Annals of the Missouri Botanical Garden 1996 YY 9 ume 8 mber 1 Volume 83, Number 1 Winter 1996 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 out- side the Garden will also be accepted. 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. Editorial Committee Henk van der Werff Editor, Missouri Botanical Garden Amy Scheuler McPherson Managing Editor, Missouri Botanical Garden Diana Gunter Editorial Assistant, Missouri Botanical Garden Vicki Couture Secretary Teresa Johnson Publications Order Processor Ihsan A. Al-Shehbaz Missouri Botanical Garden Gerrit Davidse Missouri Botanical Garden Roy E. Gereau Missouri Botanical Garden Peter Goldblatt Missouri Botanical Garden Cordon McPherson Missouri Botanical Garden P. Mick Richardson Missouri Botanical Garden For subscription information contact Depart- ment Eleven, P.O. Box 299, St. Louis, MO 63166-0299. Subscription price is $110 per volume U.S., $115 Canada and Mexico, $135 all other countries. Four issues per volume. The journal Novon is included in the subscription - price of the ANNALS. scheuler@mobot.org (editorial queries) dept] 1@mobot.org (orders) © Missouri Botanical Garden 1996 The ANNALS OF THE MISSOURI BOTANICAL GARDEN (ISSN 0026-6493) is published quar- terly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. - Second class postage paid at St. Louis, MO and additional mailing offices. POSTMASTER: Send address changes to ANNALS OF THE MISSOURI Y BOTANICAL GARDEN, Department Eleven, PO. Box 299, St. Louis, MO 63166-0299. 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 239.48-1992 (Permanence of Papen). Volume 83 Number 1 1996 Annals of the Missouri Botanical Garden Y THE SYSTEMATICS AGENDA 2000 SYMPOSIUM: INTRODUCTION! P. Mick Richardson? The fortieth anniversary of the first Systematics Symposium was appropriately celebrated with a meeting on the subject of Systematics Agenda 2000, a published report (Anon, 1994) on a global initiative to discover, describe, and classify the world’s species. Jay Savage carefully selected an array of speakers that would give the audience an overview of the diversity of systematic studies and the impact of these on the study of evolution, con- servation, food supplies, disease sources, and dis- ease prevention. The symposium began with a presentation by Brent Mishler, “Speciation, Adaptation, and Evo- lution,” that explained why sound phylogenetic frameworks are needed to formulate general evo- lutionary theories concerning basic processes such as adaptation, speciation, and extinction. This con- tribution is not published here (although see Mish- ler (1995) for further references), but the remaining six papers were submitted as manuscripts for pub- lication. Russ Monson delivered a paper entitled “Systematic s and Plant Function in a Changing En- vironment,” which illustrated the use of phyloge- netic reconstruction to test comparative hypotheses in plant physiology and developmental biology. Amy Rossman’s presentation, “Wood-rotters, Leaf- hoppers and Systematics,” coauthored by Douglass Miller, explored the role of some of the less con- spicuous but specious organisms, such as insects, fungi, nematodes, and bacteria, and their benefi- cial, neutral, and detrimental interactions with ag- ricultural and forestry systems. Mike Vecchione’s report, “Fisheries, Systematics and Marine Diver- sity,” coauthored with Bruce Collette, explained how systematists must disseminate their knowledge if it is to have practical effects in, for example, fisheries management. James Oliver's talk, “Ticks, Microbes and Disease,” sent shivers down my spine as I recalled my unpleasant experience with Lyme disease some years ago. He informed the audience of the increasing costs of tick-borne diseases and the necessity of correct identification of the tick species and the disease organisms if a rational plan of treatment and control is to be developed. A shortage of younger taxonomists in this field is a potential problem. The daytime session was brought ! This and the six articles that follow it are the proceedings of the 41st Annual Systematics Symposium of the Missouri Botanical Garden, Systematics Agenda 2000: vici s and Society. The symposium was held 30 September- l E se 1994 at the Missouri Botanical Garden in St. mposium was supported in part by ab vM. Savage, Missouri, U.S.A the National pepe Foundation under grant number BSR89-18138. 1 of the University of Miami. for selecting a fine diversity of speakers, Carol Davit for indispensable logistice al help in — the symposium, John Myers for his fine illustration "or the symposium brochure, and the rous systematists for being such an interestin me ?Missouri Botanical Carlen, P.O. Box 299, St. and diverse group. Louis, Missouri 63166, "USA. ANN. Missouni Bor. GARD. 83: 1-2. 1996. Annals of the Missouri Botanical Garden to a close by Dick Vane-Wright with a paper enti- tled “Death and the Compass: Systematics for Sav- ing Life.” Again, we saw how natural taxonomic hierarchies were the best methodology to use, in this case for determining priorities for the conser- vation of living things. Finally, the after-dinner talk, entitled “Ethnobotany for the Nineties” how the published paper was transformed for the (but see new millenium), was presented by Mike Balick; it included a fine set of examples of current ethno- botanical projects that illustrate the public’s grow- ing awareness of this branch of biological science. This was a fitting end to a stimulating meeting. I hope that you will enjoy reading the published pro- ceedings that follow. Literature Cited Anon. 1994. Systematics Agenda 2000: Charting the Bio- sphere. Technical Report. Obtainable from SA2000, He cann New us Botanical Garden, Bronx, New York 10458, U.S Mishler, B. | i. Plant systematics gi conservation: Science and society. Madroño 42: 103— 3. THE USE OF PHYLOGENETIC PERSPECTIVE IN COMPARATIVE PLANT PHYSIOLOGY AND DEVELOPMENTAL BIOLOGY" Russell K. Monson? ABSTRACT The use of phylogenetic reconstruction for the testing of comparative hypotheses is a recent ata nop in the fields lant physiology and developmental biology. In this review, several uses o including (1) the use of des trees for choo of compare phylogenetic information are discussed sing experimental systems, (2) justifying the statistic zi iun dence ed taxonomic groups, (3) identifying exaludianary direction using outgroup analysis, and (4) s tudying the ric iei tempo of physiological and developmental change. Difficulties are also discussed, especially with respect to (1) mapping continuous, physiological traits onto the discrete, binary structure of the physiological and developmental characters being mapped, and (3) re- trees from traits that are independent specting the statistical nature of phylogenetic trees. The literature that illustrate the power of a phylogenetic perspec of phylogenetic trees, (2) constructing ulk of the review is devoted to several examples from the ‘tive in pri di studies. Reviewed mns include ycolytic pathway of phosphate-stressed n in the origins of eben piae. ui in angiosperms, (3) adaptive radiation of Hawaiian Island plants into dry habitats, and (4) evolutionary pattern in the origins of C, photosynthesis. The fields of systematics reliance on phylogenetic perspective increase s, comparative PEE and developmental biology will continue to merge as their common The purpose of this review is to describe aspects of the past, present, and future relationship be- tween two historically disparate disciplines—com- parative plant systematies and comparative plant physiology and development. Researchers in plant systematics focus on the pattern of evolutionary di- versification. Researchers in plant physiology and plant developmental biology focus on the products of evolutionary diversification—functional mecha- nisms, structural attributes, and their relationships to genetic and environmental controls. The nexus among these disciplines occurs through a need to understand the role of evolutionary constraint in dictating patterns of adaptation, the timing and rate of phenotypic evolution, and the evolutionary in- teractions among phenotypic traits. Historically, one can recognize several connec- tions between the fields of comparative systematics and comparative functional biology. As examples, consider the numerous studies that originated in Sweden with Turreson (1922) and in the United States with Clausen et al. (19 948 of these studies was firmly embedded in the evo- he focus lutionary process: to uncover the cohesive and di- vergent forces that maintain the structure of taxo- nomic units while at the same time allowing for diversification and speciation. A principal compo- nent of these early studies included comparisons among plants from different geographic locations that were grown in common environments. The goa pursued in these “common-garden” studies was to partition the influence of genetic and environmental influences on character expression. During the birth of modern comparative physiology, and to a lesser extent developmental biology, scientists dis- mon-garden approach vironment, while observing inherited differences in the adaptive responses of plants (e.g., Mooney & Billings, 1961; Bjórkman & Holmgren, 1963). Fur- ! The author is grateful for the LM A comments, discussions, and unpublished material ine by Ray Huey, uel I Michael E 2 ,erdau, William Bowman, Suzanne Warw EN Dawson, and E Olmstead cl their helpful collaboration in providing t leaders of t lea Agenda Annual S ane 0004960 and BSR. 891 Ned Friedman, Richard Olmstead, Robert Teese Special thanks are due to Michael Grant, sien Farrell, and Richard the Flaveria phylogeny. Thanks are also due to the sponsors and 000 effort and the Organizing Committee of the Missouri Botanical Garden’s 41st ty ae The author's research discussed in this paper was supported by obichaux, Bill Plax ton, NSF grants BSR- artment E HE Population, and Organismic Biology, University of Colorado, Boulder, Colorado 80309-0334, U.S ANN. MISSOURI Bor. GARD. 83: 3-16. 1996. Annals of the Missouri Botanical Garden thermore, by choosing congeneric and confamilial groups for their comparisons, these mechanistic bi- ologists were using phylogenetic perspective as they observed the products of common descent, identified examples of convergent evolution, and isolated patterns of adaptation. Thus, at its incep- tion, modern comparative plant physiology had de- finitive links with comparative plant systematics. Unlike researchers in however, the focus of comparative physiologists was not the evolutionary process in relation to natural rather the nature of mecha- comparative systematics, taxonomic units, but nistic adaptations and their relationship to environ- mental extremes. Since these pioneering studies, researchers in physiology and developmental biology have strayed from their phylogenetic, comparative origins, em- bracing instead the implicit assumption that func- tional responses to the environment are the prod- When it comes to interpreting evolutionary patterns in functional traits, formal hypothesis testing has often given way to intuition and adherence to an adaptational doc- trine. This tendency has been criticized in past es- says (e.g., Gould & Lewontin, 1979). The potential contributions of phylogenetic reconstruction and the comparative method to formal hypothesis test- in physiological and developmental biology have only recently been revisited (e.g., Felsenstein, 1985; Huey, 1987; Harvey & Pagel, 1991; Garland & Carter, 1994; Lauder et al., 1995). There is little doubt that as more biologists become aware of the useful deductions that can be made using phylo- genetic data, a remarriage will be witnessed be- ucts of natural selection. tween the fields of comparative systematics and comparative physiology and development. ON THE DIFFERENT NATURES OF COMPARATIVE SYSTEMATICS AND COMPARATIVE PHYSIOLOGY/DEVELOPMENT In comparative systematics, trait variation is used to study natural relationships among different groups of organisms. In comparative physiology and development, natural relationships among organ- isms are used to study the nature and functional significance of traits. Thus, although researchers in both fields draw upon the same puzzle pieces (vari- ation in organismic traits), those in one discipline use the pieces to assemble the puzzle, while those in the other discipline use the completed puzzle to study the pieces. In focusing on adaptation, comparative physiol- ogists ask the question: does physiological variation among organisms show a pattern that is consistent with the process of natural selection (Feder, 1987)? Pattern is typically measured as the functional at- tributes of organisms native to different environ- ments. If the pattern correlates with enhancement of growth and persistence across an environmental gradient, then the pattern is ascribed to natural se- lection and is taken as adaptive. Recent studies, however, have revealed phylogenetic history to be as likely a constraint on pattern variation as is nat- ural selection (Huey, 1987; Garland & Carter, 1994; 1995). Disentanglement of phylogeny and selection is only accomplished through reliance on comparative systematics and the existence of accurate phylogenetic reconstruc- Lauder et al., tion. The integration of comparative systematics with comparative physiology and development is an exercise in mapping functional traits onto estab- lished phylogenetic trees. Several methods have been developed to accommodate such mapping and the accompanying statistical analysis (Felsenstein, 1985; Harvey & Pagel, 1991; Brooks & McLennan, 1991). There are some potential pitfalls that must be recognized before one can successfully engage in the activity of phylogenetic mapping. For example, many comparative physiologists and developmental biologists may not recognize the tentative nature of phylogenetic trees. At first appearance, such trees reflect firm, definitive relationships, especially from the perspective of mechanistic biologists who are accustomed to observing the discrete conclusions of manipulative experiments. It is important, how- ever, to recognize the statistical nature of phylo- genetic trees. A phylogenetic tree is essentially the solution of greatest parsimony, given knowledge about a particular set of measured, shared traits among an assumed set of related taxa. In many cases a tree constructed from assumptions of max- imum parsimony will poorly reproduce the alleged phylogeny (Fiala & Sokol, 1985; Rohlf et al., 1990; Lamboy, 1994). Phylogenetic trees have conclusive confidence limits that should be honored when framing conclusions about evolutionary patterns. As a second example of issues dealing with phy- logenetic mapping, comparative physiologists and developmental biologists must recognize disparities in the nature of traits that are frequently studied. Comparative systematics is based on the use of dis- crete, binary traits, the raw material of cladistic construction. Comparative mapping of physiological and developmental traits also requires that the have clearly delineated limits. There are examples of such traits (e.g., the presence or absence of the C, photosynthetic pathway), though most physiolog- ical traits, and to a lesser extent developmental Volume 83, Number 1 Monson 5 Phylogeny and Comparative Plant Biology “YO 20r A max N ; N 18 F E 18 [p --------------2 H E =. w g n€————— ÁN à (U | 2 o o P an «+ (= > o O + Iz l à L L L i S i | fi J Q 1000 1200 1400 1600 1800 2000 2200 Photon flux density (umol m^ s`!) Figure 1. nvironment and the pas Hypothetical responses of photosynthesis rate to incident phot on flux density in a high-light grown leaf af (L). The results posees variability in photosynthesis rat e as a function of the t (or growth) light environment. One commonly used index of photosynthesis rate in studies of genotypic differenc ‘es is the maximum photosynthesis rate (Ama). traits, are continuous and show plasticity in their responses to the environment. This leads to diffi- culties in the process of mapping. As an example consider the case for photosynthesis rate, one of the most commonly studied characters in comparative physiology. Within a plant, maximum photosynthe- sis rates can vary considerably (by up to a factor of 10) among leaves that develop in different parts of a plant’s canopy (e.g., Fig. 1). By studying the cause of such variation comparative physiologists can gain insight into interactions between the pho- tosynthetic process and a plant’s environment, in this case the light environment. Yet this same vari- ation creates difficulties in assigning discrete char- acter values for phylogenetic mapping. In some way, the influence of environment on character expression must be standardized. Phylo- genetic reconstruction is only meaningful if the mapped traits reflect genotypic comparisons with- out environmental influence on character variabil- ity. One way this is commonly accomplished in studies of comparative physiology is to compare maximum observed values for some process (e.g., Ana in Fig. 1). It is reasoned that if the expression of a trait is measured at its maximum for all com- pared taxa, then variability in reflect genotypic differences. In practice, difficul- the measurements ties arise in determining the species- and trait-spe- cific conditions that foster maximum expression of a trait. For some traits (e.g., resource acquisition rates) maximum expression may come with envi- ronmental conditions that maximize growth, where- as with other traits (e.g., mechanisms of stress tol- erance) it may come under environmental conditions that minimize growth. One final issue to deal with in terms of phylo- genetic mapping is the need to construct phyloge- netic trees from traits that are not mechanistically linked to those being mapped. This concern is min- mized when trees are constructed from molecular ios (though one can imagine gene sequences in which molecular variability is reflected in en- zymatic variability). Trees based on morphological traits carry greater risk in this respect. Lack of in- dependence between traits used in tree construc- tion and those mapped onto the tree would decrease the likelihood of reproducing the true phylogeny for the mapped trait. Annals of the Missouri Botanical Garden ON SPECIFIC IssUES THAT CAN BE ADDRESSED UsiNG A PHYLOGENETIC PERSPECTIVE PHYLOGENY AND THE CHOICE OF EXPERIMENTAL TAXA Phylogenies can be used to provide direction in the choice of experimental systems—allowing for the choice of taxa that are closely related or more distantly related. This allows investigators to max- imize the possibility that the traits being compared are the products of common descent in the case of closely related groups, or convergent evolution in the case of distantly related groups. Through the a priori choice of groups with contrasting phyloge- netic patterns, investigators can study the relative roles of phylogenetic inertia versus selection in in- fluencing the evolution of functional traits. Traditional choices of experimental systems in comparative functional biology have been driven by 1987; Lauder 1995). With respect to plants, the paradigm considerations of environment (Huey, et al., of choice has stated that adaptive responses to en- vironment are best observed in extreme habitats. Thus, adaptive responses to high temperature are best studied in hot, desert habitats, whereas adap- tive responses to cold temperatures are best studied in cold, tundra environments. This approach con- tains a compelling dose of intuition and logic. How- ever, it also perpetuates some dangerous and short- sighted assumptions. One principal assumption is that any intuitively beneficial aspect of a functional attribute is the product of selection in response to the most common, and potentially stressful, envi- ronmental extreme. A corollary to this assumption is that, with respect to influences on phenotype, selection in a plant's current environment has over- shadowed historical events in the acquisition of traits from past ancestors. These assumptions feed the adaptationist program discussed above. This current environment-oriented approach has little provision. for phylogenetic constraint or inertia. iis is where a phylogenetic perspective can con- tribute to adaptive analysis. By choosing groups based on criteria of both environment and the avail- ability of phylogenetic reconstructions, one can conduct truly synthetic analyses of (1) the relative roles of selection versus phylogenetic constraint, and (2) the relationship of traits to the current ver- sus past environment. PHYLOGENY AS A GUIDE TO STATISTICAL ANALYSIS Phylogenetic relationships can be used to justify statistical patterns among experimental groups be- ing compared. It is an unattainable goal for com- parative biologists to obtain groups that are com- pletely independent in terms of past history. There is always going to be some degree of common, hi- erarchical descent in a system characterized by a monophyletic cladogram. Phylogenetic connections interfere with statistical assumptions of indepen- dence (Clutton-Brock & Harvey, 1984; Felsenstein, 1985; Martins & Garland, 1991). Basically, this is “degrees of freedom” issue that becomes es- enis relevant in attempts to construct correla- tions among co-occurring traits or relate the ex- pression of a trait with environmental variability. By ignoring phylogenetic relatedness, investigators tend to inflate the potential for Type I errors (wrong- ful rejection of the null hypothesis), reduce the power of their statistical conclusions, and increase inaccuracies in estimating correlation coefficients (Martins & Garland, 1991). The only groups that appear to be immune from such phylogenetic influ- ences are those characterized by early diversifica- tion followed by long periods of phylogenetic stasis (so-called "star" phylogenies in Felsenstein, 1985) (see Martins & Garland, 1991) Several methods have been proposed to deal with the problems of phylogenetic interdependence. The simplest method, though not the most accurate, is to center the analysis on higher taxonomic units, which will presumably possess weaker phylogenetic interdependence in terms of the traits they exhibit Crook, 1965; Clutton-Brock & Harvey, 1984). This method does not completely deal with the issue of —. phylogenetic connections, however, since even the higher levels will be interdependent to some de- gree. An approach that deals more directly. with phylogenetic connections was first proposed by Fel- senstein (1985) and modified in various ways by others (e.g., Huey & Bennet, 1987; Grafen, 1989 In essence, this approach takes advantage of ex- — b pected variance of character change to compute standardized, independent contrast values from the measured comparative data. Such contrast values can then be subjected to standard statistical tests of significance. Even with this approach, however, uncertainties exist due to limited knowledge about phylogenetic branch lengths and their relationship to the rate and pattern of evolution—specifically, whether trait diversification has occurred in a grad- ual or punctuated pattern. (Martins & Garland, 1991) PHYLOGENY AND THE DETERMINATION OF EVOLUTIONARY PATTERN Phylogenetic trees represent maps from which one can polarize patterns of adaptive diversifica- tion. Polarization involves determination of the se- Volume 83, Number 1 Monson 1996 Phylogeny and Comparative Plant Biology Water-use efficiency 1 2 3 4 5 6 Species outgroup (W) outgroup (W) D igure 2. Hypothetical water-use efficiencies measured for six different plant species (upper panel). The lower two trees yee trate how evolutionary patterns in this physiological character can be polarized with respect to ancestral and derived traits using outgroup comparison. In both cases the outgroup is assumed to occur in wet habitats (W). s i exhibit higher water-use efficiencies occur in dry habitats (D). In the tree on the left, a single transition is found t occur as species radiated into dry habitats and evolved higher water-use efficiencies. In the tree on the right, a transition has occurred early during diversification of the group, followed by a later reversal. quential origins of traits relative to patterns of tax- between, or among, multiple traits. By simulta- onomic diversification. The origins of traits are neously mapping two or more traits, one can gain polarized through comparison with ancestral out- insight into how the evolutionary appearance of one groups. Without outgroup polarization, it is impos- — trait may have influenced the appearance of a dif- sible to determine whether a trait represents the ferent trait. Alternatively, the appearance of traits ancestral or derived state. This use of phylogeny is can be mapped simultaneously with major ecolog- illustrated in Figure 2 for a hypothetical study of ical shifts (e.g., shifts in habitat type, phenology, or photosynthetic water-use efficiency. Without phy- community structure). With respect to plants, this logenetic analysis, it is not possible to determine approach may be particularly useful in deducing whether high or low water-use efficiency represents the relationships between functional and life-his- the ancestral state in this group of species. Two tory traits, as well as between functional and struc- alternative phylogenies are presented, each of tural traits. leads to a different conclusion concerning The use of phylogenetic trees for the study of ihe es pattern in this trait. evolutionary patterns represents the most funda- In addition to polarizing single traits, outgroup mental of all symbiotic connections between the analysis and phylogenetic mapping also provide a disciplines of comparative systematics and com- means for studying the evolutionary interactions parative functional biology. This is the connection Annals of the Missouri Botanical Garden that is destined to draw the most attention from comparative physiologists and developmental biol- ogists. As is evident in the following sections, this approach has great power to provide new insight into the processes and patterns of functional ad- aptation. PHYLOGENETIC RECONSTRUCTION AND THE DETERMINATION OF EVOLUTIONARY TEMPO With appropriate calibration, phylogenies can be used to assess rates of evolution. Rates of evolution can provide insight into patterns of adaptive radi- ation and their underlying processes such as nat- ural selection and genetic drift. Temporal calibra- tion of phylogenies can be accomplished through paleontological dating, the use of molecular clocks (e.g.. dePeer et al., 1993), or the application of sta- tistically based trajectory models to the variance among measured characters (e.g.. Lande, 1985; Lynch, 1990; Martins, 1994). The use of dated phy- logenies to derive temporal patterns of evolution in functional traits has not been great. There are sev- eral complexities that still need resolution before such use is likely to increase. For example, molec- ular clocks are typically based on neutral traits. However, it is likely that most functional traits have, at some time in their past, been the subject of selection—a process that may have accelerated evolutionary change relative to neutral markers. Additionally, there themselves and their application to phylogenetic These uncertainties include hetero- are uncertainties in the clocks reconstruction, geneity in the rate of nucleotide base substitution depending on the DNA environment (Saccone et al., 1989), errors due to interactions between the choice of nucleotide sequence and the method of tree reconstruction (Zharkikh & Li, 1993), and the influence of certain functional traits (e.g., genera- tion time and metabolic rate) on nucleotide substi- tution rates (Martin & Palumbi, 1993). If such com- plexities can be resolved, evolutionary clocks could become valuable tools as comparative biologists ad- dress the issue of future environmental change and its influence on evolutionary patterns in the Earth’s biota ON THE Usk OF SYSTEMATIC INFORMATION SYSTEMS FOR THE DISCOVERY AND INVENTORY OF UNIQUE BIOCHEMICAL AND PHYSIOLOGICAL PROCESSES Phylogenetic trees provide maps of historical linkages among groups of organisms. As with any map, phylogenetic trees can provide direction and orientation for searches of unique functional pro- cesses. It is not uncommon to find investigators screening an entire genus to determine the extent of distribution in some functional trait or identify possible comparative systems in which the trait dif- fers. Such an approach has been taken, for exam- ple, in the identification of Chlorella species pos- sessing the ability to utilize bicarbonate as an inorganic carbon source (Miyachi et al., 1985) and the assessment of phytochelatin (heavy-metal bind- ing peptides) distribution within various groups of plants (Gekeler et al., 1989). Recognition of phy- logenetic pattern in the distribution of phytochem- icals has existed for many years. In fact, phyto- chemical distribution represents the foundation for chemosystematics, a central discipline within the broader field of comparative systematics. In many instances the chemicals of systematic interest also have roles in plant adaptation. Terpenes and alka- loids, for example, exhibit strong phylogenetic af- finities and serve an adaptive role in deterring her- bivory (Banthorpe € Charlwood, 1980; Lerdau et al., 1994). Quaternary ammonium and tertiary sul- fonium compounds are distributed along conserva- tive phylogenetic lines and have important roles in plant responses to salinity and water stress (Rhodes & Hanson, 1993), examples of phytochemicals that have dual importance to com- Numerous other parative systematics and comparative physiology can be described (e.g., flavonoids, cyanogenic gly- cosides). In all these examples the systematic ap- proach has played an obvious role in expanding the list of species that possess the chemical of interest and in establishing correlations between taxonomic and ecologic distributions. Ine recent case in which phylogenetic knowl- edge has had an obvious influence on trait discov- ery concerns the inducible pyrophosphate-depen- dent phosphofructokinase that was recently reported in Brassica nigra (L.) W. D. J. Koch (Theo- 1993, 1994). Theodorou and 1994) suggested that this enzyme confers dorou & Plaxton, Plaxton an adaptive advantage under phosphate stress by a allowing glycolytic processing of fructose 6-phos- phate in the face of reduced ATP availability. Fol- lowing a protocol typical of comparative physiolo- gists, Plaxton’s group recently screened other plant species to assess overall distribution of the induc- ible enzyme (W. Plaxton, Queen’s University, pers. comm.). Their search revealed its presence at rel- atively high constitutive levels in tobacco and to- mato cells, but the absence of phosphate-stress in- ducibility. A similar pattern of constitutive presence, without inducibility, was found in other Brassica species thought to be closely allied with B. nigra (e.g., B. oleracea L. and B. rapa L.). Thus, Volume 83, Number 1 1996 Monson Phylogeny and Comparative Plant Biology it appeared that the induction mechanism was iso- lated to B. nigra, being absent even from conge- ners. At this point Plaxton’s group initiated discus- sions with Suzanne Warwick, a systematist who had recently constructed a phylogenetic tree of Brassica and related taxa through the use of cpDNA markers (Warwick & Black, 1993). From the molecular phy- logeny it was clear that B. nigra is more closely allied with three members of the genus Synapis, than Brassica (Fig. 3). Plaxton’s group has recently surveyed the two species of Synapis, S. alba L. and S. arvensis L., which appear to share close affinities with the re-aligned B. nigra. Both Synapis species exhibit the inducible pyrophosphate-dependent phosphofructokinase. This example provides three important conclu- sions concerning the use of phylogenetic trees in studies of comparative physiology and developmen- tal biology. First, it is clear that phylogenetic trees can be useful in directing the search for novel bio- chemical, physiological, and developmental pro- cesses. Second, this example illustrates the poten- tial two-way exchange of information in collaborations between systematists and functional biologists. In this case the distribution of the pyrophosphate-de- pendent phosphofructokinase provides independent support for the taxonomic realignment of B. nigra as suggested by the molecular phylogeny. Finally, it should be clear that the effective use of system- atic information systems by comparative physiolo- gists and developmental biologists requires that taxonomy reflect phylogeny. The fact that B. nigra was traditionally classified with Brassica, rather than with Synapis, caused this biochemical search to stray from the correct phylogenetic path. ne area in which a phylogenetic approach is currently missing, but potentially beneficial, is the study of herbicide resistance in weedy, agricultural pests. In the past 25 years it has become obvious that the continued use of herbicides on agricultural fields has resulted in the evolution of several her- bicide-resistant weed species (Warwick, 1991; Holt et al., 1993). Apparently, there is considerable in- terspecific and interpopulation variability in the likelihood of evolving herbicide resistance. It is un- known to what extent such variability is due to dif- ferences in selection regime, population genetic structure, or phylogenetic constraint. A systematic approach would be of obvious benefit in partition- ing the influences of selection versus phylogeny, though to date no such approaches have been at- tempted. An example of the potential use of com- parative systematics to the question of herbicide resistance can be seen in the case of triazine re- sistance in Amaranthus. Within North America nu- merous cases of triazine resistance have been re- ported in A. powellii S. Watson and A. hybridus L., the two less widespread species of North American weedy amaranths (Hill, 1982). Triazine resistance has only been reported in one population of the more widespread species, A. retroflexus L. Although the data are incomplete, there is reason to hypoth- esize that the two minor species will usurp the eco- logical dominance of the widespread species in the face of continued heavy herbicide use (see Gressell & Segal, 1982). This issue is complicated by the fact that European populations of A. retroflexus ap- pear more likely to develop triazine resistance com- pared to North American populations, a situation common to weedy species that find themselves out- side their native ranges (Gressell & Segal, 1982). This problem begs for the inclusion of a phyloge- netic perspective. The question of whether one spe- cies is more or less constrained by its phylogeny to evolve herbicide resistance would appear to be fun- damental to understanding and predicting future patterns in biological responses to herbicide appli- cations. ON THE USE OF PHYLOGENETIC ANALYSIS TO DISCERN EVOLUTIONARY PATTERNS IN PLANT DEVELOPMENT AND FUNCTION Phylogenetic trees provide comparative biolo- gists with an important tool to uncover and polarize evolutionary pattern in functional traits. Through an examination of pattern comes insight into the evo- lutionary constraints that have influenced, and will continue to influence, functional responses to the environment. In the following paragraphs examples are provided to illustrate the use of phylogenetic trees to study evolutionary patterns in plant phys- iology and development. EPHEDRA AND DOUBLE FERTILIZATION For many years it was thought that a defining trait of angiosperms was the process of double fer- tilization during reproduction. Friedman (1990) has recently challenged this dogma through definitive observations of double fertilization in Ephedra, a non-flowering seed plant. Phylogenetic trees con- structed from numerous traits, including molecular markers, have placed Ephedra as a basal member of the Gnetales, the group of extant non-flowering seed plants most closely allied to the angiosperms (Doyle € Donohue, 1986). Using these phyloge- netic relationships as a guide, Friedman (1992) was able to demonstrate that the likely homolog to an- giosperm endosperm is the supernumerary-embryo product of the second fertilization in Ephedra (Fig. 10 Annals of the Missouri Botanical Garden Brassica rapa Brassica oleracea Brassica bourgeaui Brassica incana Brassica hilarionis Brassica montana Brassica cretica E Brassica insularis Brassica rupestris Brassica defle xa Sinapis aucheri Brassica oxyrrhina Brassica barrelieri Brassica repanda = Brassica desnottesii Brassica gravinae Brassica elongata ' Brassica nigra Sinapis arvensis Sinapis alba Brassica fruticulosa Brassica maurorum Brassica spinescens Sinapis pubescens Brassica tournefortii Figure 3. Phylogenetic relationships among members of subtribe Brassicinae (Cruciferae, tribe Brassiceae) based on cpDNA restriction mapping. The tree has been “pruned” from the more complete analysis presented in Warwick and Black (1993), such that only species in Brassica and Sinapis are shown. The arrow marks the currently known distribution of the inducible PPi-dependent phosphofructokinase as determined by W. Plaxton (Queens University, Ontario, pers. comm.). Volume 83, Number 1 Monson 1996 Phylogeny and Comparative Plant Biology o A c E: e X 9 có =Z E 3 ® o D EA = += ® ~ E Ke à ® LX < c C» < Q ©% o c = OG Wu à m « REDUCTION OF FEMALE GAMETOPHYTE TO EMBRYO-SAC FORM ADDITION OF 2ND FEMALE NUCLEUS TO 2ND FERTILIZATION EVENT MODIFICATION OF 2ND FERTILIZATION PRODUCT INTO NON-EMBRYO TISSUE PROLIFERATION OF 2ND FERTILIZATION PRODUCT INTO SUPERNUMERARY EMBRYO DOUBLE FERTILIZATION Figure 4. Phylogenetic mapping of double fertilization and the evolution of polyploid d in the Anthophytes, including the Gnetales (Ephedra, Gnetum, iugo 2. fossil groups Pentoxylon and Bennettitales, and the angi o- sperms. Following observations by Friedman (1990, 1992) it is concluded that double fertilization occurs prior to rawn from vergence of the angiosperms, but the evolution of BA endosperm occurred after divergence. Re 2). Friedman (199 4). The supernumerary embryo functions to nourish the primary embryo—an act of apparent “altruism” that raises questions about the role of kin selection in the evolution of developmental pathways (Fried- man, 1992) ADAPTIVE RADIATION IN THE HAWAIIAN ISLANDS Recent collaborative studies between Robert Robichaux (a comparative physiologist and ecolo- gist) and Bruce Baldwin (a comparative molecular systematist) provide an example of how phyloge- netic trees can direct the study of physiological ad- aptation. The Hawaiian silversword alliance in- cludes 28 endemic species that represent three genera (Argyroxiphium, Dubautia, and Wilkesia). A molecular phylogeny has recently been constructed ased on nuclear ribosomal DNA sequences. By mapping the distribution of wet versus dry habitat preference onto the phylogenetic tree Baldwin and Robichaux (1995) concluded that there have been at least five independent transitions from ancestors occurring in wet habitats to the derived species oc- curring in dry habitats. Adaptation to dry habitats in this alliance is known to include morphological and physiological traits that influence leaf energy balance (e.g.. dimensions, stomatal conduc- tance, and leaf surface properties that regulate solar reflectance) and physiological traits that influence leaf responses to water stress (e.g., cell wall elas- its influence on turgor maintenance) (Robichaux et al., 1990). Unibake the molec- ular phylogeny exhibited its lowest degree of res- olution in those lineages with the greatest physio- logical limiting the for successful reconstruction of physiological diversi- fication. This illustrates one important constraint on ticity and diversity, potential Annals of the Missouri Botanical Garden the process of mapping physiology onto phyloge- netic trees, namely that one must have a highly resolved phylogeny before gaining insight into physiological divergence. Using a similar approach, Todd Dawson (a com- parative ecophysiologist) has teamed up with Ste- phen Weller (a comparative systematist), Warren a comparative systematist), and Ann Sakai aa Wagner (a population biologist) to study another group of Hawaiian plants, Schiedea and Alsinidendron (Car- yophyllaceae). They have also observed the wet-to- dry transition in this group (Weller & Sakai, 1990; Weller et al., 1990) and are currently investigating the physiological and ecological attributes (e.g.. photosynthetic water-use efficiency, hydraulic prop- erties of the water-conducting system) of these taxa which permit them to exploit the drier habitats. Ul- timately, the study is aimed at placing these attri- butes into a phylogenetic context using character mapping procedures and the current phylogenetic tree (Wagner et al., 1995). This will hopefully pro- vide insight into whether those traits that permit certain species to thrive in dry habitats have arisen through response to selection in their current hab- itats or are historical artifacts of past selection. EVOLUTION OF C, PHOTOSYNTHESIS One of the earliest uses of phylogeny to discern evolutionary pattern in a functional plant trait in- volves C, photosynthesis. Within five years after the discovery of C, photosynthesis researchers had as- sembled phylogenetic trees showing the distribu- tion of C, plants (Evans, 1971; Moore, 1982). From this effort it was clear that C, photosynthesis has multiple origins and represents a derived trait, evolving from the ancestral C, photosynthetic path- way (the pentose phosphate pathway). More recently, Hattersley and Watson (1992) constructed an evolutionary hypothesis in the Po- aceae that reflects patterns of C, evolution and is consistent. with. well-accepted, higher-level taxo- — nomic affinities. Extensive reticulation of lineages in the Poaceae makes it difficult to reconstruct phy- logeny with high levels of confidence (though mono- phyly has been established for several subfamilies, Kellogg & Campbell, 1987). Hattersley and Watson (1992) mapped photosynthetic pathways onto their tentative phylogenetic tree, revealing several possible patterns of C, evolution. Of partic- ular importance, it appears that multiple origins of C, photosynthesis can occur within a single family and reversals of C, photosynthesis, back to the an- cestral C, type, have occurred in several groups. Such patterns suggest that the evolution and rever- sal of a pathway involving complex developmental and biochemical modific na occurs with relative ease (see also Watson et al., . This supports decisions of optimization ins p C, mapping in which addition and loss of this pathway are treated as equally likely as two additions. It has been suggested that such labile evolutionary pat- terns must be founded on a genetic architecture with one, or a few, regulatory genes controlling linked sets of C,-family structural genes (Watson et al., 1985; Monson, 1989a; Hattersley & Watson, 1992; Ehleringer & Monson, 1993). Thus, relative- ly few mutations, if they occur in the regulatory genes, can have a large influence on the evolution- ary expression of photosynthetic pathway type Further work on the evolution of C, otomí sis has been conducted in the genus Flaveria (As- teraceae). The monograph by Powell (1978) de- scribed 21 species for this genus. Physiological characterization of the component species has re- vealed that only five or six might be classified as “fully expressed” C, or C, types (Monson, 1989a; Monson & Moore, 1989). Most of the species reflect some intermediate phenotype between the C, and C, extremes (the so-called C,-C, intermediates). Some workers have described these C,-C, species as intermediate evolutionary stages on the path from the ancestral E type toward the derived C, type (Monson et al., 1984; Monson, 1989a; Brown & Hattersley, 1989). Thus, this appears to be a very active group of species in terms of photosynthetic p evolution. ¡sing the morphological descriptions provided in Powell (16 1ave constructed a phylogenetic tree of this genus (Fig. 5). This tree was used to address the question of whether the appearance of C.-C, intermediate photosynthesis always precedes the appearance of fully expressed C, photosynthe- sis. Support for such a pattern would strengthen the supposition that C,-C, intermediate photosynthesis in this group represents the antecedent to the evo- lution of C, photosynthesis. In this reconstruction, C, species were classified as those in which the majority of atmospheric CO, is assimilated through the C, pathway, and the C, and C, cycles have evolve d coordination to the point where leaves ex- hibit higher water- and nitrogen-use efficiencies— hallmarks of fully expressed C, plants. In this case, improvements in photosynthetic water- and nitro- gen-use efficiency were assumed if past gas-ex- change measurements have revealed reduced in- tercellular CO, concentrations without reduction in photosynthesis rate below the average expressed by all species in the genus. Such traits are presumably a reflection of the C, CO,-concentrating mecha- Volume 83, Number 1 1996 Monson Phylogeny and Comparative Plant Biology —Ó Sartwellia (C3) F. chloraefolia (Ca-C4) F. oppositifolia (Ca-C4) F. pubescens (C3-C4) A F. floridana (C3-C4) F. brownii (C4) F. linearis (C3-C4) F. sonorensis (C3-C4) — F. cronquistii (C3) F. vaginata (C4) F. pringlei (C3) F. angustifolia (C3-Ca) F. robusta (C3) F. ramosissima (C3-Ca) | F. palmeri (Ca) F. intermedia (Ca) pep C3: C4 Lo a | y Ca | it (0| | || [ C3-C4 F. bidentis (C4) $ F. campestris (C4) F. trinervia (C4) F. australasica (C4) Figure 5. the branch and bound option in PAUP 3.1 a consistency index (C.I.) = 0.41, a homoplasy index (H.I are based on 15 morphological traits that we or absence of pappus scales, pr heads, length of achenes, dens absence of ray florets, overall a of the re derived e Powell (197 esence or absence of ligules, a of disc y of inflorescence aggregation, pr :orolla, length of dhe corolla throat, presence or absence of connate F. anomala (C3-C4) u Phylogenetic relationships i in the genus Flaveria (Flaveriinae—Asteraceae). The tree was constructed using Swofford, i 4). The length of the tree is 33 steps and is characterized by = 0.59, and a retention index (R.I.) = 0.74. All relationships 8). Traits included phyllary number, presence orets, presence or absence of discoid ence or absence of self-compatibility, presence or leaves, presence of linear versus ovate leaves, a versus erect growth habit, and the presence or absence of leaf spiraling at senescence. nism. Species with the C, pathway are so classified because they exhibit 90% or more of their CO, fix- ation by the C, pathway. Species with C,-C, inter- mediate photosynthesis exhibit considerable CO, fixation by the C, pathway, but no advantages in terms of elevated water- and nitrogen-use efficien- cies. These designations were drawn from past re- ports of photosynthetic traits in this genus (Ku et al., 1983; Monson et al., 1986, 1987; Moore et al., 1988, 1989; Monson, 1989b; Ku et al., 1991). The tree exhibits two regions where resolution is such that the sequence of C,-C, and C, transitions can be mapped (regions A and C), and one region of inadequate resolution (region B). Using the path of greatest parsimony, the results from regions A and C demonstrate that C, photosynthesis has evolved at least twice independently. dira both cases of C, evolution were preceded by t appearance of C,-C, intermediate photosynthesis strengthening the hypothesis that this pathway is an evolutionary precursor to fully expressed C, pho- tosynthesis. The tree presented in Figure 5 supports evolutionary lability with respect to the C, pathway. Within a relatively small group of species there ap- pear to be cases of independent switches among the C4, C,-C,, and C, pathways. Caution should be used in relying too heavily on this tree to identify details of C, evolution. The tree is based on a limited number of characters and only yields moderately robust resolution. A tree based on numerous RFLP molecular markers is currently under construction, and is likely to yield greater resolution in interpreting C, evolutionary patterns in this group (P. Soltis and M. Ku, unpublished manuscript; M. Ku, Washington State University, pers. comm.). Nonetheless, the example provided Annals of the Missouri Botanical Garden in Figure 5 demonstrates the utility and guidance provided by phylogenetic perspective in the inter- pretation of metabolic evolutionary patterns. CONCLUDING STATEMENT Since Felsenstein’s seminal paper (Felsenstein, 1985), comparative physiologists and developmental biologists have observed the birth of a new exper- imental approach—the rigorous testing of compar- ative hypotheses using phylogenetic information. The approach is recent in its origins, and there are only a few examples of its successful application. However, the recent publication of several synthetic treatises of the phylogenetic approach and its re- the growing interest in this discipline (e.g., Huey, 1987; Harvey & Pagel, 1991; Brooks & McLennan, 1991; 1995). There is no doubt that com- lationship to comparative biology testify t © Lauder et al., parative biologists will come to rely heavily on As so aptly stated in Felsenstein’s paper: Phylogenies are phylogenetic perspective in future studies. fundamental to comparative biology. There is no do- ing it without taking them into account (Felsen- 1985). Systematics and comparative organismic biology stein, will be reciprocally strengthened through the shared use of phylogenetic analysis. Comparative physiologists and developmental biologists will be able to provide a stronger evolutionary interpreta- tion of functional responses to the environment. Systematists will acquire deeper insight into the evolutionary process, as well as help in resolving points of equivocation on phylogenetic trees through the use of new characters. The most press- ing limitation to these activities is the availability of adequately resolved phylogenetic maps. The achievement of a successful synthesis will require the continuation of efforts to construct classification systems that (1) deal with groups holding physio- logical and developmental interest, and (2) are ac- curate in their reflection of true phylogeny. Literature Cited Baldwin, B. G. & R. H. Robichaux. 1995. Historical bio- geography and ecology of the Hawaiian silversword al- liance (Compositae): New molec ular phylogenetic ap- prone € In W. L. Wagner (editors), Hawaiian Biogeography: Evolution on a Hot Spot Ar- chipelago. Smithsonian Institution Press, Washington, 2 V. Charlwood. 1980. The terpenoids. Pp. 18-220 in P Bell & V. Charlwood (editors), Encyclo- pedia of Plant Physiology, volume 12B. Springer-Verlag, erlin. Bjórkman, O. & D. Holmgren. 1963. Adaptability of the photosynthetic apparatus to light intensity in ecotypes iom exposed and shaded habitats. Physiol. Pl. 16: 889— 914. Brooks. A. pea nds p Bou Univ. € 1 Brown, R. H. & P. W. Hattersley. 1989. Leaf anatomy of T -C, species as related to i n of C, photosynthe- . PI. d urs 9]: 1543- 19 isa d. . D. Keck & W M. NA 1940. Experi- mental Sdn on the nature of species. I. Effect o varied environments on un dn 'an western plants. Carnegie Inst. Wash, € 1991. Phylogeny, . Experimental studies on ilie nature of species. P Environmental responses of climatic races of Achillea. Carnegie Inst. Wash. Publ. 581. Clutton-Brock, T. H. & P. H. Harvey. 1984. m approaches to investigatin g adaptatic on. Pp. 7-29 in J. H. Krebs & B. Davies (editors), Behavioral Ec dis An Evolutionary a 'h, 2nd Edition. Blackwell, Ox- ford, United ae Crook, J. H. 1965. The adaptive significance of avian social ak iat Symp. Zool. Soc. London 14: 181- 218 — - dePeer, Y. V., Neefs, P. D. Rijk & R. D. Wachter. 1993. anta ting pe shies eukaryotic small ribosomal subunit RNA s es: Calibration of the molecular clock. J. Molec. "Evol. 37: 221-232. Doyle, J. A. & M. J. Donohue. 1986. Seed plant phylog- eny and the origin of ES perms: Án experimental cla- distic approach. hes Re . 92: 321-431. Ehleringer, J. R. & R. “Monson. 1993. Evolutionary and ecological ak of Lag ede ‘ela vari- ation, Annual Rev. Ecol. Syst. 24: 411-430 Evans, L. T. 1971. Evol iu ary, ada e ve. mi environ- mental e of E Dogs) ian E athwa dés ess- ment. Pp. 130-13 Hatch, C. B. Ost O. ten io Photosy leo a Maui 'spi- ration. John Wile ey & Sons, New Yor Feder. M. E. 1987. The analysis of piilu ‘al diver- sity: The prospects for pattern documentation and gen- eral que stions in ecological oe ogy. Pp. 38-75 “4 M. E. Feder, A. F. Bennett, W. Burggren & R. Huey (dica: New sue dd in Ecological aL ogy. Cambridge Univ. Press, Cambridge Felsenstein, 1985. Phylogenies and the comparative method. Su gue 125: 1-15 Fiala, K. L. € R. R. Sokol. 1985. the accuracy A c ane estimation: computer simulation. pase 39: 60 Friedman, W. E. 199¢ uble DE nium in | Ephedra, a non-flowering seed ier Its P. on the origin of angiosperms. Ex 247: 951-95 W. E. 1992. Evidence of a pre-angiosperm or- igin of endospern rm: qoe for the evolution of ce 255: 336-339. Fac 'lors determining D m using 9— flowe UN plants. Garland, T. & P. A. er. 1994, ogy. Anal ae dun 56: 57 Gekeler, W., E. Grill, aul M. H. Zenk. 1989. E irvey E ee T nt kingdom for the ability to bind Bel metals through phytochelatins. Z. Naturf. 44: 361— oe Gould, S. J. & pepa 1979. The spandrels of San Marco and the Panglossian paradigm: A critique of the M programme. Philos. Trans., Ser. B. 205: 581-598. Grafen, A. 1989. The oo regression. Philos. Trans., Ser. B. 326: 119— 7 oo physiol- Volume 83, Number 1 1996 Monson Phylogeny and Comparative Plant Biology Gressel, J. & L. A. Segal. 1982. genie factors con- trolling the rate of ae of r The out- look for the future. Pp. 348 in B. M. LeBaron & J. Gressel (editors), Herbicide Resistance in Plants. John Wiley & Sons, New Y Harvey, P. H. & M. D. Pagel. 1991. The Comparative Method in Evolutionary Biology. Oxford Univ. Press, York. Hattersley, P. W. € L. Watson. 1992. Diversification of photosynthesis. Pp. 38-116 in G. P. Chapman (editor), srass Evolution Bug Domestication. Cambridge Univ. Press, Cambri Hill, R. J. 1982. ‘ae and biological considerations of herbicide-resistant ge hrs tolerant biotypes. Pp. 81-98 in H n & J. Gressel (editors), Herbic Me Resistance in e ida Wiley & Sons, New Holt, J. S. s. B. Powles & J. A. M. Holtum. 1993. Mech- anisms and agronomic aspec Annual Rev. Pl. Physiol. Pl. Melee: 203—229. Huey, R. B. 1987. Phylogeny. T = the compara- tive method. Pp. rd r, A. F. Bennett . W. Burggren & R. B. Huey al New Directions in Ecological B tose Cambridge Univ. Press, Ca hideo, . Bennett. 1987. Phylogenetic studies of Preferred temperatures versus temperatures of lizards. c uaa performance 8-1115. Kellogg, E. A. . S. Campbell. analyses of the Graminae. Pp. 310-322 in T. R. Sod- erstrom, K. W. H ;. S. Campbell € M. E. Barkworth (editors), Grass SEA and Evolution. Smithsonian Institution a Washington 3., optimal Evolution 41: 1987. Phylogenetic Ku, K. Monson, R. 0. Talcioha, H. Naka- moto, D. B. pone & G. E. Edwards. 1983. Photosyn- thetic characteristics of C,-C, intermediate Flaveria s in the C, and C, pathways. Pl. . Wu, Z. n »ott, C. Chu & G. E. Ed- 199]. aeara and eg rur characteristics of Flaveria species. Pl. Physiol. 96: 8-528 5l T Lamboy, W. F. 1994. The accuracy of the maximum par- simony method for phylogeny reconstruction with mor- pho d cal Virg unes Syst. Bot. 19: 489-505 Lande, 1985. Expected time for random genetic drift of a idol peer stable phenotypic states. Proc Natl. Acad. Sci. 82: 645. Peras G. V. . K. Monson & R. en. 1995. Systematics and the a: of iaa dun and function. a ence (in press). Lerdau, vak & R. Monson. 1994. Plant chem- ical defense: Monaerpne and the ges differentia- tion psc hypothesis. Trends Ecol. . 9: 5861. 990. The rate of morphologies evolution in mammals from the standpo f the neutral expecta- — Martin, A. P & S 1993. Body size, meta- San rate, — time, and the molecular cloc Proc. Natl. . Sci. U.S.A. 90: 40874091. Martins, E. a [e the rate of phenotypic evo- lution from comparative data. Amer. Naturalist 144: 193-209 —— & T Garland. 1991. Phylogenetic analyses of the correlated evolution of continuous characters: A simu- 034—551. lation study. Evolution 45: Miyachi, . Tsuzuki & Y. Yagawa. 1985. Carbonic anhydrase in various microalgae. Pp. 145-154 in W. Lucas & erry (editors), Inorganic Carbon Uptake by d uidie Phiotosylithetic Mice American Soci- ety of Plant NUM Press, Rockville. Monson, R. K. the pee letales pathways re- sulting in C, kw bibo and Crass ssula tabolism. Advances Ecol. R 1989b. cean acid me- -110 The relative pote nit of reduced photorespiration and improved water- and nitrogen-use efficiencies, to the advantages C,-C, intermediate p in Flaveria. es ologia 80: 215-221. Fi re. 1989. On the significance of C,- 4, intermediate intei to the nosmet — G.E. LI es. 19 evolution of C, d : wur 12: 689—699, E M. S. Ku. bo C.-C, intermediate photosynthesis. BioScience 34: 503-574. ———, V. S. Schu Ku. 1987. Photosyn- thesis in ra 'eria beni A. Powell, a C,-like C,- C, inte WU PL ao 85: 1063- 1067. . Moore, M. ¡dwa Cx Heme of C,- “a C,- aatan path- C,, C,, and C,-C, puis an Flaveria species Planta 168: 493—502 Mooney, H. A. & W. D. Billings. 1961. Comparative physiological ecology of arctic and pues populations of Oxyria Wo Ecol. T 31: 1-29. Moore, B. D., M. S. B. . E. i ie 1989. Ex- pression vs ¿like I o. in several species of Flaveria. PI. € zell 20 12: ipea . R. Mon M. S. B. Ku & G. E. Edwards. 1988. Activities of princ WP chan Oe and pho- ymes in leaf mesophyll and bundle Pe ath protoplasts from the C,-C, ey Flaveria ramosissima Cell heb. 29: 999-106 Moore, P. D inns Eos of hotte pathways in flowering plante Nature 295: 647—648 ell, A. 978. mática of Flaveria digg inae ure eae). Ann. Missouri Bot. Gard. 65: 59% l 986. Rhodes, . D. Hanson. 1993. Quaternary ammo- nium ul tertiary sulfonium compounds in higher Jw Annual Rev. Pl. Physiol. Pl. Molec. Biol. 44: -384. "s haux, R. H., cy. G. D. Carr, M. Liebman & R. W. Pear- Adaptive radiation of the Hawaiian Silver- sword alliance (Compositae-Madiinae): Ecological, morphologic sal, us physiological diversity. Ann. Mis- souri bra rà 7: 64-72. Rohlf, Chae R. R. Sokal & J. Kim. nee of e nine Ride: E ‘ts of tree to- pology and evolutionary model. Evolution 44: 16 1684 Saccone, C., G. Pesole & G. Preparata. 1989. DNA mi- € s be the molecular clock. J. Molec. Evol. 29: 4€ Swofford, D. i 154 Phylogenetic Analysis Using Par- simony (PAUP) 3.1s. Illinois Natural History Survey, Champaign. Theodorou, M. E. & W. C. Plaxton. 1993. Metabolic ad- aptations of plant respiration to ae phosphate de privation: PI. Physiol. 101: 339— a p Ppi- -depende nt phosphofruc Ainas by phosphate starvation in seed- Annals of the Missouri Botanical Garden lings of Brassica nigra. Pl. Cell Environm. 17: 287- 294. Turreson, G. 1922. The genotypical response of the plant species to habitat. Hereditas 3: 211—350. ios W. pi S. G. Weller & A. K. Sakai. 1995. Phy- eny, biogeography, and evolution in Schiedea and Al- ERTI “Caryophyllac eae). In W. Wagner & V. Funk (editors), Hawaiian Biogeography: Evolution on a Hot Spot Archipelago. Smithsonian Institution Press, ashington, D. C. (in press). Warwick, S. I. 1991. Herbicide resistance in weedy plants: ‘gestae ps oo biology. Annual Rev. Ecol. Syst. 22: 95-1 a D. Blac E ul Molecular relationships in subtribe deus (Cruciferae, tribe Brassiceae). Canad. di 71: 906—921. Watson, L., E: Clifford & M. J. Dallwitz. 1985. The classific WM of Pn eae: Subfamilies and supertribes. Aust 3 ta Bot. 33: —484 Weller, S. G. & A. K. Sakai, The evolution of di- cliny i in Schiedea Nene ee cu) an endemic Ha- waiian genus. Pl. Spec. Biol. : c ag R. Herbst. 1990. e evolution of dioecy i in he (Caryophyllaceae: Alsinoideae) in the Hawaiian Islands: Bose 'ographical 6-276. Ine e of the maximum- parsimony method: The case of five taxa with a molecular clock. Syst. Biol. 42: 113-125. SYSTEMATICS SOLVES PROBLEMS IN AGRICULTURE AND FORESTRY! Amy Y. Rossman? and Douglass R. Miller? ABSTRACT In forest and agricultural oe the conspicuous elements, namely the trees, crop plants, and farm animals, rn aa interactions with m ess conspicuous organis s. These less conspicuous but specious organisms such nsects, fungi, nematodes, and peines can be beneficial, even essential, or they can be utterly devastating causing billions i dollars damage. ide present knowledge of the systematics of these less conspicuous organisms is limited. For some groups even the m tion—is lacking. This paper presents e It the People’s Republic of China. Systematics i is essential in directing the collection, organization, and use of ests in eastern North America have important problems in agriculture and forestry been devastated by the n all the examples detailed in this paper, basic systematic knowledge was essential to solving Trees, crop plants, and farm animals are the most conspicuous elements in forest and agricultural ecosystems, yet these organisms have complex in- teractions with many less conspicuous organisms. he myriad of insects, fungi, nematodes, and bac- teria that are part of these ecosystems can be ben- eficial, even essential, to the development of the crop, or can be utterly devastating causing billions of dollars damage. At present our knowledge of the systematics of these less conspicuous oran : grossly limited—so limited that we often do n have even an elemental iid cedo of their existence—an inventory, a checklist, a means of identification. These are the ecosystems upon which humanity depends for survival. Within forestry and agriculture there is an in- creased interest in holistic approaches to managing the biological resources on which these industries depend. Such management strategies must allow the exploitation of biological resources to provide the immediate needs of food and fiber, but also must accommodate management approaches that minimize the impact on the environment and en- sure long-term use of these resources. Such strat- egies include sustainable agriculture, the biological control of pest organisms, integrated pest manage- ment, and the management of forests for products adequate systematic knowledge these initiatives can be successful. This paper presents examples of problems in ag- riculture and forestry that have been solved by ap- plying a systematic understanding of the organisms involved. In some cases the result was to solve short-term problems with short-term economic gain, for example, in increased international trade, while in other cases the result has been incalculable, long-term benefit, such as in the biological control of an exotic weed that was threatening to destroy an entire ecosystem. In all cases basic systematic knowledge was essential to solving the problem. AGRICULTURE AGRICULTURAL PRACTICES REFLECT THE SYSTEMATIC UNDERSTANDING OF PEST ORGANISMS Agricultural practices of previous centuries in- cluded empirically integrated management of crop ! We thank the following systematists who contributed ideas and information to this pap Carolina; Harry Evan USDA- Agricultural Research Service, Vegetable C Carolina State University, Plymouth, North pos and David Spooner, : Marc Cubeta, North atic onal Institute of Dei. Silwood, En- rops Research Unit, Madison, Wisconsin. ns, Intern SDA-Agricultural Research Service, Systematic Botany and Mycology Laboratory, Systematic Entomology Labo- ratory, Beltsville, Maryland 20705, U.S.A. ANN. Missouni Bor. GARD. 83: 17-28. 1996. Annals of the Missouri Botanical Garden Binucleate hyphae of Rhizoctonia solani AG-4 under fluorescence (left) and dark-field microscopy (right). Photo by Mare Cubeta. Figure 1. pests, often with limited success. During the last 100 years, however, these limitations on agricul- tural productivity have been lowered, resulting in an increased human population and demand for food. Despite spectacular success, 10-20% of all agricultural crops are still lost to pests and patho- gens (Anonymous, 1993). As the need to produce more agricultural commodities increases, expecta- tions have also increased for lower chemical input to agricultural systems and products. Knowledge of the systematics of the insects, fungi, nematodes, and microorganisms that consume a significant por- tion of the agricultural products provides the key to solving this dilemma. For example, until recently the fungus commonly identified as Rhizoctonia solani Kiihn. was believed to be a single, widespread species that occurred on almost every vascular plant and caused root rots, barepatch, wilts, diebacks, blights, and blotches (Farr et al., 1989; Parmeter, 1970). This fungal spe- cies produces almost exclusively vegetative hyphae (Fig. 1), albeit vegetative hyphae with distinctive morphology (Parmeter, 1970). In the past 20 years differences were noticed in the rarely formed sexual state, and relatedness was defined on the ability of strains to undergo anastomoses or hyphal fusion. Now the R. solani complex is separated into anas- tomosis groups, or AGs, based on this ability. Re- cent molecular analyses of the anastomosis groups and increased knowledge of sexual states has al- lowed systematists to characterize biologically meaningful species that correlate with such impor- tant parameters as host susceptibility (Ogoshi, 1987; Sneh et al., 1991). Within the entity previ- ously referred to as R. solani, a number of species are now recognized, some of which are pathogens specific to particular crop plants, others are my- corrhizal with orchids, while still others can be used as biological control agents of plant pathogen- ic fungi (Vilgalys & Cubeta, 1994). This crucial systematic information allows plant pathologists to recognize and seek out control strategies for the species that are pathogenic on specific crops, al- lows orchid growers to understand the positive as- pects of the presence of these fungal species, and allows biological control specialists to use these fungi in their arsenal of control agents. The system- atist has made order out of chaos. SYSTEMATICS PROVIDES THE MEANS FOR DISCOVERING AND DEVELOPING BIOLOGICAL AGENTS TO CONTROL AGRICULTURAL PESTS AND PATHOGENS Damage to agricultural crops due to fungi, both in the field and. during harvest and storage, is es- timated at more than $3.5 billion in the United States (Kendrick, 1992), while the dollar value from insect damage is equal or greater. Insects, nema- todes, fungi, and microorganisms constantly com- pete with humans for these commodities. An in- creasingly attractive alternative to chemical control of agricultural pest organisms is through biological control or the manipulation of a biological antago- nist, often a natural enemy. One suspects that con- siderable biological control exists in natural sys- tems and that the interaction and balance between organisms is extremely complex. Fungi are being explored as biological control agents of insects, nematodes, plant pathogenic fun- gi, and noxious weeds. The fungi involved are not the macrofungi with which most people are familiar, that is, mushrooms, polypores, or lichens. Rather, the fungi having the greatest impact in agriculture and forestry are microscopic in size, often fast- growing, and producing many tiny reproductive structures. Ás a group of organisms they are vastly understudied, to the extent that at least 5096 of the new species with biocontrol potential have yet to Volume 83, Number 1 Rossman & Miller 19 Systematics and Agriculture and Forestry be discovered and described. Fundamental system- atic information, such as species descriptions and understanding of relationships with known species, is needed. One example in which systematic knowledge has contributed to the development of effective biolog- ical control concerns fungi used to control soilborne fungal diseases in temperate agricultural systems. Like many Ascomycetes, these fungi are most com- monly encountered as asexually reproducing strains for which a sexual state may or may not be known. One of the most commonly used biocontrol strains was initially identified as Gliocladium virens Gid- dens, Foster oster. Although described in Gliocladium, this biocontrol fungus is morphologi- cally and biologically unlike the type and other species in this genus. Using both morphological and molecular approaches, two systematists have shown that Gliocladium virens actually belongs in the relatively unstudied genus Trichoderma (Fig. 2) Los & Samuels, 1994; Samuels & Rehner, 1993). Based on that conclusion, one would predict the related sexual state of this fungus would be an 3). Thus, strains of the closest sexually reproducing relative, ascomycete in the genus Hypocrea (Fig. : H. gelatinosa (Tode) Fr., were tested for the pro- duction of the fungal metabolite gliotoxin. Gliotoxin is correlated with potential for biological control of fungal pathogens. Strains derived from the closely related sexual state produced as much gliotoxin as the biocontrol fungus. In addition, some of these newly discovered biocontrol strains produced their sexual states in culture, allowing conventional ge- netic manipulation. Thus, increased knowledge of the systematics of the Trichoderma complex led to the prediction and discovery of more effective strains of biological control fungi. nlike the above example, agricultural problems are often solved using the brushfire approach of reacting when an emergency arises, undertaking a narrow research program on the pest causing the problem, possibly finding a solution, and going on to the next agricultural brushfire. Because compre- hensive systematic knowledge is not generated in solving an immediate problem, the short-term so- lution does not add significantly to the development — of a predictive classification system. In fact, ir many situations systematic analysis of a single spe- cies, removed from the context of phylogeny and biogeography, and not carefully integrated into a classification system, detracts rather than adds use- ful information. An example of a circumstance where comprehensive research followed a brushfire solution involved a pest in Africa. An unknown mealybug attacked cassava in West Africa (Fig. 4 — and cost farmers £1.4 billion each year (Anony- mous, 1986). A systematist described the species an: Phenacoccus manihoti) as new and, based on the systematic relationship of this species with others, suggested that natural enemy exploration be un- dertaken in Central and South America (Matile- Ferrero, 1977). Additional specimens of a mealy- bug erroneously identified as P manihoti were discovered in northern South America, and several of its parasites were imported to Africa. Unfortu- nately, none of the biological control agents were effective, and a mealybug systematist was asked to study the South American material. The systematist determined that the mealybug from northern South America actually was a second species different from P. manihoti described from Africa (Cox & Wil- liams, 1981). Eventually true P. manihoti was lo- cated further south in South America, and effective parasites were discovered and successfully intro- duced into West Africa (Herren & Neuenschwan- der, 1991). After this brushfire was put out, the International Fund for Agricultural. Development offered financial support for a study on the mealy- bugs of South America. A book has recently been published (Williams & Grana de Willink, 1992) that serves as a first step toward understanding the diverse mealybug fauna of the area and prepares the world for the emergence of the next devastating mealybug pest. INTERNATIONAL EXCHANGE OF AGRICULTURAL COMMODITIES DEPENDS ON ACCURATE SYSTEMATIC INFORMATION International exchange of agricultural commodi- ties in the United States was valued at $24 billion for imports and $42 billion for exports in 1992, accounting for a significant portion of total domestic exports (Anonymous, 1993). Regulations governing the international exchange of agricultural and forest commodities are directed at the containment of eco- Thus, change requires the accurate and rapid identifica- of both pathogens. Systematists provide the expertise and nomically damaging organisms. such ex- tion domestic and exotic pests and tools on which these identifications are made, al- lowing the existence of this multibillion dollar in- dustry. The smut fungi, Ustilaginales, cause severe dis- 5). Although relatively well surveyed and described, these obli- eases of important grain crops (Fig gate parasites are generally considered to be host specific and are identified primarily based on teli- ospore characteristics. Their identification is diffi- cult and requires taxonomic expertise, particularly Annals of the Missouri Botanical Garden e 2. SEM of Trichoderma virens, a fungus with potential for the biological control of plant pathogenic fungi. Photo ie James Plaskowitz. if the host is unknown or misidentified. Tilletia in- dica (Mitra) Mundkur, karnal bunt of wheat, is a smut fungus that occurs in limited regions of the taminated with Tilletia indica. Plant quarantine of- world (Green, 1984; Smith et al., 2; Waller & — ficials in both the U.S. and Canada became quite Mordue, 1983). Extreme vigilance is required to exc ite d, and a ban on the import of wheat from the prevent the spread of this pathogen. Recently U.S wheat (Triticum aestivum L.) imported into Canada from the United States was determined to be con- . to Canada was suggested. A systematist who Volume 83, Number 1 1996 Rossman & Miller 21 Systematics and Agriculture and Forestry Asci with 16 partspores of Mt rea gela- Figure tinosa, i: org state of Trichoderma virens was asked to study the material correctly identified the smut as Tilletia barclayana (Bref.) Sacc. & Syd. This smut fungus occurs on rice and apparently contaminated the wheat when it was stored in a warehouse that had previously contained rice. This identification was eventually confirmed with iso- zyme analysis, and a potentially costly international incident was averted (Mary Palm, 1993). Systematic knowledge of another smut fungus on pers. comm., wheat and the clarification of the circumstances un- der which it was collected have allowed the sale of f ¡a wheat from California to the People's Republic « China. Export of wheat from the Pacific Northwest of the United States to the People's Republic of China has been curtailed since 1972 because of dwarf bunt, a fungus not known to occur in China. A the presence of Tilletia controversa Kühn., common bunt o rough-spored bunt, is caused by T. tritici (Bjerk.) Wolff and occurs throughout the world wherever wheat is grown (Anonymous, 1990; Mordue & Wal- ler, 1981). Differentiating teliospores of T. tritici from T. controversa is difficult to impossible be- second bunt disease of wheat, = Rd g 1 West Africa. The tree on the iani attac ked by the mealybug da occus manihoti has not produced large, edible while the the left is free of mealybug attac ik ne oss normal tuber production. Figure 4. Cassava trees i one on cause of their morphological similarity. Thus, it is important to know if dwarf bunt is present, absent, or has ever occurred in specific regions in the U.S. from which wheat might be shipped to the People's Republic of China. The wheat-growing regions of California are free of dwarf bunt except for one re- port of T. controversa (Duran & Fischer, 1956). This report curtailed the export of wheat from California to China. The report of dwarf bunt in California was based on a specimen collected [on June 30, 1917 in Jacksonville, “California”] by a U.S. Department of Agriculture plant pathologist, H. B. Humphrey. The specimen was deposited in the U.S. National Fungus Collections and was available for study. De- spite intense efforts using morphological, biochem- ical, and molecular means, it was not possible to identify this specimen as either 7. tritici or T. con- troversa. Since identification was not possible, a new strategy was devised that again depended on systematic facilities. All specimen label data as- sociated with the specimens at the U.S. National Fungus Collections had previously been entered 22 Annals of the Missouri Botanical Garden E ure 5. Corn smut fungus, Ustilago maydis, infect- ing iig Is of field corn in Maryland. into a computerized database. Using this systematic information resource, it was possible to determine the approximate route of Humphrey on his 1917 trip (Table 1). Although this database was never intended for tracking a scientific expedition, it pro- vided the information necessary to prove that the dwarf bunt specimen was collected in Jacksonville, Oregon, not in California. This fact was confirmed by the itinerary and telegrams of H. B. Humphrey for this trip deposited and maintained at the U.S. National. Archives. Using these sources of infor- mation it was proven that this dwarf bunt specimen was not collected in California (Rossman, 1994). Table 1. Route taken by H. B. Humphrey in 1917 as J.S. National Fungus Collections. As a consequence, the People's Republic of China has lifted the quarantine on the import of wheat from California. According to the California Wheat Commission (B. Fernandez, pers. comm.), a first shipment of California wheat to the People's Re- public of China left Stockton on April 1, 1995. This first shipment of 30,000 tons is worth about $4.7 problem has million. Solving this systematic opened a multimillion dollar market. SYSTEMATICS DIRECTS THE COLLECTION, ORGANIZATION, AND USE OF VASCULAR PLANT GERMPLASM Humankind has always assumed that the genetic resources required to support agriculture would continue to exist in nature forever. As an indication of the importance of genetic resource preservation, Congress has recently mandated that the United States Department of Agriculture formulate a pro- gram to develop, store, and access genetic re- sources for all kinds of living organisms (National Research Council, 1991). The report on specifically how this could be done and how much it would cost has been presented to them. This program in- cludes not only vascular plants and animals of ob- vious importance to agriculture but also fungi, in- sects, nematodes, and microorganisms. Congress recognizes the essential role that biological diver- sity plays in human existence. The problem is how to obtain, organize, preserve, and utilize the genetic resources needed to insure that future agricultural needs will be met. The critical basis for developing and utilizing genetic resources is systematic knowl- edge (Shands & Kirkbride, 1989) For vascular plant germplasm in the United States, a large system of repositories exists. The U.S. National Plant Germplasm System contains 5— 10% of the 250,000 vascular plant species in ex- istence (National Research Council, 1993). Infor- mation on the over 400,000 accessions is based on the systematics of the organisms using the Germ- plasm Resources Information Network. Entry to ac- cession information is through the scientific name of the species, and all communications depend on determined by collection data on specimens deposited in the Specimens collected by H. B. Humphrey in 1917 deposited in the U.S. National Fungus Collections e June 24 1917 Campbell, Santa Clare County, California e June 30 1917 Medford, Oregon € June 30 1917 Jacksonville, California (actually Oregon) e July 4 1917 Pullman, Washington e July 9 1917 St. Paul, Minnesota Volume 83, Number 1 Rossman & Miller 23 Systematics and Agriculture and Forestry tes of soquicition~ — Inst ite Figure 6. John Wiersema, nomenclaturalist, sources Information Network accurate nomenclature as determined by a system- atist specializing in the classification of cultivated plants (Fig. 6). The systematist provides informa- tion on the relatedness and therefore the usefulness of the germplasm to its potential crop. Repositories of vascular plant germplasm set their priorities for collecting and utilizing germ- plasm based on systematic knowledge of crop plants and their immediate relatives. For example, the numerous wild relatives of the cultivated potato provide the genetic basis for disease resistance that is present in today's crop. Cultivated potato is con- sumed worldwide and represents the fourth most important food resource (Hawkes, 1990). In the na- tive habitat of wild potatoes ranging from the south- western United States to south-central Chile, these relatives are often obscure weeds and are not used locally as food; indeed, some are mildly poisonous. David Spooner’s research on wild potatoes (Sola- num sect. Petota) illustrates the value of system- atics information to plant breeding. Spooner's re- search has two components—-collecting for more complete germplasm representation in genebanks and systematics research to understand the species and their relationships with Solanum sect. Petota. His program, collecting in collaboration with re- searchers from South America, has resulted in the reviewing scientific names of vascular plants in the Germplasm Re- acquisition of more than 20 species—relatives of the cultivated potato—not present in the world’s genebanks. largest collection of potato germ- holds about The world's plasm at Sturgeon Bay, 5000 accessions (Bamberg & Spooner, 199 about 170 of the 232 wild potato species recognized in the latest comprehensive taxonomic treatment of Solanum sect. Petota (Hawkes, t present, there are disagreements among taxonomists regard- ing the species, hypotheses of natural interspecific hybridization, and the relationships among taxa of cultivated and wild potato. The treatment previous to Hawkes (1990) recognized only 157 species, and the species boundaries and their interrelationships have yet to be reconciled (Spooner & van den Berg, 1992). Thus, Spooner’s ongoing research program using a variety of tools to investigate the identity and relationships among these species is crucial for the enhancement of commercial potato products (Giannattasio & Spooner, 1994a, b; Spooner et al., 1993; Spooner & Sytsma, 1992). Because it takes 8-15 years from the initiation of a breeding pro- gram to a commercial variety release, considerable isc onsin, time and expense can be saved by making an initial choice of breeding material based on accurate sys- tematic knowledge. 24 Annals of the Missouri Botanical Garden FORESTRY from Germany. When the ship’s holds were opened, Forest land in the United States occupies 737 million acres, yielding products valued at about $1 billion in 1992 (Anonymous, 1993). Other forest commodities and non-commodity uses such as rec- reation and watershed resources have an even greater value. Changes in the public’s perception of the value of forest lands have led to a major shift in their management. This losophy has is in older tree stands and in new management phi- greater value of non-timber species. Long-term for- est system management has a different set of prob- lems than those associated with traditional forest management. While most forest managers measure physical and chemical aspects of their forest as well as monitor changes in the macrofauna and flora, they generally are unable to measure the biological diversity of the total forest biota. Knowledge of the systematics of non-timber components of forest eco- systems is essential to their sustainable mainte- nance. SYSTEMATIC KNOWLEDGE HELPS PREVENT THE INTRODUCTION OF EXOTIC PESTS AND PATHOGENS THAT DESTROY FORESTS A major force in the destruction of forests, aside from harvesting of trees, has been damage inflicted by introduced eae pests and pathogens. The eco- nomic loss to timber revenues is estimated at $2 billion dl (Campbell & Schlarbaum, 1994), with a much greater loss in recreational value. The infamous American chestnut blight introduced on Asian chestnut nursery stock in 1904 has altered the landscape of the eastern deciduous forest for- ever (Anagnostakis, 1987). Yet nursery stock and unrefined logs are imported into the United States without thorough knowledge and understanding of the organisms associated with them (Campbell & Schlarbaum, 1994; Redlin, 1991). The fungus causing dogwood anthracnose, Discula destructiva Redlin (Fig. 7), was described only recently (Red- lin, 1991) but its origin is still unknown because of inadequate baseline data on the fungi in the United States and the rest of the world. Its simul- taneous appearance on both coasts of North Amer- ica suggests that it was introduced on nursery stock (Campbell & Schlarbaum, 1994). Increased sys- tematic knowledge of the inconspicuous organisms that occupy forests worldwide is needed to prevent additional destructive introduction Systematic tools for the rapid Sdentificaiion of po- tentially harmful organisms also will avert future disasters. In 1993 a ship arrived in Wilmington, North Carolina, carrying a cargo of military goods moths were seen flying from the cargo areas. Ag- ricultural quarantine inspectors closed the holds and set the ship back out to sea. The insects ap- peared to be the gypsy moth but inspectors were unsure whether they were the European gyps moth, which is established on the East Coast from an earlier introduction, or the Asian gypsy moth, which does not occur in the U.S. and is notable for its flying females. The samples were shipped to Delaware, Ohio, where systematic specialists used molecular techniques to quickly identify the spec- imens as the Asian gypsy moth. The ship's cargo areas were fumigated, and a trapping program around the port was initiated to eradicate the Asian gypsy moth before it could become established. Previous development of a rapid identification method by systematists had given quarantine de- cision makers the tools to avert the establishment of a potentially damaging exotic pest. Decisions to allow the importation of agricultural and forestry commodities into the United States are generally based on whether a potentially damaging organism already exists in this country. Unfortu- nately, this assumes that these organisms are known and accurately identified in the U.S. At present only about 50% of the insects (Kosztarab & Schae- fer, 1990) and 20-40% of the fungi in the U.S. have been described (Klassen, 1986). A database has been developed of reports of fungi on plants and plant products in the United States (Farr et al., 1989). This database is one of the primary re- sources on which the Animal and Plant Health In- spection Service depends when making decisions about entry of commodities into the United States. Previously this information was scattered and dif- ficult or impossible for decision makers to obtain. Now with a single source of systematic information, more knowledgeable decisions can be made. In is information has allowed the im- portation of plant products that previously were prohibited entry because of the lack of knowledge about the organisms in the United States. Although this database is the most comprehen- sive account of plant-associated fungi in the world, listing 13,000 species (Farr et al., 1989), it is far from complete, particularly for fungi on non-crop some cases t plants. The fungi reported in the literature are usu- ally those that are conspicuous or cause damage to plants of economic interest. We now know that there are potentially pathogenic fungi that live in- side apparently healthy plant tissue (Carroll, 1988). Visual inspection of such plants does not detect these latent pathogenic fungal s As an ex- ample of the inadequate knowledge of fungi in the ecies. Volume 83, Number 1 Rossman & Miller Systematics and Agriculture and Forestry Figure 7. Conidiomata and conidia of fungus causing dogwood anthracnose, Discula destructiva, erumpent through leaf epidermis. U.S., we compared from various sources the fungi of profound economic importance. In Farr et al. 19 known to occur on one host, Chamaecyparis thyoi- ( O species are listed on this host. In the des (L.) B.S.P., Atlantic white cedar (Table 2). Al- U.S. National Fungus Collections, 50 species are though not an inconspicuous host, this tree is not represented, of which only 9 are included in Farr 26 Annals of the Missouri Botanical Garden Table 2. in Australia, the native plant populations in Mad- Fungi jon on Chamaecyparis ed Atlantic white cedar ? (Farr et al., 1989), USNFC (U.S. National pui pinna Bills & Polishook (1992). Fungi known on Chamaec yparis thyoides, Cedar Atlantic White C Bills & FOPP USNFC Polishook FOPP 40 9 3 USNFC 4l 2 Bills & Polishook 72 40 50 77 Total number of species 153 et al. (1989). B 81 fungal species are known from this host. In just one study Bills & Polishook (1992) isolated 77 fun- gal species occurring as endophytes in living tissue ased on these two resources alone, of this host. Of these, only three species were listed in Farr et al. (1989), while two species were rep- resented in the U.S. National Fungus Collections. From these three resources 153 fungal species are known on Chamaecyparis thyoides, of which only 40, or about 25%, are reported in Farr et al. (1989). These data suggest that by simply gathering infor- mation from our systematic collections, we can en- hance the knowledge base considerably. Then, in this case, by undertaking even cursory sampling activities, we can again double the number of fun- gal species known from that host. This fundamental knowledge of the organisms that occur in the Unit- ed States is needed to make enlightened decisions about the safety of allowing entry of agricultural and other plant commodities into this country. SYSTEMATIC KNOWLEDGE OF HOST AND PATHOGEN IS NEEDED TO CONTROL EXOTIC ORGANISMS Introduced organisms that threaten forest eco- systems as well as grazing lands can be controlled using biological agents. To be successful, system- atic knowledge of both the biocontrol agent and tar- get organisms is critical. For example, in northern Queensland, Australia, a noxious weed known as grazing lands and threatened ea native forests by completely covering the trees. Exotic to Australia, where it was intentionally introduced over a century ago to cover spoils from gold mines, this perennial, woody, climbing shrub still exists in relictual but threatened populations in its native habitat of Mad- agascar (Evans, 1993). To solve the weed problem agascar were examined for both fungus and insect natural enemies. Although insects were located, none were found to be host specific and thus were not considered safe for introduction into Australia. A rust fungus (Uredinales) with biocontrol poten- tial was collected by Harry vans of the Inter- national Institute of Biological Control in 1987- 1988. Unfortunately, this rust was considered to be- long in the genus Hemileia Berk. & Broome, in which is also placed the notorious coffee rust fun- vus H. vastatrix Berk. & Broome. The rust pathogen of rubbervine had been described in 1914 but lit- erally ignored for decades because of its seeming lack of economic importance. Evans (1993) under- took a systematic study of this potential biological control agent and discovered that the nuclear con- dition of supposed urediniospores was that of a sex- ually reproducing fungus, thus what appeared to be urediniospores were actually teliospores or the sex- ual spores of the rust. Under conditions of low hu- midity, such as occur in the semiarid parts of Mad- agascar, these prolific urediniospore-like teliospores were produced in abundance. However, with high hu- midity, such as in a greenhouse in the United King- dom or as occurs only occasionally in this region of Madagascar, true teliospores are produced that even- tually germinate to form a variable number of basid- iospores. This rare phenomenon reveals the evolution- history of this fungus as well as its closer relationship with another group of rusts. This rust is now placed in Maravalia Arthur in the Chaconiaceae, Uredinales, only distantly related to Hemileia. Biological control agents must be host specific. In testing for host specificity, close relatives of Cryptostegia grandiflora in the Asclepidaceae were examined for susceptibility to the rust fungus. While the target plant proved to be highly suscep- tible, related species demonstrated a range of re- sistance that reflected the relationships among the host plants. For example, within the subfamily Per- iplocoideae of the Asc M the closest re- lated genus, Gymnanther , proved to be highly resistant with a response s included hy- phal collapse after penetration. In another related plant species, Finlaysonia obovata Wall., mation of sori was initiated but pustules did not mature (Evans & Fleureau, 1993). On two species in the subfamily, Gonocrypta grevei Baill. and Cryp- tolepis grayi P. I. Forst., fertile sori did develop in greenhouse tests. Further studies on ecotypes of G. grevei demonstrated population differences with sporulation on certain plants occurring only at sat- uration inoculum levels. Based on these findings, two physiological races of the rust fungus were dis- Volume 83, Number 1 1996 Rossman & Miller Systematics and Agriculture and Forestry tinguished, one adapted to Cryptostegia, and the other to Gonocrypta. This was corroborated by the field observation that rusted Cryplostegia grandiflo- ra grew intertwined with healthy Gonocrypta grevei (Evans & Tomley, 1994). As a result of increased systematic knowledge of a rust fungus, Maravalia cryptostegia (Cummins) Ono, and of the host plant and its relatives, rubbervine in Australia is no lon- ger devastating the forest ecosystem. This success story is the result of an extremely small piece of the entire systematics puzzle; most pieces of the puzzle remain separated. If there ex- isted a thorough systematic understanding of a ma- jority of the rust fungi, the potential for controlling exotic weeds such as rubbervine would be in- creased. In some parts of the world exotic weeds threaten the extinction of more biological diversity than the threat of habitat destruction from human activity (U.S. Congress, 1993). Lack of systematic knowledge is hindering use of biological control agents to stop the destruction of endangered native habitats due to exotic weeds. Given the ability of one rust species to control a noxious weed that threatened a native forest in Australia, imagine the untapped potential for the use of fungi as biological control agents. Systematics is the key to achieving that potential. CONCLUSION Many costly problems in agriculture and forestry could be solved with increased systematic knowl- edge of the inconspicuous organisms that influence forest and agricultural ecosystems. Systematics pro- vides the means to use and benefit, rather than suf- fer, from their biological influence. The economic payoff from using systematic knowledge to solve problems in agriculture and forestry is far greater than the cost of funding a systematics research pro- gram. Literature Cited Anagnostakis, S. 1987. Chestnut blight: The classical blem of an introduced pathogen. Mycologia 79: 23— 37. Anonymous. 1986. The one billion pounds a year mealy- rings in more projects. CAB International News, October 1986: 6-7. Anonymous. 1990. Distribution maps of plant diseases. Tilletia indica. Map No. 290. 2nd ed. Int. Mycol. Inst., Kew, England. Anonymous. 1993. ep nen ey 1993. Govern- ment Printing d = i Bamberg, J. B. & ner. 51004, The United States pato AN dci station a Taxon 489— ins G. F. & J. D. Polishook. 1992. Recovery of endo- ra fungi from Chamaecyparis thyoides. Sydowia 44: -]2. Campbell, F. T. : S. E. Schlarbaum. ests. North American p and the eines xi iae Pests. Natural pop s Defense Council, ork. Carroll, G. 1988. Kuss endophytes in stems phe fe ves: From latent pathogen to mutualistic symbiont. Ecology 59: 2-9. 1994. Fading For- Cox, J. M. & D. J. Williams. 1981. An account of cassava mealybugs with a description of a new species. Bull. Entomol. Res. 71: 247-258. Duran, R. & G. W. Fischer. 1956. The genus — Washington State Univ. Press, Pullman, Washingt Evans, H. C. 1993. Studies on the rust, a. cryp- tostegiae, a potential biological control agent of rubber- vine weed, Cryptostegia grandiflora (Asclepiadaceae: Periplocoideae), in Australia, I: Life-cycle. Mycopath- ologia 124: 163-174. & L. Fleureau. 1993. Studies on the rust, Mar- avalia d a potential biological control agent of rubber-vine weed, Cryptostegia ru Ascle- dene Reriploci deac), in Australia, II: Infection. Myc i au 124: 175-194. . Tomley. 1994. Studies on the rust, Mar- m eryptosegiae a potential biological control agent of rubber-vine we Cryptostegia sina ig (Ascle- eee Pell Maidens, in Australia, III: Host 93-108. am >. P. Chamuris /. Rossman. 1989. Fungi on Plan nts x Plant P Produc ‘ts in M United e American Phytopathology Society, St. Paul, Min neso ELS amd R. B. & D. M. Spooner. 1994a. A reex- amination of species boundaries between Solanum meg- istacrolobum and S. toralapanum (Solanum sect. Petota, series E i 1 Morphological data. Syst. Bot. 19: 89- == 4b. A reexamination of species bou la bu Solanum —— a and S. toralapanum (Solanum sect. series s egistacro- in rw: data. Syst. Bot. 19. 106— Greer 84. Pests not known to occur in ds e United ida 's OF e limited distribution: Tilletia indica. PNKTO 58: 1-8. Hawkes, J. G. ). The Potato: Evolution, Biodiversity, and Genetic Mod Smithsonian Institution Press, bie es px Herren, H. P. Neuenschwander. 1991. Biological md if cassava pests in Africa. Ann. Rev. Entomol. 36: 257-283. Kendrick, B. 1992. The Fifth Kingdom, 2nd edition. Fo- cus Information Group, Newbu Klasser . Knutson (editors), poena tions for a National Bidkateal Survey. ‘pana i of Systematics Collections, Lawrence, Kan Kosztarab, M. & C. W. Schaefer. p: E uus of the North American insects and arachnids: Status and needs. Virginia Agricultural hue a Station Infor- mation Series 90: 1—247. Matile-Ferrero, D. 1977. Une cochenille Poi nuisi- ble au manioc en Afrique equatoriale, Phenacoccus ma- nihoti n. sp. Ann. Soc. Entomol. Frang. (n.s.) 13: 145- 152. Mordue, J. & J. Waller. Mycol. Inst. Descr. 719: Tilletia caries. Commonw. 1981. 1-2. Annals of the Missouri Botanical Garden National Research Council. 1991. Managing Global Ge- netic Resources: Agricultural Crop Issues and Policies. National Academy of Sciences, Washington, D.C National Research Council. 1993. Managing Global Ge- netic Resources. The U.S. National Plant Germplasm System. National Academy of Sciences, Washington, Ogoshi, A. mosis and intraspecific groups p er Kuehn. Annual Rev. Phytopathol. 25: 1 Parmeter, J. A. Editor. 1970. Rhizoctonia P Biology and erat Univ. California, Berkeley. 991. Discula destructiva sp. nov., cause of 1987. Ecology and pathogenici ity of anasto- onia solani Redlin, S. C dogwood anihrac nose. Mycologia 83: 633-042 Rehner, S. & G. J. Samuels. 1994. Taxonomy and phylogen ny of Gliocladium analysed from nuclear large subunit ribosomal DNA sequences. Mycol. Res. 98: 625-63 . Rossman, A. Y. 1994. Report p NIU controversa from ee erroneous. Pl. Dis 755-756. Samuels, G. J. & S Rohner aF oward a concept of genu M species in Kis eir die Pp. 186-188 in R. D. den & J. L. Vaughn (editors), Pest Manage- ment: Biologically Test Tee hnologies. American Can- cer Society, Washington, D.C. Shands, H. L. & J. H. Kirkbride, Jr. 1989. Systematic botany in support of agriculture. Symb. Bot. Upsal. 28: 1992. Smith, 1., D. McNamara, P. Scott € K. Harris. Quarantine Pests for Europe. CAB International with EPPO. Cambridge Univ. Press, Cambridge Sneh, B., L. Burpee & A. Ogoshi. 1991. Identification of Ri neona species. American Phytopathological Soci- ety, St. Paul, Minnesota Spooner, D. M. & R. G. van iden Berg. 1992. An analysis of recent taxonomic concepts in wild pow (Solanum sect. Petota). Genet. Res. Crop Evol. 39: 23-37. & K. es 992. iler ies of series relationship of Mexican and Central American wild po- tatoes (Solanum ii Petota): Evidence from chloro- plast DNA restriction site vana. Syst. Bot. 17: 432- , R. Castillo T. & L. i Lopez J. 1993. Synonymy within viU potatoes (Solanum sect. Petota: Solanaceae): The case of Solanum i pi Syst. Bot. 18: 209- 2 U.S Gak Office of Tec shnology Assessment. 1993. Harmful n non- xiu ee species in the United States, Md F-565. U.S. Government Printing Office, Washing- n, D. $ Vilgalys, R. & M. A. Cubeta. 1994. Molecular system- atics m eve iud of Rhizoctonia. Ann. Rev. Phytopathol. 32: 13 Waller, J & J. ord 1983. Tilletia indica. Commonw. Mycol. Inst. Deser. 7 Williams, D. J. & M. C. Gans de eae k. 1992. Mealy- vugs of Central le South America. CAB allen Cambridge Univ. Press, Cambridge. FISHERIES AGENCIES AND MARINE BIODIVERSITY! Michael Vecchione? and Bruce B. Collette? ABSTRACT In addition to the potential negative impacts on biodiversity from fishing activities, there are positive aspects as well. Fisheries agencies are amo species, ng the best equipped argeninations to examine questions from which they come. Systematics is the base from which involving marine biodiversity y on the accurate identification of iig many questions bal E must be addressed. Taxonomy is a critical tool for ecologists. Therefore, in addition to training new systematists, the systematics communi y velops and train other biologists to be profic must ‘ient in taxonomy. C oser cooperation between fisheries and systematics is evelop better ways to disseminate the information it de- nnd needed to develop the UE. and skills necessary for assessment and maintenance of marine biological diver The problem of conserving biological diversity has received so much attention that almost any sci- entifically literate person will have heard of it by now. It is rapidly becoming an international con- servation priority emphasized in both the scientific (e.g., Harper & Hawksworth, 1994; Eldridge, 1992) and popular (e.g., Sawhill, 1994) press. Govern- ments at all levels in nations around the world are debating and implementing legislative and execu- tive actions to assess and preserve biodiversity. The Convention on Biological Diversity adopted as part of the 1992 United Nations Conference on Envi- ronment and Development calls for countries to un- dertake two major tasks: (1) identify the compo- nents of biological diversity that are important for conservation and sustainable use, and (2) integrate biodiversity concerns into socio-economic plan ning. Institutions that bring together the people who manage, use, and study biodiversity are crucial for achieving long-term responsible management of bi- ological resources. The widespread debate about what biodiversity is has resulted in a consensus that three levels of diversity are included: genetic diversity within spe- cies, phylogenetic diversity (species diversity in- cluding consideration of higher-level relationships), and diversity of ecosystems. Debate continues over the relative importance of these components (Bar- ault Hochberg, 1992; Brooks et al. 1992; Franklin, 1993; Stiassny, 1992), but degradation at one level affects the other levels as well. For in- stance, reducing the phylogenetic diversity in mod- el ecosystems alters ecosystem function (Naeen et al., 1994, 1995). This is very important in the con- text of marine fisheries. MARINE BIODIVERSITY Consideration of marine and estuarine ecosys- tems generally has lagged behind terrestrial and freshwater concerns for biodiversity (Norse, 1993; Ray & Grassle, 1991). Aside from early compari- sons between tropical rainforests and coral reefs, which are spectacularly diverse and easily visited (Jackson, 1991), marine habitats have remained largely “out of sight and out of mind" at many of the colloquia on biodiversity. This is despite the fact that marine environments occupy 71% of the area and more than 9596 of the volume of the bio- sphere (Angel, 1993). A recent focus on marine biodiversity (e.g., National Research Council, 1995; Vincent & Clarke, 1995) has begun to correct this oversight. Points made in the many discussions on terres- trial biodiversity cannot simply be extrapolated to the marine environment. The nature of life in the sea is very different from that of terrestrial and freshwater environments (Peterson, 1992; Steele, 1985, 1991). This is especially true in the pelagic (Angel, 1992) and deep-sea (Grassle, 1991) realms. Many more differences in basic body plan, as rep- resented by diversity of phyla, are found in the sea ! We thank the members of the NOAA Biodiversity Budget team and its subsequent manifestation in the Healthy Coastal Ecosystems team. Aleta Hohn, Donna Wieting, and Roger Griffis were ve helpful. Richard Young and Ma ary zu provided constructive comments on a draft of the manuscri ional Oceanic and Atmosp tional M of Natural History, Washington, D.C. 2 eric n National Marine Fisheries Service Systematics Laboratory, Na- S.A. ANN. MISSOURI BOT. GARD. 83: 29-36. 1906. 30 Annals of the Missouri Botanical Garden than anywhere else (Ray & Grassle, 1991). Life history traits of marine organisms differ greatly from those on land or in fresh waters, particularly with regard to dispersal (Strathmann, 1990). Coast- al marine and estuarine ecosystems supply impor- tant services to people but suffer from anthropo- genic alterations, ironically resulting from the human attraction to the coasts (Ray, The difficulty of basic questions about the nature of biological diversity in the sea is increased by our comparative lack of knowledge about marine or- ganisms. Estimates of the number of marine species vary by orders of pa (Briggs, 1994; Grassle & Maciolek, 1992; May, 1992, 1994). Population characteristics i marine species are not easily comparable with the better studied examples on land (Palumbi, 1992). Evidence is accumulating for surprisingly high genetic variability of marine pop- ulations in currently recognized species such as the common American oyster (Palumbi, 1994) and for the presence of many complexes of morphologically very similar sibling species (Knowlton, 1993), which contrasts with the terrestrial situation. It has been argued that Recent extinctions are not very common in the sea (Culotta, 1994), and conversely, that such extinctions may be commonly occurring but we lack the knowledge to recognize them (Carl- ton, 1993). Even if there are fewer, widespread spe- cies and comparatively few extinctions, it is likely that such a situation increases the importance of each extinction for the health of the ecosystem. In- deed, understanding phylogenetic diversity in ma- rine animals with extensive fossil records, such as foraminifera and mollusks, may allow detailed in- vestigations of the history of life and the processes of diversification (Buzas & Culver, 1991; Jablonski, FISHING EFFECTS Marine fisheries are among the many human ac- tivities that impact diversity in marine ecosystems (Messieh et al., 1991). Fisheries, however, specifi- cally target biological resources for harvest. The impact of assorted fisheries varies with the methods employed (Norse, 1993). Whereas some particular- ly destructive methods, such as dynamite fishing, have been widely prohibited, other methods are a continuing source of controversy. For instance, con- cerns about the effects of trawling have been voiced for centuries (de Groot, 1984). Of particular con- cern lately have been bycatch, the incidental mor- tality of non-target species, and physical disruption of the environment (Kennelly, 1995; Hendrickson & Griffin, 1993). Numerous studies continue to show that these effects vary among habitats (partic- ularly bottom type) and target species (e.g., Hamre, 1994; Riemann & Hoffmann, 1991; Ryan & Mo- loney, 1988; Van Dolah et al., 1991). Recently developed fishing methods, such as large drifting pelagic gill nets made of synthetic materials, are controversial (Norse, 1993). The Jap- anese driftnet fishery for squid began in 1978. By as many as 36 million “tans” (monofilament gillnet panels 30-50 m long and 7-10 m deep) were being set each year by Japanese vessels (Yatsu et al., 1994a). The Japanese National Research In- stitute of Far Seas Fisheries estimated that between 1989 and 1991 the bycatch of this fishery included 57,675 cetaceans. Other bycatch of this fishery in- cluded millions of blue sharks, albacore and skip- jack tuna, pomfrets, and pelagic armorheads, as well as numerous fur seals, seabirds, sea turtles, salmon, and other fishes (Yatsu et al., 1994b). Sim- ilar numbers could be expected for the vessels of the Republic of Korea and Taiwan, which com- prised a third of the vessels setting driftnets for squid in the North Pacific (Fitzgerald et al., 1994). Reports of this bycatch led to a public outery to ban the use of pelagic driftnets, known as “walls of death.” The increasing efficiency of harvesting methods, together with increasing numbers of harvesters, of- ten has resulted in precipitous decreases in abun- dance within populations of target species (Rosen- 1993). In addition to the obvious economic problems, this can cause profound changes in the ecosystem. For example, in the fish- ery for bottom fish on Georges Bank (in the Atlantic Ocean east of Massachusetts), 67% of the fish caught in 1963 were the prized gadoids (cod and hakes) and flounders, whereas 24% was made up of unwanted dogfish sharks and skates. By the dominant catch had shifted dramatically, with 14% gadoids and 74% sharks and skates (Sissen- wine & Cohen, 1991). Such changes in populations of large predators could cause profound effects throughout the food web. Similar situations occur in both bottom and pelagic fisheries around the berg et al., world. The shift in species fished on Georges Banks is one response to the decreased abundance of some target species. Similarly, fishermen are searching deeper waters for additional species to exploit (Vec- chione, 1987), resulting in bycatch and other im- pacts in new areas. Another response has been de- velopment of methods to enhance population size by hatching and releasing the young (Omori et al., 1992). Taking this a difficult step further, some spe- cies are reared to harvestable sizes by either ex- Volume 83, Number 1 1996 Vecchione & Collette 31 Marine Biodiversity situ aquaculture or in-situ cage or raft culture (Tseng, 1992). Along with problems involving nu- trient loading, these culture methods have caused concerns about reduction in genetic variability in the cultured species (Upton, 1992). When exotic species are cultured, the introduction of these alien species (either accidentally or deliberately) into ecosystems (Carlton, 9) has caused substantial problems with serious ecological and economic re- sults, including reduction in the number of native species (Carlton, 1992). FISHERIES AGENCIES A major role of marine fisheries agencies has been to determine why catches of commercial spe- cies fluctuate widely. The overall goal has shifted from maximizing catch to achieving sustainable use of the renewable resources (Rosenberg et al., 1993). Many early efforts focused on field surveys of abundance or spawning biomass for data input in single-species population models. The resulting estimates of resource availability have been used with greatly varying success by fisheries managers to determine the amount of catch that can be al- lowed while maintaining commercially viable pop- ulations. The focus of fisheries management has pro- gressed from single species to multiple target spe- cies (e.g.. Murawski, 1993) to ecosystems (such as the large marine ecosystem approach of Sherman et al., 1990). For an ecosystem management effort to have any chance of success, information is need- ed on all abundant or ecologically important spe- cies. One aspect that has received particular atten- tion is variability in the recruitment of young stages of commercial species to the fisheries and the in- teractions of ecosystem dynamics with recruitment (Fogarty et al., 1991; Frank & Leggett, 1994). Over the years, fisheries agencies have increas- ingly had to deal with other marine resource issues. In addition to traditional foodfishes, other natural resource products (e.g., aquarium fishes, collecta- ble seashells and coral, etc.) are harvested from the sea, including some with biomedical importance (Wright & McCarthy, 1994). Also, the long history of managing marine populations made fisheries agencies the organizations of choice for protecting threatened and endangered species (Upton, ] as well as insulating the species from fishing activ- ities. In some countries, the agencies participate in the design and management of marine parks and other natural reserves. Along with the parks and the endangered species responsibilities, fisheries agencies become involved in regulating non-con- sumptive uses of living marine resources, such as whale watching off New England. Efforts to pre- serve ecosystem integrity and to protect coastal ursery areas have moved the agencies into the broad field of environmental protection and pollu- tion abatement. This in turn has forced the inclu- sion of pollution indicator species (Parker, 1991) into fisheries concerns. Often during difficult economic times, people try to supplement or replace lost income by harvesting 1987). Also, changes in strategies for managing fisheries re- natural resources (e.g., Vecchione, sources can cause widespread direct and indirect effects on the economics of coastal communities (Smith, 1995). One of the most difficult aspects of implementing new regulations is the resistance to changes in traditional fishing methods. Thus, hu- man cultural implications have had to be consid- ered in addition to attempting to manage the har- vest (e.g., Smith, These sores tasks have required the devel- opment of an extensive data-collection infrastruc- ture in addition to ongoing resource surveys. Many databases exist that contain vast detailed informa- fish, crustacean, and cephalopod populations and their rmore, many s tion about changes in abundance of many genetic variability. Furtherm pecimens have been deposited in archival museums (Collette & Vecchione, 1995). This combination of data and specimens is particularly important because histor- ical data can be found for comparisons with present or future conditions (Allmon, 1994; Tyler, 1994). Collette and Vecchione (1995) recently summa- rized the importance of systematics and taxonomy in fisheries. Many workshops and study panels have pointed to an upcoming crisis in the systematics of marine invertebrates (Winston, 1992). There is a lack of replacements for current research positions at the Smithsonian Institution and other major mu- around the world for many c ap of marine n & lime g. 1992). Over a two-decade ud (197 995), m number of fish specimens in d in the United States and Canada increased by 7796, while over the same period the number of curators/researchers respon- sible for those collections decreased by 7396 (Poss & Collette, 1995). A major reason for this world- wide decline has been a continuing decrease in funding, prestige, and number of positions in sys- tematics (Cotterill, 1995). There is a need to train additional systematists for placement in an in- creased number of positions, both in fisheries agen- cies and in the scientific community at large, so that experts are retained for every important grou of organisms. In addition to training additional sys- seums invertebrates (Feldm 32 Annals of the Missouri Botanical Garden tematists, the systematics community must find bet- ter ways to disseminate knowledge of their groups and train fishery biologists, ecologists, and others to use up-to-date taxonomy as a tool in their re- search. Resource management agencies should hire systematists to provide the agencies with needed expertise and to bear a share of the costs of funding systematics. Fisheries agencies that already are surveying for other fisheries-related problems could easily and with little added effort or expense expand those surveys to focus on questions of biodiversity. Co- ordinating these activities with museums and aca- demia would allow maximum return while mini- Hoagland, 1994). — mizing duplication of effort PROPOSED ACTIONS Some proposals made to the U.S. National Oce- anic and Atmospheric Administration (NOAA) are listed below as an example of how a federal agency can expand its efforts in marine biodiversity. We feel that these proposals could be applied to fish- eries agencies worldwide with only minor adjust- ments (Fig. 1). 1. DEVELOP INFORMATION SYSTEMS FOR BIODIVERSITY METADATA Fisheries agencies have databases on the distri- bution of most economically important organisms and some other species that live within their re- spective geographic area, in addition to concurrent environmental parameters. Museums are comput- erizing information on the specimens in their col- lections. Furthermore, visual information (e.g., vid- eotapes recorded by submersibles) has been archived and could be used to document biodiver- sity that was observed in areas difficult to sample conventionally (Felley € Vecchione, 1995). All of the data mentioned above can be accessed by me- tadata to form a marine biodiversity database that can address questions such as whether there have been changes in marine biodiversity similar to those reported in terrestrial and freshwater ecosys- tems. Biodiversity metadatabases can be constructed from minimal data: species name, locality, depth, date, and either catalog or station number to refer back to the original complete records. Accuracy of species identification and linkage to voucher spec- imens deposited in archival museums are vital to insuring taxonomic credibility of the databases. Such databases can provide a current and retro- detect spective picture of biodiversity to any changes that are occurring. Biodiversity databases will reduce duplication of collecting efforts in car- rying out marine biological inventories. Properly constructed, these databases can be integrated with other national efforts to catalog biological diversity, including all biomes (e.g., the U.S. National Bio- logical Survey and the proposed National Biodiver- sity Information Center). Such efforts are already well under way in Australia, Mexico, and Costa Rica. 2. EXPAND EXISTING SAMPLING AND MONITORING PROGRAMS Most fisheries agencies conduct field surveys to provide information for resource management. The major cost of marine sampling is putting a research vessel to sea. The cost of preserving a broad taxo- nomic suite of material for the study of diversity is comparatively much less. A team of taxonomic spe- cialists and field technicians should be added to fisheries laboratories currently carrying out re- source surveys. These personnel could be employed directly by the agencies or under contract from uni- versities, etc. They would be charged with sampling a broad array of organisms, not just those of eco- nomic importance. They would utilize additional types of gear and, if necessary, special techniques to preserve specimens. They would facilitate the flow of well-preserved voucher specimens to sys- tematic specialists at universities and museums, and study part of the material themselves. 3. DEFINE DETAILED QUESTIONS AND DEVELOP METHODS TO ASSESS AND MANAGE BIODIVERSITY Detailed achievable goals have not yet been de- fined for assessing and managing marine biodiver- sity. One such question currently being posed is whether an all-taxon survey of a marine area is fea- sible. A demonstration project, limited in time and area, should be established to identify specific. re- search and management goals and capabilities for long-term information and conservation needs. This project would involve specific sites of contrasting characteristics to define attainable goals, which then could be expanded as necessary throughout the nation’s waters. 4. INVENTORY SANCTUARIES AND RESERVES Sanctuaries and reserves often have been estab- lished based on politics rather than biology. Bound- aries have been drawn based on governmental ju- risdiction instead of knowledge about the life histories of resident organisms. Existing biodiver- sity in these areas cannot be maintained without Volume 83, Number 1 Vecchione & Collette 33 1996 Marine Biodiversity Assess and Maintain Marine Biodiversity Identify Develop y : existing biodiversity pics pn resources databases Information LEE Marine Biodiversity Management Survey and A —N : a ) Quality Control yY No, y Expand Identify Initiate existing Eos jr A additional Cooperative programs material Curatorships Outreach and Education versity Uni Fellowships Figure 1. The 10 actions proposed for fisheries agencies in the text, and how they fit into a plan to assess and maintain marine biological diversity. knowing what species live in each. In order to man- thesized. Second, these syntheses should be sup- age these areas, their species composition and plemented by additional collecting from ships and abundance must be inventoried. This task has two in-situ observations. Voucher specimens document- components. First, any existing information on the ing the occurrence of different kinds of organisms biota of each sanctuary and reserve should be syn- in the sanctuaries should be deposited at nearby 34 Annals of the Missouri Botanical Garden museums. After this, life-history information is — of the limited expertise available to identify many needed to determine whether populations are self- sustaining within the boundaries of the sanctuary or if modification of the boundaries is necessary to maintain populations of key species. 5. STUDY ADDITIONAL MATERIAL COLLECTED OR ARRANGE FOR ITS STUDY Much new material will be collected by the field- work described above and will need to be identi- fied. This will require an increased number of spe- cialists in the taxonomy of groups of marine organisms that now lack adequate systematic ex- perts. Specialists need to be added to museums to study poorly known speciose groups of marine in- vertebrates, such as small bivalves and gastropods, sponges, cnidarians, cumaceans, organisms para- sitic on fishes, various groups of worms, and mi- croorganisms. Such specialists are vital to assure the accuracy of identifications. Furthermore, a rel- atively small number of specialists will then be available to train other biologists in taxonomy as needed. In order for systematics to attract students, more positions and funding must be made avail- able. 6. DEVELOP A PROGRAM OF COOPERATIVE SYSTEMATICS CURATORSHIPS Insure that at least one expert exists for every major group of organisms by setting up a system of cooperative curatorships in museums holding major collections of marine specimens. These systematists would be hired or contracted by fisheries agencies with consultation of the museums in which they are located, similar to the National Systematics Labo- ratory at the U.S. National Museum of Natural His- tory. Agencies should insure that there are positions available for the systematists they train to identify organisms, write keys, study phylogeny, and pro- duce monographs. Taxonomic credibility must be maintained for the biodiversity program to be ef- ective 7. HELP FUND MUSEUMS HOLDING LARGE COLLECTIONS OF MARINE SPECIMENS Good collections must be maintained to avoid expensive repeat sampling. Information from a large number of such collections is needed to create marine biodiversity databases. Cooperation with museums is vital to the success of a marine bio- diversity program because museums hold most of the collections of marine organisms that serve as vouchers for species occurrence and employ most groups of marine organisms. Most museums are currently understaffed at all levels, so the necessary information will not become available for a ie time unless these institutions receive assistance Such funds could also be used to facilitate the in- corporation of important collections maintained by individual investigators at universities, marine sta- tions, and fisheries laboratories, into archival mu- seums so that the information can become part of the museum databases. 8. DEVELOP FELLOWSHIPS IN SYSTEMATICS Training additional marine systematists can be accomplished by developing a fellowship program to support students in cooperating graduate schools, similar to the U.S. Fish and Wildlife Service's Co- operative Research Program and NOAA's Cooper- ative Marine Education and Research programs (CMER) at several northeastern U.S. universities. Some universities will be associated with museums housing Cooperative Systematics Curatorships. These fellowships could provide additional training to current fisheries employees to fill some of the new positions described above. 9. DESIGN NEW WAYS TO DISSEMINATE TAXONOMIC INFORMATION New ways are needed to transfer information on taxonomy to a wide array of user groups and to simplify learning of a taxonomic discipline. Novel tools include multimedia computer keys to facilitate identification of marine biota by fishery biologists on shipboard, and fisheries agents collecting statis- tical information at landing ports. Computerized in- formation could be distributed via the Internet or on CD-ROM. These systems could best be designe in cooperation with ongoing project development such as that at the Smithsonian Institution and the Institute of Taxonomic Zoology at the University of Amsterdam. 10. PUBLISH MARINE BIODIVERSITY RESEARCH An outlet is needed for monographs related to marine biodiversity such as taxonomic revisions and the series of larval fish guides being produced by fisheries laboratories (e.g.. Matarese et al., 1989). Credible lists of species identifications and abundances in local ecosystems should be pub- lished either electronically or in journal format. Along with this, a marine biodiversity newsletter could be produced electronically for rapid dissem- ination of informal information. Volume 83, Number 1 1996 Vecchione & Collette 35 Marine Biodiversity WHAT ARE THE BENEFITS? A successful marine biodiversity program will produce information that will assist in managing living marine resources and assuring that represen- tative segments of the biota that inhabit the ocean today will be here for our children to appreciate. Fisheries agencies should take the lead in marine biodiversity research and conservation. An addi- tional benefit of a formal biodiversity program would be to demonstrate a pro-active position for fisheries agencies in understanding marine ecosys- tems for their future protection. Literature Cited Allmon, W. D. 1994. The value of natural history collec- tions. ig 37: 83-89. Angel, M 1992. Managing biodiversity in the oceans. Pp. 23- p^ in M. A. Peterson (edito Axe of Oceanic Life. An Na luative Review. a Stra- tegic and reci Studies, Washington, D.C. Sig- nificant Issues 5 Volume XIV, Number 12. 1993 a of the pelagic ocean. Con- ervation yw : 760-712. Barbault. R. E. Hochberg. 1992. dee and community level approaches to studying o international research programs. Acta Oecol. BT 146. Briggs, J. C. 1994. Species ees Land and sea com- pared. Systematic re o 43: 35. Brooks, D. R., R. L. Mayden & s . MeLennan. 1992. Phylogeny and biodive n Conserving our evolution- ary legacy. Trends Ecol. Evol. 59. Buzas, M. A. & S. J. a ot eon diversity and 41: 483— dispersal of benthic foraminifera. Bioscience 41: 489. series, Carlton, J. T the ocean: e e of near-shore Biol. : 1989. Man's role in changing the face of foe ‘al invasions and implications for environments. Conservation ij 2 Introdu ed marine and estuarine mollusks of North America: An end-of-the-20th-century perspec- tive. J. Shellfish Res. 11: 489-505. 1993. N a of marine invertebrates. r. Zool. 33: 496 PR B. B. & M. pe E tween fisheries and systematics. 20: 20-25. Cotterill, F. P. D. 1995, Systematics, biological knowledge and environmental conservation. Biodiversity and Con- 3-205. 1995. Interactions be- Fisheries (Bethesda) . Is marine biodiversity at risk? Science lridridge, N. (Editor) 1992. Systematics, Ecology, and the Biodiversity C ET Columbia Univ. ress, New York. Feldmann, R. M. Manning. tematic Pires in he age af blodiventity "J, Paleontol. 66: e - 7 isis in sys- Felley, J. . Vecchione. 1995. use by nekion on the ental slope using archived U.S. 93 videotapes from A e Fishery Bull., 73. Assessing habitat contin 262-2 Fitzgerald, 5. M., McElderry, K Hatan: abe. J. S. Park, Y. Gong & S.-Y. Yeh. = aka, l vM 1990— 199] North Pacific high seas driftnet sc uw observer programs. North Pacific Commission Bull. 5 Fogarty. M. J.. M. P. Sissenwine & E. Recruitment variability and the dynamics of exploited marine populations. Trends Ecol. Evol. 6: 241-246. ‘rank, K. T. & W. C. Leggett. 1994. Fisheries ecology in the context of ec ollo :al and evolutionary theory. An- nual Rev. Ecol. € Syst. 25: 401—422. Franklin, J. F. 1993. Preserving pena Species, ec pd or landscapes? Ecol. Appl. 3: 202-205. cs J. 99]. Deos benthic em Bio- nce ES 46 & N. J. Mac bk: 1992. Deep-sea species rich- ness: Regional and local diversity estimates from quan- titative bottom samples. Amer. Naturalist 139: 3 341. Groot, S. J. de. 1984. The impact of bottom trawling on benthic fauna of the North Sea. Ocean Management 9: 177-190. . 1994. Biodiversity and exploitation of the main fish stocks in the Norwe gian-Bar ents Sea ecosystem. Biodiversity and Conservation 3: 92. Harper, J boa Measurement and estimation preface. Philos Ser. B 345: rl Hendrickson, & W. L. Griffin. 1993. An analysis of manageme nt eh ies for reducing shrimp by-catc h in the Gulf of Mexico. North American J. Fisheries Man- agement l 3: 686-097. Hoagland, K. E. 1994. Risks and opportunities for nat- ural history collections: Moving toward a unified policy. 29-132. 73 1994. Biodiversity: . Trans ura Jablonski, D. 1993. The tropics as a source of re ary novelty through geological time. Nature 364: 144. Jackson, J. B. C. 1991. corals. i2 ience 41: Kennelley, S. J. 1995 deme Med trawl fahories. Rev. Fish Biology and Fish- 4. 2 and diversity of reef 15-48: WE issue of bycatch in Australia's eries 5: 213-2: Knowlton, N. 1993. Ce | in the sea. Annual Rev. wid a 24: 189-21€ Matarese, A. C., A. W. Kio ll, D. B. Blood & B. Vine 1989. Laboratory guide to early life hiatory stages of northeast Pacific fishes. NOAA Technical Re- port NMFS 80. i M. 1992. Bottoms up for the oceans. Nature 357: 79 1994. Biological diversity: Differences between land n sea. Pie Trans., Ser. B 343: 105-111. > | W. Rowell. m L. Peer & P. J. Cranford. D pos of trawling, dredging and ocean ebd on the eastern Canadian continental shelf sea- bed. Contine a Shelf Res. 11: 1237-1263. Murawski, S. A. 1993. Climate change and marine fish Piin a Forecasting from historical analogy. Trans. Amer. Fisheries Soc. 122: gres dirá S., L. J. Thompson, S. P. Lawler, J. H. Lawton & .M. Woodfin. 1994. Dec de Ei uersiby can alter jd performance of ecosystems. Nature 368: 734—737. & 1995. Em- paced evidence that dni lining species diversity may alter the performance of terrestrial ecosystems. Philos. c )—202. 1995. Understanding Marine 'ashington, Trans., Ser. B 347: 24 National Research Council. National Academy Biodiversity. Press, D.C 36 Annals of the Missouri Botanical Garden Norse, E. A. Editor. 1993. Global Marine Biological Di- versity. A Strategy for Puldmg Conservation into De- cision ioca Island Press, Washington, Omori, M., C. P. Norman & H Biodi- dcn ii impac ts aa fisheries id: transpor- tation. Pp. 63- Peterson Vag Diver- Center for arine speciation on a small plan- et. Trends Ecol. Evol. 7: 114-117 1994. Genetic es reproductive isola- tion, and ı marié speciation. Annual Rev. Ecol. & Syst. 5: 047— Parker, T. DN indicator species. Marine Pol- ib ii [s 01. T N. MTM 1992. Diversity of Oceanic Life. n E pen Review. Center for Strategic and International Studies, b Rows d Significant Is- sues Series n XIV. Num Poss, S. G. & 3. Collette. 1995. oet survey of fish collections à i United States and Canada. Copeia 1995: os Ray, G. C. 1 991. Coastal-zone biodiversity patterns. Bio- science 41: 490—498 & J. Pete F. Cianail 1991. Marine biological diver- n Bioscience 41: 453-457. Riemann, B. & E. Hoffmann. 1991. Ee ologic - conse- quences of dredging and bottom trawling in the Li- mfjord, Denmark. Marine Ecology Progress Series 69: 171-178. Rosenberg, A. M. J. Fogart . P. Sissenwine, J. Beddington eT G. Shepherd. er Achieving sus- tainable use of renewable resources. Science 262: 828— 829, Ryan, P. ( . Moloney. 1988. Effect of trawling on bird em m) E in the ope oe Peri |. er d region. EU wi Es Progress - Sawhill, J. C. The N versity. ae B onservancy Sherman. K., L. Ale a & B. D. Gold. 1990. Large Marine Ecosystems: Patterns, Process, anc Yields. American Association for the Advancement of Science, Washington. RS Sissenwine, M. P & E ohen. 1991. ductivity and f ies e of the northeast shelf ecosystem. 107-123 in K. Sherman, Al- exander & B. D. Gold (editors), Food Chains, Vie Ids. Nature Conservancy and biodi- 5-9, m I cz Resource pro- Models, and Management « f Large Marine Ecosystems. Westview Press, Boulder, Colorado. Smith, C. L. 1994, Connecting cultural and biological diversity ^ restoring northwest salmon. Fisheries (Be- thesda) 19: 20-26. Smith, S. (eos Social implications of changes in fish- eries regulations E commercial fishing families. Fish- eries van sda) 2 Steele, J 19 pP comparison of ee and ma- rine ec li al systems. Nature 313: 355-: eee: 1001. Can ec ological theory cross Sanden bond! P Theoretical Biol. 153: 425-436. Stiassny, M. L. J. 1992. Phylogenetic analysis and the role of sys Galante: in the biodiversity crisis. Pp. 109— 120 in N. E Idridge (editor), 1992. Systematics, Ecology. dE the Biodiversity Crisis. Columbia Univ. Press, New En ne R. R. 1990. Why ES eps evolve dif- erently in Js sea. Amer. Zool. 30: 197— id K. 1992. Mariculture on Se . Peterson (editor) a and bio- Di- es Hi st. Res. 3: 5 5 ( ries Bellos la) 17: 25. Van Dolah, R. F., P. H. Wendi & M. V. Levisen. 5 A study of the effects of shrimp trawling on benthic com- munities in two South Carolina sounds. Fisheries Res. 12: 139— $ Vecchione, fish, abd uris in the Gt s Mw 49: 14-22. Commercial fishing for gulf butter- ilf of Mexico. Marine Fish- erie Vincent, A. & A. Clarke. = d in DS marine a Ecol. Evol. 10: 55- Trond Winston, J. E. 199: En and E ce tion. Pp. 144—108 i in. N. Eldridge (editor) Sys- tematics, Ecology, and the Biodiversity C risis. gie a Univ. du Yor Wright, A. E. € P. J. McC Carthy. 1994. Drugs from the sea al ite Branch. Sea p 35(8): 10-19 atsu, A., liramatsu & S. Hay 1994a. A review of the Japanese squid driftnet fishery with notes on the cetacean bycatch. Reports of the ee Whaling Commission Special Issue 15: 365-3 y er Ma TUM Ab. alos of the Jap- on the hy-catch. 5-24. < anese squid driftnet fishery with notes North Pacific Commission Bull. 53(D): 5 IMPORTANCE OF James H. Oliver, Jr.? SYSTEMATICS TO PUBLIC HEALTH: TICKS, MICROBES, AND DISEASE! ABSTRACT Ticks are vectors of nematodes, protozoa, rickettsiae, spirochetes, other bacteria, m viruses that cause disease in humans à other animals. Complex relationships have evolved between me ular tick species, certain organisms that they harbor and transmit, and susceptible vertebrate host species. Effects of tick-borne pampa on host species vary from asymptomatic to fatal depending on the species involved and their erac It is critical that arthropod vectors and the organisms they transmit be identified accurately. € ere is a crisis in biosystematics of arthropods due to several factors noted in the introduction. Approximately 80 of the estimated 850 described tick species occur in the United States; although all feed on vertebrate blood, only a dee mall number impact people and livestock directly. These few species, however, are responsible for considerable ines and economic loss. Tick-borne Lyme borreliosis is prevalent in North America, qae and Asia. It accounts for more than 90% of all reported vector-borne disease in the United ied Totals of 9.677. 8.185. and 11.424 cases were reported for 1992, 1993, and 1994, respectively, by the Centers for Disease Control and Prevention. An overview of interactions among so spec te tick- borne i roorganisms, and hosts is provided, with major emphasis devoted to ticks of ^ Ixodes ricinus species complex and the Lyme disease spirochete Borrelia burgdorferi sensu lato. Systematics play a crucial role in ee m these angle Complex and fascinating relationships have due to several reasons, including decreasing num- evolved between particular tick species, certain or- bers of specialists competent to carry out the need- ganisms that they harbor and may transmit, and ed tasks. Difficulties are exac 'erbated. due to poor susceptible vertebrate host species. The effects of financial support. The advanced age of many sci- tick-borne organisms on host species vary from — entists involved in biosystematic research and lack asymptomatic to fatal depending on the species in- of new specialists being trained in this area suggest volved. that the condition will worsen unless attention is For obvious reasons, it is necessary to recognize focused on this problem. A recent report titled. tick species and the organisms they harbor before "Endangered Species: Doctoral Students in System- a rational plan can be developed for tick control or atie Entomology” (Daly, 1995) indicates there were prevention of disease. Likewise, if a tick bite has — 111 such students in 1992 in the United States and occurred, the clinician needs to know whether and Canada, down from 155 in 1982. Forty percent of how to treat the patient. Identification of tick spe- the departments surveyed that have insect system- cies is crucial to both decisions. Identification of — atists on their faculties have no doctoral students the transmitted organism is also desirable prior to in insect systematics. If the 28% drop recorded be- attempts at intervention, and this may involve com- tween 1982 and 1992 is extrapolated, the number plicated taxonomic considerations. The science of of such students might reach zero by 2017. These systematics, therefore, forms one of the foundations — facts do not bode well for a world experiencing new of public health, veterinary medicine. and agricul- and reemerging arthropod-transmitted and other ture. Regrettably, there is a crisis in biosystematics diseases. If the trend of neglecting systematics re- of arthropods, especially those of medical and vet- search and training continues, it will cripple the erinary importance (Oliver, 1988). Consortium Systematics Agenda 2000 goal to dis- The deterioration of biosystematics capacity is cover, describe, and classify the world’s species. It ! Appreciation is les a to Bill Black, Lance Durden, Lise Gern, Joel Hutcheson, Jim Keirans, Hans Klompen, Doug Norris, and David disi for Mie sn discussions and/or access to rie uh data. Thanks to Martha Joiner for editorial assistance. Parts of this work were supported by NIAID grant A124899 and CDC Cooperative Agreement No. U50/CC AIN, lts ien nts are ica the re sponsibility of the author and del Li necessarily represent the official views of CDC or NIH. ? Institute of Mibropedils and Parasitology, Dept. of Biology, Box 8056, Georgia Southern University, Statesboro, Georgia 30460, U.S ANN. Missouni Bor. GARD. 83: 37-46. 1996. 38 Annals of the Missouri Botanical Garden will also devastate the newly created United States National Biological Service. Clearly, there is a prac- tical and immediate need for expertise in system- atics. Ticks are obligate blood feeders in one or more active stages of their development. They serve as vectors (transmitters) of nematodes, protozoa, rick- ettsiae, spirochetes, other bacteria, and viruses that may cause disease in humans and other animals. In addition to serving as vectors of the above-men- tioned organisms, they may also cause severe prob- lems to hosts simply by feeding on them. Feeding results in blood loss and produces a puncture wound that can itself become secondarily infected. Injections of tick saliva from some species cause an immediate and/or delayed hypersensitivity. Strains of some tick species secrete a toxin that causes an ascending paralysis (tick paralysis) in their hosts that previously was confused with symp- toms due to polio. Clearly, ticks are a global med- ical and veterinary problem causing severe host re- actions and large economic losses under certain circumstances. However, only about 10% of the ap- proximately 850 described species impact humans and livestock directly. Approximately 80 species occur in the United States. Some species are host specific, with all developmental stages (larvae, nymphs, adults) feeding primarily on one host spe- cies (e.g., Boophilus microplus (Canestrini) feeds on _ cattle), has other species (e.g., Ixodes scapu- laris (Say)) may feed on at least 119 different spe- cies of reptiles, birds, and mammals (Anderson & Magnare lli, 1994), 1e eight major tick-borne diseases that infect humans in the United States today are Lyme dis- ease, relapsing fever, tularemia, Rocky Mountain spotted fever, ehrlichiosis, babesiosis, C wees tick fever, and tick paralysis (Spach et al., 1993). Of these, Lyme disease (Lyme borreliosis) is most prevalent in North America, accounting for more than 90% of all reported vector-borne diseases, in- cluding those vectored by mosquitoes, fleas, etc. Totals of 9,677, 8,185, and 11,424 cases were re- ported for 1992, 1993, and 1994, respectively, by the Centers for Disease Control and Prevention (CDC) (1994, 1995). Because of the impact of Lyme i (LD) in North America, Europe, and Asia, this disease has been chosen as an example to il- lustrate the importance of systematics to public health. Emphasis will focus on the fascinating in- teractions among vector tick species, the etiologic agent (spirochete) causing LD, and reservoir hosts of the etiologic agent. Inter- and intraspecific dif- ferences among tick vectors, among the causative spirochete species, and among vertebrates serving as tick and spirochete hosts are presented in rela- tion to enzootic cycles in nature and human LD. Prior to discussing these relationships, however, it is necessary to provide a brief background on LD to allow a better understanding of variations and interactions that will be emphasized. LYME DISEASE Lyme disease is maintained in an enzootic cycle in nature. The eycle involves spirochetes as the eti- ologic or Causative agent, rodents as the most com- mon, but not exclusive, reservoir hosts for the spi- rochetes, and ticks as the vectors of the spirochetes among the rodents or other animals. Humans are incidental to the natural cycle and are not involved in its maintenance. SYMPTOMS Lyme disease (LD) is a multi-systemic illness in humans. Early symptoms include an annular ery- thematous expanding skin lesion called erythema migrans (EM). It generally clears in the center and rash. Approximately ” is often called a “bull’s-eye 60% of patients present an EM rash, which occurs days to weeks after infection. Other early symptoms may occur with or without EM and include severe headache, extreme fatigue, joint pains, muscle pain, low-grade fever, swollen lymph glands, stiff neck, keratitis of the eyes, and a general feeling of body discomfort. These symptoms may or may not be followed by neurologic or cardiac symptoms and migrating attacks of arthritis (including pain and swelling) especially in the large joints. The neuro- logic complications include facial palsy, headaches, difficulty concentrating or sleeping, irritability, and poor motor coordination, Cardiac problems often in- clude irregular heartbeat and varying degrees of heart block that may Late or chronic stage LD often occurs months to years after cause dizziness. infection and usually involves chronic arthritis and/ or chronic neurologic effects. Clinical symptoms vary from patient to patient. Arthritis caused by Lyme disease is less common in Europe than the United States, but neurologic complications are more prevalent in Europe as is another LD skin manifestation, acrodermatitis chronica atrophicans. ETIOLOGIC AGENTS Lyme disease is caused by the spirochete Bor- relia burgdorferi sensu lato. Spirochetes are a group of motile gram negative spiral-shaped bacteria in the order Spirochaetales, which includes two fam- The ilies: Spirochaetaceae and Leptospiraceae. Volume 83, Number 1 Oliver Systematics and Public Health 1996 Table 1. Several examples of spirochetal infections." Spirochete Human disease Vector or source of infection Borrelia B. burgdorferi B. recurrentis B. hermsü A turicatae B. parkeri Leptospira Leptospira Leptospirosis (Weil’s disease) interogans Treponema Treponema Syphilis pallidum subspecies (subsp.) pallidum T. pallidum Bejel (endemic syphilis) subsp. endemicum T. pad Yaws subsp. pan T. pr Pinta carateum Lyme disease Epidemic (louse-borne) relapsing fever Endemic (tick-borne) relapsing fever Ixodid ticks Body louse, Pediculus humanus humanus Ornithodoros spp. ticks Exposure to contaminated animal urine Sexual contact, transplacental Direct contact with contaminated eating utensils Direct contact with infected skin lesions Direct contact with infected skin lesions "Slightly modified from Pavia (1994). genera Borrelia, Leptospira, and Treponema are members of these families and contain species that cause several diseases of humans (Table 1) (Pavia. 1994). The pathogenic spirochetes are long. up to 50 pm, but very thin in diameter and exhibit a spiral, helical shape. They have a relatively slow rate of growth: 20- to 33-hour division time com- pared to Escherichia coli’s 20 minutes. Most patho- genic spirochetes (Borrelia, Treponema) are mi- croaerophilic and were once thought to be anaerobes (Pavia, 1994). Borrelia burgdorferi has linear plasmids that code for outer-surface proteins. The spirochetes are extremely sensitive to elevated temperatures (238°C). Originally, Lyme disease in the United States and Europe was thought to be caused by the same Borrelia burgdorferi strain that was isolated in New York by Burgdorfer et al. (1982) and described by Johnson et al. (1984). Data now indicate that the etiologic agent of human Lyme disease worldwide is best recognized as B. burgdorferi sensu lato, comprised of B. DANA sensu stricto (Johnson , 1984; Baranton et al., , Borrelia garinit 1992), and pels afzelii (Canica et al. Bua et al., 1993). All three “genospecies” occur in Eu- nly B. burgdorferi sen- et al., rope and cause human LD. su stricto has thus far been reported to cause LI in the United States. Borrelia japonica found in Ja- pan is currently considered to be the fourth “gen- ospecies” in the B. burgdorferi sensu lato complex, but its association with human disease is presently unknown (Kawabata et al., f TICK VECTORS Borrelia burgdorferi is transmitted to humans by The main species responsible for infecting humans with LD in the United States are /. pacificus in the far west and /. scapularis in the central and eastern ticks in the Ixodes ricinus species complex. United States. In Europe, Z. ricinus is the principal vector, as is /. persulcatus in parts of eastern Europe eastward to Japan. There are areas of geographic overlap of /. ricinus and I. persulcatus in eastern Europe and western Russia. The previously men- tioned four most common vector species are oppor- tunistic, feeding on various species of reptiles, birds, and mammals. Some of the most common hosts of larvae and nymphs, such as the mouse Peromyscus leucopus in the northeastern U.S., also serve as excellent reservoirs for B. burgdorferi; whereas others, such as deer, supply a blood meal to large numbers of adult ticks, but apparently may play only a minor role (Oliver et al., 1992) or no role as reservoirs for the spirochete (Telford et al., There are ten other less common tick species worldwide assigned to the /. ricinus species com- 40 Annals of the Missouri Botanical Garden plex (Keirans et al., 1996a), and their vector status is unknown. Borrelia burgdorferi sensu lato also has been isolated from other /xodes species not consid- ered taxonomically to be part of the Z. ricinus com- plex, and some of these ticks are involved in sev- eral different enzootic cycles of B. burgdorferi in nature. Some of these enzootic cycles may impact humans indirectly by providing a source of infec- tive spirochetes. in California B. burgdorferi is maintained in a natural enzootic cy- cle in the dusky-footed woodrat Neotoma fuscipes by the maintenance vector /. neotomae. Ixodes neo- For example, tomae does not usually feed on humans, but Z. pa- cificus feeds on humans and woodrats and thus serves as a bridge vector (Brown & Lane, 1992). Most of the data available on Lyme disease in the United States originate from research conduct- ed in the hyperendemic northeastern region, where the white-footed mouse Peromyscus leucopus is the principal reservoir host for the spirochete and /xo- p des scapularis is the main tick vector (see Lane e . 1991, for review). VARIATION AND HETEROGENEITY OF I. SCAPULARIS Ixodes scapularis is extremely important from a public health and veterinary medicine perspective. It is the tick vector involved in most Lyme disease and human babesiosis cases in the United States. There has been controversy regarding its specificity for many years and this has intensified during the last decade. It provides a dramatic example of the importance of systematics to human welfare. When northern populations of /. scapularis were described as a separate species (/. dammini) and identified as the main vector of the Lyme disease spiroc ‘hete (B. burgdorferi), there was a question as to the ficiency of /. scapularis (southern sine eae to transmit the spirochete. The perceived absence of I. dammini (northern populations of /. scapularis) in the south was one of the reasons cited that Lyme disease occurred there rarely, if at all. Subsequent- ly, laboratory experiments demonstrated that south- ern populations were also vector competent (Pies- man & Sinsky, 1988; Mather & Mather, 1990). Later, when southern populations from nature were examined, specimens were found infected with B. burgdorferi (Oliver et al., 1993a). Because Lyme disease is so important, it justifies a fuller account and also serves as an example of the importance of systematics to public health. A complete taxonomic history of I. scapularis is provided by Keirans et al. (1996b). Briefly, Ixodes (Ixodes) scapularis was described by Say (1821); however, subsequently several tick experts consid- ered it a variety or a subspecies of /xodes ricinus (Nuttall € Warburton, 1911; Schulze, 1939). Later authors accepted /. scapularis as a species and Coo- ley & Kohls (1945) redescribed it. Since that time it was accepted as a species (Keirans, 1982, 1985). There appeared to be general agreement about the taxonomic status of /. scapularis for more than three decades, until 1979. At that time, a new species (/. 1979) based on populations of /. scapularis from the dammini) was described (Spielman et al., northeastern United States. Since the description of /. dammini, there has been disagreement regarding its status as a sepa- rate species distinct from /. scapularis. Recently, it was determined to be conspecific with Z. scapularis (Oliver et al., 1993b); and, since the name /xodes scapularis (Say), 1821, had priority over the name xodes dammini Spielman, Clifford, Piesman & Corwin, 1979, I. dammini was relegated to a junior subjective synonym of /. scapularis (based on Ar- ticle 23 of the International Code of Zoological No- menclature, Ride et al, 1985). Data supporting conspecificity of ZL. dammini and I. scapularis in- clude hybridization and assortative mating experi- ments, morphometric comparisons, chromosome and isozyme analyses, and laboratory life cycles (Oliver et al., 1993b). Earlier data indicated similar laboratory vector competency of B. burgdorferi (Burgdorfer & Gage, 1986; Piesman & Sinsky, 1988) and similar host preference and feeding suc- cess in the laboratory (James & Oliver, Additional support for conspecificity includes confirmatory chromosomal analysis (Chen et al., 1994) and comparison of sequence variation in the internal spacer regions of the ribosomal DNA genes among and within populations of /. scapularis and the former /. dammini (Wesson et al., 1993). These rDNA comparisons indicate that the geographic populations continually overlap and suggest that there is a continual gene flow (Wesson et al., 1993). Subsequent analysis of the variation in rDNA ITSI among eastern U.S. populations of /. scapularis in- dicates only 2396 sequence variation occurs be- tween regions (Georgia, Florida vs. North Carolina, Maryland vs. Massachusetts, New Jersey, New York), but 77% variation occurs within regions. This further supports the concept that all of the /. scapularis populations constitute a single species (McLain et al., 1995). Moreover, lack of evidence of clear differentiation between northern and south- ern populations was shown in a study on the phy- logeny of ticks based on mitochondrial 16S ribo- somal DNA sequences (Black & Piesman, 1994). While the latter work is not focused on the question Volume 83, Number 1 1996 Oliver 41 Systematics and Public Health of conspecificity or on intensive examination of ge- netic structure, the sequence substitution rate be- tween the individuals sampled from North Carolina and Massachusetts is equal to that found intra- specifically in the ticks Amblyomma americanum (L.) and A. variegatum (Fabricius). A more recent investigation dealing with molecular genetic varia- tion of the 16S and 125 mitochondrial genes of /. scapularis from northern and southern populations indicates much more genetic variation within southern populations (Norris et al., 1996). Almost 200 specimens of /. scapularis from numerous pop- ulations in the Northeast, the Southeast, and the northern and southern Midwest were examined us- ing single strand conformation polymorphism (SSCP) according to the methods of Hiss et al. (1994). Parts of the idiosoma from each specimen were used for PCR and the remaining parts (capit- ulum, coxae, and legs) were examined by J. E. Keirans (Curator, U.S. National Tick Collection) to confirm morphological identifications. Twelve to fourteen different presumptive haplotypes were ex- amined among the different regions. Genetic vari- ants were present among ticks from different pop- ulations even within the same states (for example, Georgia, North Carolina, and Oklahoma). Although the frequencies of haplotypes differed among the regions, there were no haplotypes that were unique to any one region. Moreover, morphological types were continuously distributed among the regions. An expanded multivariate morphometric analysis of specimens Minnesota, Massachusetts, Maryland, Missouri, North Carolina, progeny from reciprocal crosses between specimens from Massachusetts and Georgia, and /. pacificus Cooley & Kohls from California for comparison, in- dicates that “7. scapularis appears to be a polytypic species with a widespread geographic distribution in eastern North America” (Hutcheson et al., 1995). The analysis indicates latitudinal (Massa- chusetts—Georgia) and longitudinal (Georgia—Mis- souri; Minnesota—Massachusetts) clines. Measure- ments of 28 morphological characters of nymphs are continuous in contrast to meristic, and the pat- tern of geographic variation is overlapping in con- trast to disjunct. Hybrids resulting from MA x GA crosses appear morphologically intermediate be- tween northern and southern morphotypes. rom Georgia, F, HETEROGENEITY OF BORRELIA BURGDORFERI As noted in the section on etiologic agents, Bor- relia burgdorferi sensu lato is currently considered to be composed of four *genospecies." including burgdorferi sensu stricto, which includes the fos strain B-31 (Johnson et al., 1984; bag et al., 1992). B. garinii (Baranton et al., 1992), B. afzelii 1993), and B. japonica (Kawabata et EE] Canica et al., al.. 1993). The first three “ Europe; only B. burgdorferi sensu stricto is thus far reported from the United States. Borrelia garini, B. afzelii, and B. japonica are known from Japan; it is unknown if B. japonica causes disease in hu- mans. Prior to the recognition that B. burgdorferi several investiga- ~ genospecies” occur in consisted of four “genospecies,” tors noted that European isolates of B. burgdorferi are more Pid i bod those from the s States (Barbour et al., ; Anderson et al., à LeFebvre et al., 1990; pa hie 1991; "ud et al.. 1992). Although most B. burgdorferi strains from the northeastern United States appear to be rather similar, exceptions do occur. Exceptions include the antigenically variable Borrelia burgdorferi isolated from cottontail rabbits (Sylvilagus floridanus) and Ixodes dentatus from New York (Anderson et al., 1989); an isolate from a veery songbird (Anderson et al.. 1986; Barbour et al., 1985): a strain (25015) from /. scapularis from Millbrook, New York (Anderson et al., 1988: Fikrig et al., 1992); and perhaps others. Borrelia burgdorferi strains isolated from California appear to be antigenically quite variable (Brown & Lane, 1992) and more so than strains from the northeast- ern United States. gen D (Bissett & Hill 7; Bissett et al., strains from Z. neotomae (Lane & Pascocello, 1989) and /. spinipalpis (Craven & Dennis, 1993) are es- pecially unusual (David Persing, pers. comm.). Clinical symptoms of human Lyme disease pa- tients in Europe and the United States are similar in several respects. However, as already stated, less arthritic but more neurologic symptoms are report- ed in European than American patients. Further. an acute dermatological condition known as acro- — dermatitis chronica atrophicans occurs in some Eu- ropean patients but is extremely rare in the United States. Borrelia afzelii appears to be associated with late cutaneous manifestations (Canica et al., 1993) in European patients, and as mentioned above, has not been reported from the United States. Although not thoroughly substantiated, it has been suggested that perhaps each “genospecies” might produce some common clinical symptoms in patients and yet may also be associated more often with partic- ular symptoms not often produced by other “geno- species.” For example, in a study of European I Fas disease patients, infections due to VS B. afzelii) were associated with cutaneous symp- toms, whereas extracutaneous symptoms involved B. garinii (van Dam et al., 1993). More research is Annals of the Missouri Botanical Garden needed before this observed correlation can be con- sidered widespread. In general, there does appear to be a great deal of consistency among the B. burgdorferi isolates from the northern United States and similar clinical symptoms in patients from these areas. In contrast, persons in the south that are diagnosed with Lyme disease rarely exhibit arthritis that unequivocally can be attributed to Lyme disease. The patients may present with classic erythema migrans (EM) dermatologic lesions, flu-like and/or neurologic symptoms, yet rarely with arthritis. Since it is ac- cepted that patients in Europe have fewer occur- rences of arthritic but more neurologic symptoms and acrodermatitis chronica atrophicans, it seems reasonable to suggest that perhaps patients in the southern United States might present with some but not all of the symptoms common in patients from the north. This suggestion might be more appealing if it were shown that B. burgdorferi is geographi- cally widely distributed in the south, occurs com- monly, and is maintained enzootic ally among wild animals; and if B. burgdorferi in the south were genetic ally heterogeneous, with some strains simi- lar to those in the north and with others different. Currently, it appears that these requisites are being fulfilled. Recently, isolates from Oklahoma (Kocan et al., 1992), Texas (Teltow et al., 1991), Missouri (Oliver et al., unpublished), Georgia (Oliver et al., 1993a), Florida (Oliver et al., 1995), North Carolina (Levine et al., 1993; Levin et al., 1995), and Virginia (So- nenshine et al., 1993) demonstrate that Borrelia burgdorferi is much more widely distributed than originally assumed. Most of these isolates have yet to be thoroughly analyzed; however, the few that have been characterized, especially from Missouri, Georgia, and Florida (Oliver et al., 1993a; Oliver et al., 1995; Oliver et al., unpublished), show more variability than most northern isolates thus far re- ported. Isolates of B. burgdorferi from cotton rats, cotton mice, woodrats, and the ticks Z. scapularis and /. dentatus from Georgia, Florida, and Missouri show considerable heterogeneity among themselves and differences from most northeastern strains based on immunologic (monoclonal antibodies). protein (SDS-PAGE), and genetic (PCR) analyses (Oliver et al., 1993a; Oliver et al., 1995; Oliver et al., unpublished). Some strains also differ signifi- cantly in infectivity to mice (Sanders & Oliver, 1995; Oliver et al., unpublished). Pulsed-field gel electrophoresis (PFGE) analysis and DNA sequenc- ing of a portion of the gene encoding the large ri- bosomal subunit 23S also support the hypothesis of greater genetic heterogeneity among spirochetes isolated from the southeastern United States than that. found isolates from the northeastern and Great Lakes regions (Persing & Oliver, unpublished data). For example, a PFGE analysis of more than 200 Borrelia burgdorferi sensu lato isolates from throughout the United States, and including several from Europe, demonstrates a greater degree of ge- netic heterogeneity among southern isolates. Bor- relia hermsii, a species causing relapsing fever and transmitted by the argasid tick Ornithodoros hermsi in the western United States, is used as an outgroup for comparison to the B. burgdorferi sensu lato iso- lates. Borrelia hermsii is different from all B. burg- dorferi isolates. Among the latter, an isolate of B. garinii from Sweden (NBS16) and one from Russia (IP90) appear distinct from the others. The remain- der form two major assemblages: one composed of type strain B31, many isolates from the northeast- ern and Great Lakes regions of the United States, and a few isolates from the southeastern U.S. and California. Within the B31 assemblage, three of the subdivision groups of isolates are similar to the N40 isolate from New York, whereas three other groups of isolates are more similar to the type strain B31. Five of the subdivisions contain isolates from hu- man skin, cerebrospinal fluid, or blood. The largest number of isolates make up subdivision 5 and rep- resent the widest geographic range as sites of ori- gin. Sites include California, North Carolina, Flor- ida, Wisconsin, Illinois, Massachusetts, and Connecticut. Most of the human isolates analyzed are in this group, which also includes isolates from rodents and ticks. Subdivision 6 is composed ex- clusively of isolates from California (human skin, woodrat, the ticks 7. pacificus and I. neotomae), Georgia (cotton mouse), and two isolates from the Netherlands (human, /. ricinus). The second major assemblage contains a more genetically heterogeneous group of isolates from generally more temperate climates, i.e., the south- ern United States and California. However, this ma- jor assemblage also contains isolates from Colorado and the 25015 strain from upstate New York. All of the isolates analyzed in this large assemblage were from non-human origin. They were isolated from rodents (cotton mice, cotton rats, woodrats, kangaroo rats) and ticks (1. scapularis, I. dentatus, I. spinipalpis). Clearly, genetic analysis of B. burgdorferi sensu lato isolates by PFGE and DNA sequencing of a portion of the 238 gene support the contention that there is greater genetic diversity of strains in the southern and far western United States than present among strains from the northeastern and Great Volume 83, Number 1 Oliver 43 Systematics and Public Health Lakes regions. The trend could be even more sig- nificant than shown because relatively few southern isolates have been analyzed. It seems reasonable to consider the possibility that the heterogeneity of strains might produce variations in infectivity in humans as demonstrated in rodents (Sanders & Ol- iver, 1995). Perhaps different strains also produce variations in clinical symptoms and pathology as 1993). van Dam et al., —. well DISTRIBUTION AND ORIGINS OF IXODES SCAPULARIS, I. PACIFICUS, AND BORRELIA BURGDORFERI The origins of I. scapularis, I. pacificus, and B. burgdorferi are unknown. Ixodes scapularis was de- scribed in 1821, but it must have been present in the United States much earlier. If it were present prior to the last Ice Age, it may well have been eliminated at that time except from the southern part of its distribution, which was not covered with glaciers. Another possible origin of /. scapularis might have been by the transport of I. ricinus from Europe to eastern North America by explorers and settlers and subsequent rapid speciation of /. sca- pularis from I. ricinus. As noted above, I. scapularis was considered a variety or subspecies of /. ricinus by tick experts (Nuttall & Warburton, 1911: Schul- 1939) in the early part of the century. More recent studies (Cooley € Kohls, 1945; Keirans. 1982, 1985) list it as a separate species. We are currently investigating the species relationships among /. scapularis, I. ricinus, I. pacificus, and l. persulcatus. An hypothesis being tested is that at some point 7. scapularis evolved from European /. ricinus, and I. pacificus evolved from Asian /. per- sulcatus. We are testing the hypothesis initially by attempting to evaluate genetic relatedness of the species. Hybridization attempts are under way be- tween Georgia 7. scapularis and Swiss I. ricinus and soon will begin between California /. pacificus and Russian /. persulcatus. The I. scapularis X I. ricinus cross has produced F, hybrids that have now been reared to the adult stage. Fertility of these F, adults is now being evaluated. Analyses of DNA sequenc- es of various geographic populations of /. scapular- , L ricinus, l. pacificus, and I. persulcatus are also in progress. Morphometric analyses of several southern and northern populations of /. scapularis indicate greater longitudinal variation among the southern (Georgia—Missouri) than among the north- ern (Massachusetts—Minnesota) populations (Hutch- eson et al., 1995). Genetic sequence data of the mitochondrial 16S and 125 genes also indicate greater variability among southern populations (Norris et al., 1995). Shannon diversity indices of mitochondrial haplotypes calculated from haplo- type frequencies among the southeastern, midwest- ern, and northeastern populations indicate greatest genetic diversity in the southeast, least diversity in the northeast, and intermediate diversity in the midwest (Norris et al., There was approxi- mately three times more diversity in the southeast- ern than in the northeastern populations that were analyzed. Most phylogenetic analyses assume, and for some species it has been shown empirically (Ba- ker & Stebbins, 1965), that genetic diversity is greatest in regions from which a species or popu- lation originates. The opposite view would be dif- ficult to explain based on traditional models of mi- gration and range expansion. This suggests that the I. scapularis tick populations in the northeastern areas may have been derived from those in the southeastern United States. Using the same rationale applied above to the origin and distribution of the tick /xodes scapularis, it is noted that there appears to be greater diversity of Borrelia burgdorferi sensu lato in Europe com- pared to the United States. If this concept does not change after additional isolates from southern us western U.S. locations are thoroughly analyzed. would suggest that B. burgdorferi sensu lato may have been present longer in Europe. If the same rationale of greater genetic diversity is applied to origin and distribution of the spirochete in the Unit- ed States, it appears that B. burgdorferi may have been present in the southern United States earlier than in the north. This notion is counter to the com- monly held assumption that B. burgdorferi is cur- rently extending its geographic range from the hy- perendemic Lyme disease region of the northeast to other regions. Indeed, as already noted, until now many persons presumed, in the absence of data, that B. burgdorferi did not occur in the south. Data presented earlier in this paper and elsewhere dis- putes that presumption. Current data, much of which have yet to be published, indicate that B. burgdorferi is widely distributed in the southeastern states and is abundant in some locations along the south Atlantic coast. Data on the antiquity of Borrelia burgdorferi in North America are limited to two reports. Persing et al. (1990) reported B. burgdorferi specific DNA using polymerase chain reaction techniques) — museum tick specimens that were collected from Long Island, New York, in 1945 and in mouse spec- imens collected from Massachusetts in 1894 (Mar- shall et al., 1994). reported using ticks and mammals from the south- Similar analyses have yet to be ern United States. In any event, such analyses are likely to be based on relatively recent specimens 44 Annals of the Missouri Botanical Garden and it seems likely that the tick/B. burgdorferi cycle is one of great antiquity, dating back hundreds of years, perhaps to the last Ice Age CONCLUSIONS The discipline of systematics is one of the foun- dations of science. Taxonomy of arthropod vectors and of the microorganisms harbored and transmit- ted by them has an enormous impact on public health globally. Taxonomic identification of arthro- pod vectors and pathogenic microorganisms is nec- essary if we are to make informed decisions about strategies to control existent and emerging diseases. Intervention in disease requires a knowledge of the identity of the causative organisms involved and their taxonomic relationships to related species. So- ciety is best served when systematics and taxonomy utilize traditional and newer molecular techniques concurrently to focus on particular questions or problems. Science administrators, politicians, and the general public need to be informed of the im- portance of systematics to public health, agricul- ture, environmental issues, and other areas of im- mediate human welfare, as well as its pivotal role in inventories of biological diversity on planet ‘arth, Unfortunately, many taxonomic specialists are nearing retirement age and upon retirement are not being replaced. The alarming report, “Endan- gered Species: Doctoral Students in Systematic En- tomology” (Daly, 1995), notes that such students in the United States and Canada have declined from 155 in 1982 to 111 in 1992 (2896 drop) and that 40% of the departments surveyed that have insect systematists on their faculties have no doctoral stu- dents in this fiel Literature Cited joris J. F. 1991. y eges of Lyme borreliosis. . J. Infect. Dis., Suppl. 77: 23-34. agnarelli. 1994. Lyme disease: A tick- assoc dated disease originally described in Europe, but named i a a town in Connecticut. Amer. Entomol. 40: 217-2 & J. B. McAninch. 1988. New Borrelia burgdorferi ae variant isolated from /xodes dam- mini from upstate New York. J. Clin. Microbiol. 26: 2209-2212. , R. C. Johnson, L. A. Magnarelli & F. 1986. Involvement of birds in i. EE Ja the Lyme disease agent Borrelia burgdorferi. Infect. Immun. 51: 394-396. a A. Magnarelli, R. B. LeFebvre, T. G. Andread- s, J. B. McAninch, G. C. Perng & R. C. Johnson. 1989. Antigenic ally variable Borrelia burgdorferi isolated from cottontail rabbits and /xodes Taher in rural and urban areas. J. C Clin, D robiol. 27: 13-20. Baker, R. C. & G. ga 1965. The Genetics of idus Species. "Ao ademic Press, New Y Baranton, G., D. Postic, S. Girons, P. Boerlin, J.-C. Pif- aretti, M. Assous & P. A. D. Grimont. 1992. Delin- eation of Borrelia burgdorferi sensu stricto, Borrelia gar- ov., and Group VS461 associated with Lyme D . J. Syst. Bacteriol. 42: 378-383. Barbour, A. G., R. A. Heiland & T. R. Howe. 1985. Het- des of major iiri in Lyme disease borreliae: molecular analysis of North SEO 'an and European sie J. Infect. Dis. 152: 4 — Tessier & S. F. a 1984. Variation in a major ón protein of asi disease spirochetes. 4—10t . 1987. TNT ‘terization of Borrelia burgdorferi strains isolated from Ixodes A VM ticks in California. J. Clin. Microbiol. 25: 2296-2301 . Probert & S. Kuras iig rm An rete isolate of Borrelia burgdorferi from a tick in California. se New York Acad. Sci. 539: 369-371. Black, W. C., IV & J. Piesman. 1994. Phylogeny of hard- and soft- taxa (Acari: Ixodida) based on mitoc :hon- drial 168 rDNA sequences. Proc. Natl. J.S : 10034-10038. Brown, R. N. & R. S. Lane. 1992. Lyme disease in Cal- ifornia: A novel enzootic transmission cycle of Borrelia iu la Science 256: 1439-1442. Burgdorfer, W. & K. L. Gage. 1986. Susceptibility of the black- n tick, Ixodes scapularis, to the Lyme dis- ease spirochete, Borrelia burgdorferi. Zentralblatt an Bakteriologie Mikrobiologie und Hygiene, Series A 26: 15-2 A G. PA L. — Benach, J. P. Davis, E. Grun- waldt & S Haye 1982. Lyme disease—A tick — spiroc chs Se rience 216: 1317— a Canica, M. M., E a du Merle, anton & D. Pc ae c3 Monoclonal mien for identification of Borrelia afzelii sp. nov. associated with late cutaneous manifestations of Lyme borreliosis. Scand. J. Infect. Dis. 25: 441-448. [intem for Disease Control & Prevention. 1994. Lyme disease—United States, 1993. MMWR 43: 564—565, 571-5 : 395. Summary—Cases of specified notifiable aby United States, cumulative, week ending De- Rd 3l, bio (52nd week). MMWR: 43: 967—968. Chen, C., Munderloh & T. J. Kurtti. 1994. Cyto- s a teristics of cell lines from ier scapu- laris (Acari: hopes J. er Peas : 42 pci R. A. pu The genus /xodes in Am Qu Ins gon Bull. 184. National f ies of Health, Bethesda, Maryland. Craven, R. & D. Dennis. 1993. tds id d burg- dorferi from Ixodes spinipalpis ticks in C a and Col- orado. Lyme Disease Surveillance Summary po 1-2. 16 Endangered species: Doc pis students in Ew entomology. Amer. Entomol. 41: 55-59. Dam, A. P. van, H. Ku ian K. Vos poe md njaard, A. C. P. Ramselaar, M. D 1993. Different genospecies of Borrelia burgdorferi are associated with distinct clinical manifestations of Lyme borreliosis. Clin. Infect. Dis. 1 708-71 Fikrig, E., s, W. ole D. H. Per Kantor & R. A. vell. strain 25015 sing, X. 1992, eid bore :C dn ‘terization of ou rotein A and vaccination against ut. 1 ind: 148: 2256-2260. Hiss, R. H., D. E. Norris, C. R. Dietrich, R. F. Whitcomb, Volume 83, Number 1 Oliver Systematics and Public Health D. F. West, C. F. Bosio, S. Kambhampati, J. Piesman M. F. Antolin & W. C. Black IV. 1994. Molecular bs onomy using single srana conformation E (SSCP) an of mitochondrial ribosomal D nes. Insect Molec. Biol. 3: 171-182. Hutcheson, H. J., J. H. Oliver, ss. 1995. Multivariate aa homi discrimina- tion of post-larval stages of the black-legged tick, Ixodes scapularis (Acari: Ixodidae), principal vector of the agent of Lyme Wed in eastern North America. J. Med. Entomol. la pres James, A. M. & J. H Oliver Jr. 1990. Feeding and host preference of immature Ixodes dammini, l. scapularis, oe L. duh (Acari: Ixodidae). J. Med. Entomol. 27: i. A C., G. P. Schmid, F. W. Hyde, A. G. Steig- aT & D. J. Brenner. 1984. Borrelia a sp. DE agent of Lyme disease. Int. J. Syst Mens 34: 4 97. Kawabata, H., T Masuzawa & Y. nis nite Ge- nomic analysis of Borrelia japonica sp. nov. isolated from Ixodes ovatus in Japan. Microbiol. Immunol. 3T: 8 irans, J. E. 1982. The tick collection (Acarina: Ixo- dode of the Hon. Nathaniel Charles Rothschild de- posited in the Nuttall and general collections of the British Museum veri ui Bull. Brit. Mus. (Nat. Hist.) Zool. Ser. 42: 1-36 e Henry Falkiner Area and the Nuttall tick catalogue, U.S.D.A. Mise. Publ. 1438. . H. € r, Jr. & G. R. Needham. Td The Diodes deba ricinus complex worldwide: Distribu- tions, diagnosis, and delimits of the group. Proc. IX Int. Congr. Acarology. Ohio Natural History Survey, Colum- bus (in press). se J. on, L. A. Durden & J. S. H. Klom- pen. 1996b. rr (Ixodes) bx pe Say hen ri: Ix- odidae): Redescription of all active stages, disinbatien, hosts, geographic variation, medical and veterinary im- Entomol. (in pres ., S. W. Mukolwe, G. L. Mur y, R. W. Barker 1992. Isolation of rah burgdorferi (Spiroc hactalas Spirochaetaceae) from /xodes sc apularis and Dermacentor Ed ow ticks (Acari: Ixodidae) in Oklahoma. J. Med. Entomol. 29: 630—633. & J.A rena 1989. Antigenic char- acteristics of Borrelia Varie were 2 ixodid ticks in California. J. Clin robiol. 27: 2344-2349, , J. Peal & W. E e n Lyme bor- reliosis: Relation of its causative agent to its vectors and North America and Europe. Annual Rev. En- tomol. 36: 587—609 EIS R. B., g. J. A. Brown & R. C. Johnson. 1990. DNA and des. reme of tick- derived isolates of s o ee from California. J. Clin. Microbiol. 28: e e a )person, s E. Norris & Reaver competence of the rice rat (Rodentia: creian for Borrelia pipe es ae Med. Entomol. 32: 2 Levine, J. F., C. S. ety J. B. Strider, Jr., M. Levin, Ryan, P. Howard, W. Coughlin, M. Knight & 5. Yang. 1993. Ticks, their hosts, and Borrelia burgdorferi on the outer banks of North Carolina. Pp. 7—8 in Apperson, J. F. Levine & E. L. Snoddy (editors), E of the 2nd Workshop on Lyme Disease in ~ Southeast, Raleigh, North Carolina, 7-9 Sept., 199: rshall, W. F., M, S. R. Telford HI, : E Rys, B. J. “tuted, D. Mathiesen, S. E. Malawista, A. Spielman & D. H. Persing. 1994. Detection of Borrelia burgdor- feri DNA i in museum specimens of Peromyscus leucopus. J. Infect. Dis. 170: E 027—1032. Mather, T. N. & M. Mather. 1990. d compe- tence of three edil ir (Acari) as vectors S Ps Lyme disease spirochete. J. Med. Entomol. 2 : 650. McLain, D. K., . Wesson, J. H. Oliver, Jr., . H. T d er an in ribosomal DNA ld transcribed spacers l among eastern populations of D des scapularis (Acari: Ixodidae). J. Med. Entomol. 353-300. Norris, D. E., J. S. H. Klompen, J. E. Keirans € W. C. Black IV. 1996. Population genetics of Ixodes scapu- laris based on mitochondrial 12S and 16S genes. J. Med. Entomol. ing ress). el G. H. F. & C. Warburton. 1911. ere a ‘tion 7 The genus pe Pp. 133- a n G. H. uttall, . Warburton, W. F. Cooper & I bu its A 'ks. A Monograph of the n Part am bridge Univ. E London. Oliver, J. H., Jr. 1988. Crisis in biosystematics of arthro- pods. Science NS 967. —— de E. Stallknecht, F. W. Chandler, Jr.. James S. McGuire & E Howerth. 1992. p. tion of o burgdorferi in laboratory- sete dii dammini (Acari: Ixodidae) fed on experimentally ino ulated white- tailed deer. J. Med. Entomol. 29. 98 0— 984. W. Chandler, ES D. E, ‘Stallknecht, B . MeGuir A. Cummins mission ol the Lyme disease spirochete from oo sast- ern United States. Proc. Natl. Acad. Sci. U.S.A. 90: 7371- E R. Owsley, H. J. Hutcheson, A. M. Jan Chen. i S. Thè E. M. Dotson & D. K. McLain. 1993), Conspecificity of the ticks Ixodes scapularis and Ixodes dammini (Acari: Ixodidae). J. Med. Entomol. 30: 54— , F. W. Chandler, Jr., F p elg F. H. Sanders, Jr., H. J. Hutcheson, L. C , BLS. pe aea & R. S. Eo 1995. Nada occurrence and. characteriza- tion of the Lyme disease spirochete, Borrelia pee feri, in cotton rats E peg hips from Georgia anc Florida. : Parasitol. 81: Pavia, C p. “of the eee spiro- sheen: i Spire :hetal Tick-Borne Dis. 1: 3 AES Persing, D. . R. Telford HI, P. N. Rys, | Dodge. T. J. Me A 7. Malawista & A. nd ey 1990. Detection of Horrela burgdorferi DNA in museum spec- imens of Ixodes dammini ticks. Science 249: 1420- 1423. Piesman, J. & R. J. Sinsky. 1988. Ability of Ixodes sca- prd Dermacentor variabilis, and Amblyomma amer- icanum (Acari: Ixodidae) to acquire, maintain, anc transmit Lyme Pe ee je ‘hetes (Borrelia burgdor- feri). J. Med. OX 25:4 339. Ride, W. D. Srt , G. Bernardi & R. V. Melville. Edens, . International Code of Zoolog- ical Sciences. International Trust for Zoological Nomen- clature in association with British Museum (Natural History), London Sanders, F. H., Jr. & J. H. Oliver, Jr. 1995. Evaluation Annals of the Missouri Botanical Garden of Ixodes scapularis, Amblyomma americanum, and Der- macentor variabilis (Acari: Ixodidae) from Georgia as vectors of a Florida strain of the Lyme disease spiro- chete, Borrelia burgdorferi. J. Med. Entomol. 32: 402- AC Say, [. 1821. An account of the arachnides E Wa United States. J. Acad. Nat. Sci. Philadelphia 2: Schulze, P. 1939. Über schwedische, rumünisc us und nordamerikanische Formen von /xodes ricinus (L.) und über Leder punctata Can. et Fanz. f. autum- nalis P. v für Zoologie 31A: 1—8. Sonenshine, D. E., R. Ratzlaff, J. M. Troyer, S. M. Dem- merle, E. Demmerle, S. Jenkins € B. Annis. 1993. Tick-host associations and mainte Du of Borrelia ^d je in pan Pp. 8-9 S. Apperson, J. & E. es (edito) Tus: of the 2nd Workshop on Lym North € Carolina, 798 Spach, D. H., i Y 6: L, Campbell, R. D. E. Anderson a T. ritsche. 1993. diseases in the a ais New England J. Med. 329: 936-947, ee in the Southeast, Raleigh, ept., 1993, E. Quick, Tic a borne Spielman, A., C. M. Clifford, J. Piesman & M. D. Corwin. 979. Human o on Nantucket Island, USA: Description of the vector, Ixodes (Ixodes) dammini, n. sp. (Acarina: Ixodidae) J. Med. Entomol. 15: 218-234. Telford, S. R., HI, T. N. Mather, S. I. Moore, M. L. Wilson & A. Spie n: 1988. Inc 'ompetence of deer as reser- — voirs of the Lyme ad spirochete. Amer. J. Trop. Med. Hyg. 39: 105-10 Teltow, G. J., P. Fournier & J. Rawlings. 1991. oe of Borrelia Bal eed bun arthropods col- . Trop. Med. Hyg. 44: ga Wesson, D. M., K "Mclain, J. H. Oliver, Jr., J. Pie & F. H. Collins; 1993. Investigation of the validity of species status of Ixodes dammini (Acari: Ixodidae) us- ing rDNA. Proc. Natl. Acad. Sci. U.S.A. 90: 10221- 10225. Wilske, B., V. Preac-Mursic, U. B. Gobel, B. Graf, S. are is, E. Soutschek, E. Schwab & G. Zumstein. 1992, OspA serotyping system for Borrelia burgdorferi bcd on reactivity with monoclonal antibodies and OspA se- quence analysis. J. Clin. Microbiol. 31: 340-350. cted in Texa SYSTEMATICS AND THE R. 1. Vane-Wright CONSERVATION OF BIOLOGICAL DIVERSITY" ABSTRACT This paper concerns the role of systematics in efforts to conserve es digersity. Biodiversity is seen both as an nU science (involving ecology and population es as we systematics), and as a socio-politic pa activity (because of the strongly anthropoc entric focus of the Convention on Biological Diversity). Systematies has number of key roles to play, especially with respect to pepe our limited and fragmentary lainwledue of biogr through the predictive power of natural classification, and in helping to set priorities for conservation when, as is inevitably the case, resources are limited. After examining ways in which systematists must support the growing needs of society to know more d the Earth's biota, it is concluded that, because of their unique insights into the subject, systematists have an equally strong responsibility to take an active lead in many of the issues relating to the study. use, and conservation “of biological diversity. — Dan Janzen (1993) has asked “What does trop- — itarianism” (Porritt, 1994) is the best strategy is ical society want from the taxonomist?” In relation open to debate (e.g., Allen & Edwards, 1995; to the urgent need to know more about biological Oates, 1995), but this is where most of the force diversity, this question raises further issues—no- — behind national and international conservation ef- tably, is biodiversity only really important in the — forts is now located. The Global Environment Fa- tropics, and does taxonomy only have a supporting — cility (Glowka et al., 1994), the interim financial role in the study of biodiversity? Here I will discuss mechanism of the Convention (currently adminis- all three questions, taking the view that taxono- tered by the World Bank), is the largest single mists, and systematists in general, need to be source of funding ever made available for work on proactive as well as supportive in their work. biodiversity—a funding source which systematists seem surprisingly slow to exploit. As Robert May SHOULD SYSTEMATISTS TAKE A LEAD IN THE (1990) has observed, “Without taxonomy to give STUDY OF BIODIVERSITY? shape to the bricks, and systematics to tell us how to put them together, the house of biological science The t of Global Biodiversity Strategy (Reid — ;. 4 siesningless Mable.” If systematics provides et al., 1992), one of the key documents produced before M 1992 UNCED conference, is "Gui uidelines communicate about biological diversity, what role lor astion to save, study, and nee mae biotic should systematists seek to play in the study, use, wealth sustainably and equitably.” The earlier Car- nd conservation of the Earth's biotic wealth? ing for the Earth (IUCN/UNEP/WWF, 1991) has the subtitle “A strategy for sustainable living.” The primary objectives of the Convention on Biological Diversity are “the conservation of biological diver- An impression can readily be formed from much sity ... the sustainable use of its components ... — of the modern bioscience literature that the study and the fair and equitable sharing of the benefits of biological diversity is the preserve, not of sys- arising from the use of genetic resources” (Glowka tematics, but of ecology. The study of diversity rests et al., 1994). Thus many of these goals are un- neither with one nor the other, but with both. Bio- the foundation of our understanding and ability to WHAT IS BIODIVERSITY? ashamedly anthropocentric and primarily con- diversity exists at the interface of pattern and pro- cerned with conservation for human benefit, rather cess, as for example in the twin hierarchy envis- than preservation of wildlife for its own sake. aged by Eldredge and Salthe (1984; Table 1), in /hether or not this “rampant, unapologetic util- phylogenetics and population genetics (the distinc- ! My sincere thanks are due to Peter Raven, Jay Savage. and Mick Richardson for the invitation to attend the 41st Annual Systematics Symposium. It was a delight to be in such stimulating company, meet many new colleagues, and "the Missouri Botanical Garden. | am grateful to Dave a g mst To Qe ~=3 dE deu oo o d € — La dii] z =) Sa IN e Ri 22 < Ac. 20 RE ETE UN A ES Se = FE: =a = w = = m hà, —— fi 2 0 E Eua Q =f c ee — C7 = 3E SS ^? wm m< o 5 = = z on 232 Mn Un R Se = c "n e o I 3 = = 3 > = J- EE z; D S 3 3 D z = ? Biogeography Bow Conservation Laboratory, The Natural History Museum, Cromwell Road. a SW 5BD, U.K. ANN. Missouni Bor. GARD. 83: 47-57. 1996. 48 Annals of the Missouri Botanical Garden Table dissimilarity between species, and basic uncertain- The twin biological hierarchy (based on El- 19 — dredge & Salthe, Genealogical hierarchy Ecological hierarchy [pattern] [process] Codons Enzymes Genes Cells Organisms Organisms [individuation] [physiological function] Demes Populations Species Local ecosystems Monophyletic taxa Biotic regions All life Entire biosphere tion between phylogenetic and tokogenetic relation- ships: Hennig, 1966), and in historical versus eco- logical biogeography (Myers & Giller, 1988). Olivier Rieppel (1988) has discussed in depth a number of the scientific and philosophical issues of the pattern versus process debate. He observed that the analysis of pattern and process represents dif- ferent and incompatible “ways of seeing,” as “the first emphasises discontinuity, while the other is based on the principle of continuity.” The two are complementary because pattern analysis offers the only guide to common ancestry, and so gives direc- tion to the analysis of process, while process anal- ysis is the only way of giving pattern a causal ex- planation. He finally concluded, given we have different ways of seeing biology, that we should ac- cept that each has its merits and faults, and that each is incomplete by itself. Being complementary, we should make use of both approaches to get as complete a picture of biology as possible (Rieppel, 1988, esp. 170-171). Because of this intimate in- terrelationship, it is essential that systematists work together with ecologists and population biologists to develop a fuller understanding of biological diver- and, most ur- sity, including its potential uses gently, to develop more effective strategies for its conservation. SYSTEMATICS AND THE USE OF BIODIVERSITY Current estimates of the number of extant spe- cies of organisms vary from 3 to 100 million; Ham- mond (1992) gave a carefully considered estimate of 12.5 million, but confidence limits on such fig- ures are incalculable. There is greater agreement about the number of species that have been for- mally described (in the region of 1.5-1.8 million) but, even so, the uncertainties of synonymy ensure that the limits remain difficult to assess (Solow et al., 1e complications of heterozygosity, variation in genome size and degree of similarity/ ties about how to delimit the number of genes even in well-known organisms, compound any attempt to estimate the protean diversity that occurs at the genetic level. Undeterred, Wilson (1992) has esti- mated something in the order of 10' nucleotide pairs as specifying the diversity among species, to- gether with more than 10' gene combinations po- tentially available per typical species. Taxonomists might seem like chemists with a knowledge of 10 elements from the Periodic Table, while geneticists might be compared to physicists trying to under- stand the Universe from the behavior of just a sin- gle fundamental particle. Is our ignorance of the biosphere really so extreme? a mealybug was discovered attacking cassava in Africa. By the end of 1989 this pest was causing massive crop losses throughout the African tropics. Initial attempts to find a natural biocontrol agent failed, until it was realized that the mealybug was part of an undescribed species complex, and that it only occurred naturally in the southern Neo- tropics (Cox & Williams, 1981). An appropriate en- cyrtid wasp was then found, tested, and introduced into Africa. The wasp has now spread over more than 12 million km? and gives effective control of the mealybug throughout its African range, with an annual cost benefit estimated at $200,000,000 (Herren & Neuenschwander, 1991) This example, involving a new pest and an equally unknown biocontrol agent, tells us some- thing special about the nature of our ignorance. Bi- ology has to deal with a staggering number of spe- cifics—there are thousands of mealybugs, and tens of thousands of encyrtid wasps. To solve the prob- lem, entomologists had to find out precisely which mealybug they were dealing with, locate this spe- "es in its natural habitat, and discover precisely which of all those wasps was one that attacked it. Phe only reason this could be done quickly is be- cause the existing classification of bugs and wasps, however incomplete, is sufficient to form a valid basis for prediction. Once the mealybug had been accurately located in the system, it was possible to predict its geographical origin, and then what sort of insect to look out for as a natural biocontrol agent. Our detailed knowledge of the biology of all these myriad species inevitably lags behind our knowledge of their classification, but the natural system allows us to extrapolate what knowledge we do have. Thus our ignorance of biology, while profound, is manageable insofar as our classifications are pre- dictive. Far from being a passive pigeon-holing ac- tivity as some seem to believe, classification has all Volume 83, Number 1 1996 Vane-Wright 49 Systematics and Conservation the properties of an intelligence system, often al- lowing us to go far beyond the seemingly limited information available. Although there are literally millions of undescribed species, nearly all animals and plants can readily be placed within meaningful families. Like hypotheses, good classifications “al- low limited data to be used with remarkable effect, by allowing interpolations through data-gaps, and extrapolations to be made to new situations for which data are not available” (Gregory, 1980). One of the most frequently proposed uses for bio- diversity is biochemical prospecting. Costa Rica’s INBio has even formed contractual agreements with international pharmaceutical companies. Gámez and Gauld (1993), in describing the Costa Rican experience, suggested that the Hymenoptera, with their multitude of species and pharmacologically active venoms and other secretions, are a poten- tially excellent source of valuable new chemicals. While this is undoubtedly true, the same can be said for almost any diverse group Schulz et al. (1993) examined the male phero- mone-gland secretions of 10 African milkweed but- terflies, separating 214 substances in 14 chemical classes. Individual species had from 12 to 59 com- pounds (excluding tetrahydrofurans, which were not systematically investigated), including a high pro- portion restricted to individual species, or just a few, usually closely related species. Among these were some compounds rarely or never found in na- ture before, such as the ketone 16-heptadecen-2- one from Amauris hecate, and the monoterpene (E)- 2.6-dimethyl-5-octen-1,8-diol from Danaus chrysippus. If we wished to prospect for closely related substances in other species, the existing cladistic classification (Ackery € Vane-Wright, 1984) would give us imme- diate and obvious clues—such as the chemically un- investigated Amauris dannfelti and A. inferna for the ketone, and Danaus gilippus for the terpene. This suggests that, armed with an appropriate classification, chemical prospecting need not be “blind”: if we find something interesting, we can look at closely related species with the heightened expectation of finding more of the same or related compounds, If, on the other hand, we want to select, say, the 10 most different species out of a sample of 1000, to maximize the chance of finding radically different chemicals per unit effort, we can make use of the diversity measurements developed in system- Either way, the natural classification offers the most intel- atic conservation evaluation (see below). ligent basis for biochemical prospecting, or for any other use that depends on predicting the biological properties of organisms. The better our classifica- tions are, the more explanatory power they will have (Farris, 1979, 1983), and the more effective they will be for making such predictions. SYSTEMATICS AND CONSERVING BIODIVERSITY There are two major strands to conservation bi- ology: where to conserve living things, and how to conserve them (Caughley, 1994). The latter repre- sents a major application of ecology and population biology. Regarding the former, systematists have been closely involved with recent developments 1994), including the formu- lation of new approaches to measuring diversity e.g.. see Forey et al., — that take account of phylogenetic relationships. According to Taylor (1978), the notion of diver- sity, as an interrelation of species richness and in- dividual abundance, was first recognized by Henry Walter Bates. Bates's idea, to evaluate the diversity of a locality and compare it with that of another, was later formalized as a@-diversity (Whittaker, 1965). Whittaker (1972), while elaborating a more complex scheme (a-, B- and 6-diversity) to take account of species turnover at varying spatial scales (Magurran, 1988, gave a review of the considerable variety of procedures that have now been proposed for the measurement of “ecological diversity”), also suggested that time, in addition to richness and spatial turnover, should be included in a more com- plete expression of diversity. Time is the primary dimension of evolution along which differences be- tween lineages accumulate Conservation biologists, concerned with the need to set priorities for the preservation of genetic di- versity, have recognized the same problem and pro- posed that this could be solved, to a first approxi- mation, by mobilizing information contained within the taxonomic hierarchy: “The size of the potential genetic loss is related to the taxonomic hierarchy because, ideally at least, different positions in this hierarchy reflect greater or lesser degrees of genetic difference and hence differences in such variables as morphology, behaviour, physiology, chemistry and ecology. Although the degree of difference (the gap) between genera and between species within genera varies both within and among classes, the current taxonomic hierarchy provides the only con- venient rule of thumb for determining the relative size of a potential loss of genetic material” (IUCN UNEP/ WWE, 1980). Vane-Wright et al. (1991) proposed a diversity metric sensitive both to individual taxonomic (hi- erarchical, not formal) rank and total number of species found within an area. This index, subse- quently called root-weight, was the first formal measurement of taxic diversity. In a series of pa- 50 Annals of the Missouri Botanical Garden liverwort liverwort moss moss horsetail horsetail fern fern cycad cycad ginkgo ginkgo conifer conifer angiosperm angiosperm Figure 1. Choosing three land plants from eight: based on the cladogram, which combination would maximize biodiversity? Character ric hness (on the hs E chooses liverwort, g or angiosperm, plus an any one of the series moss- Ginkgo; character combination rick l iverwort, fern, and conifer or SHgioepenii (Assuming all character changes are associated only with r ane Ad pa a ae number of character changes occur at each node; based on Williams & Humphries, 1994; hierarchical relationships of plants from Humphries & Parenti, 1986.) See text for explanation. pers, notably those of Williams (e.g., Williams, potential genetic loss” (or gain). In practice, how- 1993; Williams et al., 1991, 1993, 1995; Hum- ever, it is impossible to measure character differ- phries & Williams, 1994), Weitzman (1992), Nixon ences directly on a large enough scale, and so the and Wheeler (1992), and Faith (e.g., 1992, 1994), distribution of characters across taxa has to be the concept of taxic diversity has been refined and modeled. It is now agreed that this should be based clarified— notably with respect to the fundamental on the genealogical hierarchy, as expressed more question of what is being measured, and why. and more accurately by taxonomic ranks, clado- Humphries et al. (1995) concluded that the cur- grams, and phylogenetic trees. Debate continues, rent goal is to assess option value. This concept however, regarding which model of character offers “a means of assigning a value to risk aversion change should be applied (e.g.. empirical versus in the face of uncertainty” (McNeely, 1988), and anagenetic versus cladogenetic), and whether or not can be related to the task of “maximising the hu- differences should then be assessed in terms of man capacity to adapt to changing ecological con- character richness only, or character combination ditions” (Reid, 1994). If this is accepted, then we | richness (Williams et al., 1995). can abandon the insoluble problem of trying to as- Some idea of the two approaches is given by Fig- sign fixed values to individual species (Ehrenfeld, ure 1. Suppose we could only select for conserva- 1988) and focus our attention instead at the level — tion three of the eight land plant species shown. If of expressible and heritable characters (genes, species richness were the only criterion, then any traits, features, etc.; Faith, 1992, 1994; Williams et of the 56 combinations of 3 species from 8 would al., 1995), which, collectively, can be considered to be equally acceptable. But if we interpret the tree represent the fundamental currency units of option — subtending the eight species as a statement about value. their mutual phylogenetic relationships, then on the Thus a set of species can be evaluated in terms basis of just this information (making the assump- of the total number of different characters they rep- tion that a comparable number of character changes resent. The impact of losing (or the addition of) any occur between each node shown), only 10 combi- particular species can be measured in the same nations will maximize character richness (liverwort way, thus fulfilling the need to assess “the size of plus conifer or angiosperm, plus any one from the Volume 83, Number 1 Vane-Wright 51 Systematics and Conservation Table 2. Complementarity: the eight plants listed in Figure | have been allocated to five hypothetical areas; | = liverwort. m = moss, f = fern, y = cycad, g = ginkgo, n = conifer, a = angiosperm, h = horsetail. (Based on Underhill, 1994; see text for explanation.) Spe- cies l m f y g n a h Areal — E + E + + + = 2 + + $ = = = — AS - o —- + + + c 1 + T F = — — — — 5 = =- == + + + series moss-Ginkgo). Alternatively, we could choose to maximize character combination richness, but then only two possible combinations are selected: liverwort, fern, plus conifer or angiosperm. (The root-weight index, if the cladogram is rooted be- tween liverwort and moss, would restrict the choice to liverwort, moss, and horsetail—but this index is no longer regarded as appropriate.) Despite some significant differences in these results, when large numbers of taxa are involved, simple species rich- ness for an area usually turns out to be a good for both character combination richness (Williams & Hum- ) approximation character richness and phries, 1 We now seem close to a satisfactory theoretical basis for at least part of what May (1990) has called the “calculus of biodiversity.” Another, and perhaps even more significant aspect of the procedure, is provided by the concept of complementarity. WHERE Is BIODIVERSITY MOST IMPORTANT? COMPLEMENTARITY AND ITS IMPLICATIONS In Table 2 the eight land plants in Figure 1 have been allocated to five areas. Suppose vou were told that you could only choose one area for conserva- tion, which would you select? Guided by species richness alone, area 1 would seem an obvious choice. What if you were then asked to add a sec- ond area? The greatest number of additional spe- cies to the five already represented by area l is two (liverwort and moss), both of which can be added by area 2 or area rea 1 plus area 2 or 4 can thus account for seven of the eight species: the eighth (horsetail) could then be added by a third area (3 or 5). If, however, you were asked to select just sufficient areas from the outset to represent all eight species, it is obvious by inspection of Table 2 that areas 2 and 3 together include all of them. giving a more efficient final solution than adding areas step by step. starting with the richest first. In reality both procedures have a place. Often it may not be possible to represent all species, veg- etation types, or land forms from the outset. in which case a step-wise procedure may be the most appropriate. In other cases, it may be possible to select from the beginning a set of areas to represent all known (or vulnerable, etc.) biological attributes in a region, in which case a set-wise procedure will usually offer a more efficient analysis. (In this par- ticular example, it is interesting to note that appli- cation of a taxic diversity index, such as character richness, identifies both areas 2 and 3 as richer than area l, based on the hierarchy given in Figure 1, illustrating the point that species richness should not be regarded as the sole determinant of taxic diversity.) Both the step-wise and set-wise procedures re- flect the idea of complementarity: the degree to which specified areas, singly or in combination, represent the species or taxic diversity of an entire group or set of groups. Complementarity, first ap- plied by Kirkpatrick (1983) and Ackery and Vane- Wright (1984), and formalized by Vane-Wright et al. (1991; see also Margules et al.. 1988; Rebelo & Siegfried, 1992; Pressey et al.. 1993: Faith, 1994: Williams € Humphries, 1994) has much in common with B- and 6-diversity but, crucially, in- stead of just reducing taxon turnover to numerical values or indices, information on the identity of axa between areas is retained. Although deceptively simple, the emergence of e the idea of complementarity has been significant for biodiversity evaluation. This is because it has shift- ed attention from assessing areas on an absolute scale (e.g., richness or scoring index) to a relational scale (taking account of spatial turnover). In this way. all areas can be seen as part of a whole. For example, while tropical forests and coral reefs may be the richest biological systems on Earth, the very distinct biota of other ecosystems, such as those of ocean bottoms or the relatively species-poor higher latitudes, also have a unique contribution to make (Table 3). Complementarity provides a basic crite- rion for efficient and goal-directed procedures of area selection. PRIORITIES IN CONTEXT If we take into account quantitative effects of biodiversity (Cousins, 1994), particularly in rela- tion, for example, to ecosystem services (Ehrlich & Daily, 1993), or the value of local biodiversity to humanity (Gadgil, 1991, 1992), it is evident that all areas of the Earth should be seen as important. Recognizing unique value for a particular area does Annals of the Missouri Botanical Garden Table 3. Essentially extratropical plant familes (from Heywood et al., 1978). Most contain only one or a few species. S. Hemisphere N. Hemisphere Bipolar Lactoridaceae Trochodendraceae Cercidiphyllaceae Eucommiaceae Lardizabalaceae Empetraceae Juncaginaceae Cephalotaceae Leitneriaceae Posidoniaceae Paeoniaceae Diapensiaceae Penaeaceae Theligonaceae Misodendraceae Geissolomataceae Calyceraceae Hippuridaceae Cynomoriaceae neoraceae Limnanthaceae Butomaceae Scheuchzeriaceae not mean that adjacent areas are unimportant, or that they can be abused with impunity. Priorities should not be seen as merely choosing a few of the richest, or even most complementary sites, but more in differential allocation of resources (Vane- Wright, in press) to do the best we can, in relation to relative importance both in terms of pattern and process, across all areas of land and sea. Never- theless, for the establishment of a network of spe- cial areas to act as reserves to ensure that as much of the irreplaceable qualitatively different (genea- logical) elements of diversity survive, and are thus available to future generations, analyses based on complementarity among areas supporting vulnera- ble attributes (gene, species, assemblages) will be of great importance if the limited resources cur- rently available for biodiversity conservation are to be used to maximum effect (Pressey et al., 1993; Pressey, 1994). The whole must be managed as ef- fectively and sensitively as possible; within such a whole, a network of special reserve or management sites for biodiversity then has special significance and can serve a useful function. WHAT SHOULD SYSTEMATISTS Do Fon Society? REACTION Commenting on his own question, “What does tropical society want from the taxonomist?”, Dan Janzen (1993) observed that “The wording of the chapter heading is the message. I do not ask ‘What does the taxonomist have to offer tropical society?’ Tropical society’s needs recently have been, can be and should be a major rejuvenating force in sys- tematics.” According to this view society and, in particular, tropical society, is looking to systema- tists to make certain things possible. First and fore- most, according to Janzen, a “cleaned up” set of names and a manageable system is needed for “fil- ing, comparing, searching, recording and working with the species ... that constitute the ... boun- tiful biodiversity resource of tropical nations.” Based on Janzen’s ideas, and those of others like Stork (1994) and Nielsen and West (1994), I list below some areas in which systematists are being called upon to respond. General attitude change. There is a need to de- mystify taxonomy and, in the process, make taxo- nomic products (such as identification systems and catalogues) more accessible (Miller, 1994). Much of this attitude change relates to gathering system- atic data in the first place, and processing it in a way that not only offers self-satisfaction (peer group approval), but also satisfies the rapidly expanding needs of non-specialists. Improved taxon sampling, recording, and stor- age. Better distributional data, including bionom- ic information such as host associations, etc., are essential (Wheeler, 1995; McNeely, 1995). Existing information is often based on ad hoc sampling pro- cedures, resulting in a partial and disconnected coverage. Systematists should become involved with radical approaches to rational and cost-effec- tive methods for data collection and spatial mod- eling, as well as application of techniques for im- proving estimates of distributional patterns based on existing data (Margules & Austin, 1991, 1994 This need includes dealing with the almost over- whelming number of (mainly species-level) taxa that remain unrecognized and undifferentiated (Wilson, 1992). Once biological samples have been Volume 83, Number 1 1996 Vane-Wright 53 Systematics and Conservation made, their continuing availability in well-docu- mented and properly curated collections is funda- mental to future work, including the extraction of information (such as DNA data) not necessarily considered at the time of original acquisition (Vane- Wright & Cranston, 1992) Better systematic analysis. Because knowledge of the Earth’s biota will inevitably remain incom- plete, I have argued above that it is essential to place what knowledge we do have in as powerful a predictive system as possible (cf. Nielsen & West, 1994). This applies not only to the primary activity of classification, where phylogenetic methodology should be pursued vigorously in building a natural system, but also to secondary disciplines such as biogeography where, for example, the possibility of recognizing areas of endemism still offers much in terms of predictive power, or information (Platnick, 1991) User-oriented databases. In order to make sys- tematic, taxonomic, distributional, and bionomic data, together with information about the uses and values of different organisms, as widely available as possible, user-oriented electronic databases must be developed and made widely available by appropriate means (e.g., Internet: Miller, 1994; Cracraft, 1995). In order to do this, it to " understood that continuing problems will occur over costs and intellectual property rights, and these problems (including cost recovery and aca- demic recognition) need to be solved. Improved use of advanced technology. As part of the inevitable change to electronic methods for storing, analyzing and making systematic data more widely available, every opportunity should be taken to make inc reasingly imaginative use of computer and video technology. One of the most obvious ar- eas lies in the development of fully illustrated, mul- tiple-entry, interactive keys (currently based, for example, on CD-ROM technology: e.g., ETI, 1 Watson & Dallwitz, 1993), but many other oppor- tunities exist, such as the production of specia checklists or other products tailored to fulfill unique needs, or the application of shape-analysis to identification. Other advances in biotechnology. leading to automated identification procedures based on blood or other tissue samples, or rapid increases in the quantity and quality of sequence data, must also be expected. The community of sys- tematists should embrace these positive and excit- ing developments because, so long as they are properly set up, such information systems will give systematists more time to develop their basic and traditional taxonomic skills—skills that will remain fundamental to further development and applica- tion of systematics to the problems of biodiversity. Training new systematists. Under this heading we must acknowledge the need not only to be in- volved in training new generations of systematists, but also with institution building, such as the cre- ation of new museums and reference collections. Assistance with the development of national bio- diversity institutes (Gámez & Gauld, 1993) is likely to represent a particular challenge. Involvement with biodiversity projects. The emergence of “biodiversity” as a topic (Harper & awksworth, 1994) raises many issues, not least of which is the fear that it may only represent a tran- sient “band-wagon,” likely to run out of steam or backfire on those who become too deeply commit- ted. On the contrary, because the concept links concerns over the preservation of nature and its use directly to the needs of human society, it represents a fundamentally new way of thinking about biolog- ical diversity. Systematists need to play their part in the support of biodiversity projects, including such diverse activities as preparing user-friendly identification systems, training parataxonomists, being involved with surveys and inventory schemes, setting up museums, reference collections, and oth- er information systems, making conservation and environmental evaluations, and so on. Biodiversity is a key social issue (Machlis, 1992), and it is vital that systematists play their part, not least by being sensitive to and catering for user needs (including local names, natural products information, etc.— things sometimes considered outside our remit). Involvement with education. Because biodiver- sity is important and will remain so in the future, systematists should expect to play a full and active role in building public awareness (Cracraft, 1995), including, in particular, the education of young people (e.g., Yen, 1994). PROACTION The activities listed above are described in terms of responding to the needs of society for better, more comprehensive, and above all more accessible information about the Earth’s biota and its signifi- cance. Appropriate reactions by systematists to the needs of society undoubtedly form part of our re- sponsibilities (especially as society at large has al- ready paid for so much of the collections, libraries, and other paraphernalia essential for our opera- tions). However, as I will argue in the last section, it is also our responsibility to be proactive—to put Annals of the Missouri Botanical Garden forward new ideas and create better attitudes to biodiversity by taking initiatives based on our own unique insights. ake just one example. If we attempt to set con- servation priorities based on separate analyses of the distributions of species belonging to more than one taxonomic group, it is apparent that this typi- cally leads to different, sometimes totally different, conclusions about what actions are needed (e.g., Prendergast et al., 1993). Can such conflicts be re- solved? Systematists have proposed two approaches to solving this problem: taxon summation (Vane- Wright et al., 1994), and the use of higher taxa Williams et al., 1994). Although further work on both methods (which are not necessarily exclusive) is needed, the point here is that systematists tend to propose very different sorts of solutions than ecologists or population biologists. Many ideas will need thorough evaluation before we settle on the most appropriate information and methods for priority areas analysis. Only by mobi- lizing systematic data, and creating the means to interpret it in a logical and systematic way, will we be able to develop efficient plans and monitoring schemes for conserving biodiversity. A primary goal is simply to make the most of limited resources that, inevitably, will never be enough to do every- thing that might be considered desirable. S atics Agenda 2000 can lead here by promoting a network of systematists to create the wide range of System- species-level and higher-category databases which, together with appropriate analytical procedures, will be needed for a comprehensive approach to conservation evaluation. OUR RESPONSIBILITIES AS SYSTEMATISTS The burgeoning human population, driven by consumerism and poverty, is having a massively deleterious effect on biological diversity, through industrial pollution, r over- cropping, and ecosystem transformation. We see lo- resource appropriation, cal and global extinctions on the one hand, and the spread of a limited range of synanthropic species on the other, leading to extinction of taxa and even entire ecosystems, extirpation of myriad popula- tions, and widespread loss of complementarity. Di- versity is being diminished and homogenized. As human demand for resources in both the de- veloped and developing world continues to grow, so the rate of human-induced biodiversity loss accel- erates. The end-point is unknown, as are the con- sequences. All we can say is that the biosphere, on which our life is totally dependent, will be called upon to provide more and more food, timber, and fiber, more clean water, more clean air, while ab- sorbing more pollutants, all from a continually di- minishing biotic base, quantitatively and qualita- tive Faced with this alarming prospect, most nations have now agreed, at least in principle, to try to take some form of corrective action, to which almost uni- versal support (at least on paper) for the Convention on Biological Diversity bears witness. This requires that each nation ratifying the Convention should pursue, through various provisions outlined, the conservation of biodiversity, the sustainable use of its components, and the fair and equitable sharing of benefits arising from the utilization of genetic resources. As I have already suggested, this approach to biological diversity is radically different from pre- vious concerns of the conservation movement, such as the protection of rare and endangered species, or the preservation of wilderness areas. The differ- ence relates to the anthropocentric focus: human needs constitute both the threat and the solution. That is not to say that conservation will be com- promised: on the contrary, conservation has shifted, in theory, from pressure-group status to being part of the fabric of human society. Conservation, and specifically the needs and uses of biodiversity, are now in the realm of what might be termed social engineering. So biological diversity is, all of a sudden, big business. In relation to biological science in gen- eral, and to the community of professional conser- vationists, ecologists, systematists, and so on, this is good news. However, the arrival of biodiversity as a political issue does not just signal new sources of funds. We need to appreciate fully (unless we allow ourselves to be prey to the worst sort of cyn- icism), first that biodiversity is couched within a truly social framework (we are thus operating be- yond the strict confines of science, in the areas of policy and socio-economics), and second, that we have new and expanded responsibilities that we, the community of systematists, must face up to. This includes the fact that much of the best data about biodiversity lies buried in our collections and libraries. In my view, it is very much part of our respon- sibilities to take control and help shape policy through our own initiatives. We hold the best in- sight into the strengths and weaknesses of taxono- my and systematics. We, very largely, are paid out of public funds, and thus hold the responsibility not only to react supportively to the needs of society at large, but also to mobilize, make use of, com- municate, and even lobby for the unique insights Volume 83, Number 1 1996 Vane-Wright 55 Systematics and Conservation that only we are in a position to formulate or bring to bear. Our responsibilities go beyond simply giv- ing support when asked: we must also ensure that our knowledge and understanding are brought to the fore, to be judged useful—or not—as others decide. Death and the Compass is the title of a short story by Jorge Luis Borges (1970). A criminal investi- gator has been set the task of trying to discover the murderer of a Jewish academic. At the scene of the crime, and in response to a suggestion that the scholar was accidentally murdered by somebody re- ally intending to rob the man living next door, the investigator replies, “Possible but not interesting. You'll reply that reality hasn’t the least obligation to be interesting. And I'll answer you that reality may avoid that obligation but that hypotheses may not. In the hypothesis that you propose, chance in- tervenes copiously. Here we have a dead rabbi; would prefer a purely rabbinical explanation, not the imaginary mischances of an imaginary robber.’ A multitude of taxa are under threat of death. In sympathy with Borges’s investigator, we should not avoid the burden of providing an intellectually sound and satisfying solution. We need a systematic as well as an ecological chart of the biosphere (Eld- 1984; Rieppel, 1988), and should develop a systematic plan to outwit as many im- redge & Salthe, minent taxic deaths as possible. In short, we must endeavor to find what we believe to be an appro- priate scientific solution, which takes full account of the principles and insights of systematics, in the same way that Borges’s investigator sought an in- teresting explanation for the rabbi’s fate. But Borges’s story carries not only this message for us, but also a dire warning. The investigator, = besotted with pursuit of an intellectual game of death, ends up as the final, ultimate victim, While our work must be interesting, systematic, academ- ically sound, it must also be timely, realistic, tical. We prac- must be vigilant to ensure that System- atics Agenda 2000, or whatever we like to call our current game plan, is not merely self-seeking, not merely more of the same, time-worn, mutton dressed as lamb, old wine in new bottles, but really to fulfill our responsibilities as true guardians of biological is oriented outward, toward society at large, diversity. Literature Cited Ackery, P. R. & R. I. Vane-Wright. 1984. Milkweed But- terflies: — Popa and Biology. Cornell Univ. p New Aller M. & S Edwards. 1995. The rear e- use ed du from IUCN. Oryx 29: 92-t Borges, J. L. 1970. Labyrinths [42 pieces by Borges, in- cluding Death and the Compass, translated by Yates from an M ud Spanish version of 19: E De guin Books, Lo Caughley, G. 1994. p ‘tions in conservation biology. J. Anim. Ecol. 63: 215-244. Cousins, S. H. 1994, Taxonomy and ae ll measurement, or, will the Ar work? Pp. 397 in P. L. Forey, C. J. Humphries & R. I. Vane- Wright artes alent s and Conservation Evaluation. Oxford Univ. Press, Oxford Cox, J. & D. J. Wilii: 1981. An account of cassava mealybug (Hemiptera: Pseudococcidae) with a descrip- tion of a new species. Bull. Entomol. Res. 71: 247-258 Cracraft, J. 1995. The urgency of building global capacity for pos aa science. Biodiversity and Conservation 4: 4 Ehre in p. 1988. Why put value on biodiversity? Pp. 212-216 in E. O. Wilson & F. M. Peter (editors), Bio- diversity. National Academy Press, Washington, D.C. Ehrlich, P. R. & G. C. Daily. 1993. $ wu extinction and m aed n Ambio 22: 64—068. Eldredge, . N. Salthe. 1984. lution. E pes Evol. Biol. (R ey. editors) 1: 1 "a rarchy and evo- . Dawkins & M. Rid- ET $ e Birds of Europe CD-ROM (Macintosh Version 0). ETI, Unive rsity of Amsterdam. aa E P. 92. Conservati on e evaluation and phyloge- Biol. Conservation 61: 1-10. . Phylogenetic lion of organismal biodiversity. Philos. 345: 45-5 Farris, J. S. 1979, The E content of the phy- logenetic ae. m. Syst. Zool. 28: 483-519. pá The logical basis p posee netic analysis Pp. irs in d I. Platnick & V unk (editors), Ad- vances in Cladistics 2. Med, Univ. netic die orsity. 994. pattern nr the quantifica- Trans.. Ser. H Press, New Yor Forey, P. L., C. J. Humphries & | Vane-Wright (Edi- ay Systematics ri e rvation Evaluation. Oxford Univ. Pres xfor Gadgil. M. 1991. Conserving India’s biodiversity: The so- cietal context. Evol. 3—8. 1992. Coe biodiversity as if pe e mat- ter: ase ra don India. Ambio 21: 266— Gámez. OR I auld. 1993. valive apo to pi study " tropical biodiversity. Pp. 329—336 . LaSalle & I. D. Gauld (editors), Hyme- noptera id Biodiveiibe C T International, Walling- LU rends I 21. 5: Costa Rica: n eN for rc pe Glowka, L. F. Burhenne-Guilmin & H. Synge. 1994. A Guide to the Convention on Biological Diversity. IUCN, Gland, Switze and, Gregory, p 980. Perce i as hypotheses. Philos. Tran -B 290: 181-19 alae n 1992. Species inventory. Pp. 17-39 in B. Groombridge (editor), Global Biodiversity. Status of the Earth's Living Resources. Chapman & Hall, London. Harper. J. L. & D. L. Hawksworth. 1994. Biodiversity: Mea 'asurement and estimation. Philos. — Trans., Ser. B 345 5-12. Hennig. W. 1966. Phylogenetic Systematics (translated y D. D. Davis & R. Zangerl). Univ. Illinois Press, Ur- bana. Herren, H. R. & P. ipe of cassava pests in Africa. 36; 257-28: Heywood, V. H.. Neuenschwander. 1991. Biological Ann. Rev. Entomol. D. M. Moore, I. B. K. Richardson & W. Annals of the Missouri Botanical Garden T. Stearn (Editors). 1978. TR Plants of the World. Oxford Univ. Press, Oxford. Humphries, C. going "1986. Cladistic Bio- e— Orlan Uini: Press, Oxford. ———k« P. H. Williams. 1994. Cladograms and trees in Te ee Pp. 335-352 in R. W. Scotland, D. J. Siebert & D. Williams (editors), Models in Phylog- eny Reconstruction. "pos Iniv. Press, Oxford. ———, — —— & R. ine-Wright. 1995. Measuring biodive e "as far: conservation. Annual Rey. Ecol. Syst. 26: 93-111 IUCN, U 2 5 &W WE. 1980. World Conservation Strat- ing Resource Conservation E Sustainable De- še aah 1U¢ N, Gland, Switzerlar ——,—— 16 Carir ng for the Earth. A Strategy for dis Living. IUCN, Gland, Swit- zerland. Janzen, D. H. 1993. What does tropic ‘al society want from the taxonomist? Pp. 295-307 in J. LaSalle & I. D. Gauld (editors), ide noptera and Biodiversity. CAB In- ternational, Walling Kirkpatrick, J. B. 1083. Án iterative method for estab- lishing priorities for the selection of nature reserves: An example from Tasmania. Biol. Conservation 25: 127— 134. Machlis, G. E. 1992. The contribution of sociology to e research and management. Biol. Conser- vation 62: -170. Magurran, A. i 1988. Ecological Diversity and its Mea- sure big Pn Helm, London. Margules, C. R. € M. P. Austin (Editors). 1991. Nature Conservation: Cost Effective. Biological. Surveys and Data Analysis. "-— Publications, Melbourne. — ——— & ——. 1994. Biological models for monitor- ing species decline: The construction and use of data bases. Philos 3 344: 69-75. a Pressey. 1988. Selecting networks of reserves to maximise biologic ial diversity. Biol. C onservation 43: 63-76. ~ = = — Ss z = — wn Ps ^ pu ~ May, ug M. ). Taxonomy as destiny. Nature 347: 129— 13 Mc Na e M J. 1988. Economics and Biologic cal Diver- sity: De lacie and Using Economic Incentives to Conserve Biological Resources. IUCN, Gland, Switzer- land. 1995. Keep all the pieces: Systematics 2000 and work aaa Biodiversity and Conservation 4: 510- inn s. i^ Pii d Ua of world identification rvices: Networking. Pp. 69-80 in D awkswort i (e ditor), The Identific 'ation nd C nue ene of Pes Organisms. CAB International, Wallingford, L Myers, A & P. S. Giller (Editors). 1988. iib 'al Biogeography: An Integrated Approach to the Study of Animal and Plant Distributions. Chapman & Hall, Lon- es >, West. 1994. Biodiversity research and biologic val oli Transfer of da Pp. 101-121 in P. L. Forey, C. J. Humphries & Vane- Wright varie Syston matics i Conservation Evalua- tion. Varr v. Press, Oxfor Nixon, K. C. & E D. Wheeler. 1992. Measures of phy- oe Anois. Pp. 216-234 in M. J. Novacek D. Wheeler (editors), Extinction and Phylogeny. Colum- bia Univ. pos ss, New York. Oates, J. E The dangers of conservation by rural ord, Taylor, L. R. of dau and Difference. Blackw —&P. Conserv at E 55: 235- e i from the forests of Nigeria. )ryx 29: 115- Platnick, N. Tom P. Y. Ladiges, € ln: E areas d endemism. Pp. n. "nl . J. Humphries & L. W. yelli Austral Bic. CSIRO En 'ations, Rice. ecological science into prac- tic > p Pp. 345—353 in P. J. Edwards, R. M. May . Webb (ilo: ipd -scale RS and Con- ed Biology. Blac pn Scientific, Oxford. i 1, J. H. Lawton, B. C. Ev- 1993; Rare species, the co- incidence of diversity hotspots and conservation strat- egies. Nature 365: 335-337. Pressey, R. L. 1994. Ad hoc reservations: Forward or ss kward steps in developing representative reserve ystems? Conservation Biol. 8: 662-668. ———, C. J. Humphries Margules, R. I. Vane- Wright & P. H. Williams. 1993. Beyond opportunism: Key principles for systematic reserve selection. Trends Eco ui 124-128. . & W. R. Siegfried. 1992. n should erves be located in the Cape Floristic Region, South. Africa? ie for the spatial añadas of a reserve network aimed at max imising the protection of floral diversity. es Biol. 6: 243-252. Reid, W. V. 1994, npe. Y ea for — nature rese Dm a 1-13 in P. L. Forey, C. J. Humphries & ight (ec editors) Saa and ism tion salt rugs Univ. Press, Oxford. — ———, C. Barber & K. Miller 1992. Global od alega Guidelines for Action to Save, Study, and Earth's Biotic Wealth INPS and Equitably. World Resources Institute, Washington, D.C. ane O. 1988. meda of Comparative Biology. Birkháuser, Base Schulz, S., M. Boppré & R. I. Vane-Wright. wi Spe- cific mixtures of secretions from male scent organs of a tene butterflies Dr Philos Trans., er. B 161-181. E ur . A. Mound & K. J. Gaston. 1995. Esti- UM the rat of synonymy. Syst. Biol. 44: 93-96. tork, N. E. 1994. Inventories of a More n a question of ied Pp. 81-100 in P. L. Forey, € J. Humphries & R. I. Vane-Wright uis Systematics and Conservation Evaluation. Oxford Univ. Press, Ox- 1978. Bates, Williams, Hutchinson—A va- id n eal. Symp. Roy. Entomol. Soc. London (9): 1 un rdum: 1994. Optimal and suboptimal reserve iie drin Biol. Conservation 70: 85-87. In press. Identifying priorities for the 382-387 in K. Gaston (editor), Biodiversity: A Biology ell, Oxford. S. Cranston. 1992. Collections and system- atics. Austral. Biol. 5: 14-18. . Humphries & P. H. Williams. 1991. to protect? oye = the agony of choice. "a Smith Hf b Kite thing. 1994. ns asse nant ae taxic diversity by rer Pp. 3 326 in P. L. Forey, C. J. Humphries & R. I. Vane- iege (editors), Systematics and Conservation E salvat: Ox- ford Univ. Press, Oxford Volume 83, Number 1 Vane-Wright 57 Systematics and Conservation Watson, L. € M. J. Dallwitz. 1993. Families of Flowering ants—Interactive Identification and Information Re- trieval CD-ROM Dan 1.0 for MS-DOS). CSIRO Publications, oe ur Weitzman, M. L. 1992. On diversity. Quart. J. Econ. 108: 157-183. Wheeler, Q. D. 1995. Systematics, the scientific basis for inventories a indicensty Biodiversity and Conserva- tion 4: 476—489. 1965. Dominance and diversity in land plant communities. Science 147: 250-260. 197 pgs and measurement of species di- versity. Taxon cel a —251. Williams, P. H. i as: Using taxonomy to measure more o Y biodiversity Pp. 194—227 in T-Y. Moon (editor), International Symposium on Bio- diversity and Conservation. Korean Entomological In- stitute, Seoul. & C. J. Humphries. 1994. Biodiversity, taxonomic lidia, and rd acm in conservation. Pp. 269— 287 in P. L. Forey, C. J. Humphries & R. I. Vane-Wright (editors), Sy Ape s and Conservation Evaluation. Ox- ford Univ. Press, Oxford. . J. Gaston & C. J. Humphries. 1995. Do con- serv omita and molecular biologists value differences between organisms in the same way? Biodiversity Lett. 2: 67-78. Humphries & K. J. Gaston. 1994. Centres of sends plant pe E The eign way. Proc. Roy. Soc. London, Ser. B, Biol. Sci. 256: ——, ——Q ene a 1991. Measuring TE sity: Taxonomic lios for conservation priorities. 2 tral. Syst. Bot. : 605-079 — ——, R. I. Vane-Wright & Humphries. Me easuring bine for c sin conservati s. Pp. 309— LaSalle & I. D. Gauld (editors), Hy- menoptera and a arta ae iibe naona, Wal- 1993. Wilson, E. 0. Group, Londor L ` 1992, The Diversity of Life. Penguin 16 94. "The role of museums and zoos in influ- encing public zd towards invert tion. Pp. 213-2 J. Gaston, T. R. New & M Samways EL Parspeo tives on Insect Conserv alie. Intercept, Andover, ebrate conserva- TRANSFORMING ETHNOBOTANY FOR THE NEW MILLENNIUM! Michael J. Balick? ABSTRACT e past several decades, the science of ethnobotany has evolved from a discipline primarily concerned with making lists of useful plants in a particular aphic region or among an individual t ribe, to a multidisciplinary ogra endeavor focused on understanding the relationship betwee en plants and peo ople. Ethnobotanists are involved in projects rangin from community / leve l ere paln development in Samoa. The Be e result of the ss of this branch of biological rary ethnobotanic al paradigm are discussed, including in-situ germplasm stern U nited States, Fin al marke Neotropics, and lize Ethnobotany Project's multidisciplinary approach to the nagement and use of Sabal sthnopharmae ra al studies linked to drug tudy and conservation of t surveys in Mexico, the ma traditional medicines and the development of an ethno-biomedical forest reserve in Belize are also reviewed. Contem- porary ethnobotanical studies have value not only for the research questions they address, but as a way of catalyzing awareness of the value of biological diversity and support for its conservation among a broad range of people. The term ethnobotany was first proposed in a lec- ture by John Harshberger to apply to the study of “plants used by primitive and aboriginal people ’ (Anonymous, 1895). This initial concept of b Vase was typified in a paper . Walter Fewkes (1896), *A Contribution to Ethnobotany,” in which wrote of the work of his student Jwens, who initiated a study on the foods E food resources of the Hopi Indians. In this paper, which was published after Owens's death, Fewkes wrote about their collaborative en- deavors, presenting a list of the common names and uses of several dozen food species, and stating, “I simply wish to call attention to the interesting field of ethnobotany which the Hopi Indians furnish the ethnologist." This work reflects the style in which early ethnobotanical studies were undertaken— compiling lists of common and Latin names of plants used by an indigenous group. Efforts previous to that time were carried out un- bapor botany,” a term 3-1875) to include “all forms of the vegetable ord which the aborig- der the heading of coined by Steven Powers (187 ines used for medicine, food, textile fabrics, orna- ments, etc." Edward L. Palmer, a botanist who made comprehensive descriptions of the flora of the western United States, was also one of the first bot- anists to investigate the cultural significance of plants to indigenous people through his fieldwork 14d i Palmer, . Prior to the above-mentioned _~ endeavors, other investigators studied the use of plants by North American indigenous peoples, of- ten focusing on the medicinal values that these conferred. Ford (1978) calculated that a total of 904 studies had been published on native North Amer- ican ethnobotany before 1977 In the past several decades, the nature of eth- nobotanical investigation began to change, becom- ing more focused on studies of the relationship be- tween plants and people in the broadest sense, and employing multidisciplinary perspectives. Exam (1974), whic h combined botanical, linguistic, and utilization in- and that of Schultes and Hoffman 73), which united ethnomedicine and phyto- chemistry. The emergence of ethnobotany as a mul- ples include the study of Berlin et al. ( formation, (19 tidisc ‘iplinary science, springing forth from its foun- dation in systematic botany, has resulted in interesting and important research questions that are being addressed. This “new” ethnobotany links diverse disciplines, such as anthropology, botany, nutrition, ecology, conservation, economics, and ! | am grateful to Brian Boom, Robert A . Bye, Jr., Javier Caballero, Paul Cox, Edelmira Linares, Steven ie bes Danie la Soleri for providing me with information on men research. I thank the traditional healers who hav with me in the Belize Ethnobotany Project and allowed me to give voice to their Bui of Ix Chel Tropical Research Foundation have been my valued companions in Belize, and we are ve collaborated endeavors. Rosita Arvigo and Gregory grateful E our many friends and ao of this project. Jay Walker was very kind to prepare Figure l, as was Elizabeth Pecchia in typing the manuscri ? Director and Philecology onim x ia Botany, The New York Botanical Garden, Institute of Economic Botany, Bronx, New York 10458-5126, ANN. Missourt Bor. GARD. 83: 58—66. 1906. Volume 83, Number 1 1996 Balick 59 Transforming Ethnobotany pharmacology, in a way that generates new ways of thinking and a new style of scholarship. Workers in this field also recognize the need to take a much more active role in addressing important issues of human society such as biodiversity conservation, food shortage, and the development of new medi- cines. Contemporary ethnobotanists are also raising concerns about the ethical implications of their re- search endeavors—for example, who should share in the benefits of new discoveries such as medi- cines or foods from plants (Boom, 1990; Cunning- 1993; Axt et al., 1993). They are also forging closer ties with government and private sector re- ham, search groups working to develop new food, drug. and energy resources. One important aspect of this renaissance in the field of ethnobotany has been the increased level of public interest in this sub- ject. Through television documentaries, movies, magazine articles, and radio interviews, the re- search of ethnobotanists is being used to illustrate the importance of biodiversity and its conservation. In promoting the objectives proposed by System- atics Agenda 2000, research topics in ethnobotany can attract the at- it seems obvious that current tention of important constituencies whose thinking we wish to influence. This paper will highlight examples of the re- search activities undertaken by contemporary eth- nobotanists. It will summarize several papers pre- sented at a recent meeting of the Society for Economic Botany that provide examples of active field- and laboratory-oriented ethnobotany projects. as well as the author's efforts in Belize GERMPLASM CONSERVATION Recently, concern has been. expressed. that. ex situ germplasm conservation may not be sufficient for preserving the genetic integrity of selected crop species or the plethora of landraces cultivated by traditional peoples. Crop plant relatives found the wild are also at risk. Iltis (1974) suggested that in situ conservation was an important adjunct for ex situ conservation of crop plants. A combination of both strategies is now seen as desirable, in order to ensure preservation of the maximum amount of genetic diversity. Ex situ conservation activities have gone far beyond the intention of serving only crop breeders; such materials are seen as valuable for supporting low-input traditional agricultural systems and insuring their capacity to recover from natural disasters, as well as restoring the genetic diversity of crop varieties in traditional systems in areas where this diversity has been lost due to changes in the socioeconomic and/or biophysical environment. Soleri and Smith (1995) investigated the results of conserving two Hopi folk varieties of Zea mays — .. Ina USDA germplasm program (ex situ) and by traditional farmers (in situ). They tested three hy- potheses: “1) Ex situ conservation of populations of a maize folk variety results in significant changes to those populations; 2) There are no significant differences between populations of a maize folk va- riety maintained by Hopi farmers today; and 3) There are significant differences between a folk va- riety maintained in situ by farming households to- day and the same folk variety maintained ex situ ” Working with two traditional corn varieties, they measured by an institutional conservation program. significant changes in morphological characters such as anther color, glume color, glume bar, plant height, and central spike length. They concluded that, in terms of meeting the needs of traditional agriculture, the greatest concern for traditional va- rieties conserved using ex situ technology “may be the changes during conservation resulting from ge- if The data collected to test the second hypothesis seemed netic shift as compared with genetic drift.” to indicate that there were no differences between populations of a Zea mays folk variety maintained by contemporary Hopi farmers. Based on their data, however, they neither accepted nor rejected the hy- pothesis and, instead, concluded that further re- search would be required. The third hypothesis, suggesting significant differences between the same folk variety conserved in two different ways, was accepted. Despite the technical difficulties in test- ing these hypotheses it was clear that the invariable maintenance of material utilizing ex situ conserva- tion methods is nearly impossible and that the two conservation strategies yield different results, de- manding a reconsideration of crop genetic re- sources, conservation goals, and assumptions. Work such as that being done by Soleri and Smith, in- corporating the ethnobotanical perspective into germplasm conservation, is clearly in the forefront of efforts to preserve erop germplasm and will, Lam certain, help reshape a great deal of the future thinking on this topic. MARKET STUDIES Linares and Bye (1987) and their colleagues at the Jardín Botánico, Universidad Nacional Autón- oma de México, have carried out exhaustive sur- veys of market plants in Mexico since the 1980s. Their studies, mostly focusing on medicinal and ed- ible plants, have involved work with traditional 60 Annals of the Missouri Botanical Garden herb vendors, farmers, botanists, educators, and local health promoters. Linares and Bye (1987) compared medicinal herbs sold in the markets of central and northern Mexico to those sold in the southwestern United States and found that local people group plants into folk “com- healers, medical doctors, plexes” of species that share common names, mor- phological and aromatic characteristics, as well as uses. For example, the complex known in this geo- graphic region as “hierba ants” comprised several genera (Tagetes L., Artemisia L., Pimpinella L., and Illicium L.), all with the characteristic odor of anise. One observation from this study is that each of the four complexes of plants investigated was “labeled” by choosing an individual plant species considered the most valuable within the complex. Each partic- ular “label” plant tended to be sold far outside of its natural range, and local plants were often sub- stituted during times it was unavailable. The la- beled plant for the “hierba anís” complex was Tag- etes lucida Cav. In this study the authors suggested that plants categorized as such are considered most effective, whereas the other plants in the complex are given secondary status. Using results of Linares and Bye and similar studies, the staff of the Jardín Botánico has been able to implement public education programs aimed at increasing the level of botanical literacy. Each year, special weeks are set aside to focus on traditional medicine, and individual healers set up stalls in the garden to advise the public on the uses of medicinal plants. There are also special work- shops on medicinal plants of Mexico geared toward those interested in alternative medicine. In addi- tion, the garden promotes the use of edible plants of Mexican origin. One project was a series of lec- tures on the botany, history, culinary preparation, nutrition, and other aspects of edible greens, both wild and cultivated. The public was invited to sam- ple these species, and a publication with botanical information and recipes was issued. The staff of the Jardín Botánico used ethnobotany to communicate important messages about biological diversity, its utilization and conservation, to the public. QUANTITATIVE ETHNOBOTANY The quantification of ethnobotanical data was first undertaken in a study of the Chácabo Indians of Amazonian Bolivia (Boom, 1987). This was a ma- jor step toward a much more rigorous methodology where a statistical approach could be utilized. Eth- nobotanical studies of 360 species of vascular plants known by the Chácabo identified uses for 305 species. In a one-hectare plot, 78.7% of the Percentage of useful species (in all categories specified) per hectare plot to the indigenous groups stud- ed. Percentage of useful tree species from Indigenous group inventory sites Ke apor 76.8 Tembé 61.3 Chácabo 78.7 Panare 48.6 Source: Prance et al. (1987). tree species and 95% of the individual trees were utilized. Balée continued this work among the Ka’apor and Tembé Indians of Brazil, and Boom later studied the Panare of Venezuela; the results are summarized in Table 1 (Prance et al., 1987). Their conclusions indicate a particular need for conserving plant families, such as Palmae, Lecy- thidaceae, Chrysobalanaceae, and Malpighiaceae, that are utilized extensively by these four indige- nous groups, as well as the terra firme forest, which most of the useful species occur. These be ies were the first attempt to quantify the value of the forest to indigenous people and thus argue for its conservation based on the percentage of species used locally. Working in Tambopata, Peru, with mestizo peo- ple, Phillips and Gentry (1993a) proposed a new quantitative method for ethnobotanical studies. They studied tree plots in seven different forest types in a total of 6.1 hectares. “Family use values” were calculated for plants employed for construc- tion, in commerce, food, technology, and medicine. Using the same data set, Phillips and Gentry (1993b) questioned whether the age of the infor- mant had any effect on his/her knowledge of plant use. They found that in some use categories, such as medicinal plant lore, the bulk of the information is held by older people, and suggested that areas such as this should be the main thrust of ethno- Other students in this field, such as Miguel Alexiades and Ana Irene Batis, are currently undertaking field- work that uses the quantitative approach in eth- botanical studies and conservation efforts. nobotany. It seems apparent that a quantitative ap- proach will allow ethnobotanists to question more precisely forest inventory, economics, forest man- agement, market studies, or other topics. RESOURCE MANAGEMENT Caballero (1994) studied the use and manage- ment of Sabal Adans. (Arecaceae) among the Yuc- Volume 83, Number 1 6 Balick 61 Transforming Ethnobotany atec Maya of Mexico, an example of a contemporary study of the relationship between plants and peo- ple. Sabal is a multi-use palm that has served as a source of construction material, firewood, food, medicine, and magic. Caballero discovered that some traditional uses of the trees had disappeared (magic, medicine), others were declining (broom making. poles, and fences), while others persisted (thatch, fuel) or increased (handicrafts). His studies showed the greatest pressure on the resource was where some 3500-5000 Sabal leaves. requiring 250-1250 trees (ranging from juvenile to for thatch, adult). are needed to thatch a house in the region. The importance of indigenous resource manage- ment as a way of maximizing future supplies of palm leaves is evident from this study, as are the dangers posed by overexploitation of the resource. Anderson (1988) studied Euterpe oleracea Mart.. another palm commonly used in Amazonian Brazil, and found that a single family received U.S. $15,532.86 from sales of agai fruit (63.1% of its annual cash income), and U.S, $2,916.79, or 11.8% of its income from sales of palm heart. Because palms are so important to tropical forest dwelling people, this plant family is a primary candidate for additional resource management studies. PHARMACEUTICAL PROSPECTING Cox (1994) estimated that at least 50 pharma- ceutical drugs have been discovered from ethno- botanical leads. These include digoxin used to treat atrial fibrillation, isolated from Digitalis purpurea L.. a plant employed in the eighteenth century to treat dropsy, an accumulation of fluid resulting from heart failure. A more recent discovery was of vin- cristine and vinblastine, both used to treat blood cancers, isolated from Catharanthus roseus (L.) G. Don, which has been used in the Caribbean to treat "sweet blood" (diabetes). Cox's ethnobotanical research in Samoa has led to the discovery of a potent anti-viral compound. prostratin, derived from Homalanthus nutans Guill. Cox identified this plant in 1984, after learning about it from several healers who used it to treat vellow fever. When tested at the National Cancer Institute, an extract of this plant exhibited powerful HIV ( National Cancer Institute is now plan- ~ in vitro activity sustafson et al.. 1992). The ning to license prostratin to a drug company for o P against additional study. Cox also identified a topical anti- inflammatory from Erythrina variegata L., utilized by healers in Samoa to treat skin inflammation. The active component, a flavanone, is now being devel- oped by Shering-Plough Corporation. Cox has also worked to ensure that a significant portion of roy- alties earned from the sales of these compounds will be returned to the Samoan people. His foun- dation (Seacology Foundation) has helped raise funds to protect 64,000 acres of tropical forest by creating four village-owned and -managed reserves. Although most ethno-directed pharmacological prospecting in the past has not returned benefits to native people or helped to conserve the forest eco- systems where the source plants are found, it is clear that ethnobotanical studies such as those of Cox and his students have done much to create a A for-profit entity, Shaman Pharmaceuticals, Inc., was founded new model for this type of research. with a corporate philosophy that creates mecha- nisms to ensure both short- and long-term benefits to traditional people, and promotes conservation of forest ecosystems in the areas the company gathers plants through The Healing Forest Conservancy, a foundation it established and supports. THE BELIZE ETHNOBOTANY PROJECT The Belize Ethnobotany Project was initiated in 1988 as a collaborative endeavor between the Ix Chel Tropical Research Foundation, a Belizean non-governmental organization, and the Institute of Botany of The New York Botanical Gar- den. The principal purpose of the project has been Economic to inventory the ethnobotanical diversity of Belize, a country with significant tracts of intact forest as well as nine different cultural and ethnic groups. Dozens of expeditions since 1988 have resulted in some 3600 collections of plant specimens, over 50% with ethnobotanical descriptions. Duplicate specimens have been deposited in the herbaria of the Belize College of Agriculture, the Belize For- estry Department, The New York Botanical Garden, and the US National Herbarium. A database main- tained at The New York Botanical Garden will be distributed to several computer facilities within Be- lize. The Project has gathered traditional knowl- edge graciously provided by over two dozen tradi- Kekchi, Maya, East Indian. and Men- — tional healers of Mopan, Yucatec, adino, Garifuna, Creole, nonite descent. The ethnobotanical inventory has been combined with the production of an annotated checklist of the flowering plants and ferns of Belize, based on col- lections of the ethnobotany project, previously available herbarium specimens, and relevant liter- ature. With the input of dozens of taxonomic spe- cialists, the checklist will help determine the com- prehensiveness of the ethnobotanical survey. Annals of the 62 Missouri Botanical Garden Number of Uses vs. Number of Collections AAA e3 AAAAAAA AAA AAAAAAAA 20 A4AAAAÀ - A ro.o.o. i A oe | p] ++ A Momordica charantia | 2 15 A ooo | — | | o 99 o i +.. | + Aloe vera | £ A * | g 10 A ds | e Agonandra sp. | 3 a ag so | A ¢ eoccccce 5 i / PS add ii A + A oe / koe E 7 0 A 9 - + 9 - --7----7 C A R - tn Oo M9 © © o V o rt - O yo ol - N QN 0 0 0 FT Number of Collections Figure 1. Graph de gie the importance of e interviews in investigating local uses of economically important plants. The r of Collect the different ways in "ic i each species is use tions rare s the number of people interviewed and the Number of Uses, . Wi ith each of the species represented in this graph, new uses are still being described after interviews with as many as 40 people THE CONCEPT OF THE MULTIPLE USE CURVE Rigorous ethnobotanical study demands accurate determination of appropriate sample size, the num- ber of collections, and number of people inter- viewed, in order to ensure that the information on a specific plant is relatively complete. Many pre- vious ethnobotanical studies have depended on one or a small number of collections as the basis for their information and conclusions. A mathematical relationship can be developed, based on the con- cept of the species area curve (Campbell et al. 1986), to assess the appropriate number of collec- tions or interviews. Figure 1 graphs the relationship between the number of different uses of three species versus the number of people interviewed, based on 42 inter- views in August and September 1994. Interviewees were, for the most part, not traditional healers, but elderly people willing to discuss the uses of ten plant species. Three species have been selected to illustrate the type of knowledge obtained. The upper curve in Figure 1 illustrates the uses of Momordica charantia L., a vining herb common to much of the Caribbean and elsewhere. The initial 13 interviews record 20 uses, but 5 additional uses were documented by the 39th interview. This curve illustrates a pattern for a “powerful plant" (Balick, 1990), one that is widely known with multiple uses. The middle curve is for Aloe vera (L.) F., a commonly known plant with a more focused series Burm. of uses, especially in the area of dermatology. The final species, Agonandra Miers ex Benth. sp., is an example of a plant that is not commonly known throughout the community and appears to have few- er uses, focused on “male” problems. It also has a more limited distribution in Belize than the other wo taxa, primarily in forests. One point illustrated s the multiple use curve for these three plants is the large number of interviews/collections neces- sary to obtain the maximum amount of data. Many ethnobotanical studies, including contemporary ones that utilize statistical methods, are based on a small number of interviews/collections per spe- cies. For some plants, such as those with specialist uses, a few collections from traditional healers may be sufficient, while more widely used plants may require several dozen interviews/collections before the total extant information can be obtained. B. VALUATION STUDIES One method of ascertaining the value of non- timber forest products in the tropical forest is to inventory a clearly defined area and estimate the economic value of the species found there. Peters et al. (1989) were the first to document the com- mercial value of non-timber forest products found within a hectare of forest in the Peruvian Amazon, but they did not include medicinal plants in their inventory. This aspect was evaluated later in Belize (Balick & Mendelsohn, 1992). From two separate Volume 83, Number 1 Balick Transforming Ethnobotany Table 2. Medicinal plants harvested from a 30-year-old valley forest plot (no. 1) in Cayo, Belize. Common name Scientific name Use? Bejuco verde nandra racemosa (DC.) Calawalla Smilax lanceolata L. China root Cocomecca Dioscorea sp. Ago ) Standl. Phlebodium decumanum (Willd.) J. Smith Sedative, laxative, analgesic “gastritis,” “gastritis,” cers, pain, ' chronic indi- di a high blood pressure, “can- Blood tonic, fatigue, "anemia," acid stomach, rheumatism, skin condi- tions Urinary tract ailments, bladder infec- tion, stoppage of urine, kidney slug- gishness and malfunction, to loosen mucus in coughs and colds, febri- fuge, blood tonic Flu, colds, constipation, fevers, stam- Contribo Aristolochia trilobata L. ach ache, indigestion, “gastritis,” parasites “Uses listed are based on disease concepts recognized in Belize, primarily of Mayan origin, that may or may not have alen states in Western medicine. For example, kidney slu gishness i is not a condition educ. d rec ognized by Western-trained physicians, but is a common complaint among people in this region. plots of e and 50-year-old forest a total of 308.6 and 1433.6 kg per hectare (dry weight) of medi- cines, dh i whose value could be judged by local market forces, was collected. Local herbal pharmacists and healers purchase and process me- dicinal plants from herb gatherers and small farm- ers for an average price of U.S. $2.80 per kg, sug- gesting that harvesting the medicinal plants from a hectare would yield the collector between $864 and $4014 of gross revenue. Subtracting the costs re- quired to harvest, process, and ship the plants, the net revenue from clearing a hectare was calculated be $564 and $3054 on each of the two plots (Balick € Mendelsohn, 1992). The lists of plants and their uses are presented in Tables 2 and 3. Not enough information is available to under- stand the life cycle and regeneration time needed for each species, thus we cannot comment on the frequency and extent of collection involved in sus- Table 3. tainable harvest. However, assuming the current age of the forest in each plot as a rotation length, we calculated an estimate of the present value of harvesting plants sustainably into the future using the standard Faustman formula: V—HR/(l—e "). where R is the net revenue from a single harvest and r is the real interest rate; t is the length of the rotation in years. Given a 30-year rotation in plot l. the present value of medicine is $726 per hect- are. A similar calculation for plot 2, with a 50-year rotation, yielded a present value of $3327 per hect- are. These calculations assume a 5% interest rate. These estimates for the harvest of medicinal plants compare favorably with alternative land uses in the region, such as milpa (corn, bean, and squash cultivation) in Guatemalan rainforest, which yielded $288 per hectare. Other commercial prod- ucts, such as allspice, copal, chicle, and construc- tion materials, in the plots could be harvested and Medicinal plants harvested from a 50-year-old ridge, forest plot (no. 2) in Cayo, Belize. Common name Scientific name Use Negrito Simaruba glauca DC. Gumbolimbo Bursera simaruba (L.) Sarg. China root Smilax lanceolata L. Cocomecca Dioscorea sp. Dysentery & diarrhea, dysmenorrhea, skin condi- tions, stomach and bowel tonic Antipruritic, stomach cramps, kidney infections, di- uretic Blood tonic. matism. skin conditions Urinary tract ailments, bladder infection, stoppage of urine, kidney sluggishness and malfunction, to fatigue, “anemia,” acid stomach, rheu- loosen mucus in coughs and colds, febrifuge, blood tonic 64 Annals of the Missouri Botanical Garden added to the total value of the medicinal plants. Thus, this study suggested that protection of some areas of rainforest as extractive reserves for medic- inal plants appears to be economically justified. It seems that a periodic harvest strategy is a realistic and sustainable method of utilizing the forest. Based on our evaluation of forest similar to the sec- ond, 50-hectare plot analyzed, it would appear that one could harvest and clear one hectare per year indefinitely, assuming that all of the species found in each plot would regenerate at similar rates. More than likely, however, some species, such as Bursera simaruba (L.) Sarg., would become more dominant in the forest ecosystem, while others, such as Dios- corea L., could become rare. This analysis is based on current market data, and estimates of the worth of the forest c X change with local market forces. For example, i knowledge about tropical herbal medicines de comes more widespread and their collection in- creases, prices for source plants would fall. Simi- larly, if more consumers became aware of the potential of some of these medicines, or if the cost of commercially produced pharmaceuticals became too great, demand for herbal medicines could in- crease, substantially driving up prices. Finally, de- struction of the tropical forest habitats of many of these important plants could increase their scarcity, driving up local prices. This scenario has already been observed in Belize with some species, indi- cating that the value of tropical forest for the har- vest of non-timber forest products will increase rel- ative to other land uses over time, as these forests become more scarce. C. DEVELOPMENT OF A FOREST-BASED TRADITIONAL MEDICINE INDUSTRY One of the primary dilemmas in the development of an extraction program for non-timber forest prod- ucts has been the long history of overcollecting and resultant decline of resources, and export and pro- cessing of raw materials at centers and countries far from their origin. Rattan is a classic example of pie in producing countries, who are closest to > e resource, receiving the smallest percentage of the profits derived from its manufacture into high- quality furniture (Dransfield & Manokaran, 1993). In Central America, as elsewhere in the developing world, locally developed brands of plant-derived medicine are now being marketed with a value- added component (production and packaging) re- maining in the country of origin of the raw material. As more of these products appear, based on the success of the original endeavors, greater demand for ingredients from rainforest species will result. This could contribute to preservation of tropical for- est ecosystems, but only if people carefully manage the production or extraction of the plant species that are primary ingredients in these products. In addition to using methods of sustainable extraction from natural ecosystems, small farmers will culti- vate some species for sale to both local herbalists and commercial companies. To address the latter possibility, our work in the Belize Ethnobotany Pro- ject has included a program with the Belize College of Agriculture (BCA), Central Farms, to learn how to propagate and grow over 24 different plants cur- rently utilized in traditional medicine. O’Brien, Professor of Horticulture at ordinated this effort, which has included the genera Achras L., Aristolochia L., Brosimum Sw., Bursera Jacq. ex L., Cedrela P. Browne, Croton L., Jatropha L., Myroxylon L.f., Neurolaena R. Br., Piscidia L., Psidium L., Senna Miller, Simarouba Aublet, Smi- lax L., Stachytarpheta Vahl., and Swietenia Jacq. ), ESTABLISHMENT OF AN ETHNO-BIOMEDICAL FOREST RESERVE The concept of the extractive reserve as a tool for conservation has received a great deal of atten- tion over the last few years. Many of these reserves are tracts of forest where non-timber products can be harvested by local individuals or groups who theoretically have a stake in the preservation of the forest's biological integrity (Allegretti, 1990). Prod- ucts such as rubber, Brazil nuts, copal resin, plant oils, fruits, fiber, construction materials, foliage and house plants for the florist trade, and other items have been selected for harvest and marketing from extractive reserves in the Amazon, Central Ameri- ca, Asia, and Africa. Numerous perspectives on these resources, both positive and negative, have ud highlighted recently (Browder, 1992; Ryan, ÉS Tre 1993, a 6000-acre, lowland tropical for- est, government-owned reserve was established for the extraction of medicinal plants, teaching, and apprenticeship, with financial support from The Healing Forest Conservancy anc e Rex Foun- dation. This particular forest, in the Yalbak region of Belize, contains a broad diversity of medicinal plant species. Also within its borders are many spe- cies of animals, including jaguar, tapir, peccary, howler monkeys, and numerous other mammals, birds, and reptiles. A unique feature of this reserve is that it has been designated for the extraction of medicinal plants used locally as part of the primary health Volume 83, Number 1 Balick 65 Transforming Ethnobotany 1996 Table 4. Credibility rating for use information collected. Category Rating Hypothetical example Collector uses or directly observed use | Dr. Smith saw these Orbignya cohune leaves being used as thatch in Belize. Informant uses or directly observed use 2 Maya healer, Don Elijio, told Dr. Smith he uses these Piper amalago roots for snakebite Informant heard/knew from a further source 3 Ethnographer on the Sioux reservation heard that the Sioux used the Aster for menstruation problems. Use reported from the literature 4 As for the IEB teaching collection, e uses will be gathered from the literature and summarized on the use label. Common knowledge 5 As, for example, a collection of a cultivar of coffee from a coffee plantation with a reported use as a stimulant beverage. Credibility of use information unknown 6 ew field botanist neglected to write down any infor- mation about his informant. care network. Accordingly, this type of extractive reserve was Classified as an “ethno-biomedical for- est reserve” (Balick et al., 1994), a term intended to convey a sense of the interaction between people, plants. and animals, and the health care system in the region. One objective of the reserve is ethnobotanical and ecological research, designed to identify the plant resources it contains and develop appropriate technologies for their sustainable extraction. David Campbell and his students from Grinnell College (lowa. U.S.A.) have constructed ecological transects in selected parts of the reserve to serve as long- term study sites. Some of these transects will mon- itor extraction, while others monitor changes in the native vegetation. Ethnobotanical inventories have begun to catalog economically important plants in the reserve, as well as in the surrounding Cayo Dis- trict. Other scientists will be invited to participate =~ in these studies t will be many years before this first ethno-bio- medical forest reserve can be judged a success or failure. A great deal of work must go into devel- oping the management plan and finding the finan- cial and human resources to implement it. Land use pressures surrounding the reserve, specifically log- ging and agriculture, as well as sociological and political factors, could endanger the long-term ex- istence of the reserve. However, in Belize there is a great deal of optimism and support, much of it at the grass roots level. Reserves for protecting me- dicinal plants recently have been established across India, in a nationwide effort to ensure the supply of these important species. CONCLUSION The discipline of ethnobotany is currently evolv- ing both in its philosophical underpinnings and methodology. Contemporary studies are incorporat- ing the most powerful techniques of molecular pharmacology and computerization in the analysis of ethnobotanical data. One recent development be- ing implemented at The New York Botanical Gar- den's (NYBG) Institute of Economic Botany (IEB) is the establishment of a “credibility rating” for in- formation that is collected on plant utilization. In the past there was little opportunity to evaluate the quality of data based on the way it was collected during ethnobotanical studies. When extracting in- formation from the ethnobotanical literature, it is rarely clear whether the investigator actually ob- served participated in the uses discussed, whether the data were collected during a casual walk through the forest with a younger person who remembers specific uses of the plants by one of his To address this, the credibility rating presented in Table 4 will be incorporated into the database at NYBG. Data with a rating of 1 have a reasonable certainty of being accurate, while those i rating of 5 or 6 may be less authoritative. forebears. r with a Although this rating is an experiment, and will cer- tainly be revised over time, it is an attempt to begin o standardize the quality of data collected and ~ evaluate its relative credibility. There is a deep. often spiritual relationship be- tween plants and people, in both traditional settings and among more acculturated societies. These re- lationships can be elucidated through ethnobotan- ical studies and used to increase biological literacy among the non-scientific community, as evidenced by work in Mexico, Samoa, and Belize. If the most crucial issues of biodiversity conservation are to be addressed successfully, we need to improve the ed- ucational system's ability to communicate the im- portance of science, not only to people in the Unit- 66 Annals of the Missouri Botanical Garden ed States but around the world. It is essential to build on the value of systematic knowledge and link it with the world in a way that inspires the largest possible constituency to appreciate the im- portance of biodiversity and the need for its con- servation. Ethnobotany has been an important tool for meeting this goal, and will continue to be in the future. In the context of Systematics Agenda 2000, it is worth recalling the words of Ralph Waldo Em- erson who wrote, “Nature tells every secret once.” It is imperative that we heed these words and be- come better prepared to understand those secrets. Literature Cited ie ipe . H. 1990, Extractive reserves: An alterna- e reconciling development and nad conservation in Amazonia. Pp. 252 2-264 in A. B. An- derson (editor), Alternatives to Deforestation: Steps To- wards Sustain: € Vie of the Antas Rain Forest. Co- ork. lumbia — Press, Ne Anderson, A. ona. Use and management of native for- ests dei by Acai palm (Euterpe oleracea cyte in the Amazon estuary. Advances Econ. Bot. 6: 154. a 1895. Some new ideas. Philadelphia Eve- g Telegraph. December Re M. L. Corn, M. Lee * D. M. Ackerman. 1993. Biotechnology. Indigenous Peoples, and Intellectual Property Rights. CRS Report se] Congress (93-4 Congressional Researc h Service, ' gress, Washington, Balick, M. J. 1990. Eihnoboluny and the identification of Hr uw €— from the rainforest. Pp. 47—55 in adwic icd (editors), Bioac ‘tive Compounds i sage John Wiley & Sons, Chichester. ——— € R. Mendelsohn. 1992. Assessing the economic value of traditional ~~ ‘ines from tropical rain forests. Conservation Biol. 6: 128-130. > rvigo & I. Romero. 1994, The ee aah of an ethnobiomedical forest reserve in Beliz role in the preservation of biological and cultural _ ma Biol, 8: 317 Berlin, B., D. lo & . Raven. 1974. Prin- n of Tzeltal Plant C TA ‘ation, Academic Press, 78A), . The Library of Con- — 316— la p. M. 1987 m Dibnobotany of the Chácabo Indians, eni, 1-68. Bolivia. adus s Econ. Bot. 4: — ———. 1990. Giving native people a share of the profits. Garden 14: 2841. Browder, J. O. 1992. Social and economic pe ia on the pala nt of market-oriented extrac reserves nazon e forests. Advances Econ. Bol, € I: 33-41. Caballero. J. hs Use and Management id Sabal Palms Among M ya of Yucatan. Ph.D. Dissertation, Uni- ia of € alifornia at Berkeley, eee of Anthro- -— =" = log al D. G., D. C. Daly, C. T. Prance € U. N. Maciel. 1986. Quantitative ecological inventory on terra firme and várzea tropical dos on the Rio Xingu, Brazilian Amazon. Brittonia 38; 369-393. Cox, P. A. The ono al approac $ to ae discovery: Strengths and limitations. Pp. 25—4 D. J. enis & J. Marsh parent Ethnobotany bie the rch for New Drugs. John Wiley & Sons, West Sussex. Cuming, A. B. 1993. Ethics, Ethnobiological Re- and Biodiversity. oe World Wide Fund for re, poA Switzerlar N andar (Editors). 1993. Plant Re- sia n 6. Rattans. Pudoc Sci kes | a A to ethnobotany. Amer. a E He Ford, R. I. 1978. la Historical diversity dt ae a 33—49 in R. I. Ford (editor), The Nat and Bau of Ethnobotany. Museum of ir University p ichigan, No. 67, Ann Arbor. Gustafson, . J. H. Cardelline, II, : B. Mc pae R. . Gulakows E J. Ishitoya, Z. Szallasi, N. E. Le M. ae O. S. Weislow, J. A. TH R. W. as k- heit, Jr., . Cragg, P. A. Cox, J. P. Bader & M. Boyd. 1992. A non-promoting phorbol from the Sa- noan medicinal plant Homalanthus nutans inhibits cell meo by HIV-1. J. Med. Chem. 35: 1978-1986. Iltis, H. 1974. Freezing the genetic landscape—The reservation of diversity in cultivated plants as an ur- gent social responsibility of the plant genetic A on plant taxonomist. Maize Genet. Coop. Newslet 19¢ 9-200. Linares, E. & R. E. Bye, Mi 1987. A study of four me- dl ‘plant comple of Mexico adjacent United States. J. Bian ceased: 19: 153-18: Palmer, E. 1871. Food products of the North American Indians. U.S. Commissioner of Agriculture Report 1870: 404-428. 878. Plants used by bes pps of zr United States, Amer. Naturalist 12: », 646—065 Peters, C. M., A. H. Gent Qi s ^i iae d 1989. Valuation of an fens rainforest. Nature 339: 655— 656. Phillips, O. & A. H. Gentry. 1993a. The useful Es 9 Tambopata, Peru: I. Statistical hypotheses tes new quantitative technique. Econ. Bot. 47: 15- "^. —— —— & —. |993b. The useful plants of Tambo- pata, Peru: IL Additional hypotheses testing in quan- titative ethnobotany. Eco 3343. Powers, S. 1873-1875. Aboricinal botany. Calif. Acad. 379. Sci. Proc. 5: 373-379 Prance, C. T.. W. Balée, B. M. ig & R. e Carneiro. 1987. Quantitative ethnobo tany and the case i for con- ión Biol. 96-310. servation in Amazonia. n, J. C. 19 Goods from the woods. ei Trees ). Hoffman. 1973. The Botany and Chemistry of Hallucinogens. Charles C. Thomas Pub- lishers, Pp ya Soleri, D. & S. nith. 1995. Morphological and phe- nological sate of two Hopi maize varieties con- served in situ and ex situ. Econ. Bot. 49: 56-77. THE PROSOECA PERINGUEYI John C. Manning? and Peter Goldblatt” (DIPTERA: NEMESTRINIDAE) POLLINATION GUILD IN SOUTHERN AFRICA: LONG-TONGUED FLIES AND THEIR TUBULAR FLOWERS! ABSTRACT A guild of 28 winter- and spring-flowering species of two plant families, Iridaceae and Geraniaceae, with intense pe to crimson flowers and extremely long and slender dees tubes, is pollinate d exclusively by two long-tongued I flies of the family Nemestrinidae. The two species of flies, Prosoeca peringueyi and P. sp. nov., are active in the late winter and spring, have large bodies, caste s 20—50 mm on and forage for necta while hovering. Plants pollinated hy these two flies share a suite of convergent floral characteristics, including a pee. or slightly curved floral tube at least 20 mm and up to 70 mm long, relatively short petals or tepals colored pé: cni dark blue- or red-purple with pale nectar guides, and anthers and stigmas exserted from the tube and. usually unilateral in orientation. With me exception, the flowers of all species secrete large amounts of nectar of relatively e E total sugar concentration, y 24-29%, and high sucrose: hexose ratio. Most members of the guild e odorless flowers. The long floral tube makes nectar unavailable to most insects, including a variety of bees, wasps, and other flies that pollinate plants which co-occur with members of the long-tubed flower guild. The two Prosoeca species have mouthparts long enough to forage effectively on these long-tubed flowers. ^ they are also effective pollinators because pollen adheres to their bodies and is transported from flower to flower. e flies visit a wide range of plants but are effective ra only of those with tube lengths greater that their uude is lengths. We have identified four mutually exclus s of pollen deposition on the insects’ bodies: when two or more members of the guild co-occur, each species em lh utilizes 1 a different pollen deposition site. This suggests that pollen contamination is detrimental to reproductive success. Differ ential pollen deposition sites may have evolved in response to selection for reduced pollen contamination. Since 27 of the 28 plant species appear to depend exclusively on these two species of Prosoeca for pollination, these flies must be considered keystone species in the ecosystems where they occur A close association between the form and color between the butterfly, Aeropetes (Meneris) tulbaghia, of flowers and pollination by a particular pollinator and late summer-flowering plant species with large bright red blossoms (Johnson & Bond, 1994). Pollination by long-tongued flies is a relatively unusual phenomenon, first documented in southern is well known. Convergence in floral morphology among species that rely on the same pollinator class led to the recognition of floral syndromes (Faegri & van der Pijl, 1979; Grant, 1981; Vogel, 1954). Africa by Rudolf Marloth (1908) and later in some- with morphologically similar flowers what more detail by Stefan Vogel (1954). Although pollination by long-tongued flies has also been re- ported in India (Fletcher & Son, 1931) and Cali- fornia (Grant & Grant, 1965), it appears to be par- ticularly well developed only in southern Africa. In the western part of southern Africa 28 species Those species that share the same pollinator species constitute a particular pollination guild, an extension of the term (Root, 1967) describing a group of species that exploits the same class of resources in a similar way. A guild is thus a functional unit independent Although a number of of Iridaceae and Geraniaceae have intensely col- of taxonomic considerations. / ored purple to crimson flowers with extremely long pollination syndromes have been identified in the southern African flora (Vogel, 1954), very few floral tubes. These species all occur in a restricted geographic area, flower between July and Septem- guilds have been described. The most striking of ber. and often occupy similar habitats. The conver- those that have been documented is the association i This research was supported by National Geographic Society Grant 4816-92. We gratefully ac po G the work of B.-E. van Wyk, Rand Afrikaans University. Johannesburg. who provided the analyses of sugar nectars. We also thank Peter atk, Se Paterson-Jones, n Kim Steiner for helpful comments during the p of Lo paper. Tu Herbarium, National Botanical Institute, Private Bag X7. Claremont 7735, South Africa. *B. A. Krukoff Curator of African Botany, Missouri Botanical Garden. P.O. Box 299, St. Louis. Missouri 63166— 0229, U.S.A. ANN. Missourt Bor. GARD. 83: 07-86. 1996. 68 Annals of the Missouri Botanical Garden gent floral morphology in this group of spring-flow- ering geophytes and small shrubs constitutes a distinct floral syndrome, and their coincident ge- ography and phenology suggests that they are mem- bers of a specific pollination guild. Some of these species belong to the genus Lapeirousia (Iridaceae) and have already been found to be pollinated by one or both of two species of long-tongued flies in the genus Prosoeca (Diptera:Nemestrinidae) (Gold- blatt et al., T i is to extend our observations to determine whether 5). The purpose of this investigation the convergence in floral morphology to the L. sil- enoides-type flower in the other species coincides with pollination by the same fly species. Our results support the recognition of a distinct pollination guild. We discuss some of the implications of such a specialized pollination system on plant ecology and evolution and consider its possible origin. METHODS PLANT SPECIES Members of the guild were initially identified during the course of field research in 1992 and 1993 in connection with a study of pollination ecol- ogy of Lapeirousia subg. Lapeirousia (Goldblatt et al., 1995). In this study species with purple crimson flowers, white to cream nectar guides, and a perianth tube in excess of 30 mm were pollinated by Prosoeca peringueyi or P. sp. nov., or occasion- ally both. We then reviewed the literature for re- cords of species with purple to crimson flowers re- corded from the west coast of southern Africa. All species having a perianth tube at least 30 mm long, or the anthers and stigmas held at least 30 mm from the base of the floral tube, were listed for further study. These species were examined in the field whenever possible to obtain observations on nectar characteristics and pollinators (Table 1). The ap- parent floral tube length was determined as the dis- tance from base to the mouth of the tube. The ac- tual floral tube length is less in some species due to the occlusion of the lower part of the tube and was determined empirically as the level down to which nectar could be freely extracted using a mi- cropipette. Functional tube length was determined as actual tube length plus the distance between the mouth of the tube and the mid point of the anthers. Measurements were made on a minimum of 10 in- dividuals per population. Complete distribution ranges of plant species were taken from the literature and supplemented by recent herbarium records. Voucher specimens were made for all populations studied. Plant vouch- ers are deposited at the Missouri Botanical Garden Herbarium, St. Louis (MO), and the Compton Her- barium, Cape Town (NBG — INSECT SPECIES Observations of insect foraging (Figs. 1-6) in- volved 4-20 hours per species and included such aspects as the density and diversity of floral for- agers and how they removed rewards from flowers. Insects observed probing the floral tube or brushing the anthers or stigmas were captured and killed in a jar using ethyl acetate fumes. Location of pollen deposits was based first on visual observation of foraging insects and later on examination of pinned specimens. Individual insects were washed of pol- len after pinning by placing the insect on a glass slide and rinsing the whole body in 100% ethanol while gently dislodging pollen loads on the frons, thorax, and sternum with a dissecting needle. The dry pollen residue was stained and mounted in 1— 1974). To prevent contamination of the body of an insect with 2 drops of Calberla's fluid (Ogden et al., pollen carried by another in the same jar, the bod- ies of insect specimens were isolated from each oth- er by wrapping them in tissue. Insect distributions Natal M seum, Pietermaritzburg, and the South African Mu. were determined from collections at the seum, Cape Town, plus our own observations and collections. Insect vouchers are deposited at the atal Museum, Pietermaritzburg. NECTAR ANALYSES Nectar volume measurements (Table 2) were made from unbagged flowers in the field and rep- resent the standing crop that will be influenced by visitation rates. Whole flowers were picked and nectar was withdrawn from the base of the floral tube with 3 wl capillary tubes after separating the ovary from the perianth (Iridaceae) or base of the hypanthium tube from the pedicel (Geraniaceae). Nectar was extracted from five or more individuals per population in most cases (Table 2). Nectar sam- ples were dried on Whatmans filter paper no. 1 and sent to B.-E. van Wyk, Rand Afrikaans University. Johannesburg. for analysis (Table 2). The percent- age of sucrose equivalents in fresh nectar was mea- sured in the field on a Bellingham « Stanley hand- (0-50%) from individuals per population. held refractometer five or more RESULTS PLANT CHARACTERISTICS A total of 28 plant species occurring along the west coast and near interior of southern Africa, and Volume 83, Number 1 Manning & Goldblatt 69 1996 Prosoeca peringueyi Pollination Guild Table 1. on the basis of geography and pollinators of co-occurring plant species (P. peringueyi and Study sites and dates of observations on pollinators are listed for those species for which pollinator observations are Species belonging to the Prosoeca peringueyi-P. sp. pollination guild. Pollinators in parentheses are inferred g [ gue | 8 I ? sp. are not sympatric). a vailable Family/species Pollinator Study site Geraniaceae a framesii L. Pelargonium cortusifolium U Herit. crassicaule V2 echinatum Gan de flowered form) incrassatum (Andr.) Sims magenteum van der Walt sericifolium van der Walt curviscapa G. Lewis dregei Baker ecklor ftella bun ex Bolus geniculata G. Lewis pubescens G. Lew sambucina var. longibracteata G. Lewis unguiculata G. Lewis Geissorhiza kamiesmontana Goldblatt Hesperantha icd (Klatt) de Vos oligantha Diels old Goldblatt Lapetrous penal subsp. dolomitica Dinter lewisiana (B. Nord.) Goldblatt jacquinii N. E. | oreogena Schltr. pyramidalis subsp. regalis Goldblatt & J. Manning = silenoides (Jacq.) violacea Goldblatt (P. peringueyi) — (P. peringueyt ~ (P. peringueyi P. peringueyi P. peringueyi 2 2 , > y P. peringueyi P. peringueyi P. peringueyi (P. peringueyi) P. sp. nov. P. peringueyi P. sp. nor. — P. peringueyi) P. peringueyi (P. sp.) (P. peringue yt) (P. peringueyi) P. peringueyi P. (P. sp.) (P. peringueyi) P. peringueyi (P. peringueyi) P. sp. nor. P. peringueyi P. sp. nov. P. peringueyi P. peringuevi P. peringueyi Steinkopf-Anenous Pass (24 Aug. 1992); Kamieskroon (24 Aug. 1993): Garies Hill (23 Aug. 1994): Spek- takel Mts. (5 Sep. 1994); Kamiesberg, Leliefontein (13 Sep. 1993 and 20 Sep. 1994) Bidouw Valley (12 Aug. 1994) Steinkopf-Anenous Pass (24 Aug. 1992) Spektakel-Naries (4 Sep. 1994) Kamiesberg. Sneeukop (12 Sep. 1993): Kamiesberg, Leliefontein (13 Sep. 1993 and 20 Sep. 1994) Hantamsberg Plateau (3 Sep. 1994) Oorlogskloof Nature Reserve (8 Sep. 1993) Nieuwoudtville Nature Reserve (8 Sep. 1992 and 4 Sep. 1994 Garies Hill (23 Aug. 1994 — Kamiesberg, Sneeukop (12 Sep. 1993) Anenous Hills (24 Aug. 1992) Oorlogskloof Farm (8 Sep. 1992); Oorlogskloof Nature Reserve (8 Sep. 1992) Botterkloof Pass (23 Aug. 1993); Farm Alpha (1 Sep. 1994) Glenlyon (23 Aug. 1993 and 4 Sep. 1994): Nieuwoudt- ville Nature Reserve (4 Sep. 1994) Trawal (5 and 11 Aug. 1994) Spektakel Pass (23 Aug. 1992): Kamieskroon to Gro- otvlei (22 Aug. 1992): Kamieskroon (2 Aug. 1993) 10 km S Kamieskroon (24 Aug. 1993): (23 Aug. 1994) Botterkloof Pass (23 Aug. 1993) Garies Hill Annals of the Missouri Botanical Garden Table 1. Continued. Family/species Pollinator Study site Romulea hantamensis (Diels) TN P. sp. nov. Goldblatt Sparaxis variegata subsp. metelerkampiae P. peringueyi (L. Bolus) Goldblatt Tritonia marlothii de Vos (P. peringueyi) Xenoscapa uliginosa Goldblatt & (P. peringueyi) anning Hantamsberg (3 Sep. 1994) Pakhuis Mts., Farm Alpha (1 Sep. 1994) mostly endemic there, were identified as converging on Lapeirousia silenoides in floral morphology (Ta- le 3). These include 22 species of Iridaceae in the genera Babiana, Geissorhiza, Hesperantha, Lapei- rousia, Romulea, Sparaxis, Tritonia, and Xenoscapa, and six species of Geraniaceae, all in the genus Pelargonium. Within this group are seasonal geo- phytes (Iridaceae and Pelargonium incrassatum) and small to moderate-sized shrubs (Pelargonium cortusifolium, P. crassicaule, P. echinatum, P. ma- 1 1cm — — Figures 1, 2. The flower.—1. P. peringueyi.—2 p. nov. genteum, and P. sericifolium). While the habits and growth forms of the species vary, their flowers share several unusual properties and may be considered to constitute a distinct floral syndrome (Figs. 7—9) The floral tube is straight or slightly curved to sig- moid, very narrow (1.5-2.5 mm diam.) and (18—-) 30—70 mm long, and the petals or tepals are shorter than the tube. The flowers are typically zygomorphic, with stamens and styles unilateral, but are actino- morphic in five species in which the stamens are sym- 2 1cm Prosoeca species responsible for pollinating plant species with the Lapeirousia silenoides-type . Psp Volume 83, Number 1 Manning & Goldblatt 71 Prosoeca peringueyi Pollination Guild Figures 3-6. Prosoeca species foraging on Lapeirousia silenoides-type flowers. In these photographs the flies are inserting their mouthparts into the floral tubes and have not yet probed deep enough to reach the nectar in the lower part of the tube or brush the anthers and stigm: is of the flowers.—3. P. > peringueyi visiting L. pyramidalis subsp. regalis.—A. P. peringueyi visiting Babiana dregei.—5. P. peringueyi and L. silenoides.—0. F. sp. nor. and L. oreogena. 72 Annals of the Missouri Botanical Garden Table 2. Nectar characteristics of species with the Lapeirousia silenoides-type flower. Fru = fructose, Glu = glucose, Suc = sucrose. Sample size for nectar volume figures are indicated i in parentheses in volume column; sample size for nectar sugar components is in the last column after sucrose:fructose + glucose ratio. Volume Mean % Mean Suc/ Family/species pl (n) sug: Fru Glu Suc Glu + Fru (n) Geraniaceae Pelargonium cortusifolium 2.6-3.1 (2) 24 — — — — incrassatum 1.5-2.2 (1) 38 0 75 25 0.33 (1) magenteum 0.6-1.8 (5) 29 ; 0.28 (1) sericifolium no measurable nectar produced Iridaceae Babiana curviscapa 2.04.4 (5) 25 12 19 69 2.23 (1) lregei 3.9—9.6 (5) 22 13-15 19-21 8 1.94. (2) ecklonii 4.3—8.9 (5) 27.7 5-11 10-18 72-85 3.29 (3) flabellifolia 3.9-9.6 (5) 27 14 79 3.76 (1) framesii 2.6—6.4 (5) 26 7-9 12-14 77-81 3.84 (3) geniculata 3.2-4.8 (5) 29 17 21 62 1.63 (1) pubescens 3.24.8 (5) 28 338 9-14 78-88 2.18 (2) sambucin ar. longibracteata 3.96.6 (3) 30 6-12 10-19 69-84. 3.48 (3) Hesperantha utifolia 0.7—1.1 (10) 24 23-29 24—30 41—53 0.94 (3) oligantha 1.1-1.8 (5) 26.4 19-23 24-25 52-57 1.20 (2) Lapeirousia dolomitica subsp. dolomitica 1.4-5.5 (5) 29 4-9 12-14 77-84 4.13 (2) subap. lewisiana 5.1-5.5 (4) 27 5-12 14-25 63-81 2.94 (5) Jacquinii 15-23 (6) 26 8 17 75 3.00 (1) oreogena 2.5-7.3 (10) 26 13 21.5 65.5 1.90 (4) pyramidalis subsp. regalis 2.6—4.8 (10) 28 4-21 12-31 48-84 2.45 (6) silenoides 1.7-3.6 (16) 27 5-8 18-27 65-77 2.45 (3) violacea 1.4—1.8 (10) 27 9-13 15-16 71-76 2.77 (2) Romulea hantamensis 3.1-5,2 (3) 20 23 27 50 1.00 (1) Spara ca subsp. metelerkampiae 1.7-2.2 (6) 28.5 1-12 4-22 70-95 3.23 (4) Tritonia marlothii 1.8-3.5 (5) 29 13-14 17-18 68-70 2.23 (2) metrically arranged around a central style (Table 3). In all species except Pelargonium sericifolium (which does not produce nectar; Goldblatt et al., 1995), nec- tar accumulates at the base of the floral tube and fills its lower third. Nectar is thus accessible only to in- sects with tongues long enough to reach at least into the lower third of the tube The flowers are mostly intensely pigmented in col- ors ranging from dark blue-purple and violet to bright red-purple or cerise, but are pale mauve in taxa from the Richtersveld (northern Namaqualand) and southern Namibia. Contrasting markings in white or cream are almost always present, usually accompanied by additional darker areas of pigmen- tation (Figs. 7-9). The markings, which may take the form of streaks or spots near the tepal bases in spe- cies of Iridaceae, are confined to the lower tepals in species with zygomorphic flowers but are present on all the tepals in species with actinomorphic flowers. In Pelargonium (Fig. 7) the pale color signal is pro- vided by the white filaments, which are unilateral and declinate, rather than by tepal coloring. What- ever the shape and color of the flowers, the anthers and stigmatic surfaces are always held outside the mouth of the floral tube in a position where they will be brushed by the body of an insect probing the Volume 83, Number 1 Manning & Goldblatt 73 Prosoeca peringueyi Pollination Guild Table 3. Floral characteristics of species with the Lapeirousia silenoides-type flower. The floral tube is closed in the lower 10-12 mm in B. curviscapa, 20-30 mm in B. dregei. and 15-20 mm in B. framesii, hence floral tube length does not reflect the distance that an insect must extend its mouthparts to reach the nectar. Z = zygomorphic: A = actinomorphic. Tube length Family/species Symmetry Scent (mm) Flowering time Geraniaceae Pelargonium cortusifolium Z 0 ca. 30 Mar.—Nov. crassicaule Z 0 15-25 (Mar.—)Aug.—Sep.(—Oct.) echinatum Z 0 3—00 (July-)Aug.-Oct. incrassatum Z 0 30—40 Aug.—Sep magenteum Z 0 3347 (June—)July—Sep sericifolium Z 0 35-60 (July—)Aug.—Sep. Iridaceae Babiana curviscapa Z 0 30—48 Aug.—Sep. dreget Z 0 50-605 Aug.—Sep. ecklonii Z 0 40-50 Sep. flabellifolia Z 0 40-65 Aug.—Sep. framesii Z l 60-70 Aug.—Sep geniculata Z 0 35445 Aug pubescens Z 0 ca. 50 (July—)Aug. sambucina var. longibracteata Z l 30-50 Aug.—Sep. var. unguiculata Z | 38—55 Aug.—Sep. Geissorhiza kamiesmontana A 0 18-25 Sep. Hesperantha latifolia A 0 15-25 Aug.—Sep. oligantha A 0 30-36 Sep.(-Oct.) purpurea A 0 ca. 20 Sep. Lapetrousia dolomitica subsp. dolomitica Z l 2545 ne Sep. subap. lewisiana Z 0 45-55 July- jacquinii 7. 0 30-40 ete oreogena A 0 50—60 Aug.(—Sep.) pyramidalis subsp. regalis Z 0 40-50 July—Aug. silenoides Z 0 40—55 July-Sep. violacea Z 0 35—40 Aug.—Sep. Romulea hantamensis A 0 50-70 Aug.—Sep. Sparax variegata subsp. metelerkampiae Z 0 34—37 Aug.—Sep. Tritonia marlothii 7. 0 2544 Aug.—Sep. Xenoscapa uliginosa Z 0 25—30 Sep.(-Oct.) tube. Pollen is often inconspicuous and of the same We have no data on ultra-violet light reflectance in color as the tepals, especially in Lapeirousia, or may — any of the species under consideration: and it is pos- be white, possibly adding to the signal provided by sible that differential reflectance in the UV light the contrasting color of the perianth. In P. incras- range may add to the visual signals evident in the satum and P. magenteum the pollen is bright orange. — visible range. 74 Annals of the Missouri Botanical Garden Figure 7. peringueyi.—A. Pelargonium magenteum.—B. P. sericifolium.—C. P. incrassatum. (Scale: full size Flowers open during the day and often close par- tially or fully at night. They are typically unscented, at least to the human nose both in the open air and when several flowers are held in a warm confined space. Some species of Babiana, however, have a light to moderately strong scent (Table 3). The flow- ering season in the guild ranges from late May (one species), with a marked rise in July and a peak in early September, and continues until early October (Fig. 10). Individual species and populations re- main in flower for at least two weeks, or for a con- siderably longer time in the case of the shrubby Pelargonium species. Individual flowers usually last three to four days, and longer in species of Iridaceae when not pollinated. Species of Iridaceae are protandrous. The pollen is shed half a day to one day before the stigmas unfold and become available for pollen deposition. Unless removed by some agent, the pollen remains in place in the an- ther thecae. Species of Pelargonium are also pro- tandrous. The deciduous anthers are shed the same day that the flower opens, whereas the stigmas only unfold the following day. Flowers of the species of the guild are almost all herkogamous (and self-in- compatible, at least in L. dolomitica and L. silenoi- des) and thus require insect-mediated pollination. The only known exception is L. jacquinii, which is self-compatible and autogamous (Goldblatt et al., * Y PS MIS, MAD nse ge m. WAT ELLA Flowers of species of Geraniaceae belonging to the Lapeirousia silenoides guild, all pollinated by Prosoeca ifoli CPi Il size.) Nectar quantities are ample, and the upper range of nectar volumes for species in the guild is 1.1 pl to 9.6 wl (Table 2). Nectar sugar concentrations are mostly 24-30%, exceptionally as low as 20% in Romulea hantamensis and 22% in Babiana dregei Table 2). Nectar sugar analyses, available for 21 species (Table 2), show a characteristic sucrose- rich to sucrose-dominant pattern in the 15 species —. of Iridaceae examined. Sucrose : hexose sugar ratios range from a high of 4.13 in Lapeirousia dolomitica subsp. dolomitica to 1.20 and 0.94 in the two spe- cies of Hesperantha for which we have data, and 1.00 in Romulea hantamensis, the only species of that genus belonging to the guild. Most species have sucrose: hexose ratios in the 2 to 3.5 range. The pattern in two species of Pelargonium, how- ever, shows hexose dominance with sucrose : hexose ratios of 0.28 and 0.33. This is a marked contrast to the spectrum for Iridaceae. POLLINATOR IDENTITY Pollinator observations were obtained for 17 out of the total listing of 28 plant species (Table 4). These species are from throughout the range oc- cupied by members of the guild. In all of these instances pollination was carried out by either Pro- soeca peringueyi or P. sp., or, rarely, both (Figs. 1, 2). No other insects were seen to visit any of the > Figure 8. pollinated) by Prosoeca peringueyi.— alis.—bD. L. dolomitica subsp. dolomitica.—E. kampiae.—H. Tritonia marlothii. (Scale: full size.) Flowers of species of Iridaceae belonging to the Lapeirousia silenoides guild pollinated (or inferred to be A. Babiana framesii.—HB. B. curviscapa.—C L ^ violacea.—F. L. silenoides.—G. Sparaxis variegata subsp. meteler- apeirousia pyramidalis subsp. re- Volume 83, Number 1 Manning & Goldblatt 75 1996 Prosoeca peringueyi Pollination Guild Annals of the Missouri Botanical Garden ORAR SA OPO LYSE PSS = aa —— _ 02 oS WS 3i d SSS RM e9. Flowers of species of E 'eae belonging to the Lapeirousia silenoides EE pollinated (or inferred to be pollinated) by Prosoeca sp. nov.— hiana E hantamensis.—E. Hesperantha E (Scale: ful species of the guild during more than 200 hours of observation time except for three anthophorid bees, which visited but did not forage on individuals of L. silenoides. lt is almost certain that the plant spe- cies for which we do not have pollinator observa- tions will prove to be one or both of these fly spe- cies. Because fly species are allopatric, we have inferred pollinator identity on the basis of the range of the plant species for which we have no obser- vations. Prosoeca peringueyi is confirmed as the pri- mary pollinator of three species of Geraniaceae and —B. Lapeirousia oreogena.—C. L. jacquinii.—D. Romulea eleven species of Iridaceae, and P. sp. of five spe- cies of Iridaceae (Table 4). GEOGRAPHY The plant species with the viii n silenoides- type flowers are restricted to coastal and near in- terior southern Africa, a semiarid region " low, pre- dominantly winter rainfall. The 28 species have a collective range that extends from extreme south- western Namibia through the western part of North- Volume 83, Number 1 1996 Manning & Goldblatt 77 Prosoeca peringueyi Pollination Guild NO. SPECIES IN FLOWER AUG SEPT OCT JUNE JULY Figure 10. Flowering times of the plant species be- longing to the Lapeirousia silenoides guild. ern Cape Province, South Africa, an area known as Namaqualand, and reaches the northwestern por- tion of Western Cape Province (Fig. 11). The dis- tribution of individual s species within this area is often highly local (Goldblatt, 1972, 1984, 1985: Lewis. 1959: van der Walt «€ Vorster, 1988). At most, seven species of the guild are present in any quarter-degree square, and no more than four guild members co-occur locally. Diversity is greatest in the Kamieskroon area of Namaqualand and in the Pakhuis Mountains of Western Cape Province. Usu- ally the ranges of two or more species overlap. Up to seven species have been recorded in a quarter- degree square of geographical latitude and longi- tude. Species are infrequent in the north of the range, where only L. dolomitica subsp. dolomitica. Tritonia marlothii, Pelargonium cortusifolium, and P. crassicaule occur, and along the coast where the putatively autogamous L. jacquinii is the only rep- resentative. The combined ranges of Prosoeca peringueyi and P. sp. (Fig. 12) fall entirely within the main range of the plant species and accord almost exactly with the area within which two or more plant species occur. Prosoeca peringueyi has the wider distribu- tion, extending from northern Namaqualand to the Pakhuis Mountains, and P. sp. has a localized range along a corridor of high country in Northern Cape Province between Nieuwoudtville and the Hantam Mountains. POLLEN PLACEMENT Observation of living and pinned insects, corrob- orated by pollen washes, confirmed that pollen of a particular plant species is consistently deposited on a limited part of the insect's body (Table 4). We have identified four mutually exclusive sites of de- position: top of the thorax or dorsum; top of the head or frons; the base of the proboscis or face; and underside of the thorax or sternum and abdomen (Fig. 13). Pelargonium flowers have declinate sta- mens so that the filaments and anthers are situated below the mouth of the floral tube (Fig. 7) and pol- len deposition is invariably sternotribic. In P. in- crassatum (Fig. 7C), which has long filaments, pol- len is deposited on the underside of the sternum and thorax (Fig. 13D), but in P. sericifolium and P. 7A, B, 13C), which have very short filaments, deposition is on the face. Species magenteum (Fig. of Iridaceae belonging to the guild have either ac- tinomorphic flowers with symmetrically disposed stamens (Fig. 9B, D, E) or zygomorphic flowers with the stamens unilateral and arcuate (Figs. 8, 9A, C). The anthers are then either held above the mouth of the floral tube or are dorsal to it. and pollen 13A, B). length is also variable, being short in species of deposition is nototribic (Fig. Filament Lapeirousia and Romulea but relatively long in Ba- biana, and the site of pollen deposition varies ac- cordingly. In species of Babiana pollen deposition is mostly on the top of the thorax (in one species on the top and sides), in Hesperantha on the ventral head, whereas in Lapeirousia and Romulea it is on the frons. Often there are at least two species of the guild, and sometimes more, co-blooming locally and vis- ited indiscriminately by P peringueyi. At such sites, particularly in Namaqualand, Lapeirousia sil- enoides, Pelargonium incrassatum, and one species of Babiana, either B. curviscapa, B. dregei, B. fra- mesii, or B. pubescens, flower together. In the Clan- william District L. jacquinii, sometimes L. violacea, Pelargonium magenteum, and a species of Babiana and/or Sparaxis commonly form part of a local plant community. At sites in the Kamiesberg, central Na- maqualand, as many as four co-blooming members of the guild were recorded. Lapeirousia silenoides, B. curviscapa, Hesperantha latifolia, and Pelargo- nium incrassatum were noted near Leliefontein, and on Sneeukop we encountered Babiana dregei, H. latifolia, and two other presumed members of the guild, Xenoscapa uliginosa (endemic there) and Geissorhiza kamiesbergensis. — A similar situation prevails with species visited At different sites Babiana fra- mesii and either Lapeirousia oreogena or L. jacqui- nii, or L. jacquinii and B. sambucina, or B. flabel- lifolia, Hesperantha oligantha, hantamensis flower concurrently and are visited in- In the above examples, pollen contamination of one spe- by Prosoeca sp. nov. and Romulea discriminately by the same fly individual. 78 Annals of the Missouri Botanical Garden Table 4. Pollinator characteristics and effective tube length of species with the Lapeirousia silenoides-type flower. Measurements of insect mouthparts were made from individuals collected on the plant species concerned. Dash in column three reflects no pollinator recorded on that species. Insect tongue Anther to base length Family/species of tube (mm) Pollen deposition site (mm) Plants pollinated (or inferred to be pollinated) by Prosoeca peringueyi Geraniac Pelargonium cortusifolium ca. 35 ventral head — crassicaule 18-35 entral head — echinatum 30—58 ventral head — incrassatum 44—52 ventral thorax 28-3: magenteum 37-53 ventral head 30-35 sericifolium 43-06 ventral head 35-40 Iridaceae Babiana curviscapa 40-55 dorsal thorax 25-28 lregei 47-53 dorsal thorax 25-28 ecklonii 52-65 dorsal thorax — framesii 45-50 dorsal thorax 30-35 geniculata 45-55 dorsal thorax — pubescen ca. 62 dorsal thorax 32-35 sambucinc var. unguiculata 38-55 dorsal thorax — Geissorhiza kamiesmontana 23-32 ventral head — Hesperantha atifolia 20—35 ventral head 20-25 purpurea ca. 25 ventral head - Lapeirousia dolomitica subsp. dolomitica 25-45 frons 30 subsp. lewisiana )J-50 frons — jacquinii 35-45 frons 32-35 pyramidalis subsp. regalis 45-55 frons 32-34 silenc 45-60 frons 3540 violacea 40-45 frons 32-35 Sparaxis variegata subsp. metelerkampiae 40-45 frons 32-35 Tritonia thii 27—46 frons — Xenoscapa uliginosa 27-32 frons Plants pollinated (or inferred to be pollinated) by Prosoeca sp nov. Iridaceae Babiana flabellifolia 50-70 40-45 framesü 45—50 dorsal thorax 4048 sambuci var. longibracteata 35-55 dorsal thorax — Hesperanthc oligantha 3040 ventral head — Lapeirousia Jacquinii 35-45 frons 40—45 oreogena 55-65 ventral head 40-48 Romulea hantamensis 60-75 frons 4045 Volume 83, Number 1 Manning & Goldblatt 79 1996 Prosoeca peringueyi Pollination Guild 12 14 18 2 30 32 34 pee HP qme s nn ]1718 Qui [1720 [1721 [722 [1743 [124 11725 28 [1730 Jes Tem [732 Ns dirás fida A am 1820" rime ue 1605 T -H | LL 9 11 ga | M T " so o sien IUE iul: d ee puts fes 118 m j tH an 1920 An lies s 1924 m es 19: E Lau E W916 Tem 918 1919 - | 927 e 1929 1193014 L Lb | " (1918 113 a hos EH t JERE lc ai i ia An [1983 es H | Lal 12017 12018 [2019 mp as 082 2023 12024 m SE 027 IEEN NN T pr ae | . E» En TL [20827 d (208 [2088 20) 20- ch gos pou |2015 ll] 4+ {| Ll —À 2123 m 2126 2177 ess Tetas [2125 £39 | 31 | St} | CN be [213 na | | 5 2220 2221 ame 2223 |2224 [2225 [2226 [2227 2228 den To | / aor t 1 ES : }— = - | : - T (M 2216 |2217 zi 2219 "TH 3 dam ne 7 & 83 I Joa E 122 L2—— | I: Lil ip T Y : | | | ux oe oe ee al LEE m dios cl ia | el isi d pm IM P sis [aste (2397 eg 8329 Ss 20 IN Caca Jaaa} 123 3 | | 1 + E ‘2 21 [2422 2a TT TT - TT Ic ido abi [2482 2420 2 4 im B zoe IET BETIS. 4 T TN L 11 wee E i SEE a Er zai | E T diis. p" da e em um 24ga F sal pa [24 bs | b. AS i | | |b a [ . [uM ssi 2516 2517 ae tte 2 za pst 2524 [oso da. ar | LL al ai I | | ATV ——— te ! E z | T———r1 | = A — iz - L i | IE SE | T NANO: je m er EP a 2623 2624 12628 aia a [2628 [2629 2630 A 26 Li T T | "pet q.s 26 ce ial 2629 bded d sjam Ld ng He ai hia 4] ——4— BE! Si Jada. a 2723 (2724, (2725 (2786 (ear Fe e 2730 xis dl AER EE mA | | ++} | ts 2619 |2820 | 2823 aula aren mur 2 la t | | H+ 28 tL LLL gu id sa 20 2 icd | an ama aa e 2830 [2831 a er pL 28 | | | | SEN L E | e A | i eet gee ias E eee as iaa Vis 2000 2 ll al es sai “ts ae | [297 323 12980 [29/6 | 7 A 1 FT | | ¡e | 6 I-II | T] , dl JEN | | | Lr | | | | mi d 7 - | Tn HM Tg sois (3020 isi o 3022 3023 mas (3827 [zed | ] 070174 30 i TH roth | ee Ji raas a 3123 sn [res is n | | = | | | 1 41 | fi T LI d Li | | [3220 [3221 3222 [3223 E (3225 (3226 gz ELA | if mu F | | ! jt EN | 432 paa 3323 3324 (9925 [3326 | E [1 EE NNENEN T TET + " ERE ee "PR | 2 | | [o mm {gn ajo (3429 za 22 3493 [a [3425 RREI m H IH 4 El | | I "i : | on L1 HE +t ++ | | NBI 1990 | 34 Figure 11. Map of southern Africa showing the distribution range of species with the Lapeirousia silenoides-type Pi Fisures indicate the total number of species recorded per quarter-degree square of geographical latitude and longitude. cies by that of another is minimized or prevented — species in the guild the mean coefficient was 0.77. by the use of different deposition sites as outlined suggesting selection for reduced pollen. contami- above. At any study site that included two or more — nation. The index greatly underestimates the pollen species of the guild, the pollen of each species was. contamination coefficient because not all of the placed on a different part of an insects body (e.g. species in a grid co-occur locally. The local co- Fig. 13). The only exception to this pattern was at occurrence of more than one species using the Botterkloof Pass (and presumably other localities same pollen deposition site is rare, and we only where these two species co-occur) where pollen de- — know of the single example mentioned above. position sites for Lapeirousia jacquinii and L. vio- lacea are identical. — | FORAGING PATTERNS A crude estimate of potential pollen contamina- tion was determined by comparing the number of Adult specimens of Prosoeca peringueyi have guild members recorded from any quarter-degree been collected from late July to late September, square with the number of pollen deposition sites — with a peak during mid August to mid September. utilized by these species. The number of placement Specimens of P. sp. have only been recorded from positions exploited in any quarter-degree square is mid August to mid September. Both species have a positively correlated with the total number of guild similar foraging behavior. The flies move rapidly members occurring in that grid (Fig. 14). A coef- — between flowers and hover for two to three seconds ficient of pollen contamination was calculated for while orienting and inserting their proboscis into grids containing more than one species of the guild — the floral tube (Goldblatt et al., 1995) (Figs. 3—6). by dividing the number of species into the number. In species with zygomorphic flowers the fly always of loading sites per quarter-degree grid. For the orients itself in the same way, approaching the flow- 80 Annals of the Missouri Botanical Garden 12 14 16 18 20 22 24 26 28 3 - eT TA - 1720 [1741 [1722 [1723 [i24 [1745 [1726 [i 1728 TT 1 T 7 743 fina J — == TY | | | | =] ] | LA ta 1063 1024 (1685 Tie = cs ll | 817 a | 8 11814 Ter sl = nae be va EF 7 ‘TT 18 181 18 : LI | LL 1l — M Hi 1921 [rege [1645 [teas [ds Nos [1037 1928 J] + sii - 1934] — = = = = = EN | "i 4-1 [2021 |2022 |2023 |2024 mt 2086 ne | = 2009 2030 i E 2081 - bo) | ba as 20% 2035 5:20 H- — 121 [2132 ada mds [m is sis 272 Kaas 2129 2 Ty sal do [n P iJ 2131 12382 2493 ase ig a 4 si (ake 2221 (2222 |2223 [2224 |2225 [2226 (2027 s |S ll | | eir sis [2219 2220 | 2221 |222 4 5 - d 2228 jabog "NE bol 1+ d ci T 2 oig 2288 2254 |a $5 122 ——]1-7 L — " i= "T e L | Sia 2322 |2323 |23b4 Tree a n 2sie |2317 2318 aso 2320 "T si ss s ag € xd ns 2390 |asjr as ads Sc 7 2416 2417 |2418 2419 2420 2421 sil 2423 2454 [24 Ji | | a T -jt de 2516. 2517 EN 2519 2380 [2521 |2522 |2523 2534 125 a arfaa 2508 2529 2580 2581 _| ma L i cs BS 520) | 2621 i Jen zh ario |2720 (2721 El | "e mn 1 | —1—1 1281828167 2018 2819 |2820 2821 |2822 [2823 |2824 28 5 |2526 apr 2828 {+--+ Pr d | AA — té 918 (2919 2020 ms [2922 |2923 [2994 [2925 |29 dd or $29 Deus gui b | l ar | — + fe — m | ee abr s A 4 TH 3019 |3020 3081 |3022 |3023 3024 {3025 3026 |3927 |322; 30: IO SSA | EN | Ni ii [tt] Ha 120 (aten [3122 [3123 S184 |3125 3126 inn 3128 [3129 "m | | | = ee E T + TT — LA AA | TN E a J : | LLL HHHH 3220 3221 [3222 |3223 a [aas 6 3227 3228 229 { | jj tf} 4} | [ | | LLLI | D DB 1l, L| Li it | te 31 [3322 [3323 |3324 |3325 [3326 |33g7| 1 cae 3317 e318 [s3}9 33 asl cad hl bs i oe A | | — + —L— t4 —+ a en cae e th Ted l3ajo [3420 [341 92193 sf 117 pa TTT - de : e 3 | | 1 Í TL 1 Figure 12. Map of southern Africa showing the distribution ranges of Prosoeca decia (closed circles) and P. sp. (open circles) recorded on quarter-degree squares of latitude and longitude. The combined ranges of plant species belonging to the guild pollinated by these two fly species is indicated by the heavy salina er directly from the front. The flies are unable to of each species located on a different part of the discriminate between flowers that have already body been visited either by themselves by other individ- uals and may repeatedly visit the same flower even DISCUSSION when all the nectar has been removed (pers. obs.) When nectar is absent visits are brief. Both fly species are active on mild to warm days Our observations indicate that plants with flow- from mid morning to early afternoon, and again in ers conforming to the following syndrome constitute the late afternoon. Foraging is most active on warm a guild adapted for pollination by the long-probos- days between 12:30 and 2:30 PM, but some for- cid flies, Prosoeca peringueyi and P. sp. nov.: peri- aging occurs at almost any time of day. Density of anth or hypanthium tube narrow, straight or slightly visitors varies considerably, ranging from 4 to 5 curved, and 30-60 mm long; tepals or petals short flies present locally at the same time, or as few as — in relation to tube length and pigmented dark pur- l or 2 over periods as long as an hour. Flies remain ple to crimson, or sometimes lilac or pale mauve at each flower for 3-5 seconds, and pollen is pas- with nectar guides consisting of white to cream sively brushed onto various parts of the head, tho- — spots and streaks and areas of darker pigmentation; rax, or abdomen, depending on the species visited. and exserted and prominent anthers and stigmas Pollen depositions are usually heavy enough to be that are presented outside the mouth of the tube so visible to the naked eye against the dark bodies of that they will contact the body of any animal that the insects. Commonly, pollen of two or three dif- probes the floral tube. Associated with these fea- ferent species can be distinguished by color, that tures is the production of nectar with a relatively THE LAPEIROUSIA SILENOIDES-TYPE FLORAL SYNDROME Volume 83, Number 1 1996 Manning & Goldblatt 81 Prosoeca peringueyi Pollination Guild Figure 13. rug placement of pollen on Pro- soeca pering eyl.— tana eurviscapa, dorsum.—B apeirousia silenoides, bus —C. Pelargonium sericifol- ium, face.—D. P. incrassatum, sternum (ventral thorax and abdomen). Hale hing indicates the site of pollen deposi- tion. (Scale: full size.) high sugar concentration and the absence of floral odor (present in two species). The plant taxa oc- curring north of 29°S, i.e., in the Richtersveld and southern Namibia, comprise a distinct subset of the guild characterized by the paler-colored lilac or pale mauve flowers. This group comprises Lapet- rousia dolomitica subsp. dolomitica, Tritonia mar- lothii, Pelargonium cortusifolium, and P. crassicau- le. Only Pelargonium sericifolium has no measur- able floral nectar: we presume that it is an example of pollination by deceit (Goldblatt et al.. 1995). In color and shape the flowers closely resemble those of Lapeirousia silenoides and P. magenteum. These species frequently co-occur with P. sericifolium and have ample amounts of nectar of relatively high sugar concentration. Two species of Babiana and one subspecies of L. dolomitica are exceptional here in having sweetly scented flowers. Presence of scent is usually considered to add to the attractive- ness of the flowers, but in Babiana we are inclined to consider it a vestigial trait in view of its rarity among members of the guild. Most of the species of Babiana sect. Babiana, to which the Prosoeca- pollinated species belong, have strongly scented flowers and are bee pollinated (e.g.. B. odorata, B. scabrifolia, unpublished observations). Nectars of moderate sugar concentration and typ- ically sucrose-rich to sucrose-dominant seem to be characteristic of plants pollinated by active insects such as bees of the family Anthophoridae and by long-tongued flies (Goldblatt et al., 1995), although not of plants pollinated by other Diptera (Baker & 1983, 1990) such as Calliphoridae, Musci- dae. and Tachinidae. The nectar sugar concentra- tions of flowers pollinated by Prosoeca and other Nemestrinidae (Table 2) are typically somewhat lower than those of bee-pollinated flowers. This may be related to the difficulty of sucking up liq- uids of higher viscosity, as is the case in lon y- tongued butterflies (Johnson & Bond, 1994). low sucrose to hexose ratios in the two species of Pelargonium (Table 2) contrast with the pattern in Iridaceae belonging to the guild. Species of Pelar- gonium are visited as avidly as any of the Iridaceae. Indeed, on the basis of the frequency of visits, Pel- argonium incrassatum appears to be one of the most important nectar sources for P. peringueyi. This leads us to conclude that nectar sugar composition is not a significant factor in the P. peringueyi pol- lination guild. Not all species with long perianth tubes and dark purple to crimson flowers belong to the guild. In some species that have flowers apparently conform- ing to the guild the lower part of the perianth tube 82 Annals of the Missouri Botanical Garden 12 14 |. 16 8 22 26 28 30 2 34 vc Rma DAE TUS [UU [UP [1719 [020 [1731 [1722 [1723 m: [1745 [1726 [1727 [1728 [1729 [1730 [1731 T1742 Yizds hi I— 71 | { | — | | : 181p |1820 1883 11834 [1825 T1 1827 |1828 |i 1830 14 1181p [1616 [1817 |181 1831 11832 i 18 Nes 11 p | | 1 18 - lui ig lisi hozo 1921 [1922 |1923 [roda frogs Node 1927 li 8 119 1030 Gi TOR lias [rota 191 T T 1992) [1933 Tig: 20- i CIN 2016 2016 2017 s Bote [2020 [2021 12022 |2023 |2024 2028 208 s 2051 20: 20 LL | i = our 26 Jano |2120 |2121 |2122 |2123 [2124 [2125 [2126 [212 2128 155 " 2130 Tara [pe ani m a d 2 | is 2131 [292 2133 [br | | | | f gprs ZE ais |2220 (2221 |2222 |2223 |2224 |2 2227 |2228 L Pj a 22 5 nt zu f 22 2283 125 22 | — 171 Iilll. | | | > [2294 [2pas 22 LLLI lis s zd as aie {2318 [osa 2321 |2322 (2023 |23b4 |2325 |2326 (2322712308 [2359 [2355 123 2s 12a I— | | - - LLÀ LT T Sune s ba y | + 2d 2415 2416 au zd 2419 |24 E 2422 |2423 [2424 [2425 2426] |2497 24 2429 [2430 24% zia da : a 24l ' | a | | | | 1 | | = me 4 L ! 2521 [2522 2523 |2524 |2 7 | A | 2514 2545 2516 2517 |2518 2519 2 4 25 E32 | 2504 | 2526 |2527 an 2529 25 E 7] EA ] | | a | T . ta i tz zeis (2610 2623 (2624 2627 2629 sif LI 26^ [BER dl and i T T ma + | NEZ; |+—4 T | | _ j l Y ! | | 2 JA are aro pmi ze eo pup re [ares [are 5 mere ma [amo oe ae FUN! UM D | nu id TAL + sionista [2018 |2819 [2820 [2821 [2822 [2823 12824. ae ee 2828 Ea y ass E ILL log | | 1 | 1^ E | : | | | —Q a: -T- gef ap egl2nie 1210 [2000 Jas [ama ‘2000 (2094 [ass 206 [2527 Taupe Papa [aso Taga PtH T | EN | | | | | | | | | Taob2 13023 130 | | | TTA ++ 1 [3022 ¡3023 [30p43 ++ dd T. ee tft 31 2 9/3120 3121 13122 [3123 [3124 |3125 [3126 13127 3188 Sido 130 T- i = T ¡ERASE | ZII ARA ++ tar 9 [320 3222 |3223 |32p4 3295 3226 [3227 32084209 +. HJ ILI — 3 BERE EB Trt tr { LI] 32 IEEE PID dd QATAR | 1 1 i "ZIITTITEL, a T -- 3 9 [3480 |34? gapa Taaa | 5 | To HHH EE | | a uns LT | | LILILLI MATES 16 18 20 22 24 26 28 30 32 34 36 igure Map of southern Africa showing the differentiation of n loading sites. Figures indicate the number of different pollen- loading sites utilized by plant s species per qua egree square of geographical latitude and lon- gitude. The total number of species occurring in each grid is tried in Figure 11. is so narrow as to effectively prevent the penetra- tion of an insect's tongue. In these cases nectar is forced into the upper part of the tube where it is accessible to a variety of insects. Long-tongued flies visiting these species may obtain nectar but will not come into contact with the pollen and stigmatic surfaces. This situation has already been described for the long-tubed species Lapeirousia montana, which superficially appears to be a member of a long-tongued insect pollination guild (Goldblatt et al., 1995). The tube is up to 43-55 mm long, but is so narrow in the lower half that the nectar is forced upward and lies only 10-15 mm from the mouth of the tube. The flowers are visited by a variety of bees, Lepidoptera, and bombyliid flies, all of which may accomplish pollen transfer. Like- wise, we have found that whereas B. sambucina var. longibracteata and variety unguiculata do conform to the guild except in their scented flowers, variety sambucina does not. Although the perianth tubes of all three varieties are 30-55 mm long, in variety sambucina the lower part of the tube is effectively blocked and nectar is forced into the top of the cylindrical part of the tube where it is accessible to a variety of insects. The situation in variety sam- bucina is no surprise because its distribution range is mostly outside that of Prosoeca peringueyi or P. sp. A comparable situation exists in Babiana dre- gei, which has a tube mm long, thus longer than the mouthparts of P. peringueyi. In this plant, too, the tube is narrowed in the lower 20-30 mm rendering the functional tube length much shorter than the external length, and the nectar is thus available to individuals of P peringueyi with tongues of moderate length. Color of the perianth and of the nectar guides may also be misleading. Babiana pauciflora G. Lewis has purple flowers with tubes 35—45 mm long, but a strong fragrance, bright yellow nectar guides, and a flowering period of June appear to exclude the species from the guild. Babiana atten- uata G. Lewis and B. truncata G. Lewis also have Volume 83, Number 1 1996 Manning & Goldblatt 83 Prosoeca peringueyi Pollination Guild perianth tubes in the 30-45-mm-long range, but the blue or mauve flowers have yellow nectar guides and, in the case of B. attenuata, fragrant flowers. We do not regard them as guild members. THE PROSOECA PERINGUEYI POLLINATION GUILD The Prosoeca peringueyi—-P. sp. pollination guild (hereafter referred to simply as the P. peringueyi pollination guild) is unique in its cian of a long floral tube with a distinctive perianth coloring, and its restricted occurrence in so small a part of southern Africa is striking. While a long-tongued fly pollination syndrome has been described for the Cape Flora (Whitehead et al., 1987). this covered only plants with cream to pink flowers with dark nectar guides. The P. peringueyi guild thus deviates significantly from other long-tongued fly pollination systems. There are at least two other guilds of plant species adapted for long-tongued fly pollination in southern Africa. Late spring- and early summer- flowering species with white to cream flowers with — red nec a guides (e.g., Lapeirousia anceps (L.f Ker-Gawl., L. fabricii (Delaroche) Ker-Gawl.) are pollinate 4 by either Moegistorhynchus longirostris (Nemestrinidae), or Philoliche gulosa, or P. rostrata (Tabanidae) (Vogel, 1954; Goldblatt et al., 1995). Summer- and autumn-flowering species with long- tubed, blue, pink, or white flowers (e.g... Nivenia stenosiphon Goldblatt (Iridaceae, Goldblatt & Bern- hardt, . Disa oreophila H. Bol. (Orchidaceae, pers. obs.). Gladiolus microcarpus G. Lewis (Irida- ceae, pers. obs.), and Zaluzianskya microsiphon (0. Kuntze) K. Schum. (Scrophulariaceae, pers. obs.)) are pollinated by Prosoeca ganglbaueri over a wide portion of southern. Africa. Both these syndromes are quite distinct from the Prosoeca peringueyi pol- lination system in the flower color, plant and insect distribution, pollinator identity, and flowering time. Tongue length is surprisingly variable in Prosoe- ca peringueyi and somewhat variable in P. sp. The variation corresponds locally to floral tube length of their nectar plants. In the Kamiesberg Mountains where Hesperantha latifolia, Babiana dregei, and Pelargonium incrassatum are major nectar sources and effective tube length is 20-40 mm, P. perin- gueyi has a tongue of 20-25 mm long. In other places in Namaqualand where L. silenoides and B. pubescens have floral tubes in excess of 50 mm, individuals of P. peringueyi have tongues 35—40 mm long. In the Olifants River Valley where L. pyr- amidalis and L. jacquinii are major nectar sources for P. peringueyi and have tubes 35-45 mm long. flies caught on these species have tongues 30-33 mm long. Clearly local variation in tube length in the flowers of food plants is tracked by the polli- nator. The disparity between the length of floral tube in plants with the Lapeirousia silenoides-type flower and the shorter length of the mouthparts of the sole pollinators is quite easy to explain (Goldblatt et al.. ecords of nectar secretion show that these flowers secrete ample amounts of fluid for insect- pollinated flowers, and it is unlikely that dominant pollinators are ever forced to extend their mouth- parts to the base of the tube unless all the nectar has been removed by earlier foragers. More impor- tantly, Darwin (1877) hypothesized that successful pollination of spurred orchids occurred when or- chids evolved floral spurs slightly longer than the tongues of their pollinators, ensuring maximum contact between the insect’s head and the orchid's column by forcing the insect to ram its head down the floral throat. This has since been shown exper- imentally by Nilsson (1988). As in the nectariferous orchids, species with the L. silenoides-type flower force their pollinators to make maximum contact with the anthers and stigmatic surfaces that block or at least encircle the entrance to the floral tube. The combined geographical ranges of all the spe- cies pollinated primarily by Prosoeca peringueyi and P. sp. (Fig. 11) are greater than the ranges so far recorded for the two fly species (Fig. 12). To the north in southern Namibia P. peringueyi may simply essen- — not have been collected yet. To the south, tially below the 33rd parallel and west of the 19th north-south parallel, the guild is represented main- ly by cies is known to be au- togamous (Goldblatt et al., 1995), although it is ac- tively pollinated by both species of Prosoeca within L. jacquinii. This spe their ranges. Presumably autogamy has enabled it to extend its range outside that of its facultative pollinators. The only other species that occurs out- side the ranges of the two flies is Pelargonium ma- genteum. The reasons for its wider distribution to the east of the range of P. peringueyi and P. sp. are unknown. The potential for pollen contamination from other members of the same guild is greatly increased for plant species that share a single specialized polli- nator species. One strategy to enhance segregated gene flow is differential placement of pollen on the insect body. This strategy is developed in Orchi- ie ‘eae (Dressler, 1968a, b; Manning & Linder, 1992), Scrophulariaceae (Steiner & Whitehead, 1988. 1990). and various other families (Grant. 1994). In flowers of species belonging to the P. per- ingueyi pollination guild, four mutually exclusive sites have been identified. In this guild the contam- suggesting — ination coefficient is never below 0.5, 84 Annals of the Missouri Botanical Garden that there is a threshold of pollination efficiency that determines the number of species that can ef- fectively use the same loading site. In view of me- chanical constraints on flower architecture in the genera involved, there appears to be a definite limit to the number of species that can enter the guild at any locality. Pollination contamination may thus be a significant factor in influencing species pack- ing in specialist systems. Locally, the niche offered by pollination by Prosoeca peringueyi or P. sp. ap- pears to become saturated by the presence of more than three co-flowering species. The presence of four similarly adapted species is rare ifferential placement of valleys; on an insect’s body demands precise orientation of the pollinator relative to anther position. Floral zygomorphy fa- cilitates this, and we suggest that this is an impor- tant factor in favoring genera with predominantly zygomorphic flowers. e Lapeirousia silenoides pollination syndrome appears to have evolved in five different lineages in Lapeirousia subg. Lapeirousia, a taxon that in- 1995). A comparably polyphyletic evolution of the syndrome cludes just 21 species (Goldblatt et al., appears to have occurred in Babiana and in Pel- argontum, as species with this syndrome in these genera are taxonomically isolated. ORIGIN OF THE PROSOECA PERINGUEYI POLLINATION GUILD The Prosoeca peringueyi pollination guild in- volves at least six genera of plants in two families and two species of Prosoeca (Nemestrinidae). Al- though the plant species belonging to the guild and their pollinators are restricted to western southern Africa, all of the genera, both plant and insect, e tend beyond the range of the guild. Other ea of Prosoeca, both long- and short-tongued, visit flowers of various colors including white, cream, pink, lilac, blue, and yellow (Johnson, 1992; Gold- blatt et al., has been observed visiting species outside the guild 1995; pers. obs.). Prosoeca peringueyi for nectar that are pale lilac or pink and green in color. In addition, the northern taxa of the guild are lilac or pale mauve and some have been confirmed to be pollinated by Prosoeca peringueyi. The orig- inal determinant for the characteristic crimson or purple color in the guild was thus apparently not directed by the innate preference of P. peringueyi for that particular flower color, and was in conse- quence presumably plant-directed. Comparative studies of each of the guild genera that pollination by long-tongued flies is apomorphic. suggests Of the guild members only Babiana has flowers in which dark blue or violet color is the plesiomorphic condition. In addition, even short- tubed species of this genus secrete fairly large amounts of nectar. Also, Babiana has flowers with a wide gullet, which makes access to the pollinator mouthparts easier than the narrow-tubed flowers in the other guild genera, for example, Lapeirousia and Pelargonium. These factors suggest to us that the first steps in the development of the Prosoeca peringueyi pollination system were through the ge- nus Babiana. Significantly, both Babiana and Lap- eirousia, the two genera that individually have the most species in the guild and together comprise 54% of the guild, are largely developed in arid hab- itats. It may be that the origin of the syndrome in Babiana and its subsequent development in spe- cies of Lapeirousia was a consequence of higher species richness in these genera in the western part of southern Africa. The predominant dark purple flower color in Babiana would explain the charac- teristic floral coloring in the L. silenoides pollination syndrome, unknown in other Prosoeca pollination guilds. Subsequently, species in other genera could be expected to enter the guild in response to the reproductive benefits derived from these pollina- - ors. The advantages to the plant species of a dedi- cated pollinator are obvious and include increased pollination success and decreased pollen contami- nation and loss. Pollination success in one popu- lation of Lapeirousia pyramidalis subsp. regalis that we investigated was 45% (SD + 25%; n = 23). To the pollinator, however, the energetic rewards of flo- ral specialization are important. For large active in- sects that hover while foraging the energy demands are likely to be high. Long-tongued nemestrinids are capable of feeding from short-tubed flowers, but these are smaller and hold far less nectar than that typically present in the long-tubed flowers in the guild. In addition, the flies are in competition with other insects that can obtain the nectar in short- tubed flowers. Long-tubed flowers can contain large amounts of nectar that cannot be collected by short- tongued insects. They are therefore an attractive energy source for insects able to exploit it. We speculate that there is reduced pressure to darken the flower color in the northern members o the guild. This more arid region supports both fewer plants and fewer pollinators, and a more faculta- tively generalist pollination system might be fa- vored. This is borne out by the greatly extended flowering periods of Pelargonium cortusifolium anc P. crassicaule, which although peaking in August and September are prolonged far beyond the flight period of Prosoeca peringueyi. During this time they Volume 83, Number 1 Manning & Goldblatt 85 Prosoeca peringueyi Pollination Guild are presumably visited by other insects, perhaps bees and bee flies. Significantly, both species have short anthers, which will contact visitors of a range of shapes and sizes. Possibly a threshold diversity is necessary before selection pressures become strong enough to favor such specialist pollination systems. EVOLUTIONARY IMPLICATIONS OF THE PROSOECA PERINGUEYI POLLINATION GUILD The Prosoeca peringueyi pollination guild ap- pears as distinct as other highly specific pollination systems in the African subcontinent, including those involving sunbirds, Nectarinia species (Re- belo, 1987), oil-collecting bees in the family Mel- ittidae (Steiner & Whitehead, 1990, 1991a), resin- collecting bees in the family Megachilidae (Steiner & Whitehead, 1991b; Armbruster & Steiner, 1992), other guilds of long-tongued flies (Goldblatt et al.. 1995; Johnson & Steiner, 1995; Manning & Gold- blatt, 1995), and the butterfly, Aeropetes tulbaghia (Johnson & Bond, 1994). Where such systems oc- cur they contribute to the particular floral charac- teristics of various plant communities. The recognition of the Prosoeca peringueyi pol- lination syndrome and the way it functions is the key to understanding the presence of a series of species with unusual dark purple or crimson flow- ers with long floral tubes in the flora of coastal and near interior western southern Africa. The ecolog- ical niche presented by these two flies is so specific that it will allow fly-pollinated and non-fly-polli- nated members of the same genus flowering at more or less the same time to coexist with little or no hybridization. The diversity of pollination systems there is one of the reasons why some 30 species of Babiana (50% of the total species) and 19 species of Lapeirousia (48% of the total) co-occur in this area. Like the members of the Aeropetes tulbaghia pol- lination guild (Johnson € Bond, 1994), strong similarity in floral morphology between members of the Prosoeca peringueyi guild. This suggests strong selection for floral conformity. This may be a characteristic of guilds in which the pol- linator is an insect that is not flower constant. there is Prosoeca peringueyi and P. sp. nov. may be re- garded as keystone species. Such species are defined operationally as those that, by their effective disap- pearance from a system, would cause (directly or indirectly) the virtual disappearance of several other species. The extinction of either P. peringueyi or P. sp. nov., but especially the former, would result in significant decreases in seed set in many of the spe- cies in the P. peringueyi guild, prevent outcrossing, and might lead to their ultimate extinction. Literature Cited Armbruster, W. S. E. Steiner. 1992. Pollination ecology of four Daléihiampia species a in northern Natal, South Africa. Amer. J. Bot. 79: 306— 313. Hallar: H. G. & I. Baker. 1983. Floral nectar A eis stituents in relation to pollinator type. Pp. 117-14 s & R. J. Little (editors), Handbook of " x- ceed Pollination Biology. Scientific and Academic Editions, New York. —— & ——. 1990. The predictive value of nectar ar to oa recognition of pollinator types. Israel . Bot. 39: 157-166 Da. 1. 1877. The Various Contrivances by which British and Foreign ^ m are Fertilised by Insects. D. Appleton, New Dressler, R. L. 1968a. Bbsenations on orchids and eug- lossine bees in Panama and Costa Rica. Revista Biol. Trop. 15: 143-183. — ———,. 1968b. A pid by euglossine bees. Evolu- tion 22: 202-21( Faegri, K. € L. van der Pijl. 1979. neige ee Third revised edition. Press, Po Fletcher, B. & S. K. Son. 1931. A veterinary ento- mology i ET Part XIV. Indian J. Vet. Sci. Anim. Husbandry 1: 192-199. Pe A revision of the genera a Pourret and Anomatheca Ker in the E r rainfall r gion of South Africa. Contr. Bolus Herb. 4: 1-111 — 19 A revision of Hesperantha eee ca in J. S. African The Principles of Pergamon ~ a ^ © = ^ £z -— — the winter rainfall area of southern Africa. Bot. 50: 15-141. 1985. Systematics of the southern African genus Geissorhiza, (Iridac eae-I[xioideae). Ann. Missouri Bot. Gard. 277 —— & i Bernhardt: 1990. Pollination biology of Niv- enia (Iridaceae) and the qune Cope e of heterostylous self- compatibility. ine a ——— ;. Manning & P Be UE 1995. in "his subgenus Lapeirousia Cr Ixioi- leae). Missouri Bot. Gard. 82: 517-534. Grant, V. 1981. Plant Speciation, edition 2. Columbia Univ. Press, New York. 1994. Modes and origins of mec prea E Natl. Acad. — o> Pollination ethologie a PE in angiosperms. Proc. i. US i —10. E . Grant. 1965. Flower Pollination in the Phlox Fa C bia Univ. . New Yor S )92. Plant auth (^ RETA Pp. JS in a ^ Cowling (editor), The Ecology of Fyn- Fire and Diversity. Oxford Univ. Press, m ES Nutrie nts, — — & V. f? Bond. 1994. Red flowers and dg pollination i in the fynbos of South Africa. Pp. 137-148 in M. Arianoutsou & R. Groves (editors), Plant- Animal Interactions in Me se aa Ecosystems. Kluwer ademic Press, Dordrech — & K. E. Steiner. 1995. Long proboscid fly polli- nation of two orchids in the Cape Drakensberg Moun- t In pres The genus pem J. S. African Bot., ains. P Syst. Evol. 1 Lewis, G. J. 1959. Suppl. 3. 86 Annals of the Missouri Botanical Garden Manning, J. C. & P. Goldblatt. 1995. Cupid comes in many guises: The not-so-humble fly and a pollination guild in the iubes Veld & Flora 81(2): 50—52 P. Linder. 1992. Pollinators and evolution in Daperit or why are there so many species? S. African J. Sci. 9. Marloth, R. 1908. Some observations on entomophilous flowers. S. African J. Sci. 5: 110-113. Nilsson, L. A. 1988. The oce of flowers with deep corolla ra Nature 3 = 147-149. Ogden, E. C., G. S. Rayn s & D. M. Lewis. 1974. Mara i Sampling een "Pollen. London. Rebelo, A. 1987. Bird sure in the Cape Flora. Pp. 83-108 in A. elo (editor), A Preliminary Syn- thesis of nis pes in the Cape Flora. CSIR, Pretoria. Root, R. B. 1967. The niche exploitation pattern of the blue-gray yer " ther. Ecol. Monogr. 37: 317-350. Steiner, K. E. V. Whitehead. 1988. The association between oil- wah ing flowers and oil-collecting bees in the Drakensberg of pup diga Monogr. Syst. Bot. Missouri Bot. Gard. 25: 2 199( ). Pollin nie adaptation to —— creting flowers—Redivi iva and Diascia. Evolution 44: 1701-1707. Oil flowers and oil bees: Fur- ther evidence for salad adaptation. Evolution 45: 01. 1493-150 & 99]b. Resin collection and the pol- lination of Dale ans capensis (Euphorbiaceae) by d cordatu um (Hymenoptera: Megachili- dae) in Sout in S. Africa 54: 67-72. Van der Walt, J. . Pelargoniums A Southern Aion oie 3. Ann. Kirstenbosch Bot. ard. 16. Voge i S. 1954. nos Wi the Typen als Elemente der Sippengliederung. Bot. Stud. 1: 1-338. Whitehead, V. B., ee A. G. Rebelo. 1987. Insect pollination in the Cape Flora. Pp. 52-82 in A. G. Rebelo (editor), A Preliminary Synthesis of Pollina- tion Biology in the Cape Flora. CSIR, Pretoria. REVISION OF PAVETTA SUBGENUS BACONIA (RUBIACEAE: IXOROIDEAE) IN CAMEROON! Stephen D. Manning? ABSTRACT Pavetta subg. Baconia (Rubiaceae tribe Pavetteae) is a group of forest shrubs in tropical and subtropical Africa. It includes species of Pavetta with bearded corolla throats, subterete Baconia, and the Cameroon species are described; five previously rec Ged speci revised here. es are placed in synonymy. Seven series us corolla tubes usually shorter than corolla lobes, terete or style tips, and m. ineo of colors other than hes k. asd. is a major center of diversity for subgenus enty-nine species are recognized, including eleven newly ed earlier to subdivide subgenus Baconia have iei abandoned owing to inconstancy of character states used to delimit most of them. Sixteen Cameroon species are presently unreported from elsewhere. Aspects of leaf morphology are useful taxonomically. habitat preferences, descriptions of each speci Infraspecific taxa are recognized in seven of the twenty-nine species. An ov and d a tive strategies. Habitat pre de erences, distributions, and distinguishing features follow view is provided of these, other aspects of morphology, Pavetta L. shrubs (less often small trees or geofrutices) in for- consists of about 400 species of est understories (less often savannas) of the paleo- tropics and paleosubtropics. Its center of diversity is in Africa. Most taxa bear characteristic bacterial nodules sensu Lersten & Horner (1976) on their leaves (Zimmermann, 1902; von Faber, 1912: Ler- 1974, 1975; 1976). Most species are infrequently collected. The similar ge- —. sten, Lersten & Horner, nus /xora L. is distinguished from Pavetta in having style tips recurved at maturity, and no species of Ixora have bacterial nodules (Bremekamp, 1934). The taxonomic history of Pavetta is recounted by Bremekamp (1934: 12-13), who produced the only comprehensive monograph of the genus (1934, 1939a, 1939b). By 1939 Bremekamp had recog- nized 377 species, 206 described by himself, and 3 subgenera: Eupavetta Bremek., correctly referred to as subgenus Pavetta (316 species, Africa, Ara- bia, Asia, tropical Australia, Sri Lanka, Melanesia, Philippines); Baconia Bremek. (58 species, all Af- rican); and Dizygoón Bremek. (3 African). e Candolle (1807) described the monotypic ge- species, all trop- ies nus jus DC. with B. corymbosa DC. (presently . Williams). He dis- tinguished Baconia by its hanson flowers not sur- Pavetta corymbosa (DC.) F. rounded by bracts, bearded corolla throat, noncil- iated stipules, projecting anthers, and "simple stigmas.” Baconia was reduced to synonymy of Pavetta by Bentham and Hooker (1873), followed (1877). Baconia as a subgenus, including members of Pav- by Hiern Bremekamp (1934) recognized etta with the following characters: bearded corolla throat: short, stout corolla tubes; minute bracteoles at flower bases: and white or colored drupaceous fruits. This group was kept by Bremekamp within Pavetta, rather than restoring it as a genus, because of its tetramerous flowers, cylindrical upper part of small stigmatic lobes, and, style, in most species, ! This manuscript is adapted from a pa toral dissertation pe. I thank John Dwyer (MO), Pe . Charlotte Taylor (MO), and Henk van der Werff (M m nd Patricia Kernan a Ellis, and Je hg a O) for gone 'an Magis and €: Yatskiese h (MO) for field suppor or othe 'rwise facilitating he puc Phyllis B ( ield (MO) a Theseus and Susan Wilson ( ‘cies State Univer completed at St. Louis University and the Missouri Botanical er sae (MO). Diane Bridson (K). Elmar Robbrecht (BR) Amy M McPherson ata, Benoit Satabié ick, nd Macon for computer assis- ielpful suggestions; Ferdinand N Jo hn in n: —Bee ia ) (YS) for illustrations: tance; and aded p n for examining specimens at G on my behalf. | grate ally s ac Meery financial support from the Missouri Botanic a Garden graduate studie Depertment. and Sigma X Research MO, P, and y A for providing work space US MPU, NY, P. n their herbaria : ON publicati ions at my s department and alumni fund, the $ , the Se ienlific Research Society. e United Republic of Cameroon for the partait for field research. Saint Louis University Biology the Ministry of Higher Education and Scientific I er a administrators o : the administrators of B. BM, BR. G. GH, HBG, | than VAG, and YA for e specimens ~ Ei cde used in ds study; the Age odds of all herbaria whose personnel searched for specimens at my reque a, ; librarians at BR, K. and MO who searched for request; and all others who helped in any way ? Department of Biology, Arkansas State University—Beebe, P. 0. Drawer H, Beebe. Arkansas 72012, U.S.A. ANN. MissoURI Bor. GARD. 83: 87-150. 1996. Annals of the Missouri Botanical Garden bacterial nodules. Since then, new taxa have been described and others ape 'ed in synonymy within subgenus Baconi .. Bremekamp, 1953, 1956; Adam, 1973; o 1978). tion remains otherwise unchanged. The defining but its circumscrip- character is the bearded corolla throat; other char- acters listed by A. P. de Candolle, Bremekamp, and others are not universal or are shared with related laxa. Bremekamp (1934: 21) noted that the species of subgenus Baconia were closely related and did not divide it into sections as he did subgenus Pavetta. He divided it into seven series based on number of ovules per locule and branching, nodule, and ves- titure characters. Added collections since then re- veal that, within series (even within species), char- acters upon which series were separated are not constant. Those series are thus not recognized here. Similarly, Bridson (1978) and Bridson and Verd- court (1988 ments of East African species. ) did not recognize the series in treat- METHODS Classical alpha taxonomy methods were used in this study, supplemented by field observations and collections in Cameroon. In addition, certain char- acters (whether more than 10% of calyx lobes over- ap at the base at maturity, whether the corolla throat beard characteristic of subgenus Baconia ex- tends onto the corolla lobes, calyx lobe shape and ength, corolla lobe and tube lengths, style exser- tion length, leaf venation pattern, and leaf venation density) were treated morphometrically. Their av- erages sometimes distinguish species despite over- lapping intraspecific ranges. At least 10 and, when sufficient material was available, 20 or more mea- surements per taxon were made for such characters to determine the average character state. Chi- square analyses were performed to determine whether taxa had more than 10% of their calyx lobes overlapping at the base at the p = .05 con- fidence level. HABIT All of the 29 species of Pavetta subg. Baconia in Cameroon are woody. The majority are understo- ry forest shrubs 1-5 m tall when flowering or fruit- ing. Representatives of seven species have been re- corded on herbarium labels as taller than 5 m. VESTITURE Vestiture, when present, consists of unbranched erect to appressed or less often reflexed trichomes from less than 0.01 to more than 0.5 mm long; tri- chomes toward the short end of this range are more common than longer ones. STEMS The main stems of Pavetta subg. Baconia are monopodial, usually with limited branching. ranches are usually monopodial but occasionally exhibit sympodial growth, e.g., in P. baconiella Bre- mekamp. The branches are often floriferous twiglets with terminal inflorescences. Limited vegetative lat- eral branching also occurs. Vestiture ranges from absent to dense, the density often increasing toward the apex LEAVES As in nearly all Rubiaceae, leaves are simple and opposite with margins entire, although a slight- ly undulate margin sometimes occurs. Leaves are almost always darker green above than below. Leaves vary from membranaceous to coriaceous and are usually chartaceous to subcoriaceous. Anisophylly. In most species, pairs of leaves immediately or one node below inflorescences are sometimes unequal in size; the degree of aniso- phylly varies within species. Anisophylly is never present at all nodes. Leaf blade size ranges de- scribed include anisophyllous leaf pairs. lade size and shape. Leaf blade size and shape are variable within species and overlap be- tween species. Used with caution, these are nev- ertheless useful and convenient taxonomic charac- ters in some species. See Manning (1991) for estimates of reliability of leaf size and other char- acters in Pavetta rigida Hiern (subg. Pavetta). Acumens. Most species have leaves with taper- ing acumens, but some leaves lack acumens in most species. Typical acumens in subgenus Bacon- ia are about 10 mm long and 5 mm across at the base. Apices. Descriptions of apices here include those of nonacuminate leaves and of acuminate leaves below their acumens, as if the directions of the two converging leaf margins below the acumen continued unchanged until they met. In most spe- cies, most leaf apices are acute and some are ob- tuse. Bases. least occasionally attenuate part or all the way to their point of attachment to the stem as seen at 10X. To the naked eye, most bases appear cuneate In all species leaf blade bases are at or subcuneate, some appear attenuate, and there is significant variation within most species. Asym- metrical bases are common in subgenus Baconia. Volume 83, Number 1 1996 Manning 89 Pavetta Subgenus Baconia in Cameroon Venation patterns. | Venation characters are de- scribed in some detail below and included in spe- cies descriptions because they are of diagnostic value, have not been emphasized previously, and are used in keys. They are reported as visible on illuminated, untreated leaves at 10X magnification. — Terminology and definitions follow Hickey (1973 except measures of venation prominence and den- sity, described below. Leaves in subgenus Baconia have brochidodrom- ous to eucamptodromous venation. Intermediates very are ndary veins usually branch and become smaller near leaf blade margins common, 1.e., seco but these branches then join the next most dista secondary vein near the leaf margin. ether leaf venation, especially fourth- and fifth-order veins, is more easily visible above, be- low, or neither when illuminated at 10X is usually characteristic of species and thus of taxonomic val- ue at the species level. Venation prominence. Venation is classified here as prominent, prominulous, intermediate, ob- scure, impressed, or invisible. Prominent and prominulous veins occur only on lower leaf sur- faces. Prominent venation is a term reserved for leaves in which veins protrude from leaf blades most conspicuously to the naked eye. Prominulous means subprominent, though prominulous veins do clearly and significantly protrude below the leaf blade. Vertical protrusion of more than 0.5 mm may qualify a vein to be prominulous; often at least 180° of arc of the vein is visible. Obscure means hard to see when illuminated at 10x. Impressed vena- tion means veins are at a lower level, as seen from above, than the upper leaf surface. The upper leaf surface usually ares downward toward the veins and upward between them. Venation is intermediate in prominence if not described as prominent, promi- nulous, obscure, impressed, or invisible. Prominulous leaf veins are very common, On in- dividual leaves, veins often grade from prominulous near the base to intermediate near the apex. Tertiary veins. ary veins vary from some- times orthogonal eee through random reticu- late to most commonly percurrent, i.e., “tertiaries from the opposite secondaries joining,” sensu Hick- ey (1973), though they sometimes branch. Tertiary veins are usually oblique to the midrib, but occa- sionally subperpendicular to it. Higher order veins. Fourth- or fifth-order veins are reticulate. With rare exceptions, the areoles sensu Hickey (1973) so formed range from 0.2 to 3 mm across; the majority are 0.5 to 2 mm across. Veins forming areoles further branch into veinlets terminating inside areoles with or without yet fur- ther branching. Areolation in most species is im- perfect sensu Hickey (1973), i of irregular shape, more or la wake | in size. *.. forming “meshes s, Venation density. Leaf venation density was de- termined by counting the number of veins or vein- lets crossed per millimeter of straight line distance on blades, then averaging at least 10 such counts per taxon counted. Veins or veinlets crossed in- clude any veins from secondaries to the lowest or- der veinlets seen on herbarium specimens illumi- nated at 10X. Including the midrib or areas very close to margins was avoided when possible. Oth- erwise measurement locations were selected sub- randomly except (1) when only small portions of leaf surfaces revealed higher order venation clearly, such portions were used and (2) measurements were intentionally spread over different specimens when more than one specimen was available. Density data in species descriptions are all based on averages of counts performed as above and range from two veins or veinlets crossed per millimeter in Pavetta grossissima S. D. Manning (very coarse) through three (coarse), four (medium), five (fine), six (very fine), to seven per millimeter (extremely fine) in P. tenuissima S. D. Manning. Density is most often medium to fine. This char- acter is of diagnostic value in species having un- usually coarse or fine leaf venation. Single density measurements are occasionally significantly higher or lower than average for a taxon, so at least several counts must be made on each plant for which this character is used diagnostically. Domatia. Along veins on lower leaf surfaces of most species are hollows, areas of denser vestiture than the rest of the leaf, or both. These occur most often in angles of branch veins along midribs, sometimes along secondaries, and occasionally along tertiaries. The descriptions of these structures here as tuft, pocket, pit, crypt, and intermediate between pit and crypt are sensu Robbrecht (1988). Although some taxa are relatively constant with respect to presence and types of domatia, others are quite variable. Even in species that usually bear domatia, not every branch vein angle along the midrib has a domatium, and whole leaves or specimens may lack domatia. Nodules. In leaves of over 80% of Pavetta spe- cies, including African and other species of sub- genus Pavetta and a similar percentage of subgenus Baconia, are black growths termed bacterial nod- ules (Bremekamp, 1934; Lersten, 1974, 1975; Ler- sten. & Horner, 1976) or leaf galls (Robbrecht, 1988, including illustrations). These vary greatly in shape from punctate or pustuliform to linear or ramified. Punctate and pustuliform nodules are Annals of the Missouri Botanical Garden Figure 1. Awned stipule. The awn is defined to in- clude only the upper 2 mm, analogously to a leaf acumen. usually scattered on blades; linear ones are often along the midrib. Similar nodules are uncommon in other Rubiaceae but are present in 74 African and Madagascan species of Psychotria L. and 12 of 17 species of African Sericanthe E. Robbrecht (Rob- brecht, 1988). In all Pavetta species in which nod- ules occur except P. urophylla Bremekamp, they appear on upper leaf surfaces. In P. urophylla, like Psychotria species, they appear on lower leaf sur- faces In subgenus Baconia, three of Bremekamp's sev- en series were separated on nodule characters; nod- ules were variously distributed in his other series. Material observed since Bremekamp's monograph reveals that nodule characters are not consistent enough to assign them as much diagnostic impor- tance as he did. Nodules have been found in all but 2 (Pavetta rubentifolia S. D. Manning and P. molundensis K. Krause) of the 29 species of sub- genus Baconia in Cameroon. In nearly half of the species, however, they are sometimes or usually ab- sent. Within specimens, some leaves may reveal a few nodules and others none. In some species, some leaves have profusions of conspicuous nod- ules while others do not. In most species, most nod- ules are subpunctate and scattered. In others they tend to be mostly linear and associated with major veins. Stipules. Single-awned, entire-margined, oppo- site interpetiolar stipules whose awns are decussate with the leaf pair at the same node are universal in subgenus Baconia. Bases of opposite stipules are connate just above petiole bases. Stipules thus sheathe the stem at each node. Stipule awns were measured analogously to leaf acumens, as in Figure in which the awn is cuspidate and 2 mm long. Care must be taken in using awn length diagnos- tically in subgenus Baconia; the tips often break off. In some species stipules are also partly decid- uous below their awns. INFLORESCENCES Position. Inflorescences in subgenus Baconia are almost always terminal on floriferous twiglets, which can be up to 37 cm long. Floriferous twiglets are absent at some nodes, rendering inflorescences at those nodes axillary. Such axillary inflorescences sometimes occur along with terminal ones in col- lections of Pavetta camerounensis S. D. Manning, P. grossissima, and P. tenuissima made since the second edition of the Flora of West Tropical Africa (Hepper & Keay, 1963) and are exceptions to the key to genera of Rubiaceae therein (p. 105), which distinguishes between genera with terminal and ax- illary inflorescences. At nodes bearing axillary in- florescences, there are typically two oppositely ar- ranged inflorescences. In no species of subgenus Baconia are all inflorescences axillary. Inflorescences are most often solitary but, in some species, some inflorescences are clustered at termini of floriferous twiglets. When inflorescences are so clustered, those on side branches sometimes overtop the central one. Bremekamp (1934) used such overtopping and floriferous twiglet internode lengths to separate three series of subgenus Bacon- ia from three others. These characters vary so muc within taxa that they are not used diagnostically ere, Architecture. Inflorescences, shape, are dichasia or modified dichasia which are nearly always compound. Although flowers are usu- ally close enough to each other for corolla lobes of adjacent flowers to overlap, the main branches of inflorescences are sometimes far enough apart that these inflorescences are described as having sub- whatever their units. Dimensions and flower number. With rare ex- ceptions, inflorescences are 20 cm or less across. Their sizes and degrees of congestion, though vari- able within species, are important distinguishing characters between some species. There are rarely only 1-5 to sometimes more than 400 flowers per inflorescence in subgenus Ba- conia. Flower number is often not a good diagnostic character because within species maximum number of flowers per inflorescence typically is about 10 times the minimum number. Phenology. Flower opening within an inflores- cence usually occurs subsynchronously (flowers to- ward the center of inflorescences tending to open slightly earlier) after an extended period in full- sized bud. Volume 83, Number 1 96 Mannin 91 g Pavetta Subgenus Baconia in Cameroon 1 D Inflorescence Branches Stipular Awn Foliar Appendage Body of Sheathing Bract Figure 2. Sheathing bract at base of an inflorescence. wing a normal stipular awn and foliar appendages. The b look like re duced leaf blades from above or below. Fruits on an infructescence also usually develop and mature subsynchronously. Sometimes only one or a few fruits develop: often many do. Complete fruits persist on infructescences after ripening. racts, peduncles, and inflorescence delimita- tion. These features are treated below in some de- tail either because previous treatments have not done so (bracts) or to clarify or simplify previous treatments (peduncles, inflorescence delimitation). These features have diagnostic value in some spe- cies and are used (bracts) or necessary to interpret characters used (peduncles, inflorescence delimi- tation) in the keys to species. racts. Bracts as described here are of four types: sheathing. foliar, other, and bracteoles. a. Sheathing bracts. Sheathing bracts are op- posite and connate, thus sheathing the axis that bears them. They are present at inflorescence bases and usually also at bases of higher inflorescence branches. They are presumably homologous to stip- ules, often look like them, and usually bear cus- pidate to linear awns in positions one would expect awns on stipules. They sometimes bear additional awns or fimbriae at right angles to the “stipular” position and occasionally in other positions. Num- bers and positions of additional bract awns some- times vary on individual plants. Connation of op- posite bracts is often so pronounced that the sheath formed assumes the shape of a cup or, if spreading, of a saucer. At times, sheathing bracts bear foliar appendages. In such bracts the sheathing tissue is adnate to the foliaceous tissue, and the foliar ap- pendages are decussate to the “stipular” position. Such sheathing bracts also bear awns or fimbriae in the normal “stipular” position, and are homolo- gous to a leaf pair and stipule pair. Figure 2 illus- trates such a bract. b. Foliar bracts. Foliar bracts are only sporad- ically present in subgenus Baconia. They are pet- iolate, unlike foliar appendages on sheathing bracts. Foliar bracts resemble reduced foliage leaves in position and shape, although they are sometimes more reduced in length, proportionately, than in width. They sometimes occur on the main inflorescence axis and sometimes on side branches of the inflorescence subtending flower clusters. Dis- tinctions between foliage leaves, foliar bracts, stip- ules, and sheathing bracts are illustrated in Figure 3. ther bracts. These are usually more distal in inflorescences than sheathing bracts. They usu- ally range from wedge-shaped to ovate, obovate, « linear and often bear 1—several fimbriae or a tuft of hairs at or near their apices. They can vary con- siderably in shape within specimens. The largest are distinguishable from sheathing bracts only in that they do not quite sheathe the inflorescence These usually subtend upper inflorescence along side = "1 axis. Smaller ones are usually The most reduced ones are fimbriae or branches. branches. awnlike structures on inflorescence branches. d. Bracteoles. These include all bracts on in- dividual pedicels. They are small and resemble the smaller of the “other bracts” described above. They often take the form of fimbriae as in Figure 4. Their positions on their axes vary. Figure 4 depicts typ- ical variation in nonfoliar bracts and bracteoles. Little and Jones’s (1980) definition relevant part, is employed here: soon The base of the Peduncles. of peduncle, “the stalk of an inflorescence. peduncle is at the node of the most apical pair of vegetative leaves and the top is at the point of at- tachment of the lowermost inflorescence branch. If there is no internode between the most apical veg- leaves and the lowermost inflorescence branch, the inflorescence is sessile. Bremekamp’s (1934) definition and discussion of peduncles are not followed here because they are more complicated and confusing (Manning, 1990). Inflorescence delimitation. Whether a branch floriferous twiglet just below the flower cluster ter- minating a major floriferous twiglet is interpreted etative — as bearing a separate inflorescence or as a branch of whether the branch has leaves subtending the flow- er cluster. If so, i If not, it is a branch of a single terminal inflores- cence. Subtending “leaves i liar bracts and the branch a part of a single ter- are reduced a single terminal inflorescence depends on bears a separate inflorescence. 4 may be considered fo- minal inflorescence if the “leaves” more than half in length compared to mature leaves on the main floriferous twiglet just below the inflo- rescence, or if they are no larger than any foliar Annals of the Missouri Botanical Garden Inflorescence ranches Stipule Figure 3. bracts on the main floriferous twiglet rachis, or if they have fallen off and left scars clearly smaller than leaf scars. The above distinction between solitary and sep- arate inflorescences is only occasionally difficult in practice, FLOWERS Flowers are almost always 4-merous in all spe- cies, though occasional 5-merous flowers occur. calyx. Calyx tubes in subgenus Baconia are often 1-2 mm long. Calyx lobe shape and size are important distinguishing characters, despite varia- Foliage Leaves Distinctions between foliage leaves, foliar bracts, stipules, and sheathing bracts in Pavetta subg. Baconia. tion in most species. Calyx lobes are up to 4 mm long. Typical shapes are illustrated in Figure 5. Al- though one or two of these shapes predominate in most species, several shapes routinely occur as less common variants. Calyx lobes in subgenus Baconia are usually val- vate in open flowers, but often overlapping at the base in bud. Some (never all) lobes remain slightly overlapping at the base in mature flowers in several species. This is used diagnostically for the few spe- cies in which it occurs in more than 10% of mature flower calyx lobes. Corolla. In Cameroon, corollas are usually Volume 83, Number 1 1996 Manning Pavetta Subgenus Baconia in Cameroon Bracts Sheathing Bract Figure 4. Variation in nonfoliar bracts and bracteoles in Pavetta subg. Baconia, showing fimbriae. Bracteole white. less often greenish white, cream-white, the upper surface of the corolla lobes or down into cream, pale green, greenish yellow. yellow, brick the upper part of the inside of the tube. Rarely, red, or with green tips, margins, or stripes. small amounts of shorter vestiture, sharply delim- Corolla exteriors are glabrous; interiors are usu- ited from the corolla throat beard, also occur on ally glabrous or subglabrous except the beard of upper surfaces of corolla lobes. In some species, hairs around the corolla throat, attached within the the beard projects outward from the corolla throat. tube near the top. In a few species the beard nor- Members of subgenera Dizygoón and Pavetta mally extends unbroken onto the proximal part of sometimes bear arachnoid vestiture inside upper 94 Annals of the Missouri Botanical Garden 2-LOBULATE COMPRESSED : COMPRESSED ROTUND C SHORT TRIANGULAR D DENTICULATE | Figure 5. parts of corolla tubes. Such vestiture is distinguish- able from that of subgenus Baconia in at least one of the following three ways: (1) it extends further down into the tube and does not form a distinct TRUNCATE Calyx lobe shapes in Pavetta subg. Baconia. OBLONG A VE ring, (2) it is less dense, or (3) the hairs are finer, shorter, or more curved than the typically bristly hairs of subgenus Baconia corolla throats. Corolla throat beards similar to those typical of subgenus Volume 83, Number 1 Manning 95 Pavetta Subgenus Baconia in Cameroon Baconia also occur in some other Rubiaceae, such as the African Morelia senegalensis A. Rich. ex DC. and the South American /sertia coccinea Vahl. Corolla tubes vary from cylindrical to twice or more as wide at the top as at the base. A constric- tion of the tube often occurs just above its base marking a position just above the epigynous nectary dise. Corolla tube length can often be used in dis- tinguishing species despite intraspecific variation. 1 ~ Corolla tubes are normally as short as 2 mm some species and as long as 10 mm in others. Av- erage corolla tube length in subgenus Baconia is about 5 mm, shorter than in subgenus Pavetta. The range of variation between longest and shortest tubes in open flowers of most species is about 2 mm. Intraspecific variation seen in this character ranges up to 7 mm; in well collected species whose ranges of variation are probably best represented, longest tubes of open flowers are about twice as long as shortest ones. Corolla tubes are often 1-2 mm wide. Androecium a. Stamens. Anthers are introrse, subbasifixed. usually sublinear, and dehisce along longitudinal slits, usually when flowers are in bud. The four sta- mens are attached to corolla tubes near their throats, alternating with the corolla lobes. When flowers open, anthers become exserted and usually hang down between corolla lobes, sometimes coil- ing when dry. b. Pollen. subglobose, and typically from 10 to 25 qm in di- Pollen is tricolporate, globose to ameter. This is usual in Rubiaceae and tribe Pav- etteae (Robbrecht, 1988). Although pollen is usu- ally in monads, tetrads were observed in Pavetta staudtii Hutch. & Dalziel (Zenker 4913, MO). Gynoecium Ovary. The inferior ovaries vary from gla- brous to pubescent. Vestiture on ovaries is similar to that of inflorescences bearing them until it thins as fruits increase in size. Inside the base of the corolla tube is an epigynous, annular or subannular nectary disc. Structural details are discussed below under “Fruits and Seeds.” Each ovary bears a The swollen part is the pollen presenter. Style-pollen b. Style-pollen presenter. single, apical style swollen toward its apex. presenters range from most often clavate to less of- ten fusiform, and from glabrous to pubescent, some- times within species. Hairs on pollen presenters are usually suberect, borne nearly perpendicularly from the surface but curved near the apex, and sometimes as long as 0.5 mm. Some species’ pollen presenters have clear vertical ridges. The part of the style below the pollen presenter is normally glabrous in almost all species. Distances that styles are exserted beyond corolla tubes vary considerably and can be diagnostic at the species level. In some species, styles are ex- serted over 15 mm: in others, only 3 mm: in most. intermediate distances. c. Stigma. Stigmas in most species are slightly bifurcated at the tip. They never become recurved as in /xora, however. FRUITS AND SEEDS Fruits. bose drupes. A shallow vertical depression around Fruits in subgenus Baconia are subglo- the outside often marks the fruit's division into two locules. Average fruit size ranges from about 5 mm diameter in some species to about 10 mm in others. Fruits are most often. white, off-white, or glaucous, but in some species are orange, orange- dull pink. black, bluish black, or white with green ver- green, brown-green, yellow-green, yellow, tical stripes: red ones have not been reported. Fruits in Pavetta are divided into two locules by a vertical septum. In almost all species, each locule bears a single ovule from near the midpoint of the septum. Often, both ovules develop into seeds (Fig. 6A). However, sometimes only one of the ovules develops into a seed. This has been observed i over half of species from Cameroon for which fruits are known. Absence of a shallow vertical depres- sion around the fruit exterior often indicates that this has happened. In most such fruits, the septum is close to or against one side of the fruit, resulting ina vestigial locule on the other side of the septum from the developed seed (Fig. 6B). In Pavetta camerounensis, although most fruits have only one ovule per locule as is typical of sub- genus Baconia, one fruit was seen in which one locule has two seeds borne side by side from a single fleshy placenta whose attachment is near the middle of the septum, while there is only one seed in the other locule (Fig. 6C). This situation occurs more often in P. /asioclada (K. Krause) Mildbraed ex Bremekamp. in which it is also common for each Fig. 6D). been observed to be attached — of the two locules to bear two seeds Placentae have more than halfway up the septum, but never api- cally, in a few species. Seeds. of their locules. They are nearly hemispherical, ex- Mature seeds occupy most of the volume cept the adaxial side is concave at maturity, with the concavity situated where the placenta was for- attached. merly The concavity is typically sur- rounded by a ringlike outgrowth of testa tissue. Annals of the 96 Missouri Botanical Garden '(pe^oura sejuaoe[d) s[noo[ qoea ur spass oM], "(YH ‘ISZ «»znojv]) DPD]9015D] 4 JO UNAY jo uonoes astaAsUBI] *(]J— (paAoura1 sejuaoe[d) a[noo] 19ujo ay) ur paas 1oze[ ouo *o[noo| auo UL spaas neus OM], (VAN “108€ P 12 PNM op) sisuauno42urpo “dsqns sisuauno42um2 J JO M pord ue wo spaas jo SMAIA perxepy 7)— (4 "grr0[ [DUADY X ¡puÁDy) pedojaaap pəəs auo quo ‘punssmua) 4 Jo nij jo uoroas ssor [euorenbz] ^qg— (ON “SEQ! Furuunpy) pdipsoimau Yq jo may jo uonoəs ssor [eUOyenby y —"DIUOIDG “3qNS ngoai jo spess pue sy “9 analy HEM 144 wnidas EN ) ODD NM unidas N WN A NN HEM yng pass pedoje^eq unideg spoog HEM 11n44 unidas Volume 83, Number 1 Manning 97 1996 Pavetta Subgenus Baconia in Cameroon Seeds that are less than full sized also occur fairly The species with the highest altitudinal range. often in full-sized fruits, however. Rarely, parthe- Pavetta hookeriana Hiern. has been collected nocarpy occurs. flower during more months (August and December The seeds in four-seeded fruits of Pavetta lasio- as well as February through June) than any other clada (Fig. 6D) are similar to others in the subge- species. It is the most widespread montane species. nus in basic structure, although constrained to fit Mature fruits of ten species of subgenus Baconia. within fruits very similar to the two-seeded ones in have been collected in Cameroon during November, other species of the subgenus. and those of nine species in December. This peak period for ripe fruits is in the dry season in most HABITAT parts of the country. In contrast, during the wet More species in subgenus Baconia have been month of August, only one species of subgenus Ba- found in wet forest than in drier forested areas. anq conta has been collected in mature fruit in Came- more occur in drier forested areas than in savanna, roon. Intermediate numbers of species have been Even in savanna areas, collections are often from collected in mature fruit during other months. The gallery forest rather than open savanna. spread of fruiting seasons throughout the year is in spite of fewer total species-months being repre- REPRODUCTIVE STRATEGY sented in fruiting collections than in flowering col- ; lections and no single species having been collect- Flowers are protandrous. The introrse anthers oe ea ee l x ed in ripe fruit in more than six months of the year. usually shed pollen onto the swollen upper part of i the style (pollen presenter) while the flower is ii bud. The pollen presenter becomes exserted as the |, m : i PAXONOMIC TREATMENT corolla opens. Pollen adheres to it and is ready foi insects to transfer it to receptive stigmas. This pol- T . | B , hologicall rae . i Mie axa in subgenus Baconia are morphologically lination syndrome is typical of tribe Pavetteae " I agn (Robbrecht, 1984, 1988) and subfamily Ixoroideae (Bremekamp, 1966; Robbrecht, 1988). March is the peak flowering month for subgenus relatively similar and are difficult to separate. Sub- genus Baconia taxa outside Cameroon show similar amounts and patterns of morphological variation to Baconia species in Cameroon. Sixteen of 29 spe- those in Cameroon. Since species of subgenus Ba- cies, and 20 of 38 taxa, including infraspecific taxa, have been collected in open flower in March. This conta are often similar in many characters, one might ask why the whole group or much of it should not be considered a “species complex” as found, for example. by Seyani (1988) in Dombeya burges- siae Gerr. ex Harv. € Sond. (Sterculiaceae). This is peak flowering is near the beginning of the main rainy season. There have been no flowering collec- tions during September and October, toward the end of the rainy season in much of the forested not done because good correlation of morphology area, and few during July and August. This cannot With geographical location has been found in sub- completely be accounted for by infrequency of col- genus Baconia, while in the D. burgessiae complex lecting efforts during this wet period; fruits have there was poor correlation of morphology with ge- been collected during every month of the year. ography. KEY TO SPECIES OF PAVETTA SUBG. BACONIA IN CAMEROON BASED ON CHARACTERS WHEN FLOWERING l Bacterial nodules on leaves more conspicuous below than above; corollas brick red to beige tan — AI 28. P. urophylla l Bacterial nodules on leaves more conspicuous above than below, or absent: corollas white, cream, green or yellow A EA AA 2(1). les 2 in. both Iosgles o ovary or 2 in one locule, 1 in the other: at least some leaf E pu veins sid or [rie rnm below . P. lasioclada 2. vule | per locule; some leaf tertiary veins prominent or agar al below or not... 0 3 3(2). Some leaf tertiary veins connecting adjacent pui iry veins to form a subrec ‘tangular gridlike pattern. conspic uous below to the naked eye (Figs. 18. 19, 23). leaves a numerous conspicuous light-colored iary veins borne from midrib redi ilarl 3 Leaf tertiary veins not forming a subrec Mei: gridlike pattern with secondary veins conspic deer naked eve unless tertiary venation also includes numerous conspicuous light-colored veins borne from midrib perpendicularly 6 4(3). Some nonfoliar br acts fimbriate as in Figure 4; inflorescences ca. 6— across; adaxial surfaces of ‘orolla lobes not pubescent near throat; corolla tubes 5-7 mm fana al pal s 6-7 mm long 51 P. namatae Annals of the Missouri Botanical Garden 4. Most or all nonfoliar bracts lacking fimbriae; inflorescences ca. 0.5—4 cm across; adaxial surfaces of some or all corolla lobes pubescent near throat; corolla tubes 2-5 mm long, corolla lobes 2-6 mm long E 5(4). Inflorescences + lax, not subumbellate, (1—)2—4 cm across; average ca. 5—6 veins or veinlets d per mm of leaf blade; longest leaf blades <18 cm; corolla tubes 3-5 mm long, corolla lobes 4-6 mm long. sse ele E AR 20. P. muiriana 5. Inflorescences - congested, subumbellate or with subumbellate subunits, 052 « across; average ca. veins or veinlets crossed per mm of leaf blade; longest leaf blades =19 cm; xls tubes 2-3 mm long, corolla lobes 2—4 mm long... cnn none 27. P. tenuissima 6(3). Inflorescences glabrous throughout (or rarely partly subglabrous); nonfoliar bracts almost all lacking fimbriae |... 7 6 At part of inflorescences puberulent or pubescent or, if inflorescences glabrous or subglabrous throughout, nonfoliar bracts fimbriate as in Figure 4 00 7(6). Sheathing bracts covering inflorescence branches for most of their lengths (Fi ig. 13). most sheathing bracts drying light green; inflorescences ici subumbellate or of tightly clustered subumbellat diia cc —c——— ————— —K————— bocasess 12. P. EEUU 7. Sheathing bracts not e mbites inflorescence branches for most of their — most sheathing bracts dry black or brown; inflorescences congested and subumbellate or not . .22. P. i 8(6). Leaf blades pubescent to ASA above (at least a few hairs scattered over the whole upper vera A PERPE IERI 8. t least part of some or all leaf blades on each : specimen n glabrous above AAN eco UN 11 9(8) Sa and higher order leaf venation above not obscure when illuminated at 10x 16. P. longibrac hiata 9. Fourth and higher order leaf venation above obscure when illuminated at 10X |... 10 10(9). Flowers with corolla lobes longer than corolla tubes; most calyx lobes ovate (Fig. 5), thei ir tips pointed or Aur) so; largest leaf blades (11.5-)14.5-22 X 4.5-9 cm; styles exserted 9-11 mm; floriferous twiglets 19.5 em long poe . 19. P. mpomii 10. vue flowers with corolla lobes shorter than corolla tubes; most c alyx lobes sub g. 5) to obo vate, their tips not pointed; largest leaf blades (16.5—)21—34 X 8.5-15.5 em; styles fie 12- 18 dar flor erous twiglets (6—)18—29 em long. D f D viridiloba 11(8) Anthers septate 00000000000000000000000000000 0 A — —— o 7. P. cellulosa 11. Anthers not septate MMM A RA P 12(11). Leaf apices aise to rounded, fol ac n most cuf blades oblanceolate « or r strongly obovate: corolla lobe average length 10-13 mm; average style exsertion 17-23 mm 0000000 13 12. Some or most leaf apices acute, most leaf blades elliptic Ay or Pos shapes but most neither oblanc eolate nor strongly obovate; corolla lobe average length <9 mm; average style exsertion <12 mm l 13(12). Most calyx lobes e iiie to compressed rotund (Fig. 5), 0.1-1 mm ionis corolla lobe — e length 10 mm, range 8-13; average style exsertion 17 mm, range 11-20 P. kupensis 13. Most calyx Pg deltoid to rotund or pentagonal, not compressed (Fi ig. 5), L 53 1 mm n long; ES lobe average length 13 mm, range (9-)12-14; average style exsertion 23 mm, range 20-30 re aa 14(12). e least some upper leaf blade surfaces on each specimen appearing dotted when eel at 10X wing to minute dd 15 14. Upper leaf blade surfaces not appearing dotted or papillate when illuminated at t 10x 17 15(14). Corolla tubes (2-)3-8 mm long, corolla lobes (3-)5-9 mm long, styles exserted (2-)6-10 mm... A E AAA P. longibrachiata 15. Corolla tubes 1-3 mm long, corolla lobes 2-41 mm ui styles exserted 24 mm 0. 16(15). Inflorescences congested with ca. 40 flowers per cm of inflorescence width; average irs p len L0 AAA > brac a 16. Inflorescences lax with c a. 10 flowers per cm ‘of inflorescence width; average corolla a length 3 A dd Ae ls ERN H P. 1 17(14). A few hairs scattered over the whole upper leaf blade s surface -- 16. P. longibrachiata 17. At least part of some or all leaf dep on each specimen mites above .. 18 18(17). Most calyx lobes subquadrate (Fig. 5)... 2 000o000r0 000er eono o oeren — 19 18. Most calyx lobes of shapes other han subquadrate (Fig. 5 JA AS 21 19(18). Calyx lobes puberulent externally; nodules, if present, at least as conspicuous on ı midrib « as elsewhere ET E A ON 5. P. calothyrsa 19, Calyx lobes glabrous or subglabrous externally; nodules, if present, more conspicuous lean than along midrib.. 20 20(19). Under 10% of calyx lobe bases overlapping after flowers open, fourth and higher order venation ‘obec cure pilin on most leaves when illuminated at 10X, inflorescences 1—6(—8) cm across with (10—)20—100(— 200) ean A RR AMOR IU IR NA 26. P. varii 20. Over 10% of calyx lobe bases overlapping after flowers open, fourth and higher order venation not obscu above when Vn nado xcept on some of the e leaves, inflorescences (4—)7-14 cm across with (75—)100—400 flower: ANE A ET EEA ET a 24. P. robusta eh Average calyx lobe qiii €] mm. ttd 22 verage calyx lobe length >1 mm. 28 m Average ca. 2 veins or veinlets cr esed p per mm of leaf blade a as seen n illuminated | alüx |. 10. P, grossissima Average (3—)4—6 veins or veinlets crossed per mm of leaf blade as seen illuminated at 10x |... 23 23(22). Longest leaf blades >10 cm including acumen; widest inflorescences always =1 cm across |... e 24 Volume 83, Number 1 Mannin 99 1996 g Pavetta Subgenus Baconia in Cameroon 23. 24(23). 24. 25(24). 25. 26(25). 26. 21(25). 27. 28(2 1). 29(28). 20. 30(28). 30. 3 1(30). 3d. 32(30). 32. ey 32). as 33). 34. »x 32). wa 35). Key TO (CALYX Longest leaf blades =10 cm including acumen; widest inflorescences not always =1 em ee: "p PERCORSO t P. baconiella Somo -inflorescences subcapitate or subumbellate; some or all leaves with >10 pairs of sec enis € A A » P. camerounensis Inflorescences not subcapitate; if subumbellate (P “Zaboni a). all leaves with =10 pairs of UD ve 25 Most calyx lobes subdeltoid to subrotund, not compressed (Fig. 5), their average length 0.75 mm, a occasionally as short as 0.5 mm = Most calyx lobes compressed (Fig. 5), their average length <0.5 mm, only « occasionally as long as 75 mm Longest stipule awns =5 mm long |. 020. . oe P. bidentata Longest stipule awns <5 mm long 2 P. gabonica Leaves maroon, stipules glabrous or subglabrous internally, inflorescences (5-15 em across Beka geese eS saa Das aces scape O AN AO 25. P. rubentifolia Leaves green, stipules pubescent internally, inflorescences 1-5-8) cm across — 3. P. brachycalyx Average corolla tube length 3—4 mm, average style exsertion 6 mm 20 Average corolla tube length =5 mm, average style exsertion 28.5 mm 30 Either most fourth and higher order leaf venation subobscure below when illuminated at 10X or veins intermittently pubescent below as in Figure 12; leaves coriaceous to subcoriaceous |. 11. P. hookeriana Fourth and higher order leaf venation conspicuous below when illuminated at 10X, veins not intermittently pubescent below as in Figure 12; leaves chartaceous to subcoriaceous lo. Pl sp sun TN Over 10% of calyx lobe bases overlapping after flowers en iso calyx lobe length 1.7-2 mm 3l Under 10% of c alyx lobe bases ove rlapping after flowers open, average calyx lobe length 1—1.5 mm 32 Average style exsertion 8.5 mm, range (5-)8-9(-12): nodules EE veins glabrous below ——— ——— MC A ene 18. P. molundensis Average style exsertion 11 mm, range 9-14; nodules present; veins subglabrous to pubescent > low ee eae es A A A E A FOE —— a A corymbosa Fourth and higher order leaf venation cle arly visible above when illuminated at 10x Fourth and e order leaf venation not ER isible above when illuminated at 10x . 35 Domatia absent along secondary veins below 23. P. owarlensis Vido pen ne peered veins palin = jus ‘a. 3 veins or veinlets crossed per mm of leaf blade. 23. P. owariensis erage ca. 4—5 veins or veinlets crossed per mm of leaf blade lo. P. juna ae and higher order leaf venation obscure or invisible below when illuminated s 10x 223. P. owariensis Fourth and higher order leaf venation clearly visible below when illuminated at LOX —_ 30 Domatia absent along secondary veins below Ls 23. P owartensis Domatia present along secondary veins below . 16. P. longibrachiata SPECIES OF PAVETTA SUBG. BACONIA IN CAMEROON BASED ON CHARACTERS WHEN FRUITING CHARACTERS INCLUDED) Bacterial nodules on leaves more conspicuous below than above Si M. P takir Bacterial nodules on leaves more conspic uous sag than below. or absen Seeds 2 in both loc ules of fruit or 2 in one locule, I in the other: at least some leaf tertiary veins ear nt or prominulous below —— 14 P. lastoc lada Seed 1 per locule; leaf tertiary veins prominent or not below 0000000 s Some leaf tertiary veins connecting adjacent n veins to form a subrectangular gridlike pattern conspicuous be low to the naked eye (Figs. 18, 19, 23). leaves without numerous conspicuous ad colored tertiary veins borne from midrib perpendicularly 4 Leaf tertiary veins not forming a subrectangular gridlike pattern with secondary veins conspic nous to tbe: naked eye unless tertiary venation also includes numerous conspicuous light-colored veins borne from midrib perpendic ularly ) Some nonfoliar bracts ‘fiabriate as in Figure 4; infructescences ca. 6-11 em across 21. P. namatae pen — I ls lac ing fimbriae; infructescences ca. 0.5— across 5 Aver. ca. 5-6 veins or veinlets crossed per 1 of le d bl: dm infrue ‘ttescences not subumbe qa o congested (1-)2— 4 cm across; longest leaf bl: Ache ss un 20. P muiriana Average ca. 7 veins or veinlets crossed per mm [leaf blade: infruc tescences subumbe i ate or with bus allato subas, congested, 0.5-2 cm across: Yi st leaf blades =19 cm . P. tenuissima Infr uctescences glabrous Tyagi (or rarely partly hak hana nonfoliar bracts ea all lack- = => s TA leas of infructescences p pes nt or D scent or, if infructescences glabrous or subglabrous throughout, nonfoliar bracts fimbriate Fig Sheathing bracts covering infructescence b n s for most of their lengths (Fig. 13), most atesti- ing bracts drying light green; infructescences congested, subumbellate or of tightly clustered subumbe T ate subunits 12. > kribie nsis Sheathing bracts not covering infructescence branches for most of their lengths, most she athing p» acts drying black or brown: infructescences congested and subumbellate or not 22. P. neurocarpa 100 Annals of the Missouri Botanical Garden 8(6). Leaf blades pon ent to subglabrous above (at least a few hairs scattered over the whole upper race) o eee A A eT SPEO ES 8. At least part of some or all leaf blades on each s specimen glabrous above s 9(8). Fourth and higher order leaf venation not obscure above when illuminated at 10X 16. P. longibrachiata 9. Fourth and higher order leaf venation obscure above when illuminated at 10X |... 10 10(9). Most calyx lobes ovate (Fig. 5), their tips pointed or nearly so; largest leaf blades (11. 5-)14. a x 4.5 9 cm; floriferous twiglets 5.5-19.5 em long 0 .P T 10. Most calyx lobes subquadrate (Fig. 5) to obovate, their tips not ‘pointed; largest leaf blades (16. ELM x 8. 5-155 5 cm; floriferous twiglets (6-)18-29 cm long 00000000000 29. P. viridiloba 11(8). Leaf apices obtuse to rounded, not acute; most leaf blades oblanceolate or strongly obovate 12 ll. Some or most - apices acute; most leaf blades elliptical or other shapes but neither oblanc eolate nor strongly obovate 00000000 12(11). Most calyx lobes short triangular to compressed rotund (Fig. 5), 0.1-1 mm long; domatia present along midrib y sometimes in external angles of two connecting secondary veins of at least larger leaves below; infructescences 2-9 em across... tette 13. P. kupensis 12. Most t calyx lobes deltoid | i rotund o or r pentagonal, not compressed (Fig. 5), 1.5-3 mm long; domatia pyle infructescences 9-13 em across s eee P. longistyla 13(11). Mature ruits crowned by uice calyces most lobes of which are >2 mm long o DETUR 14 3. ature fruits not crowned by persistent calyces, or persistent calyx lobes <2 mm long PEN 5 14(13). Most calyx lobes subquadrate AAA A SEDES 5. P. calothyrsa ; Most calyx lobes n aries but rather ovate, subrotund or rotund (Fig. 5) . —.. 23. P. owariensis 15(13). Longest stipule awns 75 mm long; leaves with nodules conspicuous along midrib 2. P. bidentata 15 Longest stipule awns <5 m n long; xar with a conspicuous ii midrib or ziol--üsually not except in P. hookeriana, whe hi is variable in this respect. 16(15). At least some upper leaf blade surfaces on each specimen appearing , dotted when illuminated at 10x, owing to minute papillae 00000000 7 16. Upper leaf blade surfaces ; not a ewan: dotted or papillate when illuminated at 10X = 17(16). Infructescences sine cm across; ind lobes longer than wide, their average length c ca. 1.0 mm. ? brachysiphon 17. Infructescences not c ongested, ca. 0. 5-9 cm ac ross; width of calyx lobes s subequal to length, their average length aa ee ene neon 18(17). Infructescences > average length of calyx debes 1.3 Whe aia 15. , : Infructescences neither lax nor congested, average length of calyx lobes 1.1 mm .. 16. P. longibrachiata 19(16). A few hairs scattered over the whole upper leaf blade surface... lo. P. uber 9. At least part of some or all leaf blades on each specimen glabrous above . 20 20(19). Average ca. 2 veins or veinlets crossed per mm of leaf blade surface as seen illuminated at 10X, mature (AAA 10. P iniu 20. Average (3—)4—6 veins or veinlets crossed | per mm of leaf blade surface as seen illuminated at 10X, mature fruits iii i or ilis Sg aper other colors (black, blue, gray, or whitish) in species for which mature fruits are known e in P. gabonica, in which they are orange, yellow, or brownish 1. - 21(20). Leaves maroon, le glabrous or subglabrous internally, infructescences ca. 0. 5-1.5 5 CM across .... PNE IIET E EE NE A A T E E E 25. P rubensifoli 21. Leaves green, stipules glabrous or subglabrous or not internally, ‘infructescences as narrow as 0.5-1.5 A ERR RR IURE AAA RR AA CORONEL E E AT 22(21). Some infructescences sube apitate | or - subumbellate; some se e larger than 10 X 2.5 cm includin acumen and some or all leaves with >10 pairs of secondary veins... . P. camerounensis 22. Infructescences not subcapitate, in most species not tubae ate either; if subumbellate, ‘eithe er largest 25(24). 25. 26(24). = blades no larger than 10 X 2.5 em including acumen or all leaf blades with — 10 pairs of secondary eins 2). Fondi. and higher order leaf venation not obscure above when illuminated at 10x, or obscure above only on some of the largest leaves 000 Fourth = higher order leaf venation obscure above when illuminated at 10X 0 34 . Average ca. 3 veins or veinlets c dee per mm of leaf blade, fourth and ded order leaf venation easily pea below when illuminated at 10X 0000000000 2: Average ca. 4-6 veins or pia 'rossed per mm of leaf blade, fourth and higher order leaf venation easily “Visible below or not dio illuminated at LOX |... v = ) Leaf blades not larger than 10 X 2.5 em inc luding acumen ... — DNE . P. baconiella Largest leaf blades larger d 10 x 2.5 cm including acumen E P. owariensis Mature fruits orange, yellow, or brownish if known; hd and higher order leaf venation only moderately sily visible below when illuminated at 10X —. 0 27 € Mature fruits gray, blue-gray, blue-green, whitish green, or blackish if known; fourth and higher order leaf venation easily visible or cisma easily visible below when illuminated at 10X . Hairy pit or pocket domatia up to several mm = i secondary veins of most leaves 7. P. cellulosa 28 Danana not along secondary veins of most leaves... . Most calyx lobes deltoid or subdeltoid (Fig. 5), he ir average length 0.75 mm, only oc casionally . as short as 0.5 mm; mature fruits orange, dull yellow, or brownish; plants widely distributed but so far ef found on Mt. Cameroon 0000000 n - . P. gabonica Volume 83, Number 1 Manning 101 1996 Pavetta Subgenus Baconia in Cameroon 28. lai calyx lobes more compressed (Fig. 5) than deltoid, their ave i length $0.5 mm, only occasionally s long as 0.75 mm; fruit color unknown; plants so far found only on Mt. Cameroon 3. P. brachycalyx 29(20). Fun "id higher order leaf venation zi visible below when mae at 10X 30 29 rth and higher order leaf venation visible but only moderately so or subobscure below when illuminated U e a aR aa br 30(29). Domatia present along secondary veins of most leaves below 31 30. Domatia absent or only occasionally present alc ong leaf secondary veins below 32 31(30). Fourth and higher order leaf venation more conspicuous Kien than below . 7. P. cellulosa 3 Fourth and higher order leaf venation not more conspicuous above than below... .P longibrac M 32(30). Most calyx lobes oe (Fig. 5), nodules scattered on leaf blades. fruits unkno 24. P robusta inn Mis various, most commonly rotund, most not subquadrate: nodules dE mature fruits blue- gray or whitish 18. P. molundensis 33(29). Nodules se is) on leaf blades en 7. P. cellulosa 33. Nodules absent or mostly along midrib 11. P. hookeriana 34(23). Fourth and higher order leaf venation easily visible below when illuminated at 10X 35 34. Fourth and higher order leaf venation not easily visible below when illuminated at 10X 37 35(34). Domatia present on secondary veins below as pubescent pits, pockets, or crypts, or veins below puberulent to pubescent a 3. P. corymbosa 35. Domatia not present on secondary veins below as pubesc ent pits, pockets, or crypts, veins below glabrous or eub A acess A ME 36 36(35). Most nodules pustuliform, scattered on blade: most calyx lobes subquadrate (Fig. : 5) 26. P. staudtii 3 lost nodules elongate and associated with secondary or higher order veins; calyx lobes various. often ovate or rotund but most not subquadrate 23. P. — re 37(34). Most nodules pustuliform, scattered on blade; most calyx lobes subquadrate (Fig. 5) — 26. P. staudtii 7 Most nodules not pustuliform; nodules absent or most elongate or associated with midrib or be veins; calyx lobes various but most not subquadrate . P. owariensis l. Pavetta baconiella Bremekamp. Repert. and subumbellate, 0.2-1.5 em across, puberulent, Spec. Nov. Regni Veg. 37: 73. 1934. TYPE: sessile, flowers 1-75; sheathing bracts saucer- to Cameroon. Between Monjala and Mole, Mild- cup-shaped, subglabrous, sometimes with a linear braed 8374 (holotype. B destroyed). NEO- awn ca. 3 mm: other bracts to 2 mm, + ovate t TYPE: Cameroon. Southwest Province: South obovate, sometimes with l—-several fimbriae ca. | Korup National Park, in rocks by Mana River, mm; bracteoles ca. 0.5 mm long resembling fim- July 1983 (fl, fl bud € fr), Thomas 2212 (neo- briae of smaller bracts occasionally borne directly type, MO: isoneotypes, K, P. WAG not seen, on pedicels. Calyx tube 0.7—1 mm long, 1-1.5 mm A; 2 other isoneotypes at unknown locations wide halfway up: lobes valvate, rotund to deltoid, [BR?, PRE?, G?] not seen). Figure 7. broadly triangular, subquadrate, pentagonal or ovate, 0.2-0.8 X 0.5-0.8 mm, puberulent, some- times carinate, rim usually lighter. Corolla yellow: floriferous twiglets 1-6 cm, 2(—4) of them some- tube evlindrical. 2-3 X 1 = Shrubs 0.8 m. Twiglets glabrous or subglabrous, mm; lobes 3-5.5 mm: times growing sympodially to a total length of up — ale clavate. subglabrous to pubescent, exserted 3— to 9 cm, the more proximal inflorescence(s) then 4 mm. Fruits 5-10 mm across, glabrous or sub- pseudoaxillary. Leaves chartaceous, sometimes — glabrous, Mature seeds 2 or 1 attached ca. halfway slightly anisophyllous: blades ovate to elliptical (or up septum, concave. less often oblong or obovate), 2-10 Xx 0.5-3 cm, glabrous except major veins subglabrous or puber- Pavetta baconiella was first collected “between ulent above and below; apex acute, subacuminate — Monjala and Mole.” Cameroon. This location is un- or with acumen 5-15 X 2-3 mm, often curved; known. It is otherwise known only from the very base cuneate to attenuate; midrib prominulous be- high rainfall Korup National Park in Southwest low toward base; secondary veins 5-12 each side, Province, Cameroon. It was found growing in riv- usually eucamptodromous, small hairy pit or crypt. erside rocks along the Mana River when the river domatia sometimes in branch vein angles of midrib level was probably higher than average. or secondary veins; nodules scattered on blade. Leaf size of Pavetta baconiella is the smallest of sometimes along midrib, few to many; third and all Cameroon species of subgenus Baconia. Sym- higher order venation + equally obvious above and podial floriferous twiglet growth such as that on below; venation density coarse. Stipules deciduous, some branches of P. baconiella has not been seen cup-shaped, pubescent internally, subglabrous ex- elsewhere in the subgenus. ternally, the linear awn to 2 mm. Inflorescences ro- The neotype is the only known extant collection tund to inverted pyramidal in outline, condensed of P. baconiella. 102 Annals of the Missouri Botanical Garden N Figure Pavetta baconiella (Thomas 2212, MO).—A. Habit.—B. Inflorescence.—C. Infructescence. of leaf venation and three bacterial nodules on upper leaf surface. D. Details Volume 83, Number 1 1996 Mannin 103 g Pavetta Subgenus Baconia in Cameroon 2. Pavetta bidentata Hiern, Fl. Trop. Africa 3: 176. 1877. Ixora rajo (Hiern) Kuntze, Revis. Gen. Pl. I: . 1891. TYPE: Equato- rial Guinea. Bioko pode formerly Fernando Po, Apr. 1860 (fl), Mann 395 (holotype. K). KEY TO THE VARIETIES OF PAVETTA BIDENTATA IN CAMEROON 1. Leaves sessile or subsessile; leaf bases founded to cordate (or rarely attenuate), some pouched a in Figure var. sessilifolia l. pe with petioles to 35 mm long, some ch longer than | mm; leaf bases cuneate (to less ofter rounded or attenuate), not pouched var. bidentata a. Pavetta bidentata var. bidentata D gee K. Schum., Bot. Jahrb. Syst. 33: 353-354. YPE: Camero Province: near Tus Pe 1899 (ib. Deistel 127 (holotype, B de- stroyed). Pavetta permode sta Wernham, J. Bot. 54: 27. 1916. Syn. nov. T : Cameroon. South Province: Bitye, Feb. (fl bud & Pavetta rcu orum Premi . Nov. Regni 78. 1934. TYPE: Nicea, Eket District, Talbot & Talbot 3316 gls La BM). s pe ert. Spec le T: 1912— 13 (fl). Pome venusta Bremek., 78. 19 Syn. : Cameroon. Prive e: Y eie Mildbraed 797 78 (holotype. B de- stroyed). Shrubs to 4 m. Twiglets glabrous or subglabrous, floriferous twiglets to 31 cm. Leaves petiolate, pet- ioles to 35 mm; blades coriaceous to chartaceous. sometimes anisophyllous, elliptic to obovate, ob- long or ovate, 31 X 0.7-10.5 cm, glabrous, at times subglabrous below near midrib, veins gla- brous; apex acute (or rarely emarginate), sometimes with acumen 3-25 X 1-10 mm: attenuate (or rarely rounded), sometimes asymmet- rical; midrib sometimes prominulous below near 2) each side, some- base cuneate to base; secondary veins 4—18(— times joined 1-10 mm from margin; pit, pocket, tuft, or intermediate between pit and crypt domatia usually in branch vein angles of midrib, occasion- ally in branch vein angles of secondary veins, pu- bescent, sometimes extending a few mm distally from branch vein angle; nodules usually elongated along midrib; if present on blade. often less fre- quent than along midrib, sometimes associated with veins; fourth and higher order venation usually more obvious below; venation density medium to fine. Stipules sometimes deciduous: lobes variable including ovate, rotund, compressed rotund, atten- uate, deltoid, or pentagonal shapes, or truncate; gla- brous; awn linear, 2-15 mm. Inflorescences pyra- midal to rotund or corymb-shaped in outline or with subunits of these shapes, 1.5-16 em across, some- times lax, puberulent to glabrous, peduncle if pres- ent to 15 mm; flowers (1-)10-100(2250); sheathing bracts with rotund, ovate, deltoid, truncate or irreg- ular lobes or unlobed, then funnel- or saucer- shaped, glabrous, at times subglabrous externally, most with narrow awns 0.5—10 mm long, upper ones sometimes with foliar appendages 1-6 mm long: foliar bracts absent or 10-20 mm long. Other bracts linear to pentagonal, ovate, subquadrate. spreading and semicircular, or obovate, to 4 mm long some- times including l-several fimbriae to 2 mm long. bracteoles re- sembling smaller bracts. Calyx tube 1-1.8 long. 1.5-2 mm wide halfway up: lobes valvate. tri- angular to rotund (or less often pentagonal, ovate. trapezoidal, subquadrate, compressed rotund (Fig. 5) or 2-pointed); (0.2—)0.5-1 X 0.7-2 mm, glabrous (to occasionally subglabrous), sometimes carinate: rim lighter colored. Corolla white; tube cylindrical or subeylindrical, 4-10 X 1-2 mm: lobes 5-10(- 12) mm. Style fusiform to clavate, glabrous to pu- exserted 5-15 or fimbriae borne from axis directly: mm berulent, mm. Stigma narrowly 2- lobed. persistent calyx lobes, glabrous, pink or white with Fruits to ca. 1 em across, sometimes with green stripes, then blackish. Mature seeds 2 or 1, attached ca. halfway or more up septum, concave. examined. CAMEROON. Southwest Province: Limbe, Apr. (fl), Maitland 1180 (B. K): Baro, near Korup National Park, Feb. (fl), Nemba, Thomas & Mambo 889 (MO); Mbu, Rumpi Hills, 10 km \ Additional ae imens V of Wone on 2 Mamfe road, Nov. (fl. Mambo & Thomas 1 (MO), Oct. = pue 627 (MO); Etam, Bak- ossi Forest Reserve. between Kumba and Tombel. Mar. (fl), Etuge & Thomas al (MO) (not 8la, which is Pavetta rigida Hiern); Bakossi Mountains W of Bangem, Jan. (fl). Thomas & McLeod 5266 (BR, MO), Oct. (fr), Etuge 535 (MO): T Fakamanda Forest Reserve, Apr. (fl), Thomas et al. i (MO). Littoral Province: between Njoke and Mal- e (possibly Southwest Province), Feb. (fl), Schlechter een (BR, K); 3 km E of Eboné, ca. 10 km S of Nkong- samba, Apr. (fl & vegetative), Leeuwenberg & Ber, (WAG Y v and 8 km W of Masok, Apr. (fl), Leeuwenberg 5414 WAG in part, the other WAG sheet being P rigida Hie rn). Centre Province: Nkol Bisson, ca. 7 k W of Yaoundé, Jan. (fl bud), de Wilde & de Wilde- Dustin e (P. WAG), Apr. (fl bud & fl). Breteler 2737 WAG). Nov. (fr), de Wilde & de Wilde-Duyfjes 1208. di WAG): hill Akondoi W of Etoug Ebé, Yaoundé, June (fr), Manning 1916 (MO): Mt. Eloumden, ca. 10 km SW of Yaoundé, Mar. (fl bud), Sonké 86 (MO), Li (fl), Manning 9 (MO) and 2/22 y (fr), Manning 2160 e p mw 2 ay E (MO): Ngoro, Massif de Ngolé be Me Band of Bafia Apr. ( a Raynal & Raynal 10675 (P, YA). South Prov- ‘ampo Game Reserve, Mar. (fl bud), Nkongmeneck ince: 409 em Bidjap. 32 km E of Nyabéssan, Mar. (fl bud fl), Raynal & Raynal 10283 (P, YA); b xi (fl), Staudt 212 (Sy Bitye (fl bud), Bates 1318 (MO). Feb. ( bud & fl). Bates 716 (BM). month inkhown g bud & fl), Bates 1203 & 1318. combined sheet (BM): hill Ebon, near Nko- Ere 104 Annals of the Missouri Botanical Garden 5 ME 2 "n AD ESA Y DA d- / SEES Se SS hee (A PHYLLIS BICO Figure 8. Pavetta bidentata var. sessilifolia (Letouzey 11557, P).—A. Habit, including flowers and fruits on the same inflorescence.—B. Node showing a pouched leaf base and long stipule awns.—C. Details of leaf venation with bacterial nodules along midrib. Volume 83, Number 1 1996 Manning 105 Pavetta Subgenus Baconia in Cameroon biyo, 25 km ENE of ip ys Mar, (fl bud & fl). Letouzey 10187 (P). Unknown Location: collector unknown. la- beled from He ye atthe E iis s Gardens (fr). SCA 2259, HNC . 5 (YA). NIG Eket District: (fl), Talbot & Hefe p (BM) b. Pavetta bidentata var. sessilifolia S. D. Man- ning, var. nov. TYPE: Cameroon. Centre Prov- ince: hill Kombeng, 8 km SSE of Matomb, ca. 50 km WSW of Yaoundé, July 1972 (fl bud, fl & fr), Letouzey 11557 (holotype, P: isotype, YA). Figure 8. varietate bidentata foliis laminis basi plerumque ro- tundatis ad cordatas, interd ut maximum | mm differt um saccatis, petiolis nullis vel Similar to variety bidentata except full-sized leaves sharply different in being sessile or with pet- ioles S 1 mm and having bases mostly rounded to cordate, sometimes pouched, and not cuneate. Variety sessilifolia is placed within Pavetta bi- dentata because its collections resemble other col- lections of P. bidentata more than any other taxon of subgenus Baconia. It is given varietal status be- cause of the sharp differences stated in the last paragraph. Additional specimen examine == CAMEROON. Centre d vince: Mt. Kala, 25 km W of Yaoundé, May (fl bud & fl). Farron 7245 (P). Pavetta bidentata var. bidentata occurs in Zaire, Equatorial Guinea (Bioko Island), and Nigeria as well as Cameroon, and thus is centered in the Low- er Guinean and Congolean subcentres of specific endemism sensu White (1979), In Cameroon, it is South, Littoral. and Centre provinces. Variety sessilifolia is so far widely distributed in Southwest, known only from a small area in southern Centre Province. Cameroon. Pavetta bidentata has been found in primary and secondary forest and at elevations from sea level to as high as 1750 m. Plants collected on rocky or thin-soiled substrates, as well as higher elevation ones, tend to be smaller leaved than those of low- land substrates not reported as rocky. Aids in recognizing this species are stipule awns usually longer than in other species, often longer than 10 mm: nodules more prevalent along the mid- rib than elsewhere on leaves: often elongated and pubescent domatia along the midrib; often narrower than average leaves for their length; and, apart from domatia and corolla throats, plants almost com- pletely glabrous. Several specimens of variety bidentata show hel- icoid inflorescence branching, which is very unusu- al in subgenus Baconia. In variety sessilifolia, the terminal leaf pair is sometimes reduced to one leaf, and the leaves of the uppermost two nodes some- times appear whorled because these nodes are sep- arated by a very short internode. Pavetta permodesta was a species of unbranched monocaulous dwarfs sensu Robbrecht (1988) whose short stature may have indicated a young stage of the plant, the result of ecological hardship, or hav- ing been cut off at ground level. The latter two situations have both been reported on herbarium labels of specimens referred to P. permodesta. Fea- tures of specimens earlier referred to P. permodesta overlap those of other members of P. bidentata var. bidentata. Although all material of Pavetta venusta seen by Bremekamp (1934) has been destroyed, and no oth- er herbarium material has been found, all P. biden- tata plants near Yaoundé fit Bremekamp's descrip- Their overlap with those of other members of P. bidentata tion of P. venusta. very well. features. also var. bidentata. 3. Pavetta brachyealyx Hiern, Fl. Trop. Africa 3: 169. 1877. Ixora brachycalyx (Hiern) Kuntze in Revis. Gen. Pl. 1: 286. 1891. TYPE: ( Southwest Province: Mt. Cameroon, Dec. (f. Mann 2159 (holotype, K; isotype. P). lameroon. 1862 Shrubs to 4 m. Twiglets glabrous, floriferous twiglets 4-22 cm. ally anisophvllous: blades elliptic to oblong or ob- ovale. (1—)6-20 X (1—)2—9 em. glabrous, veins gla- glabrous (to less often subglabrous) Leaves subcoriaceous, occasion- brous above, below: apex acute to obtuse, usually with acumen 5—10 X 3-8 mm; asymmetrical; base cuneate to attenuate, oc- asionally midrib prominulous or T prominent below, secondary veins 4-11 each side, sometimes joined (1—)3—7 mm from margin. some- times prominulous below: sparingly hairy pocket, pit or crypt domatia in branch vein angles of midrib veins; nodules scat- and occasional ly seconi lary tered on blade: fourth and higher order venation usually more obvious above: venation density me- dium. Stipules deciduous, rotund lobed to unlobed and cup-shaped, pubescent internally, glabrous ex- + falcate, 14 Inflorescences rotund or subrotund to corymb- ternally, awn cuspidate or linear, mm. shaped in outline or with subunits of these shapes. 1-8 em across, puberulent distally to subglabrous proximally, peduncle absent or to 7 10-100. Sheathing bracts with rotund or truncate lobes or unlobed, then bowl-shaped, deciduous, pu- mm, flowers bescent internally, glabrous (to at times subgla- brous) externally, the triangular or linear awns 14 x 0.5-1 mm. Other bracts linear to broadly obo- Annals of the Missouri Botanical Garden vate, to ca. 2 mm, usually several-fimbriate, fim- briae ca. 1 mm, or fimbriae borne from axis directly. Bracteoles resembling smaller bracts. Calyx tube 1-1.3 mm long, 1.5-2.5 mm wide halfway up; lobes valvate, compressed rotund (or less often short tri- angular, rotund or denticulate), 0.2—0.5(-0.8) X 1- 1.5 mm (or occasionally truncate at base), subgla- brous, sometimes carinate, rim not or narrowly lighter. Corolla creamy white; tube cylindrical or subcylindrical, 3-5(-6) X 1-2 mm; lobes 4-6 mm. Style pubescent, clavate, exserted 5-10 mm. Fruits 7-10 mm across, glabrous, mature color unknown. Seeds 2, attached ca. halfway up septum, concave. Additional specimens | examined. CAMEROON. Southwest Province: Mt. Cameroon, Jan. (fl), Dunlap 20 O); ;ameroon, Limbe (Victoria) District, Dec. (fl), ) : Cameroon S slope above : Mt. Cameroon, Buea J. District, [m (fl & fr), Maitland 2 3 (K eae — = Pavetta brachycalyx is a forest understory shrub endemic to Mt. Cameroon between ca. 500 and 1500 m. Although specimens with short calyx lobes from other locations have previously been identified as this species, they belong to other species. Among species with most calyx lobes com- pressed (Fig. 5), Pavetta brachycalyx most closely resembles P. kupensis S. D. Manning. Pavetta ku- having mostly strongly obovate leaves and venation often conspicuously brochidod- romous below to the naked eye. It also has larger pensis differs flowers. Pavetta gabonica Bremekamp is also sim- ilar but has nonc ed calyx lobes. 4. Pavetta brachysiphon Bremekamp, Repert. Spec. Nov. Regni Veg. 37: 74. 1934. TYPE: East Province: near confluence of Lom and Djérem Rivers, near Deng Deng, ca. 235 km NE of Yaoundé, Mar. 1914 (fl), Mild- braed 8536 (holotype, B destroyed; lectotype, selected here, K). Cameroon. Shrubs. Twiglets subglabrous, floriferous twiglets 8-20.5 cm. phyllous; blades elliptic to oblong or obovate, 3— Leaves chartaceous, sometimes aniso- 1-3.5 cm, glabrous below, minutely puberulent or papillate above; veins puberulent to subglabrous below, puberulent above; apex acute with acumen 4-10 X 2-6 mm; base cuneate to attenuate; midrib and secondary veins prominulous below, secondary veins (5-)8-10 each side, sometimes joined 2-6 mm from margin; tuft domatia in branch vein angles of midrib and sometimes secondary veins, some- times elongated; nodules scattered on blade and sometimes on midrib; fourth and higher order veins more obvious below; venation density fine. Stipules deciduous, sheathing, puberulent externally, pubes- cent internally; awns not seen. Inflorescences sub- rotund in outline or with subrotund subunits, 1-3 cm across, congested, pubescent to puberulent, ses- sile, flowers 50-100; sheathing bracts subrotund lobed to cup-shaped without lobes, pubescent in- ternally near base, puberulent to subglabrous ex- ternally; sheath with linear awns or foliar append- ages to 2 mm; other bracts deciduous, subquadrate to subquarter-spherical or ovate, to 2 mm, times with l—many fimbriae to ca. | mm; bracteoles ej smaller bracts. Calyx tube 0.5—0.8 mm long, 1 some- .2-1.8 mm wide halfway up; lobes cla rotund to oblong, ovate or pentagonal, 0.5-1.: 0.5-1 mm, puberulent, often carinate, rim bue 1-2 X 1 mm; lobes 2-4 mm. Style clavate, subglabrous, exserted 2—4 mm. Corolla tube cylindrical, Pavetta brachysiphon is known only from the type specimen from northwestern East Province, Cameroon, near a forest-savanna boundary. Al- though the calyx lobes of P. brachysiphon are of approximately average length in subgenus Baconia, P. brachysiphon is distinguished by its otherwise very small flowers in very congested inflorescences. Leaves are small and narrow. The following aspects of leaf morphology are also distinctive: leaves are more or less papillate on upper surfaces and tend to dry golden brown below, darker brown above; domatia are usually along lateral veins as well as the midrib; and fine venation tends to be more ob- vious below than above. Pavetta longibrachiata Bremekamp, P. laxa S. D. Manning, and P. cellu- losa Bremekamp are three species similar to P. bra- chysiphon in these aspects of leaf morphology, ex- cept fine venation tends to be more obvious above than below in P. cellulosa. 5. Pavetta calothyrsa O Spec. Nov. Regni Veg. 37: Cameroon. East Province: 3 a “Lamoko” and the “Posten Plehn,” formerly Bezirk Molundu, probably near Ndélélé, Apr. 1911 (fl), Mildbraed 4921 (holotype, B de- stroyed; lectotype, selected here, HBG) Repert. 34. TYPE: , between Shrubs to 4 m. Twiglets glabrous, floriferous twiglets (6—)20-27 cm. gla- brous, occasionally anisophyllous; blades elliptic to ovate (or less often rotund, oblong, or obovate), (5—) 9-32(-37) X 3-13 cm; apex acute to obtuse or round- ed, usually with acumen 5-20 X 3-15 neate to attenuate, sometimes asymmetrical; Leaves coriaceous, mm; base cu- midrib and secondary veins sometimes prominulous below; secondary veins (5—)7-10(-11) each side, usually eu- camptodromous; small, subglabrous pit or pocket Volume 83, Number 1 Manning 107 1996 Pavetta Subgenus Baconia in Cameroon domatia sometimes in branch vein angles of midrib; Bana, nodules uncommon, usually limited to midrib if pres- ent; fourth and higher order veins usually slightly more obvious below, at least somewhat obscure on both faces; venation density fine. Stipules sometimes deciduous, cup-shaped, glabrous externally, pubes- cent (to rarely subglabrous) internally, the some- times deciduous linear to triangular awn ca. 3-8 mm. Inflorescences subrotund to corymb-shaped or with subunits of these shapes, 3-25 cm across, pu- bescent to glabrous; peduncles absent or to 20 mm; flowers 35450; sheathing bracts sometimes decid- uous, lobes broadly triangular to rotund or unlobed, sheath then saucer- or cup-shaped, puberulent to glabrous externally, pubescent or puberulent inter- nally, sometimes with linear awns 1-7 mm or foliar appendages 2-23 mm, foliar bracts similar to slightly reduced foliage leaves occasionally present, other bracts wedge-shaped to ovate, obovate or lin- ear, sometimes truncate or deltoid at top, lobed, o with |-several fimbriae to 1 mm long at top or fim- briae borne from axis directly; bracteoles resem- bling smaller bracts. Calyx tube 1-1.5 mm long, 2— 2.5 mm wide halfway up; lobes valvate (or less often overlapping slightly at the base), subquadrate, 0.5— 1.5(-2) X 1-2 mm, puberulent, often carinate, rim shallowly 2-lobulate (or less often rotund). sometimes lighter. Corolla white, whitish green, or white with greenish apex: tube cylindrical except often constricted at or near base to ca. half its di- ameter above, 5-9 X 2—4 mm: lobes 5-12 mm, sometimes reflexed. Style clavate, pubescent to pu- exserted 8-15 narrowly 2-lobed. berulent, mm. Stigma sometimes Fruits 7-10 mm across, crowned by slightly accrescent persistent calyx, subglabrous to glabrous, whitish green. Mature seeds 2 or 1, attached ca. halfway or more up septum, concave. Additional specimens examined. CAMEROON. East Province: Boyo River of Kongolo near Bétaré Oya, 'Ay 5 km Feb. (fl bud). Letouzey 353.3 (BR E of Ber- toua. Dec. (fr). Brete ler 790 (W "5 5 km E of Moloundou. Apr. (fl). Villiers 092 (P); River Dja between Ntouo and mélima. Mar. (fl), poms zey Province: Yaoundé, Jan. or Feb. (fl bud). Mildbraed 8014 (HBG, K). month unknown (fl bud & fl). Zenker 700 Ade month unknown (fb. W): N'Kolbisson.- K upper part of : sheet, P, S). s.n. (NY), Zenker & Sa 269 W of Yaoundé, Nov. (BR, MO, WAG, YA); jee 28 km NE of Bafia, Mar. (fl bud & í). Raynal & Raynal 10557 $a YA); ET pr. (f), Raynal 2 Arie 10713 (P, 40 km SSW f Bafia, Dec. (fl Y, Ngolé near Ngoro, A YA): and Colline d Ebat. bud), Letouzey 9692 (BR, Fokam. 10 km NE of usina; Apr. (fl bud & fl), de Wilde & de Wilde- Duyfjes 2378 (MO, BR. WAG, YA): Mt. Jan. (fl). Félix 2994 (P). The northwest portion of the known range of Pav- etta calothyrsa is in Cameroon, where it occurs in widely scattered drier areas of forest in East, South. Centre, and. West Provinces. The species has also been reported from Gabon and Zaïre. It is com- monly found in gallery forest and sometimes in or on the edge of forest in otherwise savanna areas, 0 in disturbed areas. It has been reported on marshy black humus and rocky substrates. Reported. ele- vations are from 650 to 1400 m. The inflorescences of Pavetta calothyrsa are usu- ally among the largest. and peduncles are some- times among the longest, in subgenus Baconia. Most leaves and flowers are also larger than average for the subgenus. The only other taxon of subgenus Baconia from Cameroon whose fruits are known to have conspicuously large, persistent calyx lobes is P. owariensis Palisot de Beauvois var. satabiei S. D. Manning, in which they are rotund to deltoid, pen- tagonal, or ovate but not subquadrate or 2-lobulate as in P. calothyrsa. Pavetta calothyrsa and P. mo- lundensis are superficially similar but differ in calyx lobe shape. Pavetta calothyrsa resembles P. robusta Bremekamp in calyx. lobe shape. large leaves, and many-flowered inflorescences, but calyx lobes in P. calothyrsa are 1 mm or more long, those in P. robusta 0.5-1 mm long. Also. bacterial nod- ules are absent or few and usually along the midrib in P calothyrsa: in P. robusta, they are scattered on the blade more than along the midrib. Pavetta calothyrsa is restored here from synon- ymy with P. nitidula Hiern, which does not occur in Cameroon. Pavetta calothyrsa and P. nitidula are 1898: Bridson, 1978), but P. niti- dula has sessile or subsessile leaves only 3.5-16.5 similar (Hiern, 14.5 em lacking acumens, usually triangular or rotund calyx lobes, sometimes purple or pink co- rollas, many more nodules including some scattered blade, inflorescences only 2—7 cm on the and across. 6. Pavetta camerounensis 5. D. Manning, sp. nov. TYPE: Cameroon. South Province: massif de Ngovayang. 16 km W of Lolodorf. Feb. 1979 (ff bud & fl), Satabié & Letouzey 373 (holotype, P: isotype, YA). PAVETTA CAMEROUNENSIS AN KEY TO THE SUBSPECIES OF CAMEROON |. Floriferous twiglets 9-21 cm long subsp. camerounensts |. Floriferous twiglets absent or 0.5-9 cm long subsp. brevirama 108 Annals of the Missouri Botanical Garden a. Pavetta camerounensis subsp. camerou- nensis. Figure 9. Frutices. Rami glabri (vel ik ne subglabri). Rami floriferi 9-21 cm. Folia nervis secundariis utroque 8-23, domatiis nullis, aliquot venis do interdum. subter ponens vel prominulis. Inflorescentiae 0.2— e congestae. | 2.5 cm alae, sae Lobi calycini valvati, subrotundati (ad tendum uen triangulares, denticulatos ve truncatos), = 1 X 1-1.5 mm, glabri. Corolla tubo 2-4 mm, lobis 3-5 mm, qs saepe super prope faucem pu- bescentibus. Styli exserti 2-4 mm. Semina 1-2(-3). Shrubs to 2.5 m. Twiglets glabrous (to at times subglabrous), floriferous twiglets 9-21 cm. Leaves chartaceous to coriaceous, glabrous (to less often puberulent on veins below), occasionally aniso- phyllous; blades broadly to narrowly obovate, ellip- — tical, ovate (or occasionally oblong), 12-32 X (2.5- 4—12 cm; apex with acumen 5-30 attenuate, secondary veins prominent or prominulous below at acute to obtuse or rounded usually X 2-10 mm: sometimes asymmetrical; base cuneate to midrib and least near base, secondary veins sometimes im- pressed above, 8-23 each side, usually joined 1— 11 mm from margin; domatia absent; nodules ab- sent or few, linear, along midrib and secondary veins or rarely on blade; tertiary veins sometimes prominulous below; fourth and higher order vena- tion obvious above and usually below; venation density medium. Stipules often deciduous or frag- mented, apparently rotund to ovate or deltoid lobed, glabrous or subglabrous (or less often puberulent) externally, glabrous to thinly pubescent especially near base internally, at least sometimes with a lin- ear or cuspidate awn 1-5 1-3 mm. Inflores- cences sometimes subcapitate or subumbellate, ro- 0.2-2.5 congested, puberulent to glabrous, peduncle absent tund to subrotund in outline, cm across, or to 3 mm, obscured by sheathing bracts, flowers 5—60; sheathing bracts rotund to deltoid lobed or unlobed, then saucer- or bowl-shaped, glabrous to puberulent or thinly pubescent, sometimes with ovate or linear awns or foliar appendages 1-3 mm long, sometimes attached ca. halfway down the sheath, sometimes near its apex; other bracts often light cream, subquadrate to subtruncate and wedge- shaped, ovate, deltoid or linear, often at least as wide as long, to 2 mm, sometimes including 1-3 fimbriae or a linear awn ca. 1 mm, or fimbriae borne from axis directly; bracteoles resembling smaller bracts. Calyx tube 0.7—1 mm long, 1.2-2 mm wide halfway up; lobes valvate, compressed ro- tund (to less often truncate, rotund, triangular, den- ticulate or cleft near apex), = 1 X 1-1.5 mm, gla- brous, sometimes carinate, rim lighter, sometimes narrowly so. Corolla white, creamy, pale green, or greenish yellow; tube cylindrical or subcylindrical, 1-2 mm; lobes 3-5 mm, usually puberulent near base above. sometimes constricted near base, 24 X Style fusiform or clavate, pubescent to subglabrous, exserted 2—4 mm. times with persistent calyx, glabrous, dirty or glau- Fruits 7-12 mm across, some- cous grayish green, bluish, or greenish white with Mature seeds 1—2(—3), attached halfway or more up septum, concave. a bluish sheen. Additional e examined. CAMEROON. South ce: Bipindi, Jan. (fl), Manning 1408 (MO); hill Nkoltsia ca. 23 js NW of Bipindi, Nov. (fl bud), Villiers 1003 (P); 27 km E of Kribi, Jan. (fl), Bos 6173 (K, P, WAG); and 10 km E of Kribi, May (fr), Bos 4654 (WAG). Centre Province: N'kolbisson, 8 km W of Yaoundé, Nov. A iud & 2 ~ Wide Pi de Wilde-Duyfjes 1211 (WAG), fr). « e Wilde- Pi n 1184 (P. WAG), | (MO); M Fébé, near Yaoundé, 3801 (WAG y 15 km SSW o Obala, Dee, d. iyd 9776 (BR, i Centre ig Lit- toral Province: Kéllé River, 60 km NNW of Eséka, Mar. (fr), de Wilde & de Wilde Duyfjes pus (P. WAG a Kéllé of Eséka, Nov. (fl bud & i tative), de Wilde & de Wi p Duvfjes 1318 & 1318 aH e 7 1318C — ~~ < e = = = gabonica; 30 a 'NW of F ilde & de Wilde-Duyfjes 1491 (BR, WAG É uie Pos ie 1318 is a mixed collection; 1318, 1318B, and 7318C above refer to designations on herbarium = hon ole tags attached to specimens, when the des yan ae conflict b. Pavetta camerounensis subsp. brevirama S. D. Manning, subsp. nov. TYPE: Southwest Province: Barombi-Mbo village, ca. 5 km NW of Kumba, Dec. 1986 (fl bud & fl), Nemba & Thomas 417 (holotype. MO: iso- types, BR, YA not seen). Figure 10. Cameroon. A subspecie camerounensi ramis floriferis absentibus vel non plus quam 9 cm longis, inflorescentiis interdum is differt. Similar to subspecies camerounensis except flo- riferous twiglets absent or less than 9 cm long, in- florescences sometimes axillary with peduncle to 5 Although the longest floriferous twiglet of subspe- cies brevirama is as long as the shortest in subspe- cies camerounensis, floriferous twiglets of the two subspecies normally differ sharply in length (Figs. 9 and 10). Although other character states overlap, the following features of subspecies brevirama also tend to differ and may be useful in confirming identifi- cations: absence of leaves with broadly ovate blades (common in subspecies camerounensis) or rounded apices, though other shapes described for subspecies camerounensis all occur; nodule-like growths if pres- ent sometimes ramifying extensively with leaf ve- Volume 83, Number 1 Manning 9 109 Pavetta Subgenus Baconia in Cameroon NA CIA 1/2 LG ( Pr ASIS eb.) AA D Figure 9. Pavetta camerounensis subsp. camerounensis (Satabié & Letouzey 373. P).—A. Habit.—B. Details of leaf venation.—C. Flower. 110 Annals of the Missouri Botanical Garden iss os: ; OK = Figure 10. nation; sheathing bract awns sometimes up to 5 mm; other bracts sometimes obscured by inflorescence congestion; calyx lobes occasionally subquadrate, sometimes puberulent; major veins rarely impressed Pavetta camerounensis subsp. brevirama (Nemba & Thomas 417, MO). (A, C, D, MO holotype; B, MO isotype).—A. Habit—larger inflorescences terminal on short side branches, smaller ones axillary at lower nodes.—B. Nodes, one showing an axillary inflorescence.—C. Flower with corolla throat beard extending out onto corolla lobe.— D. Flower, corolla and part of calyx removed to show epigynous nectary disk. above. Subspecies camerounensis specimens often dry lighter green than those of subspecies brevirama. Also, the two taxa are largely geographically sepa- rate. Volume 83, Number 1 Manning 111 Pavetta Subgenus Baconia in Cameroon 1996 Additional ini imens | examined. d in )ON. ut hwest : Korup National i» r. (fl & g fr), Thomas ud Meleod 5724 (MO). I a ‘bud & (MO), Apr. (young fr), Manni 1742 pra (young fr he fr), Mannin (MO). Apr. (fr), Manning 1729 (MO): Ekondo Tini- Mun. - demba road, Jan. a. Thomas 4358 (MO); S of Ekumbako, Dec. (fl bud, fl & fr), Thomas 2703 (MO); Mundemba- Fabe road. Nov. (fr). Mos 336 (MO); Southern Bakundu Forest Reserve, Banga, Mar. (young fr), Brenan 9279 oe 9279A (K), W of seco Sep. (fr), Manning 101 (fr). ia dede! ib 30719 (K); inside crater rof Lake Barombi-Mb JO, 3 km 8 ~ =1 eE . Manning 849 (MO), 856 (MO): 1 3ar his Mbo, ca. 5 km NW of Kumba. Dec. ( fr), Mead 1080 (MO); Bolo-Meboka. Kumba- Mamfe road between Kumba and Konye, Sep. x : Baduma, Kumba-Mamfe road. Aug. Thomas 185 (MO): Bakolle- Mies ins d Mamfe din. May (fr). Etuge & Thomas a O). Littoral Province a km A bifurcation for Edén. ~ — Douala-Yabassi road May (fr). Farron 7286 ( So far as is known, both subspecies of Paretta camerounensis are endemic to Cameroon. Subspe- cies camerounensis has been found only in the wet western portion of South Province and in southern Centre and (7) Littoral Provinces. Subspecies bre- virama is one of the most common Pavetta taxa in and near the extremely high rainfall Korup National "ark. in southwestern Southwest Province. It also occurs north and east of there in less extremely wet forest in Southwest and Littoral Provinces. Subspecies camerounensis is a shrub at eleva- tions from 120 to 1000 m in varied habitats. usually in fully shaded forest. It also occurs in degraded and secondary forest and forest borders and has been collected both along rivers and on mountains. Reported substrates include sandy and clayish soils and loamy soil on granitic rock. Subspecies brevir- ama is a small shrub in deep shade and less often partly shaded locations. Its highest reported ele- vation is 350 m, inside the crater of Lake Barombi — Mbo, not far above lake water leve Corolla throat vestiture of Pavetta camerounensts extends onto adaxial surfaces of corolla lobes, a con- dition found in only a few other species. Pavetta camerounensis resembles P. grossissima, P. gabonica, like P has inflorescences so condensed as to and P tenuissima. Pavetta camerounensis, grossissima, often appear subumbellate: P. grossissima, however. has extremely coarse reticulation of higher order leaf venation, unlike P. camerounensis. The largest inflo- rescences of P. camerounensis resemble those of P gabonica, but in P. gabonica the corolla throat. ves- titure does not normally extend onto the lobes as it does in P. camerounensis. The number of secondary vein pairs does not exceed 10 in P. gabonica; it al- ways does on at least some leaves in P. camerounen- sis. Specimens of P gabonica often have smaller leaves than those of P. camerounensis. Pavetta ca- merounensis, particularly subspecies brevirama, re- sembles tenuissima in its condensed inflores- cences, leaf sizes and shapes, lack of anisophylly. and in that floriferous twiglets are sometimes so re- duced as to render inflorescences axillary. Pavetta tenuissima differs in that its leaves usually have ex- tremely fine mesh reticulation of their higher order veins and in that its secondary and tertiary veins have a more predominant rectangular gridlike pat- tern below on herbarium specimens. Pavetta camerounensis is so named because it is relatively widespread in Cameroon forests. 7. Pavetta cellulosa Brel me nerap, Repert. Spec. Nov. Regni Veg. 37: 1934. TYPE: Zaire. Kala, Sep. 1907 (fl bud & fl). Pynaert 1692 (holotype. BR). Shrubs or small trees with stem to at least 5 cm diam. Twiglets subglabrous or glabrous, finally pu- berulent: floriferous twiglets (3-)6-20 cm. Leaves chartaceous. occasionally anisophyllous: blades el- liptic to oblong, obovate or ovate, 5-19 X 1.5-6.5 em, glabrous, veins subglabrous or glabrous above. subglabrous below: apex acute (to less often round- ed) with acumen 5-10(-15) X 5 mm: base cuneate (to sometimes attenuate), sometimes asymme trical: midrib and secondary veins usually prominulous below: secondary veins (5—)7-11 each side. usually eucamptodromous; hairy pocket or pit domatia in most branch vein angles of midrib and usually along secondary veins, sometimes extending several mm from branch vein angle; nodules scattered on blade: fourth and higher order venation usually slightly more clearly visible above; venation density medium to fine. Stipules cup-shaped, deciduous, subglabrous to puberulent externally, pubescent in- ternally, awn cuspidate, 2 X 0.5 mm. cences subrotund to corymb-shaped in outline or with subunits of these shapes, (0.5-)1-8 cm across, pubescent to puberulent, sessile. flowers (10—)25— Inflores- 100: sheathing bracts rotund lobed or unlobed and cup-shaped, puberulent externally, pubescent in- ternally, sometimes with linear awns 2—4 mm: other bracts + ovate, to ca. 3 mm, most larger ones con- cave with 1-3 fimbriae. or fimbriae borne from axis directly: bracteoles resembling smaller bracts Calyx tube 1.5-2 mm long, 2— lobes valvate, oblong, sometimes present. 2.3 mm wide halfway up: ovate, rotund, long or short triangular (or less often subquadrate or pentagonal), 0.5-2(22.5) X 1-1.5 (2) mm, pubescent, usually carinate, rim lighter. 112 Annals of the Missouri Botanical Garden Corolla white, tube cylindrical, 2-6 X 1-2 mm; lobes 5-8 mm. Anthers septate. Style clavate, pu- bescent, exserted 6-11 mm. Prefruiting ovules 2, + reniform, attached ca. halfway up septum. Additional specimens examined. CAMEROON. a Province: Bitye, River Dja (fl), Bates 1573 (BM, Pavetta cellulosa is known only from Zaire ex- cept for the collection reported here from lowland forest in eastern South Province, south-central Cameroon. Although the collection from Cameroon differs from ones from Zaire in having some calyx lobes more elongate and in having domatia less promi- nent or common along secondary veins, infraspe- cific taxa are not created here in the absence of more Cameroon material. The anther thecae of Pavetta cellulosa are nearly all septate in open flower after dehiscence. This distinguishing feature otherwise occurs in subgenus Baconia only in P. urophylla subsp. urophylla, which does not occur in Cameroon. Other features of P. cellulosa useful for identifying nonflowering specimens are the partly deciduous stipules, which leave behind bases of stipule sheaths that some- times dry conspicuously lighter brown than the dark brown stem, and the similarity to P. longibra- chiata and P. brachysiphon in leaf morphology. For example, leaves of P. cellulosa tend to dry golden brown below, darker brown above, and domatia are usually along lateral veins as well as the midrib. Higher order venation tends to be more obvious above than below, however, unlike in the latter two species, and upper leaf surfaces are not papillate as they usually are in those two species. 8. Pavetta corymbosa (A. P. de Candolle) F. N. Williams, Bull. Herb. Boissier Sér. 2: 378. 1907. Baconia corymbosa DC., Ann. Mus. Natl. Hist. Nat. 9: 219. 1807. Verulamia cor- ymbosa DC. ex Poir, Encycl. Méth. Bot. 8: 543. 1808. TYPE: Sierra Leone. 1785 (fl), Smeathman (holotype, G-DC not probable isotype, MPU photocopy seen; re- portec for Verulamia corymbo- sa). Sn. seen; as “Stadman” Pavetta rhombifolia Bremek., Repert. Spec. Nov. R leg. 37: 34. TYPE: Sierra Leone. Highlands of Bafodya, Scott Elliot 5506 (K). KEY TO THE VARIETIES OF PAVETTA CORYMBOSA IN CAMEROON Floriferous twiglets glabrous to subglabrous; leaf veins subglabrous to puberulent below; most calyx lobe bases overlapping after flowers open x var. corymbosa . Floriferous twiglets pubese enl; leaf veins puber ulent to pubescent below 10% of s lobe us Meis LA but most c d lobes valvate af- r flowers of var. neglecta a. Pavetta corymbosa var. corymbosa Pavetta corymbosa var. glabra Bremekamp, Repert. Spec. Nov. Regni Veg. 37: 70. 1934. Shrubs to 3 m. Twiglets glabrous to subglabrous, floriferous twiglets to 25 cm, sometimes strongly curved. Leaves chartaceous to coriaceous, some- times anisophyllous, blades above and below and veins above glabrous (to less often subglabrous), veins below subglabrous to puberulent; blades el- liptic to ovate or obovate, 4-17 X 1—6.5 cm; apex acute (to occasionally obtuse) with acumen 3-8 X 2—4 mm; base cuneate to attenuate, often asym- metrical; midrib prominent below, secondary veins prominulous below, 7-10 each side, sometimes joined 1-6 mm from margin; nodules along midrib and scattered on blade; pubescent domatial tufts, pockets or pits in branch vein angles of midrib and some secondary veins; fourth and higher order ve- nation more obvious below; venation density very fine. Stipules cup-shaped, subglabrous to glabrous outside, pubescent inside, awns linear or cuspidate, + falcate, 2-5 mm. Inflorescences subrotund to corymb-shaped in outline, 5-7 em across, pubes- cent to puberulent, peduncle absent or ca. 2 mm, flowers 50-100: sheathing bracts unlobed and cup- shaped or rotund or pentagonal lobed, puberulent or subglabrous outside, at least sometimes glabrous inside, sometimes with fimbriae to 2 mm or linear awns ca. 2 mm; foliar bracts 7-50 mm usually pres- ent; other bracts obovate to ovate or linear, to 3 mm, sometimes with fimbriae 1-2 mm, or fimbriae borne from axis directly; bracteoles resembling smaller bracts. Calyx tube 1.2-1.5 mm long, 2-3 mm wide halfway up, lobe bases usually overlap- ping, lobes rotund to compressed rotund, subquad- rate, obovate or shallowly 2-lobulate, ca. 1-2 1.5-2 mm, pubescent to subglabrous, sometimes carinate, rim much lighter. Corolla white; tube cy- lindrical, mm; lobes 9-11 mm. Style cla- vate, puberulent, exserted 9-14 mm. Fruits ca. 1 cm across, subglabrous or glabrous. Seeds 2, at- tached ca. halfway up septum, concave. Additional specimens examined. SIERRA LEONE. (fl). dcs s.n. eft side of sheet). CAMEROON. Adamaoua Province(?): Dodéo, Mar. (fl), Félix 339 b. Pavetta corymbosa var. neglecta Breme- Volume 83, Number 1 1996 Manning 113 Pavetta Subgenus Baconia in Cameroon kamp. Repert. Spec. Nov. Regni Veg. 37: 70. 1934. TYPE: Ghana. Volta Region, formerly Togoland: Kpandu, 1924 (fl bud & fl), Robert- son 102 (holotype, BM; isotype, MO) Similar to variety corymbosa except that the twig- lets are pubescent, there is more vestiture on leat veins above and below, and most calyx lobes are valvate after flowers open. Other character states of Cameroon representa- tives of the two varieties overlap to varying degrees see Manning, 1990, for details). The small number — of collections from Cameroon and large number of collections from elsewhere suggest that revision of this species may result in combination of the two varieties or other changes. Additional specimens examined. CAMEROON. Northwest Province: Ndop plain between Bamali and Bambalang, Apr. (fl bud & fl), Brunt 357 (K); locality un- i Nor thwest idee e th say most like sly Metchie River 2 km N d ay : bud & fl). ilde & de e Duyfjes 2519 (BR, P. WAG). West Province: Dschang. May (fl bud & fl). lo 5204 (P). Both varieties of Pavetta corymbosa occur in sa- vanna and have a more northern distribution than any other species of subgenus Baconia except Pav- etta lasioclada. The presence of variety corymbosa Ada- maoua Province in Cameroon is part of a distribu- very near the Nigerian border with northern tion west to Sénégambia in the Guineo-Congolian/ Sudanian Transition Zone sensu White (1979). The presence of variety neglecta in West and Northwest Provinces of Cameroon is part of a distribution at least as far west as Ivory Coast and as far east as Central African Republic (Hepper € Keay, 1963). mainly in the Guineo-Congolian/Sudanian transi- tion zone sensu White (1979), despite the repre- sentatives from Cameroon occurring in the Lower Guinean subcentre of specific endemism sensu White (1979). Pavetta corymbosa var. neglecta grows in gallery forest. All plants from Cameroon for which there are field data were collected near water. There are no field data on the Cameroon sheet referred. to variety corymbosa. Pavetta corymbosa is most commonly identified by its often overlapping calyx lobe bases, a feature also found in P robusta and a few other species. Leaves are smaller and inflorescences smaller or fewer flowered than in similar species such as P. robusta and P. calothyrsa. Flowers are larger than in most species from Cameroon. Although Houttuyn purportedly described anoth- er taxon as Pavetta corymbosa in 1813 (Hiern, 1877), the name Houttuyn actually gave that taxon was Crinita capensis Houtt. Attempts to find a valid publication of Pavetta corymbosa Houtt. at BR, K. and elsewhere have been unsuccessful: all refer- ences to it I have seen have been as a synonym of later names. Thus, I follow Bremekamp (1934) treating P. corymbosa Houtt. as a nomen nudum. The currently accepted name Pavetta corymbosa (DC.) F. N. Williams was thus available when pub- lished in 1907 and is thus retained. 9. Pavetta gabonica Bremekamp, Repert. Spec. Nov. Regni Veg. 37: 77. 1934. TYPE: Gabon. Sierra del Crystal: 1°N, July 1862 (fl). Mann 1729 (holotype, K: isotype, P). Shrubs to 3 m. Twiglets glabrous or subglabrous, € floriferous twiglets (2-)5-35 cm. Leaves charta- ceous to subcoriaceous, sometimes anisophyllous: blades ovate), (1 obovate to oblong (less often elliptic or 5-)3-25 X (0.5)1-11 cm; major veins glabrous (or sometimes subglabrous be- blades and low): apex acute to obtuse (or occasionally round- ed). acumen (2-)5-25 X (1-)3-12(-30) mm, rarely absent; base cuneate to attenuate (or occasionally obtuse), sometimes asymmetrical; midrib prominu- lous below toward base; secondary veins 4-10 each side, sometimes joined 2-5 mm from margin: small tuft, pit, pocket, or intermediate between pocket and pit domatia sometimes in branch vein angles of midrib and rarely along secondary veins; nodules sometimes scattered on blade, less often along mid- rib: fourth and higher order venation usually more obvious above but also visible below, occasionally equally obvious above and below; venation density medium. Stipules deciduous, unlobed and cup- shaped to triangular lobed, pubescent inside, gla- brous or subglabrous outside; cuspidate to linear awn 1-2(-5) mm. Inflorescences rotund to subro- tund, irregular (or less often pyramidal) in outline, 1—5.5 em across, flowers sometimes congested, low- er branches and subunits often well spaced, pu- bescent to puberulent (to rarely glabrous) distally, subglabrous to puberulent (rarely glabrous) proxi- mally, usually with peduncles to 10 mm, sometimes with broadly spreading sheathing bracts along and bases of peduncles: flowers 10-150; sheathing bracts rotund lobed, partly deciduous, glabrous or subglabrous outside, pubescent to subglabrous in- ] mm, occa- side; sometimes with linear awns ca. sionally with foliar appendages 3—4 mm: other bracts + ovate to obovate, to 4 mm, often with 1— many fimbriae to 2 mm, or |-several fimbriae or a tuft of hairs borne from axis directly: bracteoles re- sembling smaller bracts. Calyx tube 0.7-1 mm long. 1.2— mm wide halfway up; lobes valvate, 114 Annals of the Missouri Botanical Garden deltoid to pentagonal, rotund (or less often sub- quadrate, subrotund, or truncate), (0.1—)0.2-1 X 0.7-1.5 mm, puberulent to subglabrous, sometimes carinate, rim lighter. Corolla at least sometimes dark glaucous green in bud and cream, yellow, yel- lowish white, or greenish later, tube usually broad- ening from base to up to ca. twice as wide at throat, otherwise 24 5)1-2 mm; lobes (2-)3—5 mm. Style clavate, pubescent (or less subeylindrical, often puberulent), exserted 3-7 mm. Fruits ca. 5— 7 mm subglabrous or glabrous, orange, brownish, or dull yellow. Mature seeds 2 or 1, at- across, tached ca. halfway up septum, concave. Additional specimens examined. CAMEROON. Southwest Province: Kumba = Johann-Albrechtshóhe (fl), Staudt 565 (S); Bakolle Bakossi on Kumba-Mamfe road, May (fr), Etuge & Thomas 146 (MO); Bakossi Moun- . Jan. (fl), Thomas & Mcleod 5296 ; 1. (fl & fr), Thomas & McLeod 5354 (MO); crest , 30 m SE of Mamfe, June (fr), Letouzey 13862 . YA). e Province: 30 km SW of Eséka, Dec. (fl), de Wilde & de Wilde- Duyfjes 1493B (BR, P. WAG); 30 km WNW of E ec. (i r), de Wilde & de Wilde- Duyfes 1493 (BR, M . YA); near Ngong 25 km of Eséka, Dec. (fl bud & EN Mad 12347 2 90 km W a negar Nov. (fr), « des et al. WAG); 23 km W of Yaoundé, June fr), oc 21 28 (MO), 21. $ (MO), 2141 (MO). I pee Province (Cen- tre?): Kéllé River, 50 km NW of Eséka, Nov. (fl bud), de ?): K Wilde & de Wilde-Duyfjes 1318 (K), (fl bud & fl), 1318C (WAG) [not 1378 or 1318B Ps - or 1318C AR which are Pavetta camerounensis]. South Province: Bipindi, n unknown (fl bud & fl), “Zenker 1204 (BM. HBG, on /AG), 2468 (BM, BR), 4744 (BM, BR, HBG, A Mar. (fl & fr), pri 268 (BR. GH, MO., P, US, WAG ). dei unknown (fl & fr), Zenker 1565 (BM, BR, . K, Ss, fr), Zenker bk } month ea V)... pea (BM, Bk HBG. 4 (K); Lolodorf (fl). urs 8 (S); Ebemvok E el | v. Pu un Mar. (young fr), Raynal & Raynal 10437 (P); Efulen (= Efoulan? or Grand nee or E d or Batanga, Gabon?), Mar. (fl & fr), Bates 451 (K); 7 N of Kribi, Nov. (fl bud). Bos 5611 (P). GABON. Environ is of Libreville, July (fl bud), Klaine 1924 Sie month unknown (fl), fees 2193 (BR); Abanga Chantier C.E.T.A., June (fl bud & ~ ^s N. Hallé 2207 (P); Gaboon e T, UN, July (fl bud & A Mann 962 (K). Pavetta gabonica occurs in Gabon and south- western Cameroon, centered in the Lower Guinean White western subcentre of specific endemism sensu (1979). South Province, southwestern Centre Province, Within Cameroon, it occurs in and the eastern half of Southwest Province. It is typi- cally a small shrub of shady wet forests, sometimes near watercourses or in swampy areas. Usually a lowland species, it also has been found at eleva- tions of 800— elevations in mountains of northern Southwest Province Pavetta gabonica has two features rather unusual JO m west of Yaoundé and similar in subgenus Baconia: orange to yellow or brownish fruits and often yellowish flowers. It is compared with P. brachycalyx following the description of that species. Pavetta brachycalyx appears to be a close montane relative of the more widespread P. gabon- M. ^ a. The type specimen from Gabon is smaller leaved and fewer flowered than most members of the spe- cies, especially those in Cameroon. Other collec- tions from Gabon bridge the morphological gap be- tween the type and Cameroon representatives. 10. Pavetta grossissima S. D. Manning, api nov. TYPE: Cameroon. a Province: 5 km NE of Mundemba, Nov. 1986 (fr), Manning 944. (holotype, MO; isotypes, BR, K, YA). Fig- ure 1] Frutices. Folia nervis secundariis utroque 5-12; reti- "nig grossissimo. d tescentiae subumbellatae, gd ulae, 0.1-2 cm latae. Lobi calyce cini in fructu valvat - ped rotundato- -compressi, ovati vel pe la ca. ] mr ructus maturi Se rdum. cum calyce X, porrat aurantici Shrubs to 1 m. Twiglets puberulent (to sometimes 4-20 cm and probably longer (less often absent or <4 cm). subglabrous), floriferous twiglets chartaceous, sometimes anisophyllous; eaves blades oblong to elliptic, obovate (or uncommonly ovate), 4-22 X 2-9 cm, glabrous; veins puberulent to subglabrous; apex acute or rounded, subacumi- x 2-11 mm; neate to obtuse (occasionally rounded), sometimes nate or with acumen 5-20 base cu- asymmetrical; midrib and secondary veins promi- nulous below, secondary veins 5-12 each side, mostly eucamptodromous; minute tuft domatia sometimes in branch vein angles of midrib, some- times accompanied by pits; nodules scattered on blade, sometimes along midrib; third and higher order venation obvious above and below; venation density very coarse. Stipules deciduous, rotund lobed to unlobed and cup-shaped, subglabrous to puberulent externally, glabrous internally; awn lin- ear, 2 mm to cuspidate, 4 X 2 mm. Infructescences subumbellate, rotund in outline or reduced and of irregular shape, 0.1-2 cm across, puberulent, pe- duncle to 5 mm or absent; fruits 1-12; sheathing bracts rotund lobed to unlobed and saucer-shaped, puberulent externally, glabrous or subglabrous in- ternally, awns deciduous, fimbriae if present ca. 0.5 mm; other bracts + obovate to linear or subquad- rate, ca. 1 mm, fimbriae if present 0.5—1 mm; brac- teoles resembling smaller bracts. Calyx seen only on fruits, tube then 0.5-1 mm long, 1-2 mm wide alfway up; lobes then valvate, + rotund, com- pressed rotund, ovate or pentagonal, 0.2-1 Volume 83, Number 1 Manning 115 Pavetta Subgenus Baconia in Cameroon AN V, RUM à a k El D e ll. Pavetta grossissima (Manning 944. MO).—A. Habit, including terminal and axillary infructescences.— B. Details of leaf venation, coarsest seen in subgenus Baconia.—C. Axillary infructescence. D. Young fruit with persistent calyx. 116 Annals of th Missouri Botanical Garden mm, puberulent, sometimes carinate, rim lighter. Other floral parts not seen. Fruits ca. 1 em across, glabrous or subglabrous, sometimes with persistent calyx, orange. Mature seeds 2, or 1 and empty sec- ond locule then very reduced, attached ca. halfway up septum, concave. Additional specimens | examined. CAMEROON. Southwest Province: 10 km S e of a Nov. (fr), Manning 970 (MO), 973 (BR, K, MO); 5 km NE of Mun- demba, Nov. (fr), Manning 906 (MO), 929 (MO); Korup National Park, in s ing fr), Manning 1710 (MO), 1722 (MO), 1767 o LM 4111 (MO), June (vegetative), «n p (BR == Pavetta grossissima is known only from in and near Korup National Park in far southwestern Southwest Province, Cameroon. It is a small shrub of extremely wet forest understory in fully and part- ly shaded areas at elevations of 120 m or less. Pavetta grossissima is distinguished principally in having the coarsest reticulation of higher order leaf veins of all species of subgenus Baconia. It is so named because of this. Also, orange fruits are known otherwise only in P. gabonica, which does not have leaf vein reticulation as coarse or inflo- rescences as condensed as P. grossissima. 11. iens Ron Hiern, Fl. Trop. Africa . Baconia montana Hook. f., J. Linn. Soc. Bot. 7: 196. 1864. /xora hookeriana (Hiern) Kuntze, Revis. Gen. Pl. 1: 287. 1891. TYPE: Cameroon. Southwest Prov- ince: Mt. Cameroon, ca. 2100 m, Dec. 1862 (fl & fr), Mann 2166 (holotype, K). > roc. Pavetta e Breme k., Repert. Spec. Nov. Regni Veg. 37: 73. TYPE: Equatorial Guinea. Bioko Island, Jan. ee (fl), Exe ll 787 (holotype, BM). KEY TO THE VARIETIES OF PAVETTA HOOKERIANA IN CAMEROON l. Leaf veins glabrous, subglabrous, or rarely puber- ulent but not intermittently pubescent : slow ar. hookeriana l. Leaf veins intermittently pubescent below (Fig. 12) var. pubinervata . Pavetta hookeriana var. hookeriana Shrubs or small trees to 10 m. Twiglets glabrous (or subglabrous in some plants from the Oku area), floriferous twiglets 3-21 cm. YT Leaves coriaceous or subcoriaceous, sometimes anisophyllous; blades glabrous, obovate to elliptical, 1.5-14.5 X 0.7-7 cm, veins glabrous to subglabrous or rarely puber- ulent above or below, apex acute to obtuse (less often rounded), acumen if present usually subequal in length and width, 1-9 x 1-9 mm; base cuneate 30050 (to sometimes attenuate), sometimes asymmetrical; midrib and secondary veins sometimes prominulous elow, secondary veins (3-)5-14 each side, occa- sionally joined 2-10 mm from margin, otherwise eucamptodromous; pubescent tuft, crypt, pocket or pit domatia sometimes in branch vein angles o midrib and rarely of secondary veins, sometimes elongated several mm above branch vein angle; nodules absent to fairly numerous, most elongate and close to midrib; fourth and higher order ve- nation seldom clearly visible, usually more visible below than above; venation density fine. Stipules cup-shaped, pubescent internally, glabrous (or at times subglabrous) externally, awn deciduous, lin- ear to cuspidate, 1.5-5 mm. Inflorescences rotund to corymb-shaped in outline or with subunits of these shapes, 0.5-8 cm across, puberulent to gla- sessile, flowers 10-85(-150); sheathing bracts rectangular lobed to unlobed, then cup-(or less often saucer-)shaped, pubescent inside, gla- brous to subglabrous outside, sometimes with fim- briae to 1.5 mm, linear awns 1-4 mm or foliar ap- pendages 5-15 mm; foliar bracts if present 1-5 cm long or resembling slightly reduced foliage leaves; other bracts absent or linear to ovate or quarter- spherical, to ca. 3 mm, with or without 1-several fimbriae to 2 mm, or a fimbria or tuft of hairs borne from axis directly; bracteoles resembling smaller bracts sometimes present. Calyx tube 1-1.5 mm long, 1.5-2.5 mm wide halfway up; lobes valvate, rotund (or less often oblong, ovate, triangular, or pentagonal), 0.7—1.5(-3 brous, ) X 1-2 mm, puberulent or sometimes often reflexed when dry, rim lighter. Corolla white or cream, tube sometimes green or lobes green tipped outside; tube usually cylindrical, 2-5 X 1—2(—4) mm; lobes 4—8 mm. Style clavate, puberulent or subglabrous, exserted 4—8 mm. Stigma sometimes conspicuously 2-lobed. Fruits ca. 8-10 mm across, glabrous or subglabrous, later black. Mature seeds | or 2 (except one empty fruit seen), attached ca. halfway up septum, concave. subglabrous, carinate, bluish green, Additional Te examined. CAMEROON. Southwest Province . Cameroon (fl a I epis 992 (B), (fl bud, fl & fr), Maitland = = TDM 2 (B, K), Mar. (fl), Boughey GC 7005 (K), Deistel 77 (G ep aha 9525 (BM, BR, P. identified also as Brenan 9525), S yis 2326 (YA), Apr. (fl), bi dii GC 6840 (K), MO); around crater Shia dea (f ies 3145 (MO). West Provin : Ds t. Ban m- boutos, June (fr), Saxer 97a, 97b (Ko: Dita Bambou- tos, Nov. (vegetative), Meurillon 496 (HBG), May (fr), Meurillon 1281 (K); Baba djou, Mt. Santa, De. (fl va : & fr), Félix 2792 (HBG, . Northwes vin Santa, Mt. Santa, June (fr), Saker 12 (K); i ela orest Reserve near nursery, Apr. (fl bud), Ujor FHI K); Bafut-Ngemba Forest Reserve, NW face of "Tj ~ Volume 83, Number 1 Mannin g 117 Pavetta Subgenus Baconia in Cameroon Figure 12. with nien and intermittent vestiture along v« Mbakakeka Mt.. Feb. (fl). Hepper 2160 (BR); Bafut-Ngem- ya Forest Reserve. ai ia Mar. (fl). Onochie FHI 34850 (Ky io Nen Forest Rese ve Abaw to Lake 4. Dec. (fl bud. fl & fr). Letouzey 13448 (BR. YA): Oku area. June (fl. Brunt 607 (K). Aug. (fl) IN 805 (P). July (f. Letouzey 8937 (P). Feb. (fl). 1380 (MO). Thomas (MO). Apr. (f). Thomas & Mcleod 6006 b. Pavetta hookeriana var. peer S. D. Manning, var pes i : Buea or Limbe. Mar. 1877 (fl). Kalbreyer 94 (holotype. K, labeled “Y fi Vic- BM, labeled * = part of Buea, top of sheet only). Figure 12. nov. “€ ameroon. west Province toria”: isotype, Cam. Bonjango” A varietate. hookeriana foliis nervis subter interrupte vibes scentibus differt Similar to variety hookeriana except leaf veins intermittently pubescent below. Pavetta hookeriana var. pubinervata has larger leaves than most specimens of variety hookeriana, which usually has small leaves and light bark and often has many short internodes. Other features of variety pubinervata include some leaf blades ovate, some leaves with a few hairs above and below, fourth and higher order venation easily visible be- Paretta hookeriana var. ee ata (Kalbreyer 91. BM). —A. Habit.—B . Details of lower leaf surface low. sheathing bracts sometimes deltoid lobed, pe- dunele absent or ca. 4 mm, calyx tube occasionally only 0.7 mm long or 1.2 mm wide halfway up. and calyx lobes occasionally as short as 0.5 mm. Variety pubinervata is given varietal status main- ly because of the sharp difference in leaf vestiture from any of the numerous collections of variety hookeriana: it is also geographically distinct. Pavetta hookeriana is a montane species occur- ring at elevations of 1500-2800 m (var. hookeriana) or 500 m (var. pubinervata — ) in or at edges of forest. It is the only widespread high elevation species of subgenus Baconia in Cameroon. It occurs only on including Mt. Mt. Oku, and other mountains in Cameroon and the offshore is- lands Pagalu and Bioko of Equatorial Guinea. In Cameroon. volcanoes. Cameroon. this distribution includes locations i eastern Southwest Province, western West Province. Va- riety pubinervata, known only from the type, was collected on Mt. and southern and central Northwest Province. Cameroon. Nodules are often more numerous on plants of this species from more northern locations than from elsewhere. Some larger-leaved specimens of variety hook- g.. Deistel 77 and Onochie FHI 34850) ertana (e. 118 Annals of the Missouri Botanical Garden look similar to P. molundensis, a species of lower elevations occurring further east. However, P. mo- lundensis has more flowers per inflorescence, longer calyx lobes, and larger leaves. Pavetta hookeriana also superficially resembles high altitude speci- mens of P. kupensis but has much shorter styles and usually rotund to ovate, often reflexed calyx lobes longer than the usually somewhat compressed (Fig. 5) ones of P. kupensis. Kalbreyer 94, the only specimen of variety pub- inervata, is probably the same specimen referred to Pavetta owariensis as "Kalbreyer 92” by Hepper and Keay (1963) and referred to Pavetta intermedia Bremekamp as “Kalkbreyer 92” by Bremekamp (1934). The handwritten numeral interpreted as a "4" here, though ambiguous, is clearly different from a “2” in the same handwriting later on the same sheets. 12. Pavetta kribiensis S. D. Manning, sp. nov. YPE: Cameroon. South Province: 25 km ENE of Campo, Mar. 1968 (fl bud & fl), Letouzey 9148 (holotype, P: isotypes, BR, YA). Figure 13 Frutices. Rami glabri. Folia glabra, nervis secundariis Ir utroque 5-12. Inflorescentiae in ambitu rotundatae ad subc orymbosas, 0.5-2 Ss onales, rotundato-compres- rotundati, (vel interdum ipee -l ).7 Corolla tubo 4—6 si vel deltoidei), O.: —].5 mm mm, lobis 4-7 mm, labis interdum super jim prope faucem puberulis. _~ Shrubs to 1 m. Twiglets glabrous, floriferous twiglets 1-19.5 cm. Leaves chartaceous or thinly subcoriaceous, glabrous, sometimes anisophyllous; blades oblong, pes: elliptical (or less often ovate), 2.5-21.5 or rounded, often asymmetrical, acuminate, acumen usually abrupt, 1-27 X 1-9 mm; base attenuate to X 1-7.5 em: apex obtuse to acute cuneate, sometimes asymmetrical; midrib and sec- ondary veins often prominulous below especially toward base, secondary veins sometimes impressed above, 5-12 each side, sometimes joined 1-7 mm from margin; domatia, if present, usually pubescent pockets or tufts in most branch vein angles of mid- rib; nodules scattered on blade, rarely on midrib; fourth and higher order venation usually more ob- vious below; Stipules deltoid to rotund, pentagonal or attenuate lobed or venation density medium. truncate, glabrous; awn sometimes deciduous, cus- pidate to linear, to ca. 10 mm. Inflorescences sub- umbellate, rotund to subcorymb-shaped in outline, 0.5-2 cm across, congested, glabrous, peduncle ab- sent or to 10 flowers (1-)5—40; sheathing bracts unlobed and cup- or bowl-shaped, truncate, or rotund, ovate or narrowly triangularly lobed, gla- brous externally and internally, linear awns if pres- ent to 4 mm, foliar appendages 2-6 mm; foliar bracts if present resembling foliage leaves, some- times reduced proportionately more in length than width; other bracts often obscured by sheathing bracts in congested inflorescences, + ovate, 1—2 mm, without fimbriae but sometimes with a + ob- long, blunt pointed awn ca. 1 mm; bracteoles re- sembling smaller bracts or absent. Calyx tube 1— 1.3 mm long, 1.5-2.5 mm wide halfway up; lobes valvate (or bases at times slightly overlapping), ro- tund (or less often pentagonal, compressed rotund (Fig. 5) or deltoid), 0.2-1 X 0.7-1.5 mm, glabrous, sometimes carinate, rim usually lighter. Corolla white or creamy white; tube usually broadened from base to apex, X 1-3 mm: lobes 4-7 mm, some- times thinly puberulent above near base. Style nar- rowly clavate to fusiform, puberulent to pubescent, exserted 4-8 mm. Young fruits dark green, ca. 9 mm across, glabrous, with persistent calyces. Seeds 2, attached ca. halfway up septum, concave. raa ipe di examined. € AMEROON. South e: 15-25 km SW of Zingui — ca. 45 km SSE of ribi. Mar. (fl bud & fl), Leones 9116 (BR, P); 12 km ENE of Kribi, Lolodorf road, fr), ri 4638 (WAG); 19 km ENE of Kribi, iene as r. (fl), Bos 4137 WAG): ee km ENE of Kribi, Lolodorf ded Jan. (fl), Bos 6202 ( p Pavetta kribiensis is endemic to western South Province, Cameroon. It occurs in wet forest, some- times along streams. he distal, nonsheathing bracts of Pavetta kri- biensis are unusual in sometimes having a more or less oblong, blunt-pointed awn rather than the thin fimbriae found in most species Pavetta kribiensis resembles P. staudtii vegeta- tively and in nodule patterns. It is distinct from P. staudtii in having rotund to deltoid calyx lobes rather than usually subquadrate (Fig. 5) ones, as well as in its more condensed inflorescences and nonfimbriate bracts. 13. Pavetta kupensis S. D. Manning, sp. nov. TYPE: Cameroon. Southwest Province: W side > Fig Pavetta kribiensis (A, B, Letouzey 9148, P; C,D, Bos 4638, WAG).—A. Habit including bacterial nod- ules. “e i of inflorescence with open flower, bracts covering inflorescence branches for most of their lengths, awns on end Domatium on lower leaf surface.—D. Fruit. Volume 83, Number 1 Manning 119 1996 Pavetta Subgenus Baconia in Cameroon 120 Annals of the Missouri Botanical Garden of Mt. Kupe, Feb. 1986 (fl), Thomas & McLeod 5481 (holotype, MO; isotypes, BR, K, P, WAG not seen, YA not seen). Figure 14. rutices. Folia laminis glabris, valde obovatis vel ob- lanceolatis (ad interdum e elliptic as vel oblongas), apice ob- secundariis utroque 5-9; dom- ryptis inte ndum puberulis secus costam et interdum in MALA externis nervorum secundariorum et aliorum nervorum sec 'undariorum vel tertiariorum (ve | raro nbus vel in see subro- sas dh visae, 2-9 cm latae. . triangularo-compressi ad rotundato- CE (ad qu rotundatos vel non profunde 2-lobulatos), 0. 1— l -2 mm, glabri ad subglabros. Corolla tubo d mm bu lobis 8-12 mm longis. Styli exserti 11—20 mn obi ca- Shrubs to 4 m. Twiglets glabrous, floriferous twiglets 2-11 cm. Leaves coriaceous, glabrous, sometimes anisophyllous: blades strongly obovate or oblanceolate (to less often elliptical or oblong), 3-19 X 1-7.5 em; apex obtuse to rounded, usually with acumen 3-10 X 2-8 mm; base cuneate to at- tenuate, often slightly asymmetrical; midrib prom- inulous below toward base; secondary veins 5-9 each side, joined 1-10 mm from margin, their con- nections often conspicuous to the naked eye; dom- atia usually crypts or pits along midrib and some- times in external angles of branching secondary veins, sometimes puberulent; nodules few, scat- tered on blade and sometimes on midrib; fourth and higher order venation obscure, sometimes slightly more obvious above; venation density fine. Stipules deciduous, rotund lobed to unlobed and then sau- cer- or cup-shaped, pubescent internally, glabrous externally, with cuspidate, faleate awn ca. 1-3 X mm. Inflorescences subrotund to corymb-shaped in outline or with subunits of these shapes, 2-9 cm across, glabrous or subglabrous, sessile, flowers 25-75; sheathing bracts triangular, rotund or com- pressed rotund lobed to unlobed and then saucer- or cup-shaped, pubescent internally, glabrous ex- ternally, with triangular, ovate, or linear awns 1-3 X 0.5-1 obovate, or triangular, occasionally carinate, to 3 mm; other bracts linear, broadly ovate, mm, sometimes with |—several fimbriae 0.5-2 mm or fimbriae borne from axis directly; bracteoles re- 1-1.5 long. 2-2.5 mm wide halfway up: lobes valvate. sembling smaller bracts. Calyx tube mm short triangular to compressed rotund (or less often — rotund), occasionally shallowly 2-lobulate, 0.1—1 1-2 mm, glabrous or subglabrous, sometimes car- inate, rim not or narrowly lighter: corolla in bud greenish or white with green vertical stripes, in open flower greenish white; tube cylindrical or sub- cylindrical, (4—)6-8 X 1-2 mm; lobes 8-12 mm. Style narrowly clavate, puberulent, exserted 11-20 mm. Additional specimens examined. CAMEROON. Southwest Province: between summits Kupe, Nov. (fl bud), Letouzey 450 (BR); Mungo River valley near Konye, 40 km N of Kumba, Dec. (fl bud), Thomas & Nem- ba 5197 Pavetta kupensis is endemic to Cameroon. It has been found only in forest on and approximately 30 km west of Mt. Kupe in eastern Southwest Prov- ince, at elevations from 300 to 2000 m. The low- land specimen has larger leaves than other collec- tions. The chief distinguishing features of Pavetta ku- pensis are that most leaves are usually strongly ob- ovate and apices never acute. Styles are exserted farther than in most spe- D. Man- ning, which also has much longer calyx lobes and brochidodromous with cies, but not as far as in P. longistyla S. larger inflorescences than P. kupensis and lacks domatia. Comparisons to other species follow the descriptions of P. brachycalyx and P. hookeriana. Collection of more material may reveal interme- diates sufficient to combine Pavetta kupensis with one or more of its congeners. It is not obviously closer to any one of them than the others. 14. Pavetta lasioclada (K. Krause) Mildbraed ex Hse Repert. Spec. Nov. Regni Veg. : 62. 1934. Chomelia e vane K. Krause, ce Jahrb. Syst. 34: 9. TYPE: Togo. Sokode, 1904, Kersting p ee B de- stroyed). LECTOTYPE: Sierra Leone. Falaba, Mar. 1892 (fl bud), Scott Elliot 5142 (lecto- type, selected here, K; isolectotype, BM). Pavetta baconia Hiern pea aiios Scott dein J. Linn. Soc., Bot. 30: 83. "YPE: Sierra Leone. Near Falaba, Mar. 1892 ^ bud). Scott “Elliot 5412 (= 5142, K. BM?). eer vg cig K. Kra TYPE: Cameroon. al: Pass Tehape. Mar. estroyed). Pavetta buries A. Chev., Explor. Bot. Afrique Occid. Franç. 1: 334. 1920. No holotype designated; spec- imens cited include one from Frenc de m (1899) and eight from French Guinea (all 1 og Pavetta tisserantii Bremek., Repert 37: 63. 1934. Syn. nov. TYPE: pub lic. Balaougu, 17 km N of Bamber Mar. 1925 (fl), Tisserant 1865 (P use, Bot. Jahrb. Syst. 40: 419. oo Province, prob- 1909 (fl), Ledermann 2835 egni Veg. Shrubs or small trees to 5(-10) m. Twiglets pu- bescent especially finally, floriferous twiglets 6.5— 26 cm. ceous), sometimes anisophyllous; blades obovate to elliptic or ovate, 4.5-2! 2-11 cm, pubescent Leaves chartaceous (to less often coria- Volume 83, Number 1 Manning 121 1996 Pavetta Subgenus Baconia in Cameroon Figure 14. Pavetta kupensis (Thomas € Mcleod 5481, MO).—A. Habit.—B. Details of lower leaf surface with conspicuous submarginal veins, obscure higher order venation, and domatia.—C. Flower. 122 Annals of the Missouri Botanical Garden below, less densely pubescent to subglabrous above, veins pubescent below, pubescent to puber- ulent above; apex acute to obtuse or rounded, acu- men 2-15 X 2-13 mm (or less often absent); base cuneate to obtuse (or less often rounded or trun- cate), sometimes asymmetrical; midrib and second- ary veins de or prominulous below, second- ary veins 7-15 ea mm from margin; pocket domatia sometimes or tuft domatia rarely present in branch vein angles of midrib (or occasionally other veins); nodules un- side, sometimes joined 1—8 common, along midrib or scattered on blade; some ertiary veins prominent or prominulous below; fourth and higher order venation invisible to obvi- ous; venation density medium to fine. Stipules cup- shaped, pubescent externally, glabrous or subgla- brous internally; awn falcate or linear, 2-10 mm Inflorescences rotund to pyramidal, subpyramidal or corymb-shaped in outline or with subunits of these shapes, 3-17 cm across, lax to congested, pubescent, peduncle absent or ca. 2 mm; flowers 20-200, fragrant; sheathing bracts rotund lobed or unlobed and cup- to saucer-shaped, pubescent ex- ternally, subglabrous internally, linear awns 2-7 mm, foliar appendages 7 mm; foliar bracts 1.5—5 cm or resembling slightly reduced foliage leaves, sometimes proportionally broader compared to length than foliage leaves; other bracts semicircular to ovate or deltoid, to ca. 5 mm, sometimes with fimbria ca. 2 mm or fimbriae borne from axis di- rectly; bracteoles resembling reduced bracts some- times present. Calyx tube 1-1.3 mm long, 2-2.3 mm wide halfway up; lobes valvate, long triangular or ade to ovate, oblong or obovate, 1.5—4 X 1-2 mm, pubescent, not or slightly carinate, rim Ta " Corolla in bud white with green apex or grayish green, in open flower white sometimes quickly becoming rust; tube slightly broadened from base to apex, 4—7 X 1-2 mm; lobes 6-11 mm Style clavate, puberulent to subglabrous, exserted 7-17 mm. Stigma sometimes 2-lobed, lobes to 1 mm. Ovules 4 in 2 locules with 2 ovules attached to a placenta ca. halfway up on each side of sep- tum, or only 2 or 3 ovules developed. Fruits ca. 8 mm across, thinly pubescent or puberulent, green, pale green, whitish green, or glaucous green be coming brown or bluish black. Seeds 4 with 2 at- tached to a placenta ca. halfway up on each side of septum, or only 2 or 3 developed, concave. € En specimens examined. CAMEROON. North Province: Sokorta Tene si ) km ENE of Bélel, Mar. (fl bud), eni 3191 (YA). A oua Province: S of Ngaoundéré, Oct. (fr), Breieler 600 (BR, GH, P, WAG, YA); 17 km S of Meiganga, Nov. (fr), de Wilde et al. 3953 (WAG); Mama, 40 km SE of Meiganga, Oct. (fr), Letouzey 6104 (BR, YA); Boubala road, Tibati area, Dec. (fl bud & fr), Letouzey 2551 (BR, YA). Centre Province: 15 km W of Méting, Matsari-Linté road, May (young as silage 545 (P). East Province: m E of Deng Deng, Jan. Breteler 998 = Letouzey 3328 (same Ei inse (BR, K, WAG, YA). CENTRAL AFRICAN REPUBLIC. Buala, abe led “Kamerun,” 6°25'N, 15%30'E, June (young fr Mildbraed 9559 (K); Bosum Namgebiet, probably zoum, yy “Neu-Kamerun,” Mar. (fl bud & fl), Tess- Y 247 (K); Balaougu, 17 km N of Bambari, Mar. (fl), iil 1865 (P) ET Of the species of subgenus Baconia, Pavetta la- sioclada and P. corymbosa have the most northern and widest distributions. Pavetta lasioclada occurs west at least to Sierra Leone and east at least to Central African Republic in the Guineo-Congolian/ Sudanian transition zone sensu White (1979). Most collections from Cameroon have been in Adamaoua Province, others in the southeastern corner of North Province and northern Centre and East Provinces. It occurs in gallery and young forest and forest bor- orth Province collection is the single known represen- ders in predominantly savanna areas. The tative of subgenus Baconia from either North Prov- ince or Far North Province. Pavetta lasioclada has been reported from 650 to 1500 m elevations in Cameroon. Bozoum, presently in Central African Republic, was in “Neu-Kamerun” when Tessmann 2247 was collected in 1914 (U.S. Office of Geog- raphy, 1962; Ade Ajayi & Crowder, 1985: Chapters 59 and 62, maps). The most distinctive feature of Pavetta lasioclada is that its fruits often bear two seeds per locule. Other aids to recognition are that the midrib and secondary veins are usually impressed above and prominent below; tertiary veins are often prominent below, sometimes even more so than in P tenuis- sima and P. muiriana S. D. Manning; plants are pubescent to puberulent on most parts, though stip- ules and sheathing bracts are glabrous or subgla- brous inside; and inflorescences are large and flow- ers fragrant. Pavetta lasioclada and P. tisserantii are here combined because the three differences used to separate them (degree of vestiture on lower side of leaves, number of secondary veins, and number of internodes of flowering shoots) are not reliable. Du- plicates or single sheets of some individual collec- tions (e.g., Breteler 998 = Letouzey 3328 and Tess- mann 2247) show much of the range of variation encompassing both former species. Scott Elliot’s description of Pavetta lasioclada as P. baconia var. hispida is the earliest for the spe- cies. This was, however, 18 years after the first de- scription of P. baconia by Hiern, including five va- rieties. None of the earlier varieties of P. baconia Volume 83, Number 1 Manning 123 Pavetta Subgenus Baconia in Cameroon belongs to P. lasioclada. Thus, P. baconia is not the correct name for this species. The specific epithet “lasioclada” (Krause, 1909) was the next attributed to this species after “Pavetta baconia var. hispida." The lectotype chosen here was seen by Breme- kamp (1934) along with the later destroyed holo- type when he transferred this species from Cho- melia to Pavetta. It is probably also the specimen Scott Elliot used in his original brief description, though he cited it as 54/2 rather than 5/42—1he locality and description fit. 15. Pavetta laxa 5. D. Manning, sp. nov. TYPE: Cameroon. South Province: near Mékomo, 8 km SW of the confluence of Rivers Dja and Lobo, Mar. 1962 (fl bud & fl), Letouzey 4581 (holotype, YA; isotype, P). Figure 15. utex. Folia laminis super papillatis vel puberulis, is glabris, nervis utrinque puberulis; apice ac uto ad vel in monades corymbosas divisae, 3—4 atae, laxae. Lobi calycini rotundati (ad ic Ran sho, obovate vel ovatos apice rotundato), 1—1.5(-2 l , pubes- centes. Corolla tubo 2-3 mm, lobis 3 mm. Styli e exserti 2— 4 mm. Shrubs ca. 5 m. Twiglets puberulent to subgla- brous, floriferous twiglets 19—20 cm. Leaves char- taceous, blades papillate or minutely puberulent above, glabrous below, major veins puberulent or subglabrous above and below, top pair of leaves strongly anisophyllous; blades elliptical to obovate (or less often ovate, oblong, or rotund) or reduced to a vestige at the top node, 6-15.5 X 4.5-6.5 cm, often asymmetrical; apex acute to rounded with x 4-8 mm; base attenuate (to less or rounded); acumen 7-13 often cuneate, cordate, midrib and secondary veins prominulous below, secondary veins 6-14 each side, sometimes joined 1-5 mm from margin; domatial tufts of hair occasionally ac- companied by pits in branch vein angles of midrib and secondary veins; nodules absent or on blade, very rare; fourth and higher order venation more obvious below than above; venation density medi- um to fine. Stipules sometimes deciduous, cup- shaped, pubescent internally, subglabrous exter- nally, awn cuspidate, ca. 1— ] mm, falcate. Inflorescences corymb-shaped or with corymb- shaped subunits, 3—4 cm across, lax, pubescent to puberulent, sessile, flowers ca. 35; sheathing bracts unlobed and cup- to saucer-shaped or ovate lobed, short pubescent, awns + linear, 0.5-1.5 mm; other bracts ovate, to 2 mm with 1-several fimbriae to 0.5 mm; bracteoles uncommon, resembling bract fimbriae borne from pedicel directly. Calyx tube 0.7-1 mm long, 1.2-1.8 mm wide halfway up: lobes valvate, rotund (or less often oblong, obovate, or x l short pubescent, sometimes carinate, rim lighter. ovate with obtuse apex), ca. 1-1.5(- mm, Corolla white; tube cylindrical or subeylindrical, 2— 3x to subglabrous, exserted 2—4 mm. Stigma with mm; lobes 3 mm. Style clavate, pubescent connivent lobes. Pavetta laxa is known only from the type from South Province, south-central Cameroon near bor- ders with East and Centre Provinces in a lowland, probably forested area. The leaves of Pavetta laxa are similar to those of P. longibrachiata and P. brachysiphon (see dis- cussion following the description of P. brachyst- phon). In. P. laxa the leaves are larger and less narrow than those of P. brachysiphon, however, and corollas are shorter and styles are shorter and ex- serted less than those of P. longibrachiata. Pavetta laxa has very small flowers except for calyx lobes — arger than average in subgenus Baconia. Pavetta laxa is so named for its very lax inflo- rescences. 16. Pavetta longibrachiata Bremekamp. Re- 1934. TYPE: Cameroon. West Province (probably): Ndonge, Ledermann 6179 (holotype, B stroyed). LECTOTY pert. Spec. Nov. Regni Veg. 37: 75. de- PE: Cameroon. East Prov- ince: near Bangué, mouth of Ba * River, B°N, 15?4 E, Feb. 1911 (fl hud). "Mildbraed 4505 (lectotype, selected here, HBG) Shrubs to 5 m. Twiglets shortly puberulent to gla- brous, floriferous twiglets 4-26 cm. Leaves char- — aceous to subcoriaceous, sometimes anisophyllous, top pair sometimes extremely so or reduced to one leaf; blades elliptical to obovate, oblong (or less often ovate or rotund), sometimes asymmetrical, 3.522 X 1.5-7.5 em, (occasionally puberulent or partly pubescent to) normally glabrous or upper surface appearing papillate, major veins puberulent below, puberulent to glabrous above; apex acute (to less often obtuse or rounded), sometimes asymmet- x 2-10 mm; base cuneate (to less often attenuate, rounded rical, subacuminate or with acumen 3-15 or cordate), sometimes asymmetrical; midrib anc sometimes secondary veins prominulous below, secondary veins 4—11(-14) each side, sometimes joined 1-6 mm from margin; domatial tufts, pockets or pits in branch vein angles of midrib and often along secondary veins; nodules absent or few, scat- tered on blade and sometimes on midrib; fourth and higher order venation usually more obvious below than above, occasionally obvious both above and 124 Annals of the Missouri Botanical Garden e l5. Pavetta iin (Letouzey 4581, YA).—4A. Habit.—B. Piedad —C. Inflorescence.—D. Leaf venation and vestiture: top half, upper surface; bottom half, lower surface.—E. Angle of secondary leaf vein below with tuft doma- ium.—F. Stipules below or neither above nor below; venation density medium to fine. Stipules cup-shaped, puberulent to subglabrous externally, pubescent to glabrous in- ternally, awn sometimes deciduous, curved, linear or cuspidate, 1-3 X <1 mm. Inflorescences rotund to corymb-shaped in outline or with subunits of these shapes, 0.5-9 cm across, puberulent to sub- glabrous, peduncle to 45 mm or usually absent, Volume 83, Number 1 Manning 125 Pavetta Subgenus Baconia in Cameroon flowers 15-200; sheathing bracts unlobed and cup-, funnel-, (or less often saucer-)shaped or with rotund to ovate, irregular, or truncate lobes, puberulent or = subglabrous externally, pubescent internally, some- times with linear awns 0.5—5 mm, sometimes with foliar appendages 2-2.5 mm or ovate appendages 0.5-1 mm; foliar bracts usually absent, 2 — mm to size and shape of normal foliage leaves if present; other bracts + broadly obovate, to 4 mm, sometimes with ]-several fimbriae or with a ca. mm awn, or fimbriae or tufts of hairs borne from axis directly; bracteoles resembling smaller bracts. Calyx tube 0.7-1 mm long, 1.2-2 mm wide halfway up; lobes valvate, triangular, ovate, oblong, pentag- onal, rotund, or subquadrate, 0.5-2 X | mm, pu- bescent to subglabrous, usually carinate, sometimes concave, rim narrowly lighter. Corolla white with green apex in bud, white in open flower; most tubes broadened from base to apex, (2338 X 0.5-1.5 mm; lobes 3-9 mm. Style clavate, pubescent to subglabrous, exserted (2-)6-10 mm. Stigma + 2- lobed. Fruits 7-8 mm across, subglabrous or gla- brous, whitish green or ash gray. Mature seeds 2 or l, attached ca. halfway up septum, broadly con- cave. CAMEROON. East Additional spec imens iva Province: 14 km E Jimako. Feb. (fl). Leeuw nr 7785 (BR, HBG, K, MO a WAG, YA); 10 km 5 of Nder ba, Mbang-Ndemba iud. May oP deg 1451 (BR. WAG). South Province: Bitye r Dja, Nov. (fl bud). D 1520 (BR. MO). Feb. d bud & M. Bates 1619 (BM, MO): Mévos-Méla, 32 km dias har Nov. (fr). Le- lorca? 8324 (BR, P). tre Pu ce: 5 km W of Mbal- mavo, Feb. (fl bud), de "Wilde & Wilde- “Dest 1807 (BR, P, WAG); Yaoundé (fl bud), bes 75 . 5), and see Ivory Coast specimens listed below o “origine Yaoundé” foot o {goro at Ap Guéré road E of Yoko. eb. (fl b ud). g Pon . 3428 ( A). Donkor Province: NW side sameroon d Koto, Mar. (fl). Thomas 4494 (MO). IVORY JAST. Centre de Recherches Agronomiques de Binger rales near Bingerville. cultivated. origin Yaoundé, Cameroon, des (fl), Bodard 1353 ( Feb. (young fr). Bodard 1356 (K 32 Pavetta longibrachiata is fairly widespread the East Province and drier forested areas of Centre and has also Although endemic to Cameroon (except for cultivated plants) and South Provinces of Cameroon. been found in the Southwest Province. based on existing collections, it is predicted from its fairly wide distribution and undercollection of neighboring countries that it will be found also south and east of Cameroon. Pavetta longibrachiata occurs in secondary and disturbed forest understory and new growth of sa- The cul- tivated specimens are less robust vegetatively and vanna, mostly at 550—1( m elevations. reproductively than most specimens from Came- roon; they represent a rare recorded attempt to cul- tivate a species of subgenus Baconta. Pavetta longibrachiata is most similar to P. laxa and P brachysiphon, based upon leaf morphology features discussed following the description of P. brachysiphon, and is compared with those species following their descriptions. It can be distinguished from P. cellulosa, which it also resembles, in having nonseptate anthers, and from P. owariensis by its papillate or rarely pubescent upper leaf surfaces, vestiture on twiglets and stipule exteriors, and usu- ally smaller flowers. No duplicates of the original holotype, Leder- mann 6179 (B), have been found. Two paratypes, Waibel 169 (B) and Mildbraed 4505 (B), have also been destroyed; Mildbraed 4505 (MBG) is here se- lected as lectotype. 17. Pavetta longistyla S. D. Manning. sp. nov. CYPE: Cameroon. West Province (probably): “route des Mbos; Mar. 1967 (fl), Meurillon 645 (holotype, P: iso- type, BR). Figure 16. probably near Dschang. rulex. Rami glabri. Folia glabra. yr rotundato. ad Freee nervis secundariis utroque 5 omatiis nullis Inflorescentiae in ambitu rotundatae ui hes icd: 9-13 cm triangularos, mm, latae. L obi calycini nai t s ad ovatos vel 1.5-3 x (0.7—)1— lobis 9-14 mm. Styli exserti 20-: 30 ined tubo 6—10 gi Shrubs. 15-18 cm and probably longer. Leaves chartaceous Twiglets glabrous. floriferous twiglets to subcoriaceous, glabrous, not anisophyllous: blades obovate (to less often elliptical), 5.5-19 X 2-9.5 cm: apex rounded to obtuse with acumen 2- 15 x 2-12 mm: base cuneate to attenuate, some- times asymmetrical; midrib prominulous below near base, secondary veins 5-8 each side, some- times joined 2-9 mm from margin; domatia absent: nodules absent or along origins of some branch veins along midrib (and less often secondary veins) or in branch vein angles of midrib, fourth and high- er order venation usually slightly more easily visi- ble above but obscure on both sides: venation den- sity medium. Stipules deciduous, cup-shaped, pubescent internally near base, glabrous externally: 2-5 mm. Inflorescences rotund to subcorymb-shaped in awn cuspidate linear, sometimes falcate, outline, 9-13 em across, subglabrous distally, gla- brous proximally, sessile, flowers 40-60: sheathing bracts sometimes deciduous, at least sometimes ro- tund lobed, pubescent internally, subglabrous ex- ternally, sometimes with linear awns or foliar ap- pendages to 3 mm; foliar bracts if present 10-25 1-3 mm, sometimes with mm: other bracts ovate, 126 Annals of the Missouri Botanical Garden Figure 16. Pavetta longistyla (Meurillon 645, P).—A. Habit.—B. Flower and corolla opened to show beard.—C ) and D. Leaf bases with obscure higher order venation and lack of domatia. several fimbriae to 2 mm, often absent at upper nodes of inflorescences; bracteoles absent. Calyx tube 1.5-2 mm long, 2-3 mm wide halfway up; lobes valvate, deltoid to pentagonal, rotund (or less commonly oblong, obovate, or ovate), 1.5-3 X (0.7—) 1-2 mm, subglabrous to glabrous, usually carinate, rim lighter. Corolla white, subcylindrical or cylindri- cal, tube 6-10 X 2-3 mm; lobes 9-14 mm. Style narrowly clavate, puberulent, exserted 20-30 mm. Ovules in open flower 2, white, + ovoid, apparently attached ca. halfway up septum. Pavetta longistyla is known only from the type collected near the western boundary of West Prov- Volume 83, Number 1 Manning 127 Pavetta Subgenus Baconia in Cameroon ince, Cameroon, in forest beside a stream at an altitude of 1100 m. Pavetta longistyla resembles taxa of subgenus Pavetta in that it lacks bracteoles and its style- pollen presenters are exserted 20-30 mm, further than in any other species of subgenus Baconia in Cameroon. Its bearded corolla throat places it in subgenus Baconia, however. Most leaves are broadly obovate and, as in Pav- etta. kupensis, leaf apices are obtuse or rounded, never acute as in most species. Pavetta longistyla’s large. lax inflorescences, lack of domatia, and less strongly brochidodromous venation distinguish it from P. kupensis. It resembles P. mollissima Hutch. & Dalziel, a western African species, vegetatively but is distinct from it in having less vestiture, larger flowers, and laxer inflorescences. It resembles P. viridiloba K. Krause and P. molundensis in inflo- rescence shape but can be distinguished from P. viridiloba by its smaller, usually obovate leaves and ack of vestiture and from P. molundensis by its larger, fewer-flowered inflorescences and obscure leaf venation. 18. Pavetta molundensis K. Krause, Bot. Jahrb. Syst. 57: 39. 1922. TYPE: Cameroon. East Province: ca. 15%22'E, 3°27'N, Bundi, former- ly in Bezirk Molundu, near Yokadouma and Neola, Mar. 1911 (fl bud), Mildbraed 4673 (ho- lotype, B: reported as Mildbraed 6473, pre- sumably in error, destroyed; lectotype. select- ed here, HBG) Pavetta insignis Yao Repert. Spec. Nov. Regni Veg. 37: 65. 1934. SYNTYPES: Uganda. E. Brown 217 (K): Mildbraed pr (B destroyed). Shrubs to 6 m. Twiglets glabrous, floriferous twiglets 3-27 cm. Leaves coriaceous or subcoria- ceous, sometimes anisophyllous; blades elliptical to obovate (or less often oblong or ovate), 3—30 X l- 12 cm, blades at times subglabrous below): apex acute (or sometimes asymmetrical, glabrous (or less often obtuse), sometimes asymmetrical, sub- acuminate or with acumen 5-15 X 4—10 mm: base cuneate to attenuate; midrib prominulous below to- ward base, secondary veins 6-11 each side. eu- camptodromous; domatia absent or small pits, crypts, or pockets present in some branch vein an- gles of midrib and less often of secondary veins; nodules absent; fourth and higher order venation + equally obvious above and below; venation density fine. Stipules usually deciduous, cup- or saucer- shaped, pubescent internally, glabrous or subglab- mm. rous externally, awn linear, ca. Inflores- cences + pyramidal, corymb-shaped or rotund in outline, 3-9 cm across, pubescent to puberulent distally, glabrous to puberulent proximally, sessile. flowers 100-350; sheathing bracts cup- or saucer- shaped or ovate lobed, pubescent internally, gla- brous to puberulent externally, with 1—several lin- ear awns to 3 mm or foliar appendages 2-5 mm: foliar bracts if present 1.5 cm or resembling less reduced foliage leaves; other bracts ovate, linear (or 3 mm, at times subquadrate or rotund), to some- times with 1—several fimbriae to 1 mm or rarely an awn ca. 2 mm, or fimbriae borne from axis directly; bracteoles resembling smaller bracts sometimes present. Calyx tube 0.7—1.8 mm long. 2-3 mm wide halfway up; lobes valvate (or less often bases slight- ly overlapping), rotund (to less often oblong. ovate. obovate, long-pentagonal, deltoid, subquadrate, o1 occasionally 2-lobulate owing to either an emargin- 1-3 not or inconspicuously ate apex or a cleft near base), 1-2 mm, pubescent, carinate, rim usually lighter. Corolla white; tube subcylindrical .5—2( 7-8(-11) mm. Style clavate, glabrous or subglabrous, exserted (5-)8-12 mm. Fruits 7-10 mm across, sub- or cylindrical, 3-5 —3) mm, lobes (5—) glabrous, blue-gray or whitish. Seeds 2, attached ca. 24 of the way up septum, concave. Additional specimens examined. (CAMEROON. East Province: 5 km E of Bertoua, Dec. (fl), Breteler 833 (BR, GH. P, WAG, YA): 50 km from Bertoua a Abdumad- jali. Dec. (fl bud and ke Nana 464 (BR. P. YA): 75 km N of Bertoua toward Deng Deng, Dec. (fl En Nena 386 (BR. P. YA); near Deng Deng, Oct. (fr), Nana 302 E YA); near confluence of Lom and Djérem Rivers, ca. 250 km NE of Yaoundé, Mar. (fl bud and fl), Mildbraed 8616 (K). nm A Most of the range of Pavetta molundensis is fur- ther east and south than Cameroon, including Su- dan, Zaire, Uganda, Tanzania (Bridson, 1978; Brid- & Verdcourt, 1988) and Central African Republic (Bridson, pers. comm.). It is thus centered son in the Congolian centre of specific endemism sensu White (1979). only in East Province. It grows in closed and open forest and savanna, including on pebbly soil and In Cameroon it has been collected on laterite. Pavetta molundensis is similar to P. calothyrsa and P. robusta except in typical calyx lobe shapes and in that it lacks bacterial nodules. Aids in rec- ognizing P. molundensis are corolla lobes averaging much longer than corolla tubes and fruits with seeds attached ca. two-thirds rather than halfway up the septum. It is glabrous or nearly so vegeta- tively, except for stipule interiors. Vestiture is pres- ent on inflorescences, however, increasingly so to- ward the apex. culminating in pubescent calyx 128 Annals of the Missouri Botanical Garden lobes. Several collections of P. molundensis have dried nearly blac Pavetta hookeriana and P. molundensis are com- pared following the description of P. hookeriana. The original description of Pavetta molundensis lists the type as Mildbraed 6473 rather than 4673. However, locality and habitat data, date, and fea- tures of 4673 match those of the description. That description is referred to on the HBG isotype of Mildbraed 4673 selected here as lectotype. 19. Pavetta mpomii 5. Manning, sp. nov. YPE: Cameroon. South Province: Nkolemen- long hill W of Ebianemeyong, near Nyabéssan, 60 km E of Campo, Apr. 1970 (fl bud & fl), Letouzey 10341 (holotype, P; isotypes, BR, YA). Figure 17 Frutices. Folia laminis super parce pubesc entibus, sub- s, nervi mm. Styli exserti 9— Shrubs to 2 m. Twiglets pubescent, floriferous twiglets 5.5-19.5 cm. Leaves chartaceous to sub- coriaceous, veins pubescent or puberulent above and below, blades thinly pubescent above, glabrous or thinly pubescent below, sometimes anisophyl- lous; blades obovate to elliptical, oblong (or less often ovate), 3-22 X 1-9 cm; apex acute (to less often obtuse), usually with acumen 5-20 X 3-10 mm; base cuneate to obtuse (or less often attenuate or rounded), often asymmetrical; midrib prominu- lous or prominent below especially near base, sec- ondary veins prominulous below, 4-13 each side, eucamptodromous (or less often brochidodromous); domatia absent or tuft or hairy pocket domatia in some branch vein angles of midrib; nodules if pres- ent few, on midrib or secondary veins (or occasion- ally scattered on blade); fourth and higher order venation more obvious below but often obscure be- low also; venation density fine. Stipules unlobed and cup-shaped (or less often compressed rotund lobed), pubescent internally and externally, awn cuspidate (or less often linear), (1-)3-7 mm. Inflo- rescences corymb-shaped to rotund in outline, 3— 7(-12) cm across, pubescent, sessile or with pe- duncle to 4 mm, flowers 20-125; sheathing bracts compressed rotund lobed or unlobed and bowl- shaped, pubescent, sometimes with foliar append- ages 14 mm long or linear awns 1-2 mm long; foliar bracts often present; other bracts linear, ovate or obovate, 0.5—1(—3) mm, sometimes with 1-sev- eral fimbriae ca. 1 mm or fimbriae or tufts of hairs borne from axis directly; bracteoles resembling smaller bracts. Calyx tube 0.5-1.5 mm long, 1.2— 2 mm wide halfway up; lobes valvate or bases over- lapping, ovate (to less often oblong, pentagonal, ro- tund, or triangular), 1.2- mm, pubescent, sometimes carinate, rim lighter. Corolla white; tube widened from base to throat, 5-8 X mm; lobes 8-13 mm. Style narrowly clavate, thinly pubescent to subglabrous, exserted 9-11 mm. Stigma occa- sionally visibly 2-lobed, lobes ca. 0.25 mm. Pre- fruiting ovules 2, + reniform, attached ca. halfway up septum. Additional specimens examined. CAMEROON. South Province: 10 km ESE of Campo, Mar. (fl bud), Letouzey 9209 (P); SE slopes of Mt. Eléphant est. 10 km SE of Kribi, Feb. (fl), Bos 6419 (WAG); Bipindi (fl), Zenker 4569 (BM, K, MO); Lolodorf (fl), Staudt 27 (S). — "n Pavetta mpomii is endemic to western South Province, Cameroon, a high rainfall area. The only elevation reported is 200 m, on a slope of Mt. Élé- phant. Pavetta mpomii is similar to P. viridiloba and the more southern P puberula Hiern. Inconspicuous tertiary leaf venation and green calyx lobes on dry specimens characterize P. mpomii and P. viridiloba and distinguish them from P. lasioclada and P. na- matae S. D. Manning. Vestiture in Pavetta mpomii is less than that of P. viridiloba, coming closer to P. puberula in that respect. Leaves and inflorescences are usually smaller than in either of the other two species. Ca- lyx lobes of P. mpomii are intermediate in shape. These three species are further compared following the description of P. viridiloba. Pavetta mpomii is named in honor of M. M. Be- noit Mpom, formerly of the National Herbarium of Cameroon, Yaoundé. 20. Pavetta muiriana S. D. Manning, sp. nov. : Cameroon. Southwest Province: Bak- ossi Mountains W of Bangem, Jan. 1986 (fl bud & fl), Thomas & Mcleod 5303 (holotype, MO; isotypes, BR, FHI not seen, K, P, PRE not seen, WAG not seen). Figure 18. Frutices. Folia costa et biis secundariis subter pro inentibus, saltem versus bases, nervis secundariis iPad (811-16; domatiis nullis; odoli nullis vel secus cos- tam et raro nervo c calycini deltoidei (ad interdum rotundato iP iind ro- tundati vel breviter triangulares), 0.2-1 X 0.5-1 mm, pu- bescentes. Corolla tubo 3-5 mm; lobis 4—6 mm, lobis su- Volume 83, Number 1 Manning 129 Pavetta Subgenus Baconia in Cameroon re 17. Pavetta mpomii (Letouzey 10341, P).—4. Habit.—B. Lower leaf surface with vestiture both on veins C. Flower. ¡gu and lamina. 130 Annals of the Missouri Botanical Garden Figure 18 with bacterial nodule along midrib.—D. Part of lower leaf surface.—E. Stip per prope faucem pubescentibus. Anthera atra, albo-vittata. Styli exserti 3-6 mm. Shrubs to 2 m. Twiglets pubescent, floriferous twiglets 5.5-10.5 cm. Leaves chartaceous, not an- isophyllous; blades elliptic to obovate (or less often oblong), (1-)7-18 X (123338 cm, blades glabrous, major veins pubescent to subglabrous; apex acute, usually with acumen 3 X 2-6 mm; base cu- neate to obtuse, often asymmetrical; midrib and Pavetta muiriana (Thomas 5449, MO).—A. Habit.—B. Inflorescence.—C. Part of upper leaf surface Stipule. secondary veins prominent below at least toward base, secondary veins (8—)11—16 each side, joined 1—5 mm from margin; domatia absent; nodules ab- sent or along midrib and rarely secondary veins, inconspicuous because of vestiture; tertiary veins prominent below; fourth and higher order venation obvious, usually slightly more so above; venation density fine to very fine. Stipules deciduous, pu- berulent internally and externally, + compressed Volume 83, Number 1 Manning Pavetta Subgenus Baconia in Cameroon 131 rotund lobed, awn linear, sometimes falcate, 2-5 Inflorescences pyramidal to subrotund (1-)2-4 + lax, pubescent, sessile, not subumbellate. mm or probably longer. or irregular in outline, cm across, flowers 20-35; sheathing bracts ovate lobed or un- lobed, then bowl- or saucer-shaped, pubescent ex- ternally. glabrous to pubescent internally: awns lin- ear, 2-9 (or less often <2) mm; other bracts broadly to narrowly ovate, 0.5-2 mm, usually not fimbriate, rarely with several fimbriae to ca. 0.5 mm: brac- teoles resembling smaller bracts. Calyx tube 0.5-1 mm long, 1.5-2 mm wide halfway up: lobes valvate. deltoid, subdeltoid (or less often compressed ro- tund, rotund, or short triangular), 0.2—1 0.5-1 mm, pubescent, sometimes carinate, rim not or nar- rowly lighter, vestiture of margins appearing lighter. Corolla white or yellow, tube cylindrical, broadened from base to throat (or less often obovate in outline). 3-5 X 1-2 mm, lobes 4-6 mm, pubescent above near throat. Anthers drying black with white stripes. Style clavate, puberulent to pubescent, ex- serted 3-6 mm. Prefruiting ovules 2, + reniform, attached ca. halfway up septum. CAMEROON. South- p aet specimen examined. 2 Kumba, Jan. (fl). ^?rovince: Lake Barombi Mbo, Thomas 5449 (MO). z Pavetta muiriana is known from the eastern half of Southwest Province, Cameroon. Its two collection localities are separated by ca. LOO km. One is near a volcanic crater lake at an elevation of ca. 300 m. the other is on a hillside at 1500 m. The higher elevation collection had yellow corollas, the lower had white ones. Pavetta muiriana is compared to P. tenuissima, a related species, following the description of the latter. The inflorescences and anthers (black with white stripes in open flower on herbarium sheets) of Pav- etta muiriana resemble those of P. neurocarpa Ben- tham. Corollas of P muiriana are similar to those of P. camerounensis, P. rubentifolia, P. neurocarpa. P. urophylla, and P. tenuissima and differ from other species in being pubescent on the upper side of Nonsheathing bracts of P. mui- P. neu- lobes near throats. riana are similar to those of P. kribiensis, rocarpa, P. tenuissima, and P. urophylla and differ from other species in usually lacking fimbriae such as those in Figure 4. Pavetta muiriana is named in honor of William Muir, former professor at Carleton College, North- field, Minnesota, U.S.A., thor’s interest in botany while he was an under- who developed the au- graduate student, and his wife, Elizabeth, for her outstanding support of his work. 21. Pavetta namatae S. D. Manning, sp. nov. ^: Cameroon. Unknown location: Mar. 1918 (fl), Gocker 18 (holotype, MO: isotype, US). Figure 19. Frutex. Folia nervis secundariis utroque (5—)10—15: venis tertiariis interdum subter prominulis. Inflorescentiae sub rotundatas vel in monades cm latae, in ambitu corymbosae ad : corymbosas vel ss 'orymbosas divisae, 6-11 plus minusve lax obi calycini vabvati, rotundati M" inte ip rivis y ovalc 0.5-1 Corolla tubo 5—7 exsertl 8-12 mm əs vel triangulares). 0.5 mm. lobis 6—7 mm. Styl 3 Shrubs ca. 6 m. Twiglets glabrous. floriferous twiglets 7-12 cm. Leaves chartaceous to coria- ceous, blades and veins glabrous above, glabrous to puberulent with minute hairs below, sometimes anisophyllous; blades obovate to elliptical, 4-21 X 5-8 em: apex acute to obtuse, often asymmetrical, 2-10 mm: neate to attenuate, sometimes asymmetrical; midrib — usually with acumen 3-20 X base cu- prominent and secondary veins prominent or prom- inulous below, major veins sometimes impressed above, secondary veins (5—)10-15 each side, usu- ally joined 2-10 mm from margin; domatia present as hairy pockets in some or all branch vein angles of midrib, secondary and sometimes tertiary veins: nodules scattered on blades; tertiary veins some- times prominulous below: fourth and higher order venation more obvious below than above; venation density fine to very fine. Stipules sometimes decid- uous, shallowly cup- to saucer-shaped, pubescent internally, glabrous (to occasionally puberulent) ex- ternally, awn cuspidate, 3-5 X + falcate. Inflorescences corymb-shaped to subrotund in out- with corymb-shaped or subcorymb-shaped sub- mm, line or subunits, 6—11 em across, sometimes + lax, glabrous to glabrous, peduncle absent or to 3 mm: flowers 40-200: sheathing bracts unlobed and cup- or bowl-shaped or obovate to compressed rotund lobed. pubescent internally, glabrous or subgla- brous externally, usually with awns or other ap- pendages, occasionally with fimbriae; awns + lin- ear. 1-2 mm: other appendages foliar, 1-5 mm or ovale, ca. 3 X 0.5-0.8 mm: ] mm; foliar bracts if present ca. 6 mm or resembling fo- liage leaves; other bracts obovate, subquadrate, lin- fimbriae to ca. ovate, to 2 mm, sometimes with l—several » | mm or fimbriae borne from axis di- ear Or fimbriae t rectly; bracteoles resembling smaller bracts. Calyx tube 0.7-1 mm long, 1.5-2 mm wide halfway up: rotund (or less often pentagonal, lobes valvate, ovate, or triangular), 0.5—1 lent to subglabrous, usually carinate, rim lighter. Corolla tube constricted near base, broadened from constriction to throat, 5-7 X 0.5-1.5 mm; lobes 6— x 0.5-1 mm, puberu- 132 Annals of the Missouri Botanical Garden 9. Pavetta namatae (Gocker 18, MO).—A. Habit.—B. Part of inflorescence.—C. Inflorescence.—D. Node Figure | with stipule base glabrous outside, hairy inside, Volume 83, Number 1 1996 Manning 133 Pavetta Subgenus Baconia in Cameroon 7 mm. Style clavate, glabrous to puberulent, ex- serted 8-12 mm. Stigma sometimes with 2 lobes ca. 0.5 mm. Immature fruits subglabrous. Immature seeds 2, attached ca. halfway up septum, + reni- form. Additional ier Ser Oia CAMEROON. Centre rovince: 7 E of Makak, June (young fr), Manning 2097 (MO). Pavetta namatae is endemic to Cameroon. The only known collection locality is in south-central Centre Province Pavetta namatae has leaf venation very similar to that of P. lasioclada, including tertiary veins of- ten. prominent. below. Many other features. differ. however. These include leaf size, number of ovules per locule in fruits, calyx lobe features, and vesti- ture of stipules, sheathing bracts, leaves. inflores- cences, and calyx lobes. Pavetta namatae resembles P. owariensis except in leaf venation pattern. Tertiary and higher order venation of P. owariensis are not prominent below Also, there are 4—10 pairs of (5-)10-15 as in P. namatae. secondary veins in P. owariensis, pairs in P. namatae. Pavetta namatae is named in honor of Ferdinand Namata of Makeke Camp near Mundemba, South- west Province, Cameroon. His knowledge of the flo- ra and fauna of Cameroon, his efforts to conserve them, and his field guidance to researchers are here recognized. 22. Pavetta neurocarpa Bentham. Niger Fl. 849. Ixora neurocarpa (Benth.) Kuntze, Revis. Gen. Pl. 1: 287. 1891. TYPE: Equato- rial Guinea. Bioko Island, formerly Fernando Po. Nov. (18417) (fr). Vogel 151 (holotype, K). Pavetta mannii Hiern, Fl. Trop. Africa 3: 169. 1877. Ixora mannii Bs rn) Kuntze, Revis. Gen. Pl. 1: 287. 1891. TYPE: Cameroon. Probably Southwest Prov- ince: ouis River or Ambas Bay, Dec. 1862 (fl). ann s.n. (Ky Paséita ans Hute h. & Dalziel. FI. W. Trop. Africa 2: 91. 1931. Syn. nov. TYPE: dee Oban District: 1911-1912 (fl), Talbot & Talbot s.n. Shrubs to 4 m. Twiglets glabrous. floriferous twiglets 2-29 cm. Leaves coriaceous to subcoria- ceous. sometimes anisophyllous; blades elliptical to oblong or obovate, (3-)7-35 X (1-)5-13 cm, gla- rounded to obtuse, acute (or rarely x 2-12 mm: sometimes strongly brous; apex emarginate), acumen if present 5-30 attenuate, base cuneate to asymmetrical; midrib and secondary veins some- times prominent or prominulous below, secondary veins (5-)8—16 each side. usually joined 2-15 mm from margin; domatia absent or occasionally pres- ent as small, sometimes hairy pockets in some branch vein angles of midrib; nodules scattered on blade, occasionally along midrib: fourth and higher order venation obvious or obscure above, much or slightly less obvious below; venation density fine. Stipules deciduous, remains apparently rotund to diamond or trapezoidally lobed, glabrous, awn de- ciduous or apparently absent. Inflorescences cor- ymb-shaped to subrotund, pyramidal or irregular in outline or with subunits of these shapes, 0.5-13 cm across, lax to congested, glabrous (or less often sub- glabrous), peduncle absent or to 20 mm, flowers 3— 200: sheathing bracts ovate lobed to unlobed and then cup- or saucer-shaped, sometimes subtruncate near apex, glabrous, sometimes with foliar append- ages ca. 1 mm, linear awns 0.54 mm, or ovate awns 0.54 X ca. 1 mm: other bracts ovate (or less often subquadrate), concave, 0.5-2 mm, usually not fimbriate, rarely with several fimbriae or an awn to ca. 0.5 mm; bracteoles resembling smaller bracts sometimes present. Calyx tube 0.7—2 mm long, 1.2— 2 mm wide halfway up: lobes valvate or almost overlapping at base, compressed rotund (to less of- ten rotund, pentagonal, subquadrate, short trian- gular, or truncate), 0.2— 8) X | mm, glabrous (or less often subglabrous), sometimes carinate, rim lighter. Corolla white or green-white. tube cylindri- (0.5—)2—4(—7) —3) mm: lobes 3-8 mm, pubescent above cal or broadened from base to throat. x (0.5-)1-2( near throat or lobe vestiture sometimes nearly absent but vestiture present on filaments near their points of insertion: anthers usually drying black with white stripes: style fusiform, usually narrowly so, pubescent —15) mm. Fruits ca. T or puberulent, exserted (3—)6-1 1 6-10 mm across, glabrous or subglabrous, white sometimes with green vertical stripes. Mature seeds 2, attached ca. halfway up septum. concave. Additional specimens examined. CAMEROON. Southwest Province: Cameroon River or ev Bay, location uncertain, Dec. (fl). Mann. s. e (K, P); 12 km NW E of Limbe, Feb. (fl), Suid 651 of Batoke, 3 a Idenao ; path da 4 km W ier has of Mt. Cameroon, Dec. (fl), Thomas et al. 512 O), Dec. (fr). Thomas et al. 5121 (MO); S slope of i Cameroon above Batoke, Dec. (fl bud), Thomas 2770 in part (MO), the remainder being Pavetta rigida Hiern and designated 2770A; NE of . Victoria = L Es igs (fl). Keay FHI 37533 (K): Koop Na Park, fl), Thomas a (K. MO. P. YA), Apr. (young fr), Manning 1711 (MO), 1769 (MO): km g of Mundemba, Nov. (fl bud & fr), Kanma 1035 fing S shore of Lake Barombi Mbo, 5 km NW of Kumba, A Manning 1779 Sar June (fr), Dee 2060 uo uth Province: hill of Nkoltsia near Mi 18 km NW "of Bipindi. Feb. (fl), Villiers 794 (P). NIGERIA. Oban District: (fl), Mr. and Mrs. P. A. Talbot s.n. (K). — w — T = — — Most collections from Cameroon are from mature, 134 Annals of the Missouri Botanical Garden fully shaded, very wet forests in Southwest Prov- ince, Pavetta neurocarpa has also been found in wet forest in western South Province. Outside Came- roon, it has been reported only from nearby Bioko and southeastern Nigeria, where the plants formerly referred to P. mannioides are centered. It is a Lower Guinean species sensu White (1979). It has been collected near a lake and reported to be fragrant. Substrates reported include volcanic lava soil and rock at a hill’s summit. Elevations reported are from 50 to 800 m, mostly toward the lower end of the range. Pavetta neurocarpa is characterized by glabrous inflorescences usually drying black with strongly concave, sometimes almost boat-shaped bracts, which are glabrous internally. Occasionally, corolla tubes are shorter than 1 mm, though normally they are 2-4 mm long. Relatively unusual features P. neurocarpa shares with a few other species follow the description of P. muiriana. Bremekamp (1934) and others maintained a sep- arate species, Pavetta mannioides, for specimens with smaller, contracted inflorescences with smaller and fewer flowers than P. neurocarpa. Recent col- lections (Satabié 651 and Thomas 4322) show vari- ation sufficient to combine these species into one taxon. 23. Pavetta owariensis Palisot de Beauvois, E d'Oware et de Benin, en tanque 1: 87, t. 1806. TYPE: Nigeria. Between Oware = Buonopozo (fl), Palisot de ene s.n. (holo- e, KEY TO THE VARIETIES OF PAVETTA OWARIENSIS IN CAMEROON l. | Fourth and higher order leaf venation obscure or invisible below... 1. — At least some leaves with fourth and higher order leaf venation not obscure or invisible below 2. Calyces not persistent in mature fruit: Apo so far known only from South, Littoral, and Centre Provinces . var. opaca + Domatia absent from secondary veins and ab- sent along most midribs var. a 3. Data ca along midrib and some se ondary vein var. owariensis Pavetta owariensis var. owariensis Pavetta dr Hutch. & Dalziel, Fl. W. Trop. Africa 2: 92. TYPE: Sierra Leone. Fundu, Jan. 1908 (fl As a 214 (K). Pavetta ea gee (a Nov. Regni Veg. 7: 15. TYPE: Nigeria. Eket District, 1912— 191: 3 n ed & Talbot s.n. ta: BM; isotype, K). rx Pavetta Pw Bremek., Repert. Spec. Nov. Regni Veg 47: 20. 1939. TYPE: Nigeria. Sapoba (fl), [niea 2346 i Shrubs or small trees to 8 m. Twiglets glabrous o puberulent, floriferous twiglets 11.5-19 cm. Leaves chartaceous, sometimes anisophyllous; blades elliptical to ovate, obovate or oblong, (2-)9— 16(-25) X (1—)3.5—6.5(-9) cm, glabrous except ma- jor veins and margins sometimes subglabrous, apex acute or obtuse, often asymmetrical, with acumen 3-12 X 2.5-7 mm; base attenuate to cuneate, sometimes asymmetrical; midrib prominulous be- low toward base, secondary veins 4—9 each side, eucamptodromous; usually hairy, crescent-shaped pocket domatia or pit or tuft domatia in most branch vein angles of midrib, sometimes along sec- ondary veins; nodules scattered on blades or near midrib; fourth and higher order venation obvious above and below; venation density coarse. Stipules sometimes deciduous, rotund to triangular lobed, pubescent internally at least near base, subglabrous externally; awn cuspidate to linear, 1-3 mm. Inflo- rescences 3-6 cm across, subrotund to subpyram- idal or cruciform in outline, sometimes lax, puber- ulent, peduncle sometimes with sheathing bracts along its length as well as at base, to 10 mm; sub- units sometimes inverted pyramidal in outline; flowers 20-125; sheathing bracts triangular to ro- tund lobed or unlobed and saucer-shaped, puber- ulent to subglabrous externally, pubescent inter- nally; awns ovate or linear, 1 mm; other bracts ovate or linear, to 2 mm, sometimes with 1-several fimbriae to 1 mm, or tufts of hair or fimbriae borne from axis directly; bracteoles resembling smaller bracts sometimes present. Calyx tube 0.7-1.5 mm long, 1.5-2 mm wide halfway up; lobes valvate (or bases less often slightly overlapping), rotund, tri- angular, pentagonal (or less often oblong or ovate), 0.5-2 X 0.5-1.8 mm, puberulent, sometimes car- inate, rim usually lighter. Corolla white; tube broadened from base to apex, 3.5—6 X 0.7-2.5 mm; obes 4—7 mm. Style narrowly clavate, pubescent or puberulent, exserted 7-9 mm. Pre-fruiting ovules 2, attached ca. halfway up septum, reniform. - ier edi Specimen examined. (CAMEROON. Sout Province: southern end of —— National 0). rovi Park, Mar (fl), Thomas & McLeod 5823 (K, M b. Pavetta owariensis var. opaca S. D. Manning, var. nov. TYPE: Cameroon. South Province: 60 km S of Edéa, S of Mboké, Mar. 1965 (fl), Leeuwenberg 5555 (holotype, P; isotypes, BR, K, MO, WAG). Figure 20. A varietate owariensi domatiis interdum nullis, inter- Volume 83, Number 1 Manning 135 1996 Pavetta Subgenus Baconia in Cameroon 7.0 cm D» ! 5d re AS aa Y MAL ESSE ADE RON Ñ / ij, a e Y SS 136 Annals of th Missouri Botanical Garden dum secus costam et rare secus nervos secundarios; reti- ita tenui, venulis Tip obscuris vel invisibilibus dif- . Inflorescentiae m latae; corolla tubo (3-)5-9 mm às id lobis 6-11 mm piles styli exserti (6—)10—15 mm. Shrubs or small trees to 15 m tall. Floriferous twiglets glabrous, to 37 cm. Leaves sometimes co- riaceous, no ovate blades seen; apex occasionally rounded; acumen to 23 X 12 mm; base sometimes obtuse, only slightly if at all asymmetrical; midrib prominent below at least toward base, domatia ab- sent or present as crypts in some branch vein an- gles of midrib and occasionally on secondary veins, often puberulent; nodules along midrib and on blade, sometimes rare; venation density fine; fourth and higher order venation obscure or invisible above and below; stipules cup-shaped, glabrous ex- ternally, awn to 4 mm. Inflorescences corymb- shaped to pyramidal or rotund in outline, 2-15 cm across, sometimes subglabrous proximally, sessile or subsessile, flowers to 350 per inflorescence; sheathing bracts sometimes with linear awns to 2 mm or foliar appendages to 3 mm; foliar bracts sometimes present; other bracts concave, irregular or obovate; calyx tube to 1.8 mm long and to 2.5 mm wide halfway up; lobes 1-3 X 1-2 mm, sub- glabrous to pubescent; corolla tube subcylindrical or cylindrical, (3-)5-9 X 1-3 mm, lobes 6-11 mm: style subglabrous to pubescent, exserted ca. (6—)10— 15 mm. Fruits ca. 8 mm across, subglobose, puber- ulent to glabrous, whitish, gray-green, or light green with darker green streaks, without persistent calyx. Seeds 2 or 1, slightly concave when immature, Other character states within ranges described above for va- riety owariensis. The most obvious distinguishing feature of vari- ety opaca is the obscure leaf venation above and elow as a result of its thick, sometimes brittle leaves. Although this also occurs in variety satabiei, the calyx does not persist in fruit in variety opaca as it does in variety satabiei. The tallest represen- tative of subgenus Baconia in Cameroon belongs to variety opaca. —— specimens examine d. Primaria e yy ovince: Bipindi (fr & vegetative), Zenker 2554 (B HBC. MO, S, W, WAG); Bitye (fl), “Bates 1210 (BM, de si “Bitye, €— ; MO, labeled E River”). Lit- eran July (fr), / e & Gar. 109 E). Littoral Province (Centre?): Kéllé River ca. 50 km NW of Eséka, Nov. (fl E & fr), de Wilde & de Wilde- Mii 1295 (K. ; G). Centre Prov- ince: Yaoundé (fl), Zenker & Staudt 278 (K), 543 (BM); from y Es éka, chantier de la SBC de Badjob, Feb. (fr), Mpom 202 (BR, YA): 40 km S of Badjob, 50 km SW of Eséka near Nyong River, Dec. (fr), de hg & de Wilde- Duyfjes 1538 (B, BR, K, MO, WAG, YA); Son Mbong, E km SW of Eséka, Dec. (fl bud & fr), Rus 1378 (BR, T o YA); 5 km W of Son Mbong, Mar. (fl), Leeuwenberg 5069 (BR, HBG, MO, WAG, YA). c. Pavetta owariensis var. glaucescens (Hiern) 5. D. Manning, comb. nov. Basionym: Pavetta glaucescens Hiern, Fl. Trop. Africa 3: 171 1877. Ixora glaucescens (Hiern) Kuntze, Revs Gen. Pl. 1: 286 E: Equatorial Guinea. Bioko Island, formerly Fernando Po, rec'd. June 1862 (fl), Mann s.n. (holotype, K). Pavetta chionantha K. Schum. & K. Krause, Bot. Jahrb. Syst. 39: 552-553. 1907. P Cameroon. South Proviride: Bipindi, Mar. 1900 (fl bud & Fi Zenker 2254 (syntypes, B al BM, HBG, S, W) Shrubs to 4 m. Floriferous twiglets 6-27 cm. Leaves membranaceous to coriaceous, blades to 25 10 cm, apex occasionally rounded or emarginate, acumen to 2] 12 mm or occasionally absent, domatia absent from secondary veins and usually from midrib, venation sometimes brochidodromous, venation density medium, fourth and higher order venation often obscure above; stipules cup-shaped, glabrous externally, awn to 4 mm. Inflorescences sometimes trapezoidal or corymb-shaped in outline, to 16 em across, sometimes subglabrous, peduncle absent or ca. 3 mm; sheathing bracts sometimes cup-shaped, their internal vestiture often extending mm and € beyond margin, with linear awns to sometimes foliar appendages 3-5 mm; foliar bracts resembling normal or slightly reduced foliage leaves sometimes present; calyx lobes mostly ovate, ca. 14 mm long, sometimes subglabrous, margin ciliate; corolla tube 5-8 mm, sometimes cylindri- cal, lobes 6-12 mm, tips green in bud; style ex- serted 9-18 mm, sometimes glabrous. Fruits and seeds similar to those of variety opaca. Other char- acter states within ranges described above for va- riety owariensis. Of the above described features, the ones most reliably distinguishing variety glaucescens from va- riety owariensis in Cameroon are: fourth and higher order venation often obscure above, domatia absent from secondary veins and usually from the midrib, calyx lobes mostly ovate, corollas larger, and styles further exserted. 7 d examined. CAMEROON. th 5 km N of Kumba on Kumba- a & Thomas 140 (MO); Forest Reserve, Mar. m, i-i et 9 (K); Southern Bakundu Bopo- Bangs ; path, Feb. (fl), Binuyo & Darme FHI 35576 (K); m Bakundu Forest Reserve, path from beacon 45—44, Apr. (fl), Ejiofor FHI 29306 (K); Southern up. Forest rob near piis e W of Kumba ( . Apr. uds fr), riso 1680 (MO); 'Barombi Kang near Kum- specimens Volume 83, Number 1 1996 Manning 137 Pavetta Subgenus Baconia in Cameroon ba, Feb. (fl), Thomas 4367 (MO), June ie), Thomas 7076 (MO); Lake Barombi Mbo ca. 5 km NW « (fl & young fr), Manning 1785 703, 792, 819, 820 (all MO), June tin, Manning 2059 (MO), Mar. (fr), Nemba & ne 76 (MO); of Korup opt b ud), devis & McLeod 5826 (BR. MO), Apr. (fl). p mae 4724 (MO); southwest- Bs i AE Park, June (fr), Thomas 8108 = = ern corner of Kor (MO): Mundemba, Mar. (fl). aires 6798 (Mí »: Hio del Rey, June (fl). Johnston s.n. oral Province: S of Nkongsamba, Dec. (fl bud). $ = 306 (WAG). Centre Province: Yaoundé (vegetative), Zenker 700 (K lower part of sheet only: not upper part of K sheet or NY, P, or S which are Pavetta calothyrsa). South DON e: Bipindi, s. W). Mar. (fl bud & fl), Zenker 2254 (BM, HBG, jara ~ oo . Pavetta owariensis var. satabiei S. D. Man- TYPE: Cameroon. Southwest Province: Butu-Dikome Balue Road ca. 34 km NW of Kumba, Mar. 1976 (fl bud & fl), Satabié 252 (holotype, BR; isotype, YA). Figure 21. ning, var. nov. A varietate opaca fructu maturo cum calyce persistenti differt Similar to variety opaca except calyces persistent in fruit. Also, in variety satabiei the tallest reported representative is 5 m, no domatia have been seen, and fruits to ca. 10 mm across have been seen. Additional specimens examined. CAMEROON. Southwest Province: around Ehumseh, Ntehol and Me- jelet, W of od June (fr). Etuge & Thomas 176 (BR. MO): path from Ndibisi to Mejelet, W of Bangem. June (fr). Etuge & Thomas 534 (MO). In Cameroon, Pavetta owariensis var. owariensis is known only from Korup National Park in south- western Southwest Province. It is centered west of Cameroon in the Upper and Lower Guinean sub- centres of specific endemism sensu White (1979). It also has been reported from Bioko and Gabon. Variety opaca is endemic to Cameroon and is fairly widespread in Littoral, Centre, and South Prov- inces. Variety glaucescens occurs in Cameroon, southeastern Nigeria, Bioko, and reportedly in Ga- bon, Cabinda (Angola), Rio Muni (Equatorial Guin- 1963). Its distri- bution thus is centered in the Lower Guinea centre ea), and Congo (Hepper & Keay, of specific endemism sensu White (1979), overlap- ping that of the more west-centered variety owa- riensis in southern Nigeria, Cameroon, and Bioko. In Cameroon, though most collections of variety glaucescens are from southern Southwest Province, others are from widely scattered locations in Cen- tre, South, and Littoral Provinces. Variety satabiei is endemic to Cameroon’s Southwest Province. Pavetta owariensis occurs in mature and second- ary forest and at forest edges. Varieties glaucescens, opaca, and owariensis are lowland taxa; variety opa- ca has also been found at 800 m. Variety satabiei has been found so far only at elevations between 800 and 1400 m. Reported heights of varieties opa- ca and owariensis include plants taller than in the other two varieties or in most other species of sub- genus Baconia in Cameroon. Intermediates nearly bridge the gaps between the varieties recognized here, which are thus a com- promise between distinctions that are too fine at the species level and a failure to recognize taxonomi- cally the differences that do exist. Although Pavetta owariensis is variable as inter- preted here, the spreading, sometimes lax inflores- cences, sometimes profuse in varieties glaucescens, opaca, and satabiei, and medium to large flowers of Pavetta owariensis span the varieties recognized and distinguish P. owariensis from otherwise similar species such as P grossissima and P. gabonica, which have smaller inflorescences. Calyx lobes are never subquadrate (Fig. 5) as in P. calothyrsa. They are never compressed (Fig. 5) and are 1 mm or more long. Their bases overlap less than 10% of the time, distinguishing P. owariensis from P cor- ymbosa. Lack of papillae or vestiture on upper leaf surfaces distinguishes it from most specimens of P. longibrachiata, and leaf venation and domatia pat- terns in the latter species are distinct from those found in any of the varieties of P. owariensis. Pav- etta namatae differs in having tertiary leaf veins often prominent below and in having more second- ary veins. The former Pavetta glaucescens, including the former P. hygrophytica and newer collections, re- mains intact at the varietal level. Variety opaca is so named for its obscure or invisible higher order venation both above and below. Variety satabiei is named in honor of Benoit Satabié, director of the National Herbarium of Cameroon, who collected the first known and type specimen of the variety. 24. Pavetta robusta Bremekamp, Repert. Spec. Nov. Regni Veg. 47: 19. 1939. TYPE: Gabon. Haute-Ngounyé, Poungui, Mar. 1927 (fl bud & fl), Le Testu 6420 (holotype, BM: isotypes, BR. MO, P). Shrubs ca. 5 twiglets 2.5-18 cm. Leaves coriaceous to subcor- Twiglets glabrous, floriferous laceous, glabrous, occasionally anisophyllous: blades oblong to obovate or elliptical, 5-28 x 2- 13 cm; apex obtuse to acute, rounded (or rarely — A emarginate), acumen if present 4-15 X 3-10 mm, base cuneate or attenuate, sometimes asymmetrical, midrib prominulous below toward base, midrib and secondary veins often impressed above, secondary 138 Annals of the Missouri Botanical Garden Q l em ] ^w SE NS PS e AS PHYLLIS BICK2 1.0 cm Figure 21. Pavetta owariensis var. satabiei (A—D, Satabié 252, BR; E, Etuge & Thomas 178, MO).—A. Habit.— B. Part of upper leaf surface with bacterial nodules along midrib.—C. Part of lower leaf surface without domatia.—D. Flower.—E. Fruit with persistent calyx. Volume 83, Number 1 Mannin 139 Pavetta Subgenus Baconia in Cameroon veins 5—9 each side, eucamptodromous or brochi- dodromous, often sparingly pubescent, often cres- cent-shaped pocket or sometimes pit domatia in some branch vein angles of midrib; nodules scat- tered on blade, rarely on midrib; fourth and higher order venation usually slightly more obvious above; venation density fine. Stipules rotund lobed, gla- brous externally and internally at least above the base, awn deciduous. Inflorescences corymb- shaped to subrotund or pyramidal in outline, 4-14 cm across, subglabrous to puberulent distally, sub- glabrous to glabrous proximally, sessile except leaf- less lateral inflorescences from node below main inflorescence, if considered separate, with pedun- cles to 4 em: flowers 75—400, fragrant; sheathing bracts rotund or ovate lobed, ziii to pubescent near base internally, glabrous externally, awns de- ciduous, + ovate to linear, =1 mm; foliar bracts if present resembling small foliage leaves except base sometimes auriculate: other bracts ovate to rotund lobed, to 3.5 mm, sometimes with l-several fim- briae to 2 mm or fimbriae borne from axis directly; bracteoles resembling smaller bracts sometimes present. Calyx tube 1-1.8 mm long, 1.2-2 mm wide halfway up: lobes valvate or bases overlapping, subquadrate (to sometimes obovate, compressed ro- tund, rotund, truncate or shallowly 2-lobulate), 0.2— -2 mm, glabrous, not or inconspicuously car- inate, rim sometimes lighter. Corolla white, tube pap from base or slightly above base to apex, 5-9 X |-2.5 mm: lobes 5-9 mm. Style cla- vate, subglabrous to pubescent, exserted (4—)7—10 mm. Stigma 2-lobed, lobes rarely spreading up to 2 mm. Ovules in open flower 2, + reniform, one on each side of septum. eee a examined CAMEROON. South Province: e (fl bud « fl), Bates 1046 (BM, MO; both labeled * in Yaunde” The only collection of Pavetta robusta from Cam- eroon is apparently from lowland forest. The spe- cies also occurs in Gabon. Its large leaves and large, profuse, many-flowered inflorescences fit its name excellently. Comparisons of P. robusta to the similar P. calothyrsa and P. molundensis follow the descriptions of those species. It is similar to P. cor- ymbosa in having calyx lobe bases overlapping more than 10% of the time. However, calyx lobes are rarely longer than 1 mm in P. robusta, rarely as short as 1 mm in P. corymbosa. Calyx lobes are usually subquadrate (Fig. 5) in P. robusta, usually rotund and rarely subquadrate in P. corymbosa. Leaves and inflorescences also are larger on aver- age in P. robusta than in P. corymbosa. 25. Pavetta rubentifolia S. D. Manning, sp. nov. TYPE: Cameroon. Southwest Province: Bak- ossi Movsiaiis W of Bangem, Jan. 1986 (fl), Thomas & Mcleod 5343 (holotype, MO; iso- type, YA not seen). Figure 22. ex. Folia marronina, coriacea, glabra; nervis secun- dariis utroque 8-13; domatiis nullis; nodulis nullis. Inflo- rescentiae in ambitu o ge ha subcorymbosas, ).5-1.5 cm latae. Lobi calycini Py aan -compressi (vel interdum eid i). ca. x | glabri. tubo 3—4 mr 3-5 mm. super prope faucem pubescentibus. Styli exserti 3-5 mm. viter triangulares, 20.5 evi ~ = Corolla : lobis Shrub 1 m. Twiglets glabrous, floriferous twiglets ca. 14 cm not or hatslly anisophyllous: blades oblong to ellip- 13-16 subacuminate or with acumen 8-15 X 3-5 mm; . Leaves maroon, coriaceous, glabrous, tic or obovate, X 3.5—4.5 cm: apex acute, base cuneate to attenuate, sometimes asymmetrical; midrib sometimes prominulous below toward base. secondary veins 8-13 each side, usually eucamp- todromous; domatia absent; nodules absent; fourth and higher order venation usually somewhat more obvious above; venation density coarse. Stipules glabrous. Inflorescences deciduous, cup-shaped, Y subrotund to subcorymb-shaped in outline, 0.5— cm across, subglabrous to puberulent except pe- duncle glabrous, peduncle 4 mm, flowers 20-35: sheathing bracts unlobed and cup- or saucer- shaped or broadly lobed, glabrous, awns linear, ca. 1 mm; other bracts + ovate to obovate or linear, to 2 mm, sometimes with 1-several fimbriae ca. 1 mm: bracteoles if present resembling smaller bracts or fimbriae borne from pedicels directly. Calyx tube 0.7-1 mm long, 1.7-2 mm wide halfway up; lobes valvate, short pide) compressed rotund (or less often deltoid), ca. 0.2-0.5 X 1 mm, glabrous, not (or inconspic dine carinate, rim narrowly lighter. Corolla white; tube cylindrical or subcylindrical, 3— 4 X 1-1.5 mm; lobes 3-5 mm, near throat. Style narrowly nita puberulent, ex- pubescent above serted 3—5 mm. Pavetta rubentifolia is endemic to eastern South- west Province, Cameroon. It is a montane forest shrub collected at an elevation of between 800 and 1600 m Pavetta rubentifolia is characterized by its ma- roon-coriaceous leaves, very small flowers, and very small but not subumbellate inflorescences. It is dis- tinct from other Cameroon specimens, but similar to P. ixorifolia Bremekamp s. str., a species occur- ring west of Cameroon, and to P. nitidula s. str.. which occurs south of Cameroon. Pavetta ixorifolia s. str. and P. nitidula s. str. lack the elongate acu- mens sometimes found in P. rubentifolia. Some 140 Annals of the Missouri Botanical Garden Figure 22. rescence. Cameroon collections have previously been identi- fied on herbarium sheets as P. nitidula and others as P. ixorifolia, but those all belong to other spe- cies. Pavetta rubentifolia resembles P. camerounensis in having corolla throat beards extending onto the upper corolla lobe surfaces. These species differ, however, in that the inflorescences of P. rubentifol- Pavetta rubentifolia (Thomas & McLeod 5343, MO).—A. Habit.—B. Details of leaf venation.—C. Inflo- ia, while very small, are not subumbellate and in the normally shorter, narrower maroon leaves of P. rubentifolia. Pavetta rubentifolia is so named because of the maroon color of its leaves as seen in the field. 26. Pavetta staudtii Hutchinson & Dalziel, Fl. W. Trop. Africa 2: 91. 1931. TYPE: Cameroon. Volume 83, Number 1 1996 Manning 141 Pavetta Subgenus Baconia in Cameroon Southwest Province: Kumba, formerly Johann- Albrechtshóhe, 1896 (fl bud & fl), Staudt 599 (holotype, K; isotypes, P. S). Shrubs to 6 m. Twiglets glabrous, floriferous )6-30 cm. brous, sometimes anisophyllous; twiglets (4— Leaves subcoriaceous to coriaceous, gla blades elliptic to obovate (or less often ovate or oblong), 2-20 x or rounded, sometimes asymmetrical, with acumen (2-)5-25(-35) X —12 mm; base cuneate (to less often rounded, attenuate, or 1—7.5 em; apex acute to obtuse usually cordate), often asymmetrical; midrib and second- ary veins prominulous or subprominulous below, secondary veins (3-)5-10 each side, sometimes joined 2-10 mm from margin; pit, pocket, or crypt domatia in branch vein angles of midrib and rarely of some secondary veins, usually pu- bescent; nodules mostly pustuliform, often nu- erous and conspicuous, scattered on blade and not usually along midrib or secondary veins; ourth and higher order venation rather obscure above, usually slightly more obvious below; ve- nation density medium. Stipules cup- shaped, gla- brous externally and above base internally, sheath of- ten + triangular lobed and extending to 3 mm above node, awn deciduous, linear to cuspidate, 1-3 mm. Inflorescences corymb- or subcorymb-shaped in out- line, 1-8 tally, subglabrous (to less often glabrous) proximally. peduncle to 20 mm usually present, sometimes with a sheathing bract ca. halfway up, flowers (10—)20— 100(-200); sheathing bracts rotund to elliptic lobed to unlobed and saucer- or cup-shaped, glabrous (to em across, puberulent to subglabrous dis- sometimes pubescent at base internally), with linear awns 1-2 mm or occasionally foliar appendages 2-5 mm; other bracts ovate to obovate (or less often lin- ear), to 2 mm, sometimes with l—several fimbriae to ca. 1 mm long, or fimbriae borne from axis directly; bracteoles resembling smaller bracts sometimes pres- ent. Calyx tube 0.5-1 mm long, 1.2-1.5 halfway up: lobes valvate or overlapping in bud, not overlapping at bases in open flowers, subquadrate (to mm wide less often rotund, pentagonal, or shallowly 2-3 lobu- late), 0.2-1 x at base), glabrous to subglabrous, sometimes carinate, 0.7-1.5 mm (or occasionally truncate rim lighter. Corolla white or creamy white, tube broad- ening from base to throat, 4—1 0.7-2.5 mm: lobes sometimes green tipped or margined, 5-10 mm. Style clavate, pubescent (to less often subglabrous). exsert- ed 6-12(-15) mm. Stigma sometimes narrowly 2- lobed, lobes 1-2 mm. Fruits ca. 7-10 mm across, glabrous, white, whitish, brown, light green, or gray with dark or glaucous green vertical stripes. Ovules 2, sometimes pendent in flower bud, seeds 2 or 1, concave, attached ca. halfway up septum. Additional examined. CAMEROON. Southwest Provi jo, W bank of Meme River on osa a ub Feb, (fl bud), Thomas & Nemiva 664 (MO); Kindonge Camp, Southern Bakundu ecu ca. 20 km SSW of Kumba, Apr. (fl), Movida 5 (MO); 5 km W of Kumba, Nov. (fr), el 739 id paik he Mejelet- Ehumseh to Mua i D — —. = . (fr), Mrrmenle 8153 (WAG); Mungo River valley o and gare de Mujuka, June (fr), Fleury 33426 (P). Centre Province: hill N > 5 a — (fr), Bamps 1339 ; 12 km W of Son Mong Mar (fl bud & fl), ¡a 5026 i Fifinda-Bella forest track 5 km E of Kribi-Edéa road, ca. 16 km E of Kribi, N of Lolodorf road, Sep. (fr), Bos 5.373 (BR, MO, a Bipindi (fl bud), Zenker 4913 (BM, BH, HBG, K, MO, S, W), (fl bud, fl & vegetative), Zenker 4427 (BM, BR, HBG, S. W), (fl), Zenker 4355 (BM, BR, HBG, , MO, S, W); hill DU ie rua on road between Biwong Boulou and mila 25 SE of pompes Jan. (fl ud), Letouzey 9839 (BR, HBG, "p WAG, YA); Meyo die m 9 km W of Sangmélima, Mar. (fl), Meijer -4 (MO); 10 km SW of Ambam, S of Ebolowa, ate de Wilde & de Wilde-Du xhes 2044 (BR, P, WAG); r ati near Congo bor der SE of Djoum, Jan. (fl bud & i. " Biholong 261 (BR, A). mts, Pavetta staudtii is one of the most widespread species of subgenus Baconia in Cameroon, occur- ring in scattered locations in Southwest, Littoral. Centre, and South Provinces. It is centered in the Lower Guinea subcentre of specific endemism sen- su White (1979) and is endemic to Cameroon, though it is predicted from collections near borders that range extensions will be discovered south and perhaps west of Cameroon. Pavetta staudtii is in dense and open primary and secondary forest understory, mostly in high but not highest rainfall areas. It has been re- ported from elevations of 50-1100 m, from lowlands. Often collections have ported to be sweet smelling. If a plant of subge- nus Baconia has conspicuous and numerous punctate, subpunctate, or pustuliform black nod- ules mostly scattered on leaf blades, leaves often long compared to their width, and inflorescences always corymb- or subcorymb-shaped and usu- ally neither large nor condensed, it is likely to belong to P. staudtii. South Province collections sometimes have more flowers and larger inflores- cences than those from elsewhere. Pavetta staudtii is distinct from P kribiensis, which it resembles vegetatively, in having subquad- 142 Annals of the Missouri Botanical Garden rate (Fig. 5) rather than rotund to deltoid calyx obes, less condensed inflorescences, and often fim- briate bracts. 27. Pavetta tenuissima S. D. Manning, sp. nov. TYPE: Cameroon. South Province: Ma’an, Nov. 1979 (fl), Letouzey 15211 (holotype, P). Figure 2: "rutices. Rami florife ari O(sic fac sientes inflorescentias secundariis a anemias: nervis ener i eons 10-15; reticulo 1, pone uus pe prope faucem pubescenti- bus. Styli pane 2-3 r Shrubs to 2 m. Twiglets pubescent, floriferous twiglets absent or to 10 cm. Leaves chartaceous, not or hardly anisophyllous, blades obovate (or less often elliptic), 14-31 cm, glabrous except margin sometimes pubescent; major veins below, costa and at times part of secondary veins above pubescent; apex acute to rounded, usually with acumen 3-20 X 3-10(-15) mm; base acute to obtuse (or occasionally rounded), sometimes asymmetrical; midrib and secondary veins prom- inent below, secondary veins 10-15 each side, usually joined 2-5 mm from margin; domatia ab- sent or dense tufts of vestiture in branch vein angles of midrib and along secondary and tertiary veins below; nodules if present linear on side veins or of irregular shapes, scattered on blade; tertiary veins prominent below; fourth and higher order venation obvious above and below; vena- tion density extremely fine. Stipules * rotund or compressed rotund lobed, pubescent near base internally and densely pubescent or puberulent externally, awn linear, 5-9 mm. Inflorescences subumbellate or with subumbellate subunits, ro- tund in outline, 0.5—2 cm across, congested, ses- sile, pubescent, flowers (10—)50; sheathing bracts rotund lobed or unlobed and then saucer-shaped, pubescent externally, + subglabrous internally, awns if present linear, 1-5 mm; other bracts + subrotund to deltoid or linear, sometimes lobed, to 3 mm, not fimbriate, awn if present ca. 1 mm; bracteoles resembling smaller bracts. Calyx tube 0.2-0.8 mm long, 1-1.5 mm wide halfway up; lobes valvate, triangular or rotund, 0.5-1 X 1 mm, pubescent at least near margin, sometimes carinate, rim sometimes lighter. Corolla white or light green; tube cylindrical, 2-3 X 1 mm; lobes 2-4 mm, sometimes thinly pubescent above near throat. Style fusiform to clavate, pubescent, ex- serted 2-3 mm. Fruits ca. 5-10 mm across, gla- brous to pubescent, bluish green. Mature seeds 2 or l, attached ca. halfway up septum, concave. ji aga a imens examined. CAMEROON. South rovince: Ma’ ae oa fr & fr), : (P, YA); Adjo ou, 17 km Raynal 10148 (P); eee 17 km SW Feb. (fr), Raynal & Raynal 9897 (P). Littoral (Centre?) Province: 50 km NW of Eséka, Kéllé River, Nov. (fr), de Wilde & de igen Duyfjes 1293 (WAG). Southwest Prov- ince: Ban Southe akundu Forest Reserve, Mar. (young fr), Brenan 9280 (K), 9280A (K). Pavetta tenuissima is only known from Came- three collections have been from ni, Equatorial Guinea, that it probably exists there as well. It also occurs in southern Littoral (Centre?) and Southwest Provinces. It usually grows in wet lowland primary forest and has been report- ed from both riverine and transiently inundated for- est and on clayish soil. Pavetta tenuissima has the finest mesh reticula- tion of higher order leaf veins in Cameroon species of subgenus Baconia. The prominent, subrectan- gular gridlike pattern formed by the secondary and tertiary veins below is conspicuous to the naked eye, and distinctive. This subrectangular, gridlike venation also occurs in P. muiriana; these two spe- cies also have similar floral morphology. Pavetta tenuissima can be distinguished from P. muiriana by its extremely fine leaf vein reticulation, smaller and more congested inflorescences, mostly larger A leaves, and smaller corollas. Inflorescences of Pavetta tenuissima occasionally (Raynal & Raynal 9897, de Wilde & de Wilde-Du- yfjes 1293) are on floriferous twiglets so reduced as to render them axillary. In this respect and in its condensed inflorescences, P. tenuissima resembles P. camerounensis subsp. brevirama and P. grossis- sima. The latter two taxa, however, do not have the extremely fine higher order leaf vein reticulation of P. tenuissima, which is so named because of this feature. 28. dep e phia Pise Bull. Jard. Bot. 6: 257. 1956. TYPE: Zaire. Équa- teur Dus km. 28-29 of road iun Bikoro to Lac Tumba, May 1936 (fl), Louis 1997 (ho- lotype, U not seen; isotypes, B KEY TO THE SUBSPECIES OF PAVETTA UROPHYLLA — . Anthers septate, most calyx lobes ecarinate : = urophylla l. Anthers not septate, most calyx oe a with carinae produced into apic ulae pues bosii Manning 143 Pavetta Subgenus Baconia in Cameroon Volume 83, Number 1 EZ A M 77 Al, EN 7 . Habit. —B. Lower surface of leaf base showing details of JWET. B Pavetta tenuissima (letouzey 15211, P).—A Figure 2: j leaf venation, the finest seen in subgenus Baconia.—C. Fk 144 Annals of the Missouri Botanical Garden a. Pavetta urophylla subsp. urophylla Shrubs ca. 3 m. Twiglets glabrous, floriferous twiglets 6-24 cm. sometimes anisophyllous; blades ovate to elliptical or oblong, sometimes asymmetrical, 5.5-29 x 2- 11 cm; apex acute to obtuse, often asymmetrical, with acumen 10-45 X attenuate (or less often obtuse); midrib sometimes l Leaves coriaceous, glabrous, 3-12 mm; base cuneate to prominulous below, secondary veins (: each side, sometimes joined 3-5 mm from margin; mostly pubescent pocket domatia typically in most branch vein angles of midrib; nodules scattered on blade, usually conspicuous and more easily visible below; fourth and higher order venation obvious above and below; venation density fine. Stipules de- ciduous, rotund or deltoid lobed or unlobed and then cup-shaped, pubescent internally, glabrous ex- ternally, awn deciduous, the one seen cuspidate, ca. ] mm. Inflorescences corymb-shaped, subcorymb- shaped, or pyramidal in outline, 2-7 cm across, + lax, puberulent distally, subglabrous proximally, 100); ovate lobed, deciduous, pubescent internally, gla- brous externally, awns linear, 1-2 mm sessile, flowers 35—60(— sheathing bracts ; other bracts obovate to linear, 0.2-2.5 mm, not normally fim- briate, sometimes with awn ca. 0.5 mm; bracteoles resembling smaller bracts sometimes present. Calyx tube 0.7-1 mm long, 1.5-2.5 mm wide halfway up; lobes valvate, subtruncate, ca. 0.1-0.5 X 1 mm, glabrous, usually ecarinate, drying brownish black, rim narrowly lighter. Corolla beige-tan; tube slightly broadening from base to throat, 2(-3.5) X 1-1.5 (-2.5) mm; lobes 4—5 mm, often thinly pubescent above near throat, reflexed. Anthers septate. Style fusiform, pubescent (to sometimes puberulent), ex- serted ca. 5-6 mm D. Man- Cameroon. South Province: top of Calvary Mountain, 28 km ENE of Kribi, Lolodorf Road, Mar. 1970 (fl & young fr), Bos 6611 (holotype, WAG). Figure 24 b. Pavetta urophylla subsp. bosii S. ning, subsp. nov. A subspeci ie urophylla foliis valde discoloribus; inflo- rescentiis 3-5 mm pedunculatis; calyce lobis breviter triangularibus, denticulatis (ad interdum rotundato- -com- pressos vel deltoideos), carinatis, carinis in apicula acuta productis; corolla lobis 2-4 mm; antheris nonseptatis dif- fert. Similar to subspecies urophylla except leaves more strongly discolorous; blades’ venation density medium; inflorescences subglabrous distally, gla- brous proximally, with peduncles to 5 mm with con- spicuous ovate-lobed sheathing bracts at their bas- es and apices; bowl-shaped sheathing bracts also present pd on inflorescences; calyx tube some- times only 0. mm, lobes short triangular, den- ticulate (or e often compressed rotund), with ca- rinae produced into sharp apiculae, rim sometimes more broadly lighter; corolla brick red; anthers not septate; style puberulent to glabrous, exserted ca. 4—5 mm. The most obvious features distinguishing sub- species bosii from subspecies urophylla are non- septate anthers and apiculae on most calyx lobes. he more strongly discolorous leaves of subspecies bosii and the absence of extremely long acumens in subspecies bosii, such as sometimes occur in Al- though the two subspecies could be considered sep- subspecies urophylla, are also noteworthy. arate species, this is not done because there are few collections and they clearly resemble each oth- er more than either resembles any other known tax- ^ n. Subspecies bosii is known only from western South Province, Cameroon, in high forest, elevation unknown. Subspecies urophylla occurs in Zaire and Congo. It has been collected at the edge of forest on sandy-clayey soil and in semideciduous, riverine, swampy, and seasonally inundated forests. Reported elevations from Zaire are ca. 370 and 470 m. Pavetta urophylla is one of the most distinctive species in subgenus Baconia. Distinguishing fea- tures are long leaf acumens, nodules more conspic- uous from lower than upper leaf surfaces, brick red or beige-tan corollas, and calyx lobes often sub- truncate near the base. Also, atypically in subgenus Baconia, the lower part of the style is not glabrous (Fig. 24) 29. Pavetta viridiloba K. Krause, Bot. Jahrb. Syst. 54: 353. 1917. TYPE: Cameroon. East Province: between Mpan (formerly Assobam) on Boumba River and Lomié, Apr. 1911 (fl), Mildbraed 5120 (holotype, B destroyed; lec- totype, selected here, HBG) — e24. Pavetta urophylla subsp. bosit d pen as i: inl halo p urther up, opposite of the Flower.—F. ower leaf surface Baconia open dl the upper leaf su RE kr FA Ashi corolla removed to show aie s Bacon nther.—D. Calyx rma ion id eae and »» ind dul, "Nodules i in > all other species of ena Volume 83, Number 1 Manning 145 1996 Pavetta Subgenus Baconia in Cameroon 146 Annals of the Missouri Botanical Garden KEY TO THE VARIETIES OF PAVETTA VIRIDILOBA 1. Twiglets glabrous, leaf blades glabrous to subgla- brous above, secon n veins only sparingly pu- bescent above (Fig. 25C r. meurillonii 1. Twiglets pubescent (to less often pube i lez E blades pubescent to subglabrous above, secondar veins densely pubescent above... var. viridiloba a. Pavetta viridiloba var. viridiloba Shrubs to 3 m. Twiglets pubescent (to less often puberulent), floriferous twiglets (6-)18-29 cm. Leaves chartaceous to coriaceous, sometimes ani- sophyllous; major veins pubescent, blades pubes- cent, sometimes thinly so, to subglabrous above, to puberulent below, obovate to elliptical, oblong (or X 1.5-15.5 cm; apex obtuse to acute, subacuminate or with acumen 3- less commonly ovate), 5— 15 X 4-10 mm, acumen often at least as wide as long; base cuneate to attenuate, sometimes asym- metrical; midrib prominent below, secondary veins prominulous below especially toward base, 6-16 each side, eucamptodromous or brochidodromous; domatia absent except branch vein angles of midrib sometimes slightly indented or slightly more hairy than adjoining parts of veins; nodules absent, elon- gated along midrib or other veins, or scattered on blade; fourth and higher order venation usually ob- scure or invisible, usually slightly more easily vis- ible below, visible on near-apical leaves on non- flowering shoot; venation density medium. Stipules deciduous, cup-shaped, pubescent internally and externally, awn linear to cuspidate, usually falcate, (€«1-)1-9 X 1-2 mm. Inflorescences subcorymb- shaped to rotund or corymb-shaped in outline or with subunits of these shapes, 3-16 cm across, oc- casionally lax, pubescent, peduncle absent or to 8 mm; flowers 20-200, fragrant; sheathing bracts un- lobed and cup- to saucer-shaped or subrotund to ovate lobed, pubescent internally, pubescent (to at times subglabrous with vestiture sometimes con- centrated above subtending leaves) externally with awns resembling stipule awns 0. mm, linear awns 0.5-1 mm, or foliar appendages 2-7 mm; fo- liar bracts if present resembling but usually smaller than foliage leaves; other bracts ovate to linear, to 1.5 mm, sometimes with 1-several fimbriae 0.5-1 mm or fimbriae borne from axis directly; bracteoles resembling smaller bracts sometimes present. Calyx tube 1—1.5 mm long, 2.5—3.5 mm wide halfway up; lobes valvate, ovate to oblong, obovate, rotund, sub- rim lighter. Corolla white; tube broadening from base to apex, 4-11 X (1-)1.5-3(-5) mm; lobes 5— 8 mm. Style narrowly clavate, glabrous to puberu- lent, exserted 12-18 mm. Fruits ca. 5 mm across, subglobose or ellipsoid, thinly pubescent, pale gray. eeds 1-2, attached ca. halfway up septum, con- cave, sometimes only shallowly so. sae GE specimens examined. CAMEROON. East Provin m E of Lomié, Sep. (fr), Leeuwenberg 6688 (WAG). pes Tace Bitye, River Dja, Mun (fl bud & fl), Bates 1666 (BM, MO); between Dja Ri Sangmélima, May (young fr), Mildbraed 5468 (HBG); Ebemvok, 55 km W of Ebolowa, ous bs bud and fl), Raynal & Raynal 10432 (P, YA). € rovince: í km E of Makak, June (young fr), Manani 1970 (MO); 7 km ESE of Makak, June (young fr), Manning 2037 (MO), 2095 (MO); 14 km SW of Yaoundé along new Yaoundé- Douala road, May (young fr), Manning 1891 (MO) b. Pavetta viridiloba var. meurillonii S. D. west Province: Fontem, Mar. 1967 ( Meurillon 617 (holotype, P). Figure 25. A varietate viridiloba ramis glabris, foliis laminis super glabris vel subglabris, nervis secundariis super non nisi parce pubescentibus, inflorescentiis pubescentibus ad arie differt. Similar to variety viridiloba except as in key to varieties. Also, pit domatia are occasionally pres- ent, the largest leaf seen has been 17 X 9 cm, and inflorescences can be partly subglabrous. ariety meurillonii is recognized because, though it resembles variety viridiloba more closely than any other taxon, it is clearly different from all the collections of variety viridiloba in the features dis- cussed above. It is also geographically distinct. Variety viridiloba is endemic to south-central Cameroon, having been collected from South Prov- ince, southwestern East Province, and south-central Centre Province. Variety meurillonii is known only from the type from eastern Southwest Province, Cameroon. Pavetta viridiloba occurs in primary and second- ary forest including disturbed forest, and in rela- tively open areas, at elevations of 6 300 m. Bremekamp (1934) reduced Pavetta viridiloba to synonymy with P. puberula, but it is restored here. The two species can be distinguished in that calyx lobes are narrowly triangular with pointed tips in P. puberula but broadly ovate, oblong, or obovate or 2-lobulate (never with a pointed tip) in P. viri- diloba. Further, there is almost always more vesti- ture in P. viridiloba than in P. puberula; calyx lobes are densely shaggily pubescent rather than only pu- berulent, = leaf blades are much more thickly pubesce Calyx pm are usually green on herbarium sheets, accounting for the specific name, owing to vestiture thereon, a feature shared with P. mpomit. Volume 83, Number 1 1996 Manning 147 Pavetta Subgenus Baconia in Cameroon Figure Part of u 7.0 cm = ie ATIE AY KEBLE LSD 25. Pavetta viridiloba var. meurillonii (Meurillon 617, P).—A. Habit.—B. Part of lower leaf surface —C. pper leaf surface.—D. Flower bud.—E. Node and part of glabrous twiglet. 148 Annals of the Missouri Botanical Garden Calyces are usually longer and wider than those of most other taxa in subgenus Baconia. Leaf blades sometimes are among the largest in subgenus Ba- conia. They are olive green below Pavetta viridiloba can be distinguished from P. mpomi as in the key to species and as follows: some of its leaves reach a larger size than compa- rably placed ones in P. mpomii, floriferous twiglets are longer, most calyx lobes are more broadly elon- gate, corolla lobes are shorter, and styles are further exserted. Occasional 5-merous flowers have been found in both varieties of Pavetta viridiloba. The original species description did not indicate where the type was stored. Mildbraed 5120 (B), seen by Bremekamp, is destroyed. Mildbraed 5120 (HBG), the lectotype selected here, is the only du- plicate of Mildbraed 5120 found. TAXON OF UNCERTAIN STATUS Pavetta sp. near brachycalyx. 34593 HNC = SCA 1938 (YA) was collected from an unknown location probably in the Southwest Province of Cameroon and possibly cultivated in the Limbe Botanic Gar- den because it bears a label of the Victoria Botanic Gardens without data. Although previously identi- fied as Pavetta corymbosa, it is actually closer to P. brachycalyx. It differs from P. brachycalyx in hav- ing larger leaves and more flowers per inflores- cence. Its origin is unknown and its status marginal either within P. brachycalyx or as a new species. CAMEROON SPECIES EXCLUDED FROM THIS TREATMENT Pavetta bangweénsis Bremekamp, Repert. Spec. Nov. Regni Veg. 37: 79. 1934. TYPE: Came- roon. East Province (probably): Bangwe (prob- ably aoe Conrau 66 (holotype, B de- stro No eed fitting the description of this species has been found. SPECIES NOT KNOWN FROM CAMEROON Pavetta ankolensis Bridson, Kew Bull. 32: 610— 612. 1978. TYPE: Uganda. Eggeling 3709 (holotype, K; isotype, EA not seen). Pavetta annobonensis Bremekamp, Repert. Spec. Nov. Regni Veg. 47: 16-17. 1939. TYPE: Equatorial Guinea. Pagalu Island, Sep. 1911, Mildbraed 6676 (holotype, B destroyed; iso- type, HBG). Pavetta dermatophylla Mildbraed, Gart. Berlin-Dahlem 13: 704. Notizbl. Bot. 1937. TYPE: Equatorial Guinea. Pagalu Island, Sep. 1911, Mildbraed 6749 (holotype, B destroyed; iso- type, HBG). Pavetta gossweileri Bremekamp, Repert. Spec. Nov. Regni Veg. 37: 64. 1934. TYPE: Angola. Nov. 1911, Gossweiler 5211 (holotype, BM). Pavetta hymenophylla Bremekamp, Repert. Spec. Nov. Regni Veg. 37: 68. 1934. TYPE: Tanza- nia. Lushoto District: Amani, Aug. 1911, Grote 3541 (holotype, B destroyed; isotypes, EA not .K finds dalei bo Repert. Spe Nov. Regni Veg. 37: 72-73. 1934. TYPE: Kenya. Mt. Kenya, Oct. 1932, Dale 3054 (holotype, K). Pavetta intermedia Bremekamp, Repert. Spec. Regni Veg. 37: 71. 1934. TYPE: Zaire. Province: Lesse, Mar. 1914, Bequaert (holotype, BR). Pavetta ixorifolia AREA Repert. Spec. Nov. egni Veg. 37: 79. 1934. TYPE: Guinea. Fou- ta Djallon: Dalaba, im 1907, Caille 18137 (holotype, P). Pavetta kasaica Bremekamp, Repert. Spec. Nov. Regni Veg. 37: 69. 1934. TYPE: Zaire. Kasai: Ipamu, June 1921, Vanderyst 10774 (holotype, BR Nov. Kivu 3136 Pavetta micheliana J.-G. Adam, Bull. Inst. Fon- dam. Afrique Noire Sér. A, Sci. Nat. 35: 87- 89. 1973. TYPE: Liberia. Mt. Nimba, Yiti val- ley, Jan. 1965, Adam 20735 (holotype, P not seen; isotype, K) Pavetta micrantha Bremekamp, Repert. Spec. Nov. Regni Veg. 37: 74. 1934. TYPE: “Centralaf- rika" (Zaire, Ituri?). Between Frumu and Ma- wambi, 1908, Mildbraed 2908 (holotype, B de- stroyed; isotype, HBC). Pavetta mollissima Hutch. & Dalziel, Fl. W. Trop. Africa 2: 91. 1931. TYPE: Ghana. 1929, Vigne 1601 (holotype, K). Pavetta monticola Hiern, Fl. Trop. Africa 3: 170. 1877. TYPE: Sáo Tomé & Príncipe. Sáo Tomé: Mann 1074 (holotype, K; isotype, in nitidula Hiern, frican Pl. Collect. by r. Friedrich Welwitsch in 1853-61. Volume ‘ part 2: 486. 1898. TYPE: Angola. Pungo Andongo, 1856-1857, Welwitsch 3189 (holo- Pavetta congensis anges Repert. Spec. Nov. Regni 37: 67. . TYPE: Zaire. Likimi, Feb. 1910, chair 64 EAN BR). Pads. coriacea o Repert. Spec. Nov. Regni Ve . 1934. TYPE: Uganda. Ruwenzori, Nov. i5 a ott x Ello 8310 (holotype, K Pavetta obanica Bremekamp, Repert. Spec. Nov. Regni Veg. 37: 77. 1934. TYPE: Nigeria. Oban, 1911, Talbot 359 (holotype, BM). Volume 83, Number 1 6 149 Manning Pavetta Subgenus Baconia in Cameroon Pavetta oblongifolia eed Repert. Spec. ov. Regni Veg. 37: 65. 1934. TYPE: Séné- gambia. 1838, me» 673 (holotype, K; iso- types, / Pavetta oresitropha Biecidbüing, Repert. Spec. Nov. Regni Veg. 37: 73. 1934. TYPE: Equatorial Guinea. Bioko Island, Nov. 1911, Mildbraed 7135 (holotype, B destroyed; isotype, HBG). Pavetta polyantra Bremekamp, Repert. Spec. Nov. Regni Veg. 47: 21. 1939. TYPE: Gabon. Up- per Ogowé: Lastoursville, Oct. 1929, Le Testu 7541 (holotype, Herb. Le Testu = BM, iso- types, Pavetta pd pius Fl. Trop. Africa 3: 171. . TYPE: Gabon. Sierra del Crystal, July 1862. Mann 1718 (holotype, K). Pavetta redheadii Bremekamp, Repert. Spec. Nov. Regni Veg. 47: 18. TYPE: Zambia. Mwinilunga, Nov. 1937, Milne-Redhead 3429 (holotype, K: isotype, B Pavetta schweinfurthii Bremekamp, Repert. Spec. — w ov. Regni Veg. 37: 66. 1934. TYPE: “Central Africa” (Sudan?). Gr. Seriba Ghaddar, Apr. 1869, Schweinfurth 1341 (holotype, K). Pavetta Mi cen Hooker f.) Hiern, Fl. Trop. Africa 3 177. PYPE: Tanzania. Bukoba yes trict: Karangwe, Feb. 1862, Speke & Grant 422 (holotype, Pavetta NN. Bret wd SAN Bull. 1954. TY 501. Kain a marked 1951. E A not seen, S oie Pavetta oo. Bremekamp, Repert. ind Nov. egni Veg. 37: 66-67. 1934. TYPE: “West Urundi” ee Meyer 1037 (holotype, B SPECIES NOT KNOWN FROM CAMEROON ORIGINALLY DESCRIBED IN BUT SUBSEQUENTLY EXCLUDED FROM PAVETTA SUBG. BACONIA Tarenna funebris (Bremek.) N. Hallé, Adansonia 7: 505. 1967. Pavetta funebris Bremek., Repert. Spec. Nov. Regni Veg. 37: 63. 1934. TYPE: Zaire. Kasai: Mpio-mpio, Aug. 1921, Vander- yst 10221 (holotype, BR). Literature Cited Adam, J.-G. 1973. Espéces nouvelles pour le Libéria et les Monts Nimba (Astéracées et so 'ées). Bull. Inst. Fondam. Afrique Noire Sér. . Nat. 35: 79-91. Ade sed i, J. F. & M. der. T Biaoril Atlas of Africa. Edu Group, Essex, England. Be mham; G. & J. D. Hooker. 1873, reprinted 1965. Gen- LENIN 2 1-151 Bre mekar ing 19 M. A n d of the genus E pert. Spec. . Reg 37: 1-208. 9a. " n of f the genus Paretta L. Additions pe emendations. Repert. Spec. Nov. Regni 47: 12 A monograph of the genus Paretta L. Additions sn dd II. Repert. Spec. Nov. Reg- —98. 17: ni Veg. jn new species of ees L; from trop- ical wa Sonik Africa: IL Kew Bull. 8: 505. — ———, 1056. Rubiacées Nouvelles du i Belge et du Baud Urundi prie Trichostachys. Pavetta). 258 Bull. e Bot. État 26: 253- 9006. bns on the PURUS the delimitation, "pu me la. of the Rubiaceae. Acta Bot. Neerl. : 1-33 Bridson, D. M. 978. Studies in Pavetta (Rubiaceae sub- fam. Cinchonoideae) for Part 2 of b : m cal East Africa: Rubiaceae.’ Kew Bull. 32: 609 Verdcourt. 1988. Rubiaceae e 2). In: R. . Polhill (editor), Flora of Tropical East Africa. A. / n ma, Rotterdam. Netherlands. dest Candolle, A. P. de. 1807. Mémoire sur le Curiera, Genre Pavetta pirungensis Bre mekamp, Bull. Jard. Bot. État 14: Nouveau de la famille des Rubiacées. Ann. Mus. Natl. 1937. TYPE: Zaire. Kivu: Virungu Moun- ist. Nat. 9: 216-222. LEG volcano, Feb. 1932. Lebrun 4952 ^ Faber. F. C. von. 1912. Das erbliche dpi sa e n von | siii BR). Bakterien und tropischen Pflanzen. Jahrb. . Bot Pavetta vanderijstii Bremekamp, Repert. Spec. Nov. 31: 285-375. egni Veg. 37: 68. 1934. TYPE: Zaire? Kik- dire E N. Es R. ` z ess ue - » q in ora O es ropica rica utenin- veil (= Kikwit?), Oct. 1920, Vanderyst 8133 son & J. M. Dalziel (editors). pies ed., Vol. 2. Govern- ments of s ria, Gold Coast, Sierra ete and The (holotype. BR). Pavetta yambatensis Bremekamp, Repert. Spec. ;ambia, Crown Agents for Overseas Governments, Lon- Nov. Regni Veg. 37: 74. 1934. TYPE: Zaire(?). don. . Yambata, Feb. 1914, De Giorgi 1718 (holo- Hickey. L. J. 1973. Classification of E^ arc ur ‘ture of . . dic ee domus leaves. Amer. J. Bot ys: type. BR: isotype, K). Hiern, W. P. 1877. Rubiaceae. Pp. 162109. in D. F. Pavetta zimmermanniana Valeton, Icon. Bogor. 2, Oliver EE Flora of Tropical Africa, Vol. E, Fasc. 3 tab 143. 1904. TYPE: Indonesia, Cul- Reeve, London. tivated in Bogor Botanic Garden as 4603, or- 1898. Pp. 485-488 in Catalogue of the African o us bs plants collected by Dr. Friedrich Welwitsch in 1853- IP HBOBPIRUM TIOE SGEN: 61. Vol. I. Part 2. British Museum of Natural History. des : m — f London. The South Cameroons, Gaboon River Speci- Krause : 1909. Rubiaceae Africanae Il. Bot. Jahrb. men of this species, Mann 961 (K). is from present- Syst. 43: 129-160. Lersten, y R. 1974. Colleter morphology in Pavetta, day Gabon rather than Cameroon. 150 Annals of the Missouri Botanical Garden Neorosea and Tricalysia (Rubiaceae) and its — to the bacterial leaf nodule symbiosis. Bot. J. Linn. Soc 39: 125- nt | pni types in Rubiaceae, especiall o to le cterial leaf nodule symbiosis. Bot. . Linn. Soc. 71: 311 319 —— & H. T. Horner, Jr. 1976. Bacterial leaf nodule symbiosis in angiosperms with emphasis on Rubiaceae and Myrsinaceae. Bot. Rev. (Lancaster) 42: 145-214. Little, R. J. & C. E. Jones. 1980. A Dictionary of Botany. Van Norstrand Reinhold, New York. 1990. Pavetta utu Baconia (Rubi- ssertation, St. Louis Uni- 3 Lo : ivory Micro- films International, Ann Arbor, Mic 'higan S.A. 1. Intraspecific variation in Pavetta rigida (Rubiac eae): Piet i of reliability of taxonomic infor- mation. Ann. Missouri Bot. Gard. 78: 535-538. Robbrecht, E. 1984. The delimitation and taxonomic po- sition of the tropical African genera Leptactina and Dic- RUE a Pl. Syst. Evol. 145: 105-118. 88. Tropical Woody Rubiaceae. Opera Bot. Belg. 1 Seyani, I a . The taxonomy of Dombeya burgessiae complex alae ane in Africa. In: P. Goldblatt & P. P. Lowry II (editors), Modern Systematic Studies in Af- rican oe Monogr. Syst. Bot. Missouri Bot. Gard. 25: 677-6 U. S. Office of Geography. 1962. Gazetteer, Central Af- rican Vir dpa lic. U.S. Government Printing Office, Wash- ington, D.¢ White. n 1979. The Guineo-Congolian Region and its Furia to other phytochoria. Bull. Jard. Bot. État 49: 11-: Zimmermann, A. Ueber Bakterienknoten in den Blüttern einiger Rubisogsn. Jahrb. Wiss. Bot. 37: 1-11. Erratum Baldwin, ba “e G., Michael J. Sanderson, J. Mark Porter, Martin F. Wojciechowski, C aoa sl S. Campbell & Mich ael J. Donoghue. 1995. dence on eras rm phylogeny. Ann. Missouri Bot. 82: 247-277 Gard. An error in the original of Figure 4 in this article has been discovered by the authors. Base 1 of the hypothetical ITS 2 secondary struc- ture in Figure 4 is the last base at the 3’ end of the 5.85 subunit (“C,” position 424 in Baldwin, 1993) adjacent to ITS 2. The numerical designation of each of the variable ITS 2 base positions in Fig- ure 4 should be increased by one (e.g., 19 becomes 20, 25 becomes 26, etc.) and the indicated loca- tions of the variable sites should be shifted two bases forward throughout the diagram. Sites 28, 76, 104, and 176 (designated as 27, 75, 103, and 175 in Fig. 4) should not be indicated as variable. Site 95 should be indicated on the diagram as a variable site (one change, in C. hooveri). The adjustments in positioning of variable sites results in a shift from six pairs to only two pairs of changes that map to directly opposing sites along a stem (sites 20 and 37 and sites 26 and 33). Based on the ITS phylog- eny of Calycadenia, changes at each of the two pairs of sites were independent, unlinked events that occurred in different lineages and do not rep- resent examples of compensatory change to main- tain ITS 2 secondary structure. The above correc- tions do not impact the conclusions about evolution of ITS 2 secondary structure presented in the paper. Volume 83, Number 1, pp. 1-152 of the ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on 25 January 1996. ANN. MISSOURI BOT. GARD. 83: 151. 1996. Experimental and Molecular Approaches to Plant Biosystematics The proceedings of the Fifth International Symposium of the International Organization of Plant Biosystematists (IOPB) Edited by Peter C. Hoch and A. G. Stephenson Twenty-three original contributions that span the breadth of biosystematics, a dynamic field of study that bridges the realms of systematics and population biology. The papers are arranged in four groups, reflecting the original four symposia of the 1992 meeting. DNA and Plant Biosystematics presents of phylogenetic analysis to problems at the species and population levels. Monographs in Systematic Botany from the Missouri Botanical Garden, Volume 53. ISBN: 0-915279-30-4. 416 pp. Illustrated. 1995. $60.00 U.S. $62.00 Non-U.S. Annals of the Missouri Botanical Garden, Volume 82 ; Number 2: Alternative Genes for Phylogenetic Reconstruction in Plants A symposium cosponsored by the American Society of Plant Taxonomists and the Botanical Society of America, organized by Pamela S. Soltis and Douglas E. Soltis, and presented at the 1993 AIBS meetings. Although the chloroplast gene rbcL has been successfully used to reconstruct plant phylogeny, many important questions of plant phylogeny and evolution cannot be addressed using it. The contributors to this issue of the Annals explore the potential of eight alternative genes or DNA regions for phylogenetic reconstruction at a variety of hierarchical levels. Both nuclear and chloroplast genes are evaluated. Three regions of the nuclear ribosomal RNA cistron are explored: the 18S gene, the internal transcribed spacers (ITS), and the 26S gene. Small multigene families from the nuclear genome may also carry phylogenetic signal: the phytochrome gene family and the small heat shock gene family. Three genes from the chlo- roplast genome are also considered: atpB, ndhF, and matK. Each paper describes the location, size, structure, and rate of evolution of the chosen gene and discusses its potential for phylogenetic study. This issue also contains: “The Comparative Pollination and Floral Biology of Baobabs (Adansonia—Bombaca- ceae)” by David A. Baum and “In Memoriam: Peter G. Martin.” Annals 82(2) 1995. 174 pages. $27.50 U.S. $28.00 Non-U.S. 7 To order, please indicate method of payment below. Checks or money orders should be in U.S. funds, payable through a U.S. bank, to Missouri Botanical Garden. Orders must be prepaid. $3.00 handling fee on all orders. Additional $2.00 handling fee will be added to orders requiring invoices. No shipments are made until payment is received. Phone: (314) 577-9534, fax: (314) 577-9594, e-mail: dept] 1@mobot.org. copy(ies) of Monograph No. 53 copy(ies) of Annals 82(2) Send order to: Please send Department Eleven, Missouri Botanical Garden Please send — P.O. Box 299 . Box St. Louis, MO 63166-0299, U.S.A. L] Check/money order enclosed Send books to: O Send invoice ($2.00 fee will be added to total) O Charge card number (MasterCard/Visa) Name Expiration date : Ads ded ur p MN Name as it appears on card Telephone number (daytime) Postal Code Country PRICES ARE SUBJECT TO CHANGE WITHOUT NOTICE 83(1) CONTENTS Systematics Agenda 2000: Systematics and Society, the 41st Annual Systematics Symposium of the Missouri Botanical Garden i Introduction to the Symposium i- -n P. Mick Richardson 1 T The Use of Phylogenetic Perspective in Comparative Plant Physiology and De- — . velopmental Biology ...— — ie Russell K. Monson 3 Systematics Solves Problems in Agriculture and Forestry ———- s Amy Y. Rossman & Douglass R. Miller Yi Fisheries ESANS and Marine Biodiversity — | Michael kea é dux B. “Collette 29 Importance of Systematics to Public Health: Ticks, eee and Disease —— E mes H. Oliver, Jr. 37 Systematics and the Conservation of Biological ee de PR R. E Vena Wright 47 Transforming Ethnobotany for the New Millennium Michael J. Balick 58 — The Prosoeca peringueyi (Diptera: Nemestrinidae) Pollination Guild in Southern Af- rica: Long-tongued Flies and Their Tubular Flowers — oa John C. Maning & Tus ‘Goldblatt 67 Kia al Pavetta Subgenus Haconiy (Rubiaceae: —: in Cameroon ..—— = — Stephen D. Manik 87 .— “Pitas for The. ITS Region of Nuclear Ribosomal DNA: A Valuable Source of Ev- — idence on Angiosperm Phylogeny — M7 ; aU Em Add G. Baldwin, Michael J. sen. J. Mark Porter, Martin F2 Foexga, Christopher S. See & Michael J. Donoghue 151 : ae Cover illustration . Pavett mshensis subsp is S. D. Manning, sp. nov DY Annals ^ ofthe — Missouri Botanical Garden a Y umber Volume 83, Number 2 Spring 1996 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 out- side the Garden will also be accepted. 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. Editorial Committee Henk van der Werff Editor, Missouri Botanical Garden Amy Scheuler McPherson Managing Editor, Missouri Botanical Garden Diana Gunter Editorial Assistant, Missouri Botanical Garden Vicki Couture Secretary Teresa Johnson Publications Order Processor 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 P. Mick Richardson Missouri Botanical Garden For subscription information contact Depart- ment Eleven, P.O. Box 299, St. Louis, MO 63166-0299, Subscription price is $110 per volume U.S., $115 Canada and Mexico, $135 all other countries. Four issues per volume. The journal Novon is included in the t acre price of the ANNALS. scheuler@mobot.org (editorial queries) dept11@mobot .org (orders). 3 O Missouri Botanical Garden 1996 The ANNALS OF THE MISSOURI BOTANICAL GARDEN (ISSN 0026-6493) is published quar- — terly by the Missouri Botanical Garden, ee Tower Grove Avenue, St. Louis, MO 63110. — Second class postage paid at St. Louis, MO and additional mailing offices. POSTMASTER: Send me address changes to ANNALS OF THE MISSOURI es BOTANICAL GARDEN, Department Eleven, P.O. ur Box 299, St. Louis, MO 63166-0299. The mission of the Mb Botanical Garden is to discover and share knowledge about eS d a, their environment, in order to preserve and enrich life. i © Th This paper meets the A 4 AAICL/ALIICC. 720 40 1004 Da à + Lo de -~ Volume 83 Number 2 1996 Annals of the Missouri Botanical Garden Y A REVISION OF DISCOCARPUS (EUPHORBIACEAE)! Sheila M. Hayden and W. John Hayden? ABSTRACT As revised here, ho a is interpreted to consist of three neotropical species: D. essequeboensis Klotzsch, D. gentryi S. M. Hay which is de 4 spend acce ena name, D. brasiliensis Klotzsch ex types are proposed for the two sf ‘ieee others, Following p previous literature. x Müll. scribed and named herein as new to s Foliar anatomy is described w spruceanus Müll. Arg. One essequeboensis. Lecto- cience, and D. is reduced to synonymy of D excluded from Discocarpus, as are h are Arg., shown to occur on both epidermides. Evidence prese 'nted supports close relationships with Lachnostylis Turcz. and M Aubl.: x & K. Hoffm little was found to support previous hypotheses concerning a relationship with Chonocentrum Pierre ex Discocarpus Klotzsch is a genus of trees found in seasonally flooded riparian habitats of northern South America, where they are components of the forest canopy. The plants are dioecious and bear small clusters of flowers in the axils of simple, al- ternate, entire leaves. Discocarpus was first described by Klotzsch (1841) who initially named, but did not describe, two species; he subsequently described one of these, D. essequeboensis Klotzsch (1843), based on Schomburgk collections from the Essequibo River region of Guayana. Omitting two nettles (Urtica- ceae) from Mexico and Nicaragua grossly misplaced in the genus, the next species of Discocarpus to be named was D. spruceanus Müll. Arg. (1863), based on collections of Richard Spruce from the Rio Ne- gro of Brazil. Some 32 years after being first men- tioned by Klotzsch, D. brasiliensis Müll. Arg. (1873) was formally named, based on a collection of von Martius from the early 19th century. Three taxa were added to the genus in the 20th century. The first addition was D. hirtus (L. f.) Pax & K. Hoffm. (Pax & Hoffmann, 1922), a consequence of syn- onymizing the South African genus Lachnostylis with Discocarpus. In current literature, however, Lachnostylis is treated as distinct from Discocarpus (e.g., 1986; Webster, 1994b). The two most recently described species, Levin, Mennega, 1987 ! Financial support for the master's thesis research upon which this study is based was gene Bus provided by the T School of Arts ¿ and Sciences at the University of Richmond. We thar ik Rafael de Sá, Miles F. Joh inson, Geoffrey . Levin, Gary Radice, Dean a Donna M. E. Ware, and a anonymous reviewer for assistance, and the curators of BM, BR, C, E, F, G, GH, GOET, ILLS, K, L, LD, M, MANCH, MICH, MO, NY, OXF, P, m RB, RSA, S, TCD, U, UC, URV, US, and W for the loan of he rba rium specimens ? Department of Biology, University of Richmond, Ric desd Virginia 23173, U.S.A. ANN. MISSOURI Bor. GARD. 83: 153-167. 1996. 154 Annals of the Missouri Botanical Garden D. mazarunensis Croizat (1948) and D. duckeanus Jabl. (1967), were based on South American ma- terial. Jablonski (1967) accepted the five South Amer- ican species noted above as distinct entities con- stituting Discocarpus, but his treatment indicated several gaps in the available data on these plants. For example, staminate or pistillate flowers were undescribed for several species. Furthermore, re- cent studies reveal that two of the species accepted by Jablonski were misplaced in Discocarpus, and presently available collections indicate the exis- tence of a previously unrecognized species (Hay- den, 1995). In addition, the issue of generic rela- tionships is unresolved. Discocarpus was classified most recently in subfamily Phyllanthoideae Asch. Wielandieae Baill. ex Ab), but earlier concepts of generic relation- ships have varied widely (see below). Moreover, re- tribe ebster, ports of foliar sclereids in Discocarpus and Amanoa (Gaucher, 1902; Levin, 1986) suggest the new pos- sibility of placement in tribe Amanoeae. This paper provides a revision of Discocarpus, including de- tailed descriptions of foliar anatomy and discussion of relationships. MATERIALS AND METHODS This revision is based on a total of 171 herbar- ium specimens of Discocarpus borrowed from 29 herbaria in the United States, Europe, and South America. Small samples of leaf tissue from the fol- lowing collections were removed for anatomical study: Discocarpus essequeboensis Klotzsch, Jan- goux & Bahia 294 (NY), Krukoff & Fróes 11974 NY), Maas et al. 7395 (U), Schomburgk 659 (U), Silva 4776 (NY), Smith 2692 (F); Discocarpus gen- tryi S. M. Hayden, Encarnación 25065 (F), Vázquez & Jaramillo 5487 (NY); Discocarpus spruceanus üll. Arg., Davidse 27631 (NY), Wurdack & Ad- derly 43349 (NY). Half of each sample was mount- ed directly on stubs and sputter coated with a gold/ palladium mixture prior to observation with SEM. The second half of each sample was rehydrated by boiling in water with a few drops of Aerosol OT, dehydrated in tertiary butanol, embedded in par- affin, and sectioned at 10 ¡um on a rotary micro- tome. Paraffin sections were stained in toluidine blue (Berlyn € Miksche, 1976) or a combination of saffranin and haematoxylin (Johansen, 1940). E SYSTEMATIC TREATMENT Discocarpus Klotzsch, Archiv. Naturg. 7(1): 201. 1841. TYPE: Discocarpus Klotzsch. Figures 14. essequeboensis Dioecious trees (or shrubs), (3-)10—30 m tall, DBH 2 silvery gray to dark purplish red; lenticels raised, 5-100 cm. Twigs glabrous to short-pilose, elongate, parallel with the axis; terminal buds acu- minate, cylindric, glabrous to tomentose, 3-7 mm long, often with two basal knoblike protrusions, Leaves alternate, simple, petiolate, glabrous, leathery; petioles 4-8 mm long, wrinkled: margins entire; base obtuse: sometimes sexually dimorphic. apex acute to acuminate; venation pinnate; ultimate veins reticulate, orthogonal. Stipules fugaceous. In- florescence axillary, l-several flowers per node; flower clusters sessile, subtended by cupulate bracts; bracts ca. | mm long, 1 mm wide, glabrous to pubescent; staminate clusters several per node; pistillate clusters one per node. Staminate flowers sessile, congested, 10-30 per node; sepals (4) 5, 1.5-2 mm long, 1-1.5 mm wide, pilose; petals (0— )5, delicate, hyaline, glabrous to pubescent, less than 1 mm long, linear, often fringed apically; disk extrastaminal, lobed; stamens (4) 5; filaments fused below the level of the disk, free portions 1.5-3 mm long; anthers elongate, 1 mm long, longitudinally dehiscent, exserted; pistillode segmented into two or three linear, pubescent, membranous filaments. Pistillate flowers 1—3(—5) per node; pedicels essen- tially lacking to 5 mm long; sepals 5, cupulate, 1.5- 3 mm long, 1-2 mm wide, densely pubescent; pet- als (0-)5, hyaline, 0.5-3 mm long, up to 1 mm wide, pubescent; disk slightly lobed; ovary 3-car- pellate, subglobose, smooth or sculpted, densely pubescent; styles 3, parted to the base or nearly so, spreading horizontally, densely pubescent below; stigmas 3, dilated, lobed, horizontal or reflexed; ovules 2 per locule. Fruits symmetrically 3-lobed or asymmetrically subglobose, 6-12 mm tall, 6-15 mm diam., longitudinally dehiscent EN a 3, or 6 mericarps, thick, hard, brittle; surface smooth to nae wale —3-seeded; pericarp ca. densely pubescent; columella persistent. Seeds subglobose, ecarunculate; testa thin, shiny. Discocarpus can be distinguished from other woody genera of subfamily Phyllanthoideae by the combination of: deciduous stipules; dioecy; minute petal-bearing flowers produced in axillary clusters; lobed extrastaminal disks; finely reticulate exine on pollen grains; and styles that are not bifid, but ter- minate in three dilated, irregularly lobed stigmas. It is noteworthy that floral merosity and presence of petals are somewhat inconstant. Most flowers are 5-merous, but 4-merous flowers are not infrequently encountered. Most flowers examined showed petals to be present and isomerous with the sepals, but sometimes fewer petals, or even none, will be Volume 83, Number 2 Hayden & Hayden 1996 Revision of Discocarpus WET IDO , 745 sae abe ^ Figures 14. Floral morphology of Discocarpus, SEM. 1, 2, Discocarpus spruceanus Müll. Arg Spruce 3527 (G). — l. Staminate flower with sepals and petals removed to uh fused filaments; bar = 1 mm. —2. Fuse d es aments; bar = 250 um. —3. Discocarpus essequeboensis Klotzsch, Martius s.n. (L). Hair on fruit surface; hat 5 1 Discocarpus M Klotzsch, Schomburgk 35 (G). Dilated, lobed, reflexed stigmas: bar = 1 mm. found. Notably, the minute size and flimsy texture ing highly regular minute areoles that lack free- of the petals render them easy to overlook. Impor- ending veinlets. tant clues to recognizing the genus in sterile con- In Webster’s (1994b) key to genera of Wielan- dition include the elongate terminal buds with — dieae, two aspects of morphology of staminate flow- paired distinctive basal knobs and fine veins form- ers attributed to Discocarpus are at variance with 156 Annals of the Missouri Botanical Garden 80 70 po 50 40 0 0 SS ë į e s e = ———— — as A (a Y > — = i ora S a T e 2 k - C 10 4 ; ir . / Si o » i ant m N m i a): o ; 3 [] LN Ua f 10 . P F \ 2 Ea EN Yo f S e Ps TUE e / Figure 5. ceanus Mill. Arg.; triangles = Discocarpus gentryi S. M. Hay the above description. Staminate flowers are sessile and bear filaments fused to the base of the pistil- lode, features also noted by Pax and Hoffmann 1922) Three species of Discocarpus occur in the Ama- zon and Orinoco River basins of Brazil, Colombia, Peru, and Venezuela, as well as smaller rivers o Guyana and Surinam (Fig. 5). Habitat is lowland rainforest below 250 m, along seasonally flooded riverbanks of vdrzea forests or occasionally inun- dated fields. Discocarpus species can form canopy trees, but samples are sometimes collected from specimens described as small shrubs. KEY TO SPECIES OF DISCOCARPUS (FIGS. 6-9) la. Terminal bud of staminate specimens lacking basal knobs, covered with dense indumentum; ovary a fruit surface smooth; fruits ed, 2 carpels usually abortive; widely sc aered in un "as and Orinoco River basins D. spruceanus Mull Arg. lb. Terminal bud of staminate specimens with two prominent basal knobs, glabrous or only sparsely pubescent; ovary and fruit surface sculpted; fruits 3-seeded, all carpels accrescent. 2a. Pistillate flowers on pedicels 45 m ovary and fruit surface deeply sculpted into N 1 Y ^ T m long; Distribution of Discocarpus. Squares = Discoc es akii Klotzsch; circles = Discocarpus spru- der long undulate ridges with sharp crests (be- neath dense indumentum); fresh staminate flowers yellow; Amazonian Peru and western razil D. gentryi S. M. Hayden . Pistillate flowers sessile or nearly so; ovary and fruit surface weakly sculpted into mur- es with rounded € staminate »uyana, Surinam, and northeastern Brazil D. essequeboensis Klotzsc h ——— essequeboensis Klotzsch, London J. : 52. 1843. TYPE: Guyana. On branch i upper Essequibo River, 'Schomburgk 35 (lec torpor selected here, BM; isolectotypes, C, K, OXF, P, U, W). Figures 10, 3, 4, 9. 873. SYNTYPES: Brazil. Bahia: near Villa do Rio de Contas, Martius s.n. (G, L, M( ahia: Martius s.n. (M) Disc rea pes brasiliensis Klotzsch ex Müll. Arg. Mart. Fl. 16 3: Trees, 10-20 m tall, DBH 25-100 cm. Terminal buds similar in staminate and pistillate specimens, glabrous, with two basal knobs. Leaves 8-22 cm long, 4-48 cm wide; apex acuminate (to 10 mm long) to merely acute. Inflorescence bracts glabrous to pubescent. Staminate flowers 10-30 per node, Volume 83, Number 2 1996 Hayden & Hayden Revision of Discocarpus 157 Figures 6-9. 18419 (F). pisa —1. Discoc Müll. Arg., Spruce 3527 (P). carpus gentryi S. M. Hay Ovary and fruit surface fe atures of Discocarpus. pu Discocarpus gentryi S. M. iln n, Gentry et al. den, Revilla 41 Mature fruit with one bd veloped and two aborted lobes. —9. Discocarpus essequeboensis 1 (NY). Mature fruit. —8. Discocarpus spruceanus + eed :h, Se erii: 706 (G). Mature fruit. All bars = 2 mm. cream-colored; sepals (4) 5. 2 mm long. 1.5 mm wide; petals (0—)5, pubescent; disk lobed; stamens (4) 5; pistillode segments 3. Pistillate flowers 1-5 per node; pedicel 0-1 mm; sepals 3 mm long, 2 mm wide, light green; petals reduced, (0—)5, 0.5-3 mm long; ovary shallowly muricate to reticulate: stigmas reflexed. Immature fruits dark red; mature 3-lobed, 3-seeded, surface weakly sculpt- fruit brown, symmetrically 8-9 mm tall. 10-15 mm diam., ed into shallow muricae Seed 6-8 mm diam.: pylar and basal surfaces somewhat flattened: testa or short undulate ridges with rounded crests. micro- red-brown. Distribution. Central and eastern Brazil, Guy- ana, Surinam (Fig. 5); on sandy soil in frequently inundated forest along rivers (várzea) and periodi- cally flooded fields. Flowers have been collected s.n. (G. L, M). from June through December; fruits, from Septem- ber through December. Common Names. “Square Wood" (in reference to shape of trunk, Anderson 408); “Oity do Campo” (Fróes & Krukoff 11974) near uu ns examined. BR nay a Bastos 201 Amazonas: Martius s.n. Maué Pires 109 A 5 ] Bahia: Villa do Rio de ase Martius Goias: Rio Araguaia at mouth of R Silva 4862 (ILLS, NY); Rio P oe Rio Alto Turiagu, Nova es Ae a 2°5'S.-45 . Jangoux & Bahia 294 (NY, RB); Rio Pin- dare B asin, We. Fróes & Krukoff 11974 7 H, MICH, NY, US); Rio Mearim-Lapela, municipality de Vitória do mearim, e ampo Coberto, Silva 4191 (RB). Mato Grosso: Rosa & do 2] 19 Ea NY). > jac hogi ira io Ja- iranha. Silva 4776 margin of Rio Juruena, Pará: Marabá, Fróes & Rio Trombetas, Ducke 8953 (BM, G); Ducke En Trom- 7988 (BM); Rio Trombetas margin, Porteria, betas margin, 158 Annals of the pes Botanical Garden Figure 10. Discocarpus essequeboensis Klotzsch. —A. Disk from staminate flower; Pires 109 (U). —B. Staminate flower, one sepal removed; Pires 109 (U). —C. Habit, staminate specimen; eine pea (NY). —D. Habit, a specimen; Rosa & Santos 2149 (MO). —E. Immature fruit: Schomburgk 35 (W). —F. Terminal bud; Krukoff 11974 (NY). Volume 83, Number 2 Hayden & Hayden 159 Revision of Discocarpus Ducke 7993 (BM); National park of Tapajós, 60 km from Itaituba-Jacarecanga at the margin of Rio Tapajós, Silva & Rosário 3992 (NY). GUYANA. Upper Essequibo River, Schomburgk 706 (BM, G, K, L, P, U, W); upper Essequibo i y (BM, E, F, C, K, L, MANC Rupununi Savanna, near Maricouba pond r Karanambo Ranch, 3°45'N, 59°19'W, Górts-van Ryn et al 388 (URV); Rupununi District, oap cae. , Kuyuwini River, forest along river, 2°5'N, 59*15'W, Jar sen- Jacobs et x 2903 (URV); Manakobi, (ae Riv- er, Anderson 408 (K), Schomburgk s.n. (L, U), Se homburgk 1237 (F), Mong og (F. G. K, P. W); Cuyuni-Ma- l- zaruni Region, Ess iver m downstream « Omai, 5°26' N, 38°42" W. Gillespie 1573 (MO): spank River, Monkey Pond landing SW of Mt. Makarapan, 3°53'N, 58°55'W, Maas et al. 7395 (P, U); basin of Es- sequibo River near mo outh of Onoro Creek, 1°35'N, Smit 2692 (F, G, NY). SURINAM. His Corantjne, B. W. 2044 (U); Tapanahoni, Kappler 9 U, W), Kappler 2143 (GOET, W), Schomburgk 459 d Notes. Klotzsch's boensis was originally based upon three collections, Schomburgk 35 (pistillate flowers), Schomburgk 659 (staminate flowers), and Schomburgk 706 (mature ruits). Since staminate material of Discocarpus re- species Discocarpus esseque- veals few diagnostic characters, and since the fruit- ing specimen bears only fragmentary label data, the pistillate collection, Schomburgk 35, is by far the best choice to typify the species. The BM specimen includes abundant flowers that prove important in defining the species (see below). The spelling of the specific epithet adopted here follows that used by Klotzsch (1843) in contrast to that of the earlier nomen nudum, ea essequiboensis Klotzsch (Archiv. Naturg. 7(1): 201. 1841). Discocarpus die iis is here placed in syn- onymy under Discocarpus essequeboensis. The two entities are virtually indistinguishable, and it seems that recent practice has been to identify material from the Guianas as D. essequeboensis and Brazilian specimens as D. brasiliensis. In the past, Discocar- pus brasiliensis was supposedly distinguished from D. essequeboensis by the presence of small bumps or muricae on ovaries and fruits of the former in contrast to the smooth ovaries and fruits of the lat- ter (Müller, 1873; Pax & Hoffmann, 1922). How- ever, the syntypes of D. essequeboensis listed by Klotzsch include a fruiting specimen, Schomburgk 706, which has a surface texture that is obviously bumpy and identical to other specimens identified as D. Talaik Further, the lectotype of Disco- carpus essequeboensis, Schomburgk 35, illustrates a range of developmental stages from very young flowers just emerging.from the bud to early fruits. The flowers on this specimen reveal a developmen- tal change in ovary surface from nearly smooth to contoured or bumpy as the ovary matures. On the basis of these observations, and given the lack of any other consistent character of either pistillate or staminate material of D. essequeboensis and D. bras- iliensis that delineates two separate species, the de- cision was made to place Discocarpus brasiliensis in synonymy under Discocarpus essequeboensis. Both names were first published by Klotzsch; however, the name D. brasiliensis, published in 1841, re- mained a nomen nudum until Müller provided it with a diagnosis in 1873. Thus D. essequeboensis, published in | 1843, is the oldest legitimate name for this spec In addin. to those named Discocarpus bras- iliensis Klotzsch, other specimens collected by Mar- tius bear the name D. bahiensis Klotzsch, but it appears that the latter name has never been pub- lished. It is noteworthy that, aside from these nearly 200-year-old collections by Martius, no other col- lections of Discocarpus have been seen from the Atlantic coastal forest of Brazil. Discocarpus gentryi S. M. Hayden, sp. nov. YPE: Peru. Loreto: Santa María de Nanay, SW of Río Nanay, Schunke V. 2443 a F; isotypes, G. GH, NY, US). Figures 11, 6. Arbor vel frutex dioeci ia, 3-14 glabra; flores staminati lutei; flores pistillati 1 (2) per nod- um; pedicellus florum pistillatorun m 4-5 mm longus; fruc- tus trilobus symmetricus, seminibus tribus; pons undulatum profunde, viride: testa cinnamome m; gemma terminalis Trees or shrubs, 3-14 m tall, DBH 35—40 cm. Terminal buds similar in staminate and pistillate specimens, glabrous to sparsely pubescent, with two basal knobs. Leaves 7-15 cm long. 3-7 cm wide; apex acute to acuminate. Inflorescence bracts glabrous to somewhat pubescent apically. Staminate flowers 15—30 per node, bright yellow; sepals 2 mm long, 1 mm wide; petals (0—)5, glabrous to sparsely ubescent; disk with fingerlike lobes; stamens 5; pistillode segments 2-3. Pistillate flowers 1 (2) per node; pedicels 4-5 mm long; sepals 3 mm long, 2 mm wide; petals 5, 2 mm long, 1 mm wide; ovary surface deeply undulate; stigmas reflexed. Fruit green, symmetrically 3-lobed, 12 mm tall, 15 mm iam., 3-seeded, surface deeply sculpted into long undulate ridges with sharp crests. Seeds ca. 8 mm diam.; micropylar and basal surfaces somewhat flat- tened; testa golden brown. Distribution. Amazonian Peru and western Bra- zil (Fig. 5); on white sand or clay soil of low, sea- sonally inundated forest along rivers (vdrzea or ta- huampa); 120—150 m altitude. Flowers have been collected from December through April; fruits, from September through February. 160 Annals of the Missouri Botanical Garden E a = E he pn ge Figure 11. Discc sige p gentryi S yder ninate flower; Rimachi Y. 3300 (NY). —B. Habit, staminate specimen; Rimachi Y. 3300 (MO). EN joie ia Ye Y. 3300 (MO). —D. Disk from staminate flower; Rimachi Y. 3300 (NY). —E. Habit, pistillate specimen; Gentry et al. 18419 (F). —F. Fruit; Revilla 411 (F). Volume 83, Number 2 1996 Hayden & Hayden 161 Revision of Discocarpus Common Names. “Ucuchahuasi” (Vásquez «€ Jaramillo 5487); “Loromicuna” (Ayala 1415). Additional specimens examined. BRAZIL. Amazo- nas: Rio Negro near Ilha Provedengia, Steward et al. 516 (NY). PERU. Loreto: Maynas, Iquitos, Río Nanay, Que- brada de Morropon, Rimachi Y. 3281 (F, MO, NY, RSA); Río Nanay, 03%51'S, 73°32'W, Vásquez et al. 7528 (F, NY); Río Nanay at Almendras, 03°48'S, 73°25'W, Vásquez & Jaramillo 5487 (F, MO, NY); lia in A gorge of the small settlement of San Pablo de Cuyana above Santa Cla- ra de Nanay, iir ae Y. 3300 (F, MO, NY, RSA); Caño Iricahua, below o Herrera, on the left margin of Río Ucayali, Encarnación 25065 (F); arie Río Nanay, 8 bends in the river above de Morona Cocha, Revilla 411 (F, MO, NY); vicinity of Iquitos, Revilla 3598 (F, MO); Río Itaya below San Juan de Muniches, 40 mins with 40 hp motor, Gentry et al. 18419 (F, gin of Zungarococha, primary forest, Ayala 1415 (MO). Notes. ferred here to D. gentryi have only been collected Peruvian specimens of Discocarpus re- within the last 30 years. The genus was not treated in MacBride’ (1951) earlier compilation of Eu- phorbiaceae for the Flora of Peru, although the presence of D. brasiliensis was predicted. When specimens from Peru with sculpted fruit surfaces first came to light they were identified as D. bras- iliensis, and they are referred to as such in Brako and Zarucchi’s (1993) checklist. However, D. bras- iliensis is herein synonymized with D. essequeboen- sis, and, further, the Peruvian collections prove to be distinct both morphologically and geographically from this species. As noted in the key, fruits of D. gentryi have pronounced surface relief, and pistil- late flowers are distinctly pedicellate. In contrast, fruits of D. essequeboensis, though somewhat sculpt- ed, are smoother, and pistillate flowers are sessile or nearly so. Additionally, mature capsules of D. essequeboensis are brown and the seeds are dark brown, in contrast to the mature capsules of D. gen- tryi, which are green and contain golden brown seeds. Careful dissection of staminate flowers of D. gentryi reveals disk lobes much more elongate than those of the other two species. Discocarpus gentryi has been collected most frequently along river- banks near Iquitos, Peru, especially in the vicinity of Río Nanay, a blackwater river. There is one ad- ditional record of the species from western Brazil, along the Rio Negro, another blackwater river. The specifie epithet commemorates Alwyn H. Gentry (1945-1993), for his many important con- tributions to the floristics of Central America and northern South America. Gentry’s collection of Dis- cocarpus from the region around Iquitos, Peru, was instrumental in recognizing these plants as new to science. Discocarpus spruceanus Müll. Arg., Linnaea 32: 78. 1863. TYPE: Brazil. Amazonas: Rio Negro above the mouth of the Casiquiare River, Spruce 3527 (lectotype, selected here, BM; iso- lectotypes, BR, C, E, F, G, GH, GOET, K, LD, MO, NY, OXF, P, TCD, W). Figures 12, 1, 2, 8 dicas eire 7n eig Phytologia 3: 34. "deis TYPE: zil. : Municipality Humayta, near Liv- ramento, on Rio 1 Avramento, Krukoff 6703 a NY; isotypes, G, US). Trees, 10-30 m tall, DBH 30-60 cm. Terminal buds sexually dimorphic, densely pubescent with very small to no basal knobs in staminate trees, usually glabrous to sparsely puberulent and with two basal knobs in pistillate trees. Leaves 5-12 cm long, 2-5 cm wide; apex acute to acuminate. Inflorescence bracts pubescent. Staminate flowers 15-30 per node, pale yellow; sepals 4-5, 1.5 mm long, 1 mm wide, pubescent; petals 4-5, glabrous to sparsely pubescent; disk irregularly lobed; sta- mens 4-5; pistillode segments 2-3. Pistillate flow- ers 1-3 per node; pedicels 1-5 mm long; sepals 1.5 mm long, 1 mm wide; petals 5, 1.5 mm long, 0.5 mm wide; ovary smooth; stigmas horizontal. Fruit brown, subglobose, asymmetrically 3-lobed by abortion of 2 (1) carpels, 6-9 mm tall, 6-7 mm diam., 1(-2)-seeded, surface smooth. Seed shape and dimensions unknown (usually shriveled in herbarium specimens); testa brown. Distribution. Widely scattered in the Amazon and Orinoco River basins of Brazil, Colombia, and Venezuela (Fig. 5); in várzea or rebalse vegetation of frequently inundated forest along rivers; often locally abundant. Flowers have been collected from January through August; fruits, in November and December. Additional specimens examined. BRAZIL. Amazo- nas: Rio Negro, northern Brazil, Spruce 3781 pro parte q BR, C, E, F, G, G . GOF , K, MO, 2 : TCD oma (RB). Pará: óes & Black 24513 (U); proximity of Conceição do A aguaia, 8°44'S, 49?26' W, Mileski ipd y B). nr OMBIA. poe 2 km S of Solano, 8 km SE of Tres Esquinas n Río Caquetá below uen of Río Ontequara Little & Little 9004 (US). VENEZUE s: Río Guain- ia between Comunidad and ene Rita, Wurdack & Ad- derley, 43349 (NY, S, US); Caño prc 25 km S of San Cargos of Río Negro, 1°38’N, 8'W, Liesner 8634 (MO, NY); Departamento Río Neg ro, lower part of the Río Baria, 1°27'-1°10'N, 66732'-66725'W, Davidse 27631 (F, MICH, MO, NY); n Atabapo, Riv- erina del Caño Yagua, 03°37'N, 5'W, Marin 479 (MO). The syntype collections of Discocarpus spru- 162 Annals of the Missouri Botanical Garden A Figure 12. Discocarpus spruceanus Müll. Arg. —A. Hi staminate specimen; Mileski 120 (RB). —C. Staminate flower, on staminate flower; Ducke 904 (F). —E. Terminal bud, pistillate specimen; Spruce 3527 27631 (F). —G. Habit, pistillate specimen; Spruce 3527 (BM). Habit, staminate specimen; Ducke 904 (F). —B. Terminal bud, e sepal removed; Ducke 904 (F). —D. Disk from M). —F. Immature fruit; Davidse E ceanus, Spruce 3527 and Spruce 3781, are often versed at G. Further, in many herbaria both curated together leading to potential confusion of | Spruce collections are mounted on the same sheet, the two. Generally, specimens labeled as Spruce and fragments from the two are often mixed to- 3781 are staminate and those labeled Spruce 3527 gether in the same packet. An additional con- bear fruits, but the numbers were apparently re- founding factor is that a portion of the duplicates Volume 83, Number 2 1996 Hayden & Hayden 163 Revision of Discocarpus of Spruce 3781 has been recognized as the type of a different plant, Chonocentrum cyathophorum (Müll. Arg.) Pax & K. Hoffm. (Indeed, one of the two collections of Spruce 3781 received from OXF was a specimen of Chonocentrum misidentified as Discocarpus spruceanus; types of Chonocentrum cyathophorum have been seen from OXF, G, and NY, confirming that these plants are not Discocar- pus.) Of the two syntypes, the fruiting collections that constitute Spruce 3527 are by far more diag- nostic than the staminate collection, Spruce 3781, and therefore the former serves better as the type. This selection also avoids possible confusion with Chonocentrum. Of the several duplicates seen, the specimen from BM is particularly representative. The sexual dimorphism in terminal buds of D. spruceanus is remarkable given that species of Dis- cocarpus are otherwise so similar to each other veg- etatively. The question whether staminate and pis- tillate material cited here truly pertain to the same species cannot be dismissed lightly. In southern Venezuela near the Casiquiare-Negro confluence, pistillate plants with the characteristic smooth-sur- faced partially aborted fruits occur with staminate plants with hairy buds. No pistillate material of oth- er species has been collected from this area. Fur- ther, staminate plants with densely hairy buds do not occur in the Guianas and eastern Brazil, where all pistillate specimens prove to be D. essequeboen- sis, and they are similarly absent from Peru, where pistillate specimens are D. gentryi. We therefore interpret staminate specimens with hairy buds and pistillate material with smooth fruits and just a sin- gle fertile carpel to be conspecific. At this writing, we have seen one record of Dis- cocarpus spruceanus from Colombia, a flowering sta- minate specimen, Little & Little 9604, collected in 1945 EXCLUDED SPECIES Discocarpus duckeanus Jabl., Mem. New Yor Gard. 17: 85. 1967. TYPE: un es S — Chaetocarpus echinocarpus (Baill.) Ducke. Jablonski (1967) based his species on a single collection bearing staminate flowers that, unlike genuine Discocarpus, has consistently petal-less flowers with eight (or more?) filaments fused into a central column, and subglobose anthers diverging at various levels. Although present in Discocarpus, fusion of filaments is restricted to the base of the flower; the filaments, never more than five, diverge at the same level and terminate in distinctly elon- gate anthers positioned at approximately the same height. In all respects, Jablonski's species matches Chaetocarpus echinocarpus (Acalyphoideae). Discocarpus mazarunensis Croizat, Bull. Torr. Bot. Club 75: 400. 1948. TYPE: Fanshawe 2124 (NY) = Chaetocarpus schomburgkianus (Kuntze) Pax & K. Hoffm When Croizat (1948) named his new species, based strictly on staminate material, he noted that its vegetative features were discordant with those of Discocarpus. Jablonski (1967) accepted D. ma- zarunensis without comment. In recent years, Mi- chael Huft annotated several specimens of Disco- carpus mazarunensis as Chaetocarpus schom- burgkianus, and Gillespie (1993), following his lead, excluded this species from Discocarpus. As in the case described above, the flowers in the type of D. mazarunensis have filaments fused into a prominent staminal column with anthers diverging at different levels. Exclusion from Discocarpus is thus justified. me mexicanus Liebm., Skr. Vide i - . Christiana, Math.-Naturvidensk. Kl. 309. 1851 = Laportea mexicana (Liebm) Wedd. (Urticaceae) (as per Pax & Hoffmann, ). 1922 Discocarpus nicaraguensis Liebm., Skr. Vidensk. - els ristiana, Math.-Naturvidensk. KI. 309. = Laportea nicaraguensis (Liebm ) ) Wedd. (Usticas eae) (as per Pax & Hoffmann, 1922). FOLIAR ANATOMY Leaf anatomy was found to vary little from spe- cies to species, hence the following descriptions pertain to all three species. Dimensions cited are average values based on 10 measurements of each feature per specimen. Epidermis (both adaxial and abaxial) uniseriate; cells irregular, partially sclerified; outer periclinal walls sclerified; anticlinal walls wavy, sclerified un- evenly, thicker toward the surface, thinner toward the mesophyll (occasional cells of adaxial epi- ermis sclerified on inner periclinal wall, thinner along anticlinal wall outward); outer periclinal walls bearing subcuticular micropapillae (best seen in Discocarpus spruceanus, Figs. 1 . 16). Ad- axial epidermal cells 15 ¡um thick, Me dms bear- ing tannin deposits; cuticle 1-2 jum thick. Abaxial epidermal cells 11 ¡um thick, occasionally bearing tannin deposits; cuticle «1 ¡um thick. Stomata re- stricted to the abaxial epidermis, densely crowded, 164 Annals of the Missouri Botanical Garden Figures 13-17. (NY). —13. Leaf cross se = 50 um. —14. Pos =í is below vein; ectic hue = 2 bar = 10 m wall of cian a i oriented randomly, widely elliptic, 18 ¡um long, 15 jum wide; anticlinal walls forming stomatal pore mi- nutely crenulate (Fig. 17); subsidiary cells brachy- paracytic. Mesophyll stratified, throughout; palisade cells well developed, some tannin deposits scattered erified adax m. —15. Sclerified ae ds m T. Discoc "pm gentryi S. M. Hayden, Rimachi Y. 3281 (NY bar — Foliar aan of Discocarpus. | us Discocarpus spruceanus Müll. Arg., Wurdack & G iiis = 10 um ial epidermis with micropapillae; Sclerified abaxial epidermis with mic e S . SEM of stomate with crenulate anticlinal —16 times lightly sclerified near the adaxial epidermis; spongy layer weakly developed, vertically oriented intercellular spaces large, druses present (Fig. 13) arge veins composed of concentric arcs of xy- lem and phloem bounded above and below with groups of fibers; small veins vertically percurrent Volume 83, Number 2 1996 Hayden & Hayden 165 Revision of Discocarpus by fibrous bundle sheath extensions sheathed with a single layer of parenchyma (Figs. 13, 20): cells of the parenc wis sheath frequently bearing pris- matic crystals reoles well developed, quadrangular, el (Fig. The combined presence of many sclerified cells, thick-walled fibers, prismatic crystals, and druses render leaves of Discocarpus physically tough and durable. While the anatomical preparations de- scribed above conform generally with previously published information, neither Gaucher (1902) nor Levin (1986) mentioned the existence of sclerified cells in the adaxial layer. DISCUSSION OF RELATIONSHIPS In this century, Discocarpus has been assigned to subtribe Discocarpinae of tribe Phyllantheae (Pax & Hoffmann, 1922, 1931). Kóhler (1965) sug- gested placement in Bridelieae. Hutchinson (1969) did not include the genus in any of his proposed tribes. Most recently, Webster (1975, 1994b) has placed Discocarpus in tribe Wielandieae, an assem- blage of primitive mostly petal-bearing phyllan- thoid genera. At the generic level, Discocarpus has been most closely associated with Lachnostylis and Chonocentrum; in Webster’s (1994b) classification, all three genera are placed in Wielandieae and key out adjacent to each other. In fact, the South Afri- can genus Lachnostylis was combined with Disco- carpus by Pax and Hoffmann (1922 The small trees and shrubs of Lachnostylis grow in much drier habitats than neotropical Discocarpus and, thus, the plants appear different superficially. However, when one looks beyond the much smaller leaves and highly branched stems, details such as flowers, areolation, and shape of terminal buds sup- port Pax and Hoffmann’s (1922) earlier view. To distinguish Lachnostylis from Discocarpus, Webster (1994b) cited thin styles, disks, and stamens adnate to the pistillode in the pubescent staminate former. However, cursory examination reveals thick styles in Lachnostylis similar to those of Discocar- pus in at least some specimens, and pubescence of the disk may be little more than a reflection of the overall hairier aspect of Lachnostylis. Most impor- tantly, as documented herein for Discocarpus (Figs. 1, 2), staminate flowers of both genera have connate filaments adnate to the base of the pistillode. Hence, the relationship between Lachnostylis and Discocarpus seems extremely close. In contrast, relationship with Chonocentrum is much less likely. Chonocentrum first became asso- ciated with Discocarpus by accident. As discussed above, type collections of Discocarpus spruceanus and Chonocentrum bear the same collection num- ber, Spruce 3781, as a result of mixing these clearly distinct plants. Chonocentrum is still known only from the type collection, so comparative data are scarce, and Webster (1994b) considered any pos- sible relationships with this genus to be uncertain. Cursory examination of several isotypes of the only species in the genus, C. cyathophorum, shows this plant to be clearly distinct from Discocarpus. The cuplike fused calyx, complete absence of petals, and large funnelform pistillode contrast he n with Discocarpus and have no counterpart in Wie- landieae. Although Webster (1994b) stated that the pollen of Chonocentrum is unknown, Punt (1962) placed the genus in his “Antidesma type,” noting that the grains are “quite different” from those of Discocarpus. Given its ament-like staminate inflo- rescence, fused calyx, and absence of petals, Chon- ocentrum keys readily to Websters (1994b) tribe Antidesmeae, a context within which further com- parative studies should prove fruitful. The cuplike calyx of staminate flowers of Hyeronima Allemao (Franco R., 1990) and the funnel-like pistillode of Cyathogyne Mill. Arg. (Pax & Hoffmann, 1931), both members of subtribe Antidesmineae, appear directly comparable to structures found in Chono- centrum, Two previously unappreciated characters may serve as synapomorphies that argue for a novel tax- onomic placement of Discocarpus (including Lach- First, neotropical Discocarpus and Amanoa share the nostylis?) near Amanoa (tribe Amanoeae). unique feature of sclereids in the epidermis, which is otherwise unknown in the Euphorbiaceae (Gaucher, 1902; Levin, 1986) and extremely rare among dicots. Presence of foliar epidermal sclere- ids is likely synapomorphous for these genera. Sec- ond, staminate flowers of Discocarpus, Lachnostylis, and at least two species of Amanoa, A. nanayensis W. (Hayden, 1990), share an androgynophore-like structure (sometimes described as filaments connate to the . Hayden and A. steyermarkii Jabl. "mi pistillode); this feature, too, is likely synapomor- phous for these genera (Webster, 19942). In addi- tion, while several genera of Wielandieae possess scalariform perforation plates in the wood, Amanoa and Discocarpus (as well as Lachnostylis) share the derived state of simple perforation plates (Mennega, 1987; Hayden et al., 1993) tures, Mennega (1987) argued for the exclusion of Based on wood fea- Discocarpus from Wielandieae. Superficially, inflorescences of Discocarpus and Amanoa appear distinctly different; however, their basic architecture may prove to be homologous. As described and illustrated here, flowers of Discocar- 166 Annals of the Missouri Botanical Garden Foliar Figures 18-20. anator (NY). Areoles between veins in | 2692 (F). —19. Prismatic crystals associated with parencl — 40 um. —20. Parenchyma cells lining areoles; bar — pus occur in axillary clusters. Substructure within these clusters is difficult to discern in the dried, pressed specimens available for study; however, their placement appears to be consistent with the sessile cyme diagrammed for Amanoa by Pax and Hoffmann (1922). Thus, while the cymes of Disco- carpus occur in the axils of foliage leaves and those of neotropical Amanoa are placed in the axils of ny of Discocarpus. —18. Discocarpus gentryi S. M. Hayden, Vásquez & Jaramillo 5487 1 paradermal section; bar = 100 um. 19, 20, Discocarpus essequeboensis Klotzsch, Smith 1yma cells lining the areoles; polarized light micrograph; bar 10 um. reduced bracteal leaves, the differences between these genera are neither great nor absolute in this regard. It should be noted that cymes of some Af- rican Amanoa are axillary to foliage leaves. While Amanoa is largely monoecious, Amanoa anomala Little (Little, 1969) is dioecious, as is Discocarpus. It is also noteworthy that Punt (1962) and Köhler (1965) distinguished the pollen of Discocarpus from Volume 83, Number 2 1996 Hayden & Hayden Revision of Discocarpus that of other Wielandieae. Although pollen differ- ences with Amanoa exist, Punt (1962) included Discocarpus as a distinct type under his “Amanoa configuration.” Thus, flower structure, foliar anatomy, wood, and pollen, all support classification of Discocarpus in tribe Amanoeae as superior to its present placement in Wielandieae. inflorescence architecture, Literature Cited Berlyn, G. P. & J. P. Miksche. 1976. Botanical micro- technique E CUA: The lowa State Univ. Press, A ee Brako, L. _ Zarucchi. 1993. th d E hos Flow- ering B and Gymnosperms of P o de las Angiospermas y Gimnospermas del er Danos Syst. Bot. Missouri Bot. Gard. 45. Croizat, L. 1948. Discocarpus. In: , Plant piii do in Guiana in 1944, 2m to sh Talelberg nd Kaieteur Plateau-IV. Bull. Torrey Bot. C Manco R., P. The genus Vien DE aceae) in South Sos. Bot. Jahrb. Syst. 111: 297- 346. Gaucher, L. phorb A . 8, 15: Gillespie, L. J. 1993. Eupborbiao eae of the Cua: An- notated species checklist and key to the genera. Brit- tonia 45: 56-94. Hayden, S. M. 1995. A Taxonomic Revision of Neotrop- ical Discocarpus (Euphorbiaceae). Master’s Thesis, Uni- versity of Richmond, Richmond, Virginia. Hayden, W. es on = ‘al Amanoa (Eu- phorbiac cag) Brittonia 42: 260— , M. P. Simmons & L. J. um 1993. m of Amanoa (Euphorbiac eae). [AWA Journal 14: 205-213. Hutchinson, J. 1902. Recherches anatomiques sur les Eu- 1969. Tribalism in the family Euphorbi- 8-758 7. Discocarpus. P. 85i 1: B. Maguire & Collaborators, The Botany of the ua Highland. Part VII. Mem. New York Bot. Gard. 17. Johansen, D. A. Hill, New York 1940. Plant Microtechnique. McGraw- Klotzsch, J. F. 1841. Neue und weniger gekannte suda- merikanische Euphorbiaceen-Gattungen. Arch. Syst. Naturgesch. 7(1): 175-204 l Discocarpus. In: G. Bentham et ses XIV.— Congibubans towards a flora of South America, —Enu- meration of plants collected by Mr. Sc homburgk i in Brit- ish Guiana. London J. Bot. 2: 52. Kahler, E. 1965. Die Pollenmorphologie der biovulaten Euphorbiaceae pe ihre Bedeutung fiir die Taxonomie. bs Palynol. 6: G. A. 1986. n foliar morphology of Phyl- "lilies us (Euphorbiaceae). I. Cons Mi souri Bot. Gard. 73: 29-85. Little, E. L. 1969. New tree species from Esmeraldas, Ecuador. * dee 18: 404418. MacBride, J. F. 1951. Euphorbiaceae. /n: Flora of Peru. Field Mus. Nat. B Bot. Ser. 13 (part 3A), no. 1: 3- JO. pectus. Ann. Mennega, A. M. W. 1987. Wood anatomy of the Euphor- biaceae, in particular of the subfamily Phyllanthoide ae. Bot. J. Linn. Soc. 94: 11-126. Müller, J. (Müll. Arg.). 1863. Euphorbiaceae. Vorläufige Mittheilungen aus dem für De Candolle's Prodromus bestimmten Manuscript über diese Familie. Linnaea 32: 26. 873. oe In: Martius, Flora Bras- liens 11(2): 1 y E. v RA 1922. pio dip uit Phyllan- aula Phyllanthese-Discocarpina anzen- reich IV. 147. XV (Heft 81): 202-205. . Euphorbiaceae. /n: A. Engler & K. Pid. Die natürlichen Vini or e Ed. 2. 19c: 11-253. Wilhelm Engelmann, | Punt, W. 1962. Pollen morphology of la Euphorbiaceae with special n ce to taxonomy. Wen : 1-116. Webster, G. 75. Conspectus of a new y classification of the E poa eae. Taxon 24: 593 ———. 4a Classification. of jan ias Ann. ed Bot. Gard. 81: 3-32. 1994b. Synopsis of the genera and um taxa of bio eae. Ann. Missouri Bot. Gard. 33-144. A REVISION OF THE FERN GENUS PHANEROPHLEBIA (DRYOPTERIDACEAE)! George Yatskievych? ABSTRACT The p and de riptions provided. Classification of the g stelar anatomy, and telas: of i species. of P. nobilis is discuss atly named rimarily neotropical fern genus Phanerophlebia is d ie taxonomic "d to include ries ig ies, with a enus within the Dryop ustification for nah ‘tion = a Um: diploid species : contrasted key the morphology o a variety ed with its tetraploid derivative, eridaceae 1s reviewe gc iston yl, P. juglandifolia, and kakaea and polyploidy within the aipa are a Phanerophlebia C. Presl is an enigmatic genus comprising eight species of terrestrial ferns of the World tropics and adjacent subtropical regions. Most pteridologists know little about the group because of the rarity with which the species are encountered in nature. The genus was revised taxonomically by Underwood (1899) and Maxon (1912). Both of these early studies were based on relatively few herbarium specimens and did not ad- equately circumscribe variation in critical morpho- logical characters, resulting in many subsequent misdeterminations. need for a new taxonomic study of this genus was further supported by disagreements and ques- tions raised by the few researchers who have ex- amined the species in recent years. Smith n Mickel and Beitel (1988), and Stolze (1981), i their detailed floristic treatments of tee of Chiapas and Oaxaca, Mexico, and Guatemala, respectively, each pointed out taxonomic problems in the group and emphasized different characters in delineating taxa. Tryon and Tryon (1982), in their review of American ferns, pointed out the confusing pattern of morphological variation in the genus and suggested that some of the traditionally accepted species might be combined, following much-need- ed, critical, future studies. Aside from questions of species-level taxonomy, Phanerophlebia has also provided problems in ge- neric classification. The distinctness of the group from Asiatic Cyrtomium and nearly cosmopolitan Polystichum has remained controversial and is symptomatic of the general difficulties of generic circumscription present in the family Dryopterida- ceae (and in several other pteridophyte families as well). The dinis popr incorporates data from several concurrent studies (Stein et al., 1989; Yatskievych, 1989, 1990, a 1993; Yatskievych & Gastony, 1987; Yatskievych et al., 1988) on aspects of evo- lution and systematics of Phanerophlebia and the results from field, herbarium, and greenhouse stud- ies on the group, in an attempt to refine our un- derstanding of the genus. An evolutionary species concept is employed, which circumscribes a set of morphologically defined taxa that are also geneti- cally separated. Genetic separation, based on var- ious estimates of genetic similarity, rather than crossing barriers, has been used, because attempts at artificial hybridization between taxa failed uni- formly (unpublished data). The varietal designation is used in this paper to denote putatively interfertile morphotypes within a species, which sharp biogeographic discontinuities, and which are sep- ' | am indebted to Gerald Gastony and other members of the faculty at Indiana University for their support and guidance during my ¢ doc John Mickel, Michael Windham. Gerald a d lea Pi n A study ne Myra from the advice, data, and reviews of she id Barrington, ryon, David Wagner, Herb and Florence Wagner, anc ael Grayum, Boone Hallberg, Richard Hevly, Tom Ranker, Greg and Carol Starr, Eckhard Wollenweber, and Ric ae Worthington are among those patient souls who put up with my idiosyncratic field behavior. The curators of the herbaria cited generously provided loans of pertinent specimens. am grateful to Big Bend National Park, Texas Parks and Wildlife Department, p various agencies of the Mexican government for generously approving collecting permits for this project, and to S aum Koch, Blanca Perez-García, and Ramón Riba for thei ir extraordinary help in obtaining permits in Mexico. This s search was supported by grants died the Indiana Univers mal Science Foundation (BSR 8 det by Sigma X the love, support, and habes nt of my wife -12660), « du epi of Biology, Indiana University Graduate Sc ‘hool, Indiana Academy of Science, and Nation: , the Se cute Research Society). Finally, this study could not have been possible without Kay, who became part of my life during its completion. A. al Academy of Science (through the Joseph jin fund, ? Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166, U.S ANN. Missouni Bor. GARD. 83: 168-199. 1996. Volume 83, Number 2 1996 Yatskievych y 169 Revision of Phanerophlebia arated by only a single, relatively trivial, morpho- logical character (whose genetic basis remains un- known). METHODS AND MATERIALS Herbarium specimens, including types, were studied from the following. collections: A, ARIZ ASC, BH, CAS, CM, CR, CU, DS, , F, GH, HAL, IND, K, LL, MEXU, MICH, MO, NY, P, viations follow those used in Holmgren et al., ). Approximately 1400 specimens representing about 670 different collections of Phanerophlebia were examined. Additionally, about 260 collections of Cyrtomium and Polystichum were examined for comparative purposes Information from herbarium specimens was sup- plemented with observations of populations in na- ture. A total of 25 populations of Phanerophlebia was located during fieldwork in the southwestern United States, Mexico, and Costa Rica. In addition to field observations and carefully pressed speci- mens, representative plants from each population were transplanted into common culture in the De- partment of Biology greenhouses at Indiana Uni- versity for further observation of phenotypic varia- tion. For detailed observations of scale morphology, small strips of clear tape (Scotch) were used to re- move samples from leaves and rhizomes with the aid of forceps and mounted on microscope slides in Hoyer's medium. Stelar observations were based on direct observation of freehand sections from liv- ing rhizomes. Details of venation patterns and epi- dermal morphology were observed from pinnae or pinna pieces cleared using the bleaching proce- dures summarized by Dilcher (1974). Following de- hydration and staining with safranin O, samples for observation of leaf venation were dried flat in a small press similar to that described by Wagner (1976), and epidermal samples were mounted on microscope slides in Permount, prior to viewing. Chromosome counts were made on representa- tives of each population. Young pinnae with devel- oping sporangia were fixed in Farmer's solution (3: 1 absolute ethanol: glacial acetic acid) for meiotic counts. Sporangia were spread using the standard acetocarmine squash technique (Manton, 1950), as modified by Haufler et al. 5). Cells were ex- amined with a compound microscope and chromo- somes were drawn using a camera lucida. New counts or those differing from the literature were also photographed, using Kodak Technical Pan film. For spore counts, single, unopened sporangia were carefully removed from field-collected (rather than greenhouse-grown), dried pinnae with a dis- secting probe and opened in a drop of lactic acid under the dissecting microscope. Ten counts from different sori were compiled for representative plants from each population. Spore sizes were also measured from material in lactic acid, using a com- pound microscope equipped with an ocular mi- crometer. Thirty measurements of the longest di- mensions of spores from at least five sporangia were made, from which the mean and standard deviation were calculated for each sample. Spore morphology was observed under the compound microscope from material in lactic acid and from dried material, as well as by scanning electron microscopy (SE The latter observations were carried out using a Cambridge Stereoscan 250-MK2 machine. Dry spores were mounted on aluminum SEM stubs us- ing double-sided tape without pretreatment and sputter-coated with gold/palladium using a Polaron E5100 unit prior to viewing. MORPHOLOGY AND ANATOMY RHIZOMES AND LEAVES Rhizomes of Phanerophlebia species are dictyo- stelic and anatomically similar to those of Cyrtom- ium species studied by Chandra and Nayar (1982) and Gibson et al. (1984). They range from 3 mm in diameter in small plants of P. pumila to ca. 20 mm in P macrosora, but are 15 mm in adult plants of most species. Rhizomes may be charac- terized as generally short-repent (erect or ascending in P. macrosora and P. pumila), unbranched (tend- ing to branch at maturity in P. auriculata and P. d and deep-seated in the substrate (ex- ept in P. gastonyi, P. pumila, and sometimes in P. © juglandifolia). In all species, the rhizomes are ensely covered with scales near the apices and with persistent petiolar bases and adventitious roots on older portions. The dense mat of petioles and roots makes the rhizomes appear larger than they actually are. A scales are discussed below, under indum Leaves ena develop equally from all sides of the rhizomes and are clustered near the rhizome apices. Mature leaves range in length from 4 cm in P. pumila to 270 cm in P. macrosora. They persist for two or more years and eventually die back to the petiolar bases, which are persistent. There is no well-defined zone of abscission along the peti- oles. The petiolar bases persist for several years, turning hard and “woody” in P. macrosora, but re- maining green and semisucculent, with green in- 170 Annals of th Missouri Botanical Garden ternal tissues, in other species. Although an iodine test was not performed, these persistent petiolar bases presumably function as trophopods (Wagner & Johnson, 1983 tiole vascularization is simpler than in many other members of the Dryopteridaceae, and the 4-8 vascular bundles (mean = 6) are arranged in an uneven ring. The petioles are curved abruptly at their bases, so as to orient the leaves vertically, except in P. macrosora. In this species the petiolar bases are usually oriented horizontally for a dis- tance of 5-10 cm, before curving upward, the le: bases thus appearing geniculate. Petioles may be shorter than or somewhat longer than the laminae, and are scaly (see below) in all species. Laminae are monomorphic and 1-pinnate (some- times simple in P. gastonyi and P. pumila), with alternate pinnae (leaflets). Rachises are adaxially sulcate, with the groove more or less confluent with the costal groove of each pinna, although, in prac- tice, this character is usually difficult to assess. The laminae have discrete terminal pinnae, but are not typically imparipinnate in that some leaves of all species have an even total number of pinnae. Such leaves are usually mixed on the same plant with those having odd numbers of pinnae, and there is no morphological evidence for the existence of “pseudoterminal” pinnae; i.e., no nodes or articu- lations are present at the tip of a rachis to suggest suppression of the laminar apex and its replace- ment by an adjacent, lateral pinna. Pinnae of all species are chartaceous to subcor- iaceous and have shallow, adaxial, costal grooves. They range in shape from ovate to linear-lanceolate and commonly have attenuate apices and asym- metrical bases. Pinna bases are normally more de- veloped acroscopically (Fig. 1), and acroscopic, basal auricles occur commonly in Phanerophlebia auriculata, sporadically in P pumila, and rarely in nobilis (var. nobilis). Pinna margins are spinu- lose-serrulate, the teeth connected by a narrow, white, marginal band of hard tissue, but the density and distribution of the teeth vary. In P. gastonyi, the serrulations are confined to the distal half of each pinna, and in P. juglandifolia the density of serrulations in proximal portions of the pinnae can vary greatly on individual leaves. In P. macrosora, serrulations tend to be somewhat more widely spaced than in other species and the teeth are often somewhat stouter. Rare individuals of most species (particularly Phanerophlebia auriculata and P. pumila) have some leaves with the pinnae deeply incised to lac- erate, particularly on the acroscopic side (Fig. 1b). The same phenomenon was described by Wagner = L] 0] * e. “>... .. Cee posos ..? ePi Figure 1. —Aa. P. auriculata (Yatskievych et « ple, dichotomous venation and sence of basal auricle. auriculata a et al. 8), showing irregular acroscopic lobing. —c. jenem ata (Yatskievych & McCrary 85-05), showing simple (non- hae eie b d venation. —d. P. gastonyi 182), showing anastomosing vena- Cleared pinnae of oe species. id —10), showing sim- acrosc na 1 ——D, e Yatskievych et « oe Scale bar — (1979) for Polystichum munitum, and a similar phe- nomenon in Polystichum acrostichoides (f. incisum (A. Gray) Gilbert) was attributed by Wagner et al. (1970) to environmental and temporal factors. In Phanerophlebia, it is of interest to note that in such pinnae the lobes each contain one vein that has branched from the costa and that the multiseriate (with respect to the costa) sori are arranged in a single series on either side of this midvein in each lobe. Such anomalous pinnae provide indirect evi- dence that the 1-pinnate leaf dissection in Phaner- ophlebia is derived from a more highly dissected leaf type. Venation in Phanerophlebia species is anadro- Volume 83, Number 2 1996 Yatskiev ych 171 Revision of Phanerophlebia mous (this often difficult to observe) and consists of numerous, arcuate, secondary veins branching sequentially from the costa of each pinna (Fig. 1). ese secondaries branch 1-4 times before termi- nating near the pinna margin. There is no vein run- ning parallel to the margins, but there is a thick- ened, narrow band of hard, white tissue along the edge, at which the veins terminate. Thus, the veins do not extend into the marginal serrulations of the pinnae. The basal 1-2 vein branches do not extend to the margins, instead terminating between the costa and margin. Vein endings are tapered (except those terminating in sori), lacking bulbous thick- enings. The venation pattern of Phanerophlebia is of two kinds. The most common type, which characterizes all taxa except P. gastonyi, P. juglandifolia, and P. nobilis var. remotispora, consists of simple, dichot- omously branched secondaries essentially lacking anastomoses (Fig. la—c). All of the free-veined taxa can produce rare, marginal anastomoses, and in P. haitiensis these can be fairly common (>2 per pin- na, but very irregularly distributed). The ultimate vein-branches in the free-veined taxa range from arcuate to nearly straight, but are usually more or less parallel to one another. In Phanerophlebia gastonyi, P. juglandifolia, and P. nobilis var. remotispora, the branching secondary veins form a regular pattern of 1-3 series of simple, submarginal anastomoses (Fig. 1d). The pattern of reticulation is that of elongate areolae with acute apices and bases. There are no free, tertiary veins included in the reticulations (as in Cyrtomium), but in most cases the basal branches of the secondaries end below the anastomoses of adjacent, more mar- ginal branches, thus appearing enclosed. Although there are no clear morphological distinctions be- tween the patterns of reticulation found in the closely related P. gastonyi and P. juglandifolia on the one hand and those of P. nobilis var. remotispora on the other, evidence from restriction site poly- morphisms in the chloroplast genomes (Yatskievych et al., 1988) and the fact that only one of the two varieties of P. nobilis exhibits reticulate venation suggest that this pattern has arisen independently twice in the genus Sori of Phanerophlebia species are either termi- nal or lateral on the secondary veins and are situ- ated in 2-4 series (sometimes only l-seriate in P pumila) between the costa and margin. In P. um- bonata, the sori are sometimes clustered submar- ginally, but otherwise they occur in a broad band between the costa and each margin. In P. gastonyt and P. juglandifolia (occasionally also in other spe- cies) the innermost series of sori may be situated less than 2 mm from the costa, but the innermost sori are otherwise normally more than 4 mm from the costa. This character has been used by some authors to differentiate P. juglandifolia from P. no- bilis var. remotispora, but is too variable for consis- tent application. The sori are more or less circular and in all but two taxa are covered by an indusium. Presl (1836) first noted the absence of an indusium in Phaner- ophlebia juglandifolia (which he used as evidence to align this species with Polypodium under the generic name Amblia), but nearly all subsequent workers ignored or dismissed this character until it was reexamined by Smith (1981). Phanerophlebia gastonyi, which has been confused with P. juglan- difolia in the past, is the other species lacking in- dusia. Part of the problem with regard to this char- acter is that in all of the indusiate species except P. umbonata the indusia are fugacious, shriveling and mostly falling off by the time the sorus is ma- ture. Thorough examination of sori in these taxa will always result in the discovery of at least a few remaining, shriveled indusia, however. In those species possessing indusia, these ephemeral structures are light tan, peltate, roughly circular in outline, erose-margined, and 0.6-1.1 papery, concolorous, and flat to slightly concave centrally, mm in diameter. They are membranous to except in Phanerophlebia umbonata, which has firmer, subcoriaceous, persistent indusia with a raised, darkened umbo in the center. Indusia of P. pumila are unusual in that the attachment point of the receptacle is sometimes acentric, and the in- dusia are often more ovate than circular in outline. It is also of interest that although diploid P gas- tonyi and its tetraploid derivative, P. juglandifolia, lack indusia, the rare, sterile, triploid hybrid be- tween the latter taxon and the indusiate P. macro- sora has rudimentary, irregular indusia less than 0.2 mm in diameter. SPORES Spores of selected Phanerophlebia species were examined cursorily (usually in conjunction with those of Cyrtomium) by several earlier authors, but the first detailed study of spore morphology in the genus appe eared as part of the generic treatment for Cyrtomium (including Phanerophlebia) in Tryon and Tryon’s (1982) broad survey of fern genera. The subsequent discussion in Tryon and Lugardon's (1991) survey of fern spores expanded slightly on this initial treatment. The present account attempts to expand upon this excellent foundation. Sporangia contain 64 spores in all fertile taxa 172 Annals of the Missouri Botanical Garden and are monolete and dark brown at maturity. Spores fall into three size classes. Those of the dip- loid Phanerophlebia gastonyi are 30—42 um in lon- gest dimension, whereas those of other fertile dip- loid and tetraploid taxa in the genus are 41—60 um. Spores of P. haitiensis, although seemingly well formed, show great size variation, even within in- dividual sporangia, and measure 36-52 um. Spore size is thus of little use in identifying polyploids, except as a means of distinguishing P gastonyi from its closest relatives. The exospore in all species is smooth (Fig. 2) and is covered by a two-layered perispore. The thin, inner perispore has a slightly undulate sur- face, often also with widely scattered papillae. These papillae may represent the rudiments of col- umellae that once connected the inner and outer perisporal layers, but actual columellae apparently no longer exist. The outer layer of perispore is much thicker than the inner layer and consists of inflated, irregularly undulate folds. The density and degree of inflation of these folds varies develop- mentally and within each species. They are thus of little use in distinguishing among them, although the mature spores of P. macrosora generally have fewer folds than those of other taxa. The surface of the outer perisporal layer also varies from nearly smooth to rugulose and contains no observable mi- croperforations. 11 fertile specimens of Phanerophlebia species possess at least some sporangia that, at maturity, contain malformed, apparently abortive spores of variable size with the outer perisporal layer rela- tively smooth and the folds winglike and uninflated (Fig. 2d). These resemble immature spores in their incomplete perisporal deposition and although not obviously collapsed or shrunken within, they will not germinate on agar or soil. Sporangia containing such spores are especially frequent in the relatively few herbarium specimens available of P. haitiensis. Whether this phenomenon has a genetic basis is not presently known, but plants placed under en- vironmental stress (such as excess sunlight or high temperature) in the greenhouse produce such abor- tive spores with greater frequency than when grown under more optimal conditions. INDUMENT Leaves of Phanerophlebia species do not possess true hairs in the sense that this term is usually applied to such epidermal outgrowths in most ferns. Instead, the uniseriate, multicellular trichomes present in all species are in reality reduced scales, and intergrade completely with the larger, pluriseri- ate structures usually associated with this term ig. 3). Such reduced scales are common in the Dryopteridaceae and have been noted by several students of the family. Daigobo (1972) believed that variations in the pattern of cellular orientation and shape made such trichomes the most stable and important characters for infrageneric classification of the 47 species of Polystichum in Japan, Ryukyu, and Taiwan, and applied the name “microscale” to them. Viane (1986), who held trichome morphology to be important in classification of Dryopteris spe- cies, adopted the name “paleaster” for such hair/ scale intermediates. Moran (1986, 1987), working with the related genera Olfersia and Polybotrya, coined the term “proscale” for them. Smith (1986), in his monograph of Cyclodium, differentiated re- duced scales from acicular hairs and septate glands also found in that genus but declined to invent a name for these structures. Barrington (1989), in his studies of neotropical Polystichum, adopted the term “microscale” and noted that most species of neotropical Polystichum shared Daigobo's (1972) Metapolystichum type of scale development. The pattern of scale development in all species of Phanerophlebia is identical and is similar to that seen in the Metapolystichum type of Daigobo (1972). The reduced scales found to some degree on the abaxial laminar and distal costular surfaces are tan to brownish and appressed. The smallest are uniseriate and approximately 3 cells long, with somewhat flattened cells and thickened endwalls. In contrast to the situation described by Daigobo (1972) for Metapolystichum, however, the terminal cells of these smallest scales are obtuse, rather than acutely attenuate (Fig. 3). Larger scales with two or more cellular series display the morphology more typical of Metapolystichum scales. In these larger, reduced scales the basal 1-2 series of cells are smaller than the other cells and are nearly square to rhomboidal in outline. Above this base, the body of the scale is broadened and the individual cells are elongate longitudinally. Processes homologous to the lateral cilia of fully developed costal and petiolar scales are first evident near the base of the reduced scales, which are otherwise entire along their lateral margins. Along the costae, the reduced scales grade abruptly into the longer and broader structures usu- ally associated with the term scale. Reduced scales are best observed on immature leaves and are ap- parently shed during development or at least easily abraded from mature laminae. The pinnae of some populations of P. auriculata and P. pumila possess the densest covering of reduced scales, but this re- tention of vestiture may be due primarily to the Volume 83, Number 2 Yatskievych 1996 Revision of Phanerophlebia Figure 2. Spores of d specie P. umbonata (Yatskievych & Wollenweber 83-87), a mixture of mature iet immature spores. —b. P. auric iun buie et al. 83-10). —c. P. gastonyi (Yatskievych et al. 85-182). —d. P. macrosora (Yatskievych & dU way 5-30), with incomplete perispore Domain characteristic of nonfunctional spores binoja in some sporangia of all species. Scale bars = H Annals of the Missouri Botanical Garden from a single Figure 3. Representative laminar scales leaf of Phanerophlebia auriculata (Yatskievych et al. 83— 0). protected habitats in which these plants usually grow. The most highly developed scales occur along the proximal portions of the petioles and on the rhizomes. The rhizome and largest petiolar scales are similar in morphology within most species, but the rhizome scales are usually less variable and often slightly smaller and more densely ciliate along their margins than those of adjacent petiole bases. The largest rhizome and petiolar scales range in length from 2.5—4.0 mm in P. gastonyi to 10-15 mm in P. macrosora. Rhizome scales are ovate to elliptic-lanceolate, but petiolar scales vary from linear to ovate in most species, due to the gradation of mixed scale types described above. Exceptions are P. gastonyi, which has very few nar- rower scales mixed with the sparse, ovate scales along the petiolar bases, and P. pumila, which has only linear-filiform petiolar scales except in the very basal portions. The margins of the rhizome and petiolar scales are ciliate, and those of Phanerophlebia macrosora are strongly erose-ciliate. In all taxa, there is a ten- dency for these cilia to break off with age. In P. nobilis, cilia tend to be abraded easily, and a dili- gent search with a dissecting microscope is nec- essary to observe the few remaining on scales at rhizome or leaf maturity. In contrast, the margi of petiolar scales in P. gastonyi, P. juglandifolia, and P. macrosora tend to remain relatively unal- tered at maturity, although the cilia become abrad- ed with age. The petiole bases of young leaves in all species are covered with a dense indument of imbricate scales except in P. gastonyi, wherein the scales are sparser and essentially non-overlapping. In P. macrosora, the dense, chaffy covering of scales is especially persistent, even with age. Rhizome and petiolar scales are concolorous, ex- cept in Phanerophlebia gastonyi and P. juglandi- folia. The former species has scales with broad, dark centers of sclerified cells and narrow, hyaline margins. In the latter species the darker, central band is of variable width, though well defined from the correspondingly broader, hyaline margins. Specimens of P. macrosora rarely have scales with slightly darker, poorly differentiated, central dire Otherwise scales of this species are light The remaining taxa produce tan to orange- bon scales; those of P. nobilis and P. umbonata are usually a pronounced orange-brown in color. HYBRIDIZATION AND POLYPLOIDY Chromosome counts for Phanerophlebia species are presented in Table 1. These i indicate that the base number for the genus is x — 41, a common number in the Dryopteridaceae. chromosome counts had been published previously for most species and appeared to indicate a rela- tively simple cytological situation in the group, with less polyploidy and hybridization than is typical in Polystichum and other related genera. The earliest reports for the genus were diploid counts for P. um- bonata (Wagner, 1963; Mickel et al., 1966). Diploid counts had also been published for P. ekina oa P. macrosora, P. nobilis, and P. remotispora (Smit & Mickel, . Two tetraploid taxa had been identified, P. auriculata (Reeves, 1978) and P. pumila (Smith & Mickel, 1977), and except for a single tetraploid count for a sample tentatively as- cribed to the otherwise diploid P nobilis var. re- motispora (Smith & Mickel, 1977, reported as P. cf. remotispora), there was no evidence of cytological heterogeneity in any of the species. The only pres- ently accepted species for which no chromosome count exists is P. haitiensis, which is known histor- ically from very few collections and is presumed Representative S extinct at its type locality in Haiti. New counts reported in Table 1 agree with most earlier counts for the genus, but document the sur- prising find that Phanerophlebia juglandifolia, as traditionally circumscribed, comprises two cyto- types (Yatskievych & Gastony, 1987). The previ- ously published count from Oaxaca, Mexico (Smith Volume 83, Number 2 Yatskievych 175 Revision of Phanerophlebia Table 1. Chromosome numbers in Phanerophlebia. All counts are from meiotic material. Vouchers for new counts are accessioned at IND and MO, with duplicates to be distributed elsewhere. Number of erotic Species Bivalents Source P. auriculata 82 U.S.A. Arizona: Chochise Co., Bass Canyon, Galiuro Mtn., 3 Jan. 1983, Yatskievych et al. 83-10; Garden Canyon, Huachuca Mtn., 17 May 1983, Yatskievych 83-161; Santa Cruz Co., Sycamore Canyon, Puhia Min., 30 Dec. 1982, bar onde & Yatskievych 82-273; Yavapai Co. (Reeves, 1978). New Mex jew Ana Co., Ice Canyon, bean Mtin., 21 July 1984, Yat- unn. et al 84-6 P. gastonyi 4] MEXICO. Chiapas: ca. 13 km N of Berriozabal, 17 July 1985, Yatskievych et al. a 182. Oaxaca: (Smith & Mickel, 1977, as P. juglandifolia). ha racruz: | km N of Pas de Enriquez on rd. from Jalapa to Misantla, Sep. 1986, Yatskievych & Gastony 86-337. P. juglandifolia 82 MEXICO. Chiapas: (Smith & Mickel, 1977, as P. cf. remotispora). COSTA RICA. San José: 6 km S of Hwy. 2 on Hwy. 12 to Santa María de Dota, 15 Mar. 1986, Yatskievych & McCrary 86-13. Heredia: 2 km SE of Sacra- mento on Hwy. 114, S slope of Volcán Barva, 18 Mar. 1986, Yatskievych & McCrary 86-31. P. juglandifolia 1231 COSTA RICA. Heredia: 2 km SE of Sacramento on Hwy. 114, S slope of X macrosora Volcán Barva, 18 Mar. 1986, Yatskievych & McCrary 86-3 la. P. macrosora 41 COSTA RICA. Heredia: 2 km SE of Sacramento on Hwy. 114, S slope of Volcán Barva, 18 Mar. 1986, Yatskiev a & McCrary 86-30. MEXICO. Oaxaca: trail to Llano Verde, 1-2 mi. NE of Natividad, 30 Dec. 1983, Yatskievych et al. 83-467; 22 mi. NE of Teotitlan del Camino on rd. to Huautla, 28 Sep. 1986, Yatskievych & Gastony 86-329; (Smith & Mickel, 1977) P. nobilis var. nobilis 4l MEXICO. México: 1-2 km E of San Rafael on W slope of Volcán Iztacci- huatl, 21 July 1985, Yatskievych et al. 85-211. Oaxaca: 35 km S of Tlax- iaco on Hwy. 125 to Putla, 27 Sep. 1986, Yatskievych & Cuin 86-327; (Smith & Mickel, 1977, as P. D var. remotispora 4l MEXICO. Chiapas: ca. 13 km N of Berriozabal, 17 July 1985, Yatskievych et al. 85-186; (Smith & Mickel, 1977, as P. remotispora). Hidalgo: 17 mi. SW of Chapulhuacan on Hwy. 85, 4 May 1983, S nde & Wollenweber 83-128; 32 km SW of pos eR on Hwy. 8 o Jac :ala, 22 Dec. 1983, Yatskievych et al. 83-353. Veracruz: ca. 3 km S of Orizaba on Hw Hur Sierra de San Cristobal, P May 1983, o & Wolle ae 3-158. P. pumila 82 En Chiapas: (Smith & Mickel, 1977). Oaxaca: Llano de Las Flores, at high point on Hwy. 175 N of Ixtlán de Juarez, 13 July 1985, Yatskie- vych et al. 85-139; NE of Ixtlán de Juarez on old rd. from Capulalpan de Mendez, near junction with newer road to Francisco I. Madero, 19 July 1985, Yatskievych & Gonzalez L. 85-209; (Smith & Mickel, 1977 P. umbonata 4.1 MEXICO. Nuevo León: SW of Monterrey on rd. to Valle de San Angel, lower slopes of Mesa de Chipinque, 1 May 1983, Yatskievych & Wollenwe- ber 83-87; km post #29 on Hwy. 58 from Linares to San Roberto, 19 Dec. 1983, Vasskieoych et al. 83-299; Cola de Caballo, W of El Cercado, ca. 20 i. S of Monterrey, 4 Jan. 1984, Yatskievych et al. 84-04; Cañon de San udi Sierra Madre Oriental, ca. 35 km SE of Monterrey, 16 Sep. 1986, Yatskievych & Gastony 86-250. San Luis Potosí: (Mickel et al., 1966). U.S.A. Texas: Bre r Co., Maple Canyon, Basin, Chisos Mtn., 12 Mar. 1985, Yatskievych ys McCrary 85-05. 176 Annals of the Missouri Botanical Garden Figure 4. Meiotic preparations illustrating new chromosome counts in Phanerophlebia. a, b, photographs; a’, b’ ings. —a. P. juglandifolia, 2n interpretive camera lucida drawing mos ` 3 LAE g Vet 4 ^ S = " Fá de b' EI = 82 Il (Yatskievych & McCrary 86-13). —b. P. juglan- difolia X macrosora, 2n = 123 1 (Yatskievych & McCrary 86-31a). Scale bars = 10 um. & Mickel, 1977), applies to the rare diploid taxon treated as P. gastonyi in the present work, whereas the tetraploid count attributed to P. cf. remotispora by Smith and Mickel (1977) applies to the taxon (Fig. 4) treated herein as P Juglandifolia sensu stricto. A further novelty was the documentation of sterile triploid hybrids (Fig. 4) in a mixed popula- tion of diploid P. macrosora and tetraploid P. jug- landifolia, the only known primary hybrid in the genus. Thus the cytological situation in this small Volume 83, Number 2 Yatskievych 177 y Revision of Phanerophlebia group is somewhat more complex than was sug- gested by earlier studies. As has been noted above, spore size did not prove to be a reliable indicator of ploidy in Pha- nerophlebia and was useful only in discriminating P. gastonyi from its relatives. However, the vari- ability of spores in the small sample of P. haitiensis and the apparent high number of malformed spores in this taxon suggest that hybridization and/or poly- ploidy may have had a role in its formation. Measurements of epidermal cells were also somewhat equivocal as a means of documenting polyploidy in the genus. Epidermal cells in Pha- nerophlebia species are irregular in size and shape, but are extremely wavy-margined (jigsaw-puzzle- piece shaped) in surface view (Fig. 5). Stomates are restricted to the abaxial surface of the pinnae and are more or less of the polocytic type (sensu Van Cotthem, 1970). Although polyploids, such as P. pumila, generally possess larger epidermal cells on both adaxial and abaxial pinna surfaces, the within- sample variation is so high that quantification of these measurements was useless. Measurements of stomates (technically, guard cell lengths) proved to e a more stable means of distinguishing diploid taxa (mean 47.7 um, range 44.9— m) from polyploids (mean 60.8 um, range 59.6-63.7 jum). Measurements from a cleared pinna of P. haitiensis (Fig. 5) fell within the range of guard cell size in polyploid taxa, a further suggestion that this species is a polyploid of some sort. A brief discussion summarizing what is known about each of the four polyploids in the genus (in- cluding Phanerophlebia haitiensis), along with speculations as to the likely parentage of each, is presented below. The sterile triploid, P. juglandi- olia X macrosora, is included in the discussion of P. juglandifolia. Phanerophlebia auriculata. This tetraploid oc- cupies the northwestern portion of the range of the genus (Fig. 6). Isozyme studies (Yatskievych, 1990; Yatskievych & Gastony, 1987) have indicated that it is probably an allopolyploid, but did not result in any firm conclusion regarding potential progen- itors within the genus. These data are, however, consistent with the speculations that follow. Anal- ysis of chloroplast DNA restriction site. mutations (Yatskievych et al., 1988) suggested that one of its parental species is P. nobilis. Morphologically, P. nobilis var. nobilis and P. auriculata are very sim- ilar, and the rare occurrence of acroscopically au- riculate pinnae in P. nobilis (otherwise unique to P auriculata) reinforces this similarity. The other morphologically similar diploid occurring in the northern portion of the generic range is P. umbon- ata (Fig. 6). These taxa all share identical rhizome and petiolar scale types and overall leaf morphol- ogy. Specimens from some localities where the ranges of these taxa overlap have been misdeter- mined by earlier workers and a few samples, such as those collected by Robert Bye in the Sierra Ma- dre of Chihuahua, Mexico (i.e., Bye 6989, 7094, 7363), were quite difficult to determine during the present research. Phanerophlebia auriculata and P. umbonata share a tendency for rhizomes to branch with age, a feature not seen in P. nobilis. Biogeo- graphically, the derivation of P. auriculata from P. nobilis and P. umbonata also seems reasonable, giv- en that these three taxa are the only ones known to occur in northern Mexico. However, mixed pop- ulations have not been found in nature. Phanerophlebia haitiensis. As noted above, mea- surements of stomata suggest that this rare taxon 1s some sort of polyploid, and its irregular spore size and relatively high number of malformed spores suggest that hybridization is involved. The irregular pattern of anastomosing venation also leads me to suspect that it is of hybrid origin. This taxon is presumed extinct in Haiti (see below, under taxo- nomic treatment) and was unavailable for labora- tory study, so no comparative information is avail- able from cytological, isozymic, or chloroplast genomic studies. What is unusual is that no other species of Phanerophlebia has ever been recorded from Hispaniola, making it difficult to imagine how this taxon might be present at a single, isolated site in Haiti. Morphologically, dep iP haitiensis is reminiscent of P. nobilis. The few specimens that exist are id in venation between free- veined variety nobilis and reticulate-veined variety remotispora, but P. haitiensis has smaller leaves and fewer pinnae than is usual for either variety of P. nobilis. It is possible that the irregular spores and reduced leaves might have resulted from inhospi- table environmental conditions, rather than from genetic causes, but this would not explain the rel- atively large stomates and the irregularly anasto- mosing venation. The existence of two morpholog- ically similar collections from Guatemala (Beaman 3056) and Honduras (Moran 5706) (see under tax- onomic treatment) suggests that further fieldwork in southern Mexico and Central America may even- tually result in at least partial resolution of the enigma surrounding this taxon. Phanerophlebia juglandifolia. Yatskievych and Gastony (1987) discussed the two cytotypes consti- tuting the traditional taxon P. juglandifolia and as- certained that the of the agrees with the tetraploid. Morphological characters specimen name 178 Annals of the Missouri Botanical Garden => 4 a Figure 5. Camera lucida drawings of epidermal cells (surface view) of Phanerophlebia species. a-d, abaxial surfaces, with stomates; a'—d', adaxial surfaces. —a. P. gastonyi (Yatskievych et al. 85-182). —b. P. nobilis var. remotispora (Yatskievych & Wollenweber 83-158). —c. P. pumila (Mickel 5377). —d. P. haitiensis (Ekman 7793). Volume 83, Number 2 1996 Yatskievych y 179 Revision of Phanerophlebia Phanerophlebia 0 auriculata E umbonata Figure 6. that separate tetraploid P. juglandifolia sensu stric- to from the diploid here treated as P. gastonyi are detailed below in the taxonomic treatment. Restric- tion site analysis of the chloroplast genomes (Yat- skievych et al., 1988) of these plants confirmed that the diploid plants formerly attributed to the species were a progenitor of the tetraploid. Isozyme studies (Yatskievych, 1990; Yatskievych & Gastony, 1987) of the tetraploids showed high levels of fixed het- erozygosity not segregating in their gametophytic offspring, which is consistent with the hypothesis that these plants are of allopolyploid origin. The identity of the other parental taxon unfortunately remains unclear, particularly because the isozyme studies did not supply convincing evidence for the identity of the second progenitor taxon. The possi- bilities for the second parent include: (a) P. nobilis var. remotispora, the other net-veined taxon in the genus; (b) a free-veined taxon, such as P. macro- Distributions of Phanerophlebia auriculata and P. umbonata, based upon herbarium specimens examined. sora; or (c) a parental taxon that is now either ex- tinct or has not been sampled. Of these three hy- potheses, the first, that P. nobilis is involved in the parentage of P. juglandifolia, seems the least likely. Morphologically, the tetraploids do not combine the critical features of P. nobilis and P. gastonyi, in- cluding rhizome and petiolar scale shape and color and the pattern of spinulose serrulations along the pinna margins. The third hypothesis, of an extinct or at least as yet unsampled progenitor taxon, may be the most likely, but cannot be evaluated from the present data. The other hypothesis, that a free- veined, indusiate species of Phanerophlebia con- tributed the other parental genome to P. juglandi- folia, was explored by Yatskievych and Gastony (1987). Surprisingly, of the other extant species in the genus, the one that possesses the most likely morphology to account for the differences between P. juglandifolia and P. gastonyi is P. macrosora. 180 Annals of the Missouri Botanical Garden The tetraploid possesses rhizome and petiolar scales with color, shape, and size somewhat inter- mediate between P. gastonyi and P. macrosora, and its pinna size, shape, and number are also plausible for this cross (Yatskievych & Gastony, 1987; see also below, under taxonomic treatment). However, P. juglandifolia bears a far closer superficial resem- blance to P. gastonyi than to P. macrosora, and lacks the pronounced odor, coarsely serrate pinna margins, and geniculate Laos bases that char- acterize the latter species Compounding the obli of establishing the parentage of Phanerophlebia juglandifolia is the existence of rare, naturally occurring, triploid hy- brids between this species and P. macrosora (Table l, Fig. 4) from Costa Rica (see also below, under taxonomic treatment). These hybrids, which pro- duce aborted spores and are therefore sterile, are of intermediate morphology between P. juglandifol- ia and P. macrosora and were relatively common at the single site where I was able to observe them. If tetraploid P. juglandifolia contained one genome from P. gastonyi and the other from P. macrosora, then the Costa Rican hybrids would represent a backcross and should display 41 bivalents and 41 univalents at diakinesis. Instead, the triploids dis- play 123 univalents (Table 1). Unfortunately, this weakens the argument that P. macrosora might have been the second progenitor of P. juglandifolia. The derivation of this tetraploid thus remains unre- solved. Phanerophlebia pumila. This morphologically re- duced tetraploid remains enigmatic. Isozyme stud- ies (Yatskievych, 1990; Yatskievych & Gastony 1987) indicate that it is probably of allopolyploid origin, as are other tetraploids in the genus, but do not assist in determining its parentage. Surprising- ly, its pattern of chloroplast DNA restriction frag- ments (Yatskievych et al., 1988) do not support any of the extant species of Phanerophlebia as a likely progenitor. Thus, the experimental evidence sug- gests that one or even both parental diploids of this taxon are either extinct or so rare as not to have been sampled by collectors thus far. Potential pro- genitors are to be sought in the mountains of south- ern Mexico and adjacent Guatemala, where overall species diversity within the genus is the highest. The Wagner oe i of the molecular data (Yatskievych et al., 1988) grouped Phanero- phlebia pumila and P. sario as sister taxa. It therefore suggested that the latter species might be involved in the parentage of P. pumila. This seems unlikely biogeographically, because the two are widely allopatric. Phanerophlebia pumila displays none of the critical morphological features of P. um- bonata, such as a tendency for branching rhizomes or umbonate indusia. Instead, based upon a qual- itative assessment of morphological characters, it seems at least plausible that P. nobilis might rather have contributed one of the genomes to form P. pumila. Although the leaves of P. pumila tend to be highly reduced in size and pinna number, and the petiolar scales are at most linear-filiform, the pinna shape and the rhizome scales, as well as the overall leaf shape, suggest some depauperate sam- ples of P. nobilis. Further, this diploid can some- times occur in sinkholes and other habitats nor- mally associated with P. pumila, and one set of replicate herbarium specimens examined during this study (Hinton 12428) consisted of a mixed col- lection of these two species. No other species of Phanerophlebia, except the northern tetraploid, P. auriculata, has been found in the specialized hab- itat type (sheltered, overhanging rock walls) to which P. pumila is restricted. PHYLOGENETIC RELATIONSHIPS Most modern systematists agree that questions of phylogeny are best approached through considera- tion of the shared, derived characters (synapomor- phies) present in various subsets of species within a clade. Such analyses require several assumptions to be accepted. First, the group in question should be monophyletic. Second, it is assumed that char- acter states for various features can in some way be polarized; that is, that there is a rational way to distinguish apomorphies (derived character states) from plesiomorphies (primitive character states). Finally, the researcher must assume that he or she can detect homoplasy (parallel or convergent evo- lution), which will tend to confuse interpretation of any analysis. There is little doubt that the species of Phaner- ophlebia are monophyletic. However, the species are so character-poor that although autapomorphies — derived characters unique to one taxon) allow sep- aration of the species from one another, as in the key to species below, there are virtually no synapo- morphies that can be documented. Further, those character states that are shared by two or more taxa, such as the rhizome and petiolar scale type of P. nobilis and P. umbonata (among the diploids), are not easily polarized. Also, it is apparent from study of independent data sets (such as restriction site analysis of the chloroplast genomes) that some characters judged by most earlier workers to be im- portant for classification in the group, particularly the anastomosing venation patterns of P. gastonyi and P nobilis var. remotispora, are homoplasious. Volume 83, Number 2 Yatskievych Revision of Phanerophlebia Thus, no formal cladistic analysis of morphological characters is attempted here. GENERIC CONSIDERATIONS The genus Phanerophlebia was first erected. by sl (1836), based on Aspidium nobile Schltdl. & Cham. Presl (1836) also described Amblia, based on Polypodium juglandifolium Humb. & Bonpl. ex Willd., and Cyrtomium, based on Polypodium fal- catum L. f. As with many of Presl's genera, accep- tance of these innovations was neither immediate nor unanimous, and various authors have treated the groups differently up to present times. Smith (1842) first submerged Amblia (which Presl had allied with Polypodium L.) in Phanerophlebia (which Presl had allied with Aspidium Sw.), while maintaining this ge- nus separate from Cyrtomium, and recognizing the relationship of both to Polystichum Roth. Moore (1857) was the first to unite Phanerophlebia and Cyr- tomium, under the latter name, but retained two sub- groups based on the two genera. Diels (1899) pre- sented the most extreme alternative to Presl’s segregate genera by submerging Amblia, Cyrtomium, dnd Phanerophlebia in his large and heterogeneous version of Polystichum. Since that time, most pteri- dologists have either maintained Phanerophlebia as a separate genus (e.g., Underwood, 1899; Maxon, 1912) related to Polystichum, or have included Pha- nerophlebia in an expanded concept of Cyrtomium (e.g.. Morton, 1957; Tryon & Tryon, 1982). No recent workers have questioned the close re- lationship of Phanerophlebia and Cyrtomium to Po- lystichum. In fact, one New World taxon, Polysti- chum dubium (Karsten) Diels, of Andean South America and adjacent Central America, continues to be placed in T pee by some authors (e.g., Tryon & Stolze, Ithough superficially similar to some udin and Phanerophlebia species, this morphologically variable taxon does not appear to be a geographically disjunct member of either group, and is known to Sion with Polystichum platy- iun (Willd.) C. Presl in Ecuador (Barrington, 1985). Its c esate: ‘ation within Polystichum was be- yond the scope of the present work, and it remains to be studied in greater detail in the future. Two hypotheses have been suggested to account for the presence of two morphologically similar “splinter genera” in the Old and New Worlds. Christensen (1930) regarded Phanerophlebia and Cyrtomium as independent derivatives from differ- ent polystichoid ancestors. By this reasoning, the strong similarities between the two groups would homoplasy. In reflect convergent evolution, i.e., contrast, Copeland (1947) suggested a single, Asi- atic origin for the entire group, with an American component having become established via a trans- oceanic, disjunctional event. He noted that such a major disjunctional pattern exists in a few other fern genera, such as Plagiogyria (Kunze) Mett. (Plagiogyriaceae), and this scenario is therefore plausible. Evidence in support of Christensen's (1930) hy- pothesis comes from several sources. An interesting general observation concerns the distribution. of apogamous taxa in the polystichoid ferns. Several authors have remarked on the exceedingly large number of apomictic taxa that have been described in Cyrtomium (e.g., Shing, 1965) and C. falcatum was the subject of classic studies on the phenom- enon of apogamy in ferns (Manton, 1950). Exami- nation of the scant information available on apog- amy in Asiatic Polystichum (Tsai & Shieh, 1985) indicates that this phenomenon also occurs in scat- tered species within the Asiatic component of the genus. In cont erti of Phanerophlebia reproduce sexually. This is also true of all species of American Polystichum that have been examined. Although hybridization and polyploidy are relatively common throughout Poly- stichum, apomixis is expressed only in the Old rast, as noted above, all fertile taxa World contingent. The morphological innovations said to unite Phanerophlebia and Cyrtomium do not survive close scrutiny. Mitsuta (1977), who studied vena- tion patterns in polystichoid ferns, noted that the areolae in Phanerophlebia were narrower (i.e., elon- gate, with more acute ends) than those in Cyrtom- ium and lacked well differentiated, free, included veinlets as are found in most species of the latter group. Wagner (1979), in a summary of reticulate veins as systematic characters, also noted that the pattern of anastomoses develops differently in the two genera. In Phanerophlebia, anastomoses are of a submarginal ontogenetic which Wagner (1979) termed “discal,” but in Cyrtomium they are “costal” in origin. Patterns of perispore deposition were said by Tryon and Tryon (1982) to be identical in Phaner- ophlebia and Cyrtomium. Superficially, the two groups appear similar, but in Phanerophlebia spe- type, cies the surface between the inflated folds is smooth to rugulose and imperforate. However. in the seven species representing Cyrtomium that I studied, the outer perispore surfaces were not only rugulose, but also had few to many microperforations (Fig. 7 Other evidence for the separate origins of Pha- nerophlebia and Cyrtomium comes from a phylo- genetic analysis of restriction site mutations in the chloroplast genomes of the two genera (Yatskievych ^l 182 Annals of the Missouri Botanical Garden Y a. C. ma- Figure 7. Spores of Cyrtomium species. crophyllum (Mitsuta s.n.), with the very few mic roped ations between the folds not easily seen. —b. C. falcatum (Yatskievych & McCrary 83-184), e microperforations more easily seen. Scale bars = 1( et al., 1988). In that study, the groups of Phaner- ophlebia and Cyrtomium species were less related to one another than either was to a group of taxa representing Polystichum. Ithough present evidence supports the hypoth- esis that Phanerophlebia and Cyrtomium were in- dependently derived from different ancestral groups, the real question that remains unanswered is whether either of these derivatives is distinct enough to merit recognition as a genus separate from Polystichum (Wagner, 1985). Yatskievych (1989) suggested that simplification of leaf dissec- tion is a trend that has occurred repeatedly in dif- ferent species groups throughout Polystichum. Taken singly, all of the characters said to separate Phanerophlebia or Cyrtomium from Polystichum also occur in other species of that genus. The venation patterns in two other Asiatic seg- regates of Polystichum, Cyrtogonellum Ching and Cyrtomidictyum Ching, actually resemble those of Phanerophlebia far more closely than do those of Cyrtomium. However, Ching (1938, 1957) clearly differentiated these segregate genera from Phaner- ophlebia, Cyrtomium, and Polystichum, based on other morphological innovations they possess. Spe- cies in both groups also have been submerged in Polystichum by some subsequent authors (e.g., Kra- mer et al., 1990) Perispore pattern was cited by Tryon and Tryon (1982) as a stable character for separation of Cyr- tomium (including Phanerophlebia) from Polysti- chum. These authors felt that the simpler, imper- forate perispore lacking columellae of Cyrtomium was more like that found in Dryopteris than the per- forate or reticulate, often columellate formation characteristic of Polystichum. Although the peri- spores found in Cyrtomium, Phanerophlebia, and Dryopteris are remarkably similar, two factors argue against placing too much emphasis on this inter- pretation. First, as noted above, at least some spe- cies of Cyrtomium possess a perispore with micro- perforations (Fig. 7). Second, Polystichum contains species with a broad range of perispore types. Mitui (1973) attempted to classify the perispore types found in 40 species of Polystichum. He differenti- ated eight perispore types, based on differences in patterns of ornamentation. Most of the species stud- ied had reticulate or perforate perispores of varying complexity, but a small group possessed what Mitui felt to be a derivative perispore type closely related to that of Cyrtomium (Figs. 7, Polystichum species with imperforate perispores are apparently present in both the Old World and New World, although few American taxa have been examined in detail. For example, the perispore type found in Polystichum lemmonii Underw , of the western United States, is quite ala to that of Phanerophlebia species in its relatively compact, imperforate outer layer. As noted above, the widely scattered papillae on the inner perisporal layer of Phanerophlebia species also may represent reduced remnants of earlier columellae. Further studies in- volving a large number of tropical American spe- cies of Polystichum are needed to resolve the pres- ent ambiguities in our knowledge of perispore evolution in the gro t is probable dh Phaneraphdebin and Cyrtom- tum will eventually be resubmerged in Polystichum. In a phylogenetic sense, recognition of these and other splinter genera results in a paraphyletic Poly- stichum, a situation clearly to be avoided. The de- cision to continue treating Phanerophlebia as a sep- arate genus in the present work was a difficult one and stems from my hesitancy in resubmerging a clearly monophyletic unit back into the poorly re- Volume 83, Number 2 1996 Yatskievych 183 Revision of Phanerophlebia Figure 8. commonly associated with the genus F. lemmonii poo Arcey 26 586), onii (D'Arcy 2686), Ns ii spore, sah inflate d folds. Scale vit = Spores of Polystichum species. —b. P. munitum solved quagmire that represents our present, lim- ited knowledge of classification in Polystichum. Particularly in tropical America, Polystichum com- prises a diversity of species clusters whose inter- relationships have not been studied in detail. We not only do not know which of these species clus- (Yatskie with a perispore type simil ar to d at found in Phanerophlebia and Cyrtomium. ichitis od et al. 2482), showing the reticulate perispore h & i / "set ‘Crary 84— with a cristate-microperforate showing smooth exine, rugulose inner perispore, and outer perispore ters is most closely related phylogenetically, or even morphologically, to the Phanerophlebia clus- ter, but we do not have a clear idea how to polarize the states for the various morphological and other characters separating these groups. Because no sta- similar to that of ble infrageneric classification, 184 Annals of the Missouri Botanical Garden Daigobo (1972) for Japan, presently exists for the New World species, it seems most prudent to avoid publishing the new combinations necessary to in- corporate all of the presently recognized taxa of Phanerophlebia into Polystichum. TAXONOMIC TREATMENT Phanerophlebia C. Presl, Tent. Pterid. 84, 1836. TYPE: Phanerophlebia nobilis (Schltdl. & Cham.) C. Presl (Aspidium nobile Schltdl. & Cham.). Amblia C. Presl, os Pterid. só pes oH Amblia juglandifolia mb. & B illd.) C. Presl [= Pha mero prs en & Bon pl. ex Willd.) J. Sm.]. Plants perennial, often long-lived, terrestrial, sometimes growing from soil pockets in rock crev- ices, or epipetric on rock faces; rhizomes dictyo- stelic, scaly, short-repent to erect, branched at maturity, usually covered with persis- sometimes tent petiole bases and adventitious roots; rhizome scales unevenly ciliate, the cilia sometimes shed during maturation; leaves clustered at apex of rhi- KEY TO THE TAXA OF PHANEROPHLEBIA monomorphic, evergreen; petioles shorter than or sometimes longer than laminae, scaly (the scales sometimes shed at maturity), with 4-8 vas- cular bundles arranged in a ring; laminae 1-pinnate (rarely simple) with conform terminal pinnae, char- subcoriaceous; zome, taceous to rachises and costae usually somewhat scaly with reduced scales, grooved adaxially, the grooves + confluent at junc- tions; pinnae asymmetrically subcordate to cuneate at base, sometimes with an acroscopic, basal auricle, spinulose-serrulate along margin, at least distally, ovate to narrowly lanceolate, often somewhat falcate, lacking trichomes, but with reduced, often hairlike scales along veins (these grading into the broader, more typical scales along the rachis and petiole); venation anadromous (this often difficult to observe), pinnate, the veins 1—4-branched, extending + to margin, free or with 1-3 series of marginal anasto- moses; sori terminal or apparently lateral on veins, round, in (1-)24 series between costa and margin; indusia absent or present, if present then peltate, with erose margins, persistent at maturity or fuga- cious; sporangia with 64 spores; spores monolete, dark brown; perispore with prominent, inflated folds; exospore smooth. Chromosome number: x — 41. (the rare, sterile hybrid, P. juglandifolia X P. macrosora is not included, but may key imperfectly to one of its parents) la. Veins commonly anastomosing toward the margins, in 1—3 series (less frequent in P. haitiensis), these some- times distributed irregular y. 2a. Sori indusiate (the indusium shriveling at maturity or fugacious); rhizome and petiole scales concolorous, reddish brown. 3a. Lateral pinnae (2-)6-17 pairs; pinna margins spinulose-serrulate nearly to base 6b. Phanerophlebia nobilis var. I. remotispora 3b. Lateral pinnae 2-4 pairs; pinna margins spinulose-serrulate only in distal half . Phanerophlebia haitiensis 2b. Sori exindusiate: rhizome and petiole s scales bicolorous, with darker, sclerotic canes and lighter m 4 s usually overlapping, 6— i with broad, dark, sclerotic centers and narrow, hyaline Phanerophlebia gastonyi on 11 mm Door with dark brown central stripes and qe da of about nai width; buds often present on rachis in axils of distal pinnae; spores m lon _ 4. Pha -— = ~ : in free, or with a few rare anastomose din habia yu ndifoli ia lla pinnae 0—3(—5) pairs; ln usually <40 cm long; broadest petiolar eigen <2 m 5b. age pinnae (2-)5-17 pairs; pues usually >40 cm a broadest petiolar scales 52m 2.7 m long; rhizome scales 1( iolar scales stramineous an; leaves to aa plants with an e skunklike scales 2.5-7.5 mm long, 7a. Indusia not shriveled at iE. + odor be 5. Phan 6b. Petiolar scales brown or reddish 2€ (rarely lighter c olored with age); leaves to 1.25 m lon wide; plants without a persistent, wide 7. Phanerophlebia pumila +15 mm er. 7 mm si ose macrosora ; rhizome unpleasant, skunklike odor Kn , fresh. “with a raised, central projection m . Phanerophlebia umbonata 7b. Indusia shriveled at maturity, often fugaci ious, flat to concave centrally, not held te. t least some pinnae wit a prominent, acroscopic auricle at base; plants of d soil pockets in rock crevices and ledges of mesic canyons; SW Un ited States ico l. io auriculata ‘ly in 8b. Pinnae rarely auriculate at base; plants of stream banks and forest Meca (rarely sinkholes or rock crevices); nearly Mexico in mountainous throughout s 6a. Phanerophlebia nobilis var. nobilis Volume 83, Number 2 1996 Yatskie vych 185 Revision of Phanerophlebia 1. Phanerophlebia rans Underwood, Bull. Torrey Bot. Club 26: 212 vyrtomium auriculatum Tu un C. V. Morton, Amer. Fern J. 47: 54. 1957. TYPE: Mexico. Chihua- hua: cool, damp cliffs, Mapula Mountains, 17 Oct. 1886, Pringle 831 (holotype, NY; iso- types, CAS, F, GH, K, LL, MICH, MO, NY, P, US) Plants not strongly scented; rhizomes to ca. 15 mm diam., deeply seated in substrate, short-repent to ascending, sometimes branched at maturity; rhi- zome scales 3.5—7.5 mm long, 1—4 mm wide, ovate to elliptic-lanceolate, ciliate, concolorous, brown (rarely lighter colored with age); leaves 10-60(-75) cm long (very short leaves sometimes fertile); pet- ioles shorter than to nearly as long as laminae; pet- iolar scales sometimes deciduous, dense and over- lapping, much like rhizome scales, the broadest ca. 4 mm wide, grading into reduced, hairlike struc- tures above; pinnae (2-)5-12 pairs, to 9 cm long, ovate to lanceolate, usually falcate, the apex usu- ally attenuate, the base obliquely cuneate to sub- cordate and usually with a prominent, acroscopic auricle (at least some present on every plant), rarely irregularly incised, the margins spinulose-serrulate nearly to base; buds absent from axils of distal pin- nae; veins free, 1-3-branched; sori in 2—4 series between costa and margin; indusia 0.6—0.9 mm diam., membranous, flat or concave centrally, not umbonate, shriveled at maturity; spores 41—60 4 long. Chromosome number: n — Illustrations. See original description; also Ditt- mer et al. (1954: 38), Knobloch and Correll (1962: 164, pinna), Mickel (1979: 163) Phanerophlebia auriculata grows in soil pockets in rock crevices (usually granite or quartzite); it is restricted to mesic, sheltered ravines and canyons, in oak and pine-oak forests; 600-2300 m; south- western United States (Arizona, New Mexico, Tex- as) and northern Mexico (Chihuahua, Coahuila, So- nora) (Fig. 6). Clausen’s (1949) report of this species from Nuevo León was based upon misde- termined specimens of P. umbonata. Isozyme studies (Yatskievyc ho 1990: Yatskievych & Gastony, 1987) have indicated that Phanerophle- bia auriculata is an allotetraploid taxon. An ex- amination of chloroplast DNA variation in the ge- nus (Yatskievych, 1988) implicated the diploid P. nobilis as one potential parent. Based on morpho- logical characters and distribution, the diploid P umbonata appears to be the most likely extant can- didate for the other parent. These three taxa are very similar in details of leaf shape and size, rhi- zome and petiolar scale shape and coloration, pinna number, and venation. Sterile specimens of the two presumed parents are virtually indistinguishable morphologically. Phanerophlebia auriculata shows no tendency toward umbonate indusia, a distin- guishing character of P. umbonata, but does have a tendency for branched rhizomes, which are com- mon in P. umbonata, but not produced in P. nobilis. The acroscopically auriculate pinnae of P. auricu- lata are not found in P. umbonata, but can be en- countered rarely in P. nobilis. Small auricles are also encountered commonly in another tetraploid, P. pumila, suggesting that this character may arise as a side effect of the combination of parental ge- nomes giving rise to these allopolyploids. e dues specimens. U.S.A. Arizona: Cochise County, short canyon near mouth of Garden Canyon, Hu- achuca Mountains, d 162-50 (ARIZ, MICH); ¢ Gila County, SE wall « rker Canyon, e Ancha, ham 0109B (ASC Y eph C ounty, J Arava the “Box” below Klondyke, Philli (ARIZ, CAS, GH, y Superstition Mountains Goodding 6146 (ARIZ); County, near the top and on the N side of Mt. Baboquivari, Gould, Darrow & Haskell 2788 (ARIZ); Santa Cruz Coun- 'amore jedes Pajarito Mountains, T. & M. Van .n. (A ; Yavapai a : Canyon, "e Sedona, TEE. 4788 (ARIZ, ; Yuma County, he arrow canyon, Kofa ser line d Hinck- ley F- 9. 39 (MICH, US). New Mex ma Ana County, Ice Canyon, Organ Mountains, Worthington 7679 (A NY, UTEP); Socorro County, Luna Park Campground, San Mateo Mountains, Reeves et al. 88-1 (NMC). Texas: Cul- ved County, Victorio Canyon, E margin of Sierra Dia- y dele Wendt & Chic ro : 10728 (CAS, LL); El y, Hueco Tanks, o Mountains, Waterfall 6635 (CH. MO, NY); Jeff da County Rose mi. N of Alpine, ir enn 21779 Canyon, 15 JL). MEXICO. Chihua- ua: age of TIER, Correll & Johnston 21772 (LL, Mojada, Stewart 2198 (G : ispe, region of Río de aan he Phillips 546 (ARIZ, GH, MICH, NY). Y 2. Phanerophlebia gastonyi Yatskievych, Novon 2: 445. 1992. TYPE: Mexico. Chiapas: ca. 13 km NW of Berriózabal, limestone outerops in evergreen tropical hardwood forest, elevation 1000 m, 15 July 1985, Yatskievych, González L., Ranker, G. Starr & C. Starr 85-182 (holo- type, MO; isotypes, ARIZ, CHAPA, IND, MEXU, NY, UAMIZ, UC). Plants not strongly scented; rhizomes to ca. 5 mm diam., superficial on substrate, short- EUREN not branched at maturity; rhizome scales 2.54.5 mm — e ng, 1.5-2 mm wide, ovate to elliptic-lanceolate, ciliate, bicolorous with broad, dark, sclerotic cen- ters and narrow hyaline margins; leaves to 45(—55) em long; petioles slightly shorter than to usually longer than laminae; petiolar scales usually decid- Annals of the Missouri Botanical Garden Figure 9. b, b'. Details of pinnae showing venation, sori, indusi scales. (a—c from Yatskievych et al. 85-18. uous with age, mostly not overlapping, much like rhizome scales, the broadest 2 mm wide, the re- structures uncommon among the pinnae (0—)1—3(—4) pairs, to 12(— sometimes duced hairlike broader scales; 15.5) em long, ovate to lance-ovate, somewhat falcate, the apex acute to attenuate, the Phanerophlebia gastonyi (a—c) and P. nonis var. remotispora (a—c ^» a, and pinna Vue —e, 2; a'—c ^ from Yatskievych 85-1 e 3 —a, a’. Representative leaves. — '. Representative series of petiolar base obliquely cuneate to rounded, nearly equilateral, the margins entire to slightly un- dulate proximally, spinulose-serrulate only in distal half; buds absent from axils of distal pinnae; veins with 1—3 series of regular marginal anastomoses, 2—4-branched; sori in (1-)2—4 series between costa sometimes Volume 83, Number 2 1996 Yatskievych 187 Revision of Phanerophlebia and margin; sori exindusiate; spores 30-42 ¡um long. Chromosome number: n = 41 Illustrations. Figure 9a—c; see also Mickel and Beitel (1988: 520, as P Juglandifolia). Phanerophlebia gastonyi grows in moist soil, less commonly on fissured limestone rock faces, in shaded understory of cloud forests and montane rainforests; 900-2200 m; southern Mexico (Chia- pas, Oaxaca, Veracruz) (Fig. 10). Specimens of Phanerophlebia rd have been identified by previous workers as acier ian but molecular studies (Yatskievych et al., 1988) ha identified it as a progenitor of that species. The two taxa are very similar morphologically, but can be separated by the characters included in the key. Dip- loid P. gastonyi is treated here at species rank, rath- er than as a variety or subspecies of P. juglandifolia, because isozyme studies (Yatskievych, 1990; skievych & Gastony, 1987) have suggested that the latter, tetraploid taxon is of allopolyploid origin, with the other diploid parent not yet identified. Representative specimens. ME ). Chiapas: 13 km N of Berriózabal, Breedlove & a 203 (DS, F, MS NY). Oaxaca: trail from San Pedro Nolasco N to the Lla- no Verde, Mickel & Hellwig 3786B (NY). Veracruz: ca. 12 km S of Misantla, Conant 726 (GH, MEXU). ow . Phanerophlebia haitiensis C. Christensen, Kongl. Svenska Vetenskapsakad. Handl. II 16: 42 + pl. 10, figs. 1—4 . Cyrtomium haitiense (C. Chr.) C. V. Morton, Amer. Fern J 47: 55. 1957. TYPE: Haiti. Massif de la Selle, Ganthier, Badeau, deep gulch above Badeau, alt. 2000 m, 28 Jan. 1925, Ekman 3119 (ho- lotype, S; isotypes, BM, US). Plants not strongly scented; rhizomes to ca. 10 mm diam., apparently deeply seated in substrate, short-repent to ascending, not branched at maturi- ty; rhizome scales 4.5-7.0 mm long, 2—4 mm wide, ovate to elliptic-lanceolate, erose-denticulate, with few, short cilia at base, concolorous, brown (rarely lighter colored with age); leaves to 50 cm long; pet- ioles slightly shorter than to longer than laminae; petiolar scales usually deciduous, loosely overlap- ping, much like rhizome scales, the broadest ca. 3 mm wide, grading into reduced, hairlike structures above; pinnae 1—4 pairs, to 9 cm long, lanceolate to lance-ovate, usually falcate, the apex acute to attenuate, the base unevenly cuneate and lacking an acroscopic auricle, the margins sometimes slightly undulate, spinulose-serrulate in distal half; buds absent from axils of distal pinnae; veins with irregular marginal anastomoses, 2—3-branched; sori in 1-2(23) series between costa and margin; in- dusia 0.7—1.1 mm diam., cave centrally, not umbonate, shriveled at maturity; spores 35-56 ¡um long, often lacking a well-developed membranous, flat or con- perispore. Chromosome number unknown. Illustrations. See original description. Phanerophlebia haitiensis grows on protected limestone cliffs; 2000-2500 m; Central America (Guatemala, Honduras), Haiti (Fig. 10). This perplexing taxon is apparently of hybrid or- igin, based on its irregularly sized spores with re- duced perispore deposition and its irregularly anas- tomosing veins. Examination of stomates from a cleared pinna fragment indicates a size range in agreement with polyploid taxa in the genus (see previous discussion). It has not been re-collected in Haiti since Ekman's original e ns (both from the same locality). John T pers. comm.) has searched for this species and failed to locate it. It is therefore presumed extinct there. Single collections from Guatemala and Honduras (see below and Fig. 10) share with the Haitian col- lections the irregularly anastomosing venation and spores of variable size with apparently reduced per- ispore deposition. The Guatemalan collection pos- sesses few mature sporangia, however, and plants at this locality require further study. Smith (1981) sug- gested affinities between this species and P. lindenii E. Fourn., which is here treated as a synonym of P. pumila (M. Martens & Galeotti) Fée. See the treat- ment of the latter species for further discussion. Specimens examined. HAITI. Massif de la Selle, Croix des Bouquets, ravine between M. Mérillon et M. Badeau, type loca ality, E EMALA 324-325 on Ruta Na- cional ON between Chemal a San Juan Ixcoy, Sierra de los Cuc humatanes, se ias 3056 (GH, TEX, UC). HON- anta Bárba km N of El Mochito on E slopes of ML Santa vee Moran 5706 (MO) E is) E E * - ^ s = © x — 4. Phanerophlebia juglandifolia (Humboldt & Bonpland ex e Smith, J. Bot. (Hook- er) 4: 187. 1841. fe de di ae Humb. & Bonpl. ex Willd., 4) 5: 195. 1810. Amblia leuis adi & Bonpl. ex Willd.) C. Presl, Tent. Pterid. 185. 1836. Aspidium prre (Humb. & Bonpl. ex Willd.) Kunze ex Klotzsch, Linnaea 20: 363. 1847. Cyrtomium juglandifolium (Humb. & Bonpl. ex Willd.) T. Moore, Index Fil. lxxxiii. 1857. Dryopteris juglandifolia Eg & Bonpl. ex Willd.) Kuntze, Revis. Gen. Pl. 2: 813. 1891. A Juglandifolium ihm & Bonpl. ex Willd.) Diels in Engler & Prantl, Nat. Pflanzen- fam. 1(4): 191. 1899. TYPE: Venezuela. Mona- gas: Caripe, Humboldt & hor ein s.n. (holo- type, B—Herb. Willd. 1 isotypes, P, NY (fragment)). 9688, sheets la, lb: 188 Annals of the Missouri Botanical Garden 3 500 km Lx % e - a : : C em AT S : E insi "T 994 d ^ ‘wy z j AQ a Phanerophiebia @ gastonyi e haitiensis @ juglandifolia » macrosora ^ Figure 10. Distributions of Phanerophlebia gastonyi, P. haitiensis, P. juglandifolia, and P. macrosora, based upon aa specimens examined. Plants not strongly scented; rhizomes to ca. 7 mm diam., generally deeply seated in substrate (some- times superficial), short-repent to nearly erect, not branched at maturity; rhizome scales 6.0—10.5 mm long, 3-5 mm wide, lanceolate, ciliate, bicolorous with broad or sometimes narrow, dark brown, some- what sclerotic centers and hyaline margins as wide as centers or narrower; leaves to 60(-85) em long; petioles shorter than to slightly longer than lami- nae; petiolar scales subpersistent, overlapping and often dense, much like rhizome scales, the broadest 4 mm wide, grading into reduced, hairlike struc- tures; pinnae 2—4(—6) pairs, to 17.5 cm long, ovate to lance-ovate, usually somewhat falcate, the apex attenuate, the base obliquely cuneate to rounded, the margins often slightly undulate proximally, spi- nulose-serrulate in distal half or more commonly in distal two-thirds; buds present on at least some leaves in each population, the gemmae in axils of distal pinnae (rarely in axils of more proximal pin- nae); veins with 1—3 series of regular marginal anastomoses, 3—5-branched; sori in 2-4(-5) series between costa and margin; sori exindusiate; spores 41-60 um long. Chromosome number: n Illustrations. Humboldt et al. (1825, 7: pl. 665, as Polypodium juglandifolium), Underwood (1899: pl. 359—360, pinna), Vareschi (1969: 368, pinna); Stolze (1981: 166, Phanerophlebia juglandifolia grows in moist soil as Cyrtomium Juglandifolium). of ravines and slopes, associated with limestone or volcanic substrate, in rainforests and evergreen cloud forests, rarely in pine-oak forest; (350—)700— 2700 m; eastern and southern Mexico (Chiapas and disjunctly in Hidalgo and Veracruz), nearly throughout Central America, to Colombia and Ven- ezuela (Fig. 10). Previous workers were unaware that Phanero- phlebia juglandifolia, as traditionally circum- scribed, was a heterogeneous mixture of a wide- spread, allotetraploid taxon (to which the type of this name can be ascribed) and one of its diploid progenitors, which is here treated as P. gastonyi. Unfortunately, the identity of the other parental tax- on remains unclear. Compounding the problem of morphological circumscription of P. juglandifolia was the discovery of rare, naturally occurring hy- brids between this species and P. macrosora. These hybrids, which are apparently restricted to the mountains of central Costa Rica, are sterile trip- loids of intermediate morphology, and were rela- tively commonly encountered the single site where I was able to observe these plants. They are Volume 83, Number 2 Yatskievych 189 y Revision of Phanerophlebia morphologically variable, though most are robust plants with the overall aspect of P. macrosora. They 'an be separated from both parents by their some- what petiolar and scales, irregularly anastomosing venation, highly reduced or absent indusia, and malformed spores of the sporangia also abort early in their de- smaller, bicolorous, rhizome many « oe The only specimens seen that re present is hybrid are both from medium elevations (1 0-2000 m) in the volcanic mountains of central Costa Rica (Alajuela, La Ventolera, on S slope of Volcán de Poás, Standley 34531 (GH, US); Here- dia, 2 km SE of Sacramento on hwy. 114 from San José de la Montafia, S slope of Volcán Barva, Yat- skievych & McCrary 86—31a (CR, IND, MO, NY)). me specimens (of P. cid i ia). MEXICO. Chiapas: above Finca Cuxtepec, mpo. Angel Sb ¢ orzo, Breedlove & Seeder 46731 (CAS). Hidalgo: 8-9 km N de M on Molango, Her- nández M., Cortés & Her idc MO). eracruz: hacia el arroyo s Toluca, reda que va a e, Nee Huayacoc votla, Palma G 67 (XAL). GUATEM . Alta Verapaz: trail between Sepacuité and Sec poria Maxon & Hay 3289 (US). El Quiché: mountain inim 8 of Nebaj, Proctor 25075 (DS, LL. TEX. US ntla: between Santa María de oes and Palfn, Standley 65 300 (F). Huehuetenango: above San Juan Sierra de Cuchumatanes, "nece 50036 (F, US). E ango: along road between La Finca Pirineos and Patzulín, aru 87078 (F, US). San Marcos: | mi. above Africa, 3.3 mi. above Finca Armenia above San Rafael, Croat 40940 e EL SALVADOR. Santa Ana: NE iw dee de los Naranjos, Tucker 1288 (BH, GH, K, LL, , NY, i HONDURA S. Lempira: Celaque E p trail from Camp 1 to Río Naranjal, Moran 5565 (MO). Santa Barbara: Cerro Santa Barbara, 10 km E de Lago a Clewell & Hazlett e wane Qu. A- galpa Bde n la miU Mat a: bed Palcila, Moreno 7014 : Cerr S Cale SW of Jinotega, "Stanley dos y Us COSTA RICA. e Tapesco de La AI- faro Ruiz, Smith 1463 (GH, NY). Ca 0: CE Brade & Brade 53 (GH). Heredia: Pea de Barve (CR). Puntarenas: e Verde, around Fa AL along Hío ese 'imal below Ta therfa, Hammel & Trainer 13825 (MO). San José: along Que ebrada Tab- lazo ae on forested ie creek, Tablazo, Grayum & Schatz 5166 (MO). riqui: mi. from Gualaca on road to Cerro Hornito, Antonio 1752 (MO). VENEZUELA. Distrito Federal: Cerro de Avila, Quebrada Chacaito, Manara s.n. (MO). COLOMBIA. Cundinamarca: Ix de Teque ndam, Bo- gotá, o de J. Triana 61 (P). Magdalena: Sierra de Perijá, 10 km ENE of Manure, 3 js from Venezuelan border, Grant 10799 (GH). — 5. Phanerophlebia macrosora (Baker) Under- wood, Bull. Torrey Bot. Club 26: 213. 1899. Aspidium juglandifolium (Humb. & Bonpl. ex Willd.) Kunze ex Klotzsch var. macrosorum Baker, J. Bot. 25: 25. 1887. Cyrtomium ma- crosorum (Baker) C. V. Morton, Amer. Fern J. 47: 55. 1957. TYPE: Costa Rica. Without fur- ther locality, 1886(?), Cooper s.n. (holotype, K; isotypes, GH, NY (fragment), US). d arie ies Underw., Bull. Torrey Bot. Club 26: 214. 1899 os Guatemala Quiché: San a Tie. a. 7000 pp., Apr. 1892, Heyde & Lux s.n. [herb aset Smith 3241 | e NY; isotypes, GH, > US). Plants with a strong, unpleasant, skunklike odor when fresh; rhizomes to ca. 20 mm diam., deeply seated in substrate, erect or ascending, unbranched at maturity; rhizome scales 10-15 mm long, 5—7 mm wide, ovate to lance-ovate, erose-ciliate, con- colorous (rarely with a slightly darkened central area); leaves 0.7-2.7 m long; petioles shorter than to nearly as long as laminae; petiolar scales persis- tent, dense and overlapping, much like rhizome scales, the broadest ca. 7 mm wide, mixed with reduced, hairlike structures above; pinnae (4-)6— 17 pairs, to 27 cm long, narrowly oblong-lanceo- late, occasionally slightly falcate, the apex attenu- ate, the base obliquely cuneate to rounded and lacking an acroscopic auricle, the margins spinu- lose-serrulate nearly to base; buds absent from axils of distal pinnae; veins free, 3—4-branched; sori in 2—4 series between costa and margin; indusia 0.6— 1.1 mm diam., membranous, flat or concave cen- trally, not umbonate, shriveled at maturity; spores 41—60 um long. Chromosome number: n Illustrations. Smith (1981: 330), Stolze (1981: 166, as i E macrosorum), Mickel and Beitel (1988: 5 Pini macrosora grows in moist soil in mesic ravines and on shaded slopes, usually with igneous or volcanic substrate, in montane rainfo- rests and cloud forests, rarely in oak forests; 1800— 3200 m; eastern and southern Mexico (Chiapas, Hi- dalgo. Oaxaca, Veracruz), through Central America, to western Panama (Fig. This interesting species » easily recognized by its large, coarsely divided, brittle leaves with a thick, coriaceous texture, by its densely and per- sistently scaly petioles, and by its pronounced skunklike odor (first brought to my attention by John T. Mickel). This pungent, disagreeable odor is noticeable from more than 10 m in the field and can be an aid in locating plants. The compounds responsible disappear during drying and have not been identified, although some mixture of volatile terpenoids is suspected. The species is of further interest within the genus (and among ferns in gen- eral) for its geniculate petioles. The petiole bases deviate nearly horizontally from the deepset, nearly erect rhizomes, then curve upward, but the laminae are again oriented nearly horizontally. The leaves thus have a pronounced sigmoidal curvature at ma- turity. For a discussion of the rare, sterile hybrid be- 190 Annals of the Missouri Botanical Garden tween this species and Phanerophlebia juglandifol- ia, see the treatment for that species, above. Representative specimens. MEXICO. Chiapas: SW side of Cerro Mozotal, 11 km NW of junction of road to Motozintla along road to El Porvenir, Breedlove 41713 S). Guerrero: El Asoleadero, 15 km W de Camotla, Rzedowski 18531 (ENCB). Hidalgo: El Potrero, carretera Metepec-Tenango de Doria, Gimate L. 975 (ENCB, F, MEXU, NY). Oaxaca: 26-29 km NE of Teotitlán del Camino, vicinty of pass at Pto. Soledad, Mickel & Hellwig 4156 (NY). Veracruz: La Pandura, camino Ingenio El Rosario a Xico, mpo. Xico, pee hig F. XAL). GUATEMALA. srg eid a S of re shi Madison 670 (GH). El Progres sb Paid Finca Pia- monte and top of Montaña da along Joya Pacayal, Steyermark 43709 (F, US). Quetz Volcán de Zunil at and above Aguas 65420 ud z Marcos: slopes of Cerro Tumblador, ca. 15 km W n Marcos, Williams, Molina R. & Williams 23058(US). P la lá: Volcán San Pedro, N-facing slopes to- ward Lago de Atitlán, above village of San Pedro, Stey- 248 H )). Suchitepéque ermark 47. z: Volcán Santa Clara, between Finca El Naranjo and epa vedi RM io 46717 (F, US). B SALVADOR. Chalaten : Cerro El C road up Pital, Seiler 418 (F). COSTA RICA. pra Volcán Irazu, 0.5 mi above Cartago, Rossbach 3079 (GH). Heredia: N of He- redia, ca. 1 km beyond Porrosatí, Lellinger & White 1679 (F, US). Puntarenas: upper Río Burú, Gómez, Chacón, Chacón & Herrera 21432 (CR, y Copay, Tonduz 11930 (US). e Chiriquí, 7.3+ mi. from Boquete, Armond 534 (CAS, above Sanatorio Duran, ca. 9 mi. ~ "T2 — 6. Phanerophlebia nobilis (Schlechtendal & Chamisso) C. Presl, Tent. Pterid. 85. 1836. As- pidium nobile Schltdl. & Cham., 610. 1830. Cyrtomium pe (Schltdl. & Cham.) T. Moore, Index Fil. ii. 1857. TYPE: Mexico. Veracruz: Hacienda de Lagu- na, Oct. 1828, Re s.n. (holotype, HAL). Linnaea 5: Ixxxiii. Plants not strongly scented; rhizomes to ca. 15 mm diam., erect or ascending, usually unbranched at maturity; rhizome scales 3.5—7.5 mm long, 2 mm wide, ovate to elliptic-lanceolate, denticulate or deciduously ciliate, concolorous, brown; leaves to 1.2 m long; petioles shorter than the laminae; petiolar scales sometimes deciduous, dense and overlapping, much like rhizome scales, the broad- est ca. 4 mm wide, grading into reduced, hairlike structures above; pinnae (2-)6-17 pairs, to 17 cm long, lanceolate to linear-lanceolate, usually at least somewhat falcate, the apex attenuate, the base obliquely cuneate, rarely subcordate and with an acroscopic auricle, the margins spinulose-serrulate nearly to base; buds absent from axils of distal pin- nae; veins free or with 1-3 series of areoles toward the margin (these sometimes distributed irregular- , 1-3-branched; sori in 2-4 series between costa hong and margin; indusia 0.6—0.9 mm diam., nous, flat or concave centrally, not umbonate, shriv- eled at maturity; spores 41-60 um long. Chromo- some number: n = membra- Two varieties are separable based on differ- ences in venation. These were accorded specific rank by earlier workers, based on the view that reticulate versus free venation was a character of fundamental importance in the classification of Phanerophlebia species. The discovery that there are no other morphological characters that reliably separate these two taxa correlates with results from chloroplast DNA analyses, which also sug- gest a close affinity between the two (Yatskievych et al., 1988). Phanerophlebia remotispora has pre- viously been classified closer to P. juglandifolia than to P. nobilis, based on the assumption that reticulate venation had arisen only once in the group. Studies contrasting P. juglandifolia and P. ee (Yatskievych, 1990; Yatskievych et ., 1988) showed that these two taxa are consis- tently and readily separable using morphological, isozymic, and chloroplast genomic characters. It is of interest that data from chloroplast DNA anal- yses also indicated that reticulate venation has arisen independently twice in the genus (Yatskiev- ych et al., The varietal designation is here used for two ap- parently interfertile morphotypes within a species, lacking sharp biogeographic discontinuities, and which are separated only by a single morphological character whose genetic basis remains unknown. The two varieties of Phanerophlebia nobilis are sympatric throughout much of Mexico, except for the absence of variety remotispora in the north- western portion of the species’ range, but the two varieties have not often been found growing togeth- er (see below). 6a. Phanerophlebia nobilis var. nobilis Venation free, or with a few, rare anastomoses. Illustrations. Kunze (1844: pl. 67, as Aspidium nobile), Mickel and Beitel (1988: 517 Phanerophlebia nobilis var. nobilis grows in moist soil, rarely on rock faces (in sinks), with limestone, igneous and volcanic substrate; shady, mesic ra- vines in oak, pine-oak, pine, or pine-fir forests, rarely in cloud forests, often associated with Alnus, or other streamside trees; 1000-2800(-3200) m; Mexico (Chiapas, Chihuahua, Distrito Federal, Guerrero, Hidalgo, Jalisco, México, Michoacán, Morelos, Oaxaca, Puebla, Querétaro, San Luis Po- tosí, Sonora, Tamaulipas, Tlaxcala, Veracruz), to be Volume 83, Number 2 Yatskievych 191 Revision of Phanerophlebia 500 km ww 4 S pem 2 o o de fe) O d ) Y à b i 95°w / 25%N [e y + o) o) * 0 o) MER O S o e gj i o 9 O Oe o E : ee , ? P nobilis: s e eo e 0 E, O var. nobilis ee e - A^ .T e var. remotispora 89899 e adjacent or mixed populations ure ll. 18 mens examinec expected in all states except the Yucatán penin- sular lowlands and the Baja California Peninsula (Fig. 11). The single collection known from Chiapas is morphologically atypical and was ascribed to Pha- nerophlebia macrosora by Smith (1981), who also noted its odd morphology. A single specimen la- beled as having originated from Brazil (Rio Grande o Sul, Porto is Reineck s.n. (GH)) either represents Alegre, Gebiisch unter der Rua da an escape from cultivation or more probably a mis- labeled specimen (Alan R. Smith, pers. comm.). Plants of this variety tend to have somewhat shorter leaves (to 75 cm) than those of variety re- motispora and commonly have fewer than 10 pairs of lateral pinnae. These differences disappear un- der common greenhouse culture, and are probably under environmental rather than genetic control. In central and southern Mexico, variety nobilis is often found at somewhat higher elevations and in conif- erous forests, although there is nearly total congru- Distributions of Phanerophlebia nobilis var. 1 nobilis and var. remotispora, based upon herbarium speci- ence of elevational ranges between the two varie- ties. Representative specimens. MEXICO. € rom Mt. Male, near Diod Matuda 4688 "s ue ^. Chi- ahua: $ a Charuco, Arroyo Hondo, ae 7988 (ARIZ, DS, MEXU. MICH, US). Distrito Federal: Cañ- ada de Contreras, cerca del Primer Dinamo, erp 27212 (DS, ENCB, F, LL, MICH). Guerrero —40 km by road W of Chilpancingo, near and above ol town of Omiltemi, Anderson & Laskowski 4363 (ENCB). Hidal- o: 4 km N de Tlahuelompa, mpo. Zac dd Riba 66 (MEXU). os l km above nt of El Isote, in pine-fir zone above or NW slopes of Nevado de Co- lima, Mc Vaugh 10125 (MEXU, MICH, MO, US). Méxic 00: San Rafael, Walmánsico, Pe 1649 Michoacán: Cuincho [( ointzio|, pres Mars shia, Arsène s.n. (F, GH, P). Morelos: mpo. Tepoztlán, Camacho G. 10-5 trail from San Pedro Nolasco N to the Hanis Verde, Mickel & Hellwig 3786a One rmn He del j dig Lyonnet 3240 (MEXU, US). Querétaro: Agua Blanca, 10 S de Pinal de iere Fe ceo & Zamudio 992 MO). San Luis Potosí: in montibus San Luis Potosí, Schaffner s.n. (NY). Sinaloa: 2.8 mi. NE of El Paraiso, > seltleme km E de S (ENCB). mpo. 2D” ~ 192 Annals of the Missouri Botanical Garden 60 mi. SW of El Salto, road between pee Deo. and Villa Union, Sin., Ownbey & Ownbey 1936 (GH, US). So- nora: das , Río Mayo, Gentry 3660 (ARIZ, E H, MO, UC). T. maulipas: above Casa Piedra on trail to Agua del rn ed i sierra de Guatemala, ca. 7 km WNW of Gomez s, Johnston 12777 (TEX). Tlaxcala: Tlax- cala, Bl “Puebla, Aen s.n. (A, F, GH, K, MEXU, MICH, MO, P, UC, US). Veracruz: head D Orizaba Val- ley, Copeland s.n. (GH, MEXU, MICH, P, UC). “y 6b. Phanerophlebia nobilis var. remotispora ( 2: 446. 1992. load hace remotispora E. Fourn., Mexic. Pl. 100. d Cyrtomium remoti- sporum (E. Fou . V. Morton, Amer. Fern J. 47: 54. 1957. pet Mexico. Veracruz: Ori- zaba, Izhuatlancillo [Ixhuatlancillo], 1865-1866 [sheet at K = Bourgeau 2349 (holotype, P; isotypes, GH, K, MICH, MO, NY, P, US tj ^. Fournier) Yatskievych, Novon Aug. 20 May 1866], BM, Venation commonly anastomosing toward the margins, with 1-3 series of areoles, these some- times distributed irregularly. For further compari- son with variety nobilis see treatment of that variety, above. Illustrations. Figure 9a'—c'; see also Stolze (1981: 166, as Cyrtomium remotisporum), Mickel and Beitel (1988: 520, as Phanerophlebia remotis- pora Phanerophlebia nobilis var. remotispora grows in moist soil, rarely on rock faces (in sinks), with lime- stone, igneous and volcanic substrate, in mesic ra- vines and on slopes, in oak, pine-oak, and pine forests, cloud forests, less commonly in montane rainforests or deciduous or subdeciduous forest as- sociations; also found along brushy roadsides, in overgrown coffee plantations, and in some dis- turbed or secondary forests; (300—)600-2300 m; southern and eastern Mexico (Chiapas, Hidalgo, México, Michoacán, Oaxaca, Puebla, San Luis Po- tosí, Veracruz), Guatemala (Fig. 11) This is the only Phanerophlebia taxon to be found in secondary forests or disturbed roadside areas. The single collection from Michoacán is a mixed gathering with the sheet cited from CM rep- resenting variety remotispora and the other sheets (at A, GH, K, MEXU, MICH, MO, P, and UC) rep- resenting variety nobilis. Representative specimens. MEXICO. Chiapas: slope of the sumidero in esp praia npo. Tenejapa, Breed- love 10768 (DS, MICH, US). Hidalgo: 5 km N de Ten- ango de Doria, harap a Hue ibi Hernández M M. 333 (ENCB, MO). M uev r. Temascalte Hinton 403 (NY. Michoacá "Án: cu jus [Cointzio], den Morelia, Arséne 5415 (CM). Oaxaca: vicinity of lumber camp on top of Cerro San Felipe, on road off Axaca—Ixtlan NX Highway, 21 km N E Oaxaca, Mind 361 (NY). Puebla: Villa Juarez, Riba s ¿XU). San Luis Potosí: 3 km NE del le de Xilitlilla mpo. Xilitla, Rzedowski 10571 ENCB, NY). Veracruz: near Fortín above era plant of Cervecería Moctezuma, Croat 39411 (MO). GUA- — TEMALA. Alta Verapaz: along road to El Ester (Lago Izabal), o mi. E of Highway 14 to Cobán, Croat 41442 E . El Quiché: falls of is las Violetas, 2.5 mi. N of Ne i Proctor 25446 (LL, US). The two varieties of Phanerophlebia nobilis have only rarely been collected in mixed populations or at adjacent sites. Specimens of intermediate mor- phology are also occasionally encountered. These are characterized by pinnae with irregularly anas- tomosing venation, the areolae few and not distrib- uted into a regular, marginal series. Such specimens appear to have well-formed spores. Intermediates between variety nobilis and variety remotispora in- clude MEXICO. Michoacán: Cuincho [Cointzio], Río Gran- de, vic. of Morelia, Arséne 5961 (GH, MO, US). Veracruz: Orizaba, Fisher s.n. (MO) 7. Phanerophlebia gr (M. Martens & Gal- eotti) Fée, Mém. Foug. 5: 282. 1852. Aspidium pumilum M. Martens & Galeotti, Nouv. Mém. Acad. Roy. Sci. Bruxelles 15: 64 + pl. 17, fig. l. 1842. Cyrtomium pumilum (M. Martens & Galeotti) C. V. Morton, Amer. Poe 47:5 1957. TYPE: Mexico. Oa no- Verde et del Carrizal, Mar. 1840, Galeotti 625] (holo- type, BR, photos, BM, LL, MICH, UC; isotype, P) dai lindenii E. de Mexic. Pl. 1: 100 + pl. 4. 1872 P E: Mexico. Chiapas: in pineto pr. Ciudad Heal [San Cristóbal de La as Casas], Mar. [1838], Linden s.n. (holotype, P). Plants not strongly scented; rhizomes to ca. 5 mm diam., usually superficial on substrate, erect or nearly so, not branched at maturity; rhizome scales 3-5 mm long, 2-3 mm wide, ovate to lance-ovate, ciliate-denticulate, concolorous, brown; leaves 304 tile); petioles shorter than to longer than laminae; 5) em long (very short leaves sometimes fer- petiolar scales usually persistent, densely or loosely overlapping, filiform, the broadest 1.7 mm wide, grading into reduced hairlike structures; pinnae 0— 3(—5) pairs, to 8(-12) cm long, ovate to lanceolate, usually somewhat falcate, the apex obtuse to atten- uate, the base obliquely subcordate (occasionally cuneate) and sometimes with a small acroscopic au- ricle, the margins serrulate-denticulate nearly to base, rarely irregularly incised; buds absent from axils of distal pinnae; veins free or uncommonly with few marginal reticulations, 2—3-branched; sori Volume 83, Number 2 1996 Yatskiev 193 vych Revision of Phanerophlebia in 1-2(-3) series between costa and margin; in- dusia 0.6-0.9 mm diam., cave centrally, not umbonate, shriveled or subper- sistent at maturity; spores 41-60 jm long. 2 membranous, flat or Con- Chromosome number: n = Illustrations. Figure 12; see also original de- scri a Underwood (1899: pl. 359—360, pinna), Stolze (1981: as Cyrtomium pumilum), Mickel and E (1988: 520 Phanerophlebia pumila is epipetric on sheltered limestone rock faces, particularly those with over- hangs, and in sinks; 2100-2950(-3700) m; south- ern and western Mexico (Chiapas, Guerrero, Mi- choacán, Oaxaca), Guatemala (Fig. 13). Phanerophlebia pumila is a morphologically vari- able tetraploid of presumed allopolyploid origin. Unfortunately, although isozyme studies (Yatskiev- ych, 1990; Yatskievych & Gastony, 1987) have sug- gested the presence of two parental genomes in this taxon, neither they nor data from chloroplast DNA variation in the genus (Yatskievych et al., 1988) have identified either of the two presumed progen- itors involved. On the basis of habitat and mor- phology, P. nobilis may have been involved, but this taxon does not account for the reduced pinna num- ber and leaf size, nor for the characteristic, narrow petiolar scales. One or both diploid progenitors of e sought this species may be extinct, but shouk by future collectors in the mountains of southern exico. Although Phanerophlebia pumila has been char- acterized by some workers as having extremely re- duced leaves with only 1—5 pinnae, the species is, in fact, extremely variable in this regard. Under greenhouse cultivation, plants with only 1—3 pinnae in the field often produced elongate leaves with up to 11 pinnae (also the largest number encountered in the field). Similar observations have been made independently by John T. Mickel (pers. comm.). Different leaves from various individuals of a single Mexican population (Oaxaca, Llano de Las Flores, near km post 4130 on hwy. 175, N of Ixtlán de Juarez, Yatskievych et al. 85-139 (CHAPA, IND, MEXU, MO, NY)) display great variation in pinna number, size, shape, and distribution (Fig. 12). In- cluded within the range of variation demonstrated in Figure 12 are leaves referable to P. lindenii E. Fourn., a rare morphotype thought to differ from P. pumila in its cuneate (rather than subcordate) pin- na bases. Larger leaves of P. pumila are extremely similar to smaller leaves sometimes displayed by plants of P. nobilis growing in ecologically subop- timal sites, such as the edges of sinkholes. Such plants can usually be determined correctly, based on differences in petiolar scales, but incomplete specimens cannot be determined without exami- nation of stomatal sizes. A single mixed collection of these two species is known (MEXIC( ichoa- cán: Sierra Torrecillas, Distr. Coalcomán, Hinton et al. 12428; P. pumila = GH, K, LL, MEXU, MO, NY, US; P. nobilis var. nobilis = NY). Representative specimens. MEXICO. Chiapas: 10 km ; of El Porvenir along road from Huixtla to Siltepec, Bicedious & Smith 31816 (DS, MICH, NY, TEX). Guer- rero: top of the Sierra Madre near Chilpanci ingo, Nelson 2222 (US). Michoacán: Sierra gue: Coal- comán, Hinton 15935 (DS, ENCB, F, > Oaxaca: trail N of San Pedro Nolasco to re L = Verde, at the high point above the Llano Verde (Las Cruces), Mickel 5377 (NY). GUATEMALA. Huehuetenango: be- tween Tojquiá and Caxín, summit of Sierra los Cuchu- matanes, ^ cea 50211 (F, US). 8. Phanerophlebia umbonata Underwood, Bull. Torrey Bot. Club 26: 211. 1899. Cyrtomium umbonatum (Underw.) C. V. Morton, ern 54. 1957, TYPE: Mexico. León: cool shaded cañons, Sierra Madre, near Monterrey, 14 June 1888, Pringle 1982 (ho- lotype, NY; isotypes, F, GH, MO, NY, P. UC, Amer. Nuevo Plants not strongly scented; rhizomes to ca. 15 mm diam., deeply seated in substrate, short-repent to asc ending, often branched at maturity; rhizome scales 2.54.5 mm long, 2-4 mm wide. ovate to elliptic-lanceolate, ciliate, concolorous, (rarely lighter colored with age); leaves to 90 cm long; petioles shorter than to nearly as long as lam- inae; petiolar scales sometimes deciduous, dense and overlapping, much like rhizome scales, the broadest ca. 4 mm wide, grading into reduced, hair- like structures above; pinnae 10-18 pairs, to 15 em long, lanceolate to linear-lanceolate, usually fal- cate, the apex attenuate, the base cuneate to nearly truncate and lacking an acroscopic auricle, the margins spinulose-serrulate nearly to base: buds absent from axils of distal pinnae; veins free, 1—3- branched; sori in 2-3 series between costa and margin, often submarginal; indusia 0.6-0.9 mm diam., firm, convex with a raised, darker umbo cen- trally, persistent and not shriveling at maturity; spores 41-60 ¡um long. Chromosome number: n = Illustrations. See original description; also Knobloch and Correll (1962: 164); Mickel (1979: 164) Phanerophlebia umbonata grows in moist soil, less commonly among rocks, with igneous and limestone substrate, in sheltered canyons and ra- 194 Annals of the Missouri Botanical Garden Y Volume 83, Number 2 Yatskievych 195 1996 Revision of Phanerophlebia 500 km 95° Ww 7 + 20°N e ? e e f P ço i v € — 1 e e P pumila Figure 13. Distribution of Phanerophlebia pumila, based upon herbarium specimens examined. vines, in oak and pine-oak forests, rarely in cloud MEXU). San Luis Potosí: pes 90, E of forests; 550—1900 m; southwestern United States (Texas, known only from the Chisos Mountains in Brewster County), northern. Mexico (Chihuahua, Coahuila, Nuevo ^m San Luis Potosí, Sonora, Ta- maulipas) (Fig. P. one umbonata is easily distinguished from related taxa by its persistent indusium with a raised, central umbo. Incomplete collections of sterile leaves of this species, however, are virtually impossible to distinguish from P. nobilis and can be difficult to distinguish from P. auriculata. The range of P. umbonata is generally to the northeast of that of P. nobilis and to the east of that of P. auriculata. The latter taxon can also usually be dis- tinguished by its acroscopically auriculate pinnae, which possess a somewhat denser indument of re- duced, uniseriate scales abaxially. The diploids P. nobilis and P. umbonata have been implicated in the parentage of tetraploid P. auriculata. Representative ANE U.S.A. Texas: Brewster County, damp places on side of Casa Grande, basin of the Chisos Mountains. Warnock 178 (ARIZ, G H, TEX). MEXICO. Chihuahua: Gu: i dre Mountains, fee s.n. rm Coahuila: i12 de Milagro, E side of Sierra de los Guajes, ca. of Hacienda de la Encantada, i 'art 1532 (GH). Nue- vo León: steep, rocky slope near Horsetail Falls, 6 km SW of Villa de Santiago, Clausen 7555 (CU, GH, ) 20.0 mi. Ciudad x Lm S en 563 (E ; Cafion ejas, Sierra Bs ‘0, Sie dental, aie T2 (MICH, UC). Tamaulipas: km 17.6 (11 mi.) SW de Cd. Victoria, carretera 101, Cowan 3728 (TEX ). EXCLUDED TAXA Phanerophlebia aurita Fée, Crypt. Vasc. Brésil 2: 0 . 1. 1873. = Polystichum au- ritum (Fée) ) Yatskie vych. Phanerophlebia caryotidea (Wallich ex Hook. € Grev.) Copel., Gen. Fil. 111. 1947. = Cyrtom- ium caryotideum (Wallich ex Hook. & Grev.) C. Presl. Phanerophlebia caryotidea (Wallich ex Hook. € Grev.) Copel. var. micropteris C. Fl. Madagascar 1: 326. 1958. micropteron (Kunze) Ching. T Jalcata ( (L. £) Copel., 111 = Cyrtomium falcatum ( Chr. ex Tard., = Cyrtomium Gen. Fil. L. f.) C. PUR Phones falcata (L. f.) Copel. var. devexis- capulae (Tag.) Ohwi, Fl. Jap. pte rid. 1957 omium fale atu . Phanerephlebia forte (J. Smith) Cope a Gen. Fil. 111. = Cyrtomium fortunei J. Smith. Tere (J. Smith) Copel. var. cliv- Figure 12. h. Phanerophlebia pumila. —a-g. Silhouettes of sample leaves from different T in a single collection. 9.) Habit. —i. Detail of pinna. —j. Rhizome scale. (a-j all from Yatskievych et al. 85— 196 Annals of the Missouri Botanical Garden icola (Makino) Ohwi, Fl. Jap. Pterid. 69. 1957. »yrtomium fortunei J. Smith. Phanerophlebia ei e Smith) Copel. var. in- termedia (Tag. Fl. Jap. Pterid. 69. 1957. — Cyrtomium Pe J. Smit Phanerophlebia fraxinella (H. C hrist) Copel., ; . 1947. = Cyrtogonellum fraxinellum (H. Christ) Ching. Phanerophlebia hookeriana (C. Presl) Copel., Gen. = Cyrtomium hookerianum (C. Gen. Presl) C. a adulto a) Okuy. ex wi, Fl. Jap. Pterid. 70. 1957. = Cyrtomium macrophyllum (Makino) Tag. Phanerophlebia macrophylla (Makino) Okuy. ex Ohwi var. tukusicola Okuy., Coll. Il. Wild Pl. Jap. 7: 41 + pl. 373, fig. 2. 1960. = Cyrtom- tum macrophyllum (Makino) Tag. var. tukusi- cola (Tag.) Tag Phanerophlebia nephrolepioides (H. Christ) Copel., ven. Fil. 111. 1947. Cyrtomium nephrole- pioides (H. ada Copel. Phanerophlebia semicordata (Sw.) Conz., Fl. Taxon. ; — Cyclopeltis semicordata Sw.) J. Smith. jie ate tachiroana Copel., 1¢ Gen. Fil., 111. — Cyrtomium hookerianum (C. Presl) C. Chr. Phanerophlebia vittata (H. Christ) Copel., Gen. Fil. 111. 1947. Cyrtomium | lonchitoides H. Christ [note that Copeland's use of the name P. vittata probably applied to C. balansae (Christ) C. Chr., but the type of the basionym C. vittatum H. Christ actually — ides (Ching, 1936)]. C. lonchito- Literature Cited | hcm D. S. 1985. Polystichum in Central and South nerica. Pp. 25-26 in D. S. Barrington (editor), The present pisei and taxonomic status of the fern uu The 1984 Botanical Society of America Pteridophyte Section Bm. Amer. Fern. peg J. 75: 22- 98° species and aa ea in tropical Amer ‘rican Pay hum (Dryopteridaceae). Ann. Missou- i Bot. Gard. 305-373. C die. P. & K Nees polystichoid cd Cyrtomium and Polack hie —403. 1982. Morphology of some sporophytes of Arac cages J. Indian Bot. Soc Ching, R. C. 1936. On the genus Cyrtomium Presl. Bull. Chinese Bot. Soc. ae 5-100. 1938. evision of the Chinese and Sikkim- Himalayan Dryopteris with reference to some the neighboring regions. Bull. Fan Mem. Inst. Biol. 8: 275-334 spec les in 195 37. Cyrtomidictyum Ching, a yet little known C hine 'se fern genus. Acta Phytotax. Sin. 6: 255-266 Christensen, C. 1930. The genus Cyrtomium. Amer. Fern J. 20: 41-52. C poe an, R. T. 1949. Ferns p the Monterrey region, Mex- Amer. AM 39: 89- Cope land, E. B. 1947. Gener >ra 1 Filic ‘um, Chronica Botan- ica, Waltham, TANE huse Cotthem, W. R. J. Van. 1970. e classification of stomatal types. Bot. t ^u Soc. 63: 235-2 Daigobo, s. Taxonomical studies on the fern genus P in geen Ryukyu, ye Taiwan. Sci. Rep. Tokyo Bunrika Daigaku, Sect. B 1 gi Diels, L. 1899. Polypodiaceae. Pp. 139- n A. En & K. Prantl (editors), Die Natürlichen OS e ed. 1, pt. 4, vol. 1. W. Engelmann, Leipzig. Dilcher, D. L. 1974. Approac thes to the identific as of angiosperm leaf remains. Bot. Rev. (Lancaster) 40: 1— 157. Dittmer, H. J., E. F. Castetter & O. M. Clark. 1954. The ferns and E allies of New Mexico. Univ. New Mexico Publ. Bot. 6: 1-139, (ibaon, A. C., H. W. Calkin & P. S. Nobel. 1984. Xylem anatomy, water flow, and hydraulic conductance in the fern C yronium eta Amer. J. Bot. 71: 564-574 Haufler, C. H., M. Windham, D. M. Britton & S. J. Robinson. 1985. T loidy and its evolutionary signif- cance in Cystopteris protrusa. Canad. J. Bot. 63: 1855- 3. 186 Holmgren, P. K., H. Holmgren & L. C. Barnett (Edi- tors). 1990. n. herbariorum. Part E The herbaria of the world, ed. 8. Regnum Veg. 120: i-x, 1—693. Humboldt, A., A. Bonpland & C. S. Kunth. 1825. Nova Genera et Species um vol. 7. Librairie Greque- atine-Alle d Ksiobloe h, I. W. & D. S “Correll. 1962. Ferns and fern allies of C oa ps 0. Contr. Texas Res. Found., Bot. Stud. e * —198. Ru. R. C. Moran & A. R. Smith. 1990. Dni nh ‘eae. Pp. 101-144 in K. Kubitzki (editor), The Families and Genera of Vascular Plants. . Pteridophytes and nee (K & P s. Green, volume editors). S Kunze, G. 1844. Die F dungen, Kosei 'icle (Heft) 7. F Manton, I. 1950. Problems of C ytology and Evolution i in the Pte per Cambridge Univ ri Maxon, W. 912. Notes on the Nort of Phane vil Bull. Torrey Bot. Club 39 Mickel, J. T. 79. How to Know n: Ferns M Fern Wm. €. Brown, Dubuque, M. Beitel. 1988. cad bado p Es xaca, Mexico. Mem. New York Bot. Gard. 46: 1— ————, V. agner, Jr. & K. L. Chen. 1966. rou mosome übservslam on the ferns of Mexico. Caryologia 19: 95-102. Kramer, Kl — mu Allies. Mitsuta, S. 1977. Evolution of simple fronds in Cyrtom- ium—A_ pattern morphological study. Acta Phytotax. Geobot. 28: 131-142. Mitui, K. 1973. genus Polystichum by se anning electron microscopy. Bull. Nippon Dental Coll., Gen. ie 2: 103-123. Moore, T. Index Filic ‘um a Genera). William Pamplin, Lo wa R. C. 1986. n ple al fe rn genus Olfersia. ner. Fern J. 76: 161-178. 1987. Monograph of the neotropical fern jm Poly botrya p eae). Bull. Ilinois Nat. Hist. urv. 34: + 1-1: Characteristics of the pe rispore of the , Synopsis of the Volume 83, Number 2 1996 Yatskievych 197 Revision of Phanerophlebia Morton, C. V. 1957. Observations on cultivated ferns, II. The proper a name of the holly fern. Amer. Fern . 47:5 Presl, C. B. 1836. Tentamen Pteridographiae, seu Genera Filicacearum Praesertim Juxta Venarum Decursum et Distributionem Exposita. Filiorum Theophili Haase, rague. Reeves, T 1978. [Miscellaneous e ee gee report 556 in A. Lóve (editor), IOPB ae aie number ar reports LIX. Taxon 27: 534 Shing, 965. A taxonomic sal ack of the genus rt Presl. Acta Phytotax. Sin. 12(suppl.): 148. Smith, A. R. 1981. adore, Pp. 1-370 in D. E. Breedlove (editor), ngs of Chiapas, Part 2. California Academy of Sciences, San Francisco. 1986. Revision of the neotropical fern genus Cyclodiam. ga Fern J. 76: 56-98. & Mickel. 1977. Chromosome counts for Brittonia 29: 391-398. An arrangement and definition of the gen affinities of Suh exic " Ten rns. bes J. era of in with observations on the : genus. J. Bot. (Hooker) 4: 38-70, 147—198. Stein, D. B., C. Yatskievych € G. J. Gastony. 1989. Chlo- roplast DNA evolution and phylogen "Y e some polysti- choid ferns. Biochem. Syst. Ecol. 101. Stolze, R. G. 1981. Ferns and fern | " Guatemala. Part II. Polypodiaceae. Fieldiana Bot., n.s. 6: i-1x = ryon, A. F. & B. Lugardon. 1991. rid of the Pteri- a oe su Re rare IN ew Tryon, R. M. & R. G. Stolze. 1991. Pieridophyia o of Peru. Part IV. E ee 'ridaceae. Fieldiana Bot., 27: i- iil, i . F. Tryon. 1982. Ferns and Allied Plants, With 2 Reference to Tropical America. Springer- Verlag, New York. Tsai, J.-L. & W.-C. Shieh. 1985. vey of "e fern family Aspidiaceae (sensu Copeland) in Taiwan. J. ET ¿ngin. (Taiwan) 22: 121-144. Underwood, L. M. 1899. American ferns—ll. The genus Phaneophha, Bull. Torrey Bot. Club 26: 205-215. Vareschi, V. ). Helechos. Vol. l(l, in T. Lasser oy boc p Venezuela. Instituto Botánico, ‘Merida, A cytotaxonomical sur- Viane, R. 1. 1986. Taxonomical significance of the leaf indument in Dryopteris (Pteridophyta): I. Some North American, Macronesian, and European taxa. Pl. Syst Evol. 153: 77-10 Wagner, D. H. 1976. Taxonomic Investigation of the Ge- nus Polystichum in Western North America. Ph.D. Dis- sertation, Washington State University, Pullman, Wash- ington. 1979. Systematics of Potene in western North America os 2 Mexico. Pteridologia 1: 1—64 Wagner, W 19 A biosystematic survey of Unit- ed States fe ide eon abstract. Amer. Fern J. 53: 1-1 1979. Reticulate veins in the systematics of modern ferns. Taxon 28 1985. Morphological variation and evolution in E Pp. 22-23 in D ria (editor), The present evolutionary and taxonomic status of the fern genus Polystichum: The 1984 Botanical Soc iety of America Pteridophyte Section Symposium. Amer. Fern. 22-28 87-95. E J. 75: 22-% & D. M. Johnson. 1983. Trophopod, a commonly — storage structure of potential systematic val- 3-269. ue in ferns. Taxon 32: 26 D. Farrar & B. McAlpin. 1970. Pte ape of the Highlands ne s cal Station area, oo Ap- Mitchell Sci. Soc. 86: 1-27. new io in South American Polystic un Amer. Fern J. 79: 26-27 —— ——. 1990. Systematics of the Fern Genus Cyrtomium in he New World. Ph.D. Dissertation, Indiana Univer- sity, E India 1997 vations in the fern genus Phanero- phlebia. Novon 2 2; : 445-446. ——. 1993. Antheridiogen response in Phane e bia and related fern 30— ——— & €. J. Gastony. 1987. Hybridization and s ploidy in the fern genus Phanerophlebia. Amer. J. Bot. 74: 715. [Abstract.] . D. B. Stein & G. J. Gastony. 1988. Chloroplast DNA en and systematics T iT bia and related A genera. Proc. Natl. Acad. Sci. U.S.A. 8 2589-256 < = un 4 x ® < mk = = 5 ~ m genera. mer. 'ern Appendix 1. Exsiccatae. A. Taxa of Phanerophlebia accepted (1) Phanerophlebia auriculata Underwood 2) P. gastonyi Yatskievych 3) P. haitiensis C. Christe (4) E e pci (Humboldt & Bonpland ex Willd- now) J. Smith (5) P macrosora (Baker) Underwood (6) P. nobilis (Schlechtendal & Chamisso) C. Presl (6a) var. nobilis (6b) var. re apo ad (E. Fournier) Yatskievych (7) P. pumila (M s & Galeotti) Fée (8 ) P. umbonata U m nde B. Collections by collector an number (or collection date where no number was indicate Anonymous (det. C. Ku nae) (Ge y (herb. Jeanpert) with- out date or number (6b); 000806 (Ga): on 26 Apr. 1917 (6b); on : lr 1919 (1). ; [= Arsène 6287] (8) Arsène 6291] (8). Aco M E 281 (6b). Aguirre C. on 4 Oct. vin (8). Alfaro in Feb. 1902 [herb. J. Donnell- pene 8074] (4 103 is rb. J. Donnell-Smith 8074] (4). A 48 (4). es derson & Laskowski 4363 (6a). Min ce (4); 2861 Sa E 534 (5 ). Arreguín 435 (6a). Arséne on : 2 1 (6a); on 7 July 1911 (8); Me 16a); 5415 (6a. b d lesion 5961 (6a, b intermediate). Ave ne R. 00063 (6b). Ballesteros & Ballesteros 371 (6a). 403 (6a). Barkley 14579A (8). Barkley, Webster (G.) & Rowell 7130 (8). Barrington 1211 (5). Bartlett 10095 (8) 10391 (8); 10406 (8); 10931 (8). Batalla & Bravo on 11 Mar. 1933 (6a). Beaman 3056 (3-vel. aff.). Benson 10947 (1). Be E & Lee 65 (1). Bocge 2737 (6a). Bonilla B. F-4322 (6b). Bonpland without date or loc m (4). Botteri a (6b). ourgeau on 20 May 1866 (6b): 45 (6b); 916 (6a); 917 (6a); 104 (6a); 1648 (6b): 2276 (6b); 2341 (6b); 2348 (6b): 2349 (6b). Boutin & Kimnach 3101 (6a). Boza without date or number (4). Brade (A.) 182 (4). Brade (A.) & Brade (C.) on 17 Dec. 1909 [herb. E. Rosenstock 161] (4); 9573 . Breckon & Christman 622 (7). Breedlove 7368 (6b) 10768 (6b); | 5242 (5); 15299 (5); 24996 (4); 25355 (6b): 25759 (6b); 26262 (5); 26778 (4): 26840 (4); 27697 (6b); 32942 (4); 33657 (6b); 34400 (4); 34615 (6b); 34692 (4): 34698A (4); 38770 (6b); 41708 (4); 41713 (5); 42564 (4): 2 — 198 Annals of the Missouri Botanical Garden 48670 (4); 53211 (6b). Breedlove & Almeda 56931 (6b). 59 7). Breedlove & Dress- 21748 (6b); 22330 (6b); 31320 (6b); 31433 (6b); 31495 (2); 31816 (7); 31881 (5); 32310 (6b); 32744 (6b). Breed- love & Strother 46731 (4). Breedlove & Thorne 21363 (6b); esti (2). Brenes in 1912 (4). Butterwick 53 (6a). jan & Osborn 2530 (1). Bye 6989 (1); 7094 (6a); c Caballero 162 (4). Camacho G. 10-58 (6a). Camp 2561 44 (6a). ery 3 (6a). Castillo C. € Vázquez 1360 7 (6b). Castillo C ; 1789 (4). Chase 7312 (6 4). Cházaro B. & € astillo Ch. 6789 (5). nd d. Chrysler & Roever 5611 (4). Cisneros 1649 Clausen 7555 9 7570 (8). Clewell & Hazlett 3975 ni S € 9 (8). Conant 726 (2). Conant, Dorante, s«& EE 727 (6b). Contreras 4956 (6 ii nid zatti ti (C) 2133 (5). Conzatti (C.), Conzatti (H.) & (T.) 2363 (5). Consti pa & Gómez (T.) 3482 (5). Die (C.) & González 1138 (6b). Cooper in 1886(?) (5). Cope- £ rrell & Sae, 23043 (1); 23241 (1). Correll & Johnson (L) 21772 (1). Cory 26488 (8). Cosson eid (6a). ae A 7778 (6b). Cowan 3728 (8). Cowan & Valdés 86 (8). Croat 39411 (6b); 39484 (6b); 39489 (6b); ae (6b); 40940 (4); 41442 (6b); 41474 See 44033 (6b); Croat 46105 (6b); 47628 (4); 48839 y 67734 (4). Croat & Hannon 65955 (1). I w & d 2059 (1). Darrow, Gould. Pultz & Phillipe (W) 2556 enport without date or number (6a). i Davis on 7 Apr. 1946 (8). Deaver 4788 (1). del Campo on 11 Mar. 1933 (6a); on 20 May 1933 (6a). Dorfler 190 (4). Dressler 1613 (4); 1947 (8). Drushel 9538 (8); 9539 (8). Dunn ee & Dunn (Don) 19082 (6b). Edwards 364 (8). Ehrenberg 868 (6a). Ekman 3119 (3); 7793 ( Fendler 233 (4). Ji i on 23 May 1988 (1). F nández & Zamudio 992 . Ferris in 1902 (1); 264-08 (1); 294 (1). Fink 62 che Fink 66 (6b). Fisher on 9 Aug 1924 (6a, b intermediate); on 14 Aug. 1926 (6b); 35305 (6b). Fournier 64 (6b). Fryxell & Aes 3588 (6b). Funck 211 (4). Galeotti 6251 (7); 6343 (6a). Gallegos H. 370 (6a). Gentry 3660 (6a); 7988 (6a); 8119 (8). Gereau & Martin 2011 (7). Ghiesbreght without date or number (7); 209 (7); 414 (4). Gilbert 34 (8). Gimate L. 975 (5). Gómez (L.) 360 (5); 528 (4); 3441 (5); 22207 (4). Gómez (L.), Chacón (L.), Chacón (R.) € Herrera 21424 (4); 21432 (5); 21489 (5); 21966 (4). Gómez H. on 3 Oct. 1982 (6a). Gómez-Pompa & Riba 381 (6b). González Q. 1641 (6b); 2419 (6a). Good- ding 392-45 (1); 162-50 (1); 138-52 (1); 57-53 (1); 9 61 (1); 3070 (1); eds a A 6146 (1). Goodding & Hinckley F-3-39 (1); F-9-39 (1). Gould 4347 (1). Gould, Darrow & el 2788 us Grant 10799 (4). Gra- yum, Poveda & Gómez-Laurito 823: Grayum & Schatz 5166 (4). Guillemin without date or number (6b); in 1866 (6a Hahn 172 (6a). Hallberg 848 (7); 1391 E 1456 (6b). eere. Stockwell & Aiello 982 (4). Hammel & Trainer 13825 (4). Hartman 578 (1). Harvey 996 (8). "Hellwig 361 (6b); ES (7). Hernández M. 3337 (6b). Hernández M. (R.), Cortés € Hernández M. (I.) 5916 (4). Hernández M. = — iim, g a = » Hernández V. 4251 (5). Herrera 3611 (5). Heyde & Lu n Apr. 1892 [herb. J. Donnell-Smith 32: 59] (6b); 224 (5). Hinton 403 (6b); 1335 (6a); 3544 (6a); 12428 (6a mixed collection); 15935 (7); 16913 (8); 17511 (8); in (8). House & Andino 961 (4). Hunnewell on 20 Jan. 1941 (5); 14599 (5); 14600 (5) Jiménez 78-21 (6a). Johnson 699 (6b). Johnston (1.) & Müller (C.) 1362 e Johnston (M.) 12777 (6a); 12782 (6a). Johnston (M.), Wendt & Chiang C. 10728 (1). Jones on 23 Sep. He [loc. zn (1); on 23 Sep. 1903[loc. b] (8). s 1911 ae in 1827 (6a). Kearney € Peebles 14485 (1). Kellerman 5774 (5). €: in Jan: 1902 e Renan in oy Kaba on 1 26 Sen: 1882 (6b). Killip 5468 (5); 5 er Knapp 1462 (4). Knobloch 1967 (8); 7011 (1). Koch 7670 (6a). Kruckeberg 4836 (8). Lacás 154 (8); 247 (8); 371 (8); 486 (8). Lara & Cha- varria 28 (4). Laughlin 96 (6b); 549 (6b). Leavenworth 93 (8); 801 (8). Leding £ 1). Lee, Berkman & Tharp 46193 (1). Leland on 26 Sep. 1896 (8). Lellinger & White 1679 (5). Lemmon in Aug. 1882 (1); on 8 Aug. 1882 (1); on 12 Aug. 1882 (1); in Sep. 1882 (1); 321 (6b). León 17 (4). Leonard & Mickel 4122 (6a). Liebmann in Aug. (6b). Linden in 1838 (7); in Feb.(?) 1838 (7); 1551 (7); 1552 UA : Ant (M.), Lint (H.) & Haskell 1005 (1). Little & Sharp 2 (7). Loomis & Peebles 5413 A Lorence & Cedillo a d 2 (6a). Lorence, Martin & Cedillo T. 3268 (6a). Lyonnet 87 (6a); 765 (6a); 1481 (6a); 1533 (6a); 1633 (6a); 1926 (6a); 3240 (6a). Madison 670 (5). sg on 27 June aie (4). Marsh 236 (8); 268 (8); 1871 (8). Martin, Moore & Titley s.n. 6a). Matuda 193 (6b); Eo (6b); 1365 te 1893 (5); 968 (5); 4688 (6a); 5230 (5); 18710 (6a); 18783 (6a); 26457 (6a); 28263 (6a); 28304 (6a); 30948 (6a). Maury 6752 (6b). Maxon 4935 (4); 527: RA ; 14238 (6a); 20437 (6a); 22314 (5); (6a). MN 231A (6a); 232 (6a). Mexia 1558 (6a). ee ogers 2581 (8); 2935 (8). Mickel 563 (8); 5377 (7); 6044 (2); 6182 (2); 7248 (6b). Mickel & Hellwig 3786A (5); 3786a (6a); 3786B (2); 4156 (5). Mickel & Leonard (5); 5384 (7); 5384A (7); 5384B (7). Mickel & Par- due 7047 (6a). Mohr 184 (6b). Montgomery & Root 7339C (8). Moore & Wood 4427 (6a). Morales 1953 (5). Moran m: (4); 4167 (4); Di (4); 5667 (4); 5706 (3). Morelos 6a). des "no 4 (4). Moritz 10955 (4). Moya R 34 oy iens er ( d (6b); d (8). Müller (C.) & Müll- er (M.) 3 . Münch in 1901 (6b); in Oct. 1901A (7); in Oct. IE (7); 61 (5); 74 a Narave F. 319 (5); 370 (5). Nava & Cruz P. 14 (6a). Nee & Taylor 26813 (6b). Nelson (A.) & Nelson ne 1206 (1). Nelson (E.) 2222 (7). Nevling & Gómez-Pompa 2308 (6b). Nicolas on 4 Dec. 1910 (6a). Niles 447 (1). Ojeda E al So Orzel 79-21 (8). Ownbey (G.) & Own- bey (F.) 1 Pa acon ha 2901 (2). Palma G. 67 (4). Palmer 450 arry, sec slow, Wright (C.) € Schott on 5 Mar. (1). Pennell 16852 (8). Perkins & Hall 3256 (8); 3294 (8). Peterson J-1751 (6a). Phillips (E.) 546 (1); 725 (1). Phil- lips (W.) & Reynolds 2946 (1). Pineda R. 607 PU Pittier 207 (4); 2982 (4). Pringle on 21 July 1884 (1); 3 (1); 831 (1); 1982 (8); 3403 (8); 5585 (6b); 13739 (8). Proctor 25075 (4); 25446 (6b). Puig 4105 (8); 4918 (6b). Purpus in June 1931 (6b); 1595 (6a); 2454 (6a); 2933 (6b); 2993’ (6b); 6427 (6b); 15414 (6b); 16024 (6b); 16476 (6b); 16578 (6b). Volume 83, Number 2 Yatskievych 199 1996 Revision of Phanerophlebia Rascon without date or number (6b). Rebolledo V. 291 — C. (R. s C. (L.) & Martínez 7168 (6b). Torres R. (5). Reeves (B.) et al. 88-1 (1). Reeves (T.) R1202 (1). Reichenbacher 1135 (1). Reichenbacher & Van Devender (T.) 759 (1). Reineck in July 1899 (6a); in Sep. 1899 (6a); on 14 Sep. 1899 (6a). Rhoads E Mar. 1899 (6b). Riba without date or number (6b); 66 (6a); 1002 (6a). Riba, Tryon (R.) & Tryon (A.) 361 (6b). Riskind 1592 (1). Rivera 00071 (8). Rollins & Tryon (R.) 586 (5). Rowell & Barkley 16M595 (8 . ud 583 (8); (8). Ruiz O. in Aug. 1947 (6a). Runyon 720 (8); 975 (8); 985 (8). Rzedowski 679 (6b); 10056 (6b); 10571 (6b); 18531 (5); 19864. (6a); 26425 (6a); 27212 (6a); 27229 (6a); 34234 (6a). Saborio 36 (4). Salinas, Rowell & Barkley 16M575 (8). Salvin 113 (4); without date or number (4, 5 mixed col- lection). Sánchez (J.) 32 (6a). Sánchez M. 164 (6a). Sán- chez S. 445 (6b). Scamman 7087 (5). Schaffner in 1850— 1855 (6b); in Aug. 1875 (6a); in 1876 (6a); 67 (6a); 82 (6a); 277 (6a); 461 (6b). Schiede in Oct. 1828 (6a). Seaton 49 (6b). Seibert 187 (4). Seigler (D.), Bohnstedt & Seigler (E.) DS-2550 (8). Sei Sessé, Mocifio & Mal- donado 3853 (6a). Shreve 5434 (1). Skutch 742 a Smith (Alan) 510 (6a); 515 (6a). Smith (Austin) F24 (4); 48/134 (4); 1463 (4). Smith (C.) on 24 Dec. 1894 (6b); 2187 (6b): 2202 (6b). Smith (R.) M329 (8). Soxman (G.) € Soxman (F.) 362 (1). Spence 27 (6b). Sperling 4958 (2); 4994 (2). Sperry 178 (8). Standley 10468 (4); 32649 (4); 34531 (4 X 5 hybrid); 57840 (5); 60049 (5); 61300 (4); 6: : 67247 (4); 84842 (4); 86672 (4); 87087 (4); 90066 (6b). Stanford, Lauber & Taylor 2085 (8). Stanford, Retherford € Northeraft 1069 (8). Stewart 523 (8); 1532 (8); 1533 (8); 2198 (1). Steyermark 36243 (5); 36871 (4); cd is (4); 43709 (5); 46717 (5); 47248 (5); 50036 (4); 50211 (7) 5130 (4). Stork 1411 (4); 1558 (4); 1732 (4): a "o Studhalter 308 (1 Taylor 364 (8). fend orio L., Ramamoorthy € Lafrankie 3655 (6b). Tharp on 4 Sep. 1915 (8); 1817 (8). Thornber, Goodding & Nelson (A.) on 16 Mar. 1935 (1). Thorne & Lathrop 41308 (6b); 41784 (6b). Ton 1990 (5); 4840 (6b); 5320 (6b); 5686 (6b); 5889 (6b). Tonduz 11930 (5). Torres = 109 a a 61 (4). Tucker 1288 (4). van der Werff & Herrera 7111 (4). Van Devender (T.) on 6 Feb. 1978 (1); on 13 Feb. 1977 (1): on 16 Aug. 1978 (8). Van Devender (T.) & M 1977 (1). Van Devender (T.) & Van Devender (M.) on 28 Mar. 1976 r 6a); 2694 (6a); 7967 (6b); 9265 (2); 13027 (6b) 15836 (6b); 16261 (6b). Verlet in (vn 1851 (8). Virlet 28 (6a). von Rozyns ski 415 (8); € Türchheim in Feb. 1856 (6b); in Sep. 1885 [herb. J. Don- nell- ma 768] (6b); in Sep. 1886 [herb. J. Donnell- en 1051] (4); in 1907 (6b). Vovides, Rees & Vázquez (T.) 6 (6b). Warnock 178 (8); 21779 (1). Waterfall 6635 (1). Waters 737 (1). Weber 6046 (5). Webster (G.) & Preston 2878 (8). Webster. (M.) 177 (8). Wendt, Lott & pror 1946 (1). 5 (5). White & Chatters 27 (8); 229 (8). Whi- rehouse on 11 Nov. 1931 (1); on 12 Sep. 19: E 3). Whitson 598 (8). Wiggins 7139 (D; (6b). Mes dh Molin: illiams (T.) 2: 3058 (5): 23802 (4). m 0059B Q): 0109B (1). o - 14 D: on 15 May 18 92 (1); on 17 Sep 3 (0); on 5 E 1902 (1); 1491 i » Wo- ronow 3020 (6b). Worihitigten 7679 (1). Wright (A. H.) & ht (A. A.) on 8 July 1925 (1). Wynd & Miller (C.) 349 (8). Yatskievych (G.) 78-361 (1); 83-161 (1). Yatskievych (G.) € Forbes 82-211 (1). Yatskievych (G.) € Gastony 86-250 (8); 86-329 (5); 89-218 (8). Yatskievych (G.) € González L. 85-209 (7). Votskiovyeh (G.), Hevly & Wind- ham 81-313 (1). Yatskievych (G.) & & McC rary 9: 5-05 (8); 86-13 (5); 86-30 (5); 2 (4); y 5). Yatskievych (G.), Crary & Worthington nal OO 84—68(1). Yatskievych (G), Ranker, González L., (G.) € Starr (C.) 85-139 (7); 85-182 (2); 85-186 pe 85-211 (6a). Yatskievych (G.) & Windham 85-296 (1). Yatskievych (G.), Windham & Ranker 83-299 (8); 83-353 (6b); 84-04 (8). Yatskievych (G.), Leger a & Hallberg 83—467 (5). Yatskievych (G La odhar livan 83-10 (1). Yatskievych (G.) & (8); 83-128 (6b); 83-158 (6b). Yatski a G : skievych (U.) 82-273 (1). Young on 4 Sep. oe "n Zolá B. 00710 (6b); 00742 (6b). — — — - 1 cER REVISION DE LAS ESPECIES Fernando O. Zuloaga? and AMERICANAS DE PANICUM Osvaldo Morrone? SUBGENERO PANICUM SECCIÓN PANICUM (POACEAE: PANICOIDEAE: PANICEAE)! RESUMEN En el presente tratamiento se realiza una revisión de las especies americanas de la sección Panicum del género Panicum. Se estudiaron treinta y un d ies de esta sección, inc yendo el análisis exomofológico y anatómico de las mismas. La sección Panicum se caracteriza por incluir especies anuales y perennes, que habitan en lugares abiertos y secos; las pr ntas poseen lígulas eg 'eo-ciliadas, láminas oblongo-lanceoladas a filiformes, inflorescencias axas as, con espiguillas solitarias dispuestas sobre pedicelos usualmente largos; las espiguillas son ovoides a largamente elipsoides, glabras, con la gluma inferior (4—)%4—34(—¥,) del largo de la espiguilla, gluma superior y lemma inferior a ig rn pálea inferior tan larga como el antecio pes a E ida o ausente, flor inferior ausente, antecio superio ve un número básico de cromosomas de x esde el punto de vista histofoliar, dos patrones ara os pea fueron hallados en el transcorte de las especies iis la sección. Se ane uten las relaciones de la sección estudiada con otras del subgénero Panicum, como así también con géneros afines de la tribu Paniceae. Se inc luye una clave de las especies analizadas, una descripción anatómica de la sección, fotografías de diez especies, descripciones morfológicas, ilustraciones y mapas de distribución de los diferentes taxones. = ^ = ABSTRACT Panicum subg. Panicum sect. Panicum is revised. Thirty-one American species are treated in this work, in which exomorphologic 'al and anatomical characters are considered. Section Panicum is defined mainly as an annual or pe- species of open and dry places, with membranous-ciliate ligules, leaves oblong-lanceolate to filiform, inflores- cences lax and diffuse, with spikelets solitary on pedicels that are usually long; spikelets ovoid to long ellipsoid, glabrous, and the lower glume (A )4-4(-¥,) the lent of the spikelet, upper glume and lower lemma ier lower palea as long as the upper aa ‘ium to reduced or absent, lower flower absent, din nthecium indurate, and a basic chromosome number of x = 9. Two anatomical dE were found in the trar sverse section of a's leaves. Possible js hips of Panicum a other sections of the typical subgenus, as well w er genera of the Paniceae, are discussed. A key to the species as well as an anatomical desc 'ription of the section ind mide cad of ten species, mache 'al descriptions of the species, illustrations, and distribution maps are presented. = c = E] El género Panicum L. es uno de los más grandes — (1878) en la serie Paniculatae y Hackel (1887) en de la familia Poaceae, contando con aproximada- la sección Eupanicum. Nash (1903) reconoce Ca- mente 470 especies distribuidas en regiones cáli- — pillaria, sin aclarar su rango taxonómico, incluyen- das y templadas del globo (Clayton € Renvoize, do en el mismo especies anuales o perennes del entro de este género la sección típica es sur de los Estados Unidos; trata dentro del mismo ni, hallándose en América, Europa, Afri- a P. capillare L., P. capillarioides Vasey, P. cogna- ca, Asia y Oceanía. Las especies de esta sección tum Schult. (= Digitaria cognata (Schult.) Pilg.), fueron tratadas por diferentes autores bajo varios P. diffusum Sw., P. filipes Scribn., P. flexile (Gatt.) nombres. Así Nees (1829) ubica a especies de este Scribn., P. hallii Vasey, P. miliaceum L., P. phila- grupo en la sección Effusa. Steudel (1855) las agru- delphicum Bernh. ex Trin., P. proliferum Lam. (= pa en las secciones Miliaria y Virgata, Bentham P. dichotomiflorum Michx.) y P. stenodes Griseb. s autores desean expresar su j des pde a Gerrit Davidse por su ayuda en la confección de parte del presente manuscrito y a Vladimiro Dudás y Andrés Bestard por la realización de las ilustraciones y el armado de los mapas. El trabajo de neg llevado a cabo en Colombia, Venezuela y Brasil fue posible gracias a subsidios de la National Geographic Society, # 3964-88, # 4594-91, 11-91, a quien los autores tambien desean a su a imiento. ? Instituto de Botánica Darwinion, retia 200 Casilla de Correo 22, San Isidro 1642, Argenti ANN. Missouni Bor. GARD. 83: 200—280. 1996. Volume 83, Number 2 Zuloaga & Morrone 201 Panicum Subg. Panicum Secc. Panicum (especie de la sección Tenera (Hitche. & Chase) Hitchcock y (1910) consideran, en su revisión de las especies ilg). Más recientemente, Chase de Panicum de América del Norte, a las especies en los grupos Capillaria y Diffusa, agrupando en el primero a los taxones anuales y en el segundo a las especies perennes: consideran a P. barbipulvinatum Nash, P. capillare, P. cayennense Lam. (especie de a sección Rudgeana (Hitche. € Chase) Zuloaga (Zuloaga, 1987b)), & Chase, exile, P. gattingeri Nash, P. hirticaule J. Presl, P. decolorans Hitche. P. miliaceum, P. pampinosum Hitchc. & Chase, P. parcum Hitche. & Chase, P. philadelphicum, P. so- norum Beal y P. stramineum Hitche. & Chase den- tro de Capillaria y las siguientes especies en Dif- fusa: P. capillarioides, P. diffusum, P. filipes, ghiesbreghtii E. Fourn., P. hallii, P. hirsutum Sw. y P. lepidulum Hitche. € Chase. Fernald (1919) real- iza un estudio de grupo Capillaria para Nueva In- glaterra, considerando en el mismo a P. capillare, P. philadelphicum y P. tuckermanii Fernald. Stapf (1920) ubica las especies en la sección Miliaceae, tratando 11 especies para la flora de Africa tropi- cal. Pilger (1931, 1940) trata a las especies dentro de la sección Virgata Hitche. & Chase e incluye dentro de la misma a las secciones Diffusa de Hitchcock & Chase, Hiantia y Miliacea de Stapf. Henrard (1941) realiza una sinopsis del grupo Di- ffusa e incluye en el mismo a P. bergii Arechav., P. campestre Nees ex Trin. (especie de la sección Rud- geana, Zuloaga, 1987b), P. capillarioides, P. ghies- breghtii, P. hirsutum, P. peladoense Henrard, P. pil- comayense Hack. (= P bergii), P. quadriglume (Doll) Hitche. Schult.; caracteriza a este grupo por incluir espe- cies perennes, cespitosas y con panojas difusas. Fairbrothers (1953, 1954) revisa las especies anua- les del grupo y reconoce como especies válidas a P. capillare, P. cayennense, P. decolorans, P. flexile, P. hillmanii Swallen, P. hirticaule, P. litophilum Swallen, P. miliaceum, P. parcum, P. philadelphi- cum (con tres subespecies) y P. stramineum. Swal- y Ja especie asiática P. trypheron len (1943) considera para Panamá a las siguientes especies: P. ghiesbreghtii, P. hirsutum y P. hirticau- le. Swallen (1955) incluye en su tratamiento de las gramíneas de Guatemala a P. furvum Swallen, P. hirsutum, P. hirticaule, P. lepidulum, P. pampino- sum y P. parcum. Hsu (1965) ubica a diversas es- pecies de estos grupos en la sección Panicum. Wal- ler (1976) estudió las especies de la sección Diffusa que crecen en América del Norte e incluyó en su tratamiento a P. capillarioides, P. diffusum, P. fur- vum, P. ghiesbreghtii, P. hallii var. filipes (Scribn.) F. R. Waller, P. hallii var. hallii, P. hirsutum, P. lepidulum, P. pilcomayense y P. tamaulipense F. R. Waller & Morden. McVaugh (1983) describe en la Flora Novo-Galiciana a P. hirsutum, P. hirticaule, P. lepidulum, P. parcum y P. stramineum (con 3 iedades: var. típica, miliaceum e hirticaule). Zu- loaga (1987a) efectúa una clasificación infragenéri- género, de las especies americanas del ~ considerando dentro de la sección típica a Capi- llaria y Diffusa e incluyendo en la misma a 22 es- pecies. Davidse (1994) trata en la Flora Mesoame- ricana a. P. furvum, P. ghiesbreghti, P. hirsutum, P. hirticaule, P. hispidifolium, P. parcum y P. stra- mineum. Tovar (1993), al describir las especies de gramíneas del Perú cita a P. hirticaule, P. quadri- glume y P. stramineum. El presente trabajo tiene por objetivo efectuar un estudio taxonómico y anatómico de las especies americanas de la sección Panicum, dilucidar las especies que integran la misma, discutir las rela- ciones específicas y de la sección con otros taxones infragenéricos de Panic um. MATERIALES Y MÉTODOS Para el estudio histofoliar se obtuvieron cortes — transversales y epidermis a la altura del tercio me- dio de la penúltima lámina de la innovación fértil. Se utilizó material proveniente de ejemplares de herbario, previamente tratado con Contrad 70 (Schmid € Turner, 1977) durante 24 a 48 hs a 20°C, o material fresco fijado en FAA. Los cortes transversales se hicieron a mano alzada, previo tra- tamiento con HF al 5% durante 24 hs. Para la ob- tención de las epidermis se siguió el método de Metcalfe (1960). Los cortes fueron coloreados con azul de meti- leno al 1% y con safranina-alcian blue y montados en ee y con safranina al 1% en alcohol 80° glicerina. Para la observación de las células clorenqui- máticas se realizaron macerados siguiendo el mé- todo de Jeffrey (Sass, 1940). de los cuerpos de sílice y células suberosas se utili- Para la identificación z6 respectivamente fenol líquido (Metcalfe, 1960) y Sudán III (Sass, 1940). plástidos de almidón, y su distribución, se realizó La determinación de los mediante unas gotas de solución iodo-iodurada (Sass, 1940) ara las descripciones histofoliares se adoptó la terminología propuesta por Ellis (1976, 1979). Las observaciones anatómicas fueron hechas con un microscopio fotónico Wild M20 con cámara de dibujo. Las disecciones fueron estudiadas con un microscopio estereoscópico Wild M5 con cámara de dibujo El estudio exomorfológico fue realizado sobre la 202 Annals of the Missouri Botanical Garden base de materiales pertenecientes a los siguientes UB E USM, UTMC, VEN, W, WIS, um SM a las siglas que figuran en Holmgren et al. (1990). Dentro del ma- terial examinado de cada especie sólo se citaron ejemplares representativos de cada país; una lista completa de los especímenes estudiados se en- cuentra ordenada alfabéticamente por coleccionista al final del texto (Apéndice). Las fotomicrografías fueron tomadas con un equipo iip o Nikon FXA, con cámara foto- gráfica )B2 35 mm, y la película utilizada Ko- dak n de 100 AS Para la obtención de fotomicrografías de epi- dermis abaxiales de la lemma y pálea del antecio superior se empleó un microscopio electrónico de barrido Jeol JSM-25 SII, perteneciente a la Facul- tad de Odontología (Universidad Nacional de Bue- nos Aires, Argentina). Con un asterisco (*) se sefialan en la cita de materiales examinados los ejemplares empleados en el estudio histofoliar. Con dos asteriscos (**) se indican aquellos especímenes utilizados en el aná- lisis de la lemma y pálea superior. Se presentan fotomicrografías MEB (Figs. 1, 2) en la descripción de la morfología; mapas de dis- tribución (Figs. 3-11) en Distribución y Ecologia: transcortes (Figs. 12-14) en Anatomía Foliar y Dis- cusión; e ilustraciones de las especies de Panicum secc. Panicum (Figs. 15-28) en el Tratamiento Ta- xonómico. MORFOLOGÍA HÁBITO Dentro de esta sección se incluyen especies anuales y perennes. El tamaño de las plantas es variable, desde especies de 2-8 cm de alto, como P. mohavense Reeder (Fig. 25G—N), hasta otras que alcanzan aproximadamente los 2-3 m de altura, como por ejemplo P. hirsutum (Fig. 21). Todas las especies son cespitosas, con cafias simples o ramifi- cadas en los nudos basales y medios, excepcional- mente con ramificaciones libres en la porción su- perior de las cañas en P. aquarum Zuloaga & Morrone (Fig. 16); las cafias son multinodes, con entrenudos muy cortos hacia la base y alargados en las ramas floríferas o paucinodes, con 1-2 nudos en P mohavense; los entrenudos son cilíndricos, huecos y pilosos a glabros. Los nudos son normal- mente comprimidos, castaños a oscuros y con pi- losidad variable. Las vainas son abiertas y estriadas y varían desde glabras a densamente pilosas con pilosidad variable, con pelos tuberculados o sin los mismos (urticantes en P. hirsutum) o glaucas como en P. hallii var. hallii, P. hallii var. filipes y P. ta- maulipense; excepcionalmente las vainas son es- ponjosas, con aerénquima, en P. aquarum. Las ligu- las son membranáceas en la porción basal y corta a largamente pestañosas hacia la porción distal. Las láminas son planas, o con los bordes involutos, y varían en forma desde oblongo-lanceoladas, lan- ceoladas, linear-lanceoladas, lineares a filiformes; la pilosidad en las láminas es variable, desde hís- pidas o hirsutas, con pelos tuberculados, a glabras, siendo los bordes escabrosos, los basales con pelos tuberculados o sin los mismos. En P hallii var. hallii, P. hallii var. filipes y P. tamaulipense las lá- minas inferiores a la senescencia persisten enrol- ladas en la base de la planta. La persistencia de hojas basales enrolladas fue previamente citada en Panicum mystasipum Zuloaga & Morrone (Zuloaga & Morrone, 1992); estos autores sugieren que po- siblemente este carácter se encuentre relacionado con la protección de las yemas a factores ambien- tales externos. INFLORESCENCIAS Las especies de esta sección poseen panojas ter- minales o terminales y axilares. El primer caso se da en las especies perennes, las que poseen una caña florífera con una inflorescencia terminal. En las especies anuales, por el contrario, se encuen- tran inflorescencias terminales en las cañas prin- cipales y de las ramificaciones basales salen a su vez inflorescencias axilares. Excepcionalmente se observa un patrón complejo de ramificación en la inflorescencia de P. aquarum (Zuloaga & Morrone, 1991) Las inflorescencias son laxas y difusas, con es- piguillas solitarias, en la mayoría de las especies, a espiciformes a subespiciformes, con las ramifi- caciones adpresas al eje principal, en P. decolorans, P. chasei Roseng., B. R. Arrill. & Izag., P. magnis- picula Zuloaga, Morrone & Valls (Fig. 24) y P. pam- pinosum. Las ramificaciones inferiores de la inflo- rescencia son verticiladas en P. bergii (Fig. 18) y P. hillmanii (ocasionalmente verticiladas en ejem- plares de P. capillare) y alternas a subopuestas en el resto de los taxones. Además, en P. bergii y P. capillare las inflorescencias son caedizas en su con- junto a la madurez, actuando la panoja como una unidad de dispersión. Vegetti & Pensiero (1993) estudian la tipología de diversas especies de la sección Panicum, con- cluyendo que la inflorescencia en este grupo es po- Volume 83, Number 2 Zuloaga & Morrone 203 Panicum Subg. Panicum Secc. Panicum litélica, con florescencia principal desarrollada, no truncada. ESPIGUILLAS Las espiguillas son bifloras, con un antecio in- ferior estéril y uno superior perfecto, a excepcion- almente trifloras, con dos antecios inferiores estéri- les y uno superior perfecto en P. quadriglume (Fig. 27). Las especies de esta sección poseen espigui- llas globosas u ovoides a largamente elipsoides, ce- rradas (abiertas a la madurez en P. aquarum, P. aztecanum Zuloaga & Morrone, P. ephemeroides Zu- loaga & Morrone, P. exiguum Mez, P. hispidifolium Swallen, P. lepidulum, P. magnispicula, P. pela- doense y P. quadriglume), glabras y pajizas o con tintes purpúreos. La gluma inferior varía en tamaño desde Y, Y, (-4,) del largo de la gluma superior y la lemma inferior (midiendo de Y, a V, en P. flexile, P. mohavense y P. stramineum); es aguda a acuminada, —5-nervia, ocasionalmente 7—9-nervia, con el ner- vio medio a menudo escabroso. La gluma superior y lemma inferior son subiguales y 7-9(13—15)-ner- vias, ocasionalmente 5—7-nervias en P. aquarum y P. ephemeroides; la gluma inferior y superior se en- cuentran separadas por un entrenudo marcado en P. aquarum (Fig. 16), P. aztecanum (Fig. 17) y P. parcum. La gluma superior es tempranamente cae- diza en un grupo de especies que incluye a P. exi- guum, P. furvum, P. peladoense y P. quadriglume. Este carácter está correlacionado con el hecho de poseer estas especies antecio superior negruzco a la madurez, el que queda expuesto por la caída de la gluma superior. La pálea inferior varía en su desarrollo, desde especies en que la misma es nula o poco manifiesta, como por ejemplo P. capillare o P. hirticaule hasta especies en las que alcanza el mismo largo que la lemma inferior, como es el caso de P. bergii, P. chasei y P. stramineum. La flor in- ferior está ausente. El antecio superior encierra una flor perfecta; la lemma y pálea están endurecidas y poseen un color pajizo a castafio, en especies como P. chasei o negruzco, tal como se mencionó ante- riormente para P. exiguum, P. peladoense y especies relacionadas. La lemma encierra a la pálea y es 5— 7-nervia; la pálea es 2-nervia. En el interior del antecio se encuentran dos lodículas truncadas, cu- neiformes, que abrazan los bordes inferiores de la pálea superior, tres estambres y un ovario con dos estilos libres desde la base y estigmas plumosos. Finalmente el cariopsis varía de ovoide a elipsoide o fusiforme, tiene un hilo punctiforme y el embrión alcanza aproximadamente Y, a Y, del largo de la cariopsis (de Y, a Y, en P. capillarioides y P. deco- orans La espiguilla de las especies aquí tratadas posee el plan estructural de las Paniceae, pero difiere por tener una zona de articulación primaria en la base del antecio superior, cayendo este último desnudo a la madurez, desarticulando posteriormente las restantes brácteas en la base de la gluma inferior junto al pedicelo. La presencia de desarticulación primaria en la base del antecio superior ha sido previamente señalada en la tribu Paniceae en Axo- nopus, Brachiaria, Echinolaena, Ichnanthus, Meli- nis, Panicum, Paspalum, Pennisetum, Rhynchely- trum, Sacciolepis, Tricholaena y Yakirra (Lazarides & Webster, 1984; Silberbauer-Gottsberger, 1984; Davidse, 1987; Webster, 1987; Zuloaga, 1987a; Morrone & Zuloaga, 1992; Morrone et al., 1995). TEXTURA Y ORNAMENTACIÓN DEL ANTECIO SUPERIOR El antecio superior es ovoide a largamente elip- soide (obovoide en P. alatum Zuloaga & Morrone, Fig. 15) dorsiventralmente comprimido, crustáceo, liso y glabro, con los bordes de la lemma enrollados y cubriendo los 7, de la superficie de la pálea. La epidermis abaxial de la lemma y pálea posee célu- las largas rectangulares distribuidas en hileras lon- gitudinales, más de tres veces más largas qu e paredes anticlinales longitudinales e an- chas, onduladas. La epidermis abaxial puede presentar, en parti- cular sobre la pálea, diferentes ornamentaciones, incluyendo papilas simples, papilas verrugosas, mi- cropelos y aguijones. epidermis de la lemma comúnmente no posee ornamentación, habiéndose observado papilas sim- ples distribuidas regularmente en toda la superficie en Panicum alatum var. alatum (Fig. 1C), P. hir- ticaule var. verrucosum Zuloaga & Morrone (Fig. 1A) y P magnispicula y en ejemplares aislados de P. stramineum (como por ejemplo en Palmer 206) o sólo junto a la porción distal de la lemma en P quadriglume (Fig. 2C). Papilas simples distribuidas irregularmente en el ápice de la pálea se observaron en Panicum alatum var. longiflorum Zuloaga & Morrone, P. alatum var. minus (Andersson) Zuloaga & Morrone, P. azteca- num Zuloaga & Morrone, P. bergii, P. capillare, P. chasei, P. diffusum (Fig. 2A), P. exiguum, P. flexile, P. furvum (Fig. 2D), P. ghiesbreghtii, P. hallii (Fig. 2E), P. hirsutum, P. hirticaule var. hirticaule, P. his- pidifolium (Fig. 2G), P. lepidulum (Fig. 2F), avense, P. mucronulatum Mez, P. pampinosum, P. peladoense, P. philadelphicum, P. quadriglume (Fig. 2B, C) y P. stramineum. Papilas simples distribui- das regularmente en toda la superficie de la pálea se hallaron en P. alatum var. alatum (Fig. 1C, D), P. mo- 204 Annals of the Missouri Botanical Garden P. hirticaule var. verrucosum (Fig. 1A, B), P. mag- nispicula y ocasionalmente en ejemplares de P. stramineum (Fig. 1H). das irregularmente en el ápice de la pálea se halla- Papilas verrugosas agrupa- ron en P. aquarum, P. capillarioides, P. decolorans (Fig. 2H), P. ephemeroides, P. hillmanii, P. hispi- difolium (Fig. 2G), P. miliaceum y P. parcum. Se encontraron aguijones en el ápice de la lem- ma junto a la pálea en Panicum alatum var. minus, P. bergii, P. ghiesbreghtii, P. lepidulum, P. milia- ceum y P. mucronulatum. Micropelos bicelulares, con célula distal globosa, se observaron en el ápice de la pálea de Panicum aztecanum, P. ghiesbreghtii, P. lepidulum y P. mu- cronulatum. La base del antecio difiere en las especies es- tudiadas. Así, Panicum aquarum, P. aztecanum, P. decolorans, P. ephemeroides, P. hispidifolium (Vig. 16), P. lepidulum, P. magnispicula y P. parcum po- seen un anillo compuesto de células menos engro- sadas que el resto de la superficie de la lemma, castaño a la madurez. Este anillo es menos manifiesto en P. capillare, P. flexile y P. phi- ladelphicum. En Panicum bergii, P. capillarioides, P. diffusum, P. exiguum, P. furvum, P. ghiesbreghtii, P. hallii, P. hillmanii, (Fig. 1B, F), cronulatum, P. pampinosum, P. peladoense, P. qua- driglume y P. stramineum (Fig. 1H) notablemente P. hirsutum, P. hirticaule P. miliaceum, P. mohavense, P. mu- se encuentran 2 cicatrices basales a ambos márgenes de la lem- ma, compuestas de células de paredes delgadas, menos engrosadas que el resto de la superficie de la lemma, y marcadamente castañas a la madurez. Panicum alatum difiere de las restantes especies por poseer 2 expansiones aliformes, carnosas en los márgenes basales de la lemma, oliváceas a la ma- durez (Fig. 1D, E); este carácter es, hasta el mo- mento, único dentro del género Panicum y se en- cuentra, dentro de la tribu Paniceae, presente en los géneros Ichnanthus P. Beauv. y Echinolaena Desv. A través de un análisis con Sudán IV (Jen- sen, 1962) se determinó en P. alatum la presencia de aceites en los apéndices carnosos. La presencia de eleosomas en gramíneas fue correlacionada con la dispersión a través de hormigas (mirmecofilia), habiendo sido previamente mencionada en diversos géneros de gramíneas por Berg (1985), Davidse (1987), Zuloaga et al. (1988) y (1991) Morrone « Zuloaga DISTRIBUCIÓN Y ECOLOGÍA Las especies de la sección Panicum se encuen- tran aproximadamente desde los 50° de latitud nor- te hasta los 40° de latitud sur, desde el SE de Can- adá y los Estados Unidos de América hasta la región central de Chile y Argentina (Figs. 3-11). Entre los taxones con distribución más amplia den- Panicum strami- de los Estados Unidos de tro de la sección se encuentran: desde el SW América y México, Islas del Caribe hasta el W de neum, Argentina; P. alatum. var. minus desde el SW de Estados Unidos de América hasta Venezuela y Ec- uador (en las Islas Galápagos), P. hirticaule var. hirticaule desde el SW de Estados Unidos de América hasta Venezuela, Ecuador y Perú, P. ghies- breghtii desde el SW de Estados Unidos de Amér- ica, México, Islas del Caribe hasta Venezuela, Co- lombia y Ecuador. Panicum hirsutum se extiende desde el SW de Estados Unidos de América, Méx- ico e Islas del Caribe hasta Ecuador, Colombia, Venezuela y N de Brasil, habiendo sido coleccion- ada ocasionalmente en el N de Argentina. Final- mente, P. hispidifolium crece desde el sur de Méx- ico hasta Colombia y Venezuela. En América del Norte habitan exclusivamente las siguientes especies: Panicum philadelphicum, P. flexile y P. capillare se encuentran en el SE de habiendo sido introducida P. capillare en América del Sur en Brasil, Uruguay, Argentina y Canadá y Estados Unidos de América, Chile. Panicum pam- pinosum y P. mohavense son endémicas del SW de Estados Unidos de América, mientras que P. ala- tum var. alatum, P. capillaroides, P. hallii var. ha- llii, P. hallii var. filipes, P. hillmanii y P. hirticaule var. verrucosum crecen desde el S y SW de Unidos de América hasta el NE y Por último, P. aztecanum, Estados centro de México. P. alatum var. longiflorum y P. tamaulipense son endémicas de México Panicum lepidulum y P. parcum son exclusivas de México y América Central y P. furvum y P. diffu- P. decolorans, — Figura l. verrucosum.—A. Apice de la lem lemma. C,D. Panicum alatis var. alatt E. Pa Jolunga & Zizumbo 44: ( . Wiggins 15160; E. Carter ng 88 Pesa 206). Apice de la lemma y pálea.—D. Bas inic ‘um atu var. longiflorum, on del antecio, con dos expar » sobre la lem anicum stramineum, base d lpse io, con dos cic atrice es Fotomicrografías MEB de antecios superiores de especies de Panicum. A,B. Panicum hirticaule var. ma y gu. con papilas.—B. rices sobre la ase del antecio superior, con dos cicat de 1 antec io, con dos expansiones F ase t isiones 1a.—G. Panicum hispidifolium, da del antecio sobre la lemma. (A, > G, Rodríguez 1981; H, carnosas.—T. — -- LI 4970; F. Pohl & Davidse 1217 Volume 83, Number 2 Zuloaga & Morrone 205 Panicum Subg. Panicum Secc. Panicum 206 Annals of the Missouri Botanical Garden sum son endémicas de Guatemala e Islas del Ca- ribe respectivamente. Nueve especies son exclusivas de América del Sur. Dentro de las mismas, las que poseen una dis- tribución más amplia son P. exiguum y P. bergii var. bergii, habiendo sido introducida esta última es- pecie en los Estados Unidos de América; P. aqua- rum se halla en el N y NW de América del Sur, P. peladoense en el S de Brasil, Bolivia, Paraguay y Argentina, P. quadriglume en Perú, Bolivia, Para- guay, Brasil y Argentina, P. bergii var. pilosissimum en el NE de Argentina y S de Brasil, P. chasei en el E de Argentina, S de Brasil y Uruguay y P. mu- cronulatum, P. ephemeroides y P. magnispicula son endémicas del Brasil. Todas las especies de esta sección habitan usual- mente en campos o sabanas, sobre suelos arcillo- sos, arenosos a rocosos, en bordes de caminos y también en laderas de montafia; algunas especies en América Central se hallan en bosques decfduos. Una excepción la constituye P. aquarum, especie que habita campos bajos inundables del norte de América del Sur. Las especies crecen desde el ni- vel del mar hasta los aproximadamente 1300 m, llegando P. lepidulum hasta los 2150 m y P. hirti- caule hasta los 2500 m en el Peri. NÜMEROS CROMOSÓMICOS Los estudios cariológicos en la sección Panicum son escasos, habiéndose registrado datos sobre la citología de 16 especies. La sección se caracteriza por tener un número básico de cromosomas x = 9 (Zuloaga, 1987a) y se han citado diferentes niveles de ploidía dentro de la misma. A continuación se resumen los námeros cromo- sómicos determinados para las especies de la sec- ción Panicum tratadas en el presente estudio: Panicum alatum var. minus, 2n — 18 (Gould, 1965, ajo P. hirticaule). Panicum bergii var. bergii, n = 18 (Dubcovsky € Zuloaga, 1991); 2n — 36 (Delay, 1950; Parodi, 1946, bajo P. pilcomayense; Núñez, 1952, bajo P. pilcomayense). Panicum capillare, n — 9 (Spellenberg, 1967, bajo P. capillare var. occidentale Rydb.); 2n = 18 (Avdulov, 1931; Gould, 1968, 1975; Spellen- berg, 1970; Ferakova, 1976; Fairbrothers, 1954; Tzvelev, 1976, bajo P. capillare subsp. barbipulvinatum (Nash) Tzvelev; Reeder, 1977; Love & Love, 1981, bajo P. capillare var. occidentale; Vahidy et al., 1987). Panicum capillarioides, 2n = 36 (Gould, 1960, Panicum decolorans, 2n = 36 (Gould, 1966). Panicum flexile, 2n = 18 (Brown, 1948; Fairbroth- ers, 1954). Panicum hallii var. filipes, 2n = 18 (Brown, 1951; uld, 1958, 1968; Waller, 1976); 2n = 36 (Gould, 1958). Panicum hallii var. hallii, 2n = 18 (Gould, 1958, 96 968; Reeder, 1971; Waller, 36 (Waller, 1976, bajo P. pilco- 1976); 2n = mayense). Panicum hillmanii, 2n = 18 (Fairbrothers, 1954). Panicum hirticaule, n = 9 (Davidse & Pohl, 1972); 2n = 18 (Fairbrothers, 1954; Gould, 1975). Panicum hirsutum, 2n = 36 (Gould, 1975; Waller, , 1976). Panicum hispidifolium, n = 27 (Dubcovsky & Zu- loaga, Panicum nul = 18 (Parfitt, 1981); 2n = 36 (Avdulov, 1931; Krishnaswamy, 1951; Frey et al., 1981), 2n = 54 (Krishnaswamy, 1951); 2n = 72 (Arenkova, 1940; Krishnaswamy, 1951). Panicum pampinosum, 2n = 18 (Reeder, 1977); 2n = 36 (Gould, 1966). Panicum parcum, 2n = 36 (Gould, 1966; Gould & Soderstrom, 1970). Panicum philadelphicum, 2n = 18 (Brown, 1948; Fairbrothers, 1954, bajo P. lithophilum). Panicum stramineum, 2n = 36 (Fairbrothers, ANATOMÍA FOLIAR Caracteres histofoliares en corte transversal (Fig. 12). Transcorte: plano, expandido o en forma de “U” o “V”; semiláminas simétricas a ambos lados del nervio medio; láminas de ancho variable, con 16-120 haces vasculares en transcorte; espesor de la lámina variable, de 91-250(-325) um en las => Figura 2. Fotomicrografías MEB de antecios superiores de especies A Monde um.—A. Panicum diffusum. Ápice de la lemma y pálea, con papilas simples hacia el ápice de la pálea y pálea.—C. Detalle de B, con papilas simples.—D. Panic | B, C. icum quadriglume.—B. Apice de la lemma um. ee de la lemma y pálea, con papilas da um furvu el ápice de la pálea.—E. Panicum hallii var. hallii. Apice de la lemma y pálea, con ioc sobre la pálea Panicum lepidulum. ae de la pálea, con papilas simples y micropelos.—G. Panic —H. Panicum decolorans. Apice de la wa pálea, con papilas v m hispidifolium. pie de l , con inia as verrugosas. (A, Eggers 76; B, n.) gosas.— C. Chase 8897; D, ad 51627; E, Hall 816; F, Pringle 497, G, Roda 1981; H, Humboldt s.n Volume 83, Number 2 Zuloaga & Morrone 207 Panicum Subg. Panicum Secc. Panicum 208 Annals of the Missouri Botanical Garden wm P. alatum var. alatum P. hirticaule var. verrucosum (9 P. mohavense e P.philadelphicum SCALE O 1006 200 300 400 500 600 700 800 900 1000 MILES 400 600 800 1000 1200 1400 KILOMETERS LAMBERT'S AZIMUTHAL EQUAL- AREA PROJECTION o PF C do | — R 120 110 100 90 80 WEST LONGITUDE GOODE'S SERIES OF BASE MAPS Prepared by Henry M. Leppard HENRY M. LEPPARD, EDITOR Duhliched h he = US £ Chi. D, Chi Winn Figura 3. Distribución de Panicum alatum var. alatum, P. hirticaule var. verrucosum, P. mohavense y P. philadelphicum. Volume 83, Number 2 1996 Zuloaga & Morrone 209 Panicum Subg. Panicum Secc. Panicum (&) P. hirsutum Æ P. mucronulatum @ P.peladoense O 200 400 600 800 1000km ane) 100 200 300 400 500 600 miles Figura 4. costillas, 40-170 um en los surcos, los márgenes planos a ligeramente enrollados; superficie adaxial plana sin costillas y surcos manifiestos o con cos- tillas de ápice redondeados a ligeramente aplana- dos y surcos poco profundos y suaves, hasta Y, del ancho del transcorte (en P. aquarum, P. aztecanum y P. ephemeroides con surcos profundos, hasta 7, del ancho del transcorte); costillas y surcos abaxia- les variables, no manifiestos o con costillas de ápi- ce redondeado a ligeramente plano y surcos suave- mente pronunciados, hasta '/, del ancho del transcorte. Haz vascular medio: de estructura varia- ble, con un haz vascular de primer orden estruc- turalmente indistinguible de los restantes laterales o con un haz vascular de primer orden estructu- ralmente distinguible, solitario o acompafiado por 1-2 haces vasculares de primer orden y 2—5 haces vasculares de segundo orden y con células paren- quimáticas incoloras localizadas hacia la cara ada- xial. Distribución de los haces vasculares: 3-16 ha- Distribución de Panicum hirsutum, P. mucronulatum y P. peladoense. ces vasculares de primer orden (21—30 haces vasculares de primer orden en P. hirsutum), (2—)3— 7 haces vasculares de segundo orden entre haces vasculares de primer orden contiguos; todos los ha- ces vasculares equidistantes de ambas epidermis. Estructura de los haces vasculares: haces vasculares de primer orden de contorno subcircular a elípti- cos, con vasos de metaxilema de contorno angular y diámetro igual o menor que las células paren- quimáticas Kranz; haces vasculares de segundo or- den, de contorno angular, con xilema y floema dis- tinguible. Vaina de los haces vasculares: vaina parenquimática externa Kranz de estructura varia- ble: (A): de contorno regular, compuestas por célu- las Kranz de paredes radiales y tangencial interna recta, con cloroplastos especializados, alargados, de posición centrípeta o (B) de contorno irregular con las paredes radiales y tangencial externa manifies- tamente arqueadas y con cloroplastos especializa- dos, discoides, localizados centrífugamente: exten- 210 Annals of the Missouri Botanical Garden e P.capillare = P. capillarioides Figura 5. Distribución de Panicum alatum var. minus, P. capillare y P. capillarioides. Volume 83, Number 2 Zuloaga & Morrone 211 Panicum Subg. Panicum Secc. Panicum *K P. aquarum € P. bergil var. bergii nor = P.decolorans Xt P. ghiesbreghtii di \ msn o 200 400 600 800 1000km O 100 200 300 400 500 600 miles 100 o Figura 6. Distribucion de Panicum aquarum, P. bergii var. bergii, P. decolorans y P. ghiesbreghtit. siones parenquimáticas de la vaina ausentes; vaina interna mestomática completa, compuesta por célu- las pequeñas de paredes uniformemente engrosa- das. Células distintivas Kranz: ausentes. Esclerén- quima: pobremente desarrollado, P A adaxiales y abaxiales, subepidérmi- en grupos cos, en contacto o no con los haces vasculares; mar- gen de la lámina con escaso esclerénquima sub- marginal. Mesófilo: radiado en una capa contínua o interrumpida adaxial o abaxialmente por trabas de esclerénquima, compacto, con escasos espacios in- tercelulares; células clorenquimáticas tabulares; 2(-3) células clorenquimáticas entre haces vascu- lares contíguos; células parenquimáticas incoloras ausentes o presentes por debajo de las células buli- formes, compuestas por 1-2-3-series células en contacto con ambas epidermis o sólo por debajo de las células buliformes, en P. furvum. “Arm cells” ausentes. Células fusoides ausentes. Células epi- dérmicas adaxiales: células buliformes presentes en los surcos adaxiales y entre los haces vasculares, en grupos pequefios, en forma de abanico, com- puestos por 2-7 células, la central usualmente de mayor tamaño que las laterales, ocupando hasta Y, del ancho del transcorte. Células epidérmicas pe- quefias y regulares en forma, excepcionalmente ar- queadas en P. aquarum, P. aztecanum y P. epheme- roides. Aguijones y ganchos de frecuencia variable. Papilas ausentes, excepcionalmente presentes en ejemplares aislados de P. stramineum. Macropelos presentes o ausentes, cuando presentes con base bulbosa, asociada a células epidérmicas sobreele- vadas o en el mismo nivel que el resto de la su- perficie. Células epidérmicas abaxiales: células bul- iformes ausentes. Células epidermicas pequeñas. Aguijones y ganchos como en la cara adaxial. Pap- ilas ausentes. Macropelos de frecuencia variable, similares a los presentes en la cara adaxial. Epidermis abaxial en vista paradermal (Fig. 212 Annals of the Missouri Botanical Garden a P.aztecanum € P. hallii var. hallii vol. 2k P. parcum : > » 1 ` eS 3 i SM A $ | SCALE © 100 200 300 400 500 600 700 800 —— À————" 4 00 a o 1000 ts o E 400 [ Mu Al 1400 KILOMETERS n | - r^ | | | i LAMBERT'S AZIMUTHAL EQUAL-AREA PROJECTION | | o | | . 1 ui d Lo — ] =a aes | " ——À en — € s. ———— — al c 120 110 | 100 90 80 WEST LONGITUDE GOODE'S SERIES OF BASE MAPS HENRY M LEPPARD, EDITOR Prepared by Henry M. Leppard Published by the Univeraity of Chi p. C. Figura 7. Distribución de Panicum azt ecanum, P. hallii var. hallii y P. parcum. Volume 83, Number 2 1996 Zuloaga & Morrone 213 Panicum Subg. Panicum Secc. Panicum Æ P. bergii var. pilosissimum © P. ephemeroides | ——| * P.hispidifolium a P.lepidulum *k P. magnispicula e P.quadriglume O 200 400 600 800 1000km Ls st 1 miles ura 8. Distribución de Panicum bergii var. pilosissimum, P. ephemeroides, P. hispidifolium, P. lepidulum, P. magnispicula y P. quadriglume. 13). Zonación: zonas costales e intercostales di- ferenciadas; zonas costales 1—6 células de ancho; zonas intercostales de 3—14 células de ancho, ho- mogéneas. Células largas intercostales: rectangula- largas que anchas, de res, más de 3 veces más paredes suavemente engrosadas, las anticlinales longitudinales onduladas, paralelas, las anticlinales transversales rectas. Células cortas intercostales: con su eje mayor transversalmente alargado, soli- tarias o en pares sílico-suberosos, de distribución homogénea en la zona intercostal, alternando con 1(-2) células largas, o raro más. Aparatos estomá- ticos: de 26—40 um de largo, 15.5-26 jum de ancho (19.5 y 45 pm de largo en P. mohavense y P. hir- sutum respectivamente), distribuidos en 2 hileras longitudinales en la zona intercostal, cada una próxima a la zona costal, usualmente en 1 hilera, de disposición central en P. peladoense; estomas se- parados por una célula larga interestomática, rect- angular, de paredes terminales cóncavas; células subsidiarias cupuliformes a ligeramente triangula- res. Micropelos: bicelulares, fusiformes, de 8 um de largo, comúnmente distribuidos en la región central de la zona intercostal, con la célula basal de paredes paralelas, más engrosadas que las de la célula distal; célula distal 2 a 3 veces más larga que la basal, decídua, de paredes muy delgadas (en P. ephemeroides y P. magnispicula la célula basal puede ser menor del mismo tamaño que la distal). Aguijones: presentes o ausentes, cuando presentes de frecuencia variable, distribuidos en las zonas costales, de barba más larga que la base. Ganchos: ausentes o presentes, de frecuencia variable, usual- mente dispuestos entre las hileras de estomas y la zona costal, de barba más larga que la base. Pap- ilas: ausentes. Macropelos: presentes o ausentes, cuando presentes unicelulares, distribuidos en la región central de la zona intercostal o en los már- 214 Annals of the Missouri Botanical Garden = SH E | 9 -y % Un p "lal 0 — 2]. onc pIEEED[ees2Mso- a = le Ji Ya : be oni J : 2 a W : ez ———20 | I Ra ORL É* T pos — ae A Y 2s i o | ° "i" < - ee = P.chasei . ERI : eZ |] T é @) P. diffusum è A T ho e P.exiguum : »* P. furvum n T. el Y f : ) Æ P. stramineum mu ) A E E e co O P. tamaulipense ES " A - iX L = vid - > By 0 dn - ae i i E Es T ^x». » MN [ONE - wa —— DNE 2 - s : bo 80 70 60 i 40 Figura 9. Distribución de Panicum genes, variable en estructura: (a) de paredes del- gadas, de 150-300 ¡um de largo, asociados a 2—4 células epidérmicas o (b) de 500-3350 jum de lar- go, rígidos, de paredes engrosadas, asociados a nu- merosas células epidérmicas sobreelevadas en re- lación al resto de la epidermis. Cuerpos de sílice costales: halteriformes, de eje longitudinal corto, menos frecuentemente nodulares, cruciformes o halteriformes de eje longitudinal alargado, distri- buidos en 14 hileras longitudinales, alternando con células cortas; cuerpos de sílice intercostales usualmente cruciformes, menos frecuentemente de contorno irregular, homogéneamente distribuidos en la zona intercostal. Epidermis adaxial en vista paradermal. Zona- ción: zonas costales e intercostales distinguibles; zonas intercostales heterogéneas, con una región central con células buliformes distinguibles de las laterales. Células largas intercostales: rectangula- chasei, P. diffusum, P. exiguum, P. furvum, P. stramineum y P. tamaulipense. res, las centrales de paredes anticlinales delgadas y suavemente onduladas, las laterales con paredes suavemente engrosadas, las anticlinales longitudi- nales onduladas. Células cortas intercostales: soli- tarias o en pares sílico-suberosos, distribuidas en bandas a cada lado de las zonas costales, rectan- gulares o de contorno irregular, transversalmente alargadas. Aparatos estomáticos: similares a los pre- sentes en la cara abaxial, distribuidos en 2 hileras en la zona intercostal. Micropelos: bicelulares, de estructura similar a los de la cara abaxial, presen- tes entre la banda estomática y la zona costal. Agui- Jones: frecuentes, raramente ausentes, distribuidos en hileras longitudinales en la zona costal. Gan- chos: frecuentes en hileras longitudinales a ambos lados de la zona costal y las bandas estomáticas, menos frecuentemente aislados o ausentes. Papilas: ausentes, sólo presentes en ejemplares aislados de P. stramineum (Allem 1479, Harley et al. 16294), Volume 83, Number 2 1996 Zuloaga & Morrone Panicum Subg. Panicum Secc. Panicum 215 f& P. alatum var. longiflorum @ P. flexile m P. hallii var. filipes *K P. hillmanii Ye P. pampinosum 200 00 600 800 1:000 1200 1400 LAMBERT'S AZIMUTHAL EQUAL- AREA PROJECTION o , EUN S. o Baer d = MS ! Ho 100 60 WEST LONGITUDE GOODE'S SERIES OF BASE MAPS Prepared by Henry M. Leppard HENRY M LEPPARD, EDITOR Published by the Un; C E Chi g p, Chi 9 Manos Figura 10. Distribución de Panicum alatum var. longiflorum, P. hallii var. filipes, P. flexile, P. hillmani y P. pampinosum. 216 Annals of the Missouri Botanical Garden pe f ^ uM 3 P > PD S Lu» ra A 7 Y a j WS PM P iu: y nme mE fo | = — | E » ; o T Oe wo 0 1 1 l A Lo j i A m _ | l | ' \ | e RENI LU 1 e P.hirticaule | | A ee Figura 11. Distribución de Panicum hirticaule. con 1-7 papilas por célula larga, distribuidas en — 16-120 haces vasculares en transcorte. Las costi- hileras longitudinales, de diámetro mayor que Y, del llas y surcos presentan diferentes grados de desa- ancho de la célula larga. Macropelos de frecuencia rrollo, desde poco manifiestos hasta Y, del ancho y estructura similar a los presentes en la cara aba- del transcorte, o con surcos adaxiales profundos, xial. Cuerpos de sílice costales: halteriformes o a ve- hasta V, del ancho, como se observó en P. aquarum, ces cruciformes, raramente nodulares; cuerpos de P. aztecanum y P. ephemeroides. El nervio medio es sílice intercostales: cruciformes o de contorno irre- también variable, con un haz vascular de primer gular, con su eje mayor transversalmente alargado, orden estructuralmente indistinguible de los restan- distribuidos entre las bandas estomáticas y la zona tes laterales o estructuralmente distinguible, inclu- costal. yendo 1-3 haces vasculares de primer orden y 2- Las especies americanas de la sección Panicum 5 haces vasculares de segundo orden, asociados o poseen una anatomía foliar variable. Las hojas son no a células parenquimáticas incoloras. Este rango planas, expandidas o en forma de “U” o “V,” si- de variación se observó dentro de una misma es- métricas a ambos lados del nervio medio, incluyen- pecie, como por ejemplo en P. bergii, P. chasei, P. do un número variable de haces vasculares desde diffusum, P. ghiesbreghtii y P. stramineum. Volume 83, Number 2 1996 Zuloaga & Morrone 217 Panicum Subg. Panicum Secc. Panicum Figura 12. una porción de la lámina.—B. Detalle de la lámina, con haces vasculares de primer y segun con cloroplastos marc dente centrífugos, flecha señalando nervio medio. 3 a, flecha sefialando nervio medio.—D. Detalle de la lámina células [9 con cloroplastos centrífugos.—E. Panicum hispidifolium, detalle de una porción de Pan de una porción de la lámin y mo orden y lámina E e detalle de una porción de lámina c n haces vasculares de primer y segundo orden y células Kranz con cloroplas on haces vasculares de primer y segundo orden y células Kranz con Anatomía foliar en corte transversal de especies de Panicum. A, B. Panicum bergii.—A. Transcorte de o orden y células Kranz anicum stramineum.—C. Transcorte n haces vasculares de primer stos centrípetos.—F. um cloroplastos centripetos. (A, B, Zuloaga et al. 2323; C, D, Zuloaga et al. 2515; E, Davidse & Pilz 31533; F, Zuloaga 3447.) & Deginani El mesófilo es conspicuamente radiado, compac- to, continuo o interrumpido por trabas de esclerén- quima, pudiéndose hallar columnas 1-3-seriadas de células parenquimáticas incoloras en contacto con ambas epidermis (en P. alatum, P. aquarum, P. capillare, P. decolorans, P. hirsutum, P. hirticaule, P oe P. lepidulum, P. magnispicula, P. m y P. tamaulipense) o sólo en contacto con ld Pe buliformes (en P. furvum La presencia de una doble vaina E de los haces vasculares, la externa con cloroplastos especializados, y por poseer mesófilo radiado, con 2(-3) células clorenquimáticas entre haces vascu- lares contíguos, nos permiten inferir que las espe- cies de la sección Panicum son Kranz, del subtipo anatómico PS (= XYMS+), de acuerdo con los ca- racteres establecidos por Brown (1977), Ellis 71) y Hattersley & Watson (1976). estructura de la vaina parenquimática de los haces vasculares presenta variaciones anatómicas, lo que permite distinguir dos grupos marcadamente diferenciados: (A) con vainas de contorno regular, con células Kranz de paredes radiales y tangencial interna rec- tas, con cloroplastos alargados, de posición centrí- peta. Dentro del mismo se incluyen a P. alatum, P. aquarum, P. aztecanum, P. ephemeroides, P. exi- guum, P. filipes, P. flexile, P. furvum, P. halli, P. hillmanii, P. hispidifolium (Fig. 12E), P. hirticaule, P. magnispicula, P. miliaceum, P. mohavense, P. pampinosum, P. parcum, P. peladoense (Fig. 12F), P. philadelphicum, P. quadriglume y P. tamauli- pense. (B) con vainas parenquimáticas de contorno irre- gular, con células Kranz de paredes radiales y tan- gencial externa arqueadas, con cloroplastos discoi- des, distribuidos centrifugamente. Se incluyen en este grupo a P. bergii (Fig. 12A, B), P. capillare, P. capillarioides, P. chasei, P. decolorans, P. diffusum, 218 Annals of the Missouri Botanical Garden ee ee in aay rt D KC e o — ee S A pg Sti : e i z p m on oe e Se Figura 13. a nie de especies de Panic chaset.—D. Panicum peladoense.—E. Panic m.—A. Panicum bergii.— ‘um quadri al ume. e.—C. Panicum 3072: B, B. Panicum capillar F. Panicum stramineum. (A, Zuloaga et al. . Zuloaga & Moran: 3060; C, ua & Zuloaga 32423; D, Zuloaga & Deginani 3447; E, Chase 8897; F, Zuloaga 9.) P. ghiesbreghtii, P. hirsutum, P. lepidulum, P. mu- cronulatum y P. stramineum (Fig. 12C, D). Las especies del grupo A poseen cloroplastos centrípetos, característicos de las especies del sub- tipo anatómico NAD y fisiológico NAD-me (Gutié- rrez et al., 1974). Las del grupo B tienen cloro- plastos de posición centrífuga, estructura típica de subtipo anatómico PCK y que tradicionalmente ha — sido asociada al tipo fisiológico PEP-ck (Gutiérrez et al., 1976). No obstante, la correlación estructural y funcional debe tomarse con precaución, dadas las excepciones halladas hasta el presente dentro de Panicum. Así, diversos autores citaron especies de Panicum con anatomía PCK y subtipo fotosintético NAD-me (Ohsugi & Murata, 1980; Ohsugi et al., 1982, Oguro et al., 1985, Prendergast et al., 1987). Dentro de la sección Panicum, P. bergii y P. cap- illare, especies que se incluyen en el grupo B y que corresponden al subtipo anatómico PCK, han sido descritas fisiológicamente como del tipo NAD- me (Gutiérrez et al., 1974; Prendergast et al., 1987) Volume 83, Number 2 1996 Zuloaga & Morrone 219 Panicum Subg. Panicum Secc. Panicum Las similitudes exomorfológicas de los dos gru- pos anatómicos con cloroplastos centrípetos y cen- trífugos requieren inmediata confirmación del tipo fotosintético a través de estudios fisiológicos. istribución en América de las especies de la sección Panicum con anatomía NAD y PCK en relación a las divisiones políticas, muestra una marcada frecuencia relativa de especies NAD en América del Norte, con un 66%, 80% y 60% de especies NAD en Canadá, Estados Unidos de mérica y México respectivamente. Cabe destacar que México presenta la mayor concentración de es- pecies de la sección, representada por el 48% de las mismas y un 10% de especies endémicas. A su vez, en las Islas del Caribe sólo se ha registrado la presencia de especies del tipo anatómico PCK. En América Central la relación entre las especies NAD y PCK es variable, predominando las especies NAD en una frecuencia del 75%, excepto en Belize y Costa Rica donde sólo se han registrado especies PCK. En América del Sur la frecuencia de especies NAD es relativamente baja, oscilando entre el 54 y el 66%, con una alta frecuencia re- lativa del 100% de especies PCK en Uruguay. Cabe destacar que en Brasil habita el 36% de las espe- cies de la sección, con un 10% de especies endé- micas. El patrón de distribución de los dos tipos anatómicos no muestra una clara predominancia de alguno de ellos en América; sin embargo serían necesario estudios más detallados de las relaciones fitogeográficas con los diferentes tipos anatómicos A . En lo que se refiere a la comparación entre los caracteres exomorfológicos e histofoliares, el estudio anatómico muestra marcadas diferencias entre especies exomorfológicamente afines. Así, parcum, especie que fue considerada como sinóni- mo de P. decolorans y que posee marcadas afini- pues con P. lepidulum, pertenece anatómicamente A, o sea que es del subtipo anatómico NAD, Testa que P. decolorans y P. lepidulum poseen el subtipo anatómico PCK. Panicum azte- canum, especie de la que varios especímenes fue- ron previamente identificados como P. lepidulum, se distingue, además de diferencias exomorfólogi- cas que se discuten bajo esta especie, por tener un subtipo anatómico NAD. Panicum diffusum, espe- cie erróneamente citada fuera de su área natural de Islas del Caribe por diferentes autores, se separa de P. hallii var. filipes por ser anatómicamente PCK. Panicum hirticaule, especie afín a P. ghies- breghtii y P. stramineum, se caracteriza por tener anatomía NAD, mientras que los dos últimos taxo- nes son PCK Algo similar ocurre con P. mohavense y P. stra- mineum, siendo la primera del subtipo anatómico NAD y la segunda del subtipo anatómico PCK. Fi- nalmente, P. capillare se distingue por ser anatóm- icamente una especie PCK, mientras que P. flexile, P. hillmanii y P. philadelphicum, especies afines de América del Norte, son uniformemente del subtipo anatómico NAD. DISCUSIÓN Las especies de la sección Panicum poseen ca- racteres anatómicos y exomorfológicos diagnósticos del subgénero Panicum. El subgénero se caracteri- za por incluir especies Kranz, del subtipo anató- mico PS (=XyMS+) (Hattersley & Watson, 1976; Brown, 1977), con 2 vainas alrededor de los haces vasculares, la interna mestomática y la externa Kranz, con cloroplastos especializados, mesófilo compacto, radiado, y con 2-3 células clorenqui- máticas entre haces vasculares. Además este sub- género posee un número básico de cromosomas x = 9 e incluye especies con espiguillas elipsoides a largamente ovoides, con gluma inferior de tamaño variable y gluma superior y lemma inferior común- mente 7—9-nervias (excepcionalmente 5-nervias en algunas especies y hasta 13-15-nervias en otros taxones). El antecio superior es uniforme en el sub- género, crustáceo, liso, con papilas simples o con papilas verrugosas en el ápice de la pálea, menos frecuentes en toda superficie (Zuloaga, 1987a). Las especies del subgénero Panicum habitan usualmente ambientes abiertos y secos, a excepción de especies de Dichotomiflora, las que crecen en bordes de ríos. Hitchcock & Chase (1910) consideran informal- mente a las especies americanas de la sección Pan- icum dentro de los grupos Capillaria y Diffusa, to- especies son anuales o perennes respectivamente. Pohl (1980), McVaugh (1983) y Zuloaga (19872) consideran que este cár- acter no resulta diagnóstico para distinguir a las especies aquí agrupadas en la sección Panicum, resultando el mismo de difícil distinción en re- giones tropicales (McVaugh, 1983). A la vez las especies consideradas aquí en la sección Panicum poseen una combinación de caracteres diagnósticos comunes: la misma incluye plantas cespitosas, con cafias usualmente simples o ramificadas en la base (excepcionalmente ramificadas en los nudos supe- riores en P. aquarum); las lígulas son membraná- ceo-ciliadas y las láminas oblongo-lanceoladas a filiformes. Las inflorescencias son piramidales, lax- as y difusas, raramente con las ramas adpresas, con mando en cuenta si las espiguillas ovoides a largamente elipsoides, diver- gentes o más o menos adpresas, dispuestas no un- ilateralmente. El largo de la gluma inferior es vari- 220 Annals of the Missouri Botanical Garden able, comúmente de Y, a Y. La gluma superior y lemma inferior son 7—9-nervias y superan en largo al antecio superior. La pálea inferior alcanza igual largo que la lemma inferior o está reducida a au- sente en algunas especies; la flor inferior es ausen- te. Finalmente el antecio superior es crustáceo y brillante y siempre lleva papilas hacia el ápice de la pálea superior. Diversas modificaciones estructurales, presentes en las espiguillas e inflorescencias de las especies de la sección Panicum, sugieren diversos mecanis- mos de dispersión. En primer lugar la presencia de lemma y pálea endurecidas encerrando a la cariop- sis a la madurez fue correlacionada a la dispersión por animales granívoros o herbívoros, los que in- gieren las diásporas accidental o intencionalmente (Ridley, 1930; 1982). (1985) sugiere que las brácteas endurecidas pro- van der Pijl, Thomasson tegen a la cariopsis del ataque de insectos o dur- ante el pasaje de la misma a través del tracto di- gestivo de los animales. Dentro de la sección Panicum se ha observado, a diferencia de lo que ocurre en el resto del género, que el antecio superior es tempranamente caedizo. Este mecanismo de dispersión está correlacionado, en Panicum exiguum, P. furvum, P. peladoense y P. quadriglume, con el hecho de que al madurar la cariopsis la gluma superior cae, dejando el dorso del antecio superior al descubierto, siendo este úl- timo de color negro y brillante. Un síndrome se- mejante al hallado en las especies americanas des- critas, fue observado en P atrosanguineum A. Rich., especie que habita en Africa y en el NW de la India (Clayton, 1989). Van der Pijl (1982) indica que la presencia de color llamativo y rigidez a la madurez en las diásporas funcionaría como un me- canismo de atracción para las aves granívoras. Adicionalmente a la caída prematura del antecio superior, en Panicum bergii y P. capillare las inflo- rescencias desarticulan en su conjunto a la madu- rez y son dispersadas por el viento. Finalmente, la presencia en Panicum alatum de eleosomas en la base de la lemma permite sugerir que las diásporas en esta especie se dispersan a través de hormigas. Beattie (1983) sugiere diversos significados adaptativos para el establecimiento de la mirmecofilia, entre las que se incluyen protec- ción al fuego, dispersión a distancia y el hecho de poseer los hormigueros un microhabitat rico en nu- trientes. Apéndices aliformes similares a los pre- sentes en Panicum alatum en los márgenes basales de la lemma, han sido señalados en especies de Ichnanthus y Echinolaena (Davidse, 1987). RELACIÓN CON SECCIONES DEL SUBGENERO PANICUM La sección típica se relaciona en América, den- tro del subgénero Panicum, con las secciones Rud- geana, Virgata Hitchc. & Chase, Urvilleana (Hitche. & Chase) Pilg. y Dichotomiflora (Hitche. & Chase) Honda. La sección Rudgeana es exclu- sivamente americana y comprende 5 especies dis- tribuidas desde Mesoamérica y las Islas del Caribe hasta Bolivia y Brasil. Esta sección es, desde el punto de vista anatómico, exclusivamente del sub- tipo NAD (Fig. 14C, D) y se distingue de las es- pecies de la sección Panicum que poseen este sub- tipo anatómico por incluir taxones con espiguillas marcadamente abiertas a la madurez y de inserción oblícua, con gluma inferior acuminada a subulada y antecio superior separado por un conspicuo en- trenudo de la raquilla. El entrenudo es dorsiven- heterogéneo, constituído J tralmente comprimido por dos porciones, una porción basal y dorsal ríg- ida, lisa y brillante, de la misma consistencia que el antecio superior, y una porción ventral junto a la pálea del antecio superior constituída por tejido membranáceo (Zuloaga, 1987b às especies PCK de la sección Panicum poseen — una anatomía similar a las especies de Dichotomi- flora; esta última se caracteriza por tener además papilas simples en la epidermis adaxial o en ambas Fig. 14A, B) (Ellis, 1988; Zuloaga et al., 1993). Papilas adaxiales fueron halladas en especímenes ja aislados de P. stramineum. Las especies de la sec- ción Dichotomiflora crecen en lugares húmedos, siendo frecuentes en bordes de ríos, con cañas erectas a decumbentes, arraigadas en los nudos in- feriores. Las espiguillas en esta sección poseen glu- ma inferior 1—3-nervia, Y, a Y,(—%) del largo de la gluma superior y la lemma inferior. Las secciones Urvilleana y Virgata difieren de las especies de la sección Panicum por tener clo- roplastos discoides, distribuidos periféricamente en las células Kranz (Fig. ¿—H) (Zuloaga et al., 1993). Las especies de Urvilleana se distinguen anatómicamente por tener haces vasculares de primer, segundo y tercer orden, estos últimos dis- tribuidos por debajo de las células buliformes, con extensiones adaxiales de la vaina desarrolladas en los haces vasculares de primer y segundo orden, y surcos profundos; más de Y, del ancho del tran- scorte. Exomorfológicamente Urvilleana se dis- tingue por incluir especies que habitan en dunas o sobre suelos arenosos con rizomas cundidores, lám- inas involutas, rígidas, densamente pilosas, espi- guillas pilosas con gluma inferior Y, a Y, del largo de la espiguilla, flor inferior presente y antecio su- perior piloso hacia la base, con largos pelos blan- Volume 83, Number 2 Zuloaga & Morrone 221 1996 Panicum Subg. Panicum Secc. Panicum EIC: xr THO HEUTE Oper rx : taot ipu XX TFA y DOE SERE ne NM ne, CO SUE mites ts E E A Sa Bon TU LAE co ALTI à Deore st X Wh ae. id BEAT AR ESE SS ar re MEET m. x RO DEUX + COPE dH 3 So A ris UE AH ky. 4 ue H 50 um A Figura 14. Anatomía foliar de especies de À ata . A, B. Panicum elephantipes.—A. Detalle en transcorte de una porción de lámina, con haces vasculares de primer y segundo orden, células Kranz con cloroplastos centrífugos y papilas sobre ambas epidermis.—B. Epidermis iura, con papilas y aguijones costales.—C. Panicum cervicatu. m gu : sie noh ph dirinbucion "HERE S nia G, H. ri racemosum.—G. Detalle en transcorte de una porción de la lámina.—H. scorte con seculares de primer, segundo y tercer orden, células Kranz con EE oisi: periféricos. (A, Z Zulo. oaga 2336; B, goes 3084; C, LA d 3837; D, Zuloaga 3978; E, Zuloaga s.n.; V. Zuloaga et al. 2304; G, H, Zuloaga 3360.) 222 Annals of the Missouri Botanical Garden quecinos. Las especies de Virgata poseen asimismo rizomas cundidores, cañas rígidas, espiguillas con flor inferior presente. Panicum maximum Jacq., especie del género perteneciente al subgénero Megathyrsus Pilg. y que anatómicamente se caracteriza por ser del subtipo PCK, difiere de las especies de la sección típica por ser una especie con gluma inferior reducida, por tener gluma superior y lemma inferior 5-ner- vias, flor inferior estaminada y antecio superior transversalmente rugoso (Zuloaga, 1987a). RELACIÓN CON OTROS GÉNEROS DE LA TRIBU PANICEAE La sección Panicum se separa, dentro de la tribu Paniceae, de géneros afines y con anatomía PCK por las siguientes características: Urochloa posee ramificaciones de la inflorescencia racemosas y an- tecio superior transversalmente rugoso, mucronado a crestado, no caedizo; Brachiaria posee ramifica- ciones de la inflorescencia con espiguillas dispues- tas unilateralmente, espiguillas vilosas con gluma inferior Y, o menos del largo de la espiguilla y an- sin papilas; Eriochloa posee una dilatación cupuliforme en la tecio superior completamente liso, base de la espiguilla, con la gluma inferior fusion- ada a dicha dilatación, y antecio superior transver- salmente rugoso. Cabe destacar que además estos géneros son desde el punto de vista fisiológico PEP- Dentro de los géneros NAD de la tribu Paniceae, Panicum se separa de Scutachne Hitchc. & Chase, género endémico de las Antillas, por tener este úl- timo gluma superior y lemma inferior coriáceas, y de Yakirra y Arthragrostis Lazarides, géneros en- démicos de Australia, por incluir el primero espe- cies con un conspicuo entrenudo en la base del antecio superior y el último por poseer articula- ciones en la base de ramificaciones de la inflores- cencia, pedicelos y espiguillas. TRATAMIENTO TAXONÓMICO Panicum L. ESPECIE TIPO: Panicum miliaceum Linnaeus (lectótipo, de- signado por Hitchcock & Chase, 1910 secc. Panicum. P. secc. Miliaceae Stapf in Prain, Fl. Trop. Afr. 9: 640. 1920 Plantas anuales o perennes, cespitosas a arro- setadas y con cañas erectas a ocasionalmente de- cumbentes, simples a ramificadas; entrenudos hue- excepto en Vainas sin cos. aquarum. Lígulas membranáceo-ciliadas. Láminas aerénquima, oblongo-lanceoladas a filiformes, planas o con los bordes involutos. Inflorescencias terminales y axi- lares, laxas y difusas con ramificaciones divergen- tes del raquis, a espiciformes o subespiciformes, con ramificaciones adpresas al eje principal. Espi- guillas globosas u ovoides a largamente elipsoides, solitarias, no estipitadas a estipitadas, glabras. Glu- ma inferior (V,—)V,—/(—V.) del largo de la espiguilla, —5(7—9)-nervia. Gluma superior y lemma inferior subiguales, de igual largo o superando en longitud al antecio superior, (5—)7-9(-15)-nervias; gluma superior tempranamente caediza (en P. exiguum, P. furvum, P. peladoense y P. quadriglume). Pálea in- erior presente, de igual largo o mayor que el an- tecio superior, a reducida o ausente; flor inferior ausente. Ántecio superior crustáceo, liso a papiloso en toda su superficie, glabro, brillante a opaco, pa- jizo, grisáceo o castafio a negro a la madurez, ca- yendo desnudo a la madurez sin las restantes brác- teas; lemma envolviendo con sus bordes a la pálea, esta 2-nervia con papilas verrugosas o simples ha- cia el ápice; lodículas 2, truncadas, cuneiformes; estambres 3; estilos 2, estigmas plumosos. Cariopsis con hilo punctiforme a oblongo; embrión —V, del largo de la cariopsis, ocasionalmente mayor. Iv Plantas C,, del subtipo fisiológico NAD-me. Nümero básico de cromosomas x — 9. Sección con 31 especies en América, distribui- das desde Canadá hasta la Repüblica Argentina. Habitan en campos, en lugares secos. CLAVE DE LAS ESPECIES l. Antecio superior negro a la madurez; gluma su- perior caediza a la madurez, Re al descu- bierto el antecio l. ere superior pajizo a castaño a la madurez; luma superior no caediza a la madurez Espiguilla triflora, con una flor bond estéril, una Paus estan y una flor superior perfecta; antec UE con el dorso de la lemma hacia n gluma inferior g 2(1). 29, P. quadriglume 2. Espiguilla _ con una flor inferior estéril y na flor superior perfecta; antecio oe con el dorso de la De mma hacia lé a superior 3 . Plantas anuales; gluma inferior y y. del largo de la espiguilla ? exiguum 3. | Plantas perennes; gluma inferior y- Y del largo illa ^a qi = I. = P. furnum alto T 'encias e 7-22 cm de jargoi EN e de (2.5-)2.7— 330 3.7) mm de largo; América del Sur, en = lua. Brasil, Paraguay, Uruguay y la Argenti 27. P. em 5(1). Plant tas de ambientes acuátičos; vainas con aerénquima; cañas eee ramificadas n los nudos superiore P. aquarum 5. Plantas de ambientes secos a EE PEE OA Volume 83, Number 2 Zuloaga & Morrone 223 Panicum Subg. Panicum Secc. Panicum htimedos, vainas sin aerénquima; cafias simples a a ramificadas en los nudos me- dios y/o bas 6(5). Infiorescenci "ias s spiiormes a idis Pca con ramifica s de primer orden adpresas al eje princ ‘ipal 6. nflorescencias laxas, difusas a contraidas, con semper "iones de primer orden divergentes del eje pri (6). Espiguillas de 2-4 mm de largo; antecio su- rior con dos cicatrices basales fi. RR de (4.2- )5-6 mm de — iion con un anillo circular en la base |... 8(7). Espiguillas de 2-2.2 mm de largo; ms de —6(—8) cm de alto, arrosetadas, con inflores- cencias subincluidas en las vainas foliares . 23. P. mohavense 8. Espiguillas de 2.8—4 mm de largo; plantas de 0 c m ii alto, no arrosetadas, con inflores- as largamente exertas 9(8). Fapiguillas apro gluma inferior Y, del largo a espiguilla; pálea inferior reducida, Y, del largo del antecio superior; inflorescencias de 5— 6 cm de largo; plantas anuales; Estados Unidos de América y México 25. P. pampinosum 9. Espiguillas anchamente ovoides; gluma inferior ¥,~¥, del largo de la espiguilla; pálea inferior tan larga como el antecio superior; inflorescen- cias de 8—24 cm de largo; plantas dapes sur de Brasil a la Argentina . P. chasei 10(7). —— s sias paucifloras de 3—6 cm de lar- o; gluma inferior Y, o más del largo de la rip da de la gluma superior por un entrenudo conspicuo; Brasil 21. P. magnispicula 10. Inflorescencias multifloras de (5-)10-20 cm de largo; gluma inferior A o menos del iod de la espiguilla, sin un tre la gluma inferior y superior; México. ... 8. P. decolorans 11(6). Inflorescencias caedizas en conjunto a la ma- durez por ruptura del pedünculo 11. florescencias no caedizas 12(11). Plantas anuales; pees inferior : ausente a v tigial, hasta de mm de largo 5. P. capillare 12. Plantas perennes; isin inferior elfptica, de E e igual largo que el antecio 13(12). nh: y ramificaciones de la inflorescencia abros; vainas y laminas esparcidamente p losas a glabrescentes _.. 4a. P. bergii var. Laut 13. Raquis y ramificaciones de la inflorescencia, vainas y láminas densamente hfspidas ergii var. pilosissimum 14(11). sha robustas Leite de 3 m de alto; pelos urticantes ! y nm 14. pias menores hasta de 1.50 m de alto; ed ~ ras a hirsutas pero sin pelos ur- 15 NAM uu perennes |... 16 Plantas anuales 4 2 iom Espiguillas de (3.3—)3.6—6.4 mm de largo _. 17 16. Espiguillas de 1.8—3.1 mm de largo 20 17(16). Espiguillas de 5—6.4 mm de largo; m e su- perior y EUR inferior 3. "m mm más lar superior s ullae 17. Espizuillas p (3. do wr mm de larg d^ or y lemma inferior 0.8-1.5 mm ás la e el antecio superior - piam duda láminas basales rizadas |... ue var. ET 15a 18. t i no glaucas, láminas fre 18(17). 19(18). Espigullla con un estípite en la base; gluma inferior E pimp separadas por un entrenu- m del largo; gluma superior 7— 9-nervia; ees lineares a fi ye 0.103 c cho .. £ : 19. Espiguilla sin un estípite en la n gluma inferior y superi ior sin un entrenudo marcado; uma superior 9—15-nervia; ip ul linear- lanceoladas, de 0.4—1.2 cm de ancho ....... P lepidulum Láminas no glaucas; láminas Sae no ri- zadas; láminas densamente pilosas 20. Láminas glaucas; láminas basales ligera a C RM rizadas; láminas glabras a es- arcidamente hirsutas Paitin de 40-90(-120) cm de alto; cuello > nsamente piloso; hojas e ascenden- P. ghiesbreghtii )15-60 cm de 2d cuello gla- 9. 20(16). 21(20). PAP CEN de (6— bro; hojas patentes |... diffusum 22(20). Espiguillas de 2.7-3.1 mm de E espi- illas adpresas sabe las ramas de la in scencia |... a 22. Espiguillas de 1.8-2.8 mm d pss sobre las ramas de la inflo- 23(22). Espiguilla de 2.1-2.8 mm de largo; gluma largas que el antecio superior; láminas de 11-23 cm de largo por 0.2-0.4 cm de ancho 15b. P hallii var. e pod 23. Espiguillas de 1.8-2.2 mm de largo; g superior y lemma inferior tan largas como s antecio superior; pig de 20-35-50) cm de largo por 0.50.8 c ancho |... CABRA 31. P. tamaulipense 24(15). Antecio superior - obovoide a la madurez, cor- tamente estipitado, con 2 expansiones car- nosas en la e la lemma 0 24. Antecio superior ovoide a elipsoide, no es- > sin expansiones basales, con 2 cica- trices basales o con un anillo en la base de Espiguillas de 3.64.5 2. de lar alarum v var. v onim Espiguillas de 2. DR. 3) mm de lar 26 Antecio superior fuertemente oe en a su superficie, opaco 25(24). Espiguillas de 3.6-4.5 mm de largo — 25. 26(25). alatum var. alatum 20. Antecio superior liso, ie con papilas hacia el ápice de la p x . P. alatum var. minus Espiguillas de 4. 561 mm E largo 2 Espiguillas de 1.6—4 mm de largo Espiguillas estipitadas, con un conspicuo en- trenudo entre la gluma superior y lemma in- ferior; antecio superior ium caedizo; panojas erectas 26. P. parcum Fapiguillns no estipitadas, sin entrenudos de la raquilla consipcuamente marcados; ante- wi Eo Ny e Annals of the Missouri Botanical Garden 29(21). 30(29) 32(31). cio superior tempranamente caedizo; panojas itantes € nu > miliaceum 'anojas parcialmente incluidas en las hojas superiores, con las ramificaciones inferiores verliciladas; antecio superior con dos cica- trices basales conspicuas, ca. 0.6 mm de lar go lo. P. nr Panojas exertas, con — aciones inferio- res opuestas o alternas; antecio superior con dos cicatrices basales, menores de 0.6 mm de largo o con un anillo en la base de la lemma 30 Pálea unos M YY, del anno del antecio supe p Pálea inferior de ida, Y, o menos del largo del antecio = o ausent Láminas lineares, de 23 cm ao largo, 0.2 cm de ancho; S.L adpresas en las ra- mas; lemma inferior 5-nervia; antecio supe- rior con un anillo en la base de la lemm na P. ephemeroides Áminas lanceoladas a linear-lanceoladas, de AS 0.3—1.5(-2) em de bns espiguillas difusas sobre las ramifica- => T 5 em de largo, an- 1-1 pud antecio superior con dos cicatrices en la ciones; lemma inferior lemma o con un anillo en P hispidifolium. con espiguillas globosas y qum inferior Y, del largo del antecio superior : Espiguillas ovoides; gluma pa y lemma inferior | mm más largas que i perior; láminas lanceoladas el antecio su- 24. P. mucronulatum Espiguillas globosas; gluma epa y lem- ma inferior tan largas como el antecio su- perior o hasta 0.4 mm más jos láminas linear-lanceoladas 3 wo Gluma inferior /,—/, del largo de la espigui- lla; gluma superior 9-1 1(-13)-nervia; pálea inferior igual o más larga que el antecio su- perior; espiguillas de 2.3— 3.2 1 mm ue largo 1 stramineum Gluma inferior Y,—Y, del largo de bi espig lla; gluma superior 7-nervia; pálea inferior w =~ A del largo del antec 10 Hn capu: > hi si ium C luma superior y lemma vod tan larg no el antecio superior o jo 0.3 mm más 28. : 3.2-3.7 mm de largo bs gas P. philadelphicum Gluma superior y lemma inferior 0.8—1.6 mm más largas que el antecio superior 35 "álea inferior ausente; gluma inferior YY, del i argo de a espiguilla; antecio Meu con un Sdn en la base de la lemma D. flexile %álea inferior ,—V, del b del antecio su- perior; gluma inferior YY, del largo de la espiguilla; antecio superior con dos cicatri- ces en la base de la lemma, de ca. 0.2 mm de largo : Antecio superior conspicuamente papiloso opaco 8b. P hirticaule var. verrucosum Antecio superior liso, brillante, con papilas junto al ápice de a ey . P. hirticaule var. hirticaule 1. Panicum alatum Zuloaga & Morrone, sp. nov. TIPO: México. Baja California Sur: 15.5 mi. S of El Arco, 27°45'N, 113°20'W, 23 Oct. 1959, Wiggins 15160*,** (holótipo, MO; isótipos, GH, LA, MEXU, SI). Figura 15. Panico hirticauli affine sed lemma superiore base bia- lata et anthoecio superiori obovoideo et stipitato differt. de 13-95 cm de alto. Cafías geniculadas a erectas, ramificadas en Plantas anuales, cespitosas, los nudos inferiores; entrenudos de 1.5-10 cm de largo, cilíndricos, los inferiores comprimidos, gla- bros a híspidos en la porción superior, con tintes violáceos; nudos glabros a cortamente pilosos. Vainas de 2-7 cm de largo, glabras a hirsutas, con pelos caedizos, un borde papiloso-pestañoso en toda su extensión o sólo hacia la porción distal, el restante glabro. Lígulas membranáceo-pestañosas, de 0.7-1.8 mm de largo; cuello glabro. Láminas de 2-18 cm de largo, 0.3-1.7 cm de ancho, más o menos erectas lanceoladas a linear-lanceoladas, ascendentes, planas, glabras o con la cara abaxial esparcidamente pilosa, de base redondeada y ápice agudo, con los bordes finamente escabriúsculos, los basales pestañosos. Pedúnculos hasta de 30 cm de largo, glabros a esparcidamente pilosos. Inflorescen- cias terminales, laxas, difusas, de 3.5-23 cm de lar- go, 1.5-11 em de ancho; eje principal y ejes de las ramificaciones glabros, escabriásculos; ramifica- ciones de primer orden alternas a subopuestas, di- vergentes; ramificaciones de segundo orden adpresas con espiguillas apretadas sobre los ejes, próximas; pulvínulos glabros; pedicelos claviformes, de 14 mm de largo, escabriúsculos, cortamente pilosos ha- cia el ápice a glabros. Inflorescencias axilares pre- sentes, similares a las terminales. Espiguillas elip- de 2.4-3(4.5) mm de largo, 0.7-0.9(-1.5) ancho, de ápice largamente acuminado, sin soides, mm de estípite basal, verdosas a cobrizas o con tintes pur- púreos; gluma superior y lemma inferior subiguales, 0.70.9(-1.8) mm más largas que el antecio supe- rior. Gluma inferior ovado-acuminada, de 1.2-1.8(- 2.7) mm de largo, YY, del largo de la espiguilla, 5— 7-nervia, el nervio medio escabriúsculo. Gluma su- perior de 2.2-2.7(-3.3) mm de largo, 7-9(-11)-ner- via; gluma inferior y gluma superior separadas por un corto entrenudo de 0.3 mm de largo. Lemma inferior glumiforme, de 2.1-2.8(3.9) mm de largo, —9(-11)-nervia. Pálea inferior ausente o reducida, cuando presente anchamente lanceolada, de 0.2— 0.9 mm de largo, 0.2-0.3 mm de ancho, YY, del largo del antecio superior, de ápice bilobado à agu- do, membranácea, hialina, glabra. Antecio superior 8(2.5) mm de argo, 0.6-0.9(-1.2) mm de ancho, crustáceo, pajizo obovoide a la madurez, de 1.5-1. — Volume 83, Number 2 Zuloaga & Morrone 225 1996 Panicum Subg. Panicum Secc. Panicum Figura 15. Panicum alatum var. alatum.—A. Hábito.—B. Espiguilla vista del lado de la gluma inferior.—C. Espiel vista del lado de la gluma superior. D-F. Pálea inferior.—G. Antecio superior visto del lado de la lemma.— H. Ant superior visto del lado de la pálea, con expansiones carnosas hacia la base.—I. Cariopsis, vista escutelar.— J. Cariopsis, vista hilar. (A-H, Gentry 14353; 1, J, Hinton et al. 6423.) 226 Annals of the Missouri Botanical Garden a castaño oscuro a la madurez, liso, fuertemente papiloso en toda su superficie o con papilas presen- tes sólo hacia el ápice de la pálea, cortamente es- tipitado, con dos expansiones carnosas basales de 3 m de largo, castaño claras a la madurez; lemma a Cariopsis ovoide, de 1—1.2(-1.7) mm de largo, 0.6(-1) mm de ancho, blanquecino; hilo punctiforme; embrión Y, del largo de la cariop- sis. Esta nueva especie se distingue de Panicum hir- ticaule por tener antecio superior obovoide, corta- mente estipitado y con dos expansiones carnosas en su base. la. Panicum alatum var. alatum Espiguillas de 2.5-3 mm de largo. Gluma inferior 1.4—1.8 mm de largo. Gluma superior y lemma in- ferior 0.7-0.9 mm más largas que el antecio su- perior. Ántecio superior opaco, fuertemente papiloso en toda su superficie. Distribución y ecología. Sur de los Estados Unidos de América y norte de México (Fig. 3), en lugares abiertos, rocosos, ocasional en bordes de arroyos en lugares húmedos, desde el nivel del mar hasta los 1220 m Material wl id examinado. ESTADOS UNIDOS DE AMERICA. Arizona: Phoenix, 25 Sep. 1904, Griffiths 7317 (US). California: overflow land along Colorado Riv- er, p a 1905, Schellenger 3 (LA). Sin estado: E of Desert , 16 Sep. 1929, Jones 24493 (MO); Gila bottom, Sl 16 (MO). MEXICO. Baja California Sur: Purísi- ma, 17 Feb. deis Brandegee 8 a JA, US); 12 mi. S of El Solito, 2099. 3°48'W, : . 1959, Wiggins 15135 , MO); T mi. S of Millers landing, 19 Oct. 1959, Weems 15097 (GH, ISC); . SW of jet. of road to Huatamote along Hwy. Mex. ? L 365 m, 29 ene. 1977, Reeder & eur 6718 (MEXU); Cuesta de Chuenque, ca. 22 km S of rs on road to een Esc cie 255 UN, beue 10 (€ 9 Carter & Ferr XU). Durango Ceballos, Zona del seni 10, p m, e: Oct. 1978, Martínez 1313 (MEXU). rrero: un Coyuca, Cutzamala, 8 Sep. 1934, Hinton et 1 al. 64 3 (GH, NY, US); Cutzamala, 8 Sep. 1934, (G Apatzingán, 4 mi. W of Apatz zingsán, 1200 1941, Lcavenwanh & Hoogstraal 1371* (M yarit: : mi. SE of Las Vacas, 60-90 m, 21 es 1960, McVaugh 19291 (US). San Luis Potosí: 12 mi. E of Ciudad Valles 17 July 1963, Mc icd et al. 840 (US); 2 km E of Santa Catarina Ys Hwy. 70, 1220 m, 23 Oct. 1978, Reeder & Reeder 7045 (LA). Sinaloa: El Quelite cut-off to Rt. 15, 34 mi. S of Elota, 10 Sep. 1962, Skorepa 116* (MO, WIS), pr. p.; Cerros Navachiste, n in Topolobampo, 26/ 30 Sep. 1954, Gentry 14353*,** (Sl); outskirts of Topo- gone 9 Oct. 1966, Gould 12110* (GH, US). Sonora: ni. S of na 215 red 7 Wet & tte 22: (MEX, MO); 35.8 mi. SE of Guaymas along 15, 22 Sep. 1974, de & d 4732 (MEXU); du. mas, 17 Sep. 1908, Hitchcock 3553 (US), 1887, Palmer 208 (LA, NY, US), 346 (US); near Guaymas, on rocky, Mante y Tampico bes of Granja Trinidad, Au 1971, Beetle et al. M-1211 (ISC, MO). Sin der ud 1897, Palmer 249 (US), 251 (GH, US). lb. Panicum alatum var. longiflorum Zuloaga & Morrone, var. nov. TIPO: México. Baja Cali- fornia Sur: Mesa del Potrero de San Javier (NE of Misión San Javier), ca. 111°32.5'W, 19 Sep. 1965, Carter 4970*,** (holótipo, MO; isótipos, GH, LA, MEXU). Panico alato var. alato affine sed spiculis 3.6-4.5 mm longis, anthoecio laevi non papillato differt. Espiguillas de 3.6—4.5 mm de largo. Gluma in- ferior de 1.9-2.7 mm de largo. Gluma superior y lemma inferior 1.2-1.8 mm más largas que el an- tecio superior. Antecio superior liso, brillante, con papilas presentes sólo hacia el ápice de la pálea. Distribución y ecología. Ocasional en el sur de los Estados Unidos y México (Fig. 10), en ambien- tes abiertos, secos. Nombre vulgar. “Zacate” (en México). Material adicional sai ea a UNIDOS DE AMERICA. California: ha, 4 July ^ by 6 (US). MEXICO. Baja California Sur: ek Teombé, N of Portezuelo Gabilán, ca. 25?51'N, 111%25'W, 1 Oct. 1965, Carter 5074** (GH, LA, MEXU, MO); Sierra de la Giganta, Carter 5548 (GH, Encinal, S side of Valle de los Encinos, 111°34’W, 1 Oct. 1967, Carter 5374 (GH, LA Sierra de la Giganta, Cerro de Barreno, S side of Valle de Los Encinos, ca. 26°03'N, d 1200 m, 30 Sep. 1967, Carter 5353 (LA, MEXU, SI). Sonora: 12 mi. W of Hermosillo, 27 Aug. 1941, Wiggins & Rollins 127 (MO, US); sin localidad, Hitchcock 3541 1/2 (US le. Panicum alatum var. minus (Andersson) Zu- loaga & Morrone, comb. nov. Basónimo: Pan- icum hirticaule J. Presl var. minus Andersson, Kongl. Vetensk. Acad. Handl. 1853: 135. 1855. TIPO: Ecuador. Archipiélago de Colón: mayo 1852, Anders- son s.n. (holótipo, no visto; isótipo, GH, proba- bles isótipos, MO-2098647, NY). “In insula Indefatigable,” Panicum hirticaule J. € var. majus Andersson, Kongl. ensk. Acad. Handl. 1853: 135. LL TIPO: Ecuador. Arc i de Colón: *Hab. in insulis Chatham et Charles," mayo 1853, yes ee s.n. (holótipo, no visto; isótipo, GH). Espiguillas de 2.4—3(-3.2) mm de largo. Gluma inferior de 1.2-2.1 mm de largo. Gluma superior y lemma inferior 0.4—0.8 mm más largas que el an- tecio superior. Antecio superior liso, brillante, con papilas hacia el ápice de la pálea. Volume 83, Number 2 1996 Zuloaga & Morrone 227 Panicum Subg. Panicum Secc. Panicum Distribución y ecología. Ampliamente distri- buida en el sur de los Estados Unidos y norte de México (Fig. 5), donde crece en suelos arcillosos o arenosos; común a lo largo de bordes de caminos. Es ocasional en El Salvador, Nicaragua y Venezuela y también se halla en las Islas Galápagos, Ecuador, donde crece en suelos rocosos. Habita desde nivel del mar hasta los 1860 m. Material adicional examinado. ECUADOR. Archi- — GH); Espafiola, landing site on north coast and area to El Chaco, 3 July 1986, Lawesson 3143 (MO); Indefatigable Island, Academy Bay, 4 May 1932, Howell 9076 (F, GH, MO, NY); Gardner Island, near Hood Island, 22 Apr. 1932, Howell 8728 (GH, US). EL SALVADOR. La Unión: La Unión, 13 Nov. 1911, Hitchcock 8794 (US). Sonsonate: Acajutla, 10 ago. 1951, Rohweder 1648 (MO). ESTADOS UNIDOS DE AMERICA. Arizona: San Ber- nardino Ranch, 28 Aug. 1892, Mearns 738 (US); Tucson, 2 Sep. 1903, Thomber 176 (LA, MO); Santa Cruz Valley near Tucson, 20 June 1881, Pringle 13940* (MO); Pata- gonia, Santa Cruz County, Hitchcock 3675 (US); Yavapai Co., near Kirkland, 5 Oct. 1930, Peebles et al. 7417 (LA). alifornia: bed of Fort Dry Lake, 24 mi. E of Blythe, 350 ft., 30 Nov. 1937, Grinnell & s 1082a (JEPS, al E of Desert Center, 16 Sep. 1929, Jones s.n. (LA 406 New Mexico: Gila, 29 Aug. 1880, Greene 13107 As plains of Río Gila, 29 Aug. 1880, Greene 258 (US); Doña Ana County, Mesilla Valley, 11 Oct. 1904, Wooton s.n. (MO-1314437). Texas: El Paso, 23/24 Aug. "1915, Hitch- cock 13333 (US); sin localidad, Nealley s.n. (US). HON- URAS. Valle: 3 km N of Jícaro Galán along highway 1 to Tegucigalpa, 5 Oct. 1986, Davidse & Pilz 31651 (MO, SI); savanna on heavy black clay, heavily grazed and dis- turbed, 3 airline km W of San Lorenzo along the road to Amapala, 5 Oct. 1986, Davidse & Pilz 31704 (SI). ME- Sierra Judrez, 2 Oct. Caquihui, Sierra de la Giganta W of Lo 110°52’W, 17 Nov. 1959, ai 15510 (GH); SW of San Ignacio, 27°14'N, 112°53’W, 25 Oct. die ir 15179 (LA); pa de la Giganta, Arroyo at NW of C N, 110°44’W, 2 Nov. 1971, Moron 18878 1/2 h cárcega in disturbed area, 21 Dec. 1972, Reeder & Reeder 6098 (MO). Chiapas: Chiapa de Corzo, El EL 5.6 ong Mexican highway 190, , 1 July ad rwr d 10630 i. NW of Mexico-Guatemal r on the ed to Comitán, 900 m, 23 Aug. ee Reeder & Reeder 2082 (LA, MEXU). Chihuahua: Chihuahua, 4800 ft., Soder- strom 896 (US); Meoquí, 24 ago. 1935, Le Sueur 032 (US); Llano de Chilicote 7 mi. E of Chilicote Station, 26 Sep. 1938, Johnston 7992*, ** (GH, US). Coahuila: El Mes- teño, La 4 de Marzo, 28 July 1976, Roig s.n. (ANSM). Guerrero: 6 km al W de Colotlipa, 20 Aug. 1978, Blanco et T 512 (MEXU). Jalisco: Guadalajara, Río Grande de ? of Guadalajara, 5200 ft., 26 Aug. 1941, 5 (MO); Puente Grande, 13 mi. SE of Guadalajara, on road to Aguascalientes, 5000 ft., 17 Sep. 1959, Soderstrom 647 (US); Tecolotlán, 1500 m, 5 Oct. 1970, Díaz Luna 2044 (MEXU). Michoacán: Apatzingán, pes at La Majada, 5 mi. W of Ape ee 000 . 1941, — & Leavenworth 1335 (US); Cherinds. 277 Sep. 1946, Hernández Xolocotzi et al. X-2791 (US); ca. 5 mi. 'N of Cotija and 22 mi. S of Jiqauil- LA, NY, Walther 63 (US); 12 km NW of A 500 Aug. 1950, Lyons Jr. 59 (US) 83 (MEXU, d Oaxaca: Tomellin, 14 Aug. 1910, Hitchcock s.n., Amer. Gr. Hb. 28 (LA, US); 12 km NE of the city of Teri bo y the Pan-American nod ing 185 & 190), 50 m, 7 July 1958, King 391 (US Luis Potosí: Palma Court, Valles, 8 Aug. 1961, ihe 9558 (ISC, LA, MO, US). Sina- loa: Culiacán, 27 Aug./15 Sep. 1891, Palmer 1545 (LA, NY, SI, W), about 15 mi. SE of Villa Unión, in disturbed area along the road, 23 Sep. 1953, 90 m, Reeder & Reeder 0 (LA, US); vicinity of Labradas, 18 Sep. 1925, Ferris & Mexía 5070 (M = a ir 27 Aug. 1891, Palm- er 1554 (F, NY). Sonora: i. S of Hermosillo, 17 Sep. 1908, Hitchcock 3541 (US); es 17 Sep. 1908, Hitch- cock 3547 (US); Sres Junction, Pinacate Region, 7.5 mE o Vidrios, 31°59’N, 113?21'W, 215 m, 14 Sep. 1986, Felger 86- 339 MEX). Tamaulipas: en los alrede- dores de Ciudad Mante, Sánchez-Ken & Nieto M. 430 (MEXU). Veracruz: Puente Nacional Veracruz, carretera Jalapa—Veracruz, Emiliano Zapata, 250 m, 10 ago. 1982, Dorantes et al. 1510 (F, MEXU). Sin estado: sin localidad, Palmer s.n. (US). NICARAGUA. León: Puente Izapa, km 60 carretera a León, 12?15'W, 86°44’W, 13 jul. 1981, Mo- reno 9838 (MO). VENEZUELA. Zulia: Guarero, Guajira, 25 dic. 1954, Vareschi 3745 (VEN). T El ejemplar Wiggins 15179 comprende plantas sumamente pequefias, de 1.5 a 6 cm de alto. Los ejemplares Hitchcock 8794, Davidse & Pilz 31651, 31704, incluidos en esta variedad, se dis- tinguen por poseer un rizoma en su base, consti- tuyendo los ünicos especímenes aparentemente perennes de esta especie. Reeder & Reeder (1971) ubican a P. hirticaule var. majus y var. minus Andersson en la sinonimia de P. hirticaule. 2. Panicum aquarum Zuloaga & Morrone, No- von 1: 185, figs. 1, 2. 1991. TIPO: Venezuela. Guárico: Dpto. Infante, Parque Nacional Aguaro-Guariquito, Congriales de la Gorra, entre La Esperanza y Mesa de Cambao, ca. 9?12'-9?16'N y 67?48'—61?60' W, 60 m, dic 1981, Delascio, Montes & Davidse 11206 (hol- ótipo, VEN; isótipos, MO, SI). Figura 16. Plantas anuales?, de 0.90—1.30 m de alto. Cañas erectas, arraigadas en los nudos inferiores y mani- fiestamente ramificadas en los nudos superiores; entrenudos de 7-19 cm de largo, glabros, estriados, huecos; nudos glabros, comprimidos, castaños. Vainas de 8-18 cm de largo, delgadas, glabras, lus- trosas en la cara adaxial y con aerénquima, los márgenes membranáceos. Lígulas ca. 1.2 mm de 228 Annals of the Missouri Botanical Garden Figura l6. Panicum aquarum.—A. Há bito.—B. E spiguilla, vista lateral.—C. E spiguilla vista del lado de la gluma inferior. —D. A A boda vista del lado de la gluma superior. —E. Pálea inferior. —F. Antecio superior visto del lado de a lemma.—G. Antecio superior visto del lado de la pálea. (Blydenstein I 8 16.) i H. Cariopsis, T escutelar.—l. Cariopsis, vista hilar. Volume 83, Number 2 1996 Zuloaga & Morrone 229 Panicum Subg. Panicum Secc. Panicum largo, cortamente membranáceas en la base y lar- gamente ciliadas en el ápice, la cara adaxial pilosa; cuello glabro. Láminas lineares, de 11-18 cm de largo por 0.2—0.4 cm de ancho, las inferiores de- ciduas, glabras, angostadas hacia la base, los már- genes escabriúsculos, con nervio medio manifesto en la zona basal. Pedúnculos hasta de 25 cm de largo, escabriúsculos. /nflorescencias terminales, piramidales, laxas, multifloras, de 35—40 cm de lar- go por 12-20 cm de ancho; eje principal cilíndrico, glabro; pulvinulos glabros; ramificaciones de pri- mer orden alternas o subopuestas, divergentes; ejes de las ramificaciones triquetros, escabriúsculos; ramificaciones de segundo y tercer orden alternas con ejes escabriúsculos; espiguillas congestas sobre las ramificaciones de tercer orden; pedicelos clavi- formes, glabros. Inflorescencias axilares numerosas, menores, similares a la inflorescencia terminal. Es- piguillas solitarias, elipsoides, de 3-3.9 mm de lar- go, 1-1.2 mm de ancho, estipitadas, glabras, paji- zas y con tintes violáceos, abiertas a la madurez; gluma superior y lemma inferior subiguales, acu- minadas, superando en largo al antecio superior. Gluma inferior ovada, abruptamente acuminada, de 1.7-2.7 mm de largo, Y, a Y, del largo de la espi- guilla, (3—)5-nervia, el nervio medio escabriúsculo hacia la porción superior; entrenudo conspicuo en- tre la gluma inferior y la gluma superior. Gluma superior de 3-3.6 mm de largo, acuminada, (5—)7- nervia. Lemma inferior glumiforme, de 3-3. de largo, (5—)7-nervia. Pálea inferior oblongo-lan- de 1.7-2.1 mm de largo, 0.6-1 mm de ancho, tan larga a ligeramente menor que el antecio mm ceolada, superior, glabra, hialina. Antecio superior elipsoide, mm de largo, 1.2-1.3 mm de ancho, crus- táceo, pajizo, con manchas negruzcas a la madurez, con un anillo castafio en la base; lemma superior 5-nervia, los márgenes involutos; pálea superior 2- nervia, con papilas verrugosas hacia el ápice. Ca- riopsis anchamente elipsoide, 1.5 mm de largo, 1 mm de ancho, blanquecina; hilo subbasal, puncti- forme; embrión Y, o menos del largo de la cariopsis. Distribución y ecología. Restringida al extremo norte de América del Sur, Brasil, Colombia y Vene- zuela (Fig. 6). Habita en lugares inundados; florece y fructifica entre septiembre y mayo. Material adicional examinado. BRAZIL. Amazonas: Mun. Humaitá, 7°31'S, 63*10' W, pn a — 3- ES t 5 > 3 - se Y S a x A A 5 Un P4 i= o lac] Francis, Steller & Jacobs 335 (LIL). Kentucky: Doyle Valley, W of Chaumont Road, 1 Oct. 1949, Lix 738 (US). Maine: York Co., Occan Park, 2 Sep. 1931, Moldenke 6377 (US). Maryland: near Potomac, NW of Cabin John, 5 Oct. 1904, Chase 2726 (US). Massachusetts: Harwich, 29 July 1919, Fernald & Long 17828 (US). Missouri: 2 A NW of Holden, Steyermark 72470 (LIL). Montana: mi. above Blendive, 20 July 1883, Ward s.n. (US- da Nebraska: North Platte, June 1896, Plank 38 Somerset Co., Watchung, 29 Aug. Creek, Metcalfe 1138 (LIL). New York: Princes Bay, 6 Oct. 1894, Kearney Jr. s.n. (US-742048). North Caroli- na: sin localidad, 1885, McCarthy s.n. (US-952946). North Dakota: Morton Co., roadside between Mandan and Bismark, Muñoz & Rollins 1006 (US). Ohio: Cincin- nati, Spring Grove, 21 Aug. 1905, Braun s.n. (US- 2663130). Oregon: MacKenzies bridge, 1740 ft., 21 July 1927, Hitchcock 23470 (US), Roseburg, 14 July 1908, Hitchcock 2798 (US). Pennsylvania: Easton, 1894, Porter s.n. (US-480434). South Dakota: Belle Fourche, 4 Aug. 1897, Griffiths 399 (US). Texas: New Braunfles, 20 June 1910, Hitchcock 5197 (US). Vermont: Passumpsic River, Norte: E of San Telmo, Rancho San José, east of San Telmo, 14 Sep. 1930, Wiggins & Demaree 4833 (F, NY, US). Coahuila: Municipio de Villa Acuña, Serranfas del Burro, Rancho El Bonito, ca. 29?01'N, 102%07'W, 1620 m, 18 sep. 1977, Valdés Reyna y Riskin 1230 (ANSM). URUGUAY. Montevideo: Montevideo, Felippone s.n. (SI). Gould (1979) indica que P. capillare fue citado por St. Vincent, en las especímenes de esta especie en las Anti Hitchcock (1951) y Davidse (1987) mencionan que en esta especie la dispersión se produce a tra- vés del conjunto de la panoja, la que es llevada por el viento a la madurez. La inflorescencia tiene en esta especie más de la mitad del largo de la planta. ntillas, pero "E no halló 6. Panicum capillarioides Vasey in Coulter, Contr. U.S. Natl. Herb. 1: 54. 1890. TIPO: Es- tados Unidos de América. Texas: Point Isabel, Nealley 634 Dom US-76925; probable isótipo, US-953165). Plantas perennes, cespitosas. Cafias erectas de 0—75 cm de alto, simples o ramificadas en los nudos inferiores; entrenudos ae die huecos, pilosos o glabros, de 1— iám., nudos pilosos. Vainas iguales o menores que los entrenudos, hir- sutas, verdes o con tintes violáceos, los márgenes pestafiosos. Ligulas membranáceo-ciliadas, la por- ción membranacea de 0.5 mm de largo, cilias de 1.5 mm de largo; cuello piloso. Láminas linear-lan- ceoladas, de 12-30 cm de largo, 0.5-0.8 cm de an- cho, erectas, ascendentes, planas, redondeadas en la base, el ápice atenuado, hirsutas, con pelos tuber- culados, a esparcidamente pilosas, escabrosas, el nervio medio manifiesto. Pedúnculos subincluidos a exertos, hirsutos. Inflorescencias terminales exertas; panojas laxas, difusas, de 15—30 cm de largo, 10— 12(-26) cm de ancho; eje principal anguloso, híspido hacia la base a glabro, escabriúsculo; ramificaciones de primer orden alternas a subopuestas, las i riores ocasionalmente subverticiladas, de 8-18 cm de largo, divergentes, desnudas hacia la base, ram- ificaciones de segundo orden divergentes; espiguillas por rama 1-2(-3); pulvínulos cortamente pilosos; pedicelos de 2-20 mm de largo. Espiguillas solita- rias, largamente ovoides, de 5-6.4 mm de largo, 1— 1.2 mm de ancho, no estipitadas, pajizas, con tintes violáceos, glabras, la gluma superior y lemma infe- rior subiguales, superando 3.2—4 mm en largo al an- tecio superior. Gluma inferior de (2-)2.5-3 mm de largo, casi Y, del largo de la espiguilla, 5—7-nervia, aguda a obtusa. Gluma superior de 5—6 mm de largo, aguda, 10—13-nervia; gluma inferior y superior se- paradas por un conspicuo entrenudo ca. 0.4 mm de largo. Lemma inferior glumiforme, aguda, 10-12- nervia. Pálea inferior de 1.5-2 mm de largo, tan larga como el antecio superior, membranácea, los márgenes escabrosos. Antecio superior ovoide, de 1.6-2 mm de largo, 1-1.1 mm de ancho, liso, pajizo, 236 Annals of the Missouri Botanical Garden con tintes castaños a la madurez, lustroso, crustáceo, con 2 cicatrices basales, castañas a la madurez, ca. 0.2 mm de largo, lemma 5—7-nervia, pálea con papi- las verrugosas en su ápice. Cariopsis de 1.5 mm de largo, 1 mm de ancho; hilo oblongo; embrión V, a Y, del largo de la cariopsis. Distribución y ecología. Sur de los Estados Unidos de América y N de México (Fig. 5); crece en lugares abiertos sobre suelos arenosos; desde el nivel del mar hasta los 1000 m; florece y fructifica entre abril y diciembre. Material adicional examinado. MCA UNIDOS DE AMERICA. Texas: Frio Co., 24 June 1941, Tharp 43066 (Moy Kleberg Co., a N of Riviera, off Hwy. 77, Jet. 1946, Lundell & Lundell soa 3 (MO); 9 miles Ru Raymondville, Gould 11456 (US); El Toro Isla, z A de la Laguna Madre, Tharp 49042 (MO, S); Kingsville, diu r s.n. * (US-558794); W of Kingsville, Swallen, 10254 ei idis ia s Division, nue Ranch, Swal- (US); 1 mile W at — c "d s E M en Y x Aa T N * = a Hitchc ui al S = ieur 1000 m, Rede osky 4618 (US). auli nda Buena VE Wooton s.n.* (US- 1061806); od Swallen 1642, 693 *(US): sin locali- dad, 2 dic. 1930, Viereck 857* (U E —. Panicum capillarioides se distingue claramente del resto de las especies de la sección Panicum por poseer espiguillas largamente ovoides, con la gluma superior y la lemma inferior superando entre 3.2 y 4 mm al antecio superior. 7. Panicum chasei Joni Pool B. R. Arrillaga & Izaguirre, Bol. . Agron. Univ. Montevideo 103: 9, fig. 2, de TIPO: Uruguay. Rocha: 3 feb. 1967, en costado del camino en suelo arenoso-arci- lloso, Rosengurtt 10804 (holótipo. MVFA r visto; isótipos, BAA, K, P, SI, US-2946507 e Ruta 9, Fortaleza Santa Teresa, Plantas perennes, de 30-80 cm de alto. Cafías erguidas, simples, paucinodes; entrenu- cespitosas, dos cilíndricos de 3-13 cm de largo, cubiertos de pelos tuberculados tiesos a esparcidamente pilosos; nudos violáceos, pilosos, con pelos blanquecinos adpresos. Vainas de 3-12 em de largo, usualmente menores que los entrenudos, cubiertas de pelos tuberculados. Lígulas membranáceo-pestañosas, de 0.7-1.8 mm de largo; cuello castaño claro, piloso. Láminas linear-lanceoladas, planas o con los bor- des involutos, erectas, ascendentes, rígidas, de 9— 25 cm de largo, 0.2-0.6 cm de ancho, angostadas en la base, con pelos tuberculados en ambas caras, más abundantes en la cara adaxial, el ápice acu- minado, los bordes escabriúsculos, ciliados. Pedún- culos largamente exertos, cilíndricos, híspidos, de 10-31 em de largo. Inflorescencias espiciformes a subespiciformes, de 8-24 cm de largo, 1-5 cm de ancho, con las ramificaciones adpresas al raquis, en ocasiones ligeramente divergentes; eje principal glabro, liso, anguloso; pulvínulos esparcidamente pilosos a glabros; ramificaciones alternas, las es- piguillas adpresas sobre los ejes; pedicelos esca- briúsculos, claviformes, de 1-5 mm de largo. Es- piguillas anchamente ovoides, de 2.8-3.8 mm de largo, 1.1-1.5 mm de ancho, acuminadas, no estip- itadas, glabras, pajizas y con tintes violáceos; glu- ma superior y lemma inferior subiguales, iguales o superando hasta en 0.3 mm al antecio superior. Gluma inferior de 1.5-2.4 mm de largo, Y, a Y, del largo de la espiguilla, 5—7-nervia, nervio medio es- cabriúsculo, cortamente pubescente en la cara in- terna. Gluma superior de 2.8-3.5 mm de largo, 7— 9-nervia, nervio medio escabriúsculo o no hacia el ápice, cortamente pilosa en la cara interna; estípite ausente. Lemma inferior de 2.7-3.5 mm de largo, glumiforme, 9-nervia. Pdlea inferior elíptica, de 1.8-2.6 mm de largo, 0.9-1.2 mm de ancho, tan larga como el antecio superior, membranácea, fin- amente denticulada o no en los bordes superiores, con el ápice bilobado a entero. Antecio superior elipsoide, de 2-2.6 mm de largo, 1-1.3 mm de an- cho, crustáceo, glabro, liso, lustroso, pajizo, castaño a grisáceo a la madurez, con dos cicatrices basales, de 0.15 mm de largo, castañas a la madurez; lemma superior 7-nervia, pálea superior con papilas sim- a ples en su ápice. Cariopsis elipsoide a ovoide, « 1.4-1.6 mm de largo, 0.8-1.1 mm de ancho; hilo punctiforme; embrión menos de la mitad del largo de la cariopsis. Distribución y ecología. | Argentina, Brasil y Uruguay (Fig. 9). Vive en campos; florece desde octubre hasta mayo. Material | adicional ~ ARGENTINA. Corrientes: Estancia Rincó Ambrosio, Schwarz 9595 (LIL, LP). Entre Rios ^ va Escocia, Cordini 20 BAA). : ASIE, Rio Gr Do Sul: * d 5* (US); Belem vlla nin 151 (BAA Pelotas, da 527 (BAA); e ruz Alta y P 3 ids 1962, Rosengurtt & Del nos ani (B AA, SI, US); o Alegre, Morro da Gloria, Valls 1646 (CEN); Fazenda dida de de Criagáo bese Swallen 7467 (US). URU- GUAY. Rivera: Subida de Pena, Sierra de la Aurora, Ro- sengurtt B-7128 (US). Salto: San Antonio, Orihuela 92 (US), s.n. (US-1648463). Soriano: Estación Juan Jackson, ~an B-242 (US cal Camino a Rivera, 32 km de Tacuarembó, 8 feb. 1981, Cabrera & Zuloaga * (SI, US). 3042344 Especie afin a P. ghiesbreghtii y P. peladoense. De la primera se separa por poseer panojas con- traidas, espiguillas generalmente de mayor tamaño y por tener láminas menores. Panicum peladoense Volume 83, Number 2 1996 Zuloaga & Morrone 237 Panicum Subg. Panicum Secc. Panicum difiere de P. chasei por sus panojas laxas, con es- piguillas con antecio superior negro a la madurez y gluma superior caediza a la madurez de las es- piguillas. Panicum diffusum Sw., especie que crece en las Indias Occidentales, se diferencia de P. chasei por poseer laminas lineares hasta de 0.2 cm de ancho, panojas laxas y espiguillas menores, de 2-2.9 mm de largo por 0.8-1 mm de ancho, con pálea inferior reducida. El ejemplar Gallinal et al. 4473 se aparta de los restantes examinados por tener inflorescencias más laxas, con ramas divergentes del raquis. . Panicum decolorans Kunth Bonpland & Kunth, Nov. Gen. 1816. TIPO: Querétaro: temperatis, apricis regni Mexicani prope Que- rétaro, alt. 995 hexap., Humboldt s.n. ** (holó- tipo P, fragmento y foto, US-80663; isótipo, P). oo in Hu a 1 México. nl in Plantas anuales, Cañas erectas de (15-)30-90 cm de alto, geniculadas o no, simples o ramificadas en los nudos inferiores, paucinodes; entrenudos de 6-12 cm de largo, cilíndricos, hue- cos, glabros, hirsutos hacia la porción superior a glabrescentes; nudos cortamente pilosos. Vainas de 3.5-7 cm de largo, comúnmente más cortas que los entrenudos, hirsutas, con pelos tuberculados cad- ucos, o glabras, los márgenes cortamente ciliados en toda su superficie o sólo hacia la porción su- perior, ocasionalmente con un margen ciliado y el restante glabro. Lígulas membranáceo-pestañosas, de 0.6-2 mm de largo; cuello glabro. Láminas lan- ceoladas, de 4.5-18 cm de largo, 0.5-1.5 cm de ancho, planas, de base redondeada y ápice agudo, glabras a hirsutas en la cara abaxial, ocasional- cespitosas. mente hirsutas en toda la superficie, con pelos tub- erculados de 1-1.5 mm de largo, los márgenes ba- sales ciliados con pelos caducos. Pedúnculos angulosos, hasta de 27 cm de largo, glabros a hir- sutos hacia la porción superior. Inflorescencias ter- minales exertas, largamente pedunculadas, multi- floras, espiciformes, de (5-)10-20 cm de largo, 5-1(-2) em de ancho; eje principal anguloso, escabriásculo; pulvínulos glabros; ramificaciones de primer orden adpresas al eje, ascendentes, en ocasiones la inferior algo divergente, alternas, con espiguillas congestas sobre ejes de tercer o cuarto orden; pedicelos claviformes, escabrosos, trique- tros, de 1. mm de largo. Espiguillas largamente ovoides, de (4.2—)5—6 mm de largo, 1.8-2 mm de ancho, no estipitadas, glabras, pajizas, con tintes violáceos; gluma superior y lemma inferior subi- guales o la gluma superior algo menor que la lem- ma inferior, 1-2 mm más largas que el antecio superior. Gluma inferior ovada, de 1.8-2.7 mm de largo, Y, o algo menor del largo de la espiguilla, aguda a obtusa, abrazando en la base a la gluma superior, 5—7-nervia, el nervio medio escabriúscu- lo. Gluma superior de (3.8—)5—5.9 mm de largo, aguda a acuminada, 10-13(-15)-nervia, la cara in- terna escabriúscula hacia el ápice. Lemma inferior glumiforme, 10—13-nervia. Pálea inferior reducida, de 1-1.5 mm de largo, /, o menor del largo del antecio superior, membránacea, obtusa, glabra. An- tecio superior ovoide, de 2.7-3.8 mm de largo por (1.1-)1.5 mm de ancho, glabro, pajizo a castaño, liso, lustroso, crustáceo, obtuso, fácilmente caedizo a la madurez antes de la caida de las glumas, la base con un anillo circular, castaño a la madurez; lemma 7-nervia; pálea con papilas verrugosas hacia el ápice. Cariopsis elipsoide, de 2-2.5 mm de largo or 1-1.5 mm de ancho; hilo oblongo; embrión Y, a Y, del largo de la cariopsis. Distribución y ecología. Endémica de México, crece en los estados de Baja California Sur, Chi- huahua, Coahuila, Guanajuato, Hidalgo, Jalisco, Michoacán, Oaxaca, Puebla, Querétaro y San Luis Potosí (Fig. 6). En flor entre mayo y noviembre; llega hasta los 1870 m. Material adicional examinado. MEXICO. Chihua- W of Carretas, 1675 m, 21 Aug. 1939, Harvey aju east of fiapuato, Sohns 210 (US). Hidalgo: Mun. Jacala, 4500 ft., mountain roadside, Chase 7090* (ANSM, GH, MO, US), 7270* (GH, US). Jalisco: swampy ground be- side Lake Chapala, near Tuxcueca, 4400 ft, 25 Aug. 1941, Leavenworth & Leavenworth 1839 (F, MO, US); 1 mile E of Tuxpán, on Mex 15, 1870 m, 20 Oct. 1976, Brunken & Perino 434 (MEXU, MO). Michoacán: Cha- vinda, Hernández Xolocotzi et al. 2793 (US). Oaxaca: km 568 of Panamerican highway, Baldwin Jr. 14321* (US). Puebla: Tehuacán, 5500 ft., 9 Aug. 1910, Hitchcock 6057 (US). Querétaro: Querétaro, Hitchcock 5822, 5864 (US), Agniel & Arséne 10269* (US); on rocky slopes back of motel 5 miles N of Querétaro, 8 Sep., Gould 11596* (US). San Luis Potosí: Cárdenas, Hitchcock 5712 (F, GH, NY, US); 2 km E of Santa Catarina along hwy. 70, 1220 m, 23 Oct. 1978, Reeder & Reeder 7042 (MEXU). Afín a P. parcum, se distingue por los caracteres de la inflorescencia (laxa y difusa en P. parcum, de 6-15 cm de ancho), por poseer gluma inferior Y, o menos del largo de la espiguillá (Y, del largo en la última especie) y por carecer de un entrenudo mar- cado entre la gluma inferior y superior (mientras que P. parcum tiene un estípite ca. 0.6 mm de largo entre ambas glumas). 238 Annals of the Missouri Botanical Garden 9. Panicum diffusum Swartz, Prodr.: 23. 1788. Panicum diffusum Sw. var. genuinum Doll in Martius, Fl. Bras. 2 (2): 199. 1877, nom. illeg. superfl. TIPO: Jamaica. “Jamaica, Hispaniola” (holótipo, S; fragmento y foto, US-76920; isó- tipos, G, M) Tabl. : “Ex Amer. merid. c D. a Aoc m P, fragmento, Us- 2903511, US-76920). Panicum ficum Sw. var. debile Nees, A Bras. Enum. PI. . 1829. Pann debile Schult. ec 2: A 1824, non P. debile Desf., 17 98, non Ell., 1816, nec Poir., 1816. TIPO: “Hispaniola” (holótipa no visto, fragmento, US- 974075). e l. l: 173. Panicum caespititum Lamarck, 1791. T x Ri- foto Plantas perennes, cespitosas y cortamente rizo- matosas. Cañas erguidas a decumbentes, algo geni- 5—60 cm de alto, ramificadas en los nudos medios o simples; entre- culadas en la base, de (6-)1 nudos cilíndricos, huecos, cortamente pilosos a gla- bros; nudos corta a largamente pilosos. Vainas es- triadas, de 2—6 cm de largo, glabras, los bordes membranáceos, cortamente pestañosos hacia la porción superior. Lígulas membranáceo-ciliadas, reducidas, de 0.6-1 mm de largo; cuello castaño, glabro. Láminas lineares, de (3-)6-15 cm de largo, 0.2 cm de ancho, patentes, los bordes involutos, aplanadas hacia la porción media, cara adaxial densamente pilosa con pelos blanquecinos, la cara abaxial glabra o esparcidamente pilosa, los már- genes basales con pelos tuberculados caedizos o sin lo mismos, el resto escabriúsculos, base angostada, el ápice subulado. Pedúnculos de 9-15 cm de lar- go, cilíndricos, glabros. Inflorescencias terminales exertas, laxas, difusas, de 3-16 cm de largo, 1.5— 8 cm de ancho; eje principal anguloso, escabriús- culo; pulvínulos castaños, glabros; ramificaciones de primer orden alternas, raro opuestas, divergen- tes y distanciadas; ramas inferiores de 3-8 cm de largo, desnudas en su tercio inferior, con espigui- llas congestas y adpresas, hacia la porción superior sobre los ejes de segundo orden; ejes de las rami- ficaciones y pedicelos triquetros, escabrosos. Inflo- rescencias axilares presentes, numerosas, similares a las inflorescencias terminales. Espiguillas larga- mente ovoides, de 2-2.9 mm de largo, 0.8-1 mm de ancho, no estipitadas, glabras, verdosas o con tintes violáceos, gluma superior y lemma inferior mm el largo del antecio superior. Gluma inferior de 1-1.2 mm de subiguales y superando 0. largo, Y, o poco menor del largo de la espiguilla, ovada, aguda, 5—7(—9)-nervia, con los nervios ana- stomosados hacia el ápice. Gluma superior de 2— 2.8 mm de largo, 11—13-nervia, cortamente pilosa hacia el ápice en la cara interna. Lemma. inferior de 2-2.8 mm de largo, 9-11-nervia, cortamente pi- losa hacia el ápice en la cara interna. Pálea inferior ovada, reducida de 1—1.3 mm de largo, 0.6—0.8 mm de ancho, ¥,—¥, del largo del antecio superior, hiali- na, glabra. Antecio superior elipsoide, de 1.5-1.8 mm de largo, 0.6—0.8 mm de ancho, fácilmente caedizo a la madurez, pajizo, liso, lustroso, crus- táceo, con dos cicatrices basales de 0.12 mm de largo; lemma 7-nervia; pálea con papilas simples hacia el ápice. Cariopsis de 1.1 mm de largo por 0.7 mm de ancho; hilo punctiforme; embrión menos de Y, del largo de la cariopsis. Distribución y ecología. Se encuentra en las In- dias Occidentales, en Bahamas, Cuba, Guadalupe, Jamaica, Martinica, Puerto Rico, República Domi- 9). Llega hasta los 1500 m; en flor durante todo el año. nicana y St. Thomas (Fig. Material respresentativo citado. New Providence, Harold Road, Britton d Prace "n nx US). CUBA. Guanabacoa, Hitchcock 2 m Camaguey, pabana N of La Gloria, d Fei na S). ipe as . 1904, Van Aiba 355 mayo 1914, Ekman 995 (G, US). Isla near Nueva Gerona, 12 May 1904, a 494* a MO, NY, P, US); near Madruga, Britton et al. 649 (NY). Ma- tanzas: aes de La Palma, León 9651 (NY). Oriente: Santiago de Cuba, in pastures N of the town, 26 June 1914, Ekman 1517 (G, NY); Mir, in the savannas at the station, d E 1916, Ekman 7511 (G, US). Pinar del ] onda, among rocks in hillside, Wilson 9443 o US); near Guanajay, in | grassy fields, P. 2 (US). Santa Clara: LUPE. Grand-terre, Richard s.n. (P); sin localidad, P. Duss 676 (P), 3181* (NY, US). HAITI. Vicinity of Bassin Bleu, 630—1500 m, Leonard & Leonard 15090 (NY, US); vicinity of Mole St. Nicolas, on bare rocks, between bay and north coast of Cap du Móle, Leonard & Leonard 13147 (US); montagnes du Iron d'Eau, between Source Matelas and Source Puantes, Ekman 2151 (US). JAMAICA. Inverness, Clarendon, 15 Oct. 1915, Harris 12164* (MO, NY, P; E warton I ne ditch in cut, Hitchcock 9463 (US). MARTINIC oteaux des Trois Ilets, ad 536 (US). PUERTO d Guanajibo, near Mayagüe 1915, Britton & Cowell 4064 (MO, NY); |: Quebradillas pockets in limestone cliff above the sea Fe 6577 (G, dde Mi > dl ae ane DOMINE Santo of Ciudad Trujillo, ilar 15744, 15902 2 (US); a Valle de San Juan, 20 ago. 1929, Ekman 13390 (G, ; Senec Valle del ad. Hato del Yaque, Ekman 16237 (US). Monte Cris- ti: foot of El Morro, near seashore, Jiménez 3620 (US). ST. THOMAS. Northside Bay, Sep. 1880, Eggers 76** (G, M, MO, P, US). T Feb. s, o Panicum diffusum se caracteriza por incluir plantas grac "les con cañas manifiestamente ramifi- cadas, láminas linear-lanceoladas e inflorescencias con espiguillas adpresas en las ramas y espiguillas Volume 83, Number 2 1996 Zuloaga & Morrone 239 Panicum Subg. Panicum Secc. Panicum de 2-2.9 mm de largo. Algunos especímenes de esta especie son glaucos y se asemejan a P. filipes, la que se distingue por el tipo de panoja y ancho y largo de la hoja y por el tamaño y forma de la espiguilla. El material citado para Brasil por Doell (1877) como P. diffusum Sw. var. subcontractum correspon- de a P. exiguum Mez. En esta obra Doell cita tam- bién un ejemplar de St. Bartholomei, Forsstrom s.n. (F, S), de las Antillas y que corresponde a Panicum diffusum. domi diffusum fue citada también para Brasil, basándose en la obra de Doell, por Beetle heey y Gould (1979 Gould (1975, 1979), Waller (1976) y Beetle (1987) indican erróneamente que la especie crece en Texas, México y Brasil, sin citar ejemplares de herbario. 10. Panicum ephemeroides Zuloaga & Morro- ne, nom. nov. Panicum ephemerum Zuloaga, Morrone « Valls, Iheringia, Bot. 42: 4, fig. 1- 8, 24 y 25. 1992, hom. illeg., non P. epheme- rum Renvoize, 1979. TIPO: Brasil. Goiás: Mun. Jataí, 51 km a oeste do Rio Claro e 9 km a este do acesso a Perolándia ao longo da rodovia BR-364, 860 m, 3 abr. 1986, Valls & Werneck 9867* (holótipo, CEN; isótipos, SI, US). Figura 19. Plantas anuales, cespitosas, herbáceas, de 60—70 cm de alto. Cañas erectas, simples, paucinodes; en- trenudos cilíndricos, huecos, glabros; nudos glabros. Vainas herbáceas, estriadas y con nervios anasto- mosados, glabras. Lígulas membranáceo-ciliadas, la porción membranácea de 0.4 mm de largo, cilias de 1.5-2.2 mm de largo. Láminas lineares, de 2 cm de largo, 0.2 cm de ancho, con los bordes in- volutos, la base redondeada y el ápice largamente atenuado, los bordes escabriúsculos. Pedúnculos de 15-25 cm de largo, cilíndricos, glabros. Inflorescen- cias terminales exertas, laxas, difusas, multifloras, ca. 30 cm de largo; eje principal anguloso, escabriús- culo a liso y glabro, con ramificaciones primarias divergentes del raquis, las inferiores alternas y las superiores opuestas a alternas; pulvínulos glabros y con tintes violáceos; ejes de las ramas escabrosos, con espiguillas adpresas a las ramificaciones secun- darias, solitarias, distantes y cortamente pediceladas sobre ejes de tercer orden; pedicelos claviformes, escabrosos. Inflorescencias axilares numerosas, simi- lares a las inflorescencias terminales. Espiguillas lar- gamente elipsoides, de 3.5—3.6 mm de largo por 1.3 mm de ancho, biconvexas, no estipitadas, glabras, verdosas y con tintes violáceos, abiertas a la mad- urez, gluma superior y lemma inferior subiguales y de ápice violáceo largamente atenuado, superando 1 mm en largo al antecio superior. Gluma inferior de 2.8 mm de largo, Y, del largo de la espiguilla, abra- zadora, glabra, de ápice atenuado, 3(—5)-nervia, el nervio medio escabriusculo hacia el ápice. Gluma superior 5(—7)-nervia, el nervio medio escabriúsculo. Lemma inferior 5-nervia, glumiforme. Pálea inferior elíptica, de 2 mm de largo, 1 mm de ancho, tan larga como el antecio superior, hialina, los bordes esca- briásculos, glabra en el resto de su superficie. An- tecio superior elipsoide, de 2.1 mm de largo, 1.1 mm de ancho, blanquecino, castafio a la madurez, crus- táceo, lustroso, papiloso sobre el ápice de la pálea superior y con un anillo castafio en la zona de in- serción con la raquilla; lemma 5-nervia; pálea con papilas verrugosas hacia el ápice. Cariopsis elipso- ide, de 2.4 mm de largo, 1.6 mm de ancho, blan- quecina; hilo oblongo; embrión Y, o menos del largo de la cariopsis. Distribución y ecología. Conocida sólo por la colección típica en el estado de Goiás, Brasil (Fig. 8). Habita en campos, llegando hasta los 860 m; en flor en el mes de abril. Especie relacionada a P. stramineum, de la que se distingue por tener esta última espiguillas dis- puestas hacia el ápice de las ramificaciones, ha- llándose los ejes de las ramas desnudos hacia la porción inferior, espiguillas globosas de 2.3—3.2 mm de largo, gluma inferior V,-V, o un poco más del largo de la espiguilla y gluma superior y lemma inferior 9—11(-13)-nervias. 11. Pani exiguum Mez, Bot. Jahrb. Syst. 56, Beibl. 125: 3. 1921. TIPO: Brasil. Minas Ge- rais: Serra de Caldas, Mosen 4571 (holótipo, B, fragmentos, BAA, US-80623; isótipo, P). Figura 20. Panicum diffusum Swartz var. subcontractum Doll in Mar- tius, Fl. Bras. 2(2): 199. 1877. TIPO: Brasil. Minas Gerais: Guarda Mor, Pohl 2435 (lectótipo, aquí de- signado, W). Plantas anuales, cespitosas, de 10-30(-50) cm de alto. Cafías geniculadas, decumbentes, arraiga- das o no en los nudos inferiores, luego erguidas a erectas, densamente ramificadas; entrenudos com- primidos, huecos, hirsutos, con pelos de base tu- berculada a glabros, de 1.5-11 cm de largo; nudos pajizos a castafio-oscuros, densamente pilosos. Vainas de 1—4 cm de largo, comúnmente menores que los entrenudos, hirsutas, cubiertas de gruesos pelos tuberculados, los bordes pestañosos. Lígulas membranáceo-ciliadas, de 0.5-1.5 mm de largo, la porción membranosa reducida, luego largamente ciliadas; cuello pajizo, piloso. Láminas oblongo- 240 Annals of the Missouri Botanical Garden ABESTARD. © Y FY, C "2 lo A Figura 19. Panicum ephemeroides.—A. Hábito.—B. Inflorescencia.—C. Espiguilla, vista lateral.—D. Espiguilla vista del lado de la gluma inferior. —E. Pálea inferior.—F. Antecio superior vista del lado de la lemma.—G. Antecio superior visto de la pálea.—H. Base de la lemma y pálea superior. (Valls & Werneck 9867.) Volume 83, Number 2 Zuloaga & Morrone 241 1996 Panicum Subg. Panicum Secc. Panicum Figura 20. Panicum exiguum.—A. Hábito.—B. Espiguilla vista del lado de la gluma inferior.—C. Espiguilla vista del lado de la gluma superior.—D. Espiguilla sin la gluma superior, dejando al descubierto el antecio superior.—E. Pálea inferior. —F. Antecio superior visto del lado de la lemma.—G. Antecio superior visto del lado de la pálea.—H. Cariopsis, vista escutelar.—I. Cariopsis, vista hilar. (Valls et al. 7693.) Annals of the Missouri Botanical Garden lanceoladas a lanceoladas, acuminadas, de 2—6(— 10) cm de largo, 0.2-0.4(-0.7) cm de ancho, pla- nas, redondeadas a subcordadas en la base, con bordes escabriúsculos y largamente ciliados, carti- laginosos, el resto de la superficie hirsuta, con pe- los tuberculados o sin los mismos. Pedúnculos sub- incluidos a exertos, hasta de 20 cm de largo, cilíndricos, hirsutos a glabrescentes. Inflorescencias 5—8(-14) cm de largo, 2-6(-8) cm de ancho, con espiguillas dis- axas, difusas, piramidales, de 2 persas sobre las ramas, estas divergentes; eje prin- cipal anguloso, hirsuto, con largos pelos blanque- cinos, a ramificaciones alternas, raro opuestas, hirsutas a escabroso; pulvínulos — glabros; escabrosas, violáceas, pedicelos escabrosos, de 2— 10 mm de largo, con largos pelos claviformes o sin los mismos. Pur largamente ovoides, acu- minadas, de (2.1—) mm de largo, 0.8-1 mm e ancho, abiertas a E madurez, no estipitadas, glabras, verdosas, con tintes violáceos; gluma su- perior y lemma subiguales y superando 0.6-1 mm en largo al antecio superior. Gluma inferior de (1.5— 1.7-2.4 mm de largo, ?,-*; del largo de la espi- guilla, acuminada, (3—)5(—7)-nervia, el nervio me- dio escabroso hacia el ápice, cara interna esca- briáscula hacia la porción superior. Gluma superior de 2.2-2.8 mm de largo, 7(-9)-nervia, el nervio me- dio escabroso hacia el ápice, este agudo a acumi- nado, la cara interna lisa a escabriúscula hacia el ápice, caediza y dejando a la vista el antecio su- perior negruzco; gluma inferior y superior separa- das por un entrenudo ca. 0.3 mm de largo. Lemma inferior glumiforme, de 2.2-2.8 mm de largo, 7(- 9)-nervia. Pdlea inferior de (0.7-)1.2-1.6 mm de largo, 0. e ancho, Y,-Y, del largo del antecio superior, ovada, pequefia, membranácea, glabra, bilobada o no en el ápice. Antecio superior elipsoide, de 1.4—1.8 mm de largo, 0.6-0.8 mm de ancho, liso, con papilas simples agrupadas en el ápice de la pálea, pajizo, negro a la madurez, cae- dizo antes que la gluma inferior y la lemma inferior, con dos cicatrices basales, de 0.1 mm de largo, castafias a la madurez; lemma superior 5—7-nervia; me^ pálea con papilas simples hacia el ápice. Cariopsis de 1.1-1.4 mm de largo, 0.7-1 mm de ancho; hilo punctiforme; embrión Y, del largo de la cariopsis. Distribución y ecología. Su centro de distribu- ción es Brasil, desde donde se extiende esporádi- camente a Colombia, Venezuela, Bolivia, Perú, y Paraguay (Fig. 9). Llega hasta los 1000 m; habita en campos arenosos o arcillosos. Nombre vulgar. “Capim do resfriado" (en el ejemplar Macedo 2192). Material representativo citado. BOLIVIA. Beni: Es- tación Biológica de Beni, 40 km al este de San Borja, 14°50'S, 66?40', 200 m, 15 abr. 1991, i sid 2834 (MO). anta Cruz: Nuflo de Chavez, San Antonio de Lomerio, 16%45'S, 61?48'W, 31 ene. 1985, Killeen 823 (F, MO, SI). Tarija: Avilés, 6 km SW of Chocloca, Beck 761 (K, LPB). BRASIL. Ceará: Sobral, Black & Avelino 88C (IAN). Goiás: Chapada dos Veadeiros, ca. 38 km N of Veadeiros, 1000 m, 16 Mar. 1969, Irwin et al. 24519 (F, MO, K, NY, P, UB, US); Chapada dos Veadeiros, 4 km al S de Tere- zinha, Anderson 7381* (MO, NY, UB); Santa Rita do Ara- guaia, Chase 11791* (US). Maranhão: Carolina to San Antonio de Balsas, Swallen 4106** (K, US); Grajaú to Porto Franco, red clay outcrops, open ground, Swallen 3822 (IAN, K, NY, RB, US); 100 m NE of main house of Fazenda "Morros," 35 km S of Loreto, 300 m, 3 Apr. 1962, Eiten & Eiten 3956 (F, M, NY, SP, US). Mato Gro- sso do Sul: Campo Grande, sand and clay, 550 m, Chase 10846 (F, GH, , US); Porto Esperanca, 28 Feb. 1930, Chase 11071 (GH, IAN, RB). Minas Gerais: Co- rinto, Fazenda do Diamante, 600 m, 15 abr. 1931, Mexta 5622* (BAA, F, G, GH, IAN, K, M, NY, P, R, US, W); Santa Terezinha, Macedo 2194* (MO, US); Serra do Cipó, km 50, 800 m, 28 Mar. 1925, Chase 9300* (F, GH, MO, NY, US). Pará: Estreito-Marabá, km I, Pinheiro & Car- valho 55, 218, 483, 634 (UB, US); Transamazonian high- way (BR-230), 6 km W of Estreito, Rio Tocantins, 6°32" S, 47°32'W, 300 m, 28 Feb. 1980, Plowmann et al. 9269* n SI). Paraiba: Campina Grande, Pickel 3799 (IAN, B, US). Pernambuco: Arcoverde, 21 jul. 1971, Andra- de Lima 71-6363 (F). Piauí: Gilbués, Filgueiras & Ro- Swallen 4859 (CEN, RB, SP). COLOMBIA. Norte de Santander: Los Estoaques, La Playa, 15 ene. 1974, Bal- ick 133 (COL). PARAGUAY. Río Apa, Hassler 10986 (G, K, MO, NY, P, US). PERU. Juniń: Colonia Perené, 590 m, 22 Oct. 1923, Hitchcock 22115* (F, GH, NY, R, US). VENEZUELA. Aragua: La Victoria, prope Colonia Tovar, en 2565 (K). Distrito Federal: Yaguara, hierba de m de alto, en sabanas altas de colinas, nov. 1940, iis 1469 (F, VEN). Lara: Dtto. Palavecino, dirt road between the quebrada “La Mata" and the east side of the Parque Nacional Terepaima, Burandt Jr. & Gantaume V0498 (MO). Trujillo: south of Valera, stony flat above to río Matalán, 2600 m, 7 Mar. 1940, Chase 12389 (NY, VEN) Especie afin a P. peladoense Henrard, de la que se separa por ser una planta anual, con cafias de- cumbentes y manifiestamente ramificadas en la base, láminas de base redondeada a subcordada y espiguillas con la gluma inferior comúnmente Y,-Y, del largo de la espiguilla. De P. hirticaule se dis- tingue por ser la gluma superior caediza, antecio superior negruzco y gluma inferior Y, a Y, del largo de la espiguilla. El ejemplar Swallen 4859 posee un porte mayor al que se presenta usualmente en esta especie, lle- gando hasta unos 90 cm de altura, con una inflo- rescencia de aproximadamente 20 cm de largo. Luces (1963) citó el ejemplar Tamayo 1469 como Panicum hirticaule. Volume 83, Number 2 1996 Zuloaga & Morrone 243 Panicum Subg. Panicum Secc. Panicum 12. Panicum flexile (Gattinger) Scribner in Kearney, Bull. Torrey Bot. 0: 1893. Panicum capillare L. var. flexile Gatt., Tennessee Fl.: 94. 1887. Chasea flexilis (Gatt.) Nieuwl., Amer. Midl. Naturalist 2: 65. 1911. TIPO: Estados Unidos de América. Tennessee: Cedar Glade near Nashville, Sep. 1888, Gat- tinger s.n. (lectótipo, designado por Hitchcock & Chase (1910) TENN, no visto, fragmento US-80552). Plantas anuales, cespitosas, erectas, de 10—70 cm de alto. Cañas delicadas, erectas, paucinodes, rami- ficadas en los nudos inferiores; entrenudos de 3-9 cm de largo, 1 mm diám., surcados, glabros a hir- sutos junto a la porción superior; nudos densamente pilosos, con pelos blanquecinos. Vainas de 3—6 cm de largo, usualmente más largas que los entrenudos, verdes a violáceas, híspidas, con los bordes espar- cidamente pilosos. Lígulas membranáceo-pestaño- sas, de 0.6-1 mm de largo; cuello glabro, blanque- cino. Láminas linear-lanceoladas, de (2.5-)7-25 cm de largo, 0.1-0.5 cm de ancho, planas o con los bordes involutos, ascendentes, erectas, predominan- temente basales, verdes o con tintes violáceos, de ase angosta y ápice agudo, con los bordes esca- briúsculos, los basales largamente pestañosos, con pelos blanquecinos, tuberculados hasta de 5 mm de largo. Pedúnculos largamente exertos, hasta de 20 cm de largo. Inflorescencias terminales angostamente elipsoides, de 5-30 cm de largo, 1-6(-10) cm de ancho, difusas; eje principal glabro, liso, cilíndrico; pulvínulos glabros; ramificaciones de primer orden ascendentes, poco divergentes, alternas a subopues- tas, ocasionalmente subverticiladas, glabras, lisas; ramificaciones de segundo orden, divergentes, laxas, difusas, con espiguillas aisladas, divergentes; pedi- celos de 0 mm de largo, glabros, escabriús- culos. Inflorescencias axilares similares a las termin- ales, de menor tamaño. Espiguillas largamente elipsoides, de 2.4—3.6 mm de largo, 0.6-0.8 mm de ancho, no estipitadas, glabras, verdosas, con tintes cobrizos, de ápice largamente agudo; gluma superior y lemma inferior subiguales, superando 0.8-1.5 mm en largo al antecio superior. Gluma inferior ovado- acuminada, de 0.8-1.3 mm de largo, V,-/, del largo de la espiguilla, 3-nervia. Gluma superior de 2.3— 3.3 mm de largo, 7-9-nervia. Lemma inferior glum- iforme, de 2.4—2.7 mm de largo, 7—9-nervia. Pálea inferior ausente. Antecio superior largamente elipso- ide, de 1.6-1.7 mm de largo, 0.6 mm de ancho, crustáceo, liso, lustroso, brillante, pajizo y con tintes negruzcos a la madurez, con un anillo basal circular; lemma 7-nervia; pálea con papilas simples hacia el ápice. Cariopsis fusiforme, de 1.2 mm de largo, 0.4 mm de ancho, pajiza; hilo punctiforme; embrión Y, del largo de la cariopsis. Distribución y ecología. Se encuentra en Canadá y en los Estados Unidos de América (Fig. 10), en planicies, sobre suelos frecuentemente húmedos, o en ambientes rocosos. Llega hasta los 200 m; en flor entre agosto y octubre. Material representativo citado. NADA. Ontario: St. Clair River, 17 Aug. 1904, Dodge 124 (US). ESTADOS UNIDOS DE AMERICA. Alabama: Franklin Co., E of T ball, 21 Sep. 1939, Harper 3759* (MO, US). Arkinsas Marion Co., 600 ft, 7 Oct. 1950, Demaree 30200 (US); Ful. ton, 16 Oct. 1901, Bush 1087 (MO). Georgia: Catoosa Co., s US. Kentucky 1926, pir abhi 358 (MO). Michi- Pai — al Park iat use area, entrance on Huy. M-25, 1.6 mi. S and W of Port Crescent Rd., 9 Aug. 1982, Gereau 1 wa py Mississippi: near Starkville, Sep. 1929, Kearney Jr. 72 (US). Missouri: Ste. Genevieve Co., Establishment tock 22 Sep. 1976, Christ Aug. 1923, Fernald et al. 141415 (US). Benson Co., Wood Lake, Tokio, 1 Aug. 1940, Stevens 477 (MO). Ohio: Cincinnati, Cox Bend of South York, 5 Sep. 1939, Braum 2648 (US). Oklahoma: Muskogee Co., 25 Sep. 1940, Bebb 5957a (US). Pennsylvania: Lancaster Co., Diverville Swamp, Sep. 1892, Small s.n. (US-742109, 742110); Lancaster Co., Diverville Swamp, 2 Sep. 1862, Porter s.n. (US-952823). Tennessee: Maury Co., ca. 2.5 mi. SE Columbia by Tenn 50, 27 Aug. 1973, Kral 51473 (MO); Johnson City, Canby 221 (US). Texas: Clarksville, Feb. 1894, Plank 10 pr.p. (US). Virginia: Rockbridge Co., about Lexington, 3 Sep. 1924, Churchill s.n. (MO-1056628); Fau- quier Co., WW slope of Bull Run Mountains, 13 Sep. 1942, Allard 10422 (MO). Esta especie se caracteriza por poseer cañas florí- feras que alcanzan % o más (hasta %) del alto de las cafias vegetativas. Las panojas son angostas y las espiguillas largamente elipsoides, con pálea inferior ausente y antecio superior con la base redondeada. Es frecuente observar que las plantas tengan en ge- neral tintes violáceos. Las inflorescencias axilares se encuentran incluidas en las hojas o son poco exertas. ejemplares examinados de Canadá poseen pedicelos de 0.8 a en los especímenes de los Estados Unidos los pedi- celos llegan a tener 5 mm de largo. 1.5 mm de largo, mientras que 13. Panicum furvum Swallen, Contr. U.S. Natl. Herb. 29: 416. 1950; Sr Davidse, Flora Mesoamericana 6: . TIPO: Guate- mala. Huehuetenango: pun. Nentón and Las Palmas, vía Yalisjao, Rincón Chiquite, Chiaquial, Guaxacana, in Sierra de los Cuchu- matanes, 30 Aug. 1942, 800—1200 m, Steyer- 244 Annals of the Missouri Botanical Garden mark 51627*,** (holótipo, US-1935086; isóti- pos, F-1202588, US-1935080). Plantas perennes, densamente cespitosas. Cañas erectas de 16-20 cm de alto, simples, paucinodes; entrenudos de 2-3 cm de largo, papiloso-pilosos junto a la porción superior, luego glabros, pajizos o con tintes violáceos; nudos castaños, pilosos, con pelos blanquecinos ca. 1.5 mm de largo. Vainas de 2-3.5 cm de largo, usualmente más largas que los entrenudos, glabras a esparcidamente pilosas hacia la porción basal, pajizas o con tintes violáceos, los bordes membránaceos. Lígulas membranáceo-pes- tañosas, ca. 0.5 mm de largo; cuello glabro. Lámi- nas linear-lanceoladas, de 2-12 cm de largo, 0.2— 0.4 cm de ancho, erectas, ascendentes, predomi- nantemente basales, planas o con los bordes invo- lutos, glabras a esparcidamente pilosas en la cara adaxial, de base atenuada y ápice agudo, los bordes inferiores largamente pestañosos. Pedúnculos exer- tos, hasta de 9 em de largo, cilíndricos, delgados, glabros a esparcidamente pilosos. /nflorescencias largamente exertas, laxas, paucifloras, de 3-6 cm de largo, 0.5-1.3 em de ancho; eje principal glabro a esparcidamente piloso junto a la base; pulvínulos glabros; ramificaciones ascendentes, poco diver- gentes, glabras, lisas a finamente escabriúsculas, pajizas y con tintes violáceos, las inferiores hasta de 3 cm de largo; pedicelos claviformes, hasta de 7 mm de largo, adpresos, glabros. Espiguillas elip- soides, de 2.4—2.7 mm de largo, 0.8 mm ancho, de ápice largamente agudo, no estipitadas, glabras, pa- jizas, con el ápice de las brácteas violáceos; gluma superior y lemma inferior subiguales, superando 0.5-0.8 mm en largo al antecio superior. Gluma inferior ovado-acuminada, de 1.4—1.6 mm de largo, Y o un poco más del largo de la espiguilla, 5-ner- via, con el nervio escabriüsculo hacia el ápice. Glu- ma superior de 2.4—2.5 mm de largo, 7-nervia, cae- diza a la madurez, dejando el dorso del antecio superior descubierto, la cara interna escabriüscula hacia el ápice; gluma inferior y superior separadas por un breve entrenudo de la raquilla de 0.2 mm de largo. Lemma inferior glumiforme, de 2.4—2.6 mm de largo, 9-nervia, con la cara interna esca- briáscula hacia el ápice. Pdlea inferior reducida, lanceolada, de 1.2-1.4 mm de largo, 0.5 mm de ancho, % del largo del antecio superior, hialina, glabra, de ápice bilobado, los bordes glabros a den- ticulados. Antecio superior elipsoide, de 1.6 mm de largo, 0.7 mm de ancho, crustáceo, liso, lustroso, glabro, negruzco a la madurez, con dos pequefias cicatrices inconspicuas ca 0.1 mm de largo en la base; lemma 7-nervia; pálea con papilas simples hacia el ápice. Cariopsis no vista. Distribución y ecología. Conocida sólo para la localidad típica en Guatemala (Fig. 9), creciendo n laderas de montaíia entre los 800 y 1200 m Swallen (1950) sefiala, al describir la especie, que la espiguilla tiene 1.5-1.6 mm de largo. Sin embargo, el análisis del material típico permitió confirmar lo establecido por Davidse (1994), en el sentido de que la longitud de las espiguillas varía entre 2.4 y 2.7 mm. Panicum furvum se relaciona, dentro del grupo de especies perennes con antecio superior negruzco y gluma superior caediza, a P. peladoense, especie sudamericana de la que únicamente se distingue por el tamafio de las plantas y espiguilla. Dentro de las especies centroamericanas se aproxima a P. hallii var. hallii, especie que tiene láminas rizadas en la base de las plantas, espiguillas mayores, de 2.7A mm de largo, con gluma superior persistente y antecio supe rior pajizo (o castafio a la madurez). Davidse (1994) indica la relación de P. furvum con especies subamericanas de la sección y menciona que, si bien su presencia en Guatemala es peculiar, no hay indicaciones en la etiqueta de colección del ejemplar tipo que indique que se trata de una es- pecie introducida. 14. Panicum ghiesbreghtii E. Fournier, Mexic. 2: 29. 1886; E. Fourn. in Hemsl., Biol. Centr.-Amer., Bot. 3: 489. 1885, nud. TIPO: México. “Absque loco,” Ghiesbreght s.n. (lectótipo, designado por Hitchcock & Chase (1910), P, fragmento y foto, US-76924). nom. Panicum ec Hitchcock, Contr. U.S. Natl. Herb. 12: 223. 1909. TIPO: Cuba: sin localidad, 1865, vie 758 (holótipo, US-559958; isótipos, MO, NY, P, US-823614, 974804). Plantas perennes, cespitosas, de 40-90(-120) cm de alto, con innovaciones intravaginales. Cañas simples a ramificadas en los nudos inferiores, erec- tas; entrenudos cilíndricos, de 6-20 cm de largo, 2-3 mm diám., papiloso-pilosos, paucinodes; nudos densamente pilosos, con pelos blanquecinos. Vai- nas de 5-15 cm de largo, las inferiores común- mente menores que los entrenudos, hirsutas con densos pelos tuberculados. Ligulas membranáceo- ciliadas, de (0.5—)1.5-2 mm de largo, con largos pelos por detrás en la base de la lámina; cuello 'nsamente piloso. Láminas linear-lanceoladas, de 30-55) em de largo, (0.5—)0.8-1.4 em de an- um planas, erectas, ascendentes, de ápice larga- mente atenuado y base redondeada, la cara adaxial densamente pilosa, pilosidad serícea, cara abaxial hirsuta, con pelos tuberculados, los bordes infer- iores ciliados, el nervio medio manifiesto. Pedún- Volume 83, Number 2 Zuloaga & Morrone 245 Panicum Subg. Panicum Secc. Panicum culos subincluidos en las vainas foliares a exertos, hasta de 40 cm de largo, glabros a hirsutos. /nflo- rescencias terminales laxas, amplias, multifloras, cortamente exertas a parcialmente incluidas, de 7— 35 cm de largo, 5-23 cm de ancho; eje principal anguloso, escabroso; pulvínulos glabros; ramifica- ciones inferiores alternas, las subsiguientes verti- ciladas a opuestas o alternas, divergentes, ramifi- caciones de primer orden desnudas en su tercio inferior, las espiguillas apareadas sobre ejes de ter- cer orden hacia la parte superior, no congestas; pedicelos claviformes, escabrosos. /nflorescencias axilares presentes, similares a las inflorescencias terminales. Espiguillas largamente ovoides, de 2.6— 3.1 mm de largo, 0.9-1. e ancho, no estip- itadas, glabras, abiertas a la madurez; gluma superior y lemma in- ferior subiguales, superando 0.7—0.9 mm en largo al antecio superior. Gluma inferior de 1.4—1.7 mm de largo, % o poco menos del largo de la espiguilla, verdosas, con tintes violáceos, 5—T-nervia, nervio medio escabriúsculo hacia la porción superior, los nervios anastomosados hacia el ápice, el ápice agudo. Gluma superior de 2.5— mm de largo, 9—-13-nervia, el nervio medio esca- briúsculo hacia la porción superior, aguda, corta- mente pilosa hacia el ápice en la cara interna. Lem- ma inferior de 2.5-3 mm de largo, 9-11-nervia, nervio medio escabriúsculo hacia la porción supe- rior, aguda, cortamente pilosa hacia el ápice en la cara interna. Pálea inferior reducida, de 0.5-1.3 mm de largo, 0.: mm de ancho, hialina, Y4-% del largo antecio superior, ovada, de ápice obtuso. Antecio superior ovoide, de 1.7-2.3 mm de largo, 0.8-1.1 mm de ancho, ?4 del largo de la gluma superior y lemma inferior, pajizo, liso, lustroso, gla- bro, crustáceo, el ápice obtuso, con 2 cicatrices ba- sales, de 0.12 mm de largo, castafias a la madurez; lemma 7-nervia; pálea con papilas simples y micropelos globosos hacia el ápice. Cariopsis elip- soide, de 1.2-1.3 mm de largo, 0.7-0.8 mm de an- cho; hilo oblongo; embrión !4 del largo de la cariop- sis. Distribución y ecología. Habita desde Estados Unidos, las Indias Occidentales y América Central hasta el norte de América del Sur, en Colombia y Venezuela (Fig. 6); crece en campos. Llega hasta os m. Florece y fructifica entre enero y no- viembre. Material ii citado. ANTIGUA. Sin locali- dad, Hitchcock 16382 (US), Wullschagel 621, 622 (M); Scotts Hill Expt. e US). BAHAMAS. Crooked Island, Sandrail Point, Brace 4812 (NY, US). BELICE. Corozal: (n. Sarteneja and Chunox, 18°17'N, 88°15'W, 10 m, 18 Mar. 1987, Davidse & Brant 32638* (MO, SI). COLOMBIA. Magdalena: Santa Marta, Smith 165 (F, G, GH, K, MO, NY, P, US). Santander: p Ma Zapatoca, Fassett 25492 (COL, US). Tolima: El Gigante, above Tolima, 500-1200 m, Lehmann 8745 (NY, US). COSTA RICA. Guanacaste: dini road W of the Cafias cemetery, 70 m, 24 June p Pohl & Davidse 10556 ul n US). San José: San José, 1100 m, Hitch- , Jiménez 530** (US); Hacienda Taboga, tud: Santa Bárbara, zas: cercanías de C 6 (NY); eue | of Herradura, , between Gerardo and Corojalito, on the railroad track, Ekman 12686 (US). Santiago de Cuba: Bayate, Ekman 2011 (G), 9667 (NY); Guaro, 20 km from Preston, pride 21 397 (US). Villa Clara: Placetas del Sur, León 6429 (US). DOMINICANA. Distrito Nacional: vicinity of Ciudad Trujillo, D 14148 (US). aie Valle del Cibao, Hato del Y: Ekman 16129 (US). ECUADOR. Chac 'ao, Miranda, 98 ago. 1939, Eggers 14949 (F). Manabí: El Recreo, Eggers 15419* (F, K, M, NY, P, US). EL SALVADOR. Acajutla, near sea level, 29 Nov. 1911, Hitchcock 8993 (SI ) ESTADOS UNIDOS DE AMERICA. Texas: Brownsville, 13 Oct. 1941, Silveus 7293 (US). GUADALUPE. Sin lo- calidad, Questel 1439 (P, US); Basse Terre, Duss 668 (P). HAITÍ. Vicinity of Pilate, 325 m, Leonard 9607 (NY, US); vicinity of Pétionville, 15 June 1920, Leonard 4901* (US); vicinity of Port au Prince, Leonard & Leonard 15754 (US). HONDURAS. Francisco Morazán: El i orano, 800 m, 2 jul. 1947, Molina 242 (F, GH, MEX U, MO); vicinity of El Zamorano, Swallen 11147 (MO, US). JAMAICA. In- verness, Lower Clarendon, 22 Nov. 1913, Shs abi 11691 (MO, NY, P, US). MEXICO. Campeche: al Sd Bolonchen de Rejón, cerca de las Grutas ds Xtacumbi- Ixunan, 'arretera vía ruinas a Campeche, 25 jul. 1986, Patter 11739 (MO). Chiapas: ca 22 km S Teopisca along Highway Comitán, pine-oak forest, O m, 11 Aug. 1975, Davidse & Davidse 9492 (MO). Mexico: Peñón, 1700 m, 8 Oct. 1933, Hinton 4413 (MO, US). M ichoacan: about 12 mi. W of Quiroga in open pine- part forest, 7000-9000 ft, Sohns 791, 797 (US). Mo- relos: Oaxtepec, Matuda s.n. (MEXU-288916). Oaxaca: Oaxaca, Hitchcock 6143 (US). Puebla: roadside along route 130, 2.7 mi. S of the Puebla-Veracruz state line, 20 May 1979, Harriman & Jansen igi (CHAPA). Quin- tana Roo: Coba, in savanna, n of Lake Coba, June 1938, Lundell & Lundell 7791 E MEXU, MO, US). San Luis Potosí: low ground along Río Tampaón, Chase 7480 Aira E^ Santos 1832 (NY). Reyes calle Palmas 1 km al W de la carretera Cár- edi», 4 . 1981, Suárez 4 (MO). Ta- maulipas: Chamal, Swallen 1717 (US); along route 85, ca. 4—5 miles south of Ciudad Mante, 18 Feb. 1961, King 3791 (F, US). binder Orizaba, "e aa 6391 (US); environs de Veracruz, Gouin 21 (P). Yucatán: Chi- chen Itzá, along roadside, Lundell & i 7427 (US). NICARAGUA. Boaco: near Río Tecolostote, 7.5 km S of Hwy. 7 at El Pa initi ca. 12°11'N, 85°40'W, 60 m, 27 June 1982, Stevens 21647 (MO). Carazo: Jinotepe, 500 m, Hitchcock 8687 (US). Chinandega: Corinto, near sea level, 10 Nov. yis Hüchcock 8755 (SI, US Esteli: kn 163 on Hwy. 1, ca. 11.2 km N of entrance to Estelí, « 13*13"N, 8023'W. 920 m, 19 mayo 1981, Stevens 20198 246 Annals of the Missouri Botanical Garden (MO); 16 km N de Esteli, 900 m, 6 jul. 1970, Pohl & Dans 12210 (F, MO). Managua: along new road from 1 (N of San Jacinto entrance) to San Francisco del zarnicero, ca. 4.8 km W of hwy. 1, 16 Aug. 1978, Stevens 10006 (MO). Matagalpa: bep) 23 jul. 1969, inn 2307 . Rivas: San Juan del Sur, r sea lev Hüchcock 8601 (US). PANAMA. Ancon, Killip 4100 (US) in the vicinity of Balboa, Canal Zone, near swamp, e h- cock 7997 (SI, US), 8014 (US). Panama: Taboga island, Gulf of Panama, Hitchcock 8065 (SI, US), 8094. (US). PUERTO RICO. Vicinity of Coamo Springs, Chase pss NY); vicinity of Cayey, road to Guayama, Chase (US). ST. THOMAS. Virgin Islands, sin colector (P). is GIN ISLANDS. Tortola, Fishlock 206 (NY). VENEZUE- . Aragua: en terrenos satincs de la Est. Exp. del Ser- vicio Shell para el Agricultor, Tamayo 3949 (VEN). Distrito Federal: seii brique E ¿hase 12607 (US); La Florida, near Caracas, 12 Mar. 8, Alston 5298 (F, e San ca arriba de Santa * Socorro, 15 km antes de la áltima población, en desvío a Zaraza, 150 m, Zuloaga et al. 4506* (SI, VEN). Lara: Cerca de Sicarigua, Burkart 16870 (VEN). Mérida: Bre- guesa: a 4 kr Q903'N. 69748' W, 200 m, : ago. 1989, Zuloaga & Ortíz 4300, 4301* (MO, SI). Yar : Bureche, antes de Bar- quisimeto, 12 abr. 1946, Ed 16508 (SI). Especie afín a P. hirticaule, de la que se separa por incluir la última especie plantas anuales, con cañas comúnmente decumbentes, hojas general- mente de menor tamaño, patentes, delgadas y de base subcordada. Hitchcock (1927) cita esta especie para Bolivia con el ejemplar Steinbach 6979, el que corresponde a P. quadriglume. Esta especie fue erróneamente citada para Su- damérica austral por Hitehcock (1936), Swallen (1943), Luces (1963), Palacios (1969), Adams (1972), Smith et al. (1982), y Longhi Wagner « Boldrini (1988); el material citado por estos autores — corresponde a P. chasei. 15. Panicum ep Vasey, Bull. Torrey Bot. Club. 11: 61. 1884. TIPO: Estados Unidos de Amé- rica. "ien sin localidad, Hall 816** (holóti- po, US-76926; isótipo, NY). glaucas. Cañas Plantas perennes, cespitosas, )20—75 cm de alto, simples o ram- erectas de (1 ificadas en la base, paucinodes; entrenudos de 1— 11 cm de largo, cilíndricos, huecos, glabros, pajizos o con tintes violáceos; nudos pilosos, cubiertos de pelos blanquecinos cortos, o glabros. Vainas de 3— 11 cm de largo, glabras a hirsutas, con pelos tub- erculados caducos, los márgenes membranáceos, glabros a ciliados junto a la región ligular. Lígulas membranáceo-pestañosas, de 0.6-1.2 mm de largo, cuello glabro. Láminas linear-lanceoladas, de 4-23 cm de largo, (0.1)-0.2-0.5 cm de ancho, ascendentes, las basales rizadas, planas o con los erectas, bordes involutos, glaucas, glabras a esparcidamente hirsutas, de base redondeada a angostada, conti- nuándose imperceptiblemente con la vaina, el áp- ice agudo, con los bordes cartilaginosos, escabritis- culos, los inferiores pestafiosos, con pelos papilosos, blanquecinos, hasta de 3.5 mm de largo. Pedúnculos hasta de 40 cm de largo, cilíndricos, glabros, verdes o con tintes purpüreos. /nflorescen- cias terminales exertas, laxas, de 9—31 cm de largo, 3-15 cm de ancho; eje principal triquetro, glabro, pulvínulos glabros; ramificaciones de primer orden alternas, divergentes, glabras, lisas; ramificaciones de segundo orden divergentes o adpresas, con es- piguillas dispersas a aproximadas, adpresas a div- ergentes; pedicelos de 1-15 mm de largo, trique- claviformes, escabritisculos. tros, glabros, Espiguillas largamente ovoides, acuminadas, de 2.1—4 mm de largo, 0.8-1 mm de ancho, no estip- itadas, glabras, verdosas o con tintes violáceos; glu- ma superior y lemma inferior subiguales, 0.3-1.2 mm más largas que el antecio superior. Gluma in- ferior ovada, de 1.2-2.4 mm de largo, 44% de largo de la espiguilla, de ápice agudo a acuminado, 3— 5-nervia, con el nervio medio escabriúsculo hacia la porción superior. Gluma superior de 2-3.5 mm de largo, de ápice acuminado, 7—9(—11)-nervia, es- cabriúscula en la cara interna hacia la porción su- perior; gluma inferior y superior separadas por un corto entrenudo de 0.3 mm de largo. Lemma infe- rior glumiforme, de 1.8-3.7 mm de largo, 9-11- nervia. Pálea inferior reducida, de 0.8-1.5 mm de largo, 0.3-0.4 mm de ancho, %—% del largo del antecio superior, membranácea, hialina, glabra, con los bordes denticulados. Antecio superior ovoide a elipsoide, de 1.5-2.4 mm de largo, 0.7-1.2 mm de ancho, crustáceo, liso, pajizo, con tintes negruzcos a la madurez, con 2 cicatrices basales de 0.2 mm de largo, castafias a la madurez; lemma 7-nervia; pálea con E simples en su ápice. Cariopsis —1.7 mm de largo, 0.6-1 mm de an- cho, VM A hilo punctiforme; embrión % del ovoide, de 1 largo de la cariopsis. 15a. Panicum hallii var. hallii Panicum es E. Fournier, Mexic. Pl. 1886. TIP éxico: sin localidad, Virlet T S. aquí oda P, fragmento, US-86967). Láminas de 4—15 cm de largo, (0.1—)0.2—0.5 cm de ancho. /nflorescencias con espiguillas aproxi- madas, adpresas. Espiguillas de 2.7—4 mm de largo. Volume 83, Number 2 1996 Zuloaga & Morrone 247 Panicum Subg. Panicum Secc. Panicum Gluma superior y lemma inferior 0.8—1.2 mm más largas que el antecio superior. Distribución y ecología. Ampliamente distri- buido en el sur de los Estados Unidos y en México (Fig. 7); crece en campos áridos sobre suelos ro- cosos, hasta los 1920 m; florece entre marzo y oc- tubre. Nombre vulgar. “Sacate colorado” (en el ejem- plar Wooton s.n.). Material representativo citado. ESTADOS UNIDOS DE AMERICA. Arizona: Santa Cruz Co., Patagonia, 21 Sep. 1908, ages 3706 (US); Santa Rita Mt re ie s 20 Oct 2, Griffiths & Thornber 238 (US); 42 i. N of Tueson, 4500 Me 15 Aug. 1915, Hitchcock 13255 S of — SW Baca Co., 4000 . New Mexico: 1400 m, 30 vi Cimarron Co., 3 mi. N of Kenton, 9 July 1947, Rogers Pen (US). Texas: Young Co., 13 mi. SE of Graham, 15 1965, Gould 11075a (US); tiva ir pea El Paso, 22/25 May 1912, Chase 5900 (US); Kerrville, Hitchcock 5293 (US); Big Spring, 25/30 Aug. 1915. Hitchcock 13411 (US); Wil- iamson Co., roadside S.H. 29, 15.3 mi. E of Burnet, 9 Sep. 1969, Waller 2163 (MO). Utah: Beaver County, Red Knoll, east side Pine Valley, 3 Sep. 1934, Hutchings & Stachmann s.n. (US-1648084). MEXICO. Aguascalien- tes: 1.5 km a Real a Asientos, 22 Oct. 1965, Atrupo 1723 (MEXU). Chihuahua: km 5 desviación a Nuevo Casas Grandes de carretera Cd. Juárez, 1600 m, 19 sep. 1955, Hernández & Mathus N-1911 (US); Sta. Eulalia Moun- tains, ago. 1875, Pringle 376* (F, GH, MO, NY, P, US); e of Cerro Campana, 79 km N of Chihuahua, 1620 m, 23 hue. 1975, Davidse & Davidse 10057 (MO). Coahuila: onclova, in vicinity of Chulavista Hotel, 23 May 1965, Gould 1187 (US); Municipio de Castafios, south of Cas- tafios, Rancho Santa Teresa, Wynd & Muller 174* (GH, MO, NY, US); Laguna de Jaco, Johnston & Muller 1111* (GH); San yd de los Alamos, eastern base of the e San Antonio de los Alamos, Johnston f Santa Elena Mines, 7 Sep. 1940, Johnston & Muller 1017* (F, GH, US); 2 miles NW of Frontera, Aug. 1938, Johnson 7170* (GH). Durango: 24 miles NW of La Zarca, 6200 ft., Soderstrom 819 (US); des- México, carretera 57, 67 km al N de Saltillo, 1600 m, 25 ago. 1978, Bernal & Cárdenas s.n. (MEXU-271683); N of Sabinas, Hidalgo, 28 Aug. 1971, Beetle 1061 (MO). Oa- xaca: Mun. Zapotitlan, 4 km e Zapotitlan Salinas, 25 sep. 1990, Sá er et al. 328 (MEXU). Puebla: ae ds de Tehuacán, por E e a Esperanza, 1 ur 989, Chiang et al 294 (MEXU). Querétaro: 2 mi. m le , Soderstrom ae (US). San Luis Po- tosí: valley of ihe Rioverde between Rioverde and S km uadalcazar, m, Rzedowsky 4931 (GH, US); Valle de Rioverde, between Rioverde and Boquilla, Sohns 1248 (F, MEXU, P, US). Tamaulipas: vicinity of San Mi- guel, Bartlett 10591 (GH); vicinity ja! Mo 320 m Palmer 554 (US); 35 km from Vict on the roa to Casas qu Soto La pda ca. "280 m "3 Oct. 1956, Mar- tinez Martinez & ando F- 2303 -— US). Sin E EL. sin n localidad. Virle 1371 ( Existen especímenes considerados aquí dentro de la variedad típica que poseen características in- termedias con la var. filipes, incluyendo plantas de mayor tamaño, con inflorescencias con ramifica- ciones abiertas, "e de 2.9 mm de largo. Entre los mismos se pueden citar a Waller 2434, arnock s.n., Bush 858, Hall 817, Palmer 1 9897, Roybal 883, Sohns 1215, Rzedwosky 5174, Martínez Martinez & Borja Luyando F-2303. 15b. Panicum hallii var. filipes (Scribner) F. R. Waller, Southw. Naturalist 19: 105. 1974. Pa- nicum filipes Scribn., Bot. Explor. S. Texas 1: 13. 1895. TIPO: Estados Unidos de América. Texas: growing in rich shaded ground in the upper part of the “Arroyo” at Corpus Christi, 30 ft., 31 May 1894, Heller 1809 (holótipo, US-2463136, fragmento y foto, US-953192; is- ótipo, NY). Láminas de 11-23 cm de largo, 0.2-0.4 cm de ancho. Inflorescencias con espiguillas dispersas, no adpresas sobre las ramas. Espiguillas de 2.1-2.8 mm de largo. Gluma superior y lemma inferior 0.3— 0.7 mm más largas que el antecio superior. Distribución y ecología. Frecuente en campos y en bordes de caminos (Fig. 10). Florece entre mar- zo y diciembre. Material age citado. ESTADOS UNIDOS DE mE, Tex an Patricio Co., near entrance to rpus Christi. State Park, 2 Oct. 1964, Gould 10972, 10973 (US); along Ennis Joslin Road by golf rse, 0.9 mi. N of Padre Island Parkway, Corpus Christi, 22 July 1960, Waller 2017 (MO). MEXICO. Coahuila: 24 km E of Don Martín Dam, 335 m, 20 mayo 1939, Harvey 950 (US); El utis near Múzquiz, Harvey 1186* (MO, US). Nuevo Leon: i. S of China on the road Foward Méndez, 7 Dec. Ts Crutchfield & age 6068 (US); € Highway from Laredo mulique Pass, 1500 ft. MO). San Luis Potosi: El Banito, Chase 7558* Fernando on the Matamoros highway, 7 Dec. 19 59, John- ston 4870 (MEXU); Chamal, an 1716 (US); Hacienda Buena Vista, Wooton s.n. (US 1061807); Laredo, 28 ago. 1971, Beetle et al. 1031 Mo Correll & Johnston (1970) citan a P. filipes como una especie afín a P. halli, de la que se distingue por el tamaño de la espiguilla (de 2-3 mm, panoja más amplia y láminas mayores en P. filipes y es- ipes y piguillas de 3—3.7 mm en P. halli). Además rela- Annals of the Missouri Botanical Garden ciona a la especie con P. pilcomayense (esta aparen- temente introducida en Texas). Se considera que el ejemplar Johnston 4870, ci- tado por Waller & Morden (1983) bajo P. tamau- lipense, pertenece a P. hallii var. filipes. Existen, dentro del material considerado de esta variedad, ejemplares, como Beetle et al. 1031, Wa- ller 2008, Hitchcock 5500, Harvey 950, que poseen características intermedias con la variedad típica, como el poseer espiguillas adpresas sobre las rami- ficaciones, siendo las medidas de las espiguillas coincidentes con el rango de variación existente en la variedad filipes. Los ejemplares Lundell & Lundell 10657, Swa- llen 1467 y Swallen 1506 se caracterizan por po- seer espiguillas menores y agrupadas en el extremo de las ramificaciones, hallándose la mayor parte de las ramas desnudas hacia la base. 16. Panicum hillmanii Chase, J. Wash. Acad. Sci. 14: 345. 1934. ue Estados Unidos de América. Texas: rillo, 11 ago. 1918, Hitchcock 16206 IE US-1037542). de 15—40 cm de Cañas erectas a geniculadas, arraigadas y Plantas anuales, cespitosas, alto. ramificadas en los nudos inferiores; entrenudos de 3-5 cm de largo, hirsutos hacia la porción superior; nudos pilosos, contraidos, castafios. Vainas de 4—6 cm de largo, usualmente más largas que los en- trenudos, verdosas o con tintes violáceos, densa- mente híspidas, con pelos papilosos, un borde pestañoso, el restante glabro. Lígulas membraná- ceo-pestafiosas, de 1-2 mm de largo; cuello piloso. Láminas linear-lanceoladas, de 5-20 cm de largo, 1 cm de ancho, planas, ascendentes, erectas, híspidas a esparcidamente pilosas en la cara ada- xial, de base subcordada a redondeada, el ápice agudo, márgenes escabriúsculos, largamente pes- tañosos hacia la base, con pelos papilosos, blan- quecinos, hasta de 5 mm de largo. Pedúnculos sub- incluidos en las vainas foliares, densamente híspidos hacia la porción distal. Inflorescencias la- xas, cortamente exertas, de 10-19 cm de largo, 10— 18 cm de ancho; eje principal anguloso, escabriús- culo, densamente híspido hacia la base o en toda su extensión; pulvínulos híspidos; ramificaciones de primer orden alternas a subopuestas, las infe- riores verticiladas, divergentes, ascendentes, esca- briúsculas, esparcidamente pilosas hacia la base; ramificaciones de segundo orden divergentes, con espiguillas adpresas sobre los ejes; pedicelos es- cabriúsculos, adpresos, de 0.3-1 cm de largo. Es- piguillas globosas, de 2.4-2.8 mm de largo, 1-1.1 mm de ancho, túrgidas, no estipitadas, pajizas o 2840 con tintes violáceos a ferrugíneas, glabras, de ápice acuminado; gluma superior y lemma inferior sub- iguales, tan largas como el antecio superior o su- perándolo hasta 0.4 mm. Gluma inferior aovada, de 1.2-1.6 mm de largo, % o un poco menor del largo de la espiguilla, de ápice acuminado, 3—5-nervia. Gluma superior de 2.4—2.7 mm de largo, 7—9-ner- via, el nervio medio escabriúsculo hacia el ápice. Lemma inferior glumiforme, de 2.2-2.7 mm de lar- go, 9-nervia, el nervio medio escabritisculo hacia el ápice. Pdlea inferior lanceolada, de 1.2-1.5 mm de largo, 0.8 mm de ancho, % del largo del antecio superior, glabra, hialina. Antecio superior elipsoide, de 1.8-2 mm de largo, 0.8-1 mm de ancho, crus- táceo, pajizo, castafio oscuro a la madurez, liso, lus- troso, con 2 cicatrices basales conspicuas, ca. 0.6 mm de largo, castafias a la madurez; lemma 7-ner- via; pálea con papilas verrugosas junto al ápice. Cariopsis ovoide, de 1.6 mm de largo, 0.6 mm de ancho, pajiza a blanquecina; hilo punctiforme; em- brión % del largo de la cariopsis. Distribución y ecología. Habita en los Estados Unidos de América, en los estados de California, Kansas, Oklahoma y Texas (Fig. 10). Crece en sue- los arcillosos en praderas; frecuente en bordes de caminos. Florece y fructifica entre mayo y octubre. Material adicional sri d aee UNIDOS DE AMERICA. California: Solar , 3 mi. S of Purah Creek, road 31, 23 June 1969, Rose a (MO); Ventura Co., Maricopa Rd. (route 399) near Meiners Oaks, Ojai Valley, 5 Oct. 1963, Pollard s.n. (MO-1970993). Kansas: Ulysses, 27 June 1893, Thompson 56 (US); Bucklin, 1892, Hitchcock s.n. (US-2434893); vicinity of Coldwater, along road 2 mi. W of town, 8 July 1929, Rydberg & Imler 734 seg Oklahoma: along railroad near Snyder, 19 Sep. 1903, Eggert s.n. (MO-2974983); Frederick, 26 July s Me 100 (MO); iy ord ~ Huntsville, 25 May | Heer ale s.n. (MO 2974984); Comanche Co., X Quanah Lake, s dà Mountains Ref- E 2 July n McMurry 978 (US); Washita Co., near . 16 June 1913, n 962 (US); Washita or Swan- od Stevens 1197 (US). Texas: Randall Co., Canyon, 2 line 1918, Palmer 13885, 14078 (MO); Brewster Co., Kokernott Ranch, 8 mi. NE of Alpine, 27 July 1941, War- nock 20925 (MO); Oklaonian, 1902, Reverchon (MO); prairie near Stanton, te (MO- 207499 0); € O); Nolan Co., Bus 7 Sep 1275630); e Oct. 1925, Kay 9180 (US); Abilene, 21 May 1902, Tracy 8295*,** (MO, US); sin localidad, 1889, Nealley s.n. (US-952982). — bce = — re A — = 8 LS © in = — m E ud Especie afín a P. stramineum y P. capillare. De P. stramineum se separa por tener ramas inferiores de la inflorescencia verticiladas, eje principal y los pulvínulos de la inflorescencia híspidos, pálea in- ferior reducida, ?4 del largo del antecio superior, y cicatrices conspicuas en la base del antecio supe- rior. De P. capillare se distingue por tener espi- Volume 83, Number 2 1996 Zuloaga & Morrone 249 Panicum Subg. Panicum Secc. Panicum guillas globosas, acuminadas, con pálea inferior de- sarrollada, % del largo del antecio superior y antecio superior con cicatrices conspicuas en su base. Cabe destacar que en algunos especimenes, como por ejemplo en Palmer 14078, la gluma su- perior es caediza, dejando al descubierto el dorso de la lemma superior. 17. Panicum hirsutum Swartz, Fl. Ind. Occ. 1: 173. 1797. TIPO: Jamaica, Swartz s.n. (holó- tipo, S, fragmento y foto, US-76922). Figura 21 Panicum chacoense ser lr 15: 102, fig. 8. 1969. TIPO: Argen Isla Bras TG 19 ene. 1963, Schulz nis ds BAA; isótipos, G, SD. Plantas perennes, robustas, hasta de 3 m de alto, cortamente rizomatosas. Cafías multinodes, postra- das, con raíces fulcrantes en los nudos inferiores, y erguidas en la extremidad, ramificadas en los nu- dos medios; entrenudos de 11—20 cm de largo, 0.4— 0.8 cm diám., cilíndricos, huecos, pajizos, con pe- los tuberculados urticantes, nudos contraidos, violáceos, pilosos. Vainas de 14-20 cm de largo, mayores o menores que los entrenudos, estriadas, hirsutas, con pelos urticantes caedizos y de base tuberculada, adpresos, verdosas o con tin- tes violáceos, de bordes membranáceos, glabros o con uno de los márgenes pestafioso. Ligulas mem- branáceas en la base y cortamente ciliadas en la porción superior, de 0.2-0.7 mm de largo, con lar- gos pelos por detrás en la base de la lámina; cuello piloso. Láminas lanceoladas, de 20-50 cm de largo, 1.5-3.5 cm de ancho, planas o con los bordes in- volutos, de base subcordada a cordada, glabras o con escasos pelos esparcidos, el nervio medio ma- nifiesto, los bordes escabriúsculos. Pedúnculos ci- líndricos, hasta de 25 cm de largo, glabros. Inflo- rescencias terminales, laxas, contraidas a difusas, multifloras, de 25-45 cm de largo, 5-15 cm de ancho; eje principal anguloso, escabroso; pulvínulos levemente pilosos a glabros; ramificación inferior verticilada, las restantes opuestas o alternas, ad- presas al raquis a divergentes, las espiguillas ad- resas, cortamente pediceladas sobre las ramas; pedicelos de 0.5-2 mm de largo, escabrosos. Es- piguillas angostamente ovoides, de 1.8-2.6 mm de largo, m de ancho, no estipitadas, gla- bras, pajizas, gluma superior y lemma inferior sub- caducos; iguales, tan largas como el antecio superior. Gluma inferior de 0.7—1.4 mm de largo, ovada, % de la longitud de la espiguilla o algo menor, 3—5-nervia, cortamente pilosa en el ápice de la cara interna o totalmente glabra. Gluma superior de 1.7-2.2 mm de largo, 7-11-nervia. Lemma inferior de 1.7-2.2 mm de largo, 7-9-nervia, glumiforme. Pálea infe- rior lanceolada, de 1.3-1.7 mm de largo, 0.3-0.4 mm de ancho, % del largo del antecio superior, membranácea, hialina, con los bordes denticulados. Antecio superior m elipsoide, de 1.2-1.6 mm de largo, 0 m de ancho, glabro, liso, lustroso, pajizo, castafio a s madurez, con 2 cica trices en la base de la lemma de 0.1 mm de ) d castafias a la madurez; pálea con papilas simples junto al ápice. Cariopsis elipsoide, de 0.8 mm largo, 0.5 mm de ancho, pajiza; hilo punctiforme; embrión menos de la mitad del largo de la cariop- sis. Distribución y ecología. Originalmente descrita para Jamaica, crece en el sur de los Estados Unidos de América y México, hallándose abundantemente en Mesoamérica y las Indias Occidentales, habien- do sido hallada en mucha menor proporción en Su- damérica en Colombia, Venezuela, Ecuador, Perú, Brasil y la Argentina (Fig. 4). Habita en lugares húmedos a lo largo de cursos de agua o en bordes de pantanos, entre los 30 y 1800 m. Nombre vulgar. “Cortadera,” “gamelote” (en Venezuela, de acuerdo al ejemplar Tamayo 1768). Material representativo citado. IGUA. Donovans pe Box 176 (US). Santa Lucia: Mabouya Valley, Box 0 (US). ARGENTINA. Chaco: Isla Brasilera, Schulz Hob (CEN, F). BELICE. Cayo: 9.5 miles S of Ge ville, on road to Agustine, 30 May 1973, Croat 23483 (MO); vicinity of Millionario between the McCal River and C y 34 (US); Hole ep. 1 5 Aug. 1980, — 2090 (MO). Or- 14 mi. E of San Antonio, 8 Dec. 1976, Leino 76105 (MO). Stann Creek: ated Bird Highway, 13 Sep. 1954, Gentle 8380 (F, G, MO, NY, US). Toledo: Pate’s Camp, Edwards Road beyond Columbia, Gentle 7120 (F, G, NY, US); Maya Mountains, canyon along Blad- en Branch from Richardson Creek to Quebrada de Oro, 16?32'N, 88?46'49"W, 100-200 m, 12 Mar. 1987, Davidse £ d 32401 (MO). Sin localidad, Smart 112 (F). BRA- ,. Para: Lower Amazon, Estate Cacanal Grande, Goeldi PA 128 (US). COLOMBIA. Bolivar: along the Caño Chacagua, N of Los Piñones, Island of Mompós, Lands of Loba, Curran 261 (US); along the Mompós-Juana Sánchez trail, Island of Mompós, Curran 250 (US); Canabetal, Pen- nell 3866 (GH). Cundinamarca: La Mesa, 1200 m, Tri- ana s.n. (US-1865276). Magdalena: alrededores de Cién- aga, Romero Castañeda 8948 (MO, NY); Santa Marta, Smith 164 (GH, K, MO, P, NY, US). Santander: Boca Sogamoso, Río Magdalena, 110 m, Pennell 3849 (GH, NY, US); prope Billete Blanco, Woronow et al. 4691 (US). Sin estado ni localidad, Mutis 2151 (US). COSTA RICA. He- redia: Finca La Selva, 100 m, 21 abr. 1982, Hammel 11795 (MO). Limón: near Puerto Limón, in plantations, Pittier 3634 (NY, US); margin of Bonilla Lakes, above Tunnel Camp, 13 Dec. 1929, Dodge et al. 5610 (MO). Annals of the rior.—C. Espiguilla vista del lado de la gluma superior.—D. CUBA. Cienfuegos: Soledad, 3 s (NY, US). Ha- a: Guanimas, on beach, 2 1929, León 14181 d ri . € 6183, 1 (US); Preston, on the railroad, 13 Nov. 1914, Ekman 3419 (NY, US). Pinar del Río: Mangas, Pueblo Nuevo, Hato del Corojal, 9 Oct. 1923, Ekman 17578 (US). 250 Missouri Botanical Garden xe Yr pea X 4 5 ORY EIN ey AE vt H WE Figura 21. Panicum hirsutum.—A. Porción de una caña florífera.—B. Espiguilla vista del lado de la gluma infe- Antecio superior visto de la pálea. (Schulz 12138.) Villa Clara: Caibarién, Ekman 16312 (US); Cienaguita, s 259 (NY). DOMINICANA. Barahona: valley of Neiba, eu 8357 (US); Barahona, Río Yaque del Sur, Ekman 5791 (US). Duarte: valle del Cibao, Pimentel, at Río Cuaba, odis 13275 (US). Samana: Sánchez, Ji- ménez 4309 (US). ECUADOR. Guayas: Panigón Planta- tion, 8 miles S of Milagro, 50 m, Hitchcock 20580 (GH, Volume 83, Number 2 1996 Zuloaga & Morrone 251 Panicum Subg. Panicum Secc. Panicum NY, US), 20601 (US). ESTADOS UNIDOS DE AMERICA. Texas: Cameron Co., vicinity a em E Aug. 1921, Ferris & eee 3175 (MO, US); Cameron Co., Southmost, 1941, dinde s.n. e 1273 3655); Río "Rico. near Río Gra ide. about 6 miles from Me dn Silveus 625 (US); Brownsville, i Apr. 1917, Nelson s.n. (US). GUADALUPE. Entre la Rivierc- Salée et St. oe Duss Mean 49430 (F). Izabal: F miles E of El Estor on Lake Izabal, 13 Sep. 1961, Popenoe 34 (F); seashore around Punta Palma, across bay from Puerto Barrios, 22 Apr. 1940, Steyermark 39802 (F). Pe- tén: Parque Nacional Tikal, aguada hotel Posada la Selva, 30 abr. 1970, Ortíz 1049 (F, NY); ruinas Plaza May- or Tikal, 333 m, 16 nov. 1965, Molina 15281 (F, NY); 35 km E de Santa Elena, brecha El Remate-Tikal, 200 m, Molina 15426 (F, NY, US); Tikal National Park, Tikal, growing around aguada at Camp, Lundell 15918 (US). GUYANA. Within 30 mi. of Georgetown, Rodway s.n. (US). HAITÍ. Massif du Nord, Ekman 4392 (US); Plaine de LArtibonite, Dessalines, Ekman 8534 (G). HONDU- RAS. Atlántida: La Fragua, 20 m, Feb. 1928, Standley 55711 (F, US); Tela, meag Lancetilla valley, 6 Aa 1927, Standley 53428 (F, US). Francisco Morazán: San An- tonio de Oriente, 22 ui 1945, Rodríguez 3126 (F). MEX- ICO. Campeche: Tuxpeña, 18 Feb. 1932, Lundell 1355 (F, GH, MO, NY, US); 70 km al E de Escárcega, 26 mar. 1982, Cabrera & Cabrera 2237 (MEXU, MO). Chiapas: slopes and small streams with Tropical Rain Forest along the ridges 6-12 km south of Palenque on the road to Oco- singo, 300 m, 12 Oct. 1972, Breedlove 28837 (MEXU, MO). Colima: Manzanillo, sea level, Hitchcock 7031 (RB, US). Oaxaca: San Antonio, 2 Sep. 1894, Pringle 5573* (F, GH, MEXU, MO, US); District of Tuxtepec, Chiltepec and vicinity, 200 m, Martínez Calderón 568 (US). Quin- tana Roo: 4 km NW of J.M. Morelos, along Hwy. 184, at the place where it crosses the nho of Laguna Chi- chancanab, 18 May 1982, Davidse et al. 20612 (MEXU, ae San care de la Montafia, 9 km S of Hwy. 186 on Hwy. to Tomás Garrido, 120 m, 9 May 1982, Davidse et al. 20271 “MEXU. MO). San Luis Potosí: bank of Río Valles, 27 June 1910, Leavenworth 199 (GH, MO). Ta- basco: Boca Cerro, Tenosique, 1 jul. 1939, Matuda 3577** (F, GH, US); km 21 de la carretera Cárdenas— Coatzacoalcos, 17 sep. 1981, Magaña 427 (MEXU, MO). Veracruz: Estación Biológica San Andrés Tuxtla, 27 nov. 1978, 150 m, Martínez Calderón 1798 (F, MEXU, MO). Yucatan: Becanche, 19 ene. 1956, Enriquez 347 (MEXU). NICARAGUA. Zelaya: vicinity of Waní, includ- ing Río Ulf, ca. 13%41'N, 84%50'W, 90-100 m, 22 ae 1978, Stevens 7982 (MO). PANAMA. Bocas del Toro km NW of Bocatorito on SE side of island, 9 Feb. E Peterson & Annable 6729 (MO). Colon: 4 km E of Buena Vista, Nee 6783 (MO). Darién: Darién, Sep. 1967, Duke 14117 (F, MO, NY); Río — Leopold 109 (MO). Pan- ama: past Chepo and Río Mamoni on road to El Llano, 25 Sep. 1972, "Tyson 6791 (MO). Zona del Canal: Be- and Tabernilla, Hitchcock 7964 (F, ia = MO, P, US). PERU. Ucayali: Canchahuayo, Vásquez 7025 (MO, SI). TRINIDAD Y TOBAGO. Little Tobago Island, 26 July 1914, Broadway 4895 o 8 miles S of Mayaro, along Mayaro-Guayaguayare road, 1 Aug. Davidse 2510 (MO); Parrott Hill, 22 July a Hoodia 4586 (MO, NY, US). VENEZUELA. Anzoategui: Pari- aguán, El Tigre, 16 jul. 1946, Oropeza s.n. (VEN). Bari- a; nas: Sistema de riego Río Boconó, Trujillo 11399 (MY). io intervenido, Ra- jillo 11492 (MY); 7 km (MO). Guárico: cie era ia he corro, 5 km antes del últim 1989, Zuloaga e et al. 4507* (MO. SI, VEN). 4 Rfo Grande Tuy, above sip ud Pittier 6332 (US). Portuguesa: Dito. Araure, en terrenos adyacentes a la vía que une Agua Blanc ca con Pu Rares Reyes 2051, de Esmujaque, cda en las orillas del Río Motatán, pin (id 176 8* (Us, VEN). Yaracuy: Carretera entre Yaritagua y Barquimi- seto, Burandt Jr. V0238 (MO, VEN); San Felipe, orillas arenosas del Río Yaracuy, 11 abr. 1946, Burkart 16447* SI). Zulia: vicinity of Mene Grande, Pittier 10620 (GH, US, VEN); San Martin, on Río del Palmar, Pittier 10534 (GH, NY, US, VEN). — Esta especie es abundante en América Central y Colombia y Venezuela, siendo ocasional en el norte del Brasil. Sólo fue registrada una colección en Argentina, en las cercanías del límite con Para- ay. Difiere del resto de especies de la sección por ser una planta de gran porte, con panojas amplias que llevan espiguillas pequefias y por poseer vainas con pelos urticantes. Gould (1975), puntualiza que los pelos silíceos presentes en las vainas pueden causar irritación en la piel al contacto con los mis- mos. 18. Panicum hirticaule J. Presl, Reliq. Haenk. 1: 308. 1830. Panicum polygamun Sw. var. hirticaule (J. Presl) E. Fourn., Mexic. Pl. 2: 28. 1886. Panicum capillare L. var. hirticaule (J. Presl) Gould, Madrofio 10: 94. 1949. TIPO: México. Guerrero: Acapulco, Haenke s.n.** (holótipo, PR t no visto, fragmento, US-80698; isótipo, MO-1837663). Plantas anuales, de 25-120(-150) cm de alto. Cafias decumbentes, arraigadas en los nudos infe- riores a erectas, ramificadas hacia la base; entrenu- dos de 4—21 cm de largo, comprimidos, surcados, hirsutos; nudos cortamente hirsutos. Vainas de 3— 10 em de largo, más cortas que los entrenudos, hirsutas, con pelos tuberculados, verdosas a púr- puras, con un margen pestafioso, el restante glabro. Lígulas membranáceo-pestañosas, de 0.9-1. de largo; cuello hirsuto. Láminas lanceoladas a li- near-lanceoladas, de 6-25(-31) cm de largo, 0.4— 1.6(-2.2) cm de ancho, planas, hirsutas a esparci- pilosas, con pelos blanquecinos, tuberculados, de base subcordada, amplexicaule, el mm damente ápice subagudo, con los márgenes papiloso-pesta- ñosos. Pedúnculos subincluidos en las vainas foli- ares a exertos, hasta de 36 cm de largo, híspidos a Annals of the Missouri Botanical Garden glabros. Inflorescencias terminales, laxas, de 9-26(— 30) cm de largo, 5-18 cm de ancho; eje principal anguloso, escabroso, glabro a esparcidamente pi- loso junto a la base; pulvínulos glabros; ramifica- ciones de primer orden alternas a subopuestas, en ocasiones subverticiladas, divergentes, glabras a esparcidamente pilosas hacia la base; ramifica- ciones de segundo orden adpresas con espiguillas adpresas; pedicelos de 0.9-2.7 mm de largo, es- cabriúsculos. /Inflorescencias axilares presentes, similares a las terminales. Espiguillas ovoides, de 1.9-2.8(-3.5) mm de largo, 0.8-1 mm de ancho, abruptamente acuminadas, no estipitadas, glabras, cobrizas a violáceas; gluma superior y lemma in- ferior subiguales, superando hasta 0.8 mm en largo al antecio superior. Gluma inferior ovado-acumi- nada, de 1.3-2.4 mm de largo, 4% del largo de la espiguilla, 3—5-nervia, nervio medio escabriús- culo a liso. Gluma superior de 1.8-2.9(-3.3) mm de largo, 7-9(-11)-nervia, con el nervio medio esca- briásculo hacia el ápice a liso. Lemma inferior glumiforme, de 1.8-2.8(-3.3) mm de largo, 9-ner- via. Pálea inferior lanceolada a ovado-acuminada, e 0.4—0.9 mm de largo, 0.2-0.4 mm de ancho, 4— Y del largo del antecio superior, membranácea, hialina, glabra. Antecio superior elipsoide, de 1.5— 2.3 mm de largo, 0.7-0.9 mm de ancho, crustáceo, glabro, liso, brillante a opaco, pajizo a negruzco, papiloso en toda la superficie, con papilas simples, o sólo presentes hacia el ápice de la pálea, con dos cicatrices basales ca. 0.2 mm de largo, castafias a la madurez; lemma 7-nervia. Cariopsis elipsoide, de 1-1.7 mm de largo, 0.7-0.8 mm de ancho, blan- quecina; hilo punctiforme; embrión la mitad del largo de la cariopsis. 18a. Panicum hirticaule var. hirticaule Panicum capillare * mee miliaceum Vasey, Contr. U.S. atl. Herb. 1: 890, non Panicum miliaceum L., 153. Panic um sonorum Beal, Grasses N. Amer. 2: 130. 1896. Panicum hirticaule J. Presl var. milia- ceum (Vasey) Beetle, Phytologia 47: 381. 1981. TIPO: México. Sonora: Lerdo, 1889, Palmer 947 (holótipo, US-2903025). Panicum hirticaulon J. Presl var. glabrescens Andersson, Kongl. Vetensk. Acad. Handl. 1853: 135. 1855. TIPO: Ecuador. Archipiélago de Colón: Isla Chath- an, Andersson s.n. (tipo no visto). Espiguillas de 1.9-2.8(-3.3) mm de largo. An- tecio superior pajizo, brillante, liso, con papilas sim- ples hacia el ápice de la pálea. “zacate de año,” “chiri-chiri” Nombre vulgar. “Triguillo,” “zacate peludo perdis” (en México); (en Perú Distribución y ecología. Abundante en el sur de Estados Unidos de América, México y Mesoaméri- ca; menos frecuente en América del Sur donde cre- ce en Ecuador, Perá y Venezuela (Fig. 11). Crece en ambientes abiertos, sobre suelos modificados, desde el nivel del mar hasta los 2500 m. bs Saher dbi citado. COSTA RICA. Guan- aste: Playa S end of bay, 6 July 1968, Pohl & pom 10053 E ISC, LA); upper slopes of bluffs above Pacific S of Playas del Coco, 5 Aug. 1966, Pohl & a 10216 (ISC, MO). Puntarenas: vicinity of Cascajal, along road from Cascajal to Pigres, 30-100 m, 6 Bias 1949, Holm & Iltis 286 (GH). ECUADOR. Ar- i o de Colon: Isabela Island, Volcano Alcedo, Fow i 8d (MO): Indefatigable Island, 6 mi. north of Acad- emy Bay, 10 Apr. 1930, Svenson 236 (F, GH, LA). Guay- as: Caperia, km 21, Guayaquil to Daule, 20-200 m, 15 Feb. 1982, Dodson & yi entry 12506, 12514 (MO). EL SALVADOR. La Libertad: e road ca. 1 km NE of ¿A-1, on road 9 km WSW of Quetzaltepeque, 18 Juno 1978, Pohl & Gabel ak (F, ISC). Santa Ana: agunita Clara, 5 km S of Metapán, 500 m, 10 June 1970, de & Davidse 11872 (F, ISC, MO); along hwy. 12, 9 km N of Texi d 550 m, 25 June 1978, Pohl & Gabel 13665 (F, ISC, MO). San Salvador: Apulo, in pasture slopes near Lake giu 18 June 1949, idi hs Mo- lina 16751 (F). Si vineia: San Andrés, ) ft. 1943, Walkins 17 (US); La Cabaña, e m, 12 juy 198, Walkins s.n. (US-2437900); west side of Lake Hopango, 2 Nov. 1911, Hite a 8924 (US). ESTADOS UNIDOS AMERICA. Arizona: Pima Co., n of trail to top of Mt. Baboquivari, j: 2d 1947, Gould uen (LA, ma Gila County, Pocket ;zanyon, Sierra Ancha Moun tains, 29 on 1946. Could & Hudson us GH, LA). Na Mexico: Cedar Spring, Kos rd 1078 (MO); 1 mi. W of Hillsboro, 6000 ft., 2 Oct. 1904, Metcalfe 1442 (GH, — < ~ 2 = — -— 3 = =. a © A, . Texas: SE f Col 1940, E vans s.n. (US-2434928); El Paso, 10 Sep. ; Jones LA-39220 on: W. Klickitat Co., € land near Bingen, Nov. 1894, Suksdorf 2330 (GH, LA, MO). GU ngos A. Izabal: Los Amantes, 9 mayo 1919, Blake 7326A (US); vicinity of Puertos Barrios, 2 July 1922, undi s.n. (US). Suchi 5 (F, US). Zacapa: rock Pepezca, 200-250 m vicinity of Zacapa, 200 m, 6 Oct. 1940, Sta ur (F). Sin E Weatherwax 79a (US); between San Andrés and Muluá, 12 Apr. 1932, Weatherwax 217 (U5). HAITÍ: Plaine Cul- d Port-au-Price, Damine, in fields, 25 mayo 1928, Ekman 999] (G, US). HONDURAS. Choluteca: 7 km road NE of Choluteca, 50 m, 4 July 1970, Pohl & Davidse 12177** (F, ISC, MO). Comayagua: Comayagua, Esta- ción Agrícola El Carao, 630 m, 30 jul. 1983, Casco 46 . Cortés: San Pedro Sula, Hacienda Monte Moria, 16 abr. 1984, Cristoff 181 (NY). Francisco Morazán: Tegucigalpa, cerro El Berrinche, 1000 m, 6 ago. 1978, Castro 69 (MEXU, MO). Valle: 4 km SE of San Lorenzo, )-5 m, 5 Oct. 1986, Davidse & Pilz 31665* (MO). MEX- ICO. Baja California Sur: km 65 east of San Antonio, 6 June 1973, Beetle M-2551 (MO); Mesa de San Geróni- mo, northerly from Rancho Viejo (on road from Loreto to San Javier), 23 Sep. 1965, Carter 5016 (LA, MEXU). ~ Volume 83, Number 2 199 Zuloaga 8 Morrone 253 Panicum Subg. Panicum Secc. Panicum Chiapas: marshes along Rio Usumacinta, 25-30 km east northeast of Palenque Junction, 30 m, 12 Dec. 1981, Breedlove 56081 (MO, NY); forest at Mirador for Chicoas- en Dam along road from Tuxtla Gutiérrez to the Chicoasen X bag m, 17 Nov. 1976, Breedlove 41540 (MO). Chi- : N end of E side of Sierra del Cuchillo Parado, 2937" N, 104°5 55'W, 1000-1300 m, 21 Oct. 1972, Wendt et al. 9788 (MO, NY); 20 Km S of Ciudad Camargo, 1220 m, 1 July 1939, Harvey 1378 (GH, MO, US); Guasaremos, Rfo Mayo, 20 Sep. 1935, Gentry 184] (MEXU, MO, US). Chi ua-Sonora: Rancho Carreteras, 1460 m, 26 ago. 1939, Harvey 1600 (MO, US). Colima: Alzada, 21 Sep. 1d €— hg (US); se July 1897, Palmer 143 (ISC, S, W). Durango: 10 km N of Ceballos and 84% S P du. state line. on the Torreon-Chihua- hua highway, 1200 m, 26°36'N, 102°12’W, Johnston et al. 12292 (MO). Guerrero: south Chilpacingo, 7 Dec. 1941 Leavenworth 966 (MO); Río Balsas, 26 ago. 1910, Orcutt 4197 (MO). Guanajuato: Irapuato, 5800 ft., 1 Oct. 1910, Hitchcock 7424 (US). Jalisco: ca. 5 road-miles SW of Santa Cruz de las Flores, 1550 m, 24 Aug. 1957, Mc- Vaugh 16326 (MEXU, NY, US); Tecalitlán, al S de Te- calitlán, rumbo a Pihuamo, 12 Aug. 1987, Ornelas U. et al. 1167 (MEXU). Mexico: Temascaltepec, Anonas, 24 July 1934, — pad (MO, NY, US); Temascaltepec, 000 m, Hinton 1062 (F, MEXU, NY, ; Michoacán, Chavinda, 27 Sep. 1946, Hernández Xolocotzi et al. x-2775 (US); at edge of tropical deciduous forest, La Majada, 8 May 1941, Leavenworth & Pap ni 1335 (GH, NY). Nayarit: 1 mile west of Mazatán, 600 m 17 Sep. 1960, McVaugh 19117 (NY, US); near ve al, 5 miles N of Compostela, 22 Sep. 1960, McVaugh 19333 (NY, US). Oaxaca: 2 km al SE de San Martín Toxpalán, por carretera a Teotitlán del Camino, 26 ago. 1980, Med- rano et al. F-1495 (MO); Los Tules Niltepec, 11 dic. 1985, Torres C. et al. 7860 (MEXU, MO). Sonora: El Rancho de la Nacha, 25 mi. W of La Angostura, 4300 ft., 14—20 Aug. 1941, Vera So 1850 (GH, LA, MO, NY); 27 miles west of Hermosillo on the road to Kino Bay, 720 ft., 28 Aug. 1941, Wiggins & Rollins 137 (GH, LA, MO, NY). Tabasco: km 6.6 de Emiliano Zapata hacia Tene sique, 5 jul. 1981, Cowan 3398 (MEXU, MO). Veracruz: La Mancha, carretera Cardel-Nautla, 1 ago. 1971, Dor- antes 246bis (F, GH); Baños del Carrizal, Purpus A (GH, LA, MO, NY). Yucatán: Chichankarrab, Gaume 1501 (F). NICARAGUA. Esteli: Kukamonga km 167, Portal de Belén, 1 ago. 1983, Moreno 21833 (ISC, MO); faldas del Chayo 13°16'N, 86°20’W, 700-1000 m, 31 jul. 1983, Moreno 21790 (ISC, MO). León: along H m S of road to Puerto Somoza, 50 m, 18 July 1970, Pohl 1977, md 2672 (F, IS SC, MO); Rte. l, Managua, | Aug. in Seymour 6275 (MO). Masaya: km 17.5 carretera Managua-Masaya, terrenos de la UCA, 12%01'N, 86?10'W, 10 jun. 1982, Sandino 3038 (MO). i Maderas, cafetal Las Cuch- as, 11°28’-29'N, 85729—30'W, 200-300 m, 23 sep. 1984, Robleto 1269 S MO, SI); along CIA, near Lago de Nicaragua, 24 SE of Rivas, 1 Aug. 1971, Pohl 12674 (F, ISC, MO). "PANAMA. Panama: Taboga Island, Dec. 1923, Standley 27960 (US). Zona - Canal: Chiva Chiva trail near Miraflores Lake, 30 . 1965, Tyson 1396 (MO); Balboa, Sosa Hill, Dec. Ps Nando 25277 (MO, US). PERU. Cajamarca: Prov. Cajamarca, alrede- dores de Choropampa, sobre la carretera de penetración Pacasmayo-Cajamarca, a la altura del km 124, anual, en bordes de campos de cultivo y laderas no cultivadas, 7 mar. 1981, Sánchez Vega 2363* (MO, SI, US). Huánuco: Huánuco, in cornfield, loose clumps, subdecumbent be- low, 28 Apr. 1923, McBride 3526 (F, US). La s in Prov. Trujillo, Chual-Cruz de Casc ots [d zii 3 1953, Lopez 1e 951 (US). P : Pro al al- rededores de a, 15 oct. 1983, Sands 1 oo (MO, ; 40 km west o Piura Horton 11361 (F, MO). WA SE corro, 15 July 1946, Hulkart 17204 (SI); quie. i di María de Ipire a El Socorro, 08°58'N, 65?40'W, 150 m, 17 ago. 1989, Zuloaga et al. 4508* (MO, SI, VEN). Especie afín a P. stramineum y P. ghiesbreghtii. De P. stramineum se distingue por poseer pálea in- ferior reducida, espiguillas cobrizas a violáceas, dispuestas densamente en las ramificaciones, sien- do estas adpresas. Panicum ghiesbreghtii se dis- tingue por ser una especie perenne, con láminas más angostas y de mayor tamaño, de 16—30(—55) cm de largo. El tamaño de las espiguillas en P. hirticaule nor- malmente varía entre 1.9 y 2.8 mm de longitud. Se observaron algunos ejemplares de México y Esta- os Unidos con espiguillas hasta de 3.5 mm de longitud, los que poseen asimismo la base del an- tecio superior redondeada. Existen también varios especimenes del Pert, de los departamentos Huá- McBride 3526, López Miranda 951 y Ridoult 1919), de mayor porte y con espiguillas de aproximada- mente 3.3-3.5 mm de largo. Hitchcock (1927) cita P. hirticaule para el Perú sobre la base de dos ejemplares: Hitchcock 22115 y McBride 2536. El a P. exiguum y el sigues a P. hirticaule con es- piguillas mayores. Swallen (1943) cita esta especie primero de estos corresponde para Bolivia, sin mencionar ejemplar de herbario; probablemente esta cita corresponde a P. strami- neum. Swallen (1964) y Tovar (1993) citan la presencia de P. hirticaule en Argentina sin mencionar ejem- plares de referencia. 18b. Panicum hirticaule var. verrucosum Zu- loaga & Morrone, var. nov. TIPO: México. Chi- huahua: between Casas Grandes and Sabinal, 4 Sep. 1989, Nelson 6355 (holótipo, US; isó- tipo, GH). Panico hirticauli affine sed ao ‘io papillato, papillis verrucosis omnino paginae diffe Espiguillas de 2.4-3.5 mm de largo. Antecio su- perior negruzco, opaco, con conspicuas papilas sim- ples en toda su superficie (Fig. 1A, B). Distribución y ecología. Sur de los Estados 254 Annals of the Missouri Botanical Garden Unidos, en Arizona, y norte de México, en los es- tados de Chihuahua, Guerrero, Jalisco, Nayarit, Sinaloa y Sonora (Fig. 3). Se halla en campos; llega hasta los 600 m. Nombre vulgar. “Zacahuastle” (en México). Material adicional examinado. ESTADOS UNIDOS DE AMERICA. Arizona: near Fresnal, Papago Indian Reservation, 30 Sep. 1934, Peebles 10366 (GH, US); Pima South Canyon, Baboquivani Mts., 15 Sep. 1931, n Mountains, 25 Aug 1904, Grif- US S); 7 miles N of Nogales, US high- > ~ > wo a Na) © SÍ N > bn | "S way 89, Ruby Aug. 1959, rius 160 (US); north slope of Santa Rita Mountains, 13/21 Sep. 1904, Griffiths 7194 (US). MEXI- Chihuahua: c asics Hemanacs & Tapia 240* o Guerrero: Cuautepec, $ sie oni Cuilutia, 10 jul. 1986, Herrera Castro 84 (MEXU). Jalisco: Amatitán, Bar- ranca Santa Rosa, orilla del M Santiago, 600 m, 4 ago. 1977, Villarreal 6661* (MEXU). Nayarit: El Nayar, Jesús María, m go. 1977, Colunga & Zizumbo 44** (MEXU). Sinaloa: alrededores del dique La Primavera, selva baja caducifolia a 10 E 2 sur de Culiacán, 60 m, si 1248 (MEXU). Sono in loc: pres 17 sep. 908, Hitchcock 3526 (US), Wiggins 6023 (US). = En el ejemplar Gilman 37 las papilas son dimi- nutas, el antecio superior es brilloso la gluma superior se cae en algunas espiguillas Una porción del ejemplar Griffiths 69, 39 1/2 co- rresponde a esta variedad (las otras dos plantas de esta colección son P. pampinosum). Hitchcock € Chase (1910) y Fairbrothers (1953) lo citan dentro de P. hirticaule. En el ejemplar Peebles 10366 las espiguillas lle- gan excepcionalmente a tener 3.8 mm de largo. 19. Panicum hispidifolium Swallen, Contr. U.S. Natl. Herb. 29: 424. 1950. Panicum de dai Swallen, Contr. U.S. . Herb. 24 1949, non G. Forst., ae TIPO: Honduras. El Paraíso: Güinope, 1430 m, l ene. Rodríguez 1981** (holótipo, US-1869144; isó- tipos, F-1166244, 1306220). Figura 22. Panicum d gre E. Fournier, Bull. Soc. Bot. France, 27: 293 30, non Steud., 1853. TIPO: Nic- a. Managua: Isla de Omotepe, oct. 1869, Levy 1166 (holótipo, P, fragmento, US-80615; isótipos, G, p) Plantas anuales, herbáceas. Cafías cespitosas, erguidas, simples, de 20-100 cm de alto, genicu- ladas o no en los nudos inferiores; entrenudos cilfn- 4—15 cm de largo, 0.1—0.4 cm hirsutos, con pelos tuberculados caedizos; dricos, huecos, de diám., nudos pilosos, violáceos. Vainas de 4—13 cm de largo, menores o mayores que los entrenudos, den- samente hirsutas, con largos pelos tuberculados, los márgenes pestañosos. Lígulas cortamente mem- branáceas en la base, luego largamente ciliadas, de 2-3 mm de largo; cuello piloso, violáceo. Láminas linear-lanceoladas, de 8-45 cm de largo, 0.8-1.3 cm de ancho, planas, herbáceas, acuminadas, sub- cordadas o de base redondeada, hirsutas, con bor- des escabrosos, ciliados hacia la base, el nervio medio manifiesto. Pedúnculos incluidos en las vai- nas foliares a exertos, hasta de 40 cm de largo, cilíndricos, hirsutos. Inflorescencias terminales ex- ertas; panojas laxas, difusas, piramidales, multi- floras, de 12-50 cm de largo, 10-40 cm de ancho; eje principal anguloso, híspido, cubierto de gruesos pelos de base tuberculada, glabro hacia la porción superior de las panojas; pulvínulos pilosos; ramifi- caciones opuestas o alternas, las inferiores ocasio- nalmente subverticiladas, divergentes, hasta de 17 cm de largo; ejes de las ramificaciones híspidos, con espiguillas distantes, no adpresas; pedicelos largos, de 0.: mm de largo, escabrosos. Inflo- rescencias axilares presentes, semejantes a las ter- minales. Espiguillas globosas, abiertas en el ápice, e 3.2-3.7 mm de largo, 1.2-1.6 mm de ancho, solitarias, no estipitadas, glabras, verdosas y con tintes violáceos, agudas a acuminadas, glumas y lemma inferior con nervios marcados; gluma su- perior y lemma inferior subiguales o superando rasta 0.3 mm en largo al antecio superior. Gluma inferior ovada, de 2-2.6 mm de largo, %—% del lar- go de la espiguilla, aguda, 5-nervia, el nervio medio escabroso, escabriúscula hacia el ápice de la cara interna. Gluma superior de aproximadamente igual largo que la espiguilla, aguda, 7-nervia, el nervio medio escabriúsculo hacia el ápice, escabriúscula en el ápice de la cara interna; gluma inferior y superior separadas por un corto entrenudo ca. 0.3 mm de largo. Lemma inferior glumiforme, 7—9-ner- a aguda, pilosa en el ápice de la cara interna. ?álea inferior ovado-lanceolada, reducida, de 1.6— i mm de largo, 0.6-0.9 mm de ancho, % del largo del antecio superior, hialina, los bordes superiores esparcidamente pilosos a glabros, glabra en el resto de la superficie. Antecio superior anchamente ovoi- de, de 2.1-2.5 mm de largo, 1.2—1.5 mm de ancho, crustáceo, glabro, liso, lustroso, con tintes grisáceos a la madurez y un anillo en la base, castafio a la madurez; lemma 7-nervia; pálea con papilas verru- gosas hacia el ápice. Cariopsis anchamente ovoide, de 1.6—1.8 mm de largo, 1.1—1.3 mm de ancho; hilo punctiforme; embrión menos de la mitad del largo de la cariopsis. Distribución y ecología. Se halla desde México y América Central, en El Salvador, Honduras, Nica- ragua, Costa Rica, hasta Colombia y Venezuela (Fig. 8). Crece en sabanas; es frecuente en borde Volume 83, Number 2 Zuloaga & Morrone 255 Panicum Subg. Panicum Secc. Panicum 1996 acia. SE ES E o.—B. Espiguilla vista del lado de la gluma infe mu hee n lado de Pálea inferior.—E. Antecio superior visto del la lemma.—F. Antec Cariopsis, vista hilar. Tm ^ al. 4363) a 22. Panicum hispidifolium.—A. Habit vista del lado de la gluma superior.—D. superior visto del lado de la pálea.—G. Cariopsis, vista escutelar.—H. 256 Annals of the Missouri Botanical Garden de caminos. Habita entre los 60 y 1430 m; florece y fructifica entre julio y diciembre. Material representativo citado. COLOMBIA. Atlanti- co: Puerto de Colombia, Hitchcock fae (US). Magdale- na: Santa Marta, Smith 2152 A. GH, 5 i . NY, P US). COSTA RICA. Guanaca km Hagens along the Carretera a ‘ana, . 100 m, E Oct. 1968, Pohl & Davidse 11291 (F, MEXU); 5 km S of La Cruz, 200 m, Pohl & Gabel 13703 (F, MO); 5 km SE of Liberia, on ins Cañas road, 1966, Harris 40* (F). EL SALVADOR. razán: 2 km W of Las Delicias, 13 Sep. 1951, 650 m, Rohweder da (MO). O Cortes: Ocote Arrancado, 5 1 N Lago de Yajoa, 600 m, bosque trop- ical pees nov. "1980, Nelson E ud 5621 (MO). El Par- aiso: Road to Yuscarán, moist p aces Sees rocks, about km 2, Swale en 11344 (US). Franci orazán: road between El Jicarito and El Pedregal. 800-950 m, 13 Nov. 1948, Sandie 92 (F); vicinity of El Zamorano, Swal- len 11264 (F, MO, US); vicinity of Las Mesas, ca. E of Zamorano, P 1986, Davidse n Pilz 31533* (MO. SI). MEXICO. Chiapas: 58 km south of Mexican Highway 190 on road to Nueva Concordia, 900 m, 11 Sep. 1974, Breedlove 37573, 37654 (MEXU, MO). Oaxaca: near road nine miles toward Tehuantepec from Zanatepec, Mor- ley id (F. e ve MO). NICARAGUA. Masaya: Piedra Quemada, NW of Volc i Santiago, n m, 28 ee 1976. m 1115 > (MO). VENEZUELA. Anzoátegui: km al oeste de El Tigre es Ta i d ge 8? d 1504* Ll 3 EN). Aragua: Ranc ho Grande, Parque Nacional Henry Pittier, 480 m, sabanas de loma pedregosa, Mon- taldo & Ramia 3304 (MY). Bolívar: road to Salto Cha- varipo, 88.7 km SW of Caicara del Orinoco, 7?7'N, 66°28'W, Steye nan et al. 131327 (MO). Cojedes: Ga- leras del Pao, ago. 1980, Ramia 7201, (VEN); San Carlos, Sabanas entre San Carlos y Manrique, 29 sep. 1978, Gar- ófalo toa (VEN). Guárico: carretera Calabozo-El Som- brero, a 7 km antes de la última población, 9*03'N, 6073 W. 250 m, Alora et al. 4363* (MO, SI, VEN); 12 km de Calabozo, Davidse 2950 (MO ne N): entre El Som- brero y Pitara Bridge, Pittier 12511 (N US, VEN). Mon zas: sin localidad, Larez 605 (VEN). D Dtto. Ospino, alrededores de carretera y comienzos de primera bajada hacia Wwe p Reyes 1773 Pee PORT); Dtto. Guanare, orilla arretera entre Las as y San Bs ael, Rámirez Reyes 1715 > (I m Trujillo: v vicinity of ividive, Pittier 10830 (GH, NY, VEN); Llanos de Taille Muller 987 (VEN) A Especie afín a P. hirticaule, de la que se diferen- cia por poseer espiguillas globosas de mayor ta- maño, no adpresas y distantes, por tener inflores- cencias difusas que alcanzan ra nte Y del largo total de la planta y lígulas de 2-3 mm de Swallen (195: nonimia de P. hirticaule a P. flabellatum, cuyo tipo coincide con P. hispidifolium. 20. Panicum lepidulum Hitchcock & Chase, Contr. U.S, Natl. Herb. 15: 75, fig. 64. 1910. TIPO: México. Chihuahua: rocky hills near Chihuahua, 22 Sep. 1885, Pringle 497*,** pin US 155163; isótipos, F, MEXU, NY, US-742174, W). Figura 23. 5) considera erróneamente en la si- Plantas perennes, cespitosas, con rizomas de en- ¡om trenudos cortos. Cañas erec tas, pauc inodes, « de (15-)25-70 cm de alto, simples, con innovaciones intravaginales naciendo en los nudos inferiores; en- trenudos de 4-17 cm de largo, cilíndricos, hirsutos a glabros, de 1-2 mm de diámetro; nudos pilosos. Vainas por lo general más cortas que los entrenu- dos, hirsutas con pelos tuberculados, usualmente con tíntes purpúreos, los márgenes pestañosos. Lí- gulas membranáceo-pestañosas, de (1—)2(-3) mm de largo; cuello piloso. Láminas linear-lanceoladas, de 8-26 cm de largo, 0.4—1.2 ascendentes, más o menos adpresas al eje, planas cm de ancho, erectas, o con los bordes involutos, hirsutas, con pelos tu- berculados, ocasionalmente glabras, de base redon- deada y ápice acuminado, los márgenes escabritis- culos, pestafiosos hacia la base, el nervio medio manifiesto. Pedúnculos hasta de 30 cm de largo, glabros a hirsutos. Inflorescencias terminales exer- tas, laxas, difusas, de 7-20 em de largo, 4-10 cm de ancho; eje principal hirsuto hacia la porción ba- sal a glabro, escabroso en el resto de la superficie; pulvínulos glabros; ramificaciones de primer orden ascendentes; ramas de segundo orden divergentes, llevando 1-3 espiguillas solitarias; pedicelos de 2— 12 mm de largo. Espiguillas largamente ovoides, de (3.3—)3.6—4.2 mm de largo, 1.2-1.5 mm de ancho, agudas, no estipitadas, glabras, verdosas o con tin- tes purpüreos; gluma superior y lemma inferior subiguales, 0.9-1.5 mm más largas que el antecio superior. Gluma inferior ovada, de 1.6-1.9 mm de argo, Y o algo menor del largo de la espiguilla, aguda, 5-nervia, el nervio medio escabriúsculo ha- cia el ápice. Gluma superior acuminada, de 3.3-3.6 mm de largo, 9—15-nervia, el nervio medio esca- briúsculo hacia el ápice. Lemma inferior acumina- da, de 3.3-3.9 mm de largo, 9-nervia, el nervio medio escabriúsculo hacia el ápice. Pálea inferior reducida, ovada, de 1.5-2 mm de largo, 0.7—0.9 mm de ancho, % del largo del antecio superior, hialina, glabra. Antecio superior ovoide, de 2.1-2.5 mm de largo, 1-1.5 mm de ancho, crustáceo, gla- bro, liso, lustroso, pajizo, con un anillo circular en la base, castaño a la madurez; lemma 7-nervia; pá- lea con papilas simples y micropelos globosos hacia el ápice. Cariopsis ovoide, de 1.8 mm de largo, 1 mm de ancho; hilo oblongo; embrión % del largo de la cariopsis. Distribución y ecología. Abundante en México y ocasional en Guatemala (Fig. 8), crece en sabanas o en bosques de robles, sobre suelos a menudos volcánicos; se halla también en bordes de caminos. Llega hasta los 2400 m: florece entre marzo y oc- tubre. Volume 83, Number 2 Zuloaga & Morrone 257 1996 Panicum Subg. Panicum Secc. Panicum » 4 B IU 4 y — adipe PAS, m Figura 23. dio um lepidulum.—A. Hábito.—B. "ep vista del lado de la gluma inferior. —C. Espiguilla, vista lateral. —D. Espiguilla vista del lado de la gluma superior.—E. Pálea inferior—F. Antecio superior visto IN lado de la lemma.— io superior visto del lado de la pálea. cim Cariopsis, vista escutelar.—I. Cariopsis, vista hilar. ie (A-G, Davidse & Davidse 9927; H, I, Arsène 2864. 258 Annals of the Missouri Botanical Garden Material representativo citado. GUATEMALA. Guate- se e i hcock 9014 (US). MEXICO. Aguascalien- mi. E of Calvillo, 6900 ft., 25 Sep. 1963, Reeder & rise Fn 5). Chiapas: Arriaga, savanna 2 km SE of Arriaga, Breedlove 36852 (NY). Chihuahua: 12.8 mi. W 16, ca. ~ - ~ ~ hd inity of La Cueva, Johusion 9104* (GH). eral: near Mexico city, Pedregal, Hitchcock 5958 rango: Durango, Palmer 525* , US); a miles NE of Durango on the Durango Torreón isad, 6300 ft., 30 S P 1959, em 812 ( US). obla ES. 13 km f San Miguel, « dional in rocky soil near ds top La Presa, . Sohns 504. (US); Santa r de Ju- ventino MO rumbo a Guanajuato, 2 sep. 1981, Beetle M-7324 (MO). Hidalgo: SR dee in 'anic outcrops at head of uk sent into Barranc titlán b ps Zo- quital and Los Venados, 2000 uus Jr. 4218 (GH US). Jalisco: Paso de la Troje, n 36, acis of Ojuelos on de to Aguascalientes, rocky slopes on ha near Cerr Campana, McVaugh 16819 (NY, p mi. SE of e dae of Hwy. 80 and 45 at La Moreno, 1800 m, 21 Aug. 1975, Dade & Davidse 9927% Be xd 3 km al W de San Cristóbal Ecatepec, base de la Sierra de Lu 2400 m, 18 ago. 1974, Duda 32145 (MEXU). Michoacán: 16 miles N of the junction of highways 37 and 15 Davidse 9912* (MO); vicinity of Morelia, Arsène 241 7* (G H, MEXU, MO, , US). Morelos: Barranca de Atzingo, 1900-2000 m, 7 mar. 1972, Vázquez 3592 (MEXU). Oaxaca: Hacienda Aguilar, 1580 m, 23 jul. 1919, Conzatti 3603 (US); 13.5 mi. SW of Sola de ue move the road to Puerto Escon- , 1920 m, 14 Aug. 1975, Davidse 9657 (MO). Pueb- la: Tehüseán; 5500 E Hitchcock 6063 (MEXU, US). Querétaro: Near San Juan del Río, 18 Aug. 1905, Rose et al, 9555 (MEXU, NY, US). San Luis Potosí: Las Can- oas, 20 July 1910, + 5756 iu 8 km al W de San Sons k retera S.L. Potosí-A. Morelos 1800 m, Fani oe US). Sonora: "Valle de Teras, near n Angostura, White 3542* (GH). de catecas: Chap- ingo, km 846 carretera Juárez, 26 km al N de Fresnillo cerca de Tropico de Cáncer, 2150 m, h sep. 1955, Her- nández & Mathus n-1667 (US). Pa "en de | E x: H near kr e O ~ n Panicum lepidulum es afin a P. pampinosum y P. parcum. Panicum pampinosum se distingue por in- cluir plantas anuales, con inflorescencias con las ramificaciones poco divergentes del eje, y por tener las espiguillas con la gluma inferior Y, del largo de la espiguilla. Panicum parcum se caracteriza por incluir plantas anuales, con espiguillas estipitadas y gluma inferior % del largo de la espiguilla. Hitchcock (1951) cita esta especie para Utah, New Mexico y Arizona y Waller (1976) para Ari- zona, sin citar especímenes de herbario. No fue posible hallar ejemplares de P. lepidulum para los Estados Unidos en los herbarios examinados de ese país, por lo que dichas citas se consideran dudosas. 21. Panicum magnispicula Zuloaga, Morrone & Valls, Iheringia, Bot., 42: 5, figs. 9-17, 26 y 27. 1992. TIPO: Brasil. Santa Catarina: Água Doce, campo graminoso em encosta fngreme com afloramentos, 7.2 km ao sul da divisa Pa- raná-Santa Catarina ao longo da rodovia BR- 153, 1160 m, 3 dic. 1987, Valls, Gomes & Sil- va 11521* (holótipo, CEN; isótipo, SI). Figura 24. Plantas perennes, cespitosas. Cañas floríferas erectas de 20-30 em de alto, simples; entrenudos cilíndricos, huecos, glabros; nudos glabros. Vainas estriadas, más largas que los entrenudos, hirsutas, con pelos tuberculados blanquecinos, los bordes superiores largamente pestanosos. Lígulas mem- branáceo-ciliadas, reducidas, ca. 0.5 mm de largo. Láminas linear-lanceoladas, de 10-12 cm de largo, 0.2 cm de ancho, de base redondeada y ápice lar- gamente atenuado, hirsutas, con largos pelos tuber- culados a híspidas, los bordes ciliados, involutos. Pedúnculos ca. 25 em de largo, cilíndricos, glabros. Inflorescencias terminales exertas; espiciformes a subespiciformes, paucifloras, de 3-6 cm de largo, l em de ancho, con aproximadamente 7-12 espi- guillas por panoja; eje principal y eje de las rami- ficaciones angulosos, escabrosos; pulvínulos pilo- sos; ramificaciones por panoja 34, las inferiores de 3 em de largo; espiguillas solitarias y adpresas sobre los ejes; pedicelos claviformes, escabrosos y con largos pelos cerca de la inserción de la espi- guilla. Espiguillas largamente elipsoides, de 5.5— 5.8 mm de largo, 1.8-2 mm de ancho, estipitadas, glabras, verdosas o con tintes violáceos, con gluma superior y lemma inferior subiguales y superando en largo al antecio superior, de ápice atenuado. Gluma inferior de 3.5 mm de largo, % o más del largo de la espiguilla, aguda, 7—9-nervia, el nervio medio escabriúsculo hacia la porción superior, se- parada de la gluma superior por un entrenudo cons- picuo de 0.5 mm de largo. Gluma superior acumi- nada, 9-nervia, cortamente pilosa hacia el ápice en la cara interna. Lemma inferior acuminada, 9-ner- via, violácea hacia la porción superior. Pálea in- ferior elíptica a lanceolada, reducida, de 2.5-2.8 mm de largo, 1-1.2 mm de ancho, % del largo del antecio superior, hialina, glabra, ciliolada en los márgenes a glabra. Antecio superior largamente elipsoide, de 3.4-3.6 mm de largo, 1.5-1.7 mm de ancho, glabro, crustáceo, pajizo, lustroso, papiloso, con papilas distribuidas regularmente en toda la superficie de la pálea superior y en menor número sobre la lemma, con un anillo castaño circular en la zona de inserción con la raquilla; lemma 7-ner- via; pálea con papilas simples en toda su superficie. Volume 83, Number 2 Zuloaga & Morrone 259 1996 Panicum Subg. Panicum Secc. Panicum Figura 24. Panicum magnispic ‘ula. —A. Habit iftos PE sy vista del lado de la gluma inferior. —C. Espiguilla, vista uel —D. Gluma inferior. —E. Gluma superior.—F. Pálea inferior. —C. Antecio superior vista del lado de la álea.—H. Cariopsis, vista hilar.—I. Cariopsis, vista esum (Valls et al. 11521.) Cariopsis elipsoide de 2.2 mm de largo, 1.3 mm de 22. Panicum miliaceum L., Sp. Pl. 1: 58. 1753. ancho, pajiza; hilo oblongo; embrión poco menos Milium panicum Mill., Gard. Dict. (ed. 8): n. de % del largo de la cariopsis. 1. 1768. Milium esculentum Moench, Metho- dus: 203. 1794. Panicum milium Pers., Syn. Distribución y ecología. Coleccionada unica- Pl. 1: 83. 1805, nom. illeg. superfl. Leptoloma mente en el estado de Paraná, Brasil (Fig. 8), donde miliacea (L.) Smyth, Trans. Kansas Acad. Sci. habita en campos del NW del estado, a 1160 m. 25: 86. 1913. TIPO: India (lectótipo, desig- Panicum magnispicula se asemeja a P. chasei, nado por Sherif & Siddiqi (1988), no visto). distingiéndose esta última por tener inflorescencias multifloras, pedicelos escabriúsculos y espiguillas Plantas anuales, cespitosas, herbáceas, de 15- e 2.8-3.8 mm de largo. 70 cm de alto. Cañas decumbentes, ramificadas o 260 Annals of the Missouri Botanical Garden no en los nudos inferiores, a erectas, ramificadas en los nudos superiores; entrenudos cilíndricos, huecos, pilosos, de 2-12 cm de largo: nudos os- curos, cubiertos de pelos blanquecinos. Vainas de 4—12 cm de largo, mayores que los entrenudos, es- triadas, herbáceas, densamente pilosas, con pelos tuberculados caducos. Ligulas de 1-1.3 mm de lar- go, cortamente membranáceas en la base, luego lar- gamente ciliadas. Láminas oblongo-lanceoladas, de 6-25 cm de largo, 0.4-1.2 cm de ancho, planas, de color verde oscuro, subcordadas en la base, con nervio medio manifiesto y bordes membranáceos. escabrosos, pilosas en ambas caras, los márgenes inferiores ciliados con pelos tuberculados. Pedún- culos cilíndricos, pilosos. Inflorescencias laxas, mul- tifloras, nutantes, de 6-20 cm de largo, 4-11 cm de ancho, incluidas cuando jóvenes en las hojas superiores; eje principal triquetro, escabroso, on- dulado, con largos pelos blanquecinos hacia la par- te basal o sin los mismos; pulvínulos glabros, pajizos; ramificaciones alternas, raro subopuestas, divergentes del eje principal y desnudas en la por- ción basal, las espiguillas solitarias sobre largos pedicelos; pedicelos escabrosos, esparcidamente pilosos, de 4-9 mm de largo. Inflorescencias axi- lares similares a las inflorescencias terminales. Es- piguillas ovoides, de 4.6—5.5 mm de largo, 1.5-2.1 mm de ancho, no estipitadas, glabras, pajizas, con tintes violáceos, gluma superior y lemma inferior subiguales, con nervios marcados y superando en largo al antecio superior. Gluma inferior de 2.8—3.6 mm de largo, % a ?4 del largo de la espiguilla, 5— 7-nervia, el nervio medio escabriüsculo en su por- ción superior, de ápice atenuado. Gluma superior de 4—5.1 mm de largo, 11-13(-15)-nervia, el nervio medio escabriüsculo. Lemma inferior de 4—4.8 mm de largo, 11-13-nervia, glumiforme. Pálea inferior ovada, reducida, de 1.2-1.6 mm de largo, 0.5-0.8 mm de ancho, % o un poco menos del largo del antecio superior, membranácea, con el ápice bilo- bado. Antecio superior elipsoide, de 2.7-3.4 mm de largo, 1.3-1.9 mm de ancho, llegando cuando ma- duro a 2.3 mm de ancho, tempranamente caedizo, liso, glabro, lustroso, blanquecino, amarillento a la madurez, con dos cicatrices basales de ca. 0.4 mm de largo; lemma 7-nervia; pálea con papilas verru- gosas junto al ápice. Cariopsis de 2-2.3 mm de 1.4-1.8 mm de blanquecina; hilo embrión aproximadamente la mitad largo, ancho, punctiforme; del largo de la cariopsis. Distribución y ecología. Originaria del Viejo Mundo, donde es cultivada desde la antigüedad. Introducida en América, donde se cultiva para ali- mento de aves, como pequefio cereal, o para pas- toreo de verano; es apreciada por su precocidad y resistencia a la sequía (Palacios, 1969 Nombre vulgar. “Mijo.” Material representativo citado. ARGENTINA. Bue- nos Aires: Punta Lara, Zuloaga 1890* (SI). Distrito Federal: sin localidad, Calderón 1672 (BAA). Cordoba: :hinga, Gi | Corrientes, Quarín 1321 (CTES). Entre Rios: Concepción del Uruguay, Burkart 24109 (SI). La Pampa: Telen, Mon- ticelli 168 (S I). La Rioja: Chilecito, pee ison er (BAA). Mendoza: Godoy Cruz, Ruíz Leal 5 AA, LIL). Sal- ta: Salta, Chachapoyas, 14 ene. Ls Noo ara Aie (M). Santa Fe: Rosario, Isla del Rowing, 6 en 8, Chris- tian s.n. (SI). Tucumán: Capital, calle De S del Hipód- romo, Tiirpe 3009 (LIL). BRASIL. Distrito Federal: Pla- no Piloto, SQN 76, Filgueiras 1204 (IBGE). Rio Grande do Sul: Sao Leopoldo, Kaine 1937 (BAA). Santa Ca- Klein 17008 (US). Sao Paulo: Pinbivitos. Gehrt 31509 (Us S). COLOMBIA. Valle: Palmira, Estación Agrostológica, Villamizar- Jaramillo s.n. (COL- 44878). ESTADOS UNIDOS E e. ‘A. California: Butte Co., on the S edge of Pentz Road, about 1 mile E of Highway 99, 10 Aug. 1988, je 6168 (MO). Florida: x, 1901, Curtiss 6867 (M Rosario, Reitz Pensacola, 8 Au setts: Dump Valley Road. n ide 10 Sep. 1916, Churchill 7 (MO). Mississippi: ca. 4 miles N of Biloxi 1971, Lasseigne eons caia to Hattiesburg, 24 July 812 (MO). Missouri: St. Louis, Linde nwood freight yard a the Frisco Railroad, 3 July 1955, Muelhenbach 667 (MO). New York: Ithaca, 21 Jul. 1916, Metcalfe 5511 . nius Dakota: 18 miles S and 15 miles W of 7 Aug. 1971, Seiler 3825 (MO). Utah: North Park, qe 8023 (MO). Map Chittenden Co., lagstaf in yard of house at Rt. jo Leisure Lane Aug. 1983, road] (MO). G i AYANA FRANCESA. Station des Not Bassion- de PArataye, Feuillet 4442 (US). HONDUR Morazán: El Za- morano, 800 m, Rodrigue 3512 (US). REPUBLICA DOMINICANA. Santo Domingo: Santo Domingo, Jardín Botánico, Liogier 21371 (NY). Especie relacionada con P. capillare y P. hirti- caule; la primera se distingue por tener inflores- cencias erectas, caedizas, con espiguillas de 2- 2.8(-3.2) mm de largo, y gluma superior y lemma inferior 7-9-nervias. Panicum hirticaule se separa por poseer inflorescencias erectas, con espiguillas cobrizas a violáceas, adpresas, de 1.9-2.8(-3.5) mm de largo, gluma superior y lemma inferior 7— 9(—11)-nervias. 23. pes mohavense Reeder, Phytologia 71: . TIPO: Estados Unidos de América. ue Co., Arizona strip, Main y, ca. 1.5 km S of the junction of the Colorado City road with the Main Street Valley road, low hills with limestone terraces, ca. 1525 m, 11 Oct. 1990, Reeder & Reeder 8630*,** (holótipo, ARIZ no visto; isótipos, K no visto, MO, US-3238288). Figura 25 G-N. Plantas anuales, de 2-6(-8) cm de alto, profu- Volume 83, Number 2 Zuloaga & Morrone 261 1996 Panicum Subg. Panicum Secc. Panicum Figura 25. A-F. Panicum mucronulatum. Br Hábito.—B. né v n vista del lado de la gluma inferior.—C. Espiguilla vista del lado de la glum pa nE Pálea inferior. —E. Antecio superior visto del lado de la lemma.— F. Antecio superior visto del lado de la pálea. G— e Panicum mohavense. TL. Hábito.—H. Espiguilla vista del lado de la gluma inferior.—I. Espiguilla vista del lado de la gluma superior.—J. Pálea inferior. —K. Antecio superior visto del lado de la lemma.—L. Antecio superior visto del lado de la pálea.—M. Cariopsis, vista escutelar.—N. Cariopsis, vista hilar. (A—F, Pinto 121; G-N, Reeder & Reeder 8630.) 262 Annals of the Missouri Botanical Garden samente ramificadas en la base. Cañas paucinodes, erectas; entrenudos 1-2, más cortos que las vainas foliares; nudos híspidos. Vainas de 0.4—1.5 cm de largo, estriadas, verdosas o con tintes violáceos, papiloso-híspidas en toda su superficie o hacia la porción basal, con los márgenes esparcidamente papiloso-pestañosos. Lígulas de 0.2-0.4 mm de lar- go, membranáceo-pestañosas, con largos pelos blanquecinos en la cara abaxial, hasta de 0.8 mm de largo; cuello glabro. Láminas linear-lanceoladas, de 1—4 cm de largo, 1-3 mm de ancho, erguidas, planas, glabras, de base recta, continuándose im- perceptiblemente con la vaina y ápice agudo, los bordes basales papiloso-pestañosos, luego glabros, escabrosos, involutos hacia el ápice. Pedúnculos in- cluidos en las vainas foliares. Inflorescencias subes- piciformes, paucifloras, subincluidas en las vainas oliares, no superando en largo a las hojas, 0.5-2 cm de largo, 0.5 cm de ancho; eje principal angu- loso, escabriúsculo sobre las aristas; pulvínulos gla- bros; ramificaciones flexuosas, escabrosas; pedice- los de 0.5-2.8 mm de largo, escabrosos. Espiguillas globosas, de 2-2.3 mm de largo, 1-1.2 mm de an- cho, glabras, de ápice cortamente acuminado, ver- de-pálidas o con tintes violáceos; gluma superior y lemma inferior subiguales, tan largas como la es- piguilla. Gluma inferior ovada, de 0.8-1 mm de largo, de ápice obtuso a agudo, membranácea, de- licada, abrazadora, 3-nervia. Gluma superior de 2— 2.2 mm de largo, 9-nervia, con los nervios mani- fiestos, el central escabriúsculo hacia el ápice. Lemma inferior glumiforme, 9-nervia. Pálea inferior ausente o reducida, cuando presente hasta de 0.4 mm de largo, hasta % del largo del antecio superior, hialina, glabra. Antecio superior anchamente ovoide, de 1.4-1.8 mm de largo, 1 mm de ancho, alcan- zando igual largo que la gluma superior y lemma inferior, pajizo, castaño claro a la madurez, lustro- so, con un anillo y 2 pequeñas cicatrices en la base, castañas a la madurez; lemma 7-nervia; pálea con papilas simples junto al ápice; lodículas 2, ca. 0.4 mm de largo; estambres 3, las anteras de 0.9 mm de largo. Cariopsis anchamente elipsoide, de 1.3 mm de largo, 1 mm de ancho, olivácea; hilo punc- tiforme; embrión % del largo de la cariopsis. Distribución y ecología. Conocida sólo para el estado de Arizona en los Estados Unidos de América (Fig. 3), donde crece en campos abiertos a 1525 m. Especie relacionada a P. stramineum, distin- guiéndose esta última por incluir plantas mayores, hasta de 125 cm de alto, con láminas de 4.5-30(2 42) cm de largo por 0.3-1.3(22) cm de ancho, es- piguillas de 2.3-3.2 mm de largo y pálea inferior tan larga o más larga que el antecio superior. 24. Panicum mucronulatum Mez, Bot. Jahrb. Syst. 56, Beibl. 125: 2. 1921. TIPO: Brasil. Bahia: sin localidad, Blanchet 104 (lectótipo, aquí designa B, isolectótipo, US-80776; probable cds NY, foto del lectótipo, K). Figura 25A-F. Panicum virgatum L. var. pilosum Doll in Martius, Fl. Bras. 2(2): 217. 1877. SINTIPOS: Brasil. Bahia: llheus, Blanchet 2961 (K, MO, US). Pará: Almeirim, Martius s.n. (M no visto). Plantas anuales, herbáceas, cespitosas. Cafias erectas de 60—100 cm de alto, ramificadas en los nudos inferiores, multinodes; entrenudos comprimi- dos, huecos, con largos pelos blanquecinos adpre- sos o con pelos tuberculados, erguidos; nudos os- curos, comprimidos, pilosos. Vainas comúnmente mayores que los entrenudos, pajizas, cubiertas de pelos tuberculados rígidos, quebradizos y urtican- tes, los bordes ciliados. Lígulas membranáceo-ci- liadas, de 1.4-1.6 mm de largo, arqueadas; cuello densamente piloso, cubierto de pelos tuberculados. Láminas lanceoladas, de 25-45 cm de largo, 1-1.5 cm de ancho, planas, hirsutas, de ápice acuminado a subulado y base subcordada, los bordes ciliados a escabrosos, el nervio medio manifiesto. Pedún- culos parcialmente subincluidos en las vainas folia- res, híspidos hacia la porción distal. Inflorescencias terminales exertas; panojas laxas, difusas, pirami- dales, multifloras, de 25-40 cm de largo, 10-30 cm de ancho; eje principal anguloso, hirsuto en la por- ción basal, luego escabroso, pulvínulos glabros; ramificaciones de primer orden alternas, raro opuestas, divergentes; eje de las ramificaciones hir- sutos a escabrosos, las espiguillas difusas y regu- larmente dispuestas sobre la ramificaciones; pedi- celos triquetros, escabrosos, de 2-20 mm de largo. Espiguillas solitarias, ovoides, de 3.3-3.6(-3.9) mm de largo, 1.3-1.6 mm de ancho, no estipitadas, gla- bras, verdosas o con tintes violáceos, glumas y lem- ma inferior con nervios marcados, subiguales, su- perando 1 mm en largo al antecio superior. Gluma inferior de 1.5-2 mm de largo, % del largo de la espiguilla, aguda, (8—)5—7-nervia, el nervio medio escabriúsculo, los nervios anastomosados hacia el ápice. Gluma superior aguda, de 2.8-3 mm de lar- go, 9-13-nervia. Lemma inferior glumiforme, 3-3.3 mm de largo, 7-9-nervia. Pálea inferior elíptica, de 2.3-2.6 mm de largo, 0.9-1.1 mm de ancho, mem- branosa, blanquecina, glabra, tan larga como el an- tecio superior, los bordes y el ápice esparcidamente pilosos. Antecio superior ovoide, de 2.2-2.4 mm de largo, 1-1.2 mm de ancho, glabro, pajizo, liso, lus- troso, crustáceo, con dos cicatrices basales ca. 0.2 mm de largo; lemma 7-nervia; pálea con micropelos Volume 83, Number 2 1996 Zuloaga & Morrone 263 Panicum Subg. Panicum Secc. Panicum globosos y papilas simples distribuidas irregular- mente en su ápice. Cariopsis anchamente elipsoide, de 1.6 mm de largo, 1.1 mm de ancho; hilo puncti- forme; embrión la mitad del largo de la cariopsis. Distribución y ecología. Brasil, en los estados de Alagoas y Bahia (Fig. 4); se encuentra en bordes de arroyos o cerca de cultivos de caña de azucar, entre los 25 y 200 m. Nombre vulgar. “Capim-de-orvalho.” Material adicional examinado. BRASIL. Alagoas: Porto Calvo, Fazenda Porto Seguro, Campelo 2154 (CEN). Bahia: sin localidad, Gardner 208 (US); Dom Macedo Costa, Fazenda Mocambo, 5 Mar. 1985, Noblick & Lemos 3955 (CEPEC); Mun. E n de Jesus, BR- 101 Cruz das Sagi essa , Coradin et al. 3394 SN); D Pinto 121 poses 306*,** (IAN, US); Cac cia. 25-50 m, coarse tufts, near streamlet, brushy border of sugar cane field, Chase 8105* (NY, RB, US), 8105bis (US); Municipio São Sebas- tiáo do Passé, Area da Estação Experimental Miranda, km 62 da Rod. BR 324, Hage et al. 1716 (CEPEC, K). Pa- raíba: Escola do Agronomia do Nordeste, Moraes 677* (SP, US). Pernambuco: Escola de Agronomia do Nord- este, Coelho de Moraes 677 (IAN, SP); Recife, Sendulsky 1434A (SL, SP); Tapera, Apr. 1928, Pickel 1393* (US); near Vitoria, Recife, Davis & Andre-Lima 61068 (IBGE); sin localidad, Gardner 1838 (K), Glocker 208 5). 2 Panicum mucronulatum es afín a P. hispidifolium y P. stramineum. Panicum hispidifolium se aparta por tener espiguillas globosas abiertas en el ápice, gluma superior y lemma inferior de igual largo que el antecio o hasta 0.3 mm más largo que el mismo, gluma superior 7-nervia y pálea inferior % del largo del antecio superior. Panicum stramineum se dis- tingue por tener inflorescencias con ramificaciones desnudas hacia la base, espiguillas globosas, de 2.3—3.2 mm de largo, gluma superior y lemma in- ferior de igual largo que el antecio superior o hasta 0.4 mm más largas. 25. Panicum pampinosum Hitchcock & Chase, Contr. U.S. Natl. Herb. 15: 66, fig. 48. 1910. Panicum capillare L. var. pampinosum (Hitche. & Chase) Gould, Madrofio 10: 94. 1949. aria dandy bi Presl var. pampi- o Arizona: 2600 ft., 1903, Thornber 193* (holótipo, US-592754; isótipo, MO-2752215). de 11-27 cm de alto. erectas a ramificadas en los nudos basales y me- dios; entrenudos de 2-3 cm de largo, glabros, vio- láceos a pajizos; nudos pilosos, con pelos blanque- Plantas anuales, Cañas cinos, sedosos, hasta de 1 mm de largo. Vainas de 1.5-2.5 cm de largo, usualmente más largas que los entrenudos, con tintes violáceos, papiloso-pilo- sas, con los márgenes pestañosos. Lígulas mem- branáceo-pestañosas, de 1 mm de largo; cuello gla- bro, blanquecino. Láminas linear-lanceoladas, de —5 cm de largo, 0.3 cm de ancho, ascendentes, erectas, poco divergentes, planas, cortamente pi- losas en ambas caras, de base redondeada y ápice agudo, con los márgenes escabriúsculos, ondulados, los basales pestañosos, nervio medio blanquecino. edúnculos hasta de 4 cm de largo a subincluidos, cilíndricos, glabros, purpúreos. inflorescencias lar- gamente exertas, terminales, espiciformes a subes- piciformes, de 5-6 cm de largo, 1-2 cm de ancho, paucifloras; eje principal triquetro, glabro, esca- briúsculo; pulvínulos glabros; ramificaciones de primer orden 3—4(—5), alternas, las inferiores de 4— 8 cm de largo, glabras, desnudas hacia la base; ramificaciones de segundo orden adpresas, con es- piguillas próximas, apretadas sobre las ramas; ped- icelos de 1-5 mm de largo, glabros, escabriúsculos. Inflorescencias axilares similares a la terminal, de menor tamaño, de 2-3 cm de largo. Espiguillas glo- bosas, 3.6—4 mm de largo, 1.4 mm de ancho, túr- gidas, no estipitadas, de ápice agudo, glabras, pa- jizas o con tintes violáceos; gluma superior y lemma inferior subiguales, superando 1.6 mm en largo al antecio superior. Gluma inferior ovado-acuminada, e 2.4-3.1 mm de largo, Y, del largo de la espi- guilla, 3—5-nervia. Gluma superior de 3.6—4 mm de largo, 7-nervia, con los nervios verdosos. Lemma inferior glumiforme, de 3.2-3.5 mm de largo, 9- nervia. Pálea inferior reducida, de 0.7 mm de largo, 0.3 mm de ancho, % del largo del antecio superior, hialina, glabra. Antecio superior ovoide, de 2-2.4 mm de largo, 1.2 mm de ancho, túrgido, globoso, crustáceo, liso, lustroso, pajizo, con 2 cicatrices ba- sales, ca. 0.3 mm de largo; lemma 7-nervia; pálea con papilas simples junto al ápice. Cartopsis no vis- a. et Distribución y ecología. Sur de los Estados Unidos de América y México (Fig. 10); crece en campos en áreas a menudo perturbadas entre 780 y 2250 m; florece entre agosto y octubre. Material adicional examinado. ESTADOS UNIDOS DE AMERICA. Arizona: Greenlee County, about 40 miles N of Clifton along Coronado Trail, 7500 ft., 25 Aug. 1971, Reeder & Reeder 5520 (US); Tucson 2 25 Aug. 1904, Griffiths 6939 1/2 pr. p. (US w Mexico: Grant County, 1880-1881, Rusby 444 fie sin dal. 1851-1852, Wright 2084 (US); Organ Mountains, 29 Aug. 1894, Wooton 2014 (US). MEXICO. Chihuahua: Miñaca, 7000 ft., 13 Oct. 1910, Hitchcock 7751 (US). Esta especie se caracteriza por incluir plantas de porte pequeño, con panojas espiciformes a subes- 264 Annals of the Missouri Botanical Garden piciformes y espiguillas globosas de 3.6—4 mm de largo, con gluma inferior Y, del largo de la espi- illa. Swallen (1955) cita esta especie para Guatemala en los departamentos de Chiquimula, Jutiapa y Za- capa, sin indicar los ejemplares de herbario corres- pondientes. No se han hallado espécimenes de di- cho país por lo que esta cita se considera dudosa. 26. Panicum pareum Hitchcock & Chase, Contr. U.S. Natl. Herb. 15: 68, fig. 53. 1910. Panicum decolorans Kunth var. parcum (Hitche. & Chase) Beetle, Phytologia 54: 4. 1983. TIPO: México. Sinaloa: Lodiego on the Culiacán River, gia side, not very com- mon, 9-15 Oct , Palmer 1657** (ho- lótipo, Du. Sca F, ISC, NY, P). Plantas anuales, cespitosas. Cañas erectas a ge- niculadas, de 20-90 cm de alto, simples, pauci- nodes, innovaciones extravaginales; entrenudos de 5-20 cm de largo, 2 mm diám., cilíndricos, glabros a hirsutos; nudos glabros, castaños. Vainas de 10— 12 cm de largo, comúnmente más cortas que los entrenudos, hirsutas con pelos tuberculados, los márgenes pestañosos. Lígulas cortamente mem- branáceas en la base y luego largamente ciliadas, de 1-1.2 mm de largo; cuello hirsuto, castaño. Láminas linear-lanceoladas, de 6—45 em de largo, 0 cm de ancho, planas, de base angostada, agudas, hirsutas, con pelos caducos, de base tub- erculada, de aproximadamente 1 mm de largo, los bordes escabrosos, pestañosos hacia la base, el ner- vio medio manifiesto. e subinc luidos en las vainas foliares a exertos, de 1 cm de largo, hirsutos en la porción distal. oso encias termin- ales exertas, de 7-30 em de largo, 6-15 cm de ancho, laxas, difusas, erectas; eje principal angu- loso, escabriúsculo, glabro a hirsuto hacia la base; pulvínulos violáceos, glabros; ramificaciones de primer orden alternas, divergentes, las inferiores de 6-15 cm de largo; pedicelos triquetros, de 5-18 mm de largo, escabriúsculos, glabros. Inflorescen- cias axilares presentes, similares a las inflorescen- cias terminales. Espiguillas solitarias, ovoides, de 4.56 mm de largo, 1.8-2 mm de ancho, glabras, verdosas y con tintes violáceos, estipitadas; gluma superior y lemma inferior subiguales, 1.2—1.8 mm más largas que el antecio superior. Gluma inferior de 3—4.2 mm de largo, % del largo de la espiguilla, aguda, 5—7-nervia, el nervio medio escabroso, se- parada por un entrenudo marcado de la gluma su- perior, ca. 0.6 mm de largo. Gluma superior de 3.9— 5.4 mm de largo, 9-1l-nervia, el nervio medio escabroso, cortamente escabrosa en la cara interna. Lemma inferior aguda, de 3.6—4.8 mm de largo, 9— l 1-nervia, escabrosa en la cara interna. Pálea in- ferior lanceolada, reducida, de 1.2-1.8 mm de lar- go, 0.4—0.6 mm de ancho, % o algo menor del largo del antecio superior, membranácea, glabra. Antecio superior ovoide, de 2.7—3.3 mm de largo, 1.5-2 mm de ancho, crustáceo, glabro, liso, lustroso, pajizo, con tintes negruzcos a la madurez, de base circular, castafia a la madurez; lemma 7-nervia; pálea con papilas verrugosas en el ápice. Cariopsis ovoide, de 2.3-2.5 mm de largo, 1.4—1.8 mm de ancho, blan- quecina; hilo punctiforme; embrión % del largo de la cariopsis Distribución y ecología. Se encuentra desde México hasta Costa Rica (Fig. 7), en terrenos ro- cosos abiertos o en bosques de pinos, entre el nivel del mar y 2070 m de elevación; florece entre julio y octubre. is Ler ial r OSTA RICA. Guana- al end of Playas e n. o, 14 Nov. 1968, Pohl e Dude 11437 (F, US). GUATEMALA. Chiquimula: grassy plains and low slopes around Chiquimula, 400 m, 20 Oct. 1939, Steyermark 30064. (F, US). Jutiapa: region of El Ta blón, northeast o r. 850-900 m, 31 Oct. 1940, Standley 75925 (F, US); vicinity of M 850 m, 24 Oct. 1940, Standley 75102a (F. US). Zacapa: near La Fragua, 200—500 m, 14 Oct. 1940, Standley 74761 (F, US); rocky hills in vicinity of Santa Rosalía, 2 mi. south of Zacapa, 200 m, 7 Oct. 1939, Steyermark 29310 (F). HONDURAS. El Paraíso: Road to Yuscarán, edge of brushy places, 800 m, 5 Nov. 1951, Swallen 11371 (MO, US): road to Danlí, not far from Río Choluteca, Swallen ll 250 (US). Francisco Morazán: ragion of Las Mesas, e hills around the Zamorano valley, rocky pine woods, 800-900 m, 2 Nov. 1951, Swallen 10765 (MO, US); San Antonio del Oriente, 1200 m, Swallen 10909 (US). ICO. Chiapas: 14 km E of Ocozocuatla, in open grassy field. Gould 12729* (MO, US); tiérrez, along road to El eye 2500 ft., 27 Oct. 1965, Miis A P ail) pes 33 (F. MEXU, US). Chihuahua: Río Ne 337, Le Siew 173 oo Colima: Alzada, Poen Aa (RB, US). Duran . SE of oo about 600 ft., e Oct. 1966, Could 12288 (US). Guanajuato: km 20 N de Silao, 16 sep. 1946, Her- nández Xolos otzi et al. X-2468 (US). Guerrero: Balsas, Hitchcock 6782 (F, MO, NY, US); Coyuca, Santa Barbara, 9 Sep. 1934, Hinton 6686 (GH, MO, NY, US). Jalisco: about 13 miles northwest of León, Guanajuato in a grassy a with scattered thorny shrubs, 2070 m, Reeder & Reeder 2288 (US). Michoacán: West-facing slopes of Cer- ro de Carboneras, ca. 22 km S of Uruapan, 3300-3700 ft., King & Soderstrom 4831 (MEXU, NY, US). Morelos: Xochiltepec, 24 Sep. a Lyonnet 2639 (MEXU, MO, US). Nayarit: Mun. Na 3.9 km al NE de Jesús María, camino a Huejuquilla, D> 16'N, 104*30'W, 530 m, selva baja iiy "ifolia, 15 sep. 1989, Flores et al. 1053 (MO). : 11 mi. W of Zanatepec, 150 m, 28 Aug. 1953, Fia e Reeder 2152 (US); near km post ae on Hwy. 190, ca. 40 miles NW of Tehuantepec, 710 m, 13 Aug. 1975, Davidse 9587 (MO). Puebla: Mesa del Boe hote 3 km al E de El Salado, bosque de Quercus, 1310 m, 20 oct. 1985, Guízar 1850 (ANSM, CHAPA). Sinaloa: Las c N e = Volume 83, Number 2 Zuloaga & Morrone 265 Panicum Subg. Panicum Secc. Panicum Mesas, Sierra Surotato, ca. 3000 ft., 15 Sep. 1941, Gentry 6667 (GH, MO, NY). Sin Estado: Alzada, 4 Oct. 1910, Orcutt 4687 (US). NICARAGUA. León: along Hwy. 12, km SE of Jct. with Puerto Somoza road, 40 m, 8 1971, ca. 12 km E of Puerto Somoza, Pohl 12707 (F, MO, dd El Velero, at mouth of Estero San José, ca. 10 km yr of Hwy. 32, 12%08'N, 86°45'W, 9 Sep. 1984, m 23096 (MO, SI). Managua: Inst. Pedagógico de Varones, Vernier 1505 (CH). Beetle (1977) considera a P. parcum como si- nónimo de P. decolorans. Esta especie se distingue de P. decolorans por tener panojas laxas con las ramificaciones divergentes del raquis y espiguillas estipitadas, con gluma inferior y superior separadas por un entrenudo marcado. 27. Panicum peladoense Henrard, Blumea 4: 504. 1941. TIPO: Paraguay. Paraguarí: Cerr Pelado, 3 abr. 1883, Balansa 4357** "os po. L no visto; isótipos, O, P, US- 1108609, 1647868, 1649644). Figura 26. Panicum campestre Nees, Fl. Bras. Enum. Pl. 2: 197. = non Nees ex Trin., 1826. Panicum cayennense ar. campestris (Nees) Pilg., Bot. Jahrb. Syst. n 132. 1901. TIPO: Brasil. São Paulo: entre Tau- baté y Pindamonhangaba, Martius s.n. (holótipo, M, fragmento y foto, US-80546, fragmento, US-974750). Panicum bergii Arec piste var. leiophyllum Hackel & Lindman, Kongl. Svenska Vetenskapsakad. Handl. 34 (6): 10, pl. 4b, 1900. TIPO: Brasil. Rio Grande do Sul: Cachoeira, 11 Feb. 1893, Lindman 1185 (is- ótipos, P, US-702326, W). Plantas perennes, cespitosas, de 20—60 cm de alto. Cafías erguidas, simples, paucinodes; entrenu- dos de 3.5-10.5 cm de largo, cilíndricos, huecos, densamente híspidos a glabrescentes; nudos violá- ceos y con pelos blanquecinos densa a esparcida- mente dispuestos. Vainas de 3.5-8 cm de largo, usualmente menores que los entrenudos, pajizas, densamente pilosas, con pelos adpresos de base tuberculada a glabrescentes, los bordes ciliados. Lígulas cortamente membranáceas y luego larga- mente ciliadas, de 0.5—1.7 mm de largo; cuello cas- taño claro, piloso. Láminas lanceoladas a linear- lanceoladas, de 7-22 cm de largo, 0.2-0.4 cm de ancho, erectas, ascendentes, planas, de base an- gostada continuándose imperceptiblemente con la vaina, densamente pilosas, con pelos tuberculados en ambas caras o sólo en la cara abaxial a glabras, los márgenes escabriúsculos y largamente ciliados, cartilaginosos. Pedúnculos subincluidos en las vainas foliares a exertos, hasta de 30 cm de largo, cilíndricos, glabros. Inflorescencias laxas, difusas, piramidales, de 7-18 em de largo, 3-13 cm de an- cho; eje principal cilíndrico, escabroso, con largos pelos esparsos o sin los mismos; pulvínulos glabros; ramificaciones alternas u opuestas, en ocasiones subverticiladas hacia la base, ramas escabrosas a pilosas; pedicelos glabros, claviformes, escabrosos, de 2-20 mm de largo. Espiguillas largamente ovoi- des, acuminadas, de (2.5)2.7-3.3(3.7) mm de largo, 1-1.2 mm de ancho, solitarias, no estipita- das, glabras, pajizas, con tintes violáceos; gluma superior y lemma inferior subiguales y superando 1-1.2 mm en largo al antecio superior. Gluma in- ferior ovada, aguda, de (1.5-)1.7-2.1(-2.4) mm de largo, % a % del largo de la espiguilla, 5—7-nervia, el nervio medio escabriúsculo hacia la porción su- perior. Gluma superior de (2.3-)2.6-3.1(-3.4) mm de largo, 7—9-nervia, caduca a la madurez de la espiguilla y dejando al descubierto el dorso de la emma superior, la cara interna escabriúscula hacia el ápice; gluma inferior y superior separadas por un entrenudo ca. 0.3 mm de largo. Lemma inferior glumiforme, de 2.5-3(-3.4) mm de largo, 7—9-ner- via. Pálea inferior elíptica, de 1.7-2.2 mm de largo, 0. mm de ancho, tan larga como el antecio superior, membranácea, hialina, con los bordes fi- namente pilosos a glabros, de ápice bilobado a en- tero. Antecio superior elipsoide, de (1.7—)1.9-2.1 (2.3) mm de largo, 0.8-1.1 mm de ancho, crus- táceo, glabro, liso, lustroso, pajizo, negro a la ma- durez, con dos cicatrices basales ca 0.1 mm de largo; lemma 7-nervia; pálea con papilas simples hacia el ápice. Cariopsis de 1.2-1.3 mm de largo, 0.6-0.9 mm de ancho; hilo punctiforme; embrión 5 del largo de la cariopsis. Distribución y ecología. Habita en campos de Argentina, Bolivia, Brasil, Paraguay y Uruguay (Fig. 4); llega hasta los 1200 m. Florece de octubre a marzo Material representativo citado. ARGENTINA. Cór- doba: 7 km W de Sta. Rosa de Calamuc bo p 2909 (CTES, SI, US). Sin localidad, 9 ene 1919, Castellanos Misiones: Dpt 8978 (BAA). BOLIVIA. Santa Cruz: Flo of Samaipata, Renvoize et al. 4044 (K); ESTE E yis .B S of Santa Cruz, Renvoize et al. 3933 ( Dis- trito p^ al: Reserva ee an jy us 1060* (SI, SP); Area da FAL, 5 m 1991, Azevedo Filgueiras 950 (MO). r hie ra Viannápolis and a Funda, Mint Pa sandy cerrado, Chase 11280 (US). ato Grosso: ita do Araguaya, on Rio Araguaya, Chase 11858; (US). "Malo Grosso o Esper- anga on Río Paraguay, Chase 11071* (US): Campo Gran- de, 7 Feb. 1930, Chase 10780* (US), between Campo Grande and Dourados, 400—500 m, 14—17 Feb. 1930, Chase 10886 (F, US). Minas Gerais: Serra do Cipó, 110 km NE of Belo Horizonte, W of Vaccaria, 800 m, 28 Mar. 1925, Chase 9279 (F, GH, MO, NY, US); Serra do Man- 266 Annals of the Missouri Botanical Garden Figura 26. Panicum peladoense.—A. Hábito.—B. Espiguilla, vista lateral.—C. Espiguilla vista del lado de la gluma inferior. —D. Espiguilla vista del lodo de la gluma superior.—E. Pálea inferior. —F. Antecio superior visto del lado de : lemma.—G. Antecio superior visto del lado de la pálea.—H. Antecio superior maduro.—l. Cariopsis, vista escute- ar.—J. Cariopsis, vista hilar. (Zuloaga & Deginani 137.) Volume 83, Number 2 Zuloaga & Morrone 267 Panicum Subg. Panicum Secc. Panicum tiqueira, between Sitio and Dr. Sá Fortes, 1200 m, 1 Mar. 1925, Chase 8696 (F, GH, K, MO, NY, US). Pará: Lagoa Aura Preta, Drouet 1953 (F). Paraná: Mun. Arapoti, o by pey boundary, Smith et al. o Grande do Sul: Alto do rr a, legre, alls & Arzivenco 2033 (CEN); Pelotas, 22 Apr. 1946, Swallen 9156* (US). Santa Catarina: 17 km NE of the S. Catarina-RGS border, BR- 116, Davidse et al. 13145 (MO, NY, SP). Sao Paulo: Fazenda Campininha, 10 km NEE of Padua ao = oe S, 47°10'W, 15 July 1960, Eiten et al. 2116 (F, G, Y, US). Sin Estado: sin localidad, jer 2155 (K), Bur- chell 6794-2 (K), Sellow s.n. (K). PARAGUAY. Amam- bay: in altiplanitie Sierra de (anie Hassler 12099 (G). Concepción: San Salvador, Rojas 2784 (BAA, US). praderas de Itaigti, près de Villa Rica, 17 dic. US). Par o Peró, pres de Paraguarí, di oct. 1876, Balansa 14 (G, K, E UGUAY. imo a ruta 29, Del Puerto 1527 (P). Tacuarembó: ac- u ruta c 1981, Cabrera & Zuloaga 32385 (SI), 32388* (MO, SI, US). Henrard (1941) menciona en la descripción ori- ginal de la especie que la misma posee vainas, nu- dos y láminas glabrescentes o con pilosidad rala. Se ha comprobado que el carácter de la pilosidad es variable, encontrándose desde ejemplares den- samente pilosos hasta glabrescentes. Esta especie fue descrita por Nees en 1829 como P. campestre, estando este nombre invalidado por existir un homónimo anterior descrito por Trinius en 1826, perteneciendo esta última especie a la sección Rudgeana (Zuloaga, 1987b). Henrard (1941) erróneamente señala que P. campestre, des- crito por Nees en 1829, corresponde a P. ghies- breghtii. Una especie afín a P. peladoense es P. exiguum Mez, distinguiéndose esta última por ser una planta anual, con cafias abundantemente ramificadas y de- cumbentes en la base, con gluma inferior ?4 o más del largo de la espiguilla. Panicum peladoense se separa de P. chasei por incluir el áltimo taxon plantas con panojas espici- formes a subespiciformes, con espiguillas ancha- mente ovoides y antecio superior castafio a grisáceo a la madurez, sin la gluma superior caediza. 28. Panicum philadelphicum Bernhardi ex Tri- nius, Gram. Panic.: 216. 1826. Panicum phi- ladelphicum Nees, Fl. Bras. Enum. Pl. 2: 198 1829, hom. illeg. TIPO: *misit. s.n., Pan. ladelphici sibi cl. Bernhardi" (holótipo, LE no visto, fragmento y foto de tipo US-80910 Panicum capillare L. var. sylvaticum Torrey, Fl. N. Middle United States: 149. 1824, non P. sylvaticum Lam., 1798. Panicum torreyi E. Fourn. in Hemsl., Biol. Centr.-Amer., Bot. 3: 497. 1885. TIPO: dadas Uni- dos de América. New York: “in dry woods near New York" em UMO no visto, fragmento y foto US- 8055 Panicum capillare L. var. Xe uni Gattinger, Tennessee FL: 1887, non Panicum campestre Nees, 1826. B oo L. var. je Nash, II. Fl. .U.S. 1: 123. 1896, hom. illeg. superfl. Panicum gattingeri Nash in Small, Fl. S.E.U.S.: 92, 1327. 1903. TIPO: Estados Unidos de América. Tennessee: near Nashville, Sep., Gattinger s.n. (lectótipo, desig- nado por Hitchcock & Chase (1910), TENN no visto, fragmento y foto US-80667) v capillare L. var. minimum Engelmann ex Gattin- r, Tennessee Fl. 94. 1887. Panicum minimum (En- aim ex Gatt.) Scribn. € Merr, U.S.D.A. Div. Agrostol. Cire. 27: 4. 1900. TIPO. “Estados Unidos de América. Tennessee: “Green Brier, Sept. 1878, Gattinge. " (lectótipo, designado por Hitchcock & Chase a 1910), TENN no visto, fragmento US- sist ail capillare L. var. geniculatum Scribner. in Kear- y, Bull. Torrey Bot. Club 20: 477 y 479. 1893, non Pan geniculatum Lam. 1798. TIPO: Estados Unidos de América. Kentucky: Bell County, collect- ed near Wasioto, Sep. 1894, Kearney Jr. 378 (lec- tótipo pee por Hitchcock & Chase, 1910, US- 823622 Panicum — sani Fernald, Rhodora 21: 112. 1919. emagog, Sep. 1859, Tuckerman s.n. (holótipo, GH, fragmento US-81287; isótipo, MO). Panicum lithophilum Swallen, Proc. Biol. Soc. Wash. 54: 43. 1941. TIPO: Estados Unidos de América. Geor- gia: Stone Mountain, rocky slope, 23 Aug. 1905 Hitchcock s.n., Amer. Gr. Nat. Herb. no 24 (holótipo, US-952898; isótipos, US-1912007, 2488400). Plantas anuales, cespitosas, de 10-60 cm de alto. Cañas erectas, delicadas, ramificadas hacia la porción basal y media, menos frecuentemente de- cumbentes, arraigadas en los nudos basales; en- trenudos de 2-5 cm de largo, 0.1 cm diám., hir- sutos en toda su extensión o sólo hacia la porción superior; nudos pilosos. Vainas usualmente más lar- gas que los entrenudos, de 2.5-7 cm de largo, ver- des a púrpuras, híspidas, con pelos tuberculados, blanquecinos, hasta de 5 mm de largo, un borde pestañoso, el restante glabro. Lígulas membraná- ceo-pestañosas, de 0.8—1.3 mm de largo; cuello pi- loso. Láminas linear-lanceoladas, de 4-12 cm de largo, 0.2-0.6 cm de ancho, ascendentes, erectas, poco divergentes, planas, hirsutas a esparcida- mente pilosas, verdes o con tintes purpúreos, de base redondeada a subcordada y ápice agudo, los bordes escabriúsculos, los inferiores largamente pestañosos, con pelos tuberculados, blanquecinos, de 0.6-3 mm de largo, el nervio medio hirsuto. Pedúnculos largamente exertos, hasta de 18 cm de largo, glabros a hirsutos junto al ápice, cilíndricos. Inflorescencias terminales laxas, difusas, de 7-2 cm de largo, 4-24 cm de ancho; eje principal glabro Annals of the Missouri Botanical Garden a esparcidamente piloso hacia la porción basal, fi- namente escabriúsculo hacia la porción distal; pul- vínulos glabros a cortamente pilosos; ramificaciones de primer orden alternas a subopuestas, divergen- tes, las inferiores de 7-15 cm de largo; ramifica- ciones de segundo orden divergentes: ramifica- ciones capiliformes, desnudas hacia la base, llevando 1-4 espiguillas; pedicelos cortos, de 0.3— 1.5 em de largo, adpresos, capiliformes, glabros, escabriúsculos. Inflorescencias axilares presentes. Espiguillas elipsoides, de 1.6-2.1 mm de largo, 0.5-0.7 mm de ancho, de ápice acuminado, no es- tipitadas, verdosas o con tintes violáceos; gluma su- perior y lemma inferior subiguales, tan largas como el antecio superior. Gluma inferior ovada, de 0.5— 0.9 mm de largo, abrazadora, % o un poco menor del largo de la espiguilla, de ápice acuminado a redondeado, 3—5-nervia. Gluma superior de 1.6—2 mm de largo, 7-nervia. Lemma inferior de 1.6-1.9 mm de largo, 7-9-nervia. Pálea inferior ausente. Antecio superior angostamente elipsoide, de 1.5—1.7 mm de largo, 0.4 mm de ancho, crustáceo, pajizo, lustroso, con tintes negruzcos a la madurez, con un anillo circular en la base; lemma 7-nervia; pálea con papilas simples junto al ápice. Cariopsis elip- soide, de 1 mm de largo, 0.4 mm de ancho, pajiza; hilo punctiforme; embrión % del largo de la cariop- sis Distribución y ecología. | Habita en los Estados Unidos de América (Fig. 3), sobre suelos graníticos en bosques abiertos o campos; florece entre abril y septiembre. Material y cie citado. ESTADOS UNIDOS DE AMERIC A- Alaba ama: iu A 8 Sep. 1897, Eg- gert 58 (US). Arkansas: Saline Cx , Quac 'hita Region, 9 Aug. 1971, Done 164207 (MC » lin, 12 Sep. 1 Woodward s.n. (US- of V NA Da 12 Aug ta Chase s.n. (MO). Putnam Co., 13 mi. Prince . Hyacinth 960 achusets: ‘Randolph, "Purkapa Pond, 1898, Churchill 103 (MO). Missouri: 5 mi. upland w den bering Mill Cree "n 23 Aug. 1937 Stey- ermark 25209 (MO); Maries Co.. upper slopes along Gas- conade Rives 2% mi. NE of Vienne, 29 Aug. 1937, Stey- 5 Sep. E of Lebanon ermark 2554 (MO w Hampshire: Sullivan Co., Unity, 20 Aug. e Seymour 20900 (MO). North ete olk Co., 1 n X Colombus, Freeman 53433** (US) Oklahoma he 21 Sep. 1894, Bush 722 de, Mus- konge; 18 Men Bebb 5796 (US). Pennsylvania Lancaster Co. wamp two miles S of Refton in Eozoic, 23 Sep. 1901. Heller s.n. (MO-2975482). S Caroli- na: Orangeburg, 16 Aug. 1905, Hitchcock 7 (US). Ten- ssee: 3 mi. of ern Creek Station, 26 Aug. 1897, Kear- Dp eed Co.. Brown & Clebsch xr ** (US). Texas: Dallas Co., Rer- erchon 1842 (MO). Vermont: Caledonia Co., Kirby Mt., Kirby, 14 Aug. 1964, eens 21936 (MO). Virginia: Portsmouth, 26 Aug. t 2, Noyes 459 (MO). West Vir- gi 1903, Steele & Steele 16 nia: sin localidad, ga MO). Wisconsin: D County, common in the Fire ane, 30 Aug. 1955, Seymour & Jones 16381 (MO). Swallen diferencia a Panicum lithophilum de P. philadelphicum por poseer inflorescencias rígidas, con espiguillas de mayor tamafio, agrupadas en pa- res sobre el extremo de las ramificaciones y por tener hojas angostas, erectas y de color purpüreo. El examen de material de P. philadelphicum per- mitió comprobar que estos caracteres no son váli- dos para mantener a P. lithophilum como una en- tidad separada de P. philadelphicum. Panicum philadelphicum está relacionado a P. capillare y P. exile, separándose por poseer espi- guillas ac uminadas de 1.6-2.1 mm de largo, alcan- zando el antecio superior igual largo que la gluma superior y la lemma inferior, con inflorescencias no caedizas y de forma piramidal (con panojas caedi- zas en P. capillare y angostamente elipsoides en P. flexile). Ocasionalmente se observan en esta especie ejemplares con cañas decumbentes y arraigadas en los nudos inferiores, brothers (1953). carácter señalado por Fair- 29. Panicum quadriglume (Döll) Hitchcock, U.S. Natl. Herb. 24: 460. 1927. cum cayennense Lam. var. quadriglume Döll, in Martius, Fl. Bras. 2 (2): 220 7. TIPO: Brasil. Minas Gerais: Caldas, 18 Mar. 1887, Regnell III-1406 (holótipo, W; isótipo, P, US- 724411, fragmento, US-80560). Figura 27. Contr. Pani- Plantas perennes, cespitosas, de 15-80 cm de alto. Cañas erectas, paucinodes, ramificadas hacia los nudos medios; entrenudos de 4—10 cm de largo, cilíndricos, cubiertos de pelos de base tuberculada a glabrescentes; nudos pajizos, cubiertos de largos pelos blanquecinos. Vainas menores que los en- trenudos, de 3—9 em de largo, densa a esparcida- mente hirsutas, con pelos tuberculados, los bordes pestañosos. Ligulas de 0.5-1 mm de largo, con la luego ciliadas; cuello pajizo, glabro a piloso. Láminas linear-lan- base cortamente membranácea, ceoladas, de 6-26 cm de largo, 0.3-0.6 cm de an- cho, erectas, ascendentes, rígidas, planas, de base angostada y ápice acuminado, densa a esparcida- mente hirsutas en ambas caras con pelos tubercu- lados caducos, los bordes escabrosos y ciliados, con pelos tuberculados caducos. Pedúnculos subinclui- dos en las vainas foliares a exertos, de 8-20 em de largo, cilíndricos, glabros. Inflorescencias termina- les exertas, laxas, difusas, piramidales, de 7-20 cm Volume 83, Number 2 Zuloaga & Morrone 269 1996 Panicum Subg. Panicum Secc. Panicum Figura 27. Panicum quadriglume.—A. Hábito.—B. Espiguilla, vista lateral—C. Espiguilla vista del lado de la gluma inferior. Bis Espiguilla vista del lado de la gluma superior.—E. Pálea inferior—F. Antecio superior visto del lado de la lemma.—G. for io superior visto del lado de la pálea.—H. Cariopsis, vista escutelar.—1. Cariopsis, vista hilar. (Steinbach 6979.) 270 Annals of the Missouri Botanical Garden de largo, 3-13 cm de ancho, cuando jovenes in- cluidas en las vainas superiores; eje principal gla- bro, triangular y escabriúsculo; pulvínulos glabros; ramificaciones de primer orden alternas, raro subopuestas o verticiladas hacia la parte superior, divergentes del raquis, escabrosas; pedicelos es- cabriúsculos, verdosos a violáceos, de 3-13 mm de largo. Inflorescencias axilares presentes, similares a las inflorescencias terminales. Espiguillas solita- rias, largamente ovoides, acuminadas, de 2.5—4.2 mm de largo, 1-1.4 mm de ancho, no estipitadas, glabras, verdosas y con tintes violáceos, trifloras, con dos lemmas inferiores, una con y la restante sin su correspondiente pálea, y antecio superior con el dorso de la lemma dispuesto hacia la gluma inferior. Gluma inferior de 1.2-2.3 mm de largo, Y del largo de la espiguilla, de ápice agudo a obtuso, 5—9-nervia, el nervio medio escabriúsculo hacia la porción superior. Gluma superior de 1.9-3.6 mm de largo, de ápice agudo a obtuso, caediza a la madu- rez, 7—9-nervia, el nervio medio escabriúsculo; glu- ma inferior y superior separadas por un corto en- trenudo ca. 0.2 mm de largo. Lemma inferior I de 2.2-3.7 mm de largo, acuminada, 5—9-nervia. Pá- lea inferior I ausente. Lemma inferior II de 2.2-3.8 mm de largo, acuminada, 7—9-nervia. Pálea inferior II elíptica, de 1.5-2.3 mm de largo, 0.7-0.9 mm de ancho, tan larga como el antecio superior, mem- branácea, hialina, de ápice bilobado, brevemente ciliada o no hacia los bordes superiores, encerran- do una flor estaminada. Antecio superior angosta- mente ovoide, de 1.6-2.3 mm de largo, 0.8-1.2 mm de ancho, 0.3-0.9 mm más corto que el largo de la espiguilla, crustáceo, glabro, liso, lustroso, pajizo, negro a la madurez, con dos cicatrices basales ca. 0.1 mm de largo, castañas a la madurez; lemma 7- nervia; pálea con papilas simples junto al ápice. Cariopsis anchamente elipsoide, de 1.6 mm de lar- go, 1-1.1 mm de ancho, pajiza; hilo punctiforme; embrión la mitad o más del largo de la cariopsis. Distribución y ecología. Argentina, Bolivia, Brasil y Paraguay (Fig. 8). Habita en campos. Llega hasta los 2600 m; florece y frucitifica entre enero y mayo. Material representativo citado. ARGENTINA. Mi- siones: Dpto. Montecarlo, Colonia Montecarlo, Schwindt 1154 (LIL, MO, US). BOLIVIA. Beni: Near R >, K, M, US); ia Pu between the Río Khara Vin ca. 3-4 km Licoma Pampa, 16%48'S, 67?14'W, 2600 m, 12 Mar. 1989, Lewis 35344 (MO, SI). Santa Cruz: Nuflo de Chávez, Comunidad Puesto ae 60 km of Concepción 16?25'S, 62°00’ W, 450—750 m, 21 Feb. 1986, Killeen 1828 (F, SI); Buena Vista, Steinbach 6979 (BAA, F, G, GH, K, MO, US, W). BRASIL. Goias: Goyandira, 800 m, campo, Chase 11560* (RB, US); M Presidente Kennedy, 12 km west of village of Presidente Kennedy, 3°25'S, 48°37'W, 1 Feb. 1980, Plowman et al. 8226 (MO, NY). Maranhao: Grajahú to Porto Franco, Swallen 3844 (IAN, K, RB, US). Mato Grosso: between Rondonópolis and São Lourenço, 16°20'S, 54°30'W, Chase 11924 (US). Mato Grosso do Sul: vicinity of Dourados, 22°S, 54°30’ W, 400 m, 18-21 Feb. 1930, Chase 10976* (F, GH, MO, NY, US); € A 10866* (R, US). Minas Gerais: Serro do Curral, SE of Belo Horizonte, trail to Pico, 1000 m, 26 Mar. 1925, Chase 9089 (F, GH, MO, NY, US); Antonio Justiniano, 17 km S of Oliveira, 900 m, 16 Mar. 1925, Chase 8897*,** (F, GH, MO, NY, RB, US, W). PARAGUAY. Amambay: Pedro Juan Caballero, Rojas 12741 (LIL, MO); in regione calcarea cursus superioris fluminis Ap. Hassler 10967 (BAF, G, K, NY, P, US). Concepción: entre el Río Apa y el Río Aquidabán, Fiebrig 5102 (F, GH, pis Cordillera: Fortín Lopez, Hassler 1942 (BAA, G, K, P); ics del lago Ipacaray, Hassler 12494 (BAA, E E H, K, NY, US). Paraguari: cerca de Sapucay, Hassler de (US). PERU. Foren valley of Río 05*19'S, 7914'W, 1700 m, 11 Apr. (US). Cuzco: tando a cen Vargas 17225 (US). Huancavelica: entre ee y Virgen Pampa, 1200 m, Tovar 4585 (US). San Martin: Alto Rfo Huallaga, 60-900 m, Dec. 1929, Williams 5800 (F, GH, NY, US). Henrard (1941) menciona que P. quadriglume sería una especie teratológica, por el hecho de po- seer 4 “glumas,” de las cuales indica que la cuarta corresponde a una gluma III y no a la lemma de la segunda flor. Sin embargo el análisis de abundante material de la especie descarta la hipótesis de Hen- rard, puesto por Palacios (1968). Esta peculiar estructura de la espiguilla es prác- ticamente unica dentro del material examinado hasta el momento, en el género Panicum, habién- dose hallado ocasionalmente una estructura similar coincidiendo así con lo anteriormente ex- en un especimen de P. crateriferum Sohns (Zuloaga « Sendulsky, 1988) En esta especie se observa, tal como en P. pe- ladoense y P. exiguum, que los antecios son ne- gruzcos y caedizos independientemente de las glu- mas a la madurez de la espiguilla. 30. Panicum stramineum Hitchcock € Chase, Contr. U.S. Natl. Herb. 15: 67, fig. 50. 1910. Panicum capillare L. var. stramineum (Hitchc. & Chase) Gould, Madroño 10: 94. 1949. cum hirticaule J. sl var. stramineum (Hitchc. & Chase) Beetle, Phytologia 47: 383. 1981. TIPO: México, Sonora: Guaymas, 1887, Palmer 206*,** (holótipo, US-592753; isóti- pos, F, NY, SI, US-742139, 823636, 823638, W). Figura 28. Panicum e Renvoize, Kew Bull. 37: 325. 1982. TIPO: Brasil. Bahia: 41 km N of Senhor do Bonfim on the BA 130 highway to Juazeiro, 50 m, 26 Feb. Volume 83, Number 2 Zuloaga & Morrone 271 1996 Panicum Subg. Panicum Secc. Panicum Figura 28. Panicum stramineum.—A. Porción de una caña florífera.—B. NE ly an vista del lado de la gluma inferior. —C. Espiguilla vista del lado de la gluma superior. —D. Pálea inferior. —E. Antecio superior visto del lado de la lemma.—F. Antecio superior visto del lado de la pálea.—G. Cariopsis, vista esc he —H. Cariopsis, vista hilar. (Palmer 206.) Annals of the Missouri Botanical Garden 1974, Harley et al. 16384 (holótipo, CEPEC; isóti- pos, K, RB, US-2955107). Plantas anuales, cespitosas. Cañas erectas de 10-70(-125) em de alto, geniculadas y ramificadas en los nudos inferiores; entrenudos de 2-15 cm de largo, cilíndricos, glabros; nudos corta a largamente pilosos. Vainas de 2-9 cm de largo, glabras a hir- sutas, con pelos papilosos, verdes a purpúreas, con los márgenes membranáceos, glabros, o con un margen pestañoso. Lígulas membranáceo-pestaño- sas, de 0.9-2.4 mm de largo, con largos pelos por detrás en la base de la lámina o sin los mismos; cuello glabro a híspido. Láminas linear-lanceola- das, de 4.5-30(42) cm de largo, 0.3-1.3(22) cm de ancho, las basales de menor tamaño, erectas, ascendentes, poco divergentes, planas, glabras a esparcidamente papiloso-pilosas en ambas caras, de base redondeada y ápice atenuado, con los bor- des escabriúsculos, glabros a papiloso-pestañosos hacia la base. Pedúnculos hasta de 24 cm de largo, cilíndricos, glabros. Inflorescencias terminales exer- tas, laxas, de 2.5-22(-38) cm de largo, (1-)4-10- 20) cm de ancho; eje principal anguloso, escabriús- culo; pulvínulos glabros a cortamente pilosos; ramificaciones de primer orden alternas a subo- puestas, divergentes, desnudas hacia la base; ram- ificaciones de segundo orden escabriúsculas, div- ergentes, con espiguillas dispersas sobre los ejes; pedicelos claviformes de 0.7-2.5 mm de largo, es- cabriúsculos, con pelos blanquecinos hacia el ápi- ce. Inflorescencias axilares presentes, similares a las terminales. Espiguillas globosas, de 2.3-3.2 mm 0.8-1.4 mm de ancho, túrgidas, abrup- tamente acuminadas, no estipitadas, pajizas; gluma de largo, superior y lemma inferior subiguales, tan largas como el antecio superior o deg 0.4 mm. Gluma inferior ovado-acuminada, de 0.7—1.2(— mm de largo, 4—% o un poco más del los. de la espiguilla, de ápice obtuso a agudo, 5—7-nervia, el nervio medio escabriúsculo hacia el ápice. Gluma superior de 1.8-3.1 mm de largo, 9-11(-13)-nervia. mma inferior glumiforme, 1.7—3 mm de largo, 9— 11-nervia. Pálea inferior ovado-lanceolada, de (1.4—)2.3-2.7 mm de largo, 0.5-0.9 mm de ancho, igual o más larga que el antecio superior, mem- branácea, hialina, glabra, con los márgenes supe- riores ciliados a glabros. Antecio superior elipsoide, de 1.5-2.5 mm de largo, 0.7-1.3 mm de ancho, crustáceo, liso, lustroso, pajizo y con tintes grisá- ceos a la madurez, con dos cicatrices basales de 0.2 mm de largo; lemma 7-nervia; pálea con papilas simples hacia el ápice, en ocasiones distribuidas en toda la superficie. Cariopsis ovoide, de 1.2—1.6 mm de largo, 0.6-0.9 mm de ancho, pajiza; hilo punctiforme; embrión Y del largo de la cariopsis. Distribución y ecología. Presente en los Esta- dos Unidos de América y México y en América del Sur desde Venezuela hasta la Argentina (Fig. 9). Crece en campos o en sitios abiertos en bordes de caminos o vías férreas. Es también frecuente en bosques decíduos o en sitios inundados, ocasional- mente apoyante sobre la vegetación; habita desde el nivel del mar hasta 1100 m de elevación. “Capim-lanudo” (en Brasil). Material representativo citado. ARGENTINA. Cata- marca: Santa Lucía, 2 abr. 1950, Brizuela 982 (GH, NY). rmosa: Dpto. Matacos, Ingeniero Juárez, 14 ene. 1957, Burkart 20275* (SI, US). Jujuy: Dpto. San Pedro, Chag- uaral, Finca Jure, 18 abr. 1983, Ahumada & Castellón 4829** (SI). La Rioja: Dpto. Gral. San Martín, Estancia La Diana, ca. 800 m, Mar. 1907, Kneucker 730 (GH, ISC, a MO, NY, SI, W). Salta: Dpto. Campo Santo, Giiemes, 3 abr. 1945, O'Donell 2629 (NY, SI). BOLIVIA. Beni: Trinidad, 9 abr. 1979, dio & Schinini* (SI); Prov. Gral. dd n en la zona de influencia del río e ‘uma, 200 1 ar. 1990, “Beck 15176 (SI). Chu- e. 1992, Saravia Toledo & Nelson Nombre vulgar. a: Isi fanus 10412 (S SI). nacio, 22 km N an José de Chiquitos, 17%35'S, 60°45'W, 320 m, 1 I 1986, Kile 1711 (F, SI). SIL. Bahia: 4 km north of Senhor de Bonfim, disturbed ground by roadside, 550 11'W, 10%27'S, 24 Feb. 1974, Harley et al. 16294 R, RB, US). Ceará: Ig- uatú, iar 4396 (GH, RB, US). Maranhao: Mun. Lo- reto, Ilha de Balsas, region between the Balsas and Par- n rivers, 26 km south "d Rd 200 m, 11 Feb. 1970, en & Bien eds o Grosso do Sul: Bas de Cerro do Urucum, 23 e S Cora 22 nov. 1977. Allem & Vieira 1479% (MO, SI). P zenda Nacional to Picos, low ground near small bn 4 abr 1934, Swal- len 4192 (RB, US). Para: Ilha do Marajó, Faz. Gavinho, i ldi 261 (NY). ESTADOS UNIDOS DE AMERICA. Ar- zona: Pima Co., 18 Aug. 1937, Proctor 2571 (US); under Mesouite bushes, White's Mills, 5 Sep. 1867, Palmer 270 MO). MEXICO. Chiapas: Mun. rcd dec iduous for- est, at Río Las Arenas, 1: : , 27 Aug. 1974, Breedlove 36779-A (MO). Lagunillas, | Aug. ean: Tancitaro, road to Hda. California, 1000 ft., 1 1941, Leavenworth & Hoogstraal 1453 (GH, MO, NY, US). Nayarit: entre Concepción y Acaponeta, 29 jul. 1897, Rose 1889* (MEXU, NY, US); entre Rosario y Acoponeta, 28 jul. 1897, Rose 1883 (US); 26.8 mi. S of Sinaloa state line i in cay 24 July 1975, Dunn et al. 21855* (MO, : eón: 12 mi. 5 of Monterrey, 14 July 1933, n 404 (US 3531 ~ S). entre Rosario y Ac E. Rose 1878 (GH, US); vic vu m Labradas, 18 Sep. 1925, Ferris & Mexía 5069 (MO, US). Sonora: Orozi.. near Río Yaqui, 7 Sep. 1935, Pennell 20221 (MO, US). PARAGUAY. Alto Par- guay: Colonia San Lázaro, cerca del Río Apa, ene. 1931, Rojas 5476a (US). Boquerón: Tinpunké, cercanía Río Pilcomayo, 28 oct. 1985, Mereles 715 (SI); entre las Es- tancias Santa Ramona y San Alberto, Puesto Sastre, 9 mar. ser arii 51 (SI). = 5 YE N o juni = o yz e) 7e Y) vi , 10 ene. a i (MO. SI). Presidente Hayes: Ruta Concepción- Volume 83, Number 2 1996 Zuloaga & Morrone 273 Panicum Subg. Panicum Secc. Panicum T Colorado, a 28 km del puente, 18 feb. 1990, €: in Departamento: San Salvador, Mis | e 11027 Aids) PERU. Tumbes: eet i ie to Tutumo, about 20 km from Tumbes, 26 abr. 1969, Simp- son & Schunke V. gm (F, US). ALL pao: entre los cafios Balza anati, carretera Mantecal-Quin- tero, 7°45'N, 69°40’ wW 5 ago. 1989, Zuloaga et al. 4340* (MO, SI, VEN); entre Achaguas y San Fernando, cerca de Bethel, 7°48'N, 67°55'W, 6 ago. 1989, Zuloaga et al. 4350* (SI); Mantecal, 17 jul. 1975, Ramia & Montes 5345* (US, VEN). Carabobo: Valencia, aprox. 4 km al E de Valencia en la autopista entre Valencia y Maracay, 8 jun. 1978, Burandt Jr. & Wingfield V0242 (VEN). Fal- cón: Paraguaná, 2 km E de Sta. Ana, 4 ene. 1980, Wing- field 7361 (VEN). Guárico: S de Calabozo, 22 jul. 1976, Castillo 583* (VEN). Zulia: Perijá, carretera Las Piedras- San José-Las Laras, en Km 9 al E del empalme con car- retera Perijá, 10 sep. 1977, Bunting 5447 (VEN). Sin Estado: Hato Becerra, a 2 km S del Río Orituco, 11 jul. 1961, Blydenstein 148 (VEN). Zuloaga (1987a) sinonimiza esta especie con P. hirticaule y posteriormente Zuloaga (1989) y Kil- leen (1990) citan material de Sudamérica um de P. stramineum bajo P. hirticaule. Recientemente Davidse (1994) correctamente revalida la especie, sefialando, entre los caracteres distintivos de la misma, que posee espiguillas más obtusas y diver- gentes con gluma inferior relativamente más pe- quefia y pálea inferior mucho más larga. Panicum caatingense, especie sinonimizada con P. hirticaule por Zuloaga (1989) y Killeen (1990) debe ser con- siderada en la sinonimia de P. stramineum. En P. stramineum la gluma inferior varía entre 14 y % del largo de la gluma superior y la lemma inferior. Esta especie es afín a P. hirticaule, de la que se diferencia por poseer espiguillas globosas, túrgidas, más o menos abruptamente acuminadas y con la pálea inferior desarrollada, alcanzando el mismo largo de la lemma inferior. Las plantas usualmente tienen láminas erectas, poco divergentes del caule. El ejemplar Simpson 829, de Perá, tiene espi- guillas similares a P. stramineum, pero la gluma inferior es un poco mas larga que % y los pedicelos son aparentemente glabros y los ejes esparcida- menten pilosos. 31. Panicum tamaulipense F. R. Waller & Mor- den, Syst. Bot. 8: 221. 1983. TIPO: México. Tamaulipas: Sierra de Tamaulipas area, be- tween La Chona and Río Santa Olaya, low deciduous forest, 50 m, 26 Sep. 1956, Martí- nez Martínez & Borja Luyando F-2161 (holó- tipo, TEX no visto; isótipo, US-2463101). Plantas perennes, cespitosas, de 50-80(-120) cm de alto. Cañas simples, glaucas, innovaciones intravaginales, erectas, paucinodes; entrenudos de 3-10 cm de largo, cilíndricos, huecos, de 1-2 mm diám. en su base; nudos pilosos con pelos cortos. Vainas de 8-12 cm de largo, iguales o mayores que los entrenudos, estriadas, hirsutas, con tintes vio- láceos, con pelos tuberculados adpresos, un margen pestañoso, el restante glabro. Lígulas membraná- ceo-ciliadas de 0.5-1.2 mm de largo; cuello piloso a glabro. Láminas linear-lanceoladas, de 20-35 50) cm de largo, 0.5-0.8 cm de ancho, erectas, ascendentes, rígidas, planas, ligeramente rizadas en la base, glaucas, glabras, de base angostada y ápice agudo, con los bordes ligeramente involutos, esca- briúsculos, el nervio medio manifiesto. Pedúnculos hasta de 50 cm de largo, glabros, lisos, pajizos o con tintes violáceos. Inflorescencias exertas, de 15— 30 cm de largo, 6-15 cm de ancho, ramificaciones inferiores hasta de 18 cm de largo; eje principal escabroso, glabro, ligeramente ondulado; pulvínu- los glabros, ramificaciones de primer orden alter- nas, divergentes, pedicelos claviformes, de 4-1 mm de largo, divergentes, flexuosos, glabros, es- cabriúsculos. Espiguillas ovoides, de 1.8-2.2 mm de largo, 0.7— e ancho, de ápice abrupta- mente acuminado, no estipitadas, glabras, pálidas o con tintes violáceos hacia el ápice, con los ner- vios verdosos; gluma superior y lemma inferior su- biguales, tan largas como el antecio superior. Glu- ma inferior ovado-acuminada, de 1-1.2 mm de largo, alcanzando Y del largo de la espiguilla, 5- nervia. Gluma superior de 1.8-1.9 mm de largo, acuminada, 7—13-nervia. Lemma inferior glumifor- me, de 1.8 mm de largo, 7—9-nervia. Pálea inferior lanceolada, de 1-1.2 mm de largo, 0.3 mm de an- cho, % del largo del antecio superior, membraná- cea, hialina, glabra, de ápice finamente denticula- do. Antecio superior ovoide, de 1.5-2 mm de largo, 0.6-1 mm de ancho, crustáceo, glabro, liso, lustro- so, con dos cicatrices basales ca. 0.2 mm de largo, castañas a la madurez; lemma 7-nervia, pálea papi- losa hacia el ápice. Cariopsis ovoide, de 1 mm de largo, 0.5 mm de ancho, blanquecina; hilo puncti- forme; embrión % del largo de la cariopsis. Distribución y ecología. Endémica de México, donde se encuentra en campos de los departamen- tos de San Luis Potosí y Tamaulipas (Fig. 9); llega hasta los 350 m; florece y fructifica entre mayo y diciembre. Material adicional examinado. MEXICO. San Luis Potosí: El Banito, 100 ft, 22 July 1939, Chase 7558 (US); Mun. El Pujal, Chase 7492 (US); 2 miles S of Huichi- hoayan, McGregor et al. 870* (US); Ciudad Valles, 5 May 1962, Beetle M-718 (US); Valles, 800 ft., 17 July 1933, Fisher 3313 (US); i way 55 to Xilitla, aepo: 1431 te to Limón, 3 May 1931, Swallen 1618 (US); 64 km from Soto La Marina on d road to Casas and Victoria, 350 m, 274 Annals of th Missouri A m Garden 4 Oct. 1956, Martínez Martínez & Borja gando F-2364 dt (US); 35 km from Victoria on the road to Casas and Soto La Marina, 280 m, 3 Oct EI p d & Bor- ja Luyando F-2333* (US); 5 n l toward > afa Soto la Marina, 800 ft., 11 Dec. a 59, ae hs eld & John- ston 4956 (MEXU). Especie afin a P. hallii var. filipes, de la que se distingue por tener espiguillas menores, de 1.8-2.2 mm de largo y láminas de 20—35(—50) cm de largo por 0.5-0.8 cm de ancho. TAXON Duposo Panicum capillare var. minus Dóll, in Martius, Fl. Bras. 2(2): 202. 1877. Doell indica que esta variedad posee panoja pau- ciflora y hace referencia a la cita de Nees de Pani- cum philadelphicum (en la que este autor incluye a P. capillare var. minus Muhl. en la sinonimia de P. philadelphicum). Literatura Citada Adams, C. D. 1972. TOS Plants of Jamaica. Univ. est Indies, Mona, Jamai Arenkova, D. N. 1940. Poly; ploid races in Millet (Pani- cum iv ird L.). Compt.-Rend. (Dokl.) Acad. Sci. URSS 332-335. Avdulov, N. P. 1931. Karyosystematischen Untersuchung der Famile der Gramineen. Trudy Prikl. Bot., Suppl. 44: 1-428 Beattie, A. J. 1983. oe of ant- ari gang E Sonderb. Naturwiss. Vereins. Hamburg 7: 241 Beetle, 1977. Tidi V grasses from Ne o V. Phytologia 37: 317—407. —— ———. 1987. Noteworthy grasses from Mexico XIII. Phytologia 63: 209—297 Bentham, G. 1878. Flara australiensis 7. Lovell Reeve, London Berg, R. Y. 1985. Spikelet structure in. Panicum aus- traliense (Poaceae): Taxonomic ecological implica- tions. Austral. J. Bot. 33: 579-58 Brown, W. V. 1948. A pe cal ind: in the Gramineae. Amer. J Bot. 35: 382-395 95 a numbers of some Texas grasses. Ball Torrey Bot. Club 78: 292-204 The Kranz syndrome and its sies in grass ee s. Mem. Torrey Bot. Club 23: 1-97. Cabre ra, A. L. 1970. Gramíneas. En: A. bes (ed- "P Fl. Prov. Buenos Aires, Colec. Cien INTA 4(2): —024. ds W. D. 1989. Gramineae, XXIV Paniceae. En: E Launert & V. Pope (editores), Flora Zambesiaca 10: 9— 40. ie Litho Printers pd Whitstable, Kent. & S. A. Renvoize. 1986. Genera Graminum. Kew Bull., ^d a P Correll, D. . Johnston. 1970. Manual of the Vascular d "E Texas. vn: C. L. Lundell (editor), Contr. Texas Res. Foundation, Bot. Stud. 6: 164—182 Davidse, G. 198 Press, cw on D. 994, a. Pp. 302-318 en G. M. Davidse, M. Sousa S. & A. O. Chater (editores), Flora Mesoame- ricana, Vol. 6, Alismataceae a Cyperaceae. ose Nacional Autónoma de México, México, D. issouri Botanical Garden, St. Louis; The Natural Tom Mu- seum, London. —— ——— & R. VW. Pohl. 1972. Chromosome numbers and notes on some Central American grasses. Canad. J. Bot. 9 3. Delay, C. 1950. Nombres c hromosomiques chez les Pha- nerogames. n pus l on d a om Joell, J. € ribe 3, Par . F. P. von Martius A t os Brasilien ansis 2 A Ne 342. Mu- nich, arg Le eu üg. Dubcovsky, J. F Zuloaga. 1991. Números cromosó- micos de especies sudamericanas de Panicum a o- : Paniceae). Bol. Soc. Argent. Bot. 27: 201-206 li R P. 1976. A procedure for standardizing compar e leaf blade as A Disusbutien of the Kranz syndrome in nthe southern s Eragrostoideae and Panicoideae ac- cording to the bundle sheath anatomy and cytology. o 9: 73-110. 1979. A proc 'edure for standardizing compara- live leaf anatomy in the Poaceae. II. The epidermis as seen in surface view. Bothalia 12: 641—671. 1988. Leaf anatomy and systematics Ai A (Paus eae: Panicoideae) in southern Africa. En zold- blatt & P. P. Lowry (editores), Modern T Sie. ies in African Botany. Monogr. Syst. Bot. Missouri Bot. Gard. 25: 129-156. Fairbrothers, D. E. 1953. era in the Capillaria group of i n um in Arizona and New Mexico. Amer. J. Bot. 40: EN in the Capillaria group of ai ‘um. Tesis Doctoral Graduate of Cornell University, Ith Td Index of c iE ud numbers of . Rerum Nat. Univ. 1976. Comenianae, Bot. 25: 1-18 Fernald, M. L. 1919. Me um § Capillaria in New En- _ gland. Rhodora 21: 14. Mizianty Mirek. 1981. C ioa cid aba rs of Polish vase a plants. Fragm. Florist. Ge bot. 27: 581-590. Gould, F. W. 1958. Chromosome numbers i in southwest- Bot. 45: 75 ern grasses. Ame . 1960. ciii IIl. Amer. J. Bot. 47: 873-8 . 1965. Mi Ra eA numbers in some Mexican grasses. PA Soc. Bot. co) 29: 49-62. —— ———. 1066. Chromosome humbe rs " some grasses. Canad. J. Bot. 44: 1683-1696 —— ———. 1968. Chromosome numbers of Texas grasses. Canad, Y n 46: 1315-1325. The Grasses of Texas. Voss. s Station, Texas. 1979. Poaceae. En: R. A. pedi (editor), Flora Chromosome numbers in southwestern T. Mexican Texas A & M Univ. ate Lesser Antilles 3: 25-220. Arnold NE Harvard University, Jamaica Plain, isetts. & T. R. Gedcom, 1970. C Sidi numbers of some Mexican and Colombian grasses. Canad. J. Bot. 48: 1633-1639 Gutiérrez, M., V. E. Gracen € G. E. 1974. Bio- chemical and Pia "i in C, plants. Planta Eu) 119: 279— . Edwards & 7 5 Brown. Edwards. 1976. PEP car- Volume 83, Number 2 1996 Zuloaga & Morrone 275 Panicum Subg. Panicum Secc. Panicum boxykinase containing species in the Brachiaria group of the subfamily Panicoideae. Biochem. Syst. & Ecol _49. Hackel, E. 1887. Cyperaceae und Gramineae. En: H. G. A. Engler & K. A. E. Prantl (editores), Die Natürlichen Pflanzenfamilien. d ce ad (Reimpresión.) Hattersley, P. W. & L. Watson. 1976. C, grasses: An an- atomical criterion for dentine between NADP- Malic enzyme species and PCK or NAD-Malic enzyme species. Austral. J. Bot. 24: 297-308. Henrard, H. Th. 1941. Notes on the nomenclature of some grasses II. Blumea 4: 496-538. Hitchcock, A. S. 1908. Types of American grasses: A study of the e species of grasses described by Linnaeus, Gronov loane, Swartz, and Michaux. Contr. xs Natl. Herb. 12: 113-158. . 1927. The grasses of Ecuador, Peru, and Bolivia. Contr. d. Natl. Herb. 24: 291—556. — ———. 1936. Manual of the Ave of the West Indies. U.S. D. ie Misc. Publ. 243: . 1951. Manual of i diy of the United States. U.S.D.A. Misc. Publ. 200: 1-1051. & A. Chase. 1910. The Ne ieee species of Panicum. Contr. U.S. Natl. Herb. 15: 1-396. Holmgren, P. K., N. H. Holmgren & L. C. Barnett. 1990. Index Herbariorum. Part I: The Herbaria of the World. New York Botanical Garden, New York Hsu, C. C. 1965. The classification of Panicum (Gramin- eae) and its allies, with special reference to the char- acters of lodicule, style-base and lemma. J. Fac. Sci. Univ. Tokyo, Sect. 3, Bot. 9: 43-150. Jensen, W. A. 1962. Botanical Histochemistry. Principles and Practice. W. H. Freeman, San Francisco Killeen, T. J. 1990. The grasses of Chiquitanía, Santa Cruz, Bolivia. Ann. Missouri Bot. Gard. 77: 125-201 Krishnaswamy, N. 1951. Origin and distribution of cul- tivated plants of South Asia: Millet. Indian J. Genet. Pl. Breed. 11: 67—74. Lazarides, M. & R. D. Webster. 1984. Yakirra (Paniceae, oaceae), a new genus for Australia. Brunonia 7: 289— Longhi-Wagner, H. M. & A. I. I. Boldrini. 1988. Gra- mineas da Estagáo Ecologica de Aracuri, Esmeralda, I^ Pos do Sul, Brasil. Iheringia, Bot. 38: 21-42. A. & D. Lóve. 1981. En: Chromosome number re- Es LXVIII. Taxon 29: 53 35. Luces, Z. 1963. Las Gramíneas M Distrito Federal. Ins- tituto Botánico, Direc. Recursos Nat. Renovables. Min- isterio de Agricultura y Cría. McVaugh, R. 1983. Gramineae. En: W. R. Anderson (ed- itor), Flora Novo-Galiciana. A Descriptive Account of the Vascular Plants of Western Mexico. Vol. 14. Univ. Michigan Press, Ann Arbor. 1960. Anatomy of Monocotyledons. I. Gramineae. Oxford Univ. Press, Oxford. Morrone, O. & F. O. Zuloaga. 1991. Revisión del género Streptostachys (Poaceae: Paniceae). Su posición siste- mática dentro de la tribu Paniceae. Ann. Missouri Bot. Gard. 78: 359-376 . 1992. Revisión de las especies sud- americanas de los géneros Brachiaria y Urochloa (Poa- ceae: Panicoideae: Paniceae). Darwiniana 30: 43-109 & E. Carbonó. 1995. Revisión del grupo camita del inn Paspalum (Poaceae: dor uan Paniceae). Ann. Missouri Bot. Gard. 82: 82— Nash, G. V. 1903. Poaceae. Pp. 48-160 en J. P ‘Small (editor), Flora of - Southeastern United States. [Pub- lished by J. Nees von juni C. G. D. 1829. Agrostologia Bras- iliensis . . . Stuttgart & Tübingen. (This is vol. 2, n 1, Gramineae, of Flora a as se Enumeratio ... ed- ted by C. F. D. von Martiu Nic E. G. 1978. Gramineae. En: M. N. Correa (edi- tor), Fl. Patagónica, Colec. Cient. INTA 6(2): 532-538. Núñez, O. 1952. Investigaciones cariosistemáticas en las gramíneas argentinas de la tribu Paniceae. Revista Fac. v. Nac. La Plata 28: 229-255 inat Leaf anatomy, post-illu- rst and NAD-malic enzyme activity of Panicum dichotomiflorum. Pl. Cell Physiol. 21: 1329— 1333. N. Chonan. 1982. C, syndrome of the species in the Dichotomiflora group of the genus Pani- m. Bot. |a (Tokyo) 95: 339-347. Palacios, R. A. 1968. Variaciones en la estructura de las espiguillas en algunas especies argentinas de Panicum. Bol. : 38-43. oc. Argent. Bot. ; Panicum. En: A. Para ges Fl. IL Entre Ríos, Colec. pou INTA 6(2): 2 Parfitt, B. D. : Chromosome dne reports LXXXI. us 30: 515.51 6. Parodi, L. R. 1946. Gramíneas Bonariensis. Clave para la Determinación de los Géneros y Enumeración de las Especies, 4 ed. Acme Agency, Buenos Aires. Pijl, L. van der. 1982. Principles of TE in Higher Plants, 3rd ed. Springer-Verlag, New Pilger, R. 1931. Bemerkungen zu Panicum ud verwand- ten Gattungen. Notizbl. Bot. Gart. Berlin-Dahlem 104: 40-245. 40. Gramineae. III: Unterfamilie Panicoideae. n: A. Engler & K. Prantl (editores), Die Natiirlichen Planznímiien 2da. edición, 14: 1-208. Engelmann, Leip Pohl, R Sw 1980. Gramineae. En: W. en Y Flora Costaricensis. Fieldiana, Bot. 4: 350-39 : N. E. Stone. 1987. xin struc- mical aio in leaf blades of C, grasses (Poaceae). Austral. J. Plant Physiol. 14: 403— 420 Reeder, J. R. 1971. Notes on Mexican grasses IX. Mis- cellaneous chromosome numbers—3. Brittonia 23: 7 . 197 Parer numbers in western grasses. Amer. E E 64: —110. ——— & C. G. o 1971. T Pp. 823-892 i rter (editores), Flora of en I. iggins Prag Islands. Stanford Univ. Press, Stanford, Cal- iforn Ridley. H. N. 1930. The Dispersal of Plants Throughout the World. L. Reeve, London Rosengurtt, B., B. R. Afrillaga de Maffei € P. Izaguirre de Artucio. 1970. Gramíneas Uruguayas. Montevideo, ruguay. Sass, J. E. 1940. Elements of Botanical Microtechnique. McGraw Hill, New York and London. Schmid, R. € D. Turner. 1977. Contrad 70, an effective softener of herbarium material for anatomical study. Taxon 26: 551-556. Sherif, A. S. & M. A. Siddiqi. 1988. Poaceae. En: S. I. 276 Annals of the Missouri Botanical Garden Ali € S EM H. Jafri da Flora of Libya 145: 1— . Morrone. 1991. Panicum aquarum, a new aateh Univ. Dept. Bot., Tripoli. South Finis species of Panicum (Poaceae: Panicoi- Silberhaueet Gottes I. 1984. Fruit dispersal and try- deae: Lia niceae). Novon 1: 185-190. panocarpy in Brazilian cerrado grasses. Pl. Syst. Evol. 1992. Panicum ray sistem, (Po- 147: 1-27. aceae: Panic e Paniceae), ys nueva especie para Smith, L. sshausen & R. Klein. 1982. Gra- mineae, part T. p 633- de en R. Reitz (editor), Flora Ilustrada Catarinense, Bra í. . ‘Chromosome numbers of plants. 970. B6 as number reports XX V. Taxon 19: 102-11 i , um En: D. Prain (editor), Flora of Tropical eres 9, L. Reeve, London. ?. G. 1855. Synopsis Plantarum Graminearum. En Synopsis Bud Glumacearum 1: 1-475. J 3. Gramineae. En Flora of Panama. 04—28( w grasses from Mexico, Central Amer- ica, and ean, Contr. U.S. Natl. Herb. 29: 395—428. 1955. Flora of Guatemala: Part III: Grasses of Guatemala, Fieldiana, Bot. 24: 1-390. Gramineae. Pp. 237-301 en F. Shreve I. L. S (editores), Vegetation and Flora of the Se nora Desert, vol. 1. Stanford Univ. Press, Stanford, Cal- ifornia. o J. R. 1985. Miocene fossil grasses: possible aptation in reproductive bracts (lemma and palea). pu nn. Missouri Bot. Gard. 72: 843-851. Tovar, O. wes Las gramíneas (Poaceae) del Perú. Ruizia 3: 347- Tzvelev, N. A 1976. Zlaki SSRR [Grasses of the Soviet ap Academiya Nauk. SSSR, Botanicheskiy Institut omarova. Leningradskoye Otdeleniye, Len- de [English Vi E published fo r the Smitisan: ian Institution Libraries and the National Science Foun- dation, 2 am D.C., by Amerind Publishing, New Delhi. 1983.] Vahidy, ^ f^ G. pico ds Y. Schigenobu. 1987. Chro- mosome counts of Missouri puc ceae and Poaceae. Ann. wee s Gar rd. 74: 4 Vegetti, A. & nsiero. 1993, cencia da Pani ‘um sección Panicum (Poaceae », Argent. Bot. 29: 7-10. Panicum section Diffusa. Pp. 469— 477 en F. W. Gould (editor), The Grasses of Texas. Texas A & M Univ. Press, College Station, Texas. 76. Systematic Study of Panicum section Dif- fusa (Poaceae) in North America. Tesis Doctoral, Grad- uate College of Texas A & M University, College Sta- tion, a de la inflores- : Pani- Texas. & C . W. Morden. 1983. Panicum Valen d (Poaceae: Paniceae): A new species from Mex s Bot. 8: 221-222. Webster. R. D. 1987. J. Cramer, Stuttgart. Zuloaga, F. O. 1987a. Systematics of the New World spe- cies of Panicum (Poaceae: Paniceae jJ. Pp. 287-306 en . R. Soderstrom, K. Hilu, C. S. Campbell & M. E. Barkworth (editores), Grass S niemals and Evolution. Smithsonian Institution Press, Washington, D.C. 198 A revision of Panicum subgenus Pani- cum section Rudgeana (Poaceae: Paniceae). Ann. Mis- souri Bot. Gard. 74: 463—478 989. El género Panic um en la República Ar- gentina. IHI. Darwiniana 29: 289—370. Nn = The Australian Paniceae (Poaceae). ee: el Brasil. ie Bot. 42: 13-2 & T. Sendulsky. 1988. na revision of Panicum subg. piana sect. e Apes eae: Pani- ceae). Ann. Missouri v Gard. 7 292. , O. Morrone & A ee M morfológico e histofoliar » las especies americanas de género Acroceras (Poaceae: Paniceae). Darwiniana 28: 191-217. Estudio exo- Jubcovsky € O. Morrone. 1993. Ir HHigeneno ganii era in genus Panicum (Poaceae: Pani- nerical analysis. Casi d. coideae: Paniceae): A nun Bot. 71: 1312-1327. Apendice. Cada especimen es citado por el primer nombre del colector en el caso en que otros coleccionistas participen de la colección. Se indica entre paréntesis el número de la especie: Panicum alatum var. alatum (la), P. alatum var. jns dora (1b), P. alatum. var. minus (lc), P. aquarum (2), P. aztecanum (3), > bergii var. pilosissimum (4b), P. ca larioides (6), P. chasei (7), P. decolorans (8), (9), P. ephemeroides (10), P. ae (11). P. flexile (12), p um (13), P. je ien (14), P. hallii var. hallii 5a), P. hallii var. filipes (15b), a He epee (16), P. hir- sutum (17), P. hi vim ie var. pes caule (18a), P. hirticaule var. verrucosur n (18b), P. hispidifolium ( (19), P. lepidulum (20), P. magnispica ula (21), P. miliaceum (22), P. mohav- ense (23), P. mucronulatum (24), P. pampinosum (25), P. parcum (26), P. peladoense (27), P. philadelphicum (28). P. quadriglume (29), P. stramineum (30) y P. tamaulipense (31). Abbott 1027 (9); Abrams 12902 (30); Agniel 10269* (8), 10393 (8); Ahart 6168 (22); Ahumada 911 (4a), 1502 (4a), 4829** (30), 5023 (30); Ainslie s.n. (12); Alberdi 12269 (22); Allard 10422 (12), 14148 (14), 15744 (9) 15902 (9); Allem 1479 (30), 2145 (30), 2233 (11); pod 5235 (14), 5298 (14); Alvarez 603 (4a); Amer. Gr. Hb. 504 (15a); An- derson 358 (12), 897, 899 (18a), 1365 I 2096 (5). 381* (11), s.n. (1c); Andrade Lima 71-6363 (11); Araujo 151 (7), 186 (7), 289 (4a), 527 (7); Archer 20 s 3968 (14); Arechavaleta 24a (4a); Aristeguieta 4301 (19), 4334. (19), 5665 (19), 5755 (19); Arséne 2417* (20), 2864 (20). 2985 (20); Artupo R. 1723 (15a); Arzivenco 640 (5), 950 (27); Avifia 1248 (18b). Badillo 7 (14), 146 (14), 690 (14); Báez s.n. s.n. (28); Balansa 12 (Aa), 13 27, 4357** (27); Baldwin Jr. 14321* (8); Balick 133 (11); dod 20 (5), 120 (5); PS 493* (29); 9 (8); Barkworth 51 5a); Barneby 2455 : s.n. (7): Bartlett 10d an 10549 (15a), 10591 (15a), 2 (8); Bebb 2 dw 5957a (22); Bain =~ 2 3 = =~ 3 => n J Y ~ ~ = es) — (28); Bennett s.n. (12): Be ma s.n. (15a): Berte Hos s.n. Bertero s.n. (9); Bertoni 5381 (Aa), 5532 (21). 5 : 51-11086 (29), 51- 11413 TE. a -1 x Black 86C (11). 88C (11), 51-11686 (29), 54-16594 id Blair 731 (11); Blake 7326A (18a); Blan ; Blanchet 2961 (24); chard s Blanco 512 (1c); als s.n. Mor Blumer 1683 (15a); Volume 83, Number 2 1996 Zuloaga & Morrone 277 Panicum Subg. Panicum Secc. Panicum Blydenstein 148 (30), 1846 (2); Boege 1858 (18a); Boelcke 11889 (4a); Boom 6189 (19); Bourgeau 2751 (14); Box 18 (14), 39 (14), 49 (14), 176 (17), 190 (17), 3703 (17), 3836 (17); Brace 4812 (14); Brade 8144 (27); Brandegee 4 nA 8 (la), 42 (1c); Braum 2648 (12), 2687 (12), s.n. (12), s. (5); Breedlove 10630 (1c), 13492 (26), 13833 (26), 27334 (18a), 28359 (26), 28837 (17), 36779-A (30), 36852 (20), 30875 (26), 37573 (19), 37654 (19), 41540 (18a), 46945 (17), 51680 (18a), 52871 (26), 54261 (26), 54117 ye 54908 (14), 55245 (17), 56081 (18a); Britton 314 (9), 382 (9), 424 (9), 649 (9), 745 (14), 1380 (9), 2356 (9), 4064* (9), 6123 (9), 6349 (14), 6859 (9), 13177 (9); Brizuela 982 (30); Broadway 2629 (17), 4586 (17), 4895 (17), 9291 9* ** (28); Brunken 241 (15a), 360 , 0); Buchtien 4174 (29); Bunting 5447 (30), 5568 (30), 5663 (30), Burandt Jr. V0238 (17), V0242 (30), V0498 (11); Burchell 6769 (29), 6794-2 (27), 6827 (11), 8813 (2); we 7875 (19); Bur- 8 (4a), 15943 (5), — 858 (15a), 1087 (12), 1156 (15a), 5120 (28), s.n. (28). Cabanillas 15 (18a), 17 (18a): Cabrera 772 (4a), 2237 (17), 2461 (17), 2484 (4a), 2630 (17), 7008 (4a), 9837 (4a), 11739 (14), 22167 (27), 25988 (4a), 2652 (30), 28697 (4a), 29057 (4a), 32328 (4a), 32385 (27), 32388* (27), 32416 (4a), 32423*,** (7), 32704 (4a), 34220 (30); Calderon 1672 (22); Calzada 4492 (17), 6250 (17); Camby 6 (1b); Campelo 2154 (24); Canby 221 (12); Canoniero 92 (3b); Cárdenas 1647 (30), 3533 (30), 3534 (30); Carter 5 (1c), 4726 (la), 4788b (1c), 4970*,** (1b), 5016 (18a), 5074** (1b), 5353 (1b), 5374 (1b), 5548 (1b); Casali 44 (4a), 47 (4a); Casco 46 (18a); Castaneda 8948 (17); Cas- tellanos 43* (27), 25909 (30), s.n. (4a); Castillo 583* (30), s.n. (5); Castro 69 (18a); Catalán 92 (26); Chase 1473 (12), 1474 (12), 1479 (12), 2726 (5), 4986 (5), 5900 (15a), 6308 (9), 6547 (14), 6567 (9), 6577 (9), 6744 (14), 6856 (28), 7089 (20), 7090* (8), 7256 (15a), 7270* (8), 7480 (14), 7492 (31), 7558* (15b), 8105* (24), 8105bis (24), 8696 (27), 88021/2 (29), 8897*,** (29), 9089 (9), 9279 (27), 9300* (11), 10151 (4a), 10450 (29), 10534 (27), 10589 (27), 10665 (27), 10697 (27), 10723 (27), 10725 (11), 10780* e Eo ‘aa? r (11), 10064 (29), 10886* (27), * (29), 9 (29), 11071* (27), 11119 (11), Med E a eo 11465 (11), 11482 11560* (29), 11638 (11), 11641 (11), 11791* (11), 29), 12115 (29), 12389 == by 294 (15a), 9201 Asi 12097 peru Christ 19 (12); Chris- tian s.n. (22); Churchill 7 (22), 103 (28), s.n. (12); Clayton 4366 (27), y des (27): Coelho de phe 677* (24); Collins 1 (5); Colunga 44** (18b); Conzatti 3603 (20); Cook 78 (9); Coradin 3394 (24); — 20 ALA Me: MES (15a), 17274 ws Va 8 (6); C 3398 (18 Crespo s.n. (22), 17 (14), 7 00. i Crit 181 (ua) Qus 23483 (17. 23664 (17), 47699 (18a); Croft 240 (6); Crowe 545B (17); Crutchfield 4956 (31), 6068 (15b); Curran 250 (17), 261 (17); Curtiss 384 (9), 494* (9), 6867 (22); Cutler D'Arcy 3830 (12); Da Silva 1 (11), 7 (11), 124 (11), 137 (11), 144 (29); Dalby 44 (4a); Darrow 3419 (15a); Davidse 2510 (17), 2925 (19), 2950 (19), 4.314 (19), 9492 (14), 9550 (26), 9587 (26), 9657 (20), 9806 (20), 9912* (20), 9927* (20), 10057 (15a), 10618 (27), 10712 (27), 11145 (27), 18161 (4a), 18851 (17), 20246 (17), 20271 (17), 20528 (17), 20612 (17), 29687 (26), 31533* (19), 31553 (19), 31651 (1c), 31665* (18a), 31704 (1c), 32004 (17), 32401 (17), 32638* (14); Davis 61068 (24), s.n. (17), s.n. (28); Dawson 400 (4a), 424 (4a), 713 (22), P1702 (22), 3355 (22); de Granville 9138 (22); de la Sagra 10 (17); de la Sota 4257 (Aa); de Wet 4343 (30), 4354* (1a), 4732 la); Deann 42277* (12); Del Puerto 685 (4a), 1527 (27); Delascio 11585 (19); Demaree 27528 (28), 30200 (12), 33174 (5), 64207 (28); Díaz Luna 2044 (1c); Dodge 1 (12), 7 (12), 124 (12), 5610 (17), s.n. (12); Dodson 12506 (18a), 12514 (18a); Dombrowski 11096 (27); Dorantes 246bis (18a), 1152 (14), 1510 (1c); Dore 46-137 (5), 10544 (5), 12887 (5), s.n. (5); Drouet 1953 (27), Ducan 100 (16), 1438 (28), 4172 (28); Duefias 2 (19); Duke 303 (17), 10142 (17), 14117 (17); Duncan 13221 (12); Dunn 21855* (30); Duss 12 (14), 536 (9), 668 (14), 676 (9), 768 (17), 3181* (9), 3184 (14), 3917 (17), s.n.* (14); Dwyer 2781 E E em Y * (15a); Eggers 76** (9), 5315 (14), 5406 m 1458 n. que ON Pes (14); Eggert 58 (28), Eggleston 22396 (12); Eiten 211 6 27. 3956 (11 n "oi aD, 10579 (30); Ekman H20 (9), 755 (17), 995 (9), 1 9), 1442 (14), 1517 (9), 1987 Pe ¿ (17) 7), 7483 (9), 7511 (9), 8534 (17), 9667 (14), 9991 (18a), 10068 (?), 11732 (9), 12036 (14), 12686 (14), pr d 2 13390 (9), prodi. 16237 (9), 16312 (17), 17578 (17); Enriquez 347 (17); Eskuche 2538-5* 27); Evans s.n. (18a); Ezcurra s.n. Fabris 1080 (5); Farwell 893 (12); Fassett 25492 (14); Felger 86-339 (1c); Felippone s.n. (5); Fendler 2565 (11); Fernald 519 (5), 9259 (28), 19754 (5), 41415 (12); 17828 (5); Ferreyra 5929 (18a); Ferris 3175 (17), 5069 (30), 5070 (1c); Fueillet 4442 (22); Fiebrig 112 (29), 5095 (29), 5102 (29); Filgueiras 1060* (27), 1204 (22), 1332 (11), 2282 (11); Fisher 3313 (31), 35314* (30); Fishlock 206 es B 1053 (26), 1197 (26); Fontana F57-7 (27), 56 (27); Fortuna 2 (4a), s.n. (5); Fosberg 27768 (29); Ps se 283 (lc); Fowler 8d (18a); Freeman 53433** (28). Gail s.n. (5); Galli 172 (4a); Gallinal PE-1290 (Aa), PE- 2362 (4a), PE-4473 (7); Gándara 2 (18a); Ga 5); Gardner 1838 (24); Garnier 833 (18a); ; r 150 e 2477 (145 Gehrt peste 81 (4a); Gau 1 ( 09 (22); Gemtc hhujnicov 3 69 (2), s.n.* (27); Gentle 7120 an 8380 (17); Gentry bio Ape: 0967 a 8349 (15a), 14353*,** (la); Gere 44 (12 0 (18a); Giardelli 256 (4a), 616 (22); Gilly be (17); ae 37 (1 2145 (19); Glaziou 13338 (4a); Goeldi 111 (12. 128 mez 8243 (30), 824 e Goodding 192-45 (30), 243 (5), Gouin 6 (17), 21 04; e y 1187 (15a), 2409 (15a), 3192 (18a), 3748 E- ), 4 $ (Ba) 4510 (18a), 4550 (18a) * Juas eo = eo ER RR IO e ed gen 25, 12084 (1c), 12110* d. 12171 (18a), 12288 (26), 12322 (20), 12729* (26), 12757 (26); Greene 258 (1c), 13107 278 Annals of th Missouri Botanical Garden lc); Greenman 5047 (12), 5101 (12); Griffiths 28 (18a), 238 (15a), 399 (5), 684 (5), 5004 (15a), 6939 1/2 (18b), 6939 1/2 (25), 7194 (18b), 7317 (la), 7808 (15a); girs s (18a); Grinnell 1082a (lc); Grisol 36 (30); Guagli- ne 270 (4a), 360 (4a), 473 (4a), 595 (4a), 667 (4a); Guardado s.n. (20); Guízar 1850 (26); Gunther 5 (5). enke s.n.** (18a); Hage 1716 (24); Hahn 1232 (9), T (80): Hall 816** (15a), 9813 (15a); Halley 69 (6); Hamilton 230 (14); Hammel 11795 (17); Hanes s.n. (12); colas E t4 (30); Harper 184 (28), 3759* (12); j 14); Harrington 3361 (15a); Harris 40* (19), inc n (14), 12161 (14), 12164* (9); Harvey 950 (15b), 1186* (15b), 1378 (18a), 1516 (15a), 1578* (8), 1600 (18a), 1653 (18a), 6667 (26); Hassler 1942 (9), 10967 (29), 10986 (11), 11682 (29), 12099 (27), 12494 (29), 13022 (29); Hatch 4615 (6), 5054 (15a); Hatschbach 23576 (29), 43533 (27), 46144 (29); Hayden 8220 (5); Heller 701 (12), 13864 (5), s.n. (28); Henderson 68-852 (12); Henrich 139 (18a); Hermann 6996 (12); Hernández 240* (18b), N-1667 (20), N-1911 (15a); Hernández Xol- ocotzi X-2304 (8), 2468 (26), X-2775 (18a), X-2791 (lc), 2793 (8), X-3379 (17); Herrera Castro 84 (18b); Hers 4000 (30); Herter 544. (4a), s.n. (4a); Hess 1459 (14); Hicken 22 (4a); Hicks 813(12); Higgins 6990 (30); Hill s.n. (12); Hin- ton 166 (18a), 1062 (18a), 1424 (14), 1428 (14), 1996 (26), 2000 (26), 2010* (3), 4413 (14), 4442 (18a), 6337 (18a), 6423 (1a), 6686 (26), 9284 (18a); Hitchcock 7 (128), 165 (28), 220 (15b), 2798 (5), 3481 (18a), 3494 (1c), 3526 (18b), 3541 (le), 3547 (le), 3553 (la), . 3604 (18a), 3631 (18a), 3675 (lc), 3706 (15a); (15b), 5197 (5), 5293 (15a), 5318 (15b), 5349 (6), 5. (15b), 5395 (15b), 5537 (15a), 5547* (6), 5605 (15a), 5712* (8), 5719 1/2 (15a), 5756 (20), 5822 (8), 5864 8 ! 5958 (20), 6057 (8), 6063 (20), 6143 (14), 6363 (14), 6391 (14), 6418 (14), 6782 (26), 6808 (26), 6810 (26), 6811 (26), 6814 (18a), 7031 (17), 7081 (26), 7089* (18a 7117 (20), 7199 (20), 7238 (20), 7424. (18a), 7510 (20), 7591 (20), 7751 (25), 7957 (17), 7964 (17), 7997 (14 8014 (14), 8065 (14), 8094 (14), 8382 (17), 8397 (17), 8435 (17), 8452 (14), 8517 (18a), 8601 (14), 8687* (14), 8755 (14), 8794 (1c), 8924 (18a), 8993 (14), 9014 (20), 9463 (9), 9927 (19), 13255 (15a), 13333 (lc), 13411 (15a), 13445 (15a) 13490 (15a), 13517 (15a), 13539 (15a), 13597 (15a), 16121 (28), 16219 (15a), 16382 (14), 20580 (17), 20601 (17), 22115* (11), 23239 (9), 23292 (14), 23396 (17), 23397 (14), 23470 (5), 35411/2 (1b), s.n. Amer. Gr. Natl. Herb 21 (12), s.n. (15a), s.n. (16), s.n. Amer. Gr. Hb. 36 (1 Amer. Gr. Nat. Herb. 28 (18a), —. x * (1c); Holm 286 (18a); Hosseus 43 (4a); Howard 8357 (17); Howell 8728 (1C), 8590 (1c), 8827 (1c), 9076 (1c), 9965 (1c), 9757 (1c); Humboldt s.n.** (8); Hunter 63 (14); Hun- ziker 12 (5); Hutchings s.n. (15a); Hyacinth 960 (28). Ibarrola 2533 (4a), 2611 (4a), 2889 (27), 3481 (27); Iltis et al. 114 " da); Irwin 24519 (11); Isely 15 (5), 4431 (28), 4596 Jack 6295 (17); Jahn 197 (Aa); Janzen s.n. (18a); Ji- menéz 530** (14), 730 (14), ue 3620 wi 4309 (17); Tonus 467 (15a), 694 (15a), 1017* (15a), 1111* (15a), 4870 (15b), 7170* (15a), 7866 (18a), 7992*, ** (lc), 8175 ee 8250 (15a), 8250a (15a), 9104* (20), 10383 (15a), 871 (15a), 12292 (18a); Jones 24493 (la), s.n. (18a) j. s.n. (lc Kappel s.n. (7); Kay 9180 (16); Kearney 72 (12), 151 (28), 960 (12), 961 (28), 12868 (5); Kearney Jr. s.n. (5); Kellog s.n. (22); Killeen 823 (11), 856 (29), 891 (29), 1563 29), 1702 (30), 1711 (30), 1770 (29), 1828 (29), 1848 (29), 2397 (29), 2834 (11); Killip 4100 (14), 4326 (18a), 12174 (17), 37046 (4a); King 771 (30), 919 (30), 3791 (14), 4679 (1c), 4831 (26); Kneucker 730 (30); Kral 33114 (12), 51473 (12); Krapovickas 1251 (4a), 2586 (4a), 2906 4a), 6681 (Aa), 17861 (27), 25201* (4b), 34700* (30); Krotkov 8709 (12); Kuhlmann 2530 (11), s.n. (29). L'Herminier s.n. (17); Labat 1186 (8); Ladd 12336 (12), 13859 (12); Laferriere 2725 (18a); Lanfranchi 186 (5), 453 22), 481 (22); Langlassé 263 (30); Larez 605 (19); Las seigne 2812 (22); Laughlin 1113 (1c); Lawesson 3143 (lo) Le Sueur 032 (lc), . 170 1 173 (26); ae enworth & Hoogstrall 1335 (18a); Leavenworth & Leavenworth 1335 (1c); Leavenworth 199 (17), 442a (30), 5b), 966 (18a), 1371* (la). 1453 (30), 1839 (8), aspi s.n. (5); Legrand 1455 17); Lenaz s.n. (11); Lin 190 (9); 923 (9 ) 2532 (9), 2559 (9), 2674 (14), 3449 (9), 3747 (14), 3913 (17), 5681 (9), 6429 (14), 8834 (14), 9138 (14), 9314 (14), 9651 (9), 14181 (17), 15339 (14), s.n. (19); Leonard 2030 (12), 4901* (14), 7449 (9), 7745 (14), 8344 (14), 9417 (14), 9607 (14), 11576 (9), 11875 (9), 11900 (9), 12237 (9), 12590 (9), 13147 (9), 13609 (9), 13744 (14), 13828 (9), 14338 (14), 15090 (9), 15754 (14), 16057 (12); Leopold 109 (17), 176 (18a); Levy 1018 (17); Lewis 565 (4a), 629 a, 35344. (29); Lindhei- mer 1266 (15b); Lindmann A-1187 (27); pd 21371 (22); Lix 738 (5); Lopez Figueiras 341 (14); Lopez Mirand 951 (18a); López Q. 26 (18a); Lot 1948 (14); rasa 17 (14), 188 (17); Lundell 1335 (17), 7427 (14), 7622 (17), 7791 (14), 8719 (6), 10657 (15b), 14683 (6), 15918 (17); Lyon- net 2639 (26); Lyons Jr. 59 (lc), 83 (1c). Martínez PRA 2 (4a), 8978 po Macedo 2192 (11), 2194* (11); Machado de Campos 209 (27); Macnab 184 (9); Macoun 26330 (12), 26331 Tex 26332 (12); szek 271 (5); Magafia 427 (17); Mally s.n. (6); Mali me 3044 (11); Malvárez 1350 (4a); Mann 60 (15a); Marsh 1278 (15a); Martínez 1313 (la), 6077 (4a), 7860 (17), 8272 (17), 8371 (17), 14613 (17); Martínez Calderon 568 17), 1798 (17); Martinez F-1709 (15a), F-2275 (15a), F- 2303 (15a), F2333* (31), F2364 (31); Matuda 3577** 17), 3640 (18a), s.n. (14); McGregor 840 (la), 870* (31); McBride 332 (5), 3526 (18a); McCarthy s.n. (5); McMurry 978 (16); McVaugh 5164 (28), 16326 (18a), 16819 (20), 17705 (20), 19117 (18a), 1929] (la), respi (18a), 19651 Vidi 2 26); isis e T), 4 17); — — — ~ eS pom Ko oe, — 5511 (22); Mexta 2740 (20), 5543a (29), 5622* (11); Mey- er 33 (4a), 6329 (27); Mick 44 (15a); Miers 3429 (4a); Miller s.n. (12); Moldenke 1372a (5), 6377 (5); Molina 242 (14), 2219 (17), 3905 (14), 15281 (17), 15426 (17); Mon- taldo 330 e (19); Montes 2018 (19), 3478 (4b), 11238 (27), 14642 (4a); Monticelli 168 (22); Moore Jr. 3924 (15a), 4218 (20); Moraes 677 (24); Moran 13664 (lc), 16483 (5), 18878 1/2 (1c); Morel 7564 (4a), 8656 (4a); Moreno 1181A 18a), 9838 (1c), 21790 (18a), 21833 (18a); Mori 243 (19); Morley 696 (19); Mosen 4469 (29); Muelhenbach 667 (22); Mueller 404 (30); Mulford 1078 (18a); Muller 987 (19), s.n. (19); Mufioz 1006 (5); Mutis 2151 (17); Myers 4189 (17). Nash — T — — — a 1297 (14); Navarro s.n. (22); Nealley s.n. (16), s.n. (le); Nease 71 (28); Nee 183 (20), 6783 (17), 18347 Volume 83, Number 2 1996 Zuloaga & Morrone 279 Panicum Subg. Panicum Secc. Panicum (17), 34307 (30), 37618 (4a); Neill 1115 (19), 2454 (18a), 910 (17); Nelson 5621 (19), 6355 (18b), s.n. (17); Nicora 1176 (4a), 2577 (4a), 2724 (4a), 2900 (4a), 6511 (5), 6806 (5), 8361 (22), 9737 (30), 18478 (30); Noblick 3728 (30), 955 (24); Norton 339 (28); Novara 8436 (22); Noyes 459 (28). O'Donell 2629 (30); Ochoa 1584 (18a); Orcutt 4197 (18a), 4687 (26); Orihuela 92 (7), s.n. (7); Ornelas U. 1167 (18a); Oropeza s.n. (17); Ortíz 1049 (17). Paget 152 (5); Palacios 1856 (30), 4368 E Pipe Ic (18a), 143 (18a), 168a (30), 206*,** (30), 208 (la), 249 (la), 251 (la), 270 (30), 346 (la), 525* ad 533 (20), 554. (15a), 750 (1c), 802 (9), 1338 (15a), 1538 (30), 1545 (le), 1554 (lc), bud (26), 13277 (15a), 14147 (15a), 14214 (15b), s.n. (1c); Parodi 590 (5), 607 (22), 1750 (5), 4490 (Ab), "512 (27), 4513 (27), 6935 (27), 8218 (5), 12363 y ae 988 (15a), 1048 (15a), ; Pedersen 1882 (27), 3748 (4a), 5428 (27), 5457 an ge (27), 5812 (27), 11109 (27); Peebles 7417 (lc), 10366 (18b); Penn s.n. (15a); Pennell 3849 (17), 3866 (17), 20033 (3), 20221 (30); Perez 8 (4a); Peterson 6729 (17); Pfister 60 (5); y 1393* (24), 1403 (4a), 3799 (11), 5154 (27), 5795 (27); Pinheiro 36 (11), 55 (11), 74 (11), 182 (11), 218 (11), 265 (11), 273 (11), 297 (11), 336 (11), 362 (11), 382 (11), 385 (11), 483 (11), 530 (11), 561 (11), 512 (11), 581 e Pee a 596 (11), 603 oe 634 (11); jd to 121 a ), 3 i E s.n. woe ttier ee er 11291 (19), 11437 (26), 11872 (18a), 12177** (18a), 0 (14), Pohl 12228 (18a), 12369 (18a), 12674 (18a), 12707 (26), 13609* (18a), 13665 (18a), 13703 (19), 14172 (19); Pollard s.n. (16); Popenoe 34 (17); Porter s.n. (12), s.n. (5); Pott 444 (11), 3862 (11), 3915 (11); Hy 197 (lo) 376* (15a), 464 (lc), 497*,** (20), 5573* (17), 13940* (lc), s.n. (30); Proctor 2571 (30); Purpus 6212 (18a). Quarín 110 (4a), 1321 (22), 1639 (4a), 1969 (27), 2866 (27), 2909 (27); Questel 1439 (14). Ragonese s.n. (5); Rambo 1937 (22), 9845 (4b), 9875* (7), 43391 (27), 43692 (27); Ramia 3330 (17), 4444 (30), 4452 (30), 4831 (19), 5345* (30), 7. 7201 (19), 8491 (30); Ramírez 51 mtrez Reyes 1715 (19), 1773 (19), 2048 (14), 2051 “17, 2054 (17), 2428 (14); Reed 517* (14); Reeder 1936 (14), 1956 (18a), 1962 (17), 2082 (1c), 2152 (26), 2164 (26), 2288 (26), 2300 (20), 2430 (1c), 3517 (18a), 3814 (20), 5520 (25), 6098 (1c), 7042 (8), 7045 (la), 7425 (18a), 8630 P13125 (15a); Reitz 17008 (22); Renvoize 3583 (Aa), . (27), 3640 (4a), 3912 (30), 3933 (27), 4044 (27); Raver: chon 1842 (28), 2230 (15b), 2490 (15a), 2840 (16), 2841 (15a), 3527 (15a); Reyes 34 (17); Richard s.n. (9); Rick- etson 677 (22); Ridoult 1919 (18a); Riedel 2155 (27), s.n. (29); Robbin 860 (5); Robleto 980 (18a), 1269 (18a), 1535 (18a); Rodríguez 16 (19), 99 (4a), 541 (19), 694 (14), 1234 (19), 1981** (19), 3056 (14), 3220 (14), 3126 (17), 3512 (22), 27191 (14); Rodway s.n. (17); Rogers 4491 ( i 4756 (15a), 6119 (15a), 6609 (6); Role: 1497 (18a), 1648 (1c), 2507 (19); Roig s.n. (1c); Rojas 105** (Aa), 2358 (4a), 2745 (29), 2762 da) 2784 (27), 5476a (30), 6654 (29), 11027 (30), 11047 (30), 12741 (29), 12819 (29); Roland 34 (12); Rollins 5813 (15a); Romanczuk 402 (4a); Romero 190 (17); Rondón 2585 (29); Rose 1878 (30), 1883 (30), 1889* o. 3281 (30), 4410 (9), 9555 (20), 18052 (15b), 6903 42 - (Lc); 5 (25); Ruth 66 (12); Rydberg 796 (20), 1136 (15a), 2047% p^ 3775 (20), 4539 (15a), 4618 (6), 4931 (15a), 5174 (15a), 6547 (15a), 6869 (15a), 17367 (3), 17423 (3), 17449 (3), 32145 (20), 39002 (8), 49933 (20), 49884 (8), 50000 (15a). Sagástegui 10912 (18a); Sampaio 2971 (4a); Sánchez Vega 2338 (18a), 2363* (18a); Sánchez-Ken 328 (15), 430 ; Sanders 9297 (18a); Sandino 3038 (18a); Santana Michel 1752 (8), 1756 (20), 1883 (20); Saravia Toledo 1243 (30), 1301 (30), 1305 (30), 10412 (30); Schinini 8190 (27), 9293 (4a), 12601 (4a), 16368 (30), 17527 (4a), 19219 (4a); Schomburgk 259 (4a), 456 (4a), 656 (4a), 701 (4a); Schott 16 (1a); Schulz 1065 (Aa), 12138* (17), 12349 17); Schwarz 1718 (4a), 1975 (4a), 6484 (27), 6499 (27), 9492 (21), 9595 (7); Schwindt 1154 (29); Seiler 3825 (22); Sellow s.n. (27); Sendulsky 1434A (24); Seymour 2829 (18a), 6275 (18a), 6033 (18a), 6284 (18a), 16381 (28), 20900 (28), 21936 (28); Shafer 385 (9), 1361 (14), 1437 (9), 1512 (9); Sherff s.n. (18a); Shreve 9821 (15a), 9897 15a); Silva s.n. (7); Silveus 625 (17), 3461 (30), 7293 (14), 7302 (14); Simpson 829* (30); Sintenis 4983 (14); Skorepa 116* (1a); Small s.n. (12); Smart 112 (17); Smith 164 (17), 165 (14); 2152 (19), 4120* (28), 14726 (27), s.n. (15a); 2 iy 492 (1c), 565 (1c), 548 (1c), 612 (1c), 648 (1c), 22 (1c); Soares 267 (7); Soderstrom 160 (18b), 363 (15a), p (8), 531 (15a), 647 (1c), 812 (20), 819 (15a), 896 (1c), 1045 (17); Sohns 210 (8), 253 (20), 285 (20), 288 (20), 504 (20), 791 (14), 797 (14), 834 (26), 836 (26), 905 (26), 922 (26), 945 (26), 960 (26), 1066 (20), 1094 (15a), 1114 (20), 1215 (15a), 1344 (15a), 1430 (6), 1431 (31); Solomon 2741 (15a), 3658 (5); Soto Núñez 4118 (18a); Spegazzini s.n. (22), s.n. (5); Standley 1767 (14), 3983 (15a), 6441 (15a), 9339 (12), 11797 (14), 12793 (19), 14492 (19), 24087 (26), 24654 (19), 24771 (18a), 25277 (18a), 27407 (18a), 27960 (18a), 29252 (19), 29357 (14), 53428 (17), 53692 (17), 55711 (17), 74383 (18a), 74761 26), 75102a (26), 75925 (26), s.n. (18a); Stanford 2338 (15a); Steele 16 (28); Steere 1632 (14); Steibel 2451 (4a), 3473 (5); Steinbach 2204 (29), 5299 (29), 5639 (29), 6979 (29), 7076bis (29), 7086bis (11); Steller 335 (5); Stergios 6935 (4a); Stevens 477 (12), 962 (16), 2672 (18a), 3750 (18a), 7982 (17), 10006 Ms m (14), 21647 (14), 23094 (18a), 23096 (26), 25813 (17); Stewart 1211 (15a), 1293 (lc), 2395 (15a); bra 2097 (28), 2554 (28). 15334 (28), 24250 (12), 25209 (28), 29053 (26), 29310 (26), 29326 (18a), 30064 (26), 39802 (17), 45998 (17), 47766 (18a), 47859 (18a), 49430 (17), 51627*,** (13), 72470 (5), 131327 (19); Stork 11361 (18a); Stuckert 14434 (4a), 14798 (Aa), 14799 (4a), 15511 (Aa); Suárez 4 (14); Suazer 1361 (14); Suskdorf 2330 (18a); Svenson 236 (18a), 9003 (28); Swallen 1079 (15b), 1467 (15b), 1506 (15b), 1555 (6), 1618 (31), 1642 (6), 1693* (6), 1716 (15b), 1717 (14), 1742 (15b), 1743 (15b), 1780 (6), 2381 (14), 2404 (14), 2621 (14), 3822 (11), 3844 (29), ` (11), 4106** (11), 4192 (30), 4396 (30), 4859 (11), (7), 8203 (27), 8311 (27), 8682 (27), 8771 (27), 8964 (27), 9097 (27), 9156* (27), 9308 (27), 9390 (29), 9427 (27), 10039 (6), 10102 (6), 10197 (6), 10242 (15b), 10254 (6), mm e — — — 280 Annals of the Missouri Botanical Garden 10290 (15b), 10300 (4a), 10314 (4a), 10765 (26), 11254 (26), 10838 (19), 10909 (26), 11147 (14), 11150 (14), 11250 (26), 11263 (14), 11264 (19), 11344 (19), 11371 (26). Tamayo 1000 (30), 1469 (11), 1768* (17), 3824 (17), 3949 (14), 4159 (17); Tellez 2630, 3414 (17); Tharp 6739 (4a), 7219 (4a), 43066 (6), 43119 (15a), 43120 (15a), 49042 (6), 49235 (6), s.n. (16); Tapes 56 (16); Thorn- ber 176 (1c), 193* (25), 208 (18a), s.n. (15a); Tirel 156 (27); Toledo T-2886* (18a); Tales 2626 (4a); Ton 6277 (17), 6220 (17); Tonduz 14586 (17); Torres 1288 (18a 3097 (14), 7860 (18a), 14027 (26); Torres Guzmán s.n. (14); Tovar 4585 (29); Tracy 7409 (16), 7941 (15a), 7945 (15a), 8200 (15a), 8295*, ** (16), 8908 (15b), 9018 (14), 9082* (9), 9111 (9), 9116 (14), 9338 (15a); Triana s.n. (17); Troiani 2827 (4a), 3902 (5), 4094 (5); Troncoso 2638 (4a); Trujillo 11399 (17), 11492 (17); Türpe 3009 (22); Tyson 1396 (18a), 6791 (17) Ugent vd (20), 5801 (20); Ule 6857 (4a), 8014 (3a): Umbach s.n. (12). Valdés m 125 (15a), 1230 (5), 1526 (15a), 1580 (15a), 1986 (15a), 2164 (15a); Valls 860 (7), 1645 (27), 1646 (7), 1929 (27), 2026 (7), 2033 (27), 2112 (7), 2769 (7), 7664 (29), 7693 (11), 8379 (11), 9867* (10), 11521* 21); Van Hermann 355 (9), 2444* (9); Vareschi 3745 (1c): 9); Vasquez 7025 (17); Vázquez 3592 (20); — = T Vargas 17225 (2€ Venturi 644. (4a), 2539 Vera Carletti 97 (17); Vera Santos 1832 (14). 1842 (18a), 1850 (18a), 1907 (14), 2052 (18a), 2256 (14): Vernier 1505 (26); Viereck 857* (6); Villamil 1863 (4a 1994. (Aa), 2331 (4a), 5889 (5); Villamizar-Jaramillo s.n. 4a), 2552 (4a), 7530 (4a), s.n. (4a): (22); Villarreal 6232 (15a), 6521 (15b), 6661* (18b); Virlet 06 (14). Wagner 4341* (20); Walkins 17 (18a), s.n. (18a); Wal- lace 292 (18a); Waller 1872 (15a), 1959 (4a), 1983 (15a), 1994 (15a), 2008 (15b), 2017 (15b), 2163 (15a), 2193 (15b), 2410 (4a), 2413 (15b), 2419 (15b), 2422 (4a), 2425 15a), 2434 (15a), 2620 (4a), 2643 (a). 29780 (15b); Wal- lies 13294 (15a); Walther 63 (1c); We n. (5); Warnock 20925 (16), s.n. (15a); eris 5:40 PE 13294 (15a); Weatherwax 79a (18a), 217 (18a); Weddell 2531 (29); Wendt 9788 (18a); Wharton 5151 (28); Wheeler et al. 122 lc); Wherry 14339 (28); White 3542* a 3658 (18a); reg 2090 (17); Whiting 608 (15a); Wiggins 127 (1b), 137 (18a), 4793 (5), 4833 (5), 6023 (18b), 7076 , 15135 (la), 15160*,** (la), 15179 15510 (1c), 18135 (1c); Wilcox 303 (1c); Williams 1010 (29), 5800 (29), 12152* (14), "I6751 (18a); Wilson 1405 (9), 9443 (9); Wingfield 5 5548 (17), 5913 (14), 6840 — Ww — — (30), 7361 (30); Woods 1 12); Woodward s.n. (28); Wooton 2014 (25), s.n. ae 1. (15a), s.n. (18a), s.n. (1c). s.n. * (6); Woronow 4691 (17 2. Wright 2084 (25), 3877 (9), 3852 (9), 3860 (9), 1540 (9); Wullschagel 621 (14), 622 (14); Wynd 174* (15a Zarazúa s.n. (30); Zelaya 2307 (14); Zuloaga 137 (: 412 (30), 457 (4a), 477 (4a), 514 (4a), 535 (4a), 547 ( 573 (4a), 924 (4a), 1889 (4a), 1890* (22), 232. 2515* (30), 2984* (30), 3060* (5), 3072* (4a), 3151 (4a), 3271* (Ab), 3273 (4a), 3447* (27), . 4300 (14). 4301* (14), 4340* (30), 4350* (30), 4363* (19), 4504* (19), 4506* (14), 4507* (17), 4508* (18a), 4722* (11). STATISTICAL SUMMARY OF SOME OF THE ACTIVITIES IN THE MISSOURI BOTANICAL GARDEN HERBARIUM, 1995 Vascular Bryophyte Total Acquisition of Specimens Staff Collections 20,162 2,145 22,307 Purchase 28,875 0 28,875 Exchange 28,773 3,006 31,779 Gifts 11,142 974 12,116 Total acquisitions 88,952 6,125 95,077 Mountings Newly mounted 56,158 12,534 68,692 Mounted when received 30,741* 0 30,741 Total specimens filed 86,899 12,534 99,433 Repairs Specimens repaired 29,948 n/a 29,948 Specimens stamped 3,973 n/a 3,973 Total repairs 33,921 0 33,921 Specimens sent On exchange 42,578 4,639 47,217 As gifts 23,083 783 23,866 Total 65,661 5,422 71,083 Loans sent Total transactions 642 54 696 Total specimens 40,439 9,534 49,973 To U.S. institutions Transactions 371 32 403 Specimens 25,491 4,989 30,480 To i institutions Tran ons 271 22 293 ee 14,948 4,545 19,493 To student investigators Transactions 73 8 81 Specimens 11,067 513 11,580 To professional investigators "Transactions 589 47 636 Specimens 29,372 9,021 38,393 Loans received Transactions 307 15 322 Specimens 22,713 1,877 24,590 * Of the 30,741 vascular plants “mounted when received,” 28,875 are specimens of Chinese plants purchased directly from China. From U.S. From abroad Total Visitors 336 85 421 n 31 December 1995 the total number of mounted, accessioned specimens in the herbarium was 4,591,434 (4, 316, 127 vascular plants and 275,307 bryophytes ). The 84,000 “mounted when received specimens” in the 1994 report should have been treated as purchases.— Marshall R. Crosby, Senior Botanist NOTICE THE 1995 Jesse M. GREENMAN AWARD The 1995 Jesse M. Greenman Award has been won by Lynn Bohs for her publication “Cyphoman- dra (Solanaceae),” Flora Neotropica. This study is based on a Ph.D. dissertation $ iw University under the di- ?. Schulte The bed eel a vom and a cash rection of Dr. prize of $1,000, is presented each year by the Mis- souri Botanical published as Monograph 63 of Garden. It recognizes the paper judged best in vascular plant or bryophyte system- atics based on a doctoral dissertation published during the previous year. Papers published during 1995 are now being accepted for the 28th annual award, which will be presented in the summer of 1996. Reprints of such papers should be sent to Dr. P. Mick Richardson, Greenman Award Com- mittee, Missouri Botanical Garden, P. O. Box 299, St. Louis, Missouri 63166-0299, U.S.A be considered for the 1995 award, reprints must be received by 1 June 1996. . In order to Volume 83, Number 2, pp. 153-282 of the ANNALS OF is MISSOURI BOTANICAL GARDEN was published on May 17, 282 Experimental and Molecular Approaches to Plant Biosystematics The proceedings of the Fifth International Symposium of the International Organization of Plant Biosystematists (IOPB Edited by Peter C. Hoch and A. G. Stephenson Twenty-three original contributions that span the breadth of biosystematics, a dynamic field of study that bridges the realms of systematics and population biology. The papers are arranged in four groups, reflecting the original four symposia of the 1992 meeting. DNA and Plant Biosystematics presents innovative work that uses the rapidly developing nucleic acid methods adapted from molecular biology. Plant Growth Patterns and Biosystematics includes comparative and developmental analyses of plant architecture and branching patterns. Plant Reproductive Strategies surveys new approaches in the anal- ysis of plant reproductive biology, an area central to both systematic and population-level studies. Phylogenetic Analysis and Population Biology emphasizes the application of the powerful new methods of phylogenetic analysis to problems at the species and population levels. Monographs in Systematic Botany from the Missouri Botanical Garden, Volume 53. ISBN: 0-915279-30-4. 416 pp. Illustrated. 1995. $60.00 U.S. $62.00 Non-U.S. Annals of the Missouri Botanical Garden, Volume 82, Number 2: Alternative Genes for Phylogenetic Reconstruction in Plants A symposium cosponsored by the American Society of Plant Taxonomists and the Botanical Sat America, organized by Pamela S. Soltis and Douglas E. Soltis, and presented at the 1993 AIBS meetings. Although the chloroplast gene rbcL has been successfully used to reconstruct plant phylogeny, many important questions of plant phylogeny and evolution cannot be addressed using it. this issue of the Annals explore the potential of eight alternative genes or DNA regions a phylogenetic reconstruction at a variety of hierarchical levels. Both nuclear and chloroplast genes are evaluated. Three regions of the nuclear ribosomal RNA cistron are explored: the 18S gene, the internal transcribed spacers (ITS), and the 26S gene. Small multigene families from the nuclear genome may also carry phylogenetic signal: the phytochrome gene family and the small heat shock gene family. Three genes from the c roplast genome are also considered: atpB, ndhF, and matK. Each paper describes the location, size, structure, and rate of evolution of the shean gene and discusses its potential for phylogenetic study. This issue also contains: “The Comparative Pollination and Floral Biology of Baobabs (Adansonia-Bombaca- ceae)” by David A. Baum and “In Memoriam: 1: Peter c. Martin." Annals 82(2) 1995. 174 pages. $27. 30 U.S. $28.00 Non-U.S. To order, please indicate method of payment below. Checks or money orders. should be in U.S. funds, payable through a U.S. bank, to Missouri Botanical Garden. Orders must be prepaid. $3. fee on all datu: Additional $2.00 handling fee will be added to orders requiring invoices. No shipments are made until perme is received. Phone: HE 577-9534, fax: Se. 577-9594, e-mail: dept] L@mobotorg Send. order to: - 3 ae Weise send copied of lead No. 53 artus Eleven, Mixes Botanical Calden ees Please send —_ — copyđies) of Annals sare 299 St. Louis, MO 63166-0299, U.S.A. O Check/money order enclosed ; | Send A to: L] Send invoice ($2.00 fee will be added to to total) TI ena; m Charge cn number. ASAS "pueda Name - : Expiration date — | z AA EA ee Name as it appears on Rer Telephone number ej) — Postal Code Country — Panic d SUBJECT TO CHANGE WirHOUT NOTICE E OS on 850) : T CONTENTS A Revision of the Fern Genus See rubeo. P Sea George Yatskievych Revisión de las Especies Americanas de Panicum Subgénero Panicum Sección Pan- icum (Poaceae: Panicoideae: Paniceae) ——— ——— m Fernando O. Zuloaga & Osvaldo Morrone Statistical Summary of Some of the Activities in the Missouri Botanical Garden Her- arum, 1005. 2 oo A Marshall R. Crosby Nared oeer o BANI GEOP E RUP PRETI E Cover illustration. Pavetta camerounensis subsp. camerounensis S. D. Manning, Linda Ellis sp. nov.. by bes ale PW E teca A I CS n ln Annals of the Missouri Botanical or. umber Volume 83, Number 3 Summer 1 contributed from the Missouri Botanical Garden, St. Louis. Papers originating out- side the Garden will also be accepted. Authors should write the Managing Editor - for information concerning arrangements for publishing in the ANNALS. acca to Authors are printed in the back of the last issue of each volume. Editorial Committee | Henk van BEE Werff Editor, |. Missouri eniin ads Amy Scheuler McPherson | Managing Editor, | Missouri Botanical Garden uia Caña Editorial Assistant, = | Missouri Botanical Garden ! Vicki TUNE : Se NA zu Teresa basan x oh soos e Publications Order Processor Annals of the 996 Missouri Botanical Garden The Annals, published quarterly, contains papers, primarily in systematic botany, Ihsan A. Al-Shehbaz Missouri Botanical Garden Gerrit Davidse Missouri Botanical Garden Roy E. Gereau Missouri Botanical Garden - Peter Goldblatt | Missouri Botanical Cordes m Condon McPherson : Msn Pob Garden — “Zo Mick Richardson) 3 : Missouri Botanical. ala peser pre dene s heuler@mobot.ong oi quere) dept (orders we ie hp Ave mobot.org | RR C : EC Missouri. Botanical Garden 1996 a ES For REN ae contact Depart- - | ment Eleven, P.O. Box 299, St. Louis, MO s 63166-0299. Su scription price is $110 per - S., $115 Canada and Mexico, $135 — tries. Four issues per volume. The riodicals postage paid at St. Louis, MO and ad- 2 i included in "ffe loser: k E Y The mission of the Missouri Botani PA E The ANNALS OF THE MISSOURI a f "GARDEN (ISSN 0026-6493) is published quar- — - terly by the Missouri Botanical Garden, 235 Tower Grove Averiue; St Louis, MO 63110. Pee | ditional mailing offices. POSTMASTER: _ dress changes to ANNALS OF THE Mis — | | c BOTANICAL GARDEN, Department Eleven, $ Box 299, St. seers ee SI their r environment, in order to preserve and enrich life. a; | z e This paper meets he requirements of ANSI/NISO 23948-1992 remanence o rope to discover and share knowledge aboot plas Volume 83 Number 3 1996 Missouri Botanical Garden A REVISION OF CYNANCHUM (ASCLEPIADACEAE) IN AFRICA! Sigrid Liede? ABSTRACT Based on the study of dried specimens and living material, a taxonomic revision of Cynanchum in mainland Africa is d denies — um xa jx La 31 species (about half the number found in Madagascar), 6 of them new. One mbination is propos ecies are neotypified. An artificial key to all taxa is presented; all species are fully described. pu taxa E nea illustrated are provided with illustrations. Names for African Cynanchum placed into synonymy under other genera are listed with their current placement. Since the works of Schlechter (1895) and Brown (1902-1903, 1908), the African species of Cynan- chum have not been revised. Asclepiadaceae have not been treated to date in most African Flora pro- jects, with exception of Bullock (1963) for West Af- rica and Huber (1967) for Namibia. Therefore, even regional treatments of Cynanchum are lacking for most of the continent. The present account is the first one covering all of mainland Africa. Vincetoxicum is understood as being separate from Cynanchum and is not considered here. There are only a very few species of Vincetoxicum in the extreme north and northeast of the African conti- nent. The separation of Vincetoxicum renders irrel- evant the question of whether Blyttia Arnold should be maintained for the treatment of Cynanchum, be- cause Blyttia could possibly be subsumed under Vincetoxicum, but certainly not under Cynanchum. A more detailed account on the Cynanchum/Vin- cetoxicum problem is in preparation (Liede, in press). To date, there is no valid infrageneric classifi- cation for Cynanchum. The present author recog- nizes only some East Asian members as sufficiently ! The continued support of my Asclepiad work by the Deutsche Forschungsgemeinschaft is gratefully acknowledged (grants LI 496/1-4). F. Weberling, G. K. Gottsberger (Ulm), and F. Albers (Miinster) provided working space and support of N. E. Newt E Apis Kenya, during fieldwork in East Africa deserves particula s and M. G. Gilbert, Missouri Botanical Garden, on an earlier draft of this manuscript are OD Steve comments from gratefully acknow gratitude to the directors of the ed. ? Abtlg. Spezielle Botanik (Biologie V), Universitat Ulm, Albert-Einstein-Allee 11, D-89069 Ulm, Germany. ANN. Missouni Bor. GARD. 83: 283—345. 1996. 284 Annals of the Missouri Botanical Garden distinct to warrant sectional status (sect. Rhodo- stegiella (Pobed.) Tsiang & P. T. Li). All African members of the genus belong to the typical section Cynanchum. In the course of a cladistic analysis, which will be published separately, it was found that the two presently recognized species of Pentarrhinum E. ey., P. abyssinicum Decne. and P. insipidum E. Mey., are closely related to some species of Cynan- chum, namely C. balense Liede, C. gonoloboides Schlechter, and C. somaliense (N. E. Br.) N Using morphological characters alone, however, ev- idence is not sufficient to either include Pentar- rhinum in Cynanchum or to transfer the species mentioned to Pentarrhinum. The thick-walled fol- licles with protuberances, shared in Africa, as far as known, only by these five species, might consti- tute a stronger indication of relationships than the highly variable corona morphology. However, there are "true" Cynanchum species with muricate fol- licles in East Asia (C. corymbosum Wight, C. mur- icatum (Blume) Boerlage). The close relationship between Cynanchum and Pentarrhinum has been taken into account in the construction of the key, which includes the two presently recognized spe- cies of Pentarrhinum. As it was found during the course of this study that Pentarrhinum species were frequently labeled “Cynanchum indet.” even in larger herbaria, this inclusion might also be of some practical value. MATERIALS AND METHODS A total of ca. 2000 herbarium specimens from BM, BOL, EA, G, GRA, BG P, PRE, SAM, SHRG, STEU, and UPS have been studied. Several collecting trips to Africa yielded living material of numerous species. Spirit material or restored material (heated to 65°C in Water: 95% Ethanol: Glycerine = 5: tergent per 200 ml) was examined under a binoc- 4:1, plus 3 drops of de- ular microscope, and, in most cases, under the SEM after Critical Point Drying. Data have been stored ELTA (version 3.06, Dallwitz. 1980; Dallwitz & Paine, 1986), and the descriptions were directly generated from this database. The key has been constructed manually using the INTKEY program for confirmation. Locality information has been de- rived from Polhill (1988) for East Africa and from various gazetteers for the remainder of Africa. De- limitation and spelling of subdivisions follows the guidelines for the major floras in the area; for East frica (Kenya, Tanzania, Uganda), the subdivisions of the Flora of East Africa according to Polhill (1988) have been used (KI-K7, T1-T8, Ul-U4, respectively). Throughout the paper, the corona ter- minology developed by Liede and Kunze (1993) has been used, in which Cs denotes the staminal parts of a gynostegial corona, Ci the interstaminal parts, and C(is) fused staminal and interstaminal parts. Indumentum terminology is following Hewson 88). Jul Cynanchum L., Sp. Pl. 212. 1753. LECTOTYPE SPECIES: Cynanchum acutum L., designated by E. Meyer, Comm. Pl. Afr. Austr. 216. 1838. Bunburia Harv., Gen. S. Afr. Pl. 416. 1838. TYPE: Bun- buria elliptica ~ Colostephanus Harv., Afr. Pl. 417. a M capensis ds rv. 1838. TYPE: Cyathella Decne., Ann. Sci. Nat. Bot., sér. 2, 9: 332. “Type. species not designated. PNEU E. Mey., Comm. Pl. Afr. Austr. 215. 1838, not. J. F. Gmel. (1791). Type species not designated. Frndotropis pe Gen. Pl. 591. 1838, not de (1825); nor 38). Type species not rae Flanagania Sc hltr., Bot. Jahrb. Syst. 18, Beibl. 1894. TYPE: Flanagania orangeana Schltr. Feratuihosienud Baill., Hist. Pl. 10: 247. 1890. Type spe- signated (Perianthostelma abyssinicum Baill. nomen nudum in se a s P). Sarcocyphula Harvey, Thes. cocyphula gerrardii = 45: 10. not des jd: 58. 1863. TYPE: Sar- Only have been considered. synonyms relevant for mainland Africa Plants commonly twining, more rarely erect or decumbent, leafless or leafy; with milky, white to yellow latex; glabrous or with an indumentum; in- dumentum always consisting of uniseriate, multi- cellular hairs. /nflorescences usually bostrychoid, with geminiflorous partial inflorescences, basally occasionally with one or two dichasial ramifica- tions; shortening of the rachis results in a sciadioi- dal inflorescence structure (inflorescence terminol- ogy following Weberling, 1989). Flowers 5-merous, small (not exceeding 1.5 cm diam., normally not exceeding 5 mm diam.); aestivation imbricate or Corolla glabrous or adaxially with sparse, multicellular, verrucose tri- contorted, always dextrorse. chomes. Corona present, of gynostegial origin, con- sisting of staminal (Cs) and interstaminal (Ci) parts fused for at least 1⁄4 of total corona length (referred to as C(is) throughout this paper—corona termi- nology following Liede & Kunze, 1993); Cs, Ci, or both differentiated in shape; each lobe of Cs with or without adaxial appendage. Anther wings con- sisting of inner and outer ridges separated by a bristle-filled cavity, outer ridge either in the same plane as the back of the anther or centrifugal to- ward the base. Pollinaria consisting of two pen- dulous pollinia (synapomorphy of the tribe Ascle- Volume 83, Number 3 1996 Liede Cynanchum in Africa 285 piadeae) and a well-developed translator apparatus. Stylar head with a conspicuous protru- sion at the upper end of the corpusculum, dividing the stylar head in a rather uniform lower and a variable upper part. Fruit of two follicles, usually one aborted, follicles normally obclavate, winged or wingless, pericarp mostly thin and smooth, rare- ly thicker and/or with protuberances (see discus- sion above). Seeds brown, elliptic to pyriform in outline, winged or wingless, smooth, sculptured or hairy, crowned with a coma of white hairs. Chro- mosome number, as far as known, 2n = 22. istry: all species studied were found to contain pregnane glycosides. em- ARTIFICIAL KEY TO THE SPECIES OF CYNANCHUM and PENTARRHINUM l. Plants twining, —— without well-developed leaves 2 1% Plants erect or twining, not succulent; with well- developed leaves 1.3 2(1) | Staminal corona parts edis gynostegium and, in young flowers, connivent over it... C. lenewtonii 2t Staminal corona parts not exceeding the gynostegium C. gerrardii 3(1). Plants erect, ibis Us. less than 30 cm high, sparsely branched to unbranched; leaves linear or at least three times longer than wide, um lobes with revolute margins; stylar head capitate at Combination of Qe otherw 4(3. Leaves absent at the time of rd leaf blades 2-8 mm wide; inflorescences with well-developed rachis; caudicles flattened. straig aecox 4’. Leaves present at the time of flowering: leaf di less than 2 mm wide; inflorescences without well- developed rachis; caudicles cylindrical, s-shap C. orangeanum 5(3). Corolla lobes adaxially with trichomes 6 95 Corolla glabrous 11 6(5) Corolla lobes only with a few basal smooth trichomes C. adalinae subsp. mannii 6”. Corolla lobes with + evenly spaced, verrucose trichomes 7 7(6). Staminal corona parts with adaxial appendages (ligules)............ 8 7% taminal corona parts without adaxial appendages (ligules) — 10 8(7). Stylar head elongated, much exceeding the gynostegium s C. umtalense 8 Stylar head not elongated, much shorter than the gynostegium 2220022202202 9(8). Leaves triangular; peduncles 2-10 mm long; inks ee n rachis 0—0.5 mm long; corolla lobes not contorted in bud . virens 9^ Leave ovate; peduncles 15-40 mm long; inflorescences with rachis 2-10 mm long; corolla lobes contorted C. abyssinicum 10(7) Plants less than 50 cm a manel branched; leaves fleshy, less than 15 mm long, elliptic; corolla lobes less than 2 mm long; Cs sed for more than % of corona length C. M 10’. Plants more Gan 2m high, richly branched; leaves herbaceous to coriaceous; 20—40 mm long, ovate; corolla lobes more than 3 mm long; Cs and Ci fused for less than !4 of total corona length... 26 sbrusifolium 11(5). Staminal corona parts fused only at base; stylar head flat or umbonate; follicles thick-walled, warty or with soft spines 11’. Corona fused for at least % of its length; stylar head of various shapes; follicles mostly thin-walled, smooth (except C. gonoloboides) 12(11). — of five fleshy, papillate, shining, yellow parts with appendages projecting toward the center of the Pentarrhinum insipidum 12’. ih rona not fleshy or papillate, without Spenge projecting toward the center of the flower 13 13(12). ribera stipitate; corona lobes prominently trifid C. somaliense '. Gynostegium sessile; corona lobes not trifid 14 14(13). Leaves leathery, venation prominent, slightly cordate at the base; corona lobes bifid almost to the base .... .. C. balense 14’. Leaves thin, venation not prominent, prominently cordate to lobate; corona lobes slipper-shaped Pentarrhinum abyssinicum 15(11). Staminal corona with prominent adaxial appendages (ligules) 16 5'. Staminal corona without prominent adaxial appendages... 17 16(15). Leaves at least aceite with indumentum; inflorescences 20-35-flowered, rachis 15-60 mm long; pe- uncles 3-10 cm; corona not exceeding the gynostegium, interstaminal corona lobes shorter than staminal corona lobes; stylar head conical C. acutum 16'. Leaves glabrous; inflorescences 1-12-flowered, rachis absent; peduncles ca. 1 cm; corona exceeding the gynostegium, staminal and interstaminal corona lobes of the same length; stylar head strongly bifurcate ... C mossambicense 17(15). Leaves clavate and longly beakec hastate; corona papillose; stylar head elongate-conical to obinfundibuliform; follicles strongly ob- ked I. edi s not distinctly hastate; if triangular-deltate, then corona not papillose and stylar head not elongate- onic: aal to obin pime follicles not so strongly eem vate and less diae e beake . Stylar he ad well ex . cou fused for more than % of its length eeding the corona; floral buds 6-6 Stylar E not or eem exceeding the corona; floral ba 3—4 mm long... C. andai subsp. ae~ avidens subsp. clavidens 286 Annals of the Missouri Botanical Garden 19." | Corona fused for less than % of its length (strongly folded coronas appear more fused than they are, especially in C. adalinae, a check nod. uuu ost trece DENN 20(19). Mat ure leaves more than 5 cm long m VM EET 21 20". ature leaves up to 5 cm dune ————————————"— 22 21(20). Leaves dark green, with nerves raised abaxially; corona white, not obscu ‘uring the long- eae. gynoste- gium; follicles thick-walled, warty |... gonoloboides 21’. Leaves lighter green, nerves not raised; corona purplish red, obscuring ; the sessile Eae follicles thin-walled, smooth C. longipes 22(20). Leaf blades about three times as long as wide, ovate, conspic ‘uously paler below; gynostegium shock: stipitate; stylar head umbonate, lower part of stylar head conspic uously bent upward in dried material ..... EEN EAE IN ERES C rungweense 22. Leaf blades about twice as long as wide (in C. falcatum very y rarely more than three times as long as wide but then triangular or falcate), the same green above and below; gynostegium almost sessile to long: stipitate; stylar head conical or flat, but lower part not conspicuously bent upward in dried material 3 23(22). Plants erect ES loaf blades less than 5 mm long, with margins thickened; corona apically papillose 10X); gynostegium sessile ee eee eee C. blyttioides 23’. Plants decumbent or twining leaf blades more than 5 mm long, margins not thic kened; corona smooth es gynostegium sessile or stipitate 000 24(23). Leaves s vius BN. B falcate; Cs and Ci differentiated with the triangular Gs higher p the bifid Ci; stylar Rt OI e uiuere metuenda e EN LUE edu Tenet iu e EE EE C. falcatum 24'. Leaves ovate; corona either annular or with only Cs differentiated; gy nostegium mostly on a bulge (sessile in C. ellipticum, ol sessile in some populations of C. altiscandens); stylar head not clavate. ES 25(24). d lobes conspicuously twisted in bud, more than 6 mm long: gynostegium long-stipitate (stipe > A A ree eee See ee C. africanum 25’. Corolla nee not a uously twisted in bud, less than 5 mm long; gynostegium sessile or stipitate, but tipe shorter than iia 26 26(25). Plants erect or ae corona fused for about Y of its length 0000 e ae et 26". lants twining; corona almost totally fused 0000000 28 27(26). Plants dwarf shrublets or decumbent, but never twining: leaves rounded or indistinc tly cordate at hase, less than 1 cm long; corona not adnate to the filaments 000000000 eyheri 215 Plants twining; leaves distinctly cordate at base, more than 3 em long; corona connate to the laments aes A A A A ed he Oe he an 28(26). dde of coastal habitats; leaves fleshy: upper margin of corona neatly five- crenate, anther wings shorter niher tea C. natalitium 28’. Plants not restricted to coastal habitats; tin not fleshy: upper margin of corona smooth or irregularly crenulate, anther wings as long as the anther 0000000000. 29(28). Planta glabrous, mature plants forming a pi nain stem, not sarmentose; pedunc les 12-20 mm long; upper corona ed irregularly crenulate; ee gium sessile (on a very short stipe in introgression forms with C. natalitium) MIA C. ellipticum 20”. Plants with a (check the main nerves of the leaves, peduncles, and pedicels), iras not forming a corky main stem; pedunc es 2-5 mm ag upper corona margin smooth; gy ios mostl on a short bulge (almost sessile in some collections) 0000 . altiscandens a Plants erect shrublets, less than 50 em high 31 Plants twining, more an Emi (0 4 ci ts iio do dat dia, -— 33 3130 Leaves ovate, 20 mm long, base rounded, margins neither thickened nor e :renulate n C.i meyeri Leaves triangular, more ai 25 mm long, base cordate to lobate, margins thickened and conspicuously E MAA A A AAEN 3 32(31). Corona fleshy, pink, Cs ‘oblong, « exc ceding the g gynostegium and Ci, Ci with conspicuously ns margins anthers not massive 0000000000000 irc oronae 32'. Corona not fleshy, white, Cs extended into a long, reflexed tip, margins of Ci diode m massiv —————————————H— CREE C. cras ow T 33(30). Corona exc seeding the gy nostegium and ee nes ‘uring it; Cs forming fi five conspic uous folds; stylar head conical; coma a the seeds attached alon t^ of seed length 33’. po na as long as the gynostegium or peas at not obscuring it, Cs not t forming folds; stylar head not onical; coma ‘of ey attached terminally 0000000000000 5 34(33). Tüfloresc 'ences sessile or very shortly pedunculate (less than 1 em) E C. adalinae subsp. ande A Inflorescences long- n ulate (more than 1.5 em)... a C. adalinae subsp. mannii 35(33). Corona as high as the gynostegium, or slightly higher 36 35'. Corona den twice as high | s the gynostegium A 37 36(35). Leaves cordate, but not bue corona adnate to - a for more than % of total corona ón lobes of Cs eae inflexed, upper margin variably toothed 0 C. sc chistoglossum 30”. Leaves lobate; corona not adnate to the filaments, camina lobes oblong, erect, with involute margins PM A IA SIA C. ledermannii SER A lobes more than 5 mm, corona more than 5 mm long; stylar head capitate olyanthum ;orolla lobes less than 5 mm, corona less than 5 mm long; stylar head umbonate. E PA Volume 83, Number 3 1996 Liede Cynanchum in Africa 1. Cynanchum abyssinicum Decaisne in Can- dolle, Prodr. 8: 548. 1844. Vincetoxicum abys- sinicum (Decne.) Kuntze, Revis. Gen. Pl. 2: 424. 1891. TYPE: Abyssinica. Quartin-Dillon s.n. (holotype, G not seen). Figure 1. Cynanchum abyssinicum var. tomentosum Oliver, Trans. Linn. Soc. London, Bot. 2: 342. 1887. TYPE: Tan- zania. Kilimanjaro: Moshi, 1884, Johnston 177 (lec- e, designated here, r a holstii K. Giuni, Bot. Jahrb. Syst. 17: 135. 1893. Cynanchum holstii (K. Schum.) K. Schum Nat. Pflanzenfam. 4(2): 253. 1895. TYPE: Tanzania. Tanga: Mlalo, Usambara, Holst 507 (holotype, B pre- sumably destroyed; lectotype, designated here, K). Plants ascending, twining, 3-4 m high, richly and irregularly branched. Shoots herbaceous, sparsely covered with appressed trichomes 0.5—0.7 mm long, along two lines, basally woody, with blackish bark (fide protologue of Vincetoxicum hol- stii); internodes 2.2-12 cm long, 0.9-1.5 mm diam. “Stipules” absent. Leaves with petioles 15-25 mm idag, leaf blades herbaceous, 35-75 mm long, 10- 35 mm wide, ovate, basally cordate, with 1-3 col- leters in the adaxial sinus, apically acute to acu- minate, adaxially isolatedly covered with appressed trichomes 0.35—0.4 mm long, evenly distributed over the whole surface to glabrous, abaxially slight- ly papillose, veins and margins isolatedly covered with appressed trichomes 0.35—0.4 mm long, to gla- brous. Inflorescences bostrychoid, 15-20-flowered, 3-5 flowers open at a time; rachis 2-10 mm long. Peduncles 15-40 mm long, densely covered with appressed trichomes 0.1-0.5 mm long. Flowers sweetly scented (Ash 2094, Maas Gesteranus 5161); floral bracts 0.6—0.8 mm long, 0.1—0.2 mm wide at the base, triangular, with trichomes; pedicels 0.8— 1.2 mm long, sparsely covered with erect trichomes 0.3-0.35 mm long. Buds 4.5-5 mm long, 1.6-1.7 mm diam., conical, with contorted aestivation. Ca- lyx basally fused, ciliate, abaxial surface with tri- chomes; lobes 2.2-2.6 mm long, 0.8 mm wide, ovate, apically acute. Corolla rotate, basally fused; 7 mm long, abaxially and adaxially brownish purple, adaxially with verrucose trichomes 0.15— 0.2 mm long, evenly distributed over the whole sur- face; lobes 1-1.2 mm wide, decurved, oblong, api- T acute. Corona ivory, abaxially glabrous, 7-8 m high, exceeding the scuring it; C(is) cyathiform, consisting of Cs and Ci fused for less than half of total corona length, Cs and Ci differentiated, Ci shorter than Cs. adnate to the filaments, with adaxial appendages; lobes of Cs basally laminar, elongate-triangular, apically filamentous, flat to producing a convex fold, apically erect; appendages of Cs elongate-tri- gynostegium but not ob- Cs not angular, apically filamentous, erect. Lobes of Ci laminar, rectangular, bifid, or apiculate, producing a pronounced convex fold along the upper two- thirds of corona length, erect to reflexed, with straight, lacerate margins. Gynostegium 2 mm high, 1.6-1.8 mm diam., sessile. Stamens without free filaments; anthers about as high as broad, rectan- gular, abaxially planar; anther wings 1 mm long, parallel to each other, extending along the whole length of the anther; adjacent anther wings parallel, in the same plane as the anther. Connective ap- pendages 0.8-0.85 mm long, 0.8-0.85 mm wide, ovate, narrower than the stamen, slightly inflexed. Pollinarium: corpusculum mm long, ovoid; caudicles 0.2-0.25 mm long, flattened, straight, horizontal, trapezoid; pollinia laterally at- tached to the caudicles, 0.5-0.55 mm long, 0.2 0.3 mm wide, ovoid, ovate in cross section. Stylar head cream, 1.5-1.7 mm diam., 1.1-1.3 mm high, upper part 0.8-1 mm high, depressed-conical. Fol- licles one per flower, pendulous, 70-75 mm long, mm diam., obclavate, obtusely deltate in cross section, apically shortly beaked, irregularly keeled, dark brown, longitudinally grooved, glabrous. Seeds 5-5.5 mm long, 3-3.2 mm wide, pyriform, dark brown, seta and aseta side sculptured with longi- tudinal ridges, marginally with 0.6-0.8-mm-wide wing with entire margin; coma 25-30 mm long. (Description of follicles and seeds from the lecto- type of var. tomentosum, Johnston s.n.) Chromosome number unknown. Distribution and habitat. Africa: Eritrea, Ethi- opia (Arsi, Gonder, Ilubabor, Shewa, Welega), Ke- nya (K3, K4, K5, K7), Tanzania (T2, T3, T7), Ugan- da (U2), Zaïre; 1700-3000 m; forest margins, savanna, open scrub, often in secondary vegetation. Widespread, but infrequent. Figure Flowering time. All year, with peak between b June and October. Vernacular names. Maneriat (Kipsangali). Used as a relish (Perdue & Kibura 1 1341); leaves used to make tea as a tonic (Mat- thews 6354). Selected specimens examined. ERITREA. Dekemehare, 2000 m, Schimper s.n. (UPS). ETHIOPIA. Arsi: Chilalo . N slope of Mt. o, 2900 m, 20 Nov. 1971, Thulin 1522 (K, UPS). Gonder: Fenter, 12 1909, 50 (FT). Hlubabor: Kombolcha, 2050 m, 12 Chiovenda 1450 Dec. Wilde-Duyfjes 8843 (K, MO, UPS). KENYA. C a- chakos, Chyulu North, 1800 m, 22 Apr. 1938, Bally 7934 288 Annals of the Missouri Botanical Garden ICONES SELECTAE. lol 5: Tab. 69. NETOS D X 84 —— — a si PR cR A Figure 1. N \ CYNANCHUM abs ssinieum, Pre., o PEC. Prodr. vol E Pug ope. Cynanchum abyssinicum Decne. Illustration from De Lessert, Icones plantarum. Vol. 5, Tab. 69. Volume 83, Number 3 Liede 289 1996 Cynanchum in Africa REED) o 2 E GUN E J 40 Ss \ I— s ae 30 \ 30 e/a" o Hu m 20 o 0 | 10 A | 20 * =) Lal ` A 30 0 100 200 300 400 300 600 ES = / | y : q P rm Figure 2. Known distribution of Cynanchum abyssinicum (dots), C. galgalense (open circles), and C. umtalense (asterisks). (K); North Nyeri, Nyeri, 2000 m, 30 Jan. 1933, Napier 2495 (K); South paii Kiandongoro forest, Nyeri-Naivasha Y mi. from Kagumo bridge, 21 Aug. 1968, Matheng (MO). Nyanza: Pas, Tinderet Fo e 376 rest a Camp alley: Naivasha, S of Aba 3000 Polhill 433 (EA, K); Nakuru, Nyahuru Thompson Falls, 23 Aug. 1981, Gilbert 6345 (EA, K); Trans Nzoia, E of Mt. Elgon, 2830 m, 8 Jan. 1955, Irwin 180 (K) TANZANIA. Arusha: Oldeani Mt., 2000 m , 10 Feb. jon) St. Clair-Thompson 612 (K). Iringa: Mufindi, escarpment above Luisenga stream, 1600 m, 17 = 1984, Bridson & Lovett 533 (K, MO). Tanga: Marungu, 1600 m, July 1893, Volkens 641 (K). UGANDA. Kigezi, ce Valley, 2330 ar. 1960, Lind 2710 (K); Mbale, Buginyanya, Bug- K). ZAI ie- ishu, 2000 m, 1 Sep. 1932, Tho x ga- Kiva Mts., 1929, ea 7604 (K); entre Kibati et Mir- agongo, Mission au Parc National Albert, Jan. 1938, Lebrun 9391 (K). Comments. The degree of hairiness varies greatly between densely tomentose and almost gla- brous. As there is neither a sharp limit between glabrous and hairy forms nor a correlation with oth- er characters, the separation of variety tomentosum Oliver does not seem advisable. This variety has never been properly published, but its name is fre- uently found on specimens. The specimen select- ed as lectotype is marked as type by N. E. Brown, but was not cited as such in, e.g., Flora of Tropical Africa (Brown, 1902-1903). Cynanchum abyssinicum is most closely related to C. umtalense and C. virens, with which it shares the trichomes on the adaxial corolla surface and the ligulate corona. 2. Cynanchum acutum Linnaeus, Sp. Pl. 212. 1753. Vincetoxicum acutum (L.) Kuntze, Revis. Gen. Bot. 2: 424. 1891. Solenostemma acutum (L.) Wehmer, Pfl.-Stoffe, ed. 2, 2: 1004. 1931. TYPE: LINN 308/3 (holotype, LINN). Figure 3. 290 Annals of the Missouri Botanical Garden Figure 3. Cynanchum acutum L. 1, 7: Mashaly s.n.; 2-6: Hort. Bot. Münster s.n.—1. Habit with inflorescence and fruit. "2 Flower, two corolla lobes removed.—3. Staminal corona lobe with ligule, pre view.—4. Gynostegium and corona, one staminal corona lobe removed.—5. Pollinarium.—6. Stylar head.—7. Seed. Drawn by Jim Conrad. Volume 83, Number 3 1996 Liede Cynanchum in Africa Cynanchum excelsum Desf., Fl. Atlant. 1: 212. 1798. Vin- cetoxicum excelsum (Desf.) Kuntze, Revis. Gen. Pl. 2: 424. 1891. TYPE: Tunisia. Tozzer, Desfontaines 489 (holotype, a Cynanchum fissum Pomel, Nouv. Mat. Fl. Atl. 81. 1874. TYPE: penis ‘Berges du Chellif, Pomel s.n. (holo- type, AL not seen; isotype el monspeliacum L., Sp. Pl. 212. 1753. TYPE: LINN 308/6 lora LINN). Only synonyms relevant for Africa have been considered. Plants ascending, twining, richly branched. Shoots herbaceous, sparsely to densely covered with flexuous trichomes 0.5—0.7 nodes 6-15 cm long, 1.5-2 mm diam. “Stipules” absent. Leaves with petioles 15-50 mm long; leaf blades herbaceous, 45—60(-70) mm long, 25-45(- 70) mm wide, triangular, basally cordate to lobate, lobes 13-20 mm long, with 1-3 colleters in the adaxial sinus, apically obtuse, with flexuous tri- chomes .6 mm long, evenly distributed, abax- ially isolatedly covered with flexuous trichomes 1— 1.2 mm long, restricted to veins and margins. Inflorescences 20—35-flowered, 10—24 flowers open at a time, basally dichasial, apically bostrychoid, rachis 15-60 mm long. Peduncles 30-100 mm long, sparsely to densely covered with flexuous trichomes 0.8-1 mm long. Flowers with a sweetish carnation- like scent, very nectariferous; floral bracts 1.5-2 mm long, 0.4—0.6 mm wide at the base, ovate (— lanceolate), glabrous; pedicels 7-25 mm long, densely covered with flexuous trichomes 0.4—0.6 mm long. Buds 5-6 mm long, 2.5-3 mm diam., conical, with contorted aestivation. Calyx basally fused, abaxial surface with trichomes; lobes 2.2— 2.5 mm long, 1.1-1.3 mm wide, ovate, apically acute. Corolla rotate, basally fused, 5-8 mm long, abaxially rose, adaxially basally purple, apically rose; lobes 1-1.3 mm wide, twisted, patent to hor- izontal, oblong, apically obtuse. Corona pink, 2.7— 3 mm high, shorter than the gynostegium; C(is) consisting of Cs and Ci fused for about two-thirds of total corona length, Cs and Ci differentiated, Ci slightly shorter than Cs. Cs basally just adnate to the filaments, appressed to the back of the stamens, with adaxial appendages; lobes of Cs laminar, tri- angular, apically erect; appendages of Cs slightly longer than Cs, laminar, triangular, erect. Lobes of Ci laminar, deeply bifid, producing a cleft in front of the guide rails and a pronounced convex fold along the upper two-thirds of corona length, erect, with straight margins. Gynostegium 3.2-3.5 mm high, 2.5-2.7 mm diam., sessile. Stamens with free filaments 0.5—0.6 mm long; anthers about as high as broad, rectangular, abaxially planar. Anther mm long; inter- wings 1.2-1.5 mm long, parallel to each other, ex- tending along the whole length of the anther; ad- jacent anther wings parallel, in the same plane as the anther. Connective appendages 0.5-0.6 mm long, 0.5-0.6 mm wide, ovate, narrower than the stamen, amd inflexed. Pollinarium: corpusculum 0.3 m long; caudicles 0.1 mm long, flat- tened, i er horizontal, triangular; pollinia lat- erally attached to the caudicles, 0.35—0.4 mm long, 0.35—0.4 mm wide, globose, elliptical in cross sec- tion. Stylar head 1.4—1.5 mm diam., 1.1-1.2 mm high; upper part 0.7—0.75 mm high, depressed-con- ical. Follicles usually one per flower, pendulous, mm long, 7-8 mm di fusiform to narrowly oblong, round in cross section, apically strongly beaked, light brown, longitudinally grooved, glabrous. Seeds 6—6.5 mm long, 2.8-3 mm wide, ovate, light brown, seta and aseta side sculp- tured with longitudinal ridges, marginally with wing 0.3-0.4 mm wide with entire margin; coma 30-3 n — 22 (voucher: 1am., mm long. Chromosome number: ex hort. Münster s.n., MSUN). Distribution and habitat. Asia. Europe. Africa: Algeria, Egypt, Tunisia; lowlands, close to water. Very widespread and not rare, but uncommon in Africa. Figure 4 Flowering time. September to October. Additional specimens examined (Africa only). EGYPT. Kom Aushim, El Fayum, 15 July 1960, Boulos s. p de Bords de Nil a K uL Aim prés P Caire, 13 Oct. 1908, Burdet 481 (G), Culhares à Koubak, 20 m, 21 Oct. dl Burdet 482 (G); Damietta, Ezbit-El-Burg, 18 Oct. 1982, Mashaly s.n. (K); Aegypto superiori, 1837, Schimper 959 (L). TUNISIA. Gafsa. Feb. 1908, Pitard 434 (L). Comments. This is the lectotype species of Cy- nanchum, designated by E. Meyer (1838). This Eurasian taxon extends into Africa only at the rthern margins. Sometimes, infraspecific taxa are recognized under this widespread and variable spe- cies; none of them, however, is based on African material. Among African species, C. mossambicense is probably the closest relative. 3. Cynanchum adalinae (K. Schumann) K. Schumann in Engl. & Prantl, Nat. Pflanzen- fam. 4(2): 253. 1895. Vincetoxicum adalinae K. Schum., Bot. Jahrb. 17: 134. 1893. TYPE: Gabon. Ogowe, 10 Apr. 1881, Soyaux 277 (ho- lotype, B presumably destroyed; lectotype, designated here, K). Figure 5. Plants ascending, twining, 2.5-5 m high, richly and irregularly branched; rhizomatous; rhizomes 2— 3 mm diam. Shoots perennial, herbaceous, sparsely 292 Annals of the Missouri Botanical Garden E à 09 200 200 400 $00 s00 T AC E SAO! ] HA ] / LRL Figure 4. Known distribution of Cynanchum acutum (asterisks) in Africa, and of C. mossambicense (dots). covered with flexuous trichomes 0.2—0.25 mm long, along a single line, basally woody, with brownish bark; internodes 7-15 cm long, 1-1.5 mm diam. "Stipules" absent. Leaf blades herbaceous to papery, ovate to elliptic, apically acuminate, apiculus 6—10 mm long, adaxially and abaxially glabrous. /nflores- cences 15-25-flowered, 5-9 flowers open at a time, basally frequently with one or two bifurcations, api- cally bostrychoid. Flowers sweetly scented (Breteler 1287); floral bracts triangular, glabrous; pedicels 3— 6 mm long, densely covered with flexuous trichomes 0.15—0.2 mm long, along a single line. Buds conical, basally with imbricate, apically contorted aestiva- tion. Calyx rotate, basally fused, ciliate, lobes ovate, apically acute. Corolla rotate, basally fused, abaxi- ally and adaxially creamish green, lobes incurved to patent, oblong to lanceolate. Corona white, 2.5-3 mm high, exceeding the gynostegium, almost entirely obscuring it; C(is) consisting of Cs and Ci fused for Y to % of total corona length, Cs and Ci differenti- ated, Ci longer than Cs. Cs not adnate to the fila- ments, appressed to the back of the stamens, without adaxial appendages; lobes of Cs laminar, oblong, producing a pronounced convex fold, apically in- flexed, with straight margins. Lobes of Ci laminar, oblong (when flattened), producing a pronounced convex fold along the upper two-thirds of corona length resulting in a cucullate shape, erect, with straight margins. Gynostegium 1.6-1.8 mm high, 1.5-1.6 mm diam., aments, anthers about as high as broad, trapezoidal, sessile. Stamens without free fil- abaxially planar; anther wings 0.9-1.1 mm long, convergent, extending along the whole length of the anther, stamens forming a triangular basal arch; ad- jacent anther wings parallel, in the same plane as the anther. Connective appendages 0.6-0.8 mm long, 0.3-0.35 mm wide, ovate, narrower than the stamen, slightly inflexed. Pollinarium: corpusculum Volume 83, Number 3 Liede 293 1996 Cynanchum in Africa SELA Z 27 VAS Cynanchum adalinae (K. Schum.) K. Schum. 1, 7: subspecies adalinae (Morton s.n.). 1'—6': subspecies mannii (Scott Elliott) Bullock. 1'-5': m 1832; 6': Deighton 2522.—1, 1’. Internodes and inflorescences.—2'. Flower, two corolla lobes removed.—3'. Gynostegium and corona, partially removed.—4". Pollinarium.—5'. Stylar head.—96'. t.—7. Seed, seta side; sd de po of the coma, which is unique in the genus. Drawn by Jim Conrad. 294 Annals of the Missouri Botanical Garden 0.18—0.2 mm long; caudicles 0.12-0.15 mm long, cylindrical, straight, declinate, thickened at the in- sertion of the pollinium; pollinia apically attached to the caudicles, 0.14—0.16 mm long, 0.13-0.15 mm wide, globose, elliptical in cross section. Stylar head ite, 0.75—0.8 mm diam., 0.95-1 mm high; upper part 0.65—0.7 mm high, conical. Follicles one, oc- pendulous, 60-95 mm fusiform, round to obtusely white, casionally two per flower, long, 7-10 mm diam., deltate in cross section, apically strongly beaked, keeled, medium brown with green mottling, longi- tudinally grooved, glabrous. Seeds 8.5-9.5 mm long, 5.06 mm wide, ovate, light brown, seta and aseta side sculptured with longitudinal ridges, marginally with wing 0.8-1.2 mm wide, distally with irregular, strongly dentate margin; coma 25-30 mm long, at- tached to the seed along about one-third of its length. Comments. The affinities of Cynanchum adal- inae are still uncertain; the closest relative is most likely among C. altiscandens and its relatives. KEY TO THE SUBSPECIES: Inflorescences sessile or very shortly pedunculate (less than 1 cm C. adalinae subsp. adalinae Inflorescences long- pedunc ui (more than 1 C. adalinae An mannii — 3a. Cynanchum adalinae subsp. adalinae ee congolense De Wild., Ann. Mus. Congo, Sér. 5, 1: 190. 1903-1906. TYPE: Congo. Wangata, De- Pa re 644 (holotype, not found). Leaves with petioles 15-30 mm long; leaf blades 50-75 mm long, 30-50 mm wide, basally cordate to lobate, lobes 6-11 mm long, with 1-2 colleters in the adaxial sinus. /nflorescences sessile to very shortly pedunculate (peduncles to 6 mm long): ra- chis 2-4 mm long. Floral bracts 0.6—0.7 mm long. 0.7-0.8 mm wide at the base. Buds 2.5-3 mm long, 1.5-1.8 mm diam. Calyx lobes 0.8-1 mm long, 0.5-0.6 mm wide. Corolla 3.5—4 mm long, lobes 1.5-1.8 mm wide, apically acute. Chromosome — 22 (voucher: Meve 903, MSUN). number: 2n Distribution and habitat. | Africa: Cameroon, Congo, Fernando Po, Gabon, Ghana, Ivory Coast, Nigeria, Zaire; 6—650 m; edges and openings of secondary forest. Widespread. Figure Flowering time. December to September. Selected specimens examined. CAMEROON. Kribi, ca. 13 km on Ebolowa rd., 14 Nov. 1968, Bos 3277 (K, MO WAG): Bertoua, near ee mission, ca. 650 m, 25 Ape 1961, Breteler 1287 (K, WAG); Oveng, near village km from Sangmélima along rd. to Yaoundé, 20 Mar. 1062. Breteler 2654 (K, WAG); Victoria, July 1904, Pede 3 K); Mt. Cameroon, above Likombe, 900 m, 27 Feb. 1995, Meve 903 (MSUN). CONGO. Léfini, Région de Kindamba, environs de Meya, sur la piste d'Hamon, ca. 100 m, 4 Nov. 1963, Desecines 11307 (MPU). FERNANDO PO. 1861. Mann s.n. (K). GABON. Estuaire, Forest de la Mon- dah, rd. from Libreville to Santa Clara, 16 Sep. 1986, Breteler, Lemmens & Nzabi 7769 (WAG ‘ N’gounié, Waka, ca. 380 m, 24 Nov. 1984, Arends. Louis & De Wilde 440 WAG): Ogoué, about 15 km SSE of Pana, 19 Oct. 1983, Breteler 6983 (WAG). GHANA. Ashanti: Mampong, 8 Dec. 1953, Morton 75 (K). Central Region: Cape Coast, 6 m, 7 July 1959, Hall 1513 (K). IVORY COAST. 2 km a PEst he E 18 June 1963, Garnier & Bouaké 18 K); ca. s W of ew cee near Gunther Fuyts house, on $^ : rie c. 4 km SE of Louga, 13 June 1963, De Wilde 200 (K, WAG). NIGERIA. Benin, Sapoba Forest res á v on Pas rd., <). 70m mi. .E of Tudor. Lamborn 303 (K); Ikom/Obudu div., ca. 14 mi. SW o River reserve, 28 May 1946, e & Cuei 18918 (K). Ondo: Akure, Idanre, ca. 500 m, : 948, Brenan : Keay 8691 (K). ZATRE. Yangambi Platea, Italowe, 9 Sep. 1938, Louis 1265 (SHRG). _~ L “~ . e Wow ez mg Comments. The type of Cynanchum congolensis De Wild. has not been found. The fairly detailed description, however, leaves no doubt that the spec- imen described belongs to C. adalinae subsp. ad- alinae. The main differences noted are the slightly smaller flowers and the almost sagittate leaf bases. As both floral size and leaf shape tend to be slightly variable, there is no reason to maintain C. congo- lensis as distinct. 3b. Cynanchum adalinae subsp. mannii (Scott- Elliott) Bullock, Kew Bull. 17: 185. 1963. Vin- cetoxicum mannii Scott-Elliott, J. Linn. Soc., Bot. 30: 93. 1894. Cynanchum mannii (Scott- Elliott) N. E. Br. in Dyer, Fl. Trop. Afr. 4(1): 394. 1903. TYPE: Sierra Leone. Bagroo River, 1861, Mann s.n. = K). Cynoctonum acuminatum Benth. in Hook., Niger Fl. 453. ¿ynanchum acuminatum (Benth.) K. Schum. t. Pflanzenfam. 4(2): 253. 1895, : 111. 1820. TYPE: Sierra Leone. Don s.n. (M da BM not seen). Leaves with petioles 25—35 mm long; leaf blades 55—65 mm long, 20-30 mm wide, ovate to elliptic, basally cordate with 3-4 colleters in the adaxial sinus. Inflorescences with peduncles 1: m long; rachis 5-15 mm long. Floral bracts 0.5—1.2 mm long, 0.4-0.6 mm wide at the base. Buds 3-4 mm long, 1.5-1.8 mm diam. Calyx lobes 1.2-1.4 mm long, 0.6-0.7 mm wide. Corolla 3-4 mm long, adaxially basally sometimes with a few smooth tri- chomes; lobes 1-1.2 mm wide, apically obtuse. Distribution and habitat. Africa: Ivory Coast, Volume 83, Number 3 1996 Liede Cynanchum in Africa 295 Figure 6. Known distribution of Cynanchum adalinae subsp. adalinae (dots), subspecies mannii (open circles), and C. ledermannii (asterisks). The one traced locality of one of the lost types is indicated by a circle around the asterisk Liberia, Sierra Leone; 50-200 m; edges and open- ings of primary forest. Not as widespread and fre- quent as the typical subspecies, but not immedi- ately endangered. Figure Flowering time. February to October. Selected bip wr IVORY COAST. 25 km SW of Gué ar. 1962, E o 3738 (K, MPU, WAC). ' LIBERIA. Grand Bassa, 20 mi. u- chanan, 18 Jan. 1969, dupl 1914 (WAG); Grand Geel about 5 mi. S Tchien, on rd. to Sinoe, 18 Jan. Jansen 1244, 1245 WAC » Montserrado, road from ae Hills to Gbama, 120 km from Monrovia, 18 Oct. 1963, Van ipis 165 (K, WAG). SIERRA LEONE. Southern Prov- e: Vevehun, between Fwendu and Potoru, 11 Apr. 1929, Becher 1649 (K). Western Area: Havelock, Freetown, 19 June 1964, Morton 1368 (K, MO, WAG) Comments. Distinguished from the typical sub- species mainly by the much longer peduncle. Found in the western part of the distribution area of the species. Cynanchum dinklagei Schltr. ex Milbr. [Repert. Spec. Nov. Regni. Veg. 41: 264. 1937, nomen nu- dum] represents this taxon, judging from original material (Dinklage 3064, K). Cynanchum adalinae subsp. mannii is illustrat- ed in Adam (1975: 971, pl. 505). 4. Cynanchum africanum (Linnaeus) Hoffmann- segg, Verz. Pfl.-Kult. 54. 1824. Periploca af- ricana L., Sp. Pl.: 212. 1753. Vincetoxicum af- ricanum (L.) Kuntze, Revis. Gen. Pl. 2: 422. 1891. Periploca pallida Salisb., Prodr. Stirp. Chap. Allerton 148. 1796, nom. superfl. (ren- aming of P. africana L.). TYPE: H.S.C. 79 (lec- totype, designated by Wijnands (1983), BM not seen 296 Annals of the Missouri Botanical Garden iii Vu rl R. Br., Mem. Wern. Nat. Hist. Soc . 1: 46. 1 b ow is pL (R. Br. Mey.. inim PL. Afr. r. 216. 1838. Cynan- ahu africanum (L.) Hoffmanns. var. crassifolium (R. Br.) N. E. Br. in Dyer, Fl. Cap. 4(1): 749. 1908. Cynoctonum crassiflor 84 I TYPE: j (BM). The fact that “Cynanchum obtusifolium, Linn. Suppl. 169?" is cited in synonymy does not render Cynanchum crassifolium R. Br. superfluous because the n mark is an expression of doubt (ICBN 52.2, l, see Greuter et a , Cynoct tonum in num Decne. in Gandolla, Prodr. 8 I 1844. Vincetoxicum dregeanum (Decne.) Kun- e, Revis a. Pl. 2: 424. 1891, nom. superfl. when published substitute name for Cynanchum crassifol- i Cy la pom , Mem. Wern. Nat. Hist. Soc. I: 10 aa 2 (R. Br.) E. Me Comm. Pl. Afr. Austr.: 216. 1838. Vincetoxicum d losum (R. Br.) G. Nicholson, Ill. Dict. Gard. 4: 160 887. late crassifolium E. Mey. n pilosum (R. Br.) Decne. in Ca adela dub 8: Papi hum RED arh var. pilosum (R. EE Se hltr., ot. rb. Syst. 18, Beibl. 45: 10. 1894. nom. su- p Periploca africana L. in epee al Cynanchum p unb. ex Decne. in Candolle, d 44. TYPE: not cab eas 2 syn um intermedium N. E. Br. in Dyer, Fl. Cap. 4(1): TYPE: South Africa. Cape: prope Port Elizabeth, Oct. 1897, West 1924 (holotype, SAM). Plants twining, sparsely branched, 30-60 cm high, sarmentose with adventitious roots along the whole lower surface of the runner; subterranean or- gans rhizomatous in older plants, rhizome 5-10 mm diam. (e.g., Bayliss 6144, PRE). Shoots perennial, 50-100 cm long, 1-1.5 mm diam., herbaceous, gla- brous or sparsely to densely covered with trichomes 0.5—0.6 mm long; in old plants basally woody with brownish bark. “ mm long, 4-6 mm wide. Leaves with petioles 2-6 Stipules” ovate, almost round, 4-6 mm long, 2-3 colleters at the base of the leaves; leaf blades herbaceous to coriaceous, 20-30 mn long, 10-25 mm wide, ovate, basally cordate or rounded, apically mucronate, adaxially and abaxi- > ally glabrous, or isolatedly to sparsely covered with erect trichomes 0.4-0.5 mm long, evenly distrib- uted over the whole surface. Inflorescences scia- dioidal, 4—7-flowered, all flowers open at a time; peduncles 10-20 mm long, glabrous, or sparsely to 0.5-0.6 mm long. Flowers sweetly scented; floral bracts 1.5-2.5 densely covered with erect trichomes mm long, 0.5-0.7 mm wide at the base, triangular, glabrous, or with trichomes; pedicels 4-12 mm long, glabrous, or isolatedly to sparsely covered with erect trichomes 0.6-0.7 mm long. Buds 5-8 mm long, 2.5 mm diam., elongated-conical; aestivation basally imbricate, apically contorted. Calyx basally fused, abaxially with trichomes, lobes 2-2.7 mm long, 1-1.5 mm wide, triangular, apically acute. Corolla rotate, basally fused, 6-12 mm long, abaxially and adaxially purple to brown, glabrous abaxially occasionally with a few trichomes); lobes lt .8-2.2 mm wide, recurved, oblong, apically acute, twisted. Corona white, tubular, 6-10 mm high, ex- ceeding the gynostegium, partly obscuring it; C(is) consisting of Cs and Ci fused for more than % of total corona length, only Cs differentiated in shape. Cs without adaxial appendages; lobes of Cs flat, tri- angular to bifid (shallowly or more deeply, then giv- ing the impression of a differentiated. Ci), erect, with straight margins. Gynostegium 0.8-1.2 mm , 1-1.5 i den without free mm diam., atop a stipe, 2—4 mm long. filaments, anthers deltoid, abaxially convex; anther wings 0.3—0.4 mm long, parallel to each other, extending along the whole length of the anther; adjacent anther wings parallel, in the same plane as the anther. Connective ap- pendages 0.75—0.8 mm long, 0.45—0.5 mm wide, triangular, equaling the stamen in width, slightly 0.2-0.25 mm long; caudicles 0.18—0.2 mm long, flattened, con- inflexed. Pollinarium: corpusculum cavely recurved, triangular; pollinia apically at- tached to caudicles, 0.4—0.45 mm long, 0.15-0.18 mm wide, pyriform, elliptical in cross section. Sty- lar head 1-1.1 mm diam., 0.8-0.9 mm high; upper part 0.5-0.55 mm high, depressed-conical. Folli- cles usually one per flower, 45—60 mm long, 9-11 mm wide, obclavate, round in cross section, api- cally obtuse, medium brown, d mua yl slightly grooved; glabrous. Seeds 6—7 mr wide, pyriform, light brown, seta and aseta side tu- n long, berculate, marginally with indistinct wing 1-1.1 mm wide with entire margin; coma 25—30 mm long. Chromosome number: 2n = 22 (Liede 2548, Liede & Meve 642, MSUN). Distribution and habitat. Africa: South Africa (Cape Province); 0-200 m, very rarely to 700 m; dunes, flats to gentle slopes, mostly on sand. Comments. The late Onno Wijnands brought to my attention the fact that the lectotypification of this species (Linné 307.5, LINN) in Liede (1993) is invalid, because it is predated by his lectotypi- fication (Wijnands, 1983). urther details, illustration, distribution map, and citation of specimens are provided in Liede (1993) 5. Cynanchum altiscandens K. Schumann, Abh. Kónigl. Akad. Wiss. Berlin 64. 1894. TYPE: Tanzania. Tanga: Usambara, Kwa Msfuza Hochwald, Aug. 1893, Holst 9078 (holotype, B presumably destroyed; lectotype, designated here, K). Figure 7. Volume 83, Number 3 297 1996 Liede Cynanchum in Africa Figure 7. Cynanchum altiscandens K. Schum. 1-7: Drummond & Hemsley 2106.—1. Rhizome and roots.—2. Habit with inflorescence and fruit.—3. Flower.—4. Gynostegium and corona, partially removed.—5. Pollinarium.—®. Stylar head.—7. Seed, seta side. Drawn by Jim Conrad. 298 Annals of the Missouri Botanical Garden Cynanchum mensense Schweinf. ex K. Schum., Abh. Kón- igl. Akad. Wiss. Berlin 64. 1894. TYPE: Eritrea: Gheleb, 1850 m, 17 Apr. 1891, Schweinfurth 1505 (holotype, B M sumably destroyed: lectotype, des- ignated here, M). Plants ascending, twining, 3-5 m high, richly and irregularly branched, sarmentose, with runners 2-3 mm diam., adventitious roots developing along the whole lower surface of the runner. Shoots her- baceous, glabrous to sparsely covered with flexuous trichomes 0.5-0.6 mm long; internodes 35-75 cm long, 1-1.5 mm diam. “Stipules” widely ovate, 8— 12 mm long, 7-10 mm wide. Leaves with petioles 5-15 mm long, leaf blades herbaceous, 20-45 mm long, 12-28 mm wide, ovate, basally rounded with 1-3 colleters adaxially, apically acute or apically acuminate, apiculus 1-2 mm long, adaxially gla- brous to sparsely covered with appressed trichomes 6—0.75 mm long, evenly distributed over the whole surface, abaxially glabrous or sparsely cov- ered with appressed trichomes 0.5—0.6 mm long, concentrated on veins and margins. /nflorescences bostrychoid, 5-20-flowered, 5-12 flowers open at a time; rachis 1-5 mm long; peduncles 2-5 mm long, glabrous to densely covered with appressed tri- chomes 0.25-0.3 mm long. Flowers with floral bracts 0.6-1 mm long, 0.6-0.8 mm wide at the base, triangular, with trichomes; pedicels 5-8 mm long, glabrous to densely covered with flexuous tri- chomes 0.4—0.6 mm long. Buds 3—4 mm long, 1.5— 2 mm diam., conical, with imbricate aestivation. Calyx basally fused; abaxial surface with trichomes; lobes 1.2-1.6 mm long, 0.7-1 mm wide, ovate, api- cally acute. Corolla rotate, basally fused; 3.5—4.5 mm long, abaxially and adaxially yellowish green; lobes 0.8-1.2 mm wide, straight, patent, horizontal or declinate, oblong, apically acute. Corona white, tubular to urceolate, 3-3.5 mm high, exceeding the gynostegium, entirely obscuring it; C(is) consisting of Cs and Ci completely fused, only Ci differenti- ated. Cs not adnate to the filaments, without adaxial appendages. Lobes of Ci laminar, triangular (when flattened), along the upper third of corona length, apically re- producing a pronounced convex fold flexed, with straight margins. Gynostegium 1.7-1.8 on a bulge of 0.4-1.3 mm length. Stamens without free filaments, anthers broader than high, trapezoidal, abaxially planar. Anther wings 0.€ tending along the aleta length of the anther; ad- jacent anther wings parallel, in the same plane as mm high, 1.8-2 mm diam., 5 mm long, convergent, ex- the anther. Connective appendages 0.8-0.9 mm long, 0.45-0.5 mm wide, triangular, narrower than the stamen, erect. Pollinarium: corpusculum 0.25— 0.3 mm long. rhomboid; caudicles 0.15—0.2 mm long, flattened, straight, horizontal, trapezoidal; pol- linia subapically attached to the caudicles, 0.35— 0.4 mm long. 0.1—0.12 mm wide, ovoid to oblon- ud ovate in cross section. Stylar head 0.9-1 mm diam., 0.9-1 mm high; upper part 0.6—0.65 mm high, depressed-conical. Follicles one, occasionally two, per flower, pendulous, 55—65 mm long, 8-10 obclavate, obtusely deltate in cross sec- keeled, brown, longitudinally grooved, glabrous or with iso- lated indumentum. Seeds 5-5.5 mm long, 2.2-2.5 ovate, medium brown, seta and aseta side mm diam., tion, apically strongly beaked, medium mm wide, sculptured with longitudinal ridges, marginally with 0.4—0.6 mm wide wing with entire margin; coma 25-30 mm long. Chromosome number: 2n — 22 (voucher: Liede & Newton 2873, ULM Distribution and habitat. Africa: Eritrea, Ethi- opia (Harerge, Shewa, Sidamo), Kenya (K2, K3, K4, K6), Tanzania (T1, T2, T3), Uganda (U2, U3, U4, U7); 1000-2600 m; forest margins, thickets, road- side shrubbery. Widespread and frequent. Figure 8. Flowering time. Almost all year, with peak be- tween September and March. Vernacular names. | Sandab-Ngingichet (Kipsan- gali); Ngobito-Ol'dorobo, Sinande (Masai). Uses. robo use stems to sew up the ends of bee hives. Browsed by all domestic stock; Wando- Specimens seen. ERITREA. Oculé Cusai, Monte Me- taten, 2500 m, 12 Sep. 1902, ae 1483 (EA); Environs de Acrour, 1900 m, 8 Apr. 18¢ d en ul & Riva 1692 (K). ETHIOPIA. Hore ; the rd. to E ca. 1800 m, 5 Dec. 197 e Wit 7277 (WAG). jede i du rum 2600 m, 7 Dec. . Meyer po (K). : 13 km S of Aghere Mar- iam on new rd., Du m, us May 1976, Gilbert & Jefford 4316 (K). KENYA. Central: e, where stream cuts direct rd. to Kija through African Inland Mission, 1 Dec. 1963, Verdcourt 3814 (B, K); Nyeri, Aberdares, near forest station, 13 Jan 1922. Fries 695 (K). North Nyeri: uki, Sweet Wa- ters Ranch, 1750 m, 26 Dec. 1964, Gillett 16569 (K); Masai, Narok, Orengitok ca. 12 mi. from Narok on rd. to Olokurto, 2530 m, 17 e 1961, Glover, Gwynne & Sam- uel 1429 (K). Rift Valley: Nakuru, Nakuru National Park, 1740 m, 20 July 1975, Gillett 20851 (K); Naivasha, Crater Lake. 1830 m. Nov. 1958, Newbould 3634 (K); Trans Nzoia, Hoey's Bridge (Moi’s Bridge), 1960 m, Sep. 1971, Tweedie < M 1 (K); Ngorongoro, 2330 2 (K). Mara: Loliondo, 10 mi. W of Klein's Camp, ca. 2000 m, 11 Nov. 1953, Tanner 1821 (K, MO); Moshi, Kilimanjaro Siid, 1650 m, 15 Jan. 1934, Schlieben 4546 (K). Tanga: rd. to Lushoto town, 3 E 1986, Kisena 623 (K). UG clang Busoga, Usoga. 1894, Scott Elliott 7227 (K); E Ankole, Mitooma, Lukiri. 1700 m, 11 Jan. 1989, ror iin 2752 (MO); Volume 83, Number 3 Liede 299 Cynanchum in Africa E 30 o 200 \c00 eoo 000 16 = / j / / 2 | Ld sb PR Figure 8. Known distribution of Cynanchum altiscandens (dots), C. longipes (asterisks), and C. rungweense (open circles). Masaka, Kabula, 14 Mar. 1936, i-i aly 1328 (K); Mengo, near Mukono, Nov. rae Diimmer 1238 (BM, mi. W of eae trading 1974, I 2249 (K); Toro, N edge of Kihabule LFR, W of Katwe, Queen Elizabeth National Park, 1030 m, 22 Apr. 1969, Lock 69/85 (EA). Comments. Cynanchum altiscandens is closely related to C. rungweense, C. ellipticum, C. natali- tium, C. africanum, and C. zeyheri, a group of spe- cies characterized by highly fused coronas and, with exception of C. ellipticum, stipitate gynostegia and the formation of runners or rhizomes just under the soil surface. 6. Cynanchum balense Liede, sp. nov. TYPE: Ethiopia. Bale: Rira, 3260 m, 20 Dec. 1959, Mooney 8359 (holotype, K). Figure 9. Cynanchum gonoloboides affinis cum foliis neuraphyl- lis, fructibus incrassatis. Differt gynostegio sessili, parti- bus staminalibus interstaminalibusque coronae gynoste- gialis non nisi basaliter connatis; partibus staminalibus profunde bifidis. Plants ascending, twining, 6-7 m high, richly and irregularly branched. Shoots perennial, herba- ceous, isolatedly glabrescent with flexuous tri- chomes 0.4-0.5 mm long, basally woody with brownish bark; internodes 35—45 cm long, 3-3.5 mm diam. “Stipules” absent. Leaves with petioles 17—40 mm long; leaf blades coriaceous, 55-110 mm long, 35—60 mm wide, ovate, basally cordate with 9-11 colleters in the adaxial sinus; apically acuminate, apiculus 7-10 mm long, adaxially iso- latedly covered with flexuous trichomes 0.4-0.45 mm long, evenly distributed over the whole surface, abaxially sparsely covered with flexuous trichomes 0 mm long, concentrated on veins and mar- 300 Annals of the Missouri Botanical Garden GA ES EN eR oe w^ Y BR Pu e WA MO Figure 9. Cynanchum balense Liede. 1-6: Mooney 8359.—1. Habit with inflorescence.—2. Flower.—3. Gynoste- gium and corona, partially removed.—4. Pollinarium.—5. Stylar head.—6. Fruit. Drawn by Jim Conrad. Volume 83, Number 3 1996 Liede Cynanchum in Africa gins. Inflorescences basally dichasial, apically bos- trychoid, 12-20-flowered, 5-15 flowers open at a time; rachis 20-50 mm long; peduncles 40-50 mm long, densely covered with flexuous trichomes 0.5— 0.6 mm long. Flowers fragrant; floral bracts 5-6 mm long, 1-1.5 mm wide at the base, ovate, with tri- chomes; pedicels 15-20 mm long, densely covered with flexuous trichomes 0.5-0.6 mm long. Buds 4.5-5 mm long, 3.54 mm diam., ovoid, with im- bricate aestivation. Calyx entirely free, abaxial sur- face with trichomes; lobes 44.2 mm long, 1.8-2 m wide, ovate, apically acute to acuminate. Co- rolla rotate, basally fused; 4-5 mm long, abaxially and adaxially greenish purple; lobes 2-2.2 mm wide, straight, horizontal to declinate, oblong, api- cally acute to acuminate. Corona pink, 2.5-3 mm high, equaling the gynostegium in height, C(is) con- sisting of Cs and Ci only basally fused, only differentiated. Cs not adnate to the filaments, with- out adaxial appendages; lobes of Cs laminar, deeply bifid, apically erect, with straight, entire margins. Gynostegium 2-2.2 mm high, 1.4-1.6 mm diam., sessile. Anthers about as high as broad, trapezoidal, abaxially convex; anther wings 1.4—1.6 mm long, convergent, extending beyond the anther proper, consisting of distal and proximal ridge, with space between distal and proximal ridge papillose, prox- imal ridge curved; adjacent anther wings parallel, centrifugal. Connective appendages 0.7-0.8 mm long, 0.8-1 mm wide, ovate, equaling the stamen in width, strongly inflexed. Pollinarium: corpuscu- lum 0.4-0.45 mm long, ovoid; caudicles ca. 0.15 mm long, cylindrical, concavely recurved; pollinia apically attached to the caudicles, 0.48-0.5 mm long, 0.15-0.17 mm wide, clavate. Stylar head 1.2— 1.3 mm diam., 0.6-0.8 mm high; upper part 0.4 mm high, umbonate. Follicles one per flower, 90— 95 mm long, 12-15 mm diam., obclavate, keeled, apically strongly beaked, medium brown, thick- walled, longitudinally grooved, glabrous. Seeds and chromosome number unknown. Distribution and habitat. Africa: Ethiopia: Bale; 3260 m (at upper forest margin); patch of forest in pasture. Only known from the type collec- tion. Probably rare and endangered, even though Rira is included in the Mt. Bale National Park (fide M. G. Gilbert). Figure 10. Flowering time. December. Comments. Cynanchum balense is a sister spe- cies of C. gonoloboides, with which it shares the characteristic dark green leaves with pronounced nervature as well as similar size and shape of the fruit. 7. Cynanchum blyttioides Liede, sp. nov. TYPE: Somalia. Sanaag, above Geei Harre, W of Gar- do airstrip, 710-840 m, 7 Oct. 1980, Beckett 428 (holotype, EA; isotype, K). Figure 11. Fruticulus erectus, habitu Blyttia et Diplostigma simi- lis, sed partibus staminalibus interstaminalibusque coro- nae gynostegialis late connatis, abaxialibus papillosis dif- fert. Plants erect, nontwining, 50-75 cm high, richly basicaulously branched. Shoots woody, with grayish bark, glabrous. “Stipules” absent. Leaves subsessile; leaf blades 3-4 mm long, 2-3 mm wide, ovate, ba- sally rounded with 1-3 colleters adaxially, apically obtuse, marginally thickened, crenulate, adaxially and abaxially glabrous. Inflorescences sessile, scia- dioidal, 2—4-flowered, all flowers open at a time. Flowers with floral bracts 0.8-1 mm long, 0.4—0.6 mm wide at the base, triangular, glandular over the whole surface; pedicels 2.5-3 mm long, glabrous. Calyx basally fused; lobes 1.2-1.3 mm long, 0.5- 0.6 mm wide, triangular, apically acute. Corolla cy- athiform, basally fused, with imbricate aestivation, 2-2.5 mm long; lobes 0.8-0.9 mm wide, incurved, triangular, apically acute. Corona cyathiform, abax- ially papillose, 1.5-1.7 mm high, equaling the gy- nostegium in height; C(is) consisting of Cs and Ci fused for ?4 to % of total corona length, Cs and Ci differentiated, Ci longer than Cs (and about twice as broad; upper halves of Ci folded in over Cs), with straight margins. Cs not adnate to the fila- ments, appressed to the back of the stamens, with- out adaxial appendages; lobes of Cs laminar, ob- long, apically erect, with straight margins. Lobes of Ci laminar, ovate, producing a slight convex fold along the whole corona length, inflexed, with straight margins. Gynostegium 1—1.2 mm high, 0.9— 1 m diam., sessile. Stamens without free fila- ments, anthers about as high as broad, hexagonal, abaxially biconcave; anther wings 0.65—0.7 mm long, parallel to each other, not extending along the whole length of the anther; the anther forming a “pseudostipe” 0.3—0.4 mm high, adjacent anther wings parallel, basally centrifugal, forming a dis- tinct “mouth” with the basal lateral margin of the anther. Connective appendages 0.36—0.4 mm long, 0.5-0.55 mm wide, ovate, narrower than the sta- men, slightly inflexed. Pollinarium: corpusculum .15-0.17 mm long, margins of the corpuscular cleft basally widened; caudicles 0.09—0.1 mm long, flattened, convexly recurved, trapezoid; pollinia subapically attached to the caudicles, 0.2—0.25 mm long, 0.08-0.1 mm wide, ovoid. Stylar head 0.5— 0.6 mm diam., 0.25-0.35 mm high; upper part 0.2— 0.3 mm high, umbonate. Follicles one per flower, Annals of the Missouri Botanical Garden AAA Figure 10. pendulous, 35 mm long, 7 mm diam., obclavate, obtusely triangular in cross section, apically strong- ly beaked, wingless, light brown, with reddish brown mottling, smooth, glabrous. Seeds 6—6.5 mm long, 3.5-4 mm wide, ovate, light brown, seta and aseta side densely covered with regularly arranged trichomes 0.5-0.7 mm long, marginally with wing 0.3-0.5 mm wide with entire margin; coma 12-15 mm long. Chromosome number unknown. Distribution and habitat. Africa: Somalia (Sa- naag); 700-850 m; fragmented marine limestone hills, thin bush with poor grass cover. Rare and localized; but probably undercollected. Figure 12. October. Sod Keh. Flowering time. Vernacular name. Uses. medicinal. Fruits edible, nutty flavor; fruit and latex Known distribution of C. balense (star), C. gonoloboides (open circles), and C. somaliense (dots). Comments. The affinities of Cynanchum blyt- tioides are unclear. In habit, it resembles strongly the genus Blyttia, though corona structure places it in Cynanchum. Within Cynanchum, it resembles the two other new Somalian species, C. crassianth- erae and C. rubricoronae, in habit and the peculiar undulated leaves. The slightly papillose corona re- minds one of C. clavidens. The specimen in EA has been chosen as holo- type because the K specimen is rather poor. Ta. Cynanchum clavidens N. E. Br. subsp. clav- idens, Bull. Misc. Inform., Kew 106: 256. 1895. Cynanchum flavidens N. E. Br., Index Kewensis Suppl. 1: 121. 1906. spelling error for Cynanchum clavidens. TYPE: Somalia. Boobi, 5 Feb. 1933, James & Thrupp s.n. (ho- lotype, K). Figure 1: Volume 83, Number 3 Liede 303 Cynanchum in Africa Figure 11. Cynanchum blyttioides Liede. 1-6: Beckett 428.—1. Habit with leaves and old fruit.—2. Leaf.—3. Flower.—4. Gynostegium and corona, partially removed.—5. Pollinarium.—6. Stylar head. Drawn by G. Hintze. Annals of the Missouri Botanical Garden * | | D: 12. Known distribution of C. blyttioides (star), Cynanchum clavidens subsp. clavidens (open circles), sub- le species hastifolium (dots), C. crassiantherae (squares), and C. rubricoronae (triangle). Plants ascending to erect, twining, 1-2.5 m high; irregularly branched. Shoots probably deciduous (O’Brien 43), apically herbaceous, glabrescent to sparsely covered with flexuous trichomes 0.3-0.35 mm long, along a single line; basally woody, with yellowish to brownish bark; internodes 1-5.5 cm long, 1.5-2 mm diam. “Stipules” hastate, 7-10 mm long, 6-8 mm wide, apically obtuse. Leaves with petioles 6-14 mm long; leaf blades herbaceous, 10-75 mm long, 4-35 mm wide, hastate, basally cordate to lobate, lobes 4-7 mm long, with 1-2 colleters in the adaxial sinus, apically acute to acu- minate, adaxially isolatedly to sparsely covered with flexuous trichomes 0.3-0.4 mm long, concen- trated on veins and margins, abaxially papillose, glabrous to isolatedly covered with flexuous tri- chomes m long, concentrated on veins and margins. Inflorescences sciadioidal, sessile, 3— 6-flowered, all flowers open at a time. Flowers with floral bracts 1-1.5 mm long, 0.4—0.7 mm wide at the base, triangular, glabrous; pedicels 3-10 mm long, sparsely to densely covered with flexuous tri- chomes 0.4—0.5 mm long, along a single line. Buds 3.5—)6—6.5 mm long, 1.5-2 mm diam., elongated- conical to conical, with imbricate, apically contort- ed aestivation. Calyx basally fused, abaxial surface with trichomes; lobes 1.6-2 mm long, 0.7-1 mm wide, linear to triangular or ovate, apically acute. Corolla rotate, basally fused; (3-J5-6.5 mm long, abaxially creamish green, adaxially green; lobes 1— 1.2 mm wide, decurved, lanceolate, apically obtuse, with revolute margins. Corona tubular, white, abax- ially apically papillose, 2-3 mm high, equaling to exceeding the gynostegium in height (except for the appendage of the stylar head), but not obscuring it; C(is) consisting of Cs and Ci fused for about % of — Volume 83, Number 3 Liede 305 1996 Cynanchum in Africa Bi oo D xn d à a ES EZ Lud ig Ti 5A -—' —mÁ 0,5 mm 0,2 mm Figure 13. Cynanchum clavidens N. E. Br. 1-5: subspecies clavidens. 1'-7': subspecies hastifolium Br.) Liede.—1. Node with E 2d Brien ig plene leaf shape (Hucks 272).—1'. Leaf shapes; left: dm Es right: Gilbert et al. 7399.— ower.—3, 3'. tegium and corona (partially removed). | 4’. Pollinarium.— 5, 9’. Stylar head. 2-5: lad i Pins PA 2-3 "Gillet 13863. 6', 7': Gilbert et al. 7399.—6'. Fruit. —7'. Seed seta side. Drawn by Jim Conrad. 306 Annals of the Missouri Botanical Garden total corona length, Cs and Ci differentiated, Ci shorter than Cs. Cs not adnate to the filaments, not appressed to the back of the stamens, without ad- axial appendages: lobes of Cs laminar, ovate, pro- ducing a pronounced convex fold, apically erect. Lobes of Ci laminar, ovate to oblong, erect to re- flexed, with laterally involute margins. Gynostegium ; mm high (without appendage of stylar head), 1-1.2 mm diam., filaments 0.7—0.8 mm long; anthers about as high sessile. Stamens with free as broad, trapezoidal, abaxially planar to convex; anther wings 0.8-0.85 mm long, convergent, ex- tending along the whole length of the anther, ad- jacent anther wings parallel, centrifugal, basally forming a distinct “mouth” with the basal lateral margin of the anther. Connective appendages 0.6— 0.7 mm long, 0.5—0.6 mm wide, ovate, equaling the stamen in width, erect, with denticulate margins. Pollinarium: corpusculum 0.3-0.35 mm long; cau- dicles 0.35-0.4 mm long, cylindrical, s-shaped, convex-concave; pollinia subapically attached to the caudicles, 0.35-0.4 mm long, 0.1-0.12 mm wide, ovoid, ovate in cross section. white, 0.6-0.65 mm diam., 1-2.5 mm high; upper part 0.7-2.2 mm high, obinfundibuliform (the ap- Stylar head pendage exceeding the corona is the main diagnos- tic feature of the subspecies). Follicles normally one per flower, pendulous, 100-120 mm long, 15-20 mm diam., obclavate, round in cross section, api- cally strongly beaked, green with dark brown mot- tling, longitudinally grooved, glabrous. Seeds ca. 7.5 mm long, 4.5 mm wide, ovate, medium brown, seta and aseta side papillose with regularly ar- regularly ar- ranged trichomes 0.15-0.25 mm long, marginally ranged papillae, and with sparse, with wing 0.6—0.8 mm wide, with denticulate mar- gin; coma 25-30 mm long. Chromosome number unknown. Distribution and habitat. Africa: Ethiopia (Bale, Harerge), Kenya (Kl, K4, K7), Somalia (Bakool, Bay, Shabeellaha Dhexe), Tanzania (T5): 230-1400 bushland. Widespread, but localized. Figure 12. m; dry savanna, Acacia—Commiphora Flowering time. January to July. Use. Fruits edible. ETHIOPIA. Harerge a, 8 B Apr 1956, Scimono "m avo East National Park, Selected specime ns e: xamine - Mt. Scillavel Park, ' [ : ks 27 ). Coast: le, Mackinnon Rd., Tara ert, 400 m, Oct. 1965 ved 3199 (K). Northern Frontier: Moyale, 10 km from Moyale barrier, 10 Dec. 1993, Liede & a ton 3160 (ULM). SOMALIA. Bakool, Wajid (17 km E of Uegit on rd. to Oddur), 415 m, 22 May 1983, Gillett & Hemming 24352 (K); Bay, Bur Heybe, 230—375 m, 28 Apr. 1985, 0 dicis 43 (K); Shabeellaha Dhexe, Johwar, SW Morajiid. km S Bulo Caano, 17 July 1988, Kilian & Lobin 1762 2 4 ; and Goma. E of Kondoa, Polhill «e o 1239 (K). 1400 m, 20 Jan. 1962, 7b. Cynanchum clavidens N. E. Br. subsp. has- tifolium (N. E. Br.) Liede, comb. nov. et ega hastifolium N. , Bull. . Inf., Kew 1895: 257. Oct. D TYPE: Ethiopia. Tigray: near Djeladjeranne, 29 Aug. 1840, Schimper 1690 (holotype, K; isotype, P). Figure 13. Cynanc Wisi hastifolium K. Schum. in Engl. & Prantl, Nat Pflanzenfam. 4(2): 253. Oct. 1895. TYPE not loc nied dem e comments). Plants ascending, twining, 0.5—1 m high, sparse- ly and irregularly branched. Shoots basally woody, with yellowish brown bark, apically herbaceous, glabrescent to sparsely covered with flexuous tri- chomes 0.3-0.35 mm long, along a single line; in- 1.5—4.5 cm long. “Stipules” hastate, 7-10 mm long, 6-8 mm wide, apically obtuse. Leaves ternodes with petioles 6-25 mm long; leaf blades herba- 12-48 mm long, 8-25 mm wide, hastate, basally truncate to cordate with 2-3 colleters in the ceous, adaxial sinus, apically acute, adaxially isolatedly covered with flexuous trichomes 0.5—0.6 mm long, evenly distributed over the whole surface, abaxially papillose, glabrous to isolatedly covered with flex- uous trichomes 0.5—0.6 mm long, restricted to veins and margins. Inflorescences sciadioidal, sessile, 5— 16-flowered, all flowers open at a time. Flowers with floral bracts 0.8-1 mm long, 0.5-0.6 mm wide at the base, ovate; pedicels 6-8 mm long, densely covered with flexuous trichomes 0.3-0.4 mm long. Buds 3-4 mm long, 1.2-1.8 mm diam., with imbricate, apically contorted aestivation. Ca- conical, lyx basally fused; abaxial surface with trichomes; lobes 2-3 mm long, 0.4-0.6 mm wide, linear to lanceolate, apically acute. Corolla rotate, basally used; mm long, abaxially and adaxially green; lobes 0.8-1 mm wide, horizontal to decurved, lin- ear to lanceolate, apically acute, with revolute margins. Corona white, tubular to urceolate, abax- ially apically papillose, 5.5-6 mm high, exceeding the gynostegium (including stylar head) and partly obscuring it; C(is) consisting of Cs and Ci fused for about % of total corona length, Cs and Ci differ- entiated, Ci shorter than to as long as Cs. Cs not adnate to the filaments, not appressed to the back of the stamens, without adaxial appendages: lobes of Cs laminar to filamentous, ovate-oblong, produc- ing a pronounced convex fold. Lobes of Ci laminar, Volume 83, Number 3 1996 Liede 307 Cynanchum in Africa very elongatedly oblong, producing a pronounced convex fold along the upper two-thirds of corona length, erect, with laterally involute margins. Gy- nostegium 44.2 mm high, 2.3-2.5 mm diam., ses- sile. Stamens with free filaments 0.8-0.9 mm long; anthers about as high as broad, trapezoidal, abaxially planar; anther wings 0.45-0.5 mm long, convergent, extending along the whole length of the anther, adjacent anther wings parallel, centrifugal, basally forming a distinct “mouth.” Connective ap- pendages 0.85-0.9 mm long, 0.85-0.9 mm wide, ovate, broader than the stamen, erect, with dentic- ulate margins. Pollinarium: corpusculum 0. 0.28 mm long; caudicles 0.2-0.23 mm long, flat- tened, s-shaped, convex-concave, trapezoid; pollin- ia subapically attached to the caudicles, 0.4—0.42 mm long, 0.2-0.22 mm wide, ovoid, ovate in cross section. Stylar head white, 1.2-1.3 mm diam., 1.4— 1.5 mm high; upper part 1-1.2 mm high, umbonate (only in Gillett et al. 22615) or elongated-conical (not exceeding the corona, in contrast to subsp. clavidens). Follicles normally one per flower, pen- dulous, 60-100 mm long, 10-15 mm diam., obcla- vate, round in cross section, apically strongly beaked, light to medium brown, longitudinally grooved, glabrous. Seeds 6.5-7.5 mm long, 44.5 mm wide, ovate, medium brown, seta and aseta side papillose with regularly arranged papillae, a with sparse, regularly arranged trichomes 0. mm long, marginally with wing 0.6-0.8 mm A ien denticulate margin; coma 22-25 mm long. Chro- mosome number: 2n — 22 (voucher: Liede & Newton 3226, ULM). Distribution and habitat. Africa: Eritrea, Ethi- opia (Gamo Gofa, Shewa, Sidamo, Tigray). (Kl, K4), Mali, Niger, Somalia (Hiiraan, Mudug, Sanaag), Tanzania (T1, T2), Upper Volta; 200—1600 m; Acacia seyal-Balanites bushland. Widespread, but localized. Figure Kenya Flowering time. March to October. Vernacular name. Gesuriat (Somali); Shubkax (Hobyo). Uses. Fruits edible. Selected specimens examined. ERITREA. E of Amba- . Gam , 16 Apr. 1985, Ep 576 (K). Shewa: 1 km W of Birrta on track to K 0 km N of Meki, 1700 m, 22 Sep. 197, Glen & se Abate 3123 (K). Sidamo: Borana, 4 of Neghelle on road to Filtu and Dolo, 1450 m, p 2 con Ash 2421 (K, MO, UPS). Tigray: the road to Maichew, ca. 2150 5 7 Ce "i Mac hakos, Ithaba, 1000 m, 15 May 1938, Bally 8357 (K). Northern Frontier: 13 km N of Isiolo on rd. to Marsabit, 1050 m, 2 Nov. 1978, Gilbert, Gathachi & Gatheri 5316 (K). NIGER. Plateau Koutou, 9 Sep. 1966, Fabrégues 2065 (P). SOMALIA. Hiiraan, 5 km W of Muk- akori on the rd. to Buloburti, 200 m, 13 June 1979, Gillett, Hemming & Watson 22615 (K); Mudug, 53.4 km N of Hobyo along Hoby-Budbud inland route, 26 May 1987, Wieland 4319 (MO); Sanaag, Dobo pass, 1330 m, 5 Feb. 1933, Gillett 4952 (FT, K). TANZANIA. Arusha: Mbulu, Mbagaya River-Ndabash, Lake Manyara Nat. Park, 1580 m, 2 Mar. 1964, Greenway & Kanuri 11282 (K). Shinyanga: Shinyanga, Koritschoner 2084 (K). UP- PER VOLTA. Markoye, 21 Aug. 1975, Toutain 46782 (P). Comments. Both Cynanchum hastifolium N. E. Br. and C. hastifolium K. Schum. were described in October 1895, so that priority cannot be estab- lished. Here, the common practice to attribute the name to N. E. Brown, who provided a detailed de- scription and a type, is followed. Schumann, in con- trast, just mentioned the taxon and did not indicate any type material. Cynanchum macinense A. Chev. [Explor. Bot. Af- rique Occ. Franc. 1: 435. 1920, nomen nudum Mali. Macina, pays de Habés, de Ko boro-Kerki à Kanikombolé, 2 Sep. 1910, Chevalier 24861 (P)] represents this taxon. The differences between subspecies hastifolium and subspecies clavidens are of such minor nature that species rank cannot be maintained. The best distinguishing character is the long-exserted stylar ead in subspecies clavidens. Mature buds and flowers of subspecies clavidens are almost always much larger than those of subspecies hastifolium, but floral size is quite variable in both subspecies. Subspecies clavidens commonly inhabits low-lying, eastern localities; subspecies hastifolium is com- monly found in the more western highlands. This pattern, however, breaks down in southern Somalia, where subspecies hastifolium is found in low-lying coastal areas. The distribution of subspecies has- tifolium is remarkable for the fact that it is one of the very few African species that shows a marked disjunction between East and West Africa. Cynanchum clavidens shows affinities to the two Somalian endemics, C. crassiantherae and C. rub- ricoronae, but is otherwise isolated in Cynanchum. e only known material of Perianthostelma Baillon, a specimen with the unpublished name P. abyssinicum in P, represents C. clavidens subsp. hastifolium. 8. Cynanchum crassiantherae Liede, sp. nov. TYPE: Somalia. Shaabeellaha Dhexe, 10-12 km N of Adale on rd. to Haji Ali, 15 m, May— June 1983, Gillett & Hemming 24513 (holo- type, K; isotype, EA). Figure 14. 308 Annals of the Missouri Botanical Garden Figure 14. Cynanchum war rs Liede. 1—6: Gillett & Hemming 24513.—1. Shoot with nue —2. Flower.—3. Flower in lateral view.—4. Gynostegium.—5. Pollinarium.—6. Stylar head. Drawn by G. Hint Volume 83, Number 3 1996 Liede 309 Cynanchum in Africa Plantae erectae, rhizomatosae, foliis carnosulis, margin- staminalibus filamentosibus apicale reflexis; gynostegio breve stipitato, antheris crassis. Plants erect, 20-25 cm high, sparsely basicau- lously branched; rhizomatous, rhizomes 2-2.5 mm diam. Shoots herbaceous, glabrous; internodes 12- 20 cm long, 1-1.5 mm diam. “Stipules” ovate, 3-8 mm long, 2-3.5 mm wide. Leaves with petioles 10— 15 mm long; leaf blades fleshy, 15-25 mm long, 10-13 mm wide, triangular, basally lobate to auric- ulate, lobes 3-5 mm long, without colleters, apically acute to acuminate, adaxially and abaxially glabrous, margins thickened, crenulate. /nflores- cences bostrychoid to sciadioidal, 10—16-flowered, 8-12 flowers open at a time. Peduncles 0-2 mm long, glabrous. Flowers with floral bracts 2-2.2 mm long, 0.2-0.3 mm wide at the base, linear, glabrous; pedicels 5-8 mm long, glabrous. Buds 2.8-3 mm long, 1.8-2 mm diam., conical, with imbricate aes- tivation. Calyx basally fused, abaxial surface gla- brous; lobes 1.6-1.8 mm long, 0.4-0.5 mm wide, ovate-lanceolate, apically acute. Corolla rotate, basally fused; 3-3.5 mm long, abaxially and adax- ially yellowish green; lobes 1-1.2 mm wide, hori- zontally spreading, oblong, apically acute. Corona cyathiform, white, high, exceeding the gynostegium but not obscuring it; C(is) consisting of Cs and Ci fused for about % of total corona length, Cs and Ci differentiated, Ci shorter than Cs. Cs not adnate to the filaments, without adaxial ap- pendages; lobes of Cs lobes filamentous, apically reflexed. Lobes of Ci laminar, ovate to triangular, flat, erect. Gynostegium 1-1. m high, 1.6-1.7 mm diam., atop a stipe, 0.4-0.5 mm long. Stamens without free filaments; anthers about as high as broad, massive, deltoid, abaxially convex; anther wings 0.4-0.45 mm long, parallel to each other, extending along the whole length of the anther, with space between distal and proximal ridge of anther wings glabrous; adjacent anther wings divergent to- ward the base, in the same plane as the anther. Connective appendages 0.5-0.6 mm long, 0.4—0.5 mm wide, triangular, equaling the stamen in width, slightly inflexed. Pollinarium: corpusculum 0.2 mm long, ovoid; caudicles 0.1 mm long, flattened, straight, declinate, triangular; pollinia subapically attached to the caudicles, 0.3—0.35 mm long, 0.1— 0.12 mm wide, ovate in cross section, oblongoid. Stylar head 0.7-0.8 mm diam., 0.6-0.65 mm high; upper part 0.3 mm high, equaling the lower part in height, conical. Fruits, seeds, and chromosome num- ber unknown Distribution and habitat. Africa: Somalia (Gal- guduud, Shabeellaha Dhexe); < 50 m; coastal dunes. Rarely collected. Figure 12. Flowering time. May. Additional specimens examined. SOMALIA. Galgud- yg A km on r , 30 May 1989, Thulin & A 2 Achab sia 13 km N of Cadale along rd. to Ceeldheer, 10 m y 1990, Thulin, Hedrén & Abdi M. Dahir 7231 (UPS) Comments. Cynanchum crassiantherae is close- ly related to another Somali endemic, C. rubrico- ronae. It is clearly distinct from this and all other species of Cynanchum by its massive anthers. 9. Cynanchum ellipticum (Harvey) R. A. Dyer, Mem. Bot. Surv. South Africa 17: 138. 1937. Bunburia elliptica Harvey, Gen. S. Afr. Pl. 416. 1838. TYPE: South Africa. Cape: near Grahamstown, Bunbury s.n. (holotype, Herb. Hook. 1867, K) Plants twining, 1.5-3 m high, richly branched; subterranean organs consisting only of fibrous roots. Shoots perennial, 200-300 cm long, 1-1.5 mm diam., herbaceous, glabrous, basally woody with grayish bark. “Stipules” ovate, almost round, 4-6 mm long, 4-6 mm wide. Leaves with petioles 5-15 mm long; leaf blades herbaceous, 20-40 mm long, 10-20 mm wide, elliptic to oblong, basally rounded with 2 colleters adaxially, apically obtuse to acute, apiculate, adaxially and abaxially glabrous. /nflo- rescences bostrychoid to sciadioidal, 2-15(rarely — 30)-flowered, 4—10 flowers open at a time; rachis to 8 mm long; peduncles 12-20 mm long, glabrous. Flowers sweetly scented; floral bracts 0.8-1 mm long, 0.5-0.7 mm wide at the base, triangular, gla- brous; pedicels 5-10 mm long, glabrous. Buds 2.5— 3.5 mm long, 1.5-2.5 mm diam., ovoid; aestivation imbricate. Calyx basally fused, abaxially glabrous; lobes 0.8-1.2 mm long, 0 mm wide, ovate to oblong, apically obtuse. Corolla rotate, basally fused, spreading to recurved, 2.5—4 mm long, abax- ially and adaxially glabrous, brown to green; lobes 0.8-1.2 mm wide, spreading to recurved, cucullate, apically obtuse, straight, or apically twisted. Coro- na white, cyathiform, 2.5-3.5 mm high, exceeding the gynostegium but not obscuring it, abaxially gla- brous; C(is) consisting of Cs and Ci completely fused; upper margin entire or irregularly crenulate. Cs without adaxial appendages. Gynostegium 1-1.4 mm high, 1.2-1.6 mm diam., sessile. Stamens with- out free filaments, anthers broader than high, trap- ezoidal, abaxially convex; anther wings 0.5—0.6 mm long, convergent, proper forming a basal arch, adjacent anther wings extending beyond the anther 310 Annals of the Missouri Botanical Garden parallel, in the same plane as the anther. Connec- tive appendages 0.35—0.4 mm long, 0.25-0.3 mm wide, triangular, narrower than the stamen, slightly inflexed. Pollinarium: corpusculum 0.2-0.22 mm long; caudicles 0.15-0.17' mm long, flattened, straight, declinate, thickened at the insertion of the pollinium, pes pollinia laterally attached to 5 mm long, 0.1-0.12 mm wide, ovate, round in cross section. Stylar head 0.8-0.85 mm diam., 0.7-0.8 mm high: upper part 0.45-0.48 mm high, depressed-conical. Follicles 45-60 mm long, 6-8 mm wide, obclavate, obtusely deltate in cross section, apically strongly beaked, keeled, longitudinally slightly grooved, glabrous. Seeds 6-7 mm long, 3.5—4.5 mm the caudicle medium brown, wide, pyriform, dark brown, seta and aseta side tu- berculate and sculptured with longitudinal ridges (less pronouncedly so on the aseta side); margins wingless, entire; coma 25 mm long. Chromosome number: 2n = 22 (voucher: Liede 2933, ULM). Distribution and habitat. Africa: Mozambique, South Africa (Cape Province, Natal, Transvaal); 0— 1300 m; flats to moderate slopes, in sand or be- tween rocks, indigenous forests and forest margins, thickets, frequently in disturbed habitats. Comments. In Liede (1993), this species was discussed under C. capense Thunb. However, Cy- nanchum capense L.f. [Suppl.: 168. 1782] was in- terpreted by N. E. Brown (1908) as a synonym of Pentatropis microphylla Wight & Arn., because the description is clearly taken from the specimen Kón- ig s.n., The other specimen cited in the protologue (Spa- which represents Pentatropis microphylla. rrmann s.n.) represents Cynanchum eur L.f. Cynanchum capense Thunb. iud : : 47. 1800; type: Thunberg s.n. , UPS 6 3. UPS] represents a later homonym of ara i capense Further details, illustration, distribution map, and citation of specimens are provided in Liede (1993) under C. capense. 10. Cynanchum faleatum Hutchinson & E. A. Bruce, Bull. Misc. Inform., Kew 1941: 145. 1941. TYPE: Somalia. Woqooyi Galbeed, boundary, 44^10'E, 8?57'N, 1290 m, 4 Oct. 1932, Gillett 4114 (holotype, K). Figure 15. Plants ascending, twining, 1—1.5 m high, richly and irregularly branched. Shoots sparsely to densely covered with appressed tri- herbaceous, chomes 0.3—0.4 mm long; internodes 3-8 cm long. 0.5-1 mm diam. “Stipules” absent. Leaves with pet- ioles 2-10 mm long; leaf blades herbaceous, 20— 40 mm long, 2-16 mm wide, triangular to falcate, basally truncate, obtuse, rounded or cordate with 4—5 colleters in the adaxial sinus, apically acute to acuminate, adaxially papillose, sparsely covered with erect trichomes 0.3-0.4 mm long, evenly dis- tributed over the whole surface, abaxially sparsely to densely covered with appressed trichomes 0.23- 0.3 mm long, evenly distributed over the whole sur- face. Inflorescences bostrychoid to sciadioidal, 5— 12-flowered, 4—10 flowers open at a time; rachis to 2 mm long; peduncles 0-5 mm long, densely cov- ered with appressed trichomes 0.16—0.25 mm long. Flowers with floral bracts 0.5-1 mm long, 0.4—0.6 mm wide at the base, ovate, with trichomes; pedi- cels 2-5 mm long, densely covered with flexuous trichomes 0.15-0.25 mm long. Buds 2.2-3 mm long, 1.5-1.7 mm diam., ovoid, with imbricate aes- Calyx fused for about % of its length; ab- lobes 0.9-1.2 mm long, 0.6-0.8 mm wide, ovate, apically acute. Co- rolla cyathiform, basally fused; 2-3 mm long, abax- ially and adaxially creamish green to yellow; lobes tivation. axial surface with trichomes: 1-1.3 mm wide, patent, oblong to lanceolate, api- cally obtuse to acute. Corona cyathiform, white, 8-2 mm high, equaling the gynostegium in height: C(is) consisting of Cs and Ci fused for % to % of total corona length, Cs and Ci differentiated, Ci shorter than Cs. Cs not adnate to the filaments, without adaxial appendages; lobes of Cs laminar, broadly triangular, flat, apically reflexed. Lobes of Ci laminar, ovate to bifid, reflexed, with straight, entire margins. Gynostegium 2-2.5 mm high, 1.4— 1.6 mm diam., ments, anthers about as high as broad, trapezoidal, abaxially planar; anther wings 0.7-0.8 mm long, parallel to each other, extending along the whole sessile. Stamens without free fila- bu of the anther; adjacent anther wings paniei centrifugal. Connective appendages 1-1.2 mm ong, 0.8-1 mm wide, ovate, broader than the sta- men, erect, inflated, with denticulate margins. Pol- linarium: corpusculum 0.2-0.25 mm long, ovoid; caudicles 0.1-0.15 mm long, cylindrical, concavely recurved, thickened at the insertion of the pollini- um; pollinia subapically attached to the caudicles, 0.45-0.55 mm long, 0.12-0.15 oblon- goid, ovate in cross section. Stylar head 0.8—1 mm diam., 1.2-1.3 mm high; upper part 0.9-1 mm high, capitate. Follicles one per flower, pendulous, 50—65 mm long, 4 mm diam., obclavate, apically strongly beaked, keeled, dark brown, with dense indument. Seeds 5.56 mm long, 3.5—4 mm wide, ovate, light to medium brown, seta and aseta side mm wide, sculptured with longitudinal ridges, margins with 0.3-0.5-mm-wide wing with dentate margin; coma 20-23 mm long. Chromosome number unknown. Volume 83, Number 3 Liede 1996 Cynanchum in Africa Figure 15. Cynanchum falcatum Hutchinson & E. A. Bruce.—1. Shoot with inflorescences (Friis et al. 3169); extra leaf of the type (Gillett 4114). 2-5: De Wilde 6331.—2. Flower.—3. Gynostegium.—4. Pollinarium.—5. Stylar head.— 6. Fruit (Friis et al. 3169). Drawn by G. Hintze. Distribution and habitat. Africa: Ethiopia (She- Flowering time. February to August. wa, Sidamo), Kenya (Kl), Somalia (Bakool, Nu- gaal); 250-1500 m; Acacia-Commiphora open bushland. Rare. Figure 16. Additional specimens examined. ETHIOPIA. Shewa: Vernacular name. | Hayab (Gessariad). 312 Annals of the Missouri Botanical Garden n- 1003 16. Known distribution of Cynanchum falcatum (dots), C. heteromorphum (asterisks), and C. polyanthum gure (open circles). 23 km N of Awash on rd. to Nazareth, ca. 1090 m, 30 Aug. 1967, Westphal & decia MS 1466 (K); 37 km NE Nazareth along rd. to Awash, 350 m, 4 Feb. 1970, De Wilde 6331 (K, WAG). Siles 40 km NE Neghelle along rd. to Filtu, 1450 m, 20 May 1982, Friis, aa & Vollesen 3169 (K). KENYA. Northern Frontier: Huri Hills, 25 Feb. 1963, Bally 12526 (K ©). SOMALIA. Taka. 22 km W of Mukwakori, 250 m, 13 June 1979, Gillett, Hemming «€ Watson 22616 (K); Nugaal, Gardo-Hudun Rd., 950 m, 20 June 1981, Gillett & Beckett 23533 (EA). Comments. Cynanchum falcatum is close to C. heteromorphum, but the leaves are triangular and the corona more highly fused. The type is an un- representative specimen with very slender, falci- form leaves, perhaps caused by disease. Most spec- imens exhibit leaves more like those widespread in the genus, while floral structure is the same. 11. Cynanchum galgalense Liede, sp. nov. TYPE: Somalia. Bari, Al Miskat Mts., ca. 15 km SW of Candala, Toh well, 800-850 m, 2 Dec. 1985, Thulin 5612 (holotype, K; isotype, UPS). Figure 17. Plantae ae carnosulis marginibus incrassatis, infer- ioribus caducis; partibus staminalibus interstaminalib- usque coronae quits tegialis late connatis, sulcatis, gynos- tegium obducenibus; capite stylorum tabulari. Plants ascending, twining, 5-30 cm high, sparsely basicaulously branched. Shoots herba- ceous, densely covered with erect trichomes 0.1— 0.12 mm long, basally slightly woody, with yellow- ish to brownish bark; internodes 10—50 cm long, 0.7-1.2 mm diam. “Stipules” absent. Leaves cadu- the upper ones sessile, cous, lower ones with Volume 83, Number 3 313 Liede Cynanchum in Africa Figure 17. Cynanchum galgalense Liede. 1: Thulin & Warfa 5612; 2-5: Newbould 1095.—1. Habit with inflores- cences.—2. Flower.—3. Gynostegium and corona, partially removed.—4. Pollinarium.—5. Stylar head. Drawn by G. Hintze 314 Annals of the Missouri Botanical Garden petioles to 5 mm long; leaf blades fleshy, 4-14 mm long, 2-9(-13) mm wide, elliptic, basally rounded, without colleters, apically acute, margins straight and thickened, adaxially and abaxially sparsely to densely covered with erect trichomes 0.1—0.12 mm long, evenly distributed over the whole surface. /n- florescences bostrychoid, 5-24-flowered. all flowers open at a time; rachis 5-50 mm long; peduncles 5—7 mm long, densely covered with erect trichomes 0.1-0.15 mm long. Flowers with floral bracts 1-1.2 mm long, 0.1-0.2 mm wide at the base, linear, with trichomes; pedicels 3—4 mm long, sparsely covered with erect trichomes 0.1-0.12 mm long. Buds 1.4— 1.5 mm long, 1 mm diam., conical, with imbricate Calyx basally fused; abaxial surface with trichomes; lobes 0.8—1 mm long, 0.4—0.5 mm aestivation. wide, ovate, apically acute. Corolla cyathiform, ba- sally fused; 1.5-1.7 mm long, abaxially and adax- ially cream-colored; adaxially with verrucose tri- chomes 0.04—0.05 mm long, evenly distributed over the whole surface; lobes 0.5-0.7 mm wide, in- curved to patent, ovate to oblong, apically acute. Corona urceolate, ivory, ca. 0.8 mm high, exceeding the gynostegium and partly obscuring it; C(is) con- sisting of Cs and Ci fused for more than % of total corona length, Cs and Ci differentiated, Ci as long as Cs. Cs basally just adnate to the filaments, with- out adaxial appendages; lobes of Cs laminar, ovate, producing a pronounced convex fold along the up- per half of corona length, apically erect. Lobes of Ci laminar, oblong, producing a pronounced convex fold along the upper half of corona length, erect to reflexed, with straight margins. Gynostegium ses- 1-1.1 Anthers broader than high, trapezoidal, abaxially convex: sile, 0.9-1 mm high, mm diam. anther wings 0.7—0.8 mm long. convergent, extend- ing along the whole length of the anther; adjacent anther wings parallel, in the same plane as the an- ther. Connective appendages 0.3—0.4 mm long, 0.3-0.4 mm wide, ovate, narrower than the stamen, strongly inflexed. Pollinarium: corpusculum ca. 0.15 mm long, elliptic, margins of the corpuscular cleft basally widened; caudicles ca. 0.05 mm long. flattened, concavely recurved, trapezoid; pollinia subapically attached to the caudicles, 0.11-0.12 mm long, 0.045-0.055 mm wide, elliptical in cross section, oblongoid. Stylar head 0.35—0.4 mm diam., 0.2-0.25 mm high; upper part 0.08—0.1 mm high, equaling the lower part in height, tabular. Follicles one per flower, ca. 30 mm long, 4 mm diam., fu- siform, apically acute but not beaked, light brown, smooth, isolatedly covered with trichomes, with pa- pery pericarp. Seeds and chromosome number un- known. Africa: Somalia (Bari, 00 m; in rock crevices of Distribution and habitat. Nugaal, Sanaag); limestone cliffs. Localized and rare. Figure 2. Flowering time. July, November to December. Vernacular name. Darjo. Additional specimens ceaminen SOMALIA. Bari, Gal- galo, 1000-1150 m, 1 Dec. 1986, Thulin & Warfa 6205 (K, UPS); above Galgalo, ca. 1150 m, 27 Nov. 1971, La- vranos 9014 (K): jud Me Tukalamis, Newbould 1095 (K); Sanaag, Surud R N of Erigavo, 1900-2060 m, 8 July 1981, Gillett & Watson 23843 (K). Comments. Cynanchum galgalense is a very unusual species, but most likely a Cynanchum with verrucose, sparse trichomes on the adaxial side of the corolla lobes and a highly fused corona. Prob- ably related to C. obtusifolium and its allies. White latex would confirm its position in the genus. 12. Cynanchum gerrardii (Harvey) Liede, Tax- on 40: 117. 1991. Sarcocyphula ella Harv., Thes. Cap. 2: 58, t. 191. 1863. TY South Africa. Natal: Tugela, Gerrard 1321 (lio lotype, TCD; isotype, BM). Cynanchum sarcostemmatoides K. Schum. in Engl., Pflan- zenw. Ost-Afrikas C. 323. Aug. 1895. CM sarcostemmoides K. Schum. in Engl. & Prantl, Nat. Pflanzenfam. 4(2): ui Oct. 1895. (orth. var.). TYPE: Tanzania. Tanga: Amboni, June 1893, Holst 2706 (lectotype, epale by Liede (1993), K). Plants ascending, twining, 0.5-2 m high, richly acrocaulously branched; subterranean organs con- sisting only of fibrous roots. Shoots perennial, semi- succulent, finely striate, obscurely glaucous, gla- brescent, isolatedly covered with appressed multicellular trichomes 0.2-0.4 mm long. basally corky, with thin, yellowish bark; internodes (2-)4— 8(-10) em long, 1.5-2.5 slightly ivory. Leaf scales often not exactly opposite, 0.8—1.2 mm long, 0.5-0.8 mm wide, apically acute. Inflorescences bostrychoid to sciadioidal, 4—7-flow- mm diam. Latex white to ered, 2 owers open at a time; rachis to 4 mm long; peduncles 0-2.5 mm long, sparsely covered with appressed trichomes 0.2-0.4 mm long. Flowers sweetly scented, nectariferous; floral bracts 0.2-0.5 mm long. 0.5-0.6 mm wide at the base, deltoid, with trichomes: pedicels 3.5 mm long, glabrous. Buds 0.4—0.5 to ovoid, with imbricate aestivation. Calyx basally mm long, 0.3-0.4 mm diam., globose fused; abaxial surface glabrous; lobes 0.3-0.4 mm long, 0.5-0.6 mm wide, ovate to triangular, apically acute. Corolla rotate, petals fused for about a quar- ter of their length; 2-3 mm long, abaxially green to white, adaxially green; lobes 1-1.5 mm wide, de- clinate, ovate, apically acuminate. Corona white, Volume 83, Number 3 1996 Liede 315 Cynanchum in Africa cyathiform, 1.2-1.5 mm high, slightly exceeding the gynostegium. C(is) consisting of Cs and Ci fused for % to % of total corona height, only Cs differentiated, Ci thinner than Cs. Cs adnate to the aments, appressed to the back of the stamens, without adaxial appendages; lobes of Cs laminar, triangular, apically erect to inflexed, with straight margins. Gynostegium sessile, 1.2-1.5 mm high, 0.8-1.2 mm diam. Stamens with free filaments 0.6— 0.8 mm long. Anthers broader than high, trapezoi- dal, abaxially planar to convex; anther wings 0.3— 0.4 mm long, parallel to each other, extending along the whole length of the anther; adjacent anther wings divergent toward the base, in the same plane as the anther. Connective appendages 0.38—0.42 mm long, 0.5-0.55 mm wide, ovate to triangular, narrower than the stamen, strongly inflexed. Polli- narium: corpusculum ca. .1-0.12 mm long, flattened, straight, declinate, tri- angular, thickened at the insertion of the pollinium; pollinia subapically attached to the caudicles, 0.4— 0.45 mm long, 0.15-0.2 mm wide, ovoid, round in cross section. Stylar head white, 0.8-0.85 mm diam., 0.25-0.3 mm high; upper part 0.2-0.23 mm high, flat to depressed-conical. Follicles one, oc- casionally two per flower, 85-120 mm long, 6-8 mm diam., elongated, round in cross section, api- cally short beaked, light brown to dark brown, smooth, glabrous. Seeds 5-6 mm long, 2-3 mm wide, pyriform, medium brown; seta and aseta side d s regularly arranged papillae and trichomes 0.3 m long, wingless, scan MIN coma 20-25 n — 22 (voucher: mm long; caudicles g. Chromosome num m lon Noltee 995, MSUN). Distribution and habitat. Arabian Peninsula: 5, K6, K7), Somalia (Sanaag, Woqooyi Galbeed), South Africa (Cape, Natal, Transvaal), Tanzania (T2, T3, T6, T7), Uganda (U1, U4), Zaire, Zambia, Zimbabwe. Mascarene Islands: Comores, Madagas- car (Antsiranana, Toliara). 0—1500 m; close to the sea or further inland, on rocky outcrops, in sandy or clayey depressions, often in slightly disturbed sites. Very widespread, but not frequent. Figure 18. Flowering time. All year; mostly after rains. Vernacular names. Debina Dir (Somalia); Ki- pagoro (Kitamba), Melktou (South Africa). Uses. Zulus eat young shoots; in East Africa used as fish poison; in Somalia eaten, much liked by pregnant women (Gillett 3939). Selected specimens examined. ETHIOPIA. Gamo Gofa: 27.5 km S of Arba Minch, 1150-1300 m, 3 Aug. 1975, Gilbert, Thulin & Aweke 330 (K, MO, UPS). Shewa: ca. 40 km E of Nazareth, ca. 1200 m, 5 Apr. 1966, De Wilde & De Wilde-Duyfjes 10519 (K, MO). Sidamo: 1 km N of Harekelo along rd. to Kebre Mengist 1450 m, 24 May 1982, des Tadesse & Vollesen 3292 (K). KENYA. Central: Kitui, 1 mi. N of Kangonde on Kangonde-Embu rd., 8 May 1960, eta 1664 (K); Machakos, Kibwezi Plaisis, 1000 m, 1938, pel 8333 (K). Coast: Kilifi, Malindi Golf DEA 10 m, 22 Sep. 1989, Robertson 5914 (K); Kwale, Kilibasi Hill, 450—827 m, 19 Nov. 1989, Luke & Robertson 2105 (K); Mombasa, Port Tudor, Mom- I thes m, Oct diy MacNaughtan 26 (K); Tana Riv- , 30 S of Garsen, 15 m, 25 Sep. 1961, Fen & k Paulo 560 (K); Taita, Voi, 660 m, 11 May 1931, Napier 1064 (K); Masai, SW of Ngong hills, 1860 m, 4 Aug. 1968, Gillet 18689 (K). Nyanza: South Kavirondo, Sand Mountain, 2 Dec. 1934, Turner 6628 (K, MO); Tur- kana, 1 km NE of Loiya on Lodwar rd., ior 75 m, 8 Nov. 1977, Carter & Stannar 264 (K). MOZAMBIQUE. Cabo g. 1983, Groenendijk & Dunge 566 N fs g l=3 O N A p N © B 1962, Wild 5877 (K, Surud, 1973, Bally & — n. (K); Woq Hargesia, 1130 m, 21 Sep. 1932, “Gillen 3939 K). SOUTH AFRICA. Cape: i N of Pluto's Vale, 11 Apr. 1954, Noel 1544 (GRA); Alexandria, Zuurkop, Addo National Park, 330 m, 18 Oct. 1951, Archibald 3854 (GRA); But- terworth, along the Kei River, Jan. 1892, Flanagan 1038 (BOL, SAM); Grahamstown, Piggott Bridge, 265 m, 13 Apr. 1978, Bayliss 8474 (MO); Jansenville, Gannahoek, 8 May 1985, Hoffman 791 (GRA); i River, 600 m, 15 Mar. 1970, Bayliss 4554 ( Elizabeth, Farm Vaalkrans, 600 m ker 4928 (NBG); Steytlerville, 2 km ward Steytlerville, 8 Dec. 1978, a 1887 (NBG); Stut- terheim, Dec. 1870, ns E Durban, Bluff, 23 Apr. 1914, Wood 12613 (PRE); Lower Tugela, Lower pP valley opposite Gnembe River con- fluence, ca. 170 m, 27 Feb. 1963, Edwards 3048 (PRE); Nkandhla, Nogeya, Cia Valley, 31 May 1967, Venter 3708 (PRE); Port Shepstone, Oriba Flats, Umzimkulu ini, 16 Apr. 1968, Strey 8128 (PRE). Tran "4 Eastern Transvaal, Rogers 2661a zr SWAZI D. Louwsburg, Maloma, ca. 500 m, 1963, pomi s.n. (NBG 71759, NBG); Mbabane, near Ci. 9 Sep. 1957, Compton 27031 (NBG); Stegi, Mhlumeni Border Station, 22 mi. NE of Stegi, 330 m, 4 June 1947, Codd & Dyer 2929 (PRE). TANZANIA. Arusha: Kemo- somu gorge, 1660 m, 19 Feb. 1971, Richards & Arasululu 26688 (K). Iringa: top of Kitonga gorge, Image mountain, 1200-1800 m, 9 Dec. 1986, Lovett & Congden 1070 (K, . Morogoro: Uluguru Gebirge, 500 m, 6 June 1933, Schlieben 3993 (B); Moshi, Kikafu River bridge, 15 Apr. 1968, Bigger 1773 (EA). Pwani: Mafia Island, Tretole, 2 Aug. 1932, Schlieben 2620 (B). Tanga: Sawa, sea level, 3 T 1956, Faulkner 1914 (K). Zanzibar: Rio sea vel, 26 June 1960, Faulkner 2624 (K). UGANDA. Bun- yoro, Hudango Forest, m 1936, Eggeling n (K); Kar- amoja, near Amudat, 1330 m, Wilson 858 (K). ZAIRE. Ruindi, Oct. 1937, fale 8000 (K); Katanda, Sep. 1937, E 316 Annals of the Missouri Botanical Garden AT á E 9 BN } 4 | “7 | 1 XT \ x I re a (ae — Wn D w A EQ EY A P / ( 4 P n AO t AC Figure 18. circles). Lebrun 7702 (K); May-yo-moto, Sep. 1937, Lebrun 9217 ); es et la Kamakaba, Jan. 1938, Lebrun 9694 (K). ZAM usaka, 8 Mar. 1971, Fanshawe 11191 (K). a pin Chuanja Hill, Lona-re Zhau Game Reserve, 27 Ma . Ngoni 134 (MO); Mandula, Whin- dale Ranch, dn ila, ‘above sable ees ca. 1000 m, 24 Mar. 1969, Leach, Biegel & ALS 28 (K, MO, SRGH); i. E Nus 5287 (K, bus r. 1969, Leach 11630 (K. SRGH); Nuane r Kap enis, 65 km NE of Malvernia, 25 Apr. 1962, Tae 7718 (K, SRGH). Comments. Throughout its history, several names have been applied to Cynanchum gerrardii. Of these, Cynanchum aphyllum (Thunb.) Schltr. is not an available name for this species, because Asclepias aphylla Thunb. [Prod. Fl. Cap. 47 (1794)], a later homonym of A. aphylla Forssk., has been lectotypified (Liede, 1991) in such a way Known distribution of Cynanchum gerrardii (dots), C. lenewtonii (asterisks), and C. schistoglossum (open that it is a synonym of Sarcostemma viminale (L.) R. Br. Likewise, Cynanchum eir pierum (Turez.) R. A. Dyer ex Bullock [Kew Bull. 10: 624. 1955] is based on Sarcostemma tetrapterum n 'z., which has been lectotypified (Liede, 1991) on a specimen repre- senting Sarcostemma viminale (L.) R. Br. sensu lato. Vincetoxicum sarcostemmoides Schweinf. ex Pen- zig [Atti Congr. Bot. Int. Genova 349 (1893), nomen nudum] also represents this C. gerrardii (fide Bul- lock, 1955). An extremely adaptable species, also easy to cul- tivate. The affinities of C. gerrardii clearly lie with a group of leafless Malagasy species. Further details, illustration, and a detailed south- ern African distribution map are provided in Liede (1993 Volume 83, Number 3 1996 Liede 317 Cynanchum in Africa 13. Cynanchum gonoloboides Schlechter, Wiss. Erg. Deut. Zentr.-Afr. Exped., Bot. 2: 543. 1913. TYPE: Rwanda: Schlechter 1617 ie B presumably destroyed). NEO- YPE: Kenya. Rift Valley: Nakuru, Doboti, in 9 mi. from Cobb's gate near the Mau Forest Reserve on the track to Nairagie Ngare, 3200 m, Glover, Gwynne & Samuel 1492 (neo- type, designated here, K; isoneotypes, EA, FT). Figure 19 Plants ascending, twining, 4—5 m high, richly and irregularly branched, rhizomatous; subterra- nean organs woody rootstocks, 7-9 cm diam. Shoots perennial, herbaceous, glabrescent, densely cov- ered with flexuous trichomes when young; basally woody, with yellowish bark, internodes 7-20 cm long, 2.5-3 mm diam. “Stip- ules” absent. Leaves with petioles 12-20(—40) mm long, leaf blades herbaceous, 55-90 mm long, 35- 60 mm wide, ovate, basally cordate with 8-12 col- leters in the adaxial sinus, apically acuminate, apiculus 4—10 mm long, adaxially densely covered with flexuous trichomes 0.2-0.25 mm long, evenly mm long, distributed over the whole surface (when young), abaxially densely covered with flexuous trichomes 0.4—0.5 mm long, restricted to veins and margins (when young). Inflorescences bostrychoid, 6—15- flowered, 3—7 flowers open at a time; rachis 5-30 mm long; peduncles 15-25(-30) mm long, sparsely covered with flexuous trichomes 0.25-0.35 mm long. Flowers pleasantly honey-scented; floral bracts 2-2.5 mm long, 0.5—0.7 mm wide at the base, linear to triangular, with trichomes; pedicels 10-22 mm long, densely covered with flexuous tri- chomes 0.15-0.25 mm long. Buds 3.54 mm long, 44.5 mm diam., globose, with imbricate aestiva- tion. Calyx basally fused; abaxial surface densely covered with trichomes; lobes 3.54 mm long, 2- 2.5 mm wide, ovate, apically acuminate. Corolla rotate, basally fused, 4-5 mm long, abaxially and adaxially greenish yellow, brown along the main nerves, often also basally; lobes 2-2.5 mm wide, horizontal to declinate, oblong, apically acute. Co- rona white, 3.54 mm high, equaling the gynoste- gium in height, C(is) consisting of Cs and Ci fused for more than % of total corona length, only Cs differentiated. Cs not adnate to the filaments, adax- ially with a basal protuberance corresponding to the lower margin of the anther proper, without adaxial appendages; lobes of Cs laminar, trifid (the medium lobe widely ovate, occupying about % of the total width, the lateral lobes na apically in- flexed. Gynostegium 1.2-1.3 m 3.5-3.7 mm diam., atop a stipe, 1.5-1.7 mm e nme with- out free filaments, anthers broader than high, del- toid, abaxially convex; anther wings 0.35—0.4 mm long, parallel to each other, not extending along the whole length of the anther; the anther forming a “pseudostipe,” of 0.2-0.25 mm height, adjacent an- ther wings parallel, in the same plane as the anther. Connective appendages 0.75—0.8 mm long, 1.4—1.5 mm wide, ovate, narrower than the stamen, strongly inflexed, with emarginate margins. Pollinarium: corpusculum 0.35—0.4 mm long; caudicles 0.8—0.9 mm long, cylindrical, s-shaped, concave-convex, horizontal; pollinia laterally attached to the caudi- cles, 0 m long, 0.24—0.26 mm wide, glo- bose to ovoid, ovate in cross section. Stylar head white to cream, 2.25-2.5 mm diam., 0.35—0.4 mm high; upper part flat. Follicles ca. 110 mm long, 15 mm diam., obclavate, round in cross section, api- cally acute, wingless, light brown, isolatedly cov- ered with short protuberances, with thick pericarp. Seeds 9-11 mm long, 4—6 mm wide, ovate, medium brown, seta side smooth, marginally with 0.4 mm wide wing, coma 20-25 mm long. Chromosome number unknown. Distribution and habitat. Africa: Ethiopia (Si- damo), Kenya (K3, K4); Rwanda, Tanzania (T6); 2800-3800 m; bamboo forest. Rather localized, but neither rare nor endangered. Figure 10. Flowering time. January to April. Vernacular name. | Ol'obito (Masai). ETHIOPIA. Sidamo: Mt. , 28 Jan. 1953, Gillett 14973 Additional specimens seen. Nyeri en geyo, Charanqui Hills, Embobut forest, in Arorr & Em- 3800 m, Jan. 1971, Tweedie 3912 top of Arawa Forest, ca. gon, 2860 m, Feb. 1961, “Tade 2103 (BR, K Comments. The type specimen [Rwanda. Kar- issimbi, ca. 2500 m, Nov. 1907, Schlechter 1617] could not be found and was most probably de- stroyed in B. The description of the species, how- ever, clearly matches Sp. A in Agnew (1974: 387) with C. gonoloboides. Therefore, Schlechter’s old name is neotypified here. Despite its highly fused corona and the distinctive stipe, the species is eas- ily identified as a sister species of C. balense. 14. Cynanchum heteromorphum Vatke, Lin- naea 40: 215. 1876. TYPE: Ethiopia. Tigray: Hamedo plain, 1530 m, 31 Aug. 1862, Schim- per 940 (holotype, B presumably destroyed; lectotype, designated here, K). Figure 20. Cynanchum fraternum N. E. Br., Bull. Misc. Inform., Kew 318 Annals of the Missouri Botanical Garden ie ere proba re 19, Cynanchum gonoloboides: Schltr. 1: Glover et al. 1492; 2-9: Liede & Newton 3157.—1. Habit with arcu :e.—2. Flower, top view.—3. Flower, lateral view.—4. Corona, staminal lobe; adaxial view.—5. Gynostegium and corona, partially removed.—6. Pollinarium.—7. Stylar head.—8. Fruits.—9. Seed, seta side. Drawn by Jim Conrad. Volume 83, Number 3 Liede 319 1996 Cynanchum in Africa Figure 20. Cynanchum heteromorphum Vatke. 1-5: Schimper 1802; 6, 7: Schimper 940.—1. Shoot with inflores- cences.—2. Flower.—3. Gynostegium and corona, partially removed.—4. Pollinarium.—5. Stylar head.—6. Fruit.— Seed, seta side. Drawn by G. Hintze. 320 Annals of the Missouri Botanical Garden 106: 256. eat TYPE: Ethiopia. Tigray: Djeladjer- anne, 24 ( 1840, Schimper s.n. (holotype, B umabl y d: ad lectotype, designated here, K). a pleianthum K. Se hum., Annuario Reale Ist. Bo 7: 39. 1898. TYPE: Somalia. Locis pal- üdosis in Pianure di Savati, 25 Nov. 1893, Ruspoli & Riva 1533 (holotype, B Bons destroyed; lec- totype, designated here, FT). Plants ascending, twining, sparsely branched. Shoots herbaceous, sparsely glabrescent with ap- pressed trichomes 0.25-0.3 mm long; internodes 6.5-10 cm long, 0.8-1.1 mm diam. “Stipules” ab- sent. Leaves with petioles 6-8 mm long; leaf blades herbaceous, 18-32 mm long, 8-14 mm wide, ovate, basally rounded to obtuse with 5 colleters adaxially, apically acute to acuminate, apiculus 0.5-1 mm long, adaxially isolatedly covered with appressed trichomes 0.25-0.3 mm long, evenly distributed over the whole surface, abaxially with appressed trichomes 0.25-0.3 mm long, concentrated on veins and margins. Inflorescences bostrychoid, 8—10-flow- ered, 3-5 flowers open at a time; rachis 1-3 mm long; peduncles 7-9 mm long, densely covered with appressed trichomes 0.15-0.18 mm long. Flowers with floral bracts 0.6-0.7 mm long, 0.3-0.5 mm wide at the base, triangular; pedicels 4-6 mm long, sparsely covered with appressed trichomes 0.15- 0.2 mm long. Buds 1.8-2 mm long, 1.3-1.5 diam., sally fused; abaxial surface with trichomes; lobes 1-1.2 mm long, 0.5-0.6 mm acute. Corolla cyathiform, basally fused; 2.5-3 mm mm ovoid, with imbricate aestivation. Calyx ba- wide, ovate, apically long; lobes 0.8—1 mm wide, patent, oblong, apically obtuse. Corona cyathiform, 3 mm high, exceeding the gynostegium but not obscuring it; C(is) consist- ing of Cs and Ci fused for a little less than % of total corona pie Cs and Ci differentiated, Ci shorter than Cs. Cs not adnate to the filaments, without ada ppendarés: lobes of Cs laminar, elongate-triangular, apically reflexed. Lobes of Ci laminar, oblong, flat, reflexed. Gynostegium 1.8-1.9 mm high, 1.8-1.9 mm diam., sessile. Stamens with- out free filaments, anthers about as high as broad, rectangular, abaxially planar; anther wings 0.9-1 mm long, parallel to each other, extending along the whole length of the anther; adjacent anther wings parallel, in the same plane as the anther. Connective appendages 0.65—0.7 mm long, 0.7— 0.75 mm wide, ovate, narrower than the stamen, erect, with dentate margins. Pollinarium: corpus- culum 0.2 mm long, about as long as broad, rhom- boid; caudicles 0.12—0.15 mm long, cylindrical, concavely recurved; pollinia subapically attached 5—0.4 mm long, 0.15-0.17 mm wide, oblongoid, ovate in cross section. Stylar head 0.7-1 to the caudicles, 0 mm diam., 0.5-0.6 mm high: upper part 0.25-0.3 mm high, umbonate. Follicles one per flower, pendulous, 50-65 mm long, 7-8 mm diam., obclavate, obtusely deltate in cross section, apically strongly beaked, keeled, light brown, longitudinally grooved, sparsely covered with trichomes. Seeds 4— 5 mm long, 2.5—3 mm wide, ovate, medium brown, seta and aseta side sculptured with longitudinal 0.4—0.5 dentate margin; coma 15-20 mm long. Chromosome ridges, marginally with 5 mm wide wing with number unknown. Africa: Cameroon, Eritrea, Ethiopia. (Ti- gray), Somalia; around 1600 m, on sandy soil. Very Distribution and habitat. Central African. Republic, rarely collected and even less so in recent years; presumably endangered. Figure 16. Flowering time. October to November. ~ Additional specimens examined. CAMEROON. Ber- toua, near Catholic mission, 5 Nov. 1960, Breteler 615 (K, WAG). CE NTR. AL AFRICAN RE PUBL IC. Krébidjé (Fort Sibut), vallée de la moyenne Tomi, 6 Oct. 1902, Chevalier 5649 (P). ERITREA. Medri od i Adi Ghebsus, 1600 m, 2 Nov. 1906, Pappi 7300 (FT Comments. The flower of Cynanchum hetero- morphum is quite similar to that of C. falcatum; however, the clavate stylar head and the pronouncedly triangular leaves of the latter are unmistakable. With C shares the unusual East African—West African dis- clavidens subsp. hastifolium, this taxon junction. 15. Cynanchum ledermannii Schlechter Bot. Jahrb. Syst. 51: 140. 1913. TYPE: Burundi. Bubanza, Mugomero (Rugazi), 2000 m, 2 May 981, Reekmans 10069 (neotype, designated here, K). Syntypes cited in the protologue: Cameroon. Northern: Banso Mtns., ca. 2000 m, Oct. 1909, Ledermann 5757; Muti-slopes, near Mfongu, 1700-1800 m, Oct.-Nov. 1909, Led- ermann 5892, 593 la (all probably destroyed in B, no isotypes found). Figure 21 Plants ascending, twining, to 3 m high, richly and irregularly branched. Shoots perennial, herba- ceous, sparsely covered with flexuous trichomes 0.4-0.5 mm long. along a single line; internodes 15-25 cm long, 1-1.2 mm diam. 5-7 mm long, 5-7 mm wide. Leaves with petioles 35-50 mm long, 4—6 colleters at the base of the leaves; leaf blades herbaceous, 60-70 mm long, ovate, “Stipules” 30-35 mm wide, ovate, basally lobate, lobes 4-7 mm long, apically acuminate, apiculus 10-12 mm long, adaxially glabrous, abaxially isolatedly cov- ered with flexuous trichomes 0.15-0.2 mm long, concentrated on veins and margins. Inflorescences Volume 83, Number 3 Liede 321 Cynanchum in Africa Figure 21. Cynanchum ledermannii Schltr. 1, 3-5: Schlieben 3436; 2: deci vdd —]. Shoot with inflores- cence.—2. Flower.—3. Gynostegium.—4. Pollinarium.—5. Stylar head. Drawn by U. Mev 322 Annals of the Missouri Botanical Garden sciadioidal, 8—15-flowered; peduncles 35-50 mm long, densely covered with flexuous trichomes 0.25-0.3 mm long, along a single line. Flowers with floral bracts 1.6—1.8 mm long, 0.3-0.5 mm wide at the base, elongate-triangular, with trichomes; ped- icels 5-7 mm long, densely covered with flexuous trichomes 0.2-0.25 mm long. Buds 2-2.5 mm long, 2-2.5 mm diam., globose, with imbricate aestiva- tion. Calyx basally fused; abaxial surface with tri- chomes; lobes 0.8-1 mm long, 0.5-0.7 mm wide, lanceolate, apically acute. Corolla rotate, basally fused; 3.54 mm long; lobes 1.7-2 mm wide, de- curved, ovate, apically obtuse. Corona 1.8-2 mm high, equaling or very slightly exceeding the gy- nostegium, C(is) consisting of Cs and Ci fused for almost % of total corona length, C(is) only Cs dif- ferentiated, cyathiform. Cs not adnate to the fila- ments, not appressed to the back of the stamens, without adaxial appendages; lobes of Cs laminar, oblong, flat, apically erect, with laterally and api- Gynostegium 1-1.2 mm on a bulge 0.5-0.7 mm cally involute margins. high, 1.3-1.5 mm diam., Stamens without free filaments; broader than high, trapezoidal, abaxially planar: mm long, convergent, extending anthers anther wings 0.7 beyond the anther proper with stamens forming a basal arch; adjacent anther wings parallel, in the same plane as the anther. Connective appendages ca. 0.5 mm long, 0.5 mm wide, ovate, narrower than the stamen, slightly inflexed. Pollinarium: corpus- culum 0.5 mm long, caudicles ca. 0.075 mm long, flattened, straight, horizontal, triangular; subapically attached to the caudicles, ca. 0.25 mm long, 0.12-0.15 mm wide, ovoid, ovate in cross sec- tion. Stylar head 0.9-1 mm diam., 0.35-0.4 mm high; upper part 0.1-0.15 mm high, umbonate. Chromosome number unknown. pollinia Africa: Burundi, 1700-2000 m: but a few Distribution and habitat. Tanzania (T6), Very rarely collected, Cameroon, Zaire; mountain forest. collections are probably erroneously hidden among the stacks of C. adalinae subsp. mannii; presumably endangered. Figure 6. schistoglossum or C. Flowering time. May, October to November. Additional specimens dirimi TANZANIA. Moro- Uluguru, Nordwestseite, ca. 1100 m, 13 Feb. 1933, Schlieben des (B) ZAÏRE. Yangambi. 21 July 1950, De- viet 539 (SRG Comments. In habit, Cynanchum ledermannii reminds one strongly of C. adalinae subsp. mannii: but, as Schlechter remarked in the protologue, the corona of the two species is very distinct. Also, the gynostegium is sessile, the stylar head conspicu- ously conical in C. adalinae, while here, the gy- nostegium is elevated on a bulge and the stylar head flat to umbonate. The corona reminds one a little of Pentarrhinum. All three syntypes cited by Schlechter (1913) were probably destroyed in B, and isotypes could not be located. As the accurate description allows an exact match of the characters of this species, it is neotypified here. 16. Cynanchum lenewtonii Liede, Kew Bull. 49: 119. 1994. TYPE: Kenya. Northern Fron- tier: Moyale, 13 Dec. 1952, Gillett 14031 (ho- lotype, K: isotype, B). Plants twining, to 2.5 m high, richly basicau- lously branched. Shoots perennial, semi-succulent, finely striate; obscurely glaucous, sparsely glabres- cent with flexuous trichomes 0.3—0.4 mm long; in- ternodes 25-60 cm long, 1.5-2.5 mm diam. Latex present, white (the note "clear juice" on the spec- imen Gillett 12964 could not be confirmed by my own field observations). Leaf scales papery, 1.2-1.7 mm long. 0.6-0.8 mm wide, ovate, apically acute. Inflorescences sciadioidal, 2—6-flowered, 2—4 flow- ers open at a time, subsessile. Flowers with floral bracts 0.5-0.7 mm long, 0.6-0.8 mm wide at the base, triangular, glabrous; pedicels 3-4 mm long. glabrous. Buds 2-2.5 mm long, 1.5-2 mm diam., ovoid to cylindrical, with imbricate aestivation. Ca- lyx basally fused, abaxial surface glabrous; lobes 0.6—0.7 mm long, 0.4—0.5 mm wide ovate, apically acute. Corolla rotate to very widely campanulate, fused between !4—^ of their length; 2.5-3.5 mm long, abaxially and adaxially greenish yellow; lobes —].2 mm wide, recurved, oblong to triangular, api- cally obtuse to acute. Corona white, cyathiform to campanulate, 2-2.5 mm high, exceeding the gy- nostegium, partly obscuring it, consisting of Cs and Ci fused for % to % of total corona length, both Cs and Ci differentiated, Ci shorter than Cs. Cs adnate to the filaments, appressed to the back of the sta- mens, adaxially with a basal protuberance corre- sponding to the filament, without adaxial append- ages: lobes of Cs filamentose, apically inflexed to erect (when young, they bend over the gynostegium and touch each other in the middle, later they open up): Ci laminar, triangular to very narrowly trian- gular, erect to reflexed; with straight upper margin. Gynostegium 1.5-1.6 mm high, 1.3-1.4 mm diam., sessile. Stamens with free filaments 0.6-0.7 mm long, broader than high, rectangular, abaxially pla- nar; anther wings 0.18-0.2 mm long, extending along the whole length of the anther, parallel to each other, adjacent anther wings divergent toward Volume 83, Number 3 1996 Liede 323 Cynanchum in Africa the base, in the same plane as the anther. Connec- ovate, narrower than the stamen, slightly inflexed. Pollinarium: corpusculum 0.18- 0.2 mm long, margins of the corpuscular cleft sin- uate; caudicles 0.1—0.12 mm long, flattened, straight, horizontal, not thickened at the insertion of the pollinium; pollinia laterally attached to the caudicles, 0.25—0.3 mm long, 0 .15 mm wide, ovoid, ovate to round in cross section. Stylar head 0.6—0.65 mm diam., 0.45-0.5 mm high; upper part 0.17-0.2 mm high, flat to depressed-conical. Fol- licles one, occasionally two, per flower, erect, 65— O mm long, obclavate, wingless. Seeds unknown. Chromosome number: — 22 (voucher: Liede & N Newton 3161, MSU Distribution. Africa: Ethiopia (Bale), Kenya (K1). Figure 18 Comments. Cynanchum lenewtonii is closely related to C. gerrardii, but is easily recognized by its long staminal corona parts closing over the gy- nostegium in young flowers. Further comments, citation of specimens, and il- lustration are provided in Liede (1994). 17. Cynanchum longipes N. E. Brown, Bull. Misc. Inform., Kew 1897: 273. 1897. TYPE: Nigeria. Lagos: Papalayito, 1895, Millen 48 (holotype, K). Figure 22. Plants ascending, twining, 5-8 m high, richly and irregularly branched. Shoots perennial, herba- ceous, sparsely covered with erect trichomes 0.7— 0.9 mm long, along a single line; internodes 10-20 cm long, 1.5-1.7 mm diam. “Stipules” ovate, acute, 10-15 mm long, 8-12 mm wide. Leaves with peti- oles 15-100 mm long; leaf blades herbaceous, 60— 115 mm long, 23—65 mm wide, ovate, basally lo- bate, lobes 10-15 mm long, with 9-11 colleters in the adaxial sinus, leaves apically acuminate, apic- ulus 8-12 mm long, adaxially and abaxially gla- brous. /nflorescences bostrychoid, 5—9-flowered, all flowers open at a time; rachis 2-7 mm long; pe- duncles 5-7 mm long, sparsely covered with flex- uous trichomes 0.15—0.2 mm long, along a single line. Flowers with floral bracts 1-1.2 mm long, 0.2— 0.4 mm wide at the base, triangular; pedicels 10— 20 mm long, densely covered with flexuous tri- chomes 0.15—0.25 mm long. Buds 5—6 mm long, 3.5—4 mm diam., conical, with imbricate aestiva- tion. Calyx basally fused, ciliate; lobes 1.4—1.5 mm long, 0.9-1 mm wide, ovate, apically acute. Corolla rotate, basally fused; 6-8 mm long, abaxially green- ish yellow, adaxially yellow; lobes 2-2.5 mm wide, horizontal, oblong, apically obtuse. Corona purplish red, 5-6 mm high, exceeding the gynostegium, partly obscuring it; C(is) consisting of Cs and Ci fused for more than ?4 of total corona length, Cs and Ci differentiated, Ci shorter than Cs. Cs basally just adnate to the filaments, without adaxial ap- pendages; lobes of Cs laminar, triangular or trifid (with two short side teeth), apically erect. Lobes of Ci laminar, triangular, erect. Gynostegium 0.35—0.4 mm high, 0.45—0.5 mm diam., Stamens without free filaments, anthers about as high as broad, pentagonal, abaxially convex; anther wings 0.24—0.26 mm long, parallel to each other, extend- ing along the whole length of the anther, adjacent anther wings parallel, in the same plane as the an- ther. Connective appendages 0.15-0.2 mm long, 0.14-0.16 mm wide, triangular, narrower than the stamen, p inflexed. Pollinarium: corpusculum 3 m long, between 1.5 times and twice as long as Sai rhomboid; caudicles 0.2 mm long, declinate, triangular; pollinia i .68 sessile. flattened, straight, subapically attached to the mm long, 0.25 mm wide, oblongoid, round in cross section. Stylar head 0.28—0.3 mm diam., 0.6-0.8 mm high; upper part 0.3—0.45 mm high, flat to de- pressed-conical to conical. Follicles one per flower, pendulous, ca. 90 mm long, 7.5 mm diam., fusi- form, apically acute, but not beaked, keeled, me- dium brown. Seeds and chromosome number un- known. caudicles, Distribution and habitat. Africa: Cameroon, Ga- bon, Guinea, Guinea-Bissau, Ivory Coast, Liberia, Nigeria, Sierra Leone, Uganda, Zaire; 250-1300 m forest, gallery forest, forest clearings and plantations. Not frequent, but very widespread. Figure 8. Flowering time. (May—)July to November. Selected seimen: examined. CAMEROON. South- west, around Kum 50 m, 16 May 1984, Thomas 3483 (MO); Bitye near River Ja, Sep. 1922, Bates 1856 (K). GABON. La Waka, Oubangui-Chari A.E.F., Glatemazé, 20 km NE Boubani, 1927, Le Testu 2285 (BM); Mbaïki, Bou- koko, 5 July 1948, Le Testu 1027 (BM); Woleu-Ntem, Min- voul, W Aug. 1923, Le Testu 234 (BM). GUINEA. Fá, 10 Sep. 1962, Guerra 3839 (K); Kinsan, Kindia, July 1937, Jacques- Félix 1810 (K, P). GUINEA-BISSAU. Bafatá, en- tre Geba e Mato de Cao, 15 Sep. 1955, Junta de Invest- gagóes coloniais 3370 (K). IVORY COAST. Bouaké, route de Sakasso, km 12, 250 m, 9 Aug. 1963, Garnier & Boua- ké 95 (K); Boukoka, 2 Aug. 1947, Tiss 102 (G). LIBERIA. Bong, Gbanga, 1926, Linder 653 (K); Nimba, Nimba Re- serve, Mount Bele Rd., 500 m, 10 Sep. 1964, Adames 512 (K). NIGERIA. West, Ife, Shasha forest Reserve, on track 26, 20 Sep. 1973, Latilo s.n. (67538 FHI, K). SIERRA LEONE. Mamaba, 2 Nov. 1914, Thomas 4553 (K); Njala, 8 Oct. 1949, Deighton 5188 (B, K); Ronietta, 250 m, 17 324 Annals of the Missouri Botanical Garden e 22. Cynanchum longipes N. E. Br. 1: Junta de investigações coloniais 3311; 2-5: Adames 512. —1. Habit with pea 'ence.—2.. Flower.—3. Gynostegium and corona, partially removed.—4. Pollinarium.—5. Stylar head. Drawn by Jim Conrad. Volume 83, Number 3 1996 Liede 325 Cynanchum in Africa Nov. 1914, Thomas 5377 (K). ZAIRE. Lilando, 15 Aug. 1938, Louis 10843 (K). Comments. Cynanchum longipes is a species of uncertain affinities, the red corona resembling the Malagasy C. papillatum alliance, but with the habit of a rather unspecialized species pl. 18. Cynanchum meyeri (Dec pod Schlechter, Bot. Jahrb. Syst. 20, Beibl. 8 noctonum meyeri Decne. in E Prodr. 8: 531. 1844. (nom. superfl. when published, based on S. ovatum E. Mey.). Vincetoxicum meyeri (Decne.) Kuntze, Revis. Gen. Pl. 2: 424. 1891. Sarcostemma ovatum E. Mey., Comm. Pl. Afr. Austr.: 216. 1838. Cynanchum ovatum (E. Mey.) Druce, Bot. Soc. Exch. Club Brit. Isles 4: 618. 1917, non Cynanchum ova- tum Thunb., Observ. Cynanch.: 6. 1821 (= Leptadenia). TYPE: Namibia. Garip, in colli- bus ad ostium fluminis, < 200 m, Oct., Drége s.n. (holotype, SAM sub SAM 18556; isotype, MO sub MO 2760941). Metaplexis mucronata Spreng., Neue Entdeck. Pflanzenk. 1: 269. 1820. Cynanchum mucronatum (Spreng.) N. We Br. in Dyer, Fl. Cap. 4(1): 745. 1908, non Cy- nchum mucronatum Andrews, Bot. Repos. 8: t. 515. 1808. TYPE: not known. Cynanchum pearsonii N. E. Br., Bull. Misc. Inform., Kew 1914: 18. 1914. TYPE: Namibia. Great Namaqua- land, Liideritz, N of Rotkuppe Station, 23 Feb. 1909, Pearson 4466 (holotype, K; isotype, BOL) Plants erect, 20-40 cm high, richly branched. Shoots perennial, 15-30 cm long, 4 mm diam., ob- scurely glaucous, woody with grayish bark, isolat- edly to densely covered with appressed trichomes 0.2-0.25 mm long, glabrescent. “Stipules” ovate, 5 mm long, 3 mm wide. Leaves with petioles 1-3.5 mm long; leaf blades coriaceous, 7-20 mm long, 4— 10 mm wide, ovate, basally rounded, without col- leters, apically acute, or obtuse and apiculate, adaxially isolatedly covered with appressed tri- chomes mm long, evenly distributed over the whole surface; abaxially glabrous. Inflorescences bostrychoid, 6-13-flowered, 2-8 flowers open at a time; rachis 1-2 mm long; peduncles 1-3 mm long, densely covered with appressed trichomes 0.15-0.2 mm long. Flowers with floral bracts 0.6-0.8 mm long, 0.7-0.9 mm wide at the base, triangular, with trichomes; pedicels 2-4 mm long, densely covered with appressed trichomes 0.15-0.2 mm long. Buds 0.8-1.2 mm long, 0.7-0.9 mm diam., conical; aes- tivation imbricate. Calyx basally fused, abaxially with trichomes; lobes 1-1.2 mm long, 0.5-0.6 mm This species is illustrated in Adam (1975: 971, 506) . wide, ovate, apically obtuse to acute. Corolla rotate to subglobose, fused for about % of their length, 1.5-2 mm long, abaxially glabrous, or with isolated trichomes, green to white; adaxially glabrous, green to white; lobes 0.5-1 mm wide, ovate, apically ob- tuse to acute, patent to recurved. Corona white, cy- athiform, 1-1.5 mm high, exceeding the gynoste- gium but not obscuring it; C(is) consisting of Cs and Ci for about half of total height, only Cs dif- ferentiated in shape. Cs without adaxial append- ages; lobes of Cs flat, triangular, erect, with straight margins. Gynostegium 0.7—0.75 mm high, 0.9-0.95 mm diam., sessile. Stamens with free filaments 0.2— 0.25 mm long, anthers broader than high, trape- zoidal, abaxially convex; anther wings 0.25-0.26 mm long, convergent, extending along the whole length of the anther; adjacent anther wings parallel, in the same plane as the anther. Connective ap- pendages 0.35-0.4 mm long, 0.22-0.25 mm wide, ovate, narrower than the stamens, erect. Pollinar- ium: corpusculum 0.12-0.13 mm long; caudicles .04—0.05 mm long, cylindrical, s-shaped, con- cave-convex; pollinia apically attached to the caudicles, 0.15—0.17 mm long, 0.05-0.06 mm wide, pyriform, elliptical in cross section. Stylar head 0.6—0.65 mm diam., 0.5—0.55 mm high; upper part 0.25-0.3 mm high, conical to depressed-con- ical. Follicles 30-35 mm long, 4—6 mm wide, ob- clavate, round in cross section, apically strongly beaked, gray, smooth, with isolated hairs. Seeds 4.5-5.5 mm long, 2.5—3.5 mm wide, pyriform, dark brown, seta and aseta side tuberculate, marginally wingless, entire; coma 15 mm long. Chromosome number unknown. Distribution. Africa: Namibia. Lüderitz; 0-200 m; flats and slopes; sand and rock crevices; fre- quently on granite. Comments. Further details, illustration, distri- bution map, and citation of specimens are provided in Liede (1993). 19. Cynanchum mossambicense K. Schumann in Engler, Pflanzenw. Ost-Afrikas C: 323. 1895. TYPE: Mozambique. Zambesia: Quilli- mane, Stuhlmann 643 (holotype, B presum- ably destroyed; lectotype, designated here, K) Cynanchum complexum N. E. Br., Bull. Misc. Inform., w 1895: 256, 337. 1895. TYPE: Mozambique. Mapanga, Oct. 1887, Scott s.n. (lectotype, designated here, K) Plants twining, richly branched. Shoots peren- nial, 1-1.5 mm diam., herbaceous, sparsely to densely covered with erect trichomes 0.5 mm long. 326 Annals of the Missouri Botanical Garden “Stipules” absent. Leaves with petioles 15-20 mm long; leaf blades herbaceous, 25-50 mm long, 15- 30 mm wide, ovate, basally cordate with 2 colleters in the adaxial sinus, apically acute and apiculate, apiculus 1-2 mm long, adaxially and abaxially nearly glabrous. Inflorescences sciadioidal, 5—12- flowered, all flowers open at a time; peduncles 8- 12 mm long, glabrous, or sparsely covered with ap- pressed trichomes 0.4 mm long. Flowers with floral bracts 1-1.2 mm long, 0.5-0.6 mm wide at the base, ovate, with trichomes; pedicels 5-7 mm long, glabrous. Buds 5.5-7 mm long, ca. 3.5 mm diam., conical to elongate-conical; aestivation slightly con- torted. Calyx basally fused, abaxially glabrous; lobes 1.2-1.5 mm long, 0.5-0.7 apically acute. Corolla rotate, fused at the base, 6— 7 mm long; abaxially and adaxially glabrous, white: lobes 1.5-2 mm wide, spreading to recurved, ob- mm W ide, ovate, long, apically obtuse, twisted. Corona white, tubular, 5—6 mm high, exceeding the gynostegium, partly obscuring it; C(is) consisting of Cs and Ci fused for little more than half of total corona length; Cs and Ci differentiated in shape; Ci as long as Cs, thinner than Cs. Cs appressed to the back of the stamens, with adaxial appendages; lobes of Cs flat, triangular, erect, with laterally involute margins; adaxial appendages shorter than Cs, erect, liguli- form. Lobes of Ci basally flat, elongatedly triangular, apically filamentose, twisted, erect, with laterally involute margins. Gynostegium 1.8-1.9 mm high, 1.8-1.9 mm diam., anthers higher than broad, sessile. Stamens without free filaments; trapezoid, abaxially planar; anther wings convergent, 0.9—0.95 mm long, extending beyond the anther proper with stamens basally slightly arched; adjacent anther wings parallel, centrifugal, outer guide rail smooth. Connective appendages 0.75—0.8 mm long, 0.47— 0.5 mm wide, triangular, equaling the stamen in width, slightly inflexed. Pollinarium: corpusculum 0.3-0.35 mm long; margins of the corpuscular cleft sinuate; caudicles 0.05—0.06 mm long, flattened, straight, horizontal to declinate, triangular; pollinia subapically attached to the caudicles, 0.42—0.45 mm long, 0.2-0.23 mm wide, ovoid to oblong, ovate in cross section. 0.8—0.85 mm diam., 1-1.1 mm high; upper part 0.4—0.45 mm high, bi- furcate. Follicles 55—70 mm long. 4—5 mm diam., Stylar head O obclavate, obtusely deltate in cross section, apically shortly beaked, keeled, medium brown, longitudi- nally grooved, glabrous. Seeds 5.5—6 mm long, 4 mm wide, ovate, medium brown, seta and aseta side sculptured with longitudinal ridges, marginally with 0.3-0.35 mm wide wing with entire margin; coma 15-20 mm long. Chromosome number unknown. Distribution and habitat. Africa: Mozambique, South Africa (Transvaal), Swaziland, Zimbabwe; 0— 1000 m; scrub, riverine forest. Not frequent, but probably on sand and over rocks, coastal dune not endangered. Figure Flowering time. March to October. Selected specimens examined. MOZAMBIQUE. In- bons O m, 14 Oct. 1906, ur s.n. (K); Manica, near aes 1000 m, 16 June Pole Evans 5226 (K); Sofala, oe Oct. 1887, ul . (K); Zambesia, N of), 20 Aug. 1962, Wild 5882 (K, . SOU TH AFRICA. Transvaal: Acornhoek, a ae Pretorius Kop, on banks of Sabi near Hippo Pool, Oct. 1931, Letty 91 (PRE, SRGH). eh ZILAND. Stega. a zi River, near Ranches, ca. 330 1 25 July 19: 98, Compton 27923 (NBG, PRE). ZIMBABWE. Chiredzi, Gona-re-zhou, 2 km from Chipindas Pools on N bank of Lundi River, on edge of research Officer’s garden, 28 May 1971, Grosvenor 554 (K, SRGH); Darwin r upper reaches of Nyatandi river, 900 m, 27 Jan. 1960, Phipps 2426 (K, SRGH); Ndanga, Sabi-Lundi-Junction, Chitsis Kraal, 270 m, 6 June 1950, Wild 3393 (K, SRGH). Comments. Cynanchum mossambicense, a spe- cies restricted to a rather small area in southeastern Africa, is probably the closest African relative of C. acutum. Lectotypification of C. mossambicense is neces- sary because the original material was destroyed in B. The isotype in K is chosen as lectotype. N. E. Brown cited two specimens (Kirk s.n. and Scott s.n., both K) as syntypes for C. complexum. The better specimen, Scott s.n., Illustration and additional information on south- is chosen as lectotype. ern African material are provided in Liede (1993). 20. Cynanchum natalitium Schlechter, Bot. Jahrb. Syst. 18, Beibl. 45: 32. 1894. TYPE: South Africa. Natal: Durban, 14 Aug. 1893, Schlechter 3082 (lectotype, designated by Liede (1993), BOL; isolectotype, GRA). Plants twining, to 30 cm high, richly branched, sarmentose, adventitious roots formed along the whole lower surface of the runner; subterranean or- gans consisting only of fibrous roots. Shoots peren- nial, 30-150 cm long, 1-2 mm diam., herbaceous, isolatedly glabrescent with appressed trichomes 0.25—0.3 mm long; in old plants basally corky, then bark yellowish. “Stipules” ovate, 5-6 mm long, 3- 4 mm wide. Leaves with petioles 7-25 mm long; leaf blades fleshy, 20-50 mm long, 15-35 mm wide, ovate to orbicular, or obovate, basally round- ed, or cuneate with 1 colleter adaxially, apically obtuse and acuminate, acumen 5 mm long, glabrous, or isolatedly covered with appressed tri- chomes 0.35—0.5 mm long, evenly distributed over the whole surface. Inflorescences sciadioidal (some- Volume 83, Number 3 1996 Liede Cynanchum in Africa times two sciadioids in dichasial arrangement), 6— 16-flowered, cles 5-15 mm long, glabrous, or isolatedly covered with appressed trichomes 0.25-0.35 mm long. Flowers sweetly scented; floral bracts 1-1.5 mm long, 0.5-1 chomes; pedicels 5-10 mm long, isolatedly covered with appressed trichomes 0.3-0.35 mm long. Buds owers open at a time; pedun- mm wide at the base, ovate, with tri- 4-5 mm long, 2-3 mm diam., elongated-conical; aestivation imbricate. Calyx basally fused, abaxi- ally with trichomes; lobes 0.8-1.2 mm long, 0.6-1 mm wide, triangular, apically acute. Corolla rotate, basally fused 3.5-5 mm ially glabrous, dull green to brown; lobes 1.5-2 mm long, abaxially and adax- wide, spreading to patent, oblong, apically obtuse. Corona white, urceolate, 3.54. mm high, exceeding the gynostegium but not obscuring it; C(is) consist- ing of Cs and Ci completely fused, upper margin deeply 5-crenate, margins of sinuses recurved; Cs without adaxial appendages. Gynostegium 0.8—1.2 mm high, 1.5-1.8 mm diam., atop a stipe 1-1.5 mm long. Stamens without free filaments; anthers about as high as broad, trapezoidal, abaxially con- vex; anther wings 0.23-0.25 mm long, not extend- ing along the whole length of the anther, adjacent anther wings parallel, basally widened, in the same plane as the anther. Connective appendages 0.5— 0.55 mm long, 0.4-0.45 i than the stamen, slightly inflexed. Pollinarium: cor- pusculum 0.3-0.32 mm long; margins of the cor- puscular cleft divergent toward the base; caudicles 0.1-0.12 mm long, flattened, straight, declinate, tri- angular; pollinia subapically attached to the cau- dicles, 0.4—0.45 mm long, 0.15—0.17 mm wide, cla- vate, elliptical in cross section. Stylar head 0.5— 0.55 mm diam., 0.3-0.35 mm high, upper part ca. 0.1 mm high, flat. Follicles 40-45 mm long, 5-8 mm wide, obclavate, obtusely deltate in cross sec- tion, apically strongly beaked, keeled to winged with 0.5-1 mm broad wing, medium brown, longi- tudinally grooved, with isolated trichomes. Seeds mm long, 3-3.5 mm wide, ovate, light to mm wide, ovate, narrower medium brown, seta and aseta side sculptured with longitudinal ridges, marginally with indistinct, 0.5 mm wide wing with entire margin; coma 15-17 mm long. Chromosome number unknown. Distribution and habitat. Africa: South Africa (Cape Province, Natal); 0-60 m; on littoral dunes, dune forest; full sun to light shade. Comments. Further details, illustration, distri- bution map, and citation of specimens are provided in Liede (1993). 21. Cynanchum obtusifolium Linnaeus f., Suppl. Pl. 169. 1784. Vincetoxicum obtusifol- ium (L. f.) Kuntze, Revis. Gen. Pl. 3, 2: 200. 1898. TYPE: Thunberg s.n. (holotype, UPS 6311, UPS; seen on IDC microfiche). puo pa Meisn., J. Bot. 2: 546 ("446"). 1843. Nom. superfl. when published, substitute name for adan obtusifolium i Plants twining, 2-3 m high, richly branched; subterranean organs consisting only of fibrous roots. Shoots perennial, 50-200 mm diam., herbaceous, glabrous, or sparsely to densely covered with erect trichomes 0.4-0.7 mm long, ba- sally woody with yellowish bark. “Stipules” ovate, almost round, 5-7 mm long, 6-8 mm wide. Leaves with petioles 5-15 mm long; leaf blades herbaceous to thinly coriaceous, 20-40 mm long, 1 mm wide, ovate to elliptic, basally rounded to subtrun- cate, or cordate, with 2—4 colleters in the adaxial sinus, apically obtuse and acuminate, or obcordate and acuminate, marginally entire or crenulate, adaxially and abaxially glabrous, or with a sparse indumentum concentrated on veins and margins; trichomes appressed, 0.3-0.45 mm long. Inflores- cences bostrychoid, 8—15-flowered, 3-6 flowers open at a time; rachis 1-3 mm long; peduncles 3- 5 mm long, densely covered with erect trichomes 0.5-0.7 mm long. Flowers sweetly scented; floral bracts 1-2.5 mm long, mm wide at the base, ovate, with trichomes; pedicels 3.5-6 mm long, densely covered with appressed trichomes 4—0.7 mm long. Buds 4-5 mm long, 2-3 mm diam., elongated-conical; aestivation imbricate. Ca- lyx basally fused, abaxially with trichomes; lobes 5-2 mm long, 0.8-1.2 mm wide, ovate, apically acute to acuminate. Corolla rotate, fused at the base, 3-4 mm long, abaxially glabrous, green, adaxially with verrucose trichomes, green; lobes 1— 1.5 mm wide, patent, ovate to oblong, apically ob- tuse. Corona white, cyathiform, 2-2.5 mm high, ex- ceeding the gynostegium but not obscuring it; C(is) consisting of Cs and Ci fused between for about Y of total corona length; Cs and Ci differentiated in shape, Ci shorter than Cs, dorsally connate to Cs. Cs without adaxial appendages; lobes of Cs flat, tri- fid (the two lateral teeth much smaller than the middle one), inflexed, with laterally involute mar- gins. Lobes of Ci flat, rectangular, erect, with straight, emarginate margins. Gynostegium 1.5-2 mm high, 1.5-2 mm diam., sessile. Stamens without free filaments, anthers about as broad as high, trap- ezoidal, abaxially convex; anther wings 0.5-0.55 mm long, extending beyond the anther proper, which does not form a basal arch; adjacent anther cm long, 1-1.5 328 Annals o Missouri Botanical Garden wings parallel, basally widened. Connective ap- pendages 0.25-0.3 mm long, 0.5-0.55 mm wide, depressed ovate, narrower than the stamen, slightly inflexed. Pollinarium: corpusculum 0.25-0.27 mm long; margins of the corpuscular cleft sinuate; cau- dicles 0.12-0.13 mm long, flattened, straight, de- clinate, triangular; pollinia subapically attached to the caudicles, 0.375—0.4 mm long, 0.16—0.18 mm wide, ovoid. Stylar head 1-1.1 mm diam., 0.7-0.75 mm high; upper part 0.5-0.55 mm high, upper part depressed-conical and umbonate. Follicles mm long, 10-12 mm wide, obclavate, obtusely del- tate in cross section, apically obtuse, keeled, me- dium to dark brown, longitudinally grooved, sparse- ly to densely covered with trichomes. Seeds 6-7 mm long, 3.5—4 mm wide, pyriform, dark brown, seta and aseta side sculptured with longitudinal ridges and papillose, marginally with 1.0—1.1-mm-wide wing with entire margin; coma 30 mm long. Chro- 22 (voucher: Liede 2924, mosome number: 2n = Distribution and habitat. Africa: Mozambique, South Africa (Cape Province, Natal); 0-250 m; dunes; mostly on sand; dune scrub and coastal veg- etation; frequently in disturbed habitats; full sun to partial shade. Comments. Periploca africana L. var. B L. [Sp. Pl. ed. 1: 211, 1753, illustrated with Burman, Rar. Afr. Pl.: p. 34, t. 14, fig. 2. 1738] also represents this taxon urther details, illustration, distribution map, and citation of specimens are provided in Liede (1993). 22. Cynanchum orangeanum (Schlechter) N. E. Brown in Dyer, Fl. Cap. 4(1): 745. 1908. Flan- agania Veios Schltr., Bot. Jahrb. Syst. 18, Beibl. 4 . 1894, TYPE: South Africa. Or- ange A a Colesberg, Orange River near Bethulie, 1330 m, Dec. 1892, Flanagan 1502 (lectotype, designated by Liede (1993), SAM; isolectotype, BOL) Plants erect, 10-20 cm high, basally sparsely branched; subterranean organs rhizomatous, rhi- zome 5—7 mm diam. Shoots 15-20 cm long, 1-1.5 mm diam., herbaceous, densely to sparsely covered with erect trichomes 0.25-0.27 mm long, basally woody with grayish bark. “Stipules” absent. Leaves sessile; leaf blades herbaceous, 30-50 mm long, —1.5 mm wide, linear, basally decurrent, without colleters, apically acute, or obtuse, marginally straight and thickened, adaxially and abaxially gla- brous, or isolatedly covered with appressed tri- chomes 0.15—0.2 mm long, restricted to the veins and margins. Inflorescences sciadioidal, 1—5-flow- ered, all flowers open at a time (one extraordinary umbel seen with 16 flowers, Hardy 6583, peduncles 0—5 mm long, isolatedly to sparsely cov- ered with erect trichomes 0.2-0.25 mm long. Flow- ers with floral bracts 1—1.5 mm long, 0.3—0.5 mm wide at the base, linear, with trichomes; pedicels 3-7 mm long, sparsely to densely covered with ap- pressed trichomes 0.25—0.3 mm long. Buds 2.54 mm long. 2-3 mm diam., depressed-conical; aes- tivation imbricate. Calyx basally fused, abaxially with trichomes (prominently so along the midrib); lobes 2-2.5 mm long, 0.8-1.2 mm wide, ovate, api- cally acute. Corolla rotate, basally fused, 4-6 mm long, abaxially glabrous (or with a few isolated tri- chomes), brown, adaxially glabrous, brown; lobes 1.5-2 mm wide, incurved, ovate to lanceolate, api- cally acute, Corona white, cyathiform, 4—5 mm high, shorter than the gynos- tegium; C(is) consisting of Cs and Ci fused for about Y of total height, both Cs and Ci differentiated in shape, Ci as long as Cs. Cs appressed to the back of the stamens; lobes of Cs flat, oblong, reflexed, without adaxial appendages. Lobes of Ci filamen- margins revolute. tous, reflexed, with straight margins. Gynostegium 1.5-2 mm high, 1.5-2 mm diam., without free filaments; anthers broader than high, rectangular, abaxially convex; anther wings 0.4—0.5 mm long, convergent, extending along the whole length of the anther; adjacent anther wings parallel, in the same plane as the anther. Connective ap- pendages 0.5-0.55 mm long, 0.65—0.7 mm wide, widely ovate, narrower than the stamen, erect. Pol- sessile. Stamens linarium: corpusculum 0.2-0.25 mm long; margins of the corpuscular cleft parallel; caudicles 0.15— 0.175 mm long, cylindrical, s-shaped, concave-con- vex, thickened at the insertion of the pollinium; pollinia subapically attached to the caudicles, 0 0.35 mm long, 0.12-0.15 mm wide, ovoid, round Stylar head 0.8-0.85 mm diam., 0.9-1 mm high; upper part 0.7-0.75 mm high, cap- itate. Follicles 40-70 mm long, 4-7 mm wide, fu- siform, round in c in cross section. ross section, apically strongly beaked, light brown, smooth, glabrous. Seeds 5—6 mm long, 3—3.5 mm wide, ovate, dark brown; seta and aseta side tuberculate, marginally with 0.6- mm-wide wing with entire margin; coma 15-20 mm long. Chromosome number unknown. Distribution and habitat. Africa: Botswana, Na- mibia, South Africa (Cape Province, Orange Free State), Zimbabwe; 1000-1500 m; on flats, in sand, frequently red Kalahari sand, between rocks, short grassland; full sun. Figure 23 Volume 83, Number 3 1996 Liede Cynanchum in Africa 329 20 0 2 ¿E A / j \ 30 E \ To VA d SEX É 20 l N " | 10 10 i e y] 0 | 10 lio NR | ; == lo \ a o o e A a D lao 30 O 100 200 200 400 $00 800 MILES a tex A bi sb "I Figure 23. Known distribution of C. orangeanum (open circles) and C. praecox (dots). Comments. Cynanchum orangeanum is closely Mussengue, Jan. 1870, Welwitsch 4222 (holotype, K; related to C. praecox, but is otherwise isolated among African Cynanchum. Further details, illustration, and citation of spec- imens are provided in Liede (1993). 23. Cynanchum polyanthum K. Schumann in Engler & Prantl, Nat. Pflanzenfam. 4(2): 253. 1895. TYPE: *Im Land der Monbuttu bei Manza," Schweinfurth 3345 (holotype, B pre- sumably destroyed; lectotype, designated here, K). Figure 24. Vincetoxicum polyanthum K. Schum., Bot. Jahrb. Syst. 17: 136. 1893, non Vincetoxicum pa es thum Kuntze, Revis. Gen. Pl. 2: 424. 1891, replacement name for Tylophora floribunda Miq. Cynanchum obscurum K. Schum. in Engl. & Prantl, Nat. Pflanzenfam. 4(2): 253. 1895. TYPE: Angola. Cuan- za Norte, Golungo Alto, Ad dumeta in Sobato de isotype, BM). Cynanchum welwitschii Schltr. & Rendle, J. Bot. 34: 99. 1896 K. Schum. (Art. 52.1 ICBN, see Greuter et al., dio batesii Wernham, J. Bot. 54: 228. 1916. TYPE: Cameroon. Bitye, 2 Dec. 1914, Bates 643 (holotype, Plants ascending, twining, 2.5—3 m high, sparse- ly and irregularly branched; subterranean organs woody rootstocks. Shoots perennial, herbaceous, sparsely covered with flexuous trichomes 0.5-0.6 mm long, along two lines; internodes 7-10 cm long, 0.5-2 mm diam. “Stipules” absent. Leaves with pet- ioles 15-50 mm long; leaf blades herbaceous, 55— 100 mm long, 35-55 mm wide, ovate, basally lo- bate, lobes 10-15 mm long, with 3-5 colleters in the adaxial sinus, apically acuminate, acumen 1-2 330 Annals of the Missouri Botanical Garden Figure 24. o oe polyanthum K. Schum.—1. Habit with inflorescence (De n 11112) and fruit (Gossweiler 4866). 2-5: De Witte 11112.—2. Flower and corona, two corolla lobes removed.—3. Gynostegium and corona, partially removed.—4. ae Stylar coa Seed, seta side (Synnott 688). eo by Jim Conrad. Volume 83, Number 3 1996 Liede 331 Cynanchum in Africa mm long, adaxially isolatedly covered with flexuous trichomes 0.55—0.6 and margins, abaxially sparsely covered with flex- mm long. Inflorescences bostrychoid to sciadioidal, 8—15-flowered, 6-10 flowers open at a time; rachis to 7 mm long; pe- duncles 35-75 mm long, sparsely covered with ap- pressed trichomes 0.5-0.6 mm long. Flowers with floral bracts 1.2-1.4 mm long, 0.4—0.8 mm wide at the base, ovate, with trichomes; pedicels 12-35 mm long, densely covered with erect trichomes 0.4—0.5 mm long. Buds 5—6 mm long, 2.5-3 mm diam., conical, basally with imbricate, apically with con- torted aestivation. Calyx basally fused, abaxial sur- face with trichomes; lobes 0.8-1 mm long, 0.6—0.7 m wide, ovate, apically acute. Corolla rotate, ba- sally fused; 7-9 mm long, abaxially greenish yel- low, adaxially greenish yellow to purple, with a very few isolated trichomes at the sinuses; lobes 1.5-2 mm wide, horizontal, lanceolate to oblong, apically obtuse. Corona white, 6—7.5 mm high, exceeding the gynostegium, partly obscuring it; C(is) consist- ing of Cs and Ci fused for more than ?4 of total corona length, Cs and Ci differentiated, Ci shorter than Cs. adaxial appendages; lobes of Cs laminar, elongated- triangular, apically erect. Lobes of Ci laminar, bifid, mm long, restricted to veins uous trichomes Cs not adnate to the filaments, without producing a pronounced convex fold along the up- per half of corona length, reflexed. Gynostegium 2— 2.5 mm high, 2-2.5 without free filaments; anthers broader than high, trapezoidal, abaxially planar; anther wings 2 mm mm diam., sessile. Stamens long, convergent, extending along the whole length of the anther, adjacent anther wings parallel, in the same plane as the anther. Connective appendages 0.5—0.6 mm long, 0.5-0.6 mm wide, obovate, nar- rower than the stamen, slightly inflexed. Pollinar- ium: corpusculum 0.3-0.35 mm long, ovoid; cau- dicles 0.1 mm long, flattened, straight, declinate, trapezoid; pollinia subapically attached to the caudicles, 0.3-0.35 mm long, 0.12 mm wide, ob- longoid, ovate in cross section. Stylar head 0.7— 0.85 mm diam., 0.7-0.85 mm high; upper part 0.4— 0.5 mm high, capitate. Follicles usually one per flower, pendulous, 80-90 mm long, 7.5-8 mm diam., obclavate, obtusely deltate in cross section, apically strongly beaked, keeled, with isolated in- dumentum. See m long, 4.3 wide, ovate, medium brown, seta and aseta side sculptured with longitudinal ridges, marginally with 1-mm-wide wing with dentate margin; coma 12-15 mm long. Chromosome number unknown. Distribution and habitat. Africa: Angola, Cam- eroon, Gabon, Uganda (U3, U4), Zaire; 1200-1600 m, moist grassland, forest margins, thickets. Very widespread, but not frequent. Figure Flowering time. September to May. Vernacular name. Molo-Busyo (Lissango). Selected specimens examined. OLA. Cuanza Nor- te, Golungo alto, Nov. 1878, Welwitsch 4200 (G, K). CAM- EROON. Yaunde, Zenker 223 (K). GABON. Mbaiki, Bou- koko 27 June 1949, Le Testu 1510 (BM). UGANDA. Buganda, Mengo, Mutingo, edge of Lake Victoria, a few miles from Kampala, 1250 m, Dec. 1935, Chandler 1496 (K); Masaka, 4 mi. from Masaka on rd. to Bukakata, 1150 1953, Drummond & Hemsley 4736 (K); East- ern, Busoga, Bugabula, Musumu swamp at crossing of Jin- ja-Kamuli rd., 10 mi. S of Kamuli, ps m, 27 May 1953, Wood 760 (K); Western, Mubende, Singo, 1-2 mi. SE of Eo 1200 m, 16 Mar. 1969, Lye 2354 (K). ZAÍRE. eni, Mutsora, Parc National Albert, 1200 m, 26 Mar. 1955, De Witte 12092 (K); Yangambi, 21 May 1938, Louis 9479 (K); Rutshuru, plaine route de Djombo, 28 Mar. 1937, Ghesquiere 3918 (K, SRGH). Comments. Cynanchum polyanthum K. Schum. can be treated as a new species, because the name was published without citing the earlier Vincetoxi- cum (Cynoctonum) polyanthum K. Schum. as bas- ionym; the latter represents a later homonym to Vin- cetoxicum polyanthum Kuntze. The affinities of C. polyanthum remain obscure. Its closest relatives are probably C. heteromorphum and C. falcatum from Ethiopia. 24. Cynanchum praecox Schlechter ex S. Moore, J. Bot. 40: 256. 1902. TYPE: Zimba- bwe. Harare, valley of Mazoe River, Sep. 1898, Rand 512 (holotype, BM). Figure 25. iE pygmaeum Schltr., Bot. n us 5l: iu 1913. TYPE: Cameroon. Bamen a. 3 mi. fro Kubo along Oku rd., 1850 m, 15 Feb. 1958, Hep- per 2011 (neotype, i so here, K); syntypes cit- ed in the protologue mann 2226, 2230 (both probably destroyed in B, no isotypes foun Plants erect, nontwining, 3—10 cm high, unbran- ched, with rhizomes 1.5-3 mm diam. Shoots her- baceous, densely covered with erect trichomes, 0.35-0.4 mm long, along a single line; internodes 3 mm diam. “Stipules” absent. Leaves absent at the time of flowering, sessile, leaf blades herbaceous, 40—60 mm long, 2-8 mm wide, linear to elliptic to ovate, basally decurrent, without colleters, apically acute, marginally straight and thickened, adaxially and abaxially glabrous. Inflorescences supported by inflorescence bracts different from the vegetative leaves, bostrychoid, 5-15-flowered, all flowers open at a time; rachis 1-2 mm long; inflorescence bracts ca. 0.9 mm long, 0.5 mm wide, ovate, with apicu- ciliate; peduncles sparsely covered with erect trichomes 0.35—0. late apex, mm long, mm 332 Annals of the Missouri Botanical Garden 0,2 mm Figure 25. Cynanchum Dab n Schltr. ex Moore . Habit (Hepper 2011). 2-6: Drummond 4893.—2. Inflores- cence with young fruits.—3. Flower, corolla partially md —A. Gynostegium and corona, partially removed.—5. Pollinarium.—96. Stylar head. Drawn by Jim Conra Volume 83, Number 3 1996 Liede 333 Cynanchum in Africa long, along a single line. Flowers with a musky scent; floral bracts 0.5-0.7 mm long, 0.2-0.3 mm wide at the base, triangular, with trichomes; pedi- cels 5-15 mm long, densely covered with erect tri- chomes 0.4—0.45 mm long. Buds 6—6.5 mm long, 3—3.5 mm diam., elongated-conical, with imbricate aestivation, dextrorse. Calyx basally fused, ciliate; lobes 1.5-2.5 mm long, 0.7-0.8 mm wide, trian- gular, apically acute. Corolla rotate, fused for about V. of total length; 5.5—6.5 mm long, abaxially and adaxially yellowish brown; lobes 0.5-1 mm wide, incurved, oblong, da acute, with revolute mar- gins. Corona white, m high, exceeding the gynostegium but not cio it; C(is) consisting of Cs and Ci fused for more than half of total corona length, Cs and Ci differentiated, Ci as long as Cs, urceolate to campanulate. Cs not adnate to the fil- aments, appressed to the back of the stamens, with- out adaxial appendages; lobes of Cs laminar, ovate, apically erect, with laterally slightly involute mar- gins. Lobes of Ci laminar, oblong, reflexed, with straight margins. Gynostegium 1-1.1 mm high, 0.8— 0.9 mm diam., Stamens without free fila- anthers broader than high, trapezoidal, abaxially planar; anther wings 0.4 mm long, parallel to each other, extending beyond the anther proper, stamens forming a basal arch, adjacent an- ther wings parallel, centrifugal. Connective ap- pendages 0.4—0.45 mm long, 0.35—0.4 mm wide, ovate, equaling the stamen in width, erect, with emarginate margins. Pollinarium: corpusculum .2-0.25 mm long, ovoid, margins of the corpus- cular cleft sinuate; caudicles 0.05—0.06 mm long, flattened, straight, horizontal, rectangular; pollinia laterally attached to the caudicles, 0.22-0.25 mm long, 0.12-0.13 mm wide, ovoid, round in cross section. Stylar head 1-1.2 mm diam., 1.5-1.8 mm high; upper part 1-1.2 mm high, capitate. Mature fruits, seeds, and chromosome number unknown. sessile ments; Distribution and habitat. Africa: Cameroon, Malawi, Nigeria, Sierra Leone, Tanzania (T1, T4, T7), Zaire, Zambia, Zimbabwe; 1500-2500 m; burnt savanna and grasslands. Rare, but very wide- spread. Probably undercollected because of its small size. Figure 23 Flowering time. August to February. Selected dpi examined. CAMEROON. Ba- menda, Kumbo, ca. 3 mi. along Oku rd., 1850 m, 15 Feb. 1958, Hepper ae (K). MALAWI. North, Mage Vipya Plateau, 3 mi. in Vipya link road, 1900 25 Sep. 1972, Pawek 5814 (K, MAL, MO); Rumpi, Nyika Plateau, Chelinda bridge, 1500 m, 10 Sep. 1976, dg ek 11793 (K, MO). NIGERIA. Ca. 5 km N of Lana Feb. 1941, Milne-Redhead 5030 0o. SIERRA LEONE. Kabala, Loma Mountains, 8 Jan. 1966, Adam 22989 ane (MO). TANZANIA. Lake. Bukoba, Bugene, 1660 m, 21 July 1947, Ford 190 (K). Southern Highlands. Rungwe, Mbogo Mtn., 2250 m, 7 Nov. 1966, Gillett 17641 (EA). Western. Buha, Kibondo, 1500 m, 7 Aug. 1950, Bullock 3096 (K); Ufipa, Mbisi, Ufipa Plateau, 2500 m, 6 Oct. 1950, Bullock 3419 (K). ZAIRE. Garamba-Bagbele parc ay Uele, 3 Feb. 1950, De Saeger 106 g (K). A. Abercorn, Kawimbe, ca. 3 mi. from on Kara rd., 1500 m, 29 Aug. 1956, Richards 6027 (K). ZIM- BABWE. Harare, Makabusi woodlands, upper Chiravra R. near entrance gate, 10 Sep. 1981, Best 1647 (MO). NS > Comments. Both syntypes of Cynanchum pyg- maeum have been destroyed in B, and duplicate specimens, or any specimens annotated “C. pyg- maeum” in Schlechter’s handwriting, were not found. Therefore, the name has to be neotypified. Bulllock (1953) presented a detailed account on C. praecox and established the synonym of C. prae- cox and C. pygmaeum. The closest relative of C. praecox clearly is C. orangeanum, which is found south of the area of C. praecox. The affinities of these two sister species, however, remain obscure. 25. Cynanchum rubricoronae Liede, sp. nov. TYPE: Somalia. Nugaal: Aska, near Las Anod, 28 Oct. 1944, Glover & Gilliland 199 (holo- type, K). Figure 26. Plantae habitu agio C. crassiantherae. Differt struc- egialis roseae succulentis, partibus oblo a NN partibus interstaminali- bus liguliformibus, reflexisq Plants erect, nontwining, 20—40 cm high, sparse- ly basicaulously branched. Shoots herbaceous, warty (Kuchar 16793), glabrous; internodes 20—40 cm long, 1-1.5 mm diam. “Stipules” ovate, 10-12 mm long, 6—7 mm wide. Leaves with petioles 15— 20 mm long; leaf blades fleshy, 25-40 mm long, 15-24 mm wide, triangular, basally lobate to auric- ulate, lobes 7-10 mm long, without colleters, api- cally acuminate, mm long, marginally thickened, crenulate, adaxially and abaxially glabrous. Inflorescences 10—20-flowered, all flowers open at a time, bostrychoid to sciadioi- dal; peduncles 0-2.5 mm long, sparsely covered with flexuous trichomes 0.4-0.5 mm long. Flowers with floral bracts 1—1.2 mm long, 0.2-0.3 mm wide at the base, linear, with trichomes; pedicels 5—7 mm long, glabrous. Buds 3-3.5 mm long, 1.5-1. acumen 0 mm diam., conical, with imbricate aestivation. Ca- lyx basally fused; abaxial surface glabrous; lobes 1.6-1.7 mm long, 0 cally acute. Corolla rotate, basally fused; 3—3.5 mm long, abaxially and adaxially cream; lobes 0.7-0.8 mm wide, incurved, ovate, apically acute to apically obtuse. Corona basally purplish red, fading to white 4—0.5 mm wide, triangular, api- 334 Annals of the Missouri Botanical Garden Figure 26. Cynanchum rubricoronae Liede. Glover & Gilliland 190.—1. Shoot with inflorescence.—2. Flow- er.—3. Staminal corona lobe (center) and ME interstaminal lobes.—4. Gynostegium and corona, partially re- moved.—5. Pollinarium.—6. Stylar head. Drawn by G. Hintz Volume 83, Number 3 1996 Liede 335 Cynanchum in Africa toward the apex, 1.8-2 mm high, exceeding the gy- nostegium, partly obscuring it; C(is) consisting of Cs and Ci fused for a little more than % of total corona length, Cs and Ci differentiated, Ci shorter than Cs, thicker than Cs. Cs not adnate to the fil- aments, apically appressed to the back of the sta- mens, without adaxial appendages; lobes of Cs lam- inar, oblong, apically erect, with straight margins. Lobes of Ci solid, massive, reflexed, flatly lingulate. Gynostegium 0.8—1 mm high, 0.8-1 mm diam., ses- sile. Stamens without filament, anthers broader than high, trapezoidal, abaxially planar; anther wings 0.3-0.4 mm long, convergent, very vaguely differ- entiated, extending along the whole length of the anther; adjacent anther wings parallel, in the same plane as the anther, basally forming a distinct “mouth” with the basal lateral margin of the anther. Connective appendages 0.3-0.4 mm long, 0.4—0.5 mm wide, ovate, equaling the stamen in width, slightly inflexed, with emarginate margins. Polli- narium: corpusculum 0.14—0.16 mm long, ovoid; caudicles 0.12-0.13 mm long, flattened, concavely recurved, trapezoid; pollinia laterally attached to the caudicles, 0.26—0.28 mm long, 0.1—0.12 mm wide, oblongoid, ovate in cross section. Stylar head 0.3—0.4 mm diam., 0.17-0.2 mm high; upper part ca. 0.05 mm high, umbonate. Follicles one per flow- er, pendulous, ca. 70 mm long, 7 mm diam., ob- clavate, apically strongly beaked, wingless, light brown, with dark brown mottling, smooth, glabrous. Seeds and chromosome number unknown. Distribution and habitat. Africa: Somalia; ca. 300 m; sandplains, hills with shrubland. From the scanty documentation in herbaria, the species ap- pears to be very rare and is probably endangered. Figure 12. Flowering time. May, October. Additional specimens examined. SOMALIA. Hiiran, Bulo Burte, 23 km from Yasoomman along rd. to Maxaas, then SE 3% km along cutline, then NE 2 km along cutline, 285 m, 1 May 1985, Kuchar 16793 (K); Noogaal, 124 km NW of Eil, on the road to Gardo, 5 Jan. 1973, Bally & Melville 15561 (K, MO). Comments. Cynanchum rubricoronae is a very distinctive plant with quite attractive flowers. Its closest relative is probably the Somalian C. cras- siantherae, with which it shares vegetative charac- ters. However, its corona morphology is very dis- tinct and is unique in the genus. The two paratypes cited are rather poor (Bally & Melville 15561 is completely without flowers) and are assigned to C. rubricoronae only tentatively. 26. Cynanchum rungweense Bullock, Kew Bull. 10: 622. 1956. TYPE: Tanzania. Mbeya: Mbeya, below Mporoto, Inkuyu, 17 Mar. 1932, St. Clair-Thompson 846 (holotype, K). Figure 27. Plants ascending, twining, 5-6 m high, richly and irregularly branched. Subterranean organs woody rootstocks. Shoots perennial, herbaceous, sparsely covered with appressed trichomes 0.4—0.5 mm long, along a single line, glabrescent; inter- nodes 3—6.5 cm long, 0.8-1 mm diam. “Stipules” widely ovate, 4-6 mm long, 6-8 mm wide, acumi- nate. Leaves with petioles 15-30 mm long; leaf blades herbaceous, 40-50 mm long, 14-20 mm wide, ovate, basally cordate to lobate, lobes 4—6 mm long, with 2 colleters in the adaxial sinus, api- cally acute, adaxially dark green, sparsely covered with appressed trichomes mm long, re- stricted to veins and margins, abaxially much paler green, papillose, glabrous. Inflorescences bostry- choid to sciadioidal, 12—16-flowered, 6-9 flowers open at a time; rachis to 2 mm long; peduncles 10— 30 mm long, glabrous to densely covered with erect trichomes 0.2-0.3 mm long, along a single line. Flowers with floral bracts 1.3-1.4 mm long, 0.4— 0.5 mm wide at the base, elongate-triangular, api- cally glandular, with trichomes; pedicels 11-15 mm long, densely covered with flexuous trichomes 0.3— 0.35 mm long. Buds 5-7 mm long, 3—4 mm diam., conical, with imbricate aestivation. Calyx basally fused, abaxial surface glabrous; lobes 1.6-1.8 mm long, 1.2-1.4 mm wide, ovate, apically obtuse to acute. Corolla rotate, basally fused; 6-7 mm long, abaxially green, adaxially greenish brown; lobes 2.5-3 mm wide, horizontal, oblong, apically obtuse. Corona urceolate, white, 4.5-5 mm high, exceeding the gynostegium, partly obscuring it; C(is) consist- ing of Cs and Ci fused for more than % of total corona length, Cs and Ci differentiated, Ci slightly longer than Cs. Cs not adnate to the filaments, with- out adaxial appendages; lobes of Cs laminar, tri- angular, apically erect. Lobes of Ci laminar, trian- gular, erect. Gynostegium 1.6-1.7 mm high, 1.8-2 mm diam., atop a stipe, 0.55-0.66 mm long. An- thers broader than high, pentagonal, abaxially pla- nar; anther wings 0.8-1 mm long, convergent, ex- tending along the whole length of the anther; adjacent anther wings parallel, basally centrifugal. Connective appendages 0.7-0.75 mm long, 0. mm wide, ovate, narrower than the stamen, slightly inflexed. Pollinarium: corpusculum 0.3— 0.35 mm long, more than twice as long as broad, obovoid; caudicles 0.45-0.5 mm long, apically cy- lindrical, then flattened, s-shaped, convex-concave; 336 Annals of the Missouri Botanical Garden e 27. Cynanchum jpa ense Bullock sett- pee 236.—2. Flower, corolla partially remov ium.—o5. Stylar head. Drawn D. P Conrad. —]. Habit iad 3269) and inflorescence (Leedal 4015). 2-5: Dow- ved.—3. ;ynostegium and corona, partially removed.—4. Pollinar- Volume 83, Number 3 1996 Liede 337 Cynanchum in Africa pollinia apically attached to the caudicles, 0.58— .6 mm long, 0. .26 mm wide, ovate in cross section, pyriform. Stylar head 1.3-1.5 mm diam., 0.3-0.4 mm high; upper part 0.05-0.1 mm high, shorter than the lower part, umbonate. Fruits, seeds, and chromosome number unknown. Distribution and habitat. Africa: Malawi, Tan- zania (T7), Zambia; 1700-2600 m; forests and for- est margins. Fairly localized and infrequent, but probably not threatened. Figure 8. Flowering time. October to March, May. Ilago (Safwa). Vernacular name. Selected specimens examined. MALAWI. North, Rum- i, Station Kyimbila, Nyassa Hochland, 2000 m, 19 Dec. 1911, Stolz 1034 (B. K, M). TANZANIA. Iringa: Mufindi, Luisenga Stream, forest path by stream, 1830 m, 3 Jan. 1987, Lovett 1303 (K, MO) M ya: Mbeya range, World’s End view, Ipinda, 2660 m, 6 Feb. 1976, Cribb, Grey-Wil- son & Mwasumbi 10576 (K). Ruvuma: near Uwemba vil- & Mwasumbi 18209 (K). ZAMBIA. Eastem (FZ), Chama, Nyika Plateau, S of Zambian Rest House, 21 May 1989, Goyder, Pope & Radcliffe- Smith 3269 (K). Comments. the more attractive, relatively large-flowered spe- cies of the genus. It belongs to the C. altiscandens group with a very highly fused corona and a stipi- tate gynostegium. Its closest relative is probably C. altiscandens. chlechter was obviously aware of this species, as it is found under the name Cynanchum stolzii Schltr. in several herbaria. However, the name was Cynanchum rungweense is one of never published, so C. rungweense Bullock is the valid name for the taxon. Cribb and Leedal (1982: 105) reported C. rung- weense from the mountains of southern Tanzania. 27. Cynanchum schistoglossum Schlechter, J. Bot. 33: 271. Sep. 1895. TYPE: South Africa. Natal. Stanger, Phoenix, Apr. 1895, Schlechter 7090 (neotype, B; isoneotypes, AMD, BM); syntypes cited in the protologue: Schlechter 7106, Taylor 1895 (both probably destroyed in B, no isotypes found). xi n phi ens N. E. Br., Bull. Misc. Inform., Kew 895: 257. Oct. 1895. TYPE: Congo, Sep. 1863, rus n s.n. (holotype, BM). Cynanchum brevidens N. E. Br. var. zambesiata N. E. Br., Bull. Misc. Inform., Kew 1895: 257. Oct. 1895. TYPE: year ai Island, July 1838, Kirk s.n. (holotyp p vagum . Bull. Misc. Inform., Kew : 257. Oct. 1895. “TYPE: Zaire. Stanley Pool, - gu ug. 1888, Hens 77 (holotype, K). Cynanchum minutiflorum K. Schum., Bull. Soc. Roy. Bot. Belg. 37: 123 (1898), nom. illeg., because Schumann cited 2 ens E the type of Cynanchum vagum N. E. Br., bubus did De Wild. & T. Wo eur Mus Congo, Ser. a Bot. Ser. 2, 1(2): 4 900. TYPE: Zaïre. a Toumbwé, 27 Lig ie Deore 904 (ree cia! here, BR). Plants twining, to 3 m high, richly branched; rhi- zomatous; rhizomes 1-2 mm diam. Subterranean or- gans woody rootstocks. Shoots perennial, 1-1.5 mm diam., herbaceous, sparsely glabrescent with erect trichomes 0.3-0.4 mm long. “Stipules” ovate, 3-7 mm long, 2-5 mm wide. Leaves with petioles 10— 25 mm long, leaf blades herbaceous, 35-60 mm long, 15-35 mm wide, ovate-lanceolate to ovate, basally cuneate or cordate to lobate with 4 colleters in the adaxial sinus, apically acute to acuminate, adaxially isolatedly covered with erect trichomes mm long, evenly distributed over the whole surface, abaxially sparsely covered with erect tri- chomes 0.3-0.4 mm long restricted to veins and margins. Inflorescences bostrychoid, 5-20-flowered, 5-10 flowers open at a time; rachis 1-2 mm long; peduncles 5-12 mm long, densely covered with erect trichomes 0.3-0.4 mm long. Flowers musky scented; floral bracts 1 mm long, 0.5 mm wide at the base, triangular, with trichomes; pedicels 3-8 mm long, densely covered with appressed tri- chomes 0.3-0.4 mm long. Buds 1-1.5 mm long, 1- 1.5 mm diam., globose; aestivation imbricate. Calyx entirely free, abaxially with trichomes; lobes 0.8- 1.2 mm long, 0.4—0.6 mm wide, triangular, apically acute. Corolla cyathiform, 1.2-3. ong, abax- ially with a few isolated trichomes, adaxially gla- brous, whitish to yellowish green; lobes 0.8-1 mm wide, incurved, lanceolate, apically acute to acu- minate. Corona white, cyathiform, 1.2-1.7 mm high, equaling the gynostegium; C(is) consisting of Cs and Ci fused for %—% of total corona length, and Ci differentiated in shape; Ci shorter and thin- ner than Cs, laterally connate to Cs. Cs adnate to the filaments for not more than % of total corona length, appressed to the back of the stamens, with- out adaxial appendages; lobes of Cs flat, equally bifid or trifid with the two lateral lobes much small- er than the medium one, inflexed with straight mar- gins. Lobes of Ci flat, triangular, erect with straight margins. Gynostegium 0.75-1 mm high, 0.8-1.1 mm diam., sessile. Stamens with free filaments 0.1— 0.2 mm long, anthers broader than high, trapezoi- dal, abaxially planar; anther wings 0.37 mm long, paralleling the anther, extending along the whole length of the anther, adjacent anther wings parallel, in the same plane as the anther; outer guide rail smooth; connective appendages 0.3-0.4 Annals of the Missouri Botanical Garden mm long, 0.4-0.5 mm wide, widely ovate, narrower than the stamen, strongly inflexed; margins emar- slightly bifid. Pollinarium: 0.14—0.15 mm long; margins of the corpuscular ginate, corpusculum cleft parallel; caudicles 0.09-0.1 mm long, cylin- drical, straight, horizontal; pollinia laterally at- tached to the caudicles, 0.2-0.25 mm long, 0.09— Stylar upper Folli- fusiform, round 0.1 mm wide, ovoid, ovate in cross section. head 0.7-0.9 mm diam., 0.3-0.35 mm high: part 0.1-0.15 mm high, depressed-conical. cles 50-55 mm long, 5-6 mm wide, in cross section, apically obtuse, medium brown, smooth, glabrous. Seeds 4.5-5 mm long, 3.54 mm wide, ovate, light brown, seta and aseta side almost smooth, with 0.4-0.5 mm wide wing with distally dentate margin; coma 20-25 mm long. Chromosome number unknown. Distribution and habitat. Africa: Angola, Bot- swana, Burundi, Kenya (K4), Malawi, Mozambique, Namibia, Rwanda, South Africa (Natal, Transvaal), Tanzania (T2, T3. T4. T6). Uganda (U2, U4), Zaire, Zambia, Zimbabwe; 0—1800 m; on clayey loam; for- est margins, thickets, grasslands; often near water. also roadsides and disturbed areas. Figure Flowering time. April and October. All year, with peak between Selected specimens examined. ANGOLA. Golungo Alto, Punto de Felix Simoes, June 1856, ju 4241 (K); Huilla, near Lopollo, Oct. 1895, ras 4251 (K). SWAT Mochudi, Jan.—Apr. 14, Rogers 6617 y PRE); Ngamiland, E bank of et river, near boundary with SWA, ca. 1020 m, 27 Apr. 1975, Müller & Biegel 2282 (K. MO, PRE, SRGH); Northern, Bushman Pits, Botlethe River at Loromoja, 22 Apr. | K, MON PRE, SRG ii Crombie unos 2 Buono med cie m, 28 (den us iz o ¿NYA. Central: North Nyeri, Nyeri, 19 Dec. 1921, Frias & Fries 139 (K). MALAWI. Between Koln and Kawanga, 2000-2300 m, June 1890, Whyte Mamitete River id E e on Lilongwe-Ft. pee Rd., 1150 m, 5 Feb. 1959, iat 1464 (K). MOZAM- IQUE. Mon iba. ee 3 July 1949, deed 459 (K); a 21 July 1950, Chase 2230 (K, SRGH). NA- . Andara, bei Dikundu. im minia E km S jenes ie 15 June 1971, Giess 11436 (PRE). RWANDA. Biumba, colline Karukwanzi, n is Mutara, pres de la river Kakitumba, 21 Mar. 1958, Troupin 6762 (MO). SOUTH AFRICA. Natal: Dota Merebank, S.W., 19 Feb. 1967, Baijnath 132 (PRE); Pietermaritzburg, Inanda, July 1880, Wood 611 (BM): Port d s Isipingo Rail (Platts Estate), ca. 13 m, 7 Apr. 1966, Ward 5548 (PRE, Stanger, near Umhlanga river, 24 aps 1895, Wood 5664 (BOL, MO, PRE); Umzinto. adan 6 Apr. 1967, Ba- ynath 287 (PRE). Transvaal: Acornhoek, 1% mi. E of Skukuza, Kruger National Park. 300 m, 5 Apr. 19 Codd 5491 (PRE). TANZANIA. Arusha: Sakila, rd. to TL Sakila swamp, 1500 m, 14 Sep. 1971, Richards 27219 (K, MO). Morogoro: Uluguru Gebirge, ca. 1200 m, 30 June 1933, Schlieben 4044 (K, MO). Rukwa: Sakalilo (nr), 1000 m, 25 May 1951, Bullock 3896 (K). Ruvuma: Il- onga, 530 m, 20 June 1967, Robertson 733 (K). Tabora: near Kisanga, ca. 700 m, 19 Aug. 1970, Thulin & " Mhoro . Tanga: Korogwe, Kisa a near Mpal altas . 30 Apr. 1971, Semsei 4239 (K). UGANDA. An- kole, Mitoma, 1500 m, Mar. 1939, ae fee 600 (K); Kigezi. Kanungu. 1830 m. June 1939, Purseglove 821 (K); da Mulange. 1430 m, Sep. 1919, Dümmer 4306 (BM, K); Toro, near Sempayo, Oct. 1924, Liebenberg 947 (K). ZAIRE. Dolo (C ongo), June 1899, Schlechter 12485 (K. 12 Apr 1968, Phiri 158 (K, S ] Lindzi Boma, 27 Apr. 1952, White 2479 (K, MO); Chilanga, Mt. dg ees 'h Station, 24 Mar. dine 3078 (K). 3WE. gi cem Oct. Gairdner 546 ( on toria Falls, S. Bank of Zambesi, 1000 m, May 1915 . Rogers 13125 uo e Comments. Neither of the two syntypes of Cy- nanchum schistoglossum could be traced and both have probably been destroyed in B. The specimen selected as neotype was collected and identified by Schlechter and can thus be considered to conform to his concept of the species. Schumann (1895) published C. minutiflorum as nomen nudum and only typified it three years later on Hens 77, the type specimen of C. vagum N. E. r. Of the two well-preserved syntypes of the de- dewevrei (Dewèvre 904 and 976a, both in BR), the one with the more precise collec- scription of C. tion data has been chosen as lectotype. Cynanchum schistoglossum is perhaps the most variable species on the African mainland. Corona dentation and degree of fusion differ considerably among the populations. The species can be recog- nized by the very small flowers (smallest-flowered species on the African mainland) and the distinct fusion of the staminal corona parts with the fila- ments, which is only found in C. gerrardii and C. lenewtonii, with which it also shares the character- istic shape of the anthers and anther wings. 28. Cynanchum somaliense (N. E. Brown) N. E. Brown in Dyer, Fl. Trop. Afr. 4(1): 398. 1903. Schizostephanus somaliensis N. E. Br. in Bull. Misc. Inform., Kew 1895: 250. 1895. Cynan- chum trifurcatum Schltr., Bull. Herb. Boissier . 1896, nom. nov. TYPE: Somalia. Boo- bi: Jane: & Thrupp s.n. (holotype, K). Figure 2 Cynanchum dentatum K. Se ‘hum., Annuario Reale Ist. Bot. 39. 1898. TYPE: Somalia. Inter Sassaber -Mil et Ogaden Riva 844 (holo- Roma 7: : et Cabaden iter icis dierum a Mi distans locis aridis silvaticis, Jan., type, F Volume 83, Number 3 Liede 339 Cynanchum in Africa Figure 28. racy somaliense (N. E. Br.) N. E. Br. 1, 2: € & Thulin 1510; 3-7: Mesfin & Vollesen 4238; 8: Friis et al. 3221.—1. Habit.—2. Node with inflorescence.—3. Flower, corolla partially remove . Gynostegium and corona, partially a. Corona lobe, leal d Pollinarium.—7. Stylar head.—38. Fruit. Drawn by Jim Conrad 340 Annals of the Missouri Botanical Garden Plants erect or ascending, twining, 14 m high, sparsely basicaulously branched, with rhizomes. Shoots herbaceous, sparsely covered with flexuous trichomes 0.3-0.4 mm long; internodes 4-11 cm long, 0.8-2 mm diam. “Stipules” ovate, apiculate, 10-12 mm long, 7-10 mm wide. Leaves with peti- oles 15-45 mm long, leaf blades herbaceous, 30— 85 mm long, 20-60 mm wide, ovate, basally cor- date to lobate, lobes 3-5 mm long, with 5-7 col- leters in the adaxial sinus, apically acute to acu- minate, adaxially and abaxially isolatedly covered with appressed trichomes 0.3-0.4 mm long, con- centrated on veins and margins. Inflorescences bos- trychoid, 10—15-flowered, 6-8 flowers open at a time; rachis 1 mm long; peduncles 10-30 mm long, densely covered with flexuous trichomes 0.2-0.3 mm long, along a single line. Flowers aromatically scented (fide Gillett 13340); floral bracts 1.4—1.8 mm long, 0.3 mm wide at the base, linear, with trichomes; pedicels 4-7 mm long, densely covered with flexuous trichomes 0.2-0.3 mm long. Buds 1.6-1.8 mm long, 1.4-1.5 mm diam., (apically widened), with imbricate aestivation. Ca- cylindrical lyx basally fused; abaxial surface glabrous; lobes 2-2.2 mm long, 0.6-1.1 apically acute. Corolla rotate; basally fused; 3.5—4 mm wide, ovate to oblong, mm long, abaxially yellowish purple, adaxially yel- low; lobes 1-1.2 mm wide, patent, cucullate, api- cally obtuse. Corona white, changing to purplish red with age, 3-3.5 mm high, equaling the gynos- tegium; C(is) consisting of Cs and Ci only basally fused, only Cs differentiated. Cs not adnate to the filaments, without adaxial appendages; lobes of Cs laminar, trifid (central lobe internally with two strong folds), apically erect. Gynostegium 1-1.2 mm high, 2-2.2 long. Stamens without free filaments, anthers broad- er than high, deltoid, abaxially wings 0.4-0.45 mm long, divergent, not extending mm diam., atop a stipe 2.2-2.3 mm convex; anther along the whole length of the anther; the anther forming a “pseudostipe” 0.6—0.7 mm high; adjacent anther wings parallel, basally widened, in the same plane as the anther. Connective appendages ca. 0.25 mm long, 0.7—0.8 mm wide, ovate, equaling the stamen in width, strongly inflexed. Pollinarium: corpusculum ca. 0.3 mm long, more than twice as long as broad, elliptic; caudicles ca. 0.8 mm long, cylindrical, straight, declinate; pollinia apically at- tached to the caudicles, 0.75-0.8 mm long, 0.25— 0.3 mm wide, ovoid, ovate in cross section. Stylar head white, 1.1-1.2 mm diam., 0.3-0.35 mm high: upper part 0-0.05 mm high, flat to umbonate. licles one per flower, 50-60 mm long, 8-10 mm diam., obclavate, apically obtuse, medium brown, sparsely covered with 3—5-mm-long protuberances. with sparse indumentum. Seeds approximately 60— 80 per follicle, marginally winged (only immature seeds known). Chromosome number unknown. Distribution and habitat. Africa: Ethiopia Bale, Haverge, Sidamo, Welo), Kenya (K1, Somalia (Hiiraan, Sanaag, Woqooyi Galbeed/Todgh- eer), Sudan, Tanzania (T1), Uganda (U1, UA); 350- 1700 shrub, —. m, Acacia-Commiphora bushland, open and grasslands. Widespread and fairly com- mon. Figure 10. Flowering time. March to November. Vernacular names. | Goriss (Boran), gasur riyoli (Ogaden), vapo (Uganda). Uses. Helps women in childbirth, also for ab- dominal pains (Dyson-Hudson 224, 225, 226). Selected es a ETHIOPIA. Bale: be- n Gaad and Harr ca. 860 m, 26 Sep. 1964, Burger 4 (EA); Haverge. Se lave; 630 m, 11 Apr. 1956, Sim- mons 141 (K). Sidamo: Borana, 46 mi. SE of Neghelle, on Filtu- s Road, 1450 m, 4 Apr. 1974, Ash 2425 (K, MO). Welo: below Back on Assab rd., 700 m, 19 Sep. 1962, Mooney 9662 (WAG). KENYA. Central: Meru, 13 km N of Isiolo on road to Marsabit, 1050 m, i Feb. din Gilbert. Gachathi & bane 5314 (K, MO); North Nyer , 1500 m, 19 i 1981. Gib 6093 (K). Northern. Frontier: M Mandera, 2 km N ) Wak, 30 Apr. 1978, Gilbert & Thulin 1265 (K). SOMA- LIA. Hiiraan, Bulo Burti, ca. 25 km along the = E Buqda Caqable, then 5 km SW along cutis, 180 7 May 1986, Kuchar 17017 (K, MO); kn m (S a Eri- gavo, 1340 m, 26 Nov. 1980, ae & Watson 3279 K); Woqooyi Galbeed/Todgheer, Boundary Pillar 93, 1100 m, 10 Apr. 1932, Gillett 4173 (K). SUDAN. Imatong Mts., just S of Ngarama, along rd. to Molongori, i northern- most E porat of cont. mountain chain, 700 m, 13 Mar. 1986, Friis & Vollesen 1189 (K). TANZANIA. un Igalu- is. Sn Mwanza, 1200 m, 18 July 1953, Tanner 1597 K). UGANDA. ML E near Emonayaben, Nabilatuk, 1200 m, 26 June 1957, Dyson-Hudson 226 (K); Mengo, Buvuma, Namunyoro, Maitland 1190 (K). T ieni — R Schlechter (1896) suggested Cy- nanchum trifurcatum as a nomen novum for Cy- nanchum somaliense (N. E. Br.) N. E. Br., that there was another species known by the name Comments. stating of Cynanchum somaliense. However, he neither in- dicated an author of this species, nor is any such species known to me as validly published. There- fore, there is no need to rename Cynanchum so- maliense (N. E. Br.) N. E. Br. Cynanchum somaliense has long been regarded as the closest relative of Schizostephanus alatus. However, while this judgment was derived from the long stipe and the seemingly similar pollinarium structure, Schizostephanus has recently been iden- tified as a member of a different subtribe (Liede, 1993). While long stipes have been developed sev- Volume 83, Number 3 1996 Liede 341 Cynanchum in Africa eral times independently, the unique fine structure of the pollinarium of Schizostephanus has been il- lustrated in Liede (1993). Liede and Nicholas (1992) have deduced that the corona in Pentar- rhinum can be interpreted morphologically as a de- velopment from the type found in P. somaliense. The close relationship of C. somaliense and Pen- tarrhinum is further supported by the wingless fol- licles with soft spines. Schizostephanus, in contrast, possesses smooth follicles with conspicuous wings. 29. Cynanchum umtalense Liede, sp. nov. TYPE: Zimbabwe. Melsetter: Chirinda Forest Margin, ca. 1200 m, Jan. 1962, Goldsmith 1/ 62 (holotype, K; isotypes, B, BR, FI, SRGH). Figure 29. Volubilis. Corollae lobis introrsum pilis singularibus or- natis; partibus staminalibus coronae gynostegialis ligula- tis; capite stylorum longe elongato Plants ascending, twining, 5-6 m high, richly and irregularly branched. Shoots perennial, herba- ceous, densely covered with flexuous trichomes 0.7—0.8 mm long; internodes 7-10 cm long, 0.5-1 mm diam. “Stipules” absent. Leaves with petioles 15-20 mm long, leaf blades herbaceous, 35-70 mm long, 15-45 mm wide, ovate, basally cordate to ob- tuse with 2-3 colleters in the adaxial sinus, api- cally acuminate (acumen 2-3 mm long), adaxially and abaxially sparsely covered with appressed tri- chomes 0.6-0.8 mm long evenly distributed over the whole surface. Inflorescences sciadioidal, 7-11- flowered, 2-5 flowers open at a time; peduncles 7— 15 mm long, densely covered with flexuous tri- chomes 0.5—0.6 mm long. Flowers with floral bracts 1.6-1.8 mm long, 0.4—0.5 mm wide at the base, triangular, with trichomes; pedicels 8-10 mm long, densely covered with flexuous trichomes 0.4—0.5 mm long. Buds mm long, 2-2.5 mm diam., conical, basally with imbricate, apically with con- torted aestivation. Calyx basally fused; abaxial sur- face with trichomes; lobes 1.4—1.6 mm long, 0.7— 0.8 mm wide, ovate, apically acute. Corolla rotate, basally fused; 6-7 mm long, abaxially with tri- chomes, adaxially cream, rose to purple along the main nerves; adaxially with verrucose trichomes 0.1-0.12 mm long, evenly distributed over the whole surface; lobes 1.8—2 mm wide, apically twist- ed, patent to recurved, linear to triangular, apically obtuse to acute. Corona white, 2.3-2.5 mm high, equaling the gynostegium (except for stylar head appendage); C(is) consisting of Cs and Ci fused for more than % of total corona length, Cs and Ci dif- ferentiated, Ci shorter than Cs. Cs not adnate to the filaments, with adaxial appendages; lobes of Cs lobes laminar, triangular, apically erect; append- ages of Cs slightly longer than Cs, laminar, trian- gular, erect. Lobes of Ci laminar, rectangular, pro- ducing a pronounced convex fold along the upper third of corona length, erect, with straight, emar- ginate margins. Gynostegium 1.6-1.8 mm high (without stylar head appendage), 1.6-1.8 mm diam., sessile. Stamens without free filaments, an- thers about as high as broad, rectangular, abaxially planar; anther wings 0.7-0.8 mm long, parallel to each other, extending along the whole length of the anther; adjacent anther wings parallel, in the same plane as the anther. Connective appendages 0.5- 0.6 mm long, 0.7-0.8 mm wide, ovate, narrower than the stamen, slightly inflexed. Pollinarium: cor- pusculum 0.2-0.22 mm long, rhomboid; caudicles 0.09-0.1 mm long, flattened, straight, horizontal, triangular; pollinia subapically attached to the cau- dicles, 0.3-0.35 mm long, 0.12 mm wide, clavate, ovate in cross section. Stylar head white, 0.5—0.6 mm diam., 2.5-2.6 mm high; upper part 2.3-2.4 mm high, obinfundibuliform. Follicles one per flower, 45 mm long, 6-8 mm diam., obclavate, keeled, apically shortly beaked, medium brown, with dense indumentum. Seeds 6.5-7 mm long, 3.5-4 mm wide, ovate, medium brown, seta and aseta side sculptured with longitudinal ridges, mar- ginally with 0.3-mm-wide wing with entire margin; coma 18-20 mm long. Chromosome number un- known. Distribution and habitat. Africa: Malawi Dee Central), Zimbabwe (Melsetter, Umtali); O m; forest margins. A fairly infrequent species with a rather limited distribution, but prob- ably not immediately threatened because it grows inside forest reserves. Figure 2. Flowering time. January, April to July. Specimens examined. MALAWI. Central. Dedza, Ded- za Mountain, 1750 m, 5 Apr. 1978, Pawek 14228 (BR, K, MAL, MO, WAG); lower eastern slopes of Domwe Hill, above Trinidad's Place, 1 Apr. 1961, Chapman 1212 (K). North. Nkhata Bay, 3 km S of Chikangawa, 1950 m, 10 July 1978, Phillips 3515 (K, MAL, MO, SRGH, WAG). ZIMBABWE. Umtali, Vumba Mts., 28 Apr. 1957, Chase 6465 (B, FI, K, SHRG), 4 June 1957, Chase 6541 (K), 2 Apr. 1958, Chase 6866 (K), 12 May 1957, Pole-Evans 5210 (K); Hawkdale, Vumba, 29 Jan. 1957, Chase 6562 (K). Comments. Cynanchum umtalense is a very distinctive species, but clearly identifiable as a member of the group with verrucose trichomes on the corolla lobes and ligules; probably closest to Cynanchum abyssinicum. 342 Annals of the Missouri Botanical Garden Figure 29. Cynanchum umtalense Liede. 1-5: Chase 6866; 6: Chase 2562.—1. Habit with inflorescences.—2 Flower.—3. Gynostegium and corona, partially removed.—4. Pollinarium.—5. Stylar head.—6. Fruit. Drawn by Jim Conrad. Volume 83, Number 3 1996 Liede 343 Cynanchum in Africa 30. Miei pa virens (E. Meyer) D. Dietrich, Syn. Pl. 2: 905. 1840. Cynoctonum virens E. Mey., C Pl. Afr. Austr. 216. 1838. Cy- nanchum virens (E. Mey.) Steud., Nomencl. Bot. (ed. 2) 1: 462. 1841, nom. superfl. Vin- cetoxicum virens (E. Mey.) Kuntze, Revis. Gen. Pl. 2: 424. 1891. Endotropis meyeri Decne. in Candolle, Prodr. 8: 546. 1844, nom. superfl. TYPE: Namibia. Garip, 19 Dec. 1832, Drége 3439 (lectotype, designated by Liede (1993), P) Not Cynanchum virens (E. Mey.) D. Dietr., Syn. Pl. 2: 906. 1840, based on Schizoglossum virens E. Mey., Comm. Pl. Afr. Austr. 219. 1838. TYPE: South Africa. Natal: inter Omsamcaba et Port Natal, Drége s.n., as cited by Schlechter, Bot. Jahrb. Syst. 20, Beih. 51: 7. 1895 (see comments). Plants twining, 0.5—1 m high, sparsely branched; subterranean organs rhizomatous; rhizome 10-15 mm diam. Shoots perennial, 50-100 cm long, 1 mm diam., herbaceous, glabrous, or isolatedly to sparse- ly covered with appressed trichomes 0.25—0.3 mm long, basally ieee with yellowish bark. * ‘Stipules’ ° ovate, 3-5 mm long, petioles 10-25 mm long; leaf blades herbaceous, 0-50 mm long, 15-30 mm wide, lanceolate, ba- sally cordate with 4 colleters in the adaxial sinus, apically acute, abaxially and adaxially glabrous to isolatedly indumented with appressed trichomes 0.25-0.3 mm long, evenly distributed over the whole surface. Inflorescence bostrychoid to scia- dioidal, 7-15-flowered, 3-8 flowers open at a time; rachis to 0.5 mm long; peduncles 2-10 mm long, glabrous isolatedly covered with appressed tri- chomes 0.25-0.3 mm long. Flowers with floral bracts 0.8-1 mm long, 0.2-0.3 mm wide at the base, linear to triangular, with trichomes; pedicels 5-7 mm long, glabrous to isolatedly covered with erect trichomes 0.25-0. long, 2.5-3 mm diam., elongated-conical; aestiva- mm long. Buds .9 mm tion imbricate. Calyx basally fused, abaxially gla- brous or with trichomes; lobes 1-1.5 mm long, 0.5— 0.7 mm wide, triangular, apically acute. Corolla rotate, fused at the base, 6—8 mm long, abaxially glabrous, greenish white, adaxially with verrucose trichomes, whitish green; lobes 1-1.5 mm wide, spreading, ovate, apically acuminate, twisted. Co- rona white, cyathiform, 5—5.5 mm high, exceeding the gynostegium but not obscuring it; C(is) consist- ing of Cs and Ci fused for ca. % of corona length, both Cs and Ci differentiated in shape, Ci shorter and thinner than Cs, dorsally connate to Cs. Cs adaxially with adaxial appendages; lobes of Cs flat, long-apiculate, inflexed, with straight margins; ad- axial appendages shorter than Cs, erect, liguliform. Lobes of Ci flat, ovate, erect, with straight, dentic- ulate margins. Gynostegium sessile, 1.4—1.6 mm high, 1-1.2 mm diam. Stamens without free fila- ments; anthers about as high as broad, trapezoidal, abaxially planar, anther wings 0.55—0.6 mm long, parallel to each other, extending beyond the anther proper forming a basal arch; adjacent anther wings parallel, in the same plane as the anther. Connec- tive appendages 0.6-0.65 mm long, 0.5-0.55 mm wide, deltate, narrower than the stamen, slightly in- flexed. Pollinarium: corpusculum 0.2-0.22 mm long; margins of the corpuscular cleft sinuate; cau- dicles 0.14—0.16 mm long, flattened, straight, horizontal, triangular; pollinia laterally attached to the caudicles, 0.32-0.35 .16 mm wide, ovoid, round in cross section. Stylar head 0.6-0.75 mm diam., 0.8-1 mm high, upper part 0.57-0.6 mm high, conical. Follicles 50-60 mm long, 15-20 mm wide, obclavate, obtusely deltate in cross section, apically shortly to strongly beaked, keeled, light to dark brown, longitudinally grooved, glabrous. Seeds 5-5.5 mm long, 3-4 mm wide, ovate, medium to dark brown, seta and aseta side sculptured with longitudinal ridges, marginally with 0.4—0.6-mm-wide wing with entire margin; coma 20-25 mm long. Chromosome number unknown. mm long. 0.14—0 Distribution and habitat. Africa: Lesotho, Na- mibia, South Africa (Cape Province, Transvaal, Or- ange Free State); 1200-2000 m, flats and gentle slopes; sandy to loamy soil; riverine vegetation; partial shade. Comments. Dietrich (1840: 905, 906) pub- lished two combinations as Cynanchum virens (E. Mey.) D. Dietr., one based on Cynoctonum virens E. Mey., the other one on Schizoglossum virens E. Mey. Schlechter (1895) is the only author using the lat- ter, because all other authors regard Schizoglossum, a member of the subtribe Asclepiadinae, as well distinct from Cynanchum. Further details, illustration, distribution map, and citation of specimens are provided in Liede 1993) — 31. Cynanchum zeyheri Schlechter, Bot. Jahrb. Syst. 20, Beibl. 51: 3. 1895. TYPE: Ecklon & Zeyer 78 (lectotype, designated by Liede (1993), SAM). 15-20 cm high, branched; subterranean organs consisting only of fibrous roots. Shoots perennial, 15-20 cm long, 1- 1.5 mm diam., herbaceous, glabrous, or isolatedly Plants decumbent, richly to sparsely covered with appressed trichomes 0.15— 344 Annals of the Missouri Botanical Garden 0.2 m long, basally woody with yellowish bark. “Stipules” 2-3 mm long, 3—4 mm wide. Leaves with petioles 5-10 mm long; leaf blades herbaceous, 0-15 mm long, 5-10 mm wide, ovate, basally rounded, without colleters, apically obtuse, or acut and apiculate, adaxially and abaxially glabrous. /n- HAlorescence sciadioidal, 2-5 all flowers open at a time; peduncles 0.5-5 mm long, glabrous to isolatedly covered with appressed trichomes 0.15-0.2 mm long. Flowers sweetly scented; floral bracts 0.7-1 mm long, 0.2-0.5 mm wide at the base, triangular, papillose; pedicels 5-10 mm long, glabrous. Buds 3.54 mm long, 1-1.5 mm diam., elongate-conical; aestivation imbricate, apically contorted. Calyx basally fused, abaxially glabrous; lobes 1-1.5 mm long, 0.5-0.8 e, apl- cally acute and apiculate. Corolla rotate, fused at the base, 3—4 mm long, abaxially glabrous, brown, adaxially minutely papillose brown; lobes ca. 1 mm wide, spreading, oblong, apically obtuse, mostly twisted, with revolute margins. Corona white, -flowered, mm wide, ovat tu- bular to campanulate, 1.5-2 mm high, shorter than the gynostegium; C(is) consisting of Cs and Ci fused for about ?4 of total length, only Cs differentiated in shape. Cs without adaxial appendages; lobes of Cs flat, ovate, erect, with straight margins. Gynos- tegium 0.7-0.8 mm high, 0.8-1 mm diam., atop a stipe, 0.7-1 mm long. Stamens without free fila- ments; anthers trapezoid, abaxially rounded; anther wings 0.25-0.3 mm long, clearly differentiated, paralleling the anther, parallel to each other, ex- tending along the whole length of the anther mar- gin; connective appendages 0.55-0.6 mm long, 0.4—0.45 mm narrower than the stamen, erect. Pollinarium: corpusculum 0.13—0.14 mm long; caudicles 0.07-0.08 mm long, flattened, straight, horizontal to declinate, triangular; pollinia 0.3-0.35 mm long, 0.12-0.13 mm wide, clavate, elliptical in cross section. Stylar head 1-1.1 mm diam., 0.5-0.55 mm high: upper part 0.47-0.5 mm high, conical. Follicles 35—45 mm long, 12-15 mm wide, club-shaped, sharply deltate in cross section, apically shortly beaked, wingless, light brown to medium brown, longitudinally grooved, glabrous. Seeds 5—6 mm long, 4—5 mm wide, pyriform, me- dium brown, seta and aseta side tuberculate (but less pronouncedly so on the seta side); marginally with indistinct wing 0.2-0.3 mm wide, with entire margin; coma 20-25 mm long. Chromosome number unknown. wide, ovate, Distribution and habitat. Africa: South Africa [Cape Province, disjunction Cape Peninsula (Lions- head)-Bredasdorp|; 0—1000 m; flats to moderate slopes; shales and limestone; fynbos, renosterveld, strandveld. Comments. Further details, illustration, distri- bution map. and citation of specimens are provided in Liede (1993). EXCLUDED SPECIES Cynanchum aphyllum (Thunb.) Schltr. = Sarco- stemma viminale (L.) R. Br. (Liede, 1991) Cynanchum arboreum Forssk. = rea (Forssk.) Schweinf. Cynanchum atropurpureum (E. Mey.) D. Dietr. = chizoglossum atropurpureum E. Leptadenia arbo- Me me i: argel ile = Solencstemana argel Delile) Hayne Cynanchum bidens (E. Mey.) D. Dietr. glossum bidens E. Mey. Cynanchum boveanum Decne. = eanum (Decne.) Decne. Cynanchum chirindense S. Moore = Tylophora sp. Cynanchum cordifolium (E. Mey.) D. Dietr. = Schi- zoglossum sp. Cynanchum crispum Thunb. = Gomphocarpus cris- jus (P. J. Bergius) W. T. Aiton Cynanchum defoliascens K. Schum. = Blyttia fru- ticulosum (Decne.) D. V. Field € J. R. I. Wood Cynanchum filiforme L. f. Schlechter (1895) Cynanchum fruticulosum Decne. = Blyttia ll culosum (Decne.) D. V. Field € J. R. L Cynanchum gossweileri S. Moore = Se T ean gossweileri (S. Moore) Liede = Schizo- Glossonema bov- = Schizoglossum sp. fide Cynanchum hamatum (E. Mey.) D. Dietr. = Schi- zoglossum hamatum E. Mey. Cynanchum heterophyllum Delile = Leptadenia heterophylla (Delile) Decne. Cynanchum lancifolium Schumach. & Thonn. = Leptadenia lancifolium (Schumach.) Decne. Cynanchum latifolium Schumach. & Thonn. = Lep- tadenia lancifolium Decn Cynanchum linifolium (Balf. D Bullock = Vince- toxicum linifolium Balf. f. Cynanchum mauritianum Bojer ex Decne. = Ty- lophora laevigata Decne. Cynanchum microstegium K. Schum. = Blyttia fru- ticulosum (Decne.) D. V. Field & J. R. I. Wood Cynanchum molle (E. Mey.) D. Dietr. = Anisotoma cordifolia Fenz Cynanchum oleaefolium Nectoux = Solenostemma Delile) Hayne Cynanchum omissum Bullock = Fockea angusti- folia K. Schum. Cynanchum ovatum Thunb. lata (Retz.) Wight = Leptadenia reticu- Volume 83, Number 3 1996 Liede 345 Cynanchum in Africa Cynanchum pendulum Poir. = Leptadenia sp. Cynanchum radians (Forssk.) Lam. = Odontanth- era radians (Forssk.) D. V. Field Cynanchum reticulatum Retz. = Leptadenia reti- culata (Retz.) Wight Cynanchum roseum Chiov. = Tylophora heterophyl- la A. Rich. Cynanchum scabrum Schumach. & Thonn. = Mars- denieae sp. Cynanchum senegalense Sieber ex Decne. = Gym- nema subvolubile (Schumach.) Decne. Cynanchum subvolubile Schumach. € Thonn. = ymnema subvolubile (Schumach.) Decne. Cynanchum eres (Turez.) R. A. Dyer = S. viminale (L.) R. B Cynanchum validum N. E. Br. alatus Hochst. ex K. Schum. Cynanchum verticillare Lam. = Schizoglossum sp. fide Schlechter (1895) Cynanchum viminale L. = Sarcostemma viminale (L.) R. Br. = Schizostephanus Literature Cited Adam, J.-G. 1975. Asclepiadaceae. Pp. 966-976, 1174- 1190 in: Flore descriptive des Monts Nimba, 3e partie, T Mus. Natl. Hist. Nat. Ser. B, 24. Agnew, A. D. Q. 1974. Asclepiadaceae. Pp. 366-396 in Flora of Lee Tanya. Oxford Univ. Press, ad a Brown, N. E. piadaceae. In: scle W. (editor), Pis er Sorel Africa 4(1): 231—503. i Reeve, eae sclepiadaceae. In: W. T. Dyer (editor), Flora Capensis AQ) 518-1036. Lovell Reeve. London Bullock, 53. Notes on Abia Atclepiadacede III. Bull Misi: Inform., Kew 8: 353-355. ———. 1955. Notes on African Asclepiadaceae VII. Bull. Misc. ap e Kew 10: 611—626. clepiadaceae. In: J. chinson & J. Flora of West Tropical Africa 2: 196 M. Did (ets on. al. 1982. Asclepiadaceae. P ntain Flowers of Southern Tanza- . Balke - Rotterdam. Dallwitz, M. J. A pine system for coding taxo- omic pek Taxon 29: 4 . A. Paine. 1986. ys guide to the DELTA syst ied general system A Wah taxonomic de- dina aire Austral. Div. Entom 13. De Lessert, B. 1846. Apocynaceae, Mdb eae. Pp. 20-38 (tab. 45-91) in Icones Selectae Plantarum, vol. ortin, Masson et Soc., Paris Diet D. 1840. Pentandria Digynia. Pp. 883-909 in: Synopsis Plantarum, vol. 2 Greuter, W. G., F. R. Begs: H. M. Burdet, W. G. Chal- oner. o Code), Adopted by the XVth Inter- national Botanical Congress, Yokohama, August-Sep- tember, 1 Hewson, H. J. 1 88. Plant Indumentum—A Handbook s Terminology. Australian Flora and Fauna Series 2: 1-27. Huber, H. 1967. 114. Asclepiadaceae. In: H. Merxmüller (editor), Prodromus einer Flora von Südwestafrika 4: 1— Liede, S. 1991. Cynanchum gerrardii—A new combina- tion for a well-known African species (Asclepiadaceae). Taxon 40: 113-117. A revision of the genus Cynanchum in southern Africa. Bot. Jahrb. Syst. 114: 503-550. —— 4 Cynanchum lenewtonii (Asclepiadaceae), new leafless species from the African mainland. Kew Bull. 49: 119-123. 1996. Cyn Ee oe cum-Tylophora: New. considerations on an old problem Taxon ^ press — —— & H. Kunze. A descriptive system for c rona analysis in the Fe eni Pl. Syst. Evol. 185: 275-284. & A. Nicholas. 1992. A revision of the genus Pentarrhinum E. Meyer (Asclepiadaceae). Kew Bull. 47: 4 Meyer, E 1838. Asclepiadaceae. Pp. 193-225 in Com- EL de plantis Africae australioris. Voss, Leip- Polhill, D. 1988. Flora of Tropical East Africa. ait of 95. Beiträge zur Kenntnis stifle ischer Asclepiadeen. Bot. Jahrb. Syst. 20, Beibl. 51: 1- 56. . 1896. Cynanchum trifurcatum. Bull. Herb. Bois- sier 4: 448. . 1913. O africanae. Bot. Jahrb. Syst. 51: 128-15 — K. M. e P. 123 in Th. Durand & Ém. De Wildemann, FP pour la Flore du Congo, deux- iéme cule. Bull. Soc. Roy. Bot. Belgique 37: 128 Weberling, F. 1989. Morphology of flowers and inflores- 983. Asclepiadaceae. Pp. 48—49 in The Botany of the Commelins. Balkema, Rotterdam. PHYLOGENY AND SPECIATION IN LAPEIROUSIA SUBGENUS LAPEIROUSIA (IRIDACEAE: IXIOIDEAE)! Peter Goldblatt? and John C. Manning? ABSTRACT cladistic analysis of southern African subgenus Lapeirousia, one of two subgenera of the exclusively African genus Lapeirousi, yielded four equally parsimonious trees, one o "which is identical with the strict consensus tree. C aracters used in the analysis included growth form, corm A a range of floral characters, and capsule and seed features, not before known to vary significantly in this genu ora aris these is an apparent re toac versa The analysis suggested : ıl of m tube length from prd e from pollination by long-tongued flies and sphinx moths to pollination by bees and noctuid moths. s n er some surprising evolutionary changes. long to short, a shift corresponding al jo ge is a shift in floral organization from zygomorphy to ac EE correlated with an acaulescent form. The reconstructed phylogeny is used here to as sess characte volution and patterns of speciation by comparison of species in terminal clades in the cladograms. The resulting comparisons suggest that speciation in the subgenus is either allopatric or the result of microgeographic differentiation and ecological diversification stimulated by edaphic diversity. Despite the variety of floral forms and pollination syndromes in the subgenus there is no evidence of sympatric or lidera n speciation syndromes between sphinx mató, two guilds resulted appears important of these pollination guilds are tw ong-tongued fly Prepollination reproductive isolation appears to be achieved by shifts in pollination o long- tongued flies, and bees. The remarkable floral divergence that has to be a consequence of nee tion for repeated entry into preexisting pollination guilds. The most uilds in which either Prosoeca (Nemestrinidae) or Moegistorhynchus (Nemestrinidae) and Philolie he (Tabanidae) are pollinators. These two guilds are also likely to have been important in promoting speciation in other genera and families in the southern African flora. The flora of southern Africa is rich and unusually diverse for an area falling AAA in tem- Some 20,400 species of native vascular plants are currently rec- ognized in the region (Arnold & de Wet, 1993), of which about 80% are endemic (Goldblatt, 1978). The major factors proposed to account for the spe- perate latitudes (Goldblatt, 1978). cies richness are climatic, edaphic, and topograph- ic diversity and a history of paleoclimatic change in the late Tertiary (Goldblatt, 1978). Although these factors may permit the existence of large numbers of species, they do not indicate the modes of speciation that have led to this diversity. One method of inferring modes of speciation is by com- paring biological attributes of closely related, and by extension evolutionarily recent, sister species. Differences in biology between such species give an indication of the factors that led to speciation. Cladistic analysis is a critical method for identify- ing sister species. A detailed phylogenetic hypoth- esis such as a cladogram makes it possible to trace backwards in time the series of evolutionary events that gave rise to the modern taxonomic distribution of ecological features even in the absence of fossil information (Armbruster, 1993). This in turn makes it possible to develop specific hypotheses on the evolution of species diversity (Manning & Linder, 992). The tropical and southern African genus Lapei- rousia Pourret comprises some 40 species segre- gated in two subgenera each with two sections Goldblatt & Manning, 1990, 1992, 1994). Subge- nus Lapeirousia (21 species) is centered in coastal and near interior southwestern Africa. This is a prm semiarid region of low to moderate winter rainfall and extreme summer drought. Two widespread in the southwestern ue southern jd of Western Cape Province, South a, and a f ther two occur in tropical Africa de 1990b). Subgenus Paniculata (19 species) comprises the largely tropical African section Paniculata (14 spe- cies). with one species in the southwestern part of southern Africa (Goldblatt & Manning, 1992), and section Fastigiata (5 species), which is restricted species are ! This research was supported by National Geographie Society Grant 4816-92. We thank Peter Hoch, Jorge Crisci, D. Snijman, and A. de Musis for helpful comments during the preparation of this paper, and Jennifer Hedin for running the bootstrap analyse . A. Krukoff Curator "1 "Afric an Botany, 0299, U.S.A. Missouri Botanical Garden. PO. Box 299, St. Louis, Missouri 63166- * Compton Herbarium, National Botanical Institute, Kirstenbosch Botanic Gardens, Claremont 7735, South Africa. ANN. Missourt BOT. GARD. 83: 346-301. 1996. Volume 83, Number 3 1996 Goldblatt & Manning Phylogeny in Lapeirousia Subg. Lapeirousia to the southwestern part of Western Cape Province. Field studies of subgenus Lapeirousia conducted over the past three years have provided a wealth of information about the ecology and biology of its 21 species, and this has enabled us to develop a de- tailed understanding of the group. This has provid- ed an objective measure of relationships among the species that in turn lays the foundation for analyses of changes in pollinator preferences and patterns of radiation and speciation. We use the reconstructed phylogeny of subgenus Lapeirousia and the inferred modes of speciation to address questions of how its species coexist and how they might have evolved. We expect these hypotheses to have more general application for plant speciation in southern Africa. MATERIALS AND METHODS CHARACTERS AND TAXA Subgenus Lapeirousia is believed to be mono- phyletic and the sister clade to subgenus Panicu- lata (Goldblatt & Manning, 1990), the outgroup for the cladistic analysis. The subgenus is regarded here as comprising 21 species (Goldblatt, 1972, 4 Goldblatt & Manning, 1994). Both subspe- cies of L. pyramidalis (subsp. regalis and subsp. pyramidalis) are included in the analysis because they differ in some important characters, and we were unwilling to assign arbitrary plesiomorphic states for the species. Morphological and anatomi- cal characters (Table 1) were assembled from the above systematic treatments, and supplemented by new data presented here. For the cladistic analysis 24 characters were ul- timately selected (Table 1; Appendix). These in- cluded all macromorphological aspects of the plants, as well as capsule and seed surface mor- phology, not known until now for most species. Seed characters of all species of Lapeirousia were inves- tigated for this study. Seeds vary both in primary and secondary sculpturing, and provided valuable characters for the cladistic analysis (Appendix: characters 6, 7, and 8). Chromosome cytology (Goldblatt & Takei, 1993) was not included in the cladistic analysis, but was used to assess the trees that were generated. Basic chromosome number for Lapeirousia is irai x — 10, but e base num- r for subgenus Lapeirousia appears to be x = (Goldblatt, 19902; Goldblatt & Takei, en E tistate characters of an additive nature, e.g., corm tunic bases lobed, lightly denticulate, coarsely den- tate, were treated as ordered (characters 2, 6, 7, and 14) ECOPHYLOGENY Data on pollination and habitat were recorded for each species in subgenus Lapeirousia and mapped onto the cladogram (Fig. 4). The ecological char- acteristics of the lower nodes were determined us- ing the operating principle of parsimony. In this way the evolutionary history of the interactions can be inferred and the sequence of evolutionary changes that have generated the current interac- tions can be determined (Donoghue, 1989; Brooks & McLennan, 1991; Armbruster, 1992). This meth- odology has been used to investigate the evolution of pollination systems in, for example, Dalechampia (Euphorbiaceae) (Armbruster, 1992, 1993, 1994). Data on pollination ecology are derived from Gold- blatt et al. (1995) CLADISTIC ANALYSIS Data were analyzed using the Hennig86 package of programs for parsimony analysis (Farris, 1988), using the mh* and bb* option, followed by succes- sive weighting. Five randomly generated taxon se- quences were analyzed, with the same result being obtained. The data include considerable homoplasy (there appear to be convergent trends for some flow- er types in different lineages) with the result that the strict consensus tree of 1322 equally parsimo- nious trees resulting from unweighted computations (Fig. 1A) shows little resolution (length 74, consis- tency index (CI) 0.48, retention index (RI) 0.67). Successive weighting, recommended by Farris (1969) for situations where unreliable (homopla- sious) characters outnumber reliable characters, is one way to improve tree resolution. The method selectively weights those characters that are more consistent at the expense of those that are homo- plasious. The method can be successful even when cladistically consistent characters are heavily out- numbered by homoplasious ones. After invoking the successive weighting option six equally parsi- monious trees (CI 0.81 and RI 0.91), but only four different topologies, were obtained. The four trees differ only in minor details and correspond in most respects to our intuitive ideas about species rela- tionships in the subgenus. The strict consensus tree obtained with successive weighting (Fig. 1B) is ac- tually identical to one of the four final trees (Fig. 2A), and discussion is framed around this tree. All the differences in branching patterns in the other trees are present in a second tree illustrated (Fig. 2B). The differences are alternative topologies at nodes 7 and 14. We generated one more tree using the data matrix and the same options, mh*, bb*, and successive weighting, but with Savanno- 348 Annals of the Missouri Botanical Garden able 1. Matrix and character list for the cladistic analysis (Figs. 1-3). The outgroups are Savannosiphon or subgenus Paniculata. The features of the latter are inferred by comparison of its constituent Li ies (no single species accords with the hypothetical ancestor). When the plesiomorphic state is uncertain for the outgroup, the character is scored as "?". Chromosome numbers were not included in the analysis, but ee a known (Goldblatt & Takei, 1993), are indicated opposite each species in Figure 4. Character 25 is used only with the analysis with Savannosiphon as the n (Fig. 3). Multistate characters 2, a 7, and 14 are treated as ordered (see Appendix 1), and 5 , 15, 17 are treated as unordered. Character number 111 111111 122222 2 Taxon 123456 789012 345678 901234 5 Savannosiphon 0720100 100000 201000 001070 0 Subgenus Paniculata 000700 000000 000000 000000 l L. pyramidalis (Lam.) Goldblatt subsp. pyramidalis 100102 011011 300101 010110 l subsp. regalis Goldblatt & Manning 100102 011011 303101 010100 l L. silenoides (Jacq.) Ker Gawl. 100101 011011 303101 110000 | L. verecunda il 100101 011011 302101 110000 l L. divaricata Br. 120101 020111 111021 110011 l L. spinosa (( Shree Goldblatt 120111 001111 111021 110071 l L. dolomitica Dinter 131101 001011 303011 010000 l L. violacea Goldblat 131101 001011 303010 110000 l L. tenuis ola) Goldblatt & 131101 00001 1 100017 110000 l L. pem N. E. Br. 110101 021011 303010 110100 l L. fabricii (Delaroche) Ker Gawl. 121101 020111 222010 111000 l L. barklyi Baker 111101 000111 023020 111000 l L. simulans Goldblatt & Manning 110101 000011 302010 110000 ] L. macrospatha Baker 110111 021011 202010 011000 l L. arenicola Schltr. 110111 021011 202000 110000 l L. littoralis Baker 100100 101011 301000 ?11010 ] L. anceps (L.f.) Ker Gawl. 100101 000011 302010 110000 l L. B Une Baker 100020 101011 301000 011010 l L. montana Klatt 100020 201011 300000 010010 l L. plicata (Jacq.) Diels 100020 101011 200000 010000 l L. oreogena Schltr. ex Goldblatt 100020 201011 303000 010000 l L. exilis Goldblatt 100111 001011 200100 ? 10000 l l. Corm tunics consisting of compacted fibers (0)—corm tunics woody (1) 2. Corm tunic bases lightly lobed (0)—bases minutely denticulate (1); bases coarsely and irregularly short-dentate (2); bases fairly regularly long-dentate (3) 3. Corm shape campanulate (0)—corm shape broadly conic (1) = Flowers cora (0) —flowers zygomorphic (1) . Plants with well-developed aerial stems (0) —plants forming fairly compact tufts (1): plants normally without aerial stems, i.e., acaulescent (2) 6. Seed surface cells unevenly colliculate-foveate and not in straight files (0)—cell surfaces colliculate and cells in files (1); cell surfaces tuberculate and cells in files (2) 7. Seed surface without secondary sculpturing (0)—surface with folds in a diffuse ruminate pattern (1); surface with a * regularly reticulate pattern (2 8. Capsules rounded in transverse section, thus without locular ridges (0)—capsules with winglike locular ridges (1); capsules with auriculate lobes decurrent on locular ridges (2 9. Branches borne well above the ground (0)—branches mostly or only at ground level (1) — 10. Lower tepals straight (0) —lower tepals geniculate (1 11. Leaf blades plane (0)—blades plicately ribbed (1) 12. Bracts soft-textured (0)—bracts firm-textured (1) 13. Perianth tube about as long as the tepals (0) —shorter than the tepals (1); 1.5-3 times as long (2); (3-)4-6 times as long (3 14. Perianth tube + cylindric throughout (not abruptly widened above) (0) —tube abruptly expanded above into a flared upper part (1): tube with the upper part wide and cylindric (2) Volume 83, Number 3 1996 Goldblatt & Manning Phylogeny in Lapeirousia Subg. Lapeirousia Table 1. Continued. . Perianth pale blue with white and dark blue markings (0)—predominantly white (sometimes with cream or blue markings) (1); cream with red markings and reddish on the reverse (2); dark red to purple or blue to violet (3) . Tepals ovate to lanceolate (0)—tepals spathulate (1) . Surface of lower tepals plane (0)—lower tepals each with a slender cusp near the base (1); lower tepals each with a toothlike ridge at base (2) 18. Upper tepal reflexed (0)—upper tepal arched (1) 19. Outer bracts without a median fold o r keel (0)—outer bracts with median fold or keel (1) 20. Inner bracts about as long as the outer (0)—two-thirds to half as long (1) 21. Flowers small (upper tepal usually less than 16 mm long) (0)—flowers large (upper tepal 18-27 mm long) (1) 22. Outer bracts acute (0)—bracts obtuse to retuse (1) (0)—flowers sweetly scented (1) d upper lateral tepals separating from the tube at the same level (0)—lower tepals joined to the 23. Flowers unscented 24. Lower an upper laterals for 3-5 mm (1) 25. Corm base rounded (0)—bases flat, corm thus campanulate (1) siphon as the outgroup, and one more character, campanulate corms, the generic synapomorphy for Lapeirousia (Fig. 3). Because Savannosiphon has been suggested to be the sister genus to Lapeirousia (Goldblatt, 1989) we were curious to see how the resulting trees would compare with those in which subgenus Paniculata is the outgroup. Trees were analyzed using CLADOS (Nixon, 1992) and the trees presented here were generated with this pro- am. The phylogenetic relationships were also recon- structed using PAUP 1 (Swofford, 1993) in or- der to utilize the bootstrap option and establish a measure of confidence in the results of the cladistic analysis (Felsenstein, 1985). A heuristic search was carried out saving minimal length trees only with the collapse zero length branches option in effect. The data were subjected to reweighting, using the maximum value of rescaled consistency indices, which is the same as successive weighting in Hen- nig86 (Swofford, 1993). A hundred replicates were run with the heuristic search option (limitations of time made it impractical to run more than 100 rep- licates), simple weighting, and with characters sam- pled randomly. The bootstrap values are presented on the strict consensus tree (Fig. 1B). RESULTS CLADISTICS On the trees illustrated (Figs. 2A, B), the char- acter distributions as mapped show subgenus Lap- eirousia to be supported by seven synapomorphic characters, four of which are autapomorphic and three homoplasious. The subgenus can thus reason- ably be presumed to be monophyletic. The auta- pomorphies for the subgenus include woody corm tunics, leaf blades with plicate ridges, floral bracts firm-textured, and inner bracts about half as long as the outer (characters 1, 11, 12, 20), all universal for the subgenus. Within subgenus Lapeirousia the strict consensus tree indicates a primary divergence into two clades, the smaller one of which (node 2) comprises all the species with derived seeds having primary sculpturing (character 7). The clade, which includes the type of section Sophronia (but not all the species assigned to it by Goldblatt & Manning (1990)), also contains the two tropical African spe- cies, L. littoralis and L. odoratissima. Within this clade a group of species are apomorphic in their acaulescent habit and actinomorphic flower, a re- versal which may well be closely associated with the acaulescent habit (Goldblatt, 1990b). In addi- tion, all species of the clade at node 2 have the derived basic chromosome number, x = 8. Al- though we see no a priori reason to believe that the clade is an artifact of the analysis, in the strict consensus tree generated using Savannosiphon as outgroup, L. littoralis is one clade of a trichotomy (Fig. 3) in which the acaulescent species comprise the second clade and the remaining species of the subgenus the third. Close association of the acau- lescent species and L. littoralis depends on how character 7 (seeds with primary sculpturing) is po- larized, and this will remain uncertain until more information can be obtained about the generic re- lationships of Lapeirousia. The second of the two primary clades (node 5) in both the trees with subgenus Paniculata as out- oup has either two synapomorphies (Fig. 2A), seeds with colliculate secondary sculpturing (char- acter 6), and bracts with a median fold or keel 350 Annals of the Missouri Botanical Garden s 2 g E z E 3: s » E s s 3 3 3 3 E 228.28 FS a z 3 ¢ #5 2 & § Ss 4 2 ^ $ $ P E E à ¢ NE $ E 28 € i P B ^ P 8 PE E > 9 89 $ $ oO = 9 S E 0 E E X S 6 Be > 9 $ d E 8 & D 3 B8 $ Be u 9 & & E à B ES d 4 4d d Jd dà d d d d d 3 d d d ài 8 d . 3 ob i = qd i [*] S E 4 ES 3 a 3 Ss g 383 8 4 2 = BE a E y E S a R 8 S8 LR » 8 5 t P &g 35 8 2 S Ec. g 5 B 83 uno 8 $ B B 8 ee B E . PF E Eg à $ 3 € BEE $ E p 5S 8 & 8 5 D wu P E & S8 du $ B à SE 8 S $ A SENEC NE WE m Da go E vt o © eo T T o NM o || a | eo <+ (ro) - ao © T Oo) N (o M N e o N E Figure l. Strict consensus trees generated from the character matrix (Table PER Tree generated using the mh* and bb* options of Hennig86.—B. Bootstrap values drawn on the strict consensus tree generated using the mh* and bb* options of Hennig86 followed by successive weighting. The bootstrap ie were calculated using PAUP and involving the maximum value of rescaled consistency index option (equivalent to successive weighting of Hennig86). (character 19), or one synapomorphy (Fig. 2B). at node 9 has either three characters (Fig. 2A): character 6. Within this clade there is again a pri- cusps on the lower tepals (17), perianth cream with mary dichotomy. Of the two resultant clades the one red markings (15), and the branching pattern (9), at node 6 is supported by one floral character, the all homoplasious; or four, the last being the outer spathulate shape of the tepals. However, the clade bracts folded or keeled (19) (Fig. 2B). All members —. Goldblatt & Manning Volume 83, Number 3 Phylogeny in Lapeirousia Subg. Lapeirousia E a * Š as S B E E i 3 E E echo 7” PME 2m0- ~ m- paa n “crt o“Egam E ons 3 a E] nus E e u a ao d O - E = = NEZ | 9 wm A fran oco s " 5 [5] a — “E wn E cea- As pad E) 3 e 2 ~e- à zae- 3 2 otia u a 2 -m — 8 L E 3 E 2 = E E = 5 : O E 4 S a E s ^" E a ar - 3 Sg m E DI E] " E | 2: me c © Y e : - 1 z = : a E E 3 d É a E: E E: 5 2 3 3 “5 E 5 E 9 3 a E aD CioYmam E a i 9 = = . 8 Sp d 8 œ he ^ - 3 - e * 2 d ~ ma o : zs d » d 10 d = d nN - | 9 caa- T B S N o o 5 ep- s E: 5 2 E a a eam 3 md 9 EL ls 2a aln 3 "T T — — Figure 2. Cladograms generated using the mh*, bb*, and successive weighting options of Hennig86 (Farris, 1988), e 1) and subgenus Paniculata as outgroup.—A. Str o the tree are shaded as follows: black = non-homopla = reversal. Numbers above character bars refer to character m mbers, those iE the des indicate the character state jacquinii shows poor resemblance to other members of the clade in gross morphology. It shares apo- morphic bracts with L. pyramidalis, has the ple- siomorphic corm type, and may have a hybrid ori- of this clade except Lapeirousia anceps, which is sister to the remaining species, and L. jacquinii are apomorphic in their corms with basal teeth or spines and chromosome numbers. In particular, L. 352 Annals of the Missouri Botanical Garden Savannosiphon _ Paniculata L. littoralis L. odoratissima -—— L. plicata — montana L. oreogena L. exilis L. silenoides L. verecunda L. pyramidalis subsp. regalis L. anceps in simulans L. jacquinii L. macrospatha L. arenicola L. dolomitica B violacea L. tenuis — 7 fabricii L. barklyi — L. divaricata Ls L. spinosa Figure 3. Strict consensus tree of seven equally parsimonious trees generated with the mh*, bb*, and successive weighting options of Hennig86 (Farris, 1988) with Savannosiphon as outgroup. The tree was generated using the same data matrix (Table 1) as the trees in Figures 1 and 2, plus one character, campanulate corms, apomorphic for all species of Lapeirousia. Scoring was changed for character 7, to reflect changed polarity. Volume 83, Number 3 1996 Goldblatt & Manning 353 Phylogeny in Lapeirousia Subg. Lapeirousia gin. The topology of the remaining species of the clade is consistent with our intuitive ideas about relationships. Terminal pairs in the strict consensus tree (Fig. 2A) and in the alternative equally parsimonious tree (Fig. 2B) are identical, except for Lapeirousia silenoides-L. verecunda and L. dolomitica-L. viola- cea (compare the topology at nodes 7 and 14 in Figs. 1 and 2). These species are terminal pairs only in Figure 2B. The number of steps to achieve the different topologies is the same (although the characters themselves differ), but we have no a priori reason to favor either alternative. The pos- sibility that the species are terminal pairs remains a reasonable hypothesis. In the tree with Savan- nosiphon as outgroup (Fig. 3) the tree topology is the same at all the major nodes except for the po- sition of Lapeirousia littoralis, discussed above. The terminal taxon pairs are the same as in Figure 2A. Bootstrap values (Fig. 1B) calculated with PAUP and based on a heuristic search with 100 replicates (Swofford, 1993) are drawn on the strict consensus tree obtained using Hennig86. Values above 70% are believed to have a 95% confidence level (Hillis & Bull, 1993). Values calculated for our trees range from 58% to 94%. The five terminal species pairs in the consensus tree have bootstrap values of 61% and higher. These values lend support to our spec- ulations about evolution and speciation in terminal clades obtained by the cladistic analyses and allow us to frame hypotheses about the mechanisms of speciation in subgenus Lapeirousia. Ultimately we hope that the hypotheses about phylogenetic rela- tionships presented here will be compared with phylogenetic reconstructions based on independent molecular methods of DNA sequencing or restric- tion enzyme analysis. The phylogenetic analysis does not support our previous division of subgenus Lapeirousia into two sections (Goldblatt & Manning, 1990). In this clas- sification species of section Sophronia fall in two major clades, those above nodes 2 and 6 (Fig. 1). Section Lapeirousia, however, corresponds to the clade above node 9. A preferable classification for the subgenus would be to restrict section Sophronia only to the species above node 2 and to recognize a third section for those at node 6, thus according sectional rank to the three major clades (Fig. 1) of the subgenus. A revised sectional classification based on our phylogenetic analysis will be included in a taxonomic revision of Lapeirousia currently in preparation. SPECIATION MECHANISMS IN TERMINAL TAXA There are five terminal pairs (Fig. 2A: nodes 4, 8, 12, 16, and 17) available for analysis on the Table 2. Distribution of differences (indicated by an asterisk, *) among taxa of terminal sister groups. Inferred differences in parentheses. 4 = L. gena-montana; TA = L. silenoides- dep 8 = cae ‘pyramidal subsp. regalis; 12 = L. macrospatha-arenicola; 14A = L. dol- mitica-violacea; 16 = L. fabricü-barklyi; 17 = L. divari- cata-spinosa. Node MA 4 7A 8 12 16 17 Allopatry * * = = > + 7 Parapatry = = » * ia T = Soi * » * * e * * Pollinator - T 1). comm 3 consensus tree obtained by successive weighting, and three of these are present on the unweighted consensus tree. The alternative hypothesis provides two more pairs (Fig. 2B: node 7A, 14A, the last one present in more than half of the most parsimonious trees). Moreover, these last two pairs are members of species trichotomies in the trees in which they are not terminal pairs. For analysis of biological differences, we compared the biology of the species at the terminal nodes to assess the factors that might have led to speciation of these pairs. The environmental determinants taken into account in this analysis are spatial separation, edaphic differ- ences, and pollinator divergence (Table 2). There is no separation in flowering time between the spe- cies in any of the pairs available for analysis (Gold- blatt et al., 1995) A geographical component (Fig. 4) is present in all of the terminal pairs and the sister species are either allopatric, i.e., separated by a significant geographic distance (nodes 4, 14A, and 17), or par- apatric, i.e., their ranges are adjacent although they do not grow intermixed (nodes 7A, 8, 12, and 16). In addition, members of most species pairs occur on a different soil type (nodes 7A, 8, 14A, 16, 17, but not those at nodes 4 and 12). Pollinator diver- gence characterizes three of the four parapatric species pairs (nodes 7A, 8, and 16) but only one of the allopatric pairs (node 4). There are floral differences between the species in all the terminal pairs except at node 17 (Lapei- rousia divaricata and L. spinosa). These differences are substantial at node 16 (L. barklyi and L. fabri- cii), involving morphology and markings, and ac- company a shift in floral type and pollination syn- drome between fly and bee pollination. Floral differences are less marked at nodes 4, 7A, 8, 12, and 14A and largely involve pigmentation and small changes in tube length. These relatively mi- Annals of the 354 Missouri Botanical Garden vVidlodv lSI1V83N39 JOrvW “Y :3vaimaaid "INOT "IN. :3VGOINIHLIS3IN3N/VSO7209 ‘Hd :3VOINV8SVL "83d "Ud :3VOINIBLS3IA3N “Yad "Ug :3VOINIHISS3IN3N (vaaioav) (3vaINvav1) (‘/9NO7 "JW :3VQGINIHLS3W3N) "83d "Hd ‘AON ‘dS "Sg :3VOINIHLS3IA3N (3vaiNvav1) "/9NO7 ‘WF :3VGOINIHLS3IN3N/VSO709 "Hd :3WOINV8V 1 ‘Yad “Yd :3VOINIHIS3IN3N OIU3132 "H :3VQIO9NIHdS (3VGINIB.LS3IW3N/3VQINV8VL) "U3d "Ug :3VOINIHIS3IN3N lSlIVH3N39 “AON ‘dS ‘Yd :3VOINIHLS3IN3N 1SITVyY3INI9 lSIlvH3N39 (3vai5NIHdS) (3vaiSNiHaS) HOLVNITIOd AV19 L1H9N SONVS ANYS IYVHV1V™ ANYS SWL ANYS SWL SONVS JILIZLYYND QNVS QNVS IH VHV1V» ANYS IHVHVT1VM ANYS SWL ANYS IHYHYIYJ ANYS SWL 3NOLSQNVS ANOLS AV'IO OOHVM 31VHS AYNISINIVIN 31INVHO VnO VINVN ANYS ONY 31IZLH VnO AVIO 31143100 AV'12 31193100 AV1IO OOHVOI ANYS IH VHV1VM ANYS IH YHV1V» AdAL NOS [ep] 8 8 8 8 8 VSONIdS VIVOIHVAIQ IA DIHvg ¡RENE VIIVIOIA VOILIWO10Q SINN3SL VIOOINAYV VHIVdSOHOVIN IININOD Vf SNVINWIS SddONV SI1V93MH 'd SITVGIIINVHAd 'd VONO23H83A S3GION3'IS SIx3 VN3903HO VNVLNOW VIVOlld VWISSILVYOdO SITVHOLILIT1 VIVINIINVA SNN3D8NS u YJJWNN S3l93dS FLY SILENOIDES-TYPE w a > - Ps o oc a < u > - u POLLINATION EQUIVOCAL SPHINX MOTH BEE/GENERALIST TIU = UJ ULL LUE Liu BES Volume 83, Number 3 1996 Goldblatt & Manning Phylogeny in Lapeirousia Subg. Lapeirousia nor differences in floral features either accompany a change in pollinator species within the same gen- eral pollination syndrome (fly pollination) (node 7A), or a shift between fly and generalist pollination (nodes 4, 8, and 16), or are not accompanied by shifts in pollinator at all (nodes 12 and 14A). Dif- ferences between the allopatric L. divaricata and L. spinosa (node 17) and the parapatric L. arenicola and A. macrospatha are mainly vegetative, and no change in pollination system is involved. Speciation in subgenus Lapeirousia, as far as can be assessed by comparing the terminal pairs in the cladogram, thus appears to be exclusively geo- graphic, either allopatric or parapatric. Speciation is usually combined with specialization for a dif- ferent soil type, and a shift in pollinator typically accompanies the speciation event in three of the four parapatric pairs but not in allopatric pairs (Ta- ble 2). Significant changes in flower architecture accompany shifts between bee and fly pollination. Only minor or no floral changes at all occur in shifts within a single pollination system. HISTORICAL ANALYSIS There are three potential problems with the eco- phylogenetic method for analyzing the evolutionary history of ecological features (Armbruster, 1992): stability of cladogram topology, completeness of ecological data, and circularity of reasoning. The reconstructed phylogeny of subgenus Lapeirousia is generally well supported (Fig. 1B), and the ecology of nearly all species is known (see Fig. 4). The edaphic hypothesis is free from the danger of cir- cular reasoning because soil preferences were not used to reconstruct the phylogeny. There is, how- ever, a possibility for circularity of reasoning re- garding the evolution of pollination types because a number of floral characters of obvious adaptive value in pollination were used to generate the phy- logeny. These are perianth tube length (character 13), perianth color (character 15), upper tepal ori- entation (character 18), and presence or absence of fragrance (character 23). In the absence of data in- dependent of the ecological data being analyzed, Armbruster (1992) has suggested that the indepen- dence of characters used can be assessed a poster- iori by checking for high consistency between the phylogeny and the ecological feature under consid- eration. In subgenus Lapeirousia the phylogeny and pol- lination system are poorly correlated suggesting that there is no circular relationship between the two. The consistency indices for the four characters related to pollination type are 0.25, 0.16, 0.33, and 0.20, respectively. These are at the lower end of the scale for morphological characters (0.20 to 1), un- related to pollination system, and also argue against any circularity in our analysis. Another challenging test of the ecophylogenetic hypothesis is to generate predictions based on the hypothesis and compare the predictions with new observations (Armbruster, 1992). In subgenus Lapeirousia our ecophylogenet- ic hypothesis includes the inference that the L. sil- enoides-type long-tongued fly pollination (Goldblatt et al., 1995) has evolved repeatedly. A prediction of this hypothesis is that the specialized corollas in these lineages differ from one another in detail, al- though they are functionally and superficially sim- ilar. This is indeed the case (see below). The eco- phylogenetic hypothesis about evolution of pollination systems in subgenus Lapeirousia is thus supported by all three tests. EDAPHIC DIVERSIFICATION Species of subgenus Lapeirousia show nearly complete substrate fidelity. It is most parsimonious to assume that coarse and sandy soil in general (four changes) rather than fine clay (five changes) was the substrate favored by the species ancestral to subgenus Lapeirousia, although it is not possible to infer whether this was Kalahari sand or Table Mountain Series sand (Fig. 4). Comparison of the habitat and distribution of subgenus Paniculata (Goldblatt, 1990b) with those of subgenus Lapei- rousia suggests the former. Ecological diversifica- tion to clay soils derived from shales or dolerite occurred within the clades defined by nodes 3 (L. plicata, L. montana, and L. oreogena), 7 (L. vere- cunda and L. pyramidalis subsp. pyramidalis), and 17 (L. spinosa). A single species, L. silenoides (node Figure 4. Pollinators, c 'hromosome numbers, soil types, and patterns of — plotted on the historical hypothesis m the cladogram shown in Figure 2. Long colonization lay substrates. Pollinators ME indicated in parentheses are inferred on the basis of floral morphology as detailed by Goldblatt et al. (1995). Pollinator names have been abbreviated as follow hynchus longirostris (Diptera: Nemestrinidae : H. = Hippotion (Lepidoptera: Sphingidae), P. = Parafidelia (Hymenoptera: Fideliidae), Ph. = Philoliche (Diptera: Tabanidae). Pr. = Prosoeca (Diptera: Nemestrinidae), M. longi. ). = Moegistor- 356 Annals of the Missouri Botanical Garden 7A), is virtually restricted to Namaqualand granite and granitic sands. Edaphic divergence character- izes five of the seven terminal species pairs and is likely to have provided the initial impetus for ge- netic differentiation between founder and parent populations. DIVERSIFICATION OF POLLINATION TYPES A well-developed floral tube and pale flower ap- pear to be ancestral in the subgenus, and pollina- tion by long-tongued insects A. be inferred to be plesiomorphic. The outgroup, subgenus Paniculat is diverse in floral morphology (Goldblatt 19906), but the floral types present in subgenus Paniculata, coupled with field observations (Goldblatt et al., 1995), suggest that pollination by bees or a com- bination of bees and Lepidoptera is predominant. Just three species in subgenus Paniculata have ex- tremely long floral tubes and display some of the same characteristics as species in subgenus Lap- eirousia that are known to be pollinated by sphinx moths or long-tongued flies. These species are, however, all believed to be derived in subgenus Paniculata, not only because of their derived floral characters but because of their apomorphic karyo- types (Goldblatt & Takei, 1993). H ence is strong that although both moth and fly pol- lination are present in the outgroup at low ence, the infer- frequencies, these pollination types are indepen- dently derived within each of the subgenera. In subgenus Lapeirousia, species richness is not a function of any one pollination strategy. Each of the major clades includes a diversity of pollination systems, and shifts in pollination system have oc- curred repeatedly. A pale flower (sphinx moth- and L. fabricii-type flowers) is evidently plesiomorphic at nodes 4, 6, and 9 (Fig. 4). The inference follows that the L. silenoides-type fly pollination (involving Prosoeca peringueyi and P. sp.), associated with viv- idly pigmented flowers, was derived four times in the subgenus, at node 4 (L. oreogena), node 7 pyramidalis subsp. regalis and L. silenoides), node 1 J. dolomitica and L. violacea), and node 11 (L. jacquinii). The L. fabricii-type fly pollination (in- volving Moegistorhynchus longirostris and Philoli- che gulosa) is inferred to have been independently derived three times within the subgenus, at node 7A (L. verecunda), and twice in the entire clade at node 9 (Fig. 4). It is most parsimonious to infer a shift to a generalist or bee-dominated system at node 13. In this scenario, a reversal to the L. fa- bricü-type fly pollination system occurred at node 16 (the L. silenoides-type fly pollination system evolved at node 14 of the clade). The L. fabricii- type fly pollination appears to be the ancestral pol- lination system for more than half the species in the subgenus (the entire clade defined by node 9). Bee/generalist pollination appears to be the most derived pollination state in the clade defined by node Thus, it appears that most of the pollination types found in subgenus Lapeirousia have evolved more than once. Bee or generalist pollination has evolved four or five times (node 3 or both nodes 3 and 4, and nodes 6, 8, and 10), L. fabricii-type fly 16), L. silenoides-type fly pollination four times (nodes 4, 6, 1l, and 14), and sphinx moth pollination once or twice (nodes 2 or 3 or both). Pollination systems pollination three times (nodes 5, 7A, anc in subgenus Lapeirousia are thus evolutionarily la- bile and prone to parallel evolution and reversal. This applies equally to shifts between bee and fly systems and to shifts between the two fly systems. e evolution of floral actinomorphy in subgenus Lapeirousia coincides with the evolution of an acau- lescent habit. Floral actinomorphy is restricted to the species at node 3, all of which are of low stature with a tufted, acaulescent habit. Except for L. odor- atissima these species grow on clay, and the colo- nization of more stable clay soils by the ancestor(s) of the remaining species in the clade may have facilitated the diversification of plants with this growth form. Acaulescence has the advantage of retaining the ovules and developing seeds at or be- ow soil level, thus providing protection from the e tufted growth form favors vertical flower presentation, and weather and predators (Burtt floral actinomorphy is a likely result. A long floral tube enables the flowers to project above the leaf and bracts. The development of long-tubed acti- nomorphic flowers in subgenus Lapeirousia is thus most likely related to plant habit and floral presen- tation and not to pollinator-driven selection. DISCUSSION Our analysis suggests that for the terminal spe- cies pairs in subgenus Lapeirousia speciation has been exclusively at the diploid level and has mostly involved a combination of shifts in substrate pref- erences when descendant species occur in adjacent habitats or geographic isolation when the descen- dant species are separated by significant distance and may or may not occur on different substrates. For the species analyzed the former situation slight- ly predominates. This pattern is consistent with the hypothesis (Goldblatt, 1978; Linder, 1985; Cowling & Holmes, 1992) that edaphic differences are the predominant cause of species richness in the south- Volume 83, Number 3 1996 Goldblatt & Manning Phylogeny in Lapeirousia Subg. Lapeirousia western part of southern Africa. The isolating effect of discontinuities in substrates (e.g., Kruckeberg, 1986) may in fact be the most significant cause of local and taxonomic diversity. It has been suggested by Linder and Vlok (1991) that each distinctive habitat may represent a “geographic” region with- out the usual physical barriers associated with al- lopatric speciation. Such microgeographic or para- patric speciation is particularly likely in the southwestern part of southern Africa where varied soils and topography accompanied by steep precip- itation gradients provide a strong selection differ- ential for population divergence and the develop- ment of edaphically isolated populations (Goldblatt, 1978; Linder, 1985) Local divergence between populations, whether linked to edaphic specialization or not, is possible only if gene flow is sufficiently restricted to permit genetic differentiation either by drift or selection. Limited gene flow is typical of most Iridaceae (Goldblatt, 1991), and short gene dispersal distanc- es are a characteristic of most plants in the south- western part of southern Africa (Cowling et al., 1992). The effect of restricted gene flow is that neighboring populations may be well defined as a result of selective differentials or random drift (Lev- in, 1981, 1993) Speciation in subgenus Lapeirousia seems to have been initiated by shifts in substrate prefer- ence. This has in most instances been accompanied by changes in floral morphology and shifts in pol- lination strategy, thereby enhancing genetic differ- entiation initiated by adaptation to different sub- strates. Edaphic differences in Namaqualand and the northwestern Cape are often associated with other physical differences such as altitude, aspect, and rainfall. They thus signify more profound niche differences than might be suggested by edaphic dif- ferences alone. The action of selection on founder populations on novel substrates should be corre- spondingly stimulated, and would enhance small- scale genetic differentiation. We anticipate that strong selective pressure for the development of prepollination isolating mechanisms would follow the development of these edaphically adapted ge- notypes at the contact zones between the parent and daughter populations. In other genera of Iridaceae and some Orchi- daceae and Amaryllidaceae small changes in color and scent are sufficient to attract different polli- nators and prevent hybridization (Johnson, 1992; Johnson & Bond, 1994; Paulus & Gack, 1990; Steiner et al., 1994). In subgenus Lapeirousia the floral changes necessary for a shift between pol- lination by sphinx moths and long-tongued flies or between the two types of long-tongued fly polli- nation primarily involve perianth pigmentation, while those necessary for a shift to bee pollination involve a shortening of the floral tube and some- times a change in flower color. Both flower color and flower shape are frequently governed by rather simple genetic systems (Gottlieb, 1984; Macnair, 989). A shorter floral tube is readily achieved through paedomorphosis, while simple hetero- chronic changes in development may account for other observed floral differences (Guerrant, 1982). In Aquilegia, for instance, many of the differences in shape that lead to the attraction of different pollinators are governed by small numbers of genes (Prazmo, 1965). or the species of subgenus Lapeirousia rela- tively simple mutational or recombinational events might result in a change in flower color from cream to magenta or violet (or vice versa), instant- ly shifting the plant into a different pollinator guild and effectively isolating it from the parent species. Potentially nonadaptive bottlenecks be- tween pollination systems are thus likely to be small, and pollinator-directed selection for differ- ences in floral morphology should be rapid. This is particularly significant in a situation where flo- ral conformity is as strongly stressed as in the long-tongued fly guilds (Manning & Goldblatt, 1996). Although shifts in pollination system in subgenus Lapeirousia involve a complete change in the pollinator species, there can be occasional overlap. For example, we have seen long-tongued fly species visiting short-tubed flowers and some bees occasionally collect pollen from long-tubed species, which they may successfully pollinate. In addition, Prosoeca peringueyi has been observed visiting a few individuals of L. verecunda at the single known site where the ranges of this species and L. silenoides meet. It is thus possible that shifts between the various pollination systems could have occurred with intermediate phases in which both old and new systems operated. Be- cause of the apparently small morphological/ge- netic changes required for shifts between the var- fal ious pollination systems in subgenus Lapeirousia, the shifts are likely to have been ra We postulate that a source of strong differedtial selection for the incipient species to develop spe- cific pollination systems was provided by preexist- ing and independently established pollination guilds involving long-tongued flies in the families Nemestrinidae and Tabanidae. An increase in tube length would enable a generalist species to enter one of these guilds by excluding insects with short- er tongues from competing for nectar resources. A 358 Annals of the Missouri Botanical Garden decrease in tube length would facilitate the devel- opment of a generalist or bee-dominated pollination system. Shifts between the two long-tongued fly pol- lination systems require only a change in flower color, and plant species can thus shift readily be- tween them. The minor morphological changes re- quired to accomplish a shift between these polli- nation systems seem to have permitted rapid establishment of reproductive isolation between the diverging populations. Early stages in floral diver- gence by pollinator-directed disruptive selection are evident in L. pyramidalis. The two subspecies are characterized by differences in floral morphol- ogy (scent, flower color, tube length), which, al- though minor, enable them utilize different polli- nation systems. Although prepollination isolating mechanisms include seasonal and habitat differences, as well as the use of different pollinators or different body parts of the same pollinator (Levin, 1978), separa- tion in flowering period between related species pairs has not developed in subgenus Lapeirousia. This is probably due to the short rainy season and hence short period for biological activity for most organisms in western southern Africa. Species of subgenus Lapeirousia with the same flower type and which belong to the same pollina- tion guild rarely co-occur. In the few instances where two species with flowers of the same general type grow together (e.g., L. jacquinii and L. violacea; L. fabricii and L. anceps), the two differ in their chromosome number and no hybrids have been lo- cated (Goldblatt et al., 1995). We suggest that one consequence of this genetic isolation was that the genotype of populations colonizing new habitats would not be swamped by neighboring populations with a different chromosome number even in the absence of different pollination systems. de- crease in chromosome number characterizes the clades defined by nodes 2 and 13. These clades may thus represent repeated bursts of parallel ra- diation from founder species that had become ge- netically isolated from the ancestral stock. Despite marked differences in floral morphology between closely related species in subgenus Lap- eirousia, comparisons of biological differences with- in species pairs in the subgenus indicate that the diversity of flower types and pollination strategies occurred after the differentiation of local popula- tions to new habitats. We hope that the hypotheses about phylogenetic relationships presented here will be compared with phylogenetic reconstructions based on independent molecular methods of DNA sequencing or restriction enzyme analysis. character oe Syst. Zool. 18: ———. 1088. Literature Cited Armbruster, W. S. 1992. Phylogeny and the oo of plant- ee interactions. BioScience 42: 12-2 —— 1993. Ev jaaa of plant pollination ae tia ad tests with the neotropical vine Dale- Mains Evolution 47: ^ 1480 )-1505. ——. 4. Early evolution of Dalechampia (Euphor- TES Insights from phylogeny, biogeography, and comparative ecology. Ann. Missouri Bot. Gard. 81: 302- 31 316. Arnold, T. H. & B. C. de Wet (Editors). 1993. Plants n > Africa. Names and Distribution. Mem. irv. S. Africa No. 62. National Botanical Institute, rin Boobs D.R. & A. McLennan. Ecology and Behavior Univ. Chi 1991. P icago Press, Chica Burtt, B. L. 1970. The evolution aa taxonomic signifi cance of a o ova ertain monocotyle- dons ae J he ). Cowling, R. P. M. Holmes. 1992. Endemism and speciation in a b od flora from the Cape Floristic Region. Biol. " ve Soc. 47: 367-383. E >. Rebelo. 1992. Plant diversity and endemism. Pp. os 112 in R. M. Cowling (editor), The Ecology of Fynbos. Oxford Univ. Press, Cape Town. Donoghue, M. J. 1989. Phylogenies and the analysis of evolutionary sequences, with examples from seed plants. Evolution 43: 1137-1156. Farris, J 969. A successive oo approach to 85. lennig86 version 1.5. ram and soft- ware doc 'umentation. Published by the author, Port Jef- ferson Station, New Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783- lí P. 1972. A revision of the genera Lapeirousia Pourret and Anomatheca Ker in the winter rainfall re- gion of c Africa. Contrib. Bolus Herb. 4: 1-111. An analysis of the flora of southern Africa: Its dd. 's, relationships, and origins. Ann. Mis- souri Bot. Gard. 65: 369—430. —— 1098 The genus Watsonia. A aed mono- Prea Ann. Kirstenbosch Bot. Gard. -148. 1990a. Cytological uns in ds African ge- nus Lapeirousia (Iridaceae-Ixioideae). Ann. Missouri Bot. 77: 375-382. Ob. Systematics of Lapeirousia acia Ixioideae) in tropical Africa. Ann. Missouri Bot. Gar 77: 430484. . An overview of the systematics, phylogeny and biology of the African Iridaceae. Contrib. Bolus erb. 13: 1-74 1 Manning. 1990. Leaf and corm tunic structure in Tones (Iridaceae-Ixioideae) in rela- tion to phylogeny and infrageneric classification. Ann. Missouri Bot. Gard. 77: 365-37 & 199; Eternas of the um Af- rican ¡AAA pms ee (Iridaceae—I plex (sect. Fastigiata) and a new species of sect. Pan- iculata. S. African J. Bot. 58: 326-336. & . New taxa and revisions to the taxonomy of southern African Lapeirousia subgenus Lapeirousia (Iridaceae subfamily Ixioideae). Novon 4: 339-346. Volume 83, Number 3 Goldblatt & Manning 359 Phylogeny in Lapeirousia Subg. Lapeirousia — & M. Takei. 1993. Chromosome cytology of the African genus oe (Iridaceae-Ixioideae). Ann. Missouri Bot. Gard. 80: 961-973. , J. C. Manning & P. Bernhardt. 1995. Pollina- tion in Lapeirousia subgenus Lapeirousia (Iridaceae: Ixioideae). Ann. Missouri Bot. Gard. 82: 517-534 Gottlieb, L. D. 1984. Genetics and morphological evo- lution in plants. Amer. Naturalist 123: 681—709. Guerrant, E. O. 1982. Neotenic evolution of Delphinium nudicaule (Ranunculaceae): A hummingbird-pollinat- ed larkspur. Evolution 36: 699-712. Hillis, D. M. € J. J. Bull. 1993. An empirical test of bootstrapping as a method for nie confidence in phylogenetic analysis. Syst. Biol. 42: -192. Johnson, S. D. 1992. Plant-animal diei Pp. 175-205 in R. M. Cowling (editor), b Ecology of Fynbos. Oxford Univ. Press, Cape Tow . J. Bond. 1994. Red neus and butterfly pollination i in the fynbos of South Africa. Pp. 137-148 Groves (editors), Plant- 1986. The stimulus of unusual ge- ologies for plant speciation. Syst. Bot. 11: 455—463. 978. The origin of ee mechanisms in flowering plants. Evol. Biol. 11 17. ——. 1981. Dispersal versus gene ens in plants. Ann. es Bot. Gard. 68: 233-253 —— Local speciation in plants: The rule not the exception e Bot. 18: 197-208. Linder, H. P. 1985. Gene flow, speciation and species diversity ieri in a species-rich area: The Cape Flora. Pp. 53-57 in E. S. Vrba (editor), Species and Speciation. Transvaal Museum Monographs 4. Trans- vaal Museum, Pretoria J. Vlok. 1991. The morphology, taxon- omy and oe Z o (Restionaceae). Pl. Syst. Evol. 175: 160. Macnair, M. R. Ds "Ts p for rapid speciation —210 ; 1992. Pollinators and evolution in Disperis, or why are there so many s cies? S. African J. Sci. 88: 38-49. — —— & P. Goldblatt. te The Prosoeca An (Diptera: Nemestrinidae) pollination syndrom southern Africa: Long-tongued flies and their d. flowers. Ann. Missouri Bot. Gard. 83: 67-86. Nixon, K. 1992. CLADOS 1.2 IBM PC-compatible ch acter analysis program. Published by he: auido Ith- aca, r Paulus, H. F. & C. Gack. 1990. dien, ee as dca linating isolating factors: Evolution and speciation in Ophrys eyes te Israel J. Bot. 39: 43-79. Prazmo, W. Cytogenetic studies in the s Aquilegia. us oO. ai of the traits distinguishing different complexes in the genus Aquilegia. Acta Soc. Bot. Poloniae 34: 403-437. Steiner, K. E., V. B. Whitehead & S. D. Johnson. 1994, Floral and pollinator divergence in two «os de- ceptive South African orchids. Amer. J. Bot. 81: 1 194. Swofford, D. L. 1993. PAUP: Phylogenetic Analysis Us- ing simu Version 3.1.1 Illinois Natural History Survey, Champa pendix. Character states and character polariza- tion. "The outgroup S character polarization i is Lapeirousia character 25, used in a secondary 3 [7 E e un © e. © = a di [a] thus and Micranthus) (Goldblatt & Manning 2. Entire to lobed corm bases E i hi species of subgenus Paniculata and several in subgen pei- rousia. Minutely serrated (denticulate) tunic ght occur in a number of species (Goldblatt, 1972), and the devel- opment of short spines (short- dentate) ww in L. divar- icata and its close ally L. spinosa, while L. dolomitica, L. violacea, and L. tenuis have long spines extending from the corm bases. The nature of the corm bases is treated as a multistate character with denticulate scored as (1), t dentate as (2), and long-dentate as (3). The degree rration, minutely toothed to short- or long-dentate, is albo as additive and hence ordered. 3. Campanulate-shaped corms occur in all species of cies of mee Lapeirousia, L. tenuis, L. dolomitica, and L. violac 4. The Boadonorphic: — in subgenus Panicu- x is probably an acti rphic flower (Goldblatt & g, 1990; ‘Goldblatt E "Takei, 1993), but we are suf- Asidi: uncertain about the state in the outgroup that we prefer to score the ae as (?). Most species of sub- genus Lapeirousia have zygomorphic flowers, but four spe- cies that also have the aerial stem reduced (thus acaules- cent) have actinomorphic flowers. Elsewhere the E pus er pas been interpreted as a reversal rom the rphic condition, attendant on the acau- occid habit “Goldblatt, 1972. 1990b). 5. The BE habit. and densely tufted growth abit chara ditions. ER growth em is unknown in sub- 6. s the png related genus Savannosiphon, possibly the sis- ter genus to Lapeirousia (Esla 1989). are lightly and irregularly colliculate-foveate and the cells are not ar- ranged in regular files (Goldblatt & Manning, 1992). I and regularly colliculate, scored as (1), o se sored a as (2), both presumably derived states. The pat- ecies Is score isna acata aee no ans in the topology of the strict consensus tree but it is one step longer). It may he argued that the irregular or regular ordering & Annals of the Missouri Botanical Garden of cells in files and the surface sculpturing are separate characters, but they are fully correlated, and hence w prefer to treat them as a single one. The extent to which the surface cell wall is raised, either moderately (colli- culate), or strongly, especially toward the center (tuber- culate), is regarded as additive and hence this multistate character is ordered. 7. Seeds with primary surface sculpturing are rare in Lapeirousia, and seeds in subgenus Paniculata are glo- bose, the plesiomorphic condition, except in L. otaviensis t in negle cta Goldblat tt & Manning, which a diffuse ruminate pattern in L. plicata and L. odoratissima, scored as (1), and h i L. oreogena an montana, scored as (2). states are treated as additive with the diffuse pu seen as a less specialized type of sculpturing than the E reticulate pattern. 8. Strongly developed winglike ridges are present on the locular sutures of the capsules of Lapeirousia sileno- ides, L. verecunda, and L. pyramidalis. These are regarded us cent species all branch from the - e by default en 'ause the aboveground internodes are c 'le 10. Geniculate lower tepals (with a sharp bend near the midline) are treated as apomorphic. In Lapeir eirousia the character i is restric sted to a idis of aperies of subgenus n siomorphic: sau for Lapeirousia i isa ee leaf. A jew . Firm- textured floral bracts are characteristic of the species of subgenus Lapeirousia and contrast with the soft- textured bra of section Panic ulata. Firm-textured bracts also occur in section Fastigiata of subgenus Pan- iculata, and appear to be derived here. Based on outgroup comparison, the firm- inte bracts of subgenus Lapei- rousia are regarded as deriv 13. The perianth tube is shoved to about as long as the tepals in most species of subgenus Paniculata, and this e tepals is scored as (1); a tube slightly longer, to 2-3 times as long as the tepals is scored as (2); a tube (3-)4—6 times as long as the tepals is scored as (3). 14. The development of a perianth tube with an abrupt- ly widened throat or gullet i is enar to a few species of subgenus Lapeirousia (L. fabricii, L. barklyi, L. divari- cata, L spinosa). Species of scies us Paniculata mostly have a uniformly cylindrical tube, sometimes gradually widened toward the apex, or a widely Minn -shaped tube. gullet in vasmontana) may be the plesiomorphic state for Lapei- rousia, and is scored as (0). The more or less cylindrical tube of the zygomorphic- flowered spec ies of subgenus Paniculata, as well as in most species of subgenus Lap- eirousia, is scored as (1) and the presence of a gullet as 2). The character is treated as ordered as we regard the trend in the elaboration of the tube as additive. 15. The predominant perianth color in the outgroup, subgenus Paniculata. is blue with — predominantly white, sometimes with blue or mauve mark- ings and/or a lilac flush on the reverse of the tepals (1); white to cream to ivory with red markings on the lower tepals, and on the reverse of the tube and sometimes of epals (2); shades of dark red to purple or violet to blue, | in either case with contrasting white markings (3). The character states are unordered. sed on outgroup comparison the more or less spathulate tepals of Lapeirousia silenoides, L. verecunda, and L. exilis are apomorphic. The plesiomorphic Eon tion, ovate to lanceolate tepals, ron 'lerizes the species of subgenus Paniculata and the other species of Sb ems i de . The development of a tooth or ridge of raised tissue rived. Filiform teeth occur in Lapeirousia p L. fa- bricit, L. simulans, L. macrospatha, and are weakly developed in L. jacquinii. Ridged teeth occur in L. pid icata, L. spinosa, and L. barklyi. We assume these struc- d uae two states of the same character; filiform a are scored (1), and ridged teeth as (2). 18. Erect to arched upper tepals, restricted to a few species of subgenus Lapeirousia, are regarded as rin Species of subgenus Paniculata and many of subgen i nih have all the tepals extended and lying in more EU ess the same plane . Bracts strongly ‘folded in the midline or keeled, Noi in several species of subgenus Lapeirousia, are treated as derived; the bracts of Ub Paniculata are neither folded nor keeled. 20. The inner floral bracts are about as long as to slightly shorter (but narrower) than the outer bracts in all species of the outgroup. Species of subgenus Lapeirousia always have the inner bracts about half as long as the outer or less, hence are regarded as apomorphic for the character. 21. Small to aue sized flowers (excluding the perianth tube) are most likely plesiomorphic for the out- group (upper tepal less Pe 18 mm long). A few species of subgenus Lapeirousia, notably L. fabricii and L. barklyi, have iD larger flowers, the upper tepals being 18—27 rphic. s are acute in other ubgenus and in ke ak Paniculata. 23. Sce ied iud rs of, for subsp. pyramidalis, regarded as derived, rd flowers are rare in subgenus Volume 83, Number 3 1996 Goldblatt & Manning 361 Phylogeny in Lapeirousia Subg. Lapeirousia Paniculata and are apomorphic within the subgenus. There is no information available about scent che] in Lapeirousia, and although scents may differ in subgenus Lapeirousia we simply score presence or absence of any scent. subgenus Paniculata (and most species of subg Lapeirousia) the lower and upper lateral tepals separate from the tube at 2 EE same level, er plesiomorphic condition. In L. diva and L. spinosa the lower and upper lateral tepals a are united for 3— 5: mm, apomorphic for these two spec 5. For the pip with Savannosiphon as outgroup. Corms with flat bases are unique in Iridaceae to Lapeirou- sia and thus scored as apomorphic (1 PHYLOGENETIC RELATIONSHIPS, SEED CHARACTERS, AND DISPERSAL SYSTEM EVOLUTION IN AMARYLLIDEAE (AMARYLLIDACEAE)! D. A. Snijman? and H. P. Linder? ABSTRACT The phylogeny of the mostly African tribe Amaryllideae is presented as a basis for classification and an enquiry into the evolutionary aj between seed and fruit characters and seed dispersal modes. is based on morphological, Three dispersal nes anemoc and the patt monophyletic s (Amaryllis, Nerine, Brunsvigia, Crossyne, Hessea, S dinae the ability of the green, chory, atelechory/rain shed, wind-tumbled infructescences in three clades espec ially in semiarid, southwestern Africa tively slow developing. The limited seed dispersal mode in Crinum a consequence of the derived indehiscent fruit, which is basal in Crininae, and the are large, cork-covered, endosperm-rich, and rela and Ammocharis appears to be a anchoring function of their persistent, lax scape The cladistic analysis eed anatomical, an cytologic val den of which seed and fruit pesi cs are illustrated. lec wash, an ern is considered in terms of functional frui and ed morphology. The new classifica ation recognizes two subtribes: Crininae (Boophone sensu stricto autochory, are optimized o e selected cladogram Crinum, Mq d and Cybistetes) and Amaryllidinae hylogeny suggests that in Amarylli- r they are i aid favored the development of rapidly In Crininae the seeds Amaryllideae, a monophyletic group consisting of 11 currently recognized genera and approxi- mately 155 species, is one of nine tribes (sensu Dahlgren et al., family Amaryllidaceae. The members are a 1985) in the monocotyledonous ul- bous herbs with distichous or rarely rosulate leaves and a lateral, solid, naked scape which is termi- nated by an umbel-like cluster of flowers. Their seeds are characteristically large and nondormant. Although remarkably uniform in floral and vegeta- tive morphology, much diversity exists in the tribe’s fruiting structures, some of which function as spe- cialized units of dispersal in wind or water. Mem- bers of the tribe inhabit grassland, savanna, and tropical forests in sub-Saharan Africa, but are most speciose in southern Africa, where habitats include semiarid, dwarf shrublands in a winter-rainfall cli- mate. The tribe’s only pantropical representative is Crinum L., which extends from Africa to Madagas- car, the Mascarene and Pacific Islands, and the tropics of America, Asia, and Australia. Several Amaryllideae have traditional and eco- The charred and crushed bulb scales of Ammocharis coranica (Ker Gawl.) Herb. have been used as an adhesive (Phillips, 1917, 1938), and a decoction of the bulbs of Boophone Herb., Crinum, and Nerine Herb. have provided medicine and poison to many African peo- ple (Phillips, 1917; Watt & Breyer-Brandwijk, 1962). The use of the highly toxic bulbs of Boo- phone disticha (L.f.) Herb. for arrow poison by early nomic uses. Brunsvigia Heist., hunter-gatherers has been particularly well docu- mented (Bradlow, 1994; Forbes, 1986; Forbes & ourke, Numerous forms and hybrids of Nerine, Amaryllis L., and Crinum are cultivated for their elegant flowers, and the spherical fruiting heads of Boophone and Brunsvigia are locally gath- ered for dried floral arrangements. Amaryllideae were first delimited as monophy- letic by Traub (1965, 1970), who recognized the presence of threadlike “fibres” and bulb scales as diagnostic for the tribe (char- acter 1). This feature has since been corroborated in torn-off leaves by four non-homoplasious synapomorphies: uniteg- ! This study i a ae cation of part of a Ph.D. thesis presented to the rra of Botany, University of Cape y m oF P. L. Perry, A. J. Mannin rson-Jones, . 8 Romanowski, sae S J. E. Wand for help during this study A. W. Meerow is ler for valuable comments on the Town, South A . The first author thanks S. E. Foster manusc is ? Compton Herbarium, National Botanical Institute, P. Ba 7, Claremont 7735, South Africa. 3 Bolus o. University of Cape Town, Rondebosch a South Africa. ANN. MISSOURI BOT. GARD. 83: 362-3806. 1996. Volume 83, Number 3 1996 Snijman & Linder Phylogenetic Relationships in Amaryllideae Table 1. Major recent classifications of Amaryllideae. Crinum, Ammo- Amaryllis, charis, Cybis- erine, Hessea, tetes, Boophone Brunsvigia Carpolyza, Strumaria Crossyne Traub (1957) Crininae Crininae Strumariinae Lae eae Crineae Crineae Hesseae ub (1963) Crineae Crineae Strumarieae e (1965, 1970) Crineae Crineae Crineae Dahlgren et al. (1985) ir pel Amaryllideae Amaryllideae D. Müller-Doblies (1985) Crinin Crininae Strumariinae Snijman & Linder Cates Amaryllidinae Amaryllidinae Amaryllidinae mic ovules (Huber, 1969) and bisulculate pollen grains (Schulze, 1984) with spinulose exine sculp- turing (Meerow, 1995), to which we add green em- bryos in nondormant, water-rich seeds (character Various subtribal classifications of Amaryllideae have been proposed (Table 1), most of which reflect the nomenclatural confusion introduced by Traub's (1957, 1962, 1963, 1965, 1970) continuous mis- application of the name Amaryllis to the unrelated genus Hippeastrum Herb. (Brummit, 1987). These systems differ only in the rank attributed to the two subtribal groups recognized in Amaryllideae (Cri- neae sensu Traub, 1965, 1970). Thus far, the rela- tionships of the genera within these groups have never been disputed. Amaryllidinae (Pax, 1887) or Crininae (Traub, 1957, 1963), the larger of the two subtribes, are characterized by the presence of minute “fibres” between the broken parts of leaves and bulb scales, a solid scape, and regular to zygomorphic flowers (Traub, 1963). Strumariinae, which are defined by no single character but by a combination of char- acters, comprise small plants with mostly two leaves surrounded by an amplexicaul cataphyll; regular flowers with a reduced perigone tube; fila- ments inserted into a sheath formed by the anther connective; and style variously connate to the fila- ments (Müller-Doblies, 1985) Essentially these interpretations were limited to basic vegetative and floral morphological data and they were phenetically based; therefore the mono- phyly of the subtribal groups was never explicitly questioned. The primary aim of this study was to use new morphological data from the infructescences, cap- sules, and seeds of Amaryllideae in a phylogenetic analysis to establish monophyletic groups of gen- era, and to assess the monophyly of the traditionally recognized subtribes Amaryllidinae and Strumari- inae. Of special interest in Amaryllideae are the unique and complex seed dispersal mechanisms by 363 wind or water. To analyze the evolution of these dispersal mechanisms, changes in seed and fruit characters are traced on the phylogenetic tree to assess whether the different dispersal modes may have been facilitated by these functionally related characters. Lastly, in an attempt to establish evo- lutionary patterns in the ecology of Amaryllideae, we use the phylogeny to analyze the interrelation- ships between seed and fruit characters, dispersal traits, phenology, and phytogeography. MATERIALS AND METHODS CHARACTER ANALYSIS Morphological data on Ammocharis Herb., Brunsvigia, Crinum, and Nerine were obtained from the literature, living collections at Kirstenbosch National Botanical Garden, and dried material at NBG. Crinum variabile (Jacq.) Herb. and all spe- cies of Amaryllis, Boophone, Crossyne Salisb., Car- polyza Salisb., Cybistetes Milne-Redh. & Schweick., Hessea Herb., and Strumaria Jacq. were studied in the field, and the South African summer-rainfall representatives of Crinum and Ammocharis were studied in cultivation. Leaves, flowers, and seeds for anatomical study were fixed in FAA (90 ml 7096 ethanol; 5 ml glacial acetic acid; 5 ml formaldehyde), subsequently de- hydrated in an ethanol-toluene series, embedded in wax, and cut with a rotary microtome at 10-15 jum. Sections were stained with safranin and alcian blue, and photographed with a Zeiss Axioskop. Seeds examined by scanning electron microscopy were fixed in FAA, dehydrated in ethanol, and crit- ical point dried before sputter-coating with Au/Pd. Photographs were taken with a Cambridge S200 SEM at 10 kV. Documentation for the taxa studied micromorphologically is given in Appendix I. The micromorphological characters reported here for C. variabile conform to published results of other Af- rican, Asian, and Australian species of the genus (Dutt, 1962, 1970; Howell & Prakash, 1990; Merry, Annals of the Missouri Botanical Garden Distribution of charac tor states among the genera of Amaryllideae and subgenera of Hessea and Strumaria. Table 2. Indeterminate states are coded as Haemantheae 00000 00000 inum 10100 00100 Ammocharis 11100 00100 Cybistetes 11100 00100 Boophone 10000 00000 Amaryllis 10000 00000 rine 10000 00000 Brunsvigia 11000 01010 rossyne 11001 01210 Hessea subg. Hess 10010 00011 Hessea subg. ona unis 10010 00011 Hessea subg. Kamiesbergia 10010 00011 Strumaria subg. Strumaria 10000 00000 Strumaria subg. Tedingea 100?0 00010 Strumaria subg. Gemmaria 10010 100?0 Carpolyza 10000 00010 00000 00000 00000 00000 0000 20000 11100 10011 10000 1110 00000 11000 10011 10000 1110 00000 11100 12011 10000 1110 00000 11000 10111 00000 1110 11000 11100 10000 00000 0010 12000 11100 11000 00111 0010 12000 11100 12000 00111 0010 ?2000 11100 12000 00111 0010 02002 11000 12000 00111 0010 01101 11000 12000 00111 0010 0201? 11000 12000 00111 0010 00000 11011 10000 00111 0011 00000 11011 11000 00111 001? 00001 1011 12000 00111 0011 00001 11001 11000 00111 0011 1937; Toilliez-Genoud, 1965; Venkateswarlu & Lakshmi, 1978). New seed data are provided for Boophone, Ammocharis, Brunsvigia, Crossyne, Hes- sea, and Strumaria. Of the 34 morphological characters identified for the phylogenetic analysis, 5 have slightly overlap- ping states: leaf number, pedicel length, anther at- tachment, staminal tube length, and scape habit af- ter seed set. The sets of different states recognized in these characters nevertheless proved to be in- formative, since the topology was considerably less resolved when they were excluded from the phy- logenetic analyses. Only one of the four multistate characters was treated as unordered. Autapomor- phies were excluded from the cladistic analyses but were later inserted onto the cladograms. Character state distribution is provided in Table 2 CLADISTIC ANALYSIS Cladistic analyses were carried out using the im- plicit enumeration (ie*) option of Hennig86 (Farris, 1988), and the tree bisection reconnection (TBR branch swapping of PAUP (Swofford, 1993). Trees were rooted by designating an outgroup (Nixon & Carpenter, 1993). Anderberg and Tehler (1990) suggested that for taxonomic studies only strict con- sensus trees, containing only components retrieved by each of the complete set of most parsimonious trees, should be used. This is a very conservative approach, as these consensus trees are substantial- ly less informative than the most parsimonious trees on which they are based (Farris, 1983), and they reveal little about process, as in character state change (West & Faith, 1990). Carpenter (1988) therefore argues that one of the most parsimonious cladograms should be preferred over a consensus tree as a phylogenetic hypothesis. To choose among the set of most parsimonious trees, we use the logic that adding to the potential homoplasy of a char- acter which is already highly homoplasious should not be as significant as adding to the potential ho- moplasy of an otherwise non-homoplasious char- acter. Consequently, we select those topologies from the set of most parsimonious trees that minimize the possible homoplasy of the least homoplasious characters. There are several approaches by which this logic can be implemented. Goloboff (1993) de- veloped a method for simultaneously downweight- ing homoplasious characters and locating parsi- monious trees, but this has been criticized by Turner (1995). Meacham (1994) proposed a method for calculating the character compatibilities in a data set, and weighting those characters that are maximally compatible. We used the successive ap- proximations weighting scheme implemented in Hennig86 using the rescaled consistency index: the product of the character consistency index and the character retention index (Farris, 1969; Carpenter, 1988). Maximally homoplasious characters, if they occur in only two taxa, would have a consistency index (CI) of 0.5, a rescaled consistency index of 0, and a weight of 0. Minimally homoplasious char- acters, with CI = 1 and retention index (RI) = 1, would have a rescaled consistency index of 1 and a weight of 10. In extreme cases Goloboff (1991a, b) has questioned the validity of using the rescaled consistency index as a measure for deciding which most parsimonious cladogram is to be preferred. Despite these reservations, successive approxima- tions weighting analysis is considered to have value when used with caution (Crisp € Weston, 1995). Our preferred most parsimonious tree was com- Volume 83, Number 3 1996 Snijman & Linder Phylogenetic Relationships in Amaryllideae pared with the majority rule tree and the strict con- sensus tree, and decay analysis (Bremer, 1988; Donoghue et al., 1992) was conducted to provide measures of strength for the phylogenetic hypoth- esis represented by the most parsimonious tree to- ology. To investigate alternative hypotheses of monophyly, the additional cost of grouping Ama- ryllis within the Boophone clade was determined using MacClade (Maddison & Maddison, 1987). Character distributions on cladograms were exam- ined, and figures were generated using CLADOS (Nixon, 1992). To analyze the number of times a particular dis- persal mode evolved and to determine its interre- lationship with fruit, seed, and phenology states in the lineage, the dispersal modes and phenological states were plotted on the terminal taxa of the cladogram and the most parsimonious interpreta- tion of the character states at each inner node was obtained using Farris optimization (Farris, 1970; Mickevich, 1982). The states coded for seed dis- persal are taken from Van der Pijl (1982): anemo- chory (wind dispersal), which is divided into ane- mogeochory (tumbling) and anemoballists (wind ballists); atelechory (limited dispersal); and autochory (dispersal by the plant itself), which is used here in the passive sense. For the analysis of plant phenology, the foliage leaves were coded as synanthous (the leafing period coincides with the flowering period) or hysteranthous (the leafing pe- riod is separated from the flowering period). All known seed and fruit characteristics were in- cluded in the data matrix for phylogeny construc- tion, but for the analysis of dispersal mode only the abscission of the infructescence (character 22) was used. Following Deleporte (1993), the primary cod- ing of such a character may not necessarily intro- duce circularity into our interpretation of dispersal system evolution. CHARACTER CODING The weightings determined by the successive ap- proximations character weighting routine are given in parentheses and autapomorphies are marked with an asterisk. 1. (10) Bulb scales: without extensible elements when pulled = 0, with numerous extensible elements when pulled = 1. Although Traub (1965, 1970) stated that “fibres” appear when the bulb scales and leaves of Amar- yllideae are torn apart, an unpublished report (P. Goldblatt, pers. comm.) that these highly extensible elements are long, spirally thickened tracheids is shown to be correct. Before elongation the tracheids are ca. 325 pm long and at all stages the wall thickenings are characteristically lightly stained by safranin. When artificially stretched the cottony ap- pearance of these elements is reliably diagnostic for the tribe. 2. (3) Outermost bulb scales: fleshy to parchment- like = 0, hard and brittle = 1 In the outgroup and most representatives of Amaryllideae, the outer tunics form brown, softly fibrous coverings. Bulbs with extremely thick cov- erings, which become hard and brittle, are found in Ammocharis, Cybistetes, Crossyne, and the Bruns- vigia species with prostrate leaves. Although the chemical substances in these thick, dark brown or tan-colored tunics are unknown, these secondarily thickened bulb tunics are coded as a single state. 3. (10) Leaf habit: annual = 0, lasting longer than a year = 1. A strictly annual growth pattern, in which just one set of foliage leaves is produced and shed each year, is found in most members of Amaryllideae and the outgroup. In contrast, Crinum species may either be evergreen or, as in Ammocharis and Cy- bistetes, they may have mature foliage leaves that die back during seasonal drought then regrow at the end of the dormant period. As Troll (1954) showed in Ammocharis coranica (Ker Gawl.) Herb., regrowth of the mature leaves is due to a well-de- veloped intercalary meristem, which results in the truncate leaf apices that characterize most species in these genera. This growth pattern is particularly remarkable in plants of Cybistetes longifolia from the arid Richtersveld, southwestern Africa, in which the leaf blades elongate and die back up to three times during a single year in response to al- ternating wet and dry sequences (Snijman & Wil- liamson, 1994). 4. (3) Leaf number: at least four = 0, two = 1. A reduced number of foliage leaves is found in all species of Hessea and Strumaria subg. Gem- maria. Most individuals in these taxa have two leaves per year, but, rarely, three leaves have been recorded (Snijman, 1994). Four or usually more fo- liage leaves per shoot occur in all other Amarylli- deae and in most genera of Haemantheae and the family. 5. *Leaf pigmentation: unmarked = 0, speckled with red abaxially = 1. Only in Crossyne are the leaves abaxially speck- led with red, an autapomorphy for the genus. 366 Annals of the Missouri Botanical Garden *Leaf surface: glabrous = 0, pubescent (at i 1 least in juveniles) = Leaves with soft, simple, uniseriate hairs (0.2— 6.0 mm long) are unique to Strumaria subg. Gem- maria. Pubescence diminishes in some species of this group at maturity (Strumaria sect. Bokkeveldia) but is always present in juveniles. This specializa- tion is treated as an autapomorphy. 7. (10) Leaf margin: unthickened thickened = 0, heavily The leaf margins of Brunsvigia and Crossyne are raised with thick-walled cells, often giving the leaves a slightly crisped, reddened outline. 8. (10) Leaf margin: untoothed = 0, with thin- walled branching teeth = 1, with thick-walled branching bristles = 2 Multiseriate, branching, cartilagenous cilia are present on leaf margins of Crinum (Arroyo & Cut- ler, 1984), Ammocharis, and Cybistetes, and these are thin-walled in comparison to the long, bristly cilia on the leaf margins of Crossyne. This multi- state character is treated as unordered. 9. (4) Pedicel length at anthesis: equaling or less than perigone length = 0, at least twice the perigone length = Two variables determine inflorescence form at anthesis: pedicel length and flower size. A clus- tered inflorescence in which pedicel length is less than the perigone length predominates in Amaryl- lidaceae, but widely spreading inflorescences, with pedicels at least twice as long as the perigone, are characteristic of Brunsvigia, Crossyne, Hessea, and Strumaria subg. Tedingea. The states slightly over- lap in Nerine and Carpolyza, but when partitioned according to the method of Almeida and Bisb (1984), the data for these taxa could be reliably scored (Snijman, 1992). 10. (10) Flower color during senescence: pigmen- tation accentuated = 0, pigmentation lost = I. Verdoorn (1973) first drew attention to color changes in the flowers of Amaryllideae when she noted that floral pigmentation in most southern Af- rican Crinum is accentuated with whereas in a few species it is lost. Enhanced col- species age, oring is common in the senescent flowers of all gen- era of Amaryllideae and Haemantheae, except in Hessea, where pigmentation is consistently lost in all species. Crossyne, whose two species have con- trasting states, was treated as indeterminate, and Crinum was scored as plesiomorphic until addi- tional data clarify the interpretation. 11. (2) Stamen position: spreading = 0, declinate Declinate stamens, absent from the outgroup, are consistently found in Amaryllis, Nerine, and Bruns- vigia. Both states occur in Crinum and Crossyne, which are treated as indeterminate. 12. (4) Staminal tube: absent = l. conspicuous = 2. 0, rudimentary = Stamens in Amaryllideae are either separate (Ammocharis, Boophone, Crinum, and Cybistetes) or proximally connate to varying degrees. The con- nation in Amaryllis, coded as rudimentary, is ex- tremely short (ca. 1 mm) and extends above the confluence of only the outer tepal whorl before be- coming free. In Nerine, Brunsvigia, Crossyne, and Hessea, the staminal tube extends shortly to well above the confluence of both inner and outer tepal whorls, and is thus coded as conspicuous. Repre- sentatives of Hessea subg. Hessea may have the lon- gest staminal tubes in the tribe, up to 4 mm long, in contrast to those of Hessea subg. Namaquanula, in which the tube is almost absent, putatively a secondary reduction (Snijman, 1994). The interpretation of the stamens in Strumaria subg. Strumaria is more complex. Although sepa- rate in S. bidentata Schinz, the stamens in the re- maining four taxa are connate. Serial sections in S. truncata, however, indicate that the outer stamens are proximally adnate to the style (Snijman, 1994). Thus the staminal connation in Strumaria subg. Strumaria is regarded as autapomorphic and treat- ed independently from the staminal tube in other genera of Amaryllideae. 13. *Filament trichomes: trichomes absent = 0, trichomes present = The filaments are smooth in all Amaryllideae and Haemantheae, except Hessea subg. Namaquanula, where they are proximally covered with short tri- chomes on the adaxial surface (Snijman, 199 14. *Filament morphology: both whorls uniform = 0. outer and inner whorls dimorphic =1 Unknown elsewhere in Amaryllidaceae are the specialized, dimorphic filaments of Hessea stenosi- phon (Snijman) D. € U. Müll.-Doblies. The short outer filaments occlude the perigone throat, where- as the inner filaments of this species are erect and club-shaped, with a structural composition highly representative of osmophores as described by Vogel (1990). The epidermal cells are densely cytoplas- mic and the inner parenchyma contains strikingly conspicuous masses of starch (Snijman, 1994), Fra- Volume 83, Number 3 1996 Snijman & Linder Phylogenetic Relationships in Amaryllideae grance emission from these structures is still to be tested. 15. (4) Anther attachment: + dorsifixed = 0, in a short connective sheath (subcentrifixed) = 1, in a long connective sheath (centrifixed) = 2. Previously described as “basifixed” (Baker, 888, 1896; Dyer, 1976; Phillips, 1951; Traub, 1963), the anthers of Hessea and Carpolyza were recently redefined in terms of the degree to which the filament tip is inserted into a tube formed by the anther connective. The relative lengths of the dorsal and ventral walls of the tubular connective enabled D. and U. Miiller-Doblies (1985) to rec- ognize three states of anther attachment: dorsifixed anthers (relative lengths less than 50%), subcen- trifixed anthers (relative lengths 50-75%), and cen- trifixed anthers (relative lengths more than 75%). Despite their apparent arbitrary subdivision, cur- rent data indicate that the states are only slightly overlapping. Anthers are dorsifixed in the outgroup and in most genera of Amaryllideae; subcentrifixed in Hessea subg. Namaquanula, Strumaria subg. Gemmaria, and Carpolyza; and centrifixed in Hes- sea subg. Hessea. The interpretation of the anther attachment in Hessea subg. Kamiesbergia is not yet resolved (Miiller-Doblies, 1992; Snijman, 1994). To reflect the hypothesized morphocline in anther at- tachment, the character states are treated as addi- tive. 16. (10) Pollen morphology : monosulcate = 0, bis- ulculate = 1 Relative to the large, boat-shaped-elliptic, mon- osulcate pollen grains in other genera of Amaryl- lidaceae (Meerow, 1995), pollen grains of Amaryl- lideae are consistently bisulculate and normally dispersed at the two-celled stage (Dutt, 1962; Erdt- man, 1966; Howell & Prakash, 1990; Schulze, 1984 17. (10) Pollen exine: reticulate = 0, spinulose = 1. Unlike reticulate exine elsewhere in Amarylli- daceae, surface ornamentation of pollen is uniform in Amaryllideae; gemmate with scattered large spi- nulae (Meerow, 1995) 18. (1) Style position: symmetrically placed = laterally displaced = The style is uniformly symmetrically placed in the outgroup, Boophone, Ammocharis, Hessea, Stru- maria, and Carpolyza. All remaining Amaryllideae have a laterally displaced style. Lateral displace- ment of the style is usually associated with decli- nate stamens but in certain species of Crinum, Crossyne, and Cybistetes the stamens remain evenly spread. 19. (10) Style form: slender = 0, proximally en- larged = 1. With the exception of Strumaria, the style in Amaryllideae and the outgroup is typically slender. Goldblatt (1976) first recognized the proximally en- larged style in Strumaria as a derived, unique char- acter. The stylar swellings are variable but take three basic forms: increasingly thickened upward then abruptly narrowed above (Strumaria subg. Tedingea); more or less equally thickened in the lower half (subg. Strumaria); or distinctly broadest at the base and gradually tapering upward (subg. Gemmaria) (Snijman, 1994) 20. (10) Nectar collection site: pooled around the style , in three discrete sites in the axils be- tween dei inner filaments and style = 1 Nectar collects in a central well around the base of the style in the outgroup and in most genera of Amaryllideae. This contrasts with Carpolyza and Strumaria, where nectar collects in three discrete sites, in the axils between the inner filaments and the style base. These separate nectar collection sites vary in depth and diameter. In most species (Carpolyza, Strumaria subg. Tedingea and subg. Gemmaria), the site is shallow, holding nectar in a single large droplet. Through lengthwise and radial extension of the confluence formed by the outer fil- ament whorl and style in Strumaria subg. Strumar- ia, the volume of the cavities between the inner filaments and style has been increased, and deep nectar wells have developed. Preliminary studies have shown that within Strumaria, nectary anatomy is quite diverse. Without exception the nectaries are septal, but considerable variation is found in the position of the nectaries relative to the locules, in their size, and their degree of convolution. Al- though the extent of septal nectary diversity in Amaryllideae does not approach that reported in Haemodoraceae (Simpson, 1993), a detailed study of floral anatomy may give further insight into the tribe’s functional floral morphology. 21. (10) Ovule: bitegmic = 0, unitegmic = 1. In contrast to the bitegmic ovules of most Amar- yllidaceae, the ovules in Amaryllideae are uniteg- mic. Representatives reported to have a solitary in- tegument are Amaryllis (Markótter, 1936; Oganezova, 1990; Schlimbach, 1924); Boophone disticha (Schlimbach, 1924); Brunsvigia minor Lindl. and several Crinum species (Hofmeister, 1861); Cybistetes (Markótter, 1936); and Nerine sar- 368 Annals of the Missouri Botanical Garden niensis (L.) Herb. (= N. curvifolia Herb.) (Goebel, 1932) Furthermore, several species of Crinum have been described as having ategmic ovules (Toilliez- Genoud, 1965; Tomita, 1931). However, the defin- itive embryological studies of Dutt (1957a, b, 1959, 1962) and Venkateswarlu and Lakshmi (1978), on a range of Crinum species, present strong evidence for the existence of an integument before its loss during the transformation of the ovule into the seed. The present study also confirms the presence of unitegmic ovules in Brunsvigia and Nerine (see Ap- pendix I) and reports their presence in Crossyne, Hessea, and Strumaria. This current data give fur- ther support to Huber's (1969) hypothesis that uni- tegmic ovules are synapomorphic for the tribe. 22. (1) Scape habit during seed dispersal: withering yack after seed release = 0, detaching at ground level after seed release from the cap- sules = 1, detaching at ground 5 before seed release from the capsule = : * Habit of fruiting cluster during seed REN: maining attached to scape = 0, detaching from scape — At maturity the scape in Amaryllideae varies in its duration and place of abscission. The scape per- sists above ground until long after seed release, then slowly withers away in the outgroup, Amaryl- lis, Ammocharis, Boophone, Crinum, and Strumaria subg. Strumaria. Elsewhere in the tribe, the scape detaches from the bulb at ground level soon after seed release (Carpolyza, Nerine, and Strumaria subg. Tedingea) or before the release of seed from the capsules (Cybistetes, Brunsvigia, Crossyne, Hes- sea, and Strumaria subg. Gemmaria). Thus through early abscission of the scape in the latter five taxa, the entire infructescence becomes the main unit of dispersal. Dispersal is similar in Boophone, but the structural unit is uniquely derived by the abscis- sion of the entire fruiting cluster from the top of the scape. Multistate character 22 is treated as or- dered, and until the abscission tissues of the scape are studied anatomically, the states are considered to be homologous. = (), indehiscent = I. 24. (10) Fruit: dehiscent In the outgroup indehiscent fruits occur in four genera, whereas a dehiscent capsule is found only in Cyrtanthus L.f. Dehiscent capsules also occur uniformly in the closely similar Hippeastreae and Lycorideae (Dahlgren et al., 1985) and in Alli- aceae, which molecular data rd as basal to Amaryllidaceae (Fay et al., 1994). Although not common in the outgroup, the existence of dehiscent capsules in one representative, as well as at a level immediately basal to the Amaryllidaceae, formed the basis for selecting capsule dehiscence as the outgroup Character state. This coding parallels Dahlgren and Rasmussen’s (1983) hypothesis that baccate fruits are secondarily derived from cap- sules in Amaryllidaceae. 25. (10) Developing fruit: never rostellate = 0, ros- llate = 1 te Ammocharis, Boophone, Crinum, and Cybistetes have fruits which, while developing, are beaked with the persistent base of the perigone tube. The beak persists on the mature fruits only in Crinum, and its final length has proved to be diagnostically valuable in many species (Nordal & Wahlstrgm, 1980; Verdoorn, 1973). None of the other members of Amaryllideae or the outgroup has beaked fruits. 6. (10) Mature fruit: + regular = 0, irregular = 1. 27. Mature fruit: not large = 0, conspicu- ously enlarged = The form of the fruit is more or less symmetrical Amaryllideae except in Ammocharis, Crinum, and Cybistetes, where it is molded by the enclosed developing seeds. As the shape of the seeds is often extremely variable (Manasse, 1990; Snijman & Williamson, 1994; Toilliez-Genoud, 1965), the fruit also assumes an irregular form. Conspicuously large, obconical to rarely spindle-shaped capsules (reaching up to 60 mm long and 25 mm across in many species), with visible transverse veining, are autapomorphic for Brunsvigia. 28. (10) Testa: without stomata = 0, stomatose = 1. Although not common in the outgroup Haeman- theae, dry seeds, crusted with phytomelan, occur in one representative, Cyrtanthus; wi other Amaryllidaceae; and consistently in Alliaceae, the family basal to Amaryllidaceae (Fay et al., 1994). In contrast, the mature seeds of Amaryllideae are water-rich and lack phytomelan in the coat. Schlim- bach (1924) was the first to report stomata on the seed coat of Nerine sarniensis (= N. curvifolia), and Boyd (1932) later recorded them in Carpolyza spir- alis (UHérit.) Salisb. A survey of the testa in this study has shown that stomata are also present on widely in the seeds of additional representatives of Nerine, in Brunsvigia, Crossyne, Hessea, and Strumaria (Ap- The stomata are anomocytic (Fig. 5B—E), as is typical for Amaryllida Arroyo & Cutler, 1984; Dahlgren & Clifford, 1982), and the sur- rounding cells are covered by a sculptured cuticle Fig. 5D, E), in which the central striations become increasingly thick and sinuous with age. pendix I). — Volume 83, Number 3 1996 Snijman & Linder Phylogenetic Relationships in Amaryllideae 29. (10) Integument in developing seed: disinte- grating = 0, enlarging = 1. 30. (10) /ntegu- ment in developing and mature seed: without chlorophyll = 0, chlorophyllous = 1 In Brunsvigia, Crossyne, Nerine, Hessea, Stru- maria, and Carpolyza the integument expands into a water-rich, chlorophyllous tissue with vascular strands. As reported in N. sarniensis (= N. curvi- folia) this is achieved by increasing the number and the size of the cells after fertilization (Schlim- bach, 1924), and when finally ripe, the integument contains rich amounts of starch and chlorophyll (Goebel, 1932) (Fig. 4B—D). In the other represen- tatives of Amaryllideae the integument remains un- differentiated and finally disintegrates (Dutt, 1970; Venkateswarlu & Lakshmi, 1978), whereas in the outgroup, the integuments of seeds shed in a dry, dormant state are crushed at maturity. 31. (10) Endosperm in mature seed: undifferentiat- ed externally = 0, formed into cork externally = 1. 32. (10) Mature seed: without chlorophyll in endosperm = 0, endosperm with chlorophyll A novel feature in Ammocharis, Boophone, Cri- num, and Cybistetes is the development of a few layers of cork tissue from the peripheral cells of the endosperm (Fig. 4E, F). Although well known in Crinum (Merry, 1937; Schlimbach, 1924; Venka- teswarlu & Lakshmi, 1978) and Cybistetes (= Am- mocharis falcata sensu Markótter, 1936), a cork- covered endosperm is reported here in two additional genera: Boophone and Ammocharis (Fig. 5F, H). Another novel feature in these genera is that chloroplasts occur in the peripheral cell layers beneath the phellogen, even when the seed is cov- ered by the pericarp (Dutt, 1957a, 1962; Schlim- bach, 1924). Amaryllis is the only taxon of Amar- yllideae in which the extensively developed endosperm does not have corky or chlorophyllous outer layers (Schlimbach, 1924) (Fig. 4A). In all other Amaryllideae the endosperm remains undif- ferentiated and comparatively poorly developed. 33. (10) Embryo: without green pigment = 0, green =] Dahlgren and Clifford (1982) suggested that the embryo of Crinum could be chlorophyllous. Since then Howell and Prakash (1990) recorded this fea- ture in Crinum flaccidum Herbert, while our study confirms the presence of green-pigmented embryos in every genus in Amaryllideae. The chlorophyll content, however, is yet to be investigated. 34. (10) Basic chromosome number: 11 = 0, 10 = 1. pygmaea Snijman, which has x = 1 Chromosome numbers in Amaryllideae are well known (Goldblatt, 1972; Gouws, 1949; Nordal & Wahlstrgm, 1980) and a karyotype of x = 11 is considered basic in Amaryllideae (Goldblatt, 1976) and the family (Inariyama, 1937; Meerow, 1984; Sató, 1938). Species of Carpolyza and Strumaria have a reduced base number of x = 10 (Goldblatt, 1976; Snijman, 1994), the only exception being S. Ithough most taxa in Haemantheae have x = 9 and 8 (In- ariyama, 1937; Ising, 1970; Vosa & Marchi, 1980) some species of Clivia have the base number of x = 11 (Gouws, 1949). OUTGROUP SELECTION Traub's (1962, 1963) tribal classification of Amaryllidaceae provides a valuable framework for selecting an outgroup that has synapomorphies shared with the ingroup (see Nixon & Carpenter, 1993). Traub’s system divides the currently recog- nized tribes (Dahlgren et al., 1985) into two infra- family groups, the Amarylloidinae and the Pancra- tioidinae. Within the Amarylloidinae, the apparent close similarity between representatives of Amar- yllideae and Hippeastreae (Brummit, 1987) is based essentially on highly homoplasious floral characters, but the pollen, fruit, and seed charac- ters are discontinuous (see Meerow, 1995). Other differences between the two tribes are root and scape structure, and spathe valve arrangement, and in these characters Haemantheae is intermediate (Arroyo & Cutler, 1984). The close relationship postulated to exist between Haemantheae and Amaryllideae, for which equitant spathe valves (Ar- royo, 1984) is a possible synapomorphy, thus ac- counts for our choice of Haemantheae as outgroup to root the cladogram. All the characters defined for the analysis of the Amaryllideae apply also to the outgroup Haemantheae. DELIMITATION OF GENERA IN AMARYLLIDEAE All the currently recognized genera of Amaryl- lideae were examined to ensure that their delimi- tation could reasonably be considered monophylet- ic. This is accomplished by identifying hypothetical synapomorphies of the ingroups relative to other taxa (Nixon & Carpenter, 1993). Amaryllis, Carpo- lyza, and Cybistetes were accepted as monophyletic by virtue of their being monotypic. Boophone sensu lato (five species) has commonly been distinguished from other genera of Amarylli- deae by large bulbs, inflorescences of more than 00 flowers, a perigone with tube shorter than the tepals, long pedicels in fruit, and triquetrous dry 370 Annals of the Missouri Botanical Garden fruits with solitary or few seeds per locule (Dyer, 1976; Nordal, 1982). Milne-Redhead and Schweickerdt (1939) first questioned the inclusion of species with both dehiscent and indehiscent fruits in Boophone and recommended that the ge- neric limits be reexamined. In agreement with Miiller-Doblies (1994), three groupings have been proposed for the five species. oophone disticha and Boophone haeman- thoides F. M. fruits, which are also found in Crinum, Ammochar- Leight. share indehiscent, beakec is, and Cybistetes. However, the species are unique- ly characterized by the abscission of the fruiting head from the top of the scape. Consequently, this species pair, which includes the type species of Boophone, is considered to be monophyletic. (2 Boophone guttata (L.) Herb. and Boophone flava W. F. Barker ex Snijman have leaves abaxially speck- — led with red, and fringed with long, rigid, branching hairs. These two unique features are taken as evi- dence of the monophyly of the species pair, and they are assigned to the genus Crossyne, which was recently resurrected by D. and U. Miiller-Doblies (1994). (3) Ponens pulchra W. F. Barker shares none of the putative synapomorphies recognized for the newly defined Boophone and Crossyne. Possible derived characters in Boophone pulchra, relative to the tribe as a whole, include the presence of ma- cropapillae on the adaxial leaf epidermis and en- larged, usually triquetrous capsules, features that occur in species of Brunsvigia. For the phyloge- netic analysis of Amaryllideae, Boophone pulchra is, therefore, treated as belonging to Brunsvigia in accordance with D. and U. Miiller-Doblies (1994). The only feature in which Brunsvigia pulchra (W. U. Miill.-Doblies differs from other Brunsvigia species is the elongation of the pedicels after anthesis rather than prior to anthesis. Al- though typical of Boophone sensu stricto, this char- acter state occurs in several other members of the tribe (Amaryllis and Cybistetes) and cannot be rec- ognized as autapomorphic for Boophone sensu stric- to. Crinum is currently regarded as the largest genus in the tribe with ca. 65 species (Fangan & Nordal, 1993). Species concepts of these ornamental plants have traditionally been too narrow (Fangan & Nor- dal, 1993; Hannibal, 1985), and thus far revisional work for the African floras has resulted in a re- duced number of taxa being upheld (Nordal, 1977). he genus is cytologically variable (Nordal & ul 1980), but vegetative and floral diver- sity is comparable to that of other genera of Amar- yllideae (Brunsvigia, Hessea, and Strumaria). Flow- ers vary from hypocrateriform to funnel-shaped or bell-shaped, with spreading or declinate stamens; the leaves, which sometimes form a false-stem, are channeled or ribbed, thin-textured or fleshy, with intact or truncate tips (Nordal, 1977). The truncate leaf apices, which also occur in Ammocharis and Cybistetes, result from a well-developed and puta- 1954), which enables the leaves to die back and regrow tively derived intercalary meristem (Troll, over several seasons. Evident synapomorphies do not exist to separate Crinum from Ammocharis; therefore, Crinum is probably paraphyletic with Ammocharis and Cybistetes embedded in it. How- ever, for the analysis Ammocharis and Cybistetes are provisionally retained, as suggested by Snijman and Williamson (1994), and following the recommen- dations of Nordal (1977) and Verdoorn (1973) ex- isting subgeneric divisions of Crinum are avoided because they are so weakly supported. Nerine, with ca. 23 species, is possibly the sec- ond largest genus of Amaryllideae. Genera of Amaryllideae are distinguished by subtle floral characters, and the extremely similar floral mor- phology shared by species of Nerine provides good evidence for the proposed monophyly of the genus. The characteristic attenuate, more or less crisped, tepal outline of Nerine species is a likely synapo- morphy for the genus. Brunsvigia (19 species) is uniquely macropapil- late on the adaxial leaf epidermis in many species Arroyo & Cutler, 1984), and large, inflated cap- sules (sometimes up to 70 mm long) with visible _ transversal veins occur in all the species. Both fea- tures provide good evidence for the inclusion of Boophone pulchra in Brunsvigia. Although not in- flated, similarly shaped, somewhat triquetrous or elongated dry fruits also exist in Boophone sensu stricto and Cybistetes. The different mode of dehis- cence in the latter genera (indehiscent as opposed to dehiscent: character 24) suggests, however, that the fruits have been separately derived. The en- larged capsule is regarded here as evidence for the monophyly of Brunsvigia. Hessea (13 species) and Strumaria (23 species), commonly distinguished by their consistently acti- nomorphic flowers, are the only genera of Amaryl- lideae currently classified to reflect explicitly es- tablished phylogenetic relationships. Hessea, with three subgenera, is defined by one unique syna- pomorphy: the total loss of floral pigmentation at senescence (Snijman, 1994). The synapomorphy for Strumaria, also with three subgenera, is the uniquely swollen or winged style base (Snijman, 1994). In order to provide maximum structure for the phylogeny of Amaryllideae, the reliably mono- phyletic subgenera of Hessea and Strumaria were Volume 83, Number 3 1996 Snijman & Linder 371 Phylogenetic Relationships in Amaryllideae — Haemantheae Ammocharis i . 1 1617182132 Cybistetes HHH — Amaryllis — Nerine 1112 11 2 792 [- Brunsvigia l 1222 28 29 30 Crossyne | A eI BuEGSE 15 21111 -F Hessea subgen Hessea 2 1213 HF- Hessea subgen Namaquanula 11 14 4 9 1011151822 R . a ee ie: HHHHHH EE Hessea subgen Kamiesbergia 1110102 4 61922 HHH- Strumaria subgen Gemmaria Ana] Carpolyza UUUNLUSIS 00011 9 22 1519 MHF Strumaria subgen Strumaria 0 0 01 e l. Strict consensus tree of Amaryllideae. Solid uu eim a a and open bars indicate reversals. also included in the analyses. All taxa included in the phylogenetic analyses are listed in Table 2. RESULTS The cladistic analyses with equal character weighting (excluding autapomorphies) gave 16 min- imal length cladograms of 54 steps, a consistency index (CI) of 0.68, and a retention index (RI) of 0.80. The strict consensus tree (Fig. 1) generated by PAUP and Hennig86 shows that all 16 clado- grams agree in the following respects. (1) Boophone, Crinum, Ammocharis, and Cybistetes appear as a monophyletic group. (2) Amaryllis resolves as basal to Nerine, Brunsvigia, Crossyne, Hessea, Carpolyza, and Strumaria. (3) Brunsvigia and Crossyne resolve ssea, Carpolyza, and Strumaria form a monophyletic group. in a monophyletic clade. (4) He Successive weighting retrieved 2 of the 16 most parsimonious trees. These differ only in the degree of resolution shown in the Hessea clade. The tree Strumaria subgen Tedingea bars indicate non-homoplasious synapomorphies; gray bars that we have chosen for interpreting character evo- lution (Fig. 2) differs from the strict consensus tree (Fig. 1) as follows. (1) The Crinum—Ammocharis— Cybistetes clade is more fully resolved. (2) Bruns- vigia, Crossyne, Hessea, Carpolyza, and Strumaria form a clade that resolved in 14 of the 16 minimal length trees. (3) A clade of Hessea subgenera, which was retrieved in 10 of the 16 minimal length trees, is present. (4) The subgenera of Strumaria are pres- ent in a clade. Support for this component is more complex. Strumaria subg. Gemmaria is basal to arpolyza and Strumaria subg. Tedingea and sub- genus Strumaria in 10 of the minimal length trees, but this component is supported only by the hom- oplasious character 22(1) (CI 0.28, RI 0.54). In contrast, Carpolyza resolves as basal to the clade of Strumaria subgenera in 6 of the trees retrieved. However, character 19 (style proximally enlarged, CI 1.00 and RI 1.00), which supports the Strumaria clade, is unique in Amaryllidaceae (Snijman, Annals of the Missouri Botanical Garden Haemantheae r- Amaryllis 1112 Nerine 111518 re 2. The preferred most parsimonious cladogram of Amaryllideae. Solid bars indicate non-homop apomorphies; gray bars indicate parallelisms and open bars indicate rev 1994), and we therefore accept the topology of the Carpolyza—Strumaria component in our preferred tree with confidence (Fig. 2). Lastly, the differences in resolution obtained in the Crinum—Ammocharis— Cybistetes clade in the strict consensus tree (Fig. 1) and the preferred most parsimonious tree (Fig. 2) do not alter our discussion of seed character and dispersal system evolution. hen Amaryllis was placed at the base of the Boophone—Crinum—Ammocharis—Cybistetes clade, as suggested by the classifications of Traub (1957, 1962, 1963), the topology was 2 steps longer than our most parsimonious tree. The plac ement of Am- aryllis in a position basal to the two major clades gave a topology | step longer than the minimal length cladogram. When parsimony was relaxed by 2 steps, 1404 trees were generated and resolution was lost in all branches except the Crinum—Am- mocharis—Cybistetes clade and the Brunsvigia— Crossyne clade. lasious syn- ersals. DISCUSSION RELATIONSHIPS WITHIN AMARYLLIDEAE All previous classifications of Amaryllideae (Ta- ble 1) placed the Boophone—Cybistetes clade, Am- aryllis, Nerine, and Brunsvigia in Amaryllidinae. Our cladistic analysis clearly shows that subtribe Amaryllidinae 1957, 1963) is paraphyletic with Crossyne and Strumari- inae (the Hessea—Carpolyza—Strumaria clade) em- bedded in it. Strumariinae, the only other subtribe 1985), is (=Crininae sensu Traub, previously recognized (Müller-Doblies, monophyletic. In order to recognize the monophyletic groups generated by the cladistic analysis we propose a new classification. It is widely accepted that no- menclature is most effective if taxon names are sta- ble. Thus prior to translating cladistic results into new classifications, Linder (1991) recommends an assessment of the nodes that are unlikely to change Volume 83, Number 3 1996 Snijman & Linder Phylogenetic Relationships in Amaryllideae with the addition of new data. Since Amaryllis is the oldest valid name available for typification at subtribal level, the stability of its position is of par- ticular importance. Our results best support the placement of Amaryllis basal to Nerine, Brunsvigia, Crossyne, Hessea, Carpolyza, and Strumaria. We have therefore chosen to join the Strumariinae cla- de with Crossyne, Brunsvigia, Nerine, and Amaryllis to form a more inclusive clade, which is formally named subtribe Amaryllidinae. The Boophone, Cri- num, Ammocharis, and Cybistetes clade is accord- ingly recognized as a second subtribe: the emended Crininae. The option of recognizing a third sub- tribe, containing only Amaryllis, was also consid- ered but we regard the formal classification given below to be more conservative in reflecting the presently understood relationship of Amaryllis to Nerine and other representatives of its clade. No- menclatural stability within Amaryllideae is thus maintained until strong support to the contrary be- comes available. REVISED CLASSIFICATION OF THE TRIBE AMARYLLIDEAE Tribe Amaryllideae Bulb scales when torn produce extensible cot- tony threads; flowers actinomorphic or zygomor- phic; pollen bisulculate; ovule unitegmic; seed water-rich with green embryo. Includes about 55 species. Mostly African but with several Crinum species in the tropics of Asia, Australia, and America, Madagascar, and the Mascarene and Pacific Islands. Subtribe Amaryllidinae Pax in Engler & Prantl, Nat. Pflanzenfam. 2, 5: 105. 1887 Scape usually abscissing at ground level; sta- mens declinate (except in Hessea, Strumaria, and Carpolyza), connate into a tube proximally (ex- cept in Strumaria and Carpolyza); seed (except in Amaryllis) with a well-developed chlorophyl- lous integument and stomatose testa. Includes 82 known species. Amaryllis Linnaeus, Sp. Pl.: 292. 1753. Flowers zygomorphic; stamens connate proxi- mally into a rudimentary filament tube; seeds colorless or pink. Monotypic. Nerine Herbert in Bot. Mag. 47: t. 2124. 1820. Flowers zygomorphic; tepals attenuate, margins + crisped. 23 species (Traub, 1967). Brunsvigia Heister, Beschr. neu. Geschl.: 3. 1755. Bulb scales usually thick and brittle; leaf mar- gins raised; flowers zygomorphic; capsule en- larged, turbinate or fusiform with visible trans- verse veins; scape detaching at ground level; infructescence dispersed as a single unit. 19 spe- cies (Dyer, 1950, 1951). Crossyne Salisbury, Gen. Pl.: 116. 1866. Bulb scales thick and brittle; leaves speckled with red abaxially, margins raised and fringed with long, stiff trichomes; inflorescence 100- or more-flowered; flowers zygomorphic, sometimes only weakly so by deflexed style; scape detaching at ground level; infructescence dispersed as a single unit. 2 species (Miiller-Doblies, 1994). Hessea Herbert, Amaryllidaceae: 289. 1837. Leaves usually 2; flowers actinomorphic, aging to brown; filaments inserted into a sheath formed by anther connective (centrifixed to subcentrifix- ed), except in H. bruce-bayeri (D. € U. Müll.- Doblies) Snijman and H. stenosiphon (Snijman) D. & U. Miill.-Doblies; infructescence disperse as a single unit. 13 species (Snijman, 1994). Strumaria Jacquin, Collectanea 5: 49. “1796” [1797]. Leaves up to 6, then usually arranged in a fan (subg. Strumaria), or mostly 2 and pubescent, at least in juveniles (subg. Gemmaria); flowers ac- tinomorphic; filaments of outer or both whorls proximally adnate to style, usually inserted into a sheath formed by anther connective (subcen- trifixed); style proximally swollen or winged; nec- tar collects in axils between inner filaments and style; basic chromosome number x = 10 (except in S. maea Snijman, where x = 11). 23 spe- cies (Snijman, 1994). Carpolyza Salisbury, Parad. Lond. 1: t. 63. 1807. Flowers actinomorphic; inner filaments adnate to style proximally, inserted into a sheath formed by anther connective (subcentrifixed); nectar col- lects in axils between inner filaments and style; basic chromosome number x = 10. Monotypic (Snijman, 1994). Subtribe Crininae Pax in Engler & Prantl, Nat. Pflanzenfam. 2, 5: 108. 1887. Flowers actinomorphic or zygomorphic; fruit in- dehiscent, rostellate at least during development; seeds lacking an integument, endosperm-rich, cork-covered, with layers of chlorophyll-containing cells below the phellogen. Includes ca. 73 species. several 374 Annals of the Missouri Botanical Garden Table 3. Distributions and favored habitats of genera of Amaryllideae. Number of species is given in parentheses. Southern Africa is defined as the area south of a line between the Kunene and Zambezi rivers. Distribution Habitat Subtribe CRININAE Boophone (2) Crinum (ca. 65) Ammocharis (5) Cybistetes (1 — Tropical East Africa to southern Africa Tropics of Africa, Asia, Australia & Ameri- ca, southern Africa, ene & Pacific Islands Tropical East Africa to southern Africa Winter-rainfall region of southern Africa Subtribe AMARYLLIDINAE Brunsvigia (19) miee (2) Hes Du — 13) eitis (2: (1 ) Carpolyza Winter- rainfall een of South Africa Southern Afr Southern Afics Winter-rainfall region of South Africa Winter-rainfall region of southern Africa Winter-rainfall region of southern Africa Winter-rainfall region of South Africa Madagascar, Mascar- Open plains or slopes Moist sites, forests, river edges, seasonal pools or salt pans In rocks or seasonally wet places Open sites Mesic places Mesic places Open iis or rocky sites Open pla In rocks and seasonally moist sites In rocks and seasonally moist sites In rocks and seasonally moist places Crinum Linnaeus, Sp. Pl.: 291. 1753. Leaves perennial, often formed into a false-stem, fringed with cartilaginous teeth, apex often trun- cate; flowers zygomorphic, sometimes only weak- ly so; fruit irregular, wall membranous to spongy. The genus has not yet been revised over its entire distribution range, but good regional T s are available for Africa (Nordal, 1977, Nordal & Wahlstróm, 1980; Verdoorn, tn In- cludes ca. 65 species (Fangan & Nordal, 1993) with ca. 40 in Africa. The genus is probably paraphyletic. Ammocharis Herbert, Appendix: 17. 1821. Leaves perennial; somewhat falcate, fringed with cartilaginous teeth, apex often truncate; flowers actinomorphic; fruit with membranous walls, ir- regular. 5 species (Milne-Redhead & Schweick- erdt, 1939). Cybistetes Milne- wea - D 'hweickerdt in J. Linn. Soc., Bot. 52: 19 Leaves perennial, a with cartilaginous teeth, apex of mature leaves truncate; flowers weakly zygomorphic; scape detaching at ground level; infructescence dispersed as a single unit. Monotypic. The genus is probably nested in Am- mocharis (Snijman & Williamson, 1994). Boophone Herbert, Appendix: 18. 1821. Leaves erect, distichous; flowers actinomorphic; fruiting head detaching from top of scape, dis- persing as a single unit. 2 species. PHYTOGEOGRAPHY AND PHENOLOGY A comparison of the present distribution patterns of subtribes Amaryllidinae and Crininae (Table : shows that Crininae is widespread in the tropical and temperate regions of sub-Saharan Africa (Fig. 3). One genus, Crinum, is pantropical, with ca. 40 of its ca. 65 species in Africa (Nordal, 1982). Ap- proximately 20 Crinum species occur in southern Africa, but only one is known from the winter-rain- fall region (Verdoorn, 1973). Only three other rep- resentatives of Crininae (Boophone disticha, Boo- phone haemanthoides, and Cybistetes) are known in the temperate winter-rainfall region of southern Af- rica, although one of these (Boophone disticha) also ranges widely into central Africa. In contrast, Amaryllidinae is confined to the temperate regions of southern Africa (Fig. 3). Most Nerine species are known from the summer-rainfall region. Brunsvigia ranges widely over both summer- and winter-rain- fall regions, whereas all the remaining representa- tives (Amaryllis, Crossyne, Hessea, Carpolyza, and all but two Strumaria species) are restricted to the winter-rainfall region of southern Africa. Most rep- resentatives of Crininae flower and fruit in summer during the leafing period. This differs in Amaryl- lidinae, where some species are summer-flowering but most flower and fruit in autumn, after the veg- etative phase of the previous winter, which offers a short growing period. SEED CHARACTERS IN AMARYLLIDEAE Throughout the tribe, the seeds are water-rich, nondormant, and have chlorophyllous embryos. As Volume 83, Number 3 1996 Snijman & Linder 375 Phylogenetic Relationships in Amaryllideae 20 10 0 10 20 30 fa 9 À b 8 EN Hr Noa | = | SD , eu | : ; : m |: TM | TT i - | tl | II | | AQ LITT Crininae Amaryllidinae = IU UH m" | 3n e NBI 1991 | | Winter-rainfall region | | Figure 3. Distributions of subtribes Amaryllidinae and Crininae in Africa. commented on by several authors (see Rendle, 1901), the seeds are large (at most 20-30 mm across in Crinum) but not necessarily heavy (Boyd, 1932; Dutt, 1962), and they germinate easily with- out additional water (Beaurieux, 1914; Howell & Prakash, 1990; Isaac & McGillivray, 1965). In comparison to other Amaryllidaceae, the embryos of Amaryllideae seeds have a cotyledon with a par- ticularly well-developed vascular system that con- tributes to seedling vigor, as reported by Boyd (1932). This can be seen in the rapid elongation of the hypogeal cotyledon, which assists the primary root to bore deeply into the soil and allows bulb formation within the first two months after germi- nation (Clark & Parsons, 1994, pers. obs.). Embryological studies (Dutt, 1962; Schlimbach, 1924; Venkateswarlu & Lakshmi, 1978) have shown that the mass attained by Amaryllideae seeds is due either to the development of the ovu- le’s solitary integument or to the endosperm. Con- sisting of up to 25 cell layers, the single integument in Amaryllidinae (except Amaryllis) is well devel- oped relative to the endosperm (Fig. 4B). As first reported by Schlimbach (1924), stomata are present in the testa (Fig. 4D), and the integument has con- siderable amounts of chlorophyll and starch (Fig. 376 Annals of the Missouri Botanical Garden Figure 4. the cells of the endosperm containing plastids and the partially et integument forming the testa.—B. vigia josephinae; part of the showing chloroplasts and a stoma in cross seclic coranica; endosperm with ype n and cork in qum Scale bars: A, C, E, F = 100 pm; B = 250 Appendix I gives details of vouc a specimens. 4B, C). The stomata are anomocytic (Fig. 5B-E and the suprastomatal cavity is overarched by lobes of the adjacent epidermal cells. The epidermis is covered by a thick, sculptured cuticle (Fig. 5D, E) in which the striations become increasingly sinuous with age. Growth of the embryo is delayed during the development of the accessory seed tissues (Goe- bel, 1932; Schlimbach, 1924), and the embryo re- mains small relative to that of Crininae when the — Seed anatomy in Amaryllideae. A-D. Species in subtribe embryo, endosperm, and well-de endosperm and integument densely packed a“ stare h grains.—D. species in subtribe Crininae.—E. Cy 81); endosperm showing mature cork and phellogen surrounding the inner ee of starch-filled ce yllidinae.—A. Amaryllis belladonna; note runs- nsvigia radula; part of the Str rumaria hapini ae layers of integument bistetes dilo ipie (Duncan ells. mmocharis veloped integument.—C. er layers; note chloroplasts and starch grains in the inner cells. nw = 25 pm . EM = embryo, EN = endosperm, IN = integument. seed is shed. Amaryllis, with its large. irregular, white to pink seeds, is exceptional in the tribe in that it has chlorophyll only in the embryo. The seed of Amaryllis possesses mainly endosperm (Hof- meister, 1861), without noticeably large amounts of starch (Fig. 4A), and the integument provides only the testa ferui 1924), which lacks stomata Fig. 5 In contrast, the solitary integument and nucellus —. Volume 83, Number 3 Snijman & Linder 377 Phylogenetic Relationships in Amaryllideae Figure 5. Seed surfaces in Amaryllideae. A—E. Species in subtribe rio note the presence of stomata except in Amaryllis.—A. Amaryllis belladonna.—B. Brunsvigi s.—C. Ner oli run ta.—E. Hessea breviflora, showing Sin of the thick sinuous cuticle. F-H. Species in e Canina note t surface at different stages of maturity.—F. Boophone disticha.—G. Crinum variabile. —H. Ammocharis coranica. Sale bars: 50 jum. See Appendix I for details of voucher specimens. Annals of the Missouri Botanical Garden of Crininae are transitory and absent from the fully developed seed (Dutt, 1962, 1970). The mature seed consists of massive amounts of starchssantabi ing endosperm. The endosperm is differentiated into cork in the outermost layers (Fig. 5F-H) and has several layers of chlorophyll-containing cells immediately below the phellogen (Fig. 4E, F). The seeds of Crininae are larger (10-30 mm diam.), more irregular and angular than the ovoidal, sto- matose seeds (4—10 mm diam.) of the Amaryllidi- nae. The embryo is well developed when the seed is shed, and the orientation of the embryo with re- spect to the seed as a whole is arbitrary (Toilliez- Genoud, 1965). Cotyledon extension is positively geotropic even within the seed (Howell & Prakash, The chlorophyllous tissue present in the integ- ument of the Nerine—Brunsvigia—Crossyne—Hessea— Carpolyza-Strumaria clade anc in the outer layers of the endosperm in Crininae implies some func- tional activity. Tests with infrared gas analysis showed that seeds of representatives from both sub- tribes (Brunsvigia orientalis (L.) Eckl. & Zeyh. and Cybistetes longifolia (L.) Milne-Redh. & Schweick.) are photosynthetically active and that those with a well-developed integument are capable of photo- synthesis at lower light levels than those of Crini- nae (Wand, pers. comm.). Furthermore, a test for durability in water indicated that seeds of C. defix- um Ker Gawl. and C. flaccidum Herbert had un- donde germination when removed from fresh wa- er four months (Clark & Parsons, 1994; Dutt, o although fresh water increases susceptibility to pathogen invasion (Manasse, 1990), and when wounded, the damaged outer portions of Crinum seeds rapidly produce fresh cork tissue (Dutt, 1962; erry, 1937). Tests on the buoyaney and viability of Crinum seeds in salt water indicate that they can t and remain viable for more than two years (ostia e phylogeny et by the morphological data ak little about the character assembly of the three seed types that exist in Amaryllideae. All the seed characters for Crininae (characters 31 and 32, Fig. 2) appear at the same branch point of the phylogeny; similarly, the seed characters for Nerine, Brunsvigia, Crossyne, Hessea, Carpolyza, and Stru- maria resolve at a single node (characters 28, 29, and 30, Fig. 2). In general, this pattern suggests that the selective association between these char- acters is strong (Armbruster, 1992; Frumhoff & Reeve, 1994), but, in the absence of additional data, our interpretation remains tentative. Relative to the bitegmic ovule and dry seeds of the outgroup, it is evident that the unitegmic ovule and fleshy seeds of Amaryllideae arose at the base of the lin- eage, through the loss of an integument in the ovule and the loss of the metabolic activity that leads to maturation drying in the seed. The solitary integ- ument persisted and thickened in the seeds of the Amaryllidinae lineage, and chlorophyll and stomata developed. In Crininae, the integument became short-lived and disappeared from the seed, but abundant endosperm and a corky covering were de- veloped. Thus, apart from a major divergence in seed morphology subsequent to the origin of the fleshy, nondormant seed in Amaryllideae, no pre- vailing sequences in the origin of the different seed characters are evident. EVOLUTION OF DISPERSAL SYSTEMS IN AMARYLLIDEAE Wind dispersal (anemochory) occurs in six gen- era of Amaryllideae (Fig. 6), and in all but one of these clades the mode of wind dispersal is ane- mogeochorous (sensu Van der Pijl, 1982). The func- tionally interrelated characters shared by the five anemogeochorous taxa are low stature, lightness of the infructescence; detachment of the infructesc- ence as a single unit before seed release (characters 22(2) and 23, Fig. 2); and long, stiff pedicels (char- acter 9, Fig. å tions. Despite the similarities in the morphology 2) that radiate outward in all direc- and manner in which the infructescences of each genus are blown over the ground, subtle differences are evident in the dispersal units” structure and tumbling, skidding, or lofting ability. Particularly good tumblers have been observed among representatives of the Brunsvigia—Crossyne clade. In many Brunsvigia species this ability ap- pears to be associated with broad, kitelike a surfaces (Fig. 7 ). a characteristic that i ognized to be aptimal for ease of T (Maddox & Carlquist, 1985; Van der Pijl, 1 is spe- cialization is further enhanced by i nae dehis- cence of the capsules, which prolongs the period over which seed is scattered. In Crossyne the in- fructescences are uniformly spherical due to the exceptionally large number of pedicels (100 or more): ane these equal the scape in length (Fig. al en released, contact with the ground is minimal and little force is required to roll the ball- like infructescence Relative to other representatives of Amarylli- deae, Hessea, Carpolyza, and Strumaria are char- acteristically diminutive in size. The narrowly dis- tributed, wind-dispersed species of Hessea and Strumaria subg. Gemmaria have infructescences in which the pedicels often remain shorter than the scape (ca. half the length of the scape) and are few Snijman & Linder Volume 83, Number 3 Phylogenetic Relationships in Amaryllideae Haemantheae peau Suningy Jo uorsstosqe s[oorpod poje3uo[o s[oorpod Kueui nay juoosrgopur pas Áy109 SIABI] SNOYIUBIDNSA Y Ts E] voupy UJ9ynos adeos Jo uorsstosqe jusumSojurt uoo18 Ei s[ootpod poye3uo[o a anemogeochory (tumbling) ozis juv[d poonpoi TIT atelechory/rain wash dispersal |] autochory anemoballist dispersal mode equivocal n LLL . E. Ammocharis S £5 5 72 > $ Cybistetes o da > o E gS 8 . e — r3 Amaryllis Ed E ; 9 & & Nerine 3 y ^ oe O n Brunsvigia Crossyne Hessea subgen Hessea Hessea subgen Namaquanula Hessea subgen Kamiesbergia Carpolyza Strumaria subgen Gemmaria Strumaria subgen Tedingea Strumaria subgen Strumaria (ALLEL: i ES Sox E Z & S o s Q 6 € E Dispersal systems, plant phenology, ee and functionally related seed and fruit characters Figure 6. in (estatico mapped onto the cladogram given in Figure in number. Although this configuration has some wind dispersal ability, it appears to be optimal for lodging in rock crevices and damp depressions, and in some species (H. monticola Snijman, H. cinna- momea (L'Hérit.) Durand & Schinz, and H. ma- thewsii W. F. Barker) the infructescences have been rved to interlock so that the seeds are locally Mducaned (E (Fig n Boophone sensu stia the more or less spher- ig shape of the infructescence is attained by the rapid elongation of the large number of pedicels after anthesis (Fig. 7G). Thereafter (ca. six weeks), the dried fruiting head breaks away from the top of the scape (character 23, Fig. 2) and is lofted or blown a Korint by the wind. Since the scape remains attached to the bulb, the cohesiveness of the released fruiting head is weakened and it may reak into subunits before lodging. The seeds, held by the indehiscent dry fruits, are gradually released as the papery walls disintegrate during transport. Annals of the Missouri Botanical Garden S M ^ CY x Infructescences of various South African winter- rainfall e ies of popa —A, B. Brunsvigia bos- Figur re maniae; WT of kitelike capsule, wind-blown infructescence.— )ssyne Hessea monticola; numerous, symmetrically arranged pedicels.—D. few pedicels.—E. Strun tescence showing frayed, worn-down fruits.—G. tkiwering "D. F-H = 10 cm. The infructescence configuration of Cybistetes is es- sentially similar to that of many Brunsvigia species; however, the indehiscent, protective, strongly ribbed, dry fruits are dependent on being worn down by rolling (Fig. 7F), a process that gradually releases the seeds. From the character distribution, the secondary loss of the abscission zone in the scape, and a re- versal to short, lax pedicels (characters 9 and 22(0), Fig. 2) have apparently given rise to one further method of wind dispersal in representatives of Stru- maria truncata; note the lax, pendulous Boophone al ZR —H. Crinum variabile; showing the mature fruits dropped close to the parent plant. Scale bars: i ntact infructescence showing a ll la ked infructescences each w Cybistetes longifolia; a wind-rolled d - note the elongation of the pu after A =] yedicels.—F. maria subg. Strumaria. In this group the fruiting head does not detach and the seeds are individually released relatively high above the ground surface. The pedicels in S. truncata Jacq., S. barbarae Ob- erm., and S. hardyana D. € U. Miill.-Doblies are exible and pendulous and swing to and fro in strong winds so that seeds are widely scattered from the parent plant (Fig. 7E). Unlike the secondary transportation of the seed in other wind-dispersed Amaryllideae, the trajectory of the seed in Stru- maria subg. Strumaria is direct from plant to Volume 83, Number 3 1996 Snijman & Linder Phylogenetic Relationships in Amaryllideae ground, and belongs to the dispersal class of ane- moballists (Van der Pijl, 1982). The sequence of character assembly in the phy- logeny suggests that the development of anemogeo- chory in Amaryllidinae possibly commenced with the hysteranthous leaf habit, which is almost basal to the subtribe (Fig. 6), followed by the develop- ment of an abscission layer in the scape at ground level and thereafter by an inflorescence with long pedicels (Fig. 6). In combination, these characters appear to have resulted in the shedding of the in- fructescence at a time when the effect of the veg- etative cover as a potential windbreak was minimal, and with a configuration that could have supported the transition of the entire infructescence into a single, complex unit, in which other characters as- sociated with wind dispersal could be indepen- dently derived. These further specializations (Fig. 6) mostly reflect small quantitative changes: in- creased capsule surface (Brunsvigia), increased pedicel number (Crossyne), and altered pedicel length (Brunsvigia, Crossyne, Hessea, and Strumar- ia). In contrast, neither hysteranthous leaves nor an abscission zone in the scape are basal to subtribe Crininae. The apparent absence of an historical transition toward the light, spherical, detaching in- fructescences of Boophone and Cybistetes (Fig. 6) suggests that the assembly of characters suited to anemogeochory in these two lineages was extremely rapid, or that the intermediate ancestral species are extinct. The differences in the situation of the ab- scission zone in the scape of Boophone sensu stricto and Cybistetes (at the top of the scape and at ground level, respectively) imply that the dispersal units of these genera are independently derived structures. This indicates that anemogeochory evolved twice in Crininae. The interpretation for Amaryllidinae, however, remains ambiguous. Two equally parsi- monious hypotheses can be proposed. Either ane- mogeochory evolved independently in three lin- eages of Amaryllidinae (in the Brunsvigia—Crossyne clade, the Hessea clade, and in Strumaria subg. Gemmaria); or anemogeochory evolved only once in the Brunsvigia—Strumaria clade and was subse- quently lost in Carpolyza and in the Strumaria clade consisting of subgenera Strumaria and Ted- ingea. Either interpretation, however, suggests that the fruiting characters associated with wind dis- persal are plastic. Dispersal by tumbling is restricted to and con- sistently present in the five genera and one sub- genus of Strumaria that we have discussed. With the exception of only Boophone, all are confined to southern Africa and, with the exception of some Brunsvigia and Strumaria species, all are concen- trated in the western half of South Africa. The high- est concentrations of Amaryllideae species with the anemogeochorous habit (based on the number of species recorded in BOL, NBG, PRE, and SAM per 15’ X 15’ grid area) are in the west of the Northern Cape (grid 3119AC, eight species), Namaqualand (grids 2917BA, 3017BB, 3118DA, and 3118DB, five species each), and the Western Cape (3319CB, six species; 3420BC, five species). This clearly in- dicates that wind dispersal in Amaryllideae is more common in South Africa than in the rest of Africa and that the semiarid, winter-rainfall region has the greatest diversity of wind-dispersed representa- tives. Ellner and Shmida (1981) found similarly high concentrations of taxa adapted to tumbling in the semi-desert areas of Israel, and Maddox and Ca- rlquist (1985) reported wind dispersal in many Cal- ifornian desert plants. Van der Pijl (1982) stressed the association between wind dispersal and biotic poverty, particularly in pioneer vegetation, whereas the studies of Maddox and Carlquist (1985) con- centrated on the importance of wind in determining the dispersal mode of desert plants. The winter- rainfall region of southern Africa encompasses two biomes (Rutherford & Westfall, 1986) in which wind and disturbance are major features. The Fyn- bos Biome offers numerous transient niches through frequent fire (Cowling, 1987) and the Suc- culent Karoo Biome, with a 15-50% canopy cover (Hilton-Taylor, 1994), offers many more persistently open sites. Wind in both biomes is reported to ex- ceed that of similar climatic regions elsewhere in the world (Deacon et al., 1992) The other major dispersal specialization in Amaryllideae is represented in the Crinum—Am- mocharis clade (Fig. 6), where the seeds are large, the fruits are variable, and the infructescences have no wind dispersal ability. Most species of Crinum and Ammocharis have green and membranous to brown and papery fruits; however, the fruit walls of a few tropical species of Crinum (C. delagoense 1. Verd., C. kirkii, Baker, C. papillosum Nordal, and C. stuhlmanii Baker) are berrylike with a red to orange, thick or spongy pericarp (Nordal, 1982). This trend corresponds with Dahlgren and Ras- mussen’s (1983) hypothesis that baccate fruits in Amaryllidaceae are secondarily derived. The some- what lax scapes and short pedicels of these genera, whose representatives commonly occupy washes, salt pans, or streambeds (Table 3), do not detach from the bulb. After anthesis the scape elongates, bends downward, and either rests on the ground (Fig. 7H) or droops into water. Until the walls grad- 382 Annals of the Missouri Botanical Garden ually disintegrate, the seeds are retained in the in- dehiscent fruit close to the parent plant. Depending on local conditions, the angular seeds germinate where they were deposited or their buoyancy allows them to be carried into seasonal washes and stream beds or the floodplains of permanently flowing riv- ers (Clark & Parsons, 1994; Manasse, 1990, pers. obs.). Limited dispersal has been regarded either as a strategy to recapture a favorable site where suitable habitats are limited (Zohary, 1962) or alternatively as a product of specialized seed-containers which serve primarily to protect and regulate the timing of dispersal and germination (Ellner & Shmida, 981). The loss of an opening mechanism in the fruit at the base of the Crininae lineage (character 24, Fig. 2) and the development of a large seed with a specialized endosperm at the same point of ) sug- gest an early evolutionary association between the the phylogeny (characters 31 and 32, Fig. indehiscent fruit and the large seeds in Crininae. The subsequent development of the anchoring function by the persistent, lax scape in Crinum and Ammocharis (Fig. 6) suggests, therefore, that tem- poral limitation preceded spatial limitation of seed dispersal in Crininae. A critical analysis of the evo- lutionary sequence of dispersal and protective functions nevertheless awaits tests on the role that the berrylike fruits of Crinum may play in protect- ing the developing seeds from possible microbial damage in moist subtropical or tropical terrestrial habitats (see Corner, 1992). The presence of both wind- and water-dispersal — mechanisms in Crininae indicates greater diversity in dispersal mode in Crininae than in Amaryllidi- nae. Since particularly m endosperm-rich seeds are basal to the Crininae lineage, this pattern sup- ports the hypothesis (Stebbins, 1970) that taxa hav- ing large seeds are particularly subject to selection pressures favoring special mechanisms for seed dis- persal. Although Clark and Parsons (1994) sug- gested that ravens may carry and drop seeds of Cri- num flaccidum, biotic dispersal of the fleshy seeds has never been recorded in the tribe. The large, water-rich seeds of Amaryllideae, which are borne near the ground and dropped at or near maturity, have no smell (pers. obs.). Except in Amaryllis, the seeds lack attractive coloring and none require burial for germination (Isaac & McGillivray, 1965; , pers. obs.). These characteristics of the syndromes that Van der Pijl (1982) has identified as adaptive for biotic dispers- al. It seems possible that the highly toxic alkaloids, Watt & Breyer-Brandwijk, 1962), may have limited seed which are widely present in the tribe (see predation (see Manasse, 1990) and the develop- ment of seed and fruit specializations primarily as- sociated with reptile, bird, or mammalian dispersal agents, and conversely, may have promoted the tribe's evolutionary development toward the exploitation of diverse abiotic dispersal mechanisms. EVOLUTIONARY PATTERNS IN THE ECOLOGY OF AMARYLLIDAEAE It is well accepted that new periods of evolution- ary vigor and new diversification coincide with the conquest of new biological niches, by means of new features (Riedl, fairly rapid departure of a group of organisms from ). Characters that permit a a preceding ecological sphere are referred to as key innovations by Larson et al. (1981). Moreover, such innovations are considered to promote the appear- ance of supportive adaptations during the ecologi- cal transition. The developmental stage at which selective pres- sures would be expected to be maximal is that of i 74). Thus for Amaryllideae we suggest that seedling vigor, as- early seedling development (Stebbins, sociated with the green embryo of the large non- dormant seed, probably contributed to the wide- spread radiation of many ancestral species of the group throughout sub-Saharan Africa during the early Tertiary when seasonally moist subtropical/ tropical conditions prevailed (see Axelrod & Ra- ven, 1978). With the subsequent aridification of Af- rica and the inception of extreme summer-aridity in southwestern Africa during the Pliocene (Deacon 92; Tankard & Rogers, 1978), a postulated key innovation for Amaryllidinae was the devel- opment of a seed with a large, green integument et al., the integument is thought to enable the seed to photosynthesize and and stomatose testa. Since ripen independently of the parent plant (Goebel, 1932), this novel feature of post-ripening is consid- ered to have facilitated the accelerated shedding of the infructescence, and so promoted the transition of the entire infructescence into a unit, highly spe- cialized for wind dispersal (Fig. 6). This occurred especially in the semiarid, winter-rainfall region of southwestern Africa where the short growing season favored the seed’s rapid release to coincide with the onset of the rainfall season, and the extreme winds favored the development of anemogeochory. With increasing aridity the phylogenetic hypoth- esis for Crininae reflects a different evolutionary strategy. The subtribe’s present distribution sug- gests that the evolution of its large seed was mainly promoted in the summer-rainfall regions of sub-Sa- haran Africa, where the duration of the growing Volume 83, Number 3 1996 Snijman & Linder Phylogenetic Relationships in Amaryllideae season imposed little limitation on seed size. Boo- phone and Cybistetes adapted to dry, open sites, but many species of Crinum and Ammocharis adopted a hydrophilic habit that allowed them to survive at the margins of permanent and seasonal water bod- ies. A key innovation that possibly permitted Cri- ninae to evolve both xerophilic and hydrophilic habits was the development of a protective, ecolog- ically versatile, corky seed covering, whose water- resistant and insulating qualities led to the evolu- tion of wind- and water-dispersal mechanisms (Fig. 6) during the climatic changes of the Pliocene. The occurrence, however, of only three wind-dispersed Crininae species in the winter-rainfall of south- western Africa area may reflect an intrinsic con- straint to the production and timely release of the lineage's large seeds in a region where the growing Furthermore, although the pantrop- ical distribution of Crinum may reflect a former Gondwanan distribution, the development of water dispersal and the tolerance of Crinum seeds for sa- line conditions supports Koshimizu's (1930) and Arroyo and Cutler's (1984) suggestion that long-dis- tance dispersal events may have been influential in the past. We propose that the concentration of a great number of Crinum species in southern Africa (Nordal, 1977; Verdoorn, 1973) is a consequence of its atelechorous/rain wash dispersal mechanism which, together with the inability to develop stored seed banks, resulted in high rates of isolation, mi- gration, and extinction among the ancestral taxa in- habiting the seasonal water systems in the increas- ingly arid regions of the west. season is short. CONCLUSION Character distribution in the Amaryllideae indi- cates constancy in floral and vegetative morphology relative to the diversity shown in the seed and fruit- ing structures. Seed characters have been shown to have greatest diagnostic value at the tribal and sub- tribal level. A single integument, at least at some stage of the seed's development, and a green em- bryo are synapomorphic for the tribe. Large cork- covered, green, endosperm-rich seeds characterize Crininae, and seeds with a green, stomatose integ- ument are characteristic of all Amaryllidinae other than Amaryllis. Elsewhere in Amaryllidaceae, fleshy seeds also occur in the African tribe Hae- mantheae and the mainly American tribe Euchar- ideae, but so far little is known about their phylo- genetic importance or ecology (Meerow, 1989). contrast, the phylogeny of Amaryllideae, combined with data on dispersal, germination, and distribu- tion patterns, suggests that much of the diversifi- cation in Amaryllideae arises from the group's nov- el seed characters. We propose that the specialized seed morphology and the unique and complex dis- persal systems in Amaryllideae reflect a series of altered compromises that developed during the cli- matic changes of the past; in particular between the selection pressures that favored large, nondormant seeds, and those that favored diverse but efficient modes of seed dispersal. The seed-dispersal mech- anisms that have evolved in Amaryllideae are all associated with abiotic agents, predominantly wind and water, whereas biotic dispersal remains un- known in the tribe. Literature Cited Almeida, M. T. & F. A. Bisby. 1984. A simple method for establishing taxonomic characters from measure- ment data. Taxon 33: 405—409. Anderberg, A. & A. Tehler. 1990. Consensus trees, a necessity in FETAN practice. Cladistics 6: 399—402. Armbruster, W. S. 1992. Phylogeny and the Aoi y of plant- T interactions Bioscience 42: 12-2 Arroyo, S. C. Contribución al i nuns de los bulbos de Amanidacene Kurtziana 17: 55-70 1984. Evolutionary and — ic aspects uf Ms al morphology in Amaryllidiehe from South America and Southern Africa. Kew. Bull. 39: 467-498. Axelrod, D. I. & P. H. Raven. 1978. Late pow and Tertiary cia history a Africa. Pp. 77-130 in M. J. A. Werg Base ai eap and Ecl of Southern Alia Jun Baker, J. G. i George Bell, London joe dena “of the Amaryllideae. 8 ryllideae. Pp. 171-246 in W. J. Thi- selton- -Dyer ey Flora Capensis 6. L. Reeve, Lon- don. Beaurieux, R. 1914. Observations anatomiques et phy- siologiques sur le Crinum capense Herb. Mém. Soc. Roy. Sci. Liege 3, 10: 1-41. Boyd, L. 1932. Er m seedlings. Trans. Bot. Soc. 1775. Tablecloth Press, Cape Town. The limits of amino acid sequence p in angiosperm phylogenetic reconstruction. Evolutio 42: 795—803. Brummit, R. K. 1987. Report of the committee for sper- matophyta: 33. te 36: 734—739. 88. Choosing among multiple equally parsimonious i sm Cladistics 4: 291—296. Clark, T. & R. F. Parsons. 1994. Ecology of Calostemma and Crinum (Amaryllidaceae) in the River Murray area, South-Eastern Australia. Proc. Roy. Soc. Victoria 106: —145. Corner, E. J. H. 1992. The C in DE Primitive or advanced? J. Linn. Soc. : 15-19. Cowling, R. M. 1987. Fire and its ide] in ae and speciation in Gondwanan shrublands. S. African J. Sci. 83: 1 12. Crisp, M. D. & P. H. Weston. 1995. Mirbelieae. Pp. 245— 282 in M. Crisp & J. J. Doyle (editors), Advances in 384 Annals of the Missouri Botanical Garden Legume Systematics 7: Phylogeny. Royal Botanic Gar- dens, Kew. Dahlgren, R. M. T. & H. T. Clifford. cotyledons: A un. Study. London. & F. N. Rasmussen. 1983. Monocotyledon evo- lution: a and phylogenetic estimation. Evol. Biol. 16: 255 .H.T a & P. F. Yeo. 1985. The Families of the Monoc pg Springer-Verlag, Berlin. Deacon, H. J., R. Jury & F. Ellis. 1992. Selecti regime and is E 7-22 in R. M. Cowling (editor), The Ecology of Fynbos: oo Fire and Diversity. Oxford Univ. Press, Cape Town. Deleporte, P. 1993. (actis attributes and tests of evolutionary scenarios. pedi 9: eo ne Donoghue, M. J., R. G. 1, J. E. LEE J. D. Palmer. 1992. e E Wn e pe based on rbcL sequences. Ann. Missouri Bot. Gard. 7 1982. The Mono- Academic Press, 33-— Dutt, B. S. M. 1957a. O of the ovule of Crinum ide Ker. Curr. Sci. 26: 22-23. —— —. 1957b. Ovule and embryo sac of Crinum asta- Tod & Cult. 22: 688-690. 1950 ;Ovulé and embryo-sac of Crinum 2 3-294 ium L., a E Ae Curr. Sci. 28: 29 16 62. A contribution to the life-history af Cri- num defixum Ker. Pp. 37-4 Symposium. Counc il for Scientific and Indus search, Biological Research Committee, New L lh. . 1970. Amaryllidaceae. Bull. Indian Natl. Sci. Acad. 41: 365-367 . 1950. A review of the genus Brunsvigia. Pl. . 1951. A review of the genus Brunsvigia Heist. (Part 2). Pl. Life 7: 44-64. 1976. Amaryllidaceae. Pp. 947-954 in R. A. Dyer (editor), The Genera of Southern African Flower- ing Plants x a »partment of Agricultural Technical Ser- vices, Pret Ellner, S. € A. "Shmida. 1981. Why are adaptations for long-range seed dispersal rare in desert plants? Oec- ologia (Berlin) 51: 133-144. Erdtman, G. 1966. Pollen Morphology and Plant Taxon- . Hafner, New York. Fangan, B. M. & I. Nordal. 1993. A comparative analysis of morphology, chloroplast-DNA and distribution within the genus Crinum (Amaryllidaceae). J. Biogeogr. 20: 55-6 pue J. S. 1969. A successive grace approach o character weightin t. Zool. 14—385. 1970. M. kj c eir eat trees. Syst. Zool. 19: 83-92. . 1983. The logical basis of phylogenetic analysis. Pp. 7-36 in N. I. Platnick & V. A. Funk (editors), Ad- vances in Cladistics 2. Proceedings of the second meet- ing of the Willi Hennig Society. Columbia Univ. Press, York. 1988. Hennig86: Version 1.5 E Admiral Street, oe E Station, New York 117 , MF, J well, L. R. Caddick, Hn W. Chase. 1994. A molecular evaluation 2 i mono- phyly E Alliaceae and Amaryllidaceae. Amer. J. Bot 81 (6, Suppl.): 154—155. [Abstract.] Forbes, x "i (Editor). 1986. Carl Peter Thunberg. Travels at the Cape of Good Hope 1772-1775. d ond Series No. E d Riebeeck Society, Cape Town. ———— & J. Rourke (Editors). tion. Cladistics 7 7: 395-40 — TN 1980. Paterson's Cape Travels 1777-1779. Brenthurst Press, Johannesbu Frumhoff, P. C. € H. K. Reeve. 1994. Using phylogenies to test hypotheses of adaptation: A critique of some cur- rent proposals. Evolution 48: 172-180. Goebel, K. 1932. mii der Pflanzen 3, 1. 3rd ed. Gustav Fischer, Jen Goldblatt, P. 1972. ose cytology in relation to classification in Nerine and ee (Amaryllida- ceae). J. S. African Bot. 38: 2 5. . 1976. Chromosome cyto ae of Hessea, Strumar- ia and Carpolyza PE ES Ann. Missouri Bot. Gard. 63: 314-320. Goloboff, P. A. 1991a. gd "i the choice among cladograms. Cladistics 7: 215- 1991b. Random data, cm and informa- 6 P-Wee, Version 2.0 NONA, Version 1.0. Computer Programs and Manual. Published by the au- thor, New Gouws, J. B. "1949. Karyology of some South African EE E Life 5: 54-81. Hannibal, L. The Asian- South Pacific Crinum. Notes Waimea eN z Hilton-Taylor, C. 1994. nn m ape Domain (Succulent Karoo). Pp. 204-217 in S. D. Davis & V. H. Heywood (editors), cao of Plant Diversity. A Guide and Strat- egy for their Conservation. Oxford Univ. Press, Oxford. Hofmeister, W. 1861. Neue Beitráge zur Kenntniss der Embryobildung der Phanerogamen 11. Monokotyledon. l th.-Phys. Cl. Kónigl. Sachs. Ges. Wiss. 5: 629— 760. Howell, G. & N. Prakash. 1990. Embryology and repro- due ecology of ps eie i Crinum flaccidum Herbert. Austral. J. B -444. Huber, H. 1969. schaftsverháltnisse der Liliifloren. Mitt. Bot. samml. München 8: 219—538. harivams S. 1937. Karyotype studies in dur eae i. Rep. Tokyo Bunrika Daigaku Sect. B, 52. : Die pides und Verwandt- Staats- W. E. & G. McGillivray. 1965. A note on the water- oring seeds num and of some other South African a greed eae ah = Amar. yllidoideae). Ann. Bot. (London) N.S. 29: —740. Ising, G. 1970. ón of karyotypes in Dus s 65: 1-28. Koshimizu, T. 1930. Carpobiological studies of Crinum asiaticum L. var. japonicum Bak. Mem. Coll. Sci. Kyoto S iol. 5: E Larson, A., D. B. Wake, L. R. M n & R. Highton. 1981. A molecular phylogenetic a on the or- igins of morphological novelties in the salamanders of the tribe Plethodontini (Amphibia, Plethodontidae). iacu 35: 405-422. Linder, H. P. 1991. Confidence limits in phylogenies: = vr d ion the African Restionaceae. Taxon 40: 25: e of two species of C Imp. Univ., Ser. B., Bio 206 Maddison, W. P. & D. ja raei 1987. MacClade version : An interactive, graphic program for analyzing phy- nid and studying ipei. de evolution. Distributed by the authors Maddox, J. C. & S. Carlquist. 1985. Wind dispersal in Californian desert plants: qiue studies and conceptual nia Aliso —96. Manasse, R. S. 1990. Seed size and habitat effects on Volume 83, Number 3 1996 Snijman & Linder Phylogenetic Relationships in Amaryllideae the water-dispersed paren Crinum erubescens (Amaryllidaceae). Amer. J. Bot. 77: 1336-1341. Markötter, E. I. 1936. Die a ae ee van sekere geslagte van die Amaryllida . Ann. Univ. Stellen- eks po Wis-Natuu wk. 2: 1-84. 4. Phylogenetic relationships at the basal radiation of angiosperms: Further study by prob- p of character compatibility. Syst. Bot. 19: 506— Meerow, A. W. 1984. ores! evolution in the Amar- yllidaceae. Herbertia 40: 154. ——. 1989. od p the Amazon lilies, Eu- TERM and pila (Amaryllidaceae). Ann. Missouri Bot. Gard. 76: 136-220. . 1995. Towards a phylogeny : Amaryllidaceae. Pp. 167-179 in P. J. Rudall, P. ibb, D. F. Cutler & C. J. Humphries y Monocotyledons: System- atics and Evolution. Royal Botanic Gardens Merry, J. 7. Formation of periderm in the endosperm of Crinum asiaticum. Pap. Michigan Acad. Sci. 22: 159- Mickevich, M. F. P oH series analysis. Syst. Zool. 31: 4614 Milne-Redhead, E. & a Schweickerdt. 1939, E new conception of the genus Ammocharis Herb. J. Linn. Soc., Bot. 52: 159-197. Müller-Doblies, D. & U. e Liliifloris notulae 2. De taxonomia subtribus Strumariinae (Amaryllidaceae). Bot. s Syst. pu 17-4 992. De bibo notulae 4. Notes on "dm nud V rS DIE. maryllidaceae of Southern Africa). Gleditschia 2 o. 1994. De Labor: notulae 5. Some new taxa and combinations i in the Amaryllidaceae tribe Amary aped from arid Southern Africa. Feddes Re- pert. 105: Nixon, K. C. p CLADOS, = 1.2. Published by the author, Trumansburg, New . Carpenter. 1993. “On outgroups. Cladis- tics 9: 413-426. Nordal, I. 1977. Revision of the East African taxa of the genus Crinum (Amaryllidaceae). Norweg. J. Bot. 24: 179- pe maryllidaceae. Pp. 1—30 in R. M. Polhill ido, Flora of Tropical East Africa. Balkema, Rotter- . Wahlstrgm. 1980. A study of the genus Cri- num (mira i in Cameroun. Adansonia, ser. 2, 20: 179-198. Oganezova, G. E. . Seed and fruit anatomy of some Amaryllidaceae in connection with their systematics and ar Bot. Zhurn. (Moscow & Leningrad) 75: 15-630 Pax, E. 1887. Amaryllidaceae. Pp. 97-124 in H. G. A Engler & K. A. E. Prantl oe Die Natiirlichen Pflanzenfamilien 2, 5. W. Engelmann, Leipzi iran E. P. 1917. A ici to the flora of the e Plateau and environs. Ann. S. African Mus. 16: “379 B. [s ——. 1938. Ammocharis coranica. T. . B. Pole- Evans asas The Flowering ed ha Sh. Af- rica 18. L. Ree je Ashford, Kent. Amaryllidaceae. Pp. 198-207 in E. P. Phillips (editor), The genera of South African flowering plants, 2nd ed. Mem. Bot. Surv. S. Africa 25. Pijl, L. Van der. 1982. DEUM of Dispersal in Higher Plants. Seeinger- Verlag, Berlin Rendle, A. B. 1901. The bulbiform 2 of certain Amaryllideae. J. Roy. Hort. Soc. 26: 89— Riedl, R. 1978. Order in Living ubi ce) : Systems Analysis of ies John Wiley, Chicheste Rutherford, M. C. & R. H. Westfall. 1986. Biomes of southern Afri E objective categorization. Mem. Bot. Surv. S. Africa 54: 1-98. Sató, D. 1938. Karyotype evolution and phylogeny. IV. Karyotypes in Amaryllidaceae with special reference to the SAT-chromosome. Cytologia 9: 203-242. Schlimbach, Von H. 1924. Beitráge zur Kenntnis der Sa- menanlagen und Samen der Amaryllidaceen. Flora 117: -54 Schulze, W. 1984. Beitrage zur Taxonomie der Liliifloren V. Der Umfang der Amaryllidaceae. Wiss. Z. Fried- rich- Pre Univ. Jena, Math.-Naturwiss. Reihe 32: 5-100 Simpson, E r3 1993. Septal nectary anatomy and phy- logeny of the Haemodoraceae. Syst. Bot. 18: 593-613, Snijman, D. A. Systematic Stidite in the Tribe Amaryllideae (Amaryllidaceae) Ph.D. Thesis, Univer- sity of Cape 994. p US of Hessea, Strumaria and Cárpolyza A Amaryllidaceae). Contr. Bolus Herb. 16: 1-162. . Williamson. 1994. A taxonomic re-assess- ment of imacharis herrei and Cybistetes longifolia (Amaryllideae: juu e dll Bothalia 24: 127-132. Stebbins, G. L. 1970. Transference of function as a factor in the evolution of seeds and their accessory structures. Israel J. Bot. 19: 59-70. Stebbins, G. L. 1974. Flowering Plants, Evolution above the Species Level. Edward Arnold, London. Swofford, D. 1993. PAUP: Phylogenetic analysis using parsimony, version 3. L. 1. Illinois Natural History Sur- vey, FIS Illino Tankard, A. J. & J. R NE: 1978. Late Cenozoic palaeo- environments on the west coast of southern Africa. J Biogeogr. 5: 319—337. Toilliez-Genoud, J. 1965. Lovule et la graine de Crinum giganteum And. (Amaryllidacées). Adansonia 5: 525— Tomita, K. 1931. Über die Entwicklung des nackten E Rep. Tóhoku Traub, H. P. 1957. Classification of the oe Subfamilies, tribes and genera. Pl. Life 13: 1962. Classification of the bis Pl. Life 18: 50-54 1963. The Genera of FU. ow The American a e de La Jolla, Californ 1965. da to Traub’s “The Cun. of the mec od peed Pl. Life 21: 88-89. Review of the Genus Nerine. The Amer- ican Plant ne Fa La Jolla, California. 1970. An Introduction to Herbert's *Amarylli- dicm. etc." 1837 and Related Works. J. Cramer, Leh- re. Troll, W. 1954. Praktische Einführung in die Pflanzen- morphologie 1: Der Vegetative Aufbau. Gustav Fischer, ena. Turner, H. 1995. Cladistic and biogeographic analyses of Arytera Blume and Mischarytera gen. nov. (Sapinda- me ón notes on methodology and a full taxonomic n. Blumea Suppl. Vol. 9: 1-230. Gii. busto J. & N 1978. A contribution 386 Annals of the Missouri Botanical Garden to the d a Crinum moorei L. Botanique (Nag- pur, India) 8: Verdoorn, I. C. un k genus Crinum in southern Af- rica. Bothalia 11: 27—52 Vogel, S. 1990. The role ol scent glands in riis Amerind, New Delhi. Translation of S. e 5 Duftdriissen im Dienste der Bestáubung. Uber e und Funktion der Cen. Akad. Wiss. Abh. Math.- Naturwiss. Kl. 10: 599-763. Vosa, €. G. & P. D. Marchi. 1980. Chromosome analysis of d eere a aa (Amaryllidaceae). PI. Syst. Evol. Watt, J. M. ^ x e e NE 1962. The Me- dia ma and Poisonous Plants of Southern and Eastern E Livingstone, nburgh. West, J. G. € D. P. Faith. 199( Due methods and as- ipie in phylogenetic liu Austral. Syst. Bot. 3: 9 Zohary, M. 1962. Plant Life of Palestine. Reinhold Press, London ndix I. mic epee NN Vouchers marked w also used for infrared gas a cen Amaryllis belladonna L.. Africa, Vredenburg, Snijman a (NBG); pde coranica (Ker Gaw Herb., Africa, Bloemfontein, rer 415 (NBC). Pob disticha (L.f.) Herb., Africa, Montagu, Snyman 1268 (NBG); gh helada F. M. Leight., South Africa, Nieuwoudtvi Hi Snijman 588 (NBG); Brunsvigia conil F. M. Leight., South Africa, Karkams, Snijman 275 (NBG); Pici ad W. F. Barker, South Te Laingsburg, Snijman s.n. (NBG 151858); Brunsvigia josephinae (Redouté) Ker Gawl., South Africa, Worcester, Compton 20484 (NBG); Bruns- ar iege: R. A. Dyer, South Africa, Jeffreys Bay, Ma- lan 137 (NBG); Branaiain minor Lindl., South Africa, farm Arendskraal, Snijman 596 (NBG); *Brunsvigia orien- talis (L.) Eckl. € Zeyh., South Africa, near Vanrhynsdorp. Amaryllideae material examined for seed an asterisk were wn = Snijman 437 (NBG); Brunsvigia radula (Jacq.) W. T. Ai- ton, South Africa, Knersvlakte, dian 1254 (NBC). Crinum variabile (Jacq.) Herb., South Africa, Bowes- orp, le Roux s.n. (NBG 892/82): Crossyne guttata (L.) D. & U. os -Doblies, South Africa, Montagu, Snyman 1259 (NBG); Crossyne flava (W. F. Barker ex Snijman) D. & U. Müll. "Doblies. South Africa, Grootvlei, Perry 1126 (NBG); * Cybistetes longifolia (L.) Milne-Redh. € Schweick., Africa, Be quc Du 4637 (NBG); Cybis- tetes 2s o a) Milne-Re : Schweick., South Af- s Bay, Duncan e pet a Hessea be Ca Herb., South Africa, Arabies, William- son 3431 36); Hessea mathewsii W. F. Barker, South Africa, farm Skaapplaas, Snijman 842 (NBG); Hessea pus- illa Snijman, South Africa, farm Perdekraal, Snijman 1072 (NBG); Hessea a on n (Snijman) D. & U. Mill.- Doblies, South Africa, Kamiesberg, Snijman 1179 (NBG). Nerine bowdenii W Watson, So uth Africa, Harrismith, Schelpe s.n. (NBG 878/75); Nerine filamentosa W. F. Bar ker, South Africa, Cathcart, MacMaster s.n. (NBG 148/85); Nerine filifolia Baker, South Africa, Harrismith, Schelpe s.n. (NBG 877/75); Nerine marincowitzii Snijman, South Africa, SE of Leeu-Gamka, Snijman 1245 (NBG); Nerine masonorum L. Africa, n i non & Mason s.n. (NBG 66355); Nerine undulata (L.) Her- bert., South Africa, Engcobo, Gibson s.n. (NBG 59983). Birmania aestivalis Snijman, South Africa, ss Perry 199] 3G); Strumaria chaplinii (W. F. Barker) Snijman, South Africa, ere ea 10239 (NBG); Strumaria gemmata Ker ( uth Africa, Baviaan- skloof, Vlok 956 (NBG); Sirumaria arc (W. F. Barker) Snijman, South Africa, Sutherland, Forrester & Leitch s.n. (NBG 145851); Strumaria Lo (D. € U. Müll.- Doblies) Snijman, d rica, Fontei ea Snijman, BG); Strumaria tenella (L.f.) rica, jd BG); Strumaria truncata Jac South Africa, Steinkopf, Perry 1057 (NBG). A TAXONOMIC SYNOPSIS OF THE GENUS SALSOLA (CHENOPODIACEAE) IN NORTH AMERICA! Sergei L. Mosyakin? ABSTRACT key and synopsis of the genus Salsola L. (Chenopodiaceae) in North America north of Mexico are presented, including the taxonomy and distribut subsp. pontica (Pallas) Mosyakin. tion of six recognized species. A new combination is proposed: Salsola kali L. The genus Salsola L. sensu lato, which includes the segregate genera Caroxylon Thunberg, Clima- coptera Botschantsev, Hypocylix Woloszezak = Dar- niella Maire, Neocaspia Tzvelev, Nitrosalsola Tzvel- ev, and Xylosalsola Tzvelev, comprises 130-150 species, which are especially numerous in the arid and coastal zones of Eurasia, including Central Asia (all sections of the genus), the Middle East, and the Mediterranean region. Some species, most- ly representatives of Salsola sect. Salsola (includ- ing sect. Kali Dumortier), regions of the world as introduced synanthropic weeds, including in North America. e considerable biological diversity of arbores- bby species of Salsola sect. Caroxylon also occur in other cent and shru (Thunberg) Fenzl (or the separate genus Caroxylon) is confined also to northern (Mediterranean), south- ern, and eastern parts of Africa (see Botschantzev, 1969, 1975a, b). Some “satellite” genera closely related to Salsola (Aellenia Ulbrich, Physandra Botschantzev, Horaninivia Fischer & gensohnia Bunge) are widely recognized by bota- nists, while some others are not. However, in most cases the latter segregate genera, usually treated as synonyms of Salsola sensu lato or recognized as infrageneric taxa, are not less distinct from Salsola sensu stricto, than, for example, such readily ac- cepted pairs of genera as Kochia Roth and Bassia Allioni, Salicornia L. and Arthrocnemum Moquin- Tandon, Corispermum L. and Anthochlamys Fenzl, Meyer, Gir- Atriplex L. and Halimione Aellen. It means that the generic boundaries of Salsola and related taxa are in need of a systematic revision. It is also evident that Salsola in the traditional sense should be re- garded as a group of genera, rather than as a nat- ural genus. The taxonomic problems related to forming a generic concept within this group have been discussed recently by Tzvelev (1993). He ac- cepted several quite distinct genera, including Ca- roxylon, which is represented in North America by only one alien species, usually referred to as Sal- sola vermiculata; all other North American taxa be- long to Salsola sensu stricto. As noted above, all species of North American Salsola should be regarded as naturalized or casual aliens native to Eurasia. In the course of preparing a taxonomic treatment of the genus for the Flora of North America it has become evident that the taxonomy, nomenclature, and distribution of some species need clarification. Extensive herbarium collections from MO, GH, NY, and US served as the base for this study. Compar- ative Eurasian material from LE, 3 , and some other Russian and Ukrainian herbaria was also examined. Types of Salsola australis R. Br. and S. caroliniana Walt. sent on loan to MO from BM were also consulted, as were photographs of Lin- naean specimens from LINN. Following is a brief synopsis of the Salsola spe- cies occurring in North America north of Mexico, 'T thank Clifford W. Crompton (Agriculture € Agrifood Canada, Ottawa, Ontario, Canada), Noel H. Holmgren (New York Botanical Garden, Au U.S.A.), Stan Washington D.C., U.S de disc 'ussion a rou comments, and Anne Botanic arden, St. Louis, MO, U. ew. of BM, BRY. GH, KW, LE. “MHA , NY, and A.), Leila M. Shultz (Utah State University, Logan, Botanical Institute, St. Petersburg, Russia), and Stanley L. Welsh (Brigham Young University, Provo, UT, U.S A h and nwyn G. Shetler (National Museum of Natural History, Smithsonian Institution, UT, U.S.A.), Nikolai N. Tzvelev rra A.) for a cheuler McPherson (both ‘Missouri eats Smith and Amy S or their help in du preparation of the manuscript. The kind help of the staff US herbaria is greatly appreciated. Special thanks are due to Nancy R. Morin, Convening Editor, and TA d of the Flora of North America Project (Missouri Botanical Garden, St. Louis, MO. US 2N, G. Kholodny Institute of Botany, 2 Tereshchenkivska Str., Kiev, 252601 Ukraine. ANN. MISSOURI Bor. GARD. 83: 387-395. 1996. 388 Annals of the Missouri Botanical Garden with some notes on their taxonomy and distribution. Unfortunately, it is impossible to present a com- plete account of the North American distribution for all Salsola species, mostly due to taxonomic confusion (especially for species of the S. kali ag- gregate) in the literature. The distribution state- ments in this account are based on reliably iden- tified herbarium specimens seen in the course of this study. For more or less common species only states, provinces, and territories are cited in the following synopsis, while for the rare aliens more detailed information is presented. is article does not pretend to solve all prob- lems concerning distribution and ecology of intro- duced species of Salsola in North America. Its pri- mary aim is to provide a more reliable taxonomic background for further studies. TAXONOMY OF SALSOLA IN NORTH AMERICA Salsola L., Sp. Pl. 1: 222. 1753. TYPE: Salsola soda L. (lectotype, selected by Britton & Brown, 1913). Annual herbs, or subshrubs [in the Old World also shrubs and small trees], glabrous or + pubes- ARTIFICIAL KEY TO SALSOLA TAXA IN NORTH AMERICA cent or hispid. Stems branched (very rarely simple), erect, ascending or prostrate. Leaves mostly alter- nate (occasionally lower ones subopposite), sessile, semiterete, lanceolate, linear, or filiform, margin entire. Inflorescences spicate. Flowers normally bi- sexual, solitary (rarely in twos or threes, but in this case lateral flowers poorly developed) in axils of bracts, with 2 bracteoles. Perianth segments 5, at maturity covering the fruit and in most species de- veloping a transverse dorsal membranous or almost coriaceous wing (sometimes only 2 or 3 perianth segments winged, or all wingless). Stamens 5. Styles and stigmas 2 (3). Seeds usually horizontal; pericarp adherent; embryo spiral. The name is derived from Latin sal, salt; or sal- sus, salty. The name Salsola was first used by Ces- alpino (1519-1603) for a plant that is now known as Halogeton sativus (L.) Moquin-Tandon (basio- nym: Salsola sativa L., published in 1762). Vernacular names. Saltwort, Russian-thistle (English); Soude, Salsovie (French); Salzkraut (Ger- man); Kali, Soda (Italian); Solyanka (Russian), So- lyanka, Kurai (Ukrainian); Solanka (Polish); Slanobyl (Czech). la. Annual herbs, glabrous or papillose-hispid; leaves and bracts with spinose (or at least muc Mee apex; perianth segments complete 2a. Leaves (especially swollen apex; plants always glabrous ene ones) mostl site, wi N [-» . Leaves all alternate (sometimes only 1—3 pairs of lower ones s subopposite), with spinose or r spines en, or in some y glabrous, or indistinctly papillose (occasionally ciliate at ma y oppc ith apex mucronulate, not spinose; i distinctly at be dus or subopposite; Enid segments normally with crenate or dd bw 5. S. soda scent species indistinctly swollen at bue > apex (sometimes papillose at margins, but never brc crenate or pectinate-ciliate); planta papillose- hispid or occ asionally glabrous. 3a. Leaves fleshy (in living plants), linear, in herbarium specimens » mostly 1-2 mm broad, + gradually apex, at maturity forming a slender columnar beak above ihe broad wings; fruiting Scd T= 12 mm diam.; a o open ands and inland saline habitats ... 4b. boa segments w maturity; wings abs with short-acuminate or r triangular apex, never forming a colum bes ent or shorter; fruiting dd 6—7(10) mm diam. or less; plants n maritime saline habitats (seashores, tidal marshes, etc l. S. k S. paulsenii at 5a. oo segments with rigid, aoi apex and distinct midvein; bracteoles not swoller "PES la. S. kali s ab kali 5b. Perianth segments with weak apex and obscure midvein; bracteoles swollen connate at o c habitats), narrowly linear to filiform, cases abruptly narrowed into w j spinose apex; inflor y ide . Leaves normally not fleshy (occasionally somewhat fleshy in plants growing in e pur alkaline n herbarium specimens mostly less than 1 mm = in most eak apical spine (mucro); Bracts appressed ne strongly imbric ate at maturity, gradually enr iam subulate rescence narrowly sp perianth segments wingless or rarely with narrow (usually less than 1 mm) e yracts reflexed or appre ssed a matu ‘ate, rather dense, not interr rupted at m: plants normally erect, branched above the base, or with a few ida branches near the "EE S. collina 6b. Brat reflexed, not imbricate at maturity, in most cases + abruptly narrowed into spinose submucronulate ape half branched from the base. x; inflorescence spicate, at maturity interrupted at least in lower erianth segments normally with membranous wing; plants erect or ascending, Volume 83, Number 3 1996 Mosyakin 389 North American Salsola 7a. Perianth segments with long- acuminate spinose apex, at maturity forming a slender beak above the wings; fruiting perianth 7-12 m 3. S. columnar -l > . Perianth segments with obtuse to wea paulsenii ly acuminate or reflexed apex, at maturity not forming a columnar beak; fruiting perianth normally less than 8-10 mm diam. . S. tragus lb. Perennial subshrubs covered with glabrous (smooth) and minutely denticulate hairs (sometimes becoming glabrous at perit leaves and bracts obtuse, without spinose apex; pean segments with + 6. S pubescent S. Mas Sek sensu lato m papillose) apex 1. Salsola kali L., Sp. Pl.: 222. 1753. TYPE: Herb. Burser XVI(2): 24 (lectotype, selected by Jonsell & Jarvis (1994), UPS not seen). See discussion below. The species is described from Europe: “Habitat in Europae litoribus maris.” Annual herbs, 5-50 cm tall, papillose-hispid or glabrous. Stems branched from the base, erect, rarely ascending; branches arcuate or sometimes prostrate. Leaves alternate, linear, fleshy, in her- barium specimens mostly 1-2 mm broad, semiter- ete, normally not swollen at base; apex + gradually narrowed into rather firm spine. Inflorescence spi- cate, interrupted at maturity. Bracts reflexed, not imbricate at maturity, alternate, narrowing into su- bulate spinose apex. Flowers normally solitary in axils of bracts or reduced upper leaves. Perianth segments glabrous, with weak or firm apex, at ma- turity wingless or with comparatively narrow wing (in subsp. pontica sometimes prominently winged), becoming connate and united with bracteole bases, or bracteoles free. Fruiting perianth ca. 3-5 (rarely up to 8) mm diam 2n = 36 (Basset & Chromosome number. Crompton, 1970). Distribution. Native to maritime coastal areas of Europe, northern Africa, southwestern Asia; in- troduced and naturalized in many other coastal regions of the world, including North America. This species is represented by two (or possibly three) subspecies Note on lectotypification. The specimen LINN 315.1 was selected as the lectotype of Salsola kali by Jafri and Rateeb (1978). T sents a typical form of the species. However, it his specimen repre- lacks a Species Plantarum number, and probably was added to the Linnaean collection after 1753. Because of that this lectotypification was supersed- ed by Jonsell and Jarvis (1994) in favor of the Bur- ser specimen cited above. The second Linnaean specimen identified as S. kali, LINN 315.2, also lacks a Species Plantarum number; moreover, it seems to be identical to S. collina Pallas. la. Salsola kali L. subsp. kali Stems normally papillose-hispid (var. kali), or rarely glabrous (var. polysarca G. F. W. Meyer). Bracteoles free, not swollen. Perianth segments with rigid, almost spinose apex and distinct mid- vein. Habitats. Seashores, salt marshes, sandy places in coastal regions and other saline maritime habitats, very rarely in ruderal inland habitats; 0-100 m. Distribution. Herbarium specimens were ex- amined from the following states, provinces, and territories: FRA Pierre et Miquelon. CAN Newfoundland, Nova Scotia, Prince Edward Island, Quebec. U.S.A.: Connecticut, Delaware, Maine, Maryland, Massachusetts, New Hampshire, New Jersey, New York, North Carolina, Pennsylvania, Rhode Island, South Carolina, Virginia. This sub- species possibly also occurs in the District of Co- lumbia, U.S.A regions of western and northern Europe and is nat- uralized in many coastal regions of the world. . It is native to coastal maritime lb. Salsola kali L. subsp. pontica (Pallas) Mo- syakin, comb. nov. Basionym: Salsola kali var. pontica Pallas, Illustr. Pl.: 37. 1803. Salsola pa de as) A. Degen, Flora Velebitica 2: . 1937. TYPE: authentic specimens in BM, joia not selected; see discussion below. Stems in most cases glabrous (var. glabra Forss- kal = Salsola pontica var. glabra Tzvelev, Ukray- ins’k. Bot. Zhurn. 50(1): 82. 1993), or sometimes papillose-hispid (var. pontica). Bracteoles swollen, connate at base. Perianth segments with weak apex and obscure midvein. Habitats. in coastal regions, rarely in ruderal inland habitats; 0-100 m. Seashores, salt marshes, sandy places Distribution. Herbarium specimens were ex- amined from the following states of the U.S.A.: Al- abama, California (one locality: San Nicolas Island, Lat. 33?15'N; Lon 30'W, U.S. Naval Radio- logical Defense Laboratory, near road above sand spit at 100 ft. elevation, R. E. Foreman (No. 42), 390 Annals of the Missouri Botanical Garden E. C. Evans III, S. C. Rainey), Delaware, District of Columbia, Florida, Georgia, Louisiana, Massa- chusetts, Maryland, North Carolina, New Jersey, New York, Oregon (only one locality: Mississippi. on ballast grounds of Albine, Portland, 29 Sep. 1910, W. N. Suksdorf s.n.), South Carolina, Texas, and Virginia; this subspecies also occurs in Mexico and seems to be common in coastal regions of South America. It is native to coastal maritime regions of southern Europe, northern Africa, and southwestern Asia, northward to Great Britain, Caspian Sea, and is locally naturalized in many eastward to the coastal regions of the world. is subspecies is closely related to S. kali subsp. kali, and in Eurasia replaces the latter on Black, Azov, Caspian Seas. Its nomenclature is complicated and seashores of the Mediterranean, and still unresolved. For a long time it was known in European botanical literature under the misapplied L.) or S. kali subsp. tragus (L.) Celakovsky (in part, excluding names S. tragus (sensu auct., non the type of the basionym). According to Botschantz- ev (1974), the southern maritime taxon of the 5. kali aggregate is conspecific with 5. caroliniana Walter (type at BM), which seems to be the earliest Tzvelev (1993), however, regarded S. caroliniana as a synonym of valid name at the species level. S. tragus sensu stricto. I have studied a small frag- ment of the type specimen of S. caroliniana sent as a loan to MO. This immature plant evidently be- longs to S. kali, not to $. tragus sensu stricto. How- ever, it is impossible to assign it with certainty to any subspecies of S. kali. Moreover, in American literature the name S. or 5. caroliniana (Walter) Nuttall, was sometimes applied caroliniana, kali var. — to other taxa of the S. kali aggregate. Because of this uncertainty in the proper identity of 5. caroli- niana, | selected as a basionym for the subspecies the name given by Pallas (1803) to the coastal Eur- asian race of the S. kali aggregate. At least, this name refers to the native Eurasian taxon, as it was noted by Degen (1937) and Tzvelev (1993). When Pallas (1803) had at his disposal specimens from coastal describing his Salsola kali var. pontica, habitats of the Black Sea in the Crimea near mod- ern Sevastopol (^. . . in littore Chersonesi Tauricae crescentes pro distincta specie habuissem, nisi lec- ” Pal- las, 1803: 37), and probably some additional spec- tae circa Maeotin et in mediterraneis Tauriae, imens from adjacent coasts of the Black and Azov Seas, where only one littoral taxon of the S. kali aggregate is known to occur. The description and illustration in the protologue are also diagnostic. | did not have a chance to select the lectotype among Pallas’s specimens deposited at BM; however, if there are any obstacles for selecting a particular specimen at BM, the lectotypification could be based on the illustration in Pallas (1803). n their secondary, anthropogenous areas of dis- tribution (particularly, in North America) both sub- species of S. kali are often represented by conver- gent or deviate forms and seem to be less morphologically and geographically separated from each other than in their native Eurasian areas (e.g., North American un imens of S. kali subsp. pontica often approach S. tragus in having quite broad wings, à character that is not as common in Eur- asian littoral plants). Due to this peculiarity of in- troduced plants, it would be reasonable for the pur- pose of nomenclatural stability to abandon some uncertain names of taxa described within the 5. kali aggregate from various parts of the world (including S. caroliniana and S. australis; see also discussion below, under S. tragus). When immature, subspecies pontica is almost in- distinguishable from S. kali subsp. kali. Unfortu- nately, many specimens of S. kali sensu lato are represented in American herbaria by immature plants. More detailed studies using fruiting material are needed to clarify the distribution of both taxa in North America. However, 5. kali subsp. pontica is certainly a more southern subspecies (in both North America and Europe), and seems to be the only race of S. kali occurring in littoral habitats from South Carolina to coastal Texas. It is also known from scattered localities as far north as Mas- sachusetts. From New Hampshire to Newfoundland only subspecies kali seems to occur, and it is quite common as far south as Virginia, being gradually replaced southward (in Virginia and North Caroli- na) by subspecies pontica. Both subspecies of S. kali were probably the first taxa of Salsola introduced to North America (evi- dently, on ship ballasts) shortly after the establish- ment of the first European settlements and the be- ginning of colonization of the continent. Salsola kali is confined mostly to coastal saline habitats; however, it rarely occurs as introduced in ruderal inland habitats, but usually not far away from the coast. 2. Salsola Eu L.. Cent. Pl. 2: 13. 1756. Sal- sola kali subsp. tragus (L.) Celakovsky, Prodr. Fl. SP us 2: 155. 1871 (see discus- sion under S. kali subsp. pontica). TYPE: LINN 315.3 (lectotype, selected here; see also Degen, 1937; Tzvelev, 1993). Salsola australis R. Brown, Prodr. Fl. Nov. Holl. 1: 81 (eme deg Botschantzev, Kew Bull. 29: id 1974). ' "S. Australia: Nuyts Archipelago, Pe- Volume 83, Number 3 Mosyakin 391 North American Salsola trel Bay, Isle St. poca 8 Feb. 1802 (fl. & fr.) Good oe Bauer” (lectotype, Im by Botschantzev , BM, cia dex . var. ei odis Tandon, Chenopo- hica Enumeratio: 136. 1840. ine: *Ad Baltam [sphalm. “Balkam” -S. M.] (Besser in herb. DC.)" (holotype, probably at P not seen; isotype in the Bes- ser memorial collection at KW). uus s . var. oup Fenzl in Ledebour, Fl. . 2: 798. 1851. TYPE: no reference to the dn in cl protologue; original specimens annotated by Fenzl at LE, lectotype not selected. i ae 5 " Ae leptophylla Bentham, Fl. Austral. 5: . TYPE locality: usmod and N. S. ed ie es.” ” TYPE: not designated Salsola tragus subsp. iberica Sennen & Pau, Bull. Acad. ntern. Geogr. Bot. (Le Mans), ser. 3, 18: 476. 1908. Salsola iberica Sennen & Pau) Botse ;'hantzev, Bot. Zhurn. (Leningr 19, cum auct. “Sen- u" (comb. invalid Salsola iberica (Sennen & Pau) Botschantzev ex Mec ead 25 Dopol- neniy i i Izmeneniy E ios SSSR”: . 1973, cum auct. “Sennen et ” TYPE: a Contille Mir- anda de Ebro (El las); Logroño, terrains vagues, près la station (Sennen)" (holotype, not seen; isotype, US). Salsola kali L. var. pseudotragus eck in Reichenbach, Icon. Fl. Germ. Helv. 25: 172. 1909. TYPE: not des- meum described from Germany, "Inprimis in terris ioribus.” Salsola ese A. Nelson in Coulter’s New Manual Bot. M 2: 169. 1909. TYPE: "Na dE ~ Salsola ruthenica Iljin in B. A. K Sonye Ras- teniya SSSR (Weeds of the USSR). 2 137. 1934, nom. illegit. Salsola kali L. subsp. ruthenic Soó in Soó et Jávorka, Magyar Nov. Kéz. 786. 1 PE: not designated. Salsola kali L. var. o T P. Aellen, jur i der Botanischen Staatssammlung München 4: 1961. TYPE: ET Lüderitz- Süd. weed in dry riverbed near farmhouse farm "Klein Aus," 28.6. 1949, Kinges 2297 (M not seen). D follow s synonymy of Botschantzev, Kew Bull. 614. 1974]. Distr. Annual herbs, (5-)10—100 cm, glabrous or sparsely papillose-hispid. Stems profusely branched from the base or near the base (rarely simple in underdeveloped specimens), erect, rarely ascend- ing; branches normally arcuate. Leaves alternate, filiform or narrowly linear, in herbarium specimens normally less than 1 mm broad, semiterete, not swollen at base; apex rather soft, subspinescent. In- florescence spicate, interrupted at maturity (at least in lower part). Bracts at maturity reflexed, not im- bricate; alternate, + abruptly narrowing into mu- cronulate-spinose apex. Bracteoles free, or occa- sionally connate at base in lower flowers. Flowers solitary or rarely 2 or 3 (in the last case lateral flowers mostly abortive) in axils of bracts or re- duced upper leaves. Perianth segments glabrous, with weak apex; at maturity distinctly winged; fruit- ing perianth ca. 4-10 mm diam Chromosome number. 2n = 36 (Mulligan, 1961; Bassett & Crompton, 1970). Russian-thistle (English); Vernacular names. soude roulante (French). abitats. Waste places, roadsides, cultivated fields, disturbed natural and semi-natural plant communities (e.g., coastal and riparian sands, semi- deserts and deserts, eroded slopes); 0—2500 m Distribution. Herbarium specimens were ex- amined from the following states and provinces: CANADA: Alberta, British Columbia, Manitoba, ew Brunswick, Nova Scotia, Ontario, Prince Ed- ward Island, Quebec, Saskatchewan. U.S.A.: Ala- ama, Arizona, Arkansas, California, Colorado, Connecticut, Delaware, Georgia, Idaho, Illinois, In- diana, lowa, Kansas, Kentucky, Louisiana, Maine, Maryland, Massachusetts, Michigan, Minnesota, Mississippi, Missouri, Montana, Nebraska, Nevada, New Hampshire, New Jersey, New Mexico, New York, North Carolina, North Dakota, Ohio, Okla- oma, Oregon, Pennsylvania, Rhode Island, South Carolina, South Dakota, Tennessee, Texas, Utah, Vermont, Virginia, Washington, West Virginia, Wis- consin, Wyoming. Judging from the limited number of herbarium specimens available from Alabama, Georgia, Louisiana, and Mississippi, this species seems to be rare (or undercollected) in the south- eastern part of the United States. Salsola tra, probably also occurs in Newfoundland, the District of Columbia (it is known from the adjacent part of Maryland), native to inland arid regions of southeastern Europe and Central Asia, and Florida; and occasionally occurs as an introduced alien in some other regions of the world (naturalized in South Africa, Australia, and South and Central America). Salsola tragus may have been first introduced to the United States in South Dakota in 1873 or 1874 in flax seed imported from Russia (Piper, 1898; Beatley, 1973; Crompton & Bassett, 1985). noxious weed now occupies almost all its dene range in North America. However, young plants are considered to be an additional forage source for livestock in arid rangelands (Welsh, 1984). The mature plant may break off at the stem base to form a tumbleweed (also called Rolly-polly in Australia). The synonymy of S. tragus is complicated, and the correct use of some names applied to this taxon still uncertain. For example, Botschantzev (1974, who had selected the lectotype of 5 S. aus- 392 Annals of the Missouri Botanical Garden tralis R. Brown (deposited in BM), considered it to be conspecific with S. pestifer and S. ruthenica. Crompton (pers. comm.; see also note in Clemants, , who had also studied the type, claimed it to be conspecific with 5. kali sensu stricto. I have also studied the type of S. australis sent on loan to MO. The herbarium sheet contains several fragments. In my opinion, four of these fragments belong to S. kali subsp. pontica, a southern maritime race of the S. kali aggregate (see discussion above). However, the lectotype of S. australis selected by Botschantz- ev (1974) is the fifth fragment, which is mounted at the right side of the herbarium sheet. This frag- ment represents a form morphologically interme- diate between 5. kali subsp. pontica and S. tragus. ne combination S. kali subsp. ae tragus (L. Čelakovský was constantly misapplied in Europe to the littoral taxon treated here as S. kali subsp. pon- tica (see discussion under that taxon). The author- ship of the former combination was often incor- rectly attributed to Nyman (Consp. Fl. Europ.: 631. 1881), who in fact Čelakovský. In North America, this taxon was subsequently treated as S. kali var. tenuifolia (with the authorship incorrectly attributed to Tausch, validated it 10 years after who published only the name, nomen nudum), S. pestifer, S. iberica (with its authorship incorrectly attributed to Sennen Pau, who published their taxon as a subspecies), and S. australis. In Europe this taxon was also known as S. ruthenica and S. kali subsp. ruthenica (Iljin) Soó. name, because, when describing his species, Iljin Salsola ruthenica is an illegitimate cited in its synonymy S. pestifer, the earlier valid name of the same rank. The subspecific name pro- posed by Soó is legitimate, but, since its basionym was illegitimate, it should be regarded as a new name (not a new combination; see ICBN article 58, Greuter et al., 1994). Recently Tzvelev (1993) confirmed that the cor- rect name for the widespread narrow-leaved weedy representative of the S. kali aggregate is S. tragus L., a name that had already been accepted for this taxon by Degen (1937). Judging from the photo- graphs of the Linnaean specimen of S. tragus (LINN 315.3), which I selected as the lectotype, this point of view is correct. This change of a name seems to be desirable, because it would guarantee more stable nomenclature for this taxon in the fu- ture. The name S. tragus was used for this species by some American botanists in the 19th century, but unfortunately, most North American botanists did not resist the temptation to accept the common European misapplication of the name. 3. Salsola paulsenii D. I. Litvinov, Izv. Turkes- tansk. Otd. Russk. Geogr. Obshch. 4, 5: 28. 1905. TYPE: chara. In arenosis subsalsis pr. Farab (ad fl. Amu-darja). 14 Sept. 1903. Legit N. Andros- sow et M. Kolow” (LE). “Turkestania. Dominium Bu- Ross. 49: No hara, in arenosis 1913—S. M.]. a^ AE ~ E ‚Litvinov, Herb. Fl. “Dominium Buc pr. Tak i Sept. 1903 [sphalm. eg. N. Androssow” (LE). 10—80(-100) cm, papillose-hispid. Annual herbs, glabrous or sparsely Stems profusely branched from the base or near the base (rarely simple in underdeveloped specimens), erect, rarely ascending or prostrate; branches straight or arcuate, often almost perpendicular to the stem. Leaves al- ternate, filiform or linear, in herbarium specimens normally ca. 1 mm broad, but occasionally slightly broader, semiterete, not swollen at base; apex spi- nose. Inflorescence spicate, distinctly interrupted at maturity. Bracts at maturity strongly reflexed, not imbricate; alternate, narrowing into spinose apex. Bracteoles free or connate near the base, spreading, spinescent. Flowers solitary, or rarely 2 or 3 (in the — ast case lateral flowers mostly abortive) in axils of bracts or reduced upper leaves. Perianth segments glabrous, with long-acuminate spinose apex, at ma- turity forming a slender columnar beak above the wings, prominently winged. Fruiting perianth 7-12 mm diam. 2n = Chromosome number. 36 (Semiotrocheva, 1983; see fig. Vernacular name. Barbwire Russian-thistle. Habitats. In sandy soil in disturbed natural and semi-natural plant communities (e.g., open sands, sand dunes, sandy waste places, semi-deserts and 0 deserts, eroded sandy slopes, etc.); 0— Distribution. Herbarium specimens were ex- amined from the following states: Arizona, Califor- nia, Colorado, Nevada, Utah. Salsola paulsenii is native to southeasternmost Europe and Centra > Asia This species is confined mostly to open sands, and rarely to saline sandy habitats. It may be ex- pected in the future in New Mexico and Texas, as well as in some Great Plains states. It was first reported from North America by Munz (1968), and the determination was made by Botschantzev. Ad- ditional details of distribution and morphology of this species have been discussed by Beatley (1973) and Fuller (1986). Salsola paulsenii is weakly differentiated from its Volume 83, Number 3 1996 Mosyak 393 yakin North American Salsola allied taxon, S. tragus sensu stricto. The interme- diate forms between them seem to be quite rare in Central Asia, but more common in southeastern- most Europe, western Kazakhstan, secondary syn- anthropic localities in East Europe, as well as in the United States. 4. Salsola collina P. S. Pallas, Illustr. Pl.: 34. 1803. inter Rhymnum et Samaram fl. a jugo Uralensi descendentium" (holotype, BM not seen). TYPE: *in tractu collium cotaceorum Annual herbs, 10-100 cm, sparsely to densely papillose or hispid (rarely almost glabrous). Stems branched above the base (occasionally with slender branches near the base), erect, rarely ascending; branches straight or slightly arcuate. Leaves alter- nate, filiform or narrowly linear, semiterete, some- times semi-amplexicaul at base; apex with rather soft bristle (rarely spinescent). Inflorescence nar- rowly spicate, not interrupted. Bracts at maturity appressed and imbricate, alternate, gradually nar- rowing into a subulate spinose apex. Flowers soli- tary or rarely 2 or 3 in axils of bracts or reduced upper leaves (sometimes flowers are also present at axils of lower leaves and branches, at maturity forming gall-like caducous balls). Perianth seg- ments glabrous, with weak, flaccid apex; at maturity wingless or with narrow erose wing, becoming con- nate and united with bracteole bases; fruiting peri- anth ca. 3-5(7) mm diam. 2n — 18 (Pohl & Gilles- Chromosome number. pie, 1959). Habitats. Waste places, roadsides and railway areas, cultivated fields, disturbed natural and semi- natural plant communities; 100-2000 (?) m. Distribution. Herbarium specimens were ex- amined from the following states in the U.S.A.: Ar- izona, Colorado, Iowa, Kansas, Michigan, Minne- sota, Missouri, Montana, Nebraska, New Mexico, North Dakota, Texas, Utah, and Vermont. The species is also reported from additional lo- calities in Iowa (Schapaugh, 1958), Utah (S. L. Welsh, pers. comm.), and Canada: Ontario (possibly also in Quebec) and Saskatchewan (Crompton & Bassett, 1985). Salsola collina is native to south- easternmost Europe, southern Siberia, and arid regions of Central Asia; it is known as a naturalized or casual alien in some other regions of Europe and Asia. This Asian species was reported for the first time for North America from Minnesota by Moore (1938). However, it had been collected in Kansas in 1923 (Brooks et al., 1976), but was misidenti- fied. Later it spread to Colorado, Iowa, and Missouri (Cory, 1948; Schapaugh, 1958; Muhlenbach, 1979). At present this species is known in North Amer- ica mostly from the Great Plains region and scat- tered localities in other states. Recently it was also discovered in Canada (Crompton & Bassett, 1985). However, its actual distribution seems to be under- estimated due to the common and constant confu- sion with deviate forms of S. tragus, which occa- sionally resemble S. collina in having narrow inflorescences and gall-like caducous flowers/fruits at the axils of lower and middle branches. In the future S. collina may be expected to occur within the present range of S. tragus throughout North America. It was also reported as a casual alien from sev- eral countries of western and central Europe and is regarded to be established or even completely nat- uralized in eastern Europe (see map of its second- ary distribution in Baranova & Khilova, 1990). 5. Salsola soda L., Sp. Pl. 1: 233. 1753. Kali soda (L.) Scopoli, Fl. Carn., ed. 2, 1: 175. 1772. TYPE: LINN 315.7 (lectotype, selected by I. Hedge in Jarvis et al., 1993). [Th cies described from southern Europe: “Habitat in Europae australis salsis."] e spe- Annual herbs, 5-70 cm, glabrous. Stems branched from the base or nearly so, erect or as- cending; branches straight or slightly arcuate (lower ones sometimes almost prostrate). Leaves (especial- ly lower ones) mostly opposite, linear, semiterete, fleshy, in herbarium specimens usually more than 1.5 mm broad, distinctly swollen or ovate at base; apex mucronulate, non-spinose. Inflorescence spi- cate, distinctly interrupted. Bracts at maturity hor- izontally reflexed, alternate or almost opposite, abruptly narrowing into mucronulate non-spinose apex. Flowers mostly solitary in axils of bracts or reduced upper leaves. Perianth segments glabrous, with crenate to pectinate-ciliate apex, at maturity wingless or with rudimental triangular tubercles, not connate to bracteoles. Fruiting perianth ca. 3- 6(7) mm diam. Chromosome number. 2n = 18 (Zakharyeva, 1985). Habitats. Coastal and disturbed saline habitats; m. Distribution. Herbarium specimens were ex- amined from California: San Mateo Co., Palo Alto Yacht Harbor near airport, altitude about sea level, 6 Oct. 1974, J. H. Thomas 17615 (MO); Santa Cla- ra Co., Palo Alto Yacht Harbor, near E end of yacht 394 Annals of the Missouri Botanical Garden basin, edge of disturbed area along salt marsh, Dec. 1975, J. H. Thomas 18062 (US). Salsola ds is native to Eurasia (mostly Atlantic Europe, the Mediterranean region, and southwestern Asia) and northern Africa. Salsola soda is known from several localities in central California, near San Francisco Bay (Tho- mas, 1975). It can be expected to spread in Cali- fornia, or to appear in inland or coastal saline hab- itats in other southern states. 6. Salsola vermiculata L., Sp. Pl.: 223. 1753. TYPE: LINN 315.20 (lectotype, selected by Botschantzev, 1975b). [The species described from Spain: “Hispania.”] Subshrubs, 20-70(-100) em, densely pubescent (especially when young) with smooth and minutely denticulate hairs; sometimes becoming glabrous at maturity. Stems branched at the woody base, branches erect or ascending, virgate. Leaves 5-8 X 0.5-1 mm, semiterete, usually pubescent, ex- panding into ovate base, bearing in their axils sev- eral reduced leaves ca. 1—4 mm long; apex obtuse. Inflorescence spicate but its primary axis some- times. paniculately branched. Bracts normally densely pubescent, obtuse. Bracteoles free. Flowers solitary or rarely 2-3 in axils of bracts or reduced upper leaves. Perianth segments sparsely pubes- cent above wings (especially at apex), sometimes becoming glabrous, with conical apex, winged at maturity. Fruiting perianth (including wings) 7-12 mm diam. 2n — 18 (Sankary, 1986; S. vermiculata L. var. villosa (Del.) Chromosome number. the record for * oq.") Habitats. Rocky slopes, clay soils, disturbed places; ca. 1000 m. Distribution. Herbarium amined from California: San Luis Obispo Co., Re- cruit Grade Pass, 5.2 km W of Kern Co. line, Tem- blor Range, elev. 979 m, 3 Oct. 1978, J. L. Johnson & G. D. Barbe 2448 (US); Temblor S ridge top. and E slope at Crocker Grade, S o ca. 2 mi. W of the county line and 7 . 3200 ft., a . (MO). miculata is native to the Mediterranean region. specimens were ex- Croc E Salsola ver- Salsola vermiculata sensu lato is known in North America only from California, as a locally persis- tent escaped weed. It is naturalized near an aban- doned experimental plot, Recruit Grade Pass, Tem- blor Range, San Luis Obispo Co. (possibly also in Kern Co.), where it was previously tested as a po- tential forage plant introduced from Syria in 1969. Salsola vermiculata is designated as a Federal Nox- ious Weed by the United States Department of Ag- riculture (Westbrooks, 1993). perennial species of Salsola sensu lato introduced It is so far the only to North America. Together with other related Eur- asian and African taxa, it should probably be seg- regated into a separate genus, Caroxylon. Salsola vermiculata sensu lato is a taxonomically complicated and morphologically polymorphic complex represented in Eurasia (mostly in the iterranean region, western and Central Asia) by several closely related races usually treated as sub- species or distinct species (Botschantzev, 1975a, b; 1984). stricto is a western Mediterranean species occur- Greuter et al., Salsola vermiculata sensu ring in southwestern Europe and northwestern Af- rica. Numerous records of S. vermiculata from Syria and other countries of the Middle East refer to other closely related species and/or subspecies, which re- place it in the eastern Mediterranean region. North American material most probably belongs to S. da- mascena sensu stricto. Ít fits the protologue and the but additional study and comparison with other Eurasian “micro- type specimen deposited at LE, species” are necessary. Some of these taxa remain little-known and poorly understood taxonomically. Because of that I prefer provisionally to place the North American plant of Syrian origin in $. ver- miculata sensu lato until further clarification of its taxonomic identity. The nomenclatural citation for Salsola damas- cena and its basic synonymy are provided below. Salsola damascena Botschantzev, Bot. Zhurn. (Len- ingrad) 60(4): 500. 1975. TYPE: “Syrie: Tallus pierreux du jardin Boustan el Nashé a Mezzé près de Damas, 9 Aug. 1856, C. Gaillardot 1627” (holotype, LE). According to Botschantzev (1975a, b), the syn- onymy of Salsola damascena includes S. rigida Pal- tenuifolia Boissier (Flora Orient. 4, 2: 968. 1879, in part) and S. vermiculata subsp. tenuifolia (Boissier) Botschantzev (Novosti Sist. Vyssh. Rast. (Leningrad) 1: 375. 1964). The names Salsola vil- losa Delile (Fl. Aegypt. Illustr.: 57, No. 309. 1813, in part, excluding the type) and S. vermiculata subsp. villosa art Fig (Palest. J. Bot., ser. J, 3, 3: 132. 1945, basionym) were a to 5. damascena sensu las var. 1 part, excluding the type of the stricto. Literature Cited Baranova, E. . V. Khilova. 1990. Materialy k geo- grafiche e cee ristike nekotorykh predstaviteley Volume 83, Number 3 Mosyakin 395 North American Salsola antropofil’nogo elementa flory Severo-Zapada evropeys- koy chasti SSSR (Materials for geographical character- istics of some representatives of the anthropophilous flo- ristic age in the Southwest of the European part of the USSR). Vestn. Leningradsk. Univ., Ser. 3 (Biol.), 1(3): 35-44. Bassett, I. J. & . W. Crompton. 1970. In: A. Lóve (ed- itor), IOPB a number reports. XXVII. Taxon 19: 437-442. Beatley, J. C. 1973. Russian-thistle (Salsola) species in western United States. J. Range Managem. 26: 225- 226. Botschantzev (Bochantsev), V. P. 1969. Rod Salsola L., ratkaya istoriya ego a i rasseleniya (The genus Salsola L.: A concise histo de — and dispersal). Bot. Zhurn. (Leningrad) 54(7): 989-1001. —— synopsis of Salsola (Chenin = I South and South-West Africa. Kew Bull. 2 E 5a. Novye vidy Salsola L. (New species of Salsola Lo Bot. Zhurn. a 60(4): 498-505. ——— Vidy sektsii Versceitlatae Botsch. sei Garon (Thunb.) Fenzl roda Salsola L. (Spe- ectionis Vermiculatae Botsch. sectionis Carox- ylon (Thunb) Fenzl generis Salsola L.). Novosti Sist. Vyssh. Rast. 12: 160—194. 2 Britton, N. L. & A. Brown. 1913. An Illustrated Flora of the Northern United States, Canada, and the British Possessions. Ed. 2. Ne Brooks, R. E., R. L. McCann & L. A. Hauser. 1976. Vascular inns new to the state of Kansas. Pp. 1-12 in: New Records of the Fauna and Flora of Kansas for 1975. Techn. Publ. State Biol. Surv. Kansas, No. pec S. E. 1992. Chenopodiaceae and Amarantha- of New York State. New York State Mus., Bull. No. 485: 1-100. Cory, V. L. 1948. Salsola collina Pall. in Colorado. Leafl. W. Bot. 5: 104. Cidlnpton, C. W. & I. J. Bassett. 1985. The biology of Canadian espe e Salsola pestifer A. Nels. Canad. Pl. Sci. 65(2): 379-388. Degen, A. 1937. ires Velebitica. Vol. 2. Budap Fuller, Th. C. 1986. Russian thistles PY of the: Sierra Nevada. Calif. ire be Soc. Newsl. 5 (3): 4-6. Greuter, W., t & G. Long om 1984. Med- Checklist. A EA Inventory of Vascular Plants of the Circum-mediterranean Countries. Vol. 1. Conser- vatoire et Jardin botaniques de la Ville de Genève, Ge- e nève. Greuter, W. G., F. R. Barrie, H. M. Burdet, W. G. Chal- oner, V. Demoulin, D. L. Hawksworth, P. M. Jørgensen, D. H. Nicolson, P. C. Silva, R. P. Trehane & J. McNeill. Editors. 1994. International Code of Botanical Nomen- clature (The Tokyo Code), Adopted by the XVth Inter- national Botanical Congress, Yokohama, August—Sep- tember, 1993. - S. M. H. & F. B. Rateeb. 1978. Chenopodiaceae. . 1- fo; in Flora of L ai Vol. 58. Al Faateh Univ., vn Sci., Heg Bot., Tripo . Barrie, p. M. Allan E L. Reveal. : 1-100. C. E. Jarvis. 1994. Lectotypification of Lin- naean names for Flora Nordica Vol. 1 (Lycopodiaceae— depu pe Nordic J. Bot. 14: 145-164. = J. W. 1938. aprisa dentata and Salsola collina n Minnesota. Rhodora 40: 135. Muhlenbach, V. 1979. C EA to the synanthropic (adventive) flora of the railroads i in St. Louis, Missouri, . Ann. Missouri Bot. Gard. 66: 1-108. Mulligan, G. A. 1961. fp a numbers of Canadian weeds. III. Canad. J. Bot. 39: 1057-1066 Munz, P. A. 1968. ppan to a California Flora. niv. California Press, Berkeley allas, P S. 1803. Illustrationes plantar rum imperfecte vel nondum cognitarum, cum centuria iconum. De Halo- phytis, seu plantis apetalis kalicis generatim. Lipsiae. Piper, C. V. 1898. The Russian Thistle in Washington. Washington Agric. Exp. Sta. Bull. 35: 1-18. Pohl, R. W. & J. P. Gillespie. 1959. Distributional and + notes on Salsola collina. Rhodora 61: 265— 267 Sankdty, = N. 1986. Chromosome number reports 91. Taxon 35: 404. Schapaugh, W. 1958. Salsola o Pall., Proc. lowa Acad. Sci. 65: 118-12 Semiotrocheva (Semiotroc zeva), N E m new to Iowa. 1983. _Kariosiste- Thomas, J. HH. 1975. Salala soda L. {Chenopodiaceae in central California. pene 23:9 Tzvelev (Tsvelev), N. N. 1993. Z ides o marevykh (Che- nopodiaceae) iii Pha Evropy (Notes on Chenopo- diaceae of Eastern Europe). Ukrayins'k. Bot. Zhurn. iei T^ Welsh, 1984. Utah flora: Chenopodiaceae. Great Basin T eem 44: 183-209. Westbrooks, R. G. 1993. Exclusion and eradication of foreign weeds from the United States by USDA APHIS. Pp. 225-241 in: B. cKnight My ad Ec Pollution: The Control ‘and Impact of Invasive Exotic Species (Proceedings of a Symposium bad a at t eh Uni- versity Place Conference Center, Indiana University— Purdue University at Indianapolis on October 25 & 26, 1991). Indiana Acad. Sci., Indianapolis. Zakharyeva, O. I. 1985. Chisla khromosom nekoto tsvetkovykh rasteniy Kavkaza i Sredney Azii (Chromo- bers of some flowering plants from the Cau- casus and Middle Asia). Bot. Zhurn. (Leningrad) 70(12): 1699-1701. RITMOS TEMPORALES DE LA Rodolfo Dirzo** INVESTIGACION Guillermina ae TAXONOMICA DE PLANTAS VASCULARES EN MEXICO Y UNA ESTIMACION DEL NUMERO DE ESPECIES CONOCIDAS! RESUMEN analizaron los patrones temporales de la investigación taxonómica en México con base en un conteo o estimación del número de taxa (específicos e —€— ‘fficos) descritos para ese p a parie jai establecimicnip del sistema de Lineo (en 1753) y hasta 1988. Se encontraron tendencias temporales qu en ¿l conteo ac ae produjo un a de 23,630 nuevos taxa mexicanos Aon: para la ciencia en 235 años de existencia de la taxonomía. Esta cifra se utilizó como base para estimar el número de especies conocidas de México hasta 1988 utilizando dos factores de corrección, a saber: (i) una estimación de la proporción de especies que, aunque ocurren en México, no fueron descritas para ese país, y (ii) una estimación pi grado de redundancia nomenc ae (sinonimia). Con este método se llegó a un ud eal de 16,870 taxa. Además, la curva de acumulación de nue que se encuentra lejos de la asíntota, y en los últimos años la tasa es mayor a 150 por año. Bajo el pao de que aún falta por reconocer aproximadamente A 20% del total tioribto del país, calculamos que el número llegaría a 20,244, cercano a la estimación de Rzedowski, de 22,800. Nuestra estimación y la de Rzedowski sugieren que la cifra ampliamente citada de 30,000 especies debe tomarse con SEM Nuestra Pat conservadora, y la tendencia actual creciente de seguir acumulando nuevos taxa sefialan a México como uno de los centros de mayor diversidad botánica del planeta. = & a ABSTRACT Temporal patterns of research on Mexican vascular plants were analyzed on the basis of counts or estimates of taxa (species and infraspec ific) described from Mexico since the establishment of the Linnaean system (in 1753) up to 1988. Temporal tendencies were found that correlate with known historical events. The cumulative count yielded a total of 23,630 Mexican taxa described in the 235 years of existence of the binomial system. This number was used as a basis to estimate the number of taxa known to Mexico up to 1988 using two correction factors: (i) an estimate of the proportion of species that, even though they are present in Mexico, they were not described from that country, and (ii) an estimate of the degree of nomenclatural ae eis (synonymy). With this method we vision at a total of 5 870 taxa. Moreover, the curve of the cumulative number of new taxa seems to be far from reaching the asymptote, and over recent years the rate is greater than 150 per year. Under the argument that approximately 20% of the total em ric naa of the country is yet unknown, we calculate that the number would rise to 20,244, close to Rzedowski's estimate of 22,800. Our estimate, and that of nada suggests Ha E widely cited figure of 30, sd species should be taken with caution. Our conservative estimate and the lack of e ce of an asymptote in the rate of ac M tion of taxa detected in this study underscore Mexico as one of the inde s of greater floristic oa on the plan La sistemática, como disciplina biológica, opera colección en el campo, conducentes al nombra- en dos principales campos de acción. El primero, miento y clasificación de los organismos. El segun- comúnmente conocido como sistemática descripti- do, como extensión lógica, concierne al estudio del va, o taxonomía, se refiere a la exploración y re- origen, evolución y mantenimiento de la diversidad ! Este estudio se llevó a cabo con el apoyo ec onómico de una beca postdoc ‘toral de la Fundación Jessie Smith Noyes literatura de gran utilidad; en particular agradecemos a Roy E. Gereau, Al Gentry, Fernando Zuloaga, William D y Dale Johnson. Jerzy Rzedowski, Peter Raven, Fernando Zuloaga y un revisor anónimo revisaron una versión preliminar y nos hicieron correcciones muy valiosas. Elizabeth Shaw nos proporcionó un listado de holótipos y lectótipos de México, a partir de las tarjetas del Gray Index, o a los pem años. 2 Missouri Botanical Garden, P.O. Box 299, St. Louis Missouri 63166-0299, U.S.A. * Dirección actual: Centro de Ecología, Departamento de Ecología Evolutiva, UNAM, Ap. Post. 70-275, México 04510, D.F., México. ANN. MISSOURI Bor. GARD. 83: 396-403. 1996. Volume 83, Number 3 1996 Dirzo & Gómez 397 Investigación Taxonómica en México biológica (Raven & Johnson, 1986). De esto se des- prende que la sistemática es una ciencia funda- mental para muchas otras disciplinas biológicas (p. ej., ecología, fisiología, evolución), pero es además una ciencia de importancia pragmática, por ejem- plo para la búsqueda e identificación de fármacos, nuevas fuentes de alimento, agentes polinizadores, agentes de control biológico, y para el diseño de unidades de conservación de la biodiversidad, etc. (Wilson, 1988a, b, c). Sin embargo, a pesar de su enorme importancia, la sistemática, aún en el pri- mero de sus dos objetivos, todavía se encuentra le- jos de cubrir su cometido. Esto es al punto tal de que, el número de especies del planeta, no se con- oce siquiera en términos de su orden de magnitud. Por ejemplo, a partir de la inauguración del sistema binomial por Lineo, en 1753, se han descrito apen- as unos 1.4 millones de especies, mientras que las estimaciones más recientes (p. ej., Erwin, 1983) sugieren que el número podría ser cercano a 30 millones. Claramente la sistemática, en sus dos ver- tientes, es un campo que requiere atención priori- taria a nivel planetario, sobre todo a la luz de las alarmantes tasas de extinción biológica contempor- áneas (véase Dodson & Gentry, 1991; Raven, 1987; Sarukhán & Dirzo, 1992; Wilson, 1988c) En esta contribución presentamos un análisis temporal de la intensidad de la investigación tax- onómica de las plantas vasculares de México (ex- ploración botánica y descripción de especies) y, como corolario, ofrecemos una estimación del po- sible número de especies (y taxa infraespecíficos) conocidos para el país hasta ahora. Nuestro primer objetivo se basa en un sondeo del número de taxa publicados para el país a través del tiempo. Para nuestro segundo objetivo usamos la misma base, con unas correcciones que más adelante se descri- ben. El territorio mexicano, de aproximadamente dos millones de km?, contiene una enorme diversidad de hábitats naturales que van de la selva tropical húmeda a las comunidades de vegetación alpina; además, por su historial geológico y por su posición latitudinal, es el sitio de encuentro de floras de origen Neártico y Neotropical (véase Rzedowski, 1978, 1991) y es asiento de una de las floras más icas del planeta (Dirzo, 1994; Rzedowski, 1978, 1991; Toledo, 1988). MATERIALES Y MÉTODOS EL CURSO TEMPORAL DE LA INVESTIGACIÓN TAXONÓMICA Para este análisis nos basamos en dos índices importantes que recopilan la información concer- niente a las especies vegetales nuevas para la cien- cia, el Index Kewensis (IK) y el Gray Index (GI). Ambos índices proveen los nombres de especies de plantas vasculares que se publican, incluyendo el nombre científico de la planta, el autor, el año en que se publicó la descripción de la especie, el or- igen geográfico y la fuente de información referente a la publicación del taxón. El GI consiste de un juego de tarjetas, cada una de las cuales contiene la información referente a un holótipo y lectótipo dado. Dicho índice se usó mayoritariamente, ya que contiene información exclusiva para las plantas del hemisferio occidental. Para nuestro análisis revi- samos todas las tarjetas del GI transcritas en los volúmenes 1-12 (correspondientes a 1 ) 968) y Pg l y 2 (correspondientes a 1965-1977) (GI, 1978), asf como las tarjetas co- rrespondientes a los últimos 11 años (hasta 1988). Para hacer el conteo de los taxa publicados antes de 1886, utilizamos los volúmenes 1 y 2 del IK, los cuales recopilan la información a partir del es- tablecimiento del sistema de Species Plantarum, por Lineo, en 1753 (IK, 5). Para este periodo de 1773-1885, y debido a que el IK aglutina la información para los taxa de todo el mundo, hici- mos una estimación del número de especies des- critas para México utilizando una muestra de 100 páginas seleccionadas al azar. Por lo tanto, para ste periodo, sólo tenemos un número estimativo global del número de taxa descritos, mientras que a partir de 1886 y hasta 1988, proveemos infor- mación detallada en términos del número de taxa (obtenido por búsqueda y conteo directo) descritos por cada año. e ESTIMACIÓN DEL NÚMERO DE ESPECIES CONOCIDAS El conteo de los taxa descritos de México se tomó como base para estimar el posible número de ama (incluyendo los taxa infraespecfficos). A dicho conteo se le incorporaron dos factores de cor- wee a necesarios: (i) adición del número de es- pecies que, aunque se distribuyen en México, no fueron descritas originalmente para ese país. Es de- cir, se requirió una estimación de la relación es- pecies no descritas : especies descritas de México y después adicionar dicho estimado al conteo inicial. (ii) A esta última cifra habría que reducirla para lidiar con la redundancia de nombres (i.e., sinoni- mia) adjudicados a una misma entidad. En otras palabras, se requirió estimar la razón sinónimos/ nombres aceptados y corregir la cifra descrita en el inciso anterior de acuerdo a ésto Para el primer caso se tomó una muestra al azar de nombres (600 en total) escogidos de doce floras, 398 Annals of the Missouri Botanical Garden flórulas o listas que aportasen la información sobre la publicación de especies. Cada nombre escogido se verificó en el IK, o en el GL o en la publicación de su descripción original, para definir el país de origen. Las floras, flórulas o listas utilizadas fueron: Arboles y Arbustos de México (Standley, 1920- 1926), Golfo de California (Johnston, 1924), Flora Novo Galiciana: Compositae (Mc Vaugh, 1984) y Le- guminosae (McVaugh, 1987); Flora Fanerogámica del Valle de México (Vol. II Dicotyledoneae) (Rze- dowski & Calderón de Rzedowski, 1985); Listados Florísticos de México: IV Flora de Chiapas (Breed- love, 1986), V Angiospermas Acuáticas de México (Lot € Novelo, 1986), VI Flórula de la Isla de Cozumel (Téllez € Cabrera, 1987), VII Estación de Biología Tropical Los Tuxtlas, Veracruz (Ibarra & Sinaca, 1987); Flora de Veracruz: Chloranthaceae (Ludlow-Wiechers, 1978), Araliaceae (Sosa, 1979), (Moreno, 1980), Cannaceae (Jiménez, 1980), Nyctaginaceae (Fay, 1980), Clethraceae Bárcena, 1981), Ebenaceae (Pacheco, 1981), Cy- atheaceae (Riba, 1981), Papaveraceae (Martínez- Ojeda, 1982), Bignoniaceae (Gentry, 1982), Con- naraceae (Forero, 1983a), Martyniaceae (Taylor, 1983), Juglandaceae (Narvae Flores, 1983), Bru- nelliaceae (Nee, 1985), Marattiaceae (Palacios, 1990); Flora de Chiapas: Malvaceae (Fryxell, 1990); Las Cactáceas de México (Bravo-Hollis, 1978; Bravo-Hollis € Sánchez-Mejorada, 1991). Para el caso de la sinonimia se utilizaron nueve —. volúmenes de la Flora Neotropica, correspondien- tes a las siguientes familias (o grupos dentro de familias; ver referencias): Bignoniaceae (Gentry, 1980); Bromeliaceae (Smith & Downs, 1974, 1977, 1979); Connaraceae (Forero, 1983b); Flacourti- aceae (Sleumer, 1980); Lauraceae (Kubitzki & Renner, 82) Lecythidaceae (Prance & Mori, 979); Moraceae (Berg, 1972); Olacaceae (Sleumer, 1984); Zingiberaceae (Maas, 1977). Una ventaja de la Flora Neotropica es que se incluyen todos los sinónimos existentes para cada nombre, no sólo los sinónimos conocidos a nivel local (como en el caso de otras floras regionales más específicas para Mé- xico). De estas floras se escogieron nombres al azar (380 en total) y a cada uno de ellos se les contó el nümero de sinónimos. Además, se utilizó el análisis de los sinónimos correspondientes a las Solana- ceae, a nivel mundial, derivado del tratamiento re- ciente de la familia por D'Arcy (1990, y com. pers.). Con las razones sinónimos/nombres acepta- dos calculadas para cada grupo, se calculó una ra- zón promedio, la cual fue utilizada como factor de corrección de la redundancia nomenclatural. RESULTADOS Y DISCUSIÓN LA INVESTIGACIÓN TAXONÓMICA A TRAVES DEL TIEMPO El muestreo aleatorio de los voltimenes I y II de IK (que abarcan de 1753 a 1885) arrojó un pro- medio de 3.07 (error estándar + 0.759) especies descritas para México, por página de dicho índice. El total de páginas (considerando sólo aquellas que contienen listados de especies) que incluyen estos volúmenes es igual a 2525 [i.e., 1268 (Vol. I) + 1257 (Vol. ID]. Por lo tanto, con base en el pro- medio, el número estimado de especies para ese periodo es 3.07 X 2525 = 7752. Un conteo directo de las primeras 100 páginas del Vol. I arrojó un valor de 310 especies; este número es considera- blemente cercano al valor de 307 que se estimaría con base en el muestreo aleatorio. Adicionalmente, dado que el IK no contiene información referente a taxa infraespecíficos, éstos se contaron directa- mente del GI, el cual sí provee la información (a partir de 1753) para dichos taxa. El total de nom- bres infraespecfficos contados fue 480, el cual su- mado al valor de 7752, produce un total de 8232 para el periodo que va de Lineo hasta 1885. Este resultado implica, crudamente, que para este lapso inicial del estudio taxonómico de las plantas vas- culares de México, la tasa de "descubrimiento" y descripción de plantas nuevas para la ciencia era de 58 por afio. partir de 1886, el conteo directo del GI nos permite ofrecer una imagen más detallada del curso temporal del estudio taxonómico para las plantas vasculares de México (Fig. 1). Para el resto del siglo XIX la tasa de descripción de especies (mas los taxa infraespecíficos) procedió a un ritmo (pro- medio) notablemente alto, 250 por afio, observán- dose las máximas oscilaciones en 1892 espe- cies) y 1894 (481 especies). La existencia de los altos valores iniciales es explicable por la extraor- dinaria y tesonera labor de algunos botánicos, entre los que sobresalen el Espafiol Martín Sessé y el Mexicano José Mariano Mociño, en particular por sus monumentales obras (en coautoría) Plantae Nova Hispaniae (1887-1891) y Flora Mexicana (1887-1897), aunque muchos de los taxa nombra- dos deben ser sinónimos. (Para una discusión de las posibles fechas correctas de la publicación orig- inal de estas obras en La Naturaleza, véase Smith, 1942.) El tercio final de este siglo es seguido por otro lapso que, bajo el nuestro sensor de nuevos taxa descritos, parece ir en caída, sobre todo a par- tir de 1910 (coincidiendo con la Revolución Mexi- cana) y por una década subsecuente. Este lapso de 1900 a 1920 acusa un promedio de nuevos taxa por afio de 170, e incluye el segundo valor indi- Volume 83, Number 3 1996 Dirzo & Gómez 399 Investigación Taxonómica en México «t "4 400 = i A p e 9 [^ M : 200 7 L 100 - aw M a 0 Ld LI x LI x LI T Y 1885 1895 1905 1915 1925 1935 1945 1955 1965 1975 1985 1995 AÑO Figura 1. Curso temporal de la descripción de especies (y taxa infraespecíficos) de México (i.e., basado en tipos mexicanos), a partir de 1885. Los valores que muestra la figura se obtuvieron por conteo directo del Gray Index (ver detalles en el texto). vidual anual más bajo (44) de todos los tiempos. El periodo de 1920 a 1950 es de cierta recuperación pero, más A de marcada oscilación, re- gistrándose aquí el mayor valor anual de taxa de toda la historia, en 1924 (612). En este periodo que Rzedowski (1981: 10), llama la “etapa heroica de la botánica mexicana” destacan las contribuciones de Mexicanos excepcionales como Maximino Mar- tínez, en conjunción con las obras monumentales de algunos extranjeros como Trelease (1924) con sus estudios sobre Quercus y la excepcional anto- logía sobre los árboles y arbustos de México de Standley (publicada en el lapso de 1920-1926). Estas contribuyen de manera significativa a la pro- liferación taxonómica observada y explican en buena medida el inusitado pico de 1924. El periodo de 1951 a 1970 observa una disminución sensible, con un promedio de 70 taxa nuevos por año, y con menos oscilación. En este intervalo calificado por Rzedowski (1981: 12) como “la etapa de estudios sobre la vegetación” las contribuciones taxonómi- cas son mayoritariamente de extranjeros. Finalmente, en las últimas dos décadas, se nota una resurgencia de la investigación taxonómica, con una notable contribución de botánicos Mexi- canos (complementaria a la aún mayoritaria contri- bución de taxonómos extranjeros) y con un impre- sionante promedio anual de 110 taxa. Es de destacar que en este periodo reciente, han surgido hallazgos inusitados como el descubrimiento de una excepcional familia, Lacandoniaceae (Orden Triuridales), con características dg sor- prendentes (véase Márquez-Guzmán et al., : Martínez y Ramos, 1989), así como el ddl ario descubrimiento del maíz diploide, perenne (Il- tis et al., 1979). ESTIMACIÓN DEL NÜMERO DE ESPECIES (Y TAXA INFRAESPECÍFICOS) Con los datos del número de taxa descritos para cada afio (cf. Fig. 1), se elaboró una curva del ná- mero acumulativo de especies (incluyendo los taxa infraespecíficos) a través del tiempo (Fig. 2). Por simplicidad, los puntos de la curva corresponden al valor sumado de cinco afios consecutivos. De hecho, en los últimos 100 años (1888-1988) se describió el doble de especies (15,398) del número descrito en el periodo correspondiente al lapso que va desde la implantación del sistema de Lineo (en 400 Annals of the Missouri Botanical Garden 25 : JA a -J = S i. Paj 5 5 4 « O 10 J ^ : 5 TT es Co ee E e a Ne 1880 1900 1920 1940 1960 1980 2000 1890 1910 1930 1950 1970 1990 AÑO Figura 2. Número acumulado de taxa descritos para México, a través del tiempo. Cada punto en la curva corres- 5 ponde al valor sumado de taxa en periodos subsecuentes de cinco años. El valor acumulado correspondiente a 188 se obtuvo a partir del muestreo del Index Kewensis (ver detalles en el texto). 1753) hasta la iniciación del GI (1886). A partir de la curva acumulativa se llega a un total de 23,630 taxa nombrados formalmente hasta 1988, aunque es evidente que la curva aún no muestra una as- intotización. Nuestro número final estimado de 23,630 tiene los siguientes atributos: (i) no toma en cuenta que la flora del país incluye especies descritas en otros países (es decir, no basadas en tipos de México), y (ii) no considera la redundancia de nombres adju- dicados a una misma entidad específica. La depur- ación de estas dos características nos permitiría of- recer una estimación del número de especies conocidas de México. La corrección referente a la relación especies descritas : no descritas para México se basa en los resultados del Cuadro 1 (ver Materiales y Métodos). s doce floras, flórulas o listas analizadas varían en la proporción de taxa descritos para México des- de 20 (Cozumel) a 84.2% (Compositae de Nueva Galicia). Dado que la proporción global calculada es influenciable por el tamaño de muestra utilizado para el cálculo de la relación de cada flora en par- ticular (por ejemplo, si una relación baja es resul- tante de un tamaño de muestra grande, la relación global tenderá a ser desproporcionadamente baja) se tomó, como criterio conservador, el promedio de las doce relaciones individuales. Con este criterio, la proporción resultante es 46.86% “mexicanas”: 53.14% “no mexicanas.” Aplicando este promedio de 46.86% como factor de corrección a nuestro val- or de 23,630, el número que se obtiene pasa a 50,427 (i.e., 23,630 X 100/46.86). Nuestro factor de corrección tiene un coeficiente de variación de 40.6%, lo cual resulta en parte adjudicable al tamaño de muestra (N = 12 floras/ flórulas/listas). Desafortunadamente no parece ex- istir otra fuente de información disponible con la cual confrontar nuestra proporción. Debido a su in- erés intrínseco y su potencial de aplicación, sería de gran utilidad calcular la proporción en cuestión con un tamaño de muestra más robusto. Finalmente, nuestra corrección referente a la re- dundancia nomenclatural (sinonimia) se basa en los resultados del Cuadro 2. La razón sinónimos : nom- bres válidos varió considerablemente entre grupos, con un ámbito de 0.62 (Connaraceae) a 4.37 (Fla- courtiaceae). Nuevamente, por las razones expues- tas para la corrección anterior, se optó por el cri- terio más conservador de utilizar el valor promedio, el cual resultó de 1.989 (coeficiente de variación = 56.2%). Con base en esta muestra calculamos que por cada nombre hay casi dos sinónimos. Apli- cando este factor de corrección se llega, finalmente, a un estimado de 16,870 (i.e., 50,427/2.989) es- Volume 83, Number 3 1996 Dirzo & Gómez Investigación Taxonómica en México Cuadro 1. Cálculos del número de especies descritas de México y de otros países, obtenidos del muestreo de doce floras, flórulas o listas. Estos datos se utilizaron para estimar el porcentaje de especies descritas con respecto a las no descritas de México, pero que se distribuyen en el mismo país. (X = promedio de los doce porcentajes; D.E. = desviación estándar.) Descritas de Flora/flórula o listado* México otros países Total % Cactáceas de México 33 15 48 68.75 Arboles y arbustos 70 29 99 70.70 Chiapas 14 30 44 31.82 Chiapas: Malvaceae 8 13 21 38.10 Acuáticas 4 14 18 22:22 Cozumel 2 8 10 20.00 Flora de Veracruz 13 22 35 37.14 Los Tuxtlas 6 14 20 30.00 Golfo de California 27 29 56 46.43 Nueva Galicia: Legum. 50 47 97 51.55 Nueva Galicia: Comp. 80 15 95 84.21 Valle de México: Dicot. 34 23 57 59.65 x + DE 46.86 + 19.04 * Las referencias para estas fuentes de información se dan en la sección de Materiales y Métodos. pecies y taxa infraespecíficos conocidos hasta 1988 para el territorio mexicano. La tendencia temporal mostrada en la Figura 2 señala claramente la ausencia de un patrón asin- tótico hacia el final del lapso computado. De hecho, en los dos últimos años (1987 y 1988) para los que se recopiló la información completa del IG, nuestra búsqueda arrojó valores de 191 y 173 taxa, res- pectivamente. Es decir, la proyección inmediata es que, de seguir la misma tendencia, los próximos años verían una tasa mínima subsecuente de más de 150 especies por año (tal vez el número sería más cercano a 200 que a 150). Desde luego, es esperable una asintotización de la curva, pero re- sulta claro que la flora mexicana aún tiene un po- tencial de incremento. Rzedowski (1991) argumen- ta que el porcentaje de las especies de fanerógamas reconocidas está por debajo del 90% del total, aunque seguramente por encima del 75%, y estima que “un complemento de 20% ... tal vez consti- tuya una aproximación razonable” (p. 6). Si agre- gamos dicho complemento a nuestra estimación de 16,870, se llega a un total de 20,244. Este valor se acerca a la predicción preliminar de Rzedowski (1978: 73) de que la flora mexicana podría sobre- pasar el valor de 20,000, así como a su predicción más reciente de 22,000 especies (Rzedowski, 1991). La información disponible, aún parcial, para el año de 1989, agrega ya otras 100 especies nue- vas para la ciencia, y la tendencia inmediata extra- uadro 2. Estimación de la relación entre el número de sinónimos y el número de nombres válidos con base en el análisis de diez familias (o grupos dentro de familias) de plantas. Sinónimos Nombres válidos Razón Familia* A (B) (A/B) Bignoniaceae 38 26 1.462 Bromeliaceae 198 150 1.320 Connaraceae 31 50 0.620 Flacourtiaceae 131 30 4.367 32 15 2.133 Lecythidaceae 46 25 1.840 Moraceae 40 1.350 Olacaceae 30 24 1.250 Solanaceae 7793 2297 3.393 Zingiberaceae 43 20 2.150 x + D.E. 1.989 + 1.118 * Las referencias para esta fuentes de informacíon se dan en la sección de Materiales y Métodos. 402 Annals of the Missouri Botanical Garden polable de nuestro sondeo (cf. Figs. 1 y 2) hace ensar que, efectivamente, el número sobrepasará las 20,000 especies. Sin embargo, nuestra estima- ción es considerablemente menor a la proyección hecha por Toledo (1988: 17) de que la flora mexi- cana “se estima en 30,000 especies de plantas vas- culares,” es decir, un 36 y un 48% adicional a las estimaciones mínimas 22,000 de Rzedowski (1991) y de 20,244 Aunque la tendencia del número de especies por describirse va en aumento (cf. Fig. 2), claramente, el grado de redundancia nomenclatural debe to- de este estudio, respectivamente. marse en consideración para tales proyecciones. Una resultados del Cuadro 2 revela que cualquier modificación del factor de corrección de la sinonimia, aun cuando pequeña, puede alterar marcadamente el número estimado. Es necesario enfatizar que nuestros re- sultados son totalmente dependientes de las fuentes de información nomenclatural la cual, al momento, dista de ser óptima o es limitada. Aún tomando en cuenta estas salvedades, es claro que la riqueza observac ión de nuestros florística del territorio mexicano es notable y so- brepasa las expectativas iniciales (véase revisión de Rzedowski, 1978: 73) y sobrepasa la de otros país- es de mayor extensión como los Estados Unidos de Norteamérica, Canadá o (Rzedowski, 1991). La India, con una extensión mayor que México (3.3 vs. 2.0 millones km?) con una posición latitudinal similar, con gran variedad la ex-Unión Soviética de ambientes (selvas a vegetación alpina), y para a cual existe una enumeración floristica reciente, suma tan sólo 15,000 (S. K. Jain, com. pers. 1990). Toledo (1988) cita otras comparaciones que igual- mente subrayan la enorme riqueza florística de México. La definición de la magnitud de la diversidad botánica de una región y finalmente del planeta, es de gran importancia académica y aplicativa (Wil- 1988a, b, c). En el caso de México, la esti- mación de unas 17,000 especies nombradas hasta 1988 y la tendencia creciente señalada, sugieren son, que la flora de este territorio bien podría acercarse a contribuir con ca. 7% de la flora total del planeta (i.e., considerando que hasta ahora se han descrito un poco menos de 250,000 especies de plantas (Wilson, 1988a)). Desde luego, contra este notable potencial de contribución florística a nivel plane- tario, habrá que contraponer la tremenda alteración de los hábitats y potencial de extinción de plantas en México (véase, por ejemplo, Dirzo & García, 1990; Flores-Villela & Geréz, 1988; Rzedowski & Calderón de Rzedowski, 1987). Una vez más, este estudio hace resaltar lo obvio: la investigaci6n tax- onómica es de importancia medular y aún tiene mucho por hacer pero, además, es éste un quehacer científico en clara carrera contra el acelerado reloj de la deforestación y la extinción biológica. Literatura Citada — ^ -— Bárcena, Flora de Veracruz, Fascículo 15: Clethraceae. Instituto i ional de Investigaciones sobre Recursos Bióticos, Xalapa, México. Berg. C. C. 1972. Flora Pi a, Monograph No. Olmedieae, oe (Mora . Hafner, New Yor D Bravo-Hollis 1978. Las Cactáceas de Meses Volú- men I, E Nac ional Autónoma de México, México. & H. Sánchez-Mejorada. 1991. Las c actáceas de n ei is II. Univ. Nacional Autónoma de Méxi- B 2; Breedlove, D. 1986. Listados Florísticos de México IV. Flora de € i ad pi de Biología, Univ. Nacional Autónoma de M éxi Y Arcy G 1990. The Solanaceae since 1976 with a review of its Taie de Pp. 75-137 en J. G. Hawkes, da Lester e & N. Estrada (editores), Solanaceae e totes -C ae Evolution (Proceedings of the Tin International Solanan ‘eae Conference). Royal Bo- ical Gardens, Kew Dirzo, R. 1994 Diversidad de la Flora Mexicana. Agru- pación Sierra | Mac éxico. —— QU | Ca fa. 1990. Los Tuxtlas, a P4 ‘al area in southwest Conserv. Biol. 6: 84—9( Dodson, C. H. & A. H. Cons tion in l poaren Ecuador. Ann. 273-: ^ 20. ` . 6G. Rates of deforestation in México. 1991. Biological e Missouri Bot. Gard. 7 Erwin, T L. 1983. Beetles sind oe insects of tropic val forest canopies at Mani a fogging. Pp. halve k (editores) Tropical Rain Forest Ecol- ogy a Management Blackwell, Edinburgh. Fay, J. J. . Flora de Veracruz, Fascíc us 13: Nycta- ginaceae. po tituto Nacional de Investigaciones sobre Recursos En Xalapa. México. Flores-Villela, & P. Gérez. 1988. Ca nservación en México. iene Conse ral io a Méx Forero, E. 1983a. Flora de Veracruz, Fascículo 25: Con- naraceae. Instituto Nac ona de Invest iones sobre Recursos Bióticos, Xalapa, Méxic 19831 Hora Neotropic a, Tana No. 36: onnaraceae. er, New York. Fryxell, P. 199 yos : nir aru 7 ne California Academy of nces, $ ;entry, A. H. 1980. Do Neotropica, Temper No. 25: MD ia Hafner, ork. — ———,. 1982. Flora de Veracruz, Fascículo 24: Bi ituto Nac m HM ~ — igno- ene sobre 1-12. G. K. Hall, usetts. 978. Supplements 1-2. G. K. Hall, Boston, Massac us Ibarra, M. G. & S. Sinaca. 1987. Listados ea de México VII. “Estaci ¡ón de Biología Tropical Los Tuxtlas, Veracruz. eee de Biología, Univ. Nac = ra ‘0, México. y Doe , R. Guzman & B. Paz doi diploperennis (Gramir iae): A xico. Science 203 6-188 1979. 1eW teosinte from KE e Kewensis. 1895. An Enumeration of the Genera Volume 83, Number 3 1996 Dirzo & Gomez 403 Investigación Taxonómica en México and Species of Flowering Plants, Volumes I & II. Clar- d. Flora de Veracruz, Fascfculo 11: Can- naceae. fngubute Nacional de Investigaciones sobre Re- cursos Bióticos, Xalapa, México. Johnston, I. M. 1924. Expedition of the California Acad- emy of Sciences to the Gulf of California in 1921. The botany = vascular plants). Proc. Calif. Acad. Sci. 12: 951-12 Kubitzki, E &S. Renner. 1982. Flora Neotropica, Mono- graph No. 31: Lauraceae 1. Hafner, New York. t, A. & A. Novelo. 1986. Listados orfsticos de México Y prp nies Acuáticas de México. Instituto de Biología, Univ. Nacional Autonóma d México, México. Ludlow- Wiechevs, B. 1978. Flora de Veracruz, Fascfculo 3: Chloranthaceae. Instituto Nacional de Investiga- ciones sobre Recursos Bióticos, Xalapa, México. Maas, P. J. M. 1977. Flora Neotropica, Monograph No. 18: Renealmia —— Hafner. r Márquez- m . M. Engleman, A. Mart E. Martínez & C. H. Ramos. 1989. percha MINE: uctiva "M oon ndonia wm (Lacandoniaceae). Ann. Missouri Bot. Gar : 124-127. artínez, & H. E 1989. Lacandoniaceae (Triuridales): Una nueva familia de México. Ann. Mis- souri Bot. Gard. 76: 128-135 Martínez-Ojeda, E. 1982. Flora de Veracruz. Fascículo 22. Papaveraceae. Instituto Nacional de Investigaciones sobre Recursos Bióticos, Xalapa, México McVaugh, R. 1984. Compositae. En: W. R. Anderson (ed- itor), Flora Novo-Galiciana, Vol. 12. Univ. Michigan Press, Ann Arbor. 1987. Leguminosae. En: W. R. Anderson a tor), Flora Novo-Galiciana, Vol. 5. Univ. Michig Press, Ann Arbor. Moreno, P. 1980. Flora de Veracruz, Fascículo 10: Cari- caceae. Instituta DNE de Investigaciones sobre Re- pee México. 983. Firs de Veracruz, Miel in M ide Flores, H. V. 1 cursos Hidticon Xa, México. - a de Veracruz, Fascf Ebenaceae. Instituto Nacional de Investigaciones sobre Recursos Bióticos, e. México. ulo 16: Palacios, M e Veracruz, Fascículo 61: Mar- attiaceae. Instituto Nacional bid E tigaciones sobre Recurs 3 nun Xalapa, Flora Neotropica, ce, G. T. & S. A. Mori. 9. Monograph No. 21: Bi thidacess Part L Hafner, New or Raven, P H. 1987. We're killing our world: The global ecósystem in Apa McArthur Found. Occas. Pap. & G. B. Johnson. 1986. ees Times Mirror/ Mosby College Publishing, St. Loui Riba, R. 1981. Flora de Veracruz, Fascí ‘fculo 17: Cyathea- gámica de México. Acta B Instituto de Ec n Mex. TUN UE -& LEN ceae. Instituto Nacional de Investigaciones sobre Re- cursos Bióticos, Xalapa, México. Rzedowski, J. 1978. Vegetación de México. Editorial Li- 1 Un siglo de la botánica en México. Bol. Soc. Bot p -14. 991. Diversidad y M de la flora fanero- . Mex. 14: 3-21. alderón a Rardowski 1985. Flora faner- ogámica del Valle de . Vol. II Dicotyledoneae. El bosque tropical ers vas a región mexicana det ùl Bajío. Trace -21. 20, Editores. 1992. éxico Ante i . Mociño. n 897. Flora Mexicana. Imprenta Tale io Escalante, México 887-1891. Plantae Novae Hispaniae. Imprenta ldbacip Ee Méx Sleumer, H. O. 1980. Fs AS Monograph No. Z2: i ve Hafner, N ork. . Flora Neotropica, Monograph No. 38: Ola- caceae. ao New York. Smith, H. N. 1942. The pu ation dates of “La Natur- "B. : 1974. Flora Neotropica, Monograph No. 14: Pitcaimioideae (Bromeliaceae). Hafner, New York. 977. Flora Neotropica, Monograph No. 14, Part 2: Tillandsioideae (Bromeliaceae). Hafner, New York. . 1979. Flora Neotropica, Monograph No. 14, Part 3: Bromelioideae (Bromeliaceae). Hafner, ora de Veracruz, Fascículo 8: T nal » Investigaciones sobre R alada, Standley, P. C. 1920-1926. Contr. . Natl. Herb. 23. Taylor, K. R. 1983. Flora de Veracruz, Fascículo 30: Mar- i ncm Instituto Nacional a n iones sobre ie Xalapa, Méx E Cabrera. 1987. pom salina de México VI. E. de La Isla de Cozumel. Nak tituto de Bi last Univ. Nacional Autonóma de México, "rà x Toledo, V. M. 1988. La diversidad biológica de México. encia y Desarrollo 81: 17-30. Trelease, | 19 The American Oaks. Acad. Sci. XX. Wilson, E. o. 1988a. Systematic botany—A key science de tropical E 'h and documentation. Symb. Bot. psal. M xico. e and shrubs of Mexico. Mem. Natl. Ma E us National Academy Press, 1 plv D.C 1988c. The « current state of biologial diversity. Pp. 3-18 en E. O. Wilson edd Biodiversity. National Academy Press, Washington, D.C DIVERSITY OF THE FLORA OF FAN SI PAN, THE HIGHEST MOUNTAIN IN VIETNAM! Nguyen Nghia Thin? and Daniel K. Harder’ ABSTRACT si Pan, situated in ee ie near the border with China and reac hing a an altitude of 3143 m, is the l highest mountain in base: s flor subtropical and temperate components. | also su[ extensions of the old. E ropic a flora of the southern Yunnan Province of China. by the following: Aceraceae (Acer); Cupressac dendron, Vaccinium); Faga ae (Castanea, Castanopsi dia, Juglans, Platycar: ceae (Fokienia hodginsii (Dumm) Henr s, Fagus, ya); Pea (Beilschmiedia, Machilus, Phoebe, Neolitsea); Magnoliaceae (L IDE Mag- nolia, Manglietia, Michelia); and Pinaceae (Abies nukiangensis Cheng & L. K. The flora is diverse and cha & Thomas); Ericaceae c (Rhodo- Lithocarpus, Quercus); Juglandaceae (Carya, Engelhar- Fu, Tsuga chinensis (Franch.) Pritz). Fan si Pan, located at 22°09’—23°30'S, 103*59'E and reaching an altitude of 3143 m, the highest point in Vietnam. It is situated in the + S mountainous northwest province of Láo Cai. Fan si Pan and its contiguous mountain range are oriented along the Red River in roughly a northwest to southeast direction and extend into the Yunnan Province of China and to the Himalayan chain to the northwest. This range of mountains is derived from rocks of gneiss and ancient granite. The cli- mate is humid or perhumid (76-96%) year-round with an average yearly rainfall of 2770 mm; the heaviest rains are concentrated in the months of July and August. The average temperature is about 15°C, cember and January are the coldest months, when with a range between —3°C and 20°C. De- snow can fall for 1-3 days each year. This paper presents a listing of plant species from this diverse region of Vietnam, discusses the main vegetation types occurring there, and relates the Sino-Himalayan affinities of specific taxa to past climatic shifts believed to have occurred in this region. FLORISTIC DIVERSITY On the basis of published works by Lecomte (1907-1951), Vo Van Chi (1975), Aubreville et al. 103°— (1960-1983), Ke et al. (1969-1976), Ho (1970— 1972, 1901-1993) and Loc (1984), from our preliminary investigations (1991-1992), and results the flora of Fan si Pan is reported and arranged according to the Brummitt (1992) system, including 1750 species belonging to 680 genera in 210 fam- ilies of 7 divisions as presented in Table 1 Table 2 shows that in only 22 families, 746 spe- cies, or 43% of the total species found on Fan si Pan, are represented. There are several widespread and well-known families with abundant taxa rep- resented in the flora. The most significant of these include the Orchidaceae (26 genera and 62 spe- cies), Asteraceae (36 genera and 59 species), Eri- caceae (6 genera and 58 species), Poaceae (30 gen- era and 47 species), and Cyperaceae (6 genera and 42 species). In the Fan si Pan flora there are also many spe- cies-rich genera with a large number of taxa. These genera are listed in Table 3 Table 3 shows that in only 26 genera, repre- senting 3.8% of the 680 known genera in the re- gion, the 388 species comprise nearly 22% of the total in the flora, suggesting that the region around Fan si Pan and the local conditions have selectively encouraged diversification in several genera. A diversity of morphologies are found in the flora ! The authors thank Vo the Royal Botanic Gardens, Kew, U.K table comments and corre Scientific Research provided by e Viet namese Gove an Chi, College of Pharmacy, Ho Chi Minh City, Vietnam, and G. T. Prance, Director of + for providing valuable references. We also thank Peter Raven, Director of the , for receiving and transmitting our manuscript to the Annals of the O Botanical ctions to our manuscript provided by H. van der woule Ls to recognize the support for this investigation by the bonn of Basic ernment. editor, ? Herbarium, Department of b nun National University, 90 Nguyen Trai Road, Dong Da, Hanoi, 10,000, Vietnam. * Missouri Botanical Garden, ANN. Missouni Bor. P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A GARD. 83: 404—408. 1996. Volume 83, Number 3 Thin & Harder Diversity of a Vietnamese Flora Table 1. Diversity of plant divisions represented in the flora of Fan si Pan Table 3. Most speciose genera in the flora of Fan si Pan ranked by total number of species. Division Species Genera Families Bryophyta 180 40 25 Psilotophyta 1 1 1 Lycopodiophyta 20 2 2 Equisetophyta 2 1 1 Polypodiophyta 300 86 27 Gymnospermae 10 8 6 Angiosperma 1237 542 148 Dicotyledon 1022 443 131 Monocotyledons 215 99 17 Totals 1750 680 210 as well. Widespread forest genera with large trunks are found, such as Beilschmiedia, Phoebe, Machilus (Lauraceae); Eberhardtia (Sapotaceae); Castanopsis, Lithocarpus (Fagaceae); Schima (Theaceae); Altin- gia (Hamamelidaceae); and Fokienia (Cupressa- ceae). Different from typical tropical trees from hot, humid forest conditions, species of these woody plants from Fan si Pan lack buttresses. Woody li- anas are for the most part absent from the flora, yet vines are common. Besides mosses, which are very common, the most common epiphytes found in the Fan si Pan Table 2. si Pan ranked by species. Most species-rich families in the flora of Fan Number of Genus Family species Rhododendron Ericaceae 40 Rubus Rosae 36 Carex Cyperaceae 36 ic orac 23 Acer Aceraceae 19 Symplocos Symplocaceae 16 Asplenium Aspleniaceae 15 iplaziu Woodiaceae 15 Lithocarpus Fagaceae 13 aris Orchidaceae 12 Vaccinium Ericaceae 12 a Violaceae 12 Castanopsis Fagaceae 11 urya Theaceae 11 Pilea Piperaceae 11 Ardisia Myrsinaceae 10 Begonia Begoniaceae 10 Dryopteris Dryopteridaceae 10 Lepisor Polypodia 10 Melios Sabiaceae 10 Polygonum Polygonaceae 10 Pteris Adiantaceae 10 Michelia Magnoliaceae 9 Clemati. Ranunculaceae 9 Querc Fagaceae 9 Tetrastigma Vitaceae 9 otal 388 Family Genera Species Rosaceae 12 80 Orchidaceae 26 62 Asteraceae 36 59 Ericaceae 6 58 Poaceae 30 47 Cyperaceae 6 42 Lamiac 23 42 Rubiaceae 12 38 Urticaceae 10 37 Fagaceae 5 33 Moraceae 4 28 Theace 10 27 Gesneriaceae 11 23 Myrsinaceae 4 23 Fabaceae 10 21 Lauraceae 8 21 Acanthace 9 20 Commelinaceae 10 20 Aceraceae 1 19 Symplocaceae l 16 Magnoliaceae 4 15 Melastomataceae 10 15 To 248 746 flora are in the following families: Acanthaceae: Staurogyne petelotii R. Ben.; Ericaceae: including Rhododendron poilanei Dop, Vaccinium poilanei Dop; Gesneriaceae: Lysionotus petelotii Pell., Ly- sionotus spp., and Aeschinanthus; Loranthaceae: Loranthus parasiticus Druce; Orchidaceae: Liparis spp., Dendrobium spp., and Aerides spp. Distinctive from other typically tropical and subtropical floras are epiphytic species in the Ericaceae, Acantha- ceae, and Gesneriaceae. The herbaceous lant composition varies with Acanthaceae, Aristolochiaceae, Begoniaceae, Ges- neriaceae, Myrsinaceae, Piperaceae, and Urtica- ceae. Under the dense forest canopy in deep shade the few plants that can survive are those that do not depend on light for growth but, rather, obtain nutrients from decaying plant matter, such as the saprophytic Balanophora pierrei Van Tiegh. It is worth noting here that taxa of Aristolochiaceae, Be- goniaceae, Gesneriaceae, and Urticaceae are very 406 Annals of the Missouri Botanical Garden common in the flora of Fan si Pan, while some fam- ilies that could be expected here because they are common in many other tropical forests are mostly absent, such as taxa in the Araceae and Euphor- biaceae. A great number of species occur at Fan si Pan that have attractive flowers and leaves or interesting leaf shapes, particularly species of Ericaceae: Rho- dodendron (40 species), with various-colored flow- ers; Aceraceae: Acer (19 species); Hippocastana- ceae: Aesculus wangii Hu; Rosaceae: Sorbus, with leaves becoming red in winter; Trilliaceae: Paris (3 species), with attractive flowers and leaves. Many of these species could be valuable to horticulture as subtropical and temperate ornamentals. DIVERSITY OF FLORISTIC COMPONENTS The flora of Fan si Pan is also interesting for its apparently high endemism. Preliminary estimates from the available information accumulated for this study indicate that up to 30% endemism occurs in the flora. For example, the Orchidaceae have 19 endemic species reported for northern Vietnam; 18 of them can be found at Fan si Pan. The genus Carex of the Cyperaceae has 7 species endemic to northern Vietnam: 6 of these have been collected at Fan si Pan. According to Vo Van Chi (1975), 25 genera and 39 species were discovered and initially described from collection records from Fan si Pan. The Fan si Pan flora supports tropical, subtrop- ical, and temperate elements. TROPICAL ELEMENT Most of the tropical elements in the Fan si Pan flora are distributed well below an altitude of 1500 m, yet some extend as high as 2000 m. These taxa are members of the Annonaceae, Araceae, Clusi- aceae, Combretaceae, Cyatheaceae, woody Euphor- biaceae, Gnetaceae, Hernandiaceae, Icacinaceae. Polygonaceae, and Proteaceae. Such tropical gen- era and species as Actinodaphne, Aleurites, Arto- carpus, Calamus, Caryota, Citrus, Duabanga, and Wendlandia are relatively common elements of the flora and constitute a significant component of the forest. This is true not only of the forest trees but also of the tropical understory shrubs and herbs that are common, such as species of Acorus, Anotis, Balanophora, Calamus, Dichroa, Dioscorea, Dios- pyros, Disporum, Macaranga, Mallotus, Pasania, Sarcandra, Vernicia, and Wendlandia. SUBTROPICAL ELEMENT The subtropical component of the flora of Fan si Pan possibly represents a Tertiary floristic element from North Vietnam-South China suggested by the following families: Aceraceae (Acer with 19 spp.). Cyperaceae (Carex with 36 spp.). Fagaceae, Ges- neriaceae, Hamamelidaceae, Lauraceae, Magnoli- aceae, Myrsinaceae, Symplocaceae (Symplocos with 16 spp.). Theaceae, and Urticaceae. Most of the subtropical species found at Fan si Pan are located near 1600 m elevation, which is characterized by an average temperature of 15.7°C and nearly 2732- 2778 mm of precipitation per year. Characteristics of the elevation, climate. and geomorphology of this region combine to maintain this subtropical ele- ment of the flora. Subtropical taxa are represented by the following genera and species: Hydrangea (Hydrangeaceae); Embelia (Myrsinaceae); Polygonum capitatum F. Ham. ex D. Don, P. flaccidum (Meissn.) Steud., P. palmatum Dunn, P. thunbergii Sieb. & Zucc. (Po- lygonaceae); Anemone (Ranunculaceae); Huoden- dron, Rehderodendron (Styracaceae); Crawfurdia, Gentiana (Gentianaceae); Schisandra, Kadsura (Schisandraceae); Buddleja, Fagraea (Loganiace- ae); Cornus (Cornaceae); Allomorpha, Anerincleistus, Bredia, Fordiophyton, Oxyspora, Plagiopetalum, Sarcopyramis (Melastomataceae); Viola (Violaceae); Adinandra, Anneslea, Gordonia, Pyrenaria (Thea- ceae); Exbucklandia, Rhodoleia (Hamamelidaceae); Liriodendron, Manglietia, Magnolia (Magnoliaceae); and Sorbus, Potentilla (Rosaceae). TEMPERATE ELEMENT At altitudes above 2000 m, the Fan si Pan flora contains temperate floristic elements characterized by Paris (Trilliaceae); Alnus, Betula (Betulaceae): Enkyanthus, Leucothoea, Pieris, Rhododendron, Vac- cinium (Ericaceae); Celtis, Ulmus (Ulmaceae); Cas- tanea, Fagus, deciduous Quercus (Fagaceae); Aes- culus (Hippocastanaceae); Juglans, Platycarya (Juglandaceae); Crawfurdia, Gentiana (Gentianaceae); Tsuga (Gymnospermae); Huperzia spp. (Lycopodiophyta); Coptis, Ranunculus (Ranunculaceae); Panax (Araliaceae); Hypericum (Clusiaceae); Salix Abies, (Salicaceae). These genera are speciose, and many are quite frequent and widespread in this region, including, in particular, Rhododendron with 40 spe- cies, Vaccinium with 12, and Quercus with 9. DIVERSITY OF VEGETATION TYPES Although the diversity of vegetation types occur- ring at Fan si Pan is not yet fully understood, we can preliminarily report on the main types encoun- tered. Their distributions appear to be associated with elevation. Volume 83, Number 3 1996 Thin & Harder 407 Diversity of a Vietnamese Flora 1. TROPICAL VEGETATION BELT The tropical vegetation belt formerly contained forest vegetation that has mostly been destroyed by shifting cultivation, overexploitation, and forest res. It has subsequently been replaced by second- ary forest along the valleys and by savanna on the slopes in areas not converted to cassava, rice, or maize cultivation. 1.1. Secondary forest type. This derived vege- tation type is extensive and distributed along the many river and stream valleys. It is characterized by species of Lauraceae, Fagaceae, Meliaceae, Sap- indaceae, Fabaceae (including Mimosaceae and Caesalpiniaceae), Magnoliaceae, and Burseraceae. 1.2. Savanna. The savanna vegetation type is also extensive and is derived spontaneously after cultivation is abandoned in a shifting agricultural system. It is characterized by the grasses Miscan- thus floridulus (Labill.) Warb. ex K. Schumann & Lauterb. and Saccharum arundinaceum Retz., with some invasive and widespread trees; Mallotus pan- iculatus (Lam.) Muell.Arg., Macaranga denticulata Muell. Arg., Trema orientalis (L.) Blume, with Musa spp. in wetter places, and Imperata cylindrica (L.) Raeusch., Cratoxylon polyanthum Korth., Rhodo- myrtus tomentosa (Aiton) Hassk., and Melastoma candidum D. Don in drier places. 2. SUBTROPICAL VEGETATION BELT This subtropical zone can be delimited roughly between 1000 and 2000 m elevation. The temper- ature between these altitudes is extremely stable with a range between 15° and 17°C. Many of the subtropical taxa occurring in this zone have affin- ities with the Tertiary flora of northern Vietnam and southern China. 2.1. Dense evergreen forest. Dense evergreen forest was once widespread in this zone; however, due to extensive overexploitation and shifting cul- tivation, it can now be found only in isolated and inaccessible steep valleys and on steep slopes. The characteristic families of the forest trees are rep- resentatives of the Betulaceae, Fagaceae, Hama- melidaceae, Lauraceae, Magnoliaceae, Sapotaceae, and Theaceae 2.2. Subtropical savanna. The derived sub- tropical savanna is a common type of vegetation found at these altitudes at present. It is secondary, arising after the destruction of subtropical forest most commonly by overexploitation or shifting cul- tivation. This widespread vegetation type is char- acterized by herbaceous species such as the grasses Arundinella nepalensis Trin., Imperata cylindrica, Microstegium sp., and Miscanthus floridulus; shrub species including Buddleja spp., Clematis leschen- aultiana . Litsea cubeba (Lour.) Pers., Luculia intermedia Hutchinson, Osbeckia crinita Benth., Oxyspora paniculata (D. Don) DC., Polygonum chi- nense L. var. scabrum, P. paniculatum Andrz., Por- ana racemosa Roxb., Rubus ellipticus Smith, and Viburnum cylindricum Buch.-Ham. ex Don; and some trees, including Acer campbellii Hook. & Thoms. ex Hiern., Alnus nepalensis D. Don, Itoa orientalis Platycarya kwangtungensis un, Sauravia bir DC., and Wightia spe- ciosissima (D. Don) Merri 2.3. Subtropical grassland. Subtropical grass- land is also derived after shifting cultivation, tram- pling and grazing by a and forest fire. Species of Poaceae, Cyperaceae, Fabaceae, and Pteridium, and including ud shrubby species of Melastoma, Osbeckia, Rubus, and Artemisia, pre- dominate. 3. TEMPERATE VEGETATION BELT Temperate vegetation is found at altitudes over 2000 m and supports temperate species indicated by species in such genera as Abies, Acer, Adinan- dra, Aesculus, Agapetes, Alnus, Altingia, Coptis, Cornus, Crawfurdia, Embelia, Enkyanthus, Fagus, Fokienia, Hydrangea, Huodendron, Liriodendron, Magnolia, Oxyspora, Primula, Quercus, Rehdero- dendron, Rhododendron, Rhoiptelea, Sorbus, Tern- stroemia, Vaccinium, and Valeriana, among others. 3.1. Temperate forest. Temperate forest vege- tation composed of the above-mentioned genera is found in the deep valleys and on steep slopes of Fan si Pan. In general, this vegetation is relatively widespread at high altitudes, where it is dominated by species in the Aceraceae, Hippocastanaceae, Fagaceae, Magnoliaceae, Lauraceae, Cupressaceae, Podocarpaceae, Pinaceae, and Taxaceae. 3.2. Montane cold savanna. This very local- ized vegetation type occurs on the slopes of the highest summits of Fan si Pan, which are believed to have been originally covered by forest. Again, conversion of the original forest to its present veg- etation type was a result of human cultivation and burning of these areas. The montane, cold savanna supports many species of shrubs and herbaceous species belonging to the Poaceae, Cyperaceae, Lil- iaceae, Hypoxidaceae, Zingiberaceae, Gentiana- ceae, Ericaceae, Rosaceae, Melastomataceae, Lami- aceae, Hypericaceae, Gesneriaceae, and Pteridophyta. 408 Annals of the Missouri Botanical Garden These species are light-loving and cold-enduring, with Arundinaria spp. as dominant. ARCHAIC CHARACTER OF THE FAN SI PAN FLORA Fan si Pan contains an archaic element in the flora represented by the number of monotypic fam- ilies or endemic genera that are present in the re- gion. Families such as Bretschnederaceae, Penta- phyllacaceae, Rhoipteleaceae, and Sargentodoxaceae are present at Fan si Pan. The high level of endem- ism, particularly for the genera and species in the Orchidaceae, and diversity in the genus Carex, also support the contention that this region represents a refugium for these limited distribution taxa. dia, Fokienia, Rhodoleia, Rhoiptelea, and Sargen- todoxa (Wu Cheng Yi, 1965) occur in this region. These genera represent extensions of floristic ele- ments from a more northern flora, particularly the subtropical flora of southern China and adjacent areas, and suggest that the more tropical climate existed across this area in the past. Fan si Pan appears to have served as a protected area for many of these taxa during glacial extensions across this region. Because of the high diversity in the flora, this has long been recognized as a center for an- giosperm radiation (Takhtajan, 1969). The interac- tion of climate, geomorphological features (includ- ing soil), and geology has supported an enrichment of the flora. Archaic characters of the geological structure of Vietnam support the presence of prim- itive vegetation and a high number of relictual spe- cies, which have occurred there for a long time and still survive today (Thai Van Trung, 1978 Literature Cited Aubreville, A., M. L. Tardieu-Blot € J. E. Vidal. 1960— 1983. e du Cambodue, du Laos et du Vietnam. Fas. 1-20. Pa Brummitt, R. K 1992. Vascular Plant Families and Gen- era. Royal Botanic Gardens, Kew Chi, V. V. 1975. Flora and Vegetation of Sapa (Lào Cai Province). Hanoi Univ. Ho, P. H. 1970-1972. An Illustrated Flora of South Vi- etnam. I-II. South Vietnamese Ministry of Education, Ho, P H. 1991-1993. An Illustrated Flora of Vietnam I- III. South Vietnamese. Printed by Mekong Printing, al. " al. 1969-1976. Common Flora in Vietnam. Sci. Techn. Publ. House, Hanoi. Lecomte, H. 1905-1954. Flore Général de l'Indo-Chine. I- VII. Masson Editeur, Paris. Loc, P. K. 1984. Species belonging to Pinopsida in the Vires en Biol. J. 6(4): 5-10. Takhtaja Flowering Plants: Origin and Dis- pera Taste by C. Jeffery. Oliver Be Boyd, Ed- inbur rung, T V 1978. Ms aaa oe Vietnam Forest. Ed. 2. Sci. Techn. Publ. Hous Wu, Cheng Yi. 1965. On Ew Tropie 'al Affinity of the Chi- nese Flora. Science Press, Beijing. Ie van der Werff? and G. Richter? TOWARD AN IMPROVED CLASSIFICATION OF LAURACEAE! ABSTRACT Published suprageneric classifications of Lauraceae and the characters used in these classifications are briefly concluded that androecial characters Cassytha, the other including all other genera. The latter group is divided into three tribes, ne Lau ureae, Perseeae, and Cryptocaryeae, based on characters of wood and bark anatomy and inflorescence structu Lauraceae form a large, predominantly tropical family of trees and shrubs, with the exception of Cassytha, an herbaceous parasite. The family is best represented in the American and Asian trop- ics, and has also a rather large number of species in Australia and Madagascar, but is poorly rep- resented in Africa. About 50 genera are currently recognized, with 2500-3000 species. Economically, Lauraceae are an important group. Many species yield high-quality timber, others spices or aromatic oils, and Persea amer- icana Miller is cultivated worldwide for its edible fruits. Ecologically, Lauraceae are, in the New World, a very important group. They are present in wet forest at any elevation (from sea level to páramos) and are frequently the most common or one of the most common tree families, especially in the foothills and at middle elevations of the Andes. In spite of their importance, Lauraceae are, in respect to classification and species numbers, poorly known. Our lack of knowledge of species numbers and distribution is no doubt related to the fact that many species are tall trees with small, inconspicuous flowers, difficult to locate and to collect. This is clearly shown by a recent floristic treatment (Australia: 115 species, of which 46 were new, Hyland, 1989), recent revi- sions (Nectandra: 114 species, of which 33 were new, Rohwer, 1993a; Pleurothyrium: 40 species, of which 20 were new, van der Werff, 1993), and the fact that in the most recent monograph of An- iba (Kubitzki, 1982) not a single collection was recorded from Ecuador, while currently 11 spe- cies are known from that country. More intensive collecting will hopefully correct this lack of knowledge. Lauraceae have, with a few exceptions, trim- erous flowers. Flowers are bisexual or unisexual. There are two whorls of three tepals; the whorls are usually equal in size and shape, but in some cases the whorls are unequal. If the whorls are unequal, the outer whorl is usually smaller than the inner one, although the reverse can also be the case. Flowers have four whorls of three sta- mens, but in most genera, one, two, or three as are reduced to staminodia. The anthers open by two or four valves. The ovary is generally superior, with one locule and one ovule, and the fruit, a one-seeded berry, sits either free on a pedicel, is partially enclosed by persistent tepals or the receptacle, or is entirely enclosed by the receptacle. CLASSIFICATION OF LAURACEAE Strictly speaking, there is no lack of suprage- neric classifications of Lauraceae. All have in common one characteristic: they are not widely accepted. We will present a brief review of these classifications and list the main characters used n making them. The position of Cassytha in the different classifications will not be discussed; it is always separated from the other Lauraceae be- cause of its herbaceous, parasitic habit, and we place it in its own subfamily, the Cassythoideae. ! John Myers assisted in the preparation of the figures. We thank Tom Wendt for critical comments on an earlier t. version of the manuscrip 2 Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. 3 Institut für Holzbiologie und Holzschutz, Bundesforschungsanstalt fiir Forst- und Holzwirtschaft, LeuschnerstraBe 91, 2050 Hamburg 80, Germ ANN. Missouni BOT. GARD. 83: 409—418. 1996. 410 Annals of th Missouri Am Garden Characters Used in Nees (1836) Classification 1) Leaves deciduous vs. evergreen. 2) Inflorescence umbellate or glomerate. a) Inflorescence umbellate, involucrate. b) Inflorescence glomerate or subracemose, arising from perulate buds. ) Inflorescence paniculate. ) Anthers opening apically. ) Anthers opening below tip, equal. ) Outer anthers petaloid. ) Anthers with distinct filaments. 4) Fruits covered by perianth tube. 4) Fruits not protected by perianth tube. 5) Staminodia lacking or, if present, without capitate apex 5) Staminodia with triangular head. 6) Tepals entirely persistent 7) in a solid cu 7) spreading, not thickened 6) Tepals largely deciduous 7) Truncate base,only, persistent. 7) Entirely deciduous. N U UY WwW C2 Figure l. Main characters used in Nees’s (1836) clas- sification. It has been suggested that Cassytha is closely related to Cryptocarya (Rohwer, 1993b); however, the main characters discussed in this paper (in- florescence structure and wood and bark anato- my) will not elucidate the relationsips of Cassy- tha. The classifications by Nees (1836; Fig. 1), Meissner (1864; Fig. 2), Bentham and Hooker (1880; Fig. 3), Pax (1889; Fig. 4), Mez (1889; Fig. 5), and Hutchinson (1964; Fig. 6) are all based on the following characters: inflorescence paniculate versus umbellate; number of anther cells (2 vs. 4); number of stamens; fruit enclosed in perianth versus seated in a cup or free; and owers unisexual or bisexual. These classifica- tions are strongly influenced by the choice of the most important character, and differences be- tween the classifications are a result of such choices and are not based on new or better data. For instance, Pax used 2- versus 4-celled anthers as the most important character, while Mez and Nees used inflorescence paniculate versus race- mose. None of these authors defends or explains his determination of the importance of the char- acters, and all classifications are in some aspects confusing. Several of these classifications include genera no longer recognized or which were based on faulty diagnoses, but such details are of his- torical interest only. Meissner 1) Suborder Laurine 2) Suborder Os ro from Lauraceae 3) Suborder Cassytheae, Cassy Laurineae: A. Inflorescence paniculate, racemose or spicate. No involucres. Tribus Perseaceae: F lowers hermaphrodite. Stamens 9. Cupule present or lacking. ed. amens we 4-celled, inner 3 extrorse; 7 genera eas ec eurothyrium b) Stamens free, 4-celled, al introrse; 2 genera (Sassafras, Sassafridiu c) Stamens fee, 2-celled, um 3 extrorse; 1 genus Goeppertia) d) xen fied flowers hermaphrodite; 2 genera ds Synandrodaphne) pci hrodite, fruits enclosed in calyx. ) Flowers pines . Adenostemum, excluded. 2) Flowers 3-merous, 2-celled. 10 genera, including: Cryptocarya, Aioue ea, Ampelodaphne 3) Flowers 3-merous, capitate, 2-celled; stamina monadelphic. Misantheca 4) Flowers 3-merous, 4-celled, stamens free. 4 genera B. Flowers late or glomerulate. Involucrum present. Tribus Litseaceae Subtribus T Tetranthereae. Anthers 4-celled; 5 genera Subtribus Daphnidieae. Anthers 2-celled; 5 genera igure 2. Main characters used in Meissner's (1864) classification. Kostermans (1957) published a new classifica- tion, in which he recognized five tribes (Fig. 7). One tribe was recognized by its involucrate inflores- cence, the other four non-involucrate tribes by the development or lack of cupules. One tribe was rec- ognized by a complete absence of a cupule (for ex- ample. Persea and Beilschmiedia), the second by the presence of a more or less cup-shaped cupule (Ocotea, almost completely enclosed by the cupule (Cryp- Nectandra), the third by having the fruit tocarya, for example), and the fourth by having a truly inferior ovary and the fruit entirely enclosed by the hypanthium (only Hypodaphnis). Further di- vision within the tribes is primarily based on num- ber of anther cells. In comparison with the contem- E (1964), Kostermans’s classification is clearly superior, not porary classification of Hutchinson because the characters used for the classification are sounder, but because he knew the Lauraceae well. Thus, he excluded a number of weak genera recognized by Hutchinson, and avoided errors that Hutchinson, less experienced with Lauraceae, made. Kostermans’s classification has found general acceptance during the last 30 years, although sev- Volume 83, Number 3 1996 van der Werff & Richter Classification of Lauraceae Bentham and Hooker (1880) 3 TRIBES 1) Perseaceae. Stamens of whorl III opening extrorse, with 2 basal glands, inflorescences lax, pedunculate. a) Anthers 2-celled. Fruit included in perianth. b) Anthers 2-celled. Fruit with / without cupule. subdivided by number of stamens. c) Anthers 4-celled.Fruit with/without cupule subdivided by number of stamens. 2) Litseaceae. Trees or shrubs. All stamens opening introrse. Inflorescence dense, short, subsessile (except Sassafridium). a) Inflorescence lax or imbricate - bracteate. b) Inflorescence umbellate or capitate, included in an involucre. Subdivided by number of anther cells. 3) Cassytheceae. Leafless vines. . Main characters in Bentham & Hooker's " s d a Stamens 9 Stamens III Extrorse x Stamens 3 Anthers 4-celled Stamens III Introrse Leafless--Cassytha Anthers 2-celled Leafy Stamens III Extrorse All stamens Introrse Stamens 3 Stamens 6 or 9 Receptacle shallow Receptacle deep cup-shaped Figure 4. Main characters used in Pax's (1889) clas- sification. Fig (1964) ahe Mez Herbaceous parasitic vine; inflorescence indeterminate m Shrubs or trees, inflorescence ania T Pues rd Inflorescences Ee exinvolucrate ..... erseeae rs of outer two whorls 2- he or mee 5 EL. Anthers of ae two whorls 4-celled ....... Inflorescences racemose, ssc more Litseeae Anthers 2-celled ....... Anthers 4-celled ....... Figure 5. Main characters used in Mez's (1889) clas- sification. eral workers have pointed out difficulties with ge- = circumscription and classification (Hyland, 1989; Rohwer et al., 1991; van der Werff, 1991). mit a (1981) published the results of his study of wood and bark anatomy of Lauraceae, in which he found three large groupings of genera (Fig. 8). One of the groups corresponds with the tribe including genera with involucrate inflorescences, but the oth- er two groups have no counterpart in the existing classifications. For instance, Richter placed Cryp- tocarya and Beilschmiedia in the same group, while in Kostermans's classification they occupy very dif- ferent positions. e most recent classification is by Rohwer (1993b). He recognized two main groups, based on inflorescence type, one involucrate and one exin- volucrate. Further divisions were based on fruit and floral characters, but because these characters were Hutchinson unisexual Anthers 4-celled Fl diae s owers perfect Inflorescence 2 or more flowers "d in involucre 1 flower in each involucre enclosed in bracts Anthers 2-celled ruit enclosed in ain be Anthers 2-celled Fruit not enclosed in calyx tube Inf] E enclosed in bracts All anthers introrse Anthers 4-celled ll Anthers III extrorse Main characters used in Hutchinson’s 412 Annals of the Missouri Botanical Garden mi S / ^ ^ Hypodaphnis \ Tribus Hypodaphneae Kosterm LI || V i ` 4 N Fd * e eo ENS a ~ 5 N Eusideroxylon N i oq l | Tribus Cryptocaryese Meissn. Potoxyion A = we ewe eee —T -- \ y = ~ ~. / E — Neolitsea i bow Pd wen _ a soe ae e _ e EN — — dedu a M N Pd -—7 / - Laurus Y P Dicypelliu fy Y” Phyllostemonodaphne Y Py Systemonodaphne de P Pleurothyrium Linders Pd / Clinostemon A Licaria Tribus Litseese Mez. 7 / -—T / Ocotea — oo” / / Nectandra Fi Cinnadenia Urbanodendron Aniba A - -— licReria 227 Tribus Cinnamomese Baill — = A ” / "OT A Ie inco = / TC Aiouea oJ / ... TENE Tribus Perseeae Mer. Figure 7. Classification of Kostermans (1957). Reprinted with permission used with some hesitation, no formal classification STRENGTH OF CHARACTERS USED IN PUBLISHED was proposed. Keys to genera were recently pub- CLASSIFICATIONS lished by van der Werff (1991; for genera of the New World) and Rohwer (1993b; for genera world- A robust classification demands that the char- wide). acters used are reliable: that is, there are no or few 413 van der Werff & Richter Classification of Lauraceae Volume 83, Number 3 1996 (3Y32NDILSIMAM) § auydepeaz! auydepoutioy 9v1jvsses vTi1e[n[Tequn ‘ded (ze 9s8ebepen) t" wÁ1e203dÁ12 “(T86T) 1914214 Jo uongogisse[)) uaadeiney] ur uasuniatddnay AIYISTUOJPUY :7¢ 'qqy ueqoaduepufy suyo /i (eputyu 1epo/pun ZTOH) 2unuwy3sufexsqy eTTetzszed errrrrrsrror. (ePurH) uxesej3seg euuo exeL === (ztoH) wXkyoueredpueg j3ru exey (d) e1spuTT NOM LLLI M . ———S a A ee Sem => SaS 99817 T0O9N “Te Y STSUDUTYI we wo vəs) "IM (Ww) vaapuga1 NL VA \ sniney ,.*" N 2 M wnuwoureuuT209N M /pezeaueacy * vÁ10203d415 . /qetaeneuy E > : E d E MS E cad Pree TTS rat Pr ad : ^ ananez ZA . t E “7 ; : : Vine : $ : y! . ` uo1puepoueqin : Y STRUAUTALOI . : 4 stedouydepodred 4 auydepovowa73 9 As . . A °t eesse UOTAXOI9PTSNZ . . wnj[TadA»jq i E um FIAYIOINETd tes. s Il $ / à wiJeuaTQpu3 «4 € 9 3 0 à o. i j/ e vanojv ^ Ho Pag A S ET n s Il k^ . . à PL S PE . A syuydepodAH se” vipuv32aN M LS X (IVIIWIONVNY 3H) 'g 310314 414 Annals of the Missouri Botanical Garden Table 1. Genera with variation in number of anther cells. Aiouea normally nine 2-celled, rarely six or three 2-celled Aniba normally nine 2-celled, rarely six 2-celled Aspidostemon either six 2-celled or three 2-cellec Beilschmiedia ormally nine 2-celled, rarely six 2-celled or nine 1-celled Ca wy 'odaphnopsis Cassytha 2: 'innamomum ?^otameia n normally nine 4-celled, rarely nine 2-celled or six 2-celled normally nine 2-celled, rarely six 2-celle normally nine 4-celled, rarely nine 2-celled normally three 2-celled, rarely six 2-celled ncn nine 2-celled, rarely six 2-celled + ~ 4-celled nine 2-celled or six 4-celled + three 2-ce BD nine 4-celled, rarely nine 2- clado or six 4-celled three 2-celled normally four 2-celled, rarely four 1-celled or two 1-celled Urbanodendron normally nine 4-celled, rarely nine 2-celle exceptions to the conditions characteristic for a giv- en taxon. An analysis of the characters used most frequently in the generic and suprageneric classi- fications of Lauraceae will allow us to estimate how well these taxa are founded. One frequently used character refers to the in- florescence. It is phrased in slightly different ways Nees (1836) and Meissner (1864) contrasted umbellate versus pa- in the various classifications. niculate inflorescence, with and without an invo- lucrum; Hutchinson (1964) stressed the presence or absence of bracts; Kostermans (1957), decussate bracts; and Rohwer (1993b) mentioned “some kind of involucre.” Based on the senior author's experi- ence, the character states of involucrate, racemose versus exinvolucrate, paniculate inflorescence are reliable; we know of no genera in which both kinds of inflorescence are represented, and we accept the inflorescence differences as reliable generic char- acters. The importance of inflorescence types in the classification of Lauraceae will be discussed further in this article. The next set of frequently used characters are those of the androecium, i.e., the number of fertile Table 2 related species or species groups are placed in different “Genus” pairs in which apparently closely genera due to generic circumscription by anther cell num- ber 4-celled 2-celled Cinnamomum Aiouea Cinnamomum Temmodaphne Aiouea Ocotea cotea Endlicheria Rhodostemonodaphne Endlicheria Williamodendron Mezilaurus Litsea Lindera Parasassafras Sinosassafras stamens and number of anther cells on each sta- men. Possible variation of these characters can, of course, best be studied in genera defined by some- thing other than these androecial characters. This variation is considerable (Table 1). For instance, among the neotropical species of Caryodaphnopsis, defined by having opposite leaves and unequal te- pals, are species with nine 4-celled stamens, nine 2-celled stamens, and six 4-celled stamens plus three staminodia. Likewise, most species of Pota- meia, defined by having dimerous flowers, have four 2-celled stamens; a few have four 1-celled stamens and one species, as yet undescribed, has two l- celled stamens. Neotropical species placed in Per- sea mostly have nine 4-celled stamens, but some have nine 2-celled stamens or six 4-celled and three 2-celled or six 4-celled and three staminodia. Looking at the genera defined by 2-celled stamens, the similarity between most Aiouea species (2- celled) and Cinnamomum (4-celled) is striking and seems more than convergence; however, Aiouea vexatrix van der Werff is very similar to some sym- patric Ocotea species, as are A. lundelliana Allen and A. costaricensis (Mez) Kostermans (van der Werff, 1987a, 1988; Rohwer et al., 1991). situation is found in Endlicheria (two-celled). Some of its species are strikingly similar to Rhodoste- similar monodaphne or Ocotea species (Rohwer et al., 991). A third generic pair is formed by M (2-celled) and Williamodendron (4-celled); species of Williamodendron were initially described as Me- Werff, 1987), quently recognized as a distinct genus (Kubitzki & Richter, 1987). A few other examples are presented | Table 2. On the other hand, celled genera that do not have a 4-celled counter- zilaurus (van der but were subse- there are also 2- part, such as Cryptocarya, Beilschmiedia, Aniba, and Licaria. These examples indicate that the an- droecial characters often vary within genera and are Volume 83, Number 3 van der Werff & Richter 415 Classification of Lauraceae unreliable characters in classification at generic and higher levels. The only character of the gynoecium frequently used is the degree to which the fruit is covered by the hypanthium—from no cover and fully exposed to a completely inferior ovary and the fruit fused with the hypanthium. In most genera this character is con- stant; exceptions occur in Ocotea, which includes spe- cies with pronounced cup-shaped cupules and spe- cies with very small, platelike cupules. In general though, the gynoecium character promises to be use- ful in generic and suprageneric classifications, be- cause of its constancy at the generic level. Earlier classifications were attempts to order the taxa being studied and were, in fact, frequently keys translated into a hierarchical system. A clas- sification was a system enabling one to make iden- tifications, and if that goal was met, the classifica- tion was acceptable. ore recently, the idea that classifications should reflect relationships and evolution of the taxon to be classified has found wide acceptance. Whether or not a phylogenetic classification is helpful in the identification process is less impor- tant. It is important to be aware of the dual pur- pose of a classification—on the one hand a path to identification, on the other a reflection of the phylogeny. For purposes of identification, the an- droecial characters are very useful because they are well defined and readily observed. On the oth- er hand, characters such as number of stamens or number of anther cells are variable in several gen- era. This can only be observed in genera that can be defined by other characters. For instance, Car- yodaphnopsis can be recognized by having oppo- site leaves and strongly unequal tepals; it also has very distinct wood anatomical characters. Within Caryodaphnopsis are species with nine 4-celled, nine 2-celled, and six 2-celled stamens. Other genera, for example, Ocotea, lack non-androecial characters (Ocotea is defined by having nine 4- celled stamens, with the cells in two horizontal rows), and species that resemble Ocotea very closely, but with 2-celled instead of 4-celled sta- mens, are placed in me genera (van der Werff, 1988; Rohwer et al., Caryodaphnopsis, whose species share several non-androecial characters, can be expected to be zenera such as monophyletic, but genera such as Ocotea, whose species only share androecial characters, are not likely to be monophyletic. Problems with classi- fication of Lauraceae exist at two levels: there is a need for better defined, monophyletic genera, and a need for a phylogenetic classification at the suprageneric level. The focus of the rest of this paper is a more natural classification at the su- prageneric level; it is acknowledged that this will not immediately lead to monophyletic genera, or to easier identifications. OUTLOOK FOR A PHYLOGENETIC CLASSIFICATION As has been discussed, the existing classifica- tions are largely based on floral characters. The an- droecial characters vary frequently within genera and are therefore a poor choice as main characters for a generic and suprageneric classification. The gynoecial character emphasized by Kostermans does not vary within genera and holds more prom- ise. However, the classification based on this char- acter differs greatly from the generic groupings us- ing wood and bark anatomy. It seems unlikely that a thorough reexamination of floral and fruit characters will yield data with which a more robust classification can be constructed. In- corporating new data sets in building a classification looks like a more promising approach. Such an ap- proach requires extensive collaboration between par- ticipating specialists. A few years ago, such a project was proposed and initiated by B. Hyland and the se- nior author, and will incorporate data from DNA stud- ies, wood and bark anatomy, leaf oils, leaf venation and leaf cuticles, fruit anatomy, pollen, inflorescence types, and the traditional flower and fruit morphology into a new classification. The published results of the study of wood and bark anatomy by the junior author (Richter, 1981), and the senior author's observations of inflores- cence structures, both indicate that the Lauraceae are divided into three groups of genera. Wood and bark features employed are of an exclusively qual- itative nature, quantitative characters being exclud- ed as less reliable for their intrinsically high vari- ation. They were selected and weighted in a function of their diagnostic value (identification) and discriminatory power (classification) within the specific context of Lauraceae. The set of secondary xylem characters includes primarily those relating to axial parenchyma distribution, fiber morphology, inorganic compounds, and vessel morphology. Sec- ondary phloem characters considered as highly di- agnostic and discriminating refer mainly to me- chanical tissues, i.e., presence versus absence and morphology of phloem fibers and sclereids. These features were employed both in the positive (pres- ent) and negative affirmative (absent) sense. Group definitions are never based on any single feature, but on a combination of lead characters supported by secondary features of lesser diagnostic and/or discriminatory value. 416 Annals of the Missouri Botanical Garden YY Ü SEE Sk > y 3S y LANAN | e Figure 9. Inflorescence types of group 1. In the following, observations on inflorescence racemose inflorescence; each flower has one brac- types are described and complemented by evidence teole at the base of the pedicel. Frequently, the derived from wood and bark structure: inflorescence axis is shortened, with the inflores- l. Tribe Laureae. A number of genera have a cence appearing umbellate. The inflorescences are Volume 83, Number 3 1996 van der Werff & Richter 417 Classification of Lauraceae Figure 10. Inflorescence types of group 2. often protected by a number of bracts (decussate or alternate). This group, with some modifications, has been recognized in nearly all classifications. It in- cludes genera such as Litsea, Lindera, Laurus, and Sassafras, for instance (Fig. 9). In terms of wood and bark structure, the is characterized by the absence of marginal paren- ma and, in most instances, of septate fibers. Conversely, phloem fibers are always present. Fur- ther subunits can be recognized, for example the genus Sassafras on account of its accentuated growth ring structure, unique in Lauraceae and re- flected in both secondary xylem (“ring porous”) and hloem (distinct layering by early and late formed tissue strata). ribe Perseeae Nees. niculate-cymose inflorescence. The initial branch- ing of the inflorescence is paniculate, with alter- nate or opposite branches, while the flowers are arranged in cymes. The lateral flowers of a cyme are strictly opposite. At some point along the ped- icel, two opposite bracts are present, frequently near the middle, but sometimes near the base. In- cluded in this group are most neotropical genera (e.g., Ocotea, Nectandra, Aniba, Licaria, Pleuroth- yrium) and some neotropical/Asian genera (e.g., Persea, Cinnamomum, Phoebe, and Dehaasia) (Fig. This group has a pa- Wood and bark structure depicts a group of gen- era characterized by the absence of marginal pa- Inflorescence types of group 3. Figure 11. renchyma and the ubiquitous presence of septate fibers (secondary xylem). Phloem fibers constitute part of the secondary destin of nearly all taxa at- tributed to this group ex some species of Aniba, Licaria, and Ocotea ide 19 Tribe Cryptocaryeae Nees. The third group is formed by genera with a paniculate-+ cymose inflorescence. At first glance these inflorescences look much like those of group 2, but the ultimate divisions are not strictly cymose. The lateral flowers of a “cyme” are not quite opposite, and flowers can appear individually placed along an inflorescence axis. The placement of bracts along the pedicels is variable in this group. Sometimes only one bract is present, sometimes two alternate or (sub)opposite ones; further observations are needed. This group includes such genera as Beilschmiedia, Cryptocar- ya, Endiandra, Potameia, and Triadodaphne (Fig. 11 Wood and bark structure supports this circum- scription of the Beilschmiedia/Cryptocarya assem- bly. All taxa share a number of distinctive features, such as the presence of marginal parenchyma, non- septate fibers with conspicuously bordered pits, and exclusively simple vessel perforations in the sec- ondary xylem. Conversely, in the secondary phloem the lack of fibers combines with characteristic sclereid formation. s far as wood and bark structure is concerned, not all taxa can be satisfactorily accommodated in the three groups described above. Cinnamomum and Persea, for instance, appear to be transitional 418 Annals of the Missouri Botanical Garden between Group 1 and Group 2, with closer affinities to the latter. Similarly, Mezilaurus (including Cli- nostemon), an easily defined and recognized taxon, shares diagnostic bark characters with Group 3 and diagnostic wood characters with Group 2. Other, mostly small genera with a very distinctive wood and bark structure do not fit well with any of the three groups, though certain affinities can be rec- ognized, for instance, in the case of Caryodaphnop- sis, Eusideroxylon/Potoxylon, and Hypodaphnis with Group 3, of Aspidostemon and Chlorocardium with Group 2, and of /teadaphne with Group Biogeographically, this division in three groups of genera is more logical than the generic alliances proposed in earlier classifications. The Laureae, with racemose inflorescences, are best represented, at the generic level, in the Northern Hemisphere (Laurus, Umbellularia, Parasassafras, Litsea, Lindera, Neolitsea), although several genera are well represented in the Asian tropics and a few extend into Australia. Most genera with unisexual flowers (about 10) belong to this group, and it in- cludes genera with four and two anther cells. The Perseeae are mostly neotropical, with the genera in the Persea—Cinnamomum— Sassafras, oebe complex also present in tropical and subtropical (Northern Hemi- sphere) Asia; Ocotea is also present in Africa and Madagascar. Only three genera in this group have unisexual flowers; one of these, Ocotea, also in- cludes many species with bisexual flowers. Both genera with 2-celled and 4-celled stamens are part of this group. The Cryptocaryeae are best repre- sented in the Southern Hemisphere (Cryptocarya, Beilschmiedia, Endiandra, Potameia), but are also present in the Northern Hemisphere. All genera in this group have bisexual flowers. The core genera (Cryptocarya, Beilschmiedia, Endiandra, Potameia, Triadodaphne) have all 2-celled stamens; Hypoda- phnis, Eusideroxylon, and Potoxylon, provisionally placed in this group, have 4-celled ant Although not all genera can be satisfactorily placed in Richter's system, it avoids several anom- thers. alies present in Kostermans's classification, such as treating Endiandra, Mezilaurus, Beilschmiedia as close relatives. Persea, and ius, the two recent classifications of Lauraceae differ greatly from one another. Kostermans’s (1957) classification is based mainly on one character only, the position of the gynoecium relative to the hy- panthium, while Richter’s ( 1) classification is based on several characters from bark and wood and is supported by observations on inflorescence Lauraceae. Ann. —— —. 198 New World ————. 1993 types. Although not all genera can be placed in our proposed tribal groupings (data are not yet avail- able for some small genera and some genera have small, few-flowered inflorescences, making an in- terpretation of the inflorescences difficult), the fact that two greatly different sets of data support this classification makes this the best classification at hand, and the one to be tested when additional data become available. Literature Cited Bentham, G. € J. D. Hooker. 1880. Laurineae. Pp. 146— 168 " G. Bentham & J. a Hooker, Genera Plantarum, vol. : Reeve, vus Hutc La J. 19€ era of Flowering Plants (Dic WOW dT S I. Clarendon Press, Oxford. Hyland, B. P. M. 1989, A revision of Lauraceae in Aus- tralia e luding Cassytha). Austral. Syst. Bot. 2: 135— 207. A. J. G.H es. Inst. 57 Kubitzki, K. 1982. Lauraceae: Aniba. Fl. Neotrop. Mon- ogr. 31: 1-84. Kostermans, 1957. Lauraceae. Comm. For. Re be — —— & H. G. Richter. 1987. ki & Richter (ro of neotropical Lauraceae. Bot. Jahrb. Syst. 109: 49-8 Meissner, C. F. 1864. Williamodendron Kubitz- Lauraceae. /n: A. de ae (ed- itor), Prodromus Systematis a 15: 1-26 Mez, C. ee Lauraceae Americanae. Jahrb. b Bot. Gart. Berlin 5: 1—556. Nees von Esenbeck, F. C. 1836. Systema Lauri- narum. Sumtibus Veitii et sociorum, Berlin. Pax, Lauraceae. Pp. 106-126 in A. K. Prantl (editar) Die natiirlichen PAanzenfamilien. vol. TIL Engelmann, Leipzig. Ric hee i G. 1981. Anatomie des sekundüren Xylem und der Rinde der Lauraceae. Sonderb. Naturwiss. Vor eins eea 5: 1-148. ive X je anatomy Mears sabia 3ull. n.s. 6: 187 Rohwer, J. G. ña. je 'eae: dA Fl. Neotrop. Monogr. 60: is ————. 1993b. m 'eae. Pp. —391 in K. En bitzki, J. G. Rohwer & V. Bittric x bon ) The F ilies and Genera of Vascular Plants Il. Sa lag, B Berlir a — c of Lauraceae Il. —199, . Richter & H. van der Werff. 1991. Two new genera e neotropic ‘al Lauraceae and critical re- bes on the q delimitation. Ann. Missouri Bot. Gard. 78: 388—40( Werff, H. van der. ton Six new species of neotropical Missouri Bot. Gard. 74: 401—412. 7b. A revision of grag (Lauraceae). Mun Bot. Gard. 74: 1988 Ann. "ight new species i one new combi- nation of | meoteopical Lauraceae. Ann. Missouri Bot. Gard. 75: 4 19. 19 )9]. A key to the genera of Lauraceae in the Ann. Missouri Bot. Gard. 78: 377-387. 3. A revision of the genus Pleurothyrium is pe Ann. Missouri Bot. Gard. 80: 39-118. THE USE OF CUTICULAR FEATURES IN THE TAXONOMY OF THE LAURACEAE’ D. C. Christophel?, R. Kerrigan’, and A. I. Rowett’ ABSTRACT Lauraceae, with more than 2500 species in ov r 50 genera, and with the general lack of agreement in their The recent classification n, make an ideal ae for a obal study based on leaf cuticles. A technique for preparing such as Endiandra and sige nd and in sorting out biogeographical anomalies, such as fi junct species within one where the cuticles confirm placement of dis enus. Potential tak la posed by larger genera such as Litsea, a Cinnamomum, Ocotea, and Nectandra are discusse Two of the most recent classifications of the Lau- raceae are those of Kostermans (1957) and Rohwer (Kubitzki et al., 1993). The former is based on his own life-long studies, and the latter relies heavily on Kostermans and on the wood anatomical studies of Richter (1981), as well as including studies (pri- marily South American) by the author himself. The problems associated with these and other classifi- cations of the Lauraceae are discussed by van der Werff and Richter (1996, this issue). From that re- view it is clear that further work is required to ar- rive at a satisfactory, or at least widely acceptable, classification of the family. Cuticular, and to a smaller degree, leaf architectural studies form one module in a multifaceted study of the generic de- limitations of this family currently in progress. The aim of this study is to analyze cuticular fea- tures for selected Lauraceae species whose generic positions have been predetermined by others, and in some cases questioned. At this time, no attempt has been made to use these features to confirm or disprove the evolutionary position, or indeed the “monophyly” (sensu cladistics), of these groups of Lauraceae. Consistent with that, individual features are not weighted or equalized, but are simply ex- amined in light of the hypotheses generated by the placement of the species in genera by earlier work- ers. Characters that have shown consistency in this study are being included by the first author in the global generic revision of the Lauraceae, which fol- lows on from this project. Even in that study, the rimary aim will be to place groups of species in consistently recognizable genera, and not necessar- ily to propose phylogenetic relationships (cladistic or otherwise). Cuticular studies of the Lauraceae are not novel. As early as 1926, Bandulska presented a study of fossil and extant Lauraceae and described some species of the former based on cuticular similarities to certain extant species. More recently, Hill (1986) discussed those features of the cuticle that allowed identification of Lauraceae, and placed the 12 spe- cies of fossil Lauraceae found in the Eocene Ner- riga deposit within the form genus Laurophyllum based on these cuticular features. He made no at- tempt to place them relative to extant genera. Christophel and Rowett (in press) present descrip- tions of leaf architecture and cuticles of all species of leafy Lauraceae found in Australia. They also present a dichotomous key to cuticular features that allows identification to the generic (or to a species group within genus) level. hile most of the genera occurring in Australia can be clearly defined (extant leafy genera found there include Beilschmiedia, Cinnamomum, Cryp- tocarya, Endiandra, Lindera, Litsea, and Neolitsea), a notable exception is Cryptocarya. Although its species do not intermingle with those of any other ! The authors thank Linda Allen and Kevin Celand for cuticle preparation, and Heidi Wittesch for aiding with photo preparation. B. P. M and for discussions. R. S. Hill supported by grants from the Australian Researc . Hyland (QRS) and H. van der Werff (MO) are thanked for "— leaves and identific TP ill and J. G. Rohwer provided helpful criticisms of the 'h Council and the Wet Tropics paa Agency nuscript. e project w 2 Botany Department, University of Adelaide, Adelaide, S.A. 5005, Australia 3 derum Section, South Australian Department of Minerals and Energy, Greenhill Road, Adelaide, S.A. 5063, Australia ANN. Missourt Bor. GARD. 83: 419—432. 1996. 420 Annals of the Missouri Botanical Garden genus, no set of characters has been found that holds all species together and yet separates the ge- nus from others (Christophel & Rowett, in press). The result is that five groups of species key out to Cryptocarya in different parts of the key. Unfortu- nately, these five groupings do not correspond to those of Hyland (1989). This discrepancy indicates that further study of Cryptocarya is warranted, with the possibility existing that the genus as now de- fined is not natural and perhaps even polyphyletic. The heterogeneity, and perhaps artificiality of this large genus, has been previously suggested based on other features (e.g., Rohwer in Kubitzki, 1993). In both the systems of Kostermans (1957) and of Rohwer (1993), the genera occurring in Australia are well spread across the major subfamilial group- ings, and hence the mainly successful generic sep- arations of Christophel and Rowett (in press) sug- gest that their character set represents a good starting place for the global generic revision. Di chotomous keys in general (e.g., Christophel & Rowett, in press) suffer from the requirement that certain features must be used before other (perhaps more obvious) ones can be applied. Therefore, the results of the global revision will be presented as a computerized, multi-entry key. If genera such as Cryptocarya do not fall out more reasonably when a better sample of th e polyphyletic nature of the genus will be sup- ported. eir species is included, then METHODS The majority of the leaf samples used in this and the ensuing complete study have been provided, or at least have had their identifications confirmed, by either H. van der Werff, Missouri Botanical Garden (MO) or B. P. M. Hyland, Herbarium Australiense— Atherton (QRS), who are both participating in the global revision of the Lauraceae. Hence the con- sistency of identification is maximized. Samples were only taken from mature leaves, and 1-cm? pieces were taken from the near basal margin on the lefthand side of the leaf with the adaxial surface upward, to eliminate variation that might have been caused by sampling from different areas. At least three samples from two different collections were used whenever possible, although for some species only one collection could be obtained, in which case several leaves from it were used. A list of specimens studied is presented in Table 1. While the technique for cuticle preparation is similar to that used by Christophel and Rowett (in press), it is described here for completeness. The pieces of leaf are placed in small test tubes and soaked in 70-95% ethanol for approximately 18 hours. Experiments have shown that no observable difference in either quality of preparation or speed can be detected in samples by varying the ethanol strength within the above range. The ethanol is then decanted off and 10 drops of 4096 w/vol H,O, and five drops of 9096 ethanol are added to the test tube, which is then heated to gentle boiling in a water bath for 2—48 hours. The cooking can be judged complete when the sample has turned light yellow to white and when the cut edges of the cu- ticle can be observed to be peeling back from the leaf surface like the two covers of a book. At that point the sample can be removed from the test tube and placed in a watch glass or petri dish of water, where the “book covers" are pried open and the "pages" of the book, or the actual cellular material, can be brushed away with fine artists’ brushes. removal of cellular contents specimens at this stage proves difficult, the specimens can be returned to 70-90% ethanol for a further 12 hours, which usu- ally renders them much easier to clean. While no chemical reaction involving the ethanol and H,O, is apparent, leaf pieces pre-soaked in this manner clear between three and four times faster than those not pretreated. It may be that the ethanol acts to break surface tension on the cells, thus fa- cilitating entrance of the hydrogen peroxide into the tissue. This is supported by the fact that prepara- tion of fresh leaf material, which is naturally hy- drated, is not enhanced to the same degree by the pretreatment in ethanol as are dried specimens. Cleaned cuticles can be quickly rinsed in 296 ammonia to adjust pH, and then stained in 0.196 Crystal Violet for 20-120 seconds, or until the cu- ticle appears uniformly light purple to the naked eye. Under the compound microscope, anticlinal cell walls should appear bright to dark purple, as should thickened cuticular ledges, but the lumens of the cells should not appear stained, or with only a slight tinge of purple. Overstained samples can be destained by rinsing in 90-100% ethanol. In general, understained slides are preferable to over- stained ones, as various techniques, including in- terference contrast, can be used to compensate for lack of stain, but little can be done to observe ob- scured detail in overstained cuticles. Stained cuticles can be mounted in phenol glyc- erin jelly. When the slides have set and excess jelly is cleaned from the coverslip with warm water and a razor blade, the coverslip can be ringed with nail varnish (polish) to retard dehydration. If more per- manent mounting is desired, the stained cuticles can be dehydrated and mounted in a resin-based medium. However, the first author has had slides Volume 83, Number 3 1996 Christophel et al. 421 Cuticular features of Lauraceae Table 1. Specimens examined. Species Country Collector & number Aiouea guianensis Aublet Aiouea guianensis Aublet Beilschmiedia bancroftii (Bailey) C. White Beilschmiedia bancroftii (Bailey) C. White B. brunnea B. Hyland B. castrisinensis B. Hyland B. castrisinensis B. Hyland B. castrisinensis B. Hyland B. collina B. Hyland B. collina B. Hyland B. elliptica C. White & Francis B. obtusifolia (F. Muell. ex Meissner) F. Muell. B. obtusifolia (F. Muell. ex Meissner) F. Muell. B. oligandra L. S. Smith B. oligandra L. S. Smith B. oligandra L. S. Smith B. peninsularis B. Hyland B. peninsularis B. Hyland B. tooram (Bailey) B. Hyland B. tooram (Bailey) B. Hyland B. volckii B. Hyland B. volckii B. Hyland Caryodaphnopsis henrii A. Shaw C. tonkinensis (Lec.) A. Shaw C. tonkinensis (Lec.) A. Shaw /. Sp. nov. Chlorocardium venenosum Rohwer et al. Chlorocardium venenosum Rohwer et al. Cinnamomum baileyanum (F. Muell. ex zn Francis C. laubatii F. Muell. C. laubatii F. Muell. C. laubatii F. Muell. C. oliveri Bailey C. oliveri Bailey C. propinquum Bailey C. propinquum Bailey C. sellovianum (Nees & Mart.) Kosterm. C. sellovianum (Nees & Mart.) Kosterm. C. virens R. Baker Cryptocarya sp. Cryptocarya sp. Cryptocarya sp ndiandra acuminata C. White & Francis Endiandra acuminata C. White & Francis ndiandra acuminata C. White & Francis E. anthropophagoram Domin. : anthropophagoram Domin. v. bellendenkerana B. Hyland B c B. Hyland E. bessaphila B. Hyland x bessaphila 3 Hand collinsii B. Hyland P collinsii B. Hyland Brazil Philippines Peru Colombia Ecuador Australia Australia G. T. Prance 4401 Jansen Jacobs 1986 ra Hyland 10104 Gray 1605 Hyland 4128 PREIS! Qe ) Q Q Q Q ) ) S < on No] on Kunming Herb. 645874 Kunming Herb. 0125 Ramos 1667 Jaramillo 11930 Cogollo 5153 Zak 3913 B. Hyland 2648 B. Gray 315 B. Gray 3923 B. Gray 3925 B. Hyland 3303 B. Hyland 12456 . Gray 686 B. Gray 921 Harley 25261 Duarte 9295 B. Hyland 12340 Miller 3725 Miller 3955 Armbruster 3145 B. Gray 2951 B. Gray 2953 B. Gray 2072 B. Gray 1644. B. Hyland 10650 B. Hyland 3570 B. Hyland 11134 422 Annals of the Missouri Botanical Garden Table 1. Continued. Species Country Collector & number E. compressa C. White Australia B. Gray 1753 E. compressa C. White Australia . Gray 3723 E. cooperana Australia B. Gray 3209 E. cooperana Australia B. Gray 3565 E. cowleyana Australia B. Gray 1818 E. cowleyana Australia B. Gray 4123 E. crassiflora €. White & Francis Australia B. 1 no ~ E. dichrophylla F. Muell. Australia B. Gray 29 E. dichrophylla F. Muell. Australia B. Gray 37 i E. dielsiana Teschner Australia B. Hyland 13176 E. dielsiana Teschner Australia B. Gray 3890 E. discolor Benth. Australia B. des 2996 E. discolor Benth. Australia p. )2 E. floydii B. Hyland Australia B. non 4619 E. glauca R. Br. Australia B. Gra E. glauca R. Br. Australia B. a 14039 E. globosa Maiden & E. Betche Australia B. Gray 3173 E. globosa Maiden & E. Betche Australia B. Gray 4152 E. grayi B. Hyland Australia B. Gray 3157 E. grayi B. Hyland Australia B. Gray 3726 E. grayi B. Hyland Australia B. Gray 4084 E. grayi B. Hyland Australia B. Gray 4162 E. hayesii Kosterm Australia B. Hyland 4553 E. hypotephra V. Muell. Australia B. Gray 456 E. hypotephra F. Muell. Australia B. Gray 3969 E. hypotephra F. Muell. Australia B. Gray 4001 E. impressicosta Alenn Australia B. Gray 2454 E. ibid ie Alenn Australia B. Gray 3101 E. insignis (Bailey) Bailey Australia B. Gray 1641 E. insignis (Bailey) Bailey Australia B. Gray 3714 E. introrsa C. White Australia B. ai 12338 E. jonesii Australia B. Gray 2255 E. jonesii Australia B. Gray 2869 E. kingiana Gamble Sarawak B. Hyland Hlo E. leptodendron B. Hyland Australia B. Gray : E. leptodendron B. Hyland Australia B. Gray 3953 E. limnophila B. Hyland Australia B. Hyland 12375 E. limnophila B. de land Australia B. Hyland 12377 E. longipedicellata C. White & Francis Australia B. Hyland 6256 E. d C. White & Francis Australia B. Gray 3243 E. microneura C. Whit Australia B. Gray 3822 E. m iru B. Hyland Australia B. Hyland 12584 E. monothyra B. Hyland Australia B. Gray 1647 E. monothyra B. Hyland Australia B. Gray 1652 E. monothyra B. Hyland Australia B. Gray 3284 E. monothyra B. Hyland Australia B. Cray 3073 E. montana C. White Australia B. Gray 2825 E. montana C. White Australia B. Gray 3127 E. muelleri Meissner in DC Australia B. Hyland 12330 E. muelleri Meissner in DC Australia B. Hyland 12957 E. muelleri Meissner in DC Australia B. Hyland 12972 E. palmerstonii (Bailey) C. White Australia B. Hyland 11721 E. phaeocarpa B. Hyland Australia B. Gray 3297 E. phaeocarpa B. Hyland Australia B. Gray 3873 E. pubens Meissner in DC Australia B. Gray 297] E. sankeyana Bailey Australia B. Gray 3579 E. sankeyana Bailey Australia B. Gray 4103 Volume 83, Number 3 Christophel et al. 423 1996 Cuticular features of Lauraceae Table 1. Continued. Species Country Collector & number E. sideroxylon B. Hyland Australia B. Gray 3904. E. sideroxylon B. Hyland Australia B. Gray 4148 E. sieberi Nees Australia B. Gray 1504 E. virens F. Muell. Australia B. Hyland 4602 E. wolfii B. Hyland Australia B. Gray 1328 E. wolfii B. Hyland Australia B. Gray 3962 E. xanthocarpa B. Hyland Australia B. Gray 1614 E. xanthocarpa B. Hyland Australia B. Gray 5155 V. sp. a B. Hyland 14217 Litsea costalis (Nees) Kosterm. Sembilon B. Hyland 14130 Ocotea lentii Burger Costa Rica Moraga 171 Ravensara sp. Madagascar aps 3835 avensara sp. Madagascar . Mc d da 14428 Rhodostemonodaphne grandis (Mez) Rohwer French Guiana iine Rhodostemonodaphne grandis (Mez) Rohwer Peru Kroll "A Rhodostemonodaphne grandis (Mez) Rohwer Peru Kayap 3541 R. kunthiana (Nees) Rohwer Ecuador Zak 4452 R. kunthiana (Nees) Rohwer Costa Rica Herrera 2397 mounted in glycerin jelly for 20 years without harm befalling them, and the advantages of that medium are that it is inexpensive, quick to use, and the cuticles can be retrieved by soaking the slide in hot water if other preparations or stain adjustments are to be made later. For long-term storage in a dry environment, dehydration will be better retarded by ringing slides in a wax. Safranin staining also pro- vides good visual detail of the cuticles, but the red color necessitates the use of a green filter for max- imum contrast with black-and-white photography, while the purple staining of Crystal Violet does not require this. All cuticular photographs (Figs. 1-18, 20-25) were taken using a Zeiss automatic 35mm camera on a Zeiss Axioplan microscope with a 40X objec- tive and 10X eyepiece. CUTICULAR CHARACTERS Leaf cuticular characters used to delimit genera in the Australian Lauraceae (Christophel & Rowett, in press) fall into three broad categories: (1) those of the walls of normal, non-stomatal cells, (2) those of the stomatal complex (including subsidiary cells), and (3) those of specialized cells such as trichomes, trichome bases, and hydathodes. The Lauraceae appear to be uniformly hypostomatic, and hence separate character sets can be used for the adaxial (upper) and abaxial (lower) cuticles. In practice, leaves of many species of Lauraceae ap- pear to differ in their two cuticular surfaces, not only in presence/absence of stomata, but also in features of the indument and of general epidermal cell size and shape (e.g., Figs. 1, 2). Considering first the features of non-stomatal cells, the periclinal walls can be seen to vary from densely papillose (Fig. 1) to lightly papillose (Fig. 3), granular (Fig. 2), punctate (Fig. 10), slightly striate (Fig. 6), or smooth (Figs. 14, 18). To be con- sidered granular, the periclinal walls must be reg- ularly and coarsely so. There are probably several subtly different states between coarsely granular > Figu Abaxial cuticle of Eum henrii A. Shaw, from L509, Botany Department Herbarium, Adelaide eer collected from the unming Botanical Institute Herbarium (#645874), Kunming, China. Collector and number in Chinese and not readable. Collection from Southern China. Magnification as per Figure 24.—2. Adaxial ae of Caryodaphnopsis Caryodaphnopsis sp. nov. en from L317, Botany Depart at MO, from Peru; magnification as per from L508, Botan rii A. Shaw. All information and magnification as per Figure ment Herbarium, Adelaide Üniversity, from Jaramillo 11930 Figure 24.—4. Abaxial cuticle of Caryodaphnopsis tonkinensis (Lec y Department Herbarium, Adelaide University, from the Botanical Institute of Kunmin —3. Abaxial cuticle of .) A. Shaw , China, (sheet #0125); collected in southern China; magnification as per Figure 24.—5. Abaxial cuticle of Chlorocurdium venenosum Rohwer, Richter & van der Werff taken from L398, Botany Department Herbarium, Adelaide University; originally from Cogollo 5153 at MO; collected in Colombia. Scale EN as per Figure 20.—6. Adaxial cuticle of Chlorocardium vene- nosum. All details and magnification as per Figure : 424 Annals of the Missouri Botanical Garden Cuticular features of Lauraceae Christophel et al. e © 2 E 2 Z e © o E 3 o > 426 Annals of the Missouri Botanical Garden and truly smooth, but, in practice, artifacts such as remains of epicuticular waxes, stain granules, and bits of adhering cell wall complicate such deter- minations. These difficulties could have been over- come by rigorous and prolonged chemical treat- ment, and by observation using SEM; however, one of the major purposes of this study is to develop a character set that is readily usable by general bot- anists and ecologists, and hence such highly spe- cialized treatment was not employed. Several separate features of the anticlinal walls can be characterized. The uniformity of thickness is one such character. Anticlinal walls can be reg- ularly uneven (beaded, e.g., Figs. 7, 9), or if sinu- ous in outline can be thickened in the troughs (but- tressed, e.g., Fig. 11), or can be irregularly thickened or smooth, which is the most common state. The straightness of the anticlinal walls is also important, and cuticles are scored for the majority of cells being angular (Fig. 13), regularly or irreg- ularly rounded (Fig. 5), undulate (one trough and peak per wall, e.g., Fig. 18), or sinuous (more than one peak and trough per wall, e.g., Fig. 11). This anticlinal wall shape was placed in a seven cate- gory index by Wilkinson (1979), and this alterna- tive method of scoring is also being tried for the global revision, to be followed by analysis to deter- mine the better system. Uniformity of cell size and maximum dimension are also features being in- cluded. An example of between-species variation of average cell size can be seen in Figures 7 and 11. Of equal import is the variation apparent in some species in cell size and shape between the adaxial (upper) surface and the abaxial (lower) surface. Relative uniformity of features may be seen in Fig- ures 7 and 8, while variation in shape between sur- faces can be seen in Figures 11 and 12, and vari- ation in cell size between surfaces may be seen in Figures 5 and 6. While the number of subsidiary cells (2 Lauraceae and their arrangement (paracytic) are ) in the uniform across the family, and all species are hy- postomatic, the stomatal complex still provides a rich character set, likely including many of the fea- tures that will prove critical in the delimitation of supraspecific taxa. As illustrated in transverse sec- tion in Figure 19, and as shown by Hill (1986) in his figure 1, the stomatal complex in the Lauraceae has sunken guard cells with over-arched subsidiary cells. Although the family is truly paracytic, to the non-specialist observing the cuticle in surface view the subsidiary cells appear to be guard cells, and the unspecialized cells in the adjacent row are con- sidered to be subsidiary. Their variable arrange- ment leads to identifying the subsidiary pattern as virtually nonexistent, or anomocytic. Possession of uneven-sized subsidiary cells (Fig. 9) has proven common only in certain genera, such as Beilschmiedia, and could be of taxonomic utility. By far the most variable feature of the complex, and hence possibly one of the most useful at various levels, is the nature of the two cuticular ledges (Fig. 19). sometimes outward in the stomatal pore, and arises The upper (Fig. 19 UL) protrudes upward and from the lateral walls of the subsidiary cells. In some taxa, the lower ledge (Fig. 19 LL) protrudes between the guard cells (Fig. 19 GC) and the sub- sidiary cells (Fig. 19 SC), sometimes forming into various shapes inc ae non that look somewhat like butterflies (Figs. nomic feature is i dy to riis by the fact that these . Àn additional taxo- expanded ledges are variously deciduous. Some genera, like Endiandra, have both ledges visible, although not expanded, and this “double” ledge can thus be used to distinguish this genus in Australia (Christophel & Rowett, in press). Some taxa, including many not found in Australia, have other ledge configurations, and in extreme cases (e.g.. Fig. 24. Aiouea guianensis Aublet) the entire stomatal complex may be obscured by a thickened ring of cuticle. Finally, some features of cells ad- jacent to the stomatal complex may also be modi- fied, such as the papillae in Caryodaphnopsis (Figs. T » The final category of taxonomically useful char- acters is derived from the presence/absence and features of various specialized cells. In the Laura- ceae, all trichomes thus far reported in the litera- 3823 hou gnific ation as per barium, Adelaida eas om B. Gray miedia volckii. All lobo < ma White & Francis from L643, Bota QRS; collected in Australia; m mation and magnification as per Figure 7.— Muell. from L644, Botany Pepaument learn Adelaide University, fro in Australia; magnification as magnification as per Figure > per Fig agnific ation » Hm Figure 7.—10. Abaxial cuticle of Beilsc nd Mu ue (F. 7.—12. Adaxial cuticle of Beilschmiedia a. ee cuticle of a volckii B. Hyland taken mn L650, Botany Department Her- at QRS; collec w in Australia. —8. Figure 7 Aba y Department Herbarium, a University, from B. Hyland Adaxial cuticle of Beilsch- dal peru of JHeichmiedia elliptica C. 4128, housed at of Beilschmiedia elliptica. All infor- Muell. ex nes ad r) F. 9 housed at QRS; collected All ONEA and Adaxial cuticle 428 Annals of the Missouri Botanical Garden 950550 000 D UL oC v GC 19. Diagram of transverse section of typical Lauraceae stomatal complex (modified ies Hill, 1986, fig. 1). Figure Stippled area shows actual cuticle, as would be seen in a cuticular preparation. Black are guard cells (GC), over-arched by the subsidiary cells (SC). Upper cuticular ledge (UL) a ea represents sunken, embedded nd br cuticular ledge (LL), which vary in prominence across the family, are also shown. Leaf surface toward top in Mou ture or observed by the authors have been simple (e.g.. Litsea costalis, (Nees) Kosterm. Fig. 13). Gen- erally their bases can be characterized as poral, meaning that they consist not of a set of specially modified cells, but of a pore, or hole, at the junction of several non-stomatal cells. In some species, e.g., Ocotea lentii Burger, giant trichomes are found that clearly have rings of modified cells surrounding the trichome base. The trichome and its base provide separate suites of characters, as the trichomes are variously deciduous from species to species, and even in those species where they are not deciduous, they can be easily broken in preparation. Obvious features such as trichome density and trichome po- sition relative to veins or areoles are also of poten- tial taxonomic utility. auraceae also display regions of thick- "v ell (5-20) that stain densely and are likely glandular. It should be noted that wound tissue has a similar appearance, but that interpretation can be discounted when the areas are regular in their lo- cation across a cuticle slice (and indeed duplicated on additional specimens). The final type of special epidermal structure that has been observed in the Lauraceae is the hydathode. Wilkinson (1979) sug- gested that structurally there are several types of hydathodes, but within this study the term is re- stricted to giant stomates. These structures are usu- ally 2-2.5 frequently occur over veins. It must be noted that times the size of normal stomates, and some Lauraceae species have variable-sized sto- mates, and care must be taken to avoid labeling a large stomate at the extreme of a size range as being a hydathode. PRELIMINARY RESULTS When groups of related species or genera are examined for the expression of the above-described characters, some interesting preliminary results can be seen. For example, Christophel and Rowett (in press) noted that Endiandra species were charac- e Figures 13-18. ment Herbarium, Adelaide University from B. Hylar hairs and poral hair bases.—14 3.—15. Abaxial cuticle of xpo did glauca R. Cuticle taken from L671, i 13.—16 L705, Botany Department Herbarium, Adelaide University, from B. Gray 1328 hous Abaxial cuticle of ida sieberi Nees sed m L 703, fh Department Her- magnification as per Figure 13.—17 barium, Adelaide University, from B. Gray 1504 housed a 13.—18. Abaxial cuticle of Endiandra kingiana Camble ded “L528, . Abaxial cuticle of Endiandra montana C. Herbarium, Adelaide University, from B. Gray 2825 housed in Br. Photographed with ment Herbarium, Adelaide Uniendo om LB Cre —13. Abaxial cuticle of Litsea costalis (Nees) Kosterm. Cuticle taken from L190, Botany Depart- id 14130 housed at QRS; iin ted in Sembilan. Note den se simple e taken from L691, Botany Department n QRS; collec isis in jio jdn ation as pe er dui 3211 bu in QRS Abaxial cuticle Vi usd w oli B. Hyland. Taken from S: collected in Australia; ollected in Australia; magnification as per Figure Botany Department Herbarium, Adelaide 13. University, from B. Hyland 14146 housed at QRS; collected in Sarawak; magnification as per Figure Volume 83, Number 3 1996 Christophel et al Cuticular feature s of Lauraceae 429 430 Annals of the Missouri Botanical Garden terized in Australia by having relatively straight, thin, double stomatal ledges. These may be ob- 15, 16, and 17, where it can be seen that a thin, inner (upper) ledge is visible served in Figures 14, just inside the stomatal pore, and another parallel set of ledges (lower) can be seen outside of these, separated by approximately one-third of the sub- sidiary cell, and parallel to its long axis. While the other three Australian species illustrated (Figs. 14, 16, 17) are photographed with normal transmitted light, Endiandra sieberi Nees (Fig. 15) was photo- graphed with interference contrast (Nomarski) mi- croscopy, to compensate for unsatisfactory staining of that particular specimen. Not only does this clearly show the pairs of ledges, but the faint out- lines suggest the position of the sunken guard cells may also be seen. Beilschmiedia, the sister genus to Endiandra, lacks these straight, double ledges. While straight, double ledges appear a good de- scriptor for Endiandra, it should be noted that other features are variable. For campa the anticlinal walls of Endiandra montana C. White, E. wolfu Hyland, and £. glauca R. Br. are straight, whereas E. sieberi Nees has irregularly rounded to undulate anticlinal walls. A measure of the success of this character must be its effectiveness on extra-Australian species. To date 20 such species have been examined, and all show the double ledges. Some (e.g., Endiandra kin- giana Gamble from Sarawak—Fig. 18) even illus- trate the feature more clearly than the Australian species. Beilschmiedia is considered to be closely related » to Endiandra by most authorities (e.g., Hylan 989; Kostermans, 1957; Rohwer, 1993). In gen- eral, it can be separated easily by the fact that its species normally possess nine fertile anthers. How- ever, just as in Endiandra, in Australia a group of Beilschmiedia species (4) possess only six anthers. While this would qualify them for membership in Allen’s Brassiodendron, Hyland (1989) has placed them in Beilschmiedia based on a suite of other features. Christophel and Rowett (in press) were able to separate Endiandra and Beilschmiedia in Australia quite easily in that no species of Beilschmiedia had the nearly straight, double ledge that Endiandra stomates possess. Beilschmiedia itself is character- ized by all species having some degree of uneven anticlinal wall thickening (buttressing or beading) (Figs. 7, 9, 11). Some species, such as Beilschmie- dia obsta (F. Muell. ex Meissner) F. Muell. (Figs. 11, 12), show a marked degree of difference in the degree of thickening between the adaxial and abaxial walls. This feature, however, is not restrict- ed to Beilschmiedia, as some Endiandra species show anticlinal wall thickening. In addition, all Australian species appear to have some degree of unevenness in the size of the two subsidiary cells (e.g., Beilschmiedia elliptica C. White & Francis, Fig. 8). This is not as reliable a feature for generalized identification, however, as in many species it is not visible on stomates in some regions; hence a random leaf piece might not show them. Cuticular analysis can also be used on species groups, with potential biogeographical significance. One of the most obvious disjunctions in the family occurs in Caryodaphnopsis, which has species re- corded from South America and also from Southeast Asia. Examination of the cuticles of Caryodaphnop- sis species shows not only some unique features, but also highlights the fact that both geographic groups possess them. While the adaxial (upper) cu- ticular surface of Caryodaphnopsis species usually has smooth to granular periclinal walls (Fig. 2), the abaxial surface invariably has papillae present to some extent on the periclinal walls. The density of the papillae varies from species to species, but in all species examined papillae are clearly present on the cells surrounding the stomatal complex, and even over-arch it to a degree. This feature can be | Figures 20-25. s. Specimen citation and scale bar huntiana (Nees) Rohwer showing butterfly- os ad lo 4452 a t MO; collected in Ecuador (Nees E M rt.) Kosterm. BT a Adelaide University. Prepared from Harley 252 —20. Abaxial cuticular surface of y Hele bur (Mez) Rohwer showing butterfly- shaped lower cuticular ledge (dark stained). Cuticle from s spaci men , Bot Pu from Mori 20904 at MO; collected in French Guiar as per Figure 20 = I--e y Department Herbarium, Adelai —21. 1 Merida surface of Rhodostemonodaphne —22. Abaxial cuticular surface of Nt iir a ii ower cuticular ledges. Ape with Nomarski i f j ontrast microscopy; prepared from L106, Botany sce aps Herbarium, Adelaide r; magnification as per Figu showing cuticular collar nee n apreta taken from L124, Botany Department University. Originally from Zak e 20.—23. Abaxial cuticle of Cinnamomum sellovianum 261 at MO; collected in Brazil; magnification as per Fi igur 0.—24. Abaxial cuticle of Atouea guianensis Aublet from L113, Botany Department Herbarium, Adelaide U a from Jensen Jacobs 1986 at MO; « magnification as per Figure 24. 'ollected in Guyana.—25. Adaxial cuticle of Aiouea guianensis. All information anc Volume 83, Number 3 1996 Christophel et al. 431 Cuticular features of Lauraceae seen in the three species illustrated for this paper (Figs. 1, 3, 4); two of these (Caryodaphnopis henrii A. Shaw and C. tonkinense (Lec.) A. Shaw) were collected in China and the third, an as-yet-unde- scribed species, occurs in Peru. DISCUSSION The use of cuticular features for various taxo- nomic delimitations has been successful in various diverse groups (e.g., Araucariaceae—Stockey & Ko, 1986; Myrtaceae—Christophel & Lys, 1986). It has also proven satisfactory for family identifi- cation (Hill, 1986) and local generic differentiation (Ow et al., 1992) within the Lauraceae. Preliminary investigations have now shown in this paper that, based on genera such as Endiandra, Beilschmiedia, and Caryodaphnopsis, the use of a suite of cuticular characters applied to a global range of genera and species has great promise. Concerning the examination of Endiandra, of particular interest among the Australian species are the two that have six anthers (all other species in the genus possess three anthers). Hyland (1989) commented on them and noted that while some workers (e.g.. Allen, 1942) placed them in a sep- arate genus, Brassiodendron, he (Hyland) consid- ered anther number by itself to be a non-definitive character, and the Endiandra based on a majority of other features. Endiandra montana C. White (Fig. 14) is one of those six-anthered species, and is shown to clearly belong with the other Endiandra species because of the double ledge cuticle feature. The other six- anthered species, while not illustrated here, also has the double ledge, and hence supports Hyland's placement in Endiandra. It should be noted that no attempt is made here to infer that the double ledge character is more taxonomically important than anther number, but it does support Hyland's placement of those species in Endiandra, which was based on a suite of morphological characters. The four species of Beilschmiedia in Australia included two species in that possess six anthers have been examined, and show no obvious differences from the remaining species in the genus. This can be seen by compar- ing one of ES a Beilschmiedia volckii B. Hy- land (Figs. 7, 8), to the nine-anthered species A obtusifolia (F. Muell. ex Meissner) F. Muell. and B. elliptica C. White & Francis (Figs. 9, 10, 11, 12). Perhaps of even more significance is the fact that they are clearly distinct from the six-anthered species of Endiandra in their stomatal ledge structure. The few species of extra-Australian Beilschmiedia examined thus far have agreed with the features found by Christophel and Rowett (1995) for that genus in Australia. Following from the results of the study of Cary- odaphnopsis cuticles, it should be noted that in the Lauraceae, papillae are not restricted to that genus. For example, they can be found in a few species of Beilschmiedia and are reported to be common in Aniba (Kubitzki & Renner, 1982). However, it is the similarity of the structure of the papillae, and not only their presence, that supports the place- ment of these disjunct species of Caryodaphnopsis in the single genus. he exclusive use of cuticle characters within the family is not without some problems, or at least challenges, however. Christophel and Rowett (in press) have shown that within Australian species of Cryptocarya, good generic delimiting features are not obvious. Examination of approximately 20 ad- ditional species from this genus occurring outside of Australia has not solved this difficulty, although it has become clear that the decision to sink Rav- ensara from Madagascar into Cryptocarya s.l. is a good one, based on epidermal similarities. Another group that should prove challenging is Cinnamomum. It can be seen that South American species of this genus have distinctive cuticular sig- natures (Fig. 23) and that they are clearly different from the Cinnamomum species of Australia figured by Christophel and Rowett (in jn has e the American Cinnamomum spec a placed until recently in Phoebe Blas 1993), Asian species of the genus, including the type spe- cies, form a third distinctive group, and further study may well lead to the conclusion that the group is not natural (and is perhaps polyphyletic) as it is currently defined. The tripli-nerved, or acrodromous venation pat- tern is common in both Crypotcarya and Cinna- momum. However, no positive correlation has been found between groups of species possessing that character and those sharing common cuticular fea- tures. Basic venation characters are nonetheless being initially included in the global revision in order to test them on the full spectrum of species in the family. Hyland (1989) suggested that the generic com- plex of Litsea, Neolitsea, and Lindera is also in need of further detailed study. Christophel and Rowett in press) had little difficulty separating the three genera based entirely on Australian material, but those species represent only a small proportion of the total for the three genera, and the lack of single outstanding features to separate them, similar to the suggests that the ~ double ledges of Endiandra, group could provide a challenge on a global level. 432 Annals of the Missouri Botanical Garden Finally, some of the largest genera (e.g., Ocotea and Nectandra) occur in South America, and that material has been the least thoroughly studied by us to date. A strong possibility exists that both of these genera are currently unnatural (H. van der Werff, MO, pers. comm.) and hence much work needs to be done here. Preliminary observations on material from various genera in South America, however, suggest that some features that are highly diagnostic in Australian material (e.g., the butter- y-shaped cuticular ledges prominent in Crypto- carya and to a lesser extent, Cinnamomum) are in fact widespread in other geographic regions and in other currently prescribed taxa. The potential util- ity of such characters will only be assessable once thorough studies of the material have been made. Literature Cited Allen, C. K. 1942. Studies in Lauraceae IV. Studies in Papuasian species collec E a the Arnold Expeditions. J. Arnold Arbor. 23: 112- obiter H. bet. On n iie of some fossil and ent Lauraceae. de Lin c. Bot. 47: Christophel, D. c. S. D. pen 1986. Mummified ER of two new species d gc from the Eocene of Victoria, Australia. A l. J. Bot. 34: 649-662. ———— & A. I. Rowett. "1996. A Leaf and Cuticle Atlas of Australian Leafy Lauraceae. Flora of Australia Sup- dd Series. Australian Gov. Printer, Melbourne, Hill. R. " 1986. Lauraceous leaves from the Eocene of Wales. Alcheringa 10: 327-351 89. A revision of Lauraceae in nus tralia (excluding Du. Austral. J. Syst. Bot. 2(2, 3): 135-367. Kostermans, : 193-256. Kübiteki, K. & S. Renner. 1982. Lauraceae l (Aniba and Aiouea). Flora Fe a: )w, C., J. Hsiao & C. Liao. 1992. Cuticle Micromor- phology of Cinpamomese (Lauraceae) from Taiwan. Bull. par Forest NCHU 14(2): 1-30. Richter, H. 81. Anatomie des sekundaren Xylems und der Bus der Lauraceae. Sonderb. Naturwiss. Ver- eins Hamb. 5 Rohwer, J. G. 1993. Lauraceae. Pp. 366-391 in K. Ku- bitzki, J. G. Rohwer & V. Bittrich (editors), The Fami- lies 3 Genera of Vascular Plants II. Springer-Verlag, Ber Stoc fen R. A. & H. Ko. 1986. Cuticular mic romorphol- ogy of em 'ariaceae Jussieu. Bot. Gaz. 147(4): 508— 548. A. J. G. 1957. Lauraceae. Reinwardtia Werff, H. van der & H. G. Richter. 1996. Toward an improved classification of Lauraceae. Ann. Missouri ot. Gard. 83: 409-418. Wilkinson, H. P. 1979. The plant surface. Pp. 97-162 in ;. R. Metcalfe € L. Chalk, Anatomy of the Dicotyle- dons, 2nd ed. Clarendon Press, Oxford. Volume 83, Number 3, pp. inis of the ANNALS OF THE oo BOTANICAL GARDEN as published on August 8, 199 Experimental and Molecular Approaches to Plant Biosystematics The proceedings of the Fifth International Symposium of the International Organization of Plant Biosystematists (IOPB) Edited by Peter C. Hoch and A. G. Stephenson Twenty-three original contributions that span the breadth of biosystematics, a dynamic field of study that bridges the realms of systematics and population biology. The papers are arranged in four groups, Annals of the Missouri Botanical Garden, Volume 82, Number 2: Alternative Genes for Phylogenetic Reconstruction in Plants y A symposium cosponsored by the American Society of Plant Taxonomists and the Botanical Society of — America, organized by Pamela S. Soltis and Douglas E. Soltis, and presented at the 1993 AIBS meetings. Although the chloroplast gene rbcL has been successfully used to reconstruct plant phylogeny, many - important questions of plant phylogeny and evolution cannot be addressed using it. The contribut this issue of the Annals explore the potential of eight alternative genes or DNA regions for phylogenetic reconstruction at a variety of hierarchical levels. Both nuclear and chloroplast genes are evaluated. Three — regions of the nuclear ribosomal RNA cistron are explored: the 18S gene, the internal transcribed spacers (ITS), and the 26S gene. Small multigene families from the nuclear genome may also carry phylogenetic signal: the phytochrome gene family and the small heat shock gene family. Three genes from the chlo- - roplast genome are also considered: atpB, ndhF, and matK. Each paper describes the location, size, structure, and rate of evolution of the chosen gene and discusses its potential for phylogenetic st dy. This © issue also contains: “The Comparative Pollination and Floral Biology of Baobabs (Adansonia-Bombaca- — -ceae)" by David A. Baum and “In Memoriam: Peter G. Martin." Annals 82(2) 1995. 174 pages. MA US, A a O frien A REEL To order, please indices method of payment below. Checks or money- orders should be in U.S. funds, ; payable through a U.S. bank, to Missouri Botanical Garden. Orders must be prepaid. $3.00 handling — fee on all orders. Additional $2.00 handling fee will be added to orders requiring invoices. No shipments — are made until payment is received. Phone: (314) 577-9534, fax: (314) 577-9594. e-mail: - Wight aaah: 0 es iol ng Pre geet pee ped LI Dri irr Soe Eleven, Missouri Botanical Garden crt Please send - pas copy(ies) of Ann als 82(2) - $ - St Louis, MO 63166-0299, U.S.A. E MEE : O Check/money order enclosed mM de ex Send books te: n oy Dee ES -O Send invoice ($2.00 fee will be added to total) —— — E Charge card number (MasterCard/Visa) = —— Expiration date M T Oe P a + ear NE. CONTENTS A Revision of Cynanchum (Asclepiadaceae) in Africa — ————— Sigrid Liede Phylogeny and Speciation in Lapeirousia imm Lapeirousia (Iridaceae: Ixioideae) Bee Ginter Sy Peter Goldblatt & John C. Manning Phylogenetic Relitionthips; Seed AS and Dispersal System Evolution in Amaryllideae (Amaryllidaceae) .... 2 .. D. A. Snijman & H. P. Linder A Taxonomic Synopsis of the Genus Salsola a (Chenopodiaceae) i in North America -.- S Mo ie Che nad ua te LL TS a Sergei L. Mosyakin Ritmos ales do la SENSSA Taxonómica “de aes Yesidgies en México y una Estimación del Número de Especies Conocidas Rodolfo Diso & Calar Dind Diversity EN TE Flora of Fan Si Pan, the Highest Mountain in Vietnam ...........-.---—— guyen Nghia Thin & Daniel K. Harder Toward an Improved ftandihpaton ol Lama A ee _ Henk van der Werff & H. G. Richter The Use of Cuticular Features in the as of the Lauraceae .. ES Fe Sb bse D. C. Christophel, R. Kerrigan & p 1 Rowett ee P ECESO, Pastia c camerounensis subsp. camerounensis S. D. Manning, sP: nov., L i É Linda Ellis. - | a i AIT EA | Annals of the Missouri Botanical - Garden 1996 Y Number 4 Volume 83, Number 4 Winter 1996 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 out- side the Garden will also be accepted. 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. Editorial Committee Henk van der Werff Editor, Missouri Botanical Garden Amy Scheuler McPherson Managing Editor, Missouri Botanical Garden Diana Gunter Editorial Assistant, Missouri Botanical Garden Vicki Couture Secretary | Teresa Johnson i ; - Publications Order Processor 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 P. Mick Richardson Missouri Botanical Garden For subscription information contact Depart- . ment Eleven, P.O. Box 299, St. Louis, MO .. 63166-0299, Subscription price is $110 per . volume U.S., $115 Canada and Mexico, $135 ... all other countries. Four issues per volume. The journal Novon is included in the subscription . price of the ANNALS. * amcpher@admin.mobot.org (editorial queries) . deptl1@mobot.org (orders) __ http://www.mobot.org © Missouri Botanical Garden 1996 - The mission of the Missouri Botanical Garden is to discover and share knowledge about plan | their environment, in order to preserve and enrich life. : AE e ii BA oder o The ANNALS OF THE MISSOURI BOTANICAL GARDEN (ISSN 0026-6493) is published quar- terly by the Missouri Botanical Garden. 2345 Tower Grove Avenue, St. Louis, MO 63110. Pe- riodicals postage paid at St. Louis, MO and ibo ditional mailing offices. POSTMASTER: Send ad- dress changes to ANNALS OF THE MISSOURI BOTANICAL GARDEN, Department Eleven, Box 299, St. Louis, MO 63166-0299 EI VETS A aë ARMRICI /JAMIIC AZ 7520 405 10607 /DB, vi ^ ^, WEG WE PO. tsand — — A eere II O E: ” ee e m EEEE A —— a — ee EE EE AA Volume 83 Annals Number 4 of the NZ 1996 Missouri Botanical Garden ALWYN HOWARD GENTRY, 1945-1993: A TRIBUTE This issue of the Annals of the Missouri Botanical Garden is dedicated to the memory of Alwyn H. Gentry. The first section contains contributions by some of those scientists who knew him. It begins with an overview of his life by James S. Miller, including lists of Gentry’s publications and of his students. A transcript of the presentations given at the memorial service held at the Missouri Botanical Garden on 20 August, 1993, follows. The speakers were: Theo- dore M. Barkley, Hugh H. Iltis, Walter H. Lewis, Enrique Forero, Mark Plotkin, Oliver Phillips, and Ricardo Rueda. The tribute closes with the eulogy delivered by Peter H. Raven. ANN. Missouri Bor. GARD. 83: 433—460. 1996. 434 Annals of the Missouri Botanical Garden ALWYN HOWARD GENTRY, 1945-1993 Al Gentry with flowering branch of Tabebuia in Costa Rica. The Central American species of Tabebuia were the subject of his Master's Thesis, and Costa Rica was the first tropical country that he visited. Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute 435 THE LIFE AND WORK OF AL GENTRY On August 3, 1993, a small airplane on a Conser- vation International research mission crashed into an isolated mountain ridge in western Ecuador, near Guayaquil. Alwyn H. Gentry and Theodore A. Parker III, arguably the leading authorities on the botany and ornithology of tropical America, died in the crash along with the pilot and Eduardo Aspiazu, an Ecua- dorian ecologist and President of the Guayaquil chap- ter of Fundación Natura. Three other researchers sur- vived the crash. This event prematurely ended the career of one of the greatest plant explorers and bot- anists of all time. Gentry’s knowledge of neotropical plants was unsurpassed, and Parkers extensive ex- perience with Latin American birds mirrored Gentry’s knowledge of its plants. Anyone familiar with Gentry knew that he worked tirelessly, was ambitious, con- ducted ground breaking and insightful scientific stud- ies at a tremendous rate, and was passionate in his quest for understanding and conservation of the most diverse natural areas in the world. Al Gentry was born on January 6, 1945, in Clay Center, Kansas. He graduated from Clay County Com- munity High School in 1963, where he was both val- edictorian and outstanding science student, and he began his university studies at Kansas State Univer- sity in the fall of that year. He graduated in 1967 with two degrees, one a bachelor of arts in physical science and the other a bachelor of science in botany and zoology. He spent the summer of 1967 in Costa Rica, his first immersion in the tropics, on an Organization of Tropical Studies (OTS) Introductory Course. Here he began his work on Bignoniaceae, which was to remain his primary monographic interest for life. In the fall, Gentry moved to Madison, Wisconsin, to begin his master’s studies in the botany department with Pro- fessor Hugh H. Iltis. He spent the summer of 1968 at the University of Miami in a seminar on tropical botany, finished his thesis on Tabebuia (Bignoniaceae) in Central America in the fall semester, and graduated with an M.S. degree in January of 1969. With support from the Davis Fund of the University of Wisconsin, he left immediately for fieldwork in Costa Rica, Pan- ama, and Mexico from January to March. In the fall of 1969, Al began his doctoral studies at Washington University in St. Louis, Missouri, under the direction of Professor Walter Lewis. With support from a National Science Foundation pre-doctoral dis- sertation improvement grant, Gentry left for field stud- ies in Costa Rica and Panama in May of 1971. He returned to Panama in August, working for a year as curator of the Summit Herbarium, a position managed on a one-year rotation basis by the Missouri Botanical Garden and generally filled by graduate students working on the systematics of Central American plants. His Ph.D. thesis, completed for graduation in December of 1972, was entitled *An Eco-evolutionary Study of the Bignoniaceae of Southern Central Amer- ica. In October 1972, Gentry was hired as an Assistant Curator by the Missouri Botanical Garden, where he spent his entire career. In 1974 he made his first trips to South America, South Africa, and Madagascar, and late in the year he made his first trip to Peru, the country that would remain the major geographical fo- cus of his studies. During his career at the Missouri Botanical Gar- den, Gentry’s interests in systematics and ecology continually expanded to include a wide variety of top- ics, but most of his time and energy focused on three: the systematics of Bignoniaceae; tropical floristics, particularly the flora of Peru; and studies of the cor- relation between environmental variables and patterns of forest diversity and composition. Gentry became interested in the systematics of Big- noniaceae during his first trip to the tropics on the OTS course in Costa Rica in 1967. Although he was initially attracted to the brilliant-flowered genus Tab- ebuia and chose it as the subject of his master's stud- ies, his interest rapidly shifted to the vining genera that comprise the majority of the species. Throughout his life he maintained interest in the family, and he continued to publish profusely on it. His graduate studies led to his first ten publications on Bignonia- ceae, and he published the first part of a Flora Neo- tropica treatment of the family in 1980 and the second in 1992. In addition, he contributed treatments of the family for nine floristic volumes and had five treat- ments of the family in press at the time of his death. His interest in the family was not solely concemed with alpha taxonomy, however. He authored and co- authored studies on the ethnobotany, palynology, cy- tology, and chemistry of the family, and even as early as his doctoral studies he began investigating polli- nation mechanisms and the ecological importance of the family. At the time that Gentry joined the Missouri Botan- ical Garden's research staff, the group was deeply committed to completion of the Flora of Panama, an ongoing project that had begun with fieldwork in the 1920s and had been appearing as fascicles in the Annals of the Missouri Botanical Garden for nearly three decades. The 1970s at the Garden marked a decided effort to bring the project to completion, and Gentry played a major role, contributing the Flora of anama treatments for Bignoniaceae, Buxaceae, Hu- miriaceae, Rafflesiaceae, and Sabiaceae. He also made a significant contribution by routinely identify- 436 Annals of the Missouri Botanical Garden Š 2 Figure l. Gentry in Peru with € ing the steady stream of incoming collections, and he published notes and new species in eight additional families. This had a variety of affects on Gentry. First. it was the beginning of the development of what were to become unequaled skills in general plant identifi- cation, Second, he began his studies of the other plant groups, such as Sabiaceae, Buxaceae, and Passiflor- aceae, that continued to remain secondary interests in later years. He also became adept at assembling floristic information and acquiring the basic skills that later enabled him to coauthor two florulas in Ecuador, Río Palenque with €. H. Dodson, and Jauneche with C. H. Dodson and F. M. Valverde, and one of the Chocó, Colombia, with Enrique Forero. However, it was Peru that ultimately became Gen- try's primary floristic focus. After his first trip in 1974, he returned there in 1976 and at least once a year for every year until his death, making 33 trips. He published extensively on the Peruvian flora and de- veloped collaborative relationships with many of his Peruvian colleagues, particularly Rodolfo Vasquez and Camilo Díaz. Gentry became famous for his use of transect sam- ples to assess vegetation. He collected his first tran- sects as part of his OTS training in 1967, probably with little more in mind than comparing two adjacent vegetation types, but perhaps his greatest research arose from this auspicious start. As part of his doctoral e uo O amilo Díaz and Rosa Ortiz de Gentry. studies, Gentry laid out transects at Madden Lake, Pipeline Road, and Curundu in Panama, this time to document the importance of Bignoniaceae as the ma- jor liana family in tropical forests and to study species phenology and pollination. It was not until later that he began to view the transects as a tool that allowed a relatively quick method of measuring diversity, tax- onomic Composition, and structure of forests. Up until he began his studies, no one had ever collected a data set with any geographically signifi- cant coverage. Most of the ecological sampling that had been done used a variety of different methods and was thus not comparable from site to site. Gen- trys brilliance was to devise a method that allowed a representative sample to be captured in a brief period of several days. This, coupled with his ever-increasing ability to identify tropical plants, particularly from vegetative characters, allowed him to amass a data set that, while with some limitations and assumptions, was comparable from site to site. Of course, another important character was that Gentry simply knew no limits. There are few people who would embark on explaining patterns of diversity and taxonomic com- position of forests on a worldwide basis expecting to collect all of their own data. Gentry, trained as a tax- onomist, took on an ecological problem that had been beyond the imagination of the ecological research community, Volume 83, Number 4 Alwyn Howard Gentry: A Tribute The method involved sampling all of the woody species greater than 2.5 cm diameter at breast height (DBH) in one-tenth of a hectare. He stretched a 50-m line through the forest and sampled all plants within one meter of either side. This he repeated ten times, and the sum of ten 2 X 50-m transects totaled 0.1 ha. With his knowledge of tropical plants, Gentry could rapidly make decisions about how many indi- viduals encountered represented new species. Any- thing questionable was vouchered for later study in the herbarium, and each species was vouchered at least once. The final analysis could only be completed after all of the specimens had been identified. Thus even in the richest forests, he could collect the site in a week, and with another week in the herbarium, complete his analysis. The vast majority of plants encountered at a given time in any forest that Al censused were sterile. Iden- tifying them thus required either the effort of multiple visits to a given site to obtain the majority of species in fertile condition or the ability to accurately identify sterile specimens. Gentrys bold determination led him to the second option; this required that he de- velop and refine the capacity to work with sterile specimens. For years he carried a notebook with a draft key to families and genera of neotropical genera based on vegetative characters, and as he encountered new taxa he continually revised his text. This effort led to Gentry’s remarkable capacity to identify neo- tropical plants from all sorts of specimens, and to the publication of his Field Guide to the Families and Genera of Woody Plants of Northwest South America. Gentry’s contributions to science were phenomenal. His more than 200 published works, with many ad- ditional articles still in press, cover a wide variety of subjects. He radically changed our modern concepts of the classification of Bignoniaceae and contributed to our knowledge of many other families as well. The publications resulting from the data that Gentry and his colleagues collected from 226 different 0.1-ha transects revolutionized the study of patterns of plant diversity. The transect studies that Gentry began as a graduate student expanded to more than 200 sites worldwide, and the focus of the studies shifted from attempting to study the importance and phenology of ignoniaceae to attempting to understand patterns of plant diversity and the taxonomic composition of for- ests. At the same time that he amassed such an im- pressive publication record, however, he also had an enormous impact as an explorer, plant collector, and teacher. Gentry as Explorer/Plant Collector. Gentry made more than 80,000 plant collections, and hundreds of new species have been based on these specimens. Many were collected in the course of ground-breaking expeditions to remote regions of tropical America, some of them never before visited by botanists. Gen- try’s stamina and enthusiasm on these expeditions, as well as his near obliviousness to physical discomfort, were legendary. On one occasion (recounted in A Par- rot Without a Name, by Stap, 1990), Al and Camilo Diaz became lost for over three days in a remote lo- cation in the Peruvian Amazon, finally re-emerging back at camp haggard and disheveled. Despite having wandered without food or shelter (they survived on nuts, snails, and, finally, palm heart, retrieved from a tree Al cut down with his jacknife), Al paused only long enough for a brief meal before resuming collect- ing, protesting over the objections of fellow expedition members that now he would have to work extra to make up for the time lost. Even those who worked with him under more mundane circumstances can at- test to his tireless persistence, having suffered through late nights of pressing plant specimens amassed dur- ing long and strenuous days of collecting. But that tirelessness was a reflection of Gentry’s belief in the fundamental importance, even urgency, of under- standing tropical forests. How could one rest when there were so many marvels to be uncovered, so many undescribed species, so many vanishing before they could even be described? Gentry as Teacher. Despite his notoriously frenetic schedule, Gentry found time—and indeed, considered it his mission—to encourage and instruct promising young botanists. Over the course of his career, he supervised the theses of about 20 students. At the time of his death, he was supervising nine graduate students, and many more had applied or were plan- ning to apply to graduate schools in St. Louis, spe- cifically so they could work with Gentry. Many of these were from the tropical countries he so loved. Perhaps one of Gentry’s greatest legacies is his in- spiring and launching the careers of a cadre of the next generation of tropical biologists (see list of stu- dents following publications). At the time of his death, Al Gentry was involved in a myriad of projects. His monographic work on neotropical Bignoniaceae had continued, although the manuscripts had not yet been written, and he was well into a study of Malagasy members of the family. The database that he had developed contained records for some 55,000 herbarium collections he had examined from more than 122 herbaria around the world. Wil- liam G. D'Arcy has assumed primary responsibility for the completion of Gentry’s studies of Bignoniaceae, and Warren D. Hauk has joined the Garden staff to help with this task. This work will include completion of revisions for those genera that were not treated in the two volumes of Bignoniaceae Gentry contributed 438 Annals of the Missouri Botanical Garden to the Flora Neotropica. Gentry had envisioned two additional volumes for treatment of the family. He also left an unpublished manuscript treating the Bignoni- aceae of Colombia, which will be published in the ra olombia series. During his career Gentry assembled the world's richest herbarium collection of Bignoniaceae, including his own collections and those he received as gifts from other institutions. These col- lections and his database are readily available; the Missouri Botanical Garden encourages other research- ers to study and use these resources. A “Gentry In- vitation Series” has been established for taxonomic publications by people who are relying heavily on ma- terial left by Gentry or who wish to pay tribute to the contributions he made to their work. Those interested should contact W. G. D'Arcy at the Missouri Botanical Garden. The transects occupied an ever increasing per- centage of Gentry’s time and were undoubtedly the part of his work he planned to focus most on in the future. In 1990, the Rapid Assessment Program (RAP) Team was founded at Conservation Internation- al, and Gentry began working half-time with Louise Emmons, Robin Foster, and Ted Parker to conduct quick inventories of areas deemed to be of great con- servation significance. Gentry's 0.1-ha transects fit the AP Team philosophy perfectly, and his participation helped serve his need for even more opportunities to install transects and collect data from an ever increas- ing number of sites. | Gentry certainly never imagined that the data that he and his collaborators had amassed from 226 0.1-ha transects were, in fact, ready to be summa- rized. At the time of his death, data from only about 90 of his sites had been computerized, and these were in drastic need of revision to include names of newly identified plant specimens. Gentry generally worked from hand-calculated summaries of each site. Due to advances in software over the long course of his work on transects, the data that had been computerized was transferred more than once from one program to an- other, resulting in data sets in a variety of formats and often with extraneous information needing removal. At the Garden, Oliver Phillips undertook the difficult task of reorganizing previously computerized data and insuring that information from the other sites was en- tered into the same format. All of the data was either entered or converted to consistent computerized for- mat by Nancy Hediger. The Garden will make all of the transect data avail- able via the World Wide Web (http://www.mobot. org), so that researchers can either access information from individual sites or download regional sets or the entire collection. Oliver Phillips and I are preparing a book to present the results of the data analysis, with an introduction covering the historical background and the rationale and methodology of the transect data. The bulk of the book will consist of single-page data summaries for each transect site, and the volume will be published in the Monographs in Systematic Botany rom the Missouri Botanical Garden and entitled Glob- al Patterns of Plant Diversity: Alwyn H. Gentry’s Forest Transect Data Set. It will serve as a companion vol- ume to the computerized set of transect data available on the World Wide Web. The list of Gentry’s publications that follows is as complete as possible. However, at the time of his death, he was actively involved in a large number of collaborative projects, and many of his collaborators continue to include him as a coauthor. He also had contributed treatments of various taxa to a variety of floristic projects that have yet to be published. Thus, his list of publications will continue to grow beyond the number included below, and will include contri- butions in the Gentry Invitation Series. To document the extent of Gentry's involvement in the training of botanists, a list of all his students and their projects, compiled by W. G. D'Arcy, follows the publications list. Literature Cited Stap, D. 1990. A Parrot Without a Name. Alfred A. Knopf, New York Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute Alwyn H. Gentry: Publications Gentry, A. H. comparison of some leaf charac- teristics of ME dry forest and pum wet forest in Costa Rica. Turrialba 19: 4194: Gentry, A. H. 1969. Tabebuia: The tortuous history of a cita name. Taxon 18: 635-652. Gentry, A. H. 1970. A revision of Tabebuia (Bignonia- ceae) in Central America. Brittonia 22: 246-264. Gentry, A. H. 1971. Studies in Bignoniaceae III. Two new Panamanian Ba rege of Bignoniaceae. Ann. Missouri Bot. Gard. 58: 93-96. Gentry, A. H. DE Note on Gibsoniothamnus. Fieldiana, Botany 34: 55. Gentry, A. H. 1971. Bignoniaceae of OTS lowland study sites. Section 26-16 in C. E. Schnell (editor), Handbook Bignoniaceae of S ia D. Thesis, Washington University, St. Louis, Missou- em A. H. 1972. Vb au (Bignoniaceae): A cri- tique. Taxon 21: 113-1 H. 1972. The type species of Bignonia L. 664. : 659 Gentry, A. H. 1973. Studies in Bignoniaceae VII. Den- drosicus, Enallagma, and Amphitecna. Taxon 22: 637— 640 Gentry, A. H. 1973. Generic delimitations of Central American tege Brittonia 25: 226-242, Gentry, A. H. 1973. Rafflesiaceae. /n: Flora of Panama. Ann. Missouri Bot. Gard. 60: 17-21 Gentry, A 1973. A new species of Proteaceae from Panama. Ann. Missouri Bot. Gard. 60: 571-572. Gentry, A. H. 1973. Schlegelia ari pde 2 familial Gentry, A. H Studies in Bignoniac ie New species of Dal and Pachyptera. Phytologia 26: 84 Gentry, A. H. 1973 [1974]. s In: ad of Panama. Ann. Missouri Bot. Gard. 60 Gentry, A. H. 1974. Flowering phenology ft TAS in tropical Bignoniaceae. Biotropi Gentry, A. H. 1974. Studies in ey: ae XI: A syn- opsis of the genus Distictis. Ann. Missouri Bot. Gard. 61: 494-501. Gentry, A. H. 1974. Studies in Bignoniaceae XII: or noteworthy species of South po y i eae. Ann. Missouri Bot. Gard. 61: 872-88 Gentry, A. H. 1974. Key to the genera Lu Guatemalan Bignoniaceae. /n: Standley & L. Williams, Bignonia- ceae in Flora of Guatemala. Fieldiana, Bot. 24(10): 155-158. Gentry, A. H. 1974. . Coevolutionary patterns in Central Americ i ] 6— an Bignoniaceae. Ann. Missouri Bot. Gard. 61: 9 759. Gentry, A. H. 1974. Gibsoniothamnus toas) in Panama. Ann. Missouri Bot. Gard. 61: Gentry, A. H. 1974. Notes on gr CB Ann. Missouri Bot. Gard. 61: Gentry, A. H. . Raven. ipe The Missi Botan- ical Garden. Pl. is. Bull. 20: 34-38. m A. H. 75. Humiriaceae. /n: Ann. Missouri Bot. Gard. 62: 35-44. Gentry. . Bignonia crucigera: A case of mis- id desk ‘Reon 24: 121-123. Flora of Panama. Gentry, A. H. 1975. Additional Panamanian Myristica- ceae. Ann. Missouri Bot. Gard. 62 4-479, Gentry, A. H. 1975. Studies in Bignoniaceae 17: Kige- lianthe: A synonym of Fernandoa (Bignoniaceae). Ann. Missouri Bot. Gard. 62: 480—483. H. 1975. AR of Vellozo's Bignonia- 975. Ballarni bad -Dendrosicus: ifferent view of proposals 46, 47, and 48. Taxon 24: 9 391-392. Gentry, A. H. 1976. Amphitecna-Enallagma-Dendrosicus revisited. Taxon 25: 108. ag A. H. 1976. Studies in Bignoniaceae 18: Notes n S. Moore's Mato Grosso Bignoniaceae. Ann. Missouri Bot. Gard. 63: 42-4 Gentry, A. H. 1976. Budi in Bignoniaceae 19: Addi- tional new or noteworthy Sanh American Bignoniaceae. Ann. Missouri Bot. Gard. 6 0. Gentry, A. H. 1976. A new Panamera Sterculia with taxonomic notes on the genus. Ann. Missouri Bot. Gard 63: 370-372 Gentry, A. H. 1976. Bignoniaceae of southern Central erence: Distribution and ecological specificity. Biotro- -131. ceae. ei “Missouri Bot. Card. 63: 341-345. Gentry, A. H. 1976. Relationships of the Madagascar Bignoniaceae: A striking case of convergent evolution. Pl. Syst. Evol. 126: 255-266 Dodson, C. H. & A. H. Gentry. ` 1977. Epiphyllum ji lanthus (L.) Haw. (Cnciacone) and its allies in Ecuador. Selbyana 2: 30-31. Dodson, C. H. & A. H. Gentry. 1977. New species in the Urticaceae and SE from the Rfo ues: Sci- ence Center, oa Selbyana 2: 7. Endangered plant species and hab- 6-149 in G. Prance & T. Elias, Extinction is Forever: New York Botanical Garden, Bronx, New York. Gentry, A. H. 1977. A new and (Bignoniaceae) from p and Peru. Ann. Missouri Bot. Gard. 63: 8- Gentry, A. A 1977. Phyllarthron bilabiatum: A new spe- i agascar. Ann. Missouri Ne ew species of Gibsoniothamnus Serophulracaignaniacean and Tournefortia (Bor- rn Panama and the Chocó. Ann. Missouri Bot. Card. ot Latas. 77. Studies in Bignoniaceae 25: New sens and combinations in South American Bignoni- aceae. Ann. Missouri Bot. Gard. 64: 311—319. Gentry, A. H. 1977. Studies in Bignoniaceae 26: New taxa and combinations in northwestern South American New species of Leguminosae, Lau- raceae, Monimikcede, and new combinations in Bignoniaceae from western Ecuador. Selbyana 2: 39— 45. Gentry, A. H. 1977. Notes on Middle American Bigno- niaceae. Rhodora 79: 430—444. Gentry, A. H. 1977. New records of Apocynaceae yk Panama and the Chocó. Ann. Missouri Bot. Gard. 6 320-323 440 Annals of the Missouri Botanical Garden ui A. H. 1977 -172. Bignoniaceae. /n: Flora of Ecuador Dodson. ¿ A. H. Gentry. 1978. Heliconias (Mu- saceae) of the es 2 RUNS Science Center, Ecuador. aga 2i rg Dodson, C. . Gentry. dy Flora E e ud Palenque Scie lence ‘Centex Selbyana 4: i-xxx; Gentry, A. H. 1978 Dives dde + e regeneragao a poeira do INPA, com referéncia especial as le ceae. Acta Amazonica 8: 67-70. Gentry, A. H. 1978. Floristic knowledge a needs in acific p eres America. Brittonia 30: 53. Gentry, A. H. 19 8. Bignoniac eae. 3e Botany ] the Gua- yana Hilo. Mem. sard. 29: 245-283. Gentry, A. H. Buxac eae. a rae of ene Ann, Missouri 2 (oun 65: Gentry, A. H. 1978 Ma pallinatom for mass-flowering plants? Biotropica 10: 68-69. Gentry, A. H. 1978. Saisie in Bignoniaceae 31: New species ^on combinations from Mpeg dg and Brazil. Missouri Bot. Gard. 65: 725-7: Gentry, re H. new Pidas (Theac us Rom the Panama/Colombia border. Ann. Missouri Bot. Gard. 65: 773-7 Kaastra, 'R C. . H. Gentry. 1978. A new Erythrochi- ton (Rutac one from Ecuador. Selbyana 2: 287-2 Gentry, A. H. 1979. rea and conservation of plani species in tropical America: A prp 'al per- spective. Pp. 100-126 in n Hedbe erg (ec or), System- atic Botany, Plant Utilization, and I iem ere Conser- vation. Almqvist & Wiksell "sire Stockholm. Gentry, A. H. 1979, Distribution patterns of neotropical Bignoniaceae: Some phytogeographical implications. Pp. 339-354 in K. Larsen & I. B. Holm-Nielson (ed- itors), Tropical Botany: Academic Press, London, New i ork, Gentry, A. H. 1979. Transfer of E speci ies of Dialyanth- era to Otoba bs aceae). Taxon 27: Gentry, A. H. 1979. irme e 2 in Big- noniaceae. Ann, jd Bot. Gard. 66: 778—787 Gentry, A. H. & A. S. Tomb. 1979. ee implica- tions p M eas eae palynology. Ann. Missouri Bot. Gard. € X111. Goldblatt. T à A. H. Gentry. ra Cytology of Bigno- niaceae. Bot. Not. Ts 475-4 Kinzey, W. H. & A. H. Gentry. P 79. Habitat utilization in two species of Callicebus. Pp. 89-100 in R. Sussman (editor), Primate Ecology: Problem- a Field Stud- ies. Wiley & Sons, New Yor Gentry, A. H. 1980. TE 3 in Bignoniaceae 37: New species of Bignoniac eae from eastern South America. Phytologia 46: 201-215 Gentry, A. H. 1980. Bienonixe eae, Part 1 (Crescentieae and a Flora Neotropica Monograph No. 25. (EE 15 Ge toes " H. 1980. The m of Peru: Fieldiana, Bot. n.s. 5: 1— Gentry, . H. 1980. Sim species of Apocynaceae, Big noniaceae, Passifloraceae, and Piperaceae from nies Colombia ana Eeuador Phytologia 47: 97-107. Gentry, A. H. 1980. Sabiaceae. Flora of Panama. Ann. Missouri Bot. eg 67: 949— Gentry . López- See 1980. ds and inc asad owing of the Upper Amazon. Science 210: 1354-1356 Gentry, A. H. in Palmae, Araliaceae, Apocynaceae, and Bignoniaceae A conspectus. 1981. New species and a new combination from the Chocó and Amazonian Peru. Ann. Missouri Bot. Gard. 68: 112-121. entry, A. H. 1981. Distributional patterns and an ed ditional species of the Passiflora vitifolia complex azonian species diversity oh to pride ally enn: ated communities. Pl. Syst. : 95-10 Gentry, M 1. 981. a eds 'o de amans ivas de Conservación. 3644 in T. patada G. (editor), Seminario sobre Proy- ~ : w species of Myristicaceae, Com- bretaceae, and U rticaceae from coastal Colombia and Ecuador. o 48: 233-237. sentry, A. H. & R. Foster. — A new Peruvian Styl- oceras atis Discovery of a Apr A Fr. missing link. Ann. Missouri ‘Bot Gard. 68: 122-124. c el Prance (editor), Biological Divemitea ‘ation in the Trop- ics, Columbia d Press, New Yor Gentry, A. H. 1982. New or noteworthy spec ies of Middle American TE eae. Wrightia 7: 8: Gentry, A. H. 1982. Samhain In: Flora de Veracruz 24: 1 = x DI Gentry, A. H. 1982. s of neotropical plant species diversity. e Biol. 15: 1-84 entry, A. H. 1982. n culti ied species of Tabebuia with notes on de e un Misi eae. Pp. 52-79 in Proc. 3rd Menninger Flowering Tree Conference. Gentry, A H. 1982. fNeotsie al floristic diversity: Phy- togeographical connections between Central and South America, Pleistocene cumani fluctuations, or an acci- dent of the Andean orogeny? Ann. Missouri Bot. Gard. 9 O 9: Bad 3. Emmons, L. H. & A. H. Gentry. 1983. Tropical forest ructure oe the distribution of gliding and prehensile- tailed eres Amer. Naturalist 121(4): 513 Gentry, A. H. 3. Bignoniaceae. /n: Flora de Venezue- la. VII, sis js 7433. Fundación Instituto Botánico de Venezuela, Cara Gentry, A. H. 1983. Plagioceli (Ulmaceae)— fluous genus. Taxon ¢ : 460. H. 1 A super- Gentry, A. . A new combination for a problematic aes d 'an Apocynaceae. issouri Bot. Gard. 70: 205. Gentry, N z 1983. Alstonia (Apocynaceae): Another Pa- laeotropical ani in Central America. Ann. Missouri Bot. Gard. 70: 2 Gentry, A. H noniaceae. Sonderb. 99 083, Dispersal and distribution in Big- Naturwiss. Vereins Hamburg 7: 7-199. Gentry, A. H. 1983. pene ec ology and diversity in neotropical forest communities. Sonderb. Naturwiss Vereins Hamburg 7: 303— 31 Gentry, A. H. 1983. Lianas gail the “paradox” of con- trasting latitudinal ary in n and litter produc- tion. Tropical Ecology 24: 6 Gentry, A. H. 1984 n ies from Jauneche, Ec ados Inga jaunechensis (Leguminosae and Annona hystricoides (Annonaceae). Phytologia 54: 475-478. Gentry, A. H. 1984. Klainedoxa biomas ae) at Mako- kou. Gabon Three sympatric species in a putatively w Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute 441 monotypic genus. Ann. Missouri Bot. Gard. 71: 166— 16 . 1984. Bignoniaceae. /n: Flore du Came- 6-04. 84. New species and combinations in Apocynaceae from Peru and — Amazonia. Ann. Missouri Bot. Gard. 71: 1075-10 984. The demise of the bcn rain for- mental rss 84. The cultivated species of Tabebu. Florida Nurseryman 31: 8-10. Gentry, A. H. 1984. An overview of nig i phyto- geographic patterns, with an emphasis on Amazonia. Pp. 19-35 in 1? Simpósio de Trópico Hamide Proceed- d Vol. II. Departamento de Difusáo de Tecnologia, Brasilia D.F., Brazil. Gentry, A. H., auta & E. de S. F. da Rocha. 1984. Ta bebuia mia. (Vahl) Nicholson, ipé-amar- elo (Bignoniaceae) no símbolo da Sociedade Botánica do Brasil. Atas Soc. Bot. Brasil, secc. Rio de Janeiro 2: -~ 77-80. Gentry, A. H. & K. Cook. ceae): A widely used eye medi Eos of Sout rin 11: 337 We .H. 1984. Martinella epi America. Forero, E. Gentry. 1984. New w phanerogam spe- cies from Fut Colombia eel 55: 365-371 Dodson, Gent Valverde. 1985. lora of Tannache. Los Rios, Ec adt Banco Central del Ecuador, d Gentry, A. 1985. Studies in Bignoniaceae 48: New South American species of Bignoniaeae. Phytologia 57: 248. 1985. An ecotaxonomic : dne of Pana- A2 i Arcy & M. Correa A. (editors), The Botany and Mo History d Panama: La Botánica e Historia Natural de Panamá. Monogr. Syst. Bot. Missouri Bot. Gard. 10. Gentry, A. H. 1985. Contrasting Mice deaur patterns of ae ae yes and Panamanian plants. Pp. 147- 160 in W. G. & M. eie (editors) The Botany and M n of Panama: La Botánica e Historia Natural de Panamá. Monogr. Syst. Bot. Missouri Bot. Gard. Gentry, A. H. 1985. ud distribution and diversity pat- erns in Amazonia: . Nat. Geogr. Soc. Res. Rep 1979: 245-252. Gentry, A. H. 1985. Plants. Pp. 67-75 in the Biosphere atalogue. Gentry, A. H. 1985. Bignoniaceae. /n: Flore de Gabon 27: 19-61. 985. Algunos resultados preliminares de estudios botánicos en el Parque Nacional del Manu. Pp 2 in M. Rios (editor), Reporte Manu. Universi- dad Agraria, La Molina, Peru qum A. H. 1986. Exploring the Mountain of the Mists. p. 124-139 in Science Year: The World Book Science ns ual. Geotry, A A. H. . Notes on Peruvian Palms. Ann. Mis- souri Bot. Tu 13: 158-165. Gentry, 5 E 1986. Endemism in tropical vs. tempera munities. Pp. 153-181 in M. Soulé pd ee Biology. Sinauer Press, Sunderland, Mas- sachusetts. Gentry, A. H. 1986. Sumario de patrones fitogeograficos neotropicales y sus hop aciones para el desarrollo de la Amazonia. e cad. Col. Cienc. 16: 101 -116. Gentry, A. H. 1986. "pets richness and floristic com- position of Chocó region plant communities. Caldasia 15: 71-91 Gentry, A. H. 1986. New neotropical species of Meliosma (Sabiaceae). Ann Missouri Bot. Gard. 73: 820-824. Gentry, A. H. 1986. An overview of neotropical phyto- geographic patterns. Proc. of Ist simposio do Tropico Humedo, Belem, Brazil 2: 19-35. Gentry, A. H. 1986. Rasgos fitogeoleráficos del Neotró- picos: Implicaciones en la conservación del medio nat- ural en Ecuador. Cultura 8: (24). Gentry, A. H. € R. H. Wettach. 1986. Fevillea—A new oilseed from Amazonian Peru. Econ. Bot. 40: 177-185. (editors), Priorités en Matiére de Conservation des Es- pèces à pEr ar. IUCN, Gland. Gentry, A. H. & C. H. Dodson. 1987. Contribution of non- -tree es to species richness of tropical rain forest. Bio- entry, A 1987. Diversity and bio- geogra aphy of neotropical vascular epiphytes. Ann. Mis- souri Bot. Gard. 74: 233. Gentry, A. H. & L a 1987. Geographical vari- ation in iy a nd composition of the understory of neotropical forests. Biotropica 19: 216-227. entry, A. H. . Steyermark. 1987. A revision of Di- lodendron (Sapindaceae). Ann. Missouri Bot. Gard. 74 —538. Vasquez, R. & A. H. Gentry. 1987. Limitaciones del uso e nombres vernaculares en los inventarios forestales de la Amazonia Peruana. Rev. For. Per. 09-120. Forero, E. & A. H. Gentry. 1988. Noctiópical piena dis- tribution patterns with emphasis on northwest South America: A preliminary overview. Pp. 21—37 in LP Van- zolini & W. Heyer (editors), Proc. Workshop Neotropical "v3 Patterns. Acad. Bras. Cienc. Rio de Janei- Gentry, A. H. 1988. Tree species richness » Upper Am- azonian forests. Proc. Nat. Acad. Sci E 56-159. Gentry, A. H. 1988. Changes i in plant co co sity and i. composition on environ Gent Madagascar Bignoniaceae. 185 in P. Goldblatt & P. P. Lowry (editors), Modern Systematic Studies in ETE Botany. Monogr. Syst. Bot. Missouri Bot. Gard. e A. H. New species and a new combination for plants En E South America. Ann. Mis- souri Bot. Gard. 75: 1429-1439. . 1988. Where have all the tropical forest resource. Forest Ecol. Managem. 22: 73— 76. — M, D. Kindack, B. A. vere d -C. Ethier, D. V. Awang & A. H. Gentry. 1988. Na quinone con- $m of Tabebuia spp. J. Natural Tol 51(5): 1023-1024. Forero, E. & A. Gentry 1989. Lista anotada de las plantas el Departamento del Chocó, Colombia. Bibl. José Je- rónimo Triana 10: 1-142. Gentry, A. 1989. Northwest South America (Colombia, Ecuador, Peru). Pp. 391—400 in D. Campbell & H. D. Hammond (editors), Floristic Inventory of Tropical Countries. New York Botanical Garden, Bronx, New York 442 Annals of the Missouri Botanical Garden , A. H. 1989. Three new Hispaniolan species of bebía, Moscosoa 5: Gentry, A. H. 1 phiques du domaine néotropica uelques sinaditiss phytogéogra- urs implications pour e e ae en Equateur. Equateur 1986, vol. l: 1 Ed. ORSTOM, Coll. Colloques et Semi- e. ate 1989. Lista Anotada de las Plantas del 1-142. Inst. Ciencias Naturales, Univ. Nac. Col., Bo- gota. Gentry, A. H. 1989. Speciation in tropical forests. Pp. 113-134 in 7 Holm-Nielse Balslev editar): Tropical Forests: Botanical a Speciation, and Diversity. Academic Press, London. Gentry, A. H. 1989. A new species of Allomarkgrafia and notes on the genus. Ann. Missouri Bot. Gard. 76: 923- 92 entry, A. H. 1989. Diversidad del bosque tropical vs. desarollo: o 0 Pate erp Pp. 1-7 in Mem. 5 y io Di y Madera, Universidad del idad Floristica y dea ica de la VR Pp. 65-70 in G. García Li. al (editors), Investigación y Manejo de la Am Peters, C. M., A. H. Gentry & R. O. endalok, 1989. Valuation of an Amazonian forest. Nature 339: 655— 656. Vasquez, R. & A. H. Gentry. 1989. Use and misuse of forest- harvested fruits in the Iquitos area. Conservation Biol. 3: l-12. Blaney, C. L. . H. Gentry. 1990. Have Your Forest and Fat It Tow, Pores ld (Summer 1990): 40-47, Gentry, A. H. 1990, Solinginhila (Bignoniaceae), a new genus from the Paraguayan Chaco. Syst. Bot. 15: 277 279, Gentry, A. H. 1990. A new species of Vantanea (Humi- riaceae) from Amazonian Peru. Contribution to the s of the flora and vegetation of Peruvian Amazonia. XX. pacc - 379—380. Gentry, A. H. d patterns in neotropical Bignoniaceae. ee Prance & Gottsberger (editors), Modes of Pise: tion and Evolution of Woody Angio- sperms in Tropical Environments. Mem. New York Bot. Gard. 55: 118-129. T Gentry, A. H. (Editor). 1990. ded O ‘al Rainfo- rests. Yale Univ. Press, New Gentry, A. H. 1990. Floristic lid and differences between southern Pra a ‘a and upper and cen- tral Amazonia. Pp. 141— in A. Gentry (editor), Four Neotropical Forests. Yale Univ. Press, New Haven. Gentry, A. H. 1990. Tropical forests. Pp. 35-43 in A React eth E and Ecology of Forest Bird :ommunitie Academic Publishing, Netherlands. Gentry, A. H. a . Herbarium taxonomy vs. field knowl- edge. e there an ee solution? Flora Malesiana Bull. Gentry, " A a Selva Humeda de Colombia. Pp. 1— 198. Villegas Editores Bogotá. Gentry, A. H. . Blaney. 1990. Alternative to de- struction ae P biodiversity of tropical forests. Western wii 16(1): 2-7. & J. Terborgh. 1990. Composition and dy- namics a the Cocha Cashu mature floodplain forest. Pp. 542-564 in A. Gentry (editor), Four Neotropical Rain- forests. Yale Univ. Press, New I Peixoto, A. L. & A. H. Gentry. aven. 1990. Diversidade e com- posição florística da mata de tabuleiro da Reserva Flo- restal de Linhares (Espfrito Santo, Brasil). Rev. Bras 19-25. a J. eal, C. E. Jarvis & A. H. Gentry. 1991. On the alie ium of Bignonia cruc de 7 ae noniaceae). n nn. Missouri Bot. Gard. 7 Dodson, C. H. . H. Gentry. 1991. Bd. in lion in western Ec 'uador. Ann. Missouri Bot. Gard. 78: 273-295 esito les: D. & A. H. Gentry. 1991, The struc- and diversity of rain forests at Bajo Calima, Chocó Region, western Colombia. Biotropica 23: 2-11. Gentry, A. H. 1991. The distribution and evolution of climbing plants. Pp. 3-49 in J. Putz & H. Mooney (ed- itors), The Biology of Vines. Cambridge Univ. Press, idge Gentry, A. H. 1991. o and dispersal systems of lianas. Pp. 393-423 in z & H. Mooney (editors), The a of Vines. Cambridge Univ. Press, Cam- bri onan p H. 199]. A iden A usen or po le? Pp. Vicente Rodríg ánchez Páez (editors), Me- morias del E bd Ecobios Colombia 88. INDERENA. entry, A. H. 1991. El Jardín del Mundo. Credencial (Colombia) 9 v 42-52 Gentry, A. H. .. Bignoniacene, Pp. 71-73 in M. M. Fuiza de Me lo et fi (editors), Flora ai da Ma do Cardoso. Vol. 1. Instituto de Botánica, Sáo Pau- ~ Tropical forest diversity vs. devel- 169-185 in J. ~ TN A. H. 1991. Vegetación del Bosque de Niebla. Pp. 12-51 in Cristina Uribe Hurtado (editor), Bosques de Niebla de Colombia. Banc '0 de Occidente Creden- cial, Santafé de Bogotá, D.C Gentry, A. H. 1991. Tropic a forest diversity vs. sien opment: Obstacle or opportunity? Proc. ITTO Sy sium on Alternative Products from Tropical er ad Gentry, A. H. 1991. T s. devel- opment: obstacle or opportunity? Diversity 7(1-2): 89— 90. Hegarty, M. E. Hegarty & A. H. Gentry. rie Secondary ae in vines with an emphasis those with defensive functions. Pp. 287-310 in J. Putz & H. Mooney (editors), The Biology of Vines. Cam- bridge Univ. Press, Cambridge. Parker III, Foster, L. H. _Emmons, A. H. Gentry, S. Bec 5, Estenssoro & E H nojosa. 1991. A biologic a TEA of the Alto Madidi region and ad- jacent a of northwest Bolivia. /n: T. A. Parker III $ B. Bailey pene RAP Working Papers 1. Conserva- tion International, Washington, Gentry, A. H. 1992. A synopsis of Bignoniaceae ethno- botany and economic botany. Ann. Missouri Bot. Gard. 19: 53-04. sentry, A. H. 1992. Bignoniaceae—Part II (Tribe Teco- meae). Flora Neotropica Monograph 25(2): 1-370. Gentry, A. H. 1992, Tropic ‘al forest biodiversity: Distri- butional Xin and their conservational significance. Oikos 63: 8. Gentry, A. H. p 992. Exarata (Bignoniaceae), a new genus from the Chocó region of Ecuador and Colombia. Syst. Bot. 17: 503-507. Gentry, A. H. 1992. Four new species of Meliosma (Sa- biaceae) from Peru. Novon 2: 155-158. sentry, A. H. 1992. Six new Am dn - adi eae from Upper Amazonia. Novon 2: ~ 2 Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute 443 Gentry, A. H. 1992. Distribution patterns of Central pe and West Indian Bignoniaceae. Pp. 111-125 S. P. Darwin & A. L. Welden d udi er oí Mn a Tulane University, rle Gentry, A. H. . New species of ak "os from Amazonian Peru. Novon 2: 333—338. Gentry, A 992. Diversity and floristic composition of Andean forests of Peru and adjacent countries: Im- Fre for their conservation. Mem. Mus. Hist. Nat., U.N.M.S.M. (Lima) 21: 11-29 Gentry, A. H. 1992. New non-timber forest products from western South America. Pp. 125-136 in M. Plotkin & L. Famalore (editors), Sustainable Harvest and Market- ing of Rain Forest Products. Island Press, Washington, Góntry, À A. H. & R. Ortiz. 1992. A new species of Aptan- dra (Olacaceae) from Amazonian Peru. Novon 2: 153- 1 Paker III, T. A., R. B. Foster, L. H. Emmons, A. H. Gentry, J. L. Carr, L. Albuja V., A. Almendáriz, C. Jos- se. P. Yanez & A. Luna. Status of forest remnants in the RAP Working Papers 2. Conservation International, Washington, D.C Steyermark, J. A. & A. H. Gentry. 1992. Sabiaceae. Flora de Venezuela 5(1): n Fundación Instituto Bo- tánico de Venezuela, eun Appanah. S., A. H. Gentry & J. X Lafrankie. 1993. Lia diversity and species prs of Malaysian rain Gal J. Tropical Forest Science 6: 116-123. Gentry, A. H. 1993. A Field Guide to the Families and Genera of Woody Plants of Aries outh America (Colombia, Ecuador, Peru) W ington, D.C. Lien by Univ. entry, A. H. 1993. vide "d floristic ie pron of lowland Tee in Africa a uth America. Pp. 547 in P. Goldblatt rese Biogeography of. Africa and South America. Yale Uni s, New York. Gentry, A. H. 1993. Vistazo ipa a los Bosques Nub- lados Andinos y la Flora de Carpanta. Pp. 67-79 i Germán I. Andrade (editor), hut Selva Nublada y Páramo. Fundación Natura Colombia, Bogo Gentry, A. H. 1993. Tropical forest iver and the po- tential for new medicinal plants. Pp. 13-24 in A. Douglas Kinghorn & Manuel F. Balandrin (editors), Human Medic- inal Agents from Plants. A.C.S. Symposium Series 534. American mor Society, asking, j Gentry, A. H. Six new species of Adenocalymna posee ua eastern South America. Novon 137-141. e. = Gentry, A. H. 1993. El significado de la Vg ao Fondacién Alejandro Escobar. Bogotá, Gentry, A. H. & G. Aymard C. A new species of Styloceras (Buxaceae) from Peru. Novon 3: 142-144. Gentry, A. H. & R. Ortiz S. 1993. Patrones de Compos- ición Florística en la Amazonia Peruana. Pp. 155-166 in R. Kalliola, M. Puhakka & W. Danjoy (editors), Ama- zonia Peruana— Vegetación Hümeda Tropical en el Lla- no Subandino. je y ONERN, Jyväskylä. Keel, S., A. H. Gentry & L. Spinzi. 1993. Using vege- tation analysis to im the selection of conservation sites in eastern ruri Conservation = 7: 66-75. "e n . H. Gentry, R. B. Foster, L. H. Em- & J. y. ein Jr. 1993. The d Dry For- ests of Santa Cruz, Bolivia: A Global Conservation Pri- ority. RAP Working Papers 4. Conservation International, Washington, D . Phillips, O. & A. H. Gentry. 1993. The useful plants of Tambopata, Peru: I. Statistical hypothesis tests with a new quantitative technique. I Bot. 47(1): 15-32. Phillips, O. & A. H. Gentry. 1993. The useful plants of Tambopata, Peru: II. iia hypothesis testing in quantitative ethnobotany. Econ. Bot. 47(1): 33-43 Awang, D. V. C., B. A. Dawson, J. C. Ethier, A. H. Gentry, M. Girard & D. Kindack. 1994. Naphthoquinone con- stitutents of commercial Lapacho/Pau d'arco/Taheebo products. J. Herbs, Spices, and Med. Plants. 2: 2743. E R. B., T. A. Parker III, A. H. Gentry, L. H. Em- dt uae Hi Schulenberg, L. e G. D. ochea, W. Wust, M. Romo, J. A. Castillo, D "Philips C. Reynel, A. Kratter, P. K. Don ahue & L. J. Barkley. 1994. The Tambopala- Candamo reserved zone of southeastern Perá: A biological as- sessment. RAP Working ome 6. Conservation Inter- 1994. Increasing turnover through time in tropical PRIME Science 263: 954—958. Phillips, O., entry, C. Reynel, C. Wilkin & P. Galvez-Durand B. 1994. Guanticaire ethnobotany and Amazonian conservation. Conservation Biol. 8: 225- 248 Phillips, Vasquez. 1994. O. L., P. Hall, A. H. Gentry, S. A. Sawyer & R. Dynamics and species richness of trop- ical rain forests. Proc. - l. Acad. Sci. 91: 2805-2809. Barringer, K. & A. H. Gen 5. New species of Gib- soniothamnus ie Schlegelieae). Novon 5: 0-124. Clinebell II, R. R., O. L. Phillips, A. H. Gentry, N. Stark & H. Suuring. 1995. Prediction of neotropical tree and liana species richness from soil and climatic data. Bio- diversity and Conservation 4: 56— Gentry, A. H. 1995. lige and floristic composition of neotropical dry forests. In: S. H. Bullock, H. A. Moo- ney & E. Medina (editors), Seasonally Dry Tropical For- ests. Cambridge Univ. Press, New York. Gentry, A. H. 1995. Patterns of diversity composition in neotropical 126 in S. P. Churchill, H. Balsle Luteyn (editors), Biodiversity and Conserv tropical Montane Forests. The New York Botanical Gar- den, New York. Vasquez, R. & A. H. Gentry. 1995. Use and misuse of forest-harvested fruits in the Iquitos area. Pp. 96-107 in D. Ehrenfeld (editor), Readings from Conservation Biology. Blackwell Scientific Publications, Cambridge, and floristic 996. Species extirpations and T e. Pp. 17-26 i Szaro (editor), Biodiversity in "Mariaged yide 13 Theory and Practice. Oxford Univ. Press, New York. Works by Gentry in Press or in Preparation Arrabidaea dara a new species of Bignoniaceae from Bahia. Hook., l Bignoniaceae. In pae ns of Hawa Flora of Capeira and ri Guay Region. Banco Na- cional, Quito. [With Roundtable on Malo Forests. | [With C. Pe- & R. Mendelsohn.] Nicole In: Flora de Nicaragua. 444 Annals of the Missouri Botanical Garden Apocynaceae. "i Flora de s 'aragua. Humiriaceae. "lor: i di Passifloraceae. j^ Bignoniaceae. jd Bignoniaceae. / Bignoniaceae. / Flora de San Juan. : Flora of the Venezuelan Guayana. : Flora de Colombi: ciet VR diversity and inde dan: The consery 1al signifo ance of B-dive ersity and explosive specia- tion. /n: G. Prance (editor), Proc. Manaus Works hop on Amazonian Conservation. Biodiversity of Peruvian Amazonia: Problems and oppor- tunities. /n: K. Mejia (editor), HAP Symposium Volume, Iquitos Species richness and floristic composition of Chocó region * Leyva (editor), Co- Y Williams (Bignonia- cea LK. aes | Diversidad Florística Seip a: =i y estudio de 'onservae ión de la flora. In: R. Dirzo, D. Piñero & M. rroyo (editors), ipn en America Latina. E ‘eae. In: = ioe E = Flora de Paraguay. Figure 2. Gentry in his office at the Missouri Botanical Garden in 1993. Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute 445 GRADUATE STUDENTS SUPERVISED BY GENTRY THESIS TITLE) (WITH Henry Rodriguez, M.S., Washington University, 1977 [de- gree not complete Michael Zimmerman, / ; of Polemonium University, 979. David Lorence, A Systematic and Eco-evolutionary Study of the Monimiaceae in the Malagasy Region. Ph.D., Washington University, 1980. Franklin Ayala, No thesis. M.S.. Washington University, 19 Analysis of the Reproductive Strat- in Colorado. Ph.D., Washington Valerie Kaos; An Ecological Investigation of Scle idis in Two Tropical Forests. Ph.D., Washington University 1982. César Barbosa, Revisión Taxonómic o de la Sección Cau- lanthera (Zygia Browne) del Género Pithecellobium dope (Leguminosae-Mimosoi da 'ae) en Colombia. , Universidad Nacic E ba Colombia, 1984. (Co- E with Enrique Forer Clement Hamilton, An Ecosy toalla Study and Revision of Psychotria pd Psychotria (Rubiac eae) in Mex ico and Central America. Ph.D., Washington ¡es Dóhald Faber-Langendoen. Combining Conservation and Forestry in a Colombian Rain Forest: An E o cal As- sessment. , Saint Louis University, Ricardo B Beso of the Genus C d ( Ver- benaceae) in Mesoamerica. Missou- ri-St. Louis, 1989. Systematics rand Evolution of the Ge- nus Petrea, Ph.D., University of Missouri-St. 993 . University of Louis, Washington Galiano, The Flora of Yanacocha, a Tropical High-Andean Forest in oe Peru. M.S., University of Missouri-St. Louis, Gracielza Dos Santos, E oai Wood Anatomy of Te- comeae Es eae). M.S., University of Missouri-St. Louis, 199% Ric hard Clinebell, ks Biology of Exotic Invasion into the ade Tr: ct Longleaf Pine Savanna. M.S., Uni- rsity of Missou uri- St Louis, 1991. Oliver Phillips. Comparative bestr of Tropical Forests, Ph.D., Washington University, 199: Current students (at time of death): Brad Boyle, Washington University. Studying now under P. Mick Richardson Laura Marsh, Washington University (co-directed with R. Sussman) Carlos Reynel, University of Missouri- St. Louis his Ph.D. under P. Neotropical Zanthoxylum (Rutaceae) w on the Wood Anatomy of the Genus, Ph. 3 of Missouri-St. Louis, January 1996 Jennifer Hedin, Washington University. Studying now un- der P. Mick Richardson Ivón Ramírez. University of Missouri-St. Louis. Received er Ph.D. under P. Mick Richardson, Syrena Phy- Rh. and Chromosome Number Evolution in Cryptan- thus (Bromeliaceae), June 1996. Jason Bradford, pce University. Studying now un- der Paul Berr . Received niversity —James S. Miller, Missouri Botanical Garden, St. Louis, Missouri. Figure 3. Gentry and friend in Madagascar. 446 Annals of the Missouri Botanical Garden THE MEMORIAL SERVICE It was my pleasure to be Al Gentry’s first teacher in the world of botany. Al was born and grew up in Clay Center, Kansas, a small town in the north- central part of the state. He graduated from the Clay Center high school in 1963, and as is common in that part of the world, he came that fall to Kansas State University in Manhattan, Kansas. He arrived with superb recommendations, for he had been re- garded as a brilliant student in high school. Al was not sure what he should study in college, but being a capable boy in a small town carried certain as- sumptions, and one was that he should major in physics. Al took courses in physics and did very well; he also did well in chemistry and molecular biology, where for a while he had an undergrad student job in a laboratory. He was spin in anthropology. but he earned a “B” i anthro- pology course and so concluded that np dE was not his fi Al was ics with the then-popular “honors program," and there is where I made his acquain- tance. He was a part of an honors seminar group that was centered on tropical biology, the role of the tropics, and related matters. In the course of the term, he and I began to talk about plants, he began to hang out in the Herbarium, and he be- came fascinated with what plant taxonomists do. Students who were juniors in the honors program were encouraged to take on an “honors project, and especially a project that related to some on- going research program in the University. Al looked into the Herbarium for a possible honors project. and particularly for one that might involve his background in the physical sciences. At that time, Kansas State had a program to study radiation and radiation shielding that were a part of nuclear pow- er plants, and there was a radiation shielding lab- oratory in the rolling hills just a few miles west of town. The people who directed the research at the site were interested in learning if the radiation from the site affected the nearby biota, so Al formulated a project wherein he collected plants, identified them, mapped their distributions on the site, and compared his data to those from undisturbed or non-radiated tall grass prairie. The results were presented as his honors seminar, but as he was wrapping up his work, he wondered if the radiation caused the observable differences in plant distri- butions, or if the differences were attributable to disturbances from the construction of the site. He could not solve that particular question, but he had done some research on his own that involved a de- kid finable problem, a clear set of procedures, the tech- niques to carry them out, and some support from the Department of Nuclear Engineering. It all fit together nicely, and it led Al into a career in sys- tematic botany. Al graduated from Kansas State University in June 1967, with two degrees, a Bachelor of Arts and a Bachelor of Science, both magna cum laude. He was capable, articulate, ready to talk about any- thing, perhaps a bit ornery, and obviously going places in systematic botany. We at Kansas State assisted in sending him to the graduate program at the University of Wisconsin, to work with Professor Hugh Iltis, and, as may be said, “The rest is his- Dd ory Some years ago, a fellow who had just returned from a stint in the tropics came to the Herbarium at Kansas State, and he brought with him a sack of mangoes. We sat at the lunch table and chopped up the mangoes and ate them. Somebody said, “Hey, taste this one—it REALLY is good,” and sure enough, that one mango was truly excellent. One of the Old-Heads there said, thing about mangoes, for once you find a really “There is a funny good one, you will remember it forever, and you will spend years searching and waiting for another one that is as good." Al Gentry was our Good Man- go. We look forward and wait for another good man- go, or for another Al Gentry.— Theodore M. Barkley, Kansas State University On November 21, 1966, I got a letter from one Al Gentry, a Manhattan, who informed me that his botany pro- senior at Kansas State University in fessor, Theodore Barkley, suggested that he might like to do graduate work with me in Wisconsin, that he had a G.P.A. of 3.9 and ... “do we have a fel- lowship” for him? I wrote him right away and said that would Mr. Barkley write me a letter outlining all his sins and virtues, and for him to tell me what he was interested in: the tropics? Wisconsin flo- ristics? a summer in Costa Rica? whatever? Three days came the reply, in Gentry’s very small hand- writing: “I have enclosed a copy of the personal essay that I wrote for the Woodrow Wilson Fellow- ship, I think it will cover all of your questions, even though Dr. Barkley about flipped the number of ‘I(s)’ in this essay, there are lots of *I(s) in there.” Indeed, Alwyn Gentry did not hide his capabil- ities, then or later. Nevertheless, this was a very interesting essay, six pages long and exceptionally well written. To him, first of all, the beauty of nature and the stimulation of science were his chief intel- lectual concerns: “I need to be out of doors as much as possible, just to be close to the sheer beauty of Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute the natural world . . . to feel the mysterious perfec- tion of the organisms, the how and the why." “Bot- any, intellectually, is pleasing to me because of its great emphasis on evolution and ever since I have been a preschooler I have been a taxonomist at heart His second consideration, he writes, *is my de- sire for physical adventure with tropical plants, tropical field trips, the chance to be away from civ- ilization exploring little- mown jungles and discov- ering exotic new specie But then, also, in a cal switch, “I feel a very strong desire to be something more than just a sci- entist.... I want to be able to justify the impor- tance of my work to myself and others, and as a college teacher I want to stimulate others to ques- tion why they are students and consider their debt to society.” Now, I am a bit of a moralist myself and, as an old environmentalist, I have fought many an envi- ronmental battle. But now here comes this young man from the flatlands of Kansas, this boy, barely 21 years old, writing a professor in the flatlands of Wisconsin, not only about physical adventures in little-known jungles away from civilization (that alone would pique my interest), but also professing there, for one so young, quite astonishing senti- ments. And yet, at the very end he returned to his main theme: “The collecting and field study of tropical plants also presents the challenge of physical ad- venture in tropical jungles. . . citement of unknown places excites me very much, beyond anything else. [And now, get this!] After receiving my doctorate degree, I hope to reevaluate . The mystery and ex- the whole taxonomic system now in use so as to give it a better portrayal of true evolutionary rela- tionships. When I read this, I said to myself, “My God,” I mean, who does he think he is? What have you done, Ted Barkley, to this innocent kid? I showed Gentry’s letter and fellowship application to Paul Allen, our Chairman, who scribbled on the top of it the understatement of the year: “This chap has possibilities!” Well, I finally did meet up with this Wunderkind, when, arranged by Barkley, I gave a speech at Kan- sas State on April 6, 1967 (Iltis, H. H. 1969. A requiem for the prairie. Prairie Naturalist 1(4): 51— 57), with Gentry in the audience. In that rather inflammatory talk, I urged the students in Manhat- tan to become environmental activists, to fight for a series of 1-million-acre Prairie National Parks, to stage “UC Berkeley” protests on agricultural cam- puses to force intellectual change, to revolt against Figure 4. the Pan ama/Colombia border. Gentry conducted the first botánical sir du of the mountain in 1975 and discovered many new species. Gentry on Cerro Tacarcuna along the environmental ignorance of their stodgy teach- ers. The chancellor sat through the whole thing, and looked like had been eating sour apples. Af- terwards, Al was introduced to me, a thin young man, pale, hungry-looking, intense, and, I felt, somewhat uptight, and very, very ambitious. He liked the speech. And so it came about that Al Gentry chose the University of Wisconsin to be- come my graduate student. And when he first vis- ited Madison for a week in June of 1967, we ar- ranged for him to take the OTS in Costa Rica, which, for one who wasn't even a student yet at the UW, was very unusual. But somehow we finagled travel money for him to get to the tropics, finally, and the rest is history. I need to add here something about his letters of recommendation. According to Prof. Lloyd Hulbert, Al was “exceptionally intelligent, among the most gifted students" he has ever had but *. . . in the minor details in his work, because in fact careless this is the result of the feeling that such things are unimportant for him to bother [with] in the big 448 Annals of the Missouri Botanical Garden scheme of things.” And this is of course the para- doxical glory, in a way, of much of Al Gentry’s work. That he saw the big picture and became this fine taxonomist and global ecologist. Ted Barkley, his major advisor, wrote that, “If there is one word to criticize him it would be that he is perhaps over confident or even somewhat conceited. However, this may be only a natural ee of being a bril- liant and competent young n Well, I had other a in that fall of 1967, I got divorced, that's always difficult, there was al- ways something going on; in short, I didn't pay much attention to Al. But he got all A's the first semester, and he got all A's again the next semester. While visiting us that June, Al was looking for a research project. | wrote my friend, Russ Seibert, a graduate of the Missouri Botanical Garden, and then Director of Longwood Gardens in Pennsylva- nia, who once upon a time studied Bignoniaceae, and asked him if he or anybody else was now work- ing with this family. And when he said “no,” I sug- gested to Al that he might have a look at it once he got to Costa Rica. Well, as many of you know, Al fell in love with the Bignoniaceae. He started to work on a revision of Central American Tabebuia in the spring semes- ter of 1968, worked very fast and furious, and soon told me in no uncertain terms that he's going to be done with his master's thesis by the end of the sum- mer. That declaration caused our first big blowup. I told him that, “At UW, it takes two years, so there is no way you can do it in one." "Oh no, you can't." “Oh yes, I can.” I mean, I said, “I am the pro- fessor here and you are the student [genius or not]. and it takes time to write a good thesis." But any- body who ever knew Al Gentry may appreciate that confrontation; he was enormously stubborn, abso- lutely confident that whatever he wanted to do he was going to do it. Well, I have similar attributes, and I was older, which had its advantages. So, in- stead, that summer we shipped him off to Miami, Florida, to take a tropical botany course. In the fall of 1968, he was really going at his Tabebuia, fin- ished his thesis in record time, and by Christmas, finally got his master's degree in plant taxonomy. But Al, from early on, was also an ecologist, and the very first paper he ever published, for OTS and while still at Wisconsin, was a comparison of leaf shapes between dry and wet forests in Costa Rica. But ominous events were about to overtake him. The war in Vietnam was in full swing. people were getting drafted, including Al. We talked a lot about this because he did not believe in the war, did not really feel that he would like to collect plants in Vietnam, at least not under those circumstances! In the meantime, that fall, he brought a lovely young lady to Madison, his childhood sweetheart, ulie. Professor Lenn Thien and his wife and I threw an engagement party for them, a very inter- esting affair as it turned out, because they were both so shy. I bought some bottles of champagne to think Al came from a "dry county" in Kansas that didn't permit alcohol, and my strategy didn't work: they hardly touched the stuff, and both just sat there, totally serious, shyly looking at each other, for three whole long hours, no matter how many jokes I cracke So, that Christmas, right after Al graduated with his master's degree, the couple got married, took a loosen things up a little. quick honeymoon trip to Costa Rica and Panama to collect more plants, and then off into the Army. By April 7, 1969, he was in Fort Leonard Wood, Missouri, a little bit later in Fort Sill, Oklahoma, where his wife by then had joined him. Eventually he ended up at Officer’s School in Fort Belvoir near Washington, and, was offered a commission as a second lieutenant. But this pro- on completion, motion he turned down, for several reasons; in any case, he did not want to go to Vietnam. By then, we were all writing letters to the military that Al had to support his widowed mother, and also that he was needed to teach botany in Madison. And so, not any too soon, we got him out of the Army and back into botany. It was now the summer of 1970, and Al was un- decided whether to return to the UW or not. It all depended on getting a fellowship. To make a long story short, he was bribed (I hate to put it this way) by Walter Lewis of the MBG to come to Missouri, with money we could barely equal. After many let- ters, back and forth, about the needed support, Walter then offered him not only a fellowship but also travel funds, funds “unlimited travel did it! I didn’t believe it, but what else could I do except wish Al apparently is what he promised. That well? In retrospect, it was a good decision. It was time for us to part. Let me wrap up this sad duty with a note that I wrote to the library here, and to Peter Raven: I am giving a set of all my 25 years of correspondence with Al Gentry to the MBG library, sending it to the place where he not only earned his Ph.D. deal- ing with his beloved trumpet vine family, but also where he was able to really spread his wings and learn to fly like a botanical Jonathan Livingston Seagull; that is, do impossibly productive feats in all too short a ate was very kind to this bril- liant, complex Me contradictory man, who not only escaped by hairs breadth being sent to Vietnam and possibly death, but who, were it not for the Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute 449 tolerance, or wisdom, of the Missouri Botanical Garden might well have suffered the fate of many another maverick biological genus and disappeared somewhere into relative obscurity. But the lives of Gentry, perhaps the greatest tropical ecotaxonomist, and Peter Raven, arguably the greatest botanical organizer, intersected in their planetary orbits just at exactly the right time, in 1972 or so, shortly after Raven became Director of the Garden and Gentry needed a permanent position. a platform from whence to take off. And now, all too soon, we mourn Al Gentry's untimely death and our great loss. What could he not have accomplished in the next 20 years! There will not be another one like him in our lifetime, and we are all grateful for what he has given to the world. And we must also say thanks, finally, to Pe- ter Raven and to the Missouri Botanical Garden for putting up with and supporting. this driven and dif- ficult small-town school on yearned for excitement and dreamed of ansas boy, who from grade adventure, and found more, finally, than he could have ever hoped for.—Hugh H. Iltis, University of Wisconsin Al Gentry’s interest in the Bignoniaceae was kin- dled while at the University of Wisconsin. It was during this time, in the fall of 1968, when he was completing parts of his thesis at the Smithsonian Institution, that I first met this enthusiastic and en- ergetic student of plant biology. Thanks to Hugh Iltis’s many stories of St. Louis, some great and some not so great but undoubtedly absolutely true, we were fortunate enough to bring Al here in 1969 at typical warp drive even then. I feel certain that Al would have revised the whole 112 genera of Big- noniaceae if some brakes had not been put on his Ph.D. research scope. Even so, using detailed eco- logical studies of mostly new data on edaphic limits of distribution, floral phenology, floral morphology, pollinators, and fruiting strategies, his systematic treatment included a detailed analysis of no fewer than 85 Central American species. A study of the family and a number of little-used taxonomic char- acters, apparent only from field observations, were also incorporated into his dissertation in arriving at systematic and phylogenetic conclusions The dissertation was 769 pages long, plus 184 full-paged figures: without hard covers it weighed 9-3/4 pounds and was 4-1/2 inches thick. Lest you think we only weigh and measure dissertations when awarding degrees at Washington University, be assured that at least in Al’s case his research did indeed involve extraordinary scholarship, both in quality as well as mass. I perhaps dwell too long on Al the Ph.D. student, but I think that what he accomplished then epitomized what he was to do in the future: his great interest in plants as inter- active parts of complex ecosystems, his acute ob- servational powers rarely if ever lost from his mem- ory bank, and last, but not least, his ability to impart this knowledge with clarity of thought. Suc aspects were clear to me and other members of his committee when Al graduated in December of 1972 to take up his position as a curator at the Missouri Botanical Garden. I must add one anecdote about Al’s dissertation known to very few, except Julie and perhaps their children, and one which graduate students in par- ticular might appreciate, for pitfalls befell even Al Gentry when he was a student. Quoting from his letter to me of June 30, 1972, he said, “I feel there must be a cloud of bad luck hovering over me: the [original] copy of the thesis I had with me was sto- len along with a suitcase full of clothes in Copen- hagen [Al was visiting European herbaria examin- ing types of Bignoniaceae]. Now with the fire at Summit, I may really be wiped out [referring to the Garden’s herbarium and living quarters in the old Panama Canal Zone where Al was acting curator while doing his fieldwork, which burned while Al was in Europe]. I'll let you know as soon as we get back to Summit whether you’ve still got a terminal grad student or one who is more or less starting over again.... Believe me we've got our fingers ell, it turned out that Al did not have to redo his research, for the only carbon copy sur- vived the fire in a file cabinet with only limited water and smoke damage, but it had to be retyped crossed.” by Julie; remember, this was the age before com- puters. Over the years Al's interest in teaching and di- recting graduate students’ research expanded with his career. He was deeply committed to our aca- demic program in biology at Washington University as well as later at the University of Missouri-St. Louis. Whenever possible he participated in our faculty meetings as an Adjunct Professor in Biol- ogy, but his greater contribution came with his ded- ication to graduate students seriously interested in plant systematics and ecology. Al taught them well individually and through his two courses, “Phyto- geography” and “Structure & Composition of Trop- ical Forests,” and a seminar in Floristic Taxonomy, one 1 developed many years ago, which Al contin- ued to give from the 1970s on. When not in the field, he was available for in-depth discussions and reviews of classwork and research, and he gave his time freely for determining specimens to genus and 450 Annals of the Missouri Botanical Garden Lo, r Figure 5. Gentry with José Schunke in species, even very incomplete ones, with amazing skill and speed. This does not suggest that Al en- joyed being pestered, especially by those who did not learn after the first or second attempt, for Al did not “suffer fools gladly.” An aspect of our association, which will not be found in our professional curriculum vitaes, is that we belonged to the same undergraduate fraternity. Alpha Tau Omega. I mention this for two reasons: first, because Al received at least one coveted award from the fraternity as an academically top senior at Kansas State, a fact which comes to no surprise to anyone here, and second, because i provided a very special link in our relationship which meant a great deal to both of us. For his family, for his associates at Washington University, the Missouri Botanical Garden, and elsewhere, and for me, Al’s death is very much a - compounded loss, because we have lost not only a friend, we have also lost a truly exceptional col- league, who devoted his life to the making of a brighter and greener planet for humanity.—Walter H. Lewis, Washington University ocache Nuevo, eru, in 1979, Al Gentry’s death shocked the scientific com- munity. This is shown in a message I received from Dr. Ariane Luna Peixoto of the Universidade Fed- eral Rural do Rio de Janeiro, Brazil, one of Al's closest friends in Latin America. She says: “For Brazilian botanists, particularly those working on the Atlantic coastal forests, the passing of Al Gen- try constitutes a source of undescribable sorrow. Gentry was our friend and our teacher. A devoted teacher who would teach us from far away through his very detailed letters as well as through his pub- lications, and from close by in our classrooms and tropical forests. Speaking in a mixture of English, Spanish, and Portuguese, in field trips and in the classroom, he never neglected anything. He was al- ways willing to share his knowledge. For us Latin Americans, he took the mystery out of many things. He showed us that plant identification in the field can be simple and easy as long as you have the commitment, as long as you develop the capacity to observe and, above all, as long as you love what you are doing. He showed us that phytosociological sampling in tropical forests can be carried out in a Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute 451 quick, efficient, and correct way, as long as you are willing to go beyond your own physical abilities. He showed us that inventories and phytosociology are also taxonomic tasks that can be used to great advantage in monographic research. Al Gentry was a role model. With his departure, we have a difficult but important task before us: to try our best to reach his level in our own professional lives.” I knew Al during his entire professional career. We both received our Ph.D.s in 1972. It was in late 1972, during the Annual Systematics Symposium, that I met him here in St. Louis. At that time, and at Peter Raven’s suggestion, Al and I agreed to de- velop a joint research project in the Chocó region of Colombia. It would be a cooperative effort be- tween the Missouri Botanical Garden and the In- stitute of Natural Sciences of the National Univer- sity in Bogotá, where 1 worked. It didn't take long for Al to show up in Colombia. Early in January of 1973 I received a phone call from St. Louis. It was Al, telling me that he would be in Bogotá in a couple of days (actually, on Jan- uary 5th. I was on vacation!), and that he would like to be in the field in Chocó, with me, by the 6th! Somehow I managed to get plane tickets for both of us for the 6th. He arrived on the 5th, spent the night in my apartment, and we flew to Quibdó the following day. We spent the next five days col- lecting in central Chocó. This was the beginning of what we considered at the time one of the most successful joint programs in botanical research be- tween a North American institution and an insti- tution in a developing country. Al and I were to- gether in the field in Colombia perhaps as many as five times. But the program lasted for twelve years. During this time we provided training for several Colombian and U.S. students, helped establish a herbarium in the city of Quibdó, initiated some small-scale field research by local botanists, col- lected over 17,000 numbers (more than 80,000 her- barium specimens) between the two of us, and brought the attention of the scientific community to this poorly known but extremely rich region in terms of biodiversity. Al’s eagerness to be in the field is well known. He was not particularly interested in public rela- tions, meeting people, or spending time in big cit- ies. He wanted to be in the field. In fact, for at least the first two years of his involvement in Colombia he seldom spent more than a day in Bogotá. As a result, no one at the Colombian National Herbari- um knew him. I had a hard time explaining to my colleagues that Al Gentry really existed! Finally, I decided to “kidnap” him for a few days so that he could meet the local botanists. He was extremely strong in the field. I remember a time when Al, Andrew Sugden (a British student), Douglas Daly, and myself were in Chocé, exploring the foothills of the little known Cerro Torrá. Basi- cally without food we traveled for hours, until most of us were exhausted. Andrew, Doug, and I actually decided to give up our collecting. We sat down to rest and to try to recover our strength. Al kept on going, and several hours later came back having found a place for us to spend the night. We then were up until 2 a.m. pressing plants. Pressing plants at night was something he seemed to enjoy. Even with our poor performance in the field that day, we had managed to collect nearly two hundred numbers! To be sure, Al and I had many differences of opinion. A couple of examples of a fairly mild na- ture will serve to illustrate this point. One of the many positive results of our work in Chocó was the publication of the “Chocó Checklist” (Forero & Gentry, 1989). However, bringing this checklist to completion was not an easy task be- cause, as we all know, Al was stubborn when it came to taxonomic matters. He would not accept, for example, the subdivision of the genus Cassia into three genera as proposed by Irwin and Barneby (1982). As someone interested in the family Legu- minosae, I was happy with their classification. Al wasn’t. In the end I was forced to give up and in- clude all the pertinent species under Cassia in or- der to see the checklist published. Not two days had passed since the checklist was published when I received a letter from Rupert Barneby asking me if we were out of touch with taxonomic literature! Al and I used to argue because he blamed me for rating my students too high. Who did not hear him say that his students were all wonderful, the best!? The truth is that we both had excellent stu- dents. Our lives followed somewhat different but par- allel paths. My life has been more devoted to the administration of science. His was devoted to his research. I am glad that I was able to contribute to his scientific accomplishments as we worked to- gether in Colombia. I am also grateful for the op- portunities I had, as an administrator both in Co- lombia and at the Missouri Botanical Garden, to help him in some way to continue to fulfill his sci- entific dream. But more than that, I am glad that I can remember him as a wonderful colleague and as a friend. Literature Cited Forero, E. & A. Gentry. 1989. Lista anotada de las plan- tas del Departamento del Chocó, Colombia. Bibl. José Jerónimo Triana 10: 1-142. Annals of the Missouri Botanical Garden Figure 6. Museum of Natural History, unloading equipment at Yaviza, Panama. on one of Gentry’s expeditions into the little known Darién. Irwin, H. S. & R. Barneby. 1982. The American Cassi- inae. A synoptical revision of Leguminosae Tribe Cas- sieae subtribe Cassiinae in the New World. Mem. New York Bot. Gard. 35, parts 1-2: 1-918. —Enrique Forero, Universidad Nacional, Bogotá, Colombia When I gave my first presentation at the Missouri Botanical Garden in 1983, I showed some slides of medicine men and said that these people can iden- tify every species in the forest without looking at the fruit or the flowers and that no university- trained botanist could ever hope to do that. Almost before I was even finished this skinny guy had come racing down to the podium, shook my hand, said Pm Al Gentry and I can do that. So I handed him a pile of these miserable sterile specimens that the Indians had given me: a piece of bark here. a piece of leaf there, and within minutes Al had iden- tified everything to species. I gave a talk at the American Horticultural So- ciety two days after the accident. I started out by asking how many people knew that a basketball player died a month ago of a heart attack. and al- most every hand in the audience went up. I then Gentry with Rudolfo Hindo and Charles Myers, head of the department of herpetology at the American asked how many people knew that the greatest trop- ical biologist who ever lived died in a plane acci- dent two days ago. and about 20 hands went up. I truly believe that says something terrible about our culture and who our heroes really are. We worship sports figures and politicians, and we have guys like Al Gentry and Ted Parker who don't care about money, don't care about clothes—they just want to make the world a better place. When people die in the military it is said that they gave their lives for their country, but what did Al and Ted give their life for? They gave their lives for everybody and for the planet. They were always pushing the edge of the envelope and they paid for it with their lives. Al and Ted were heroes in an age without heroes. The most incredible display of biological exper- tise I have ever seen was in a meeting at the World Wildlife Fund about eight years ago. WWF had in- vited 30 top scientists to discuss Andean priorities. Despite all the charismatic megavertebrates around the table, Al and Ted started out in northern Ven- ezuela and walked down the Andes peak by peak, valley by valley. This is top priority, the narcos are here forget it. Nobody’s ever been here, we have got to go there. They made the rest of these guys superfluous. In fact, it was humiliating to be a sci- Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute 453 entist in the room with Ted and Al that day because if they had taken those other 28 guys and pushed them outside, it wouldn’t have made any difference. When they died, Russ Mittermeier said 75% of the unpublished information that exists on the biodi- versity of Andean forests was lost with Ted and Al. Jim Duke just finished a book on Amazonian eth- nobotany with a dedication to Al that Pd like to read: “Our book has been enhanced by the taxo- nomic foundation provided by the arduous and intensive perseverance of the late Dr. Gentry. He would not rest until he named an unknown forest We respectfully but sadly dedicate this friend of the forest and teacher to many of us, trying to save the forest that survives him. The Amazonian Center for Environmental Education and Research (ACEER) will dedicate their 250,000-acre forest . The forest that kept him going like a robot swallowed him up. But his spirit lives on, and will help in the difficult efforts to save the forest of today for the children of tomorrow. Few reserve to his memo of us can view any attractive bignoniaceous vine without recalling the ethereal spirit of Al Gentry and the forests he represented. Long live the forest and spirit of Al Gentry” (Duke & Vasquez, 1994). en I spoke at Ted Parker's funeral service last weekend, I read a line from ‘Tennyson: “Those whom the gods love die you I don’t know if the gods put Al and Ted, but I sure did. Literature Cited Duke, J. A. & R. Vasquez. 1994. Amazonian Ethnobo- tanical Dictionary. CRC Press, Boca Raton, Florida. —Mark Plotkin, Conservation International I graduated from Washington University and Al- n Gentry was my doctoral advisor for five years. Scientists like Al are very rare—he has tackled so much, taking everything to the limit and then be- yond, exploring new fields and new horizons, and through his field guide unlocking the mysteries of the neotropical forests for us all. A scientist that special who is a great teacher as well, who has devoted so much time and energy and enthusiasm to their students around the world, is truly unique. People like Al and Ted are rarer than gold, and we know what a terrible loss this is. It would be hard to overstate Al's influence on botany students in the Americas. In 21 years as a curator, he had at least 20 doctorate and master's students at all three of the universities affiliated with the Garden—-Washington University, Univer- sity of Missouri-St. Louis, and St. Louis University. Although as a researcher Al had been going full steam ever since his undergraduate days, as a teacher his responsibilities were snowballing. When he died he was supervising theses by no less than nine students, and had recently taught grad- uate courses in St. Louis and Latin America (Peru and Colombia). Several qualities come to mind when I think of Al Gentry the teacher, qualities that made him truly special. There was his tremendous enthusiasm, ex- citement, and knowledge about plants and forests that drew like-minded people to him. Al went all over the world, botanizing like a man possessed, and took students in each country into the field wit im. Here was someone who could convince even the most confirmed urbanite of the essential wonder of the natural world. During these field trips it was impossible not to be inspired by his excitement, and so from all over the world students came to St. Louis to learn from Al Gentry and his colleagues here! Other students came to meet the man whose botanical zeal they had only heard about through some of the Gentry legends. Ted Parker used to tell the story of how Al once fixed his sights on col- lecting an unusual looking epiphytic bromeliad, 20 meters up a tree. Climbing up, he put his hand in it from below, and felt a sharp pain. Cursing the spiny plant that he thought had pricked him, he put his hand in again only to get it cut again. After this happened a few times, eventually he realized that rather than being cut by the plants’ leaves, he was being repeatedly bitten by a venomous fer-de- lance! He survived that encounter of course, to the great benefit of his students. Al cared deeply about his students, not just ac- ademically, but also about our welfare. For students coming from abroad, the culture shock arriving in the United States can be pretty severe. But he was always there if you wanted him for personal advice. Since his office was always open, people would troop in, in a fairly uninterrupted stream, and he managed to find time for all of them. One of the strongest memories I have of Al is of him wryly smiling, as he sits at his desk which is piled two, three, maybe even four feet deep with specimens, and manuscripts, and papers, 'normous amounts of work, and a small queue of students and curators in his office wanting to bend his ear. There he was with dozens of projects on the go at the same time, and he was still willing to talk with you about what you wanted to talk about. The description of Al as a fallen giant forest tree is so right, perhaps it is something we have all Annals of the Missouri Botanical Garden Figure 7. thought about. He was always searching for new light, for new growth, and new inspiration. He was a forest giant and most of us who studied under him and who worked with him, really were privi- leged to be around this forest giant. When he died it crashed down and left a huge gap in the forest, an enormous gap. When we first learnt of his pass- ing, we thought that we could never fill this huge hole, it is too big. But the more we think about Al and what he has done to teach others, of the many students not only here but all over the world, thousands and thousands of people he has inspired to dedicate their work to the tropical forests, we hope that perhaps we can fill that gap eventually. But we had better do it quickly for the sake of giants like Al who cared so passionately.—Oliver Phillips, University of Leeds I was a graduate student with Al Gentry for six years, and I graduated two days before he left for his last trip. He taught field and classroom courses at many universities throughout all of Latin Amer- ica. On the field courses and collecting trips, he was a tireless plant collector and a minute-to-min- ute teacher. He was not always a serious person. Once in a while he would tell you a quick joke and then return to talking about plants. The teaching Al Gentry teaching a group of students at Von Humboldt in Peru. would include anything related to plants, such as nomenclatural problems, phytogeography (one of his favorite topics), ecology, and so on and so forth. He realized that one way to do conservation of bio- tropical diversity was teaching and learning from others. Whenever you talk to people who have met Al, they will tell you how good he was as a botanist and as a teacher, and this was because Al believed in just teaching and passing on his experience gained from years of travel and research in Latin America. In the field, he was always accompanied by stu- dents and botanists from local institutions and he would always talk to them about ideas and projects. He encouraged many people to do research in the tropics, and in some cases he suggested topics to the students. He also encouraged many of us in the tropics to study abroad. He helped students get en- rolled at foreign universities, and in many cases he was also our thesis advisor. As a major professor, he was one of the best, helping students despite the fact that he was always very busy. For example, to come to his office to talk about the thesis or about any scientific problem no student needed an ap- pointment. When new the dates he would be out of town, he would come to his students and tell them, so that they could come to his office to clear up urgent situations. In some cases, when he went Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute 455 away for two or three months and the students were working on their theses, he would say, “Write all your questions down and we will talk about them when I come back.” If the students had a document ready to be revised by the advisor, Al would take it and put it in his briefcase and read it at the airports while waiting for a flight, and then would send it back by mail a couple of days later. That was Al! If the students for some reason did not come to see him after he returned from a trip, he would come to the students’ cubicles to see how the thesis work was going. He was also very warm—he would ask you about your thesis and about your personal problems that he could help to solve. Another way Al would help his students was let- ting them use his grant money to travel for their own research activities and for publications. Many of us as students made our first formal publication because Al encouraged us to do it. He would say, “Its so easy, and after the first one you will be a professional. For example, to publish a paper on a new species is a piece of cake. Just take the Bo- tanical Latin book and write the Latin description and then write a complete description in English, bring it to me and I will give it a quick look, and that is it.” It took me a while to believe this, but one day he came to me and said, “Hey Ricardo, I got an Aegiphila from Colombia, I believe it is a new species. Please check it. You know, if it is a new species this is your opportunity for your first paper. There is only one problem, if it is new, I need a name soon since it is a collection from my plot project in Colombia.” Taking into account his great experience, I checked everything about the collection, but I knew from the beginning that it would be new. I wrote the paper and he helped me through the process. When it came out in Novon he came to me and said, “Did you see the paper?” I said, “Yes, I am so happy." And he said, “Now you can go on without problems.” Another Al tradition was to celebrate the theses defenses of his students. Everybody in the Garden knew about the famous pizza party after the stu- dent’s defense. I remember my party, two days be- fore he left on his last trip, and I remember Al talking very highly about his students. He was very proud of them, and we were also proud of having as our advisor the best modern tropical botanist, the man that had a continental view of the flora. Unfortunately, we have lost Al and we are all sad, but we also know that if he were in this room, he would say it is not the time to be sad, it is time to work. We say to you, Al, that you are alive in every one of us. You trained many people in Latin Amer- ica and we will keep your scientific legacy alive with us forever.—Ricardo Rueda, Universidad Na- cional Autónoma de Nicaragua-Leon 456 Annals of the Missouri Botanical Garden EULOGY In thinking about Al Gentry at 22 years old, go- ing to the tropics for the first time after longing to do that for several years, and with so much love for biology, 22 years old on his way between the plains of eastern Kansas and the forest, perhaps, of Mad- ison, | was thinking about how he must have felt, and I was reminded of Darwin. I looked up what Darwin said on February 29, 1832, when he was on the voyage of the Beagle and he got to Bahia in Brazil, and Darwin said, “The day has passed delightfully. Delight it- self, however, is a weak term to express the feel- ings of a naturalist who, for the first time, has been wandering by himself in a Brazilian forest. Among the multitude of striking objects, the gen- eral luxuriance of the vegetation bears away the victory. The elegance of the grasses, the novelty of the parasitical plants, the beauty of the flow- ers, the glossy green of the foliage. all tend to this end. A most paradoxical mixture of sound and silence pervades the shady parts of the wood. The noise from the insects is so loud, that it may be heard even in a vessel anchored sev- eral hundred yards from the shore: vet within the recesses of the forest a universal silence appears to reign. To a person fond of natural history, such a day as this, brings with it a deeper pleasure than he can ever hope again to experience. After wandering about for some hours, | returned to the landing-place; but, before reaching it, I was overtaken by a tropical storm. I tried to find shel- ter under a tree which was so thick, that it would never have been penetrated by common English rain; but here, in a couple of minutes, a little torrent flowed down the trunk. It is to this vio- lence of the rain that we must attribute the ver- dure at the bottom of the thickest woods: if the showers were like those of a colder clime, the greater part would be absorbed or evaporated be- fore it reached the ground. I will not at present [and this is a sentence I think is particularly appropriate] I will not at present attempt to de- scribe the gaudy scenery of this noble bay, be- cause, in our homeward voyage, we called here a second time, and I shall then have occasion to remark on it” (Darwin, 1839: 11, 12) Or perhaps on a little less philosophical but very beautiful mode I was reminded of Douglas Botting's words in his book on Alexander von Humboldt, who with his friend Bonpland was the first ever to un- derstand the richness of tropical vegetation. I think this says it very well too “At nine in the morning on 16 July 1799 the William Ramirez, an entomologist from the on a field trip with Figure 8. University of Costa Rica, and Gentry leaves of Gunnera that are locally used as disposable um- brellas. Pizarro anchored in the Port of Cuminá and even the surviving typhoid victims managed to drag themselves up on deck to have a look. After three. weeks on the high seas, the land seemed The high mountains of New An- line the horizon; breathtaking. dalusia, half veiled by mists, the town in its fort glimmered between tall co- conut palms along the banks of a river; rose- coloured flamingoes, snowy egrets and huge brown pelicans the size of swans foraged along the shore at the edge of a motionless green sea; even at that early hour, the light was dazzling. Humboldt and Bonpland hurried ashore and looked around them. The place bowled them over, instantly and totally. In the first flush of his Alexander wrote to his brother: extravagant country enthusiasm, What we're in! Fantastic plants, electric eels, armadil- many, real, a fabulous and los, monkeys, parrots: and many, half-savage Indians. "What trees! Coconut palms 50 to 60 feet high: Poinciana pulcherrima with a big bouquet of wonderful crimson flowers; pisang and a whole host of trees with enormous leaves and sweet smelling flowers as big as your hand, all utterly new to us. As for the colours of the birds and fishes—even the crabs are sky-blue and yellow! Volume 83, Number 4 Alwyn Howard Gentry: A Tribute Up till now, we’ve been running around like a couple of mad things: for the first three days we couldn't settle to anything; we'd find one thing, only to abandon it for the next. Bonpland keeps telling me he'll go out of his mind if the wonders don't cease soon.’ *[n this exotic new world the explorers seem to have experienced a kind of sensory ecstasy. Nothing—no shape, no form, no voice, no colour, no smell—was familiar to them. Nothing would readily fit into their existing pattern of memory and experience, therefore, everything seemed to demand equal attention. The Indians—naked! Their huts—bamboo and palm leaf! Their chairs—branches of coral washed up on the shore! Their plates—half a coconut shell! This way and that they charged, confronted at every turn with brilliant and startling new visions, like men in a mescalin trance—here a quama tree loaded with silvery blossoms, there a castle moat full of crocodiles. It required some effort of will to turn their attention to more mundane but nec- essary matters. For first they had to present their passports to the Governor, and then they had to find somewhere to live" (Botting, 1973: Al Gentry lived from January 6, 1945 to em 3, 1993. He had a wonderful life. It came to an abrupt and sad end, and when his life ended Al Gentry joined the ranks of other botanists who died through the years in the pursuit of the same kind of knowledge, the same kind of ecstasy that he un- doubtedly felt on that day in 1967 when he first plunged into the forest of Costa Rica. Banister in Virginia, Forskáhl in Arabia, Lófling in Venezuela, David Douglas in Hawaii, Jeffery in the Colorado desert, Leitner in Florida, and Orton in Bolivia at Lake Titicaca are among the botanists whom Al now joins who gave their lives in the same quest in which he was engaged. Al brought his enormous energy to learning about plants educating people and about sharing his enthusiasm for them. Tropical plants and tropical nature are the most exuberant expression of biology on earth and it’s the part of that expression that it is hardest for us all to un- derstand, most wondrous to us all, and most diffi- cult to deal with. As we have heard, hardship in the field was perfectly acceptable to him and the amenities were always forgotten in his relentless drive for new information and new experience. He really loved his students. We set a rule at the Gar- den that no one had to teach students, anybody could do exactly what they wanted. For Al there was no choice. He loved dealing with students at Washington University, Saint Louis University, the University of Missouri-St. Louis, and as we have heard all over Latin America. He loved to think about plants and to learn about plants and com- munities, the relationships between them, the in- ventories of the biodiversity, the basic information that would make conservation possible. Our loss, which we feel very deeply, is shared by biologists and people who knew him and people who had even a dim understanding of his work from all over the world, and we have received countless messa- ges of condolences, especially from his Latin Amer- ican colleagues. Among the things that we will be doing in his memory will be the establishment of a fund to support education for Latin American bot- anists, the Flora of Peru Checklist, which will be published within the next two months will be ded- icated to his memory, the first complete inventory of the plants of a South American country. The March issue of Conservation Biology, which will have a major article by Oliver Phillips and Al in it, will be dedicated to his memory. The Amazonian ethnobotanical dictionary by Jim Duke and Rodolfo Vasquez which is in press will, as we’ve heard from Mark Plotkin reading Jim Duke’s words, be dedi- cated to his memory. Along with Conservation In- ternational, we will sponsor a symposium or collo- quium on Rapid Assessment and publish the results as a contribution to continuing that effort. And here at the Garden, we will find the necessary means to finish his taxonomic work in progress on the Bignoniaceae and other fronts. And to sum- marize and organize and publish his work on some 250 forest plots, and you have heard about the pos- sibility of dedicating reserves in the memory of Al and Ted Parker. What does it all mean? Very, very hard to say. I have been wondering about that question for quite some time. I believe deeply and profoundly that one can be happiest if one looks on life and all the subparts of life not as a destination, not as arriving at the destination, but as the journey to that des- tination. What happens along the way is the whole point. I think about Al growing up in a small Kan- sas town in the late 1940s, being fascinated by a leaf in the sun or by a butterfly, by a sparrow, or by a little dog, the same sorts of things that anybody growing up in the country would be fascinated by and would feel and note. It is experiences like that when you are a little kid that add up to create your own style. To give you a feeling for what is impor- tant to you and what you want to do. What you want to make out of it all. What you want your journey to be like. Each of us has a completely different set of experiences and each of us puts together those experiences in his or her own way and they add up to the kind of person we are, the things that 458 Annals of the Missouri Botanical Garden Figure 9. receive the Crafoord Prize Museum, at Uppsala ie we enjoy and the things that allow each of us to make a unique contribution. Listening to those who spoke this morning and thinking about the feelings that we all share, I think we can all be very proud of our botanical family, our biological family, our Missouri Botanical Garden family, the network of people who care about life on earth and want to do something about it, but let's say in a humbler way the network of those of us who care about one an- other. It should have been crystal clear in thinking about Al’s life and in listening to these remarks that the best things in life don't amount to putting on an act for one another or living up to some sort of general expectations about what you should do in order to be able to have perhaps a wonderful eulogy at the end of it all. The best thing in life is living day to day, it’s doing it. If we can learn to the extent that each of us makes a unique contribution, to the extent that we can learn to cherish and appreciate the contribution that each one of us make, we wil be able to fulfill our own lives best. I don't want to mince words, and the speakers have made it clear that in some respects Al was a difficult person. We 0, Gentry installed one ed j $ $ Ne f Gentry pos and identifying plants on Olaf Hedberg’s dining room table. While in Stockholm to 19 hi = is 0.l-ha transects, with Hedberg from the Botanical are all difficult people. We just have different kinds of diffculty, but the genius is the love that we feel for one another, the cherishing that we can do for one another, the fraying that we can give one an- other to accomplish great things together and ful- filling things together come from understanding, ap- preciating, and cherishing even those difficulties. It seems that at a time like this the most appropriate single emotion that we can share is the desire to cherish and love and respect and tolerate one an- other more. The rewards of that are great and the penalties for not doing that are correspondingly great. Let's spending six months in Brazil. It is true that he did conclude with what Darwin said after stop again after circumnavigating the globe (re- member he was about the same age as Al was when he first got to the tropics), he did stop by briefly on his way home, but it wasn't the big celebration that he expected it to be. So let's hear what he said in conclusion, what Darwin said after six months in Brazil. first time around. The big time, the big ex- perience, the thing that you are doing now is always Volume 83, Number 4 1996 Alwyn Howard Gentry: A Tribute 459 the big experience—that's the thing to remember, don't look for some kind of future wonderful thing that is going to make everything all right, just do it. On the 23rd of June in 1832 Darwin said, * Again I went to the forest, which so often has been proved so fruitful in all kinds of animals.— It is in all probability the last time I shall ever wander in a Brazilian forest.—lI find the pleasure derived from such scenes increases, instead as might have been expected, diminishing. Today instead of the rude tracks, I followed a brook, which in a narrow ravine flowed amongst the huge granitic blocks.—No art could depict so stupendous a scene.—The decaying trunks of enormous trees scattered about, formed in many places natural bridges; beneath & around them the damp shade favoured the growth of the Fern and Palm trees.—& looking upwards the trees in themselves lofty, thus seen, appeared an almost incredible height.—I soon found even by creep- ing, I could not penetrate the entangled mass of the living € dead vegetation.—On coming out of the forest, the effect without any exaggeration is that of the full light of the sun breaking on a person who has just left a darkened room" (Keynes, 1988: 76). Our deepest sympathies to Al's family and those he leaves behind. Our shared memories of him will be with us always. I thank you all for coming. I appreciate the wonderful and heartfelt comments by all the speakers on behalf of our sorely missed colleague. Thank you. Literature Cited Botting, D. 1973. Humboldt and the Cosmos. Harper & Row, New York. Darwin, C. 1839. Journal of Researches into the Geology and Natural History. Henry Colburn, London. [Quoted from Facsimile reprint of the first edition. Hafner, New York. 1952.] Keynes, R. D. . Charles Darwin's Beagle Diary. Cambridge Univ. Press, Cambridge. — Peter H. Raven, Director, Missouri Botanical Gar- en 460 Annals of the Missouri Botanical Garden Exarata chocoensis A. Gentry, Syst. Bot. 17: 503. 1992. “This remarkable plant seems to be morphologically intermediate between tribe Schlegelieae . . . and tribe Teco- meae.” Gentry had been aware o iun s problematic new genus since at least 1979 but had refrained from describing it for some years "due in part to uncertainty about its familial placement and in part to confusion about whether it is a tree or a as a Fieldwork condue ted in vide ith a RAP team in eastern Esmeraldas, Ecuador, proved that the nt was a tree and supported its generic segregation from Schlegelia. The aos is by John Myers, who has pde cn of plants for die Mies pie al Garden research effort, and is reproduced with the permission of Systematic Botany. TAXONOMIC REVISION OF CRUCKSHANKSIA AND OREOPOLUS (RUBIACEAE: HEDYOTIDEAE)! Charlotte M. Taylor? ABSTRACT 0 reopolus is here distinguished based on its ternate leaves, triangular entire to shortly mucronate or bilobed stipules, bilobed pee 2-7-lobed in fruit; mation are presentec lobes that do not enlarge in fruit; thus circumscribed it only O. glacialis (Poepp.) Ricardi. Crückshatlata 3 is distinguished by ie, ce to ibshemita leaves, erose :d inflorescence bracts that resemble seven species are recognized here. Keys and complete descriptions, nomenc Fait and distributional infor- ed. the leaves, and calyx lobes that are markedly enlarged The taxonomy and study of Oreopolus Schltdl. and Cruckshanksia Hook. & Arn. (Hedyotideae; Puff, 1988; Robbrecht, 1988) have been closely in- tertwined since these genera were first described. The type species of Oreopolus was first described in Cruckshanksia, and two other species, C. ma- crantha Phil. and C. palmae Clos, have been alter- natively treated in Oreopolus by some authors (e.g., Ricardi, 1963a). Both of these genera are endemic to temperate South America and share generally similar habits, corollas, and fruit, including several unusual features, and they appear to be closely re- lated. These plants are characteristic of cool dry sites from near sea level to as high as 3500 m in the northern end of their range. Cruckshanksia is found at low to high elevations in arid central to northern Chile, with two species, C. macrantha and C. hymenodon Hook. & Arn., cent Argentina. Oreopolus glacialis (Poepp.) Ricar- di is found in the Andean Cordillera from central extending into adja- Chile and Argentina south into Tierra del Fuego. These two genera share papery to membrana- ceous capsules with loculicidal dehiscence and two ovules per locule, with the ovules borne on an elon- gated pseudoseptum (Fig. 1A, B). This pseudosep- tum is attached peltately to the true septum and extends at right angles from it into the locule. The genera also resemble each other in ally geophytic or cryptophytic habit and distylous flowers with bright yellow salverform corollas with their low, usu- slender tubes. However, Oreopolus can be separat- ed by its usually ternate leaf arrangement, entire triangular stipules that are usually imbricate by the poor development of internodes, floral bracts that resemble the stipules, and calyx lobes that do not enlarge markedly in fruit. In contrast, opposite to occasionally subalternate leaf arrangement, erose usually deeply bilobed stipules separated by well- developed internodes, inflorescence bracts that re- semble the cauline leaves, and calyx lobes typically enlarging markedly in fruit distinguish Cruckshank- sia. Additionally, Cruckshanksia is distinguished by its inflorescence bracts and in some species cauline leaves that are deeply 2—6-lobed, an unusual fea- ture. Several species are also notable for the pro- longation of one or more of the calyx lobes into a stipitate petaloid appendage. of species treated here were first de- scribed in Cruckshanksia, and were retained in that genus until recently (e.g., Schumann, 1891; Mufioz, 1966). This group was studied in detail by Ricardi (1963a, b, 1968, 1973; Ricardi & Quezada, 1963), who first separated Oreopolus with only one species (Ricardi, 1963a). Later Ricardi (1973) restricted Cruckshanksia to seven species with well developed petaloid calyx lobe appendages and transferred two species that lack these structures to Oreopolus. GENERIC LiMrrs Regardless of how they are delimited from each other, Cruckshanksia and Oreopolus share a dis- I thank several colleagues for valuable comments and helpful discussions, notably M. Muñoz, J. Pipoly, P. Pefiailillo, ica N. Bacigalupo and C. Marticorena. This work was supp ool were invaluable field companions. Special ported in by a grant from the National Science Foundation to the Missouri Botanical Garden for the Flora of Chile project (DEB-9201097). ? Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. ANN. MISSOURI Bor. GARD. 83: 461-479. 1996. 462 Annals of the Missouri Botanical Garden Figure 1. Cross section of ca ele styled flower, partially diee: ted. A, B from Ricardi (1963a: fig. 3C, F); , D to same scale. seeds.—B. sule containing two dev tinctive and apparently unique ovary arrangement, as well as distylous flowers with similarly shaped salverform yellow corollas. Based on the ovary ar- rangement, these genera appear to be closely re- ated. As circumscribed by Ricardi (1963a), Oreopolus included three species and was distinguished by its capsules dehiscent into two valves and flowers lacking petaloid calyx lobe appendages, while Cruckshanksia included seven species (two of which are not recognized in the current treatment) and was distinguished by its capsules dehiscent into four valves and flowers each with one or more petaloid calyx lobe appendages (Ricardi & Que- zada, 1963; Ricardi, 1963a). However, two species that Ricardi placed in Oreopolus, “O. macranthus” (C. macrantha) and “O. palmae” (C. palmae), were not known in fruit according to his own descriptions (Ricardi, 1963a), placed in Cruckshanksia, C. lithiophila, is so sim- ilar to C. macrantha that these can be separated only by the size and shape of the mature calyx lobes. Mature capsules of C. macrantha and C. pal- and another species that mae are now known. In both species the capsules are dehiscent into four or five valves, which either necessitates the transfer of these two species to Cruckshanksia and the separation of the genera only by the number of capsule valves, or, if these species are retained in Oreopolus, the separation of the genera only by the presence versus absence o petaloid calyx lobe appendages. Fruit and floral bie of Oreopolus glacialis.—A. Dehisced capsule that contained two AO eloped seeds.—C. Long-styled flower, partially dissected.— C, D from Ricardi (1963a: fig. 2B, D Neither of these alternatives takes into consid- eration several other unusual features, in particular the ternate leaves and membranaceous stipules of Oreopolus glacialis and the 3—7-lobed foliaceous inflorescence bracts found in the remaining spe- cies. Although Ricardi illustration of “O. macran- thus" (1963a: fig. 5) clearly shows its three-lobed foliaceous inflorescence bracts, this feature was not included in his descriptions of either this species or “O. palmae," although it is present in both taxa, nor was it mentioned in the genus description he presented for Oreopolus. All of the species studied here except O. glacialis also share well-developed calyx lobes that persist and usually enlarge mark- edly on the fruits, and filamentous erose append- ages on the calyx limb and also on at least some of the stipules and inflorescence bracts. These un- usual features support the circumscription of an ap- parently monophyletic group comprised of all the species studied except O. glacialis. Therefore, Or- eopolus is here circumscribed to include one spe- cies, O. glacialis, while the other two species in- cluded in this genus by Ricardi are here assigned to Cruckshanksia. The two species here transferred back to Cruck- shanksia, C. macrantha and C. palmae, do not bear petaloid calyx lobe appendages but do share with the other Cruckshanksia species relatively large ca- lyx lobes that are narrower at the base than at the middle and are enlarged in fruit, in contrast to the smaller triangular lobes of Oreopolus, which are Volume 83, Number 4 1996 Taylor 463 y Oreopolus and Cruckshanksia broadest at the base and do not enlarge at all in fruit. This revised classification also eliminates the problem of placing two very similar species, C. ma- crantha and C. lithiophila, in separate genera, even though these are so similar that when they are in flower they cannot always be separated with confi- dence. The characters that are used here to distinguish Oreopolus from Cruckshanksia are summarized in the key to the genera presented in the taxonomic treatment, below. In the following morphological survey, Oreopolus and Cruckshanksia are compared as they are delimited in the taxonomic treatment. MORPHOLOGY OF OREOPOLUS Habit. Plants of Oreopolus are perennial hem- icryptophytes originating from woody taproots. They form dense low cushions to 75 cm in diameter in open sand or exposed hard soils or rubble. Stems. The stems are stout and the internodes are usually very shortly elongated at most, usually to less than the length of the stipules, so that these overlap. Vesture and crystals. The vegetative portions of the plants range from completely glabrous to mod- erately or sometimes densely puberulous with short unicellular or uniseriate trichomes. The plants con- tain raphides, although these may be difficult to observe in succulent tissues. Leaves. The leaves are ternate or rarely oppo- site, shortly petiolate, simple, and lack domatia. The venation is generally pinnate, although often only the midrib is visible. The cauline leaves are all similar. The stipules are interpetiolar and fused to the bases of the petioles (Fig. 2A). They are generally triangular in shape, though frequently shortly mu- cronate or sometimes shortly bilobed at the apex, with entire margins. They are membranaceous and usually nearly hyaline. Inflorescences. The flowers are terminal and solitary, or terminal and axillary and borne in con- gested groups of two to six. Each flower is sub- tended by bracts that resemble reduced stipules, and are triangular, membranaceous, and entire along the margins, but are acute rather than bilobed at the apex. The leaves at the distalmost nodes ex- ceed and usually to some extent enclose the inflo- rescences. ers. The flowers are hermaphroditic and A (Fig. 1C, D). Long-styled and short-styled forms are represented on herbarium specimens in approximately equal proportions, and frequently are found mixed in the same collection when more than one plant was sampled. These forms differ in relative style lengths, point of insertion of the sta- mens, degree of development of the filaments, and form and pubescence of the corollas. The stigmas are exserted in all mature long-styled flowers seen (in contrast to the apparently delayed elongation of the style in Cruckshanksia; see below). The ovary is inferior and bilocular, turbinate in shape, and pilosulous. Ovary pubescence is not correlated with pubescence of the vegetative parts. Each locule contains two ovules, which are at- tached to elongated axile placentas that extend per- pendicularly from the septum into the locules to form pseudosepta that almost completely subdivide each locule (Fig. 1B). The ovules are anatropous. e calyx limb may be glabrous or puberulous, and is deeply five-lobed with the lobes generally similar in size and shape. The lobes are triangular to lanceolate, with an acute or sometimes shortly emarginate apex. Colleters have not been seen. he corollas are salverform with slender cylin- drical tubes and five triangular lobes with valvate aestivation. They are pale to usually very bright yellow throughout. The shape of the tube is dimor- phic, with the tubes of the short-styled form uni- formly cylindrical in contrast to those of the long- styled form, whic swollen for several millimeters at the top to accommodate the included anthers (Fig. 1C, D). This swollen portion of the long-styled form was referred to as the "garganta" by Ricardi and Quezada (1963). The pattern of pu- bescence is also dimorphic: both the long-styled and short-styled forms are similarly puberulous to pilosulous externally (abaxially) in the tubes and on the lobes, but the short-styled form is glabrous internally (adaxially) throughout the tube while the long-styled form is glabrous in the lower portions of the tube but barbate in the throat. The five stamens are inserted at the top of the corolla tube in short-styled flowers, and at the base of the throat swelling in long-styled flowers. The anthers of short-styled flowers are exserted on slen- der, usually flattened filaments, while those of long- styled flowers are subsessile and included. The an- thers of both floral forms are narrowly oblong and dehiscent by longitudinal slits. The style is slender and filiform, in the short- styled form extending to only a little beyond the middle of the corolla tube and in the long-styled orm shortly exceeding the corolla tube. The paired stigmas are linear and papillose on the adaxial sur- face but glabrous on the abaxial surface in both floral forms. The disk is not evident (*not devel- oped," Ricardi & Quezada, 1963, my translation from the Spanish). are Annals of the Missouri Botanical Garden The infructescences The to chartaceous-walled obovoid Infructescences and fruit. do not differ in form from the inflorescences. is a papery- capsule that is somewhat flattened laterally. The ruit capsule is dehiscent loculicidally and apparently basipetally into two valves. The calyx limb is per- sistent on the capsule, frequently becoming split completely between the lobes. The seeds are planoconvex to somewhat obovoid and somewhat flattened, with a smooth brown to black seed coat. A median sulcus on one side marks the site of attachment to the placenta, and at dispersal some placental material frequently re- mains attache similar to those from species of Rubiaceae with ere. The seeds of Oreopolus are only one ovule and seed per locule and two per fruit; however, because the placenta is enlarged -— perpendicularly to the septum in Oreopolus the seeds are borne at right angles to the true septum, so the sulcus is on the lateral face and the true adaxial face is smooth. MORPHOLOGY OF CRUCKSHANKSIA Habit. sometimes annual geophytes or hemicryptophytes Plants of Cruckshanksia are perennial or originating from ropelike woody taproots and un- derground stems with corky, usually red-brown bark, except for those of C. pumila, which are typ- ically annuals with a smooth gray to brown epi- dermis. The perennial plants generally form loose circular clumps to as much as 40 cm in diameter, with the above-ground stems weak to reclining. Most or all stems die back to the ground during dry or cold periods. The annual plants of C. pumila may grow to 15 cm tall and branch several times, but may also flower at only 3—4 cm tall, with the flowers produced from the second stem node above the cot- yledons. Individual plant size in this species may be correlated with microsite water supply. Stems. The taproots of the perennial species usually produce four to ten or more separate slen- der stems. These may be borne above or below ground. Some species, notably Cruckshanksia ma- crantha and C. lithiophila, typically have twenty to irty or more of these stems. Axillary buds are often well developed and frequently densely pu- bescent. The above-ground stems are generally slender. with well-expanded internodes. More basal inter- nodes are generally quadrate while those at more distal nodes are subterete to irregularly angled or channeled. Although Ricardi and Quezada (1963) used the relative lengths of the more proximal in- ternodes to distinguish species, this character ap- pears to vary with local environmental conditions and is not considered taxonomically reliable here. Plants of Cruckshanksia are generally puberulent to densely villosulous or Vesture and crystals. shortly pilosulous, with short unicellular or unise- riate trichomes found on most or all parts of the plant distal to the hypocotyl. The plants contain raphides. Leaves. The cauline leaves are opposite or oc- casionally subopposite to alternate at more distal nodes, and lack domatia. The venation is generally subpalmate, with one to two pairs of strongly as- cending secondary veins arising in the basal third or half of the blade. In species with very narrow leaf blades (Cruckshanksia verticillata, C. monti- ana), only the midrib may be evident and some- times even this cannot be distinguished. In annual plants of C. pumila, the cotyledons are usually per- sistent on the flowering and sometimes also the fruiting plants. The cauline leaves at the more basal nodes are simple or they may be deeply two- to three-lobed in some individuals of Cruckshanksia verticillata. The leaves of more distal nodes may be consistently simple and distinct from the lobed "floral leaves" or foliaceous inflorescence bracts (“hojas florales” and sensu Ricardi and Quezada, 1963) in C. palmae, C. macrantha, and C. lithio- phila, or in the remaining species the cauline “pseudotrifolios” leaves at the upper stem nodes may vary from sim- ple to lobed and intergrade with the floral bracts. Simple leaves taper to an acute, usually short-pet- iolate base; the lobed leaves are cuneate to rounded at base and sessile to subsessile. The lobed cauline leaves of several species (notably C. montiana and C. hymenodon) often bear linear multicellular ap- pendages on the margins near the bases of the lobes, and sometimes in the sinuses between the lobes. These structures were apparently referred to " by Ricardi and Quezada (1963), and are perhaps comparable to the erose marginal pro- as "escamas jections found on the stipules (see below). Stipule morphology typically varies on an indi- vidual plant. At more basal nodes bearing simple leaves and at the cotyledon-bearing nodes of annual plants, the stipules are interpetiolar and may be free or fused to the petioles (Fig. 2B). The inter- petiolar portion may be triangular or shortly bi- lobed, and entire or somewhat erose. At progres- sively more distal nodes, the stipules typically are progressively more deeply bilobed and more erose, with the linear marginal appendages up to several millimeters long (Figs. 2C, E). At stem nodes clos- est to the inflorescence, the atipules are sometimes completely divided into two lobes, which may be Volume 83, Number 4 Taylor 465 Oreopolus and Cruckshanksia Ne st A d + Ae CX vr DS Vi NETRAA] Ji P d D RMM Ve! Noja és / E ha A A ud " «t o d Ge P IA VL 15 = AS nf p Vj y a [5 " d A KP ad 14, VA ji: E e MIL C fa VES. ^ Bark TRAN dac; ee in Us ES / p t, ^T ” @ A N :. F aN 30 lat’ p Ti X, P AN N EP. SPD (x DUAE JA NU wy: S. Are S T s Wes es FTE B Figure 2. Ravi uoce | of Oreopolus and species of Cruckshanksia. —A. O. pore node near middle of C. hyme stem. —B. n, node near base of stem. —C. node inediscly halo acerco. —E. C. " menod y Donat 172, MO; B, E based on Taylor 10782, MO; € connected by an interpetiolar line or may show no connection at all (Fig. 2D). These lobes are fre- quently fused to the petioles, and typically have strongly erose margins. In several species (notably Cruckshanksia hymenodon and C. pumila), the stip- ules of distal stem nodes are often irregular: only one lobe on one side may be developed, or one lobe may be well developed and free while the other lobe is reduced and fused to the leaf. Irregularly lobed cauline leaves are subtended by irregular stipules, and three-lobed cauline leaves lack stip- C. montiana, node mear middle of stem on, node near middle of stem. All to same aile. A based on 2 based on Gay s.n., MO-393317; D based on Taylor 10781, MO. . hymenodon, ule lobes. Ricardi and Quezada (1963) used degree of stipule lobing to distinguish species, but this fea- ture seems to vary with developmental stage on an individual plant, and several degrees of lobing are typically found on a single stem. This feature is not considered taxonomically reliable here. The correlation between irregular stipule devel- opment and lobed leaf shape suggests that the lobed leaves could be formed by a fusion of ex- panded stipule lobes and the leaf blade, as in Gal- ium L. and some Knoxieae (Robbrecht, 1988). In 466 Annals of the Missouri Botanical Garden this case the lateral leaf lobes would be equivalent to expanded stipule lobes. This possibility is sup- ported by the lack of stipule lobes subtending these leaves, and their sessile rather than petiolate bases. The linear appendages on lobed leaves would then be equivalent to the erose projections characteristic of the stipule margins. More detailed study is need- ed to evaulate this possibility. Inflorescences. The inflorescences are terminal, dichasial to somewhat irregular, strongly congested to subcapitate cymes. The flowers are sessile to subsessile (Fig. 3A). Ricardi and Quezada (1963) reported that the flowers of Cruckshanksia monti- ana are borne on pedicels to 2 mm long, but the structures they measured are similar to those they called inflorescence branches in other species, and are considered inflorescence branches here. ach flower is subtended by an unlobed or usu- ally three- to seven-lobed “floral leaf” or foliaceous inflorescence bract that encloses the flower bud un- til anthesis (Fig. 3A, B). These structures have been variously referred to as leaves or bracts by different authors; they are here considered bracts based on their position, subtending the individual flowers of a cymose inflorescence, and usually dif- ferent form from the cauline leaves. This interpre- tation agrees with that of Jansen (1979) for Ka- Jewskiella Merr. & Perry, in which inflorescence bracts are frequently subopposite or alternate and bear colleters. Jansen considered these bracts to be homologous to leaves, although with the problem that colleters, which are typically found on inflo- rescence bracts of Kajewskiella, are normally found on stipules but not leaf blades or petioles. The pres- ence of colleters in the inflorescence bracts of Cruckshanksia suggests that not only are the bracts generally homologous to leaves, but that they are homologous to the entire leaf, including the stip- ules, as proposed also for Gaertnera Lam. (van Beu- sekom, 1 The inflorescence bracts of Cruckshanksia are generally similar in color, size, and pubescence to cauline leaves, and in those species that occasion- ally have lobed cauline leaves (notably C. hymen- odon and C. pumila), leaves and bracts may inter- grade (see above). Flowers. The flowers are hermaphroditic, and are distylous in all species of Cruckshanksia. (Al- though Ricardi & Quezada (1963) reported that several species had only one floral form, both have now been found for all species.) Long-styled and short-styled forms are represented in living popu- lations of C. pumila and C. hymenodon (pers. obs.) and in herbarium specimens of all species in ap- proximately equal numbers. The forms differ in rel- ative style length, point of insertion of the filaments, and form of the corolla. In long-styled flowers the style apparently elongates after anthesis; this one floral form may have a stylar or “ixoroid” pollen presentation mechanism (Robbrecht, 1988), which implies that these flowers are also protandrous. The presence of this mechanism has not previously been seen in distylous flowers and must be con- firmed by field study, but such a mechanism is sug- gested by the pilosulous pubescence on the upper- most portion of the style and abaxial stigma lobes, to which pollen grains are found adhering in most herbarium specimens that have the style elongated and the stigmas exserted. The ovary is inferior and bilocular, turbinate or somewhat ovoid or obovoid, and usually relatively densely pubescent. The ovary and ovules are sim- ilar to those of Oreopolus (see above). The calyx limb is pubescent similarly to the veg- etative portions of the plant, and deeply to nearly completely divided with a tube portion 1.5 mm long or shorter. Calycine colleters have not been seen. In some species (C. palmae, C. macrantha, C. lith- tophila), the calyx lobes are consistently five, equal or nearly so, and elliptic to oblanceolate (C. pal- mae, C. macrantha) or slightly to strongly stipitate with the upper portion expanded i elliptic to ovate appendage (C. lithiophila). In the remaining species the calyx lobes vary from two to ve per flower on a single plant, and are usually strongly unequal (Fig. 3A). In these species each flower bears one to four stipitate, appendaged, bright yellow or pink petaloid calyx lobes along with usually one or more shorter triangular lobes, into a narrowly and sometimes also one or more elliptic to oblan- ceolate lobes similar to the lobes of the inflores- cence bracts (these last structures were not de- scribed by Ricardi & Quezada, 1963). These petaloid calyx lobe appendages were variously re- ferred to as “sépalos petaloides" and “ufias” by Ri- cardi and Quezada (1963). The calyx typically also bears erose or linear appendages in the sinuses be- tween the lobes (referred to as “escamas calicina- les" by Ricardi & Quezada). These are similar to the erose appendages found on the bracts and stip- ules. Strongly and irregularly unequal calyx lobes are found in several genera of Rubiaceae (e.g., Ron- deletia L., Pentas Benth.), and stipitate calyx lobe appendages are also known from other genera of this family (e.g., Mussaenda L., Warszewiczia Klotszch, Pogonopus Klotszch). Such appendages are often brightly colored and are thought to func- tion in attracting pollinators, in which case they are considered semaphylls (Robbrecht, 1988). The ap- Volume 83, Number 4 1996 Taylor Oreopolus and Cruckshanksia 467 Infl letails of Cruckshanksia hymenodon. with i ule A based on Mahu s.n., MO-3272960; B based on nine 18484, MO. Figure 3. Bu to same sc de pendages found in Cruckshanksia are membrana- ceous, reticulately veined, and generally yellow, or pink to occasionally white in C. hymenodon. The form of the appendages changes markedly during development of the flower. In bud they are narrowly elliptic and acutely to obtusely angled at both apex and base. At anthesis in most species they are ovate or elliptic-oblong and rounded to truncate at the base and apex or with a triangular projection and short mucro at the apex. In some species the calyx lobe appendages continue to enlarge markedly as the fruit develops (see below). Previously published descriptions of most Cruckshanksia species include measurements of fruiting rather than flowering ca- lyx lobe appendages, probably because the plants usually flower and fruit concurrently and the ap- pendages are larger and more striking in fruit. Ricardi and Quezada (1963) used the shape of the petaloid calyx lobe appendages to distinguish species. However, this feature is variable, and at best seems to distinguish only developmental stages of an individual flower. Material cited by them fre- quently shows both of the conditions they used to separate a pair of species on different flowers of the same plant, and this feature is not considered tax- onomically reliable here. The corollas of Cruckshanksia are similar in form to those of Oreopolus (see above). They are bright to deep yellow and frequently marked with darker yellow on the lobes. The corollas of both floral forms are glabrous internally and on the adaxial surfaces of the lobes, and typically pubescent throughout the external (abaxial) surface. Herbari- um specimens of some species, notably Cruck- shanksia hymenodon, show marked variation in co- 1 cm B. Floral bract. —A. Cymule with imr rolla length; this may be real morphological variation, perhaps due to water availability, or it may include differential shrinkage during prepa- ration of dried specimens. The stamens, style, and stigma are similar to those of Oreopolus (see above). Infructescences and fruit. The infructescences are similar to or often somewhat more expande than the inflorescences, with the cyme branches elongating to separate the developing fruits (Fig. 3A). The fruit is a papery- to chartaceous-walled, subglobose to ovoid or oblong capsule. The capsule is dehiscent loculicidally, acropetally, and frequent- ly irregularly into two to five valves. In species with variously two to four appendaged calyx lobes, the number of capsule valves often equals the number of appendaged calyx lobes. The illustration pre- sented by Schumann (1891) suggests that the de- hiscence is basipetal, but this probably was drawn from a specimen and based on a capsule that was crushed during the drying process; similar capsules can be found on specimens that also have capsules dehiscing from the base. The placentas appear to become fleshy as the capsule matures, and ma sometimes be dispersed with seeds still attached. The calyx limb is persistent on the capsule, and the calyx lobes typically enlarge and sometimes also change shape as the fruit matures. In species in which the calyx lobes are all similar, the lobes typically retain their general shape but become en- larged and papery, and sometimes more strongly narrowed at the base. In species in which the calyx lobes are unequal, the smaller lobes do not change noticeably but the appendaged lobes may enlarge and change form markedly, typically becoming sub- 468 Annals of the Missouri Botanical Garden orbicular to reniform, cordate at the base, and emarginate at the apex, with the triangular apex and mucro becoming inrolled or reduced. At ma- turity, these appendages are dry, papery, and usu- ally brown, and may function in dispersal or at least dehiscence of the fruit; such structures are consid- ered pterophylls (Robbrecht, 1988). The seeds are similar to those of Oreopolus (see above), except the seed coat may be smooth to ru- gose. TAXONOMIC TREATMENT This work is based on study of herbarium spec- imens from A, , CONC, CORD, CTES, E, F, GH, K, LP, M, MO, NY, SGO, SI, and UC, and field observations in northern Chile in 1991 and 1993. Measurements given in brackets are taken from Ri- cardi (1963a, 1963b) and Ricardi and Quezada (1963) but were not observed on specimens stud- ied. KEY TO THE GENERA la. Leaves ternate or rarely opposite, thickly suc- or usually imbricated by the limited development of the internodes; leaves dr enlarging in fruit; corolla pubescence dimorphic, differing in long-styled and short-styled forms; capsules basipetally dehiscent into 2 valves o ia Oreopolus lb. Leaves opposite or sometimes s subopposite to al- by simple to usually 3-7-lobed foliaceous bracts similar to the cauline leaves; calyx lobes 2-5(6), elliptic to diu. subulate, or sometimes pro- longed into stipitate petaloid appendages, usu- ally enlarging in fruit; corolla pubescence similar in long-styled and short-styled forms; capsules acropetally dehiscent into 2—5 valves : Cruckshanksia Oreopolus Schltdl., Linnaea 28: 493. 1857. TYPE: Oreopolus citrinus Schltdl. [= Oreopo- lus glacialis (Poepp.) Ricardi]. Perennial, rather succulent, hemicryptophytic low shrubs from a woody taproot, glabrous to pu- bescent, with raphides; stems quadrate to terete, internodes usually not or only shortly expanded. Leaves ternate or rarely opposite, simple, shortly petiolate; blades elliptic to lanceolate, thickly suc- culent, drying coriaceous, without domatia; vena- tion pinnate but usually not evident; stipules interpetiolar, fused to petioles, imbricated by poor development of the internodes, membranaceous, triangular, entire to shortly mucronate or bilobed. Inflorescences terminal, capitate, sessile; flowers 2— 6, distylous, each subtended by a bract resembling a reduced stipule, this sometimes deeply divided with age; hypanthium turbinate; ovary inferior, bi- locular, ovules 2 per locule, anatropous, borne on placentas attached to septum and prolonged into a partial pseudoseptum; calyx limb deeply 5-lobed, lobes equal, generally triangular; corolla slenderly salverform, yellow, in short-styled form the tube glabrous internally and uniformly cylindrical, in long-styled form the tube swollen and barbate at top, lobes 5, triangular to lanceolate, acute, valvate; stamens 5, in short-styled form inserted at top of corolla tube, exserted, with flattened filaments, in long-styled form inserted at base of enlarged throat of corolla tube, included, subsessile, anthers in both forms basifixed, narrowly ellipsoid to oblong, dehiscent by longitudinal slits; styles filiform; stig- mas 2, linear, papillose, in short-styled form borne near middle of corolla tube and abaxially glabrous, in long-styled form exserted by several mm an abaxially pilosulous; disk rudimentary. Fruits cap- sular, obovoid, laterally somewhat flattened, papy- raceous to chartaceous, dehiscent loculicidally and basipetally into 2 valves, calyx limb persistent, similar to that of flowers; seeds planoconvex to ob- ovoid, somewhat compressed, sulcate at attach- ment, brown, smooth to somewhat rugose. One species of dry cool regions in the Andean Cordillera, central Chile and Argentina to Tierra del Fuego = glacialis (Poepp.) Ricardi, Gayana 6: oepp., in 1 . IX Region.’ Prov. Malleco: Vol- cán Antuco, 8500 ft. [2750 m]. Mar. 1829, Poeppig 59 (holotype, B not seen, destroyed; isotype, CONC-28779). Oreopolus citrinus Schltdl., Linnaea 28: 492. 1857. TYPE: Chile. X Region. Prov. Valdivia: É sig de 2r co, in tierra Pehuelchorum," Dec. 1854, Lech- * For map of regions in Chile see Flora de Chile, Vol. 1, edited by Clodomiro Marticorena and Roberto Rodrí- guez, published in 1995 by Universidad de Concepción, Chile. Volume 83, Number 4 1996 Taylor 469 Oreopolus and Cruckshanksia ler/Hohenacker 2895 (holotype, HAL not seen; isotypes, CONC, K, NY, photo (neg. #2479) SGO). Oreopolus patagonic ce Secs , Rev. Fac. Agron. Veterin. La Plata 3(30-31): 525. 1897. Cruckshanksia patagon- ica Speg.) Macloskie, Rep. Princeton Univ. Exp. Pa- : 740. 1905. TYPE: Argentina. "Prov. : San Julián, río Deseado, o s.n. (holotype, LP not seen; isotypes, = Oreopolus glacialis var. pilosus Ricardi, Rev. Fac. Ci. E: Argentina. Prov. San m, 9-17 Jan. 1954, A. Ruiz Leal 15678 (holotype, herb. H. Ruiz Leal, not seen; isotype, CONC). Plants forming dense lawns or cushions to 75 cm diam., glabrous or moderately to densely pu- berulous, usually 3-8 times branched, from tap- roots to 15 mm thick, with red somewhat corky bark; stems 2—10(20) cm long, suberect. Leaves with petioles 2-7 mm long; blades 8-15 mm long, mm wide, at apex acute, at base acute to rounded; stipules 1-3 mm long, smooth to some- what costate, acute to mucronate or bilobed with mucro or lobes to ca. 0.5 mm long. /nflorescences ca. 1 em long and wide excluding corollas; flowers with hypanthium 1-2 mm long, pilosulous; calyx limb membranaceous, glabrescent, (1)3-4 mm long, divided for 4—%, lobes (0.5)1-2 mm long and wide, triangular to lanceolate, acute to emar- ginate or shortly bilobed; corolla pale to bright yellow, puberulous to pilosulous externally (abax- ially, tube 13-16 mm long, 0.3-0.5 mm diam. near middle, enlarged throat of long-styled form 1.5-2 mm long and in diam., lobes 4-5.5 mm long, 1-1.5 mm wide at base; anthers 1-1.5 mm long, filaments 1-1.5 mm long in short-styled form; stigmas 0.3-0.5 mm long, in long-styled form exserted by 1-5 mm. Capsules 5-8 mm long, 3-6 mm wide, pilosulous; seeds 4-6 mm long, 2— 3 mm wide. Illustration: Ricardi (1963a: fig. 3). Distribution and habitat. Southern Andean Cordillera, central Chile and Argentina to Tierra del Fuego (32°-53°S), at 2500-3500 m in the northern part of its range, to as low as 300 m in the southern part, usually in open sand, rubble, or hard soil. Collected in flower October through March, in fruit December through March. The leaves and stems of Oreopolus glacialis range from smooth to markedly papillose, with en- larged epidermal cells, throughout its range. In general, surface texture is correlated with plant size: plants with relatively small leaves and short internodes are usually smooth, while those with relatively large leaves and expanded internodes are papillose. No geographic pattern or correlation with other characters is evident for plant size, which probably varies primarily with microhabitat. The unpublished names “Cruckshanksia oblon- ga F. Meigen,” “Conchospermum oblongum Phil.,” “Conchospermum ovatum Phil.,” “Conchospermum nudicaulis Phil.,” “Oreocaryon nivalis Kuntze ex K. chum.," “Oreopolus oblonga Phil. ex F. Meigen,” and “Oreopolus oblongus Phil.” have all been used for Oreopolus glacialis. Although some of these names have been listed in synonymy with Oreopolus elacialis in previous treatments, none of them has been validly published. Ricardi originally separated variety pilosus Ri- cardi from the typical variety by its “leaves pubes- cent on both sides, without glandular punctations,” in contrast to the “leaves glabrous with brown glan- dular punctations abundant on the lower surface” of variety glacialis (Ricardi, 1973: 218; my trans- lation from the Spanish text). Occasional plants of Oreopolus glacialis from throughout its range are pubescent on vegetative parts with trichomes 0.1- 0.2 mm long, and these seem to be the plants that Ricardi intended to segregate as variety pilosus. No difference is evident in the density, character, and distribution of pubescence on flowers and fruits of plants he placed in the two varieties. All of the representatives of “var. pilosus” were collected near San Carlos de Bariloche in southern Argentina, where vegetatively glabrous plants have been col- lected apparently sympatrically (e.g., Buchtien s.n., CONC) e structures that Ricardi (1963a) de- scribed as “glandular punctations” are found on oc- casional specimens from throughout the range of this species. These structures may be found on gla- brous or pubescent plants, on only the abaxial leaf surface, on both surfaces, or throughout the vege- tative parts of the plant. They do not appear to be actual glands, but rather the collapsed walls of en- larged epidermal cells, and are found only on spec- imens with a markedly papillose epidermis. However, some plants from near Bariloche, in- cluding the type collection of variety pilosus, are unusual in having a reduced calyx limb, ca. 1 mm long (e.g., Boelcke 10049). This feature was not cit- ed by Ricardi as distinctive, but does seem to be locally frequent in the populations that Ricardi in- tended to separate in his variety. This variety is not recognized here, though further study may support its separation (N. Bacigalupo, pers. comm.). Representative specimens examined. ARGENTINA. Chubut: Departamento Languifieo, 50 km al S de Tecka, Correa et al. 10390 (BAB). Mendoza: San Carlos, Laguna 9992 (BAB). Rio Negro: Parque Nacional Nahuel Huapi, 470 Annals of the Missouri Botanical Garden subida al Granito, Boelcke & Correa 5936 (BAB). Santa Deseado, Picada Río Deseado-Bahía Nodales, Cor- rea & Nicora 3443 (BAB). Tierra del Fuego: Estancia Sarmiento, foothills to N of Sierra Beauvoir, Goodall 4380 (BAB, MO). CHILE. METROPOLITAN c Prov. Cordillera: Paso de los Peladeros, Jan. 1933, oa ; n. Grandjot (MO-1160900). VI REGION. Prov. Colchagua: Co edil. lera y ae a Pirion 79 (CONC, GH). VII REGION. rov. Curi pius of Lake Planchon, Zóllner 8 (MO). 9696 (MO). P v. Talca: entre Paso Pec ehinecha es y Laguna del Maule, "Ricardi et al. 980 (CONC). VIII RE- GION. Prov. Biobío: faldeos de Volcán Antuco, 5.5 km al S del Refugio de Sky, Miu » Baeza 11092 (CONC). IX REGION. Prov. Malleco: Volcán Loquimay, e Constance 10936 (CONC). XI SECUN Prov. Aisén a Ricardi & Matthei 515 (CONC). XII REGION, Prov. M anes: Estancia Penitente, Río Penitente, near ridge crossing river to E of St. Palermo, Moore 2264 (BAB). Prov. ranza: Las Cumbres, Bagu- ales, Ricardi & Matthei 378 (CONC). ] Cruckshanksia Hook. & Arn., Bot. Misc. 3: 361. 33, nom. cons., not Cruckshanksia Miers, Trav. 2: 529. 1826, Iridaceae, Type = Cruck- shanksia graminea Miers; not Cruckshanksia Hook., Bot. Misc. 2: 211. 1831, Geraniaceae or Ledocarpaceae, Type — Cruckshanksia cis- tiflora Hook. [= Balbisia peduncularis (Lindl.) Don]. TYPE: Cruckshanksia hymenodon Hook. & Arn Rotheria Meyen, Reise um die Erde 1: 402. 1834. TYPE: sin Visage Meyen (= Cruckshanksia hy- nodon k. & Arn.). Annual or perennial herbs or shrubs, usually geophytic or hemicryptophytic from a woody tap- root, glabrescent to usually pubescent, with raph- ides; stems quadrate to subterete or irregularly channeled, internodes generally expanded; coty- ledons usually persistent and photosynthetic on annual plants of C. pumila. Leaves opposite or sometimes subopposite to alternate at more distal nodes, simple or sometimes deeply 2—3-lobed, shortly petiolate to subsessile; blades elliptic to narrowly so, without domatia; venation subpalmate with 1-2 pairs of strongly ascending secondary veins or sometimes not evident; stipules interpe- tiolar, free or fused to petioles or leaf bases, on more basal nodes generally triangular to shortly bilobed with margins entire to shortly erose, on more distal nodes more deeply to completely bi- lobed with margins erose, the erose appendages sometimes elongate (“escamas” sensu Ricardi & Quezada, 1963) or the stipules reduced or absent when leaves lobed. /nflorescences terminal, con- gested-cymose to subcapitate, branched (0)2-8 times; flowers (1)3-15, distylous, subsessile, each subtended by a foliaceous bract, these simple to deeply 3-7-lobed, similar to cauline leaves in size, texture, color, and venation, sessile, deeply 3-7-lobed, entire to erose with prolonged linear appendages; hypanthium cylindrical to turbinate or subglobose, densely pilosulous to villous; ovary inferior, bilocular, ovules 2 per locule, anatropous, borne on placentas attached to septum and pro- longed into a partial pseudoseptum; calyx limb deeply 2—5(6)-lobed, lobes often erose with linear appendages (*escamas calicinales" sensu Ricard & Quezada, 1963), equal (C. macrantha, C. pal- mae, C. lithiophila) or strongly unequal (remaining species), narrowly triangular to elliptic or some- times prolonged into a NADIE petaloid append- age, this elliptic to ovate, membranaceous ulate-veined, puberulent to glabrescent, brahily colored (semaphylls; “uñas,” “sépalos petaloides” sensu Ricardi & Quezada, 1963); corolla slenderly salverform, bright to deep yellow, externally (abax- ially) puberulent to pilosulous, internally gla- brous, in short-styled form the tube uniform, in swollen at top 1963), valvate; long-styled form the tube (“garganta” sensu Ricardi & Quezada, lobes 5, triangular to lanceolate, acute, stamens 5, in short-styled form inserted at top of corolla tube, exserted, with flattened filaments, in long-styled form inserted at base of enlarged throat of corolla tube, included, subsessile, an- thers in both forms basifixed, narrowly ellipsoid- oblong, dehiscent by longitudinal slits; styles fili- form; stigmas 2, linear, papillose adaxially, pilose abaxially, in short-styled form borne near middle of corolla tube, in long-styled form exserted by several mm; disk rudimentary. Infructescences sometimes somewhat more expanded and open than inflorescences but otherwise similar; fruits capsular, subglobose to ellipsoid-oblong or ovoid, somewhat didymous, papyraceous to chartaceous, dehiscent loculicidally, acropetally, and often ir- regularly into 2-5 valves, calyx persistent, often enlarging, with petaloid appendages (pterophylls) becoming ovate or orbicular to reniform, rounded to truncate or emarginate at apex, rounded to usu- ally cordate at base, dry, papery, whitened to yel- lowed or brown; seeds ellipsoid to ovoid, somewhat brown to compressed, sulcate at attachment, black, smooth to verrucose. Seven species of arid regions from the Pacific coast to the Andean Cordillera in northern Chile, two species (Cruckshanksia hymenodon, C. macran- tha) extending into adjacent Argentina. Volume 83, Number 4 1996 Taylor Oreopolus and Cruckshanksia KEY TO SPECIES OF CRUCKSHANKSIA la. Calyx lobes equal, n none bearing petaloid appendages or all lobes bearing them, elliptic to oblanceolate or with stipitate elliptic to ovate blades or appendages in flower, these sometimes becoming orbicular to broadly elliptic in fruit. 2a. ig blades 15-40 mm long, sharply acute at apex; stipules 1.5-2 mm long; calyx lobes in flower 10-15 . pal m long, in fruit 12-23 mm long, without stipes 2 Leaf blades 3.5-1 a mae 8-20 mm long, obtuse to rounded at apex; stipules 0.8—6 mm long; calyx lobes in flower ] mm Ael including pt when present, in fruit 7— E mm long, elliptic to dilanceabite and without oup or 9-13 mm long, ovate, and with stipe s 10-15 m 3a. Calyx lobes. elliptic to perdes hr stipes, in Sd. 3.5-5 mm long, in fruit 7-17 mm long 2 C. macrantha 3b. Calyx lobes in flower composed of stipes 3-6 mm long each bearing an elliptic to ovate appendage 3-5 mm long, in fruit the stipes 10-15 mm long with appendages ovate, 9-13 mm lon x lithiophila — c the remaining (1-4 lobes elliptic to 4a pum near middle 3. C. . Calyx lobes unequal, 1—3(4) bearing peo appendages, these orbicular to broadly elliptic i in flower, and r or subulate s annual or rarely eatin perennial; corolla lobes 1-2.5 mm long; corolla tube 0.1-0.3 5. C. pumila 4b. Plans perennial; corolla lobes 2-4.5 mm long; corolla tube 0.3-1 m Petaloid calyx appendages pink or white (frequently fats salle leaves E VE mm ins diam. near middle. nodon 5b. der calyx appendages dip es 0.8—4(6) m . E auline leaves simple or wide. mes 3-lobed; alos calyx lobe open 6-10 mm long and 8-15 mm wide in dowel to 12 mm long and 18 mm wide . C. montiana 6b. pain leaves simple to usually 2-3- EMO petaloid calyx lobe appendages 3—6 m and 2-7 mm wide in flower and fru 1. Cruckshanksia palmae Clos, in Gay, Fl. Chil. 3: 194. 1848. bie Sieh palmae (Clos) Ricardi, Gayana, Bot. 6: 1963. TYPE: Chile. IV Region. Prov. dic iie ‘cerros arenosos de Guanta [sic; Huanta] en el valle de Coquimbo a una altura de 6 6 7,000 pies” [1935-2260 m], Nov. 1836, C. Gay s.n. (holotype, P not seen; isotype, SGO). Perennials, densely puberulous to villosulous throughout, from well-developed taproots to 5(10) mm thick, with red bark frequently peeling in plates; stems usually 5-10, to 15 cm long, weak. Leaves simple, subsessile; blades elliptic to narrow- ly so, 1540 mm long, 4-7 mm wide, sharply acute at apex, acute to attenuate at base, subcoriaceous; stipules at more basal nodes deltoid to broadly tri- mm long, acute to shortly bifid or bilobed, entire to slightly erose, at more distal nodes deeply bilobed, lobes triangular to ligulate, 2.54 mm long, acute to acuminate, erose. Inflo- rescences 1-4 cm long and wide excluding corollas; bracts with lobes narrowly elliptic to oblanceolate, acute to acuminate at apex, acute at base, centr. lobe 24-35 mm long, 3.5-10 mm wide, lateral lobes 2, 14-28 mm long, 2-4 mm wide; flowers with hypanthium ca. 1 mm long; calyx limb pilosulous, lobes (4)5, equal, elliptic to oblanceolate, 10-15 mm long, 2-5 mm wide, acute at apex and base, entire or erose with minute appendages; corolla ex- ternally moderately to densely pilosulous, tube 18— d. iG. Meses * 8 24 mm long, 0.4-0.8 mm diam. near middle, en- larged throat of long-styled form 2-3 mm long and 2.5-3 mm diam., lobes (4)5, 4-5 mm long, 1-2 mm wide at base; anthers 2-2.5 mm long, filaments 1.5-2 mm long in short-styled form; stigmas 0.5—1 mm long, in long-styled form exserted by 2-3 mm. Capsules ca. 5 mm long and wide, persistent calyx lobes 12-23 mm long, 3-5 mm wide; seeds 2-3 mm long, 1-2 mm wide. Illustrations: Ricardi (1963a: fig. 4, as “Oreopolus palma”), Squeo et al. (1994: fig. 78, as “Oreopolus palmae"). Distribution and habitat. p Cordillera of central Chile (30?—33?30'S) at 2000—4000 m, in open sand or rubble. colla in flower November through March, in fruit January through March. This species is distinguished by its simple cau- line leaves that are acute at the apex, and five calyx lobes that are all similar in form and size, elliptic to oblanceolate, and 10-15 mm long in flower be- coming 12-23 mm long in fruit. It is similar veg- etatively to Cruckshanksia hymenodon, which can be distinguished by its usually two or three petaloid calyx lobe appendages, and to C. macrantha, which can be distinguished by its leaves, which are 8-15 mm long and obtuse to rounded at the apex. The consistently trilobed inflorescence bracts were not included in Ricardi’s (1963a) description of this species. The fruits are described for the first time here. The specific epithet commemorates Sefior Ga- 472 Annals of the Missouri Botanical Garden briel Palma, so although it was originally published The was written s “palma,” it must be corrected to “palmae.” name “Cruckshanksia virescens Clos” on labels of at least one of Gay’s collections of this species, but was never published. Representative specimens examined. CHILE. IV RE- GION. Prov. Choapa: 2 hrs. by horse E of Cuncumen, fd e (CONC, GH). Prov. Elqui: at dam, La . 108 km from Rivadavia, Johnston 16388 (GH, K. MO). “Prov. Limari: Río Blanco, Jiles 1124 (CONC, M). V REGION. Prov. Cordillera: in valle San Ramón, Grandjot s.n. (CONC-1074, CONC-55546, M, MO- 1063540, MO-116091). N . Cruckshanksia macrantha Phil., Linnaea 33: 97. 1864. ea i macranthus (Phil.) Ricar- di, Gayana, Bot. 963. TYPE: Chile. IV Region. Prov. Cola Quebrada Escondida, 1860-1861, Volckmann s.n. (holotype, SGO- 26866; isotype, CONC-43455) Perennials, densely and finely puberulent to vil- losulous throughout, forming mats to 1 m or more diam., from well-developed taproots to 5(10) mm thick with red bark often peeling in large flakes: stems 3-30, to ca. 15 cm long, weak. Leaves simple, subsessile; blades narrowly to rather broadly ellip- tic to lanceolate, oblanceolate, or ligulate, 8-15 mm long, 2-5 mm wide, obtuse to rounded at apex, acute to rounded and then abruptly attenuate at base, subcoriaceous; stipules at more basal nodes broadly triangular, 0.8-1 mm long, acute to usually bilobed or shortly bifid, entire to slightly erose, at more distal nodes deeply bilobed, lobes triangular to ligulate, 0.8-1 mm long, acute to rounded, erose. Inflorescences 1-2 cm long and wide excluding co- rollas; bracts with lobes narrowly elliptic to oblan- ceolate, acute to rounded at apex, acute to attenuate at base, central lobe 4-11 mm long. 1-5 mm wide, lateral lobes 2, 4-10 mm long, 1-2 mm wide; flow- ers with hypanthium 1.5-2[2.5] mm long; calyx limb densely pilosulous, lobes (4)5, equal, elliptic to oblanceolate, 3.5-5 mm long. 0.8-2 mm wide, acute at apex and base, entire or erose with ap- pendages 0.5-1 mm long; corolla externally mod- erately to densely pilosulous, tube 18-23 mm long, .3-0.6 mm diam. near middle, enlarged throat of long-styled form 2-3 mm long and 1.5-2 mm diam., lobes 5, 4—6 mm long, 1-2 mm wide at base; anthers 1.5-2 mm long, filaments 1-1.5 in short-styled form; stigmas 0.5-1[2] mm long, in mm long long-styled form exserted by 1-3 mm. Capsules 3— 5 mm long, 3.5-8 mm wide, persistent calyx lobes 7-17 mm long, 2.5-10 mm wide; seeds 2.5—3 mm long, 1.5-2 mm wide. Illustrations: Ricardi (19632: fig. 5. as "Oreopolus macranthus"); Squeo et al. (1994: fig. 77, as "Oreopolus macranthus"). Distribution and habitat. Andean Cordillera of northern Chile and adjacent Argentina (26?30'— 32°15'S) at 1100-4300 m, most frequently collect- ed above 3000 m. Collected in flower November through February, in fruit January through Febru- This species is distinguished by its simple cau- line leaves that are obtuse to rounded at the apex, and by the five calyx lobes that are similar in form and size, elliptic to oblanceolate, and 3.5-5 mm long in flower becoming 7-15 mm long in fruit. It is similar to Cruckshanksia lithiophila, which can be distinguished by its calyx lobes all bearing pet- aloid appendages with slender stipes; these species may not be distinct (see discussion under that spe- cies). It is also similar to C. palmae, which can be distinguished by its leaves which are 15—40 mm long and acute at the apex. The consistently tri- lobed inflorescence bracts were illustrated for this species by Ricardi (1963a: fig. 5D) but not includ- ed in his description. The fruits are described for the first time here. ded diio Md specimens examined. ARGENTINA. San Juan: Junta del Río Cura, Río de la Tagua below its confluence ien the Río de la Sal, Johnston 6162 (CONC, F, GH). CHILE. III REGION. Prov. Chañaral: cerca del Salar de Pedernales, Zóllner s.n./herb. Garaventa 5324 CONC), Zóllner 786 (CONC). Prov. Copiapó: Cordillera Río Turbio, Cerro Cadillal, Werdermann 936 (CONC, E, F, GH, M, MO, NY). Prov. Huasco: Laguna Valeriano, ei 6065 (CONC, F, GH, K). IV REGION. Prov. a: Río Piuquenes, San Román s.n. (CONC-29868). Prov. "Elqui. Baños del Toro, Doña Ana, Werdermann 222 CONC, E, F, GH, K, MO). — = 3. Cruckshanksia lithiophila Ricardi, Gayana, Bot. 7: 3. 1963. TYPE: Chile. III Region. Prov. Copiapó: Quebrada Vizcachas, a 43 km de La Puerta, 3100 m, 1 feb. 1963, M. Ricardi, C. Marticorena Matthei 659 (holotype, CONC-28752; isotype, 5GO-73933). Perennials, puberulent to densely villosulous throughout, from well-developed taproots to 2.5 cm thick with red corky bark; stems 10-30, to ca. 10 cm long, weak. Leaves simple, subsessile; blades lanceolate to elliptic, 5-20 mm long, 2-7 mm wide, obtuse at apex, attenuate at base, subcoriaceous; stipules at more basal nodes triangular, 1-2 mm long, acute to usually bilobed, erose, at more distal nodes deeply bilobed, lobes triangular, 1-6 mm long, acute, erose. Inflorescences 1-2.5 cm long and 1—4 cm wide excluding corollas; bracts with lobes narrowly elliptic to oblanceolate, acute to rounded Volume 83, Number 4 1996 Taylor 473 Oreopolus and Cruckshanksia at apex and base, central lobe 9-21 mm long, 3-6 mm wide, lateral lobes 2, 9-18 mm long, 2-5 mm wide; flowers with hypanthium 1-1.5 mm long; ca- lyx limb villosulous, lobes 5, equal, entire or erose with appendages 0.5-3 mm long, with stipes 3-6 mm long, appendages elliptic to ovate, 3-5 mm long, 2-4 mm wide, obtuse to rounded at apex, cu- neate at base, somewhat petaloid, color unknown; corolla externally pilosulous, tube 10-14.5 mm long, 0 8 mm diam. near middle, enlarged throat of long-styled form 1.5-3 mm long and 1.5- 2 mm diam., lobes 5, 2-3 mm long, 1-1.5 mm wide at base [Ricardi (1963b) said 2-2.5 mm long, but his illustration shows the lobes ca. 4 mm long if the proportions are correct]; anthers 1-1.5 mm long, filaments 1-1.5 mm long in short-styled form; stigmas 1-1.5 mm long, in long-styled form exsert- ed by 1-2 mm. Capsules 44.5 mm long, ca. 4 mm wide, persistent calyx lobes with stipes 10-15 mm long and appendages ovate, 9-13 mm long, 5-8 mm wide, acuminate; seeds 3—4 mm long, ca. 2 mm wide. Illustrations: Ricardi (1963b: figs. 1, 2); Ri- cardi and Quezada (1963: fig. 2). Distribution and habitat. Andean Cordillera of northern Chile (27%-27%50'S) at 3300-3600 m, in open rubble. Collected in flower and fruit generally concurrently in November, January, and March. This infrequently collected species is distin- guished by its simple cauline leaves that are obtuse at the apex, and by the five calyx lobes that are each composed of a stipe and petaloid appendage. It is similar and probably closely related to Cruck- shanksia macrantha, which is distinguished by its elliptic to oblanceolate calyx lobes that are nar- rowed but not stipitate at the base and generally shorter (3.5-5 mm long in flower and 7-17 mm long in fruit). These two species are maintained here, although considering the variability found in other species of Cruckshanksia, C. lithiophila may rep- resent only a well-marked variant of C. macrantha. The difficulty of distinguishing these species is demonstrated by Werdermann 936: the specimens at F and NY were annotated by Ricardi as “Oreo- polus macranthus” (i.e., C. macrantha), while the specimen at GH, which appears similar in all re- spects, was identified by him as C. lithiophila. The petaloid calyx lobe appendages are usually not well developed at anthesis but enlarge mark- edly as the fruit develops. Ricardi’s description (1963b) does not make this distinction; his mea- surements apply to the fruiting stage. epg sae Sapp examined. CHIL a III REGION. pó: quebrada Chinches, a Chinches, Prov. Cop T " y 537 (CONC); Ber EL Colorado, a mitad de falda, Marticorena et al. 582 (CONC); qq de Vizcachas, Ricardi & Marticorena 3779 (C ino internacional de Copiapó a Tinogasta, quebrada Co- docedo, Villagrán & VEU 4668 (CONC). 4. Cruckshanksia hymenodon Hook. & Arn., Bot. Misc. 3: 361. 1833. TYPE: Chile. IV Re- gion. Prov. Coquimbo: Coquimbo, Cruckshanks 3 (lectotype, designated here, E). SYNTYPES: Chile. IV Region. Prov. Coquimbo: Coquimbo, Cuming 861 (K), = Cruckshanksia montiana; Cuming 875 (K), = Cruckshanksia montiana. Rotheria lanceolata Meyen, Reise um die Erde 1: 402. 1843. TY nknown. Cruckshanksia bustillosi Phil., Linnaea 28: 696 C gion Metropolitana: Prov. Santiago: Cordillera de la prov. de Santiago, 1857, Bustillos s.n. (holotype, SGO-56889, photo GH Perennials, densely villous to pilosulous throughout, from well-developed taproots to 1(2) cm thick with corky reddened bark; stems 3-10, 15— 30 cm long, weak. Leaves simple or rarely 2-3- lobed, petioles 1-3 mm long; blades elliptic to nar- rowly so or lanceolate, 7-30[40] mm long, 2- 9.5[15] mm wide, acute at apex, attenuate at base, membranaceous to subcoriaceous; stipules at more basal nodes triangular, 2-4 mm long, acute to bi- lobed, entire to erose, at more distal nodes deeply bilobed, lobes triangular, 2-4 mm long, acute to acuminate, erose with appendages to 1 mm long. Inflorescences 1-3 cm long and 2-10 cm wide ex- cluding corollas; bracts with lobes elliptic to nar- rowly so, acute at apex and base, central lobe 10— 13 mm long, 3-7 mm wide, lateral lobes 2(4), 8— 12 mm long, 1-5 mm wide; flowers with hypanthi- um 1-1.5 mm long; calyx limb pilosulous to gla- brescent, lobes (4)5(6), erose with appendages 0.4— 5 mm long, unequal, (1)2-3(4) lobes with stipes 2— 9 mm long, appendages ovate to suborbicular or oblong, 3-11 mm long, 2.5-11 mm wide, rounded to truncate at apex and base, petaloid, pink or sometimes white (often drying yellow), remaining lobes subulate and 4-10 mm long or infrequently similar to lobes of bracts and 3-8 mm long by 1- 2 mm wide; corolla externally moderately to dense- ly pilosulous, tube 9-23[26] mm long, 0.3-1 mm diam. near middle, enlarged throat of long-styled form 1.5-3[4] mm long and 1.5-2 mm diam., lobes 5, 3-4 mm long, 1-2 mm wide at base; anthers 1.5— 2[2.5] mm long, filaments 1-2 mm long in short- styled form; stigmas 1-2 mm long, in long-styled form exserted by 1-3 mm. Capsules 3—4 mm long, 3.5—4 mm wide, persistent calyx lobes with stipes 4—9 mm long and appendages 5-11 mm long, 4— Annals of the Missouri Botanical Garden 11 mm wide; seeds ca. 2 mm long, ca. 1 mm wide. Illustrations: Ricardi and Quezada (1963: fig. 3), Schumann (1891: fig. 8Q, R). Distribution and habitat. Northern Chile and adjacent Argentina (26°-33°50'W) in arid regions at 20-2950 m. Collected in flower and fruit usually concurrently, at low elevations in October and No- vember, at high elevations December through March. This is the most commonly collected species of Cruckshanksia. It is distinguished by its dense vil- lous to pilosulous pubescence on all vegetative parts, usually simple cauline leaves, and calyx with usually 2-3 lobes bearing pink or white petaloid appendages. Label data often describe the color of either the petaloid calyx appendages (pink or white) or the corollas (yellow) without specifying which structure is described, and this has been the source of some confusion. Additionally, the pink petaloid appendages frequently become yellow when dried. In the original description of Cruckshanksia hy- menodon, Hooker and Arnott noted that two differ- ent forms could be distinguished, “a. foliis incanis" and “B. foliis minus pubescentibus," and cited a collection made by Cruckshanks for the first and two collections by Cuming for the second. These two forms correspond to what are here considered two distinct species, C. hymenodon and C. monti- ana, respectively. Although all previous authors ” have interpreted these taxa in this way, no one pre- viously has clearly selected a lectotype. This is done here. Even though both of Cuming's collec- tions have several duplicates and are more widely distributed than the single Cruckshanks specimen, this last specimen is chosen here in order to fix the application of this name to the species to which it was applied by Clos (1848) and all subsequent au- thors. Ricardi and Quezada (1963) recognized two va- rieties in this species: var. hymenodon and var. bus- tillosii. They distinguished variety bustillosii in their key based on its “linear-lanceolate leaves, se- pals 2, corollas 16-20 mm long, and plants with a tendency to form lawn 99 s," in contrast to “leaves mostly oblonglanccolate: sepals 2-3(5), and corol- las 26-30 mm long” in variety hymenodon (Ricardi & Quezada, 1963; my translation from the Spanish; no habit information given by them for var. hymen- odon). They distinguished variety bustillosii in their diagnosis and discussion only by its “chamaephytic habit, linear-lanceolate leaves, and emarginate apex of the sepals” (Ricardi & Quezada, 1963; my translation from the Spanish; in their usage, “sepal” denotes only calyx lobes that bear petaloid ap- pendages). However, many specimens bear some characteristics of both varieties and cannot be clas- sified according to these criteria, including speci- mens studied by Ricardi and Quezada. For exam- ple, Werdermann 158, placed by them in variety hymenodon, has corollas 18 mm long and petaloid calyx lobe appendages that vary from obtusely an- gled to truncate or in a few flowers shortly emar- ginate, and Wagenknecht 18505 (CONC, GH, MO), also placed by them in variety hymenodon, has sev- eral petaloid calyx lobe appendages that are as deeply emarginate as those shown in their illustra- tion of variety bustillosii, along with others that are obtusely angled. Several leaves on specimens iden- tified as variety bustillosii (e.g., Worth & Morrison 16568; Morrison 16998) are similar in shape to those of typical variety hymenodon, although they are generally shorter than the average (but not the extreme) for variety hymenodon and some other leaves on the same plant are relatively narrower. The plants segregated as variety bustillosii are all from relatively higher elevations, but specimens of variety hymenodon are cited from throughout this same range, so no geographic or ecological differ- ence is evident. The plants classified as variety bus- tillosii are generally markedly smaller in stature, and on collections with abundance indicated they are usually said to be rare or infrequent, in contrast to the usual description of the larger, more robust plants of variety hymenodon as common or abun- dant. It seems likely that these plants are no more than reduced representatives of this species, prob- ably due to microhabitat conditions, and this vari- ety is not recognized here. Répresentattue specimens examined. ARGENTINA. San Juan: Departamento Calingasta, al S de Barreal c Kiesling et in 8051 (MO, SI). CHILE. III RE- GION. 19 = al interior id Llanta, Ri- cardi et ab 1563 (CONC). Prov. Copiapó v Copiapó along Rte. 5, ca. 2 km N of ie entrance to Mina Flor del Llano, Taylor et al. 10781 (CONC, MO), 10782 (CONC, MO). Prov. Huasco: Alto del Carmen, oe 158 (E, GH, K, M, MO). IV REGION. Prov. : e a Vega Escondida, 3 a a Pe 16998 ( rov. Elqui: Baños del” Toro. Werdermann 211 (E, F. F, GH, 5 M, MO), Zóllner 10392 (MO). Prov. Limari: Tos Molles, Jiles 4762 (CONC, M). V REGION. Prov. Petorca: Cerro Chache, 5 hr. by horse SE of Patagua Mine, 18 km E of La Ligua, Morrison 17025 (GH, MO). REGION METROPOLITANA: Prov. Cordillera: Lagunillas, Zóllner 10644 (MO) 5. Cruckshanksia pumila Clos, in Gay, Fl. Chil. 3: 196. 1848. TYPE: Chile. IV Region. Prov. Coquimbo: “vecindad de Arqueros, 8bri 1836," C. Gay 242 [holotype, P not seen, pho- to (neg. #4646) SGO; isotypes, P not seen, photo (neg. 44647) SGO, probable isotype F- 972004] Volume 83, Number 4 1996 Taylor 475 Oreopolus and Cruckshanksia p tripartita Phil., Viage Des. Atacama 200. cl E: Chile. III Region. Prov. Chañaral: Pan car, R. A. Philippi s.n. (holotype, SGO—56883; sip SCO- 43336/herb. F. Philippi 951). Cruckshanksia chrysantha Phil., Anales Univ. Chile 41: 730. 1872. TYPE: Chile. III Region. Prov. Huasco: Yerbas Buenas, Oct. A T King s.n. (holotype, SGO-56904; isotypes, E, K-ex herb. E. C. Reed, photo (neg. 2472) SCO, SGO-43356/herb. F. Phil. 948b, SGO—43348/herb. F. Phil. 948a, photo we GH). Crckhansia geisseana Phil., Anales Univ. Chile 85: . 1894. TYPE: Chile. III Region. shy) Copiapó: e Bandurrias, 1885, Geisse s.n. (lectotype, designated here, SGO-43365/herb. E Phil. 1935, photo SGO; isotype, GH). Cruckshanksia darapskyana Phil., Anales Univ. Chile 85: 738. 1894. TYPE: Chile. II Region. Prov. Antofa- gasta: Taltal, 1889, Dr. L. Darapsky s.n. (holotype, SGO-56884; isotype, SGO-43346/herb. F. Phil. 2318, photo GH). Annuals or rarely apparently perennials, puber- ulent to densely villosulous or pilosulous through- out (above hypocotyl), from a solitary short to well- developed taproot 1-2.5 mm thick with smooth gray or brown epidermis; stems to ca. 15 cm long, usu- ally 1-4 times branched but occasionally more or not at all, ascending to weak; cotyledons sessile, narrowly oblanceolate to ligulate, 8-16 mm long, 1.5-3 mm wide, at apex truncate to rounded, at base attenuate, fleshy drying coriaceous, glabrous, with stipules interpetiolar, fused to bases of coty- ledons, membranaceous, interpetiolar portion trun- cate to broadly triangular, 1-2 mm long, entire to shortly bifid or erose. Leaves simple or rarely 2-lobed, petioles (0.5)2-7 mm long; blades elliptic to usually narrowly so or oblanceolate, (4)12—30 [40] mm long, (2)3-8 mm wide, acute at apex, acute to usually attenuate at base, membranaceous to subcoriaceous; stipules at more basal nodes deeply bilobed, lobes triangular to narrowly so, 1—4 mm long, acute to acuminate, erose with appendages to ca. 1 mm long, at more distal nodes similar or more deeply lobed to completely divided. /nflorescences 1-2.5 cm long and 1-6 cm wide excluding corollas; bracts with lobes elliptic to narrowly so, acute at apex and base, central lobe 8-34 mm long, 3-4 mm wide, lateral lobes 2(4), 5-15 mm long, 2-4 mm wide; flowers with hypanthium 0.5-1.5 mm long; calyx limb pilosulous, lobes 4—5, erose with appendages 1—4 mm long, unequal, (0)1-2 lobes with stipes 2-7 mm long, appendages elliptic to suborbicular, 2-10 mm long, 5-10 mm wide, ob- tusely angled to rounded or subtruncate and usually mucronate at apex with mucro 0.5-1 mm long, rounded to truncate or somewhat cordate at base, petaloid, bright to deep yellow, remaining lobes su- bulate, 2.5-8 mm long; corolla externally moder- ately to densely pilosulous, tube 9-12 mm long, 0.1-0.3 mm diam. near middle, enlarged throat of long-styled form 1-1.5 mm long and diam., lobes 5, 1-2.5 mm long, 0.5-1 mm wide near base; an- thers 1-1.3 mm long, filaments 1-1.5 mm long in short-styled form; stigmas 1-2 mm long, in long- styled form exserted by 1-1.5 mm. Capsules 2-3 mm long, 2-3.5 mm wide, persistent calyx lobe ap- pendages to 13 mm long, 15 mm wide; seeds ca. 1.5 mm long, ca. 1 mm wide. Illustrations: Ricardi and Quezada (1963: Fig. 5, as “C. tripartita,” and ig. 9 but in Fig. 9A the stipules at more distal nodes are not entirely accurate and Fig. 9B shows a stipule of a relatively basal node). Distribution and habitat. Northern Chile (23° 30’-30°10'S) at 10-1900 m, in arid regions. Collect- ed in flower and fruit concurrently, September through December. This frequently collected species is distin- guished by its annual habit, usually simple cauline leaves, calyx with usually 1-2 lobes prolonged into yellow petaloid appendages, and relatively small flowers with slender tubes. Ricardi and Quezada (1963) separated Cruck- shanksia tripartita from C. pumila, commenting that Johnston (1929) had combined them but the taxa seemed distinct. Philippi in his original description separated C. tripartita based on its larger, more branched inflorescence, but this feature intergrades continuously with plants he placed in C. pumila and appears to represent only an advanced devel- opmental stage. Ricardi and Quezada commented in their discussion under C. tripartita that these species are separated by their “habit, pubescence, stipules, and floral characteristics” (my translation from the Spanish), but in their key to species they separated these only by the stipules “non-interpe- tiolar, setaceous, entire, to 4 mm long” in C. tri- partita in contrast to “interpetiolar, triangular-long acuminate, laciniate-ciliate, to 2 mm long” in C. pumila. As discussed in the morphology section (above), these stipule forms represent developmen- tal stages that may be found on the same stem, and both conditions are found on most specimens cited for each of these species by Ricardi and Quezada. This stipule distinction thus serves primarily to separate plants that flower precociously from those that show more vegetative development. When the distinctions in this key are applied, the individual plants of one population are sorted into two species (e.g., Taylor 10764, CONC, MO; 10793, CONC, MO). No pattern is evident in any other features, and C. tripartita is here combined with C. pumila. Ricardi and Quezada (1963) cited “Huanta, “in 476 Annals of the Missouri Botanical Garden collibus arenosis, 8bri 1836", leg. C. Gay (SGO),” as the “isotype” of Cruckshanksia pumila. These data correspond to two specimens, CONC-28802 and SGO-56885, which additionally both bear the collection number 1929. However, this information does not agree with the locality cited in the original species description, “en los arenales porfíricos de la vecindad de Arqueros,” so thes parently are not types. Ricardi and Quezada (1963) listed several Geisse collections made in 1885 and 1886 and deposited at SGO as “types” and one at GH as an “isotype” of C. geisseana. Philippi gave the type locality for this species as “prope Ban- durrias haud procul a Chañarcillo detexit orn. Gu- EE e specimens ap- lielmus Geisse," and judging from annotations at SGO most likely based his description on most or all of the set of specimens sent to him by William Geisse with collection dates of 1886 or earlier. These specimens are here considered syntypes. From them, the one best represented in several her- baria is chosen here as the lectotype. These collec- tions have been variously attributed to *W. Geisse," "Guill. Geisse," “Guillermo Geisse," and “G. Geisse." Many of the "perennial" specimens cited by Ri- cardi and Quezada (1963) represent annual plants with well-developed taproots, or have here been re- ferred to Cruckshanksia hymenodon. The few ap- parently truly perennial specimens of C. pumila (Jaffuel 2628, CONC, GH; Jiles 5370, CONC; Ri- cardi & Parra 77, CONC) are additionally char- acterized by relatively small leaves and petaloid calyx appendages, although these all fall within the range of sizes found among the annual plants. These plants are all from the Taltal area, and may represent a distinct population. Only the apparently perennial habit distinguishes them, however, and they are here provisionally included in C. pumila. UE nu specimens examined. CHILE. II RE- GIO - Pro ntofagasta: Taltal, near Paposo, Cerro roe 5560 (CONC). III REGION. E : : Carret tera Panamericana, entre Las Bombas y . Ricardi et al 1430 (CONC). Prov. . Co nite Vallenar y a a 39 km de Copiapó, Ricardi et We 663 (CONC). IV REGION. Prov. Elqui: Rivadavia, Montero 11687 (CONC), Ricardi 2178 Li imari: Corral Quemado, Jiles 3484 6. Cruckshanksia montiana Clos, in Gay, Fl. Chil. 3: 195. 1848. TYPE: Chile. IV Region. Prov. Coquimbo: “dunas cerca de La Serena, 7bri 1836," C. Gay 1931 (holotype, P not seen; isotypes, CONC-43456, GH, K, NY, SGO- 56880, and possibly F-635148 Cruc prs capitata Philippi, Anal. Univ. Chile 41: 731. 1872. TYPE: Chile. e Region. Prov. Huasco: ta Es Bajo, T. King s.n. (holotype, SGO-56879; isotype, SCO- 43367/herb. E Phil. 944, photo GH). Cruckshanksia densifolia Philippi, Anal. Univ. Chile 41: 1872. Cruckshanksia capitata var. densifolia palo Reiche, Anal. Univ. Chile 106: 973. 1900 TYPE: Chile. III Region. Prov. Huasco: Carrizal Bajo, 1871, T. King s.n. (holotype, SGO-56882, pho- to GH; isotype, E). Perennials, puberulent to densely villosulous throughout, from well-developed taproots 1.5 em thick with red corky bark; stems 5-15, to ca. 20 cm long, weak. Leaves simple with petioles 1-5 mm long or sometimes 3-lobed and sessile; blades nar- rowly elliptic, 8-23 mm long, 0.8-4(6) mm wide, acute and sometimes mucronate at apex with mucro to ca. 0.5 mm long, attenuate at base, membrana- ceous to subcoriaceous; stipules at more basal nodes triangular, 1-3 mm long, acute to shortly bilobed, entire to slightly erose, at more distal nodes deeply bilobed, lobes triangular to narrowly so, 1.5-7 mm ong, acute, erose with appendages to ca. 1 mm long. Inflorescences 1-3.5 cm long and 2-4.5 cm wide excluding corollas; bracts with lobes narrowly elliptic to narrowly oblanceolate, acute at base and apex, central lobe 10-17 mm long, 1.5—4 mm wide, lateral lobes 2-4, 8-15 mm long, 1-2.5 mm wide; flowers with hypanthium 1.5-2 mm long; calyx limb pilosulous, lobes (4)5(6), erose with linear append- ages 0.5—4 mm long, unequal, (1)2-3(4) lobes with stipes 6-10 mm long, appendages orbicular to ob- long or somewhat reniform, 0 mm long, 8-15 mm wide, rounded to usually truncate and mucro- nate at apex with mucro 0.5-1 mm long, rounded to truncate at base, petaloid, bright to deep yellow, remaining lobes subulate and 4-10 mm long or oc- casionally similar to bracts, 3-8 mm long, 0.5-1.5 mm wide; corolla externally depu to densely pilosulous, tube 10-13 mm long, 0.3-0.6 mm diam. near middle, enlarged throat of long- styled form 1— 1.5[3] mm long and diam., lobes 5, 2—4.5 mm long, 1-2 mm wide at base; anthers 1-1.5[2] mm long, filaments ca. 1-2 mm long in short-styled form; stigmas ca. 1 mm long, in long-styled form exserted by 1-2 mm. Capsules 2.5-3[4] mm long, ca. 3[4] mm wide, persistent calyx lobe appendages to 12 mm long, 18 mm wide; seeds 1.5-2 mm long, ca. 1 mm wide. Illustrations: Ricardi and Quezada (1963: fig. 6. fig. 7, as "C. capitata") Distribution and habitat. Northern Chile, in arid regions at 10-500 m. Collected in flower and fruit usually concurrently, September through No- vember. This species is distinguished by its usually sim- ple narrow cauline leaves with the secondary ve- Volume 83, Number 4 1996 Taylor 477 Oreopolus and Cruckshanksia nation not evident and calyx with usually 2-3 lobes prolonged into yellow petaloid appendages. It is similar to Cruckshanksia verticillata, as discussed under that species. Ricardi and Quezada (1963) separated Cruck- shanksia capitata based on the form of the stipules and petaloid calyx lobe appendages. However, as discussed in the morphology section (above), both of these features vary with developmental stage, usually along a single stem. The stipule form by which they distinguished “C. capitata” is that char- acteristic of more basal nodes and the form of the petaloid calyx lobe appendages they used is that found on the flowers, while the forms by which they distinguished “C. montiana” are those of the most distal nodes and the fruiting stage, respectively. Most of the specimens they cite show both of the conditions they used to separate these species, and consequently C. capitata is here combined with C. montiana. Cruckshanksia montiana was named in honor of Sefior Manuel Montt, but originally published with the spelling *montiana." This epithet has been im- properly corrected by various authors to “montti- ana," and has also been misspelled as “montteana” and “montana.” As discussed in detail under the treatment of Cruckshanksia hymenodon, Hooker and Arnott originally based their description of that species on three specimens, two of which represent C. montiana. Representative specimens examined. CHILE. III RE- GION. Prov. Copiapó: Quebrada del León, Billiet & Ja- din 5356 (BR, MO), Werdermann 434 (CONC, E, GH, K, M, MO). Prov. Huasco: Huasco, Werdermann i (CONC, E, GH, K, MO). IV REGION. Prov. Elqui: r from La Serena to Punta Teatinos, West 3918 (CONC, GH, 0). 7. Cruckshanksia verticillata Phil., Anal. Univ. Chile 85: 737. 1894. TYPE: Chile. II Region. Prov. Antofagasta: Bandurrias, 1886, W. Geisse s.n. (holotype, SGO-56873/herb. F. Phil. 1936; isotype, SGO-56878, photo GH). Cruckshanksia paradoxa Phil., Anal. Univ. Chile 85: 738. Cruckshanksia capitata 2 auc (Phil.) Reiche, Anal. Univ. Chile 106 TYPE: Chile. II Region. Prov. "ead Sonde Nov. 887, W. Geisse s.n. (lectotype, designated here, SGO-72374). Perennials, puberulent to villosulous or pilo- sulous throughout, from well-developed taproots to 6 mm thick with somewhat corky reddened bark; stems 2-10, to 10 cm long, weak. Leaves simple to usually 2-3-lobed, subsessile to sessile; blades narrowly to very narrowly oblanceolate, 7-18 mm long, 1-2 mm wide, acute and frequently falcate at apex, attenuate at base, membranaceous to sub- coriaceous; stipules usually none or with 1 or 2 lobes, these separate, triangular, 1-1.5 mm long, 0.8-1 mm wide, acute to rounded, erose with ap- pendages to ca. 1 mm long. Inflorescences 1-2 cm long and 1-3 cm wide including corollas; bracts similar to cauline leaves of most distal nodes; flowers with hypanthium 0.5-1 mm long; calyx limb villous to pilosulous, lobes 2, erose with lin- ear appendages 0.5-3 mm long, equal, with stipes mm long, appendages elliptic-oblong to sub- orbicular, 3-6 mm long, 2-7 mm wide, rounded and frequently mucronate at apex with mucro ca. 0.5 mm long, acute to cuneate at base, petaloid, yellow; corolla externally moderately to densely pilosulous, tube 12-14 mm long, 0.3-0.5 mm diam. near middle, enlarged throat of long-styled form 2-3 mm long and 1.5-2 mm diam., lobes 5, 3-3.5 mm long, 1-1.5 mm wide near base; anthers 1-1.3[2] mm long, filaments ca. 1 mm long in short-styled form; stigmas ca. 0.5 mm long, in long-styled form exserted by 1-2 mm. Capsules ca. 3 mm long and wide, persistent calyx lobe ap- pendages enlarging slightly at most; seeds ca. 1.5 mm long, ca. 1 mm wide. Illustration: Ricardi and Quezada (1963: fig. 8). Distribution and habitat. Northern Chile, in open sand and rubble of arid regions near Taltal at 100-1000 m. Collected in flower and fruit October through November. This infrequently collected species is distin- guished by its cauline leaves with usually two or three narrow lobes, and by the calyx limb reduced to two lobes, both with petaloid appendages. The cauline leaves are generally indistinguishable from the inflorescence bracts. It is similar and probably closely related to Cruckshanksia montiana, which can be distinguished by its usually simple cauline leaves and petaloid calyx lobe appendages 7-15 mm wide. Most features of C. verticillata are con- sistently smaller than those of C. montiana except the corollas, which are usually slightly longer, al- though these measurements overlap. These two spe- cies are maintained here, although considering the variability found in other species of Cruckshanksia, C. verticillata may represent only a well-marked variant of C. montiana The designation of type specimens for Cruck- shanksia verticillata here follows annotations by Ri- cardi (in herb.) and citations in Ricardi and Que- zada (1963). The lectotype chosen here for C. paradoxa is the only one of the syntypes that does not have some confusion attached to the label data. 478 Annals of the Missouri Botanical Garden Additional specimens E CHILE. III REGION. Prov. Copiapó: Bandurrias, 1885, "ers s.n. [SGO- 43368/herb. F. Phil. 1918. pholo G H], 5-1886, W. Geisse s.n. (GH, SGO-56875); Atacama, com. ic 2/1888 [K not seen photo (neg. #SGO-2478) SGO]. IV REGION. Prov. ajonales, Nov. 1886, W. Geisse s.n. (SGO— 661) Literature Cited Beusekom, C. F. van. 1967. A revision of the Malesian and Ceylonese ssa of the genus Gaertnera Lamk. m Blumea 15: 359-391 Clos, D 848. Rubiaceas. Pp. 176- 212 in: C. Gay, His- toria se y n de Chile. a Tomo 9. A revision of the genus Kajewskiella (Rabiacsaii, [p 25: 283-204. h M. 1929. Papers on the flora of northern 1. The coastal flora of the departments of Chañ- pel ond Taltal. Contr. Gray Herb. 85: 1-138. Muñoz Pizarro, C. 1966. Sinópsis de la flora chilena. Ed- iciones de la Universidad de Chile, Santiago Puff, C. Observations on Carphalea Juss. (Rubi- aceae, Hedyotideae), with particula Madagascan oon nd its taxonomic position. Bull. Jard. Bot. Belg. 58: 3. Ricardi, M. Té chain del género Oreopolus Schlecht. Gayana, Bot. 6: 3-16. 1963b. Una nueva especie de Cruckshanksia (Rubiaceae). Gayana, Bot. 7: 3-7. ——. 1968. Nota sobre Oreopolus edd: Revis- ta Fac. Ci. Agrar. Univ. Nac. Cuyo 13(1-2): 3-7. 19 Addenda al nas Oreopolus (Rubi- scene), Bol. Sie: Biol. Concepción 46: 217-221. & M. Quezada. 1963. e = Cruckshanksia (Rubiaceae). ao Bot. 9: Robbrecht, E. Mis al o Rubiaceae. Opera Bot. Belg. 1 Schumann, K. P NN Naturl. Pflanzenfam. Beibl. 4(4): 30, fig. 8Q, R. Squeo, F. A., R. Osorio & G. Arancio. 1994. Flora de Los Andes de Coquimbo: Cordillera de Dofia Ana. Ed- iciones Universidad de La Serena, La Serena, Chile. Appendix I. Index to numbered collections examined. Identifications are indicated by numbers in parentheses following the collection number, as below. Asterisks in- dicate a type collection. Oreopolus glacialis (Poepp.) Ricardi C ruckshanksia hymenodon Hook. & Arn. | — HOMO i = ~. ~ ~ ~ n JON Sw oll an . pumila Clos C. verticillata Phil. Arroyo, M. K. 81002 (4), 81042 (4), 81096 (4), 81112B (6), 81227 (2), 81619 (4), 81634 (4), 81681 (2), 83370 (2), 83403 (2), 83464 (4), 84555 (4), 83624 (2), 841003 (1), 850831 (1). Arroyo, S. C. 150 (1), 225 (1), 321 (1), 2243 (1), 2376 (1), 2513 (1), 3339 (1). Bayern 352 (7). Betfruend 12438 (1). Billiet 5356 (5), 5570 (2). Boelcke 1694 (1), 3382 (1), 3410 (1), 3411 (1 3546 (1), 5936 (1), 7228 (1), 10049 (1), 11114 (1), 11300 (1), 11669 (1), 13797 (1), 15106 (1), 15837 (1). Bridges =Y 1207 (1), 1283 (1 ), 1299 (5). 1300 (5), 1301 (7), 1302 2) Cabrera 12624 (7). Castellanos 7908 (1). Cei 24176 (1). Comber 263 (1), 841 (1). Cordini 210 (1). Correa 2645 1), 3059 (1), 3443 (1), 3566 (1), 3 (1), 9091 (1), 9241 (1), 9992 (1), 10056 (I), 10390 (1), 10442 (1). Covetto 7140 (1). Crespo 1645 ( ; 2104 ^ 2287 (1). Cuming 861 (5), 875 (5), 1301 (7), 1302 ee 2624 (1). Diem 3224 (1), 177 (1), 364 (1). Elliot 22 (7). Eskuche Flores 25 (7). Furlong is (1). Garaventa 4231 (7), 4400 (7), 4732 (2). Gardner 4535 (1). E i 5), 412 (2), 413 (6), 865 (5), 1255 dk ling 5756 (1). Gigoux 217 (7), 220 (8). 627 (2). ( 2 D Goodall VT ). 3955 (1), 4283 (1). 4380 (D). 2043 (7), 2109 (7), 2118 (7). 2522 (2), 2540 (7), (1). Greninger 14 (2). man 43 (7). Hosseus 245 (1), 1332 (1). Hunziker 4845 (4), 7086 (1). Ilin 43 (1), 55 (1), 138 (1), 456 (1), (1), 5736 (1), 5762 (1), 18146 (1). Jaffuel 1158 (7), 2628 i. x. 1124 (6), 1199 (6), 1301 (7), — — — (1). Donat 77 (1), 172 G rau 2897 re — 601 (1), 1324 (1), 5524 (1). 5526 1167 (2). T t uc WA NN ^3 B GANN JUN wa qn. IN = e Ce nh - RB N <4 > p nN 25 d ^ Ee 23% dit 2 B2 w =3 al ec ms O (2). 5: 5803 (7), 5812 (2). 6354E (5), 6396 (2), «Johnston 3671 (2), 4784 (7), 4818 (2), , 4974 (2), 5106 (7), 5160 (7), 5301 (7), 5560 (7), 5595 . 5711 (7), 5930 (4), 6065 (4), 6162 (4). Kansel 3747 (7). Kiesling 8051 (2). King 26 (7), 51 (2). Krapov 4128 (1), 4288 (1). Krapovickas 5698 (4). 5726 (2). Kurtz 9570 (2). Koslowsky 115 (1). 5528 (1). Landero 630 (1). Le e 2895* (1). Leuenberger 3403 (1). Lobb 435 (5), 436 ( s 6: : re :orena ] (2), 63 (1), ag (2), 244 ), 248 (7), 328 (7), 495 (2), 537 (3), 582(3), 1030 (1), 1264 O, 1387 (1), 1656 (7), 1678 (2), io (2), 1699 (7). 9810 (2). Merxmüller 24996 (1). Meyer 9637 (1). Mo- . Montero 11687 (7). Moore 2264 (1). Mo- rong 54 (2). 55 (7), 107 (8), 1106 (2), 1109 (2), 1303 (5). Morrison 16963 (2), 16998 (2), 16999 (6), 17025 (2), 17443 (2). Muniez 159 (1). Muñoz 2726 (2), 2837 (7), 2955 (7). O’Donell 3748 (1). Pereyra 2166 (1), 5306 (1). Pérez 23404 (2). Peterson 270.33 (1). Pirion 79 (1), 126 (1). Pisano 3660 (1), 4174 ah 4730 (1). 4781 (1), 5604 (1). Poeppig 59* (1) Rechinger 63338 (7), 63339 (2), 63476 (7), 63494 (7). Ricardi 12 (2), 77 (7). 378 (1), 484 (1), 515 (1), 533 (2), ) c as NN eA = 3S T3 — -l 552 (2), 663 (7), 980 (1), 1102 (2), 1131 (2), 1136 (5), 1250 (7). 1430 (7). 1462 E 1489 i 1563 (2). 1569 ( (6 (7), 2232 (2), 2275 (7). : (1 (2), 2586 (7), 2627 (1). 2034 (2), 3779 (3), 3970 (5), 4359 ( (7), 5663 (1). Rodraguéz 10 (2), 77 (5). Argis 40 a 1616 (7). Roig 11973 (2), 13019 (2), 13032 (2), 13043 1). Rosas 1012 (7), 1169 (2), 1248 1844 (1). Rose 19338 (7). Rossow 1310 (1), . 1463 (1), 1872 (1), 1617 (1), 1651 (1), 2662 (1), 4565 (1). Ruiz 1. 3131 (1), deg 11689 (1), 11727 (1). 15678 (1), 15736 (1). 16833 (1). 21441 (1). Sánchez 293 (1), 442 (1). 552 (1). Santesson 1272 (1). Volume 83, Number 4 1996 Taylor 479 Oreopolus and Cruckshanksia Schlegel 937 (6), 2415 (6), 5716 (5), 5886 (2). Seibert 3 69 (1). Silvestri 5744 2459 58 | =" — un o 3 ya] o ^ -Ó $ wo © N — Un c 3 3 1), 3067 (1). (1), 207 (1), 339 (1). Squeo (4), 88097 (4). Stuchart 18146 (1). Stuessy 11092 Taylor 10679 (7), 10764 (7), 10781 (2). es (2), 10793 (7). TBPA 455 (1), 539 (1), 740 (1), 1669 (1), 2174 (1), 2330 (1), 2717 (1), 2854 (1), 3619 (1), 3758 (1). Teil- lier 557 (7), 657 (7), 731 (7), 968-969 [sic] (2). Vervoorst 5626 (1). Villagrán 4668 (3). Volckmann 64 (2). Wagenknecht 18120 (2), 18484 (7), 18505 (2), 18553 7). Werdermann 112 (7), 133 (5), 158 (2), 211 (2), 222 (4), 434 (5), 810 (7), 936 (4). West 3918 (5), 4778 (1). Witte 35 (1). Worth 15835 (7), 16196 (2), 16233 (5), 16270 (2), 16388 (6), 16398 (2), 16565 (6), 16568 (2), 6682 (2). — Zöllner 787 (2), 786 (4), 5324 (4), 6015 (7), 6168 (7), 6196 (7), 6503 (7), 8455 (1), 8720 (7), 9312 (7), 9696 D, ius (2), 10392 (2), 10644 (2), 10928 (5), 11376 (7), 14049 (2), 14819 (2), 18646 (7), 18693 (7), 18906 (2), 19124 (7), 19136 (7), 19138 (2). mm A SURVEY OF THE E. Nic Lughadha? and C. Proença’ REPRODUCTIVE BIOLOGY OF THE MYRTOIDEAE (MYRTACEAE)! ABSTRACT The My aria usually present small, epigynous, 4—5-merous, polystemonous flowers that last one day. Bee- -polli- nation in dics ollen is the sole reward is the dominant pollination system. Nectar has been best documented in Syzygium but bbs also occurs in other bee-pollinated genera. The most common bee visitors are Apidae: Meli- poninae and Bombinae. Bird- and peste pollination occur in Old World Syzygium with nectar as the primary reward. Bird-pollination with petals as the reward occurs in New World Acca and Myrrhinium. General floral morphology is very uniform, while inflorescence hoe s and flowering strategies are very diverse. Stigmas are dry and ovules are anatropous, hemicampylotropous or anacampylotropous and have an outer 2-6-layered integument a an inner 2-lay- ered integument or a single integument. The ovary usually contains more ovules than will form seeds. Flowering strategies vary from mass-flowering types, in which the flowering episode typically takes only a few days, to steady- state types of up to 90 days duain Flowering at dry/wet season transition is common in seasonal climates, and fire- uced flowering is found oce MA a Outbreeding i is probably widespread, although both self-compatible and self- incompatible species exist. The self-incompatible species have self pollen tubes penetrating the micropyles, so preferential outcrossing may be maintained by a late-acting mechanism. Cryptic dioecy, in which female flowers have mimic” sterile anthers occurs in several genera. Apomixis occurs in Syzygium and this has been reported to be linked to the polyembryony found in this genus. Myrtoid fruits are fleshy berries or drupes, dispersed by birds, bats, loys au small mammals. Fruit size, color, texture and number of seeds are all very variable. Seed coats may be absent to bon but have a smooth s E een ndosperm i is mainly digested by the developing embryo. Early embryology is nett uniform but final embryo morphology varies widely across the genera. Germination times vary from 10 days to over 2 years and seed viability seca from 15 days to 1 yea The reproductive biology of the Australian Myrta- — sense to embrace virtually all the fleshy fruited Myr- ceae has recently been the subject of a thorough re- — taceae. Thus we include Johnson and Briggs’s (1984) view by Beardsell et al. (1993). The alliances tradi- monophyletic Myrtoideae sensu stricto, and their Ac- tionally assigned to the subfamily Leptospermoideae mena alliance (including Syzygium), although phylo- are the major focus of their account, reflecting not genetic analysis has suggested that this latter group only the predominance of these groups in the Austra- is more closely related to the leptospermoid Eucalyp- lian flora but also the widely acknowledged dearth of tus alliance than to the Myrtoideae s. str. (Johnson & information on the reproductive biology of the Myr- — Briggs, 1984). Conversely, we use the term Leptosper- toideae in general (e.g., van Wyk € Lowrey, 1988: moideae in its broadest sense to include all Myrtaceae Proenga & Gibbs, 1994), with capsular fruit. Where names have been changed The present survey was compiled to complement — to fit modern taxonomic concepts, the name that ap- that of Beardsell et al. (1993) by bringing together the peared in the original publication is cited in paren- scattered and often fragmentary data on the Myrtoi- — theses, e.g., Syzygium paniculatum (as Eugenia pan- deae. We use the term Myrtoideae in the traditional — iculata). ' This survey is based largely on a portion of the doc we thesis of the first author, funded by the Royal Botanic Gardens Kew, UK, with substantial contributions from the ee thesis of the second author, funded by the Conselho Nacional de Desenvolvimento Científico e Tec -nológic o, Bre azil oth theses were prepared under the supervision of Peter Gibbs of the University of St. Andrews, Se n and re authors gratefully acknowledge his guidance and encouragement in the field and in the laboratory, as well as his critical reading of the present manuscript. Pranom Chantaranothai, Martin Cheek, John Wyndham din Keith Ferguson, Madeline Harley, Ray Harley, Bruce Holst, Jim Jarvie, Leslie Landrum, John Parnell, Alan Paton, Susanne Renner, Paula Rudall, Peter G. Wilson, Elizabeth Woodgyer, and Rea PU OR "e read and made improvements to various versions of this paper, while J. W. Dawson, Keyt Fischer, Maria A. S. Alves, and Anthony Raw provided unpublished data; we thank them all for their patience their unfailing Viae and support and, in particular, to Ann McNeil who facilitated an online search of the BIOSIS bibliographic cal databas ? Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3AE, United Kingdom. * Depto. de Botánica, C. P 4631 Universidade de e ane 970, Brazil. ANN. MISSOURI BOT. GARD. 83: 480-503. 1996. Volume 83, Number 4 1996 Nic Lughadha & Proenga Reproductive Biology of Myrtoideae The pantropical Myrtoideae as a whole has not received any synthetic taxonomic treatment since Niedenzu’s (1893) synoptic account. There is no Old World equivalent to McVaugh’s (1968) invalu- able review of the American Myrtoideae. As a sub- family the Myrtoideae are rather less morphologi- cally diverse than the Leptospermoideae. Ap- proximately 60 genera are recognized for the esti- mated 2375 species of Myrtoideae, while the Lep- tospermoideae with less than half the number of species (1300) includes some 72 genera (species numbers from Schmid, 1980; generic estimates re- flect the present authors’ view of current taxonomic opinions). While this discrepancy may be inter- preted as an artifact of our ignorance of the former group, most myrtologists would agree that the per- ceived homogeneity is real and has contributed to, rather than resulted from, the comparative taxo- nomic neglect which this subfamily has indisputa- bly suffered. McVaugh (1968) described the species of American Myrtaceae as “distressingly alike in aspect and in most individual characters, making identification and classification of both genera and species a correspondingly difficult and tedious mat- This homogeneity is portrayed in more concrete terms in Schmid’s (1980) table of the “distinguish- ing organographic characters of the subfamilies of the Myrtaceae,” the majority of which are based on reproductive organs. For each character a statement of the prevailing state within each subfamily is fol- lowed by a summary of the exceptions. Schmid ex- pressed the estimated percentage of exceptions us- ing precise and clearly defined qualifying terms: very rarely? ere < 2% of species), “rarely” (2-4%), * onally” (5-10%), “occasion- ally” (5-30%) icd “often” (31-60%). Hence a rea- sonably accurate impression of the variability of particular characters within the subfamilies can be gained. The qualifier “very rarely” is applied to the majority of the exceptions listed for the Myrtoideae. In the Leptospermoideae, however, exceptions are more or less evenly distributed across the range from “rarely” to “often” with “very rarely” being applied in only two instances. Thus the Myrtoideae emerges as a much more homogeneous group than the Leptospermoideae. INFLORESCENCE STRUCTURE Inflorescences may be apical, subapical, axillary, supra-axillary, ramiflorous or cauliflorous in posi- tion. The authoritative analysis of inflorescence structure in the Myrtaceae by Briggs and Johnson (1979) documented an equally impressive range of structural form within the Myrtoideae. They range from many-flowered panicles and thyrsoids through dichasia, botryoids, metabotryoids, triads, and me- taxytriads to monads and metaxymonads. In their brief description of adaptive syndromes, these au- thors drew attention to correlations between habitat, niche, and inflorescence arrangement. In particular, large panicles or aggregations of smaller inflores- cences are reported by them as typical of forest and woodland species of Myrtaceae (myrtoid examples include species of Myrcia and its allies, Blepharo- calyx (as Temu), Pimenta, Xanthomyrtus, species of Syzygium and its allies). However, in the under- growth or understory of non-sclerophyllous com- munities, they noted that flowers are often in mo- nads and/or triads that are not massed (myrtoid examples include Myrceugenia, Psidium spp., Ugni, Myrteola, Myrtus sensu stricto, Rhodamnia, Lophomyrtus, Neomyrtus, Campomanesia, Luma, and Eugenia spp.) (Briggs & Johnson, 1979). Ram- iflorous or cauliflorous inflorescences have been linked to bat pollination (Crome & Irvine, 1986), and to bat or marsupial dispersal (see Fruit Char- acteristics and Dispersal Agents). FLOWER STRUCTURE Flowers of the Myrtoideae are hermaphrodite (or very rarely unisexual), epigynous (or very rarely semi-epigynous), with (3—)4—5(—9) sepals and (0-)4— 5(-6, 12) petals per flower. The perianth parts are free in bud or occasionally form a calyptra or oper- culum or rupture irregularly at anthesis. In several genera the individual calyx lobes or the hypanthial remnants are deciduous after anthesis. Petals are white, cream, pale pink, or (rarely) deep pink or red. Stamens are usually numerous but, when few, the androecium may be diplostemonous, haploste- monous, or obhaplostemonous (i.e., a univerticillate androecium with only antipetalous stamens; Schmid, 1980). Stamens are generally free and at- tached directly to the hypanthium, but fascicled stamens occur in a few species of Syzygium some- times segregated as Pareugenia (Schmid, 1972a). Most Myrtoideae have a secretory cavity at the apex of the anther connective but this feature is lacking in a handful of species in various genera (Schmid, 1972a). Carpels, when present, are 2 or occasion- ally 3-18 per flower, and the number of ovules per ovary is usually many, occasionally few, and very rarely one. General flower structure varies little among species as compared to other larger families, but flower size may vary by an order of magnitude. 482 Annals of the Missouri Botanical Garden ANTHER DEVELOPMENT AND MICROSPOROGENESIS The anther is tetrasporangiate and dehiscence is by longitudinal slits. In Gomidesia the thecal mar- gins are often strongly inrolled, so that the interior of the sacs and the pollen are only exposed over a small area at the apex (and often another small area at the base) of the anther. Gomidesia anthers have often been termed poricidal; however, they are more accurately described as longitudinally dehiscent and functionally poricidal. Terminally dehiscent anthers have also been reported in Acmena. The anther sacs, which are almost globose and some- what divaricate, open by a terminal slit and are one of the diagnostic features of this genus (Merrill & Perry, 1938). Pollen grains of all Myrtoideae (and all Myrtaceae) studied to date are two-celled at an- se thesis (Schmid, 1984[1985]) POLLEN GRAINS Patel et al. (1984) reviewed the palynological lit- erature on the Myrtaceae. Pike's (1956) survey of 300 species in 71 genera remains the most com- prehensive work on the subject to date. Although her study focused on the Leptospermoideae of the southwest Pacific area she also investigated repre- sentatives of the Myrtoideae, including some South American genera, and was forced to conclude "There appears to be no particular feature that sep- arates the pollen of the Myrtoideae from that of the Leptospermoideae" (Pike, 1956: 46) Barth and Barbosa (1972) confirmed the steno- palynous nature of the Myrtoideae on the basis of their survey of 140 species in 19 genera. They de- scribed myrtoid pollen as follows (translated ‘from the Portuguese original with modern preferred terms (Punt et al., 1994) inserted [in brackets] after outmoded terms): “grains small to medium, oblate to peroblate, tricolporate with lalongate ora [en- doapertures], goniotreme [angulaperturate] with tri- angular amb [polar view], surface of apocolpia and mesocolpia granular, smoother near the apertures. On average sexine is twice as thick as nexine 2 [endexine], which is constant in thickness. The sex- ine is tectate, the bacula [columellae] being re- sponsible for the granular surface of the grains; un- dulations on the surface originate from the tectum. Nexine 1 [foot layer] is limited in occurrence; ir general it is absent or reduced to fragments in ' (Barth & Barbosa, - grains with a thicker exine 1972: 468—469). All these features common to the Myrtoideae are also characteristic of the Myrtaceae as a whole. Both Pike (1956) and Patel et al. (1984) recognized three pollen types in the Myrtaceae: (1) longicol- pate, (2) syn- or parasyncolpate, and (3) brevi- or brevissimicolpate grains. All members of the Myr- toideae studied were reported to have grains of one or the other of the first two types, grains of type (3) being confined to certain members of the Chame- laucium group. However, Barth and Barbosa (1972) reported grains of all three types in the Myrtoideae and listed some 25 instances where both longicol- pate and syncolpate grains were found in the same species, usually in the same collection, as most species in this study were represented by a single specimen. Similar variation was reported in about one third of the 18 species studied by Lieu and Melhem (1973), while most of the remaining spe- cies sampled had consistently longicolpate grains. Pollen is released as free monads except in Myr- tus communis L. and Psidium cattleianum Sabine (as Psidium littorale Raddi) where there is a mix- ture of free monads and tetrahedral tetrads (Patel et al., 1984). Dimorphic pollen occurs in associa- tion with cryptic dioecy in Decaspermum parviflo- rum (Lam.) A. J. Scott (Kevan & Lack, 1985) and in the South African species of Eugenia (van Wyk & Lowrey, 1988; van Wyk & Dedekind, 1985). Male flowers produce normal pollen grains and fe- male (apparently hermaphrodite) flowers produce morphologically abnormal unviable pollen grains. STIGMA AND STYLE STRUCTURE Beardsell et al. (1993) described stigmas in the Myrtaceae as unspecialized and generally of the “wet” type, citing the Heslop-Harrison and Shivan- na (1977) review of the receptive surfaces of the angiosperm stigma. In fact, Heslop-Harrison and Shivanna (1977) recorded dry stigmas with unicel- lular papillae in five of the seven genera of Myr- taceae studied, including all three of the myrtoid genera in the survey (Acca (as Feijoa), Eugenia, and Syzygium). Dry stigmas were also reported for the six genera studied by Proenga and Gibbs (1994) (Blepharocalyx, Campomanesia, Eugenia, Myrcia, Psidium, and Siphoneugena) and in Gomidesia (Nic Lughadha, pers. obs.) and thus appear to represent the norm in the Myrtoideae. Stigmas are undivided or very rarely divided. Bi- fid or occasionally trifid stigmas are reported in the African species Eugenia ancorifera Amshoff (Am- shoff, 1974), E. aschersoniana F. Hoffm., and E. mossambicensis Engl. (Amshoff, 1958). Schmid (1980) drew attention to the style and stigma o Campomanesia guazumifolia (Cambess.) erg, which has been variätialy described as “shortly bi- fid,” “broadly capitate,” and “peltate.” Most Myr- toideae have punctiform or capitate stigmas. Proen- Volume 83, Number 4 1996 Nic Lughadha & Proenca 483 € Reproductive Biology of Myrtoideae Table 1. Correlations between stigma diameter, ovule number, and pollen tubes.* Stigma Max. diam- number number pecies eter of pollen (number of flowers per obs.) (mm) ovules tubes Blepharocalyx salicifolius (8 fls) 0.1 12 +35 'ampomanesia pubescens (6 fls) 0.74 78 +20 ampomanesia velutina (4 fis) 0.57 29 +50 Eugenia dysenterica (3 fls) 0.1 8 +15 j linearifolia (3 fls) 0.1 4 17 Myrcia bain. p fls) 0.1 6 10 Psidium firmum (3 0.65 298 +120 Siphoneugena densiflora (3 fis) 0.1 13 +15 * Adapted from Proenga (1991). ca (1991) reported both types in a study of eight species and went on to demonstrate a positive cor- relation between stigma diameter and ovule number (Table 1). This was explained in terms of the need for the stigma of a multi-ovulate flower to support more germinating pollen grains. The dioecious Pi- menta dioica (L.) Merr. and the putatively dioecious P. guatemalensis (Lundell) Lundell are exceptions, having very low ovule numbers (1—4 per ovary) and large peltate stigmas. Proenga (1991) argued that the females of these dioecious species no longer need to minimize stigma size in order to avoid de- position of self-pollen and consequent inbreeding depression (in the case of a self-compatible spe- cies) or occupation of all the ovules by self-pollen tubes and subsequent abortion of the pistil (in the case of a self-incompatible species) and suggested that expanded stigmas optimize pollen collection from the pollinator’s body, the only pollen source for a totally female plant. Carpolepis elegans (Mon- trouzier) J. Wyndham Dawson, the only known di- oecious Leptospermoideae, also has a large peltate stigma in contrast to the punctiform stigmas of its hermaphrodite congeners (Dawson, 1992). OvARY STRUCTURE The ovary comprises 2(3-18) fused carpels. The locules are generally multi-ovular, but the number of ovules is often reduced and very rarely each loc- ule may contain only a single ovule. Axile placen- tation is the norm but parietal, basal, and apical placentation have also been reported (Schmid, 1980). Where placentation is axile a compitum (a passageway connecting the loculi) is often present. Carr and Carr (1961) commented that, because of their relationship to the stylar canal(s), such com- pita may increase the chance of fertilization where pollination is limited. Pollen tubes were observed ET through compita in Siphoneugena densiflo- O. Berg and Myrcia linearifolia Cambess. Plein, 1991). OVULE STRUCTURE AND MEGAGAMETOPHYTE DEVELOPMENT The extremely uniform embryology of the Myr- toideae was noted by Mauritzon (1939). Recent re- views by Tobe and Raven (1983) and Beardsell et al. (1993) have shown that subsequent studies have added few genera to the list studied by Mauritzon and have done little to dispel the original impres- sion of uniformity within the Myrtoideae. Ovules of the Myrtoideae (as of the Myrtaceae as a whole) are generally described as anatropous, crassinucellate, and bitegmic. Mauritzon (1939) commented “I hardly exaggerate when I say that there is no other great family in the plant kingdom in which the ovules of the species so regularly have two invariably two-layered integuments.” He then proceeded to describe the fusion of these two-lay- ered integuments to form a single integument of four cell layers in material which he referred to as Eugenia paniculata Banks ex Gaertn. (almost cer- tainly Syzygium paniculatum Gaertn.). Unitegmic ovules have also been reported in Syzygium cumini (L.) Skeels (Narayanaswami & Roy, 1960a; Roy & Sahai, 1962 (as S. caryophyllifolium (Lam.) DC.)), Syzygium fruticosum DC. (Roy, 1961 (as Eugenia fruticosa L.)), Syzygium jambos (L.) Alston (van der Pijl, 1934 (as Eugenia jambos L.)), Syzygium ma- laccense (L.) Merr. & L. M. Perry (van der Pijl, 1934; Roy, 1960 (as Eugenia malaccensis L.)), and S. myrtifolium (Roxb.) DC. (Roy, 1962 (as Eugenia myrtifolia Roxb.)). Van Wyk and Botha (1984) com- mented that those Eugenia species for which only a single integument has been described in the lit- erature can probably all be referred to Syzygium. Tobe and Raven (1983) discussed the possible or- igin and significance of unitegmy in Syzygium. Few would dispute their conclusion that it represents a derived feature within the basically bitegmic Myr- tales. However, their proposition that unitegmy “probably will be found in other genera when the family is better known embryologically” (Tobe & Raven, 1983: 86) has yet to be substantiated. Mauritzon’s generalization about myrtaceous in- teguments seems to hold true insofar as it concerns the inner integument: where this is distinct from the outer integument it is nearly always two cell layers thick (locally three cell layers thick in some South African Eugenia species (van Wyk & Botha, 1984)). However, the outer integument in the Myr- Annals of the Missouri Botanical Garden toideae has proved more variable than Mauritzon’s survey led him to suspect. Petit (1908, quoted in Mattos, 1989) reported 4 cell layers in the outer integument of Luma apiculata (DC.) Burret (as Myrceugenia apiculata (DC.) Nied.) and comment- ed that the outer integument may comprise even more layers in some species of Eugenia judging by the considerable thickness of the integument of Myrcianthes pungens (O. Berg) D. Legrand and Eu- genia uniflora L. (cited as Eugenia pungens O. Berg, and Stenocalyx michelii O. Berg, respective- ly). Van Wyk and Botha (1984) described an outer integument four to six layers thick over the greater part of its free length in their detailed study of South African Eugenia. This latter group of Eugenia also seems to rep- resent an exception to the generalization that ovules of the Myrtoideae (and the Myrtaceae) are anatro- pous. Van Wyk and Botha (1984) reported hemi- campylotropous ovules (occasionally tending to be ana-campylotropous) and suggested that campylo- tropous ovules have frequently been taken as anat- ropous during cursory investigations. In all Myrtoideae studied to date the micropyle is formed from both integuments. As in other Myr- taceae, embryo-sac formation in the Myrtoideae fol- lows the Polygonum-type pattern and the antipodal cells are ephemeral (Tobe & Raven, 3) FLOWERING FLOWERING SEASONALITY Data on the time of initiation of flowering in the Myrtoideae have been compiled and discussed by Proenga (1991) and Proenga and Gibbs (1994). A tendency to flower at the dry/rainy season transition was distinguished, a pattern exhibited by 6 of the 8 species (in 6 different genera) studied by them in the Distrito Federal, central Brazil, and 7 out of 10 species of Myrtoideae investigated in other South American forest communities by other re- searchers (Frankie et al., 1983; Ferreira & Merona, 1987; Morellato et al., 1989). This pattern was clearly demonstrated in a floristic study of the Myr- taceae of the Serra do Cipó, Minas Gerais, Brazil, in which Kawasaki (1984, 1989) reported that the majority of the 50 species (in 11 genera) recorded or the area were found flowering in September and October (spring), after the first rains in the region. Two of these species had a second flowering period in February and March (late summer/early autumn), and a further five species flowered only during this latter period. In Costa Rican dry forest Eugenia salamensis Donn. Sm. was the only mass-flowering species (out of 21 mass flowerers studied) to bloom at the beginning of the wet season (Frankie et al., 1983). The other species (not Myrtaceae) bloomed during the long dry season. Ruiz and Arroyo (1978) reported flowering by Eugenia sp. at the beginning of the wet season in a Venezuelan deciduous forest. Proença and Gibbs (1994) proposed that flowering may be cued by abrupt increases in humidity, which are much more frequent at the dry/rainy sea- son transition than during the remainder of the year. Landrum (1986) commented that most species of Campomanesia flower in the spring, usually in October in southeastern Brazil and adjacent regions. In contrast, most species of the temperate and subtropical South American genus Myrceugen- ia flower during the summer and autumn with a relatively small number of species flowering during late winter and spring (Landrum, 1981). Chilean Myrtaceae flower during the dry spring and summer (Landrum, 1988). Southern African Myrtoideae exhibit a strong conformity to spring dry/wet season flowering, with 12 of the 15 species of Eugenia studied by van Wyk and Lowrey (1988) conforming to this pattern. The remainder flower in early summer (two species) or in winter (one species). Spring flowering is also re- ported as the norm in southern Australian Myrta- 1993; O'Brien & Calder, 1993), though this generalization may be more ap- ceae (Beardsell et al plicable to the capsular dry-fruited Myrtaceae than to the Myrtoideae, as no specific examples from this latter group are cited. In fact, two detailed studies of Australian species of the genus Syzygium illus- trate the variation in the time of flowering within this large genus. In Kuranda, northern Queensland, the rainforest tree Syzygium tierneyanum (F. Muell.) T. G. Hartley & L. M. Perry flowered in January (Hopper. 1980 ) while S. cormiflorum (F. Muell.) B. Hyland, another northern Queensland rainforest tree, had an extended flowering period ranging from at least late July to mid-November (Crome & Irvine, 1986). In Indonesia, however, at least one species of Syzygium does appear to demonstrate the dry/wet season transition spring flowering pattern discussed by Proenga and Gibbs (1994): in Sulawesi trees of Syzygium lineatum (DC.) Merr. & L. M. Perry (as Syzygium syzygioides (Miq.) Merr. & L. M. Perry and subsequently corrected) flowered in February and March (spring) toward the end of what ap- peared to be “a fairly dry season"— precise infor- mation on climate was not available (Lack & Kev- an, 1984). In the same area Decaspermum parviflorum (Lam.) A. J. Scott flowered in January and February, apparently toward the end of a short dry period (Kevan & Lack, 1985). Volume 83, Number 4 1996 Nic Lughadha & Proenca 485 Reproductive Biology of Myrtoideae FLOWERING DURATION, PATTERN, AND SYNCHRONY Meager though the data on flowering seasonality may appear, information on flowering strategies is scarcer still. Proenga and Gibbs (1994) described four flowering strategies observed in the Myrtaceae of the Distrito Federal, central Brazil. Of these, three corresponded to types described by Gentry (1974) viz., “big bang,” “cornucopia,” and “steady state” flowering, and the fourth, for which they coined the term “pulsed bang,” is a modification of the big bang strategy from which it differs in its discontinuity: flowering days may be followed by intervals of several days when no flowers are open. Such pulsed bang flowering was reported in Myrcia rhodosepala Kiaersk. and in Blepharocalyx salici- folius (Kunth) O. Berg in central Brazil. Kevan and Lack (1985) described a similar pattern in Decas- permum parviflorum in Sulawesi. In this species flowering occurred regularly once in every two or three days with all plants in synchrony over the peak blooming time from late January to mid-Feb- ruary. The total flowering period of ca. 4 weeks was, however, considerably shorter than the ca. 8-week periods estimated for the central Brazilian pulsed- bang species. Also in Sulawesi, Syzygium lineatum had a total flowering period of approximately the same dura- tion as the nearby Decaspermum parviflorum (ca. 4— 5 weeks) but, as far as we could tell, appeared to exhibit a typical cornucopia pattern (Lack & Kev- an, 1984). Siphoneugena densiflora in central Brazil also exhibited the cornucopia strategy with many flowers being produced per plant over a period of a month or so and a population flowering period estimated at 8 weeks. The distinction between the cornucopia strategy and the steady state strategy in which relatively few flowers are produced each day over a long period of time may be somewhat arbitrary in some in- stances. The central Brazilian shrubs Psidium fir- mum Berg and Campomanesia pubescens (DC.) Berg, with flowering periods of ca. 12 weeks, were considered steady state (Proenga & Gibbs, 1994). The Amazonian tree Psidium acutangulum DC. had an even longer flowering period. Flowering was re- ported all year round with relatively few flowers being produced from January to March (Falcáo et al., 1992). Similar patterns were reported for other Amazonian fruit crops studied by these authors, in- cluding Eugenia stipitata McVaugh and Myrciaria dubia (Kunth) Mc Vaugh (Falcão et al., 1988, 1989). The term steady state seems inappropriate for these latter species, as monthly totals for flower produc- tion exhibit spectacular peaks and dips. A better example of steady state flowering is provided by the northern Queensland species Syzygium cormiflo- rum, which has a rather long flowering period (> 14 weeks) and apparently relatively few flowers open per day. Crome and Irvine (1986) commented that “flowers are not numerous” and that their ex- perimental program was limited by “insufficient flowers.” Syzygium tierneyanum, also from northern Queensland, appears to represent the opposite ex- treme of flowering strategy, viz., the classic big bang. The species is described as bursting into flower and reaching its peak over a 12-day period. Hopper (1980) counted 334 flowers per cubic meter of lower canopy and estimated that a 20-m plant could carry ca. 300,000 flowers at its seasonal peak. The central Brazilian species Eugenia dysen- terica DC. and Campomanesia velutina (Cambess.) O. Berg also exemplify the big bang strategy, albeit on a smaller scale (Proenga € Gibbs, 1994), as does Eugenia salamensis, a dry forest species in Costa Rica (Frankie et al., 1983). Two Costa Rican dry forest shrubs, Eugenia oer- stediana O. Berg and Psidium guineense Sw., ex- hibit a multiple bang strategy, flowering several times a year in synchronized flowering episodes lasting less than two weeks (Suarez € Esquivel, 1987). LONGEVITY OF INDIVIDUAL FLOWERS One-day flowers are the norm in the Myrtoideae. Species in which individual flowers last one day or less were reported from Blepharocalyx, Campoma- nesia, Decaspermum, Eugenia, Myrcia, Myrciaria, Psidium, Siphoneugena, and Syzygium (Hopper, 1980; Kevan & Lack, 1985; Peters & Vasquez, 1986/87; Proenga & Gibbs, 1994). Stratton (1989) reported a mean floral longevity of 1.27 days, and a range of 1.1-1.5 days for three myrtaceous spe- cies in Costa Rican cloud forest. In this study of 110 species, in 35 families, taxonomic constraints at family level were the most important determinant of floral longevity, explaining 73% of the variance. Singh and Sehgal (1968) reported that the stigma of Psidium guajava L. is receptive for 2 or 3 days. Primack (1985), however, included the Myrtaceae in a list of families with long-lived flowers, which typically last 4-19 days. This generalization was very likely based on examples from the Leptosper- moideae in which long-lived flowers appear to be common, e.g., flowers of Leptospermum scoparium J. R. Forst. & G. Forst. may last up to 3 weeks (Primack, 1980). Such longevity is rare in the Myr- toideae and where it occurs appears to be positively 486 Annals of the Missouri Botanical Garden associated with large flower size, e.g., Syzygium cormiflorum has flowers of ca. 4 cm in diameter, which do not brown until one week after anthesis (Crome & Irvine, 1986). EFFECTS OF FIRE AND DROUGHT ON FLOWERING In the Australian Myrtaceae flowering is not en- hanced by fires. In fact, the flush of vegetative rowth in recovering, fire-resistant species gener- ally inhibits flowering for several seasons (Beardsell et al., 1993). In contrast, van Wyk and Lowrey (1988) reported that in southern Africa, grassland fires promoted new growth and flowering in the rhi- suffrutex Eugenia albanensis similar zomatous geoxylic Sond., and these authors considered that : effects were likely in E. cf. mossambicensis and | E. Br. Cesar (1980) reported three Myr- taceae (one Campomanesia sp. and two undeter- mined Myrtaceae) among 50 species that presented some form of fire-induced flowering in central Bra- zilian savanna grassland. Eugenia myrcianthes Nied. flowers very soon after burning in the same habitat (C. Proenga, pers. obs.) Sanaiotti and Magnusson (1995) studied the ef- fects of annual fires on the production of fleshy fruits in a Brazilian Amazonian savanna. In Eugen- ia biflora DC. and Myrcia sylvatica (G. Mey.) DC. fruiting was dramatically reduced after fire, but re- covery was rapid with fruiting reaching 50-90% of normal levels in the first year after burning and 50— pusilla N. 100% in the second year. Van Wyk and Lowrey (1988) considered drought conditions a major factor in delayed flowering in uthern African Eugenia, while Falcão et al. (1988, 1989, 1992) reported flowering periods co- inciding with precipitation minima in three Ama- zonian species, i.e., Eugenia stipitata, Myrciaria dubia, and Psidium acutangulum. BREEDING SYSTEMS DICHOGAMY Beardsell et al. (1993) commented that all her- maphroditic species of the Myrtaceae studied so far, except for one protogynous species of Verticordia, have protandrous flowers, although there may be some overlap of male and female phases providing potential for self-pollination. Intervals of days or even weeks separating male and female phases are not uncommon in the Leptospermoideae, but the potential for such conspicuous protandry is signif- icantly reduced in the Myrtoideae where flowers are generally short-lived. Furthermore, the detection of dichogamy in the Myrtoideae is rendered more dif- ficult by the dry nature of the stigma, which does not permit visual assessment of receptivity. From a theoretical viewpoint it seems unlikely that temporal separation of male and female func- tions should be widespread but undetected among the Myrtoideae with short-lived flowers which offer only pollen as a reward. Female-phase flowers would offer nothing to reward visitors and, as Ren- ner (1989) commented, only mimicry or deception might account for visits to such flowers. That di- chogamy should co-occur with nectar production seems more likely and this does in fact appear to be the case in at least three of the four genera where the existence of dichogamy has been postu- lated. Dichogamy seems most likely to occur in Syzy- gium, where many species offer nectar as a reward; interestingly, this genus includes species with rel- atively long-lived flowers possibly affording more scope for temporal separation of male and female functions. In the context of the Myrtoideae the flow- ers of Syzygium cormiflorum, as already mentioned, are remarkably long-lived: styles do not brown un- til, on average, a week after anthesis, and they are not shed until about three weeks later (Crome & Irvine, 1986). According to these authors a flower could be successfully pollinated at anthesis, but they were unable to tell whether the stigma was actually receptive on the first day or whether the pollen retained viability until the stigma matured later. Thus oes is possible but not confirmed in this spec The mulie reliable report of protandry in the Myr- toideae that we have encountered to date is that of Grifo (1992) for Myrcianthes. In some species of this genus the stigma remains introrsely curved un- til after the a len through piany slits and are leaning out- nthers have released much of the pol- wards, away from the now erect and apparently receptive cee Grifo (1992) also reported nectar see Attractants and production in Myrcianthes Rewards below. Peters and Vasquez (1986/87) have documented protogyny in Myrciaria dubia, where the stigma is exserted first and the filaments of the numerous an- thers expand later. These authors stated that anthe- sis occurs early in the morning, that the flowers are receptive to pollination for a period of 4—5 hours, and that by the time the anthers emerge to release the pollen the stigma is no longer receptive to pol- lination. The method of assessment of stigma re- ceptivity was not discussed but in the accompa- nying diagram the stigma appears shrivelled in the flower with erect stamens. After pollination the sta- mens start to wilt and the whole hypanthial cup, Volume 83, Number 4 Nic Lughadha & Proenca 487 Reproductive Biology of Myrtoideae with sepals, petals and stamens, abscisses on the following day, leaving only the ovary. Grifo (1992) commented that the monotypic Amomyrtella is also decidedly protogynous. The large peltate stigma of this species is exserted before the petals unfold (herbarium label for Solomon 11018). Stylar extension post-anthesis is common in the Leptospermoideae. It appears to coincide with the onset of stigma receptivity and has been variously interpreted as a mechanism to favor outcrossing (Moncur & Boland, 1989) or as a mechanism to reduce interference from self-pollen (Lloyd & Webb, 1986; Webb & Lloyd, 1986). The phenom- enon appears to be rarer in the Myrtoideae but has been reported from two species of Syzygium. In the flowers of S. cormiflorum the style expands to max- imum length 4-5 days after opening and after the stamens are shed (Crome & Irvine, 1986). In con- trast, in the one-day flowers of S. lineatum, the style reaches its full length 4 to 6 hours after anthesis, often while the anthers still have much pollen on them (Lack & Kevan, 1984). ANDROMONOECY AND DIOECY Andromonoecy, which is common in the Lepto- spermoideae (Beardsell et al., 1993), has not been detected in the Myrtoideae. Interestingly, Beardsell et al. (1993) suggested that andromonoecy in the Australian Myrtaceae (Leptospermoideae) may rep- resent a response to soils of low fertility and to drought, allowing optimal resource allocation for re- production. The Myrtoideae, in contrast, though tol- erant of poor soils appear to be drought avoiders, reaching their maximum ecological importance (as estimated by density and/or basal area) in areas of South America where there is a combination of coolish temperatures, a steady supply of water, and a poor, weakly acidic soil (Proenga, 1991). The term cryptic dioecy has been applied to di- oecious breeding systems in which one or both of the functionally unisexual morphs appear to have perfect hermaphroditic flowers, making the dioecious condition difficult to detect (Mayer & Charlesworth, 1991). In populations of most cryptically dioecious species, plants with staminate flowers (male) co-exist with apparently perfect-flowered but functionally pistillate (female) plants whose anthers produce ster- ile or abnormal pollen, or are indehiscent. This sit- uation is sometimes mistaken for androdioecy, in which functionally male and functionally hermaph- rodite individuals co-exist. Nic Lughadha (1994) summarized current knowl- edge on dioecy in the Myrtaceae, which has until recently been considered a relatively rare phenom- enon (van Wyk & Lowrey, 1988). In the Myrtoideae, cryptic dioecy has been reported in Pimenta dioica (Chapman, 1964), in Decaspermum parviflorum (Kevan & Lack, 1985), and in all 15 species of Eu- genia native to South Africa (van Wyk & Lowrey, 1988). With the exception of Pimenta dioica, which has structurally hermaphroditic flowers in both sex- es, all of these species have male flowers with greatly reduced pistils, while female flowers appear perfect but generally have fewer stamens whose anthers do not produce viable pollen. In the typically small myrtaceous flower with its many stamens, both types of cryptic dioecy (apparently hermaphrodite and ap- parently androdioecious) may easily be overlooked by the casual observer and are probably more com- mon than the few literature reports suggest. Reduced styles and abortive ovaries have been reported for Pimenta guatemalensis (Lundell, 1968) and for three species of Calyptranthes, C. fascicu- lata O. Berg (Berg, 1857), C. longifolia O. Berg, and C. speciosa Sagot (Mc Vaugh, 1958). In the Flo- ra of Peru, McVaugh (1958) also noted short styles and apparent imperfect development of the hypan- thium in Myrcia aliena McVaugh and suggested that this species may be partially or completely di- oecious. A further two imperfect flowered species are among the thirty species of Myrcia described as new by McVaugh (1969). Myrcia imperfecta McVaugh and M. myriantha McVaugh (both from Mount Ayanganna, Guyana) were each based on a single collection with male flowers only and no ves- tige of a style. McVaugh commented that this was a condition rarely noted in Myrcia and that its sig- nificance was unknown, though the imperfect flow- ers of M. imperfecta could be abnormal. Myrcia al- masensis Nic Lughadha from the Pico das Almas, Brazil, was based on apparently male material with a vestigial style and brought to four the total num- ber of putatively dioecious species of Myrcia. Of course the possibility that some of these collections may represent andromonoecious species cannot as yet be discounted. Interestingly, many of these pu- tatively dioecious neotropical Myrtoideae are from high-altitude areas. Sobrevila and Arroyo (1982) have discussed the abnormally high incidence of dioecy in a Venezuelan montane cloud forest as compared to other tropical forest communities. Classical dioecy, where female flowers lack sta- mens completely, is unlikely to occur in the Myr- toideae. In flowers offering pollen as a sole reward to pollinators, which is the case for most Myrtoi- deae, selection tends to favor female flowers which mimic male flowers by retaining stamens, even though the pollen therein may be sterile (Lloyd, 1982). A possible mechanism for the evolution of 488 Annals of the Missouri Botanical Garden dioecy in pollen-only taxa is outlined under At- tractants and Rewards below. Dioecy is unknown in the Australian Myrtaceae (Beardsell et al., 3), and we consider it highly improbable that it has been extensively overlooked there. The only published example we have encoun- tered of dioecy in the Leptospermoideae occurs in the small genus Carpolepis, which is endemic to New Caledonia (Dawson, 1992). dioecious and, apparently, more or less cryptically so. Male flowers include fertile stamens, a morpho- logically normal, but non-functional ovary, a full- length style, and a peltate stigma; female flowers are Carpolepis elegans is similar but the stamens are replaced by recurved staminodes whose anthers bear no pollen. The other two species of Carpolepis bear hermaphrodite flowers with punctiform stigmas. Dioecism has also been discovered in some species of the Myrtoideae of New Caledonia, and these are shortly to form the basis of a new genus (J. W. Dawson, pers. comm.). The fact that dioecy is extremely rare in the Lep- tospermoideae yet appears to have arisen indepen- dently on at least half a dozen different occasions in the Myrtoideae can be interpreted as further ev- idence in support of the non-random association between dioecy and dispersal mode, which was de- scribed and discussed by Bawa (1980) and Givnish (1980, 1982). These authors argued that unisexu- ality is more likely to establish itself in taxa with large few-seeded animal-dispersed fruits than in taxa with other modes of dispersal. In species that produce nutrient-rich fruit an increase in female reproductive effort produces a disproportionate in- crease in female fitness due to selection for spatial and temporal peaks in fruit production. They con- sider that such an advantage to increasing female effort could, under certain circumstances, favor in- dividuals that invest only as females or only as OUTCROSSING RATES In most Myrtoideae, especially the small-flow- ered species, many flowers are produced per tree per flowering day. The potential for geitonogamy would therefore seem great. However the low values [Pre-emergent reproductive success, the percentage of all ovules maturing into seeds ex- pressed as: % natural fruit set X no. of seeds per 1987] typ- ically found in the Myrtoideae (Proenga & Gibbs, 1994; Nic Lughadha, unpublished data) are sug- fruit/no. of ovules per fruit, Wiens et al., gestive of predominantly outbreeding species. Electrophoretic studies have demonstrated the ex- istence of high outcrossing rates in various species of Eucalyptus and Melaleuca with similar flowering patterns (Phillips & Brown, 1977; Moran E pn ; Griffin et al. 7; Butcher et al., Beardsell et al. (1993) interpreted these high ae of outcrossing in spite of ample opportunities for e mw geltonogamy as reflecting the operation of a barrier to self-pollination such as self-incompatibility. Proenga (1991) suggested that high levels of out- breeding could be maintained even in self-compat- ible species through flowering strategies that favor trap-lining or opportunistic behavior by pollinators. In cultivated Psidium guajava, a relatively low lev- el of outcrossing of 25-41% (mean = 36%) was reported by Soubihe Sobrinho and Gurgel (1962). SELF-INCOMPATIBILITY Beardsell et al. (1993) considered that self-in- compatibility is probably widespread in the Myr- They cited cases ranging from partial self-incompatibility to complete self-sterility with no seed production after selfing. No myrtoid ex- taceae. amples of self-sterility were reported, but they con- sidered it likely that the same mechanism of self- incompatibility operates throughout the Myrtaceae and acknowledged the need to examine more spe- cies in order to determine whether reduced seed- set after selfing is due to the expression of lethal recessive genes in the zygote or to a late-acting self- incompatibility system. The latter explanation was favored by Proenca and Gibbs (1994) in their study of the reproductive biology of central Brazilian Myrtoideae. Of the eight species studied by these authors, three (Blepharocalyx salicifolius, Campo- manesia velutina, and Siphoneugena densiflora) were apparently strictly self-incompatible and set no fruit when selfed. A further two species (Myrcia linearifolia and Campomanesia pubescens) exhibit- ed partial self-incompatibility with ISI values [In- dex of Self-Incompatibility, expressed as the ratio of fruit set from selfed vs. crossed flowers] of 0.12 and 0.09, respectively. Self-pollen tubes were ob- served to penetrate ovules in all of these species and no differences were detected between self- an cross-pollinations with respect to the mean number of penetrated ovules per flower at 24, 48, or hours after pollination. The time of abscission of selfed pistils varied between species from one week to one month after pollination, but rejection was synchronous within species, and no ovary enlarge- ment was detected in crossed or selfed pistils up to the time of abscission of the latter. This repre- sents the most detailed study of “self-incompatibil- ity” in the Myrtoideae published to date but, as no material was fixed beyond 72 hours after pollina- Volume 83, Number 4 1996 Nic Lughadha & Proenca 489 Ç Reproductive Biology of Myrtoideae tion, we can only speculate as to the nature of the rejection mechanism in operation. More recent studies on Gomidesia have included fixations up to 28 days after pollination, and results obtained to date suggest that preferential outcrossing is sus- tained by a post-zygotic mechanism or effect (Nic Lughadha, unpublished data). Complete self-sterility has also been reported in Eugenia sp. in secondary deciduous forest in Ven- ezuela (Ruiz & Arroyo, 1978), in Eugenia sp. (pre- sumably different) in montane cloud forest in Ven- ezuela (Sobrevila & Arroyo, 1982), and in Syzygium lineatum in Sulawesi, Indonesia (Lack evan, 1984). Bullock (1985) reported almost complete self-incompatibility in Mexican Psidium sartorian- um Nied. (ISI 0.02). However, in none of these cases is any indication given of the timing or nature of the reproductive barriers operating after self-pol- lination. Reports of self-compatibility in Myrtoideae are almost as numerous as those of self-incompatibility. Proença and Gibbs (1994) recorded three com- pletely self-compatible species (Eugenia dysenter- ica, Myrcia rhodosepala, and Psidium firmum) that set statistically equal numbers of fruits after self- and cross-pollination. For the cultivated guava, Psidium guajava, self-pollination in isolated trees has been registered between 64% and 90% (Sou- bihe Sobrinho & Gurgel, 1962). In Peru, Myrciaria dubia showed 91% fruit set after geitonogamous pollination (Peters & Vasquez, 1986/87). In Vene- zuela, Myrcia fallax (Rich.) DC. showed signifi- cantly diminished fruit set after self-pollination as compared to cross-pollination; however, its ISI of 0.24 exceeded the threshold value of 0.2 arbitrarily set for self-incompatible species (Bawa, 1974; So- brevila & Arroyo, 1982), and so this species was classed as self-compatible. Syzygium cormiflorum from northern Queensland also exhibited partial self-compatibility with 18-36% fertilization (esti- mates based on ovule enlargement) in self-polli- nated flowers and 72.7-86.7% fertilization in cross-pollinated flowers (Crome & Irvine, 1986). In another northern Queensland tree species, Syzy- gium tierneyanum, bagged inflorescences set fruit in the absence of cross-pollination (Hopper. 1980 Beardsell et al. (1993) interpreted this result as demonstrating autogamy (in the sense of automatic self-pollination of a self-compatible species) but did not discuss the possibility of apomixis. Chan- taranothai and Parnell (1994) obtained similar re- sults in a study of breeding systems of Thai species of Syzygium. They found that all four species stud- ied in detail were apparently self-compatible with self-pollination appearing to enhance seed-set. — However, they further demonstrated that two of these species were actually apomictic (see below). Purseglove (1968) documented self-fertility in Psid- ium guajava. Schroeder (1947) reported results ranging from complete self-compatibility to almost complete self-incompatibility in different varieties of Acca sellowiana (O. Berg) Burret (as Feijoa sel- lowiana O. Berg). APOMIXIS Davis (1966) reported apomixis as occurring widely within the Myrtoideae. Where apomixis oc- curs it is generally by adventitious embryony, al- though there is some evidence indicating the pos- sibility of apospory. The resulting polyembryony is discussed in more detail under Embryo and Seed Development (see below). Rye (1979) postulated a positive correlation be- tween the widespread occurrence of apomixis in the Myrtoideae and the high frequency of polyploidy in this group, but commented that the only specific example was Syzygium jambos (cited as Eugenia jambos). A further example is Syzygium cumini, which exhibits adventitious polyembryony (though apparently not consistently: Narayanaswami & Roy, 1960a; Roy & Sahai, 1962 (as S. caryophyllifol- ium); Tiwary, 1926 (as Eugenia jambolana Lam.); van der Pijl, 1934 (as Eugenia cumini (L.) Druce); Chantaranothai & Parnell, 1994) and in which sev- eral authors reported varying levels of polyploidy (Mehra, 1976; Bir et al., 1980; Singhal & Gill, 1984; Singhal et al., 1984, 1985; Gill et al., 1989: as E. jambolana in each case, gametophytic counts 11, 22, and 33, sporophytic counts 22, 44, and 66). Whether this association extends to other polyem- bryonic members of the genus Syzygium and, more importantly, whether it is continued outside the ge- nus remains to be investigated. Luma apiculata, the only other polyembryonic species of the Myrtoideae for which chromosome counts are available, lends no support to Rye's hypothesis as it is reported to have n = 10 (Titow de Tschischow, 1956), or about 2n = 22 (Landrum, 1981). However, as there is circumstantial evidence that polyploidy may not be ubiquitous in this species, it is possible that these diploid counts may derive from non-polyembryonic material. The correlation cannot be tested further in the absence of chromosome counts for any other myrtoid genera in which polyembryony has been documented. POLLINATION BIOLOGY ATTRACTANTS AND REWARDS Petals and/or stamens may act as the visual at- tractants in flowers of the Myrtoideae, but the sta- 490 Annals of the Missouri Botanical Garden mens are generally the most conspicuous structures in the open flower. Scent also appears to play a role in attraction. The odors produced are generally de- scribed as sweet but flowers of Syzygium cormiflo- rum have a faint unpleasant smell (Crome & Irvine, 1986). Grifo (1992) described the flowers of Myr- cianthes as smelling either sweet and similar to an apple-blossom or rather sour and similar to “well- seasoned sneakers.” Chantaranothai and Parnell (1994) found that floral buds of Syzygium SS and S. megacarpum (Craib) Rathakr. € N. were sweetly fragrant during the period of nalli AN before anthesis. In southern African Eugenia spe- cies van Wyk and Lowrey (1988) did not discover any osmophores but found that the strong sweet odor was emitted either by the anther tisue and/or by the pollen grains. Pollen is the principal reward available to visi- tors of most Myrtoideae flowers. Van Wyk and Lo- wrey (1988) found that the pollen of southern Af- rican Eugenia tested positive for lipids (in the form of oil droplets) and negative for starch. Moncur (1988) commented that sugars in pollen are polli- nator rewards in Acca sellowiana (as Feijoa sellow- tana). In several dioecious species, such as the southern African Eugenia and Decaspermum par- viflorum, female flowers produce non-viable pollen- like material which may represent an important re- source to insect pollinators (Kevan & Lack, 1985; van Wyk & Lowrey, 1988). Kevan and Lack com- mented that D. parviflorum is unusual among di- oecious species in presenting pollen (and sterile pollen) as the main food reward and cited Vitis and Solanum as other examples of this phenomenon. However, pollen-only flowers also occur in other dioecious species such as Actinidia chinensis Planch. (Schmid, 1978), Rosa setigera Michx. (Kev- an et al., 1990), and Saurauia veraguasensis Seem. (Haber & Bawa, 1984). It is noteworthy that female flowers of these species accrue neither of the major advantages proposed by Bawa (1980) for female flowers of dioecious species: they must still allocate resources to the production of stamens and sterile pollen and still run the risk of their stigmas becom- ing clogged with their own sterile pollen. This could be interpreted as an indication that the avoidance of inbreeding may represent a more important se- lective force in the maintenance of dioecy in these species than does the reallocation of reproductive resources (Mayer & Charlesworth, 1991). it should be borne in mind that reallocation of re- sources may take place on a whole-plant level, with However, male plants producing a more prominent floral dis- play, more flowers, and/or more pollen per flower than female plants. This is certainly true of D. par- viflorum (Kevan & Lack, 1985), and there is cir- cumstantial evidence that it may be the case in m Pa dioecious Myrtoideae (Nic Lughad- 94). Givnish (1980) discussed the potential ob sexual selection in the evolution of dioecy in pollen-only taxa. Where pollen is the only re- ward then disproportionately many pollinators may be attracted to plants with heavy pollen loads. This advantage could drive the evolution of dioecy if male and female flowers mimic each other, as is the case in all dioecious Myrtoideae known to date (see discussion of dioecy above). Besides the pollen and sterile pollen offered as the principal reward, Decaspermum parviflorum is also reported to produce minute quantities of nectar (Kevan & Lack, 1985). Most other reliable reports of nectar production that we have encountered refer to species of the genus Syzygium. Schmid (1972b) described an ovarian annular nectary in S. aro- maticum (L. ) Merr. € L. M. Perry, S. jambos, S. malaccense, and S. paniculatum and cited Werth's (1901) account of the secretion of nectar in copious amounts into the concavity formed by the depressed top of the ovary in flowers of S. jambos (as Jambosa vulgaris DC.). Crome and Irvine (1986) measured copious but variable flow of weak nectar in S. cor- miflorum. Small quantities of nectar of rather high sugar content are produced in flowers of S. lineatum (as S. syzygioides; Lack € Kevan, 1984). Nectar production is also reported from S. tierneyanum (Hopper, 1980) and S. samarangense (Blume) Merr. & rry (Chantaranothai & Parnell, 1994). Studies of genera of Myrtoideae other than Sy- zygium have generally failed to find any evidence of nectar production. Nectar is absent in flowers of southern African Eugenia species (van Wyk & Lo- wrey, 1988) and, apparently, in most South Amer- ican Myrtoideae (Landrum, 1986; Proenca & Gibbs, 1994; Nic Lughadha, unpublished data). Exceptions include the report by Pirani and Cor- topassi-Laurino (1993) of bees collecting pollen and nectar from flowers of Plinia glomerata (O. Berg) Amshoff and that of Peters and Vasquez (1986/87) on nectar production in Myrciaria dubia. The other notable exception is the genus Myrcian- thes, in which sweet-tasting nectar is secreted at the base of the staminal disk (Grifo, 1992). Grifo's descriptions of visitor behavior indicate that nectar is the principal reward offered by these species. Small insects may approach the flowers and feed on the nectar while entirely avoiding the numerous stamens, but larger insects become covered with pollen as they brush the anthers in search of nectar. Nectar is widespread in the Leptospermoideae, and the nectar-producing Chamelaucium group Volume 83, Number 4 1996 Nic Lughadha & Proenca Reproductive Biology of Myrtoideae emerged as sister to the Myrtoideae in cladistic analyses by both Johnson and Briggs (1984) and Grifo (1992) (along with the Leptospermum group in the former analysis). Grifo’s (1992) cladistic anal- ysis also indicated that Myrcianthes is “much more closely related to Syzygium than previously sup- posed” and that these genera are basal to the Eu- geniinae in the Neotropics. We can therefore hy- pothesize that nectar production is basal in the Myrtaceae with the numerous pollen-only genera of the Myrtoideae arising as a result of secondary loss(es) of this facility. Renner (1989) has proposed a similar scenario to explain the distribution of nec- tar production among the neotropical melastomes. Her observation that the Melastomataceae that offer nectar are pollinated by a broader range of polli- nator classes than the pollen-only species seems also to hold true for the Myrtaceae on the basis of the evidence available to date. The role of the secretory cavity at the apex of the anther connective does not appear to have been investigated in any member of the Myrtoideae. In Thryptomene calycina (Lindl.) Stapf, a member of the Chamelaucium alliance (Leptospermoideae s.l.), Beardsell et al. (1989) demonstrated that the anther connective glands secrete a lipid-rich fluid which solidifies and serves as a food reward for pollinating insects. Although a similar role is pos- sible for the anther connective glands of the Myr- toideae it seems unlikely that any visitor small enough to be attracted to the minute quantities of secretion that these glands might afford would con- tact the stigma regularly while collecting. Renner (1989) discussed a comparable paradox offered by the staminal glands of Mouriri (Memecylaceae) and noted Morley’s (1976) suggestion that these glands may be involved in odor production. Van Wyk and Lowrey’s (1988) observations on fragrant anthers and pollen in southern African Eugenia represent the only evidence we have encountered to support the hypothesis of an odor production role for the anther connective glands in the Myrtoideae. Of course, another possibility is that these structures serve no function in the flowers of the Myrtoideae and are simply vestigial remnants of an organ which is more or less ubiquitous in the Myrtaceae. The scattered examples of species whose anthers lack secretory cavities undoubtedly represent sec- ondary losses of this feature. The role of petals as an attractant and reward to visitors to Acca and Myrrhinium is discussed under Bird Pollination below. BEE POLLINATION Bees (Apoidea) appear to be the most common pollinators of Myrtoideae, as they are of Myrtaceae as a whole. There is a strong association between the Myrtaceae and the short-tongued Colletidae, which are considered to be the most primitive flow- er-visiting bees (Michener, 1979). This relationship is most notable and best documented in Australia where nearly half of the bee species belong to the Colletidae and most of these are restricted to or collect pollen primarily from flowers of Myrtaceae (Armstrong, 1979). These include Eugenia and Sy- zygium as well as numerous non-myrtoid genera. Bees reported to collect pollen from Eugenia and/or Syzygium include Euryglossina and Hylaeus (Col- letidae) as well as Homalictus and Lasioglossum (Halictidae) (Michener, 1965). When current taxo- nomic concepts are applied, most of these records are referrable to the genus Syzygium; only one spe- cies of Eugenia s. str. occurs in Australia (Hyland, 1983). Proenca and Gibbs (1994) suggested that the Myrtaceae—Colletidae association, so striking in Australia, probably exists to some extent in the Neotropics. They cited two neotropical examples of this association, viz., rapid pre-dawn pollen remov- al from Eugenia salamensis by Ptiloglossa spp. (Colletidae: Diphaglossinae) in Costa Rica (Frankie et al., 1983) and pollination of Siphoneugena den- siflora by Ptiloglossa sp. in central Brazil (Proenga, 1992). Colletidae are not particularly common in the Neotropics: e.g., in central Brazil, where the Proenca and Gibbs study was carried out, only 5% of bee species are Colletidae (A. Raw, unpublished list based on 12,000 collections). Therefore one could argue that such reports are unlikely in the absence of some degree of association. However, a careful compilation of 33 reports of bee visitation in neotropical Myrtoideae (Table 2) showed that most registered visits are by Apidae: Meliponinae (14), followed by Apidae: Bombinae (9), Halictidae (4), Anthophoridae (4), Colletidae (2). In defense of the neotropical Myrtaceae—Colletidae association, it should be noted that both reports were of pre-dawn visitation, so that this phenomenon could be wide- spread but largely overlooked in studies based purely on diurnal observations. The first report of buzz-pollination in the Myr- taceae was published by Proença (1992). She de- scribed buzz-pollination in Siphoneugena densiflora by Ptiloglossa sp., in Myrcia torta DC. (as M. dic- tiophylla (O. Berg) J. R. Mattos & D. Legrand) by Augochloropsis? sp. (Halictidae), and in Myrcia rho- dosepala and Blepharocalyx salicifolius by Bombus spp. (Apidae: Bombinae). Bombus spp. (including B. atratus and B. morio) were considered the prin- cipal pollinators of a further five species of central Brazilian Myrtaceae (Campomanesia pubescens, C. 492 Annals of the Missouri Botanical Garden Table 2. Species of neotropical Myrtoideae with presumed pollinators or insect visitors to flowers. Species Visitors Behavior Reference Blepharocalyx salicifolius “visitors” Proenga & Gibbs (1994) Apidae: Bombinae Apidae: M Halictidae Apidae: Bombinae Apidae: Bombinae Apidae: Bombinae Apidae: Meliponinae Colletidae Apidae: Meliponinae Apidae: Meliponinae eliponinae Campomanesia pubescens Campomanesia velutina Eugenia dysenteric Eugenia salamensis Eugenia stipitata Eugenia ] Halictidae Apidae: Meliponinae Apidae: Bombin Apidae: Bombinae Eugenia sp. 2 Eugenia spp. Myrcia linearifolia ibinae Myrcia rhodosepala Anthophoridae: Xylocopinae Halictidae ? Apidae: Meliponinae Myrcia torta Myrcia sp. Myrciaria dubia 1 Apidae: Meliponinae Myrciaria dubia 2 Apidae: Meliponinae Pimenta dioica Anthophoridae: Xylocopinae “visitors” Proenga & Gibbs (1994) Proenga & Gibbs (1994) Proenga & Gibbs (1994) “visitors” “visitors” "orging" Frankie et al. (1983) visiting" Falcáo et al. (1988) "aggressive Roubik (1989) Ruiz & Arroyo (1978) Absy & 977) "collect pollen" "visitors" Proenca & Gibbs (1994) “pollinators” Proença & Gibbs (1994) “pollinators” Proença & Gibbs (1994) “pollen in Vit € D'Albore (1994) oney” Falcáo et al. (1989) visiting Anthophoridae: Anthophorinae Halictidae Plinia cauliflora Apidae: ias Apidae: Bombinae Apidae: Meliponinae Plinia glomerata Psidium acutangulum Apidae: Meliponinae Psidium firmum Apidae: Bombinae Apidae: Meliponinae Apidae: Bombinae Apidae: Meliponinae Psidium guajava | Psidium guajava 2 Colletidae Siphoneugena densiflora Myrtaceae indet. spp. Apidae: Meliponinae I Anthophoridae: Xylocopinae “pollinators” Peters & Vasquez (1986/87) “visiting” Chapman (1965) “inten Guibu et al. (1988) exploitation” “foraging” Pirani & Cortopassi- Laurini (1993) “visiting” Falcáo et al. (1992) “presumptive Proenga & Gibbs (1994) pollinator” “foraging” Camillo & Garófalo (1989) “intensive Guibu et al. (1988) exploitation Proenga (1992) Vit & D'Albore (1994 ~w Note. Visits by honeybees, introduced to the Neotropics, velutina, Eugenia dysenterica, Myrcia linearifolia, and Psidium firmum), but no buzzing behavior was noted during observation of these species. The same species of Bombus exhibited a preference for Psidium guajava in secondary vegetation in south- eastern Brazil (Camillo € Garófalo. 1989). Bombus spp. were reported to collect pollen and nectar from Plinia glomerata in Sáo Paulo (Pirani & Cortopas- si-Laurino, 1993). Visits by Melipona quadrifascia- ta, pican angustulata (both Apidae: Meli- poninae), and Apis mellifera were also recorded. Guibu et al. (1988) documented intensive exploi- are excluded from this table. tation of Myrtaceae by Melipona quadrifasciata. They listed Psidium guajava s.l. and Plinia cauli- flora (Mart.) Kausel s.l. (as Myrciaria cauliflora rt.) O. Berg s.l.) among the species visited by these bees in Sáo Paulo, Brazil. Other species of the bee genus Melipona may be important pollinators of Myrtaceae in Amazonian Brazil. Melipona rufiventris and M. seminigra col- lected pollen of Eugenia spp. and other unidenti- fied ce near Manaus (Absy & Kerr, 1977; Absy et al., n the same area Falcáo et al. (1988, 1989, on recorded Melipona lateralis and Volume 83, Number 4 1996 Nic Lughadha 8 Proenca 493 Reproductive Biology of Myrtoideae M. pseudicentris visiting cultivated Eugenia stipi- tata, Myrciaria dubia, and Psidium acutangulum. In the nearby Reserva Ducke, Roubik (1989) ob- served aggressive behavior of several species of Tri- gona (Meliponinae) on Eugenia sp. In Peru Myr- ciaria dubia is commonly pollinated by Melipona de ily and Trigona postica (Peters & Vasquez, 6/87). A recent study of pollen spectra in the holler of 48 species of Melipona in Venezuela (Vit & D’Albore, 1994) provides additional evidence of the importance of Myrtaceae to these bees: Myrta- ceae spp. and Myrcia sp. were respectively the fifth and eighth most common pollen types in a list of 13. Ruiz and Arroyo (1978) reported visits to a Ven- ezuelan Eugenia sp. by Augochloropsis fulvofim- briata (Halictidae) and by Apis mellifera and Tri- gona testiceicornis (Apidae). Chapman (1965) saw solitary bees belonging to the genera Ceratina (Anthophoridae, Xylocopinae), Exomalopsis (Anthophoridae, Anthophorinae), and Halictus (Halictidae), as well as honeybees, visiting flowers of Pimenta dioica for pollen in Jamaica. In a population of Decaspermum parviflorum in a forest clearing in Indonesia, Kevan and Lack (1985) recorded vigorous pollen collection during a 45-minute peak period by a variety of bees, mostly Apis dorsata (Apidae: Apiinae) and Nomia spp. (Halictidae: Nomiinae). Nomia and Trigona contin- ued to glean during the rest of the day and per- functory visits by Xylocopa spp. were also reported. In a nearby forest Syzygium lineatum (as Syzygium syzygioides and subsequently corrected) was visited sparingly and mainly for nectar (Lack & Kevan, 1984). Another study notable for the paucity of po- tential pollen vectors was that of van Wyk and Lo- wrey (1988). These authors provisionally consid- ered that the 15 southern African species of Eugenia studied might be bee-pollinated, as hon- eybees were observed visiting Eugenia spp. in Pre- toria and were also active in natural populations of Eugenia capensis Harv. BIRD POLLINATION Bird pollination appears to be much less fre- quent in the Myrtoideae than in the Leptospermoi- deae. Ford et al. (1979) reported six species of birds visiting three species of Syzygium in Austra- lia, but in no case was a bird actually seen to be carrying pollen. Honeyeaters were reported as the most commonly observed vertebrate visitors to flow- ers of Syzygium cormiflorum and S. tierneyanum in Australia (Hopper, 1980; Crome & Irvine, 1986). Hopper (1980) suggested that honeyeaters may be the most important pollinators of S. tierneyanum but acknowledged that the role of bats was neglected in his study. Crome and Irvine (1986) demonstrated experimentally that in S. cormiflorum birds were far less effective pollinators than bats, accounting for less than 25% of all successful pollinations. Both of these species have white or cream flowers, but Ford et al. (1979) noted a tendency for red stamens among large-flowered, bird-pollinated Australian Myrtaceae in general (mostly Leptospermoideae). Two small neotropical myrtoid genera, Acca and Myrrhinium, have stiff red stamens reminiscent of those of many Australasian bird-pollinated Myrta- ceae. Landrum (1986) noted that there is no nectar in the fresh flowers of Acca sellowiana and that in both genera the petals change color and become sweet and juicy just as the anthers dehisce. There are various reports of visiting birds eating the petals of Acca sellowiana (Kiaerskov, 1893; Knuth, 1906; McGregor, 1976; Vogel et al., 1984). In view of their isolated position within the Myrtoideae, Lan- drum (1986) hypothesized that the use of petals as an attractant in these two closely related genera (which he called the Myrrhinium complex) is very ancient. Schroeder (1947), however, observed ex- tensive bee visitation in Acca sellowiana cultivated in California and found that flowers so visited set fruit ca. 16 times as well as flowers protected from visitation. Free (1993) also reported that bees are frequent visitors to the flowers of this species and assumed that they were responsible for most of the pollinations. In Myrrhinium atropurpureum Schott recently opened flowers have deep wine-red petals with a normal petal texture, crinkled filaments, and closed anthers; putatively receptive flowers have in- flated, very pale lavender petals, stiff filaments, and dehisced anthers; the pale petals provide a dra- matic contrast to the rest of the inflorescence, which is a uniform deep wine red, even to the axes, and the petal texture is unlike that of any other flower known to us, aerenchyma-rich, resembling the hollow structure of certain rubiaceous fruits such as Coccocypselum (C. Proenga, pers. obs.). It seems likely that Acca and Myrrhinium evolved un- der selection for bird pollination (perhaps in areas with an impoverished bee fauna) without totally los- ing their adaptations to bee pollination, so that in cultivation Acca sellowiana at least still responds to bee pollination. MAMMAL POLLINATION Long-tailed pygmy possums (Cercartetus cauda- tus) feed on Syzygium cormiflorum in northern Queensland (Hopper, 1980). Beardsell et al. (1993) suggested that the cauliflorous flowering of several 494 Annals of the Missouri Botanical Garden Syzygium species might facilitate access to the flowers by larger marsupials, which move up and down the trunks and larger branches. These authors also stressed that bats, being nocturnal and not eas- ily studied, might be more significant pollinators of some Myrtaceae than had been thought previously. Syconycteris australis (Queensland blossom bat) ap- peared to be one of the minor pollinators of the night-flowering rainforest tree Syzygium tierneyan- um (Hopper, 1980) and also made nocturnal visits to the flowers of the cauliflorous S. cormiflorum (Crome & Irvine, 1986) as did Macroglossus lago- chilus, another small blossom bat. In S. cormiflorum flowers open at any time of the day or night, but bats were the single most important pollinators, al- though visits by birds to the flowers were more fre- quent and more numerous. Start and Marshall (1976) reported the bats Eonycteris spelaea and Macroglossus minimus feeding on S. malaccense in West Malaysia. WIND POLLINATION Grifo (1992) considered wind-pollination possi- ble but unlikely in Myrcianthes. Peters and Vasquez (1986/87) mentioned the possibility of wind-polli- nation in Myrciaria dubia and Moncur (1988) the same for Acca sellowiana (as Feijoa sellowiana), but in both of these studies bees were considered to be the most important pollinating agents. FRUITS FRUIT CHARACTERISTICS AND DISPERSAL AGENTS The typical fruit of the Myrtoideae is a fleshy single-seeded, usually orange, red, or black berry, and the Myrtoideae have long been characterized as a fleshy-fruited subfamily or tribe. However, semi-dry to dry berries (i.e., leathery or pithy) have also been recorded, especially in Eugenia and re- lated genera. Drupaceous fruits are reported for Myrtella and Stereocaryum. Subdrupaceous fruits, with variable levels of woodiness of the endocarp, are found in some species of Eugenia s.l., Rhodo- myrtus (Schmid, 1980), and in Acmena (Hartley & Craven, 1977). Drupoid fruits with 1—3 pyrenes oc- cur in Eugenia and Hexachlamys (Berg, 1857; Rot- man, 1982). A curs in Campomanesia. drupoid in that the single seeds developing in each of the locules adhere to the endocarp, as in a true drupoid fruit, but it differs in the texture of the endocarp, which can be woody or merely glandular- verrucose, simulating a false seed-coat (Landrum, 1982). The more primitive subtribe Myrtinae in- very specialized “drupoid” fruit oc- It may be considered cludes many species that have green or yellow, sev- eral- to many-seeded fruits. Our knowledge of dispersal agents is mainly based on deductions from fruit morphology; few ac- tual dispersal events have been reported. Fleshy perigynia have clearly evolved in association with zoochory. Briggs and Johnson (1979) highlighted two common secondary developments that may oc- cur independently or in conjunction, viz., increase in fruit size and reduction in seed number. Proença (1991) suggested that single-seeded fruits, i.e., "packaging each seed separately," would enable the plant to abort inferior zygotes on an individual basis and to mature fruits at different rates, thus permitting more dispersal events for the same num- ber of seeds. Various dispersal mechanisms have been sug- gested for many-seeded fruits. Landrum (1986) considered that most species of Campomanesia best fit the mammal dispersed syndrome as defined by Janson (1983) based on an Amazon rainforest com- munity, but Snow (1981) registered Campomanesia as forming part of the diet of specialized frugivorous birds. Motta Jünior et al. (1994) reported that Dus- icyon thous, a small nocturnal Canidae, feeds on fruits of Psidium sp., Campomanesia sp., and Pli- nia cauliflora (as Myrciaria cauliflora). Seeds of Campomanesia sp. recovered from scats germinated successfully both in situ and in the laboratory, in- dicating that this small mammal is a potentially effective dispersal agent. The genus Psidium is cit- ed by Snow (1981) as having fruits that form part of the diet of specialized frugivorous birds, but Psidium firmum is apparently consumed by small rodents (Proenga, 1991). Evidence to this effect in- cluded a gnawed-through peduncle of an unripe fruit, marks of superficial bites on unripe fruit, ripe fruit torn open and the flesh partly eaten, and near- by scats which were identified as pertaining to Rhipidomys or Oryzomys. Cultivated Psidium gua- java is eaten by parrots ( , pers. comm.), other birds (Advani, 1981). and bus s (Fun. milayo, 1980; Advani, 1982). Other multi-seeded fruits mentioned by Snow (1981) include those of Decaspermum and Rhodamnia, which are exploited by specialized frugivorous birds, and Calycolpus, exploited by specialized and unspecialized frugiv- orous birds. Kevan and Lack (1985) observed mis- tletoe-birds (Dicaeidae) feeding on the fruits of Decaspermum parviflorum and discussed bird dis- persal in this dioecious species in relation to the selective pressure for the production of energy-rich fruits. Ant-dispersal by Messor minor (André) has recently been reported for Myrtus communis, the Volume 83, Number 4 1996 Nic Lughadha & Proen 495 ¢a Reproductive Biology of Myrtoideae seeds of which bear elaiosomes (Aronne & Wil- cock, 1994 Johnson and Briggs (1984) suggested that in di- verse lines of Myrtaceae, fruits with large, single seeds seem to be adapted to vertebrate dispersal in tropical forest. In the large genus Syzygium, one- seeded fruits are the norm, but mature fruits from different species may differ in diameter by an order of magnitude (Hyland, 1983). K. Fischer (pers. comm.) compared the morphology and nutritional and secondary chemistry of fruits of New Guinean species of Syzygium eaten by birds and by bats. In the species of Syzygium studied by her, fruits eaten by bats were both longer and wider than those con- sumed by birds but did not differ in nutrient com- position or secondary chemistry. Among 16 wild fruits edible to man that were studied in Malawi, Syzygium guineense Guill. & Perry had the highest levels of iron and magnesium (Saka & Msonthi, 1994). Beardsell et al. (1993) noted that in many species of Syzygium the infructescences are borne at some distance from the tips of the branches thus allowing larger animals access to the fruits. They cited the extreme case of Syzygium cormiflorum, the berries of which develop on the main trunk and larger branches, and suggested that this arrange- ment enables larger marsupials to eat the fruits dur- ing tree to tree movements. Fruits of a cauliflorous species of Myrciaria are eaten by bats (Semir, 1984). Eugenia s.l. is cited as consumed by un- specialized frugivorous birds (Snow, 1981). Two un- identified species of Eugenia s. str. from Costa Rica are eaten by Pharomachrus mocinno (Trogonidae) and by Aulacorhynchus prasinus (Rhamphastidae), respectively (Wheelwright et al., 1984). Fruits of Eugenia punicifolia (Kunth) DC. are probably con- sumed by pheasant-like Tinamidae in central Bra- zil, and Siphoneugena densiflora fruits are eaten by Miyarchus swansonii (Tyrannidae), an opportunistic fly-catcher, and by other birds (Proenca, 1990). Fruits of Myrcia torta DC. (as Myrcia dictiophylla) are eaten by Neothraups fasciata (Emberizidae), a generalist low-foraging tanager (Alves, 1992). Ob- servations of Blepharocalyx salicifolius revealed that the fruits are swallowed whole by migratory Elaenia chiriquensis and Tyrannus melancholicus (both Tyraniidae) and by resident Neothraups fas- ciata (Emberizidae), all opportunistic frugivores that also feed on insects (Paes, 1993). Snow (1981) listed Acmena, Eugenia, Myrcia, and Syzygium as forming part of the diet of specialized frugivorous birds, while fruits of Eugenia and Myrcia are also exploited by non-specialized birds. He commented that while at least 13 genera of Myrtoideae provide fruits that are eaten by frugivorous birds, special- ized and unspecialized, none seems to provide the staple diet of any frugivorous bird. In this respect the Myrtoideae resemble the Euphorbiaceae but differ from the Lauraceae, Burseraceae, and Pal- mae. Grifo (1992) saw monkeys of the genus Al- ouatta consuming the fruits of Myrcianthes pungens in Corrientes, Argentina. However, her observations suggested that the fruits, including the seeds, were chewed, thus considerably reducing the possibility that the monkeys were agents of dispersal. OVULE-SEED RATIOS AND REPRODUCTIVE CAPACITY In the fruits of most Myrtoideae, including the many-seeded ones, the number of mature seeds is much smaller than the original number of ovules in the ovary from which the fruit is derived. Even in genera in which the ovule number is reduced to two in each of two (or three) locules, most ovules do not develop into mature seeds. This superfluity of ovules may simply represent a relictual situation, with these species being derived from taxa in which most or all ovules within the ovary would have ma- tured into seeds (Caspar & Wiens, 1981). For ex- ample, Landrum (1981) envisaged an ancestor of Myrceugenia that developed all or nearly all its ovules into seeds while in present-day species few ovules develop into seeds. Alternatively, the super- fluous ovules may provide scope for the exercise of female choice (Stephenson & Bertin, 1983) and/or allow for the wastage of ovules inherent in a post- zygotic “self-incompatibility” system. In the dioe- cious Pimenta dioica (and the putatively dioecious P. guatemalensis) the number of ovules scarcely ex- ceeds the number of seeds per fruit (Landrum, 1986). We suggest that in these species there is no need for superfluous ovules, since all pollen tubes must originate from non-self pollen. In the Australian Myrtaceae (largely Leptosper- moideae), Rye and James (1992) identified reduc- tion in ovule numbers and competitive selection among fertilized ovules as mechanisms leading to reduction in seed number. These trends were in- terpreted as the result, in part at least, of selection for larger seed size. In addition, they found that plant size is positively correlated with ovule num- ber and seed set, and that the effects of all these factors tended to reinforce each other in determin- ing the net reproductive capacity of the species. Another factor that has been linked to reproduc- tive capacity is chromosome number. Rye and James (1992) found that high reproductive capacity is associated with high dysploid chromosome num- bers in four genera including Eugenia. 496 Annals of the Missouri Botanical Garden EMBRYO AND SEED DEVELOPMENT EMBRYO DEVELOPMENT The Myrtoideae apparently follow the onagrad pattern of embryo development but, as in the Myr- taceae as a whole studied. The deant of information in this area is rudimentary , the process has scarcely been acute, even in the context of our knowledge of Myrtoideae biology in general, and may be attributable to the extreme difficulties ex- perienced in fixation for histological studies in the family (Mauritzon, 1939; Davis, 1968; Beardsell et 989, 1993). SEED-COAT Beardsell et al. (1993) considered a seed-coat formed from both integuments as “a feature of all Myrtaceae so far examined.” marize the situation in the Leptospermoideae it is While this may sum- far from true with respect to the Myrtoideae. Van Wyk and Botha (1984) reported testa formation from the outer integument only in some southern African Eugenia species and a pachychalazal seed- coat in others. Narayanaswami and Roy (1960b) de- scribed the testa in Psidium guajava (and in P. cu- javillus Burm. f., now considered synonymous) as being formed from the outer integument only. The seed-coat appears to be entirely absent in Acmena, Acmenosperma, and Waterhousea (Hartley & Cra- ven, 1977; Hyland, 1983). McVaugh (1968) indicated that the testa varied in thickness from group to group in the neotropical Myrtoideae and may be membranous, leathery, car- tilaginous, or bony. Where the seed-coat is hard (i.e., in most Myrtinae apart from Landrum’s Cam- pomanesia complex), an operculum in the seed- coat, located directly over the basal end of the hy- pocotyl, allows the embryo to emerge. Landrum and Stevenson (1986) detected a strong correlation be- tween seed-coat texture and embryo structure but admitted the possibility that changes in embryo morphology naturally follow changes in seed-coat texture. Whatever the thickness, the testa is usually quite smooth and unsculpturec ampomanesia the thick glandular locule wall adheres to the delicate membranous testa of the single seed maturing inside the locule, as described above. The locule wall thus serves as a false seed- coat, to which Landrum (1982) attributed a protec- tive function, hypothesizing that frugivorous birds or mammals would tear apart the fruits without bit- ing into the glandular turpentine-smelling walls. Some field observations support this theory, as par- tially eaten fruits were found to have the locules intact (Landrum, 1986) but, as discussed above, some seeds are clearly ingested by small mammals (Motta Júnior et al., 1994). Proença (1991) sug- gested a possible co-adaptive role for the pseudo- testa, interpreting it as a defense to prevent devel- oping larvae of Tephritidae flies from migrating from infected to healthy locules. The situation in Campomanesia, where each locule harbors a single developing embryo, was compared to that in a spe- cies of Berberis in which high mortality of Tephrit- idae larvae was found to occur in single-seeded fruits when compared to several-seeded fruits (Her- rera, 1984). Tephritidae are notorious predators of (Malavasi & Morgante, 1981; Tan & Lee, my yrt aceous fruit crops ; Morgante & Malavasi, D Burk, 1983). = 2 EMBRYO STRUCTURE Embryo structure has been considered funda- mental in the classification of the Myrtoideae since the time of De Candolle ( ). He distinguished three basic embryo types commonly referred to as eugenioid (with thick, fleshy cotyledons and a rel- atively insignificant hypocotyl), myrcioid (with leafy cotyledons that are much broader than the hypo- cotyl), and myrtoid/pimentoid (with a well-devel- oped hypocotyl and relatively small, narrow coty- ledons). used these differences as the basis for their clas- sifications, little attention has been paid to the bi- ological significance of these structures. One no- Although most subsequent authors have table exception is the discussion by Landrum and Stevenson (1986) in which the various embryo types are interpreted as different responses to the same selective pressure for increased food storage in the embryo. Essential oil-secreting schizogenous glands occur in the embryos of Pimenta racemosa (P. Miller) J. W. Moore (as Myrcia acris DC.), Luma apiculata (as Myrceugenia apiculata), Myrceugenia glaucescens (Cambess.) D. Legrand & Kausel (as Eugenia glaucescens Cambess.), Myrcianthes pun- gens (as Eugenia pungens), and Eugenia discolor DC. and are absent from Myrtus sp., Psidium, Acca (as Feijoa), and Eugenia uniflora (as Stenocalyx michelii) (Petit, 1908). In a study of southern Af- rican Eugenia, four species had conspicuous glands and five “apparently eglandular but usually with a few obscure glands species were described as mainly associated with the radicular protuberance” (van Wyk, 1980). Hyland (1983) recorded oil glands in the embryos of 4 out of ca. 50 species of Australian Syzygium studied, i.e., S. macilwraith- ianum B. Hyland, S. sayeri (F. Muell.) B. Hyland, S. velae B. Hyland, and S. wilsonii B. Hyland. In Volume 83, Number 4 Nic Lughadha & Proenca 497 Ç Reproductive Biology of Myrtoideae the Leptospermoideae the embryo lacks oil glands (Petit, 1908). ENDOSPERM The endosperm is nuclear and is generally de- scribed as being digested by the developing embryo and therefore absent from the mature seeds (Tobe & Raven, 1983). However, studies by Petit (1908, cited in van Wyk & Botha, 1984) have shown that traces of endosperm are present in many myrta- ceous seeds including those of the genera Acca (as Feijoa), Myrtus, and Psidium (all Myrtoideae). Van Wyk and Botha (1984) reported the presence of en- dosperm in mature seeds of some species of south- ern African Eugenia. They suggested that variation in the amount of endosperm formed in the young seed may be of taxonomic importance, whereas the quantity of endosperm that remains in mature seeds is variable within species and therefore less likely to prove taxonomically significant. POLYEMBRYONY Tobe and Raven (1983) described polyembryony as usual in Syzygium but our literature survey sug- gests that, though frequent, it is by no means the norm in this genus. The earliest reports are by Ti- wary (1926), who concluded that all (six) of the species of Syzygium (reported as Eugenia) exam- ined by him showed “the existence of some kind of polyembryony.” However, as he interpreted early bifurcation of the plumule as evidence of polyem- bryony, he may have overestimated the frequency of this latter phenomenon within the genus. In only two of the species he examined (S. cumini and S. jambos, as Eugenia jambolana and E. jambos, re- spectively) was the presence of more than one em- bryo in the seed actually demonstrated. Subsequent studies have confirmed the existence of polyembryony in Syzygium jambos, disputed its occurrence in S. cumini (van der Pijl, 1934; Hen- derson, 1949; Roy, 1953; Chantaranothai & Par- nell, 1994) and in general indicated that the con- dition was far from ubiquitous in the genus. Van der Pijl (1934) reported polyembryony in S. aqueum (Burm. f.) Alston (as Eugenia aquea Burm.f.), S. jambos (as E. jambos), S. javanicum Miq. (as E. javanica Lam.), and S. malaccense (as E. malaccensis), but found no evidence for poly- embryony in S. cumini (as E. cumini) on the basis of the examination of specimens of this species from four different localities and fruits obtained from the local market. Henderson (1949) found polyembryony in “some species” of the 138 in the genus Eugenia s.l. in Malaya, and Merrill and Perry (1938) reported that of the 45 species of Syzygium known from China, 5 are regularly polyembryonic carpum (as S. latilimbum (Merr.) 4 Perry), S. forrestii Merr. & L. M. Perry, and S. han- cei Merr. & L. M. Perry). Hyland (1983) described ca. 50 species of Syzygium occurring in Australia and recorded consistent polyembryony in Syzygium paniculatum and the presence of “two-seeded fruits” in S. aqueum (polyembryony? see S. aqueum above). Chantaranothai and Parnell (1994) found no evidence of polyembryony in S. cumini, nor in S. megacarpum or S. formosum, but confirmed the ex- istence of polyembryony in S. jambos and S. ma- laccense and reported production of 2-6 seedlings from 1-3-seeded fruit in S. siamense (Craib) P. Chantaranothai & J. Parn. We concur with van Wyk and Botha's (1984) conclusion that most reports of polyembryony in dia are probably referrable to the genus Syzygiu Johnson (1936) Rm 2-22 embryos in seeds of Luma apiculata (as Eugenia hookeri Steud.). However, Landrum (1986) revised Luma apiculata and, although he had obviously dissected seeds and examined embryos, did not mention the occurrence of polyembryony. Thus we may infer that polyem- bryony is not constant in this species. Gurgel and Soubihe Sobrinho's (1951) study of myrtaceous fruit-bearing trees in Brazil is one of the most complete studies of polyembryony in the Myrtoideae to date. Sixteen species were investi- gated, representing five genera native to Brazil plus Syzygium, which is introduced and widely cultivat- ed. The following species were found to be consis- tently monoembryonic: Campomanesia sp. (as Myr- tus mucronata Cambess.), Psidium guajava, Psidium guineense (as Psidium araga Raddi), Eu- genia uniflora, Eugenia tomentosa Cambess., Eu- genia uvalha Cambess., Eugenia myrcianthes (as Myrcianthes edulis O. Berg), Eugenia brasiliensis Lam., Eugenia lucescens Nied. (as Phyllocalyx lu- schnathianus O. Berg), and Gomidesia reticulata O. Berg. Gurgel and Soubihe Sobrinho (1951) found a high percentage of polyembryony in Syzygium ma- laccensis 97%) and Syzygium cumini (76— 95%). One to eleven embryos were reported per seed. Repeated sampling of these species in suc- cessive years suggested the existence of a phenetic as well as a genetic component to the level of poly- embryony. A low to medium percentage of poly- embryony (33-7096), with one to five embryos re- ported per seed, was found in the following native Brazilian species: Plinia cauliflora (as Myrciaria cauliflora), Plinia trunciflora (O. Berg) Kausel (as 498 Annals of the Missouri Botanical Garden Myrciaria trunciflora O. Berg), and Plinia edulis (Vell.) Sobral (as Eugenia edulis Vell.), as well as a species cited as Myrciaria trunciflora, which Mat- tos (1989), based on the common name, claimed to be Myrciaria coronata Mattos. Sobral (1993) ex- cluded this species from Myrciaria in his recent revision of that genus and stated that it is a Plinia. Rotman (1982) registered consistent polyembryony in Guapurium peruvianum (possibly = Plinia) but registered no polyembryony in Plinia trunciflora. Thus all low to medium polyembryonic species may be members of genus Plinia. Traub (1939) had al- ready noted that polyembryony was usual in Plinia cauliflora (as Myrciaria cauliflora ) Florida. Observations on the origin of the embryos in polyembryonic seeds are few and disparate. Tiwary (1926) reported embryos originating from the egg cell (presumbly sexual), from the synergids and from the nucellus in S. cumini (as Eugenia jam- bolana). Narayanaswami and Roy (1960a) and Roy and Sahai (1962) also studied S. cumini, and the latter authors found that the fertilized egg degen- erates along with the synergids. The upper half of the nucellus becomes proliferative and produces cultivated in several embryos, one or more of which may survive, producing mono- or polyembryonic seeds in this species. Roy and Sahai (1962) also noted the oc- casional production of supernumerary embryo-sacs (by apospory or by the activity of more than one archesporial cell) but reported that ovules in which this occurs eventually degenerate. The multiple embryos in seeds of S. jambos are also of nucellar origin (van der Pijl, 1934), and Johnson (1936) ten- tatively suggested a nucellar origin for the embryos of Luma apiculata (as Eugenia hookeri). Gurgel and Soubihe Sobrinho (1951) confirmed that in S. jam- bos the sexual embryo often dies immediately after fertilization. However, they considered the adven- titious embryos to be derived from the inner integ- ument. Van der Pijl (1934) reported an integumen- tary origin for the multiple embryos of Syzygium malaccense (as E. malaccensis). REGENERATION FROM SEED Beardsell et al. (1993) highlighted the lack of published data on the regeneration of the Myrtoi- deae of the tropical forests. They suggested that germination requires breakdown of the pericarp by weathering or ingestion by animals including birds, and reported that some species of Syzygium ger- minate after fermentation of the fruits. Tidbury (1949) also advocated fermentation of fruits. When fruits of Syzygium aromaticum (Clove Tree) were left to ferment for about three days the seeds were easily hulled, and such hulled seed produced “somewhat better seedlings” than unhulled. Ger- mination was rapid, occurring within 12-14 days of sowing, and germination percentages were high, almost always above 90%. Hyland (1983) included seedling germination periods in his descriptions of some 50 Australian species of Syzygium. Germi- nation periods ranged from 10 days to almost 10 months (300 days) and there was considerable with- in-species variation in germination time. In the same paper, species of Acmena were cited as having even longer germination times, ranging from 24 days to well over 2 years (860 days), while values for Waterhousea tended to be lower at 10-60 days. Eugenia reinwardtiana (Bl.) DC. and Acmenosper- ma claviflorum (Roxb.) E. Kausel were reported as having germination periods of 30-50 and 40-90 days respectively. Ferreira (1982) studied germination of seeds of Psidium acutangulum, which had been treated with a 1% solution of a proprietary systemic fungicide after being removed from the fruit pulp. He found that germination occurred 30-100 days after sow- ing and that ca. 90% of seeds planted in loam (the most successful medium) germinated eventually. Lorenzi (1992) reported a similar range of germi- nation times among 16 native Brazilian species (in eight genera) of Myrtaceae studied by him. The most rapid germination (10 days after sowing) was observed in Eugenia pyriformis Cambess., while some seeds of Plinia edulis (as Marlierea edulis) took up to 100 days to germinate. Gurgel and Sou- bihe Sobrinho (1951) reported germination times between 18 days (for Psidium guajava) and 73 days (for Campomanesia sp. as Myrtus mucronata). Lev- els of germination varied considerably (40-80%). Lorenzi (1992) also reported seed viability peri- ods. These were generally short (as little as 2 weeks in Eugenia involucrata DC.), but species of Psidi- um and Acca had longer seed viabilities of several months. Viability of seeds of Psidium guajava may exceed one year. Blepharocalyx salicifolius seeds lost viability within two months (Matos, 1994). CONCLUSIONS As might be predicted from their morphology, the Myrtoideae emerge from this review as a rather uni- form group with respect to their overall pollination and dispersal syndromes. Most species diverge lit- tle from the apparently highly successful combi- nation of small, short-lived flowers offering pollen to attract bees as pollinators and fleshy fruits adapt- ed to endozoochory. However, it is possible to ob- Volume 83, Number 4 1996 Nic Lughadha & Proen 499 nga Reproductive Biology of Myrtoideae serve that more morphological diversification has arisen in characters linked to seed dispersal and seedling establishment than in those linked to pol- lination biology. For such a large group, the species exhibit an impressive uniformity with respect to general flower structure and pollen morphology. he rather narrow pollination spectrum is remi- niscent of other large groups with faithful bee pol- lination, e.g., Malpighiaceae (Anderson, 1979), Me- lastomataceae (Renner, 1989), and Solanaceae (Symon, 1979). Renner (1989) has suggested that such a lack of diversification may be the result of being “stuck on an adaptive peak.” However, the data available suggest that the Myrtoideae run the gamut of flower-visiting bee taxa and fully exploit differences in bee behavior, such as buzzing abili- ties and trap-lining, the more highly evolved Api- dae bee family being dominant. This is reflected in the considerable diversity in flower size, flowering phenology, and especially flowering strategies en- countered. The breeding systems of the Myrtoideae vary from dioecy and complete self-sterility (with late-acting variants) through to self-compatibility and apomixis, suggesting that the Myrtoideae have successfully capitalized upon their bee pollinators, and diversified their reproductive strategies accord- ingly, to establish themselves as important elements in many different life-zones and habitats. In this respect, the Myrtoideae are similar to the Leptos- permoideae, although this latter group has coevol- ved primarily with the most primitive flower-visiting bee family, the Colletidae, while simultaneously developing a secondary evolutionary line of bird pollination, which is very rare in the Myrtoideae. Certain specialized features of the floral biology of the Leptospermoideae are absent in the Myrtoideae, e.g., secondary pollen presentation, wet stigmas, and the presence of ovulodes in the ovary. Con- versely, buzz pollination is, as far as we know, lim- ited to the Myrtoideae. Within the Myrtoideae s.l. nectar-rewarding bird pollination systems, post- anthesis stylar extension, unitegmic ovules, and apomixis have only been registered in Syzygium (of the Acmena alliance) and may be restricted to this group (although the possibility of polyembryonic Plinia of the Myrtoideae s. str. also being apomictic cannot be discarded as yet). The endozoochorous fruits of the Myrtoideae seem to be adapted to many different classes of dispersers, and this characteristic is probably the most important ecological difference as compared to the largely wind-dispersed Leptospermoideae. The Myrtoideae show a great variety in size, color, texture, and seediness of fruits, with the seeds also varying significantly in size, seed coat structure, and embryo morphology. Pseudo-testa formation from the endocarp, testa formation from the outer integument only, oil glands in the embryo, and polyembryony are apparently restricted to the Myr- toideae. Our knowledge of the reproductive biology of the Myrtoideae is still far from complete, and the fol- lowing questions merit further investigation: Are in- florescence structure and flowering strategy corre- lated with habitat? Is flowering cued by abrupt fluctuations in humidity? How common is buzz pol- lination, and what are the adaptations and bees in- volved? Which birds are involved with bird polli- nation systems where petals are offered as a reward, why has this mechanism evolved, and how does it differ from nectar-rewarding bird pollination sys- tems? How common is cryptic dioecy and what are the selective pressures maintaining this apparently expensive breeding system? What are the main agents of fruit dispersal and, in particular, how im- portant is the role played by bats and other small mammals? Has co-evolution with parasites affected fruit and seed morphology? What is the nature of the mechanism which maintains preferential out- crossing and does more than one type of self-in- compatibility exist in the family? Is there any re- lationship between apomixis, polyembryony, and chromosome number? The answers to such ques- tions as these may have important and diverse im- plications. First, they may help to resolve some of the complex taxonomic problems in the Myrtoideae by helping to elucidate the functional significance of characters of taxonomic importance. Second, studies of breeding systems in particular should aid systematists, population geneticists, and ecologists in the interpretation of patterns of variation within and between populations and species. Third, this information may facilitate the genetic improvement of the several economically important species. Fi- nally, a deeper understanding of the reproductive biology of a group that is of great ecological im- portance in several endangered tropical ecosystems may represent an invaluable contribution toward their conservation. Literature Cited Absy, M. L. & W. E. Kerr. 1977. Algumas plantas visi- tadas para obtengao de pólen por operárias de Melipona seminigra merrillae em Manaus. Acta Amazon. 7: 309- 315. ——, ezerra & W. E. Kerr. 1980. Plantas nec- tartferas E por duas espécies de Melipona da Amazó Acta Amazon. 10: 271-2 Advani, RO 1981. Field observations on the ecology and biology of birds of economic importance in the southern 500 Annals of the Missouri Botanical Garden biome of Rajasthan (India). Z. Angew. Zool. 64: 385- 1982. Feeding, reds. and roosting behaviour of the fruit eating bat and damage to fruit crops in Ra- jasthan and Gujarat (India) Saeugetierk. Mitt. 30: 46— 48. Alves, M. A. 992. Dieta e táticas de forrageamento de Vola fasciata (Passeriformes: Emberizidae). Ararajuba 2 ius E Notes on Myrtaceae. VII. Myr- rir of French iud Africa. Acta Bot. Néerl. 7: 53-5 Pn Un nouvel Eugenia du Cameroun (Myr- bibas); pepe Sér. 2, 14: 481-483. Anderson, W. R. 1979. Floral conservatism in neotropical Malpighiaceae. Biotropica 11: 219-223. Armstrong, J. A. 1979. Biotic pollination mechanisms in the Australian flora—A review. New Zealand J. Bot. 17: 7-508. Aronne, G. & C. C. Wilcock. 1994. First evidence of i f the Mediter- A. F. Barbosa. 1972. Catálogo sistemáti dos pólens “= eid idein do Brasil meridional XV. Myrtaceae. Mem. Inst. Oswaldo Cruz 70: 467—495. Bawa, K. S. 1974, Se systems of tree species of a lowland tropical community. Evolution 28: 85-92. ——, 19 volution of dioecy in flowering plants. Annual Rev. Ecol Pg 11: 15-39. Beardsell, D. V., E. G. Williams & R. B. Knox. 1989. ecretory tissue of the flowers of Thryptomene ca- lycina (Lindl.) Stapf a ie J. Bot. 37 65—80. , S. P. O'Brien, E. G. Williams, R. B. Kno M. Calder. 1993. Pup a aa of ia yrtaceae. Austral. J. Bot. 4 Berg, O. 1857-1859. aa ES E P. von Martius editor), E Brasiliensis 14: 1—655 Bir, S. S., . Gill, T. S. Bedi & V. K. Singhal. 1980. ae status E 4 woody taxa of Garwhal Hi- malaya. Pp. 81-96 in P. K. Khosla (editor), Improve- ment of Forest Biomass. Indian Society of Tree Scien- | L. A. S. Johnson. 1979. d in the Myrtaceae: Evidence from inflorescence structure. Proc. Linn. Soc. New South Wales 102: 157-256. Bullock, S. H. 1985. Breeding systems in the flora of a vii E :iduous forest. Biotropica 17: 287-301 Burk, T. Behaviour ecology of mating in the Carib- bean ay ay Anastrepha vu Lo as Tephritidae) ew. Florida Entomol. 6 Butcher, P. A., J. C. Bell & d F Fani of genetic diversity and nature of the in Melaleuca alternifolia (Myrtaceae). Austral. 365-375. Camillo, E. & C. A. Garófalo. 1989. Analysis of the niche 1992. Patterns breeding system J. Bot. . de. mus Syste- matis Naturalis Regni Vegetahilia 3: 207 7-296. Paris. Carr , S. G. & D. J. ; The functional signifi- cance of syncarpy. Phytomorphology 11: 249-256. Caspar, B. B. & D. Wiens Fixed rates of random ovule abortion in Cu jon (Boraginaceae) and its possible relation to seed dispersal. Ecology 62: 866— 869. Cesar, H. L. 1980. Efeitos da TM. e corte sobre a vegetacáo de um campo sujo na Fazenda Água Limpa, Distrito Federal (M.Sc. Thesis, Unpublished). Univer- sidade de Brasflia, Brasflia. Chantaranothai, P. & J. A. N. Parnell. 1994. The breed- ing yesh E some Thai Syzygium species. Trop. Ecol. : 199 Chapman, T i 1964. Some aspects of dioecism in pi- mento n Ann. Bot. 28: 451—458. — new development in the agronomy of pd c. Quarterly (Trinidad) 2: 2-12. rome, F. H. J. & A. K. Irvine. 1986. “Two Bob Each Way": The pollination and breeding system of the Aus- tralian rain forest Em Syzygium cormiflorum (Myrta- ceae). pea 18: 125. Davis, G. L A oe of the Angio- sperms. dn Wiley & Sons, New York. — —. 1968. Floral morphology and the sg ts of tametopytes ka as lyptus melliodora A. Cunn Austral. J. Bot. 9-35. Dawson, J. W. T Flore de la Nouvelle-Calédonie et Dépendances. 18. ceae—Leptospermoideae. Mu- séum National d' egi Naturelle, EE Falcáo, M. de A do erreira, W. B. Chavez Flores & C. R. e. 1989. ed fenológicos e ecológicos do camu-camu (Myricaria dubia ( ura se Anais. Sociedade Brasileira de Fruticultura, Fortalez ment, T. C. dos Santos & R. de Souza. Crescimento e fenologia de Araçá- Pera (Psidium aaaneulum DC.) Acta Amazon. 22: 5—293. — ———, V. B. Chavez- jin F. S. A. do N. Ferreira, C. R. Clement, M. J. B s, J. M. C. de Brito & T. C. T. dos Santos. 1988. Peine fenológicos e dae dioe do “Aragá- boi” (Eugenia stipitata McVaugh) na Ama- zónia Central. I . Plantas juvenis. Acta Amazon. 18: 27- Ferreira, L. V. & J. R. de Merona. 1987. Aspectos sobre a ecologia da família Myrtaceae em uma po de ter- rme na Amazônia Central. Pp. 290 in Congresso Nacional de a 38, Resumos. Lüvenidade de "1982. Observação da germinação e araça- p (Psidium acutangulum D.C.). Acta Amazon. 12: 503-507. Ford, H. A., D. C. Paton & N. Forde. 1979. Birds as pollinators of Australian plants. New Zealand J. Bot. 17: 509-519. Frankie, G. W., W. A. Haber, P. A. Opler & K. S. Bawa. 1983. C ingacietistos and organization of the large bee Snares o in ag Costa dn dry forest. Pp. 441—447 . E. Jone tle (editors), Hand- book of Experimental tec Biology. Van Nostrano Reinhold, New York. Free, J. B. 1993, Insect Pollination of Crops. Academic Press, London Funmilayo, O. . Ecology of the straw-coloured bat (Eidolon helvum) in Nigeria. Rev. Zool. Afr. 93: 589— Gentry, A. H. 1974. Flowering phenology and diversity 1989. Genetic vari- ed in trees of Social Turcae id mk. e Pl. Sci. 11: cn T. J. 1980. Beolosio d constraints on the evo- Volume 83, Number 4 1996 Nic Lughadha & Proen Reproductive Biology of Myrtoideae 501 lution of breeding systems in seed plants: dioecy and wage > gymnosperms. Evolution 34: 959-972 =, utcrossing versus eco ological constraints in the cuis of dioecy. Amer. Naturalist 119: 849— 65. Griffin, A. R., G. F. Moran & Y. J. Fripp. 1987. Prefer- ential Pipes: in Eucalyptus regnans F. Muell. Aus- tral. J. Bot. 35: 465-475. Grifo, F. T Pes A revision of Myrcianthes Berg (Myr- taceae). Ph.D.Thesis. University of Cornell, Ithaca. [Un- published, available from UMI microfilms, Ann Arbor, Michigan, U.S.A.] M. Ramalho, A. Kleinert-Giovannini & V. L. Imperatriz- -Fonseca. 1988. Exploragáo de recursos florais por colónias de Melipona quadrifasciata (Apidae, ICI Revista Brasil. Biol. 48: 2 Gurgel, J. T. A. & J. Soubihe Sobrinho. 1951. Poliem- brionia em Mirtáceas. Bragantia 11: 141-163. Haber, W. A. € K. S. Bawa. 1984. Evolution of dioecy in Saurauia (Dilleniaceae). Ann. Missouri Bot. Gard. 71: 289-293. Hartley, T. G. € L. A. Craven. 1977. A revision of the Papuasian species of Acmena (Myrtaceae). J. Arnold Ar- bor. 58: 325-342. i rare M. R. 1949. The genus Eugenia er Malaya. ee Bull. Singapore 12: 1-293 2, C. M. 1984. Selective pressures on fruit seedi- ness: Differential predation of d abi on the fruits of Berberis hispanica. Oikos 42: Heslop-Harrison, Y. & K. R. P 1977. The recep- tive surface of the Angiosperm stigma. Ann. Bot. 41: 1233-12 Hopper, S. D. 1980. Pollination of the rain-forest tree Syzygium tierneyanum (Myrtaceae) at Kuranda, North- ern Queensland. Austral. J. Bot. 28: 223-237. Hyland, B. P. M. 1983. A revision of Syzygium and allied genera üortacese) in Australia. Austral. J. Bot., Suppl. Ser., 9: Janson, C. H. 983. Adaptation of fruit morphology to dispersal agents in a neotropical forest. Science 219: 187-189. Johnson, A. “i 1936. Polyembryony in Eugenia hookeri. Amer. J. Bot. 23: Johnson, L. S & B. G. Briggs. 1984. Myrtales and Myrtaceae—A phylogenetic anih Ann. Missouri Bot. Gard. 71: 7 à Kawasaki, M. L. 1984. A oe Myrtaceae na Serra do Cipó, Minas Gerais, Brasil. M.Sc. Thesis, Universidade de Sáo Paulo, Sáo Paulo. "Unpublished. ] 1 A. J. Lack. 1985. Pollination in a cryp- tically oe plant, Decaspermum parviflorum (Lam.) A. J. Scott (Myrtaceae), by pollen-collecting bees in Su- . J. Li — ———, D. Eisikowitch, J. D. Ambrose & J. R. Kemp. 990. Cryptic dioecy and insect pollination in Rosa ti y pong cn a rare plant of Carolinian Canada. Biol. J. L ae me H. "1893. Enumeratio Myrtacearum brasilien- arming (editor), Symbolarum ad Floram Brasiliae Centralis Cognoscendam 39: 1-200. Copen- Knuth, P 1906. Handbook of Flower Pollination (Enghien translation from German]. Clarendon jion s, Oxford Lack, A. J. & P. G. Kevan. he reproductive biology of a canopy tree, ada jia (Myrta- ceae), in a rain forest in Sulawesi, Indonesia. Biotropica 16: 31-36. Landrum, L. R. 1981. A monograph of the genus Myr- ceugenia (Myrtaceae). Fl. Neotrop. Monogr. 29. New York Botanical Garden, New York. 1 t of the fruits and seeds of Campomanesia (Myrtaceae). een 34: 220-224. —. 1 Campomanesia, wR Blepharocalyx, Legrandia, Acca, Myrrhinium, and L ma (Myrtaceae). Fl. Neotrop. Monogr. 45. New York Botanical Garden, New York. 1988. The Myrtle ree (Myrtaceae) in Chile. Proc. ‘California Acad. Sci. 4 7-317. . Stevenson. 1986. Varia bility of embryos in the subtribe Myrtinae (Myrtaceae). Syst. Bot. 11: 155— 162. Lieu, J. & T. S. Melhem. 1973. Palinologia em Myrtaceae. Hoehnea 3: 1 Lloyd, D. G. 1982. Selection of combined versus separate sexes in seed plants. Amer. Naturalist 120: 571-585. C. J. Webb. 1986. The avoidance of interfer- ence between the presentation of pollen and —" in Angiosperms. I. Dichogamy. New Zealand J. B 135-162 Lorenzi, H. 1992. Árvores brasileiras: Manual de iden- tificacáo e cultivo de plantas arbóreas nativas do Brasil. Editora Plantarum, Nova Odessa. ciae Studies of tropical American a 4: 31-51. Malavasi, A. & J. S. presi 1980. Biology of the fruit flies (Diptera, Tephritidae) 2: Indices of infestation in different hosts and localities. Revista Brasil. Biol. 40: R. B. 1994. Efeito do fogo sobre os regener- es de Blepharocalyx salicifolius (H. B. € K.) Berg e em cerrado aberto, Brasília, D. F. M.Sc. esis. apre of Brasília, Sigo IUe ] Mattos, J. R. 1989. Myrtaceae do Rio Grande do Sul. Convénio Secretaria da aaa ta T EMBRAPA, Porto Alegre. Mauritzon, J. 1939. om to ha ana Rosales nd und E 2) 35. Aig, a ud in flowering plants. Trends Eco 20-32 McGregor, S. E. 1976. In wack pollination ig A RP plants. U.S.D.A. Agricultural Handbook no. 496. Unit- ed States Department of Agriculture, Washington, D.C. McVaugh, R. 1958. Myrtaceae. In: Flora of Peru. Field Mus. T Hist., Bot. Ser, 13: 567-818. . The genera of American Myrtaceae—An interim E Taxon 17: 354—418. ——. . Myrtaceae. /n: The Botany of the Guayana Highland—Part VIII. Mem. New York Bot. Gard. 18: 55-286. Mehra, P. N. 1976. Cytology v Himalayan Hardwoods. Sree xA. Press, Calcut Merrill, E. & L. M. Perry. "1938. The Myrtaceae of China. J. Ail Arbor. 19: 191-247. Michener, C. D. 1965. A classification of the bees of the Australian s Pacific regions. Bull. Amer. Mus. Nat Hist. Et 1-362. 979. or of the bees. Ann. Missouri Bot. ea 66: 271-3 Moncur, M. W. 1988. d Development of Tropical and Subtropical Fruit and Melbourne. & D. J. Boland. 1989. Floral morphology of Eu- calyptus melliodora A. Cunn. ex Schau. and compari- 502 Annals of the Missouri Botanical Garden sons P other eucalypt species. Austral. J. Bot. 37: 5-13 Moran, G. E & J. C. Bell. 1983. Eucalyptus. Pp. 432- 441 in S. D. Tanksley & T. J. Orton (editors), Isozymes in Plant Breeding, Part 8. Elsevier Science Publishers, Amsterdam. Morellato, L. P., R. R. Rodrigues, H. F. Leitáo Filho & C. A. Joly. 1989. Estudo comparativo da fenologia de es- pécies arbóreas de floresta de altitude e floresta Mir. semidecídua na Serra do Japf. Revista Brasil. Bot. 5-98. Morgante, J. S Malavasi. 1981. Adult and larval BA ae of Anastrepha fraterculus and its relationship to host availability, Environm. Entomol. 10: 5 Morley, T. . Memecyleae (Melastomataceae). Fl. Neotrop. Monogr. 15. New York Botanical Garden, New York. Motta Júnior, J. C., J. A. ad & S. A. Talamoni. 1994. Notes on ed (Dusicyon thous) eating for seed dispersal and ies habits in poA cia Brazil. Mam- malia 58: 156- Narayanaswami, : ys S. K. Roy. 1960a. Embryo sac de- velopment and polyembryony in Syzygium cuminii. Bot. Not. [1960]: 273— E . 1960b ped of the genus Psidium. J. Indian Bot Soc. P Nic Lughadha, E. 1994. VE on the Myrtaceae of the Pico das Almas, Bahia, Brazil. Kew Bull 49: 321—329. Niedenzu, F. 1893. Myrtaceae. Pp. 57-105 in A. Engler K. Prantl (editors), Die natürlichen Pflanzenfamilien, l. 3, Abt. 7. Leipzig. O'Brien, S. P. & D. M. Calder. 1993. Reproductive bi- ology and floral phenologies of the sympatric species d myrsinoides and L. continentale (Myrta- ceae). Austral. J. Bot. 41: 527—539. m M 1993. Utilização de frutos por aves em a área de cerrado do Distrito Federal. M.Sc. Thes Universidade de Brasília, Brasília. [Unpublished. ] Patel, V. C., J. J. Skvarla Raven. 1984. Pollen charac dam in relation to the pia of Myrtales. esis. nn. Missour P Gard. 71: 858-96 Peters, € asquez. 1986/87. Eatudioė ecologicos de camu u (Myrciaria dubia). Producción de frutos en oer naturales. Acta Amazon. 16/17: 161— Petit, L -A. 1908. Recherches sur la structure anato- mique du fruit et de la graine des Myrtacées. Thesis. an de Paris, Paris. Lucien Dec e Lons-le Phillipe. M. A. . H. D. Brown. 1977. Mating T and brida in eens pauciflora. Austral. J. Biol Sci. 30: -344 a van n 1934. Über die Polyembryonie bei £u- genta. Recueil Trav. Bot. Néerl. 31: 113-187. Pike, K. M. 1956. Pollen morphology of i eae $4 the dw jus Pacific area. Austral. J. Bot. 4: 13-5 Pirani, J. . Cortopassi- Ramee p es e ns em ga Paulo. Editora Universidade de Sao Paulo, Sao Paulo. Primack, R. B. 1980. Variation in the phenology be nat- ural populations of montane shrubs in New Zealand. J. Ecol. 68: 849-862. 1985. tán of individual flowers. Annual Wi Bool. Syst. 16: 15-37. Bu C. 1990. A revision of Pd a Berg (Myr- taceae). Edinburgh J. Bot. 47: 239-27 nensis (Ac s 'eae). Bot. ——— 198 . 1991. The Reproductive Biology and Taxonon ny of the Myrtaceae of the Distrito Federal (Brazil). Ph.D. Thesis. University of St. Andrews, St. Ni el [Un- published. | 1992. Buzz- polliuation-— Older: and more wide- spread than we think? J. Trop. Ecol. 8: 115-120. — —— & P. E. Gibbs. 1994, do biology of eight sympatric Myrtaceae from central Brazil. New Phytol. 126: 343-354. unt, W., S. "Bla ckmore, S. Nilsson & A. Thomas. 1994. Glossary of Pollen and Spore oia, LPP Contributions Series, 1. Utrecht. Purseglove, J. W. 1968. Tropical Crops, Vol. ledons. yla London. Renner, S. S. 1989. A survey of reproductive biology in neotropical Melastomataceae and Memecylaceae. Ann. Missouri Bot. Gard. 76: 496-518 Rotman, A. D. 1982. Los géneros cia Hexach- lamys, Myrciaria, Paramyrciaria, Plinia y Parita gena en la o Argentina. Darwiniana 24: 157-18 Roubik, D. W. 1989. Ecology and Natural dn of Tropical Bees Cambridge Univ. Press, Cambridge. oy. S. K. 1953. Embryology of Eugenia jambos. Curr. Sci. K bid 621. . 1960. as of Eugenia malaccensis. Curr. Sel. P vit 196 pore s eae fruticosa L. Proc. Natl. Pies Su, India 31: . 1962. P of ae nia myrtifolia Roxb. Sei. & Cult. 28: 378. 1, Dicoty- R. Se e embryo sac and embryo of Syzygium o J. Indian Bot. Soc. 41: 45-51. Tad T. Z. & M. T. K. Arroyo. ogy of a secondary deciduous forest. 221-230. Rye, B. L. 1979. Chromosome number variation in the yrtaceae and its taxonomic implications. Austral. J Bot. 27: 547-73. |. J. James. 1992. The relationship between dysploidy and reproductive capacity in Myrtaceae. Aus- tral. J. i 1. 40: is Saka, J. 1 . Msonthi. 1994, Nutritional value of edible fruits af Nen wild trees in Malawi. For- est Ecol. Managem. 64: 245-248. 1978. Reproductive ecol- Biotropica 10: Sanaiotti, T. M Magnusson. 1995. Effects of annual fires on the paduko of fleshy fruits eaten by birds in a Brazilian Amazonian Savanna. J. Trop. Ecol. 11: 53—65 Schmid, R. 1972a. A Me de of the Eugenia-Syzygium cai (Myrtaceae). Amer. J. Bot. 59: 423-436. ———. 1972b. Floral pees of Myrtaceae I. gium. Tel Jahrb. Syst. 92: 433—489. ——.. 1978. Reproductive anatomy of Actinidia chi- Jahrb. Syst. 100: 149-195. ). Comparative anatomy and morphology of Peilowyion and Heteropyxis, and the subfamilial and tribal Er ‘ation of Myrtaceae. Taxon 29: 559-595 84[ 1985). Reproductive anatomy and mor- SE i Myrtales in ara lo systematics. Ann. Missouri Bot. Gard. 71: 832-8 Schroeder, C. A. 1947. da eure of the "eijoa. ps Amer. Soc. Hort. Sci. 61-16 Semir, J. lo. Encontro de seu e em (Oral Communi cation). Ilha Solteira, Brazil. Singh. R. & O. P. Sehgal. 1968. Studies on the blossom- biology of Psidium guajava L. (guava). II. Pollen stud- Syzy- Volume 83, Number 4 1996 Nic Lughadha 4 Proenca 503 Ç Reproductive Biology of Myrtoideae ies, stigmatal receptivity, pollination and fruit-set. In- dian J. Hort. 25: 52- Singhal, V. K. & B. S. Gill. 1984. SOCGI plant chro- mosome reports II. J. Cytol. Genet. 19: 115-117. & S. S. Bir. 1984. Cytology of d coda species (Polypetalae). Proc. Indian Sci. Cong Assoc. 71: 143-144 ; & ———. 1985. Cytology of woody species. Proc. Indian Acad. Sci., Sec. B, 94: 607-617. Snow, D. W. 1981. Tropical frugivorous birds and their food plants: world survey. Biotropica 13: 1-14. Sobral, M. 1993. Sinopse de Myrciaria (Myrtaceae). Na- l. paea Sibrevila, C. & M. T. K. Arroyo. 1982. Breeding systems in a montane cloud forest in Venezuela. Pl. Syst. Evol. : 19-37. Soubihe Sobrinho, J. S. & J. T. A. Gurgel. 1962. Amount of natural cross-pollination in guavas. Bragantia 21: 15— 20 Start, A. N. & A. G. Marshall. 1976. Nectarivorous bats as dpud of trees in West Malaysia. Pp 141—150 in J. Bur . T. Styles (editors), Tropical Trees. Vari- ation, jae and Conservation. Academic Press, London. Stephenson, A. C. & R. I. Bertin. 1983. Male competi- tion, female choice and sexual selection in plants. Pp. 109-149 in L. Real (editor), Pollination Biology. Aca- demic Press, London Stratton, D. A. 1989. Longevity of individual flowers in a Costa Rican cloud forest. Biotropica 21: 308-318. Suarez, P. & C. Esquivel. 1987. Fenologia del Guisaro (Psidium guineense Swartz) en Barva de Heredia, Costa Rica. Brenesia 28: 97-105 Symon, D. E. 1979. Sex forms in Solanum bringe E the role of pollen collecting insects. Pp. 3 . G. Hawkes, R. N. Lester & A. D. M. (edi- lod The Biology pi Taxonomy of the Solanaceae. Ac- ademic Press, Londo Tan, K. & S. Lee. 1982. Species diversity and abundance of Dacus (Diptera, Tephritidae) in five ecosystems of EN West Malaysia. Bull. Entomol. Res. 72: 709— Tidbury, G. E. 1949. The Clove Tree. Crosby Lockwood, London Titow ia Tichisho ow, N. 1956. Nümero de cromosomas de algunas plantas chilenas. Bol. Soc. Biol. Concepción 31: 145-1 47. Tiwary, N. K. 1926. On the occurrence of polyembryony in the genus Eugenia. J. Indian " Soc. 5: 124-136. Tobe, H. & P. H. Raven. 1983. mb ryologic al analysis of the Myrtales: Its definition me characteristics. Ann. Missouri Bot. Gard. 70: 71-94. Traub, H. P. 1939. Polyembryony in Myrciaria cauliflora. Bot. Gaz. 101: 233-234. Vit, P. & G. R. D'Albore. 1994. Melissopalynology for stingless bees (Apidae: Meliponinae) from Venezuela. J. a Vogel, S., T. B. Westerca 1984. Orni- thophily on the Canary e SS "PL Syst. Evol. 146: 225-248. Webb, C. J. & D. G. Lloyd. 1986. The avoidance of in- terference between the presentation of pollen and stig- mas in Angiosperms. II. Herkogamy. New Zealand J. Bot. 24: 163-178. Werth, E. 1901. Blutenbiologische Fragmente aus Ostaf- rika. Rune Verhandlungen 42: 222-260. Wheelwright, N. T., W. A. Haber, K. C. Murray & C. Guin- don. 1984. Tonio fruit- -eating birds and their food lants: A survey of a Costa Rican lower montane hab- itat. Biotropica 16: 173-192. Wiens, D., C. L. Calvin, A. Wilson, C. I. Davern, D Reproductive success, spontaneous embryo abortion, and genetic load in flow- ering pe Oecologia 71: 501-509. Wyk, A. E. v 980. A note on the seed morphology of the genus "uita L. (Myrtaceae) in Southern Africa. J. S. African Bot. 46: 115-119. & R. Botha. 1984. The genus Eugenia (M yrta- ceae) in southern Africa: Ontogeny and taxonomic value of the seed. S. African J. Bot —80. Dedekind. 198. 5. The genus Eugenia in Southern Africa: Morphology and taxonomic value of pollen. S. African J. Bot. 51: 371—378. & T. K. Lowrey. 1988. Studies on the reproduc- tive biology of — a . (Myrtaceae) in Southern Af- rica. Monogr. Syst. Bot. Missouri Bot. Gard. 25: 279- 293. THE PHYLOGENETIC Neil Snow: UTILITY OF LEMMATAL MICROMORPHOLOGY IN LEPTOCHLOA S.L. AND RELATED GENERA IN SUBTRIBE ELEUSININAE (POACEAE, CHLORIDOIDEAE, ERAGROSTIDEAE)! ABSTRACT Mic ‘romorphological features of the lemma were investigated in the subtribe Eleusininae (Poaceae) using sc anning electron microscopy. Ninety-two taxa were studied, which included 48 genera and all currently recognized species of Leptochloa s.l. All species “of ies hloa and a majority of genera in Eleusininae have cork cells, but silica cells are mostly absent in Leptochloa and most related genera in Eleusininae. Enneapogonoid-type microhairs are reported for n macrohair occurring in three species of Coelachyrum is described for the first time and probably represents a syn- apomorphy in that genus. The analysis of m suggests that micromorphological dere vary little within a genus, and thus have the potential to serve as phylogenetic markers at the generic level. ainty concerning the Vica relationships of papillae, Vern prickles, and macrohairs is discussed in light of this study and previous literat RESUMEN Las caracterfsticas mic romorfológicas de la lema en la subtribu Eleusininae (Poaceae) se examinaron utilizando microscopía electrónica de barrido. Se estudiaron 92 taxa, 48 géneros, y todas las especies de Leptochloa hasta ahora reconocidas. Todas las especies de Leptochloa y la avarl de los géneros de Eleusininae tienen Se beri 'adas, pero las células silicfferas están ausentes en Leptochloa y en la mayoría de los taxa de Eleusininae. Se reporta por segunda vez en Eleusininae la presencia de micropelos bicelulares de tipo enneapogonoid para Cladoraphis jode is y Pies wiseana. Aun oque los. micropelos bicelulares se presentan en casi todos los taxa, variac iones en e dee dum por primera vez, y mund represente una sinap moria ph cierto nivel en ese Mein, El aan de pies hloa indica que las características micromorfológicas varían poco, y por tanto tienen valor para estudios filogenéticos su- praespecíficos. La dificultad sobre las relaciones homólogas de papilas, ganchos, aguijones, y macrotricomas se discute con base a lo hallado en este estudio y con datos anteriores de la literatura. Cladistic estimations of phylogenetic history with easily defined character states. This tendency may morphological characters tend to favor the use of reflect the potential pitfalls discussed by Chappill distinct qualitative characters, or those possessing — (1989) and Stevens (1991) in the uncritical use of ! Mike Veith facilitated the work with eee at the Electron Microscopy Laboratory at Washington PT I the curators of B, BM, BRI, CANB, , PRE, and US for sending loans and for granting permission to samp directly from herbarium specimens. S. Hatc ns seni an SERI version of Morden's dissertation, and helpful discussions ANN. MISSOURI Bor. GARD. 83: 504—529. 1996. Volume 83, Number 4 Snow 505 Lemmatal Micromorphology quantitative characters (but see Thiele, 1993). Used alone, and due to the perception of high levels of homoplasy, characters of gross morphology have been considered inadequate as phylogenetic mark- ers in the grass family (Thomasson, 1978; Hilu & Wright, 1982; Kellogg & Campbell, 1987; Davis & Soreng, 1993; Clark et al., 1995). Not surprisingly, revisionary studies of grasses are turning to micro- morphological characters for additional phyloge- netic data. Epidermal micromorphological characters of grasses have systematic value between the ranks of subfamily and species (Prat, 1932; Tateoka et al., 1959; Metcalfe, 1960; Ellis, 1979). Descriptive studies of micromorphological features in grasses have focused on the surfaces of the leaf (Prat, 1932; summary in Ellis, 1976; see also Morden & Hatch, 1987; Davila & Clark, 1990; Peterson & Annable, 1990; Barker, 1993; Scholz, 1993; Chen et al., 1993), glumes (Lucas, 1979; Molina, 1993), the lemma (Hsu, 1965; Thomasson, 1986; Peterson, 1989; Soderstrom & Zuloaga, 1989; Kellogg, 1990; Zuloaga & Judziewicz, 1991; Valdés-Reyna & Hatch, 1991; Naredo et al., 1993; Ball et al., 1993), and palea (summary in Consaul & Aiken, 1993: 1651). The phylogenetic application of micromor- phological characters has been more limited (Pe- terson & Annable, 1992; Barker, 1993; Visser & Spies, 1994; Guala, 1995). AN OVERVIEW OF MICROMORPHOLOGICAL CHARACTERS Several categories of micromorphological char- acters have been recognized. Short cells (which in- clude cork cells and silica cells), long cells, bicel- lular microhairs, papillae, hooks, prickles, and macrohairs are all examples of micromorphological characters (Metcalfe, 1960; Ellis, 1979). (See Ellis, 1979, for a summary of alternative terms for hooks and prickles.) Short cells and long cells are readily distin- guished on the basis of relative size, and (when both are present) may constitute a synapomorphy uniting Joinvilleaceae and Poaceae (see also Doyle et al., 1992; Kellogg & Linder, 1995). The differ- ences between short cells and long cells have been documented thoroughly (Kaufman et al., 1969; Kaufman et al., 1970) and are evident in grass fos- sils of Miocene age (Thomasson, 1978, 1984). The elongation of long cells follows the initial differ- entiation of long cells and short cells (Kaufman et al., 1969; Kaufman et al., 1970). In addition to their extended length, long cells (unlike short cells) frequently have sinuous margins (Metcalfe, 1960; Clifford & Watson, 1977; Ellis, 1979; Chen et al., 1993). However, the distinction between long and short cells is not always absolute. Thomasson (1978: fig. 1d) illustrated long cells, identifiable by the sinuous margins, having similar dimensions to short cells for a fossilized species of the genus Nas- sella (tribe Stipeae). The two types of short cells recognized are cork cells (sometimes called suberin cells), which accumulate suberin (Kaufman et al., 1970), and silica cells, which accumulate silica into readily observed silica bodies. Microhairs are bicellular structures that require high magnification for detection (Tateoka et al., 1959). (For exceptions to the bicellular condition see Dahlgren et al., 1985; Renvoize, 1985; and Zu- loaga et al., 1989.) They are present throughout the family except in subfamily Pooideae (Johnston & Watson, 1976). Microhairs have been classified as “chloridoid,” “panicoid,” or “enneapogonoid,” based on typological variants (Tateoka et al., 1959; Amarasinghe & Watson, 1988; Watson & Dallwitz, 1992). Only rarely is their ontological status as a distinct character uncertain. For example, as dis- cussed by Barker (1993), the “long slender papil- lae” (Clayton & Renvoize, 1986: 165 several genera in Arundinoideae, given that they are reported to occasionally have the remains of a small apical cell (Renvoize, 1986: 328), may be interpretable as microhairs. Papillae are short, undifferentiated processes that arise from the outer cell wall. Hooks have been considered processes having a rounded base and an apex that is at least slightly pointed (Ellis, 1979), but are not recognized as a category in this study (see Discussion). Prickles are basally swollen processes having short, sharp api- ces that typically point toward the apex of the struc- ture (leaves, glumes, lemmas, or paleas). common to were provided by T. Filgueiras, D. Nicolson, G. L. Stebbins, and F. Zuloaga. The careful reviews of J. T. Columbus i a manuscript significantly. F. Lorea assisted with preparation of the abstract in Spanish. d from: Missouri Botanical Garden (Andrew W. Mellon Foundation); Wash- is). Biology, P.O. Box 1137, St. Louis, Missouri 63130, U.S.A., and Missouri ingt Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166, U.S.A 506 Annals of the Missouri Botanical Garden Macrohairs are generally unicellular structures (but see Kabuye & Wood, 1969) visible to the na- ked eye. Among lemmatal micromorphological structures, only microhairs have not been reported to accu- mulate silica. The presence of silica has been re- ported for cork cells (Kaufman et al., 1972); long cells (Ball et al., 1993); papillae (Terrell & Wergin, 1981; Consaul & Aiken, 1993); prickles (Soni et al., 1970; Kaufman et al., 1972 (as "trichomes"); Terrell & Wergin, 1981; Valdés-Reyna & Hatch, 1991; Ball et al., 1993); and macrohairs (Consaul & Aiken, 1993). (Stomata occasionally occur on lemmas. They are sufficiently infrequent in the taxa examined thus far to be considered an abnormal constituent of the lemma. In a study of the leaf of Avena, Kaufman et al. (1972) were unable to find silica in stomatal cells.) OVERVIEW OF LEMMA MICROMORPHLOGY The lemma is the lower of two bracts that (usu- ally) subtend each flower in grasses. The upper bract (the palea) is not homologous to the lemma (Clifford, 1987). The lemma is a transformational serial homology (de Pinna, 1991) of the leaf by vir- tue of its “ontogenetic individualization” (sensu Wagner, 1989a, b) from the leaf (Philipson, 1934; an, 1973; Kellogg, 1990). Its transformational derivation from the leaf is supported by rare ata- visms in mutants to a leaflike progenitor having sheath, ligule, and blade (Philipson, 1934), its dim- inuitive size relative to regular leaves, its restricted occurrence in the inflorescence, the occasional presence of lemmatal stomata (common to the leaf), and the frequent infraspecific differences of micro- morphological features between the leaf and lemma (Thomasson, 1978: 977; Terrell & Wergin, 1981: 706; Snow, unpublished). The most extensive work on lemma micromor- phology in Eleusininae was that of Valdés-Reyna and Hatch (1991). Their survey in Eragrostideae of 27 species (representing 30 genera) suggested such characters might be useful phylogenetic markers in the subtribe Eleusininae. This study extended their generic survey of micromorphology (Valdés-Reyna & Hatch, 1991) to include genera of Eleusininae not previously examined, including a few other gen- era in Chloridoideae, and to characterize these fea- tures for all species in Leptochloa. AN OVERVIEW OF THE SYSTEMATICS OF CHLORIDOIDEAE The grasses are composed of at least four mono- phyletic subfamilies and contain a number of tribes with gia phylogenetic affinity (Dahlgren et al., 1985; Kellogg € Campbell, 1987; Doyle et al., 1992; Davis & Soreng, 1993; Barker et al., 1995; Clark et al., 1995; Kellogg & Linder, 1995). The subfamily Chloridoideae Rouy has about 1300 spe- cies, which tend to occupy dry, warmer climates (Hartley & Slater, 1960). (Taxonomic concepts fol- low Clayton & Renvoize, 1986.) The largest tribe, Eragrostideae Stapf, has about 80 genera and 1000 species. Within Eragrostideae, the most diverse subtribe is Eleusininae Dumort, and only the high- est latitudes are excluded from the overall range of its approximately 55 genera. Relatively few genera in Eleusininae occur outside of a single continent —~ e.g., Coelachyrum, Eragrostis, Leptochloa, Tricho- neura, disregarding minor exceptions and a few cosmopolitan weeds). Excluding Eragrostis, with its approximately 350 species (Clayton & Renvoize, 1986; Van den Borre & Watson, 1994), most of the approximately 600 species in the subtribe occur in relatively small genera (10 species or fewer). Lep- tochloa P. Beauv. is a nearly worldwide genus of about 35 species, although many continue to seg- regate the genus Diplachne (McNeill, 1979; Phil- lips, 1982; Watson & Dallwitz, 1992; Simon, 1993; Nicora, 1995). Recent intergeneric studies in Eleusininae dif- fered in the taxa, characters, and analytical meth- odologies, which hampers meaningful comparisons between the results (Phillips, 1982; Hilu & Wright, 1982; Campbell, 1985; Hilu & Esen, 1993; Duvall et al., 1994). The position of Leptochloa in the ge- neric arrangement of Clayton and Renvoize (1986: 192) suggests three non-mutually exclusive possi- bilities: (1) that Leptochloa has many close generic relatives, (2) that Leptochloa is evolutionarily basal to related genera, and (3) that Leptochloa may be paraphyletic as currently recognized. The mono- phyly of most chloridoid groups, including Lepto- chloa, remains largely untested. The purpose of this paper is to create a more extensive data set for inferring phylogenetic rela- tionships in Leptochloa and Chloridoideae (Snow, in prep.). In addition, terminological inconsisten- cies throughout the literature, coupled with the re- sults herein, provide a basis to question the onto- logical status of several characters, and suggest the need for a critical reevaluation of the characters. MATERIAL AND METHODS The adaxial surface of the lemma was studied for 92 taxa using scanning electron microscopy (SEM) at the Electron Microscopy Laboratory at Washing- ton University (Appendix 1). Samples were re- Volume 83, Number 4 1996 Snow 507 Lemmatal Micromorphology moved directly from herbarium specimens. To as- sure the use of semaphorants (Wiley, 1981: 119), I generally sampled lemmas from spikelets that con- tained mature caryopses. This precaution was im- portant, because spikelets in early stages of ontog- eny may be present even when the inflorescence is well exserted from the leaf sheath. Three specimens of most taxa were observed. Some samples were sonicated in xylene for 30 minutes to remove the epicuticular waxes that can obscure surface fea- tures, but this treatment was not always effective. After being placed on aluminum stubs the speci- mens were coated with gold using a Polaron E5000 sputter coater. Samples were observed with 0° tilt at 20 kV on a Hitachi S-450 SEM and photo- graphed using Polaroid 55 positive-negative film. Except for Figures 10 and 42, all photomicrographs have the apex of the lemma directed toward the right. Attempts to standardize the scale of photo- micrographs proved uon since various magnifications were nec The format differs iei Valdés- Reyna and Hatch (1991) in that microhairs and macrohairs are in- cluded, my interpretation of papillae on short cells is different (see Results), and the degree of undu- lation of the long cells with papillae was not re- corded. Valdés-Reyna and Hatch (1991) discussed their findings based on four variations of cork cell occurrence: cork cells adjacent to silica cells; cork cells not adjacent to silica cells; cork cells papil- late; or cork cells not observed. This study merely recorded cork cells as a binary character (present or absent). I will refer to taxa in Eleusininae be- sides Leptochloa as “Other Eleusininae Taxa” (OET), whereas other taxa in Eragrostideae (fide Clayton & Renvoize, 1986) will be designated “Other Eragrostoid Genera” (OERG). This study focused on the presence or absence of characters. No attempt was made to measure quantitative variation of the characters, but distri- bution and frequency on the lemma is sometimes discussed. RESULTS A summary of lemmatal microcharacters ob- served is presented in Table 1. For comparison, the results of Valdés-Reyna and Hatch (1991) are in- tercalated therein. Unless noted otherwise these re- sults accord with those of Valdés-Reyna and Hatch (1991) for taxa examined in both studies. Figures are presented at the end of the text, arranged in alphabetical order by genus name. Figures are not presented for all taxa and characters observed or discussed, but virtually every character recorded (except those with question marks, Table 1) was verified with a photomicrograph. Cork cells. Cork cells were abundant in the in- tercostal zones in all species of Leptochloa, and al- ternated with long cells. As was true of all micro- morphological characters studied, their distribution could vary significantly on a single lemma. There was a general tendency for cork cells to decrease in frequency near the apex. ong OET, cork cells were absent in Desmos- tachya (Fig. 14), Ectrosia gulliveri (Fig. 17), Ectro- siopsis, Odyssea mucronata (Fig. 60), and Psam- magrostis (Fig. 65). Unlike its congener, Odyssea paucinervis has cork cells. The cork cells of Apo- chiton (Fig. 2), Halopyrum (Fig. 23), and Neyraudia (Fig. 58) were not noticeably darkened, as is com- mon for mature cork cells under SEM observation (Valdés-Reyna & Hatch, 1991). However, since their size and location suggested cork cells, they were recorded as present (Table 1). Cork cells were present in all OERG except Spartina (not shown). With the exception of a sparse distribution in Tragus (not shown), cork cells were common in OERG. Some taxa had short cells resembling cork cells in size, and from which other structures could be seen developing, such as bicellular microhairs (e.g., Bewsia, Fig. 3), prickles (e.g., Diplachne gigantea, Fig. 38, Leptochloa rupestris, Fig. 48), and macro- hairs (e.g., Leptochloa fascicularis, Fig. 36, L. nee- sii, Fig. 45). Many taxa such as Eragrostiella (Fig. 20) and Indopoa (Fig. 26) had short cells whose outer walls were rounded in profile, and which were ontogenetically destined to become cork cells. (Val- dés-Reyna & Hatch (1991: 536) referred to these as papillate cork cells.) That they were destined to become cork cells became evident by examining different stages of development on a lemma, in which cells at the apical and basal portions were in different stages of development. The initially rounded outer walls of short cells appear to collapse coincident with or prior to suberin deposition; this process was evident in the shrunken tissue of the outer wall, coupled with darkened lumens of the short cells (e.g., Leptocarydion, Fig. 28, Leptochloa uniflora, Fig. 51, and L. virgata, Fig. 52). Silica cells. Silica cells were absent in all spe- cies of Leptochloa except for L. monticola (Fig. 41), which had abundant silica cells in all specimens examined, and one specimen (Snow 5811-A) of L. fascicularis (Fig. 35), which had a few silica cells present in the intercostal regions. A majority of OET genera had silica cells. In Halopyrum (Fig. 23, not visible) they were poorly developed (minimal silica deposition?), obscured 508 Annals of the Missouri Botanical Garden Table 1. Summary of lemmatal micromorphological characters for species of Leptochloa, most genera in Eleusininae, and a few genera in Eragrostideae outside Eleusininae (fide Clayton & Renvoize, 1986). Genera not included in Eleusininae are indicated with a double asterisk. Taxa are arranged alphabetically. Taxa examined by Valdés-Reyna and Hatch (1991) but not examined herein are followed by single asterisk; their data are intercalated for convenience by: (1) relying on Watson and Dallwitz (1992) for the type of microhair (as they determined from leaf blades); (2) examining herbarium material for the presence of prickles and macrohairs (Appendix 1); and (3) by leaving a question mark in the column for papillate short cells, since their interpretation of papillae differed (see Results). The generic abbreviation for species tentatively placed in Leptochloa that apparently lack a valid combination in the latter is D. (for Diplachne) + specific epithet. Abbreviations: + = present; — = absent; BM = bicellular microhairs (C = chloridoid, P = panicoid, E = enneapogonoid); Macrohairs (N = normal, CC = clavicorniculate, A = apiculate, CR = crispate); PLC = papillate long cells; PSC = papillate short cells. — PSC Prickles | Macrohairs - - + iz z a” = e Taxon Cork Silica Acrachne racemosa = Apochiton burttii Bewsia biflora Bouteloua curtipendula** Brachychloa schiemanniana Chloris paniculata** Chloris verticillata** ladoraphis cyperoides Cladoraphis spinosa | 222/222 | Eb Ge ee ee d 4ReRoRROX4A4 + + | | t++etteeeeeoetepeet | ~ | | T 2 E Z S S X S = = S à R S | + > 3 o 3 2 i 3 S = R ++++ Drake-Brockmania somalensis Ectrosia gulliveri Ectrosia leporina Ectrosiopsis lasioclada Eleusine indica Eragrostiella bifaria | VTAANANAAQAAAAAANA (44000000 | | + ++ 44 ++ ++ 44444 | | | | — E | ~ ragrostis spp. * Erioneuron spp.* Gouinia virgata Habrochloa bullockii Halopyrum mucronatum Harpachne schimperi IH ZZZ Z | Heterachne abortiva Indopoa paupercula Kengia serotina Leptocarydion vulpiastrum Leptochloa aquatica D. caudata Leptochloa chinensis +++ 000 PAA PA ORO” | | Leptochloa chloridiformis ptochloa ciliolata Leptochloa coerulescens D. cuspidata Leptochloa decipiens Leptochloa digitata Leptochloa divaricatissima deo odeboReodoe deba oe be AAA | | et op te ey ARA AAA 2222222222222 Volume 83, Number 4 1996 Snow Lemmatal Micromorphology 509 Table 1. Continued. Taxon Cork Silica iz z "S c e d a Prickles Macrohairs Leptochloa dubia D. eleusine Leptochloa fascicularis Leptochloa fusca D. gigantea Leptochloa ligulata a Leptochloa monticola qu mucronata D. mue Tebtachloa: nealleyi Leptochloa neesii Leptochloa obtusiflora Leptochloa panicea Leptochloa panicoides iflora Leptochloa rupestris Leptochloa scabra Leptochloa sp. nov. (Snow, in prep.) Leptochloa squarrosa Leptochloa uniflora Myriostachya wightiana Neesiochloa barbata Neyraudia reynauldiana Ochthochloa compressa Odyssea mucronata Odyssea paucinervis Orinus thoroldii Oropetium aristatum 3 wees fleckii Pogoneura biflora Psammagrostis wiseana Scleropogon ager Sohnsia filifolia Spartina pectinata** Steirachne barbata Tragus pedunculatus** Trichoneura CREME Tripogon major Triraphis andropogonoides Vaseyochloa multinervosa* Viguierella madagascariensis +++++++++++++++++++ ++++i+i++i++i++++i+t++++++++4 +++++++++ ME +++++++ Todd CA a OOO 1010 10061616 | | "9 | | eo "wonodococooooooocooododootm:scgTI30tro0t.tw OO 060 (C301 ie ie. DE + HAHAHA AFA AAA 4444444 BORORRRRBK€BA BAR FFAA B b RB +44 ++ ++ 44+ ++ ¢¢4 ++ 4+ 4+ 4+ N CC ZZZZZZZZZZZZZZZZZ | | ZAZRZZZZZZZZ | 4222222222422 | ZZEZZZZZ|Z21| Zi Zi 510 Annals of the Missouri Botanical Garden by epicuticular waxes, or both. Kengia (Fig. 27, not visible) and Eleusine (not illustrated) had a few sil- ica cells that were localized over the costal zones. The proximity of silica cells to cork cells was a character noted by Valdés-Reyna and Hatch (1991). Silica cells were not adjacent to cork cells in Desmostachya (Fig. 14), Ectrosia gulliveri (Fig. 17), Ectrosiopsis (Fig. 19), Myriostachya (Fig. 56), Odyssea mucronata (Fig. 60), and Steirachne (Fig. 69). Silica cells were adjacent to cork cells in Cla- doraphis (Fig. 7), Ectrosia leporina (Fig. 18), Har- pachne (Fig. 24), Odyssea paucinervis (not shown), Psilolemma (Fig. 66), Richardsiella (Fig. 67), and Sclerodactylon (Fig. 68). In Pogonarthria (Fig. 63) and Triraphis (Fig. 71) their location varied (cork cells adjacent to silica cells not shown for Trira- phis). It was uncertain whether the dark, narrow bands adjacent to some silica cells in Heterachne (Fig. 25) were artifacts of preparation (e.g., shrink- ing of cell walls adjacent to the silica bodies noted by Kaufman et al., 1972: fig. 6, and Terrell & Wer- gin, 1981) or whether the narrow bands were cork cells that were partially obscured by overarching silica bodies; the uncertainty was enhanced be- cause not all silica cells had adjacent bands. Since a few short cells with dark lumina were visible ad- jacent to some silica cells, cork cells were scored as present for Heterachne. 1 did not observe silica cells on Leptocarydion, but given their observation by Valdés-Reyna and Hatch (1991: fig. 24, tab. 2), they were recorded as present (Table 1). Among OERG, silica cells were observed Spartina and Tragus, being more common in the latter. Bicellular microhairs. Chloridoid were observed in all species of Leptochloa except L. ciliolata, Diplachne cuspidata, and L. decipiens. Microhairs generally were scattered, but in Lepto- chloa caudata and L. longa they were common near the apex. The basal cells of microhairs of Lepto- chloa panicoides (Fig. 47) and L. viscida (Fig. 5: were atypically thick and nontapering toward their Additional study is needed to eval- microhairs Wo — basal insertion. uate whether the shape of the basal cell can be hypothesized as a separate character. Among OET, seven eleusinioid genera had pan- icoid microhairs (Ectrosia gullivieri, Habrochloa, Heterachne, Neyraudia, Steirachne, Triraphis, Vi- guierella), whereas the remaining OET had chlori- doid microhairs. Microhairs were not observed in Cladoraphis spinosa or Viguierella, the latter of which needs further study, since only one specimen was available for study (Table 1). As with species of Leptochloa, microhairs varied in abundance and were generally scattered; only near the apex in Apo- chiton and Triraphis were they common. Nicora (1962: Eleusininae for Neeragrostis, which Clayton and Renvoize (1986) included in Eragrostis. The en- neapogonoid microhairs observed here for Clado- raphis cyperoides and Psammagrostis wiseana (Fig. 65) represent only the second report for this type in Eleusininae. All microhairs in Psammagrostis (Fig. 65) had swellings distal to the base of the microhair, a condition not observed elsewhere. All OERG had chloridoid microhairs. Papillae. Papillae occurred singly on short cells or singly on the distal ends of long cells. 9) illustrated enneapogonoid microhairs in Three clarifications are needed prior to presenting observations of papillae. irst, the distal ends of the outer walls of epi- dermal long cells can be noticeably swollen, as in Eragrostiella (Fig. d Indopoa (Fig. 26), and Lep- tochloa virgata (Fig. 52). These swellings were rec- ognized by yk eee and Hatch (1991: fig. 23) in at least one instance as papillae. As I interpret them, outer walls that are merely swollen are dis- tinct from true papillae. The distinction between long cells possessing a single, distal papilla (e.g.. Leptochloa coerulescens, Fig. 31) and long cells that are merely swollen can be seen with Leptochloa chi- nensis (Fig. 29), in which a single papilla is evident atop the distal swelling of a long cell, and with L. fusca (Fig. 37), in which long cells are either pa- pillate or merely swollen. Second, the outer wall of short cells prior to differentiation can appear round- ed (papillate, fide Valdés-Reyna € Hatch, 1991) when viewed in profile (e.g., Eragrostiella, Fig. 20; Indopoa, Fig. 26). The rounded outer walls should not be confused with papillae, which are localized processes arising from the outer wall. Third, the apical extensions of the lateral walls of long cells of Leptocarydion (Fig. 28) and Steirachne (Fig. 69) are distinct from true papillae; Palmer and Ger- beth-Jones (1988: 94) have noticed a similar dis- tinction in Phacelurus (Panicoideae, Andropogo- neae), did Peterson (1989) in Muhlenbergia (Chloridoideae, Sporobolinae). Thus, for this study papillae are taken to be apically rounded, undif- ferentiated processes that arise from and are local- ized on the outer cell wall, and which can become silicified (Clark & Gould, 1975; Terrell & Wergin, 1981; Consaul & Aiken, 1993). (Ellis (1979) cited Metcalfe (1960) that papillae can become cutin- ized. However, Metcalfe (1960: 668) indicated only one species (Trikerata hookeri (tribe Stipeae)) as having cutinized papillae. Furthermore, Metcalfe (1960) did not indicate how he determined the presence of cutin.) It is important to stress that oth- er structures on grasses called papillae may not be Volume 83, Number 4 1996 Snow Lemmatal Micromorphology homologous to the papillae described here (Clark $ Gould, 1975; Zuloaga, 1987: fig. 26.2c; Dávila & Clark, 1990; Filgueiras et al., 1993; Jenks et al., 1994). Papillae on short cells. Papillae on short cells occurred in several species of Leptochloa (L. fas- cicularis, L. fusca, Diplachne gigantea, D. muelleri, L. panicoides, L. uninervia, and L. viscida). In most cases they had a collapsed appearance, which may have been an artifact of preparation, or, since the silicification of papillae has been documented (Clark & Gould, 1975; Terrell & Wergin, 1981; Consaul & Aiken, 1993), perhaps was an indication that silica deposition had not occurred. Collapsed papillae on short cells are frequent in the literature (e.g., Palmer & Gerbeth-Jones, 1986: pl. 15a—f). Relatively few OET had papillae on short cells (Dinebra spp.; Drake-Brockmania spp., Fig. 16; Halopyrum, Fig. 23; Orinus, Fig. 61). Only in Hal- opyrum (Fig. 23) were the papillae large. Unlike Valdés-Reyna and Hatch (1991) I did not detect papillae on short cells of Eleusine (not shown). Among OERG, only Tragus (not shown) had pa- pillae on the short cells. Papillae on long cells. In the 13 species of Lep- tochloa having a single papilla on the long cells (Table 1) the papillae were more frequently present near the apex of the lemma. Leptochloa fusca (Fig. 37) and Diplachne parviflora (not shown) were in- consistent in this feature even over localized por- tions of the lemma; some long cells had papillae, others did not. Leptochloa viscida (Fig. 53) had only a few papillae near the lateral veins. The size of the papillae arising on long cells was more or less constant within Leptochloa, with the exception of L. coerulescens (Fig. 31), in which the papillae were noticeably longer. In some cases the papillae were weakly developed, as in Diplachne prae (not shown) and Diplachne gigantea (Fig. Although many OET had long cells zem swollen ends (e.g., Bewsia, Fig. 3; Coelachyrum yemenicum, Fig. 12; Trichoneura, Fig. 70, in part; compare with Coelachyrum poiflorum, Fig. 10, in which the long cells have no swellings), relatively few had a single papilla on the long cells (Dinebra, Fig. 15; Drake- Brockmania, Fig. 16; Kengia, Fig. 27; Odyssea, Fig. 60, not visible; Orinus, Fig. 61; and Oropetium, Fig. 62). The significantly greater basal diameter of pa- pillae in Halopyrum differs from other taxa in Eleu- sininae. Long cells with papillae were absent in the Prickles. cies of Leptochloa and tended to increase in fre- quency toward the apex. Among Leptochloa species Prickles were observed in every spe- prickles were rare in Diplachne cuspidata, L. dig- itata, L. mucronata, L. rupestris, L. uniflora, occa- sional in most species, or common, as in L. cau- ata. Prickles were observed for all OET. A single prickle was observed on Psilolemma; however, since only one was observed on three specimens, and because its apex was directed toward the lem- matal base (almost universally they point toward the apex in Eleusininae), it was considered an ab- normality and recorded as absent. Prickles were most abundant in Cladoraphis spp. (Figs. 6, 7), Ec- trosia spp. (Figs. 17, 18), Gouinia (Fig. 21), Halo- pyrum (Fig. 23), and Oropetium (Fig. 62). Unlike Valdés-Réyna and Hatch (1991), prickles were observed (with a dissecting microscope) on herbar- ium specimens of Tridens spp., Vaseyochloa multi- nervosa, and Triplasis purpurea, although they were occasionally only infrequently present at the apex. Prickles occurred on all OERG. Spartina (not shown) had prickles of widely different sizes. Prick- les were generally most abundant on the awns of species bearing these structures (e.g., Lintonia, Fig. 54); in Pogoneura, prickles were restricted to the Macrohairs. All species of Leptochloa had at least some macrohairs on the lemmas. They were not observed using SEM on Diplachne gigantea, but analysis of an isotype (Vesey-Fitzgerald 1551; BM) with a dissecting microscope revealed the presence of short macrohairs along the edges of the midrib of some lemmas. In some species they are rare, such as the sparse basal occurrence on Lep- tochloa digitata (Fig. 32). Most macrohairs ob- served in Leptochloa were typical in having smooth edges and a rounded or acute tip. However, a “clav- icorniculate" type (see Discussion) was found in D. eleusine, in which the subapical portion was no- ticeably clavate, and above which occurred a cor- niculate tip. This feature was noted earlier by Phil- lips (1974, 1982) for D. eleusine, Coelachyrum yemenicum, Lintonia, and Trichoneura. The hairs shown for Tribolium obliterum by Visser and Spies (1994: fig. 3b) may be clavicorniculate, but this is uncertain since only their apices are shown. A majority of OET had macrohairs on the lemma; they were absent in Acrachne (Fig. 1), Cladoraphis spinosa. (Fig. 7), Desmostachya (Fig. 14), Ectrosia leporina (Fig. 18), Ectrosiopsis (Fig. 19), Eleusine (not shown), Eragrostiella (Fig. 20), Heterachne (Fig. 25), Myriostachya (Fig. 56), Psilolemma (Fig. 66), Sclerodactylon (Fig. 68), Steirachne (Fig. 69), and Viguierella (Fig. 72). Only two macrohairs were observed at the very base of one specimen (Faden et al. 74/613, Appendix 1) of Harpachne schimperi. 512 Annals of the Missouri Botanical Garden The study of several herbarium sheets at MO re- vealed that a few macrohairs occasionally occur on the edges near the base of some lemmas. However, they are at most infrequent and are always rela- tively short. Moreover, given their length, one could just as easily designate them as (relatively long) prickles. Examination of specimens of H. bogdanii (MO: Heady 1466, Bogdan 4524), the other species in the genus, also revealed a few short hairs near the base of some lemmas. Given their sporadic oc- currence and their questionable status as hairs (vs. relatively long prickles), I have recorded them as present or absent (Table 1). The abundant macro- hairs on Apochiton (Fig. 2) made observation of oth- er features difficult. The clavicorniculate hairs of Coelachyrum yemenicum appear identical to those of Diplachne eleusine. A crispate macrohair, iden- tifiable by the irregular (“crisped”) surface and ap- ean reported here for the first time in grasses . 1962; Metcalfe, 1960; Ellis 1976, 1979), was al for Coelachyrum brevifolium (Fig. 8), C. poiflorum (Figs. 9, 10), and C. stoloniferum (Fig. 11). The crisping is expressed most thoroughly to- — ward the apex, and somewhat less so basally. Some — macrohairs of Trichoneura grandiglumis (Fig. 70 were swollen at the base. Among the OERG surveyed, only Spartina lacked macrohairs. Some specimens of Cynodon nlemfuensis (Fig. 13) had macrohairs with distinctly "apiculate" (sensu Peterson, 1989) tips. DISCUSSION This study originated from morphologically based cladistic studies of Leptochloa using various combinations of likely sister genera, as hypothe- sized by Clayton and Renvoize (1986). The sole use of gross morphology gave poorly resolved consensus trees and. clades with low support values (Snow, unpublished). Moreover, the use of a single data set may have resulted in artificial groupings, a poten- tial problem Hilu and Wright (1982) alluded to in their phenetic study of chloridoid grasses. These factors suggested that additional characters might lead to more "accurate" (sensu Hillis & Bull, 1993 estimations of the phylogenetic relationships. Given — that micromorphological characters have known systematic value in grasses (references in Introdue- tion), this study was undertaken to enlarge extant data sets for purposes of phylogenetic inference. Cork Cells. chloa, cork cells are of no phylogenetic value in Given their universality in Lepto- the genus. owever, the variable presence of cork cells in Eleusininae and Eragrostideae suggests they are phylogenetically informative at the generic level. Cork cells were recorded as present only when the short cell was substantially darkened (as viewed by SEM), whereas Palmer and associates generally re- corded cork cells as present when any undifferen- tiated short cell occurred adjacent to a silica cell e.g.. Palmer et al., 1985: pl. le, 2e, 3e, 4c, and many others therein). I agree with J. T. Columbus (pers. 1995) that the extent to which suberin deposition actually occurs, and the extent to which it can be observed, needs further investigation. If suberin deposition in short cells is T comm., January eventually demonstrated to be invisible to SEM, or it is shown that “cork cells” on lemmas are lacking in suberin, then their use as phylogenetic markers will need reconsideration. These considerations aside, I have followed others (e.g., Valdés-Reyna & Hatch, 1991) by recording cork cells as either pres- ent or absent. Silica Cells. absent from Leptochloa, and thus have minimal in- Lemmatal silica cells were mostly frageneric phylogenetic value. Since only one spec- imen of L. fascicularis (Fig. 35) had a few silica cells, their occurrence on that specimen is probably similar to lemmatal stomata, which occasionally reappear as atavisms from the transformationally antecedent leaves. However, in Leptochloa monti- cola (Fig. 41) silica cells are common, which re- quires a reassessment of earlier research. Usin several lines of evidence, Valls (1978) suggested that L. monticola was generically misplaced. Based on leaf anatomy and citing Clifford and Watson (1977), Valls (1978) suggested a possible alliance with Chionochloa Zotov (Arundinoideae, Arundi- neae). To assess a possible alliance with Arundi- neae, | sampled lemmas from Chionochloa conspi- cua (Forst. f.) Zotov subsp. cunninghamii (Hook. f.) Zotov, C. flavescens Zotov, Danthonia dominguensis Hack. & Pilg., and Rytidosperma pilosum (R. Br.) Connor & Edgar (Appendix 1; data not shown). Un- ike L. monticola, which had chloridoid microhairs, the microhairs of C. conspicua var. cunninghamii and R. pilosum were panicoid; furthermore, micro- omin- guensis. Of relevant note, Watson and Dallwitz (1992) reported panicoid microhairs (for the abaxial leaf surface) for Danthonia, which does not accord hairs were not seen for C. flavescens or D. with the chloridoid microhairs of Leptochloa mon- ticola. In addition, the species of Chionochloa lacked cork cells, which were present in L. mon- ticola. Morever, the dumbbell-shaped silica bodies in L. monticola differed from the saddle-shaped sil- ica bodies of C. flavescens and ominguensis. Whereas each of these taxa has at least one dis- crepancy when compared to L. monticola, lemmatal Volume 83, Number 4 1996 Snow 513 Lemmatal Micromorphology micromorphology alone does not support an obvious relationship of L. monticola to Arundineae. Saddle-shaped (Clifford & Watson, 1977; Valls, 1978), dumbbell-shaped (Valls, 1978), and cross- shaped (Metcalfe, 1960) silica bodies have been reported for leaf blades in Leptochloa. Valls (1978: 82) has shown that both dumbell-shaped and sad- dle-shaped silica bodies occur in Leptochloa dubia. Ovate and saddle-shaped silica bodies were ob- served for Odyssea mucronata. The presence of two shapes of silica bodies on L. dubia and O. mucron- ata cautions against their uncritical use as diag- nostic or phylogenetic markers. Recognition of shape differences of silica bodies as distinct char- acters (or character states) thus seems premature until it can be demonstrated that infraspecific vari- ation in silica body shape is minimal. This caveat is supported by studies in Oryza (Whang & Kim, 1994) and Zizania (Terrell & Wergin, 1981) (both genera in tribe Oryzeae), which documented exten- sive infraspecific variation in the shape of silica bodies. Ignoring for the moment silica body shape, and focusing on mere presence or absence, silica cells appear to be important phylogenetic markers, since they were not observed in over half of the OET and OERG examined. Their presence on the lemma is probably symplesiomorphic, given the near univer- sal occurrence of silica bodies in grass leaves, and that the lemma is transformationally homologous to the leaf. Testing for parallel loss or gain, however, will require a cladistic approach (Snow, in prep.). Bicellular microhairs. The lack of microhairs (e.g., Leptochloa ciliolata, Diplachne cuspidata, L. decipiens) is not unequivocal evidence for their ab- sence, since a single microhair was often all that was visible for three or more specimens of a given taxon. Microhairs were not restricted to certain por- tions of the lemma, but, when infrequent, were usu- ally found near the apex. In some taxa, such as Lintonia, they extend onto the awn. Given a simple presence or absence, microhairs probably would be of little phylogenetic value in Eleusininae. How- ever, two features of microhairs suggest their utility as systematic markers. First, Amarasinghe and Watson (1988) demonstrated that ten chloridoid genera, including Leptochloa, have partitioning membranes in the basal cell of the microhair. More- over, the membranes were limited to the subfamily Chloridoideae (Amarasinghe & Watson, 1988: 307). Of the genera they surveyed in Eleusininae, Era- grostis, Pogonarthria, and Triraphis lacked the membranes. The use of partitioning membranes as a potential phylogenetic marker in the subfamily merits further investigation over a wider range of taxa. Second, with few exceptions (e.g., Eragrostis, Watson & Dallwitz, 1992), microhairs can be readi- ly assigned to one of three morphological types: chloridoid, panicoid, or enneapogonoid, based on morphological differences of the basal and distal cells and their length ratios (Tateoka et al., 1959; Jacobs, 1987; Amasaringhe & Watson, 1988; Wat- son & Dallwitz, 1992; but see below also). The elongated aspect and thin wall of the distal cells in the microhairs of seven OET (e.g., Habroch- loa, Fig. 22) characterize them as panicoid micro- hairs (Table 1), which in all cases accords with the observations of Watson and Dallwitz (1992) for leaves. The discovery of enneapogonoid microhairs on the lemmas of Cladoraphis cyperoides and Psam- magrostis wiseana represents the third report of en- neapogonoid microhairs outside the tribe Pappo- phoreae (Chloridoideae) and the second in Eleusininae. They were first reported for Amphi- pogon strictus (Arundinoideae) by Amarasinghe and Watson (1988). Earlier, Watson and Dallwitz (1992: 232) were unable to find microhairs on leaf blades of C. cyperoides from photographic material provid- ed by R. P. Ellis. Of the bicellular microhairs il- lustrated by Tateoka et al. (1959), the microhair of C. cyperoides seems to most resemble those of Spo- robolus vaginiflorus (Tateoka et al., 1959: fig. b,64) and Pappophorum elegans (Tateoka et al., 1959: fig. b,155). It also somewhat resembles a microhair re- ported by Peterson (1989: fig. 2,d) for two species of Muhlenbergia. The swelling distal to the base of the microhairs in Psammagrostis (Fig. 65) was ab- sent elsewhere in Eleusininae. Although not identified as such, an enneapogo- noid microhair was recently shown for the arundi- noid species Pentameris distichophylla (Barker, 1993: fig. l.c) The recognition of three mor- phological categories of microhairs admittedly rep- resents typological thinking, which can misguide the understanding of biological reality (Mayr, 1982). Thus, when making preliminary hypotheses of homology (de Pinna, 1991), one might consider whether the “enneapogonoid” microhairs of Cla- doraphis cyperoides and Psammagrostis wiseana are homologous with those occurring in Enneapogon, or whether the relative elongation of the basal cell is a “character” itself, the elongation having been achieved independently from the enneapogonoid type. Overall, the distinct morphologies of micro- hairs in Eleusininae and their tendency to be re- stricted generically together suggest their potential value as indicators of phylogenetic relationships. Papillae. The characterization and interpreta- tion of papillae were initially confounded by several Annals of the Missouri Botanical Garden factors (see Results). Despite the more restricted criteria given above, the presence or absence of papillae is probably of phylogenetic value in Lep- tochloa and other genera in Eleusininae. The pa- pillae on short cells were almost always collapsed, and thus could be confused with the collapsing out- er wall of short cells, which apparently occurs dur- ing suberin deposition of cork cells (see above That these structures could rightly be called papil- lae only suggested itself after repeated comparisons — between numerous photos of taxa with and without this feature, and seeing similar photos of collapsed papillae in the literature. The papillae on short cells are perhaps best revealed on Drake-Brock- mania somalensis (Fig. All members of the ne hloa fusca group (ten- tatively: Diplachne cuspidata, L. fascicularis, L. fus- ca, Diplachne gigantea, D. muelleri, D. parviflora, L. uninervia) have a single papilla located distally on long cells, as well as having some short cells with a single papilla. Other characters that suggest a close relationship among its members include an apically attenuated ligule, relatively long spikelets, spikelets with frequently five or more florets, the perisperm of the caryopsis being only weakly ad- nate to the endosperm, and a pronounced lacuna in the midvein of the leaf blade (Snow, in prep.). Leptochloa chinensis and L. coerulescens, which both have papillae on long cells, have not been previously hypothesized to be closely related. Many species of Leptochloa have never been apportioned into subgenera or sections, but based on Clayton and Renvoize (1986), papillate species that would be placed in section Diplachne include the L. fusca group, L. panicoides, and L. viscida, whereas other papillate species would be placed in section Lep- tochloa. Papillae on short cells are also known from Festuca (Consaul € Aiken, 1993) and other genera in Pooideae (Thomasson, 1986). the Eleusininae genera studied herein re- ported to have papillae (Valdés-Reyna & Hatch, 1991), I observed them only in some species of Leptochloa. This is due to the different interpreta- tion of papillae used for this analysis (Results). The papillae of Halopyrum, which have a larger basal diameter than related genera, approximate a con- dition found in some genera of other subfamilies (e.g.. Brachiaria (Panicoideae, Paniceae), Zuloaga & Soderstrom, 1985; fig. 2d; Olyra (Bambusoideae, Olyreae), Soderstrom € Zuloaga. 1989; fig. 21c). in which the entire outer wall swells outward, thus making more difficult the distinction between pa- pillae and swollen outer walls. Among OERG, I was unable to determine wheth- er Tragus had papillae on short cells, since distin- guishing between long and short cells was difficult; this uncertainty is reflected by the question mark in Table 1 Prickles. Whereas Valdés-Reyna and Hatch (1991) did not observe prickles in their sample of Leptochloa, Y found a universal occurrence of prick- les in the. genus. Given their universality and as- suming that Leptochloa is monophyletic, they have no phylogenetic value. Others have suggested that the location of micromorphological characters (e.g.. costal vs. intercostal zones) should be noted (Ellis, 1979; Consaul & Aiken, 1993). However, no ob- vious differences in costal and intercostal zones were noted for Leptochloa. Peterson (1989) differ- entiated between regular prickles and apiculate prickles in some species of Muhlenbergia, but this distinction did not hold in Leptochloa, OET, and OERG (but see comments below under macrohairs). The phylogenetic value of prickles is similar for OET, since only Psilolemma was lacking this char- acter. Species of Trichoneura and Tripogon studied by Valdés-Reyna and Hatch (1991) lacked prickles, but I observed them in other species of these gen- era (Table 1). As documented so thoroughly by Dá- vila and Clark (1990) for leaf blade epidermal mi- cromorphology, all species need to be sampled to fully assess the occurrence of prickles and other micromorphological features within a genus. That the size of prickles could vary on a lemma is most evident in Spartina (not shown), in which prickles on the margin of the lemma dwarfed those in the intercostal zones. Sizable differences were also evident in Apochiton, Cladoraphis spinosa, Di- nebra retroflexa (Fig. 15), Leptochloa longa, L. neal- leyi (Fig. 44). and L. Others have found discrepancies in lemmatal or laminar prickles with- in a species (Prat, 1948: his large and small “spic- ules”; Davies, 1959: his “asperities” and “incipient asperities”; Ball et al., 1993: compare fig. 1 with fig. 8: Thomasson, 1986: fig. 15, and noie the in- 5, and 7 scabra. creasing size of prickles between figs. 2, Barker, 1993: fig. 2a—d). In several cases prickles were flat at the base or had weak conduplicate basal folding (present but not shown in: Orinus, Leptochloa fascicularis, L. monticola, Diplachne muelleri), a condition illus- trated oe 1987: fig. 26.4d; Thomasson, 1986: figs. ; Valdés-Reyna € Hatch, 1991: fig. 44) and dise anne (Consaul & Aiken, 1993: 1656) ear- lier for other taxa. The conduplicate condition may be an early expression of prickle ontogeny in which the outer wall evaginates prior to its expansion, af- ter which silica deposition may occur and the ri- gidity of the prickle is achieved. The accumulation of silica in the tips of prickles (Consaul & Aiken, Volume 83, Number 4 1996 Snow 515 Lemmatal Micromorphology 1993) probably explains their brittleness and pro- pensity to break (e.g., Gouinia, Kengia (Fig. 27)). Two categories of prickles (prickles and hooks) have historically been recognized (Metcalfe, 1960; Ellis, 1979). The tendency to distinguish between hooks and prickles reflects the frequent occurrence on specimens of two distinct sizes of projections of the outer wall (e.g., Palmer & Gerbeth-Jones, 1986: pl. 10a,c & pl. 12c,f; Dávila & Clark, 1990: fig. 34). Since the distinction between them seems ab- solutely arbitrary within Eleusininae (e.g., note gra- dations in Dinebra retroflexa, Fig. 15), 1 have cho- Metcalfe (1960) suggested that hooks were homologous with prick- sen not to recognize hooks. les. His use of the term homology was probably invoking the concept of intermediate forms (Re- mane, 1952; Sattler, 1984), which illustrates their common developmental trajectories, and which ac- cording to some intepretations of homology (de Queiroz, 1985) establishes their homology. Ellis (1979: 668) indicated that prickles on leaf blades originate from short cells, but prickles clearly also arise from long cells of lemmas on Gouinia (Fig. 21), Kengia, Leptochloa monticola (Fig. 41), Linton- ia (Fig. 54), and Pogonarthria (Fig. 63). If prickles continue to be recognized as characters distinct from macrohairs (but see below), it will be neces- sary to recognize that prickles can arise from either short cells (Ellis, 1979) or long cells (this study). Macrohairs. Ellis (1979) has stated that mac- rohairs arise from long cells. Kellogg (1990: 1983 noted that the arachnose lemmatal hairs of Poa ap- — pear to initiate from short cells, but only because those cells have not yet elongated into long cells. My observations indicate that macrohairs can arise from long cells or short cells. Their origin from short cells is most evident in Bewsia (Fig. 3), Lep- tochloa fasicularis (Fig. 36), L. neesii (Fig. 45), and Neesiochloa (Fig. 57). On long cells, they can arise from the proximal end, as is evident in Coelachy- rum yemenicum (Fig. 12), Diplachne eleusine (Fig. 34), Trichoneura (Fig. 70), Triraphis (Fig. 71), or rarely on the distal end, as for example in L. mu- cronata (Fig. 42). In Leptochloa spp., normal macrohairs were gen- erally restricted to the lower and middle portions of the nerves. A clavicorniculate type was observed for Diplachne eleusine (Fig. 34), which was also present in Coelachyrum yemenicum (Fig. 12). Phil- lips recognized clavicorniculate hairs for both D. eleusine and Coelachyrum yemenicum (as Cyphole- phis yemenica), calling them “club-shaped” (Phil- 1974) and later (Phillips, 1982: 155). Since the hairs have a distinctly narrowed lips, “clavate” portion above the subapical swelling, the clavicor- niculate designation seems to best describe their morphology. (The assistance of D. Nicolson in mat- ters of terminology is here acknowledged.) Clayton and Renvoize (1986) distinguished Leptochloa from Coelachyrum in part by the broadly elliptic, con- cavo-convex, rugulose caryopsis of the latter. The inflorescence of Diplachne eleusine resembles that of C. yemenicum in having few, relatively short, and more or less erect spicate branches. In addition, the caryopsis characters of the two species are sim- ilar in outline and cross-sectional shape, and in having a smooth pericarp that is weakly adnate to the endosperm (Snow, unpublished). Like most spe- cies of Leptochloa and unlike most species of Coe- lachyrum, the grain of C. yemenicum is not rugu- lose. Watson and Dallwitz (1992) segregated C. yemenicum into the monotypic genus Cypholepis and mentioned the presence of clavicorniculate hairs. Phylogenetic studies are needed to determine the generic affinites of D. eleusine and C. yemeni- cum, but the above evidence, along with cladistic studies by Van den Borre (1994: 216), suggests they may be phylogenetically close. This study is the first to document the crispate macrohair, which occurs in Coelachyrum breviflo- rum (Fig. 8), C. poiflorum (Figs. 9, 10), and C. sto- loniferum (Fig. 11). It is characterized by an irreg- ular wrinkling (crisping) of the surface, which is expressed most fully toward the apex. That the wrinkling was found repeatedly in these taxa and not elsewhere suggests it was not an artifact of preparation. In grasses, the morphological “type” most resembling the crispate type is that of Dan- thoniopsis viridis (Panicoideae, Arundinelleae), which has a sparsely papillate shaft (Palmer & Tucker, 1981: fig. 9f). It differs from the crispate type in that significant portions of the shaft are rel- atively smooth. Infrequent and irregular papillate bulges on the shaft of macrohairs have been noted sporadically on the lemmas of Neostapfia colusana Davy and Orcuttia pilosa Hoover (Snow, unpub- lished), both members of the tribe Orcuttieae (Chloridoideae). Trichomes having micropapillate wall sculpturing were reviewed by Uphof (1962) and recently have been noted in Rubiaceae (Sul- livan, 1986) and Verbenaceae (Rueda, 1994). The exclusive occurrence of crispate macrohairs in three species of Coelachyrum "cr it is a syn- apomorphy at some level in that gen As mentioned above, Peterson (1989) used the term “apiculate” to distinguish between the apices of prickles in some species of Muhlenbergia. This distinction is useful for the apices of some macro- irs. The apices of macrohairs in two specimens ha (Crook 2115, Gereau et al. 3490) of Cynodon nlem- 516 Annals of the Missouri Botanical Garden fuensis (Fig. 13) were noticeably apiculate in com- parison to other taxa (Appendix 1). The possible systematic significance of apic ‘ulate tips warrants further attention. Peterson (1989) also recognized “swollen base” macrohairs in some species of Muhlenbergia. Some, but not all, macrohairs of Trichoneura grandiglumis (Fig. 70) had this feature. Macrohairs with swollen bases may arise from long cells, which have a larger foundation relative to short cells. Future research should evaluate whether macrohairs have normal or swollen bases. It is important to distinguish between the base of the macrohair and the surrounding cells that can swell outward to buttress the macrohair at its base (Ellis, 1979: 65 matal macrohairs have been called (collectively) tu- 7). Swollen cells adjacent to lem- bercles or papillate bases, and were noted in OET only for Richardsiella (Fig. 67) and in OERG for Tragus (not shown). The tubercles were much larger in Richardsiella. ike prickles, the basal portion of some macro- hairs was distinctly flattened or somewhat condu- plicate, as in Bewsia (Fig. 3), Chloris verticillata (Fig. 5), Coelachyrum yemenicum (Fig. 12), Lepto- chloa neesii (Fig. 45), and L. squarrosa (Fig. 50). This condition is evident in some macrohairs of the unrelated genus Olyra (Soderstrom & Zuloaga. 1989: fig. 6c). elongation can be seen with three closely spaced hairs on Chloris verticillata (Fig. 5). As with prick- les, the outer wall of the cell appears to evaginate Ontogenetic stages of macrohair in a conduplicate manner and flatten as the macro- hair begins development. Although most fully de- veloped macrohairs lack the folded base. the base sometimes remains flattened (Coelachyrum yemen- icum, Fig. 12; Leptochloa neesii, Fig. 45; L. squar- rosa, Fig. 50: and Neesiochloa, Fig. 57). The base of macrohairs can be hollow, as evidenced by ba- sally severed hairs in Leptochloa fusca (Fig. 37) and Ochthochloa (Fig. 59) Phylogenetic utility of lemmatal micromorpholo- gy. Micromorphological characters generally vary little in Leptochloa. With few exceptions, cork cells were present, silica cells were absent. chloridoid bicellular microhairs were present, prickles were present, and macrohairs were present. The occur- rence of papillae on long cells and short cells was more variable, suggesting their presence or absence may be phylogenetically informative within the ge- nus. In addition to Leptochloa, seven genera were studied for two or more species. Overall, little in- trageneric variation was generally observed for giv- en characters. For example, only Odyssea varied in its expression of cork cells (Table 1). As so thor- oughly documented by Dávila and Clark (1990) for the leaf blades of Sorghastrum, a complete survey of all species is necessary to fully characterize a genus. So, although further study is needed, it ap- pears that lemmatal micromorphological characters within genera of Eleusininae are largely conserva- tive. However, as evident from Table 1, significant variation occurred between genera. As such, mi- cromorphological characters are probably reliable phylogenetic markers at the generic level. Micromorphological characters and problems of homology. The micromorphological characters in this study have been reported and discussed fol- lowing the tradition of Metcalfe (1960) and Ellis (1979). That is, cork cells, silica cells, bicellular microhairs, papillae, prickles, and macrohairs have been treated separately, without questions having been raised about their ontological status as dis- tinct characters. However, abundant evidence from the literature suggests that papillae, hooks, prick- les, and macrohairs often represent arbitrary stages along one or two developmental trajectories. (If we recognize the fundamental difference between long and short cells, and if we therefore choose to dif- ferentiate between the outward extensions of the outer cell walls of short cells vs. long cells (into papillae, hooks, prickles, macrohairs), then one might argue that there are separate developmental trajectories for outer cell wall extensions of short and long cells.) Confounding matters has been an inconsistency in the previous use of terminology. As one of many examples, the cupulate papillae in Muhlenbergia (subfamily Chloridoideae; Peterson, 1989) resem- ble the hooks of some Melica spp. (subfamily Pooi- deae; Thomasson, 1986). This may simply reflect differences in papillae at the subfamilial level, or suggest that papillae between the two subfamilies are not homologous. As a second example, most agrostologists have seen the gradual transition that occurs in the axils of ys > paman branches on the rachis from “prickles” to “macrohairs.” Such tran- sitions are easily boni. on individual specimens. Statements such as . . it seems probable that those [prickles] that even- tually become pointed may pass through an unbarbed phase [i.e.. papillae and/or hooks] ean their ontoge- netic ibat ment but this p ded further in- vestigation (Metcalfe 1960: and The distinction between mac m and prickles is often not clear ... (Ellis 1979: € suggest that the distinction between character and Volume 83, Number 4 1996 Snow 517 Lemmatal Micromorphology character state is ambiguous, and illustrate the need for reconsidering the relationship between ho- mology and ontogeny for these features. Morden (1985: 36), upon examination of micro- morphological characters on the lemma and palea of some species of Muhlenbergia, stated “The pres- ence of papillae on the epidermis of the floret ap- pears to represent the initiation of villous [i.e., ma- crohair] growth, whether or not elongation follows.” he photomicrographs (Morden, 1985: figs. 12- 19, 24-27) reveal that the papillae on immature florets (lemmas and paleas) are present in regions where macrohairs develop on mature florets. The thinning of epicuticular waxes on the elongating pa- pillae (Morden, 1985: fig. 25) suggests that wax- covered papillae can elongate into macrohairs (giv- en, it is aao, some sort of genetic signal). However, since s an have several pa- pillae per cell Metcalfe, 1960: Ellis, 1979), it is doubtful that all papillae represent early stages in the ontogeny of macrohairs. To be used as phylogenetic markers, there must be no ambiguity surrounding the ontological status of characters. Regarding papillae, hooks, prickles, and macrohairs, it remains necessary to question which entitites can be meaningfully compared as homologous characters (Haszprunar, 1992), since gradations between “characters” can be seen on in- dividual specimens: e.g., (1) between papillae and hooks (Thomasson, 1986: figs. 1, 4, 7, 12, 14); (2) between papillae and prickles (Davila & Clark, 1990: fig. 29); (3) between papillae and macrohairs (Morden, 1985; discussed above); (4) between hooks and prickles (Thomasson, 1986: figs. 4, 15; Davila & Clark, 1990: fig. 3); and (5) between prickles and macrohairs (Palmer & Gerbeth-Jones, 1988: pls. 9c, 10a, 28f; Valdés-Reyna & Hatch, 1991: fig. 36; Consaul & Aiken, 1993: fig. 23; Bar- ker, 1993: fig. 2b). Given these observations, the distinctions between papillae, hooks, prickles, and macrohairs have been somewhat arbitrary at best. Although the uncertain homology status of sev- eral micromorphological characters and their ques- tionable value as phylogenetic markers should now be evident, it would be premature to assert that distinctions between them are always arbitrary, and that they thus hold no phylogenetic value. Rather, previous research and the results herein indicate the need for a thorough review of the ontological status of papillae, hooks, prickles, and macrohairs. If these features are not ontogenetically individu- alized structures that are comparable across taxic hierarchies, then their utility as phylogenetic mark- ers is suspect. Their phylogenetic application needs to be evaluated in light of theoretical considerations of ontogeny and homology (e.g., Patterson, 1982; Roth, 1984, 1991; de Queiroz, 1985; Wagner, 1989a, b; de Pinna, 1991; Hall 1992, 1994) and will be the basis of a subsequent paper (Snow, in prep.). Literature Cited Amarasinghe, V. & L. Watson. 1988. Comparative ultra- P npo of microhairs in grasses. Bot. J. Linn. Soc. 98 303-31 Ball, T. B. d. D. Brotherson & J. S. Gardner. 1993. A typologic and morphometric study of variation in phy- toliths from einkorn wheat (Triticum monococcum). Can- ad. J. Bot. 71: equus Barker, N. P. 1993. A biosystematic study of Pentameris (Arundineae, lt :eae). Bothalia , H. P. Linder & E. H. Harley: 1995. Polyphyly of ED (Poaceae): Evidence from rbcL se- quence data. Syst. Bot. 20: 423-435. Bor, N. L. The Grasses of Burm and Pakistan (Excluding jpg Rep tion. Otto Koeltz — Koenigstei Borre, A. Van den. 1994. Taxonomy of ile "Chloridoideae e With nl Reference to the Genus Era- . Thesis, Australian National University, h Ce India ted Edi- & L. Watson. 1994. The infrageneric classifica- tion of Eragrostis (Poaceae). Taxon 43: 383—422. Campbell, C. S. 1985. The subfamilies and tribes of Gra- mineae (Poaceae) :i the southeastern United States. J. Arnold Arbor. 66: —299, Chappill, J. A. B Quantitative c characters i in phylo- genetic analysis. Cladistics 5: 217-234. Chen, S.-L., Y.-X. Jin & Z.-J. m 993. Micromorphol- ogical Atlas of Leaf Epidermis in de ao Jiangsu Science and Technology Press, Nanjin Clark, C. A. & F. W. Gould. 1975. Some ‘evidential char- actersitics of pem of ng Panicum, and ochloa. Amer. J. Bot. 62: 743-748. Clark, T. G., W. Zhang & D Wendel. 1995. A phylog- eny of the grass family pci based on ndhF s quence data. S em Bot. 6-460. Clayton, W. D. E Pp. 397-398 in Flora of West Tropical jet Crown Agents for Oversea Gov- ernments and RM Lon E & S. A. Renvoize. era Graminum Grasses of the World. Kew Bull. ms Ser. XIII. Royal Botanic Gardens, Kew. . M. Phillips & S. A. Renvoize (Editors). 1974. Craminene (Part 2). Flora of Tropical East Africa. Crown Agents for Oversea Governments and Adminis- trations. Whitebriars Press, London. ecu is T. 1987. Spikelet and floral morphology. Pp. 30 in T. R. Soderstrom, K. W. Hilu, C. S. Campbell a Barkworth (editors), Grass Ssutematics and Evolution Smithsonian Institution Press, Washington, DC. cing L on. 1977. Identifying Grasses: Data, raga "i ¡ilustran Univ. Queensland Press, St. Luc Consaul L. L. € S. G. Aiken. 1993. Limited taxonomic value of palea intercostal = ee in North Am ican Festuca (Poaceae). Canad. . 71: 1651- 1659. Cope, T. A. 1985a. Key to Sinon of mi Arabian Pen- 518 Annals of the Missouri Botanical Garden insula, Arab Gulf J. Res., Special Publication No. 1, Vol. 1 —.. 1985b. Key to Somali ( National Her- barium, Mogsdahn, Dahlgren, R. M. T., H. T. Clifford & P. F. Yeo. 1985. The Families of the ns iR Structure, Evolution, and Taxonomy. Springer- Verla ag, Berlin. Davies, I. 1959. The use of ed characteristics for the identification of i in the leafy stage. J. Brit. Grassland Soc. 14: 7-16. - Clark. 199 croscopy survey rdi de p eni of yi eei (Po- eae: Ren: er. J. Bot. 77: 499-511. Dave J. R. J. Soreng. 1993, a structure in the grass family (Poaceae) as inferred from dgs Grasses. = ) Scanning electron mi- plast DNA restriction site variation. Amer. J. Bot 1444-1454 Doyle, J . R. J. Soreng, D. Garvin & M. J. Anderson D p qs DNA inversions and the da of i grass family (Poaceae). Proc. Natl. Acad. , US pv 77 a Duvall, M. “4 P. M. Peter 1994. Alliances s of Miler (Pose eae) within New World . H. Christensen. Elffers, J. & J. Ke ne O'Byrne. 1957. Notes on African grasses: XXIV. Richardsiella, a new genus of grasses m en val Africa. Bull. 11: 455-462. Ellis, R. P. 1976. A proc ses for stondandivine compar- ative a pro in the Poace eae. I. The leaf-blade as viewed in : . A gies bn E compara- The epidermis as seen in sur ría ace view. Bothalia 12. 641-671. Filgueiras, T. S., Zuloaga. 1993. A new spec ies of Streptostac hys (Pose eae: Paniceae) from -251. a) a z N 1 ~ © z © = 3. Watson, M. Koekemoer, L. Smook, N. P. Barker. H. M. Anderson & M. J. Dallwitz. 1990. Grasses of Southern Africa. Mem. Bot. Surv. S. Africa i 58, Gould, E Univ. ES ress, College St Sta Guala, G. E, H. 1995. A put analysis and revision of the genus Apoclada yo eae: Bambusoideae: Bam- busodae). Syst. Bot. 20: 207-223. Hall, B. 1992. Pats Developmental Biology. Chapman and Hall, London ——. 1994. Homology: The Hierarchical Basis of Comparative Biology. Academic Press, San Diego. Hartley, W. & C. Slater. 1960. Studies in the origin, evo- lution. and distribution of the reir none IH. of subfamily Eragrostoideae. Austral. J. Bot. 276. Haszprunar, G. 1992, significance for evolutionary biology and phylogenetics. J. Evol. Te 5: 13-24. Hillis, D. M. & J. J. Bull. 1993. An empirical test of Dante as a dA n for WERE HA in Sra du i s. Syst. Biol. 42: 182-19 Hilu, K. W. & A. Den 1993. ae and imn NUM ic al wow in the n = Subfi Amer. J. Bot. 80: 104-1 & K. Wright. T E Re of Gramineae: study. Taxon 31: 9-36. The classification of Panicum (Gramineae) ~ a QE S "Eu o — - - at = D Dp c ~ S sale Botanic Gardens. e rasses of Texas. Texas A&M The tribes 8: 256— The types of homology and their amily Chloridoi- deae. Cluster analysis Hsu, €. 1965. and its allies with special reference to the characters of lodicule, style base "s oe J. Fac. Sei. Univ. To- kyo, Sec. 3, Bot. 9: Jacobs, S. W. L. rl Syste matics of the chloridoid grasses. Pp. 277-286 in T. R, Soderstrom, K. W. Hilu, C. S. Campbell & ] «worth (editors), Grass ; Sys- tematics and Pad Smithsonian Institution. Press, bep ide C Jenks, A. P J e i cork cells wax filaments on y Sorghum bicolor (L.) Moench. Int. Plant Sci. 15 Johnston, C. R. = L. NN 1976. Microhairs: A uni- versal characteristic of non- e stucoid grass genera? Phstomorpolos 26: 297-301. Judziewicz, E. 1990. Poaceae. "es En (editor), Iba of the Galanas: Books, Germ es Kabuye, & D. Wood. 1909. multic gllilar Sonn hairs in the Gramineae. Bot. J. Ann. Soc. T p Kaufman, P. i. Poleras & P. A. Adams. 1969. Regulation A x and cellular diffe 'rentiation in de- veloping Arena ig eie by gibberellic ac e and in- dole-3-acetic acid. r. J. Bot. 56: 918— ———, ———— & J. A "Smith. 1970. reee tral de: ve lapseni of ain a cell ne in Avena internodal epidermis. Bot. Gaz. 131: 173 —————. J. D. Lacroix, J. J. "iini L F. Allard € W. C. Bigelow. 1972. Scanning electron microscopy and elec- tron microprobe analysis of silicificat florescence bracts of Avena sativa. 1018-1025. Kellogg, E. A. 1990. Ontogenetic studies in florets of Poa (Gramineae): Allometry and heterochrony. Evolution 44: 1978-1989 Rich id E. N. Ashworth. 1994. In- the secretion of epicuticular = Górts-van Koeltz Scientific A first record of tion patterns in in- Amer. J. Bot. ™ AS al = & ea her 1987. Phylogenetic analyses of the eae ae. Pp. 310-322 in T. R. Soderstrom, K. W. Hilu, C. S. Campbell & M. E. Barkworth (editors), Grass SERA and E volution. Smithsonian Institu- tion Press, Marsan n, D. — ——- & H. P. Linder. 1995. Phyloge ny E Poales. Pp. 511—542 in A J. Rudall, P. J. Cribb, ù Cutler & C. Humphrie s (editors), Monoc ree Systematics and Evolution. Royal Botanic Gardens, Kew. Lucas, M. A. 1979. Hooked spikelet emergences of Pseu- doechinolaena (Gramineae—Panicoideae) I. C ona je ae and iiA Iselya 1: 115-139. 982. The Growth of Biologic ‘al Thought: Div- erstiy, Evolution and Inheritance. The Belknap Press of Harvard Univ., Cambridge. Me eh " "1979. Diplac dates e d ‘hloa (Poaceae) in rth America. Brittonia 1940: Me oe . 1983. itor), Flora Novo-Galiciana, vol. 14. Univ. Press, Ann Arbor. Metcalfe, C. R. 1960. Anatomy of the I. Gramineae. Clarendon Press, Oxfor Molina, A. M. 1993. Las especies k del. g (Gramineae: Poeae) de Sudamérica. Parodiana 8: 67 - Ser pm V. E Anderson (ed- Michigan MEUM: énero Koeleria 37- Marlen: C. W. A Biosystematic Study of the Muh- lenbergia repens ub. (Poaceae). Ph.D. Dissertation, Texas A&M University, College Station. 5 Ha tech. 1987. Anatomical study of the A Mida repens pue x "i eae: Chloridoideae: 347-3509. Eragrostideae). Sida 1 Volume 83, Number 4 Snow 519 1996 Lemmatal Micromorphology Naredo, M. E. B., D. A. Vaughan & F. ps Cruz. 1993. Riedl, R. 1978. Order in Living Organisms. John Wiley ee spikelet morphology of Oryza schlecteri Pilger and related species of Leersia ind Oryza (Po- aceae). J. Pl. Res. 106: 109-112 Nicora, E. G. 1962. Revalidación del género de Grami- neas “Neeragrostis” de la flora Norteamericana. Rev. Argent. Agron. 29: 1-11. —— 1995. Los géneros Diplachne y Leptochloa (Gra- mineae, Eragrosteae) de la Argentina y países limitro- fes. Darwiniana 33: 233-256. Ortíz, J. J. 1993. Estudio sistemático del género Gouinia (Gramineae, Chloridoideae, Eragrostideae). Acta Bot Mex. 23: 1-33. Palmer, P. G. & S. Gerbeth-Jones. 1986. A scanning elec- tron microscope survey of the epidermis of I African grasses, IV. Smithsonian Contr. Bot., No. 6 & 1988. A in d im micro- scope survey of the epidermis of East African grasses, V, wow s African supplement. Smithsonian Contr. Bot., y E Tucker. 1981. A scanning electron mi- croscope survey of the o of East African grass- es, I. pg iene Contr. Bot., No. 49. ;erbeth-Jones & S. uis es 1985. A scan- ning ders microscope survey of the epidermis of East African grasses, II. Smithsonian Contr. Bot., No. 2 EN C. 1982. ing rg ate characters and ho- mology. Pp. 21-74 in: K. A. Jos . E. Friday (ed- itors), Systematics Association Spec ‘ial Volume No. 21, Problems in a logenetic Reconstruction. Academic Press, » Peterson, P. Mo 989. Lemma micromorphology in the annual Muhlenbergia (Poaceae). Southw. Naturalist 34: b . R. Annable. 1990. A revision of Blepharo- neuron i eae: Eragrostideae). Syst. Bot. 15: 515- 525. —— 1992. A pn of Chaboissaea (Po- ae: Eragrostideae). gre Philipson, W. R. 1934. The morphology of the lemma in 71. i 1974. Pp. P : W. D. Clayton, S. S. A. Renvoize (editors) Flora of Tropic al East Africa. EP (Part 2). Crown Agents for Oversea Governments and Administrations. Whitebriars Press, Lon 82. A numerical analysis by the Eragrosto- TEES (Gramineae). E" Bull. 37: —162. Pinna, M. C. C. de. 1. Concepts E tests of ote in the cladistic oim Cladistics 7: 367—394. © Prat, H. epiderme des graminées. Bul anato- mique et systématique. Ann. Sci. Nat., Bot., ser. 10, 14: 117-324. . reais saa singe e E epidermis in Zea mays. Ann. Missouri Bot. Gard. 35: 341—351. dies E de. 1985. 2 Bode method for deter- mining character polarity and its relevance to phylo- netic systematics. Syst. Zool. 34: 280—299 52. Die Grundlagen des natürlichen Sys- tems, der vergleichenden Anatomie und der Phyloge- netik. Akademische Verlagsgesellschaft Geest & Portig, eipzig i dee Y A survey of leaf-blade anatomy in EH des Bull 40: 737-744. A survey of og blade anatomy in grasses VIII. od Bull. 41: 323: & Sons, Chichester. 84. On homology. Biol. J. Linn. Soc. 22: 1991. Homologies and hierarchies: Problems solved and unresolved. J. Evol. Biol. 4: 167-194. Rueda, R 994. Systematics and exclusion of the genus Petrea (Verbenaceae). Ann. Missouri Bot. Gard 81: 610-652. Sattler, R. 1984. Homology—A continuing challenge. Syst. Bot. 9: 382-39 Scholz, H. 1993 Elytrigia arenosa (Gramineae)—Ein mitteleuropáischer Relickt-Endemit. Bot. Jahrb. Syst. 115: 351- gA Simon, B. K. A key to the Australian Grasses. 2nd ed. Queensland Department of Primary Industries, Bris- ban ÓN TR. the genus Olyra and the new segregate genus Parodi- olyra (Poac = nee Olyreae). Smiter Col Bot., Soni, S. L., P. E er & W. C. Bigelow. 1970. Elec- tron microprobe analysis of the distribution of silicon in leaf epidermal cells of the oat plant. Phytomorphology 20: 350-363. Stevens, P. F. 1991. Character states, continuous varia- tion, and phylogenetic analysis: A review. Syst. Bot. 16: 553—583. & F. O. Zuloaga. 1989. A revision of Sullivan, G. A. 1986. Remijia chelomaphylla (Rubi- aceae), a new species from Peru. Syst. Bot. 11: 298— 01 301. Tateoka, T., S. Inoue & S. Kawano. 1959. Notes on some grasses. IX. Systematic rdg eet of bicellular micro- hairs of leaf epidermis. Bot. Gaz. 121: 80-91. Terrell, E. E. € W. P. Wergin. 1981. Epidermal features and silica pravon in lemmas and awns of Zizania (Gramineae). Amer. J. Bot. 68: 697—707. Thiele, K. 1993. The holy grail of the perfect character: he cladistic treatment of ae data. Cladistics 9: 275-304. Thomasson, J. R. 1978. Epidermal patterns of the lemm in some fossil and living p and their oblea significance. Science 199: € 77. 1984. Miocene grass (Gramineae: Arundino- ideas) leaves showing external micromorphological and internal tasa features. Bot. Gaz. 145: 204-209 —— —. 1986. ma epidermal features in the North American species s of Melica and selected species of Bri- za, Catabrosa, Glyceria, Neostapfia, Pleuropogon, and Schizachne alae Syst. Bot 53-262. Tran, V. N. . Sur la valeur niccoliclgique des lem- mes de BS Bull. Mus. Hist. Nat. ser. 3, 128, Bot. 8: 33-57. Tsvelev, N. N. 1976. Grasses of the Soviet Union. Vols. I and II. (Nauka: Leningrad). English translation 1983. Oxonian Press, New Delhi Tutin, T. G. 1980. Gramineae. /n: T. G. Tutin, V. a Hey wood, N. A. Burges, D. M. Moore, D. H. Valentine, s. M. Walters & D. A. Webb (editors), Hora Europaea. vol. 9. e Univ. Press, Cambri Uphof, J. C. Th. 1962. Plant Hairs, nd ed. Gebriider- AS rur Nikolassee. Valdés-Reyn Hatch. 1991. Lemma micro- ol in e [pr (Poaceae). Sida 14: 531-549. Valls, J. F. M. 1978. A Biosystematic Study of Leptochloa With Special Emphasis on Leptochloa dubia (Gramin- 520 Annals of the Missouri Botanical Garden eae: Chloridoideae). Ph.D. Dissertation, Texas A&M University, College Station. Visser, N. C. & J. J Spies. 1994, Cytogenetic studies in - LA Ven (Poac eae: pae I. A taxo- 'al overview. S. African J. Bot. 60: 127-131. Wagner, G. “19894. The biologic val is concept. Ann. Rev. Ecol. Syst. 20: 51-69. . 1989b. The origin of morphological characters and the biological basis of homology. Evolution 43: 1157-1171. Watson, L. E. & M. J. Dallwitz. 1992. The Grass Genera of the World. CAB International. Univ. Press, Cam- bridge. Whang, S. S. € K. Kim. vie M phytolith morphology in ric ] Pl. Biol. : Wiley, E. O. 1981. Td ARA The Theory and Prac- tice A ents Systematics. John Wiley & Sons, Zuloaga. E 0. 1987. Systematics of New World species of Panicum er eae: Paniceae). Pp. 287—306 in T. R Soderstrom, K. W. Hilu, C. S. Campbell & M. E. Bark- worth (edite de Grass Systematics and Y dE Smithsonian ng Press, Washingt on, D. — ; ewicz. 1991. A revision of pum (Poaceae: anise tea Olyreae). Ann. Missouri Bot. 94.1. Gard. 78: 928— . Soderstrom. 1985. Classification of the outlying spec ies of New World d (Poaceae: Pan- iceae). Smithsonian Contr. Bot., 59. O. Morrone & J. iia 1989. Exomor- phe logie ‘al, anatomical, and cytological studies in Pan- icum validum (Poaceae: Panicoideae: Paniceae): tematic position within the genus. Syst. Bot. 14: 220— 230. Its sys- Appendix l. Species, authorities, puntry of origin of specimens studied. Order s siphaballe cal. Not all taxa are in subfamily Chloridoideae (see Discussion). Taxa scattered within Leptochloa with the generic abbreviation “D.” (for Diplachne) lack a valid combination in Leptochloa. All vouchers at MO unless an collectors, and of taxa 1 Pied 972; Clayton et al., 1974; Cope, 1985a, b; Elf- s & i. nedy-O’Byrne, 1957; Gibbs Russell et al., 1990: Gould, 1975; Judziewiez, 1990; Me Vaugh, 1983: Ortíz, 1993: Simon, 1993; Tsvelev, 1976; T 1980. . Hubb.: Tanzania; Gris pnway & Kanuri Bewsia biflora (Hack.) Gooss.: Strid 2911, Zambia; Dav- idse 5968, South Africa; Wild 1685, Zimbabwe. Boute loua curtipendula (Michx.) Torr.: Palmer 31445, U.S.A.; Tho- mas 107107, U.S.A.; Goodding 2463, U.S.A. Brachychloa schiemanniana | (Schweic Phillips: Balsinhas ozambique; Elffers S-n., Bur ur (US Chloris paniculata Scribn.: Stewart 260; Cocos blend (Costa Rica). € eed rerticillata Nutt. Snow 5779, U.S.A: Henderson 69-307, U.S.A S.A. Chondro- sum dim pi Kon Beetle M- 5579, Mexico; Garcta co; Thieret & hal ass 53242, U.S.A. P dsl e conspic ua (Forst. f.) Zotov us cunning- hamü (Hook. f.) Zotov: Gardner 3307. / Zealand. Chionochloa flavescens Zotov: Mason 2920, New Zealand. Cladoraphis cyperoides (Thunb.) S. M. Phillips (85): Dav- idse 33370, South Africa; Lavranos & Pehlemann 19640, Namibia; Goldblatt 4244, South Africa. ade i nosa (L. f.) S. M. Phillips: Davidse 33301, Bayliss 5732, South Africa; Davidse 3 Coelachyrum brev aes Hochst.: 28a, Saudi Arabia. Coelachyrum poiflo- Getahun 11906, E Ahiopia. Coelachyrum stoloniferum C. E. lubb.: Wieland & Gahair 4010, Somalia; Wieland 4098, aie Coelachyrum yemenicum (Schweinf.) S. M. Phil- lips: Ellis 2643, Botswana; Gilbert & Gilbert 2335, Ethi- opia; Leistner 1315, South Africa. Cynodon ie Vanderyst: (89) Gereau et al. 3490, Tanzania; Grant s Kenya; Crook 2115, Zimbabwe. Danthonia dominguensis Hack. & Pilg.: Zanoni et al. 37703, Dominican Republic. Desmostachya bipinnata (L.) Stapf: Fosberg 56909, Pakistan; Matamir 7/8, Egypt; Dwyer 13120, Saudi Arabia. Dinebra polycarpha S. M. hl Phillips: Ndegwa 639, Kenya. Dinebra retroflexa (Vahl) Panz.: Tanner 1430, Tanzania; Faden & Faden 74/741, Kenya: Schimper 1610, Ethiopia. Drake-Brockmania haareri (Stapf & C. E. Hubb.) S. M. Phillips: Mbano 5753. Tanzania; Drummond = Hemsley 2287, Tanzania; Bogdan 5434, Kenya (US). 'e-Brockmania somalensis Stapf: Heady 1815, e Faden & Faden 74/991, Kenya; Rob- ertson 1755, Keny Ectrosia zulliveri F. Muell.: Lazarides 4745, Australia; Blake 13574, Australia. Ectrosia leporina R. Br.: Clarkson 4892, Australia; Blake 18655, Australia; Clarkson 6059, Australia. Ectrosiopsis lasioclada (Merr.) Jansen: Lazarides 4787, Australia. pue Viris (L.) Gaertn.: Snow 5777, S.A.; Irwin et al. Brazil; Auquier 2142, Zaire. Fragrosiela hri vehi} Bor: Davidse & Sumithraar- achchi 899 si Lanka Ash 648, Ethiopia; Faden & Fa- den 74/967, Gouinia (s Presl) Scribn.: Stevens & Moreno 5, Nicaragua; M durs 9239, Mexico; Seymour 3305, E ~ Nicaragua. Habrochloa bullockii C. E. Hubb.: Webster T234, Tan- zania. Halopyrum mucronatum (L.) Stapf: Ash 708-A, Ethi- opia; Davidse & Sumithraarachchi 8224, Sri Lanka; hd shoff 6408. Ethiopia Heierae hne abortiva (R. Br.) Hughs: Must 1213 ralia; Clarkson 3700, Australia; Lazarides 9250, Aus stia a Indopoa pauperc ula (Stapf) Bor: eon rae 10287, India; Santapau 10269, India; Saldanha 17978, India Kengia serotina a .) Packer: FA 655, Czech Bsnibiln (MO 1743346), Czech Republic; Clemens Leptocarydion vulpiastrum (De Not.) Stapf: Ndegwa 655, Kenya: Crook 818, Zimbabwe; Van Jaarsveld 412, South Africa. Leptochloa aquatica Scribn. & Merr.: Hitch- exico; Arséne 5760-A, Mexico. D. caudata K. : Lewys Lloyd 1, Tanzania; Kingollah 39, Kenya. L. diinn (L.) Nees: Yao Kan et al. 79280, Japan (Chi- na?) Hsu 1392, Taiwan; Davidse & pep iecit 9006, Sri Lanka: L. rub pla Hack. ex Stuck.) Par- odi: Pensiero de Vegetti 2752, Argentina; Venturi 5724, Ar- gentina. L. ET Usa ) S. T. Blake: Anderson 738, Australia; Blake 5860, Australia; Hubbard 3217, Austra- lia. L. mo e "ns oy Adam 14030, Senegal; Adam id 75, Senegal; Adam 27799, Liberia. D. cuspidata Lau- : Giess & van dor Wait 12632, Namibia. L. decipiens n Br.) Stapf ex Maiden: Blake 12707. Australia. L. dig- S E a No? Volume 83, Number 4 1996 Snow 521 Lemmatal Micromorphology itata (R. Br.) Domin: Chippendale & Constable 19009, Australia; Clemens 21, Australia; Lazarides 3797, Austra- lia. L. divaricatissima S. T. Blake: Blake 10517, Australia; Johnson 410, Australia; Blake 19152, Australia. L. dubia (Kunth) Nees: Snow 5857, U.S.A.; Pringle 1027, Mexico; Renvoize & Cope 4240, Bolivia; Stuckert 17257, Argentina. D. eleusine Nees: Maguire 8398, South Africa; Giess 8422, Namibia; Brown & Shapiro 166, South Africa. L. fasci- cularis (Lam.) A. Gray: Snow 5811-A, U.S.A.; Davidse a al. 11647, Brazil; Wingfield 5867, Venezuela. L. fusca (L P. Beauv. ex Roemer & Schult.: Davidse & a achchi 9174, Sri Lanka; Leippert 4606, Namibia; Davidse & Sumithraarachchi 9119, Sri Lanka; Clemens 1615, Chi- na; Hubbard & Winders 6405, Australia. D. gigantea Lau- sen p (BRI); Anderson 704, Australia (BRI). L. longa jriseb.: Croat 16903, Panama; Hitchcock 672, Trinidad; A nin 2612, Trinidad. L. marquisensis (F. Br.) P. M. Pe- terson & Judz.: Wagner & Lorence 6222, Marquesas; Lor- ence et al. 6230, a: Perlman 10242, Marquesas. L. monticola Chase: Leonard 4751, Haiti (B); Ekman 3075, Haiti (US); Ekman 1576, Haiti (US). L. mucronata i : A, U.S.A.; Raveill 1813, Lanka; Davidse 7466, Sri wan 657. L. obtusiflora Hochst.: Mhoro 1868, Tanzania; De Wilde 6449, Ethiopia; Verdcourt 1101, Ke- nya. L. panicea (Retz.) Owhi: McGusker 229, Tanzania; Reekmans 8750, Burundi. L. panicoides (J. Presl) Hitchc.: Snow 5790, U.S.A.; Summers & Hudson 4596, U.S.A.; Nelson et al. 3213, Honduras. D. parviflora (R. Br.) Benth.: Lazarides 4400, Australia. L. rupestris C. E. Hubb.: Wood 2848, Yemen (BM); Pappi 2960, Eritrea (US). L. scabra Nees: Snow 5795-A, U.S.A.; Pohl & Dav- idse 12062, Honduras; Davidse 261 7, Trinidad. L. sp. nov. (Snow, in prep.): Davidse & Sumithraarachchi 9066, Sri Lanka. L. squarrosa Pilg.: Mwasumbi & Mponda 12671, Tanzania; Bon P Tanzania; Greenway 5096, Tanzania. L. uniflora Hochst. ex A. Rich.: Davidse & Ellis 5925, South Africa; ALT. et al. 14075, Tanzania; Ellis 2780, Zimbabwe. L. uninervia (J. Presl) Hitche. & Chase: Snow 5789, U.S.A.; McDaniel & Rimachi 23074, Peru; Garcia . Beauv.: Irwin et al. 11332, Brazil. L. Siscida (Scribn.) Beal: Snow 5817, U. P A.; Thurber 45, U.S.A.; Wright 2044, U.S setopkifa P M. Peterson & aan Decker 637. Marquesas Lintonia nutans Stapf: Faden & Faden 74-747, Kenya; Ndegwa 599, Kode Ngoni 319, Botswana. Lophacme digitata Stapf: Smook 1453, South Africa; Anderson 36, So Myriostachya wightiana puo ex Steud.) Hook. f.: Dav- dse Sumithraarachchi Sri Lanka; Davidse & Sumithraarachchi 9051, br pite Griffith 6621, India. Pilg.: Harley et al. 16293, ~. raudia reynauldiana (Kunth) Kena ex Hitchc.: 890113, China; Levine 1812, China; Maxwell 86-218, Thailand. Ochthochloa compressa cie Hilu: Dwyer 13928, Saudi Arabia; Mun Ethiopia. O. paucinervis (Nee ia; Leistner 2262, South Africa: Ellis 2767, South Africa. Orinus thoroldii (Stapf) Bor: Younghusband 147, Khama- jong. Oropetium aristatum (Stapf) Pilg.: Fay 5932, Central African Republic; Adam 31611, French Guinea; Adam Ax Sudan. Pogonarthria orne (Hack.) Hack.: Rodin 9073, Na- nibia "Goldblat 1901, lagoa Giess 11285, Namibia. neura bifora Napper: Greenwa J. Stretch s.n., ralia (CANB). Psilolemina jaegeri at S. M. Phillips: yis 10288, Tanzania; Green- way & Kanuri DE anzania; Bullock s.n. (MO acces- sion 1819125), Tanz Richardsiella tun Elffers & Kenn.-O'Byrne: Phipps & ET itzgerald 3227, Zambia; Webster A340, Zambia (two s in i sampled accession 2324532). pora pilum (R . Br) Connor & Edgar: Gardner 1372, New EE macrostachyum (Benth.) A. Camu Croat 30943, Madagascar; F osberg 48752, Aldabra Island ra a agus pedunculatus "ie: Dinter 3698, South Africa; Ellis 2700. i^» Africa (Nees) Stapf & C. Hubb.: poe 733, DE dh Ferreira F168, South ry ioe Brain 1242, Zimbabwe. Tripogon ma- jor Hook. f.: Jacques-Georges 22087, Sierra Leone; Bogdan 3932, Kenya; Thulin & Mhoro 3040, Tanzania. Triraphis andropogonoides (Steud.) E. Phillips: Smook ibbs-Rus- sell 2192, South Africa; Smook 3008, South Africa; Schee- pers 1313, South Afric bees. ase tent A. Camus: Gentry 11882, Madagasc Note did: in Re-examination of ¿AA obtusiflora suggests some hairs on the lemma are clavate or clavicorniculate (Mhoro 1868; Verdcourt 1101; both at MO). ichoneurd a 522 Annals of the Missouri Botanical Garden Figures 1-9.* —1. Acracne racemosa, —2. Apochiton burttii. —3. Bewsia biflora. —A, Brachychloa i naan —5. Chloris verticillata; from left to right. arrows illustrate sequential ontogeny of macrohair. —6. Cladoraphis ¢ cype- roides; bar = 5 pm. —7. Cladoraphis UR —8. Coelachyrum brevifolium; note crispate macrohairs. a Coela- chyrum poiflorum, with crispate macrohair ' Figures 1-72. SEM photomicrographs of lemmas of Leptochloa spp. and related genera. Unless indicated otherwise, scale bars equal 50 jm. Abbreviations: AELW = apical extensions of lateral walls of long cells; BM = oe mic icrohair; CC — cork « ell; DSLC = distally swelling long cell; MH = mac ian PL = ick : papillate long cell; le; PSC = papillate short cell; SC silic ‘a cell. For details see text. 524 Annals of the Missouri Botanical Garden Figures 19-27.* —19. Ectrosiopsis lasioclada. —20. Eragrostiella bifaria; note swollen imma ture cork cells. — 21. Gouinia virgata; prickles arising from long cells. —22. Habrochloa bullockii. panicoid microhair; bar = 5 wm. — 23. Halopyrum mucronatum. —24. Harpachne schimperi. —25. Heterachne abortiva: bar = 5 um. —26. Indopoa pauperculata; note swollen distal ends of long cells and swollen immature short cells: bar = 5 um. —27. Kengia serotina. Volume 83, Number 4 Snow 525 1996 Lemmatal Micromorphology 3 VEA Figures 28-36? —28. i nt vulpiastrum; note apical extensions of lateral T4 P long cells; bar m. —29. Leptochloa chinensis; note papillae occurring on distal swelling of long cell; ba —30. Leptoc hloa ciliolata. —31. Leptochloa coerulescens: chloridoid microhair, papillate long s with id ga papillae; bar = 5 pm. —32. Leptochloa digitata. —33. Leptochloa dubia. —34. olack eleusine, with clavicorniculate macrohairs. —35. EE fascicularis, with a few sporadic silica cells. —36. Leptochloa po macrohair arising from ell short c 526 Annals of the Missouri Botanical Garden Figures 3745.* —37. Leptochloa fusca; bar = —38. Diplachne gigantea, with pric kles arising from short cells; bar = 5 pm. es Tou ‘hloa ligulata. 40. P T monticola; note prickles arising from long cells and their conduplicate bases. —41. Leptochloa monticola with silica cells. —42. Leptoc hloa mucronata; macrohairs arising from distal end of long m (lemmatal apex towan d top). “3. Diplac hne muelleri; papillate short and long cells. . Leptochloa i "yi; note different prickle sizes. —45. Leptochloa neesii, with basally flattened macrohairs arising ae short cells; bar = 5 um. Volume 83, Number 4 Snow 1996 Lemmatal Micromorphology j ALA n 4 / T 7. i . á di Dig GOGO QG & LARS. M : ` Figures 46-54. — —46. Leptochloa panicea with intercostal macrohairs. —47. Leptochloa panicoides; note thick basal cell of microhair. —48. Leptochloa rupestris. —49. Leptochloa scabra. —50. Leptochloa squarrosa; note basally flattened macrohair; bar = 5 wm. —51. Leptochloa uniflora; note atypically shortened long cells. —52. Leptochloa virgata, with swollen distal portions of long cells, and cork cells with collapsing outer walls. —53. Leptochloa viscida; note thick basal cell of microhair. —54. Lintonia nutans, with prickles arising at base of dorsal awn. 528 Annals of the Missouri Botanical Garden Figures 55—63." —: 5. Lophacme digitata. —56. Myriostachya wightiana. —57. Neesiochloa n —58. Ney- raudia reynauldiana ah panicoid microhair. —59. Oe hthochloa compressa, showing hollow, basally severed mac te bar = 5 pm. —60. Odyssea mucronata; bar = 5 pm. —61. Orinus thoroldii. —62. Oropetium aristatum. —63. ea fleckit; prickles arising from long ce n and short cells. FURTHER STUDIES ON John A. Beutler,? Ada Belinda Alvarado- PHORBOL ESTER Lindner,? and Thomas G. McCloud? BIOACTIVITY IN THE EUPHORBIACEAE! ABSTRACT Six hundred and thirty-four organic extracts, including 36 previously untested genera of the Euphorbiac eae, were examined in a phorbol dibutyrate receptor binding assay. Phorbol bioactivity was newly detected in the genera An- un Blachia, Borneodendron, Dichostemma, Spirostachys, Tapoides, Trigonostemon, and Wetria. The activity of phorbol esters in the National solvent evaporated in vacuo at 40°C. All crude ex- Cancer Institute (NCI) primary anti-HIV screen tract samples were dissolved in DMSO at a con- (Gustafson et al., 1992; Erickson et al. centration of 10-20 mg/ml for initial examination, prompted us to investigate the taxonomic distribu- — then stored at —20°C. For dose response studies, Y Ke) Yu tion of phorbol esters, as measured by a simple these samples were further diluted in DMSO. competition binding assay. We have extended our previous study (Beutler et al., 1989) to detect the PHORBOL DIBUTYRATE BINDING ASSAY presence of phorbol esters in crude organic extracts ' he assay was performed as previously reported T (Beutler et al., 1989). Each extract was tested at an initial concentration of 100 pg/ml. If displacement at this concentration was greater than 60 percent, of Euphorbiaceae. Our aim has been to examine as many tribes and genera as possible to map the dis- tribution of phorbol ester bioactivity in the family. it was taken as a positive result. The definition was MATERIALS AND METHODS i : increased from the value of 50 percent displace- SAMPLES ment used in our previous report, due to exami- . nation of nonspecific. binding effects of the crude The plant samples for this study were collected F B under the auspices of the NCI Developmental Ther- apeutics Program. Critical plant determinations were made by Michael Huft, Missouri Botanica Garden, D. Doel Soejarto, Field Museum, and John Burley, Arnold Arboretum. The taxonomic frame- work adopted was that of Webster (1994), except for the distinction of two species of Wetria, as main- extracts. Initial examination and dose response were performed in duplicate, with at least four con- centrations used for dose response. The incubation was terminated by filtration over Whatman GF/B glass fiber filters in a Brandel cell harvester. A re- cent modification of the assay used for a minority of samples involved a reduction of the assay volume o 250 pl in 96-well polypropylene microtiter plates, with harvest onto GF/B glass fiber paper, which was then dried in vacuo and counted on a tained by Soejarto. Voucher specimens (see Appen- dix 1) are located at the Field Museum (Southeast Asia), the Missouri Botanical Garden (Africa), or the New York Botanical Garden (Americas) except where indicated in the appendix. Specimens with- Betaplate counter. . RESULTS AND DISCUSSION out numbers (s.n.) are not vouchered. Dried plant materials were processed by percolation in a 1:1 The cumulative total of active genera including mixture of methylene chloride: methanol, and the — our previous results (Beutler et al., 1989) are tab- ' We thank the staff and collaborators of the NCI Developmental Therapeutics Program for their assistance, especially Gordon Cragg for management of the plant collection contracts, Michael J. Huft, D. Doel Soejarto, and John Burle ey for pe erforming critical botanical detarinaliona, and Peter M. ee = helpful discussions. Research supported by the National Cancer ane S, under contract with SAIC Fre ? Laboratory of Dru :overy Research and Development, ed Therapeutics Program, Division of Cance Treatment Dinos à ei Cen nters, National Cancer Institute, Frederick Cancer Research and Development Cane Frederick, Maryland 21702-1201, U.S * Chemical Synthesis and Analysis Laboratory SAIC Frederick, NCI-Frederick Cancer Research and Development Center, Frederick, Maryland 21702- 1201, U.S.A. ANN. Missouni Bor. GARD. 83: 530-533. 1996. Volume 83, Number 4 1996 Beutler et al. 531 Phorbol Ester Bioactivity in Euphorbiaceae Table 1. Distribution of phorbol bioactivity by genus I1 7 1994), including previous results (Beutler et al., 89). Samples active/ Taxon samples tested I. jussa 1/199 nti esmeae 1/88 1/42 II. Oldfieldioideae 0/14 III. Acalyphoideae 4/169 30. Acalypheae 4/76 4/4 IV. Crotonoideae 65/118 34. Manihoteae 1/3 Cnidoscolus 1/2 38. Jatropheae 8/8 Jatropha 8/8 39. Codiaeae 10/13 Dimorphocalyx 2/2 Codiaeum 6/6 Blachia 2/2 40. Trigonostemoneae 9/9 Trigonostemon 9/9 42. Crotoneae 26/40 Croton 23/37 Eremocarpus 1/1 Fahrenheitia 2/2 44. Aleuritideae 11/21 Aleurites 1/1 Borneodendron 1/1 Cyrtogonone 5/5 Crotonogyne 2/2 Ta, s 2/2 V. Euphorbioid 83/134 45. Stomatocalyceae 1/10 Plagiostyles 1/3 46. Hippomaninae 46/77 Duvigneaudia 2/3 Excoecaria 8/8 Maprounea 5/8 Omalanthus 6/15 19/25 in 4/5 Stillingia 2/2 49. Euphorbieae 36/47 Anthostema 4/4 Dichostemma 6/6 Euphorbia 25/33 Synadenium 1/1 ulated (Table 1) according to Webster's (1994) ar- rangement of genera. Thirty-six genera that we had not previously examined were tested (Amanoa, An- thostema, Austrobuxus, Blachia, Borneodendron, Chamaesyce, Clei- Cephalomappa, Chaetocarpus, Dichostemma, Elaterios- dion, Clutia, Conceveiba, permum, Erythrococca, Koilodepas, Lingelsheimia, Monadenium, Neoscortechinia, Omphalea, Pausan- dra, Pera, Pimelodendron, Plukenetia, Pseudolach- nostylis, Ptychopyxis, Sampantaea, Sauropus, Savia, Sebastiania, Senefeldera, Strophioblachia, Synaden- ium, Tapoides, Trigonostemon, Wetria, and Zimmer- manniopsis), bringing the total number of genera tested to 111. Further species and plant parts of previously tested genera were also examined. Rep- resentatives of a total of 39 of the 49 tribes distin- guished (Webster, 1994) have been examined. In total, 106 species in the subfamily Phyllanthoideae, and 2 species of the Oldfieldioideae have been ex- amined with negative results. A single positive test of Antidesma nigricans Tul. (Phyllanthoideae) seeds (Beutler et al., 1989) prompted us to test a sample from each of the 26 species of Antidesma in the NCI repository. The lack of further positive results with these samples leads us to suspect that the sin- gle positive sample may have been misidentified. Seventy-six species of Acalyphoideae were tested, with 2 species of Wetria being positive (see below). Thirty-two of the 51 species of Crotonoideae ex- amined appear to contain phorbol esters (63%), while 43 of the 63 species of Euphorbioideae tested are active (68%). While the results of the *H-PDBu binding assay reflect only the presence of phorbol esters which bind to protein kinase C, there appears to be a good correlation between the bioactivity data and the chemical literature on phorbol diterpene distribu- tion. Our laboratory has demonstrated in three spe- cific instances (Homalanthus nutans (Forster) Pax (Gustafson et al., 1992), Excoecaria agallocha L. (Erickson et al., 1995), and Maprounea spp. (Beu- tler et al., 1995)) that the activity detected in our samples is due to conventional phorbol esters. Iso- lation of the compounds responsible for *H-PDBu binding activity in the newly identified genera (An- thostema, Blachia, Borneodendron, Dichostemma, Spirostachys, Tapoides, Trigonostemon, and Wetria) should be pursued to positively identify the com- pounds responsible for the observed bioactivity. These results indicate that phorbol ester bioac- tivity is primarily limited to subfamilies IV and V in Webster’s (1994) scheme. The only exception found in our data is that for two species of Wetria, which is placed in the Acalyphoideae. The overall pattern of distribution is consistent with previous chemotaxonomic data (Kinghorn, 1979) and lends biochemical support to Webster’s arrangement of genera. Further, our data are consistent with em- bryological data (Jensen et al., 1994), and legumin- like protein distribution (Kapil € Bhatnagar, 1994), 532 Annals of the Missouri Botanical Garden which both separate the Phyllanthoideae and Old- fieldioideae from the other three subfamilies. he observation of *H-PDBu binding activity in Wetria insignis and W. macrophylla, and the isola- tion of phorbol esters from Pycnocoma by Bergquist et al. (1989) may appear anomalous, but would be consistent with the proposed derivation of Croto- noideae and Euphorbioideae from Acalyphoideae if the biosynthetic capability to form phorbol esters arose in a common ancestor of the Acalyphoideae, Crotonoideae, and Euphorbioideae. Hecker and Adolf have perceived a similar chemotaxonomic pattern to that which we observed (Adolf & Hecker, 1977). It must be emphasized that no definitive work has been done on the biosynthesis of any phorbol ester in any plant. Such a study would fa- cilitate interpretation of the chemical data for tax- onomic or phylogenetic purposes. Alternative ex- planations for our pattern of results are that Wetria and Pycnocoma have been misclassified or mis- identified; however, we believe the identifications to be correct. It is notable, however, that Wetria and Pycnocoma are placed in different tribes, and that their close relatives such as Cleidion or Sampan- taea, and Ptychopyxis, respectively, have not dis- played *H-PDBu displacing activity. Patterns of diterpene occurrence and irritancy have been previously examined by Schmidt (1986), who analyzed the occurrence of all diterpene hy- drocarbon skeletons (e.g., crotofolane, lathyrane, casbane, etc.) in the Euphorbiaceae. Since that re- view, other novel diterpene skeletal types have been discovered, and known types have been found in previously unexamined genera (e.g., Kashman et al., 1993; Bernart et al., 1993). We conclude that phorbol bioactivity and phor- bol ester biosynthesis are more widespread throughout the genera of the Euphorbiaceae than has been previously reported. The chemotaxonomy of this family can be explored by application of the 3H-PDBu binding method to more taxa, and by bio- assay-guided fractionation to determine the chem- ical structures responsible for these results. The *H-PDBu binding assay allows rapid testing, generating semiquantitative data that can be used to prioritize isolation of the compounds responsible for the bioactivity. It is more selective than such assays as mouse ear irritancy and does not require the use of whole animals. The distribution of *H- PDBu binding activity may also be of use in as- sessing the potential toxicology of Euphorbiaceae used as phytomedicines. Literature Cited Adolf, W. € E. Hecker. 1977. Diterpenoid irritants and cocarcinogens in Euphorbiaceae and Thymelaeaceae: Structural relationships in view of their biogenesis. Is- rael . Ch . 16: Keine Bietogulst. .. H. Obianwu & B. Wickberg. 1989. Iso- lation and structure ponds doc of a novel phorbol derivative in an intramolecular diester macrolide. J. Chem. Soc., Chem. Comm. 183-184. Bernart, M. W., Y. Kashman, M. Tischler, J. H. Cardellina I & M. R. Boyd. 1993. Bershacolone, an unprece- dented diterpene prar from Maprounea afri- cana. Tetrahedron Lett. ae Re 4464. Beutler, J. A., A. B. Alvarado, T. G. McClou G. M. Cragg. 1989. Distribution of phorbol ester bioactivity in the Euphorbiaceae. Phytother. Res. 3: 188-192. ashman, M. Tischler, J. H. Cardellina II, G. N. Gray, M. Currens, M. E. Wall, M. C. Wani & M. R. 1995. A a of Maprounea triterpe- nes. J. Nat. Prod. 58: Erickson, K. L., J. A. Ec H. Cardellina II, J. B. MeMahon, D. J. foem d M. R. Boyd. 1995. A novel phorbol ester from Excoecaria pores J. Nat. Prod. -112. Dun: K. R., J. H. Cardellina II, J. B. McMahon, R. J. Guo J. Ishito oya, Z. Szallasi, N. E. Lewin, P. M. Blumberg, O. S. apad J. A. Beutler, R. W. Buck- ipee G. M. Cragg, P. A. € nonpromoting phorbol from the Sam lind plant Homalanthus nutans inhibits RE ‘killing HIV-1. de Med. Chem. 35: 1978-1986. a U., I. Vogel-Bauer & M. Nitschke. 1994. Leg- um minlike prota and the systematics of the Euphor- . Missouri Bot. Gard. 81: 160-179. Y A. K. Bhatnagar. 1994. The contribution of embxyaloes to the systematics of the Euphorbiaceae. Ann. Missouri Bot. Gard. 81: 145-159. Kashman, Y., M. W. Bernart, M. Tischler, J. H. Cardellina II & M. R. Boyd. 1993. Koumbalones A and B, new casbane diterpenes from Maprounea africana. J. Nat. Prod. 57: 426—430. Kinghorn, A. D. 1979. Cocarcinogenic irritant. Euphor- biaceae. /n: A. D. Kinghorn (editor), Toxic Plants. Co- lumbia Univ. Press, New York. Schmidt, R. J. 1986. Biosynthetic and chemosystematic aspects of the Euphorbiaceae and Thymelaeaceae. In: F. J. Evans e ia Occurring Phorbol Esters. CRC Press, Boc Webster, G. 1994. "agam of the genera and suprage- neric taxa of Euphorbiaceae. Ann. Missouri Bot. Gard. 3-144. Specimens of Euphorbiaceae examined. refer to the *H-PDBu screening Appendix The characters + and — i e for the specimen. Acalypha lancetillae Standley, Balick 1924 (NY) (—); Alchornea cordifolia (Schum. & Thonn 8575 (MO) (—); A 15094 (MO) (—); Anthostema aubryanum B: Pherson 15212, Wilks 2626 (MO) (+); Antidesma bunius L.) Spreng., Burley 68 (F) (=); Antidesma celebicum Miq., Burley 3558 ell. Arg., Soejarto 7375 (F) (—); Antidesma cuspidatum Muell. Arg., Soepad- mo 110 (F) (—); Antidesma ghaesembilla Gaertn., Vara- darajan 1528 (F) (—); Antidesma gymnogyn q offm., Soejarto 5878 (F) (—); Antidesma leptocladum Tul., Burley 173 (F) (—); Antidesma leucocladum Hook. f., Volume 83, Number 4 Beutler et al. 533 Phorbol pd Bioactivity in Euphorbiaceae Soepadmo 111 (F) (-); Antidesma leucopodum Miq., Soe- Schmidt 1137 (MO) (—); Antidesma menasu Miq., num Bl., Soepadmo 59 (F) desma pendulum H desma pentandrum (Blanco) Merr., Soejarto 7937 (F) (—); Antidesma petiolare Tul., Miller 3741 (MO) (—); Antidesma ry Shaw, Takeuchi 4444 (F) (—); Antidesma , Takeuchi 4388 (F) (— Antidesma salicinum Ridl., a 261, Ong 172 (F) (—); Antidesma sarcocarpum Airy Shaw, Takeuchi 7048 (F) (—); Antidesma tomentosum Bl., Burley 53 (F) (—); Anti- sma venosum Tul., Schmidt 1236 (MO) (—); Aporusa ar- borea (Bl.) Muell. Arg., Burley 1727 (F) (—): Aporusa plan- choniana Baill., Soejarto 5950 (F) (—); Aporusa prainiana 2 (F) (—) Austrobuxus nitidus e i F) (—); Bis- Eus sp., Frodin 2069 (F) E); TN castanicarpus (Roxb.) Thw., Soepadmo 308 (F) (—); Chae- tocarpus globosus (Sw.) Fawc. & Rend., Zanoni 45133, Garcia 2635 (NY) (—); d sp., Chafoar 4510 (F) (—); Claoxylon longifolium (Bl.) Miq., Burley 382 (F) (—); Cleidion lanceolatum Merr., Soejarto 7975 (F) (—); Clei- iaeum variegatum (L.) A. Juss., Harini 135 (F) (+); Con- ceveiba rhytidocarpa Muell. Arg., Beck 1049 (NY) (—); Croton argyratus Bl., Soejarto 6190, Burley 1468 (F) (+); Croton lechleri Muell. Arg., Boom 7792, Bennett 4531 (NY) (+); Croton leiophyllus Muell. Arg., Soejarto 6455 Del., Gereau 2576 (MO) , Spjut 11129 (b) (—); nsis Merr., Soejaria 6476 (F) (+); pe punctatus nee Beutler s.n. (+); Croton spp., Daly 56 (+), Williams 656 (+), Takeuchi 4508 (—) (F); ed none argentea (Pax) Prain, hes 13735 (MO) (+). Dichostemma glaucescens Pierre, McPherson ag! Di- bata 935 (MO) (+); Drypetes longifolia (Bl. Pax & K. Hoffm., éd 755 (F) (—); Drypetes madagascariensis (Lam.) Humbert & Leandri, Zarucchi 7459 (MO) (— petes parvifolia (Muell. Arg.) Pax & K. ; Drypetes paxii Hutch., Martin 509 (MO) (—); P 3029 (F) (—); Drypetes similis Hutch., Nemba 551 (MO) (9). lateriospermum m Bl., Burley 3175, Soepadmo 193, Frodin 2022 (F) (—); Endospermum diadenum did ) Airy Shaw, Burley 1450 (F) (—); Erythrococca ulugurensis A. R.-Sm., Gereau 2949 (MO) (—); Euphorbia laurifolia Juss., Mena 61 (NY) (+); Euphorbia maculata L., Beutler s.n. (—); Euphorbia tirucalli L., Harini 150 (F) (+); Excoecaria phillipensis Merr., Soejarto 6553 (F) (+). Flueggea virosa (Willd.) Voigt, Rulangaranga 7 (MO) Glochidion pomiferum Airy Shaw, Takeuchi 4076 (F) iz fimbric erl., Soejarto 6321 (F) (—); Synade + Homonoia riparia Lour., Ong 170 (F) (—); Hyeronima uell. Arg. ado IR (NY) (-). 07 (F) (~). _Lingaltheina f frutescens Pax, McPherson 15084 (MO) Macaranga conifera (Zoll.) Muell. Arg., Meijer 119104 (-); anga denticulata (Bl.) Muell. Arg., Stone 15911 (F) (—); Macaranga lowii King ex Hook. f., Burley 1397 (F) (—); Macaranga pleioneura Airy Shaw var. vel- utina Whitmore, Takeuchi (F) (—); Macaranga ram- iflora Elm., Stone 15849 (F) (—); Mallotus cuneatus Ridl., ie 791 —); Mallotus appr bd + (Lam.) ell. Arg., Soejarto 6230 (F) (—); Mallotus riciloides (Pers .) Muell. zn Burley 79 (F) (—); dor dis- coidea (Baill.) Webster, codes 2574 (MO) (—); Monaden- ium laeve Stapf., Kayombo 1027 (MO) (-). Neoschortechinia kingii (Hook. f.) Pax & K. Hoffm., 0 (F) (—); Neoschortechinia nicobarina Hook. f., Burley 2478 (F) (—); Neotrewia cumingii (Muell. Arg.) Pax & K. Hoffm., Madulid 6798, Burley 3587 (F) (—). (—); Omalanthus novaguine 4004 (F) (—); Omalanthus leschenaultianus A. Juss., Nich- olson 3 (c) (+); Omalanthus rotundifolius Merr., Nicholson 4 (c) (~). usandra sp., Daly 6034 (MO) (—); Pera arborea Mu- tis, Balick 3036, Balick 3066 (NY) (—); Pera benesis Rus- 2: ae 109 (NY) (—); Pimelodendron amboinicum , Takeuchi 4236 (F) (—); Plukenetia sp., Acevedo 7 691 (NY ) (—); Pseudolachnostylis maprouneifolia Pax var. maprouneifolia, Gereau 2749 (MO) (—); Ptychopyxis philippina Croiz., McDonald 3745 (F) (—). Sampantaea amentiflora Airy Shaw, Soejarto 8202 (F) (—); Sapium biloculare (S. Wats.) Pax, Wirt s.n. (d) (+); TAY laurocerasus Desf., Acevedo 2181 (NY) (+); Sap- armieri Huber, Daly 5634 (MO) x Sapium sp., Sogjarto 6193 (F) (+) Sauropus androgynous Merr., Har- ini 35 (F) (—); Sav a platyrhachis Bail, Harder E MO) (=) ionia brasiliensis Spreng., Saldias NY) (—); Securinega virosa € ) Baill. "s ay 8060 hio) —); pe ra cf. macrophylla Ducke, Beck 1 056 (NY) —) Spi. achys africana Sond., Muhoro 6252 (MO) (+); sling bu ica L., Spjut s.n. (b) (+); Strophioblachia nium TT Rulangaranga 179 (MO) (+). Tapoides pon (Merr.) Airy Shaw, Meijer 128772 (F) (+); Thecac El x A. Juss. var. montana Leandri, Zarucchi 7384 (MO) (—); Trigonostemon s tranus Hoffm., Burley 689 (F) (+); Tavdotie- mon viridissimus (Kurz) Airy Shaw, McDonald 3401 (F) (+ pine. + E Wetria insignis (Steud.) Airy Shaw, Takeuchi 6386 (F) (+); Wetria macrophylla (Bl.) J. J. Sm., Soejarto 7770 (F) +). immermanniopsis uzungwae A. R.-Smith (ined.), Kay- ombo 610 (MO) (—). © Voucher: Central Drug Research Institute, Lucknow, Ind (b) Voucher: World Botanical Associates, Laurel, Mary- land, U.S.A (c ) Voucher: Botanic Pli Smith College, Northamp- ton, Massachusetts, (d) Voucher: Dept. of ht & Evolutionary Biology, University of Arizona, Tucson, Arizona, U.S.A. NATIONAL BIOLOGICAL SERVICE: INTRODUCTION! P. Mick Richardson? “The establishment of a biological survey by the Department of Agriculture, under a recent act of Congress, should mark the beginning of a new era in the botanical field-work of the United States.” One might think that this sentence from an editorial in a botanical journal was written to celebrate the formation of the National Biological Survey in 1993 (National Research Council, 1993). It was actually written one hundred years earlier (Anonymous, 1896). What happened to the proposed biological survey in the intervening century? Perhaps it met resistance of the type that has confronted its belat- ed but necessary successor (Babbitt, 1995; Macil- wain, 1995). One day we shall hopefully have a full account of the part of biological history that con- cerns political opposition to the gathering of basic knowledge and information. 1e day before I sat down to write this introduc- tion I noticed a supplement to the journal Bio- Science on the current periodicals shelf in the li- brary at the Missouri Botanical Garden. It was entitled “Science and Biodiversity Policy" and con- tained a variety of articles about the development of a national strategy on biological diversity, arti- cles that began as papers delivered at the 1994 American Institute of Biological Sciences (AIBS meeting in Knoxville, Tennessee (Mooney & Ga- briel, 1995). I immediately wondered whether we ar had been scooped and the Missouri Botanical Gar- den symposium proceedings rendered redundant. However, my fears were unfounded: the two sym- posia are largely complementary, and, in fact, the BioScience publication makes an interesting but rather long introduction to our symposium. e papers are discussed here in the order in which they were presented on the Saturday of the Systematics Symposium. Ronald Pulliam of the Na- tional Biological Service (NBS) began the sympo- sium by outlining the history of the NBS agency in an excellent talk entitled “The National Biological Service: Building a National Partnership.” He de- scribed how the agency came into being as the Na- tional Biological Survey and metamorphosed under political pressure into the National Biological Ser- vice. S works with others to provide the scientific understanding necessary to support the sound management and conservation of the nation's biological resources. The paper is not reprinted in these proceedings, but his paper in the BioScience volume (Pulliam, 1995) describes the birthing pro- cess 0 ter the coffee break, Meredith Lane (University af Kansas Natural History Museum and National Science Foundation), in a paper entitled talked about how these collections have related to society at large during a long and evolving history. She urged an abandonment of personal competition in “Roles of Natural History Collections,” order to complete a compilation of the earth’s biota. The morning session was completed by Anne Fron- dorf, a second speaker from NBS. In the paper “Systematics Information as a Central Component in the National Biological Information Infrastruc- ture,” she and coauthor Gary how NBS is working with various partners to make aggoner describe data more easily available to those who need it when they are making decisions on resource man- agement. The Internet is a powerful tool and es- sential for the distribution of information that is being generated and/or maintained in a range of biological fields and institutions. ter lunch we were treated to an update in our current understanding of the diversity and relation- ships of microscopic forms of life. Diana Lipscomb (George Washington University) gave a paper enti- tled “Biodiversity and Microorganisms,” published here as “A Survey of Microbial Diversity.” A mul- tikingdom system is necessary not only to accom- modate the increasing diversity of microbes re- vealed by new technologies (my Grammar School biology teacher, Mr. Malkin, frequently informed us that molecular biology and the electron microscope had complicated and thus ruined the study of bi- ! This and the five articles that follow it are the proceedings of the 42nd Annual Systematics Symposium " the | Missouri Botanical Garden, National Biological Survey (now National Biological S The symposium was held 6-7 Oc tober 1995 at the Missouri Botanical Garden in St. 2 n ve 3 3 m. = = = e Y y Z © 3 = at T a S ao} PE a cc "E => a cA c S 5 B Uu 3 O = ^ 9) = Z 3 a E E z 5 =: ervice; see Nevling e is Issue). S , Missouri, U.S r gr iini number Dun earn I elp in organizing and ipei ips = podra Yevonn Wilson-Ramsey for her fine illustration of three Papaver species for the sym- m brochure, and the symposium registrants for being such a trouble- ei and pleasant group. 2 Missouri Botanic al Garden, P.O. Box 299, St. Louis, Missouri 63166, U.S.A. ANN. Missourt Bor. GARD. 83: 534—535. 1996. Volume 83, Number 4 1996 Richardson 535 Introduction ology), but also to replace the simple four- or five- kingdom systems used in teaching, and in my opin- ion, misleading, students of biology. This paper is an excellent introduction to recent developments in this field and will be especially useful to those of us who have allowed higher organisms to fully oc- cupy our minds for too many years. The next paper, unfortunately not published here, was given by Rahmona Thompson (East Central University) and was about “The Flora of North America, the Na- tional Biological Service, and Electronic Technol- o e warned botanists that we must be aware of changing methods of communication and we must also keep up with the revolutionary changes taking place in the publishing process, emphasizing the use of hypertext and the flexibility of digital formats. The final talk of the afternoon reflected the international aspects of the day’s program. Nation- als of six countries were in the audience of 350 participants. The paper, entitled “An International View of National Biological Surveys,” was delivere by Jorge Soberón (CONABIO), and the published form is coauthored by Jorge Llorente and Hesiquio Benítez. The natural history collections in indus- trialized countries hold large amounts of data in the form of labels on specimens collected in developing countries. The authors advocate a policy of sharing this data in order that the scientists in developing countries will have access to the existing informa- tion about their national biota. This is not a com- pletely straightforward task but requires the inter- national negotiation of agreements for the use and ownership of data. A thought-provoking after-dinner address enti- tled “The Last Species” was given by Lorin Nevling (Illinois Natural History Survey). It was accompa- nied by more than a hundred slides. The words and pictures together earned the deliverer a standing ovation. You must read the article for yourself to see what he said, but the gist of it is that he chal- lenged the audience to widen their views of their responsibilities as educators if we would like the world to be still recognizable in the year 3000. Besides the seven formal papers described above, three other significant events took place dur- ing the weekend of the symposium. First, the Mul- ford Commemorative Luncheon was held on Friday, 6 October 1995, in the Spink Pavilion. A sandwich buffet was followed by two short papers concerning female botanists. The first paper, by Carol Prentice, a botanical historian living in Boise, Idaho, was on the subject “A. Isabel Mulford: Pioneer Botanist” and concerned an interesting woman who taught at Vassar College, gained the first-ever Ph.D. degree awarded by Washington University for her work on Agave undertaken at the Missouri Botanical Gar- den, undertook three expeditions out West in the 1890s, taught high school for a few years, and then seems to have vanished into obscurity until her death. This remarkable history will be forthcoming as further details about Isabel Mulford's life are un- earthed. The second paper was more general. It was entitled “Botany: A Place for Women in Science” and was given by Barbara Ertter (University of Cal- ifornia-Berkeley), who also plans to publish an ex- panded version of her talk. The second event took place at lunchtime on Saturday, 7 October 1995. This was the dedication of the Edgar Anderson Me- morial Area, a section of the Garden planted with various taxa of Buxus. Charles Heiser (Indiana Uni- versity) spoke about some of the exploits of Dr. An- derson, with whom he had done his Ph.D. work. At the third event, which took place immediately prior to the talk delivered by Lorin Nevling referred to above, an award was presented to John Dwyer in recognition of his services to taxonomy at Saint Louis University and the Missouri Botanical Gar- den. Altogether, the 42nd Annual Systematics Sym- posium was an interesting, exciting, and historic meeting. Literature Cited Anonymous. 1896. Editorial. Bot. Gaz. 22: 57-58. Babbitt, B. 1995. Science: ar "E me se chapter of conservation history. Science 267: 1955. Macilwain, Tu Sc iento pal to al against e 373: 646. Mooney, H. & C. 1995. Toward a national strategy on biological diversity. Preface in BioScience (The Science and Biodiversity Policy Supplement). National Research Council. A Biological Survey for the Nation. National Academy Press, Washington, D.C. Pulliam, H. R. 1995. The birth of a federal research agency. BioScience e Science and Biodiversity Pol- icy Supplement): S91-9 ROLES OF NATURAL Meredith A. Lane? HISTORY COLLECTIONS! ABSTRACT atural history collections have always contained a wealth of data: genetic and phylogenetic information stored as an hee D part of the samples of organisms themselves, and biogeographic, ecological, and biogr: aphical information ored in the labels that are affixed to them. Together. a preserved organism and its label are a scientific specimen et has great intrinsic value. Separately, the label is a piece of paper with meaningless insc riptions upon it, and the plan spider, microbe, mushroom, or bird, though carefully preserved, is just so much dead organic matter, Natural Vimus collections are the repository of the vouchers for the documentation of what we know about the diversity of living things—what species exist and where, what their habitat requirements are, what ecological associations they have with other species, what useful biochemical products they might generate, and who collected them and has studied them Before the advent of computers, natural history collections were physical databases from which geo ographic or ecolog analyses and reports could be extracted by human visitation and transcription, usually a laborious and time- “consuming task. However, such analyses are invaluable for land-use planning, pharmacognosy, conservation biology, range man- agement, forestry, agriculture, and a host of other applications, including scientific studies of the ecology and ii of the species being examined. Computerization of label data md such reports on distribution and ecology of species more readily available to potential users; they add value to the data. Interconnecting the databases brings robustness o the information that natural history collections can UE to ae making bodies; appreciation of robust data will i appreciation of the collections from which those data were taken. Interconnec tivity requires that collec- tions personnel abandon competition in favor of achieving a common goal: the dise 'overy and desc 'ription of the world's — e z = o iota When I was first asked to write this paper, I was of plants. It is true that botanists have been slow told that the topic I would cover should be “the to give up the name “herbarium,” a word that de role of natural history collec ‘tions in relationship to Tournefort and Linnaeus used to mean a collection the National Biological Service." I have taken the — of preserved specimens of plants. 1 suspect that we liberty, while keeping biological surveys in mind, botanists might much more readily have become natural historians of plants had our collections con- a of expanding my topic to include the roles (plural of natural history collections, not only in relation tinued to be called, as they were before de Tour- to the NBS but also in relation to society at large. — nefort and Linnaeus (Radford et al., 1974), hortus and to each other. siccus or hortus mortus! However, 1 for one consider I begin with a broad-brush statement about the herbaria to be natural history collections, and refer activities of biological surveys in general, and their to them as such except in the discussion of the relationship to collections. Next, I touch on the his- history of the two sorts of collections. tory of natural history collections—what their roles have been in the past, and how those roles have Missions OF BIOLOGICAL SURVEYS changed (or not) over the years. Then, I present the SERRA "ERU X t : : IDENTIFY NATURAL (LIVING) RESOURCES perceived roles of natural history collections as they stand and as I think they need to become. Most biological surveys around this country have Finally, I come full circle to the relationship be- as their charge the documentation of the biota of tween natural history collections and biological sur- — the state or other geographical entity to which they veys, and the National Biological Service in partic- belong. The political entities that established the ular. surveys expect them to provide information to pol- Throughout, when I use the term “natural history icy-making bodies that will allow for “better deci- collections" I mean collections both of animals and sion-making" with regard to those living resources ' Craig Freeman assisted with many aspects of this paper, John Simmons pointed me to literature about the histo ory of collections, and Vivian Nolan retrieved it from various libraries. incluc ling those of the National Science Foundation and the sala Institution. I thank = the people and the libraries. 2 Uni of Kansas Natural History Museum, Bridwell Botanical Laboratory, 2045 Constant Ave., Lawrence, Kansas 66047- 3729, U.S.A. loi, edu). For communication purposes only before 30 June 1997, Division of Environmental Biology, Suite 635, National Science Foundation, 4201 Wilson Blvd.. Arlington, Virginia 22230, U.S.A. (malane@nsf. gov) ANN. Missourt Bor. GARD. 83: 536—545. 1996. Volume 83, Number 4 1996 Lane 537 Natural History Collections (as defined in human economic terms). To be able to provide the information requested, biological surveys personnel must extract it from the litera- ture, or perform original research (C. C. Freeman, pers. comm.; Cameron, 1929; Kim & Knutson, 1986; Rautenbach & Herholdt, 1990). PERFORM TAXONOMIC/ECOLOGICAL STUDIES It would be desirable, from the political stand- point, for the literature to contain reports of studies that have already been done that would answer the questions of the legislatures. Very frequently, this is not the case, and biological surveys personnel undertake original investigations. Often, the requests for information involve ques- tions about the ecology of certain organisms. The ecological research, in turn, often involves ques- tions about the taxonomy and phylogeny of those organisms. The answers to these questions rest squarely in natural history collections that the bi- ological surveys personnel may maintain them- selves or have access to in natural history museums (Chernoff, 1986; Pettitt, 1994). f there is no collection, the first step in the re- search is to make one. Even if the consulted lit- erature does contain a taxonomic, phylogenetic, and ecological analysis of the organisms of interest (so that the biological survey person does not have to do original research), that study would have been based on natural history collections that voucher and document the work of the author(s) of the paper (e.g., Yochelson, 1969; Conference of Directors of Systematics Collections, 1971; Pettitt, 1994). EXAMINE ENVIRONMENTAL IMPACTS Biological surveys often examine the environ- mental impacts of human actions (Kim & Knutson, 1986; Pettitt, 1994) within their own purview or those of neighboring states. Again, it is natural his- tory collections that document the spread of intro- duced taxa and decline of native species, the cur- rent distribution of taxa, and the relationships among organisms, which live, not in clades but rather in interconnected, interoperable habitats. The phylogenetic studies that are done by system- atists using the same collections provide the his- torical context for the evolutionary emergence of the organisms within those habitats. DEVELOP MONITORING, MANAGEMENT, AND CONSERVATION PROTOCOLS Within museums lie solutions to problems in conserving natural resources (Shropshire & Shrop- shire, 1991). The biological surveys are often ex- pected to develop plans to evaluate, monitor, and mitigate all manner of ecological phenomena. Their reports are vouchered and documented by speci- mens placed in natural history collections; without those vouchers, the facts that their studies have un- covered are in question. These protocols must be developed in light of collections; otherwise there is no basis for the pursuit of one management method over another (Conference of Directors of Systemat- ics Collections, 1971; Rautenbach & Herholdt, 1990; Pettitt, 1994). RELATIONSHIP OF BIOLOGICAL SURVEYS TO ATURAL HISTORY COLLECTIONS From these examples, it is clear that natural his- tory collections are fundamental to all that biolog- ical surveys do. Where there is no collection, the first job of a survey is to form one. Such was the situation in 1804-1806, when President Thomas Jefferson, having just acquired the huge Louisiana Territory for the fledgling United States, sent Meri- wether Lewis and William Clark and their compan- ions to make a survey of the headwaters of the Mis- souri River and territories thereabout, to include observations of the minerals, soils, climate, peo- a and animals in their diverse kinds, as well as . the dates at which particular plants put forth or ides their flowers or leaf, times of appearance of particular birds, i i 1969). Jefferson was himself a naturalist of note, who kept detailed records of the requirements of many sorts of plants that he attempted to grow in his experimental garden, and of observations of the Eq of animals on his plantation (Martin, 952). Lewis and Clark took their charge from the dd seriously, as it was meant, and brought back bales of specimens of various sorts of organ- isms, sometimes at the risk of life and limb (Cut- right, 1969). This leads us to the discussion of the history of natural history collections. I shall in a moment return to Lewis and Clar HisTORY OF NATURAL HISTORY COLLECTIONS The geneses of natural history collections are, 1 have found, in general understood differently by botanists and zoologists. The instigation for col- lecting plants, and the simple method of their pres- ervation, led to a different beginning for herbaria than for collections of animals, for which specimen- conservation techniques development is still in full swing (Ouellet, 1985; Hawks, 1990). The origins of natural history collections extend back through Western history to Ptolemy's academy 538 Annals of the Men Botanical Garden in third-century B.C.E. Alexandria, which was a palace that contained botanical and zoological gar- dens for teaching purposes, and cloisters and lec- ture rooms for students and teachers (Bateman, 1975). Pliny's first-century C.E. Historia Naturalis not only describes this academy, but g ence its name (Porter, 1991). The d oan of collecting between that time and the Renaissance, unfortunately, is one of war, looting, and plunder, vives our sci- but, by the fifteenth century, palaces in Europe contained collections of books, art (much of which had natural subjects), and the odd natural oddity (narwhal tusks interpreted as unicorn horns, for ex- ample) (Ritterbush, 1969; Bateman, 1975) ZOOLOGICAL COLLECTIONS Through the sixteenth century and well into the seventeenth, that con- tained shells, fossils, minerals, casts from nature, and botanical and zoological art, became a common possession among European nobility (Alexander, 1979; Impey & MacGregor, 1985; Hooper-Green- hill, 1992). Some of these cabinets were actually whole rooms, others were truly pieces of furniture designed to contain objects that represented the en- tire world in miniature. These objects, however “cabinets of the curious,” neatly stored, were not scientifically arranged, nor were they available to the public for study—they were private collections amassed for marily aesthetic, purposes (Ritterbush, 1969). Al- though I do not know this, I would wager that the owners entered into collection- building competi- rivate, pri- tions: “mine is bigger than yours”—the same syn- drome that is perpetuated among some museums and collectors today. The first of the public museums of natural history (which included, from its inception, two herbaria was that in Paris, founded in 1635; the first uni- versity museum was established at Basel in 16 (Bateman, 1975). The Ashmolean Museum at Ox- ford University, established in 1683 (Alexander, 1979; Lintz, 1991), seems to be the first to have official curators, a catalog, and a set of regulations — for visitors. These early museums, however, were more of art than of actual natural objects, because of the difficulty of preservation of animals. When Charles Wilson Peale, an American, worked out the use of arsenic in taxidermy in the late 1700s, that obstacle was overcome, and natural history collec- tions began to acquire, for the first time, long-last- ing collections of birds and mammals (Porter, 1991). Peale’s Philadelphia Museum, started in 1786, does not by many years postdate the establishment of the British Museum of Natural History in 1753 Lintz, 1991) or the first formal natural history col- lection in North America, the Charleston Library Society, started in 1773 (Alexander, 1979). Eve the venerable Paris Museum only began to acquire birds at the very end of the eighteenth century; by 1793, that collection included 493 skins. —. HERBARIA Modern herbaria, in the post-Renaissance sci- entific sense, have a history that is at least a cen- tury longer than zoological collections. If we accept herbaria as natural history collections (which in this day and age they certainly are), then the oldest formal natural history collection for scientific pur- poses of any sort (and this, most zoologists do not know) is the oldest herbarium in existence, which dates from 1523 (Ogilvie, 1985). The herbarium of Gherhards Cibo dates from 1532 (Radford et al., 1974), and the herbarium of the University of Pad- ua, begun in 1545, was the first institutional her- barium (Shetler, 1969), and therefore the first uni- versity natural history collection. In the New World, the establishment of the first herbarium (at Win- ston-Salem, North Carolina, in 1772) predates by a year the oldest zoological collection on this conti- nent (Shetler, 1969; Porter, 1991). The original use of the word “herbarium” is much older even than the sixteenth century. Her- baria once were rooms in medieval monasteries, in which were kept, usually hanging from the rafters, bundles of dried herbs that were to be used for flavoring food, for counteracting mildew and body odors in linens and clothing, and, of course, for medicinal purposes. For an image of such a room, think of the garden and monastery in Franco Zef- ferelli's lush cinematic production of Shakespeare's Romeo and Juliet —in the herbarium, Juliet is pre- sented with the sleeping draft that had been com- pounded from the herbs hanging overhead. In ad- dition, these examples of dried herbs would have been used in teaching younger monks about the plants and their uses. Thus, the monastery herbar- ium was both the forerunner of the modern teaching collection of plants, and continuation of the *herb- alist" period of botanical history (Radford et al., 1974) urpose in painting this sketchy history of natural history collections is three-fold: First, that the term "natural history collections" as a desig- nation of scientific resources should be understood e to include herbaria; we should not use the phrase “natural history collections and herbaria” to refer to the collections enterprise, because that distinc- Volume 83, Number 4 1996 Lane 539 Natural History Collections tion sets us apart rather than bringing us together. Second, that biological surveys have always made and relied upon natural history collections. Third, I stress the connections between herbaria and col- lections of other sorts of organisms because it is the history of herbaria that is rooted most firmly in the service to society ethic (i.e., the medicinal and ag- ricultural uses of plants). It is this linkage of her- baria, and thereby collections of other sorts of or- to societal service that will serve systematics well as it grows into the roles outlined for it in Systematics Agenda 2000. ganisms, ROLES OF NATURAL HISTORY COLLECTIONS IN THE PAST Originally, natural history collections made it possible for their viewers to have some notion of the biota and artifacts of distant places that they themselves could not visit (Lintz, 1991; Hooper- Greenhill, 1992). Collections were made for enter- tainment value as “curiosities” (Impey & Mac- Gregor, 1985), and hidden from competitors until their value could be estimated. Plants and shells were collected for their aesthetic value; the early collections of objects were valued originally as the subjects of paintings and drawings, rather than for themselves (Ritterbush, 1969). Collections were made because interest in the natural world was a major preoccupation of Renaissance learning (Im- pey & MacGregor, 1985), and, according to Albert Bickmore (first director of the American Museum of Natural History), for “teaching our youth to ap- preciate the wonderful works of the Creator” (Al- exander, 1979). The education function evolved lat- er into a view of nature as an open book, from which “the Birds and Beasts will teach thee!” as stated the tickets to Peale’s Philadelphia Museum (Lintz, 1991). Only gradually were repositories for the natural history collections made by survey expeditions such as that of Lewis and Clark identified and estab- lished. Despite his intense interest in natural his- tory (Martin, 1952), and his charge to Lewis and Clark to make collections of the organisms they en- countered, Jefferson (with, for him, unusual lack of foresight) did not make provision for the storage and curation of the specimens thus acquired. Many of the specimens that Lewis and Clark collected were dispersed (Porter, 1991); only a few were re- turned to an appropriate U.S. repository. Though the Congress continued to send out many more ex- peditions that collected specimens during the next several decades, it was not until 1879 that the of- ficial U.S. repository for such collections was es- tablished: the United States National Natural His- tory Museum (Cowan, 1969; Porter, 1991). Meanwhile, some of the states had initiated natural history collections. Here, New York is a telling ex- ample: The New York State Natural History Survey was started in 1836, and that Survey generated a Cabinet of Natural History (that is, collections) as its first task. The cabinet was physically located in Albany in 1843, seven years after the institution of the Survey itself (Porter, 1991). Only after the col- lections were made were ecological studies insti- tuted. In the nineteenth century it was recognized that collections must be made because they were critical to the progress of science, but provision for the conservation and preservation of the specimens was a harder sell. Today, it is not only the latter but also the former that must be promoted, although natural history collections today occupy prime real estate: for examples, the California Academy of Sci- ences in Golden Gate Park (San Francisco), the Smithsonian Institution on the National Mall (Washington, D.C.), the Field Museum of Natural History on Lakeshore Drive (Chicago), and the Mis- souri Botanical Garden on Shaw and Tower Grove Avenues (St. Louis). The taxpayers and contributors that provided this real estate, and the curators who through the aan have built the collections, de- serve an appropriate return on their investments (Cowan, 1969; S Allmon: 1994; Shetler, 1995). PRESENT ROLES OF NATURAL HISTORY COLLECTIONS Today, the roles of natural history collections, as is often stated in the literature, are two-fold: re- earch (Lemieux, 1981; Edwards, 1985; Lintz, 1991; Stansfield, 1994), sometimes called the “in- ner museum function,” and education, the “outer museum function” (Humphrey, 1991; Allmon, 1994). Some authors add service to this list as a third role (Laerm & Edwards, 1991; Pettitt, 1994). The things that natural history collections do, how- ever, may be summarized in a few statements. Of course, it is possible and reasonable to expand this list into its myriad component parts (as, for exam- ple, in Systematics Agenda 2000), and there are contexts in which I do, too. However, for the pur- poses of this paper I will use the “executive sum- mary” form. Natural history collections record the world’s bi- ota in space and time, and document what we do and do not know about that biota (Michener et al., 1970; Conference of Directors of Systematics Col- lections, 1971; Lamanna, 1976; Dessauer & Haf- ner, 1984; Duckworth et al., 1993; Anonymous, 540 Annals of the Missouri Botanical Garden 1994; Federal Biosystematics Group, 1995). Further, natural history collections are the funda- mental and indispensable resource for biological surveys, which study living organisms to under- stand ecosystem dynamics and conservation of liv- ing diversity (Irwin et al., 1973; Edwards & Grotta, 1976; Kim & Knutson, 1986; Duckworth et al., 1993; Anonymous, 1994; Federal Biosystematics Group, 1995). Natural history collections voucher economically important organisms, research in sys- tematics, population biology, ecology. genetics (and a host of other fields), and provide specimens and knowledge for education and exhibits (Michener et al., 1970; Irwin et al., 1973; Lemieux, 1981; Che noff, 1986; Duckworth et al., 1993; Allmon, 1994; Anonymous, 1994; Federal Biosystematics Group, 1995). Natural history collections are the basis for public and formal education programs (Lemieux, 1981; Allmon, 1994; Anonymous, 1994; Federal Biosystematics Group, 1995). Natural history col- lections are irreplaceable assets of the greatest val- ue (Brain, 1990 It is this repository-of-knowledge function, I sus- pect, that keeps most systematists constantly de- fending and promoting natural history collections. We understand, and we must constantly be vigilant that others understand, that human beings learn about new things only in the context of what they already know. Natural history collections are the context for what we already know about the diver- sity of living things on this planet. Collections are the vouchers for the knowledge to be passed to suc- ceeding generations (Lamanna, 1976). It is esti- mated that that knowledge is exceedingly skimpy: Perhaps only 1 to 5, or at most, 15 percent of the species on Earth have been apprehended by sci- ence. In our research function as the describers of new species, we despair of our ability to discover and describe the remaining 85 to 99 percent (which means tens of millions to a hundred million or more species) in the 40 or so years we have (if we are lucky) before they are mostly destroyed (Anony- mous, . But think! How much more difficult would that task be without the comparative re- search collections we already have of the things that are known to science? Thus, natural history collections personnel have several goals that must be reached. First, we must preserve against the forces of entropy the heritage that is represented by the some two billion speci- mens (Duckworth et al., 1993) in natural history collections around the world (approximately 400 to 500 million in the U.S. alone). This goal requires constant improvement of our techniques of speci- men conservation (Lemieux, 1981; Hawks, 1990), because many if not most of these specimens could not be collected again—too many habitats have been destroyed, and the cost of collecting expedi- tions has become too high to replicate something someone else has already done. Second, we must continue to collect, to gather new knowledge, and to process that knowledge for the benefit of society and biodiversity itself. Third, and most importantly, we must constantly educate—not just youth, the public, and politicians—but also ourselves. As natural history collections curators or collec- tions managers or researchers, we must school our competitive instincts into cooperative ones. No lon- ger should the goal be to demonstrate that Kew has more specimens than Paris, the American Museum more than Berlin, my own herbarium in Kansas more than Oklahoma's or Colorado's; in short, to demonstrate that: “mine is bigger than yours.” We must teach ourselves better public relations (Grove, 1970; Irwin et al., 1973), political maneuvers (bi- ologists and particularly systematists are especially lax in this area), and yes, even better business practice (Malaro, 1994). We must develop and fol- low long-range plans and careful acquisitions pol- icies (Hoagland, 1994; Malaro, 1994) that are co- ordinated among museums to avoid duplication of effort. We are together in a race against time: *Gen- tlepersons of Collections, rev your engines!" But first, examine your motives. There is no time for the pettiness of competition in this race. Our common goals must be to partition the work of diversity dis- covery and collection (Anonymous, 1994), conser- vation and preservation (Hawks, 1990; Herholdt, 1990), and become much better ambassadors both for the work we do, the collections within which we work, and the uncounted millions of species we serve. Those are lofty, admirable, and incredibly diffi- cult goals. But how do we attempt them, and from whence will the funding come? I believe that we have a start. It is a beginning that needs improve- ment and modification, not to mention a “pedal to the metal," but a beginning nonetheless. Twenty to twenty-five years ago, a number of re- ports and working papers produced by committees of the fledgling Association of Systematics Collec- tions and similar bodies (Manning, 1969; Confer- ence of Directors of Systematics Collections, 1971; Edwards & Grotta, 1976; Humphrey & Clausen, 1977) suggested that computerization for collec- tions management was a direction that natural his- tory collections ought to take. During the interven- ing two decades, a number of things have happened in this area: (1) quite a number of natural history collections have digitized at least part of their spec- Volume 83, Number 4 1996 Lane Natural History Collections imen data, (2) the realization of the need for data standards has occasioned the development of some of these within certain collections communities (Bi- ological Collections Data Standards Workshop, 1992), (3) a number of monographers have realized the value of computerizing the data of the speci- mens they borrow, (4) there have been workshops about natural history collections computerization and networking, and (5) the Research Collections in Systematics and Ecology program of the National Science Foundation has instituted a special cate- gory for collections computerization that supports several database-development projects around the country. Most curators and collections managers now recognize the value of digitized collections in- formation, and are willing to expend effort to com- puterize (Owen, 1990; Allen 1993; Cohn, 1995). owever, progress in this area has been slowed by two factors: (1) the learning-curve difficulties ex- perienced by collections curators and managers in trying to understand fully relational database man- agement systems and Internet connectivity, and (2) the lack of cooperation and agreement within the larger community on database structure, semantics, and syntax—that is, what are collectively called standards, which are important especially as we in- terconnect our databases. We must get past these barriers. As Shetler (1995) has so aptly said, “If our generation doesn’t figure out how to provide better access to the information stored in our ex- isting collections, then the next generation may not be able to defend keeping these collections.” INFORMATION MANAGEMENT NOT ONLY FOR COLLECTION MANAGEMENT Many collections have begun the databasing pro- cess with excellent intentions: information provi- sion on a case by case basis to other systematists and natural historians who work within a collec- tions environment. Such databases allow the home collection to develop loan-management software that accommodates the idiosyncratic needs of that particular museum. While such database capabili- ties are useful within a small sphere, they do little for outreach beyond the systematics collections community. NOT ONLY FOR THE PURPOSES OF SINGLE STUDIES Many monographers have databased many spec- in files accessible only by outmoded and outdated programs. There also have been databases pro- duced specifically to answer certain ecological questions. But these databases alone do little to contribute to the store of human knowledge if they are left in storage on a floppy disk in a desk drawer. They contain data plus keystrokes plus analysis— the value-added components—that turn raw data into information. Why should this be withheld from others? Why not share it? In so doing, investigators fulfill both a scientific and a societal obligation. NATURAL HISTORY COLLECTIONS DATA INFORM AND ILLUMINATE POLICY DECISIONS Natural history collections have for decades claimed that the data contained in the specimen labels, vouched for by the specimens themselves, are an inestimable information resource for land and resource management, range science, agricul- ture, pharmaceutical chemistry, DNA sequencing for phylogenetic studies, and myriad other fields. This claim is, in fact, true. However, to put those data to use has for the same duration of years been a time-consuming and tedious task of transcription and collation of the data, requiring that a thorough investigator visit one to several collections or obtain the specimens on loan. Those persons who need these data, but who are not inclined to sit for many hours in quiet among the cases, tend to disbelieve the claim of value of specimens because of the te- dium involved in extracting the information. Bad decisions have sometimes been made about the use of biological resources because individuals found it easier to guess or to follow preconceived notions than to obtain real information, even though the needed data were right at hand in the closest nat- ural history collection. Computerized collections catalogs. In the 1970s, when farsighted curators began to apply computer tools to collections management tasks, the argument for doing so was that natural history collections could more easily track loans, care for specimens, and keep catalogs (e.g., Humphrey & Clausen, 1977). he computerization effort was seen as a boon in- ternal to the systematics community. The accessi- bility of the data for ecologists and others was re- garded as a useful byproduct. Collections were to support systematic research first; managing collec- tions more efficiently by computerizing was to be for the benefit of collections personne Value-added data. Formal curatorial catalogs of natural history collections date at least from the establishment of the Ashmolean Museum; most probably they are older than that, stemming from the jumbled days of the “curiosity cabinets.” Orig- inally, catalogs were listings of the kinds of objects 542 Annals of the Missouri Botanical Garden in a collection. With the advent of computerized “catalogs,” the meaning has mutated, become much more expansive, and the catalogs, which we now call databases, infinitely more valuable. There are those who fear that the rapid accessibility of infor- mation in computerized catalogs will render, in the minds of policymakers and funding bodies, the specimens themselves obsolete. In fact, exactly the opposite is true. provides readily accessible information, renders the specimens that voucher it even more valuable. Re- member? In the days before computerization, the specimens and their data were often being ignored altogether. Now, the data cannot be ignored, but likewise they are worthless unless vouchered by well-cared for specimens. People who use data e “evolved catalog,” because it want to know that they are real data; those people are therefore more likely, rather than less, to ap- preciate natural history collections for the inesti- mably valuable resource that they are. COLLECTIONS AS CONSERVATION TOOLS In fact, to humans, knowing something almost always makes that something more valuable. Make it possible to use the knowledge and the valuation skyrockets. Think, for instance, of oil. Petroleum. Nasty black stuff that bubbled out of the earth; once, it was to be avoided! And now look—humans value “black gold” as much as precious metals, are ready to mortgage entire economies, and even go to war for it, because we can use it because we know its properties. I submit that the same could be true of “biodiversity”—humans must come to know and use biodiversity (and it is to be hoped that that use will be sustainable, unlike the use of petroleum). Then, and only then, it will be valued and therefore saved. Natural history collections have an important role to play in this scenario. As the caretakers of the human record of biodiversity and what we know about it in the formal sense, presumably we care about the biota of Earth and want to see biodiver- sity saved. The task of finding ways to get humans to know, and therefore value, and therefore save biodiversity is one of the great challenges to human ingenuity in our time. Natural history collections contain a vital portion of the information on ways of knowing and utilizing biodiversity. We must have the strength of our convictions that our kind of science has a necessary place among those ways of knowing and utilizing, but we must get beyond our hubristic notion that our science, as we have practiced it in the past, is sufficient to supply all of the needed “ways of knowing and utilizing.” We must become entrepreneurial thinkers; we must find ways to add value to our data, and we must make it available and rapidly accessible. By providing high-quality information (value-added data), we will increase the value of our collections, and of biodiversity, in the estimation of our publics. And so, the question becomes: How do we in- crease the value of our data and thereby our col- lections and thereby biodiversity? I believe that finding answers to this question is a role that nat- ural history collections must play now and in the foreseeable future. The data from a single collec- tion, even a large one, cannot by itself answer the big questions. However, those data, interconnected with the data from other collections can together provide a robust answer. We must be continuously conscious of the need to provide outputs of our sci- ence that are useful to society. That, recursively, will make our collections and biodiversity more valuable. This means turning our thinking outward. We must get out from behind our specimen cases—either that, or invite society in, or both. One way or another, we must engage our publics by making it easy for them to know, understand, use, value, and save. I think that this will be easier together than it is for any one institution, or even a group of nine or ten institutions, alone. Natural history collections have in the past five years taken the first, hesitant steps on the road to interconnec- tivity. It is my purpose here to encourage us all to get in shape and begin the marathon. INTERCONNECTIVITY: I am borrowing "interconnectivity" from cyber- space and bringing it into the very human level of relationships among natural history collections per- sonnel because the word “community” has been overworked and used to mean something far less grand than I propose. Natural history collections must together become a nationwide museum: a group of interconnected, interrelated, interactive, interdependent nodes on the Internet. As Benjamin “We must all hang together, or we shall surely all hang separately." We must make ourselves indispensable—together we are, individ- Franklin said, ually we are not. We must view ourselves as work- ing not at Missouri Botanical Garden or the Uni- versity of Kansas or the Smithsonian or Montana State University or the Paleontological Research In- stitute or the Field Museum or the University of Wyoming or the New York Botanical Garden or Louisiana State University, but rather in the great- est collection of all: the collection of all natural history collections. The good part is that to build Volume 83, Number 4 1996 Lane 543 Natural History Collections this grand museum we do not need to change phys- ical location, we merely need to get better at co- operating and appreciating each other's strengths. We do not have to become homogenized; in our diversity is our strength. We can easily grow beyond the “mine is bigger than yours” syndrome, because ours will be, virtually, enormous. BENEFITS FOR NATURAL HISTORY COLLECTIONS In the 1970s, the thought was that computeriza- tion would be a good collections management tool g., Conference of Directors of Systematics Col- lections, 1971; Humphrey & Clausen, 1977), and that analysis of the data by other scientists would be a useful byproduct of the effort. I hope that I have made clear that this notion is actually a better concept if it is turned inside out. We should be computerizing and interconnecting our data for use by other scientists and society. Collections manage- ment is a minor, though useful, byproduct of that effort. The benefits of outward-looking and inter- connected data sets for natural history collections science itself are many: The generation of more robust answers to sys- tematic, phylogenetic, and biogeographical ques- tions, because we wi data sets as well. Shared, combined information is good information. Good information drives good sci- ence; the two are reciprocally illuminative. ) “Gap analysis” of our knowledge—what do we know about which taxa? Are data hiding in some small but precious collection somewhere, under- used and unappreciated and out of context? To- gether with information from other collections, seamlessly combined via the Internet, small collec- tions become immediately more valuable, and large ones can fill in interstices without needing more cases, more compactors, more supplies for the adoption of orphaned collections. The large collec- tions, by sharing data and software with the smaller ones, can actually preserve those collections (and the education that is based on them) in place. More education means more knowledge, more appropri- ate use, more tendency to save. (3) In turn, knowing what we do and do not know will help us make better acquisitions policy deci- ave access to ecological sions, so that systematics professionals can cover new ground rather than unknowingly repeating our (4) Interconnectivity will allow each natural his- tory collection to achieve greater outreach than any one could alone. Some collections have no exhibit space—their collections could be featured at one that does. Others may have magnificent exhibit space but might be missing a component needed for a really spectacular display—and that small collection across the country might have the needed contribution. These are win-win situations. 5) Interconnectivity, because it will allow us to make better decisions, will also reduce our tenden- cy to reinvent the wheel, will reduce competition for scarce resources, and will promote the possi- bility (because we will be providing robust infor- mation that is valuable to society) of finding new resources. Together, we can achieve an efficiency of ae that not even the largest institution can do All of these benefits of interconnectivity together allow us to move forward with the primary task: discovering, classifying, and understanding the world's biota. Again, efficiencies of scale come into play. Dare I dream that we might one day avoid the production of nomenclatural synonyms and hom- onyms, and the waste of precious time that such represent, because we can quickly and efficiently look beyond the physical collection and library at hand to see if what we are about to do has been done before? BENEFITS FOR SOCIETY The benefits to society of interconnected natural history collections lie in the value-added compo- nent. Good information, that is, a more robust data set with value added, drives good policy decisions, allowing for the greatest value to be *mined" from biodiversity i in a sustainable fashion. “Good” here provide complete information about biodiversity anytime in the near future, if we have thirty to a hundred million more species to discover. But, we do know that there is more information in all nat- ural history collections about any one known spe- cies than there is in any one natural history collec- tion about that species poses of Directors of Systematics Collections, 1971). Certainly, the in- formation about a adn species in all natural history collections has synergistic, emergent prop- erties that cannot be imagined based on a single collection. The educational value to society of having all the information from all the natural history collections at the touch of fingertips on a keyboard cannot be estimated. This education, on the Internet, can take many forms: classroom study, individual explora- tion, discussions with curators via CU-See Me or 544 Annals of the Missouri Botanical Garden even more cutting-edge technologies, or informal education opportunities set up and designed to be Web-accessible, to name but a few. Together, we have the wherewithal to put in motion the excellent ideas of individuals who alone may be unable to realize them. I would like to make one final point here that brings our future role full circle to one of the past roles of collections: entertainment. The “cabinets of curiosities” were for the entertainment and personal satisfaction of their owners. People like to be en- tertained. They like to feel that they have owner- ship in something larger than themselves. If they are entertained, they will support and contribute. If there is any doubt of this, just take a look at pro- fessional athletic teams. The Internet offers us the opportunity to make natural history collections en- tertaining. Let us not lose the chance to gain new contributors because we are too focused on formal educational tools, formal scientific endeavor, and the esoterica of which only systematists are capa- ble. An easy leap, once we grow past our tendency to compete among ourselves, is to let our enormous “curiosity cabinet” be their enormous “curiosity cabinet"—and I believe that people will feel that they have ownership in something larger than them- selves, and will contribute to it. Certainly a nation- wide virtual museum will be an entity larger than any of us, but we will all have ownership in it. Go collections! Go connections! NATURAL PARTNERS Biological surveys in general are, of course, the most natural of partners of natural history collec- tions in this enterprise. There are many users of collections data, and these also will be served by natural history collections interconnectivity. But, it is the work of biological surveys, with their con- nections to natural resource users, that are most understandable to our common general public and the decision-making bodies that rely on them. In turn, biological surveys cannot provide good infor- mation to drive good decisions without natural his- tory collections (LaRoe, 1995). Biological surveys and natural history collections are already inextri- cably interconnected. Both sides of the partnership benefit from ready accessibility of information and appreciation, one by the other, of the role that each plays in the societal arena. The National Biological Service of the United States has a slightly different character than the state-level biological surveys about which I have been speaking, in that it sees itself as an organizer of existing information, and a catalyst for needed work throughout the nation. It plans to be a clear- inghouse, a networking node that will put potential collaborators in touch with each other and with needed information. Again, interconnectivity. Nat- ural history collections have everything to lose if we do not engage in the several partnerships that will be thus formed, and everything to gain if we do become a strong, interconnected web that stud- les and conserves an increasingly fragile web of — ^ Literature Cited Alexander, E. P. 1979. Museums in Motion: An Intro- duction to the History and Functions of Museums American Association for State and Local ion Nashville. Allen, W. H. 1993. gos rise of the botanical database. BioScience 43: dr 276 Allmon, W. D. tions. Curator se Anonymous. 1994 E T Agenda 2000: E the AE Systematics Agenda 2000, New York. eon : 1975 The functions of museums in bi- seums J. 74: 159-164. Biological Collections Du: Standards Workshop. 1992. An Information Model for Biological Collections: Report of the Biologic E Collections Data Standards Workshop, August 18-24, 199 Assoc iation of Systematics Col- s. Washing D.C 1990. Forewond. P. iii in E. M. Herholdt (editor), Natural History sae ne Their Management and Value. Transvaal Museum, Pretoria. Cameron, ih 1929, The cuius ut Biological Survey: Its History ies and Organization. Johns Hopkins The vale of natural history collec- T. © Press, New Yor Cher " 1986. rele oe and long-range ecologic rese >p. 29-44 in K. C. Kim & L. Knutson, Foun- pee for a National Biologi cal Survey. Association of Systematics Collections, Kahiin. D.C. Cohn, J. P. 1995. Connecting by computer to collections. BioScience 45: 518-521. Conference of Directors of Systematics Rig ‘tions. The Systematic Biology Collections of the An Essential Resource (A Re ence Foundation). Part 1. Nature, Importance, Condition, and Future. New York _ Botanic al Garden, Bronx R. 1969. The nauena collections as biological standards. Proc. Biol. Soc. Wash. 82: 611—61 Cutright, P. R Lewis and Clark: Pioneanni Nat- uralists. Univ. Illinois Press, Urbana. Dessauer, H. C. & M. S. Hafner. 1984. Collections of Value, Management. Field and Labo- ratory Procedures, and Directory of Existing Collec- NS. e n of Systematics Collections, Lawrence. Duc era . Genoways & C. L. Rose. 1993. bs a Ne ahural deal 'e Collechipas: Chronicle of Our Environmental Heritage. National Institute for the gs. ation of € ior Property, Washington, D.C. n R. 1985. Research: A museum cornerstone. H. 1971. Jnited State es: Frozen Tissues: von (editor), Museum Collections: Their B s aM Future in Biological Mies British Co- lumbia Provincial Museum, Victor Edwards, S. R. € L. D. Grotta (I zditore) 1976. System- Volume 83, Number 4 1996 Lane 545 Natural History Collections atics 75: A Report on a Workshop Assessing the Cur- of Systematics Collections. Association of Systematics: An Impendi risis. Canadian Museum of Nature, Ottawa Grove, seums and media: A status report ERIC Meis et on Media and Technology Series One epartment of Health, Education & Welfare, Washington, Hawks, C. A bt advances in the conservation of natural s scienc cio s. Pp. 53-59 in E. M. Her holdt (editor), Natural History Collections: E Mark agement ransvaal Museum, Pretor Herholdt, Vertebrate collections m ment: A onal perspective with special reference to Southern Africa in E. M. Herholdt (editor), Natural History Collections: Their Management and apo Pretoria . Guidelines for Institutional Poli- cies and aes in Natural History Collections. As- sociation of res Collections, Washington, D. Hooper-Greenhill, E. Museums and the Shaping of Knowledge. a jane & New York. Humphrey, P. S. 1991. The nature of university natural ato Jones : Directions for oc . C. Clausen. 1977. Automated Cataloging for Museum Collections: A Model for Decision and a Guide to Implementation. Association of Systematics Collections, Lawence. Impey, O. & A. MacGregor. 1985. The Origins of Mu- seums: The Cabinet of Curiosities in Sixteenth and Sev- Ke e Century Europe. Clarendon Press, Oxford. Irwin, H. S., Bea ie yanas Pre (Editors). 1973. erica’s Systematics Collections: A National Plan. social of Systematics Collections, -—M ; L. Knutson. 1986. ea basis = a load Biological Survey. Pp. 3-19 i & L. oen dera: Fourtitions for a National Biolog- ical S Association of Systematics Collections, , Vashington, : rm, J. & A. L. Edwards. seum of natural history? Pp. 13-39 in P. S. Cat Jones (editors) Natural History Museums: ibas for ri . Texas Tech Univ. Press, Lubbock. Lamanna, C. 1976. Role of culture collections in the era of molecular biology. /n R. R. Colwell (editor), The Role of Culture Collections in the Era of Molecular Biology. American Society for Microbiology, Washington, D.C. 1991. What is a state mu- Ecosystems. .S. Department of the Interior, "uaa Lemieux, L. 1981. The functions of museums. ‘Gancite (Quart. Canadian Mus. Assoc.) 14: 4-9, Lintz, D. 1991. Are we going in circles? Pp. 239-249 in P. S. Cato & C. Jones (editors), Natural History Muse- ums: Directions for Growth. Texas Tech Univ. Press, ubbock 1994. Museum Governance: Mission, Malaro, M. C. Ethics, Policy. Smithsonian Institution Press, Washing- ton Manning, R. B. Proc. Biol. Soc. Wash. 82: Martin, E. T. 1 dia Thomas o, Scientist. H. Schu- 1969. qe in museum collections. 71-685. man, New Michener, C. D. J.O Corliss, R. S. Cowan, P. H. Rave C. W. Sabro sky, D. F. Scuipes s & G. W. Wharton. 1970. Systematics in oth of Biological Research. National Research Council Division of Biology and Agriculture, 85. Botanical collections in museums. H. Miller (editor), Museum Collections: Their Roles and Future in Biological bis s British Co- lumbia Provincial Museum, Victo Ouellet, H. 1985. Museum colle nda Perspectives. In E. H. Miller (editor, Museum Collections: Their Roles and Future in de vm Research. British Columbia Provincial Museum, Victoria Owen, R. D 1990. Databas se computerization and con- sortium development for vertebrate collections—a col- 5 in E. M. tors), Manual of Natural History i ignc Museums and Galleries Commission, London Porter, C. M. 1991. Natural Lugd in the 20th century: An oxymoron? Pp. 221—237 in P. S. Cato & C. Jones bud Natural History Museu Directions for ech Univ. Press, Lubbock. Radford, A , Y. c. Dickison, A. B. Massey & R. Bell. 1974. CRM Plant Suid. Harper & Row, New ork. Meilen bach, I. L. & E. M. Herholdt. 1990. Curators and the research collections of natural history museums— Their epum to nature conservation in South Af- rica. P 151 in E. M. Herholdt (editor), Natural €: ¡el Their Management and Value Transvaal Museum, Pretoria. Kinar. P. C. 1969. Art and science as influences on the Ep development of natural history collections. . Soc. ag t 82: 561-577. Seld. S yA 1969. e herbarium: Past, present, and future. Proc. Biol. Ss alinean 82: 687-749. 1995. Association of Systematics |o 07] str rategic planning: Diverse institutions, common go Association of Systematics Collections Newsletter 23. 49- Shropshire, C. C. & R. T. Shropshire. 1991. Natural his- tory museums as a function of wildlife management. Pp. 41—47 in P. S. Cato & C. Jones (editors), Natural His tory Museums: Directions for Growth. Texas Tech Uni Press, Lubbock Stansfield, G. 1994. de pad and t f natural history museums. Pp. 1-9 in G. Stansfield, J. Mathias i% G. Reid (editors), Manual of Natural History Cura- orship. Museums and Yochelsan, E. L. 1969. Fossils—The collecting and storing. Proc. Biol. Soc. Washington 82: 5—601. SYSTEMATICS INFORMATION AS A CENTRAL COMPONENT IN THE NATIONAL BIOLOGICAL INFORMATION INFRASTRUCTURE Anne Frondorf" and Gary Waggoner? ABSTRACT The National Biological Service (NBS) is nta the development of the National Biological Information Infrastructure T (NBII), a distributed network of biologic val data a atics information in the N a mation infrastimett 'ture are descri II is dicun by presenting a visio current initiatives of the NBS that direc KA contribute to developing such a systematics infor- ibed. information E e electronically over the Internet. The importan n of a "national systematics information The mission of the National Biological Service (NBS) is to work with other agencies and organi- zations to provide the scientific understanding and technologies needed to manage the Nation's (i.e., the United States’) biological resources. A funda- mental part of this mission is to make data and information on biological resources more accessible for more people to use in making resource man- agement decisions. (For the purposes of this dis- cussion, the term inally collected or developed, while “information” “data” refers to material as orig- refers more broadly to material that has been pro- cessed, or integrated with other material, or dis- played or presented in certain ways. The objective of the NBS and the National Biological Information Infrastructure is to make both data and information more broadly accessible.) A key element of the NBS program is development of a national partner- ship for sharing biological information: the National Biological Information Infrastructure I). NBII concept is a significant component of the rec- ommendations made by the National Research Council in its 1993 report entitled “A Biological Survey for the Nation.’ The goal of the NBII is to establish a distributed "federation" of biological data and information sources, relying on a network of partners and co- operators to make the data they generate and/or maintain available to others throughout this feder- ation, using the Internet. The basic NBII philoso- phy is to encourage and facilitate biological data stewardship. Under this philosophy, the manage- ment of databases remains with those who are re- sponsible for the data, using the "information su- perhighway" to interconnect and disseminate these data to others. In addition to biological data and information, new types of software tools providing new capabil- ities will also be available over the NBII network, to help users work at their own local computer with the actual data from several different remote sources, or to work collaboratively with others at remote sites by using the network (i.e., virtual lab- oratories). The NBII will also point to sources of biological science expertise: people and organiza- tions that users can contact to get assistance in c and understanding biological data. BII is being implemented in three sequen- tial De. The NBII Directory is currently point- ing users to biological databases and information sources, within NBS and from other agencies or organizations. The NBII Clearinghouse is also now in operation, with several sites serving metadata de- scribing and documenting their data holdings and, in some cases, also the data themselves. (The term "metadata" refers to data which serve to describe other data, including such things as how, when, where, and by whom the data were collected; the subject matter of the data; indications of data qual- ity; and information on how to obtain or access the data. In a library, for example, if the books are considered “data,” the information in the card cat- alog would be the “metadata.”) The NBII Distrib- uted System is the ultimate goal of the NBII and is expected to be operational by 1998. In this phase, users will be able to directly access, retrieve, com- ! National Biological Service, U.S. Department of the Interior, 1849 C S S.A. US. Department of the Interior, Tec Ius Transfer Center, P.O. Box 25387, tule S.A. 2 National Biological Service, 810, DFC, Denver Colorado 80225, U.S . NW, Washington D.C. 20240, ANN. MISSOURI Bor. GARD. 83: 546—550. 1996. Volume 83, Number 4 1996 Frondorf & Waggoner National Biological Information Infrastructure bine, and analyze biological data from different sources at different locations (i.e., a virtual national biological database). CURRENT EMPHASES IN NBII DEVELOPMENT NBS is working simultaneously on three fronts in implementing NBII: (1) making the most signif- icant biological data and information products from NBS research and inventory programs electronical- ly accessible; (2) supporting the efforts of many public and private partners to make their signifi- cant biological data and information accessible to others over the NBII; and (3) working cooperatively with other agencies and organizations on the tools and standards needed to provide the infrastructure components of NBII. Examples of significant biological data and in- formation currently available from NBS through NBII include the North American Breeding Bird Survey, data from the nationwide GAP analysis pro- gram, wildlife health bulletins, and data on non- indigenous aquatic species, including the zebra mussel. We continue to work with the NBS science centers, cooperative research units, and other of- fices to identify NBS biological data and informa- tion holdings, and to systematically proceed with cataloging, describing, and electronically serving the most significant of these on NBII. Partner agencies and organizations, including federal and state agencies, private groups, univer- sities, museums, herbaria, and libraries, have by far the bulk of the biological data and information in the country. Thus, we are working with many partners to make the biological data and informa- tion they have more accessible over NBII. Key ar- eas of interest for NBS in developing these part- nerships are with other federal agencies, with state resource agencies, and with the systematics and collections management community. Many of these partners already are making data and information accessible over the Internet, and NBII is linking to these providers, such as the Na- tional Wetlands Inventory database of the U.S. Fish and Wildlife Service, the Global Change Master Di- rectory of data and information, and metadata and data in the National Spatial Data Infrastructure. Other agencies and organizations are interested in working with NBS to help automate, describe, and serve their information, some of which may not cur- rently be available in electronic format. NBS is also working with partners on developing, adapting, and refining the types of software tools, protocols, and standards that are needed to allow users to access, compare, exchange, and use a va- riety of different data and information sets through- out a widely distributed electronic federation of sources. NBS has taken a leadership role in developing a draft metadata content standard for use in docu- menting biological resource data and information in NBII. A metadata content standard specifies what metadata items (fields) to use in describing a data set and how these fields should be formatted and presented. Standardizing the format and presenta- tion of metadata in this way makes it easier for people to compare and contrast among different distributed data sources. The draft standard pro- vides an overall framework for all biological me- tadata, linking with the Federal Geographic Data Committee standard for spatial data, as well as with the U.S. machine readable cataloging (USMARC) standard for library materials/publications. NBS has contracted with the American Institute for Bi- ological Sciences (AIBS), which assembled a dis- tinguished panel of biological scientists that has provided peer review and recommendations on this draft standard. The final AIBS report has been com- pleted, and NBS will be testing and evaluating this standard and will provide it to other partners who might also be willing to test the standard and pro- vide comments for NBS to use in further revision. We are also developing and testing new types of software tools and approaches for serving and ac- cessing biological data and information over the In- ternet. This includes tools for distributed searches, distributed computing, and development of new controlled vocabulary systems to use in searches. IMPORTANCE OF SYSTEMATICS DATA AND INFORMATION IN NB Systematics data and information are of funda- mental importance in building the NBII. System- atics data and information are necessary to support biological inventory, monitoring, and research. Sys- tematics collections data can be applied in long- term biological or ecological research or to analyze long-term environmental trends. Use of standard- ized taxonomic nomenclature supports data sharing and comparison between agencies and organiza- tions. Museum collections also provide essential outreach and education services to raise public awareness and understanding of biodiversity. Systematics data are collected and maintained throughout a widely distributed network of collec- tions, museums, herbaria, botanical gardens, uni- versities, natural history surveys, and other public and private agencies and organizations. In many cases, availability of different systematics data and 548 Annals of the Missouri Botanical Garden information may not be fully known by the people who would be able to use it. Also, much usefu data, such as the information on biological speci- mens housed in museums, are not currently avail- able in an automated format and therefore are not easily accessible. In addition, integration of infor- mation from two or more different sources to help answer questions does not always occur because the data from different sources cannot be easily compared or combined. It is clear that systematics represents a micro- cosm of all of the issues involved in implementing e NBII. Building on the overall NBII effort, our goal is to help develop and support a “national sys- tematics information infrastructure” in which data, information, and tools are electronically accessible in a distributed network for a broad array of differ- ent applications by a variety of users. OBJECTIVES IN DEVELOPING A NATIONAL SYSTEMATICS INFORMATION INFRASTRUCTURE . Support the development and use of data and árida standards, protocols, and tools or tech- nologies to increase the ability to access, under- stand, share, compare, exchange, and use system- atics data and information from many different data and information providers. 2. Support the development of comprehensive surveys and resource directories of the sources, sta- tus, and extent of systematics expertise, data, and information. 3. Develop and implement a program that com- bines small, targeted tactical investments in certain areas where there are opportunities for pilot or demonstration projects, with larger, more strategic investments on major fronts. 4. Actively pursue and develop partnerships in all areas to help meet mutual needs. CURRENT NBII INITIATIVES IN THE SYSTEMATICS INFORMATION ARENA INFRASTRUCTURE DEVELOPMENT NBS has several activities under way to support development of the tools and standards needed to support a nationally (and internationally) distrib- uted systematics information infrastructure. NBS is an active participant (along with several other fed- eral and state agencies, private organizations, and international i ann in the development of the Interagency Taxonomic Information System (ITIS). ITIS is the first comprehensive national da- tabase to provide quality information on the sci- entific nomenclature and taxonomy of all U.S. flora and fauna. ITIS can be accessed on the World Wide Web at http://www. itis.usda.gov/itis. As described above, NBS has developed a pro- posed metadata content standard for use in docu- menting non- E biological data and infor- mation on As this proposed standard is refined, tested, and released over the next year, NBS is working to ensure that it includes appro- priate elements on systematics/taxonomy and that the proposed standard is beta tested on actual se- lected systematics data sets to review its utility and applicability. NBS will also work to disseminate in- formation on applicability and use of a metadata standard to the systematics community. is also working with numerous partners on several fronts to develop a Standardized National Vegetation Classification System. The major effort is in conjunction with two NBS vegetation mapping programs, the National Park Vegetation Mapping program and the National GAP Analysis Program. is working closely with The Nature Conser- vancy and the network of State Heritage Programs in the development of this draft classification. NBS is also actively working with other federal agencies on the Vegetation Subcommittee of the Federal Geographic Data Committee to develop and adopt national standards for vegetation classification and analysis. Development of these standards will sup- port the coordinated production of uniform statistics on and mapping of the nation's vegetation among all the federal agencies and other cooperators. By sharing in a common reference system for how veg- etation cover is classified and described, agencies and organizations that collect and analyze data about vegetation and distribution, trends, etc., can more easily share and exchange their data. And finally, NBS serves on a special panel of the Eco- logical Society of America charged with assisting in the evaluation and development of a standard- ized national vegetation classification system. BS is also working to foster development, en- hancement, and transfer of tools, technologies, and approaches to support the national (and interna- tional) systematics information infrastructure. Key : (1) cost-efficient approaches to large-scale automation of specimen data that are not currently automated; (2) georeferencing of specimen data that are not currently georeferenced; and (3) visualiza- tion and imaging tools to allow for electronic *view- ing" of specimens. One example of the work that NBS is supporting in this area is a project to convert vertebrate dis- tribution records of the original Biological Survey in the Smithsonian’s National Museum of Natural Volume 83, Number 4 1996 Frondorf & Waggoner 5 National Biological Information Infrastructure History to digital format, with digital geographic co- ordinate data from a digital gazetteer that links geo- aphic place names to geographic coordinates. The locality database will be available electronically throug . Next, records from this biological survey locality system will be matched to localities associated with specimen records in the National Museum’s vertebrate specimen databases. This will add geographic coordinates to all matched verte- brate specimen records in the National Museum’s database. Under the NBS State Partnership Program, a competitive program through which NBS supports state-level biodiversity research, inventory, and in- formation transfer activities, NBS is supporting work at the Florida Museum of Natural History to catalog uncataloged ichthyological materials, auto- mate specimen data, and make the ichthyological specimen data accessible electronically over NBII. INFORMATION RESOURCE DIRECTORIES NBS is currently supporting development of four comprehensive information resource directories that are pertinent to the systematics community. NBS and the Association of Systematics Collections are cooperating in the development of two major directories of taxonomic resource information for North America. One is a directory of taxonomic ex- perts and their respective areas of expertise, the other is a directory of research systematics collec- tions, including information on the electronic ac- cessibility of each collection’s specimen data. Both of these directories will be served on NBII and will include direct links to any data or information products that are already available on the Internet. NBS and The Nature Conservancy (TNC) are co- Vr adi n development of a Natural Heritage “node” on the NBII, which will include Je on each of the 50 state natural heritage programs in the United States, plus national-level data summaries and information products from TNC, available on-line. NBS and the Fish and Wildlife Information Exchange are cooperating on development of a directory of state-level biodiver- sity databases, which will also be accessible over Each of these directories will obviously provide an accessible source of information on systematics resources (e.g., systematics expertise, systematics collections, state-level biodiversity databases) for people to use. More importantly, each will also pro- vide a baseline view of the current situation in these areas (i.e., what information resources are available to what extent in what areas) and lead to the identification of additional measures needed to achieve the overall goal of a distributed systematics information infrastructure. For example, they can highlight major taxonomic groups for which signif- icant systematics data are not yet electronically ac- cessible, or new types of tools or protocols needed to support a distributed systematics data network. This information will be used to guide future NBII priorities in the systematics arena. Under a Mem- orandum of Understanding with the National Sci- ence Foundation, NBS hopes to pursue joint efforts to support new projects that focus on these priority areas. DEVELOPING PARTNERSHIPS s noted above, effective partnerships are key to successful development of the NBII and its system- atics component. Partner agencies and organiza- tions maintain most of the significant systematics data; partnerships are needed to jointly develop new software tools and new data and metadata stan- dards; and partnerships are needed to provide the funding resources to support more strategic invest- ments in this area. One recent example of an exciting partnership in this area is a proposed pilot project among NBS, the Smithsonian's National Museum of Natural His- tory, and the Mexican Commission for the Under- standing and Use of Biodiversity (CONABIO). The objectives are to demonstrate and test the feasibil- ity of automating and georeferencing data on se- lected specimens in the Smithsonian collections that were collected in Mexico, as the basis of de- veloping a distributed “international electronic cat- og of Mexican specimens.” This project will be done as a pilot to test the methodology and ap- proach and evaluate costs and benefits to the par- ticipating agencies. If successful and cost-effective, the approach could be used on other portions of Smithsonian natural history collections and/or in other institutions with Mexican specimens. CONCLUSIONS The White House Interagency Subcommittee on Biodiversity and Ecosystem Dynamics has identi- fied systematics as a research priority that is fun- damental to ecosystem management and biodiver- sity conservation. This primary need identified by the Subcommittee requires improvements in the or- ganization of, and access to, standardized system- atics information on nomenclature, classification, and collections. Working with many partners, NBS is striving to make the valuable data and informa- tion embodied in the systematics community, in- Annals of the Missouri Botanical Garden cluding the Nation’s natural history collections, more aie to land use decision makers, re- source and the public in pedir Through championing the Na- tional Biological Information Infrastructure con- cept, NBS is providing leadership to those collect- ing and using biological data and information in the anagers, researchers, students, pursuit of informed decision making about the wise use and conservation of the Nation's biological her- itage. Literature Cited National Research Council. 1993. A Biological Survey for the Nation. National Academy Press, Washington, D.C. A SURVEY OF MICROBIAL DIVERSITY! Diana Lipscomb? ABSTRACT The development of new technologies is increasing our understanding of prokaryotic and eukaryotic microbial di- versity in terms of abundance, diversity, and ecological function. New studies reveal greater diversity than was previously heterogeneous group of diverse taxa and is either paraphyletic or polyphyletic. Alternatively, a multi-kingdom system provides a categories such a: more puce view ie the diversity of life. Concomitant with this is , protozoa, or fungi can no longer the realization that some traditional be considered distinct phylogenetic groups. An overall scheme n clacaifieation xu reflects our growing databases will hardly look like those followed in many text- books, but it will reflect more accurately the relationships of the unicellular microorganisms. Today, most people are aware of the economic and ecological importance of multicellular animals, plants, and fungi. But this appreciation does not usually extend to the importance of biological di- versity of microorganisms. Possibly this is due to the scale at which most microorganisms live. The majority are too small to observe with the naked eye. Many are able to move and constantly change position and so cannot be easily tracked. They have short lives ranging from a few hours to a few days, but they can also form protective cysts and remain in the environment in an inactive state for years. Microorganisms can respond rapidly to changing conditions in their environment, becoming active and growing, or inactive and encysted within min- utes or a few hours. All of this makes them more difficult for scientists to study and for many people to relate to and understand. Nevertheless, aware- ness of microbial diversity is important both be- cause humans depend on microbes for many eco- logical services (production of oxygen, mineral and nutrient cycling, etc.) and because some forms are agents of disease. It is the purpose of this paper to try to convey something about the diversity of the microbial world: what is known and what is uncer- tain, and the prospects and challenges for survey- ing microbial diversity. ScoPE OF MICROBIAL DIVERSITY Microorganisms include both prokaryotes and unicellular eukaryotes (the protists). They are ubiq- uitous and in great abundance in all natural water, sediment, or soil, and are in symbiotic association with multicellular organisms. Information about the size and nature of microbial populations is very in- complete because of difficulties in detecting and extracting them from the environment. Neverthe- less, by piecing together various sources of infor- mation (and taking into account that activity and biomass measurements can be artifacts of the meth- ods used) we are beginning to understand microbial diversity in terms of abundance, diversity, genetic variability, and ecological function. ABUNDANCE Microbial populations may show marked differ- ences over very small distances (McAlice, 1970; Krumbein, 1971; Ashby € Rhodes-Roberts, 1976) or change in intervals as short as a few minutes during environmental flux (Erkenbrecher & Steven- son, 1975). Even taking this into account, in many environments numbers of microbial cells are high. For example, Fenchel (1992) recently estimated that a one-centimeter core of coastal marine sedi- ment contained 4 X 10'” bacterial cells, 10* het- erotrophic flagellates and amoebae, 10* chlorophyll a-containing microorganisms, and 2000 ciliates. DIVERSITY Microbial communities can also be made up of many different species. Many eukaryotic microor- ganisms can be identified by a taxonomic expert using standard microscopical methods (e.g., Fois- ! Support from U.S. National Science Foundation grant (DEB-9305925) is gratefully acknowledged. ? Department of Biological Sciences, George Washington University, Washington D.C. 2005 2, U.S.A. ANN. MISSOURI Bor. GARD. 83: 551-561. 1996. 552 Annals of the Missouri Botanical Garden sner, 1991). Using silver stains and light micros- copy, for example, Finlay et al. (1993) studied the kinds of ciliated protists (phylum Ciliophora) in- habiting the sandy sediment of a Spanish stream. They sampled this stream on just one day in winter (water temperature was 4°C at the warmest time of day). They marked out 1 square meter, and took from it 13 random sediment samples by pushing a 3-cm-diameter tube 3 cm into the sand, capping it, and extracting the sand. From this small amount of sand, they identified 65 species of ciliates belong- ing to 50 genera, from 17 orders. Considering the broad diversity of ciliate habitats available within the area, the importance of physical transport pro- cesses in the river basin, and the fact that many ciliate species have a cosmopolitan distribution, it is probable that the species richness they recorded is representative of the sandy sediment of the river in winter. Analysis of the diversity of prokaryotic microor- ganisms is more difficult because they often lack enough morphological complexity to be accurately identified using microscopic methods. Traditionally, culturing techniques were used to recover prokar- yotes from the environment and to characterize them by their physiological and nutritional require- ments. Because none of these culturing methods permit the isolation and characterization of all pro- karyote species, microbiologists attempted to obtain meaningful diversity data from the use of a wide range of techniques, each of which permits the study of a small fraction of the total marine biomass (Austin, 1988; Button et al., 1993). a small percentage of the prokaryotes that actually exist in a habitat are revealed by these methods (Brock, 1987; Staley & Konopka, 1985 More recently, modern molecular genetic tech- niques have enabled microbiologists to detect more microbial taxa (Stahl et al., 1989; Bull et al., 1992; Embley € Stackebrandt, 1994; Amann et al., 1995). These molecular tools allow microbiologists to extract and analyze nucleic acid sequences (pri- marily rRNA sequences) directly from the environ- ment, thus circumventing the need to culture pro- karyotes before identifying them. The first studies characterizing prokaryotes directly from environ- mental samples used sequences of electrophoreti- cally separated 5S rRNA molecules (Stahl et al., 1984; Lane et al., 1985; Stahl et al.. 1985). These techniques were refined for 16S rRNA (Olsen et al., 1986) and made easier to perform with the advent of polymerase chain reaction (PCR) technology. This modified approach was used by, for example, Giovannoni et al. (1990) to characterize the diver- sity of bacterioplankton in the Sargasso Sea and ¿ven so, only revealed evidence for a previously unrecognized photosynthetic bacterium and a unique heterotro- phic bacterial group. n alternative molecular approach involves clon- ing cDNA transcribed from 16S rRNA using a primer complementary to the universally conserved region of rRNA (Ward et al., 1992; Weller & Ward, 1989). Ward et al. (1990) used this method to study the thermophilic (~50°C) microbial community from Octopus Spring in Yellowstone National Park. Although this community had been previously stud- ied by more traditional methods and was believed to be relatively simple (Ward et al., 1987), the anal- ysis of the cDNA revealed eight distinctive bacte- rial groups that did not match sequences known for cultured species including those previously isolated from hot springs. Since these first studies, such molecular methods have facilitated our exploration of the free-living diversity of marine and estuarine waters (e.g., Schmidt et al., 1991; DeLong, 1992), organisms colonizing solid surfaces (Amann et al., 1992), or- ganisms in water treatment activated sludge (Wag- ner et al., 1993), and organisms in soil (Torsvik et al., 1990; Masters et al., 1991; Stackebrandt et al., 1993). These methods have also been used to de- scribe microorganisms living in association with other organisms, such as the chemoautotrophic bac- teria symbiotic in invertebrates living near hydro- thermal vents (Stahl et al., 1984; Distel et al., the cellulolytic nitrogen-fixing symbiotic bacteria that enable shipworms (wood-boring mol- luscs) to use wood as a principal source of food (Distel et al., 1991), the prokaryotic symbionts of various protists (Embley et al., 1992; Springer et al., 1992; Embley et al., 1993), the bioluminescent bacterial symbionts of fish (Haygood & Distel, 1993), and the cellulolytic symbionts in the rumen of mammalian herbivores (Angert et al., 1993), and to identify both plant and animal pathogens (e.g., elman et al., 1990; Gundersen et al., 1994). Caution must be used in interpreting results be- cause the new molecular methods can be problem- atic. In addition to technical problems, such as probe sensitivity in certain conditions, it can be difficult to retrieve sequences of rare organisms Amann et al., 1995). Furthermore, PCR probing of mixed DNA samples can suggest the presence of — nonexisting organisms by forming chimeric se- quences assembled from sequences of different species (Liesack et al., 1991). Nevertheless, it is obvious that molecular techniques provide extreme- ly useful tools for analysis of microbial diversity (Stahl & Kane, 1992) Volume 83, Number 4 Lipscomb 553 Microbial Diversity ECOLOGICAL FUNCTION Microorganisms fill diverse ecological roles, and practically all key environmental processes are driven by microbial diversity. Microorganisms are essential to biological nutrient cycling, sulfur oxi- dation and reduction, ammonification, methanoge- nesis, and methane oxidation. Symbiosis with mi- croorganisms is a major route by which many multicellular eukaryotes have gained access to complex metabolic activities (Douglas, 1994). These include nitrogen fixation, production of es- sential amino acids and vitamins, cellulose degra- dation, and photosynthesis. Microorganisms are es- sential to all major food webs. A breakdown of the annual carbon fixation and oxygen production by an individual photosynthetic microbe group is not available. It has been estimated that up to as much as 80% of the production of the open ocean is con- tributed by photosynthetic protists (Platt et al., 1983; Takahashi & Beinfang, 1983; Andersen, 1992), and in lakes chrysophytes (protists) and cy- anobacteria contribute up to 40% of the primary productivity (Konopka, 1983). In ecosystems that lack direct photosynthetic in- put, microorganisms are even more important as primary producers. Chemoautotrophic bacteria in environments such as deep sea vents and marine sediments provide energy to support a rich assem- blage of heterotrophic microbes and animals (Jan- nasch & Taylor, 1984; Cavanaugh, 1994). Despite the enormous production of autotrophic microbes, continual grazing by heterotrophs keeps their numbers in check. Some of the biomass pro- duced by photo- and chemo-autotrophic microor- ganisms is consumed directly by animals (e.g., the echinoderm Holothuria atra obtains 10—40% of its carbon from bacteria; Moriarty et al., 1985), but the majority is consumed by heterotrophic microorgan- isms either from pools of dissolved organics, or as decomposers and primary consumers (Pomeroy, 1974; Williams, 1981; Fenchel, 1982, 1986, 1987; Azam et al., 1983; Ducklow, 1983; Hagstrém, 1984). These heterotrophs include prokaryotes, protists, and their protistan predators (Azam et al., 1983, Pomeroy, 1984; Williams, 1984; Sherr & Sherr, 1988). Ciliated protists (phylum Ciliophora) are especially important trophic links in these mi- crobial food webs in that they are major consumers of planktonic bacteria, pico- and nano-planktonic autotrophs, diatoms, dinoflagellates, other ciliates, and heterotrophic flagellates and amoebae, and they are eaten in turn by animals such as zooplank- ton and planktivorous larval fish (reviewed in Pierce & Turner, 1992). SYSTEMATIC DIVERSITY As more is being learned about microbial diver- sity, it is becoming clear that commonly used clas- sification schemes (e.g., Jahn & Jahn, 1949; Whit- taker, 1969, 1977; Margulis & Schwartz, 1988) are oversimplified and simply inadequate for describ- ing the true nature of the diversity, and that mul- tiple kingdoms are needed (Leedale, 1974; Taylor, 1978; Lipscomb, 1985, 1989, 1991; Woese et al., 1990). Certainly the changes and rapid prolifera- tion of new classification schemes of microorgan- isms are confusing to many, but they represent an invigorating period in microbial research from which a new, and hopefully more realistic and sta- ble, system of microbial classification will emerge. PROKARYOTE SYSTEMATIC DIVERSITY In contrast to plants, animals, and even eukary- otic microorganisms, the morphology of prokaryotic microorganisms is, in general, too simple to serve as a basis for classification (Olsen & Woese, 1993; Amann et al., 1995). Classification traditionally de- pended on isolation by culturing followed by char- acterizing organisms according to physiological and biochemical traits. However, much of our under- standing of prokaryote phylogeny is quite recent, due largely to the recent advent of modern molec- ular genetic sequencing methods. Sequences from ribosomal RNA, for example, have revealed two distinct groups of prokaryotes. This distinctiveness caused Woese et al. (1990) to propose a new su- perkingdom category, the domain, and to create three domains: Archaea (for the archebacteria), Bacteria, and Eucarya (for the eukaryotes). Despite criticism of this system (e.g., Cavalier-Smith, 1993), it is in widespread use today. Archaea and Bacteria are both prokaryotic, but the archebacteria have no peptidoglycan in their cell wall, have membrane lipids composed of branched carbon chains attached to glycerol by es- ter linkage, use methione as the start signal for pro- tein synthesis, lack the rRNA loop that binds ri- bosomal protein in the eubacteria, and lack the common arm sequence (guanine-thymine-pseu- douridine-cystine-guanine) of the tRNA found in all eukaryotes and eubacteria. Least squares analysis of ribosomal RNA indi- cates two major groups of Archaea (Embley et al., 1994). One branch, the Crenarchaeota, includes a homogeneous group of sulfur-dependent extreme thermophiles. The other group, the Euryarchaeota, is more heterogeneous and includes methanogens (anaerobes with the ability to get energy by anaer- obically biodegrading substances to methane—an 554 Annals of the Missouri Botanical Garden economically important saga! used world- wide both to reduce waste and generate fuel-grade biogas [Reeve, 1992]), extreme halophiles (Archaea living in high salt environments), and miscella- neous thermophiles (Woese, 1987). This classifi- cation is controversial and has inspired a limited debate over the best way to analyze DNA sequence data, as well as the appropriateness of relying on a single molecular sequence (in this case the ribo- somal RNA) (see, for example, Lake, 1987, 1989; Rivera & Lake, 1992). Hopefully, as more data from different molecules are accumulated and sys- tematists of prokaryotes become more sophisticated about phylogenetic analysis, a better substantiated hypothesis of archebacterial relationships will emerge. The true Bacteria, or Eubacteria, have a pepti- doglycan cell wall, membrane lipids composed of straight carbon chains attached to glycerol by ester linkages, an rRNA loop and the common arm se- quence of the tRNA, and they use formylmethion- ine as a start signal for protein synthesis. Neighbor joining analysis indicates that there are approxi- mately 13 groups of eubacteria (Embley et al., 1994), but because not all h those needing further investigation are temporarily placed in descriptive categories rather than formal taxa: ave been surveyed, TJ . Aquifex-Hydrogenbacter—organisms that oxi- dize H, or reduce sulfur compounds at extreme temperatures (up to 95°C). Some authors have suggested that these may be phylogenetically intermediate between the remaining bacteria and the Archaea. N Thermotogales—two genera from geothermally heated marine sediments . Green nonsulfur or filamentous bacteria —al- though most (e.g., Chloroflexus) are photosyn- thetic, a few of these thermophilic bacteria are nonphototrophic. . Planctomycetales—this group contains bud- ding organisms that are often found attached to surfaces by holdfasts. These organisms have some unusual features for bacteria; at least one species has membrane-bound nucleoid (Fuerst & Webb. 1991), and their rRNA operon is different from other bacteria in that the 16S and 23S genes are separate aem 'k & Stacke- brandt, 1989; Liesack et al., E mE iene dd group includes just two genera: the radiation resistant Deino- w — m coccus and the chemoorganotrophic thermo- phile Thermus, both of which have atypical cell a ~] oo c — © walls containing ornithine in place of diami- nopimelic acid in its peptidoglycan. . Spirochaetes—this group, which was originally recognized on the basis of its unique morphol- ogy and mode of motility, has been confirmed by analysis of rRNA . Chlamydia—this group consists thus far of only one genus, Chlamydia, whose members are all intracellular parasites responsible for a number of sexually transmitted diseases and trachoma (a form of blindness) . Actinomycetales—gram-positive bacteria which may form branching chains that superficially resemble the filamentous bodies of fungi. They are notorious for species that cause tubercu- losis and leprosy, but most are free-living soil microbes of interest to pharmaceutical com- panies for their ability to produce antibiotics that retard the growth of other bacteria. . Proteobacteria—this large, diverse group in- cludes the purple sulfur bacteria and their rel- atives. Many of the organisms in this group are phototrophic and us chlorophyll, which abo. longer wavelengths e the pigment bacterio- than other chlorophylls. It has been hypothe- sized (but not yet rigorously tested using phy- logenetic methods) that the ancestor of this group was photosynthetic and that physiologi- cal diversity characteristic of the group arose as a result of the exchange of photosynthetic capacity for other physiological processes (such as sulfate reduction) as the bacteria ex- panded into new ecological niches. The group thus contains a diversity of bacteria including, in addition to the purple bacteria, sulfate- and sulfur-reducing bacteria, fruiting myxobacteria, enteric bacteria, free-living and symbiotic N,- fixers, and sulfide-oxidizing taxa. . Bacteroides-Flavobacterium—this group con- sists of several genera forming a major clade of gram-negative bacteria. It includes a mix- ture of physiological types including obligate anaerobes (e.g., Bacteroides) and obligate aer- obes (Sporocytophaga) . Gram-positive bacteria with low G+C%—in- cluded here are the endospore formers, lactic acid bacteria, most cocci, the coryniform bac- teria, and mycoplasms. . Cyanobacteria—the blue-green algae, like eu- karyotic plants, use chlorophyll a and two pho- tosystems in tandem to transfer electrons from water to NADP* releasing O, as a waste prod- uct. . Green sulfur bacteria—these bacteria are pho- tosynthetic but use only one photosystem to Volume 83, Number 4 1996 Lipscomb 555 Microbial Diversity transfer electrons from H,S (rather than water) That these groups are natural phylogenetic lin- eages of Archaea and Bacteria still needs to be tested with cladistic analysis. Furthermore, the ability of the rRNA molecule to give significant support for the deep branches of these trees needs to be investigated. EUKARYOTE DIVERSITY Unicellular eukaryotic microorganisms are usu- ally collectively referred to as protists. For many years, protistologists complained that an under- standing of the phylogenetic relationships of the unicellular eukaryotes was hampered by the pau- city of their fossil record and their microscopic size (see Corliss, 1974). These barriers are coming down, resulting in great research activity from which a new picture of protist relationships is emerging. The reasons for recent advances are two- fold and obvious: First, with the widespread use of transmission electron microscopy, confocal micros- copy, and molecular genetic sequencing, micro- scopic size is no longer a hindrance to gathering character data on these organisms. In fact, unicel- lular eukaryotic microorganisms possess a level of cellular and molecular diversity that far exceeds that found in multicellular eukaryotes, and this di- versity provides a wealth of information for recon- structing their phylogeny. Second, systematists have refined their empirical methods—largely thanks to the cladistic revolution—and it is possi- ble to sort through and analyze the new characters in meaningful ways. One of the major changes that has emerged from these studies is the realization that the group is not a cohesive, taxonomic group that stands as an an- cestral hub from which the other eukaryotic king- doms are derived. Instead, the protista is paraphy- letic, not monophyletic, and contains a hetero- geneous mix of organisms at intermediate levels of organization and equivocal boundaries with multi- cellular taxa. A paraphyletic taxon is united by the possession of shared primitive characteristics (sym- plesiomorphies) rather than uniquely derived char- acteristics, and its members thus do not have an- cestors unique to just themselves and lack unique individual histories (Hennig, 1966). When unicellular forms are determined to be an organizational grade along lineages leading to mul- ticellular taxa, the unicellular forms should be in- cluded in the group with their multicellular rela- tives. For example, boundary between the unicellular chlorophytes and there is no kingdom-level the green plants, the choanoflagellates and the sponges, or the chrysophytes and the multicellular brown algae (Lipscomb, 1989). Not all unicellular forms are related to one of the multicellular king- doms. There is evidence that some unicellular taxa (e.g., the kinetoplastids, euglenoids, or ciliates) are independent lineages and not related to the multi- cellular organisms, and these forms should be placed in their own separate taxonomic categories. Thus, it has become clear that commonly used 5-kingdom classification schemes and traditional categories (such as zooflagellate, phytoflagellate, and sarcodine) are oversimplified and simply in- adequate for describing the true nature of diversity, and that multiple kingdoms are needed. Determining what and how many of these cate- gories there might be is not easy (see history of the field in Lipscomb, 1991). Cladistic analyses of the ultrastructural and biochemical features of the cell (Lipscomb, 1985, 1989, 1991, in prep.) confirmed the presence of at least thirteen major groups of protists, all of which, incidentally, had been pro- posed individually by earlier, more traditional sys- tematists (e.g., Smith, 1951; Jeffrey, 1971; Leedale, 1974; Edwards, 1976; Taylor, 1976, 1978). Some of these groups have traditionally been called algae because they are photosynthetic. Algae are not all phylogenetically related. In fact, the oc- currence of at least six distinct lineages of algae provides immediate evidence for their biodiversity (Andersen, 1992): 1. Rhodophyta. The red algae, primarily because they can be multicellular and reach large body sizes, are persistently considered to be plants by many biologists even though their biology is quite different from that of green plants. They are characterized by chloroplasts bound- ed by two membranes containing separate thy- lakoids and lacking an external coat of endo- plasmic reticulum (Dodge, 1973; Pueschel, 1989). They have chlorophyll a , a- and B-car- otene, leutein, and zeaxanthin. Other pigments include the water-soluble allophycocyanin, phycocyanin, and phyrierythrin localized in phycobillosomes found on the thylakoids of the chloroplast. Floridean starch is the major stor- age product and it is found in the cytoplasm rather than the chloroplast. This carbohydrate has a primary chain of a-(1,4)-linked glucans with a 1,6-linked branched chain (Raven et al., 1990). There are no flagella, centrioles, or bas- al bodies (Pueschel, 1989). Without centrioles, it is not surprising to find that mitosis in red algae is different from that in green algae and 556 Annals of the Missouri Botanical Garden N ^ higher plants. The spindle forms on a unique nucleus associated organelle (called the NAO), which is located on each division pole. Fur- thermore, the nuclear membrane does not break down but remains intact except for small openings at the pole for the spindle to pass through (Scott, 1983). Red algae have the old- est fossil record of all the algae, with the ban- giophytes found in 1.25-billion-year-old rocks (Butterfield et al., much of this because the fossil record for most . One cannot make too early eukaryotes is very poor and some other protist group may have predated the red algae but left no record. . Chlorobionts. This clade represents one of the major groups of photosynthetic organisms. It includes the chlorophytes, charophytes, prasi- nophytes, and the multicellular plants. It has also been referred to as the chlorophyte series (Taylor, 1978) or the Viridiplantae (Cavalier- Smith, 1983). All of these taxa have chloro- phylls a and b, and a-(1-4)-linked glucan (am- ylose/amylopectin) as a food storage in their chloroplasts. The chloroplast is bounded by two membranes, as in the rhodophytes, but the thylakoids are in many-layered grana. Motile cells with flagella have a stellate flagellar tran- sition region and a cruciate flagellar root sys- tem. Most unicellular chlorobionts have a cell wall or scales, but not always made of cellu- lose. Cryptomonads) can be char- acterized as having chlorophyll a and c,, phy- cobillins, and the xanthophyll alloxantin in Cryptophytes (— chloroplasts surrounded by four membranes and thylakoids in stacks of two. The inner pair of membranes forms the plastid envelope and the outer pair forms the plastid endoplasmic reticulum. There is an expanded space be- tween the plastid endoplasmic reticulum and the plastid envelope on one face. This expand- ed area contains ribosomes, starch grains, and the nucleomorph (a unique double membrane bound structure containing DNA). The nucleo- morph has been postulated to be a vestigial nucleus belonging to a photosynthetic eukary- otic symbiont (Ludwig & Gibbs, 1987; Douglas et al., 1991). The food storage is the a-(1-4)- linked glucan glycogen, and it is stored in the nucleomorph. The mitochondrial cristae are flattened tubes. The flagella have tubular mas- tigonemes in two rows on one flagellum and one row on the other. They also possess a unique extrusome, called the refractile ejecto- . Chromobionts (= some, and a periplast of proteinaceous plates underlying the cell membrane Stramenopiles) are a grou including both photosynthetic and heterotro- phic forms. These are the heterokont unicel- lular organisms with tubular cristae in their mitochondria, mastigonemes in at least two rows on one of the flagella, and a B(1-3)-linked glucan as a food storage product. Those forms that are photosynthetic have chlorophyll a, cl and c2, thylakoids in stacks of three, and four membranes surrounding the chloroplast with the outermost membrane continuing around the nucleus. The placement of the oomycete and hyphoch- ytrid fungi and the labyrinthulids with the chromobionts is not surprising and was sug- gested by mycologists (e.g., Barr, 1981, 1992). hey share with the Raphidophyta a similar flagellar root structure that consists of a sheet of microtubules extending over the surface of the nucleus. All these taxa share with the ga- mete of foraminiferans the presence of masti- gonemes in two rows on the forward projecting flagellum and a naked trailing flagellum. he heliozoan Actinopodea appear to be de- rived from the chromobionts, specifically from the chrysophyte order Pedinellales. The Helio- zoa were traditionally considered to be sarcod- ines because they have pseudopodia. These pseudopods, usually called axopodia, are long and slender and are made rigid by a core of microtubules. Ultrastructurally identical struc- tures form an anterior ring around the flagella of the Pedinellales. Thus, presence of axopodia is a synapomorphy that unites the Heliozoa with the chrysophytes. This relationship has been reported before by protozoologists (e.g., Patterson, 1989) but often ignored by phycol- ogists. Loosely allied with the chromobionts are the Eustigmatophyta and the bicoecids. The Eus- tigmatophyta are photosynthetic and, as with other chromobionts, they have four membranes surrounding their chloroplast with thylakoids in stacks of three and two rows of mastigone- mes on one flagellum. However, there is only chlorophyll a (no c) and the outermost chloro- plast membrane is not continuous with the nu- clear membrane. Whether the eustigmatophy- tes are primitive chromobionts or derived forms with many secondary losses cannot be deter- mined at this time. The colorless, phagotrophic bicoecids have been considered to be related to the chromobionts by other protistologists be- Volume 83, Number 4 1996 Lipscomb Microbial Diversity v (eN cause they also have two rows of mastigonemes on one flagellum, and their flagellar root struc- ture is similar to that of the gametes of brown algae and the xanthophyte Vaucheria (Moestrup Thomsen, 1976). However, they have evolved many unique features that are perhaps obscuring information that would allow us to place them definitively as the sister taxon to one of these chromophyte groups. . Euglenozoa. The euglenids and the kinetoplas- tid flagellates were first recognized as related taxa by Leedale (1967). The morphological fea- tures these taxa share include linked micro- tubules underlying the cell membrane, and discoidal cristae in the mitochondria. Some members of the Euglenida (or Euglenophyta in the botanical literature) are autotrophic and possess chlorophyll a and b. It has been hy- pothesized that these chloroplasts are the rem- nants of endosymbiotic chlorobionts or chlo- robiont chloroplasts (Lefort-Tran, 1981; What- ley, 1981). Like many chlorophyte chloro- plasts, those of the euglenids lack light-har- vesting carotenoids (Rowan, 1989) and a girdle lamella. In addition, each chloroplast is sur- rounded by an additional single membrane (Dodge, 1973), which is consistent with the idea that they are remnant symbionts. Euglen- ids have a unique storage product, paramylon, which is a B (1,3)-linked glucan chain stored as grains in the cytoplasm. Phagotrophy is common in both photosynthetic and colorless forms, and the microtubules associated with the base of the flagella play a role in feeding (Triemer & Farmer, 1991). . Alveolates. The Dinoflagellata and Ciliophora have been linked as sister taxa on the basis of an alveolar membrane system and the presence of microtubules lining the cytopharynx. Be- cause they also have alveolar membranes, the parasitic Apicomplexa are also included in this lineage. This alveolar membrane system con- sists of a layer of membrane-bound sacs lying just beneath the plasma membrane. In the di- noflagellates, these alveoli contain the theca. Some dinoflagellates are photosynthetic and so are sometimes called algae. These photosyn- thetic forms contain chlorophyll a and c, com- plexed with a unique xanthophyll, peridinin. The Ciliophora make up one of the largest groups of protists and, as has already been dis- cussed, play a major role in microbial food webs as predators and bactivorous organisms. Ciliates generally have rows of cilia with a unique system of two microtubular and one fi- brous roots. They also possess two kinds of nu- clei: a micronucleus that functions in genetic exchange, and a macronucleus that functions in protein synthesis. The Apicomplexa are all symbiotic and in- clude many major disease-causing organisms (e.g., Plasmodium, which causes malaria). The anterior end of the cell contains a unique com- plex of organelles that presumably function in attachment and penetration of the hosts’ cells. The remaining groups consist almost exclusively of heterotrophic organisms. Many of these have been called protozoa, but this shared mode of nu- trition is not sufficient to unite all of these forms into a taxonomic category. 7. eo 9. Parabasalida. These heterotrophic flagellates all possess a distinctive flagellar root structure, which includes a rod of microtubules that ex- tends from the basal bodies around the surface of the nucleus (= axostyle). An elaborate stack of golgi vesicles is often associated with this root between the nucleus and the basal bodies. These protists are almost exclusively symbiotic and are found in many different hosts. Some species, such as Trichomonas, are parasites of humans, but other species, such as Barbula- nympha and Trichonympha, inhabit termites and the woodroach Cryptocercus where they aid in the digestion of cellulose. . Metamonadida. The majority of these flagellat- ed protists are intestinal symbionts (e.g., Giar- dia), but some are free-living. They were once thought to be closely related to the parabasa- lans with whom they share the absence of mi- tochondria and storage glycogen, but they lack the axostyle and golgi apparatus. The meta- monads are characterized by the presence of three bands of microtubules associated with the base of their flagella: a supranuclear band, an infranuclear band, and a band paralleling the recurrent flagellum and oral inpocketing. Animals. The taxa that are included in the an- imal clade are all united by mitosis in which the nuclear membrane breaks down by frag- mentation, septate junctions, choanocytes, col- lagen, spermatozoa, and basal bodies at right angles to each other in flagellated cells. Group- ing with the animal taxa are a protozoan taxon (the choanoflagellates) and the Chytridiomy- cota (chrytid fungi). The chytrids are grouped with the sponges and choanoflagellates be- cause all have microtubules radiating perpen- dicularly to the basal body, which are inter- connected by concentric rings of electron- 558 Annals of the Missouri Botanical Garden po © TJ E — pe — Yo dense material (Lipscomb, 1989, 1991; Barr, 1992). Although additional data is needed to be cer- tain, it is probable that the Myxozoa, a group of protists with multicellular spores and spe- cialized structures called polar capsules, will belong to the animal lineage. Like the true metazoans, they have a separation of somatic and germ cells and cell junctions. The polar capsules share many striking characteristics with the nematocysts of cnidarians 1990), and these similarities appear to be syn- apomorphies that unite the Myxozoa with the om, cnidarian clade. . Amoeboflagellates. The amoeboflagellates and the cellular slime molds form a clade in whic the taxa form limax amoeba that move rela- tively rapidly by the production of eruptive pseudopodia that lack fine extensions. This re- lationship has been suggested by other proto- zoologists (Page, 1976). The amoeboflagellates are also able to transform to a flagellated stage. . Rhizopods. The rhizopods with blunt lobose pseudopodia can be only tentatively grouped together. They have a simple cell structure and lack many of the characteristics used in the cladistic analysis. As the analysis is expanded, we may discover that this group is polyphylet- ic. Opalinids and Proteromonads. These possess similar arrays of subpellicular microtubules underlying the folds in their cell membranes and a unique form of pinocytotic heterotrophy. Mignot (cited in the paper by Brugerolle & Joy- on, 1975) was the first protozoologist to suggest that these forms were related. . Microsporidia. The members of this group have a number of features that are not apparently homologous to features found in other protists. They are small (2-7 um) intracellular parasites that infect many kinds of eukaryotes, including other protists. The spore is the developmental stage with the longest duration, and at the light microscopic level it is the only life cycle stage that can be identified as belonging to a micro- sporidium. Not surprisingly then, most of our morphological information comes from the spore stage, which has a characteristic polar tube and cap used to inject sporoplasm into a host cell where it can grow and divide. Phy- logenetic analysis of the small subunit rRNA often shows the microsporidians diverging first off of the eukaryote tree. This is most conser- vatively interpreted as meaning that the mi- crosporidians are unique eukaryotes well adapted to intracellular parasitism. But some have interpreted this to mean that the micro- sporidians are direct descendants of primitive eukaryotes (e.g., Cavalier-Smith, 1993 In conclusion, a new classification consisting of at least 13 groups of eukaryotes is emerging, but a robust phylogeny of the eukaryotes will not be re- solved until more taxa are examined with both mor- phological and molecular techniques. First, cell bi- ology, particularly electron microscopy, has provided the basic data on which this multi-king- dom system is based, and many of these characters turn out to be convergent, or reversed, or both. For this reason, some branches are defined by very few characters and cladograms have several unresolved branch points where several lineages appear to di- verge simultaneously rather than form a more in- formative branching pattern. Second, analyses do not yet include many parasitic and structurally unique forms. Thus they do not address phyloge- netic placement of all of the taxa. CONCLUSION An appreciation and understanding of the natu- ral world depends on a description of the systematic and ecological diversity of microorganisms as well as the better-known plants, animals, and fungi. There can be little doubt that the biodiversity of microorganisms is still poorly known. Yet as our knowledge has grown it has become increasingly clear that these organisms are an essential and di- verse part of the ecosystems of the wor here is no consensus on the phylogenetic re- lationships of the various unicellular organisms at the phylum and kingdom level and, as long as new taxa and characters are discovered, innovations in classifications of the kingdoms will likely continue. It is obvious that classifications that divide organ- isms into just two, three, four, or five kingdoms are too simplistic, and a multi-kingdom system pro- vides a more realistic view of the diversity of life. Concomitant with this is the realization that some traditional categories such as monera, algae, pro- tozoa, or fungi can no longer be considered distinct phylogenetic groups. An overall scheme of classi- fication that reflects our growing databases has not yet completely emerged. Undoubtedly, it will not resemble those followed in many textbooks today, but it will reflect more accurately the relationships of the unicellular microorganisms. Literature Cited Amann, R. I., J. Stromley, R. Devereux, R. Key & D. A. Stahl. 1992. Molecular and microscopic identification Volume 83, Number 4 1996 Lipscomb 559 Microbial Diversity of sulfate- reducing bacteria in multispecies biofilms. Appl. Environm. Microbiol. 58: 614-623. , W. Ludwig € K.-H. Schleifer. 1995. Phyloge- netic identification and in situ detection of individual microbial cells without cultivation. Microbiol. Rev. 59: 143-169. Andersen, R. A. 1992. Diversity of eukaryotic algae. Bio- diversity Conserv. 1: 267-292, Angert, . D. Clements & N. R. Pace. 1993. The largest acters Nature 362: 239-241. Ashby, R. E. & . Rhodes-Roberts. 1976. The use of IU of variance to examine the variations between amples of marine bacterial populations. J. Appl. Bac- Biol. 41: 439-451. Austin, B. 1988. Marine Microbiology. Cambridge Univ. Press, Cambridge. Azam, F., T. Fenchel, J. G. Field, J. S. Gray, L. A. Meyer- Reil & F. Thingstad. 1983. The ecological role of water column microbes in the sea. Mar. Ecol. Progr. Ser. 10: 257-263. Barr, D. J. S. 1981. The phylogenetic and taxonomic im- plications of flagellar rootlet morpho ied among zoo- sporic fungi. BioSystems 14: 359- 370 . Evolution and the kingdoms of organisms [3 the perspective of a mycologist. Mycologia 84: 1— "n T. D. 1987. The study of microorganisms in situ: Progress and problems. Symp. Soc. Gen. Microbiol. 41: 1-1 Brugerolle, G. & L. Joyon. 1975. Étude y e ul- trastructure des genres Proteromonas et Karotomorpha (Zoom patignptionss 1 Proteromonadida Graba 1952). Pro- I Bull, I as a source of .M. Goodfellow & J. H. Slater. 1992. Biodi- phage in see dla An- noll & K. Swe 1990. A ban 04-107. chut, P. Quang, R. Martin & B. R. . Viability and isolation of marine bac- procedures and initial 6-kingdom je erage and a uni logeny. en W. ; Hammes any pik ak es etr T Intracellu- lar Space as Oligogenetic Ecosystem. Walter de Gruyter, erlin. ] 3. Kingdom Protozoa and its 18 phyla. Mi- crobiol. a 57: 953-994. Cavanaugh, C. M. 1994. Microbial symbiosis— Patterns 1 diversity in the marine environment. Amer. Zool. 34: 79-89. Corliss, J. O. 1974. Time for evolutionary biologists to take more interest in protozoan phylogenetics? Taxon 23: 497-522. DeLong, E. F. 1992. Archeae in coastal marine environ- ments. Proc. Natl. Acad. Sci. USA 89: 5685-5689 Distel, D. L., E. F. DeLong & J. B. Waterbury. 1991 Phylogenetic characterization and in situ localization of bacterial endosymbionts of shipworms (Teredinidae: Bi- valvia) by using 16S rRNA sequence analysis and oli- godeoxynucleotide probe hybridization. Appl. Envi- ronm. Microbiol. 57: 2376-2382. , D. J. or G. J. Olsen, S. J. Giovannoni, B. Pace `N. R. Pace, D. A. Stahl & H. Felbeck. 1988. Sulfur-oxidizing Dod endosymbionts: Analysis of phylogeny and png DR by 16S rRNA sequences. J. Bacteriol. 170: 2506-2510 Dodge, J. D. 1973. The Fine Structure of Algal Cells. Academic Press, London. Douglas, A. 1994. Symbiotic Intereactions. Oxford Univ. d. Douglas, S. E., C. A. Murphy, D. F. Spencer € M. W. Gray. 1991. Vb: ial are evolutionary chi- maeras of two piriosepeticalr distinct unicellular eu- aryotes. Nature 48-1 yrs H. W. 1983. Production and fate of bacteria in the oceans. aie 33: 4 Ediands, P A y dro: ls of plants into higher taxa based on a and biochemical criteria. Tax on 25: 529- Embley, T. M. & E. Stackebrandt. 1994. The use of ri- bosomal RNA sequences in microbial ecology. In R. Pickup, J. Saunders & G. A. Codd (editors), V Molecular Approaches in Environmental Microbiology. Ellis Hor- wood, London. , R. P. Hirt & D. M. Williams. 1994. Biodiversity at the molecular level: The domains, kingdoms and phy- la of life. Philos. Trans., Ser. B 345: 21-33. & R. A. D. Williams. 1993. Re- duced thermophilic bias in the 16S rRNA sequence from Thermus ruber provides further support for a re- lationship between Thermus and Deinococcus. Syst. 2 Microbiol. 16: 25-29. . Finlay, R. H. Thomas & P. L. Dyal. 1992. The u use of rRNA sequences and fluorescent probes to investigate the phylogenetic positions of the anaerobic ciliate Metopus palaeformis and its archebacterial en- dosymbiont. J. Gen. Microbiol. 138: 1479-1487. Erkenbrecher, C. W. . H. Stevenson. 1975. The in- fluence of tidal flux on microbial biomass in salt marsh creeks. Limnol. € Oceanogr. 20: 618-625. Fenchel, T. 1982. Ecology of heterotrophic "— lates: * pq to heterogeneous environments. Mar. e Ecol. Pro The Fo jut of Meterctropie microflagel- lates. Advances Mar. Ecol. 9 — ———. 1987. Ecology of Aa Springer-Verlag, Ber- pre can sai ss learn from microbes: centimetre of sediment surface. Fun Finlay, B. T C. Tellez & G. Esteban. 1993. Diversity of ee og oA in the sandy sediment of a Spanish trea r. J. Gen. Microbiol. 139: 2855-2863. Foleiet W "1091. Basic light and scanning electron mi- iy methods for taxonomic studies of ciliated pro- tozoa. Eur. ^ dirae 27: 313-330. Fuerst, J. A. & R. I. Webb. 1991. Membrane bounded nucleoid in E eubacterium Gemmata obscuriglobus. Sci. U.S.A. 88: 8184-8188. ED; Britschgi, C. L. Moyer & K. G. Field. 1990. Genetic diversity in Sargasso Sea bacter- ioplankton. - 345: 60—63. Gundersen, ., I.-M. Lee, S. Rehner, R. E. Davis & D. T. E 1994, Phylogeny of mycoplas malike or- ganisms (Phytoplasmas): A basis for their classification. J. Bacteriol. 176: 5244-5254. Hagstróm, Á. 1984. Aquatic bacteria: measurements and significance of growth. In M. G. Kluge & C. A. Reddy (editors), Current Perspectives in e ae Ecology. American Society for Microbiology, Washin Haygood, M. G. & D. L. Distel. 1993. Bioluminesobel 560 Annals of the Missouri Botanical Garden symbionts of flashlight fishes and deepsea anglerfishes form unique ra related to the genus Vibrio. Nature 363: D Hennig, W. o Ji i Systematics. Univ. Ili- nois Press, Urba William C. Brown, Dubuque; Iowa. ; . D. Taylor. 1984. Deep-sea micro- biology. Annual E Microbiol. 38: 487—514. Jeffrey, C. 1971. Thallophytes and kingdoms: A critique. Kew Bull. 37: 403-416 Konopka, A. 1983. Eilimnetic and ier primary production in an Indiana r lake. Canad. J. hardwat 98. 1971. TE microbiology and grain- size distuübunon as related to tidal movement during the first mission of the West undersea laboratory ere hg Biol. 10: 101— 112. Lake, J. A A rate independent technique for anal- ysis of s eic d sequences: Evolutionary parsimony. Molec. Biol. Evol. 4 -191. ——.. 1989. Origin of eukaryotic nucleus: wu and eocytes are genotypically un Canad. J. Micr biol. 35 109-118. Lane, D. J., D. A. Stahl, G. J. Olsen, D. J. Heller & N. R. Pace. 1985. Phylogenetic analysis of the genera id deni and Thiomicrospira by 5S rRNA sequenc- s. J. Bacteriol. 163: 75-81. adale, G. F. 1967. The Euglenoid Flagellates. Prentice- Hall, New Jersey. 74. How many are the kingdoms of organ- isms? Taxon 23: 67—89. Lefort-Tran, M. 1981. The triple layered organization of the Euglena eal ary envelope (significance and functions). Ber. Deutsch. Bot. Ges. 94: 463-476. Liesack, W. € E. Sind a 1989. Evidence for un- linked RRN operons in the oe ete Pirellula ma- rina. J. Bacteriol. 171: 5025-503 — Weyland & E. Stac ae risks of gene amplification by PCR 16S rRNA analysis of a mixed culture of strict baso- philic bacteria. Microbial Ecol. 21: 191-198. . Sóller, T. Stewart, H. Haas, S. Giovannoni & E. Stackebrandt. 1992. d influence of trachytelically (rapidly) evolving sequences on the topology of phylo- genetic lenire ud delatum a and the phylo- genetic position of Planctomycetaceae as revealed by comparative analysis of 16S ribosomal RNA sequences. Syst. Appl. iecur 15: 357-362. . The eukaryotic kingdoms. Cladis- 1991. Potential 1989. Relationships among the eukaryotes. /n B. Fernhalm: K. Bremer & H. Jornvall Sue Hier- archy E Ld Exerpta Medica, Amsterdam The kingdoms and the protozoa. /n J. P. Kester nia R. Baker (editors), Parasitic Pn Vol. 1, 2nd Edition. Academic Press, New Yor . Phylum Myxozoa. [n L. Marinis: J. O. J. Chapman (editors), Handbook of Protoctista. Jones - ed oston. e the nucleomorphs Chlorarachnion the vestigal nuclei of eukaryotic ee Ann. New York Acad. Sci. 503: 198- Eu Margulis, L. & K. V. Schwartz. 1988. The Five Kingdoms: An Illustrated Guide to the Phyla of Life on Earth. 2nd Edition. W. H. Freeman, San Francisco. Masters, C. L, R. G. E. Murray, B. E. B. Moseley & K. W. Minton. 1991. DNA polymorphism in new isolates of Deinococcus radiopugnans. J. Gen. Microbiol. 137: 1459-1469 McAlice, B. J. 1970. Observations on the small scale distribution of estuarine phytoplankton. Mar. Biol. 7: 11. l Moestrup, O. & H. Thomsen 1976. Fine structural studies on the ma genus Bicoeca: |. Bicoeca maris with particular emphasis on flagellar apparatus. Proto- plasma 12: 101- E ou D. J. W., P. C. Pollard, W. G. Hunt, C. M. Mor iarty & T. J. Wassenberg. 1985. Produc stivity of acuda and micro-algae and the effect of grazing by holothuri- ans in sediments on a coral reef flat. Mar. Biol. 85: 293— Olsen: G. J. & C. R. Woese. 1993. Ribosomal RNA: A key to phylogeny o 7: 113-123. sane, S. J. Giovannoni, N. R. Pace & D A. "m 1986. Microbial ecology and evolution: A ri- bosomal RNA approach. Annual Rev. Microbiol. 40: 337-365. Page, F. C. 1976. A revised classification of the Gym- Es (Protozoa: Sarcodina). Zool. J. Linn. Soc. 58: 61-77 Patterson, D. J. 1989. Stramenopiles: Chromophytes from . J. T. Turner. 1992. Ecology of planki EA in marine food webs. Rev. Aquatic Sci. 6: 139— 18 Platt, a D. V. Subba Roa & B. Irwin. 1983. Photosyn- thesis of picoplankton in the oligotrophic ocean. Nature 301: 702-704. Pomeroy, L. R. 1974, The ocean's food web, a changing pon BioScience 24: 499-504. 4. Significance of microorganisms in carbon ;. Kluge and energy deine in marine ecosystems. /n M. ( & C. A. Reddy (editors), Current Perspectiv Mi- crobial Ecology. American Society for Microbiology. Washington, D.C. Pueschel, C. M. 1989. An expanded survey of the ultra- bei of red algal pit plugs. J. Phycol. 25: 625—636. n, J. A., A. M. Johnson & J. J. MacFarlane. 1990. ‘Carbon Make In K. M. Cole & R. G. Sneath (ed- itors), Biology of the Red Algae. Cambridge Univ. Press, New York. ie J. N. 1992. d biology of methanogens. ual Rev. Microbiol. 46: 165-191. S iin D. A S. Loutit, D M. Schmidt, S. Falkow & L. S. Tompkins. 1990. Agent of bacillary angiomitosis: An approach to the anlealineation of uncultured patho- 573-1580. 1 1992. Evidence that eu- otes and eocyte prokaryotes are immediate rela- tives. Science 257: 74-76. Rowan, K. S. 1989. Photosynthetic Pigments of Algae. rd e Univ. Press, Cambri Schmidt, T. M., E. F. DeLong & N. R. Pace. 1991. Anal- ysis of a marine picoplankton community by 16S rRNA gene cloning and sequencing. J. Bacteriol. 173: 4371- 4378. Scott, J. 1983. Mitosis in the freshwater red alga Batra- chospermum ectocarpum. Protoplasma 118: 56—70. Sherr, E. B. & B. F. Sherr. 1988. Role of microbes in Volume 83, Number 4 1996 Lipscomb 561 Microbial Diversity pelagic food webs: A revised concept. Limnol. & Oceanogr. en ag eA Smith, G. M. The classification of algae. Jn G. M. um vedi, oul of Phycology. McGraw-Hill, New Yor Springer, N., W. Ludwig, V. Drozanski, R. Amann & K.-H. Scheifer. 1992. The phylogenetic status of Sarcobium cr an obligate intracellular bacterial Ct of mall amoebae. F. E. M. S. Microbiol. Lett. 96: 199— 02 20 1 Stackebrandt, E., W. Liesack & B. M. Goebel. 1993. Bac- terial diversity in a soil sample from a subtropical Aus- tralian environment as determined by 16S rDNA anal- M. D. Kane. 1992. ethods of microbial identification, mae ivl and monitoring of function. Curr. Opin. A 3: 244-252. ne, G. J. Olsen & N. R. Pace. 1984. kas. of eccl vent-associated symbionts by ribosomal sequences. Science 224: 409411. y E . Characteriza- tion of Yellowstoné hot springs microbial community 5S rRNA sequences. Appl. Environm. Microbiol. 49: 1379-1384. . Devereux, R. I. Amann, B. Flesher, C. Lin & J. Staley. 1989. Ribosomal based studies of natural microbial diversity n ecology. /n T. Hattori, Y. Ishida Y. Maruyama, R. Morita & A. Uchida (editors), Recent Sides in | Microbial Ecology. Japan Scientific Press, Tokyo, Ja Staley, J. T. & A Konopka. 1985. Measurement of in situ activities of nonphotosynthetic microorganism aquatic and ja habitats. Annual Rev. Microbiol. 39: 321-346. Takahashi, M. & P. K. Beinfang. 1983. Size structure of phytoplankton biomass and ee uh in subtrop- ical Hawaiian waters. Mar. Biol. 76: 201-211. Taylor, F. J. R. 1976. Flagellate eios A study in conflicts. J. Protozool. 23: 28—40. 1978. Problems in the development of an ex plicit hypothetical phylogeny of the lower pl BioSystems 10: d Torsvik, V., J. Goksoyr & F. L. Dane. 1990. High diver- sity of DNA of soil (E Appl. Environm. Microbiol. 56: 782-787. < P Triemer, R. E. & M. A. Farmer. 1991. An ultrastructural comparison of the mitotic apparatus, feeding a apparatus, flagellar apparatus and Kine Pure in euglenoids and Tee vitis ig sma 164: 91-104. agner, , H. jose & K. H. Schleifer. 1993. Probing cave: sludge with protobacteria-spe- ee’ oligonucleotides: Inadequacy of culture-dependent thods for describing microbial community structure. pay ora Microbiol. 49: 1379-1384. Ward, D. Tayne, K. L. Anderson & M. M. Ba- Tr Community structure and interactions (editors), Ecology of Microbial Communities. Cambridge Univ. Press, Cambridge. eller & M. M. Bateson. 1990. 16S rRNA sequences reveal numerous uncultured organisms in a natural community. Nature 345: 63—65. . Bateson, R. Weller & A. L. Ruff-Roberts. 1992. Ribosomal RNA analysis of microorganisms as they occur in nature. Advances Microbial Ecol. 12: 219-286. Weller, R. & D. M. Ward. 1989. Selective recovery of 16S rRNA sequences from natural microbial commu- nities in the form of cDNA. Appl. Environm. Microbiol. 55: 1818-1822. Whatley, J. M. 1981. Chloroplast evolution-Ancient and modern. Ann. New York Acad. Sci. 361: 1 Whittaker, R. H. 1969. New concepts of kingdoms of organisms. Science 163: 150-1 —— 1977. Broad ciber The kingdoms and the y oed In J. P. Kreier (editor), Parseilic io cademic Press, New York. “ad P. J. 1981. Incorporation of microheterotrophic processes into the classical paradigm of the planktonic food w E Kieler Meersf. 5: 1-28. y LE production in the marine food chai dna eror's new suit of clothes? Jn M. J. R. Fall (editor), eal is jon t mr Materials in Ma- rine Ecosystems. Woese, C. R. 1987. aoan salia Microbiol. Rev. 51: 221-271. , O. Kandler & M. L. Wheelis. 1990. Towards a bdbural system of organisms: Proposal for the domain Archaea, Bacteria, and Eucarya. Proc. Natl. Acad. Sci. U.S.A. 87: 4576-4519. AN INTERNATIONAL VIEW OF NATIONAL BIOLOGICAL SURVEYS! Jorge Soberón?-?, Jorge Llorente? *, and Hesiquio Benítez? * ABSTRACT Museums and herbaria in industrialized countries hold a huge amount of data in the form of labels associated with specimens collected in developing countries. These data represent a significant part of the existing information on biodiversity available for most developing countries. In this paper we first discuss the usefulness of the label information, reign museums, but the international scientific community as well, due to the increased scientific dos that large databases coupled with modern computer technologies provide. Surveying the biological diversity of a nation in- volves the inventorying, cataloging, and mapping of ecosystems, species, populations, and genes. It may also include research on the dynamic aspects of population biology, community and ecosystems ecology, and human uses of the different compo- nents of biodiversity. The task of biological survey- ing is among the oldest that biologists have faced (Badiano, 1552; Hernández, 1790). However, the increasing rate of ecosystem destruction with its as- sociated species extinction and genetic erosion (McNeely et al., 1990; Wilson, 1992; Ehrlich 1995; Whitmore & Sayer, 1992), and the realization that perhaps 90% of the species on the planet re- » main to be named, not to mention mapped or un- derstood (May, 1990; Groombridge, 1992; Stork, 1988), have added urgency to the problem. To ad- dress this urgency, new methods of documenting biodiversity are currently being developed—such as Rapid Assessment Programs (RAPs) (Conserva- tion International, 1991, 1992, 1993a—c), the di- vision of labor and information management meth- ods of Costa Rica's Instituto Nacional de Biodiversidad (INBio) (Janzen et al., 1993), All Taxa Biotic Inventories (Janzen, 1993), new sam- pling methods including the so-called BioRaps (Margules & Austin, 1995), where extrapolations of distributions are obtained prior to the sampling (Margules & Redhead, 1995), and advanced meth- ods using sensors and pattern recognition software. The reasons different societies and cultures may have for cataloging biodiversity run the spectrum from the strictly utilitarian to the philosophical and religious (Norton, 1987; Wilson, 1992). In countries like Mexico, which has a large, diverse rural pop- ulation (Sarukhán et al., are very important. Indeed, in Mexico more than 3000 species are used as medicinal plants (Argueta et al., 1994), many other species have industrial applications (Sarukhán & Dirzo, 1992), such as jo- joba for high-grade oil, and there are hundreds of varieties of corn, beans, chile, and other species that are used for food. Still, non-utilitarian reasons for preserving biological diversity are expressed very forcefully by different segments of the public, such as outcries for protecting the monarch butter- fly [Secretaría de Desarrollo Social (SEDESOL), 1993], waterfowl [Commission for Environmental Cooperation (CEC), 1995], marine mammals (Sali- nas & Ladrón de Guevara, 1993), and other con- spicuous species. A good example in plants is the recently discovered Lacandonia schismatica (Mar- tínez & Ramos, 1989), which despite being appar- ently useless and inconspicuous has been protected by the Chole Indian community of San Javier, Chia- pas, at the cost of not clearing the remaining patch- 1996), utilitarian reasons ! We thank Peter Raven and the bios of the as iii for the invitation to participate. Many ideas presented here have been po ssed extensively with Alejandro Pelae m se involvement we gratefully acknowledge. We also thank Mary Car Navarro and Rafael Caballero for the CIS w and species accumulation curves analyses, an Lupita Bermejo “for helping with the references. Exequiel Eu developed the algorithms for ei analysis of c ia points. We are grateful to Amy S uer McPherson for very det ailed and kind editoria work. misión Nacional Para el Conocimiento y Uso de la Biodiversidad (CONABIO) Fernández Leal 43, Coyoacán, Mis D.F. 04020, Mexico. 3 Centro de Ecología, UNAM. Apdo. Postal 70-275, México 04510, Mexico * Museo de Zoología “Alfonso L. Herrera,” Fac. Ciencias, UNAM. Apdo. Postal 70-399, México D.F. 04510, Mexico. ANN. MISSOURI BOT. GARD. 83: 562-573. 1996. Volume 83, Number 4 1996 Soberón et al. 563 International View XIX and XX Centuries Traditional collecting methods Taxonomy Biogeography New methodologies Specimen collections at Natural History Museums RAP" : Pus Conservation ATBI Data Label Actions 8 i 6 ? Remote sensing 10 specimens of 10 species GIS analyses Presence of species Statistics on hot spots Empirical relationships | - Protected regions Social, economic, demographic, geographic comercia Restoration plannin information | - Planning for surveying and prospecting - Assessment and monitoring Figure 1. Traditional and new methods of collecting provide information about presence of species, mostly as labels on specimens. This information may ultimately have practical applications for conservation, alone or coupled with other types of data. es of rainforest where the first population was dis- covered (Coello et al., 1993) There is no doubt, then, that both in developed and underdeveloped countries different segments of the public, for different reasons, desire to preserve or sustainably use biodiversity. Consequently, gov- ernments have begun to design national programs for surveying ps ee diversity [National Re- search Council (NRC), 1993; Maddox & Gee, 1994; Gámez et al., 1993; Environmental Pares In- formation Network ( N), http .gov. au/general/], and to create idibus]: agencies or agreements that support or encourage such ef- forts (e.g., Global Environment Facility, Convention of Biological Diversity). In this paper we will concentrate on surveying species from the point of view of the Mexican bio- diversity agency, the Comisión Nacional de Biodi- versidad (CONABIO). First, we will present argu- ments for inventorying existing collections and exploring the possible uses of the specimen label data. Second, we will describe our experiences and points of view on the role that foreign institutions play in such efforts. Third and finally, we will il- lustrate these two points using specific examples from CONABIO databases. THE USEFULNESS OF LABEL DATA During the last two centuries major efforts were made to explore, collect, and inventory the natural world. Museums and collections are the deposito- ries of the results, and it is because of the existence of the specimens located in those depositories that organized biological studies are possible (Janzen, 1993; Wheeler, 1995) The museums and herbaria of the world hold in the order of 10* specimens of an order of magnitude of about 10° species (Chalmers, 1996), collected in localities that, to a certain gross spatial scale, cover most of the surface of the planet and an important part of its waters. Therefore, the information these specimens provide should represent the base for national inventories. However, for most of these specimens, we know little more than the date, lo- cation, collector, and name, and much of this in- 564 Annals of the Missouri Botanical Garden USA 214,963 CANADA 847 Figure 2. Graphic representation of the Bird Atlas id Mexico database showing the pr oportions of Mex xic specimens that are housed i in collections in difi formation has some degree of uncertainty, both tax- onomic and geographic. One obvious question is what use such apparently meager data has for con- servation biology. Despite the serious limitations, such data have constituted the basis for much of classical biogeography, and it is the common ground where systematics, biogeography, and ecol- ogy come together to contribute to conservation bi- ology (Fig. 1). From the viewpoint of conservation biology, the basic results that label data from museums produce are lists of taxa and the distribution areas of taxa. Both types of biogeographical information are es- sential to the practical applications of conservation policies. In many countries, choosing areas to be preserved or finding locations of certain important species are the most common “scientific” tasks for conservationists (not to mention the economic, so- cial, and political actions that in the end are the sine qua non of conservation). In countries with high alpha and beta diversities but with a lack of knowledge about that diversity, having reliable lists and ways of extrapolating data rom a collection locality to a larger area allow, among other things, the following: (1) For a given site, determination of the likely presence of endemic, protected, scientifically inter- esting or oe important species (Bojór- quez et al., (2) e paani statistics such as number of species and percentage of endemics. This also allows the location and mapping of hot and cold spots for better ecological planning (Prendergast et al., 1993; International Council for Bird Preserva- tion, 1992; Nelson et al., 1990). (3) Development of empirical relations between Mexican collections 130,348 Other countries 509,519 Figur re 3. Graphic representation of CONABIO’s all- species databases showing the prop portions of Mexican specimens housed in collections in different countries. the above and easily measured climatic, geologic, and biological puis rs to allow extrapolation (Margules & Austin, 1995; Had les & Redhead, 1995; aa; et ane (4) Restoration planning using historical records or empirical rules for hypothetical distribution ar- eas (Allen & Wilson, 1991). (5) Efficient planning of surveying and prospect- ing, both for areas lacking previous fieldwork using “hypothetical species lists” and for species of sci- entific or economic interest (Margules & Redhead, 1995; Sittenfeld & Gámez, 1993). (6) Assessment and monitoring of global climate effects on the areas of species distributions (Chap- man & Busby, 1994). Despite their usefulness, floristic and faunistic studies, and the development of procedures for the production of reliable distributional areas, have been the subject of far less theoretical and meth- odological development than other approaches such as, e.g., the population biological methods that un- derlie Minimum Viable Population Analysis (Soulé, 1987), and which are, at least in many underde- veloped countries, often far removed from the ac- tual practice of conservation. There is no doubt, however, that the proper use of the huge amount of existing label information, in conjunction with mod- ern computer systems, introduces many method- ological and even theoretical issues, some of which we will discuss below, and therefore this should be an area of active researc Of course, any len coming from muse- ums, even assuming it is taxonomically adequate, may require field corroboration and is not a sub- stitute for detailed ecological studies. But label in- formation, because of the sheer abundance of it, its Volume 83, Number 4 1996 Soberón et al. 565 International View REMIB SNIB CONABIO eports External Users: Central Node ment Data base administrator NGO's General Public STEN STEERING COMMITTEE INSTITUTION Curator — Data base custodían Duis beso cibipdisé Data base administrator ; ADVISORY BOARDS as bane Sar INSTITUTION Curator Data base custodian Data base administrator Figure 4. Structure of the Mexican network of biodiversity information (REMIB). The steering committee and advisory boards are academic bod ies in charge of overseeing standards and membership to the network. SNIB is the national system of biodiversity information. The function of REMIB is to provide a mechanism for updating information maintained in the different institutions. historical importance, and its direct relevance to important conservation questions, should have high priority in any national biological survey. SURVEYING THE MUSEUMS OF THE WORLD In this section we will rely heavily on the ex- periences of CONABIO that we believe are most amenable to generalization. Although CONABIO has supported a certain number of field studies, mainly on protected areas, and a pilot parataxono- mist program, modeled on the experience of INBio, has been started, most of the efforts have been di- rected toward computerization of national collec- tions and sharing information with foreign collec- tions. When CONABIO vas created, its main objective was to promote the inventorying of Mexican biodi- versity in computer data banks that could be up- dated regularly. There are several problems in ful- filling this task: (1) Clearly, to inventory all species inhabiting the Mexican territory is a task for gen- erations. Priorities must be set. (2) Most of the specimens that have been collected in Mexico, with e exception of plants, are not held in Mexican collections. Sharing of data is necessary. (3) Gath- ering and updating the data are tasks for profes- sional (or, in some cases, advanced amateur) tax- onomists working in museums and herbaria. Policies and methods for determining the quality of data, acceptance and rejection criteria, accessing and updating the information, ownership, copy- rights, economic compensation, etc., should be agreed upon between taxonomists and their insti- tutions. (4) The size and distribution of the data make its capture in electronic formats expensive and time-consuming, requiring long-term commit- ments. (5) Given the dynamic and complex nature of taxonomic information and the lack of accepted standards for a taxonomic “data model,” the crea- [mad > tion of a suitable computer system was not trivial. (6) In order to be useful to a heterogeneous pub- lic—including government officers, scientists, members of non-governmental organizations (NGOs), and perhaps the general public—statistical, analyti- 566 Annals of the Missouri Botanical Garden Figure 5. degree per side square of of altitude over sea level. Most collecting has been restricted to a limited range of altitude, suggesting that the sampling the Pacific Coast of of ecological conditions is insufficient of Sciences, National University of Mexico cal, and display tools that help to organize the data in user-friendly and interesting ways should be de- vised and implemented. SETTING PRIORITIES Setting taxonomic and geographic priorities for the biological inventories was intensely discussed by CONABIO and its advisors. The final, pragmatic option we chose was to support work on groups in which Mexican experts were willing to work, or where there were ample and easily available data. In a second stage, results of these discussions will be used to pinpoint areas of both taxonomic and geographic ignorance. This many of the to be oriented toward ver- leads grants from tebrates (terrestrial and aquatic), butterflies, some beetle families, and other important invertebrate groups, like certain families of parasitoid wasps and The first results of those projects are now being used to plan field trips to mollusks and annelids. fill in gaps, as will be described below. Distribution of collecting localities of two families of butterflies (Papilionidae and Pieridae) in a one Mexico. The dots represent collection sites, and the shadings categories The database used is the Butterflies Database of the Zoology Museum, School SHARING INFORMATION WITH FOREIGN INSTITUTIONS It is well known that many countries with great biological diversity do not house the best collec- tions of their own flora and fauna. For example, about 90% of the bird specimens collected in Mex- ico are housed in foreign museums, mostly in the United States, Canada, and the United Kingdom (see Fig. 2). This is the case for almost any group considered at higher taxonomic categories, with the exception of vascular plants, for which there are two very large Mexican collections and several me- dium-sized ones. CONABIO is supporting the work of Mexican taxonomists to obtain information from the labels of specimens housed in foreign collec- tions. In fact, by September 1995 almost 80% of CONABIO’s data bank came from specimens housed in foreign collections (Fig. 3). To obtain the information a variety of procedures are used, from requesting computerized catalogs from curators to manually recording the information attached to each specimen. In this way, in the last two years Volume 83, Number 4 Soberón et al. 567 International View (20°, 104°) (20°, 103°) Area= 1 degree square Perimeter = Number of points = 27 | Chi-square = 53.56 D. freedom = 15 P = .000003 e e e e e e pa e j o e e *e $ 5 ee @ e e e o L (19°, 103?) (19°, 103°) Figure 6. Spatial analyses of the data from Figure 5. Spatial distribution was analyzed for randomness of pattern, using standard techniques (see text). The null hypothesis of a random distribution of localities within the square was rejected. information on hundreds of thousands of speci- mens, mostly of vertebrates and butterflies, have been assembled in databases now kept in some Mexican institutions, and eventually most of the in- formation will be available through the Internet, via CONABIO’s home-page. Several interesting points can be learned from CONABIO’s experience. When the work is done by taxonomists acquainted with the proper handling of collections, almost all foreign museums have been open to and helpful with the effort of computerizing labels. Among other things, this sometimes benefits the museums that lack computerized catalogs or the resources to computerize large collections. How- ever, there were a number of concerns related to the idea as a whole. Most concerns of providers are associated with the precise conditions for releasing information to the public domain. In many cases this issue has been resolved by carefully stating the rights of the providers following ideas already used by the Australian Environmental Resources Infor- mation Network (ERIN), or contained in the pub- lished Association for Scientific Collections’ (ASC, 1993) “Guidelines for Institutional Database Poli- cies.” These points can be formalized by signing letters of agreement or Memoranda of Understand- ing between the national biodiversity agency and the foreign museum. In our case, C BIO has already signed a Memorandum of Understanding with the British Museum of Natural History, and is working on other such agreements. POLICIES ABOUT OWNERSHIP AND UPDATING Ownership of information is a subject closely re- lated to the above. Although who is the “owner” of information in a public museum is open to discus- sion, there is no doubt that the museums, the cu- rators of the collections, and the taxonomic experts are the actors directly involved in creating and up- dating the information, and they should be the au- thorities on the use of it. CONABIO is following the experience of ERIN regarding the concept of cus- todianship of the information. This means that we agree that only the providers of the databases and the curators of the collections have the authority to 568 Annals of the Missouri Botanical Garden 98.81 | Asimptote 80 - 9 70 4 . 9 60 * "T d So 40} ` + 30 - 20 - 10 T T T T oe 200 400 600 800 1000 Specimens Figure 7. Species-effort analyses of the data from Figure 6. Using CONABIO’s database, an accumulation of number of new species plotted as a functio curve shows that for the area sample di n of ee effort (number of specimens in fiv sa decreasing rate of new registers, and the asymptote of the curve (about e-year intervals) was done. The 99 species) is near the last value of different species (86) collected in the area. See text for further explanation. modify the databases. It also means that if part of the information is regarded as sensitive (for exam- ple, precise georeferencing of populations of com- mercially valuable endangered species) or not yet ready to use by non-experts (for example, if it is not stable taxonomically), the providers or curators have the authority to mark the fields or records as access-restricted. NABIO, as well as some other national bio- diversity agencies (INBio, ERIN), supports a policy of open access to most of the basic information on biodiversity, that is, the label data. However, this is not universally agreed upon, and therefore before releasing any information to the public domain the formal agreements and permits should state at least what parts of the information are considered open, the disclaimers and pointers in the databases that identify the original sources and dates of the infor- mation, the responsibility (or lack thereof) of the sources, the procedures to report to the sources the amount of use of their information, and any updat- ing mechanisms for the database. The periodic updating of databases containing label data from many tens of thousands of speci- mens located in dozens of institutions in many countries is not a simple task. However, the exis- tence of the Internet makes the task feasible. Pio- neering efforts are already under way by many in- stitutions that are making their catalogs available on-line through the World Wide Web. Among the many entry points, Cornell University Biodiversity and Biological Collections is a useful one. (Its URL is: http://muse.bio.cornell.edu.) Nevertheless, a tru- ly distributed database of museum information is still lacking. A major international effort will be required to reach this goal (Systematics Agenda 2000, 199 In Mexico. CONABIO is coordinating a network of institutions in an effort called Red Mexicana de Información sobre Biodiversidad (REMIB), in which the partner institutions will have access to the information held in each collection using widely available Internet navigators. Although short of full interoperability of the databases, this will allow the Volume 83, Number 4 1996 Soberón et al. 569 International View [a By CAN RD es e 2» sy Figure in southern Mexico. Lines are ro along roads or around biological AM partners to retrieve information on-line and to gain experience with the problems of creating a distrib- uted database (Fig. 4). LONG-TERM COMMITMENTS Any country that attempts to create national in- ventories must be prepared to make a long-term commitment. Not only are lists incomplete even for vertebrates, but the detailed georeferencing re- quired for modern studies (Margules & Redhead, 1995) is available only for certain groups or in a few countries. In the developing world, this will require the creation and strengthening of national capacities in terms of physical and human infra- structure (at the technical, professional, and re- search levels) to do conventional exploring and tax- onomic work, but since most of the already existing data are based in museums in the developed coun- tries, some new commitments are required. As an example, we can take the Birds of Mexico database that was compiled by A. Navarro and H. Benítez of the Museo Alfonso L. Herrera, at the School of Sci- 8. Dots represent pora points in the Bird Atlas of Mexico database, near the Isthmus of Tehuantepec, s. The figure displays the “Collector's Syndrome,” in which dots tend to accumulate ences of the National University of Mexico, and Townsend Peterson, of Kansas University. This da- tabase took more than five years and visits to 40 museums in seven countries to be assembled. The georeferencing of the locations has been going on for three years, and the debugging of the geograph- ical fields alone has involved some 30,000 of the total 280,000 records. Maintaining and updating this database will be a permanent task that may require a budget to allow for trained personnel, traveling, computer resources, etc. Because scien- tific grants are difficult to obtain for these efforts, governments may have to allocate money specifi- cally for the compilation and maintenance of these databases. All the other databases of a national scope that CONABIO is supporting (mammals, fish- es, reptiles and amphibians, butterflies, and scar- abeid beetles) have taken many years to be com- piled and georeferenced and will require con- tinuous efforts by their custodians at universities to maintain, debug, and update. This effort will cost resources and institutional commitments at national and international levels. Annals of the Missouri Botanical Garden N. ^v. Neer 9 om m.n MG igure 9. Collecting localities for Abies religiosa. Data are from herbarium specimens deposited at the Instituto Nacional de Investigaciones Agropecuarias y Forestales (INIFAP). The commitment also means participating in the many international forums that deal with this sub- ject. Since the issue of sharing and updating col- lections information on a global scale has presented a great many questions, it is important that many opinions and points of view be taken into account when agreeing on policies, standards, methods, etc. Currently, there are several more or less interna- tional efforts addressing these problems, groups like the International Organization for Plant Infor- mation (IOPI), the Working Group on Taxonomic Databases (DW), the b nod Union for Bi- ological Sciences! group S. and initia- tives like Systematics Agenda 2000, Species 2000, and the Clearing House of the Biological Diversity Convention. Regular participation from member countries is needed in order to maintain an insti- tutional memory DESIGN OF A DATA MODEL To date, there is no single agreed-upon data model that can be used for an effective exchange of information among data holders around the different sources efficiently and consistently. Several data models are in use in institutions around the world, for example, the databases MUSE (http: //muse.bio.cornell. sao); SMASCH (htt erkelev.edu/ h.html) and \ r TROPICOS (http://www.mobot.org/MOBOT/re- search/datamodel.html) and others accessible through the Biodiversity and Biological Collec- tions Home page (http://muse.bio.cornell.edu/). However, this effort is not complete, and becomes even more complicated when other themes are added to the purely taxonomic information. CON- ABIO has dedicated significant time, money, and human resources to define a data model that can adequately address the diverse needs of govern- ment institutions dealing with the manifold as- pects associated with biological information. many cases incompatible data formats among, museums highlight the additional standardization effort that is required to be able to use information Volume 83, Number 4 1996 Soberón et al. 571 International View Figure 10. Data on altitude, precipitation, and average yearly temperature were obtained for the collecting localities in the previous figure. The geographical information system was then used to plot other localities with similar conditions (see text). from different sources. Taxonomists and scientific societies throughout the world must participate more actively in order to propose and enforce more effective “standards” for taxonomic information, since curatorial information represents the “core” of any biodiversity information system. ANALYTICAL TOOLS Amassing large databases of label information requires the development of many tools for its anal- ysis. This includes, first, display and exploratory tools. Since the information and its interpretation are basically geographical, Geographical Informa- tion Systems (GIS) are an essential tool for dis- playing and organizing the data. However, due to large gaps of knowledge, inferential and predictive tools are also needed. Given the non-randomness of these samples (in space, time, and taxa), and the fact that most label information is just presence data, this presents some interesting statistical prob- ems. As an example of display tools, in Figure 5 we present the distribution of collecting localities of Papilionid and Pierid butterflies (the database comes from 15 museums in Mexico, the United States, Canada, and Europe) in a one degree per side square on the Pacific Coast of Mexico. At a glance one can tell that most of the highlands are not well sampled. An analysis of the randomness of the spatial distribution (Pielou, 1969) confirms that the points are far from being randomly distrib- uted (Fig. 6). In Figure 7 an accumulation curve (Soberón & Llorente, 1993) is displayed. This curve shows that the effort concentrated on the area cov- ered by the points has resulted in a significant drop in the discovery of new species. The two explora- tory tools show that although the lowlands in the square may be adequately sampled, more effort is required in sampling other areas. In another example, in Figure 8 we show the collecting localities for birds in Mexico. The map clearly displays the artifactual pattern created by overcollecting along roads or near biological sta- tions and points out gaps where few or no collecting efforts have been made. This type of information is already being used by CONABIO to evaluate the 572 Annals of the Missouri Botanical Garden information about location of species provided in some environmental impact assessments. On the matter of predictive methods, empirical modeling of the presence data as a function of eas- ily measured, climatic or physiographic variables holds great promise (Butterfield et al., 1994). In igure 9 we display the points where the fir Abies religiosa has been collected, according to the data of the Instituto Nacional de Investigaciones Fores- tales, Agrfcolas y Pecuarias (INIFAP) herbarium in Mexico. In Figure 10 the shaded areas show regions where similar conditions of average temperature, precipitation, and elevation occur. Similar regions were defined using a boxcar method (Carpenter et al., 1993). It is quite possible that the large shaded areas may include non-reported localities for Abies religiosa, and perhaps for other species of Abies as well. Inclusion of the information of other national and foreign herbaria would lead to a much better analysis. Leaving aside the work of Soto and Gó- mez-Pompa (1990), empirical modeling is just be- ginning in Mexico, and although the Australian ex- perience shows that the tool has predictive power, more work is needed in the mountainous regions of Mexico. The possibilities that empirical modeling opens for bioprospecting, planning, and restoration are significant. The National Reforestation Program in Mexico will be using this tool for planning their actions in the immediate future. CONCLUSIONS The use of label information that is currently housed in the many museums around the world will be extremely useful for ecological planning and monitoring. The ability to assemble large label da- tabases, together with the computing tools required to analyze them, is increasing the value of collec- tions, museums, and the work of taxonomists around the world. Besides the practical implica- tions for prospecting for important species, resto- ration and reforestation, ecological planning, locat- ing priority areas, and monitoring and evaluating environmental impact statements, the scientific ap- plications of the databases are quite interesting. Biogeographical and ecogeographical analyses pre- viously done on the scales of thousands of square kilometers may be refined a thousandfold. Empiri- cal modeling should conduce to forming hypotheses and suggesting mechanisms for explaining the cor- relations found. For most countries all the above requires more than just national efforts. The collaboration of mu- seums all over the world, together with scientific societies and many multilateral agencies and pro- grams, will be needed to computerize and distribute the information, agree on policies and standards, develop methods and techniques for analyses, and provide funds. Literature Cited Allen, R. B. & J. B. Wilson. 1991. A method for deter- mining indigenous vegetation from simple environmen- tal factors, and its use for vegetation restoration. Biol. Conservation 56: ipa Argueta, A., o & M. E. Rodarte. 1994, dm de las Plantas de j Medicina Tradicional Mexicana. Three volumes. Instituto Nacional Indigenista, México D.F. Association for Scientific Collections (ASC). 1993. ASC nal d guidelines for institutio atabase policies. ASC Newsletter 21(6): 61— Badiano, J. 1552. Libellus de i di Indorum ec manuscript from rding to Latin introduction by Juan Badiano dies e Instituto Mexicano del Seguro Social, Mexico D. Bojórquez, L., I. Azuara, E. Ezcurra & O. Flores. Identifying conservation priorities in Mexico, through GIS and m Ecol. E UN 5: 215-231. Butterfield, B. R, B. Csuti & J. M 1994. Modeling vertebrate distributions for gap ci Pp. | Ronald Miller (editor), Pine the Diversity of Nature. Chapman & Hall, Lond Carpenter, G., N. Cillison & J. Winter. 1993. DO- MAI flexible modeling procedure for mapping po- ls. Biodiversity tential distributions of plants and animals and Conservation 2: 667-680. 996. Monitoring and sfc biodi- versity: FOSTER data and training. Pp. 171-179 in F. Di Castro & T. Biss es, Biodive ersity, Science and Development: Towards a New ae ip. CAB Inter- national, Wallingford, United Kin Cha apman . & J. R. Busby. io “Linking plant spe- cies un to continental biodiversity inventory, eR modeling and environmental monitoring. Pp. 79-194 in Ronald Miller (editor), Mapping the Diver. of Nature. Chapman & Hall, London. Coello, G., A. Escalante & J. Soberón. 1993. Lack of genetic variation in Lacandonia schismatica (Lacandon- laceae: Teiluridales)3 in its only known locality. Ann. Mis- souri Bot. Gard. 80 1. Commission for Gaui Cooperation (CEC). 1995. on the P6. h of Migratory Birds at 95). Commission for Envi- Pa International (CI). 1991. A Biological As- sessment of the Alto Madidi Region and Adjacent Areas of Northwest Bolivia. Conservation International, Wash- ington, D.C. — ———. 1992. Status of Forest Remnants in the Cordil- of S outhwestern bia Roof Forest Reserve, ee District, Belize. Con- a esla Washing 93b. The Lowland De Forests of Santa Cruz, Bolivia: j Global Conservation Priority. Conservation Volume 83, Number 4 1996 Soberón et al 573 International View International, Fundación Amigos de la Naturaleza, ashi gpi D. A Biological Assessment of the Kanuku Mountain Region of Southwestern Guyana. Conserva- tion International, Washin 995. ngton, D.C. Ehrlich, R. P. 1 The scale ‘of the human enterprise and biodiversity loss. Pp. 214-226 in J. Lawton & M. May (editors), Extinction Rates. Oxford Univ. Pros Oxford. Gámez, R., A. a, A. ahi E. León, J. Jiménez & Piv 1993. Costa Rica's conservation program and the National Biodivesiy Institute (INBio ). Pp. 53- 68 in W. Reid, enfeld, S. Laird, D. poat E: Meyer, M. E 2 [inis diversity Prospecting: Using Genetic Resources tainable Development. World Resources Institute, Washington, D.C. Groombridge, B. 1992. Global ne Status of the rvation Monitor- , F. 1790. Historia Plantarum T Hispaniæ. In: Historia Natural de Nueva España. 1960. Volumes I-III. Facsimile Edition. Universidad Nacional Autón- oma de México. México Iterations Council for Bird iia (ICBP). 1992. g Biodiversity on the Map: Priority Areas for Global Conservation. International Council for Bird Preservation, enc United Kingdom. Janzen, D. H. 1 nomy: Universal sd essential infrastructure for en and management of trop- ical wildland biodiversity. Pp. 100-113 in O. Terje & P. J. Schei (editors), Proceedings of the Norway/UNEP Expert Conference on Biodiversity. Directorate for Na- ture Management, Trondheim, Norway. ——, Y. 2 s R. Gámez, A. Sittenfeld & J. Ji- ménez. 1993. Research management po olicies: Permits for collecting iud research in the tropics. Pp. 131-157 in W. Reid, A. Sittenfeld, S. Laird, D. Janzen, C. Meyer, M. Gollin, R. Gamez & C. Juma (editors), Biodiversity Prospecting: Using Genetic Resou Sustainable Development. World Resources Institute, Washington, C D.C. Maddox, J. & H. Gee. 1994. Mexico's bid to join the World. Nature 368: 789—804. M. P. Austin. 1995. Biological models . M. May (editors), Extinction Rates. Oxford Univ. Press, Or- ford. & T. D. Redhead. 1995. BioRap. Guidelines for Using the BioRap Methodology and Tools. CSIRO, Dickson, Australia. Martínez, E. & C. H. Ramos. 1989. Lacandoniaceae (Triuridales): Una nueva familia de México. Ann. Mis souri Bot. Gard. May R. M. 1990. How many species? Philos. Trans. Roy. Soc. London, B330: 292-304. McNeely, J. A., K. R. re W. V. Reid, R. Mittermeier serving the World’s Biological Diversity. alberi Union for the Conservation of Nature and the World Resources Institute. Gland, Swit- zerland. National Research Council (NRC). 1993. A Biological Survey E the Nation. National Academy Press, Wash- ington Nelson, B. Es C. A. C. Ferreire, M. F. da Silvan & M. L. Kawasaki. 1990. Endemism centres, refugia and bo- tanical collection density in Brazilian Amazonia. Nature 345: 714—716. Norton, B. G. 1987. Why nn B ec Variety? Princeton EE Press, Prince Pielou, E 1969. An ribs: o Manita Ecology. John urs & Sons, New York. Prendergast, J. R., R. M. Quinn, J. L. Lawton, B. C. Ev- ersham & D. W. Gibbons. 1993. Rare ibis the co- incidence of diversity hotspots and conservation strat- E. rón de Guevara. 1993. Riqueza y o mamíferos m marinos. Ciencias, Facul. tad de Ciencias, Univers idad Nacional Autónoma de México. Special Sarukhán, J. & R. Dirz rzo. 1982 México ante los retos de la biodiversidad. Comisión Nacional para el Conoci- miento y Uso de la Biodiversidad, México , J. Soberón & J. Larson. 1996. Biological con- setvatión in a high Beta-diversity country. Pp. 171-179 in F. di Castri & T. Younes, Biodiversity, Science and Development: Towards a New Partnership. CAB Inter- national, Wallingford, United Kingdom. Secretaría de Desarrollo Social (SEDESOL). 1993. Reu- nión sobre la mortandad de la mariposa monarca en México. Avándaro, México. Sittenfeld, A. & R. Gámez. 1993. Biodiversity prospect- in eby INBio. zb 69—67 in W. Reid, A. Sittenfeld, S. rd, D. Janzen, C. Meyer, M. Gollin, R. Gamez & C. jui (editors), Er Prospecting: Using Genetic Resources for Susta cna X © World Re- sources n Washin Soberón, J. & pre 1993. e use of species accumulation Bs for the prediction of species richness, Conservation Biol. 7: 480—488. Soto, M. & A. Gómez-Pompa. 1990. Bioclimatología de la d. x pou No. 1. Instituto de Ecologfa A.C., Xalap Soulé, M. (Editor) xs kisa Populations for Conser- vation. Cambridge Univ. Press, Cambrid Stork, N. E. 1988. Ins da aia Fact, fiction and spec- ulation. Biol. J. Linn. Soc. 35: m Systematics Agenda 2000 (SA2000). 994. Systematics A harting the Hu. Technical Re- au by the Association of Systematics Collections, New Wheeler, Q. D. 1995. Systematics, the scientific basis for inventories of biodiversity. Biodiversity and Conserva- tion 6489. Whitmore, T. C. & J. A. Sayer. 1992. Tropical Defores- tation and Species Extinction. Chapman & Hall, Lon- on. Wilson, E. O. 1992. The Diversity of Life. The Belknap Press of Harvard Univ. Press, Cambridge, Massachu- s. THE LAST SPECIES! Lorin I, Nevling? ABSTRACT The Last Species is an effort to examine the cho cae that confront us today that will determine where this nation will be in the year 3000. This presentation addres Service, its formation a the broader national political and environm nd evolution, and its politic al liabilities The lat i ar A shift i ent dis tress wit ithin the scientific community. Some practical suggestions the o E role for the National Biological ceives special attention because it fits into Ale meaning of critical words and phrases will be are made on how to engage and educate seca officials and the public at large. The sunset is bold, harsh, colored by an acrid, hot atmosphere. The waning sun fails to diminish the heat, absorbed during the long day by a black- ened scorched landscape that returns it to the at- mosphere with equal force as night falls. A pair of tall pink spires droop and slowly, ever so slowly, come to reside, recumbent, on the coarse sand. Al- though in full flower, no seeds will result for there are no pollinators except the hot wind. They are destined to dry, to wither, and to die. hat made the earth unique exists no more. The last species is gone and the earth has become only another of billions of lifeless hulks hurtling through the cosmos. Because no eyes are around to witness the final extinction, no one cares. Is this the earthly scene that we want to mark the millennium begin- ning in the year 3000? I think not. If you agree, then we need to examine the present state of our national affairs and take action to assure a different uture. Many of us are reasonably sure that a meteor collision with the earth caused the last great ex- tinction. For decades, we thought the likely cause of the next major extinction event would be a nu- clear war. Remarkably, that seems far less likely today. We appear to be at the beginning of the sixth great extinction, but this time the cause is of an entirely different nature. The threat is human pop- ulation with its egosystem affecting virtually every ecosystem. If not now, there soon will be too many of us and too many of us doing the wrong things. The number of individuals in a species does not protect that species from extinction; we need only to consider the fate of the passenger pigeon. The United Nations projects that the world population will grow from 5.6 billion to 11 billion by the year 2045. Most of this growth will be in the developing nations. At the end of World War II there were about 55 sovereign nations—today there are 207 (Anonymous, 1995a). Of these, approximately 90 are exploiting their natural resources beyond sus- tainability. Political fragmentation is reflected in numerous areas of armed or political contention throughout the world, often pitting one “culture against another. If we look at the number of dissim- ilar groups that exist, we should anticipate that this fragmentation and accompanying unrest will only increase. The United States makes up approximately 4.5% of the world's population but consumes 25-30% o the resources. Meanwhile, 1.3 billion people live in dire poverty, that is, with less than $1 a day of income. Consider further what a small percentage of its budget this nation spends on research and it becomes clear why the funding of research has be- come increasingly difficult. .S. spends more than villion for advertising, exceeding the gross national product of Australia. The value of our arms exports vastly exceeds that of foreign aid. Common sense dictates that we cannot continue on this path over the long term. It is imperative that we acknowledge that all commerce takes place within the parameters of our natural resources. If we fail to make this acknowledgment, and under- stand and appreciate the limits it places on our actions, we are doomed to self-destruction. It is clear that the world's tropical forests are being irretrievably lost, with 6996 being converted to marginal agriculture, charcoal, local use, or ex- port to developed nations. Much of the tropical for- ' This paper was presented as an after-dinner address at the Garden's 42nd Annual Systematics Symposium. It wa accompa anied by nearly 120 slides that are not reproduced her e. | am especially grateful to Michael Jeffords and Des Crowder of the Illinois Natural History aes for advice ini leiste ral, technical, and electronic support during the proparation and delivery of this presentat 2 Illinois Natural History Survey, 607 a Peabody Drive, Champaign, Illinois 61820, U.S.A. ANN. MISSOURI Bor. GARD. 83: 574-580. 1996. Volume 83, Number 4 1996 Nevling 575 The Last Species est is being devoted to the manufacture of plywood in some developed nations. This plywood, in turn, will be used primarily to make construction forms. Closer to home, the oak logs cut in southern Illinois are used primarily to construct pallets. We are los- ing a vast array of plant and animal species that will remain forever unknown. The worldwide collapse of fisheries is a fact and is due to overexploitation, various pollution factors, and the application of ill-founded hatchery, stock- ing, and aquaculture practices. If you do not be- lieve that there is a collapse, check your local mar- ket prices and species availability. Eighty countries (Anonymous, 1995b) have mod- erate to severe water scarcity, and these countries are home to 40% of the world’s population. The global demand for water is estimated to double ev- ery 21 years. Irrigation consumes 90% of all water used in poor countries with up to half of it being wasted. It is estimated that 10 million deaths an- nually are water related. Drinking-water quality is a problem even in this country with federal stan- dards sometimes being exceeded by nitrates and the herbicide atrazine, as well as several other pes- ticides, in the Midwest (Taylor, 1994). For example, the regulatory level for atrazine is three parts per billion with the highest actual sample level at more than six times the regulatory level. Recent research has shown that DNA modification occurs in hamster tissue at the regulatory level, that is, at three parts per billion (Gunset, 1995). The organism Cryptosporidium is scarcely mon- itored by water finishers even though it is an in- creasing problem. A check of untreated intake wa- ter in 66 water treatment plants in the U.S. identified the presence of Cryptosporidium in 87% of the plants (LeChevillier et al., 1991). The 1993 outbreak in Milwaukee of cryptosporidiosis is es- timated to have caused 400,000 illnesses (Mac Kenzie, 1994) resulting in a loss of $37,000,000 (Smith, 1995) in unearned wages and productivity. Given the large number of activities of the bio- logical community that range over an absolute gal- axy of interests, some centralized coordination may be necessary if we are to make overall progress in protecting the environment. What others are either unable or unwilling to do is an appropriate and logical function of government. In this case, no one has been able to function as an organizer, so it falls to the government to become the organizer and co- ordinator. It might be helpful to briefly review the recent history of a biological survey for the U.S. in order to better appreciate the present political and bio- logical situation. e 1977. Canadian entomologists begin to plan for a Canadian National Biological Survey. Michael Kosztarab, Virginia Polytechnic Institute and tate University, initiates discussions on the need for and utility of a biological survey for the 1982. The Association of Systematics Collections (ASC) endorses the biological survey concept and is soon followed by the endorsement of 28 other scientific societies. @ 1984. A major unscheduled discussion of the bi- ological survey takes place at the ASC annual meeting hosted by the Illinois Natural History urvey. @ 1985. The ASC annual meeting results in the publication Foundations for a National Biologi- cal Survey (Kim & Knutson, 1986). e 1986-1992. A bewildering array of meetings with federal agencies and members of Congress takes place. Legislation was drafted but never advanced. A last-minute attempt in 1992 to form the National Biological Survey (NBS) by Presi- dential Order failed to materialize. @ 1993. Department of the Interior Secretary Bruce Babbitt forms the NBS by Secretarial Order. He also poses a series of questions to the National Research Council (NRC) about the organization and direction the Survey should take. A 19-member committee, under the leadership of Peter Raven, was assembled and issued its re- port A Biological Survey for the Nation in Octo- ber 1993. This report served as the major guide- line for the development of the Survey. One of the stated reasons that Secretary Babbitt formed the Survey was to prevent “environmental train wrecks.” In order to accomplish this he thought it necessary to “develop data that everyone could agree was accurate.” In turn, this would lead to an increase in the quality of natural resource management decisions. In a parallel statement, Illinois Governor Jim Edgar, referring to the three Illinois surveys, stated, “Surveys play a very important role in helping us in government make better decisions because we have the facts from the Surveys.” @ On February 24, 1994, the Association of Sys- tematics Collections and the Department of the Interior entered into a Memorandum of Under- standing. As a result, a Systematics Resources Working Group, consisting of representatives of ASC and NBS, together with representatives from the Environmental Protection Agency, Na- tional Marine Fisheries of the National Oceanic and Atmospheric Administration, the U.S. De- partment of Agriculture and, more recently, the 576 Annals of the Missouri Botanical Garden National Science Foundation, entered into a planning process to assist in the implementation of the MOU. e In June 1994, Ronald Pulliam is named Director of the N Given all this orderly and logical progression of organization and activities, what changed to place us in our current circumstances, that is, NULL to defend the enforcement of environmental law: The scientific community was so buoyed up a what we interpreted as positive trends and devel- opments that we became complacent, even self-sat- isfied, and failed to recognize the controversy that had been developing for many years. The signs were there—some were even discussed at profes- sional meetings—but we did not appreciate the depth of negative feeling toward regulation. Unfor- tunately, Stan Shetler (1995) may have had it right when he said, “I believe, however, that reaction to federal environmental protection and conservation efforts and the funding of science has been setting in for a long time and that the change in American attitudes about regulations and tax support is not a passing fancy.” at happened? In my opinion, the language changed. Language is a living, dynamic thing and words do change. In Through the Looking Glass (Carroll, 1977), “When / use a word,” Humpty Dumpty said, in a rather scornful tone, “it means just what I choose it to mean—neither more nor less.” “The question is,” said Alice, “whether you can make words mean so many different things.” The answer is yes, one can make words mean dif- ferent things at different times. One technique of developing pejorative meaning in words is demon- ization, which is a favored format of religious and political leaders. Even a casual examination of the political debates of our time indicates that religion and politics seem to be merging with increasing ,* frequency and ferocity. Reversing the connotation of a word is accom- plished with relative ease through the use egregious examples based on falsehoods or near-truth. Let me translate a few code-word examples: natural re- source abuse — wise use; that science with which exploiters agree — good science; review by the ex- ploiter community — peer review; obstructionist — environmentalist. "Traditional family values" falls into this category. ittle genealogical re- search by some zealots might convince them that their antecedents were not all that great. An ex- ample of our own making is “theory of evolution.” Theory has vastly different meanings for scientists and the general public. This difference in meaning also opened the door for scientific creationism. Another set of catchwords that remains confusing to the sci- entific community is “ecosystem management." My father was always interested in the local St. Louis issues of bribery and graft. After bribery fell out of fashion, he recognized that it was resurrect- ed, legitimized, and institutionalized and called Po- litical Action Committees or PACs. Aside from the protestations of elected officials, is there any reason to think that contributions do not influence legis- lation? Until this is addressed, that is until PACs are eliminated, legislative power is going to remain in the interest of ia contributors to the detriment of the public at lar As the bill for D: National Biological Survey moved through the House of Representatives in 1993, several real or perceived issues surfaced. The very word Survey in the title conjured up in the minds of some representatives swarms of “Au- dubon Gestapo” sweeping across the nation to find new dickey birds in order to regulate the use of private property. The main issues were access to private property, written permission to access pri- vate property, and the use of volunteers. What I am going to do is briefly outline each problem and re- late the Illinois experience. Access to private property is a complex issue in that it is difficult to keep isolated from the Endan- gered Species Act (ESA). Under the ESA, there has been a number of highly publicized incidents, often untrue, that purport to have “infringed” on individ- ual freedom. The words used in these cases are “unreasonable restraint.” The word of these so-cal- led incidents has spread like ripples on a pond. Mark Twain observed, “A lie is halfway around the world before the truth puts on its shoes.” The fact is that the Fish and Wildlife Service obtained in- junctive relief under the ESA only four times in six years (Beattie, 1995). Governmental regulation is butting heads with a new sense of land ownership rights as perceived and defined by the owners. The perception is that access to private property will lead to new and bur- densome regulation which, in turn, is seen as a form of uncompensated land-taking. These senti- ments are mostly found in rural areas that have not experienced the regulation, that is, zoning, neces- sary to make urban areas operable and livable. Since 1917, the Illinois State Geological Survey, Natural History Survey, and State Water Survey have had authority, in statute (1992 Chapter 96 1/2, Y 7403(b)(1)), to enter all lands of the state: x a To investigate and study the natural resources of State and to prepare printed reports and furnish onan fundamental to the conservation and devel- Volume 83, Number 4 1996 Nevling 577 The Last Species opment of natural resources and for that purpose the officers and employees thereof may, pursuant to rule adopted by the Department, enter and cross all lands in this State, doing no damage to private property.” The wording has changed slightly over the years, but the overall authority remains intact. Staff are expected to contact landowners prior to entering because a plastic identification card provides little protection against the rare event of physical or armed intervention. After hearing the reasons for entering property, the landowner often accompanies the scientist and actually assists with the research. Thus, a landowner is converted into a volunteer and one with an enhanced understanding of the impor- tance of our natural resources, particularly on his or her property. Were there any instances of problems? Of course. I had to deal with one that was precipitated by staffers who failed to clear entry with the landown- er. The landowners were senior citizens who were frightened by the presence of unannounced per- sons. Their concern was exacerbated by recent cat- tle-rustling activities in the area. The written permission to access private property issue is, at least in part, a liability question. Written rmission is one way of being certain of who is on dle land, and liability is considerably less in the case of a trespasser. Obtaining written permission prior to gaining access to private property is pos- sible, but only to a limited extent. Why? More and more rural landowners are proving difficult to iden- tify because of the shift of population toward urban and suburban areas or their relocation to other states or other countries. We live in Piatt County, Illinois, a relatively small county that is largely ag- ricultural in nature. The largest landowner is a for- eign corporation with the land in the hands of a local manager. In central Illinois, on a per-acre ba- sis, close to 75% is operated by someone other than the landowner (Gucker, 1995). Alaska has 71% of its farmland under lease. The lowest percentage of leased farmland is in Maine, New Hampshire, and Massachusetts with 17 to 23%. Many biological events are of a short duration and may be long over before it is possible to locate the landowner, much less receive written permission. The efforts of volunteers are not opposed, per se. The concern is expressed in each of two steps: data gathered by volunteers may be inaccurate because of a lack of expertise; inaccurate data could be used to formulate policy or additional regulation. The ar- gument basically is that volunteers are not as well trained as professional scientists, and even profes- sional scientists have been known to err from time to time. There is also a concern that volunteers may have subtle “white hat” prejudices that would in- fluence the data-gathering process. Early on, the House of Representatives passed an amendment to the NBS bill that explicitly for- bade the use of volunteers. This amendment was not in the tradition or practice of the majority of the adult population of the United States. Although some languages do not even contain a word for vol- unteer, volunteerism is deeply seated in the Amer- ican way of life. It is often said that approximately 80% of American adults serve as volunteers. If vol- unteerism was curbed by legislation, how would our organizations prosper? Consider the impact on churches, scouting, 4-H, hospitals, museums, ser- vice organizations, even political campaigns! To deny the NBS access to volunteers via legislation was both discriminatory and distinctly unAmerican. There are numerous positive examples of data- gathering by volunteers in natural history. NBS ex- amples include the Christmas Bird Count, Breeding Bird Census, Atlas Programs, Project FeederWatch, Hawk Migration Data, Adopt-A-Refuge, Save-Our- Streams, Natural Heritage Data, and the Fourth of July Butterfly Count. It is imperative for us to un- derstand the concerns of some, and special mea- sures must be taken. Data gathered by volunteers must be reviewed by professional scientists, anom- alies investigated, and occasionally data sets may have to be discarded. All interpretation of data must be made by professionals. In November 1994, along came the Contract with America (CONWAM) and, suddenly, we were con- fronting a horse of another color. Environmental leadership from the White House collapsed, the en- vironmentally sensitive Vice President disappeared in a flash of white light, and new leadership emerged in the Senate and the House. As we look forward to the next presidential election, leading presidential wannabes and a third-party hopeful do not hold promise of anything better. It seems as though a fifth Horseman of the Apocalypse should be added—elected officials. The NBS problem areas became increasingly dif- ficult and on January 5, 1995, in an attempt to dampen overheated sensibilities, Secretary Babbitt changed the name to the National Biological Ser- vice. Subsequently, both houses of the Congress terminated the NBS, but the Senate’s version of H.B. 1917 transferred NBS programs to the Natural Resources Science Agency. Unfortunately, the Sen- ate version provides that no funds may be used for surveys, including aerial ones, on private property unless with the explicit written permission of the owner. What must we do to restore the worth of our sci- 578 Annals of the Missouri Botanical Garden ence in the eyes of the public and legislators? I have a few suggestions to offer for your considera- tion. We must capture the data held in our collections in such a way as to make it usable and useful to a broad spectrum of users. The National Biological Information Infrastructure (NBII) was conceived precisely to make this possible through intercon- nectiveness. For nearly a quarter of a century too many in the profession have been dragging their feet and using an inability to devise a commonly agreed-upon data format to justify inaction. Full computerization of collections data cannot wait un- til my grandson grows up, even though at ten months he has an early start. There has been a continuing debate about the best hardware and soft- ware. As the technology becomes increasingly transparent, this is no longer a matter for valid dis- cussion. Concerns do remain—proper specimen identification, for example. I am torn on this point: one side says make the data available, even if it is not robust, because it will get corrected; the other side says get it right. In 1972, I suggested that data entry could be made at the time of loan return, thus assuring a higher than normal level of reliability. Had that been done consistently by all, the ento- mologists, who are facing a monstrous task, would e further along. A second problem area, no pun intended, is that of geographic locality. Older collections, where the locality data is obscure or too broadly drawn, may have to be put aside even though they are of im- mense historical importance. For current collec- tions I see no reason why global positioning systems should not be employed to provide the level of ac- curacy necessary for GIS applications. Some members of the Congress would prefer to eliminate not only aerial surveys but also electronic data-gathering. At first, this seems curious, but even with limited reflection it is clear that what they wish to eliminate is incontrovertible evidence. Just as I once said that GIS would blow your socks off, the same is now true of the blending of satellite imagery and landcover analysis. Before the end of the year we expect to have complete coverage for Illinois. We completed a 10-year change compari- son for one county and found that the changes were not what everyone imagined. A loss of farmland was projected, but proved untrue; the actual loss was in forested lands. Lack of electronic, that is, Internet, accessibility rovides an additional and major problem for some organizations. Those not on the Internet with a me page are excluded from a major information interchange. If you do have a home page, keep track of where the queries are originating and the kinds of organizations making the inquiries. For those of us in the governmental realm, it is clear that we must make our data available. Those among you from private institutions may experience a problem at several levels. The curator may wish to make collections data freely available, but the di- rector does not. The director's position is that the institution has an obligation to try to recoup part of its investment in developing the database. This po- sition honors fiduciary responsibility. The reverse situation may occur if the curator refuses to provide collections data. Wayne King refers to this as “cu- ratorial barony.” For those among you who believe that your data can be a major source of institutional revenue and that it can support computer opera- tions, you are advised to reconsider for you will experience disappointment. Computer operations are better regarded as part of the infrastructure support necessary to do science rather than as a revenue stream The majority of you are systematists and your institution is, I hope, a member of the Association of Systematics Collections. I mentioned the System- atics Resources Working Group (Working Group) previously, and I would like to share some aspects of our first interim report with you. That report will be issued shortly with six recommendations. The Working Group recognizes the necessity to focus on products rather than process in order to demonstrate the relevance of systematics collec- tions, systematic biology, biodiversity, and biotic in- ventories. The first recommendation, to establish a home page for ASC, has been accomplished. Two additional recommendations have been implement- ed by ASC and NBS, the development of a Taxo- nomic Resource Expertise Directory and a Di ory of Research Systematics Collections. directories will be available on the Internet through — The fourth recommendation is to automate es- sential collections data. The given is that there are insufficient funds to automate all collections data over the short term. The Working Group identified high-priority groups that could serve as demonstra- tion projects for data automation. The high-priority groups are amphibians and reptiles; conifers; fresh- water crustaceans; freshwater fishes; mycorrhizal fungi; lepidoptera; lichens; unionid, hydrobiid, and pleurocerid mollusks; nematodes; and orchids. The Working Group fully realizes that this list will not enjoy universal appeal and, therefore, suggests that the list be subject to regular review and revision. It is of interest that the NSF has awarded more than Volume 83, Number 4 1996 Nevlin 579 g The Last Species sixty grants in FY 93 and 94 in these priority ups. The fifth recommendation is to develop a public document detailing practical applications of sys- tematics as a supplement to Systematics Agenda 2000. Eventually, examples of applications of sys- tematics should be made available on the ASC Home Page. If you would like an outstanding ex- ample of efforts along these lines, see Reinventing Systematics (Becker, 1 The final recommendation is the development of a list of priority needs for training and professional positions. This list will be developed from identified gaps in scientific expertise in critical biological groups with input from the NSF Partnerships for Enhancing Expertise in Taxonomy (PEET). Look for opportunities to provide baseline infor- mation to both the scientific and lay communities. In Illinois we did this through Our Living Heritage; The Biological Resources of Illinois (Page & Jef- fords, 1991) and The Changing lllinois Environ- ment: Critical Trends (Illinois Department of Energy and Natural Resources, 1994). NBS recently did the same in Our Living Resources (LaRoe et al., 1995). Once this information is in the public do- main it can be used for demonstrating change over me. The Office of the President appears to have little power beyond that of persuasion and veto. There- fore, it may be foolish to concentrate exclusively on the presidency with regard to elections. It may be far more important to devote energies to congres- sional elections. Forget party labels and support those that support environmental issues because the current environmental battlefield is the appro- priations committees. Be very careful on this point because some may be prohibited from taking ad- vocacy positions, and others may be with institu- tions that find it unacceptable. Proposed new leg- islation linking federal grants and advocacy are an added institutional threat (Anonymous, 1995c). You are required to learn at least the rudiments of politics and the Byzantine pathway that leads to legislation. It is said that lawmaking is similar to sausage making—you should never see it done. It is important to realize that elected officials are in- terested in science only as a tool to help formulate policy and legislation. You must not ignore the state-level lawmaking because most land-use law is drawn at the state or county level. Lawmaking is complex and politics, or the lack of knowledge, is apparent at virtually every step, and "it isn't over until it's over." Write, fax, call, or e-mail your elected officials on issues that are important to you. Be persistent. Check to see if your state has a delegation office in Washington. If it does, find out how to make contact and get to know the staffers. They have systems to contact the state delegation in short order. You may be able to convince them to transmit your message to the entire delegation. I would like to pause to personally thank the Trustees of the Garden for supporting Peter Raven's effective advocacy on behalf of a host of biological matters over the years. Their support did not hap- pen by accident, but through Peter's continuing ef- forts to educate the Trustees. Peter, for you I offer my first modern-day environmental holy card. Most of you believe you are educators—educate! If you do not know where to begin, start with 5th and 6th graders. It is imperative to capture their attention before the glands begin to work and they focus on one another. There are a multitude of ways to approach children—exhibits, theater, coloring posters, curriculum development, field trips, and so on. The second priority should be elected officials. Lure them to your institutions and tour the collec- tions—if you cannot attract the legislators, settle for legislative aides. Show them what you have ac- complished and share with them what remains to be done. Stress the importance of their ongoing support. Follow up because you need to let them know when something good happens and not just when you desperately need their assistance. Try the same approach on that reluctant dean or vice chan- cellor for research. Once in a while you may wish to try an “exper- imental" field trip. We tried one that examined wet- land delineation techniques in a number of differ- ent environmental settings. Invitees were a local state legislator, representatives from the Farm Bu- reau, U.S. Fish and Wildlife Service, the Soil and Water Conservation Districts, U.S. Army Corps of Engineers, and the state Department of Conserva- tion. The learning and understanding process was generated by the field trip participants and not en- tirely by the trip leaders. In the end, you must educate everyone. Do not neglect your trustees or board members. Do you need ideas? Walk around this garden because it absolutely abounds in teaching and learning op- portunities. You are familiar with all the reasons for preserv- ing biodiversity, ecosystems, and the environment. They have almost become a mantra and have put some of us in a hypnotic state, even depression, rendering us incapable of action. Our mantra has not worked, not because our reasoning is not logical but precisely because it is. Stephen Meyer (1995) has noted, “The practice of what most academics 580 Annals of the Missouri Botanical Garden consider to be good science is largely antithetical to the practice of politics.” It is a lesson that I, and I suspect far too many scientists, forget with aston- ishing regularity. We, you and I, are privileged to be participants in the ultimate detective story—understanding the course of evolution of living things. Could there be anything more challenging or more rewarding? I think not. We are also fortunate in that we often have the opportunity to enjoy and appreciate the sheer beauty of the living world. Because we are so privileged and have some higher level of under- standing of where the world is headed, we, you and I, have a responsibility level far above that of the average citizen. Do you want the story to end in desolation, or do you want the next millennium to herald a new sun- rise? Tonight, now, I place the responsibility for the shape of the next millennium into your hands. Literature Cited 1995a. The Statesman’s Year-Book. 1994— acMillan Press, ndon oo l . Washington Environmental Protec- n Report. Water rae Looms, World Bank Says. Re- e No. 698, Aug. 15. p.1, 2. ipei 19950. TRA Adopts Limitations on Ad- ocacy. Aviso. American Association of Museums, Sep., Anonymous. 99 Beattie, M. 1995. Preserving po Myths and Reali- 6( 995. dui. Posten Agricultural Research epu United States Department of Agri- culture 43(5): Carroll, L. 1977. d the Looking-Glass. St. Martin's Press, New York. Gucker, D. 1995. Survey reveals Alaska ahead of Illinois in leased farmland. Piatt County Journal Republican. July 26, p. 1 Gunset, G. 1 Health risks may out top weedkiller. Chicago Tribune. Sep. 10. [News account of unpub- lished research of A. L. Rayburn Illinois Department of Ener, and Natural Resources. The werd Illinois Environment: Critical Trends Summary Report and Volumes 1-7 Technical Rep Kim, i i & L. Knutson (Editors). 1986. Foundations for a National Biological Survey. Association of System- atics Collections. Lawrence, Kansas. LaRoe, E. T., G. S. Farris & C. E. Puckett. 1995. Our Living Resources. A Report to the Nation on the Dis- tribution, Abundance, and Health of U.S. Plants, Ani- mals, and Ecosystems. U.S. Department of the Interior, National Biological Service, Washington, D oe M. W., W. D. Norton, R. G. Lee & J. B. Rose. Giardia and Semone an Pp. xvi-xvii in Wa- e dd da Executive Summary. esearch and American Water Works Association. . 1994. A Massive Outbreak in Milwaukee. Cryptosporidium Infection Transmitted Through the Public Water m The New England Journal of Medicine 331: 161-167 Meyer, S. M. 1995. The Role of Scientists d in the *New Politics.” The Chronicle of Higher Education. May 26, p. Bl, 2. National Research Council. 1993. A Biological Survey r the Nation. National Academy Press, Washington, Page, L. M. & M. R. pus (Editors). 1991. Our Living Heritage: The Biol i ederal Government Museums: ASC Smith, V. 1995. Complacency caused Milwaukee’s crypto outbreak. Viewpoint. Journal American Water Works Association. 87: 8, 10. poc - G. 1994. Josie in rial Public Water : A Year of ce Monitoring. Presenta- n $t “he 1994 Illinois all Pesce Confer- nce, University of Illinois, Urbana-Champaign. ANNALS OF THE MISSOURI BOTANICAL GARDEN: CHECKLIST FOR AUTHORS 1. General Instructions O Text is in English or Spanish on numbered pages. O Manuscript is typed on one side of nonglossy 8% X 11 in. paper. O At least 1 in. is left as margin all around, except on the first page, which has 3 in. left blank at the top. O Three copies of double- or triple-spaced printed manu- script, including abstract, legends, tables, specimen lists, Literature Cited, and footnotes, are enclosed. O Manuscript is also submitted on MS-DOS/WINDOWS 3% in. diskette, as WordPerfect® [preferable] file. O Right-hand margin is not justified, and words are not broken there. . O Special typefaces are not used. Common Latin words or phrases are not italicized (e.g., et al., i.e., sensu, etc.). O Only names at the rank of genus and below are itali- cized. O Correct accents, umlauts, and other diacritical marks have been include O All figures and tables are cited in the O Photocopies of the figures are included han each copy of the manuscript. 2. Style L] Recent issue of the Annals is used as O Chicago Manual of Style, latest editign. i is ad as a reference. 3. First Page [] Footnotes are typed as double-spaced paragraphs on n on granting agencies, herbaria that loaned specimens, and the name of the artist. The second footnote is the author's address. Ad- dresses for additional authors are each separate foot- notes. No other footnotes occur, except in tables, where unavoidable. 4. Abstract O A one-paragraph abstract precedes text. Papers in Spanish al an English abstract in addition to a ages resumen. O The Ei iun is concise (1 paragraph) and includes brief statements about the paper's intent, materials and methods, results, and significance of findings. 5. Tables O Tables are neat, double-spaced, and easily understood rather than long and complex. L] Tables do not contain vertical or horizontal lines. [Ed- itor and/or printer will add them as needed.] O Captions are typed double-spaced as paragraphs at the tops of the tables O Each table starts on a separate sheet and is double- spaced. 6. Abbreviations O Periods are used after all abbreviations (which are minimized) except metric measures, compass direc- tions, and herbarium designation O When dates are given as part of colled information, three-letter month abbreviations are used, except for months with four letters, which are spelled out in full. O States are not abbreviated, and cities are spelled out. [St. is acceptable.] O Periodicals are abbreviated according to B-P-H (Bo- tanico-Periodicum-Huntianum) and to B-P-H/S (Bo- tanico-Periodicum-Huntianum/Supplementum). L] Authors’ names are abbreviated ro isi to Authors of Plant Names or spelled out in O Book titles are abbreviated par et to Taxonomic Lit- erature, edition 2, but with initial letters capitalized. Book titles are spelled out in the Literature Cited. O Authors’ names in major taxonomic citations, such as the formal citation at the head of a taxonomic treat- ment, are spelle O Herbaria are abbreviated according to Index Herba- riorum, edition 7. Taxonomic Treatment O One paragraph per basionym is used as ee Tax- on author, literature citation, type citation, e.g., Pleu- rothyrium amplifolium (Mez) Rohwer, Mitt, Inst. Allg. 3. 1986. Nectandra amplifolia Mez, Arbeiten Kónigl. Bot. Gart. Breslau 1: 131. 1892. TYPE: Brazil. Rio de Janeiro Alto Macahé, Glaziou 17731 (holotype, B; isotypes, B, G, K, NY, P). Figure 11 O Lectotype designations are included together with an indication of where they were designated, the year, and the author. This reference is listed in the Literature Cited. If the author of the paper submitted is making the lectotypification, the phrase “selected here” is used O Exclamation points are not used for specimens exam- ined. Types not seen are indicated as such (e.g., MO, US not seen) O A brief Latin diagnosis for each new taxon is provided rather than a complete Latin description. O When relevant, nomina nuda, misapplied names and superfluous names are included in the discussion fol- lowing the description, but are not part of the formal synonymy. 8. Specimens Examined O If many specimens were examined, those cited in the text are limited to one collection per geographical area. O Long lists of specimens examined are placed at the end of the paper, following the Literature Cited. They are arranged alphabetically by collector, followed by pipes ue followed by the number of the taxon in the Specimens « are cited in the text as follows: Additional specimens examined (or Selected s aaie examined or Paratypes). "MEXICO, Oaxaca: Sie n Pedro No- lesco, Talea, 3 Feb. 1987 (fl), podia 865 (BM, G, K, US). [Dates iis d iei status are optional.] Countries are run together in the same rhe e.g., COUNTRY A. Major oli d division: ... COUNTRY B. Major political division: . . ,u 9. Vouchers and Genetic Sequences C) If paper presents original data, it includes citation of vouchers. [Depending on the type of paper, reference to the original wild source may also be required.] 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 O Genetic sequences have been banked, and accession numbers are provide O Author accepts responsibilty for establishing the ac- curacy of information provi O 10. Keys s are clear and have been checked carefully for consistency with the descriptions. 11. Literature Cited O The Literature Cited contains all references cited in the text O All entries in the Literature Cited are cited in the text. O Spelling of author(s) name(s) and years of publication have been double-checke O All entries have been verified against original sources, especially journal titles, accents, diacritical marks, and spelling in languages other than English. Periodicals are listed as follows: author's last name, initial(s). Year. Full title of article. Journal abbreviated as in B-P-H/S. Volume: pages. No parenthetical part numbers after volume (ric tin are given unless each part is paginated na ately. For more than on iniri this style is followed: a thors last name, initial(s) second author's initial(s), last name & third author's initial(s), last name O Books appear as follows: author's last name, initial(s). Year. Full Unabbreviated Title (edited by Editor), 3rd ed., Vol. 2. Publisher, City of P Publication O Citations of work “in prep.,” unpublished theses and dissertations, and similar references to inaccessible sources have been eliminated or kept to a minimum. They are not included in the Literature Cited. = O 12. Illustrations O s name(s) and hee number(s) are written in ncil on the back of each figure or plate, on both originals and review copies. C] Scale bars appear on illustrations, photographs, and maps. m r E tee are not indicated in cap- tion O All isa — are mounted on stiff card- board no thic in. O Illustrations are prisa to fit either es width (2% in. or ca. 68 mm) or full page width (5% in. or ca. 140 mm), or illustrations no larger than 11 X x18 in. (7 29 X 40 cm). [Maximum size for printed illustrations is 514 X 8% in.] Oversized or delicate figures are sub- mitted as photostats. [Line drawings may be submitted as quality photocopies. Figures are numbered in Arabic numerals in the order of their citation in the text. Parts of figures are labeled with capital letters. O Photographs are crisp black-and-white prints. [] Figures are grouped into composite plates when pos- sible; edges of photographs are abutted. O No stripping is inserted between plate or figure seg- ments. [Printer will insert stripping. O Edges of figures are squared. C] Maps include reference to latitude and longitude and are bounded by a fine border. O Scanning electron micrographs are free of conspicuous charging. O Axes on graphs are all labeled O Captions provide all explanatory text. No text appears on the figures. Captions are separate from other text, one paragraph for each group of figures, and following the style in current issues of the Annals O Symbols on maps are legible and reduction has been taken into consideration [1 13. What to Submit O Full mailing and e-mail addresses, oi and fax numbers are subm si in cover lette O Three copies of manuscript are eid ed. O ree Cae e original figures are enclosed with first submission. Copies of half-tone figures (pho- tographs) and plates must be of sufficiently good qual- ity to show all the detail of the originals. [Send origi- nals or photostats/photocopies of originals when requested—usuall O An MS-DOS/WINDOWS 3% in Perfect® ita a à is subiailted in ‘addition to the printed document SCOPE OF THE ANNALS The Annals publishes original articles in systematic bot- any and related fields. Papers who se is the est lishment of new nomenclature in vascular plants and bryo- phytes are not accepted. Rather, they should be submitted to Novon for consideration. (Novon instructions are avail- able from the managing editor. CONDITIONS FOR PUBLICATION Authors are urged to contact the managing editor prior to submitting manuscripts. Manuscripts sent already re- viewed and revised are generally subject to the full normal review process initiated here. Manuscripts not properly prepared may be returned for revision prior to review. It is the author's saci ER to supply missing biblio- graphic Ten the editor does not perform library research except by prior arrangement with authors not having access to crucial reference materials. Communi- cation with the editor at all points of the publication pro- cess is encouraged. PAGE CHARGES Page costs are $80 per page, although charges may be reduced under certain circumstances. Corrections in proof will be billed nonnegotiably to the author at $3 per line changed. CORRESPONDENCE Amy Scheuler McPherson, anaging Editor Scientific Publications Missouri Botanical Garden P.O. Box 299 St. Louis, MO 63166-0299, U.S.A Tel: (314) 577-5112 Fax: (314) 577-9594 Electronic mail: amcpher@ admin.mobot.org Volume 83, Number 4, pp. 433-583 of the ANNALS OF id MIssouRI BOTANICAL GARDEN was published on November 27, ANNALS OF THE MISSOURI BOTANICAL GARDEN VOLUME 83 1996 Colophon This volume of the ANNALS 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 8 point type. This volume has been printed on 70# Vintage Gloss. This is an acid-free paper designed to have a shelf-life of over 100 years. Vintage Gloss is manufactured by the Potlatch Paper Company. 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. The ANNALS is printed and distributed by Allen Press, Inc. of Lawrence, Kansas 66044, U.S.A. © Missouri Botanical Garden 1996 ISSN 0026-6493 VOLUME 83 ALVARADO-LINDNER, ADA BELINDA. (See John A. Beutler, Ada Belinda Alvarado-Lindner & Thomas G. McCloud) BALDWIN, BRUCE G., MICHAEL J. SANDERSON, J. MARK PORTER, MARTIN F. WOJCIECHOWSKI, CHRISTOPHER S. CAMPBELL & MI- CHAEL J. DONOGHUE. Erratum for The ITS Region of Nuclear Ribo- somal DNA: A Valuable Source of Evidence on Angiosperm Phylog- eny BALICK, MICHAEL J. Transforming Ethnobotany for the New Millenium -...... BENÍTEZ, HESIQUIO. (See Jorge Soberón, Jorge Llorente € Hesiquio- Benítez) . BEUTLER, JOHN A., ADA BELINDA ALVARADO-LINDNER & THOMAS G. MCCLOUD. Further Studies on Phorbol Ester Bioactivity in the Eu- phorbiaceae CAMPBELL, CHRISTOPHER S. (See Bruce G. Baldwin, Michael J. Sander- son, J. Mark Porter, Martin F. Wojciechowski, Christopher S. Campbell & Michael J. Donoghue) CHRISTOPHEL, D. C., R. KERRIGAN & A. I. ROWETT. The Use of Cutic- ular Features in the Taxonomy of the Lauraceae COLLETTE, BRUCE B. (See Michael Vecchione & Bruce B. Collette) — DIRZO, RODOLFO & GUILLERMINA GÓMEZ. Ritmos Temporales de la Investigación Taxonómica de Plantas Vasculares en México y una Esti- mación del Námero de Especies Conocidas DONOGHUE, MICHAEL J. (See Bruce G. Baldwin, Michael J. Sanderson, J. Mark Porter, Martin F. Wojciechowski, Christopher S. Campbell & Mi- chael J. Donoghue) FRONDORF, ANNE & GARY WAGGONER. Systematics Information as a Central Component in the National Biological Information Infrastruc- ture GOLDBLATT, PETER & JOHN C. MANNING. Phylogeny and Speciation in Lapeirousia Subgenus Lapeirousia (Iridaceae: Ixioideae GOLDBLATT, PETER. (See John C. Manning & Peter Goldblatt) —__ GOMEZ, GUILLERMINA. (See Rodolfo Dirzo & Guillermina Gómez) — HARDER, DANIEL K. (See Nguyen Nghia Thin & Daniel K. Harder) — HAYDEN, SHEILA M. € W. JOHN HAYDEN. A Revision of Discocarpus (Euphorbiaceae) HAYDEN, W. JOHN. (See Sheila M. Hayden & W. John Hayden) —— KERRIGAN, R. (See D. C. Christophel, R. Kerrigan & A. I Rowett) —— LANE, MEREDITH A. Roles of Natural History Collections LIEDE, SIGRID. A Revision of Cynanchum (Asclepiadaceae) in Africa — LINDER, H. P. (See D. A. Snijman & H. P. Linder) 1996 530 530 151 419 29 396 151 LIPSCOMB, DIANA. A Survey of Microbial Diversity LLORENTE, JORGE. (See Jorge Soberón, Jorge Llorente & Hesiquio Benítez) MANNING, JOHN C. € PETER GOLDBLATT. The Prosoeca peringueyi (Dip- tera: Nemestrinidae) Pollination Guild in Southern Africa: Long-tongued Flies and Their Tubular Flowers MANNING, JOHN C. (See Peter Goldblatt € John C. Manning) MANNING, STEPHEN D. Revision of Pavetta Subgenus Baconia (Rubiaceae: Ixoroideae) in Cameroon MCCLOUD, THOMAS G. (See John A. Beutler, Ada Belinda Alvarado-Lindner & Thomas G. McCloud) MILLER, DOUGLASS, R. (See Amy Y. Rossman & Douglass R. Miller) ....... MILLER, JAMES S. Alwyn Howard Gentry, 1945-1993: A Tribute. MONSON, RUSSELL K. The Use of Phylogenetic Perspective in Comparative Plant Physiology and Developmental Biology MOSYAKIN, SERGEI L. A Taxonomic Synopsis of the Genus Salsola (Che- nopodiaceae) in North America NEVLING, LORIN I. The Last Species NIC LUGHADHA, E. & C. PROENCA. A Survey of the Reproductive Biology of the Myrtoideae (Myrtaceae) OLIVER, JAMES H., JR. Importance of Systematics to Public Health: Ticks, Microbes, and Disease PORTER, J. MARK. (See Bruce G. Baldwin, Michael J. Sanderson, J. Mark Porter, Martin F. Wojciechowski, Christopher S. Campbell & Michael J. Donoghue) PROENCA, C. (See E. Nic Lughadha & C. Proenca) RICHARDSON, P. MICK. Systematics Agenda 2000: Systematics and Society, the 41st Annual Systematics Symposium of the Missouri Botanical Gar- den. Introduction RICHARDSON, P. MICK. National Biological Service, the 42nd Annual Sys- tematics Symposium of the Missouri Botanical Garden. Introduction ..... RICHTER, H. G. (See Henk van der Werff & H. G. Richter) ROSSMAN, AMY Y. & DOUGLASS R. MILLER. Systematics Solves Problems in Agriculture and Forestry ROWETT, A. I. (See D. C. Christophel, R. Kerrigan & A. I. Rowett) SANDERSON, MICHAEL J. (See Bruce G. Baldwin, Michael J. Sanderson, J. Mark Porter, Martin F. Wojciechowski, Christopher S. Campbell & Mi- chael J. Donoghue) SNIJMAN, D. A. & H. P. LINDER. Phylogenetic Relationships, Seed Char- acters, and Dispersal System Evolution in Amaryllideae (Amaryllida- ceae 951 151 362 SNOW, NEIL. The Phylogenetic Utility of Lemmatal Micromorphology in Lep- tochloa s.l. and Related Genera in Subtribe Eleusininae (Poaceae, Chlor- idoideae, Eragrostideae) SOBERON, JORGE, JORGE LLORENTE & HESIQUIO BENITEZ. An Inter- national View of National Biological Surveys TAYLOR, CHARLOTTE M. Taxonomic Revision of Cruckshanksia and Oreo- polus (Rubiaceae: Hedyotideae) THIN, NGUYEN NGHIA & DANIEL K. HARDER. Diversity of the Flora of Fan Si Pan, the Highest Mountain in Vietnam VANE-WRIGHT, R. I. Systematics and the Conservation of Biological Diver- sity VECCHIONE, MICHAEL & BRUCE B. COLLETTE. Fisheries Agencies and Marine Biodiversity WAGGONER, GARY € ANNE FRONDORE. (See Anne Frondorf & Gary Wag- goner WERFF, HENK VAN DER & H. G. RICHTER. Toward an Improved Classi- fication of Lauraceae WOJCIECHOWSKI, MARTIN F. (See Bruce G. Baldwin, Michael J. Sander- son, J. Mark Porter, Martin F. Wojciechowski, Christopher S. Campbell & Michael J. Donoghue) YATSKIEVYCH, GEORGE. A Revision of the Fern Genus Phanerophlebia (Dryopteridaceae) ZULOACA, FERNANDO O. & OSVALDO MORRONE. Revisión de las Es- pecies Americanas de Panicum Subg. Panicum Secc. Panicum (Poaceae: Panicoideae: Paniceae) 504 562 461 404 47 29 546 409 151 200 Experimental and Molecular Approaches to Plant Biosystematics The proceedings of the Fifth International Symposium of the International Organization of Plant Biosystematists (IOPB) Edited by Peter C. Hoch and A. G. Stephenson Twenty-three original contributions that span the breadth of biosystematics, a dynamic field of study that bridges the realms of systematics and population biology. The papers are arranged in four groups, reflecting the original four symposia of the 1992 meeting. DNA and Plant Biosystematics presents innovative work that uses the rapidly developing nucleic acid methods adapted from molecular biology. Plant Growth Patterns and Biosystematics includes comparative and developmental analyses of plant architecture and branching patterns. Plant Reproductive Strategies surveys new approaches in the anal- ysis of plant reproductive biology, an area central to both systematic and population-level studies. Phylogenetic Analysis and Population Biology emphasizes the application of the powerful new methods of phylogenetic analysis to problems at the species and population levels. Monographs in Systematic Botany from the Missouri Botanical Garden, Volume 53. ISBN: 0-915279-30-4. 416 pp. Illustrated. 1995. $60.00 U.S. $62.00 Non-U.S. Annals of the Missouri Botanical Garden, Volume 82, Number 2: Alternative Genes for Phylogenetic Reconstruction in Plants A symposium cosponsored by the American Society of Plant Taxonomists and the Botanical Society of America, organized by Pamela S. Soltis and Douglas E. Soltis, and presented at the 1993 AIBS meetings. Although the chloroplast gene rbcL has been successfully used to reconstruct plant phylogeny, many important questions of plant phylogeny and evolution cannot be addressed using it. The contributors to this issue of the Annals explore the potential of eight alternative genes or DNA regions for phylogenetic reconstruction at a variety of hierarchical levels. Both nuclear and chloroplast genes are evaluated. Three regions of the nuclear ribosomal RNA cistron are explored: the 18S gene, the internal transcribed spacers roplast genome are also considered: atpB, ndhF, and matK. Each paper describes the location, size, structure, and rate of evolution of the chosen gene and discusses its potential for phylogenetic study. This issue also contains: “The Comparative Pollination and Floral Biology of Baobabs (Adansonia—Bombaca- ceae)” by David A. Baum and “In Memoriam: Peter G. Martin.” Annals 82(2) 1995. 174 pages. $27.50 U.S. $28.00 Non-U.S. To order, please indicate method of payment below. Checks or money orders should be in U.S. funds, payable through a U.S. bank, to Missouri Botanical Garden. Orders must be prepaid. $3.00 handli fee on all orders. Additional $2.00 handling fee will be added to orders requiring invoices. No shipments are made until payment is received. Phone: (314) 577-9534, fax: (314) 577-9594, e-mail: . dept] 1@mobot.org. ts ca à ETAT. : “Send order tor. >. ae 0 Please send —— copy(ies) of Monograph No. 53 Department Eleven, Missouri Botanical Garden Please send —— copy(ies) of Annals 82(2) P.O. Box 299 St. Louis, MO 63166-0299, U.S.A. m Check/money order enclosed — - ; Send books to: O Send invoice ($2.00 fee will be added to total) - O Charge card number (MasterCard/Visa) | ane Expiration date meee si cca Address - Name as it appears on card Telephone number (daytime) MI Sad TR Postal Code Country - PRICES ARE SUBJECT TO CHANGE WITHOUT NOTICE 83(4) CONTENTS Alwyn Howard Gentry, 1945-1993: A Tribute James S. Miller et al. 433 Taxonomic Revision of Gruckshanissia and Oreopolus (Rubiaceae: Hedyotideae) ... Charlotte M. Taylor 461 A Survey of the Reproductive Biology of the CRORE (Myrtacea e) ic Lughadha & C. Proença 480 The Phylogenetic Utility of Lemmatal oaei in sph s.l. and Related Genera in Subtribe Eleusininae (Poaceae, Chloridoideae, Eragrostideae) — Neil Sie 504 Further Studies on Phorbol Ester Bioactivity in the Euphorbiaceae m E __. John A. Beutler, Ada Belinda Alvarado-Lindner & Thomas G. McCloud 530 National Biological Service, the 42nd Annual Systematics Symposium of the Missouri Botanical Garden Introduction to the Eu P. Mick Richardson 534 Roles of Natural History Collections - Meredith A. Lane 536 Systematics Information as a Central Component in the National Biological In- formation Infrastructure Anne Frondorf & Gary Waggoner 546 A Survey of Microbial Diversity .- | ass MANDO Peon 551 An International Yi of National Biological Surveys . Jorge Soberón, E Llorente & Hesiquio Benítez 562 The D SE E Lorin I. Nevling 574 Checklist for Authors Ste hn 581 Ea dineritión. Pavetta camerounensis subsp. camerounensis S. D. Manning, sp. nov., by Linda Ellis. Ce SIUS SR NI T RR a a RARE