A systematic revision of the genus Manekia (Piperaceae)

Presentation on theme: "A systematic revision of the genus Manekia (Piperaceae)"— Presentation transcript:

1 A systematic revision of the genus Manekia (Piperaceae)
Heather Schubert

2 Outline Introduction of Piperaceae Taxonomic history of Piperaceae
(Bornstein) Introduction of Piperaceae Taxonomic history of Piperaceae Introduction of Manekia Materials and methods Results Discussion Piper obliquum

3 Piperaceae (Agardh) Pantropical family Consists of 5 genera
2,400-3,600 species (Callejas 2001; Bornstein 2005; Wanke et al. 2007; Smith et al. 2008) Largest genera Peperomia and Piper Piperaceae is a large pantropical family with major centers of diversity in northern South America and Central America in the New World, and Malaysia in the Old World. The family consists of 5 genera (Piper, Peperomia, Manekia, Zippelia, ???) with species estimates in the range of 2400 to The largest genera in the family are Peperomia and Piper. Nearlly all taxa in this family have been placed in one of these two genera at one time or another.

4 Piperaceae Aromatic Perennial Shrubs, trees or vines (pipers)
(Bornstein) Aromatic Perennial Shrubs, trees or vines (pipers) Herbs (peperomias) Leaves Alternate, opposite, whorled or basal Sessile or petiolate Members of the Piperaceae are aromatic due to glands with ethereal oil cells or volatile cell. They can be perennial (a plant that lives more than two years) shrubs, trees or vines (Piper and relatives) and herbs (Peperomia). The leaves are variously arranged either alternate, opposite, whorled or basal. They can be sessile (attached directly by the base or without a stalk) or petiolate (with a stalk), the petiole often sheathing or vaginate. P. arboretum

5 Piperaceae Inflorescence Spicate or racemose Terminal or leaf-opposed
(Bornstein) Their inflorescence (flower-cluster of the plant) can be either spicate (arranged in a spike that is more or less elongated with sessile or subsessile flowers) or racemose (a more or less elongated inflorescence with pedicellate or stalked flowers). This structure can be terminal (at the end of a brach) or leaf-opposed. P. arboreum (Bornstein)

6 Piperaceae Flower Fruit
(Bornstein) Flower Small, perfect (Neotropical) or imperfect (Old World tropics), lack a perianth Subtended by a glabrous to pubescent bract Unilocular ovary with single, basal ovule Fruit drupe Members of the Piperaceae have small, perfect (has both male and female parts) or imperfect (has either male parts or female parts, but not both) flowers. They also lack a perianth (lack all of the sepals and petals). Each flower is subtended by a bract that can be glabrous (without hairs) to pubescent (with hair). If female reproductive structures are present the flower will have a unilocular ovary with a single, basal ovule. The fruit is a drupe (fleshy fruit with a firm endocarp that permanently encloses the usually solitary seed). P. abalienatum

7 P. schiedeanum P. pseudofuligineum
Scanning electron microscope (SEM) photos In fruit or flower. Left has somewhat glabrous to slightly pubescent bracts. Right has extremely pubescent bracts. (Bornstein) (Bornstein) P. schiedeanum P. pseudofuligineum

8 P. arboreum Infructescence. Can see individual fruits (drupes).
(Bornstein) Infructescence. Can see individual fruits (drupes). P. arboreum

9 Inflorescence immature with white flowers, mature with greenish brown fruits
(Bornstein) P. arboreum

10 Piper nigrum Black pepper Spice/seasoning
A few species of Piper are known for their economic value. The most important of these is Piper nigrum (black pepper). Drupe is used to make black pepper. (Bornstein)

11 Piper betle “Betle leaf” chewed with the areca nut
Used as herbal medicine and for narcotic properties (Bornstein) The leaf of Piper betle is wrapped around the areca nut and chewed on. Used as a stimulant, antiseptic, and a breath-freshner.

12 The Problem in Piperaceae
Circumscription Large number of described species Based on characters of questionable taxonomic value, or fragmented material lacking flowers Based on occurrence in a distinct political distributional unit Circumscription is a major problem in the Piperaceae, especially at the species level. Taxonomic issues are further complicated by the large number of described species, combined with descriptions that have often been based on characters of questionable taxonomic value, or on fragmentary material lacking flowers and especially fruits necessary for identification (Bornstein 1989). New species have also been established based on their occurrence in a distinct political distributional unit, despite their similarity with previously described species (Yuncker 1958), leading to an excessive number of described species.

13 Circumscription in Piper
Several generic segregates established (Kunth 1839; Miquel ) Reduced to synonymy or recognized at sectional level (C. de Candolle 1869) Additional genera created (Trelease and Yuncker 1950; Yuncker 1958) Phylogenetic analysis confirmed Manekia generic status (Jaramillo et al. 2004) The problem of circumscription within Piper began when Kunth (1839) and Miquel ( ) established several generic segregates from Piper, which were later reduced to synonomy or recognized at the sectional level by C. de Candolle (1869). Additional genera were later created by Trelease and/or Yuncker (1950; Yuncker 1958), including the current taxa of interest Sarcorhachis and Manekia. A phylogenetic analysis of DNA sequence data confirmed that Sarcorhachis deserved generic status (Jaramillo et al. 2004), along with Piper, Peperomia and Zippelia.

14 Circumscription: Manekia vs Sarcorhachis
Sarcorhachis and Manekia recognized (Trelease 1927) Placed in tribe Pothomorpheae Single species of Manekia, M. urbanii Haiti Six species and two varieties of Sarcorhachis (Steyermark 1971) Costa Rica to Colombia, Ecuador, Peru, Venezuela, Brasil and the Lesser Antilles In 1927, William Trelease recognized Sarcorhachis and Manekia and placed both in the tribe Pothomorpheae, based on their axillary inflorescences. Sarcorhachis was recognized in his treatment of the Piperaceae of Panama. Manekia was established in his treatment of the Piperaceae of Hispaniola, two months prior to Sarcorhachis. He recognized a single species, M. urbanii, from Haiti. No other species were ever described in this genus. In 1971, Steyermark recognized six species of Sarcorhachis with a range in Centeral America and northern South America.

15 Sarcorhachis and Manekia
Shared a multitude of features Taxonomic synonyms (Bornstein 1996) Principle III of the International Code of Botanical Nomenclature Earlier name deserves priority Bornstein (1996) proposed conserving the name Sarcorhachis Arias et al. (2006) provided most new name combinations Upon examination, Bornstein (1996) realized that Sarcorhachis and Manekia shared a multitude of features supporting taxonomic synonymy. According to principle III of the International Code of Botanical Nomenclature, the earlier name Manekia deserves priority over Sarcorhachis. Bornstein (1996) proposed to conserve the name Sarcorhachis against Manekia to minimize the number of name changes when merging the two genera, since it was the more established name in botanical literature. The proposal was rejected, therefore resulting in the valid taxa of Sarcorhachis requiring transfer to Manekia. Arias et al. (2006) provided most of the new combinations and postulated that further analysis might lead to taxonomic synonymy, reducing the number of species in the genus.

16 Manekia Scandent habit Alternate, palmately veined leaves
Petiole vaginate Spikes usually axillary, occasionally terminal (solitary or paired) Stamens 4 Stigmas 3-5 Fruits laterally compressed and immersed in fleshy rachis Rachis pubescent Members of Manekia have a scandent or climbing habit. Their leaves are alternate and palmately veined (with three or more veins arising from a common point). They have an expanded petiole that is torn open upon emergence of the enclosed shoot apex and inflorescence, initially vaginate with deciduous wings. The inflorescence is usually spicate and can be either axillary (located in or arising from the point of the angle between a stem and a leaf) or occasionally terminal. The spikes can be solitary or paired. The flowers are perfect (meaning they have both male and female parts). Each flower has 4 stamens (male) and 3-5 stigmas (upper female reproductive structure). Fruits are laterally compressed and are immersed or embedded in a pubescent fleshy rachis (main axis of the inflorescence).

17 M. urbanii M. incurva M. venezuelana M. naranjoana M. sydowii
M. obtusa Like mentioned earlier, the members of Manekia are found in Central America and northern South America. Manekia incurva=Lesser Antilles: Dominica, Guadeloupe, Martinique, Puerto Rico M. naranjoana=Costa Rica, Panama M. obtusa=Brasil M. sydowii=Colombia, Equador, Peru M. venezuelana=Venezuela M. urbanii=Haiti ( and southamerica)

18 Systematic revision of Manekia
Morphological study Examination of type specimens and other herbarium material Molecular investigation Examined three DNA regions Internal transcribed spacer (ITS) psbj-petA intergenic intron rpl32 spacer Phylogenetic analysis pending For my thesis I worked on a systematic revision of Manekia. This study was two-fold: a traditional morphological study combined with a modern molecular investigation. The morphological study included the examination of type specimens and other herbarium material to analyze qualitative and quantitative features. The molecular analysis examined three DNA regions: The internal transcribed spacer (ITS) within ribosomal DNA located in the nucleus. Plus the psbj-petA intergenic intron and rpl32 spacer, both part of the chloroplast genome. A phylogenetic analysis using the molecular sequences is pending and will be used to create a working hypothesis in order to evaluate the evolution of morphological features that can be used to discriminate between species. In addition to the phylogenetic diagram, a formal taxonomic revision will be created, including a key to the species, type citations and synonymy, and associated species descriptions based on significant morphological data.

19 ITS = a piece of non-functional RNA situated between structural ribosomal RNAs (rRNA)

20 Circular structure of the chloroplast genome of Lolium perenne
Lolium perenne = Ryegrass Diekmann, K. et al. DNA Res :dsp008v1-8; doi: /dnares/dsp008

21 Herbarium Materials A, BM, C, ILL, F, G, GH, K, MO, NY, P, RB, US and WIS Destructive sampling MO, NY, US Collected after 1980 Wide distribution For my study I used several specimens from herberia all over the world. These letters are acronyms for the different herberia. A=Arnold Arboretum BM=British Museum in London England MO=Missouri Botanical Gardens Herbarium I recorded label information for each specimen used in my study. I did destructive sampling (removed vegetative material from the packet on the sample or actually broke some of the leaf off) on some of the specimens from MO, NY, and US. The specimen had to be collected after 1980, this meant it would likely have viable DNA than older specimens. The specimens choosen were also from a wide distribution for each species to see genetic variablity within a species.

22 Morphological Characteristics
Vegetative Reproductive Leaf Shape Base Apex Venation Pubescence Texture/Color Stipule Prophyll Stem Inflorescence/Infructescence Position Rachis Bract Stamen position Anther Carpel Stigma Fruit I recorded notes on a multitude of vegetative and reproductive structures. Vegetative characteristics included leaf, stipule, prophyll and stem morphology. Reproductive features included inflorescence morphology, like position (whether it was axillary, terminal) rachis pubescence, bract pubescence, stamen position, anther morphology, carpel morphology, stigma number, fruit morphoogy.

23 Morphological Measurements
Internode length Prophyll length Leaf blade length and width Petiole length Petiole vagination distance Inflorescence/ Infructescence length Peduncle length Anther length Fruit length In addition to taking notes on state, or presence or absence I took several measurements on vegetative and reproductive features. Mostly was length and width measurements.

24 Manekia naranjoana Petiole vagination
Can see that the leaf shape is ovate, base is obtuse, apex is somewhat acuminate. Leaf venation is palmate with 7 main veins. Petiole is vaginate the entire length and extends onto the blade. You can also see the propyll and could take notes on it as well. Internode length. Manekia naranjoana Petiole vagination

25 Methods: Extraction and amplification
Quiagen Dneasy extraction kit PCR/Amplification ITS psbj-petA rpl32 Amplifications run in MJ Research programmable thermal cyclers PTC and PTC Now I will begin to discuss the molecular investigation materials and methodology. To extract the DNA is used a Quiagen Dneasy extraction kity and followed protocol. The genes I was amplifying was ITS, psbj-petA and rpl32 markers. Amplifications were run in a MJ Research programmable thermal cycler.

26 Methods: PCR and cloning
Visually verified PCR products on 0.7% agarose gels Gel purification using Millipore Montage DNA Gel Extraction kit Cloned using a Promega PGEM T-easy cloning kit PCR was run in 0.7% agarose gels and the products were then visually verified. On this gel you can see the primer dimer band as well as the DNA that I was trying and sucessfully amplified. To the right of each row is the 100 base pair ladder to determine if the gene is falling out in the correct location. The bands were cut out of the gel and purified using a Millipore Montage DNA gel extraction kit using manufacturer’s protocol. Cloning was competed for each sample using a Promega PGEM T-easy cloning kit. Why clone????????

27 Methods: Sequencing Li-Cor LongreadIR 4200 automated sequencer
Verified with BLAST searches Edited using Geneious Pro 4.7.6 Aligned using CLUSTAL Submitted to Genbank upon publication After cloning the samples were loaded into the Li-Cor automated sequencer. Results were edited using Genious and aligned using Clustal Each sequence was verified with BLAST searches and submitted to Genebank upon publication.

28 Statistical and Phylogenetic Analyses
Canonical Discriminant Analysis Maximum Likelihood Phylogenetic Analysis with Bootstrap values (Bornstein) An Average Linkage Cluster Analysis with Unweigthed Pair Group Method Analysis was performed on the averages of 41 morphological variables to demonstrate the similarity between species. A Canonical Discriminate Analysis to discriminate between taxa, was performed on 41 morphological variables and revealed 5 canonical axes at the p<0.05. P. friedrichsthalii

29 Canonical Discriminant Analysis
Axis Eigenvalue Cumulative Percent Likelihood Ratio Approximate F Value Num DF Den DF p 1 4.7047 43.82% 5.47 175 506.14 <0.0001 2 2.5564 67.63% 4.27 136 408.69 3 1.6936 83.40% 3.50 99 309.27 4 1.1061 93.70% 2.86 64 208 5 0.6763 100.00% 2.29 31 105 0.0010 Results from the canonical discriminant function analysis showing the 5 significant canonical axes all with a p<0.05. Direct attention to cumulative percent. The 1st canonical axis represents almost 44% of the variation among species. The 1st and 2nd canonical axis combined represents nearly 68% or 2/3 or the variation among the species. I will focus most of the attention on the 1st and 2nd canonical axes since they represent most of the variation for simplicity.

30 Petiole vagination stops prior to blade
7 main veins Range for each species plotted on canoncical axis 1 and 2. The dominant discriminating variables of canonical axis 1 were petiole vagination and bract ciliation. petiole vagination extends onto the blade petiole vagination stops prior to reaching the blade bract ciliate (hairs along entire margin) bract partly ciliate (hairs along a portion of the margin of the bract) Dominant discrimant variables of canonical axis 2 was the number of main veins ^ mv=7 v mv=5 You can see that M. obtusa and M. urbanii are falling out separately. M. incurva also seems to be falling out separate based on these two axes. Naranjoan, sydowii, and venezuelana are clustering together suggesting they are very similar. M. venzuelana is based on one individual therefore the range cannot be determined. Only one individual was used because most of the venezuelana were missing flowers and fruit. So, looking at this figure is seems that there are four distinct taxa: M. obtusa, M. urbanii, M. incurva, and the sydowii, naranjoana, venezuelana complex. 5 main veins Petiole vagination stops prior to blade Bract partly ciliate Petiole vagination extends onto blade Bract ciliate

31 Leaf base obtuse, deltoid, oblique
5 or 9 main veins Leaf base cordate 7 main veins Now we are looking at canoncial axis 2 with the number of main veins being the dominant discriminating variable and canonical axis 3 dominated by if the number of main veins is 7 and whether the leaf base is cordate. Once again there is good resolution separating out M. urbanii. Obtusa fall out somewhat separate but overlaps a little, and incurva separates but also overlaps. Naranjoana, sydowii and venezuelana all overlap and do not show any resolution. Still only the one venezuelana individual. Again you can see that there are perhaps four taxa, but the resolution is not quit as distinctive. 5 main veins 7 main veins

32 naranjoana (Costa Rica) rpl32 spacer 550 nucleotides sydowii (Peru)
67 sydowii (Peru) sydowii (Peru) sydowii (Peru) 92 sydowii (Colombia) sydowii (Ecuador) 52 naranjoana (Panama) venezuelana (Venezuela) naranjoana (Costa Rica) 86 venezuelana (Venezuela) sydowii (Peru) sydowii (Peru) venezuelana (Venezuela) incurva (Dominica) 100 incurva (Dominica) 69 obtusa (Brasil) This is a likelihood tree with bootstrap values created in PAUP using the chloroplast gene region, rpl32. What you see here is that using this marker the individuals were broken up into two clades which has a bootstrap value of That means that 100% of the bootstrap trees created by PAUP had these two clades broken apart. In the lower clade you can see that incurva and obtusa fall out there. Incurva has a Bootstrap value of 100 and obtusa has a BS of 69. In the upper clade you find naranjoana, sydowii and venezuelana. You also find a couple naranjoana and sydowii in the lower clade. The reason for the two separate clades is due to an apparent inversion in the rpl32 gene region. The lower clade has the inversion and the upper clade does not. This suggests that incurva and obtusa have retained this inversion from an ancestor. So the question arises of why would some naranjoana and sydowii have the inversion and not the others? One possibility is that they were misidentified. Therefore, we had a look at them and found that not to be the case. Another possibility is that the inversion was widely distributed throughout the ancestor population and the inversion was obtained in obtusa and incurva and sometimes in naranjoana and sydowii. Therefore, if you took several samples from their distribution you would find a mixture of individuals with and without the inversion allele. The inversion is not being selected against and can be carried throughout the population because it does not affect reproduction. If you remember rpl32 is a spacer and is not expressed in any way it is basically junk DNA that gets coded. So looking at this tree you can infer that incurva is a separate taxon, and obtusa fall out separately but does not have good resolution. However, the morphological data supported that it is indeed a separate taxon. And similar to the morphological data the upper clade is a complex of naranjoana, sydowii and venezuelana. Therefore, there are three separate taxa represented by this tree. M. Urbanii was not included in the phylogenetic analysis because no viable DNA was obtained form the herbarium specimen. obtusa (Brasil) 100 75 naranjoana (Panama) 0.5 substitutions/site sydowii (Peru) sydowii (Colombia) naranjoana (Costa Rica) obtusa (Brasil) obtusa (Brasil) naranjoana (Costa Rica)

33 Discussion 4 distinct taxa M. incurva M. obtusa M. urbanii
M. naranjoana Basionym: Piper naranjoanum C. DC., Linnaea 37: Basionym: Sarcorhachis sydowii Trel., Repert. Spec. Nov. Regni Veg. 48: Basionym: Sarcorhachis venezuelana Steyermark, Pittiera 3: So, from the morphological and molecular analyses it seems that there are 4 distinct taxa.

34 Discussion Better resolution Future Studies
Use ITS, psbj-petA and rpl32 individually and combined M. venezuelana - morphological analysis M. urbanii – genetic analysis Future Studies Establish phylogenetic and biogeographic relationships between Manekia and Zippelia Better resolution may be obtained using the other two genetic markers and then combining all three. What I presented today was just preliminary genetic data. It would be better to have more venezuelana specimens with reproductive features to include in the morphological analysis to determine how much it overlaps with naranjoana and sydowii. It would be great to get DNA sequences for M. urbanii. This species is only found in Haiti and it is unkown whether it still exists due to the vast deforestation in Haiti. If you remember this is a vine that grows up trees. It is found in montane regions therefore it still may exist. Also, would be great to get M. urbanii specimens for the genetic analysis.

35 Acknowledgements I thank Southeast Missouri State University for allowing me to conduct this research. I acknowledge Dr. Allan J. Bornstein and Dr. James F. Smith for obtaining specimens. I extend thanks to Boise State University (BSU), especially Dr. James F. Smith, for allowing me to use the necessary equipment and teaching me the techniques needed for this study. Also, I thank Dr. Michael S. Taylor for allowing me to use the necessary equipment in his lab. I thank Dr. Bornstein, Dr. Smith and Dr. Taylor for comments and discussions; Dr. John S. Scheibe for assistance with statistical analyses; my committee members, Dr. Bornstein, Dr. Taylor and Dr. Michael Aide, for assistance and encouragement throughout this study; and Jay Zimmers from BSU for helping with molecular work. This study was funded by the College of Graduate Studies, the College of Science and Mathematics, the Department of Biology, Dr. Allan J. Bornstein, Dr. Michael S. Taylor, and Dr. James F. Smith (BSU). I thank Southeast Missouri State and my committee (Dr. Bornstein, Dr. Taylor and Dr. Aide) for encouragement, discussions and comments. Dr. Bornstein for obtaining specimens Dr. Taylor for allowing me to use his lab Dr. Scheibe for assistance with statistical analyses Boise State University and Dr. Smith for teaching me techniques and using equipment I thank the College of Graduate Studie, College of Science and Mathematics, Department of Biology, Dr. Taylor, Dr. Smith and Dr. Bornstein for funding this project.

36 References Arias, T., R. Callejas, and A. Bornstein New combinations in Manekia, an earlier name for Sarcorhachis (Piperaceae). Novon 16: Bornstein, A Taxonomic studies in the Piperaceae—I. The pedicellate pipers of Mexico and Central America (Piper subg. Arctottonia). Journal of the Arnold Arboretum 70: The Piperaceae in the southeastern United States. Journal of the Arnold Arboretum, Supplementary Series 1: Proposal to conserve the name Sarcorhachis against Manekia (Piperaceae). Taxon 45: Piperaceae. Pp in A Tropical Garden Flora, eds. G. Staples & D.R. Herbst. Honolulu: Bishop Museum Press. Callejas, R Piperaceae. Pp in Flora de Nicaragua, Tomo III, eds. W. D. Stevens, C. Ulloa, A. Pool, and O. M. Montiel. St. Louis: Missouri Botanical Garden Press. Candolle, C. de Piperaceae. DC. Prodr. Syst. Nat. Veg. 16(1): Committee for Spermatophyta (1223) Conserve Sarcorhachis Trel. Against Manekia Trel. (Piperaceae). Taxon 47: 870. Holmgren, P. K., N. H. Holmgren and L. C. Barnett (editors) Index Herbariorum Part I: The Herbaria of the World, 8th edition. New York Botanical Garden, Bronx, NY. Jaramillo, M. A., Callejas, R., Davidson, C., Smith, J. F., Stevens, A. C. and E. J. Tepe A phylogeny of the tropical genus Piper using ITS and the chloroplast intron psbj-petA. Systematic Botany 33(4): Jaramillo, M. A., P. S. Manos, and E. A. Zimmer Phylogenetic relationships of the perianthless Piperales: Reconstructing the evolution of floral development. International Journal of Plant Science 165(3): Kunth, K. S Bemerkungen über die Familie des Piperaceen. Linnaea 13: McNeill Miquel, F. A. G Systema Piperacearum. Rotterdam. Möller, M and QCB Cronk Origin and relationships of Saintpaulia H. Wendl. (Gesneriaceae) based on ribosomal DNA internal transcribed spacer (ITS) sequences. American Journal of Botany 84: Shaw, J., Lickey, E. B., Schilling, E. E. and R. L. Small Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: The tortoise and the hare III. American Journal of Botany 94(3): Smith, J. F., A. C. Stevens, E. J. Tepe and C. Davidson Placing the origin of two species-rich genera in the late cretaceous with later species divergence in the tertiary: a phylogenetic, biogeographic and molecular dating analysis of Piper and Peperomia (Piperaceae). Plant Syst Evol 275: Standley, P. C. and J. A. Steyermark Piperaceae. In: Flora of Guatemala. Fieldiana, Bot. 24(3): Steyermark, J. A Notes on the genus Sarcorhachis Trel. (Piperaceae). Pittieria 3: Trelease, W The Piperaceae of Panama. Contributions from the United States National Herbarium 26(2): Trelease, W. and T. G. Yuncker The Piperaceae of northern South America. 2 vols. Urbana. Wanke, S., L. Vanderschaeve, G. Mathieu, C. Neinhuis, P. Goetghebeur, and M. S. Samain From forgotton taxon to a missing link? The position of the genus Verhuellia (Piperaceae) revealed by molecules. Annals of Botany 99: Wen, J and EA Zimmer Phylogeny and biogeography of Panax L. (the ginseng genus, Araliaceae): inferences from ITS sequences of nuclear ribosomal DNA. Molec Phylogenet Evol 6: Yuncker, T. G The Piperaceae—A family profile. Brittonia 10: 1-7.

37 Questions? (Bornstein) P. bredemeyeri

38 Herbarium codes for herbaria loaned from for this study.
INSTITUTION LOCATION A Arnold Arboretum, Harvard University U.S.A. Massachusetts. CAMBRIDGE. BM The Natural History Museum U.K. England. LONDON. C University of Copenhagen Denmark. COPENHAGEN. ILL University of Illinois U.S.A. Illinois. URBANA. F Field Museum of Natural History U.S.A. Illinois. CHICAGO. G Conservatoire et Jardin botaniques de la Ville de Genève Switzerland. GENÈVE. GH Harvard University K Royal Botanic Gardens U.K. England. KEW. MO Missouri Botanical Garden U.S.A. Missouri. SAINT LOUIS. NY New York Botanical Garden U.S.A. New York. BRONX. P Muséum National d'Histoire Naturelle France. PARIS. RB Jardim Botânico do Rio de Janeiro Brazil. Rio de Janeiro. RIO DE JANEIRO. US Smithsonian Institution U.S.A. District of Columbia. WASHINGTON. WIS University of Wisconsin U.S.A. Wisconsin. MADISON.