunea Hatt.. Scaphophyllum Inoue, and Hattoria R. M. Schust., before and formed their modern distribution patterns might have had a wider range because of extinctions in other regions. For example, the two-species genus Nipponolejeunea is endemic to northeastern Asia, but another amber fossil species from the Oligocene was found in Europe (Grolle, 1981). Thus, excessive extinctions of taxa between Ptilidium and Such endemism seemed to be a relict. the Trichocoleopsis-Neotrichocolea group might have occurred, which could explain the morphological gap between them. Indeed, Trichocoleopsis and Neotricho- colea may retain many primitive characters compared to Ptilidium, waler sacs anc such as possessing endosporous protonemata. POSITION OF THE PTILIDIUM-(TRICHOCOLEOPSIS- NEOTRICHOCOLEA) CLADE Based on different sampling density (taxon and methods, previous molecular 2004; He-Nygrén et al., 2004, 2005; Forrest et al., 2006) have Neotrichocolea close to gene) and analysis studies (e.g.. Davis, 2006; Heinrichs et al., identified Ptilidium “Leafy I" or “Leafy IL," and/or or as a part of basal leafy polytomy. However, none of these studies obtained support for these placements. Recently, Hendry et al. (2007) reporte Herzogianthus R. M. Schust. the Prilidium—Neotrichocolea clade, but the position that the New Zealand endemic genus seemed lo be related to of this elade and the endemie genus within liverworts remained unresolved. According to our analyses, the Piilidium-(Trichocoleopsis-Neotrichocolea) clade was dicas with moderate support as a member of “Leafy " based on the ML and MP analyses of the four-gene is matrix, and was placed together with “Leafy I” in the ML significant analyses of the matrix without The Trichocoleopsis did not provide enough new informa- nine-gene support. addition of the genus o definitively resolve the tion position of the Ptilidium—(Trichocoleopsis—Neotrichocolea) clade. on This problem, and its implications for understand- ing morphological evolution in the leafy liverworts, have been discussed in detail in recent studies (see 2006). Both Heinrichs et al. (2005) and He-Nygrén et al. (2006) suggested that “Leafy 1” and the “Ptilidium clade” Porellales, Forrest et al., should be grouped in the order even though this relationship was not supported in their molecular analyses. Various shared morphological characters were cited to support this viewpoint, such as the mutual Frullania branching incubous leaves, type, lack of ventral branching, Yl 8 Ptilidium and some chemical components. endogenous spore germination (however, lacks this trait). — CONCLUSIONS Molecular and morphological evidence showed that Trichocoleopsis and Neotrichocolea represent a mono- phyletic group. The family Neotrichocoleaceae, which includes the two genera as suggested by Inoue (1974). should be recognized. The family is characterized by having incubous leaves with cilia on the lobed leaf margins, water sacs, a paraphyllose coelocaule, and extremely thin-walled interior capsule cells. Molecu- that allied to Ptilidiaceae, but except for the similar ciliate lar results indicated Neotrichocoleaceae was and lobed leaves, there are almost no known synapomorphies shared by the two families. On the o contrary, Neotrichocoleaceae differ from Ptilidiaceae in many characters. Thus, merging the three genera into one family is premature based on the current evidence. The exact position of the Neotrichocolea- ceae—Ptilidiaceae clade remains unresolved in the present study. A better sampling of both sequence data and taxa is required before the problem can be resolved. Literature. Cited Asakawa, Y., R. Matsuda & C. 1981a. Pinguisane-Ly pe sesquiterpenes from Ptilidium pulcherrimum. Phytochem- istry 20: 1427-1428. Toyota, T. Takemoto & R. Mues. of the live Mes hocolea and Trichocoleopsis. — r 0. Suire. 1981b. 'rworts in the foe atic esters and terpenoids genera "Trick 10€ colea, Phytoc pad 20: Boisselier-Dubay ) Lambourdiére & H. Bischler. 2002. Molec kp "leone nies aun s Mes ical reductions in the liverwort sube archantiidae . Molec. a ogen. a vol. 24: : vt l. . B. & X. Li. report of "i fist national red list m 2 se 205 . Guo, C. Gao, P. Wu edu angered bryophytes. J. Hattori Bot. Lab. 99: : Capesius, L & Bopp. 1997. New rm of liverworts 2 on molecular and morphological data. PL. Syst. Evol. 207: 87— 97. Chen, P. C. & P. a = u. . The pre ‘liminary es of the p. 1-59 1 Hw aa Hsu orulam la ‘am. Shanghai Sones: and Technology Press, Shangha ; V. R. Guerke. o ntal f Jubula Dum. (Hepaticae). Bryologist 83: 179-201. « R. Stotler. 2000. Morphology and classification of the Marchantiophyta. Pp. 21-70 in A. J. Shaw € B. Pound (editors), Bryoy nem Biology. Cambridge Univer- sity Press ge. dales B. J, L. L. Forrest & R. E. Stotler. 2005. Evolutionary. trends in simple thalloid liverworts (Mar- chantiophyta, Jungermanniopsida subclass Metzgeriidae). Taxon 54: 299-316, Volume 95, Number 3 2008 Liu et al. 46 Phylogeny of Trichocoleopsis and Neotrichocolea Davis, E. C. 2004. A molecular phylogeny of bo liverworts (Jungermanniidae: Marchantiophyta). Pp. 61-86 in B. Goffinet, V. C. Hollowell & R. Magill Pee Molecular se of Bryophytes. Monogr. Syst. Bot. Missouri . Gard. 98. Evans, A. i 1905. A remarkable Ptilidium from Japan. Rev. Bryol. 32: 57-60 2e 39. The lite of the Hepaticae. Bot. Rev. 5: 49-96. Felsenstein, J. 1985. Confidence limits of phylogenies: An : 783-191, . Crandall-Stotler. 2004. A phylogeny of the simple iballoid liverworts (Jungermanniopsida, Metz- DN as inferred from five chloroplast genes. Pp. n B. Goffinet, V. C. Hollowell & R ut Mie ar E of Bryophytes. Monogr. Syst. Bot. Missouri Bot. Gard. 005. Progress wee ny for the liverworts, with particular focus on the ud ipu J. Hattori Bot. "E 97: 127-159. avis, D. G. E B. J. Crandall-Stotler, A Clack r "n L. Hollingsworth. 2006. Unraveling the evolutionary history of the liverworts Vea ta): ultiple taxa, genomes and analyses. Bryologist 109: -334 Magill towards a robust — Frey, W. & M. Stech. 2005. A morpho-molecular classifica- tion of the liverworts (Hepaticophytina, Bryophyta). Nova ae 81: 55-78. Furuki, & M. Mizutani. 1994. Checklist of Japanese aic and Anthocerotae, 1993. Proc. Bryol. Soc. Japan 6: Gao, C. & T. Cao E 2000. Flora Yunnanica, Vol. 17. Science Press, Beijing. Gao, C. H. & T. Cao. 1988. A study of Trichocolea Dum A e and e Hatt. (Hepasi- ca China. Invest. Stud. Nat. 8: 24— Gasc ui O. 1997. BIONJ: An pid version of the NJ algorithm based on a simple model of sequence data. Molec. Biol. Evol. 14: 685—695. Grolle, R. e a der Lepidolaenaceae. J. Hatt. Bot. La J^ 1981 d ees fossil in Europa. J. Hatt. Bot. Lab. 50: 143-157. 8 omina generica Hepaticarum; references, ypes and Viu nae Acta Bot. Fenn. 121: 1-62. Guindon, S. & uel. 2003. A simple, f algorithm to estimate ost duce i maximum 1e E p 52: 696—704. Hall, T. 999, te iP user-friendly ence > age editor and analysis s for Windows 95/98/NT. Nucl. Acids Symp. Ser. 41: 9 and accurate biological Hattori, S. 1947. Contributio ad floram om ES as II. att. Bot. Lab. 2: 1-26 He- Nygrén, X., Ahan nen, A. Juslén, D. Glenny & S. Piippo. 2004. o. of the liverworts—Beyond a leaf and a thallus. P 118 in B. Goffinet, V. C. Hollowell & R. Magill (editors), Molecul ar Systematics of Bryophytes. Monogr. Syst. Bot. Missouri Bot. . 98. , A. Juslén, I. Ahonen, D. Clenny & S. Piippo. 2006. Illuminating the evolutionary history of liverworts (Mar- chantiophyta)—Towards a natural classification, Cladis- l. a tics 22: 1— o » S. R. Gradstein, R. Wilson & H. Schneider. 2005. A a natural classification of liverworts (Marc htop) 1 on the chloroplast gene rbcL. Cryptog. Bryol. 26: -150. Hendry, T. A., B. Wang, : Yang, E. C. Davis, J. E. Braggins, R. M Qiu. 2007. de ieu phyloge- New Zealand, Goebelobryum > liverworts from Jackiella unguiculatum, and Herzogianthus vaginatus, chloroplast genes. Bryologist 110: 738-751. Hentschel, E R. Wilson, M. Burghardt, H. J. Zündorf, H. Schneider & J. Heinrichs. 2006. Reinstatement of m 'oleaceae (Junge rmanniopsida) based on chloroplast rbcL data: Exploring the importance of female cs of Jungermanniales. Pl. curvata, using three hee for the systematic . Evol. 258: 211-226. S, E P. Frahm & I. Theisena. zoophag liverwort ps rea. Blog 108: 212-21 Hong, . The leafy sta ae of South Korea and dh litori om especially to the flora 26. Evidence of 2005. species, Hess, second Pleurozia of North America. Bryologist 69: Horikawa, Y. 1934. Monographia Japonicarum. J. Sci. Hiroshima Univ., 2: 101-325 Inoue, H. 1956. Studies on spore germination of Hepaticae 1. Trichocoleopsis sacculata (Mitt.) Okam. J. Hatt. Bot. Lab. 55-58. AAR Australi- Ser. B, Div. 2, Bot. 1963. Young stages of the development of ==. sacculata. J. Jap. Bot. 38: 250-251. Illustrations of Japanese Hepaticae. Tsukiji Shokan tr Pu Tokyo. À., Konstantinova, N. A., A. D. Potemkin & R. N. Schljakov. 1992. Check- ve of a a ae and Anthocerotae of the De USSR. Arctoa 1: 87-127. Koponen, T., T. ` ao, S. Huttunen, A. Juslén, C. Peng, S Piippo, P. Rao, J. Váňa & V. Virtanen. 2004. Bryophyte flora of Hunan Province e, China. 3. Bryophytes from Taoyuandong and Y Badagongshan and i e Reserves, with additions to floras of Mangshan Nature Reserve and W deca uan Global Cultural Heritage Area. Acta Bot. Fenn. 177: 1-47. Lewis, L. A, B. D. Mishler € Vilgalys. 1997. Phylogenetic relationships of the liverworts (Hepaticae), ophyte lineage, inferred from nucleotide e ee a oroplast gene rbcL. Molec 7-393. asal embryc sequence data Phylogenet Evol, Liu, Y., Y. Jia, W . L. Qiu. 2005. A taxonom eassessment of o inferred from alee lar and morphological evidence. Bryologist 108: Mitten, w. 1891. On the species of Musci and Hepaticae recorded from Japan. Trans. Linn. Soc. London ser. 2, 3: 153-206. Mizutani, M. 1972. ios ae (Hepaticopsida). in Z. Iwatsuki M. Mizutani (editors), Colou Illustrations of Bryophytes of Japan. Hoikusha Publishing sa ES. Hawi 1969. Check list of Japanese Hepaticae and Cac t us. Bryol. Lichenol. 5: 33-43. Pp. 267—369 ured Müller, K. 1951- . Die Lebermoose Europas. /n Ra- ens 8 Ko pane Flora, 3rd ed 6 : 1-1365. Nakai, T. )rdine genera, sectiones pe emalis Tokyoensis a annos -1926-1 941 Tokyo us 5. 1911. Neue Beitráge zur Moosflora Japans. Bot. g. (Tokyo) 25: 159. 162. Annals of the 470 Missouri Botanical Garden u, . L. Li, B. Wang, Z. Chen, V. Snoop, M. Groth- Ee F. 1897. Hepatica Japonica. Bull. Herb. Boissier 5: ns O. Dono J. Lee. L. Ker est, G. F. 8. . Hendry, D. W. Taylor, C M. Testa, M. Son D. L. 2003. PAUP*: Phylogenetie Analysis Using Parsimony, Vers. 4.0b 10. Sinauer Associates, Sunderland, ¿stabrook, T. : Ambros, B. eum Stotler, R. J. Duff, M. Stech, W. Frey, C. Davis. 2006. The deepest divergences tz D. Quandt & C. in land plants inferre dl D phylogenomic evidence. Proc Natl. Acad. Sci. U.S.A. 103: 15511-15516. Schuster, R. M. 1957. Notes on o Hepaticae. XV. He iiir jar Bry ol. Liché nol. 26: 123-145 ran Anthocerotae of North Eon studies on 21-405. Pp. 41-82 in Bryophyte ' Phyloge ne E and e J. Hatt. Bot. Lab. : . 1979. The 2a 25 of the Bs aticae. G. C. S. Clarke € C. Duckett (editors), spo Academic Nile New York. 3. Phytogeography of the Bryophyta. Pp. 463— 626 in t M. Schuster an New Manual of Bryology. Vol. 1 e Botanical Laboratory, Nichina 984. Evolution, Ed and classifica ation of the He ms ae. Pp. 892-1071 in R. M. Schuster (editor), New Manual of ossa Y 2: Hattori B Laboratory, Nic hinan. Stech, Frahm, H. 2000. Molecular Pci of Dd Goebel (1 abla 'eae, ilger & W. Frey Treubiopsida) and high taxonomic level classification of the Hepaticophytina. Studies in austral A rain forest bryophytes 6. Nova Hedwigia 71: 195-208 Tromp ni, J. D., Wheeler, JA Massach uselts. T. J. Gibson, F. Plewniak, F. ne G. Higgins. 1997. The Clustal X windows inter Flexible strategies for M sequence alignment Acids Res. Xm aided by quality analysis tools. Nucl. 4876—4882. 2000. Molecular phylogenetic reconstructions of the marchantioid liverwort radiation. Bryologist 103: 314-333. Wu, P. C. genera of 99-117. genera and endemic Asiatic Bryobrothera 1: China. 1992. The east the bryophytes in — -. : M. Z. Wang. 1997. Conservation biology of rare and re red = of ea s in China. Pp. 37— 41 in P. C. Wu, B. Tan & Y. we (editors), Chenia, 3-4. Énculedss a ublisbing Hou Gan K. € Z. Iwatsuki. s alo of the Hepatics of Japan. J. Hattori Bot. Lab. € Yang, Z. 1994. Maximum like diet po netic e stimation ^ rales over 38: , Beijing. < from DNA sequences Re methods. J. Mol. Evol. 39: 306-314. Zhang 1982. Eine Neue Trichocoleopsis in Qin Ling 12. "ee Acta Bot. Yunnanica 4: 171- A TAXONOMIC REVISION OF THE Anthony R. Magee,?* Ben-Erik van Wyk,” SOUTH AFRICAN ENDEMIC GENUS ARCTOPUS (APIACEAE, SANICULOIDEAE)! Patricia M. Tilney,? and Michelle van der Bank? ABSTRACT The genus Arctopus L. is revised. It is an anomalous member of the family Apiaceae, with several unusual morphological characters such as the prostrate spiny leaves, thick resinous tuberous roots, sessile female perianth, and pseudanthia formed by distinctive brac reproductive morphologies, being easily Rs ud by the large involucral bracteole pseudanthia. The anatomy of the petiole as be useful diagnostic characters and are etd and discuss . A formal taxonomic treatment is presented, including a key to the specie quence dat ce both morphological and DNA se correct nomenclature, ae desc ie illustrations, < ITS, IUCN Red List, Key words: Apiaceae, Arctopus, phylogeny, Saniculoideae, flowers with a homochlamydeous ‘teoles. The three species of Arctopus were found to differ mainly in their s that surround the female as the morphology of the inflexed spines and leaf margins were also found to ed. Possible phylogenetic selationsbipe were explored E and distribution ma ps- taxonomy. Arctopus L. is endemic to the Cape Floristic Region (CFR) of South Africa and consists of three dioecious e important medicinal plants with a 2007). Arctopus was described by Linnaeus (1753) in Species Plantarum, from figures by Plukenet (1691-1705: tab. 271, fig. 5) and Burman (1738: tab. 1). The name foot," species. They ar rich ethnobotanical history (Magee et al., Arctopus is derived from Greek, meaning “bear's due to the broad simple leaves that remain appressed to the ground and are armed with large inflexed spines. The genus was originally thought to be monotypic, consisting only of A. echinatus L., whose virtue as an important medicinal plant was well known to the Khoisan people (Pappe, 1857). It also became popular among the early Cape Dutch settlers and ecame known by the vernacular names “sieketroost” or “platdoring” (Smith, 1966). Two additional species, A. dregei Sond. and A. Sond., were later described by Sonder (1862) in Flora Capensis. Wolff last revised the genus in 1913 but monacanthus Carmich. ex reported male material of A. monacanthus as un- Because the based solely on emi PESE, known. ecles circumscriptions were the identification of male plants has been difficult, resulting in much confusion. This paper is aimed at presenting a complete revision of the genus, including the circumscription, nomenclature, typification, formal descriptions, and illustrations of diagnostic. characters for the three species that are recognized. A key to aid identification of female and, for the first time, male non-flowering material is provided. The known distribution of each species 1s recorded. Morphological, anatomical, and sequence data (ITS) from Arctopus are also evaluated in order to explore possible phylogenetic relationships at both generic and species level. TAXONOMIC POSITION Arctopus is an anomalous genus of geophytes belonging to the family Apiaceae. There have been differences of opinion as to the correct taxonomic position of the genus within the family. Drude (1897— 1898) divided Apiaceae into the three subfamilies Apiodeae, Saniculoideae, and Hydrocotyloideae. He placed Arctopus in the subfamily Saniculoideae based on the presence of toothed leaves, surface outgrowths on the mericarps, and intrajugal oil ducts. In Wolff's (1913) Arctopus was also placed in Saniculoideae. both Froebe (1964) and Magin (1980), in their studies of the inflorescence treatment, However, structure of the Saniculoideae, found that Arctopus shares many characteristics with the Hydrocotyloi- deae. Pimenov and Leonov (1993) listed it in the ! We thank the directors and staff of the ue herbaria for kindly making specimens available: BM, BOL, K, NBG, PRE S, and SAM. We are es specially of South Africa is gratefully acknowledgec y grateful to L. Agenbach, N assessment for Arctopus dreget. Financial n. from the University of Johannesburg and the National Research Founc dged. Raimondo for providing the IUCN Red bs ation elme, and D. = * Department of Botany and Plant Biotechnology, University of Johannesburg, Kingsway Campus, P.O. Box 524, Auckland Park 2006, Johannesburg, South Africa. * Corresponding author: a.r.magee@hotmail.com. 4 doi: 10.3417/200517 ANN. Missouni Bor. Garp. 95: 471-486. PUBLISHED ON 23 SEPTEMBER 2008. 472 Annals of the are Botanical Garden Table l. l, lamina; md, midrib; mfr, mature fruit: p. petiole. Material of Arctopus species used for anatomical studies (all material kept at JRAU). f, flower: ifr. immature fruit: Taxon Voucher specimens Locality Plant parts studied A. dregei Magee & Boatwright 2 Magee & Boatwright 5 Magee & Boatwright 31 Magee & Boatwright 4 => . echinatus Magee & Boatwright 6 Van Wyk & Viljoen 3715 Van Wyk 4128a Van Wyk s.n., 14 Oct. Winter et al. 170 Magee & M d 34 Van Wyk 3 Van Wyk f Van Wyk 4141a Van Wyk 4161a Van Wyk s.n., 12 Van Wyk, 1993 = . monacanthus July 1997 Winter & Tilney s.n.. 9 Oct. Rondeberg p. |, md, mfr Darling p. md, f, mfr Malmesbury p. mfr Rondeberg p. md Yzerfontein p. l, md, mfr Oorlogskloof p. l, md Nieuwoudtville md, f Arendskraal p, mfr Du Toitskloof f Pakhuis Pass p. mfr Piketberg p. ifi Nieuwoudtville md Elandskloof p, l, md, f, mfi Gifberg p. md Jonaskop p. md 1993 McGregor Rd. p subfamily Hydrocotyloideae but were uncertain about under incertae sedis). However, recent molecular data (Plunkett & Lowry, 2001; Chandler & Plunkett, 2004; Plunkett et al., 2004) show Hydrocotyloideae to be polyphyletic. A the tribal affinity (listed new subdivision of the family into Apioideae, Saniculoideae, Azorelloideae, and Mackinlayoideae 2004). in which a close relationship between Arctopus and the 5 (2003). in structure of has been proposed (Plunkett et al., Sanicu- their loideae is suggested. Liu et al. evaluation of the fruit Saniculoideae, proposed a broader saniculoid concept that includes Arctopus and some other anomalous African genera (viz. Lichtensteinia Cham. & Schltdl., Polemanniopsis B. L. Burtt, and Steganotaenia Hochst.). Calviño and Downie (2007), using chloroplast sequence data (trnQ-trnK 5': tion of the subfamily, proposed two tribes within exon region) to assess the circumserip- Saniculoideae, namely a redefined Saniculeae and a newly deseribed Steganotaenieae (comprising Polem- anniopsis and Steganotaenia). In this analysis, A. echinatus is placed within Saniculeae as a sister group to Eryngium L., Sanicula L., Hacquetia Neck. ex DC., Petagnaea T. Caruel, Astrantia L., and Actinolema Fenzl, with Alepidea D. Delaroche successively sister and the most basally diverging lineage within the tribe. MATERIALS AND METHODS The three species were studied in situ and observed throughout much of their known distributional range. BOL, JRAU, K, NBG, PRE, S, and SAM was studied (abbreviations follow Herbarium material from BM, Anginon Ral.). 1990). From this material, with geographical information from Leistner and Morris (1976), all the species verified and mapped. The Holmgren et al., together the recorded distributions of were carefully conservalion status of each species was assessed using IUCN Red List criteria (IUCN, 2001). For anatomical procedures, leaf (petiole, lamina, and midrib) and fruit material were fixed in formalin, acetic acid, and alcohol (FAA) for at least 24 hr. and then treated according to the method of Feder and O’Brien (1968) for embedding in glycol methacrylate (GMA), but modified so that the material in the first two changes of GMA was infiltrated for a minimum of 24 hr., followed by a final infiltration of five days. Staining was done according to the periodic acid Schiff/toluidine blue (PAS/TB) staining method (Feder & O'Brien, 1968). A list of voucher specimens for the Table 1. were made with the aid o anatomical study is given in Drawings, all — done by the first author, a camera lucida attachment on a Zeiss (Göttingen, Germany) compound microscope. Total DNA was extracted from 0.5-1.0 g of fresh leaf material following the CTAB method of Doyle and Doyle (1987). The concentrated extracts were cleaned and QlAquick silica columns according to the manufacturer's protocol (Qiagen Inc., Hilden, Germany). Genera of the Saniculoideae belonging to both tribes Saniculeae (Alepidea, Astrantia, Erygium, Hacquetia, Petagnaea, and Sanicula) and Steganotae- nieae (Steganotaenia) were included in the analysis. genera of the & Sehltdl. Thunb. together with two representative Apioideae (Heteromorpha Cham. and Hermas villosa was selected as the outgroup based on the sister relationship Volume 95, Number 3 2008 Magee et al. Revision of Arctopus (Apiaceae) Table 2 Data matrix of character states in the genus Arctopus, used to construct the cladogram in Figure 4. For details of the characters and polarization of character states, see Appendix 2. Characters and character states Taxa 1 Alepidea amatymbica Arctopus echinatus l2—c ao E A 0 Arctopus dregei 0 1 1 Arctopus monacanthus 0 0 0 0 0 0 0 0 0 0 0 0 | 0 0 1 1 1 0 0 0 0 l l 0 0 ] l 1 l l l 0 0 1 the Hermas L. clade and the Apioideae and Saniculoideae found by Calviño et al. (2006). sources of plant material, between Voucher specimen information (including GenBank accession num- bers, species names, and their author citations) is listed in Appendix 1. The ITS of nuclear ribosomal DNA was amplified using the Sun et al. (1994) ABIOI (5'-ACG AAT TCA TGG TCC GGT GAA GTG TT-3’) and ABIO2 (5'- TAG AAT TCC CCG GTT CGC TCG CCG TT-3’) primers. Amplified polymerase chain reaction (PCR PCR according to the — products were purified using a QlAquick (Qiagen Inc.) manufacturer's instructions and directly sequenced purification. kit on a 3130 x/ Genetic Analyzer (Applied Biosystems City, California, U.S.A.) using BigDye Terminator version 3.1 (Applied Biosystems Inc.). Inc., Foster Complimentary strands were assembled and edited using Sequencher version 3.1.2 (Gene Codes Corpo- ration, Ann Arbor, Michigan, U.S.A.), and aligned manually in PAUP* version 4.0b1 (Swofford, 2002), with gaps positioned so as to minimize nucleotide mismatches. Characters and character states used for the cladistic analysis of morphological data are given in Table 2 and Appendix 2. Character states were polarized using the method of outgroup comparison, with Alepidea amatymbica Eckl. Zeyh. as the outgroup. Where both states were found to co-occur in a single taxon, it was coded for the plesiomorphic state. Phylogenetic analyses were conducted using the Mor- phological characters were treated as ordered and 1970), whereas for the molecular data, all character trans- likely (Fitch Tree searches were per- parsimony algorithm of PAUP* version 4.0b1. equally weighted (Wagner parsimony; Farris, formations were treated as equally Fitch, 1971). formed using a heuristic search with 1000 random parsimony; sequence additions, tree — bisection-reconnection (TBR) branch swapping, and the MULPARS option in effect. A limit of 10 trees per replicate was set to reduce the time spent on swapping in each replicate. Internal support was assessed with 1000 bootstrap C ade of replicates (Felsenstein, 1985) using TBR swapping and holding 10 trees per replicate. Only values greater than 50% are reported, and the following scale was applied for support percentages: 50%-—74%, weak: 15%-84%, moderate; and 85%-—100%, strong. Bayesian. analysis was performed on the ITS sequence data, using MRBAYES v. 2.01 (Huelsen- »ck € Ronquist, 2001; Ronquist & Huelsenbeck, 2003). The TrN+G model, indicated to be the best model by MODELTEST v. 3.06 (Posada & Crandall, 1998), was implemented. The analysis was performed with one million generations of Markov chain Monte Carlo and a sampling frequency of 100. The resulting trees were plotted against their likelihoods in order to determine where the likelihoods converge on a maximum value. All the trees before this convergence were discarded as the “burn-in” phase. The remaining 4.0b1, majority rule consensus tree was produced in order to trees were imported into PAUP* y. and show the posterior PM (PP) of all observed bi-partitions (only PPs above 0.5 are reported on the tree in Fig. 4). The following scale was used to evaluate the PPs: 0.50—0.84, low; 0.85—0.94, moder- ate; 0.95—1.0, strong. MORPHOLOGICAL AND TAXONOMIC CHARACTERS HABIT Arctopus is unlike any other genus found within the Apiaceae. The species are summer deciduous geo- phytes not exceeding 15 em in height when in flower. Their broad simple leaves, which arise directly from a tuberous rootstock, are spiny and sprawl out in a rosette, remaining appressed to the ground. The plants are distinctly dioecious, and each sex is immediately apparent when in flower or fruit. The three species show no obvious differences in growth form. LEAF SHAPE The leaves of Arctopus species are simple and vary from ovate to semi-orbicular or B 7A-F, 940; greatly in shape, rhomboidal (Figs. 5A, with only Annals of the Missouri Botanical Garden those coastal populations of A. echinatus that occur between Knysna and Hermanus bearing leaves that are pinnatipartite to pinnatifid (Fig. 7D, E). In all the are incised into three segments, A. monacanthus and A. echinatus, the lobes are promi- nently toothed, giving them a very distinctly dentate appearance (Figs. 7A, C-F, 9A—C), in A, dregei, the teeth are inconspicuous and thus the lobes species, the leaves each of these being two- or three-lobed. In whereas appear crenate (Fig. 5A, B) LEAF VESTITURE The teeth on the leaf margin terminate in either flexuose setae (the presumed plesiomorphic state found in Alepidea) as in Arctopus dregel, or spinose selae as in A. echinatus and A. monacanthus. Such marginal hairs or setae are a characteristic feature of nany genera belonging to the subfamily Saniculoideae (van Wyk & Tilney, 2004). Within Arctopus, leaf vestiture was found to be a useful diagnostic character. Vertical spines are usually present in the leaf In A. these spines are attached to an inflexec recesses between the segments. mona- om canthus, laminar tooth (Fig. 9D—I), which is often conspicu- ously broad with marginal setae almost invariably in A. echinatus, the inflexed present. However, laminar tooth is either absent or very narrow and almost always without marginal setae (Fig. 7L-N). Inflexed laminar teeth and vertical spines are very rarely present in A. dregei (if present, then small, attached tọ a narrow laminar tooth, without setae). The leaves of A. markedly more spinescent than those of A. echinatus, sparse, and monacanthus are with up to six stellate spines in each leaf recess being common. In A. echinatus, the spines are often solitary or in threes; however, populations from the Cederberg in the north and around Port Elizabeth in the Eastern Cape have up to four or rarely five stellate spines per recess. LEAF ANATOMY While the three species are very similar in their a vegetalive anatomy, certain characters (as discussec value, later) have been found to be of diagnostic a particularly in the field where they are easily observec with a 10X hand Arctopus echinatus and A. monacanthus, which are if lens. This is especially true of often difficult to identify when female flowering or fruiting material is not available. Arctopus dregei a shows much variability in its vegetative anatomy anc often shares similar character states to A. echinatus and A. monacanthus. Because the macromorphology of this species is relatively distinctive, it is not easily confused, even when flowering or fruiting material is of characters to distinguish it is therefore superfluous. unavailable. The use anatomical Transverse sections of the lamina were found to be markedly similar in all three species and were therefore of little diagnostic value. The mesophyll is distinctly differentiated into palisade and spongy parenchyma, with the palisade parenchyma compris- ing about half of the mesophyll thickness in Arctopus echinatus and A. monacanthus, and less than half in A. dregei. The midrib of the lower third of the leaf, when studied in transverse section, was found to be far more useful (Fig. IA-F). The extent of sub-epidermal collenchyma and the arrangement of the vascular bundles are diagnostic. Sub-epidermal collenchyma is found on both the adaxial and abaxial surfaces. On the abaxial surface, the collenchyma extends at least halfway along the midrib in Arctopus monacanthus (Fig. 1E, F) and, extent, dregei Fig. 1A, B) and A. echinatus (Fig. 1C, D). At least two to a lesser in a “ope large, oppositely arranged vascular bundles can B Pl À 8 usually be found. Of these, the adaxial bundle is solitary (Fig. 1A-C) or further split dorsiventrally Fig. 1B—D), that is, into a dorsal and a ventral bundle in A, echinatus and in A. dregei. In A. monacanthus, bundle is always collaterally split — the adaxial (Fig. LE, F), that is, into two lateral bundles, although this has also been found to occur occasionally in A. dregei (Fig. 1B). One or two additional smaller adaxial bundles are often found in the midribs of larger leaves in all three species. The transverse sections of the petiole are similarly The petioles are adaxially ~ liagnostic (Fig. 16-1). flattened and prominently. raised abaxially. They extend laterally into wings, which are always prominently (Fig. 1G). Sub-epidermal collenchyma is present along the abaxial surface in all three species. In A. echinatus (Fig. 1H) and A. monacanthus (Fig. 11), collenchyma ut in A, monacanthus, it is ribbed in Arctopus dregei — is also found in the wings, continuous, or almost so, with the abaxial collenchyma strand. Large cavities between the vascular bundles can be found in A. echinatus (Fig. 1H) and in the late season growth of A. dregei, but never in monacanth us. FEMALE INFLORESCENCE The bracteoles of the involucel are of diagnostic value and render a female flowering plant quite unmistakable. They are prominent and accrescent in all three species, giving the umbellule the appearance of a single flower, called a pseudanthium (Figs. 5C—F, 7O-R, 9J-M). This is commonly found in other Volume 95, Number 3 2008 Magee et al. 475 Revision of Arctopus (Apiaceae) | im Transverse sections through the midribs (A-F) and petioles (G-I) of Arctopus species to illustrate diagnostic Pa l. anatomical differences. In the midribs, the a Ron collaterally split in A. monacanthus (E, F T (C) are characteristic. Cavities a th Viljoen 3715, JRAL A. monacanthus a id SM, eaten JR s.n., Arendskraal, JRAU). daxial vascular bundle can be solitary (C and a combination of all o s of A. echinatus (H) and A. monacanthus (I), and the markedly ribbed wings in A. dregei e vascular bundles of h —A. A. dregei ape & pne 2, JRAU). —B. A. dregei ede & ae 5, JR natus E & Boatwright 6, JRAU). —E. . AU). —G. A. dregei (Magee & pes right 2, JR —]. A. monacanthus (Van Wyk 3594, JRAU). c ) or dorsiventrally split (D) in A. f these in A. dregei (A, B). In the petioles, the ealthy, mature leaves are ie nt only in A. echinatus (H). A, rue (Van Wyk & A. amicus van Wyk 4141a, JRAU). —F. . —H. A. oe (Van Wyk c, collaterally split GE bundle; cav, cavity; col, collenchyma; dv, dorsiventrally split adaxial bundle; s, solitarily adaxial bundle. members of the Saniculoideae such as Alepidea, Astrantia, and Eryngium (Burtt, 1991) In Arctopus, the pseudanthial bracteoles are fused to the base of the fruit and, as in Alepidea, to one another. The bracteoles of A. dregei (Fig. 5F) are obovate and involute, with an obtuse spine-tipped apex and one or two inflexed, spinescent hairs along M), they become very large, widely obovate and foliose, with an the margin. In A. monacanthus (Fig. 9L, obtuse, spine-tipped apex. Their foliose appearance is enhanced by the prominent reticulate venation. The margin is usually entire, but one to three marginal spines may rarely occur. In A. echinatus (Fig. 7R), the bracteoles are ovate, boat-shaped and keeled, with an acute spine-tipped apex plus one or two inflexed and overlapping spines along the margin. They are markedly thickened along the margin and midrib. The pseudanthium of A. dregei consists of three to six bracteoles that become white when dry, whereas A. echinatus and A. monacanthus have either four or five dry. In monacanthus, the pseudanthium is brittle and papery at the fruiting stage and tends to break apart easily, so that the bracteoles separate from one another with the ripe fruit still attached to them. In A. dregei and A. echinatus, the pseudanthium becomes leathery at the bracteoles that become brown when fruiting stage and does not break apart easily, so that intact pseudanthia are found with the dying leaves at the end of the growing season (November—December). FLOWERS The male flowers are pentamerous with five large petaloid sepals (Figs. 5L, 7H, 9Q) that are linear to oblong and about the same size as the petals. According to Drude (1897-1898), large sepals are Sanicula, commonly found in Saniculoideae (i.e., Hacquetia, Actinolema, and Eryngium). There are five oblong to 476 Annals of the Missouri Botanical Garden Figure zm ni el C 1mm Transverse sections through the po d monoc ed DUE Tun of Arctopus species. n all three species, both ).T ifi y). The nnn re M ap o mesode Y bur ne y as E J I Li | id i c ontinuens between n both mericar (A A. monacanthus (C \. dre p ae, E Pom right 2 monacanthus (Van Wok p dla, JR AU). « oblanceolate petals with inflexed tips (Figs. 5M, 7G, OR, S). On the inner surface, a thin septum joins the he inflexed petal, basal and apical portions of preventing the petals from straightening. The stamens are twice as long as the petals, and the filaments are inflexed at their tips. Although the ovary is absent, a flattened, Saniculoideae) is present along with reduced styles. papillate stylopodium (common to many The female flowers are sessile (as in Alepidea) and tubular (Figs. 5G, 7U, 90), with a homochlamydeous perianth (a character unique to this genus within the family) consisting of 10 persistent triangular lobes (Figs. 5J, 7V, 9N). Due to the typically pentamerous nature of the male flowers and the almost petaloid sepals, these perianth lobes are here ees as five sepals and five sepaloid petals in the female flower, but further ontogenetic studies are required to = resolve the homology of these peculiar structures. Stamens are absent and the stylopodium is papillate, cushion-shaped, and swollen so that it almost completely envelops the two style bases. The styles are conical in shape and longer than the sepals. The ovary is bilocular. FRUIT MORPHOLOGY AND ANATOMY The fruit of Arctopus are unusual in Apiaceae because they become pseudo-monocarpellate with the abortion of one of the two mericarps early in fruit development (Fig. 2). However, in some populations of A. Hottentots Holland Mountains, both mericarps may develop. In A. echinatus and A. monacanthus, the fruit echinatus around Vanrhynsdorp and in the are rostrate (Figs. 75, 9L, M) and brownish, whereas in A. dregei they are white and unbeaked (Fig. 51). The fruit surface is spinescent in A. echinatus and A. monacanthus, while in A. dregei it has wart-like protuberances. Surface outgrowths on fruit are a common feature in the Saniculoideae (Drude, 1897— 1898). In mature fruit, the abortive mericarp separates JR RS e, carpophore; Im, lignified mesocarp; rd, rib oil duct: vb, vascular bundle. —B. A. echinatus (Van Wyk s.n | JRAU) ap from the fully developed one in A. echinatus and A. monacanthus. In A. dregei, the abortive carpel are slightly fused along their margins (Fig. 2A) so that they do not separate at maturity. A detailed study of mature fruit in transverse section (Fig. 2) revealed that the epidermis remains continuous between the two mericarps in A. dregei (Fig. 2A) but becomes discontinuous in A. echinatus Fig. 2B) and A. monacanthus (Fig. 2C). Much of the mesocarp becomes lignified in the mature fruit of Arctopus dregei (Fig. 2A) and echinatus (Fig. 2B) but less so in A. monacanthus Fig. 2C). The innermost one or two cell layers of A. monacanthus (Fig. 2C) and one to five layers of A. dregei g. 2A) and A. (Fig. 2B) are lignified. Des was interpreted by Liu et al. (2003) as consisting of a single lignified endocarp layer with the developed carpel and — SN echinatus adjacent lignified mesodermal cells and therefore different from the multicellular lignified endocarp characteristic of Drude’s Hydrocotyloideae (currently treated as Azorelloideae and Mackinlayoideae). In all three species of Arctopus (Fig. 2), the small rib oil ducts are always external to and associated with the numerous vascular bundles. PHYLOGENETIC. RELATIONSHIPS CLADISTIC ANALYSIS OF MORPHOLOGICAL DATA In order to explore character evolution in Arctopus, a cladistic analysis was performed in which 15 characters were polarized (Table 2, Appendix 2). The African endemic saniculoid Alepidea amatymbica was selected as the outgroup because both Alepidea and le da share similar kaurenoic acids (van Wyk et al., 1997), a well as numerous morphological characters (e.g.. geophytes with simple leaves, toothed and occasionally inflexed leaf margins ending in setae, fused pseu- danthial bracteoles and sessile female flowers). A single, shortest-length tree was produced with a tree Volume 95, Number 3 Magee et al. 477 Revision of Arctopus (Apiaceae) B [] X reversals synapomorphies without reversals synapomorphies with reversals ——————9—9C—— Arctopus echinatus 4 10 @ autapomorphies — e 5 6 10 15 _@—_@—@—@— Arctopus monacanthus 2(2) 7 8 9 9) e Arctopus dregei 13 14 dea amatymbica Figure 3. The morphological ¿eros (Table single shortest-length tree (TL = 7 steps, 2) to explore possible relationships within the genus Arctopus. Characters and character states CI = 0.94, RI = 0.83) obtained by a cladistic analysis of 15 are numbered as in Appendix 2. Underlined numbers within the branch nodes are bootstrap percentages. length (TL) of 17 steps, a consistency index (CI) of 0.94, and a retention index (RI) of 0.83 (Fig. 3). The mono- phyly of Arctopus 1s strongly supported (100 bootstrap percentage [BP]) by four rather convincing and un- ambiguous generic synapomorphies, namely the dio- ecious habit, the homochlamydeous perianth in female flowers, the the lignified SD The pseudo-monocarpellate fruit, a generic synapomorphy, was scored as a reversal in A. k seudo-monocarpellate fruit, and echinatus p mature fruit in which both mericarps had developed. Arctopus echinatus and A. monacanthus are strongly supported as sister taxa (95 BP) by five synapomor- phies. Arctopus dregei can be hypothesized to have retained more of the ancestral character states. There populations were found with also appears to be a trend toward complete mono- spermy in A. dregei (perhaps an autapomorphy), with the carpels not separating from one another when mature and the epidermis remaining continuous between the two mericarps. CLADISTIC ANALYSES OF ITS SEQUENCE DATA The total number of characters from ITS-1, ITS-2, and the 5.8S gene was 687, of which 337 were variable (49.05%) and 232 (33.7796) potentially informative. Parsimony analysis produced six short- est-length trees (TL = 715 steps, CI = 0.72, RI = 0.73). The same overall topology was recovered in both the Bayesian (BI) and parsimony (MP) analyses, with the BI differing only in that Sanicula is more resolved, with a poorly supported clade (PP 0.75) consisting of Sanicula arguta Greene ex J. M. Coult. & Rose and S. purpurea H. St. John & Hosaka. A strict consensus of the six trees from MP is shown in It is that ITS the same overall topology as was found in a Figure 4. interesting to note data produced previous analysis (Calviño & Downie, 2007) based on chloroplast data (trnQ-trnK 5': exon region). The tribe Steganotaenieae (represented here by Steganotaenia araliacea Hochst.) forms the earliest branching lineage within Saniculoideae (76 BP; PP 0.84) and Alepidea (represented here by Alepidea amatymbica) orms the earliest branching lineage within Saniculeae (50 BP; PP 0.98). In both analyses, the monophyly of Arctopus 1s strongly supported (100 BP; PP 1.0), but species relationships are unresolved. TAXONOMIC TREATMENT Arctopus L., Sp. Pl.: 1058. 1753. TYPE: Arctopus echinatus L. Apradus Adans., Fam. Pl. 2: 102, 519. 1763, nom. illeg. Annals of the Missouri Botanical Garden Sanicula a Hacquetia Eryngium Eryngium Astrantia 100 1.0 A, 0.98 p 0.84 100 Sanicula purpurea Sanicula arctopoides Petagnaea gussonei Astrantia major Arctopus echinatus Arctopus monacanthus Arctopus dregei rguta epipactis campestre giganteum Saniculeae Saniculoideae minor a amatymbica araliacea 100 m Ins Strict consensus tree of six shortest-length 715 13). Numbers above the branches ar lities obtained by Bayesian Figure 4. region (CI — 0.72, RI — the branches are posterior "proba summer-deciduous + 35 spreading in a geo- diam., Dioecious, acaulescent, phytes; roots long, tuberous, resinous. Leaves simple, prostrate, rosette with the inner smaller; ovate to semi-orbicular or rhomboidal, incised, 3-fid, each segment again 2- or 3-lobed, lobes dentate to crenate; margins toothed, leaf base truncate to sub- mm setae spinose-flexuose; truncate or cuneate; glabrous on both sides; inflexed spine(s) usually present in leaf recesses between segments, often attached to an inflexed laminar tooth; venation palmate, 5- to 7-veined, sunken adaxially but prominently raised abaxially, secondary venation reticulate; petioles adaxially flattened, membrane- winged. Male inflorescence pedunculate, often with up to 5 lateral umbels; involucre spiny, of many linear or occasionally foliose bracts; rays | to 8; involucel spiny, of many to lanceolate bracteoles; umbellule with more than 20 flowers. Male flowers 5, linear-oblong, + same size as ovate pedicellate; sepals m Heteromorpha arborescens Steganotaenieae Apioideae paniculatum ] Hermas clade 5-step trees derived from the parsimony analysis of the nrITS e bootstrap percentages above 50%, and those indicated below — inferenc petals; petals 5, white, oblong to oblanceolate, with inflexed tips, septum on inner face; stamens 5, twice as long as petals, tips inflexed; stylopodium flat, papillate; styles 2, short; ovary absent. Female inflorescence shortly pedunculate, often with up to 5 compound umbels; involucre as in male; rays 2 to 10; involuce accrescent, of 3 to 6 prominent, ovate or obovate, spiny bracteoles, forming a pseudanthium; bracteoles per- manently fused to the fruit. Female flowers sessile, 3 to 5 per pseudanthium, tubular; sepals 5, triangular; petals 5 (the 10 perianth lobes on the female flowers are interpreted here as 5 sepals and 5 sepaloid petals; further ontogenetic studies are required), — Ja however, triangular, sepals and petals persistent in fruit; stamens absent; stylopodium cushion-shaped, raised above the style bases, papillate; styles 2, long, conical, longer than the sepals; ovary bilocular. Fruit usually rostrate, base adnate to the bracteoles, usually pseudo-mono- carpellate due to abortion of inner mericarp. Volume 95, Number 3 , 2008 Magee et al. Revision of Arctopus (Apiaceae) 479 Distribution. The genus is endemic to the Cape Floristic Region of South Africa. The rainfall is generally restricted to the winter months (Goldblatt & Manning, 2002) with only Arctopus echinatus extend- ing into the almost year-round rainfall in the Eastern Cape Province. The three species share a sympatric distribution (Figs. 6, 8). Phenology. The plants are dormant in summer and form new leaves in early winter (April). Flowering commences in June through September. Mature fruit are borne around November, at which time the leaves have begun to die back, followed by complete dormancy throughout the dry summer months. Notes. acters in this genus make it immediately distinguish- The numerous unique and unusual char- able from any other in the family. The species are dioecious geophytes with large, appressed, that lack between sepals and petals, and pseudo-monocarpel- late fruit that spiny, involucel bracteoles. Adanson (1763-1764) superfluously described the genus Apradus from the same Plukenet (1691-1705: tab. 271, fig. 5) and Burman (1738: tab. 1) figures used earlier by Linnaeus (1753) when he described Arctopus. The name Apradus is, therefore, considered illegitimate here. spiny leaves, female flowers differentiation are surrounded by large, KEY TO THE SPECIES OF ARCTOPUS la. Female material (flowers sessile, leaves not required). 2a. alee bracteoles flat or keeled. nvolucel bracteoles ovate, leathery, boat- shaped and inn with an acute APEX. pore kesipuan dia bare ede . echinatus 3b. Involucel bracteoles widely obovate, foli- ose and papery, flat, not keeled, with an obtuse apex A. monacanthus 2b. Involucel bracteoles strongly involute, not eeled ...... ee eee eee . dregei Ib. Male material (flowers pedicellate) or leaves only. 4a. Marginal teeth conspicuous so that lobes appear dentate; inflexed spines in leaf recess- es typically present (except in a population around Milnerton, which has only sparsely spinescent leaves), involucel bracteoles + as long as pedicels, 5-10 mm long. yı w Laminar tooth of inflexed spine(s) in leaf ent or narrow, 0.5—-1.0(—4.0) mm wide, setae CES t always absent on tooth n transverse s recesses absen e Pa with cavities in ctic A. ec chinatus wal > - Laminar e of n xed spine(s) in leaf recesses usually broad, (2.0-)4.0— 10.0 mm wide, Seine usually present on tooth margins; dM without cavities i transverse sectio m AoE 4b. Marginal teeth inconspicuous so that lobes appear crenate; inflexed spines in leaf recess- es usually absent, involucel bracteoles usually shorter than pedicels, 1-5 mm long ... A. dregei l. Arctopus dregei Sond. in Harv. & Sond., Fl. Cap. 2: 565. 1862. TYPE: South Africa. Western Cape Province: Agter de Paarl, sandy hills, s.d., J. F. Drége 7649 (holotype, S!). Figure 5. blades 40— 100 mm long and broad, not thorny; lobes crenate; Leaves widely ovate or rhomboidal, margins inconspicuously toothed; setae flexuose, l- 5 mm; spine(s) in leaf recesses between segments usually absent, when present then inflexed, attached to an inflexed laminar tooth, inflexed part of lamina absent or narrow, less than 1 mm wide, setae res vertical spines O(to 3), 5 mm; petioles 20-8 7 mm; wings broad, 1-4 mm wide. Male cs — with pedicels longer than bracteoles, pedicels 1— 10 mm; bracteoles lanceolate, spiny, 1-5 mm. Female inflorescence with rays 4-15 mm; pseudanthial brac- teoles relatively small; 3 to 6; obovate, involute; ca. 8 X 4 mm at flowering stage, accrescent, becoming 9— 20 X 5-10 mm at fruiting stage; apex obtuse, with a sharp spine, spine ca. 2 mm; margin with 1 or 2 inflexed, spinescent hairs; pseudanthial bracteoles becoming leathery at fruiting stage, reticulate vena- tion not markedly prominent, becoming white when dried, not separating. Fruit not rostrate; l-seeded; surface pustulate, white; developed mericarp ca. 8 X 5mm; not abortive mericarp small, separating in mature fruit. Distribution. Arctopus dregei was originally con- sidered to be endemic to the Lowland Renosterveld of the Western Cape Province, from Paarl to Malmes- bury, Yzerfontein, and Darling, and extending north to Hopefield. However, two recent collections of outlying populations around Vredenburg and Koekenaap in 2006 and 2007 have significantly extended the known geographical range for this species (Fig. 6) and shown for the first time that A. dregei is not restricted to Renosterveld, but may also occur in both Strandveld and Sandveld fynbos. It thrives in seasonally moist sandy soils and is often found in sparse vegetation dominated by Restionaceae (pers. obs.). IUCN Red List dregel is considered Vulnerable (VU Blab[i,11,111,1v,v]) accord- ing to IUCN Red List criteria (IUCN, 2001). Renosterveld (the vegetation to which A. dregei is category. Arctopus gory F Lowland largely restricted) is considered to be one of the most threatened habitats in the Cape Floristic Region, with over 80% of the area completely transformed and currently only a few patches formally protected (Mucina & Rutherford, 2006). I ~ is subject to 480 Annals of the Missouri Botanical Garden UTER A8 0 ™ 2 mm 2 mm 2 mm | 2 mm Figure 5. Arctopus dregei. —A, B. Leaves, adaxial view (A: Magee « Boatwright 2, JRAU; B: Magee & Boatwright 5, JRAU). C-J. Female flowers and m (Magee & Boatwright 31, JRAU). C-F. Pseudanthium. —C, D. Top view. —E. Lateral view. —F. P pee sudanthium bracteole. —G. Female flower. —H. Fruit, lateral view. —I. Fruit, dorsal view. —J. Sepals and petals. K-M. Male flowers (Boucher & Shepherd 4408, NBG). —K. Bracteoles. —L. Sepals. —M. Petals. ei Dm LUE Y "X zd abe [m * Arctopus dregei Pacto @ Arctopus monacanthus | 12 rs HEIGHT ABOVE SEA LEVEL m. : i | 20 O 20 40 60 80 100km " B | * +` & j ea: SS ENS se 1 2 eN Q sun LA aa a y à P o 6 T 18° 19" 20° 2f 22'E 23 24 25 26 ar 28 Figure 6. The known geographical distribution of Arctopus dregei and A. monacanthus. Volume 95, Number 3 2008 Magee et al. 481 Revision of Arctopus (Apiaceae) extensive pressure in the form of agricultural expansion and seepage (drift sprays of herbicides and pesticides, as well as fertilizer runoff), and the spread of alien invasive species and urban encroach- ment and, as a result, is severely fragmented today. has Renosterveld; however, e species also been recorded outside of only two such populations are known. One of is known to occur in Strandveld vegetation near Vredenburg and is threat- ened by proposed industrial expansion (Helme 4181 [NBG], in Raimondo (in prep.) have recommended that A. dregei be considered Vulnerable (VU) according to IUCN Red List criteria (IUCN, 2001) as it has a fragmented distribution and an extent of occurrence estimated to be 12,388 km? schedula). Consequently, Helme and severely is a distinctive Notes. (Fig. 5) species that can easily be distinguished by the crenate Arctopus dregei lobes of the leaves due to the inconspicuously toothed margins and the usually absent inflexed spines in the leaf recesses. The female bracteoles are involute. The male flowers have bracteoles that are shorter than the pedicels. Additional specimens examined. SOUTH AFRICA. Western Cape Province: NW of Koekenaap, Farm Graaf- waler, Goldblatt & Manning 12875 (NBG); 7 km SE of Vredenburg on Langeberg 187, N of R79, E Sands plant, Helme 4/81 (NBG); betw. Darling, Magee & Boatwright 5 (JRAU); Hopefield, E 12699 (BOL. K); Garside 1608 (K); Hopefield, Enkelv Farm, Goldblatt & Manning 11414 (PRE); Yzerfontein, 1 la y Farm, Louw 2529 (NBG); pud Allotment area, Helme 2165 (NBG); Malmesbury, Kli n 598, Helme | (NBG); Maines, Magee & terc 31 (JRAU); Burgers Farm, Boucher & ! Rustenburg Farm, Helme 2719 (NBG); Malmesbury, Lor de Helme 2763 (NBG); Magee & Boatwright 1, 2, . RAU); Paarl, Acocks 2440 (K, PRE). Precise locality LE Drége s.n. M). 2. Arctopus echinatus L., Sp. Pl: 1058. 1753. TYPE: South Africa. Western Cape Province: Cape Town, Signal Hill near Kramat, 16 July 2003, J. Manning & G. Reeves 2845 (proposed conserved type, designated by van Wyk et al. 2006: 541, NBG). Figure 7 Y Ns IT a A => Leaves widely to very widely ovate, semi-orbicular or rhomboidal, blades 20-80 mm long and broad, thorny (sparsely thorny in the Milnerton population); lobes dentate to doubly dentate, rarely crenate (Milnerton population); margins conspicuously toothed; setae spinose, rarely flexuose, 3-10 mm; spine(s) in leaf recesses between segments inflexed, occasionally attached to an inflexed laminar tooth; inflexed part of lamina absent or narrow, 0.5-1.0 — y (—4.0) mm wide, setae absent, vertical spines 1 to 3(4), stout, 6-15 mm; petioles 20-200 X 5-7 mm; wings narrow, Male umbellules with pedicels + as long as the bracteoles; bracteoles ovate to lanceolate, spiny, 5-10 mm. Female inflorescence with rays 10-40 mm: pseudanthial bracteoles rela- tively small, 4 or 5, ovate, boat-shaped and keeled; ca. 6 X 3 mm at flowering stage, accrescent, becoming 10-18 X 5-8 mm at fruiting stage; apex acute, with a sharp spine, spine 2-3 mm; margin with | or 2 inflexed and overlapping spines; pseudanthial bracteoles be- coming leathery and firm at fruiting stage, prominently thickened along periphery and midrib, becoming brown when dried, not separating. Fruit rostrate; usually 1- seeded, rarely 2-seeded; surface usually spiny, brown; developed mericarp ca. 10 X 5 mm: abortive mericarp small, separating in mature fruit. Distribution. Arctopus echinatus is the most wide- [om spread species of the genus, occurring throughout the Western Cape Province and extending both north to Nieuwoudtville in the Northern Cape Province and east along the coast to Port Alfred in the Eastern Cape (Fig. 8). seasonally moist sandy soils (pers. obs.). Province Plants are usually found in IUCN Red List category. considered Least Concern (LC) according to IUCN Red List criteria (IUCN, 2001). Arctopus echinatus is Notes. Arctopus echinatus (Fig. 7) can be distin- guished from A. dregei by the dentate lobes on the — eaves, due to the conspicuously toothed margins, and the presence of inflexed spines in the leaf recesses. More notably, the female bracteoles are ovate, boat- shaped and keeled, with an acute apex. The male flowers have bracteoles that are equal in length to the pedicels. This species differs from A. monacanthus by the narrow or absent inflexed laminar tooth of the spine in the leaf recesses, which usually lacks setae. The leaf that of A. monacanthus in that the petiole in transverse section anatomy also differs from has cavities and the adaxial vascular bundle of the midrib is either simple or dorsiventrally split. Burtt (in Jarvis et al., 1993: 20) designated the Burman plate (tab. 1 in J. Burman, Rar. Afric. Pl. 1738) as lectotype of Arctopus echinatus. Unfortunate- was not — ly, the plant illustrated in this plate echinatus of current usage, but rather as A. Van Wyk et al. (2006) therefore proposed the conservation of A. echinatus identifiable as A. monacanthus. a conserved type in order to maintain the Y wi A usage of this name. This proposal has been recommended by the Nomenclature Committee (Brummit, 2007: 1293). Additional specimens examined. SOUTH AFRICA. East- ern Cape Province: Uitenhage, Drége s.n. (K), Zeyher s.n. 482 Annals of the Missouri Botanical Garden \ NA A Y hm ZI, oF Figure 7. Arctopus echinatus. . Leaves, adaxial view Esterhuysen 12231, end : Adamson 2517, BOL; C: Magee & Boatwright 6, JRAU; D: "i ree & urs 15, JRAU; E: Theron 1764, Pipe y 257, JRAU). G-K. Male flowers (G, H: Van Wyk & Viljoen 3715, JRAU; I: Rycroft 1939, 3 36; J: e IL e G; K: Adamson 2865, PRE). —G. m Petals. —H. Sepals. —I-K. Bracteoles. —L 2 Inflexed lamina spines (L: Taylor E PRE; M: Magee & eae te d JRAU; N: Brink 767, PRE). O-V. Female flowers and fruits (Magee € pent 6, JRAU). O-R. Pseudanthium. —O, P. Top view. —Q. Lateral view. —R. Single pseudanthium bracteole. —S. Fruit, dorsal view. —T. Fruit, lateral view. —U. Flower. — V. Sepals and petals. (BM); Paterson, Retief 471 (PRE); Port Elizabeth, Bolus 2246 Taylor 5099 (PRE): Langebaan Lagoon, Axelson 443 (NBG); (BOL, K), Borle 28 (PRE), Patterson Ed 5 (PRE); Port Darling Flora Reserve, Rycroft 1939 (NBG); betw. Yzerfon- Elizabeth, Victoria Park Lands, Long 618 (K, PRE): Bothas tein & Darling, Magee & Boatwright 6 (JRAU); Sea Point, Hill, Rodgers 4438 (BOL): reia Galpin 10762 (K, MacOwan 1624 (BM, S), Phillips s.n. (BM); Clifton, Acocks PRE), Wau 1445 (PRE); Alexandria, Bushmans River 4566 (S), Barker 1528 (NBG); Rondebosch Common, Barker Mouth, Archibald 3634 (K, PRE), Marais 409 1529 (NBG); Claremont, Salter 9231 (BM); Wynberg Hill, Alexandria, Bokness village, Brink 767 (PRE); Casona. Salter 9224 (BM); Wynberg, Gamble 22123 (K); Table Mtn., Tilney 257 (JRAU); Kowie, Britten 2135 (PRE), Tyson 19229 Fries 3133 (S), Peter 50305 (K); Signal Hill, Blake s.n. (PRE); Bathurst, Henrici 12 (PRE); Port Alfred, Hutton s.n. (NBG); Marloth 5562 (PRE); Lions Rump, Thode A105 (PRE). Northern Cape Province: Van Khyns Pass top, (NBG, PRE), Worsdell s.n. (K); Milnerton, Adamson 2517 Bond 1149 (NBG), Taylor 2908 Ap Ms der Schiff 7179 (BOL), 2865 (BOL, PRE), Hanson 2517 (BOL), Salter 8199 (PRE); Nieuwoudtville, Van Wyk 4128a (JRAU); Aarend- (BOL); Killarney, Lussem 17 (NBG); Porterville, Grootwin- skraal, Van Wyk s.n. (JRAU); E Van Wyk & terhoek Mtn., Barker 262 (PRE); Stellenbosch, Caride 1506 yan vd JRAU); Oorlogskloof s Tuinl: aagle, Magee (K), Kerfoot 5842 (PRE) [atypical, possible hybrid with et al. 93 (JRAU). Western Cape vince: Cederberg, dian ous O hu] Moss 4046 (BM); Gydo Pass, Johnson betw. ume rg hut € Crystal Pool, a ee (BOL); | 505 (NBG); Ceres, C. M. Van Wyk 3319 (PRE): Ceres, Cederberg, Scorpionspoort, Esterhuysen 12231 (BOL): Ce- le Hanekom 662 & 663 (PRE); Worcester Veld darberg, Thode A2003 (PRE); 11 km from Op-Die-Berg to rve, Olivier 61 (PRE); Worcester, Van Breda 71 (PRE): Cedarberg, Stirton 9162 (PRE); Rondegat River valley. Z da ester, Langerug koppie, Walters 1991 & 1992 (NBG): Maitre 337 (PRE): Sneeuberg, Cedarberg Forest tt Du Tooits Kloof pass, Winter et al. 170 (JRAU): Botrivier - s Y — Volume 95, Number 3 2008 Magee et al. Revision of Arctopus (Apiaceae) 3 ces | Si h E Arctopus echinatus HEIGHT ABOVE SEA LEVEL Over 1500 m EJ 900 - 1500 [EZ] 300 - 900 m C Under 300 m 20 O 20 40 60 80 100km 4 Frnt Po S A, ZB Figure 8. vlei, Loubser s.n. (NBG); Clock Peaks, Wurts 148 (NBG); Constantianek, Young 26521 (PRE); Simon's Town, Alice & Godman 67 (BM); Betty’s Bay, Dawidskraal, Boucher 1379 (NBG); Kleinmond, Burman 1171 (BOL), De Vos 607 (NBG); Palmiet, De Vos 719 (NBG); Hermanus, Fernkloof Nature Reserve, Orchard 558 (K, NBG, PRE), Magee de Boatwright 15 (NBG); Hermanus, Westcliff, Williams 303 (K); Pearly Beach, Kleinhagelkraal, C. M. Wyk 616 (NBG, PRE); Zoetendalsvlei, Smith 3098 (PRE); De Hoop flats near Die Mond, Burgers 2576 (NBG); De Hoop, Van der Merwe 800 (PRE); Soetmelksfontein, Muir 5204 (PRE); Riversdale, Still Bay, Bohnen 7540 (NBG, PRE); Riversdale, Muir 24400 (BOL) Knysna, Theron 1764 (PRE) Kranshoek lookout point, near Harkerville, Boatwright, Magee & Van Wyk 4. (JRAU); Humansdorp, Blouleliesbos, Geldenhuys 99 (PRE). (S) Precise locality unknown: Sieber 141 (S). — 3. Arctopus monacanthus Carmich. ex Sond., Fl. ap. 65. 1862. TYPE: South Africa. Cape Province: s.d., Captain Carmichael s.n. (holo- type, S!). Figure 9. Leaves widely to very widely ovate, semi-orbicular or rhomboidal, blades 25-120 mm long and broad, thorny; lobes dentate to doubly dentate; margins conspicuously toothed; setae spinose, 3-7 mm; spine(s) in leaf recesses between segments inflexed, attached to an inflexed laminar tooth; inflexed part of lamina usually broad, (2.0-)4.0-10.0 mm wide, setae present, vertical spines 1 to 5(6), stout, 2-20 mm; petioles 20-100 X 3-7 mm; wings narrow, ca. 1 mm + wide. Male umbellules with pedicels as long as The known geographical distribution of Arctopus echinatus. bracteoles; bracteoles lanceolate or occasionally ovate, spiny, 5-10 mm. Female inflorescence with rays 10-50 mm; pseudanthial bracteoles large, 4 or 5, widely obovate, foliose, ca. 10 X 3 mm at flowering stage, massively accrescent, becoming 20-40 X 15- 50 mm at fruiting stage; apex obtuse, with a sharp spine, spine ca. 4 mm; margin entire, rarely with 1 to l and papery at fruiting stage, reticulate venation 3 spines; pseudanthial bracteoles becoming brittle prominent, becoming brown when dried, breaking away (with the ripe fruit attached). Fruit markedly rostrate; l-seeded; surface occasionally spiny, brown; abortive developed mericarp ca. 5 mm mericarp small, separating in mature fruit. Distribution. Arctopus monacanthus has an ex- treme western distribution occurring from Somerset West in the Western Cape Province, north toward the Northern Cape Province (Fig. 6). It is an inland species occurring in seasonally Nieuwoudtville in moist clay soils or, less often, on sand (pers. obs.). IUCN Red List category. considered Least Concern (LC) according to IUC Red List criteria (IUCN, 2001). Arctopus monacanthus is Arctopus monacanthus (Fig. 9) is similar to A. echinatus, but differs in having a broad inflexed laminar tooth in each of the leaf recesses with setae usually present on the margins of these teeth. The 484 Annals of the Missouri Botanical Garden DA oy l1 mm | 2mm | 2mm | " Obl Figure 9. Arctopus an —A-C. Leaves, adaxial view (A: Acocks 17286, PRE; B: Mauve & Hugo 49, PRE: C: Van Wyk 4141a, JRAU). —D-I. Inflexed lamina spines (D: Van Wyk iiid JRAU: E: Barker 9882, NBG; F: Loubser 3348, NBG; G: Van Zyl 31: > Ne i soe 511, NBG; I: Van Wyk 3522, JRAU). J-P. Female flowers and fruits (J, M, P: Van Wyk 4161a, JRAU; N, O: Magee el al. 96, JRAU). J-M. Pseudanthium. —J. Lateral view. —K. Top view. —L, M. Single P udanthium bracte va ss, with fru Ae al view. —N. Sepals. —O. Flower. —P. Fruit, ple view. Q-V. Male flowers (Q, ~ Boucher 4389, NBG; U: peur 512, NBG; R, V: Parker 3518, PRE). —Q. Sepals. S. Petals. —T-V. Bracteoles. female bracteoles are widely obovate and foliose with 3/40a (NBG); Malmesbury, N side of Hercules Pillar. an obtuse apex. In transverse section, the petiole is Adamson 3168 (B OL): Dassenberg, summit of Kanon- : suis 4 > kop, Boucher 4389 (NBG); Tygerberg Nature Reserve, itl 'avities a » adaxial vascular : A I g 8 i "e eee ui the id ixial vascular bundle of the Loubser 3348 (NBG); Langverwacht above Kuils River, midrib is collaterally split. Olivier 4686 (NBG); Bottelaryberg, Acock 2511 (S); Stellen- . . "UR NER bosch. Hos 328 (PRE). Gillett 583 (NBG); Stellenbosch, Additional ipe e examined. SOUTH AFRICA. Vlottenberg station, Stegmann 10566 (BOL, PRE): Straat- Northern Cape Province: Nieuwoudtville, Van Wyk kerk, 3km W of Tulbagh, Mauve & Hugo 49 (PRE): 4128b (JRAU); Nie mapu Oorlogskloof Nature. Re- Elandskloof pass, Van Wyk 4141a (JRAU); Jonaskop Van serve, Pretorius 181504 (NBG); Oorlogskloof farm, Tuin- — Wy , (JRAU); McGregor rd., Van Wyk, Winter & Tilney laagte, Magee et al. x. (JRAU): Heerenlogement, Barker E (JRAU); Somerset West, Boucher & Mauve 4948 (PRE). 9882 (NBG); Gifberg, Van Wyk 3522 8 ela d Parker 3518 OI NBG, PRE); Somerset West, Lokenburg, Acocks 17286 (PRE). Western Cape . Runnalls 511 & 512 (NBG); Piqueniers Kloof, Oskolski et al. 35-06 (JRAU), Se ite Somerset in eee Holland Mtn., Ecklon & Zeyher s.n. 10756 (BM, BOL, K); Piquetberg, De Hoek, Steyn 540 NBG); - m D 3 E d: S). Precise locality unknown: Malmesbury-Clanwilliam rd., Pique here Laers 19629 (BOL), Stephens & Glover — Alice & Godman 514 (BM). 4 K), Van Wyk 3594 (JRAU); Pakhuis Pass top, Wisura 2982 (NBG), Magee da Boatwright 34 (JRAU); Clanwilliam, 20 km along rd. to Algeria, Scott 393 (BOL); Wupperthal, MacOwan 3230 do: Cedarberg, Middelberg West, Viviers 514 (PRE); Middleberg pass, Van Wyk s.n. (JRAU); Oliphants River Valley, opposite Warm Bath, Adanson, M. 1703-1764. Familles des Plantes. Vincent, Stephens 7246 (K); Durbanville, Klipheuwel Farm, Van Zyl Paris. ~ co X Ed = c — w” Literature Cited Volume 95, Number 3 2008 Magee et al. Revision of Arctopus (Apiaceae) Brummit, R. K. 2007. Report of the Nomenclature Committee 12 for Vascular Plants: 59. Taxon 56 89-1206 1738. Burman. J. Rariorum Africanarum eee Am- sterdam. Burtt. B. L. 1991. Umbelliferae of southern Africa: An introduction and annotated checklist. Edinburgh 48(2): 184. & S. R. Di 2007. Circumscription and phylogeny of Apiaceae eit amily Saniculoideae based on c =o last DNA sequences. Molec. Evol. 44: 9]. Notes Bot. Gard. Calvino, C. I vnie. Phylogen. ———, P. M. Tilney, B.-E. van Wyk € S. R. Downie. 2006, A molecular phylogenetic study of southern African any Amer. J. Bot. 93: 1828-1847. Chandler, G. T. & Plunkett. 2004. Evolution in Apiales: Nuclear. and Moroplast markers together in (almost) perio! harmony. Bot. J. Linn. Soc. 64: us Downie. S. R. € D. S. Katz-Downie. 1996. A molecular phylogeny of Apiaceae subfamily Apioideae: Evidence from nuclear ribosomal DNA A transcribed spacer sequences. Amer. € Bot. 83: 234-25 Doyle, J. J. & J. Dovle. 1987. A m procedure for 2 mr s of fresh leaf tissue. Bull. Bot. Soc. Amer. 19: 11-15 897 belli ferae. Pp. 134-135 in rantl (editors), Engelmann, DNA isolation Phytol. Engler & K. P Die naturlichen Pflanze ui milien 3(8). Farris, J. 1970. Methods for SEDE UN Wagner trees. Syst. Leipz Zool. 34: 21-24 Feder, N. & T. P, YBrien. 1968. Plant microtec eu me principles Me new methods. Amer. J. 55: 1235 phyloge mies: An approach using p rd Evolution 39: 783-791. Fitch. W. M. . Towards defining the course of e ed oe e ndn for a specific tree topology. Syst. Zool. 2 106 Froebe, "i A. 1964. Die Blutenstande (U S Be itr. Biol. Pflanzen 40: 381— — . Manning. 2002. lant diversity of the rape ee al pube om Mic :a. Ann. Missouri Bot. Gard. 281-302. K.. N. Index HOMME: i Felsenstein. J. 1985. Confidence limits on > der Sanic "uloideen 383. Holmgren, P. . Holmgren & L. C. Barnett. 1990, The Herbaria of the World. 8th ed. New York ie al Garden, Yonquist. 2001. MRB Bay ivesian inference of phylogeny. ie a s 17: Huelsenbeck. J. P. n F. | ic N. 2001. IUCN Red List e and Criteria Version Prepared by the IUC IUCN, Gland. cies Survival Commission. Lene and Cambridge, United ^ R. Barrie. D. M. st of Linnaean ^ui neri Allan & J. L. Reveal. Names and 1993. Their Types. Regnum Veg. 12 Leistner. O. A. & J. ML ak 197 6. pur African place names. 2: 565 Ann. Cape Prov. Mus. Nat. 2: 565. Linnaeus, C. 1753. Species taa: eS Salvii, Stock- holm. Liu, M. R., B.-E. van Wyk € P. M. Tilney. 2003. The taxonomic value of fruit structure in the subfamily Saniculoideae and related African genera (Apiaceae). Taxon 52: 261-270. Magee, A. R., B.-E. van Wyk & F. van Vuuren. 2007. miner activily of oe Bot. 73: 159-1€ a topus specie Ethnobotany (Arct African J. Magin, N. 1980. Eine be rhis ‘he Analyse der O8 rst Lagoeciaeae (Apiace: Evol. 133: 239-259, Mucina, L. € M. 3 LS E Wer 2006. a of South Lesotho and Swaziland Strelitzia. Vol. 19, South African Biodiversity. Institute, Pre toria. 1 M. F. Watson. 2004. Phylogenetic re claims i in Bupleurum (Apiaceae) based on ds ania DN Bot. 319- 'S sequence data. Ánn. e. L. 1857. Florae Capensis Medicae Prodromus. 2: 19. Cape Town. Pimenov, M. C. & M. V. Leonov. 1993. The Genera of the Umbelliferae. m al pus al Gardens e E, 91-17 Phytographia, y a Almagesti miel DN TAn G.M. & a II. 2001. Re A LE anions ‘ancient araliads’ and their significance for the of Api: ae Molec. Phylogen. Evol. 19: 259-27 6 Chandler, P. P. Lowry II, S. M. Pinney p E l. Recent NE Apiales and a revised classification. S. African J. Bot. 70: 371-381. Posada, D. & K. A. Crandall. the model of DNA 817-818 Honquist, F. & J. P. Huelsenbeck. 2003. MRBAYES 3: Bayesian. phylogenetic inference ee Lp ¡stematics advances in 1998. substitution. MODELTEST: Testing 14: Bioinformatics under mixed models. Bioinformatics 19: 1572—157¢ Smith, C. A. 1966. Common Nuaies of South African Plants. Memoirs of the Botanical Survey of South Af rica, Vol. 35. a a v 964—565 in W. de Sun. Y i Skinner, ( 1994. ed taxa using internal transcribed E ers of nuclear ribosomal DNA. Theor. a senet. 89: 26-32. Swofford, D. L. 2002. PAUP*. parsimony (and other adn Vers. Sunderland. Massachuse Valiejo- Roman, C. M., M. c. Tien E. 1.1 Down D. S. Katz-Downie € A. V. Molecular of Umbelliferae: rDNA internal transcribed spacer se > H. Liang & S. H. Hulbert. Phylogenetic analysis of sorghum and re e Phylogenetic analysis using 4.0 b10. Sinauer, 'erentieva, S. R. Troitsky. 1998. Using nuclear quences to resolve issues of e Peu relationships. Bot. Zhurn. (Moscow & E ad) 8 systematics , E. I. Te inim T. H. Samigullin € M. G. Pimenov. 2002. A lutionshoos among genera in Saniculoideae and selected Apioideae ono y inferred from nrITS S. Taxon 5l Ss. & P. M: es m Diversity of Apiaceae in Africa. S. African J. Bot. 70: 433—445 Oudtshoorn € N. Gericke. ————, B. 1997. Medicinal Plants of uh A frica. Briza Publications, Pretoria & M 2006. Arctopus echinatus ———— agee, C. E. Jarvis Pp. Tilney. (17 29) bord to conserve the name with a conserve red type (Apiaceae). Taxon 55: 54 as, P., B. G. Baldwin & L. Constance. 1998. juclear ribosomal DNA evidence for a western North i n igin of Hawaiian and South ae an species of Natl. Acad. Sei. U.S.A. 95: Sanicula (Apiaceae). Proc. 235-240. Wolff. H. 1913. Umbelliferae-Saniculoideae. Pp. A. Engler (editor), Das Planzenreich. Vol. 61. Leipzi 274-27 Enge Cuin 486 Annals of the Missouri Botanical Garden APPENDIX T. List of the voucher specimens of taxa used in the molecular analysis, with GenBank accession numbers. NEW SEQUENCE DATA Azore nese Magee & J. Hermas villosa Thunb.. A. R. (JRAL S. Boatwright 2 I ), AM748815. pS miculoideae: Vepidea amat ymbica Eckl. & Zeyh., A. R. Magee & J. : Aoi pe (JRAU), AM158945; oe dregei Sond., 4. . M os oy Boat right 2 (JRA . AM158942; Arctopus a. 6 (JRAU) AM 158943: Vs vu tum D armich. ex Sond., B.-F. van Wyk 5522 (JRAU), AM 158944; Hochst., O. Maurin 566 (JR AU). dL L., e Magee & , Steganotaenta araliacea AM748814. SEQUENCE DATA FROM GENBANK Burtt. Apioideae: Anginon and ( Thunb.) & Heeroma arbor- AF467922 (Neves Watson, 2004): & Schltdl., Saniculoideae: escens Cham. U27578 (Downie 1996). (Downie & AF337183, gium campestre Ly 4 = J = Astrantia Katz-Downie, 1996): AF337191 (Valiejo-Roman et al., 2t FO77887 (Valiejo-Roman et al., 1998 tum giganteum Bieberd., AF337182, AF33 719 hielo Román et al., 2002): a epipact AF077892 (Valiejo-Roman et al., 1998): CHEN gussonel AM403487, che (De Eryng (Spreng.) Rauschert. Castro, AF031974 Hinpuplene’) Sanicula ie wat Hook. & Arn., 1 998); Sa Coult. icula ae me ene ex J. et al., 1998); Sanicula purpurea AFO3 1971 ee et al., 1998). = & Rose, AF031975 (Vargas H. St. John & Hosaka. APPENDIX 2. Morphological characters and character states used for cladistic analysis of Arctopus. l. Setae on leaf margins: 0 = flexuose: | spinose. 2. Inflexed laminar teeth in the recesses between the leaf segments: 0 = absent; | = present, narrow (< 2(—4) mm); 2 = present, broad (> 4 mm). 3. Collenehyma in petiole wings: O = absent; I = present. l. Petiole cavities: O = absent; | = present. 5. Sexual system: 0 = monoecious; | = dioecious. O. Sepals and pet " of fem flowers: 0 = distinct sepals and petals: 1 = homoc lame sous Jer rianth. 1. Size of la braeteoles: 0 = small: | large. 8. Shape of pseudanthium apex: 0 = acute: | = obtuse. 9. Texture of pseudanthium: 0 = leathery; | = papery. 10. Fruits: O = bicarpellate: 1 = peen 'udo-monocarpellate. 11. Fruit shape: 0 = not rostrate; | = rostrate. 12. Fruit surface: 0 = protuberant; | = spinescent. — separating: 13. Mature fruits: — not separating. l4. Fruit. epidermis between mericarps of mature fruits: O = discontinuous; | = continuous. 15. Endocarp: 0 = parenchymatous: | = lignified. A NATURAL HYBRID BETWEEN Yuezhi Pan,? Suhua Shi,” Xun Gong,+* and LIGULARIA PARADOXA AND L. Chiaki Kuroda“ DUCIFORMIS (ASTERACEAE, SENECIONEAE) FROM YUNNAN, CHINA’ ABSTRACT V putative natural hybrid in t n Cass., which is morphologically intermediate between L. paradoxa Hand.-Mazz. and L, duc ond: (C. Winkl.) Hand.-Mazz.. was found on Mt. Maoniu in northwestern Yunnan, China. We employed analyses of morphology, seed vigor, somatic c -hromosome numbers, meiotic behavior of chromosomes, inter-sin mple sequence repeat (ISSR) markers, and trnL-F sequencing to test the hypothesis that the unidentified taxon is a natural interspecific hybrid between L. found in seeds of the putative natural hybrid. The meiotic behavior of the putative natural hybrid was abnormal. Among 10 ISSR primers that we tested in all three > laxa, four primers produc ed different ISSR markers between the putative parents. The paradoxa and L. duc iformis as bo nts of the putative natural hybrid. Ligularia a inire was determined as the female parent of this natural hybrid by nucleotide sequence from the chloroplast trnL-F region. Therefore, this taxon is confirmed to be a natural hybrid between £. ue and L. duciformis and is described herein. Key words: Asteraceae, IUCN Red List, Ligularia, natural hybrid, Senecioneae. Hybridization is common in many groups of examples (Barton, 1989: Wendel, 1989; Soltis et al., flowering plants. (Stebbins, 1959; Grant, 1981: 199]: Harris & Ingram, 1992; Soltis & Soltis, 1993; Ellstrand et al., 1996) and is increasingly recognized Brochmann et al., 1996; Sun, 1996; Sang et al., 1997; as an important source of evolutionary novelty Padgett et al., 1998; Wolfe et al., 1998; Aparicio et (Arnold, 1997; Rieseberg, 1997). Hybrids can arise al., 2000; Barton et al., 2001; Choi et al., 2001; Allen, in any situation where genetic divergence is not 2002). Molecular data can yield many insights into the LI closely accompanied by development of a reproduc- nature and parentage of hybrid species and, as further tive barrier, thus creating the potential for gene flow cases are investigated, into the circumstances that between entities that are taxonomically recognizable favor allopolyploid speciation. They are most com- and morphologically distinct. The consequences of monly observed in outcrossing, perennial taxa, hybridization may include introgression affecting one typically those with some mode of vegetative propa- or both taxa, formation of hybrid (especially alloploid) gation. species, and development of reticulate patterns of The genus Ligularia Cass. (Asteraceae, Senecio- evolution within a group (Arnold, 1997). neae) consists of about 130 species mainly distributed Although hybridization in combination with poly- in eastern Asia, 112 species of which are known in .. 1994). Sixty-seven species were speciation in plants (Stebbins, 1971; Levin, 1983), the recognized in the Hengduan Mountains in southwest- increasing use of molecular approaches has facilitated ern China; of them, 61 species are endemic to this understanding of this process and revealed many new area (Liu, 1989). The Hengduan Mountains have been ploidy has long been accepted as a mechanism of China (Liu et al 'This study was supported by grants from the Chinese National Natural Science Foundation (NSFC, grant numbers 3007008 ps 30130030) and a erant from the Scientific Foundation of Yunnan Province (2001C o Vouc het specimens are deposited in the herbarium of Kunming Institute of Botany, Chinese Academy of Sciences (KUN). We thank Qitai Zhang, Qing Yin, and Zhiyun Yang (Kunming Institute of Botany, Chinese Academy of Sciences) for their assistance in collec ing materials, and Zhiyun Su (Kunming Institute of Botany, Chinese / e my of Sci iences) for his assistance in writing the Latin diagnosis. ? Kunming Institute of Botany, Chinese N of Sciences, Kunming 650204, Yunnan, People’s Republic of China. sein E author: gongxun(Omail.kib School of Life Sciences, Sun Yat-Sen Unive Guangzhou 510275, Guangdong, People's Republic of China. ‘Department of Chemistry, Rikkyo University, Nishi-Ikebukuro, Toshima-ku, Tokyo 171-8501, Japan. doi: 10.3417/2006034 = ANN. Missourt Bor. Garp. 95: 487—494. PUBLISHED ON 23 SEPTEMBER 2008. Annals of the Missouri Botanical Garden considered a main center of evolution and diversifi- cation of Ligularia (Liu et al.. 1994). Ligularia is a typical alpine genus and a key taxon to studies on the origin and development of the Sino-Himalayan alpine has been taxonomically Handel-Mazzetti (1938), Hu (1967), (1985, 1989), no molecular phylogenetic studies have been reported floristic elements of flora. Although Ligularta studied by and Liu yet. The authors are doing a comprehensive study on the phylogeny of Ligularia using various methods to investigate its morphology, cytology, and molecular biology. field Mountains i in Yunnan, we found a mixed population During collection trips to the Hengduan of three taxa, namely Ligularia paradoxa Hand.- Mazz.. L. duciformis (C. Winkl.) Hand.-Mazz.. and a putative natural hybrid, occurring on Mt. Maoniu. This mixed population grows along a stream in the Po} 1200 m. The nothospecies was considered to be a natural hybrid forest at an altitude of putative between £L. paradoxa and L. duciformis based on its intermediate morphological appearance and the partial overlap of the flowering period of two putative parents. Given this anthesal overlap, cross-pollination between two putative parents could occur naturally under sympatric conditions. The aim of the present study is to clarify. the taxonomic status of the putative natural hybrid of Ligularia duciformis and L. paradoxa found on Mt. Maoniu, and to determine systematic relationships among these three taxa using various taxonomic approaches, including comparison of external mor- phology, seed vigor, examination of somatic chromo- some number and meiotic behavior, inter-simple sequence repeat (ISSR) markers, and trnL-F sequenc- ing. MATERIALS AND METHODS Morphological observation was carried out in the field during the 2001 Leaf form and pappus color were selected among and 2002 flowering seasons morphological characters as diagnostic characters. For studies on seed vigor, achenes of the three taxa were collected in October 2001. Embryos of all three taxa light weights were measured based on 1000 achenes, and were observed under a microscope. Achene the germination rates of achenes, sown in soil, were counted. Somatic chromosome numbers were examined by the conventional aceto-orcein squash method using root tips. Root tips were collected from five living the field pretreated with 0.1% colchicine solution for individuals per taxon in and then were 3 hr. After fixation with a 3:1 mixture of 95% ethanol and glacial acetic acid at 4 C for 15 min.. root tips were macerated in a 1:1 mixture of 45% acetic acid and | M hydrochloric acid at 60°C for stained with 1% 3 min., and then € aceto-orcein. To study meiosis, young calathides of the three taxa were fixed in Carnoy's solution (3:1 ethanokacetie acid) in the field, then transferred to 70% ethanol and refrigerated. Squashes were made in a drop of aceto-orcein. For study of both ISSR markers and the sequences of the chloroplast trnL-F region, the young ane of fL paradoxa, and 16 of the putative natural e were in the field. Total genomic DNA was extracted from the silica gel-dried leaf tissue using CTAB methods (Doyle & Doyle. 1987). All 90 ISSR primers were used for screening all individuals. and 10 primers (University of British Columbia, Canada [UBC] primers set #9: 807. 808, 809, 815. 820. 823. 827, 828. 834. 848) were selected for amplifying bands. The polymerase chain reaction 56 individuals (20 of Ligularia duciformis. 20 « collected and dried in silica gel z (PCR) included the Tace cycles: preheat, 94^C, 5 min. 45 cycles, denaturation, 94°C, 1 min; annealing, 52°C, 1 min.; extension, 72°C, 1.5 min.; final extension, 72 C, 7 min. The amplified. bands from three taxa were scored as (1) common to both putative parents, (2) shared with one of the putative parents, or (3) unique to the putative natural hybrid. The chloroplast trnL-F region was amplified using the universal primers trnC. and trn (Taberlet et al.. 1991). Amplification conditions were as follows: | cycle, 94°C, 4 min.; 28 cycles, 94°C, 45 sec.; 58°C, 45 sec.; 72°C, 2 min., followed by 1 cycle, 72°C, 8 min. The PCR products were purified by electro- phoresis through a 1.2% agarose gel followed by use of an E.Z.N.A. Gel Extraction Kit (Omega, China). All chloroplast. trnL-F Guang- zhou, accessions were subjected to sequencing, with amplification primers in an ABI 3700 DNA automated sequencer with the BigDye Terminator Cycle Sequencing Kit U.S.A.) comparison L. (AYO98701). have been deposited in GenBank (accession numbers DQ104429-DQ104431). The sequences were aligned (Applied Biosystems, Foster City, California, Boundaries were determined by with Vicotlana paniculata All sequences and compared in SeqMan (DNASTAR, Beijing. China). RESULTS MORPHOLOGICAL ANALYSIS Ligularia duciformis series Retusae S. and £. paradoxa belong t Liu (Liu, 1989). The morphological differences between the two species are The main leaf form and pappus color (Table 1, Fig. Volume 95, Number 3 Pan et al. 489 A Ligularia Hybrid from Yunnan, China Table 1. Morphological comparison of Ligularia paradoxa, L. duciformis, and the nothospecies L. Xmaoniushanensis. Characters Taxon Habitat Stems Leaf blades Pappus L. paradoxa perennial proximally glabrous, distally orbieular or broadly brown, shorter than tubular corolla herbs arachnoid-puberulous ovate, palmatisect and longer than its tube L. duciformis perennial glabrous or distally short reniform or cordate, white, as long as tube of tubular herbs yellow pilose margin irregularly corolla dentate L. Xmaoniushanensis perennial glabrous or sparsely broadly ovate, wine-colored, white at base, shorter herbs pubescent above palmately parted than tubular corolla and longer than its tube xd T . Figure 1. Photos of leaves, florets, and chromosomes of Ligularia paradoxa, L. duciformis, and L. Xmaontushanensis. —A. Leaf of L. paradoxa. —B. Leaf of L. duciformis. —C. Leaf of L. Xmaoniushanensis. —D. L. ng L. duciformis, and L. Xmaoniushanensts in a mixed population. —E. Florets of L. paradoxa (El). L. duciformis (E3). and L. Xmaoniushanensis (E2). —F. Somatic chromosomes of L. dip id prs Somatic chromosomes of L. duciformis. —H. Somatic chromosomes of L. Xmaoniushanensis. Scale bars: A-C mm, E = 4 mm, F-H = 10 um 490 Annals of the Missouri Botanical Garden Table 2. The weight per 1000 seeds and germination rates of Ligularia paradoxa. L. duciformis. and the nothospecies L. Xmaontushanensis. L. paradoxa L. Xmaoniushanensis L. duciformis 1.17 9.0 Weight per 1000 seeds (g) Germination rate (£6 0.79 2.2 0 56.5 leaves of L. paradoxa are 3- to 8-palmatipartite (Fig. LA, D), and those of L. duciformis are cordate or reniform with serrate margins (Fig. 1B, D). The pappus is white in L. duciformis (Fig. 13). and brown in L. paradoxa (Fig. 1E1). The morphological characters of leaf form and pappus of the putative natural hybrid are as follows: dissected peltate leaves IC, E2). The putative natural hybrid shares with eight to 15 lobes (Fig. D) and wine-colored pappus (Fig. | the morphological characters of its two putative parents in leaf form and pappus color. SEED VIGOR Embryos were observed in seeds of both Ligularia paradoxa and L. duciformis. No embryo was observed in the seeds of the putative natural hybrid. The weight 1000 seeds are listed i and germination rales per Table 2. There are distinct differences among the er 1000 seeds No germination was observed for and in seed three taxa in the weight | germination rales. seeds of the putative natural hybrid, with a sample size of 1000 seeds investigated. SOMATIC CHROMOSOME NUMBERS The two putative parents and the putative natural hybrid are diploid and share the same chromosome 2x — 58 (Table 3, Fig. 1F-H). Although the karvologies of the two putative parents are not. well number, investigated, it is possible that hybridization. takes place among taxa sharing the same chromosome number. MEIOTIC BEHAVIOR Chromosomes for the two putative parents dis- played a normal meiotic process, characterized by the formation of 29 bivalents in diakinesis (Fig. 2A, D). Table 3 X maoniushanensis from Mt. Maoniu, Ninglang Co.. Chromosome numbers and meiotic behavior in Ligularia paradoxa, I Yunnan. with chromosome migration to two poles in anaphase I Fig. 2B, E), and the formation of tetrads (Fig. 2C, F). However, the putative natural hybrid showed meiotic — abnormalities in almost all phases, especially in diakinesis of meiotic prophase | with univalents, in and H bridges. anaphase | with lagged univalents and and in telophase H with free 2G—J). chromatid chromosomes (Fig. ISSR MARKERS AND TRAL-F SEQUENCING Although clear taxa using 10 primers, banding patterns of only four the parents. The banding patterns of ISSR are summarized bands were amplified from three primers showed differences. between putative 13 bands were Of both putative in Table 4. For example, a total of amplified from three taxa using the primer 807. these, nine bands were common parents, two bands were unique to Ligularia para- doxa, and two bands were unique to L. duciformis. The putative natural hybrid shared 12 bands with the putative parents, and no additive bands were observed (Fig. 2K). : The sequences of the chloroplast trnL-F region generated from the Ligularia hybrid and its parents were consistently 852 bp in length. No intraspecific sequence variation was observed for either parental species, although there were three nucleotide substi- lutions between the parental species (Table 5). All the sequences with L. paradoxa. hybrid had identical Thus, £. considered the female parent of this natural hybrid accessions of putative paradoxa is because chloroplast DNA is maternally inherited Ligularia (Zhang et al., 2003) DISCUSSION Speciation by hybridization has been well docu- mented in plants (Abbot, 1992: Arnold, 1992: duciformis, and the nothospecies L. China. Taxon Voucher and origin Chromosome number Meiotie behav lor X. Gong 22350 (KUN) X. Gong 22355 (KUN) X. Gong 22357 (KUN) [o] L. paradoxa L. duciformis L. Xmaoniushanensis 2n — 58 Normal 2n = 58 Normal 2n — 58 Abnormal Volume 95, Number 3 Pan et al. 491 2008 A Ligularia Hybrid from Yunnan, China Figure 2. Meiosis of Ligularia paradoxa, L. duciformis, ee the nothospecies L. Qs dE A-C. Ligularia t paradoxa. —A. Diakinesis. —B. Anaphase ay Tetrad. D-F. Ligularia Ceos: —D. Diakine . Anaphase I. —F. Tetrad. G-J. Ligularia FEL. —G. Diakinesis e po n . —H. ii es | with lagge rod univalents and chromatid bridges. —I. Anaphase II with ie univalents and chromatid bridge ss. —J. Tetrad with free chromosomes. — Banding ISSR patterns om primer 807 in L. a (Lp), L. duciformis (Ld), the Bees cies L. Xmaoniushanensis (L m), and marker banding (M). Scale bars: A-J = um. Annals of the Missouri Botanical Garden Table 4. CTC TCT CTC TCT CC), 827(ACA CAC ACA CAC The banding patterns of ISSR amplified from the four primers 807/AGA GAG AGA GAG AGA GT), 823(1CT ACA CG), and 884(AGA GAG AGA GAG AGA G(CT)T) in Ligularta paradoxa, L. duciformis, and the nothospecies L. Xmaoniushanensis. AB, number of ISSR bands additive to those for the putative natural hybrid; TB, total number of ISSR bands: UB. number of ISSR bands unique to one of the putative parents. ISSR primer 807 823 2 834 Total Voucher and Taxon origin TB UB AB TB UB AB TB LB AB TB UB AD TB UB AB L. paradoxa X. Gong 22356 1] 2 - 8 1 - 7 2 - 5 p - 31 5 - (KUN) L. Xmaoniushanensis X. Gong 22357 12 = 0 I0 = 0 l5 - 0 9 - 0 lo = 0 (KUN) L. duciformis X. Gong 22355 |l 2 - 9 2 - l5 8 E 9 | - 12 16 - (KUN) Rieseberg, 1995). Gottlieb (1972) discussed several criteria for testing whether a particular diploid taxon originated through hybridization. These features included a geographical distribution in the region of parental sympatry, morphological intermediacy in several characters, partial fertility, biochemical addi- individuals tivity, and experimental synthesis of resembling the hybrid. Although no single criterion can provide a clear means for testing a hypothesis of that be fulfilled provides a higher level of support for a hybrid origin (Gottlieb. 1972). It that criteria be evaluated carefully because morphological hybridization, each criterion can is also essential hybrid intermediaey and sterility are not invariably associ- 1995) and may result from entirely Les & Philbrick. 1993). Nevertheless, Gottlieb’s criteria (1972) provide the natural hybrid between Ligularia duciformis and L. ated with hybrids (Rieseberg, separate processes a convenient avenue for discussing putative paradoxa. In the present study, the morphological characters of leaf shape and pappus color were clearly intermediate with those of Ligularia paradoxa and L. duciformis, and the putative natural hybrid satisfies Gottlieb’s (1972) criterion of morphological interme- diaey. Ligularia paradoxa, one of the only two species with palmatipartite leaves in Ligularia, is restricted to Yunnan and Sichuan provinces, China, while Z. Table 5. the nothospecies L. Xmaontushanensis. Variable sites of the nucleotide sequences of chloroplast trnL-F region in Ligularia paradoxa, duciformis is distributed in Yunnan, Sichuan, Hubei, Gansu, and Ningxia provinces, China (Liu. 1989: Liu et al., 1994). species in geographical distribution. The two putative parents are sympatric The somatic chromosome numbers of the three taxa are identical (2n = 58). This number also agrees with reports from other species such as Ligularia hookeri (C. B. Clarke) Hand.-Mazz.. L. Hand.-Mazz.. L. hodgsonii Hook. (Franch.) Hand.-Mazz. (Gong et al.. vellerea (Franch.) and L. kanaitzensis 2001). the putative. nothospecies is most likely a natural Therefore. hybrid diploid. Both putative parents have displaved normal meiotic processes, producing seeds that germinated, suggesting that it is possible for hybrid- ization to take place between L. paradoxa and L. duciformis. ISSR markers. are inherited. in a dominant or codominant Mendelian pattern (Gupta et al... 1994: Tsumura et al.. 1996) scored as diallelic. and n w absent. The absence of a band is interpreted a primer divergence or loss of a locus through > deletion of the simple sequence repeat site or chromosome rearrangement (Wolfe et al.. 1998). In Lo ISSR bands were amplified from three taxa using four primers. Twenty- the present study. a total of six bands were common to both putative parents, and unique bands were distributed: five bands for Ligularia paradoxa and 16 bands for £. duciformis. L. duciformis. and Variable site Taxon (no. of individuals) Voucher and origin 152 214 191 GenBank accessions L. duciformis (20) X. Gong 22355 (KUN) I T ( 001044209 L. paradoxa (20) X. Gong 22356 (KUN) ( C \ DO 104430 L. Xmaontushanensis (16) Y. Gong 22357 (KUN) ( C A DO 104431 Volume 95, Number 3 008 493 A Ligularia Hybrid from Yunnan, China No additional bands were observed for the putative natural hybrid, but all 46 common bands occurred in the putative natural hybrid. Therefore, ISSR markers clearly indicate that the putative natural hybrid shares a relationship with both £. paradoxa and L. duciformis. Judging from the chromosomal behavior in meiosis and the absence of seed vigor, the putative natural hybrid may not produce sexual offspring and intro- gression is not likely among the three taxa. No asexual reproduction of the putative natural hybrid was noted in the field, but individuals of the putative natural hybrid at various ages were observed in the mixed population. Thus, hybridization between Ligularia paradoxa and L. duciformis has occurred more than when once as observed in their zone of contact, populations co-occur. In conclusion, the putative natural hybrid found on Mt. Maoniu is well supported as an interspecific hybrid between L. paradoxa and L. duciformis, and L. paradoxa is the female parent. TAXONOMIC TREATMENT Ligularia Xmaoniushanensis X. Gong & Y. Z. Pan, nothosp. nov. QLigularia paradoxa Hand.-Mazz. X COLigularia duciformis Winkl. Hand.- Mazz. TYPE: China. NW Yunnan: Ninglang Co., MO 4200 m, 16 July 2001, X. Gong 22357 (holotype, KUN!; isotype, MO not seen). E i gus pos e Ligularia paradoxa Hand.-Mazz k -M duciformis (C. Winkl.) Hand.-Mazz. genita, ab ae laminis ean magnitudine inter parentes mediis, pappis < oe vinaceis sed basi albis, seminibus abortivis acile differ 1.2-1.5 m tall, glabrous or sparsely pubescent above; basal and Herbs perennial; stems simple, stout, lower cauline leaves peltate, broadly ovate, 30—40 X 45-52 cm, 5- to 8-lobed, abaxially fulvous pubescent; 7 to 12 nerves, petioles 7- to 12-palmatipartite, 30 em, winged, clasping at base; upper leaves smaller. Inflorescence corymbose, peduncle 2-9 cm, densely fulvous pubescent, involucre campanulate, 10-12 X 2-3 mm, narrowly lanceolate, acute, externally densely fulvous bracts in 2 series, 8 to 10 bracts, pubescent. Ray florets absent; disc florets 5 to 7, bisexual, corolla tubular, 8-12 mm. Achenes cylin- drical, glabrous; pappi wine-colored, white at base. Chromosome number 2n = Distribution. Known only from the type specimen collected from Mt. Maoniu. hybrid occurs Habitat. The natural within a mixed population of Ligularia paradoxa putative and £. duciformis within an area of about 2 km” along forest about 4200 m a stream bank in spruce-fir altitude on Mt. Maoniu in Ninglang, Yunnan, China. When we collected the specimen for the first time in July 2001, we observed 122 individuals, 116 in October 2001, and 130 in early August 2002. IUCN Red List category. We propose that Ligularia Xmaontushanensis be recognized as Endangered (EN), according to a ed List criteria, because of the small population size (< 250 mature individuals) in a small distribution area (< 2 km”) (IUCN, 2001). Phenology. | Observed as flowering from July to August; fruiting from September to October. Paratype. CHINA. Yunnan: Ws Co., Mt. s 4200 m, : July 2003, Q. T. Zhang, Y. Z. Pan & Z. ing 082306 (KUN). Literature Cited Abbot, R. J. 1992. Plant invasion, interspecific a urs and the evolution of new plant taxa. Trends Ecol. E 401—404. Allen, C. A. 2002. Hybrid (Liliaceae), a new allote pen species from Washington speciation in Erythronium State. Syst. 7 26(2): 263-2 Aparicio, G. Albaladejo, M Porras & G. Cebollas. 2000 » evidence for natural hybrid in Phlomis : lybrid pu of the rare P. Xmargaritae. Ann. Bo. “Oo 85: 7-12. Arnold, M process. ae Rev. Ecol. Syst. 23: 237-261. . 1997, University pes New Yor Barton, N. H. . M. TM 1989. EE speciation and hybrid zones. nome 341: 992. Mei hybridization as an evolutionary Natural Hybridization and Evolution. Oxford Y 01. The ale u hybridization in 10: 551—568. M. T. Gabrielsen. 1996. A classic example of postglacial allopolyploid speciation re- Ecol Nilsson & cantar Mol. Brochmann, C., ' examined using R markers and nucleotide sequences: Saxifraga osloensis (Saxifragaceae). Symb. Bot. Upsal. 31: 75-89 Choi, K., J. S. Kim € J. H. Pak. 2001. Natural hybridiza- een Pseudostellaria davidii and Pseudostellaria 39- Biol. tion betwe palibiniana (Caryophyllaceae). Pl. Spec. 47. = D bs J. J. & J. Doyle. 1987. A rapid DNA een procedure for sl Em of fresh leaf tissue Phytoc hem. dam -]5. Ellstrand, m ^ DE & L. H. Rieseberg. 1996. Natl. Distibution E spontaneous plant hybrids. Proc. Acad. Sci. U.S.A. 93: 2 90—5093. Gong, X., Z. J. Gu, X. Y. Lu & C. karyotypes of seven pud in Ligularia. Yunnan. E 216-222. Gottlieb, L. Indica in plants. Ann. —446. Q. Zhang. 2001. The Acta Bot. . 1972. Levels of confidence in the analysis of 9: Gard. Missouri. Bot. Gun V. 1981. Hybrid Speciation. Columbia University z New York I ress, 494 Annals of the Missouri Botanical Garden Romero-Severson & J. Gupta, M., Y. S. Chyi, J. 1994. Pa ene of DNA markers. from ela diverse genomes using simple-sequence repeats. Theor. Appl. Genet. 89: 998—1006 Handel-Mazzetti; H. 1938. Die Chinesischen Arten der 69: 95- Molecular systematics lybridization in a British n v Gattung Ligularia. Bot. Jahrb. Syst. . Heredity ~ B < 1907. The (or uelit of China. Quart. J. Taiwan Mus IUCN. 2 je N Red List Categories and Criteria Version 3.1. Prepared by the IUCN Species aris Commission. IUCN, C pes a rland, and Cambridge, United Kingdom. Les, D. H. Philbrick. 1993. Studies of hybridization and c osten number variation in ue angiosperms: Evolutionary implications. Aquatic Bot. 44: 181-228 Levin, D. A. 1983. Poly do oe novelty in flowering plants. Amer. Naturalist 122: Liu, S. W. 1985. n Ligularta (L.) ( — | 989. Ton Vol. 791 ). à „l laxonomic system of the genus , Harbin. 5(4): 63-80. Popularis Sinicae, . Beijing. 1994. Origin, evolution, and Acta Phytotaxon. Sin. T au. PN of 20 (L.) Cass. 32(6): 51 ae Padgett, D. J., D. H. Les & G. E. Crow. Amer. J. er 1468-1476 Rieseberg, L. A 995 Old wine Amer. J. Bot. 82: in new n . 1997. Hybrid origins of plant species. Annual Rev. Ecol. Syst. 28: 359-380 1998. Evidence for the hybrid origin n of Nuphar Xrubrodisca (Nymphaeaceae). . The role of hybridization in evolution: 44-953. Sang. T.. D. J. Crawford & T. F. Stuessy. 1997. Chloroplast DNA phylogeny, reticulate ds and biogeography of Paeonia (Paeoniaceae). Amer. 7 Bot. 84: 1120-1136. Soltis, P. S.. D. E. Soltis & P. G, 1991. Allozymic chloroplast DNA analyses of M in Polystichum zd. The origins of Polystichum aa 16: 245-250. (Dryopteridaceae cume and Polystichum se deri Syst. Dot. Soltis. D. I Soltis . Molecular data and the dynamic nature S E Crit. Rev. PL Sei. 12: 243-273. Stebbins, G. L.. Jr. 1959. The se o NE in Amer. Philos. Soc. 2: 51. evolution. Proc. ————. 197 P a E Plants. sd Acn Pabien Ltd., M. 1996. The allopolyploid origin of Spiranthes , Pongkongensis (ERE iceae). Amer. J. Bot. 83: 252-260. berlet, P., L. Gielly, G. Paulon € 1991. ae aa primers it amplification. of "ies non-coding Biol. 17: 1105-1109. ~ romosome | ondon. Sun, 1 ). he Bouvet. regions of chloroplast DNA. PL Mol. Tsumura, Y., K. Ohba inheritance of inter-simple sequence phisms in Douglas fir (Pseudotsuga menziesii) i nd pe Theor. Appl. Genet. 92: 40—45. Wendel, J. H ). New world te crap cottons contain old world DE Proc. Natl. Acad. Sci. U.S.A. 86: 4132 4130. Wolfe, A D., Q. Y. Xiang & 5. R. Kephart. 1998. Assessing (Sero- phulariaceae) using hype rvariable intersimple sequence repeat (ISSR) bands. Mol. E se : 1107-1125. Zhang, Q., Y. Liu € Sodmergen. 2003. Examination of the rola DNA in male re a. cells to determine die pep d for En PaE in 295 angio- 941-951. hybridization in natural populations of Penstemon sperm species. Pl. Cell Physiol. A REVIEW OF THE GENUS PYROSTEGIA (BIGNONIACEAE)'? Amy Pool” ABSTRACT Pyrostegia C. Pres two leaflets a sl is a genus of four species in the tri nd a terminal tendril, e campanulate calyc stamens, t compressed, linear capsules apparently hummingbird pollinated and have flowers 1 narrow tubular- MR a corollas. The fi c E O = v E Un e =g e 3 B a = a E > © S EX © X 3 EN & =] — 5 Ur O A eN aat dehisce o to the septum, and thin, bialate seeds. us are = eS a) $ E = Z s a i p jæ e Bignonieae. All the species are lianas with EA inis with narrow | corollas with lobes valvate basal y in bud, very similar in appearance with re Gav uL JM ap eM. ornamental, b nonia iene / ll. B. nee ll Rusby, P. cinerea illose lassl., and Tynanthus igneus Barb. Rodr. A key to the species, species descriptions, and a species distribution map are a and the relationships of the species are discussed. Key words: Arrabidaea, Bignoniaceae, Bignonicae, hummingbird pollination, Pyrostegia. Pyrostegia C. Presl is a small but diverse genus of four species in the tribe Bignonieae (Bignoniaceae). venusta (Ker Gawl.) Miers frequently cultivated throughout the tropics. The genus is native to South America, with P. Like most members of the tribe Bignonieae, Pyrostegia species are lianas with the terminal leaflet EB of their compound leaves usually modified as : tendril, and the fruits dehiscent parallel to the septum. AII multiples of four phloem wedges (Santos, the species of Pyrostegia stems with 1995), eaves with two leaflets and a terminal tendril, calyces have — campanulate with five minute denticules, fairly narrow, moderately thick-textured corollas with the narrow lobes valvate at their bases in bud, four exserted stamens, compressed, linear capsules, and thin, bialate seeds. Three of the species are apparently hummingbird pollinated and have flowers that are very similar in appearance: red-orange (rarely yellow), narrow tubular-infundibular corollas with the stamens the However, two of these species (P. dichotoma Miers and P. fields absent, tendrils apically trifid, and inserted below the middle of corolla tube. ex K. Schum. venusta) have nodes with elandular e the corolla tubes externally glabrous, while the thirc species (P. cinerea Bureau ex K. Schum.) has nodes with glandular fields present, simple tendrils, and ^ densely pubescent corolla tubes. The presence or absence of interpetiolar glandular fields is a useful character in recognizing genera in the Bignonieae (Sandwith, 1938; Seibert, 1948; Gentry, 1993) and has frequently been used as a generic criterion. However, Gentry (1980) warned that the presence of inter- petiolar glandular fields can be variable even within a species and should not be used alone as an absolute generic character. Similarly, leaf tendrils can be an extremely useful generic character, with some genera having only simple tendrils, while other genera have 1916: 1993). However, there are genera and even species where only trifid tendrils (Urban, Gentry, both types of tendrils are found (Gentry, 1980), so again, this character should not be used as an absolute in defining genera. Finally, corolla pubescence can also be an important character for differentiating genera (Gentry, 1980) and is frequently employed in keys to genera (Sandwith, 1938; Gentry, 1982), but again, it can be variable within a genus. Nonetheless, retaining P. cinerea within the genus Pyrostegia is somewhat suspect, and alternative placement is discussed in the taxonomy section of this paper. The fourth species in the genus (P. millingtonioides Sand- ' This paper is number Bignoniaceae made by Alwyn H. FTG, providing digitized images. Gentry. G, GH, GOET, L, M, rs NY, S, ee U, thank W. UPS, US, W . Stevens and L study and who entered a the National Science Foundation (grant : ? Missouri Botanical Garden, P.O. B doi: 10. 341 7/2003090 ANN. Missouri Bor. the staff of the following herbaria for providing loans of herbarium n imens: AAU, úcia Lohmann for th manuscript, Bee Gunn for preparing the e was d. and Duan Bills for help in preparing the species distribu give particular thanks to the technical staff at the Missouri Botan 14 of the Gentry Invitation Series, in acknowledgment of the contributions to the study of the BM, BR, C, CAS, C V and Timothy Harris, Peter ie and Lúcia La P Rae for e a c = advice and helpful comments on ition "o I val Garden whose work made the specimens available for large perce i Us of D specimens into m TROPICOS database. Financial support was provided by 98). ox 299; St. Louis, Missouri 63166-0299, U.S.A. amy.pool@mobot.org. Garb. 95: 495-510. PUBLISHED ON 23 SEPTEMBER 2008. Annals of the Missouri Botanical Garden with) has nodes with glandular fields absent, tendrils apically trifid, the elabrous, but has flowers that are quite different in and corolla tubes externally appearance from the other three species. It has strongly fragrant, white, salverform corollas with stamens inserted near the mouth of the corolla tube and is probably moth pollinated. Gentry (1990a: 125) found that switches from one pollinator to another within a genus are relatively rare and suggested that “such changes appear mostly to be associated with major genetic reorganization that accompanies generic differentiation." Placement of P. millingtonioides is discussed in more detail in the taxonomy section of this paper. HISTORY C. Presl established the genus Pyrostegia (Presl, 1845). genus and the type species, Bignonia ignea Vell. thorough description of the P. venusta), which he transferred into his new genus. He provided a However, he made no mention of the aestivation of the corolla lobes, a character made much of by other authors. De Candolle (1845) had treated B. ignea as a synonym of B. venusta Ker Gawl., placing emphasis in his description on the valvate aestivation of the corolla lobes. Miers (1863: 188) also emphasized the corolla lobe aestivation, describing the lobes as “almost valvate in aestivation, a feature quite peculiar to the genus." Miers (1863) transferred B. venusta and B. longiflora Cav. to Pyrostegia and listed 10 additional names as new species, but did not provide descrip- tions for the new taxa. Schumann (1894) established the new genus Macranthisiphon Bureau ex K. Schum., into which he transferred B. longiflora. Monotypic, Macranthisiphon is similar to most of the species of Pyrostegia in having narrow, orange-red corollas with narrow corolla lobes and exserted stamens; it differs in the corolla lobes being imbricate and in having large (1894) published two more species of Pyrostegia, P. cinerea and P. dichotoma, in his treatment of Bignoniaceae for Die the valvate aestivation of the corolla lobes as an early lead in his key to genera in the tribe Bignonieae. In their treatment of Bignoniaceae for Flora Brasiliensis, Bureau and Schumann (1896, 1897), treated P. ignea (Vell.) C. Presl as a synonym of P. venusta, added the foliaceous pseudostipules. Schumann also Natürlichen | Pflanzenfamilien, and used new species P. tubulosa Bureau € K. Schum. (= P. dichotoma), and again emphasized the valvate aesti- vation of the corolla lobes by using this character to separate Pyrostegia and Glaziova Bureau from other genera of the Bignonieae with 3(2 floral transferred B. tecomiflora Rusby D-foliolate leaves, (1916) — P. dichotoma) filiform tendrils, and dises. Urban to Pyrostegia. Sandwith added P. millingtonioides to the genus and, while commenting that his new species was unusual in the genus both in corolla color and shape and in the level of insertion of the stamens, nonetheless expressed that it could be placed in this 1962: 465). This was the last addition to the Macbride (1961) placed P. dichotoma in synonymy with P. venusta in the Flora of Peru, but Gentry (1982) recognized the two as close but separate species in his genus “with some confidence” (Sandwith, genus. treatment of Bignoniaceae for the Flora de Venezuela. ANATOMY Santos (1995) studied 31 genera of Bignonieae and divided them into four groups based on the type of cambial variant. Three species of Pyrostegia, P. cinerea, P. dichotoma, and P. millingtonioides, were included in her study. Pyrostegia falls into her group 2, which is characterized by possession of stems with multiples of four phloem wedges. Santos found all the genera in her group 2 to be so similar that it was difficult to separate them anatomically. All the genera in this group were found to have septate fibers, cylindrical stems and absence or rare occurrence of perforated ray cells, and scanty to vascicentric paratracheal parenchyma. Other genera in group 2 are Amphilophium Kunth, Anemopaegma Mart. ex Meisn., Clytostoma Miers ex Bureau, Cydista Miers, Distictis Mart. ex Mansoa DC., Mussatia Bureau ex Baill., Phryganocydia Mart. Meisn., Haplolophium Cham., Periarrabi- daea A. Samp., ex Bureau, Roentgenia Urb., and Tanaecium Sw. PALYNOLOGY The pollen of Pyrostegia is 3-colpate (or sometimes 4-colpate in P. venusta) and finely reticulate (Urban, 1916; Sandwith, 1962; Gentry & Tomb, 1979), a pollen type common in the tribe Tecomeae and found in several probably unrelated genera in the Bigno- nieae (Gentry & Tomb, 1979). similar pollen in the Bignonieae include Roentgenia (ca. 4-colpate, and exine somewhat warty), Potamo- ganos Sandwith, Stizophyllum Miers, Pachyptera DC. ex Meisn., and some species of Tanaecium (Gentry & Tomb, 1979). In Pyrostegia, only P. dichotoma and P. venusta have been examined using SEM (Gentry & Tomb, 1979: fig. 3). ther genera with O” m CYTOLOGY Goldblatt and Gentry (1979) found a base number of x = 20 for all but three of the 23 genera of the tribe Bignonieae known cytologically. Their count for Volume 95, Number 3 2008 Pool Review of Pyrostegia (Bignoniaceae) Pyrostegia cinerea (voucher: Gentry 12773) of 2n = 40 is consistent with this general tribal base number. Joshi and Hardas (1956) reported a count of 3n = 60 for P. ignea (= P. venusta) from a non-fruit-setting population cultivated in India. Other species (and populations) of Pyrostegia have not been investigated cytologically. REPRODUCTIVE BIOLOGY Flowers of Pyrostegia cinerea, P. dichotoma, and P. venusta are typical of those pollinated by humming- birds: odorless and the corolla usually bright red- orange, of fairly thick texture, with a narrow tube and wider mouth, and more or less glabrous internally (Gentry, 1980. 1990a), while the aromatic flowers of P. millingtonioides, with white, narrow, tubular, and thickly textured corollas, are probably moth pollinated (Gentry, 1980, 1990a). Hummingbirds have been reported visiting P. venusta (Gobatto-Rodrigues & Stort, 1992; Galetto et al., 1994). Gobatto-Rodrigues and Stort (1992) found P. venusta to be self- compatible with facultive outbreeding. The humming- birds Eupetomena macroura Gmelin and Phaethornis pretrei Lesson & Delattre were reported as visiting and effectively pollinating P. venusta, while a number of insects were found to visit the flowers as nectar or pollen thieves. Galetto et al. (1994) reported the hummingbirds Chlorostilbon aureoventris Orbigny & Lafresnaye and Sappho sparganura Shaw visiting flowers of P. venusta. Their study illustrated the pattern of nectar secretion in P. venusta. Gusman and Gottsberger (1996) compared the usual orange- flowered P. venusta with the rarer yellow-flowered individuals in respect to floral morphology, nectar composition, carotenoids, and flavenoids, and theo- rized about how these differences could affect nectar robbing by bees and visits by hummingbirds. Gentry (1990a) hypothesized that all moth- and hummingbird-pollinated species of Bignoniaceae bloom for a period of several weeks to a few months, building up to a distinct peak of flower production followed by a gradual decrease in number of flowers. He referred to these as cornucopia species. Galetto et al. (1994) and Gobatto-Rodrigues and Stort (1992) confirmed this pattern for the populations of Pyroste- gia venusta observed in their studies, but with the flowers being produced over a time span of several months (Galetto et al., 1994) or generally for about six onths with a two-month peak (Gobatto-Rodrigues & a. 1992). particularly useful in confirming the patterns of is herbarium study did not prove phenology. Only two flowering specimens of millingtonioides were seen, these collected in July and August. Six flowering collections of P. cinerea and May (1), June (2), July (2), November (1). Pyrostegia venusta has been seen in were observed: flower throughout the year (except December) but peaks between June and September. Pyrostegia dichotoma is more complicated. For its whole range, it has been seen in flower all year, but in Colombia only in January; in Guyana and Suriname in February, September, and October; in Venezuela peaking in March; and in Peru, Bolivia, and Brazil peaking from June to October. Economic USES Pyrostegia venusta is a very popular ornamental, cultivated throughout the tropics. Menninger et al. (1970) included it (as P. ignea) on his list of the most beautiful flowering climbers in the world and ranked it as the most popular of all in the tropics. MATERIALS AND METHODS Gentry established a private database that compiled label information from herbarium specimens at MO and other herbaria that he examined personally. Gentry’s private database has been incorporated into the Missouri Botanical Garden database-management sys- tem TROPICOS, , which now contains label information for all Pyrostegia specimens housed at MO in addition to those received on loan from AAU, B, BM, BR, C, CAS, CM, F, FTG, G, GH, GOET, L, M, MICH, NY, S, TEX, U, UPS, US, and W. Specimens determined by this author were selected from this database to generate the Index to numbered exsiccatae (Appendix 1) and the distribution maps, the i Dots on the maps (Fig. 1) represent a subset of the specimens latter with the assistance of Duan Bills. examined by this author, including specimens with geographic coordinates provided on collection labels and a set of specimens, for which the author estimated the geographic coordinates, selected to illustrate the full geographic range for each species. TAXONOMIC TREATMENT Pyrostegia C. Presl, Abh. Kónigl. Bóhm. Ges. Wiss., ser. 5, 3: 523. 1845. TYPE: Pyrostegia ignea (Vell.) C. Presl [Based on Bignonia ignea Vell.]. Lianas; branchlets subangulate with 6 to 8 inconspicuous ribs, lacking, or less frequently with, interpetiolar glandular fields, not lenticular; pseudo- stipules usually inconspicuous, small and subulate. Leaves 2(3)-foliolate, 1 apically trifid or less frequently simple, often with a tendril; tendril filiform, deciduous. Inflorescence a terminal or axillary pani- 498 Annals of the Missouri Botanical Garden cle; bracteoles minute, subulate, and inconspicuous; flowers usually numerous. Flowers with the calyx campanulate, truncate to undulate, the ribs extending apically to form 5 small denticules, these sometimes corolla tubular-infundibular, inconspicuous; narrow moderately thick, red-orange (rarely yellow), and odorless or, less frequently, salverform, thick, white, and fragrant; tube externally glabrous, or less frequently densely pubescent throughout, internally glabrous except at, and sometimes, below the level of insertion of stamens; lobes narrow, + valvate at base in bud, pubescent at least along margins; stamens 4, exserted, anthers glabrous, pendulous, dithecal, the thecae subparallel divaricate; staminode present, small; disc annular-pulvinate; ovary linear-tetragonal, lepidote, bilocular, placentation axile, ovules biseriate stigma exserted, less n each locule; broadly, or frequently, narrowly, bilamellate. Fruit compressed, linear, the valves parallel to the septum, smooth, medial vein sometimes slightly raised: seeds thin, bialate, wings brown with hyaline margins. KEY TO SPECIES OF PYROSTEGIA I. Corollas white, narrowly salverform, ca. 3 mm wide at mouth of tube, lobes elliptic-oblong, at least 1/2 as wide as long, 0.6-0.7 em long: stamen filaments 0.15-0.2 em long; interpetiolar ridge present Ces ku pu ge E ers Bie milingloniórdes I". Corollas red-orange (rarely yellow), narrowly tubular-infundibular, 6-13 mm wide at mouth of tube, lobes oblong, less than 1/2 as wide as long. 1-1.8 em long; stamen filaments 2.6-5.5 em long; p interpetiolar ridge absent. N Corolla tube externally densely red-puberulent; leaflets with pubescence densely silvery- o ferruginous-lomentose, adaxially glabrescent; tendrils simple; interpetiolar glandular fields present; stigma lobes subulate ..... P. cinerea Corolla glabrous to densely N tube externally glabrous; leaflets short-pilose; tendrils apically trifid; interpetiolar glandular fields stigma lobes orbicular, ovale, or broadly oblong. 3. Staminodes inserted at same level as stamens (rarely to 0.4 cm above insertion trichomes mi- of stame ns); inflore 2SCence nule, appressed to ascending; inflores- cence generally open; mature fruits gen- erally drying brown to dark brown rare ly 0.8) 3. Staminodes inserted (1 2-1.6 cm above insertion of stamens: Sae trichomes longer, initiating at right angles to surface; inflorescence generally crowded with calyces overlapping in dried. speci- mens: mature fruits generally drying with olive « P. venusta |l. Pyrostegia cinerea Bureau ex K. Schum. in Engl. Nat. Pflanzenfam. 4(3b) 223. 1894. TYPE: Brazil. Minas Gerais [cultivated at Rio de Janeiro: Quinta da Boa Vista]. s.d. [28 Feb. 1892]. A. Glaziou 14124 (lectotype, designated here, [barcode P00481542] not seen, P digitized imagel: duplicates, P [barcode P00481543] not seen, P digitized image!, C!, photo F neg. 22143 Vellosia, ed. 2, 1: 50, , non Pyrostegia ignea (Vell.) C. Presl, Brazil. Amazonas: Arvensis ad ripas Kio Manaos.” ser. P pn 10 in Rodrigues, Vellosia, ed. 2, vol. 3 (lec NM) a igneus Barb. Rodr., 1: 50, 3: ser. 2: tab. 10. 189 1845. TYPE: Negro. prope Barbosa type, designated n Ib!). here, tab. 10 in Barbosa c Branchlets silvery- to ferruginous-tomentose, in- terpetiolar glandular fields present, interpetiolar ridge absent. Leaves 2-foliolate with a simple terminal tendril; petioles 0.5-2.2 cm, tomentose: .5-1.2 em; leaflets 4—12.5 nous, 4 or 5 pairs of lateral veins prominent below, vetiolules 0 oblong or ovate, | 8 subinequilateral, 2.5-6.5 em, membra- silvery- to ferruginous-tomentose and adaxial sur- face eventually glabrescent, scattered pellucid- lepidote, base rounded, apex acute-mucronulate or briefly acuminate-mucronulate. Panicle terminal or terminal and axillary, narrowly elongate with caly- ces not overlapping in dried specimens, | to 3 times branched, peduncle, rachis, aad bracteoles silvery- to ferruginous-tomentose; calyx excluding denti- cules 4-5 4-5 mm at apex, silvery- to ferrugi- nous-tomentose, occasionally with few lepidote scales; corolla narrow tubular-infundibular, orange (rarely yellow); tube 2.5-3.5 cm long, 2.4-3 mm wide at base, 7-10 mm wide at mouth, internally sericeous al and below insertion of stamens, externally densely red-puberulent; lobes oblong, 1— 1.5 X 0.3-0.4 cm, internally red-puberulent apical- ly s marginally, externally densely red-puberu- stamens and staminode inserted tube, filaments 2.7-3.2 cm. lent; approxi- mately same level in corolla 1-1.2 cm from base of tube, stamen thecae porrect-divaricate, 2.3-3 mm, staminode 6-12 mm; dise 1.2-2 X 12-2 mm; pistil 4—5.5 em, ovary 2— 2.8 X 0.6-1 mm, stigma lobes subulate. Capsule 17.5-19 X 0.9-1 cm, midvein slightly raised, base and apex acute; seeds 0.7-0.9 X 2.2-2.5 cm (Fig. 2C-F). glabrous, drying brown, Brazil, (white sand areas) near Manaus, State of Amazonas, Distribution. disturbed areas or campinas presumably at low elevations (Fig. 1). Phenology. Flowering May to July and November; ruiting November. Volume 95, Number 3 Pool 499 Review of Pyrostegia (Bignoniaceae) 10°0'0"= 0°0'0"= -10*0'0*4 -20°0'0"— ABO OO [| -80%0'0" -70?0'0" Figure 1. cinerea Bureau ex K. Schum Gawl.) Miers (A). Schumann (1894) published Pyroste- in his key to Discussion. gia cinerea Bureau ex K. Schum. Pyrostegia without any indication of type collection. It is assumed here that the species was included in Schumann's key based on information from Bureau and that the type collection was at P. Two collections at P are annotated in what I believe to be Bureau's tc handwriting. A. Glaziou 14124 (barcode P00481542 and with — is annotated by Bureau as "Nouveau," potential specific epithets (discolor and cinerascens) with a line drawn through them and finally “cinerea o discolor." Jobert 452 is annotated by Bureau as I -60%0'0" | ' -50%0'0" -40°0'0" Distribution of species of Pyrostegia, excluding cultivated and apparently naturalized specimens. Pyrostegia . (©), P. dichotoma Miers ex K. Schum. e). P. millingtonioides Sandwith (O), and P. venusta (Ker . “Pyrostegia cinerea n. sp. B concolor." The Glaziou collection is chosen here as the lectotype in part because | suspect that Bureau would have been more likely to indicate the type variety as variety & versus variety B, and because duplicates of Glaziou 14124 are extant; in fact the sheet at C also seems to be annotated in Bureau's hand. Unfortunately, while P. cinerea is endemic to Amazonas, Brazil, the lectotype is labeled as from Minas Gerais. However, Glaziou is known to have changed localities (and sometimes also the collection) on some specimens (Wurdack, and/or date of 1970). collector, collection number, 500 Annals of the Missouri Botanical Garden — Figure 2. A, B. Pyrostegia millingtonioides Sandwith (tori G. S. Pinheiro 19, MO). ower bud. —B. Longitudinal section of corolla with androecium, style, and stigma. CF. Pyrostegia cinerea Bureau ex K. Ss hum. (from A. Ducke 532, MO). —C. Flowering branch. —D. Stem node with glandular fie ld. E. Style and stigma. —F. Longitudinal section of corolla with androecium. ~ == Volume 95, Number 3 2008 Pool Review of Pyrostegia (Bignoniaceae) The duplicate of A. Glaziou 14124 at C is without locality information. The second collection of A. Glaziou 14124 at P (barcode PO0481543) is labeled as made from cultivated material at Quinta da Boa Vista, Rio de Janeiro, collected 28 Feb. 1882. Glaziou (1911) reports both his numbers 14/24 and 19666 (the latter cited by Bureau & Schumann, 1897, as 19665, from Minas Gerais) as cultivated at Quinta da Boa Vista, S. Hio de therefore assumed that the lectotype A. 14124 (barcode P00481542), while labeled Minas A. Glaziou 14124 (barcode P00481543), and was made from cultivated material Christováo, Janeiro. [t is Glaziou Gerais, is a duplicate of e Janeiro. A. Glaziou P00481543) is it was formerly in Glaziou's private 1907, at Quinta da Boa Vista, Rio d 14124 candidate as (barcode not a lectotype herbarium, not donated to P until and not annotated by Bureau. According to Stapf (1909: 225), made his descriptions and illustrations from living Barbosa Rodrigues — plants with an “almost complete absence of speci- Stafleu and Cowan (1983: 829) were unable to locate any types or other herbarium specimens of that the important protologue materials for Barbosa Rodrigues? mens." Barbosa Rodrigues and concluded most new laxa were to be found in his published and unpublished drawings. Barbosa Rodrigues” excellent 10 (Barbosa Rodrigues, 1891b) is thus chosen here as the lectotype of Tynanthus igneus. plate, ser. 2, tab. The quality of the illustration and description leave no doubt that 7. igneus is synonymous with Pyrostegia cinerea, a conclusion previously reached by Bureau and Schumann (1896: 198). While T. igneus Barb. Rodr. (Barbosa Rodrigues, 1891a) has priority over P. cinerea Bureau ex K. Schum. (1894), P. ignea (Vell.) C. Presl (1845) blocks the transfer of the Barbosa Rodrigues name to Pyrostegia. The simple tendrils, interpetiolar glandular fields, and externally pubescent corollas of this species are unexpected in the genus ppc where they are otherwise not known (Fig. 2C, D). These characters suggest an affinity. with ja genus Arrabidaea DC. (1968: 405-406) established a set of 'haracters, which when found in combination, could Sandwith 2 suggest inclusion in the genus Arrabidaea: “branch- lets terete or subterete, not definitely tetragonous; tendrils simple; corolla tomentose outside, at least on the limb, nearly always pink, purple or white, very rarely creamy yellow, never full yellow; disk conspic- uous; anther thecae parallel, divergent or divaricate, glabrous, not curved or contorted; pollen-grains simple, 3-furrowed; ovary oblong; ovules 2 to 4 seriate in each loculus; capsule narrow, linear-oblong, compressed. Glandfields may be present or absent at the nodes, though commonly present; trichomes of the indumentum may be short and simple, long and multicellular, sometimes gland-tipped, branched and 'dendroid" (sometimes forming a dense fur), but not stellate; the foliage in two species is for the most part bipinnately 2-3 ternate; foliaceous pseudosti- pules are rarely, but sometimes present; the inflores- cence is often a pyramidal thyrse, but may be short and axillary and is sometimes found on the old wood; the calyx is extraordinarily variable in shape, size and texture, and in the presence or absence of teeth or lobes; the corolla varies from hypocratiform in species with very small flowers to campanulate-funnelshaped in most of those with larger flowers; the stamens are usually included but in one species the anthers are conspicuously exserted; the capsule is usually smooth, ut in two widely distributed species is densely prickly-tuberculate; the seeds usually have membra- nous wings, but in certain water-borne species are wholly corky." Pyrostegia cinerea differs from this typical picture of Arrabidaea in its large, orange, narrow tubular-infundibular corolla and exserted stamens. Sandwith's concept of Arrabidaea included no species with true yellow or orange corollas; A. elegans (Vell. A. H. Gentry does have though Gentry (1975) suggests that this and A. bilabiata (Sprague) Sandwith however, yellow to orange corollas, might be better placed in a separate genus. Arrabi- daea lauta Bureau & K. Schum. is very similar to P. cinerea and was initially placed in Pyrostegia by Miers (1863). Not pubescent, narrow tubular-infundibular corolla with only is it similar florally, with a oblong lobes and exserted stamens, but it is also similar vegetatively, with interpetiolar glandular fields This might suggest that A. lauta would fit more comfortably and 2-foliolate leaves with a simple tendril. in the genus Pyrostegia, rather than supporting a relationship between P. cinerea and the genus Arrabidaea. However, A. lauta differs from species of Pyrostegia in the imbricate aestivation of its corolla lobes, more delicately textured, red corolla, strongly recurved anther thecae, and relatively broader calyx. he pollen of P. cinerea was examined by Urban 1916) using light microscopy and found to be more similar to that of P. dichotoma and P. venusta, that is, 3(or 4)-colpate and finely reticulate (Gentry & Tomb, 1979), than to species of Arrabidaea, 3-colpate and more or less psilate or smooth (Gentry & Tomb, 1979). Anatomically, P. cinerea fits well with the other two species of Pyrostegia examined (Santos, 1995) into Santos” (1995) group 2, characterized by stems with multiples of four phloem wedges in transverse section. Species of Arrabidaea belong to Santos’ (1995) group 1, characterized by stems with only four phloem wedges in transverse section. Crystals were found in the vessels of most of the species of 502 Annals of the Missouri Botanical Garden Arrabidaea examined, but not in any of the species of Pyrostegia. Pyrostegia cinerea is maintained for the time being in Pyrostegia primarily on the basis of the corolla lobe aestivation, valvate for the basal half to two thirds. This appears to be an unusual character state in the Bignoniaceae. The narrow tubular-infundibular corol- la shape found in Pyrostegia appears to have evolved repeatedly in the family and in the tribe Bignonieae in groups pollinated by hummingbirds (Gentry, 1974, 1980); however, this does not seem to be true of the corolla lobe aestivation. Gentry (1990a) lists the following genera of the Bignonieae as strictly hummingbird pollinated: Dolichandra Cham., Frider- icia Mart., Gardnerodoxa Sandwith, Macranthisiphon, Martinella Baill., All of these have the corolla lobes clearly imbricate in bud. and Piriadacus Pichon. Additional ape imens examined, BRAZIL. Northern Bra- > not seen, P digitized i Image). Amazonas: Manaus. Schwacke 111-278 (GOET); Manaus, lg. da ‘achoeira nad do Tarumã, J. Chagas s.n. INPA-1198 (U); da de Aleixo, NE Manaus: ca. km to Leper Colony, A. Lasseigne P22582 (MO); along Aleixo-rd., P. J. | Mun & i zil, Jobert Maas 326 (MO); ig of Manaus. rd. to | t Gentry 12857 (MO); Manaus, loco Estrada do Aleixo, h Ducke 532 (MO), 5. uy (US); grounds of INPA at Manaus, A. H. Gentry & G. T. Prance 11198 (MO); vic. of Manaus rd. toward Rio Negro, 10 km N from Mood on Estrada Aleixo, A. H. Gentry et al. 12773 (MO); Km 14 on Estrada Mauá, G. T. Prance & Prance 21012 (NY); Reserva Florestal Ducke, Km 26, Floresta de ( Lohmann et al. 34 (MO). Rio de Janeiro: |Culti Quinta da Boa SR S. Christovao|, A. Glaziou 19666 (C, P digitized image). Manaus-lItac ur vampinarana, £. G. ivated at ) nol seen, 2. Pyrostegia dichotoma Miers ex K. Schum. in Engl, Nat. Pflanzenfam. 4(3b): 223. 1894. TYPE: Peru. San Martín: Tarapoto, 1855-1850, R. Spruce 3930 (lectotype, designated here, G!, photo F neg. 262031; isotypes, BM! BR! G-DC not seen, GH! K not seen, NY!, P not seen. WD). Bignonia tecomiflora Rusby, Mem. Torrey Bot. Club 6: 101. 1 e oe (Rusby) . Deutsch. Bot. Ges. 34: 746. IO, lees hag K. Schum. ex Urb., 1916. TYPE: Bolivi a. p Paz: Mapiri, o 1892, M. Bang 1510 (lectotype, designated he NY Pec 278050]: duis ae A!. BM!, MICH! MO! 204088 |!, Y [barcode Branchlets sparsely lepidote and/or sparsely to densely covered with small curved-ascending to appressed trichomes, or glabrous, interpetiolar glan- dular fields absent, interpetiolar ridge absent. Leaves trifid terminal 2-foliolate, often with an apically tendril, or leaves 3-foliolate; petioles 1.5—4 cm, with adaxial canal; petiolules 0.5— 2.5-12 chartaceous (membranous), 3 to 5 curving trichomes in 3 em; leaflets ovate, slightly subinequilateral, X 1.4-6.7 cm, pairs of lateral veins prominent abaxtally, sometimes with trichomes along midvein on adaxial surface and/or at base of abaxial surface, pellucid-lepidote, often especially conspicuous abaxially, with large elands in the axils of lower lateral veins, base rounded briefly r truncate, apex acuminate-mucronulate or acuminate-mueronulate. Panicle usually terminal, sometimes axillary, open with calyces not overlapping in dried specimens, 2 to 4 times branched, peduncle, rachis, and bracteoles nearly glabrous to densely covered with minute, ascending-curved to appressed trichomes, sometimes also lepidote; calyx excluding denticules 4.5-8 X 4.5-10 mm at apex, with sparse lepidote scales, sometimes minutely puberulent, apex ciliate; corolla narrow tubular-infundibular, orange or reddish orange; tube 3.5-7 cm long, 3-5 mm wide at base, 6-13 mm wide at mouth, internally sericeous al and below insertion of stamens, externally glabrous; 1-1.8 0.3—0.6 cm, and marginally, externally sometimes lepi- lobes oblong, puberulent apically dote; stamens and staminode inserted at approximate- ly same level in corolla tube (or staminode up to 4 mm above higher stamens), 1-3 em from base of tube, 3.05 gent or subparallel, 34.5 mm, staminode (sometimes stamen filaments .5 em, thecae slightly diver- developing an anther) 2-25 mm; disc 1.5-2 X 2- 3 mm; pistil 5—7 cm, ovary 3-7 X 1 mm, stigma lobes ovate, broadly ovate, or broadly oblong. Capsule (rarely 11)18-33 X 0.8-1.2 em, glabrous, drying brown, midvein apparent but not conspicuous, base seeds 0.6-1 2.6—4.5 cm acule, apex aristate: (Fig. 3C-F: Gentry, 1982: fig. 38). Distribution. Guyana, Suriname, Venezuela, Colom- Bolivia, and Amazonian Brazil; 30-1500 m elevation, generally found below 1000 m. It is frequently bia, Peru, reported to be found in disturbed areas, in savannas with islands of larger trees, or in dry, moist, or wet forest. The soil type is often described as sandy (Fig. 1). a Phenology. For whole range, flowering and fruit- ing throughout year. In Colombia flowering in January: in Guyana and Suriname in February, September, and October; in Venezuela peaking in March: and in Peru, Bolivia, and Brazil peaking from June to October. Venezuela: Barqui (Gentry, Tango (Xena 257); Peru: Paccha huasca Bolivia: Lluvio de Oro (Guillén et Common 1982). (Ferreyra 5070); al. 4106). names. Discussion. Miers (1863) published the name Pyrostegia dichotoma, without description, citing one collection, Spruce 3930 from Tarapota [Peru]. Schu- mann (1894: without directly specifying the type or any collections, Miers ` indirectly indicated Spruce 3930 as the 23) validly published P. dichotoma ub by citing “Pyrostegia dichotoma aust Ostperu, Volume 95, Number 3 2008 Pool 503 Review of Pyrostegia (Bignoniaceae) type. Bureau and Schumann (1897) cite the same collection as the only specimen of P. dichotoma. A number of specimens of this collection number were seen from a variety of herbaria (BM, BR, G, GH, NY, W, also at G-DC, K, and P, but these not seen), and it is assumed that Schumann also had a specimen at B, and that specimen was the holotype and is no longer extant. The specimen at G is chosen here as the lectotype. When Rusby published Bignonia tecomiflora, he was able to base his description upon the complete sets of Bang 1510 and 1596. Bang 1510 (NY. barcode 278056) is selected here as the lectotype based on its excellent condition and completeness: flowers, flower buds, capsules, and seeds are all present on this sheet. Gentry (1982) cited Pyrostegia tecomiflora and B. tecomiflora as synonyms under both P. dichotoma and P. The syntypes of B. tecomiflora have the inflorescences more crowded than is generally the venusta. case for P. dichotoma; however, the type of inflores- cence that generally found in P. dichotoma and not found in P. trichomes, minute and appressed, is venusta. The staminode, in one flower dissected, was at the level of the stamen insertion and, in the other flower dissected, 3 mm above the level of insertion, better fitting the pattern of P. dichotoma than venusta, in which the staminodes are inserted (rarely .2-1.6 cm above the stamens. The holotype of Pyrostegia tubulosa Bureau & K. Schum. (in Martius, Fl. Bras. 8(2): 231. 1897, Richard Schomburgk 969, from Kanuku, Guyana) is presumed to have been at B and destroyed. No isotypes were seen in this study or by Gentry (1982), and none are in the database of Schomburgk collections compiled by The Nationaal Herbarium Nederland (http://www. nationaalherbarium.nl/>). However. three specimens of Pyrostegia made in the Kanuku Guyana (Jansen-Jacobs et al. 369, 437, and Wilson- Browne 217) were studied and not found to be Mountains. of different from other P. dichotoma. Bureau and Schumann (1897) used the margin of the calyx, undulate versus truncate, to help separate P. tubulosa from P. dichotoma. However, the calyx margin is very variable in P. dichotoma, and the margin exhibited by the collections from Guyana fit well into that range. Bureau and Schumann (1897) also used the nature of the leaf apex, rostrate versus acuminate to somewhat obtuse, of specimens examined from Guyana did not show a o separate the taxa. — significant difference from those of other collections of P. dichotoma. The corolla described by Bureau and Schumann (1897), 6.8-7.3 cm that of Wilson-Browne 217, 6.5-7 cm long, are larger than the dissected flower of long, and those otherwise seen. However, Jansen-Jacobs et al. 437 had a total corolla length of 6.2 cm, well within the range for P. dichotoma. Based on these observations, P. tubulosa is treated here as synonymous with P. dichotoma (following Gentry, 1982 Pyrostegía dichotoma and P. venusta are very similar in general appearance. The most reliable p = ‘ay to separate the two is by the level of the staminode insertion. In P. dichotoma, the staminode is inserted with the stamens (Fig. 3F), and in P. venusta, it is inserted considerably (at least 8 mm) higher in the corolla tube (Fig. 3B). Bureau and Schumann 1897: fig. 98) and Lohmann and Pirani (1998: fig 10D) provide accurate and useful illustrations of the P relative position of the stamens and staminode in P. venusta. Pyrostegia dichotoma almost always has the inflorescence trichomes minute and appressed to ascending. while in P. venusta the trichomes are perpendicular to the surface at least at the base (Fig. 3A). Fabris (1965: fig. 29) provides an excellent illustration of the trichome type found in P. venusta. However, it is not uncommon for individuals of both species to have nearly glabrous inflorescences, particularly after the bracteoles have been lost. Additionally, one specimen was seen, Daly et al, 2271, which had trichomes of the type generally found in P. P. dichotoma. Pyrostegia dichotoma tends to have a venusta, but had all the other characteristics of more open, often elongate, inflorescence that is 2 to 4 times branched (Fig. 3D; Gentry, 1982: fig. 38), while P. venusta generally has a very dense panicle, which is often subcorymbose and either unbranched or 1 or 2 times branched (Bureau & Schumann, 1897: fig. 98; 1998: fig. 10A). However, some 1 as Yoshiizumi s.n. ( ESA- 21606 (from Lohmann & Pirani, specimens of P. venusta, suc 6363) (from Sao Paulo, Brazil), Harley Bahia, Brazil), and Hassler 11784 (from Central, Paraguay), can have very open inflorescences, while some specimens of P. dichotoma, such as Ducke 725 (from Amazonas, Brazil), Mathias € Taylor 6080 (from Loreto, Peru), and Gentry et al. 77514 (from Beni, Bolivia), can have fairly crowded ones. The two species do not appear to be sympatric. While P. dichotoma is known from Guyana, Suriname, Vene- zuela, Colombia, Peru, Bolivia, and Amazonian Brazil, P. venusta is native to Atlantic and southern Brazil, from Piauí to Rio Grande do Sul, southern Paraguay, and northeastern Argentina. The latter species is very widely cultivated and possibly naturalizes in some areas. Selected specimens examined. BOLIVIA. Beni: Prov. Vaca Diez, E side yeralta, J. C. Solomon 18 (MO). Pa Sorata, ang 1596 (syntypes o tecomiflora, A, BM, . K not seen, M, MICH. d NY, Pr ungas, E orocoro, 12 km NE of Caranavi, P Yu H. Gentry et al. 44347 (MO). Pando: Prov. Manuripi, 30 km 504 Annals of the Missouri Botanical Garden Figure 3. A, B. Pyrostegia venusta (Ker Gawl.) Miers (from E. W. Schupp 258, MO). —A. Flower bud. —B. Longitudinal section of corolla with androecium, style, and stigma. C-F. Pyrostegia dichotoma Miers ex K. Schum. —C. Branch with tendrillate leaves (from B. Stergios et al. 8868, MO). —D. Inflorescence. —E. Inflorescence node (D, E from J. A. Steyermark et al. 109875, MO). —F. Longitudinal section of corolla with androecium (from S. F. Smith et al. 1051, MO). Volume 95, Number 3 2008 Pool Review of Pyrostegia (Bignoniaceae) N de Puerto America, A. Jardim et al. 2431 (MO). Santa Cruz: Prov. Velasco, Parq. Nac. Noel Kempff Mercado, Asseradero El Chore, R. Guillén et al. 4108 (MO). Acre: Mun. de Senador Guiomard, BR 317, Km 23, C. A. Cid & B. Nelson 2821 bns Amazonas: Serra near Namorado Abuna, G. T. Prance et al. 14705 (MO). 364, Cuiabá-Porto Velho, Distr. de Patronal, Vila Bela da Santíssima Trindade, C. A. Cid et al. 4409 (MO). Pará: Near Embrapa station at Km 23 on rd. Altamira—Itaituba, C. Berg et al. BG755 (MO). COLOMBIA. Caquetá: Km 90 on rd. from Neivo to Florencia, ca. 6 km NW of Florencia, A. H. Gentry et al. 9067 (MO). Meta: Cordillera La Macarena, mesa del Río Sansa, J. Idrobo et al. 1274 (MO). GUYANA. ssequibo: Kanuku Mtns., M. J. Jansen-Jacobs et al. 437 (MO). PERU. Loreto: Pucallpa-Aguaytía rd., Km 37 W of Tournavista, M. E. Mathias & D. Taylor 6080 (MO). Madre de Dios: Iberia, Río Tahuamanu, R. J. Seibert 1952 (MO). Pasco: Prov. Oxapampa, Panjil, 12 km by air from Puerto Inca, D. N. Smith & R. Foster 2416 (MO). San Martín: Prov. San Martín, Tarapoto, F. Woytkowski 35089 (MO). SUR- VENEZUELA. Amazonas: Cerca del pueblo de Macuruco, P. Berry 2123 (MO). Bolívar: Mun. Sucre, carr. desde Jabillal a Ciudad Bolívar, E. Sanoja 2743 (MO). Amacuro: Dept. Tucupita, betw. Los Castillos de Guyana and the town of Sierra Imataca, G. Davidse & A. González 16437 (MO). 3. Pyrostegia millingtonioides Sandwith, Kew Bull. 15: 464. 1962. TYPE: Brazil. Pará: Obidos, in capoeira of terra firme, 23 July 1912, A. Ducke s.n. (MG 12046) (holotype, digitized image!; isotype, MG not seen). K not seen, K Branchlets with dense to scattered curved-ascend- ing trichomes, interpetiolar glandular fields absent, slight interpetiolar ridge present. Leaves 2-foliolate, often with an apically trifid terminal tendril; petioles cm, with curving trichomes at least in adaxial canal; petiolules 0.5-5 cm; leaflets ovate, slightly subinequilateral, 4-11 X 2-6 cm, chartaceous, 3 to 5 pairs of lateral veins prominent abaxially, both surfaces densely to sparsely pilose to pilose only on veins or glabrous, pellucid-lepidote, often especially conspicuous abaxially, with large glands in the axils of lateral veins, base rounded or truncate, apex briefly acuminate-mucronulate, or acuminate-mucronulate. Panicle terminal, densely corymbose or paniculate with calyces overlapping in dried specimens, 2 to 4 times branched, peduncle, rachis, and bracteoles covered with minute, ascending-curved to appressed trichomes; calyx excluding denticules 3 X 5.5- mm at apex, with sparse lepidote scales, sometimes minutely puberulent, apex ciliate; corolla narrowly salverform, white; tube 4.1 wide at base, ca. 3 mm wide at mouth, internally —5.5 em long, 2-2.3 mm glabrous or pubescent near mouth, externally gla- brous; lobes elliptic-oblong, 0.6—0.7 X ca. 0.4 cm, puberulent apically and marginally; stamens and staminode inserted at approximately same level in corolla tube, 3.9—4.4 cm from base of tube, stamen filaments 0.15-0.20 cm, thecae subparallel, 1.7— mm, staminode ca. 1 mm; disc 1.5-1.75 X ca. l mm; pistil ca. 4.4 cm, ovary 2-3 X ca. 0.8 mm, stigma lobes reniform-suborbicular. Fruit not known (Fig. 2A, B Brazil, in the states of Pará and Maranháo; in dry secondary growth; elevation not Distribution. recorded on labels (Fig. 1). Phenology. Flowering in July and August. Discussion. This species, with its strongly fra- grant, white, salverform corollas, and stamens inserted near the mouth of the corolla tube, seems oddly placed in Pyrostegia, which otherwise has probably odorless (Gentry, 1990a), orange, narrow tubular-infundibular corollas with the stamens inserted below the middle of the corolla tube. Pyrostegia cinerea, P. dichotoma, and P. venusta are typical of hummingbird-pollinated Bignoniaceae flowers: odorless, the corolla bright red-orange (or deep red-violet), of fairly thick texture, with a narrow tube and wider mouth, and more or less glabrous internally (Gentry, 1980, 19902), while the fragrant flowers of P. millingtonioides, with white, narrow, tubular, and thickly textured corollas, are probably moth pollinated (Gentry, 1980, 19904). Gentry (1990a) reported that change of pollinator within a genus of Bignoniaceae is very rare and that only 1846 of the Neotropical genera of Bignoniaceae have more than one type of pollen vector, and most of the species in these genera are overwhelmingly bee pollinated with one or two species switching to hummingbird pollination. Only two other genera of Neotropical Bignoniaceae, Arrabidaea and Tabebuia Gomes ex DC., have both moth- and hummingbird- pollinated species, and most of the species in those two genera are bee pollinated (Gentry, 1990). Pyrostegia is only genus reported (Gentry, 1990a) to pollinated species, with no species pollinated by bees. On the other correlation between dry habitats (such as that of have both hummingbird- and moth- hand, there may be a positive Pyrostegia millingtonioides) and moth pollination. In a comparison of pollinators of lianas in different types of forests, Gentry (1991) suggested that a larger percent- age of hummingbird-pollinated species is found in moist and wet forests (2%) or pluvial forests (7%) than in dry forests (1%), and a larger percentage of moth- pollinated species is found in dry forests (296) versus moist and wet forests (0%) and pluvial forests (1%). The insertion of the stamens above the middle of the corolla tube appears to be associated with the moth-pollinated syndrome. Gentry (19902) listed three Annals of the Missouri Botanical Garden the exclusively moth Barb. Rodr., and Sphingiphila A. H. Gentry (listed as new genus). genera in Bignontaceae as pollinated: Tanaecium, Leucocalantha — Leucocalantha obs.), Tanaecium (Macbride. 1961), and Sphingiphila (Gentry, 1990b) all have the stamens inserted above the middle of the corolla (pers. tube. Therefore, the main differences between Py- rostegia millingtonioides and P. dichotoma and P. venusta appear to be associated with pollen vector. The three species have many morphological charac- ters 1 ~ | common: absence of interpetiolar glands, presence of generally 2-foliolate leaves with an apically trifid tendril, leaflets similar in shape and indument, calyx of similar shape, corolla tubes glabrous externally, corolla lobes more or less valvate at base in bud, stamens and stigma exserted, and a relatively wide stigma. In addition, as previously discussed, the pollen of all four species of Pyrostegia is 3(or 4)-collate and finely reticulate (Urban, 1916; 1962; Gentry & 1979), and P. millingtonioides, P. dichotoma, and P. cinerea have (Santos, 1995). millingtonioides is not Sandwith, Tomb, similar wood anatomy Alternative generic placement of P. suggested at this time. BRAZIL. Maranhão: Al Acailándia-Santa Inés, Km 100, G. S. Pinheiro 19 (MO): Florestal P Buriticupu, proprie- dade E Floresta Rio Doce S.A. 52 Santos et al. 41 (MO). 42 (MO). P (RB 16322) (NY). Additional specimens examined. km de Buriticupu, G. dos Ducke s.n. ará: Obidos. 4. 4. Pyrostegia venusta (Ker Gawl.) Miers, Proc. Roy. Hort. Soc. London 3: 188. 1863. Basionym: Bignonia venusta Ker Gawl.. Bot. Reg. 3: tab. 249. 18 Universel 5: 18. Tecoma venusta (Ker Gawl.) Lem., Hort. 1843. TYPE: tab. 249, illustration of greenhouse plant cultivated at Combe Wood, England, from seed originally from Rio de Janeiro. Brazil (lectotype, designated by Sandwith € Hunt, 1974: 75, tab. 249 in Ker Gawler, 1818). Bignonia ignea Vell, Fl. Flumin. 244. 1825 [19829]. d yrostegia ignea (Vell.) C. Presl, Abh. Konigl. Bohm. . Wiss ss., ser. 5, 3: 523. 1845. TYPE: tab. 15 in i Món Fl. Flumin. : ‘ones 6. 1827 [1931 | ito 15 in Ve llozo, 1827 11831]. Herb. : Paraguay. desig e here, tab. Pyrostegia venusta var. villosa Hassl. Si Bull. D: 1905 [190 4]. Capibary, Dec., E. Hassler : here, G!; duplicates, A!, BM!, W!). oissier, sér. 2, 5: 04. 5940 (lectotype, ae signale 5 K not seen, MOL NY! Branchlets glabrous, or with few trichomes at the nodes, or scattered to densely short-pilose to puber- ulent, interpetiolar glandular fields absent. interpe- Leaves 2-foliolate, often with an (the or leaves tiolar ridge absent. ends rarely 3-foliolate: apically trifid terminal tendril branched again, bifid or trifid), petioles 1-4 em, densely pubescent, pilose in the adaxial canal or glabrous; petiolules 0.4-3 em: leaflets ovate (rarely lanceolate), slightly subinequi- 1.26 cm, membranous), 3 to 5 pairs of lateral veins prominent aleral, 2.5— chartaceous (rarely below, densely short-pilose to glabrous, pellucid- lepidote, often especially conspicuous abaxially, with large glands in the axils of lower lateral veins, base rounded or truncate (rarely cordate), apex briefly acuminate-mucronulate, or acuminate-mucronulate obtuse-mucronulate or acuminate). Panicle usually terminal or axillary, generally dense or subcorymbose dried. specimens, 3) times ds, and bracteoles nearly glabrous to with calyces often overlapping unbranched or l or 2(rarely branched. peduncle, rac densely puberulent or pilose, the trichomes initially perpendicular to the surface; calyx. excluding denti- cules 4-8 X 4.5-10 mm at apex, with sparse lepidote scales, glabrous to densely short-pilose to puberulent, apex ciliate; corolla narrow || tubular-infundibular, orange or reddish orange (rarely yellow); tube (rarely 8-13 mm t and 2.7)4-7 cm long, 2-5 mm wide at base, sericeous c below wide at mouth, internally and staminode, 1- 1.8 ulent apically and marginally: stamens inserted insertion of stamens externally 0.3-0.7 em, puber- n "corolla tube, staminode inserted glabrous; lobes oblong, 3.5 em from base o (rarely 0.8)1.2-1.6 em above insertion. of higher stamens, stamen filaments (rarely 2.6)3.2-5.2 cm. 4—6.3 mm, staminode |— (rarely 3 8 mm (rarely developing an anther and then similar in 1-3 x 2- ] mm, thecae subparallel, length to, and inserted with, stamens); disc 3 mm; pistil 4.6-8.5 cm, ovary 4—6.5 X ca. stigma lobes broadly ovate, ovate, orbicular, or broadly oblong. Capsule 13.5-30 Xx 0.8-1.5 cm. glabrous, drying with olive cast, midvein apparent, but not conspicuous, base acute, apex aristale: seeds 0.7—1.4 X 2.8-4.5 cm (Fig. 3A, B; Fabris, 1965: fig. 29; Bureau & Schumann, 1897: fig. 98: Lohmann & Pirani, 1998: fig. IOA—E). Distribution. Atlantic and southern Brazil. from Piauí to Rio Grande do Sul. northeastern Argentina; 70-1300 m elevation, 1000 m. erowing in disturbed semi-evergreen forest or cerrado. southern. Paraguay, and gener- below Frequently reported. as ally found Cultivated as an ornamental throughout the tropies and subtropics and possibly naturalizing in some areas (Fig. 1). Phenology. Collected in flower in every month except December, peaking June to September. Fruiting July to December. Brazil: Cipó de São João (Sand- 11982). Common names. with & Hunt, 1974; Heringer Cipo Caititu Volume 95, Number 3 2008 Pool Review of Pyrostegia (Bignoniaceae) (Cutler 8404), Cipó Tingá (Carauta 408), Dedo de Moça (Emperaire & Campelo 2748); Argentina: Pico de Tucán, Flor de San Juan (Fabris, 1965); Peru: Lluvia de Oro (Seibert 2333); Guatemala: Chiltote, Chorro de Oro, Chorro (Standley € Williams, 1974); El Salvador: San Carlos (Standley € Williams, 1974); Costa Rica: Triquitraque (Standley € Williams, 1974); U.S.A.: Flaming-Trumpet, Golden-Shower (Bailey et al., 1976). Flame Flower, Discussion. The Vellozo plate 15 (Vellozo, 1827) was chosen as the lectotype of Bignonia ignea, as no Vellozo Bignonia specimens have been found (Gentry, 1975). Based on the Vellozo plate and description (Vellozo, 1825), B. ignea is placed in synonymy of Pyrostegia venusta, following Bureau and Schumann (1897), Sandwith and Hunt (1974), and Gentry (1982). The six syntypes and 15 of t — he isosyntypes of Pyrostegia venusta var. villosa were seen, and £. Hassler 5940 (G) is chosen as the lectotype, based on the quality of the specimen and its large number of widely distributed duplicates. Hassler 1905) distinguished P. typical variety based strictly on presence or absence (Sprague, venusta var. villosa from the of pubescence, the typical variety defined as glabrous to glabrescent and variety villosa as the whole plant more or less villous to pubescent. The type material of P. venusta var. villosa examined in this study had densely pubescent to pilose young branchlets, leaflet surfaces, and pedicels. In this study, specimens with similar pubescence were seen only south of Bahia, Brazil. Plants intermediate in pubescence, that is, with scattered trichomes on the pedicels but with the young branchlet internodes and the leaflet surface between the veins glabrous, were also restricted to this same geographical range. Nearly all specimens throughout the entire species range showed some degree of pubescence, with at least some trichomes present in the adaxial canals of the petioles and petiolules and on the bracteoles of the inflorescence. The trichomes are of the same type as those found on the more pubescent specimens. The less pubescent plants are found throughout the entire range of the species and, in fact, appear to grow in the same localities as the pubescent individuals. For example, the following pairs of pubescent and less pubescent = lected at the same localities and on the same dates: Zardini 14343 and Zardini 14365 and 7193 7131. MH appropriate at this point not to formally recognize plants were co Zardini and Zardini seems more the variety. Pyrostegia dichotoma has been treated as a synonym of P. venusta by some authors (Macbride, 1961). The two are treated here as separate but closely related species, as discussed below P. dichotoma. Menninger et al. (1970) includes Pyrostegia ignea = P. venusta) on his list of the most beautiful flowering climbers in the world, and ranks it as the most popular of all in the tropics. It is very widely cultivated and possibly naturalizes in some areas (Gentry, 1973). The specimen Ferreira 192 indicated that the plant is considered toxic to cattle and has been used as a tonic and anti-diarrheic, of this but I have not seen a published Williams & 6963 also recorded that the flowers are poisonous to cattle account use. Assis Selected examined. ARGENTINA. rientes: Dep. Ituzaingó, Isla Apipé Grande, Pane o-cué, Schinini & R Vanni 15€ 81. 3 (MO). BR AZIL. 14-17 km a hi ine Cor- ) & (MO). lun. de Marliéria, ae Estadual do Rio Doce, E. e 13984 (MO). Paraíba: Mun. de Maturéia, Pico do Jabre, 2 erra de Te r araná: Faz. / ia h 21599 (MO). S Santa Genebra Forest Reserve, Barão Geraldo, near Campinas, A. H. Gentry 59074 MO). PARAGUAY. Atira, E. Hassler 3022 (syntype of A. BM, K not Pyrostegia venusta var. NY, W), 3022 Indios (Cianorte), G. 5ào Paulo: : — “~ villosa, G; isosyntypes, -a eme of Pyrostegia venusta var. villosa, G; isosyntypes, ot seen, NY): Chololo, E. Hassler 6663 (syntype of [Mtn venusta var. G): Ipe-hu, E. Hassler 5244 (syntype of Pyro stegia venusta var. villosa, G; isosyntypes, BM, NY, Wy ssle 6. NU = NE seen, villosa, Bs of patea venusta var. . Caaguazu: 10: E x Strossner, A. " 59440 Guairá: Cordillera de Ybytyruzú, 15 km N of Atena on rd. Melgarejo-Atena, E. Zardini & R. Velásquez 13555 (MO). Itapúa: "dn CEDEFO, W. Hahn & L. Pérez de Molas 2760 (MO). San Me. 12 km al NE de Choré, E. Zardini & C. Benitez 3401 (MO). Literature Cited LH. & E. Z. Bailey Liberty Hyde Bailey Hortorium. 1976. Pyrostegia. P. 9: in Hortus Third, a Concise Dictionary a Plants Cultivate : in the United States and Canada. Bailey, I , revised and ^ s by staff x P Macmillan General efe . New York. Barbosa Rodrig rues, J. 189la. Vellosia, 2nd ed., Vol. 1 Imprensa Nac sional, Rio de Janeiro. 91b. Vellosia, 2nd ed. Vol. 3: Estampas ind a. ae nsa Nacional, Rio de Janeiro. mcn E. & K. 1890. 30 in C. F. P. Martius (editor), Flor Bignoniaceae. 1 Pp LEE. i Schumann. von fasc. 118). Lipsiae apud Fried. Fleischer in Comm., Munich. —— — & — 1897. Bignoniaceae. ii. Pp. 229-452 in C. F. P. von Martius (e ditor), Flora Brasiliensis 8(2, fasc. 121). Li e apud F ried. Fleischer in Comm., Munich. de. 184 j Ex oa nap *andolli . Bignoniaceae. Pp. 142-2 (editor), Prodromus 9. Sumptius ns Misión: Paris. Fabris, H. 1965. Bignoniaceae. /n Flora Argentina. Revista Mus. La Plata. Sece. Bot. 9(43): 273—419 Annals of the Missouri Botanical Garden Galetto, L., L. M. Bernardello & H. Juliani. 1994. Characteristics of secretion of nectar in Pyrostegia venusta (Ker-Gawl.) Miers New Phytol. 127: 465—471. Gentry, A. z ceae. In |n (Bignoniaceae — E 3 [1974]. Part IX. Family 172. Bignonia- ;. Woodson Jr. & R. W. Schery bd, Flora p Missouri Bot. Gard. 60: 781— 74. € oevolutionary patterns in diis d dr ‘an Ann. Missouri Bot. Gard. 61: 728-759 75. Identification of TEREA dion i. a 7-344. . Bignoniaceae 25(1): MEER of ene . 19 Bignoniaceae. 975 Vellozo's Part 1. Fl. Neotrop. Monogr. 32. Venezuela, Fundación de Educación Ambiental, Caracas. a. Evolutionary is rns in PUE a Bigno- . Gard. 55: 118-12 . 1990b. Sphingiphila (Bignonia E a new genus yst. Bot. 15: 277-279. . Breeding and pA systems of lianas Pp. 393-423 in F. E. Putz & H. A. Mooney (editors) The Biology of Vines. Tm University Press, Cambridge. 1993. A Field Guide to the Families and Genera of Woody Plants of Northwest South America. Conservation Inte rational, Washington, D.C. & A. S. Tomb. 1979 [1980]. Taxonomic implications of Bignoniaceae palynology. Ann. Missouri Bot. Gard. 66: Bignoniaceae. /n Flora de 198 EU ad ie niaceae. £s m. New York Bo from the Paraguayan chac 1991. pei 156 G e " F. . 1911. Plantae Brasiliae centralis d laziou P poe Bot. France 58(Mém. 3f): 1-656 Gobatto- Baar aves, A. A. & M. N. 5. Stort. 1992, ne: floral e reprodugáo de Pyrostegia venusta (Ker Gawl.) Miers P cues Revista Brasil. Bot. 15: 3741. Goldblatt, A. . 1979. Cytology of Bignonta- b Gusman, A. B. & G. lectae — ceae. ` 1996, Differences in floral floral nectar constituents, flavonoids in petals of orange and yellow P (Bignoniac pus flowers. Phyton (Horn) 3 36: 161-171. Joshi, A. B. .W.1l 1956. Ploidy in two Big- noniaceous ni n de rs. Indian J. Genet. Pl. Breec 16: 57-5 Goltsberger. morphology, carotenoids, and lardas. l ohmann, L. G. & J. R. Pirani. 1998. Flora da Serra do Cipó, nas Gerais: io Bol. Bot. Univ. Sáo Paulo 17: 127-153. Macbride, J. F. Pe ua "ue. a Peru. Field we Nat. Hist., Bo n 13(5c). Menninger, E. 50 € als 1970. Flowering Vines of the ere ido Press Inc., New York. Miers, J. 1863. Report on the d collected by Mr. Weir, especially the Bignoniaceae. Proc. Roy. Hort. Soc. London 3: 179-202. 1845. Gess. T Sandwith, N. Y. 193t a (editor), Flora of SM 'ol. 4(2 o x A 3-103 in Flora of Bemerkungen. Abh. 431—584. Pp. 1-86 in A. Pulle ). J. H. de Bussy Ltd., Botanische Kónigl. ser. 5, 3: 962. Contributions to the flora of tropical America: oak Kew Bull. 15: 459—406€ Notes on Bignoniaceae: XXVI. 1968, Contributions to the [lora of MUN cal America: XXIX LXXVI. Notes on Bignoniaceae: rrabidaea in Mo s “Flora Brasilie nsis’ and subseque un Kew Bu 22: 403—420. —— —— & D. R. Hunt. 1974. Bignoniáceas. Pp. 1—172 in P. R. Reitz (editor), Flora Ilustrada Catarinense, Vol. 1 p (Bignon.). Itajaí, Santa Catarina, Brazil. Santos, C. dos. 1995. Wood Anatomy, Chloropast DNA, and j. Ph. D. Reading, United Flavonoids of the Tribe Bignonieae (Bignoniaceae) Thesis, The University of Reading. Kingdom. Schumann, K. 1894 [1894-1895]. Bignoniaceae. Pp. 189— 252 in A. Engler (editor), Die naturlichen Pflanzenfami- SA Wilhelm Engelmann, Le ipzig. 948. The use consideration a the family Bignoniaceae. Bot. Gard. 35: 123-137. Sprague, T. A. 1905 [1904]. Bignoniaceae. Pp. 78-88 in R. Chodat & E. Bassler die Plantae Hasslerianae. Bull. Herb. Boissie `T, SÉT. Stafleu, F. E LR.» Cowan: ol. 4. Bohn, mn, and Dr. of glands in a taxonomic Ann. Missouri 1983. Taxonomic Literature, Scheltema & Holkema, Utrecht/ W. Junk Publishers, The Hague/ ee ae y, n C. & L. O. Williams. 1974. Bignoniaceae. /n Flora of Guatemala Fieldiana, Bot. 24(10): 153-232. ) Miscellaneous note Poy Barbosa . Inform. Kew 1909 —226. . 1916. Über Ranke n und Pollen des Bignoniaceen. . 34: 728-758, pl. 21. . da Conceição. 1825 due Flora Fluminensis. > p asconcellos & S 7 [1831]. Flora F Te nsis, e ‘ae in Urbe Sendefelder, Paris. W ord k, J. J. 1970. Erroneous poa in n. laziou collections of Melastomataceae. Taxon 19: 911— — a Brasilias, Rio de Janeiro. Icones 6. Pub- = aeque Fluminensi praefectus, APPENDIX 1. Index to Exsiccatae. Collections are alphabetical by collector name. After the collectors number, the number in | parenthe ses corresponds to the species number in the List of Species. Type appear in boldface. All collections cited here were examined by this author as part of this study. collections List Or SPECIES l. Pyrostegia cinerea Bureau ex K. Schum. rs ex K. Schum. 3. Pyrostegia pees ed Sandwith 4. Pyrostegia venusta (Ker Gawl.) Miers. A Pyrostegia dichotoma Mier Ackermann s.n. (4). Agostini 1501 (2). Agra et al. 1955 (4), 4398 (4), 4480 (4). Aguayo 117 (4), 303 (4), 562 (4), 594 (4). Aguilar 217 (4). Alston 6402 (4). Amaral et al. 1165 (2). Anderson 72 (4). Andrade et al. 214 (4), Lima & Lima 39 (4). Angeli 358 (4). Annable 2823 (4). Araujo 384 (4). Arbo al. 2267 (4), 2656 (4). o 1181 (4). Arroyo et al. 730 (2). Aymard et al. 6303 (2), 7582 (2), 7567 2). 7618 B. Bahia 100 (2). e PS (4). Balapure 656 (4). Ball s.n. Andrade- (4). Bang 1510 (2). 6 (2). Barreto 787 (4), 790 (4). Barth 1233 (4). A 709 (4), 1032 (4), 1556 (4), Basualdo s.n. (4). Belanger 204 (4). Belém 1163 (4), 1554 (4). 1603 (4), 3876 (4). Belshaw 3431 (2). Berg et al. BG755 (2). Bernardi 19584 (4). L. Bernardi 18049 (4). Berry 1654 ( ). 212 3 (2). Bertoni 861 (4), 1477 (4), 2801 (4), 3744 (4). 3772 (4). Billiet & Jadin 3351 (4). Blanchet 2563 (4) Boom Axa (4). Boone 1240 (4). ie & 559 = Cunningham s.n. (4). Breedlove 23552 (4). Brenes 23234-A (4). e 4099 (4), d N. L. Britton & E. G. Britton 9289 (4). Brunner et al. 914 (4). Buchtien 1314 (2). Bunting & A Holmquist 4613 e 4618 (2). Burchell 4768 (4). Volume 95, Number 3 2008 Pool 509 Review of Pyrostegia (Bignoniaceae) Cabral 34. (A). ale & Sanz 29226 le a « Watkiss 73 (2). C. E. Calderon et al. 2765 (2), 2835 (2). S. Calderón 573 (4). M. 1634 (4), 2156 (4). ous 199 (4). Capelosa 7 (4). Carauta 408 (4), 854 (4). 871 (4), 888 (4). Carauta et al. 4738 (4). Cardenas 2969 (4). Cardona s. s.n. (2). Carnier 1140 (4). Carr s.n. (2). Ce oq. 23372 (4), 24978 (4). Castillo 763 (2). Catharina 8 (4). Pedra do Cavalo 777 (4). Chagas s.n. (INPA 1198) Chodat 176 (4), 176-b (4), 177 (4), 178 (4). Chun (4), 6925 (4). Cid & Nelson 2821 (2). Cid et al. 944 (2), 4409 (2). 4770 (2). 6267 (2), 6278 (2). Claussen MD, 108 a 190 (4), 489 (4), 8902 (4), Ce s.n. E (INPA 1730) (2). Coélho et al. s pr et al. 2055 (4), 5724 (4). J. PS & n hbach 336 (4 M. R. Cordeiro 598 (2). Cosme d a Cristobal & pee 1761 (4). Grori 15 du Cutler 8404. (4). edi s.n. (4). Peu n 901 (4), 1 (2). Davidse 4451 (2), 4471 (2). Dandie & Conade 16329 (2), 16437 el pu S Huber 15457 wor 1453 (4). 1511 (4), 1590 (4). Degen & Zardini 466 (4). Degener 7991 (4), 7992 (4). De gel : dod B-1717 (4). Deng 10351 (4). Dickason 5827 H. Dodson & P. M. Dodson 11696 (4). Dombrow Fa 201 o Dryander 144 (2). Duarte & Castellanos 254 id Dubs 1617 (2). Ducke 532 (1), 5321 (1), 725 (2), ee (MG 12046) (3), Ducke s.n. (RB 16322) oe m A- 906 (4). Dusén 709 (4), 766-a s 14121 (4), 1 8 (4). Echeverria 92 (4). G. n & L. Eiten poe (4), 2194 (4). Elcoro 646 (2). Elia & rra a 82 (4). Ellenberg 2595 (2). (4). Emygdio et al. 3452 (4). = = = = aes E Campelo 2748 Eugeni 2 (4). Ci ie et al. 1022 (4). Faurie 1025 (4). Fernandes 162 (4), 410 a ). A. d 913 (2), 5070 (2). M. D. Fernandez 4 (4). S . Fernández 9 (4). Fernandez Casas & Molero 3543 (4). 4). A ^yra 5070 (2), 17392 (2). Ferrucci et o 685 bu Fiebrig 18-pp (4), 5442 (4). Filho m 232 (4). Flynn 2747 (4). Forbes s.n. (4). Foster P fe n 15150 ren Fotius 3575 (4). Frazas s.n. (4). Fung 2-65 t). Ganev 779 (4). Gardner 543 (4), 1768 (4), Gardner s.n. (4). Caudichaud 558 (4), e (4). 998 (4). Genelle & Fleming 452 (4). Gentry 472 (4), 12122 (4), 12857 (1), 16273 (2), 21495 (4), 43734 (2), 2 (4), 66146 (4). Gentry & Berry 14913 (2). 15065 (2). Gentry & Franco 59065 (4). Gentry & Prance 11198 (1). Gentry & Young 31971 (2). Ge entry et al. 9067 (2), 10434 (2). 10757 (2). 10941 (2). 12773 (1). 42098 (2), 44248 (2), 44347 (2), 50177 (4). 51899 (4), 51942 (4), 52291 (2), 59428 (4), oy E ), 77514 (2), 77664 (2). Gibbs et al. 5334 . Glaziou d (4), 4696 (4), 12976 (4), 6 (1). Gomes 203 (4). Gómez é "n Ciida 54 Gold * (4). 1022 (2). een el e 166 (4). Guillén & Choré 0 (2). Guillén et al. 4108 (2). Guimardes 29 (4). — el del 955 (2), 1124 — ) y © z a En a ~~ Ie] E He age et al. 298 (4). Hahn 2 ^ Molas 2760 (4). 2507 (4), 2605 (4), 2653 (4). Handro 493 (4). Harley 21606 (4) Hassler 341 (4), 341-a (4), 1097 » 1097-a (4), 022 (4), 3022-a (4), 4908 (4), 5244 5245 (4), 5940 (4), 6663 (4), 7256 (4), 7363 (4). n (4), 10218 (4), 11784 (4), 11784-a (4). ue 1442 (4), 6304 (4), 14274 (4), 14429 (4), 21599 (4), 32549 (4), 43011 (4). Hatschbach et al. 52275 (4). Heiner s.n. a ^ He Me (4). Henschen 1-365 (4). Herbst & e 5246 (4), 5366 (4). Heringer 5264 (4), 9491 7 11982 (4), 195 pu W. (4). Holst et Hochne 475 (4), Hoehne 3653 (4), W. nee s.n. (4). Holst 6057 al. 1994 (2). E. ). Holway & M. Holway 1958 (4). Hu 12941 (4). Hu & is 19980 (4). Huber 225 (4). Hunt 6310 A), 6380 (4). Idrobo et al. 1274 (2). Irwin 2047 p Irwin et al. 55538 (2). 57636 (2). Itaipu d 419 Jansen-Jacobs et al. 369 (2), xd m E et al. 2431 (2). Jobert 452 (1). 2 700-A (4). Jorgensen 3456 (4). Jouvin 232 (4). Kahn & Llosa 2150 (2). Kaspiew 1174 (4). Khan et al. 900 1). Killeen et al. 5411 . Killip 37626 (2), 45630 (4). Kirkbride 4405 (4), 5315 (4). Klein d (4), 5509 (4), 5742 (4). Klug 2969 (2). Koyama et al. T-32176 (4). Kral 45027 (4). Krapovickas 41877 ( (4). Krapovickas et al. 25475 (4). 25775 (4), 26167 (4), 26402 (4). 46164 (4). Krukoff 10208 (2), 10209 (2), 10: 10499 (2). Kuhlmann & Hoehne s.n. (SPF 10278) ( Kummrow & Golte 606 (4). Kunkel 8115 (4). Kuntze s.n. (4). Labbiente 7 (4). Lambert s.n. (4). Lanna S. 1107 ( Lasseigne 22582 (1). J. Lau 1930 (4). S. Y. Lau & To 10862 (4). Laurenio et al. 154 (4). Leeuwenberg 10197 (4). Leite 1776 (4). Leschenault s.n. (4). Lesmero 124 (4), Lesmero s.n. (4). Lewis 88837 (4). Lhotsky s.n. (4). Liesner & González 5633 (2). Lillieskóld s.n. (4). Lin 89-067 (4). Lindberg 150 (4). Lindeman & Haas 1150 (4). Lindeman et al. 8867 (4). iude) 1587 (4), 2315-1/2 (4). A. Liogier & P. Liogier I (4). A. Liogier et al. 30274 (4). Llatas Quiroz 2299 (4). Lofgren 20 (4). Lohmann et al. 34 (1). Lopez-Palacios 239 (4). Loureiro et al. s.n. (INPA 37895) (1). von Luetzelburg 2033 (4), 2070 (4), 12511 | V). 26210 (4), d M ). Lund. 790 (4), 2012 (4). Lund s.n. (2). Lurvey Ps P. J. Maas & H. Maas 326 ux el al. 1560 (4). Magalhães 3154 (4). Maguire & Ec 34754 (2). Maguire et al. 56480 (2), 56855 (2). Maisch 13791 (4). Malme 878 (4 Mandon s.n. (2). Marcano-Berti 19-12-76 (2). March 2065 (4). E. Marin 1591 (2). G. Marín & Jiménez 213 (4). Mariuzzi s.n. (4). Márquez R. 1042 (4). Martin et al. 1852 (2). Marques Meyer s.n. (4). Martius s.n. (4). Mathias & Taylor 6080 (2) Mattos & Magnanini s.n. (4). Matzenbacher 19 (4). Maxwell 88- m (4). McFarlin 4198 (4). McKee 39307 (4). McVaugh 8 (4). Meader 12 (2). Medan et al. 93 (4). Meebold 21962 (4). 26640 (4). Corréa de Mello 27-g (4). Corréa de Mello s.n. (4). Mello Silva et al. s.n. (4). Menandro 123 (4). Mexia 4817 (4). Meyer & Mazzeo 10461 (4). Midence Jr. 803 ( 1). Mikan 15 (4). Millspaugh 33 (4). Mimura 415 (4), . Mizoguchi 675 (4), 1661 2 (4). A. Molina 31685 (4). D. Molina f ; a 14783 (4), 15904. (4), 16069 (4). Moraes 239 D. 5 o Rd (4). Mori et al. 10423 (4), 12294 (4). Mn et al. 3 . Morong 745 (4). Mosén 953 (4). Mota 21 (4). M 68051909 (4). Mwangangi 1769 Nagel 8096 (4). (2) MÁ ~ - — E 3 NE ~ Nans 64005 (4). Nee 35258 (4), 41163 B. Nelson a ). J. B. Nelson 493 (4). Nicolson 1656 (4). Núñez et al. 6779 (4). B. De & F. debeo es ). Occhioni 5634 (4). A. A. Oliveira et al. 13004 (4). S. J. M. Oliveira 7 ( Orozco C. 178 (4). Ortega 4367 (4), 7324 (4). M. Ortiz 255 4), 273 (4), 611 (4), 640 (4). Osorio s.n. (4 Pabst 7005 (4). Pabst & Pereira 5744 (4). Padilla 51-a (4), 255 D 434 (4). Palacios 108 p S & Ruíz 878 (2). Parvey 145 (4). Peckolt 23 (4 n 1 (4). Pedersen 9250 (4). Perdone 76 (4). Pereira 308 (4). B ) E. Pereira 5616 (4), 7629 p Eo. 102 (4), 13 = (4). Petersen & Hjerting 293-A (4). Pickel 1515 (4). Piedade s.n. (4). Pinheiro 19 (3). Pirani 3-77 (4). Pire & eiu en ). J. M. Pires & Santos 16310 (2). M. J. Pires et al. 676 (2). Pizziolo 145 (4). Plowman 1852 (2). Pohl 81 (4). ps (4). m ma — 510 Annals of the Missouri Botanical Garden 2011 (4). Poloni et al. 1783 (4). Prance 6277 (2). Prance & Prance 21012 (1). Prance & Schaller 26296 (4). Prance et al. 755 (2). 5139 (2), 5951 (2). 6642 (2). 7167 (2 14705 (2), 59181 ( de Queiroz 16: b 4), Quirós 135 (4). Raben 500 (4). ~ (2); 4722 (4). Quevedo et al. 2479 (2). 501 (4). 893 (4). Raben s.n. (4). Rambo 29512 (A), 42050 (4). 42287 (4), 42403 (4). 42466 (4), 42509 (A). 12625 (4), 43028 (4). 47195 (4). 47286 (4). 48722 (4). (4). Ratter et al. RI154 (2). Rauscher 36 1348 (d) Reekmans 7326 (4). Regnell 150 (A), I- 363 (8), IH- 163 (4). Reichardt 55 (4). Reineck & Czermak 605 (4). Reiss 82 (4). Reitz 1174 (4), 1502-b (4), Reitz & Klein 2051 (4), 6846 (4), 9021 (A), 12642 (4). 12815 (4). 12855 (4). 13106 (4). Renson 132 (4). Renvoize et al. 3305 (4). Reznicek et al. M38 (4). Ribas et al. 2482 (4). B. G. S. Ribeiro 1074 (2). R. T 78 (4). Riedel s.n. (2). Ritter 53450 (4). Roberston & Austin I (4). de pue et al. 639 (4). Rodrigues et al. 82 (4). -A (4). pi s.n. (4). Rosa & da 2). Rose et al. 14085 Rusby 1149 (2). 2725 (2). Catan 1851 (4). Sagastegui A. 6880 (2). Saldías et al. 2837 (2). 5014 (2). rs 2743 (2). Santoro 415 (4). 587 (4). Santos et al. 2 (4). G. dos Santos et al. 41 (3). 42 ). R. S. Santos po ( ^ R. S. Santos & Castellanos 24121 — Sampaio s.n. (4). 5 E dos 2001 a 9105 p 13281 (4). Schinini a Vanni 15815 (4). 11251 (4). Schlieben 9979 (4). Schomburgk 902 (2). Sn wen ). Schuck s.n. (4). Sone V. 9795 a Schupp 258 Schwarz 074 (4), 2727 (4), 2784 (4). 2927 (4), 2967 (4), pem (4). 3140 (4). 3191 (4). 3308 (4). B (4). 1695 (4). 6077 (4). id 1139 (2). Segovia 28 (4). A. ji ^hnem 3807 (4). P. A. Sehnem 1627 ue 5. J. Selirem. a P, Sehnem 3305 (4). pu 1420 (4 D (2), 2333 (4). Seligson 690 (4). d discs s.n. D Se (4). Sendulsky 1009 (4). Shepherd et al. 5817 (4) eh s.n. (4). Sick B871 (4). D. S. Silva et al. 36 (4 T Silva et S 7108 (2). Silvestre 61 (4). Singh 204 . N. Smith & Foster 2416 (2). L. B. Smith & Reitz 5939 (4). S. F. Smith et al. 1051 (2). Solomon 6148 (2). 17076 (2). 18218 (4). Sorta 549 (4). 1661 (4). Soto Nunez & Schinini et al. — Soria et al. Martinez S. 6170 (4). D. S. de Souza et al. 93 (4). V. C. Souza 2593 (4). Sperling » did 6416 (2). Spruce 3930 5190 (4), 19243 (4). 19244 (4). 24629 (4). 933 (4). Stergtos a (2). Stergios et al. 8808 ; 85373 (4). Steyermark el al. 109875 (2). 124583 (2). Stone 7414 (4). Strang & Castellanos 1055 (4). Streemadhavan 5716 (4). Sturrock 86 (4). Stutz 503 (4). 504 (4), 648 (4). Sucre 7780 (4), Sucre s.n. (RB 140961) (4). 17 (4). Tang 10351 (4). Tavares 650 (4). Taylor Vio . Teague 406 (4). Teixeira et al. 1348 (2). Tenerife a 4 m. 5402 a Pod 3988 (2). Thorne & Ochoa 9702 (4). Toro 108 (: Pu Torres 1 (4). Tscha (4). von Pue T 2 wo mt hu J. Valerio R. 3089 (4). f Valerio Ss 26 (4). Valverde 409 785 (4), 819 (4), 965 (4). van der Veen s.n. (4). Venturi 3425 (4). Vieira et al. i (2). Villalobos C. 55 ). Wagner 1142 (4). 12157) (4). Warming s.n. (4). Wasum s.n. (U b & Maly 486 (4). pore 166 P Weinberg 546 (4). Weir 8 (4). van der us 5 (4). West 140 (4). 1 (4). Widgren s.n. (4). L. edu e: (2). L. " J. Williams 4 (4). L. O. Williams k bn 6965 (4). 7495 (4). R. S. Williams 755 (2). 879 (2). 1574 (2). Wilson- Browne 217 (2). Windisch 652 (2). Woolston 530 (4). Wortley 1847 (4). Wovtkowski 35089 (2). M rar & Wright 406 (2). Yena 257 (2). Yanagizawa s.n. (SP 8988) (4). pis s.n. (ESA- 6363) (4). Yuncker 17917 (4). Yuncker et al. 6092 (4). Edu AS VN 11909 (4). Zardini 4410 a 4512 (4), 3 (4). 5423 (4). aa 5898 (4), 6445 (4). 6678 (4). E (4). 6777 (4). 3 (4). 7131 (4). 2000. 7213 (A), 7274 (4). 137 a pred 14305 (4). 14728 (4). Zardini & Aguayo 14394 (4), 14846 (4). 14913 (4). Zardini & Benitez 4). e & Ortiz 3070 (4). Zardini & C. Velásquez 1250 T ), 12547 (4), 13027 (4). 13857 (4), 13896 (4). 14266 (4), d 142687 (4), dM - ele & : Velázquez 12415 (4). 15212 (4). 13223 (4). 1 (4). TA 14223 (4), 14498 (4), H1 G " - Zardini ig S. 3358 (4). 3401 ( . Velázquez 15197 (4). dl el ui 12238 (4). Zika 2508 jn Zuloaga et al. 2122 ANATOMY AND HISTOCHEMICAL Andrea S. Vega,” María A. Castro,? and LOCALIZATION OF LIPID SECRETIONS IN BRAZILIAN SPECIES OF PANICUM SECT. LOREA (POACEAE, PANICOIDEAE, PANICEAE)' Fernando O. Zuloaga” ABSTRACT This paper constitutes the first report of the is ies of Panicum L. sect. Lorea Zuloaga (Poa Lipid secretions were found in leaf sheaths surrounding the bases of the macrohairs, and i anatomically here. Lipid secretions are now also 1 re ae d for envolze, P. chnoodes Trin., P. cipoense Reny & Send.. molinioides Trin., P. poliophyllum Renvoize & Zulo aga ey words: Gramineae, herbivory, lipids, Panicum. o n the mesophyll. vaginiviscosum Renvoize € Zuloaga and P. acicularifolium Renvoize & Zulc ; und P. trinii Kunth plant defenses, che mical nature, and histoc pene localization of secretions in sn tudy also includes an anatomical analys oa ace as in specialized epidermal cells, at times Viscous secretions have been previously reported for P. oaga, species in section Lorea that are analyzed T first time in id other species of section Lorea: P. bahiense P. durifolium Renvoize & Zuloaga. P. euprepes Renvoize. P. Pas eae, secreting epidermal cells. The genus Panicum L., s.l. includes nearly 400 species and exhibits a worldwide distribution (Clayton & Renvoize, 1986). As presently circumscribed in its strict sense (Aliscioni et al., 2003), the genus should be limited to only subgenus with five sections and approximately 100 species worldwide Panicum, and 68 taxa in the Americas. The other previously considered subgenera Dichanthelium Hitchc. & Chase, Megathyrsus Pilg., Phanopyrum (Raf.) Pilg., jon) and Steinchisma Raf. have been raised to the generic level or their species have been transferred to other 1998: Simon & Jacobs, 2003: 2003). Also, according to Aliscioni et al. (2003), there are several species or genera (Zuloaga et al., Aliscioni et al., sections of dubious taxonomic position within the Paniceae. i.e., species previously grouped in subgenus Agrostoides (Hitche. € Chase ex Hsu) Zuloaga and Phanopyrum. Species of section Lorea Zuloaga (included by Zuloaga [1987] i currently considered among these taxa declared incertae sedis 2003). section Lorea includes | subgenus Phanopyrum) | are (Aliscioni et al.. As presently. circumscribed, 27 species, of which five are endemic to Venezuela and the Guayana Highlands: also. 21 species are restricted to central and southern Brazil. Here they inhabit the highlands of Bahia and Minas Gerais, the cerrados (s.l.) of central Brazil to coastal regions of Rio de Janeiro and Bahia, where a few species occur in specialized cerrados, known as restingas, such as Panicum restingae Renvoize & Zuloaga. P. sacciolepoides Renvoize & Zuloaga. and P. marauense Renvoize € Zuloaga (Zuloaga, 1987; Renvoize & Zuloaga, 1995). Panicum chnoodes Trin. is the only species known to be disjunct between the Guayana Highlands and central Brazil. Panicum sect. Lorea is distinguished by its tussock habit, pungent leaf blades, and usually indistinct junction of sheath and blade (Renvoize, 1978; Renvoize & Zuloaga. 1984). Plants are caespitose, perennial, with basal persistent sheaths, with their blades lanceolate or linear-lanceolate, either pilose or glabrous (Renvoize & Zu been mentioned previously to occur in the sheaths anc -r oaga, 1984). The presence of viscose elements has [om around the ligular region in only two species of this section, P. raginiviscosum Renvoize & Zuloaga and P. acicularifolium & Zuloaga Zuloaga, 1995). Until now, no anatomical or histochem- Renvoize ho (Renvoize 'The authors thank Gerrit Davidse for a critical review of this contribution, and graphic designer Tomás E assistanc e; Cátedra de Botánica Agrícola, Buenos Aires, ‘ ? Laboratorio de Naona Vegetal. Departamento de Biodi y Naturales, Univ ersidad de I "Instituto de Botánica fzuloaga@darwin.edu.a doi: 10.341 7/2006084 reentina. avega@; agro uba.ar = Darwinion. Labardén 200, Cas Facultad de Agronomia, Universidad de Buenos Aires, iversidad y Biologia Experimental, Suenos Aires, C1428E£ HA, Buenos Aires, illa Aversa for his Av. San Martín 4453, C141 7DSE Facultad de Ciencias Exactas Argentina. macObg.fcen.uba.ar. le 22. BlI642HYD, Gorreo Buenos Aires, Argentina. ANN. Missouri Bor. Garp. 95: 511—519. PUBLISHED ON 23 SEPTEMBER 2008. Annals of the Missouri Botanical Garden ical studies have been made on any species of this group. We the epidermal cells, and the lipid nature of their secretions describe the anatomy of secreling is histochemically confirmed. Lipid secretions are anatomically confirmed for the first time in 10 species of Panicum sect. Lorea: P. acicularifolium, P. bahiense Renvoize, P. chnoodes, P. cipoense Renvoize & Send., durifolium Renvoize & Zuloaga, P. euprepes bo P. poliophyllum Renvoize & 2 P. molinioides Trin., Zuloaga, P. trinii Kunth, and P. vaginiviscosum. Essential oils in the Poaceae have been previously reported from roots, stems, leaves, or inflorescences of species of Anthoxanthum L. (= Hierochloe R. Br.). Bothriochloa Kuntze, Chrysopogon Trin. (= Vetiveria Bory), Cymbopogon Spreng., Elionurus Humb. & Bonpl. ex Willd., and Melica L. (Guenther, 1950; Arber, 1965; Pinder € Kerr, 1980; Watson & Dallwitz, 1992; Kaul & Vats, 1998). Most of these aromatic grasses follow the Cy metabolic pathway of carbon fixation, and approximately 8896 are members of the Andropogoneae, a tribe that spread globally by the late Miocene, possibly the period of greatest evolutionary diversity for aromatic grasses (Kaul & Vats, 1998). When ingested, making them more vulner- these compounds slow the growth of herbivores, able to predators and parasitoids (Coley € Barone, 1996), constituting a selective agent for plant defense i Also, Ellis (1990) reported the presence of tannin-like substances in 104 species of against herbivory. Poaceae, with many representatives in the Andropo- goneae and Arundinelleae, and occasionally in a few members of the Paniceae (two species of Digitaria Haller, three species of Echinochloa P. Beauv., and one species of Panicum: P. coloratum L.). Ellis (1990) mentioned that, chemically, tannins are a heteroge- neous group of phenol derivatives. Previous studies carried out in the Poaceae showed that secretory substances are produced and accumu- ated in microhairs (Amarasinghe & Watson, 1988), in a the single cells adjacent to photosynthetic and non- photosynthetic tissues and between vascular bundles 1998), Davidse, in elaiosomes (Bresinsky, 1987; different types of multicellular 1990). The multicellular gland is composed of epidermal cells (Lewinsohn et al., 1963; Berg, 1985; 2000), or glands (Linder et al., Morrone et al., by four first forming a pad of cushion-based macrohairs (Bowden, 1971; Ellis, 1979); the second type is formed by numerous basal cells constituting the gland but not associated with macrohairs (Nicora, 1941; Davidse, 1988; Linder et al., 1990); the third type is comprised of multicellular glandular hairs (Kabuye & Wood, 1969); and the fourth type occurs as multicellular stalked glands with a central depression (Davidse, 1988; Zuloaga & Sendulsky, 1988 type of Two types of glandular macrohairs were reported in Panicum (Kabuye € Wood, 1969): unicellular with swollen tips. present in P. deustum Thunb., and multicellular clavate glandular hairs, present in six Both types contain a Multicel- lular glands were also reported on the lower lemma in . Stolonifera Hitche. € Chase 1988), species of tropical East Africa. yellowish and sticky substance at maturity. species of Panicum secl ex Pilg. (Zuloaga & Sendulsky. activity was not demonstrated. but secretory The objective of this paper is to determine the nature and histochemical localization of the secretion and to describe the secretory epidermal cells in Brazilian species of Panicum sect. Lorea. MATERIALS AND METHODS PLANT MATERIALS rts Herbarium. specimens of 16 species of Panicum sect. Lorea were studied in order to detect specialized secreting epidermal cells (see Appendix 1). ANATOMICAL STUDIES a Sheaths and blades were fixed in FAA (1 formalin: 4 alcohol 96%: 1 acetic acid) for 48 hours, and then 70% ethanol, herbarium specimens and rehydrated in 5% Contrad stored in or they were removed from 70 (Decon Laboratories. King of Prussia, Pennsylva- nia, U.S.A.) (Schmid & Turner, 1977) over 24. hours at 20 C. The material was desilicificated in hydrofluoric acid (596) for 24 hours. Then the material was washed in distilled water, dehydrated in an ethanol series, and Transverse and longitudinal embedded in paraffin. sections 10 um thick were cut on a rotary microtome. Sections were double stained with safranin-fast green. Preliminary observations of the leaf material stained with Sudan IV (Gahan, 1984) showed that they are rich in lipids. Stained transversal sections of the leaf sheath were analyzed to detect and determine the localization of lipid secretions in the different tissues. We used only herbarium material for this study. Light microscope (LM) studies were made using a ZEISS Phomi HL microscope (Zeiss, Oberkoc hen, Germany), with a 35 mm photographic camera and Kodak (Rochester, New York, U.S.A.) Gold ASALOO film. For observation with SEM, pieces of 1 — ae middle portion of the leaf sheaths were mounted on stubs, carbon coated in a vacuum evaporator, and coated with a palladium alloy. The observations were made using a ZEISS DSM 940 A microscope at the Instituto Argentina. Complementary de Botánica Darwinion, observations of secretory structures were made with an Volume 95, Number 3 2008 Vega et al. 518 Lipid Secretions in Panicum Sect. Lorea Figure 1. Panicum eat inte growing in Caraca, tic A. showing d leaf sheaths and young innovations. — Fonseca & Alvarenga 2109, Sp d i evidences 7 fire (arrov (Zuloaga et al. 4é structure 350, SI). with secretory environmental SEM in an XL30 ESEM (Phillips, Eindhoven, The Netherlands) at the Instituto Nacional de Tecnología Industrial (INTI), Argentina. RESULTS FIELD OBSERVATIONS Plants of Panicum sect. Lorea collected from Brazil grow in campos rupestres, rocky grasslands, at 1070— 1600 m. culms, covered with young leaves, are protected by In all of the examined material, the distal ;. Herbarium specimen of P. subtiramulosum (C s (arrows) in the cataphvll. Minas Gerais, Brazil. —B. Base of the plant of P. poliophyllum, , Brazil; P. der i from Balls Brazil Scale bars: A, B = 10 « D= eS araça, Minas Gerais y). —D. Herbarium specimen of je! = ] cm; cm. abundant and dry basal sheaths, constituting an imporlant source of combustible material during fires (Fig. LA, B). In herbarium specimens, burned leaf remains can be in the cerrado landscape some readily observed, and burning is a regular phenom- enon in cerrado environments in central Brazil. When these grasses are burned, young culms are protected at their base by cataphylls and remnants of basal burned sheaths (Fig. 1C, D). Field observations of some species of Panicum sect. Lorea showed the presence of large amounts of sticky secretions, with a and smell. strong disagreeable 514 Annals o Missouri Botanical Garden Table 1. Localization of lipid secretions for 10 of the 16 species investigated in Panicum sect. Lorea with Sudan IV in transverse sections of leaf sheath. + = present, — = absen Epidermis . Specialized secreting Species Sheath indumentum Mesophyll Abaxial Adaxial structures P. acicularifolium villous + $ + P. bahiense villous = + =+ + (unicellular) P. chnoodes villous zn + ++ P. cipoense villous + + = = P. durifolium villous + + ++ = P. euprepes subelabrous + + + + (unicellular) P. molinioides villous + ++ + + (multicellular) P. poliophyllum subglabrous + + = + (unicellular) P. trinii villous + + ++ P. vaginiviscosum subglabrous + + + These secretions are estricted to leaf sheaths and sheaths, e.g.. P. acicularifolium, P. bahiense, P. ligular regions, which turn viscid. Usually, secretions — chnoodes, P. cipoense, P. durifolium, P. molintoides, are more abundant in cataphylls (Fig. 1D) and basal and P. trinii, show conspicuous costal and intercostal sheaths while young plants are growing, but they are zones (Fig. 2A). Macrohairs are restricted to intercos- almost absent in upper leaf sheaths. Basal sheaths are tal areas (Fig. 2A-C) and are unicellular, stiff, with strongly attached to the culms, and marginal hairs are slightly thickened walls, and their bases are sunken impregnated with sticky secretions that enhance the and surrounded by a group of secreting epidermal adhesion of sheaths to the culms. cells slightly raised above the general level of the sheath surface (Fig. 2C). This type of macrohair is known as a € cushion hair (Metcalfe, 1960: Ellis. 1979). Species in Panicum sect. Lorea with subglabrous ANATOMICAL STUDIES Basal sheaths are villous all over or pilose on the basal le as in P. euprepes, P. poliophyllum. and margins and subglabrous on the surface (Table D). — P. raginiviscosum show few undulations associated Species in Panicum sect. Lorea with villous basal with vascular bundles. Solitary or grouped secreting | | | i A | | i | Q Figure 2. SEM microphotographs of the leaf sheath in Panicum sect. Lorea. —A. General view of the abaxial surface = showing the position of macrohairs at aie zones. —B. Macrohairs with specialized epidermal cells in young leal specialized e epide rmal cells. —F. Groups of speci d epidermal cells after secretion. A. B. D-F: P. euprepes (Zuloaga et al. 4783, SI); ( speck s d enide ia cells. Scale bars: A-C = ds um: D. sheaths. —C. Macrohairs with dt di specialized epidermal cells after secretion in mature leaf sheaths. —D. E. Groups of D ¿Po vaginiviscosum ies et al. SID. ez. costal zone; iz. intercostal zone; mh, macrohairs: p. prickle: se. AC = 100 um: F = 2 ne] -— Lum Volume 95, Number 3 2008 Vega et al. 515 Lipid Secretions in Panicum Sect. Lorea Figure 3. 2 ‘tive cells stained with Sudan IV. Note e .D. E. sec reting structures in leaf laioplasts at a ctive cells stained with safranin-fast green. ip. Ta sheath transverse sect —G. transverse tion. a ae ee & on 4701, SI). metaxylem; ph m; G — 50, — pem specialized epidermal cells are present in both surfaces, although they are more frequent on the abaxial surface and often surround the bases of the (Fig. 2D—P). cells although secretions are detected in fresh macrohairs These are not so abundant, and herbarium material. All species examined in transverse section show a unistratified epidermis composed of quadrangular to LM photographs of the leaf sheath in Panicum sect. Lores 1. —A. Tr: : P. vaginiviscosum inre et a 4781, Sl): B-E: P. ¡A ab, abaxial epidermis: ad, adaxial epidermis pd m: s, sclerenchyma; se, specialized epidermal cells; vb, y 1sverse section stained with Sudan IV. q ictive and Vignified cells, staining negatively to Su m —F. Multicellular- E a multicellular- secreting e in leaf sheath a et al. 4783, SD; F-G: P ;e E me, oe mx, ba -F ansverse section. —E. Longitudinal section. Detai ; euprepes vascular d s; x, xylem. Scale ba tabular cells (Fig. 3). On both surfaces, the epidermal cell The chlorenchyma has isodiametric to irregularly shaped cells have a thick outer tangential wall. cells. In viscid sheaths, only mesophyll cells contain abundant elaioplasts with lipids. The sclerenchyma is subepidermic and discontinuous, and forms adaxial and abaxial girders associated with primary and secondary vascular bundles. Fibers are lignified. 516 Annals of the Missouri Botanical Garden Transverse sections of the leaf sheaths observed from herbarium material of Panicum sect. Lorea show a positive reaction to Sudan IV. Lipids are accumu- lated in the mesophyll and also in the xylem and — phloem parenchyma of the vascular bundles (Fig. 3A). Lipids are especially abundant inside the specialized epidermal cells (Fig. 3B). However, we did nol observe specialized pores through which lipids are secreted. Specialized epidermal cells. once trans- formed in inactive cells, are persistent on the sheath surface and easily detected under magnification. by their lignified and brownish-colored walls (Fig. 3C). In transverse and longitudinal sections of the leaf sheaths, specialized secretory epidermal cells are elobose to claviform in shape, according to their density on the surface (Fig. 3D, E). Their walls are thin and lignified, apparently not interrupted by specialized pores. The epidermal origin of these secreting cells is clearly demonstrated. The cytoplasm of these cells ts dense, with abundant vesicles of different sizes. The nucleus is prominent and of variable placement with Fig. 3D, E). The presence of lipids was detected in 10 species: one or two visible nucleoli Panicum acicularifolium, P. bahiense, P. chnoodes, P. cipoense, P. durifolium, P. euprepes, P. molinioides, P. poliophyllum, P. trinii, and P. vaginiviscosum but not in the remaining six analyzed in section Lorea: P. Renvoize, P. lagostachyum Renvoize & Tunt. oo Swallen, P. restingae, and P. subtiramulosum Renvoize & Zu- animarum Zuloaga, P. loreum lutzit loaga. Lipids were detected in mesophyll and both epidermides with different accumulations in the adaxial and abaxial surfaces according to the species (Table 1). Lipids appear abundant on adaxial epider- mis of P. bahiense, P. durifolium, P. trinii, and P. vaginiviscosum. The presence of unicellular special- ized epidermal cells was observed in P. bahtense, P. and P. specialized secretory cells are distributed in intercos- euprepes, poliophyllum. ln P. bahiense, tal zones and are associated with macrohairs. Panicum molinioides possesses multicellular secretory struc- tures. distributed in abaxial costal and intercostal zones that are generally not associated with macro- hairs. In P. molinioides, glands are composed of a group of cells with a slightly concave central area. In transverse. sections, these cells are axially elongated and are all surrounded by a distinct and thick cuticle (Fig. 3F, G) DISCUSSION Fire has occurred in tropical savannas for thou- sands of years, shaping the landscape and selecting for adapted flora and fauna (Ramos Neto & Pivello, 2000). Many studies have concentrated on how plant » the Brazilian cerrados species adapt to fire i (Coutinho, 1990). However, the herbaceous layer of the cerrado has not been studied as intensively as the woody laver (Mistry, 1998), and more knowledge is needed about the mechanism put in place in native herbaceous plants to survive herbivory after fire. The tropical humid climate of the cerrado region has a dry winter when grasses dry out, and a wet summer. The burning season occurs from May to September, when the herbaceous vegetation is dry and more flammable. In the early wet season (September— October). fire occurrences decrease. although the vegetation is still capable of maintaining a fire (Ramos Neto & Pivello, 2000). Soils under cerrado vegetation are generally very poor in mineral nutrients, with toxic levels of aluminum and high acidity (Coutinho, 1982). g : Alter fires, the ash is highly beneficial to the growth of herbaceous and undershrub plants with superficial rool systems, since they are provided with a large significant 1990). The ground layer plants are xeromorphic, usually with quantity of mineral nutrients. and a reduction in aluminum toxicity (Coutinho, rather stiff leaves with smooth or harsh surfaces, although some grasses and forbs have soft, hairy leaves (Fiten. 1982). Plants growing in a cerrado sensu stricto (with tree and serub forms, where trees do not form a continuous canopy, Eiten, 1982) are mostly perennials, have a eray and dusty appearance, and possess hard siliceous leaves, sometimes even densely hairy, and with underground organs well adapted to burning. Further- more, in species of Panicum sect. Lorea, basal sheaths are strongly attached to the culms, and marginal hairs are impregnated with sticky secretions, enhancing the adhesion of sheaths to the culms. These characteris- lies may represent a strategy to decrease the entrance of insects inside the leaf sheath, and perhaps also to reduce dehydration and combustibility. Fresh green leaves are produced shortly after burning: intense flowering can also be observed d few days or weeks alter 1982: Gottsberger, 1984; Miranda et al., 2002). According (2002), a peak in abundance is associated with the flush of new leaves cerrado fires (Coutinho, Silberbauer- to Marquis et al. herbivore during the initial part of the wet season. Furthermore, cerrado plants most commonly attacked by insects can evade these herbivores by evolving chemically novel toxins as deterrents to distinguish themselves from 1990) postulated that the presence of tannin-like substances their neighbors (Marquis et al., 2002). Ellis (TLS) in species of grasses is primarily associated with sour grasslands and savannas of southern. Africa, and that this feature may be related to a chemical defense mechanism as a response to damage caused by herbivores. Volume 95, Number 3 2008 Vega et al. 517 Lipid Secretions in Panicum Sect. Lorea In species of Panicum sect. Lorea, the production of lipids, combined with a repellent substance, might well be associated with the sprouting of fresh green leaves, making the foliage unpalatable for grazing animals or predators. That association would account for the occurrence of secreting cells only on the cataphylls, leaf sheaths, and ligular regions of basal — foliage, but not in leaves of the upper culms. In 10 o the 16 analyzed species of section Lorea, of a total of 27 species included in this section, lipid secretions are concentrated in epidermal cells, especially at the bases of the macrohairs, and in mesophyll. Globose- secreting cells are not so abundant, although secretions were detected in fresh material during Although the presence of sticky substances and specialized secret- fieldwork and in herbarium material. = ing epidermal cells was not detected in three of the studied species of this section, the vast majority contained both and are likely to occur in other unstudied taxa of section Lorea. In general, species of this section occur in areas seldom visited by botanists, making collections scarce (Renvoize, 1978), but it is important to collect them with the complete basal portion to locate secretory structures or substances. We have that Panicum sect. Lorea have important differences in shown the analyzed species of the presence and type of secretory structures. Lipids [em are present in the epidermis and mesophyll in 10 of the 16 species studied. Specialized cells are present or absent and, when present, are unicellular or grouped in multicellular structures. Little information has been published in relation to interactions among other genera of lipid-secreting grasses and herbivores. In some species of the genus Pentaschistis (Nees) Spach inhabiting southern Africa and Madagascar, shoots are aromatic (or fetid) due to the presence of multicellular glands at the base of the macrohairs on the abaxial leaf blade surface, sheaths, pedicels, or glumes (Linder et al., 1990; Watson & Dallwitz, 1992). as species of Prio- nanthium Desv., that are sticky and/or produce an Plants, such unpleasant smell originating from an unknown volatile substance, are postulated to have an anti-herbivore 1988). type of macrohair to that described in some species of mechanism (Davidse, Furthermore, a similar pm Panicum sect. Lorea was reported in Andropogon gayanus Kunth var. bisquamulatus (Hochst.) Hack. (Andropogoneae) by Bowden (1971), but this author interpreted. that the epidermal cells that surround macrohair bases are nectaries. Most of the aromatic grasses studied up to the present are Andropogoneae and follow the C4 pathway 1998). The Panicum, s.l., includes all photosynthetic types: Ca, (NAD malic enzyme [NAD-ME]. Y carbon fixation (Kaul € Vats. genus »hosphoenolpy- | | p? ruvate carboxykinase [PCK], or NADP malic enzyme [NADP-ME]. and also C3/Cy intermediate species. This analysis shows that many species of Panicum which includes only C4 species, have sect. Lorea, essential oils. Aliscioni et al. (2003) have suggested that Panicum sect. Lorea should be segregated from Panicum, because the species of this section that were studied were placed in an independent and strongly supported clade separate from Panicum s. str. As previously mentioned, species of this section share a similar habitat and distribution, as well as a suite of morphological characters. The presence of secretory tissue in species of this section may represent another unique character for this group. Nevertheless, the presence of secretory tissues should be analyzed in more species of Lorea and in new material of those species where glands were not found in this study, to confirm if this character represents a distinguishing feature for the group. ADDENDUM While this paper was in press, a new contribution (Sede et al., of section Lorea, which led to the segregation of two 2008) presented a phylogenetic analysis new genera in the Paniceae. Species discussed here were rearranged in both Apochloa Zuloaga & Morrone and Renvoizea Zuloaga & Morrone. Literature Cited Aliscioni, 5. S. L. M. Giussani, F. O. Zuloaga & E. A. Kellogg. 2003. A molecular phylogeny of Panicum (Poaceae: Paniceae): Tests of monophyly and phylogenetic placement within the Panicoideae. Amer. J. Bot. 90: 796— 821. Amarasinghe, Watson. 1988. M bush ultra- structure of microhairs in grasses. Bot. J. Linn. Soc. 98: 303-319. Arber, A. 1965. The Gramineae. A study of cereal, bamboo and grass. Historia i6. Naturalis Classica XLI. J. Cramer, Weinheim, New York Berg, R. Y. 1985. Spike let structure in Panicum australiense (Poaceae): Taxonomie and ecological implications. Aus- tral. J. Bot. 33: 579-58: Bowden. B. N. 1971. Studies on Andropogon gayanus Kunth VI. The claries nt of Andropogon var. bisquamulatus (Hochst.) Hack. (Gramineae). gayanus t. J. Linn. = Soc. 64: 77-80 Bresinsky, A. 1963. Bau, Entwicklungsgeschichte und rmekochorer Bot. ps Zur mv Biblioth. Inhaltsstoffe der Elaiosomen. Studie Verbreitung von Samen und Früchter n. A. Renvoize. 1986. Genera Graminum. y Bulletin additional series Clay 'lon, W. Grasses i ne lor 1996. Herbivory E oos Ecol. m2 Goles e D. & J: An de fe snses in tropical forests. 305- Annual Rev. Annals of the Missouri Botanical Garden Coutinho, L. M. 1982. Ecological effects of xi in Brazilian 273- 29! in B. J. Huntley & B. H. Walke . Ecology of Savannas. ded [ E ger ` Tropical S Verlag. - 1990. Fire in the ecology of the Brazilian cerrado. /n Biota: sses, and Global Challenges. Springer- Goldammer (editor). Fire in the Tropical Ecosyste m. Proc ul diepe "sal in the Poaceae. S.C b I ^ A Cra ass Systematics and E volution. Washington., la . Hilu. ( p med (editors). Sudan Institution. Press, 1988. revision of the genus el Bothalia 18 3153. (Poaceae: Arundineae). Eiten, G. 1982. Brazilian savannas. EN 20-17 in B. Huntley & B. Walker (editors). "urs Springer-Verlag, New Yor Ellis. I leaf anatomy in the Poaceae. Bothalia 12: dae J. Ecology of Tropical 1979. A procedure standardizing comparative The epidermis as seen in surface view. O JO. o like asa ‘es In grass leaves. Mem. Jot. Surv. South Africa 59: 1— p p. 3. 1984. e Histoe mistry and Cytochemistry An ell J> Introduction. Experimental Botany, Series of Es ay .. 18. Academic Press. London. Individual Essential Oils of the Plant Families al "ac Guenther. E. 19% a Lauraceae, Burseraceae, Myrtaceae. Umbelliferae and Geraniaceae, Vol. 4. The Essential Oils. D. Van Nostrand Company, Ine., New York Kabuye, C. H. S. € D. Wood. 1969. A first record ax es glandular hairs in the Gramineae. Bot. Linn. Soc. 62: 69—70. Kaul, V. K. & S. K. Vals. 1998. Bothriochloa pertusa We ny letic relationship in aromatic Essential oil composition of erasses. Biochem. Syst. Ecol. 26: E | pee Lewinsohn. E.. N. Dudal. A ‘admo, l. Katzie, U. Ravid, E. Putievsky & D. M. Joel. 1998. o mical localization of citral accumulation in Lemongrass leaves ( e citratus (DC.) Stapf, Poaceae). Ann. Bot. (Oxford) 81: 39. Linder, H. P., The occurrence, anatol ny, and s & M. 1990. ystematic implications of Thompson, R. P. Ellis Perold. the glands in Pentaschistis and O (Poaceae, Arundine sae). Bot. Gaz. 151: 221-233. . Morais & I. R. "i 2002. Inte ractions among cerrado plan and oe herbivores: typical? Pp. 306-328 in sd Oliveira & R. (editors). The Cerrados razil. Ec History of a Ne poopie an Aes dinoides "dae. Marquis, R. J.. H. logy and Natural Savanna. Columbia University 1960. Anatomy of the Monocotyledons. I. lare ndon Pre . United Kingdom. C. Miranda. 2002. Oliveira & R. J. Marquis (editors), Natural a Neotropical Savanna. Columbia University Press, New . Oxforc ista The Cerrados. of Ecology and History — cT rk. Mistry, J. 1998. Fire in the cerrado (savannas) of Brazil: An ecological review. Progress Phys. Geogr. 22: 425-448 Morrone, O., A. S. Vega € M. Maier. 2000. Urochloa paucispicata (Poaceae: Anatomy and « Nicora. E. 6. 19 elándulas e pidé rmicas de algunas Darwiniana 5: 316-321. i e A. R. & S. K. Kerr. 1980. The + : Bothriochloa spec les, N Elaiosomes in Panicoideae: ee iceae): hemical eds Flora 195: 303—310 . Gontribución al estudio inci o de las especies de Eragrostis. olatile esse ntial De of | ‘ 11- LS Phytochemistry Ramos Neto, M. B. & V. R. Pivello. 2000. Lightning fires in a Brazilian savanna national park: Rethinking management 26: 075-084. strategies. Environm. Managem. senus Panicum group. Lorea (des 1978. The g Renvoize. ‘ Ke w Bull 32: 419-428 0 (Gramineae). & F . The genus ÓN group Bull 30: 185- — . Zuloaga. Lorea (Gramineae). Kew á 1995. . Three new species of. Panicum eroup Lorea \Cramitte ae) from the Pico das arin Bahia. Brazil. E W Pur 50: 161-164. Schmid. ies 1977. Contrad 70, an a softener a s bud material for anatomical study. Taxon 26: 55]- Sede. 5. M. c Morrone, L. M. Giussani & F. O. Zuloaga. 2008. ie ean studies in the Paniceae (Poaceae): A re ae nt of section Lorea of Panicum. Syst. Bot. 33(2): 300. nie ee r-Gottsberger, 1. 1984. Fruit dispersal and rypa- Pl. Syst. Ex 17 nocarpy in Brazilian ee grasses, Pl. Syst. Evol. | 1-27 Simon. n. K: . Jacobs. 2003. Megathyrsus. a new generic name = Pbi subgenus Megathyrsus. Aus- trobaileya 6: 571-57 . & M. J. Descriptions. 19924). G Illustrations, Ballwin: ( rass Genera of the Identification. and Information Retrieval: including Synonyms. Morphology. Physiology, Phytochemistry, Cytology. Cl: World and Local Distribution, and , ac- \natomy, assi i- Pathoge ans; cessed 18 August 1999, muons E O. 1987. Systematics of New World species of (Poaceae: Pd Pp. 287-306 in T. R. Hilu. C. S. Campbell & M. E. Grass Systematics and — = Soderstrom, K. Barkworth (editors). Evolution. Smithsonian Institution Press, Washington, D.C. Å € \ ^ I Sendulsky. 38. A revision of Panicum subgenus i a section Eo (Poaceae: bd . Missouri Bot. Gard. 20455. 1998. análisis cladisiten de car cas rai A Missouri Bot. LM . 8. Vega & L. B Giussani. ms Gard. Panic oide: wae: Paniceae). Ann. O 631-65 APPENDIX D. species studied to detect specialized secreting epidermal Herbarium specimens of Panicum sect. Lorea P. acicularifolium Renvoize & Zuloaga. BRAZIL. Bahia: Mun. Pico das Almas. vertente leste, subida do pico do campo do Queiroz, Harley et al. 26321 (isotype. SI). Rio de Contas: P. animarum Renvoize. BRAZIL. Bahia: Mucugé, Giulietti et al. 1537 (SD: Mun. Rio de Contas. subida a Pico das Almas, 13 35'5 'W, 1450 m, 15 Feb. 1994, Zuloaga et al. 4848 (SI bahiense Renvoize. BRAZIL. Bahia: Mun. Rio de Pico das Almas, Zuloaga et al. 1850 (SI). lrin. BRAZIL. Minas Gerais: Serra do à “po, : Morro do Pilar, Vidal 5996 (Sl) send. BRAZIL. Minas Gerais: rdeal Mota a Co onceigao do Mato Dentro, BR- fuloaga & Morrone 4694 (SI): Mpio. pue da rodavia Lagoa Santa. Conceição d o Mato Dentro, Diamantina, Senduls shy et al. 389 T type; i : : duifolium Renvoize & We BRAZIL. ll k "i de Contas: m ao e) En N ^ zo P. aed: Bahia: N da cidade na estrada Mun. para 10) Volume 95, Number 3 2008 Vega et al. 519 Lipid Secretions in Panicum Sect. Lorea pe de Mato Grosso, 13 30'8, 41 52'W. Harley et al. 3 (isoty ye SI). Palmeiras ^ Zuloaga el BRAZIL. Bahia: nacio, camp Aa e s Renvoize. Q = A ON ). Se rra n rn ‘po, Km 118 Ey = © m a a ^. o — — > = ~ o = a — ON Xe) O = | Renvoize € Zuloaga. BRAZIL. Bahia: Pereira 2139 (SI). BRAZIL. Minas Gerais Dentro, E ROOM: zi BR-O10, ao lon ovi c onceição do Mato Dento D amanina, Sd et al. 448 SI). P. lutz Castellanos s.n. (SD. ii Swallen. BRAZIL. Rio de Janeiro: Pedra Bonita. P. molinioides Trin. BRAZIL. Minas Gerais: Rodovia d Cardeal Mota a Conceição do Mato Dentro. a O10, Cipó, Km 121, Zuloaga & Morrone 4701 | Serra do P. perap un Renvoize & Zuloaga. P AZIL. Minas verais: Ouro Branco, Zuloaga s.n. (SI). P. re. DA ndn & Zuloaga. BR. ZIL. E spirito Santo: o Km 32 da ESO60-Setiba, Silva 579 (SI). bi on d Renvoize & Zuloaga. BRAZIL. Distrito E ral Reserva Ecol. do IBGE, próximo ao Córrego Taquara, Fonseca & . nm arenga 2109 (SI); APA Ga G nate fee r, Filgueiras & Zuloaga 2132 rinii Kunth. BRAZIL. Bahia: Mun. Rio de Contas: Pico y et al. ma—C Cabeca de ves ido, 2 (SI). que liga Serra Grande, ramal 13 que lleva ao Campinho Cheiroso, Amorim et al. 727 (CEPEC). Distrito Federal: Restinga de Jacarepaguá, Duarte 5014 (SI). aginiviscosum Renvoize & Zuloaga. BR s Bahia: Mun. Ilheus, ruta de Olivença a Una, 17 km de Olivença, en borde de selva, Zuloaga et al. 4853 (SI); Mun. Palmeiras, Pai us io. Morro do Pai Inacio, campo rupestre, Zuloaga el n 781 (SD: Mun. Rio de Contas. Pico das Almas, verte Sites n Faz. Silvina-Queiroz, Had et al. 25772 (isotype, EB A SYNOPSIS OF AN EXPANDED Ji-pel as 2% Hang Sun,” Jian-Hua Li,' and SOLMS-LAUBACHIA Ihsan A, Al-Shehbaz” (BRASSICACEAE), AND THE DESCRIPTION OF FOUR NEW SPECIES FROM WESTERN CHINA! ABSTRACT Sequence data from the nuclear ribosomal ITS region and the chloroplast trnL-F were used pi examine the generic delimitations of Solms- laubachia Muschl.. Christolea Cambess., Desdera Pamp., Leiospora (C. A. y.) F. Dvořák, and Phaeonychium O. E. Schulz. Solms-laubachia, Desideria, and P. jafrii Al-Shehbaz formed a well- sipped monophyletic clade, with Christolea as sister group. However, both Solms- laubachia and Desideria were polvphyletic, as they appeared in more than two positions in that clade, within which P. jafrii was embedded. The results are consistent with those based on sequences of plastid maturase (matK) and the nuclear chalcone rape in SEM survey further reveals that the seed epidermis micromorphology of D. baiogoinensis (K. C. Kuan & C. H. \l-Shehbaz is most similar to that of S. lanata Botsch. Based on these findings, a comprehensive synopsis of an e onu Seine. laubachia with 26 species is presented; all species of pie and P. jafrii are transferred to Solms-laubachia resulting in 12 new combinations: 5. baiogoinensis (K. C Kuan & C. H. An) J. P. Yue, M-Shehbaz € H. Sun, S. flabellata (Regel) J. P. Yue, Al-Shehbaz & H. Sun. S. haranensis (M- E J: E Yue, Al-Shehbaz & H. Sun. 5. himalayesi (Cambess.) J. P. Yue. Al-Shehbaz & H. Sun. S. incana (Ovez.) J. P. Al-Shehbaz € H. Sun, 5. jafrii (Al-Shehbaz) J. | . Al-Shehbaz & H. Sun, S. linearis (N. r J.P. Yue, Al- She hbaz ^ n Sun, 5. a e x d hbaz) J. P. Yue, Al- EE z & H. Sun, 5. mirabilis Camp ) J. P. Yue, Al-Shehbaz € H. Sun, 5. nepalensis (H. Hara) J. | , Al-Shehbaz € H. Sun, S. prolifera (Maxim) J. P. Y Al-Shehbaz & H Sun, and 5. stewartii (T. Anderson) J. P. Yue, Al- > ibus & H. Sun. Furthermore, four new species of Solms- laubachia are described from western China: S. angustifolia J. P. Yue, Al-Shehbaz & H. Sun, S. grandiflora " d E Al-Shehbaz € H. Sun, and S. sunhangiana J. P. Yue € Al-Shehbaz (all from Sichuan Province), si 5. calcicola J. P. Yue, "As Shehbaz € H. Sun (from Xizang). Both S. linearifolia (W. W. Sm.) O. E. Schulz and S. mirabilis are lectotypifie n Key words: Brassicaceae, China, Christolea, Desideria, Leiospora. ITS, IUCN Red List. Phaeonychium. Solms- laubachia. trnL-F. Solms-laubachia Muschl. (Brassicaceae) is a Sino- & Cheo, 1981; Al-Shehbaz & Yang, 2001). Except for Himalayan endemic genus characterized by the S5. platycarpa (Hook. f. € Thomson) Botsch., which perennial habit, entire leaves, trichomes absent or also occurs in Bhutan and Sikkim, the other eight simple, latiseptate fruits detached at maturity from the species previously described in the genus are highly pedicel, entire capitate stigmas, and rounded replum restricted to the alpine areas of northwestern Yunnan, concealed by the strongly angled valve margins (Lan western Sichuan, and eastern Tibet, where they grow ! Portions. of this paper are taken from a thesis entitled “Phylogeny of the genus Solms-laubachia (Brassicaceae)" an submitted by the first author in partial fulfillment of a doctorate degree to the De do nt of Phytogeography and Ecology of Kunming Institute of Botany, Chinese Academy of Sciences. The new species were collected during field investigations in the a ngduan Mountain region of China, with support from the U.S. National Science Foundation (grant no. DEB-0321846 to Javid E. Boufford) 2 the National Natural Science Foundation of China (grant nos. 30625004, 40332021. 40771073. and SUIS to H. Lab work was partially supported by the Innovation Project of the Chinese Academy of Sciences (KSCX2-Y W-Z-030) en a Mercer Fe ‘Hlowship from the Arnold m tum to J. P. Yue. Fieldwork in Xinjiang by L Al-Shehbaz was supported by the National Geographic Society (grant 7405-03). We thank Jason R. Grant for providing DNA of three accessions of Desideria and one of Leiospora. We are grateful to a ictoria C. Hollowell, Beth Parada, and Allison Brock for their editorial advice, to Barbara Ne B^ r and Karol Marhold for their critical reviews of the manuscript. and to Fred Keusenkothen for le ‘ttering the seed figures. Last, but certainly not least, thank the directors and curators of the herbaria cited. y Laboratory of Biodiversity and Biogeography, Kunming Institute of Botany, Chinese Academy of Sciences, Heilongtan. Kunming, Yunnan 650204, People’s Republic of China. hsun@mail.kib.ac.en. ! Present address: Botany Department, The Field Museum, 1400 S. Lake Shore Drive, Chicago, Ilinois 60605, U.S.A. Qo ele ea Org. "Arnold Arboretum and. Harvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02318, US.A.: College of Life Sciences, Zhejiang University, A 310029, People! s Republic of China. jlioeb.harvard.edu. "Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63106-0299, U.S.A. Author for correspondence: ihsan. al-shehbaz@mobot.org. doi : 10.34.17/20062 14: Es ANN. Missourt Bor. Garb. 95: 520—538. PUBLISHED ON 23 SEPTEMBER 2008. Volume 95, Number 3 2008 Yue et al. Solms-laubachia from Western China on scree slopes and in rock crevices (Lan & Cheo, 1981; Al-Shehbaz & Yang, 2001). Solms-laubachia was placed by Schulz (1936) in the tribe Matthioleae, a disposition followed by several Chinese taxonomists (e.g., Kuan, 1985; Lan. 1987; An, 1995; Li, 1995; Huang, 1997; Tan et al., 1999). However, Schulz's tribal classification of the Brassicaceae has been considered as highly artificial (Hedge. 1976: Al-Shehbaz, 1984; Koch et al., 2003; Mitchell-Olds et al., 2005; Al-Shehbaz et al.. 2006). Janchen (1942) united the Matthioleae with the Hesperideae, a by Al-Shehbaz (1984) who adopted the earlier tribal name Ancho- nieae and further suggested the exclusion of Solms- position followed laubachia and other genera from the combined tribe (Al-Shehbaz. 1988). In molecular phylogenetic study on the Brassicaceae, Beilstein et al. (2006) (with Sun as the representative) falls within a their. broadly sampled showed that Solms-laubachia Al-Shehbaz & H. well- S. zhongdianensis J. P. Yue, £ supported monophyletic clade containing species of Braya Sternb. & Hoppe, Desideria Pamp., Euclidium W. T. Aiton, Malcolmia W. T. Neotorularia Hedge & Sisymbriopsis Botsch. & Tzvelev, Rhammatophyllum O. E. Schulz, Shangri- laia Al-Shehbaz, J. P. Y Bunge. Based on the molecular results and morpho- logical evidence, Al-Shehbaz et al. (2006) tentatively others within the Aiton, Léonard, ue & H. Sun. and Tetracme assigned these 10 genera and 15 tribe Euclidieae. However, the generic delimitation within this tribe needs further study since some Neotorularia, Sisymbriopsis) have been 2004, genera (e.g.. shown to be polyphyletie (Warwick et al., 2007 Yue et al. (2006) used sequences of the nuclear a chalcone synthase (Chs) and plastid maturase (matK) to study the phylogenetic relationships of Solms- laubachia within the Brassicaceae. They found that the four Desideria species, D. baiogoinensis (K. C. Kuan & C. H. An) AL-Shehbaz, D. himalayensis (Cambess.) Al-Shehbaz, D. linearis (N. Busch) Al- Shehbaz, and D. stewartii (T. Anderson) Al-Shehbaz, as well as Phaeonychium jafrii Al-Shehbaz were nested within Solms-laubachia in a well-supported clade. Therefore, their data support an. expanded Solms-laubachia, including Desideria and P. jafrii. The present paper aims to further test the relation- ships of Solms-laubachia, Desideria, and P. jafrii using sequences of the nuclear ribosomal ITS and the chloroplast non-coding trnL-F region. Additionally, seed epidermis of 13 Solms-laubachia species was SEM. Finally, Solms-laubachia is presented in this surveyed under a comprehensive synopsis of paper based on the molecular and morphological data. MATERIALS AND METHODS PLANT MATERIAL Plant material used in this study is listed in Table 1. The ITS sequences were obtained from 19 species (22 accessions), including 13 of Solms- laubachia (l7 accessions, including four new species), two species of Desideria, one species of Christolea Cambess., one accession each of Leiospora pamirica (Botsch. & Vved.) Botsch. & Pachom. and Phaeonychium jafrii. Published ITS sequences of D. — incana (Ovez.) Al-Shehbaz and D. prolifera (Maxim. Al-Shehbaz (Table 1). The trnL-F sequences were obtained from were also included in the analyses 21 species (24 accessions), including 12 species of Solms-laubachia (16 accessions), four species of Desideria (five accessions), one accession of P. jafrii, species of Christolea, and two species of Leiospora (C. A. Mey.) F. Dvorák. Leiospora species one were chosen as outgroups because this genus was sister to the clade containing species of Christolea, Desideria, and Phaeonychium O. E. Schulz (Yue et al., 2006). SEM marked with an asterisk. Solms-laubachia, Taxa whose seeds were are listed in Table 1 and observed under DNA EXTRACTION, PCR, AND SEQUENCING Total DNAs were extracted from silica gel-dried leaves using a Qiagen DNeasy Plant Mini Kit (Santa Clarita, California, U.S.A) following the manufacturers protocol. Genomic DNA one of Leiospora were genomic from three [2m accessions of Desideria anc kindly provided by Jason R. Grant (Université de Neuchatel, Switzerland). Nuclear ribosomal DNA ITS. The nuclear ribosomal DNA (nrDNA) ITS region was amplified using primers ITS-4 of White et al. (1990) and ITS-leu of Baum et al. (1998). (PCR) polymerase buffer A 25-ul polymerase chain included 2.5 ul of 10X Taq , 4 ul of dNTP (2.5 mM), 3 ul of MgC lə (25 mM), | ul of each primer (10 uM), 0.3 ul of Taq polymerase "6 units/ul), 2 ul of dimethyl sulfoxide (DMSO), and 20-50 ng DNA. The PCR program included a 3-min. hotstart and 34 cycles 72°C. Each cycle consisted Dad sec. denaturing at 94^ C, 70 sec. reaction = followed by an additional 10 min. at annealing at 55°C-57°C, and a l-min. extension at "C. CR products were gel-purified using a Qiagen Gel Purification Kit, following the manufacturer's protocol. For sequencing, the same primers as for PCR were used; for some taxa, ITS-2 and ITS-3 of White et al. (1990) were used as well. =e) 522 Annals of the Missouri Botanical Garden Table | Vouchers and sources of species used in this study. An asterisk (*) epidermis under SEM. Dagger symbol (+) indicates sequence obtained from the GenBank database. indicates the taxa also studied for seed Taxon and accession Voucher Source ITS wrab-k *Solms-laubachia angustifolia J. P. Yue, Al-Shehbaz & H. Sun *Solms-laubachia oo K. C. Kuan & C. H. An) J. P. Yue. |= Desideria iuto *Solms-laubachia calcicola J. P. Yue. M-Shehbaz € H. Sun *Solms-laubachia eurycarpa (Maxim.) Botsc ne Solms-laubachia eurycarpa (Maxim. de Botsch.. *Solms-laubachia grandiflora J. P. Yue Al-Shehbaz & H. Sun Solms-laubachia laa (Cambess.) J. P. Yue. Al-Shehbaz € H. Sun |= Desideria himalayensis | y ue, un [7 Desideria incana] (Al-Shehbaz) Yue, Al-Shehbaz & H. Sun = Phaeonychium jafru . Ja Solms- laubachia incana (Ovez.) J. P. Shehbaz & H. 5 laa jafrii ( pe *Solms-laubachia lanata Botsch.. | Solms-laubachia lanata Botsch., 2 * Solms- taubachia linearifolia (W. W. Sm.) Schulz Solms- nw linearis (N. Busch) J. P. Yue, Al-Shehbaz & H. Sun. I |= Desideria linearis] Solms-laubachia linearis (N. Busch) J. P. Yue, Al-Shehbaz & H. Sun, 2 Desideria linearis] *Solms-laubachia minor Hand.-Mazz. *Solms-laubachia platycarpa cid ES Thomson) Botseh., 1 Solms-laubachia platycarpa (Hook. f. & Thomson) Botseh.. 2 rs laubachia prolifera (Maxim.) J. P. Yue, iehbaz & H. Sun [= Desideria prolifera] mo. pulcherrima Muschl. *Solms-laubachia retropilosa Botsch., | Solms-laubachia faa de Botsch.. Solms-laubachia stew (T. Anderson) J. P. Yue, Al-She Y & H. Sun [^ D is stewartii Solms- d sunhangiana J. P. Yue & \l-Shehbaz "olmo Du hia xerophyta (W. W. Sm.) Comber E laubachia eps J. P. Yue. \l-Shehbaz & H. Outgroup taxa Christolea crassifolia Cambess. Letospora exscapa (C. A. Mey.) F. Dvořák Leiospora pamirica (Botsch. & Vved.) Botsch. Pachom. t M-Shehbaz & E Sun Yue 0246 McBeath 2105 (E) Boufford et al. Yue 0250 (KUN) Yue 0246 (KUN) Boufford et al. 31975 (A) Yue 0249 - N) Yue 0158 (KUN) Boufford et al. 30727 (A) WeBenth 1486 (F) Yue 0233 (KUN) Yue 0234 (KUN) Yue 0237 (KUN) Yue 0157 (KUN) ado ^w et al. 9 (MO Stainton 3055 (E) Yue 0379 (KUN) Yue 0256 (KUN) Yue 0239 (KUN) Yue 0153 (KUN) Yue 0162 (KUN) (KUN) 33404 (A ) Yue 0251 (KUN) Yue 0156 (KUN) Bartholomew et al. 99 (MO) Murray et al. 457 ( A l LA ) Bartholomew et al. 790 (MO) China. Sichuan, Daocheng China. Tibet. Mozhugongka China, Tibet. Riwoqe China, Sichuan, Yading China. Yunnan, Deqing China, Sichuan Xiangcheng Nepal China. Tibet. Lhasa Tibet. Lhasa ‘bel, China, Yunnan, Deqing China, China, 1 Lhasa China, Xinjiang Pakistan China, Sichuan, Yanyuan China, Tibet, Dangxion China. Tibet. Dangxion China. Yunnan. Lijian China, Sichuan, Niancheng China, Tibet, Zuogong India China, Sichuan, Jiulong C hina, Yunnan, Zhongdian China, Ninjiang Russia China, Xinjiang DQ523420 DQ523319 DQ523416 DQ523315 DQ523421 DQ523320 DQ523404 DQ523405 DQ523419 DQ523303 DQ523304 DQ523318 2 DQ523327 00523422 D0523321 DQ523409 DQ523 DQ523 DQ523308 10 DQ523309 4 DQ523313 DQ523417 DQ523316 DQ523329 DQ523418 DQ523408 DQ523317 DQ523307 DQ523407 DQ523306 FAJO28333 = DQ523411 DQ523413 DQ523310 DQ523312 DQ523412 DQ523311 DQ522332t EU 186027 zx: DQ5234060 DQ523305 DQ523415 DQ523314 DQ523423 DQ523522 — D0523330 DQ523424 DQ523323 Volume 95, Number 3 2008 Yue et al. 523 Solms-laubachia from Western China Plastid trnL-F. al. (1991) was used to amplify the trnL intron and the trnL-trnF spacer. A 50-ul PCR reaction included 50-100 ng DNA, 5 ul of 10X Taq polymerase buffer, 4 ul of dNTP (2.5 mM), 4 ul of MgCly (25 mM), ? each primer (10 uM), and 0.5 ul of Taq The PCR profile used to the trnL-F region included a 3-min. o The primer pair e and f of Taberlet et intergenic 2 ul of polymerase (5 units/ul). hotstart and 34 cycles (94 C for 1 min., 55 C for 1 min., 72 € for 2 min.), and a final extension at 72 C for 7 min. PCR described above. Sequencing of the region was done amplify products were purified using the method using the PCR primers as well as primers d and e of Taberlet et al. (1991). Sequences were analyzed using an automated ABI 3100 (GMI, Inc.. sola, U.S.A.), then edited using Sequencher (version 3.0, Gene Codes Corp.. Ann Arbor, U.5.A.). and then aligned using ClustalX version 1.83 (PC version) (Thompson et al., 1997), Genetic Analyzer Ramsey. Minne- [2m Inc., Michigan, followed by manual adjustments. PHYLOGENETIC ANALYSES Phylogenetic analyses were performed for the data sets using maximum parsimony (MP) (Farris et al., 1970; Fitch, 1971). MP analysis were conducted using PAUP 4.0b10 (Swof- 2002) sequence addition with 1000 replicates and 10 trees The heuristic searches for ford, with the following conditions: random held at each step, tree bisection-reconnection (TBR) MULPARS on, and the steepest Character state changes were unordered — branch swapping, descent off. and equally weighted in the analysis. Gaps were treated as missing data. Indels in the trnL-F alignment were coded following Simmons and Ochoterena (2000). Bootstrap (BS) and jackknife analyses of 1000 replicates were performed respectively to the heuristic tree search options as in the MP analyses. estimate support for individual clades using It was not possible to obtain the ITS sequences for Desideria himalayensis, D. stewartii, D. linearis, and A. Mey.) F. Dvorak, produce the trnL-F sequences for D. prolifera, D. Leiospora exscapa (C. nor incana, and Solms-laubachia sunhangiana J. P. Yue & Al-Shehbaz. Except for these seven species, the ITS and trnL-F sequences of all other exemplar taxa were combined into a data set, and the incongruence 1994, as implemented in PAUP) was used to assess potential length difference (ILD) test (Farris et al., conflicts between the phylogenetic signals from the different DNA data sets. 1000 replicates were analyzed with a heuristic search, two For that test, each with 10 random sequence addition replicates. To ensure the accuracy of the P value in the ILD pare est, only parsimony informative positions were used. SEED EPIDERMIS Seeds of each species were mounted on aluminum stubs using double-sided adhesive and sputter-coated with gold-palladium. Observations were made using a KYKY-10000B (Science Instrument Company, Beijing, 15 kV. terminology of seed-coat epidermis follows Murley (1951). electronic China microscope at Descriptive RESULTS AND. DISCUSSION ITS SEQUENCE DATA Twenty-four sequences of the ITS region, including he The ipi alignment of the ITS region produced a 85 gene. were used in the analysis (Table 1). wal — data set of 622 bp, of which 514 bp were uninforma- live, 65 bp were variable but not parsimony informa- live, and only 43 bp were potentially parsimony informative. The ITS-1 region was variable at 59 of the 277 sites (21.3%). the ITS-2 was variable at 45 of the 221 sites (20.4%), and the 5.85 gene was variable at only 4 of the 124 sites (3.2%) The sequences were similar in length, and the alignment only required six indels with one, two, or ITS 5.7%-9.2% between the ingroup and outgroup and three bp. sequence divergence ranged from 0%-7.5% within the ingroup. The sequences of two accessions of Solms-laubachia lanata Botsch. were identical, and those of the single accessions of S. Al-Shehbaz & H. Sun and sunhangiana were indistinguishable. calcicola J. P. Yue, MP analysis yielded three most parsimonious trees (MPTs) of 150 steps (confidence interval [CI] = 0.82, retention index [RI] OAs, consistency index [RC] = 0.63) (Fig. f Solms-laubachia, ~ and rescaled 1). All sampled species of Desideria, and Phaeony- chium jafrii formed a monophyletic clade with 98% BS support. Within this clade, D. linearis was basal to the remaining taxa. Eleven species of Solms-laubachia formed a core clade with 71% BS support. Within core Solms-laubachia, two accessions of S. retropilosa Botsch. and one of 5. linearifolia (W. W. Sm.) O. E. Schulz lay basal to a weakly supported (51%) clade that consisted of three groups, with weak support. Desideria prolifera was basal to core Solms-laubachia, whereas D. incana and P. jafrii were basal to these other species, but with weak support. TRNL-F SEQUENCE DATA Twenty-five accessions were included in the analysis. (Table 1). — The multiple alignment of the 524 Annals of the Missouri Botanical Garden 91 m Solms-laubachia eurycarpa | 78 Lo Solms-laubachia eurycarpa 2 Solms-laubachia angustifolia 5] Solms-laubachia pulcherrima Solms-laubachia zhongdianensis 75 ~ S 63 Solms-laubachia minor Solms-laubachia xerophyta Solms-laubachia grandiflora 63 | Solms-laubachia calcicola 73 a Solms-laubachia sunhangiana Solms-laubachia retropilosa | Solms-laubachia retropilosa 2 Solms-laubachia linearifolia ———4 Desideria prolifera 98 Solms-laubachia platycarpa | 94 Solms-laubachia platycarpa 2 PIE, 98 Solms-laubachia lanata | Solms-laubachia lanata 2 ~ Desideria baiogoinensis 98 Phaeonychium jafrii 98 Desideria incana Desideria linearis Christolea crassifolia Leiospora pamirica Figure 1. Strict consensus tree of three MPTs based on ITS sequence. Numbers above and below branches are BS and jac ‘knife fe percentages (value s lower than 50% nol shown). Volume 95, Number 3 2008 Yue et al. 525 Solms-laubachia from Western China trnL-F region required 42 indels. The data set had 1106 characters. However, a segment of 64 bp (bp 302-365) sequence alignment. Of the remaining 1042 charac- was excluded because of ambiguous ters, 885 were uninformative, 75 were variable but not and 82 parsimony | informative, were potentially parsimony informative. The trnL-F 3.4%-5.0% between the ingroup and outgroup and sequence divergence ranged from 0%-3.3% within the ingroup. Identical sequences were detected between two conspecific accessions of Solms-laubachia platycarpa. The MP analysis yielded 2565 MPTs of 193 steps (CI = 0.84, RI = 0.81, and RC = 0.68; cf. Fig. 2). Twelve Solms-laubachia species, together with the four Desideria species and Phaeonychium jafrii, formed a monophyletic clade, with Christolea crassi- folia Cambess. as sister group with 10096 BS support. Within this, Desideria himalayensis and D. stewartii formed a basal clade with 91% BS support. The remaining species formed less resolved polytomy with only 5896 bootstrap support. The two accessions of S. lanata were grouped with 5. angustifolia J. P. Yue, Al-Shehbaz & H. Sun at 60% BS support, while D. baiogoinensis occupied a solitary position in the Solms-laubachia polytomy. -OMBINED SEQUENCE DATA x The ILD test returned a P value of 0.014, which is lower than the general threshold of 0.05 (Farris et al., 1995). However, Sullivan (1996) argued that the significance threshold of 0.05 may be too conservative for the ILD test. Cunningham (1997) pointed out that combining data improved or did not reduce phyloge- netic accuracy whenever a P value from the ILD test was greater than 0.01. Therefore, we performed an analysis on the combined data set. The combined data set consists of 1664 characters, 146 were variable but not parsimony informative, and 99 were among which 1419 were uninformative, potentially parsimony informative. Phylogenetic anal- ysis of the combined data set yielded two MPTs of 324 steps (CI — 0.81, RI — 0.74, and RC — 0.57), which have a topology nearly identical to the ITS tree, differing only in the position of Desideria baiogoinen- sis (not shown). In the combined tree, D. baiogoinensis clustered with Solms-laubachia platycarpa rather than ITS strict consensus. However, no relationships received con- immediately with anata as in vincing support. With the exception of the basal which received 84% support, the relationships among other ingroups were position of D. linearis, not convincingly resolved in the combined analysis. Both ITS and trnL-F sequence data revealed that Solms-laubachia, Desideria, and Phaeonychium jafrii (Figs. 1, 2). neither Solms-laubachia nor Desideria is monophylet- formed a monophyletic clade However, ic, appearing in more than two positions within the core clade. Therefore, sequences of the ITS and trnL- F results support an expanded Solms-laubachia that includes Desideria and P. jafrii, as suggested by Yue et al. (2006) combined — However, neither the separate nor analysis could resolve their terminal relationship convincingly. The present study shows rather low-level sequence divergences of ITS (from 096—5.196; mean 2.4%) and trnL-F (0%-2.6%; mean 1.33%) among Solms-laubachia and Desideria. Comparisons of the phylogenetic trees generated from the ITS (Fig. 1) and the trnL-F (Fig. 2) showed some discrepancies for the position of a few taxa on either tree. For example, in the ITS tree Desideria linearis was basal to all other ingroup taxa, whereas in the trnl-F tree it supported polytomy. Solms-laubachia lanata formed with D. phylogeny (7796 BS), but switched its sister group relationship to S. angustifolia (60% BS) in the trnL-F phylogeny. These discordances between nuclear and embedded in a weakly baiogoinensis in the ITS chloroplast trees may give some clues for exploring speciation mechanisms within Solms-laubachia. The species with such conflicting relationships between the ITS and trnL-F trees might be the result of hybrid speciation followed by inheritance of chloroplast DNA from one parent and fixation of ITS sequences from the other. However, to test this hypothesis one needs to compare well-resolved phylogenetic relationships generated from both chloroplast and nuclear data sets and sequence characteristics. Due to the low sequence divergence and less resolved relationship in the present study, we could not draw reliable conclusions in the speciation context. To better understand the phylogenetic relationship and speci- — ation mechanism within Solms-laubachia, both addi- tional independent gene phylogenies and the appli- cation of more sensitive molecular markers are needed, and both approaches are in progress. SEED EPIDERMIS Seed epidermis of 12 Solms-laubachia species and Desideria batogoinensis was studied under SE — Figs. 3—5). For each species, we checked three seeds from different individuals. Seeds are rugose in S. platycarpa, S. lanata, S. retropilosa, and D. baiogoi- nensis (Fig. 3A—H) and reticulate in the other species (Figs. 4, 5). Within these two seed types, there is a sculpturing of the substantial variation in the anticlinal cell walls and the fine ornamentation on 526 Annals of the Missouri Botanical Garden 100 Solms-laubachia eurycarpa | 90 Solms-laubachia eurycarpa 2 Solms-laubachia platycarpa | 100 100 Solms-laubachia platycarpa 2 Solms-laubachia pulcherrima Solms-laubachia linearifolia Solms-laubachia zhongdianensis Solms-laubachia minor Solms-laubachia grandiflora Solms-laubachia xerophyta Solms-laubachia angustifolia m - 100 Solms-laubachia lanata 1 ~ 100 Solms-laubachia lanata 2 Solms-laubachia calcicola 97 poe Desideria linearis | 93 MEM Desideria linearis 2 100 Phaeonychium jafrii 73 E Solms-laubachia retropilosa 1 68 ees Solms-laubachia retropilosa 2 Desideria baiogoinensis a p[—— Desideria himalayensis 84 Desideria stewartii Christolea crassifolia Leiospora exscapa — 100 [— — ro Ll Leiospora pamirica Figure 2. Strict consensus tree of 2565 MPTs based on trnL-F sequence. Numbers above and below branches are BS and jackknife percentages (values lower than 50% not shown). the periclinal cell walls. The seed micro-sculpturing linearifolia with simple walls and minutely papillate is useful in distinguishing different species with cell surface (Fig. 5A, B). The seed epidermal similar gross morphology. For example, S. grandiflora micromorphology of D. baiogoinensis (Fig. 3G, TH) J.P. Yue, Al-Shehbaz & H. Sun and S. linearifolia are was similar to that of S. Janata (Fig. 3C, D). with both morphologically similar; however, the seed surface of — showing striped periclinal cell walls. This common S. grandiflora has the compound walls and smooth feature supports the merger of Desideria in Solms- cells (Fig. 5C, D), distinguished from that of S. laubachia. Such a conclusion is consistent with the Volume 95, Number 3 Yue et al. 527 2008 Solms-laubachia from Western China Figure 3. Seed epidermis of Solms-laubachia and Desideria under SEM. —A. B. S. platycarpa (X 250; X 1000). —C, D. S. lanata (X250; X 1000). —E. F. S. retropilosa (X150; X 1000). —G, H. D. baiogoinensis (X150: X 1000). Voucher sources for the taxa are represented by asterisks in Table 1 Annals of the Missouri Botanical Garden y y NS A YT t A o eum LIRA") Figure 4. Seed epidermis of Solms-laubachia under SEM. — A, B. S. xerophyta (X150: X 1000). —C. D. S. angustifolia (X150; X 1000). —E. F. S. pulcherrima (X 150: * 1000). —6. H. S. minor (X150: X 1000). Voucher sources for the taxa are represented by asterisks in Table 1. Volume 95, Number 3 Yue et al. 529 2008 Solms-laubachia from Western China =a 100 IT Semen 10um ‘igure 5. Seede up rmis of Solms-laubachia under SEM. —A, B. 5. d (X250; X1 200). —C, D. S. grandiflora (x250; X 1000). —E, calcicola (X250; X 1000). —G, H. S. eurycarpa (X150; * 1500). —I, J. S. mer (X250; X 1000). Voucher sources 2 e: laxa are aud by asterisks in Table 1. 530 Annals of the Missouri Botanical Garden ITS and triL-F data and previous findings of Yue et al. (2006). TAXONOMIC. CONSIDERATIONS Solms-laubachia has been considered by Al- Shehbaz (2001) and Al-Shehbaz and Yang (2001) to be most closely related to Desideria (12 species: Al- Shehbaz, 2005) and Letospora (six species; Botschan- 1955). Brassicaceae by n the lzev, ME three genera are unique having fruits tardily dehiscent apically, rectangular in cross section, and readily detached at maturity from the pedicel. Furthermore, the fruit valves are adnate apically to the replum, are strongly angled at the margins, completely conceal the replum, and have prominent marginal veins. Al- Shehbaz et al. (2006) placed these three genera in the tribe Euclidieae, a disposition supported by the recent molecular findings of Warwick et al. (2007). \ccording to Al-Shehbaz (2001). Solms-laubachia differs from Desideria by having pinnately (vs. 3- to 9 [to II |-toothed) leaves and linear-oblong (vs. ovate to palmately) veined, entire (vs. apically oblong) anthers. It differs from Letospora by having wingless (vs. winged) seeds, equal (vs. unequal) sepals with the lateral pair non-saccate (vs. saccate), and entire or 2-lobed stigmas, but the lobes are neither decurrent nor connivent (vs. decurrent and connivent). However, molecular data (Yue et al.. 2000; this study) strongly indicate that Desideria is polyphyletic and nested within Solms-laubachia, and Solms-laubachia and Leiospora are independent but (2007) also found Desideria to be polvphyletie in relation & closely related genera. Warwick et a > Solms-laubachia, but this study included only two species each of these two genera. Thus, the above- mentioned differences in leaf. morphology between Solms-laubachia and Desideria are not consistent with the evolutionary relationships reflected by the molecular data. The four molecular markers used in (2000) that Phaeonychium jafrit is embedded in the Solms- this study and in Yue et al. show laubachia—Desideria clade to form one monophyletic group. Therefore, we transfer below P. jafrii and all Desideria sensu Al-Shehbaz (2001. 2005) to Solms- laubachia and recognize them as S5. jafrii (Al- Shehbaz) J. P. Yue. Al-Shehbaz & H. Sun. S. batogoinensis (K. C. Kuan € C. H. An) J. P. Yue. Al-Shehbaz & H. Sun, 5. flabellata (Regel) J. P. Yue, Al-Shehbaz & H. Sun. S. haranensis (Al-Shehbaz) J. P. Yue, Al-Shehbaz € H. Sun, S himalayensis (Cambess.) J. P. Yue, Al-Shehbaz € H. Sun, S. P. Yue, Al-Shehbaz € H. Sun. 5. linearis (N. Busch) J. P. Yue, Al-Shehbaz & H. Sun. S. mieheorum (Al-Shehbaz) J. P. Yue. Al-Shehbaz & H. incana (Ovez.) J. on Sun, 5. mirabilis (Pamp.) J. P. Yue, Al-Shehbaz & H. Sun, S. nepalensis (H. Hara) J. P. Yue, Al-Shehbaz & H. Sun, 5. prolifera (Maxim.) J. P. Yue, Al-Shehbaz & H. Sun, and S. stewartii (V. Anderson) J. P. Yue. Al- Shehbaz & H. Sun. An expanded generic description of Solms-laubachia and a synopsis and key to all species are given below. Solms-laubachia Muschl. in Diels, Notes Roy. Bot. Gard. Edinburgh 5: 205. 1912. Type: Solms- laubachia pulcherrima Muschl. Desideria Pamp.. Boll. Soe. Bot. Hal. 1926: 111. 1926. TYPE &: Desideria mirabilis Pamp. Ermaniopsis H. Hara, J. Jap. Bot. 49: 198. 1974. TYPE: ¿rmantopsis pumila M. Hara. Oreoblastus Suslova, Bot. Zhurn. (Moscow & Leningrad) 57: 618. 1972. TYPE: Oreoblastus flabellatus (Regel) Suslova. Herbs with branched caudex covered with petioles of previous perennial, | sometimes pulvinate, years: trichomes absent or simple, rarely short- stalked, 2-rayed: stems absent or present and simple, leafy or leafless. Basal leaves petiolate, rosulate, simple, entire or 3- to Oto lI)-toothed, pinnately « palmately veined; cauline leaves absent or similar to basal ones and short petiolate to subsessile, not auriculate. Flowers solitary on long pedicels originat- ing from center of rosette, or in (3 toJ6- to 30-flowered, bracteate or ebracteate corymbose racemes elongating or not in fruit. Sepals oblong to ovate, free or united, persistent or deciduous, erect, equal, base of inner pair not saccate, margin membranous or not; petals purple, blue, pink, or rarely white, suborbicular, obovate, to spatulate, apex obtuse to emarginate; claw sepals; stamens 6; subequaling or longer than filaments free. dilated or not at base: anthers oblong-linear to ovate, nol apiculate al apex; nectar glands 2 and lateral, or confluent and subtending filaments: median nectaries absent o 30(to 70) per dehiscent silique or silicle, linear, oblong, bases of all present; ovules 6 to ovary. Fruit ovale, o anceolate, latiseptate, sessile, readily detached at maturity from pedicel, rectangular in cross. section; valves papery, reticulate veined, with a prominent midvein and marginal veins, glabrous or pubescent, smooth, adnate with replum at fruit apex, margin angled: replum rounded, concealed Is connate valve margins; septum complete or rarely perforated. o reduced to a rim, membranous, translucent, rarely absent: style obsolete, rarely to 1 mm long; stigma capitate, entire or 2-lobed, lobes not decurrent. Seeds unisertate or biseriate, wingless, broadly ovate to suborbicular or oblong, flattened: seed coat reticulate, rugose, or papillate, not mucilaginous when wetted: col ledons accumbent. Volume 95, Number 3 2008 ue et al. 531 Solms-laubachia from Western China Solms-laubachia are distributed in Bhutan, China (Gansu, Qinghai, Sichuan, Xinjiang Xizang [Tibet], Yunnan), India (Himachal Pradesh, Kashmir, Punjab, Sikkim), Kyrgyzstan, Nepal, Paki- Distribution and habitat. The 26 species of stan, and Tajikistan. They Afghanistan, — (3200—)4000—6200 m and are found at altitudes of grow primarily on scree slopes and in rock crevices in the alpine areas of the Himalayan, Karakorum, Pamir, and Hengduan moun- lain ranges. KEY TO THE SPECIES OF SOLMS-LAUBACHIA la. b. Flowers in racemes; stems mostly leafy. 2a. Leaves enti irt l- or Ja. Stems 2-leaved; leaf trichomes simple only; petals blue to purple, 10-17 mm long; fruit gee s ovate; 8-17 mm widé- 12... gieds cte ond teed a dee e des l an 3b. Stems leafless; leaf trichomes a mixture of simple and branched; petals white to pink, 6.5—10 mm long; fruit "hg A A ES d 10. S. jafrit 2b. Leaves dentate Aa. sepals ie D until or after fruit dehiscence; septum reduced to a rim. 5a. s 11-13 X 5-6 mm; raceme 2- to 4-flowered, appearing solitary; Nepa 5b. Pe al 5-8 x 1.5-3 mm; raceme with 5 or more flowers 4b. Sepals free, caducous or perforate >d apically. y; Nepal ........ 7. S. nepalensis : China, Kashmir, Tajik i o S. mirabilis rarely persisting until about ful maturity; septum mum leto Or rare £5 Racemes bracteate throughout. Ta. Stem and pedicel trichomes forked 22s 23. S. stewartii 7b. Stem and pedicel trichomes exclusively simple or absent. 8a. Fruit lanceolate to linear-lanceolate, (3—)4—6 mm wide; petals (6—)6.5-8 e te mm seeds biseriate, (1.52)1.8-2(-2.3) X 1-1.4 mm ......... llus. laa 8b. Fruit linear, (0.8-)1-1.7(-2) mm wide; petals 4—5(—5.5) X 1.5-2.5 ; see uniseriate, 0.8—1.1 X 0.5-0.8 mm : 6b. Racemes ebracteate. 9a. Filaments flattened, subapically toothed; petals 6.5-8 mm long Ob. Filaments terete, toothless; petals 10-18 mm long. 10a. Plants canescent; leaf trichomes almost exclusively branched 10b. Plants not canescent; leaf trichomes exclusively sim} le. a. Style 5—7 mm long; stigma entire; petals blue; ne 10 to 16 per ovary IIb. Style obsolete; stigma prominently 2-lobed; petals purple; MALY gute CPC TEM . S. flabellata Flowers solitary from a basal rosette; scapes leafless. 2a. Le eaves dentate. O 9. S. incana 14. S. mieheorum : ovules 14 to 24 per Fruit ovate to broadly lanceolate, 6-9 mm wide, prominently reticulate veined Fruit linear to linear-lanceolate, 2-5 mm wide, obscurely veined. 14a. Leaf trichomes forked and simple; pelals 6-7 mm long ee 2. S. baiogoinensis 13b. replum retrorsely ciliate; valves glabrous; sepals 3—4 mm long; Pe be a USUS ORS ER ae ius umila (Kurz) F. Dvorak valves pubes ent; sepals 6—7 mm N 14b. Leaf trichomes exclusively simple; replum spreading ciliate; ong; petals 11—14 mm long 12b. Leaves entire. — — 5. prolifera 5a. Leaves gray, densely lanate: seeds ru l6a. Leaf eae anceolate- to oblanc colt linear, rarely oblanceolate, retrorsely lanate, 1.5 37) mm wide; fruit valves obscurely veined ........0.0000 0c eee ee . S. retropilosa 16b. Leaf “ee broadly spatulate, bons to obovate, spreading or antrorsely lanate, (5 7 oi mm wide; fruit valves prominently veined 2.2.2.2. . S. lanata g 17a. Sepals 2-3 mm long; fruits 2.8-3.2 mm wide... llle 24. S. sunhangiana 17b. Sepals (4.5— 11) mm lona: fruits (4—)5—18 mm wide 18a. ding of previous seasons papery; longest leaf blades rarely to 1 em long ........ tie Coot Pd Oi bee gd gris Pp i8 Vua oa d bed deed 6. 5. zhongdianensis 18b. Pe o of previous seasons remaining thickened; longest leaf blades (1.5—)2—4.5(—5.5) cm long 19a. Leaf blade 0.3-1(—1.5) mm wide, grooved a ially. 20a. Petioles not ciliate; iA blades (2—)2. (7.7) em long; fruits glabrous, narrowly oblong to linear ......... lll llle eere S. angustifolia 20b. Petioles ciliate; leaf blades 0.7—2(-2.5) em long; fruits pubescent, Jane eolate to linear-lanceolate ............... 2 5. S. xerophyta (W. W. Sm.) H. F. Comber 19b. Leaf blade (1.5-)2-16(-23) mm wide, flattened adaxially. 2la. Leaf blade (7—)10-16(-23) mm wide; petiole strongly o kened, + corky; sepals WC EN eurycarpa (Maxim.) Botsch. 21b. Leaf blade (1.5)2-5(-7) mm wide; petiole FEN ys he um sepals free etimes united in S. linearifolia) E etioles of previous seasons papery "TETTE 12. 5. linearifolia 532 Annals of the Missouri Botanical Garden 22b. Petioles of previous seasons thickened. 23a. bn ls pink Leaf blade to l em long, glabrous, not ciliate 22... iis. ae 15. 5. minor Hand.-Mazz. 24b. Leaf blades (1.2—)1.8—4.5(—5.4) cm long, glabrous or Sp ursely ubescent, ciliate 22.2222 6. S. grandiflora pt 23b. a tals turquoise blue. 2! Seeds orbicular to broadly ovate; leaf trichomes erisped: perm loosely branched; plants of scree slopes ........ onn herrima 25b. Seeds oblong; leaf trichomes if present then straight: iudex compactly branched; plants of rock crevices ...... 3.5. S, calcicola l. Solms-laubachia angustifolia J. P. Yue, Al- — minor—S. angustifolia clade and the S. pulcherrima— ~ Shehbaz & H. Sun, sp. nov. TYPE: China. — S. zhongdianensis clade. Cytological studies (Yue et al., Sichuan: Yading Xian, Xianuoduoji, 28 23'N, — 2004) on S. angustifolia (as S. xerophyta therein) and S. 100 23'E, 4530 m, 10 Sep. 2002, J. P. Yue 0250 xerophyta also found distinct differences in their (holotype, KUN!; isotype, MO!). chromosomal karyotypes. Furthermore, all previous Herba perennis pulvinata, 2-6 cm alta. Folia basalia ide i uen pla ia lan e anen; rosulata, linearia, (2-)2.5-3.5(-7.7) em X 03-1(-1.3 1981; Tan et al., 1999; Al-Shehbaz & Yang. 2001) sparse pubescentia, integra: petiolis 04-1. 5cm longis, incrassatis; folia caulina nulla. Pedicelli fructiferi solitarii, occupy different provinces. = = = = represent S. angustifolia, and therefore the two taxa 1-1.7 em longi. Flores ignotes. Fructus lineares vel anguste oblongi, 1.6—4.2(—4.8) em X aan —8 mm, glabri. Semina Paratypes. CHINA. m Daocheng Xian, J. F. € late ovata vel suborbiculata, 2.1—3.4 X 1.2-2.8 mm. i 16419 (GH), J. F. C. Rock 16857 (MO, PE). T. T. Yü 3043 (KUN, PE) ndn Xizang Team 5818 (PE). Herbs perennial, pulvinate, 2-6 em tall; caudex covered with petioles of previous vears. Basal leaves 2. Solms-laubachia baiogoinensis (K. C. Kuan & rosulate; blade linear, (2-)2.5—5.5(-7.7) em X 0.3- C. H. An) J. P. Yue, Al-Shehbaz & H. Sun, comb. Basionym: Christolea baiogoinensis K. C. => 1.3) mm, not ciliate, grooved adaxially, sparsely nov. petioles Kuan € C. H. An in C. Y. Wu, Fl. Xizangica 2: 388. 1985. Desideria baiogoinensis (K. C. Kuan & C. H. An) Al-Shehbaz, Ann. Missouri Bot. Gard. 87: 561. 2000. TYPE: China. Xizang: — pubescent, base attenuate, margin entire; 0.4—1.5 em, thickened, not ciliate; cauline leaves absent. Fruiting pedicel solitary from basal rosette, 1— 1.7 em. Flowers not seen; ovules 14 to 22 per ovary. Fruit linear to narrowly oblong, 1.6—4.2(—4.8) em X Baiogoin, 5100 m, 18 June 1976, K. Y. Lang (4—)4.5—8 mm; valves glabrous, obscurely reticulate 9460 (holotype, PEL: isotype, PE). veined; septum complete; style (1-)1.2-1.6 mm; Distribution. China (Qinghai, Xizang) (see Al- stigma entire to slightly 2-lobed. Seeds biseriate, broadly ovate to suborbicular, 2.1—3.4 X 1.2-2.8 mm. — Shehbaz, 2000). Chromosome number: 2n = 14 (as Solms-laubachia xerophyta in. Yue et al., 2004). 3. Solms-laubachia calcicola J. P. Yue, Al-Sheh- "T "TT o baz € H. Sun, sp. nov. TYPE: China. Xizang: "oi eda Narrowly endemic to Sichuan Prov- Riwoqe Xian, Machala, 31°24’N, 96°40’E, ` 4770 m, 10 Aug. 2004, D. E. Boufford, J. H. IUCN Red List category. Critically Endangered Chen, S. L. Kelley, J. Li, R. H. Ree, H. Sun, J. P. (CR) (IUCN, 2001). Yue & Y. H. Zhang 31975 (holotype, KUNI; : l , TET isotypes, Al, MOD). Discussion. Solms-laubachia angustifolia — is one closest morphologically to S. xerophyta and both Herba perennis pulvinata, 2-5 cm alta. Folia basalia have narrow leaves grooved adaxially and with rosulata, linearia vel lanceolata, (1.8—)2.3—4. 1(—4.4) « B . . re . ( thickened petioles. However, it differs by having l. 2- 2.5 mm, glabra vel sparse pubescentia, integra; petiolis m : : a ao ae 6.5—8.9(-9.2) mm longis, incrassatis; folia caulina non-ciliate petioles, leaf blades (2-)2.5-5.5(- 7.7) DIS ) Ap En es . MEN Pedicelli fructiferi aua 3—3.5(-5) em longi. Sepala em long, and glabrous, narrowly oblong to linear libera, oblonga, 0.2-6.9 2.6-3.2 mm; petala turcosa, fruits. By contrast, S. xerophyta has ciliate petioles, — obovata, 1.5-1.8 X "ea em; ovula 18 ad 28. Fructus 5 3.9 Y pl hu leaf blades 0.7-2(-2.5) em long, and pubescent, anguste ovati vel lanceolati, 1.1-2.3(-2.5) « lanceolate to linear-lanceolate fruits. Our previous "d 1: ps el ps sparse pubescentes. Semina oblonga, study (Yue et al., 2006) revealed that 5. angustifolia abere EUR. is sister to 5. minor, whereas 5. xerophyta falls in a Herbs perennial, pulvinate, 2-5 cm tall; caudex separate branch of a polytomy that included the S. compactly branched, covered with petioles of previous Volume 95, Number 3 2008 Yue et al. Solms-laubachia from Western China 533 — years. Basal leaves rosulate; blade linear to lanceo- late, (1.8—)2.3—4.1(—4.4) cm X 1.9-2.5 mm, glabrous or sparsely pubescent with straight trichomes, not ciliate, base attenuate, margin entire; petioles (5.7— 6.5—8.9( Fruiting pedicels solitary from basal rosette, 3—3.5(—5) oblong, 6.2-6.9 X 2.6-3.2 mm: petals turquoise blue, obovate, 1.5-1.8 X 0.6-1 cm: median filaments 3.8—4.5 mm; 1-1.8 mm: ovules 18 to 28 per ovary. Fruit narrowly ovate to 1.1-2.3(-2.5) cm X 5.1-8.9(-9.2) mm, glabrous or sparsely pubescent; valves prominently —9.2) mm, thickened; cauline leaves Lo cm. Sepals free, anthers lanceolate, reticulate veined; septum complete; style 0.5-1.5 mm: stigma slightly 2-lobed. Seeds oblong, uniseriate to biseriate, 2-2.5 X (1.2)1.3-1.5(-1.7) mm. Distribution. Narrowly endemic to Xizang Prov- ince, China. IUCN Red List category. (CR) (IUCN, 2001). Critically Endangered Discussion. Solms-laubachia calcicola resembles S. pulcherrima in being the only species of the genus with turquoise petals and in having thick petioles and similar flower size, but it differs by having oblong versus orbicular to broadly ovate seeds with coarser seed sculpture (Fig. 5E, F), as well as a thick and compactly branched (vs. slender and loosely branched) caudex. It grows in rock crevices instead of scree slopes, to which 5. pulcherrima is restricted. Furthermore, molecular data (both ITS and trnL-F) clearly show that S. calcicola and S. pulcherrima fall into two separate clades in both analyses (Figs. 1, 2). The occurrence of turquoise petals in these two somewhat remotely related species would appear that this flower color evolved independently in them. Paratype. CHINA. zang: 31724'N, 9640' E, 4770 m, p Y. Tu px Y. a a an 4. Solms-laubachia eurycarpa (Maxim.) Botsch., Bot. Mater. Gerb. Bot. Inst. Komarova Akad. Nauk S.S.S.R. 17: 169. 1955. Basionym: Parrya eurycarpa Maxim., Fl. Tangut. 1: 56. 1889. TYPE: China. Tibet: near Jagem-Gol, 20 July 1884, N. M. Przewalski s.n. (holotype, LE!; isotype, PE!). Solms-laubachia pulcherrima Muschl. var. latifolia O. E. Schulz, Notizbl. Bot. Gart. Berlin-Dahlem 11: 229. 19; Solms- laubachia latifolia (O. E. Acta Phytotax. Sin. Peoia A Kenlecling, 4475 m , June-Aug. 1928, F. C. Rock 16870 (lectotype, designated by Al- Shéhbaz & Yang, 2001: 375, B!; isotypes, E!, US!, W!). Solms-laubachia aia Se Y. Z. Lan € T. Y. Cheo, Acta hytotax. Sin. 19: 477. ne TYPE: China. Sichuan Zoige, 4300 m, 4 Aug. 1960, Sichuan Drug Plant Expedition 20279 (holotype, NAS [as HJ]!). Solms- laubachia eurycarpa var. brevistipes Y. Z. Lan & T. Y. Acta Phytotax. Sin. 19: 476. 1981. TY E. China. Oiga Nanggen, 4400 m, 9 July 1965, ^ Yang HNWP [as HQ]!; isotype, a Solms- ula eurycarpa var. a R. F. Huang in S. Lic Es PE: China. 21 (holotype, AB ]: 510. Qinghai: Hen Xian, Waisxan Need 9 July 1972, 4100—4200 m, P. n Kuo 9908 E HNWP [as NWBI!). Solms-laubachia o Al-Shehbaz & G. Yang, Harvard ap. Bot. 5: 380. 2001. TYPE: China. Yunnan: De "qen, Beima Shan. A side of rd., limestone scree, 28°23'N 9901'E, 4700 m, 30 June 1994, Alpine Garden Society Expedition to ca ACE 855 (holotype, K!) m Distribution. China (Gansu, Qinghai, Sichuan, Xizang, Yunnan) (see Al-Shehbaz & Yang, 2001). 5. Solms-laubachia flabellata (Regel) J. P. Yue, Al-Shehbaz & H. Sun, Basionym: Parrya flabellata Regel, Bull. Soc. Imp. Natur- alistes Moscou 43: 261. 1870. Ermania flabel- (Regel) O. E. Schulz, Bot. Jahrb. Syst. 66: 1933. Christolea ra (Regel) N. Busch , Fl. URSS 8: 330. 1939. Oreoblastus flabellatus (Regel) lora: e. Zhurn. (Moscow & Leningrad) 57: 651. 1972. Desideria flabellata (Regel) Al-Shehbaz, Ann. 87: 558. 2000. TYPE: [Kyrgyzstan]. Tian Shan, Dschaman-Daban, s.d., comb. nov. in e marov Missouri Bot. Gard. — southern Sewerzow s.n. (holotype, LE! Christolea pinnatifida R. F. Huang, Acta da Sin. 35: e 1997. TYPE: China. Qinghai: M . Anyemaqen . 4800 m, ae = 1981, R. F. Huang CG-81-154 eee HNWP Distribution. Afghanistan, China (Qinghai, Xin- jiang), Kyrgyzstan, Tajikistan (see Al-Shehbaz, 2000). 6. Solms-laubachia grandiflora J. P. Yue, Al- Shehbaz & H. sp. TYPE: China. Sichuan: Xiancheng Xian, Rizhaosheng Shan, 29^6'N, 99°41'32"E, 4650 m, 15 July 2004, D. E. Boufford, J. H. Chen, S. L. Kelley, J. Li, R. H. Ree, H. Sun, J. P. Yue & Y. H. Zhang 30727 (holotype, KUN!; isotypes, A!, MO!). Sun, nov. 3-7 cm alta. Folia ep ria vel lineari-lanceolata, (1.2-)1.84.5 3(2.9) mm, glabra vel sparse pubescentia, integra; petiolis (5.2-)6.8-8.2(-10.2) mm e incrassatis; folia caulina nulla. Pedicelli fructiferi solitarii, 1.6-2.5(-3.2) cm longi. Sepala libera vel connata, oblonga, 6.1-7.7 X 2- 2.9 mm; petala purpurea, obovata, 1.6-2.5 X (0.6—)0.8— 0.9(-1.2) cm; ovula 14 ad 24. Fructus oblongi vel oblongo- 1-3.4(—4.8) cm X 4-11 mm, glabri vel sparse pubescentes. Semina late ovata vel suborbiculata, 2.4—3.6 X (1.6—)1.8-2.8(—3) mm. Herba perennis pulvinata, rosulala, linea em X 1.5-2. Ei lineares, , 3-1 em tall covered with petioles of previous years. Basal Herbs perennial, pulvinate ; caudex uo eaves Annals of the Missouri Botanical Garden rosulate; blade linear to linear-lanceolate, (1.2—)1.8— 4.: cm X 1.5-2.3(-2.9) sparsely pubescent, ciliate, base attenuate, margin entire; petioles (5.2-)6.8-8.2(-10.2) mm, thickened; cauline leaves absent. Fruiting pedicels solitary from mm, gl abrous or basal rosette, 1.6-2.5(-3.2) em. Sepals free or united, oblong, 6.1—7.7 X 2-2.9 mm; petals purple, obovate, 1.6-2.5 X (0.6—)0.8—0.9(—1.2) cm; median filaments 4.5-5.6 mm; anthers 1.3-2 mm; ovules 14 to 24 per 1-3.4(4.8) cm X 4-11 mm, glabrous or sparsely pubescent; valves ovary. Fruit oblong to oblong-linear, obscurely reticulate veined; septum. complete; style 1.1-2.1 mm; stigma slightly 2-lobed. Seeds biseriate, broadly ovate to suborbicular, 2.4—3.6 X (1.6-)1.8- 2.8(—3) mm. Distribution. Narrowly endemic to Xiancheng Xian in Sichuan Province, China. IUCN Red List category. (CR) (IUCN, 2001). Critically Endangered Discussion. | Solms-laubachia grandiflora most closely resembles S. linearifolia, especially in leaf morphology and flower color, but it differs by having thickened instead of thinner, papery petioles, narrow- er leaves (1.5-2.3[-2.9] vs. 1.7—4.1[|-5| mm wide), smaller seeds (2.4—3.6 X [1.6-]1.8-2.8|-3] mm vs. 3.54. X 2.5-3.5 mm), and shorter (1—3.4[—4.8] cm vs. [3-]4.5—6.5[-8] cm), less pubescent fruits. Further- more, they differ in grandiflora (Fig. 5C, D) smooth cells, whereas 5. linearifolia (Vig. 5A, B) has seed morphology, and S. has compound walls and simple walls and a minutely papillate cell surface. The molecular data from ITS (Fig. 1) show that 5. grandiflora is clustered with S. calcicola and S sunhangania, whereas S. linearifolia falls in a separate, more basal clade. Paratype. CHINA. Sichuan: Xiancheng Xian, Ri- zhaosheng Shan, 29 6'N, 99 41'32"E, 4650 m, 18 Oct. 2005, J. P. Yue 5120 (KUN, MO). 7. Solms-laubachia haranensis (Al-Shehbaz) J. P. Yue, Al-Shehbaz & H. Sun, comb. nov. nym: Desideria haranensis Al-Shehbaz, Missouri Bot. Gard. 87: 559. 2000, replacement name for Ermaniopsis pumila H. Hara, J. Jap. Bot. 49: 200. 1974, non Solms-laubachia pumila (Kurz) F. Purkyne Brne, Biol. Basio- Dvorák, Folia Prirodoved. Fak. Univ. 13(4): 24. 1972. TYPE: Nepal. ca. 5 mi. SW of Saldanggaon, very loose 19,500 ft. [5944 m], 26 June 1952, N. Polunin, W. R. Sykes & L. H. J. Williams 24 (holotype, BM!; isotypes, A!, BM!, EN. Scree, Distribution. Nepal. 8. Solms-laubachia himalayensis (Cambess.) J. P. Yue, Al-Shehbaz & H. Sun, comb. nov. Basio- Cheiranthus himalayensis Cambess. in , Voy. 1844. himalayensis (Cambess.) O. E. Schulz, Notizbl. Bot. Gart. Berlin-Dahlem 9: 1080. 1927. Oreo- blastus himalayensis (Cambess.) Suslova, Bot. Zhurn. (Moscow & Leningrad) 57: 652. 1972. Desideria himalayensis (Cambess.) Al-Shehbaz. Ann. Missouri Bot. Gard. 87: 555. 2000. TYPE: [W Xizang], “in declivitate orientali jugi vulgo Kioubrung-ghauti in Tartariá sinensi,” V. Jac- quemont 1782 (holotype, P!: isotypes, K!, P!). nde 4 Ermania Distribution. China (Qinghai, Xizang), India (Hi- Punjab), Nepal (see A machal Pradesh, Kashmir, Shehbaz, 2001). 9. Solms-laubachia ineana (Ovez.) J. P. Yue, Al- Shehbaz € H. Sun, Cleator incana Ovez., & 51. comb. nov. Basionym: Sovetsk. Bot. 194.1(1 Ermania (Ovez.) Botsch., Bot. Mater. Gerb. Bot. Inst. Komarova Akad. Nauk S.S.S.R. 17: 164. 1955. Oreoblastus incanus (Ovez.) Suslova, Bot. Zhurn. (Moscow & incana Leningrad) 57: 652. 1972. Desideria incana (Ovez.) Al-Shehbaz, Ann. Missouri Bot. Gard. 87: 558. 2000. TYPE: Tajikistan. Darvaz: Mt Masar, glacier Abdul Gassan, 11,000-12,000 ft. [3353-3658 m], 23 July 1899, V. 1. Lipsky 1936 (holotype. LE!). Distribution. ‘Tajikistan. 10. Solms-laubachia jafrii (Al-Shehbaz) J. P. Yue, Al-Shehbaz € H. Sun, Phaeonychium jara Al-Shehbaz, Nordic J. Bot. 20: 160. 2000. TYPE: China. Xizang [Tibet]: hills E of Lhasa, 14,000 ft. [4207 m], on rock faces, 14 June 1942, F. Ludlow & G. Sherriff 8714 (holotype, BM!). comb. nov. Basionym: Distribution. Bhutan, China (Xizang), Nepal (see Al-Shehbaz, 2000) 11. Solms-laubachia lanata Botsch., Bot. Mater. Gerb. Bot. Inst. Komarova Akad. Nauk 5.S.S.R. 17: 171. 1955. TYPE: China. Tibet [Xizang]: near Yerpa Monastery, 14,000 ft. [4267 m]. Aug. 1821, R. S. Kennedy 9 (holotype, K!). Distribution. China (Xizang). 12. Solms-laubachia linearifolia (W. W. Sm.) O. E. Schulz, Notizbl. Bot. Gart. Berlin-Dahlem 9: 477. 1926. Basionym: Parrya linearifolia W. W. Sm., Notes Roy. Bot. Gard. Edinburgh 11: 219. — Volume 95, Number 3 2008 Yue et al. 535 Solms-laubachia trom Western China 1919, non P. linearifolia Pavlov, Vestn. Akad. Nauk Kazakhsk. S.S.R. 1: 29. 1949. TYPE: China. Yunnan: Beima Shan, Mekong—Yangtze 28 20'N, 14,500 ft. [4420 m]. 1914, G. Forrest 13,235 (lectotype, designated here, E!; isotype, E!). divide, Aug. Solms- laubachia linearifolia var. leiocarpa O. E. Schulz, Bot. Gart. Berlin-Dahlem 9: 477. 1926. TY PE: China. NW Yunnan: Mt. Peimashan, Mekong-Yangtze divide betw. Atuntze & Pungtzera, June 1923, J. F. C Rock 9304 (holotype, B!; isotypes, E!, GH! US). Z Distribution. China (Sichuan, (see Al-Shehbaz & Yang, 2001). Xizang. Yunnan Xizang, Y 13. Solms-laubachia linearis (N. Busch) J. P. Yue, Al-Shehbaz € H. Sun, Christolea linearis N. Busch in Komarov, Fl. URSS 8: 636. 1939. Ermania linearis (N. Busch) Botsch., Bot. Mater. Gerb. Bot. Inst. Komarova Akad. Nauk S.S.S.R. 17: 166. 1955. Oreoblastus linearis (N. Busch) Suslova, Bot. Zhurn. (Moscow & Leningrad) 57: 652. 1972. Desideria linearis (N. Busch) Al-Shehbaz, Ann. Missouri Bot. Gard. 87: 556. 2000. TYPE: Tajikistan. Pamir: Schug- nan, Abchary, 2 Aug. 1904, B. Fedtschenko s.n. LEN. comb. nov. Basionym: (holotype, Ermania cel O. E. Schulz, Repert. Sp. Nov. Regni Veg 31: 333. 1933. Christolea parkeri Pa E. Schulz) Jafri, Notes Roy. Bot. Gard. Edinburgh 22: 52. 1955. Oreoblastus parkeri (O. E. Sc 'hulz) Suslova, Bot. Zhurn. (Moscow & Leningrad) 57: 653. 1972. TYPE: [India.] Kashmir: ag. l ms rwas, 13, 000 ft. [3962 m], 11 Aug. 1928, R. R. Stewart 9874A (holotype, B! SU e che Dar & Naqshi, J. Bombay Nal. Hist. 87: 274. 1990. TYPE: [India.] Kashmir: arit cds (Sind Valley), 3900 m, 20 Aug. 19 ». H. Dar 7786 (holotype, KASH not seen). Ermania kachraoi Dar & Naqshi, J. ard Nat. Hist. Soc. 277. 1990. TYPE: [India.]| Kashmir: Baltal, Sonamarg (Sind Valley), 3200 m, 2 Sep. 1982, G. H. Dar 3954 (holotype. KASH not seen; isotypes, KASH not seen, MO!). Distribution. China (Xinjiang, Xizang), India (Kashmir), Nepal, Tajikistan (see Al-Shehbaz € Yang, 2001). 14. Solms-laubachia mieheorum (Al-Shehbaz) J. P. Yue, Al-Shehbaz € H. Sun, comb. nov. Basionym: Desideria mieheorum Al-Shehbaz, : 1.2005. TYPE: China. Tibet [Xizang]: N of Sangsang, 20 41'N, 86 43'E, pioneer on frost debris, 5480-5600 m, 15 Sep. 2003, G. € S. Miehe 03-112-21 (holotype, MO!). Novon 15 Chang La, Distribution. China (Xizang). 15. Solms-laubachia minor Hand.-Mazz.. Akad. Wiss. Wien, Math.-Naturwiss. K 246. 1922. TYPE: China. Sichuan: Mt. Holoscha, 27°48'N, betw. Yenyuen & Kwapi, 4325 m, 18 May 1914, H. F. Handel-Mazzetti 2318 (holo- WU; isotypes, E!, P!, W!). Anz. 59 — type, Distribution. China (Sichuan). 16. Solms-laubachia mirabilis (Pamp.) J. P. Yue, Al-Shehbaz € H. Desideria mirabilis Pamp., Boll. Soc. Bot. Ital. 1926: 111. 1926. Christolea mirabilis (Pamp.) Jafri, Fl. W. Pakistan 55: 160. [| Pakistan. Kashmir:] Karakorum, above Caracash Valley, Chisil Gilgha Pass, 5360 m, 28 June 1914, G. Dainelli £ O. Marinelli 2 (lectotype, designated here, Fl not seen, FI photo at HB!). Sun, comb. nov. Basionym: Christolea a Jafri, Notes Roy. Bot. Gard. Edinburgh 22: 5. TYPE: [Pakistan.] Kashmir: Shaksgam Valley, m, Jl 1926, R. C. ae f olope K). Christo karo Ronumenps Y H. C. H. An, Acta . 1994. m a China. Xinjiang: Pishan [Cuma], Sher ]nxtanwan, 5250 m, 25 July 1989, Rune Kunlun Expedition 5100 (holotype, HNWP [as I]!). ylolax. Sin. Desideria pamirica Suslova, Novosti Sist. Vyssh. Rast. 10: W 163. 1973. Christolea suslovaeana Jafri, . Wesl Pakist an 55: 158. 1973, replacement names for D. pamirica non Christolea pamirica end. Mém. Acad. Imp. Sci. Saint Pé tersbourg, ser. 8, 4: 89. 1896. TYPE: Tajikistan. Pamir: asks. near Zor, 4900 m, 10 Aug. 1970, T. slo: s.n. E. LE”). Distribution. China (Xinjiang), Pakistan (Kash- mir), Tajikistan (see Al-Shehbaz, 2001). 17. Solms-laubachia nepalensis (H. Hara) J. P. Yue, Al-Shehbaz € H. Sun, comb. nov. Basio- nym: Desideria nepalensis H. Hara, J. Jap. Bot. 50: 264. 1975. TYPE: Nepal. Barum Valley, 17.700 ft. [5394 m], 26 May 1954, L. W. Swan 71-72 (holotype, BM!). p Distribution. Nepal. 18. Solms-laubachia platycarpa (Hook. f. & Thom- son) Botsch., Bot. Mater. Gerb. Bot. Inst. Komarova Akad. Nauk S.S.S.R. 17: 171. 1955 Parrya Ms Hook. f. & nn J. Proc. Linn. Soc., Bot. 5: 136. 1861, non P. Mid . Basionym: Rydb., Bull. ies Bot. Club 39: 326 TYPE: [India.] Sikkim: alpine region, on ft. [5182 m]. s.d., J. D. Hooker s.n. (holotype, K!). Re red. os Bull. Misc. Inform. Kew 1927: 247. . TY bet ien along C Vh ae torrent, ca. E^ mi. a . Everest, 17,000 ft. [5182 m], 8 June 1922, F. F ee 41 (holotype, K!). 536 Annals of the Missouri Botanical Garden Solms-laubachia orbiculata Y. Z. Lan « Cheo, Acta Phytotax. Sin. 19: 473. ia PYPE: bins: Xizang: Cona, 5020 m, 20 July 1975, C. Y. Wu € S. K. Chen 75-1129 (holotype, HNWP [as DS isotype, KUNI). Distribution. kim). Bhutan, China (Xizang), India (Sik- 19. Solms-laubachia prolifera (Maxim.) J. P. Yue, Al-Shehbaz € H. Sun, Parrya e. Maxim., Fl. Basionym: Tangut. 56. Ermania prolifera (Maxim.) O. E. Schulz, Bur Jahrb. Syst. 66: 98. 1933. Christolea prolifera (Maxim.) Ovez., Sovetsk. Bot. 1941(1 € 2): 151. 1941. C. prolifera (Maxim.) Jafri, Notes Roy. Bot. comb. nov. =. eS Gard. Edinburgh 22: 53. 1955. Desideria pro- lifera (Maxim.) Al-Shehbaz, Ann. Missouri Bot. Gard. 87. 559. 2000. TYPE: China. Tibet [Xizang|: Kon-chun-ua, 14,500 ft., 3 July 1884, N. M. Przewalski s.n. (holotype, LE!; isotypes, K!, P!, PEN. Distribution. Shehbaz, 2000). China (Qinghai, Xizang) (see Al- 20. Solms-laubachia pulcherrima Muschl. in Diels, Notes Roy. Bot. Gard. Edinburgh 5: 206. 1912. TYPE: China. NW Yunnan: E flank of Lichiang Range, 27°20'N, 12,000 ft. [3658 m]. May 1906, G. Forrest 2164 (holotype, Bl: isotypes, BM!, E!, P!) Pegaeophyton sinense (Hemsl.) Hayek & Hand.-Mazz. var. stenophyllum E. Sc hulz, Notizbl. Bot. Gart. Berlin- Dahlem 9: 477. 1926. T hina. Yunnan: Yangtze watershed, Prefect. Distr. of Likiang, E ae of L ikiang , 0300 m, 11 Aug. 1922, J. F . Dis isotypes, E!, GHI, P!, US " pou AEE pulcherrima f. angustifolia O. E. Schulz, izbl. Bot. Gart. Berlin-Dahlem 9: 477. 1926. TYPE: ;. Rock 5719 (hololyy Je, ^ bus Yunnan: Yangtze watershed, Likiang snow range, 18 000 ft. [3962 m], 30 May-6 b 1922, J. F. ock p Gas BL isotypes, GH!, P! US!, ES laubac hia a f. atrichophylla Hand.-Mazz., Akad. Wiss. Wien, Math.-Naturwiss. Kl. 62: 24 1926. TY PE: Sichuan: Mt. Gonsc eee Muli monastery, Yungning toward Dschungdien, 4500 m, H. F. Handel- Maa 7503 (holotype, WU!; isotype, "Wn. Parrya ciliaris Bureau & Franch., J. Bot. (Morot) 5: 20. 1891. Solms-laubachia ciliaris (Bureau & Franch.) Botsch., Bot. Mater. Gerb. Bot. Nauk SSSR 17 Tibet du M. Bonvalot & China. Komarova Akad. x China. 12 May 1890, (holotype, P!). Inst. rte. de Batang, d'Orléans s.n. Distribution. China (Sichuan, Xizang, Yunnan). 21. Solms-laubachia pumila (Kurz) F. Dvorak, Folia Prirodoved. Fak. Univ. Purkyne Brne, Biol. 13(4): 24. 1972. Basionym: Parrya pumila Kurz. Flora 55: 285. 1872. Vvedenskyella pumila (Kurz) Botsch., Bot. Mater. Gerb. Bot. Inst. Akad. Nauk S.S.S.R. 17: 176. 1955. pumila (Kurz) Jafri, Fl. W. 1973. Desideria pumila (Kurz) Al-Shehbaz. ins Missouri Bot. Gard. 87: 560. 2000. TYPE: E wo 15,000-18,000 ft. [4572— 5486 m]. s.d., F. Stoliczka s.n. (holotype, CAL not seen; M KI). Komarova C "Nd Pakistan 55: 15 Ermania koelzii O. E. 332. 1933. Schulz. Repert. Spec. Nov. Regni Veg. TYPE: [Pakistan] Kashmir. Rupshu. Kyensa La, 19,000 ft. [5791 m]. 9 July 1931, W. Koelz 2231 (holotype, B!). Ermania che Botsch., Bot. Mater. Gerb. Bot. Inst. Akad Nauk S.S.S.R. 17: 164. 1955. Oreo- blastus bifarius (Botsch.) Suslova, Bot. Zhurn. (Moscow : Leningrad) 57: 652. 1972. TYPE Kuen-Lun, Humboldt Range, June 1894, W. Roborowski s.n. omarova Á $: China. Xinjiang: Ulan- Bulak. 4200 m, 30 (holotype, LE!). Distribution. China (Xinjiang, Xizang), Pakistan (Kashmir). 22. Solms-laubachia retropilosa Botsch., Bot. Mater. GE Bot. Inst. Komarova Akad. Nauk $ : 171. 1955. TYPE: China. |Xizang] Tibet its Sikkan): betw. Toutan « Alabosan (Ngarolak), 27 May 1893, V. PEN. Lamaja, Kasch- karov s.n. (holotype, LE!; isotype, Pu RUM floribunda Y. - Lan & T. Cheo, Acta ylotax. Sin. 19: 475 TYPE: he Xizang: Zan. 5100 m, 3 July pn Qinghai-Xizang Expe- dition 12173 (holotype, PE [as HC]*: isotypes, HNWP', KU N. NAS). Distribution. China (Sichuan, (see Al-Shehbaz € Yang. 2001). Xizang, Yunnan) 23. Solms-laubachia stewartii (T. Al-Shehbaz & H. nym: Cheiranthus stewartii T. Anderson in J. D Hooker, Fl. Brit. India 1: 132. 1872. Ermanta stewartii (T. Anderson) O. E. Schulz. Bot. Syst. 06: 98. 1933. Christolea , Notes Roy. Bot. 1955. Ureobiastus stewartii (T. Anderson) Suslova, Bot. Zhurn. (Moscow & Leningrad) 57: 653. 1972. Desideria (T. Anderson) Al-Shehbaz, Ann. Gard. 87: 556. 2000. TYPE: [India] Kashmir. Ladak, 15,000-16,500 ft. [4572-5029 m], s.d., J. L. Stewart s.n. (holotype, K!: ED. Anderson) J. P Yue, Sun, comb. nov. Basio- Jahrb. stewartii (T. Gard. Edin- stewartii Missouri Bot. isolype, Distribution. China (Xizang), India (Himachal Pra- desh, Punjab, Kashmir) (see Al-Shehbaz & Yang, 2001). = 24. Solms-laubac m = J. P. Yue & Al- ‘hina. Sichuan: Jiulong Shehbaz, sp. nov. TYPE g Xian, Tanggu Xiang, NW of city of Jiulong, Wuxu Volume 95, Number 3 08 Yue et al. 537 Solms-laubachia from Western China Hai Abies, understory & open, (Wuxu Lake) 29%9'11N, 101 24'25"E, Picea, Quercus forest with Rhododendron dry, and boggy meadows, crevices of rocks, 4175 m, 22 June 2005, D. E. Boufford, J. H. Chen, S. L. Kelley, J. Li. R. H. Ree, H. Sun, J. P. Yue & Y. H. Zhang 33464 (holotype, KUN!; isotypes, A!, MO!). grazed, Herba perennis pulvinata, 1-3 em alta. Folia basalia rosulata, linearia vel lineari-oblanceolata, 0.8-2.5 em X l- 2.5 mm, sparse pubescentia, integra; petiolis 3.5—5.5 mm ongis, incrassatis; folia caulina nulla. Pedicelli fructiferi solitarii, 0.8-1.2 cm longi. Sepala libera, oblonga, 2-3 X 1— 1.2 mm, sparse pilosa; petala ca. 6 mm n longa: Dee ad 12. Fructus lineares vel lineari-lanceolati, 1.6-2.5 cm X 2.8- 3.2 mm, Semina pa odi ca. 2.5 1.5 mm pubescentes. caudex 1-3 em tall; covered with petioles and intact leaves of previous Herbs perennial, pulvinate, blade linear to linear- years. Basal leaves rosulate; oblanceolate, 0.8-2.5 cm X 1-2.5 mm, grooved prox- imally, sparsely pubescent, not ciliate, base attenuate, margin entire; petioles 3.5—5.5 mm, thickened; cauline leaves absent. Fruiting pedicels solitary from basal rosette, 0.8—1.2 cm long. Sepals free, oblong, 2-3 X 1— 1.2 mm, sparsely pilose; petals ca. 6 mm; anthers ca. | mm: ovules 8 to 12 per ovary. Fruit linear-lanceolate 1.6-2.5 cm X 2.8-3.2 obscurely reticulate veined; septum complete; style to linear, mm; valves pubescent, 0.6-1 mm; stigma strongly 2-lobed. Seeds ovate- oblong, ca. 2.5 X 1.5 mm mm Distribution. y endemic to Jiulong Xian in Narrow Sichuan Province, China. IUCN Red List category. (CR) (IUCN, 2001) Critically Endangered Discussion. Solms-laubachia — sunhangiana is named in honor of Sun Hang. deputy director of the of Botany, for his continuous Kunming Institute support of fieldwork on the genus. This novelty is easily distinguished from the remaining species of the genus with entire leaves by having much smaller flowers (sepals 2-3 mm vs. [5.5-]5-8|-11] mm long) and narrower fruits (2-3 mm vs. [4-]5-18 mm wide). It is closely related to 5. xerophyta, which it resembles by having persistent, linear, narrow (less than 3 mm wide) leaves grooved proximally. Such leaves are also found in 5. zhongdianensis, a species unique in the genus by its thinner, papery petioles instead of thickened ones. 25. Solms-laubachia xerophyta (W. W. Sm.) H. F. Comber, Notes Roy. Bot. Gard. Edinburgh 18: 249. 1934. Basionym: Parrya xerophyta W. W. Sm., Notes Roy. Bot. Gard. Edinburgh 12: 217. 1920. TYPE: China. Yunnan: NE of Chungtien, 2755'N, July 1918, C. o 16444 (holotype. E!; isotypes, E!, K!, P!, Distribution. China (Sichuan, Yunnan) (see Al- Shehbaz € Yang, 2001). 26. Solms-laubachia zhongdianensis J. P. Yue, Al-Shehbaz & H. Sun, Ann. Bot. Fenn. 42: 156. 2005. TYPE: China. Yunnan: Shangri-la Co., Mt. Shika, scree, sandy areas, 27°47'N, 99°35'E, ca 4450 m, 27 Sep. 2001, J. P. Yue 154 (holotype, KUN!; isotype, MO"). Distribution. China (Yunnan). CONSERVATION STATUS The four new species described above (S. angusti- folia, S. calcicola, S. grandiflora, and S. sunhangiana) are known from only their type specimens or a few additional collections. Following the IUCN guidelines , 2001), and because all four have very narrow = distribution ranges, our preliminary conservation assessment would indicate that they are Critically Endangered (CR). Literature Cited Al-Shehbaz, I. A. 1984. The tribes of Cruciferae (Brassica- ceae) in the southeastern United States. J. Arnold Arbor. 65: 343-373 1988. The (Cruciferae; Brassicaceae) in the States. " Arnold Arbor. 69: 193-212 000 spend»: in the Brassicaceae genera of Anchonieae (Hesperideae) United southeastern review of gan "e a revision of Desideria, b a critical rod of related genera. Ann. Missouri Bot. Gard. 87: 549-563 . 20 Desideria mieheorum secar. a new species from Tibet. Novon 15: & G. Yang. 2001. A a Harvard Pap. Bot. . Beilstein & E. A. Ke Lo 2006. E malics and ven a the Brassis raceae (Cruciferae): An overview. . Evol. 259: 89-120. An, Z. X. [An C. i 1995. Cruciferae. Pp. 38-229 in Z. M. Mao (editor), Flora Xinjiangensis, Vol. 2(2). Xinjiang Science, Pub revision Lor Solms- ene 371-3 =— Technology, and Hygiene ishing House, Urumqi. Baum, D. A. R. L. Small & J. F. Wendel. 1998 Bioge ography and floral evolution of baobabs (Adansonia, 'aceae) as inferred from multiple data sets. Biol. 47: ed Beilstein, M. A., I. Al-Shehbaz « E. Brassicaceae olsen and trichome evolution. Bot. 93: 607 Botschantzev, 7 E Mater. Gerb. Bot. Le Komarova Akad. 160-178 Cunningham, ( 1997. Can predict when data should be combined? Molec. 14: 733-740 A. Kellogg. 2006. Amer. J. De Cruciferis notae criticae. Bot. Nauk S.S.S.R. 17 inc ongruence tests Biol. Evol. three Annals of the Missouri Botanical Garden 1970. Sy st. A numerical Zool. 19: ALG. Kluge & M. J. Eckardt. oa h to phylogenetic 17 -191. Farris, J. S systematics. Källersjö, A. G. Kluge € C. Buil. 1994. Testing C ucc b 315-319. . 1995. Constructing a Syst. Biol. 44: significance of incongruence. — , N significance test for incongruence. 570-572. Fitch, W. M. 1971. Minimal change for a specific tree topology. Syst. Zool. 20: 406— a Hedge, 1. C. . Å systematic and Ee0 gps 'al survey of the de i: orld Cruc ie rae. Pp. | J. MacLeod & B. M The Biology and Chemistry of Cruc s Tae. Academic Press, R. F. 1997. Cruciferae. Pp. Toward defining the course of evolution: —45 in J. G. Vaughan, A. J. . Jones (editors), London 410-505 in S. W. 1 (editor), Flora Qinghaiica, Vol. 1. Qinghai People’s Publishing House, Xining. IUCN, 2001. IUCN Red List Categories and Criteria Version 3.1. Prepared by the IUCN Species Survival Commission. IUCN, Gland, Switzerland, and Cambridge, Unit . 1942. Das system der Cruciferae. Oesterr. Bot. Koch. M.. I. Al-Shehbaz Molecular systematics & K. Mummenhoff. 2003. s, evolution, and population biology in the br family (Brassicaceae). Ann. Missouri Bot Gard. 90: 151-171. Kuan, K. js 1985. Solms-laubachia. Pp. 381-384 in C. (editor), Flora Xizangica, Vol. Lan, Y. Z. 1987. Be ditor). Science Press, Beijing. — o T. Y.( 1981. On the Chinese genus Solms- laubachia Muschler (Cruciferae). Acta Phytotax. Sin. 19: 412-480. Y. Wu Science Press, Beijing. 326-339 in T. Y. Popularis Sinicae, Solms-laubachia. Pp. : gi Flora R eipublic ae Vo sheo. - Solms- laubachia. Pp. 96-100 in C. Y. Wu, ı (editors), Flora Yunnanica, Vol. 6. Al-Shehbaz, M. Koch & T. pps 25 € rue iler evolution in the post-genomic era. . 119- in R. J. Henry (editor), Plant Diversity and os uU snolypic and Phenotypic Variation in Higher Plants. CABI United Kingdom. Publishing, Oxfordshire and Cambridge, Murley, M. R. 1951. Seeds of the Cruciferae An northeastern North vi rica. Amer. Midl. Naturalist 46: 1-81. Sc us O. E. 1936. Crue de rae. Pp. 227-0! vu in A. Engler & fme (editors). Die natürlichen Pflanze bl `n, a 17B. 5 ag von Wilhelm Engelmann, Leipzig Simmons, M. P. & H. Ochoterena. 2000. Gaps as c tame ‘ters in sequence us d phylogenetic analyses. Syst. Biol. 49: 309-381. Swofford, D. L. 1998. PAUP* Parsimony, Vers. 4.0. Sin Sullivan, J. of among- w^ 'arialion. Syst. Biol. 45: 375-380. Taberlet, P., L. Gielly, G. Pauton & J. Bouvet. 1991. Universal primers * amplification. of Vile non-coding regions of M ee 2 s PI. Molec. Biol. 17: 1105-1109. Tan, Z. Z. Q. Zhang, Y. "n ao $i S. D. Zhou. T PA P A a G (editor), uanica, m . Sic au | Press, Chengdu. Gibson, : Phylogenetic Analysis Using Phylogenetic Analy Using auer, Sunderland, Massachusetts. ». Combining data with different distributions 1996, Coml g data with diff t distribut 1999. Flora E Plewniak, F. Eanmougin & : luétal* Windows interface: Flexible strategies for multiple se a nce inne nt aided by quality PEU tools. Nucl. Acids 24: 48 870—4882. Warwick, Al-Shehbaz, C. J Koch. TA Pirog ny of Braya and daa (Brassicaceae) based on iclear Thompson J.I pue D. G. b 1997. The des er, . Harris & ribosomal | internal transcribed spacer and dou trnl. intron sequences. Canad. J. ue 2: ig )2. auder, Al-Shehbaz & F. 2007. P by fee netie re E in the tribes Anchonieae, Jacquemoud. oe ae, Euclidieae, and Hesperideae (Brassicaceae) based on nuclear gear Missouri D Gard. 94: 78. White, T. J., T. Bruns, s e & J. W. Amplification and. direc DNA sequences. Ann. 1990. s ncing of fungal ribosomal 3 Taylor. e genes for S cn netics P 15- Viu in M. A. Innis D. H. Gelfai . J. Sninsky & (editors), PC R cd A Guide to Methods i dela ation. Academic Press, New York Yue. J. P. Z. J. Gu LA. Al-Shehbaz & H. Sun. Cytological studies on the / o 2004. Sino-Himalayan endemic related (Brassicaceae). Bot. J. Linn. Soc. 145: 77-86. , H. Sun, LA. Al-Shehbaz € J. H. Li. 2006. Support for an expanded Solms-laubachia (Brassicaceae Solms-Laubachia Muschler and two genera Evi- dence from sequences of chlor and nuclear genes. Ann. Missouri Bot. Gard. 93: 402—411. Volume 95, Number 3, pp. 405-538 of the ANNALS or THE Missouri BOTANICAL GARDEN ished on September 23, 2008. was pub issouri Botanical Garden Libra www.mbgpress.org CONTENTS A Revision of the Solanum havanense Species Group and New Taxonomic Additions to the Geminata Clade (Solanum, Solanaceae) Sandra Knapp Phylogenetic Relationships of Two Endemic Genera from East Asia: Trichocoleopsis and Neotrichocolea (Hepaticae) ___ Yang Liu, Yu Jia, Wei Wang, Zhi-Duan Chen, E. Christine Davis & Yin-Long Qiu A 'Tizonamie Revision of the South African Endemic Genus Arctopus (Apiaceae, Saniculoideae) Anthony R. Magee, Ben-Erik van Wyk, Patricia M. Tilney & Michelle van der Bank A Natural Hybrid Between Ligularia paradoxa and L. duciformis (Asteraceae, Senecioneae) from Yunnan, China Yuezhi Pan, Suhua Shi, Xun Gong & Chiaki Kuroda A Review of the Genus Pyrostegia (Bignoniaceae) Amy Pool Anatomy and Histochemical Localization of Lipid Secretions in Brazilian Species of Pani- cum Sect. Lorea (Poaceae, Panicoideae, Paniceae) _ Andrea S. Vega, María A. Castro & Fernando O. Z q A Synopsis of an Expanded Sohne ouha (Brassicaceae), and the Description of Four New Species from Western China A Ji-pei Yue, Hang Sun, Jian-Hua Li & Ihsan A. Al-Shehbaz Cover illustration. Archaefructus reconstruction by David L. Dilcher and K. Simons, Florida Museum of Natural History Annals of the Missouri Botanical Ga m Volume 95 umber Annals of the Missouri Botanical Garden Volume 95, Number 4 December 2008 The Annals, published quarterly, contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, independent reviewers. Authors should write the Managing Editor for information concerning arrangements for publishing in the Annals. Instructions to Authors are printed in the back of the last issue of each volume and are also available online at www.mbgpress.org. Editorial Committee Victoria C. Hollowell Scientific Editor, Missouri Botanical Garden Beth Parada Managing Editor, Missouri Botanical Garden Allison M. Brock Associate Editor, Missouri Botanical Garden Tammy Charron Editorial Assistant, Missouri Botanical Garden Cirri Moran Press Coordinator, Missouri Botanical Garden Patricia M. Eckel Latin Editor, Missouri Botanical Garden Ihsan A. Al-Shehbaz Missouri Botanical Garden Gerrit Davidse Missouri Botanical Garden Roy E. Gereau Missouri Botanical Garden Peter Goldblatt Missouri Botanical Garden Gordon McPherson Missouri Botanical Garden Charlotte Taylor Missouri Botanical Garden Henk van der Werff Missouri Botanical Garden For subscription information contact ANNALS or THE Missouni BOTANICAL GARDEN, £o Allen Mar- keting & Management, P.O. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2008 is $170 per volume U.S., $180 Canada & Mexico, $205 all other countries. Four issues per volume. The journal Novon is included in the subscription price of the Annals. annals@mobot.org (editorial queries) http://www. mbgpress. org THE ANNALS OF THE Missouri BOTANICAL GARDEN (ISSN 0026-6493) is published quarterly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. Periodicals post- age paid at St. Louis, MO and additional mail- ing offices. Postmaster: Send address changes to ANNALS OF THE MissourI BOTANICAL GARDEN, Yo Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. The Annals are abstracted and/or indexed in AGRICOLA (through 1994), APT Online, BIOSISO, CAB Ab- strac/Global Health databases, ingenta, ISIO databases, JSTOR, Research Alert®, and Sci Search®. The full-text of ANNALS or THE MISSOURI BOTANICAL GARDEN is available online though BioOne™ (http:// www.bioone.org). © Missouri Botanical Garden Press 2008 The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and their environment, in order to preserve and enrich life. This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Volume 95 Annals Number 4«" of the 20004 ¿e 08% ^^ Missouri m art oo Botanical E Garden CHROMOSOME NUMBERS IN D. C. Albach,? M. M. Martínez- Ortega VERONICEAE (PLANTAGINACEAE): : | H. oi cun REVIEW AND SEVERAL NEW id mE COUNTS! ABSTRACT Chromosomal evolution in Veronica L. and related genera (W ae Jacq., 4 ulfeniopsis D. Y. Hong, Paederota L., Lagotis Gaertn., Picrorhiza Royle ex Benth., and Veronicastrum Hei ex Fabr: Veroniceae, Plantaginaceae; formerly Scrophulariaceae) is prese ented. To this end, we conducted an extensive ie iure survey of more than 400 publications covering ca. 300 out of 500 species in the tribe. We also report 44 new chromosome counts. Chromosome numbers of Veronica hispidula Boiss. € Huet. (2n = 18, 36) and V. reuteriana Boiss. (2n = 28, 42) are reported for the first time, and both species exhibit intraspecific ploidy level variation, Other new counts confirm chromosome numbers reported previously. The evolution a ie, of chromosome numbers in Veroniceae is discussed in light of recent results i INA-based phylogenetic analyses. Most of the subgenera of Veronica exhibit only one single basic number, i.e., x = 6, 7,8. 9, 12, 17, or 20/21. In this genus, the putative ancestral base number of 9 has been reduced s al times to 8 and 7, respec rs (aneuploidy/dysploidy), often associated with transition to annual life history. In contrast, no » mambi iguous increase of chromosome base nu es has been inferred. A table that includes all species of Veroniceae and, if known, their chromosome number and full sectional and subsectional lassification in Veronica is provided. For this purpo pos ew combinations have been MEME (Veronica. sect. Vni (Rómpp) Albach, Veronica sect. Glandulosae " (Rümpp) lbach. and Veronica subsect. Cochlidiosperma (Rehb.) Albach). Key words: Annual habit, chromosome numbers, dysploidy, Plantaginaceae, polyploidy, Veronica, Veroniceae. The tribe Veroniceae within the Plantaginaceae distributed mainly in temperate. regions of the (sensu Angiosperm Phylogeny Group, 2003; formerly Northern Hemisphere and Australasia. By far, the S105} yrogens | ] I j part of Scrophulariaceae) comprises about 500 species — largest genus in the tribe is Veronica L., with ca. 450 ! We thank the Austrian Science Foundation (FWF) project P-15336 for funding research by DCA. The input of MMMO and LD was partly supported by the Junta de Castilla y León through the research project 5A048A05 and partly by the project Flora Iberica VI (RE N2002-04634-C05- 02). Help from K. Marlowe and Y os in getting literature, translations by Y.-P. Guo and S. von Mehring, and loan of voucher specimens by the curators of G, M, MS de U, NCU, and NEU are also gratefully acknowledged. We "iso thank colleagues who have helped collect material for this "Institut für Spe zielle a a und Botanischer Garten, Johannes Gutenberg-Universitil ME Bentzelweg 9b, 55099 * Departamento m Botánica ven: idad de Salamanca, E-37007 Salamanca, Spain. mmo@usal.es; Idelsan@usal.es. ‘Department of Systematic and ANO Chona Botany, Faculty Center Botany, Universität Wien, Bennwee 14, 1030 Wien, ° Department of Biology, Fac culty of Science and Arts, Yüzünci ü Yil Un niversity, 65080 Van, Turkey. doi: 10.3417/2006094. ANN. Missouni Bor. Garb. 95: 543—566. PUBLISHED ON 30 DECEMBER 2008. Annals of the Missouri Botanical Garden species, which has recently been recircumseribed to Australasian species of the Hebe 2004): include (again) the complex (Albach et al., Garnock-Jones et al., 2007). Growing information on a relation- ships in the tribe (e.g., Albach & Chase, 2001, 2004; Albach et al., 2004a, c) has highlighted A need to summarize available information for characters im- = portant for the evolution of the tribe and to point out critical taxa for which this character has not yet been One ol Since the first publication of investigated. [these characters is. chromosome number. chromosome numbers in Veronica by Heitz (1926), many authors have published chromosome numbers, including some exten- 19506; 2001) and other extensive sive regional surveys (e.2.. Iceland, Love & Love, Belarus, Dzhus & Dmitrieva, surveys of smaller groups within Veronica (e.g... subsection Acinifolia (Rómpp) Stroh; Fischer, 1972). Albach et al. (2004a) gave the first overview of the evolution of chromosome numbers in Veroniceae. The principal result was that x = 9 is the ancestral chromosome number in Veronica, with just one reduction lo x = 8 but four independent reductions to x = 7 (one from x = 8, but three times from x = 9) (Albach et al., 2004a). transition to annual life history (Albach et al., 20042). Dysploidy i is. in most cases, associated with a Here, we investigate the evolution of chromosome detail and highlight numbers of Veroniceae in more those species or populations that need further investi- galion. Evolutionary relationships detected in major contributions to floras (e.g., Turkey and Iran, Fischer, 1978, 1981: China, Hong & 1998: Martinez-Ortega et al., in press) combined with results Albach et al., Fischer. Spain, from molecular systematic analyses (e.g.. 2004a) form the basis of the discussion. The present contribution is based on previously published chromosome numbers, but also includes 44 new reports of chromosome numbers. This review presents a first attempt. to survey all chromosome numbers worldwide for the entire tribe Veroniceae, and although we are aware that our list is not vet complete, we hope that this review will allow some evolutionary conclusions and also stimulate chromo- some studies of those species that are the most interesting but not yet counted. Finally, we want to strongly emphasize the great need for good quality figures accompanying. the publication of any new chromosome numbers and the necessity to provide voucher information for each count, MATERIALS AND METHODS have found than 400 publications We including chromosome numbers in Veroniceae repre- more senting more than 2600 populations. Several compre- hensive reviews of chromosome numbers for regional floras (e.g.. the former Soviet Union, Agapova et al., 1993: Austria, Dobeš € Vitek. 2000) as well as worldwide (e.g.. Goldblatt & Johnson, 2006) do exist and greatly. facilitated. this work. Concomitantly, frequent errors. have been encountered in these surveys. Whenever feasible, we checked the original source, and eventually we checked ca. 350 original publications (including ca. 2500 studied. populations within the Veroniceae). Here, Albach et al. (2004b) and the species in Veroniceae we use the supraspecific classification of currently accepted by us (Albach, Martinez-Ortega & Fischer, unpublished data). Author names of taxa are given in Appendix. l. Original chromosome counts were conducted mostly using seeds collected in the field and. germinated in the laboratory. For Veronica bellidioides l., actively growing root tips of. plants cultivated in the Botanical Garden of the University of Vienna (Austria) were used. Root tips were pretreated with 0.002 M 8-hydroxyquinoline for 2 hr. at room fixed in 3 temperature and 1.5 hr. at 4 C, and acid for 24 hr. at room temperature and stored al —20 C or 4 C until use. For V. persica Poir. as well as other taxa studied by LD and MMMO, in the field and fixed ethanol:l acetic flower buds were collected modified Carnoy’s solution (4. chloroform:3 absolute 70% ethanol, and stored at 4 € until use. For chromosome ethanol: glacial acetie acid). transferred. to counts by HWS and DCA, material was hydrolyzed in 5N HCI for 20 min., washed, and Feulgen-stained in Schiffs reagent (Fukui € Nakayama, 1996) for l- Material analyzed by MMMO and 1954). All squash preparations were made in a drop of 45% 1.5 hr. in the dark. LD was stained in 2% acetic orcein (La Cour, acetic acid. Chromosome numbers were analyzed under light microscopy (Carl Zeiss AG, Oberkochen, HWS and DCA or Nikon Optiphot Tokyo, Japan) by MMMO and LD and black-and-white Germany) by (Nikon Corp., documented (Fig. 1). For mum of five) with well-spread chromosomes were used with photography each individual, several cells (a mini- for chromosome number determination. Chromosome counts by MAF followed the methods given in Fischer Vouchers DCA MAF deposited. in the herbarium of. the University of Vienna (WU) and those from MMMO and LD at University of Salamanca (SALA). In an electronic appendix (available at ), we have summarized ca. chromosome counts from Ca. haploid and diploid counts were considered. published as well as the currently accepted name are given. Volume 95, Number 4 2008 Albach et al. Chromosome Numbers in Veroniceae (Plantaginaceae) 2n), levels To whether chromosome number (n, chromosome significantly differ between annuals and perennials, the Mann-Whitney U test has been used. lest base number (x), and ploidy To avoid an overly large influence from the large radiation. of Australasian species in the Veronica, which are likely of hexaploid origin with respect to other Veronica, statistical were conducted with and without Australasian species as well as those species Picrorhiza comparisons with x = 17 (Veronicastrum Heist. ex Fabr., Royle ex Benth., Veronica subg. Pseudolysimachium (Opiz) Buchenau). RESULTS AND. DISCUSSION Forty-four chromosome counts reported here for the first 13 significant reports, photographs are presented (Fig. 1). time are summarized in Table l, and for — Currently, chromosome numbers for 330 of 509 species (65%) in Veroniceae have been published (Appendix L ing at ). However, some groups are Cochlidiosperma see electronic appendix for a comprehensive list- well known (e.g. Veronica subg. (Rchb.) M. M. Mart. er & Albach and Veronica subg. Synthyris (Benth.) M. M. Mart. Ort., Albach & M. A. Fisch.—all known species have been counted), whereas other groups include many species for which no count is available. Most notably, chromosome numbers of less than one fourth (2396) of the species of Lagotis Gaertn. and only about one third (3296) of the species of Veronica subg. Stenocarpon (Boriss.) M. M. Mart. Ort., Albach & M. A. Fisch are known. Chromosome base numbers (Table 2) and ploidy levels (Table 3) have been summarized, arranged by subgenus in Veronica. Albach et al. (2004a) discussed the ancestral chromo- some number in Veroniceae, which is inferred to be x = 9. although x 1964: Albach & Chase, 2004). They further inferred a single 8 clade). — 5 is also possible (Lepper, — origin of x — 8 within Veronica (the x However, a more detailed analysis shows that indepen- dent origins of x — 8 occurred in several instances independently (Fig. 2: see below). Albach et al. oe) ~ also reported four independent changes to x (Fig. 2), whic two from x 1 now needs to be revised to six changes = 9. optimization) because of overlooked numbers in V. densiflora Ledeb. (Veronica subg. Stenocarpon) and V. Y W. D. No unambiguous chromosome — three from x = 8, one ambiguous — pu magna M. A. Fisch. (Veronica subg. Chamaedrys J. Koch) Buchenau). number increases between x = (Albach et 2004a; and mue by information presented here). The evolution of an annual life history in Veronica has 7,8, and 9 are inferrec al., been studied in various aspecls in recent years. Molecular phylogenetic analyses have shown that an annual life history arose several times independently (Albach et al., 2004c). Veronica has demonstrated that a low genome size is correlated with selfing rather life (Albach & Greilhuber, 2004) as previously 1972). A higher DNA substitution rate, however, is not correlated with selfing, but rather Analyzing. genome size in than with annual history suggested (Bennett, with an annual life history in Veronica (Albach & Müller, in prep.). supporting more general results for all angiosperms (Bousquet et al., 1992). Therefore, it seems appropriate to analyze whether chromosome numbers differ between annuals and perennials. Several studies (e.g.. Levin, 2002) have suggested a reduction in chromosome number associated with the shift from perennial to annual life history. Standard statistics are not as powerful as statistical methods incorporating phylogenetic evidence due to the confounding effect of shared evolutionary history on the results (Harvey & Rambaut, 1998). However, a fully resolved phylogeny for all species of Veroniceae for which chromosome numbers are available does not exist and is not even possible to generate due to the reticulate history especially of polyploid taxa (e.g.. Albach, 2007). The results of the statistical analysis nevertheless appear robust in showing annual species to have a significantly lower chromosome base number but not a lower ploidy level. The inclusion of Australasian species of Veronica and those species with x = 17 (Veronicastrum, Picrorhiza, Veronica subg. Pseudolysimachium) in an analysis has the effect that annual species have a significantly lower chromosome number and ploidy level. Without those perennial species, annual species have on average a ploidy level of 3.3, whereas perennial species have a ploidy level of 3.4, which is not different (P = 0.85). Chromosome base differ significantly, with a lower number found in annuals (P < 0.01). We could not check whether this is an indirect effect of a correlation significantly numbers, however. of selfing with a lower chromosome base number, as is the case in the evolution of genome size (Albach & Greilhuber, 2004), due to the paucity of information on breeding systems in the Veroniceae. Both reduction in chromosome number and selfing may reduce recombi- nation rates and, therefore, it appears likely that both are selected for instances in which a low recombination rate is advantageous. The chromosome number is the product of ploidy level and chromosome base number, and its distribution is not significantly different (P = 0.19) within the Veroniceae. VERONICEAE EXCLUDING VERONICA Veroniceae in the circumscription of Albach et al. (2004b) comprise nine genera, including two mono- Table 1. Information on chromosome counts in Veronica reported here for the first time. Counts performed by DCA. HWS. LD. MAF, and MMMO. Species Chromosome number Counted by Origin Voucher . alpina L. 2n — 18 LD. MMMO Spain. Huesca: Torla. gen N face of Pico o "Lo M. M. Martínez-Ortega 300 (SALA) alpina 2n — 18 LD. MMMO Spain. Huesca: Benasque, La Renclusa, 7 Aug. 199 M. M. Martínez-Ortega s.n. (SALA) . anagalloides subsp. heureka M. A. 2n — 18 HWS. DCA Georgia. Greater Caucasus: Cross Pass D. C. Albach 299 (WU) Fisch. V. aphylla L. n = 9, 2n = 18 LD. MMMO Spain. Navarra: Isaba, Portillo de Arrasarguiat E aa 337 ( SALA 110613) V. arguteserrata Regel & Schmalh. 2n = 42. Fig. 1A HWS, DCA Turkev. Bitlis: Murat valley betw. Hamur and Tutak D. C. Albach 685 (WU) V. bellidioides L. 2n = 36 HWS. DCA Bulgaria. Mt. Pirin D. Albach pe (WU) V. bellidioides 2n — 36 HWs Spain. Pyrenees de Se edo 146 de (WU) V. bombyeina Boiss. & Kotschy subsp. 2n = 16 MAF Cult. Bot. Gard. Vienna, 13 May 1969 Fischer s.n. (V bombycina V. bozakmanii M. A. Fisch. 2n — 28 HWS, DCA Turkey. Bitlis: Kucuksu D. C. Albach 611 (WU) V. bozakmanii 2n = 28, Fig. 1B HWs, DCA Turkey. Bitlis: W of Kuzgunkegan Pass D. C. Albach 667 (WU) V. campylopoda Boiss. 2n — 42, Fig. 1C DCA Turkev. Van: W of Güzeldere Pass D. C. Albach 653 (WU) V. campylopode 2n = 56. Fig. 1D DCA Turkey. Van: Catak (cultivated at Technical University of Jensen IOK-6/2003 (WU) Denmark. Lyngby) + chamaedrys L. subsp. chamaedrys 2n = 32 LD. MMMO Spain. Oviedo: Quirós. toward Pico Gamoniteiro L. Delgado 708 & E. Rico (SALA 110656) | chamaedrys subsp. chamaedrys 2n = 32 LD. MMMO Spain. Teruel: from Fontanete to Valdelinares E. Rico 7078 (SALA 110655) V. ciliata subsp. cephaloides (Pennell) 2n = 16, Fig. IK HWS, DCA China. Xizang (Tibet) G. Miehe 98-16717 (GOET) D. Y. Hong V. cymbalaria Bodard 2n — 36 MAF Greece. Crete, Omalos: Xvloskalon, 6 May 197] H. Malicky s.n. (WU) | cymbalaria 2n = 30 MAF Greece. Crete, Ida mtns.: near Ideon Antron cave, 9 May H. Malicky s.n. (WU) [971 . cymbalaria 2n = 54 MAF Greece. e an e lands. Kefallinia: area betw. Sami & M. & G. Fischer s.n. (WU) Dhihal Apr. 1974 + cymbalaria 2n = 54 MAF Greece. ain Islands, Kefallinia: Mt. Rudi, 13 Apr. 1974 M. & G. Fischer s.n. (WU) . cymbalaria 2n — 54 MAF Greece. lonian Islands, Kefallinia: area betw. Sami & M. & G. Fischer s.n. (WU) Dhihalia. 14 Apr. 1974 | filiformis Sm. 2n = 14 Tanja Lessel, DCA Austria, Graz: Schubertstrasse, in front of Botanical Garden D. C. Albach 857 (MJG) | fruticans subsp. cantabrica M. Laínz n = 8, 2n =16 LD, MMMO Spain. Cantabria: Hermandad de Campoo de Suso, base of M. M. Martínez-Ortega s.n. (SALA) Pico Tres Mares, 11 July 1996 + fruticans subsp. cantabrica 2n = 16 LD, MMMO Spain. Segovia: El Cardoso de la Sierra. Pico del Lobo. 2 M. M. Martinez-Ortega s.n. (SALA) July 1996 | fruticans subsp. cantabrica 2n = 16 LD. MMMO Spain. La Rioja: Canales de la Sierra. Pico Gatón M. M. Martinez-Ortega s.n. (SALA) 9v9 əy} jo sjeuuy USPIBO) [jeoiuejog unossilA Table 1. Continued. Species Chromosome number Counted by Origin Voucher V. fruticans subsp. cantabrica 2n — l6 LD, MMMO Spain. Ávila: Solana de Ávila, Lagunas de El Trampal, N E. Rico 7105 (SALA 110610) slope V. fruticans subsp. cantabrica 2n = 16 LD, MMMO Spain. Lérida: Pla de Beret L. Delgado 344 & I. Soriano (SALA 609) V. gentianoides Vahl 2n = ca. 72 DCA Georgia. Greater Caucasus: W of Truso Gorge D. C. Albach 350 (WU) V. gentianoides 2n = 36, Fig. II HWS. DCA Georgia. Greater Caucasus: meadows above Zaminda Sameba D. C. Albach 341 (WU) V. gentianoides 2n = ca. 72 DCA Georgia. Greater Caucasus: beech forest below Zaminda D. C. Albach 318 (WU) Sameba V. gentianoides 2n = 46-48, Fig. IH. DCA Georgia. Greater Caucasus: below Dzuta D. C. Albach 333 (WU) V. hispidula Boiss. & Huet. 2n = 18 (5 seeds HWS, DCA Turkey. Van: Degitmenkóy D. C. Albach 635 (WU) ig. IE; 2n = (1 seedling) Fig. 1G V. micrantha Hoffmans. & Link n = 8, 2n = LD. MMMO Spain. Salamanca: Penaparda. Mostejal brook oe 34, X. MA Pos Rico Fig. IM 6-VI- e Ss ALA 1 V. nummularia Gouan 2n — 16, Fig. 1J LD, MMMO Spain. Huesca: Ansó. Collado de Petrachema, N side L 2 & I. Soriano p VI- S 110612) ent M. Bieb. 2n = 16-18 DCA Georgia. Greater Caucasus: Kazbe D. C. Albach 325 (WU V. persica Poir 2n — 28 HWS Azerbaijan. Talvsh: forest reserve Quan! dii SW of G. Schneeweiss Geo01 132/28 (WU) Lankaran V. ponae Gouan 2n — 54 LD, MMMO Spain. Huesca: Panticosa, toward lakes M. Santos Vicente 455. L. P. Gavilán ias & L. M". Muñoz Centeno (SALA 110608) V. prostrata L. 2n = 16 LD, MMMO Czech Republic. Beroun M. M. Martínez-Ortega 1331 (SALA) V. reuteriana Boiss. 2n — 28, Fig. IL HWS, DCA Turkey. Van: Catak D. a Albach 691 (WU) V. reuteriana 2n — 42, Fig. IF HWS, DCA Turkey. Bitlis: 5 of Tutak D Albach 676 (WU) V. scheereri (J. P. Brandt) Holub 2n = 32 LD. MMMO Spain. Huesca: Ansó, on the way to Collado de Petrachema L. o & I. Soriano (SALA V. scheereri 2n 2 LD, MMMO Germany. Baden-Württemberg: Schwübische Alb M. M. Martínez-Ortega 1093 (SALA) V. scutellata L. 2n = 18 LD, MMMO Spain. Salamanca: Fuenteguinaldo, meadows of El Potril, M. A Ka Ortega s.n. & E. Rico banks of Río Agueda. 17 June 1996 V. sp. indet.. aff. beccabungoides 2n — 36 HWS. DCA Georgia. Greater Caucasus: Kazbegi D. C. a 329 (WU) V. trichadena Jord. & Fourr. 2n — 36 MAF Turkey. Mugla: Sakar pass, ca. 15 km S of Mugla, 28 G. & M. A. Fischer s.n. (WU) Mar. 1978 (eeooeui&ejue|a) 9992lUOJ8A Ul SISQUINN eulosouloJu2 800c p JequinN 'S6 euinjoA ‘Je Je uoeqiv ¿vs Annals of the Missouri Botanical Garden vo 4 3 Figure 1. Photos of rd counts. Scale bars = 10 um. —A. bozakmanii (Albach 667, WU). —C. V. campylopoda (Albach 65. n U). V. hispidula (Albach 635, Uo. RE V. Veronica arguteserrata (Albach 085, WU). —B. E —D. V. o (Jensen IOK-6/2003, ca reuteriana (Mbach 676, WU). —G. pidula (Albach 635, WU). —H. A J). —A. V. gentianoides (Albach 341. WU). —J. V. MN te (Delgado 332, SALA). —K. V. ixi iu cephaloides (Miehe 98-I6717. GOET). —L. V. reuteriana (Albach 691, WU). —M. V. micrantha (Delgado . SALA). Volume 95, Number 4 Albach et al. 549 2008 Chromosome Numbers in Veroniceae (Plantaginaceae) Table 2. Chromosome base number for species arranged according to major clades in Veroniceae. Ancestral character state for each clade is underlined, based on information given in Albach et al. (2004a) and Wagstaff and Garnock-Jones (1998). x77 x=8 x= x=17 x=20 x=2l x = other Unknown Veroniceae excluding Veronica 2 6 6 7 40 Veronica subgen. Veronica 2 j 19 1 20 Sube. Beccabunga 8 2 13 10 Subg. Pseudolysimachium 18 11 Subg. Synthyris I 18 Subg. Cochlidiosperma 10 d l Subg. Pellidosperma l l 2 Subg. Stenocarpon ll 22 Subg. Pocilla 14 14 Subg. Pentasepalae 33 37 Subg. Chamaedrys l ll l Sect. Derwentia l 4 4 11 3 Sect. Hebe 69 46 3 l6 Other species of Veronica l m 2 — Total (out of 509 species) 26 04 50 25 13 50 43 178 typic ones for which no chromosome numbers are (K. Koch) Maxim. (two counts), L. takedana Miyabe & known (Scrofella Maxim. from China, Kashmirta D. Y. — Tatew. (one count), L. cashmeriana (Royle) Rupr. Hong from the Himalayas). (three counts), and £. glauca Gaertn. (five counts) are Lagotis is the least studied group in Veroniceae diploids (2n = 2x = 22). whereas L. integrifolia to (Willd.) Schischk. (three counts) is tetraploid (2n. = = despite its wide occurrence from eastern Turkey A Alaska. Although 35 chromosome numbers are 4x = . Lagotis minor (Willd.) Standl. has been available, only seven of 31 species in the genus have — studied bn eh (16 counts), with more than half nine counts) being tetraploid (2n = 2x = 44) and the been studied. All reported chromosome numbers are based on x = 11, except for 2n = 54 for L. brevituba rest diploid (2n = 2x = 22). Most of the counts for L. Maxim. (Huang et al., 1996). Unfortunately, no figure minor are from Russia, with only one diploid count is included in the latter publication and, therefore, from Canada. The taxonomic concept of L. minor and this number requires confirmation. Lagotis stolonifera related species is complex, and confusion with L. Occurrence of polyploidy in various taxa of Veroniceae and among pine and perennial species of the tribe. In the line labeled “Perennial species.” note that Australasian taxa and those with x = 17 ae been excluded. 2x 4x Ox 8x 10x 12x 16x/18x Non- eones Veroniceae 10 10 l l l Subg. Verc 14 6 6 1 ENDS Becc ao 13 14 5 l l Subg. P. Se um hium 16 9 Subg. Synth yris 16 5 Subg. Cochlidiosperma 7 3 2 Subg. Pellidosperma 4 Subg. Stenocarpon 10 l Subg. Pocilla 6 7 3 1 Subg. Pentasepalae 22 7 5 6 2 Subg. Chamaedrys 10 2 l Sect. Derwentia 15 5 Sect. Hebe 87 26 15 Others 2 l'otal 117 68 118 8 | 20 16 Annual species 30 16 8 l Perennial species 86 35 15 8 3 All perennial species 86 56 117 18 3 31 16 Annals o the Missouri Botanical Garden Lagotis Wulfenia x z 11 x-9 x - 17 x=8 x=17 X I E Veronicastrum V. urticifolia V. subsect. Veronica V. subsect. Alpinae V. E gunae V. subsect. Beccabunga V. subsect. CAA ubsect. Gentianoides iifoli idula V. subgen. ue um V. subgen. Synth V. subgen. eo V. subgen. Cochlidiosperma anica V. subgen. Mute scent V. subgen. Stenocarpon V. arvensis V. verna V. V. subsect. Chamaedrys | subgen. V. laxa Cham- V. magna (——— — V. subgen. Pocilla V. subgen. Pentasepalae x=9 . x= Veronica subgen. Veronica 7, 8* 2n=4 Veronica subgen. Beccabunga x=17 x=12 x=7,8,9 =9 8 2n = 38-42 or 48* 6 X = 8 |veronica subgen. Chamaedrys x=7 Figure adogram showing the relationships of Veroniceae based on information from Albach et al. (2004a, e g Vm ;: Albac " eor d data). glauca in the east and £. integrifolia in the south of 1930). ly, voucher specimens for many counts are not cited the distribution is likely (Hultén, Unfortunate- clearly. A detailed analysis using morphological, karyological, and molecular methods would. therefore. be necessary to disentangle the taxonomie problems ) B | around L. minor. Wulfenia Jacq. (from the southeastern Alps, western Balkan Peninsula, and southern Turkey) and Paeder- Asterisks mark counts that are considered dubious (see text for details). ota |. (from the southeastern nS) have x = 9, Wulfenia being diploid (2n = 2x = 18: W. Jacq.. W. blecicti Lakusié. W. orientalis Boiss.. : baldacit Degen with the the latter three counted just once each) and Paederota V. subgen. Pseudo- veronica = He Sannie carinthiaca first counted 11 times, : Volume 95, Number 4 Albach et al. 2008 Chromosome Numbers in Veroniceae (Plantaginaceae) being tetraploid (2n = 4x = 36), although a single one report. Clearly, more analyses for this species are hexaploid plant of P. lutea Scop. (2n = 6x = 54) was found by Fischer (1969). Wulfeniopsis D. Y. Nepal) includes two species that were both reported Hong (from Afghanistan to mostly as diploids (2n = 2x = 16) based on x = 8 (W. Hong, 11 nepalensis (T. Yamaz.) D. Y. Hong, | count), with only amherstiana (Benth.) D. Y. counts; W. one plant being tetraploid (2n = 4x = 32; amherstiana). Veronicastrum (from eastern Asia) and Picrorhiza (from the Himalayas) have been demon- strated to be tetraploid hybrids of Widfentopsis with Wulfenia and the common progenitor of Veronica plus 2004). Therefore, 2xp¡. and 2xy are 4x with respect to the Paederota, respectively (Albach & Chase, rest of the tribe. Chromosome numbers are only available for one of three species of Picrorhiza (P. kurrooa Royle ex Benth, single count; 2n = 2xp¡. = 34) and six of 18 sibiricum (L.) Pennell (7 counts). V. (3 counts), V. V. japonicum (Nakai) T. species of Veronicastrum, with V. virginicum Farw. brunontanum (Benth.) D. Y. Hong, and Yamaz. (both counted once) = 34). V. villosulum (Miq.) being diploid (2n = 2x Vs T. Yamaz. tetraploid On = j4xy4 = 08: one count). and V. liukiuense (Ohwi) T. Yamaz. octoploid (2n = 8xva = 136: one count) based on x = 17, which is the sum of the chromosome numbers of their inferred parents (xpi /xy,, = 8 + 9). Veronica s.l., according to Albach et al. (2004b) and Garnock-Jones et al. (2007). consists of 12 subgenera. l. VERONICA SUBG. VERONICA Veronica includes 45 species with — Veronica subg. chromosome numbers available for 25 (Appendix 1). (Albach € b) have Molecular systematic Chase, 2001; Albach et al., early branching analyses 2004a, that species of Veronica, as well as V. identified an the seutellata L.. and clade includes African a group of montane to subalpine Chinese-Himalayan species plus V. montana L. as consecutive sisters to a — crown group of montane to subalpine species from the Northern. Hemisphere including the type species V. officinalis L. Fig. 2). African species are characterized by high chromo- (within. subsection. Veronica, some numbers. The two closely related species V. glandulosa Hochst. ex Benth. and V. gunae Schweinf. ex Engl. are both hexaploid (2n = 6x = 54) based on x — 9 (the most common base number in the group: two counts each), although only the extreme western populations of V. glandulosa from Cameroon, corre- sponding to subspecies mannii (Hook. f.) Elenevsky, have been counted. The annual Y. 21 Fresen. also seems to be hexaploid (2n = 6x = 48; one count) based on x = 8; however, the number is based on only desirable. Chromosome numbers of V. scutellata have been investigated from 27 localities throughout its entire distribution range in the northern parts of the Northern Hemisphere. All counts demonstrate that it is a diploid species (2n = 2x = The third section in subgenus on Veronica sect. Montanae (Boriss. ex Elenevsky) Assejeva, includes a group of Chinese-Himalayan species and montana, the the (Albach, unpublished data). montana is diploid (2n = 2x = ils distribution in Europe, with only a single tetraploid (2n = 36) found Verlaque in 1985). group in section Montanae 1s very poorly investigated. — with latter as sister to resl Veronica 18: 10 counts) almost throughout "ance (Seidenbinder & Lóve. The Chinese-Himalayan Only four of more than 20 species of subsection Canae (T. Yamaz.) Elenevsky have been investigated for chromosome numbers, and no clear pattern is present. A base chromosome number reduction to x = 8 seems to have occurred in V. henryi T. Yamaz. (tetraploid, 2n = 4x = 32) and V. miqueliana Nakai (hexaploid, 2n 6x = 48) based on single reports for each. Two reports are available for V. cana Wall. with 2n = 50 and 2n = 52, which may be cases of increased chromosome number from x — 8, reductions from x — 9. or accompanied by a figure, which could have resolved miscounts. neither count is Unfortunately, the conflict. The chromosome number of the fourth species that has been investigated, A2 for robusta (Prain) T. Yamaz.. is also Mr to explain, but again may represent a miscount or an aberrant aneuploid individual, much less likely an indepen- dent reduction to x — 7. The scarcity and ambiguity of the data indicate the need to reinvestigate chromo- some numbers coupled with DNA sequence data in order to infer the evolutionary trends in this group. Veronica can be divided in Veronica Benth. Veronica (5 species), and five other species across Veronica sect. subsect. Alpinae (8 species), V. subsect. four other subsections. All these species have chromosome numbers based on x = 9 with the exception of V. baumgartenii Roem. & Schult. (x = 7; see below). Within Veronica subsect. Alpinae, V. alpina L., V. stelleri Pall. diploids (2n. = 2x = 18; 37, nine, one, three, and one nutans Bong., V. copelandii Eastw., V. ex Link, and V. nipponica Makino are count, respectively, with one a ipee cylo- type mixture: 2x and 4x in alpina in Norway: Knaben & Engelskjon, 1967), Roem. & Schult. is exclusively tetraploid (2n = 4x = Esa V. wormskjoldii 36; four counts). Veronica bellidioides is mostly tetraploid (2n = 4x = 36: 36 counts including ours from the Pyrenees and Bulgaria), but diploid plants (2n = 2x 18; seven counts) are known from the Annals of the Missouri Botanical Garden 1968, 1974). ade quate in most species of subsection Alpinae except weslern Pyrenees (Küpfer. Sampling is = or V. stelleri, which has not been sampled in Alaska. as California and Nevada, and V. nutans, which has only been studied in U.S.A.. but not in the northern Cascades or the phylogeo- ., 2006) southern A. Gray 1s more as well as in Canada graphically important (Albach et a Rocky Mountains. problematic, Veronica cusickii with only two populations from the distributional extremes studied, one diploid (2n. = 18) from California and one octoploid (2n. = 8a 72) from southwestern British Columbia. Veronica officinalis L. has been studied intensively throughout Europe but not beyond the Siberian part of its range. Most chromosome numbers reported are tetraploid (2n = 4x diploids (2n = 2x = 18) found i 2007), the east 1994). the only count outside Europe. 32 or 34 (Gadella € Kliphuis, 1963, 1966; Kliphuis € Wieffering, 1972: Rossitto et al., 1983) likely represent miscounts (voucher specimens Albach. ished data). However, this is difficult to prove 36: more than 80 counts), with Muñoz- (Böcher, Mountains n Portugal Centeno et al., ua 1944), (Stepanov, and in Siberian Sayan Re sports of 2n = for the first three studies have been checked: unpub since only Rossitto et al. (1983) presented a figure, which arguably shows 35 chromosomes. The other four species in subsection Veronica have a fairly restricted distribution. Veronica allionii Vill. from the south- western Alps is diploid (2n = 2x = 18: three counts). Franch. & Sav. tetraploid (2n = 4x = 36; dabneyt Hochst. from the whereas V. onoei from Japan is two counts). Veronica Azores and V. morrisonicola Hayata from Taiwan have not been studied yet. The five remaining e in subgenus Veronica . has been are more problematic. Veronica qd reported as diploid ( (on = 2x = 11 counts) in the Alps, Pyrenees, and Tatra mountains. The closely related V. grandiflora Gaertn. from around the Bering Strait has been reported to have 2n = 48-50 that be associated with a regular ploidy level based on x = chromosomes, a number cannot 9. Veronica urticifolia Jacq. is diploid (2n = 2x = 18; eight counts). Reports of 2n = 64 (Meskova, 1965; Nilsson & Lassen, 1971) probably refer to misidentified V. teucrium L., although at least the voucher of one study (Nilsson € Lassen, 1971) was confirmed to be V. urticifolia (Fischer, unpublished data). A count of 2n — [6 by Mattick (in Tischler, and impossible to verify since no voucher specimen is 1950) appears dubious indicated. Three studies have published chromosome numbers for V. ponae Gouan from the Pyrenees and northern Spain. It was reported as hexaploid (2n = Ox = 54; Kiipfer, 1972, 1974) in five populations (also confirmed by us, see Table 1) and diploid (2n = 2a 16-18: Huber. 1927) in voucher specimen for the diploid count, which was one. Unfortunately. no grown in a botanical garden (Dilger-Endrulat, at TUB, is known. Finally, V. pathians) is inferred pers. comm.). It may be a misidentified V. urticifolia. baumgartenii (from the Car- to have the deviating chromo- some base number x = 7 (2n = 2x = 14: two counts n although it is mostly considered closely related to V. aphylla, a relationship not seen in DNA sequence analyses (Albach, unpublished data). IL VERONICA SUBG. BECCABUNGA The aquatic species of section Beccabunga (Hill) Dumort. This subgenus consists of three sections. are sister to sections Acinifolia (Rómpp) Albach and Serpyllifolia G. Don (Albach et al.. 2004a, 2005a: ‘ig. 2). Whereas species of the former section have a rT] the plesiomorphie 2004a) and aes ies of the other two sections have a base number chromosome base number of x = 9, number for the subgenus ( Albach el al.. of x = 7. Our results of 2n = 2x = 18 and 2n = 2x = 36 (x = 9) for V. hispidula Boiss. & Huet. are important. in that respect because it has been considered to be a member of section Acinifolia (Rómpp. 1928; Fischer, 1972). The different chromo- some base number, however, agrees with its position in analyses of DNA sequences in which it is sister to the rest of the subgenus (Albach, unpublished data: see Fig. 2). It is therefore considered as having an uncertain position within the subgenus here. Species boundaries in section Beccabunga have interpreted differently by different (Borissova, 1955 1981): treated here in a narrow sense with 12 species in the been authors Fischer, the species are section. Vagueness of species boundaries is exempli- fied by and based on—the apparent existence of different ploidy races in several species, which could ther be due to cryptic taxa not rec ognize sd on basis of ] T ids . IUC ntilie alion Ol uses or due to mi: misinter- pretation of taxa. In subsection Beccabunga, Veronica beccabunga L. is diploid (2n. = 2x = 18: subspecies beccabunga l., more than 50 abscondita M. A. MUSCOSA counts; subspecies Fisch., three counts: subspecies Korsh.) Elenevsky, one count), but tetraploid numbers (2n = 4x = 36) have been reported for subspecies beccabunga from Poland (Sokolowska in Skalinska, 1964), Italy (Ferrarella et al., 1981), and Sweden (Lókvist € Hultgard, 1999) and most likely represent autotetraploids. Counts of 2n = 2x = 16 for subspecies beccabunga (Davlianidze, 1980) and subspecies muscosa (Sokolovskaja & Strelkova, 1939 1993) are americana Schwein. according to Agapova et al., probably miscounts. The close relative V. ex Benth. is tetraploid (2n = 4x = 36; 12 counts). Volume 95, Number 4 2008 Albach et al. Chromosome Numbers in Veroniceae (Plantaginaceae) Veronica anagallis-aquatica L. and its relatives are morphologically more difficult to differentiate, which is reflected in incorrectly determined specimens used Veronica — or chromosome counts. lysimachioides Boiss. seems to be a diploid species (2n. = 2x = 18: six counts), whereas V. anagallis-aquatica, V. michauxii Lam., and V. catenata Pennell are tetra- ploids (2n 4x = 36; more than 70, five and 16 Wall. is Veronica counts, respectively) and V. undulata hexaploid (2n = 6x = 54; 12 oxycarpa Boiss. seems to be another tetraploid species ai = 4x = 36; six counts), but a diploid count (2n = = 18) is reported by MeSkova (1965; see Oztiirk & Fi ischer [1982] for doubts regarding this count). The — counts). correct chromosome number for V. poljensis Murb. is unknown. Marchant (1970) states that it is diploid (2n = 2x = 18), but his figures show it to be tetraploid. Öztürk and Fischer (1982) report a tetraploid number, but state that their specimen is intermediate between poljensis and V. anagalloides Guss. Dzhus and Dmitrieva (2001) report a diploid from Belarus, but the species is not known from that area or near it and, be a misidentified V. anagalloides. therefore, may This species is diploid (2n = 2x 18; subspecies anagalloides Guss., Fisch., 10 several tetraploid counts have been pub- 10 counts; subspecies heureka M. counts including the report here). However, lished. MeSkova’s (1965) tetraploid count (2n = 4x = 36) from Ukraine has been considered improbable by Oztiirk and Fischer (1982). and no voucher could be The voucher specimen for the tetraploid 1969) does not resemble the typical subspecies heureka located. count from Afghanistan (Podlech & Dieterle, (Oxtiirk & Fischer, 1982), but the one from Yemen (Podlech, 1986) does (Albach, pers. obs). The voucher specimens from Romania (Vasudevan, 1975), Pakistan (Khatoon, 1991, in Khatoon & Ali, 1993). and Iran (Saeidi-Mehrvarz € Kharabian, 2005) have not been checked. Information on the voucher is only given for the number from Iran. Based on this information, it is probable that tetraploid taxa of V. anagalloides subsp. heureka (2n = 4x = 36) exist in southwest Asia. Sánchez-Agudo, Delgado, and Martí- nez-Ortega (in prep.) recently found tetraploid chromosome numbers (2n = 4x 36) from Spain. Their status needs further attention. The same is true for tetraploid counts of V. anaepliouies a heureka from Russian Far East (2n. = 4x = 36; Probatova et al., 1996), which is by far the easternmost occurrence Differentiation. of this taxon is often of this taxon. difficult as plants often resemble diminutive V. anagallis-aquatica. Veronica scardica Griseb. is another species for which two ploidy levels are (1970) 1 some numbers (2n = 2x = 18) from several countries reported. Marchant ( eports diploid chromo- without further documentation. This ploidy level was also reported by Oztürk and Fischer (1982) from Turkey. Strid and Franzén (in Lóve, 1981) and van Loon and van Setten (in Lóve, 1982), however, report tetraploid plants (2n = 4x = 36) from the Balkan Peninsula. The identity of the voucher specimen for the first tetraploid plant has been confirmed; the one for the second shows a mixture of characters and is not clearly identifiable but is definitely not pure V. scardica. Veronica scardica is a serpentinophyte that might be only an ecological race of V. anagalloides or 1984), could explain the different ploidy levels reported. V. anagallis-aquatica (Fischer et al., which Detailed morphological and karyological analyses together with cultivation experiments will be neces- sary to solve the systematic questions regarding this species. Finally, the annual Veronica peregrina lL., the only member of subsection Peregrinae Elenevsky, was also shown to be a member of section Beccabunga (Albach & Chase, 2001; Albach et al., 2004a). Eleven counts confirm its chromosome number of 2n = 6x = 52, which was hypothesized to be derived from 2n — 6x — 54 by chromosome fusion (cf. Hofelich, 1935; Albach el al., 2004a). Heckard (1992) from Oregon is interesting in that that chromosomes The count of 2n = 54 by Chuang and because it lineage the Alternatively, it may represent a the be a respect diverged before fusion of occurred. may miscount, but no figure is included in the publication. Section Acinifolia comprises eight species (without Veronica hispidula, which we exclude). Previously, chromosome numbers were known for four species, the diploid (2n = 2x = 14 a, V. Kotschy € Boiss., and V. syriaca Roem. & Schult. (11 three and three counts, respectively) and the tetra- ploid V. bozakmanii M. A. Fisch. (2n = 4x = 28; six We reuteriana Boiss., V. acinifolia | dn — counts including the report here). add here information on a fifth species, V. for which within one population two plants showed the tetraploid level (2n 28) and one plant the hexaploid level (2n = 6x = 42). Intraspecific ploidy reuleriana is level variation in hispidula and V. noteworthy. No morphological character correspond- ing to different ploidy levels has been noted but should be looked for in future investigations. Chromosome numbers in other rare species from southwest Asia are still unknown. Sister to section Acinifolia is section Serpyllifolia in subgenus Beccabunga, a group including only peren- nial taxa, which probably shares the ancestral chromosome base number of x = 7 (Fig. This subsection includes 11 species. nas numbers are available for five taxa. These include the diploid V. telephiifolia Vahl (2n = 2x = 14; one count). V. Annals of the Missouri Botanical Garden 70 including repens (Clarion ex DC.) Hartl in Hegi from Corsica), and V. nevadensis var. langei (Lacaita) arl. Ort. € E. Rico (six counts), whieh form subsection. Serpyllifolia (G. Don) Stroh. Mattick (in 1950) and Peev (1975) report. tetraploid 4x — 29) for V. serpyllifolia from Austria and Bulgaria, respectively, serpyllifolia L. (more than counts; subspecies Tischler, counts (2n = which may be spontaneous autotetraploids, but a voucher specimen for the first report is not indicated and those from Peev have not A count of 2n = 2x = Sokolovskaya (1990) Unfortunately, no figure is presented. been checked. 16 by Probatova and is probably a miscount. The polyploid complex of Veronica gentianoides Vahl, classified as subsection Gentianoides (G. Don) Assejeva, is without a doubt the most complex group karyologically in the genus. Eight different. ploidy levels on two different base numbers have been published, and our report of 2n = 36 presents the ninth ploidy level (with a range of 2x—10x). Tumadja- nov has studied this group extensively (e.g., l'umad- 1977). but partly due to the limited janov & Beridze, 1969: Tumadjanov et al., sull documentation of chromosome counts in questions remain, his studies. Tumadjanov et al. (1977) proposed an intraspecific x=8 predominating in the Turkish and Armenian distribu- 12 in the Georgian Lesser Caucasus. base number swite 2 from x = 810 x = 12, with tion area and x Greater. Caucasus. and However, this distinction is refuted by his own data. The ancestral diploid race (2n — 2x = 16) is found in the northern Colchis, a suggested refugium for tertiary. forest species (Tu- madjanov & Beridze, 1969). Other chromosome numbers based on x = - found in the Greater Caucasus are: 2n = 4x = 32 (throughout the species range; Tumadjanov & E 1969; Tumadjanov et al., 1972), 2n = 5x = 40 (Tumadjanov et al., 1972), 2n = 7x = 50 (Western Caucasus; Magulaev, 1984), 2n — = 04 (Azerbaijan: Tumadjanov & Beridze. 1969 from Central Caucasus; Kliphuis in Love, 1979), and 2n = 80 (Georgian parts of the Lesser Caucasus; Tumadjanov & Beridze, 1969; Tumadjanov el al., 1972). Other the Greater Caucasus are more compatible with a base chromosome numbers of number of x = The related V. schistosa E. Busch (2n = 24; six counts) is either a diploid (x = 12) ora triploid (x = 8). Zakharjeva (in Agapova et al.. 1993 counted 2n = ly) for 48 (tetraploid or hexaploid, respective- this species. but no voucher specimen is Therefore, we cannot be sure about its identity, and it may be a count for the closely related n which this number is V. gentianoides common, 1977), gentianoides 24 (either 2x or 3x), According to Tumadjanov et al. (1972 the predominating the ( polyploid races in V. eater Caucasus have 2n = = 48 (either 4x or Ox), and 2n = 72 (Ox or Ox), the latter two confirmed by us. Our report of 2n = 36 is, however, the first. that is incompatible with a hypothesis of a basic chromosome number of x = 8 in V. gentianoides in the Greater Caucasus and suggests instead x = The predominating polyploid races in the Lesser Caucasus (especially Armenia) have 2n = 32 and 48, although data from Turkey are lacking. Various species have been segregated from V. gentianoides by some authors. (e.g. Kemularia- Nathadze, 1955), but no information on ploidy level for these segregates is available. IH. VERONICA SUBG. PSEUDOLYSIMACHIUM This subgenus includes 26 species, with chromo- some numbers known for 18. In contrast to the case 1 Veronicastrum and Picrorhiza, the derived chromo- ase number x = some the 18 with a 17 likely originates from combination of two genomes with 2n = subsequent reduction to 2n 17 based on its phylogenetic position distant to any taxon with x = (Fig. 2) and karyological observations by Graze (1935). The origin of the base chromosome number x 7 Albach et al. the fact that it is a derived base has been further discussed by (2004a). To highlight chromosome number Pseudolysi- lor Veronica subg. machium alone, it will subsequently be denoted as xp sl (= 2x) Seven of the species in the subgenus are diploid (2n = 2xpy = 34), two are tetraploid (2n = Axpa = 08), and seven include both diploid and tetraploid n The tetraploid level for V. (Peev ı 197: because Ba ly voucher spuria L. 1 Lóve, 2b) needs confirmation information given in the publication and specimens in the herbarium do not match (M. A. numbers of the European species have recently been reviewed by Travnícek et al. (2004) and Albany and (2003). In ) of V. daurica Steven, V. linartifolia Pall. ex Link. V. kiusiana subsp. Nakai & nakatana Ohwi. V. Fischer, pers. obs.). Chromosome E ischer Asia, only diploids (2n = 2 xpy = miyabei ( Hondo) Yamaz., V. pinnata L.. and V. schmidtiana Regel and tetraploids (2n. = 4 xp = 68) of V. ktustana subsp. maritima (Nakai) T. Yamaz., V. ornata Monjuschko, and V. subsessilis (Miq.) are known mostly from single counts. C Carriere — iromosonmes of the species in this subgenus are ex sticky, the chromosome numbers difficult (cf. Weiss et al.. 2xpa1 mE 34 or 2n = remely small and of 2002). 4 XPsl = 68 by one or two chromosomes are consequently more which makes correct determination Deviations from 2n = = the first publication of chromosome numbers in Veronica (Hertz, 1926) already includes a dubious count, 2n = 48 for V. common than in other subgenera. Unfortunately, azurea Link (a synonym of V. longifolia L.). Volume 95, Number 4 2008 Albach et al. 555 Chromosome Numbers in Veroniceae (Plantaginaceae) The figure shows 48 to 50 chromosomes, but clearly not the expected 34 or 68 chromosomes. No voucher information. is included, so this count remains doubtful. Other doubtful records that require further LIÉ grandis Fisch. (2n = 56; Zhukova, 1967, according to Agapova et al., 1993), V. olgensis Kom. (2n — 24; Gurzenkov, 1973, likely to be a printing error for 2n = 34), and V. sajanensis Printz (2n — 18; Malachova, 1971; Krasnoborov, 1976: both according to Krogulevich and Rostovtseva, 1984) and those by Androshchuk (1988). The Androshchuk (1988) appear to have a different investigation are those for V. counts by systematic error because nine species have been investigated and all counts are based on x — 9 and as expected for subgenus Pseudolysi- 90 for V. from Yakutia by Krogulevich and Rostovtseva Future European species of this subgenus needs to focus on nol on x — 17 longifolia 1984) research on the machium. A count of 2n = ca. also appears dubious. the origin of tetraploids, whereas more fundamental information on chromosome numbers and morphology is necessary for the Asian species of the subgenus. IV. VERONICA SUBG. SYNTHYRIS Chromosome numbers are known for all species of Veronica subg. Synthyris. They all share the chromo- € T = some base number of x . which is rare Veroniceae with all extant relatives suggested by phylogenetic analyses (e.g.. Albach et al. 2004a; Fig. 2) 9. Lepper (1964), based on karyotype that the represents a diploid level with aneuploid chromosome having x = analysis, suggested number number increase. However, the alternative that the number is derived from an ancestor with 2n = 2x = 14 followed by reduction to n 6 in gametes and perhaps production of educed gametes and subsequent polyploidization s should not be discarded because of the frequent parallel reductions to x = 7 in Veronica (see above). Most species are diploid (2n = 2x = 24), but V. ritteriana (Eastw.) M. M. Mart. Ort. & Albach and V. canbyi (Pennell) M. M. Mart. Ort. & Albach are tetraploid (2n. = 4x = 48), and, in V. Mart. on Albach, V. missurica subsp. stellata (Pennell) M. M. Mart. Ort. & Albach, V. plantaginea E. James, and V. wyomingensis (A. Nelson) M. M. Mart. Ort. & Albach, both diploid and tetraploid plants have been found in £c missurica subsp. major (Hook.) M. M. different. populations. V. VERONICA SUBG. COCHLIDIOSPERMA Veronica subg. Cochlidiosperma is the group with the most counts per species. With the exception of V. sibthorpioides Deb.. Degen & Hervier, all species share the chromosome base number x = 9. Veronica sibthorpioides, a species from southern Spain and Morocco, however, has 2n. = 30 as confirmed by 14 counts from different localities, although Sánchez- Agudo et al. (unpublished data) recently reported two counts of 2n = i 2x = 18): V. crista-galli Steven (three counts), J stewartii Pennell (two counts), V. triloba Opiz (25 A. Fisch. & Greuter (one Lehm. 28. Several species are diploid (2n = stamatiadae M. lycica E. B. J. Tineo (six counts), counts), V. count), V. (one count). V. panormitana and V. trichadena Jord. & Fourr. (seven counts). A count (Peev in Love, 1972a) from Bulgaria of V. hederoides M. A. Fisch. (a synonym of V. stewartii from the Himalayas; 2n = 2x — |8) refers to V. triloba (Fischer, 1984). Counts of 2n = 2x = 18 for V. sublobata M. A. Fisch. and V. hederifolia L. by Nordenstam and Nilsson (1969) and Guo and Liu (2001), respectively, probably also refer to V. The proposed separation. of V. triloba, but the vouchers have not been checked. sublobata from V. hederifolia s. str. (Fischer, 1967), partly based on its lower ploidy level (tetraploid vs. hexaploid), has led to numerous new reports (more than 100 populations counted for each taxon) for these two species, especially in The Netherlands (de Jongh & Kern, 1973; Gadella & Kliphuis, 1976) and northern Europe & Nilsson, 1969; Fischer, 1975b). rather perfect correlation between tetra- (Nordenstam showing a ploidy in V. sublobata and hexaploidy in V. hederifolia separation between these str. and a clear-cut species. The few early deviating counts in northern Europe (Nordenstam & Nilsson, 1969) were due insufficiently careful identification of the investigated specimens. Morphological differentiation between V. sublobata and V. difficult where chromosome counts are still scarce (Sánchez- hederifolia s. str. seems to be more , however, in some southern regions of Europe Agudo et al., unpublished data). Veronica sublobata is a tetraploid (2n = 4x = 36) derivative of V. triloba (Fischer, 1975b; ization between V. Albach, unpublished data). Hybrid- triloba and V. sublobata gave rise to the hexaploid (2n. = 6x = 54) V. hederifolia (Albach, A count of 2n = 56 by Meskova (1965) is probably a miscount, but unfortunately, this — unpublished data). publication was inaccessible to us. Counts of 2n = 26 to 28 (Sorsa, 1963: 1966) refer to V. persica, which often grows side Gadella & Kliphuis, probably by side with V. hederifolia. No voucher is indicated by Sorsa (1963) and the voucher from Gadella and Kliphuis (1966) has not been checked. count (2n = 4x = 36) for V. hederifolia (Saeidi-Mehrvarz & Kharabian, 2005) is of interest sublobata is not known from that region. A tetraploid from Iran because V. This may represent a different taxon and requires further confirmation. A publication from Guo and Liu 556 Annals of the Missouri Botanical Garden (2001) provided a range of chromosome numbers for species, less than five counts per species have been V. hederifolia from its introduced range in Nanjing, — published (V. ciliata, four counts including the report China, including diploids (2n = 2x = 18), tetraploids here; V. densiflora, four counts; V. erinoides, one (2n = 4x = well as more unexpected numbers: 2n 36), and hexaploids (2n = 6x = 54), as — 22, 32. Their drawings in the paper indicate the reported chromo- some numbers. It is not clear how an aneuploid chromosome race could have arisen in the short time since its introduction, and a mix-up of the samples with another weedy species with 2n — 32 during the collection or processing is considered most likely. Veronica cymbalaria Bodard is karyologically more difficult. Tetraploid (2n = 4x = 36) and hexaploid (2n = 6x = 54) plants are morphologically indistinguish- able (Fischer, 1975a), and both ploidy levels originated more than once from V. panormitana and trichadena (Albach, 2007). A map of the origins of 38 known tetraploid and 24 two different clades of V. 7 known hexaploid plants of V. cymbalaria will be published elsewhere. Counts of 2n = 2x 18 by Hofelich (1935) and Nilsson probably refer to V. trichadena or V. panormitana, and Lassen (1971) but a voucher for the first could not be found (Dilger- Endrulat, at TUB, pers. comm.; Albach, pers. obs.) and the other has not been checked. VI. VERONICA SUBG. PELLIDOSPERMA This subgenus is a small group of seven annual species. Chromosome numbers for Veronica donii Dorfl., are not available but would be highly aznavouril and V. samuelssonii hómpp. V. Rech. desirable because, amazingly, three different base numbers (x — 7, 8, 9) have been reported so far in this a subgenus with all taxa reported as diploid. Veronica triphyllos L. has 2n = 2x = 14 (18 counts); V. praecox All. and V. glauca Sibth. & Sm. have 2n = 2x = 18 (19 and five counts, respectively); and V. mazander- anae Wendelbo, a species endemic to Iran, has 2n = 2x = 16 one count). m V I. VERONICA SUBG. STENOCARPON This subgenus includes 31 montane to alpine species of Eurasia and Mexico. Chromosome numbers — are known for all eight European species but only for two Asian species (V. ciliata Fisch., V. densiflora). All chromosome numbers are based on x = 8 and represent diploids (2n = 2x = 16) with the exception of the tetraploid (2n = 4x = 32) V. contandriopouli Quézel from Greece, which is possibly just an aberrant specimen of autotetraploid V. erinoides Boiss. & Spruner. Veronica fruticans Jacq. is the best studied species (31 counts throughout its distribution area in the European high mountain ranges). For all other count; V. fruticulosa L., three counts; V. mampodrensis osa & P. three counts including the report here; V. saturejoides nummularia Gouan, onts., one count; V. thessalica Benth., two counts 16 for V. southeasternmost distribution range confirms. those Vis., three counts: V. Our count of 2n = 2x = ciliata from the from other extremes of the distribution. VHL VERONICA SUBG. POCILLA Veronica subg. Pocilla (Dumort.) M. M. Mart. Ort., Albach & M. A. Fisch. includes 27 annual and one perennial species including the well-known cosmo- politan weed V. persica. The chromosome base number 7. although it is not for the 15 species counted is x = clear for V. cardiocarpa (Kar. € Kir.) Walp., for which the only publication states 2n = 14-16 (Hofelich, 1935), and V. campylopoda Boiss. (see below). Chro- half of the 8 mosome numbers are known for over species, especially the weedy species (V. persica, N 4 counts: V. polita Fr., 38 counts; V. agrestis L., 2 counts; V. filiformis Sm., 12 counts; and V. opaca Fr., 11 counts). Diploids (2n = 2x = 14) include V. polita (sometimes under its synonym V. didyma Ten.), V. cardiocarpa (one count), V. ceratocarpa C. A. Mey. three counts), V. filiformis, V. francispetae M. A. PE Fisch. (two counts), and V. siaretensis E. B. J. Lehm. (one count). Veronica persica, V. agrestis. V. opaca, V. biloba L. counts), and V. rubrifolia Boiss. (two counts) are tetraploids (2n = 4x = 28 cially in Asia, confused V. agrestis with V. polita for a (three counts), V. capillipes Nevski (two ). Taxonomic history, espe- — long time (Lehmann, 1910, 1940), and, consequently, several publications from the same authors in India slate a diploid level for V. agrestis (e.g.. Bir & Sidhu in Love, 1978). Probably all of these publications refer to V. polita, which is common in Asia, rather than V. agrestis, which occurs exclusively in Europe. Á report of 2n — 2x — 18 for V. polita from Hungary (Borhidi, 1968) is probably due to confusion with V. triloba, but no voucher specimen is indicated. Regarding Veronica sect. Subracemosae (Benth.) Assejeva in subgenus Pocilla, our count of 2n = 6x = 42 is the third report confirming the hexaploid level for V. Regel € Schmalh. campylopoda is the only species in this subgenus and arguteserrata Veronica one of four in the genus for which more than two ploidy levels have been reported in the literature. Fischer (1981) reported a tetraploid (2n = 4x = 28) plant from Iran. Nine counts of hexaploid plants (2n = Ox = originate from Central Asia to the Mediterranean, with our count being the first from Turkey. Our report of an Volume 95, Number 4 2008 Albach et al. 557 Chromosome Numbers in Veroniceae (Plantaginaceae) octoploid plant (2n = 8x = 56) is the second for that ploidy level in V. campylopoda, with both notably from Turkey. Reports of 2n = 2x = 18 from Afghanistan (Podlech & Dieterle, 1969) and 2n = 36 from Utah. U.S.A. (Bell, 1965), introduced, appear dubious in light of other reports, but voucher lx = ca. where it is specimens for the first count have been confirmed by the first author. The voucher specimen for the second belongs to V. arguteserrata. However, such morpholog- ical intergradation is typical for well-watered speci- mens of V. campylopoda (Albach, unpublished data). IX. VERONICA SUBG. PENTASEPALAE Subgenus Pentasepalae (Benth.) M. M. Mart. Ort., Albach & M. A. Fisch. is by far the largest subgenus of Veronica in the Northern Hemisphere, with 72 species most common in southwest Asia. Chromosome numbers are available for 32 taxa including all European species. All species investigated have a base chromo- some number of x — 8, including 19 diploids (2n — 2x 16), five tetraploids (2n = 4x = 32), one hexaploid (2n = 6x = 48). and three octoploids (2n = 8x = 64). European and Siberian species are classified. within Veronica subsect. Pentasepalae Benth. and represent several, often ancestral, diploid species (V. crinita Kit. ex Schult., V. krylovti Schischk., one count; V. orbelica (D. Peev) D. Ten., 14 20 counts including the reports here; V. two counts; V. kindlii Adamovic, one count: Peev, one count; V. orsiniana counts: V. prostrata. L., rhodopea (Velen.) Degen ex Stoj. & Stef., one count; V. rosea Desf., 13 counts; V. tenuifolia Asso, 16 counts; V. & Stef., aragonensis Stroh, V. turrilliana Stoj. three counts). Polyploid species include V. scheereri (J. P. Brandt) Holub (2n = 4x respectively), and V. sennenit (Pau) M. M. Mart. Ort. & E. Rico (2n = 8x = widespread species of Veronica subsect. Pentasepalae, = 32; six and 24 counts, 64; six counts). For the three most several ploidy levels have been reported. Veronica feucrium is octoploid (2n = 8x = 64) with one publication reporting diploid plants (2n = 2x = 16) from southern France (Kiipfer, 1969), possibly belong- ing to V. orsiniana, and one hexaploid plant (2n = 6x = 48) from Germany (Lippert & Heubl, 1989). Voucher specimens for the hexaploid plants have been checked and represent V. teucrium rather than the sympatric V. austriaca L., which is mostly hexaploid except for some specimens from southwestern Germany and north- 1952, 1961). austriaca are western Switzerland (Brandt, However, intermediates between V. teucrium and V. known from southwestern Germany; therefore, these populations need further study in a wider context. The third. species of the subsection; which is widespread and karvologically polymorphic, is V. jacquinii Baumg., for which octoploid plants (2n = 8x = 64) are reported 1986b). Other reports for this species are mostly hexaploid (2n = 6x = 48; 13 counts), with a diploid race (2n. = 2x = from Albania (Baltisberger, 1988) and tetraploid (2n = 4x = 10x = 80) reported from Bulgaria (Peev, 1972). Pinnatifid leaves from Greece (Strid in Love, 16) reported 32) and decaploid races (2n have evolved multiple times independently in Veronica (Albach et al., independently in taxa of different ploidy levels, but all 2004c) and may even have evolved are considered here under V. jacquinii. The same is true for Veronica multifida L. of Veronica subsect. Orientales (Wulff) Stroh, a taxon with species mainly found in Turkey. Diploid plants (2n = 2x = 16) of V. multifida have been reported from Bulgaria (Peev in Love, 1972b) and southern Turkey (Fischer, 1970), tetraploids (2n = 4x = 32) from Armenia (MeSkova, 1965) a southern Turkey (Fischer, 1970), hexaploids (2n — — 48) from southwestern Turkey (Fischer, 1970), i decaploids (2n. = 10x = 80) from Armenia 1990). Veronica multifida is a but no morphological (Gukasian & Safarian, highly polymorphic species, correlate to the different ploidy races has been found (Fischer, 1970). In this respect, V. multifida resembles — . orientalis Mill., another morphologically polymorphic species from southwest Asia with several ploidy levels reported. A tetraploid plant (2n. = 4x = 32) has been reporte d from Armenia (Meskova, 1965) and octoploids 1986a) and kurdica x = 64) from Iran (Ghaffari in Love, 2000). Benth. is hexaploid (2n = 6x = 48: one count). Other (2n — = Ox Israel (Pazy, The closely related V. species of the subsection are mostly diploid (2n = 2x = 6; V. bombycina Boiss. & Kotschy, two counts including the report here; V. caespitosa Boiss., two & Balansa, dichrus cinerea Boiss. two counts; V. Don, Schott & Kotschy, three counts; V. farinosa Hausskn.. one count; counts; V. cuneifolia D. one count; V. V. macrostachya Vahl subsp. sorgerae M. A. Fisch., one count; V. pectinata L., one count) with only V. elmaliensis M. A. Fisch. being octoploid (2n = 8x = 64, one count). Chromosome numbers for the ca. 30 species of the five remaining subsections of Veronica subg. Penta- sepalae are lacking except for two, the diploid taxon V. peduncularis M. Bieb. (2n = 2x 16, three counts) and the tetraploid V. microcarpa Boiss. (2n. = 4x = 32, one count). Future research should especially focus on southwest Asia and the Caucasus since those taxa for which no chromosome number is yet available occur in Turkey (18 species), Iran (12 species), and the Greater Caucasus (10 species). X. VERONICA SUBG. CHAMAEDRYS Veronica subg. Chamaedrys is one of the best studied groups in Veronica including one of the best studied species (V. chamaedrys L.. with more than 90 Annals esa a Garden counts). All species have chromosome numbers based 8 except for V. magna, which appears to have x = 7 (see below). Only V. sartoriana Boiss. € Heldr., a Greek endemic closely related to V. arvensis L., has nob been studied. The annual members of this subgenus, V. arvensis (42 counts), V. verna L. (17 counts), V. dillenii Crantz (12 counts), and V. brevistyla Moris (two counts), are all diploid (2n. = 2x = 16). A count of 2n = 14 for V. arvensis was reported by Lóve and Love (1956) from Iceland. Unfortunately, neither a figure of the chromosomes is given nor a voucher specimen indicated. Therefore, we cannot check our assumption of a miscount or a mix-up with V. polita. The European perennial species are also. diploid with the exception. of. V. chamaedrys subsp. chamaedrys, which is tetraploid (Zn = 32). subsp. chamaedrys, However, within V. chamaedrys some diploid taxa are insuffi- ciently 1973: see Dobes & Vitek. 2000 for a map) and several diploid plants cannot be clearly the taxa (Bardy & Albach, unpublished data). Finally. the Caucasian V. known (Fischer. assigned to described magna seems to be hexaploid (2n = 6x = 42) and the Asian V. laxa Benth. has been reported to be either tetraploid (2n = 4x = 32; China) 6x (22) = 46; Japan. India). More information on the or hexaploid (2n = latter would be desirable to investigate whether tetraploids and hexaploids can be differentiated. and whether diploid progenitor taxa still exist within the species. XI. VERONICA SUBG. PSEUDOVERONICA (HEBE COMPLEX) Veronica subg. Pseudoveronica J. B. Armstr., estab- Armstrong (1881). correct name for the Australasian species of Veronica lished by is the nomenclaturally formerly grouped in Hebe Comm. ex Juss. and related genera (Garnock-Jones et al., 2007). Chromosome numbers for species of Parahebe W A R. B. Oliv on New Zealand (Garnock-Jones € Lloyd, 2004), Hebe 20006). and the 2006) we will only Bayly € Kellow, Australian species Dayly | (Briggs & reviewed. Ehrendorfer, have recently been comment on the the Therefore, New The Australasian species constitute a numbers of Guinean species and more general picture. monophyletic group nested within the x = 8 clade (Fig. 2) and apparently derived from a single poly- the the subgenus is denoted as (= Guinea, ploidization event. Thus, derived chromosome XpHebe Detzneria) tubata base number in (= 6x, below). Veronica (Diels) Albach New to the remaining Australasian species in a cladistic 1984) and in some of the DNA-based phylogenetic analyses by Albach et al. (20052). see from which analysis of morphological characters (Hong. has 2n = 48 (Borgmann. is sister n with a hexaploid origin of = 3842 (Briggs & 906). which is more in line with reports 1964). This would fit the Australasian species or 2n Ehrendorfer, 2t for other species in the subgenus with chromosome = 2 numbers derived from xq, = 20 and xq. Wagstalf and Garnock-Jones (1998) 21 is ancestral to xq. = have shown shat Hebe — 20. which evolved twice, once in the main clade of Hebe on New Zealand and once in Australia. Further phylogenetic analyses (Wagstaff et al., 2002: Albach et al.. 2005a) have demonstrated that two further changes lo Xepe = 20 need be assumed for decora (Ashwin) Garn.- Jones in New Zealand and V. A hexaploid origin of the subgenus would tonantha Albach in New Guinea. an inferred. loss of six to 12 therefore require chromosomes in the various species of the subgenus. Our understanding of phylogenetic relationships of which shows much greater the New is unfortunately too limited to allow further Veroniceae Australia. karyological diversity than species ii Zealand, inferences. The phylogenetic relationships of those taxa for which DNA sequences are available, inferred y Waestall (2002). numbers are 1 Figure 2. The Parahebe in New Guinea likely originate from a single from New Zealand (Albach et a). chromosome numbers are available from a 1904) for two species N el al. and their chromosome 12 species shown immigration of a species al., 2005 single publication (Borgmann. (= P. ciliata (Pennell) P. n (V. ionantha Royen Ehrend.): 2n = 2x1. = 40: albiflora: 2n = 2X Hebe = 12). XIL VERONICA SUBG. TRIANGULICAPSULA This taxon includes two morphologically enig- difficult to place using 2004a). is. discernible matic annuals. Thev are molec ular data (Albach et al., Even karyo- no relationship because both species, the diploid Veronica grisebachii Wal- Turkey the pithyoides Lam. from Spain, have the unusual chro- lers from and tetraploid V. chamae- mosome base number of x = 6. THE REMAINING SPECIES IN VERONICA Chromosome numbers are known for one species that has not been assigned to any subgenus. Veronica on = 16, Chamaedrys with which it shares some morphological javanica Blume is similar to subgenus especially seed ultrastructure (Muñoz- 2000). plastid characters, However, neither nuclear DNA PE ) ria] with the x = 8 Centeno el al., ribosomal nor sequences show a (Albach et al., 2005a: Fig. 2), which suggests that it has gained this clade ine Sume independently. Volume 95, Number 4 2008 Albach et al. Chromosome Numbers in Veroniceae (Plantaginaceae) Chromosome numbers of the other species in Veronica that to date have not been placed confidently Don, V. simensis Fresen., V. in any subgenus (V. himalensis D. monticola ruprechtii Lipsky. V. tibetica D. Y. highly desirable. Trautv.. V. Hong, and V. viscosa Boiss.) would þe NOMENCLATURAL CHANGES FOR VERONICA For the purpose of providing a taxonomically ordered overview of chromosome numbers in Veroni- ceae, we provide a list (Appendix 1) resolved down to the subsectional level in Veronica including all species currently accepted by us. Three new names at the supraspecific level are used here for the first time, and these names is given here. The description of a new subsection for V. anagallis-aquatica and relatives will be published in the future. Veronica sect. Acinifolia (Römpp) Albach, stat. nov. Acinifolia Rómpp Repert. Spec. Nov. Regni Veg. Beith. 50: 60. 1928 [Veronica Ver- wandtschafisgruppe|. TYPE: Veronica acinifolta I Sp. Pl., ed. 2, 1: 19. 1762 The clade is equivalent at rank to Veronica sect. Beccabunga and Veronica sect. Serpyllifolia but has not been used at the sectional level vet. Veronica sect. Glandulosae (Römpp) Albach. stat. nov. Glandulosae Rómpp. in Repert. Spec. Nov. Veg. Beih. 50: 33. 1928 Verwandtschafisgruppe |. VY PE: ex Benth., Regni | Veronica Veronica glandu- (DC.) 10: 48: losa Hochst. Prodr. 846. The same rationale as above applies in this case. Rehb.) Al- Cochlidiosperma Rehb., FL 1831-1832 [Veronica (un- ranked infragenus)]. TYPE (designated by Pou- zar, 1964): Veronica hederifolia L.. Sp. Pl. 1753. pa Veronica subsect. Cochlidiosperma bach, stat. nov. Germ. Excurs.: 365. Veronica hederifolia has traditionally been placed in subsections. whose type now belongs in subgenus Pocilla. Veronica sect. Cochlidiosperma from the white-flow- To distinguish the blue-flowering species of ering species (Veronica subsect. Cymbalariae). a new subsection needs to be established. Literature. Cited N. D., K. Zemskova & L. V. Tar \rkhar ova, L. L Vakhtina. . 1993, Numeri c ee S: Moraceae—Zygophylla- ceae. Russian Academy of Science, St. Petersburg Agapova. a Flora URS the correct. nomenclatural publication of Albach. D. C. 2007. The use of AFLP the phylogenetic lieet polyploids: Multiple origins of and sequence data i Veronica cymbalaria (Plantaginaceae). New Phytol. 176: 1406 48 8. & M. W. Cha 2001. Paraphyly of Veronica (Veroniceae; Se phar eae): Evidence from the inter- nal transcribed space (ITS) sequenc of nucle ribosomal DNA. J. PL Res. 114: 9-18. ———— & M. AFLP- analyses: Gontebution lo he taxonomy of Fischer. 2 and genome size ` Veronica subg. Pseudolysimachium sect. Pseudolysimachion (Plantagina- ceae), with a key to the European taxa. Phytol. Balcan. € 401—424. . Chase. 2004. Incongruence in Veroniceae ism ut eae): Evidence from two plastid and a nue lear A region. Molec. Phylogen. Evo ribosomal D 183-197. & J. Greilhuber. melee Genome size variation and evolution in Veronica. Ann. Bot. 94: a l. Martínez- s M. Evolution. of Missouri Bot. Gard. € & —— 2004b. A new v M. W. . Fisc ue rá 2004a. perspective. Ann, Chase. eroniceae: 2 logenetic —302 classification of the 1 ribe Veroniceae—Problems and a possible solution. s 53: 429—452. 3 & V . Chase. 2004c. Veronica: Parallel morphological e lation and phylogeography in the Me dite 'rranean. Pl. Evol. 246: 177-194. ui uh, d J. Wagstaff. 20052. Tenna Cu A formerly Scrophulariaceae) on New Games, Syst. Bot. 30: 412—423. S Özgökce & R. Veronica: C AM al characters for th« Origin of 2005 > support of phyloge- Jensen, F. Grayer. netic relationships based on nuclear os 7 T DNA s Piae nce data. Biochem. Syst. Ecol. 33: 1087-1100. n C VIROS Schóns boue & A. Tribsch. phyloge slo the Verts ere -complex in Europe and North ee: rica. ero . Ecol. 15: 3269-3286 l eM re pe cies of n genus a (Scrophulariaceae) in the nei numb ers in some aine. Bot. Zhurn. 73: 454-455. 2003. p classification for the orders "ER An Phylogeny Group. An update of Angiosperm Phylogeny Grou and ois Towering plats APG II. Bot. J. Linn. Soc. 141: 399-430. Non J. B. 1881. A synopsis of the New Zealand species of Veronica, i Linn., with Ha on new species. Trans. New Zealand Inst. Technol. 344—359. Baltisberger, M. 1988. Chromosome ders i einiger Pflanzen ws . M. Ber. ETH Stiftung Rúbel os Geobot. Inst. a Bavly. . new Zealand. Bell, C. R. 1965. Documented plant chromosome numbers. Sida 2: 168-1 i. . Ke A 2006. New E: 2 Hebes, an S i s Te Papa Pr ellington, New : . Nuclear DNA conte H and minimum generation time bs DRE o roc. Roy. So Lond.. Ser. B.. Biol. . 181: 109-13 Bócher. T. W. 1944. P. ze of Veronica officinalis in relation to genetic and environmental factors. Dansk Bot. 1-20. 1e leaf si Ark. 11: 1-2 Borgmann, E. 1964. Anteil der Polyploiden in ps Flora des Dismar¢ kee rbi ges von Ostneuguinea LL Bot. : : 118-1 72. Borhidi, A. European plant species, L Acta Bot. Acad. Sci. 14: 253-260. t studies on qeu Hung. 560 Annals of the Missouri Botanical Garden Borissova, A. G. d Veronica. Shishkin & E. Pp. 373—557 in B. A . Bobrov odia Flora SSSR. Vol. Akademii N ca SSSR. Mos P Leningrad. Eco J., S. H. Strauss, A. H. Doerksen & R. A. Price 1992. Extensive variation in e era in of rbcL gene sequences among seed plants. Proc. Nath Acad. Sei U.S. id 89: Brandt, J.-P. Veronica. 7844-7848. 952. Contribution à la ope du genre dr Soc. Neucháteloise Sci. Nat. 61. € ytotaxinomie et eytogéographie de Veronica mom E s . Neucháteloise Sci. Nat. 84: 35-88, Briggs, B. G. N F. E e. ndorfer. 2006. Chromosome numbers of Australian and New > rOnica lantaginac cae E | lopea 11: 294-298. ‘ & L. R. Heckard. of some North nun rican Scrophulariaceae, Madroño 39: 137-149, Davlianidze, M. po species of Ve D -J Chuang, 1992. Chromosome numbers mostly Cali- lornian. 1980. Numeri chromosomatum nonnularum Zametki Sist. 30: e) E z plantarum caucasicarum. Geogr. 15-7 ). ux J..-H. Veronica teda alia L. N Kern. 1973. De variabiliteit van in Nederland. Gorteria 6: 202-203. 2000. number checklist of Austrian vascular plants. Verlag des de Jongh, 5 Dobeš, C. & E. Vitek. Documented) chromosome Naturhistorischen Museums Wien, Wien. Dzhus. M. A. & S. A. Dmitrieva. 200 o čisla vidov rot Veronica (Se n ie eae Belorussii res > l: (Chromosome mbers the species the genus P Veronica (Ser solidos eae) from Byelorussia). Bot. Zhurn. 86: 144-1: " Ferrarella, A., F. Grisafi, F. Lentini & M. R. Melati. 1981. Numeri eromosomici per la flora italiana. Inform. Bot. Ital. 13: 189-194. Fischer, M. A. 1967. Beiträge zur Cytotaxonomie der Ne dis À Veronica hen -Gruppe (Sero iceae). Osterr, Bot. Z. 114 22 | 169. E Chromosomenzahlen aus den Gattun- gen Veronica, Pseudolysimachion, Paederota. W Es und Lagotis (Serophulariaceae-V eronicinae). Österr. Bot. Z. 116: 430—143. 1970. Zur Cytotaxonomie der Verwandtschalts- gruppe um Veronica orientali er emend. Ait. in der Türkei. Osterr. Bot. Z. at 3l- . 1972. Neue i iud und System- atik von Veronica subsect. Acinifolía (Rómpp) Stroh. Österr. Bot. Z. 121: 413-437 1973. Notizen zur Systematik hl und Verbreitung. einiger Veronica-Sippen in Kärnten. PAGINA 83: 379 i. Untersuc i n | übe T dis n xu O x posce (5 97-105 — Veronica agg. l . oyst. Evol. 123: 97-105. EA The Veronica hederifolia group: Taxonomy, 48-00 in S. M. Walter Studies. Bol. e :colog V, dI nd phyloge ny. Pp. < (editor), E ‘rope an Floristic and Taxonomic Soc. Brit. ls. Conf. Rep. 15. 19 m Veronica L. Pp. 689-753 in P. H. Davis (editor), Flora of Turkey, Vol. 6 University Press, Edinburgh. 1981. Veronica. Pp. 52-165 in K. H. Rechinger (editor, F lora pe a, Vol. 147. Akademische Druck- und a alt, 84. de "c *horologíe und Systematik der dnd m Gruppe in Jugoslawien. Akad. | Nauka i iU a Bo 1 Hercegov. Radovi 76, Odjelj. Prir. i Mat 55-77 Veljović € B. Tatić. 1984. Veronica scardica—A uds species of pi Se LE Flora. Glas. Inst. Bot. Baste Univ. Be coer, 18: : Fukui. K. & dtum. Laboratory Methods. CRC W. 1996. Plant Chromosomes. Boca Raton. Flori Press, Florida. md T. J. & E. Kliphuis. 1963. Chromosome numbers ‘flowering plants in the Netherlands. Acta Bot. Neerl. 12: 1 iden & ———— 96. Chromosome numbers of flowering Pr Ned. Akad. JO, Kon. plants in pa Nee IL Wetensch: 1 i —556. p 6. Some critic remarks on Han hederifolia L. in the Netherlands. Proc. Kon. Ned. Wetenseh. C. 79: 61-73. Garnock- n , P.J. & D.G. Lloyd. 2004. A taxonomic revision o Bi arahebe (Scro - Meroe eae) in New sn. New Ze and J. : 181- .D. Mbach & B. ut Southern p »misphere Veronica. (P lantaginace Cue He be. hular lateat Briggs. i rà names in Sect. Derwentia, sect. Detzneria, and sect, Faxon 56: 511—582. Goldblatt. P. & Chromosome Missouri ae — Index to Plant Monogr. Syst. Bot. E. p 2000. e 2001-2003. Gard. 106. , . Weitere: Ghromosomenuntersuchungen bei UR NE der Sekti Pseudolysimachia Koch. Jahrb. Wiss. Bot. 81: Eaten. Gukasian. A. G. & A. B. numbers of some representatives of the Armenian Flora. Biol. Zhurn. dat anil 43: 259-260. Guo, S. & XZ. Liu. 2001. The chromosome number o Safarian. 1990. Chromosome Veronica hederaefolia L. : China and its ecological significance. Guihaia 21: 111-112. Gurzenkov, 1973. ie s of chromosome numbers of plants from the south of the Soviet Far East. Komarovskie Ctenij q 47-62. . & A. 1998. Phylogenetic extinction Sane me o Proe. Roy. Soc. London. Ser. B. Biol. Sei. : 1691-1696 aa der Rambaut. C EE Ver- gleichende Studien a ihre Zahl. Größe und Form im Pflanzenreich l Z. Bot. 18: 625-081. 5. Die n ktion Alsinebe Griseb. der Gattung UNDER in dues Grundlagen. Jahrb. iss. Bot. 81: 541—572. Hong. D. P M. A. Fischer. 1998. Veronica. Pp. 65-80 in ^.-Y. P. We u TEN Flora of China, Vol. 18. Science NM 2 ‘jing, and Missouri Botanical Garden Press. St. Louis. Hong, D.-Y. 1984. Taxonomy and evolution of the Veroni- ceae (Se places eae) with special reference to palynol- Huang, R.. S. Shen P X. Lu. 1996. Studies of the chromosome numbers and polyploidy for some plant the north-east AU pu Xizang plateau. Acta Bot. Boreal.- Occid. est 310—318 Huber, A. 19 ug träge zur Klärung verwandtschaftlicher Bezie nias `n in der Kernunter- Jahrb. Wiss. Bot. Gattung Veronica. l. Die suchungen in der Gattung Veronica. 359-380. Hultén, E. 1930. Flore islands. IV Dicotyledonae Pyrolaceae—C amos Kungl. Svenska B. 2 of Kamtchatka and the adjacent ele o landl.. ser. 3, 8: no. 2. Kemularia-Nathadze, L. 1955. Generum Se puri dao Leal Veronicae L. 48-57 E species novae. Zametki Sist. Geogr. Rast. 18: Volume 95, Number 4 2008 Albach et al. Chromosome Numbers in Veroniceae (Plantaginaceae) Khatoon, S. € S. L 93. Chromosome Atlas of the Angiosperms of Paltistan. pes ent of Botany, Univer- Karachi. Pakistan. Kliphuis, E. & J. H. Wieffering. 1972. Chromosome numbers of some angiosperms from the south of France. Acta Bot. Neerl. 21 Knaben, G. & T. Engelskjon. 1967. Chromosome numbers of Se E arctic-alpine plant species. I. Acta Boreal., A. 2]: 1-57 Krogulevich, n E. € T. S. 1984. somnye čisla cvetkovych rastenij Sibiri 1 Dalnego Vostoka. Akad. Nauk SSSR € Novosibirs Küpfer, P. 1968. Nouvelles prospections caryologique s dans a flore orophile des Pyrénées et de la Sierra Nevada. Bull. Soc. Neuc Er loise Sel Nat. 91: 87-1 1969. Recherches cyto dicun NS sur la flore des Soc. Neuchá- sity of Karachi, Rostovtseva. Chromo- un ^ Sibirskoe, montagnes de la peninsule ibérique. Bull. > Sci. Nat. 92: 31-48. 972. Cytotaxonomie el cytogéographie de quelques groupes d’ orophyte s du bassin occidental de la Méditen Tra- teloisc née et des Alpes. Compt. Rend. Hebd. Séances Acad. 5 275: 1753-1756 ———— 1974. Recherches sur les liens de parente entre la flore e des Alpes et celle des Pyrenées. Boissiera 23: 12315. La Cor L; F. Lab. Prac stique 3: < ed E. 1910. Über Merkmalseinheiten in der Veronica- Sektion Aliinalie: Z. Bot. 2: 577-602. 940. Polyploidie und geographische Verbreitung der h 89: 461—542. 1954. Smear and squash techniques in plant 26—330. Asien der Gattung Veronica. Jahrb. Wiss. Bot. Lepper, L. 1964. Die Cytotaxonomie der Gattung Wulfenia Jacq.—ein Beles Kleistoploidieproblem. | Ph.D thesis, Pun Schiller-Universitát, Jena. The Role of Chromosomal Change in New Yor zum Levin, Plant Edi d University Lippert, W. & G. R. Heubl. 1989. Chromosomenzahlen von Pflanzen aus Bayern und angrenzenden Gebieten. Ber. . Bot. Ges. 60: ist, "B. & U.-M. Huligárd. 1999. Chromosome numbers in South Swedish vascular plants. Opera Bot. 137: 1—42. 1956. Cytotaxonomic al conspectus of the Ice xe ~ Acta Horti Gothob. 20: 65-291. .] e ^B Chromosome number reports XXXVI. Taxon a d prn c hromosome number reports XXXVII. Taxon 21: 495-500 78. IOPB Chromosome number reports LX. Taxon 27: 223-231. —————. 1979. IOPB Chromosome number reports LXIV. Taxon 28: 391—408. ——. 1981. IOP : Chromosome number reports LXXIII. Toron 30: 829-86 19 82. m senda number reports LXXVI. 8. Press, . Lóve. Taxon 31: 574—5 — ———. 1985. end number reports LXXXVI. Taxon 34: 159-164. . 1986a. Chromosome number reports XCI. Taxon 35: 4044 10. ———. 1986b. Chromosome number reports XCII. Taxon 35: 897- 903 Magulac A. Y. 1984. Cytotaxonomic study in Qu plants of the North Caucasus. Bot. Zhur. 69: 511-517. Marchant, N. G. 197 section O Griseb. Ph.D. thesis, University of Cambridge, Cambridge. . Experimental Taxonomy of Veronica D i Martínez-Ortega, M. M., J. A Ir . Sánchez-Agudo & E. Rico. Verc mica. In C. ic & A. Herrero 13. Real Jardín Botánico de o. J. Güemes Madrid, CSIC, Meškova, L. Z. 196 nija k vM d roda Veronica L. Kazan. Muñoz- Ce enteno, d M., D. C. Albach, J. M. M. Martinez- jr 2006. seed morphology in Veronica (Plantaginaceae). genetic i ap Ann. Bot. (London) 98: 335—35t Madrid Ga press). 5. Primenenie kariologiceskogo isledova- Avtoref. Kand. Diss. S A. Sánchez- -Agudo & Systematic d , L. Delgado-Sánchez . Santos-Vicente & M Toont of Veronica L. subsect Martínez- e 2007. À he western Mediterranean. Veronica (P iptin dried in t ot. J. Linn. Soc. 155: 65-81. Nilsson, Ó. & 1971. vascular plants from Austria, Mallorca and Yugoslavia. Bot. Not. 124: 270-27 Nordenst: B. & O. Nilsson. 1969. Taxonomy P. Econ Chromosome numbers of and di: buon of Veronica hederifolia s. lat. (Scrophular- iaceae) in Sc ue Bot. Not. a 233-247 Öztürk, A. & M. A. Fischer. 1982. Karyosystematics of Veronica a sect. us. (Se e lace id with special reference to the taxa in Turkey. Pl. Syst. Evol. 140: 307—319. Pazy, B. | Flora Chromosome numbers. ee 7T- us oa in me new laxa and chromosome numbers. Phytology 2: 42—46 Peev, v taxa and ploidy leve s of some puri Veronica a species. Compt. Rend. Acad. Bulg. Sci. 811-81: Podlech, q 1986. G hromosome nstudien an Pflanzen des Saharo-Sindischen Trockengebietes. Mitt. Bot. Staats- samml. München 22: 5-20. & A. Dieterle. 1969. Cl tudien an afghanischen Pflanzen. Candollea 24: 185-243. Pouzar, Z. o Nomenclatural remarks on some generic Filip Max- Tei e validly aii d by 1990. families P. Sokoloy saya, Chromosome A ios in some representatives of the Ascle- Porciones eae, )nagraceae, dedos eae, Asteraceae, Chenopodiaceae, Lamiaceae, Oleaceae, Scrophulariaceae, Solanaceae, Urticaceae from the Soviet Far East. Bot. Zhurn. 75: 161 p 162: , E. G. Ruda & A. P. Sokolovskaya. 1990. ee mm in synanthropic plants from the Russian Bot. Zhurn. 81: 98-101. cid H. 1928. Die Verwandtschaftsverhültnisse in der attung Veronica. Repert. Spec. Nov. Regni Veg. Beih. 50: 17 e n d í Roseto. M., D. Ottonello & Fici. 1983. Numeri cromosomici per la Flora ium 993- d Inform. Bot. Ital. 15: 188-194. Saeidi-Mehrvarz, S. & A. Kharabian. 2005. Chromosome unts of some Veronica (Scrophulariaceae) species from Iran. Tur is sh a Bot. 29: 263-207. Skalinska, 964. Additions to chromosome numbers of Polish oce (fifth contribution). Acta Soc. Bot. ue = 45- Sorsa, V. 1963. Chromosomenzahlen finnischer pA ten. II. nia Acad. Sci. Fenn., Ser. A, IV, Biol. 68: 1-14. d N. V. P 394. Chromosome numbers ii some nemoral species of the west Sayan (Krasnoyarsk region). - Zhum. 79: 125- 562 Annals o Missouri Botanical Garden Tisehler, G. 1950. Die € hromosome nzahlen der Gefáss- pflanzen Mittele ae Dr. W. Junk, s-Gravenhage. Travnicek, B., M. A. Lysák, J. C halkoi & J. Dolezel. 2004. enus Pseudolysimachion (Serophulariace a in the Czec Ñ 2 pore and Slovakia vlotax. iu) 73-2 5 R. K. Berk Ia UM On the speciation Karyo-taxonomic o. of the Folia Geobot. ere da l. : ithin the series ge mnie Boriss. of the genus Veronic L. Bot. Zhurn. 54: 1660. N é i Pogosvan. 1972. Experience in analysis of MALUM gentianoides Vahl s. L populations along the profile of the Zhurn. 57: 1495-1515. Minor Caucasus mountains. Bot. — — —,. 1977. Populational structure, Videns netic re oli and origin of polyploid complex Veronica ge "ntianoldes Vahl ager. (Scrophulariaceae). Bot. Zhurn. 62: 1548-15 Vasudevan, K. N. 197 and c yloge ography a he Flora of the . Contribution to the cytolaxonomy Western Himalayas (with an attempt to compare il with the F m of the Alps). Part IL. "p Schweiz. Bot. 85: 210-25 Wagstaff, S. J. & P. J. ea 1229; | E solution and uralensis Schischk., L. yesoensis Tatew.. V. wardii W. W. Sm.. L. Sm L. yunnanensis W. Wulfenia Jacq. =í x + baldacii Dege n, 2x = 18 le blecicti La 16, 2x = 18 W. id e 2x = 16 W. orientalis Boiss.. 2x = 18 ds ls B of the Hebe complex ophulariaceae) inferre = from ITS sequences. New Ze dm J. Bot. : 125— —; ei J. Bayly, P. J. Garnock-Jones & D. C. Albach. 2002. Classification, origin, and diversification of the New Zealand pi CALAIS iceae). Ann. Missouri Bot. Weiss. H., B. pe Sun, T. F. St W akabayashi to lung . Kim, H. Kato & M 2002. ol: de do species endemic (Korea) ionis Korea and Japan. Bot. J. 93-105. relatives in 138: Island and selected Linn. Soe. APPENDIX l. List. of all species in the Veroniceae for numbers which chromosome are known. C ru numbers are o as 2n values m refer the d level. Subscri iu (x YHehe) refer to de fact lows classification of the tribe as outlined in Mbac h et al. (2004b) and Garnock-Jones et al. (2007). Lagotis Gaertn. x= ll L. brevituba Maxim., ?x = 54 L. cashmeriana dn | finge, 2x — 22 L. glauca Gaert = 22 L. integrifolia (Y iei ) Se pu rs lx = (43) 44 , minor (Willd.) Standl., 29, 44. E si ra (K. Koc 2 d 2x — 22 „n takedana a & Tatew., 2x = 22 Species not necio Lagotis alutacea W. b Sm., L: angustibracteata b B H. E. A rae d . Mill. a dni "bs m gregorjevt Krassn., . C. Tsoong & H. P. Yang. 2 aede W. W. Sm., L. kongboensis T. tegel € Sehmalh.) Maxim., L. z nth.) Rupr. nepale NSIS A poe mbica R. ssifolia Prain, stachya Maxim., L . uh Hook. globosa E a hultenii a L. humilis ikonnokovii hischk., Yamaz.. E e Ú es (Rovle ex ala Batalin. L. spectabilis. (Kurz) Hook. f.. P 1. oum I. — 36 P. lutea Scop., hi 6x — ns 54 P. Xchurchillii Huter, 4x = 36 oe D. Y. Hong r= W. A herstiana (Benth.) D. Y. Hong, 2x, 4x 16, 32 W. nepalensis (T. Yamaz.) D. Y. Hong, 2x = 16 V i Heist. ex Fabr x=17 (8 + 9; see Albach & C hase, 2004) V. ru MUS (Benth.) D. Y. Hong, 2vy. = 34 V. japonicum (Nakai) T. Yamaz., 2xy4 = 34 V. liukiuense (Ohwi) T. Yamaz., 8xy = 136 V sibiricum (L.) Pennell, 2x44 = 34 . villosulum (Miq.) T. Yamaz., V. virginicum Farw., 2xy4 = 3 Axva = 08 | Species nol investigated: Veronicastrum axillare (Siebold 8 Luce amaz.. V. caulopterum ix ance) T. Yamaz., formosa (Masam.) T. Yamaz.. kitamurae (Ohwi) T. Yam: ” latifolium (He msl.) T. Pid naz., V. End amaz., ees (Hand.-Mazz.) P. C. l'soong, V. robustum (Diels) D. Y. Hong, V. Men. o o (Hemsl.) T. Yamaz.. V. dee (Ohwi) T. Yamaz.. tubiflorum (Fisch. & C. n.) T. Yamaz. A. Mey.) H. Hara, V. vunnanense (W. : Picrorhiza Royle ex Ben : see Albac i s Chase, 2004) P. kurrooa Royle ex Benth., Axpi = 34 x=17(8+9 Species not investigated: Picrorhiza scrophulartiflora Pen- nell, Neopicrorhiza minima R. R. Mill. Veronica l.. I. oe ca subgen. Veronica sect. Glandulosae (Rómpp) Albach la. nee subseet. Glandulosae (Rómpp) Stroh x=8(?),9 : V. abyssinica Fresen.. 6x = 48 . Veronice V. glandulosa subred mannii (Hook. f.) Elenevsky, 6x = »l Y gunae Schweinf. ex Engl., 6x = 54 2. M ind sect. Scutellatae G. Don . Veronica subsect. Seutellatae Benth. x= o scutellata L., 2x = 18 — 3. Veronica sect. Montanae (Boriss. xo Elenevsky) Assejeva Volume 95, Number 4 2008 Albach et al. 563 Chromosome Numbers in Veroniceae (Plantaginaceae) 3a. Veronica subsect. Canae (T. Yamaz.) Elenevsky ? S RAS c. ; cana Wall., ?x= 50, 52 henryi T. miqueliana Nakai, 6x = 48 © robusta (Prain) T. Yamaz., 6x = 42 Species not investigated: Veronica chayuensis D. : Hong, V. deltigera Wall. ex Benth., V. fargesii Franch., P igo Diels, V. japonensts Makino, V laxissima D. ae ong, V. longipetiolata D. Y. Hong, watae T. Yamaz. V. riae H. J. P. oligosperma E i V. piroliformis Franch. V. Win V. secat ica Batalin, V. laiwanica T. Yamaz., a D. Y. Hon Hong, V. vande llioides Maxim.. Vy Hong. ic huensis. Ta = g, V. tsinglingensis yunnanensis D. Veronica subsect. Montanae Boriss. ex Elenevsky T F montana L., 2x, 4x = 18, 36 4. Veronica sect. Veronic 4a. Veronica subsect. dE Benth. 9 V. alpi nina L., = 18 iude L. 2x. 4x — 18, 36 . copelandii Eastw., 2x = 2x, 8x = 18, 72 . nipponica MO 2x — 18 . cusickii A. Gr . ex Link, 2 La pupil Roem. & Se a 4x = 36 PEPPER 4b. Veronica subsect. Gouani (Rómpp) Stroh x=9 V. ponae Gouan, 6x = 54 4c. Veronica subsect. Urticifoliae Boriss. ex Ele- nevsky x= V. urticifolia Jacq., 2x = 18 4d. Veronica subsect. Veronica V. allionii Vill., 2x = 18 $ pee 36 Sav., 4x 6 V. Xtournefortii (Vill.) F. a Schmidt, 3x = 27 (— V. allionii X V. officinalis) dabneyi Hochst., V. not investigated: Veronica 0 morasonicala Hayata. 4e. Veronica subsect. Aphyllae (Rómpp) Stroh x= V. aphylla L., u^ V. e drin — 48-50 Af. Veronica subsect. Carpathicae Elenevsky V. baumgartenii Roem. & Schult., 2x — 14 . Veronica subgen. Beccabunga (Hill) M. M. Mart. "n Albach & M. A. Fisch. Incertae sedis V. hispidula Boiss. € Huet., 2x, 4x = 18, 36 l. Veronica sect. Beccabunga (Hill) Dumort. la. Veronica subsect. Beccabunga (Hill) Elenevsky ET V. americana Schwein. ex Benth., 4x = 36 V. bes L. subsp. beccabunga, 2x, 4x = 2: 36 V. beccabunga subsp. abscondita M. A. Fisch., 2x — 2 V. beccabunga subsp. muscosa (Korsh.) E ond = 18 Ib. Veronica subsect. ined. x=9 V. anagallis-aquatica L., = 36 V. anagalloides Guss. o anagalloides, 2x, 4x = 18, V. anagalloides subsp. heureka M. A. Fisch., 2x, 4x — 18, V. catenata Pennell, 4x — 36 V. lysimachioides Boiss., 2x = 18 V. michauxii Lam., V. oxycarpa Boiss., 2x, m = - 18 36 V. p Murb., 4x = 30 V. V. spec. in 4x = 3 V. undulata Wall.. V. Xmyriantha m l'anaka, | “be TUE dun 54 ^ anagallis-aquatica X V. ire eT = 36 Species not investigated: Veronica kaiseri Tückh. le. Veronica subsect. Peregrinae Elenevsky 9 — 2) 52 Veronica peregrina L., Ox . Veronica sect. Acinifolia (Rómpp) Albach Acinifolia (Rómpp) Stroh = EN Veronica subsect. Y. V. acinifolia L., 2x V. syriaca Roem. & Sc heli = 14 Species not investigated: Veronica balansae Stroh, V. debilis Freyn, V. oetaea Gustavsson, V. yildirimlii Óztürk. 3: Veronica secl. Serpyllifolia G. Don a. Veronica subsect. ue (G. Don) Assejeva x=8(? V. gentianoides Vahl, 2-10x = 16, 24, 32, 36, 40, 48, 56 64, 72, 80 V. schistosa E. Busch, : Species nol investigated: pue P Öztürk & M. A. Fisch. 3b. Veronica subsect. Serpyllifolia (G. Don) Stroh 7 nevadensis (Pau) Pau var. nevadensis, 2x 14 /. nevadensis var. langei (Lacaita) M. M. Mart. Ort. & E. BR 2x — 14 V. serpyllifolia L. subsp. serpyllifolia, 2x. 4x = V. Lir subsp. repens (Clarion ex DC.) m in Hegi, 564 Annals of the Missouri Botanical Garden Species not investigated: Veronica archboldii Pennell, V. ^, porphyriana Pavlov, 2 xp. (flow cytometry) platycarpa Pennell. | V. spicata L. subsp. spicata, 2 xpg. 4 xp, = 34, 68 V. spicata subsp. euxina (Turrill) Stoj. & Stef., 2 x V. \ 3c. Veronica subsect. Telephiifolia M. M. Mart. Ort. & E. me subsp. fischeri Trávn., 4 xp, = 68 Rico . spicata subsp. lanisepala (Travn.) Albach, 2 xp = 34 mr V. telephiifolia Vahl, 2x = 14 2f. Veronica subsect. Spuriae (Holub) Elenevsky Xp = 17 (9 +9-— l) D cies not inve nene d: Veronica daranica Saeidi & Ghahr., V. linarüfolia Pall. ex Link, 7 Pads (34 . davisii M. A. Fisch. V. rotunda Nakai, 2 xp. = V. spuria L., 2 xpp 4 xp = an 68 III. Veronica subg. Pseudolysimachium (Opiz) Buchenau — T l IV. Veronica E Synthyris (Benth.) M. M. Mart. Ort., l. Veronica sect. Schmidtianae (Boriss. ex. T. Yamaz.) Albach b M. A Fisch: ad . Veronica see Schmidtiana Boriss. ex Elenevsky , Ez M. M. M; art. Ort. & ap 2x ET 24. PIC idu SO a besseya M. M. Mart. Ort. & Albach, 2x = 24 P los m V. bullii (Eaton) M. M. Mart. Ort. & "e h, 2x = 24 V. schmidtiana Regel, 2x = 34 V. californica M. M. s Ort. € Albach, 2x = 24 V. canbyt (Pennell) t . Mart. a i Albach, 4x = 48 Pd sect. Pseudolysimachium W. D. J. Koch V. dissecta (Rydb.) N w Mart. Or Albach. 2x = 24 . Veronica subsect. Alatavicae Boriss. ex Elenevsky V. idahoensis M. M. M. Ort. & (du h Oe = 4 f p l V. missurica Raf. : e missurica, B = và Species not investigated: Veronica alatavica Popov, V. jui subsp. major (Hook.) M Mart. Ort. & qingheensis Y. Z. Zhao. Hm h. 2x. 4x = 24. 48 | f . . e MISSUFICA nn stellata (Pennell) M. M. Mart. Ort. & 2b. Veronica subsect. Dahuricae (Holub) Elenevsky ma dle 24. Ag n= ITZ ( . Iii 3 * 9 | 34 > V. Wine (Pennell) M. M. Mart. Ort, & Albach, 2x — 24 A Apai = (printing error?) ec a nnell € L. O. Williams) M. M. Mart. Ort. & . olgensis Kom., 2 xp4 — 24 Albach. 2x = 24 V. plantaginea E. James, 2x, 4x V. ranunc pu ennell) M. M. Mart. n É mes h, 2x = 24. V. regina-nivalis M. M. Mart. Ort. & Albach, 2x = 24 V. ritteriana pour M. M. ka Ort. & Albach, 4x = 48 V. rubra (Douglas) M. M. Mart. Ort. & Albach, 2x = 24 | | | Species not investigated: Veronica ogurae (T. Yamaz.) Albach, V. pyrethrina Nakai. 2c. Veronica subsect. Pinnatae (Holub) Elenevsky xpa = 17 (9 +9 — I) E, A Seg M. M. vhs po & Albach, 2x = 24 l EM ^. utahensis M. M. Mart. Ort. & Albach, 2x = 24 Species not investigated: Veronica laeta Kar. & Kir., V. ^ wyomingensis (A. Nelson) M. M. Mart. Ort. 8 Albach, sessilifTora Bunge ex Ledeb. lr = 24, 48 . missurica X Vo rubra, 2x = 24 2d. Veronica subsect. Longifoliae (Holub) Elenevsky roy : I Cochlidi (Rehb.) M. M - f : “eronica subg. Cochlidiosperma chb. . M. xp = 17 (049 L > o A Mart. Ort. & Albach i bachofenii Heuff. = - ktusiana subsp. Mies N if ai) T. Yamaz., 4 xp = . Veronica sect. Diplophyllum (Lehm.) Boriss M e = ee : x=9 V. hiusiana subsp. miyabei (Nakai € Hondo) T. Yamaz., 2 —— V. crista-galli Steven, 2x = 18 A a” vifolia L., = 34, 68 LV . . " Species not inve stigi lec eronica simensis Fresen. . subsessilis ( Tm ) C ;jarriere, , xpa = 08 l ‘ : , y . : v ws 2. Veronica sect. Cochlidiosperma (Rchb.) Be Species not investigated: Veronica ovata Nakai, V. sachali- i ; s mE tios] i ma (R : i nth. nensis T. Yamaz., V. sajanensis Printz, V. steboldiana Miq., a. Veronica subsect. Cymbalariae Benth Laigischensis Stepanov. 21-9 V. e Bodard, - x — 36, 54 2e. Veronica subsect. Pseudolysimachium (W. D. J. V. lyeica E. B. J. Lehm EN = 15 Koch) Elenevsky V. Vince | e 2x — 18 xp 7 17 (949 1) V. stamatiadae M. A. Fisch. € Greuter, 2x = 18 V. barrelieri H. Schott ex Roem. & Schult. subsp. V. trichadena Jord. & Fourr., 2x, da = 18, 36 barrelieri, 2 xp, 4 xp = 34, 6t V. barrelieri subsp. crassifolia Wierzb.. 4 xp. = 68 2b. E subsect. Cochlidiosperma (Rehb.) Albach V. pl lieri subsp. prodanii (Degen) Albach, 4 xp4 = 08 x=09, V. incana L., 4 xp = 00 V. si ids Deb.. a Decor & Hervier, ?x = 28, 30 V. incana ME = 34 V. hederifolia L., Ox = 54 V. incana subsp. NE, (Host) ee h. : xp = 08 V. stewartii E nne sell 2x — 18 V. orchidea Crantz, 2 xpa. 4 xp, 34. 08 V. sublobata M. A. Fisch., 4x = 36 V. ornata Monjuschko. 4 xp, = - 68 V. triloba Opiz, 2x = 18 Volume 95, Number 4 2008 Albach et al. Chromosome Numbers in Veroniceae (Plantaginaceae) VI. Veronica subg. Pelidosperma (E. I Mart. Ort., Albach & M. A. Fisch. gm Sibth. € Sm. subsp. glauca, 2 glauca subsp. chaubardii (Boiss. & pue Maire & pos 2x — 18 V. glauca bp. peloponnesiaca (Boiss. & Orph.) Maire & Petitm., 2x = 18 — V. a Wendelbo, 2x = 16 V. praecox All., 2x = 18 V. triphyllos L., 2x — 14 Species not investigated: d aznavourtt. Dórfl., V. doni Rómpp, V. samuelssonii Rech. Veronica subg. Stenocarpon (Boriss.) M. M. Mart. Fisch. VII. Ort., Albach € M. A. =8 V. ciliata Fisch. subsp. ciliata V. ciliata subsp. cephaloides (P enne "m D. XS 16 Hong, 2x = . contandriopouli Quézel, 4x = 32 . densiflora Ledeb., 2x = 16 . erinoides Boiss. & Spruner, 2x = 16 . fruticans Jacq. subsp. fruticans, 2x = fruticans subsp. cantabrica M. Laínz, 2x = 16 fruticulosa L., 2x = 16 l = ) mampodrensis Losa a p: Monis- 2x = 16 numm iularia G ouan, => saturejoide s. 2x = 16 | \ | V. V. V. V. kellerert Depen à & Urum., V. y. V. V. . thessalica Benth, 2x = 1 Veronica cachemirica Gand., V. Yamaz. V. eriogyne H. J. Species nol investigated: Royle ex Benth., V. capitata dd T. De Trautv.. V. em mdi] . Yamaz., V. >. Winkl., V. fedtsc henkoi Boriss., V. filipes P. C. Tsoong, V. gorbunovii Gontsch., a D. Don, V. Eara Benth., V. . lanuginosa Benth., V. lanosa o x Benth., luetkeana Rupr macrostemon ju ex Ledeb., f Zakirov, V. Trautv., V. ma icrostemonoides mexic ana a M. A. Fisch., V. : 2 Regel, V. tianschanica monticola Li, V. ruprechtii Lipsky. ! Lincz. VIII. Veronica subg. Pocilla (Dumort.) M. M. Mart. Ort., Albach & M. A. Fisch. Veronica amoena Steven ex M. V. violifolia Hochst. ex Benth. Species not investigated: Bieb., l. Veronica sect. Subracemosae (Benth.) Assejeva la. Veronica subsect. Subracemosae Benth. x=7 V. arg arguteserrata Regel € Schmalh., 6x = 42 V. biloba L., 4x = 28 V. campylopoda Boiss., 4x, 6x, Ox = 28. 42, 56 V. capilli Nevski, 4x = 28 V. arguteserrata X V. campylopoda. 6x = 42 a ies not investigated: V. bucharica B. Fedtsch., V. nevskii Boriss.. V. V. tenuissima Boris ramosissima Boriss., V. stylophora Popov ex Vved., Ib. Veronica subsect. Cardiocarpae Boriss. ex Ele- nevsky 2 x= 7 or 8? 3. J. Lehm.) M. M. V. cardiocarpa (Kar. & E ) Walp., 2x = Species not lc. Ve =7 = investigated: Veronica intere Now Bornm. ¡sk y ronica subsect. Brevistylae Kleney V. rubrifolia subsp. Fisch., 4x — 28 respectatissima M. A. Species nol investigated: Veronica avromanica M. A. Fisch., V. ferganica Popov, V. macropoda Boiss., V. viscosa Boiss. Pocilla Dumort 2a. Veronica subsect. -~ . Veronica sect. Agrestes Benth. x=7 agrestis L., SA 4x — 28 , ceratocarpa C. a oMa 4 2x = 14 V. filifo V. francispetae M. rmis Sm., y Fisch.. = 14 V. opaca Fr., 4x = 28 V. persica Poir., 4x = 28 V. polita Fr., 2x = 14 V. siaretensis E. Bd Lehm., 2x = 14 Species not investigated: Veronica bungei Boiss., V. longi- pedicellata Saeidi. Veronica subgen. Penlosapalne (Benth.) M. M. Mart. d Albach & M. A. Fisch. Incertae sedis Boiss., V. rniakowskiana Monjuschko, Species not Paus s Veronica aucheri e V. io ees a Czerniak — chionantha gaubae Born B. Fedtsch., » paederotae Bois , V. V a n minuta C. A. Mey., V. mirabilis Wendelbo, | la. Veronica subsect. Pentasepalae Benth. X = V. aragonensis pie 4x — 32 ^ austriaca I... = / 4 > austriaca sis dostai (Domin Dostal, 6x — 48 . crinita Kit. Schult. , Jacquinii D , 2x, 6x, Bx, 10x = 16, 48, 64, 80 . Jacquinii e wicdfi Degen, 4x = 32 cindlii Adamovic, 2x = 16 . orsiniana Ten.. 2x = 16 . prostrata L. subsp. prostrata, a = 16 . prostrata subsp. sibirica Watzl, 2x = 16 . rhodopea (Velen.) Degen ex Stoj. e Stef., 2x = 16 rosea Desf., 2x = 16 scheereri (]. P. V ens 4x — 32 : sennenil Fæ M. M. s 8 E. e x = 64 ^. tenuifolia / subsp. ten vuifolia Dm 2x — 16 pea cea fontqueri (Pau) M. M. m Ort. & E. Rico, 2x — V. mand prd javalambrensis (Pau) Molero & A. E nus s 2x, 4x — 16, 32 V. teucrium L., ^d x — 48, 64 V. Neb ee Stoj. & Stef., Veronica Xgundisalvi Rennen 2x — 16 b. Veronica subsect. Armeno-persicae Stroh x=8 V. armena Boiss. & Huet.. 2x = 16 Annals of the Missouri Botanical Garden V. farinosa Hausskn., 2x = 16 V. microcarpa Boiss., Ax = 32 Species not investigated: Veronica _acrothe ca Bornm. & EN Link, V. inensis K. Koch.) Fisch., Gauba, V. montbretii, M. A. : alieni Wo Oronow. le. Veronica subsect. Orientales (Wulff) Stroh x= V. bombycina Boiss. & POE 2x --—3]6 V. caespitosa Boiss.. 2x = S O V. cinerea Boiss. & Bal.. = 16 V. cuneifolia D. Don, 2x = 16 V. dichrus Schott & Kotschy, 2x = 16 V. elmaliensis M. A. Fisch., 8x = 04 V. kurdica Benth.. 6x = 48 V. macrostachya subsp. sorgerae M. A. Fisch., 2x = 16 V. multifida L., 2x. 4x, 6x. 10x = ie . 32. 48, 80 V. multifida X V. dichrus, 6x = 6 V. orientalis oe lx, Ox = 32, 04 V. pectinata Lo, O V. thymifolia PE & Sm., 2x = 16 Species not investigated: Veronica allahue kberensis Öztürk. V. antalyenst A. Fisch., Erik € Sümbül, V. cetikiana Öztürk. | fragilis Boiss. & Haussl n... V. fridericae M Fisch., ae reyn € Sint. V. galathica Boiss.. | Dl Boiss. ds Benth.. V. polium P. H. Davis . rechingeri M. A. Fisch. RS ae Boiss. & Balanse, V. Wil "na v. ees Bornm., V. thymoides P. H taurica avis ld. Veronica subsect. Petraea Benth. x= V. peduncularis M. Bieb., 2x = 16 nol investigated: Veronica baranetzkii Bordz.. V. bogosensis Tumadz., borisovae Holub, V. caucasica M. 5 V. filifolia Lipsky, V. Bieb.) Steven, V. Bieb.. V. vendetta-deae Albach Species => petraea (M. | ieb. a umbrosa M. X. Veronica subg. Chamaedrys (W. D. J. Koch) Buchenau l. Veronica sect. Alsinebe Griseb. la. Veronica subsect. Microspermae (Rómpp) Stroh =0 V. arvensis L. = 16 Species not investigated: Veronica sartoriana Boiss. & Meldr. ». Veronica subsect. Microspermoides Albach x= V. brevistyla Moris, 2x = 16 . dillenii Crantz, 2a Fe verna l., = 16 2. Veronica sect. Chamaedrys W. D. J. Koch . Veronica subsect. Astachamaedrys Albach x= i. ? V. laxa Benth.. dx. Za V. magna M. A. Fisch.. Ox = 32, 46 = 42 2b. Veronica subsect. Multiflorae Benth. x= V. chamaedryoides Bory & Chaub.. 2x = 16 V. chamaedrys L. subsp. a 2x, 16, 32 V. chamaedrys subsp. micans M. “Isc h.. 16 V. krumovii (Peev) Peev, 2x = ae V. micrantha Hoffmanns. o od = 16 V. orbelica (D. Peev) D. "16 V. vindobonensis (M. A m isc B ) " A. Fisch., 2x = 16 XI. Veronica subgen. Pseudoveronica J. B. Armstr. D EE sect. Detzneria (Schltr, ex Diels) Albach x V. on (Diels) Albach. 6x = 38-42 or 48 2. Veronica sect. Derwentia (Ral) B. G. Briggs XHebe = 19, 20, 21 24 species, see Briggs and Ehrendorfer (2006) for review. 3. Veronica sect. Hebe (Juss.) Benth. XHebe = 20. 21 For 121 species from New Zealand and adjacent islands, see Bayly and Kellow (2006) and Garnock-Jones and Lloyd (2004) for a review Eun from New Guinea: | albiflora (Penne H) \Ibach, 2 xg = f ionantha Albach, 2 xg. = 40 N Species nol "ban Veronica brassii (Pennell) Albach. vam, V. diosmoides Royen & Ehren Wernham, V. carminea Albach. V. carstensensis Wern Schltr., V. inflexa B. ich; V. papuana (P. Albac f strigosa Abash, V. a Albach. vandewateri MI. Veronica dcs Triangulicapsula M. M. Mart. Ort. l. Aba h 5 M. sel f I hamapiihyoides 1 Lam., 4x = 24 V. grisebachii Walters, 2x = 12 Incertae Sedis within Veronica x=8 2x = 16 . javanica Blume. TOWARD A PHYLOGENY OF C. Bessega.? H. E. Hopp.? and R. H. Fortunato? MIMOSA (LEGUMINOSAE: MIMOSOIDAE): A PRELIMINARY ANALYSIS OF SOUTHERN SOUTH AMERICAN SPECIES BASED ON CHLOROPLAST DNA SEQUENCE! ABSTRACT The pantropic al and subpantropic “al genus Mimosa L. comprises more than 500 species, of which nearly 480 are re ponte ad for the American Continent. Mimosa is subdivided into five sections, four of which are represented in southern South America: Built EN DC., Habbasia DC., Calothamnos Barneby. and Mi . Previous taxonomie studies of the species from the austral region have found classification conflicts among (a) sections Be Bo a and Habbasia, (b) sections Calothamnos and Mimosa, and (e) series and subseries within section Mimosa. This paper reports a preliminary phylogenetic analysis of chloroplast nucleotide sequences of the trnL intron and the trnL-rnF intergenic spacer from 34 species of Mimosa and related genera. Key gical characters were mapped onto the phylogenetic hypothesis and discussed. Sequence analysis indicates that the genus Mimosa is monophyletic; it is derived from Piptadenia viridiflora (Kunth) Benth. The four sections proposed by Barneby are not natural groups. The cladogram retrieved indicates that the representatives of Mimosa sect. Batocaulon are not clustered, the xerophylous representatives of this section are basal. and the remaining spec ies are related to those species of sectior Habbasia, suggesting that section Batoe ilon ser. Stipellares is more recently derived. The species of Mimosa sect. C Do that were analyzed are nested in section Mimosa. The results seem to support retention of this section within section Mimosa as was noted previously by Bentham. P chloroplast sequence data suggest that the representatives from sections Calothamnos and imosa share a common ancestor with those from section Habbasia and section Batocaulon ser. Stipellares. ey words: cpDNA, Mimosa, a trnL intron, trnL-trnF intergenic spacer The pantropical and subpantropical genus Mimosa flowers, each organized into infrasectional ranks. The L. has about 500 species. Nearly 480 are reported for generic definition of Mimosa as proposed by Bentham the American Continent (Barneby, 1991, 1993: Turner has found nearly universal acceptance. Its limits were 1994a, b; Fortunato € Palese, 1999: Queiroz et al... only challenged by Britton and Rose (1928) in North 1999: Grether, 2000: Silva & Secco, 2000: Izaguirre & American Flora, with the introduction of a novel Beyhaut, 2002). The first comprehensive taxonomic division. They segregated Mimosa as another genera revision of this diverse genus was proposed by based on species that had different pod type and leaf Bentham (1875) in his global monograph of the reduction (= phyllodes). However. their artificial O o | 3 | J Mimoseae and later supplemented in Flora Brasilien- classification has largely been ignored. In South sis (1876). Bentham (1841-1842, 1875) recognized in America, the only revision of the Argentinean species the genus: (a) section Mimosa, with haplostemonous was published by Burkart (1948). His treatment was flowers; and (b) Habbasia DC., with diplostemonous largely in accord with Bentham's infrageneric classi- We are grateful to the curators of the herbaria listed under Materials and Methods for permission to study and/or loans of specimens. We wish to thank María Marcela Manifesto and Jim Luteyn for the critical reading of the manuscript. We are also indebted to the anonymous reviewers and the scientific editor, Victoria C. Hollowell, for their valuable comments is led to signific ant iux ments in the manuscript. We want to thank Consejo Nacional de Investigaciones Científicas y Técnicas (CONICE r the wi toral fellowship given to C. Bessega. This research was supported by grants 522307 (Instituto Nacional ES Tecnología Agropecuraria [INTA]) and PID267 (Agencia Nacional de Promoción Cientifica y Tecnológica [ANPCyT]) to H. E. Hopp. and by the Myndel Botanica o Collection tip 8 erants 2002 and 2004 to R. H. Fortunato. ? Laboratorio de Genética, De ‘partamento de Ecología Genética y Evolución, Facultad de Ciencias Exactas y Naturales. | 128 Buenos Aires. Argentina: Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET). pui nlina. Unidad Integrada de Investigación y Docencia, Facultad de Ciencias Exactas y Naturales (FCEyN) de la Unive ipe $ Buenos Aires (UBA): lali de Genotipificación Molecular, Instituto de Biotecnología CICVy A, INTA, Castelar. 17 Prov. de Buenos Aires, Argenti Instituto de Recursos Biológicos. INTA, Castelar, 1712 Prov. de Buenos Aires, Argentina: Consejo Nacional « Investigaciones Científicas y Técnicas (CONICET), Argentina: Universidad de Morón, Prov. de Buenos Aires, Argentina. Author for correspondence: fae ns dor: 10.34.17/20060 12 ANN. Missouri Bor. Garb. 95: 567—579. PUBLISHED ON 30 DECEMBER 2008. Annals of the Missouri Botanical Garden fication, even though eight additional species and seven varieties were described. Although there has been no global evaluation of Mimosa since Bentham (1875, 1876), Barneby (1991) presented a new taxonomic treatment of Mimosa in the New World that makes up approximately 90% of the total species. He noted two chief foci of speciation in Brazil araguay, Americas: (a) south of Amazonia and P the adjacent areas of Argentina, and Uruguay; and (b) central and southern Mexico, as well as other minor ones in Cuba-Hispaniola, and the Orinoco basin. In addition to this, Barneby expanded Bentham's generic definition, but he also included the genera Schrankia Willd. and Schranckiastrum | Massl. He reorganized the sectional and infrasectional classifica- tion, established new categories, and reordered the old ones. This classification recognized five sections in Mimosa: Batocaulon DC., Hab- basia, Calothamnos Barneby, and Mimosa (Table 1). Mimadenia Barneby, In the phylogeny proposed by Barneby, Mimosa was 1901: He suggested that section Mimadenia is a derived from Piptadenioid ancestors (Barneby, 15-16). morphological group intermediate between the Pipta- denioid ancestor and the remaining species of Mimosa. The novel position he gives to section Mimadenia is based on the extra floral nectaries that resemble both genera, and the ovate anthers that are intermediate between the orbicular connective, girdled by incurved anther sacs of Mimosa and the oblong connective, which is flanked by straight parallel anther sacs Piptadenia Benth. The anthers of the species of Mimosa sect. Mimadenia always lack the terminal gland present The dium type of some species of Mimosa sect. Mimadenia 1 Adenopodia C. Presl and Piptadenia. craspe- is not different from that of Adenopodia and Entada Adans., but is not related to the valvate pod present in 199]: 25). species of Piptadenia s.l. (Barneby, All included in section Habbasia and divided by Bentham (1875) into two groups: (1) those with setose pubes- and (2) Table 1). However, in diplostemonous Mimosa were cence those with less frequent setose me] pubescence Barneby’s classi- fication (1991), only the first group was maintained in this section. The other group was recognized as section Batocaulon. Further, Barneby (1991: 24) proposed that section Batocaulon is a derivative of Mimadenia. Bentham’s concept of section Batocaulon ae and was mainly based on the presence of cauline se the setiform cilia on leaflet margins, instead of stamen which he formerly used to characterize both Habbasia. M mention that there are exceptions in the species number, sections Mimosa and is important t admitted by Barneby in Batocaulon and now, with his definition, it has become the most heterogeneous section of the genus. Barneby's sections Calothamnos and Habbasia were postulated to be derived from section Batocaulon. Section Calothamnos (sect. Mimosa ser. Lepidotae Benth. [Table 1]) was defined by Barneby (1991) as an independent sectional rank on the basis of the stellate setae. He proposed that the stellate seta type was clearly derived from the plumose one by shortening of the primary axis and elimination « some or all of the branches. According to Barneby's position (1991), the survival of the rudimentary inner sel filaments in some species of Calothamnos suggests that the origin of the section could be closer to the. diplostemonous and stellate flowers from section Batocaulon ser. Leiocarpae. Although Mimosa lanuginosa Glaz. ex Burkart and M. diversipila Micheli have stellate seta indumentum, in (1991), they were excluded from section Calothamnos and transferred to Barneby’s taxonomic. classification section Mimosa ser. Mimosa subser. Polycephalae (Benth.) Barneby Brevipedes and section Mimosa ser. Mimosa (Benth.) Mimosa lanuginosa shares habit type and the palea- subser. Barneby, respectively. ceous calyx with other members of subseries Poly- cephalae, although both features are unknown in section Calothamnos. In contrast, the calyx of M. diversipila is reduced to a rim as is the case in many species of section Calothamnos, but it is a virgate subshrub with mostly simple, densely foliate stems extending into long efoliate racemes of capitula, a condition absent from Based Barneby's proposal, the plumose indumenta present in other members of section Calothamnos. on several infrageneric groups could be considered a convergence in an evolutionary scenario. The haplostemonous section Mimosa was considered to be derived from Habbasia. Barneby (1991) excluded series Lepidotae and Spiciflorae Benth. from section (Table 1). to Batocaulon Mimosa The Spiciflorae species were transferred ser. Plurijugate Karsten, although they have either a tetramerous or pentamerous pertanth and lack the setiform trichomes characteristic 1991). Sect (199] habit leaf of section Mimosa (Barneby, ion Mimosa sensu Bentham (1875), or Barneby ). shows the most derived. features such as the (mainly subshrubs to herbs), reduction of formula, craspedodrome leaflet venation, and reduction of petals and androecium numbers (Barneby, 1991: 23) Barneby America: Four of the five sections recognized by 1991) are represented southern South Habbasia, During the taxonomic analysis of the species from that Batocaulon, Calothamnos. and Mimosa. austral region (Fortunato, unpublished data), conflicts of delimitation were found between: (a) sections Batocaulon and Habbasia, (b) sections Calothamnos (c) within section Mimosa. and Mimosa. and among series and subseries Volume 95, Number 4 2008 gal O) o) Bessega et al. Phylogeny of Mimosa Thus, Barneby’s hierarchical treatment of Mimosa has raised some controversial issues. In an effort to examine these, we propose to use information from genomic data, which are still lacking for Mimosa species. Noncoding chloroplast DNA (cpDNA) regions have been used extensively for plant phylogenetic analyses and have proven to be particularly informa- tive at the infrageneric level or when comparing related genera (Taberlet et al., 1991; Bohle et al., 1994; Gielly € Taberlet, 1994; Sang et al., 1997; Baker et al., 1999; Murphy et al., 2000; Hughes et al., 2003; Luckow et al., 2003, 2005; Simpson et al., 2004; Jobson & Lukow, 2007). This paper reports preliminary phylogenetic anal- | intron and the trnL-trnF intergenic spacer from 34 — > ++ — yses of chloroplast nucleotide sequences of the trnL species of Mimosa and related genera. The main purposes of this study are to: (1) examine the evolutionary relationships within Mimosa and among (2) 1 Mimosa and other members of the Leguminosae; determine if Mimosa is a monophyletic group and, i the case that it is, what the monophyletic units within the genus are; (3) determine whether the phylogeny trnL-trnF spacer data is congruent with the current classifica- based on trnL intron and the intergenic tion and the evolutionary sequence proposed (Bar- neby, 1991); and (4) gain further insight into the subsequent phylogenetic studies of the genus Mimosa. MATERIALS AND METHODS The trnL intron and trnL-trnF. intergenic 1991) were sequenced spacer EA regions (Taberlet et al., Or a total of 34 taxa (Table 2). We included species that represent sections Mimosa, Habbasia, Calothamnos, and Batocaulon distributed in Argentina, Paraguay, and Brazil. We analyzed both fresh leaflets that were collected and dried in the field with silica gel, and leaflets from herbarium Morphological observations for the species were made at the Instituto specimens. de Recursos Biológicos, Instituto Nacional de Tecno- logía Agropecuraria (INTA), Castelar (BAB) on herbarium loans from B, BA, BAA, BAF, BH, BM, ES, E, F, G, GH, HB, K, LIL, M, MO, NY, P, PY. RB, SI, SP, US, W. Six representing different mimosoid tribes were selected and outgroups on the basis of morphological and molecular criteria (Luckow et al., 2003, 2005; Jobson & Luckow, 2007). axa, voucher details, and GenBank accession numbers are listed in Table Genomic DNA was isolated with a CTAB protocol (Doyle & Doyle, 1987) or with the DNA Easy Extraction kit (Qiagen, Valencia, California, U.S.A.) according to the manufacturer's instructions. Two cpDNA regions were amplified from purified DNA via polymerase chain reaction (PCR) using the four primers, c (5'-CGAAATCGGTACACGCTACG-3^), d (5'-GGGGATAGAGGGACTTGAAC-3), e (5'-GGTT CAAGTCCCTCTATCCC-3), and f (5'-ATTTGAA CTGGTGACACGAG-3') (Taberlet et al., 1991). PCR reactions were of 50 ul volume, including 2.5 units of Taq-DNA polymerase (Invitrogen, Carlsbad, California, U.S.A.), 0.2 mM each dNTP (Invitrogen), 50 ng each primer (synthesized by AlphaDNA, Montreal, Canada), 3 mM MgCl (Invitrogen), and 50 ng genomic DNA. Amplification was performed on a Biometra (Goettin- T-gradient thermal cycler using the following steps: 95°C for 4 min., 35 cycles of 94°C for 20 sec., 55°C for 30 sec., C for 2 min., additional final extension step of 10 min. at 72°C Germany) gen, and 72 with an Amplified products were separated on 1% agarose Tris-acetate-ethylenediamine tetraacetic acid (TAE) gels and purified using QlAquick Gel Extraction kit Qiagen) and used as a template for direct sequencing. In some cases, direct sequencing failed and cloning was performed using the pGEM T-Easy Vector System I U.S.A.) with DH5« competent bacterial strain. Sequencing reactions were performed with an ABI 3730 XL DNA sequencer by Macrogen Inc. (Seoul, Korea). The same (Promega, Madison, Wisconsin, automatic primers or the universal T7 and SP6 primers were used to make the primer extension for the direct sequence or cloning fragments, respectively. Sequences were edited with the computer program BioEdit (Hall, 1999), initial alignment was performed using ClustalX version 1.8 (Thompson et al., 1997) with the default parameters, and minimal manual adjustment was made. Both DNA regions were merged in WinClada (Nixon, 1999). the primer annealing sites and regions in which the alignment Uncertain. positions located near was ambiguous were excluded from the data set. Parsimony analysis was conducted with NONA (Goloboff, 1998) included in the WinClada software (Nixon, 1999) using 1000 random addition sequences, ree bisection-reconnection (TBR) holding 100 trees per replicate, and attempting to swap to completion. — Characters were considered unordered and equally weighed. The bootstrap analysis used 1000 replica- tions each with 10 random additions holding 10 in each replicate, with a maximum of 100 trees saved per replication and mapped to the selected tree. Based on principal morphological criteria for delimitation of sections, 10 morphological characters were selected for optimization on the trnL intron and the trnL-trnF intergenic spacer phylogeny to allow for discussion of the evolution of these characters in a phylogenetic context. Morphological features (listed in Fig. 3) were scored from specimens in the collection of the BAB herbarium. Nine of these characters have two states and one is multistate. 570 Annals the HRS Botanical Garden Table 1. Bentham (1875) and Barneby (1991) taxonomic proposal in Mimosa L. Bentham (1875) Barneby (1991) Section Series Section Series Subseries Habbasia DC. Setose pubescence Occasional lose pubescence Mimosa Habbasia Js Ga: 80 Mimadenia uneby, ca. 16 Glanduliferae Benth. Rubicaules Benth. Acanthocarpae Benth. S. Ephedroide ae Benth. 85 Batocaulon DC Cesalpinitfoliae Benth. Leptostachyae Benth. (= Leiocarpae Benth., Stachyomima Benth.. Tomentosae Benth. sensu 841-1842) Bentham, | Lepidotae Benth. i ond DC. Habbasia Pachycarpae Benth. Setosae Barneby Neptunioideae Barneby Rondonianeae Barneby Rojasianae Barneby Pseudocymosae Hass Piresianae Barneby Glanduliferae Benth. Nothacaciae Barneby Revolutae Barneby Rubicaules Benth. Acanthocarphae Benth. Ephedroideae Benth. Supellares Benth. Paucifoliate Benth. Cesalpinüfoliae Benth. — etocarpae Benth. Distachyae Barneby (= Leiocarpae Benth: Grether, 2000) Andinae Barneby Acantholobae Barneby Boreales Barneby Bimucronatae Barneby Hucaenioideae Barneby Fagaracanthae Barneby Bahamenses Barneby Farinosae Barneby Echinocaulae Barneby Glandulosae (Benth.) Barneby Auriculate Barneby Ceratontae Barneby Cordistipulae Barneby Campicolae Barneb Filipides Barneby Quadrivalves Barneby Plurijugae Karsten Calothamnos Barneby, ca. 24 Myriophyllae Benth. Mimosa, ca. Sensitivae Benth. 80 Pectinatae Benth. Pudicae Benth. Pedunculosae Benth. Meticulosae Benth. Hirsutae Benth. Obstrigosae Benth. Myriophyllae Benth. Mimosa Mimosa Pectinatae (Benth.) Barneby Pudicae (Benth.) Barneby Pedunculosae (Benth.) Barneby Dolentes Barneby Polycephalae (Benth.) Barneby Hirsutae (Benth.) Barneby Obstrigosae (Benth.) Barneby Volume 95, Number 4 2008 Bessega et al. Phylogeny of Mimosa Table 1. Continued. Bentham (1875) Barneby (1991) Series Subseries Section Series Section Castae (Benth.) Barneby Castae Benth. Mimosa, ca. 180 Modestae Benth. Modestae Benth. Mimadenia Barneby Batocaulon DC. Spiciflorae Benth. Dolentes Barneby Polycarpae Barneby Teledactylae Barneby (= ser. Teledactylae (Barneby) R. Grether; Grether, 2000 Lactifluae Barneby (= ser. Lactifluae (Barneby) R. Grether: Grether, 2000) Bolivianae Barneby ^y = Lundianae Barne Trichocephalae Barneby Ramosissimae Barneby Serrae Barneby Sparsae Barneby Morongianae Barneby Bipennatulae Barneby Affines Barneby Simplicissimae Barneby Dicerasteae Barneby Discobolae Barneby Brevipedes Barneby Thomistae Barneby Pogocephalae Barneby Microcarpae Barneby Lanatae Barneby Macrocalycinae Barneby Reptantes Barneby Ramnetaceae Barneby Axillares Barneby Dryandroideae Barneby Dutranae Barneby Widgrentanae Barneby Diffusae Barneby Myriadeniae Barneby Plurijugae Karsten RESULTS The aligned length for the trnL and the trnL-trnF region is 633 bp and 514 bp, respectively. In the ingroup, DNA sequence lengths of the intron ranged from 525 bp in Mimosa paupera Benth. to 541 bp in M. somnians Humb. & Bonpl. ex Willd. The outgroup species had shorter sequences (496-524. bp). For the IrnL-trnF. intergenic spacer, DNA sequence lengths varied from 358 bp in M. oligophylla Micheli and M. balansae Micheli to 388 bp in M. paupera. A matrix of 582 bp length for the trnL intron and 458 bp for the trnL-trnF intergenic spacer was analyzed. Gaps in the ingroup varied from 1 to 17 bp in length for the intron and from 1 to 36 bp for The data matrix that was used the intergenic spacer. consisted of 1040 positions, from which 124 (11.9%) were potentially informative characters. l'a- Sequence characteristics are summarized in ' ble 3. Parsimony analysis resulted in 92 trees consisting of 470 steps with a consistency index (CI) of 0.68 and RI) of 0.65, excluding uninforma- p a retention index tive characters. The majority consensus tree and one of the most parsimonious trees are shown in Figures | and 2, respectively. In. general terms, the preferred topology is in agreement with the consensus tree. The species of genus Mimosa analyzed here constitute a single clade in the majority consensus tree (Fig. 1) and in the preferred topology cladogram (Fig. 2, BS — 572 Annals of the Missouri Botanical Garden Table 2. Species names for DNA used in the present study of trnL and trnL-trnF sequences, vouchers (deposited at BAB | Herbarium Inst. Recursos Biológicos, INTA, Argentina], and GenBank accession numbers. Taxa are in alphabetical order. RHF = Reneé H. Fortunato: JAG = Julián A. Greppi; MAL = Melissa A. Luckow; EMZ = Elsa M. Zardini. Herbarium irnL GenBank trnL-trnF Species voucher numbers GenBank numbers Mimosa adenotricha Benth. RHF 8450 DQ344579 DQ344613 Mimosa balansae Micheli RHE 7534 DQ344552 DQ344585 Mimosa bifurca Benth. var. bifurca RHF 7556 DQ344553 DQ344586 Mimosa bonplandii (Gillies ex Hook. & Arn.) Benth. JAG 92 DQ344581 DQ^3440615 Mimosa brevipetiolata Burkart JAG 127 DQ344582 DQ344616 Mimosa candollei R. Grether RHF 7555 DQ344555 DO0544588 Mimosa debilis Humb. & Bonpl. ex Willd. RHF 8085 DQ3445061 10344594 Mimosa detinens Benth. MAL 4491 DQ344558 DQ344591 Mimosa diversipila Micheli var. diversipila RHF 8810 DQ344584 DO544618 Mimosa flagellaris Benth. RHF 7887 DQ344557 110344590 Mimosa hexandra Micheli MAL 4584 DQ344556 10344589 Mimosa hirsutissima Mart. RHF 7962 DQ344.562 DQ3544595 Mimosa leimonias Barneby & Fortunato EMZ 41158 DQ344575 110344609 Mimosa maguirei Barneby RHF 8418 DQ344576 DQ344610 Mimosa obstrigosa Burkart RHF 8077 DQ344568 DQ344001 Mimosa oligophylla Micheli RHF 8074 DQ344574 110344008 Mimosa paupera Benth. RHF 8018 DQ344505 10344598 Mimosa pigra L. JAG 38 DQ344560 105344593 Mimosa pilulifera Benth. JAG 39 DQ344573 10344607 Mimosa polycarpa Kunth RHE 8019 DQ344.566 110344599 Mimosa radula Benth. RHE 8413 DQ344577 DQ34401 1 Mimosa sensibilis Griseb. RHF 8385 DQ344580 DQ344614 Mimosa setosa Benth. var. setosa RHF 8445 DQ344578 DQ344612 Mimosa somnians Humb. & Bonpl. ex Willd. RHF 7953 DQ344563 10344596 Mimosa strigillosa Torr. & A. Gray RHF 8949 DQ344567 103440600 Mimosa tweedieana Barneby ex Glazier & Mackinder RHF 7949 DQ344572 10344006 Mimosa uliginosa Chodat & Hassl. RHF 8059 DQ344564. DQ344597 Mimosa xanthocentra Mart. var. xanthocentra RHF 7650 DQ344559 DQ344592 Outgroup — Anade bane colubrina (Vell) Brenan var. cebil RHF 7583 DQ34457 1 DQ344605 laxa (Griseb.) Altschul Calliandra tweediei Benth. RHF 8853 DQ344583 DQ344617 Mimozyganthus carinatus (Griseb.) Burkart RHF 75607 DQ344570 DQ3440604. Neptunia pubescens Benth. RHF 7923 DQ344551 DQ344003 Parapiptadenia excelsa (Griseb.) Burkart RHF 7669 DQ344569 DQ344602 Piptadenia virtdiflora (Kunth) Benth. RHF 7586 DQ344554. DQ344587 80%), indicating that this group is monophyletic. Both In both trees, the species of section Batocaulon are cladograms indicate that Piptadenia viridiflora not clustered together, and Mimosa hexandra Micheli (Kunth) Benth. is the species most closely related to and M. detinens Benth. are separated from the rest and Mimosa (Fig. 2, BS = 87%). constitute a basal clade (Fig. 2, BS = 94%). The Table 3. Sequence characteristics for the trab, intron and the trnL-trnF intergenic spacer sequenced in this study. trab intron trnL-trnF intergenic spacer Total Length range (bp) 496—541 358-388 — Aligned length (bp) 633 514 1147 No. of sites excluded (bp) 5l 56 107 Analyzed region (bp) 582 458 1040 Indels 27 26 53 Size of indels (ingroup) (bp) 1-17 1-36 1-36 Base substitutions 31 40 71 Total potentially informative characters 58 66 124 Bessega et al. 573 Volume 95, Number 4 2008 Phylogeny of Mimosa ear ec are ia pubesc Chae nanthera colubrina Outgroup diei iptadenia Piptadenia ls Mimosa hexandra M. pet: s Batocaulon r—— M.candollei M.strigillosa . M.somnians Habbasia 86 M.pigra .tweedieana 100 M.bifurca " NEL | Batocaulon a um M.maguire 69 M. eN Habbasia M.setosa M.paupera M.xanthocentra = e M.debilis Mimosa Calothamnos Mimosa M. bonplandii Calothamnos Figure 1. Majority consensus tree obtained from 92 equally : arsimonious trees (length 470, CI — 0.68, RI = 0.65) based on trab intron and trnL-trnF intergenic sp g branches are the percentage of trees in which The sections of the genus are indicated ata to Barneby (1991). acer sequence data. Numbers above the clade is supported. T remaining species of section Batocaulon, M. uliginosa In both figures, the species of sections Mimosa and Chodat & Hassl., and M. bifurca Benth.. the section Habbasia. From the polymorphic Batocaulon, M. candollei R. Grether shares more characters (Fig. 2, BS = 80%) are placed Calothamnos constitute a single clade. In Figure 1, with species of Mimosa paupera is sister to the remaining species; section however, in the consensus tree (Fig. 2), this species clade formed by M. hirsutissima Mart. Benth. In a similar way, M. sister (Fig. 2. BS < 50%) to & Bonpl. is sister to the and M. xanthocentra Mart. is the clade constituted by M. debilis Humb. with section Habbasia than do M. hexandra and M. radula detinens. In Figure 2, M. bifurca and M. uliginosa are more recently differentiated and clustered with high ode support within section Habbasia (Vig. 2. BS ex Willd. and M. sensibilis Griseb. (Fig. 2, BS = 00%). However, this last clade is not highly — 7196). In the consensus tree, M. xanthocentra, the supported (Fig. 2, BS < 50%) and can partially cluster formed by M. debilis and sensibilis, the explain the polytomy found in the majority consensus cluster formed by M. hirsutissima and M. radula, and tree (Fig. 1). Mimosa maguirei Barneby, M. adeno- the group that includes the remaining species from tricha Benth., and M. setosa Benth. (sect. Habbasia sections Mimosa and Calothamnos that were ana- BS A cluster formed by M. ser. Pachycarpae Benth.) are clustered (Figs. 1, 2, = 94%). and the relation among clades integrated by representatives from Batocaulon and Habbasia is not solved (Fig. 1). On the the 2) indicates that the M. maguirei, M. other hand, preferred topology (Fig. adenotricha, and M. setosa clade is sister in relation to the cluster constituted by M. tweedieana Barneby ex Glazier & Mackinder and M. pigra L. (sect. Habbasia ser. Habbasia), M. strigillosa Torr. & A. Gray (sect. Habbasia), and M. Habbasia ser. Bipinnatae) (Fig. : Habbasia ser. somnians (sect. lyzed constitute a polytomy. pilulifera Benth. gosa Burkart (sect. Calothamnos) and M. obstri- both (sect. Mimosa) can be seen trees. From the remaining species, the same topology is found in both trees (Figs. 1, 2). Mimosa brevipe- tiolata Burkart is basal (Fig. 2, BS = 71%), followed by M. flagellaris Benth. (Fig. 2, BS = 79%), and a group that is split into two clades (Fig. 2, BS = 72%). The first one (weakly supported) consisted of second is a M. diversipila and M. oligophylla. The clade (Fig. 2, BS = 50%) formed by two groups; the Bootstrap Pm s greater than 50% are shown at the nodes. The | first one (Fig. 2, BS = (Gillies ex Hook. & Arn.) Benth. (sect. and M. Barneby & Mimosa), second includes M. aen and 2. BS = 85% 11%) includes M. bonplandii Calothamnos) leimontas Fortunato (sect. and the M. polycarpa Kunth (Fig. . both from section Mimosa. Discussion The present study represents the first molecular approach toward a laa of Mimosa. This is a preliminary reconstruction. based on the trab intron trnL-trn F from southern South and intergenic. spacer, including species America. Although sampling of taxa is limited, the four sections found in the southern center of diversity of the genus (south of Amazonia in Brazil and adjacent areas of Paraguay, Argentina, and Uruguay) (Barneby, 1991) are represented. Noncoding regions have been presumed be more useful lower taxonomic ranks because they are less func- tionally constrained and are free to vary, thereby potentially providing more phylogenetically informa- tive characters per unit of sequencing effort (Clegg et sections of the genus are indicated according to Barneby 574 Annals of the Missouri Botanical Garden Mimozyganthus carinatus Neptunia pubescen Anadenanthera colubrina Outgroup Calliandra tweedie Parapiptadeni Pi dide lil ee Batocaulon M. detin tin dá — M candollei rei M.adenotricha tos Habbasia ra M.tweedieana M.strigillosa somnia M.bifurca Batocaulon M.uliginosa Mimosa Calothamnos M.brevipetiolata flagellaris Mimosa M.diversipila M.oligophylla M. O | Calothamnos M.leimon M balansae. | Mimosa 85 M.polycarpa Figure One of the most parsimonious trees of 470 steps based on trab intron and trnL-2trm F intergenic spacer sequence data. (1991). 2002). In Mimosa. both intron and the trnL-trn F al.. 1994: Gonzalez & Vovides. the tral exhibited. a intergenic spacer similar level of informative characters (12% on average) to most of the other plant species assayed with comparable methods (Baker et al., 1999; Bayer et al., 2000; Murphy et al., 2000; 2002). Despite the fact that the regions for the alignments were excluded, a large data Gonzalez & Vovides, ambiguous set (1040 characters) was obtained, and 124 informa- live sites could be retrieved for the sampled species. A low level of conflict among data ensured that many of the relationships recovered were highly supported and are addressed in this contribution. The epDNA allowed us to make a phylogenetic indicates that the four hypothesis. The cladogram sections of Mimosa included in this study may be considered a monophyletic taxon, sister to Piptadenia viridiflora (Fig. 2. BS = 80%). conflicts of As discussed in the introduction, delimitation were found between (a) sections Batocaulon and Habbasia, (b) and (c) among sections Calothamnos and Mimosa. series and subseries within section Mimosa, and are discussed below. Volume 95, Number 4 2008 Bessega e Phylogeny » moss SECTIONS BATOCAULON AND HABBASIA Section Batocaulon was defined as a highly polymorphic group (Barneby, 1991) mainly based on plain, cauline setae and setiform cilia on leaflet margins. However, these types of indumenta are als found in some representatives of sections Habbasia and Mimosa. The species analyzed here are grouped M. detinens and M. hexandra (basal and highly supported). and into two clusters: the first one includes the second includes M. bifurca and M. uliginosa (more Our suggests that species of section Batocaulon may have recent and also highly supported). analysis arisen from two speciation events; the first one gave the Benth. (M. Bimucronatae Barneby (M. hexandra), series Farinosae and later event gave rise to series Stipellares (M. bifurca and M. uliginosa). that the Batocaulon recognized by Barneby (1991) cannot be considered a natural group. detinens) — and This result indicates section The position in the tree for Mimosa detinens and M. hexandra can be partially explained by the xeroph- ilous habit and the bitetrad type pollen (Caccavari, 1985. 1986a. b, 1989; Martínez-Bernal et al., 2005) as shown by the optimization in the phylogeny hypothesis indicated in Figure 3 (characters 9 and 10). Moreover. the association between M. detinens and M. hexandra could also be explained by the flowers white filament (Burkart, 1948: Barneby, 1991). Series Stipellares is characterized by its particular foliar morphology having paraphyllidia at the base of pinna-rachis elliptic, reniform to suborbicular, simulating diminu- live leaflets; no interpinnal spicules; and, in some species, pinnae and leaflets alternately inserted along the leaf axes. We mapped three types of paraphyllidia (Fig. 3, that the foliaceous type could be considered a derived feature character 2) and determined from the ovate and subulate types. Within the clade in which all species of section Habbasia sensu Barneby are represented together along with section Batocaulon ser. Stipellares, two clades may be distinguished (Figs. 1, 2). The first includes the species of series Pachycarpae (Mimosa M. setosa). Barneby described the series Setosae in order to include the polymorphic M. setosa (1991: 352-353), which shows Habbasia pods with relatively narrow replum that differentiate it from of Although include few representatives of series Pachycarpae maguirel, adenotricha) and series Setosae (M. transitional traits between sections and Batocaulon, but exhibits. articulated those series Pachycarpae. we I and Setosae, the molecular results presented here suggest maintaining them as a single series. When the indumentum type typical of both series was mapped (Fig. 3, character 3), the bulbosus or flagelliform type appears at the base of the clade integrated by / maguirei, M. adenotricha, and M. setosa; however, one might assume a parallel derivation of the valvate fruit type in M. adenotricha, M. maguirei, and M. candollei (Fig. 3 | constituted a moderately supported clade (BS . character 4). Figure 2, Mimosa tweedieana and M. pigra 11%). Nevertheless, although it is in the same series, M. strigillosa Habbasia humifuse, strigillosa is placed in another group. Mimosa members of series life basally monadelphous stamens; differs from other by its herbaceous form, with stems radicant: and smaller, 2- to 5-seeded pods. Moreover, this species has a bicentric U.S.A. Paraná and Paraguay rivers in Paraguay to northeast- distribution (southeastern to northeastern Mexico and the basins of the in the remaining species of series Habbasia. Although in the majority consensus tree (Fig. the ern Argentina and western Uruguay) that is unusual 1) it is not solved, in (Fig. 2), M. strigillosa is closely related to M. somnians (Bipinna- selected phylogeny hypothesis tae); despite this, the clade is weakly supported. The association. between the latter species and the taxa included from section Batocaulon ser. Stipellares (M. bifurca and M. uliginosa) is also weak. However, this l 8 alter relation may be partially. explained by the absence of interpinnal spicules and the plausible of (Fig. 3 Given that all the species of series rogression o character 2). paraphyllidia | morphology Bipinnatae have striate nerved corollas not found in of Habbasia might hypothesize that it is a natural group. Nevertheless, other members species, one the tendency could not be tested in this study based on the limited number of taxa included. SECTIONS MIMOSA AND CALOTHAMNOS In our phylogenetic hypothesis, the species of sections Mimosa and Calothamnos are derived from a common ancestor (Figs. 1, 2). The position of the Calothamnos species can be explained by the presence of haplostemonous flowers (Fig. 3, character 1) as was proposed by Bentham (1841-1842, 1875). Although previous works (Grether et al., 2005) confirmed the monophyly of section Calothamnos, in our study, the two representatives of section Calothamnos ana are not clustered together. = yzed It may be that the section is that The haplo- underrepresented, but this analysis indicates section Calothamnos is not monophyletic. stemonous species with the indumentum plumose or of stelliform setae were recognized by Bentham (184.1) as section Mimosa ser transferred consisting . Lepidotae and Barneby (1991) to a new section Calothamnos (Table hy 1). However, when the indumen- tum plumose or stelliform setae trait is mapped onto the Annals of the Missouri Botanical Garden ms Mimosa hexandra M.detinens —— M. candollei M- maguirei M. eus M.seto W Maa i M.tweedieana M. strigillosa M.somnians M.bifurca 0 M.uliginosa M.paupera HL M.hirsutissima Morphological characters 4 M.radula Lg M.xanthocentra 1- Flowers: a aa (0) or 1 M.debilis haplostemono aja 2- Para phyliidia: d ME (0), ovate (1), or M.sensibilis subulate/setifor m (2) M. dM 3- Indumentum bulbous or bo O based: so (0) or pr 1) M.obst trigos 4- Craspedium: absent (0) or present (1) M. la 5- Indumentum plumose or stelliform setae: absent (0) or present (1) - flagellaris 6- Virgate stems: absent (0) o present (1) M. diversipila 7- Leaf 1-jugate ai leaflets ed ate : absent o): or pre M.oligophylla 8- Aculei a belok the node: absent (0) bonplandii or prese 1) í . - Xerophilous habit: present (0) or absent (1) M.leimonias 10- Pollen type: bitetrad (0) or tetrad (1) M.balansae 1 M.polycarpa Figure 3. Optimization of 10 morphological characters over the preferred topology. Changes are represented by boxes Above branches, numbers indicate the characters according to the list, while the trait states are indicated below tree, a parallel derivation might also be considered (Fig. 3, are of limited phylogenetic value. character 5), indicating that these characters Mimosa pilulifera (sect. Calothamnos) constitutes a clade with M. Obstrigosae (Benth.) ct. Calothamnos) is clustered. with (Fig. 2, BS = 71%) (ser. sae (Benth.) Barneby). pected because these species exhibit different morpho- Mimosa subser. and M. M. leimonias obstrigosa (ser. Barneby), bonplandii Mimosa subser. Pedunculo- The yielded groups are unex- logical characteristics, such as the type of indumentum D J | branched trichomes for M. pilulifera and M. bonplandii. H l I retrorsel y apressed setae for M. obstrigosa, and glabrous to setulose-ciliolate for M. leimonias). life form (shrub in M. pilulifera, M. bonplandit, and M. obstrigosa vs. herb in M. leimonias), division of leaves (pinnae |-jugate in M. leimonias, M. pilulifera, and M. obstrigosa vs. pinnae 2- to 9-jugate in M. bonplandii), and androecium type and color of the filament (monadelphous and pale - yellow in M. bonplandii vs. filaments free to the base and pink in the remaining species). The results suggest that these traits also exhibit low value from a phylogenetic standpoint. As pointed out in the introduction, Mimosa diversi- pila shares the calyx and pubescence types with species of section Calothamnos, and has the virgate life form present in section Mimosa ser. Mimosa subser. 1991: 708—710). In cladograms, M. diversipila is clustered with M. oligophylla from ma ipedes (Barneby, = section Mimosa ser. Mimosa subser. Pedunculosae. Although t expected because both show the simple virgate habit — iis association is weak, it can be partially from the remaining species from subseries 3, character 6). absent Pedunculosae analyzed here (Fig. SERIES AND SUBSERIES OF SECTION MIMOSA = Two well-supported groups are obtained: in the first, the species of subseries Mimosa (M. debilis and M. sensibilis) are clustered. together in accordance with the Bentham (1875) (series Sensitivae Benth.) and Barneby (1991) subdivision based on the pinna l- Volume 95, Number 4 2008 Bessega et al. Phylogeny of Mimosa Table 4 List of Mimosa species sequenced for this study. Taxa are organized according to Barneby's classification (1991). Barneby (1991) Species Sect. Sect. Habbasia ser. Bipinnatae Habbasia ser. Habbasia Habbasia Habbasia . Pachycarpae . Habbasia ser. . Habbasta ser. Sect. Habbasia ser Sect. Habbasia ser. Pachycarpae Sect. Sect. Habbasia ser. Setosae Batocaulon ser. Stipellares Sect. Batocaulon ser. Stipellares Sect. 9 el. Batocaulon ser. Batocaulon ser. Farinosae Bimucronatae . Mimosa ser. Mimosa subser. Brevipedes " cl. Mimosa ser. Mimosa subser. Hirsutae . Mimosa ser. Mimosa subser. Mimosa "e t Mimosa ser. Mimosa subser. Mimosa Sect. Mimosa ser. Mimosa subser. Obstrigosae Sect. Mimosa ser. Mimosa subser. Pedunculosae Sect. Mimosa ser. Mimosa subser. Pedunculosae Sect. Mimosa ser. Mimosa subser. Pedunculosae Sect. Mimosa ser. Mimosa subser. Pedunculosae Sect. Mimosa ser. Mimosa subser. Polycarpae Sect. Mimosa ser. Mimosa subser. Polycarpae Sect. Mimosa ser. Mimosa subser. Polycephalae Sect. Mimosa ser. Mimosa subser. Pudi Mimosa ser. Mimosa subser. Reptantes Sect. Ç xus 108 Sect. Calothamnos Sect. Batocaulon ser. Quadrivalves Mimosa somnians Humb. & pun ex Willd. & A. Mimosa tweedieana Barneby ex P a & Mackinder Mimosa strigillosa Torr. Mimosa pigra Mimosa adenotricha Benth. Mimosa maguirei Barneby Mimosa setosa Benth. var. setosa Mimosa uliginosa Chodat & Hassl. Mimosa bifurca Benth. Mimosa detinens Benth. Mimosa hexandra Micheli Mimosa diversipila Micheli var. diversipila Mimosa hirsutissima Mart. Mimosa debilis Humb. & Bonpl. Mimosa sensibil ex Willd. is Griseb. Mimosa obstrigosa Burkart Mimosa brevipetiolata Burkart Mimosa flagellaris Benth. imosa leimonias Barneby & Fortunato Mina oligophylla Micheli Mimosa balansae Micheli Mimosa polycarpa Kunth Mimosa radula Benth. Mimosa xanthocentra Mart. var. xanthocentra Mimosa paupera Benth. Mimosa ida (Gillies ex Hook. & Arn.) Benth. Mimosa pilulifera Benth Mimosa candollei R. Grether jugate and leaflet 2-jugate present in all members of subseries Mimosa (Fig. 3, character 7). The second eroup is formed by M. balansae and M. polycarpa, both from subseries Polycarpae Barneby, which share some common traits, in particular the aculei at or below the nodes (Fig. 3, character 8). Because the species of subseries Polycarpae are at the tip in the molecular tree, it is suggested that they are the most recently differentiated group. However, the position obtained for the species of with section Mimosa does not agree completely Barneby’s classification. Examples of these discor- ) I dances are the groupings of M. hirsutissima and M. radula that constitute a highly supported clade (Fig. 2, BS = 100%), although the first one belongs to eds Hirsutae (Benth.) Barneby and the second Table 4). The species from subseries Pedunculosae (M. flagellaris, to subseries Polycephalae (Fig. 2, oligophylla, M. brevipetiolata, and M. leimonias) are nol clustered, which shows that this subseries is an artificial group (Figs. 1, 2). The low support of some clades confirms the idea that Mimosa is a recently diverged genus. Low divergence is expected in neutral markers, such as cpDNA introns and intergenic spacers that would not have had enough time to be fixed in different species and yield poorly supported nodes (Soltis & Soltis, 1998). Alternative explanations that cannot be ruled out are ancestral polymorphism in cpDNA or past introgression events. The latter, combined with strong directional selection in certain environments creating flow postulated by Orr and Smith (1998), seems also to morphological «divergence despite gene as be a plausible explanation for Mimosa. In the center of diversity in south Amazonia in Brazil and adjacent areas in Paraguay, Argentina, and Uruguay, high hybridization among species from sections Batocaulon and Mimosa is described (Morales et al. 2007: Fortunato, unpublished data). The occurrence of natural hybrids followed by stabilization processes is necessary in speciation by hybridization (Grant, 1971), so the hybrids in the wild can be interpreted as evidence to suggest that interspecific hybridization plays an important role in the evolution of the genus levels ca. 30 species (Withus & 1981; Seijo, 1993, 1999, 2007). Morphological, cytoge- and molecular criteria indicated that hybrid- Mimosa. Moreover, different ploidy described in Mimosa ii Berger, 1947; Goldblatt, 2000; Morales et al., netic, ization and polypoloidy processes have been impor- Annals of the Missouri Botanical Garden tant mechanisms in the evolution of (Hughes et al., 2002). importance of each process that could explain the other legumes However, a studv on the relative evolution of Mimosa is still lacking. FINAL REMARKS In this paper, we tested the monophyly of Mimosa and determined that it is a genus derived from Piptadenia viridiflora. Recently, Jobson and Luckow (2007) suggested that P. viridiflora should be removed from the genus Piptadenia based on chloroplast genes trnL-trnF and tri K/matk. from section Mimadenia were not Although re Buh s included i ls study, representatives from sections a Habbasta, single clade. Mimosa, and Calothamnos constitute a Moreover, the sections analyzed here, accepted by Barneby (1991), are not monophyletic; in this regard, the only species that constitute a well- supported clade are those from sections Mimosa and Calothamnos. The most recent group. exhibits. the derived morphological features, as was proposed by (1991). interpreted as a progressive reduction—from diploste- Barneby The evolutionary. sequence. can be monous and free filaments to haplostemonous and an androecium with formula, in habit life form (trees and shrubs to herb). and in pollen type (bitetrads to tetrads). However, in order to determine whether this molecular hypothesis is congruent with Barneby's proposal for relationships, a study adding representatives from Mimadenia is needed. To conclude, ours is the first documented molecular The cpDNA regions examined here provide sufficient information to resolve study of the genus Mimosa. many relationships within the genus. This hypothesis, based on tral and trnL-trnF intergenic spacer regions, constitutes a starting point for further analysis. It will be interesting to see whether more exhaustive sampling and additional genetic evidence (1.e.. nuclear evidence or other cpDNA regions) will support the working hypothesis developed in this study. Literature. Cited Dransfield & shy logenetic K of the palm Asmussen, S. C. Barrow, J. 1999. A | des (Palmae) based on e pl ist DNA sequences from the trnl, Sy s -o Es W.J., C.B He 7 d lerson. uid region. Pl. Syst. Evo Qc v 1 , Sensitivae censitae the panua Me sa Linnaeus (Mimosac eae) in the New . Mem. New York Bot. Gard. € ——. 1993. ceae) from South America. NT A 8-332. Bayer, R. C. F. Puttock ¢ A. 2000. Phylogeny of South African ds (Asteraceae) 835. Increments to the genus Mimosa (Mimosa- Ke hne vi n oe Ji on 1 noncoding chloroplast sequences. Amer 59-272 monadelphous filaments—in leaf V description of Bentham, G. 1841-1842. Notes on wu ue with short ynopsis of species. J. Bol. (Hooker) 4: 243-418. ——. 18 Num of the e Mimoseae. Trans. Linn. Soc. ndn 30: 335-604. . 1876. Mimosa & Schrankia. Pp. 296—391 in C. F. P. von a (editor), Flora Brasiliensis, Vol. 15. PL 2. Miine £i "n Bohle. L . Milger. R. Cerff € W. 1994. Non- coding c a DNA for plant cum systematics : the infrageneric level. Pp. 391403 in Streit, G. P. Wagner & R. Desalle Ecology and Evolution: Approaches and Birkhauser, Basel. Britton; N. L. & J. N. Rose. 1928, Pleromimosa, Veomimosa, ea Mim lOS Opsis, Acant hop- ps. ud J. H. Barnhart (editor). 3. New York Botanical Martin. B. se hier water, D. Molec P ar \pplications. (editors). Leptoglottis, Mimosa, teron. Haitimimosa. Pp. | North Ame | Flora “Vol 1948. Las especies de Mimosa de la flora Wreentina. Darwiniana 8: 9-23 1, Caccavari, M. A. 1985. Granos de polen E Leguminosas de es mnie IV. Genero Mimosa. Bol. Argent. Bot. 24: 1-16 m E studio de los caracteres del polen en las Vino pate Pollen & Spores 28: 29-42 — ———. 1980bh. Nuevos aportes al conocimiento de las Letrades pu prs (Leguminosae, Mimosoidae). Bol. Soc. Argent. Bot. 24: 333-342. —— 1989. l i raestructura del polen de ende Leguminosae). Pollen 3 206 Mimosa & Spores 30: 275- Clegg hie T... BOS. Gaut. G. H. Learn Jr. & B. R. Mort 1994. Rates and eie rns of chloroplast DNA li pn Natl. Acad, Sei. U.S.A. 91: 6795 ge J. & J. L. Doyle. 1987. A rapid DNA isolation quantities of fresh leaf . Soc. Amer. 19: 11-15 1999. Una especie nueva del s Chaco boliviano: V. craspedisetosa Fortunato & Palese. Contribu- DO” oeni ure from ET tissues. Phytochem. i (Fabaceae-Mimoseae) para e ción al estudio de la flora y vegetación del Chaco. XUL Candollea 54(1): 83-87. a & P. Paberlet. 1994. The use of e pun ad DNA to resolve plant phylogenies: E ersus rbe s quences. Molec. Biol. Evol. 11: 277. Goldblatt, P. 1981. Cytology and EN e iy of Legumi- Pp. 127—164 in R. M. Polhill N D. a Raven E Gielly. | Advances in Le gume Syste matics. Vol. Gardens. Kew. Goloboff, P. A. 1998. PiWe/NONA. Manual and program distributed by the author: / phylogeny/Nona-PeeW « > > > ~ = zE vi a yaoshanensis Yen C. P. H. Huang eL. Litsea L. glutinosa (Lour.) C. B. Rob. L. cubeba (Lour.) Pers. L. kingii Hook. f. L. rubescens Lecomte L. sericea (Wall. ex Nees) Hook. f. Neolitsea (Benth.) Merr. N. aurata (Hayata) Koidz. var. aurata N. aurata var. chekiangensis (Nakai) Yen C. Yang & P. H. Huang N. brassii C. K. Allen N. cambodiana Lecomte var. glabra C. K. Allen N. dus H. W. N. chuii Merr. N. confertifolia (Hemsl.) Merr. N. dealbata (R. Br.) Merr. . K. Allen N. kwangsiensis H. Liu N. levinei Merr. N. lunglingensis H. N. ovatifolia Yen C. pS ang N P. H Huang var. ovatifolia N. homilantha C V. ovatifolia var. puberula Yen C. Yang & P. H. Huang N. pallens (D. Don) Momiy. & H. Hara N. phanerophlebia Merr. N. pingbienensis Yen C. Yang & P luang N. pinninervis Yen C. Yang & P. H. Huang V. polycarpa H. Liu N. pulchella (Meisn.) Merr. N. sericea (Blume) Koidz. Neolitsea sp. N. sutchuanensis Y. C. Yang var. sutchuanensis N. sutchuanensis var. gongshanensis H. W. Li N. undulatifolia (H. Lév.) C. K. Allen 1. Tomingodaphne (Blume) Hook. f. Li J. 2002199 coe C) Li H.-W. 19 (HITB Li J. 2002146 (HITBC) Nia Y.-M. s.n. (HITBC) Nia Y.-M. s.n. (HITBC) Nia Y.-M. s.n. (HITBC) Li J. 2002028 (HITBC) Li J. 2002198 (HITBC) Li H.-W. 21 (HITBC) Li H.-W. 28 (HITBC) Li J. 2002170 (HITBC) Li J. 2002094 (HITBC Li H. et al. 15299 (KUN 0789681) Li J. 2002181 (HITBC) Zhang S.-Y. 5482 (KUN 0162041) Gray, B. 05911 (KUN 0793628) Li X.-G. 202474 (IBK 00009945) Wu S.-G. 7095 (KUN 0106438) Li J. 2002065 (HITBC) Xi X.-Y. 414 (PE 1272040) Gray, B. 03993 (KUN 0793630) Li J. 2002071 (HITBC) Wu S.-J. B (IBK 00010186) Li H.-W. 29 (HITBC) Li J. 2002058 (HITBC Wu S.-J. 3246 (IBK 00010360) Mao P.-Y. 03875 (KUN 0108307) Qinhai-Tibet Exped. 5972 (KUN 0108358) Deng L. 7511 (KUN 0108338) Mao P.-Y. 04139 (KUN 0108220) Li J. 2002187 (HITBC) Zhou Z.-K. et al. EXLS-0252 (KUN 0695075 Li J. 2002105 (HITBC K. TA 2180 (KUN 0108215 Li J. 2002070 (HITBC Zhao Z.-X. f (KUN 0108178) Feng G.-M. 6987 (KUN 0108134) Li J. 2002203 (HITBC) ) Yunnan, China Guangxi, China Chin: Yunnan, c Yunnan, Yunnan, China Yunnan, China Guangxi, China Yunnan, China Yunnan, China Guangxi, China Yunnan, China Yunnan, China Guangxi, China Zhejiang, China Queensland, Australia Guangdong, China Yunnan, China Yunnan, China Hunan, China Queensland, Australia Yunnan, China Hong Kong, China Yunnan, China Yunnan, China Hong Kong, China Yunnan, China Tibet, China Guangdong, China Yunnan, China Guangxi, China Yunnan, China ;uangxi, China o Japan Yunnan, China Sichuan, China Yunnan, China Guangxi, China DQ120532 DQ120533 DQ120534 DQ120535 DQ120536 DQ120537 DQ120538 AY934883 DQ 120523 DQ120524 DQ120525 DQ120526 DQ120557 DQ120558 DQ120559 DQ 120560 DQ120561 DQ120562 DQ120563 DQ 120564 DQO 120565 DO120566 AY934884. DQ120567 DQ120568 DQ120569 DQ120570 DQ12057 DQ120572 DQ120573 DQ120574 DQ120575 DQ120576 DQ120581 DQ120577 DQ120578 DQ120579 AY 2054.05 DQ120606 DO 120607 DO 120608 DO 120609 DQ120610 DQI200611 AY265403 AY265402 DQ120599 DQ120600 DQ120601 DQ124270 DQ124271 DQ124272 DQ124273 DQ124274 DQ124275 DQ124276 DQ124277 DQ124278 DO 124270 AY265401 DO 124280 DQ124281 DQ124282 DQ124283 DQ124284 DQ124285 DQ124286 DQ124287 DQ124288 DQ124289 DQ 124294. DQ 124290 DO 124291 DO 124292 Volume 95, Number 4 Li et al. 585 2008 Phylogenetic Relationships of Litsea and Laureae Table 1. Continued. Taxon Voucher Source ETS ITS N. wushanica var. pubens Yen C. Yang Liu L.-H. 15149 (KUN 0162057) Hunan, China DQ120580 DQ124293 & P. H. Huang Parasassafras D. G. Long P. confertiflora (Meisn.) D. G. Long Qian Y.-Y. 682 (KUN 0104558) Yunnan, China AY934885 AY205395 Sinosassafras H. W. Li S. flavinervia (C. K. Allen) H. W. Li Liu Y.-H. s.n. (HITBC) Yunnan, China AY934886 AY940451 Outgroup taxa Cinnamomum pittosporoides Hand.-Mazz. 7 H. 5252 (KUN 0108156) Yunnan, China DQ120554. DQ 124209 Sassafras tzumu (Hemsl.) Hemsl. H.-W. 15 (HITBC) Yunnan, China DQ120555 AY265391 Umbellularia californica (Hook. & van der Werff s.n. (MO) North Ámerica DQ120556 AY265393 Nutt. implemented in PAUP*, and because the PHT was data with resampling using all characters equally, — non-significant, the data sets were combined for all regardless of weight. subsequent analyses. Most parsimonious trees were The 19 morphological characters were analyzed obtained from 10,000 replicates of random taxon using unweighted parsimony, with all characters addition using equally weighted (EW) characters treated as unordered. Tree search was performed with (Fitch, 1971) and tree bisection-reconnection (TBR) 1000 replicates of random taxon addition and TBR branch swapping (MULPARS off), followed by branch swapping (MULPARS on) in PAUP* version swapping on the shortest trees from this analysis with 4.01b10 (Swofford, 1999). The me ‘al data MULPARS on. These trees were then used to re- were then combined with the ITS a S data and weight the characters according to the best fit of their analyzed using the same settings A were used in the rescaled consistency indices (Farris, 1989). New | morphological data set. Clade support was estimated searches were performed with 1000 replicates using with bootstrap resampling (Felsenstein, 1985) for successive weighting (SW) until equilibrium was 1000 replicates with TBR, EW, and MP optimality. reached (Farris, 1969). Clade support was estimated Morphological character state changes were then using bootstrap resampling (Felsenstein, 1985), with plotted on one of the resulting most parsimonious 1000 replicates, TBR, EW, and MP optimality criteria combined analysis trees using MacClade (Maddison € performed on the combined weighted ITS and ETS Maddison, 2000). Table 2. Morphological characters and character states traditionally considered to be taxonomically important at the generic and sectional level in Laureae. Habit: evergreen (0), deciduous (1) Leaf venation: pinninerved (0), triplinerved (1), trinerved (å — l 2. Leaf arrangement: alternate along stems (0), alternate and crowded at branchlet apices (1), verticillate or subverticillate (2) 3 | Inflorescence type: thyrsoid, without vegetative ae bud in the main axis (0), short shoot (brachyblast) with vegetative terminal bud in the main axis (1) 5. Inflorescence: terminal or subterminal (0), axillary (1) 6. Inflorescence arrangement: panicle (0), raceme (1), fasciculate clustered (2 7. Inflorescence: sessile (0), stipitate (1) 8. Flower number per inflorescence: >1(0), 1(1) 0), present (1) 10. Involucres: large (0), minute (1) 9. Involucres: absent I1. Involucres: imbricate (0), decussate (1) 12. Involucres: early deciduous (0), persistent (1 13. Flower sex: bisexual (0). unisexual (1) 14. Basie floral number: dimerous (0), trimerous ~ E — 15. Perianth segment: present, perfect (0), imperfect, absent or early deciduous 16. Anthers: two-locular (0), four-locular (1) T. Pollen sacs of the third whorl: latrorse (0), introrse (1), extrorse (2) 18. Fruit shape: globular or oblate (0), ovoid or ellipsoid (1) 19. Fruit cupule shape: flat or discoid (0), cup-shaped (1) 586 Annals uae Eo nos Garden Bayesian phylogeny reconstruction of the combined data was performed with MrBayes 3.0b4 (Huelsenbeck & Ronquist, 2001; 2002). The program Modeltest (Posada & Crandall, 1998) was used Huelsenbeck et al.. the 56 predicted models of DNA Modeltest indicated that to test which of substitution best fit the data. TN + 14 7 rate heterogeneity and among-site rate variation from Tamura and Nei (1993) best fit the two nrDNA regions, and the ML parameters MrBayes (MB) were ~ inveamma.” The Markov e model with Ist nst = 6” and “rates = vain Monte Carlo process was : 500.000 OO generations, set so that four chains ran simultaneously for generations, with trees sampled every giving a total of 5000 the Likelihood value plots for the four chains showed that trees in initial sample. stationarity had occurred by the 600th tree. Therefore. the first 600 trees were discarded as “burn in.” and the posterior probabilities of the phylogeny and its branches LOO trees. analysis was also performed using neighbor were determined from the remaining 4 ME joining on the unweighted data set using MEGA 3.1 (Kumar et al., 2004), with bootstrap support calculat- ed on 10,000 replicates. ML analysis of the combined, performed using the DNAML option in DAMBE version 4.13 (Xia, 2000: Xia € Xie, 2001). unweighted data was RESULTS SEQUENCE CHARACTERISTICS IT variable between distantly related taxa. UN regions are quite so only regions that could be aligned unequivocally were used in this analvsis, making our phylogenetic estimates conservative. For the ingroup laxa, the length of the ITS regions. including the 5.85 region, lo 027 venerated a data set of 689 characters of which bp. and the alignment 156 G plus C ranged from 508 (22.04%) were parsimony informative. content ranged from 04.61% to 74.19% Compared to ITS, the approximately 400 bp of the 3 end of the ETS between the ETS-1 internal primer were relatively easy to amplify 185 subunit and the and sequence. For the ingroup taxa, the ETS region varied in M d from 350 to 393 bp. and alignment resulted in a matrix of 393 characters of which 99 (25. 199%) were parsimony ao ith G plus € content ranging from 46.5 to 55.23%. All were submitted to pe (Table sequences and the data matrices for both sequenced regions are available from the primary author upon request. PHYLOGENETIC ANALYSIS SW parsimony. Results of phylogenetic analyses of the ITS and ETS regions performed separately showed no hard incongruences (i.e.. there were no contradictory clades supported by bootstrap greater than 60%: The PHT (Farris et: 1995) indicated that the two regions were congruent (P 0.01: 1990; Cunningham, Because of data not shown). discussions in Sullivan. 2000). ITS unit see 1997; Farris et « ETS and transcriptional this and because the regions occur within the same and show evidence of a similar and interdependent role in the 1997), we will only present results for the combined analyses. The EW analysis produced 30,600 trees (length [L] = 1171, 0.306. retention index [RI] = 0.589). all all minor collapsed under strict consensus, SM to stabilize the tree topology. SW reduced this to 12 trees (L = 211.5, Cl = 0.576, RI = 0.834). producing an almost fully maturation of ribosomal RNAs (Good et al.. consistency index [Cl] = but because main and virtually branches was used to try resolved strict consensus tree (Fig. The SW analysis of the nrDNA data produced series of clades within a monophyletic core Laureae (93% bootstrap support). These clades are referred to informally as the Neolitsea—Actinodaphne, and Litsea, bd and Aperula clades. These sat above a basal Palminer- grade of Lindera obtusiloba Blume (sect. then Lindera communis Hemsl. (sect. The Aperula clade contained the three sampled species of Lindera sect. Aperula (L. latifolia, K. Allen. and £L. metcalfiana C. viae) xd Polyadenia). L. longi dun ulata C. K. Allen) grouped in a terminal pairing with Litsea cubeba and L. kingii Hook. f. (both section Tomingo- daphne) above a subclade of Lindera megaphylla sect. Cupuliformes) and Actinodaphne forrestii (C. K. Allen Although the whole was unsupported (< 50% Kosterm. Aperula clade as a bootstrap). all the > 70%) us branches within it had moderate lo strong support (> 909). Sitting above Sinosassafras and Parasassafras was (15% bootstrap support). with two the Lindera clade subclades. The first of these represented Lindera sect. Lindera (2 spp.). plus Litsea species from sections Tomingodaphne and Conodaphne (part). Sister to this was a subclade consisting of L. fruticosa Hemsl. (sect. Sphaerocarpae). Iteadaphne caudata (Nees) H. W. Li. and the three species of Lindera sect. Daphnidium. again with most terminal branches moderately supported, The representing members of sections Litsea, Conodaphne Litsea clade consisted of a terminal lineage (part), and Cylicodaphne, but also including Dodeca- Nees, all Actinodaphne lecomtei C. KR. denia grandiflora siling above Laurus nobilis L. and Allen. The clade included. within it three separate lineages. Section Litsea and Litsea monopetala (Roxb.) Pers. (sect. Conodaphne) formed a strongly supported pair Volume 95, Number 4 2008 Li et al. 587 Phylogenetic Relationships of Litsea and Laureae (99%), sister to a subclade to three species of Litsea sect. Cylicodaphne: L. dilleniifolia P. Y. Pai & P. H. Nees) Huang, L. garrettii Gamble, and L. panamanja Hook. f. (here called Cylicodaphne 1), and then to an L. variabilis Hemsl. var. variabilis and var. oblonga all moderate to strong bootstrap support. The other Lecomte (sect. Conodaphne) subclade, with branch in the Litsea clade was the Cylicodaphne Vl subclade, representing the remainder of Litsea sect. Cylicodaphne but including an embedded Dodecade- nia grandiflora, again with most branches showing a least moderate support. Within the Neolitsea—Actinodaphne clade, Actino- daphne (except A. forrestii and A. lecomtei) formed a with basal grade to a well-supported Neolitsea (89% the latter divided into two subclades (Veolitsea | and Neolitsea M) above N. chrysotricha H. W. Li and N. pallens (D. Don) Momiy. & H. Hara. Neolitsea showed little clear support for the internal branches. all the Neolitsea 1l bootstrap support > 50%. whereas branches in showed Bayesian analysis. Bayesian analysis of the unweighted Laureae ITS + ETS showed moderate relationship resolution (Fig. 2), with the Bayesian tree corresponding well with much of the SW tree in terms of recovered major lineages. Although not as well resolved as the SW topology, terminal SW clades with high bootstrap support were also present in the Bayesian. topology with strong posterior probability support, and both analyses included the Neolitsea— Actinodaphne clade and many of the major SW subclades. Nevertheless. there were differences between the results for the two approaches. In the Bayesian tree, the Lindera clade was not recovered, with section Daphnidium falling instead as part of a polytomy separate from the remainder. Similarly, Litsea gluti- ` nosa (Lour.) C. B. Rob. (sect. Litsea and the type species for the genus) and £. monopetala (sect. Conodaphne) were separated from the rest of the Litsea clade seen in the SW analysis. ME and ML analyses. The ME (neighbor joining) tree (Fig. 3) was well resolved and similar in major clade structure to the SW and Bayesian. cladograms. Neolitsea | and I (albeit the latter reduced) were again terminal above an Actinodaphne grade, although here A. paotingensis Yen C. Yang & P. H. Huang was with N. brassii €. K. Allen, N. wushanica var. pubens Yen C. embedded in a basal Neolitsea subclade Yang & P. H. Huang. and N. sutchuanensis Y. C. Yang. Laurus was basal to the Actinodaphne—Neolitsea clade, and sister to this group was a clade consisting of part of the MP and MB Aperula clade and an expanded Litsea clade including the Litsea and well and H Dodecadenia and Headaphne. Below subclades, as as Cylicodaphne | this was a lineage containing the Lindera, L. fruticosa, and Daphnidium subelades, and then a polytomy consisting of L. communis paired with Sinosassafras, and an “Aperula II? clade of Tomingodaphne, Cupiliformes, and Actinodaphne forrestii. Lindera obtusiloba and then Parasassafras were basal to the remainder of Laureae. md Unlike most of the other analyses, in the ML tree "6g. oc ), Lindera obtusiloba and L. communis were paired and sister to the Aperula clade, with this lineage placed above Laurus and below the Actino- daphne—Neolitsea clade. Neolitsea | and Il and the Actinodaphne grade were all present, although A. paotingensis was placed inside Neolitsea l, similar to the ME tree, and a subclade of N. aurata (Hayata) and N. Veolitsea as a whole. Below all of these was a major Koidz. kwangsiensis H. Liu was basal to lineage consisting of a Lindera clade. L. fruticosa, and the Daphnidium clade (with [teadaphne basal) and a Litsea clade (including Litsea and the Cylicodaphne | and I subclades). Actinodaphne lecomtei was placed well inside the Actinodaphne grade, whereas A. forrestit was still sister to Lindera megaphylla, but this latter pair was placed between Sinosassafras and Parasassafras at the base of the Laureae. Combined molecular and morphological analyses. Analysis of the morphological matrix by itself resulted in 317 equally parsimonious trees (L = 72, Cl 0.3056, RI — 0.8214), but these collapsed completely to an unresolved polytomy under strict consensus. When the morphological and molecular data were combined, two trees of 1271 steps (CI = 0.3021, RI = 0.6086) resulted, one of which is shown in Figure 5. This tree also recovered the Neolitsea—Actinodaphne clade, Lindera clade, and Litsea clade, which were seen in the molecular analysis, but the Aperula clade was now split into two separate entities, with Aperula | representing Lindera sect. Aperula s. str. and Aperula II representing Litsea cubeba and L. kingii (sect. Tomingodaphne), Lindera megaphylla (sect. Cupuli- formes). and Actinodaphne forrestii. The inclusion of the morphological data changed the bootstrap support for most clades, generally lowering it from the SW analysis, although there was still reasonable bootstrap support for the terminal branches and still little or no deep-branch support. When the morphological character state changes are plotted on the combined analysis tree (Fig. 5), the unique synapomorphy for the Laureae is the presence with of inflorescences with short shoots (brachyblasts a vegetative terminal bud on the main axis. Within the Laureae, the Actinodaphne—Neolitsea—Parasassafras— 588 Annals of the Missouri Botanical Garden Table 3. Data matrix of important morphological characters in Laureae for species used in the molecular analyses. m l'axa E © 4 T pii NO n D = Un p OV 17 18 H O Actinodaphne cupularis Actinodaphne Jorre stii Actinodaphne henryi Actinodaphne kweichowensis Actinodaphne lecomtei Actinodaphne obovata Actinodaphne ometensis Actinodaphne paotingensis Actinodaphne pilosa Actinodaphne trichocarpa Actinodaphne tsaii Cinnamomum pittosporoides Dodecadenia grandiflora Iteadaphne caudata Laurus nobilis Lindera chunti Lindera communis Lindera fruticosa Lindera kariensis Lindera latifolia Lindera longipedunculata Lindera megaphylla Lindera metcalfiana Lindera obtusiloba Lindera pulcherrima Lindera reflexa Lindera villipes Litsea aculivena Litsea cubeba Litsea dilleniifolia Litsea elongata Litsea garrettii Litsea glutinosa Litsea kingii Litsea liyuyingi Litsea longistaminata Litsea monopetala Litsea panamanja Litsea rubescens Litsea sericea Litsea umbellata Litsea variabilis var. variabilis Litsea variabilis var. oblonga Litsea yaoshanensis Neolitsea aurata var. aurata Neolitsea aurata var. chekiangensis Neolitsea brassii Neolitsea cambodiana var. glabra Neolitsea chrysotricha Neolitsea chuii Neolitsea confertifolia Neolitsea dealbata Neolitsea homilantha Neolitsea kwangsiensis Neolitsea levinei Neolitsea lunglingensis (QOQOQOQOQOQOQOQOOQOOOQOpmnpbnB|BboooQOIrILnooouoIioormponmooooconpmpbooocooooooooooooodocdo!|mnm QO ND mLB|PRPFHHMPKPBPOP'PKPPOCOOOOOOOOOOOOOOOFTdGAOOOOOFTnGOOOOOOOOOCGOCHP.hF.&PK.F.CDOoNBKBON PRNOPROPRO|RP2 I2 p pp pDB ppB QOO L2 2 QI E :- .O OO OQ OO OO OO QOO OO QOO ODOPKOQOhNPFOOOOQOLP|O]QM.:ornPorPcoooococonmPooodocccojtu QOOOOoOooOooOooOoOoOoOooOnm[pmmiPÀÁnrÁrnpr|bprHrpnpnrpnrnm£ÓBBmmpgpgEZFHSmppnbHnpBrp|brinhprprnLnppmgrpbÍbLDbctczcooococococoocoooonmnoco!|ne-e 5 1 1 i 1 1 1 1 iu di 1 1 0 d 1 1 1 1 T T 1 1 1 1 ii 1 1 1 1 1 1 id 1 i 1 de 1 1 1 ji uD I 1 E 1 1 I 1 1 I L 1 L 1 it 1 1 6 2 2 1 2 2 1 2 2 O 2 2 0 É 1 1 2 2 1 2 1 1 1 i 2 2 2 2 2 2 1 2 T 1 2 1 1 2 1 P P 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Om|bOOOOoOoOoo0]0nmnbnbpbmbonmnmbprpprmupwpunLiunhroppnuL;iiormbomnnpmnumnD.gpomnbmbormnmmnboomnooccomooomooonmooo!'- QOQOOOoOoOOOOOOoOoOoOoocoooooooooooodoodoooccooocoooooooomnm[Dbmnmooocooocoooocoodo!/!|co a O 00000000000 OOOOOoOo0o0o0o0000000000o0000o0o0o0ooooooooooocoooooooooc0o!|! OOOOOoOoOoOoOoOooGOo rrerrrrrrerrrrrrrrerrrrrrrrrrrrrrrrrrrrrrrrrrrr | OO OO CO CO OO O joe I2 pO B pp. p. p. B. p. p. p. B. B. B. EPR iB. iS. i. pp. po. B. i. po. po. pp. p. o. i. ipo. pp. pp. pp. p e eo e e e p eo po. p. ja | PRPRPRPERPEPEP RPE REE PERE PE PEP PEP PEPE ppp PEP SO ops jo pp pops DODDODCDODDODDOOOBRBHEEPBE EPP PEP PEP PEP PEP PEEP PEEP POPPE EEE EEE EP jp pa QOOOOOOOOOOOoOooocooocomnoocomnooocooooooocooccoocooococooocoocoocoococoocoocooocoo PRPE RPP EE PE EPP PEPE REPRE RE REPRE REF ODDODOCOOOOCOCOOBRBRBRE EEE HP ERE EF, PRP pp pp. B. B. p. p. pp. B. B. HEHEHE iS. i3. p. pp jp po. po. pp ja jp p. pp B. E. E. E. io. io. io. io ip. E. CO (CO PO E. i o. p. PR p e o i. iS. rS E|oomb|comn|nmnbmbogocoomnbmn|BococoooomnemnmneuoocomenmecoonrnebmnormermneomnbococoomneoconrnmoemnPmbPbomnbomnmocuomnb|bmnooonmnmn QOOOOOoOoOoOoOOOOOnLnoomnoonooc*x-oomnmmnmmeomnpoocooomnoocoooocoooniÍmmpnmoomnpbpoomnbpmpmpbmpmrmp Volume 95, Number 4 2008 Li et al. 589 Phylogenetic Relationships of Litsea and Laureae Table 3. Continued. Taxa | O rp E n N pas [vv LA D m Ui = Oo p ~J p. ee) | Ie) Neolitsea ovatifolia var. ovatifolia Neolitsea Mo ie var. puberula Neolitsea paller 3 ta Neolitsea phanerophlebia Neolitsea pingbienensis Neolitsea pinninervis ea polycarpa ulchella Neolitsea sericea Neolitsea Neolitsea sutchuanensis var. Neolits Nedlite ^a pu a sp. OOOooOooOooooooo!|nse LpbÁbmmnmppmpppopropopmbnmpcmdbpbdTIu NNNNNNNNnNnn ao 3 1 i i 1 I O 1 1 1 L 1 00000000000 CRPORPRPRPEPNOPFPE|N sutchuanensis © Neolitsea sutchuanensis var. ©1012 gongshaner Neolitsea indus lia Neolitsea wushanica var. pubens Parasa iind i Sassafras Sinosassa a "A avinervia OO OO: 2 Oo I 00012 O 1 111 10001 OL td 2 0 0-0 1 1 Umbellularia californica 0000 rroocooool|-J © PREROO 20000o0o0ooooooolo berrea 0 200000000000 here ppp PRP RRP PP Pepe PRP RP PPP ppp OOOooooooooo OOOoooooooo0o0 PREP RPP RPP PEE PREP P PEPE EE PROOFOFGDGDCO oooooo0oo0ooo0oo Oo p O = = = O O = = = o 0 O O O O 0 |BDbBombpbmnmn On|on|ooc Ooooonmnpnmn oOoomn|omnmo On|bmbmpbmpma PRREROO 000000 Orr rre NOom|onmbnbn í|Doooomu ooooonmn Sinosassafras clade is united by the presence of imbricate, early deciduous involucral bracts, although Neolitsea itself has decussate persistent bracts as one of its synapomorphies. Parasassafras and Sinosassa- fras are supported as a pair by the possession of triplinerved leaves, minute involucral bracts, and latrorse pollen sacs for the third whorl, whereas the monophyly of the Neolitsea—Actinodaphne clade is supported by the synapomorphies of verticillate or subverticillate leaves and thyrsoid inflorescences that ack a terminal bud. Although there are few characters supporting the main branches within the Actinodaphne grade, Neolitsea is well supported, with triplinerved, alternate, crowded leaves at branch apices, decussate Within Neolitsea, there were no morphological synapomor- persistent bracts, and dimerous flowers. phies supporting clades I and II, but within I, the & P. H. (Fig. 5la) have verticillate or species above N. ovatifolia Yen C. Yang Huang var. ovatifolia subverticillate leaves as a synapomorphy (albeit with some reversals), whereas in clade II there are two lineages: Ila, with ovoid to ellipsoid fruits, and IIb, defined by alternate leaves (again with a reversal in V. brassii). In contrast, there were relatively no morphological synapomorphies supporting the Lindera or Aperula I or II clades, although the Laurus—Litsea clade had racemose inflorescences, and the Litsea clade was defined by four-locular anthers and the L. glutinosa— L. monopetala pair by the possession of imperfect, absent or early deciduous perianth segments. Within the Litsea clade, the Cylicodaphne II subclade was characterized by cup-shaped cupules and united with Dodecadenia on fasciculate, clustered inflorescences the Lindera clade, section Daphnidium was united by and ovoid/ellipsoid fruits. Similarly, within trinerved leaves and ovoid-ellipsoid fruits. DISCUSSION PHYLOGENETIC UTILITY OF ETS IN THE LAUREAE Several studies have shown that greater resolution and support for phylogenetic estimation are achieved by increasing character number and/or taxon repre- sentation (Graybeal, 1998; Hillis, 1998; Soltis et al., 1998; Bremer et al., 1999), and our study supports the importance of adding data from both more taxa and more sequence regions to help resolve issues in i of the molecular systematics has been suggested previously Baldwin & Markos, 1998; Bena et al, 19 Clevinger & Panero, 2000; Linder et al., 2000), although the sequenced ETS segment is maeh shorter Laureae. The usefulness S region in — in. than the segment in the ITS (393 bp vs. 689 aligned bp) in Laureae, it nevertheless produces a slightly higher percentage of informative sites (25.1996) than that in the ITS (22.64%). The PHT for the two data sets showed congruence, and the combined analysis of the ITS and ETS sequences provided greater resolu- tion and increased support for the relationships than either sequence by itself (trees not shown). This is consistent with the results of other combined ITS/ETS investigations (e.g., Li et al., 2002; Becerra, 2003; Lee et al., 2003; Morgan, 2003; Saar et al., 2003; Urbatsch Annals of the Missouri Botanical Garden eol sea cambodiana var roel e sea dealb sea a prenerien ulchel s pi sea levinel tifoli. ovatifolia sea bras sea a jos var. aur. ea aurata var choklangensis Neolitsea/Actinodaphne Clade 52 ERE 55588 | bes lunglingensis sea confertifolia litsea undulatifolia sea chuii Iitsea sutchuanensis var DES sea wushanica v litsea pingbi shale sea carpa sea homilantha sea Sp. sea pinninervis litsea chrysotricha allens nodaphne paotingensis nodaphne cupularis nodaphne kweichowensis nodaph nodaphne trichocarpa ne pilosa 3 3 Litsea Clade ' Lindera Clade ' rode obovata sea gari sea Parra a sea dillenifolia sea var rabies var. die sea "arabi llis blonga Figure 1. Striet consensus from 12 sequence data. Bootstrap values Panero. 2004; il. Plovanich & Friar, 2003: 2004). el c RELATIONSHIPS AMONG MAJOR CLADES Laureae are a well-supported monophyletic group based on our molecular data. and Roalson & they trees resulting for Laureae alter succes: rarely (2004), subgroups recognized i share £ € ss greater than 50% are indicated on branches. The r and t racemo: Pa s confertiflora S roses flavinervia ndera latifolia Y w 3 22 © in aphylla eee forrestii indera communis indera Ena assafras diia a pom amomum pittosporoides Ive weighting of the con se ` panic ulate inv a ol to the ITS + matk phylogeny of Li et clades that he four n stu litsea sutchuanensis var. sutchuanensis our analyses are, | 295/0928 PeSIIOSN hanensis l| 2as/09N Actinodaphne Grade Cylicodaphne | 7] Litsea Cylicodaphne |l 8 Dodecadenia Actinodaphne Laurus | Daphnidium Iteadaphne Sphaerocarpae Lindera E Sinosassafra Aperula ] Tomingodaphne Palminerviae [ouoroup ibined ITS and ETS morphological a of pseudo-umbellate, olucrate inflores- in general, al. at Laureae and other dy more or less correspond to clades recovered here. Increased taxon Volume 95, Number 4 Li et al. 591 2008 Phylogenetic Relationships of Litsea and Laureae ded lycarpa 7] 7] ea roll es r4 antha o "e SUN nensis var. sufchuanensis > sea ssuchiaenas v var. Vau iE o itsea wushanica v = sea undulatifola 2 sea confertifolia & olitsea chuii = fitsea pinni i > sea lunglingensis D litsea ch nd sotricha 2 eo == [^] olitsea ea phanerophiebla Es sea pulchella z levine! D sea pi var. ovatifolia e olítsea brassii pd D ltsea cambodiana var. ed T sea ovatifol r. pub sea dealbata P (sea kw wangsiensis = nodaphne cupularis odaphne OR Actinodaphne/ -— ictinodaphne ome . ly” Actinodaphne gioca a Neolitsea Clade | Ema i bc Actinodaphne | | Actinodaphne tsai Grade Cylicodaphne | tsea a Litsea I| Clade 56 esr i x s z a Cylicodaphne II Lits - L sion vill Dodecadenia pa a ne lecomtei Actinodaphne lteadaphne cau at Iteadaphne n hash tifolia 109] E ¿naera Pep ui Aperula 34 Lindera metcalfiana 100 cu " 7 DUI He : Litsea kingii ] Tomingodaphne t00 Lindera megaphylla Cupuliformes / dio ls elle Actinodaphne u Lindera kar =] - Litsea eee Lindera Clade oft Litsea umbellata Lindera Lur ndera reflexa 100 7 E on = — Ai 100 Lindera pulcherrima Daphnidium Clade sm Lindera villi Daphnidium Litsea | Clade ee = 5 100 - Litsea glutinosa Hay L Litsea a petala {Litsea Tir € eo UT Sinosassafras E ds 2s us tag Lindera fruticosa Sphaerd carpae Pi fertifl Parasassafras Lindera obtusiloba Palminerviae aliforn Outgroup Sassafras tzumu Figure 2. Phylogenetic majority rule tree for Laureae from Bayesian inference analysis of the unweighted combined ITS and ETS sequence data showing posterior probabilities on branche sampling as well as the additional sequence data from the ETS region and morphology have largely improved bootstrap, but mainly in more derived regions, with deep branch relationships still unresolved. This suggests that although main lineages are being identified and supported, the these clades are unstable under different analysis procedures (e.g., Lindera =— aureae; the isolated, gener the ally basal position of L. obtusiloba basal to positions of many of communis; and the clades Lindera and Aperula less derived than the Litsea, Laurus, and Actinodaphne— Veolitsea clades). The precise relationships between major clades are, thus, still uncertain, suggesting that the sequence regions are evolving too rapidly to provide 392 Annals eon Aa Garden 96) Neolitsea aurata va du "T. cg Neolitsea aurata var chekiangensis olitsea cam aoe var TA Hin olitsea ovatifolia var. puber 2 Neolitsea dealba d i Neolitsea a ang iensis = ea 2ovatioa var. ovatifolia 2 an c E vinel ex bep phanerophlebia itsea jn aff d > 85r- Neolitsea ung! ingens E] pe Neolifsea pallens S HA TA chrysotricha 8 Neo n sea er a Neolitsea chu! — Neoltsea vague tifolia 2 Neolitsea confertifolia oO Neolitsea homik in 2 sea polyca a Neol E ea hmmm c Neolitse = pes as paotingensis Actinodaphne COIHISEd Drassi r pubens Neolit itsea sutoh 'uanensis var. sutchuanensis gongshanensis E . Actin odaohnehe nryi Actinodaphne/ => = obovata Neolitsea Clade Actir it dad Isa lr Act tinodaphne pilo Actinodaphne t Actinodaphne cupularis Grade Actinodaphne trichocarpa U Acunn phne kweichowensis || t phne lecomt - urus nobilis Laurus Lindera longipedunculata Aperula | 98 Lindera metcalfiana Aperula Clade = Lindera latifolia > Litsea acutiven , Lits oshanensis Cylicodaphne I Litsea € " " Dodecadenia ga Pind D odecadenia Litsea ly, SS ey Litsea Cl ade sea monopetala d UNIES E lteadaphne M Lindera . Litsea fiyuying! Lits Sa EE aminata. a variabilis var. variabilis Litsea variabilis vat ona Lifsea dillenifolia 99) Litsea garrettii ! Litsea panamanja Lindera a a Linder. s B 74 Litsea umbellata 50 Sinosassafras flavinervia CINECA COMM sia megaphylla TORIS forresti 99 (oi sea kingii Parasassafras eire lora CINU a ob Cinnamomum pittosporoides L—3À4 0.005 Figure. 3. Sassafras CZ UTIL Umbellularia cal EIS Cylicodaphne | Daphnidium Sphaerocarpae Lindera Sinosassafras Outgroup Minimum evolution (neighbor joining) tree for Laureae of the unweighted combined ITS and ETS sequence data. Bootstrap values (10,000 replicates) receiving support greater than 50% are indicated on branches. deep-branch resolution; and sequences from slower- of evolving gene regions may be needed to stabilize and support relationships between the major clades. However. despite this caveat. our study representative laxa from all genera and most sections of L aureae, and the resulting clades support the hypothesis includes Li et al. (2004) that inflorescence features and ontogeny are important for helping to understand evolution and improve classification within the Laureae. The Sinosassafras flavinervia and Lindera communis (sect. Polyadenia) in the MB and ME analyses moderately supported association. between also suggests Volume 95, Number 4 8 Li et al. 593 Phylogenetic Relationships of Litsea and Laureae eolitsea sutchuanensis var. gongshanensis eolitsea sutchuanensis var. Su EUER puben 3 o sea coral O chuii eolitsea sp. tee eolitsea homilanthe eolitsea a pinn eolits a lnghgens || Basyjoay vatifolia var. puberula se: cambodiana var. glabra sea ee is eolitsea chrysotricha $ eS FEST Jl ovatifolia L Ich I sea pulchella eolitsea phanerophlebia [a 2 EESTE Actinodaphne paotingensis Actinodaphne eolitsea brassii KEOS sea es var. pri lit aurata Ne yitsea proni = Actinodaphne lecomtei Actinodaphne kweichowensis ctinodaphne pilosa phne cupularis Actinodaphne/ Neolitsea Clade oda f , Actinodaphne trichocarpa Actinodaphne odaphne omeiensis Grade anhn fear e isa r Actinodaphne obovata L Aperula Clade Actinodaphne henryi zi] ai Lindera longipedunculata ndera latifolia Aperula Lindera metcalfiana E tsea kingii 7 | È ‘sea ism Tomingodaphne | E rviae Lind: ndera obtusiloba Palmine unis Polyadenia rus nobilis Laurus sea panamanj 7 sea garr LA Litsea dilleniifo Litsea indu is var. oblonga Cylicodaphne | sea variabilis Litsea ease ate 9 sea paces d Litsea acutivena Litsea elongat: Dodecadenia gar fora pele i cadenia 7 Litsea Daphnidium & lteadaphne eadaphne caudata — indera fruti Sphaerocarpae a rubescens ndera reflexa Lindera Clade pei umbellata Lindera l E itsea sericea ndera kariensis f f fertifl Parasassafras r- Actinodaphne forrestii e | L Lindera megaphylla Cupuliform | Sihosassafras [ Sassafras tzumu I Figure 4. Maximum likelihood tree for Laureae of the unweighted combined ITS and ETS sequence data. that the former may not be as closely related to bears several pseudo-umbels on axillary short shoots. Parasassafras as previously suggested (Rohwer, whereas L. communis only has a single or rarely two(to 1993), although these two genera were placed three) pseudo-umbels clustered in leaf axils. Never- successively as basal to the Lindera clade in the SW analysis (albeit without bootstrap support) and formed a clade in the combined molecular/morphology analysis, with three morphological synapomorphies. Sinosassafras flavinervia has a single or two(to three) pseudo-umbels clustered in the leaf axils and always theless. in both taxa, the terminal buds on the axillary short shoots are poorly developed or reduced. The Aperula 1 clade contained two subgroups (none of which actually belong to Lindera sect. Aperula s. str.): (1) Litsea cubeba and L. kingii: and (2) Lindera megaphylla and Actinodaphne forrestii, and three of Figur 05. 5 i+. o 125 999 Sassafras tzumu Outgroup 13 inan 00 Lindera obtusiloba Palminerviae 7 ; Lindera Polyadenia >) omingodaphne Lindera megaphylla 2 Actinodaphne forrestii A Lindera fruticosa oO Lindera pulcherrima 7) Lindera villepes Lindera reflexa Litsea rubescens 119 Litsea umbellata 618 0 1 Litsea sericea Laurus nobilis Lindera metcalfiana Lindera latifolia Cup Actinodaphne Sphaerocarpa Lindera kariensis Litsea liyuyingi uliformes Daphnidium Lindera Laurus Aperula Litsea longistaminata 1 Litsea Clade . oblonga Litsea dilleniifolia of the two trees resulting for Laureae from unweighted combined ITS, ETS, and morphological data sets. Morphological character state changes are also plotted. Numbers above branches are characters: those below are character states from Table 2. Filled circles are unique state changes: open circles represent homoplasious changes. Bootstrap values greater than 50% are indicated in brackets below branches. variabilis Cylicodaphne | Iteadaphne Lise Dodecadenia Cylicodaphne II 20 Actinodaphne/ Neolitsea Clade Parasassafras conferiflora Parasassafras Sinosassafras flavinervia Sinosassafras Actinodaphne henryi 178] ctinodaphne obovata > S6 Actinodaphne tsaii S Actinodaphne trichocarpa à E ctinodaphne pilosa 3 Actinodaphne omeiensis 3 0 Actinodaphne cupularis 0 tinodaphne kweichowensis 0 Actinodaphne paotingensis 3 Actinodaphne lecomtei 0 r— Neolitsea kwangsiensis eolitsea aurata var. aurata Neolitsea aurata var. chekiangensis l 2 3111214 Ver 11110 Neolitsea dealbata 3 1 Neolitsea sericea Neolitsea phanerophlebia Neolitsea sutchuanensis var. sutchuanensis eolitsea sutchuanensis var. gongshanensis | & 63J//O ƏN ll &o35g//0 ƏN [IE USPIBO) [eoiuejog unossilA ves au) Jo sjeuuy Volume 95, Number 4 2008 Li et al. 595 Phylogenetic Relationships of Litsea and Laureae these (Litsea cubeba, Lindera megaphylla, and A. Jorrestit) also formed a clade in the study by Li et al. (2004). Litsea cubeba and L. kingii (sect. Tomingo- daphne) are united by being deciduous, with naked terminal buds, and differ mainly in the absence of leaf pubescence in L kingii (Long, 1984). Li et al. (2004) also noted that there was a close relationship between Lindera (sect. Cupuliformes) and A. megaphylla Jorrestii, and that there were micromorphological grounds for the splitting of Actinodaphne s.l. This is further supported by a preliminary molecular study of Actinodaphne, which found the genus to be polyphy- 2006). Although Li Tsui (1987) had previously letic within Laureae (Li et al., (1985) and Lindera sect. Cupuliformes as being possibly related regarded to Litsea sect. Cylicodaphne because of their similar fruit cupules (despite belonging to different genera). this was not supported by our study, in which section Cylicodaphne itself was polyphyletic. Furthermore, Lindera megaphylla has large leaves aggregated near unusual character in the top of branchleis (an Lindera), making it morphologically more similar to some Actinodaphne species. Lindera megaphylla also = bears a pair of pseudo-umbels on each side of an axillary short shoot with a vegetative terminal bud. whereas A. forrestii has several sessile pseudo-umbels clustered on an axillary short shoot that also produces that L. . forrestii may share a common a vegelalive le ure bud. This suggests megaphylla and . inflorescence ia and could help to explain why A. forrestii was separated from other Actinodaphne species in the analyses of Li et al. (2004, 2006) and the current study. The largely basal and isolated position of Lindera This and its pseudo-umbels are obtusiloba (sect. Palminerviae) is unusual. species is deciduous, borne in an axillary, mixed bud (leaves and inflorescences together) covered by scales. This kind of mixed bud also occurs in Sassafras. According to this character and the later appearance of L. obtusiloba in the fossil record, Tsui (1987) suggested that Miocene, possibly in response to the onset of cooler, Lindera evolved from Sassafras during the more seasonal climates, although this assumes the correct assignment of fossils to extant genera, and generic definition based on vegetative anatomical and morphological features, which is an area of ongoing research. Neolitsea consistently formed a terminal monophy- pi letic lineage within the Neolitsea—Actinodaphne clade, (2004). The present study (with many more taxa) also shows that flowers (apparently a reduction from a trimerous A agreeing with the matK analysis of Li et al. Neolitsea as a clade is defined by dimerous ancestral condition), clustered/verticillate leaf ar- rangement, triplinerved venation, and decussate, persistent. involucral bracts, and that splitting the genus just on leaf venation differences, as suggested as a possibility by Li et al. (2004), is not warranted. ( Neolitsea | and Il) were found in all the analyses, the precise composi- Although two main subelades ( tion of these varied, with some taxa moving between them or to a basal grade position depending on the analysis used, possibly due to what is still a relatively small sample size for such a large genus. Similarly, the absence of previously defined totphelog ry-based sections within Neolitsea makes “representative” sampling more difficult. As they currently stand, neither Neolitsea | nor I has definitive synapomor- differ venation or inflorescence features. Leaf arrangement phies, and they do not consistently in leaf seems to be important for two of the larger subclades. with the species in la being verticillate, and those of IIb being alternate. Similarly, Ha above N. penninervis shows ovoid-ellipsoid fruit, although this feature also occurs in some Neolitsea species from subclades | and Hb. Accordingly, the composition of the clades within Veolitsea and the nature of any supporting morpho- — logical characteristics are the focus of ongoing research. Similarly, the inclusion of Actinodaphne paotingen- the ME and ML This placement may be related to the sis inside Neolitsea by analyses warrants study. instability of the number of floral parts, as there can be six to eight perianth lobes and nine to 15 fertile slamens in this species. The combined molecular and morphological phy- that Actinodaphne, despite previous morphological studies logeny indicates Neolitsea is terminal above suggesting that Neolitsea is closest to Litsea (Koster- mans, 1957; Hyland, 1989; Rohwer, 1993). Van der Werff (2001) that the inflorescence of Actinodaphne is unlike that of other Laureae, and noticed A. pilosa (Lour.) Merr., A. henryi (Nees) Blume) paniculate or racemose inflorescence enclosed. by several species (e.g Gamble, and A. obovata have a dec iduous involucral bracts. In our study, the Neolitsea-Actinod pl imbricate, early clade was defined by the possession of verticillate leaves (later becoming terminally clustered in Neolitsea) and thyrsoid inflorescences lacking a terminal bud. This latter inflorescence feature is very different from the brachyblast-type short shoot seen in Laurus or the Litsea, Lindera, and Aperula clades, which display instead a pseudo-umbellate inflorescence with a vegetative terminal bud and decussate, tardily deciduous involucral bracts that enclose each pseu- Rohwer (1993: fig. 87C—E) suggested that reduction from paniculate or racemose inflorescences led to the do-umbel. sessile pseudo-umbellate inflorescences Annals of the Missouri Botanical Garden seen in Neolitsea, and this agrees with its position as a derived terminal above Actinodaphne. Furthermore, the clustered pseudo-umbels seen in Actinodaphne and Neolitsea differ from those of Litsea or Lindera, as the former are thyrsoid, lack terminal buds, and are clustered in leaf axils. whereas the latter are arranged along leafless short shoots and bear a vegetative terminal bud. Within Neolitsea, the separation of N. ovatifolia var. Yang & P. H. may just ovatifolia from variety puberula Yen C. Huang was unexpected. H reflect the widespread distribution and/or regional differentiation within this species, but certainly warrants further study to clarify species and varietal limits in this taxon. Li (1985) and Tsui (1987) suggested an evolution- ary series for the inflorescences in Litsea and Lindera. and our results concur with their hypotheses. The flowers occur in pseudo-umbels enclosed by decus- sate involucral bracts and are arranged along « leafless axillary short shoot with a terminal bud that can grow into a normal vegetative branch after flowering. A raceme bearing pseudo-umbels arising from normal growth of the peduncles and growth of the internodes in the short shoot was a synapomorphy for the = Laurus—Litsea—Lindera—Aperula | clade, with reduction in several distal branches to create fasciculate clusters. In other. basal groups. the peduncles and internodes of the short shoot are frequently reduced, so that the flowers are arranged in spikelike pseudo-umbels (the sessile inflorescence plesiomorphic in the condition appears to be Laureae). In teadaphne and Dodecadenia, although they follow this basic racemose pattern, the number of flowers per involucre or pseudo-umbel is reduced to one. In both cases, they seem to represent reduced members of otherwise pseudo-umbellate clades. suggesting that the pattern is convergent. Previous morphology-based studies recognized considerable variability in both Litsea and Lindera and variously subdivided them into sections (Ben- tham. 1880: Hooker. 1890: Li et al.. 1984; Tsui. 1987). Hooker (1890) and Li et al. (1984) recognized four sections within Litsea (Litsea, Conodaphne, Cylicodaphne. and Tomingodaphne) based on habit, leaves. floral characters. inflorescences., and fruit cupules. Lindera was similarly divided into eight sections (Lindera, Sphaerocarpae, Palminerviae, Aper- ula, Cupuliformes, Daphnidium, Polyadenia, and by Tsui (1987). Although traditional generic delimitations based on two-celled versus four- Uniumbellatae) celled anthers were not supported by our study, several monophyletic subclades are evident that do correspond, in part, to some of these previously recognized sections, and help to shed light on their character phylogenetic relationships, as well as evolution, within the clades. For example, the Litsea clade was synapomorphie for four-locular anthers, albeit with /teadaphne embedded within it as a reversal. Within the Litsea clade, L. glutinosa (sect. Litsea and type species for the genus) and L. monopetala resolved as A (sect. Conodaphne) were consistently sister taxa, with moderate support in the ME and Bayesian analyses. Although clustered on both molecular as well as combined data, the pair is characterized by a lack or incompleteness of the perianth lobes (absent in L. glutinosa and early deciduous in £L. monopetala) (How. 1956). The Cylicodaphne | and H subelades, correspond- ing to members of Litsea sect. Cyclicodaphne s.l.. were well supported as separate in most of the analyses, and, even in the Bayesian analysis where they formed Cylicodaphne | single clade, was supported and terminal above an unsupported Cylicodaphne Il grade. Although £. variabilis was placed in section Con- odaphne by Li et al. (1984). it member of section Cylicodaphne by Hooker (1890) was treated as a and our results support its return to that section in Cylicodaphne |. Within Cylicodaphne 1, although some species have eight perianth lobes, they all have pseudo-umbellate racemose inflorescences arranged along leafless short shoots, a feature that they share with the Litsea subclade (at least in part) and lieadaphne, although there were no morphological synapomorphies to define Cylicodaphne 1. In Cylico- daphne ll, the species have one to several clustered pseudo-umbels on an axillary short shoot and ovoid- elliptical fruits, supporting the inclusion of Dodeca- denia. Lindera fruticosa (sect. Sphaerocarpae) was related n the SW Suse despite its possession of an umbel with an o [teadaphne and section Daphnidium elongate peduncle and its deciduous habit. Neverthe- ess, it was placed as a less-derived member of the Lindera clade in the ME and ML analyses, with which it shares a deciduous habit and possession of a single or few pseudo-umbels clustered on the short shoot. Its lo the intermediate both basal position may be related situation of leaf venation, as can possess triplinerved and pinninerved leaves. However, as none of our analyses showed bootstrap support for the ts associations must be position of this taxon, regarded with caution for the present, pending further studies. The Lindera clade consists of species from section Lindera and members of Litsea sections Conodaphne and Tomingodaphne. In deciduous habit, Lindera kariensis W. W. Sm. (sect. Lindera) is very like Litsea sericea (Wall. ex Nees) Hook. f. daphne), except for the difference in number of anther (sect. Tomingo- Volume 95, Number 4 Li et al. 597 Phylogenetic Relationships of Litsea and Laureae 2008 cells. Nevertheless, the placement of the evergreen Litsea umbellata (Lour.) Merr. (sect. Conodaphne) within the group is unusual, as all other members of this clade are deciduous. Litsea umbellata also appears unusual within the clade, as its cupule bears persistent tepals. Nevertheless, it is possible thal deciduous species that bear cupules without persis- tent tepals may have evolved from evergreen ancestors with persistent cupule tepals, and in section Con- odaphne s. str., there are species with and without persistent cupule tepals. Similarly, section Tomingo- daphne contains two vegetative terminal shoot bud forms in the axillary short shoots: naked versus scaly = — z 5 o m em , = Richter large portion of what we now (1993) and van der Werff and (1996) cues regard as the Laureae, including Actinodaphne, Do- decadenia, Litsea, Neolitsea, Iteadaphne, Lindera, Parasassafras, and Sinosassafras. Although Sassafras and Umbellularia were also included, they should be removed based on the results of Chanderbali et al. (2001), Li et al. (2004), and the present study. further supported by the possession of racemose This is inflorescences in — Sassafras and the presence of bisexual flowers with extrorse anther cells in the third whorl in Umbellularia. study, we sampled species representing all genera and most sections in the core Laureae, and it is clear that a major disparity exists between our molecular phylogenetic results and more tradi- tional morphology-based taxonomic concepts of ge- neric and infrageneric classification in the tribe. For example. both current and earlier Christophel, 2000; Li et al., two- versus four-celled anthers to separate Litsea s.l. analyses (Li & 2004) duum that using from Lindera s.l. results in polyphyletie or paraphy- etice genera (Rohwer et al.. 1991; Rohwer, 1993; van der Werff & Richter, 1996; Li & Christophel, 2000; Li et al., 2004). therefore, be used with caution in Laureae classifica- This character should, tions, although it does seem to be useful for defining some of the higher-level clades in the group (e.g., the Litsea clade s. str.). The significance of other traditional characters often used at the generic level, such as dimerous or trimerous flowers, also needs re-evaluation. For example, Laurus and Neolitsea have been related previously based on their dimerous flowers, but not in our analyses. Similarly, the nature and arrangement of basal inflorescence involucral bracts (ie. early deciduous vs. tardily deciduous; imbricate vs. decus- sale) are traditionally important characters for delim- iting genera. However, whereas Actinodaphne, Para- — sassafras, and Sinosassafras can be distinguished from other Laureae by the possession of early deciduous, imbricate bracts, these character states are conver- gently homoplasious in our analyses. This suggests that although potentially diagnostic for relating the latter phylogenetic two genera, the character is of limited value, especially given the position o Actinodaphne as a basal grade below Neolitsea. Nevertheless, the clades identified in this study do provide opportunities to examine evolution of specific e, a task and Christophel (2000), particular interest will be studies morphological characters in Laure: p initiated by Li and of of inflorescence development. The most common inflorescence form in Laureae is pseudo-umbels clustered in the leaf axils, but our analyses show that there are apparently at least two different ways to ape this structure. One, 1985; Tsui, 1987) a Lindera, suggested by previous studies (L the Aperula | and H clades, results from shortening of the and seen in Laurus and and internodes in the short shoot with a vegetative terminal bud. This vegetative terminal bud may be normal and scaly or naked, well developed and large, poorly developed, or even reduced. Sometimes the terminal vegetative shoot along with one or two lateral fertile short shoots with pseudo-umbels merge to form a single axillary mixed bud, and the peduncle of pseudo-umbels and the internodes of the short ane may be developed or reduced. Given the positions of these taxa toward the base of the trees, this feature appears to be plesiomorphic in Laureae. The second, derived condition, seen in the Neolit- sea—Actinodaphne clade, results instead from shorten- thyrsoid inflorescence and lacks a ing of a axis vegetative terminal bud. However, given the lack of deep-branch support for the major lineages in this study, definitive conclusions about the phylogeny of inflorescence ontogeny must await further studies. study indicates the need for In conclusion. our caution in the use of morphological similarity for assessing affinities between taxa in the Laureae. Traditional characteristics habit, leaf venation, inflorescence, and floral structure appear in many cases to have been the result of convergent and/or parallel evolution and, therefore, may not be indica- tive of evolutionary affinity or useful for taxon delimitation at higher levels. Added to this is the possibility that some features such as pseudo-umbels the identifies areas in which future research may help to may nol be homologous. Nevertheless, study clarify or correct problems of homology and ontoge- netic convergence. It provides a hypothesis for possible phylogenetic relationships in Laureae, albeit based on a single, rapidly evolving genome, and gives direction. for future studies using multiple indepen- dent and possibly more conservative markers to assess the phylogenetic hypotheses that our results indicate. 598 Annals o Missouri Botanical Garden foundation. for a revised the This will provide the phylogeny-based classification of Laureae in which reliable synapomorphies are backed by data from a range ol sources. Literature Cited & 5. Markos. 1998. Phyloge netic utility of the rDNA: oes nia 14940! FW ojci A ~The ITS region of valuable source of evidence \nn. Bot. Baldwin, B. G. external transcribed spacer. (ET: and Pe Congruence of ETS C Me ). Molec. Sanderson, J. M. Porter, M. C. S. Cam bell & M. J. nuclear ribosomal DNA: / Taie dh Ol ee rm phylogeny. Missouri yard. 82: 247— fecus n X. 2003. Evolution of Mexican Bursera (Burseraceae) inferred from ITS, ETS. and 55 nuclear ribosomal DNA sequences. Molec. Phylogen. Evol. : 300—300 Bena, G., J. M. Prosperi, B. a e & | Olivieri. 1998, Evolution of annual species of e genus Medicago: ^ molecular phylogenetic approach. Molec. Phylogen. vol 9: 552-559. Bentham, ee 1880. Laurineae. Pp. 146-168 in G. Be ntham & J. D. Hooker (editors), Genera Plantarum, Vol L Reeve, i ondor Bremer, R. K Jansen, B. Oxelman, M. Backlund, Lantz & Es J. Kim. 1999, More characters or more taxa for a robust phylogeny—A case study from the coffee family (Rubiaceae). Sys iol. 48: 413-435. Chanderbali H. van der Werff € 5. 5. Renner. 2001. Phylogeny and historical biogeography of Lauraceae: from the Mew a E and nuclear. genomes. Gard. 88: 10 Evidence Ann. Missouri Bot. Clevinger, Y 2000. of NM m and subtribe Engelmanninae He dne ae » base d on ITS and ETS sequence data. Amer. . 87 5€ 5-512. C das C. W. 1997. Can tree rediet when data should be combined? Molec. ] l: 733-740 L. Panero. es netic analysis (Ast leraceae: me ongruence Biol. les E Doyle, J. J. & J. S. Doyle. 1987. A rapid DNA isolation procedure for a oe of fr leaf tissue. Phytochem. Bull. 15; Farris, P. G. Orti Meyer. 2000. Total evidence: dk ules, nep and the phylogenetics of cichlid ishes. J. Exp. Zool. 288: 76-92. d J. S. 1969, \ successive approximations approach to character weighling. Syst. Zool. 18: 374-385. 1989. The index and consistency index. Cladistic 17-419. I. Kallersjo, A. G. Kluge & C. Bult. 1995. Testing es ance of incongruence. Cladistics 10: 315—319 1985. Confidence limits on pe s: An approach using ‘the bootstrap. Evolution 39: 783-79 ‘itch, M. 197 rete nion | the rescaled Felsenstein, J. ward defining the course of c" e "d a specific tree topology. Syst. Zool. 20: 406—416 "hind Lo R. V. Inine & R. Nazar. 1997. Interdependence in the processing of eaa RNAs in Se UNE E pombe. J. Molec. Biol. 273: 782-7 H Graybeal, A. 1998. Is it better to add taxa or joe ler rs loa diffic rn Ped netic problem? Syst. Biol. : 9— Hall, T. A. 1999, Biokdit: A biological user-friendly sequence alignment editor and analysis program for \ doit 95/98/NT. Nucl. Acids Sym. Ser. 41: 95-96, Hillis, D. M. 1998. Taxonomic sampling, phylogenetic accuracy, and investigator bias. Syst. Biol. 47: 3-8. Hooker, J. D. 1890. Laurineae. Pp. 116-189 in J. D. Hooker British India. Vol. 5. L Reeve, nes How, F. C. 1956. Lauraceae. Pp. 86-94 in F. C. How (editor). Flora E angakoni. Science Press, Beijing. Huelsenbeck, J. & F. Ronquist. 2001. MRBAYES: Bayesian infere mee of phylogenetic trees. Bioinformatics 17: 754— la non h. Miller & F. 2002. — . Mil Ronquist. Pote denia application aid ee of Bayesian inference of T —088 phylogeny. Syst. Biol. 5 1e Cen era of Flowering Plants x. Th (Docotyledonae " ad Hyland, B. P. M Hutchinson. J. 19€ . Glarendon Press. Oxford. A revision of Lauraceae in Australia (excluding casa Austral. Syst. Bot. 2: 135-307. Kostermans, A. J. C. H. 1957. Lauraceae. Pengum. Balai Besar Penjel. E Indonesia 57: 1-64. Kumar, S.. K. Tamura & M. Nei. 2004. MEGA3: Integrated software for molecular io ary geneties analysis and sequence alignment. Briefings in Bioinformatics 5: ae Lee. J.. B. C. Baldwin & L. D. Gottlieb. 2003. Phylogenetic re indi among the primarily North American genera of Cichorieae (Compositae) based on analysis of 205 nue A ar DNA ITS and ETS sequences. Syst. En 20: 616-0206 Li, H.- ` 1085. Parallel evolution in Litsea and Lindera of Lauraceae Acta Bot. Yunnan. 7: 129-13 The origin and Mire of Litsea genera group DU Lauraceae. Acta Bot. Yunnan. 1% 251-254. . P. Y. Pai, S. K. Lee , F. N, Wei, Y. T. V abc yas P. H. Huang, H. D Tsui, m Shia & J. à . 1984. Lauraceae. Pp. 1-4 in .-W. a pa Flora Reipublicae Fb Sinicae, a Science Press, Beijing LJ & D. within the [s a complex (Lauraceae): A cladistic analysis Austral. Christophel. 2000, Systematic relationships based on morphological and leaf cuticle data. svst. Bot. 13: 1-13 & H.-W. Li. 2004. Litsea complex ~ J: lati within the 0. — nran Phylogenetic (Lauraceae) inferred from sequences of the chloroplast P gene math and nuclear ribosomal DNA ITS regions. Syst. p 246: 19-34. Li. J. He. Alexander & D. L. Zhang. 2002. Paraphyletic za POR aceae): Evidence from sequences of nu clea ‘ar ribosomal DNA ITS and ETS regions. Syst. Bot. 592— 501. Li, Z.-M., J. bi & H.-W. I 4. 2006. Polyphyly of the (Lauraceae) inf ferred from the analyses of Acta Phytotax. Sin. 44: 1., genus Re LE nrDN A ITS and ETS sequences. 5 . Rl. Le R. Goertzen, B. V. Heuvel. J. Francisco- (de ega 8 R. K. Jansen. . The no transeribed spacer of 185-205 rDNA: p rnal Vor. and phylogenetic utility at low taxonomic le ees 1 Asteraceae and closely allied families. Molec. Phyloge n. Ev I 1 4: 285-303. Long, D. G. 1984. Notes relating to i bos of Bhutan: VHI. sard. Edinburgh 41: 25. Notes Bot. ( Volume 95, Number 4 2008 Li et al. 599 Phylogenetic Relationships of Litsea and Laureae Maddison, W. P. € D. 2000. MacClade: logeny and character evolution, version Maddison. nalysis of phy . Sinauer Associates, Sunderland, Massachusetts. Markos. S. & B. G. Baldwin. 2002. Suae and phylogenetic. utility of the 5 external. transcribed. spacer of. 188-265 DNA in Lessingia Compositae, Astereae). Molec. Phylogen. Evol. 23: 214-228 Morgan. D. R. 2003. nrDNA external transcribed spacer (ETS) reticulate evolution, and the sy E of : 179-101 À ler ny A the Stru euro, molecular gion of the sequences data, Machaeranthera (Asteraceae). E Var 28 Plovanic wh, A E. € J. L. Panero. TS and ETS for lee Use raceae: Heliantheae). Molec. Phy loge n. Evol. 31: Posada, D. & A. Pinal 1998. Modeltest: Testing the mode | of DN, : e Bioinformatics 14: 817-818. . Anatomie des seukundáren Xylems und Vereins Sonderb. Naturwiss. o 3 — > de .Auraceae. Phylogenetic Roalson, S. H & E. A. 2004. relationships and biogeographic patterns in North Amer- Friar. ican members of Carex section Acrocystis (C ~ raceae) using nrDN, A ITS and ETS sequence data. Pl. Syst. Evol. 243: 17 a Rohwer, J.G . Lauraceae. Pp. 366-391 in K. Kubitzki. J. G. Rohwer ^ £s Bittrich (editors), The Families and Genera /ol. 2. i of Vase ‘ular Plants, Springer-Verlag, a erlin. | ` Werff. . Two new S = = = n eec am E M e = = Ke) genera of Ne solropic cal Lauraceae anc d « 'Tritic i re marks on their generic delimitation. Ann. Missouri. Bot. Gard. 78: 388-A0( Saar, phylogenetic ana 2. N. O. Polans & P. D. 2003. A the genus Dahlia (Asteraceae) Sorensen. = — ysis 0 based on internal and external transcribed pus er regions of nuclear ribosomal D uU Syst. Bot. 28: 627-0: Soltis; D. E.. P. S. Soltis; M. E. Mort, M. W. Chase. V Savolainen, S. B. Hoot & 3 M. complex phylogenies using parsimony: An empirical Pup hh using three large DNA data sets for angiosperms. Biol. 47: 32-42 Starr. J. R.. $ . Harris & D. A. Simpson. 2003. Potential of the 5’ añil 3 eus of the intergenic spacer (IGS) of rDNA in the Cyperaceae: quences for lower-leve lew seq phylogenies in sedges wi an example from Uncinia Pers. 3-221 Int. J. Pl. Sei. 164: 213-227. Sullivan, J. 1996. Combining data with different distributions of among du variations. Syst. Biol. 45: 375-380. Swofford, . 1999. PAUP* - Bisloze netic n Using and Other Met hods). Vers l. os ‘husetts & 23. Esumaton of the number of A Sinauer Associates, n ud M. Nei. 19 R Tamura, K nucleotide Hrs cue die in the control region of mitochon- ie DNA in humans and chimpanzees. Molec. Biol. Evol. y 512- a T. J. Gibson, F. Plewniak. F. nad J-l Jeanmougin & D. G B 1997. The Clustal X windows interface: Flexible strategies for multiple sequence alignment aided by od analys tools. Nucl. Acids Res. 24: 4870-4882. ‘sui, H. 1987. A study on the system of Lindera. Acta Palos Sin. 25: 161-171. Urbatsch, L. E.. R. ae & V. Phylogenetic ie of Xylothamia. > Y Karaman. 2003. Gundlachia, and 1 van der Werff, H. 2001. / ey ) region. pumas in the Flora Malestana Blumea 406: 125-140 y Richter. 1996. Toward an improved Da of Tan ‘eae. Ann. Missouri Bot. Gard. 83: 109-418. White. T. J., . Bruns, S. B. Lee & J. Taylor. 1990. ere ation e. direct. sequencing a E NA genes and the internal transcribed spacer in jungi P e 3l ninsky 5-322 in M. A. Innis, G. H. Gelfand, J. J. 5 Vhi PC ee and Applica- "E LA : ~ Menal: Ac ademic Press, San Diego. 1 2( aloes and E oon: eni Academic Publishers, Dordrecht. & Z. Xie. 2001. DAMBE: Data analysis in molecular Hered. 92: 371-373. biology and evolution. J. COMPARATIVE POLLINATION Stacey DeWitt Smith,’ Steven J. Hall?” BIOLOGY OF SYMPATRIC AND Pablo R. Izquierdo,’ and David A. Baum? ALLOPATRIC ANDEAN IOCHROMA (SOLANACEAE)! ABSTRACT ield studies were conducted for 15 Re s of fochroma Benth. and the nested genus Acnistus Schott to quantify the Be sity of pollination systems and to s the potential contribution of pollinator behavior to the pe rsistence of closely related species in sympatry. We e T measures of pollinator visitation and pollen deposition to estimate the importance of major groups of pollinators for each species, and we a ulated proportional similarity in the pollinator assemblage among species. We found that 12 species of lochroma, encompassing a range of flower colors and sizes, were principally pollinated by hummingbirds and, in many cases, by the same hummingbird species. The remaining species were either pollinated by a mix of hummingbirds and insects (lwo species) or exclusively ie Insects (wo species). Based on proportional similarity values, the overlap in pollinator assemblages was found to be higher for sympatric species than for allopatric ones, reflecting iem A local pollinator fauna. However, observations of indis fe pili fidelity. perhaps related to territorial interactions among hummingbirds, suggested ih ab pollinators may still contribute to the reproductive isolation. of sympatric congeners. Nonetheless. because interspecific pollen. flow does occur, the maintenance of species boundaries in sympatry probably requires postmating reproductive isolating mechanisms Key words: Aentstus. flower color. a pollination, lochroma, pollen deposition, pollinator importance. pollinator visitation, reproductive isolation, sympatry. The rich floristic diversity of the Neotropics has remains the paucity of detailed studies of pollination ften been attributed to the complex and specialized ecology, particularly those that catalog not only the — interactions between plants and animals, typically range of visitors but their effectiveness as pollinators the form of herbivory or pollination (Faegri & van der (Kay & Schemske, 2004) Pijl, 1966; Janzen, 1973: Johnson & Steiner, 2000). Mere. we investigate the pollination biology of For example. Gentry (1982) noted that the Andean- — /ochroma Benth., an Andean genus of approximately centered families, such as the Ericaceae, Gesner- 25 species of Solanaceae (Smith € Baum. 2006). iaceae, and Campanulaceae, which account for a large Several authors. (e.g... Lagerheim, 1891; Cocucci, proportion of the Neotropical species diversity, are 1999) have speculated that the showy tubular flowers biotically pollinated and often appear to be special- of Jochroma species are pollinated by hummingbirds: ized for particular groups of animals such as however. no previous field studies of pollination exist hummingbirds or bats (Perret et al., 2001: Luteyn, — for this group. While several Jochroma species (e.g.. 1. 2002; Muchala, 2006). While geographical patterns in fuchsioides (Humb. & Bonpl.) Miers and £ gesner- pollinator specialization have become the subject of ijoides (Kunth) Miers) are indeed a close fit to the debate (Ollerton & Cranmer, 2002: Olesen & Jordano, — classic hummingbird syndrome flower, namely red, 2002), the largest barrier to understanding the role of — scentless, and tubular, most species vary from this pollinators in the diversification of Neotropical taxa suite of traits, suggesting pollination by other groups The authors gratefully acknowledge support for this study from National Science Foundation (grant ( 03 0931 0). the Marie l Wi HRS Pap. r Fellowship. the University of Wisconsin Tinker-Nave Fund, the University of Wisconsin Department of Jotany O. i Memorial Fund. and a postdoctoral fe "ern in Evolutionary Genomics and Molec i ar Evolution from Duke a niversity and the Duke ge for Genome Sciences & Policy. We thank W. Quizhpe. A. Rodriguez. V. Zak. and S eiva G. for assistanc e in the field, R. Bleiweiss for help with nb determinations, 5. rawik or higher-level insect identification, C. Rasmussen for de le rminations of Apidae and Halictidae, R. Jeanne for the determination of Polybia, and R. A. Smith for rena We thank A. Tye for arranging study of us Mo d endemic fochroma ellipticum. The manuscript was improved by comments from J. W. p s . Hollowell. . Spooner. K. J. Sytsma, N. Waser, P. Wilson. and two anonymous reviewers * Department ‘of m lany, University of Wisconsin, Madison, Wisconsin 53706. U.S.A. "Charles Darwin Research Station, Puerto Ayora, Santa Cruz, eaa Islands, Ecuador. 5 "Current address: Department of Biology. Duke University. Durham, North Carolina 27708. U.S.A. a e du. Current address: Ecosystem Sciences Division, ESPM, 137 Mulford Hall #3114. Berkeley. oe. 94720, US stevenhallOberkeley.edu. doi: 10.341 7/2007037 ANN. Missouri Bor. Garp. 95: 600-617. PUBLISHED ON 20 DECEMBER 2008 Volume 95, Number 4 Smith et al. 601 2008 Pollination Biology of lochroma D d E ja na c -— I. calycinum E A Pais. Rol. use" " pr LAN J n — on I. gesnerioides e “Gu ajalito"* E ME ad - z e TE : 1862 m Alun man E CIN > 2615 m +} en > ice: : i PA. arborescens y von. E cyaneum, l Malpote à T > 1825 me END ʻ Sei S E 34 f dá j| Coptambo gy . ae e . confertiflorum b di mE $ MM - — ifi ifl l í e ca NU f E . Cisne *,2310 m 7 s — I. fuchsioides Ayabaca? 2688 n 4 4 Ld 2t Sy. Guzmango Ki P MV 42563 m P PC - -3-7 Ak ae I. edule Be r hd A Le” -—— mmm I. umbellatum I. cornifolium eS "Negallpampa I. peruvianum 2 cm ac eR nec riu Figure 1. Study sites in Ecuador and Peru in northwestern South America. Country borders are in dark grey. and provinces or departments are bounded in light grey. Approximate dis de of the Amotape-Huancabamba zone (from its Weigend. 2002) are indicated with blue lines. Galápagos Islands (upper left) are not to scale. Study indicated with dashed lines, x flowers are shown to scale (bottom right). Su s with sympatric laxa are marked with stars. y specie es al each site are and the names of the principally hummingbird-pollinated taxa (as determined by this study) are in boldface of animals (Fig. 1). For instance, the flowers of 7. confertiflorum (Miers) tubular, and scented, traits more commonly associated with moth pollination (Faegri & van der Pijl, 1966). Also, many species exhibit the peculiar combination Hunz. are greenish white, — of long, tubular, blue or purple flowers, a combination that does not correspond to any known syndrome (Faegri & van der Pijl, 1966). Given the diversity of floral morphologies present in /ochroma, we predicted that the composition of visitors and their effectiveness as pollinators would vary substantially between species. Differences in pollination system among /ochroma species may carry important implications for the maintenance of species boundaries. In some parts of the Andes, Jochroma species occur in sympatry and mat flower together during the rainy season, creating the potential for interspecific pollen flow. Although some hybrids have been documented in zones of sympatry (Smith & Baum, 2006). not rampant, and many taxa coexist with no observed interspecific hybridization is (Smith, Pollinators contribute to reproductive 2000). isolation may of sympatry either by forming cre relationships hybrid formation in areas with particular /ochroma species or by exhibiting constancy during foraging bouts, ne that interspe- cific gene flow is limited (Jones, 1978; Campbell & Motten, 1985: Waser, 1980). In the present study, we assessed the specialization of pollinators on 15 Jochroma species and the nested — genus Acnistus Schott, and, in areas of sympatry, we examined the importance of pollinator behavior in interspecific pollen flow. In order to characterize the pollination system for each species. we measured pollinator visitation rates and pollen deposition for four major of pollinators (hummingbirds [Trochilidae], as well as hymenopteran, lepidopteran. groups and dipteran insects), and we calculated a composite variable, pollinator importance, for each group. Using the visitation rates for each pollinator, we compared the similarity in pollinator assemblage for allopatric and sympatric species pairs to determine if sympatric Annals of the Missouri Botanical Garden taxa tend to show more divergence in their pollinator use than is typical for allopatric taxa. Finally. we compiled observations of individual pollinator move- ments in sympatric areas to examine the possibility of preferential visitation by individual pollinators. MATERIALS AND METHODS STUDY TAXA Our range of floral variation within the genus. taxon sampling of /ochroma represented the lochroma flowers may be red, orange, yellow. white, green, blue, and purple, and all of these colors were included in our study group. Flower form in fochroma ranges from campanulate to narrowly tubular. Here, the campan- ulate form is represe ted by Aenistus arborescens (L.) Schltdl. the sole the (Hunziker, 2001), which is nested within the core clade of lochroma (Smith & Baum, 2006). Corolla tube length varies more than eight-fold in /ochroma. and in member of genus Acnistus our study taxa ranges from less than 1 em to more than 6 em (Fig. 1). Only two Jochroma species (1. ellipticum (Hook. f.) Hunz. and Z confertiflorum) produce any Their light, sweet scent is very similar to that of their close noliceable scent, and both were included here. relative 4. ärborescens, si scent apparently derives from a mixture of 3,5-dimethoxytoulene, jasmine, anisaldehyde, and aa anthranilate (Kaiser, 2000). The selection of study the phylogenetic (Smith & Baum, 2006), genus lochroma is not monophyletic. taxa also took into account recent lochrominae that the but is instead three major the fall the large core clade of 13 lochroma species (Smith € Baum, 2006), which also includes analysis of which indicated divided among clades. Eleven of studied taxa into Acnistus arborescens. Outside of this core additional species, Í, parvifolium (Roem. & Schult.) D'Arey and I, umbellatum (Ruiz & Pav.) Hunz. ex D’ Arey (Smith & Baum, 2006). taxa, Z. group of iochromas, we sampled two Also, we studied two recently named ayabacense S. Leiva and stenanthum 5. Leiva, Quipuscoa & N. W. Sawyer, which appear to be of hybrid origin (Smith & Baum, 2006). Phylogenetic analyses have suggested that Z. ayabacense is a hybrid between /. lehmannii Bitter and /. cyaneum (Lindl.) M. a Green, and / stenanthum is probably a hybrid between Z. cornifolium (Kunth) Miers and A. (Smith & 2000). collected for all studied populations: these are given arbor- escens Baum. Vouchers were Appendix STUDY SITES lochroma species are mainly distributed in the Andes of Colombia south to Peru, where they occur as sparsely distributed large shrubs or treelets in scrub or cloud forest between 2200 and 2900 m. The greatest species richness occurs in the Amotape- Huanacamba zone at the border of Ecuador and Peru (Smith & Baum, 2000: Fig. y zone lochroma species are typically found in allopa- l 1). Whereas outside of this y. geographic ranges frequently overlap within the Amotape—Huanacamba zone, and up to four species can co-occur within a single l-km” area. A few species pairs (e.g.. 7. cyaneum and £. cornifolium) hybridize in areas of contact, but many do not (e.g., L. umbellatum and /. edule S. Leiva) (Smith, 2006). differences among taxa in microhabitat preferences light, lochroma "ps There are some (e.g.. soil, and moisture conditions), but generally. species are found in areas of moderate disturbance, such as forest gaps, trails, dry stream beds, or field edges. The 11 study sites were typically located on the outskirts of rural towns o villages, in areas of mixed secondary vegetation and small-scale agriculture; the locations are shown in Figure l and listed in Appendix 1. lochroma species flower throughout the year, but the rainy April in the regions of Ecuador and Peru where these studies took peak flowering occurs during season. from roughly December & Andean place. Flowering peaks slightly earlier in northern Ecuador (December) than in southern Ecuador and Peru (January to February). During these months. the weather is typically sunny in the morning and cloudy or rainy in the afternoon and evening. Pollinator observations for each species were made over a three- to four-day period during the rainy season: a few studies were extended for a fifth day when extremely rainy conditions persisted for several days or when the overall rate of pollinator visitation was low. The dates of each study are given in Appendix 1. To the extent possible, the pollination studies were conducted in sites that were well within the geographic and altitudinal range of the species and contained mans 200). Up to 14 individuals in each population were incorporated into arge, flowering individuals (20 to each study (Appendix 1), although the scarcity and/or poor accessibility of some taxa limited the number of available Pu als (e.g.. four plants for Z. peruvia- (Dunal) J. F. Macbr.). num FLORAL BIOLOGY Stigma receptivity was judged using the hydrogen 1993). collected from flowers throughout anthesis (from bud peroxide test (Kearns & Inouye, Styles were to a wilted flower) and dipped in hydrogen peroxide. The (production of bubbles) was taken to indicate the youngest stigma to yield a positive result onset of stigma receptivity. Volume 95, Number 4 2008 Smith et al. Pollination Biology of lochroma 603 Nectar volume was measured with calibrated glass micropipettes, and the percentage of sugar was estimated using a temperature-compensated hand refractometer (QA Supplies, Norfolk, Virginia, U.S.A.) accurate to 0.2% between 0% and 32% sugar by volume. Nectar was extracted by pressing the micropipette into the spaces between the filaments a where the liquid accumulates in /ochroma. Samples were taken from first- or second-day flowers (covered with I-mm mesh bags before anthesis) at 3-hr. intervals from 0700 to 1900 hr. sampled once. Flowers were samp Each flower was only =x Ks, ed from five to 10 a individuals per study population. VISITATION Pollinator visits were recorded for two to three days during three observation periods: morning (0600 0900 hr.), midday (1100 to 1400 hr.). and evening (1700 to 2000 hr.). Two hours of observations were recorded during each period, but with the time required to move between study individuals, each period spanned 3 hr. The time periods were chosen based on previous observations that suggested that these are periods of high pollinator activity. During 0600 hr. Observations were made while sitting 2.5 to the studies, the sun rose al ca. and set al 1900 hr. 3.5 m from a subject plant to minimize the distraction of visitors. Both the number of visits to the subject plant and the number of legitimate flower visits (sensu Jones & Reithel, 2001) were noted. Bird visitors were identified to the lowest possible taxonomic level using field guides and consultation with experts. Visiting insects were collected in ethanol (70%) and later identified by comparing them to reference collections at the University of Wisconsin-Madison, the Univer- sity of I[llinois-Urbana, and the Charles Darwin Research Station. All insect specimens were depos- ited in the University of Wisconsin-Madison Insect Research collection. calculated Plant and flower visitation rates were from raw observation data by dividing the total — number of visits by a given pollinator by the time observed (in flower hours to correct for differences in display size among individuals, sensu Dafni, 1992). Flower hours were calculated from a series of observations (1 to n) as follows: FH = (Dj X Tj) + Ta), where FH equals the (Də X Tə) +... (Dn X number of flower hours, D is the display size (number of flowers on the plant), and T equals the number of hours the plant was observed. Visitation rates were calculated both for individual species and for four major groups of pollinators (hummingbirds, as well as and dipteran insects). hymenopteran, lepidopteran, Flower visitation rates were used to compute the proportional similarity (PS) of pollinator assemblages between all pairs of /ochroma species (Schemske & Brokaw, 1981; Kay € Schemske, 2003). ' takes into account both the number of pollinator l'his measure species shared and their visitation frequency; PS for a pair of plant species is 1 — Y È |P Pp; |. where 2 al — ai and Pj; are the proportions of the total visitation rate made up by pollinator species i for plant species a and b, and differences in P and P; are summed at across all pollinator species, 1... i. PS values range from 0 to 1, with higher valved indicating greater between two overlap in the pollinator visitation species. INTERSPECIFIC POLLINATOR MOVEMENTS We assessed the potential for interspecific pollen flow in sympatry by observing pollinator visitation to individual plants from different species growing side by side. The closest pairs of plants, typically 1 to 5 m apart, were selected for observation, and movement of the pollinators within the plants and between the plants was recorded. We examined the visitation patterns for bias toward particular Jochroma species by a given pollinator by comparing the expected to the observed number of visits with y^ analysis (as in Schemske, 1981). Expected values are the number of visits expected if visits are directly proportional to Also, sympatric plant pairs, we compared the number of display size. using the observations of the plant visits that involved movement between species to those that were restricted to one species as an additional measure of pollinator fidelity. POLLEN DEPOSITION AND POLLINATOR IMPORTANCE To measure pollen deposition by different pollinator species, virgin flowers (covered with green or black 1- mm mesh bags before anthesis) were presented to pollinators and then re-bagged after a single visit by a single pollinator. The pollen loads on visited stigmas were compared to stigmas from flowers that were bagged for the duration of the experiment to determine if these species require biotic pollination. Also, we collected and examined from stigmas unbagged flowers to assess the typical pollen deposition for ¡om flowers exposed to unlimited visits. To test for nocturnal pollination, stigmas were collected from flowers that had been bagged during the day but left unbagged at night (1900-0600 hr.). Additionally, for 13 study species (excluding Acnistus and /. gesnertoides), arborescens, lochroma cyaneum, stigmas were collected from flowers that were baggec at night but unbagged during the day (0600-1900 hr.). Where there were sufficient flowers, all four treat- 604 Annals of the Missouri Botanical Garden ments (always bagged, never bagged, bagged during RESULTS night only, bagged during day only) were completed pt or each individual. Styles were removed from flowers that experiment lo ensure were already open at the beginning of the that only virgin flowers were included. Styles from visited or treated flowers were collected 6 to 12 hr. formalin, acetic after replacement of the bag, fixed and alcohol (FAA) for 12 %) for storage. Lo acid, 24 hr. and transferred to ethanol (70 The fixed stigmas were examined for the presence of germinating pollen grains using a modified version of Martin's (1959) protocol. IM NaOH to soften (10 min.). washed 50 mM KPO, buffer (5 min. per wash), stained. with 0.05% im 50 mM KPO, buffer for 5 min., squashed on a microscope slide, and The styles were soaked three times in aniline blue decolorized viewed under ultraviolet fluorescence microscopy. 100 grains, than 100 (e.g o. When the style contained fewer than ca. When there divided all were counted. were greater grains, the style was into. sections halves, quarters). and a count from a representative Non- fluorescing grains (nol germinating and/or not Sola- | estimate the total load. section was used naceae) were excluded from counts. fochroma species do not differ markedly in pollen morphology, so it was pollen to species. To relate not possible to identify pollen load to ovule availability, the average number of ovules per flower in each species was estimated by ovules flowers bring the counting from enough standard error to less than 10% of the mean. when sufficient flowers were available. Pollen. deposition. (quality) was combined with stimate ee 1983: Mayfield (quantity) to give an overall « (Waser Q Herrera. 1987: visitation. rate of pollinator importance Schemske & Horvitz. 1964: et al.. 2001). Here, pollinators was calculated as the product of the relative visitation rate and the proportion of available ovules potentially pollinated by a single visit (using the previously deseribed estimates of ovules per flower). If the de | posite ad by average number of pollen erains a single visit exceeded the estimated number of ovules, the proportion was set to 1.0. We used this scaled deposition for two reasons. First. i allowed us to take into account the differences in and, thus, in potential ovule number per-visil Second, this that typically levels off quickly with increasing pollen load (Silander « 1978; & Waser, 1985). so pollen deposition that greatly exceeds the effectiveness across ae species. approach accommodates the fact seed sel Primack, Kohn number of ovules (as observed here for many hymenopterans) is unlikely to result in a proportional increase in fitness. Aenistus the importance of each group of FLORAL BIOLOGY Measurements of stigma receptivity showed that and all species of /ochroma. except 1. umbellatum., are protogynous. Stigmas are receplive when the flowers open, and the anthers dehisce » 3 hr. later except in 4 umbellatum. in which the anthers are already dehisced when the flower opens. /ochroma flowers open asynchronously throughout the day, but with some tendency toward opening in the morning, and they do not close at night. The stigmas remained receptive until the flowers wilted. two to three days after opening. Nectar-standing crop varied widely across taxa, but did not show strong diurnal patterns, in large part because of the staggering of flower maturation during the day. For comparison across taxa, we pooled measurements from all sampling times and computed averages for nectar volume and concentration (Ta- ble 1). duced the lowest volume of nectar (0.: The small-flowered: Aenistus dale pro- 0.1 ul) and presented the smallest. reward, buo on a pee and per-plant basis. The large-flowered. fochroma 38.4 + calycinum Benth. produced the most nectar 3.0 yl), although it did not offer the highest per-flower re ward because of its low sugar concentration (14.5 + 0: The most rewarding species was 4 loxense Ron Miers due to its high nectar volume (37,0 + 2.7 ul). high sugar concentration (24.0 + 0.5%). and arge display size (120.3 + 38.1 flowers per plant). VISITATION 264 hr. of observation, 47 pollinators were observed legitimately visiting the 16 Bees (Diglossa Chapman g | Over the course of ca. study species (Appendix 2). and species of lower-piercing birds dae]) were frequent illegitimate visitors. robbing nectar by making holes in the sides of the corollas. These illegitimate visits were not included in visitation rates. Legitimate visitors included hum- mingbirds (Apodiformes, Trochilidae) and a wide variety of insects from the Hymenoptera, Diptera, and Lepidoptera (Appendix 2). Hummingbird visits were evenly spread across all observation. periods. hymenopteran visits were most period (1100-1400 hr.). morning common in the midday dipteran visits occurred. primarily in the (0600—0900 hr.) moths) were most frequent in the evening period (1700-2000 hr.). 2.6 and 2.1 whereas Lepidoptera visited 4.5 flowers on average and lepidopteran visits (mainly Hymenoptera and Diptera visited flowers per plant visit, respectively. and hummingbirds 15.8 flowers. Volume 95, Number 4 Smith et al. 605 2008 Pollination Biology of lochroma Table 1. Nectar rewards across study taxa.' Flowers Average nectar Average percent Average reward per Average reward per Species sampled volume (ul) + SE sugar + SE flower + SE plant? + SE Acnistus arborescens l6 0.5 + 0.1 14.4 + 2.5 0.1 = 0.02 2.3 X 0.6 lochroma ayabacense 25 11.8. 1 22.5 + 0.3 2.7 = 0.2 459.9 + 128.1 I. calycinum 20 38.4 + 3.0 14.5 + 0.3 EDO 169.3 + 63.9 I. confertiflorum 37 Fa Mo 20.9 + 0.6 3.8 + 0.4 124.4 + 148.6 I. cornifolium 20 21. E 32 18.4 + 0.3 6.9 + 0.6 603.0 = 121.8 lI. cyaneum 29 17.7 + 1.9 23.1 + 0.7 4.1 = 0.4 177.8 + 42.6 I, edule 28 10.9 + 1.1 20.7 + 0.4 2:9 50:2 704.7 = 167.6 I. ellipticum 20 1.8 + 0.3 76 + 1.4 0.2 + 0.05 10:22 I. fuchsioides 32 20.2 + 1.8 26.4 + 0.7 5.2 + 0.4 114.8 + 79.7 I. gesnerioides 11 10.0 + 2.4 16.3 = 0.66 17+ 0.4 841.7 182.2 I. lehmannii 26 7.0 + 0.6 27.2 + 0.8 L8 4 0.2 515.3 + 229.6 I. loxense 30 37.0 + 2.7 24.0 + 0.47 8.7 + 0.5 1048.1 + 338.6 l. parvifolium 17 17.5 > 1.1 20.0 + 0.7 35503 342.6 + 133.7 I. peruvianum 20 6.9 X 0.7 21.3 + 0.0 1.4 + 0.2 122.6 + 29.0 I, stenanthum 23 21.4 4 2.5 19.5 = 0.4 4.2 X 0.5 880.5 + 243.7 l. umbellatum 32 2.2 X 0.44 16.4 = 0.6 0.4 + 0.08 31.1 = 17.1 All values are shown with AD error (SE). ? Reward per flower is the product of volume per flower and percent sugar. * Reward per plant is the n of average reward per flower and average display size (Appendix 1). ! Variation in display size across the population was not measured in /ochroma donum thus no standard error was calculated. Based on relative flower visitation, we classified the overlap. in pollinator assemblage among species species into three broad classes: principally hum- (Fig. 2). We observed that species with different mingbird pollinated, mixed hummingbird/insect pol- pollination systems (Table 2) did not have signifi- linated, and exclusively insect pollinated (Table 2 — cantly lower proportional similarity than species with Taxa for which greater than 75% of the total visitation the same pollination systems (mean PS = 0.13 + 0.02 was concentrated on a single group of pollinators were — vs. 0.18 + 0.03, unpaired t-test: P = 0.16). That is, considered to be specialized for that group (Fenster el two principally hummingbird-pollinated species did al., 2004). With this criterion, 11 of the study species not necessarily show greater overlap. in pollinator were considered to be principally hummingbird assemblage than a principally hummingbird-pollinat- pollinated (Table 2). fochroma calycinum had only ed and an insect-pollinated species. However, species 65% hummingbird visitation, but was also classified studied in allopatry had lower PS than those in as principally hummingbird pollinated because most sympatry (mean PS = 0.13 + 0.02 vs. 0.51 + 0.10: P insect. visitors were found to be poor pollen vectors < 0.00001) and species living in different regions had (described below: Table 3). Two species, Z. peruvia- — significantly lower proportional similarity than those num and L umbellatum, were visited almost equally living in the same region (mean PS — 0.08 + 0.01 vs. by hummingbirds and hymenopterans and were 0.36 + 0.04; P < 0.00001 J b : | 290 Ud: . . considered to have a mixed bird and insect pollination system. Acnistus arborescens and 1. ellipticum Wele ^ docosensemit POLDNITOR MOVEMENT visited frequently by all three groups of insects but no hummingbirds were observed (Table 2). so these two To better understand. pollinator activity in areas of laxa were considered insect pollinated. Thus, we did — sympatry, we examined patterns of pollinator visitation not observe a diverse array of specialized systems, at to pairs of plants from different sympatric species. Our least at the level of major pollinator groups. observations revealed biased patterns for most pollinator lochroma species appeared more divergent in species. We observed that the larger hummingbirds pollination system at the pollinator species level, —(e.g.. Coeligena iris Gould and Colibri coruscans Gould) but these differences may be driven more by — visited the species that provided the greater nectar geography than by specialization for certain pollina- reward significantly more frequently, in every case with tors. As described previously, we calculated the a sufficient sample size (Table 4). The smaller hum- pairwise PS metric from the pollinator visitation rates mingbirds Adelomyia melanogenys Fraser, Myrtis fanny Appendix 2) and used these values to assess the Lesson, and Polyonymus caroli Bourcier often showed Table 2. Plant and flower visitation rates across study taxa.'? All visitors Trochilidae Hymenoptera Lepidoptera Diptera Panem Plant Flower Plant Flower Plant Flower Plant Flower Plant Flower pollinators Acnistus arborescens 0.130 0.428 0 0 0.024 0.082 0.010 0.091 0.096 0.255 Insects lochroma ayabacense 0.010 0.186 0.010 0.186 0 0 0 0 0 0 birds I. calycinum 0.050 0.140 0.008 0.090 0.017 0.024 0 0.025 0.025 birds I. confertiflorum 0.037 0.221 0.011 0.173 0.024 0.042 0.002 0.006 0 0 birds I. cornifolium 0.016 0.185 0.012 0.177 0.004 0.008 0 0 0 0 birds I, cyaneum 0.075 0.299 0.041 0.238 0.027 ).040 0.008 0.021 0 0 birds I. edule 0.012 0.177 0.011 0.175 Lx 10 0.001 2x 10 0.001 0 0 birds l. ellipticum 0.018 0.022 0 0 0.007 0.008 0.007 0.008 0.005 0.005 insects 7 fuchsioides 0.032 0.304. 0.008 0.235 0.020 0.049 0.003 0.020 0 0 birds I. gesnerioides 0.00€ 0.165 0.006 0.148 0.001 0.010 0.001 0.006 0 0 birds I. lehmannii 0.007 0.341 0.006 0.338 0.001 0.003 0 0 0 0 birds I. loxense 0.040 0.283 0.017 0.25] 0.023 0.032 0 0 0 0 birds l. parvifolium 0.023 0.336 0.016 0.309 0.007 0.027 0 0 0 0 birds l. peruvianum 0.033 0.107 0.008 0.047 0.025 0.060 0 0 0 0 birds/insects I. stenanthum 0.012 0.263 0.010 0.259 0.002 0.004. 0 0 0 0 birds 1. umbellatum 0.034 0.298 0.008 0.154. 0.022 0.140 0.001 0.001 0.002 0.003 birds/insects MI species 0.032 0.232 0.011 0.174 0.012 0.032 0.002 0.009 0.008 0.017 ! Rates are presented as the numbers of visits per plant or flower per flower hour (see Methods, Visitation). * Pollinator groups are hummingbirds (Trochilidae) and the three orders of insects observed. uopier) |eoiuejog unosstlA 909 au) Jo sjeuuy Table 3 Pollen deposition by pollinators and during treatments. !? Trochilidae Hymenoptera Lepidoptera Diptera Day control? Night control* Bagged control Un gged control 8002 y Jaquinyn ‘GG euin|oA Je 19 YYWWS Average Average Average Average Average Average Average Average grains grains grains grains grains grains grains grains deposited Sample deposited Sample deposited Sample deposited Sample deposited Sample deposited Sample deposited Sample deposited Sample + SE i SE i + SE i + SE size + SE size Jo size + SE size + SE size Acnistus arborescens 898 + 128 11 266 +: 118 5 149-59 22 f 718 + 241 4 0x0 4 811 + 204 3 lochroma ayabacense 264 + 56.0 21 285 + 149 5 [7-312 n ] 1 8 114 + 89 13 I. calycinum 579 + 428 3 2248 + 727 | 139 + II 2 2058 + 322 11 60 + 38 8 134 + 83 5 1208 + 170 14 Í. confertiflorum 971 + 217 7 389 = 75 d 212+ 106 3 1262 + 178 17 111 = 48 19 70 + 45 7 1462+ 178 12 I. cornifolium 1436 + 215 3 1558 = 603 7 1396 + 331] 27 910 + 193 9 “152.33 6 2209 + 242 27 Í. cyaneum 1076 + 332 T 132x720 2 88 l 7 11527 14 24 + 11 l4 775 X* 110 26 Í. edule 1290 + 240 22 2355 £137 2 2144-944 2 2422 + 404 16 878 + 312 11 ] +1 6 1650 += 140 28 I. ellipticum T&D] 3 1326+387 5 257+ 78 860 £151 16 534 + 174 11 29 + 13 10 1769 + 464 6 I. fuchsioides 279 + 194 7 1151 +343 4 31719 4 1447 + 265 13 0=0 10 2.2 8 1099 + 177 19 I. gesnerioides 1107 + 791 3 2780 + 301 11 T * IZ] D 824 19 1423 + 114 33 I. lehmannii 3083 + 577 1] 595% 2 1155 — 853 10 122+ 119 4 23 t 13 4 4367 + 487 14 1. loxense 407.1 + 259 5 47205 6 804 + 142 2] 238 + 91 12 2:102 16 790 + 106 28 I. parvifolium 14 +7 5 7 l 109 +5 9 lee 8 00 13 438 + 86 21 Í. peruvianum 1545 | 3328 X334 4 2614 + 608 9 173 + 151 5 142 * 65 6 1956 € 473 12 I. stenanthum 166 + 59 9 14 l 7161 + 237 14 263 + 181 6 l=] 5 686+ 114 26 I. umbellatum 58 l 3727-493 2 T 0 l 438 + 255 5 365 + 177 6 17 +17 13 937+ 132 22 All taxa 940 + 121 105 1472+ 157 64 7124194 20 151 +49 27 1453 + 122 154 262 + 42 135 28 + 6 144 1323 + 67 310 = standard error. i For each plant species, the average number of pollen grains deposited by each group of visitors is presented; blank cells indicate that the pollinator group was not observed visiting that plant specie : The Eek size is the number of visited stigmas counted for each pollinator group or treatment. 3 ' Day control refers to flowers unbagged only during the day, and night control to those unbagged only during the night. * Treatment not completed for this species. t * Although this class of visitor was recorded for that species, no pollen counts were obtained. euo1yoo] jo ABojorg uoneut|og 409 608 Annals o Missouri Botanical Garden 2 a m z = | < T < 4 & N N < & = s — Q y p: 3 2 5 o b. S. =| a x = Bla le SS 2 Y 215 312 813 t Ela Sle E E 8 2 | + $5: 32 8 8/8 5.5 3 3 3 EJS 2|$ 2|y $ SI] 313 $ 3 > = = = E EY =] f Š I E E E = E $ S 313 E39 s Els 3|5 s 3 3 Se 3 S > S EY a, x E 2 a S pa Species (site) B =~ M ~ ~ - ~ - = = = = = M M lock llipticum (GAL) 0 — Q0 o ofo olo « 0 ¢ 0 0 tcnistus arl (ALL) 0 0.06] 0.02 0.09 | 0 6.01) 0.01 008 O11) 0 0.01) 0.11 003 0.01 N. ECU 1. calycinum (GUA) 0.02 ( 0.07 ( 0 0 0 0 0 0 I. gesnerioides (PUL) 0.26 0.09 0 0.01 | 0.19 0.23 0,31 | 0.24 025 || 031. 0.28 0.26 C. ECU I. confertiflorum (MAL) 39} 0 0.01 | 0.19 0.19 043 | 041 0.42 | 043 030 0.42 Dicis I. fuchsioides (COJ) 0 0.01 | 0.01 0.08 0.09 0 0.01 | 0.09 0.03 0.01 S. ECU L cyaneum (CIS) 0.7 |o 0 0 Ej ERUIT 17 f I. loxense (LOJ) 0.001 0.01 0.01 0 0.01 | 001 001 0.01 I. edule (AGA) 0.76) 0.20 | 0.19 022|022 0.19 022 I ifoli. 0.25 | 0.17 0.18 | 0.27 0.20 0.18 L umbellatum (AGA) 0.48 0.44 0.88 0.31 3 . I. ayabacense (AYA) 0.9 Y N. PER M I. lehmannii (AYA) I. peruvianum (GUZ)| I. cornifolium (GUZ) I. stenanthum (GUZ) Figure 2. Pairwise similiarity in pollinator assemblage across /ochroma species. The names of the principally hummingbird- Morea taxa are in boldface. Study site names are abbreviated as follows: AGA = Agallpampa. AYA = yabaca, CIS = Cisne, COJ = Cojitambo, GUZ = Guzmango, S = bl MAL = -M lalpote, and PUL = Pululahua. Re Ions are abbreviate P as follows: N. ECU = northern Ecuador, C. ECU = tral Ecuador, S. ECU = southern Ecuador, and N. PER = northern Peru. Values for taxa in the same region are ind d black lines. Shading of cells denotes degree of similarity significant preference for a particular species, although, in some cases, for the less rewarding species. For instance, over the course of three bouts, A. melanogenys visited 49 flowers on the low-reward fochroma umbella- tum and no flowers on Z. edule (P < 0.001). Insect visits were typically too infrequent to s — low any significant pattern, but often they tended toward the less rewarding species (Table 4). the of individual We also considered movement pollinators between these sympatric plant pairs. Although pollinator species could be observed visiting multiple sympatric plant species, we rarely saw movement between species by individual pollinators (Fig. 3). For instance, when observing an individual of lochroma cyaneum and 1. lehmannii side by side, only one of 60 visits involved movement from /. lehmannii to I. cyaneum and five involved movement in the reverse direction. The smaller hummingbirds (mentioned above) were largely responsible for these occasional interspecific movements. However, the overall rarity of interspecific movements between these individual plant pairs points to some individual-level pollinator fidelity. Due to the sparse distribution of /ochroma plants, we could not explore the frequency with which individual pollinators continued to be constant to a particular species after moving beyond the observed pairs. in wis black lines, aad taxa in sympatry are boxed in thick POLLEN DEPOSITION AND POLLINATOR IMPORTANCE Results from the bagged and unbagged controls suggested. that the study species require biotic — pollination. Styles from bagged flowers typically had very few pollen grains, but in three cases (/ochroma calycinum, l. cornifolium, and L peruvianum) they 152, and 142 grains, respectively (Table 3). These loads, which are had larger loads, with means of 134, small relative to those typically on visited flowers. could be due to very small insects penetrating the likely deposited during anthesis. If the observed loads mesh, but were more due to self-pollen bagged flowers of 7. cornifolium, I. peruvianum, and 1. calycinum are indeed self-pollen, they might not result in fertilization as crossing studies indicate that most lochroma species are self-incompatible (Smith & Baum, 2007). Comparison of the night- and day-bagged flowers indicated thal, on average, 5.5 Umes more pollen was deposited during the day (0600-1900 hr.) than during the night (1900-0600 hr.; 1 lochroma ellipticum, I. umbellatum, and Í. cornifo- ‘able 3). In three species, lium, pollen tube counts for the night-exposed flowers were greater than 50% that of day-exposed flowers. lochroma ellipticum received the most visits during the Table 4. Visitation patterns to sympatric species in Guzmango, Ayabaca, and Agallpampa. Peru. Pairs of plants of sympatric taxa were observed in the study localities (see Appendix 1), with interplant distances (d) in meters and the time observed in hours (hr.). For pairs of plants greater than 3 m apart, only hummingbird movements could be recorded. Species names are abbreviated to the first letters of the specific epithet. Although the larger study of lochroma cyaneum involved an allopatric population, it was found in sympatry with other taxa in Ayabaca, Peru, and so was ga Y^ analysis with Yates” correction. Expected values (in parentheses) are the number of visits using her, it would be expected to receive twice included in these observations. Observed plant and flower visits were compared to expecte visits expected if visits were directly proportional to display size. For instance, "i one of ihe two BR wing observed has twice as many flowers as the ot as many visits. Significant Y^ values are noted with asterisks where * is P < 0.05 and " is P < . Boldface values are significant after Bonferroni correction Site: Guzmango, Peru Hummingbird visitors Insect visitors Coeligena iris Adelomyia melanogenys Apis mellifera Vespid wasp Plant pair observed: time observed, interplant distance per cor Y per cor X per cor X per cor x L. mo with 120 flowers and /. cornifolium Plant visits 0 (3) 4 (1) 6.5 3 (4) 3 (2) 0.4 7 (5) 0 (2) 1.7 2 (1) 0(1) 0.02 with ers; 3.6 hr., l.5 m Flower visits 0(19 27(8) 61.4 18 (36) 33 (15) 28.9 14 (10) O0(4 45 2 (1) 0(1) 0.02 p dune s i5 fin and Í. cornifolium Plant visits 1 (1) 0 (0) 0 (1) 1 (0) 0.3 3 (2) O0(1l) O.1 25 flowers; 4 hr., Flower visits 2 (2) 0 (0) 0 0 (2) 2 (0) 2.7 5 (4) 00) 0.6 Hummingbird visitors Insect visitors Site: Ayabaca, Peru Coeligena iris Adelomyia melanogenys Bombus cf. Plebeia Vespid wasp Plant pair observed; time observed, interplant distance leh cya X leh cya X leh cya X! leh eya y¥? leh cya Y I. lehmannii with 60 p and I. cyaneum Plant visits 1 (6) 16 (11) 4.3 (2) 5 (5) with ers; 3 hr. ower visits 2 (66) 203 (139) 91.07 445(30) 46(61) 11.0 I. lehmannii id 144 one a I. cyaneum ud visits 0 (7) 9 (2) 32.1 20 (18) 2 (4) 0.9 1(1) 0 (0) 2 (2) 0 (0) 3 (2) 0 (1) 0.02 vith 35 flowers; 3 hr., ‘lower visits O (29) 36 (7) 143.0" 224(184 5(45) 42.8 10(8) 0(2 13 4(3 0() 01 5(4 0(1) 0.3 Insect visitors [om Site: Agallpampa, Peru Hummingbird visitors Adelomyia Colibri coruscans melanogenys Myrtis fanny Polyonymus caroli Bombus Apis mellifera 8004 p JequinN ‘S6 SUNJOA ‘ye 19 ujus Plant pair observed; time observed, 2 2 2 par edu y% par edu Ye interplant distance par edu X par edu yt par edu X par edu X. I. parvifolium with 220 flowers 5 I Plant visits 0 (1) 2 (1) 0.3 2 (2) (2 2(1) 0(1) 0.9 edule with 3 owers; 2 hr., m Flower visits 0 (7) 16 (9) 19.57 15 (21)34 (28) 13.6” 7 (3) 0 (4) 7.3" I. parvifolium with x ied Es L Plant visits 2(5) 10(7) 9 1(4 2(5 2.6 edule with 60 flowers; 4.75 hr., 5 m Flower visits 9 (66) 136 (79) 88.5" 125 (80) 50 (95) 46.6 I. parvifolium with 175 flowers and 7. Plant visits 1 (6) 16 (11) 5.7 1(0)0(1) 0.1 6(2) O0 (4 7.87 edule with 300 flowers; 5 hr., 2.5 m Flower visits 34 (108) 258 (184) 78.6" 2 (1)0 (1) 1.3 26 (10) 0 (16) 41.9" umb edu Ne umb edu y umb edu Y umb edu XY umb edu y? umb edu X I. umbellatum with 160 flowers and /. Plant visits O (6) 18(12 687 3(1) 0(2 38 9 edule with 350 flowers: 2.3 hr.. 9 m Flower visits 0 (39) 125 (86) 55.77 4 (15) 0 (34) 104^ Pwosyoo) jo KBojoig uoneuiod 609 Annals of the Missouri Botanical Garden Qu SRI I. cyaneum I. lehmannii Sau AYABACA I. cornifolium 8.1% GUZMANGO Fi percentage of interspecific movement observed between pairs of individual plants from different species indicate no observed interspecific. pollinator movement. Sites where I. peruvianum n s I. edule | I. parvifolium NS. tl AGALLPAMPA edule I. umbellatum AGALLPAMPA Observed interspecific pollinator movements among sympatric species in mixed du i Arrows indicate and er ros ssed arrows | pairs were observed are given be is w pairs in capital letters. The percentage between Jochroma peruvianum and I. cornifolium is based on data from 23 visits to two pairs of plants, between /. lehmannii and Í. cyaneum on 60 visits to two pairs of plants, / . edule and I. parvifolium on 52 visits to three pairs plants, and 7. edule and f. umbellatum on 21 visits on one pair of plants. Display sizes for these pairs are given in Table evening period (data not shown); thus, it was not surprising to find substantial night pollination in this species. Although no evening visits were observed for /. umbellatum, it could be visited by the same night-flying moths that visited the sympatric Z. edule in the evening. It is unclear what could account for the night pollination in Z. corntfolium, but the pollen loads (large relative to bagged flowers and comparable to visited flowers implicate unidentified, nocturnally active animal visi- tors. Overall, however, the relatively small amounts of nocturnal pollen. deposition in most. taxa provide assurance that our largely diurnal pollinator observa- tions covered most of the pollinator activity. Thirty-four of the 47 pollinator species observed during the study visited virgin (previously bagged) flowers, allowing for estimation of pollen deposition. Visits by all species but one (an unidentified syrphid fly) resulted in pollen deposition (data not shown). Overall, hymenopterans deposited more pollen per visit on average than other elasses of pollinators (Table 3). Dipteran visits resulted in the smallest deposition, on average (Table 3). Considering that the flowers have 50 to 500 ovules per flower (depending on species: Table visits by most pollinator classes, except for dipterans, 5). our deposition estimates suggest that single would result in enough pollen deposition to potentially fertilize all of the ovules. For instance, the average hummingbird visit deposited 264 viable pollen grains this on a stigma of fochroma ayabacense and, since species has on average 124 ovules per flower, a single visit could potentially fertilize all the ovules. Combining pollen deposition (quality) with visita- . we estimated the importance of — Hon rale (quantity each class of pollinators. For /ochroma gesnerioides and /. umbellatum, pollen counts were not available for the lepidopteran visitors. In the case of. /. gesnertoides, we used the counts from its closest o estimate — l. fuchstoides, lepidopteran For f. lepidopteran pollen deposition from Z. confertiflorum, relative, importance. umbellatum, we used the average because for these two species, a hesperid butterfly the pollinator importance estimation, species was sole lepidopteran visitor. In. this we scaled pollen deposition to equal the proportion of ovules poten- tially fertilized by a single visit. Since this proportion was 1.0 in most cases, pollinator importance values were similar to relative visitation (Table 5) as in Olsen (1997). Thus, hummingbirds appeared to be the most important pollinators for most lochroma species, with only a few having either mixed bird-insect pollination or exclusively insect pollination (Table 5). DISCUSSION SPATIO-TEMPORAL CONSIDERATIONS Interactions. between plants and their pollinators are subject to temporal variation within a day, within a 1988: Schemske € season, and across years (Herrera, Volume 95, Number 4 Smith et al. 2008 Pollination Biology of lochroma Table 5. Relative visitation rates and pollinator importance. Sealed pollen Relative flower visitation helative portar Average deposition! ales? importance? ovules per Species flower + SE Tro Hym Lep Dip Tro Hym Lep Dip Tro Hym Lep Dip Acnistus arborescens 52.0 + 3.5 l | l 0 0.19 0.21 0.60 0 0.19 0.21 0.60 lochroma ayabacense 123.5 + 13.1 | — — E 1.00 0 0 0 1.00 0 0 0 I. calycinum 188.0 + 29.3 | l — 0.28 0.65 0.17 0 0.18 0.74 0.20 0.00 0.06 I. confertiflorum 152.3 + 15.1 | l | — 0.78 0.19 0.03 0 0.78 0.19 0.03 0 I. cornifolium 171.0 + 22.6 l l — 096 0.04 0 0 0.96 0.04 0 0 I. cyaneum 368.1 + 20.1 l l 0.24 — 0.80 0.13 0.07 0 0.84 0.14 0.02 0 ÍI. edule 414.7 + 15.6 | | l — 0.98 0.01 0.01 0 0.99 0.01 0.01 0 I. ellipticum 53.4 + 13.0 — | l | 0 0.38 038 023 0 0.38 0.38 0.23 I. fuchsioides 462.7 X 7.6 0.60 l 0.69 0 0.77 0.16 0.06 O 070 0.24 0.07 0 Í. gesnerioides 204.7 + 19.8 l l | - 0.90 0.06 0.04 0 0.90 0.06 004 0 I. lehmannii 69.3 + 5.6 | | _ — 0.99 0.0] 0 O 0.99 0.01 0 0 I. loxense 254.7 + 21.7 l | 0.89 0.11 0 0 0.89 0.11 0 0 Í. parvifolium 09.1 +60.9 0l l — - 0.92 0.08 0 O 0.70 030 0 0 I. peruvianum 230.3 + 11.6 | l 0 0 0.44 0.56 0 0 0.44 0.56 0 0 Í. stenanthum 146.0 + 12.5 | l — — 0.99 0.01 0 0 0.99 0.01 0 0 I. umbellatum 134.3 + 7.2 | l l 0 0.52 0.47 4X 10% 0.01 0.32 0.67 0.01 0 Tro = Trochilidae, Hym = Hymenoptera, Lep = Lepidoptera, and Dip = Diptera. ' Pollen deposition was scaled to equal the proportion of ovules pote ntially fertilized by a single visit. ? Relative visitation rates are the numbers of flowers visited per flower hour, rescaled to sum to * Relative pollinator importance is the product of the relative 2005). but we 2003; Price et al., he did not sample in multiple years or times of the year Horvitz. 1989; Here, we sampled across 1 Ivey et al., > times of the day, and each study took place during a three- to five-day period. The issue of temporal variation is miligated by the fact all studies took place in the same season (the rainy season), during which aa conditions did Also, pollinator fauna in the study areas are Me as 1960), and, thus, rom year lo year, not vary substantially from day to day. opposed to migratory (Greenewalt, [ Casual observations in the same locality across years might be expected to shift less (8. D. Smith, unpublished data) showed some variation in the animal species visiting a particular /ochroma but not in the ied class of pollinator . bird or insect). Associations of plants with ind of pollinators are probably more robust tH the pollinator assemblage. across time than the specific composition of Plant-pollinator associations may also vary across .. across sites (Boyd, 2004; Price et al., 2005) or along environmental gradients (Scobell & Scott, 2002: Herrera, 2005). Here, we have conducted studies of each species at a single site. Observations spatial scales, e.g of several study taxa in other sites suggested that relative visitation of different classes of pollinators is similar across the species range despite variation in pollinator fauna (S. D. Smith, pers. obs.). For instance, in El Cisne. Ecuador. visited by the hummingbird Amazilia amazilia Lesson, lochroma cyaneum is mostly genys. across siles, 1.0. visitation and sealed deposition, rescaled to sum to 1.0. it is mainly visited by the whereas in Ayabaca, Peru, hummingbirds Coeligena iris and Adelomyia melano- Thus, while pollinator composition may vary be principally cyaneum appears pollinated by hummingbirds across its range. FLORAL DIVERSITY AND POLLINATOR RELATIONSHIPS One goal in undertaking this study was to determine if the floral diversity in /ochroma corresponds to a diverse set of pollinator systems. At the broadest level, we observed three basic modes of pollination (bird, ). Unlike other Andean taxa for which comparative pollination studies have been undertaken (Kay & Schemske, 2003; Pérez 2006). appeared only weakly related to 2008). axa Acnistus arborescens and 1. ellipticum were both mixed bird/insect, and insect pollinated in lochroma = et al, pollination systems floral differences (see also Smith et al., The insect-pollinated l white, scented, and offered a low reward, but varied in shape and size (Fig. 1). The mixed bird-/insect- pollinated species /. peruvianum and 7. umbellatum differed in flower color (one green, one orange), but l intermediate flowers and 1 shared two traits, small 'able 1). The greatest the 12 colors included red, rewards (on a per-plant basis) variation observed bird- t white, yellow, blue, and purple and whose corolla size ). among bird-pollinated species, however, was a large floral was among pollinated taxa, whose flower varied nearly three-fold (Fig. 1). One common feature Annals of the Missouri Botanical Garden axa had e nectar reward; all hummingbird-pollinated higher rewards than the mixed or exclusively insect- pollinated taxa (see also Smith et al., 2008). This observation is in accord with other studies showing that the amount of reward is more important than color) in determining lias & Collias, 1968; Melendez-Ackerman et al., visual cues (e.g. flower hummingbird visitation (Co Stiles, 1976; We also considered the possibility that the diversity of flower form among species sharing the same pollination system could reflect lower-level speciali- zation, e.g., for particular pollinator species. However, ~ we found no evidence to support such an explanation. A single hummingbird or insect species (e.g., Adelomyia melanogenys and Apis mellifera L.) was observed visiting multiple species of fochroma, and, conversely, à visited by gB single plant species was multiple pollinator species (Appendix 2). For exam- ple, an average bird-pollinated Jochroma species was visited by 2.4 hummingbird species, and an average insect-pollinated insect species by 8.0 species. Furthermore, measurements of pollen deposition suggested that the vast majority of these visitors were effective. pollinators. Thus, it appears that /ochroma species do not have tightly coevolved, specialized pollination systems. Despite visits by many pollinator species, it is possible that a given /ochroma species could be specialized on a guild or functional group of pollinators, which collectively explain the particular floral traits seen. If this were the case, one might expect a lack of pollinator sharing among geograph- ically proximate but florally distinct /ochroma spe- cies. However, our analysis of pollinator assemblage similarity showed that diverse taxa from the same geographic region and from the same study site shared pollinator species significantly more often than those (Fig. 2). This i consistent with the idea that /ochroma species are from different regions or sites 7 generalists within a broad class (such as humming- birds) and that they tend to be visited by whichever pollinator species are locally abundant. POLLINATOR BEHAVIOR AND REPRODUCTIVE ISOLATION The overlap in pollinator assemblage among sympatric taxa has significant implications for the maintenance of species boundaries. Considering that nearly all pollinators were capable of transferring loads of pollen in excess of the number of ovules on any given visit, any foraging bout that included visits to multiple species would almost certainly result in interspecific pollen flow. All sympatric species studied here shared at least two pollinator species, and some as many as four (Appendix 2), making interspecific pollen. flow in potentially On the other visitation rates sympatry common. hand, subtle differences in and patterns appear to restrict interspecific pollen flow. First, PS values for sympat- ic laxa, even those with the same broad. pollination system, were typically much lower than | (range, 0.20-0.96; mean, 0.51), reflecting differential visita- tion by pollinator species. Second, we observed that individual pollinators do not tend to move between sympatric species even when the plants are growing This could be explained E individual preferences (Jones € Reithel, 2001) « optimal foraging (Heinrich, 1976; Waser, 1986), but is perhaps better explained by territoriality. side by side (Fig. 3). = Areas of lochroma sympatry contained several hummingbird species, including small birds (e.g., Adelomyia melanogenys and Polyonymus caroli) and larger birds (e.g.. Colibri coruscans and Coeligena iris). As mentioned previously, smaller hummingbirds tended to visit less nectar-rewarding species in mixed populations, even though there is no mechanical barrier preventing them from retrieving nectar from the more rewarding species. Larger, more aggressive birds dominate the more rewarding species and defend individuals of the rewarding species from the 1978; Sules, territorial smaller birds (Feinsinger & Colwell. 1981). One that might prevent or reduce gene flow among plant can envision behavior species in mixed patches, and that this effect would be enhanced by differences in reward (Table 1). The combination of local spatial separation of populations, perhaps due to microhabitat specialization, and hummingbird territoriality might then reduce the potentially frequent interspecific pollen. flow in sympatry. The low incidence of hybridization among sympatric Jochroma might reflect individual pollinator fidelity driven by interactions among hummingbirds. additional pre-fertilization mechanisms (e.g.. pollen competition), and/or post-fertilization mechanisms. Although artificial interspecific crosses suggest wide crossability among /ochroma species (Smith & Baum, 2007; S. D. Smith, unpublished data), none of the spec ies pairs that grow in sympatry without observed hybrids have yet been examined. Thus, additional crossing studies and tests of hybrid fitness will be required to understand how /ochroma species coexist in northern Andean communities. Literature Cited Boyd, A. E. 2004. Rolas eae) and Breeding syster m of Macromeria pad Minator 1985. The pollination between two mechanism of asse Campbell, D. R. & A. F. competition for Motten. forest herbs. “cology 66: 554-5€ 63. Volume 95, Number 4 2008 Smith et al. Pollination Biology of lochroma 613 Cocucci, A. 1999. Evolutionary radiation in Neotropical Solanaceae. Pp. 9-2 a M. Nee, D. E. Symon, R. & J. P. J I lit Solanaceae IV: Biology and URA Royal Botanic United Kingdom Collias. N. E. & E. ;. Collias. 1968. Anna's Hummingbirds trained to select Er colors in ) l. z4 Lester Advances i In Kew. Carde ns, feeding. Condor 7 273-27: Dafni. A. 1992. ford University Press, Pollination E yu Se Practical Approach. )xford, United Kingdom. & L. van der Pijl. 1966. he Prine iples E n ted King v R. K. Colwell. 1978. aad od organiza- Zool. de K. Ecology. qued Press, Oxford, Uni ee Feinsinger, P. tion among Neotropical nectar-feeding birds. Amer. 79-795. 18: 77 Fe ao r, C. B., W. S. Armbruster, P. Wilson, M. R. Dudash & J Thomson. p Pollination de and floral spec 5d Annual Rev. Ecol. Syst. 35: 375—403. Gentry, A. H. 1982. Neotropical floristic diversity —Phytogeo- g puris 'al connec tions between Central and South. America, "ne cli oe or an accident of the hea Jer 976. a specializations of individual Ecol. Monogr. 46: 105-128. Heinrich, B bumblebees. Herrera, C. M. 1987. ía analysis of a dive Oikos 50: 79-90. ———. 1988 mosaic of a pollinator assemblage. Biol. J. Linn. Soc. 35: 95-125. Components of io quality”: insect asse e . Variation in mutualisms: The spatio-temporal 35 Plant generalization on pape e ies property or o al " henomenon? A J. 92: 13-2( A. T. 2001. T © mer. . Bot. Hunziker he genera of Solanaceae. 1 Ganter Ve rlag K. = Ruggell, Li jec Ea n. Ivey, C. T., P. Martinez & R. Wy 003. Variation in pollinator effectiveness in swamp pea Asclepias incarnata (Apocynaceae). Amer. J. Bot. 90: 214—225. Janzen, D. H. 1973. Comments on host-specificity i: tropical ie and its relevance to species richness. Pp. 201— 211 in V. H. Heywood fenton) ). Taxonomy dud Ecology. Press, New ‘or . E. Steiner. spec lation in plant pollination systems. Evol. Jones, C. P La Academic Johnson, S. D. 2000. Generalization versus Trends Ecol. 8. Pollinator constancy as a pre-pollination isolating mec cee oe en sympatric species of Cerci- dium. Evolution : ( N. & Reith L 2001. Pollinator-mediated selection on a flowet color polymorphism in Jones, B ental d of Antirrhinum (Scrophulariaceae i Amer. J Bot. 88: 447—45 Kaiser, R. 2000. Scents from rain forests. Chimia 54: Kay, K. M. € D. W. Schemske. 2003. and visitation rates for 11 Dm of Neotropical Costus 98-207. eographic patterns in E Ecology 85: "Aperi im 346—302. Pollinator assemblages (Costaceae). o 35: ] 004. network: ~ pollinator rie Comment. 875-878. Kearns, C. A. € D. W. Inouye. 1993. Techniques for Pollination 4 Biologists Unive rsity of Colorado Press, Boulder. UN s Kohn, J. ; N. M. Waser. 1985. Th inen Per on seed set in E ect of Delphinium Ipomopsis a X EC a competitor for hummingbird pollination. Amer. J. Bot 1144-1148. Lagerheim, G. 1891. Zur biologie de Jochroma macrocalyx. Deutsch. Bot. T 9: 348-351. Luteyn, J. L. 20( EDR adaptation, and endemism in Ne D oie al ricaceae: Biogeographical patterns in the /accinieae. Bot. Rev. 68: 87. Martin, F. W. 19 the style by 68: 55— 59. Staining and observing pollen uos in means of f| 5 19g uorescence microse opy. l'echnol. 34: 125-128 Mayfield, M., N. M. Waser & M. Price... 2001. Exploring ‘the st effective po aia principle’ with al complex flowers: cae and /pomopsis a Ann. Bot. 88: 591-5 Mele Boe z-Ackerman, E., D. R. Campbell & N. M. Waser. . Hummingbird behavior 2 mechanisms of selec- tion on flower color in /pomopsis. Ecology 78: 2532-2541. Muchhala, N. 2006. The pollination biology of Burmeistera (Campanulaceae): 2m ialization and syndromes. Amer. J. Bot. 93: 1081-10 Olesen, J. M. & P. DEM 2002. Geographic patterns in plant- Rn mutualistic networks. Ecology 83: 2416-2424, Ollerton, J. & L. Cra ". 2002. Latitudinal trends in plant- pollinator i i ractions: e tropical plants more special- ised? Oikos —350. Olsen. K. M. If "€ nen effectiveness and pollinator importance in a population of Heterotheca subaxillaris (Asteraceae). Oecologia 109: 114—121. Pérez, F., M. T. K. Arroyo . Hershkovitz. 2006. Ancestral reconstruction of over es and pollination systems in Schizanthus (Solanaceae). Amer. Bot. 93: 1029-1038. Perret, A. pes R. Spichiger, M. Peixoto & ae sn. 2001. Nectar sugar composition in relation to ). Ann. e od sy Me in Sinningieae (Gesneriaceae Bot. 87: 267-273. Price, vs V., N. M. Waser, R. E. Irwin, D. R. Campbell & A. K. Brody. 2005 bond and spatial variation in pollination of a montane herb: A seven-year study. Ecology 86: 2106-2116 Schemske, D. W. 1981. Floral convergence and pollinator sharing in two bee-pollinated tropical herbs. Ecology 62: 946—954. Brokaw. 1981. Tree a and the d of understory iem in a tropical forest. Ecolo —945 & C. C. Horvitz. 1984 visitors in ee ability; a precondition for mutualism Science 225: 519-521. "1989. Temporal variation in selection on a floral Pu ter, E lin 43: 461—465 Scobell, S. A. & P. E. 2002. Visitors aid floral traits of a B s e cactus (Echinocereus coccineus) m janon among e | ecialization. show only minor variation along an elevational gradient. Amer. Midl. ist 147: 1-15 Silander, J. A. & . Primack. 1978. Pollination intensity and seed set n is vening primrose (Oenothera fruticosa). Amer. ow Naturalist 100: 213-216 Smith, S. I . 2006. Floral Diversification and Pollination Biology o of the scien Clade loc e (Solanaceae). Ph.D. Thesis, Univer & D. A. Baum. RD elements of the florally diverse PE an clade lochrominae (Solanaceae). 2. Bot. 93: -115 A of Wisconsin dison. Amer. 2007. Systematics of lochrominae MM CEDERE (Solanaceae): Patte 'Tns in vue diversity and interspecific crossability. Acta Hort. : 241-2: . Ané & D. A. A 2008. "The role of pollinator shifts in the floral diversification of Jochroma (Solanace- ae). Evolution 62: 793-806. 614 Annals of the Missouri Botanical Garden ——— & M. V. Price. 1983. Optimal and actual outcrossing in PN m the nature of. plant-pollinator interaction. Pp. 341—359 in C. E. Jones J. Little: (editors). . 1981. Geographical aspects of bird-flower coevolu- Handbook i Ex ne rime i al Ann. Nos Kus p pe r Weigend, p da cause and due is baca zone in northern Peru. Bot. Re )3—002. 68: 38-54 Taste preferences, color preferences, 78: 10- Stiles, F. G. 1976, in hummingbirds. Condor 78: and flower choice 2 QS R. Pollination Biology. Van lion, aah particular e nce to Central America. Missouri Bot. Gard. € 23—95]. Waser, N. M. 1986. F bu ər constancy: measurement. Amer. Naturalist 127: Yo a rvalions on the biogeography of the lev. Appendix 1. Description of plant populations used in pollination studies in Ecuador and Peru. Plants Avg. display Hours Flower hours Species Voucher Study dates studied per plant + SE observed observed! Acnistus arborescens (L.) Schltdl. Ecuador, Alluriquin, 0. 132 S 18.99 W. S. D. Smith 209 (MO, QCNE, WIS) 26—28 Dec. 2002 8 2144 12.1 208.0 lochroma ayabacense S. Leiva a eae 4.63 8 19.71 W. S. D. Smith & S. Leiva G. 337 (F. HAO, 17-19 Jan. 2004 6 186 + 45 14.9 3156.3 NY. USM, WIS) I. calycinum Benth. p. Guajalito, 0.25 8 76.81 W, S. D. Smith 471 (F, QCNE, WIS) 28-31 Dec. 2004 6 30 x II 18.7 784.6 Il. confertiflorum (Miers) Hunz. Ecuador, Malpote, 1.91 5 78.97 W, S. D. Smith & L. Lopez 482 (MO, ! NY, 16-18 Jan. 2005 6 112 + 37 14.1 1770.8 OCNE, WIS) I. cornifolium (Kunth) Miers Peru. Guzmango. 7.39°S 78.90 W, S. D. Smith, S. Leiva G., S. J. Hall & A. 27-31 Jan. 2004 9 87 = 16 23.9 2215.2 Rodriguez 337 (F. HAO, MO, NY, USM, WIS L cyaneum (Lindl.) M. L. Green Ecuador, Cisne, 3.86 5 79.43 W, S. D. Smith 227 (MO, QCNE, WIS) 7-11 Jan. 2003 14 12 + 6 36.6 1354.0 . edule S. Leiva i^ Pus 1.95 5 18.56 W, S. D. Smith, S. d G., S. J. Hall 6-8 Feb. 2004 6 311 + 66 15.1 1773.2 . Rodriguez 359 (F, HAO, MO. NY, USM IS) I. ellipticum (Hook. f.) Hunz. E Ls Los Gemelos, 0.63 5 90.38 W, P. R. E rdo 15022 (C piss 8-10 May 2004 10 507 12.0 200.0 I. fuchsioides (Humb. & Bonpl.) Ecuador, Cojitambo, 2.75°S 78.89°W, S. D. Smith & L. Lopez 488 (F, N 21-23 Jan. 2005 6 79 + 14 21.1 1792.6 iers CNE, WIS I. gesnerioides (Kunth) Miers Ecuador, Pululahua, 0.039°N ion 50" °W, S. D. Smith 200 (MO, QCNE, WIS) 19-21 Dec. 2002 5 505 + 253 12.0 7640.0 I. lehmannii Bitter? Peru, Ayabaca, 4.63°S 79.71 W, 5. D. Smith & S. Leiva G. 330 (F, HAO. 21-23 Jan. 2004 5 275 + 119 13.2 3916.7 MO, NY, USM, WIS) I. loxense (Kunth) Miers Ecuador, Loja, 3.98°S 79.23 W, p Smith 499 (OCNE, WIS) 26-28 Jan. 2005 8 120 + 38 13.4 1822.1 I. parvifolium (Roem. & Schult.) p. dus 1.95 8 18.56. i. » D. | pU. 5. Leiva G., 5. J. Hall 4-8 Feb. 2004 6 98 + 38 15.1 1806.1 D'Arcy A. Rodriguez 303 (F, HAO, S NY, USM, WIS) I. peruvianum (Dunal) J. F. "n Tux s 7.398 78.90 W, S. D. n a S. J. Hall 353 (F, HAO, 29—3] Jan. 2004. 4 85 + 18 11.7 830.0 Macbr. F. MO, NY. USM, WIS) I. stenanthum S. Leiva, Peru, Guzmango, 7.39°S 78.90°W, S. D. Smith, S. Leiva G., S. J. Hall & A. 29-3] Jan. 2004 5 209 + 52 12.0 2511.0 Quipuscoa & N. W. Sawyer Rodriguez 313 (F, HAO, MO, NY, le WIS) I. umbellatum (Ruiz & Pav.) Peru, Agallpampa, 7.95 8 18.56 W, S. D. Smith, S. er G.. S. J. Hall 6—8 Feb. 2004 4 8l + 40 17.8 1550.8 Hunz. ex D’Arcy & A. Rodriguez 360 (F, HAO, m NY, USM, WIS) Total 263.58 36331.24 ! Flower hours were calculated from a series of observations (1 to n) as follows: FH = (Dj X Tj) + (Do X Ts) +... (D, X Th), where FH equals the number of flower hours, D is the display size (number of flowers on the plant), and T equals the number of hours the plant was observed. ? Variation in display size across the population was not recorde di in lochroma ellipticur ? We treat Z. lehmannii and the recently described /. squamosum S. Leiva & Quipuscoa as synonyms. 800€ p JequinN ‘S6 euin|oA euioJ420| jo KBojoig uoneurpod ‘JE 19 uius S9 616 Annals of the Missouri Botanical Garden 2, Pollinator taxa and visitation rates. Animals are listed by order and identified to species when possible, and Appendix 2. only legitimate pollinators (which made contact with the reproductive organs) are included. Each unidentified visitor is given a unique number within its taxonomic group, and these identifiers are oe across plant species. For example, both lochroma cornifolium and I. peruvianum were visited by Vespidae, sp. indet. 1. Only pollinators observed during the individual species studies are included; additional visitors observed during subsequent experiments (e.g., Myrtis fanny; Table 5) are not listed here. For each pollinator, plant and flower visitation rates (per flower hour) are listed (separated by a slash). Acnistus arborescens Hymenoptera Apidae, ipa mellifera L., 0.014/0.048 sp. indet. 1, 0.010/0.034 Lepidoptera sp. indet. 2, 0.010/0,091 Diptera ode . sp. indet. 1, 0.024/0.058 Tipulidae, sp. indet. 1, 0.043/0.115 sp. indet. 1, 0.019/0.048 sp. indet. 2, 0.005/0.019 sp. indet. 3, 0.005/0.014 lochroma ayabacense Apodiformes Trochilidae, Adelomyia melanogenys Fraser, 0.005/0.089 Trochilidae, Coeligena iris Gould, 0.005/0.097 l. calycinum Apodiformes Trochilidae, Coeligena torquata Boissoneau, 0.001/0.013 Trochilidae, Phaethornis Swainson, 0.001/0.045 Trochilidae, sp. indet. 1, 0.005/0.033 Hymenoptera Apidae, Parapartamona ipee Moure, 0.001/0.001 Apidae, Plebia sp. 2. 0.01 7/0.023 Diptera Drosophilidae, sp. indet. 1, 0.025/0.025 I. confertiflorum Apodiformes Trochilidae, Adelomyia melanogenys, 0.007/0.000 Trochilidae, Helian w viola d 0.004/0.082 Hymenoptera Apidae, Apis mellifera, 0.002/t Apidae, Parapartamona vittigera, ( oe Halictidae, earn sp. indet. 1, 0.008/0.010 Lepidoptera Hesperidae, sp. indet. 1, 0.002/0.006 l. cornifolium Apodiformes Trochilidae, Adelomyia melanogenys, 0.004/0.051 Trochilidae, Coeligena iris, 0.008/0.125 Hymenoptera Apidae, Apis mellifera, 0.003/0.006 Vespidae, sp. indet. 1, 0.001/0.001 lI. cyaneum Apodiformes Trochilidae, Amazilia amazilia Lesson, 0.032/0.188 Trochilidae, Coeligena iris, 0.008/0.049 Hymenoptera Vespidae, Polybia sp. e 0.027/0.040 Lepidoptera Sphingidae, sp. indet. 1, 0.001/0.007 sp. indet. 1, 0.007/0.0 a l. edule Apodiformes Trochilidae, Adelomyia melanogenys, 0.002/0.033 Trochilidae, Colibri coruscans Gould, 0.003/0.031 Trochilidae, Polyonymus caroli Bourcier, 0.005/0.089 Hymenoptera Apidae, Apis in de 0.0004/0.001 Lepidoptera Noctuidae, sp. indet. 1, 0.0003/0.001 1. ellipticum Hymenoptera Formicidae, P A 0.002/0.002 Formicidae, Wasmannia auropunctata Roger, 0.002/0.002 Vespidae, Pachodynerus galapagoensis W saree 0.003/0.005 Lepidoptera Geometridae, Oxydia lignata Warren, 0.002/0.4 Noctuidae, Agrotisia williamsi Schaus, 0.002/0. a Noctuidae, sp. indet. 2, 0.002/0.003 Diptera Syrphidae, Nanthandrus agonis Walker, 0.005/0.005 l. fuchstoides Apodiformes Trochilidae, Coeligena torquata, 0.001/0.021 Trochilidae, Heliangelus viola Gould. 0.008/0.2 14 Hymenoptera Apidae, Apis mellifera, 0.004/0.027 Halictidae, Caenohalictus sp. 2, 0.016/0.022 Lepidoptera Danaidae, sp. indet. 1, 0.003/0.020 Volume 95, Number 4 2008 Smith et al. 617 Pollination Biology of lochroma Appendix 2. Continued. I. gesnerioides I. lehmannii I. loxense Í. parvifolium Í. peruvianum ¡A stenanth um I. umbellatum Apodiformes Hymenoptera Lepidoptera Apodiformes Hymenoptera Apodiformes Hymenoptera Apodiformes Hymenoptera Apodiformes Hymenoptera Apodiformes Hymenoptera Apodiformes Hyme noptera Lepidoptera Trochilidae, Adelomyia melanogenys, 0. en 040 Trochilidae, Coeligena torquata, 0.0001/0.004 vescens pus s, 0.004/0.102 a Trochilidae, Boissonneaua fla Trochilidae, sp. indet. 3, NE dd 03 Apidae, Apis mellifera, 0.001/0.0 Sphingidae, sp. indet. 2, Mes Adelomyia ami: 0.002/0.148 Trochilidae, Coeligena iris, 0.005/0.190 Apidae, Apis ie ce redd 003 Apidae, Plebeia, 3/0. ya ud : ae 251 Trochilidae, ^ roc w Apidae, Apis mellifera, 0.001/0.0 en Chilicola cf. peinada 0.005/0.005 Vespidae, Polybia, 0.01 Trochilidae, Adelomyia IA 0.004/0.057 Trochilidae, Colibri coruscans, 0.001/0.001 Trochilidae, Polyonymus caroli, 0.010/0.230 Trochilidae, sp. indet. 2, 0.001/0.021 Apidae, ae i. 0.007/0.027 Vespidae , 0.001/0.001 Troc ia. a. Mi d 0.008/0.047 Apidae, Apis mellifera, 1/0.055 Vespidae, sp. indet. 1, o 05 Trochilidae, Adelomyia rd 0.006/0.192 Trochilidae, Coeligena iris, 0.004/0. Apidae, id mellifera, 0.002/0.00 "n Trochilidae, Adelomyia melanogenys, 0.007/0.153 Trochilidae, Colibri coruscans, 0. o 001 Apidae, Apis mellifera, 0 0.1 Anthophoridae, PME 0. "m "d Hesperidae, sp. indet. 2, 0.001/0.0 Syrphidae, sp. indet. 2, er = © c A SYNOPSIS OF THE GENUS TACHIGALI (LEGUMINOSAE: CAESALPINIOIDEAE) IN NORTHERN SOUTH AMERICA! Henk van der Werff? ABSTRACT Aubl., Suriname, synoplic treatment of the Tachigali van der Werf erl are pus ree. have been taken dr again. One new combination, are designated for three es, T. alba Ducke, T. glauc di including nd Amazonian | i ¿ E, a van der Werff, "ug ¿ sr Werff, T. fusca vi rff, T. inconspicua van der - Werff. . a overlooked names qu by Tulasne (T. melanocarpa, and T. plumbea Ducke. Vogel, Ecuador, in northern South America (Colombia. razil, and Peru) is presented and contains briel ies recognize in in northe rn South America. Ten a cies, T. barnebyt van der Werff, van der Werff, T. Vert, T. sein van der Werff, and T. vaupesiana | chrysaloides glauca, T. poeppigiana, and ‘ T. melanocarpa (Ducke) van der Werff, is proposed. Lec "e E The following species are placed i synonymy: Sclerolobium D cent B. san i bracteosa (Marms) Zaruechi E Pipoly: T. tessmannii a in T. p ume ‘ke T T. myrmec Aue (Ducke) Ducke in 1 ih.) Zaruechi € Here br paniculata A b ] T. sulcata Benoist and T. in T. tinctoria (Benth.) Zaruechi & Herend. guianensis (Be S. radlkoferi R Pad D: | grandiflora Huber, subbullatum Duc in 18, p Le uleanum Harms, T. ee Harms. 7. chra Dwyer in T. bracteolata Dwyer in richardiana: and T. reticulosa (Dwyer) Zanicchi & Herend. Key words: Northern South America, Sclerolobium, Tachigali, taxonomy. TAXONOMIC HISTORY stamens, with three being thicker and backwards curved and seven longer and straight, and the ovary The genus Tachigali was described by Aublet attached to the side of the hypanthium, while in (1775). who recognized two species, 7. paniculata Aubl. and T. trigona Aubl., both from French Guiana. The main difference between the two species given by Aublet was the position of the leaflets: opposite in 7. paniculata and alternate in T. trigona. Jussieu (1789) introduced the orthographic variant “Tachigalia”; all later authors followed this until Lewis (1987) returned to the original spelling “Tachigali.” Tulasne (1844) designated T. paniculata as the type species of the genus. Subsequently, the name T. trigona largely disappeared from use; Dwyer (1954) mentioned it as a synonym of 7. paniculata, a disposition followed 1837, Sclerolobium, with three species from central Brazil. Dwyer (1957a) « type this treatment. In Vogel published the genus esignated S. denudatum Vogel as the Bentham (1840) T. pubiflora Benth.. (1844) treated Tachigali and included five species, species. published a new species, from Guyana. Tulasne four newly described, in the subgenus Tachigali, and i The dimorphic lwo new species his subgenus Cosymbe Tul. subgenus Tachigali was characterized by subgenus Cosymbe, all 10 stamens are equal and the ovary is attached to the bottom of the hypanthium. Tulasne also treated two species in Sclerolobium. one of which, 5. The names published by Tulasne have generally (1845) Tachigali and one i sericeum Tul., was newly described. been ignored by later authors. Poeppig described two species, one in Sclerolobium, both of which had been described by Tulasne (1844) with the same collections as types. In a report on the plants collected by Spruce Para, Bentham (1850) gave a summary of Sclerolobium, in which he recognized nine species, three of which were newly described. In Flora (1870) listed 10 species species Brasiliensis, Bentham Sclerolobium and three of Tachigali. Taubert, in Engler and Prantl (1892). accepted five species in Tachigali and 12 species in Sclerolobium. Taubert divided Sclerolobium in two sections, section Eusclerolobium (10 species) with linear petals and section Platypetalum (two species) with more or less broad petals. One of the two spec ies in. section. Platypetalum, S. aureum (Tul. ! Loans from BBS, BM, COL, A HB, IAN, INPA, K, MBM, MEL, MER, MG, NY, P, R, RB, SP, US, and VEN : gratefully acknowledged. I partie bu want to thank G. Lewis for his the detailed review; his comments have greatly improve " the manuscript. Falso thank the curators and staff at BM, K, L, Mye rs prepared the illustrations. Fred Keusenkothen prepared the images 2 i^ nes Missouri 63166, “Missouri Botanical Garden, P.O. Box 299, St. Louis, doi: 10.3417/2007159 ANN. Missouri Bor. and P for their assistance and hospitality duis my visits. Johi y species, . henk.vanderwerff@mobot.org. Garb. 95: 618-660. PUBLISHED ON 30 DECEMBER 2008. Volume 95, Number 4 2008 van der Werff Synopsis of Tachigali — Baill., had been placed by Tulasne (1844) in subgenus Cosymbe of Tachigali. Rusby (1896) described a new species of Sclerolobium from Bolivia. From about 1900 onward, a sizable number of new species were described by two botanists working in Brazil, Ducke. Huber (1908) described two new species of Tachigali followed by three new species of p (Huber, 1910). Ducke (1915, 1922, . 1938, 1944) described nine See and six spe Huber and first cles oO knew and species of Tachigali. Ducke was an excellent collector, the field contributed greatly to our knowledge of Tachigali. species well from his experience, His observations on habitat preference of the various species are very valuable. For most species de- Ducke, type specimens. seribed by he used his own collections as the same period, Harms (1903, 1906, 1915, 1922, 1926, 1940) described five species of Seler- olobium and five species of Tachigali. Harms studied collections made by other botanists (Ule, Tessmann, Weberbauer, Rusby, Melinon) and had not seen the plants in the field. Several of the species described by him have been placed in synonymy by later authors. Benoist (1919) added a species of Sclerolobium and of Tachigali (Benoist, 1925), both based Macbride (1943) described S. rigidum J. F. Macbr. based on a Klug collection from Peru. The next major contribution was later a species on his own collections. made by Dwyer, who published a revision of Tachigali in 1954, species and a new variety, Sclerolobium (Dwyer, 1957a) with the descriptions of one new species and three new varieties from the area including the descriptions of three new followed by a revision of covered by this synopsis. A key to the largest group of Sclerolobium was published separately (Dwyer, 1957b). Dwyer (1958) published an additional species of Tachigali, Since 1960, five species of Sclerolobium have been described: S. froesii Pires (Pires, 1960), S. dwyeri R. S. Cowan, 5. pimichinense R. 5. Cowan (Cowan, 1961), 5. micranthum L. O. Williams (Williams, 1965), and 5. prancei H. S. Irwin & Arroyo (Irwin & Arroyo, 1974). Only one species of Tachigali has been described Pipoly T. schultesiana Dwyer, from Colombia. Dwyers publications: T. vasquezii 1995). With the large number of species in it became increasingly since (Pipoly, Sclerolobium and Tachigali, clear that the differences between the two genera were tenuous, and there is now a general consensus that the two genera should be merged under the older name Tachigali. Several species of Sclerolobium were recently transferred to Tachigali or given new names in three floras and a checklist (Zarucchi € Heren- deen, 1993; Pipoly, 1995; Barneby, 1996; Zarucchi, 1998). TACHIGALI VERSUS SCLEROLOBIUM Aublet listed, among others, the following characters: In the original publication of Tachigali, (1775) stamens unequal, with three short, erect and pressed against the upper petals, and seven slender, resting on petals unguiculate; three the lower petals; upper petals reflexed, the lower two spreading; four sepals erect and one bent down. Vogel (1837) did not mention unequal stamens or differences in the orientation of the sepals and petals in his description of Sclerolobium; he did mention that the petals were nearly so ("fila angustissimo- linear o In his description of Tachigali, Vogel thread-like, linearia"). mentioned unequal petals, unequal stamens, and the attachment of the stipe of the ovary to the upper side The descriptions were based represent the extremes of of the hypanthium. species on which these the range of floral variation found in the Tachigali— Sclerolobium complex. Tulasne (1844) recognized two subgenera in Tachigali: subgenus Tachigali, charac- terized by unequal stamens and the pistil inserted along the upper wall of the hypanthium, and subgenus Cosymbe, with equal stamens and the pistil inserted at the bottom (or center) of the hypanthium. The latter subgenus was placed in Sclerolobium by later authors (as section Platypetalum [Taubert, 1892)) 1957a). The place of attachment of the pistil was a Tulasne and became later — (Dwyer. character first correlated with the shape of the calyx cup: symmet- rical when the pistil is inserted at the bottom of the usec N hypanthium, oblique when the pistil is attached along the side of the hypanthium. During the past two decades, several authors have indicated a close relationship between the two genera. In 1993, Gentry stated Sclerolobium * adequately distinct from Tachigali." Herendeen (1993) merged Sclerolobium in Tachigali, as did Barneby and Heald in Mori et al. (2002), Pennington et al. (2004), Lewis in Lewis et al. (2005), and Silva and Lima (2007). Detailed merging Sclerolobium and Tachigali were, “is probably not Zarucchi and reasons for however, not presented in any of these publications. The present study finds that the species of Tachigali can be divided into five groups based on the characters used by Aublet, Vogel, and Tulasne (petals linear or broad; stamens equal or unequal; hypanthium sym- metrical or asymmetrical; and sepals and petals erect These species groups are or spreading/reflexed). © presented in Table 1. The first and largest group corresponds with Sclerolobium sensu Vogel; the second group with Tachigali subg. Cosymbe sensu Tulasne (= Sclerolobium sect. Cosymbe sensu Dwyer): the third group is close to subgenus Cosymbe, but differs in having flowers with spreading sepals and 620 Annals o Missouri Botanical Garden Species groups in Tachigali. Names in boldface refer to species that are not monocarpic (collections exist with leaves and fruits); italicized names refer to species with swollen domatia in the leaf stalk. Pable 1. ls broad, stamens »roup 5: peta ( unequal. calyx asymmetrical, sepals/ Group 3: petals broad, stamens Group 4: petals broad. stamens Group 2: petals broad, stamens equal, calyx asymmetrical. alyx symmetrical. o equal, equal, calyx symmetrical, Group 1: petals linear, stamens equal. sepals/petals erect sepals/petals spreading sepals/petals spreading petals spreading epals/petals erect S calyx symmetrical, alba T T. candelabrum T. macbridei T. amplifolia T. barnebyi T. dwyeri T. catingae T. argyrophylla T. micrantha T. cavipes T. micropetala PF. ferruginea melinonii T. cenepensis T. grandistipulata T. formicarum T. odoratissima T. p T. chrysaloides T T. colombiana T. macropetala araensis chrysophylla T. macrostachya T. poeppigiana T. schultesiana T. fusca T. goeldiana T. T. pimichinensis T. prancel guianensis T. ptychophysca T. hypoleuca T. pubiflora F. richardiana P T. tinctoria T. inconspicua T. rigida T. vaupesiana T. leiocalyx T. venusta . vulgaris petals; the fourth group has previously been consid- ered part of Tachigali; and the fifth group corresponds with Tachigali sensu Aublet. None of the groups differs in more than one character from its neighboring group(s). If Sclerolobium and Tachigali were both to be recognized at the generic level, they would be separaled by a single character only. For instance, if Sclerolobium would be defined by linear petals, species in group | would be placed in Sclerolobium and those in groups 2, 3, 4, and 5 in Tachigali. However, if Tachigali would be defined by an asymmetrical hypanthium, species in groups 4 and 5 would form Tachigali and those in groups 1, 2, and 3 would be placed in Sclerolobium. Recognition of such single character genera is not recommended. There- fore, in this study, as in other recent treatments mentioned above, Sclerolobium is included in Tachi- galt. CHARACTERS USEFUL IN. IDENTIFICATION FLOWERS Flowers are almost indispensable for the identifi- cation of Tachigali specimens. The indument of the calyx is rather uniform in most species, but a few have glabrous calyx lobes. The shape of the hypanthium, symmetrical or asymmetrical, is of more importance. In about half the species the petals are linear; the other half have broad, more or less showy petals. The linear petals can be difficult to recognize because they closely resemble the filaments of the stamens, but they can usually be told apart by a slight difference in color. The linear petals are densely pubescent on the distal half in a number of species, and the pubescence is usually of yellow hairs, but in one species, 7. paraensis (Huber) Barneby, the hairs are white. There is also variation among the species with broad petals. In some species, the petals are erect, shorter than, o as long as the sepals and largely covered by them. while in others, the petals are spreading to reflexed and easily seen. This study did not find the amount of pubescence on the inner surface of the petals to be a useful character for discriminating species. The number of stamens is generally 10, with one exception: T. macrostachya Huber is described having 15 (16 to 19 this treatment, the number of stamens is approximate- stamens. In the flowers seen for ly 15, but because it is not uncommon for a few stamens to break off, the original number may well have been higher. Stamens may be equal or unequal in length: in the latter case, two or three stamens are shorter, thicker, and curved. The lower part of the filaments is nearly always densely pubescent. This study did not find the distribution or the density of Volume 95, Number 4 2008 van der Werff 621 Synopsis of Tachigali pubescence on the ovary to be of taxonomic significance; even in very young fruits the pubescence falls early. STIPULES Stipules have not been seen for 17 of the 54 species included in this treatment. In a number of species. they were difficult to find or rarely present. Three main types of stipules are found in Tachigali. In 23 species, stipules were pinnately divided, with one to three pairs of lobes; the lobes were flat and of varying width but never thread-like. This is the foliose stipule type (Figs. LA, B, 4D). In nine species, the stipules were pinnately divided but with thread-like segments. This is the pectinate stipule type (Fig. 1C. D). The remaining six species had entire stipules (rarely with a single pair of lobes); these stipules were strongly revolute and subeylindrical in cross section (Fig. 1E, F stipules were present in the inflorescences. Such occurred in flowering specimens of T. (Ducke) L. F. Gomes da Silva & H. that . In some cases, small bracts resembling revolute or bracts eriopetala rs Lima, but in sterile. specimens o species, revolute stipules at the base of the leaves were up E to 5 cm long. Because the foliose stipule type occurs in quite a few species, they are not very useful for identification. Revolute stipules, however, are quite useful; sterile specimens of 7. ertopetala, for instance, can be easily identified by the combination of stipule type and indument type on the lower surface of the leaflets. Within a species, there may be some variation in the size of stipule but not in the type of stipule. DOMATIA Domatia are swollen and hollow segments of the leaf stalk inhabited by ants; they are present in 26 species and absent in 29 species. Presence or absence — of domatia seems to be characteristic for any given species, with the exception of Tachigali paniculata, where domatia can be present or absent. Tachigali paniculata has been included in both categories in the count. Shape and position of domatia sometimes aid ig. 2A—C) or, when flat on the upper side of the leaf stalk, semiterete (Fig. 2D, E). In one species, T. bicornuta Werff, n T. physophora (Huber) Zarucchi & identification. Domatia can be terete (Fi van der the domatia form two blunt horns (Fig. 4C). Herend., — domatia are terete, often longitudinally and almost never more than 1 em from the the leaf stalk (Fig. 2A). In (Poepp.) Zaruechi € Herend., domatia are usually the leaf stalk ribbed, base of T. chrysophylla more than 4 cm from the base of (sometimes even 10 em) and can be up to 4 cm long (Fig. 2B). Table 1 lists species with domatia in italics and shows that the majority of species belonging to the traditional concept of Tachigali (with an asymmetrical hypanthium and spreading petals, and dimorphic or equal stamens) possess domatia. Tachigali longiflora Ducke and 7. macrostachya, two species listed in this group as lacking domatia, have hollow leaf stalks that are inhabited by ants, but in these species the leaf stalks are not swollen. Presence of domatia is much less common among species belonging to the traditional concept of Sclerolobium. LEAFLETS Leaflet characters such as size, number of lateral veins, pubescence, and shape of base and apex are of limited use in identification, and it is rarely possible to identify a specimen based only on leaflets. Characters that are uncommon in northern South American Tachigali such as leaflets with two to four pairs of secondary veins in T. ptychophysca Benth. and pi T. loretensis van der Werff, or a cordate base in T. ^ macrostachya, merely serve to confirm a determina- tion based on other characters. FRUITS - Some species of Tachigali are reported to be monocarpic (Foster, 1977), but it is not known how many display monocarpy. In known monocarpic species, fruiting is the last phase of life after the leaves have fallen, and it is not possible to connect fruiting specimens with flowers or leaves of other collections. Fruits on their own cannot be identified due to a lack of diagnostic characters. Occasionally, field observations indicate that a species is monocar- pic; for instance, O. Poncy (pers. comm.) collected richardiana Tul. im a flowering material of T. permanent plot and, on returning a vear later, Eod that all the trees had died. In the case of formicarum Harms, another monocarpic species, a collection consisting of fruits and a leaf rachis could only be identified because of the characteristic shape of the that (known from collections with fruits and leaves) are domatia. Species are not monocarpic listed in boldface in Table 1 and are fairly equally distributed among the species groups. The species of which fruits are not yet known might be monocarpic that T. catingae Ducke is and/or are possibly rare so fruits remain uncollected. For instance, known only from the flowering type collection and two additional sterile collections: therefore, in this case it would be premature to consider this species as monocarpic. 622 Annals of the Missouri Botanical Garden Figure |. Stipules. —A. ey e physophora (Huber 530, US). -h. T. cavipes (Pires 7920, LAN). — idi: ora 107178, US). —D. T. ferruginea (Neill 12932. MO). —E. | (Rickson B-44A-85, MO). T. chrysaloides (Dorr & Barnett 5 wd NY Volume 95, Number 4 van der Werff 623 2008 Synopsis of Tachigali Figure 2. Domatia. "i Tachigali physophora (Huber 530, US). —B. T. b A cd (Steyermark 107178, US). glauca (Prance et al. 24720, US). —D. T. cavipes ae 7920, | AN). » 5255, US). Scale mm at left applies to A-C: scale e at right applies to D. Annals of the Missouri Botanical Garden SCOPE or THIS PROJECT The treatment of Tachigali presented here was begun during preparation of an account for the Flora del Río Cenepa (Vásquez Martínez et al., in press). H soon became clear that identification of herbarium specimens was very difficult. Because large loans of specimens from an earlier attempt to study Tachigali were still at MO, it was decided to identify these pc as far as possible, write a key to the species, and combine this with brief notes i synoplic treatment. This account only deals with the South French species from northern America (Colombia. Am- Roraima, Venezuela, Guyana, Guiana, Suriname, azonian Brazil [states of Pará, Acre], in central and southern Brazil. Amapá, Amazonas. Ecuador. and Peru The species Bolivia, and Paraguay are, for the most part, taxonomically different from those in northern South America, and, because nol many collections from the southern area were at hand. this project was limited to species from northern South the 1300 included in treatment, : click on the link to the Excel spreadsheet. This America. Identifications of more than collections are nol this but are available at account is not a full monograph, and a number of problems remain to be studied in more detail. especially in the variable and widespread species T. paniculata and T. guianensis (Benth.) Zaruechi & | prenc TAXONOMIC TREATMENT Tachigali Aubl., Hist. PL Guiane 372. 1775. TYPE: Tachigali paniculata Aubl. Sclerolobium \ oge J, Li olobium denudatum Vogel. Innaea 395. 1837. TYPE: Seler- Small Leaves alternate, singly pinnate, paripinnate. leaflets | large trees, some species monocarpic. opposite; leaf stalks sometimes swollen and harboring stinging ants; stipules present or caducous, pectinate, pinnate or. entire, sometimes strongly revolute. Inflorescences terminal or lateral racemes or panicles: bracts lanceolate or subulate, often caducous. Flowers while: the asymmetrical; petals 5, linear yellow, rarely sepals Dy equal Or unequal, basally united, hypanthium symmetrical or broad, sometimes densely pubescent: stamens 10, rarely 15 to 16. equal or with 3 shorter, thicker and curved filaments usually densely pubescent on the basal half; ovary stipitate, attached in the center of the hypanthium or along the upper side, pubescent. Fruits strongly laterally com- pressed, dry, l- to 2-seeded. Distribution. A genus of 60 to 70 species from Costa Rica to southern Brazil and Paraguay. Key ro TACHIGALI SPECIES IN. NORTHERN SOUTH ÁMERICA la. Domatia in the leaf petioles with 2 blunt, horn-like projections, these ca. 1 em long 5. T, bicornuta Ib. Domatia absent or present; when present, cylin- drical and never with horn-like eee 2 Leaves with scattered, small (ca. 0.2 p stellate ?amelinonit 2b. / Leaves without scattered, stellate hairs: eae nl. If present. of simple hairs ....0.0.00.0.....000. 3 3a. Petals longer than sepals, more or less spreading, readily visible in open flowers; or, if petals about as long as sepals, then petals obvious, about as wide asthe sepals ooo... cece cee eee 33 3b. Petals equal to or shorter. than the sepals. inconspicuous, linear or nearly so, clearly much narrower than the sepals ooo... la. Leaf petiole or rachis with swollen. hollow, ant- inhabited domalia dis 5 Ib. Leaf petiole or rachis without ant-inhabited domala mein E TA EE O l oa. — Sepals glabrous except the ciliate margin: twigs and inflorescences fuscous-tomentulose/tomen- NOG sh et CLP 19. T. fusca ob. Sepals uniformly pubescent; indument of twigs and inflorescences variable, but never fuscous 6 Oa. — Lower surface of the leaflets completely covered Dye indümenl cia za euer x Goes Y EXER ES 7 Ob. Lower surface of the leaflets partially covered hy the indument or glabrous 2... sess. 11 Ta. Leaflets 15-30 em long, with 18 to 25 pairs ol secondary veins: base of leaflets on sym- metrical ooo Vi. T. ferruginea 7b. Leaflets to 20 cm long. with lo 15 pairs of secondary veins; base of leaflets obtuse lo acule, symmelrical or oblique Ba. Indument of twigs and inflorescences consisting T. macbridei 8b. Indument of twigs and inflorescences consisting predominantly of appressed hairs, sometimes a few erect hairs presentas well oo. 00.00.00... 000% Ya. Base of leaflets asymmetrical: stipules pectinate or foliose, Ob. Base of entire, strongly revolute flat 10 symmetrical: Md small, —Ó 2. T. vaupesiana Stipules pectinate: domatia more ibn 4 em from base of leaf, 2-4 em long: er trees of terra T. chrysophylla P. s foliose, flat: domatia ca. ; em from base f leaf, 1—2 10b. em lone: trees of flooded Tores or | physophora lla. Inflorescences and leaf rachises A or nearly soz base of leaflets asymmetrical: domatia usually longitudinally ridged ... 40. 7 physophora IIb. Inflorescences and leaf rachises densely puber- aflets smooth. not longitudinally ridged ulous: b: ise of symmetrical: domatia Volume 95, Number 4 2008 van der Werff 625 Synopsis of Tachigali 12a. 12b. 13a. 13b. la. l 14b. 15a. 15b. 16a. lob. 17 m | ib. QU o 18b. 19a. 19b. 20a. 20b. N X Leaflets to 4 em wide: lower surface of leaflets glabrous or very sparsely appressed js nt: . odoratissima w C trees of flooded forest ....... Leaflets 6-11 cm wide; lower surface 4 leaflets with scattered. erect. hairs; trees of non-flooded [ES ae cs Ires atea e da 10. T. Sepals glabrous, at least on the distal half cenepensis Sepals uniformly and usually densely pubescent ... l6 Lower surface of leaflets rather densely pubescent, 3 the hairs ferruginous, erect ...... 4. T. micrantha Lower surface of leaflets glabrous or nearly so ... 15 Leaflets oblong to oblong-obovate, the margin often recurved; sepals 5 mm long; petals about half as long as sepals, broad, glabrous D DA ea E 35. T. micropetala Leaflets elliptic to elliptic-ovate, the margin flat; sepals 2-2.5 mm long; petals about as long as the narrowly spatulate, pubes- sepals, filiform to CONO Ls nM Er ES T. leiocalyx Lower surface of leaflets moderate ly to dense ly pubese ent with whitish hairs, these see mingly radiating from a central jaink in various direc- HONS- ae aare arse ee dde ee 50. T. setifera s. str. Lower surface of leaflets without hairs radiating from a central point ..........ooooo.ooo cmo. li Lower surface of leaflets glabrous to moderately pubescent, the surface always partially or entirely Visible, sve etos rds dc oe lee aes a 18 Lower surface of leaflets completely covered by the indument, this varying from sericeous to tomentulose 2... 26 Indument on twigs, leaves, and inflorescences pilose, consisting of erect hairs, these sparse or moderately dense .... llle Indument, when present, appressed ......... 20 Base of leaflets asymmetrical: stipules pectinate, the segments flat, linear; petals ui gla- brous . T. guianensis Base of leaflets symmetrical: a l- to 3- foliose, the segments revolute; petals linear, distally densely yellow pubescent ....... T. setifera s.l. Lower surface of leaflets with dark brown scurf or minute scales (ca. 0.1 mm), this varying from dense to 50. T. setifera s.l. Lower surface of leaflets glabrous or variously sparse; individual hairs not visible pubescent, but never with dark brown scurf; if pubescent, individual hairs visible under a dissecting scope... sse nn 2] Petals pubescent; base of leaflets symmetrical or asymmetrical da S veg E ea y Petals glabrous: base of leaflets asymmetrical ... 25 Petals white pubescent... 2.0... 38. T. paraensis 3 Petals yellow pubescent 5... 2: Stipules caducous; small (to 5 mm), entre, stalked, strongly revolute bracts Mu present in. the inflorescences: base leaflets eufléate add de e 32. E melanocarpa Stipules foliose, sessile. often more than | em long and wide, persistent; entire, strongly. revo- lute bracts lacking; base of leaflets obtuse. .... 24 24a. 27b. 31b. 32b. Base of leaflets symmetrical or slightly asymmet- to I] pairs; young leaflets 25. T. Base of leaflets asymmetrical; secondary veins 5 to rical; secondary veins 8 with some appressed hairs ...... inconspicua T pairs; young leaflets silky pubescent on E lower surface T. amplifolia Flowers pedicellate 54. T. vulgaris Flowers sessile or nearly so ..... 5l. T. tinetoria Indument on young twigs and leaf rachis predominantly GIGGL 12.2% uos urs ee ca cee 27 Indument on young twigs and leaf rachis A hs bth e SR beoe ders eee Bah 28 3X longer than wide; petals pilose: floral bracts not longer Leaflets oblong to oblong-ovate, 2 to than mature buds 2.2... l.l. I4. T. prancet Leaflets broadly e lliptic to e epN ‘obovate, up to 2X as wide as long, then leaflets less than 10 em. long; petals glabrous: floral bracts much longer id. mature A e d peo eb ances | bracteosa TEILT 41. T. pimichinensis Tips of leaflets acute or acuminate 2.2.2.2... ¢ Pétals pilose- ge dpe hie eRe oS me SS 30 Petals pla Dious 22e tated does pe 32 Stipules flat, 3-foliose or entire: trees from A UAE M E 24. T. hypoleuca Stipules strongly revolute, entire: trees from terra Leaflets elliptic: secondary veins 6 to 10 on each side; tertiary venation immersed, not visible 6. T. eriopetala seaflets oblong: secondary veins 14 to 20 on each side; tertiary venation raised, visible on the lower surface of the leaflets T. chrysaloides Base of leaflets symmetrical; floral bracts twice as long as the floral buds; flowers sessile PS De eG oa a oy t Re 21. T. goeldiana Base of leaflets asymmetrical: floral bracts caducous, nol present at base of larger. buds: flowers pedicellate ............ 39. T. peruviana Trees from the Pacific lowlands in Colombia 13. T. rolormnrana Trees or shrubs east of the Andes Sepals with a fringe of hairs “1008 ii margin, otherwise glabrous : PINE Sepals uniformly and densely pubescent Indument on inflorescences and flower buds very dark brown: hypanthium in young fruits ca. 1 cm 1 27. T. longiflora Indument on inflorescences and flower buds light brown or grey-brown: hypanthium in young fruits less than 7 mm long. usually less than 5 mm long ..... 36 Secondary veins of each leaflet 2 to 3 pairs ...... 37 Secondary veins of each leaflet 5 or more pairs ... 38 Stamens dimorphic; petals 5-6 mm long: long stamens 10-12 mm long ..... 45. T. ptychophysca Stamens monomorphic: petals ca. 3 mm long: stamens ca. 5 mm long 28. T. loretensis 626 Annals of the Missouri Botanical Garden 39b. la. 10b. hla. Ith. 14a. lib. $ I5b. CO Lower surface of leaflets densely appressed pubescent tọ sericeous: sometimes erect hairs present as well 2.2. .02 0.2.0. eee as 30 Lower surface of leaflets sparsely pubescent or Oa e ETEA hee eRe Bye IE RUN UR d Domatia absent: hairs on inflorescences B leaf axes partly erect 16. T. pubiflora Domatia present; hairs on inflorescences am leaf axes appressed or erecb oo. 10 Hairs on inflorescences and leaf axes erect or ascending. s iex sy ex ee S H Hairs on. inflorescences and leaf axes strictly ee Geena ane T MEA E 12 Stamens monomorphic. 4-6 mm long: stipules 2— 5 em long, to 3 em wide. foliose. entire or with one basal lobe ooo... 22. T. grandistipulata Stamens dimorphic, the longer ones 1-2 em lone: stipules often lacking. when present up to 2 X A IN 18. T. rigida Stamens monomorphic; petals ca. d mm long: domatia terete in cross secon... ee 3 Stamens dimorphic: petals 5-7 mm long: domatia semiterete in cross section Inner surface of the petals pubescent for at least 4/5 of their length; only known from the lower Rio STE Inner surface of the petals with a few hairs near Negro, Parana de Anavilhana argyrophylla few collections from \mapá, Roraima, and Guyana oo... 20. T. the base: mostly from Pará. glauca Flowers sessile or nearly so. a distinct pedicel lack- mg: stipules with filiform segments ..... 53. T. venusta stipules with filiform segments Flowers with a distinct pedicel, lmm long; segments of stipules narrowly broadly ovate, widest in the middle ooo E Hypanthium distinctly asymmetrical: along Rio Vaupes and tributaries in Brazil and Colombia and around Iquitos, Peru ooo... 9 cavipes Hypanthium slightly asymmetrical: lower Rio egro. downstream from Rio Jauaperi T. plumbea [flowers ca. Diameter o Dmm; stamens mono- MOP lea aa AT Diameter of flowers 10 mm or more (rarely only 8 mm): stamens dimorphic ............08% 53 Leaves with 10 or more pairs of leaflets o... 48 Leaves with up to 6 pairs of leaflets ooo... 50 Inner surface of the petals densely pubescent with long hairs, the hairs extending beyond the margins of the petals: inflorescences ca. 30 em. with longitudinal ridges due to the decurrent bases of the flowers T. candelabrum r surface of the petals pubescent. but hairs not extending beyond the margins: inflorescences terete, rarely exceeding 20 em. not longitudinally PURE? uS neces, Lalas ados Ueda dae eae 19 Hairs on calyx appressed, grey-brown: py ena um asymmetrical o... 15. T. poeppigiana Hairs on calyx minute, ascending lo erect medium brown: hypanthium symmetrical or arly so ooo 19. T. schultesiana Ma. 50b. 5la. 51b. 53a. |. Tachigali alba Ducke, Discussion. Domalia semiterete; petals densely yellow pu- bescent on the inner surface, the hairs long and extending beyond the margin of the petals To E E e e ae 18. T. Jormicarum Domatia cylindrical lacking: petals sparsely pubescent on the inner surface. the hairs not extending beyond the margin of the petals 5] Domatia lacking: inflorescences and rae ars densely reddish brown pubescent... 14. T. davidsei Domatia present: inflorescences and flowers densely brown puberulous or moderately to sparsely pubescent Inflorescences and flower buds densely puber- ulous and completely covered by indument: inner surface of the petals pubescent, the hairs covering 2 about 3/4 of the surface 3. T. loretensis Inflorescences and flower buds moderately to sparsely pubescent. the surface always partially visible: inner surface of petals with a few hairs near the base, otherwise glabrous... . 15. T. dwyert f leaflets cordate; margin of leaflets rev- ¡ue viis: i aflets bullate: stamens 15 or more a EREE E EEEE Pee eee 31. T. Base of leaflets obtuse to cuneate; margin of leaflets macrostachya not revolute; leaflets plane: stamens 10 ooo... 54 Leaves with 3 (rarely 4) pairs of leaflets: leaflets twice as long as wide: domatia present i.i... 8. T. catingae Leaves with more than 5 pairs of leaflets: leaflets more than twice as long as wide; domatia present Leal rachis sharply triangular: ae a | prese nt or nbSenb eh icu bork halk 4 d. paniculata Leal rachis terete or nearly so: mn absent ... 56 Stamens monomorphic: secondary veins 8 to 10. often slightly impressed on the upper surface Stamens dimorphic; secondary veins 4 to 8, immersed on the upper surface Stipules caducous: lower surface of leaflets sparsely to moderately appressed Pr ent, nol O E richardiana Stipules persistent, foliose; lower MM of leaflets glabrous, densely gland-dotted 4. T. barnebyi Arch. Jard. Bot. Rio de Janeiro 3: 92. 1922, Tachigali paniculata var. alba (Ducke) Dwyer, Ann. Missouri Bot. Gard. Il: 240. 1954. TYPE: Brazil. Pará: "prope Gurupa.” Ducke s.n. (MG 17227) (lectotype, designated here, MG not seen, photo Fl: duplicates. BM! G not seen, photo FH P!). Tachigali alba is a tall (30-40 m) tree from terra firme forests. Domatia are lacking, stamens are monomorphic, stipules are caducous and have nol been seen. Leaflets have an asymmetrical base, are generally oblong with slightly impressed lateral veins on the upper surface. nearly 5060. The leaf Flowers are smaller than rachis is terete or n T. paniculata, with petals 4—1.5 mm and sepals 4 mm. The ultimate Volume 95, Number 4 2008 van der Werff 627 Synopsis of Tachigali divisions of the inflorescences are rather. loosely flowered and linear, while in the similar 7. paniculata the flowers are more compactly clustered and the flower-bearing part is narrowly pyramidal, clearly wider at the base than at the apex. Tachigali alba is 'Tences morphologically close to T. richardiana: diffe are the dimorphic stamens in T. richardiana, fewer pairs of lateral veins (8 to 10 in T. alba, 4 to 7 in T. richardiana), and the presence of a patch of hairs near the base on the outer surface of the petals, while the outer surface of the petals is glabrous in T. alba. To prevent nomenclatural ambiguity, the collection Ducke s.n. (MG 17227) in MG has been designated as lectotype. The other syntypes are Ducke s.n. (MG 17075) and Ducke s.n. (MG 17110); duplicates of these collections have been seen in BM and represent Tachigali alba. Distribution. Tachigali alba is known from Amapá, Amazonas, and Pará in Brazil and a single collection from Guyana. Selected specimens examined. BRAZIL. Amapá: Pires el al. 51242 (F. GH, TAN, MG. US). Amazonas: Manaos. Ducke 932 (F. MG. R. RB. US). Pará: Rio Pracupi, Froes 32850 (IAN, US). GUYANA. al Greal Clarke 6565 (K). Rewa River. Falls, 2. Tachigali d (Ducke) Barneby, Brittonia 48: 182. . Basionym: Sclerolobium amplifolium Ducke, Arq. pus m Veg. 2: 43. 1935. TYPE: Sao pus de Olivença, . Ducke s.n. (RB. 24295) K!, P!, US!). Brazil. Amazonas: * px silva non inundabili." (holotype, RB not seen: ua Discussion. According to the original description, this species is recognized by its large leaflets that are densely pubescent; linear petals: large and persistent, foliose stipules; and lack of domatia. The description from Ducke (1935) also mentioned that the leaflets were silky pubescent and golden-shiny. but these characteristics are restricted to young leaflets. Petals are pubescent on the upper half. Leaflets have an asymmetrical base. Four collections have been seen and == from western Brazil (Acre, Amazonas. Rondónia several from Guyana, French Guiana, and Suriname. The collections from the (to 25 X 11 em) and longer floral bracts than the When recorded on Guianas have larger leaflets collections from western Brazil. field labels. habitat is given as forest on terra firme. Good flowering material of Tachigali amplifolia has seen from Guyana, French Guiana, or Suriname in this study, and it is not certain if the specimens from the Guianas should be included in T. amplifolia. The flowers of the specimens from western Brazil have relatively long sepals in relation to the hypanthium, and the sepals are ae narrow and do not overlap. Selected Lima, specimens BRAZIL. Acre: Mun. Cid Ferreira al. 10026 (INP/ Y). Sao Paulo de a Krukoff 8879 (F, MO, km E of Rondónia en route to 50799 (C c F, NY, US). FRENCH Sabatier 1546 (MO. NY). SUR- Irwin et al. 55597 (GH. examined. > lancio Amazonas: NY. US). Vilhena. Maguire et al. GUIANA. Haut Camopt. INAME. Wilhelmina NY) Rondónia: Geberal le. 3. Tachigali argyrophylla Ducke, Bol. Teen. Inst. Agron. N. 2: 14. 1944. TYPE: Brazil. Parana de Anavilhana, Ducke 936 eh RB INPA!, K!, MO!, R!, Amazonas: not seen: isolypes, s the Discussion. Characteristic of the species combination of the sericeous lower surface of the leaflets. terete domatia, and rather small, broad petals (4 mm long) with a pubescent inner surface. The base of the leaflets is asymmetrical. Stamens are monomorphic. Stipules are unknown. This species is he currently known from the type collection from lower Rio Negro in terra firme forest and possibly a second collection from Loreto. Peru. It is surprising that more collections of this species do not exist, because the lower Rio Negro is relatively well collected. A collection from the Jenaro Herrera in Peru (Spichiger 4686) is tentatively placed here; i keys to Tachigali argyrophylla based on flower size, monomorphic stamens, indument on the inner surface of petals, terete domatia, and sericeous lower surface of leaflets. This is an unusual disjunct distribution. Dwyer (1954) listed Ducke 936 and 937 since Ducke only mentioned Ducke as syntypes. This is an error, 936 when he described this species. 4. Tachigali barnebyi van der Werff, sp. nov. TYPE: Brazil. Rondónia: basin of Rio Madeira, margin of Mutumparaná airstrip. 25 Nov. 1968, . T. Prance, W. A. Rodrigues, J. F. Ramos & L. G Farias 8844 (holotype, INPA!: isotypes, COLL! FL GH!, MG! R). Figure 3. Tachigali richardianae Tul. similis, sed ea stipulis ab persistentibus, foliolis subtus glabris punctatisque et petalis extus glabris recedit. o 25 m; twigs terete, minutely puberulous. Tree, glabrescent; stipules foliose, gland-dotted on the lower terminal lobe (1.5-2.5 X l- small (4-5 x 2 short (1-2 mm) stalk. surface, with a large 1.5 em) and | or 2 2—3 mm) lateral lobes, the lobes with a Leaf rachis 7-13 cm, terete, minutely puberulous, domatia lacking; M 4 to 5 pairs, oblong to ovate-oblong, 7-10 X 2-3 cm. glabrous, gland-dotted on the lower surface, " base obtuse, symmetrical or nearly so, the apex acute: secondary veins 7 to 8 pairs per leaflet, 10— buds immersed and poorly visible. Inflorescences 20 em, densely brown puberulous. Flower Annals of the Missouri Botanical Garden Figure 3. Tachigali barnebyi (isotype). copyright reserved Testes. bec ey vas dee Me Pf Dete: " Henl /, Missouri Botanical Gatden jou 3 Ths Ly ps tt COL, ONT Volume 95, Number 4 2008 van der Werff Synopsis of Tachigali 629 densely pubescent; bracteoles 4-5 mm, 1/2 to 1/3 the length of mature buds; pedicels of open flowers ca. 5 mm, hypanthium oblique, sepals 5-7 X ca. 4 mm, — puberulous on 1 inner surface, spreading to reflexed: petals similar in size and shape to the sepals, glabrous on the outer surface and pubescent on the inner surface, spreading to reflexed; stamens 10, dimorphic, 7 longer stamens 1.2 em, 3 lower E ee; shorter, ca. 8 mm, all pubescent on the 4 mm; pistil ca. 1.2 em, ovary ca. 5 mm, — densely pubescent, the style ca. 7 mm, becoming sparsely pubescent toward the stigma; stigma entire o with 2 short lobes. Immature fruits ca. 1 em, densely pubescent; mature fruits not seen. with 7. domatia, The petioles are especially rare among the species with The differs from T. richardiana in its persistent stipules Discussion. Tachigali barnebyi shares richardiana terete petioles, lack of and arge flowers with dimorphic stamens. terete large (4-7 X ca. 4mm) petals. new species (caducous in T. richardiana) and its glabrous, gland- dotted moderately appressed pubescent and without gland dots in T. lower surface of the leaflets (sparsely to richardiana). Additional but weaker differences are the following: floral pedicels are 3 mm); the base of the 48) longer in T. barnebyi (5 vs. leaflets is more asymmetrical richardiana; and the leaflets of T. barnebyi are more coriaceous and the The weaker differences are best seen when specimens of venation is less visible than in 7. richardiana. both species are at hand. The distribution of the two species does not overlap: T. barnebyi is known only richardiana from French Brazil. also has terete petioles, has showy petals, and lacks from Rondónia and 7. Guiana and adjacent Para Tachigali alba domatia. However, it has caducous stipules, smaller (4—4.5 mm), slightly impressed secondary veins. According to the petals monomorphic stamens, and label of the type collection, isotypes have also been deposited in NY, K, S, U. and P. Etymology. It is a pleasure to name this species after the late Rupert Barneby (1911-2000), who was truly a scholar and a gentleman and contributed so much to our understanding of Neotropical Leguminosae. IUCN Red List category. known from three collections in a poorly known part of Tachigali barnebyi is the Amazonian rainforest and is therefore listed Data Deficient (DD) (IUCN. 2001). RO- | Petersen, B. W. Nelson, Mo a A A MO. US); Porto Maciel & S. Rosario 1722 PE Rodovia a Paratypes. BRAZIL. 13 km de Vilhena, i G: J. F. Ramos & C. D. Velho, hio (MG) Jamari, 1e outer surface, pubescent on the formicarum: its 5. Tachigali bicornuta van der Werff, sp. nov. TYPE: Peru. Amazonas: Bagua Province, distr. Imaza, comunidad Aguaruna de Wanas, cerros Chinim, 800-850 m, 31 Aug. 1996, C. Diaz, A. Pena, L. Tsamajain & M. Roca 8100 (holotype, HUT not seen: isotype, MO!). Figure 4. ongeneris domaliis petiolaribus bicornutis et ramulis le rmin: nalibus infruc tesce niiarum inflatis rece adit. 12 m; | weaves ca. Trees, twigs terete or slightly angular, elabrous. 50 em, with 6 to 10 pairs of leaflets, rachis + round, minutely puberulous, the hairs ca. 0.1 mm, erect, not visible with the unaided eye; a domatium present ca. 5 cm from the base of the leaf. this shaped as 2 blunt horns ca. 1 cm and with a round tip, a pair of small (ca. 6 X 3 cm) leaflets 2 horns of the 12-20 somewhat asymmetrical, the attached to the sinus between the 2 domatium: leaflets narrowly elliptic, 4.5 em, the base obtuse, apex lapering to a sharp point, the lamina gland- dotted on the lower surface, glabrous, midrib and secondary veins minutely puberulous as on the rachis, secondary veins 8 to 11 pairs per leaflet, petiolules 3— 5 mm: detached stipules ca. 3 cm long. with 3 pairs of lanceolate segments, these 1.5-2 cm X 2—5 mm. Flowers and inflorescences unknown. Detached infructescences ca. 25 cm, densely and minutely puberulous, 1- or 2-branched, base of infructescence solid. 3-6 mm diam., ultimate branches of infructes- cences 8-18 em, distally swollen and hollow, ca. | em diam., with | or 2 perforations near the base: calyx cup + symmetrical, not strongly oblique; fruits strongly laterally compressed, ca. 5 X 2 cm, the margin papyraceous, I-seeded. Discussion. The new species Tachigali bicornuta can be easily recognized by the peculiar domatia on the leaf petioles, the finger-like inflated terminal branches of the infructescences, and the minute, erect indument on leaf rachis, the main veins of the and the infructe: leaflets, cences. According to the label information of the type collection, the petiolar domatia are occupied by fierce ants. A collection from Brazil, Acre, Serra do Moa (Cid Ferreira et al. 5105, INPA, MEL, MO, terminal branches. of the inflorescences, NY) has similarly inflated which are 1-1.5 em diam. This collection differs in several characters from the new species described here and is not placed in T. bicornuta. Ms domatia are much 5 X 1.5 em) and are flattened T. elabrous and lack the longer than wide (ca. on the upper surface of the rachis, much as 1 leaflets are minute puberulous indument on the main veins; the 3 X 10 em): and the indument on the inflorescences is appressed, not leaflets are larger and wider (up to 28 630 Annals of the Missouri Botanical Garden Figure 4. Tachigali bicornuta (Diaz, Peña, Tsamajain & Roca 8100. MO). —A. Leaf. —B. Infructescence. —C. Domatia. dorsal view (top) and lateral view (bottom). —D. Stipule. —E. Base of leaflet. Volume 95, Number 4 van der Werff 631 2008 Synopsis of Tachigali puberulous or minutely erect. The hypanthium at the sheets; isotypes, MEL', MG!, MO! NY!) base of the mature fruits is symmetrical. Apart from Figures 5 and 6. the thickened terminal branches of the inflo- Tachigali poeppigianae Tul. similis; sed ab ea lobis rescence, the collection agrees quite well with 7. formicarum. IUCN Red List category. been collected once in a botanically poorly Tachigali bicornuta has known region of Peru and is listed as Data Deficient (DD) (IUCN, 2001). 6. Tachigali bracteosa (Harms) Zarucchi & Pipoly. Sida 16: 411. 1995. Basionym: Sclerolobium bracteosum Harms, Bot. Jahrb. Syst. 40: 169. 1907. TYPE: Brazil. Amazonas: Cachoeiras des Marmellos, Ule 6094 (holotype, BT; isotypes, K!, Li). u o froesti Pires, Bol. Tecn. Inst. Agron. N. 38: 23 300. Tachigali nee (Pires) L. F. Gomes da Silva & 399. 2007 5 C. Lima, Rodriguésia 58: 399, 2007. Syn. nov. TYPE: Brazil. Amazo Rio Canuma. Froes 33744 RB not seen). as: (holotype, TAN not seen: isotype. Discussion. Tachigali bracteosa has broadly elliptic to obovate leaflets, tomentulose on the lower as surface, long bracts subtending the buds, and glabrous, linear petals. Domatia are lacking. Stipules of this species have not been seen in this study. The base of the leaflets is rounded to cordate. Only a photograph of the type of Sclerolobium froesii (in Pires, 1960) has been seen. The photograph and description leave no doubt that 5. froesii is a synonym of T. bracteosa. Dwyer 1957a) cited Macedo 4063 (MO, US) from the state of Pará under 5. bracteosum; however, these specimens M lack the tomentulose indument on the lower surface of the leaflets. In shape and size, the leaflets agree with 7. bracteosa. It is possible that these fruiting specimens represent an undescribed species closely related to T. The US Macedo 4063 pectinate stipules. bracteosa. specimen of has been Brazil (Amazonas, Mato Grosso, Tocantins) and Bolivia. Distribution. Tachigali bracteosa has reported from flooded forest or riverbanks Selected specimens examined. BOLIVIA. Beni: Prov. Yacuma, Beck € de Michel 20919 ex Santa Cruz: Parque Noel. Kampfi. Quevedo et al. 2585 (NY). BRAZIL. Axinim. Zarucchi et al. 2922 (INPA, MG Ratter et 1. Lagoa da Confusao, N of Rio Suia-Missu Ferry, 'antins: Mu Mendorea el N 3962 (MO). 7. Tachigali candelabrum van der Werff, sp. nov. TYPE: Brazil. Mun. de Rod. Transamazonica Km 302, Rio Aripuena Amazonas: Manicore, entre a Rod. e a Cachoeira Matamata, 24 Apr. 1985. Cid Ferreira 5780 (holotype, INPA! [2 stipularum linearibus, petalis intus dense longeque pu- bescentibus, inflorescentiis xa racemis porcatis et = c domatiis semiteretibus recec 12 m; smooth, glabrous; stipules 1-1.5 Trees, to twigs terete or slightly angled, 5 cm, pinnate, with 3 pairs of linear segments, these ca. l mm wide, glabrous or with a few scattered hairs. Leaves 25— 40 em, with 10 to 15 pairs or leaflets, the leaf axis sharply angled, glabrous or with some appressed 1airs. leaflets in middle of the leaves 10-15 X 3- 4 em, often conduplicate, basal leaflets smaller, ca. 5 X 2.5 cm, leaflets glabrous, the base strongly asymmetrical, the larger lobe auriculate, the tip acute; secondary veins 6 to 8 pairs on each leaflet. inconspicuous, tertiary venation raised on the lower surface, immersed on the upper surface; domatia 2— 3 cm, semiterete, not strongly thickened, distally with a pair of scars from fallen leaflets. Inflores- 30—40 cm, pubescent near the base, cences paniculate, sparsely appressed the terminal racemes 20— 25 em, densely appressed pubescent and with — pronounced ridges from the decurrent bases of the flowers; bracteoles ca. 5 mm, densely pubescent, linear; flowers densely and minutely ap- pue 3 mm, ovate-elliptic, densely pubescent on the outer yellow. — pressed pubescent; pedicels 1-2 mm; sepals surface, sparsely pubescent on the inner surface: petals slightly longer and narrower than the sepals, glabrous on the outer surface and pubescent on the L^ inner surface, sometimes densely so toward the tip; stamens ca. 10, monomorphic, united at the base, 4— 5 mm, pubescent on the lower half, ovary ca. 3 mm, the l mm: hypanthium asymmetrical, but ovary inserted densely and uniformly pubescent, at the base of the cup, with the stipe appressed against the lower side. Fruits not known. Discussion. Tachigali candelabrum can be recognized by the leaves with many pairs of leaflets: the sharply angled leaf rachises; the dense pubescence on the inner surface of the broad petals, especially on the distal half; and the many ridges on the terminal racemes formed by the decurrent bases of the flowers. The long, upward-arching racemes give candelabrum, and the chosen jeb] the impression of ¢ ec epithet "candelabrum" is a noun in apposition. Tachigali poeppigiana and T. schultesiana are the only other species of Tachigali with 10 or more pairs of leaflets. They differ in having terete domatia and foliose stipules and stamens, which are only pubescent near their base. Tachigali candelabrum is only known from three collections in southern Annals of the Missouri Botanical Garden MISSOURI BOTANICAL GARDEN 10 8 po ve] T 9 copyright reserved 6 5 wW Werff 4 van der db Ly tee PVG Se SEN Hen! Vor Wert? r Werff, 200 PROJETO FLORA AMAZÓNICA de Excursao a Rodovia Transamazonica - 1985 INPA NO 127.244 CAES. chigalia do Transamazonica km 302 a Rod. e a Cachoeira 4O'W. Mata de igapo, solo arg. umido. € [e] tr 60 Árvore de 12 b« I m x l0cm jes esverdeados. N.V: de diam., flores amarelas, i "Taxi. , C.A. Cid Ferreira N9 5780 24/14/85 CONVENIO CNPQ/NSF (INPA MEMBROS DA EXCURSÃO: C.a. CID FIJAEIBA, O.A HENDERSON, F. RICKSON, D. PASEA, J.L. DOS SANTOS b J.A Figure 5. Tachigali candelabrum (holotype, lea). Volume 95, Number 4 van der Werff 633 Synopsis of Tachigali MM STE TERN VA a DM : Pd N A (= "ML BARIO EN " MISSOURI BOTANICAL GARDEN 2] [c2] [2] A copyright reserved Ed hi iy cor ee 3 r v g an der Werff vale kipe : PA Determined by: Henk van der Werff, 2007 Figure 6. Tachigali candelabrum (holotype, inflorescence). 634 Annals of the Missouri Botanical Garden Amazonas, east of Humaitá. One collection came from erra firme forest, one from igapo. and the third had no habitat information on the collection label. The holotype consists of two sheets, one with a leaf, the other with an inflorescence. Both sheets have the INPA specimen. same herbarium number and form a single IUCN Red List category. is only known from three collections. Tachigali candelabrum Without further Data Deficient (DD) information, it is best listed (IUCN, 2001). BRAZIL. Rod. d “da * 13 5487 (INPA, MEL, Transamazonica i " ra Amazonas: Mun. de Humailá. m de Humaitá., fl.. Cid f NY): 300 km E of Humaitá pus Cid Ferreira 5925 (MO, Paratypes. Ferreira 8. Tachigali catingae Ducke, Arq. Inst. Biol. Veg. hs 12. 1938. TYPE: Brazil. Amazonas: Ducke s.n. ( RB 55421) (holotype, RB not seen: isolypes, INN P!, USD. Discussion. Tachigali catingae can be by the recognized combination of glabrous leaves and twigs. while the inflorescences are rufous-brown tomentose 2 or tomentulose: its leaflets are 3 to 4 pairs and 10 em present, semiterete, and stipules are foliose with wide, stamens are dimorphic. Domatia are pair of lateral lobes: the terminal lobe is ca. 2 em long. Petals are elliptic to obovate. ca. 2 mm wide. The distal leaflets have a symmetrical, obtuse base. This species has relatively broad leaflets. The lower surface of the leaflets is densely gland-dotted. a character otherwise only seen T. barnebyi and | bicornuta. Distribution. Tachigali catingae is known only from the type collection from the upper Rio Curicuriary and two. sterile collections from the upper Rio Negro and the Rio Uaupes. As the name indicates, it occurs in caatinga forest. \dditional specimens examined. BRAZIL. Amazonas: Taracua, Rio Uaupes, Pires et al. 7512 (LAN) Upper Rio Negro. Varacua. Rodrigues 1144 ANPA). MacBr.. Field Hist. Bot. Ser. 13: 127. 1943. Benth.. FL Brazil. 9. Tachigali cavipes (Benth.) J. F. Mus. Nat. Tachigali paniculata var. cavipes 15: 229. 1870. TYPE: Panure ad Rio Uaupes.” Bras. “prope Spruce 2555 (holotype. Kl; isotype, P!). This Tachigali plumbea Ducke: Discussion. species is very similar lo the only difference is the "s distinctly asymmetrical hypanthium In eavtpes (d plumbea. The distributions of the two species do not versus the slightly asymmetrical hypanthium i overlap: T. caripes is known from the Rio Vaupes and tributaries in Brazil and Colombia, one collection from o Guinía a the R Venezuela and a few collections around Iquitos in Peru, while 7. plumbea occurs downstream along the Rio Negro and Rio Amazon. Petals are to 3 mm wide. Both species have semiterete domatia, sericeous leaflets, foliose stipules, and dimorphic stamens. Tachigali glauca Tul. and T. y venusta Dwyer are also similar; T. glauca differs in its more erect or ascending indument on inflorescences and flowers and its monomorphic stamens. and 7. venusta. differs in its pectinate stipules and sessile flowers. BRAZIL. Amazonas: Rio affluent of Rio Uaupes, Kawasaki 225 (INPA, MO). Selected specimens examined. Tiquié, ae Rio Miritiparana, Schultes & 10403 (GH. Vaupes: Rio d Davis Werff & Hio Guai- . ul iln 40: Jenaro Herrera, van der Vasquez 9989 (MO). VENEZUELA, Amazonas: nía, ee et al. 11709 (F. GH, IAN, MO, US 10. Tachigali cenepensis van der Werff, sp. nov. TYPE: Peru. Amazonas: “Quebrada chichijam entsa," 1100 ft., 24 Apr. 1973, Ernesto Ancuash 291 (holotype, MOL isotype, US!). Figure 7. Quoad domatia longa Tachigali chrysophyllae ( Zarucehi & | inflorescentiarumque erecto et basi foliolorum symmetrica differt. oe pp.) lerend. similis. sed ab ea indumento is 20-25 m; thick, densely pubescent with short, erect, straight, ferrugi- 3-lobed, the em long and wide. Tree, twigs angular, 1—1.5 em z o nous hairs: stipules foliose, flat. 2- or lobe the terminal largest, |— Leal rachis 30-40 em, densely ferruginous puberu- lous, with hairs similar to those on the Iwigs. domatia 8-10 em. 10 em from the base of the rachis and above the first present, evlindrical, 7-9 mm diam., 8- pair of leaflets: leaflets ca. 5 pairs, the basal pair ca. 10 X 6 om, the distal pairs larger, to 19 X 11 em, the ase symmetrical or nearly so, obtuse to rounded or subcordate, apices not seen (damaged on the type). sparsely pubescent with short, erect hairs on the lower surface, sometimes mixed with some longer, appressed hairs. glabrous on the upper surface; secondary veins 9 lo 12 per leaflet pairs immersed on the upper surface, raised and pubescent on the lower surface. Inflorescences paniculate, 20-30 em, densely ferru- ginous puberulous, the main axes with evlindrical leaf branches of the domatia similar to those of the rachises, bul ultimate Mature mm long, the hypanthi- ca. 2 X 3 mm, sepals 22.5 X ca. 1.5 mm, petals as long as the sepals, shorter (5-8 em long): inflorescences not swollen. flowers densely pubescent. the pedicels ca. ] um cup-shaped, symmetrical. line "ur. distally densely yellow pubescent; stamens 4— o mm, pubescent near the base, otherwise glabrous: Volume 95, Number 4 van der Werff 2008 Synopsis of Tachigali 635 MISSOURI BOTANICAL GARDEN HERBARIUM No 2812812 MISSOURI BOTANICAL GARDEN o T È o a” 2 = = > E 2 2 o 0 J by Henk van der Werff JIT c erm Missouri Botanical Gatden 1 PERU Department of Amazonas ,EGUMINO Ernesto Ancuash 29] Date: 24 April 1973 MISSOURI BOTANICAL GARDEN HERBARIUM Figure 7. Tachigali cenepensis (holoty pe). 636 Annals of the Missouri Botanical Garden ) young fruits ca. 8 mm, pubescent with spreading, dark brown hairs: floral bracts not seen. Mature fruits nol seen. Discussion. Tachigali cenepensis is a very distinel species based on the combination of long, cylindrical domatiaz; the dense, ferruginous indument consisting of short, erect. and straight hairs on twigs and inflorescences; and the sparse, erect. indument on the lower surface of the leaflets. The large leaflets with an obtuse, rounded, or subcordate base are also distinet. The species keys out together with T. odoratissima (Benth.) Zaruechr € Herend. and 7. physophora because of the presence of domatia, uniformly pubescent sepals, and a sparse indument on the leaflets. However, these two species are nol closely related to T. cenepensis. They differ in having smaller, narrower leaflets, smaller domatia, and different indument; both are species from flooded forests in Venezuela, Brazil, and Colombia. Rather long. cylindrical domatia occur in T. chrysophylla and Zarucchi & species have glabrous petals. It is more likely that the T. macbridei Herend.. but these two relationship of T. cenepensis rests with a group of (Ducke) Zaruechi & T. chrysaloides van der Werff, including T. setifera mm species Herend., vásquezii, . F. Gomes da and T; prancet (H.S . Irwin & Arroyo Silva & H. C. leaflets with a symmetrical base, pubescent petals, Lima. These species share oblong and, when present, large foliose or revolute stipules. Tachigali cenepensis is the only species in this eroup 8 y 5] grou| with domatia. Among the sterile or fruiting collections at MO are several collections from lowland Peru that closely resemble T. cenepensis in size and shape of leaflets and the dense, erect indument on the twigs and leaf stalks as well as one flowering. collection resembling 7. cenepensis in leaflet characters, bul with appressed indument on the lower surface of the leaflets. Other collections from the the valley of the same area, Río Paleazu in Pasco. have short. erect hairs on the lower surface of the leaflets. Domatia are absent in all these specimens. H is possible that the domatia in the type of T. mens specimens cenepensis are atypical and that the speci- without domatia, but with similar leaflets, belong to T. the lype of T. differ cenepensis. The specimens similar to cenepensis but without domatia from F. setifera s.l. in their larger and wider leaflets and leaves, thicker leaf stalks. and the dense, ferruginous indument on twigs and leaf stalks. Additional flowering specimens are necessary in order to resolve relationships in this group. The specimens resembling 7. cenepensis are included the exsiecatae list as 7. aff. cenepensis: they had been previously identified as T. vasquezii, Sclerolobium bracteosum, and rugosum Benth. and may have been distributed under those names. IUCN Red List category. This species is only known from the type collection. For lack of additional information, it is listed as Data Deficient (DD) IUCN, 2001) — — . Tachigali chrysaloides van der Werff. sp. nov. TYPE: Ecuador. Morona-Santiago: Cordillera de Cutucu, along Patuca-Santiago rd., 600-1000 m, 23 Oct. 1988, L. Dorr & L. Barnett 5824 (holotype, NY: isotypes, MEL!, MO!). Figure 8. ;omes da . Irwin & Arroyo) L. F. ( P ira ups (I. Silva & . Lima si s sed ab ea stipulis. magnis revo d el a nto breviore adpresso rece m a T. setifera indumento sericeo et stipulis magnis diversa e Trees, to 30 m; twigs angular, minutely appressed pubescent; stipules entire or with one basal lobe, the main lobe or entire stipule 2—4 em, strongly revolute, the lateral lobe when present 1—1.5 em. also strongly revolute, appressed pubescent on the outer surface, densely so on the inner surface; the secondary veins strongly impressed and the whole structure appearing rachis 20-40 em, canaliculate on the upper surface. densely — like a pupa of a butterfly. Leaf lerete or and minutely appressed pubescent, domatia lacking, I leaflets 5 to 7 pairs, 11-20 5-8 em, oblong to oblong-elliptic, the base symmetrical, rounded to subcordate, the apex acute or short-acuminate, the upper surface glabrous except for appressed pubes- cence on the primary and secondary veins, lower surface densely yellow-brown pubescent, ment appressed and completely covering the surface, somewhat sericeous; secondary veins 14 to 20 pairs per leaflet, immersed on the upper surface, raised on the lower surface; tertiary veins scalariform, weakly raised on the lower surface. Inflorescences paniculate, 20-30 cm, the main axis densely appressed pubes- cent, the minor axes also with longer, curly hairs; € 0.5—1 mm, floral bracts subulate, ca. 2 mm, densely pubescent, only present at base of young buds and quickly caducous; flowers 4-5 mm, uniformly appressed pubescent, the an ca. | mm long turbinate, symmetrical, sepals ca. 3 mm, inner surface Fi petals linear, ca. 2 mm, densely yellow pubescent; stamens ca. 4 mm, pubes- half, 3 mm, pubescent, the hairs stiff, brown, ascending. cent on the lower distally glabrous; ovary ca. Fruit samaroid, X 3.5 em, I-seeded. Discussion. Tachigali chrysaloides is closely related to T. prancei and T. setifera. |t differs from both species by the large, revolute stipules. Stipules are generally absent in T. prancei; the one stipule seen Volume 95, Number 4 van der Werff Synopsis of Tachigali Holo Mi | b r Werff, 2007 Figure 8. Tachigali chrysaloides (holotype). THE NEW YORK 0) "nsn PLANTS OF ECUA 9 10 copyright reserved 8 7 © MISSOURI ECUADOR. Prov. Morona Santiago. de Cutucú. Patuca (c 2°40'S378°1 road, E of the "us cts Km 2 Elev a. 600-1000 m. Tree, 25-30 m tall. Flowers yellow. L. J. Dorr & C. Barnett ) 2 -san ntíago October 1988 Annals of the Missouri Botanical Garden al. In T. setifera, st are also frequently absent; when present, they are to in this study is foliose and f ipules m The cm long and revolute, as in chrysaloides. indument on twigs and leaf rachises of T. prancei consists of short, erect hairs and is not appressed as in T. chrysaloides. The indument on the lower surface of the leaflets of T. setifera is very characteristic; it consists of short, pale, appressed hairs that radiate in hairs in chrysaloides all directions. Similar generally point in the same direction and give the leaflets a sericeous appearance. Tachigali vasquezii. a species known from two fruiting collections made in the Cenepa Basin in northern Peru, is also morpho- It differs from T. lower logically similar. chrysaloides in its leaflets is the indument: the surface of the tomentulose, not appressed pubescent, and midrib on the lower surface has short, erect hairs. Flowers are not yet known and stipules are not present on the two existing fruiting collections. Common (Shuar name in Ecuador) and Bolivia). annotated as names are “wantsun” “cano jihui” (Chacobo name in Some specimens had been previously Sclerolobium chrysophyllum Poepp.. S. rugosum, and ysopr PI 8 — T. vasquezii, and duplicates may have been distributed under those names. Distribution. — Tachigali | chrysaloides has been reported from the following areas: western Brazil Rondônia), Bolivia, and along the eastern foothills of the Andes in (Acre, Mato Grosso, and adjacently in Ecuador; a single collection is known from the Department of Loreto in Peru. IUCN Red List category. The few known collections of this species are scattered over a large area, and a listing of Least Concern (LC) seems appropriate (IUCN, 2001). Vac a Diez. vic. of BOLIVIA. Beni: Alto Ivon, Paratypes. Chacobo village prov. Brasiléia. s Waltir 920 (MO); Reserva Extr. Chico Mende ‘s, Cid et al. 10140A (INPA Se nador Guiomard, Cid Ferreira et al. 10265 (INP E Gomes & rana Rondônia: Mun. Guajir Novos, Cid Ferreira 8768 (INPA, “MEL MO). ADOI Morona-Santiago: Bomboiza, d ue de de. Mision Palacion & Cerón 7441 (N Baker. ! : Manza zanares 15212 Indanza, valley of Rio C pens Veill ¢ Nambija, fl., Granda & Cobas 899 (MO); i in the vic. of the mu i i / d & Quizhpe Haya, ung camp at ff. 19523 (MO Loreto: Prov. Maynas, Rio Vasquez & Jaramillo 5740 (MO) 12. Tachigali chrysophylla (Poepp.) Zarucchi & Herend., Monogr. Syst. Bot. Missouri Bot. Gard. 1254. 1993. Basionym: Fab ind chryso- phyllum Poepp., Nov. Gen. Sp. Pl. 3: 60. 1845. TYPE: Brazil. Poeppig 2066 os W not seen; isotype, P!). dd sericeum Tul., Arch. Mus. Hist. Nat. 4: 12 844, non a sericea Tul. TYPE: Brazil. Pane 2666 (holotype, Discussion. Tachigali chrysophylla is besi recognized by the combination of its domatia (these evlindrical and moderately swollen and often beyond the first pair of leaflets) and leaflets with a strongly unequal base and sericeous lower surface. Stipules, when present, are pectinate. Petals are glabrous and linear. Most collections are from terra firme forest. The Venezuela, majority of collections are from Peru and with single collections from Brazil and Ecuador. This distribution pattern suggests that other collections from Colombia and Brazil remain unidentified or are misidentified. It is also possible that more than one species is filed under this name: the specimens are rather variable. Poeppig (1845) was not aware of Tulasne's publication (Tulasne, 1844). Both Poeppig and Tulasne recognized Poeppig 2666 as an undescribed 8 PPLE species and both described it independently with the same collection as type (but with different holotypes). Selected examined. BRAZIL. Amazonas: Mun. Jutai, Cid et al. 8313 ANPA, MO). ECUADOR. Easfagn: vm tococha, Gudiño 1382 (MO). PERU. Amazo- Cenepa, Ancuash 1242 (MO, US). Loreto: “ NY. US). San Martín: Lamas. below Belshaw 3477 (GH. MO, NY, . Carretera Pue p an Alejandro, 30 (NY). VENEZUELA. Amazonas: Huber & Dem 1460 (NY, US, Blanco 647 [^^ ime Hs English. Evangelica Mission, US). = Ucayali: Km 3 Simpson & Schunke | Alures, Puerto Ayacucho, VEN). Bolívar: Reserva Forestal La Paragua, (NY) 13. Tachigali colombiana Dwyer, Brittonia 14: 56. 1962. TYPE: Valle: Bahia de ventura, Cuatrecasas as (holotype, F not MO, US Colombia. Buena- seen: Isolypes, Discussion. This is the only species of Tachigali occurring in the Pacific lowlands of Colombia west of 25-30 m. acking and stipules were absent from all specimens the Andes. It is a tall tree, Domatia are seen. On the leaflets, secondary veins are impressed on the upper surface and. prominently raised on the lower surface. The nia is strongly. oblique. Petals are ca. 5 X 2 mm and pubescent near the base. Stamens are dimorphic, with three stamens shorter and thicker than the others. The flowers are sessile or nearly so. Tachigali colombiana is known from Valle and Chocó departments at altitudes from 10-150 m Volume 95, Number 4 2008 van der Werff 639 Synopsis of Tachigali en ed COLOMBIA. bdo-Istmina Rd., 28-35 km S of Quibdo, Gentry « i. 36600 (INPA, MO). Valle: Bajo C alins: Murph el P 532 (MO, US) specimens examined. Chocó: 28 kı 14. Tachigali davidsei Zarucchi & Herend.. Fl. 1998. Venez. Guayana 4: 116. Sclerolobium aureum var. grandiflorum Dwyer, Lloydia 20: 1957, non Tachigali grandiflora Huber, 1909. TYPE: jon 217 (holotype, P!). Venezuela. Río Orinoco, Chaffan- Discussion. Tachigali davidsei can be readily identified by the golden-brown indument on twigs, ack of raised tertiary venation on the upper surface of inflorescences, and flowers; the domatia: the leaflets: the pectinate stipules (when present); and the symmetrical hypanthium. The base of the leaflets is symmetrical. Stamens are monomorphic; petals are ca. 5 X 2 mm. Distribution. Venezuela This species is fairly common in tributaries Delta grows mostly along the Orinoco and its (Amazonas, Anzoátegui, Apure, Bolívar, Amacuro, Guárico, and Monagas) and along rivers or on the banks of lakes. Selected specimens examined. pied Amazo- Ayacucho, Maguire et al. 36183B (NY, US). Anzoátegui: Caño Mamo, Colonello 1239 oM. Apure: Río Davidse & Gonzalez 13387 (MO, US). Bolívar: km W of Caicara, Davidse 4362 (MO, NY, US). Delta al. 6778 (MO). 5111 ~ Amacuro: Mpio. Casacoima, Diaz el i Estación Biológica Calabozo, Aristeguieta (COL, F, NY, VEN). y . Tachigali dwyeri (R. 5. Cowan) Zarucchi & Herend., Fl. Venez. Guayana 4: 116. 1998. Basionym: Sclerolobium | dwyeri R. S. Cowan, Mem. New York Bot. Gard. 10: 83-85. 1961. TYPE: Venezuela. Amazonas: Río Guai- nia, Cano Pimichin, Maguire et al. 41831 (holotype, US not seen; isotypes, GH! MO!, NY!, PD. Discussion. Tachigali dwyert is best recognized by the combination of small flowers with petals (ca. 2 X 1.5 mm) about as large as sepals, sparse indument on inflorescences and flowers, terete domatia, and 4 to 5 pairs of leaflets. The base of the leaflets is somewhat asymmetrical but not strongly so. Stipules are pinnate with linear lobes but infrequently present. Stamens are monomorphic. The petals have a few hairs in the center near the base but are otherwise glabrous. Tachigali dwyeri is known from three collections in Amazonas, Venezuela: one from the Río Guainía, one from the Río Yatua, and one from the Río Casiquiare. The latter is sterile but has the characteristic stipules of T. dwyeri. > VENEZUELA. ea nas: Rio Pacimoni, Maguire et al. 41652 (G E MO, R); Solano, Bajo Casiquiare, Williams 14734 (US Additional Spas examined. 16. Tachigali eriopetala (Ducke) L. F. Gomes da Silva & H. C. Lima, Rodriguésia 58: 399, 2007. Basionym: Sclerolobium eriopetalum Ducke, Arq. Inst. Biol. Veg. 2: 41. 1935. TYPE: Brazil. Amazonas: “prope Manaus silva non inundabili circa cataractas fluminis Taruma.” Ducke s.n. (RB 24296) (holotype, RB not seen; isotypes, K!, P!) — Discussion. This species is best recognized by its — elliptic leaflets with dense, appressed to sericeous indument, absence of domatia, and, in flowering specimens, absence of stipules; small, entire, revolute bracts are sometimes present in the inflorescences. Petals are linear and densely pubescent. The base of the leaflets varies from obtuse to rounded and is only slightly Reserva Florestal de Curua-Una in Pará are also asymmetrical. Specimens collected in the included here. These specimens have larger, oblong leaflets with more pairs of secondary veins (6 to 8 in typical Tachigali eriopetala, ca. 10 in specimens from Curua-Una) and large, revolute stipules at the base of the leaves. Most of these specimens are sterile. The few fertile collections from this area have leaflets much like typical T. eriopetala and, with the exception of Cavalcante 1510, do not have stipules at the base of the leaves. Tachigali eriopetala has been reported from terra firme forest in Amazonas and Pará. It is similar to 7. hypoleuca (Benth.) Zarucchi € Herend., which differs in having foliose stipules and being restricted to flooded forest. It can also be confused (Huber) L. F. Gomes da Silva & H. C. Lima, a species with glabrous petals, floral bracts " T. goeldiana with longer than the buds, and leaflets with more (10 to 15) pairs of secondary. veins, and that grows in flooded forest. Several specimens placed in T. eriopetala previously had been annotated by others as Sclerolobium chrysophyllum, a species with domatia and pectinate stipules. Selected specimens examined. BRAZII Amazonas: Mun. Maues. Zarucchi et al. 3117 (INPA. MG, MO). Para: Santarem, Cavalcante & Silva 1510 (IAN, MG, NY) 17. Tachigali ferruginea van der Werff, sp. nov. TYPE: Ecuador. Pastaza: 40 km 8 of Puyo, along rd. to Macas, near Pitirishka village, 01 48'S, 77°47'W, 26 Nov. 2000, D. Neill 12932 (holotype, MO!; isotypes, K!, NY!). Figure 9. congeneris foliolis magnis basi rotundatis vel cordatis, nervis lateralibus numerosis et indumento ferrugineo recedit. Tree, 20 m, twigs ferruginous pubescent, the hairs partly erect, partly appressed ascending. 640 Annals of the Missouri Botanical Garden MISSOURI BOTANICAL GARDEN HERBARIUM DULL LLL LUC N? 5301396 "Pu ed by Henk Dete 1 W. 199 " Missouri Botanical Gatden 1997 Vole kane ECUADOR AAA m E TESE L——NEN —— 2 3 4 5 Md 1 6 1 copyright reserved MISSOURI BOTANICAL GARDEN HERBARIUM (MO) Figure 9, Tachigali ferruginea (holotype). Volume 95, Number 4 2008 van der Werff Synopsis of Tachigali Supules pectinate, the lower lobes again pectinately leaf leaflets. the divided, densely pubescent. Incomplete with rachis ca. 40 cm, 5 pairs of rachis ferruginous pubescent, strongly sulcate on the lower half, and hollow, but domatia not abruptly swollen, the 10 cm base leaflets from 16 X 7 em (basal ones) to 30 X (c symmetrical, = istal ones), oblong or oblong-obovate, the rounded to cordate, the tip abruptly acute, the upper surface sparsely appressed pubes- cent, the indument denser on the midrib and this also with scattered, erect. hairs, lower surface densely ereenish brown pubescent, the hairs appressed and completely covering the surface, the primary and secondary veins with some erect, ferruginous hairs: lateral veins 20 to 25 on each side. Incomplete inflorescences to 30 cm, ferruginous pubescent as on the twigs, paniculately branched, the ultimate, flower- 1-6 em: bracts nol caducous. bearing parts seen, Flowers densely grouped together, sessile, symmetri- cal, hypanthium and lobes densely pubescent, the 3 X 2 mm, the cup petals linear, lobes erect, ovate, 5-2 mm, symmetrical: glabrous, 3-4 mm, sta- mens 10, equal, ca. 4 mm, with some stiff, ferruginous hairs near the base; ovary ca. 2 mm, densely ferruginous pubescent, style slender, ca. 2 mm, glabrous. Ovary attached in the center of the hypanthium. Fruit laterally compressed, ca. 16 X 4 em. with | central seed. Discussion. Tachigali ferruginea is only known from the flowering type collection, gathered in secondary forest along a road; a second, fruiting 7598) fl Owers are collection (Neill & Ceron possibly belongs The old The ector noted that the leaf petioles are inhabited here. reddish brown. col by fierce, stinging ants. The new species can be readily recognized by its large leaflets with many pairs of secondary veins, its ferruginous indument, its pectinate stipules, and its symmetrically based leaflets. The fruiting collection consists of fallen fruits and leaflets and thus does not show the leaf stalk, domatia, or stipules ferruginous indument on the twigs presence or absence of characters required to confidently name the collection to species. It is largely placed here because of the number of secondary veins (20 The is dense, but partly worn 25) and the symmetrical base of the leaflets. indument on the old leaflets off and paler than on the type specimen. According; to the collectors, the fruiting tree appeared to be monocarpic and was dying. Several other species in Peru and adjacent Brazil also have large, densely pubescent leaflets, linear petals, and a symmetrical differ pilose petals, hypanthium, but these species n stipule type (entire and revolute), general indument type, lower number of secondary veins per leaflet, or lack hollow, ant-inhabited leaf rachises. The type collection was initially identified as Tachigali formicarum Marms by the collector, and duplicates may have been distributed with this identification. IUCN Red List category. S only known from two recent collections and is listed as Data Deficient (DD) (IUCN, 2001) Tachigali ferruginea Paratype. ECUADOR. Napo: Reserva Biologica Jatun Sacha. O1 04/8, 77 36'W, 17-23 Jan. 1987. D. Neill & C. (MO). Ceron 7598 18. Tachigali 5 Vereins TYPE: type, Verh. Bot. Brandenburg 48: 164, 1907. Peru. Loreto: Tarapoto, Ule 6538 (holo- Y: isotypes, K!, L!). formicarum Harms. Prov. Berlin- Tachigalt tessmannii Harms, Notizbl. Bot. Gart. } ru. Loreto: Dahlem 9: 967. 1926. Syn. nov. TYPE: Pe Tessmann 4753 (holotype, BT. photo MO!). Discussion. This species is easily recognized by the large, semiterete domatia, flowers with petals about the same size as sepals, petals with long yellow hairs, and the generally large leaflets. The yellow hairs on the petals are readily visible in opening flowers. The base of the leaflets is asymmetrical. Stamens are monomorphic. The hypanthium is symmetrical. Stipules are caducous and were absent from all this study. A fruiting, leafless collection (Morawetz 15-24988, belong here, suggesting the species is monocarpic. Although inhabited domatia, the presence of aggressive ants specimens seen in K) seems to many species of Tachigali have ant- is, in most species, only infrequently mentioned, but nearly all collections of T. formicarum mention the presence of stinging ants ("leaves full of tangarana ants which sting painfully,” Woytkowskt 34445; “hormigas que pican fuerte,” Schunke V. 5255). thus 8 | | the epithet of this species is well chosen. Harms distinguished 7. tessmannii from T. formicarum by the more asymmetrical base of the leaflets of the latter. Collections cannot be separated based on this character alone, and therefore T. tessmannii is here placed in synonymy. PERU. Amazonas: Río Huashikat 2306 (MO). an Alejandro, Schunke Selected. specimens examined. Comunidad C Santiago, calerpiza, : Distr. Calleria, Km 90 S tio Marañon, just above Pongo de 7 (F, GH, US). Paseo: Pe E. pei "Achille 10213 (MO). 88, Pucallpa—Tongo Maria Rd., ileazu, Ucayali: Km Gentry et al. 36368 (MO). TYPE: Tarauaca, D. Daly. 19. Tachigali fusca van der Werff, sp. nov. Acre: Mun. Universo, 20 Sep. Tarauaca, Rio 1994, Brazil. seringal Annals of the Missouri Botanical Garden Lima & N. Goldstein Figure 10. Silveira, R. Saraiva, L. A. 8264 (holotype, NY: isotype, MO!). dis. linearibus praeditas Inter species. domatiis et pel floribus glabris et indumento ramulorum inflorescentiarum- que fusco diversa est. Tree, to 22 m: twigs angular, densely fuscous tomentulose to tomentose, the surface covered by the indument. Leaves (often incomplete on herbarium 70 cm. i petiole 4-6 mm diam., specimens) with 4 to 12 pairs of leaflets: the domatia ca. 4 em from the base, swollen, cylindrical, 8-10 mm diam., 2.5-8 cm. below the first pair of leaflets; leaflets 15-18 5 em, oblong, the base unequal, with | lobe cordate, the other obtuse, the tip acute, the upper surface glabrous, lower surface densely appressed pubescent and with erect fuscous hairs along the major veins; secondary veins 12 to 15 a similar indument to that on the leaf rachis. Stipules » pairs per leaflet: petiolules 4—5 mm, with nol seen. Inflorescences profusely branched, densely fuscous pubescent, 15-25 em, the ultimate flower- bearing branchlets 4—7 em: bracts not seen. Flower buds turbinate, with a few hairs at the very base, otherwise glabrous; pedicels less than | mm, pilose. Hypanthium ca. | X 1.5 mm, glabrous. Sepals glabrous on both surfaces except for the ciliate margin, ca. 2 mm concave; pelals linear, slightly widened in the distal half, glabrous; stamens equal, ca. 3 mm, moderately brown pubescent on the proximal 1/3. Pistil ca. 2 mm, densely fuscous pubescent, the style slender, glabrous, about as long as the ovary. Hypanthium symmetrical or nearly so. Fruits (immature) ca. 7 X 1.5 em. Discussion. Tachigali fusca is easily recognized by the combination of glabrous flowers, large, densely pubescent leaflets, the size of the domatia. and the fuscous indument. The few other species of Tachigali with. glabrous flowers all have much smaller leaflets that are glabrous or sparsely pubescent: none have a fuscous indument on the twigs and inflorescences. Only T. ferruginea shares large, oblong, densely pubescent leaflets, linear petals, a symmetrical hypanthium, ae the presence of domatia in the leaf petioles with 7 fusca, but the former has uniformly pubescent flowers, a symmetrical base of the leaflets. and a ferruginous indument, among other differences. Tachigali fusca is only known from three collections. The two paratypes are in fruit and have a less dense indument on the s. Both have a fuscous indument, long domatia. and leaflets with an unequal base. One of the paratypes has hairs on the while the some hypanthium, hypanthium is glabrous in the other paratype. IUCN Red List category. only known from three recent collections, it is listed as Data Deficient (DD) (IUCN, 2001). Because this species is Lima, margen direita 10052 (NPA, NY): Mun. de Cruzeira do Sul. seringal Iracema., Cid orn et al. 10781 (MEL). Mancio Paratypes. BRAZIL. Acre: Mun. ao alto rio Moa. Cid Ferreira et al. 20. Tachigali glauca Tul.. Arch. Mus. Hist. Nat. 4: 165. 18944. TYPE: “Guiana Gallica” [French Guiana isotypes, LL P!). . LePrieur 336 (holotype, Pt: Selerolobium koy Ducke, Arch. Jard. Bot. Rio de Janeiro l: 30. . Tachigali kaa n ke) Duc ke Me h. iun m Rio de Janeiro 3: 91. 1922. Syn 2 Brazil. Pará: Belem, Pa s.n. (MG 15659) dl MG not seen; isotypes, BM! R! USD). Discussion. Characteristic is the sparsely pubescent inner surface of the petals, with only some hairs in the basal. central part. Other species wilh a sericeous lower surface of the leaflets and domatia have a dense indument on at least the lower half of the petals (often on the lower 2/3). Tachigali glauca has slightly dimorphic stamens, foliose stipules (although these are infrequently present). terete domatia, and petals ca. 4 mm. The base of the leaflets species is quite similar to F. argyrophylla and differs s asymmetrical but not strongly so. This mainly in the sparse indument on the inner surface of the petals. The name Tachigali glauca has not been used since Tulasne (1844); Dwyer (1954) differences between this species and T. myrmecophila did not mention it. No were found in this study, and therefore T. myrmeco- phila is placed in synonymy. Distribution. Tachigali glauca has been reported from Para, with two collections from Amapa, two from French Guiana, and one from Roraima. BRAZIL. Cowan 38220 (LAN, S) Para: Jaci 27000 (IAN. US). Roraima: da carat, Km 343, Steward et al. 106 (INPA, P). GUYANA. « Selected specimens examined. Amapá: Serra do Navio, C [AN Tocantins, Froes NA. Leprieur 330 (L. 20 km S of Mabura Hill, Polak 321 (MO). 21. Tachigali goeldiana (Huber) F. Gomes da Silva & H. C. Lima, Rodriguésia 58: 399, 2007. Basionym: Sclerolobium goe a E Bol. Mus. Paraense Hist. Nat. 6: 78. TYPE: Brazil. Pará: Rio Capim. Huber. s.n. e" 6092) (holotype. MG! isotype. BM!). Discussion. — Tachigali goeldiana is a small tree of riverbanks and flooded forests that is best recognized by its suff leaflets with a symmetrical base and a dense. appressed, golden or silvery lower surface of the leaflets. Domatia are absent. Stipules are pectinate but rarely present. Petals are linear and glabrous. The inner surface of the sepals is glabrous. For their size, the leaflets have many lateral veins (10 to 15 on each side). Tachigali goeldiana has been confused with Volume 95, Number 4 2008 van der Werff 643 Synopsis of Tachigali Figure MISSOURI 9 10 copyright reserved BOTANICAL GARDEN 10. co YA ew BE Missouri desi Gatden Tachigalt fusca (holotype). MET Ke Mol pe Tachigalt SA NORTA kB ical Gard As E a Flora of Acre, Brazil Caesalpiniaceae Aubl. Brazil. Acre. Mun. TUE Rio Tarauacá, river at low water Sering ral Universo. $ 267 30"S, 7 1'12"W Recent ae relatively level terrain, low-lying areas flooded for several months in winter (Nov-Mar). Primary forest stamens yellow. N.V.: tachi preto. N.V. tachi preto 20 Sep 1994 Daly, E NYdb M. Silveira, R. Saraiva, L. A. Lima, N. 8264 Annals of the Missouri Botanical Garden hypoleuca, also a species of flooded. forests with sericeous lower leaflet surface. However, T. hypoleuca differs in its foliaceous stipules, pilose petals, and fewer (up to six) pairs of lateral veins. Tachigali goeldiana is known from Pará. Amazonas, and Roraima in Brazil. Selected specimens e ele BRAZIL. Amazonas: [ha da Costa, Arirarra, Poole J (INPA, MG. MO, NY. US). Para: Rio Capim, lg. Caju, Froes 24113 (IAN, NY, RB, US). Roraima: Rio Branco, al mouth of Rio Hapera, Mori et al. 20414 ANPA, MO, NY). 22. Tachigali Königl. Bot. Gart. Brazil. Roraima: Mairary, Ule 83 [fragm.], KN E). Notizbl. TYPE: Serra de grandistipulata Harms, Berlin 6: 304. 1915. Rio Branco, Surumu, 99 (holotype, BT: isotypes, F! Discussion. This species resembles Tachigala rigida Ducke, but differs in its smaller flowers (petals and mono- 'a. 5 mm vs. petals 8-9 mm in T. rigida) morphic stamens (4-6 mm). The base of the leaflets is in 7. rigida. Domatia are present and semilerete. Stipules are obtuse to rounded and more asymmetrical than foliose and large, to 5 X 3 em in the specimen at K. The lower leaf surface is completely covered by the appressed indument. The stamens are about as long as the petals, while in 7. rigida stamens are clearly longer than the petals. Tachigali grandistipulata is only known from the lype collection from Roraima. ZLaruechi « 23. Tachigali guianensis (Benth.) Herend., Monogr. Syst. Bot. Missouri Bot. Gard. 45: 1254. 1993. Basionym: Sclerolobium guia- nense Benth., Hookers J. Bot. Misc. 2: 237. 1850. TYPE: Guyana. Rob. Schomburgk, 2nd coll. 598 (lectotype, designated by Zarucchi & Herendeen in Brako «€ Zarucchi, 1993: 1254. K not seen: duplicates, BM! PD. Kew Gard. Sclerolobium radlkoferi Rusby, Mem. Torrey Bot. Club 6: 26 Se er 189 rolobium guianense var. radlkoferí (Rusby) ee Lloydia 20: 98. 1957. Syn. nov. TYPE: Bolivia. Betw. eas mi E Guanat, Bang 1690 (holotype, NY not isoly GH! KY) Verh. Si ead Sc Imoli an ulea anum TOA . Vereins Branden- X burg 48: 168. 1907. tinctorium var. Lloydia 20: 95. 1957. Zarucchi & Missouri Bot. Gard. 45: 1254. Ule 6450 (Harms) Dwyer ] (Harms) uleana Herend., 3 | Tarapoto, (ollas, Sclerolobium balla Due M 1935. | Arq. Inst. pio Veg. 2: 42. (Ducke) L. F. rc Rodriguésta 58: o le a subbullata Silva & H. C. Lima, nov. uM Brazil. Amazonas: ostium Javary,” Ducke s.n. (RB isotypes, KI, PH USD, fluminis (holotype, RB not seen: Discussion. Tachigali recognized by the pilose indument on twigs, leaves, guianensis is readily and inflorescences; this is especially noticeable on voung growth. The base of the leaflets is asymmet- rical, stipules are pectinate, and the great majority o the specimens turn black upon drying. Petals are linear and glabrous, and domatia are lacking. Collections from Peru and Bolivia have a denser indument and have been described as Sclerolobium radlkoferi (from Bolivia) and S. uleanum Harms (from Peru). In this study, these specimens with a denser indument are included in T. guianensis. Ducke said 5. subbullatum differed from T. guianensis in its larger less. bullate-rugose leaflets. However, leaflets and more oi and specimens with larger wider occur throughout the range of T. guianensis, and while some collections from Ecuador and Peru (Neill 14618, Grandez 1395) have bullate leaflets, others have flat leaflets. The concept of T. guianensis here presented broad. and more detailed studies is thus fairly may lead to the reestablishment of some synonyms as distinct spectes. Distribution. Tachigali guianensis has a wide distribution and has been collected in Brazil (Amapá, Amazonas, Para, and Roraima). Guyana, French Guiana. Suriname, Venezuela (Amazonas, Bolívar, and Delta Amacuro), Colombia (Putumayo), Peru (Cusco, Loreto, Madre de Dios). and Bolivia. Selected Diez, (10 km 8 of Puerto America, Jardim Noel Kempf, Quevedo 2697 Camaipi, on rd. to Rio Preto, specimens examined, BOLIVIA. Beni: Vaca Boom 1308 (MG, MO, NY) " Paz: Copacabana of Mapiri), Krukoff 11020 - Pando: 35 km N 2442 (MO, ux Santa Cruz: PN (MO, ae BRAZIL. Amapa: Mori et al. 16545 (MG, NY). Amazonas: n Humayta, near Livramento, Arukoff 6903 (MO, NY. US). Para: Rio Jipuru. Hha do Marajo, Rabelo & Rosa 3605 (MO. NY). Roraima: Ilha de Maraca, Rattler et al. 6516V (NY). C n. AMBIA. Putumayo: P uerto Asis, entre Umbria y uatrecasas 10543 (COL, F. NY 5 entro Shuar Yunkumas. s (MO). FRENCH Moron tiago: C © Dendrology Course pp o SP nd n 59 (MO). Lore ‘Lo: Varpa. Grandez 1395 (INPA). Madre de Dios: Tambopata, Explorers Inn, 51539 (MO). SURINAME. Blakawatra creek, 10 km E of Jodensavannne. Se WY) VENEZUELA. NUN Puerto Galantapo, Castillo 2191 (MO, . Bolivar: 7 Piar, Liesner & Gonz ia ? 185 (MO, US). Amacuro: Di Tucupita, Ll km ESE of Los C lid de Guayana, Fine & Gonz i m 206 (MO, NY, Gentry et al. Ayacucho, Rio Ciudad 24. Tachigali hypoleuca (Benth.) Zarucchi & Herend., Monogr. Syst. Bot. Missouri Bot. Gard. 45: 1254. 1993. Basionym: Sclerolobium hypoleu- cum Benth.. Hookers J. Bot. Kew Gard. Mise. 2: 236. 1850. TYPE: Amazonas: woody hills al Barra do Rio Negro, Riedel s.n. Brazil. (holotype, K!). Volume 95, Number 4 2008 van der Werff Synopsis of Tachigali Discussion. This species is best recognized by the sericeous indument of the leaflets, its foliaceous stipules, the nearly symmetrical base « the leaflets, and the linear, pilose petals. Domatia are absent. Nearly all collections are from flooded forest. It is similar to Tachigali ertopetala, which is a tree from terra firme forest with caducous stipules and which has small, revolute bracts in the inflorescences. Differences with T. goeldiana are discussed under that species. Tachigali hypoleuca is a rather common Roraima in Brazi species from Amazonas and collection from Venezuela Stergios 9703, from boe leaf stipules it agrees with T. hypoleuca, but it has small IS tentatively placed here: in characters and swollen domatia at the base of the leaf rachis: flowers are too old to assist in identification. Selected specimens examined. BRAZIL. ped Manaus, Igarape da Cachoeira Grande, Ducke 523 (F, GH, Y. R). Roraima: Rio Xeriuini Pires et i I. a 4 AN, US). 25. Tachigali inconspicua van der Werff. sp. nov. TYPE: Ecuador. Napo: Estación Biológic “a 1-31 Dec. 7905 (holotype, Jatun 1986 (fl). D. Neill & MO!: Sacha, 450 m, 2 M. Asanza isolype, KI). Figure 11. Tachigali amplifoliae similis. sed ab ea foliolis oblongis vel. elliptico-oblongis basi plus minusve symmetricis nervis lateralibus 8 ad 11 (nee 5 ad 7) recedit. 30—40 m: Stipules persistent, l- Large trees, twigs angular, glabrous. 3-foliose, 1-2 1-2 cm, the 2 basal lobes when present smaller. Leaf rachis 20-30 cm, sparsely appressed pubescent, somewhat canaliculate, especially just below the attachment of leaflets sparsely appressed pubescent, the leaflets: domatia lacking: 5 to pairs, petiolules 5-8 mm, 15 X 2.5-5 om, charta- ^^ oblong to elliptic-oblong, ceous, the base + symmetrical. apices acute or acuminate, acumen ca. | em, secondary veins 8 to 11 pairs per leaflet. Inflorescences paniculate, moder- ately to densely appressed pubescent, 15-20 cm, the 10 cm: pedicels 0.6-1 mm, buds uni- ultimate, flower-bearing racemes lo bracts caducous, nol seen: formly grey appressed pubescent. Hypanthium turbi- nate, symmetrical, ca. | 1.5 mm. sepals ca. 2 mm, elliptic, concave, DNE inner surface appressed pubes- cent; petals ca. 2 mm. linear, densely yellow pubescent; stamens equal, 4-5 mm, the filaments yellow pubescent on the basal quarter; ovary. 1.5- stiff, samaroid, ellipsoid, ca. 8 X 2mm, covered with rown hairs. Legume 2.5 om, l-seeded. Discussion. As the name suggests, Tachigali inconspicua does not have striking features. It belongs in the group of species with linear, pilose plane, petals, and a symmetrical hypanthium that lacks domatia. In this group it stands apart by the combination of rather large, foliose stipules, yellow hairs on the petals, and a more or less symmetrical hase to the leaflets. It is closest to T. amplifolia, which broader leaflets (to 20 X 11 em) 7 vs. 8 to Il) and a Tachigali known from the Andean foothills from central Ecuador differs in its larger, with fewer secondary veins (5 to more asymmetrical base. inconspicua ds (Jatun Sacha) to the Río Cenepa Basin in northern Peru. All collections known to date are from tall trees. most estimated as 30m and one as 40 m tall. Eighteen duplicates were made of the type collection, all but the holotype were distributed T. paraensis. IUCN Red List category. Given its distribution in Amazonian Ecuador and Peru, this species is listed as Least Concern (LC) (IUCN. 2001) ECUADOR. Palacios 3179 ME the Dicaro community. Paratypes. de Coca. W. Reserve, vie. « (MO). a 8 km de Napo: Orellana Canton, 5 km N (MO Huaorani Ethnic V. Pitman et al. 4305 Morona Santiago: carretera Bomboiza—Gualaquiza, . J. Zaruma 359 (MO). ed dn downriver a Mision, Zamora- from Miazi to Palacios 9708 (MO); same (MO). PERU. S of Rio Cenepa, fl., B. Chinchipe: Río Pachicutza, fl.. D. Neill & W. locality, fr, D. Neil & W. Palacios 9707 azonas: 5 ol oar trail. Berlin. "1660 (MO, US): e ew de Mamavaque. p MO). Loreto: Fundo cion E 1010 (MO, NY, US): Iquitos. 1814 (F, IAN, MG 5 = = E. = ai uu prov.. El Cenepa. . Ro A. Pe na & E. Chávez Alto do Maja. A. Ducke Gomes da Rodriguésia 58: 399, 2007. Ducke, Bol. TYPE: Brazil. Ducke 1028 here, RB not MO! RD). 26. Tachigali leiocalyx (Ducke) L. F. m » Silva & H. C. Basionym: Lima, Sclerolobium i alyx N. 2: 19. 1944. Sao Paulo e ies Teen. Inst. Agron. Amazonas: (lectotype, designated duplicates, GH!, K! seen: Discussion. This species is easily recognized by the glabrous sepals, linear, pubescent petals, and the coriaceous, e Hiptic loe Iliptie -ovale leaflets. The base of the asymmetrical. leaflets is obtuse. to rounded, weakly Domatia are. lacking, and stipules have not been seen on the few known specimens. known collections According to the labels, the three were all made in caatinga forest near Sáo Paulo de Silva and Lima (200 holotype is in RB: unfortunately, a lectotype has to be Olivenca. 7) indicated that the selected because Ducke (1944) cited two collections Two distributed under One as syntypes. were Ducke 1026 species consists of the e. and isolectotypes of Tachigali leiocalyx (in GH, R. and RB). while the other Gn MO ed US) represents 7. inconspicua. 646 Annals of the Missouri Botanical Garden avg NY1Of LINOSSIW N Nad Wo! panesel 1uBuAdoo OL — MISSOURI BOTANICAL GARDEN HERBARIUM No 05083783 Vac gods o Co Sp E S à wed Determined by Henk san der Werff 194 Use syrup Missouri Botanical Gatden zest Vole kyi ECUADOR 1 par aber ) Barneby de D.A Neill, 297 Napo: Tena Canton a Estacion Biológica Jatun Sacha 2110 Napo, 8 km abajo de Misahualli. Bosque Muy Húmedo 450 m 1 Jy vestigio Marið, riores amarliias. o tà é ('Ormigas habitantes en los Decaioio2. Flores en FAA. 24 Dec 1986 31 Dec 1986 David Neill, M. Asanza 7565 HERBARIO NACIONAL DEL ECUADOR (ACNE) MISSOURI BOTANICAL GARDEN HERBARIUM (MO) Figure 11. Tachigali inconspicua (holotype). Volume 95, Number 4 2008 van der Werff Synopsis of Tachigali Additional specimens examined. BRAZIL Mp Sao Paulo de Olivença, Ducke 1498 (NY, R): Duna et al. 2 (M ). 27. Tachigali longiflora Ducke. Arq. Inst. Biol. Veg. 2: 38. 1935. TYPE: Brazil. Amazonas: “Sáo Paulo de Olivença, silva non inundabili,” Ducke s.n. (RB 24291) (holotype, RB not seen; isotypes. KI PL USD. Discussion. | Tachigali longiflora is a striking spe- cies because of its very dark brown inflorescences and flower buds. The flowers are rather large for the genus (to 1.8 em), and the hypanthium is unusually long and curved; this is best visible just after anthesis and after the petals have fallen. Stamens are dimorphic. Petals are large, ca. 8 X 4 mm. Stipules have not been seen for this study: Ducke mentioned he had seen one incomplete stipule, this pinnate, with narrow segments. Domatia have also not been seen in + this study, though Ducke mentioned the presence o small ants in the leaf rachis. The leaflets have an — asymmetrical base and vary from ovate-elliptic (basa pairs) to oblong (distal pairs). Additional specimens examined. PERU. Loreto: naro 5358 (MO), Daly et al. MO). Spichiger 1490 (MO). Herrera, Huamantupa et al. (MEL. 28. Tachigali loretensis van der Werff, sp. nov. PYPE: Peru. Loreto: Prov. Maynas, Alpahuayo, 12 Nov. 1984 (fl), R. Vasquez J. Ruiz & N. Jaramillo 5888 (holotype, MO!; isotypes, Fl, K! '). Figure 12. Tachigali ptychophyscae Benth. similis. se ed ab ea floribus minoribus et staminibus uniformibus recec Tall 40 m; densely appressed pubescent. Stipules not seen; leaf rachis 10—20 cm. appressed pubescent, trees, twigs terete, minutely and minutely and moderately densely terete; domatia present, te- — rete, 1.5—3 em, 4-6 mm diam., less than | em from the base of the leaves; leaflets 4 to 6 pairs, 5-13 X 3-6 em, ovate to elliptic, chartaceous, the upper surface glabrous, lower surface inconspicuously and the moderately to sparsely appressed pubescent, o acuminate, base asymmetric, the apex acute acumen to ] em, secondary veins 3 to 8 pairs per leaflet. Inflorescences — paniculate, 15-25 em, densely and minutely appressed pubescent, the ultimate racemes to 10 cm, bracteoles ca. 5 mm, narrowly ovate to linear, falling off as buds mature and absent from the base of the mature flowers; pedicels to | mm, buds uniformly and densely grey pubescent. Hypanthium asymmetrical; sepals broad- ly ovate, 2-3 X ca. 1.5 mm, pubescent on the inner surface; petals slightly longer than the sepals, 2 mm wide, ovate, pubescent on the inner surface, glabrous on the outer surface, stamens 10, mono- morphic, ca. 5 mm, pubescent near the base: ovary ca. 1.5 mm, densely pubescent, style ca. 2.5 mm, pubescent near the base, becoming glabrous distally. Fruits unknown. Discussion Tachigali loretensis Both terete domatia Vegetatively, closely resembles T. ptychophysca. species have leaflets with few lateral veins, close to the base of the petiole, and a similar leaflet shape. The specimens from Colombia of T. loretensis have more lateral veins, up to eight pairs per leaflet. However, T. ptychophysca differs in its larger flowers with sepals 4-5 mm, petals 5-6 mm, and stamens 10— 12 mm, and in its dimorphic stamens. The appressed indument on Mhe lower surface of the leaflets is also denser in and almost ptychophysca appears sericeous. Tachigalt loretensis known from three flowering collections made in non-flooded forest on white sand near Iquitos and two flowering collections TI^ rom Vaupes, Colombia. Seedlings and juvenile trees have distinctly acuminate leaflets, with an acumen up to 4 cm. IUCN Red List category. not have a Tachigali loretensis does Peru adjacent Colombia). However, sterile collections from wide distribution. (northeast and forest plots suggest that it is locally common and it is therefore listed as Vulnerable (VU) (UCN, 2001). Paratypes. COLOMBIA. Vaupes: Río Kananari. Cerro Isibukuri, (COL. GH): Río Schultes & Cabrera 14091 anay, fl.. Gentry et al. 39315 (MO): Alpahuayo, Vasquez & Criollo 5761 (MO). Prov. Maynas, 29. Tachigali macbridei Zarucchi & Herend.. Monogr. Syst. Bot. Missouri Bot. Gard. 45: 1254. 1993. Replaced synonym: Sclerolobium Field Mus. Nat. Hist., Bot. non Tachigali rigida Ducke, rigidum J. F. Macbr., 13: 201. 1943, 1938. TYPE: Peru. Loreto: Pumayacu. Alug 3239 (holotype, F not seen: isotypes. GH! MO!) Discussion. This species can be easily confused with Tachigali chrysophyllumz; it is best separated from this species by the erect indument on the leaf rachis. In this character, it approaches T. guianense, but the latter lacks the sericeous indument on the lower surface of the leaflets. Stipules are pectinate. Domatia are often present in large leaves, but lacking The Petals are linear and glabrous. leaflets is Most in small leaves. base of the asymmetrical. collections are from upland forest. Tachigali macbridei is known from six departments in Peru, with a few collections from adjacent Brazil (Acre). Annals of the Missouri Botanical Garden MISSOURI BOTANICAL GARDEN HERBARIUM MISSOURI BOTANICAL GARDEN N? 6016792 copyright reserved Tachigali loretensis van d Werff olak yy Determined by: Henk van eh Weill ae PERU Prov. Mayne aS» Dpto. Loney dh pahuay« ( stacion IIAP). ores amarillas ve 12 dee 1984 Pulz Jaramiltc li ISSOURI Qut E GARDEN pom (MO) Figure 12. Tachigali loretensis (holotype). Volume 95, Number 4 van der Werff 649 2008 Synopsis of Tachigali Several collections from Peru (Amazonas) have Trombetas-Rio Mapuera area in Pará, Brazil. Ducke nearly sessile leaflets with a cordate base clasping the rachis. Selected mouth of m examined. BRAZ Acre: near Krukoff 54.69 pam MO, NY). maza, Comunidad Agauruna de pecas ns Mac auhan, O). Cusco: Distr. Echarate. Kiteni, Huamantupa et al. 7225 (MO). Huanuco: Tingo Maria, Asplund 1 (NY). Loreto: Río Nanay, carretera de Pieuruyacu, Rimachi 4034 (MO, NY). Madre de Dios: Parque Nacional Manu, Foster et al. 11850 (MO, Martín: Schunke | NY). San Huicte. Caseria Nueva Union abajo de Puerto 7936 (MO, US). 30. Tachigali macropetala (Ducke) L. F. da Silva & H. C. 399. 2007. Sclerolobium macropetalum Ducke, Arq. Inst. Biol. Veg. 2: 41. 1935. TYPE: Brazil. Amazonas: “Rio Negro superioris inter Gomes Lima, Rodriguésia 58: Basionym: Camanaos et ostium fluminis Curicuriary, Ducke s.n. (RB 23528) (holotype, RB not seen: isotypes, K!, P!). Discussion. Tachigali macropetala is a very distinct species and is easily recognized by the essentially glabrous sepals (but with a fringe of hairs along the margin), the symmetrical hypanthium, rather long pedicels (3-6 mm), elliptic petals clearly longer than the sepals, and leaflets with a symmetrical base and immersed venation on the upper surface. Domatia are lacking. Stipules are 2-foliose, 1-2 cm, but rarely All but one of the collections are from riverbanks or flooded forest along present. Stamens are monomorphic. the upper Rio Negro. BRAZIL. Rio Amazonas: Ducke Selected a examined, Curicuriary, NY) 31. Tachigali macrostachya Huber, Bol. Mus. Paraense Hist. Nat. 5: 387. 1909. TYPE: Brazil. Pará: "in ripis insulae Veneza fl. Mapuera,” Ducke s.n. (MG 9050) (holotype, MG not seen; BM! — isolype, Discussion. This is a very distinct species because leaflets cordate, symmetrical base, and large flowers with 15 of its large, oblong-ovate, bullate with a [evi or more stamens. The petals are 7-8 mm long and pubescent on the inner surface. Stamens are dimorphic. The leaf petioles are not noticeably swollen, but are hollow and inhabited by ants. Leaf rachises are strongly | to The hypanthium is — angular. Stipules are to 3 em long and foliose wit pairs of lobes, but caducous. strongly asymmetrical. Tachigali macrostachya is a small tree growing on riverbanks. Dwyer (1954) listed the Río Mapuera in possibly misled by Venezuela as the type locality, the actual type locality given by Huber ("in ripis insulae Veneza”). all from the Rio collections seen in this study are (1949) also reported it from the Rio Jamunda (border of Pará and Amazonas). Selected | specimen examined. BRAZIL. Pará: Mun. Oriximina, Rio Trombetas, Cid et al. 1256 (INPA. MG, 32. Tachigali melanocarpa (Ducke) van der Werff, Sclerolobium melanocar- Arq. Inst. Biol. Veg. 2: 43. 1935. Brazil. Amazonas: “prope Borba, Rio r." Ducke s.n. (RB 23330) (lecto- type, designated here, RB not seen; duplicates, KL PI. comb. nov. Basionym: pum Ducke, TYPE: Madeira infe Discussion. Tachigali melanocarpa is very similar to T. paraensis in leaflet shape (both have a cuneate, more or less symmetrical base), and the two are easily confused. The clearest difference seems to be found in the densely hairy, linear petals: with yellow hairs in 7. On average, the leaflets are larger in T. paraensis. Also, in fall whereas in 7. paraensis stipules a melanocarpa and white hairs in T. paraensis. T. melanocarpa | stipules early from young branchlets, are frequently persistent, unifoliate, and foliose, to 2.5 cm. In both species, small, revolute bracts are occasionally present along the inflorescences. Domatia are absent. The hypanthium is symmetrical. Ducke (1935) also stated that the rachis and leaf petiole are canaliculate in 7. melanocarpa, although this was not apparent on material seen in this study. The leaf rachis is somewhat flattened in 7. melanocarpa and terete in T. paraensis, but this single character difference hardly warrants recognition of two separate species. Tachigali melanocarpa has been reported from Amazonas. and Pará, Brazil, and occurs in terra firme forests. The protologue cites Ducke RB 23339 as one of the syntypes, but the K and P sheets carry the number Ducke RB 23330. Ducke cited his description of Sclerolobium melanocarpum, two collections in one with flowers and one with fruits. The RB specimen of the flowering collection is designated here as the lectotype. Selected specimens examined. BRAZIL. Amazonas: Reserva 1501, Projecto gue Florestais, 90 km NNE de Manaus, Oliveira et al. 561 (INPA, MO, NY). Pará: Jari, Estrada entre Monte a e Mungaba, Silva 1978 US). 25 melinonii Syst. 33. Tachigali Herend.. 45: 1254. (Harms) Zarucchi & Bot. Missouri Bot. Basionym: Sclerolobium melino- nii Harms, Bot. Jahrb. Syst. 33 (Beibl. 72): 24. 1903. TYPE: French Guiana. Maroni [Saint- Laurent-du-Maroni|: Melinon s.n. in 1861 (holo- type, Bf; isotype, P!). Monogr. Gard. 1993. 650 Annals eagle aed Garden Tachigali species, but is Discussion. melinonii is an incon- spicuous easily recognized by its sparse, stellate indument. No other species o Tachigali has stellate hairs. Stipules of this species Petals are have not been seen. Domatia are lacking. slightly longer than the sepals, glabrous, linear « slightly dilated, but fall off readily and are thus often lacking on herbarium specimens. The base of the leaflets is asymmetrical. Habitat data, when given by collectors, indicate this is a species of terra firme forest. It has been reported from Suriname, French Guiana, Brazil (Amapá, Amazonas). and Peru (Loreto). but is uncommon in herbaria. Selected y lil examined. BRAZIL. Amapá: Mun. lva 5525 (INPA). Mazagão, Silva 25 (INP Amazonas: Reserva 1501. Projeto Fragmentos > E lore s ri et al. 20017 (INPA MO). FRENCH GUIANA. Crique Passoura, Région d 3793 (NY). PERL 50370 (MO). Sabatier & Prévost Loreto: Jenaro Herrera, Gentry et al. Kourou, O. Williams) Zarucehi Syst. Bot. Missouri Bol. 34. Tachigali mierantha (L. & Herend., Monogr. Gard. 45: 1254. 1993. Basionym: Sclerolobium micranthum Williams, Fieldiana, Bot. 31 32. 1965. TYPE: Peru. Huánuco: Gutierrez 44 (holotype, F not seen; isotypes, K! USE, WIS not seen). Discussion. This species is easily recognized by the combination of glabrous sepals, small, obovate petals shorter than the sepals, and the rather dense, erect indument on the lower surface of the leaflets. Domatia are lacking: stipules are laciniate, to 2 em, and the obtuse to subcordate and The clearly pedicellate, with the pedicels about as long as base of the leaflets is somewhat asymmetrical. flowers are small and the flowers. For this study, only the K and US isotypes and the fruiting paratype (Gutierrez 96) have been seen. 35. Tachigali micropetala (Ducke) Zarucchi & Pipoly, Sida 16: 787. 1995. Basionym: Selerolo- bium | micropetalum Ducke, Bol. Inst. N. 2: 20. 1944. TYPE: Brazil. Manaus, "silva terris altis argillosis ultra Flores. Ducke 1219 (holotype, TAN not seen: isotypes. O! US». Teen. Agron. Amazonas: Discussion. This species is easily recognized by the combination of small flowers and the distally glabrous sepals. Petals are elliptic; to | mm. Stamens are monomorphic. Stipules have not been seen on any of the flowering specimens studied, but are present on two sterile specimens (Ducke s.n. [RB 20343], Evans & Peckham 2941); they are densely pubescent and pectinate. Domatia are lacking. Leaflets near the base whereas the The of the leaf have an asymmetrical base, distal leaflets have a more symmetrical base. f ihe leaflets are inrolled. Tachigali micropetala is known from the area near margins ( frequently Manaus. from Pará in Brazil, Guyana, French Guiana, Suriname, and Venezuela. Selected examined. BRAZIL. Estrada Manaus—Hacoatiara, Km 13: 9510 (INP A). Miguel, Silva . Limonade. M specimens Amazonas: GUY ANA. rs “Greek, Mapanagebied, Delta Amacuro: Marcano Berti 655 (MO). Fanshawe 7703 (GH, NY). norestes de El Palmar, 36. Tachigali odoratissima (Benth.) Zarucchi & Herend.. Fl. 1998. Basionym: Sclerolobium odoratissimum Benth. in Mart.. Fl. Bras. 15(2): 48. TYPE: Brazil. \mazonas: Spruce 3057 (holotype, K!: isotype. T Venez. Guayana 4: 116 Discussion. Tachigali odoratissima is best recognized by the narrow leaflets with à symmetrical base and glabrous or sparsely appressed pubescence on the lower surface. Stipules are l- to 3-foliose and ca. Lem. Petals are linear and densely pubescent. Stamens are equal. Domatia are terete and generally 24 cm. Nearly all collections seen in this study are from flooded forests or were made along rivers. It is a common species in Brazil (Amazonas, along the Rio Negro and tributaries: one collection is known from the Rio Japura) and Venezuela (Amazonas, along the Río Negro and the Río Orinoco and their tributaries). There are a few records from Colombia (Guainía and Vichada). Amazonas: COLOMBIA. ina, en el Río Inirida, Espina 338 (COL). Vichada: Puerto Nuevo, Daniel 127 (COL). VENEZUELA. Amazonas: lower part of the Río Daria; Davidse 27598 (MO, N Y). examined. BRAZIL. Duarte 7252 (F. HB, INPA, US). Cano Mi Selected specimens Mura elos. Guainía: . Hist. PL Guiane |: Aublet s.n. 37. Tachigali paniculat 942. 1775; TYPE: (holotype, BM!). French Guiana. Taupa. irigona Aubl., Hist. Pl. Guiane |: 115 its . m Nat. 4: 163. 1844. s desig- French Guiana. ud sericea TU \rch. Mu TY n N. "Ph duplicate, L! , . Mus. Hist. Nat. A magnas Ega, Poeppig 273 x1. 184 Tachigali seni Tul., : TYP 7 (holotype, E: Brazil. P!; isotype, L!). Tachigali maet Miq.. Stirp. Surinam. Select. 13. 1851. Tacl oie. E Dwyer, Gard. 41: 1954. ^E: 1951 holotype ! nol seen). mi eror Huber, Bol. Mus. 388. Sy YPE ugali paniculata var. a Missouri Bot. = riname. Kappler 5: n. nov. silvis ripariis p Hi Mapuera." Ducke s.n. (MG 8905) | (holotype. MG not seen: isotype. BM!). Volume 95, Number 4 2008 van der Werff Synopsis of Tachigali Berlin 6: Amazonas: Rio Notizbl. Kónigl. Bot. Gart. Syn. nov. TYPE: Brazil. Sao E Ule 6042 (holotype. Bt: isotypes. MG Tachigali ulei Harms, 306. 1915 cea El 1 u a lost Bot. e Syst. Beibl. 72: 20. 19023. S . TYPE: Venezuela. o. '0: ashy 12 10loty pe Br es GH!, MO!). Tachigali min ro Dwyer, Ann. Missouri Bot. Gard. 41: 249. TYPE La 54. Syn. nov. TYPE: Venezuela. Bolívar: Alto Río ^uragu di 1180 (holotype, US!) Tachigali paniculata var. comosa Dwyer, Ann. Missouri Bot. Ll: 240 1954. Syn. nov. TYPE: Brazi zonas: S Fanlo de Olivença, Arukoff 665 CH MO! isotype, GH?). Discussion. Here, Tachigali paniculata is accepted as a wide-ranging and variable species, known from the Guianas to Peru. It is also the most frequently collected species of Tachigali. Domatia are mostly absent, but sometimes present and semiterete. Of 16 collections checked from the Iquitos area, all from flooded forest, two had well-developed domatia, two had some duplicates with domatia and others without, two had the petiole bases not thickened (but ant holes were present in the petiole). and the rest had no traces. of domatia or ant habitation. Therefore, presence or absence of domatia is not considered to be a diagnostic character of T. paniculata. Stamens with and are dimorphic, three stamens shorter and i faleately curved, this character separates 7. paniculata from T. alba, which has monomorphic stamens. Tachigali alba and T. richardiana, another similar species, also differ in their terete leaf rachises; these are sharply angular in T. paniculata. Stipules, when present, are foliose, with one or The asymmetrical. Leaflets are often glabrous or nearly two pairs of lobes. base of the leaflets is so, but sometimes may be moderately to densely appressed pubescent on the lower surface. Nearly all flooded forest, and most trees 0 m). However, there is so much morphological variation collections come from are less than 15 m tall (occasionally reaching 2 in Tachigali paniculata (as recognized in this study) that additional analysis may reveal further data sufficient to segregate more taxa from among the current synonymy. Of the two syntypes of Tachigali sericea, Poeppig 3039 has flowers and is selected as lectotype, while the specimen from the Richard herbarium (French Guiana. Herb. Richard s.n. |P]) has fruits. The latter has a denser indument on the leaflets. Tachigali paniculata in grandiflora differs from typical T. and several qualitative quantitative characters. The leaflets are thinner, have an acumi- nate tip, and have the secondary and tertiary venation raised on the lower surface. The flowers are rather arge, larger than average for 7. paniculata. However, other collections placed T. paniculata can have some of these characters, although not all on the same specimen. To date, not a single specimen has been seen that matches the type of T. 21 therefore, has been included (with some hesitation) as a i synonym of T. paniculata. Ule 6042 was first identified by Harms (1907) as Tachigali paniculata and in 1917 was promoted, without comment, to the type of T. ulei. In this study, Harms's first identification is accepted, and thus T. ulei is considered a synonym of T. paniculata. Tachigali rusbyi Harms was said to differ from 7. paniculata in its smaller flowers. The flowers certainly do not differ in size by enough to warrant recognition of a separate species. Stamens of T. rusbyi are less most collections of T. clearly dimorphic than in paniculata, but they are not equal either. Dwyer (1954) emphasized the sharply angled petioles of T. rusbyi, but specimens with such petioles occur throughout the range of T. paniculata. Tachigali pulchra was recognized solely based on its long-stipitate ovary. Dwyer's description ("stipes ovarii 3-8 mm longus... ovario 1-2 mm longo. .. fruc- tus non visus”) suggests that the ovaries are long- stipitate in the flowering stage. The holotype does not show such long-stipitate ovaries in the flowers, but does have some young fruits (to 3 cm) and these are e] indeed rather long-stipitate. According to Dwyer, ii other characters 7. pulchra was indistinguishable from T. rusbyi (the types of both species were collected in the Orinoco drainage in Venezuela), and it seems best to place T. pulchra as a synonym of T. paniculata. Most collections of Tachigali paniculata are from medium-sized trees in flooded forest and range from Amazonian Brazil, to. Vene- the Guianas, throughout zuela and Peru. examined. BOLIVIA. Pando: Río Abuna, Gentry & Perry 77998 (MO). Selected Negro near je t. with I a aes — ^d Ni Waz Uu = = GY Rond ) "de wer Río Zarucchi 2151 (MO, US) FRENCH GUIANA. Riviere Grand Inini, de Granville y (HB). GUYANA. Cuyuni River, Akarabice Creek, | 440 (US). PERU. Loreto: Maynas, Caserio Casococ us ou. 1527 (MO, US). SURINAME. Lower slope s of Juliana Top, Irwin et al. 54827 (F. US). VENEZUEI Amazonas: Alto Río Orinoco betw. Tama-Tama and E ea la, Maguire et al. 41550 (F, NY). olívar: Alto Río Paragua, Cardona 1182 (US). Delta Amacuro: Río ca uro, upstream from San Victor, Steyer- mark 87233 (F, US Kubitu 38. Tachigali paraensis (Huber) Barneby, Brittonia 48: 182. 1996. Basionym: Sclerolobium paraense 652 Annals of the Missouri Botanical Garden Paraense Hist. Nat. 6: 79. Pará: Huber, Bol. 1910. TYPE: ad viam ferream inter Mus. Brazil. “in silvis paraensibus Bragança,” (MG 9642) BM!, Fl, K!, capitalem: et Rodolpho Siquiera Rodrigues s.n. MG not seen: pali ISOLYpes, LOL Se e pud Benoist, Bull. Soc. Bot. France 66: 19 Tachigali albiflora (Benoist) Zaruechi & oe e l. TYPE: French Guiana. Benoist 1074 (holotype, P!). enez, Guayana d; 115, 1998 Discussion. This species is best recognized by the leaflets with a cuneate, more or less symmetrical base, linear, densely white pubescent petals, and large stipules. Occasionally, small, strongly revolute bracts are present in the inflorescences. Domatia are absent. Stamens are monomorphic. This species is very similar to Tachigali melanocarpa, and differences between the two are discussed under that species. There has been some confusion about the type specimens. On the label of the MG holotype (photo at F), the collector is given as Rodolfo S. Rodrigues and the type locality is listed as “Estação agr. Peixe Boi.” Isotypes distributed by RB (no. 5620), with the type in type locality as Peixe Boi, and R. Siqueira as collector. — ormation retyped, gives the These labels (on specimens in F, K, and US) also give the MG herbarium number (9642). MO has a sheet with handwritten RB 5020 and citing Rodolpho Siq. Rodrigues as an older, abel annotated. with the number collector. The type locality is still given as Peixe Boi. Dwyer, his revision of Sclerolobium (1957), stated that he had not seen the type of Sclerolobium paraense, but cited among the specimens he had seen “Estação de Peixe Boi, Rodrigues 9642” (MO) and “Peixe Boi, R.R. between Belem do Para € Braganca. Sigueira Herb. no. 5620” (K, NY, US), all of which I consider isotypes. Distribution. Tachigali paraensis is known from Brazil, and eastern. Venezuela (Bolívar, the states of Maranhào and Pará in French Guiana, Suriname, Delta Amacuro). Selected specimens examined. BRAZIL. Maranhão: Mun. Caruta pera, Balee & Ribeiro 2692 (MO). Para: $n 93 Rodovia Belem—Brasilia, Kuhlmann & Jimbo 42 (LA MG. NY, SP) FRENCH GUIANA. Saúl: La fae Mort & Hartley 18160 (MO). SURINAME. i . Elburg LBB 9462 (MO). VENEZUELA. Ba ll Dor: ado, en el drenaje del Río Cuyuni, Steyermark & uen 104447 (MO, NY, US). Delta Amacuro: este-noreste de El | Berti 216 (INPA, MBM. MO, NY " Mountain. Trail, almar, Marcano 39. Tachigali peruviana (Dwyer) Zarucchi & Herend., Monogr. Syst. Bot. Missouri Bot. Gard. 45: 1254. peruvianum Dwyer, Lloydia 20: 94. 1957. 1993. Sclerolobium paniculatum var. TYPE: San Martín: Alto Río Huallaga. Williams 5701 (holotype, F not seen). Peru. Discussion. Tachigali peruviana can be recognized by the combination of the sericeous lower surface of the leaflets, the linear, glabrous petals, pedicellate flowers, asymmetrical base of the leaflets, and lack of domatia. Stipules are caducous, and none were seen in this study. It differs from T. vulgaris L. F. Gomes da Silva & H. C. Lima in the dense, sericeous indument of the lower surface of the leaflets. Tachigali subvelutina (Benth.) semi-erect indument on the leaflets and twigs. Oliveiro-Filho is very similar: it has a denser, This is rachis of E peruviana. Two collections from Humaitá, Amazonas (Teixeira et al. 199 & 912), are provisionally placed leaflets, best seen on the inflorescence axes and leaves, which are strictly appressed pubescent it here; they have larger and wider and the flower-bearing branches of the inflorescences are much longer and flexuous. The type of T. peruviana has not been seen in this study. Tachigali subvelutina has a distinct geographical distribution, a different indu- ment, and occurs in cerrado vegetation and, therefore, should not be confused with 7. peruviana. However, the differences between the two species are otherwise weak BOLIVIA. Beni: Km 29 of Maas et al. 6675 (MO acional Noel Kempff, Guillen & Selected specimens examined. rd. from Riberalta to Guayaramirim, Parque N Killeen 2710 (MO. NY) BRAZIL. Amazonas: Mun. Humaitá, Teixeira et al. 912 (INPA. MEL. MG. MO). Rondónia: Km 5 da rodovia Vilhena-Porto Velho. Silva & Pinhetro PERU. Junín: Chun- 4112 (INPA, MG, NY). chuyacu, Wovtkowski 5578 (MO, NY). 40. Tachigali physopho ea (Huber) Zarucchi « Guayana 4: 116. 1998 Sclerolobium | physophorum Huber, Hist. Nat. 6: 80. 1910. Herend., Fl. Venez. Basionym: Bol. Mus. Paraense TYPE: Brazil. inferioris," Ducke s.n. Amazonas: "ad ripas fl. Japura (MG 6751) (holotype, MG not seen: isotype, BM!). cylindrical | domatia Discussion. The swollen. the base of the leaf petiole are char- Very ribbed. close do acteristic in this species. frequently the domatia are longitudinally The leaflets are strongly unequal at the base and are sparsely to moderately appressed pubescent on the lower surface. Stipules are usually persistent, l- to 3-foliose, and 1— 2 cm. monomorphic. The Petals are linear and glabrous. Stamens are hypanthium is symmetrical. Tachigali physophora is a species of flooded forests. Most collections are from Brazil (Amazonas) with a few collections from adjacent Colombia (Guainía) and Venezuela (Amazonas). Volume 95, Number 4 2008 van der Werff 653 Synopsis of Tachigali Selected pura examined. BRAZIL. ros Anavilhanas Island, et al. 21276 (INPA, Y). COLOMBIA. a Río Inirida, Mur a 20809 Amazonas: on bank of Rio Berry 1604 (MO). (COL, F, Manaplare ie VENEZUELA San Juan & Morrocoy, Cowan) Zar- 120. 1998. Basionym: Sclerolobium MM h. Cowan, Mem. New York Bot. Gard. 10: 85. 1961. TYPE: Venezuela. Amazonas: Río NEM along Cano Pimichin, Maguire et al. 41830 (holotype. US not seen; isotypes, GH! MO!, P!). 41. Tachigali pimichinensis (R. S. ucchi & Herend., Fl. Venez. Guayana 4: Distinctive this Discussion. species are the obtuse to rounded tips of the leaflets. Other characters include the sericeous indument on the lower surface of the leaflets, the occasional presence of small, entire, revolute bracts along the inflorescences, the absence of domatia, and the linear, pilose petals. The hypan- thium is symmetrical. Stamens are monomorphic. The leaflets have a symmetrical base. Tachigali pimichinensis is only known from the type collection. 42. Tachigali plumbea Ducke, Bol. Teen. Inst. Agron. N. 2: 15. 1944. TYPE: Brazil. Ducke 817 (lectotype, Amazonas: Manaus, A. designated here, NY not seen; duplicates, IAN! RD. EPI r YPE: Ducke s.n. (RB 24289) (epitype. designated here, US!; duplicates, INPA!, K! PH. Discussion. Differences between this species and the similar Tachigali cavipes are discussed under the alter. Tachigali plumbea has sericeous lower surfaces of the leaflets, which have an asymmetrical base, semiterete domatia, dimorphic stamens, and foliose stipules that are often lacking. This species has been reported from the lower Rio Negro and the Rio Madeira. Ducke description of Tachigali plumbea, four flowering and 1944) included five collections in his one sterile. Two collections came from Porto Velho. the other three from around Manaus. Dwyer (1951) cited Ducke 817, the sterile collection, as type. He listed two specimens of this collection al NY and US. The lectotype should be chosen from those two, the NY the lectotype. The online image shows as nice a sterile and specimen is here selected as specimen as one can wish for. The specimens from Porto Velho (Ducke 228 according to the herbarium sheets, Ducke 288 according to the protologue, and Ducke s.n. (RB 55419)) differ in minor aspects from the specimens collected around Manaus (Ducke 817, Ducke 818, Ducke s.n. (RB. 24289)). (1954) excluded the Porto Velho specimens from 7. and dumbea and placed them in T. cavipes. In this Į | ! treatment. all five collections are included in T. plumbea, while acknowledging that further research Dwyer Given (RB may well show that two species are involved, the US sheet of Ducke s. ~ this wa 24289), from the same locality as the lectotype. is cn here as epilype. It is a good flowering specimen, displaying domatia, stipules, and both surfaces of the leaflets Selected specimen examined. BRAZIL. Amazonas: Re- MO serva Florestal Ducke, Martins & Assunção 91 (MO). 13. Tachigali poeppigiana Tul., Arch. Mus. Hist. Nat. 4: 168. 1844. TYPE: Brazil. Amazonas: Ega, Poeppig 2657 (holotype. P!; isotype, L!). Nov. \mazonas: Ega. Poeppig P! Gen. Sp. Pl. 3: 60. 1845. 2837 (holots pe. Tuc higali polyphylla Poepp.. TYPE: Brazil. W nol seen: isotypes. El Distinctive in this species are the leaflets (often Discussion. IO. pairs) and the many pairs of appressed, grey-brown hairs on the calyx. The hypanthium is asymmetrical. Petals are 2-3 X ca. 1.5 mm. Stamens are monomorphic. Domatia are present but may be slender. Stipules are foliose with one pair of lobes, caducous. and occasionally present in sterile specimens. Bracts are sometimes present in the inflorescences and resemble the stipules but are smaller. The base of the leaflets is strongly asymmetrical. This species can be confused with Tachigali schultesiana, but the latter differs in the characters mentioned in the key. Dwyer (1954) used the name 7. polyphylla for this species. but the name T. poeppigiana has priority. Tachigali poeppigiana is known from a few collections in Amazonas and one from Rondônia. Brazil. Most collections, but not all, come from riverbanks flooded forest. Poeppig (1845) was not aware of Tulasne’s publication (Tulasne, 1844). 72837 as an undescribed Both Poeppig and Tulasne recognized Poeppi pig species, and both described it independently with the same collection as type (but with different holotypes). Selected specimens e. xamined. BRAZIL. Amazonas: Humaitá. Arukoff 7233 (GH. MO, US). eas a 13 km de Vilhena, Vieira et al. 914 (INPA, MO, US Irwin & Arroyo) L. Rodriguésia 53: 44. Tachigali prancei pe Gomes da Silva & H. € 400. 2007. Basionym: Sclerolobium Miis H. š rittonia 26: 208. TYPE: Rondônia: Serra dos Tres AN N bank -—- Lima, Irwin & Arroyo, Brazil. of the Rio Madeira. Prance et al. 5524 (holotype, NY not seen; isotypes, COLL FL GH! INPA!, MG!, RI, RB not seen). Discussion. — Tachigali prancei is easily recognized by its large, oblong to ovate-oblong, tomentulose leaf- 654 Annals of the Missouri Botanical Garden lets with a rounded base, tomentulose inflore- scences, and linear, pilose petals. Domatia are lacking. Stipules are foliose, flat, ca. 2.5 cm, but rarely present. The apex of the leaflets tapers to an acute or acuminate tip. Floral bracts are about as long as. or shorter than, the buds. Three collections of this species have been seen, all from Rondônia. Tachigali prancei can be confused with 7. chrysaloides, which differs in having an appressed indument on twigs and leaf rachis and in its large, revolute stipules. \dditional DE examined. BRAZIL. Rondônia: Mun. Jaru. Cid et a 987 (MEL. MG. MO): Mun. \riquemes, Cid et al. 5058 (MO, NY) 15. Tachigali ptyehophysea Benth.. FI. Bras. 15: 220. 1870. TYPE: Brazil. Amazonas: "prope Panure ad Rio Uaupes.” Spruce 2644 (holotype, K!; isotypes. GH! P^). Discussion. Although the leaflets of this species are nol particularly small (to 11-12 em long in the GH isotype, to 16 em long fide Dwyer), they have only two or three pairs of lateral veins and this character alone Domatia are present and of the Bracteoles at the allows easy identification. cylindrical. Stamens are dimorphic. The base leaflets is symmetrical or nearly so. base of the flowers are ca. 3 mm wide, wider than in most other species. Stipules, when present, are foliose with one pair of lobes; lobes are to 1 X 0.6 em. The few flowering collections seen in this study all have the lower surface of the leaflets densely appressed (somewhat sericeous) and pubescent. Tachigali plychophysea is known from the upper Rio Negro and This along the Rio Vaupes in Brazil and Colombia. species is similar to T lorelensis; differences are discussed under the latter. examined. BRAZIL. Negro, Pires de Amazonas-Vaupes: Rio Apaporis, Schultes & Cabrera 13488 (COL, F., Selected Amazonas: Taracua, bacia do alto Rio (IAN). COLOMBIA. - x GH, specimens Jirijirimo, 46. Tachigali pubiflora Benth., J. Bot. (Hooker) 2: 94. 1840. TYPE: Schomburgh 43 (holotype. KI; isolypes, GH!, L!, P). Guyana. Discussion. This species is best recognized by the combination of densely pubescent leaflets, absence of domatia, dense, erect indument on twigs and inflorescences, oblique ay pana narrowly foliose stipules, and dimorphic stamens. The base of leaflets is symmetrical to slightly asymmetrical. The species can be confused with Tachigali rigida, which differs in the presence of domatia. Tachigali pubiflora is known from Guyana and adjacent Venezuela. Most indument on the lower have a dense leaflets: specimens surface of the however, an occasional specimen (for instance, Clarke 277 from Guyana) has a sparse indument. ~ — — Selected specimens examined, Kaieteur Pla- teau. fore - along Potaro River, Cowan & ia 2089 (F, US). VENEZUELA. Bolívar: ( "ede nto Ucaima, sureste de Canaima. Dew 106423 (US) Arch. Mus. French Hist. Guiana. 47. Tachigali richardiana Tul., Nat. 4: 166. 1844. TYPE: Richard Herbarium s.n. (holotype. P!). Mus. Hist. Nat. (Paris) 31: : French € 1), Missouri Bot. Gard. 41: Guiana. Martin s.n. Tae hr posent Be noist, es . 192 . yn. nov. Guiana. UE - uode 1574 cen Tachigali bracteolata Dwyer. Ann. 233. 1954. Syn. nov. TYPE: French (holotype. F). Discussion. — Tachigali — richardiana is best recognized by the combination of terete leaf rachises, dimorphic stamens, a basal patch of hairs on the outer surface of the petals, and relatively few lateral veins (4 It is closely related to 7. alba, to 7 pairs per leaflet). from which it differs in the characters given above and in the slightly larger petals (5-7 mm long vs. 4-5 mm long in T. alba). Domatia are lacking, and stipules have not been seen on any of the collections. Occasionally, a collection of T. richardiana may not have all the characters mentioned above. For instance, Sabatier 987 has five pairs of lateral veins and dimorphic stamens, but the petals lack the hairs on the outer surface. Tachigali barnebyt is also similar, but differs from T. richardiana in its persistent stipules and the gland- dotted richardiana has been reported from terra firme forests lower surface of the leaflets. Tachigali French Guiana and Brazil. Benoist considered sulcata different from T. richardiana in its slightly denser indument and a different shape of bracts. Dwyer 1954) published T. bracteolata as a distinct species based on its long bracts. Here, these species are treated as synonyms of T. richardiana. Imens lected spe vamined. BRAZIL. Pará: Estrada do M 3075 (VAN). FRENCH GUIANA. RN2 Cayenne-=Régina, pk. 79 Camp Hervo, Sabatier & Prevost 5050 (MO) eds. 48. Tachigali rigida Ducke, Arq. Inst. Biol. Veg. 4(1): 12. 1938. TYPE: Brazil. Amazonas: Igarape Macacuny prope Cucuhy, Ducke s.n. (RB 35423) holotype, RB not seen; isotype, K!). Mur rigida var. argentata Ducke, Arq. Inst. Biol. Veg. 12. 1938. TYPE: Brazil. Amazonas: Rio Curicuriari, un Rio Negro. Dun s.n. (RB 35422) (holotype, RB!) Discussion. This species is easily recognized by the combination of the densely appressed pubescent the symmetrical base of lower surface of the leaflets, Volume 95, Number 4 2008 van der Werff Synopsis of Tachigali 655 the leaflets, the presence of domatia, and the large flowers with dimorphic stamens. The hypanthium is oblique. Stipules are foliose, with one pair of lateral lobes. The characters used by Ducke to recognize variely argentata (absence of reddish hairs on lower surface of leaflets, absence of hairs on the upper surface of mature leaflets) are variable, and the variety is here included in synonymy. Tachigali rigida resembles T. pubiflora; the two species can be distinguished by the presence of domatia in T. rigida and their absence T. pubiflora. Tachigali rigida is also similar to 7. grandistipulata, but the latter has smaller flowers with monomorphic stamens, these about as long as the petals. Tachigali rigida has stamens much longer than the petals. Tachigali rigida — is known from the upper Rio Negro in Brazil, from one collection from Vaupes, Colombia, and from Amazonas in Venezuela, frequently growing along rivers in seasonally inundated forest. BRAZIL. Selected specimens examined. Amazonas: en- 1649 tre Camanaus e Curicuriari. Silva et al. 16 (INPA). COLOMBIA. Vaupes: at confluence. of Ríos Guainia and Casiquiare, Schultes & Lopez 9389 (V. GH, TAN, US). VENEZUELA. 'acimoni and Yatua, US). Amazonas: frequent along Ríos Maguire et al. 41578 (GH, IAN, MO, — 49. Tachigali schultesiana Dwyer, Bot. Mus. Leafl. 8: 152. 1958. TYPE: Colombia. Amazonas— Vaupes: Río Apaporis, Schultes & Cabrera 14045 (holotype, MO!; isotypes, GH!, US!). Discussion. Characteristic in this species are the many pairs of leaflets (up to 15 pairs), the dense, puberulous, brown indument on inflorescences and flowers, and the nearly symmetrical flowers. Stamens are monomorphic, stipules are foliose but rarely present, and domatia are present and terete. Petals 4 X 2 mm. asymmetrical. The are elliptic, ca The base of the leaflets 1s strongly type has flowers and young fruits; the young fruits (15 mm long) are seated ud on a slender stipe, 5 mm long. Tachigali poeppigiana is very similar, but can be distinguished by mentioned in the key. Tachigali differences schultesiana is known from Colombia (Vaupes, Caqueta, Amazonas) and Peru (Loreto). Selected specimens examined. COLOMBIA. ae Vaupes: Río Apaporis, Garcia-Barriga Pe (E US). Caqueta: Araracuara, Care "nas et ui 442 Pu. ea Loreto: Jenaro Herrera, Revilla 1220 (MG. MO). Zaruechi & Missouri Bot. Gard. 50. Tachigali setifera (Ducke) Herend., Monogr. Syst. Bot. 45: 1254. 1993. Basionym: Sclerolobium ferum Ducke, Arq. Inst. Biol. Veg. 2: 42 TYPE: Brazil. coliniae Campos Salles seti- 1935. “silva non inundabilis Ducke Amazonas: prope Manaos,” sn. (RB 23529) (lectotype, designated by Zarucchi & Herendeen Brako € Zarucchi, 1993: 1254, US!; duplicate, P!). Discussion. The type has a characteristic, rather dense indument on the lower surface of the leaflets, hairs that radiate consisting of appressed, whitish from a central point in various directions. This indument has not been found on any other species. Stipules are 3-foliose, with the segments revolute and are up to 3 em long on young shoots. Domatia are lacking. Petals are linear and pilose. The base of the obtuse to subcordate and is leaflets varies from symmetrical. Relationships are with a group of species including Tachigali rugosa, T. prancei, T. and 7. setifera in indument type. vasquezii, chrysaloides. These species all differ from T. A number of collections that, in all other characters, match Tachigali setifera do not have the indument of short, appressed, white hairs. Instead, the lower surface of the leaflets is glabrous or has Both forms have a similar geographical distribution: the scattered, short, erect, hairs or a brown scurf. form with the appressed, white hairs is known from Bolivia, Peru, Ecuador, and Brazil (Amazonas, Mato Rondônia) and one collection from Grosso, and Colombia, and the form lacking these hairs is known from Bolivia, Peru, and Brazil (Acre, Amazonas, and Rondônia) and single collections from Ecuador and Colombia. Apparently, the two forms do not occur together. Eight collections from the Tambopata area in Peru all have the short, white indument, while eight collections from the Bosque Nacional Alexan- der von Humboldt, Ucayali, Peru, all lack the short, Both largely because of their overlapping distributions and white hairs. forms are placed in T. setifera great similarity in all characters except the indument on the lower surface of the leaflets. In the species identification list, specimens with short, white hairs are identified as T. while specimens T . setifera s.l. setifera s. str., lacking these hairs are identified as o setifera s. Selected specimens examined of Tachigali str. BOLIVIA. Pando: Manuripi, cerca Puerto America, Jardim 1117 (MO, NY). BRAZIL. Amazonas: Mun. Borba Zarucchi 2857 (INPA, MG). Rondônia: yank of Rio Madeira, 2 km below mouth of Rio Abuna, Prance et al. 6037 , US E SUADOR Foster 9295 (MO). . Lopez oil camp. p. el al. 18951 ( Mi )). Madre de Dios: Nature Reserve, Gentry & EUG 57661 (MO). Selected specimens ‘achigali setifera s.. BO- LIVIA. i Diez, Boome 1394 (MO, NY). Santa Cruz: Parque Nacional Noel K lk et al. 644 (MO). BRAZI : Aluisio 28 (INP A). Mato Grosso: margen esquerda de Rio Juruena, Rosa & Santos 2074 (MO, NY). Rondénia: Mun. Tambopata < Beni: Prov. Vace Kempff, > = mi = z N = © = = m ts Y = ^ jJ: Edo 656 Annals of the Missouri Botanical Garden Costa Marques, Cid Ferreira 8736 (INPA, MEL, MO). to 5 pairs of leaflets; leaf petiole with a swollen de Taisha, Amazonas: Quebrada chichijam ECUADOR. Morona-Santiago: 560 km SE Ortega U. 214 (US). PERU. entsa, Ancuash — (MO). Huánaco: (US). ied de Dios: Km 15 de « Ruiz C. 27 (MO). U cayali: one el Sacramento, Clones R 43 (F, NY, US Tingo Maria, Burgos carretera lIberia-Inapari. 51. Tachigali tinctoria (Benth.) Zarucchi & Herend., Monogr. Syst. Bot. Missouri Bot. Gard. 45: 1254. 1993. Basionym: Sclerolobium tinctor- ium Benth., Hooker's J. Bot. 2: 236. 1850. TYPE: Spruce s.n. (holotype, K!; isotype, P!). Kew Garden Misc. Brazil. Para: prope Caripi, Sclerolobium | reticulosum Dwyer, Lloydia 20: 98. e Tachigali reticulosa s r) Zaruechi & He d Venez. Guayana 4: 120. 1998. Syn. nov. T . Br e Amazonas: E 7188 a MOL oe m isotypes, BM!, GH!, Discussion. This species is best recognized by the absence of domatia and the sparsely appressed pubescent lo elabrous, oblong leaflets with an asymmetrical base. Petals are linear and glabrous. Supules, when present, are pectinate and appressed pubescent. It is very similar to Tachigali vulgaris, the difference being the obviously pedicellate flowers in T. vulgaris versus sessile or very shortly Most collections P: pedicellate flowers in T. tinctoria. of this species are from the Brazilian. state of Pará, with a few collections from Amazonas and Venezuela. specimens placed by No differences between the Dwyer (1957a) his Sclerolobium reticulosum and specimens of T. tinctoria have been found in this study. Sclerolobium reticulosum is consequently placed in synonymy under 7. tinctoria. Selected specimens examined. BRAZIL. Amazonas: Rio Negro, Padauiry, Froes 22718 (COL, US). Pará: Mun. Almeirim, Mt. Dourado, Pires & Silva 1376 (INPA, MG, NY). Rondónia: Mun. € da » arques, 123 km de Costa marques, Cid Ferreira 8072 (INPA, MEL, MO). e "aima: Ducke 1595 (V, Ne NY). VENEZ Amazonas: alrededores de Puerto Ayacucho, e n» 6 (NY, US). Caracaral, 52. Tachigali vaupesiana van der Werff, sp. nov. TYPE: Colombia. Vaupes: Abiu e. E Mitu, 20 Nov. 1945, P. Allen 3366 (holotype, MO!; isotypes, GH!, NY!). Figure 13. Sclerolobium odoratissimum var. latifolium Dwyer, Lloydia Amazonas: Sáo Paulo de MO; 1957. TYPE: Brazil. Olivença, Krukoff 8988 (holotype, CH!, K!, US). isotypes, Fl, Tachigali Ru iod (Poepp.) Zaruechi & Herend. similis, sed ab ea foliolis basi aequilateralibus et stipulis li integris margine re E (nec pec tinalis) recedit. Tree, to 35 m; twigs minutely appressed pubescent, becoming glabrous with age. Leaves 15-25 cm, with 2 domatium 1.5-2 em from the base, this 2.5—4 cm, the first pair of leaflets attached just above the domatium: leaf stalk finely appressed, grey pubescent; leaflets 7-15 X 3-6 cm, nearly so, the tip acute, the upper Pa: e deja the base obtuse, symmetri- rn lower surface densely appressed, seri- ceous, pubescent, petiolules 3-5 mm, with a Jue indument as the leaf stalk; stipules small, 6-9 mm, with a short (to 3 mm long) stalk, entire, the margin revolute, infrequently present; bracts similar to the stipules occur in the inflorescences. Inflorescences to 25 cm, paniculate, finely and densely appressed pubescent, the flower-bearing branchlets slender, flexuous, to 12 em long. Flowers with a short (ca. l mm) pedicel, densely grey pubescent; hypanthium symmetrical, l mm; sepals elliptic, ca. 2 mm; petals ca. 2 mm, linear, densely yellow pubescent; stamens monomorphic, ca. 2.5 mm, with some yellow hairs near the base, the distal 2/3 glabrous; ovary ca. 2 mm, moderately pubescent with stiff, brown hairs, centrally attached in the hypanthium. Fruits not seen. Discussion. Tachigali vaupesiana is only known from three collections: the type, a second collection from São Paulo de Olivença, and a third from São Pedro on the Rio Vaupes, both in Amazonas, Brazil. The type specimens had been previously annotated as T. chrysophylla or Sclerolobium alf. chrysophyllum. Tachigali vaupesiana differs notably from that species in the symmetrical base of the leaflets and in its small, enlire stipules with a revolute margin. The stipules are bracts are found in the inconspicuous; similar inflorescences. They are present in all four duplicates of the type collection. Another difference between the new species and T. chrysophylla is in the position of the domatia: in T. vaupesiana the domatia leaflets, chrysophylla the domatia often occur distal to the first 7 occur below the first pair. of while in iir of leaflets. Similar small, entire bracts are also — present in T. eriopetala, and this species is here considered to be very similar (leaf shape, indument) to e vaupesiana, the main difference being the absence T are from the Manaus Thus, T. eriopetala is here considered as the closest relative of domatia in T. eriopetala. All collections of id seen in this study region or from Pará and none have domatia. T. vaupesiana, and T. vaupesiana is recognized as a distinct species based on the presence of domatia and odoratissimum var. its distribution. Selerolobium latifolium is placed synonymy of T. vaupesiana. lts type, Krukoff 8988, agrees well in leaf and indument characters with T. vaupesiana, and it has the same domatia. Stipules are lacking in the two duplicates of Krukoff 8988 seen in this study; this is probably due to Volume 95, Number 4 van der Werff 2008 Synopsis of Tachigali { DET ; E 2030451 E M E / A aar" ERU a Y MISSOURI BOTANICAL GARDEN 10 copyright reserved FABACEAE Tachigali a Det. D.A. Neill, 1997 Missouri Botanical Garden Herbarium (MO) PLANTS OF COLOMBIA No. 3356 m November ; ) / EaP. Sclerolobium «€ Chneo hy fens Doom + f du Tree, : we! yellow. mmon : Tack 9 ES Determined by H RAPES A n A 4 Wee ff T g Missouri Botanical Gatden = Collected by PAUL H, ALLEN and distributed by THE MISSOURI BOTANICAL GARDEN Figure 13. Tachigali vaupesiana (holotype). 658 Annals of the Missouri Botanical Garden these specimens being in young fruit by which time the stipules have commonly fallen off. Because of the importance of floral characters, a flowering collection is chosen as the type of T. vaupesiana. Sclerolobium odoratissimum var. latifolium is not raised to species rank: otherwise, a collection with young fruits would be the type collection. IUCN Red List category. only from three This species is known collections and has been given the listing of Data Deficient (DD) (IUCN, 2001). Paratype. BRAZIL. Amazonas: Rio Waupes, Mun. São Y) Gabriel da Cachoeira, Damião 2998 (INP 523. Tachigali venusta Dwyer, Ann. Missouri. Dot. Gard. 41: 235. 1954. TYPE: Brazil. Pará: Tapajoz, Boa Vista. Capucho 418 (holotype. F not seen Discussion. — Tachigali venusta is very similar to T. plumbea and T. cavtpes in Hs sc riceous pube scence, semiterete domatia. dimorphic stamens, and asymmetrical base of the leaflets. lt can be separated from the other two species by its stipules with filiform segments and the nearly sessile flowers. Stipules are not always present, but in that case the nearly sessile flowers allow identification. The upper surface of the leaflets is almost as densely pubescent as the lower surface, while T. plumbea and T. cavipes the upper surface is often less densely The seen, and the species concept is based on the paratype pubescent. type of T. venusta has not been Ducke 1989. Tachigali venusta has been reported from erra firme forest on sandy soil in the Brazilian states of Amazonas, Para, and Rondônia. examined. BRAZIL. Manaus. Reserva Florestal Ducke. US). Pará: 27 km NE of Itaituba, Prance et al. 2 MG, MO, US). odon Mun. Porto Velho. 7451 (INPA, MEL, MO Selected specimens Amazonas: Rodrigues 5431 (F, INPA, 25804 (IAN, Cid Ferreira 54. Tachigali vulgaris L. : Gomes da Silva & H. 400. 2007. synonym: Se perd paniculatum Vogel, Replaced C. Lima, Rodriguésia : an- naea 11: 397. 1837, non Tachigali paniculata Aubl., 1775. TYPE: Brazil. Mato Grosso: in regione Cujaba, Manso & Lhotzky s.n. (holotype, BF; isotype, MON). Discussion. Tachigali vulgaris is a common and widespread species, best recognized by moderately dense, appressed indument on the lower surface of the leaflets, caducous stipules and domatia, the glabrous, linear petals, and the unequal base of the leaflets. It is very similar to T. tinctoria, which differs in its sessile flowers: flowers of T. vulgaris are distinctly. pedicel- ate. The type of this species is from a collection by Manso and Lhotzky from the Cuiabá region in Mato 1837). (1870) varieties of Sclerolobium paniculatum: variety Benth.. The latter two varieties Grosso (Vogel, Bentham recognized three and paniculatum, variety rubiginosum (Tul.) variety subvelutinum Benth. have not yet been reported from the Amazon forests. Dwyer (1957a) published an additional variety aa peruvianum Dwyer, from Peru. These varieties of ; paniculatum have now been raised to species, as = peruviana. T. rubiginosa (Tul.) Oliveira-Filho, and 7. subvelutina. Specimens of T. vulgaris are known from Pará. Mato Grosso. Rondônia. and a few collections from Amazonas and Bolivia. Many specimens from Pará differ in indument from specimens from Mato Grosso and Rondônia: the indument on twigs is less dense and predominantly erect, and the leaflets also have partly erect indument. However, there are several intermediate specimens, and T. vulgaris is here accepted as a species with a somewhat variable indument, occurring in dry forest (campos, campinas) Amazon rainforest from Pará to Bolivia. Both Lemee (1952) and Oliveira 5: along the southern margin of the Filho (2006) mention that S. paniculatum and melinonii are probably conspecific. Here, S. melinonii las T. distinct species, largely based on the presence of the melinonii) and T. vulgaris are considered characteristic small, stellate hairs on T. melinonit. INCOMPLETELY KNOWN SPECIES Following is a list of validly published species of which no specimens have been seen or which are only known from fruiting collections. Also included is a likely undescribed species. Sclerolobium herthae Harms. Notizbl. Bot. Gart. Berlin-Dahlem 15: 46. 1940, TYPE: Ecuador. P: taza: Mera, Schultze-Rhonhof 2922 (holotype, B). S- The type and only collection was presumably destroved in the Berlin herbarium during World War IL 9 " Bras. 15 sylvis Capoeiras pr. 229. Tachigali multijuga Benth., Fl. 1870. TYPE: Brazil. Uananuca ad Rio Negro." Spruce 2022 (holotype. K not seen). similar specimens Ducke (1949) stated that he had not seen specimens from the upper Rio Negro, the around Rio de Janeiro. Dwyer (1954) cited, in addition Bentham also mentioned very collected near Rio de Janeiro. type locality, but that it was frequent to the Spruce collection, six collections from southern (om Brazil. Bentham compared it with Tachigali poeppigi- ana, from which it differs in having much larger flowers. Specimens identified as T. multijuga from Volume 95, Number 4 2008 van der Werff 659 Synopsis of Tachigali southern Brazil indeed have very large flowers, but no specimens like it (characterized by large flowers and 9 to 15 pairs of leaflets) have been seen from the Hio T. The specimens from n Amazonas, and it is not certain if Negro multijuga occurs in Amazonas. southern Brazil formerly placed in 7. multijuga are H.-C: now included in T. paratyensis (Vell.) Lima. Sida 16: 408. 1995. PYPE: Peru. Amazonas: Valle del Río Santiago, Quebrada Caterpiza, Huashikat 1910 (holotype. MO!) Tachigali vasquezii Pipoly, E V The holotype is a fruiting specimen. The lower surface of the leaflets is densely tomentulose, leaflets are oblong, with a rounded base, tertiary venation is raised on the lower surface, and the hypanthium is symmetrical. Domatia and stipules are absent. A second collection, Huashikat 1408 (MO), Both collections were made 2-3 km from the is also in fruit. community of Caterpiza. The sterile Tachigali setifera s. str. The differences in indument between 7. and T. intermediates have been seen. vasquezii are here considered to be 7. setifera vasquezii are striking, and no Nevertheless, | con- sider the two species closely related, based on the oblong leaflets with a rounded, symmetrical base. Tachigali sp. 1 Tall trees (height not given), resembling Tachigali chrysophylla in the sericeous lower surface of the eaflets and shape and position of the domatia (3— f stalk). It differs having pilose (not glabrous) petals and a 81 8 | = 5 em from base of leaf rom that species symmetrical base of the leaflets that is recurved. Leaflets are 9-20 cm, elliptic to oblong. One sterile collection has stipules; these are ca. 1.5 em, pinnately divided, the segments linear and distally widened in a small pouch (Fig. 1G). No other species from northern South America possesses such stipules. The hypan- thium is symmetrical. The NY specimen of Pires 3096 has old flowers in a pocket. This is very likely an undescribed species, but because good flowers are lacking. it is not described here. Specin imens Carne BRAZIL. : .. F. Rickson B- 44A- 85 (MO order area o rá, and | "Pires 3696 (COL. IAN, INPA, NY); Rio Sáo E Para—Mato Grosso, fallen fruits and leaflets, Pires 3828 (TAN) Literature Cited Aublet, F. Plantes de la Guiane Françoise. Paris. Barneby, R. C. 1996. Ne copa val ~ s at NY: Asides and oversights. Brittonia 48: 174-187. 1775. Histoire des paratypes of Amazonas: 400 km E of Be f ——— & S. V. in S. A. Mori, (editors), ( TM to the Vascular Plants of Central French Guiana, Vol. 2. New York Botanical Garden, Bro ee , Contribution à Guyanes. Bull. . Bot. France 66: PPS 2002. Caesalpiniaceae. Pp. 167-181 . Cremers, C. A. Gracie & J. de Granville l'étude BT la flore des 381l- 925. Da SC pen d'especes le de Phaner- Mus. Hist. Nat Benoist, R. ogames de la Guyane française. Bull. (Pa 18) 3l: 408—409. Be mham G. 1840. Enumeration of ae collected by Mr. — in British Guiana. J. Bot. (Hooker) 2: 38-102. 1850. Report on the Dried en collected by Mr. Spruce in the neighbourhood of Para in the months of July, Vugusl em z b imber 1849. Hookers J. Bot. Kew Gard. Misc. 2: 244 : 187 - eguminosae . Pp. 1-502 in C. F. P. Martius (editor). n Brasiliensis E N m R. 1961. Leguminosae—Cae mom ud Pp. 70-86 vee ire & J. Cm Blinde IV(2). Mem. 10(4). Ducke, A. 1915 Wurdack (editors), The Botany of the New York Bot. Gard. Plantes nouvelles ou peu connues de la region amazonienne. I. Arch. Jard. Bot. Rio de Janeiro I: 1-57. . 1922. ua nouvelles ou peu connues de la region Arch. Jard. Bot. Rio de Janeiro 3: 1—269. i. s nouvelles ou peu connues de la region amazonienne. amazonienne. VIII. Arq. Inst. Biol. Veg. 2: 27-72 3. Plantes nouvelles ou A u connues de la region X. Arq. Inst. Biol. : 1-64. Leguminosae of the amazonienne. 1944. New or ad di Brazilian Amazon. Bol. 1 Inst. Agron. “ld 49. Notas ee a F lora Ne otropic a n Bol. Tec n. Inst. Ae N E 1-248 Dwyer, J. D. | Au . The ipie ‘al American genus Tac ipe jl. (Cae ee eae). Ann. i 200 Missouri Bot. Gard. 223— i 1957a. The tropical American genus Sclerolobium Vogel (Cae a Lloydia 20: 67-117 ———. 1957b. > tropical American genus i Vogel (Cae ven eae). Llovdia 20(Suppl.): 266— . 1958. Tachigali. Pp. 152-156 in R. E. 5 os (editor). Plantae Fina ricanae X. Bot. Mus. Lea 18. Foster, R. B. 1977. Tach versicolor is a suicidal igali neotropical tree. Nature 268: 624—6: Gentry, A. H. 1993. A Field Guide to e Woody Plants of South International, Washington. D.( Harms, H. 1903. 20-25 1906. Leguminosae. /n E. U Flora der Sammlungen von Ule's Amazonas-Expedition. Verh. Bot. Vereins Prov. Branden- burg 48: 159-173. . 1915. Mj xad C i ea nigl. Bot. Gart. Berlin 6: 310 1922. Le EH americanae ie hegni Veg. 18: 232-238. . E. In J. Mildbread (editor), Plantae Notizbl. Bot. Gart. Mus Northwest America. Conservation Loris Bot. Jahrb. 33. Beibl. le (editor), Beitrüge zur Hylaea nach den Notizbl. Kó- novae. Repert. Spec. m Peruvianae III. erlin- ne m 9: 966-976 LO. Leguminosae. " Notizbl. Bot. Gart. Mus. Berlin- => Dahlem e 46-47. Huber, J. 1908. Plantae Duckeanae austro-guyanenses. Bol. Mus. 14306. “Maciek: para a Flora Amazonica. VII. M Para- ense Emilio Goeldi 5: 294-43 660 Annals of the Missouri Botanical Garden 1910. Novitates florae Amazonicae. Bol. Mus. d milio Goe Idi 6: 60-90, Irwin, H. & M. T. rrovo. 1974. Three new a 'gume species pe South America. Brittonia 26: 264— IUCN. 2001. IUCN Hed List Categories and C 3.1. Prepared by the IUCN Species Survival Commission. Paraense riteria E on IUCN, Gland, Switzerland, and Cambridge, Unite Kingdom Jussieu, A. L. de. 1789. Genera Plantarum. Paris. Lemée, A. 1952. Flore de la Guyane française, Vol H. Imprimerie Commerciale & Administrative, Brest, France Lewis, G. P. 1987. Legumes of Bahia. Royal Betania Gardens, Kew B. Se ibis B. Mackinder & M. Lock. 2005. Legume ss of the World. Eee Botanic Gardens, Kew. Lima, H. 1995. Leguminosas da Flora F anie asi. M. da C. de zo—lista ET M das especies arboreas. Acta aT e 123-14 Macbride, J. F. Pr Bot. Ser. pera ME Oliveira-Filho, A. T. 2000. Arvores Nativas de Field Mus. Nat. Hist. Tachigali. Pp. 140-141 in Minas Gerais: Mapea- mento e Inventario de Flora Nativa e dos Reflorestamentos Catalogo das de Minas Gerais. Editora Universidade Federal de Lavras. Brazil. T. D., C. Re Guide to the Trees of P, eru. Lavras, el & A. Daza. 2004. Ilustrated DH Books, Sherborne, Unitec Pennington, 1995. ) from western Amazonia. Sida 16: 407—411. A new Tachigali ipd Me nioideac) 1960. Plantas novas da Amazonia. Bol. Teen. N. 38: 23-43 Poeppig, E. 184 Pires, J. M. Inst. Agron. Nova Genera ac Species | Plantarum, 3: = -9 E eed Friderich Hofmeister. zig the plants e ile ected in Rusby, . 1896. An enume a e " Mina Bang. . Torrey Bot. Club 6: 1-130. Silva. L. F. G. da & » n de Lima. 2007. Mudanças nomenclaturais no gênero rius Aubl. (Leguminosae— Caesalpinioideae) no Brasil. CT 58: 391—401. Taubert, P. 1892, Leguminosae. Pp. 385 in A. Engler & © Prantl (editors), Die MS is ii. Pflanzenfamilien. Vol. HI. Wilhe A Engelmann, Le pag ig. Tulasne, L.-R. 1844. Légumineuses arborescentes de l'Amerique du A Arch. Mus. Hist. Nat. 4: 65-1906. Vásquez Martínez, R., R. Hojas Gonzáles € H. van der Werff. Flora del Río iur das afa Monogr. Syst. Bot. Missouri Be ard. n press). Vogel, T. 1837. De Cae ene is ud dn Linnaea. l1: 381—422. Williams, L. O. 1965. oe American Plants. VI. Fieldiana, Bot. 31(2): : Zarucchi, J. 1998. A. Steyer oe Pp. 114-120 in J. \ & B. K. Holst (editors). Flora of the uns Venezuelan Guayana, Vol. 4: Caesa Ericaceae. mark, P. E. Berry i Missouri Botanical Garden Press, St. P. Herendeen. 1993. yd al a In L. ww & J. Zarucchi (editors). a of the Flowering Syst. Bot Plants and psa: of . Monogr. Missouri Bot. Gard. < 54-1255. THE TROPICAL FLORA OF SOUTHERN YUNNAN, CHINA, AND ITS BIOGEOGRAPHIC AFFINITIES! Zhu Hua? ABSTRACT The flora of Xishuangbanna in southern Yunnan, sad stern China, consists of 3340 native seed plant species belonging to 1176 genera and 182 families. Tropical floristic eleme flora of southern Yunnan, of which the dominant geograph sal the generi c level form a major contribution (78.3%) to the toa ements are those of tropical Asian distribution. The tropical Asian flora of Xishuangbanna i is similar inc PRA lo ne r tropical floras from Yunnan, especially in the families with the These regional floras milaritie generic level, but share only 43%-50% ly at the spec and Malesia (Malay Peninsula) flora mainland Southeast Asian and Velen floras. most spec ies ric 'hness a more tha ic level. € veals that most of the peur families from southern The floristic similarities tropical Asia are more than 80% at the family level and more than 64% at the generic level. ' in 89% at the family level and more than 76% at the comparison with mainland Southeast Asia (Thailand) Yunnan are 1 the flora of southern Yunnan and those of also dominant in ; between his suggests that the tropical flora of southern Yunnan has a close affinity with tropical Asian flora and supports the idea that the flora of southern Yunnan, outheast Asian flor together with mainland S as suggested by Takhtajan, or the Malaysian ME of situated at the northern margin of tropical Asia, ‘the Paleotropical kingdom as suggested by flora of southern Yunnan comprises less strictly u ra, belongs to the Indo-Malaysian floristic subkingdom of the Paleotropical kingdom Wu and Wu. However, compared to Malaysian flora and, consequently, represents only a marginal type of Indo- Malay sian flora. The tropical flora of southern Yunnan is supposed to be derived from tropical Asian flora with the formation of f the « the Tertiary. Key words: Biogeographic affinity, southern Yunnan, astern monsoon climate after floristic composition. The tropical area of southern China is located at the northern edge of tropical Asia and is composed of the extreme southeastern part of Xizang (Tibet) among the lower valleys of the southern Himalayas, southern Yunnan, southwestern Guangxi, southern Taiwan, and Hainan separately. The largest tropical area still covered by forests is in southern Yunnan, the most southwestern region of China, which is also a key area in biogeography and a hotspot for biodiversity (Myers, 1998). margin of mainland Southeast Asia, with This is a mountainous region at the northern a monsoon climate and a slightly lower annual mean temperature (ca. 21°C) and lower annual precipitation (average, 1500 mm) in lowlands and valleys compared to the e.g., The tropical forest in main tropical rainforest areas of the world, 1996). southern Yunnan is therefore intermediate between Malaysia (Richards, classic tropical rainforests and monsoon forests as defined by Schimper (1903), or a type of subtropical rainforest, which differs in various aspects from the true tropical rainforests described by Richards (1952). After the Chinese-Russian expedition to remote areas of southwestern China, including southern in the late 1950s, papers on the tropical 1958; Wang, 1961) and tropical flora (Fedorov, 1957; Wu, 1965) of this part of China were published. It Yunnan, rainforest vegetation (Fedorov, was basically accepted that biogeographically real tropical rain- forests exist in southwestern China, but these were considered to be a different type from those in Indo- Malaysia because of the lack of representatives from Dipterocarpaceae, which dominate the rainforests of tropical Southeast Asia. Botanical interest was rekindled in the 1970s by the discovery of a dipterocarp forest in southern Yunnan, thus the Indo-Malaysian affinity of the tropical flora of Yunnan was reconsidered. Further results from biogeographic and ecological studies of the vegetation and flora of tropical southern Yunnan revealed that it does in fact comprise a part of the Indo-Malaysian flora (Zhu, 1992, 1993a, b, 1994, 1997, 2004; Zhu et al., 1998a, b, 2001, 2003; Z — mu 4 Ip prepared by Li Hong-mei. The author thanks David E. and Chris ( thanks to the reviewers fo ;larpenter > B University of California for improving t his project was funded by the National Natural Science Foundation of China (30570128, 30770158). Figure 1 was i Hor Boufford of Han ard University Herbaria for his constructive comments elish grammar within the manuscript. | also give many their impor rtant and constructive comments. Xishuangbanna Tropic 'al Lu dus Garden, Chinese Academy of Sciences, Xue Fu Road 88, Kunming 650223 Yunnan, 7 og People's Republic of China. zhuh@xt doi: 10.3417/2006081 ANN. Missouri Bor. GARD. 95: 661-680. PUBLISHED ON 30 DECEMBER 2008 662 Annals of the Missouri Botanical Garden 97? 00' 99° 00' 101° 00' 103° 00 1052 00' 107° 00' | dl A 1 d ^h " VÍ t) » OA, à no 4, % S ru | i $ £ P N » A A, di / ui E m ES ^ | Lo Ar + \ H ( | E L "is N ¿ > a d z S id ut d AC é ! D A ¿ S o en hang "I 2) NA ES ] t2 a + 2 NEM Le n : 5 tropical|SE Yunnan K E: cu (site|3) j ~ Xishuang AIN P NT (sitel1) * to ` 159) oO o e o a 140 Kilometer: 2 97° 00 99° 00’ 101° 00' 103° 00 105° 00’ 107° 00 Figure 1. Map showing the locations of the tropical areas in southern Yunnan. Site 1: Xishuangbanna administrative region in south Yunnan; site 2: Dehong administrative region in southwestern Yunnan; sile 3: tropical southeastern Yunnan. Roos, 2004). Also, T. C. Whitmore, from his visit tọ south direction, decreasing in elevation southward. Its southern Yunnan, observed that the birds in the elevation ranges from 480 m at the lowest valley tropical rainforest there sang the same songs as the — bottom in the south (Mekong River) to 2429.5 m at the n the north. The Mekong birds in the tropical rainforest of Malaysia (Whitmore, highest mountain summit 1982). and he confirmed that there is true evergreen River runs through the region from the northwest to rainforest present in the southern fringe of China the southeast (Xu et al., 1987). (Whitmore, 1984 To better understand the affinities of southern CLIMATE Yunnan’s flora, the floristic composition was concisely enumerated, and its geographic elements were Phe region has a typical monsoon climate. The annual mean temperature varies from 21.7 C at an elevation of 550 m to 15.1 C at 1979 m. and the 20 C isotherm is equal to the 850 m elevation isoline. The analyzed at family and generic levels. Also, the floristic similarities between southern, southeastern, and southwestern Yunnan (Fig. 1, sites 1-3), as well i ; as between mainland Southeast Asia (Thailand) and hottest month is June, with a mean temperature of 17.9 C at 1979 m and 25.3 € at 550 m, while the Malesia (Malay Peninsula), were compared using coldest month is January, with a mean temperature revised floristic inventories and checklists. eee ranging from 8.8 C-15.0 C. Annual precipitation increases from 1193 mm at Mengyang in the central ‘oan MEOGRAPHY ; E idc part of the region at 740 m to 2491 mm at the summit TopscREES of Nangongshan in the southern part at 1979 m. More than 80% of the precipitation falls during the rainy Xishuangbanna in southern Yunnan lies between season from May to October. In areas of lower hills 21 09'N and 22 360'N, 99 58'E and 101 50'E and valleys covered by tropical rainforests, the annual (Fig. 1). The region occupies an area of 19,690 km”. mean temperature is about 21 C and the annual It borders Burma and Laos and has a mountainous cumulative temperature (the sum of daily temperature topography with mountain ridges running in a north- means > 10 C) is 7600 C-7800 C. Annual precip- Volume 95, Number 4 Zhu 663 2008 Tropical Flora of Southern Yunnan itation is more than 1500 mm, relative humidity is ceae species during this period (Song et al., 1976, about 80%, and frost has never been recorded (Xu et 1983: Song, 1984: Wang, 1996; Penny, 2001; al., 1987). Dense fog is frequent throughout the dry season on the lower hills and in the valleys, averaging 146 foggy days per year and 1 mm precipitation per foggy day in Mengla Xian in southern. Xishuangbanna. The fog is thought to compensate for the insufficient precipita- tion so that a tropical moist climate can form locally despite relatively low annual precipitation (Zhu, 1997) SOIL There are three main soil types in the region. Lateritic soils developed from siliceous rocks such as granite, and gneiss occurs at 600-1000 m with a deep solum, but thin humus horizon. Lateritic red soils derived from sandstone substrate occur in areas above 1000 m. Limestone hill soils are derived from a hard limestone substrate of Permian origin with a pH of 6.15. occurs mainly on lateritic soils with pH values of 4.5— The tropical rainforest. in southern Yunnan 2.5. GEOLOGICAL HISTORY Southern Yunnan is at the junction of the Indian and Burmese plates of Gondwana and the Eurasian 1987; Hall, 1998). the area was part of the Tethys plate of Laurasia (Audley-Charles, Before the Mesozoic, and later margin, some fragments from Gondwana were combined. Since the Tertiary, the region had gone through several stages of rising and descending with. the intermittent. uplift of the Himalayas, and gradually formed the modern topography in the mid- Tertiary (Shi et al., 1998, 1999) During the late Cretaceous, the region had a hot, dry climate, based on fossil records from Mengla in which a relatively high proportion of Ulmipollenites 1979). the region went through a and Ephedripites were found (Sun, From the Paleocene to the Eocene. rising stage with the uplift of the Himalayas and was influenced by a dry climate with a high deposition of the the region descended and formed a series of basins with a the region has again experienced rapid uplift, associated salt minerals. From Miocene to Pliocene, wel and warm climate. Since the Quaternary, alternative climatic changes of wet and dry 1986). Paleobotanical data from the Tertiary in southern with periods (Liu et al.. Yunnan (Xishuangbanna) are extremely insufficient. Available information from neighboring regions sug- gests that southern Yunnan could be a subtropica evergreen broad-leaved forest characterized by Faga- Mehrotra et al., 2005). 1e present tropical rainforest in southern Yunnan is at the elevational and latitudinal limits of tropical rainforests in the Northern Hemisphere. It is believed that the tropical moist climate in southern Yunnan did not form until the Himalayas uplifted to a certain elevation after the late Tertiary (Zhu, 1997). Thus, the tropical rainforests in the region were developed after the later Tertiary. The fundamental topography and climate of the region have been strongly affected by the uplift of the Himalayas and the formation of the eastern monsoon climate (Shi et al., 1998, 1999). METHODS Based on identifications of ca. 60,000 specimens from southern Yunnan in HITBC, acquired during more than 40 years of intermittent field collections, 1990) was revised and updated, and 3340 native species of the former List of Plants in Xishuangbanna (Li, 1176 genera and 182 families of seed plants have been recognized (see Appendix 1). Circumscription of families and species nomenclature follows w"TROPI- COS of the Missouri Botanical Garden (http://www. tropicos.org>). Floristic and geographic attributes of the flora of southern Yunnan were analyzed. Patterns of seed plant distribution of the flora were quantified at the family and generic levels based on C. Y. bs 199]: Wu et al., 2003). The inventory data of the floras of southwestern on documentations (Wu, from the Dehong administrative region (Zhu et al., 2004) and southeastern Yunnan (Shui, 2003) (see Fig. l, sites 2 and 5, respectively) were compared to demonstrate the floristic variation in tropical Yunnan. The revised checklist of the vascular plants of Malay 1995) and the checklist of Thai plants (Smitinand, 2001) were also used to compare Peninsula (Turner, the floristic affinity between the tropical flora of southern Yunnan (Fig. l, site 1) and that of mainland Southeast Asia and Malaysia. FLORISTIC COMPOSITION A total of 3340 native species of 1176 genera and 182 families Xishuangbanna in ~ Y seed plants were recognized from The with highest species richness include Orchidaceae (94 genera/328 species), Fabaceae (65/211), Rubiaceae (45/142), Poaceae (63/132), Euphorbiaceae (36/117). Asteraceae (59/106), Moraceae 14/ 70), Urticaceae (14/72), Zingiberaceae (17/72), Acanthaceae (32/68), Lamiaceae (28/61), Asclepia- daceae (26/58), and Apocynaceae (23/51) (Table 1). southern Yunnan. families — 7/73), Lauraceae — 664 Annals of the Missouri Botanical Garden Table L The flor (Xishuangbanna) are included. Family size is from the Missouri Botanical Gardens w"PROPICOS (). No. of genera No. of species No. of genera No. of species Families with >100 spp. Orchidaceae 94 328 Poaceae 63 132 Fabaceae 05 211 Kuphorbiaceae 30 117 Rubiaceae 45 142 Asteraceae 59 106 Families with 51-100 spp. Moraceae 1 73 Veanthaceae 32 08 Lauraceae 14 70 Lamiaceae 28 61 Urticaceae 14 T2 Asclepiadaceae 26 58 Zingiberaceae 17 72 Apocynaceae 23 5l Families with 21-50 spp. \nnonaceae l5 19 Serophulariaceae I5 33 Verbenaceae 11 18 Sterculiaceae 12 33 Cucurbitaceae 17 16 Commelinaceae 10 31 Rosacea 18 16 Menispermaceac 14 30 Vitaceae 7 15 Gesnertacead 13 29 Fagaceae 5 44 Myrsinaceae 4 28 Araceae 17 43 Melastomataceae 9 27 Arecaceae o 10 Araliaceae 10 26 Cyperaceae 14 40 Rhamnaceae 9 26 Rutaceae 14 40 Solanaceae 5 25 Meliaceae 12 37 Myrtaceae 2 23 Convolvulaceae 13 36 Oleaceae 7 23 Liliaceae 19 29 Dioscoreaceae l 22 Malvaceae 10 34, Theaceae 9 22 Piperaceae 3 33 Families with 6-20 spp. Polygalaceae I 20 Bignoniaceae 7 10 Polygonaceae I 19 Lythraceae 4 10 Apiaceae 8 18 Ulmaceae 5 10 Boraginaceae 9 18 Myristicaceae 3 9 Elaeocarpaceae 2 18 Primulaceae 9 Loranthaceae 6 18 Symplocaceae l 9 Anacardiaceae 12 18 Violaceae l 9 Campanulaceae g 17 Actinidiaceae 2 8 Tiliaceae 5 17 Burseraceae 3 8 Celastraceae | 16 Proteaceac 2 8 Flacourtiaceae O 16 Sabiaceae 2 8 Loganiaceae 6 15 Sapotaceae 3 8 Smilacaceae l I5 Balsaminaceae l n Caprifoliaceac D 14 Caryophyllaceac 6 Y Capparaceae | 14 Hernandiacca | 7 Ranunculaceae | 14 Juglandaceae 3 7 Amaranthaceae 7 13 Lentibulariaceae | 7 Clusiaceae 5 13 Styracaceae | 1 Ericaceae | 13 Viscaceae 2 7 Magnoliaceae 8 13 Aristolochiaceae l 6 Sapindaceae 11 13 Ebenaceae l 0 Aquifoliaceae l 12 Malpightaceae 2 6 leacinaceae 8 12 Musaceae 2 6 Begontaceae | 11 Onagraceae 2 6 Combretaceae | 11 Passifloraceae 2 6 Cornaceae | 11 Pittosporaceae l 6 Hippocrateaceae 3 || Santalaceae 5 6 Volume 95, Number 4 665 Zhu 2008 Tropical Flora of Southern Yunnan Table 1. Continued. No. of genera No. of species No. of genera No. of species Families with 1-5 spp. Brassicaceae 3 s Podostemaceae 2 2 Elaeagnaceae | 5 Portulacaceae 2 2 Gentianaceae 5 5 Potamogetonaceae l 2 Hydrocharitaceae 4 5 Rhizophoraceae 2 2 Oxalidaceae 3 E Saururaceae 2 2 Simaroubaceae 3 5 Stemonaceae l 2 Staphyleaceae 2 5 Thymelaeaceae 2 2 Balanophoraceae l | Xyridaceae l 2 Betulaceae 3 | Berberidaceae | l Connaraceae 4 | Cephalotaxaceae | l Dilleniaceae 2 | Ceratophyllaceae l l Iridaceae 2 | Crypteroniaceae l l Lemnaceae 3 | Cupressaceae | l Nyssaceae 2 | Datiscaceae l l Podocarpaceae 3 | Dichapetalaceae | | Saxifragaceae 3 | Droseraceae | l Aizoaceae 2 3 Erythroxylaceae l l Alismataceae 2 3 Geraniaceae l l Amaryllidaceae 2 3 Grossulariaceae l l Burmanniaceae l 3 Hydrangeaceae l l Chloranthaceae 2 3 Hydrophyllaceae l l Cycadaceae l 3 Juncaceae l l Gnetaceae l 3 Lecythidaceae l l Marantaceae 3 3 Menyanthaceae l l Olacaceae 3 2 Myricaceae l l Plantaginaceae ] 3 Najadaceae l l Pontederiaceae l 3 Nvetaginaceae l l Aceraceae l 2 Orobanchaceae l l Bombacaceae | 2 Pandanaceae l l Butomaceae 2 2 Papaveracea l l Buxaceae 2, 2 Pinaceae l l Chenopodiaceae 2 2 Rafflesiaceae l l Crassulaceae 2 2 Salicacea l l Daphniphyllaceae 2 2 Sonneratiaceae | | Dipterocarpaceae 2 2 Sparganiaceae | | Elatinaceae 2 2 Sphenocleaceac | | Eriocaulaceae ] 2 Stachyuraceae l | Fumariaceae l 2 Stylidiaceae l l Hamamelidaceae 2 2 Taccaceae l | Lardizabalaceae l 2 Trapaceae l | Linaceae 2 2 Valerianaceae l l Nymphaeaceae 2 2 Zygophyllaceae | | Opiliaceae 2 2 l'otal: 182 families, 1176 genera, and 3340 species including infraspecific taxa The family Orchidaceae has higher species richness ceae (4/11), Ulmaceae (5/10), and Myrtaceae (2/23). in Xishuangbanna than in Laos and Cambodia (Chen The families Dipterocarpaceae (2/2), Datiscaceae & Tsi, 1996). (1/1), Myristicaceae (3/9), Clusiaceae (5/13), Icacina- Some families have only a small number of species, — ceae (8/12), Linaceae (2/2), and Sapotaceae (3/8) have but are very abundant in the tropical forests of even fewer species, but are also very abundant in southern Yunnan, such as Sapindaceae (11 genera/13 some forest types. species), Anacardiaceae (12/18), Burseraceae (3/8), At the generic level, Ficus L. has the most taxa (58 Elaeocarpaceae (2/18), Ebenaceae (1/6), Combreta- species including infraspecific taxa). Other genera 666 Amnals MEA En Garden Table 2. Geogra hi ( le ments of seed ants at the famil Bra | 2 level from the flora of southern Yunnan (Nishuangbanna). Table 3. Geographic elements of seed plants at the generic level from the flora of southern Yunnan (Nishuangbanna). No. of Geographic elements families (£6) s Cosmopolitan 52 (28.57) 2. Pantropic 77 (42.31) 3. Tropical Asia and tropical America 10 (5.49) disjunctions 4. Old World Tropic t (2.20) 5. Tropical va to tropical Australia 5 (2.75) 6. Tropical Asia 3 (1.65) 7. North temperate 24 (13.19) 8. Fast Asia and North America disjunctions 3 (1.65) 9. Old World temperate 1 (0.55) 10. East Asta 3 (1.05) Total 182 (100.00) with high species richness include Dendrobium Sw. 10), Piper L. Ur y Syzygium P. (43 species). Bulbophyllum Thouars (27), Calamus lL. (24). Dioscorea l. (22). Browne ex Gaertn, (22), Eria Lindl. (22). Litsea Lam. (21), Pilea Lindl. (17). Lithocarpus Blume (16). Millettia Wight & Arn. (16), Castanopsis (D. Don) Spach (16), Tetrastigma (Miq.) Planch. (16), Elaeo- (16). Elatostema J. R. Forst. & G. Forst. (14). Amonum Roxb. (14), Clerodendrum L. (14). Ardisia v. (10). Lasianthus Jack (12), Dysoxylum Blume (10). carpus L. Yi Fissistigma Griff. (9). Pometia J. R. Forst. & G. Forst.. Terminalia l... Antiaris Lesch., Gironniera Gaudich., Pouteria Aubl., Pterospermum Schreb., and Tetrameles R. Br. have fewer species, but are highly abundant in the dominant tree layer of the tropical rainforests. Lasiococca Hook. f., Garcinia L., Mitrephora Hook. f. & Thomson, Alphonsea Hook. f. & Thomson, Cleidion Blume, Sumbaviopsis J. J. Sm.. Trigonostemon Blume, and Pittosporopsis Craib are abundant in the lower tree layer (Zhu et al., 1998a; Zhu. 2006). In the tropical montane rainforest of southern Yunnan, Castanopsis, Lithocarpus, Machilus Nees, Litsea, Phoebe Nees, Anneslea Wall.. and Schima Reinw. ex Blume are the most abundant genera (Zhu et al., 2000). GEOGRAPHIC ELEMENTS GEOGRAPHIC ELEMENTS AT THE FAMILY LEVEL In the flora of southern Yunnan, families of tropical distribution contribute 54.49 of the total flora (Table 2). These families include those of pantropic distribution (contributing to 42.31% of the total flora). such as Acanthaceae, Anacardiaceae, Annonaceae. Apocynaceae, Araceae, Arecaceae, Burseraceae, Clu- siaceae, Combretaceae, Myristicaceae, Sapotaceae. No. of Geographic elements at the generie level genera (%) l. Cosmopolitan 59 (5.02) 2. Pantropic 251 (21.34) 3. Tropical Asia and tropical America 30 (2.55) disjunctions 4. Old World Tropics 112 (9.52) 5. Tropical Asia to tropical Australia 76 (6.40) 96 (8.10) 55 (30.19) 60 (5.10) 32 (2.72) 24 (2.04) 5 (0.43) 2 (0.17) Tropical Asia to tropical Africa Tropical Asia 8. North temperate 9. Fast Asia and North America disjunctions 10. Old World temperate 11. Temperate Asia 12. Mediterranean region, western to central Asia 13. Central Asia 1 (0.09) 14. East Asi: 62 (5.21) 15. Endemic to China 1 (0.94) Total 1176 (100) and Icacinaceae; Old World Tropic distribution, such as Pittosporaceae, Pandanaceae, and Musaceae: tropical Asian and tropical American disjunct distri- bution, Araliaceae, including Elaeocarpaceae, Ges- neriaceae, Lardizabalaceae, Staphyleaceae, and Styr- acaceae; and the tropical Asian distribution, such as Fam- Crypteroniaceae, Rafflesiaceae, and Sabiaceae. ilies of mainly distribution contribute 17.04% of 1 temperate distribution such as Caprifoliaceae, Pina- temperate he total flora. including the ones of north ceae, Aceraceae, Betulaceae, Buxaceae, Cornaceae, Hamamelidaceae, Juglandaceae, and Sali- North distribution, such as Magnoliaceae, Fagaceae, caceae; East Asia and America disjunct Nyssaceae, and Saururaceae; and East Asia distribution, such as Actinidiaceae, Cephalotaxaceae, and Stachyuraceae. GEOGRAPHIC ELEMENTS AT THE GENERIC LEVEI Patterns of seed plant distribution of the flora at the generic level are enumerated in Table 3. The genera of tropical Asian distribution, such as Alphonsea, Amoora Roxb.. Pterospermum, Murephora, Mycetia Reinw.. Aganosma (Blume) G. Don, Chukrasia A. Juss., Crypteronia Blume, and Anema Lour., show the highest percentage among all distribution types. contributing to 19% of the flora. Genera of pantropie distribution, such as Gnetum l.. Beil- schmiedia Nees, Cryptocarya R. Br. Capparis l.. Piper, Croton L., Dioscorea, Uncaria Schreb., La- stanthus, Morinda l., Ardisia, Bauhinia L.. and Marsdenia R. Br.. contribute to 21.349 of the flora. Volume 95, Number 4 Zhu 667 2008 Tropical Flora of Southern Yunnan lable 4. Similarity coefficients at family, generic, and species levels for three tropical regions of Yunnan.” Southern Yunnan” Southwestern Yunnan? Tropical southeastern Yunnan? Shared/similarity Shared/similarity Shared/similarity Compared regional flora coefficient, % coefficient, 96 coefficient, 96 Similarity coefficients at family level Southern Yunnan 100/100 Southwestern Yunnan 180/98.4. 100/100 Tropical southeastern Yunnan 163/89. 1 179/96.2 100/100 Similarity coefficients at generic level Southern Yunnan 100/100 Southwestern Yunnan 94.6/80.4 100/100 Tropical southeastern Yunnan 939/799 1037/76.6 100/100 Similarity coefficients at specific level Southern Yunnan 100/100 Southwestern Yunnan 1834/47.9 100/100 Tropical southeastern Yunnan 1773/50.7 2165/43.2 100/100 ' Data are from Zhu et al. (2004) for southwestern Yunnan and Shui (2003) for tropical southeastern Yunnan. There were 4933 species of native seed plants in 1432 and 199 families recognized from southwestern Yunnan (Dehong), and 4996 species of native seed plants in 1355 and 186 families recognized from tropical southeastern Yunnan based on Shui’s list. ? The similarity coefficient between two regions equals the number of taxa shared by both regions divided by the lowest number of taxa of region 1 or 2, multiplied by 100% ? Southern Yunnan is defined as Rm s iva region (site ] in Fig. 1). * Southwestern Yunnan is defined as Dehong administrative region (site 2 in Fi zJ): ? Tropical southeastern Yunnan is defined as the tropical area of hee rn part of Yunnan (site 3 in Fig. 1). Next are the genera of Old World tropical distribution, — taxus Siebold € Zucc. ex Endl., Choerospondias B. L. such as Thunbergia Retz., Dracaena Vand. ex L.. Burtt & A. W. Hill, Gardneria Wall., Hovenia Thunb., Pandanus Parkinson, Ventilago Gaertn., cian Pegia Colebr., Skimmia Thunb., Stachyurus Siebold $ Lour., Fissistigma, Polyalthia Blume, Barringtonia J. Zucc., and Pterocarya Kunth. Only 11 genera are R. Forst. & G. Forst., Carallia Roxb., Canarium L., endemic to China, including Biondia Schltr., Camp- Chasalia DC., and Uvaria L. Genera with distributions totheca Decne., Craspedolobium Harms, Cyphotheca from tropical Asia to tropical Australia include Diels, Dichotomanthes Kurz, Eleutharrhena Forman, Ailanthus Desf., Hoya R. Br., Argyreia Lour., Dillenia Nouelia Franch., Paramomum S. Q. Tong, Styrophyton L., Lagerstroemia L., Loeseneriella A. C. Sm., Murraya S. Y. Hu, Tapiscia Oliv., and Thyrocarpus Hance. J. König ex L., and Toona (Endl.) M. Roem. Genera with tropical Asian to tropical African distribution COMPARISON wrra TROPICAL FLORAS OF SOUTHWESTERN AND include Bombax L., Flacourtia Comm. ex D'Hér., SOUTHEASTERN YUNNAN Quisqualis L., Bridelia Willd., Premna L.. Urophyllum Jack ex Wall., Strophanthus DC., Mitragyna Korth., The tropical area in southwestern Yunnan is located Garcinia, Anogeissus (DC.) Wall. ex Guill., Perr. & A. mainly in the Dehong administrative region (Fig. 1, site 9 Rich., and Cymbopogon Spreng. Genera of tropical 2), which borders id Rs lies between 23 50'— distribution in all (Table 3. types 2-7) comprise 25 20'N and 97 31'-98 43'E. The region is 18.2% of the total number, while genera of temperate — 11,229 km? in area and a a monsoon climate with distribution in all (Table 3, types 8-14) contribute an annual mean temperature of 19.2 C and an annual 15.82% of the total genera, including genera with precipitation of 1540 mm. There have been 4933 north temperate distribution, such as Artemisia L., species of 1432 genera and 199 families of native seed Carpinus L., Betula L., Salix L., Swida Opiz, Corydalis plants recognized from the region (Zhu et al., 2004). ag) DC., Pinus L., and Sorbus L., those with a disjunct The tropical area in southeastern Yunnan referred distribution in East Asia and North America, such as — to in this study (Fig. 1, site 3) is located between the Schisandra Michx., Photinia Lindl., Nyssa L., Os- Tropic of Cancer and the Yunnan-Vietnam border, manthus Lour., Magnolia L., Mahonia Nutt., Illicium 22"26'-23 26" N and 104°27'—108°48'E. The region, L.. and Castanopsis, those of Old World temperate including six counties, such as Pingbian, Hekou, distribution such as Ajuga L., Elsholtzia Moench, — Jingping, Luchun, Yuanyang, and Honghe, is Herminium L., Inula L., Ligustrum L., and Paris L., 14,389 km? in area. It also has a monsoon climate and East Asian distributions such as Actinidia Lindl., with an annual mean temperature of 22.8°C and an Belamcanda Adans., Aspidistra Ker Gawl., Cephalo- annual precipitation of 1764 mm. From Shui (2003), 668 Annals of the Missouri Botanical Garden Table 5. Top 20 Jamie Ss with the highest species richness among the tropical floras of southern, southwestern, and southeastern Yunnan." Flora of tropical southeastern Flora of southern Yunnan? Flora of southwestern Yunnan! Yunnan” No. of No. of No. of Family species (Jo)? Family species (9)" Family species (%)" Orchidaceae 328 (9.82) Fabaceae 297 (6.02) Fabaceae 285 (5.70) Fabaceae 211 (6.32) Poaceae 209 (5.45) Orchidaceae 276 (5.52) Rubiaceae 14.2 (4.25) \steraceae 207 (5.41) Rubiaceae 232 (4.64) Poaceae 132 (3.95) Orchidaceae 259 (5.25) Poaceae 219 (4.38) Fuphorbiaceae 117 (3.50) Rubiaceae 166 (3.36) Asteraceae 196 (3.92) Asteraceae 106 (3.17) Rosaceae 153 (3.10) Lauraceae 141 (2.82) Moraceae 73 (2.19) Euphorbiaceae 110 (2.2 3) Urticaceae 134 (2.08) Lauraceae 70 (2.10) Lauraceae LOL (2.05) Euphorbiaceae 126 (2.52) Urticaceae 72 (2.16) Cyperaceae 97 (1.96) Rosaceae 124 (2.48) Zingiberaceae 72 (2.16) Urticaceae 90 (1.82) Fagaceae 109 (2.18 Acanthaceae 08 (2.04) Moraceae 89 ( Moraceae 104 (2.08) Lamiaceae 61 (1.83) Ericaceae 5 2) Kricaceae 96 (1.92) Asclepiadaceae 58 (1.74) Acanthaceae 77 me Lamiaceae 91 (1.82) Apocynaceae 51 (1.53) Theaceae 71 (1.44) Cyperaceae 89 (1.78) Annonaceae 19 (1.47) Liliaceae 71 (1.44) Theaceae 81 (1.62) Verbenaceae 18 (1.44) Serophulari 69 (1.40) Gesneriaceae 79 (1.58) Cueurbitaceae lo (1.38) l'agaceae 08 (1.38 Acanthaceae 77 (1.54) Rosaceae l6 (1.38) Araliaceae 66 (1.34) Celastraceae 65 (1.30) Vitaceae 15 (1.35) Apiaceae 66 (1.34) Araliaceae 64. (1.28) l'agaceae 14 (1.32) Asclepiadaceae 64 (1.30) Verbenaceae 62 (1.24) Sum of the top 20 families 839 (55.10) 2535 (51.38) 2050 (53.04) All other families ES (44.90) 2308 (48.02) 2346 (46.6) Total flora 3340 (100) 4933 (100) 4996 (100) l Data are from Zhu et al. (2004) for southwestern Yunnan and Shui (2003) for the tropical southeastern Yunnan. ? Families in boldface are the top 20 families, which can be found only in the flora of Xishuangbanna, in both the Moras of Dehong and tropical southeastern Yunnan, or in one of them, respectively. ' Southern Yunnan is defined as Nishuangbanna administrative region (site Iin Fig. 1). ! Southwestern. Yunnan is defined as Dehong administrative region (site 2 in Fig. 1). "Tropical Southeastern Yunnan is defined as the tropical area of southeastern part of Yunnan (site 3 in Fig. 1). ^ Number of species for each family in the respective floras, and their percentage within the respective floras. we compiled a list of 4996 species of 1355 genera and Apocynaceae, Annonaceae, and Cucurbitaceae are in 186 families of native seed plants from the region in the top 20 ranking families of the flora of southern southeastern. Yunnan Yunnan (Fig. 1, site 1), while Ericaceae, Theaceae, — Floristic similarities at the family, generic, and and. Araliaceae are in the top 20 ranking families o species levels between the floras of southern Yunnan these floras of southwestern and southeastern Yunnan (Xishuangbanna), southwestern Yunnan (Dehong), (Fig. l, sites 2 and 3, respectively and southeastern. Yunnan (Fig. |. sites 1, 2. and 3. Comparison of geographic elements at the generic a Table 4. The floristic level from these regional floras revealed that the respectively) are detailed in similarity between these regional floras of tropical tropical elements (Table 6, types 2-7) contribute Yunnan is more than 89% at the family level and more than 6496 of the total genera in these regional more than 76% at the generic level, but it is lower ato floras of Yunnan, and the highest proportion. of the the species level (439-50%). tropical elements occurs. in. the flora of southern The comparison of the 20 families with the most Yunnan (making up 78.3% of all the genera). species among these tropical floras of southern, southwestern, and southeastern Yunnan IS enumeral- COMPARISON WITH TROPICAL FLORAS OF MAINLAND ed in Table 5. SOUTHEAST ASIA AND MALESIA The top-ranking families in all three regional floras are basically similar. These three regional floras A Catalogue of the Vascular Plants of Malaya belong to the same floristic unit and represent the (Turner, 1995) and Thai Plant Names (Smitinand, tropical flora of Yunnan. The families Zingiberaceae, 2001) provide recently updated and relatively com- Volume 95, Number 4 2008 Zhu Tropical Flora of Southern Yunnan Table 6. southeastern Yunnan. Comparison of geographic elements at the generic level from the tropical floras of southern, southwestern, and Southern Southwestern Tropical southeastern Yunnan! Yunnan? Yunnan Geographical elements of genera No. of genera (96)! No. of genera (%)' No. of genera (£6)! |. Cosmopolitan 59 (5.0) 76 (5.3) 62 (4.6) - Pantropic 251 (21.3) 267 (18.6) 239 (17.6) 3. Tropical Asia and dps 'al America disjunctions 30 (2.6) 30 (2.1) 29 (2.1) .. Old World Tropic 112 (9.5) 120 (8.4) 116 (8.6) i Tropical Asia to HRT Australia 16 (6.5) 74 (5.2) 82 (6.1) 6. Tropical Asia to tropical Africa 96 (8.2) 94. (6.6) 95 (7.0) 7. Tropical Asia 355 (30.2) 340 (23.7) 371 (27.4) 8. North temperate 60 (5.1) 157 (11.0) 111 (8.2) 9. East Asia and North America disjunctions 32 (2.7) 54 (3.8) 17 (3.5) A Old World temperate 24 (2.0) 50 (3.5) 32 (2.4) . Temperate Asia 5 (0.4) 11 (0.8) 6 (0.4) b Me ocean region, western to central Asia 2 (0.2) 5 (0.3) 3 (0.2) 13. Central Asia 1 (0.1) |. (0.3) 1 (0.1) 14. East Asia 02 (5.3) 125 (8.7) 120 (8.9) 15. Endemic to China 11 (0.9) 25 (1.7) E (3.0) Total 1176 (100) 1432 (100) 355 (100) ' Southern Yunnan is defined as Xishuangbanna administrative re gion (site | in Fig. 1). ” Southwestern Yunnan is defined as Dehong administrative region (site 2 ir 1 Fig. 1). * Tropical southeastern. Yunnan is defined as the tropical area of the southeastern part of Yunnan (site 3 in Fig. 1). t Number of genera for each geographical element in the respective floras, and their percentage within the respective floras. plete data on the regional flora of mainland Southeast Asia and Malaysia, respectively. Comparisons of the floristic composition and similarities at the family and generic levels between southern Yunnan (Xishuang- Thailand, the Malay compared in Tables 7 and 8. banna), and Peninsula are Most of the top 20 families with the most species richness in the flora of southern Yunnan (Xishuang- banna) are shared by the floras of Thailand and the Malay Peninsula as the top 20 families (Table 7). The families Rubiaceae, Orchidaceae, Euphorbiaceae, auraceae, Moraceae, Apocynaceae, Annonaceae, Zingiberaceae, and Acanthaceae are large tropical families shared by both southern Yunnan (Xishuang- banna) and Thailand-Malaysia. However, those families with strongly tropical characteristics such as Dipterocarpaceae, Melastomataceae, Myrtaceae, Arecaceae, and Clusiaceae are also dominant in the Thai-Malaysian flora, while Urticaceae, Lamiaceae, Cucurbitaceae, Rosaceae, and Vitaceae, which are of tropical to temperate distributions, are dominant in southern Yunnan (Xishuangbanna). The families with strongly tropical characteristics have a small number of species in the flora of southern. Yunnan, but are highly abundant there (Zhu, 1997). The similarity between the flora of southern Yunnan and the floras of Thailand and Malaysia exceeds 82.5% at the family 65.6% at (Table 8). floristic level and the generic level DISCUSSION Tropical genera comprise a majority (78.2%) of the southern Yunnan (Xishuangbanna) flora. The domi- nant genera are those of tropical Asian distribution (Table 3). the Yunnan is tropical in nature and has strong tropical This reveals. that flora of southern Asian affinity. The 1 and tropical floras from southwestern, southern, (Fig. 1, 1-3) families with the mosl species ric 'hness southeastern. Yunnan sites are similar in and have high similarities at family and generic levels 89% at the family level and more than 7696 at the generic level: leui (more than see Table 4). The tropical elements comprise more than 64% of the total genera, of which the tropical Asian element contributes more than 23.7% of the total genera in these regional floras of tropical Yunnan. These patterns suggest that these three regions of Yunnan belong to the same floristic unit and are part of tropical Asian or Indo-Malaysian flora. In Yunnan (Xishuangbanna) to the tropical floras of Guangxi i southern China and Hainan in southeastern. China (Zhu & Roos, 2004), the top-ranking families are similar, except that the family Cucurbitaceae is in the comparing the flora of southern top 20 ranking families of the flora of southern Yunnan only, and Theaceae, Myrsinaceae, Liliaceae, and Myrtaceae (in Hainan) are in the top 20 families When of those of southern and southeastern. China. 670 Annals of the Missouri Botanical Garden Table 7. mpa al families with the most species richness among the tropical floras of southern Yunnan, Thailand. and the Malay Peninsula." Flora of southern Yunnan (Xishuangbanna) Flora of Thailand Flora of the Malay Peninsula No. of No. of No. of Family species (Zo)! Family species (£6)* Family species (%)' Orchidaceae 328 (9.82) Orchidaceae 530 (8.07) Orchidaceae 853 (11.16) Fabaceae 211 (6.32) Fabaceae 398 (6.51) Rubiaceae 502 (1.35) Rubiaceae 142 (125) — Euphorbiaceae 305 (5.97) | Euphorbiaceae 309 (4.81) Poaceae 132 (3.95) Rubiaceae 357 (5.84) Fabaceae 298 (3.90) Euphorbiaceae 117 (3.50) — Poaceae 227 (3,71) | Poaceae 238 (3.11) Asteraceae 106 (3.17) Cyperaceae 222 (3.03) Myrtaceae 215 (2.81) Moraceae 73 (2.19) Arecaceae 167 (2.73) Lauraceae 214 (2.80) Lauraceae 70 (2.10) Araceae 151 (2.47) Annonaceae PO? (2.64) Urticaceae 72 (2.10) Apocynaceae 120 (1.96) Arecaceae 198 (2.59) Zingiberaceae 72 (2.16) Annonaceae 117 01.91) Gesneriaceae 189 (2.47) Acanthaceae 68 (2.04) Zinglbs raceae 112 (1.83) Melastomatz 172 (2.25) Lamiaceae 61 (1.83) Serophul 103 (1.69) Cyperaceae 162 (2.12) Asclepiadaceae 58 (1.74) Moraceae 103 (1.00) — Acanthaceae 158 (2.07) Apocynaceae 51 (1:523) Asleraceae 100 (1.64) Dipterocarpaceae 156 (2.04) Annonaceae 49 (1.47) l'agaceae 100 (1.64) Zingiberaceae 153 (2.00) Verbenaceae 48 (1.44) Myrtaceae 96 (1.51) Araceae 141 (1.84) Cucurbitaceae 46 (1.38 Verbenaceae 83 (1.36) Moraceae 138 (1.81) Kosaceae 46 (1.38) Lauraceae 82 (1.34) Clusiaceae 120 (1.57) Vitaceae 45 (1.35) Acanthaceae 00 (1.31) Apocynaceae 119 (1.56) l'agaceae 44 (1.32) Convolvulaceae 18 (1.28) Asclepiadaceae 116 (1.52) Sum of the top 20 families 1839 (55.10) 3591 (58.75) 1773 (62.42) All other families 1501 (44.90) 2520 (41.25) 2804 (37.58) Total flora 3340 (100) 6111 (100) 1637 (100) ! Data are from Smitinand (2001) for Thailand and from Turner (1995) for the Malay Peninsula. There were OLLI species a native seed plants in 1573 genera and 198 families recognized from Smitinand's list, and 7037 species of native seed plants in 554 genera and 191 families a from Turner's list. Cireumseription of families and nomenclature follow the Missouri Botanical Garden's w'VROPICOS ( Gawl. (7), Paris L. (1), Peliosanthes Andrews (1), Polygonatum Mill. Reineckea Kunth (1), Tupistra Ker Cawl. (2) Ixonanthes Jack (1). Reinwardtia Dumort. (1) Buddleja L. (6), Fagraea Thunb. (1), Gardneria Wall. (2), Gelsemium Juss. (1), Mitreola L. (1) Strychnos L. (4) Dendrophthoe Mart. (1) ducc (Blume) Blume (1). Helixanthera Lour. (4). Macrosolen (Blume) Blume (4), Seurrula L. (7), Taxillus Tiegh. Ammannia L. (3), len mia L. (4), Rotala L. (2), Woodfordia Salisb. ue Alcimandra Dandy (1), Hicium L. (2). Kadsura Juss. (1). Magnolia L. (1). Manglietia Blume (2). a (2). Parakmeria | & W. C. Cheng (1), Paramichelia (1), Schisandra Michx. (3) — REC i Juss. ex Endl. (3), Hiptage Gaertn. (3) Abelmoschus Medik. (5), Abutilon Mill. (2). Cenocentrum Gagnep. (1), Hibiscus L. (4), Kydia Roxb. (3). Malva L. (1), Malvastrum A. Gray (1). Sida l. (11), Thespesia Sol. ex Corréa Ha Urena L. (5) Donax Lour. (1), Phrynium Willd. (2). Stachyphrynium K. Schum. (1). Allomorphia Blume (2), Cyphotheca Diels (1), Medinilla vine h. ex DC. (4), Melastoma L. (3), Memecylon L. (3), Osbeckia L. (6). Oxyspora DC. (2). Sonerila Roxb. (5). Styrophyton S. ee Hu (1) Aglaia Lour. (2), Amoora Roxb. (7). Aphanamixis Blume pi . Chisocheton Blume (1). Chukrasia A. 1). 2). Cipadessa Blume (1). Dysoxylum Blume (10), Melia L. (2). Munronia Wight (1). Toona (Endl.) M. Roem. (5), Trichilia P. Browne (2). Walsura Roxb. (2) Aspidocarya Hook. f. € Thomson (1). Cissampelos L. (1), Cocculus DC. (2). Cyclea Arn. ex Wight (6). Diploclisia Miers (V), Eleutharrhena Forman (1). Hypserpa Miers (1). Pachygone Miers (1). Parabaena Miers (1). Pericampylus Miers (1), Pycnarrhena Miers ex Hook. f. & Thomson ( Stephania Lour. (9), Tinomiscium Miers ex Hook. f. & Thomson (1), Tinospora Miers (3) p n — Nymphoides Ség. (1) Antiaris Lesch. (1), Artocarpus J. R. Forst. € G. Forst. (5). Broussonetia L Hér. ex Vent. (2). Ficus L. (58), Maclura Nutt. (3). Morus L. (2). Streblus Lour. (2) Ensete E (1), Musa L. (5) Myrica L. (1 o Willd. (4). Anema Lour. (4), Myristica Gronov. (1) Ardisia Sw. (10), £ "ie Burm. f. dr Maesa Forssk. (8), Myrsine L. (2) Decaspermum J. R. Forst. & G. Forst. (1), Syzygium P. Browne ex Gaertn. (22) Najas L. (1) Pisonia L. (1) Brasenia Schreb. (1), Nuphar Sm. Camptotheca Decne. (1), Nyssa L. Erythropalum Blume (1), Olax L. (1 i Chionanthus L. (1), Fraxinus L. (2), Jasminum L. (10), Ligustrum L. (4 Olea L. (3), Osmanthus Lour. (2) Epilobium L. (1), Ludwigia V. (5) Opilia Roxb. (1). Urobotrya Stapf (1) Acampe Lindl. (2), Acanthephippium Blume ex Endl. (2), Acriopsis Blume (1). Aerides Lour. (2). Agrostophyllum Blume (1), Ania Lindl. (3), Anoectochilus Blume (3), Anthogonium Wall. ex Lindl. Arundina Blume (1), Schoepfia Schreb. (1) Linociera Sw. ex Schreb. (1). — 1). Arachnis Blume (1), (1), Aphyllorchis Blume (2), poti Blume Ascocentrum Schltr. ex J. J. Sm. (1), Brachycorythis Lindl. (1). Bulbophyllum Thouars (40). Calanthe R. Br. (5), > Called Baris a, Cephalantheropsis Guillaumin (1), Ceratostylis Blume (1). T a 678 Annals of the Missouri Botanical Garden APPENDIX Continued. Families Genera Orobanchaceae (1) Piperaceae (3) Pittosporaceae (1) Plantaginaceae (1) Poaceae (63) Podocarpaceae (3) Podostemaceae (2) Polygalaceae (4) Polygonaceae (4) Pontederiaceae (1) Portulacaceae (2) Potamogetonaceae (1 Primulaceae (1) Proteaceae (2) Rafflesiaceae (1) Ranunculaceae (4 9) Rhamnaceae ) ) Cheirostylis Blume (2), Chiloschista Lindl. (V). Chrysoglossum Blume (1), Cleisostoma Blume (10), Coelogyne Lindl. (9), Corymborchis Thouars (1). Cymb Put Sw. (10), Dendrobium Sw. (43) Diploprora Wook. f. (1). Epipactis Zinn (1), Epipogium J. Gmel. ex Borkh. (2). Eria Lindl. (22). Eriodes Rolfe (1), Erythrodes Blume (1), Esmeralda Re ^ i (1), Eulophia R. Br. ex UM (1) Flickingeria A. D. Hawkes (6). Galeola Lour. (2). Gastrochilus D. Don (5). Gastrodia R. Br. (1). Geodorum Jacks. (2). Goodyera R. Br. (3), Habenaria Willd. (12). Hemipilia I. indl. (1). ne L. (D. Hetaeria Blume (1), Holcoglossum Schltr. (2), Hygrochilus Plitzer (1), Kingidium P. F. Hunt (1), Lecanorchis Blume (1), Liparis Rich. (1 | Pas Gaudich. (4), Malaxis Sol. ex Sw. (9). Nervilia Commons ex Gaudich. (2), Oberonia Lindl. (13), Ornithochilus (Lindl.) Wall. ex Benth. (1). Otochilus Lindl. (2). Panisea Lindl. (1). s Plitzer (2). Papilionanthe Schltr. (2). Parapteroceras re v. (1), Pecteilis m (2). Pelatantheria Ridl. (3), Pennilabium J. J. Sm. (1), Peristylus Blume (4). Phatus Lour. (6). Phalaenopsis Blume (V). Pholidota Lindl. ex Hook. (4). Phreatia Lindl. (1), Platanthera Rich. (1). Pleione D. Don (1), Podochilus Blume (1). Polystachya Hook. (1), Porpax Lindl. (1), Pteroceras Hasselt ex Hassk. (1). Rhynchostylis Blume (1), Robiquetia Gaudich. (1), Sarcoglyphis Garay (1), Schoenorchis Blume (2). Spathoglottis Blume (1). Spiranthes Rich. (1), Staurochilus Ridl. ex Pfitzer (2). Sunipta Lindl. (5). Taeniophyllum Blume (2), Tainta Blume (2), Thelasis Blume (1). Thrixspermum Lour. (2). Thunia Wehb. f. (0). Prichoglottis Blume (1), Tropidia Lindl. (2). Unc sin Lindl. (1). Vanda R. Br. (4). Vandopsis Pfitzer (2). Vanilla Plum. ex Mill. (1), Zeuxine Lindl. Aeginetia L. (1) \verrhoa L. (1), Biophytum DC. (3). Oxalis L. ( Pandanus Parkinson (1 Argemone Lo Adenia Forssk. (2). Passiflora V. (4) Pinus l. ( Peperomia Ruiz & Pav. (5), Piper L. (27). Zippelia Blume Pittosporum Banks ex Gaertn. (6) Plantago L. (3) leroceras Stapf (1), Apluda L. (1), Arthraxon P. Beauv. (1), Arundinella Raddi (2). Arundo L. (1). Bambusa Schreb. (3). Bothriochloa Kuntze (1). Brachiaria (Vrin.) Griseb. (1). Centotheca Desv. (1). e hyum Munro (2), Chimonobambusa Makino (2). Chloris Sw. (1), Chrysopogon Trin. (1). (1). Cymbopogon Spreng. (2). Cynodon Rich. (2). Cyrtococcum Stapf (1), Dendrocalamus Ne: es vs Digitaria Haller (4). Echinochloa P. Beauv. (2). Eleusine Gaertn. (2). Eragrostis Wolf (6) Erianthus Michx. (1), Eriochloa Kunth (1), Fargesia Franch. (1), Gigantochloa Kurz ex Munro (6). Hackelochloa Kuntze (1). Hymenachne P. Beauv. (B). Hyparrhenia Andersson ex E. Fourn. (1). Imperata Cirillo (1), Indosasa McClure (3). [sache R. Br. (1). Ischaemum L. (1). Leptochloa V. Beauv. (1), Lophatherum Brongn. (1). Melocalamus Benth. (3). Microstegium Nees (1). Miscanthus Andersson (1), Neyraudia Hook. f. (1), Oplismenus P. Beauv. (2). Oryza L. (2), Panicum L. (3). Paspalidium Stapf (1), Paspalum L. (5). Phyllostachys Siebold & Zuec. (3), Pletoblastus Nakai (1). Poa L: (I. P n P. Beauv. (1), Pseudechinolaena Stapf (1), Pseudostach yum. Munro (1). Rottboellia V. f. (1), Saccharum L. (2). Sacciolepis Nash (3). Schizostachyum ^es (3), Setaria P. Beauv. (4). S Moench (2), Sporobolus R. Br. (1). Stenotaphrum Trin. (1). Themeda Vorssk. (5), Thyrsostachys Gamble (2). Thysanolaena Nees (1). Urochloa P. Beauv. (1). nn Keng f. (1) Dacrycarpus (Endl.) de Laub. (1), Nageía ce (2), Podocarpus L Hér. ex Pers. (1) Cladopus H. Moller (1). Hydrobryum Endl. Polygala L. (13). o Lour. (4), Sec is a V. (1), YXanthophyllum t (2) ). Polygonum L. (16), Reynoutria Houtt. (1). Rumex L. Fagopyrum Mill. Monochoria C. D (3) Portulaca L. (1), Talinum Adans. (1) Potamogeton |. (2) PS O) Lysimachia l.. Helicia Lour. (6). Heliciopsis Sleumer (2) Sapria Griff. (1) Inemone L. (2), Clematis L. (9), Naravelia Adans. (1), Ranunculus L. Berchemia Neck ex DC. (3). S et Pit. (1), Gouania Jacq. (4). as Thunb. (1). Rhamnus l. Sageretia Brongn. (2). Scutia (Comm. ex A. DC.) Brongn. (1), Ventilago Gaertn. (5). Ziziphus Mill. (6) 2), Volume 95, Number 4 Tropical Flora of Southern Yunnan APPENDIX 1. Continued. Families Genera Rhizophoraceae (2 Rosaceae (18) Rubiaceae (45) Rutaceae (14) Sabiaceae (2) Salicaceae (1) Santalaceae (5) Sapindaceae (11) Sapotaceae (3) 2 2) Saururaceae Saxifragaceae (3) Serophulariaceae (15) Simaroubaceae (3) Smilacaceae (1) Solanaceae (4) Sonneratiaceae (1) “panganan eae (1) Sphenocleaceae (1) Stachyuraceae (1) Staphyleaceae (2) Stemonaceae (1) Sterculiaceae (12) Stylidiaceae (1) | Theaceae (9) Thymelaeaceae (2) Tiliaceae (5) Trapaceae (1) Ulmaceae (5) ; Agrimonia L. Carallia Roxb. (1), Pellacalyx Korth. (1) (2), Cerasus Mill. (1). Chaenomeles Lindl. (1), Dichotomanthes Kurz (1), Docynia Decne. . Duchesnea Sm. (2), Eriobotrya Lindl. (2), Laurocerasus Duhamel (4), Neillia D. Don (1), e Lindl. (5), Potentilla L. (3). e um Gaertn. (3), Pyracantha M. Roem. (1), Pyrus L. (2), Re » (1), Rubus L. (10). Sorbus L. (5), Stranvaesia Lindl. (1) Aidia she (2), Borreria G. Mey. (1), fee EE Hook. f. (2), Canthium Lam. (3), Cephalanthus L. (1), Chassalia DC. (1), Dentella J. R. Forst. € G. Forst. (1). Diplospora DC. (2), Duperrea Pierre ex Pit. Gardenia J. Ellis (1). Geophila D. Don (1), — L. (14), Hymenodictyon Wall. (3), T (1). Hyptianthera Wight & Arn. (1). /xora L. (6). Knox (0. Lasianthus Jack (12), Litosanthes Blume (1), Metadina Bakh. f. (1), Mitragyna Korth. (1). n L. (3), Mussaenda L. (7), Mycetia Reinw. (9), Myrioneuron R. Br. ex Kurz (1), Neanotis W. H. Lewis (4), Neolamarckia Bosser i Neonauclea Merr. (3). Ophiorrhiza L. (13), Oxyceros Lour. (3), Paederia L. (2), Pavetta L. (4). Prismatomeris Thwaites (1). Psychotria L. (6), Randia L. (1). Rubia L. (2). Saprosma Blume (1). Schizomussaenda H. L. Li (1). Sinoadina Ridsdale (1), Spiradiclis Blume (2). Tarenna Gaertn. (1), Tarennoidea Tirveng. € Sastre (2). Uncaria Schreb. (5), Urophyllum Jack ex Wall. (1), Wendlandia Bartl. ex DC. (11), Xeromphis Raf. (1) Acronychia J. R. Forst. & G. Forst. (1), Atalantia Corrêa (1), Boenninghausenia Rehb. ex Meisn. (1), Citrus L. (1), Clausena Burm. f. (5). Euodia J. R. Forst. € G. Forst. (8), Fortunella Swingle (1), Glycosmis Corrêa (6). Mieromeliun Blume (3). Murraya J. s ex L. (2), Paramignya Wight (1). Skimmia Thunb. (1). Toddalia Juss. (1). Zanthoxylum L. Meliosma Blume (5), Sabia Colebr. (3) Mcd Salix L. (1) pd Miq. (1). Osyris L. (1), Phacellaria Benth. (2). Pyrularia Michx. (1), Scleropyrum Arn. (1) Allophylus L. (3), Arytera Blume (1), Cardiospermum L. (1), Dodonaea Mill. (1). Harpullia Roxb. (1), n zu (1). Mischocarpus Blume (1), Nephelium L. (1). Pometia J. R. Forst. & G. Forst. (1). Sapindus L. (1). Xerospermum Blume (1) Pouteria Aubl. (1), Sarcosperma Hook. f. (4), Xantolis Raf. (3 — in Thunb. (1), Saururus L. (1) ench (1), Hea L. (2). dai (1) RE E Br. (3), Alectra Thunb. (1), Bacopa Aubl. (1), Brandisia Hook. f. & Thomson (2). Centranthera R. Br. (2), Dopatrium CR UR ex Benth. (1), Limnophila R. Br. (4), Lindenbergia Lehm. (2), Lindernia AM. (10), Mazus Lour. (1). Microcarpaea R. Br. (1), Pieria Lour. (1). Torenia L. (2). Veronica L. (1), Wightia Wall. (1) Ailanthus Desf. (2), Brucea J. F. Mill. (2). Picrasma Blume (1) Bergenia = — Smilax L. (15) Lycianthes (Dunal) Hassl. (6), Nicandra Adans. (1), Physalis L. (2), Solanum l. (15) Duabanga Buch.-Ham. (1) Sparganium bL. ( Sphenoclea Gaertn. (1) Stachyurus Siebold & Zucc. (1) Tapiscia Oliv. (1). Turpinia Vent. (4) Stemona EN (2) 1). Byttneria Loefl. (3), Ertolaena DC. (4), Erythropsis Lindl. ex Schott € Endl. (2), Ambroma l. f. ( Helicteres : (6), Heritiera Aiton (1). Melochia L. (1). Pterospermum Schreb. (5), Pterygota Schott € Endl. (1), Reevesta Lindl. (3), Sterculia L. (5), Waltherta L. (1) Stylidium Sw. ex Willd. (1 Alniphyllum Matsum. (1), Bruinsmia Boerl. € Koord. (1), Huodendron Rehc er (1), Styrax L. (4) Symplocos Jacq. (9) Tacca J. R. Forst. € G. Forst. (1) Adinandra Jack (1), Anneslea Wall. (1), Camellia L. (6). Eurya Thunb. (5), Gordonia J. Ellis (1), Py yrenaria Blume (3), Schima Reinw. ex Blume (2), Sladenia Kurz (1), Ternstroemia Mutis ex L. f Aquilaria Lam. (1), Eriosolena Blume (1) Colona Cav. (2), Corchorus L. (3), Grewia L. (6), Microcos L. (2), Triumfetta L. (4) Trapa L. (1) Aphananthe Planch. (2), Celtis L. (3), Gironniera Gaudich. (1). Trema Lour. (3), Ulmus L. ( jaa — 680 Annals of the Missouri Botanical Garden APPENDIX l. Continued. Families Genera Urticaceae (14) Valerianaceae (1) 11) Verbenaceae (1 Violaceae (1) Viscaceae (2) Vitaceae (7) Xyridaceae (1) Zingiberaceae (17) — Zygophyllaceae ( Hosen we (13), Debregeasia Gaudich. (4), Dendrocnide Miq. (3), Elatostema J. R. Forst. & * dm 1), Maoutia Wedd. (1), Oreocnide Miq. 7), Potkilospermum Zipp. ex Miq. (2), Pouzolzia Gaudich. ( ; 14), Girardinia Gaudich. (1), Gonostegia Turez. 6), Pilea Lindl. (1 Procris Comm. ex Juss. (1), Urtica ie Gaudich. (€ Triplostegia Wall. ex DC. (1) Callicarpa L. (11), Caryopteris Bunge (1), Clerodendrum L. "letcher (1), Gmelina L. (1), Phyla Lour. (1), Premna L. Roxb. (1), Vitex L. (7) Viola L. (9) Korthalsella Viegh. (1), Viscum L. (6 Ampelopsis Michx. (4). Cayratia Juss. (4), Cissus L. (9), Leea D. Royen ex L. Planch. (16), Vitis L. (4), Yua C. L. Li (1) Xyris L. (2) (14), Congea Roxb. (2). Garrettia H. R. (8). Sphenodesme Jack (1). Symphorema a (7), Tetrastigma (Miq.) Alpinia Roxb. (11), Amomum Roxb. (14), Boesenbergia Kuntze (1), Cautleya Hook. f. (1), Costus L. (3), Curcuma L. (9), Curcumorpha A. S. Rao & D. M. Verma (1). Etlingera Giseke (1). Globba L. (4), Hedychium J. Konig (7), Kaempferia L. (3), Paramomum S. Q. Tong (1), Pommereschea Wittm. (2). ), Stahlianthus Kuntze (1), Zingiber Mill. (11) Rhynchanthus Hook. f. (1), Siliquamomum Baill. (1 L. ( Tribulus L. (1 Acknowledgment of Reviewers The following individuals are thanked for their Holger Kreft collegial reviews in 2007. This peer commitment of Henrik Lantz time and effort is sincerely appreciated by the Annals. Magnus Lidén Ihsan Al-Shehbaz Pieter Baas Volker Bittrich Josef Bogner Lynn Bohs Kjell Bolmgren David E. Boufford Priscilla Burgoyne David Chamberlain André Chanderbali Rocio Cortés-B. Thomas B. Croat Edmond de Langhe Miriam L. Denham Steven Dessein Eric Dinerstein Laurence J. Dorr Cecilia Ezcurra Leandro Freitas Peter Fritsch Roy E. Gereau Peter Goldblatt Rafael Govaerts Alan Graham Patrick Herendeen John Herr Michael Hesse Paul Hiepko Peter Hoch Suzy Huysmans Steven Jansen Joel Jérémie Iván Jiménez Carel C. H. Jongkind Joachim W. Kadereit Kent Kainulainen Michael Kiehn Marcus Koch Sigrid Liede-Schumann David H. Lorence Martin Lysák Robert E. Magill Karol Marhold Michael S. Mayer Simon J. Mayo Gordon D. McPherson Thomas Mione Anna K. Monfils Michael Moody Timothy J. Motley Klaus Mummenhofl —— — Barbara Neuffer Sachiko Nishida Iris Peralta Fernanda Pérez Claes Persson Carolyn Elinore Barnes Proenca Sylvain Razafimandimbison Antonio Salatino Neil W. Sawyer Marlies Sazima George Schatz Robert A. Schlising David S. Seigler Peter Stevens John L. Strother Thomas Stützel Charlotte Taylor Mats Thulin Tim Utteridge Henk van der Werff Michael A. Vincent Linda A. Vorobik Warren L. Wagner Suzanne I. Warwick George Yatskievych María del Carmen Zamaloa Felipé Zapata —Votis gratias agamus. ANN. Missourt Bor. Garp. 95: 681. PUBLISHED ON 30 DECEMBER 2008. ANNALS OF THE MISSOURI BOTANICAL GARDEN VOLUME 95 2008 684 Annals of the Missouri Botanical Garden Colophon This volume of the Annais of the Missouri Botanical Garden has been set in APS Bodoni. The text is set in 9-point type while the figure legends and literature cited sections are set in 9-point type. This volume has been printed on 70% Sterling Ultra Gloss. This is an acid-free paper designed to | B pal 8 have a shelf-life of over 100 years. Photographs used in the ANNALS are reproduced using 300 line screen halftones. The binding used in the production of the ANNALS is a proprietary method known as Permanent Binding. f Lawrence, Kansas 66044, U.S.A. The ANNALS is printed and distributed by Allen Press, Inc. « O Missouri Botanical Garden 2008 ISSN 0026-6493 ANNALS OF THE MISSOURI BOTANICAL GARDEN: CHECKLIST FOR AUTHORS The Annals publishes original articles in systematic botany and related fields. Papers whose purpose is the establishment of new nomenclatural entities in vascular plants and bryophytes are not accepte ather, they should be submitted to Novon for a (Novon on available on the Web or from the instructic are managing editor.) CONDITIONS FOR PUBLICATION ause the electronic file will be used by the printer for E it is important to adhere to the items listed in the format section of the checklist. red by the editors and sent to the printer for formatting. Every electronic file sent by the author should be clearly labeled with the first author's name, the first four letters of the taxon involved, and Authors’ electronic files are pre an ade ‘ation of the file’s contents (i.e., text, figure, table, etc.). Manuscripts that have been reviewed befor submission will be subject to the full normal ee process initiated here. Manuscripts not properly prepared may be returned for revision prior to review. It is the author’s responsibility to supply missing bibliographic information; the editor does not perform sd research with authors not except by prior arrangement M Commu nication access to crucial reference materials. with the editor at all points of the publication process is 1 encouragec PAGE CHARGES ge costs are $80 per page, although charges may sometimes be reduced or waived. Chan by authors will be billed non-negotiably to the author at $3.50 per line changed. nges in proof made CORRESPONDENCE Submissions should be addressed to: Beth Parada Managing Editor Annals of the Missouri Botanical Garden Missouri Botanical Garden Press P.O. Box 299, St. Louis, MO 63166-0299 U.S.A. Managing Edito Annals of the MSIE Botanical Garden (Street address) Missouri Bonis al Garden Press Monsanto Center 500 S lvd. St. Louis, MO 63110-2291 U.S.A. Tel: (314) 577-9473 ext. 6519 Fax: (314) 577-9594 Email: annals@mobot.org l. General O Text is in English or Pn on numbered pages as Microsoft Word file LJ] Manuscript is submittec via email if esate or by ftp [contact annals@ O O O O O O O 0 o mobot.org for instructions or for alternate electronic media]. MC file of manuscript is labeled with first our letters of the taxon involved, and an indication of what the file contains, ., Celis Cipu text. = Celis Cipu Figure 1.tif, Celis Cipa Tablel.doc, Electronic. scans P E are included with the author's last name, first submission, with the understanding that these may have to be revised based on review. Low-resolution JPEG are not acceptable for final submission to t See images are suitable for the review process, but the printer. the printer's electronic illustration requirements. Go to http://www.mbgpress.org/; click on the Publica- tion ri at upper Ies of E page; click gital Art, Allen P p ecd to Brummit & on Guide tc a Powell's pud of Plant Names, are provided for species” names the first time they are mentioned in the text; OR hey are provided in a table that includes names of nail o is enclosed o the author’s al mailing RERUM telephone number, fax number, and e-mail address ee 2. Format of Text Right-hand margin is not justified, and words are not hyphenated there. e space is inserted after periods, colons, question marks, and all other punctuation marks. First sentences of paragraphs are indented with a tab. Two hyphens with no space between them are used to indicate ranges of numbers, for example, page numbers, 20-23; measurements, 5-7 mm; altitudes, 1500-2500 m. A single hyphen is used in hyphenated words, for example, 9-nerved, oval-triangular. Common Latin words or phrases are not italicized (e.g., et al., i.e., sensu, etc. Only taxon names at the eal of genus and below are italicized. Correct accents, umlauts, and other diacritical marks have been included. All figures and tables are cited in the text and are numbered in the order in which they are to appear. 3. Style Recent issue of the Annals is used as a model. Chicago Manual of Style, latest edition, is used as a reference. 4. First Page Footnotes are typed as double-spaced paragraphs on the first page. Th edgments, inc a: information on granting agencies, herbaria that loaned specimens, and the name of the e first footnote contains acknowl- artist. The ae footnote is the author’s address. 686 Annals of the Missouri Botanical Garden ] O C] l Addresses for additional authors are No other f where unavoidable. each separate footnotes. ootnotes occur, except in tables, 5. Abstract & Key Words A one-paragraph abstract precedes the text. Papers in Spanish have an English abstract in addition to a Spanish resumen. The abstract is concise (1 paragraph) and includes brief statements about the papers intent, materials results, and significance of findings. key a methods, A brief Abstract. list of words immediately follows the 6. Taxonomic Treatment ecies entries are organized as follows: jee vernacular name(s), Latin diagnosis (if necessary! description, distribution, summary, discussion, speci- parallel and follow mens examined. The discussions are the same order, e.g., diagnostic e distine- tion from similar species variation, distribution. and ecology, nomenclature and typification, uses. One paragraph per basionym is used as follows: Taxon author, literature citation, n citation, e.g.. Beilsch- Sal: latifolia (Nees) S ishida, Ann. Missouri Bot. Gard. 86: 680. 1999. Hufelandia e Nees, TYPE: Peru. Locality 1835 (fl), Matthew 14583 (holotype. BM! El, K! LE not seen, OXF Synonyms based on different types are placed in nol DIE d iue 074. 1836 indicated, isotypes, nol seen). separate paragraphs, each beginning with the basio- nym, followed by other combinations (if appropriate), and citation of the type. A brief Latin diagnosis for each new taxon is provided rather than a complete Latin description. For species with infraspecific taxa: Description and discussion are composite (incorporating all infraspe- cific taxa) and parallel with other scriptions. Descriptions of infraspecific parallel with one another (in the same species). All synonyms are listed under the appropriate infraspecific taxon. Descriptions: Descriptions within : are parallel, given rank. All measurements are metric. Hyphens parenthetical extre “peduncle 31.9) are nol expected: ovary with (2)4(6) locules. are used for (8.2—)14.3—28.0(— values mes: em long,” unless intermediate Length X width are A in the following manner: amina 36.4-82.8 X 9.1-16.8 « When relevant, nomina nuda, "alit d names, and discussion the superfluous names are included in following the dud but are nol pu of formal synonymy. 7. Citation of Types Exclamation points are used for specimens examined, and types nol seen are indicated as such (e.g.. MO! US not seen). Lectotype designations are included together with an indication of where they were designated [author, year, [ page number, and herbarium of deposition: e.g.. C. J. W. Schiede 159 (lectotype, designated. by Stevens (2000: 2 in the Literature Cited. 56), Pl: isotypes. .. a Phis reference is listed f the author of the paper submitted is making the ma ation, the phrase “designated here” is used. 8. Tables Create tables using the Microsoft. Word “Table” feature Tables are neat, double-spaced, and clearly presented. Captions are typed double-spaced as paragraphs at l tables Each table starts on a separate sheet. the tops of the 9. Abbreviations Periods are used after all abbreviations (which are minimized) except metric measures, compass direc- tions, and herbarium designations. When dates are given part of collection in- formation, three-letter month abbreviations are used, except for months with four letters, which are spelled out in full. States are not abbreviated, and cities are spelled out. [St as in St. Louis, is acceptable. | according to B-P-/ and to B-P- E (Botanico-Periodicum- anaes ose mentum). Authors” names are abbreviated according to Brummit & Powell's Authors of Plant Names. Book titles are Literature, edition 2, but with initial letters e apitalize xd. Book titles If an item does not appear in B-P-H or TL-2 Periodicals are i des (Botanico-Periodicum-Huntianum) — abbreviated according to Taxonomic are spelled out in the Literature Cited. 1 M references are not available, its title is fully spelled out. Herbaria are abbreviated according to the most recent edition of Index Herbariorum. Abbreviated forms are not used for references in id text, except when citing the names of plants. If i necessary to cile a particular page in the text. e (1998: | form Smith 12) is usec I0. Specimens Examined If many specimens were examined, those cited in the text are limited to ca. 12 manuscript pages. If there are a large number, an index to specimens examined is placed at the end of the paper, following the Literature Cited. It is arranged alphabetically by collector, followed by collection number, followed hy the number of the taxon in the text. Names (including mae la le of first and second collector are provided, etal.” if three or more. ect are cited in the text as follows: Additional ro examined (or Selected ee examined). MEXIC daxaca: Sierra San Pedro Nolesco, Talea, 12 de 85 I4'W, 950-1100 m. 3 Feb. 1987 (fl). Jergensen 665 (BM. G. K, US). [Dates and reproductive status are optional but are omitted from longer lists. | Countries are run toge ather in the same paragray ph, « QE. COUNTRY A. Major political division: ...COU Volume 95, Number 4 2008 Lj O Ol O O [] L ] TRY : rale paragraphs are used for major ned regions Major political divisi Sepa within major political divisions. Specimen Vouchers and Genetic Sequences If the paper presents original data, associated her- barium vouchers are cited. [Vouchers for seed and/or other collections should be included where pertinent. Dependent on the paper, reference to the original wild source may be required.] Vouchers are also cited from common names, and uses are taken from specimen labels. Herbarium vouchers state the collector and number, herbarium in which the voucher is located, and a clear annotation. that the material represents the voucher for the study in question. Nucleic acid or protein sequences corresponding to equal or greater than 50 nucleot ide 's are id into an appropriate data bank, e Bank/F The accession numbers are MU before es [Long sequences (exceeding two pages) will not be roulinely published. | Author accepts responsibility for establishing the accuracy of information provided. 12. Keys are clear and have been checked carefully for consistency with the descriptions. Leads of each couplet are parallel. Dichotomous keys are indented. Infraspecific taxa are keyed separately, not in species 13. Literature Cited The Literature Cited contains full citations of all references cited in the t All entries in the Literature Cited are cited in the text. Spelling of author(s) name(s) and years of publication have been double-checked. All entries especially journal titles, accents, diacritical marks, have been verif ie d against original sources, and spelling in languages other than English. Periodicals are listed as follows: author's last name, initial(s Full ated as in B-P-H/S. Volume: pages. No parenthetical ). Year. title of article. Journal abbrevi- part numbers after volume numbers are given unless each part is paginated separately. this style is followed: For more than one author, authors last initial(s), second author's initi- al(s), last name & third author's initial(s), last name. name, Lal O O Books appear as follows: author's last name, initial(s) Year. Full Unabbreviated Title (edited by Editor), 3rd ed., Vol. 2. Publisher, For an article within a larger followed: Author(s). Year. Name of Pp. 00-00 in Name of the editor(s), Larger Work. Citations of work City of Publication. work, this style is Publisher, City of Publication. “in prep.," unpublished theses and dissertations, and similar references to inaccessible sources have been eliminated or kept to a minimum. They are not necessarily included in the Literature Cited. 4. Illustrations Electronic figures are labeled with the first aithor s last name, the first four letters of the taxon, and ‘ vig- . Figure2.tif”, The clearly aiea what type of file it is. urel.tif”, elc. file extension Scale bars appear on illustrations, photographs, and maps. Magnifications/reductions are not indicated in cap- Lions. Electronie scans may be pre-sized to fit either column width (2-5/8 in. or ca. 68 mm) or full-page width 140 mm) rl size for printed X 8-1/4 i Figures are numbered in ida ume in the order (5-1/2 in. or ca. illustrations is 5-1/2 X of their citation in the text. Parts of figures are labeled with capital letters. Photographs are prepared according to the printer's requirements. [See Guide to Digital Art, Allen Press through the Publication Submissions link on the opening page of http://www.mbgpress.org/. | Figures are grouped into composite plates when utted possible; edges of photographs are ab No stripping is inserted between plate or figure segments. [Printer will insert stripping.] Edges of figures are squared. Maps include reference to latitude and longitude and are bounded by a fine border. Scanning electron micrographs are free of conspicu- ous charging Axes on mon are all labeled. Captions provide all explanatory text. Captions are separate from other text, one paragraph for each group of figures, and following the style in current issues of the Annals. Symbols on maps are legible, and reduction has been taken into consideration. Author Index: Annals of the Missouri Botanical Garden Vol. 95! A Albach, D. C. Schneeweiss, F. some Numbers in Veroniceae (Plantaginaceae): Review 543-500 520-538 Delgado, H. Weiss- Chromo- .M.M. Martínez- MALA L: zgökce . Fischer. and Several New Counts, Al-Shehbaz, L A. see Yue et al. B Barnhart, M. C. see Serb € Barnhart. 248—261 Baum, D. A. see DeWitt Smith et al. 600—617 Bessega, C., H. E. Hopp & R. H. Fortunato. Phylogeny of Mimosa (Leguminosae: Mi Toward a > Mimosoidae): / Preliminary South Species Based on Chloroplast DNA Sequence, 579 Analysis of Southern American 567- ei ‘eller. Assumption O Analysis: Botta, S. M. see dun et al. Brooks, D. R. é E oa Studies in the Age of Complexity, 201-223 and Go Seek: What Does Presence 51-11 Burnham, R. J. Hide Mean in the Fossil Record?, + G ack ‘ll, D. R. Pollinator Shifts and the Origin and Loss of Plant Species, 264-27 Castro, M. A. see Vega et i 511—519 Chan, Y-H. see Knowles & Chan. 224—231 Chen, Z.-D. see Liu et al. 459—470 Christophel, D. C. see Li et al. 580—599 Conran, J. G. see Li et al. 580-599 Crepet, W. L. 34—42 Crepet, W. L. The Fossil Record of Angiosperms: Requiem or Renaissance?, 3-33 Crepet, W. L. € M. A. genomics Era: Introduction, 1-2 see Gandolfo et al. Gandolfo. Paleobotany in the Post- D Davis, E. Delgado, L. C. see Liu et al. 459—470 543-506 Denham, S. S. see Freie et al. 338— DeWitt Smith, S., S. J. Hall, P. R. Pollination Biology of. Sympatric 600— see Albach et al. 390 Izquierdo € D. A. Baum. Ron and Allopatric Andean Jochroma (Solanaceae), 617 DiMichele, Deep » ime, . & R. A. Gastaldo. Plant Paleoecology in ur J8 E Endress, P. The Whole Between Floral Architecture and Floral Organ Shape, and the Parts: Relationships Repercussions on the 101-120 and Their Fragmentary Floral Fossils, Interpretation of F Fischer, M. A. see Albach et al. 543-506 Fortunato, R. H. see Bessega et al. 567—579 A. M. Taxonomic Revision of Roldana (Asteraceae: Senecioneae), Southwestern U.S.A., Mexico, and Central America, 282-337 Funston, a Genus of the G Gandolfo, M. A. see E & > 1-2 Gandolfo, M. A., . Nixon & W. Fossils for Ad of vius Dating Models, 34— 42 a Crepet. Selection of see Ps he le & Gastaldo. 144-198 an et al. 494. tiae, H. A. ig, X. P see H Hall, S. J. see DeWitt Smith et al. 600-617 Hashimoto, G. see Watanabe et al. 199 Hendricks, J. R. see Hermsen € Hendricks. 72-100 Hermsen, E. J. & J. R. Hendricks. W(h)ither Studying Morphological Character Evolution in the Age Fossils? of Molec ian Sequences, 72-100 Hertweck, b L. see Pires & Hertweck. 275-281 Hoch, P. C. The Impact of Peter Raven on Evolutionary and TNAM ee in the 20th and 21st. Centuries: Introduction, 262— Hopp, H. E. see ene el a EA 561—579 l Izquierdo, P. R. see DeWitt Smith et al. 600-617 J Jia, Y. see Liu et al. 459—470 o 7 see Walanabe et al. | Knapp. \ Revision of the Solanum havanense Species G »roup imi New Taxonomic Additions to the Geminata Clade d Solanaceae), 405-458 ¿ Y-H. Chan. Resolving Species id nies — Knowles, L. of Recent p volutionary Po aia 224— Kuroda, C. see Pan et al. 487— ! Lev-Yadun, S. see Rothwell et al. 121-134 . H.-W. see Li et - 580—59 7 Je. J: 3 Conran, >. Christophel, Z.-M. Li, L. H.-W. Li. B. netic Relationships of the un Complex and Core Laureae (Lauraceae) : ~ ITS and ETS Sequences and i 580-56 Li, J.-H. see Yue et al. 520-536 Li, L. see Li et al. 580-599 ' 95(1) pp. 1-200, 95(2) pp. 201-404, 95(3) pp. 405—542 , 95(4) pp. 543-092. Volume 95, Number 4 2008 Li, Z.-M. see Li et al. 580-599 Liu, Y., Y. Jia, W. Wang, Z.-D. Chen, E. C. Davis & Y.-L. Qiu. Phylogenetic Relationships of Two Endemic Genera from East Asia: Trichocoleopsis and Neotricho- colea (Hepaticae), 459—47 . Tilney & M. van der Bank. African Endemic . 471-486 M Magee, A. R., B.-E. van Wyk, P. M n Taxonomic Revision of the South enus . dicen (Apiaceae, Saniculoideae) eles col, V. see Rogers et al. 391-404 e UN M. M. see Albach et al. 543—566 Mülgura de Romero, M. E. see Peralta et al. 338—390 Nagatani, Y. see Watanabe et al. pe Nakazawa, M. Nickrent, D. L. Nixon, K. C. see Gandolfo et al. : Nixon, K. C. Paleobotany, Evidenc 'e, and Mole cular Dating: ) see Watanabe et al. An Example from the Nymphaeales, 4 see Albach et al. 543-566 O Ozgúkce, F. Kuroda. P Pan, Y., S. Shi, X. Gong & C. A Natural Hybrid Between Ligularia paradoxa and L. duciformis (Aster- aceae, Senecioneae) from Yunnan, China, 487-494 R. Life History Patterns and Biogeography: An Parenti, L. Interpretation " Diadromy in Fishes, 232-247 Peralta, P., M. E. Múlgura de Romero, S. 5. Denham & S. M. Botta. xs e Género Junellia (Verbenaceae), 338—390 Pires, J. C. & K. L. Hertweck. A of Cytogenetics: vhídiss in Polyploidy and Chromosomal Renaissance Evolution, 275-281 Pool, A. A Review of the Genus Pyrostegia (Bignoniaceae), 95-510 -— see Liu et al. 459—470 Q Qiu, Y.-L. R Rogers, Z. S., D. L. Nickrent & V. Malécot. Staufferia and Pilgerina: Two New Endemic Monotypic Arborescent Genera of Santalaceae from Madagasca car, 391—404. Rothwell, C. W., H. Sanders, S. E. Wyatt € S. Lev-Yadun. A Fossil Record for Growth p The Role of Auxin in Wood Evolution, 121-13 Sanders, H. see Rothwell et al. 121-134 Serb, J. M. & M. C. Barnhart. Congruence and Conflict Between Characters ie Western Molecular and Reproductive When Assessing Biological Diversity in tl Fanshell Cyprogenia aberti (Bivalvia, Unionidae), 248— 261 Shi, S. see Pan et al. 487—494. Soejima, A. see Watanabe et par 199 Sun, H. see Yue et al. 520— RUE Fossils, and Northern aphy: The Role of Physiological Uniformitarianism, ae , P. M. see Magee et al. 471-486 | Tiffney, Hemisphere Biogeog Tilney V van der Bank, M. see Magee et al. 471—486 A Synopsis of the Genus Tachigali Northern. South an der Werff, (Leguminosae: Caesalpinioideae) in America, 618-660 van Veller, M. G. P. see Brooks & van Veller. 201-223 van Wyk, B.-E. see Magee et al. 471—486 Vega, A. S., Castro & F. O. Zuloaga. Anatomy and o in azilian Species of "uum Sect. Lorea (Poaceae, 19 Localization of Lipid Secretions ime] Panicoideae, Paniceae), 51 W Wang, W. see Liu et al. 459—470 Watanabe, K., T. Yahara, G. Hashimoto, Y. Nagatani, A. Soejima, T. Kawah Chromosome Numbers and Karyotypes in Asteraceae, 199 Weiss-Schneeweiss, H. see Albach et al. ara & M. Nakazawa. Erratum for 543—500 Wyatt, S. E. see Rothwell et al. 121-134 y Yahara, 2. see w atanabe et al. 199 Yue, J.-P., 1, J.-H. Li & L A. Al-Shehbaz: A Synopsis of an eor Solms-Laubachia (Brassicaceae), and the Description of Four New Species from Western China, Zhu, H. The Tropical Flora of Southern Yunnan, China, and Its i Rd e Affinities, 661—680 11-514 Zuloaga, F. ee Vega et al. 5 Subject Index: Annals of the Missouri Botanical Garden Vol. 95'? \ lenistus, 600—617 Actinodaphne, 580-599 Africa, 391-404. see also specific countries Annonales, 275—281 Apiaceae, 471—480 Archaeopteris, 121-134 Arctopus, 471—486 Argentina, 338-390 Asia, 391—404, 459-470, 601-680. see also specific countries Asteraceae, 282—337 B Bignonta ignea, 495—510 Bignonia tecomiflora, 495—510 Bignonieae, 495-510 Bolivia, 338-390 Brassica, 215-281 215-281 19 Brassicaceae, Brazil, 51 1-5 Amazon, 618—660 G Cacalia nutans, 282-337 Cacalia peltata, 282-337 Central America, 282-337 C nile 338-390 Sichuan, 520-538 487—494, 661—680 520—538 Chromolaena pedunculosa, 199 Cineraria plantanifolia, 282-337 Clarkia, 275-281 Colombia, 618-660 Cylicodaphne, 580—590 Cyprogenta, 248-261 Cyprogenia aberti, 248—201 Yunnan, Christolea, D Desideria, 520—538 Desideria baiogoinensis. ~ 538 Diostea filifolia, 3383 E Ecuador, 618—660 Eupatorieae, 199 p French Guiana, 618-660 G Guyana, 618—600 I lochroma, 600—617 Ipomopsis, 204—274 Ipomopsis tenuituba, 204—214 J Junellia, 338—390 sect. Dentium, 338-390 sect. Junelliopsis, 338—390 subg. Junellia, 338—390 subg. Thryothamnus, 338—390 Junellia aspera var. aspera, 338-390 var. longidentata, 338-390 Junellia caespitosa, 338-390 Junellia connatibracteata, 3: Junellia echegarayt, 338-390 var. cordifolia, 338-390 338-390 338-390 var. puberulenta, Junellia lavandulaefolia, Junellia longidentata var. glandulosa, 338-390 Junellia scoparia, 338-390 Junellia spathulata var. glauca, 338—390 var. grandiflora, 338-390 Junellia toninii var. mulinoides, 338-390 Junellia tripartita, 338—390 L Lagotis, 543—506 Laureae, 580—599 Laurus, 580-599 Leiospora, 520-538 Lepidolaenaceae, 459—470 Ligularia, 487-494. Ligularia duciformis, 481—494 Ligularia paradoxa, 487-494. Lindera, 580—599 sect. Aperula, 580-599 sect. e d 580—599 Litsea, 5980—5* M Madagascar, Malesia, 391— lalay eR, 061-680 Mexico, 282-33 561—5 79 391-404 Mimosa, sect. Batocaulon, 567-579 sect. Batocaulon ser sect. Calothamnos, 507—57¢ 'O5(1) pp. -200, 95(2) pp. 201-404, 95(3) pp. 405-542, 95(4) pp. 543-692, “The subject index was compiled from volume 95 abstracts. . Stipellares, 507-579 Volume 95, Number 4 2008 sect. Habbasia, 567-579 sect. Mimosa, 507-579 N Neolitsea, 580—599 Neotrichocolea, 459—470 Neotrichocoleaceae, 459—470 North America, 391—404. see also specific countries Nymphaeales, 43—50 Okoubaka, 391404 Onagraceae, 275-281 Paederota, 543—566 Paniceae, 511—519 Panicoideae, 511—519 Panicum, 511-519 sect. Lorea, 511—519 Panicum acicularifolium, 511—519 Panicum bahiense, 511—519 Panicum chnoodes, 511—519 Panicum cipoense, 511—519 Panicum durifolium, 511—519 Panicum euprepes, 511—519 Panicum molinioides, 511—519 Panicum poliophyllum, 511—519 Panicum trinii, 511—519 Panicum Pape 511—519 Parasassafras, 580-596 eru, 338-390, Ren Phaeonychium, 520—538 Phaeonychium jafrii, 520—538 Picrorhiza, 543—566 Pilgerina, 391—404 Piptadenia viridiflora, 567-579 Plantaginaceae, 543-566 Poaceae, 275-281, 511-519 ei 282-337 Psacaltopsis pinetorum, 282-337 Pul CENA 459—470 Pyrostegia, 495—510 Pyrostegia cinerea, 495—510 Pyrostegia ee 495-510 Pyrostegia venusta var. villosa, 495-510 Pyrularia, 391—404 R Roldana, 282-337 Roldana acutangula, 282-337 Roldana albonervia, 282-337 Roldana aliena, 282-337 Roldana ned NN 282-331 Roldana ehrenbergiana, 282—337 Roldana gilgii, 282-337 Roldana hartwegii var. carlomasonii, 282-337 var. durangensis, 282-337 var. subcymosa, 282-337 Roldana heracleifolia, 282-337 Roldana heterogama, 282-337 Roldana kerberi, 282-337 var. calzadana, 282—331 var. manantlanensis, 282—337 Roldana langlassei, 282—337 Roldana lanicaulis, 282—337 Roldana lobata, 282-337 Roldana petasitis, 282-337 var. cristobalensis, 282-337 var. oaxacana, 202— xd var. sartorit, 202—3: S Santalaceae, 391—404 Sclerolobium, 618—660 Sclerolobium froesii, 618—660 Sclerolobium radlkoferi, 618—660 Sclerolobium subbullatum, 618—660 Sclerolobium uleanum, 618-660 Scleropyrum Scrophulariaceae, 543—566 Senecio acerifolius, 282-337 Senecio brachyanthus, 282-337 Senecio canicidus, 282-337 Senecio chapalensis var. areolatus, 282-337 Senecio chrismarii, 282-337 Senecio ghiesbreghtii var. pauciflorus, 282-337 Senecio diae var. glabrior, 282-337 Senecio et 282-337 Senecio E 282-337 Senecio lobatus, 282-337 Senecio RU 282—331 Senecio prainianus, 282-337 Senecio rotundifolius, 282—337 Senecio schumannianus, 282- à: 7 sect. antes 405-458 sect. d 405-458 hylla, 405—458 P oun acropterum, 405—458 Solanum argentinum, 405—458 Solanum chalmersii, 405—458 Solanum conocarpum, 405—458 Solanum elvasioides, 405—458 Solanum evonymoides, 405—458 Solanum havanense, 405—458 Solanum hookerianum, 405—458 olanum humboldtianum, 405—458 Solanum monanthemon, 405—458 Solanum myrtifolium, 405—458 Solanum naucinum, 405— Solanum pseudodaphnopsis, 405—458 Solanum sagittantherum, 405—458 Annals of the Missouri Botanical Garden Solanum sumacaspi, 405—458 Solanum troyanum, 405—458 Solms-laubachia, 520—538 Solms-laubachia angustifolia, 520-538 Solms-laubachia batogoinensis, 520-538 Solms-laubachia calcicola, 520-538 Solms-laubachia flabellata, 520-538 Solms-laubachia grandiflora, 520-538 Solms-laubachia haranensis, 520-538 Solms-laubachia himalayensis, 520-538 Solms-laubachia incana, 520-538 Solms-laubachia jafrii, 520 Solms-laubachia lanata, 5: Solms-laubachia linearifolia, 520-538 Solms-laubachia linearis, 520-538 Solms-laubachia mieheorum, 520-538 Solms-laubachia mirabilis, 520-538 Solms-laubachia nepalensis, 520-538 Solms-laubachia prolifera, 520-5: Solms-laubachia stewartii, 520-538 Solms-laubachia lcd ean 520-538 South America, 338-390, 495-510, 567—579, 618—600. see also specific countries Staufferia, 391—404 Suriname, 618—660 T Tachigali, 618-600 Tachigali alba, 618-600 Tachigali barnebyt, 618-060 Tachigali bicornuta, 618—600 Tachigali bracteolata, 618-660 Tachigali bracteosa, 618—660 Tachigali candelabrum, 618-060 Tachigali cenepensis, 018-660 Tachigali chrysaloides, 618—600 Tachigali ferruginea, 618—600 Tachigali formicarum, 618-060 Tachigali fusca, 018-660 Tachigali glauca, 618-660 Tachigali guianensis, 618—600 Tachigali inconspicua, 618-660 Tachigali loretensis, 618—600 Tachigali melanocarpa, 618—600 Tachigali myrmecophila, 618—660 Tachigali paniculata, 618—660 Tachigali plumbea, 618-660 Tachigali poeppigiana, 618-660 Tachigali pulchra, 618-6060 Tachigali reticulosa, 018-660 Tachigali richardiana, 618-660 Tachigali rusbyi, 018-000 Tachigali sulcata, 618-060 Tachigali Tachigali ulei, 618—600 Tachigali vaupesiana, 018-0660 Thailand, 661—680 Tomingodaphne, 580-599 essmannil, 618—600 y Trichocoleaceae, 459—470 Trichocoleopsis, 459—470 Tynanthus igneus, 495—510 U Unionidae, 248—261 United States, 282-337 V Venezuela, 618—660 Verbena, 338—390 sect. Junellia, 338—390 sect. Verbenaca, 338—390 ~ Erinaceae, 338—390 n Minutifoliae, 338-390 Pauciflorae, 338-390 re Seriphioideae, 338-390 ‘r. Seruoguiudeae, 338-390 r. Thymyoltae, 338-390 ser. Verbenaca, 338-390 Verbena asparagoides, 338-390 Verbena bisulcata, 338—390 Verbena cole a 338-390 L Y eA VA v2 L Verbena comberi, 338—: Verbena nrc oA oe Verbena digitata, 338-390 Verbena dolichothyrsa, 338-390 Verbena juniperina, 338-390 Verbena ligustrina, 338-390 Verbena lorentzii, 338-390 Verbena minima, 338-390 Verbena ourostachya, 338-390 Verbena pygmaea, 338-390 Verbena scoparta var. puberula, 338-390 Verbena selaginoides, 338—390 Verbena spathulata var. grandiflora, 338-390 var. parviflora, 338-390 Verbena thymifolia, 338-390 Verbena tridactylites, 338-390 Verbena tridens, 338—390 Verbena uniflora var. glabriuscula, 338-390 Veronica, 543—566 secl. Acinifolia, 543-506 sect. Glandulosae, 543-500 subsect. Cochlidiosperma, 543-500 Veronica hispidula, 543-560 Veronica reuteriana, 543—560 Veronicastrum, 343-5006 Veroniceae, 543-506 y Wulfenia, 543-5 Wulfentopsis, ae 506 X Xishuangbanna, 661—680 . 1-200 of Axxars or THe Missouri BOTANICAL GARDEN was published on 11 April 2008 . 201—404 of ANNA LS or THE Missourt BOTANICAL GARDEN was published on 18 June 2008 . 405—542 of Anxars or rite Missouri BOTANICAL GARDEN was published on September 23 543-092 of ANNALS or THE Missou rt BOTANICAL GARDEN was published on 30 December 200€ /olume 95, Number |. pp Vol 95, Number |, pj Volume 95, Number 2, pp Volume 95, Number 3, pp Pl Volume 95, Number 4, pp. . 2008. go WwW W WwW "ib opress or a CONTENTS Chromosome Numbers in Veroniceae M Review and Several New Counts D. C. Albach, M. M. Martínez-Ortega, L. Delgado, E H. Weiss-Schneeweiss, E. Ozgúkce & M. A. Fischer Toward a TEIRA of Mimosa ERES Mimosoidae): A Preliminary Analysis of Southern South American Species Based on P LE DNA Sequence. C. Bessega, H. E. Hopp & R. H. For ortunato Phylogenetic Relationships of the EE D and Core Laureae (Lauraceae) Using ITS and ETS Sequences and Morphology — — Jie Li, John G. Conran, David C. (MM Zhi- Ming Li, Lang Li & Hsi-Wen Li Comparative Pollination Biology of Sympatric and Allopatric Andean /ochroma (Sola- naceae) Stacey DeWitt Smith, Steven J. Hall, Pablo R. Izquierdo & David A. Baum A Synopsis of the Genus Tachigali (Leguminosae: Caesalpinioideae) in Northern South America | — Henk van der Werff The Tropical Flom of cem P NER Chine. add Its Biossomathis Affinities. Zhu Hua Acknowledgment of Reviewers Checklist for Authors Author Index Subject lider yi 5 543 Cover illustration. Archaefructus reconstruction by David L. Dilcher and K. Simons, Florida Museum of Natural History.