Annals of the Missouri Botanical Ga rden umber Volume 90, Number 1 Winter 2003 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. All manuscripts are peer-reviewed by quali- fied, independent reviewers. Authors should write the Managing Editor for informa- tion concerning arrangements for publishing in the ANNALS. Instructions to Authors are printed in the back of the last issue of each volume and are also available online at www.mobot.org/mbgpress. Editorial Ccmmittee Victoria C. Hollowell Editor, Missouri Botanical Garden Amy Scheuler McPherson Managing Editor, Missouri Botanical Garden Diana Gunter Associate Editor, Missouri Botanical Garden Aida Kadunic Senior Secretary Barbara Mack Administrative Assistant 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 Charlotte Taylor Missouri Botanical Garden Henk van der Werff Missouri Botanical Garden For subscription information contact ANNALS OF THE MISSOURI GARDEN, % Allen Marketing .. & Management, Р.О. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2003 is .... $8145 per volume U.S., $155 Canada & Mexico, $180 all other countries. Four issues per vol- - .. ume. The journal Novon i is included i in the: sub- p seription price of the ANNALS. _annals@mobot.org (editorial PIE e —htp/www. mobot. org/mb gpress. T © Missouri Botanical Garden 2003 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. Periodicals postage paid at St. Louis, MO and additional mailing offices. POSTMASTER: Send address changes to ANNALS OF THE MISSOURI BoTANICAL GARDEN, % Allen Marketing & Management, P.0. Box 1897, Lawrence, KS 66044-8897. .. The mission of the Missouri Botanical Garden is to саг and share ere: about plants e — their скори! їп order to preserve and enrich life. | © This paper meets he requirements o ANSI/NISO 229 48-1992 Permanence ol Pape: Volume 90 Annals Number 1 of the 2003 Missouri Botanical Garden THE PARAPHYLY Nigel P. Barker,? Н. Peter Linder,” OF CORTADERIA Cynthia M. Morton," and Mark Lyle? (DANTHONIOIDEAE:; POACEAE): EVIDENCE FROM MORPHOLOGY AND CHLOROPLAST AND NUCLEAR DNA SEQUENCE DATA! ABSTRACT The genus Cortaderia Stapf еи s approximate ‘ly 24 species occurring in South America, New Zealand. and New Guinea. Cortaderia is placed in the subfamily Danthonioideae. Several species in the genus are known to be gyno- dioecious, a feature unusual for ils subfamily. The genus his been divided into four sections, two of which are Menon p This study utilizes nrDNA sequence data from the IFTS region and the grass-specific insert in the chloroplast rpoC2. gene to elucidate the relationships of 15 species of the genus, including the anomalous Cortaderia archboldii (Hitehe.) Connor & Edgar. Results suggest that the genus is not monophyletic. and there are two clades that correspone to continental areas. The South American species of Cortaderta (including Lamprothyrsus Pilg.) form one clade, while the abris species form a second clade. Within Cortaderia s. str., section Bifida is polyphyletic, while section NPB thanks the Royal Society. London, for financial assistance in the form of a short-term visiting grant to CMM (which enabled him to conduct research at Reading University, U.K.). Rhodes University Joint Research. Council. and the National Research Foundation for financial cree We thank Mark Chase, Royal Botanic Gardens. Kew, for providing the DNA extract of Cortaderia archboldii. and Henry Connor (University of Canterbury, New Zealand) and Rowan Buxton (Landcare Research, New Zealand) for provision of material of New Zealand species of Cortaderia. We thank M. Buys and A. Barker for their careful and critical reading of this manuscript, and Henry Connor, Surrey Jacobs. and an anonymous reviewer for their thoughtful comments on this work. Contributions to this article by Mark Lyle are based on a doctoral study in the faculty of Biology. University of Hi une ; эгин nl Systematics Facility, Department. of Bolus Rhodes University, Grahamstown, 6140, South Africa. n.barker&?ru.ac.za. De xdv: nt of Botany, University of € 'ape Town. P Bag. Rondebosch, 7700, South Africa. Present address: Institute for Systematic Botany. Zollikerstr. 107. CH-8008 Zurich, Switzerland. plinder(systbot.unizh.ch. ! Department of Botany, University of Reading, Reading, RGO 2AS, UK. Present address: Director of the Herbarium. Biological Sciences, 101 Life Science Building, Auburn University, Auburn, Aishia. 308 19, U.S.A. mortocy@mail. auburn.edu. ` Universität Hamburg. Institut fiir Allgemeine Botanik und Botanischer Garten, Ohnhorststrasse 18. burg. marklyle@t-online.de. 226009, Ham- ANN. MISSOURI Bor. GARD. 90: 1-24. 2003. Annals of the Missouri Botanical Garden Cortaderia is Pe lc and sister to Lamprothyrsus. Cortaderia archboldii shows no immediate affinity with either of these two Key жуга Total Evidence. biogeography, Cortaderia, ITS, molecular systematics, paraphyly, phylogeny, Poaceae, rpoC2 ades, but data indicate a possible relationship with Danthonia DC. . Supertrees, The grass genus Cortaderia (including the or- (Schult.) Asch. & Graeb) has a Gondwanan distribution, be- CN namental “pampas grass," C. selloana ing found in temperate and Andean South America. from Costa Rica to Patagonia (including the Falk- land Islands), New Zealand (including Chatham and Pitt Islands), and New Guinea. Cortaderia is placed in the subfamily Danthon- GPWG, 2001). It was previously included in the polyphyletic subfamily Arundinoideae (Davis 1993; Barker et al., 1995; Clark et al., 1995). and some past classifications place it in its ioideae ( & Soreng, own tribe, the Cortaderieae (Zotov. 1963. emended 1987). on sequence data from the grass-specific insert in 1999) found no support for the recognition of the tribe Cortad- by Conert, Results of an initial study based the plastid rpoC2 gene (Barker et al.. erieae, as Cortaderia is included within a large group of genera that approximately agreed with the Watson Dallwitz These genera have now been elevated. to tribe Danthonieae sensu and (1992). subfamily status as Danthonioideae (GPWG, 2001). This subfamily is defined by the presence of haus- 1994; GPWG, 2001), and all species of Cortaderia examined to torial synergid cells (Verboom et al.. date possess haustorial synergid cells (Philipson. 1977; Philipson & Connor, 1984). TAXONOMY Although no recent species-level taxonomy exists 1974) and Connor (1983a) have reviewed the nomenclature. for the entire genus, Connor and Edgar ( 5 D The New Zealand species have been well studied — axonomically (Connor & Edgar, 1974. 1987: Edgar & Connor, 2000), but the South. American species have only been studied on a regional basis (e.g.. Davidse & Pohl, 1994; 1995, 1996). The number of species in the genus is at present. uncertain: Clayton and Renvoize (1986) and Watson and Dallwitz (1992) considered the ge- \stegiano et al., nus to comprise 24 species, but Lyle (1996) con- sidered there to be only 22 species, and Astegiano et al. (1995) recognized 25 species. However, at least four South American species may not be dis- tinct, and one is considered to be of hybrid origin (C. sericantha X nitida, Lyle, unpublished). Most species of the genus are found in South America (19 species), but five are known from New Zealand and nearby islands, and the anomalous C. archbol- dii is from New Guinea. Cortaderia is presently delimited by the pres- ence of a gvnodioecious breeding system and floret morphology. The genus is divided into four sections Bifida, Vutica, tata) on the basis of sexual dimorphism and lemma morphology (Conert, 1961: Connor & Edgar, 1974: 1986). are monotypic, while section Bifida is the most spe- ME New Zealand and several South can species are placed in section Bifida, while sec- (sects. Cortaderia. and Monoaris- Clayton & Renvoize. The latter two sections close. Ameri- tion Cortaderia is restricted to South America. Morphologically, Cortaderia has been considered to intergrade with the New Zealand genus Chion- ochloa Zotov. archboldii intermediate between the two. genera (Clayton & 1986). Hitchcock (1936) originally described Cortaderia archboldii as a spe- with the New Guinean Cortaderia Renvoize, cies of Danthonia. During the 1960s the broadly circumscribed genus Danthonia was broken up into numerous segregate genera. In Africa, all species formerly in Danthonia were placed in Merxmuellera Conert, Karroochloa Conert & Tiirpe, Dregeochloa Conert, and Pseudopentameris Conert (Conert, 1966, 1970, 1971: Conert & Türpe. 1969). In New Zealand. Votodanthonia Zotoy, Ery- thranthera Zotoy, and = Pyrranthera Zotov were 1963). In. Australia Blake (1972) resurrected Plinthanthesis Steud. and Monachather, Chionochloa, erected (Zotov, and Rytidosperma Steud. was recognized in South America (Nicora, 1973). Rytidosperma and Noto- danthonia have subsequently been synonymized (Connor & Edgar, 1979), although there has been some dispute about the most appropriate name (Veldkamp. 1980: Jacobs, 1982). the validity of rec- genus distinct. from 1982). and the circumscription of Rytidosperma (Clayton & Ren- voize, 1980). (1996) and Linder (1997) have once again revised ognizing Rytidosperma as ^ a Danthonia (Conert. 1975: Jacobs. More recently, Linder апа Verboom the Australasian taxa. Danthonia archboldii has not escaped the fragmentation of Danthonia as it is clearly not Danthonia sensu Zotov. Connor and Ed- gar (1974) transferred the species to Cortaderia, while virtually simultaneously Conert (1975) trans- ferred it to Chionochloa. The matter of where this species should be classified has remained. unre- solved. Volume 90, Number 1 2003 Barker et al. 3 Paraphyly of Cortaderia Table 1. Known chromosome counts for species of Cortaderia and Lamprothyrsus. 2n Species Reference 36 C. pilosa Moore (1967) cited by Connor & Edgar (1974 72 C. araucana 12 C. archboldii 7: C. colombiana (published as C. roraimensis) Ta C. selloana 72 C. speciosa H 90 ulvida 90 C. richardii 90 C. splendens 90 Loetc 108 C. jubata (published as C. atacamensis) 108 C. rudiuscula —136 Lamprothysus peruvianus Beuzenberg, pers. comm., Porgmann (1964) cited by Connor & Edgar (1974) Connor (1965a): . E. Hair & Beuzenberg ( Hair & pensar (1966) cited uh mm & Edgar Connor (19 air & eee (1966) cited by € ae & Edgar Connor (1965a) cited by Connor & Edgar (1974) ) cited by Connor & Edgar (1974) (1967) cited by Connor & Edgar (1974) Connor & Edgar (1974) comm. 1966) cited by Connor & Edgar Connor, pers. (1974) (1974) 715 € nnor & Edgar (19 (1974) Connor & Dawson (1993) Connor & Dawson (1993) CYTOLOGY AND BREEDING SYSTEM The reproductive biology of some species of Cor- taderia has been studied extensively, as they dis- phy a gynodioecious breeding system (Connor, . 1965a. b. 1970. 1973. 1979, 1983b. Кылган, 1984: Connor & Charlesworth, 1989: & Dawson. 1993; Connor et al., 2000). Some species are apomictic (Connor, 1973: Philip- 2000). exist for several species (summarized in Table 1), Connor son. 1978: Connor et al.. Cytological data and have been used by Connor and Dawson (1993) to suggest that at least some species are related to the South American Lamprothyrsus. Connor et al. (2000) also noted that there is incomplete dioecism in Lamprothyrsus. On the basis of these chromo- some counts, the base chromosome number (x) for .g.. Connor & 1996) to be nine Cortaderia is considered by some (e.g Dawson, 1993: = 9). Astegiano et al.. PHYLOGENETIC RELATIONSHIPS Early molecular studies indicated the possible Barker et al. (1999) in- cluded two species in their study based on se- polyphyly of Cortaderia. quence data of the grass-specific insert in the chlo- roplast RNA polymerase subunit C2 gene (rpoC2). The South American species was sister to Lam- prothysus and the New Zealand species was allied to the Australian genera Votochloe Domin and Plin- thanthesis. A subsequent investigation using Inter- nal Transcribed Spacer (ITS) sequence data from a number of species of Cortaderia supported the re- sult obtained from the rpoC2 study (Barker et al.. 2000). A combined analysis was conducted. by Barker et al. (2000), but this analysis had numer- ous taxa with missing data as there was little taxon overlap between the two molecular data sets. De- spite this, these results imply either two indepen- dent origins of gynodioecy within Danthonioideae. or a single origin and multiple subsequent losses in different genera. These conflicting evolutionary scenarios cannot, however, be readily tested, as the breeding systems of the other genera in the lineage have not been investigated in detail. Also, phylog- enies based on plastid genes (such as rpoC2) reflect the phylogeny of the organelle and need not reflect organismal phylogeny (Doyle, 1992: Brower et al., 1996). This could explain the unexpected result of a polyphyletic Cortaderia. This study uses evidence from both chloroplast (rpoC2) and nuclear (ITS) genes from the majority of the species of Cortaderia to test the monophyly of this genus within the context of the seven infor- mal due ола mioid genera outlined by Bark- er et al. 0). Relationships of Cortaderia arch- boldii are kn clarified. MATERIALS AND METHODS TAXON SAMPLING, DNA EXTRACTION, AMPLIFICATION, AND SEQUENCING This study includes place-holders of all the sev- en informal lineages identified by Barker et al. (2000). Table 2 details the taxa sampled in this study, as well as voucher and GenBank information. Leaf material was dried in silica gel (Chase & Hills, 1991). DNA was extracted from these samples us- ing the hot CTAB method of Doyle and Doyle (1987). The entire ITS region was amplified by the Polymerase Chain Reaction (PCR) using the prim- ers 17SE and 265Е published by Sun et al. (1994). 4 Annals of the Missouri Botanical Garden Table 2. Voucher details of species sequenced in this study. The species of Cortaderia are presented at the top of the | l | ) table, sorted according to the section to which they have been traditionally placed. Additional taxa sequenced are listed below those of Cortaderia. Note that taxa for which both ITS and rpoC2 data have ә All those species listed as “ex Cult. Ne w Ze roo were provided by ilready ‚ been published are not included. Buxton from the Botanic Garden at Lincoln, New Ze piena d (G = Garden number). Key: M = DNA extract e provided by M. Chase (Jodrell Laboratory, RBG Kew); NPB = collected by N. P. Barker, n = ЕТ by Н. P. Linder, CHR = Christchurch herbarium. Voucher, origin rpoC2 ITS Genus and species Section & herbarium GenBank No. GenBank No. Cortaderia archboldii (Hitehe.) Bifida J. Marsden 115 (= MWC AF355998 AF367620 connor & Edgar 79) (K) Cortaderia bifida Pilg. Bifida es 1 Ecuador (HBG, AF355988 AF307009 L) Cortaderia colombiana (Pilg.) Bifida Ly h 2 ares 920; Venezuela AF355991 AF307012 Pilg. (HBG, Cortaderia fulvida (Buchan.) Zo- Bifida G 5088; ex 8 New Zealand Barker et al. (1999) AF307015 tov (BOL) Cortaderia hapalotricha (Phil.) Bifida Lyle 1525; Ecuador (HBG, AF355989 AF307010 Conert BOL) Cortaderia nitida (Kunth) Pilg. Bifida Lyle 1434; Ecuador (HBG, AF355990 AF307611 OL) Cortaderia richardii (Endl.) Zotov Bifida G 3816; ex Cult. New Zealand AF355996 AF367618 (BOL) Cortaderia sericantha (Steud.) Bifida Lyle 1128; Ecuador (HBG, AF355980 AF367006 Hitche. BOL) Cortaderia splendens Connor Bifida G 10872; ex Cult. New Zea- AF355994 AF307010 land (BOL) Cortaderia toetoe Zotoy Bifida G 5042; ex Cult. New Zealand AF355997 AF307019 (BOL) Cortaderia turbaria Connor Bifida › 17358: ex Cult. New Zea- AF355995 AF367617 land (BOL) Cortaderia araucana Stapf Cortaderia G 7162; ex Cult. New Zealand AF355993 AF367014 (BOL) Cortaderia jubata (Lem.) Stapf Cortaderia Lyle 1515; Ecuador (HBG, AF355987 AF307008 BOL) Cortaderia rudiuscula Stapf Cortaderia (& 11157; ex Cult. New Zea- AF355992 AF307013 land (BOL) Cortaderia selloana (Schult.) Cortaderia Robinson s. n. (BOL) Barker et al. (1999) AF367607 Asch. et Graeb. "RN let auriculata (J. HPL 5569: Australia (BOL) Not sequenced AF307604 Black) Linder Chaetobromus involucratus NPB 1715; South Africa Not sequenced AF307599 (Schrad.) Nees subsp. involu- (GRA) cratus Chionochloa macra Zotoy CHR 475278; ex Cult. New Barker et al. (1999) AF307595 Zealand (BOL) Chionochloa pallens Zotov CHR 475 279; ex Cult. New Not sequenced AF207590 Zealand (BOL) Chionochloa rigida (Raoul) Zotoy HPL 5710: New Zealand Not sequenced AF307597 (BOL) Lamprothyrsus peruvianus Hitch. G 11154; ex Cult. New Zea- Barker et al. (1999) AF367605 land (BOL) Merxmuellera cincta (Nees) Co- NPB 1160; South Africa (BOL) Barker et al. (1999) AF367593 nert subsp. cincta Merxmuellera cincta (Nees) Co- VPB 1545; South. Africa AF355985 AF367594 nert subsp. sericea N. P. Barker (GR. Merxmuellera davyi (C. E. Hubb.) VPB 912; South Africa (BOL) Barker et al. (1999) AF307590 Conert Merxmuellera decora (Nees) Co- NPB 1168; South Africa AF355984. AF307592 nert (BOL) Volume 90, Number 1 Barker et al. 5 2003 Paraphyly of Cortaderia Table 2. Continued. Voucher, origin rpoC2 IT Genus and species Section & herbarium GenBank No. GenBank No. Merxmuellera disticha (Nees) Co- VPB 1002: South Africa Barker et al. (1999) AF367600 nert (BOL) Merxmuellera lupulina (Thunb.) HPL 7004: South Africa (BOL) AF355983 ronert Merxmuellera rufa (Nees) Conert NPB 1149: South Africa Barker et al. (1999) AF207591 (BOL) Pseudopentameris macrantha HPL 5470; South Africa (BOL) Barker et al. (1999) AF307598 (Schrad.) Conert Votodanthonia gracilis (Kirk) Zo- HPL 5683: New Zealand Not sequenced AF367003 toy (BOL) Tribolium hispidum (Vhunb.) VPB 1740: South Africa Not sequenced VE 367602 Desv. (GRA) Tribolium pusillum (Nees) H. P. HPL 5402: South Africa (BOL) Barker et al. (1999) AF307001 Linder & Davidse The grass-specific insert in the chloroplast rpo C2 on. ABI 373, ABI 377, and ABI3100 automated gene was amplified using the primers “rpoC2-UP” DNA sequencers. and “rpoC2-DOWN” (details in Barker et a 1999). s SEQUENCE ALIGNMENT PCR was carried out in a 100 ul volume using : Р ho Ren : В All component sequences from each PCR prod- PROMEGA Tag and buffer at a concentration range uct were assembled, checked, and corrected where of 1-4 mM magnesium chloride. Thermal cycling "ap" ipee | ву ub dips necessary using Sequencher version 3 (Gene Codes conditions were: an initial denaturing stage of | min. at 97°C, followed by 30 to 40 cycles (template dependent) using an annealing temperature of 52°C 700 Corporation). Completed sequences were then im- ported into the existing data set of danthonioid ITS and rpoC2 sequences, and aligned using DAPSA (DNA And Protein Sequence Alignment: E. H. Harley, Dept. Chemical Pathology. University of (l min.). extension at 72°C (3 min.) and denatur- ation at 97°C (1 min.). A IO-minute extension pe- riod at 72°C followed the 30 cycles, prior to cooling p и i ee ee toin. » I 5 ^ Cape Town). DAPSA was also used to generate data at 4°C. The ITS region of Cortaderia archboldii was RAT que . E < bond das iles for subsequent phylogenetic analysis. amplified as two separate products, ITS] and ITS2, А р ` з ` € P У ^1 ATQ € 1 "n: by me ans of the flanking primers and the internal | PHYLOGENETIC ANALYSIS (sequencing) primers described below. The rpoC2 and ITS data were analyzed both sep- All PCR products were cleaned and concentrat- ) ed into 20 wl of Tris-EDTA buffer or water using arately and in a combined analysis using РАШ LOL › > > version 4.0b3a (Swofford, 2000). Centropodia glau- the QIAGEN QlAQuick PCR Product Purification К: ( ) I pu Kit. Sequencing reactions were carried out using ©@ (Nees) Т. A. Cope and Merxmuellera rangei the ABI PRISM or ABI PRISM BIG DYE cycle — (Pilg.) Conert were used as outgroups in all analy- sequencing kit according to the manufacturers in- 565- These species have previously been shown to structions. The ITS PCR product. was sequenced be related to the danthonioid lineage (Barker, 1995: completely in both. directions, initially using the Barker et al., . 2000). Two hundred random primers 175E. 5.05, 5.8К. and 26SE published by Sun et al. (1994). Owing to primer mismatch. three input order searches were conducted (TKEEP = 1, MULPARS OFF). The most parsimonious trees found using this search were subjected to a full heuristic search (TBR option; MULPARS ON). Fast Bootstrap analyses using 10.000 replicates were additional internal. primers were used to sequence the PCR product: “ITSE? (Hsiao et al.. 1998), “Danth 5.88F" (5'-GAC TCT CGG CAA CGG-3'), and “Danth5.8SR” (5'-TTT GGC GTG ACG CCC- 3'). The rpoC2 PCR product was sequenced using conducted on the individual and combined data sels. the internal primers “rpoC2-1” and “rpoC2-2” Combined molecular data set. Arguments for (Barker et al., 1999). Sequencing was carried out and against when and how to combine data sets Annals of the Missouri Botanical Garden have been discussed at some length (e.g.. de Quei- roz, 1993; de Queiroz et al., 1995; Huelsenbeck et al., 1996; Nixon & Carpenter, and Kellogg (1996) discussed various ways in which conflict between data sets can be assessed, and advised on when it is suitable to conduct such tests. data sets is possible if the trees do not conflict, « if conflict receives low bootstrap support. Thus eac А node on each of the two trees is tested for congru- ence against the other tree. Where the nodes are congruent (that is, do not contain conflicting infor- mation, even if they are not identical), thev are ob- viously combinable. Where they are incongruent, the bootstrap support for each of the conflicted nodes is examined. If the support is less than 70%. the incongruence is interpreted as being due to chance, and that there is no hard conflict. This log- ic is derived from Mason-Gamer and Kellogg (1996), and ultimately from Rodrigo et al. (1993 and was recently applied by Eldeniis and Linder — (2000). Tests for incongruence (such as that pro- posed by Farris et al.. 4. 1995) were thus not conducted, as the trees obtained here show no well- supported conflict. To reduce the amount of missing data in the combined analysis, a combined data set was cre- ated from only those taxa for which both ITS and rpoC2 data were available. Unfortunately, four genera are not represented by a species common to both data sets, and we thus had to create four “fictive taxa" (Kellogg & Linder, 1995) or “hy- (Wiens & Reeder, 1995) that repre- sented the four genera for which different species brid taxa" or subspecies had been sequenced for each gene. These fictive taxa thus represent genera, rather than species. This approach is acceptable when the monophyly of the higher group is known or 1995). taxa represent Danthonia (a combination of D. accepted (Wiens & Reeder, These fictive spicata (L.) Beauv. ex Roem & Schult. [rpoC2] and D. californica Bol. [ITS]), Pentameris Beauv. (a combination of P. thuarii Beauv. [rpoC2] and P. macrocalycina (Seud.) Schweick. [ITS]). Chaetobromus Nees (a combination of C. involu- cratus (Schrad.) Nees subsp. involucratus [ITS] and C. geanus (Nees) Verboom [rpoC2]) and Tribolium (a combination of T. uniolae (L. f.) Renvoize [rpoC2] and T, hispidum (Thunb.) Desv. [ITS]). Pentameris has been recently revised (Barker. involucratus. (Schrad.) Nees subsp. dre- 1993) and is clearly a monophyletic lineage. How- ever, the monophyly of Danthonia has not been ex- amined critically and. for the purposes of this study. it is viewed in a narrow context, comprising mostly 1996). Mason-Gamer Following their suggestion, combination of Northern Hemisphere species. It is possible that the genus is not monophyletic, but as the two spe- cies combined here are both North American, they are considered to be sufficiently closely related to allow data combination. The third fictive taxon rep- resents a single species in a monotypic genus. As suggested by their rank, the two subspecies of Chaetobromus involucratus are obviously closely re- lated, and were once considered as distinct species (see Verboom & Linder, 1998, for a summary). The two subspecies share a unique morphological char- acter of a tuft of hairs below each spikelet, corrob- orating the monophyly of the species. The cytology, ecology, and taxonomy of this genus has been ex- 1990; Verboom & 1998), and these two entities can be readily tensively studied (Spies et al., Linder, combined. Tribolium has been recently revised (Linder & Davidse, 1997) and now includes the previously monotypic Urochlaena Nees (= T. pus- illum (Nees) Linder & Davidse). Although data for this latter species are available for both data sets and thus could be used to represent this genus, a broader sampling of Tribolium was considered ap- propriate, hence the combination of the two species of Tribolium. With the use of fictive taxa, the combined data set comprised 45 taxa (including outgroups). and included representatives of all the seven informal (2000). “Total 1995; Huel- 1996) analysis of the combined data lineages recognized by Barker et a Evidence" (TE: see de Queiroz et al., senbeck et al., set was carried out as described above for the in- dividual data sets. Supertrees. A supertree approach to combining the results of the ITS and rpoC2 analyses was taken to include all species that were sampled for one of these two regions. The supertree was calculated from strict consensus trees of the two molecular data sets, using matrix representation. parsimony (MRP: Baum, 1992; Ragan, 1992). Each consensus tree was described as a matrix, with each node con- stituting a character. Species described by that node were scored as 1, and species not included in the node as 0. Species absent from the tree were scored as unknown (?) for these nodes. In order to reflect the support each node had within its data set, we weighted the characters (nodes) relative to their bootstrap support levels: bootstrap support of 0.7—0.79 was weighted X 3; 0.8—0.89 weighted X 6 and 0.9-1 weighted X 9 (Ronquist, 1996). Since the supertrees search for resolution, and not for nodes supported by all possible trees of a given level of parsimony, we used a semi-strict Consensus tree to indicate the result, as semi-strict consensus Volume 90, Number 1 200 Barker et al. E of Cortaderia Centropodia glauca У РЯ Cortaderia fulvid и Joycea pallida Figure 1. 1 = 0.724. Solid dots indicate nodes that « ce 62 (77 Merxmuellera cincta ssp. Merxmuellera cincta ssp. One of the 610 most parsimonious trees үм ое from the rpoC2 ollapse in th cincta sericea data (length = 5 eps. ).190. consensus tree. Numbers above ihe мея as аге о values greater than 50% from 10.000 Fast Bootstrap re plicate We used both Fitch optimization and optimization pro- trees will retrieve all uncontradicted nodes. hibiting reversals, to prevent shared absences from grouping. as suggested by Bininda-Emonds and Bryant (1998). MORPHOLOGICAL DATA As part of a larger study (Barker et al.. 2000), morphological and anatomical characters were cod- ed from herbarium specimens. This data set was edited by removing uninformative characters and taxa for which sequence data were unavailable. The data were compiled by HPL from observations of both herbarium and fresh collections. In addition. published revisions and anatomical observations were consulted. These characters and their states are listed in Appendix 2, which also gives literature citations to relevant sources of some data. The data Annals of the Missouri Botanical Garden set is provided in Appendix 1. Phylogenetic search- es were done using PAUP (200 random taxon-entry = replicates were conducted, keeping five trees а most from each search, these trees then used in a second search, in which all most parsimonious trees were kept). This search was limited to 10,000 trees. Further signal was sought by successively weight- ing the data by the rescaled consistency index, and repeating the above search protocols. RESULTS All sequences are deposited in GenBank, and the accession numbers are indicated in Table 2 (or the earlier works indicated there). The morpholog- ical data set and alignments are also available from the senior author upon request. THE RPOC2 DATA SET The alignment used in this analysis is identical to that presented by Barker et al. (1999), and the new sequences were placed within the existing alignment without creating any additional inser- tions. Some of the new sequences did have dele- tions, but these do not affect the final alignment length. This alignment contained 50 taxa and re- sulted in 98 phylogenetically informative sites. The maximum parsimony analysis found 9710 аап) parsimonious trees of length 165 steps (C.I. = 0.715, КІ. = 0.896). parsimonious tree is shown in Figure 1, while the A randomly selected most strict consensus tree is shown in Figure 3. THE ITS DATA SET The IT lesser extent. Pseudopentameris) were difficult to S sequences from Chaetobromus (and to a align in two regions: one at the beginning of the ITS] region (a deletion in the Chaetobromus se- quences), the other at the beginning of the ITS2 region (an insertion in the Chaetobromus sequenc- es). Alignment of the remaining sequences was un- but incorporated тапу small problematic, gaps. mostly one to five bases in size. The alignment of 5l sequences produced 168 phylogenetically infor- mative sites. Phylogenetic analysis resulted in 640 most parsimonious trees with a length of 555 steps (С.І. = 0.490, R.I. = 0.724). most parsimonious tree is shown in Figure 2, and А randomly selected the strict consensus tree in Figure 3. COMBINED DATA SET The rpoC2 strict consensus retrieved 22 nodes and the [TS strict consensus 35 nodes. Of these. nine (40.9% of the rpoC2 nodes, 25.76% of the ITS nodes) are identical (Fig. 3). Bootstrap support for these nodes ranges from 63% (Merxmuellera rufa, M. decora, and M. lupulina clade in the rpoC2 tree) to 100% (numerous small clades in both trees). Of these nine common nodes, five comprise small clades of two to three taxa, and (with the exception of the Pentaschistis clade) none of the nodes = common correspond to the lineages recognized by Barker et al. (2000). There is thus some congruence between the two data sets, and, more importantly. there are no conflicting nodes where bootstrap sup- port is greater than 65%. The data sets were thus considered to be combinable. The TE analysis in- cluded 262 phylogenetically informative sites, and analysis of these data resulted in 32 gag par- simonious dg of length 713 steps (C.I. = 0.534, H 0.737). The strict consensus of 8 'se trees is shown in yo 4. MORPHOLOGICAL DATA SET The unweighted analysis located more than 10.000 trees, with a length of 242 steps. consisten- cy index of 0.261, The strict consensus tree of this set was poorly re- and retention index of 0.644. solved. The consistency and retention indices sta- bilized after three rounds of re-weighting. and this weighted data set resulted in 686 trees, with a con- of 0.593 and 0.859. The strict consensus of the most. parsimo- sistency index retention index. of nious trees obtained following re-weighting is shown in Figure 5. Bootstrap support values were not calculated owing to the high degree of homo- lasy, poor resolution, and the very high number of initial most parsimonious trees found in the above procedure. Furthermore, because of these reasons, the morphological data were not combined with the molecular data. In contrast to the molecular trees the consensus tree obtained from the re-weighted analysis shows a single clade for Cortaderia, including C. arch- boldii and Lamprothyrsus.. Cortaderia is sister to Chionochloa, and both are nested within a large group of Merxmuellera species. Although many of the species groups are similar to those obtained the relationships Although no support analyses were conducted, the low reso- from the molecular analyses, among these groups are quite different. lution in the unweighted trees suggests that there is little support for this particular topology in this data set. SUPERTREE ANALYSIS The semi-strict consensus supertree based on the molecular data sets is shown in Figure 6. This has Volume 90, Number 1 Barker et al. 9 Paraphyly of Cortaderia Centropodia glauca Merxmuellera rangei Merxmuellera davyi Merxmuellera macowanii Merxmuellera rufa Merxmuellera decora Merxmuellera setacea Merxmuellera cincta ssp. cincta Merxmuellera cincta ssp. sericea Chionichloa macra лїї = со e Notochloe micr Cortaderia мда Cortaderia toetoe Cortaderia splendens Danthonia californica fe wf A — O O O Bow OQO о о = wd 0 Ф CO O Oo -— s о OQ TE a8 8 S Q p Cortaderia rudiuscula Cortaderia araucana Cortaderia sericantha T | э] 0 o Ф © a [e] o Ф = 3 ® = о З D Q a Chaetobromus Dco ssp. involucratus Merxmuellera stricta Tribolium pusillum Tribolium hispidum Austrodanthonia caespitosa otodanthonia gracilis Joycea pallida Merxmuellera dura ADD» ccc 3 3 30 сас = оо оо o No O O0 m —-0oo235 = = — E = шел 4 كع ع‎ за осо o oS a 3 Pad Ф O حا‎ = = E (D =. Ф =| Ф [e] >) O o ш “uE Q 2 @ 60/77 Pentaschistis aspera Prionanthium ecklonii Figure 2. One of the 9710 most parsimonious trees obtained from the ITS data (length = 165 steps, СА]. = 0.715. R.I. = 0.896). Solid dots indicate nodes that collapse in the consensus tree. Numbers above the nodes are abc ine values greater than 50% from 10,000 Fast E IHE replicates. The “H” in parentheses after Cortaderia selloana indicates the sequence from Hsiao et al. (19€ Annals of the 10 Missouri Botanical Garden 48661) [E 19 CRISP] шолу 3ouonbos au sajeorput юиро}]]әў D112pD1107) 49)je. sosoujuoied ur H. oup pep szsmpospjuad = цох ‘PL nuusdsopily = орпќу ‘apep snumjuadopansq = pnosq ‘IPL рор: 201014) = Oly") *a3e[quiassy D49]]anuix4opy үе = VAG :SMo[[oy se рәүәчер pue soxoq ay] ur pajeorpur әле (000g) [E 19 94eg Aq рәшрпо se sozeour[ 4ofeur oup чәріо пошшоә € ш EXP] ƏY} 1233 0] I9pJo ш S3o[odoi Zgoda ou ut po[p^t«s 40 рәвволә uooq элец тец sououraq gpu! səpə uado seat] yoq ur punoj әле yey} sapou a3jeorpur юр pr[os 7221] CT] 241 ur s22u213jJtp Je: ordopodoi эці 2jepouruo228 OF 924] со эц UL sso12 yey} sem. URIQ әјвәгри SƏN uad() аці) Jes Vep SLI pue (әр 19s вер zgodi sy] шолу вәәлр snsuasuoo рош a] СЄ әл SLI Ццә$д - сой Г ч е орцАч e -c—[ mes ]|—-e—o eueiquiojoo euapeuoo recae euapeyo) [SS пріоацое еџәреџроо eJoyipunoes eiuoujue( ejeoids eiuoujueq uopoJolul 90|u20]ON E E es ехорезеа siseujueujur|d suapuajds euapeyo? upieyou епәреног) eueqin} епәрецог) 30}30} euapeuo^? epiA[nj euapeyo-? pee ЕЕ LUE Oly VINg8 | | | апочоіпо _] Volume 90, Number 1 03 Barker e Paraphyly of Cortaderia Centropodia glauca erxmuellera rangei Merxmuellera davyi Merxmuellera macowanii Malt pij ufa dads vidi us Merxmuellera setace 100 Merxmuellera cincta 8 cincta erxmuellera cincta ssp. sericea PENTAMERIS Pentaschistis aspera Prionanthium ecklonii Plinthanthesis paradoxa odon Cortaderia fulvida Cortaderia Cortaderia toeto Cortaderia a M DANTHONIA | Cortaderia archboldii Lamprothyrsus peruvianus Cortaderia oe Cortaderia colombia жане giae macrantha 99 CHAETOBROMUS Merxmuellera stricta Merxmueller ra Me enira disticha cain. pusillum OLIUM La AMI laevis Joycea pallida Rytidosperma pumila Rytidosperma nudiflorum Figure 4. The obtained from the combined rpoC2 ITS data. Taxa in ' strict consensus tree of 32 most soap trees (len Chionochloa macra th = 713s егсаѕе are fictive taxa. [. = 0.737) nodes are = 0.534. R. T ars above the bootstrap values greater than 50% aa 10,000 F Fast Situ: imd es. substantially more resolution than the strict consensus tree. Weighting the nodes according to their bootstrap support in the source trees added only a single node in the Rytidosperma clade. Treating all characters as irreversible for the unweighted matrix produced. the same result as obtained from the weighted matrix with Fitch parsimony, while the weighted matrix with ir- reversible characters located a further node that links Plinthanthesis and Notochloe. Including C. archboldii leads to the collapse of the three included Danthonia species, C. archboldii, and the Plinthanthesis clade. to form a polytomy with the two Cortaderia clades. the Rytidosperma clade, and the Chaetobromus clade (nodes marked in Fig. 6). 12 Annals of the Missouri Botanical Garden Centropodia glauca Merxmuellera arundinacea Chaetobromus пудам: лы, Joyce Е: Austrodanthonia auriculata S rh] ооз . Z. o 8 = 8 3 © w a a аа apalricha ат lombiana Lampr othyrsus ози Tribolium hispidum Figure 5. The strict consensus tree derived from the successively re-weighted morphological data. Analysis of the 59 593, К.І. A ~~ eA we ie d data retrieve d 686 most parsimonious trees ( DISCUSSION into the alignment. Indels correspond to evolution- "m" ary events, and as such are a source of phylogenetic RPOC2 SEQUENCE DIVERGENCE n : s e information. However, the use of indels in phylo- As noted above, the alignment used here was genetic analyses is still somewhat contentious, and identical to that presented by Barker et al. (1999), numerous coding methods have been proposed and no additional indels had to be incorportated for example, Giribet & Wheeler, 1999; Simmons & see, — Volume 90, Number 1 Barker 13 ا‎ Cortaderia Centropodia glauca Merxmuellera rangei Merxmuellera davvyi Merxmuellera macowanii Merxmuellera setacea Merxmuellera arundinacea upulina < ® x 333 c Ф % ога Merxmuellera cincta ssp. cincta babes ask Sd a a ssp. sericea Chionochloa At oS | === оо -F 3 ооа ани mop P Danthonia spica Danthonia ^in NR ade le эта paradoxa Notochloe rodon Cortaderia Деш ЫЙ саза lv ue Cortaderia turbar ri Salad e сыйын Cortaderia C ortaderia хотин 1А С C C C Cortaderia nitida Cort й i Cortader na Pseudopentameri rantha мис енеси а. Саан ssp. involucratus Chaetobromus involucratus ssp. dregeanus Tribolium hispidu rxmuellera guillarmodae pes pallida ш Notodanthonia gracilis Aus d onia та Pentameris thuar Pentaschistis sanii Figure 6. The semi-strict consensus “supertree” sets, by means of matrix representatii treate are е И | with solid dots. and the position € As de- Ochoterena, 2000, and references therein). i (1990), the grass-specific in- scribed by Тео et al. sert in the rpoC2 gene comprises a series of 21- Many of the indels in this alignment are thus 21 bases in base pair repeat motifs (heptameric repeats). eversible. and with Cortaderia urchboldii excluded. Nodes that collapse when C. . archboldii assumes in the tree is indicated by a broken line. DOT Ф FE Ф 3 a Ф о o о ш O pa: } 3 [e] t -— o =| Ф Prionanthium ecklonii as calculated from strict consensus trees of the two molecular data using parsimony, weighted according to bootstrap support. with the characters archboldii is included length, suggesting that the gain and loss of these repeats is under considerable functional constraint. The creation of indels is thus not random as would be expected a non-coding region of DNA. An initial study on the use of indels as phylogenetic Annals of the Missouri Botanical Garden characters for this region at the family level showed high levels of homoplasy, with reversals (deletions of heptameric repeats following insertions) being frequent (Barker, 1995). Because of the possible non-random nature of indel gain and loss, as well as the findings that multiple losses (deletions) fol- low gains (insertions), indels were not coded for inclusion in these analyses. The rpoC2 sequences for the New Zealand spe- cies are identical, with the exception of C. turbaria, which has a deletion corresponding to two hepta- meric repeat motifs. This deletion is in the same region as a deletion in Lamprothyrsus, emphasizing the homoplasy of indel data, particularly deletions, for this region. The sequence similarity between Lamprothyrsus and C. turbaria is 97.4%. An examination of rpoC2 sequences from species in the South American clade (including Lamproth- yrsus) reveals a similar lack of sequence variation among the species, with similarity ranging from 99.2% to 100%. and C. hapalotricha (Phil.) Conert are retrieved as Cortaderia colombiana (Pilg.) Pilg. a clade (in all analyses) within the South American group, but these sequences are not identical (99.3% Among the South species, C. sericantha (Steud.) Hitehe. is unique in similarity). American having a deletion of 42 bp (two heptameric repeats). but the remaining sequence (i.e., excluding the re- gion with the deletion) is identical to some of the other species, including Lamprothyrsus. The lack of resolution of relationships among the species of Cortaderia by these data is disappoint- ing, and is caused by high sequence similarity of species in each of the clades. This may be as a result of recent speciation and radiation within the clade. or that the grass-specific rpoC2 insert mu- tates more readily by slipped strand mis-pairing (producing insertions and deletions) than by sub- stitutions, and that the actual sequence is con- strained by the as-yet-unknown function of this re- gion. However, other rpoC2 data from species of the African genus Pentaschistis (Nees) Spach (Barker & Gilbert, may undergo both substitutions and insertion—de- unpublished) suggest that this region letion events. Thus the lack of sequence diversity the South New Zealand clades of Cortaderia may be the result of a recent in. both American and origin and diversification. ITS SEQUENCE DIVERGENCE With the exception of C. C. turbaria, sequences of all the species of Cortad- splendens Connor and eria from New Zealand are identical. The former species is 99.5% similar to the remaining taxa, and the latter species s identic al, but with a two-base pair deletion in the ITSI region. In addition, this group of species and Раа and Notochloe share a two-base pair insertion in ITS Ы Of the South. American (аха, C. colombiana and C. hapalotricha share a one-base pair insertion in ITSI. and have a sequence similarity of 98.4%. These two species are consistently resolved as sis- ter taxa in all analyses. Field observations indicate that these two species overlap morphologically and geographically: C. colombiana is functionally gyn- odioecious and occupies lower altitudes, while C. hapalotricha is apomictic, and is found at higher altitudes. It is possible that these two "species" are extremes of a single species that shows altitudinal variation (M. Lyle, unpublished). Sequence similarity between Lamprothyrsus and the South American Cortaderia species ranges from 90.8% (C. selloana sequence from Hsiao et al., 1998) to 98.9% (C. bifida). The least similar se- quences are those of C. selloana (Hsiao et al., 1998) and C. colombiana (94.296), and the most similar sequences are 99.8% similar (C. selloana (this study)- C. jubata (Lem.) Stapf, and C. jubata-C. ar- aucana Stapf comparisons). and implica- ITS tions. alignment—problems The ITS a lesser extent Pseudopentameris, were difficult to sequences of Chaetobromus, and to align, and were unexpectedly divergent from the remaining data set. This high divergence might in- dicate an accelerated mutation rate in this lineage or paralogous copies of the gene. Data for C. in- volucratus subsp. dregeanus show at least two dif- ferent copies of the ITS region in the genome, which may reflect the polyploid nature of this sub- species (Spies et al., Verboom & Linder, 1998). This second subspecies was excluded from the analyses discussed here because of this varia- tion, as the sequences were not usable for a portion of the ITS region. However, in regions where the sequences from these two subspecies are alignable, there is nonetheless a substantial amount of se- quence divergence. This suggests that, despite their present status as subspecies (and even their past status as species, for that matter), the ITS region is presently undergoing substantial re-arrangements in this taxon, possibly as a result of ongoing poly- ploidization events. These events may have been detected here due to the incomplete and/or ongoing process of concerted evolution in these lineages (Arnheim et al., 1980; Zimmer et al.. 1980; Arn- heim, 1983), the results of which will ultimately be a uniform copy of rDNA loci within the taxon. This has been shown to be possible in other polyploid Volume 90, Number 1 2003 Barker et al. Paraphyly of Cortaderia l.. 1995, and The Даай grasses, and plants (see. for example. Sang et i Wendel et al.. 1995). in particular. species of Cortaderia. are widely known to have eris and sometimes high ploidy levels (Table 1). Despite this there was almost no evidence of polymorphic bases in the sequences from Cortaderia or other taxa sampled. This sug- vests that the ITS region in these taxa has been homogenized through concerted evolution. Thus. al- though many of these grasses are polvploids or of polyploid origin, they all appear to be stable poly- ploids (with the exception of the excluded Chato- bromus involucratus subsp. dregeanus). | must. however. be noted that the theoretically exponential nature of PCR amplification would not detect dif- ferent loci unless they were initially present in ap- proximately equal numbers. Different and rarer loci be detectable if quencing approach were used. The methodological would only PCHR-cloning-se- limitations of the approach used here may explain the conflict between the ITS and other (especially Only based topologies to those from a second data set chloroplast) data. comparison of the ITS- derived from a single сору nuclear gene would in- dicate nodes with conflicting histories possibly caused by problems associated with the ITS data and sequencing methodology used here. DATA COMBINADILITY analv- — The results from the Total Evidence (TE sis contain no nodes that are not found in one or both the trees from the separate analyses. However, there are some noteworthy observations: The ITS data do not retrieve a monophyletic Tri- Ети (< the гроС2 data do (100% bootstrap support). The 50% bootstrap support), but combined analysis resolves this genus as mono- phyletic with 90% bootstrap support. The ITS data group Plinthanthesis and Noto- chloe together (81% bootstrap support) but гроС2 data group Plinthanthesis with the New Zealand Cortaderia clade (04% bootstrap sup- port). The combined analysis does group Plin- thanthesis and Notochloe as sister genera, but with 71% bootstrap support. The ITS data find a paraphyletic Cortaderia (< 50% bootstrap support). and the rpoC2 data re- trieve a polyphyletic Cortaderia (< 50% bool- strap. support). The combined analysis agrees with the ITS data. but also at a low (< 50%) bootstrap support. Cortaderia archboldii is sister to the fictive Dan- thonia taxon (69% bootstrap support). in agree- pm ment with the rpoC2 topology. but in conflict with the ITS data. Кенге taxa. vses is that the position of the fictive “Danthonia” One of the surprises of these anal- taxon appears to be uncertain. As a fictive taxon in the combined analysis. no clear relationship for this than its sister entity is retrieved, other Cortaderia archboldii. This is seemingly “midway” between the position it occupies in the individual trees: basal to the Australasian Cortaderia clade i the ITS analysis (less than 50% bootstrap cai: or basal to the Australasian Cortaderia and allies clade in the rpoC2 analysis. This might indicate that this fictive entity is not a monophyletic con- (and not mutually struct. There are three possible exclusive) reasons for this: l. North American Danthonia may not be mono- phyletic. 2. Danthonia may be the result of a past hybrid- ization event, and may comprise a maternal lin- eage (cpDNA) and nuclear lineage with different histories. 3. Danthonia may (possibly of hybrid origin). in which concerted include a polyploid nucleus evolution has resulted in the fixation of a single copy of the ITS region. the history of which is independent of the plastid genome. It seems that even in the restricted sense in which Danthonia is regarded now. there might still be taxonomic problems in this genus. However. this seems unlikely from a morphological perspective: apart from a couple of exceptions the basic chro- mosome number is 36 throughout (ote n & Loe- ve, 1941; Myers, 1947; De Wet. 1954: Bowden & 1962; Schwartz & Baessler. ps t). the leaf anatomy is quite uniform, and there is minor vari- Senn. ation in the lemma indumentum (Tomlinson. 1985). It is thus difficult to explain the results obtained f the Dan- thonia taxa sampled, and additional sampling and here pertaining to the relationships « studies on this genus are warranted. Supertree. of resolution The remarkable degree in the supertree (Fig. 6) shows the general concor- dance between the two data sets. However, this res- olution is only obtained by weighting nodes ac- cording to their bootstrap support, treating characters as irreversible, and calculating a semi- strict rather than a strict Consensus tree. The con- cordance is also borne out by the comparison of the two trees in Figures 1—3. The supertree differs from the TE 1). Firstly, the TE much more resolution than the supertree, and some tree. (Fig. tree locates of this resolution is well supported. For example. Annals of the Missouri Botanical Garden the Merxmuellera setacea—M. cincta clade receives a bootstrap support of 87%, but is not retrieved in the supertree. The supertree is sensitive to any con- tradictory evidence, resulting in substantial loss of resolution when C. archiboldii is included. Of more interest are nodes that conflict between the TE tree and the supertree. For example, the supertree finds (Pentaschistis clade, (Chionochloa, (Cortaderia, Ry- tidosperma))). while the TE tree obtains (Chionoch- loa, (Pentaschistis clade, (Cortaderia, Rytidosper- ma))). Curiously, neither analysis of individual data sets located the node found in the TE tree, which must thus be regarded as the novel node, suggest- ing that the supertree might be more conservative than the TE node combining Rytidosperma pumilum (Kirk) H. P. Linder, R. nudiflorum (P. Morris) Connor & Edgar, f.) Conert & Türpe, No- tree. The other conflict involves the Karroochloa purpurea (L. todanthonia gracilis (Kirk) Zotov, and Austrodan- thonia caespitosa (Gaudich.) H. P. Linder: the prob- lem lies with the inclusion of K. purpurea within this clade. However, this agrees with the result of the ITS analysis, and again the TE analysis located nodes not found in the individual analyses. There are, however, some positions of uncertainty: 1. The presence of a "Cortaderia clade" with Dan- thonia, Plinthanthesis, and Notochloe included. This is retrieved both by ITS and the combined analysis, but always with less than 50% boot- strap support. However, the conflict. of rpoC2 and ITS, and the low level of support, indicate that this grouping cannot as yet be taken seri- ously. Furthermore, the supertree is unable to resolve the relationships of any of the species of Danthonia, Plinthanthesis, and Notochloe. The other fictive taxa survived well: Pentameris remained firmly in the Pentaschistis clade in the supertree; Chaetobromus retains its position as sister to Pseudopentameris; and the TE analysis shows Tribolium to be monophyletic (the super- tree does not). So is it worth using supertrees? They might be useful as quick summaries of different trees, and also to include all the species that may be missing from many partial data sets. They can also be used as consensus trees to indicate where nodes on the TE trees deviate from the nodes common to the source trees. But there are also shortcomings to this approach. This includes the lack of resolution on some nodes, and, more importantly, the lack of any indication of support for nodes. For example. the supertree places Chionochloa in a basal position, just above the Pentaschistis node. There is probably little character support for this (and it conflicts with the TE result), but on the supertree it is just an- other node, the support for which is not known. MORPHOLOGY The lack of resolution in the unweighted data set indicates that the morphological data set lacks de- cisiveness. Yet the main groups extracted аге re- markably similar to the molecular trees, although there are numerous differences in detail. There ap- pears to be no morphological support for a biphy- letic Cortaderia, Parsimony analysis of the morpho- logical data set retrieved a monophyletic Cortaderia clade. which includes Cortaderia, Lamprothyrsus, and Cortaderia archboldii. This clade is based on the gynodioecious breeding system. THE PARAPHYLY OF CORTADERIA All molecular analyses clearly show that Cortad- In the TE taderia is retrieved as one clade, but with less than eria is not monophyletic. analysis, Cor- 50% bootstrap support, and with Danthonia, No- tochloe, Plinthanthesis, and Lamprothyrsus embed- ded in it. This clade is divided into three subcla- Australasian clade, the South American and the C. The Australasian clade includes Plinthan- des: the clade, archboldii-fictive Danthonia clade. thesis and Notochloe (88% аан support in the 74% in ITS rpoC2 analysis, but also SW 's Danthonia). The combined. analysis; analysis; in South American clade includes Lamprothyrsus and receives 68% bootstrap support in the TE (slightly up from 60% in the rpoC2 analysis, 51% in the ITS analysis). The C. archboldii-fictive Danthonia clade receives 69% support in the TE analysis. Re- lationships between these clades and the Rytidos- perma lineage (60% bootstrap support in the TE analysis) and the Pseudopentameris clade (1009€ support) are not well supported. There are thus crit- ical nodes in the topology that are poorly supported, and additional data, both molecular and morpho- = logical, are required before the relationships o these lineages can be resolved with some certainty. However, it is interesting to note that the resolution of Cortaderia into the South American and New Zealand lineages is supported (at least in part) by chemotaxonomic data. Connor and Purdie (1981) reported that Cylindrin, a triterpene methyl ether, was restricted to some species of section Cortaderia (excluding C. selloana), and absent from the South bifida (sect. Bifida) and all the New Zealand species tested. Furthermore, the New Zea- American C. land species also possessed a range of other triter- | | 8 pene methyl ethers that were not found any South American species, but which are known from Volume 90, Number 1 2003 Barker et al. 17 Paraphyly of Cortaderia species of Chionochloa. another New Zealand grass (Connor & Purdie, 1976). Within the South American clade, the species from Cortaderia sect. Cortaderia are more closely related to Lamprothyrsus than to species from Cor- taderia sect. Bifida. This result also supports Con- nor and Dawsons (1993) hypothesis of a close re- ationship of Lamprothyrsus to Cortaderia based on cytological and breeding system evidence, and il would seem appropriate to consider synonymizing Lamproth yrsus with Cortaderia. However, owing to incomplete sampling this step is not taken here. Within field suggest that currently recognized "species section Cortaderia, observations ° might well be one dioecious species (Cortaderia selloana) and one additional apomictic species, with varying populations in the process of speciation and diver- sification through loss of sexuality. Both Lamproth- yrsus and most of the species in section Cortaderta are apomictic weeds, inhabiting drier areas and mostly at lower altitudes of southern South Amer- ica. Only C. jubata is widespread. possibly as а result of anthropogenic factors. The species of Cortaderia sect. Bifida are all found in perhumid high-altitude areas, are seldom weedy, and share a well-defined. biogeographical distribution (paramo). In addition, most species in this section still show some gynodioecious traits. such as large staminodes and the presence of mor- phologically gvnodioecious populations. Unfortunately, without corroboration from molec- ular data from the species in the other two sections in Cortaderia (sects. Monoaristata and Mutica). it would be premature to suggest any further conclu- sions concerning the relationships and composition of the sections of the genus. However, morpholog- ical and anatomical evidence indicates that the species of sections Monoaristata and Mutica are not supported, and that the species in these sections belong in Cortaderia sect. Bifida (M. Lyle. unpub- lished). In addition, we feel it would be premature to formalize the status of the New Zealand species of Cortaderia, as no morphological synapomorphy has vet been found to unite these species. THE RELATIONSHIPS OF CORTADERIA ARCHBOLDIT The New Guinean Cortaderia archboldii is mor- phologically intermediate between Cortaderia and Chionochloa (Clayton & Renvoize, 1986). When the species was originally described. it was includ- ed in the then much more broadly circumscribed (Hitchcock. 1936). With the break-up of the old, broad Danthonia into numer- genus Danthonia ous segregate genera, confusion arose on the correct classification of D. archboldit. Connor and Edgar (1974) placed the species in Cortaderia sect. Bift- da, while Conert a year later put it into Chionoch- loa (Conert, 1975). Conert appears to have. been unaware of the paper by Connor and Edgar, and did not discuss the possibility that. this species could belong in Cortaderia. He cited similarities in spikelet and culm anatomy between D. archboldii and several. species of Chionochloa. Tomlinson (1985) used the same arguments to place the spe- cies in Cortaderia. Cortaderia archboldii did nol group with the three Chionochloa species in either molecular anal- ysis. The more extensive morphological data set places C. archboldii in Cortaderia, and not in Chionochloa. Unpublished analyses of Linder show that Chionochloa is monophyletic, defined by the unique morphology of the adaxial microhairs with minute apical cells; which are overlapping, restrict- ed to the base of the furrows of the leaf blades. In addition. five other characters, which are also found occasionally in other genera or are variable within Chionochloa, can be used. They are disarticulating leaf blades. presence of leaf-margin. prickles. elumes shorter than the basal lemmas, the presence of long dense hairs on the palea margins. and the presence of adaxial prickle hairs on the leaf blades. Cortaderia archboldii lacks most of these charac- ters. and specifically the adaxial microhairs are “typical.” in that they are scattered along the mar- eins of the leaf furrows, are not overlapping, and have relatively large apical cells. In addition, the palea margins are usually glabrous, and leaf prick- les are absent. The only derived character shared with Chionochloa is the presence of disarticulating leaf blades, a character that also occurs in several species of Rytidosperma, Cortaderia bifida, C. ju- bata, and Lamprothyrsus. In contrast, as pointed out by Tomlinson (1985). there are several characters in common between Cortaderia archboldii and al least some other species of Cortaderia. The adaxial leaf furrows are much wider than in Chionochloa. and have large bulliform cells (a feature in common with Cortaderia hapalotricha and C. pungens Swal- len), especially toward the relatively well-devel- oped midrib. Furthermore. C. archboldii also shares a stoloniferous habit with C. bifida, a feature that is not found in other species of Cortaderia. The gynodioecious breeding system of C. archboldii was studied by Connor (1970). who showed that this relatively rare system was found largely in Cortad- eria (Connor, 1979, 1981). Conert (1975) suggested that similar systems are also found in Chionochloa. but a detailed survey (Connor, 1991) documented it only from Chionochloa bromoides (Hook. f.) Zotov. Annals of the Missouri Botanical Garden The precise position of C. archboldii in the broad Cortaderia clade thus remains unclear. rpoC2 data suggest an association with Danthonia and the Aus- tralasian Cortaderia clade (including Plinthanthesis and Notochloe), while ITS data suggest that it is associated with yet other species of Danthonia and the South American Cortaderia clade. The TE anal- ysis links C. archboldii with the fictive taxon Dan- thonia; this could be the result of the rather limited sampling of Danthonia species in this study. The result of this conflict is that the supertree loses res- olution between the two clades of Cortaderia and the species of Danthonia when C. archboldii is in- cluded. Although bootstrap support for the position with the Australian species of Cortaderia is stron- ger in the rpoC2 tree than the support for a position with the South American species as suggested by the ITS tree, node support in different parts of the tree indicates that it would be simplistic to suggest that the rpoC2 results are more trustworthy. Cur- rently there is therefore strong evidence that C. archboldii belongs in the broad Cortaderia, rather than Chionochloa, clade, but no further progress can be made in determining where in this clade it is best placed. It is tempting to argue for the recognition of Cor- taderia archboldii as a monotypic genus, related to the Australasian species of Cortaderia and to Dan- thonia. However, the morphological evidence is not strong enough, nor is the molecular evidence cur- rently convincing. Even if Cortaderia archboldii finds a position sister to the Australasian segregates of Cortaderia, then there would still not be ade- quate grounds for erecting a separate genus for this geographically isolated species. MORPHOLOGY: CHARACTERS AND DATA SET CONFLICT The division of Cortaderia into two lineages. with several species of softly herbaceous species of Dan- thonia associated with the one lineage. and the small, heathland genera Notochloe and Plinthan- thesis associated with the other, means that several “typical” characters of Cortaderia must have either evolved twice, or been lost twice. These characters include the massive inflorescences, spikelets with linear-lanceolate, 3-veined lemmas with abundant indumentum at the base, the reductions in the lem- ma setae, parallel loss of bulliform cells and the formation of tough, xerophyllous leaf anatomy, and the rare gynodioecious sexual system. These char- acters could be clustered into two functional types. The first involves the formation of tough, long-lived plants, with tough leaves. This has happened in parallel in all three southern continents, and has resulted in remarkably convergent leaf anatomies. While it is generally possible to distinguish the leaves of southern African Merxmuellera from New Zealand Chionochloa, and these from Cortaderia, the remaining similarities are still quite convincing. The second character group involves the evolution of the gynodioecious breeding system. Possibly the arge inflorescences and long indumentum at the base of the lemmas is associated with this. Some features are found largely or only in one of the Cortaderia segregates. In many South Amer- ican species of Cortaderia there is a variously de- veloped layer of collenchymatous cells below the abaxial epidermis. In some species this can take up almost half of the leaf width, while in other spe- cles it may be much narrower. It is sometimes linked to the outer bundle sheath via well-devel- oped extension cells. However, this is also found in one species, C. fulvida (Buchan.) Zotov, from New Zealand. The South American species also have characteristically villous inflorescence branches, a character that is remarkably homoplasious in the grasses. A divided phloem is typical of the C. sel- loana clade, but is absent from most other species of South American Cortaderia. Abaxial ridging of the leaves has a similar distribution. Most of the New Zealand Cortaderia species pos- sess curious islands of clear cells in the chloren- chyma, a feature also found in Plinthanthesis and Votochloe. However, this character is missing in C. splendens, as well as C. archboldii, but this could be interpreted as a secondary reversal. In addition, all New Zealand species of Cortaderia have a well- developed and prominent leaf midrib, a feature rare in the danthonioids, but also found in the South American C. rudiuscula Stapf. Over and above these features, it is worth revis- iting the cytological data in light of the above re- sults. The New Zealand species all have 2n = 90 9). while the South American species have variable chromosome num- (i.e.. are decaploid, if x bers, ranging from 2n = 36 (tetraploid), through 2n = 72 (octoploid), to 2n = 108 (duodecaploid). As- suming the base chromosome number is x = 9, then the South American species can be viewed as having 2n = 4x = 36, 2n = 8x = 72, and 2n = 12x = 108. The chromosome count of the New Zea- and species of 2n = 90 may be a decaploid inter- 108, suggesting that these species could have arisen mediate between that of 2n = 72 and 2n = from an allopolyploid ancestor, which was a hybrid 72 and 2n = 108 parents. While resorting to an ancient hybridization event as an between 2n = explanation may be viewed as convenient and pos- Volume 90, Number 1 2003 Barker et al. Paraphyly of Cortaderia sibly undemonstrable, such a history could explain the conflict found here between the nuclear ITS and plastid rpoC2 data sets, with morphology reflecting a mixed or mosaic history, further confusing the phylogenetic signal in the analyses of multiple data sels. SUMMARY The analyses of the molecular data sets indicate that Cortaderia comprises two lineages: one Aus- tralasian, the other South. American. The relation- ships of the New Guinea species C. archboldii are not satisfactorily resolved, other than to allow us to state that it is not a member of either Cortaderia lineage. These relationships are not supported in analyses of morphological data, but the available morphological data are still poor, and it is therefore possible that there are undiscovered morphological markers for these molecular clades. Morphology is thus an area that needs further research. ample. aspects such as caryopsis morphology need morphological studies For ex- investigation. In addition, are complicated by the presence of both bisexual and female-only spikelets, creating sampling prob- lems. Having said this, we realize that the molec- ular sampling is also incomplete, since most of the Cortaderia species sampled here are Andean (аха, and the lowland species have not been included. Only once more taxa and samples have been in- cluded in the molecular data sets, and an exhaus- tive field-based taxonomic study of the species (and their breeding systems) in both lineages has been conducted. can any formal nomenclatural changes be made. Literature Cited MR N. bee Concerted e volution of р fam- ilies. Pp. 38-61 in M. Nei & R. »ehn (editors), Evolution a bag and Proteins. ane Sunderland, Massachusetts. ——, M. Krys tal, В. tas 'hmickel, G. Wilson, O. Ryder & E. A. заи 1980. Molec me унай е for genetic exchanges among о. genes on non- — 1 chromosomes in man and apes. Proc. Natl. Acad. Sci. U.S.A. 7 rg Astegiano. | E., А. M. Anton & H. E. Connor. 1995. Sinopsis del género Cortaderia (Poaceae) en Argentina. Түш 33: 43—5 А п. Flora = puc Argentina, Fasci iculo 22. Poace par Tribu ah ROA ORA s со ЕТ). 1 A biosystematic study of а ris Beauv. (Arundinoideae, Poaceae). Bothalia 23: 25—47. —— 1994. External fruit morphology of i tue Af- rican ` Атиййшеде (Arundinoideae: Poaceae). Bothalia 24: 55-66. 1995. A Molecular Phylogeny of the Subfamily Arundinoideae (Poaceae). Ph.D. Thesis, University of б own. . P. Linder & E. H. Harley. 1995. deo did of Pie Tos sie (Poaceae): quence data. Syst. Bot. 20: 423 ————, ————& —— —. 1999. Sequences of t Evidence from rbcL s 35. grass-specific insert in the me rpoC2 gene e ir done generic relationships of the Arundinoideae (Po- aceae). Syst. B pa 23: 327-350. —— ——— & C. Morton. 2000. نب —— The Danthonieae: Generic Mense and relationships. Pp. 221-231 in | L. Jacobs & J. Everett (editors), Grasses: System- atics ка pou CSIRO Publishing. koe in Baum, B. R. . Combining trees as a way of combin- ing data sels ari phylogenetic inference, and the desir- 0. ability of combining gene trees. Taxon 41: 3— Bininda- Emonds, О. a l. Bryant. 1998. Prop- erties of pi x representation with parsimony analyses. Syst. B iol. 97—508. d они and Danthonia and a es of Fairs 'hloa (Grami- neae). Contr. Sg ни Her : 1-19 Borgmann, E. 1964. Anteil 2 ү ыыы n in der Flora des Bismare енеке von Ostneuguinea. Zeit. Е. Bot. 52: 118-17 Bowden, W. E & H. A. Senn. 1962. bers in 28 grass p from South America. Сапа‹ Bot. 40: 1115-1124. Brower, A. V. Z., “4 De Salle & A. Vogler. 1996. trees, species trees, and systematics: A cladistic per- spective. Annual Pu Ecol. Syst. 27: 423—450. Chase, M. W. & H. H. Hills. 1991. Silica gel: A material for field. reservation of leaf samples for DNA e ж-з 40: 215—2 Clark, L. W. Zhang & J. T Wendel. 1995. А phylog- eny of a grass ird yes eae) based on ndhF se- 60 Chromosome num- ad. J. Gene n ideal < quence data. Syst. Variation in de jamie of cleistogamy the grass Danthonia. Clay, K. 1983. within yd re specie Ап mer. J. Е b. 835-843. envoize. 1986. Genera Gramin- 7 Londo n. Conert, H. J. 1961 hs е und Anatomie der мар J. Cramer, Wienhe 19 V Dregeochloa, eine neue Gattung der Gra- mineen i т ndinoideae, Danthonieae). Senc жа Biol. 47: 335-343. —— 1970. Merxmuellera, eine neue Gattung der Gra- mineen s oaceae: Arundinoideae). Senckenberg. Biol. 51: 129-133. — ———, 1971. The genus €— in Africa. Mitt. Bot. баланны. München 10: 299—308. 1975. Die d Arten von Australien und Neuguinea. Senc siis I 56: 153-164. 1 С px con n the systematics of the Arundinoideae. Pp. Prep in T. R. Soderstrom, K. Hilu, С. S. Campbell & M. E. Barkworth (editors), Grass den matics and Ei Smithsonian Institu- tion Press, Washington, D.C —— ——— & А. M. Türpe. 1969. Karroochloa, Gattung der Gramineen (Poaceae, oo Dan- thonieae). Sie kenberg Biol. 50: 289— 318 Connor, Н. E. 1963. Breedi ing systems in New Zealand grasses IV. [ме нү in Cortaderia. New Zealand J. Bot. |: 258-264 бза. Вак systems in New Zealand grass- eine neue ~ Annals of the Missouri Botanical Garden es V. — species of Cortaderia. New Zealand J. Bot. 17-23. TR Breeding systems in New Zealand grass- es VL Control E B СК ‘ism in Cortaderia ric hard.. New oC J. 233-242. 970. metn in Danthonia archboldii. Bot. 18: 233—236. 97]. Cortaderia re Connor s tenis ае): New Zealand 519-525. ————. 1973. Breeding sy t in Peres (Grami- rene: p 27: 663— Austral. 1 nov. —. - ———— 79. Breeding systems in grasses: New Acai J Bot. 17: 547-57: —— ———. 1981. Evolution of reproductive mn in the m a Ann. Missouri Bot. Gard. 68: 48—74. 1983a. Names and Pp in QUEM. Stapf (Gramineae) II. Taxon 32: 633-634. — ———. 1983b. Cortaderia: Interspec fic hybrids and the breeding system. Heredity 51: 395—403. — ———. 1987. Reproductive biology in the grasses. Pp. 117-132 in T. R. Soderstrom, К. W. Hilu, C. S. C: bell & M. E. Barkworth (editors), Grass Systematics and А survey. Ju c Evolution. Smithsonian Institution. Press, Washington, ).C. ————, : ы Zolov (Gramineae) in New ао New Zealand J. Bot. 29: 219-283. Charlesworth. 1989. Genetics of male-ste- rility in gy sane a scious Cortaderia (Gramineae). Hered- ity 63: 373-382. & M. I. Dawson. 1993. Evolution of Vo ab in о (Arundineae: Ann. Mis sour ri Bot. ard. 80: 512—517 Г. Крл. 1974. eria Stapf (Gramineae). ' & ———. 197‹ Gramineae). Names and types in Cortad- Taxon 23: 595-605. — Rytidosperma Steudel (Noto- iri ia Zoton) i in New Zealand. New Zealand J. Bot. on gı Name changes in the indige- 1960—1986 nomina nova iV. 1983 —1986. New Zealand J. Bot. 25: 115-170. W. Purdie. 1976. Tr B methyl ether E in Chionochloa (Gramineae). New Zea- land J. . 14: 315-326. ——— —. 1987. Nee Ze un Нога, & 1981 Triterpene methyl ethers in some South American ит les of Cortaderia (Grami- zu ew y rut 19: 171-172. . Anton CR E Pu e 2000. Dioecism in wees in жеөө ps 287-293 in S. W. L. Ja & J. Everett (editors), Gras lution. CSIRO Publishing: Collings Davidse, G. & R. W. P нЕ 199 odds Stapf. 251—252 in G. Davidse, S. Sousa x A. O. Chat itors), Flora Чыл ‘ana, Vol. 6. Alismataceae a Cy- peraceae. Universidad Nacional i d de México, D.F.; Missouri dis i Garden, St. Louis; The Natural History us шп, Davis, J. 1. & R. "1993. Phylogenetic structure in the grass tuti (Poaceae) as inferred from chloro- restriction site Amer. J. Bot. 80 acobs : Systematics and Evo- ^p. iter (ed- Londor . Soreng. variation. 1993. 2: 368-37: Donogh & J. Kim. 1995. Separate versus combined analysis of phylogenetic evidence. Annual ev. Ecol. 26: 657—681. De Wet, J. M H 1954. The genus DAE in grass phylogeny. Amer. J. Bot. 41: 204—211. For consensus (sometimes). Syst. Dobrenz, A. К. & A. A. Beetle. Danthonia. J. Range Managem. le, J. J. 1992. Gene 1966. oo in 19: 292—296 trees and species trees: Molec- Syst. Bot. МЈ m systematics as a 1-сһагасіег taxonomy. 17 кеу „ Doyle. 1987. A rapid DNA isolation pro- cedure a E M of fresh leaf tissue. Phyto- che m. Bull. —15 Edgar, E. & "n TA 2000. Flora of New Zealand, Vol. V. a Manaaki Whenua Press, Lincoln, New Zealand. Eldenás, P. & H. P. Linder. 2000. Congruence and com- plementarity of morphological and trnL-F sequence. m the phylogeny of the African Restionaceae. Syst. . 25: 701. E llis. P Р. Hen Leaf anatomy of the South African Dan- thonieae (Poaceae). V. Merxmuellera macowanii, M. davyi and i Mae cna e 493-500. Farris, J. S., Kaller . G. Kluge & C. Bult. 1994. Tesni с апсе к incongruence. Cladistics 10: 315-319 . Constructing l. 44: a dione test for RENI A ion Bio 70-57: m ee 1999. On gaps. Molec. 143. W. b Phy үч GPWG (The Grass loge ny Working ey 2001. Phylogeny and шна classification of the grasses (Poaceae). Ann. Missouri Bot. Gard. 88: 373—457. Hair, J. B. & К.) чо 1966. Contributions toa chromosome = of the New Zealand flora—7. New Zealand J. Bo d e 266 Hitchcock, A. S. Ст al results of the Archbold expedition. | os yuan grasses collected by L. J Brass. oo _ AES 30. "n C., S. W. L. Jacobs, N. P. Barker & N. J. Chatterton, )98. n molecular рро of the subfamily Arun- iden (Poaceae) based on sequences of rDNA. Aus- Syst. Bot. 11: 41—52. Е J. P., J. J. Bull & C. W. Cunningham. 1996, Combining = in n phylogenetic analysis. Trends Ecol. Evol. 11: 152-15 lgloi, G. L., A. jo A I. Dory & H. Kossel. 1990. Nu- cleotide sequence of ike maize c Faces rpo B/C1/C2 operon: Comparison between the derived protein pri- mary structures from «жш organisms with re “pe lo functional domains. Molec. Gen. Genet. 221: Aig 94. W. 982. Jacobs, S. L. 1 omment on Proposal 52 ote servation of Noda toni Zotov багы. l'axon 31: 737-74: Kellogg, E. A. & H. P. Linder. 1995. Pd of Poales. {дб —542 in P. J. Rudall, P. J. Cribb, D. F. Cutler 2. J. Humphries (editors), Monoc mede System- wie and Evolution. Roval Botanic Gardens, Kew Linder, H. P. 1997. Nomenclatural corrections in the Ry- tidosperma complex (Danthonieae, Poaceae). Telopea 7: 269-274. i4 . Davidse. 1997. ium UNA e ae: 119: 445—507 & R. Р. Ellis. 1990. Vegetative morphology and terfire survival strategies in the Cape Fynbos grass 103. edes 20: ‹ ка & The systematics of Tribol- Poaceae). Bot. Jahrb. Syst. erboom. 1996. Generic limits in the d кы. Poaceae) complex. Telopea —028. 6: : ——— T F. Thompson, R. P. Ellis & S. M. Perold. 1990. Volume 90, Number 1 20 Barker 21 EO Р Cortaderia The occurrence, anatomy, and systematic implications of the glands in Pentaschistis and Prionanthium (Po- aceae, Arundinoideae, Arundineae). Bot. Gaz. 151: я 1967. IOPB Plant chromosome number reports XI. CU 16: vage: Lyle. 1996. Change in name and status of a pampa grass imn Poac eae: Arundinoideae) from Boliv- ia. Novon 6: 72-7 Mason-Gamer, . A. Kellogg. 1996. Testing for phylogenetic ми among d шаг ici sets in the tribe Triticeae (Gramineae). Syst. Biol. 45: 524—54: Moore, D. M. 1967. анан н E m Islands angiosperms. Bull. Brit. 9— 82. Antarc. Surv. Myers, 1947. grasses. = Bo. Rev. 13: : Nic ora, E. 1973. و‎ agrostológicas patagóni- о 18: 80-106 коп, K. C. & J. M. Puro 1996. On simultaneous analysis. Cladistics 12: 221-241. Philipson, M. N. 1977. Haustorial ik ele in Cortad- eria (Gramineae). New Zealand J. Bot. 15: 777-778. 1978. б" in Cortaderia jubata (Grami- — hos Zealand J. B —59. . E. Connor. re Haustorial synergids in ыз vise Bot. Gaz. 145: agan, M. A. 1992. Phylogenetic Н а оп E trix il of trees. Molec. Phylogenet. Evol. 53-58. and genetics of forage —422. кено E. R. 1984. Naturalised species of Yn Ss eae) in southern Africa. S 3: 343- 6. S. African J. Bot. go, A. G., M. Kelly-Borges, P. R. Bergquist & P. L. . 1993. A dd test of the null hy- parametric phylogenetic tree. New Zealand J. Bot. 31: 257-268. Ronquist, F. 1996. Matrix р = trees, redun- dancy, and weighting. Syst. Biol. 45: —253. Sang, T., D. J. Crawford & T. F. тач 1995. Docu- mentation of reticulate evolution in peonies (Paeonia) using Internal Transcribed Spacer sequences of nuclear ribosomal DNA: Implications ym biogeography and con- certed = Proc. Natl. Acad. Sci. U.S.A. 92: 6813-6 du ^k & M. Baessler. 1964. Danthonia (Sieglin- gia) decumbens, ein bemerkungswerter Fall о maler and oekologischer Divergenz. Oesterr. Bot. Z. 11: 193-207 Simmons, M. P. & H. Ochoterena. 2000. Gaps as char- acters in sequence-based phylogenetic analyses. Syst. Biol. 49: p "ue Spies, J. J.. u Plessis, N. P. Barker & S. M. C. van Wyk. T^ Deve. pour in the genus Chaeto- bromus (Poaceae: Arundineae). Genome 33: 646—658. Stebbins, G. L. & R. M patie 1941. A cytological study of California ۰ T E J. Bot. E) 371-383 Sun, Y., D. Z. Skinner, G. H. Liang & S Hulbert. 1994. Ru Ausus a Surki ee pee taxa using Tube transc pem hear ers of nuclear ri- bosomal DNA. ' r. Appl. G : 26—32. Swofford, D. L. 2000. "PAUP: Phylogenetic Analysis Us- ing Parsimony. Vers. 4 beta 3. Sinauer, Sunderland, Massachusetts. Tomlinson, K. L. 5. Comparative anatomical studies in Danthonia sensu S (Danthonieae: Poaceae). Aliso 11: 97-114. ise J. F. 1980. Conservation of Notodanthonia Zo- yw (Gramineae). Taxon 29: 29 ж йш . G. & H. P. Linder. A re-evaluation of species Toii in Сыноо е нөм Роасеае Nordic J. Bot. 18: 57—77. N. P. Barker. 1994. Haustorial syn- : An important character in the systematics of ndi grasses e EY Poaceae). Amer ). . J. Dallwitz. 1992. The Grass Genera of i idge Wendel, : 1995. Bidi- sectional inicios us concerted evolution following allo- ploid speciations in cotton. Proc. Natl. Acad. Sci. U.S.A. <“ 284. Wiens, J. J. & T. W. Reeder. 1995. Combining data sets with pun Nune. of taxa for phylogenetic analysis. Syst. Biol. 4 اي‎ Zimmer, ^" . M. Beverley, Y. W. Kan & A. un. 1980. pu duplic ation and loss of genes c cline for the a chains АДА Proc. Natl. Acad. S S.A. 77: 2158- Zotov, V. D А 63. Synopsis of ae grass subfamily Arun- dinoideae M New Zealand. New Zealand J. Bot. 1: 78— 136 22 Annals of the Missouri Botanical Garden Appendix 1. The data matrix used in the morphological analysis. І 1 1 1 1 1 1 1 1 з 4 5 6 7 8 9 0 1 2 3 4 6 в 9 Centropodia glauca 0 1 0 0 0 0 0 0 1 0 0 0 0 1 2 1 2 = Austrodanthonia auriculata 0 ] ) 0 0 0 0 0 1 1 1 1 0 1 2 4 0 0 Austrodanthonia caespite 0 1 0 0 0 0 0 0 1 ) 1 1 0 1 2 4 0 0 анон lar apis 0 ] 0 0 C 0 0 0 1 1 1 0 1 2 4 0 0 0 1 0 0 0 0 ) 1 0 ) ] 2 Е 0 a "ni аста 0 0 0 0 0 0 1 ) í 0 0 1 1 2 1 ) 0 Chionochloa aa 0 1 ? ? 0 0 0 0 1 0 0 1 1 2 0 0 Chionochloa rigida 0 1 ? ? 0 0 0 0 1 1 0 ) 1 ? 2 1 ) 0 Jortaderia araucana 0 1 0 0 0 ) 1 0 0 ) 1 ) 0 0 ) ? 2 = Cortaderia bifida 0 1 2 2 1 0 1 0 1 0 1 0 0 1 0 2 3 0 0 Cortaderia colombiana 0 0 2 1 0 0 1 0 1 0 1 0 0 1 1 1 0 0 Cortaderia fulvida 0 1 0 0 0 ( 1 0 ( 0 1 0 ) 0 0 0 j 2) 0 Cortaderia ا‎ ha 0 1 2 2 ( 1 0 1 0 0 1 0 2 3 0 0 e jubata 0 1 0 0 0 | 1 0 ) 1 0 à 0 Cortaderia nitida 0 1 0 0 0 {01} 1 0 1 ) 01 0 1 1 0 Cortaderia richardiana 0 1 0 0 0 0 1 0 0 ) 0 0 0 1 7 aleta rudiuscula 0 1 (01) 0 0 ) 1 0 0 ) 1 0 0 ) 0 0 2 - Cortaderia selloana 0 1 0 0 ( ) 1 0 0 0 1 ) 0 0 0 0 3 2 vortaderia icantha 0 1 0 0 1 1 0 1 1 0 0 1 1 1 j 1 0 Cortaderia splendens 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 Cortaderia toetoe 0 1 ? 0 1 0 0 0 0 0 0 0 0 0 1 0 Cortaderia archboldii 0 1 ) 0 0 1 0 1 0 1 0 0 1 ) 2 ) 0 Danthonia californice 0 1 ( 0 1 1 0 0 ) 1 2 Б Danthonia secundiflora ( 1 0 ) 1 1 0 1 2 2 5 Danthonia spicata ) 1 0 ) 1 1 0 $ 0 0 1 2 2 ) oycea pallida 0 1 ( 0 ) ) ) 1 0 1 2 2 5 ) Karroochloa purpurea 0 1 0 0 1 l 0 1 0 1 0 1 2 2 2 0 Lamprothyrsus peruvianus 0 1 1 0 0 1 0 1 ) ? 0 0 1 1 1 5 0 0 шеЙега winacea 0 0 0 0 0 0 0 1 0 0 0 0 1 2 5 1 0 Merxmuellera cincta subsp. cincta 0 1 0 0 ( 0 0 0 1 ( 0 0 1 2 2 | 0 erxmuellera cincta subsp. sericea 0 1 0 9 0 ) ( 0 1 0 ) 1 2 2 2 0 Merxmuellera davyi 0 1 0 0 0 ) ) 0 ] ) 1 2 0 Merxmuellera decora 1 1 0 0 0 ) ] | 1 2 1 Merxmuellera disticha ) 1 0 0 ¢ 0 0 1 0 1 0 1 2 2 Merxmuellera dura 0 1 0 0 0 0 1 1 ) ) 1 2 6 ) Merxmuellera UNE 0 1 0 0 í ) ( 0 L 0 ? 0 0 2 2 2 1 ) Merxmuellera lupuli 1 1 0 0 0 0 0 0 1 | | 0 0 1 2 2 6 2 0 erxmuelle Sen 0 1 0 0 0 ) 0 0 1 0 ? 0 1 2 2 0 0 Merxmuellera rangei ) L 0 0 3 0 0 0 1 1 ) ) ) 1 2 2 : у 0 Merxmuellera rufa 1 1 0 0 0 ( 0 { 1 1 0 ( 1 2 0 iuellera setacea С 0 ) $ 1 ? 1 2 2 2 Merxmuellera stricta 1 0 0 0 1 2 2 tochloe microdon 0 1 0 1 ) 0 1 1 ) Notodanthonia gracilis 0 1 0 0 0 ) 0 0 ( 1 2 2 : ) Pent ; [үс 0 1 0 0 0 ) 0 1 ) ) 1 1 2 5 ) 1 Pentameris thuarii 0 1 ) 0 0 0 ( 0 1 0 0 0 1 1 1 2 5 0 1 rige e) asperà 0 1 ) 0 1) 0 0 0 1 0 1 ) 1 ? 1 5 0 1 Р 1 0 1 ( ) ( 0 1 0 0 0 1 1 1 0 1 Plinthanthesis prid 0 1 0 0 0 0 1 0 ) 1 1 0 Prionanthium ec ‘klon 0 1 0 0 1 0 0 1 ? 1 2 - / ) 1 ) 0 0 1 0 0 1 1 {12 2 5 ) 1 Rytidosperma nudiflorum 0 1 0 J ) 1 1 0 1 2 2 2 ) 0 Rytidosj pumil 0 1 0 0 0 ( 0 1 (€ 0 о 1 1 1 5 2 Schismus barbatus 0 1 0 0 ( 0 0 0 1 ] Í 0 0 1 2 0 ) - Tribolium pusillum 0 1 0 0 0 0 0 0 1 1 1 0 0 1 ? 0 5 Tribolium hispidum 0 1 0 0 0 0 C 0 ] ] 1) 0 ? 0 23 Barker et al. Volume 90, Number 1 2003 Paraphyly of Cortaderia Extended. Appendix 1. oo о ++ =ч о о о ++ О da Qua NO a o cC € d і ©” о о e Le et © e Oud 4 D с. 0 с e ооооосо c a. a C C e oo a з=. di oO ооо ч adoude a pt = ooooooo ooooooo oocoaoco بے ہے بے ہے ہے‎ e © чо чочо 6 ч чооосо ~ ч чоо © ч ( " DOA ч [s © = c4 ec c © 1 ооооосо ~m c4 Ort 4 сч CO ИЩ = ec ^ O ca tO OO c [; e ہے ب ب‎ n n C n оооооо о O © ¢ C C =, CO TITTEN. m 4 e. 4 а 4 c1 N 4 INA 4a 4 ‹ € [^ е e. осоо о 6 © © 4 о о o ч о 4 ч сч о обоо о єз e oco чоо ч ч оо чо о і adeo HAH с о о 1 «i1 One vd c4 n nd + = 4 efc چچ‎ aA езе ex c чс ос oo. ЄЎ O's ч ч о о о чі cd О еч rn ooo 6 e €» €» c € co C 0 Re 0 0 e. 4 A Pad ) € e € - ее oo od od e eo oN oo oo e o e o0 e о oO Е ~ 4م‎ oN es Od Od ÖGG eo - =ч da r N [m Ж ж © сос oo о о oc о о A © о о о о о о єч © ч о С» Єў ч о 4 о 1 N dO oc ч о ad oe о о чс. 4 с 4 + der e C. 0 © с. Oe 0 С e Go 3 оббо 6© © Ooooornroooco ooo сэ ce ос © оо о о соо о о чочо о б с 4 NAN ообобоомоо то ©» сз бз Єз cot с.с O NONON~ OOOO оо о 3 оі О о dà 5-8 = сі OO О г Ona © oH OO Of On OO о о сх ч no nd O 400 4 = od 2 ro i > 4 › AN 1 €» с в MA N e € © «we oo So €x e soo о ce» cr 4 C e t ч с> с єз А p" ч ч о ax с ра, чо dà оч оч ооо ч оо CO MO O су Єз Сә Ө Oc Oe Oe Oe бф © QC {+ n4 С d с e Ge et e = Жз 4 4 4 c4 4 ex Od с AN ? ce eo о о ч © о о о о eon о о о о о о cx c dad (eC rx dad bo сч о о о о © о о ті і со с oud “ч e 4 Е =ч о ч о dad dao GE = о C. Oe 0 H Ce ооо о о чоо чо оооо о оооо о e O4 О ч n оооо о чоо mH о O OOOO O GC OOO ooo oer ooa = & 3 O cà dua ed oed c4 c4 oc Eo (P 3 & › © сч сч о сч Сч C$ € OC C'S оо чо о о оосо со о e 4 rbv ou ооо со ооо оо oO с С 4 о 1 > SH OX = r {ad wq wet 3 c4 › e О с, OO e e © с> ч о ad о о о о ч о о о e с о о eo о о د‎ © m і Oud No O O о © о о ті 1 c > єз © 4 сз 4 4 о о 4 24 Annals of the Mis souri Botanical Garden APPENDIX 2. CHARACTERS AND THEIR STATES USED IN THE MORPHOLOGICAL AN ALYSIS. ADDITIONAL NOTES AND LITERATURE CITATIONS ARE PROVIDED WHERE NECESSARY. HABIT, Фл Ro ш N -l ^^ e! 4. "eod sheath качар ны (0) straight; (1 . Basal cleistogenes: (0) abs FLORAL, AND FRUIT CHARACTERS . Plants with bulb-like structures: (О) no; (1) yes. [Swol- | ‘ > danthonioids were described by Linder & Ellis (1990).] . Basal sheaths: (0) persisting as tangled fibers: (1) not persisting as tangled fibers, : 0) fragmenting transversely; fragmenting lengthw — spiraling; (2) twiste Bout leaf margins: (0) glabrous; (1) villous. Leaf apices: (0) soft; (1) pungent. lant sexual system: (0) bisexual; (1) gynodioecious. ent; (1) present. [Described 1983) and Dobrenz & Beetle (1966). p by Clay 9. Inflorescence size: (0) huge (> 30 ст long); (1) smaller < 30 cm long). 10. ee ences: (0) paniculate: (1) capitulate. 11. Inflorescence лое (О) glabrous; (1) villous. 12. Callus relative length: (О) ca. equaling rachilla: (1) much larger than rachilla. 13. Number of florets per spikelet: (0) 3-8: (1) 2. 14. Lemma shape: (0) linear-lanceolate; (1) narrowly ovate. 15. Lemma veins: (0) 3; (1) 5—7; (2) 9 or more. 16. Lemma apex: (0) entire; (1) obscurely bilobed: (2 deeply bilobe 17. Hp e (0) absent; (1) d ines; (2) tufted; (3) scattered at base; (4) tufted i two rows: e. even felt on back: (6) short in basal half. then row of long hair. 18. Setae: (0) substantial 1 ) minute; (2) abs 19. Setae position on aes : te antral on lobe s; (1) in sinus between lobe and a 20. den marginal hairs: (0) dos or absent; (1) long and dens 2]. Palea keels: (0) folded; (1) flat. 22. Palea length relative to lemma: (0) taller than lemma: (1) equaling lemma; (2) shorter than lemma. 23. Lodicule bristles: (0) absent; (1) present. 24. Lodicule microhairs: (0) absent; (1) present. 25. Caryopsis shape: (0) linear; (1) obovate. 26. Caryopsis hilum: (0) pa tate; (1) linear, at least half as long as caryopsis. This character, for the African members, is largely based on the observations by Barker (1994).| . Fruit apical indumentum: (0) absent; (1) woolly or with long hairs LEAF ANATOMY CHARACTERS 28. Lamina asymmetrical: (0) no; (1) yes. [Lamina asym- metry is described in detail by Ellis (1981). | 29. Adaxial ridging: (0) absent; (1) у» i 30. Abaxial ridging: (0) absent; (1) pre 31. Clear cells in chlorenchyma: (0) enm nt; i present. 32. Clear cells red abaxial epidermis: (0) absent; (1) sparse (1—2 layers): (2) thick (+ 3 layers). 33. Bulliform a (0) absent; (1) ihi flanking the mid- rib; (2) in most furrow Adaxial papillae: (0) А int; (1) present. Adaxial papillae: (0) all over; (1) basally on ridges only. Adaxial prickle hairs: (0) absent; (1) present. 31. Abaxial zonation: (0) absent; (1) present. it Phloem: (0) entire; (1) divided. Extension cells: (0) absent; (1) p i Vase ular bundles: (0) + round, or e (1) taller than И. Midrih: (0) ii other vascular bundles; (1) raised and massive; (2) smaller than ыз A les. Multicellular glands: (0) absent: (1) present. [Multi- cellular. glands were describe] by Linder et al (1990). | Adaxial microhairs: (0) not overlapping; (1) overlap- ping in furrows. Adaxial иен инш сар: (0) anchor-shaped; (1) lens-shaped: (2) pyramidal. 15. Outer bundle sheath of pimay vascular bundles: (0) parenchymatous; (1) thickene 16. Scattered large cells in РИНЕ РЕР (0) absent; (1) present. 47. Cin ian spaces (leaf mesic/xeric ): (0) absent; (1) present. 18. Cushion-based macrohairs: (0) absent: (1) present. CYTOLOGICAL CHARACTERS 49. Basic chromosome number: (0) 6; (1) 7. ENDEMISM IN THE MEXICAN FLORA: A COMPARATIVE STUDY IN THREE PLANT GROUPS! Claudio Delgadillo M..? José Luis Villaseñor Ríos,” Patricia Dávila Aranda? and ABSTRACT Endemism values are not equivalent among the Mexican Musci, Poaceae, and Asteraceae. The number of endemic species varies from one group to the next in various type age, peculiarities of their life cycle, dispersal ability, =: ки response to selective pressures three major plant groups include 2373 endemic taxa among which of vegetation or geographical areas, perhaps in response to . In Mexico, the rasses, and 2030 are 86 are mosses, composites, Cluster analysis of a similarity matrix shows relationships between neighboring states and among pom ically related state groups. Along the Маре 'апїс Belt the e pairs does no Key words: re are areas « always get a geographic al basis. From the standpoint of the pene of of endemism in western and cen o, but est area in Mexico Asteraceae, endemism, Mexico, Musci, есщ Although Mexico, with 1,972,544 km, is the fourteenth largest country in the world, it ranks third in biological diversity (Mittermeier, 1988). It harbors approximately 30,000 species of vascular plants, including more than 21,600 in about 2,500 genera of flowering plants (Rzedowski, 3). Among these, more than 300 genera and between 50 and 60% of the species are endemic to this country (Ramamoorthy & Lorence, 1987). There e 49 Mexican species of pines, representing more ce 50% of the total for the world (Styles, 1993), and 900 to 1000 fern species (Riba, 1993). The bryophytes include about 1700 species (cf. Sharp et al., 1994; Fulford & Sharp. 1990), and among them, the mosses compose about 25% of the Neo- tropical moss flora. High plant diversity and the large endemic ele- ment are features that set apart the flora of Mexico. Information on the number, origin, and distribution of endemics (e.g.. Rzedowski, 1978; Sharp, 1953) is still imprecise, but current data suggest their concentration in certain areas such as the Neovol- canic Belt, a mountain range bisecting the country between 19 and 20°N, and the Sierra Madre de Sur. along the southern Pacific coast, which are considered centers of endemism for many groups (Ferrusquía, 1993). Pertinent literature for vascular plants includes contributions by Rzedowski (1962, 1991a, b). in which the endemic taxa and their geo- graphical ranges are identified. Preliminary observations indicate that the num- ber of Mexican endemics is associated with cli- mate- and geography-dependent factors. Thus, for instance, in the lowland moist areas of southern Mexico the percentage of endemic vascular plant genera is the lowest in the country, while their num- bers increase toward the drier (Rzedowski, 1978) and cooler areas. On the highest mountains, the extreme climate may have caused many species to become narrowly adapted to the environment of the alpine meadows and subalpine elevations. Beaman and Andresen (1966), in a survey of the vascular flora of the summit of Cerro Potosf in northern Mex- ico, detected 27 of 64 species (42.2%) endemic to the Sierra Madre Oriental; 13 of them were restrict- ed to Cerro Potosi. High endemism values have the Tehuacán Valley (Smith, 1965) where endemism approaches 17% (Villasefior, 1993). The significance of these observations cannot be been detected in the dry lands o fully evaluated for the entire flora. The main lim- iting factors are the lack of complete data on the ! Grateful thanks are extended to Isela Rodíguez Arévalo who assisted in the compilation of the lists of endemic taxa and bibliography on the subject, and to CONACYT-Mexico, through a grant from Sistema Nacional de Investigadores to the first author, which provided support for Mrs. Rodríguez. piam Ortíz Bermüdez prepared the map of for this m Victoria iden and deni Pruski reviewed a comments for its improvement. We than and two ano 21 xico version of the manusc ript and made useful nymous r 'onstructive criticism Je e de Botánica, E de Biología. UNAM, bd Postal 20. 233, 04510 Мехіс о, D.F. moya@ servidor.unam.mx and vrios@ibiologia.unam.mx * Facultad de Estudios ee ores Iztacala, Av. de los Barrios no. 1, Los Reyes Iztacala, 54090 Tlalnepantla, México. pdavilaa@servidor.unam.m ANN. Missouni Bor. GARD. 90: 25—34. 2003. Annals of the Missouri Botanical Garden geographical ranges of taxa and incomplete inven- tories or checklists of the major groups in the Mex- ican flora. Although these will not be attained for many years yet, the use of an alternative strategy may still permit reasonable estimates of endemism, its distribution in Mexico, and how endemism re- lates to the geography of the country. In this con- tribution, we make floristic comparisons among taxa for which preliminary lists and geographic ranges are available. As specialists, we have produced and compiled information on the Mexican Musci, As- teraceae, and Poaceae that may serve to illustrate patterns of endemism for the entire country, or for such specific areas as the Neovolcanic Belt of cen- tral Mexico where the flora is best known. For Mexican mosses, in addition to data in a 1994), relative importance of the endemic species have recent flora (Sharp et al.. the number and been published in several contributions. In the de- ciduous forests of eastern Mexico eight species were once considered endemic (Delgadillo. 1979). but further study has shown virtually no endemism in these communities, as is the case of the moss flora of the Yucatan lL (Delgadillo. 1984). In the dry lands of Zacatecas (Delgadillo & Cár- 1987) and the Tehuacán Valley & Zander, 1984), the proportion of endemics is low. denas, (Delgadillo 5 but appears comparatively higher than in the trop- ical lowlands of southern Mexico. Only five species are recognized as endemic in Zacatecas (4.3% of the moss flora), and four in the Tehuacán Valley (7% of the moss flora). In the alpine areas there are 19 endemic species that account for 17% of the 1987). The studies on the Poaceae of Mexico are mostly moss flora there (Delgadillo, 1971. floristic in nature, but many contain reliable data on the distribution of species in the country (e.g.. Hernández X.. 1959, 1964: Johnston. 1940: Miran- da. 1960: Rzedowski, 1962. 1965, 1975. 197% 1993; Sharp. 1953). A valuable discussion on grass endemism was contributed by Valdés and Cabral (1993), who indicated that a total of 272 species (309€ of the grass flora) are endemic to Mexico. The Chloridoideae have the highest number of endem- ics, with 73 followed by the Panicoideae with 46, and the Pooideae with 43. According to species, these authors, the states with the highest numbers of endemic grasses are Jalisco, México, Veracruz, and Oaxaca. Despite the cosmopolitan nature of the Poaceae, their distribution patterns are well defined and are known to be correlated with edaphic and climatic features. With regard to the Asteraceae, preliminary esti- mates by Villaseñor (1993) include 1813 endemics out of 2861 species for 63.4% endemism in the Mexican flora: these belong in 371 genera. 67 of which are endemic to Mexico (Villaseñor et al.. 1998). The number of species known from Mexico in this family was expected to rise to about 3000 and, with this, an increase in the number of en- demic taxa in certain areas; Villaseñor (1993) sug- gested a trend toward higher endemism values in states located in the drier northern and southern areas or in the mountain region of Mexico. Unpub- — lished data for the Valley of Tehuacán recognize a present, in addition to the taxa endemic to the country, 32 species restricted to the valley ош of 358 Asteraceae for the local flora, and in Zacatecas l restricted taxa out of the total 488 species. The flora of Nayarit comprises 447 species of Astera- ceae, 15 of which are restricted to the state (Ortiz- 1998). insula and Tabasco there are 7 restricted endemics Bermúdez et al.. while in the Yucatan Pen- from a flora of 252 species (Villaseñor, 1989). MATERIALS AND METHODS A Microsoft state distribution and the names of species of Mus- ACCESS database containing the ci, Asteraceae, and Poaceae restricted to the polit- ical limits of Mexico was compiled from biblio- graphie sources and support from herbarium specimens. Sharp et al. (1994) and Delgadillo et al. (1995) were the main sources for mosses. In addi- tion to numerous monographs, information маз compiled from publications such as Davidse et al. (1994). MeVaugh. (1983). Valdés-Reyna and Davila (1995) for the Poaceae, and MeVaugh (1984). Rzedowski and Calderón (1995), Strother (1999), and Turner (1997), as examples, for the As- іегасеае. The main herbarium sources include MEXU for the mosses, and МЕХ, ENCB. IBUG. MICH, and US for the Poaceae and the Asteraceae. and Specimen data were used to complement the taxon distribution. Database information was used to compute Jac- card's Index of Similarity and a cluster analysis to determine the floristic relationships of the Mexican states. The database file information was exported to Microsoft EXCEL tables as the first step to use an NTSYSpe version 2.02 software package (Rohlf, 1998). Geographical Units. i.e.. А ргехепсе-арѕепсе OGUs (Operational states) matrix served to calculate the index of similarity and to produce a The UPGMA (unweighted pair- group arithmetic averages method) dendrograms (Figs. 2—1) were generated by the SAHN-clustering command in NTSYS-pe. Similar procedures were similarity matrix. used to review the relationships of individual groups or smaller areas in Mexico, e.g., the Neo- Volume 90, Number 1 2003 Delgadillo et al. 27 Plant Endemism in Mexico volcanic Belt states. As a whole, the present anal- ysis concerns 2373 endemic species, including 15 subspecies and 339 varieties, in the three plant groups studied. The database and the similarity matrices are available on request from the authors. For reference, the political subdivision of Mexico and the location of certain geomorphological fea- tures cited in the text are shown in Figure 1. The Neovolcanic Belt states discussed elsewhere in the text are Nayarit, Jalisco, Colima, Michoacán, Que- rétaro, Hidalgo. México, Distrito Federal, Morelos, Tlaxcala, Puebla, and Veracruz. The density values shown in Table 2 represent the computation of a simple density index of ecology textbooks, i.e.. number of species per unit area. RESULTS ENDEMISM IN MEXICO From our records, there are 86 endemic moss species in Mexico, 4 of which are represented by subspecific taxa. Most endemic species are known from below 2800 m, but there is a group of about 18 species known only from above 3000 m in el- evation. Among the species from the higher eleva- tions, Astomiopsis X altivallis Delgad. is conspic- uous for its presumed hybrid origin between А. amblyocalyx C. Muell. and A. exserta (Bartr.) Snid- er; Archidium acauloides Schwab, a cleistocarpic species, is also of interest because it represents a form with limited dispersal ability. Because of the small number of species involved and their narrow ranges, no distinct geographical trends of state dis- tribution are shown by the cluster analysis. How- ever, most states along the Neovolcanic Belt are grouped together, and harbor, along with Oaxaca and Tamaulipas, more than 10 endemic taxa (Table 1). Despite the disparity in group size, the Poaceae and Asteraceae show similar behavior, i.e., they are best represented in certain adjacent states, in the Neovolcanic Belt states, and in Oaxaca. The values for all three groups seemed to confirm this trend (Table 1). With respect to the Poaceae, a total of 257 en- demic species—including 12 subspecies and vari- eties—out of 950 grasses, have been registered for Mexico (Tables 1, 3) for 27% endemism. Some spe- cies, such as Festuca hintoniana Alexeey, are known only from one or a few localities, while oth- ers are exclusively known from the type locality, as is the case of Schaffnerella gracilis (Benth.) Nash. By contrast, many endemic species, including Bou- teloua scorpioides Lag., Muhlenbergia gigantea (Fourn.) Hitche., M. firma Beal, Bothriochloa hir- tifolia (J. Presl) Henrard, Panicum decolorans Kunth, and Urochloa meziana (Hitche.) Morrone & Zuloaga, are widespread in Mexico. Except for Ta- basco, there are endemic grasses known from every Mexican state, mostly distributed at intermediate elevations (ca. 1500-2800 m). The highest number of endemic species is found in Jalisco, México, and Michoacán, with 55 or more species, but the states of Chiapas, Chihuahua, Durango, Guanajuato, Nue- vo León, Oaxaca, Puebla, San Luis Potosí, and Ve- racruz have between 32 and 49 endemic species. In contrast, Baja California, Campeche, Quintana Roo, ' demic species. In contrast to the results reported by Valdés and Cabral (1993), the present study in- cludes species with a strictly Mexican range only. Tlaxcala, and Yucatán have less than 10 en- If the grasses restricted to the southwestern United States (California, Arizona, New Mexico, and Texas) and Mexico were included, the number of endemic species would increase to about 300, with the high- est number of them occurring in the semiarid hab- itats and the alpine grasslands. Endemic Poaceae are present in low numbers in the states of Cam- peche and Quintana Roo, and are unknown from Tabasco. The flora of Mexico includes about 3003 Aster- aceae; 1972 of them, or 65.7%, are endemic to the country. However, for the analysis. 2030 species, subspecies, and varieties of endemic Asteraceae were accepted, i.e., incorporating 58 taxa not fully documented and increasing the percentage to 67.6 (Table 3). The endemic taxa include 10 subspecies and 328 varieties. The known altitudinal interval for the Mexican Asteraceae places many of the en- in the intermediate elevations (ca. 1500- 2800 m), and their individual ranges are fre- quently broader than those of mosses and grasses. Some species of Asteraceae have narrow ranges that depend on the presence of special habitats, e.g.. Geissolepis suaedifolia B. L. Rob. or Stephan- odoria tomentella ( ob.) Greene that are en- demie to gypsophilous grasslands in San Luis Po- tosí; other species, suc (Kunth) Cass., hua and Durango south to Puebla and Oaxaca, as Psacalium peltatum whose range extends from Chihua- demonstrate a comparatively broad distribution in Mexico. In terms of the states, those with the largest number of endemic species are Jalisco, México, Mi- choacán, Oaxaca, and Durango, with 385 to 526 species in each state. А second group, formed by Guerrero and Puebla, contains between 317 and 365 species (Table 1). Cluster analysis of similarity data for the Aster- aceae indicates that many Mexican states are re- lated to their neighbors; the overall analysis for all three groups (Fig. 2) clearly shows this trend at the Annals of the 28 Missouri Botanical Garden € 8. т а Аду t1 46 == 66 VOA чуен, = LE A) Sno = OF ASSI e uw =. es hof = мы чыгъ "x M i кч «ОЈО EVO а . Е 4 4 f. ue) sedi[neureg = 8с (q*p) o»seqep = 2% (uos) рлоиос = QZ (ш) eo[PutS = се А715) 180104 sip ues = pz (¥0) ооу вившп() = cc '(01()) олетәләп() (xeo) вәвхе() = OZ IN) uoo] o*9nN = GT (хем) LIBRARY = QI (ION) SO[240]N = 21 Чоу) ирәвоцоцү = 9[ (хәр) IXW = CT (IPP oosref = py ) = ZI (019) ownfeueny = рр (08р) o3uean(q = QT (Aq) Pipaq osiq = 6 (07) ewop = 8 "(qeor)) Byinyeory = 2 "(unp) engengiyp = Р (S94) Ang erusoyyey вівя = c (94) euge?) eleg = а '(s3y) SQUITRISENSY = | соохәр Jo UOISIAIpgns [воо] ср әлі | 06 00, ои | e[eoxe[L = 67 (sd = ZZ (mnq)epmongq = [cc "(o3H) ospeprg = ер (019) оләлләп‹“ 9 (su) sederyy = c *(duie7?)) eyooduey = Volume 90, Number 1 2003 Delgadillo et al. Plant Endemism in Mexico Table 1. Number of endemic taxa per state for each major plant group investigated in Mexico. Number of species endemic to a state are given in parentheses; Neovolcanie Belt states are shown in bold. State Musci Asteraceae Poaceae = Aguascalientes 0 (0) 154 (0) 19 (0) 173 (0) Baja California 1 (0) 140 (32) 5 (2) 146 (34) Baja California Sur 1 (0) 103 (48) 12 (4) 116 (52) Campech 0 (0) 4 (0) 5 (0) 9 (0) dau 6 (3) 135 (37) 38 (6) 179 (46) Chihuahua 2 (0) 275 (58) 14 (8) 321 (66) Coahuila 1 (1) 223 (46) 1 (10) 265 (57) lima 0 (0) 113 (10) 20 (6) 133 (16) Distrito Federal 12 (3) 180 (1) 25 (0) 217 (4) Durango 4 (1) 385 (59) 18 (1) 437 (61) Guanajuato 0 (0) 213 (2) 38 (0) 251 (2) Guerrero 2 (1) 365 (54) 29 (0) 396 (55) Hidalgo 10 (0) 271 (15) 28 (0) 309 (15) Jalisco 10 (3) 526 (64) 80 (15) 616 (12) México 21 (2) 391 (15) 67 (7) 479 (24) Michoacan 17 (1) 460 (25) 55 (4) 532 (30) Morelos 6 (0) 230 (5) 22 (0) 258 (5) Nayarit 4 (2) 271 (15) 27 (5) 311 (22) Nuevo León 4 (2) 257 (41) 32 (4) 293 (47) Oaxaca 17 (5) 473 (115) 40 (3) 530 (123) Puebla 23 (3) 317 (13) 49 (1) 389 (17) Querétaro 1 (0) 218 (6) 18 (0) 237 (6) Quintana Коо 0 (0) 6 (0) 2 (0) 8 (0) San Luis Potosí 8 (1) 249 (23) 48 (5) 305 (29) Sinaloa 1 (1) 224 (25) 21 (1) 246 (27) Sonora 1 (0) 160 (27) 23 (0) 184 (27) Tabasco 0 (0) 6 ( 0 (0) 6 (0) Tamaulipas 11 (2) 166 (16) 24 (2) 201 (20) Tlaxcala э (0) 119 (0) 9 (0) 133 (0) Veracruz 25 (3) 271 (21) 38 (6) 334 (30) Yucatán | (0) 9 (2) 8 (0) 8 (2) Zacatecas 5 (0) 208 (5) 22 (1) 235 (6) TOTAL 86 (34) 2030 (780) 251 (91) 2313 (905) regional level. Aguascalientes, Zacatecas, Guana- juato, Querétaro, Hidalgo, and San Luis Potosí form the first block of neighboring states that share nu- merous endemic taxa. The states in the peninsulas of Baja California and Yucatan stay together in the endemism dendrogram (Fig. 2) as do groups of states in northeastern (Coahuila, Nuevo León, and Tamaulipas), northwestern (Chihuahua and Duran- go), and central Mexico (Distrito Federal, Tlaxcala, Puebla, Veracruz, Guerrero, México, Michoacán, and Morelos). The position of certain states does not conform to geographical vicinity as, for in- stance, in the case of Oaxaca, which is closer to Morelos and Michoacán than to Puebla and Guer- rero, which limit it to the north and west; the en- demic flora of Chiapas remotely links that state to the rest of the country. The data set for mosses and grasses modifies the value of the similarity coeffi- cient and the relative position of many states in the dendrogram (Fig. 3). Such states as Aguascalientes, Guanajuato, San Luis Potosí, Zacatecas, Chihua- hua, and Durango from the first block in Figure 2 have a different pairing arrangement in Figure 3. Also, individual analyses for mosses and grasses fail to produce reliable dendrograms, as indicated by the lack of similarity among neighboring states, perhaps induced by the low number of records and, in mosses, by the absence of endemic records for about six states. The number of endemic taxa in the overall anal- vsis seems indirectly related to the size of each state; thus, for instance, Aguascalientes, Colima, and Tlaxcala are among the smallest states in Mex- ico and have some of the lower numbers (Table 1). By contrast, the low numbers exhibited by the states of the Yucatan Peninsula (Campeche, Yuca- 30 Annals of the Missouri Botanical Garden Table 2. Endemism along the Neovolcanic Belt of сап states. The percentage of endemism., with about Mexico. The second column gives the number of species — 62 shared endemic moss species in the Belt (8.5%). oe ed to at pare ud in the last ri shows is nearly as high as that for moss flora of the entire the corresponding density index also in parentheses. Den- . nm | i , | н ра country (8.8%), as shown in Table 3. A distinction sity = Number of endemics/surface area X 100. ‹ Е M must be made between "shared" and “restricted” Surface endemics; in this contribution the former refers to State endemics (km?) Density species distributed in two or more states while the latter are known from a single state. Colima 133 (16) 5,191 2.56 (0.31) The Aster cin 1 bv 1640 s - Jalisco 616 (12) 80,836 0.76 (0.01) Nee earn енен UI нарына Navari 311 (22) 26.979 1.15 (0.08) and infraspecific taxa along the Belt, or nearly 55% Distrito Federal 217 (4) 1.479 14.67 (0.27) of the Mexican Asteraceae. About 1095 of them are Tlaxcala 133 (0) 4.016 3.31 (0) endemic to Mexico and 190 are restricted to a sin- Hidalgo 309 (15) 20,813 1.48 (0.07) gle Neovolcanic Belt state (Table 3); the percentage Querétaro 237 (6) 11.449 2.07 (0.05) of endemism nationwide (67.6%) is nearly the same México H79 (24) 21,355 2.24 (0.11) as that for the Belt (66.8%). By contrast, there are баса 532 (: 59.928 ‹ 5 НЕЕ : e s А А Michoacán 032 (30) 59,928 0.89 (0.05) 222 species of Poaceae (23% of all Mexican grass- Morelos 258 (5) 4.950 5.21 (0.10) M | í ENS pd es) along the Belt states, 162 of which are shared Puebla 389 (17) 33.902 1.15 (0.05) dert бу еза к: Veracruz 334 (30) 71.699 0.47 (0.04) with other states (73% of the Neovolcanic Belt Po- aceae), and 44 of them restricted to this mountain i dd (Table 3). The Neovolcanic Belt may be con- | | dered an area of high diversity (4935 in these tán, and Quintana Roo) may not be dependent on a1c А Е | en groups) and high endemism (1319 endemics, their surface area as each one has between 38,000 and 50,000 km. On the other hand, the similarity and relationship among states along the Neovolcan- including 251 narrow endemics, Table 3) and might be recognized. by these criteria. as а separate flo- ? . : istic province. Rzedowski (1978) treated it as part ic Belt (Table 1. Fig. 2) suggest areas of endemism inox ? of the Southern Ranges province (see Fi g. 1). that require further analysis. т т à The dendrogram in Figure 4 shows the overall relationship of endemism among the states along ENDEMISM ALONG THE NEOVOLCANIC BELT | the Neovolcanic Belt with a general trend in a west- Mosses represent an important element in the east direction. Jalisco and Nayarit. on the western flora of the Neovolcanie Belt. The Belt occupies coast, are very similar to each other, with about 240 portions of the states of Colima, Jalisco, Nayarit, shared endemic taxa; the states of México and Mi- Distrito Federal, Tlaxcala, Hidalgo, Querétaro, | choacán also share numerous taxa (300) and to- México, Michoacan, Morelos. Puebla, and Veracruz, gether constitute a separate area of endemism. de- thus extending the width of the country (Fig. 1). spite the geographical vicinity with states on the About 728 moss species are known from this area. western coast. Colima and Tlaxcala have the lowest By virtue of this number, the states along the Belt endemism numbers along the Belt (Table 2), and тау be considered bryologically diverse, for they this is attributed in part to their small surface area. include about 74% of the mosses known from Mex- The latter state. however, is floristically more sim- ico. The Belt states are easily accessible, and their ilar to Distrito Federal (Fig. 4) than to Puebla or collecting record is better than that of other Mexi- Veracruz that surround it. Otherwise. the close flo- Table 3. Number of endemic species and percentage of endemics among Musci, Asteraceae, and Poaceae in Mexico and in the Neovolcanic Belt. Musci Asleraceae Poaceae Total Mexican species 982 3003 950 1935 Mexican endemics 86 2030 257 2313 Percentage in Mexico 8.8 67.6 27 48 Neovolcanic Belt species 728 1640 222 2590 Percentage from total 74.1 54.6 23.4 52.5 Shared endemies in Belt states 62 1095 162 1319 Percentage in Belt states 8.5 66.8 13.0 50.9 Restricted to one Belt state 17 190 44 251 Volume 90, Number 1 2003 Delgadillo et al. 31 Plant Endemism in Mexico Ї 0.00 0.17 Figure 2. іегасеае. Poaceae, and Musci. ristic relationship between members of most state pairs (Fig. 4) may be due to their geographical proximity. As with the overall analysis, species endemism is onlv indirectly dependent on size of the area 0.34 Jaccard's Index of Similarity UPGMA dendrogram of the floristic relationships among Mexican states as illustrated yy endemic As- Belt. pographic differences, and climatic regimes may along the Neovolcanic Geologic history, to- provide better explanations for the number of en- demics in a given area. By these measures, the states of Jalisco, Michoacan, and México may be ا L‏ ofo | | 014 Figure 3. and Musci. T obs T Jaccard's Index of Similarity ¥ LJ T T on o "056 UPGMA dendrogram of the floristic relationships among Mexican states as illustrated by endemic Poaceae 32 Annals of the Missouri Botanical Garden [ PUE EM LVR 013 "ой 007 ` 0.45 LI LI T 059 LI LI T Jaccard's Index of Similarity gure 4. UPGMA dendrogram of the s re PARA of the states along the Neovolcanic Belt of Mexico as usc Fis illustrated by endemic Asteraceae, Poaceae, and M considered the most important areas of endemism Belt. constitute the largest portion of the Belt, and most along the Neovolcanic Together these states endemic taxa in the three major groups are repre- sented in one or the other states. Nevertheless, the computation of the number of endemic taxa per 100 km? indicates that Distrito Federal (with a density index of 14.67). Morelos (5.21). and Tlaxcala (3.31. see Table 2) are, by far. the most important areas of endemism in the Belt. A more accurate measure of the relative importance of each state may be ob- tained by determining the size of the range of en- demic taxa and their numbers per unit area. In this study, the density index calculated for the endemic species restricted to each Belt state (Table 2) is not meaningful because their general range size is un- known. DISCUSSION The use of a similarity coefficient and. cluster analysis has shown that the floristic relationships of states may be established with certain degree of accuracy. The resolution of the analysis, however, depends on the amount of field or herbarium infor- mation and an adequate taxonomie background. The input of data from other major plant groups should assist in refining the scheme of relationships among such states, particularly among those that show little similarity with their neighbors. In this contribution, the unusual position of a state may indicate areas of endemism or lack of distributional data, but other explanations may be sought in re- gional peculiarities or in the biological traits of the plant groups. The study of endemism by resorting to assem- blages of species from widely different plant groups may be advantageous because the combined. study of patterns may be a measure of the response of the entire flora to environmental factors that operate over broad geographical areas. The obvious disadvantage of this approach is that the taxa under investigation differ in size, evolutionary history, and biological at- tributes and, thus, in their response to selective pres- sures. For these reasons, it could not be assumed that endemism values in mosses would be similar to those of Asteraceae or Poaceae in the same region. Each taxon, by virtue of a differing life cycle or eco- logical preferences, does not operate under the same has differing re- selective pressures and, in fact. sponses. The dependency on running water for fer- tilization in mosses contributes to their minor rep- 'esentation desert areas. Nevertheless, the low number of moss endemics in other habitats may be due to relatively rapid spread of taxa following spe- ciation, Mosses are usually considered to be slow- evolving organisms, but the effect of somatic muta- tions on evolution and their rates of spread in local populations are unknown. From the standpoint of the life eyele, there are theoretical considerations. by which at least some populations may undergo rapid evolutionary change. For instance, if polyploids are produced by diplospory or apospory, or if a somatic mutation is retained in an otherwise haploid organ- ism, with the aid of asexual reproduction these pro- cesses may yield an independent and distinct taxon from one generation to the next. Rapid speciation, however, may not be the rule: new taxa may appear by slow evolutionary change and disperse gradually, Volume 90, Number 1 2 Delgadillo et al. 33 Plant Endemism in Mexico Thus, endemic mosses should be comparatively scarce in Mexico and elsewhere. This is supported by current phytogeographic and geologic information suggesting that the moss flora of Mexico has not evolved in isolation. In addition to the examples giv- en in the introduction, a recent study in the lowland areas of Chiapas (Delgadillo & Cárdenas, 2002) re- ports a single endemic species, Pylaisiadelpha shar- pii Crum, for the Lacandon Forest and more than 130 species shared with other continental areas. The broad geographical patterns exhibited by Mexican moss species and the age and North-South orienta- tion of mountain ranges agree with the hypothesis of rapid dispersal of newly evolved species in Mexico. The exploration of poorly known areas is not ex- pected to produce a sharp increase in the number of endemic species, but rather the decrease in per- centage endemism values as the distribution of de- scribed species is better known or as modern taxo- nomic evaluations result in synonymy. А few years ago, Delgadillo (1994) calculated nearly 11% moss endemism in Mexico; this contribution records only 9%, while the known number of species has in- creased from 943 in 1994 to about 982 in 2002. Lompared to mosses, grasses and composites represent heterozygote systems where sexual repro- duction, the length of the life cycle, and dispersal retard evolutionary events. Assuming similar rates of speciation, but differences in dispersal ability and age of taxa, vascular plants would be expected to be geographically limited, genetically stable, and narrowly distributed, more so than mosses. Long- lived moss species have been documented in the fossil record (Frahm, 2000; Miller, 1984), and broad continental abd intercontinental ranges are common among mosses (cf. Sharp et al., 1994). This may not be the case in vascular plants where, in addition, selection does not immediately elimi- nate mutant genotypes. Phenotypic expressions rec- ognized as endemic taxa may remain for a long time in recombinant diploid populations, even under negative selective pressures. In Mexico, although the number of moss and grass species are similar, the groups differ in their proportion of endemic taxa; the explanation for this difference may be sought among the biological attributes cited above. The present study illustrates how endemism values may not be equivalent between similar taxonomic categories, but does not support such differences in the taxonomic hierarchy. Local climates certainly act as strong selective forces for every plant group. Mosses, grasses, and composites, however, show differential responses to climate. The distribution of the first group in the drier areas of Mexico does not apparently follow obvious patterns. In fact, there are few endemic species in the desert areas of Zacatecas (e.g., Cur- viramea mexicana (Thér.) Crum and Jaffuelibryum arsenei (Thér.) Thér.; Delgadillo & Cárdenas, 1987), but in the alpine areas where mosses are dominant with lichens and grasses, an important endemic el- ement appears. Rzedowski (1962, 1991а) discussed the importance of the dry areas for endemism among vascular plants. In the higher elevations these plant groups may also increase their endemic representation. High UV radiation, daily tempera- ture fluctuation, low organic nitrogen and phospho- rus in the substrate of alpine and subalpine areas seem strong selective forces for all plants, including mosses, grasses, and composites. There are other differences that are evident in the present analysis. Degree of endemism varies among groups, and there are disparities in their altitudinal and latitudinal gradients and in the types of vegetation they occupy. Contributions by Delgadillo (1979, 1984) and Delgadillo and Zander (1984) attest to the uneven distribution of endemic mosses in the deciduous forests of eastern Mexico, in the Yucatan Peninsula, and in the Tehuacán Val- ley, respectively. Endemism values observed for the Asteraceae and Poaceae also indicate differences among geographical areas, even among those of similar surface area. On a national scale, these pre- liminary findings may be the basis for the identi- fication of areas of high endemism and the selective forces in operation. They will also assist in unveil- ing the history of the flora in Mexico and its rela- tionships to other floras in the American tropics. The Neovolcanic Belt, as an example, has been shown to be one such region where portions of the mountain range (Distrito Federal, Jalisco, México, and Michoacán) have higher endemism concentra- tions than the rest. However, a detailed floristic knowledge of less known or undercollected parts of the country, or even of adjacent areas in other coun- tries, is essential:to determine whether the differ- ential distribution of endemics is not an artifact derived from historical collecting preferences. Literature Cited Beaman, J. H. . W. Andresen. 1966. The "м, eloristica and phytogeography of the summit of C Potosi, Mexico. Amer. Midl. Naturalist 75: Davidse, G., M. Sousa & A. Chater (editors). 1 476-484 in Flora Mesoamericana, Vol. 6. ri ais 'eae а bE eae. Universidad Nacional Au- tónoma de México, México D.F; Missouri Botanical Garden, St. Louis; and The Natural History Museum, don. De ‘Igadillo M., 1971. Phytogeographic n on al- pine mosses Mexico. Bryologist 74: 331—346. 34 Annals of the Missouri Botanical Garden . 1979. Mosses and phytogeography | je т siio Tues of Mexico. Bryologist —— Mosses of the Yucatan TOM Mes "xico. Ш. ОТ Bryologist 87: 12-16 ل‎ . Moss distribution and the phytogeograph- ical iod ance of the Neovoleanic Belt of Mexico. J iin 14: 69-78. 1994. Endemism in the Neotropical moss flora. Biotropica 26 12-16 u „ûrdenas S. 1987. Musgos de Zacatecas, México. _ e eee ci. Bol. Soc. Bot. Méx- ico 47: 13-24. ———— . 2002. The Lacondon Forest (Chiapas, d A ye snchmark area for tropical mosses. Bryol- a pom 1984. The mosses of the huacán Valley, Mexico, and notes on their distribution. id ud 319-322. Jello B. & A. Cárdenas S. 1995. LATMOSS, a aie of Е un Mosses. Monogr. Syst. Bot. Missouri Bot. Gard. Vol. : Ferrusquía V., I. 19€ du of Mexico: A a Pp. 3-107 in T. P. Ramamoorthy, R. Bye, A. Lot & J. vie al Diversity of Mexie о: Origins and Press, New York. Mex au 105: —— & К. ~ Ww Fa (eio, RR E ua Univ. ка J.-P. 2 New and interesting mosses from Bal- tic and Sona amber. Lindbergia 25: 33-39. Fulford, M.&/ Sharp. 1990. The leafy Hepaticae of Mexic ‚0: One de and twenty-seven years jc ah sche. Mem. New York Bot. Gard. 63: Tune X., E. 1959. Patrones de (uv ue 1бп T al- gunos zacales mexicanos. Chapingo [México] 12: 392— 398 1964. Los pastos y pastizales. Pp. 97—133 in Las zonas áridas del centro y noroeste de México. Ed- iciones e uus ‘ano de Recursos Naturales Re- novables. Méxic Johnston, I. M. ub common to North and South nold Arbor. 21: 356—: MeVaugh. R. 1983. (side: In W. R. Anderson (ed- itor), Flora Novo-Galiciana. A Descriptive Account of the Vascular Plants of West tern. Mexico. 14: 1-436. Univ. Michigan Press, ———-—, 1984. e A ое pane 12: 1- The floristic signific ance of shrubs American deserts. J. Ar- ОҢ. Anderson (editor), of the 157. Univ. Flora Novo-Galiciar Vasc ular Plants of Nom Mexico. Arbor. 1984. Tertiary and Quaternary fossils. Pp. 1194—1232 in R. М. т : pes ' ditor), New Manual of Bryology. ue Bot. ‚ Nichinan. Miranda, F. с ано асібп del porcentaje de géneros bic ОР en América tropical. Anales Inst. Biol. Univ. Nac. México 30: 117-150. 1988. gs diversity and the trop- ical forest. Pp. 145-1 54 in К. Wilson (editor), Bio- diversity. Nationa Ac a Press Washington, D.C. Ortíz- num údez, E., J. L. Villaseñor & O. Téllez. 1998. La familia ‘ier ‘eae en el estado de Nayarit (México). Acta Bot. Mex. 44: 25-57. anam. T P. & D. H. Lorence. 1987. Species vi- ari n the Mexican flora and a description of a new species ‘of Salvia (Lamiaceae). Adansonia 2: 167-175 Riba, R. 1993. Mexican pteridophytes: Distribution and endemism. Pp. 379—395 in T. P. о К. Вуе, A. Lot & J. Fa (editors), Biological Divers y of We cuni Origins and Distribution. d Univ. Pres oss, New Rohlf, F. J. 1998. NTSYSpce. Numerical Taxonomy and in clog! Analysis Sem. Vers. 2.0. Exeter Soft- ware, or Rredowski. J: 1962. Contribuciones a la fitogeograffa flor- Algunas consideraciones fstica e histórica de México. |. acerca del elemento endémico en la flora mexicana. 2—65. Bol. Soc. Bot México 27 — ———. 1965. Relaciones geográficas y posibl fg de ie flora de México. Bol. Soc. Bot. Maden 29: 121- 177. ——. 1975. An ecological and phytogeographical analysis of the grasslands of Mexico. Taxon 24: 67-80. — ——. 1978. Vegetación de México, 1“ edición, p. 432. Editonal LIMUSA, México, D.F. 991a. Diversidad y orígenes oe a 2 fane- usw S: México. Acta Bot. Mex. 14 к ». El endemismo en la flora ш а те a ana: Un apreciación analítica preliminar. Acta Bot. 15: 47—64. 1993. Diversity and Ws of the phanerogamic flora of Mexico. Pp. 1 4 in T. P. Ramamoorthy, R. Bye, A. Lot & J. Fa (ec Tees (e npe of Mexico. Origins and Distribution. Oxford Univ. Pres New York. — ——— & С. Calderón de R. 1995. Familia Compositae, Bajío у de P na adya- entro Re- Tribu seo Flora del centes. . 38. Instituto de Ecologia, А.С. p dd Вајо, Pátzcuaro, Michoacé Sharp, A. . Notes on the flora " Mexico: World rd ORA of the woody dicotyledonous families and ES origin of the modern vegetation. J. Ecol. 41 т — ———. Н. A. Crum & Р. M. Eckel (editors). 1994 The em Flora of Mexico. Mem. New York Bot. Gard, 69: -11 ira с С Jr. 1965. Flora, Tehuacán Valley. Fieldiana, 01—143. 1999. Compositae-Heliantheae s.l. Pp. 1— 232 in T. F. Daniel (editor), Flora of Chiapas. Part 5 California Academy of Sciences, San Francisco. Styles, B. A 1993. Genus Pinus: А Mexican Purview. > : n T. P. Ramamoorthy, R. Bye, A. Lot & J. F (e 6 n Biolo ogical Diversity of Mexico. Origins and по — Univ. Press, New York. The Comps of Mexico. A systematic account 1 198 К Asteraceae, Vol. 1. Eupatorieae. — : А 21 Chorology of Mexican Grasses, Pp. 4: 1 T. P. Ramamoorthy, R. Bye A. Lot & J. Fa die Biologic ‘al Diversity of Mexico: Origins and eg Oxford Univ. Press, New York. — ——— & P. Davila. Clasificación de los géneros ps gramíneas (Poac сае) "Mene 'anas. Acta Bot. Mex. 33: 7-50. Villase fior, J. L. 1989. Manual para la identificación de las Compositae de la Península de Yucatán y Tabasco. Technical Report No. 4. = ho Santa Ana Botanical Garden. are Califor ——— 93. nilia e en México. Revista His. | a ol. Esp. 44: 117—124. — ——, G. Ibarra & D. Ocaña. 1998. Strategies for the conservation of Asteraceae in Mex Conserv. Biol. 12: 1066-1075. Soc. xic 0. YUCCAS, YUCCA MOTHS, Olle Pellmyr? AND COEVOLUTION: A REVIEW! ABSTRACT The obligate pollination ias d between yuccas не өсү апа der ca moths (Lepidoptera, Prodoridao), in which ne of th лене relatiot iships, ecological relationships, origin and reversal of the mutualism, and the potential for analyzing patterns of co-speciation and the historical role of coevolution on specific traits in driving diversification in the inter- action. Major novel developments in recent years include the recognition of a large sius complex of pollinators, н eh thought to be one polyphagous species; a majority of all moth species are monophagous. Considerable life history diversity has been unveiled, d mec hanisms that maintain a mutualistic equilibrium by preventing over- exploitation documented. Phylogenetic and ecological с including data from other, п ewly discovered facul- tative pollinators in the Prodoxidae, have been ee to erect a hypothesis for the evolution of obligate mutualism. Application of a molecular clock to phylogenetic data suggests dc the plant-moth association arose at least 44 е that the obligate mutualism evolved vey quickly after this event. Two separate events of reversal of sii other pollinator species. This appears to have happened not through selection for cheating, but rather as a byproduct of a phenological shift to an unexploited seed resource, in which case pollination behavior became redundant. Analyses of parallel diversification and character coevolution are hampered А incomplete phylogenetic information at the species level, especially for the plants, but also for the pollinators. Available data indicate considerable deviation from strict co-speciation, and no evident examples of this process. Analyses of the role of coevolutionary processes in driving the diversification of yuccas and yucca moths will be рее once fully resolved phylogenies become available. Key words: coevolution, Hesperoyucca, mutualism, Parategeticula, Prodoxidae, Tegeticula, Yucca, yucca moth. Coevolution, in the sense of reciprocally induced — plant-feeding insects often have increased rates of evolution, is one of the major processes driving di- diversification compared to sister groups with dif- versification and speciation (Farrell & Mitter, 1993; — ferent life habits; thus one or more life history as- Thompson, 1994, 1999a, b). Since first applied in pects of these groups appear to be important in plant-animal interactions as a hypothesis to explain driving diversification and speciation. This might diversification among butterflies and flowering involve, for example, chemical, physiological, and plants (Ehrlich & Raven, 1964), it has been ap- morphological arms races between the interacting plied successfully in comparative analyses to test organisms. rates of diversification in ecologically defined Our understanding of coevolutionary processes at groups, such as plant-feeding insects and parasit- — populational and species levels is still in its infancy oids (Mitter et al., 1988; Wiegmann et al., 1993; because identification of proximal factors of diver- Becerra, 1997; Farrell, 1998; Becerra & Venable, sification relies on strong phylogenetic hypotheses 1999). Several such studies show that plants and for the interacting organisms (Barraclough et al., ! [ thank the Missouri Botanical Garden for inviting me to speak at the 46th Annual Symposium, and to its archives for providing access to "nca notes on Riley's experiments. My studies of yuccas and yucca moths have been funded by de National Science Foundation and by the National Geographic Society. This work has been very much a team effort over the years. Co еен and invaluable information sources outside the lab have included John №. Thompson, Donald Davis, Jerry Powell, Richard Harrison, Jonathan Brown, Karen Clary, and Manuel Balcázar-Lara. progression of postdocs, graduate students, and undergraduate students in my lab have been wonderful collaborators over the years. They include Jim Leebens-Mack, Deborah Marr, David Althoff, Chad Huth, Kari Segraves, Joshua Groman, Beau Crabb, Mary Ann Feist, Mark Brock, Lindsey Elms, Jeff Keyes, Rachel Roberts, Andrea Farley, Allison Outz, James Goldmeyer, Xn Grimes, and Eric Weiss. Goggy Davidowitz aided with a reference in Hebrew. Colleagues too numerous to list here have helped greatly in providing samples and locality information over the years. Michael Long provided helpful information on George Engelmann's original iiec. Finally, I thank Annals editor Victoria Hollowell for her exceptional po e with me st the long gestation of this review. This paper a p to the neay of е Schmidt Nielsen, n, who revolutionized our understanding of basal Lepidoptera phylog 2D ent. of Biologic a gods es, University of Idaho, Box 443051, Moscow, Idaho 83844- "3051, U.S.A. iun udaha. edu. ANN. Missourt Bor. GARD. 90: 35-55. 2003. Annals of the Missouri Botanical Garden 1998; Pagel, 1998, 1999a, b), extensive life history data, and ultimately experimental testing of emerg- ing candidate traits (Armbruster et al., 19 ten the phylogenetic frameworks are missing, ond there are M few instances where all these criteria are A recent exception is the study of Becerra Te who used data for members of the plant genus Bursera (Burseraceae) and a group of spe- cialist herbivorous beetles (Blepharida; Chrysome- lidae) to suggest that chemical defenses and detox- ification traits have been a major evolutionary factor in their diversification. One of the most often cited cases of coevolution is the obligate mutualism between yuccas (Yucca and Hesperoyucca, Agavaceae) and yucca moths (Tegeticula and Parategeticula, Prodoxidae, Lepi- doptera). In this association, the plants rely on adult moths for their pollination, while the moth larvae require developing seeds to complete their development. This association was first recognized 1872; Ri 1872), and then served not only as an example of over a century ago (Anonymous, iley, remarkable pollination mutualism, but also as one of the first and strongest examples of evolution by means of natural selection. Together with a few oth- er models of obligate mutualism that involve seed- eating pollinators, the yuccas and yucca moths form a class of associations that are excellent systems for studies of coevolution, as well as of evolutionary and ecological dynamics of mutualism and its dis- solution. This stems in part from the relative sim- plicity of measuring fitness costs and benefits in these interactions; in most instances both plant cost and benefit can be measured in seeds. Second, in contrast to most other plant-pollinator associations, these are relatively exclusive associations, often with a single pollinator species per plant species, making it easier to measure reciprocal effects than when webs of many, simultaneously interacting taxa have to be analyzed. Considerable progress has been made in under- standing this unusual type of obligate pollination mutualisms in the past 15 years. This is certainly true of the long-recognized yucca-yucca moth and 1872; Weiblen, 2002). The two other documented examples of such L. [Ran- unculaceae| and Chiastocheta flies | Diptera: Antho- myidae], and Lophocereus schottii (Engelm.) Britton fig-fig wasp associations (Riley, obligate associations (Trollius europaeus & Rose [Cactaceae] and the moth Upiga virescens (Hulst) [Lepidoptera: Pyralidae]) were actually first documented during this period (Pellmyr, 1989, 1992; Fleming & Holland, 1998; Després & Jaeger, 1999; Jaeger et al., 2001). Here I will review our current understanding of the association between EA Figure |. Tegeticula yuccasella female collected by С. Engelmann on the ae night that he observed ps on Y. filamentosa flowers. Label likely written by C. V. Riley, who received the inn From the collections of USNM yucca moths and yuccas, and discuss its utility in exploring more км questions of coevolution. Since the latest reviews of this interaction, by Bak- er (1986) and Powell (1992), tematics, phylogenetic relationships, and life his- information on sys- tory has increased dramatically, especially in the moths, and the complexity of the association at dif- ferent hierarchical levels is now quite different. EARLY HISTORY OF STUDY or THE PLANT-MOTH INTERACTION The first observation of the yucca moths was made by George Engelmann in St. Louis in 1872 (Engelmann, 1872). Engelmann asked Charles Ril- ey, then state entomologist of Missouri, to explore the relationship between the moths and the plants. Baker (1986) provided a passage from Engelmann’s notes written on 13 June 1872 about the initial ob- servations the previous night, and one of the moths observed from that initial set of observations ap- pears to have survived. Riley donated his very large insect collection to the United States National Mu- seum of Natural History (Smithsonian Institution), where it became the nucleus for the creation of the Department of Entomology. Among his yucca moth specimens is one female Tegeticula yuccasella (Ril- ey) specimen labeled “found in Yucca flower—En- gelm. June 12/72” (Fig. 1). This date coincides with that of Engelmann's original observations at the Missouri Botanical Garden, and is obviously a moth given to Riley by Engelmann. Although this would have been an obvious candidate for holotype, it is not. Riley, a driven and opinionated worker, never bothered to designate or label type material for any of the many species that he described, but instead would mention in his descriptions the number of specimens used for the description (Davis, 1967). A lectotype having already been designated for T. yuccasella, the surviving moth from Engelmann's Volume 90, Number 1 2003 Pellmyr 37 Yuccas and Yucca Moths original observations has now been labeled to in- dicate its historical significance. Charles Riley was to dominate the field of yucca moth studies up until his sudden death in 1895, despite this being a sideline in his job as the first federal entomologist (Sorensen, 1995). One of his most important contributions was his involvement in the salvation of the French wine industry (Smith, 1992). I mention it here because it indicates Riley’s general understanding of the process of plant-insect coevolution. By the early 1870s, North American grape phylloxera aphids (Daktulosphaira vitifoliae (Fitch)) accidentally introduced in central Europe caused massive mortality of European grape culti- vars by attacking their roots. Riley (1871) reasoned that American Vitis species had coevolved wit phylloxera, and thus might tolerate them better. A grafting program with European cultivars and American roots proved highly successful in reduc- ing phylloxera impact, and carried the industry to financial survival; Charles Valentine Riley may be the only individual to have received the French Le- gion of Honor for contributions to coevolution. An extraordinary observer and able thinker, Ril- ey unfolded the basic natural history of the plant- moth mutualism and documented the life histories of the pollinator Tegeticula yuccasella and the bo- gus yucca moth Prodoxus decipiens Riley within a decade of the initial discovery (Riley, 1880, 1 In contrast to the records of most of his contem- poraries, there are very few inaccuracies in his ac- counts, simply because of his reliance on empirical observation. In this, he arguably belonged in the exclusive group of exceptional naturalists with whom he regularly corresponded, such as Charles Darwin, Alfred Russell Wallace, Henry Walter Bates, Thomas Belt, Fritz and Hermann Miiller, and Asa Gray. As one of the early protagonists of evo- lution by natural selection in the United States, Ril- ey went beyond natural history to use the relation- ship between the yuccas and the moths in discussing more general issues such as mimicry and animal pollination (Riley, 1871, 1892). relationship between yuccas and yucca moths, characterized in an 1877 letter from Charles Darwin as “the most remarkable example of fertil- isation ever published" (Burkhardt & Smith, 1994), drew the attention of many other naturalists in the first 15 years after the discovery. Riley was chal- lenged on numerous occasions regarding the ac- curacy of his observations. This included the ar- gument from P. C. Zeller, a German entomological authority whose experience with yucca moths was limited to three pinned specimens given to him, that it was simply too уе to be true (Zeller, 1875: 340—342). Others charged not only that Riley was incorrect but that the very phenomenon of in- sect pollination was a dubious notion in the first place (Boll, 1876; Мееһап, 1876); Boll went on to state that active pollination “belongs in the land of fables." Yet other critics challenged that his argu- ments about exclusivity of moths in pollinating yuc- cas were overstated (e.g., Hulst, 1886). Riley re- sponded to his critics with experimental results, not sharp pen (Davis, 1967). A prolific writer, with some 2400 entries in his bibliography (Ho & Yuille, 1990), Riley used the empirical data as he knew them to rebut and often scold his critics (e.g., Riley, 1877, 1881, 1887), and occasionally even stooped to ridicule.? Following Riley's death, a hiatus arose in the em- pirical study of the association. Trelease worked with Riley on behavioral and botanical aspects, performing extensive fieldwork, and published de- tailed observations on pollinator behavior as well s plant morphology and systematics in his works (Trelease, 1893, 1902). Considerable collections of both moths and plants were made by Susan Mc- Kelvey for her monographs on southwestern Yucca (McKelvey, 1938, 1947). Busck (1947) attempted a reassessment of moth-plant associations based on McKelvey’s insect material; his conclusions when correct generally followed those of Riley, but Busck misinterpreted morphological variation that he was the first to document among pollinator yucca moths, cheater yucca moths, and bogus yucca moths. Since the 1960s, information about the associa- tion has accrued at an accelerating pace from sev- eral lines of investigation. This includes systematic and phylogenetic studies of the organisms, as well as the ecological and evolutionary studies of the interactions between the moths and the plants. NATURAL HISTORY ORGANISMAL DIVERSITY The yuccas. The yuccas are part of the North and Central American family Agavaceae (Fig. 2). Recent data suggest that the sister group of Aga- vaceae may be the small family Camassiaceae, con- fined primarily to mesic habitats of western North America with the exception being one species in eastern North America (Pfosser & Speta, 1999). з V, Т, Chambers, an amateur lepidopterist, mistakenly used the first non-pollinating bogus yucca moth to chal- lenge Riley's description of pollinator yucca mo tham- bers, 1877). In a rebuttal, Riley (1880) untangled the con- fusion and use ambers’s moth to erect the new genus Prodoxus (Gr., “judging of a thing prior to experience”). 38 Annals of the Missouri Botanical Garden Funkiaceae Camassiaceae Hesperaloe Furcraea Beschorneria Agave in part Agave in part Manfreda in part Manfreda in part Polianthes Prochnyanthes . Tentative genus-level consensus phylogeny for Agavaceae and its sister families, adapted from Pfosser and Speta (1999), Bogler and Simpson (1996), and Eguiarte et al. (2000). The only two genera pollinated by prodoxid moths are indicated by gray box. Figure 2 Volume 90, Number 1 2003 Pellmyr 39 Yuccas and Yucca Moths Generic-level relationships are partly unresolved within the family, but are robust among the taxa involved in the obligate mutualism with yucca moths (Bogler & Simpson, 1996; Clary, 1997). The monobasic Hesperoyucca, originally described as a distinct subgenus based on such features as a cap- itate stigma (Engelmann, 1871), was long consid- ered a section within Yucca (Baker, 1986). Recent analyses show Hesperoyucca to be the sister group of Hesperaloe, a small genus of the Sonoran and Chihuahuan deserts (Bogler & Simpson, 1996). Im- portantly, Hesperaloe taxa are not associated with the yucca moths, and instead rely on hummingbirds (Pellmyr & Augenstein, 1997) and probably bats (Engard, 1980) i peroyucca and Hesperaloe constitute the sister group of all remaining yuccas (Bogler & Simpson, 1996). Yucca is divided into three sections: spongy- fruited section Clistocarpa, the fleshy-fruited sec- for their pollination. Jointly, Hes- tion Sarcocarpa, and the capsular-fruited section Chaenocarpa. Section Clistocarpa consists solely of Yucca brevifolia Engelm., whereas the two other sections consist of no more than 20 to 25 species each (Clary, 1997). Section Clistocarpa is charac- terized by the single autapomorphy of a thickened exocarp, as observed by Trelease (1893). Its posi- tion relative to the other yuccas is uncertain but possibly tied to the series iin nee of capsular- fruited species (Clary, he longstanding in- terest in yuccas and their анна: іп тапу notwithstanding, Yucca taxonomy and systematics remain in a state of flux, ological communities with much need for a modern revision. Revisionary work is complicated by the relative scarcity of her- barium material, caused in part by the logistic problems of preparing specimens from these large. succulent plants. Horticultural interests in the group also have contributed to a plethora of names, with many taxa narrowly delineated using in effect a typological species concept (sensu Mayr, 1963). Observed variation frequently has been attributed to assumed hybridization and introgression (e.g.. McKelvey, 1938, 1947; Webber, 1953), but this should be considered speculation as there is only one example where genetic evidence for introgres- sion between two yucca species is provided (Han- son, 1992). Phylogenetic analyses are limited thus far, but appear not to violate assumptions of mono- phyly of both section Sarcocarpa and section Chaenocarpa (Clary, 1997). The use of horticultural material or yucca cultivars of unknown origin in some studies may contribute to historical confusion about relationships. e Hesperoy SOM NOU CN a clade is na- tive to North America (Fig. 3), and its contiguous range has been extended into Central America and northern South America through the cultivation of lease, 1902; species have an in cultivation on other conti- nents, including in Europe since the late 1500s (Gerarde, 1633), but yucca moths have never been found either south of Mexico or on other continents. Riley (1881) attempted to establish them by send- ing batches of pollinator larvae in their cocoons to Darwin and Stainton in England, Planchon in France, H. Miiller in Germany, and Asa Gray in Massachusetts, for establishment on cultivated yuc- cas. Miiller (1874) reported that moths hatched, but no local ornamental plants were in flower. Darwin had no yuccas in cultivation, and forwarded his co- coons to Joseph Hooker at Kew, where their sub- sequent fate is unknown. The two larger Yucca sections, section Sarcocar- pa and section Chaenocarpa, have wide ranges that overlap in areas north and south of the border of Mexico and the United States (Fig. 3). The fleshy- fruited section Sarcocarpa is primarily southern, ranging throughout the Megamexico-1 biogeograph- ic region of Rzedowski (1993), and extending in one species northward to southern Colorado. The aberrant Y. aloifolia L. occurs in the northern Ca- ribbean and along the U.S. Mexican Gulf and southern Atlantic coasts; it reproduces vegetatively ut is not known to have a native pollinator. Its origin is unclear, and I will return to it later. The capsular-fruited yuccas are more northern in dis- tribution, ranging from the northern edge of the Great Plains in southern Canada southward to the Mexican High Plains. Whereas yuccas generally are associated with shrub desert, chaparral, or grasslands, many Mex- ican species often grow in pine-oak woodland (Ma- tuda & Pifia Lujan, 1980; Gentry, 1982). Packrat midden data from the Wisconsin glacial show that species such as Y. rostrata Engelm. ex Trel. that currently inhabit shrub desert grew in pine-oak woodlands in areas such as the Big Bend region of Texas during wetter periods (Van Devender, 1990). The most unusual habitats are those of the south- ernmost yuccas, Y. elephantipes and Y. lacandonica G. Pompa & Valdés. Both occur in rainforest, with the former having a terrestrial habit whereas the latter is epiphytic or epilithic (Matuda & Pifia Lu- jan, 1980; C. Beutelspacher, pers. comm.). The yucca moths. The yucca moths belong to the Prodoxidae, a basal family within Lepidoptera of 78 described species (Davis, 1998; Pellmyr, 2002) and at least 15 additional undescribed spe- 40 Annals of the Missouri Botanical Garden Figure 3. Approximate limits of natural distributions of Hesperoyucca and the sections of Yucca. Hesperoyucca, dark gray (California, Baja California, Arizona); Yucca sect. Clistocarpa, horizontal lines; section Chaenocarpa, diagonal positive-slope lines and medium gray; and section Sarcocarpa, diagonal negative-slope lines and black squares (latter along southeast U.S. coast). Only published collection records from the wild, herbarium records from UNAM and MO a Moss (1959), Steyermark (1963), Powell and Mackie (1966), Great Plains Flora Association (1977), Rowlands (1978), . . ‘ , open circles give actual sites). F 1 were used; for records from the Antilles (Trelease, 1902), specific locations and circumstances of each collection are unknown. Land areas south of Mexico have been excluded. cies (Frack, 1982; Nielsen, 1982; Pellmyr & Bal- lecular data (Brown et al., 1994; Pellmyr & Lee- cázar-Lara, in prep.). The sister family Cecidosidae — bens-Mack, 1999) together suggest that the mono- consists of gall-makers feeding mostly on Anacar- basic Prodoxoides, the only southern hemisphere diaceae (Nielsen, 1985), and it shows a typical — prodox i i ;ondwanan distribution. The presence of sister (Fig. 4). Greya is a diverse genus of boreal and genera in Africa and South America of these moths, temperate humid to semiarid areas of western North which are highly sedentary, strongly indicates an America (Davis et al., 1992), with the exception of origin of this family, and by inference the Prodox- idae, before the South Atlantic breakup 95-100 million years ago (Pellmyr & Leebens-Mack, 1999). Morphological (Nielsen & Davis, 1985) and mo- id moth, is the basal genus in the family a few basal members recently documented from easternmost Asia (Kozlov, 1996). Tetragma is con- fined to North America, whereas the large genus Lampronia is holarctic in distribution. These gen- Volume 90, Number 1 Pellmyr 41 2003 Yuccas and Yucca Moths at Extant Host families distribution S. America Cecidosidae Anacardiaceae . Africa New Zealand Prodoxides (1) Myrtaceae S. America Greya (18) Saxifragaceae W N. America Apiaceae E Asia Tetragma (1) Rosaceae W N. America Rosaceae . Grossulariaceae | Lampronia (27) Saxifragaceae Holarctic Betulaceae Cecidosidae- . Prodoxidae split Mesepiola (1) Nolinaceae W N. America Mya Prodoxus (11) Agavaceae N. America Woody monocots colonized 44.1+10.6 Mya Parategeticula (4) Agavaceae S N. America Pollinators . ——* 1.7+11.1 Mya | Tegeticula (13) Agavaceae N. America Figure 4. Genus-level phylogeny for Prodoxidae, based on mtDNA and morphological data, with information on plant host families а extant distribution. Estimated minimum dates for seminal events and trait origins are based on a molecular clock, calibrated based on biogeograp ic data from the sister family Cecidosidae. Tree information and dates from Pellmyr and Leebens-Mack (1999). Numbers in parentheses are numbers of described species era use a remarkable variety of host plants, includ- ing species of the Myrtaceae, Apiaceae, Rosaceae, Grossulariaceae, and Saxifragaceae, i.e., represen- tatives from four plant orders (APG, 1998). In cases where immature stages are known, the larva feeds inside plant tissue during early instars, and then from the outside while concealed inside folded leaves or cases during the final c of devel- opment (Davis, 1987; Davis et al., Colonization of woody monocots, woe observed in Mesepiola, coincides with colonization of arid habitats and a concurrent change in life habit to having larvae that feed inside host tissue until feed- ing is complete (Davis, 1967; Frack, 1982). Mese- piola feed on members of Nolinaceae, whereas the three yucca moth genera Prodoxus, Parategeticula, and Tegeticula feed on members of Agavaceae. Pro- doxus (the “bogus yucca moths” of Riley (1880)) coexist with the two other genera, but feed on tis- sues other than the seeds. They are not involved in pollination. Virtually all yuccas host Prodoxus spe- cies that feed inside the inflorescence scape, and most fleshy-fruited and spongy-fruited yuccas also host species that feed inside hardening galls in the exo- or mesocarp portion of the fruit. The recently described Prodoxus phylloryctus Wagner & Powell is so far unique within the genus in feeding as a communal gall-maker in fleshy yucca leaves (Wag- ner & Powell, 1988). In addition, the peduncles of at least six Agave species are used (Frack, 1982) by some Prodoxus species. I will not deal with them further here, as they are not directly involved in the pollination mutualism. The pollinating yucca moths belong in the genera Parategeticula and Tegeticula. Parategeticula, with four described species (Pellmyr & Balcázar-Lara, 2000), is unique in having lost the linear cutting ovipositor of prodoxid moths used for inserting eggs 42 Annals of the Missouri Botanical Garden into plant tissue, and instead their thick blunt ovi- positor is used in creating a groove on the surface where eggs are laid (Davis, 1967; Powell, 1984). In species with known biology, they also differ in that the larva bores into the young fruit, where it causes the formation of a gall-like structure (“суѕ of Pow- ell, 1984) formed from modified placental tissue and a few immature seeds that in effect fuse and are consumed from within. Tegeticula was until re- cently held to consist of three species (T. maculata (Riley), T. synthetica (Riley), and T. yuccasella) with broadly similar life histories (Baker, 1986). phological variation had long been reported within T. yuccasella but considered as intraspecific varia- tion (Busck, 1947; Davis, 1967); Davis (1967: 53) stated that more than one may delimitation on the Miles (1983) used morphometric data to demonstrate the pres- or- “biological entity” exist, but refrained from grounds of insufficient information. ence of at least three unnamed host-specific enti- ties. Further studies using morphological and mo- lecular tools have so far led to the description of 13 species (Pellmyr, 1999), and several additional taxa remain to be described (Pellmyr & Balcázar- Lara, in prep.). ically and molecularly highly divergent and may well consist of several biological species (Powell & Mackie, 1966; Segraves & Pellmyr, 2001), and Т. circumscribed contains two Tegeticula maculata is morpholog- synthetica as currentl species (Pellmyr, in prep.). All species consume seeds as larvae, but there is variation in oviposition timing and location. Pollinators oviposit at the time of flowering, but Tegeticula species, sometimes re- delay oviposi- ^ ferred to as “cheater yucca moths, tion to the fruit stage and have independently lost the behavioral and morphological traits of active pollination (Pellmyr et al., 96a; Pellmyr & Krenn, 2002). Intrageneric phylogenetic informa- tion for Parategeticula and Tegeticula is relatively well established (Pellmyr & Leebens-Mack, 2000), with the major remaining uncertainties revolving around a rapid burst of radiation creating most lin- eages within the T. yuccasella complex and the in- clusion of remaining undescribed species primarily from the southern portion of the range. A note of caution about older publications involving the s moths of the T. yuccasella complex is indicated: because of the historical lumping, many studies must be interpreted very cautiously and are some- times of little value, as studied species are not identifiable and because as many as three coexist- ing species may have been treated as one. Basic ecology of the plant-pollinator interac- tions. There is considerable variation in the eco- logical aspects of interactions among both yuccas and the moths, and here I only outline major shared elements. The female yucca moth of both pollinator genera is equipped with unique tentacular mouth- parts that she uses for pollen handling (Riley, 1892; Davis, 1967; Fig. 5A). She collects pollen from yucca flowers by dragging her tentacles across the anthers. The pollen is embedded in copious pollen kitt, almost to the point of floating in a semiliquid matrix in Hesperoyucca, and often comes off as a unit from the anther. The moth compacts the pollen using the tentacles, and then stores it as a batch underneath her head (Fig. 5A, B). The pollen is kept in place by adhesion alone, and the tentacles play no part in holding it in place. This load can be substantial, reaching nearly 10,000 grains in Te- geticula yuccasella females, and constituting nearly 10 percent of the moths body weight (Pellmyr, 1997). Pollen collection can recur on an occasional basis during the active life of the female, so her pollen load may consist of multiple pollen geno- types. Following pollen collection, the female seeks out flowers of her host species for the purpose of finding suitable oviposition sites. Under most cir- cumstances, only first-night flowers tend to be ac- cepted but under some circumstances older flowers may also be subject to oviposition (Riley, 1889). In Tegeticula, the female first walks around the ovary, and her decision whether to oviposit is influenced not only by the flower itself but at least in some species also by its visitation history (Addicott & Tyre, 1995; Huth & Pellmyr, 1999). In T. yucca- sella, females deposit a host-marking pheromone during oviposition, and subsequent visitors perform a crude estimation of pheromone quantity (Huth & Pellmyr, 1999). Visitors become increasingly un- likely to accept a flower with increasing number of prior visits. In one case of two coexisting pollina- lors, one species responded to visitation history whereas the other made oviposition choices inde- pendent * number of prior visits (Addicott & Tyre, 1995). e female decides to oviposit, she posi- tions mere in a species-specific location on the ovary and cuts into it (Fig. 5C). Most species pen- etrate the ovary wall and lay eggs inside the locule, but a few species oviposit very superficially under the epidermis. The female then uses the tips of her tentacles to scrape off a small amount of pollen from her batch, walks up to the stigma, and places the pollen on the papillose internal surfaces of the ate style using a series of 10-20 bobbing E). The only exception in this perfora movements (Fig. 5C, regard is T. maculata, which pollinates the capitate stigma of Hesperoyucca whipplei Torr. using the same scraping behavior as is used for pollen col- Volume 90, Number 1 2003 Pellmyr 43 Yuccas and Yucca Moths sure 5. —A. Head of Tegeticula carnerosanella female. arrows. i ross sec X^ of tentacle 1.0 mm tentacle and и 18 paci ated by black and white compacting pollen just collected from a ү. filamentosa stan ale T. interme dia positing intc ovipositing into (right) a Y. e ovary. —D. Fem fr uit; note constriction caused Carrière flower. Moth wing ndi in panels B-E 10-11 1 carnerosana (Trel.) McKelv. fiuit showing feeding path of 1 yuccasella ovipositions. —E. T. carnerosanella larva that has destroyed seven s P) held below the head, with left Large pollen load ( Female T. yuccasella males of T. yuccase E pollinating (top) and a. 8-day-old Y. filamentosa reculeanella sii pollinating Y. treculeanc L duse iie section through locule of mature Y. seeds; fruit a —F. nm. length 73 mm. For a set of color pictures of T. E behavior on Y. filamentosa, see Murawski (1997) lection. A female may repeat oviposition and pol- lination many times on a flower, especially if she started on a virgin flower. In 7. yuccasella, polli- nation almost invariably happens following the first oviposition on a flower, but females then become increasingly likely to skip pollination during sub- sequent oviposition bouts, and they also deposit less pollen per pollination event (Huth & Pellmyr, 1999). Females of the species that encounter a flower visited by one other female first. typically perform about half as many ovipositions and pol- linations as the first female (Hut Pellmyr, 1999), and a smaller yet significant reduction was observed in T. altiplanella Pellmyr (Addicott & Tyre, 1995, referred to as *deeps"). Once a female moves on, she usually walks to adjacent flowers and inspects them for suitability, then visits other side branches, and eventually she flies off to other in- florescences. Consequently, females perform both еч and xenogamous pollinations (Riley. uller, 1990; Dodd & Linhart, 1994; Pellmyr et Гы 1997; Marr et al., 2000); there is no ехрег- imental evidence of plant self-incompatibility and fruit set readily occurs following both types of pol- lination, but selfed fruits are highly susceptible to abscission when they develop in 2 with outcrossed fruits (Pellmyr et al., 1997; Richter & Weis, 1998; Huth & Pellmyr, rid Eggs of Tegeticula hatch within a few days, and larvae of species that lay eggs inside the locule start feeding on seeds immediately. In species that ovi- posit superficially, the larva first burrows in the ovary wall before entering the locule to feed on seeds (Wilson & Addicott, 1998; Pellmyr & Lee- bens-Mack, р number of seeds (Fig. 5F), depending on the spe- cies and factors such as the presence of abortive seeds that can reduce per capita consumption (Powell, 1984; Ziv & Bronstein, 1996; Bronstein & Ziv, 1997). Upon completion of feeding, the larva creates an exit path. It preferentially exits during rain, either night or day (Whitten, 1894), but per- haps more commonly at night (Groman & Pellmyr, unpublished data), and can spend extended time waiting inside the fruit for optimal conditions (Pow- ell & Mackie, 1966). The larva burrows into the arvae consume a variable Annals of the Missouri Botanical Garden ground, where it creates a silk-lined cocoon cov- ered with soil or sand particles. The exact location in the ground has never been reported, but from lab trials Riley (1873) reported depths of 7.5-10 em and Rau (1945) 2.5-7.5 em for T. yuccasella and perhaps also T. intermedia Pellmyr. Powell (1984) reported depths of 1-3 cm in shallow con- tainers for T. maderae Pellmyr. The larvae of five Tegeticula species (T. yuccasella, T. intermedia, T. cassandra Pellmyr, T. treculeanella Pellmyr, T. car- nerosanella Pellmyr) reared in my lab commonly created their cocoons at a depth of 20 ст where they reached the impenetrable bottom of the rearing canisters. The variation in reported depths among Tegeticula species may at least in part reflect depth of rearing canisters. The larva enters diapause inside the cocoon and pupates a few weeks before emergence. This may happen after a one-year diapause, but the larvae can remain in diapause in lab conditions for at least four years (Riley, 1892). Very high fruit set during mass flowering episodes in yucca populations that then effectively cease flowering almost completely for several years (Pellmyr, unpublished data) sug- gests that the moth larvae are capable of diapausing for several years in the field as well, and that there are unidentified cues that trigger completion of de- velopment and adult moth emergence. This is not to suggest that moth emergence is perfectly syn- chronized with host flowering—we know it is not (Frack, 1982)—but rather that a sufficient number has remained in diapause to emerge at the time of mass flowering to cause high levels of pollination. The life history of Parategeticula is known in less detail than that of Tegeticula, but oviposition and larval biology of one species, P. pollenifera Da- vis, has been described in detail by Davis (1967 and especially Powell (1984). The most obvious dif- — ference is that Parategeticula females oviposit on pedicels and in petals, rather than into the ovary. In this case, the larva chews its way into the ovary, and then proceeds to feed on partly modified seeds as described above. Larvae of P. pollenifera pupat- ed at 1—3 cm in shallow containers (Powell, 1984), and Р. elephantipella Pellmyr & Balcázar-Lara formed their cocoons at 2-4 cm depth in 15 cm of loose soil (Pellmyr & Balcázar-Lara, 2000). tegeticula pollenifera from southern Arizona invari- ara- ably required two years to complete development (Powell, 1984), whereas the tropical P. elephanti- pella emerged in the lab without a diapause (Pell- myr & Balcázar-Lara, 2000). Patterns of host specificity. In the traditional recognition of four species of pollinating yucca moths, three species were monophagous and the fourth species (T. yuccasella s.l.) was held to pol- linate all other yuccas. This appeared somewhat paradoxical, as most phytophagous insects show relatively high levels of host specificity (Ehrlich & Raven, 1964; Price, 1980; Farrell & Mitter, 1993; Thompson, 1994), especially when the phenological window for the insect to successfully oviposit is very narrow. Yucca moths, which only live for a few days (Kingsolver, 1984; Powell, 1984), must access the plant during the short flowering period, so moth populations would have to be locally adapted for the flowering periods of different hosts. For exam- ple. in areas such as the Big Bend region of the Chihuahuan desert, four yucca species coexist and have largely non-overlapping flowering periods spread out from February to early June. If a single pollinator species were to utilize all four species, this would require intraspecific polymorphism in emergence phenology with four distinct peaks in the moths. Busck (1947) and Davis (1967) specu- lated that T. yuccasella may be a complex, but suf- fered from a dearth of material available for study. The first solid. data supporting the hypothesized complex were provided by 083), who showed that the pollinators of three sympatric yuc- cas in southern New Mexico differed greatly in morphology. She described the entities but did not ormally name them. Addicott (1996) likewise pro- vided morphometric data suggesting the existence of several more host-specific species, and Pellmyr et al. (1996a) provided molecular phylogenetic data indicating the presence of a large complex. Thir- teen species, including eleven pollinator species, have since been described (Pellmyr, 1999 Given the revised moth species delineation, diet breadth among the pollinators is now more uniform Fig. 6). Using the yucca species delineations used in Pellmyr (1999), members of the 7! yuccasella complex have been recorded from 17 host species. —. Seven of the eleven pollinator species within the complex are monophagous, one has two hosts, two have three hosts, and one has six recorded hosts. Thus more than 7096 of all pollinator taxa are mo- nophagous, and the most oligophagous species uses six host species. The reason for this level of spec- ificity remains to be explored, but certainly involves phenological specialization on hosts and probably also selection for specialization on plants with crit- ical differences in ovary morphology. Interestingly. the two derived non-pollinating yucca moth species are known to use four and six hosts, respectively, giving them a significantly broader host range than the d with which ds coexist (Kruskal- Wallis test, — 5.68, 0.017). Proximal rea- Volume 90, Number 1 2003 Pellmyr 45 Yuccas and Yucca Moths | [] pre 1999 || pollinators 2000 = o L 8- = cheaters 2000 Number of moth species T TT T—T T 5 7 9 11 13 Number of host species Figure 6. Number of recorded hosts of described Te- geticula and Parategeticula species. The two « give data for the four species recogniz Black bars give number of hosts for described pollinator species and striped bars derived non-pollinator species as of 2000. Cheater species have significantly more — per species than pollinator species (Kruskal-Wallis test, x? = 5.68, p = 0.017). sons for the wider diet of non-pollinators remain to be explored. Because the non-pollinator larvae feed on seeds side by side with pollinators, diet is an unlikely explanation, but plausible hypotheses to test include a broader phenological window for spe- cies that oviposit into fruits, oviposition modes that are less likely to select for specialization, or higher potential for establishment on novel hosts through colonization because of limited resource competi- tion with resident pollinators (Pellmyr & Leebens- Mack, 2000; Marr et al., 2001). Alternatively, it may reflect species age; because the non-pollina- tors are among the younger species in the complex, they have simply had less time available for poten- tial diversification through host specialization. The role of copollinators of yuccas. Suggestions of pollinators other than yucca moths appeared shortly after the original description of the plant- moth interaction, and this argument has resurfaced in the last decade, leading one monographer to the unfounded conclusion that *when moth populations are low, the fly Pseudocalliope may be an important alternate pollinator [of yuccas]” (Verhoek, 1998). For this reason, it is worth revisiting the support for this untested hypothesis. will discuss first all species other than Yucca aloifolia, which is a special case in this regard. The earliest claim of copollinators was made by Meehan (1879), who presented fruits resulting from geiton- ogamous hand-pollination on a cultivated Y. glauca Nutt. as evidence that other animals could serve as pollinators. In the absence of hand-pollination, however, no fruit set occurred. Hulst (1886) was the first to use the fallacious argument of abundance of a particular visitor as evidence of pollinator func- tion when stating that honey bees (Apis mellifera L.) can be copollinators. Frustrated with such un- tested hypotheses presumed to be true, Riley used a range of experimental and observational ap- proaches to gather data to test them. Whereas he presented his conclusions in print (Riley, 1887, 1889, 1892), the original data were never pub- lished. Tabulated result sheets found in the ar- chives of the Missouri Botanical Garden showed that inclusion experiments using two of the most common flower visitors, A. mellifera (25 bees, 72 hr.) and the soldier beetle Chauliognathus pensyl- vanicus de Geer (Cantharidae) (36 beetles, 24 hr.), in separate gauze bags containing single Yucca fi- lamentosa L. inflorescences failed to result in any fruit development, whereas control inclusion ex- periments with yucca moths caused fruit produc- tion. Riley (1889, 1892) and Trelease (1893) fur- ther argued against copollinators on the basis of xtensive visitor behavior observations. For exam- ple. honey bees were found to mostly lap the floral exterior for water and honeydew exudates, and when inside the flower probed the ovary base far away from the stigma. Similarly, other visitors also rarely moved close to the stigma. Riley (1881, 1892) also emphasized that plants in areas without moths, for example where plants recently had been put into cultivation, never had been found to set fruit, even though a wide array of other insects were found on the flowers. He also noted that in areas where an introduced yucca species coexisted with a native, moth-inhabited species with different flowering phenology, fruit set was never observed, whereas rare flowering coincidence of individual plants with a native yucca had been known to result in fruit set. Speculation about copollinators was raised anew by Dodd and Linhart (1994). A lauxaniid fly (Pseu- docalliope sp.) found in abundance on Yucca glauca flowers, with some individuals found to carry mod- est quantities of pollen on their bodies, was sug- gested as a possible vector. No attempt was made to test experimentally whether the flies cause pol- lination. There is reason for skepticism, because, as Riley (1892) first pointed out, lack of fruit set is common in yucca populations for a variety of rea- sons, yet flower visitors other than moths are often common in those same populations. Further, even if occasional modest pollen transfer were to take place through visitors other than the moths, it would likely be of little or no ecological and evo- lutionary significance. This follows because flowers Annals of the Missouri Botanical Garden that receive small pollen loads or self pollen are highly susceptible to selective abscission (Richter & Weis, 1998; Huth & Pellmyr, 2000); thus a poor vector is expected to contribute minimally to plant fitness. The century-old hypothesis about existence of copollinators could readily be settled by the proper experiments. A simple experimental ap- proach could exploit the size differences between yucca moths and proposed copollinators by using selective screens that permit entry to smaller visi- tors (such as the lauxaniid fly) but exclude the larg- er Tegeticula moths; this approach worked well in determining contributions to pollination by flies and bumblebees selectively screened on Trollius 1989). ( of genetic diversity in yuccas (Feist, 1995; Massey europaeus (Pellmvr, Given very high levels & Hamrick, 1998), routine genetic analyses of any resulting seed progenies could also provide infor- mation about selfing and outcrossing rates Yucca aloifolia is the single exception to the lack of evidence for pollinators other than the moths. Introduced as a garden plant in Europe no later than 1596, in Australia Ьу 1885, and in Melanesia by 1880, it has been reported on several occasions to set fruit in locations outside North America even though there never have been any coincident moth reports. For example, Engelmann (1873) saw fruit- ing plants in Italy, Layard (1880) in ж of New Caledonia оп what undoubtedly was (MacKee, 1994), Riley (1891) conveyed a report from Australia, and Galil (1969) reported fruit set in a cultivated plant in Israel. The plant historically Y. aloifolia was scattered along the southeastern North Amer- ican Atlantic and Gulf coast, especially along sandy shores from central North Carolina to eastern Lou- isiana (Fig. 3). Occasionally plants set fruit in parts of that range, typically as a result of colonization by Tegeticula увео and T. cassandra from co- existing anc usly flowering Y. filamen- 1873; Riley, 1873; Pellmyr, 1999), but fruiting plants without oviposition scars tosa (Engelmann, or larval damage have also been reported (Riley, 1892; Groman, 1999). Riley hypothesized from flo- ral structure that it may have resulted from autog- amy, but Trelease (1893) found that he could pre- vent fruit set by excluding all floral visitors with a gauze bag in a plant that previously had produced fruit. This is the only reported experiment for any yucca that provides even moderate support for other visitors as copollinators. It needs to be replicated with reasonable sample size, and with inclusion as well as exclusion treatments. The situation is more complex as Yucca aloifolia can have fruiting and non-fruiting inflorescences within populations and even individual plants (Pellmyr, unpublished obs.). In sites with moths on other yuccas, it is generally explained by moths visiting only Y. aloifolia inflorescences that coin- cide with those of the native host (Riley, 1892). Elsewhere this cannot be the case. Conceivable ex- planations include intrapopulational variation in visitor guilds, and possibility of autogamy or gei- tonogamy, but they are relatively unlikely expla- nations in the first place, and no observational or experimental data exist to explore these or any oth- er hypotheses. It is also possible that Y. aloifolia is under limited selection for maintaining sexual re- production, as it reproduces very vigorously by veg- etative propagation. This happens both through rap- id clonal extension and establishment by broken-off plant parts (Brown, 1959); in coastal North Caro- lina, local residents disseminate the plant by cut- ting stems in 10-cm pieces that are tossed on the ground in disturbed sandy sites (J. Groman, pers. comm.). This habit of elevated vegetative propaga- tion, absence of an endemic pollinator, core loss in the fruit, and poorly synchronized flowering spread across many months suggest that Y. aloifolia may be an escaped cultivar. Described from European gardens, Trelease (1893) referred to it as a species without a known geographical origin, and this is still the case. It is most closely related to Y. ele- phantipes and Y. lacandonica, which are tropical forest dwellers along the Gulf side of Mexico from Veracruz to Yucatan and into northern ur (Ma- tuda & Piña Lujan, 4 within this range, they set fruit through the actions Javidse et al., 1994); specific yucca moth (Pellmyr & Balcázar-Lara, 2000, unpublished data) whereas plants are sterile elsewhere. This is most evident in Y. elephantipes, which is widely cultivated throughout Mexico and southward at least to. Panama for its comestible flowers. Yucca aloifolia has been reported from Mexico (Matuda & Piña Lujan, 1980). but exami- nation of available herbarium collections at UN for their records indicates that these refer to cul- tivated specimens and to Y. elephantipes (Pellmyr & Balcázar-Lara, unpublished data). In addition to its distribution along the shoreline of southeastern North America, Y. aloifolia is reportedly also estab- lished on Cuba, Jamaica, the Bahamas and Ber- muda (Trelease, 1902), where pre-Hispanic cul- tures are suggested to have used its roots for soap (Engelmann, 1873). This use, together with a dis- junct geographic range from the remainder of the genus and traits characteristic of cultivated plants, makes plausible a hypothesis that Y. aloifolia orig- inated from Y. elephantipes as a cultivar selected for its high vegetative propagation. If correct, phy- logeographic studies are predicted to show a ge- Volume 90, Number 1 2003 Pellmyr 47 Yuccas and Yucca Moths netically depauperate Y. aloifolia nested within Y. elephantipes. This would to perform from the perspective of the plant-polli- be an important analysis nator association, as possible corroboration would imply that occasional fruit set in the absence of moths in Y. aloifolia is irrelevant to understanding diversification and coevolution of the plant-moth mutualism. EVOLUTION OF THE MUTUALISM AND ITS ECOLOGICAL CONTEXT The origin of the mutualism long remained un- known, and was subject to little speculation for more than a century after its initial discovery. Two limiting factors loom important in this context. First, life history differences and variation in out- comes of yucca-yucca moth interactions had yet to be discovered. The interactions between the four historically recognized yucca moth species (Davis, 1967; Powell, 1992) and yuccas were held to be obligate mutualisms, so there was no apparent transformation series to analyze. Second, and per- haps more important in retrospect, the phylogenetic framework of the yucca moths at the family and genus level was not determined until the 1980s (Frack, 1982; Nielsen & Davis, 1985). At that time, life history data also started to appear for the close- ly related dir of prodoxid moths (Frack, 1982; Davis et al., 1992). A timeline for establishment of the yucca-yucca moth mutualism. The fossil record is quite poor for these plants and effectively absent for the moths, providing little assistance in dating the es- tablishment and diversification of the plant-polli- nator association. The only pre-Pleistocene yucca macrofossil is a 14-My old trunk segment described as Protoyucca shadishii Tidwell & Parker from Ne- vada, most resembling the extant Yucca brevifolia (Tidwell & Parker, 1990). Fossil pollen described as Agave has been described from the mid Miocene (Axelrod, 1979; Palacios & Rzedowski, 1993). Bre- mer (2000) used clocklike behavior in rbcL to es- timate the minimum age of the Funkiaceae, which is the sister family of Agavaceae + Camassiaceae (Pfosser & Speta, 1999), at 21 My, whereas Eguiar- te (1995) provided an independent rbcL-based es- timate for the Agavaceae of 14 My. For the moths, a mitochondrial DNA sequence data set was used to estimate age of their diversification, using bio- geographic events for calibration (Fig. 4; Pellmyr & Leebens-Mack, 1999). Colonization of yuccas as a host was estimated at having occurred about 41.7 Mya, with the diversification of the three genera that inhabit yuccas being so rapid that their dates overlap. Importantly, this narrow time window in- cludes the split between the two pollinator genera as well as the most basal split within Tegeticula (between T. maculata and all other species), show- ing that the pollination habit was established in a common ancestor very close in time to the coloni- zation of the yuccas by prodoxid moths. Given this rapid diversification of the moth lineages, we can infer that a basal radiation of yuccas was in exis- tence by this mid Eocene date, pre-dating current independent estimates for the plants. For compar- ison, there is strong molecular data from several fig wasp lineages suggesting that the obligate mutual- ism between figs and fig wasps had originated by the late Cretaceous, some 90 Mya (Machado et al., 2001). PATTERNS OF TRAIT EVOLUTION IN PRODOXIDAE LEADING TO OBLIGATE MUTUALISM The obligate mutualism between the moths and yuccas could have originated either through exten- sive trait-level evolution, or it could have been pre- cipitated in interactions where only minor quanti- tative changes in preexisting traits would be required to switch the outcome from antagonism to mutualism. This would happen, for example. by variation in outcome based on ecological context. Studies of the biology of more basal prodoxid moths that do not feed on yuccas lend support for this model of pre-adaptations, and also reveal two sep- arate origins of pollination mutualism between three members of the genus Greya and their saxi- fragaceous hosts (Pellmyr et al., 1996a). The first studies were made of Greya politella (Walsingham), a specialist of several species of Lithophragma (Saxifragaceae) (Pellmyr & Thomp- son, 1992; Thompson & Pellmyr, 1992). The female moth oviposits into the ovary through the tubular hypanthium, and pollen often is transported on an elongated abdominal segment. Experiments showed that oviposition was a highly effective pollination behavior. At study sites in Washington, an exten- sive guild of copollinators of mostly bombyliid flies and solitary bees also provided cross-pollination. Although none of the copollinator species was as effective per visit as the ovipositing moths, their relative abundance and far higher rate of visitation made them important contributors to pollination in the study population. In two years of study, G. pol- itella was estimated to have contributed 0.8-2% of all seed set in the study population. Their positive effect was effectively masked, as there was no sig- nificant effect of moth oviposition on net seed set. Their negative effect through larval seed consump- Annals of the Missouri Botanical Garden tion was also masked by other sources of variation in seed production. The outcome of this moth-plant interaction is thus strongly dependent on copolli- nator contributions, as there will be no selection on moths for improved pollination efficiency unless it is tied to fitness differences. The same result was found for Greya enchrysa Davis & Pellmyr, a highly effective pollinator of its hosts in Heuchera, where abundant bumblebee visitors masked beneficial ef- 1996b). In these interactions, variation in outcomes can be expected fects on seed set (Pellmyr et al., across the ranges of the species (Thompson & Pell- 2000), potentially leading to sustained selection for a stronger mutu- myr, 1992; Gomulkiewicz et al., alistic equilibrium between the moths and plants. The third case and second origin of pollination in Greya involves G. mitellae Davis & Pellmyr, a spe- cies whose larvae feed inside the flowering stalk and in leaf peduncles of Mitella stauropetala Piper. Moths pollinate while drinking nectar from the flowers. Whereas virtually all pollination was pro- vided by the moths in study populations, no selec- tion on increased pollination efficiency is expected in this interaction as larval fitness is unaffected by the incidental seed production during adult nectar- ing bouts. This case indicated that there must be a direct link between female pollination efficiency and progeny fitness to cause selection toward in- creased pollination efficiency and potentially obli- 1996a). Mapping of several life history traits that were gate mutualism (Pellmyr et al., necessary prerequisites for the origin of the mutu- alistic behavior by yucca moths indicated that most traits were basal to prodoxid moths or at least had evolved before the lineage leading to the common ancestor of the pollinator genera. Hence this sup- ported a scenario in which the life habits of pro- doxid moths commonly have states that make pol- linator function easy to acquire. At the same time, obligate mutualism that requires novel traits for highly effective pollination has only arisen once in the family, in the true yucca moths. Why did this happen in the yucca-yucca moth association, but not in the others? Pellmyr et al. (1996a) used an- cestor reconstruction of the yuccas to erect a hy- pothesis in which highly effective pollination in the moths evolved first, followed by exclusion of an- cestral copollinators through effective cessation of nectar production in the plants. А general feature of the Agavaceae is resource-limited fruit set (Suth- erland, 1982), where only a minor fraction of all flowers give rise to mature fruit. As prodoxid moths colonized yucca ovaries, they thus encountered a major new mortality factor for their progeny, be- cause all eggs inside pollinated flowers subsequent- ly abscised will perish. Floral abscission is highly selective, with fertilized flowers resulting from small pollen loads or self pollen having a much elevated risk of abscission (Pellmyr et al., 1997; Richter & Weis, 1998; Huth & Pellmyr, 2000). For this reason, variation in pollination efficiency pro- vided by female yucca moths can result in differ- ential abscission of flowers containing moth eggs, as females providing large amounts of pollen de- crease the risk of abortion. Importantly, this trait could evolve in the females against a background of relatively inefficient, nectar- and pollen-consum- ing floral visitors. In a second step, reciprocal spe- cialization in the plants on the increasingly effec- tive yucca moths is expected as the net fitness contributions attributable to the ancestral nectar- consuming visitors relative to energetic investments in the nectar reward became negative. Both selec- tive abscission as a mortality factor and high cost of nectar production were novel traits to the yucca- yucca moth association in the sense that they are not present in the plant-moth interactions imme- diately basal to it, and they may point to factors that could facilitate similar transitions in other as- sociations. Consistent with this prediction, much reduced nectar production and low fruit: flower ra- tio are characteristic of the recently described ob- ligate mutualism in the Sonoran desert between the columnar cactus Lophocereus schottii and its polli- nating moth, Upiga virescens (Holland & Fleming, 9) 199‹ REVERSAL OF MUTUALISM Mutualistic interactions contain an underlying evolutionary conflict in that the interacting partners are under selection for increased. exploitation. of each other (Trivers, 1971; Bull & Rice, 1991; Pel myr & Huth, 1994). In a plant-pollinator relation- ship, this might manifest as selection for higher ef- ficiency in reward extraction among pollinators, and smaller or more inaccessible rewards in the plants. In facultative relationships, such conflicts may re- sult in arms races that shut out excessive exploit- ers. For example, a decreasingly rewarding plant species may be abandoned by flower visitors that have a choice, while plant traits that reduce losses to poor pollinators in theory can evolve to complete exclusion. In obligate mutualisms that involve a single pollinator and plant, this conflict has а po- tentially different dynamic. In such instances, the evolution of a cheating mutant with a fitness ad- vantage over mutualist individuals is expected to lead to reciprocal extinction of the mutualists, at least at the population level and possibly on a spe- Volume 90, Number 1 2003 Pellmyr 49 Yuccas and Yucca Moths cies scale, depending on patterns of gene flow. For this reason, obligate mutualisms such as those be- tween yuccas and yucca moths were long consid- ered evolutionary dead ends (Soberon Mainero & Martinez del Rio, 1985; Bull & Rice, 1991). This is clearly not the case under all circumstances, as two distinct species of non-pollinating cheater yuc- ca moths derived from pollinating ancestors have been identified (Pellmyr et al., 1996a). The two de- scribed species, Tegeticula intermedia and T. cor- ruptrix Pellmyr, oviposit directly into fruits at dif- ferent stages of development, and the larvae consume seeds in coexistence with larvae of the pollinator species (Fig. 5D). Their presence can be very costly for host seed production; in one study of Y. filamentosa, seed destruction was tripled in populations where cheater moths coexisted with pollinator moths (Pellmyr et al., 1996a). Ecological data did not reveal any competition between co- existing larvae of the pollinator T. yuccasella and the cheater T intermedia (Marr et al., 2001), so coexistence is evidently not a problem, but the sep- arate issue of an evolutionarily stable origin of the cheater life habit remains to be explained. Phylogenetic analyses based on mitochondrial DNA sequence data suggest that the two species originated separately around 1.26 + 0.9 Mya; thus these are not ephemeral lineages Ls et al., 1996a; Pellmyr & Leebens-Mack, . А simple solution to the problem of escaping ri evolutionary dead end of obligate mutualism is coexistence of two or more mutualists on a shared partner. For example, if two yucca moths were to coexist on one yucca species, one moth species could evolve the cheater habit without causing failure of sexual re- production in the yucca. In this situation, recipro- cal extinction is only expected if both mutualists independently abandon the pollinator habit. With the recent recognition of a large number of polli- nator species, it has become apparent that coexis- tence of pollinator moth species is not uncommon, with at least five documented instances of two pol- linators sharing a host in all or part of its range (Davis, 1967; Powell, 1984; Tyre & Addicott, 1993; Pellmyr, 1999; Pellmyr & Balcázar-Lara, 2000; Pellmyr Leebens-Mack, 2000). One of those sympatry zones is implicated in the origin of T. in- termedia. 'This species is most closely related to the pollinator T. cassandra, and available data suggest that it may have evolved where 7. cassandra came into coexistence with 7! yuccasella in part of its range. The pollinating sister species of both T. in- termedia and the other cheater species oviposit in a way that distinguishes them from all other polli- nator species, and they have a characteristic ovi- positor that allows them to oviposit into either a flower or a young fruit. Thus, these pollinators may be preadapted for a switch to oviposition into fruit once a sympatric pollinator species is available to perpetuate pollination. Because of a selective ab- scission mechanism in the yuccas that causes flow- ers with many moth eggs of most pollinator species to be abscised within a few days of pollination, a large proportion of the seeds are simply not acces- sible for larval consumption by these pollinator species. Hence, a pollinator species that can delay oviposition by a few days and oviposits directly into young fruits can bypass the plant's abscission pe- riod and exploit a rich seed resource. In this sce- nario, the phenological shift can be an adaptive step into a novel niche that precedes the loss of pollination habit, which becomes redundant once fruits become the target of oviposition. Available data thus suggest that the origin of cheater yucca moths from pollinators did not result from selection for cheating per se, but rather as a byproduct of selection for exploitation of a previously untapped seed source (Pellmyr & Leebens-Mack, 2000). By analogy to evolution of non-cooperative pol- linators, it is in theory possible that cheating plants could arise in an obligate mutualism. In the case of yucca plants, that would entail the evolution of mechanisms that maintain pollination but prevent This could happen through mechanisms such as prevention of seed destruction by pollinator larvae. successful oviposition, or killing of the eggs or lar- vae. If an alternative, cooperative host species ex- ists in the area, such cheating by plants could be evolutionarily stable, whereas evolution of cheating plants in a single plant-single pollinator scenario 15 predicted to lead to extinction (Bull & Rice, 1991). The only proposed case thus far involves a popu- lation of Yucca baccata Torr., where Bao and Ad- dicott (1998) reported that the fruits of a substantial proportion of all plants lacked evidence of larva damage, and speculated that this might be evidence of a cheating mechanism in these plants. They did not speculate regarding a mechanistic basis, but mentioned that fruits without larvae had a distinc- tive shape. Further studies will be needed to de- termine whether a cheating mechanism indeed is in place It should be emphasized that the strongest, most direct evidence for selection for cheating in a m tualism would be direct evidence of individual life- time fitness gains. Such data are wanting for both yuccas and moths, and it is difficult to accrue such data. The major obstacle in measuring moth life- time fitness has been difficulty to track them during extended flight; it is likely a matter of time before 50 Annals of the Missouri Botanical Garden ЦУ г « = 6 8 = е & % © Ф = © = w 9 ct 8 - c © Ф ~ = © Ф t EF = q КЫ © c & = эы ~ Ф S м ы. Ф = 2 © з €& 5 Ф Ф = Q м > о t © c © o $ 8 б S Š 9? w 8 a o E Н E 5 Q E a a a a M M M к m [ | | @ E: Figure 7 name indicates that it is a cheater species. Two. thr › or three origins of y ~ Pellmyr 1999) Bud Pellmyr and Balcázar-Lara ( suitable technological tools will be available to solve this problem. In the yuccas, longevity of de- cades or centuries (McKelvey, 1938; Webber, 1953; Matuda & Рїйа Lujan, 1980; Webb, 1996; Comanor & Clark, 2000), with iteroparity in all but one spe- cies, and also different magnitude and possible plasticity in vegetative propagation, makes it diffi- cult ever to measure lifetime fitness. In conse- quence, surrogale measures, such as intact. seed production in the plants, are the best available op- tion. PARALLEL SPECIATION AND THE ROLE OF COEVOLUTION IN PLANT-MoTH DIVERSIFICATION When species are tightly associated, there is a probability that they may co-speciate (Kichler, 1948; Huelsenbeck ci al., 2000). Such parallel di- versification may result either from vicariance- based divergence or from coevolutionary processes between the species (Page, 1994). For this reason, obligate pollination mutualisms between seed-par- asitic pollinators and their hosts should be good candidates for parallel diversification, as potential divergence may derive from linked host speciali- 87. mojavella es of ale Sipe and Tegeticula, tracking host use each species. ee, or four origins ect. Саен агра иѕе. е from Pellmyr and Lecbens-Mac k (2000), host records from 2000). o з © - % = з - 9 x В 8 3 Ф Ф А m Ф с P ~ & = Q [^] c S © S = t S © © E & о à Фф Ф 3 о Ф P = о ۳ © Q I o Ф 2 % Q %* Ф e © > o = M M M ы M M ы M M m a " " m " 1 С] Hesperoyucca Section Clistocarpa (spongy-fruited Yucca) ШШШ Section Sarcocarpa (fleshy-fruited Yucca) ШШШ Section Chaenocarpa (capsular-fruited Yucca) Ez е аан and Chaenocarpa EL equivo A "(c)" following the of Yucca . Sarcocarpa use are indicated and zation in the pollinators and pollen-mediated gene 1995; al., 1996a). The fig-fig wasp associations and yuc- flow in the plants (Bogler et al., Pellmyr et ca-yucca moth associations are sufficiently speciose that analysis of parallel diversification is possible. Analyses of the association between figs and fig wasps have indeed indicated a high level of parallel diversification at the level of fig genera and sub- genera (Herre et al., 1996), while there is emerging evidence that this pattern breaks down to a fair degree at lower taxonomic levels (Lopez-Vaamonde et al., 2001; Machado et al., 2001). Analysis for the yuccas and yucca moths is still rudimentary as the yucca phylogeny is incompletely resolved, and the unresolved polytomy in the moth phylogeny also limits analysis (Fig. 7). Given the current unre- solved plant relationships, there are no strong can- didates for parallel diversification, although this may change with increasing phylogenetic informa- tion. Meanwhile, several lines of evidence indicate that there are numerous instances where coloniza- tion has occurred. The most obvious instance in- volves recent colonization by Tegeticula yuccasella of Yucca aloifolia. Similarly, T. baccatella Pellmyr, Volume 90, Number 1 2003 Pellmyr Yuccas and Yucca Moths which feeds on a fleshy-fruited host, is nested amid species that feed on capsular-fruited yuccas (Fig. 7), and thus supports a past shift assuming that monophyly of fleshy-fruited yuccas is upheld. The cheater T. corruptrix, also arising from an ancestor on a capsular-fruited yucca (Fig. 7). now utilizes both fleshy-fruited and capsular-fruited species. Second, the coexistence on a host of non-sister taxa of Tegeticula pollinators cannot be explained by parallel diversification; in principle, coexistence of a Parategeticula and a Tegeticula pollinator on a host could reflect two independent parallel diver- sifications with the hosts, but there is very little support from published host data for this explana- tion (Fig. 7). Third, instances where a pollinator species utilizes more than one host species (Fig. 6) cannot reflect parallel diversification, although they may possibly reflect an ancestral association with subsequent unilateral diversification in a monophy- letic group of hosts. Co-speciation does not require coevolutionary processes, and coevolution can act on organisms regardless of their history of association; thus the role of coevolution in driving diversification. be- tween the plants and the pollinators is an altogether separate matter. Selection on plant and moth traits that vary among species may arise either from the interacting partners or from factors extrinsic to the interaction. For example, traits likely to affect moth oviposition success, such as floral ovary morphol- ogy and moth ovipositor morphology. may be strong candidates for reciprocal selection as they directly affect plant and pollinator fitness. Meanwhile, traits such as petal shape and color may be more likely to be under selection based on a wide range of antagonistic interactions with other herbivores, as well as abiotic factors. To determine the historical role of coevolution in the diversification of an in- teraction, variation in divergent traits must be par- titioned to remove extrinsic components, 1.е., in es- sence to remove background evolution in the interacting groups attributable to other factors. This obviously requires groups of plants and pollinators that have coexisted during much or all of their di- versification, as is the case for yuccas and yucca moths. It also requires well-resolved phylogenies, preferably with estimates of internal branc lengths. This criterion is not yet met for the yucca- yucca moth association, nor for any other similar plant-pollinator association. Such analyses will be highly useful in evaluating the historical role of co- evolution in driving diversification and speciation in plants and pollinators. CONCLUSION It may seem a somewhat subdued note on which to end, that we cannot yet perform rigorous tests of the role of coevolution in the diversification of yuc- cas and yucca moths. But the reason is simply that a large amount of information about morphology, ecology, natural history, and phylogeny is required for any one association before analyses of the his- torical impact of coevolution can be explored. Most, but not all, of these requirements are now largely met. The last 15 years have seen a dramatic in- crease in our understanding of organismal diversity, especially among the insects, although much infor- mation from the Mexican range of the yucca-yucca moth associations remains to be published. Ecolog- ical and evolutionary dynamics have also become far better understood in the last decade, including the expansion into the realm of reversal of mutu- alism. Phylogenetic information is now arguably the primary limiting factor for analyses of coevolution and several other major questions, but there is rea- son to hope that robust information soon will be available for both groups. Ongoing parallel projects on subsets of fig-fig wasp associations (e.g.. Lopez- Vaamonde et al., 2001; Machado et al., 2001; Wei- en ush, 2002) as well as other mutualisms involving seed-parasitic pollinators (Després et al., 2002) also offer possibilities for grander compari- sons across mutualisms in the next few years. Whatever generalizations about factors mediating plant-pollinator mutualisms emerge from these еј highly specific associations can soon be used as а template in analyzing other, more complex plant- pollinator mutualisms. Literature Cited Addicott, J. F. 1996. Cheaters in yucca/moth mutualism. Nature 380: 114—115. ———— & A. J. Туге. 1995. Cheating in an obligate mu- tualism: How often do yucca moths benefit yuccas? Oi- kos 72: 382-394. — 1872. Insects shaped by the needs of flow- s. Nature 6: 442-444. APG [Angiosperm Phylogeny Group]. 1998. An ordinal classification for EL In of Towering plants. Ann. Missouri Bot. Gar 31-553 Armbruster, B xa ard, T. P. Clausen, E. "pe. JC: Log uvam, M. Matsuki, B. н у & Е. Andel. 1997. Do biochemical exaptations link evo- lution of plant defense and pollination systems? Histor- ical hypotheses and experimental tests with Dalecham- ia vines. Amer. Naturalist 149: 461—484. Axelrod, D. I. 1979. Age and origin of Sonoran i vegetation. Occas. Pap. Calif. riis Sci. 132: 1 Baker, H. G. 1986. Yuccas and yucca nudes m ic ical commentary. Ann. Missouri Bot. d 73: 556- Bao, Т, & J. К. Addicott. 1998. Cheating in mutualism: Annals of the Missouri Botanical Garden Defection of Yucca baccata against its yucca moths. aca d 55-159. Barraclough, T. G., A. P. Vogler & P. H. Harvey. 1998. Revealing is = tors Ey [он speciation. Philos. : 241-2: Becerra, J. X. chemical i nds in host use. Scier x 253-256. & a Venable. 1999. Mac меи ма of insect- lant associations: The relevance of s eue d to host affiliation. Proc. Natl. Acad. . U.S.A. 6 ^J 3. Simpson. 1996. Phylogeny of Aga- (A sequence variation. Amer. . Neff & B. B. Simpson. igins E the yucca-yuccé nes к= iation. cad. Sci. U.S.A. 92: € 686 Boll, J. 1876. Ueber die Вегас ios e: rcp و‎ schen Yucca-Arten. Entomol. Zeitun —405 Bremer, K. 2000. Early Cretaceous ineages a monocot flowering plants. Proc. Natl. Acad. U.S.A. 97: 4707—4711. 1995. Multiple or- Proc. Natl. . Е Millspaugh. 1962. The Bahama 1997. Costs of two non-mutu- alistic species in a Tune ca moth mutualism. Oec ologia 112: 379-38: Brown, C. A. 1959. e Carolina. Louisiana Re ree T M., 'ellmyr, J. N. Thompson & R. G. rison. TR Ps hondrial DNA phylogeny of ro od doxidae (Lepidoptera: Inc pian indicates a i ecological dive 'rsific ation of the yucca moths. Ann. tomol. Soc. s т. 87: 795-802. Bull, J. J. & W. R. Rice. 1991. Distinguishing mecha- nisms for the evolution of co-operation. J. Theor. Biol. 149: 63-74. Burkhardt, F. & S. Smith (Editors). 1994. Cale ve of the Correspondence of Charles ыс ы 1809—1882, with Supplement. Cambridge Univ. Pre COMES Busck, 1947. The yucca er Py. 180-185 in S. McKelvey (editor), Yucc Banks of Baton Vegetation of the Outer State Univ. Press, n- "ES f the ا‎ d Sn d States, Vol. 2. Amold Arboretum. Jamaica Pla Chambers, V. T. na Е Colorado. J.S. Geol. ee Survey Terr. : Clary, K. H. 1997 Phyloge ny, не jer e ism and Bs of Yucca L. (Араз vaceae) as Inferred from Plant Morphology and Sequences of the Internal Tran- scribed Spacer (ITS) Region of the Nuclear om DNA. Ph.D. Thesis, University of Texas, Aus бөкөн P. L. & W. Н. Clark. 2000. E нен он rates and posed age-form classification. for the Ia Tree, Yucca brevifolia (Agavaceae). Haseltonia 46 Bull. a propos шоо G., M. Sousa & A. O. Chater (Editors). Flora M ue. sricana, Vol. 6 1994. ceae. Universidad Nacional Autónoma de Mé; co D.F; Missouri Botanical Garden Press, St. fans The Natural History Museum, London. Davis, D. R. 1967. A revision of the moths of the sub- family dp Rs I шш Incurvariidae). Bull. U.S. Natl. . 255: 1-17 —. 19 og 1 de Pp. Stehr (editor), i е 354—362 in К. W. Kendall- Hunt. a Immature Insects. The Monotrysian Heteroneura. Pp. 65-90 s "els on plants: Macroevolutionary 1 N. P. Kristensen (editor), Handbook of Zoology, IV Arthropoda: Insecta. Lepidoptera, Moths and Butter- flies, Vol. 1. Walter de Gruyter, Berlin, New York. 1992. Biology nus (i epidoptera: Prodoxidae). Smithsonian Contr. Zool. 524: 1-88. Després, L. 1999, Evolution of oviposition strategies ue speciation in the globeflower flies C hias- toc het spp. (Anthomyiidae). J. Evol. Biol. 12: 83 Pettex, V. Plaisance & F. Pompanon. 2002. Speciation in the slobeflower fly Chiastocheta spp. (Pus tera: Anthomyiidae) in relation to host plant species biogeography, and ensi ab Molec. Pelen, Evol. 22: 29 Dodd, К. J ОВ. Linhart. 1994. Reproductive con- sequences of interactions between Yucca glauca (Aga- vaceae) and Tegeticula к (Lepidoptera) in Col- orado. Amer. J. Bot. 81: 815-8 Eguiarte, L. Е. 1995. A a s) vs. Huber y Dahlgren (Asparagales): Análisis molec ee sobre la Ша. y evolucíon de la familia Agavaceae sensu utc о ur de las mcn о. Bol. Soc. Bot. MS Mura pun & — Silva- кузн 2000. ción dè la familia e : Filogenia, biología re- productiva y genétic 'a de S iones. Bol. Soc. Bot 31-150. Evolu- 1964. Butterflies PA plants: A study in coevolution. Evolution 18: 586—608. Eic hler ч D Som ne rules in ectoparasitism. i ag . Hist. :5 8. Egan n. "1980. бо Potential crop for aridlands. Pp. 115-123 in Yucca. Centro de a Aplic she Saltillo, Mexico. Engelmann, G. 1871. Yucca and Hesperaloé. Watson's Botany of the th Parallel. Report of the United States. Geological Exploration of the Fortieth Parallel = ce King, Geologist in Charge. V. Suppl.: 496- Investigación en —— ———. 1872. The ده‎ Yucca and its fertilization. Bull. Bot. Club 3: 3 1873. Notes on he genus Yucca. Trans. Acad. Se i rd 3: 17-54. Farrell. D. 1998. чина fondness” oe Why are du so many beetles? Science 281: 555—559 ; C. Mitter. 1993. Hm determinants of insec d UE diversity. Pp. 253-266 in R. E Ricklefs ‚ Schluter (editors), Species Diversity in Ecologic fi i ial Univ. Chicago Press, Chicago. Feist, M. A. 1995. Population Genetics of Yuccas: Their Mating System and its Consequences on Variation and tic Structure. M.S. Thesis, University of Cincin- olland. 1998. The evolution of obligate pollination mutualisms: Senita cactus and sen- ita moth. Oec . а 114: 368-375. Frack, D. С. 082. A Systematic Study Moths к. Prodoxinae) and their Hosts (Agava- ceae), g br ‘riptions of the Subfamilies of Adelidae of Prodoxine E lat.). Thesis, ome State Polytechnic Uni- UNDE Pons Californ Fuller, € ). 1990. Factors | Affec ‘ting the Balance of Co- PNE H dnd Conflict Between the Yucca Moth, Tege- ticula yuccasella, and its Mutualist, Yucca aS Ph.D. Thesis, University of New Mexico, Albuquerque. Volume 90, Number 1 Pellmyr y 53 Yuccas and Yucca Moths Galil, J. 1969. Intentional pollination in yucca flowers. Teva Va'aretz 11: 4—12 [in Hebrew]. García-Mendoza, A. . Una nueva especie de Yucca ти eae) de бад y Puebla, México. Acta Bot. . 42: 1-5. E H. S. 1982. The € SUE in Sonora. U.S.D.A. Serv. Agric. Handboo Gerarde, J. 1633. Plantes. Adam Norton and Richard Whitakers, London. — R. . J. N. Thompson, R. D. Holt, S. L. Nuis- . Hoc aa 2000. Hot spots, eol spots, id s ge ари mosaic theory of coevolution. Amer. Naturalist 156: 156-174. Great Plains a Association. 1977. Atlas of the Flora of the Great Plains. lowa State Univ. Press, Ames. Groman, J. D. 1999. dio sae и and Life His- tory С e nces of a Rec st Shift in the Bogus Yucca Moth P. me ا‎ (Prodoxidae). ` B. “The ssis, Vanderbilt University, Nashv ille, Tennes- s M. A. 1992. Dispersed Unidirectional Introgres- sion from ds schidigera into Yucca baccata (Agava- ceae). Ph.D. Thesis, Claremont Graduate School, Po- mona, ami erre, Е. A., С. A. Machado, E. Bermingham, J. | son, D. Wi idsor, S. S. McCafferty, W. таг ih & K. Bachmann. 1996. Molecular و‎ figs and their wasps. J. Biogeogr. 4: 521-530. o, J. & W. Yuille. 1990. The Papers of Charles Valen- tine Riley: A Register of his Papers in the National Agric cultural Library. U.S.D.A. Natl. Agric. Er Bibli- ographies and Literature of Agriculture No. Holland, J. N. & T. H. Fleming. 1999. Mutua a inter- actions ени n. Upiga virescens (Pyralidae), a Bona me seed-consumer, and Lophocereus schottii (Cacta ceae). Ecology “ж 2074—2084 a J. B. Rannala & B. Larget. 2000. Bayesian the analysis of cospeciation. Evolution 54: 352 Hulst, G. D. 1886. [Soci e news.] Entomol. Americana 2: 18 Huth, C. J. & O. Pellmyr. 1999, Yucca moth 775, дени апі pollination behavior is affected by past flower vis itors: Evidence for a host-marking pheromone. Oec kia gia 119: 503-509 & . Pollen-mediated selective abortion pido evolutionary stability of mutualism een yuccas and yucca moths. Ecology 81: 1100— 1107 Jaeger. N., F. Pompanon & L. Després. 2001. Variation in predation costs Chiastocheta egg number on Trollius europaeus: How many seeds to pay for pollina- tion? Ecol. gero 26: 56—02. Kingsolver, 984. dtd Biology of a Mutu- alistic la iE Yucc uca and Tegeticula yuc- casella. Ph.D. Tx seq: University of Kansas, Lawr- ence. Kozlov, M. L. 1996. Incurvariidae and Prodoxidae (Lep- idoptera) from Siberia and the Russian Far East, with descriptions of two new species. Entomol. Fenn. 7: 55- ө, E. E. 1880. Yuccas under cultivation. Nature 22: 606—607 Lopez- Vasmonde, C., J. Y. Rasplus, G. D. Weiblen & J. M. Cook. 2001. Molecular phylogenies of fig wasps: Partial cocladogenesis of pollinators and parasites. Mo- lec. Phylogenet. Evol. 21: 55-71. The Herball or QM Historie of Mac iw C. A., E. Jousselin, F. Kjellberg, S. G. Compton & E. linating wasps. Proc. Roy. Soc. London 268: 1-10 Kee, 5. 1994. Catalogue des Plantes Introduites t Cu уе vées en bg qu S T Histoire Naturelle, Par Marr, D. L., J. Leebens- Mack, L. Elms & O. Pellmyr. . Pollen pn т in Yucca filamentosa (Agava- ceae): The paradox of self-pollination behavior by Te- geticula yuccasella (Prodoxidae). Amer. J. Bot. 87: 670- 671. nie. Museum National ock & O. Pellmyr. 2001. mda 1 — Explori Coexistence of g the impact of cheaters on the yucca-yucca moth ade Oecologia 128: 454463 Massey, L. L. Hamrick. 1998. Genetic diversity and populos sisi of Yucca filamentosa (Agava- ceae). Amer. J. Bot. 85: 340—345. Matuda, E. & I. Pifia Lujan. 1980. Las Plantas Mexican- as del Genero Yucca. Serie Fernando de Alva Ixtlilxo- chitl, Collecion miscelanea estado de México, Toluca, México. Mayr, E. 1963. Animal m and Evolution. Harvard Univ. Press, Cambri cKelvey, S. D. 1938, 1947. Yuccas of the Southwestern Unita States, Vols. 1, 2. Arnold Arboretum, Jamaica Plai MeVaugh. R. 1983. Yucca. Pp. 281-283 in W. R. An- derson (general editor), Flora Novo-Galiciana, A De- scriptive Account of the Vascular Plants of Western Mexico, Vol. 15. Bromeliaceae to Dioscoreaceae. Univ. Michigan Herbarium, Ann Arbor. Meehan, T. 1876. On self- fertilization and cross-fertiliza- tion in flowers. Penn Monthly 7: 8 46. 1879. On the ion of Yucca. itil 1: 33-36. Miles, N. J. 1983. Variation in host specificity in the yucca moth, Tegeticula yuccasella (Incurvariidae): A morphometric approach. J. Lepidopterist’s Soc. 37: N. Amer. г, ell & B. Wiegmann. 1988. The phylo- genetic stud ^ of adaptive zones: Has phytophagy = moted plant diversification? Amer. Waitin 132: 107 128. Mohr, C. 1901. dee Life of Alabama. Contr. U.S Herb. 6. Washin 1959. Flora of Аена. Univ. Toronto Press, . Natl. Murawski, D. A. 1997. Moths come to light. Natl. Geogr. g. March: 40-57. "ond Н. 1874. Gegenseitige Abhängigkeit von Blumen und sie befruchtenden Insecten. Zoologische Garten (Frankfurt) in ИДЕ эе Nielsen, E. S. . Incurvariidae and Prodoxidae from the alea area ое Incurvarioidea). In- secta Matsumurana N.S. 26: 187-200. 1985. Primitive (non-ditrysian) Lepidoptera of the Andes: Diversity, distribution, biology and D genetic relationships. J. Res. Lepidoptera Suppl. 1 16. R. Davis. 1985. The first southern hemi- sphere prodoxid and the кн of the Incurvarioi- dea. Yes Entomol. 10: 3 Page, R. Maps o» trees and cladistic e of historical associations among genes, organ- isms, and areas. Syst. Biol. 43: 58-77. Annals of the Missouri Botanical Garden Pagel, M. 1998. ano Zool. S Script. 26: 331— The maximum likelihood approach to re- 48: construc diim ancestral character states. Syst. Biol. 612-622. 1999b. Inferring the historical ue rns of bio- logic al E Nature 401: 877—88 Palacios, I . Rzedowski. 1993. m palinológico de las b feles del mioceno inferior y principios del mioceno medio de la region de Pichucalco, Chiapas, Turc Bot. Mex. 24: 1—96 Pellmyr, The cost of mutasse Interactions between bna europaeus and its pollinating parasites. Oecologia 78: : -5 ——. 1% e phylogeny of a mutualism: Evolution and pedibus between Trollius and its see 'arasitic о Biol. J. Linn. Soc. 47: 337-365. — ———. 1997. Pollinating iced eaters: Why is active pol- jiniaiion so ae Ecology 78: 1655-1660 —— —. 1999. > оа revision of the yucca n. in the Tegetic p) orc complex north of Mex Syst. Entomol e 4: 243- bd: rende in.‏ س He. edie part Йога (‏ 182-1 :42 { "ur Pollination biology of (Agavaceae). Southw. Naturalist a = 4 ázar-Lara. 2000. Systematics of the yucca moth genus nj ad. a (Lepidoptera: Prodox- idae), with description of three Mexican Ann. n species. Entomol. Soc. Amer. 93: 432-439 Huth. 1994. Evolutionary stability of mu- tualism between yuccas and yucca moths. Nature 372: 251-200. & C.J . Krenn, 2002. Origin of a complex key innovation. in an obligate plant-insect mutualism. Proc. atl. Acad. Sci. U.S.A. 99: 5498—5502. — ——— & J. Leebens-Mack. 1999. Forty million years of mutualism: Evidence for Eocene origin of the yucca- yucca moth association. Proc. Natl. Acad. Sci. USA. 96: 9178-9183. р Ю. Reversal of mutualism as а mec ۰ for ада клк radiation in yucca A ng st 156: 562-576. N. Thompson. 1992. Multiple engins of mu- tualism in the yucca moth lineage. Proc. Natl. Acad Sci. U.S.A. 89: 2927-2929. =g (С moths. 1996a. Non-mutualistic yucca moths sun ш evolutionary consequences. ture 380: Veni ——, Thon E Brown & R. Harrison. 1996b. bus of porn ation and Moneta in the yucca moth lineage. Amer. Naturalist 148: 827-847. ———., L. Massey, M. A. Feist & & J. L. Hamrick. 1997. (ent consequences ¢ of spec ‘ialization: Yucca moth be- havior and self-pollination in yuccas. Oecologia 109: 273-278. Pfosser, M. & F. Speta. 1999. Phylogenetics of Hyacin- thac eae bue on plastid DNA sequences. Ann. Mis- 52-875. &. Biologic al interrelationships of Lr and Yucca ed Univ. Calif. Publ. Entomol. 100: 93. 992. M ea g yuccas and yucca moths, Trends Ecol. Eve 0-15 —— & R. A. Mackie. iologic ‘al interrelation- ships of mathe and Yucca whipplet (Lepidoptera: Gele- chiidae, рве шш Prodoxidae). Univ. Calif. Publ. Entomol. 42: 1— Inferring evolutionary processes from 349. cieli: d Howlands, P. G. 1978 Rzedow ski. Segraves, К. & О. Pellmyr. 2001. Sorensen, osa. S amd: J. А. Sutherland, S. 1 Price, Р. W. 1980. Evolutionary Biology of Parasites. Princeton Unix Press, Princeton, New J Rau, P. 1945. The yucca plant, Yucca filamentosa, and the yucca moth, Tegetic ula (Pronuba) yuccasella Riley: i ec e o-behavior study. Ann. Missouri Bot. Gard. | ersey. T E K . Weis. 1998. Inbreeding and out- crossing in A 10 bons Consequences for the repro- uctive success of plant and pollinator. Ecol. Let 21: Ў Riley, C. V. 1871. Third Annual Report on the Noxious, iiec E and Other Insects, of the State of Missouri. x. Jefferson City, Missouri. 1872 The fertilization of the yucca pes by Pronuba о Canad. Entomol. 4: — ———. 1873. On е mn of the yucca moth. dmn Naturalist : 619 — ——. 1877. Furt rther с. оп Ргопиба ко and on the pd of Yucca. Trans. Acad. Louis 3: 568—573. 1 — mu = ee) N g The true and bogus yucca moth, with re- marks on the pollination of Yucca. Amer. Entomol. 141-145. Further notes on the pollination of Yucca and o on Pomii and Prodoxus. Proc. Amer. Assoc. Ad Sci. : 617—6: „ ААУ, Mr. Tile s Re dcr on Pronuba yuc- Americana 2: 233-236. 9. Notes on Pronuba a Yucca pollination. 50-155. Proc. Taomil. Soc. Wash. 1 . 1891. Fertilization of Yucca in Australia. Insect Life 4: 74. 1892. The yucca moths and 2 pollination. Aamin Rep. Missouri Lu Gard. 3: 99—158. * Vegetation S namics of the Joshua Tree (Yucca Ее ifolia Engelm.) in the South- western United States of America. Ph.D. ersity of California, Riversic e. J. 19 Thesis, Uni- iversity and origins of the pum erogamic lon м Мех 0. Pp. 129-144 in T. P. Rama moorthy, e, A. Lot & J. Fa (editors), Biologic al Diversity af VU Origins and. Distribution. Oxford Univ. Press, New Yor Phylogeography of the yucca moth Tegeticula maculata: The role of historical biogeogra] shy in recone iling high genetic structure with limited spec Bs ug Molec. Ecol. 10: 1247-1253 Smith, E. H. The ve ai celebration of its own, ie PER 212-221. Soberon Mainero, J. & С Marines del Rio. 1985. Cheat- ing and Ens advantage i utualistic associatior Pp. 192-216 in D. H. Bo Pins puse The ени of Mutualism. Oxford Univ. Press, New Yor d 995. Brethren of the Net: American ои 1840-1880. Univ. Alabama as Tusca 1963. Flora of Missouri. Iowa State Jniv. Press, Ames. 1982. The Pollination Biology of Panicu- mportance of Male Fit- . Ph.D. Dissertation, University of Arizo- ate Agaves: Documenting t e ness in dian na, Tucs Thompson, T N. 1994. The Coevolutionary Process. Univ. Chicago Press, Chicago. 1999a. Specific Ee on the uis o mosaic of coevolution. Amer. t 153: SI-S 1999b. What we bus e ‘do not know tae Volume 90, Number 1 2003 Pellmyr 55 Yuccas and Yucca Moths coevolution: Insect herbivores and plants as a test case. Pp. 7-30 in Н. ОШ, V. К. Brown & R. Н. Drent (edi- tors), Herbivores: Between Plants and Predators. Black- well, Oxfor llmyr. 1992. Mutualism in е gna seed parasites amid co-pollinators: Constraints on spe- cialization. Ecology 73: 1780—1791. Tidwell, W. D. & L. R. Parker. 1990. Protoyucca shadi- shii gen. et sp. nov., an arborescent monocotyledon with secondary growth from the middle Miocene of north- western Nevada, USA. Rev. Palaeobot. Palynol. 62: 79— 95 Trelease, W. 1893. Further studies of yuccas and their pollination. Annual Rep. Missouri Bot. Gard. 4: 181— 226. The — Annual Rep. Missouri Bot. ES 13: ae + Trivers, R. L. 1971. The xiu of reciprocal altruism. Phi s & T. L. Burgess. 1995. Son- oran Dunt asia Ге Ecological Atlas. Univ. Arizona P ucso A. J. & J Е Addicott. 1993. Facultative non-mu- alis behaviour by an obligate mutualist-cheating by moths. Oecologia 94: се 175. uate peter ke ted States an The Last 40,000 Years of Biotic Change. Univ. Arizona Press, Tucson. Verhoek, S. 1998. Agavaceae. ^a 60—70 in K. Kubitzki (editor), The Families and Genera of Vascular Plants III: Flowering Plant, индин i Lilianae (except Orchidaceae). Springer, Berlin. Wagner, D. L. & J. A. Powell. 1988. A new Prodoxus from Yucca baccata: First report of a leaf-mining pre- doxine (Lepdioptera: Prodoxidae). Ann. Entomol. Soc Amer. 81: 547—553. Webb, В. Н. 1996. Grand Canyon: А Century of Change: Rephotography of the Am Stanton Expedition. niv. ee =, Tuc - Yuocas "d the Southwest. U.S.D.A. Entomol. 2 299— ——4 € - Bos. 2002. Speciation in = pollinators and d Molec. Ecol. 11: 1573-1578 Whitten, J. B. 1894 e emergence of Pronuba from yueca capsules. Ann. Rep. Missouri Bot. Gard. 5: 137— 138. Wiegmann, ., C. Mitter & B. Farrell. 1993. Diver- sification s carnivorous parasitic insects—Extraordi- nary radiation or specialized dead end. Amer. Naturalist 142: 737-754. Wilson, R. D. & J. F. Addicott. 1998. Regulation of mu- tualism between yuccas and yucca moths: Is oviposition behavior responsive to selective abscission of flowers? Oikos 81: 109-118 Zeller, P. C. 1875. Beitrüge zur Kenntniss der nordam icanischen Nachtfalter besonders der “шна а-ай en, Dritte E Verh. Zool.-Bot. Ges. Wien 25 207-360, pls. 8-10 Ziv, Y. & J. L. Боо. 1996. Infertile seeds of Yucca schottii: A beneficial role for the plant in the yucca- yucca moth mutualism? Evol. Ecol. 10: 63—76 Internet Resource Pellmyr, 0. 2002. Prodoxidae. The Tree. of Life Web 15 N ж йн. CHROMOSOME REPORTS FROM SOUTH AMERICAN HYPOCHAERIS (ASTERACEAE)! Hanna Weiss,” Jürke Grau, and Carlos M. Baeza‘ Tod F. Stuessy,? ABSTRACT Eighty-nine new chromosome counts are reported from 85 populations of 15 species of Hypochaeris (Asteraceae, Lactuceae) from ben America, including first reports from nine sc roides, 2. LI n H. tenuifolia, H. d Most are 2n = 8 w ly 16) is documented for the scorzonerae, H. s ‚Н. sonch asymmetric i s ves. Tetraploic H . scorzonerae, and H. tenuifolia. js summary of previous e n species are known (H. incana and H. stenocephala). Five ta meyeniana, H. scorzonerae, H. sessiliflora, and karyotypic diffe тепе p among Түз taxa (2n = many species of the New World (c: from European ancestors rather iu the reverse. It a into many h: me chromosomal change. The absence of aneuploidy and relatively low frequency of polyploidy, plus few known st a weakened role for hybridization as an ev present interspecific hybridization, sugge group. Data so far suggest diploid recombinational speciation a than dramatic cytological alterations, pli Key words: H. чү а Dien. therefore, occurs in abou American taxa. The South Amer rican representatives » Par ge Ж аге all based on x = 4 in contrast t )pears that abitats from sea level to over 3000 m, but that this radiation has not often been accompanied by conspicuous f . apargioides, H. O NIE, palustris, ı bimodal and rst time in some E 15 of H meyeniana, ew counts reveals that ¢ 1 also show infraspecific a loi 10% of the South This pattern of greater cytologic al MT MN among erivec ypochaeris in S America radiated explosively cases o volutionary mechanism in the s the dominant mode accompanied by genetic rather 1s occasional polyploid de 'rivatives. Asteraceae, chromosome numbers, Hypochaeris, karyotype, South America, speciation. Although much is known about mechanisms of speciation and species-level biogeography in the floras of Europe and North America (e.g., Stebbins, 971; Stebbins et al., 1953; Grant, 1981), little is understood about these same processes in the flora of South America. To help reveal these dynamics in the temperate regions of southern South Ameri- ca, we have begun a series of investigations on Hy- pochaeris (Asteraceae, Lactuceae; Hypochoerts L., orthogr. var.). This is a genus of approximately 60 species, with 9 in Europe (DeFillips, 1976) and ca. 50 confined to South. America, many of which are localized in the Southern Cone (Bortiri, 1999). Hypochaeris is a good genus to extract information on modes of speciation and patterns of biogeogra- phy because of the large size and small number of chromosomes (Stebbins, 1971). phologies of the taxa varying from small acaules- cent alpine forms to broad-leaved large-headed species, and distributions in very different ecolog- ical zones ranging from sea level to over 3000 m (Lack, 1979). Apparent rapid and recent speciation the diverse. mor- of the genus in South America, therefore, combined with favorable karyological aspects, offers a usefu system in which to examine evolutionary processes. Chromosome counts as well as detailed karyo- logical studies have been investigated for European species of Hypochaeris (Parker, 1976; Mugnier & Siljak-Yakovlev, 1987; Barghi et al., 1989; Siljak- Yakovlev et al., 1994; Cerbah et al., 1995, 1998a), which appear to possess symmetrical karyotypes, and which show general correlations of different chromosome numbers (2n — 6. 8, 10, 12) with tax- onomic Molecular phylogenetic studies using ITS regions among these European taxa (Cer- bah et al., 1998b), and including four species from South America, have also revealed useful insights on broad evolutionary patterns within the genus that. correlate, to some considerable degree, with sections. previous sectional limits. Despite the large number of species of Hypo- chaeris from South America, they have been poorly studied karyologically. Chromosome numbers, sometimes accompanied by photographs, have been ' The study was supported by FWF project No. P13055-BIO to T. F. Stue epartment of Higher Plant Systematics and Evolution, Institute of Bonny University of Vienna, Rennweg 14, ? De A-1030 Vienna, Aasia; tod.stuessy@univie.ac.at. * Institute for Systematic Botany and Botanical Garden, Germa ANN. Missouni Bor. University of Munich, Menzinger Strasse 67, 80638 Munich, any. ' Departamento de Botánica, Universidad de Concepción, Casilla 160-C, Concepción, Chile. GARD. 90: 56—63. 2003. Volume 90, Number 1 2003 Weiss et al. 57 Hypochaeris Chromosome Reports Table 1. List of previous chromosome counts of Hypochaeris from South America (excluding H. glabra and H. radicata). Species 2n n Reference H. acaulis (Remy) Britton 4 Wulff (1998) H. arenaria Gaudich. 8 Moore (1981) H. brasiliensis (Less.) Griseb.! 8 Stebbins et al. (1953); Parker ec Ruas et al. (1995) 4 Coleman (1968); Turner et al. (1979) Н. chillensis (Kunth) Hieron. 8 Tomb et al. (1978); Siljak-Yakovlev et al. (1994); Cerbah et al. (1995); Cerbah et al. (1998a) 4 Wulff (1992, 1998) H. chillensis X H. megapotamica 4 Wulff (1992 H. chillensis X H. microcephala var. 4 Wulff (1992) albiflora H. chondrilloides (A. Gray) Cabrera 4 Cherubini (1982) 8 Wulff (19€ H. elata (Wedd.) Griseb. 8 Krapovickas (1951) 4 Bernardello (1986) H. foliosa ( pue ) Reiche 8 Stebbins et al. (1953) Н. gardner 5 Coleman (1968) H. glauca (Р этү Reiche? 8 Stebbins et al. (1953) H. halophila (Hauman) Cabrera? 8 Schnack & Covas (1947) H. incana (Hook. f. & Arn.) Macloskie 8 Moore (1981) H. megapotamica Cabrera 8 Saez ee Siljak- Yakovlev et al. (1994); Cerbah et al. (1995); Ruas et al. (1995); Cerbah et al. (1998a) 4 Wulff (1992, 1998) H. meyeniana (Walp.) Griseb. 8 Diers (1961) H. microcephala (Sch. Bip.) Cabrera var. 8 Saez (1949); Siljak-Yakovlev et al. (1994); Cerbah et al. albiflora (Kuntze) Cabrera (1995); Ruas et al. (1995); Cerbah et al. (1998a); Wulff (1998) 4 Wulff (1992) H. pampasica Cabrera 8 Cerbah et al. (1995); Ruas et al. (1995); Cerbah et al. (1998a) H. parvifolia H. Kost. 8 Diers (1961) H. petiolaris (Hook. f. & Arn.) Griseb. 1 Гигпег et al. (1979) H. rosengurttii Cabrera 8 Ruas et al. (1995) H. sessiliflora Kunth 8 Olsen (1980) 6 Turner et al. (1967) 4 Jansen & Stuessy (1980) H. stenocephala (A. Gray ex Wedd.) 16 Stebbins et al. (1953); Diers (1961) H. taraxacoides (Walp.) Benth. & Hook. 8 Parker (1971) [3 H. tweediei (Hook. f. & Arn.) Cabrera! 8 Saez (1949) H. variegata (Lam.) Baker 8 Ruas et al. (1995) '3 Treated as synonyms of H. chillensis', H. chondrilloides?, and H. stenocephala? by Bortiri (1999). published for 23 species (e.g., Saez, 1949; Wulff, ies on Hypochaeris, therefore, the purposes of this 1992: Ruas et al., 1995; Cerbah et al., 1998a; Ta- paper are to: (1) summarize all available previous ble 1), but detailed karyotypes exist for far fewer ^ chromosome counts for South American species of taxa (only ca. 6). Most chromosome numbers so far Hypochaeris; (2) report original counts for taxa col- documented are 2n — 8 with un two cases of lected in Chile and other Andean regions (Argen- uniform tetraploidy known (H. i ana, Moore, tina, Bolivia, Ecuador, Peru, and Venezuela); and 1981; H. stenocephala, Stebbins et Ed 1953; Diers, (3) comment on the chromosomal variation docu- 1961), and two cases of infraspecific tetraploidy mented within South America, particularly with ref- (i.e., H. chrondrilloides, Wulff, 1998; H. sessiliflora, erence to the already well known cytological pic- Olsen, 1980, both 2n and 4n reports). ture of European species. Several populations of To set the stage for additional evolutionary stud- each taxon have often been analyzed so that infra- 58 Annals of the Missouri Botanical Garden Table 2. New chromosome counts of Hypochaeris from South America [all plants from Chile (regions given by Roman numerals) unless indicated otherwise]. Counts are mostly from mitotic preparations; haploid counts (meiosis I and 1“ pollen mitosis) are indicated by *; populations for which both diploid and haploid counts were made are marked Mi ** Abbreviations of collectors: CB = C. Baeza; DC = D. Crawford; JG = J. Grau; PS = P. Stuessy; TS = T. Stuessy; HF = H. Foérther: MW = M. Weigend. Vouchers on deposit in WU and CONC (collections by TS and id аш» and López) and M (all other collectione) ++ = first report(s) for taxon; + = new chromosomal level - 3 axon, voucher, and chromosome number H. acaulis (Remy) Britton 2n = 8 VII. Prov. Talca, Laguna del Maule, Laguna sin Puerto, s.d., JG Hyp-50. УШ. Termas de Chillán, Valle de las ides 15 Jan. 1999, TS et al. 15565. IX. 20 km E of Chilean Aduana toward Paso ed "a hado, 20 Jan. 1999, TS & € 15587. ARGENTINA. 1 km E of Paso Pino Hachado, 20 Jan. 1999, TS & CB 1 ++H. apargioides Hook. f. & Am. 2n = ҮШ. Parque Nacional I vom La Laja, 25 Jan. 1998, TS & DC 15470; Las Trancas, 30 Jan. 1998, TS 15481, 15485, 13 Jan. 1999, TS et al. 15. ‚ Valle Hermoso, 15551, 15554. IX. Reserva Nacional Malalcahuello, és Feb. 1998, TS et al. 15509% 3 km w p kon 13 Feb. 1998, 15513; 10 km E of Puente Lonquimay, TS et al. 15515*, 17 km » bs Pbi nte e 15516, 29 km E of pops о 15517, 8 km E of Chüléan das 13 Feb. 1998, З km X Lonquimay, 19 Jan. 1999, TS & CB 15576*, 8 d: E of Chilean Aduana, 20 Jan. 1999, 15583, rs Rs of E ilean Aduana, 15594**, 25 E of dna 15595**, 17 Feb. 2000, 15596, Volcán Llaima, 21 Jan. 1999, 15602*, Volcán Villarrica, Piedra de Aquila, 22 Jan. 1999, 15612*. to Laguna Huinfiucá, 23 Jan. 1999, 15621*, rd to Lonquimay, 16 Feb. 2000, /58094. + +H. clarionoides (Remy) Reiche 2n = 8 Región Metropolitana de Santiago, 6 km W of La Parva, 18 Feb. 1998, TS & PS 15527, 2 km W of La Parva, 15529*. 15531* H. elata (Wedd.) Griseb. 2n = 8 BOLIVIA. La Paz, Puerto Pérez, Lake Titicaca, s.d., Karus Hyp-43. +H. meyeniana (Walp.) Griseb. 2n = 8 PERU. Dept. Cajamarca, Prov. Contumazá, Bosque de Cachil, s.d., MW 98/554. +H. meyeniana (Walp.) Griseb. +2n = 16 PERU. Dept. Tacna, Prov. Taranta, S of Volcán Tutupaca, 20 km above Candarave, s.d., MW & HF 97/688. + +H. palustris (Phil.) De Wild. 2n = 8 p de Las d Termas de Chillan, 15 Jan. 1999, TS et al. 15566*. IX. 20 km E of Chilean Aduana, 20 У ¢ ( ҮШ. Va Јап. 1999, СВ 15568*, Volcán Villarrica, Ж + 1999, 15606, 15607, rd to Laguna Huinfiucá, 23 Jan. 1999, 15622, oe ‘ап Casablanca, e Jan. 1999, 15628, H. radicata L. 2n = VIII. Cerro Ponpón, eh jon 1998, ТУ 15450H, 15450], Parque ias Concepción, 15451; 2 km W of Angol, 24 Jan. 1998, TS & DC 54, 2 km into Parque Nacional Nahuelbuta, 15458*, Parque Nacional Nahuelbuta, 15461, Piedra ч Aguila, s p^ 17 km ; -A е 25 Јап. 1998, Й 5468, 15469*, Parque Nacional Laguna La Laja, ‚ 15473; Lota, Se Bek 27 Jan. ‚ TS et al. 15477, Termas de Chillán, 31 Jan. 1998, 15487. IX. 12 km E of б. 'autín, 12 Feb. 1998, ТУ et i ue 5 km 5 of Lonquimay, 12 Feb. 1998, 75508, 8 km E of Chilean Aduana, 13 Feb. 1998, 15519; 14 km E of Cherquenco, 21 Jan. 1999, TS & CB 15597*, | km W of Pucón toward Volcán Villarrica, 22 Jan. 1999, 15604, 11 km SE of Aguas Calientes, 24 Jan. 1999, 15626*. Región Metropolitana de Santiago, Santuario de la Naturaleza Yerba Loca, 18 Feb. 1998, TS & PS 15533. + +H. scorzonerae (DC.) F. Muell. 2n = 8 Region IV. 5 km E of Huentelauquén, 16 Oct. 1999, Lopez 503, 3 km N of Puerto Oscuro, 505, 6 km N of Puerto Oscuro, 507. t +H. scorzonerae (DC.) F. Muell. 2n = 16 V. Prov. Quillota, Cerro campana, Mirador de los Guanacos, s.d., JG 87-29. H. sessiliflora Kunth 2n = 8 ECUADOR. Quito, Paso Guamani, s.d., JC Hyp-37; Carchi, Páramo El Angel, 38, Bolívar, Chimborazo, s.d., MW ‹ + +H. setosa (Wedd.) Rusby 2n = 8 VENEZUELA. Mérida, Parque Nacional Sierra Nevada, s.d., Gaviria 1048. Volume 90, Number 1 2003 Weiss et al. Hypochaeris Chromosome Reports Table 2. Continued. Taxon, voucher, and chromosome number ++H. sonchoides Kunth 2n = 8 ECUADOR. Pichincha, Nono, NW of Quito, s.d., JG Hyp-36. + +H. spathulata (Remy) Reiche 2n = 8 IX. N of Pucatrihue, 25 Jan. 1999, TS & CB 15633*. H. stenocephala (A. Gray ex Wedd.) Kuntze 2n = 16 PERU. Dept. Puno, Prov. Puno, Ruins of Sillistani, s.d., ++H. tenuifolia (Hook. f. & Arn.) Griseb. 2n = 8 MW & HF 97/119. VII. Prov. Talca, Laguna del Maule, s.d., JG s.n. ҮШ. Prov. Nuble, Nevados de pei ok agen del Diablo, s.d., 492C, JG Hyp-45; Termas de Chillán, 31 Jan. 1998, TS et al. 15486, 15490A, 1 Feb. 1998, 15498, 15505. IX. Volcán Lonquimay, 19 Jan. 1999, Т5 & CB 15577-3*; Volcán Llaima, 21 Jan. 1999, rm Volcán Villarrica, 22 Jan. 1 5605 + +H. tenuifolia (Hook. f. & Arn.) Griseb. — 2n = 16 VIII. Termas pe hin 31 Jan. 1998, TS et al. 15489. 1998, TS et al. + +H. thrincioides 2n = 8 TM Remy) Reiche VIII. Cerro Ponpón near Concepción, 21 Jan. 1998, TS 1998, TS & DC 15456** IX. 5 km into Reserva Nacional Malalcahuello, 12 Feb. 154501. IX. 10 km W into Parque Nahuelbuta, 24 Jan. specific euploidy or dysploidy might be revealed and better interpreted. It must also be stressed that the present taxon- omy and nomenclature of Hypochaeris in South America needs comprehensive revision. Despite the very helpful treatments on Argentinean species by Cabrera (1971, 1974, 1976, 1978) and more re- cently Bortiri (1999), there still remains confusion in proper limits and correct names for many taxa. Future studies will have to resolve these important issues. MATERIALS AND METHODS MATERIALS Bud and seed materials of Hypochaeris were col- lected in South America (Table 2). Collections were of populational samples in all instances. Vouchers are on deposit at CONC, M, and WU. METHODS For mitotic chromosome counts, root meristems were obtained from surface-sterilized seeds germi- nated for two days on wet filter paper or from ma- ture plants in cultivation at M. Root tips were pretreated with 0.1% colchicine for 2 hrs at room temperature in darkness, fixed in 3:1 (ethanol : acetic acid) for 24 hr., and stored until use at —20°C. For meiotic flower buds were fixed in modified Carnoy’s solution (4:3:1; chloroform: absolute ethanol : glacial acetic chromosome counts acid), transferred to 70% ethanol, and stored at 4°C until examination in the laboratory. Chromosome preparations were made using Feulgen staining with Schiffs reagent following standard methods (Fukui & Nakayama, 1996). tips and flower buds were washed in distilled water to remove the fixative, hydrolyzed in 5N НСІ for 30 min at 20°С, and washed and stained with Schiff's reagent in darkness for 1—2 hr. Squash preparations oot were made in a drop of 45% acetic acid. After cov- bie; removal on dry ice, preparations were dried г 24 hr. at 37°С and mounted in some numbers were determined from at least 20 romo- cells of at least two florets for meiotic counts and of at least three seedlings for mitotic observations. RESULTS AND DISCUSSION Eighty-nine new chromosome counts are report- ed from 85 populations of 15 species of Hypochaer- is from South America (Table 2) including first re- ports for nine taxa: H. apargioides, H. clarionoides, H. palustris, H. scorzonerae, H. setosa, H. sonchoi- des, H. spathulata, H. ior ur and H. thrincioi- des. All counts give = 8, with infraspecific tet- raploidy (2n = 4x = 16) also documented in H. scorzonerae and H. tenuifolia. A new tetraploid pop- ulation has also been detected in H. meyeniana. These new data, in correlation with previously published reports for Hypochaeris in South America (Table 1), yield 32 species (of ca. 50) now counted from 138 populations. A brief combined analysis Annals of the Missouri Botanical Garden г ~ ` -—" оф г сь ТИ = 7 1 — |9 „% |= 4 be 4 $ ae ^ * * 4 | 5% a E Т / / i” 4 "e \ C 4 "^ P Ф Li 7 Фә j e 7 ~ ||8 — [9 д ч - } 2 d + d i S ; E " 11 2 0 پگ‎ кк | ME „ % چ‎ "id ` d ө |р" 13 —— || 14 ___|| 15 a Volume 90, Number 1 003 Weiss et a 61 Hypochaeris Chromosome Reports shows 25 uniformly diploid (2n — 8) species, two species uniformly tetraploid (2n = 16; H. incana and H. stenocephala), and five species with both diploid and tetraploid cytotypes (H. chondrilloides, H. meyeniana, H. scorzonerae, H. sessiliflora, and H. tenuifolia). One species, H. gardneri, has been reported previously as n = 5 (Coleman, 1968), and another, H. pela has been documented as л — 6 (Turner et al., Because of the = inir in chromosome levels among South American species of Hypo- chaeris (all n = 8), further comments on these two deviating reports are in order. The n = 6 report (Turner et al., 1967) for H. sessilifolia con- trasts with more typical n = 4 (Jansen & Stuessy, 1980) or n = 8 (Olsen, 1980). Examination of the voucher of H. sessilifolia reported by Turner et al. (1967; Wurdack 437, X) with camera lucida drawing of meiotic bivalents attached, suggests that n = 4 is probable, with some homologous chro- 4orn = mosomes being pulled apart in metaphase l/early anaphase I earlier than the others. Turner, in fact, wrote in 1977 (handwritten note on the voucher): “In hindsight and with more thought this could be n = АП!” The voucher for the count of n = 5 for H. gardneri has not yet been located despite an herbarium search. It is worth mentioning that the European Н. glabra with п = 5 is also known to be adventive in South America (e.g., Matthei, 1995; Bortiri, 1999). Because of the morphological vari- ability of taxa of Hypochaeris, and hence difficulties with identification, an examination of the voucher for H. gardneri will be essential to reveal whether this represents a new aneuploid level in native spe- cies of the continent. Karyotypes of all newly analyzed South Ameri- can species of Hypochaeris are bimodal and asym- metric and similar in overall morphology to kar- yotypes of species analyzed previously (Figs. 1—15 see also references in Table 1). In general, they consist of two large and two small chromosome pairs. Two of these pairs, one large subtelocentric and one smaller acrocentric, appear to bear satel- lites. The other two pairs are most often acrocentric. Despite this general uniformity of karyotype and a reasonably stable chromosome number (2n = 8), consistent karyological differences among taxa do exist. The main differentiation of karyotype con- cerns chromosome size changes and the presence of satellites (Figs. 1—6, 8—15). A detailed analysis of karyotypes of South American Hypochaeris and their evolutionary importance will be reported else- where. Although the emphasis in this paper is on new cytological reports of native South American spe- cies of Hypochaeris, additional reports for the in- troduced H. radicata are also included. This taxon is abundant in Chile (Matthei, 1995; pers. obs.), growing from sea level to over 2000 m and often found intermixed with native species. Because of the possibility of hybridization between H. radicata and native congeners, which could confuse inter- pretations of patterns and processes of evolution, sampling of this taxon was also included (Table 2). Hesults show the typical chromosome level (2n — 8; Fig. 7) for this species and symmetric karyotype and no irregularities, effectively excluding hybrid- ization. Judging by chromosome number, the other weedy species of the genus, H. glabra, (distinctive with 2n = 10, Stebbins et al., 1953), is much less common in South America, at least within Chile, and no representatives of this species have been examined cytologically during this study. Tetraploidy (2n = 4x = 16) has been previously reported for four South American Hypochaeris spe- cies: H. chondrilloides (Wulff, 1998), H. incana (Moore, 1981), H. sessiliflora (Olsen, 1980), and H. ies rur (Stebbins et al., 1953; Diers, 1961). . chondrilloides and H. sessiliflora, both ploidy jen i (2x and 4x) were reported. The present paper adds three more species, Н. meyeniana (Fig. 5), Н. scorzonerae, and H. tenuifolia (Figs. 13, 14), in which infraspecific polyploid cytotypes are known. Karyological data provide some suggestions on the mode of origin of these tetraploid cytotypes. The 4x races in H. tenuifolia and H. stenocephala possess karyotypes consisting of four equal-sized sets of chromosomes. It is probable, therefore, that ies polyploids are of autopolyploid origin. We suspec this to be the case also in H. sessiliflora (Fig. 3) and perhaps also for H. chondrilloides that contain both 2x and 4x cytotypes, although no detailed kar- yotypes are available for the latter. Understanding the origin of the apparently uniformly tetraploid e Figures 1-15. Mitotic chromosomes of South American species of Hypochae eris. Scale bar = 5 jum. —3. Н. ө TS & PS TS et al. 15565.—2. H. apargioides, TS 15465 H. пора MW 98/554. —6. Н. palustris, TS et al. 15566 Hyp-. stenoc ۶ la, thrincioides, TS 15450 8.—9. H. setosa, JGa 1048.—10. H. sonchoides, JG Be MW & HF 97/119.—13. H. tenuifolia, TS & CB ip Pai H. tenuifolia, TS et al. Forme. * — 1. Н. acaulis, 5527.—4. Н. elata, КК Нур-43.—5. же: radicata, TS & PS 15533.—38. H. к» i 6.—11. H. spathulata, TS & CB 15633.—12. Annals of the Missouri Botanical Garden Hypochaeris stenocephala and possibly H. incana (only one count known) will be important for doc- umenting evolutionary mechanisms in the group. ecause of the close morphological and karyo- logical similarity and evolutionary relatedness of South American species of Hypochaeris, special at- tention was given to searching for meiotic irregu- larities that might signal hybridization. Meiotic counts showed exclusively four regular bivalents in all examined populations. Previous reports on mei- otic chromosome numbers have shown similar re- sults (Wulff, 1992, 1998). Only experimentally ob- tained hybrids between H. and H. megapotamica showed some irregularities in ho- mologous chromosome pairing with 4П and occa- sionally ЗИ + 21 (Wulff, 1992). Presumptive nat- ural hybrids of H. chillensis and H. microcephala var. albiflora, however, showed regular bivalent for- mation suggesting the possibility of recent origin of chillensis these species and high degree of karyotype simi- larity (Wulff, 1992). Two instances in our own field collections suggested possible hybridization based upon morphological features: Н. apargioides Х tenuifolia (TS et al. 15554) and H. palustris X Н. tenuifolia (Т5 & CB 15607). Because some differ- ences in chromosome size and presence of satellites are obvious in the case of H. apargioides (Fig. 2) and H. tenuifolia (Fig. 13). hybrid individuals be- tween these taxa should be detectable. These po- tential hybrids were found to be diploid, however. with no differences in size of chromosomes of the two haploid sets. Based on karyomorphology. pos- sible hybrid origin of these populations is unsup- ported. In contrast to broad cytological diversity among nine European species of Hypochaeris (2n. = 10, 12; Mugnier & Siljak-Yakovlev, 1987; Cirbah et al., 1998a), cytological uniformity of the New World members of the genus suggests several as- pects regarding evolution of the group. First, be- cause the karyotype of South American taxa rep- resents only one general pattern, in contrast to several found among European species, it can be hypothesized that the former evolved from out of the latter. This hypothesis is also corroborated by recent molecular phylogenetic studies of nuclear (ITS; Cerbah et al., 1998b; R. Samuel et al.. prep.) and chloroplast (trnL; R. Samuel et al., in prep.) genes. of South different growth forms and occurring in so many diverse hab- Second, the abundance American laxa representing s ~ | many itats is strongly indicative of rapid and recent adap- tive radiation. Third, the mechanisms of speciation accompanying this explosive evolution have clearly not been driven by gross macro-cytological alter- ations, such as dysploidy, euploidy, and marked karyotypic change. Hypochaeris represents a genus, therefore, in which perhaps more minor amounts of karyotypic, and certainly genetic, change has ac- companied speciation. Because of this situation and due to the large and few chromosomes, the genus in South America provides an excellent opportunity to map the genome and determine minor karyotypic changes during speciation, within context of the relatively stable 2n 8 karyotype, as has been done successfully for Helianthus in North America (Rieseberg et al., 1995; Rieseberg, 2001). Literature Cited Barghi, N., С. Mi ugnier & S. Siljak-Yakovlev. 1989. Kar- yological amalesi in some А ‘haeris species from Si- cily. Pl. Syst. Evol. 168: 49-57 Bernardello, L. M 1986. Nine aros cromosómicos en As- teraceae de Córdoba. Darwiniana 27: 16—178. Bortiri, Е. 1999, 280. Asteraceae, parte 14. Tribu XIII. cde 'eae: i "hoeris. Flora Fanerogámica Argentina, Fasc. 63: 1-2! rers A. L. 197 Hypochoeris. Pp. 397—409 in M. N. n (editor), Flora Patagónica, Parte VII. Composi- uo Buenos Aires. 374. Hypochoeris. 512-525 in A. Burkart nro “ы Ilustrada de Ente Rios (Argentina), Parte VI: Dicotiledoneas Metaclamideas (Gamopétalas). B: Rubiales, Cucurbitales, ee (Incluso Com- ees INTA, Buenos jns — ———. 1976. Materiales para una revisión del género Нурос hoeris : Hypo herr chilensis (H. B. K.) Hieron. Darwiniana 2 2-i 1978. Рейт Pp. 671-686 in А. L. hu (editor), Flora de La Prane ча de Jujuy, DUAE x. M enter INTA, Bu s Aire erba . J. Coulaud, В. Godelle & 5. Siljak-Yakovlev. i к anome size, fluorochrome banding, and karyo- type evolution in some Hypochoeris species. Genome 38: 689-095. — ———, J. €oulaud & S. Siljak-Yakovlev. 1998a. rDNA organization and evolutionary lie ы in the genus Hypochaeris bine en eae e s leredily 89: -3 — ——, Т. Souza-Chies, Jibier, B. is ejeune 24 5 Siljak- Yakovlev. p ate TM phylogeny of the T genus Hypochaeris using id transcribed spacers of nuclear rDNA: Inference 'é for chromosomal evolution. olec. Biol. Evol. 15 “354, d G. C. (Editor) 1997. El Altiplano: Ciencia y Conciencia en los Andes. Univ. de Chile, Santiago. Cherubini. C. 1982. Nümero de cromosomas de ا‎ espermatofitas de la flora Mendocina. Revista Fac. Ci. Agrar. Univ. Nac. Cuyo 22: 23-25. Coleman, J. R. 1968. Саша numbers in some Вга- zilian С Wo Rhodora 70: 228—240. DeFillips R. A. 1976 als 308—310 in T. T . D. M. Moore & p. p Webb (editors, 4. Cambridge Univ. Press, Cam- Hy en hoeris. . Hey ipn . Bu D. H. Valentine ‚25. М. Кога Боа, Vol. bridge. Diers, L. 1961. tationsgürteln 37—18, CUM Der Anteil an Polyploiden in den з der westkordillere Perus. Z. — pe A e] = 37— Volume 90, Number 1 2003 Weiss et al. 63 Hypochaeris Chromosome Reports ‚ К. & S. Nakayama (Editors). 1996. Plant Chro- p тоин) Methods. СКС Press, Boca Raton. Grant, V. Plant Speciation, 2nd ed. Columbia Univ. hon. м Ча. Jansen, R. К. & Т. F. Stuessy. 1980. Chromosome counts of Compositae ftom Latin America. Amer. J. Bot. 67: Krapovic kas, A. 1951. Nümeros cromosómicos de tres гоа riojanas. Bol. Soc. Argentina Bot. 4: 105— "s "н. W. 1979. The subtribe Hypochoeridinae (As- terac eae, La the ctuceae) in the tropics anc Southern 6 in K. Larsen & L.-B. Holm- germ Fs itors), Tropical Botany. Academic Press, Londor Matthei, Manual de Las Malezas que Crecen en Chile. Pili by the author, Santiago. 1 Moore, D. 8l. pa enc numbers of r angiosperms. ry Soc. Brot., Ser. 2, 53: 995- г 5. Siljak- “Yakovlev, 1987. ss Yugoslavian populations of Hypochoeris (Compositae). Cainea 40: 319-32 Olsen, J. 1980. /n: Chromosome numbers reports LXVII. T. 36 The Control of Recombination. Ph.D. is, Universi of des 1 e B-chromosome system of Hypoc Pp maculata. Я В- iui, meiotic D and i heritance. Chromosoma (Be Rieseberg, L. Н. 2001. Chromosomal и S and speciation. Trends Ecol. Evol. 16. 35 m un , С. R. Linder & G. J. т 1995. Chromosomal nic barriers to introgression in qn lianthus. Ge- Lus 141: 1163-1171. , P. M. Ruas, N. I. Matzenbacher, G. Ross, C. Bernini & A. L. L. Vanzela. 1995. Cytogenetic studies of some Hypochoeris species (Compositae) from Brazil. 9-375. 1949. Estudio citológico comparativo de al- gunas especies del género Hypochoeris (Compositae) de la America del Sur. Lilloa 19: 97—104. 5с 'hnack, B. & G. Covas. Ne Estudios cariológicos en antofitas. Haumania 1: Siljak- Yakovlev, S., e eh oli, E Roitman, N. Barghi & Mugnier. 1994. Etude caryplogique de trois especes d'Hypochoeris originaires d'Argenti microcephala var. albiflora et H. megapotamica. Canad. J. Bot. 72: 1496-1502. Stebbins, С. L. 1971. Chromosomal Evolution in Higher Plants. we Arnold, London A. Jenkins & M. Walters. 1953. Chromosomes and wee in the Compositae, tribe Cichorieae. Univ. ae Bot. 26: 401—430. Tomb, A. © „ Chambers, D. W. Kyhos, A. M. Powell | и еп. 1978. Chromosome numbers іп the Ma. XIV. Lactuceae. Amer. J. Bot. 65: 717- 721. Turner, B. L., А. M. Powell & J. Cuatrecasas. 1967. Chro- mosome numbers in Compositae. XI. Peruvian species. Ann. Mis —— а pes Urbatsch & B. Simpson. 1979 Chromosome mmber in South American мей trae J. Ae 1992, tes tee natural entre especies american ж Hypochoeris (Asteraceae). Darwini- an us 17 - Eotudion cariológicos en Asteraceae. VIII. Darwiniana 35: 374: INVASION BIOLOGY: AN EMERGING FIELD OF STUDY! Sarah Hayden Reichard? and Peter S. White? ABSTRACT Biologic 'al invasions are increasingly recognized as a key problem for the conservation of LAN al diversity. However, rec ognition that some spec ies, at least the writings of Charles I the scientific r species, goes back ) wrote the first book attempting to desc cribe the biology к invasive organisms. when ie 8, шу their native range, caus a decline in ا‎ 1950s inb: Bri tei biologist Charles Elton, It was not until t s, however, that This emergence has resulted se two forces: the . In the > development of the : scientific e for invasion biology based: on a substantial and accumulating literature, and the urgency of the ir Key words: nvasive spec ies ue because inc moaned world trade and travel are increasing the frequency of invasions. invasion bon scientific history, Systematics Symposium. THE HISTORICAL CONTEXT Over the last 20 years biological invasions have gained growing attention from ecologists. From the rapid increase in both scientific and popular arti- cles and books written about invasions one might conclude that invasions are a relatively new phe- - Yomenon. In fact, as humans first began to move around the earth, they took familiar plants and an- imals with them for use as food, medicine, or tech- nology (Fritz, 1994). We know that maize was found in eastern North America when European explorers first arrived in the New World, yet this species is known to have originated in Mesoamerica. Its pres- ence so far from its origin and its common use by Native Americans in eastern North America suggest that it was likely traded and carried by indigenous people to the region prior to the arrival of Euro- peans. And some species are known to “hitchhike” along with human travelers. The Polynesian rat (Rattus exulans) was probably brought to islands by the early Polynesians (Merlin & Juvik, 1992), most likely as an accidental introduction. Charles Darwin provided the necessary context for understanding the biological invasion problem almost 150 years ago. One of the observations that stimulated Darwin to conceive of evolution by nat- ural selection was that each new continent or island he visited, despite similar environments, had dif- erent species. He wrote (Darwin, 1859: 343): “In considering the distribution of organic beings over the face of the globe, the first. great fact. which strikes us is that neither the similarity nor the dis- similarity of the inhabitants of various regions can be wholly accounted for by climatal and other phys- ical conditions. There is hardly a climate or condition in the Old World which cannot be par- alleled in the New [yet] how widely different their organic productions [that is, their species]!” Today we understand fully that the diversity of spe- cies on our planet is, in part, the result of conti- nental separation, producing the geography of life that so intrigued Darwin and all biologists since. However, this geography sets up one of the most profound threats to the Earth’s diversity. Because the continents differ more in species than climates, each continent has the potential to provide invaders to other places. Humans are a potent force for as- sisting that invasion. In their natural settings, species are found with friends and enemies. Darwin realized that this eco- logical context provided the ultimate check on un- fettered population growth. Thus, there is another important perspective to the invasion problem: we ' This and the five articles that follow it are the proceedings of the 48th Annual Systematics Symposium of the Missouri Botanical Garden, Biological Invasions. The symposium was held 12-13 October 2001 at the Missouri Bo- tanical Garden in St. Louis, Tema uri, U.S./ was supported in part t by r Raven and Mick Richardson, to whom Peter White would like to thank each and e the National Science Foundation under we are grateful. The ery symposium grant number DEB-9981642, symposium organizers Sarah Reichard Victoria Hollowell d the — = parlicipant as well as Annals editorial staff. Jeany Davidse and Sandy p provided cd eh 'al suppor ? Center. for Urban Horticulture, University reic chard@u. washington.edu. North Carolina Botanic 2 Garde n, Campus Box 3375 North Carolina 27599-3375, U.S.A. peter.white@une Rm ANN. MISSOURI Bor. of Washington, Box 354115, Seattle, Washington 98195, U.S.A. . University of North Carolina at Chapel Hill. Chapel Hill, GARD. 90: 64—66. 2003. Volume 90, Number 1 Reichard & White Invasion Biology often transport species without their coevolutionary context. This may result in poor performance (e.g., if a plant lacks a pollinator), but it may also result in the potential for rapid population growth (e.g., i natural enemies or s = the species is free from constraints). Darwin (1859: that some introduced species may threaten native species. In his chapter on geographical distribu- . many European productions 370) also recognized tions he reported, “ cover the ground in La Plata, and in a lesser degree in Australia, and have to a certain extent beaten the natives...." This may be the first scientific comment on invasions Although there were studies of invasions in the years since Darwin, biologists and ecologists did not focus on the magnitude of the growing problem until 1958 when Charles Elton ({1958] 2000) pub- lished a book that has come to be regarded as the seminal volume in this field of study. The Ecology of Invasions by Animals and Plants is a slim book borne from his observations as an animal ecologist. Three 1957 BBC radio broadcasts he developed on he subject of “Balance and Barrier” apparently stimulated Elton to write a book on invasions, aimed at a lay audience, that laid out what are still the fundamental issues in invasion biology: that each continent has its own unique flora and fauna, that human migration and trade were breaking down the barriers that had led to the uniqueness of the biota, and that this breakdown of the barriers could have severe consequences for the mainte- nance of diversity. The book is amazing in its science and yet was in large part underappreciated until the 1980s, when it became widely recognized that the invasion of non-native species was one of the biggest threats to naturally occurring species re- and ecosystems. In 1982 the Scientific Committee on Problems of the Environment (SCOPE), a committee of the In- ternational Council of Scientific Unions, met in a general assembly in Ottawa. At that meeting the invasive spread of plants, animals, and micro-or- ganisms introduced by humans outside their native ranges was identified as a problem of global con- cern, amenable to interdisciplinary synthesis. This determination led to a number of symposia held around the world in the mid 1980s and resulted in two important books (MacDonald et al., 1986; Moo- ney & Drake, 1986) that both address three im- portant questions: (1) What are the factors that de- termine whether a species will be an invader or not?; (2) What are the site properties that determine whether an ecological system will be relatively prone to, or resistant to, invasion?; and (3) How should management systems be developed using the knowledge gained from answering these ques- tions? Each symposium consisted of distinguished biologists, most of whom worked in related areas of study, though not specifically invasions, attempting to answer these questions from their understanding of the biota of their continent. These volumes (e.g., MacDonald et al., 1986; Mooney & Drake, 1986) serve as “state of the knowledge” documentation and are notable for their general lack of specific data on invasions. This was because there were few studies specifically on invasions from which to draw conclusions. Reading the chapters, it appears that more questions were developed than answered in the course of the symposia and book production. These books fueled the newly emerging field of in- vasion biology, as a generation of graduate students read the chapters and seized the opportunity to at- tempt to answer the questions derived during the SCOPE process. In the approximately 15 years since the publi- cation of the SCOPE books the field of invasion biology has not only influenced the content of many biological journals, but it has also developed its own journal (Biological Invasions, Kluwer Press), its own set of terminology, and its own set of raging scientific debates. While the SCOPE volumes pro- vided the initial fuel for the scientific field of in- vasion biology, it was the mounting evidence of se- vere environmental degradation that focused the attention of so many people on invasions. For in- stance, invasive species were identified as signifi- cant threats to biodiversity at more Nature Conser- vancy preserves with completed conservation plans than any other type of threat, including develop- ment, fire suppression, and altered hydrology. In fact, 94% of those sites responding as of the sum- mer of 2000 listed invasive species as a serious problem (J. Randall, The Nature Conservancy, pers. comm.). In 1998, a study found that invasive spe- cies were second only to habitat destruction and fragmentation in threatening endangered species in the United States (Wilcove et al., 1998). This study reviewed listing information for species proposed as endangered or threatened under the Endangered Species Act and found that 49% of the imperiled species were in that condition at least in part be- cause of invasive species. FUTURE CHALLENGES With the recognition that invasive species are one of the most serious conservation concerns to- day, there has also been the recent realization that the problem is getting worse very rapidly. Global- ization of trade and advances in technology mean Annals of the Missouri Botanical Garden that species are moving around the earth more fre- quently and are coming from some places, such as China, that have been closed off from most of the rest of the world during the last several decades. For instance, the Asian long-horned beetle (Ano- plophora glabripennis), which is native to China and considered to be a pest there, was first detected in the Greenpoint neighborhood of Brooklyn, New York, in 1996. The United States Department. of Agriculture subsequently determined that the in- sect arrived in solid wood packing material on goods imported from China. Not only are new path- ways opening, but the journeys that used to take weeks by ship may now take hours by plane, allow- ing more organisms to survive the trip. Trade agree- ments and organizations such as the North Ameri- can Free Trade Agreement (see МАКТА website, 2002), launched in January 1994, and the World Trade Organization (see WTO website, 2002), formed in January of 1995, limit the restrictions that signing countries can place on the entry of trade goods. Many invasive species are introduced either as trade goods themselves, in the case of some plants and animals, or as contaminants of trade goods, such as insects found in shipping dun- nage and pallets. Just as international trade has increased, so has pleasure travel. Tourism has become a major sector of the U.S. economy, with current figures of about $110 billion a year (up from about $26 billion in 1986). Over 46.5 million international visitors en- tered the United States in 1996, with a projected annual growth of 3—4% (Doggett, 1997), although tourism may decline over the next several years as a result of fears regarding safety in traveling. Trav- elers often inadvertently carry invasive hitchhikers on their person or property, but they also may intentionally bring in species. For in- stance, 16,997 international passengers checked during one week in May 1990 at the Los Angeles International Airport were found to be carrying species as 1357 lots of fruits and vegetables and 325 lots of animal products, for a total of 2635 kg of contra- band material (U.S. Congress, 1993). They may bring species with them for personal use or as gifts for friends and family. To address the problems caused by invasive spe- cies and the many pathways by which they enter, we must work with the deliberate intention of re- ducing the entry and impact of such species. As scientists, we will need to ensure that our science is not only viewed by our peers, as is traditional, but that agencies managing invasive species are also aware of our findings. As scientists we must learn from the past, examine the present, and plan for the future. As field of invasion biology moves into its adolescence, it will continue to test theories basic to ecology and to form new hypoth- eses lo address the novel situations that arise fol- lowing the introduction of new species. These dis- coveries, if implemented in management and policy practices, may play a substantial role in lessening environmental degradation through invasions. Literature Cited Darwin, C. 1859. The Origin of Species by Means of Natural Selection. [See Internet Resources, below. | Elton, ( [1958] 2t The Ecology of Invasions by Animals and Plants. Reprint, with a new аш һу D. ira Univ. Chicago Press, Chica Fritz, G. L. Precolumbian Cuc кү argyrosperma ssp. argy And (Cucurbitaceae) in mE easte m 1 wood- lands of Nor rth А Econ. Bot. ( Mac ‘Donald, I. Kruger & А. The Ecology and Management of Biologic val Invasions in n e. ‘a. Oxford Univ. Press, Cape Town. Merlin, M. I O. Jurvik. 1992. Relationships among native a nd plants on Pacific Islands with and with- out significant Varr а ‘е and feral каш. Pp. 597—024 in C. P. Stone, C. W. кае & 2 T. Tunison (editors), Alien i Invasions i in Nativ Hawaii: Management and Research. Unive 'rsity of Ha- waii Cooperative National Parks Resources Studies Unit. Mooney, H. A. & J. A. Drake. 1986. Ecology of Biological Invasions ie Noah America and Hawaii. Springer-Ver- rk. An eric a. lag, New Yo U.S. Ces [Office of Tec 'hnology Assessment]. 1993. Ha mfal indigenous Species in the United States, OTA-F-565. U.S. Government Printing Office, Washing- ton, D. Wilcove, . Rothstein, J. Dubrow, A. Phillips & E. Losos. 1998. E threats to ipea species in the United States. BioScience 48: 607-615. Internet Resources Darwin, 1859. Online version of “Оп the Origin of Spec (http//pages.britishlibrary.net/charles.darwin/ lexi Hori iml od L. R. J1. Tourism’s rol changing econ- . ITA vena Industries. tain tia.doc/about/ жа x. html) as of February 28, NAFTA [North American Free Tr y РОТ ment]. 200 Available at Gillen Paral nafta.html). WTO [World Trade Organization]. (http://www.wto.org/). 2002. Available at THE THREAT OF INVASIVE ALIEN SPECIES TO BIOLOGICAL DIVERSITY: SETTING А FUTURE COURSE! Elizabeth A. Pas ae and John M. Randa ABSTRACT Over the past decade, mounting evidence has shown the pervasive and escalating harmful impacts of invasive alien species on native species and ecosystems. Thousands of non-native ue ies are established in the United States and many more worldwide. Few areas aj species’ ranges pear immune to inv slowly expand across the landscape. А | asions, son lethora of ана al effects have been айг of which unfold over decades to centuries as ibuted to invasive species, and other кы change processes and wide mE habitat destruction will likely multiply these effects. Many conservationists now consider invasive identifying pragmatic, effective solutions. They know native species лев to deliberately introduce certain species for e species among the top restora limits of today's ecological knowledge and our ability to Birds de long-term ecological con threats to biological diversity and are grappling with г that the и will involve uc assemblages of native and non- and have important questions about which inv asive species to tackle which ones to ignore, and e ssing vow: ' issues will push мй aginn the Addres «qu Better envisioning the conservation goals, though. could help to guide the science. Key words: biological diversity, conservation, global change. invasive species, management, policy, research needs. “Nowadays we live in a very explosive world, and while we may not know where or when the next out- burst will be, we might hope to find ways of stopping it or at any rate damping down its force. Its not just nuclear bombs and wars that threaten us, though these other sorts 0, Ecological explosions differ from nol Bore such a loud noise and in taking longer lo appen ... but they can be very impressive in their offects. .. —Charles Elton ([1958] 1972: 15) Thus did the British ecologist Charles Elton, in his 1958 book The Ecology of Invasions by Animals and Plants, forcefully make the case that species translocations due to human activities are trans- forming the biological world and help inspire the now burgeoning study of invasion processes and impacts. Elton built a series of revealing case stud- ies, describing, for example. how chestnut blight, Cryphonectria parasitica (Murrill) Barr, virtually eliminate merican chestnut from eastern U.S. forests during the early 20th century and how parasitic sea lampreys, Petromyzon marinus (L.), native to the Saint Lawrence River and Lake On- tario, migrated up the Welland Canal after 1829 to decimate the Great Lakes’ trout fis Today, Elton’s case has been id by hun- dreds of more recent invaders, like common cru- pina (Crupina vulgaris Cass.), Asian tiger mosquito (Aedes albopictus Skuse), and the zebra mussel (Dreissena polymorpha Pallas) (Craig, 1993; Mac- Isaac, 1996; Roche & Thill, 2001; Strayer et al., 1999). However, although е quote above may be just as true as it was in 8, the context of bio- logical invasions has T авай: Rates of species invasions have escalated, and new patterns of species translocation have emerged. For example, the end of the Cold War, globalization of trade, and free trade agreements have connected previously isolated parts of the world and increased rates at which cargoes move among them (e.g.. OTA, 1993; Mc Neely, 2001). Improved transport technologies and increased numbers of species in trade mean that more species are being shipped and shipped more rapidly around the globe and, consequently, that more are surviving in transit and becoming established in new areas. Accelerating rates of habitat destruction, climate change, and ! We appreciate the help of Barry Rice and Liz ee "у abt у». and analyzing data on The Nature EL. vancy's experience with invasive species. Several tap in colleagues including Alan Holt, Maggie Coon, Ran s paper were developed during discussions with T Coni сй зы Ciruna, and Elliot Marks. Our thanks also go К, Peter Raven and the nogi Botanical Garden for foc ‘using the 2001 Annual Systematics Symposium on the important issue of invasive spec ? Author for Micaela, ence. P.O. Box 22665, Carmel, California 93922, U.S Nature Conservancy. eam, The з Wildland Invasive Species ‚А. echornesky@sbcglobal.net. . Department of Vegetable Crops & Weed Science, Uni- du. versity of California, Davis, California 95616, U.S.A. jarandall@uc 'davis.e ANN. MISSOURI Bor. GARD. 90: 67—76. 2003. Annals of the Missouri Botanical Garden other global change phenomena are disrupting na- tive species assemblages and creating new invasion opportunities (e.g., Mooney & Hobbs, 2000, and papers therein; Chapin et al., 1997) At the same time, there is reason for optimism. Worldwide efforts to reduce international transfers of invasive species, particularly those that harm agricul- ture and other economic interests have grown signif- icantly over the past half century (e.g., McNeely, 01). A broad commitment to the conservation of biodiversity has emerged as manifested in interna- tional treaties such as the Convention on Biological Diversity [see Internet. Resources] and in national laws, such as the U.S. Endangered Species Act of 1973 [see Internet Resources]. Reducing the ecolog- ical threat that invasive species pose to biodiversity is an important part of these efforts, and our technical capacity to do so has grown. In comparison to the 1950s, today's richer scientific understanding of in- vasion processes and impacts provides a much stron- ger basis for targeting efforts and for designing effec- tive prevention and control strategies. For those engaged in the conservation of biolog- ical diversity, then, the picture is one of urgency and hope: Urgency, because the threat is imminent and largely irreversible if left unchecked; Hope. because the opportunities for actions that would contain this threat have never been greater. IMPACTS OF INVASIVE SPECIES ON BIOLOGICAL DIVERSITY “Invasive species” here refers to non-native spe- cies that become established in new locations, spread, and then cause ecological or economic harm or threaten human health. This definition is consistent with recent uses in U.S Executive Order 13112—Jnvasive Species—signed by President Clinton in 1999 (Clinton, 1999). Note, however, that it differs from strictly ecological def- initions of “invasions” and “invasive,” such as the ones provided by Richardson et al. (2000) and Re- jmanek et al. (2002), which focus on the biological attributes of a species that enable it to spread and . policy, such as become established in new locations. Rough estimates are that invasive species, as de- fined here, comprise about 10-20% of the non-na- tive species that have become established outside of human cultivation in free-living populations, or more than 650 to 1750 species of plants, animals, and plant pathogens in the United States (cf. [Wil- liams & Meffee, 1996]; ОТА, 1993; Williamson & Fitter, 1996). Their ecological impacts range from local suppression of single native species to species extinction and wholesale changes in the functioning of ecosystems (for detailed summaries of impacts see ОТА, 1993; [Williams & Meffe, 1996]; Mack et al., 2000; Ewel et al., 1999; Randall, 2000). Perhaps the most frequently documented effects of invasive species are their suppression of native spe- cies populations through predation, competition, par- asitism, or disease. A recent example is the Asian swamp eel, Monopterus albus (Zuiew), a fish originally from tropical to warm temperate portions of east Asia that was first detected in Florida and Georgia in 1996, possibly following escape from aquaculture facilities Ben- or intentional releases of aquarium specimens son et al., 2001]. Populations of native fishes appar- ently have declined in areas where this generalist predator has become established [Benson et al., 2001]. The eventual full impacts of the Asian swamp eel in the U.S. are as yet uncertain, but the fish's capacity to tolerate extended periods out of water and at low temperatures suggest its range and associated impacts will expand. Certain invasive species hybridize with congener- ic native species. At a minimum, this alters the gene pool of the native species, sometimes in maladaptive ways. At a maximum, introgression can eventually allow genetic hybrids or the invasive species to com- pletely displace the native species. A recent well- documented example is the invasion of San Francis- co Bay salt marshes by smooth cordgrass (Spartina alterniflora Loisel.), originally from the Atlantic and Gulf coasts of the United States (Anttila et al., 2000: [San Francisco Estuary Invasive Spartina Project, 2002]; Vila et al., 2000). Spartina alterniflora hy- bridizes with the native Spartina foliosa Trin., al- though rarely, and the resulting hybrids outcompete the native cordgrass, spread clonally, and interbreed with both parental species (Anttila et al., 1998, 2000; Ayres et al., 1999; Vila et al., 2000). Where they have displaced the native cordgrass, the non- native species and the hybrids alter marsh architec- ture, because, in comparison to the native species, they grow taller, thrive in deeper water, and form » denser rhizome mats that trap sediments and raise marsh elevation. These attributes enable the invad- ing species and hybrids to expand into large areas that previously were open mudflats, displacing com- munities of algae and invertebrates associated with mudflats and reducing foraging resources for certain migratory waterfowl and other animals (Callaway & Josselyn, 1992) The impacts of certain invasives on native spe- cies are indirect—mediated by the altered behavior or other characteristics of intermediary native spe- cies that are not themselves harmed. Invasive plants that have showy floral displays, like purple loosestrife (Lythrum salicaria L.), have the potential Volume 90, Number 1 2003 Chornesky & Randall 69 Threat of Invasive Alien Species to alter the abundance or behavior of insects that pollinate native plants and thereby indirectly affect native plant reproduction and fitness (Grabas & Laverty, 1999; Brown et al., 2). Certain invasive shrubs may indirectly reduce the reproductive suc- cess of native songbirds, because their architecture and lack of sharp thorns allow predators greater access to songbird nests than do native shrubs (Schmidt & Whelan, 1999). The most harmful invasive species are those that have system-level impacts, fundamentally altering the ecological processes that structure communities and ecosystems (Vitousek, 1990). A well-known ex- ample is melaleuca, Melaleuca quinquenervia (Cav.) S. T. Blake, an Australian tree that has spread and become locally abundant in Florida's Everglades. The tree forms dense thickets that trap sediments and debris and elevate the topography. Because of the plant's high rates of evapotranspiration, mela- leuca appears to draw down water levels and alter the hydrology of areas where it becomes abundant (Schmitz et al., 1997; Hofstetter, unpublished data). Melaleuca is pyrogenic and increases the temper- ature and duration of fire. Overall, the tree is fun- damentally changing habitats across Florida's “Riv- er of Grass" and diminishing recruitment in these wetland habitats of native plants and their associ- ated faunas (e.g., Gordon, 1998). The zebra mussel, Dreissena polymorpha, is an- other invasive species that is causing broad system- level effects (MacIsaac, 1996; Strayer et al.. 1999). The mussel directly suppresses native freshwater bivalves by overgrowing and smothering them. Dense accretions of zebra mussels on hard and soft lake bottoms create new habitat space for benthic invertebrates. Because the mussels’ efficient filter feeding clears the water of algae that previously supported pelagic predators, zebra mussels have undamentally transformed the food web dynamics of places like the Great Lakes. Most production and biomass now occur in benthic parts of these sys- tems, whereas pelagic production and biomass once dominated (Maclsaac, 1 Certain invasive species facilitate the establish- ment of other non-native species or exacerbate their harmful effects. The laurel fig, Ficus microcarpa L. f., for example, was introduced into Florida in 1912 as an ornamental tree, but only became invasive about 45 years later when its natural pollinator, a fig wasp specific to E microcarpa, was introduced. The laurel fig has now been reported from 18 of Florida's natural areas, including the Big Cypress National Preserve and the Loxahatchee National Wildlife Ref- uge (Langeland & Burks, 2000). Simberloff and Von Holle (1999) have hypothesized that “invasional meltdown”—accelerating impacts on native ecosys- tems—occurs because of the synergistic interactions among repeated species invasions. By suppressing native species populations and altering habitats and ecosystems, invasive species have contributed to the imperilment of nearly half States (Wilcove et al., 1998). According to an anal- ysis by Wilcove et al. (1998), non-native species are the second most frequent cause of species im- perilment; the most frequent cause is habitat deg- radation and loss. In addition to the nearer-term ecological impacts discussed above, species invasions also may be po- tent drivers of evolutionary change (e.g., Grosholz, 2002; Mooney & Cleland, 2001; Sakai et al., 2001). Some invading species may pass through an evo- lutionary “bottleneck,” for example, if certain phe- notypes survive the pathway of species transit or if the receiving environment imposes strong selective pressure on invading populations. Conversely, suc- cessful invaders may cause genetic changes in pop- ulations of native species by hybridization or by qualitatively or quantitatively altering selection pressures through ecological interactions or through changes in important habitat qualities or ecosystem processes. Many examples exist of hybridization between invasive non-native plants and closely re- lated native species yielding fertile, and in some cases, highly competitive hybrids or new species that are reproductively isolated from both parents by mechanisms such as polyploidy (Vila et al., 2000; Mooney & Cleland, 2001). IMPLICATIONS FOR CONSERVATION Rich case histories and a large body of anecdotal evidence have dominated our understanding of in- vasive species impacts on native species and eco- systems. Synthesizing this information into a more cohesive picture has proven challenging for con- servationists seeking to ensure the long-term sur- vival of native species and ecosystems. This has contributed to the relative slowness with which the environmental community has recognized the im- portance of invasive species as a major threat to biodiversity, and one whose impacts can be signif- icantly reduced by effective action. Here we describe how one major conservation organization, The Nature Conservancy (TNC), has developed an overarching assessment of the threat that invasive species pose to biological diversity. TNC is one of the world's largest place-based con- servation organizations, working at thousands of lo- Annals of the Missouri Botanical Garden cations around the world. The organization's meth- ods for selecting and designing conservation sites have evolved significantly over the past 12 years to include the identification of conservation threats and their solutions at multiple spatial and ecolog- ical scales. From this new methodological rigor, a new picture has emerged of how pervasive and sig- nificant invasive species are in the organization's work and for biodiversity conservation in general. The Conservancy selects sites for conservation ac- tion through the process of “ecoregional planning"— a stepwise process that uses ecoregions as the basic planning unit (Groves et al., 2002; [TNC, Ы). Each ecoregional plan identifies a collection of “conser- vation areas” required to fully protect the ecoregion's biological diversity. Invasive species were identified as among the top threats to TNC's conservation tar- gets in 15 of 18 completed ecoregional plans for North American ecoregions in a summary completed in 2000 (Groves & Valutis, unpublished data). Hab- itat destruction was the only other threat that was cited more frequently in the plans. A site plan is prepared for each of the conser- vation areas where TNC works or intends to work. Each plan formally assesses the relative sev erity of known threats to the area's biodiversity—such as unsustainable timber harvest, altered hydrology. second home development, or invasive plants—and sels priorities for actions to abate or mitigate these threats. Examination of plans for 89 conservation areas from 39 North American ecoregions showed the majority (94%) listed invasive species as a crit- ical threat to biodiversity at the site [TNC, a, 2000]. Among these plans, invasive species were the site- level threat most often assigned high priority due to significance and irreversibility. Implementing a conservation area plan often in- volves direct ecological management of natural communities. In 1998, The Conservancy's Invasive Species Program conducted a survey of all TNC staff in the U.S. with land management responsi- bilities [TNC, с]. A total of 110 responses were received from staff working at 916 conservation ar- eas encompassing more than 2 million acres. The managers identified 216 invasive plant species in- festing 810 (8896) of the conservation areas covered by the respondents. Table 1 provides an illustrative list of invasive plants considered a serious threat at many of these places or over a large geographical area. The particular invasive plants considered problematic varied greatly among TNC's conserva- tion areas. Of the 216 species, managers identified 175 asa area. Most respondents (over 80%) considered in- serious threat to at least one conservation vasive plants one of their top 10 management con- Table 1 rious threat to many sites or across a large geographical Examples of plant species identified as a se- area by TNC managers. Lonicera japonica Thunb. (Japanese honeysuckle) deb шш australis (Cav.) Trin. ex Steud. (common Bel eee salicaria L. (purple loosestrife) Alliaria petiolata (M. Bieb.) Cavara & Grande (garlic mustard) Cirsium arvense (L.) LE e dn thistle) Rhamnus cathartica L., «ша L. (syn. Frangula alnus Miller) (buc p Euphorbia esula L. (leafy spurge) * This Table includes species identified based on the number of conservation a acres they affect and by how fre i ACAS anc servation actions. The list is biased by the geographic dis- рни of survey respondents ds relative size of differ- ent conservation areas. Oth highly ranked species inc luded tamarisk (Tamarix spp. + ed canarygrass (Phal- aris arundinacea L.), and spotted Tucci (Centaurea maculosa Lam.). cerns, with 49% of the respondents identifying in- vasive species as one of their top 3 management concerns, Garlic mustard, Alliara petiolata (M. Bieb.) Ca- vara & Grande, is one example of the kind of in- vasive species that TNC is tackling. The plant, first detected in 1868 in Long Island, New York, now occurs in 20 eastern and midwestern states and has recently been observed in Washington, Colorado, Idaho, and Alaska (Nuzzo, 1993, [2000]; [Tu & Rice, 2001, 2002]: Rice & Johnson, in press). Gar- lic mustard invades forest understory, where it ap- parently displaces native herbaceous species, par- ticularly spring ephemerals, and reduces survival of tree seedlings in some areas (Anderson et al. 1996; McCarthy, 1997). weed also reduces the fitness of at least two native Evidence suggests the butterflies because, although garlic mustard attracts these insects, eggs laid on the plant have lower survivorship than those laid on native host plants 1971: 1994). Of additional worry n | to tionists is that controlling the plant is (Bowden, Porter, рае оне и often difficult, and some- times impossible | Nuzzo. due Feral pigs (Sus scrofa L.) provide a graphic « ex- ample of the kinds of problems сопѕет face with invasive animals. The pigs cause dura: by uprooting and killing native vegetation. and thereby opening areas for colonization by invasive plants that further alter the system. Feral pigs cause severe problems in conservation areas across the U.S ., ranging from the Great Smoky Mountains Volume 90, Number 1 Chornesky & Randall 71 Threat of Invasive Alien Species National Park in Tennessee and North Carolina to Santa Cruz Island off the coast of southern Califor- nia to Haleakala and Hawaii Volcanoes National Parks in Hawaii (Aplet, 1990; Bratton et al., 1982; Hone & Stone, 1989; Hone, 1995; Singer, 1981). In Hawaii, for example, subsequent recruitment of non-native plants, such as the nitrogen-fixing Por- tuguese fire tree (Myrica faya Aiton), into areas dis- turbed by pigs alters fundamental ecosystem pro- cesses and initiates а cascade of other species changes (Aplet et al., 1991; Vtorov, 1993 The compelling picture emerging from The Na- ture Conservancy's efforts at all levels is that in- vasive species are a high-priority threat to biolog- ical diversity that must be addressed now. Many in the organization believe that invasive species pose a fundamental risk to the biological diversity that TNC and partner organizations once thought was already successfully conserved on millions of acres and in numerous aquatic and marine systems. In response, TNC is presently mounting a major, mul- ti-year organizational initiative to address the in- vasive ерене threat. (For wer information see cdavis.edu/is or contact the жеч x en Ann Bartuska, at abartuska@tnc.org.) SETTING A FUTURE COURSE A picture has emerged over the past decade of the full impacts of invasive species on the world’s biological diversity, economies, and human health. Paradoxically, the very breadth and depth of these impacts may hold the key to a solution as the many constituencies affected by species invasions—such as conservation, agriculture, forestry, tourism, pub- lic health, horticulture, and others—combine their energies in finding common solutions to shared challenges [e.g.. National Invasive Species Council, 2001]. Understanding has also grown over the past decade of the potential policy and management strategies for reducing the invasive species threat as well as a set of key questions that can only be solved by scientific research. This section identifies several critical science challenges that lie ahead. PREVENTION, EARLY DETECTION, AND RAPID RESPONSE Improved prevention—at scales ranging from the continental to the small nature preserve or sub-wa- tershed—Xwill be the most effective way to minimize the numbers and impacts of future species inva- sions. This will require scientific and technical ad- vances that enhance our ability to understand and predict invasion pathways, to assess the risks of proposed species introductions, and to detect and manage new incipient species invasions. It also will require national and international policies that flex- ibly and rapidly integrate these advances into op- erational practices [e.g., National Invasive Species Council, 2001]. A new level of collaboration among ecologists and other researchers, engineers, and policy experts thus will be essential to the success of future prevention efforts. Issues related to international trade demonstrate the need for tighter linkages between scientists and policymakers to improve prevention. Because of the dual trends of globalization and trade expansion, incorporating measures to minimize new invasions is both more urgent and more difficult. Inspections, quarantines, and other forms of screening must rapid and compatible with free-trade agreements already in force or pending. This will require im- proved technology, transparent rationales for ex- cluding certain species, and control actions that complement any exclusion actions within each na- tion or trading bloc. Policies under development (e.g.. the new U.S. National Management Plan [Na- tional Invasive Species Council, 2001 J) include sci- ence and research agendas specifically designed to meet current and emerging information needs that derive from policy innovations. Developing criteria and systems for accurately predicting which species are most likely to become harmful invaders if introduced to a given area is one of the most important technical challenges ahead. Weed Risk Assessment systems to evaluate new species proposed for intentional introduction have already been implemented in Australia (Phel- oung et al., 1999). In addition, representatives of the U.S. nursery industry and other user groups have recently moved to develop a voluntary “code of conduct” for reducing the risks of new plant in- troductions (Randall et al., 2002). These and other systems proposed by researchers (e.g., Reichard & Hamilton, 1997; Rejmanek, 1996) are being scru- tinized for possible modification and use in Hawaii and other parts of the world (Daehler & Carino, 2000). Further analysis of their performance will help improve these systems and offer insights into how systems for other taxa and other areas could be devised (Kolar & Lodge, 2002). Even with improved prevention systems, at least some new invasions will continue. Many might be contained or even eradicated if detected early enough, and general agreement exists that better early detection systems are needed [e.g., National Invasive Species Council, 2001]. Research discov- eries that improve our ability to search for and de- tect new invasions will be crucial to making such 72 Annals of the Missouri Botanical Garden plans work. For example, spatially explicit models that predict where new invasions are most likely to occur and how they would spread from sites of ini- tial establishment could greatly improve the effi- ciency of searches (e.g., Higgins et al., 2000; Wads- worth et al., 2000). Easily accessible and networked databases that accept reports of new species and locations and that automatically notify officials in the affected area would allow such in- formation to be acted upon rapidly. Accessible tax- onomic keys that have multiple entry points or per- haps even on-line services that can identify species from scanned images would allow many thousands of land managers and other interested citizens to integrate their observations into these early detec- tion systems (e.g., Brooker et al., 1999). A FUTURE OF MIXED SPECIES ASSEMBLAGES Even with well-supported prevention programs in Many harmful species are already established. Others presently under human cultivation may some day start to spread beyond cultivation, become firmly established in the wild, and cause problems after long lag periods. And at least a few new invaders are likely to elude even the most exemplary pre- vention efforts. Consequently, invasive species are not a problem that can be “solved” or “engineered away"; rather this is a problem that will require continuous management into the future. place, a need will exist to manage invaders. atural resource managers and conservationists routinely face chronic shortfalls of money and staff time in the face of urgent needs. Since they cannot do everything, they must set priorities and tackle the most important problems first, often leaving lesser problems unaddressed altogether. In. many natural areas, the number of established invasive species or the area infested already far exceeds lo- cal management resources. Many natural area man- agers therefore take a “triage” approach and at- tempt to set priorities by identifying which species, if any, are most detrimental to the native species, communities, and ecosystem processes they are seeking to protect (see, for example, the weed con- trol plan for TNC’s Cosumnes River preserve in California available at [TNC, dJ). Several decision-support frameworks аге avail- able to help managers set priorities. These frame- works operationalize current ecological knowledge about individual invasive species their impacts, current and projected distribution range, and ease of management. For example, The Nature Conser- vancy's Weed Management Plan Template contains guidance on how to set priorities among invasive species and among specific invaded portions of a preserve [TNC, c]. The U.S. National Park Service (NPS) supports an online invasive plant ranking system that calculates а score used for assigning relative management priority to each invasive plant species at a site (Heibert, 2001). Importantly, the NPS system evaluates species’ impacts relative to the conservation goals for a specific site and ex- plicitly balances projected impacts of invasive spe- cies against management feasibility and costs. These systems warrant further testing and improve- ment, and similar priority-setting is needed for oth- er groups of invaders (e.g., insect invaders, fish, aquatic invertebrates). An implicit assumption underlying such priority- setting approaches is that conservation in the fu- ture, for some sites, will involve managing mixed assemblages of native and non-native species. Yet the focus of many site managers remains primarily on how to eliminate or sharply reduce the abun- dance of particularly harmful invaders. In some systems and with some invaders effective control has proven to be extremely difficult and sometimes impossible, however. For example, attempts to con- trol of cheatgrass (Bromus tectorum L.) across large expanses of the intermountain West have had very limited success despite many and varied attempts dating back over 50 years (Mack, 1986; Upadhaya et al., 1986; Young & Evans, 1985). In contrast, we believe conservation practitioners must allocate their resources first toward sustaining nalive species and communities and retaining, re- storing, or mimicking critical ecosystem processes, such as fire regime or periodic flooding (e.g., Rand- all et al., 1997; Stromberg & Chew, 2002a). Some- times this may mean allowing certain invasive spe- cles — o persist. In others, it may managing a non-native species to maximize the conservation value of an area. For example, in some cheatgrass-infested areas, rather than seeking to control the cheatgrass, limited management dollars might better be used for planting other native grass- es and shrubs, preventing frequent fires, and scat- tering seed periodically to ensure reliable, season- long availability of foods required by native rodents and their predators. Modeling, demonstration pro- jects, and monitoring and adaptive management will be necessary for this approach to succeed. Managers are already using this tactic in Hawai'i Volcanoes National Park, where fire-promoting in- vasive grasses have converted much of the season- ally dry woodlands to alien grass savanna (Tunison et al., 2001). Rather than seeking to directly control the non-native grasses and restore the original woodlands, management actions emphasize “reha- Volume 90, Number 1 2003 Chornesky & Randall 73 Threat of Invasive Alien Species bilitation”—the identification and establishment of fire-tolerant native trees and shrubs that potentially will persist and spread in the savannas now domi- nated by alien species. Approaches like these that involve managing mixed assemblages of native and non-native spe- cies test the limits of existing ecological knowledge. Practitioners rarely have information about the like- ly long-term impacts of the management choices they make today on the future structure and func- tion of ecological communities and ecosystems. This is one of the chief shortcomings of current approaches and of the information base that has been supplied by prior research. The field of bio- logical control, a particularly rich discipline in in- vasion biology, provides an instructive example. Al- though recent advances in selecting introductions of control agents promise to reduce ecological im- pacts on non-target species (McEvoy & Coombs, 1999), methods do not yet exist for assessing effects of biocontrol introductions on ecosystem functions, such as the recently documented impacts in Hawaii on food web dynamics (Henneman & Memmott, 2001). An emphasis on long-term research and the development of spatially explicit models and fore- casting approaches to predict the long-term ecolog- ical impacts of today’s management choices and de- faults could help to remedy this gap. CONSERVATION AT LARGE SPATIAL SCALES An important paradigm shift is under way in the spatial scale of conservation efforts. Increasingly, conservationists are focusing on conserving areas that are large enough to sustain dynamic ecological processes and functioning ecosystems in addition to populations of rare species and examples of nat- ural community types. This shift, from preserves of hundreds of acres to conservation areas that are tens of thousands to millions of acres in extent, brings new challenges in addressing invasive spe- cies. For example, controlling local infestations of invasive plants and restoring native vegetation are relatively straightforward for small conservation ar- eas like The Nature Conservancy's 76-acre Blowing Rocks Preserve in Florida where mechanical re- moval, spot applications of herbicides, and direct planting of native species are standard practices (Randall et al., 1997 onsider how infeasible such approaches are for large areas, such as the 250,000-acre Demonstration Weed Area along the Snake River that encompasses parts anagement of Oregon, Washington, and Idaho! The profound mismatch that currently exists be- tween the large spatial scale required to conserve dynamic ecosystems and the small spatial scale of standard methods for detecting, monitoring, and managing invasive species has and should continue to fuel interest in technologies and management methods that can operate at very large spatial scales. Remote sensing data applied in GIS for- mats, for example, promise to provide cheaper and more effective ways to detect and track the spread of new weed infestations—and therefore effective ways to eradicate or contain invasions early (e.g., Everitt et al., 1995; Andrews, 2001). Landscape analyses of invasion patterns may help identify emerging pathways of species spread and provide a way for targeting resources to reduce species tran- sit and to monitor for new invasions (Thompson, 1999). Creating “nodes” of native species that serve as source populations and seed banks might en- hance the recruitment and persistence of native species in certain systems dominated by invasive species (Tunison et al., 2001). interest exists in whether or not management pre- Finally, considerable scriptions, such as large-scale patterns of grazing, prescribed fire, or flooding and hydrology, that tar- get abiotic or biotic ecosystem processes will pro- vide cost-effective ways to suppress populations of invasive species and yield conditions that favor na- tive species (e.g., Koebel, 5; Randall et al., 1997; Thysell & Carey, 2001; Stohlgren et al., 1999: Stromberg & Chew, 2002b) INTERACTIONS WITH OTHER GLOBAL CHANGE PROCESSES Ours is a time of dramatic and escalating change in the world's biota. Over the near-term, invasive species and habitat loss and fragmentation due to land use change and changing resource exploitation patterns. will continue to be the major drivers of such change (Vitousek et al., 1996; Mooney & Hobbs, 2000, and papers therein). Less well un- derstood, but at least equally important over the long-term, will be the biological changes wrought by nitrogen deposition, enhanced CO, levels, and changing temperature and precipitation patterns (e.g., Mooney & Hobbs, 2000). The magnitudes and qualitative details of most global change processes have as yet proven difficult to predict. Nevertheless, all of these global change process- es will interact with one another to yield cumulative impacts on the distribution and abundance of spe- cies, the composition of ecological communities, and the functioning of ecosystems. Climate change, nitrogen deposition, and ambient CO, levels all will have profound effects on the pool of potential in- vading species, potential invasion pathways, and Annals of the Missouri Botanical Garden the suitability of native habitats for colonization (Mooney & Hobbs, 2000, and papers therein). The shift in CO, concentration from the pre-industrial level of 280 ppm to 364 ppm today may already have begun to benefit certain plant invaders like cheatgrass (Bromus tectorum), and levels are ex- pected to rise further to 560 ppm within 100 years (Dukes, 2000). Climate change is expected to result in the dis- assembly of many terrestrial and aquatic ecological communities as species ranges and abundances shift in response to changing temperature and pre- cipitation patterns and associated changes in fire éd and other disturbances (e.g., Chapin et 1997). This will greatly enhance opportunities for new invasions (e.g., Walther, 2000). The process also is likely to change the ecological impacts of certain invasions, their management potential, and even the possible and desired endpoints of conser- vation management activities. An important, and largely unanswered, question, in light of the com- bined impacts of various global change processes, is whether some of the invasive species that we choose to control today will become acceptable or even desired species at some sites tomorrow. CONCLUSION Invasive species have emerged as one of the key threats to the world's biological diversity. Today's richer scientific understanding of invasion process- es and of management and policy approaches for reducing impacts forms a strong foundation for tak- ing effective action to reduce this threat. Tighter linkages between scientists and policymakers will be essential over the near term to improve the pre- vention of new invasions and to enable more rapid detection and response to the new invasions that do occur. The critical science challenges ahead in- clude developing ecological forecasting approaches to help set management goals for mixed species assemblages, building methods to effectively ad- dress invasive species and promote native species and communities at large spatial scales, and deter- mining how the interactions among various global change processes should alter how conservationists consider and respond to species invasions. Literature Cited Anderson, R. C., T. C. Kelley & S. S. Dhillion. 1996. Aspects of the ecology of an invasive plant, garlic mus- Э rd (Alliaria Ce in central Illinois. Restoration col. 4: 181-191. жо. M. M. 2001. Use of AVIRIS Hyperspectral Re- mote Sensing to Map Mustard (Brassica nigra) and Ке п- nel (Foeniculum vulgare). ri aii M.S. Thes University of California, Davi Anttila, C. . Daehler, N. E. Rank & D. R. Strong. 1998. E sib fitness of a rare invader (Spartina alterniflora, Poaceae) threatens a common native (Spar- tina foliosa) with hybridization. Amer. J. Bot. 85: 1597— 1601. ————, R. А. King, С. ‚ D. R. Ayres & Strong 2000. Reci iproc a hybrid formation of Spartina 67 Francisco Bay. Molec. Ecol. 9: 765—771. AE G. H. 1990. Alteration of dim community biomass by the alien Myrica faya. Oecologia 82: 414— 416. Anderson & C. P. Stone. 1991. Association‏ حت be FU s pig disturbance and the composition of‏ some alien plant assemblages in Hawaii Volcanoes Na-‏ tional Park. Vegetatio 95: 55—02.‏ Ayres, D. R., D. Garcia-Rossi, H. G. Davis & D. R. Strong: 1999, Extent and Кы» of hybridization be- tween exotic (Spartina alterniflora) and n (S. folio- sa) cordgrass (Poaceae) in California, USA determined by random era о SR QUE Ds). Mo- lec. col. f ыо S. a pif pore white butterflies (Pieri- dae) and ARRA food p . Lepid. Soc. 25: 6-12. Bratton, S. P., E. Harmon & P. S. White. 1982. Pat- terns of European wild boar rooting in the western Great Smoky Mountains. Castanea 47: 230-242. Brown, B. J., R. J. Mitchell & S. A. Graham. 2002. Com- petition for pollinstion between an invasive specie (purple loosestrife) and a native congener. Ecology ax 2328-23 36 6. Callaway, J. C. & M. N. Josselyn. 1992. The introduction and spread of smooth cordgrass (Spartina oe in south S ancisco Bay. Estuaries 15: 218— Chapin, F. " HL iyd ar, R. 4 Hobbs, D. U. bs г, Ј А 0. E. Sala & D. Tilman. 1997. Biotie c ontrol vh er the func onines fe ecosystems. Science 277 500-5) х 1. 1999. Executive Order 13112 of I RA ): Invasive Species . Fed. Reg. 64: 618 196 9З. Тһе diaspora of the A / pum tiger о 2000. Pret caine plants: Prospects for a gen ing system based on current re ER ui. Biol. de 2: 93-102. Dukes, J. 5. 2 Will the increasing апарар CO, concentration v t the success of invasive species? Pp. 95-114 in Н. A. Mooney & R. J. Hobbs (editors), In- vasive Speci ies in a Changing World. Island Press, Washington, D.C. Elton, C. S. [1958] 1972. The Ecology of Invasions by Animals and Plants. Reprint. Chapman and Hall, Lon- ral sere . Anderson, D. E. Escobar, M. R. Dav ОВ. ae er r& H. J. Andrasciket. 1995. Use of r re- mote sensing for detecting and puri. = spurge (Euphorbia Pus Weed Technol. 9: Ewel, J. J., D. J. O'Dowd, J. Bergelson, Fd adiu E M. D'Autonio: i D. Gomez, D. R. Canon. R. Hobbs. A. Hol, K. R. Hopper, C. E. Hughes, M. LaHart, R. ML D ыыы но, tions of species: Research needs. BioScience 49: 619 630. Gordon, D. R. 1998. Effects of invasive, non-indigenous Volume 90, Number 1 2003 Chornesky & Randall 75 Threat of Invasive Alien Species plant species on ecosystem processes: Lessons from Florida. Ecol. Applic. si мү 989. Grabas, G. P. & T. M. La 1999, The effect of purple loosestrife (Lythrum psi L.; Lythraceae) on the pol- lination and reproductive success of sympatric co- e ering wetland plants. Ecoscience 6: 230-242. Grosholz, E. 2002. Ecological and evolutionary conse- quences of coastal invasions. Trends Ecol. Evol. 17: 22-27. Groves, C. R., D. B. Jensen, L. L. = H. Redford, Wel Sha м t, J. V. Bau . Hig .M.W.B "C: Piero, 79002; Pisin for cir aig: conserv iat n: Putting conservation sci- ence into practice. BioScience 52: ج‎ . 200 i nin Practice, Vol. 2. No. 1. [Ranking criteria also available online. at (http://www. npwre usgs Henneman, M. L. & J. Memmott. 2001. бсо of a Hawaiian community by и biological control agents. Sci EY ge 1314-1: Higgins, S. L, . Ric ik R. M. Cowling. 2000. Using a T1 landscape model for planning the management of alien plant invasions. Ecol. Applic. 10: 1833-1848. Hone, J. 1995. Spatial and temporal aspects of vertebrate pest purs with emphasis on feral pigs. J. Applied Ecol. : 11-319 I c P. Stone. 1989. A comparison and evalua- tion of feral pig management in two national parks. Wildlife Soc. ug 17: 419-425. Koebel, J. W . An historical perspective on the imet River restoration project. Restoration Ecol. 3: 149-159. Kolar, C. S. & D. M. Lodge. 2002. Ecological prediction and risk veri for alien fishes in North America. Science 298: 12: Maclsaac, Н. J. eh Potential —— and biotic im- pacts of zebra mussels on the inland waters of Nort America. Amer. Zool. 36: 287-299. Mack, R. N. 1986. Alien plant invasion into the inter- лаана west: А case history. Pp. 191-213 іл Н. А. Mooney & J. A. Drake (editors), Ecology of Biological Invasions of North America and Hawaii. Springer- Ver- lag, New York. mberloff, W. M. Lonsdale, H. Evans, M. Clout & F. A. Bazzaz. 2000. Biotic invasions: Causes, epidemiology, es consequences, and control. Ecol. Applic. 10: 689-71 McCarthy, B Response of a forest understory com- munity to experienta же val of an invasive nonin- digenous ay d а Вгаѕѕісасеае). Рр. 117-130 in J. € ken & J. W. Thieret (editors), As- sessment and онаа nt of Plant [nvasions. Springer- Verlag, New Yor McEvoy, P. B. & E. M. — 1999. Biological control of plant invaders: Regional patterns, field exper — and structured she wee models. Ecol. Applic. 1-401 McNeely, J. A. (Editor). 2001. A Global Strategy on In- vasive Alien Species. IUCN Gland, Switzerland, and Cambridge, U.K., in collaboration with the Global In- vasive Species Sb рш IUCN Publications Services Unit, Cambridge. Mooney, Н. A. & К. impact of invasive species. U.S.A. 98: 5446—5451. Е. РТ 2001. Proc. The evolutionary Natl. Acad. Sci. R. J. Hobbs (Editors). 2000. rain v in a onc vet Island Press, Washing Nuzzo, V. Distribution and spread bei inva- sive lenia Air ier уге Сауага & Grande] i Amer Pp. 115-124 in B. McKnight (editor), Biological "i sisi The Control id m of Invasive Exotic Species. Indiana Academy of nce, Indianapolis. OT А (Office of Tec bbs: Assessment). 1993. pyes Non-indigenous Species in the United States. U.S. Go ernment Printing Office, Washington, D.C. Porter, A. 1994. Implications of introduced garlic mus- tard (Alliaria peona in the habitat a м virgi- niensis (Pieridae . Lepid. Soc. mg 17 Phe M id illiams . R. Malloy. ed sk assessment model ы easa Pii iol evaluating plant introductions. 1. Баке, Мапарет. 57: 239— Randall, J. " 2000. Improving management of nonnative invasive plants in wilderness and other natural areas. U.S.D.A. Forest Service Proc. 5: 64—7: ‚ R. R. Lewis, IH & D. B. Jensen. ` 1997, Ecolog- ical restoration: Pp. 205-219 in D. Simberloff, D. C. Schmitz & T. C. Brown (editors), Strangers in Paradise: me and Management of Nonindigenous Species in Florida. Island Press, Washington, D.C. Reichard, S. H. & C. W. Ha т-а 1997. Predicting in- vasions of woody plants introduced into North America. быка Biol. 11: 193-203. Rejmanek, M 6. A theory of seed plant invasiveness: The first sketch. Biol. Conservation 78: 171-181. . Ric qs M. G. Barbour, M. J. Crawley . Е Hn ‚Р. B. Moyle, J. M. Randall, D. Simberloff Š M. Williamson. 2002. Biological invasions: Politics and the discontinuity of ecological terminology. Bull. Ecol. Soc. Amer. 83: pers Rice, B. A. & P. Johns In Press. Noteworthy collec- Alliaria petiolata (Bieb.) Cavara & Grande in Pysek, M. Rejmanek, M. С. Bar bour, F. D. Panetta & C. J. West. 2000. Naturalization and invasion of alien plants: Concepts and definitions. е апа Distributions 6: 93—107. Roche, C. T. . Thill. 2001. Biology of common crupina | and oe ‘tarthistle, two Mediterranean win- ter annual invaders in western North America. Weed Sci. 49: 439-447. Sakai, A. K., F. W. Allendorf, J. S. Holt, D. M. Lodge, J. Molofsky, K. A. With, S. R. J. Cabin, J. E. ©. McCauley, P. O'Neil, I. Thompson & S. G. Weller. 2001. The population bg of invasive species. Annual Rev. Ecol. S c yy 332. сеа; k. ;. J. Whelan. 1999. Effects of exotic Lonicera "e trad on songbird nest predation. Con- — = 13: 1502-1506. Schmitz, D. C., D. Simberloff, R. H. Hofstetter, W. Haller & D. а igit The ecological impact of nonindig- enous plants 9—61 in D. Simberloff, D. C. Schmitz & T. qned nia ud Strangers in Paradise: Impact and Management of Nonindigenous Species in Florida. Island Press, iege n D.C. sine D. & B. Von Holle. 1999. Positive interac- ons of nonindigenous species: Invasional meltdown? Biol. ее 1: 21-32. Singer, F. J. . Wild pig gem in national parks. Environm. poen 5: 363-3 = "9 { = е | 76 Annals of the Missouri Botanical Garden Strayer, D. L., N. F. Tei co, J. J. Cole, S. Findlay & M L. Pace. 1999, шайлар. B freshwater ecosys- tem Ans bivalves тыда 'lence -21. P T. J., Schell & B. V үп 1999. How grazing and An s affect native and exotic plant diversity in rocky mountain grasslands. Ecol. Applic. 9: 45—64. Stromberg, J. C. & M. K. Chew. 2002a. Foreign visitors in riparian corridors of the American Southwest: Is xen- ophytophobia justified? Pp. 195-219 in B. Tellman (ed- itor), Invasive Exotic Species in the Sonoran Region. Univ. Arizona Press and the Arizona-Sonora Desert Mu- seum, Tucson, Arizona —- & . 2002b. Flood dina and restoration of riparian vegetation in the American Southwest. Pp. 11-49 in B. Middleton йк, Flood Pulsing in Wetlands: heb the Nat a Hydrological Balance. 1999, Bun finds adjacent land uses are key to predicting the number and type of exotic species in forest gaps (Maryland). Ecol. Restoration 17: 159-160 Thysell, D. arey. 2001. DN of density of Наш иы mennziesii canopies: Prelim effects on understory vegetation. Canad. J. Forest Re. ЗІ: 1513- ые Tunison, J. T., С. M. D'Antonio & R. K. Loh. 2001. Fire and invasive Jupe in Hawai'i Volc ‘апоез National Park. Pp. 122-131 in K. E. M. Galley Wilson (edi- ча Ргос ae ve of the Bos Species Workshop: The Role of Fire in the Control and Spread of Invasive Species. Fire Conference 2000: T di First National Con- gress on Fire Ecology, Prevention, and Management. Misc. Publ. No. "a Tall Timbers e h Station, Tal- {. Turkington & D. Mcellvride. 1986. The biology of C ы. weeds 75. Bromus tectorum 1.. Canad. J. Pl. Sci. 66: are 9. Vila, M., E. Weber & C. M. D. D'Antonio. 2000. Con- servation implications ж invasion by plant hybridiza- tion. Biol. Invasions 2: 207-217. Vitousek, P. M. 1990. Biological invasions and ecosystem processes: Towards an poit of population biology and ec osystem studies. Oikos 57: 7— . M. D'Antonio, L. L. p & R. Westerbrooks. ns as global environmental w Feral pig removal: Effects on soil e m du a Hawaiian rain forest. J. Wild 2 Y. Collingham, S. С. Willis, B. Huntley & P : jo 2000. Simulating the spread i i s: Are they 38. 2000. Climatic forcing on the dispersal of exotic speci ies. Phytocoenologia 30: 409—430. Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips & E. in the United Stat ioScien | Williamson itter. 1€ The varying success of du Ec iology 77: 1661- 1666. Young, J. A. & R. A. Evans. 1985. Demography of Bro- mus tectorum in Artemesia communities. Pp. 489-502 in J. White (editor), The Population Structure of Vege tation. Dr. W. Junk Publishers, Dordrecht, The Neth- erlands Internet je 'es [Вепѕоп, A. J., P. L. Fuller . C. Jacono. 2001]. Sum- mary Report P nd Aquatic Species in U.S Fish and Wildlife Service Region 4. U.S. Geological Survey, Florida Caribbean Science Center. (http://ww fesc.usgs.gov/R4finalreport.pdf); viewed 30 September 2002. M M., . Connors & A. V. Slee. 1999]. F CLID: Euc valypta of South-East Аааа: DE anbg.gov.au/cpbr/euc Шш lid.html); viewed Septem- ber 2002. [Convention on Biological Dive sity. (http://www. biodiv. Moe onvention/articles.asp); viewed 30 September See Pam (2001) in L nales Cited. [Langeland, K. A. & addock Burks (editors). 2000]. Identification N pos of Non-Native Plants in Flori- da's Natura . IFAS Publications, No. 257, Uni- (http://aquat] .ifas.ufl.edu/identif. 02. [National Invasive Species Council. 2001]. Meeting the Invasive Species Challenge: о Invasive Species Management Plan. (http: beer aot ies.gov/council/ mpfinal.pdf); ar 20 September 2002 [Nuzzo, V. A. 2000]. Element Stewardship Abstract for Alliaria Heros ш officinalis) garlic mustard. (http://tnewee ur nnts/allipet. html); viewed : e ‘tober 2002. ). Raven, P. H. Raven, S. Reichard & . Linking EY d horticulture dhttn-// Карате Т tej ds.ucdavis to prevent plant i i ss); viewed 30 September 2002. [San Francisco Estuary Invasive Spartina Project]. uu //www.spartina.org.htm); viewed 2 23 April 2002. [TNC, a. 2000]. Me ensures across це : А рге liminary summary. (htt )00/12/b/ mossum;inte е »tion=buildframes action); viewed 30 September 2002 [TNC, Ы. Нож we жаг Setting priorities. ا‎ RE / nl); viewed September 2 002. [TNC, Invasives on the web. (http: бы us eds. ucdavis. ШЫ ш. html); viewed September [TNC, d]. Weed Management Plan for The qe Riv- er Preserve, Galt, California, 2001-2005. (http:// Incweeds.ucdavis. не ts/plans/CRP-Plan.doc); vie gs ы uw 20 (Tu, M. & Rice. dur 2002]. Early detection “ quic i du of a new invader in southeast Alas (http: //tneweeds.ucdavis.edu/success/ak001. html): vie s 7 October 2002. J.S. Endangered Species Act of 1973] (http:// айаш. fws.gov/esa.html); viewed 30 September 2002. = [Williams, J. a . Meffe. 1996]. Nonindigenous )ecies. Stat Nation's Biological Re 'sources. On- line publication of the U.S. Geological Service. (http://biology.usgs.gov/s+t/SNT/index.htm); viewed 20 ede 2002. PLANT NATURALIZATIONS Richard М. Mack? AND INVASIONS IN THE EASTERN UNITED STATES: 1634—1860! ABSTRACT Plant immigrants to North America arrived from Europe with the first human immigrants, products of the intense incentive early colonists felt to transplant European agriculture into the Western Hemisphere. iiy early deliberate and accidental introductions were species that would soon become naturalized in eastern hele America: Artemisia absinthium, Hyoscyamus niger, Plantago lanceolata, and Taraxacum officinale. The naturalized flora grew as species for food, ا‎ seasonings, and medicine were imported, cultivated, and escaped the bounds of c бее мы fields. Importation of what become the most common category of naturalized species, erstwhile ornamentals, had a modest beginning by the mid Ih century. The first recorded invasion, the spread and un of | Linaria opea: in the Mid-Atlantic colonies, was recognized by the mid 18th century, and Berberis vulgaris was ran ipant in southern New gn before 1800. Botanical records, including published € became much more common in the first decades of the 19th century and reveal a naturalized flora in the d that was quite similar in composition to the agricultural weed flora of Western Europe. Many ruderals and agricultural weeds were widespread in the eastern U.S., but probably not invasive by 1860, and included Bromus secalinus, (ences officinale, Galium рагы, and Senecio p vularis Other alien species had, however, become invasive by the 1 s, such as Echium vulgare in Virginia. Species that were to form pron invasions in the peii States from 1860 onward (e.g., Bromus tectorum, Euphorbia esula, Lonicera japonica, Me laleuca quinquenervia) had e not arrived by 1860, were undetected, or were not reported as having escaped from cultivation. Growth of the a яя flora and the ae mber of invasive taxa was certail any facilitated, and probably sparked, by the enormous growt of railroads and rail-borne commerce in the late 19th century Key words: Berberis discum biological invasion, John. Bartram, Echium vulgare, Linaria vulgaris, naturalization, ornamental plants. ‘<. . Behold, 1 have given you every herb bearing = World. From the outset, each group transferred ш is upon the face of all the earth and every tre in the which ts the fruit of a tree yielding seed; to you í ў А : it shall be for meat. based their resolve to establish European agricul- — Genesis 1: 29. Bible, King James version ture in the New World (Mack, 2001, and references therein). These determined attempts by European colo- crops and domesticated animals upon which they The first European human immigrants to North America had special reason to believe fervently їп ias and their American descendants to introduce these words of Scripture. In transporting their fam- ilies to a New World about which they knew little, they needed to believe that a Higher Authority would provide for their welfare. Most also firmly non-indigenous plants in what was to become the United States have had profound and lasting influ- ence on the floristic composition of North America’s natural communities, and have largely determined which species have proved troublesome in agricul- ture, forests, rangelands, and inland waterways and European colonists to North America, whether the lakes (Westbrooks, 1998). The size of the current 16th century Spanish or Portuguese colonists (Cros- believed that God provided through the deliberate planning and industry of the faithful. Each band of Бов, | naturalized, i.e., permanent non-indigenous, flora of by, 1972), 17th century English colonists to Mary- ihe U land (Hall, 1910: 92), or the Pilgrims to New Eng- land (Young, 1846: 42), did not trust their survival to the happenstance acquisition of food in the New .S. is not known precisely but exceeds 2500 species (Kartesz & Meacham, 1999). Of these, a small minority have become invasive, i.e., prolific and occupying significant new (i.e., non-native) ! L thank Sarah E. Reichard, Peter S. White, Peter Н. Raven, and Mick Richardson for organizing the excellent 48th Missouri Botanical Garden Symposium that ни the opportunity for this paper. I thank the staff of the libraries of Washington State University, the Academy of Natural Sciences (Philadelphia), the New York Botanical Garden, the Library of Congress, and the National Agric cultural L ibrary, as well as R. A. Black and C. L. Kinter for discussions of the ideas that are developed here. The manuscript жү. from the conscientious comments of an anonymous reviewer, pie Craddock Burks, Victoria C. Hollowell, and A. McPherson. School of Biological Sciences, Washington State Unive rsity, Pullman, Washington 99164, U.S.A. rmack@mail. i ANN. Missourt Bor. GARD. 90: 77—90. 2003. 78 Annals of the Missouri Botanical Garden ranges. These species, along with many other nat- uralized species, have caused enormous environ- mental and economic damage that ranges from threatening native biodiversity to changing the op- eration of eg ecosystems (Mack et al., 2000; Pi- mentel et al., 2000). European ae ae in North America arose from fledgling colonies of immigrants that stemmed from different countries in western Europe and were established at different times and at widely spaced locales (Quinn, 1990). This mixture of cir- cumstances and events suggests that the current naturalized flora has had a diverse origin and var- ied circumstances surrounding its eventual persis- tence. This diversity spawns a series of questions. What has been the chronology of plant naturaliza- tions? More importantly, what can be deduced from that chronology about the circumstances that led some species to become naturalized early in the Europeans’ history in North America, while other species became established much later? Which species, if any, formed pre-1860 invasions, and what sparked the demographic transitions among those few naturalized species that became invasive (Crooks & Soule, 1999; Kowarik, 1995)? What role have species introduced long ago into the U.S. con- tinued to play in their new ranges? For instance, have they been supplanted by more recently arriv- ing immigrants (Mack, 1989)? I chose to examine these issues for the period between 1634, the first records I am aware of for plant naturalizations, and 1860, before onset of the American Civil War and the enormous changes in the economy and trans- portation that so affected the U.S. for the rest of the 19th century and beyond (Kuznets et al., 1960; Meinig, 1986). These questions about plant entry, establishment, and their potential proliferation in new ranges must be considered because species’ past immigration histories form part of the infor- mation upon which the prediction of future inva- sions will continue to rely (National Research Council, 2002). METHODS Any assessment of species that became natural- ized long ago is intrinsically retrospective, and the record is incomplete and often fragmentary. Four general sources that vary substantially in potential accuracy and detail can be used in assembling this record. (1) Herbarium specimens provide unequivocal evidence for accurate species identifications. Consequently, they may form the best cate- gory of evidence, provided the collection dates and locales are accurate. Unfortunate- ly, the number of existing herbarium speci- mens across the time period under consid- eration here is meager. Few pre-1860 U.S. specimens still exist; even fewer have sur- vived from pre-1800. Their use remains an unfulfilled goal here. Pollen records can provide some supplemen- tal information (Brugam, 1978) but are lim- ited in their usefulness because many pollen N м taxa are impossible to identify with light mi- croscopy and for some families, such as Po- aceae, only a few genera are possible to iden- tify in this way (Moore et al., 1 Contemporary regional floras became much more common after 1800, e.g., Pursh (1814) and Torrey (1819), and can be enormously RN o — informative. 4) The most common tool examined here con- sists of contemporary records and correspon- dence of those who provided first-hand ac- counts of the species they saw, their ranges, and abundances (e.g., Josselyn, 1672; de 1832; Dwight, [1821] 1969; —. Schweinitz, Gray, 1842). Herbarium specimens, floras, and other records vary substantially in accuracy and scope, and until the later part of the 18th century. ipd all em- ployed only common names. | have cross-checked the common names (e.g., Fernald, [1950] 1987; Kar- tesz & Meacham, 1999) and avoided a species ci- tation unless I could be confident of the identifica- tion. І have been similarly conservative in assigning an areal extent to any non-indigenous species. For- tunately, the same common names have been used consistently for many European ruderals and medic- inal plants for more than 400 years, e.g., elecampane (Inula helenium L.), henbane (Hyoscyamus niger L.), horehound (Marrubium vulgare L.), and St. John's wort (Hypericum perforatum L.) (Harvey, 1974). In addition, a surprisingly high percentage (> 75%) of the species encountered in these early records were first described by Linnaeus, and many of these names have been retained in plant systematics. Giv- en the limitations in the information that can be re- liably deciphered about pre-1860 naturalizations, the resulting record will, however, remain a mini- um estimate of the full scope of plant entry and establishment. Kartesz and Meacham's (1999) Syn- thesis was employed as the nomenclatural standard for plant names and authorities. FIRST PLANT ARRIVALS: PRE-1700 Preparations for colonization of new lands North America illustrate the care and forethought Volume 90, Number 1 2003 Mack 79 Plant Naturalizations and Invasions that the immigrants placed in the undertaking. Iso- lated on a continent newly discovered by Europe- ans, connected to Europe only by infrequent and highly unreliable ship traffic, each colony's plan- ners knew that they needed to become self-suffi- cient almost from the outset. Acquiring sustained sources of food and medicine figured foremost in their planning (Leighton, 1970). The most reliable early information we have of plans for colonizing North America stems from the establishment. of Plymouth Colony, beginning in 1620. Although we do not know which species were in the original manifest, several lists of plants ordered in Britain from 1628 and 1631 have survived. The Endicott expedition in 1628 was to include “Wheat, rye, barley, oats, a hoghead of each in the ear, beans, pease, stones of all sorts of fruits, as peaches. plumes, filberts, cherries, pear, apple. quince, ker- nels, pomegranates, woad seed, saffron heads, li- quorice seed, (roots sent and madder roots.) pota- toes, hop roots, hemp seed, flax seeds against winter, coneys, currants plants ...” (Young, 1846: 42-43). This intent to send woad seed (/satis tinctoria L.), hemp seed (Cannabis sativa. L.), and flax seed (Lin- um usitatissimum L.) forms the earliest record we have of the introduction of non-indigenous species that remain in the naturalized or at least the adven- tive flora of the U.S. Dyer's woad, /satis tinctoria, has not apparently become widely naturalized in New England, despite its exceptionally early arrival (Ma- gee & Ahles, 1999: 552), but it has become invasive in Utah and other arid regions (Farah et al., 1988). Cannabis sativa is widely naturalized in the eastern half of the U.S. north of the 37° latitude (Haney & Bazzaz, 1970). In contrast, flax appears commonly as an adventive in abandoned fields but may be nat- uralized only locally. n 1631, just three years after the Endicott Ex- pedition, John Winthrop, Jr. bought a detailed group of seeds from a London grocer with the intent of transporting these seeds to the small colony at Plymouth. Winthrop’s (Massachusetts Historical So- ciety, 1943: 47-48) list also included species that have since become naturalized, including “Sorrell” (Rumex crispus L. or Rumex acetosella L., or both), “Tansy” (Tanacetum vulgare L.), “Rockett” (Eruca vesicaria subsp. sativa (P. Mill.) Thellung [Eruca “Buglos” (e.g., Anchusa arvensis (L.) “fennell” (Foeniculum vulgare P. Mill.), sativa |). Bieb.), “dill” (Anethum graveolens L.), ruonum majorana L. or Origanum vulgare L.), | di nipp" Nepeta cataria L.), or adventives, such as “summer sauory" (Satureja hortensis L.), “Clary” (Hyssopus officinalis “sweet maioram” м (Salvia sclarea L.), “hysopp” L.), *marigold" (Calendula officinalis L.) and “hol- lihocks" (Althaea rosea L.) (Fernald, [1950] 1987; Magee & Ahles, 1999). The identification of other species on his list is more equivocal, such as “mal- low," which L., Malva neglecta Wallr., Malva verticillata L., or Malva parviflora L. The listing of “popey” (Mas- sachusetts Historical Society, 1943: 47) may refer to Papaver somniferum L., may have referred to Malva moschata the opium poppy, but could also be referring to Papaver rhoeas L., which had reputed medicinal properties. Winthrop’s list also includes “pursland” or purslane (Portulaca oleracea L.). It is often considered non-indigenous to North America (Magee & Ahles, 1999), but ap- pears in the fossil pollen and seed record of Ontario in pre-Columbian levels (Byrne & McAndrews, 1975). Although these species appear to have ar- rived with the first waves of colonists in New Eng- land, the earliest date when they became natural- ized or even adventive is not known. Our most comprehensive picture of the species introduced by 17th century colonists in New Eng- land is derived from Josselyn’s 1672 publication New-England's Rarities Discovered and its 1674 se- quel An Account of Two Voyages to New-England. Based on Josselyn’s accounts of two visits, 163 1639 and 1663-1671 (Josselyn, [1674] 1988: xiii), he appears to have been a keen observer of the condition of the colonies in New England. Further- more, he deliberately categorized the species that he encountered in his travels between Massachu- setts and Maine, noting those that were apparently native to New England, those species also found in England (and apparently introduced), and even those introduced species that did not thrive in their new locales. Among species that he reported that later became naturalized or at least adventive were many that are not among the manifests of Endicott or Winthrop. The naturalized species include cel- andine (Chelidonium majus L.), goose-grass or *clivers" (Galium aparine L.), *Our English Clover- grass" (perhaps Trifolium repens L.), speedwell chickweed (Veronica arvensis L.), stitchwort (Stel- laria graminea L.), St. John’s wort (Hypericum per- foratum), sweet brier or eglantine (Rosa eglanteria L.), toadflax (Linaria vulgaris P. Mill), (Artemisia absinthium L.), and yarrow (Achillea mil- lefolium L.). Among species now adventive are blueflowered pimpernel [Anagallis arvensis subsp. foemina (P. Mill.) Schinz & Thellung], “egrimony” (Agrimonia eupatoria L.), herb Robert (Geranium robertianum “Oak of Hierusalem” (Chenopodi- um botrys L.), speed-well (Veronica officinalis L.), spurge time (Polygonum persicaria L.), “Rew” (Ruta graveolens L.), watercress (Rorippa nastur- wormw Annals of the Missouri Botanical Garden tium-aquaticum (L.) Hayek), and wild-mint ы tha aquatica. L.) (cf. ii? ^ Ahles, 1 thermore, Josselyn (167 85) pa на step of noting eus h species had ar- ved *. . . since the English Planted and kept Cat- "i in hae England” that were already occurring spontaneously without cultivation. In this list, Jos- selyn provided explicit information on some of the too first naturalized species (see below). The diversity of species in Josselyn’s list reveals that by 1671 (his last year in North America), and perhaps much earlier, the small lists of species that Endicott and Winthrop had ordered for the settlers had been expanded severalfold. Most prominent are species that had reputed medicinal value (Artemisia absinthium, Hypericum | perforatum, Inula helen- tum), as well as plants for seasonings (Anethum graveolens, Foeniculum vulgare P. Mill, Salvia sclarea). Perhaps most surprising is that the colo- nists were beginning to feel confident enough about their survival that they permitted themselves the luxury of importing a few ornamental species, Lin- aria vulgaris and Rosa eglanteria (Leighton, 1970). Both of these species would become naturalized, and L. vulgaris would become a scourge by the mid 18th century. The arrival of these species is also significant because this early entry of ornamental species signals a trend that would grow substan- tially over the following 350 years as ornamental species became the largest single functional cate- gory of imported species (Mack & Erneberg, 2002). EMERGENCE OF THE FIRST NATURALIZED SPECIES: PnE-1700 Our ability to detect the earliest naturalizations is severely limited by the fragmentary character of the first accounts of European colonists in North America. For example, the late 16th century Span- ish settlements in Florida included gardens and cultivated limes, lemons, and oranges (Lyon, 1996: 25); contemporaneous introductions of oranges soon appear in other Spanish subtropical and even tem- perate colonies (Gade, am, however, un- aware of any information on the fate of these spe- cies that were cultivated in 16th century Florida. Maine also received European colonists very early: successive temporary settlements were made begin- ning in 1604. The record referred to as the Edger- ton Manuscripts (Lockwood, 1931: 139) is intrigu- ing in describing the author's to a long-abandoned settlement in Maine *on the Per- maquid River Alderman Alsworth of Bristole set- tled a co., of people in 1625. ... In 1675 I found the Roots and Garden Herbes, and some old walls visit there when I went first over, which showed it to be the place where they had been." Finding garden herbs many years after the site's abandonment strongly suggests that these species had become naturalized. The first, admittedly tenuous, records of natural- ized species in what was to become the U.S. appear in the decades after establishment of English col- onies in New England. Among the praiseworthy features of these English immigrants was their com- mitment to recording a wide range of information about their nascent settlements, including the fate of their crops. Only 14 years after the first colonists landed in Plymouth Bay, William Wood (Wood, [1634] 1977) had published his account of the "New England's Prospect of information that would have proven invaluable to anyone intending to immigrate to New England. Wood's comments entitled *Of the Herbes, Fruites, Woods, Waters and Minerals" prove particularly in- formative about both the European crops being planted and also those already observed to persist ": a comprehensive report outside cultivation. Wood’ ([1634] 1977: both native and introduced, reveals species that had already been brought under cultivation. and their status: "The ground, affords very good kitchen Gardens, for turneps, Parsnips, Carrots, Radishes, Muskmillions, Isquoutersqashes, 36) listing of the crops, and Pumpkins, Coucumbers, Onyons, and whatever grows well in England grows well there, many things being better and larger; there is likewise all manner of Herbes for meate, and medicine, and that not only in plant- ed gardens, but in the Woods, without eyther the art or the help of man, as sweet Marjoram, Purselane, Sorrel, Peneriall, Yarrow, Mirtle, Saxifarilla, Bayes, etc." (italics added). The italicized statement strongly suggests that several of the species that the colonists had planted in the previous decade had already escaped cultivation and were growing freely in the surrounding forest. These species include sorrel (Ките and Rumex crispus) and yarrow (Achillea millefolium). These are the earliest accounts | am aware of for any naturalizations in North America. Identification of *sweet Marjoram" is equivocal. Wood may have been referring to Origanum majorana, which is still referred to as x acetosa L. sweet marjoram, but has rarely escaped cultivation and is not considered peg in New England today (Magee & Ahles, : 898). Alternatively, he may have seen SE DE now termed wild pip which is naturalized (Magee & Ah- les, 3). “Peneriall” in Wood’ list may refer to the native species Hedeoma pulegioides (L.) Pers. rather than the European species, Mentha pulegium Volume 90, Number 1 2003 Mack 81 Plant Naturalizations and Invasions L., for which there is no record of naturalization in the U.S. These earliest references to the naturalized sta- tus of some newly arrived European species in New England are substantially corroborated and expand- ed by Josselyn (1672). His list *Of such Plants as have sprung up since the English Planted and kept Cattle in New-England" is the best account from which to determine the European species that had become naturalized by 1671 and probably much earlier (Josselyn, 1672). Many of the species that he reported in terms indicating their newfound per- sistence in the New England flora are still promi- nent today (Magee & Ahles, 1999), such as Hyo- scyamus niger, Stellaria media (L.) Vill., Taraxacum officinale G. ebber ex Wiggers, Senecio vul- garis L., and Urtica dioica L. Josselyn's (1672: 86 report that plantain (Plantago major L.) was re- “English-Man’s м ferred to by Native Americans as foot" has been quoted repeatedly. Plantain was al- ready spreading along paths and roads well ahead of English settlements and served as a harbinger to the aborigines of the coming of the new settlers (Darlington, 1859: 219). Aside from the immediate needs for importing plants for food and medicine, the early colonists had definite interest in ensuring that their livestock had suitable forage. They soon found, however, that the coastal environments offered their animals little nutritious forage. А common complaint is exempli- fied by one colonist in Massachusetts who stated that his livestock “. . . grew lousy with feeding upon it, and are ichs out of heart and likeing (Hutchinson, [1764] 1972: 483). Help was on the way. These transplanted English farmers and herds- men were well acquainted with an array of “English grass" that would meet their livestock's needs, and they actively sought seeds of these English pasture species for introduction into their farms. Species identifications in these 17th century accounts are confounded by the common reference to a mixture of grasses and clovers as “English grass,” and it is likely that some of these species could have arrived alternatively as seed contaminants in lots of seed or in the ballast or debris off-loaded from ships with livestock, or both (Bidwell & Falconer, 1925). Spe- cies that were introduced through the 17th century were die de aviculare L., Holcus mollis L., Poa pratensis L., and possibly Agrostis capilliaris L. (Carrier, 1923: 241), and Holcus lanatus. Not only were these species becoming naturalized in New England, but they were also being actively spread in colonies to the south. In praising the growth of cattle and other livestock on Long Island (New York), Denton (1670: 5) commented that “the Is- land likewise [was] producing excellent English grass, the feed of which was brought out of Eng- land, which they sometime mow twice a year.” And Budd (1685: 10) commented on the practice of us- ing sheep to disperse these grasses “but if we sprin- kle a little English grass Hay-Seed on the Land without Plowing, and then feed Sheep on it, in a little time it will so increase, that it will cover the Land with English Grass, like unto our Pastures in Englanc Манса ра arising from the introduction of forage species were not restricted to grasses. The advantages of sowing pastures with clovers, espe- cially Trifolium repens (white clover), were common knowledge in England by the mid 17th century, and this knowledge was widely transferred to North America. As early as 1635 a tract written for those who planned to immigrate to Maryland advised that they bring “Good store of claver grasse seede, to make good meadow" (Hall, 1910: 98). This sound advice had also been followed in New England: Jos- selyn ([1674] 1988: 131) commented that “Our En- glish clover grass sowen thrives very well." It was both deliberately sown and spread as sheep were moved among fields, as Scot (1685: 187—188) re- corded for New Jersey: *As soon as any of the land here comes to be cultivated, it over-runs with small Claver-grass, by the pasturage and dunging of the cattle, and so supplants the naturall grass and hearbs.” The among the colonies was extending the new ranges movement of livestock and seed for these European pasture species. By the mid to late 17th century at least 20 spe- cies were observed to have already become natu- ralized in New England; many of these had likely become established further south in Pennsylvania. The actual number of naturalized species was likely much larger. For example, Josselyn (1672) listed under his category *Of such Plants as are common with us in England" many species that were already in New England and for which we have records of naturalization in the 18th century. How long before 1700 th cies in this category of 17th century introductions ey were naturalized we do not know. Spe- that were to become permanent residents include Artemisia absinthium, Cannabis sativa, Сайит aparine, Hypericum perforatum, Inula helenium, Nepeta cataria, and Xanthium strumarium PLANT ARRIVALS AND NATURALIZATIONS IN THE 18TH CENTURY Any chronological divisions, such as those be- tween centuries, that could be used to delineate the growth of a naturalized flora are arbitrary. The forc- 82 Annals of the Missouri Botanical Garden es that brought new species to North America and contributed to their naturalization were obviously not so coincidentally partitioned as to conform to even decadal, much less centennial breaks. Nev- ertheless, there is some justification for distinguish- ing between the naturalizations in the 17th and the 18th centuries. Leighton (1976: 1) has argued that the necessity of establishing self-sufficient colonies in the 17th century was a powerful incentive to ensure that almost all the species introduced before 1700 were for utilitarian purposes. As noted above, some plants introduced even before 1650 may have had no purpose other than ornamentation, such as К. eglanteria. Ви! even eglanteria may have been deliberately introduced to form hedges (Jos- selyn, 2: 90). In contrast, the assurance that the colonies would not only survive but also grow and prosper sparked much incentive in the 18th century for the introduction of species for the widest range of uses, especially for ornamentation. This market in ornamental species grew steadily throughout the 18th century and thereby created opportunity for more naturalizations, and even invasions (Lock- wood, 1931: 12; Leighton, 1976). Perhaps the best single view we have of the growth of this naturalized flora in the 18th century was a retrospective written in the early part of the 19th century but clearly drawn from information gathered in the late 18th century. Rafinesque, an itinerant French botanist, collected plants widely in the new United States. In 1810 he assessed the naturalized flora of the Middle Atlantic states (Raf- inesque, 1811). His work appears to be based largely on Muhlenberg's (1793) flora of the area surrounding Lancaster, Pennsylvania, /ndex Florae Lancastriensis, which Rafinesque augmented and annotated to include his own observations for the North American region that stretched south from New York State to Maryland. The most specific lo- cation information is given for species occurring near Lancaster, Pennsylvania, and those found in the vicinity of Baltimore, Maryland, Rafinesque's home in the U.S. In a useful pattern that was to be repeated by other authors in the 19th (1811) categorized the species on their mode of in- century, Rafinesque troduction: plants introduced by agriculture, plants "totally plants or weeds, i.e., those accidentally introduced. introduced by gardening, and useless' Plants in a fourth category for him had varied modes of introductions and were not native to Eu- rope. А great advantage of this Muhlenberg cum Rafinesque list is the consistent use of binomial names, most of which have been retained in modern plant systematics. In addition, Rafinesque gave his assessment of the abundance of the species, mak- ing it clear that a few (Convolvulus arvensis L., Leu- canthemum vulgare Lam. [which he termed Chry- santhemum leucanthemum], | and Verbascum blattaria L.) were both widespread and abundant. mong deliberately introduced species are Can- nabis sativa, Hordeum vulgare L., Linum usitatis- simum, Plantago lanceolata L., and Trifolium pra- tense L., which were all listed as common to very common, usually in ruderal site f the more than 300 species in Rafinesque's (1811) list, he considered the largest single group (193 taxa) to have been introduced in gardening, i.e., species grown in small plots and presumably cultivated much more assiduously than field crops (e.g., Gleochoma hederacea L., Sinapis alba L.). The second category reveals the extensive naturalization of medicinal plants and those used for seasonings in the 18th century: Asparagus officinalis L., Cy- noglossum officinale L., Digitalis purpurea L., Inula helenium, Marrubium vulgare, Nepeta cataria, and Tragopogon porrifolius L. Most of these species were listed as common and confined to roads, near dwellings and gardens. Among this large list of spe- cies are also those such as Cichorium intybus L. (chicory), which was listed as “very common—in EE] fields, roads, cultivated grounds. . . ," and Lamium amplexicaule L., considered, “every where com- mon—in fields.” Many more species had been imported for strict- ly aesthetic reasons in the 18th compared with the 17th century, and among these were some that soon became naturalized. Ornamental species that ha already become naturalized ca. 1800 in Rafin- esque's opinion included Euonymus europaea L., Ligustrum vulgare L., Rhamnus cathartica L., and Syringa vulgaris L., illustrating that woody orna- mental species were also becoming persistent. The remainder of the species noted by Rafinesque were considered accidentally introduced, such as Bro- mus secalinus L., Echium vulgare L., which was to become much more conspicuous later, Agrostemma githago L., Chenopodium album L., Convolvulus ar- vensis, and Spergula arvensis Rafinesque (1811) may have applied the appel- “naturalized” rather liberally, as he used it to mean those species that “... now grow sponta- neously ...,” not nec ا‎ persistently. For ex- ample, he listed some species, such as Fagopyrum esculentum Moench (buckwheat), as naturalized and common, yet today it is probably only adven- tive as an escape from cultivated fields. Alterna- tively, its status may have indeed changed in the last 200 years (e.g., from diminished cultivation), or he may have been simply noticing volunteer Volume 90, Number 1 2003 Mack Plant Naturalizations and Invasions buckwheat that was residual in fields. Another spe- cies he considered rare but nonetheless naturalized was Rubia tinctoria L. (madder), a species referred to among some of the earliest plant import mani- fests in the 17th century (Young, 1846: 42) but which is considered as only a rare escape from cul- tivation today (Magee & Ahles, 1999). Neverthe- less, the bulk of the species that he noted as nat- uralized are indeed persistent today, and his assessments were corroborated by his contemporar- ies. EMERGENCE OF THE FIRST PLANT INVASIONS: 18TH CENTURY The major consequences of non-indigenous spe- cies result from the small minority of naturalized species that become prolific in the new range, i.e., these species become invaders. Their abundance and aggressive growth bring about environmental damage to the native species and alteration of na- tive environments (Mack et al., 2000). These spe- cies often invade arable fields and pastures and consequently result in severe economic damage (Bridges, indigenous specie 1992). Given the growing array of non- s that were being introduced both accidentally and гети from the early 17t century onward, it is not surprising that a few were eventually reported in terms that we would equate with plant invaders. The earliest invasion of which I am aware re- sulted from the introduction of Linaria vulgaris P. Mill (yellow toadflax). In the extensive correspon- dence that John Bartram, the doyen of 18th century American botany, maintained with his colleagues Peter Collinson and Philip Miller in Britain (see Darlington, [1849] 1967), his remarks about L. vul- garis are revealing. In a report that both colleagues apparently received accompanying a letter in 1758, Bartram stated, “It was first introduced as a fine garden flower; but it was never more heartily cursed by those that suffer by its encroachment” (Darling- ton, [1849] 1967: 384). He added, “It is the most hurtful plant to our pastures that can grow in our northern climate. bipes: the spade, plough, nor oe, can eradicat when : is spread in a pas- ture" (Darlington, ides 7: 383). Bartram ex- plained that so desperate were farmers to control yellow toadflax that they would even ignite log piles in a field in the hopes of destroying it in the soil but to no avail. Bartram reported that by 1758 L. vulgaris had “. . . spread over great part (sic) of the inhabited parts of Pennsylvania" (Darlington, [1849] 1967: 384)—a clear indication that it was not simply a local problem. Linaria vulgaris had formed an invasion. It apparently spread further afield: Pursh (1814) noted that it had become * one of the worst and most troublesome weeds in several parts of Pennsylvania and Virginia." Barton (1818) went even further in his estimation of its spread as he rated it (which he termed as Antirrhi- num linaria) as “. . . extensively naturalized, in the United States. On roadsides, commons, wastes, and the borders of fields, very common and abundant." Yellow toadflax had reached North Carolina by 1832 because de Schweinitz (1832) reported that a few years after L. vulgaris (as Antirrhinum linaria) was introduced into a garden it had *. . . contami- nated the whole vicinity for many miles." Other naturalized species were also causing se- rious problems, although the extent of their new anges is more difficult to determine. Bartram lived in n Philadelphia but had numerous correspondents along the East Coast, so some of of the damage from naturalized species may have is assessments been drawn in part from the reports of others. For example, dence listed other non-indigenous species in what appears to be a declining order of prominence. He described en perforatum as a "very perni- which eastern Pennsylvania, interfering with the growth of cious weed." had spread over pastures in pasture grasses and causing injury to horses and sheep (Darlington, [1849] 1967: 384). This plant was apparently introduced repeatedly in New Eng- land and the Middle Atlantic colonies (Josselyn, 1672: 44; Haughton, 1978: 348), so it may well have formed an invasion by the mid 17th century. Hypericum perforatum was also proving to be troublesome plant in fields in New England. Eliot. writing in his fourth essay on agriculture, which was first published in 1753, complained about the difficulty of eliminating St. John's wort in fields (Carman & Tugwell, 1934: 94). Even though Eliot lived in Connecticut, he traveled widely in New England, and his essays were meant to be advice gleaned from decades of observation in the region (Carman & Tugwell, 1934). If H. perforatum had not reached the status of an invasion by the mid 18th century, it was at least a widely distributed naturalized species. inaria vulgaris and Hypericum perforatum be- deviled 18th century farmers across a broad region from Pennsylvania to New England, but these spe- cies' prominence appears to have since declined. Darlington (1859: 225) reported that L. vulgaris was “extensively naturalized” and a “vile nuisance in our pastures and upland meadows." He described H. perforatum as a "rather troublesome weed on our farms," but did not state it in terms that suggest an 84 Annals of the Missouri Botanical Garden invasion. He added the intriguing note that in Chester County, Pennsylvania, he noticed that the plant was not detected at all in 1842, and was rare in 1843, but had become “as common as ever" it subsequent years (Darlington, 1859: 55). This statement may allude to the prominence reported by Bartram in the 18th century (Carman & Tugman, 1934). Although widely naturalized in the northeastern es, — i; oads ein, 3; , 1999), neither H. perforatum nor L. pears invasive today. However, these reduced roles are largely the result of active control measures. Linaria vulgaris is controlled with the herbicide glyphosate in agricultural fields (Saner et al., 1995), while the abundance of H. perforatum has been ef- fectively curbed in much of its new range through biological control (Julien & Griffiths, 1998, and ref- erences therein). The difficulty in deciphering from early accounts whether species had become inva- vulgaris ap- sive relates to the specific interests of the observer. Eliot and Darlington were primarily interested in species that were hazards to agriculture. Even Bar- tram's observations appear often influenced by his concern about agriculture. Alien plants that were extensive exclusively in non-agricultural settings may not have been commented on. As damaging as L. vulgaris and H. perforatum could be, colonial farmers already had a much greater scourge to cope with—Berberis vulgaris L. (common barberry), the alternate host for the stem rust (Puccinia graminis f. sp. tritici), a devastating parasite of cereal crops (Peterson, 2001). Berberis vulgaris had certainly reached the level of impact to qualify as an invader by the late 18th century and had probably reached that status over a century earlier. Our knowledge of its probable entry into North America by the mid 17th century is through a combination of its direct mention and reports of the occurrence and spread of stem rust as it in- fected wheat across New England. Common bar- berry was almost certainly introduced deliberately in the 17th century, as it was valued in sauces and as a medicinal plant (Gerard, [1633] 1975: 1326). Josselyn (1672) referred to “barberry trees” in a list of introduced fruit trees and also described the rust (termed “wheat blast" until the 20th century) on wheat in New England in the 1660s. A more de- tailed account of the incidence of stem rust was provided by John Winthrop in 1668: “generally through all the plantations, both of ye Massacheu- setts colony, Plymouth, & this also [the colony of Connecticutt] insomuch that the croppe of wheat hath failed divers yeares in most plantations. The corne flourished well till it came to be eared, and the eares also would appeare faire, and as if full, but no corne in them. There have beene thousands of acres in that maner every yeare. What the cause was, whether naturall, or a blasting fro heaven we know not. Our old husbandmen of England, some of them thought it a meldow .. .” (Bidwell & Fal- coner, 1925: 13). Given the aia link between wheat, stem rust, and barberry, it seems a safe in- ference that where rust was attacking wheat, bar- erry was nearby. As further evidence for the spread and impact of Berberis vulgaris, Connecticut passed legislation in 1726 to control barberry, followed by Massachu- setts and Rhode Island in 1755 and 1772, respec- tively (Fulling, 1943). These measures failed, as Dwight ([1821] 1969) provided direct observation in 1795 or 1796 of the extent of B. vulgaris across much of eastern. Massachusetts and coastal New Hampshire. Within the approximately 3000 km? area that Dwight circumscribed in his travels he noted “... the barberry bush is spread, not uni- versally, but in spots, and those often extensive. In some fields they occupy a sixth, fifth, and even a fourth of the surface" (Dwight, [1821] 1969: 276). Clearly, barberry was exerting a major influence on wheat production across eastern. Massachusetts in the 18th century to the point that bread made from wheat had disappeared from farmers' diets in much of New England (Bidwell & Falconer, 1925: 92). Common barberry would later spread much further across the U.S., and by the time control efforts were fully implemented against it early in the 20th cen- tury, it was extensive in a 13-state area in the North Central region of the U.S. (Hutton, 1927). Other biotic invasions were growing in North America by the late 18th century, e.g., the spread of Trifolium repens, even if the proliferation of an alien clover does not produce the usual anthropo- centric connotations. As stated previously, T. repens had been spread both deliberately and accidentally through New England and colonies, such as Penn- sylvania and New Jersey, in the 17th century. The resulting transformation of pastures, for which it was valued, represents substantial increases in the soil nitrogen pool in these sites. Such change can precipitate a host of other environmental changes, including a facilitation in the establishment of other alien species. For example, introduction of the ni- trogen-fixing Firetree, Morella faya (Ait.) Wilbur, so raised the amount of biologically available nitro- gen in Hawaiian soils that Firetree has favored the persistence of other non-indigenous species (Vitou- sek et al., 1996). Volume 90, Number 1 2003 Mack 85 Plant Naturalizations and Invasions GROWTH OF A NEW NATION'S NATURALIZED FLORA: 1800—1860 > the incentives and stimuli that had operated before 1800 for the introduction of non-indigenous plants expanded substantially with the growth of the new nation’s commerce and transportation in the 60 years leading up to the Civil War in 1861 (Meyer, 1917). New species were actively sought out that would contribute to the national economy, and for the first time these searches were not left simply to private enterprise. In an often-quoted proclamation, President John Quincy Adams in 1827 instructed U.S. consular offices to gather use- ful species and U.S. naval ship captains to provide for the transport of these living cargoes to the (Hodge & Erlanson, 1956). Annual reports of the Commissioner of Patents, who was responsible for federal involvement in agriculture before 1863, routinely chronicled federal interest in introducing new species in the U.S. for potential use Ei ^ U.S 28th Congress lst Session, 1 36th Congress, 2nd Session, Ex. Federal actions were dwarfed, Сне by the private sectors economic incentive to import spe- cies new to the U.S. for all manner of use, partic- ularly as ornamentals. Establishment of commercial nurseries and seedsmen in major cities, particularly Philadelphia, even before the American Revolution (Lockwood, 1931: 12), growth in this cottage industry through the first half of the 19th century (Leighton, 1987: 67; Mack, 1991, were remarkably diligent, not only acquiring new species from overseas but also in building clien- was followed by a huge and references therein). These seedsmen teles that were not restricted to the immediate vi- cinity of their businesses and gardens. By 1804 Bernard M'Mahon, a Philadelphia seed merchant, was advertising that he had within his nursery col- lection species from such far-flung locales as the “South-Sea Islands,” Asia, Africa, and Europe M'Mahon, 1804). The product of such industry was the availability of several hundred s ies for sale that had not before entered the U.S. (Leighton, 1987; Mack, 1991). Many of these species were to become naturalized by 1900, such as Casuarina equisetifolia L., Cyperus esculentus L., Hedychium gardnerianum Shepard ex Ker-Gawl., Lonicera ja- ponica Thunb., Lysimachia es TET L.. Morella faya, Tamarix spp., and Ulex europaeus L. (Mack, 1991). —. po single events are recorded that y have spawned a naturalization. Cytisus scopar- 34 L) Link, Scotch broom, is a notorious sprawl- ing shrub that rapidly covers new range through a combination of vegetative propagation and local seed dispersal from explosive capsules. Once it oc- cupies a site, it can form an impenetrable thicket that reduces the prevalence of other plants and in- terferes with the movement of livestock (Peterson & Prasad, 1998). Although there are apocryphal iuis of earlier introductions (Lockwood, 1934: 32), the earliest clear reference to the shrub in Vir- ginia arises in a letter (Anonymous, 1921) detailing the misguided generosity of J. M. Сай, a visitor to the farm of John Cocke in Mount Pleasant, Virginia, in 1803. Learning that Cocke intended to introduce an unidentified species to his farm as sheep fodder, Galt wrote with news about what he deemed a su- perior choice: “When I was at your house you men- tioned your Intention of Cultivating the Pride of China for feeding sheep. This will answer for the winter months very well. It did not occur to me then to recommend to your notice the cultivation of Scotch Broom, which affords an ample food for be- tween two or three summer months for sheep and 2” Dr. Galt gave more than advice; he also “I have sent you hogs... sent along seeds of C. scoparius: seed sufficient to plant all your hill sides that you do not mean to cultivate in grain." The seeds had been imported by Galt from a farm in Warwick, England, a few years earlier. His account of the plant's behavior on the Warwick farm seems partic- ularly ominous in retrospect *... it was originally planted as a hedge by an old Englishman—from which it has spread over some hundreds of acres of land by the Birds." To ensure that the spread of Scotch. Broom would be complete, Dr. Galt help- fully added the following tip: *in England they have a method of Expediting Vegetation of Broom— Hawthorn and Holl—by mixing the seeds with the feed of their horn'd Cattle & keeping the Cattle up until they have passed the seed—they then sprin- kle this over their Land & plough it in, in the fall season, in the spring the seed will vegetate ...” (Anonymous, 1921). Unknown is whether the же owner, John Cocke, actually sowed the seeds given by Galt. However, C. scoparius remains naturalized in Virginia in scattered. locales (Harvill et al., 1992). Records of the escapes and naturalizations of these new immigrant species, as well as confirma- tion of the naturalized status of many other species imported much earlier, are established through the proliferation of published local and regional floras along much of the East Coast of the U.S. and even at newly established inland settlements (Sullivant, 1840). Pursh (1814) and Torrey (1824) exemplified ambitious early attempts to record floras that were not confined to urban seaports. Their records are Annals of the Missouri Botanical Garden invaluable because these early U.S. botanists con- fidently assigned Latin binomials to their collec- tions and often reported in unambiguous terms whether these species were naturalized. Thus, Pursh (1814) was able to describe the noxious alien grass Eleusine indica (L.) Gaertn. as occurring in sandy soil from New Jersey to Florida, Festuca ela- tior [probably Lolium pratense (Huds.) S. J. Dar- byshire] as occurring in wet meadows in Pennsyl- vania and New England, and Urtica dioica as found from Canada to the Carolinas. Even if the geograph- ic range was restricted, the notes on the status of these species is nonetheless valuable, e.g., Barton’s (1818) Compendium florae philadelphicae, which was restricted to sites within 10 miles of Philadel- phia. For example, Barton (1818) described Ra- nunculus bulbosus L. as so abundant that * whole fields are often rendered yellow by the pro- fusion of the plant" and Allium vineale L. as com- — — mon, pestiferous, and *. . . impossible to eradicate.” As valuable as published floras of the early 19th century are in tracing the earliest record and fate of non-indigenous species, authors such as Pursh, Torrey, and Barton were not concerned specifically with these species and their effect in the U.S. Com- ing approximately 30 years apart, the accounts of de Schweinitz (1832) and Darlington (1859) pro- vide extraordinarily valuable benchmark accounts that deal explicitly with the scope and status of non-indigenous species along the East Coast of the U.S. Although the detail of their investigations and observations differ in geographic range. these two accounts provide perhaps the best guides we have on the growth of the naturalized flora across the first half of the 19th century. De Schweinitz (1832) explicitly dealt with spe- cies that had become naturalized in the U.S., in- cluding a grouping of species by their mode of in- troduction: those introduced deliberately for cultivation and those ostensibly introduced as seed contaminants (“Introduced fortuitously with agri- cultural seeds") (de Schweinitz, 1832: 151). De Schweinitz provided a separate list of plants that he considered naturalized in smaller areas, e.g., in- dividual states or urban areas. Furthermore, he pro- vided an unambiguous definition for “naturalized” are regularly reproduced, and species: “... which gradually extending themselves, without present cultivation . . ." (de Schweinitz, 1832: 149). His list of deliberately introduced species included many that had been introduced at least a century earlier: Anthoxanthum odoratum L., Nepeta cataria, Plan- tago major, Taxaracum officinale, and Verbascum thapsus L. But this category also included apparent newcomers, such as Barbarea vulgaris Ait. f., Poa annua L., and Raphanus raphanistrum L. His list of accidentally introduced species included Allium vineale, Cerastium vulgatum [probably Cerastium fontanum subsp. vulgare (Hartman) Greuter & Bur- det]. and Lolium perenne L. These species may have been introduced in the 18th century as there is no mention of them before 1700. Species that were naturalized locally, i.e., without the extensive new range occupation he observed for others, included Briza media L., Bromus hor- and /nula helen- Anagallis arvensis L., daceus L., Dactylis glomerata L.. ium. Unfortunately, de Schweinitz did not describe these species in terms of their abundance and im- pact, especially in terms that would allow evalua- tion of any invasive role. Darlington’s (1859) American Weeds and Useful Plants provides probably the best overview we have of the composition and impact of the non-indige- nous flora ca. О. Darlington was explicitly con- cerned with those species that were troublesome in agriculture; then as now most agricultural weeds are non-indigenous (Bridges, 1992). He recorded about 400 non-indigenous taxa that were estab- lished in the eastern third of the country, but it is clear that his collections and observations are drawn from the Middle Atlantic States, the collect- ing area of Rafinesque 50 years earlier. Darlington repeatedly referred to spec “naturalized” e.g., Sisymbrium officinale (L.) Scop.). in contrast to other species that were merely present in the U.S. by the 1850s. For other species, it certainly appears that he also considered them as natural- in referring to the Scotch thistle Onopor- "... very common along = Ф ies as —. ized, e.g., dum acanthium L. as road-sides and in waste places in New England" (Darlington, 1859: 199). Interesting in this regard is his assessment of all Galium species, which he . not sufficiently important even as 3 dismissed as weeds to require notice" (Darlington, 1859: 164). He reported that the flax dodder Cuscuta epilinum Weihe had become quite rare because of the de- cline in the cultivation of Linum usitatissimum, its host. He also recognized that some species, such as Ailanthus altissima (P. Mill.) Swingle (tree-of- heaven), had both beneficial and detrimental qual- ities: providing urban shade but also escaping to vacant lots and even emerging from pavement. PLANT INVASIONS BY THE Мір 19TH CENTURY An invasion is commonly dependent on the im- migrant species being transported to many suitable localities in the new range (Moody & Mack, 1988). The disseminules of some species are readily car- ried by wind, water, or animals (Ridley, 1930) and Volume 90, Number 1 2003 Mack Plant Naturalizations and Invasions can rapidly fill a new range with little or no human assistance. But the spread of many others within a potential new range is greatly facilitated by human agencies. Consequently, as the network of roads, canals, and railroad routes grew in the 19th century U.S., the spread of non-indigenous species also ex- panded. Some of these routes or pathways were be- ing developed even before 1800. For example, there was а comprehensive network of national postal roads that linked the country from north to south by 1804 (Paullin, 1932). Additional roads, supplemented by canals and established barge traf- fic along major rivers, such as the Hudson, the Ohio, and the Delaware, extended this network (Meyer, 1917). Even by the early 19th century, commerce, including seeds and seed-contaminated cargo, was moving routinely throughout the new na- tion. The extent of the American commercial network, as well as the volume of goods moved in the interior of the country, increased markedly with the growth of railroad lines. The first U.S. railroad routes were built in the 1830s. New lines were added rapidly, and most importantly, these lines became linked, so that goods could be moved hundreds of miles in days, not the weeks or months that were needed even along the national trunk roads. From 1830 to 1850, the total length of the railroad system grew from 117 to more than 14,200 km in 27 states in the eastern half of the nation (Meyer, 1917: 573). Growth over the following ten years would dwarf even this total (Meyer, 1917, plate 5). At least one plant invasion appears to have been added in the early 19th century to those that had begun earlier. Asa Gray, who was to become the doyen of American botany in the second half of the 19th century, was by his estimation the first bota- nist to explore the Shenandoah Valley (Gray, 1842). Upon reaching Winchester, Virginia, at the north- ern end of the valley in June 1841, he traveled south. Throughout the broad valley for over a hun- immense dred miles Gray was amazed to see amounts of Echium vulgare L. (vipers bugloss), a Eurasian biennial, occupying many sites, including cultivated fields. Arriving in late June Gray saw vipers bugloss in full flower and described how it formed a *... broad expanse of brilliant blue” (Gray, 1842: 13). Gray's account of the geographic spread and prominence of E. vulgare at this time leaves little doubt that he was describing an inva- sion. He was surprised that farmers had allowed the plant to overrun their fields. Their reluctance to remove it may have stemmed in part from the difficulty of handling it, as it causes contact der- matitis (Magee & Ahles, 1999). Gray (1842) further reported that in the northern states he had seen it only as an occasional roadside plant. Darlington (1859: 242) later reported that he had seen it in “considerable quantities” in Maryland and in abun- dance in New York. Vipers bugloss may have been introduced deliberately among these widely sepa- rated sites because it was valued as a medicinal plant (Parsons & Cuthbertson, 1992: 332) It is intriguing that this plant, which was so prominent in the mid 19th century in Virginia and elsewhere, would be viewed today as locally abun- dant but not invasive anywhere in the U.S. (Lorenzi & Jeffrey, 1987: 245). The abundance of a plant invader can decline precipitously, e.g., Agrostemma githago in Britain (Clement & Foster, 1994), through a change in agricultural practices. Given the need to control plant invasions, understanding the demise of invaders such as E. vulgare in Vir- ginia becomes an important topic for experimen- tation. The ability to identify this species’ pollen in 19th-century sediments could aid in this investi- gation (cf. McGlone & Basher, 1995). e list of invasions under way by 1860 likely included more species than Cytisus scoparius and Echium vulgare, although the strength of evidence for the others is more circumstantial. Darlington (1859) was concerned primarily with the spread and damage of weedy species in agriculture and only incidentally with those species’ occupation of other sites. Nevertheless, he did describe the range and impact of several dozen species in such terms that suggest these were invaders. In fact, he de- scribed Aegopodium podagraria L. (goutweed) as an invader that “... should be carefully watched and its spread arrested” (Darlington, 1859: 151). The strength of the descriptors he used for a few species is a guide to their impact. For example, Darlington (1859: 197) referred to Cirsium arvense (L.) p as "... perhaps, the most execrable weed that has yet invaded the farms of our country." Similar lan- guage was applied to the spread and impact of Cy- perus rotundus L. in its role on cultivated ground in the South, especially in sandy fields апа sand drifts and along the seacoast. Leucanthemum vul- gare Lam. (which he termed as Chrysanthemum leucanthemum) may have also reached the status of an invader because pn (1859: 189) de- sc ribed it not only as a “great nuisance in our coun- un " but also as having “їп some district X- dase possession of their pasture fields. Tt is apparent in Darlington’s description of the status of Echium vulgare and Berberis vulgaris that these two earlier recognized invaders had maintained their role until at least the mid 19th century. Collectively these invasions had already sparked Annals of the Missouri Botanical Garden attention and admonishments to farmers to apply diligence in keeping their fields free of these pests and to sow and trade crop seeds that were free of these damaging contaminants (Darlington, 1859: 242). Even if farmers had universally accepted the advice, it would have been difficult to carry it into practice. Threshing equipment in the early 19th century was a poor match against the seed mimics (e.g., Avena fatua L., taminated crop seeds, and a great wave of new in- troductions were headed to the U.S. (Mack, 1991). s a result, many of the naturalized species that Bromus secalinus) that con- were to become ruderals in the U.S. were not only in the country Ьу 1800, they had been spread throughout much of the eastern half of the U.S. by 1860. In contrast, many other species that have be- come invaders in the U.S. were just being detected (Bartlett . 2002) and Lonicera japonica (Schierenbeck 1994) or had yet to be detected (e.g., Salsola kali L.). In a sense, the damage caused by plant by this date, such as Bromus tectorum L. et al. et al., invaders in the first 200 years or more of European colonization along the eastern coast of the United States would be far outweighed by the damage brought about by species introduced or deliberately spread post-1860. CONCLUSIONS Several timely observations can be gleaned from tracing the growth of the naturalized flora in the U.S. between the early 17th and mid 19th centu- ries. From the beginnings of European colonization їп North. America, the list of plants that became naturalized was shaped strongly by the species hu- man immigrants selected for their transplanted ag- riculture. Even if a naturalized species did not owe its new status to deliberate introduction, it likely of a deliber- ately selected species; Pursh (1814) maintained arrived as a contaminant in the see that Anthoxanthum odoratum and Festuca elatior [probably Lolium pratense (Huds.) S. J. Darbyshire] arrived in this manner. Although the link between naturalization and the deliberate introduction of species for food, forage, and medicinal use contin- ued, even in the 17th century some species appear to have been imported for aesthetic reasons. The list of species selected as ornamentals has expand- ed ever since; in fact, it dwarfed the number of new immigrant species in more utilitarian categories by at least 1860, if not earlier (Mack & Erneberg, 02 The link between agriculture in western Europe and the establishment of colonial agriculture was so strong that by ca. 1850, the weed and ruderal floras of western Europe and eastern North America were quite similar. Through a combination of delib- erate introductions and seed contaminants in im- ported seed and other cargo, Europe's colonizing flora had been largely transplanted to form much of North naturalized Although formed by happenstance, this link remains. America's flora. As introduced species were traded or acciden- tally spread in commerce among the eastern colo- nies beginning in the 17th century, a few species became so widespread and naturalized that they formed invasions. Although the known list of ap- parently invasive species by 1800 is modest, other species were probably playing that role but were not described in terms that we can decipher as in- vasions. The frequency with which widespread and perhaps invasive species were reported through the first half of the 19th century suggests that these species’ opportunity to spread and consequently proliferate was tied to the growth of all forms of commerce and the forms of transportation that fos- tered the spread of commerce. Furthermore, the connection between which spe- cies received cultivation in their new range and those that became naturalized appears high (Mack & Erneberg, 2002). This historic link between cul- tivation and naturalization (Mack, 2000) provides an important harbinger for the future. If the history of plant naturalization between 1634 and 1860 is any guide, future naturalizations will be largely shaped by (1) the often idiosyncratic human moti- vations for importing alien species, (2) the degree of cultivation provided to these species upon their entry, and (3) human industry in transporting those species to many new locales and habitats in new ranges, thereby enhancing the opportunities for es- tablishment on sites where they can persist without further human assistance. Literature Cited ко 1921. Scotch Broom. Letter of Dr. zalt. Dec. 9, 1803. Tylers Quarterly Historical and e naga еи 2 (3): 246-241. Bartlett, S. J. Novak & R. N. Mack. 2002. Genetic variation in Promis tectorum (Poaceae): Differentiation in the eastern United States. Amer. J. Bot. 89: 602- 612 e W. P. C. 1818. e үл Philadelphi- : Containing a Description of the d anc [o alized Plants Found within a Circuit of Ten Miles Around m Vols. 1 & 2. M. Carey & Sons, Philadelpl Bidwell, p! Y. & J. . Falconer. 1925. History of Agri- culture in the us United States, 1620-1800. Car- negie Institution of Washington, Washington, D.C шл. D. (Editor). 1992. Crop Losses Due to Weeds} in e United States. Society of America, С ны Illinois Weed Science Volume 90, Number 1 2003 Mack 89 Plant Naturalizations and Invasions Brugam, R. B. 1978. aie indicators pen -use change in southern Conn uat. Res. 9: 349-362. Budd, T. 1685. бы rios a E in Pennsilvania & New-Jersey in America. W. Bradford, Philade Iphia. (Microfilm. Ann Arbor, Michigan, University Microfilms [1966] March of America facsimile series, no. 32 Byrne, R. & J. McAndrews. 1975. Pre-Columbian purslane (Portulaca oleracea L.) in the New World. Na- ture 253: 726—727. Carman, H. J. & R. G. Tugwell (Editors). 1934. Essays Upon Field Husbandry in New England, and Other Pa- pers, 1748-1762 by Eliot Jared. Columbia Univ. Press, New York. oes L. 1923. The ee of Agriculture in Amer- a. McGraw-Hill, ipa ог Clement, E. J. & M.C er. 1994. Alien Plants of the British Isles. dea Society of the British Isles, Lon- n. Crooks, J. A. & M. E. Soulé. 1999. Lag times in popu- lation explosions of invasive species: Causes and im- plications. Pp. 103-125 in O. T. Sedit P. J. Sche & A. Viken (editors), Invasive Species and Biodiversity Manage Ei Kluwer, Dordrecht, The Netherlands. Crosby, A. W. 1972. The Columbian Exchange; Biological and Cultural Consequence es of 1492. Greenwood, West- port, Connecticut. Darlington, W. [1849] 1967. EPI of John Bartram and Humphrey Marshall. Facsimile of the edition of 1849, with a new introduction by J. Ewan. Hafner, New York. 59. American Weeds and Useful Plants, 2nd ейге Ьу G. Thurber). А. О. Moore, New York. Denton, D. 1670. A Brief Description of New York: For- merly called New-Netherlands. J. Hancock and W. Bradley, London. (Microfilm. Ann Arbor, Michigan, University Microfilms [n.d.] American Culture a Heel 7 Dwight, т [1821] 1969. Travels in New England and New York, Vol. 1. Reprint. B. M. Solomon (editor). Belknap Press, € ambridge, Massachusetts Farah, Tanaka & N. E. West. 1988. Aut- ec iy me population biology of Dyers Woad (Isatis tinctoria). Weed Sci. 36: 186-193. Fernald, M. L. [1950] 1987. Gray's Manual of Botany: A Handbook of the Flowering Plants and Ferns of the Central and Northeastern United States and Adjacent Canada, 8th ed. Reprint. Dioscorides Press, Portland, Oregon. Fulling, E. H. 1943. Plant life and the law of man. IV. Barberry, Atos and gooseberry, and cedar control. Bot. ib 483-591. Gade, E 1916. Naturalization of plant aliens: The nter orange in Paraguay. J. Biogeography 3: 269— 21 Gera J. [1633] 1975. The Herbal: or General Histor Plants. Reprint, revised and enlarged by T. iE ET New York. Gray, A. 1842. Notes of a botanical excursion to the mountains of North Carolina, &с; with some remarks on the botany of the higher Alleghany Mountains. Amer. J. Sci. & Arts, Ser. 1, 42: 1—49. Hall, C. C. 1910. Narratives of Early Maryland. Scrib- ners, New York Haney, A. & F. A. Bazzaz. 1970. Some ecological impli- cations of the distribution of hemp (Cannabis sativa L.) in the United States of America. Pp. 39-48 in C. R. B. Joyce & S. H. ы is The Botany and Chem- istry of Cannabis. Churchill, London aries A us Fav Nurser, Phillimore: ry ‚ T. R. Bradley, C. E. Ste F. Wie- old M 'E. Ware, m W. Ogle. G. Y. FX & G. rginia. : Plants that ransformed America. Harcourt Brace Jovanovich, New C. O. Erlanson. 1956. Federal plant on—A review. Econ. Bot. 10: 299—334. assachusetts-Bay, Vol. 1. Facsimile. Arno Press, New York. Hutton, L. D. 1927. Barberry eradication reducing stem 4—118 rust losses in wide area. Yearb. Agric. pp. 1 Josselyn, J. 1672 . New-E ngland's Rarities, Discovered in i s that "E Ф =] = LP UE 1 2 =" © =, 1674] 1988. Account of Two Voyages to New- England. John Rl Colonial Traveler: A Critical dno of Two Voyages to New-England. Reprint, edited and with an ae by P. J. Lindholdt. Univ. Press New England, Hanove Julien, M. H. & M. W. Griffiths 1998. Biological uA of Weeds: A World Catalogue of Agents and their Tar, Weeds, Ath ed. CABI Pub., Oxon, U.K. Kartesz, J. T. & C. A. Meacham. 1999. Synthesis of the North American Flora, ver. 1.0. North Carolina Botan- ical Garden, kya Hill. Time lags in biological invasions with re a to e success and failure of ici species. Pp. 15-38 in P. Pysek, К. Prach, M. Rejmanek & M. Wade (editors), Plant Invasions. General pee and Special Problems. 77 ix erdam Kuznets, S. Miller & R. A. Easterlin. 1960. Foalalo Ww es and economic growth: United States, 1870—1950. 2. Analyses of economic change. Mem. Amer. Philos. Soc. 51 Leighton, А. 1970. Early American Gardens; “for Meate or Medicine." Houghton Mifflin, Boston. 1976. American Gardens in the Eighteenth Cen- tury: “For ван or for Delight." Univ. Massachusetts ‚ Amher pec Gardens of the Nineteenth Cen- tury: “For Comfort and Affluence.” Univ. Massachusetts Press, Buc Lockwood, A. G. B. (Editor). 1931. Gardens of Colony and State; Саб and Gardeners of the Am erican Col- onie s and of the Republic Before 1840, Vol. LG C. Scrib- 1 Sons, New (Editor). 1934. Gardens of Colony and State; Gardens and Gardeners of the American Colonies and of the Republic Before 1840, Vol. 2. C. Scribner’s Sons, New York. Lorenzi, H. J. & L. S. Jeffery. 1987. Weeds of the United States and Control. Van Nostrand Reinhold, New bu E. 1996. Settlement vid survival. Pp. 40—61 i Gannon (editor), The New History of Florida. Univ йз as Gainesv ille. Mack, 1989. Temperate grasslands vulnerable to vue invasion: Characteristics and consequences. Pp. 155-179 in J. A. Drake, Н. A. Mooney, К. di Castri, R. 90 Annals of the Missouri Botanical Garden H. Groves, F. J. Kruger, M. Rejmanek & M. Williamson Environmental and economic costs of nonindigenous NE Biologie al Invasions: A Global Perspective. sp in the United States. BioScience 50: 53-65. iley & Sons, New Yor i 1814. jaa is 'ae Septentrionalis. White, . 1991. The commerc ial io trade: An early dis- О Londor perser of weeds. Econ. Bot. 51- фил, 1990. ee and Colonies, 1500-1625. 2000. Cultivation dae gant о by Hambledon. London. seducing environmental stochasticity. Biol. Invasions 2(2): 111-122. 2 i Motivations and consequences of the hu- man dispersal of plants. Pp. 23-34 in J. A. McNeely (editor), The Great Reshufling: Human Dimensions in Invasive. Alien. Species. International Union for the Conservation of Nature, Gland, Switzerland, and Cam- bridge, & M. Ermebérg. 2002. The U.S. naturalized flora: Largely the product of deliberate introductions. Ann. Missouri p Gard. 89: 176-189. —— mberloff, W. M. Lonsdale, H. Evans, M. Clout F. а Bazzaz. Biotic invasions: Cause ме Ls consequences and control. p Appl. 10: 6 А Magee, D. №. ¢ E. Ahles. 1999, Flora of the North- east. А Mi of the Vascular Flora of New England and Adjacent New York. Univ. Massachusetts. Press, Amherst. Massachusetts Historical Society. 631-1637. 1943. Winthrop Pa- Massachusetts Historical So- McGlone, M.S T L. R. Basher. 1995. The Pr) of the upper Awatere catchment, inland Kaikou Range. erates South-Island, New-Zealand. New Zealand J. Ecol. 19: 53-66. einig, D. W. Us The Shaping of America: A Geo- graphical Perspective on 500 Years of History, Vol. 3. Transcontinental 1850-1915. Yale Univ. Press, New en Meyer, B. H. . History of Transportation in the Unit- м States a 1860. Carnegie Institution of Washing- n, Washington, D. M’ Mahon, B. 1804. “A T of American Seeds, &c . ..” Philadelphia. Moody, M. E. & R. N. Mack. 1988. Controlling the spread of plant invasions: The importance of nascent foci. J. Ecol. 25: 1009-1021. Moore, P. D., J. A. Webb & M. E. c иа Analysis. Blac kwell Scientific, Oxford. . Index bis үк astriensis. Trans. il. 3: 157-184. National Re: il. 2002. Predicting Invasions of Nonindigenous Plants and Plant Pests. (U.S.) N al Research Counc ‘il, Washington, D.C. Parsons, W. T. E. G. Cuthbertson. eeds of Australia. Inkata Press, Melbourne. Раш, С. О. 1932. Atlas of the Historical Geography of the United States. J. K. Wright (editor). Carnegie Insti- tution of Washington & American Geographical Society rk. America, 1991. Pollen ation- 1992. Noxious of New York, Washington, E " New Yo Дә. D. J. . Prasad. The biology of Ca adian weeds. 109. = scoparius (L.) Link. Canad: . Sci. 78: 497-5 Peterson, P. D. 2001. sem Rust of Wheat: From Ancient ‘пету to Modern Foe. APS Press, St. Paul, Minnesota. Pimentel, D., L. Lach, R. Zuniga & D. Morrison. 2000. Rafinesque, C. 5. 1811. Ап essay on the exotic plants, mostly European, which have been naturalized, and now grow тыы in the Middle States of North cal о 3" Hexade, 2: 330-34. 1993. Тһе Vascular Flora mer- . Klein. ican cg Pim Society, Philadelphia. Ridley, Н. | The Dispersal of Plants Throughout the boe F Reeve Ashford, p - E Е M. Clements, M. ). J. Doohan Е б: п. 1995 The hi. d Canadian AUN 105. Linaria vulgaris Mill. Canad. J. Pl. Sci. 75: 525- Sc hierenbeck, . Mack & R. R. Sharitz. 1994. ects of einn on ape and biomass allocation in native ie Quse ed species of Lonicera. Ecology 75: 1661-16 Schweinitz, L. A de. 1832. Remarks on the plants of Europe which have become naturalized in a more or less degree, in the United States. Ann. Lyceum Nat. . New York 3: 148-155. Scot, x 1685. The Model of the Government of the Prov- ince of East-Jersey in America; and Encouragements for Such pde to be Concerned There. J. Reid, Edin- bun: J.K. (Microfilm. Ann Arbor, Michigan, Univer- sity Microfilms [n.d.|] American Culture Series, Reel 3.86. Sullivant, W. S. 1840. A Catalogue of Plants, Native or Naturalized. 4 in the Vicinity of Columbus, Ohio. Scott, Columbus, Ohio. Torrey, J. 1819. A Catalogue of Plants, Growing Sponta- neously within 30 Miles of the City of New York. Ly- ceum of Natural History of New York, Albany, New Yor Torrey, JA A. 1824. Flora of the Puis and Middle Sections Eu. United States, Vol. 1. T. & J. S. Swords, New Yo rk. U.S. 28th Congress, Ist Session. 1844. Senate. Report of the лагите of apr The Operations of the Pat- ent Office during the year 1843, Washington, D.C US . a: L ^ A ш = 2 of Patents for de year 1860. Agriculture. Washington, Vitousek, P. M., C. M. D'Antonio, L. L. Loope & R. Westhivoks. 1996. y oai 'al invasions as edo en- vironmental change. Amer. Sci. 84: 468—478. Westbrooks, R. G. 1998. Invasive Plants. Changing the Landscape of America. Federal Interagency Committee for the Management of Noxious and Exotic Weeds, ashington, D.C. Wood, W. [1634] 1977. New Ag ce Prospect. Reprint, with an introduction by A. T aughan (editor). Univ. Massachusetts Press, Amhers Pune. “A. 1846. Chronicles of the First Planters Ы Ше Colony of Massachusetts Bay, from 1623 to 1636. Now First Collected from Original Records and Contempo- raneous Printed Documents, and Illustrated with Not Little & Brown, Boston. INVASIVE ANTS: UNWANTED PARTNERS IN ANT-PLANT INTERACTIONS?! Lori Lach? ABSTRACT Ж = invasive ants — their interactions with plants are inevitable and наке рези great implic ations for culture and conservatio to hypotheses about how invasive ants may differ Homoptera, and interactions decus to invasive ants can in the Acum: of i attributes of other organis a determine the cons nts will be at high and low contexts in id la nectaries, Homoptera, invasiv adverse outcon . The potentia for effec ts of i invasive ants on plants to counteract, and the complexity and context- eve ha ll ark «а nt-plant interactions generally, preclude dr. Irawing simple pcm about the net impacts abundance, aggression, byproduct mutualism, e ants. from native ants in protecting plants from herbivores, tending of plant ye cen tion. Exa 'actions ти equences for the ering traits of the invasive ants, plants, and sas beret that нка with the p f monstrate that all three of these traits common invasive ants an mples den carbohydrates, context-dependent outcomes, extrafloral Interactions between ants and plants range from facultative, loose associations to species-specific, obligate mutualisms, and innumerable questions have been posed about the costs, benefits, and evo- lutionary implications of these relationships (Beat- tie, 1985; Bronstein, 1994; Jolivet, 1996 vasive ants spread, their interactions with plants are inevitable and have potentially great implica- tions for agriculture and conservation. s in- o what ex- tent are ant-plant interactions altered by the intro- duction of these novel partners? Invasive ants possess a distinct combination of traits relative to native ants (Holway et al., 2002). In this paper, I hypothesize that some of these traits will affect how invasive ants associate with plants and provide a basis for predicting the outcome of these novel interactions. I focus on the potential for differences between native and invasive ants in the way they protect plants from herbivory, tend Ho- moptera, and interfere with plant reproduction. These examples encompass the more facultative, ubiquitous interactions in which invasive ants are increasingly likely to participate as they spread around the globe. Although I explore the differences between na- tive and invasive ants for each type of interaction separately, the same ant can interact directly and indirectly with the same plant in multiple ways. Invasive ants that benefit the plant in one interac- tion may be detrimental to the plant in another. The net outcome for the plant will depend on the rela- tive balance of a range of interactions, all of which in turn will be influenced by the same set of vari- ables that typically influence ant-plant interactions. Ultimately, we will want to know whether invasive ants have the ability to tip the balance of interac- tions toward consistently positive or negative out- comes for the plants, and how these outcomes can be manipulated to achieve land management goals. INVASIVE ANTS the approximately 10,000 species of ants in the world, about 150 have been introduced around the world with the help of humans; these may be termed exotic, or tramp ants (McGlynn, 1999). In- vasive ants are the small subset of introduced ants that are able to establish and penetrate areas out- side of human-modified habitats (Holway et al., 2002). Well-known examples in the United States include the red imported fire ant, Solenopsis invicta, thank the organizers of the 48th annual Missouri Botanical Garden ptura Symposium for the invitation to contribute this paper. J. Ness, D. Holway, T. Yong, D. Pimentel, R. Root, and an ano iewer provided helpful comments. This work was supported in part by an EPA STAR fellowship ps an NSF dissertation enhancement grant to the author. ? Cornell University, Ecology and Evolutionary Biology, Corson Hall, cornell.edu ANN. MISSOURI Bor. New York 14853, U.S.A. ljl13€ Ithaca, GARD. 90: 91—108. 2003. Annals of the Missouri Botanical Garden and the Argentine ant, Linepithema humile. The lit- tle fire ant, Wasmannia auropunctata, the big-head- ed ant, Pheidole megacephala, and long-legged ant, Anoplolepis gracilipes, also appear to be highly suc- cessful invaders, although they are far less studied. See Table 1 for a summary of distributions. One hallmark of invasive ants is their ability to become extremely abundant in their new habitats. Pitfall traps in areas invaded by Solenopsis invicta (Porter & Savignano, 1990), Linepithema humile (Holway, 1998), and Pheidole megacephala (Hoff- man et al., 1999) have all shown that the number of invasive ant workers greatly exceeds that of na- tive ants in nearby un-invaded areas. Several mechanisms are likely contributing to invasive ants? achievement of numerical superiority including es- cape from natural enemies and competitors and changes in colony structure toward vane n D multi-nest supercolonies deo lack intraspecific ag gression (Holway et al., : і The physiology and una of invasive ants also likely play a role in their tendency to achieve high abundance (Holway et al., 2002). Ecologically dom- inant ants, including invasive ants such as Linepi- thema humile, have modified crops that allow them to take in more liquid foods, including floral and extrafloral nectar and homopteran honeydew (Eisner, 1957; Davidson, 1998). The ability to harvest these carbohydrate-rich resources may be especially im- portant in fueling the high tempo activity of a large workforce, thereby maintaining a high dynamic den- sity (ants/area/time) (Tennant & Porter, 1991; David- son, 1997; Davidson, 1998). Individual nests can be highly vagile, allowing invasive ants to move in re- sponse to the availability of resources (Passera, 904). Even Solenopsis invicta, which nests in mounds, will create satellite nests at the base of plants when tending aphids (Kaakeh & Dutcher, 1992). A high level of aggression is another char- acter common to invasive ants and likely enhances their success as predators (Holway et al., 2002). Ac- cess to prey and carbohydrate-rich resources and a large workforce are both the cause and effect of su- perior exploitative and interference competition abil- ities through which native ants and other competitors may be displaced (Holway et al., 2002) Displacement of native ants is the most com- monly documented consequence of an invasive ant introduction (Holway et al., 2002). ah enn hu- mile, for example, has displaced several species of epigaeic ants in California (Erickson, 1971; Ward, 7; Holway, 1995; Human & Gordon, 1996), Portugal (Cammell et al., 1996; Way et al., 1997), and South Africa (Donnelly & Giliomee, 1985). Wasmannia auropunctata has expanded its territory Native and introduced ranges of some invasive ants. Table 1. References Introduced range Native range Common name Species Veer- esh & Gubbaiah (1984), Fellowes (1999), Green et al. (1999 78), ). Haines & Haines (19 í Fluker & Beardsley (19 Asia, Australia, Pacific and Indian in dispute long-legged or Anoplolepis gracilipes (F. Ocean islands crazy ant Smith) (2001) . Young et al. Haskins & Haskins (1965), — Fluker & Beardsley (1970), Ma- Argentine ant Argentina inepithema humile (Mayr) Ji I Mediterranean-type climates of all continents, Hawaiʻi, Bermuda , Room (1975), Fowler et al. 2000) 71) ); Haskins & Haskins (1965), Greenslade (19 tropical Africa big-headed ant South America, Asia, Australia, Pheidole megacephala (Fa- Fluker & Beardsley (1970), Majer (1985 (1990), Fellowes (1999), Deyrup et al. Vinson & Greenberg (1986), Holway et al. acific Ocean islands. S Bermuda U.S., Australia bricius) — (2002) Brazil red imported fire Solenopsis invicta (Buren) ant 8), 7 2), Fabres & Brown (19 Williams (1994), Conant & Hirayama (2000) Spencer (1941), Entwistle (19 Hawaiʻi, little fire ant neotropics Wasmannia auropunctata Africa. continental U.S Galapagos Islands, New Cale- (Rogers) donia Volume 90, Number 1 2003 Lach 93 Invasive Ants at the expense of native ants on the Galapagos Is- lands (Lubin, 1984). Pheidole megacephala has di- minished the native ant populations of several hab- itats in Australia (Majer, 1985; May & Heterick, 2000; Hoffman et al., 1999; Vanderwoude et al., 2000). Solenopsis geminata (F.) and other ants na- tive to the southern U.S. have become much less common following invasion by S. invicta (Porter & Savignano, 1990; Gotelli & Arnett, 2000). nvasive ants also affect other invertebrates and even vertebrates, usually negatively. Anoplolepis gracilipes is blamed for diminished populations of various invertebrates in the Seychelles (Haines & Haines, 1978) and is generating ecosystem-level changes through its impact on the red land crab on 1999). Linepithema humile is associated with decreases in some insects 1992; Bolger 2000). The great reduction in endemic in- sects in lowland Hawai'i at the end of the 1800s has been attributed to invasion by Pheidole mega- Christmas Island (Green et al., in California and Hawai'i (Cole et al., et al., cephala (Zimmerman, 1970). Solenopsis invicta has been linked to declines in a number of terrestrial invertebrates, birds, and mammals in the southern United States (Holway et al., 2002). Wasmannia au- ropunctata has eliminated or reduced terrestrial in- vertebrate populations in the Galapagos Islands (Lubin, 1984) and New Caledonia (Jourdan, 1997). The mechanisms for these effects are not always clear but likely involve some combination of pre- dation and a via direct or indirect inter- actions (Holway et al., PREDICTING THE NATURE AND OUTCOMES OF INVASIVE ANT-PLANT INTERACTIONS A wealth of studies has established that the na- ture of ant-plant interactions is dependent on many variables including ant behavior, ant colony size and stage, host plant attributes, and the abundance and behavior of other organisms in the system (Buckley, 1982; Beattie, 1985; Keeler, 1989; Cush- man, 1991; Davidson & McKey, 1993; Bronstein, 1994, 1998; Jolivet, 1996). As explained above, in- vasive ants tend to have larger colonies and exhibit more aggressive behavior than native ants. More- over, invasive ants appear to have a stronger pre- dilection than many native ants for carbohydrate resources, which are invariably obtained directly or indirectly from plants. Since native ants are fre- quently displaced when invasive ants enter a new habitat, it will be traits of the invaders that influ- ence future outcomes for the plants. Applying ex- isting models of ant-plant interactions, | develop hypotheses about how invasive ants’ elevated abun- dance, aggression, and affinity for carbohydrate- rich food affect how they might protect plants from herbivores, tend Homoptera, and interfere with plant reproduction. ANT PROTECTION FROM HERBIVORES Protection of plants from herbivores is perhaps the oldest recognized effect of ants on plants. As long ago as A.D. 304, Chinese citrus growers facilitated the spread of the yellow citrus ant, which preyed on herbivores in their orange orchards (Huang & Yang, 1987). Since then, ants’ ability to prey on, or simply harass, would-be herbivores has been noted in many systems (Beattie, 1985; Hólldobler & Wilson, 1990). In some cases, plants entice ants by offering food or shelter. For example, it is thought that one of the primary reasons for the presence of extrafloral nec- taries is to attract ants in exchange for protection from herbivores (Bentley, 1977). A number of traits associated with invasive ants have been linked to plant defense: large colony size, high levels of worker activity, and aggressive- ness (Davidson & McKey, 1993; Bronstein, 1998 A large number of very active workers may result m in increased plant visitation by ants leading to de- creased herbivory (Gaume et al., 1997; Bronstein, 1998; Gaume et al., 1998; Linsenmair et al., 2001) because herbivores are located more rapidly and have a shorter residence time on the plant (Duarte Rocha & Godoy Bergallo, 1992). Benefits to the plant probably level off or even decline at some colony size, however. An extremely large colony may impose a cost on the plant if the ants are steal- ing floral nectar, tending Homoptera, interfering with the natural enemies of herbivores, or excavat- ing the plant’s root system. Research to determine where the cost-benefit trade-offs lie with respect to ant colony size has yet to be done. Aggressive behavior also aids in и the etour- 1983; Bronstein, 1998). Bentley (1977) re- lated ant aggression on plants to three factors: pred- p behavior or the defense of territorial boundaries (Way, 1963), and swarming plant from herbivores (Fiala et al., 1989; neau, atory behavior, ownershi behavior or the ability to rapidly recruit workers. АП three behaviors may be enhanced at the colony level by having a large workforce. Aggression is also negatively correlated with proximity to an ants’ nest; disturbance near an ants’ nest or food source will elicit a stronger response than a threat en- countered farther away (Way, 1963). The vagile, po- lydomous nesting behavior of invasive ants may al- low them to nest in closer proximity than a native Annals of the Missouri Botanical Garden 94 efns А plant = 5 H D * = T" 4 y ants herbivores A . l. A schematic of the food-for- Mig mu- tualism between ants and extrafloral nectaries. The ex- trafloral nectaries provide ants with food and ihe ants de- ter herbivores. The direction of the arr show the type of effect. A = effectiveness of ant-defense, =e ectiveness ч кн r defenses, Н = intensity of herbivory, 1, = estment in nectar а nectaries, and efns = т. nectarie ant would, possibly facilitating a greater aggressive response. Some ants provide protection from herbivores while collecting extrafloral nectar. Keeler (1981: 490) modeled the conditions under which the mu- tualism is favored: pl A(1 = D)H] > I, where p — probability that ants will find the plant, A — effectiveness of ant-defense, D — effectiveness of other defenses, Н = intensity of herbivory, and I, = investment in nectar and nectaries. From the plants’ perspective, the mutualism is maintained when the left side of the equation, the benefits to the plant, exceeds the right side, the costs of at- tracting the ants (Keeler, 1981). A schematic of the relationship is shown in Figure 1. Given the numerical superiority, attraction to carbohydrate resources, and aggressive character of > pa Since D is primarily a factor and D invasive ants, we might expect that p and Aii. > of the pláni and the herbivore(s), D invasive native* Invasive native are not likely to differ. H may а predation is a fac ‘tor, 1.е., if ants prey on ` interfere with the natural enemies of herbivores (Eubanks. 2001). However, Ну. if invasive ants diminish herbivore popula- may be less than native tions via direct or indirect interactions that affect herbivore eggs, larvae, pupae, or adults away from the plant. If extrafloral nectar production is an in- ducible defense (Koptur, 1989; Ness, 2001) I, will increase with H. For example, if H than H duced more often when native ants are present, and I sus, IS greater then extrafloral nectaries will be in- invasive? anaie МШ be greater than I... If on balance, the left side of the equation in- creases and 1, stays the same, then there will be a greater benefit margin for the plant when invasive ants are present. If, however, the presence of in- vasive ants lowers the left side of the equation, then — to maintain the mutualism, may also need РРА to be less than I, „с: For invasive ants that are aggressive, abundant, and attracted to carbohy- drate-rich resources, the model predicts mainte- nance of the mutualism if the potential for herbiv- ory is not diminished (e.g., through interactions away from the plant). ANT TENDING OF HOMOPTERA Ant tending of honeydew-producing Homoptera, including scale, mealybugs, aphids, and membra- Buckley, Davidson & McKey, 1993). Homoptera extract phloem from the host plants and excrete it as honeydew. Ants feed- ing on the honeydew often protect these Homoptera from parasites and predators (Way, 1963; Buckley, 1987). T rimental for the plant, robbing it of phloem and he ant-Homoptera mutualism may be det- sometimes leading to mold accumulation (Bach, 1991; Lewis et al., mission (e.g., pineapple mealybug wilt disease 1982), (Evans, 1973), swollen shoot virus 1976) or phytopathogen trans- ~ И В rot г Ade- doyin, 1978)) while conferring no SEEN in ex- change (Buckley, 1987). action may benefit the plant if ants deter other Beardsley et al., Less commonly, the inter- herbivores while tending Homoptera (Carroll & Janzen, 1973; Messina, 1981; Compton & Robert- 1988). As with plant protection from herbivores, the son, characters common to many invasive ants will af- fect their Homoptera-tending abilities and conse- quently the outcome for the plant. The need for carbohydrate resources and ability to harvest co- pious amounts of liquid provides the basis for the attraction of invasive ants to Homoptera. Ant ag- gression, particularly ownership behavior, is an ef- fective deterrent to most would-be parasites and predators of Homoptera (Way, 1963). Polydomy and nest vagility may allow invasive ants to reside in closer proximity to these carbohydrate resources than would native ants. Ant abundance, however, may be the primary trait affecting invasive ant-Homoptera mutualisms, as it may affect the interaction through several mechanisms. Higher ant numbers may translate into a higher probability of locating the tendees, Lach Volume 90, Number 1 2003 Invasive Ants a b Homoptera Cp plant Homoptera Ch plant A — A A на Н D H D + + = ES + + РЕ Ри v = v v 4 ants ` herbivores ants » herbivores p A A Figure 2. The ant- Homoptera пасая when it results in a byproduct mutualism with the plant (2а) and when it is a parasite on the plant (2b). The direction of the arrows indicates the direction of the effect and the positive and negative signs show ts type of effect The thickness of the arrow indicates the strength of the effect. Dotted lines illustrate weak or no effect. C, = the cost of hosting the Homoptera, H, A, 04 D are as defined in vis l. In the detrimental effect of the ants on yv herbivores and subsequent decline in herbivore attack outweighs the cost to the plant of hosting the Homoptera. In 2b, the ants fail to deter ч rema so the plant bears the ш y hosting the Homoptera and attack by other herbivores. and a higher ant: homopteran ratio may make ants more effective at deterring their partners’ enemies (Steyn, 1955). Tending ants are a limiting resource to honeydew-producing Homoptera in many sys- tems (Addicott, 1978; Sudd, 1987; Cushman & Ad- dicott, 1989; Breton & Addicott, 1992; Fischer et al., 2001), and well-tended Homoptera may grow faster, reproduce more rapidly, and produce more young than untended ones (Way, Bristow, 1984; Morales, 2000). ere the ant: Homoptera ratio becomes too low, Homoptera may become ant prey or be more easily parasitized or preyed upon Ri وت‎ enemies (Way, 1963, and references ; Breton & Addicott, 1992; Sakata, 1994; 1996). The superabundance achieved by invasive ants may e tender:tendee ratio from falling below this threshold. In such a case, we would expect limits to the abundance of both mutualists to be imposed by the host plant. Impacts on the plant, however, also may depend on whether the ants tending Homoptera deter other herbivores. In this byproduct mutualism (Yu, 2001), the plant benefits indirectly from the ants’ pres- ence; the ants benefit from the Homoptera hosted by the plant and possibly the herbivores, if they are captured as prey. In this scenario, Homoptera are analogous to extrafloral nectaries and Keeler’s mod- el (1981) can be modified as: РАП = D)H Da хамай! > С, The right side of the equation (formerly the in- vestment in nectar and nectaries, [,) becomes the cost to the plant of hosting the Homoptera, C,, in- cluding direct effects of lost phloem and indirect consequences such as mold accumulation and sus- ceptibility to phytopathogens. Other terms remain probability that ants will fin effectiveness of ant-defense e effectiveness of the same: p = plant, A (non-homopteran) herbivores, D = other defenses, intensity of (non-homopteran) herbivory. Considering the traits of invasive ants that favor high Шек: Б-Т populations, it is prob- able that in many cases С, ,,,,,.;,. will be greater than C, us The probability that the ant finds the plant, p. may vary not only with the abundance of the ant, but possibly with the ability of different jet, n to recruit ants (Del-Claro & Oliveira, 1996). unclear, though, whether invasive ants may ipie to attractant cues differently than native ants. Hy- pothesized differences in A, D, and H for native and invasive ants would be as explained above. As with the food-for-protection mutualism involving extrafloral nectaries, the byproduct mutualism will only be maintained if the benefits, in terms of de- terred non-homopteran herbivores denoted on the left side of the equation, exceed the costs exacted on the plant by the Homoptera, the right side of the equation (Fig. 2a). If costs to the plant exceed benefits, the ant-Homoptera interaction will tend toward a parasitic relationship with its host (Fig. 2b), possibly resulting in reduced fitness of the plant. Studies that fail to detect any change in plant fitness with Homoptera outbreaks associated wit invasive ants may be observing a balance between the two types of outcomes. As with interactions in- volving extrafloral nectaries, for invasive ants that are aggressive, abundant, and attracted to carbo- hydrate-rich resources, the model predicts main- tenance of the byproduct mutualism if the potential Annals of the Missouri Botanical Garden for herbivory is not diminished (i.e., through inter- actions away from the plant). ANT IMPACTS ON PLANT REPRODUCTION Seed set. . Yano, 1994; Puterbaugh, 1998), ants commonly have been re- garded as unwanted guests in flowers (Kerner, 1878; Buckley, 1982). Attracted by floral nectar, ants may damage floral structures, and depress pol- With few exceptions (e.g len viability with their antibiotic secretions (Kerner, 878; P 991; Galen, 1999). As with Homoptera and extrafloral nectary tending, super- abundance and an affinity for carbohydrate resourc- eakall et al., es combined with high levels of aggression would lead to the expectation that invasive ants would be exceptional at recruiting colony members to flowers and exploiting floral nectar. If numerous aggressive ants are present in flow- ers, their presence may affect other floral visitors. Pollinators may be forced to reposition frequently to avoid attack by the ants, and this may result in increased transfer of pollen and consequently high- er seed set. Alternatively, pollinators may avoid the flower altogether, or reposition too superficially for effective pollen transfer, ultimately resulting in de- creased seed set (Wyatt, 1980). A plethora of floral antagonists, including pollen consumers, and other florivores, however, may also be deterred, perhaps c posen for any adverse effects on pollinators. ants’ ultimate impacts on seed set will be didi on the attraction and availability of floral nectar in the context of the ants’ seasonal diet pref- erences, activity patterns, and abundance, the num- ber of pollinators and floral antagonists and their susceptibility to ant deterrence, floral structure and defenses against ants, and the pollination require- ments of the plant (Kerner, 1878; Koptur, 1979; Vinson & Greenberg, 1986; Huxley & Cutler, 1991; Klinkhamer & de Jong. 1993; Lanza et al., 1993; Koptur & Truong, 1998; Puterbaugh, 1998), as well as the relative difference between ants’ effects on pollinators and floral antagonists. For example. 1 would expect nectar-robbing by invasive ants to have little effect on seed set across a plant popu- lation in an invaded area if there is little compe- tition for the resource, e.g., if nectar is available in excess of demand by pollinators. | would expect invasive ants to be more likely to decrease seed set in a plant species that has few floral antagonists and is dependent on a few species of small polli- nators, than in a species that suffers from many floral antagonists and has a diverse pollinator guild. Moreover, flowers in which nectaries are in close proximity to the stigma and anthers relative to the ants’ body size may be more affected by nectari- vorous ants than those in which nectaries are far- ther away. Seed predation and harvesting. Impacts of ants on seed dispersal and seed predation have received more attention than impacts on seed set. In myr- mecochorous mutualisms, ants disperse seeds away from the parent plant, often burying them in their nest, in exchange for the lipid-rich eliaosome at- tached to the seed (Buckley, 1982; Beattie, 1985). The dispersed seeds therefore may escape compe- tition with siblings and parents and are less sus- ceptible to predation and other threats, such as fire (Buckley, 1982; Beattie, 1985). Ants may also be seed harvesters, consuming the seed itself. Seed harvesters, however, do not eat all the seeds they collect, accidental seed 1982; Holldobler & Wilson, and therefore may act as dispersers (Buckley. 1990) Various species of ants may be attracted to eliao- somes and opportunistically take part in myrme- cochorous mutualisms, whereas harvester ants tend to have special adaptations for harvesting, consum- ing, and storing seeds (Hólldobler & Wilson, 1990; Keeler, 1989). In both cases, the ants need to (a) discover the seed, (b) recognize the seed as a re- source, and (c) be able to carry the seed back to the nest. Seed discovery will be dictated by ant foraging patterns, seasonal and temporal overlap between ant activity and seed availability, and the attractiveness of the eliaosome, if present (Beattie, 1985; Keeler, 1989). Seeds are typically high in lipids and proteins; recognition of the seed as a desirable resource тау depend on the relative abundance and composition of other food sources in the environment and the dietary requirements of the colony when the seeds are available (Beattie, 1985). In every myrmecochorous system studied, seed-dispersing ants comprise only a small subset of the ant species present, and their behavior to- ward seeds can vary from one day to the next (Beat- tie, 1985), suggesting that recognition of the seed as a valuable food resource is not constant among ant species or over time for a single species. The ability to carry the seed back to the nest will de- pend on the ants’ foraging behavior and size and shape complementarity between seed and ant (Kee- ler, 1989). Harvester ants must also have the man- dibular strength and agility to ingest the seeds (Beattie, 1985). How do we expect invasive ants to compare to native ants in these requirements? Traits of invasive ants that have formed the basis for hypotheses about other types of interactions discussed in this Volume 90, Number 1 2003 Lach 97 Invasive Ants paper, namely elevated abundance, aggression, and attraction to carbohydrates, may not have as much influence on seed predation and dispersal. Numer- ical superiority of invasive ants may increase the probability that they will discover seeds. However, common traits of invasive ants offer little capacity for predicting how invasive ants will respond to the seeds they encounter. Other traits that may be linked to seed discovery and dispersal may not vary consistently between invasive and native ants. So- enopsis invicta and Pheidole megacephala are at- tracted to oily, lipid-rich food sources (Vinson & Greenberg, 1986; Sanders et al., . and seed harvesters occur in both genera (Hélldobler & Wil- son, 1990). Thus, we might expect these invaders to be attracted to seeds, and even have some of the mandibular adaptations helpful for ingesting seeds. but perhaps no more so than native ants. OTHER VARIABLES AFFECTING ANT-PLANT INTERACTIONS As noted above, numerical superiority, aggres- sive behavior, and resource acquisition abilities of the ants are unlikely to account for all differences between invasive and native ant-plant interactions. Other biological traits that may vary substantially among ants, but are unlikely to vary consistently among invasive and native ants such as tempera- ture tolerance, daily activity patterns, colony cy- cles, and seasonal preferences for food types will certainly affect ants’ relationships with plants. Even for a single ant species, associations with plants may change spatially or temporally. For example, the degree of plant protection is largely linked to ant foraging patterns, which may change depending on the nutritional requirements of the colony (e.g.. in relation to reproductive cycles), and the avail- ability of carbohydrate and protein resources that require less foraging effort than those on the plant of interest (Ali & Reagan, 1985; Vinson & Green- berg, 1986; Stein et al., 1990; 1995; Cornelius & Grace, 1997). Attributes of the potential prey items, Homop- app & Salum, tera, and the host plant will also play roles in de- fining ant interactions with these organisms (Way, 1963; Cushman, 1991; Huxley & Cutler, 1991; Bronstein, 1994), and invasive and native ants may diverge in their responses to these variables as well. Many herbivores (Van Der Goot, 1916; Kaak- eh & Dutcher, 1992; Pavis et al., 1992; Gunawar- dena & Bandumathie, 1993; Cornelius & Bernays, i Montgomery & Wheeler, 2000; Brinkman et . 2001) and flowers (Kerner, 1878; er & ы 1997; Ghazoul, 2000) have Eos yeah ical or physical defenses against ants. Whether dif- ferences in ant size, recruitment ability, or toler- ance to phytotoxins, for example, will enable certain groups of ants to evade these defenses is yet to be seen. In cases where herbivores and flow- ering plants have no evolutionary history with ants and therefore might lack specific defenses against ants (e.g., Hawaiʻi), invasive ants might not need to be especially aggressive or adept at thwarting de- fenses to deter herbivores or steal floral nectar. TESTING THE PREDICTIONS WITH FIELD EVIDENCE Do observations of invasive ant-plant interac- tions support predictions that the distinct combi- nation of invasive ant traits will affect the outcome of their interactions with plants? Specifically, be- cause of their combination of traits, are invasive ants more likely than native ants to deter herbi- vores, tend Homoptera to the detriment of host plants, and interfere with plant reproduction? There have been few direct tests. Most research on ant interactions with herbivores has been done in agroecosystems, has focused on how pest popula- tions and crop yields are affected by changes in ant fauna, and has not explored any particular ant at- tributes influencing the interactions. Still less is known about how invasive ants affect seed set, pre- dation, and dispersal. Nonetheless, examples below provide some evidence of the importance of the common traits of invasive ants while also under- scoring the influence of ant biology and the role of other organisms in the system in determining the outcome of these interactions for the plant. ANT PROTECTION FROM HERBIVORES There are dozens of examples of invasive ants preying on or harassing would-be herbivores (see Holway et al., 2002, for a review). Almost none of these compare invasive to native ants, however, and rarely do they pinpoint any particular characteristic of the ant responsible for effects on herbivores. Some insights can be gained from cases in whic the invader fails to deter herbivores. For example, Anoplolepis gracilipes affords protection from the sucking bug, Amblypelta cocophaga (China), in Sol- omon Islands coconut when it reaches high abun- dance in the trees. But when prey are abundant on the ground, A. gracilipes fails to forage in the trees and premature nutfall ensues due to A. cocophaga 1971). In Keelers (1981) terms, in this case a high A, ability to deter her- damage (Greenslade, bivores, is inconsequential if the probability of for- aging on the plant (p) is low. Extrafloral nectaries may serve to increase p. 98 Annals of the MESE Botanical Garden a b efns la C. bignonioides efns lz C. bignonioides —— à » = A Pw - ^ A н: | D i ^w н: [р + e с + КЭ 2 = + + tone МАТ б a C LLLI Co — y | native ant > C. catalpae inyi ; A ; N + i+ herbivore enemies herbivore enemies ‘igure 3. The food-for-protection mutualism between the tree d Кастав and the dominant native ant (3a) and with Solenopsis invicta (3b). The direction of the arrows indicate direction of the effect and the positive and negative signs show the type of br The thickness of the arrow indic a the stre al of the effect. Dotted lines illustrate weak or no effect. I. H, A. and D are as described in Figure 1, efns = extrafloral nectaries. In За, the native ant and other enemies of the main flee. Ceratoma ea are attracted to extrafloral nectar and are effective reducing the herbivore attack on the plant. In 3b, S. la is not attracted to the extrafloral nectaries, deters C. atalar other natural enemies, but still decreases Ше шр of En herbivore on the plant. In both scenarios, the extrafloral nectaries are induced by herbivory. See text for reference Several studies have shown that different invasive invasive to native ants interactions with the tree ants are attracted to extrafloral nectar and can de- and reveals specific attributes of the invader that crease herbivory on extrafloral nectar-bearing are responsible for the differences. The production plants (see Holway et al., 2002, for a review). The of extrafloral nectar is an inducible defense in this high population density of Pheidole megacephala system: extrafloral nectar production increases in and its tending of extrafloral nectaries are blamed response to herbivory, thereby attracting native ants for the difference in populations of shrubs between and parasitic wasps (Ness, 2001). Field experi- P. megacephala and native ant-inhabited sites in ments have shown that S. invicta displaces native Australian rainforest. The native Urena is L. ants and preys on the parasitic wasps of the pre- and the introduced Senna obtusifolia (L.) H. S. Ir- dominant herbivore in the system, the caterpillar win & Barneby occur in dense stands and Е Ceratoma catalpae (Boisduval). Moreover, S. invicta little from folivory in areas invaded by P megace- visits extrafloral nectaries much less frequently phala, whereas in areas with native ants, the shrubs than native ants, most likely due to peak extrafloral are small, isolated, and heavily attacked by herbi- nectar production coinciding with the stage of col- vores (Hoffman et al., 1999). Other studies dem- ony cycle when workers prefer protein-rich resourc- poni that Solenopsis invicta (Fleet & Young, es. Solenopsis invicta is apparently as effective a 00), Linepithema humile (Koptur, 1979), and predator of C. catalpae as are the native ants be- ss auropunctata (Meier, 1994) also are at- cause it preys on pupae and pre-pupal instars as tracted to extrafloral nectaries on plants in their well as larvae, and because of its exceptional ag- introduced habitats, sometimes to the benefit of the — gressiveness when prey аге encountered. Thus, plant. But these studies, as with most, do not com- while S. invicta disrupts the mutualism by pre- pare the behavior of invasive ants to that of native venting other natural enemies of C. catalpae from ants, nor do they explore specific attributes of in- visiting extrafloral nectaries and protecting Catalpa vasive ants that may be influencing the interaction bignonioides, it does not do so at the cost of in- or its outcome. creased herbivory (Ness, 2001 Research exploring the interactions of Solenopsis Putting the example in the context of Keeler's invicta with Catalpa bignonioides Walter provides (1981) model (Fig. 3), even without the lure of ex- an interesting case study because it compares the — trafloral nectar, A,,, equals or exceeds А. I, sn Volume 90, Number 1 2003 Lach Invasive Ants need not be as high as L, „w; when Solenopsis in- victa is present, neither the invasive ant nor the parasitic. wasps visit the extrafloral nectaries as much as they are visited when native ants are pre- sent. In fact, because the extrafloral nectar is in- ducible. if S. invicta decreases the herbivore load more than native ants do, I, s;„ will be lower than ous Lhe difference between the benefits of re- duced herbivory and the investment in extrafloral nectar production will be higher when S. invicta is — present than when native ants are present. It is worth noting that p. the probability of lo- cating and foraging on the plant, may be influenced both by the abundance of the ant and its attraction to the nectar reward offered. As such, an extremely high density of Solenopsis invicta may offset its weak attraction to extrafloral nectar (Agnew et al., 1982). Indeed, high abundance of S. invicta has been offered as an explanation for the lack of any difference in its foraging frequency between nec- {апей and nectariless isolines of cotton (Agnew et al., 1982). Analogously, it is possible that high den- sity of S. invicta in the invaded Catalpa bignonioi- des stand (Ness. 2001) facilitated the ants’ foraging on the plant, notwithstanding the absence of its at- traction to extrafloral nectar. In these cases. if abi- ойс or other conditions ever result in diminished abundance of S. invicta, the invader may become less effective at deterring herbivores than ants that may not be as abundant, but are lured to the plants by the nectar reward. ANTS TENDING HOMOPTERA Anoplolepis gracilipes, Linepithema humile, Phei- dole megacephala, and Wasmannia auropunctata all have been noted for their ability to cause Ho- moptera outbreaks in various parts of the world (Holway et al., 2002). Solenopsis invicta may some- times obtain carbohydrates directly from plant tis- sue (Vander Meer et al., 1995), but also has been associated with increased Homoptera populations (Lofgren, 1986; Holway et al., 2002). The data col- lected in these studies often fail to discern any par- ticular trait of invasive versus native ants that are responsible for Homoptera outbreaks. While some evidence of the importance of abundance, aggres- sion, and attraction to carbohydrate resources in invasive ant interactions with plants is seen in the examples below, evaluation of my prediction that these common invasive ant traits play a major role in determining their interactions with plants will be enhanced by further detailed study. However, some studies do allow us to examine the relative impact on the plant of the invasive ant as Homoptera ten- der versus herbivore deterrent. These studies reveal that both types of outcomes outlined above occur, those in which the ant-Homoptera mutualism is parasitic toward the plant and those in which the ants tending Homoptera enter into a byproduct mu- tualism with the plant by deterring non-homopteran herbivores. Linepithema humile’s abundance. aggression, and tending of aphids have all been related to its ability to control populations of the pine proces- sionary moth (Thaumetopoea pityocampa Den. & Schiff.) in Portugal pine plantations (Way et al., 1999). The aphids attract foraging L. humile to pine tree crowns. Later in the season, pine processionary moth larvae on trees with L. humile are fiercely attacked, whereas those on native ant-occupied trees are ignored. Consequently, parts of plantations that are inhabited by L. humile escape the severe defoliation caused by T. pityocampa in native ant- inhabited areas. Figures 4a and b contrast the in- teractions with the plant when native ants versus L. humile are present in terms of Keeler’s (1981) modified model derived above. Reviewing interactions of invasive ants and Ho- moptera in cacao (Theobroma cacao L.) is a study in contrasts, and illustrates the possibility for dif- ferent outcomes to occur between the same plant and ant. In West Africa, Wasmannia auropunctata is actively spread among cacao plantations because it effectively deters pestiferous mirid bugs despite its association with high levels of scale and psyllids (Entwistle, 1972). In contrast, in its native Brazil, W. auropunctata tends the mealybug Planococcus citri (Risso) and many other Homoptera in cacao and fails to control pest herbivores possibly be- cause it does not always achieve dominance in the ant mosaic (De Medeiros et al., 1995; De Souza et il., 1998). In the parlance of the modified Keeler (1981) model. in West Africa, C,. the cost to the plant of hosting the Homoptera associated with W A. the auropunctata, is outweighed by the high ability of W. auropunctata to deter key herbivores, and the ant can be said to have entered a byproduct mutualism with cacao. But in Brazil, W. auropunc- tata tending P. citri imposes a high С, and a small A, resulting in a cost to the plant that exceeds the benefit. Similarly, Pheidole megacephala in its native West Africa tends the mealybugs, Planococcus citri and P. njalensis (Laing), which are associated with swollen shoot virus in cacao (Taylor & Adedoyin, 1978; Campbell, 1994). As part of its tending be- havior, P. megacephala transports soil from the ground to create tent shelters for the homopterans and thereby acts as a vertical vector of Phytophtho- 100 Annals of the Missouri Botanical Garden a b aphid C, „ pine tree aphid Ch „ pine tree | 4 EE 6 = - A H D H D = = + + = AS y 4 _ y native ants a. pityocampa L. humile T. pityocampa » 3 А Figure 4. The native ant— d pityocampa— pine tree relationship (4a) and the Wi nig humile—T. жан атра—р\пе tree relationship (4b). The direction of the arrows indicates the direction of the « the positive nd negative signs show the ie of effect. “The thickness of the arrow indicates the strength of the "effect. Dotted lines illustr te weak or no effect. C,, H, A, and D are as described in Figures 1 and 2. In 4a, the native ant is not attracted to the tree by aphids, fails to prey on the major herbiv vore, T. pityocampa, and the pine trees suffer from severe defoliation. Figure 4b illustrates the byproduct а between L. humile and the pine tree. A чынан humile is attracted to the trees by aphids and preys on T. pityocampa to the benefit of the plant. See text for references ra spores, the etiologic agent of black pod rot (Evans, 1973). In this case, C, comprises not only the direct cost of hosting the mealybugs, but also the associated increased likelihood of being infect- ed with two of the worst diseases of cacao. While no cacao studies have ever reported relationships between P. megacephala and non-homopteran her- bivores, it is unlikely that strong deterrence of her- bivores (a high A) would have gone unnoticed by farmers intimately aware of any effect on their yields. Pheidole megacephala's frequent association with high Homoptera outbreaks to the detriment of plant fitness places it firmly in the pest category in West Africa (Taylor & Adedoyin, 1978) In contrast, a low C, and high A characterize the interaction between Homoptera, cacao, and Anoplo- lepis gracilipes in Papua New Guinea. The ant re- lies on honeydew-producing Homoptera to maintain its abundance (Baker, 1972). It displaces the native tent-building ants that facilitate transmission. of black pod rot (Room & Smith, 1975; McGregor & Moxon, 1985; Way & Khoo, 1992). Moreover, un- like the ants it displaces, A. gracilipes harasses adult cocoa weevils and persistently disrupts egg laying and foraging of several mirid and coreid bugs, all of which are major pests (Entwistle, 1972; McGregor & Moxon, 1985; Way & Khoo, 1992). But the same ant fails to forage in trees in Malay- sian cacao and so is not effective against herbivores (Way & Khoo, 1989), indicating that a low p. prob- ability of finding and foraging on the plant, can undermine any ability to deter its herbivores. The examples above suggest that ant-Homoptera mutualisms that are parasitic on the host plant are more likely to occur in the invaders’ native range, and those in which the mutualism benefits the plant may be more likely in the introduced range. If this were the case, we would expect that studies that report on invasive ant-Homoptera mutualisms but do not explicitly evaluate the effects of the ants on herbivores, would nonetheless observe improved plant fitness. But this is not the case; when plant fitness is noted, it is usually reported as decreasing (Beattie, 1985; Holway et al., 2002), suggesting that in most cases the invasive ant-Homoptera mutual- ism is parasitic for its host plant. Beattie (1985) further points out that the vast majority of data de- scribing homopteran damage to plants is from agroecosystems, or other heavily manipulated en- vironments where the population of natural enemies of Homoptera may be quite depauperate. Discern- ing whether there are differences in how the inva- sive ant-Homoptera mutualism affects plants in the ants’ native versus introduced habitats and in ag- ricultural versus less manipulated environments is worth additional research ANT IMPACTS ON PLANT REPRODUCTION Куеп more so than with invasive ant interactions with extrafloral nectaries and Homoptera, the dearth of research into invasive ant impacts on seed sel, seed dispersal, and seed predation precludes attempts to test the predictions outlined above at present. A brief review of the current state of knowledge can, however, point to early trends and identify areas for future research. Although several species of invasive ants have been observed to forage in various flowers (Adams, 1986; Lofgren, 1986; Buys, 1990; Hara & Hata, Volume 90, Number 1 2 Lach 101 Invasive Ants 1992; Nicolson, 1994; Hata et al., 1995), little re- search has explored invasive ants’ attraction to flowers and their interactions with pollinators. Li- nepithema humile is known to exploit floral nectar otherwise taken by bees (Buys, 1987), is associated with lower insect diversity in Protea nitida Mill. flowerheads (Visser et al., ‚ and is hypothe- sized to reduce seed set by deterring pollinators from some crops (Potgieter, 1937; Durr, 1952). to date, no studies have been published that com- pare L. humile, or any other invasive ant, to native ants with respect to their attraction to flowers, in- teractions with floral visitors, and any subsequent impacts. The impact of invasive ants on seed predation and dispersal has received a little more researc attention. Linepithema humile in the South African fynbos is slower to discover eliaosome-bearing seeds than pu i ants that it displaces (Bond & Slingsby, 1 found, the invasive ants eat the eliaosomes but fail . Moreover, when the seeds are to disperse and bury the seeds, leaving them vul- nerable to predation and fire, and resulting in re- duced seedling emergence in invaded sites (Bond & Slingsby, 1984). A recent experiment reported that the disruption of this myrmecochorous mutu- alism results in plant community changes (Chris- tian, 2001), but this study did not take into account other potential effects of the ant on the plants (e.g.. pollination) that also might have been responsible for the observed differences between invaded and uninvaded sites. In Corsica, L. humile is more like- ly to find and remove Anchusa crispa Viv. seeds with eliaosomes than without, whereas the domi- nant native ants do not show a preference between seeds with and without eliaosomes in the seeds they remove. It is unclear what impact the differences in behavior between the invasive and native ants may have on A. се population dynamics (Quil- ichini & Debussche, 2 Pheidole megacephala ME Solenopsis invicta also appear to be attracted to seeds in their adopted habitats. Pheidole megacephala collects seeds of Acacia concurrens Pedley in Australia with un- known consequences for the tree (Majer, 1985). So- lenopsis invicta is a predator of seeds of numerous crops and other plants (Ready & Vinson, 1995; Morrison et al., ) and is attracted to most eliao- some-bearing seeds, but often destroys or scarifies them (Zettler et al., tata is a poor seed disperser in its native Mexico (Horvitz & Schemske, 1986), but there are no pub- lished studies that examine its seed harvesting or 2001). Wasmannia auropunc- dispersal tendencies in the habitats it has invaded. As with Linepithema humile, further study is nec- essary to determine the impacts of P. megacephala S. invicta, and W. auropunctata on seeds relative to native ants, and any implications for plant com- munity dynamics. WHY SHOULD WE STRIVE FOR PREDICTIVE CAPACITY? The ability to predict the nature and outcomes of interactions between invasive ants and plants has the potential to yield many rewards. Agricultural- ists and conservationists, in particular, may benefit by knowing whether a potential invader is likely to have impacts that will thwart their management goals. This foresight may provide further impetus for putting policies in place to prevent the invasion. cases where the invasive ant is already pre- sent, identification of the specific ant traits that in- fluence the outcomes of invasive ant interactions with plants may reveal options for mitigating un- desirable effects. If a key to many effects of inva- sive ants is their extreme abundance and penchant for carbohydrates, complete eradication from an area may not be necessary to prevent outcomes that are counter to land management goals. Agricultur- alists have realized the link between invasive ants and Homoptera outbreaks, for example (Flanders, 1951; Prins et al., 1990; Reimer et al., venting the ants from foraging in trees via use of sticky barriers or other means decreases the Ho- moptera population below pest levels and deprives anls о 1990). present they no longer contribute to yield loss via source of carbohydrates (Samways, Трек although the ants may still be their relationship with homopterans. Deprived of their carbohydrate source, they may not be abun- dant enough to effect other outcomes on the plants or native ant community (Addison & Samways, 2000). Foraging conditions can also be manipulated naturally so that the ants are less pestiferous, or even aid in achieving management goals. In Zan- zibar coconut, Pheidole megacephala is a pest be- cause it does not prey on the primary agent causing premature nutfall, the bug Pseudotheraptus wayi ~ Brown), as much as the native weaver ant it dis- places. Pheidole megacephala is attracted to the palm crown by nectar, pollen, and various homop- terans. Retaining ground vegetation in palm plan- tations provides P. megacephala with enough for- aging opportunity that it does not ascend the palm trees in search of food. The weaver ant is then free to inhabit the canopy and prey on P. wayi (Rapp & Salum, 1995). Similarly, in Louisiana sugar cane, allowing broadleaf weeds to grow in the early part Annals of the 102 | Garden issouri Botanica M pon ien 10и si— “Coatueas jue es sasuaJap цим ѕрәәѕ səənpoıd— лив 10 :ѕрәәѕ Bul -euep Jo Seu. 10и— 'spaas о} P3198118 10и — 151108 ӘЛІЅРАЦ| рәищші|-рәәѕ si— рче "AIOATUEIS ив jsurese ѕәѕиәјәр әлцәәујә INOYILM spaas soonpoud— йч pue tura ZuBe -wep Jo ajqedero sr pue ѕрәэѕ 0] рәјәвлце si— que oAIsPAU[ [ps1odsip paas 10у sue uo juop -uədəp Аүәлциә jou si— пир 10 ‘SUE әлцеи 3ursiodsip-paas paoeydsip A[|91o[duroo jou sey pue ‘spaas 0] рәве jou si— jue әлїввАЦ] [esıadsıp рәәѕ 10] sue uo juapuadap si— -siodsip pu? suljpuey рээѕ 1e 100d si pue spaas 0} pa1oeErje SI— 10 ‘sue әлпви Зиѕлә1р pees paoe[dsip 4[o1o[d -W09 sey pue ѕрәәѕ 0] рәјәвле jou si— jue OAISPAU] poium-uoj[od you si— Jo *o]oduioo jou op siojeurjjod pue SUE jeu] os тедәп Hiis soonpoad— ‘SUE LINE 0] IP]22u әр 00] 10 ел sey— 10 ‘SUE jsulese ѕиәЈәр [елор 24129jJ2 A[peo4q seu— чавд 10 SJUR QAISPAUI JO әәцәѕәла эці ur uo[[od ләувирд 01 әре [[ns— 10 “әѕләлір pue juepunqe— EAH S10BU1[ [Od 10 :ua[[od SULLIOJSUB.L] шолу s10]eurj[od ләләр jou вәор— 10 *sayRAPAYOGIRD SUNS jou si— 10 *110]J2. Suises0j ssa] saiimnboa 10 9A19P.P әлош st wy 321008 ƏPIPÁYOqILÐ Р 0} ss92oP seq— ЛИВ әлївюли] рэишц-иәцоа si— pue *Ajr[rqepmeae unipour JO 1e]29u VATPORIYE seu— pue ‘ив Jsureze sosuojap үелор ou sey— jue|d pue ‘SUE 9AISSAJd2P AG xo]? 0] o[qu1ou[na— 10 *juepunqe jou— тәле SJOU Od pue :1ојеит[оа jo үиәлдәтәр ƏANƏƏJJƏ ue— pu? чәмор әң 01 pojogue— pue *sajeap&ugoqaeo 3urqoos— pu? 'sioquinu цат ur juasa1d— 151 jue oArsPAu] ашриә -JUR цим әввәләш jou op— IO ‘SUE 0] ƏAPLIPE JOU әле— Rio} dowoy 10 ‘Sər -9ua uv19]douiou jo qual -19]3p 34123jJ2 ир JOU si— 10 *S9]P1p -Ayoqurs BULYIIS JOU SI— 10 “моурә Surses0y ssa] soumboa 10 әлцәвцүр әлош ST Jey} 221n0s әјелр -&goqueo в 0] ssa2oP sey— jue OATSRAUT gouRpunge вләзаошон Jo BIBEJE) 19410 10 suasoyyedoyAyd p21eroossp-P19] -dowo 0} 3[qg1ou[n4 81— пир ‘зшриә JUR 2AISPAUI шолу ]jouaq— pue ‘мәрләиоц 2onpo1d— R19) dowo py pur E -әцә [R1njeu вләзаошоң JO їчәлә]әр әлцәәрә ue— pue "s9]PapAuoqaeo 3ubaas— pue ‘sdaquinu uar ur juasa1d— 151 jue QAISPAU] Ыл БП jsulese Səsuəş -Əp OAT). 29]]2 19410 sey— ШЫК 10/PUE PLENE Ju? 0] 3[qe A9u[n A— 10 * ITB I— isl IIOATGIOH 94129]JJ2 әле Jeu] SUE 12410 əəejdsıp jou saop— jue o4rspAu] УӘЈОА -1{19Y jsure3e. ѕәѕиәјәр 241]23]J2 112110 zue] sI— jue] pue КЫТ Jsureae səsuəjəp seu— :810]22]04d 3941]23]J2 әлә y} Sue 1910 soopdsrp— pue “Х\шәпцә ив se 3104 -IG19Y oziuzoo21 JOU вәор— 10 "jue[d 0} pojogajje jou si— que ӘЛІЅВАЦ jue[d 10у 3UIO2]nO 3813۸ -pe jo ysu мо jue]d 10} 3UIO2]nO ƏSIƏA -ре Jo ysu үң uonppaad рәәс [esiedsip рәәс 195 рәәс ашриәј вләјаошоң SITOATGIOY шол) U01]29]044d 'SUOT]J2EI9]UI juejd- que Jo sad] элу 10] JUB JAISE: Aut ue чим UOIT]ETOOSSE шолу ]youoq Аеш 10 әшоәэпо эѕләлре ив Je уз MOJ 10 ча Р ӘАРЦ Аеш шега [Pnpr^rpur ue Ч: Uu M Aopun S}X9JUOD pezisoyjoda yy Ка 9199], Volume 90, Number 1 2003 Lach 103 Invasive Ants Continued. Table 2. Seed predation Seed dispersal Seed set Homoptera tending Protection from herbivores Invasive ant: Invasive ant: [nvasive ant: Invasive ant: Invasive ant: Potential for plant —is not attracted to —displaces other ants that —displaces other ants that are better —is attracted to seeds, —is highly attracted to to benefit seeds, or —is not capable of dam- and —is able to disperse and nectar-robbers and pollinator deter- are better Homoptera ten- ders; or —enters a byproduct mutu- plant, and —is effective at displacing aging seeds; and —displaces other ants —is seeking carbohydrates, and —is attracted to the flower, and —does not deter pollinators from herbivores shelter them more ef- fectively than any dis- placed ants alism with the plant (see text) that are seed harvesters 5 transferring pollen; and transfer more pollen in the presence of the invasive ant (1.e.. creased repositioning frequency) through in- of the season facilitates an increase of S. invicta populations because it provides abundant prey (Ali et al., 1984). As the weeds die back with the clos- ing of the sugarcane canopy, the ants transfer their foraging to the sugarcane where they are effective predators against the sugarcane borer, Diatrea sac- charalis (Е) (Ali & Reagan, 1985) SYNTHESIS AND CONCLUSIONS Traits associated with invasive ants, namely el- evated abundance, aggression, and attraction to high-carbohydrate resources directly and indirectly affect outcomes of interactions with plants. Abun- dance of invasive ants tends to be higher than na- tive ants, but as shown in the examples, it is where. how, and when the bulk of these abundant workers tend to forage that ultimately influences interaction outcomes. As described above, while both Anoplo- lepis gracilipes and Pheidole megacephala can be abundant in coconut palm plantations, outcomes for coconut palms are quite different depending on whether conditions favor ground or arboreal forag- ing of the ants. However, the relative scarcity of Solenopsis invicta on Catalpa bignonioides appears not to be a factor in the protection of the plant from its main herbivore because it is offset by the ant's foraging efficiency. Invasive ant colony cycles and needs for a particular resource play a major role in determining foraging behavior, and may or may not coincide with plants’ or Homoptera carbohydrate production and need for protection. In the case of S. invicta on C. bignonioides, had it been the time of year when S. invicta seeks carbohydrate-rich re- sources, or had the extrafloral nectar contained a profusion of amino acids, perhaps visitation fre- quency would have been higher and an even great- er decline in herbivory witnessed (Ness, 2001). For host plants that suffer from a less palatable or vul- nerable dominant herbivore or interact with a less aggressive ant, phenological overlap between ants’ and plants! needs and rewards may be more critical to the outcome of the interaction (Bronstein, 1998). The other invasive ant traits that I have focused on, aggression and attraction to carbohydrates, also deserve further examination. In predicting effects of invasive ants on plants, it will be useful to know what biochemical or other cues trigger the ants' ag- gressive behavior. Why, for example, does Anoplo- lepis gracilipes displace the coconut bug Amblypelta cocophaga in Solomon Islands coconut (Greensla- de, 1971), but fail to deter the coconut bug Pseu- dotheraptus wayi in Zanzibar coconut (Way, 1953)? ikewise, carbohydrate resources can vary greatly in composition and therefore in their attraction of Annals of the Missouri Botanical Garden ants. In any plant community, an ant may have the opportunity to choose among various floral nectars, The options and preferences may vary over the season extrafloral nectars, ог homopteran exudates. and result in different outcomes for the individual plants and the plant community as a whole. For example, aphids attract Solenopsis invicta to nec- taried cotton plants early in the season, but later in the season extrafloral nectar is preferred and the ants are seldom observed tending aphids (Agnew et al., The amount and types of amino acids and sugars, and the balance among them have all variously been suggested as affecting attraction of ants to nectar and homopteran exudates (Lanza, 1991; Lanza et al., 1993; Koptur, 1979; Vander Meer et al., 1995; Koptur & Truong, 1998). Further research may reveal whether invasive ants are at- tracted to certain carbohydrate sources more than native ants, and why an invasive ant may be more attracted to one carbohydrate source over another, as well as how any observed preferences change with other variables (e.g., availability of other re- sources, ant colony needs). Elucidating the attractiveness and availability of different carbohydrate resources is likely the key to predicting when and where invasive ants will be found on plants and, therefore, the potential for a host of subsequent effects. Of the three attractants, Homoptera, extrafloral nectar, and floral nectar, it appears Homoptera are the most important lures for the invaders because of their near ubiquity and broad attractiveness to most invasive ants. Extra- floral nectar, while considered generally attractive to ants (Carroll & Janzen, 1973), is not as widely available. Moreover, plants that possess inducible extrafloral nectaries may offer nectar too inconsis- tently to appease the sweet tooth of invasive ants. Floral nectar is certainly common, but is perhaps not as apice) attractive to ants unless floral defens n be thwarted (Guerrant & Fiedler, 1981; (шл. 2000). Once attracted to the plant, the invasive ant may deter herbivores, as part of the food-for-protection mutualism with extrafloral nectaries, or as part of a byproduct mutualism, an indirect effect of the ants’ association with Homop- tera. But an analogous byproduct parasitism, al- though not yet reported in the literature, is not im- plausible; invasive ants lured to plants by Homoptera or extrafloral nectaries may deter pol- linators or a key herbivore enemy. Since evolution did not play a role in shaping the interactions be- tween invasive ants and plants in their adopted habitats, we should not limit ourselves to consid- ering only those outcomes that would be evolution- arily plausible. That invasive ants can exert both positive and negative effects on plants precludes any general- izalion about the ultimate impact of invasive ants on plants. Indeed, the same ant may affect different parts and processes of the same plant in different ways at different times (Lofgren, 1986). Nonethe- less, the existing data offer a starting point for pre- dicting the context in which one outcome may be more likely than another for each type of interac- tion. In Table 2 I offer predictions about under what conditions an individual plant may face a high or low risk of an adverse effect or may benefit from association with an invasive ant for interactions in- volving plant protection, Homoptera tending, seed et, seed dispersal, and seed predation. A plant 2 may simultaneously have a high risk of adverse out- comes in some categories and a low risk or poten- tial to benefit in others. Net effects will depend on whether impacts combine or compensate for each other. For example, the reproductive capacity of a plant may be severely curtailed if an invasive ant both disrupts pollination and interferes with seed dispersal. Alternatively, there may be no net effect on a plant's reproductive success in the presence of an invasive ant that increases seed set and in- terferes with seed dispersal (e.g., Horvitz & Schem- ske, 1984). Plants may have their own compensa- tory mechanisms; angiosperms that rely on a few specialized pollinators may be clonal or extremely long-lived, thereby decreasing dependence on seed production (Bond, 1994) The innumerable combinations of effects and vaslly different potential net outcomes demonstrate the need for research that takes into account the multiple mechanisms through which invasive ants may affect plants, and the consequences for the The complexity and context-dependency of ant-plant in- plant at population and community levels. teractions make predicting the net effects of inva- sive ants on plants a formidable challenge. None- theless, incorporating characteristics common to invasive ants into pre-existing models of ant-plant interactions provides a framework for generating testable hypotheses about how invasive ants may interact with plants and the consequences for the plant. Testing these hypotheses will contribute di- rectly to conservation and agriculture, and provide insights applicable to both invasion ecology and our understanding of ant-plant interactions. Literature Cited Adams, C. T. 1986. Agric woe and а al impact of the imported fire ants. 57 in ofgren & K. Vander Meer (editors), Fire and Lat Cutting Ants: Volume 90, Number 1 2003 Lach Invasive Ants 105 Biology and Management. Westview Press, Boulder, Colorado. € J. К. 1978. Paige for mutualists: Aphids n Canad. J. Zool. 56: 2093—2096. deinen ‚ Samways. 2000. A survey of ants oe Formic idae) that forage in vineyards in the Western = Province, South Africa. African En- tomol. 8: 251—260 Agnew, C. W., №. L. p D. A. Dean. 1982. In ence of cotton nectar on red imported fire ants and "ad poo ек Entomol. 11: 629—634. J). & T. Ali, A E. Reagan. 1985. Vegetation manipulation dede on predator and prey populations 1 in Louisiana, USA sugarcane ecosystems. J. Econ. Entomol. 78: 1409-1414. . Flynn. 1984. Influence of se lec ted weedy and dint free sugarcane habitats on diet composition and foraging activity of the imported fire ant (Hymenoptera: Formicidae). Environm. Entomol. 13: 1037-1041. Bach, C. E. 1991. Direct and indirect interactions be- tween ants (Pheidole megacephala), scales (Coccus vir- idis) and plants (Pluchea indica). Oecologia 87: 233— 239 Baker, G. 1972. The role of Anoplolepis longipes Jerdon (Hymenoptera: Formicidae) in the entomology of cacao in the northern district of Papua New Guinea. 14th In- en MON Congress, August 22—30, Can- berra, Australia. Beardsley, J. 29 Jr., T. H. Su, F. L. McEwen & D. Gerling. 19t ield Vicar e of the interrelationships of the big-headed ant, Pheidole megacephala, the gray pineapple seen Dysmicoccus neobrevipes, and = pineapple mealybug wilt disease in Hawai‘i, USA. P Beattie, A. J. Plant U. The Evolutionary Ecology of Ant- ved ied Cambridge Univ. Press, Cambridge, Bentley, B. L. 1977. Extrafloral nectaries and pe by pugnacious bodyguards. Annual Rev. Ecol. Syst. 8: 401-421. TL . V. Suarez, К. R. Crooks, 5. A. Morrison | T. J. Case. 2000. pe ame in ao habitat frag- vali a, nd edge ef- fects. Ecol. Applic. Bond, W. J. 1994. Do mutualisms jatte r? Assessing the impact of ти and disperser disruption on plant — Philos. Trans., Ser. B 344: 83-90. P. Slingsby. 1984. Collapse of an ant-plant mu- раг: he Argentine ant (/ridom прик humilis) -4 myrmecochorous Proteaceae. jq a : 1031-10: Breton, L. M. & J. F. Addicott. ig eres аа + ce ipee in ап aphid-ant siio Ecology 2175-2180. i M. A., W. A. Gardner & G. D. Buntin. 2001. Effect of red impo eg fire a ut bosco Formici- dae) on Rhinocyllus conicus (Coleoptera: Curculioni- dae), a ape pe control agent of musk thistle. Biol. Control 30: 612—616. Bristow, C. M. es та benefits ian ant atten- iid to two species of Hom ew York iron- ed. J. Anim. Ecol. 53: 715-72 Bronstein, J. L. 1994. Du RE outcomes in mutual- istic interactions. Trends Ecol. Evol. 9: 214-217 1998. The contribution of ant- sn protection ides to our understanding of mutualism. Biotropica 30: 150-161 the trade-off. Ecol. E^ Buc kley, R. C. 1982. Ant-plant interactions: A world re- view. Pp. 111-142 in R. C. Buckley (editor), Ant-Plant Interactions in Australia. Dr W Junk, The Hague. 987. and кең Annual Rev. Ecol Buys, B. 1987. Competition for nectar ey Argentine ants (Iridomyrmex humilis) and hone s (Apis melli- fera) on black DC Sr i LN S. Af- ric m > ol. 2 Moa bi cea between A ti ts and honeybees i in South Africa. Pp. 519-524 in R. K. Van- & Dendeno (editors), Applied Myrmecol- ogy: А World Perspective. Westview Press, Oxford. Cammell, M. E., M. J. Wa D M. R. Paiva. 1996. Diver- sity and structure of ant communities associated with oak, pine, eucalyptus me arable habitats in Portugal. Insectes Sociaux 43: ت‎ Campbell, C. A. M. tera associated with the ants Crematogaster clarivenris Pheidole megacephala and Tetramorium aculea ирин Formicidae) on cocoa in rige Bull, Entomol. Res. 84: 313-318. Ca rroll, ( С. R. & D. J. Janzen. 1973. n of foraging by ants. Annual Rev. Ecol. Syst. 4: 231— Christian, C. E. 2001. Consequences of a po in- vasion reveal the importance of mutualism for plant communities. Nature 413: 635-6: Cole, F. R., A. C. Medeiros, L. L. Loope & W. W. Zuehlke. 1992. Effects of the Argentine ant on arthropod fauna of Hawaiian high-elevation shrubland. Ecology 73: 1313-1322. е S. G. & Н. С. Robertson. 1988. Complex in- actions between mutualisms: Ants tending homopter- ans protect fig seeds and pollinators. Ecology 69: 1302— 1305. Conant, P. & C. Hirayama. 2000. Wasmannia auropunc- tata (Hymenoptera: res an Established on the Is- land of Hawaiʻi. Mus. мру + 64: 21—22. Cornelius, М. ТА, Вегпауз. The effect of plant chemistry on the а if caterpillar prey (Hymenop- Mn of bro on the ийнана preferenc es of the tr tropical ant species, Phei- dole бк усен v dio ser ыан p (Mayr). J. Entomol. Sci. 3 Cushman, J. H. ct T X NA mediation of insect mu- ualisms: “о outcomes in herbivore-ant interac- tions. pi 01: 8-144. & J. es 1989. Intra- and interspecific өйне s for mutualists: Ants as a limited and lim- iting = ч Jos gees Oecologia 79: 315-321. Davidson, D. 997. The role of resource imbalances in the evolutionary ecology of tropical arboreal ants. Biol. J. m Soc. 61: 153-181. . Resoufce discove ination in pete A fun versus resource dom- ctional ч for breaking & D. McKey. 1993. die ا‎ чаре ecology of symbiotic ant-plant relationships. J. Hymenoptera Res. 2: 13-83. De Medeiros, M. A., H. G. Fowler & O. C. Bueno. 1995. Hym., Formicidae) mosaic stability in Bahian co- coa crm — VES for management. J. Appl. Entomol. 119: 4 4. ‚Н. С. Delabie & Н. C. Fowler. De Souza, A. 1998. се spp. (Hym., Formicidae) and insect 106 Annals of the Missouri Botanical Garden damages to cocoa in Brazilian farms. J. Appl. Entomol. 122(6): 339—341. Del-Claro, К. & P. S. Oliveira. 1996. Honeydew flicking by oe iss cues to potential lending ants. Anim. Be cs : 1071-1( э 5. Deyrup, M., Di avis & S 2000. Exotic ants in Florida. s Amer. E. E. 126: 293—320. Donnelly, D. & J. Н. Giliomee. 1985. Community struc- ture of epigaeic ants (Hymeno ын "onis 'idae) in fyn- ча vegetation in the Jonkershoek Valley. J. Entomol. rica 48: 247-257. шо Roc һа, C. F. & Н. Godoy Bergallo. 1992. Bigger ant colonies reduce herbivory and herbivore residence time on leaves of an ant-plant: Azteca muelleri vs. Coe- lomera ruficornis on Cecropia pachystachya. Oecologia 9]: 249-252. 1952. The Argentine. Par ome т. Farming South Africa 27: 381—384. Eisner, T. 1957. А comparative оне 2 study of the Qui ulus of ants (Hymenoptera: Formicidae). Bull. Mus. Comp. Zool. 116: 429—490. Entwistle, P. F. 1972. Pests of Cocoa. Longman Group, Londor Eric bu. T M. 1971. The displacement of native ant species by the introduc ed Argentine ant Iridomyrmex humilis Marr Psyche 78: 257-266. Eubanks, M. D. 2001. Estimates of the direct and indi- rec er ‘ts of red imported fire m on biological con- oli in field Biol. Control 21: 3543. Evans, H. C. 1973. Invertebrate vectors of oe palmivora causing black pod disease of cocoa in Ghar Biol. 75(3): 331-345. W. Bon. Jr. 1978. The recent intro- duction of the pest ant Wasmannia ak halon into New Caledonia. EA Austral. Entomol. Soc. 17: 139-142 . Exotic ants in Asia: Is the mainland di case fee Hong Kong. Aliens 9: 5-6. 1 Maschwitz, T. Y. Pong & A. J. Helbig. 1989. Studies a a south east Asian ant-plant association: Pro- tection of Mac phu trees by Crematogaster ой най, зе сү ТӨ: —470. Fischer, a Hoffmann & W. Völkl. 2001. Com- petition s youtinlisis à in an ant- -homopteran interaction mediated by hierarchies of ant attendance. Oikos 92: 531—541. Flanders, S. E. 1951. The role of the ant in didis е conia of homopterous insects. Canad. Entomol. —98. Fleet, R. R. & . Young. 2000. Facultative mutualism between ond fire ants (Solenopsis invicta) and a legume (Senna occidentalis). SouthW. Naturalist 45: 2 a Fluker, S. . W. ә eardsley. 1970. Sympatric asso ciations e dc ants: Iridomyrmex humilis, Pheidole му Bd rii. “ое in Hawaii. Ann. Entomol. Soc. Am )–129( Fo Es H. G., J . Bernardi, J. С. Delabie. L. C. Forti Pereira- bus 1990. Pad ant ш of South America. Pp. 3-14 i nder Meer & А. Dendeno (editors), е reus A World Per- spective. Westview Press, Oxford. Galen, С. 1999, Flowers and enemies: Predation by nec- tar-thieving ns in relation to variation in floral form of an alpine wildflower, Ош. viscosum. Oikos 85: 426-434. Key & M. Anstett. 1997. Benefits con- Gaume, L., D. Mc ferred by “timid” ants: Active anti-herbivore protection Huxley, C. R. of the rainforest tree Leonardoxa africana by the minute ant Petalomyrmex phylax. Oecologia 112: 209-216. — ——, — — & S. Terrin. 1998. Ant-plant- homopter- an mutualism: How the third partner affects the action between a plant-specialist ant and its myrmeco- la 2 е Roy. Soc. London, Ser. B, Biol. G ало, үн 2000. tive. Bull. Brit. s A N.J.&A т. 2 PE EAM effects red fire “ усы Ecol. Lett. : к P. T., YDowd & P. S. bis )99. Alien ant invasion ix ecosystem "ri on ош Island, 9: Re mae Wi rs keep plants attrac- 4, Phi 1971. pu competition a frequency оке among ants in Solomon Islands conut due J: n Ecol. 8: 323—352. Guerrant, & P. L. Fiedler. 1981. Flow e. against nec ea by ants. Biotropica (Suppl.) 1 25-33. ) 6 1993. De- g Leptocorisa oratorius Fa- (Hemiptera: Coreida e) А unique chemical combination and its toxic, = жее, and alarm prop- erties. J. Chem. Ecol. 19: 851—861. Haines, I. H. & . Haines D. Pest status of the crazy ant, Anoplolepis longipes (Jerdon) (Hymenoptera: Formicidae), in the Seychelles. Bull. Entomol. Res. 68: 027—638. Hara, A. H. & T. Y. ros in Hawaii. y iu : 155-163. Haskins, C. . FE ene 1965. Pheidole mega- cephala and „зе humilis in Bermuda: а uilib- rium or slow replacement? Ecology 46: 736-74 Hata, T. Y., A. H. Hara, B. K. S. Hu, R. T. L. Tenbrink. . K. Bandumathie. ensive secretion of rice bu bricius, Gunawardena, fe 1992. Ant control on protea Kan is & V. 1995. Excluding pests from red ginger flowers with insecticides and pollinating, T or 397. polyethylene ue J Econ. Entomol. 88: 393- Hoffman, Andersen & G. J. E: ш. 1999, Impact of an роде 'ed апі оп native forest inverte- brates: Pheidole megac ephala in beue Australia. Oecologia 120: 595-604. Hélldobler, B. & ЖЯ О. к 1990. The Ants. Belknap Press, Cambridge, Massachusetts Holway, D. A. 1995. The ыш еа of the Argentine ant мш: humile) in northern California. Con- servation 2 jl. 634-1637. ———, n 'tors controlling rate of invasion: A nat- ural En using Argentine ants. Oecologia 115: 06-212. — ———, L. Lach, A. V. Suarez, N. D. Tsutsui & T. J. ( 2002. The causes and eae et es of ant invasions. Annual Rev. a Syst : 181-233. Horvitz, C. C. . Se bd 1984. Effects of ants and an ant- d herbivore on seed production of a Neotropical herb. C 65: 1369-1378. — —— & — ——. 1986. Seed dispersal of a Neotropical myrmecochore: Variation in removal rates and dispersal distance. Biotropica 18: 319-323. "hung Н. T. & F pee . BioScience 37: 665-67 Human, K. G. & D. M. E 1996. Exploitation and i e edt between the invasive Arge n- tine ant, ‚айр Шей humile, and native ant species. 40 ea The ancient cultured cit- l. Oecologia 105: ). F. niler: 1991. Ant- Plant Interac- tions. Oxford Univ. an ress, Oxford, U.K Volume 90, Number 1 2003 Lach 107 Invasive Ants Jolivet, Р, 1996. Ants and Plants: An Example of Coevo- lution. Backhuys, Leiden, Netherlands an H. 1997. Threats on Pacific islands: The spread of the tramp ant Wasmannia auropunctata og ). Pacific Conservation Biol. 3: 61—6 W. & J. D. Dutcher. 1992. aging pre radi e of red imported fire ants (Hymenoptera Formicidae) among three species of s er cover crops and their extracts. J. Econ ат "85: 389-394. Keeler, K. H. 1981. A model of selection for facultative nonsymbiotic mutualism. Amer. Naturalist 118: 488- . 1989. Ant-plant interactions. Pp. 207-242 W. G. Abrahamson (editor), Plant-Animal debo MeGraw-Hill, New York. ‹ 1878. Flowers and Their Unbidden Guests. Paul, London. > L. & T. J. de Jong. 1993. Attractive- ness to айар A plant's dilemma. Oikos 66: 180— 184. Koptur, S. 1979. Facultative mutualism between weedy vetches bearing pp qe nectaries and weedy ants in California. Amer. : 1016-1020. —— ———. 1989. Is extn ape е tion an induc- ible defense? Pp. 325—339 in J. H. Bock & Y. B. Lin- hart (editors). The luna оное? of Plants. West- vie w Pro ss, Boulder, Color Truong. 1998. p ant-plant inter- actions: Necta ar sugar preferences of introduced pest ant species є South Florida. Biotropica 30: 179-189. Lanza, J. . Response of fire ants (Formicidae: Sole- nopsis invicta and S. geminata) to artificial nectars with amino ac ‘ids. Ecol. Entomol. ай 203-210 І Vargo, S. Pulim 1993. ыбы of the fire ants bao invicta and S. грин (Hymenoptera: Formicidae) for amino acid and sugar components of extrafloral nectars. Environm =з 22: 411-417 Letourneau, D. 1983. Passive aggression: An alter- native hyy thesis р the Piper-Pheidole association. Oecologia 60: 26. Lewis, T., J. M 2 З LI ® -— E sein I. Haines, J. B. Haines & P. L Май. 1976. The crazy ant, reri longipes (Jerd.) (Hymenoptera, Formicidae) in Seychelles, and its chemical control. Bull. Entomol. Res. 66 11 Linsenmair, e E., M. Heil, W. M. Kaiser, B. Fiala, T. Koch & W. Boland. 2001. a to biotic and abiotic stress: oe -ant plants timize ment in biotic defense. . Bot. 52: 2057— Lofgren, C. S. 1986. The ec is importance Li con- trol of —— fire ants in the United States. Pp. 227— 256 in S. B. Vinson (editor), Economic Impact and Con- trol " Be sial Insects. Praeger, New York. Lubin, Y. D. 1984. Changes i in the native fauna of the els Islands following invasions by the little red asmannia auropunctata. Biol. J. Linn. Soc. invest- Majer, J. D. 1985. Hec ие by ants of rehabilitated mineral sand mines on North Stradbroke Island, Queensland, Australia, vi. e ular reference to seed = Austral. J. Eco 48. 994. Spread of Argentine ants мые co hu- mile) ens special reference to Western Australia 163-173 in D. F. Williams (editor), ору. Impact and Control of Introduced Species. West- view чеч T oulder, Colorado. May, J. F . E. Heterick. 2000. Effects of the coastal brown ant Pheidole megacephala (Fabric л оп a ant fauna of the Perth metropolitan region. TT lia. — Conservation Biol. 6 81-8 McGlynn, 1 The worldwide MP of ants: Ge nio. distribution and ecological invasions. J Biogeogr. 26: 535—548. McGregor, . E. Moxon. 1985. Potential for bio- мМ вов) of tent building species of ants associ- ated with Phytophthora-palmivora pod rot of cocoa in Papua New -o nea. Ann. Appl. Biol. 107: 271-278. Me С К. E. 1994. Coexisting = and foraging be- havior of introduced and native ants (Hymenoptera For- mic ав in the Galapagos Ко (Ecuador). Pp. 44— 61 D. Williams (editor), Exotic Ants: Эа ый Pon ct and Control of Introduced Species. Westvi Press, Boulder, Colorado. Messina, F. 81. Plant protection as a EEE of an ant- speculate ча mutualism: Interactions on go enrod vine pi Ecology 62: 1433—1440. Montgomery, S. Wheeler. 2000. Antipredatory activity of the weevi i ЫЬ vitiosa: А biological control "re of Melaleuca quinquenervia. J. Insect Behav. 13: 915—926. Morales, ustra- M. A. 2000. Mechanisms and density depen- dence of benefit in an ant-membracid mutualism. Ecol- us D. К. Williams, D. Н. 7. Damage to dry crop seed by a Tene fire ant (Hymenoptera: Formic idae). J. Econ. Entomol. 90: 218-222. Ness, J. H. 2001. The Catalpa bignonioides Food Web: esencia of Variable Interactions Among Four Tro- p Levels. Ph.D. Thesis, University of Georgia, Ath- Nicolson. 5. W. 1994. Eu — nectar: domi P availability, composition and os conse the larva of the e -— nectar Г е Ме бам ta. S. African J. Sci. 9 —80. Passera, L. ues teristics of tramp specie à 23-43 in D. F. Williams (editor), Exotic Ants: Bio a Impact and Control T Introduced Species. Westview Press, Boulder, Colora t F. adr K. Jaffe & of first-instar larva of rootstalk borer weevil, Di (Coleoptera: Curculiondae) to geminata F) (Hymenoptera: Formicidae). J. Chem. Ecol. 18: 3 re Peakall, N. Handel & A. J. Beattie. sept D and pane of, ant pollinatio n. Pp. 421—429 in C. R. Huxley & Cutler ec y res Plant гони Oxford E LT Oxford. 990. Invasion of po- 1991. The . The Argentine ant. Farming South Prins, A. J., Н. z Robertson & A. Pri 1990. Pest ants in urban and атор агеаѕ “ is Per Africa. Р 25-33 in R. К. Van e . Dendeno (editors). Applied ань А World Perspective. Westview Press, Oxford. Puterbaugh. M. N. 1998. The roles of ants as flower vis- itors: Experimental analysis in three alpine plant spe- cies. Oikos 83: 36—46. Quilichini, A. & M. Debussche. 2000. Seed dispersal and germination patterns in a rare Mediterranean island en- 108 Annals of the Missouri Botanical Garden demic nue crispa Viv. Boraginaceae). Acta Oecol. 21: 303-31 Карр, G. & M . Salum. 1995. Ant fauna, pest damage and yield i in re Pan to the density of weeds in coconut sites in Zanzibar, Tanzania. J. Appl. Entomol. 119: 45— Ready, C. C. & S. B. Vinson. 1995. Seed selection by the red imported fire ant (Hymenoptera: Formicidae) i in the laboratory. ааа Entomol. 24: 1422-1431. Reimer, N., i4 G. Jahn. 1990. Pest ants in the Hawaiian islan ids. . 40-50 in R. K. Vander Meer & A. Dendeno (editors), Applied Myrmecology: A World Perspective. Westview Press, Oxford. Room, P. M. 1975 Diversity and и of the ground foraging ant faunas of forest, grassland and tree crops in Papua New Guinea. Austral. J. Zool. 2 23: 71- 89 . S. C. Smith. 1975. Relative abundance and disinbubon of insect pests, and other components Е the cocoa ecosystem in Papua New Guinea. J. Appl. Ecol. 23 ). Sakata, Н 1994. How an ant SPR 2 prey on or to attend aphids. i Populat. Ecol. 6: 45—51. Samways, M. J. 1990. Ant кы. structure and eco- logical management in citrus and Med iE al fruit or- chards in Southern an Pp. Vi der Meer A. о pa ditors), Applied Myrmecology: A = оа. Westview Press, Oxford. Sanders, . S. Chang, A a & N. Nomura. 1992. Food : acce dades and ыш on in the colony of the bigheaded ant, Pheidole megacephala (Fabr.) (Hymenoptera: Formicidae). Proc. Hawaiian Entomol. Soc. 99-72. Spencer, H. 1941. The small fire ant Wasmannia in citrus groves—A preliminary report. Florida Entomol. 14: 6-14 Stein, M. B., . Thorvilson & J. W. Johnson. 1990. Seasonal ads es in bait preference by red imported fire ant, So aiam invicta (Hymenoptera: ү lae). Florida E A. 73: 117-123. Steyn, J. J. The effect of mixed ant populations on red scale aurantii Mask.) on citrus at Le taba. J. En We . S. Africa 18: 93-105. Suarez, A. V., pem T. J. Case. 2001. Patterns of spread in ds. 'al invasions dominated by long- distance jump T aed from (тве line ants. Proc. Natl. Acad. S ©. : 1095—1100. Sudd, J. Н. 1987. аз ا‎ n 35-365 in A Minks & P. Harrewijn (editors), Aphids: Biology, Natural Enemies and Control, Vol. A. Elsevie Amsterdam Taylor, B. & $, Е. Adedoyin. 1978. Тһе abundance and interspecific relations of common ant species (Hyme- noptera: Formicidae) on cocoa farms in Western Nige- ria. Bull. m Res. 68: i Tennant, L. E. Porter. 1991. Comparison of diets of two fire ant species (Hymenoptera: Formicidae): Solid and liquid eru 1. Entomol. Sci. 26: 450—465 Van Der Goot, P. 1916. Further inve stigations regarding the economic mete of the Gramang-ant. Appl. Entomol. 5: ERA з Meer, R 5. Lofgren & J. A. тет 9. Specificity p the red imported fire ant (Hyme эк Formicidae) phagostimulant response to car- bohydrates. Florida Entomol. 78: 144—154. Vanderwoude, C., L. A. Lobry De Bruyn & P. N. House. 19 2000. Response of an open-forest ant community to invasion by w introduced ant, Pheidole megacephala. Austral. 5: 253-259. Veeresh, G. К. Gubbaiah. 1984. A report on the ‘crazy ant (inoploleps longipes Jerdon) menace in Karnataka. 4: 65-73. J. Soil. Bi ^ jl. vs 1. 4 Vinson, S. B. & L. Cisenbere: 1986. ке UR I iology, and ecology of imported fire ants. Pp. 26 . B. Vinson (editor), Economic Шо em i2 of e ial pe Praeger, New York. Visser, D., M. G. Wright & J. H. Giliomee. 1996. The effect ul ie Argentine ant, d humile (Mayr) (Hymenoptera: Formicidae), on flower-visiting insects ‹ : Protea nitida Mill. (Proteaceae). Afric ‘an Entomol. < 285-287 Ward, P. S. 1987. Distribution of the introduced Argen- tine ant (/ridomyrmex humilis) in natural habitats of the lower Sacramento River Valley x its effects on the indigenous c fauna. Hilgardia 55: 1—16 Way, M. J. 1953. The ш айп, [m ween certain anl species with particular reference to biological control of the coreid, Theraptus sp. Bull. Entomol. Res. 45: 66° 691. —— 63. Mutualism between ants and eres preduci ing Homoptera. Annual Rev. Entomol. 8: 307 344 K. C. Khoo. 1989. Relationships between Hel- qui Mieooromas damage and ants with special refer- ence to sian cocoa smallholdings. J. Pl. Protect. Tropics 6: l- D =% ют 992. Role of ants in pest manage- Annual Bev Entomol. 37: 479— 04. . M. E. mell, M. R. fiae & C. A. Colling- Mood. 1997. оо and dynamics а the Argen- tine ant Linepithema ae kunde e (Mayr) in re- lation to vegetation, soil conditions, topography and native competitor ants in Portugal. Insectes Sociaux 44: 15—133. ment. . M. R. Paiva & M. E. Cammell. 1999, Natural biologic ‘al control of the pine processionary moth Thau- metopoea pityocampa (Den & Schiff) by the ju p ant Linepithema bsp (Mayr) in Portugal. Agric. For- est. Entomol. Aa "Exotic Ants: Biology, Impact, and C bol of йй ш. ed Species. Westview Press, Boulder, Colorado. Willmer, P. G. & G. N. Stone. 1997. How aggressive ant- eis assist seed-set in Acacia flowers. Nature 388 —167. 16 Wyatt, R. 1980. The impact of nectar- -robbing ants on the pollination system of Asclepias curassavica. Bull. Torrey Bot. Club. 187: 24—28. Yano, S. 1994. Flower nectar of an autogamous perennial Rorippa indica as an indirect defense mechanism ^ herbivorous insects. Res. Populat. Ecol. 36: 63- Young G. R., G. A. Bellis, G. R. Brown & E. S. C. Smith. 2001. The crazy ant Anoplolepis ж "d Pe) (Hy- menoptera: Formicidae) in East Arr апа, Austra- i . Entomol. 28: 97— Yu, D. W. a om of иийй Biol. J. Linn. Soc. 72: 529—5 Zettler, J. A., T. Р ш C. R. Allen. 2001. Ant-seed mutualisms: Can red imported fire ants sour the rela- tionship? Biol. Conservation 249-253. Zimmerman, E. С. 1970 Adaptive radiation in Hawaii with special reference to insects. Biotropica 2: 32-38. MOLECULAR SYSTEMATICS John F. Gaskin? AND THE CONTROL OF INVASIVE PLANTS: A CASE STUDY OF TAMARIX (TAMARICACEAE)! ABSTRACT origins of invasive haplotypes and tests for the presence standing invasions through molecular systematic and population-level studies will prove to be powerful tools in many control scenarios. Key words: logeny, tamarisk, Tamarix, saltceda biodiversity, biological control, conservation, invasion, molecular systematics, population structure, phy- ar. The invasion of habitats by non-native organisms is considered the second largest threat to biodiver- sity worldwide behind habitat destruction (Wilson, 1997). In the United States exotic plants now rep- resent 17.3% of the flora (Kartesz & Meacham, 1999), and approximately 400 of the 972 plants and animals listed by the Endangered Species Act are at risk primarily due to competition with and predation by non-native species (Stein & Flack, 1996). For these reasons, the control of invasives is becoming an integral part of ecosystem steward- ip. Methods of controlling invasive plants include manual removal, fire, herbicides, biological control, and legislation of import and sale. Effective control of invasive plants often requires explicit character- ization of the invasion at the family, species, and/ or population levels. Several species of the genus Tamarix L. (com- mon name: saltcedar or tamarisk, family Tamari- caceae) are, as a group, considered one of the worst plant invasions in the southwestern U.S. (TNC, . This invasion is the subject of localized manual, chemical, and legislative control efforts and a large-scale biological control project con- ducted by the United States Department of Agri- culture. Additional legislative control may be re- quired, as cultivars of Tamarix are still available from numerous horticultural suppliers. The effec- tive implementation of biological control projects of Tamarix (e.g., DeLoach et al., has been in- fluenced by phylogenetic concerns at the following levels: (1) FAMILY LEVEL Phylogenetic relationships of the invasive plant's family are important when biological control is pro- posed. Control agents must be tested for their risk of host-switching by confronting the control agent with plant species from closely related plant fami- lies. In the past, Tamaricaceae usually were placed in the plant order Violales of the Dilleniidae (e.g., Cronquist, 1981), but recent molecular sequence data analyses have altered the traditional ordinal placement of many plant families, and Tamarica- ceae are now included in the Caryophyllales (APG, 1998). These changes will alter the plant taxa to be ! The author thanks C. J. DeLoach, V. Ivlev, I. Mitayaev, J. Schulte, and J. Tracy for sending plant material included in the population-level study of Tamarix. Sarah Parson s and anonymous reviewers supplied helpful comments for this manuscript. This research was supported by USDA Cooperative State Research, Education, and Extension Service grant 32000-0083 6 to B. Schaal and J. Gaskin, National Geographic Society Committee for Research and Exploration grant F #6663-99 to J. Gaskin, the Mellon Foundation support of Missouri Botanical Garden graduate students, and an EPA Science To Achieve Results (STAR) graduate fellowship to J. Gaskin. 2 Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. Present address: USDA-ARS- NPARL, P.O. Box 463, Sidney, Montana 59270, U.S.A. jgaskin@sidney.ars.usda.gov. ANN. Missouni Bor. GARD. 90: 109-118. 2003. 110 Annals of the Missouri Botanical Garden Table 1. Putative U.S. Tamarix invasives complied from McClintock (1951), Baum (1967), and Crins (1989). Putative U.S. invasive Taxonomic and morphological notes T. africana Poir. T. aralensis Bunge rarely « T. aphylla (L.) H. Karst. T. canariensis Willd. T. chinensis Lour. T. gallica L. T. juniperina Bunge synonym of T. chin T. parviflora D.C. T. pentandra Pall. synony T. ramosissima Ledeb. T. tetrandra Pall. T. tetragyna Ehrenb. naturalized in eastern U.S morphologically similar to T. gallica (Crins, 1989 1 of T. ramosissima (Baum, 1978 ie AR cally similar to 7: canariensis and T. gallica in n aestival floral form (Baum, 1978) cultivated, not extensively naturalized (Baum, 196 morphologically dissimilar to all other U.S. Tamarix < morphologically similar to T. ramosissima (Crins, 1989) morphologically similar to 7! canariensis (Crins, 1989) ensis (Baum, 1978) расе n ds to all other invasive U.S. Tamarix — Sd lode similar to 7: chinensis (Crins, 1989) U.S. invasive specimens with this name considered to be T parviflora (Baum, 1967) .. not yet invasive (Crins, 1989) tested. in the risk analysis of biological control agent host-switches. (2) SPECIES LEVEL The Tamarix invasion consists of many species, some of which are morphologically very similar. The specific identities of taxa involved in the in- vasion are controversial (Crins, 1989), in part, be- cause most Tamarix species cannot be distin- guished without flowers or fruit present. Precise species identity is needed to determine the geo- graphic origin of the invasive species and its co- evolved biological control agents. Tamarix is one of the more taxonomically challenging genera among the angiosperms (Baum, 1978), and intermediate states exist for some morphological characters used in species identification. These character states can oflen vary on a single individual from season to 1949), play a role in the taxonomic confusion (Rusanov, 1949; Wilken, 1993). Improper species identifica- tion could lead to searches for biological control season (Rusanov, and hybridization may agents perhaps outside the native range of the in- vasive plant. The failure of morphological data to elucidate the identities of invasive Tamarix species necessitates the use of molecular data as an addi- tional source of taxonomic information. (3) POPULATION LEVEL Population-level investigation of any of the in- vasive Tamarix may be necessary if the biologic al control agents are extremely host-specific, and if the invasive plant species has widespread origins. Initial biological control tests show that imported insects have differential effectiveness on what pu- tatively appears to be a single species of Tamarix (T. ramosissima Ledeb.) collected from different re- gions of the U.S. (DeLoach & Tracy, 1997), raising the issue of infraspecific geographical variation. = any species of Tamarix are widespread in Eurasia 1978), and it is unlikely that much of the genetic diversity of any one species was imported (Baum, to the United States. Historical records do not re- veal precise origins or genetic information concern- ing the introductions (Horton, 1964). The control agents being tested (e.g.. saltcedar leaf beetle. Diorhabda elongata) may not have evolved with the invasive, and thus might result in ineffective or sub-optimal control. For these reasons, it would be useful to know how many genotypes are represented in the U.S. invasion, and to what degree we can pinpoint their Eurasian origins. Additionally, Tamarix is still being horticultur- ally distributed in the United States. Policy makers need to determine if the genotypes currently being promoted predominate in the invasive populations. Based on their similar morphology. invasive Ta- marix is often indistinguishable from cultivars. Mo- lecular evidence of contemporary cultivars contrib- uting to the Tamarix invasion could greatly influence future policy decisions regarding the sale and distribution of these plants. BACKGROUND Tamarix is an Old World genus of approximately 24 species (Baum, 1978). Eight to twelve of these (Table 1) were imported to the United States from southern. Europe or Asia in the 1800s to be used for shade and erosion control (Baum, 1967), and an subset has aggressive overtaken more than 1.000.000 riparian acres (Brotherson & Field, 1987). This infestation is expanding by 40,000 acres per year (DiTomaso, 1998), eroding the bio- diversity of many western U.S. natural areas, in- cluding major river systems and national parks. Tamarix species initially invade by germinating Volume 90, Number 1 2 Gaskin Case Study of Tamarix during wet periods or in riparian areas. Once es- tablished, they can tolerate drought by utilizing deep groundwater sources. They also exude excess salt from salinized water sources from glands in their scale-like leaves (Neill, 1985), which are sea- sonally dropped, forming a thick saline duff on the soil surface that inhibits the germination of other plants. In the U.S., Tamarix species are avoided by most avian frugivores and insectivores (Brotherson & Field, 1987), and only two mammal species (the desert wood rat and desert cottontail) are known to feed upon them, with minimal damage to the plants (DiTomaso, 1998). Tamarix invasions lower biodi- versity levels by displacing typical Southwestern ri- parian vegetation such as cottonwood and willow (Hughes, 1993), as well as the insects, birds, and mammals that these native trees support (Neill, 5). Their profuse growth alters stream and river dynamics by narrowing channel width (Robinson, 965), and invasions can extend over 1 km on each side of a river (e.g., Gila River, southwestern Ari- zona, and Colorado River south of Blythe, Califor- nia, U.S.; askin, pers. obs.). Dense stands of Tamarix, with their high rates of transpiration, can substantially lower the water table, have caused perennial springs and creeks to dry up. in some cases threatening regionally rare or federally listed species such as the desert pupfish and the desert slender salamander (Kerpez & Smith, 1987). Tamarix invasions have proven difficult to con- trol. These plants cannot be killed easily by fire. by cutting at ground level, or by herbicide applied to the foliage alone. Effective removal is both ex- pensive and potentially damaging to the habitat, re- quiring mechanical uprooting, or cutting at ground level with application of a systemic herbicide to the stump. Repeated treatments are often necessary (Neill, 1985). Control is possible on a small scale, but land managers are often forced to live with and large invasions due to prohibitive control costs (Stein & Flack, 1996). For these reasons biological agents were proposed as an alternative means of control. Well-researched biological control projects often come under heavy public scrutiny due to the po- tentially dire effects of control agent host-switches (Thomas & Willis, 1998). Therefore, biological con- trol researchers. must. unambiguously know the identity of the invasive Tamarix and its relation- ships to native species. Improper taxonomic iden- tification may lead to searches for control agents outside the native range of the invasive species and thus efforts or less-effective biocontrol agents. Improper identification of the invasive could also lead to the collection of biological con- wasted trol agents that have historic ties to sympatric con- gener plant species or to genotypes with a different phenology or developmental timing, again yielding ineffective biological control. Considering that the average ge control research program spans many years at a cost of hundreds of thousands to millions of dollars (Gillot, 1995), it is logical and economical to predicate a biological control project with precise taxonomic knowledge of the invasive plant. Knowing the genotype of an invasive plant is es- pecially important when choosing a fungal, bacte- rial, or viral control agent involved in a gene-for- gene resistance/virulence interaction (Kerr, 1987). Even insects are often species-specific, and in some cases, host-specificity can reach to the level of the plant genotype. An example is the differential her- bivory of the Hessian fly (Mayetiola destructor) on different genotypes of wheat (Triticum aestivum L. (Schoonhoven et al., 1998). Also, differential her- bivory on plant populations has been detected in willow trees (Salix) under natural conditions (Rank, 991). The saltcedar leaf beetle (Diorhabda elongata) from western China is already being investigated as a potential Tamarix control agent in quarantined and field releases (C. J. DeLoach, pers. comm. ). In no-choice tests, newly hatched D. elongata larvae were placed in vials, each with leaf material from a different plant specimen. The plants were col- lected from different areas of the United States and grown in common garden plots. Using morphology, — all specimens were determined to be the same spe- cies (T. ramosissima). os on the feeding and life span of the insects were recorded, and sur- vival of the insects to ie co ipiis on н ет plant specimens varied from 3 % (DeLoach & Tracy, 1997). The reduced | on several of the T. ramosissima plants may have in part been caused by less than optimal physice condition of some of the plants (DeLoach & Tracy, 1997), but genotypic differences in the plants abe may have influenced the results. The search for Tamarix biological control agents continues, as the Agricultural Research Service of the United States Department of Agriculture does not expect that the current control agents will achieve satisfactory control of saltcedar in all areas, and perhaps as many as 8 to 12 additional insects as specific herbivores will be required (DeLoach & Tracy, 1997). This is based on biological control of other invasive plants, such as cacti, lantana, and leafy spurge, which have required up to 15 or more insect. species introductions (DeLoach & Tracy, 1997) for effective control. c to 112 Annals of the Missouri Botanical Garden CULTIVARS SPECIES LEVEL Tamarix ramosissima is commonly sold today as an ornamental plant. Cultivars of 7! ramosissima include mer Glow’. The most common of these is T. ramo- ‘Pink Cascade’, ‘Rosea’, ‘Rubra’, and ‘Sum- sissima ‘Pink Cascade’, known for its dense, dark pink plumes of flowers (due to mostly compound, not simple, inflorescence racemes) and finely tex- tured bluish foliage. These cultivars and invasive populations are almost identical in floral and veg- etative structures, and may only differ in the inten- sity of flower color, density of inflorescences, and foliage color. Invasive 7! ramosissima is highly var- iable in flower color within some populations, rang- ing from deep red to white. On a single invasive plant, both simple and compound racemes can be found, making the density of inflorescences also highly variable. Invasive foliage color can vary within populations from dark green to the blue-gray found in the 7! ramosissima ‘Pink Cascade’ cultivar — (J. Gaskin, pers. obs. As a weedy species, the Tamarix cultivars are easy to grow and tolerant of poor soils. They are and from internet sales (e.g, Gertens Online Shop, available through many nurseries, catalogs, www.gertens.com). Tamarix ramosissima is not le- gally available in Colorado, Nevada, Washington, and Wyoming, where it is listed as a noxious weed (USDA, 2002). MOLECULAR SYSTEMATICS AND THE CONTROL OF TAMARIX FAMILY LEVEL Biological control agents are assessed for their risk of host-switching by placing them on U.S. na- tive plants that are closely related to Tamarix. Ta- maricaceae, along with the sister family Frankeni- aceae, had historically been placed in the order Violales (Cronquist, 1981). Therefore, U.S.D.A. re- searchers tested if the Tamarix control agents would feed and reproduce on plants from other fam- ilies in this order, such as Frankeniaceae and Fou- quieriaceae (DeLoach & Tracy, 1997). Recent DNA sequence data analyses strongly suggest that Ta- maricaceae and Frankeniaceae actually belong to- gether in the order Caryophyllales (APG, 1998), closely aligned with families such as Droseraceae 1998). This phy- logenetic rearrangement requires a substantially — and Polygonaceae (Lledó et al., different set of test plants in the greenhouse, which may provide significantly different assessments of the risk of host-switching. A recent study used DNA sequence data to de- termine how many invasive Tamarix species were naturalized in the United States and to see if the molecular data were congruent with the morpholog- ical distinctions currently used to segregate taxa (Gaskin & Schaal, in press). The taxonomy and morphology of the 12 putative U.S. naturalized Ta- marix species were investigated (Table 1). Three of the species names had been designated as syno- nyms, and two were not yet considered invasive, leaving seven putative invasive taxa. А molecular phylogenetic analysis of these and other selected species in the genus was performed from samples collected in the western U.S., Argentina, and wild native populations across Eurasia and southern Af- rica (voucher information is listed in Appendix 1). Phylogenies from both nuclear ribosomal ITS and chloroplast trnS-trnG intergenic spacer se- quence data were constructed and compared. Por- tions of the final phylogenies presented in Figure 1 illustrate incongruence with earlier taxonomic un- derstanding of the genus. For example, note that T. chinensis Lour. and T. ramosissima, thought to be- long in different sections of the genus (sects. Oli- gadenia and Tamarix, respectively), have identical placement on both phylogenies. Additionally, the most recent sectional classification of the genus (Baum, 1978) was not significantly similar to either the chloroplast or nuclear topologies found in Gas- kin and Schaal (in press). For many samples there was incongruence be- tween the chloroplast and nuclear evolutionary his- tories. For example, in the nuclear phylogeny of Figure 1, T ramosissima specimen Schulte 1 was in a clade with all of the other 7! ramosissima, but in the chloroplast phylogeny it appeared in a clade with T. canariensis Willd. (Gaskin 3049) and T. gal- lica L. (Gaskin 3039). Similarly, a T. canariensis specimen (Kirk 2) was in a chloroplast clade with another T. canariensis (Gaskin 3020), but in the nu- clear phylogeny it was found far from specimen Gaskin 3020, as the sister to the 7! ramosissima clade. These incongruences of chloroplast and nu- clear evolutionary histories, which were significant based on the Templeton test (Templeton, 1983), supported a hypothesis of hybridization (Whitte- more & Schaal, 1991; Soltis & Kuzoff, 1995). The study concluded that morphology within Ta- marix is often misleading as a means of identifying specimens. Also, though not all putative invasive species could be distinguished with molecular data, there was enough phylogenetic resolution to rec- ognize four invasive Tamarix entities in the U.S.: Volume 90, Number 1 Gaskin 113 2003 Case Study of Tamarix Species Country Voucher v + v T. chinensis US T. chinensis US J.L.Tracy 4 oo T. chinensis CN T. chinensis CN DeLoach 25 ges N-Chi 75 T. ramosissima US T.ramosissima US Gaskin 103 kZ=Kazakstan T. ramosissima KZ T. ramosissima KZ DeLoach s.n. ра! 29 T. ramosissima AR T. ramosissima AR Schulte 1 78 TN- Tunisia States 66 T. gallica SP T. canariensis TN Kirk2 f Ameri 199 T. canariensis FR T. gallica SP Gaskin 3039 T. gallica FR T. canariensis FR Gaskin 3049 a T. canariensis US T. gallica FR R. Sobhian 13 T. canariensis US T. canariensis US Gaskin 36 99 T. canariensis US T. canariensis US Gaskin 34 T. gallica US T. canariensis US DeLoach 00-01 62 T. canariensis US T. gallica US DeLoach 00-15 T. canariensis TN T. canariensis US DeLoach3 95 T. canariensis SP T. canariensis SP Gaskin 3020 sp. CN Myricaria sp. CN Feng 10 trnG-trnsS aioe apace Figure 1. of 0.95 and. a R. С. of ( Chloroplast and nuclear marker phylogenies. On the left о а marker (trnS-trnG intergenic iine er), 2 .95. On the s the strict consensus o 2 179 мара in length with a C.I. is th ITS 1-2 e single most mo tree for the 18 steps in length, with a C.I. a most Vu AE rm trees for the nuclear sequence ep (ITS1— ).88. Numbers below lines are bootstrap va d an R.C. of sam specimens were used in each analysis, and are connected by lines in between the two а о from Gaskin and Schaal (in press). (1) T. aphylla (L.) Н. Karst, (2) T. parviflora DC... (3) T. canariensis/T. gallica, and (4) T. chinensis/T. ramosissima. Additionally, there was evidence of introgression between T. ramosissima, T. canarien- sis, and T. gallica, which is a likely source of con- fusion in the characterization of some Tamarix in- vasions (Gaskin & Schaal, in press). POPULATION LEVEL To examine the Eurasian origins and relation- ships of T. chinensis and T. ramosissima invasive genotypes, and to investigate the presence of cul- tivated haplotypes in the invasion, the highly var- iable 1001 bp chloroplast trnS-trnG intergenic spacer is analyzed using the primers of Hamilton (1999). A gene tree, which infers genealogical re- lationships of their relationships (see Fig. A total of 59 cultivated, invasive, and native T. ramosissima or T. chinensis specimens was colle ed, with 33 samples from the New World and 26 from the Old World. The identities of most speci- (1978) mor- phological descriptions and keys. Voucher infor- mens were determined using Baum's mation is listed in Appendix 1 In the chloroplast sequence aligned data set, 93 (9.3%) of the sites are variable. There are 12 (1.2%) single bp changes, three single base inser- tion/deletions, one 2-bp indel, and three prominent A sequence haplot is constructed to represent the populations and es (alleles), ct- indels that vary from 8 to 55 bp in length. All in- dels are treated as a single event (a fifth base). A most parsimonious gene tree (or minimum spanning network) of 22 steps was assembled by hand, rep- resenting the fewest mutations that explain the re- lationships of the specimens (Fig. 2). The molecular analysis presents population-level information that is unobtainable using morphology alone. For example, the T. ramosissima species is represented by a total of seven haplotypes, marked through С, on the gene tree (Fig. 2). The speci- mens and their origins are also presented in the boxes. The lines separating the gene tree boxes rep- resent single point mutations or indel events. The small circles represent inferred intermediate hap- lotypes that may be extinct, may not have been collected during sampling, or may not have ever existed if mutations did not accumulate in single steps. Interesting results include the following: Haplotype A is ve (1) Of the seven haplotypes found, four are rep- resented in the western U.S. Tamarix invasion. common, representing 46 (789€) of the specimens sampled. The native hap- lotype A specimens were collected in the Republic of Georgia, Iran, Turkmenistan, Kazakstan, China, and South Korea. The naturalized U.S. specimens were collected from California, east to Texas, north to Kansas, and west to Washington. The widespread nature of this haplotype will not facilitate pinpoint- ing its invasive origins in Eurasia. Finer resolution 114 Annals of the Missouri Botanical Garden So SÍA eae c 9 8 IT. с ramosissima B T. ramosissima Е o 3 | 46 spec A CT = Via evi Miyae 19 121 23 кышны in Asia 563 220 T. ramosis U.S., CA: Gaskin 69 23 widespread in U.S. Kazakstan: ер s.n. T 356 C deletion 809] 2 base insertion 510 T. ramosissima U.S., AZ: Gaskin 62 627 T. ramosissima ., CA: Gaskin 70, 73 655 U.S., CO: Gaskin 100* , KS: Gaskin 10 U.S., MO: Gaskin 1253* U.S., OR: Gaskin 1209* insertion 2 base deletion Insertion *Cultivated specimens found in gardens. All others found in the wild. deletion A F T. ramosissima Italy: Gaskin 3065 insertion E T. ramo ine + 2 Figure 2. Single most parsimonious gene gene 'alogy of the c hloroplast sequence marker trnS—irnG intergenic space for T ramosissima and morphologic 'ally similar species. The gene tree is 22 steps (mutations) in length. The чнае (allele) designation is in each box. along with о on the number and distribution of specimens with that haplotype. The lines separating the haplotype boxes are e point mutations or insertion/deletion events. The small circles represent intermediate haplotypes not recovere ve i cai analysis. The gene tree is interpreted in the following manner: Haplotype A differs from haplotype B ii two mutations. Ones of these is a single nucleotide mutation at site ee along the trnS-trnG intergenic spacer. where haplotype A has an adenine (A) and haplotype B has a cytosine . The other difference is at site #420, where ios A has a thymine (T), and haplotype B has lost this thymine a deletion event (-). Volume 90, Number 1 2003 Gaskin Case Study of Tamarix 115 markers are needed to distinguish if there is un- revealed population structure or if the haplotype A بس = plants are genetically similar across Eurasia. these plants are genetically similar across the na- tive range, collection of insects from any area of Eurasia will be equally likely to find control agents that have evolved with this haplotype. (2) Haplotype C is rarer than A, representing only two of the specimens (Fig. 2), and was found once in southern California and once in Kazakstan. This haplotype is only one mutation different from the common A haplotype, but that mutation is a prominent 9 bp indel event that was not found in any other samples. This presents evidence that at least a small S of the invasion may have its or- igins in Kazakst (3) Hp D was found once, in Arizona. — The plant containing this haplotype (Gaskin 62 morphologically resembled 7: ramosissima. In a dif- ferent study, this haplotype was found to be com- mon in another species, T. parviflora (Gaskin & Schaal, in press). Tamarix parviflora is an invasive species with tetramerous floral structure, morpho- logically very distinct from the pentamerous floral structure of T. ramosissima and T. chinensis. This incongruence between morphology and haplotype may be due to hybridization, as was found in the genus-wide study (Gaskin & Schaal, in press). (4) Haplotype E was found twice in Argentina but never in the U.S. haplotype is genetically quite distinct from the invasions by Tamarix, This common A haplotype, differing by 11 mutations, including two notable 8 and 55 bp indel events. This genotype has not been found in Eurasia, in- dicating that further sampling of native Tamarix populations is needed. (5) All cultivated U.S. specimens of Tamarix contain haplotype G. This haplotype was not recov- ered in Eurasian sampling. Haplotype G differs from the common haplotype A by at least seven mutation events. The haplotype G, as representative of cultivar introgression, was found once, as an in- vasive, near the Salton Sea in California (Gaskin 70). Even though the presence of this genotype is not common in the invasion, its ability to invade is now confirmed. Any presence of cultivar haplotypes in invasions should serve as a strong forewarning in future policy decisions regarding the horticul- tural use of invasive taxa. The preceding chloroplast sequence marker data allow us to begin to delve into the genetic structure This preliminary analysis is of a small sample size, and of the T. ramosissima/T. chinensis invasion. a more in-depth population analysis is in prepara- tion, using highly variable nuclear DNA sequence markers such as phosphoenolpyruvate carboxylase introns. I plan to continue sequencing selected Ta- marix that exhibit resistance to biological control agents. If they are determined to be genotypically distinct from the susceptible Tamarix, their Eur- asian origins will be provided to the biological con- trol exploration project. Knowing the number of haplotypes that comprise a plant invasion, their or- igins, and the ability of cultivars to contribute to the invasion are powerful tools to document and control problematic exotic plant species. CONCLUSION Molecular analyses will have an increasing role in invasive plant control efforts. At the family level they will enable more accurate risk assessments of biological control host-switching. At the species level, molecular systematics will help elucidate in- vasive species identities and any morphologically cryptic hybridization events. At the population lev- el. molecular systematics will allow the unprece- dented characterization of invasive taxa as geno- types, allowing precise matching of biological control agents with their targets, and elucidating links between cultivars and invasions of plants. These advances in understanding plant invasions will enhance control efforts and contribute to the protection of native biodiversity. Literature Cited APG (Angiosperm Phylogeny Group). 1998. An ordinal classification for the families of DUUM plants. Ann. Missouri Bot. Gard. 85: 531—553 Baum, B. 1967. Introduced and studios d tamarisks in the United States and Canada. Baileya 15: 19-25. 1978. The Genus Tamarix. Israel Academy of Sciences and Humanities, Jerusalem Brotherson, J. D. & D. Field. 1987. Tamarix: Impacts of a suce uk ved Rangelands 3: 110-112. Crins, W. J. . The Tamaricaceae of ha southeastern United es Й Arnold Arbor. 70: 403—425. са А. An Integrated System of Classifica- x of суеп Plants. Colombia Univ. Press, New Delo oach, . Tracy. 1997. The Effects of Bio- logical я al Salteedar ры ramosissima) on Endangered Species. Biological Ass sessment raft. USDA Agricultural Research Service. ; D. Smith. å Proceedings of the X International Symposium on Bio- logical Control of Weeds. Montana ме University, Bozeman. DiTomaso. J. M. saltcedar (Tamarix spp.) 1 States. —-— Technol. 129) : Gaskin. J. К. & B. A. Schaal. aim pod Шаг ا‎ netic investigation of U.S. invasive Tamarix. Syst. (in press). 1998. pem biology, and ecology of eec United 116 Annals of the Missouri Botanical Garden Gillot, C. 1995. Entomology, 2nd ed. Plenum Press, New ork. Hamilton, M. 1999, Four primer pairs for the amplifica- tion of chloroplast intergenic regions with intraspecific variation. Molec. Ecol. 8: 521—523. 5. 1964. Notes on the Introduction of Decid- uous Tamarix. U.S. Forest Service, Fort Collins, Colo- rado. Hughes, L. E. 1993. ' lands 15: oF 55 Kartesz, J. T. & Mecham. 1999. Synthesis of the North American vers. 1.0. North Carolina Botan- ical Garden, Chapel H Hill. Kerpez, T. A. & N. S. Smith. 1987. Saltcedar Control for Wildlife Habitat in ement in the Southwestern United States. U.S. Fish and Wildlife Service, Resource d pce 169. Washington, D.C Kerr, A The impact of molec alae gene zy s on plant pathology. Annual Rev. Phytopathol. 25: 87-110. Lledó, M., 'espo, K. Cameron, M ma & M. Chase. 1998. Systematics of Plumbaginaceae based upon cla- distic analysis of rbcL sequence data. Syst. Bot. 23: 21— — — ^ © ~ © = =) E — T The Devil's Own—Tamarisk. Range- McClintock E. 1951. Studies in California Отата! plants. 3. The tamarisks. J. Calif. Hort. Soc. 6-83 Neill. W. v 1985. Tamarisk. Fremontia 12(4): 22— = N 23 Rambaut, A. 1996. Se-Al Sequence Alignment Editor. Oxforc K. Rank, N. E. 1991. Effects of plant chemical variation on a specialist Pipe Willow leaf beetles in the eastern Sierra Nevada. In: . Hall, V. Doyle-Johnes & B. Widawski (editors), nd History of Eastern Califor- nia and High-altitude Research 3: 161—181. University of California, White Mountain Research Station. Hobinson, T. W. 1965. Introduction, Spread, and Aerial Extent of Saltcedar (Tamarix) in the Ree States. U.S. Geological Survey professional paper (Stud- les in eva potranspiration), U.S. Se Printing Office, Washington Rusanov, F. 949. ned de Tamariksi. Tash- kent. [' [апайы of Central Asia. Schoonhoven, L., y & J. van — 1998. Insect- Plant T C haan and Hall, New York. Soltis, D. E. & R. K. Kuzoff. 1995. а between nuclear p chloroplast ssa in E Huechera group (Saxifragaceae). Evolution 49: 727-742. Stein, B. А. & S. R. Flack. 1996. Americas Least за = Alien species invasions si U.S. ecosystems. The N re Conservancy, Nov.-Dec.: 17-23. Templeton. A. 1983. К inference from restric- tion кше lease cleavage site maps with particular reference to the evolution of humans and the apes. Evo- lution 37: ua —244. Thomas, М. & А. J. Villis. bit 0 SUR but necessary? Trends rac Evol. 325-3 Whittemore, A. T. & . Schaal. [Eod ым ЕТА gene flow in саа К Proc. Natl. Acad. U.S.A. 88: 2540-2544. Wilken, D. H. 1993. Tamaricaceae. 1080 i C. Hickman (editor), The Jepson Manual. U niv. California Press, Berkele ey. Wilson, E. O. 1997. Strangers in Paradise. Island Press, Washington, D.C Internet Resources TNC (The Nature Conservancy). 2002. (http://tncweeds. ucdavis.edu/worst.html). USDA (United States Department of Agriculture). 2002. (hti // hi usda / lk /otat ht 15 ` T F o I ~ Gaskin 117 Volume 90, Number 1 2003 Case Study of Tamarix 96206T.IV UtOcI (OW) [POGI Чу) VA “SN ‘MN “әрә vunmssisouma VUVUDI V 986206TV.1V T6 ON) og uns) WN “SNA MN дәрә] nunssisouma хишт), V 96206T.IV L8 OW) Z2 uno?) VO “S'NA CAN qope'] punssisouma xunum][ V Л АД 08 OW) 88 urjsv?) VO “SN : MN дәрә punssisourp4 VUVUVI V 96206TAV bL OW) ©8 urgsv?) VO “SNA MN qope'] PUNSSISOWDAI ximung V 96206TV.1V GL OW) Ip 13505) XL “S'N ‘AN qope'] punssisouma халш V 86L06TAV 6S OW) 66 414509 OD “S'NA ‘MN qape' nunssisouma ximung V TLLVSTAV 86L06TAV SS (OW) £01 иту) OD “S'N MN черет тшй апош ү 86L06bAV £70 (4) €£-00 2072d XL “S'N MN әрә” punssisouma XUDUD] V 86L06V1V cro (x) &F-00 u»voToq XL “S'N MN әрә Dunssisowps хыюшт] ү 96L06VAV IO (x) [7-00 Popa XL “SN AN 'qaper] DUNSSISOWDA XUDUD], y 86206714V 8є`0 (x) 8£-00 YP PaA WN “S'A MN "qape PUNSSISOWDA XUDUDI V 86L06T.V LTO (x) ZZ-00 Po PA AN “S'NA MAN "әрә" PUNSSISOWUDA ximum[ V 86L06TAV 9c'0 (x) 92-00 YPO PA WN “SA ‘MN әрә” punssisoumna4 XUDUDI, V 86L06TAV Cc 0 (x) ©2-00 420012] VD “S'NA ‘MN әрә” punssisouma4 xipui[ V 986206T.IV vcO (x) F&-00 Yoo PA VD “S'N MAN "әрә" DUNSSISOWDI XUDUDI, V 86206T4V ETO (x) 62-00 Popa WN "SAN “әрә” puassisouma. XUDUDI, V 986206T.1V IZO (x) IZ-00 Popa SN “SA MAN "qaper] DUNSSISOWDA xiu V 86L06PAV се} (OIN) "ws aaja А UP]SXPZEN MO "әрә" punssisouma xiu V 86L06PAV ceb (ON) "ws aoap A UPISYPZEY MO “әрә” punissisouma XUDUD], V 96106F.1V 6lt (OIN) us aaja 4 UP]SXPZPN MO "qoperq punssisouna хри), V 86L06PAV ec (x) OZ aovAnpy ^q 7I UPISYPZPY MO "әрә" punissisouma XUDUDI, V 86L06PAV се (OIN) 80€ 1380) в18109) MO "әрә" punssisouma ХШ?) ү 86L06TAV OLE (OW) sos шух e1831099 MO OPE p UulsstsoumL ршн, V 86L06PAV cóc (OIN) 642 1140 8181095) MO "әрә" punssisouma XUDUDI, V 86L06PAV 1011 (OIN) ZOLI 105) uejsiuaumxng MO "qape punissisouma "JO XUDUDI, V 86L06PAV IPI (x) ‘u's yopo pq PUIYD (MO “әрә” punssisouma "JO хишт), V 92LY8V AV 86L06bAV сс (x) Р Оор 7] f XL “SN MN MOT sisuaunjo хашр, V 86L06TAV TUO (x) FI-00 42072 XL SA MN "no sisuauno XUDUDI, V 86206t1V coc (OW) ZOZ Чтв) вәоу "S MO uno] sisuaunp хош V 96206VV 1106 (x) 67-00 VASA вищо ‘MO апор туши VOLUI V 86L06TAV OTT (4) "wis цэротә( вищо (MO "uno SISUIUIYI XUDUD], V OLLVSVAV 986L06V1V £c (x) SG uovoTeaq eun) ‘MO ano] sisuaunpo XUDUDI V 86L06V AV его (x) £I-00 ovoTo(q BUI) ‘MO "noT SISUIUIYI ximum] V 86L06PAV PIL (x) ws уәротә( вищо ‘MO "qapeT punssisouma "JO XUDUDI, V сї SLI DUIS ug # чәшәәйз үү # oyoo ШЕ, saroadg od4jo[deu UOISS922E8 Xuequaz) UOISS292D8 Xyuequa:) “yudy-S ил] "921 Yorwasay 1998/4 pu? [ros pue[sse1) Suv-vasrn ‘POM M3N = AN ‘PPA PIO = = (4) ‘Ssdequinu uorssaooe yuegua<) Surpuodsa1102 pue ‘3urouanbas WNC ut pasn sre[duroxo 10у MO WSN ‘sexag, ‘a[dua], S19uonoA `1 xipuaddy Annals of the 118 | Garden issouri Botanica M Ор! ВАУ 91 (x) 01 2uo4 umf биру euro “MO JUS sapio1n2ado]p Did INIA PY ООВҮВӯЯУ Осо (OW) 0c0£ urs) ureds :MO PITA POPHATHRIDA Хнюшәр 962 V8TAV 9/1 (OW) visrun[-c улу етти, mo 'PIIUA $5usriDupo xumumnr 8L2TS8TAV St (м) 10-00 yoq XL SA MN “PIM 55иәт1рирә хиошт I82Y8T AV SUO (x) €1-00 ywq XL SA MN "1 PNP? хошо GSLV8VAV Ре (ж) £ uopoToq VI SA MN РИИ: SISUOHDUDD мыр, LO8P8PAV S9 (OW) FE urjsp:) VT SA MN PIRA SSHOUDUD?D иша СОВҮВӯЯ У 89 (OW) 9£ шту) VT SA AN PIU S75u24DUuDO ХШ, SZLVSVAV се (x) E7 uvnjgog M оне 4 MO "| Pornos хош, 909T9T.I V СОЄ (OW) 6f0£ urs) зов MO "PIEM SISUIUDUDI een LO8V8PAV 6£0t (OW) 6£0£ чир) ureds ‘MO `1 20403 мети C8LOOPAV а (OW) ESZI urfso?) OW SA ‘MN BPS | зориг VURUR] 9 c9206TAV 60cl (OW) 60c1 414509 HO SA MN PIT Peer! VURUR 9 КЛ ЖА 06 (OIN) £4 шуға) VO SA MN “eRe Ss e EN 9 ZBLO06PAV 68 (OW) OZ шуо) V) “SA MAN "qoper] DUISSISOWUDAI харш с) c9206T. AV 26 (OIN) 001 14%) 02 “SA MAN "әрә" punssisouma XUDUDI, 9 C8LOOVAV eS (OW) SOL urfsp?) SA SA MN AOPE r: 9 LE806TAV COVE (OW) €90£ usm) (е ‘MO qepeq PUSS ISOLDE хир t 689! 06®АУ pet (ON) Z 2imuyos eunuoziy MO "qaper] punssisouma. XUDUDI, d [91 T9, AV 691L06T.1V Ott (OIN) 1 AMY puruoz1y MO "qaper] puimssisouma. xiupuin q 8660ESAV O11 (OW) 29 unjsp?) ZV "SY MN Чәр ELS MED a 96206V.1 V V8 (OW) 69 unrgsp?) VO SN AN "qope] punssisouma хошо, 9 96206t4V I£ (x) 61 Any "(1 I UP]SXEZEN MO "qoper punssisouma хош? 7) QvLVOT AV 92 106v. AV [CF (м) u's 420012( UP]SNPZEN MO "әрә" punssisouma xiunum[ q 86L00TAV ЗҮР (a) VE100 voa] XL SN MN AOPE ошошл SHOUD ۷ 96206YT.1V 91 (OW) 9111 4809 uejsiuourgng MO “әрә” punssisouma "Jo XUDUDI V 86206741V 2801 (ON) 2801 иту) uejsiuaurpin] MO “әрә” PUNSSISOWUDA "Jo xiunum V 86L06PAV 196 (OW) £96 10) ШАГ MO “әрә” punssisouma "JO VIIDUD], V 86206VAV 696 (OW) 296 urjspz) uea] ‘MO Ape] PEE OME уо HDD] V 86L06PAV cre (OW) Сс umso.) 018109) MO ‘әрә puissisouma "Jo XUDUDI V 86L06PAV VLOG (x) 61-00 voor AM SA MN чәр J PUNSSISOUTI XUDUDI V 86L06TAV £10G (x) 8Р-00 ywq OD "SN MN "SENSU | ens XUDUDI V 86L060AV CLOG (x) 91-00 y0q XL “SA MN Чәрәт Rees оин ү 86L06PAV 600c (a) I€-00 ovoToq XL SA MN "qaper DUISSISOUDL VUDUDI, ۷ 86206VAV Мега! (OW) ISGI чухт) MO SA MN "QOP J 2ш155:50ш01 Xumuny V Z-I SLI «уш ua] # чэшоәйѕ WNC] # uonoo[o? Urdu) satoeds ad&jojdeu UOISSQ)O? NuPq[Uua*) UOISS2298 Nuequo«) 2yu4r-S ua] 'penuguo? сү xipuaddy WEEDS IN PARADISE: THOUGHTS ON THE INVASIBILITY OF TROPICAL ISLANDS! Julie S. Denslow? ABSTRACT Tropical island ecosystems appear to be especially vulnerable to invasive species as indicated by the often high numbers and percentages of exotic species on oceanic and continental islands. ns I reexamine hypotheses offered to account for чау ине а two interacting processes ose resource the apparently high invasibility of tropical islands and su 'ailability, propagule supply, and мше оор abilities of exotic and iind species. This r sh net resource availability and poor ability of native spec ies to preem —Xmake island communities vulnerable to the establishment and spread of alien species. In addition, ggest historically high rates с introduction have provided opportunity in the form of a diverse and abundant propagule rain 'ombi of exotic species. T nation produces a scenario that is not an optimistic one for island ecosystems. It suggests 8 at these native ecosystems on islands are particularly vulnerable to naturalizing exotics growing on their borders, and while disturbance from a dominants increases the aggressively managed and alien propagule pressure ч they will be highly modified by are dir crm Tropic ‘al islands are an effective early wa species invasions may cause to isolated ecosystems. variety of causes, including pigs, fire, grazing, and natural dieback opportunities for exotic incursions of the canopy tact forests are not immune. Unless бее forests ‚ even in pa plant ling system of the impacts that successi f exoti ч mainland natural area fragmented, degraded and оар they acquire many of the ecological attributes of islands, including limited habitat area, missing functional r oups, dec lining species diversity, and disturbed habitats. A better understanding of invasions on islands may improve our attempts to protect both mainland and island ecosystems from the i impacts of exotic species Key words: munities, tropical islan alien species, exotic species, extinctions, invasive species, invasibility, island ecosystems, plant com- ds. *He who admits the doctrine of the creation [ar eac h not created | PEREIS stoc ocked them far more fully and perfectly than did n —C. Darwin ([1859] 1972: 347—348) Island ecosystems appear to be especially vul- nerable to invasive species. Reviews cite the high numbers and percentages of exotic species on heavily visited oceanic and continental islands (e.g.. Vitousek et al., often are implicated in species extinctions in island ecosystems (Simberloff, 1995; D'Antonio & Dudley, 1995). Tropical islands as well often are character- 1997), and invasive species ized by high alien species densities, unlike tropical mainland ecosystems where the incidence of alien species is low (Rejmánek, 1996). The apparently high invasibility of islands ки suggest that is- land ecosystems hold few lessons for the prevention and management of exotic species in continental ecosystems. However, as mainland natural areas become fragmented, degraded, and depauperate, they acquire many of the ecological attributes of missing functional groups, declining species diversity, and islands, including limited habitat area. disturbed habitats (Laurance & Bierregaard, 1997). A better understanding of invasions on islands may improve our attempts to protect both mainland and island ecosystems from the impacts of exotic spe- cies. ere | reexamine hypotheses offered to account for the apparently high invasibility of tropical is- lands and suggest a simple synthesis based on re- source availability, propagule supply. and relative competitive abilities of exotic and island species. My focus is on invasions of exotic terrestrial plants into native island ecosystems. In the following dis- cussion, I describe some of the hypotheses pro- posed to account for variation in community inva- sibility; however, in developing the synthesis I have relied on few assumptions about the presence of vacant niches or the strength of competitive inter- actions in equilibrium communities. Rather, I as- sume that ecological communities are open to the ! | am grateful to F. sica S. Cordell, and S. DeWalt discussions during its writin ? Institute of Pacific Talande Forestry, USDA Forest jdenslow@fs.fed.us. ANN. MISSOURI Bor. Service, for comments on an earlier draft of this manuscript and for 23 E. Kawili St, Hilo, Hawaii 96720, U.S.A. GARD. 90: 119-127. 2003. 120 Annals of the Missouri Botanical Garden establishment and evolution of new species and that they are constantly changing with variation in the environment and in the abundance of compet- itors, mutualists, diseases, and predators (Hubbell, 2001). The composition and structure of plant com- munities seem largely attributable to differential re- sponses of individual species to resource availabil- ity, habitat conditions, and pest pressures. Beyond these individualistic patterns, assembly rules for plant communities are poorly understood and dif- 1999). ' sume that some degree of invasibility Irem dA ficult to demonstrate (Wilson, .las- all communities and that our challenge is to un- derstand why some communities appear to be more open to the establishment of new species than oth- ers. As D'Antonio and Dudley (1995) observed, gen- eralizations about the vulnerability of islands to in- vasive species often fail to distinguish among in- vasibility, opportunity, and impact, any or all of which may contribute to observed patterns: (1) is- land communities may be more invasible, that is, with similar opportunity exotic species may be more likely to establish new populations in island than in mainland communities; (2) the opportunity for new colonists may be greater on islands because islands may be exposed more frequently to inputs of seeds and other propagules than mainland hab- itats; and (3) the impact of exotic invasions on is- land species and ecosystems may be more substan- tial than on similar mainland communities. Both opportunity and impact are high on many tropical islands. Because islands often lack critical sources of food, forage, and fiber, colonists from early Polynesian voyagers to modern government agencies have promoted plant introductions, in- cluding pasture grasses, timber trees, food crops, ornamentals, and sources of fuel and fiber (Mueller- Dombois & Fosberg, 1998). In addition, rates of unintentional introductions are often high because islands have been important provisioning stations for transoceanic shipping traffic, exposing island habitats to species from wide and heterogeneous sources. Moreover, invasive species have strong im- pacts on islands, because the spatial extent of eco- systems and the population sizes of species are nec- essarily small and thus vulnerable (Simberloff, 1995, 2000; Sakai et al., 2002). Here I focus on the invasibility of tropical island ecosystems and ask whether community processes on islands make them particularly vulnerable to the establishment and spread of alien species and, if so, whether these observations have broader implications for our understanding of invasion ecology. ARE ISLAND COMMUNITIES MORE INVASIBLE? Ecological and evolutionary theorists have sug- gested several factors that may make island com- munities particularly vulnerable to the establish- ment of novel species. Communities with low native species diversity, missing functional groups, dis- harmonic community composition, poorly competi- tive species, and low pest pressures are seen to provide few barriers to the establishment of main- land species adapted to ecosystems with higher pressures from competitors, predators, and pests (Rejmánek, 1996; Simberloff, 2000; Mack et al., 2000). LOW DIVERSITY ON ISLANDS The effect of native species diversity on the in- vasibility of island communities is predicated on two hypotheses: (1) that there are fewer species on islands than in mainland habitats of comparable size, and (2) that high species richness is a deter- rent to the establishment of alien species (Elton, 1958). The proposition that islands are less diverse than mainland areas of comparable size comes largely from island biogeography theory (MacArthur & Wilson, acting processes affect species richness on islands: 1967), which proposes that two inter- The distance of islands from mainland source pools affects immigration rates of new species, and island size affects the number of species of minimum vi- able population sizes that can be accommodated. Thus diversity on small and/or remote islands is likely to be lower than in mainland habitats of com- parable size (MacArthur & Wilson, 1967 Island—mainland comparisons of native plant species diversity аге scant, however. Frequently cited support for the pattern of low plant diversity on islands is still MacArthur and Wilson's original treatise, although often the generalization is regard- ed as sufficiently widely recognized as to need no documentation. However, many factors affect native plant diversity on islands, and their effects are not easily separated (Carlquist, 1974). Geological age, latitude, elevation, habitat diversity, productivity, and proximity to source pools all affect rates of im- migration and diversification as well as patterns of 1993; Simberloff, 2000). Early human settlers on oceanic species coexistence (Ricklefs & Schluter, islands not only brought new species, but caused such widespread extinction and habitat alteration that native lowland floras in particular are difficult to. reconstruct (Steadman, 1995; Kirch & Hunt, 1997; Mueller-Dombois & Fosberg, 1998). factors interact to produce high variation in species These richness among islands and mainland sites and to Volume 90, Number 1 2003 Denslow Weeds in Paradise obscure effects of isolation and area on diversity. At least one study suggests that islands may not be significantly less rich than mainland sites. Lonsdale (1999) found no difference in the species richness of island and mainland native floras from 104: is- land and mainland sites after area was taken into account. Similarly, syntheses of earlier studies sug- gest little difference between species-area curves for island and mainland sites (Rosenzweig, 1995), and low regression coefficients in species-area plots of island plants suggest that correlates of species diversity are more varied than can be accounted for by area and isolation alone (Gilbert, 1980). Island floras may not always be more species poor than those of mainland habitats. If species diversity is assessed at the patch or stand level where most ecological interactions take place, low diversity may be a salient aspect of is- land ecosystems nevertheless. Island floras often are characterized by high levels of endemism. Even within archipelagos, ranges of congeneric species are often distinct and restricted to single islands, mountains, or valleys (Carlquist, 1974; Eliasson, 1995; Wagner et al., 1 are widespread and polymorphic, many have highly though a few species restricted, non-overlapping ranges. For example, speciation within Hawaiian Cyrtandra (53 species, Gesneriaceae), a genus of deep ravines and gulch- es, appears to have been driven by the isolation produced by the dissected topography of highly eroded islands (Carlquist, 1974). In most mainland tropical and subtropical forests, in contrast, within- stand diversity is high (e.g., Heaney & Proctor, 1990), characterized by high diversity within fam- ilies and co-occurrence of many congeneric species (Croat, 1978; Hartshorn & Hammel, 1994). Thus at a stand level, if not at a regional level, species diversity is likely to be low in island ecosystems. he thesis that species-rich communities are less invasible than species-poor communities was first suggested by Elton (1958) and has become a cen- tral tenet of invasion ecology (e.g., Mack et al., 000). However, empirical support for this hypoth- esis seemingly has been ambiguous. At the land- scape and regional scales, several recent analyses have shown that the most diverse communities of- ten also have the largest number of exotic species. Analyses of islands (Lonsdale, 1999), riparian eco- systems (Levine, 2000), and rangelands (Stohlgren et al., 9) all show strong positive correlations between native and exotic species diversity. The authors conclude that likely the same factors that promote a rich assemblage of native species— largely availability of limiting resources—also fa- cilitate the establishment of exotics. In contrast, ex- perimental manipulations of species in plots or microcosms suggest that more diverse assemblages may exploit resources more efficiently and resist establishment of new species (e.g., Levine & D'Antonio, 1999; Levine, 2000; Naeem et al., 2000; Tilman et al., 2001; Kennedy, 2002). These apparently inconsistent results from experimental and observational studies may be resolved by con- sidering process and scale (Levine & D'Antonio, 1999). Both net resource availability and species richness may increase along a resource supply gra- dient if supply increases faster than uptake by the community (Fig. 1). Thus the effect of diversity on biotic resistance to invasions should be assessed in the context of resource supply and demand; net re- source availability rather than species richness per se likely determines community invasibility (Shea & Chesson, 2002). For example, low diversity for- ests at Semliki Forest Reserve, Uganda, were found to be no more invasible than high diversity stands (Rejmánek, 1996). Similarly, observations of Davis et al. (1998) showed that competition intensity be- tween tree seedlings and herbaceous vegetation was correlated with net resource supply rather than bio- mass or gross resource supply. These studies sug- gest two correlates of community invasibility: (1) new species are most likely to become established where limiting resources are available; (2) more re- sources are likely to be used or preempted when species richness is high than when it is low. Under conditions of constant resource supply, high species richness may reduce invasibility (e.g., Levine & D'Antonio, 1999; Tilman et al., 2001). On oceanic islands, low stand-level diversity likely contributes to low levels of resource use, high resource avail- ability, and poor resistance to the establishment of new individuals (e.g., Kitayama, 1996; Kitayama & Itow, 1999). DISHARMONIC FLORAS In addition to their effects on species diversity, long-distance oceanic dispersal and novel environ- mental conditions constitute a strong ecological fil- ter оп island biota. As a result, insular floras are often depauperate in certain taxonomic lineages, 2 tional groups, Ше forms, dispersal character- s, or environmental adaptations, a pattern that Peyas (1974) described as disharmonic. For ex- ample, rain forests on oceanic islands may lack tall, large-seeded, shade-tolerant canopy trees that dom- inate many mainland rain forests. In Hawaii, native palms are confined to a single genus (Pritchardia); in contrast, dwarf, understory, climbing, and clonal palms are common in mainland forests. Similarly, 122 Annals of the Missouri Botanical Garden К, © = R = sp O o 2 К, Environmental Gradient Figure Hypothetic al patterns. of resource supply and demand across an environments gradient. R, = total resource availability; = resources consumed or preempted by species present. The increased resource ам reflects combined effec ts of all co-occurring species, whic in species richness, functional group diversity, and/or redundancy v ted в ап іпсгеаѕе vailable or net 5 1 may often, but not always, be associa vithin functional groups. i unused resources; these may remain unused or be exploited by additional native or alien species there are few species of native lianas in Hawaii. In both of groups can have important effects on forest struc- mainland ecosystems, these functional ture and regeneration processes (Dewalt et al., 2000; Schnitzer et al., 2000; Farris-Lopez et al.. ms.). The availability of vacant niches arising from this disharmony in the structure of island biota is cited frequently as providing opportunity for inva- sive species (e.g.. Mack et al., 2002; Shea & Ches- son, 2002). Among plants, however, the concept of functional groups may be a more useful model than niches, since unique habitat and resource require- ments are difficult to describe among plants. Spe- cies in a functional group share traits that similarly affect ecosystem and community processes (Fow- 1995; faya Aiton (Myricaceae) is an alien, N-fixing tree Addition of its N-rich litter increases nutrient supply rates nes, Denslow, 1996). For example. Myrica that invades recent lava flows in Hawaii. to other species and thus alters successional dy- E namics on this new substrate (Vitousek et al.. 1987). The scarcity of native species that fill this role suggests that this functional group is missing in Hawaii (Fownes, 1995), although the presence of widespread N-fixing shrubs in the fossil record (James, have been filled. Indeed the ability of alien invasive 1995) indicates that this role once may species to alter nutrient supply, disturbance re- gimes, light environments, productivity, and other ecosystem properties may be taken as evidence of unexploited opportunity and missing or poorly rep- resented functional groups. Similarly, Kitayama (1996) and Kitayama and ltow (1999) suggested that low stand productivity and low above-ground biomass in spite of high resource availability were linked to low species diversity and missing func- tional groups in Hawaii and the Galapagos. How- ever, a deterrent effect of native species on the es- tablishment of alien species with similar habitat requirements, resource needs, and growth forms may be difficult to demonstrate. In Hawaii, for ex- ample, much of the diversity among woody plants resides in understory shrubs, yet this group of plants also is well represented in the naturalized exotic flora. High levels of apparent functional redundancy among plants in mainland rain forests suggest that there may be few barriers to the coexistence of many species playing similar functional roles. For Volume 90, Number 1 2003 Densl 123 nslow Weeds in Paradise example, about 1700 species of vascular plants have been recorded in 1536 t the La Selva Biological Station, Costa Rica, of which there are 4A species of Piper (Piperaceae), 39 species of Psy- chotria (Rubiaceae), and 25 species of Miconia (Melastomataceae) (Wilbur, 1994), which are understory shrubs and small trees. There ы =, almost all are 107 species of lianas and 323 species of trees (Hartshorn & Hammel, 1994; Wilbur, 1994). Dif- ferences among these species, for example, in light requirements, growth form, or climbing mecha- nisms, are small, suggesting strong overlap among them in habitat requirements. In addition, studies on the 50-ha permanent plot in Panama fail to re- veal significant habitat differences among the ma- jority of coexisting trees and shrubs there (Hubbell et al., 1999; Harms et al., 2001). These patterns in relatively homogeneous forests suggest that the presence of many ecologically similar species is not a strong deterrent to the occurrence and persistence of plant species in mainland tropical forest. In this context, it also seems unlikely that the presence of native species would be a significant barrier to the establishment of ecologically similar alien species in island ecosystems. NATIVE SPECIES ARE POOR COMPETITORS Native species on islands often appear to be poor competitors (Darwin, [1859] 72; Carlquist, 1974). For example, wet and mesic forests through- out the Hawaiian archipelago are dominated by a single highly polymorphic species, Metrosideros po- lymorpha Gaud. (Myrtaceae). These Metrosideros forests are characterized by relatively open cano- pies, widely spaced crowns, and inefficient light ab- sorption (Cordell & Goldstein, 1999). As a result, considerable light reaches the forest understory (Cordell & Goldstein, 1999), where native and alien grasses, herbs, and shrubs are able to establish. In addition, net CO, growth rates of M. polymorpha are generally low (Burton, 1982; Burton & Mueller-Dombois. 1984), and M. polymorpha shows little plasticity in re- assimilation, leaf turnover, and sponse to increases in light or nutrient supply (Cor- dell et al., 2001; Austin & Vitousek, 2000). Where growth or carbon fixation rates have been com- pared, lower rates are often measured in native Ha- walian than in comparable alien species (Pattison et al., 1998; Baruch & Goldstein, 1999; Durand & Goldstein, 2001). Similarly, the exotic Himalayan raspberry, Rubus ellipticus Sm., is replacing the na- tive ‘akala, Rubus hawaiiensis A. Gray, in the tree- fall gaps both require for establishment in montane rain forest. Although the two species have similar dispersal mechanisms and habitat requirements, the alien К. ellipticus exhibits faster growth rates, a more efficient canopy configuration, and greater seed production increasing the likelihood that it eventually will replace the native species in these forests (Denslow, unpublished data The reasons why island species undi be poor competitors are various. Loss of resilience in the gene pool may be a consequence of founder effects, small population sizes, strong post-establishment selection, drift, and low vagility of pollinators and dispersers (Carlquist, 1974; Loope & Mueller-Dombois, 1989; Kaneshiro, 1995). short distances that characterize high tropical is- lands may constrain development of specialized ad- inbreeding depression, The steep environmental gradients over aptations. For example, Kitayama (cited in Den- slow. 2001) has suggested that the dominance of oceanic island forests by species with wide ecolog- ical ranges may be linked to their low productivity and high invasibility. In his study of Hawaiian and Bornean rain forests (Kitayama, 1996), he found that species on Borneo had narrower elevation ranges than did species in Hawaii. He suggested, therefore, that the Bornean species may be better adapted to their environments and thus present stronger barriers to alien species than the more broadly adapted species in Hawaii. The links among adaptive ranges, competitive ability, and re- sistance to invasion deserve further exploration in this regard. A large proportion of island endemics are threatened or endangered due in part to their small ranges (Simberloff, 2000; Sakai et al., 2002), and the positive correlation between range size and local abundance has been widely acknowledged (e.g.. Brown & Maurer, 1984; Kelly. 1996). These studies suggest that on islands, depauperate and disharmonic floras and poorly competitive species may result in low productivity and high ecosystem invasibility. Loss of dispersal efficiency in island species may contribute to their poor competitive abilities. The generality of this pattern in many unrelated. plant groups (Carlquist, 1974) pressure to avoid loss of reproductive output and suggests that selective reduce investment in dispersal mechanisms is strong оп islands. Compounding the evolutionary loss of dispersability has been the historic loss of avian frugivores in Hawaii (James, 1995), loss of pollinators, and lack of a persistent seed bank in many native species (Drake, 1998), all of which contribute to low seed availability. As a result, pop- ulation growth of native species may be strongly dispersal-limited in some habitats. Indigenous spe- cies thus may be less likely to fully occupy suitable Annals of the Missouri Botanical Garden habitats than aliens with better dispersal mecha- nisms, greater seed output, and larger seedbanks, and less likely to reach and occupy critical, but ephemeral, establishment sites such as treefall gaps and nurse logs. Dispersal limitation contributes to less-than-full exploitation of limiting resources in all ecosystems (Hubbell, 2001), but may be partic- ularly important on islands. EFFECTS OF LOW PEST LOADS The enemy release hypothesis proposes that alien plant species in their introduced ranges ex- perience lower pest loads than co-occurring native species and than they do in their native ranges (Keane & Crawley, 2002). As a result, populations of exotic species may be released from control by natural enemies, enhancing their competitive po- sitions relative to native species, which may remain under pressure from specialist and generalist ene- mies. There have been few appropriate tests of this hypothesis and in a few cases only were significant impacts of either generalist or specialist herbivores on the exotic plant species recorded (Keane & Crawley, 2002). Classical biological control —the introduction of specialist pests and pathogens to control invasive species in their introduced rang- es—is based on this premise. There are several ex- amples of intentional and accidental introductions of pests limiting the populations of their hosts (Lou- da et al., 1997; Strong & Pemberton, 2000), and Louda (1982) and DeWalt et al. (unpublished ms. offer evidence that pests limit both growth and hab- М itat distributions of weedy plants. In addition, data presented by DeWalt et al. (unpublished ms.) sug- gest that survival of Clidemia hirta (L.) D. Don (Me- lastomataceae) is more strongly affected by insect and pathogen impacts in its native than in its in- troduced range. The role of pest and pathogen pres- sure in island plant invasions is still poorly under- stood. Furthermore, indigenous island species, which also originated as waifs, are thought to be under low pressure from natural enemies as well (Carl- 1980). I isolated habitats, insular floras are thought to have quist, n the process of dispersal to new and left their specialized pests behind. While subse- quent diversification has produced many special- ized associations of plants and their natural ene- mies (Swezey, 1954), the common lack of defensive compounds and structures among indigenous island species suggests that plant pathogen and herbivore impacts are not high. As a result, island plants may be particularly vulnerable to the introduction of ex- olic pests such as ungulates (Mueller-Dombois & Fosberg, 1998; Richardson et al., 2000). If islands are depauperate in natural enemies or dominated by specialist herbivores, both exotic and native spe- cies may be under little control by their natural enemies. The net effect of low pest loads on native and alien plant species on islands is then difficult to predict. We lack critical information on the role of natural enemies in limiting plant population growth, abundance, and distribution in both main- land and island habitats. On balance, however, the more rapid growth rates and plastic physiological responses of alien species may result in their great- er population growth where pest pressures are low (Keane & Crawley, 2002). SYNTHESIS Net resource availability is ап important com- ponent of community invasibility. The role of re- source pulses in the spread of alien species has been noted widely (Vitousek et al., 1997; D'Antonio et al., 1999; Mack et al., 2002). The generality of this relationship can be extended usefully to in- clude chronic as well as temporary resource avail- ability, such as that provided by disturbances, and to affect native as well as alien species. Seedlings of both native and alien species are likely to be- come established where resources are chronically or temporarily under-used, as following natural or anthropogenic disturbance; under conditions of nat- urally high levels of resource supply; where re- sources are incompletely exploited by the existing а. communily; or where resource supply has been augmented. For example, it is widely recognized that both natural and anthropogenic disturbances facilitate the establishment of alien species (Rej- mánek, 1989; Horvitz et al., 1998; D'Antonio et al., 1999). in part because disturbance reduces com- petition and increases local resource availability. In intact native communities, ant mounds, treefall gaps, and riparian habitats are important sites for seedling establishment (Platt, 1975; Denslow, 1987; Levine, 2000). Similarly, where disturbances are ecologically novel or unusually frequent or in- lense, we can expect that some native species as well as some exotic species may find suitable en- vironmental conditions. Diversity and abundance of both alien and native species also are often highest in resource-rich habitats (e.g.. Stohlgren et al., 1999). and, conversely, communities characterized by resource limitation (in Hawaii, alpine and sub- alpine ecosystems, communities on young lava flows, and dryland forests) are often notable for scarcity of exotic invasives. Other attributes pro- posed to increase community invasibility, such as Volume 90, Number 1 2003 Denslow 125 Weeds in Paradise vacant niches, novel anthropogenic disturbances, or disequilibrium conditions, also address the avail- ability of resource availability, but might be seen as special examples of a more general phenomenon. In isolated tropical island habitats, resource availability likely is high because indigenous spe- cies do not effectively use or preempt them. Low local species richness, low diversity of functional groups, and low redundancy within functional groups suggest that few species may be available to take advantage of establishment opportunities or unused resources. Dispersal limitation, character- istic of all plant communities, may be particularly strong on islands, reducing the occupancy rate of suitable habitats and leaving sites and resources available for others. Where propagule pressures are high, the vulnerability of island communities to es- tablishment of exotic species will be particularly apparent (Rejmánek, 1989; Drake, 1998; Levine. 000). Invasibility, however, is not a unilateral function of habitat characteristics, but rather the interaction of habitat and species. Limiting resources are avail- able relative to the physiological and morphological capacity of plants to exploit them. Low growth plas- ticity in response to resource availability in many native species contributes to the underutilization of resources and thus invasibility of some island com- munities, and exotic species with more plastic growth responses may compete more effectively for natural establishment sites such as treefall gaps and decaying logs and take better advantage of nov- el disturbances such as pig foraging sites. Assuming that fundamental tradeoffs between rapid growth rate in high light environments and persistence in shaded environments characterize alien and native species (Reich et al., 1997: H bell, 2001), both alien and native species should be able to thrive across a full spectrum of environ- ub- ments. Invasive species often are assumed to have characteristics of disturbance-adapted species (e.g.. rapid growth in resource-rich environments and co- pious production of well-dispersed seeds); however, in part this pattern may reflect historic pathways of introduction associated with development of agri- culture and rangelands (Mack, 1992), rather than inherent characteristics of invasive species. А ris- ing global horticultural trade is introducing species that are shade-tolerant, spread vegetatively, and/or have large seeds (Reichard & Hamilton, 1997) and thus are potentially invasive in intact forest eco- systems This review suggests that fundamentally two in- teracting processes—high net resource availability and poor ability of native species to preempt those resources—make island communities vulnerable to the establishment and spread of alien species. His- torically high rates of introduction have provided opportunity in the form of a diverse and abundant propagule rain of exotic species. The combination produces a scenario that is not an optimistic one for island ecosystems. It suggests that these native ecosystems on islands are particularly vulnerable to naturalizing exotics growing on their borders, and that while disturbance from a variety of causes, in- cluding pigs, fire, grazing, and natural dieback of the canopy dominants, increases the opportunities for exotic incursions, even intact forests are not im- mune. Unless these forests are aggressively man- aged and alien propagule pressure reduced, they will be highly modified by new exotic introductions. There are implications of this scenario for main- land ecosystems as well. With rising human pop- ulations, increased incursions into wilderness are- as, and global environmental change, mainland preserves will assume many island-like character- istics. Habitat fragmentation and large-scale habitat loss will increase habitat isolation, reduce the ex- tent, and increase the edges. Fragmentation of for- ests is widely associated with increased distur- bance, alteration of edge and sometimes interior habitats, loss of species, and increased invasibility 1997). Effects of population reductions and species extinctions will simplify ecosystem structure, reduce population sizes, produce disharmonies in species assemblag- e.g., Laurance & Bierregaard, ~ es, and reduce functional redundancy. Loss of high- er trophic levels, in particular, may result in in- creased populations of herbivores and edge species. Increased intensities and rates of exploitation will free resources (light, space, nutrients) and increase opportunities for exploitation by novel species. Ris- ing rates of human traffic through suburban expan- sion, road extensions, global trade, ecotourism, and population movements will provide rising and re- peated exposure to a diverse seed rain from exotic species. Tropical islands are an effective early warning system of the impacts successive waves of exotic species invasions may cause to isolated eco- systems. As island managers develop strategies for preventing and controlling these invasions, they will be watched by many on the mainland. Literature Cited Austin, A. T. & P. M. Vitousek. 2000. OMNE. and litter decomposability of Me dil ig in native forests on Hawaii. J. Ec 129-138. Baruch, Z. & G. Goldstein. 1999. Leaf construction cost, nutrient concentration, and net CO, assimilation of na- Precipitation, de- trosideros ol. 88: 126 Annals of the Missouri Botanical Garden tive and invasive species in Hawaii. Oecologia 121: dud А Brown, J. H. Maurer. 1984. On the relationship & aes aa 'e and distribution of species. Amer. Naturalist 124: 255-27 Burton, P. J. 1982. The effect of temperature and light on Metrosideros i seed germination. Pacific Sci. 36: 229-240. & D. Mueller-Dombois. 1984. Response of Me- trosideros к ТРА to experimental canopy opening. Ecology 65: 779-791. Carlquist, S. 1974. Island Biology Columbia Univ. New York. 1980. Hawaii: A Natural m Pacific Trop- ical Botanie al Garden, Lawai, Ha Cordell, 1999, Light distribution and н of Metrosideros polymorpha forests i both ends of a ipe anc Hawaii. Selbyana ES 50-356. Press, Goldstein. substrate age gradient in ya "inzer . Vitousek. 2001. Morphologic ‘al Аре? al i ыа to N and P ertilization in nutrient-limited но canopy trees in Hawaii. Tree Physiol. 43—: Croat, T. B. 1978. Flora of Barro Colorado E Stan- ford Univ. Press, Stanford, California. D'Antonio, C. M. & T. L. Dudley. 1995. vasions as аре ents of change on islands lands. Pp. 103-121 in P. М. Vitousek, L. L. H. мыны (editors), Islands: Biological Diversity and Ecosystem Function. Springer, 2 г ————, ———— & M. 9c о in- 19 -Doniba 'e and bi- ый invasions: Direct уан ‘ts and feedbacks. Pp. 413—452 in L. R KE (editor), Ecosystems of Dis- turbed Ground. Elsevier, Amsterdam. Darwin, C. R. [1859] 1972. The Origin of Species, Oth Ed. m Murray, London. Davis, M. A., K. J Wrage & P. B. Reich. 1998. Com- petition between tree seedlings and herbaceous vege- lation: ee for a theory of resource supply and de- mand. J. Ecol. 86: 652-661. Denslow, J. E 987. Tropical esie a and tree Annual Rev. Ecol. S : 431-451. 96. Functional A КАЙ E ی‎ to a Pp. 127- s. R. Dirzo & J. Н. Cushman (editors), Rivers не Ped m Processes in Tropical Forests. Springer, New York. X The ecology of ae biotas. 23-424. species diversity. 1 Trends Ecol. DeWalt, үз Schnitzer & J. Density and hera of Evol. .J., 2000. lianas ac a oa nce in g ce Pi Panamanian lowland forest. J.” . Ecol. 16: 1- Drake. e A 1998. Relationships among the seed id ieod e and = getation of a Hawaiian forest. J. Sei. 9: 103-11 Dur nd L. a " С, 2001. Photosynthesis, d and nitrogen use efficiency in ers nd invasive tree ferns in Hawaii. Oecologia 126: : 1. . Denslow. Goldstein. 35: Eliasson, U. 1995. Patterns of diversity in island plants. Pp. 35-50 in Р. M. Vitousek, L. L. Loope & Н. Adser- sen (editors), Islands: Biological Diversity and Ecosys- ger, Berlin. The Ecology of Invasions of Animals and Plants. Methuen, London, England. Fownes, J. H. 1995. Effects of diversity on produc ivi Quantitative distributions of traits. Pp. 177—186 in M. Vitousek, L. L. Loope & Н. Adsersen (editors), Is- Biological Diversity and Ecosystem Function. Springer, Ber Gilbert, F. S. 1980. The RUM theory of island bio- geography: Fact or fic че . Biogeogr 7: 2090-235 Harms, K. E., R. Condit, ОР. Hubbell & R. B. Foster 2001. Habitat assoc faf ans of trees and hap in a 50- о]. 89: € - 994. i Um and floristic iEn. Pp. 73—89 in L. A. McDade, К. S. Вама, Н. А. тн ж чае & G. S. Hartshorn (edi- tors), La Selva: Ecology and Natural History ms a Neo- tropical Rain Forest. Univ. age Press, Chicago. Heaney, A. & J. Proctor. 1990, Preliminary ie on forest structure and ae on Volean Barva, Costa Trop. Ecol. € -319 Horvitz, C. C.. J. B. bel Rus cMann, A. Freedman & R. H. Hofstetter. 1998. Func tional roles of invasive non-indigenous plants in hurricane-affected subtropical hardwood i oe Ecol. Applic. 8: 947-97: Hubbell, 2001. The Duifiad cona Thear of Bio: diversity d ia Princeton Princeton, New Univ. Press А sey. —— S. T. O'Brien, K. E. Harms, R. Condit, B. Wee hsler, S. J. Wright & S. Loo de Lao. 1999, Light-gap disturbances, recruitment. limitation, and tree diversity in a Neotropical forest. Science 283: e James, H. F. 1995. Prehistoric e pp ж ecological Dad on oceanic islands. Pp. 87-102 in P. M. Vitou- sek, доре Adsersen qud ишк В1- ologic al Diversity and Ecosystem Function. Springer, Berlin. Kaneshiro, D. Saa K. Y. 1995. Evolution, spec lation, the genetic structure of island populations. Pp. 3 in P. M. Vitousek, L. L. Loope & H. Adsersen edil жч Is- lands: Biological Diversity and Ecosystem Function. Springer, Berlin. Keane, R. M. & M. J. Crawley. 2002. Exotic plant in- vasions y bie enemy release hypothesis. Trends Ecol. ol. x 70. Ta oe plant functional types us- data bases: Ecolog ical correlates of plant M Sci. ae an ri T. A., S. Naeem, К. M. Howe: J. . Tilman & B в. Reic h. p Ke Ily, ing bus Н. Knops, Mori as a bar- rier to ec ologic 'al invasion. Nature 417: &' 38 Kirch, P. V. . L. Hunt. 1997, nen i Ecology in the Pacific Islands. Yale Univ. dian New Haven. Kitayama, K. 1996, Patterns of species diversity on an oceanic vs a continental island mountain: A hypothesis on i ies diversification. J. Veg. Sci. 7: 879-888 . Ном. 1999, Abov eground biomass and soil nutrient pools of a Scalesia deg ulata montane forest on Santa С s. l ruz, Galapagos. Ecol. Re : 405— Laurance, W. F., Jr. & R. O. Bierregaard. 1997. Tropic 'al Forest Remnants. Univ. Chicago Press, Chica Levine, dd M. 2000. Species diversity and hiologic al in- nisu local process to community pattern. Sci тепсе н 352-854. C. M. D'Antonio. 1999. Elton revisited: A re- view к ca nce linking diversity and invasibility. Oi- kos 87 Lonsdale, = м. "1999. Global patterns of plant invasions and the concept e invasibility. Ecology 80: 1522-1536. Loope, L. L. & D. Mueller-Dombois. 1989. Characteris- tics of invaded in with special боксе to Ha- Volume 90, Number 1 Denslow 127 Weeds in Paradise waii. Pp. 257-280 in J. A. Drake, F. DiCastri, R. H. Groves, F. J. Kruger, Н. A. Mooney, M. Кејтапек & M. Williamson (editors), Biological Invasions: Global Perspective. Wiley, New York. Louda, 5. Distribution ec ology: Variation in plant recruitment over a gradient in relation to insect seed dw cm vu Monogr. 52: 25-41. . D. Kendall, J. Connor & D. Simberloff. 1997. Ecological P awe ts ad an insect introduced x = bio- logical — of weeds. Science 277: 1088— MacArthur, R. H. & E. O. Wilson. 1967. The = Se of Bland Mabel Princeton Univ. Press, Princeton, I New Jersey. Mack, R. N. 1992. е of invading plant spe- cies. Pp. 42— W. . P. Stone, J. T. Tunison & € Smith айн, poe Plant spen in Native Eco- systems of Hawai‘i: Managen and Research, Uni- ` Hawaii ксн Fr Park Resources онар : nit, Honolulu, Hawaiʻi. Simberloff, X M. Lonsdale, H. Evans, i Clout . Bazzaz. 2000. Biotic Invasions: Cause Bpidemicloay Global Consequenc es and Control. Feo. logical E iety of America No. 5: Issues in Ecology. . Simberloff, W. M. Lonsdale, Н. Evans & M. Ёш. ps Biotic invasions: Causes, epidemiology, E consequences and control. Ecol. Applic. 10: 689—710: Muelle r-Dombois, D. & F. R. Fosberg. 1998. зе Tropical Pacific Islands. Springer- Чейай, New Vegetation a © =, - = — ҮогК. Naeem, S., J Knops, D. Tilman, K. M. Howe, T. Kennedy & 8. Cala: 2000. Plant dinalis increases resistance to invasion in bs absence of covarying ex- trinsic fac i Oikos 91: Pattison, R. R., G. vhs & А. din s. 1998. Growth, biomass i 'ation and photosynthesis of invasive and native Hawaiian rainforest. species. Oecologia 117 59. . 1975. The colonization and formation of equi- librium plant species associations on badger distur- pr ‘es in a tall-grass prairie. Ecol. Маг. 15: 285 uo. P. B Walters & D. S. Ellsworth. 1997. From — to qiie Global convergence in е functioning. Proc. Natl. Acad. Sci. U.S.A. 94: 1373 13734. Heichard, 5. . W. Hamilton. 1997. Predicting in- vasions of d planis ۰ :ed into North America. C idet ta Biol. Rejmánek, M ion Imai of plant communities. Рр. 369— 388 i in J. A. Dra © DiCastri roves, F. J. Kruger, M. Re jmánek & M. Williamson (editors), Biologic al мене A Global Perspective. Wiley. New York —— ———,. 1996. Species richness and resistance to inva- sions. Pp. 153-172 їп G. Orians, R. Dirzo & pos (editors), Biodiversity and "psum m Bin: Tropical Forests. Springer, New York. M. D Ше 5. Milton I ^ sses in а иф, D. M., М. Allsopp, С. & M. Rejmánek. 2000. Plant invasions—The role of mutualisms. Biol. Rev. Cambridge Philos. Soc. 75: 65— 93 Ricklefs, В. E. & D. Schluter. 1993. Species Diversity in Ecological Communities. Univ. Chicago Press, Chi- cago go. Rosenzweig, M. L. 1995. Species Diversity in Space and Time. Cambridge Univ. Press, Cambridge. Sakai. A. K.. W. L. Wagner & L. A A. Mehrhoff. 2002. Pat- terns of endangerment in the Hallan flora. Syst. Biol. 51: 276-302. Schnitzer, S. A., J. W. Dalling & W. P. Carson. 2000. The impact of lianas on tree regeneration in tropic ‘al forest canopy gaps: Evidence for an Dre tive pathway of gap-phase regeneration. J. Ecol. 88: 655—666. K.& P. C Shea, hesson. 2002. пен ecology theory as a meno =ч biological invasions. Trends Кс ol. : 170-176 sibi D. 1995. Why do introduced species appear evastate islands more than mainland areas? Pacific Sci. 49: 87—97. . 2000. Extinction-proneness of island species— Causes “ management implications. The Raffles Bull. Zool. 48: 1—9. Tim р. W. Prehistoric extinctions of Pacific Island. birds: iod meets zooarchaeology. Sci- ence 267: d 132. cr T. J.. D. Binkley, g e ткы М. А. mele x A. Bull, suki, G. New "n ш & Y. OA exotic plant species invat 2n nes of native plant diversity. Ecol. Monogr. 60: — BUDE D. R. & R. W. Pemberton. 2000. Biological con- hel: e s oria species—Risk and reform. Science 288: 1906€ dye: ENT in Hawaii. Bishop‏ .954 ا Spec id ot -266. Tilman, D., P. B. Reic ^ T Koo . D. Wedin, T. Mielke & ;. Lehman. aint Diversity ind productivity in a long- term grassland рени. Science 294: 843—845. Vitousek, P. M., L. R. Walker, L. D. Whiteaker, D. Muell- er-Dombois & P. A. 4 1987. Biological invasion by Myrica faya alters ecosystem ы Bka in Ha- waii. Science 238: 802-804. yG: "Antonio, L. L. ‚ M. Rejmánek E R. Westbrooks. 1997. Introduc = ресе: А significar иан nt to diu caused global change. New 4 land J. Ecol. 21: 1-16 Wagner, W. . Herbst & S. H. Sohmer. 1999. Man- ual of the n ring ele of Hawaii, Revised ed. Univ. Hawaii Press, Honolulu. Wilbur. R. L. 1994. Vascular plants: T interim check- list. Pp. ieri D A. McDade, К. S. Bawa, Н. a Hespenheide & Hartshorn не s) La Selva Ecology br Natural History - нз ‘al Rain For. Meroe ‘ago Press, Chica M йв, ]. n 1999. Assembly "E in plant communities. Pp E — < 30-165 in E Weiher & P. Keddy (e тве :с0- logical Assembly Rules: Perspectives, Advances, Re- treats. Cambridge Univ. Press, Cambridge. Volume 90, Number 1, pp. 1-128 of the ANNALS OF THE 2T RI BOTANICAL GARDEN was published on February 17, 200 Ш [| ШШ J = 1753 00296 5 :ANNALS OF THE MISSOURI BOTANICAL GARDEN == | AND MISSOURI BOTANICAL GARDEN ANNUAL REPORT ARE NOW AVAILABLE IN JSTOR! JSTOR’, a not-for-profit organization, is an important endeavor dedicated to helping the scholarly community take advantage of advances in electronic technologies. The JSTOR Archive makes available the complete backfiles of Annals of The Missouri Botanical Garden, Missouri Botanical Garden Annual Report (1890-1912), and more than 100 other scholarly journals to researchers through participating libraries. JSTOR has recently announced the release of its Ecology & Botany Collection. Developed in cooperation with the Ecological Society of America, the Collection includes the archives of twenty-nine journals, offering online access to more than 1 million pages of important historic literature. The Ecology & Botany Collection American Journal of Botany American Midland Naturalist American Naturalist Annals of the Missouri Botanical Garden Annual Review of Ecology and Systematics Biotropica Brittonia Conservation Biology Diversity and Distributions Ecological Applications Ecological Monographs Ecology а Evolution Functional Ecology Global Ecology and Biogeography Letters International Journal of Plant Sciences Journal of Animal Ecology Journal of Applied Ecology — Journal of Biogeography Journal of Ecology Journal of Tropical Ecology Journal of The Torrey Тем. Society . Limnology and Oceanography Missouri Botanical Garden Annual перон New Phytologist Paleobiology. Quarterly Review of Biology Systematic Biology Systematic Botany Information regarding J STOR is available at http://www.jstor.org 120 Fifth Avenue, New York, NY 10011 CONTENTS The Paraphyly of Cortaderia (Danthonioideae; Poaceae): Evidence from Morphology and ; Chloroplast and Nuclear DNA Sequence Data __. Nigel P. Barker, H. Peter Linder, Cynthia M. Morton & Mark Lyle 1 Endemistx i in the Mexican Flo A Comparative Study in Three Plant Groups — — — Claudio Delgadillo M., José Luis Villaseñor Ríos & Patricia Dávila dE 25 ўба, Улда Moths, and Coevolution: A Review |... — Olle Pellmyr 35 Chromosome Reports from South American Hypochaeris Uaia ==. ue — Hanna ens e Е. Stuessy, Jürke Grau & Carlos M. Basa 56 Biological ш the 48th A ES f the Missouri Bot 1 Garden — — Invasion Biology: An Emerging Field of Study > Баай Hayden Reichard & Peter S. White 64 | The Threat of Invasive Alien Species to Biological Diversity: Setting a Future Course | —— Elizabeth А. Chornesky & John M. Randall 67 E Plant Naturalications and Invasions i in the Eastern United States: 1634-1860 — | Richard N. Mack 77 Б Ants: Unwanted Partners in Ant-Plant Interactions? LoriLach 91 Molecular Systematics and the ыы of Invasive Plants: А mic Study of Tamarix — ¬ (Tamaricaceae) — — John F. Gaskin 109 —.. Weeds i їп Paradise: Thoughts on the Invasibility of Tropical Islands dup) S. Denslow 119 = tava ERES maris pentandra. Plate LXXIX i in F. lora Rassias. saad TI PS. Pallas, нс Pre К.Е Friedrich, and J.J. prr cias Petropoli, Leningrad; 1784-1788. A book — cc hand-colored copper engravings of plants: нана" is ‘European and Asian areas of the са Russian topi: and one. ы а d collection. of such. books in the ыу of the Ише. = ' tanical Саг QK ( A757 C. | of the Missouri Dotanical ттр Volume Annals of the Volume 90, Number 2 i 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. All manuscripts are peer-reviewed by quali- fied, independent reviewers. Authors should write the Managing Editor for informa- tion concerning arrangements for publishing in the ANNALS. Instructions to Authors are printed in the back of the last issue of each volume and are also available online at www.mobot.org/mbgpress. Editorial Committee Victoria C. Hollowell Editor, Missouri Botanical Garden Amy Scheuler McPherson Managing Editor, Missouri Botanical Garden Diana Gunter Associate Editor, Missouri Botanical Garden Kevin Brown MBG Press Assistant, Missouri Botanical Garden Barbara Mack Administrative Assistant 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 Charlotte Taylor issourt Botanical Garden Henk van der Werff Missouri Botanical Garden For A URS information contact ANNALS OF THE MISSOURI GARDEN, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2003 is $145 per SOET U.S., $155 Canada & Mexico, $180 mn other ие Как шша pe vol- ume. scription pite of the ANNALS. annals@mobot. org (editorial quedes - http://www.mobot -org/mbgpress 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. | Periodicals postage paid at St. Louis, MO and additional mailing offices. POSTMASTER: Send address changes to ANNALS OF THE MISSOURI BOTANICAL GARDEN, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. | . The Annals are abstracted and/or indexed i in AGRICOLA йош АРТ Online, BIOSISG, . ingenta, ISI databases, JSTOR, Researc h Alert®, and Sci Se © Missouri Botanical Garden 2003 a The mission of the Missouri Botanical Garden is to discover and share knowledge don Bis and | E laor environment, in order to preserve and enrich life. | Ө This paper meets the requirements of ANSI/NISO 239.48-1992 (Permanence of Papei). Volume 90 Annals Number 2 of the WZ 2003 Missouri Botanical Garden PHYLOGENETIC ANALYSES Michael J, Zanis,* Pamela s. Soltis, AND PERIANTH EVOLUTION кк Ap Piae Zimmer,’ IN BASAL ANGIOSPERMS' ERA NUN ABSTRACT Using a compartmentalization approach, we conducted phylogenetic analyses of the basalmost extant angiosperms using sequences from six genes (over 12,000 bp per taxon) from all three genomes (c hloroplast—aipB, r cL; nuclear— 18S rDNA, 268 rDNA; mitochondrial—matR, a Trees pita from parsimony and maximum likelihood analyses of the compartmentalized data are identical. We find strong support (100% for each node) for the earliest-branc hing angiosperms: Amborellaceae, Nymphaeaceae, and an Austrobaileyales clade (Illiciaceae, Schisandraceae, Trimeniaceae, Austrobaileyaceae). Whereas most recent studies using multiple genes provided poor resolution and support for rela- tionships among the remaining basal angiosperms (C barê ме, Chloranthaceae, Canellales ( = Winterales), i perales, monocots, Magnoliales, Laurales), with compartmentalization, we find high levels (> 90%) of boot for relationships — these clades. Canellales and Piperales form a strongly supported (100%) sister казы! that і is, in turn, sister to a well-supported (100%) clade of Laurales and Magnoliales. Canellales + Piperales and Magnoliales + Laurales form : a well-supported magnoliid clade. Ceratophyllaceae are strongly ki pais (100%) as sister to the monocots; the monoc brc ne tophyllaceae clade is well supported (8696) as sister to all remaining angiosperms (Chlor- anthaceae, the magnoliid clade, and eudicots). The addition of entire 268 rDNA sequences clearly contributed to this increased internal support. We examined the diversification of perianth phyllotaxis, merosity, and йыда using our phylogenetic hypothesis for angiosperms. Ancestral perianth phyllotaxis and merosity are equivocal for each node of the Amborellaceae, Nymphaeaceae, Austrobaileyales grade; however, an undifferentiated perianth is rec Nescit ! We thank Mark Fishbein, Larry Hufford, Robert Kuzoff, and Shelley McMahon for useful comments on early versions of this manuscript; Robert Kuzolf for technical advice with sequencing and phylogenetic analyses; Mark Fishbein for technical айне е оп phylogenetic analyses; and Youngbae Suh and Sangtae Kim for providing additional 2 sequences. We would like to hank Ed Schneider, Peter Endress, and James cha for providing helpful sa ila regarding und phyllotaxis, merosity, and differentiation in the Nymphaeaceae. We thank the vue anonymous ers for critical comments. This work was DM by NSF grant DE B- 9707781 to P. S. S andD. Е. 5, а Fulbright Distinguished P eam to D. E. S. and P. S. S., a doctoral اا‎ мари DA pis to M. Z. and Ws os from the Laboratory of Molecular Systematics, Smithsonian Institutio chi 01 of Biological Sciences, Washington State University, Pullman, Washington 99164, orida Museum of Natural History and the Genetics Institute, University of Florida, Сому Florida 32611, Department of Ecology and oe Biology, University of Michigan, Ann Arbor, n :higan 48109, U.S.A. 5 Laboratory of Molecular Sys tics, Smithsonian Institution, Washington, D.C. 2056( 6 Department of Botany and “ке ipm 's Institute, University of Florida, Gainesville, Flarida 32611, USA. 7 Author for correspondence. ANN. Missouni Bor. GARD. 90: 129-150. 2003. 130 Annals of the Missouri Botanical Garden ded ancestral state for the angiosperms. Trimery and ien perianth phyllotaxis have played a major role in basal iosperm perianth evolution and represent the ancestral sta and Austrobaileyales. A differentiate d perianth has apparently evolved multiple times. basal angiosperms, floral evolution, me rosity, perianth, phylogeny m Amborella, Nyn Key words: mphaeaceae, es for the large clade comprising all angiosperms other Understanding phylogenetic relationships among extant basal angiosperms is critical for reconstruct- ing the evolution of numerous traits, including wood anatomy, chromosome numbers, pollen mor- phology, and floral morphology (Endress, 1986, 1994a; Endress & Hufford, 1989; Endress & Ig- ersheim, 1997; Igersheim & Endress, 1997; Qiu et al. 2000; P. Soltis et al., 1999a; Thorne, 1992; Walker & Walker, 1984). In recent classifications, most basal angiosperms have been recognized as subclass Magnoliidae (Cronquist, 1981; Takhtajan, 1969, 1997; Thorne, 1992). However, these groups have typically formed a grade (Donoghue & Doyle, 1989; Doyle, 1994; Nandi et al., 1998; Savolainen et al., 2000; D. Soltis et al., 2000; P. Soltis et al., 19993), and only rarely a clade (Chase et al., 1993; Qiu et al., 1993), in phylogenetic analyses of an- giosperms. Previous analyses of relationships among basal angiosperms based on morphology have generally provided conflicting results regarding the early branches of angiosperm phylogeny. Phylogenetic analyses of morphological data (Donoghue & Doyle, 1989) suggested that Magnoliales were sister to all other angiosperms, although in trees only one step longer Nymphaeaceae appeared in this position. In that same study, paleoherbs (Piperales, Nymphae- aceae, and monocots) formed a clade, Chlorantha- ceae and Austrobaileyaceae were nested within Laurales, and Winteraceae were allied with Austro- baileyaceae. In contrast, Loconte and Stevenson's (1991) phylogenetic analysis of morphological data placed Calycanthaceae as sister to all remaining angiosperms; as in Donoghue and Doyle (1989), Chloranthaceae were nested within Laurales, and Winteraceae were included within Illiciaceae and Schisandraceae, but Austrobaileyaceae appeared within Magnoliales. There have been several attempts to analyze phy- logenetic relationships among basal angiosperms using both RNA and genetic analyses using 185 rRNA sequences (Ham- by & Zimmer, 1992) suggested that Nymphaeaceae, NA sequence data. Phylo- Piperales, and monocots were groups to all remaining angiosperms. In agreement with the broad analysis of 500 rbcL sequences from 1993), Qiu et al.’s (1993) analyses of rbcL sequences focusing on basal an- seed plants (Chase et al., successive sister giosperms suggested that Ceratophyllaceae were sister to all other angiosperms, with the remaining basal angiosperms forming a clade sister to the eu- dicots. Qiu et al. (1993) also recovered four major lineages within this large basal angiosperm clade: Magnoliales, Laurales, Piperales, and Nymphae- aceae. Combined analyses using morphological and rRNA sequence data (Doyle et al., 1994) Nymphaeaceae, monocots, Piperales, and Aristolo- ound chiaceae to be successive sisters to the remaining angiosperms; the rest of the Magnoliidae formed a clade sister to the eudicots. Analyses of complete 18S rDNA sequences with greater taxon sampling than employed by Hamby and Zimmer (1992) Ilici- aceae, Schisandraceae, and Nymphaeaceae as the placed Amborellaceae, Austrobaileyaceae, earliest-branching angiosperms (D. Soltis et al., 1997). The analysis of a large atpB data set (Sa- volainen et al., 2000) also placed Amborellaceae, Nymphaeaceae, and Austrobaileyales as the earli- est-branching angiosperms. In all of these analyses, internal support (as measured by bootstrap values) at deep levels within the phylogenies was low (less than 50%). In addition, some of the above studies may have experienced long-branch attractions of highly divergent taxa or clades. By combining data from multiple genes and, in some cases, sampling numerous taxa, recent studies have greatly enhanced our understanding of rela- tionships among basal angiosperms. Recent studies not only agree on the same groups of early-branch- ing angiosperms, but also provide strong internal support for many relationships (Graham & Olm- stead, 2000; Mathews & Donoghue, 1999, 2000; Parkinson et al., 1999; Qiu et al., 1999, 2000; D. Soltis et al., 2000; P. Soltis et al., 1999a). investigations used sequences from a variety of These genes representing the chloroplast, mitochondrial, and nuclear genomes. In addition, different ap- proaches were employed, including parsimony, maximum likelihood, and a duplicate gene rooting strategy (Mathews & Donoghue, 1999, 2000). These analyses reveal that Amborellaceae, Nymphaeaceae (comprising Cabombaceae and Nymphaeaceae the two families constituting Nymphaeales 1999), and a clade of Illici- aceae, Schisandraceae, Austrobaileyaceae, and Tri- s. str, following Les et al., meniacae (hereafter referred to as Austrobaileyales) Volume 90, Number 2 2003 Zanis et al. 131 Perianth Evolution in Basal Angiosperms are successive sisters (with strong internal support) to the rest of the angiosperms. However, despite generally strong support for this branching order, an analysis employing RASA (Lyons-Weiler et al., 1996) indicated that Nymphaeaceae + Amborel- laceae may form a clade sister to all other angio- sperms (Barkman et al., 2000). In contrast, analy- ses of a large multigene data set found strong bootstrap support for Amborella as sister to all other angiosperms using both maximum parsimony and maximum likelihood methods, and a hypothesis test using parsimony rejected the Amborella + Nym- phaeaceae topology; however, a hypothesis test us- ing maximum likelihood could not reject the Am- borella + Nymphaeaceae topology (Zanis et al., 2002) Although recent studies have clarified the early branches of the angiosperm tree, relationships among the remaining basal angiosperms remain un- certain. Аз reviewed above, the remaining basal an- giosperms form a number of very strongly supported clades: Piperales, Winteraceae/Canellaceae (Canel- lales; referred to as Winterales by some), Magno- liales, monocots, Laurales, and Chloranthaceae (or- dinal APG 2003). Relationships among these clades were uncertain composition sensu in many previous nin om but are becoming clearer (e.g., Qiu et al., 1 2000; Zanis et al., 2002): a magnoliid clade wii Magnoliales + Laurales and Canellales + Piperales, although the positions of the monocots and Chloranthaceae relative to this magnoliid clade are not well supported. The place- ment of Ceratophyllaceae has varied among studies. Ceratophyllaceae were sister to all other angio- sperms in analyses based on rbcL sequences (Chase et al., 1993; Qiu et al., 1993; Savolainen et al., 000); however, in the three-gene, 567-taxon anal- ysis (D. Soltis et al., 2000; P. Soltis et al., 1999a), Ceratophyllaceae appeared as sister to the eudicots (with 53% jackknife support). Other studies have placed Ceratophyllaceae closer to the monocots (Qiu et al., 1999, 2000; Savolainen et al., Zanis et al., 2002) or as sister to Winteraceae (Par- kinson et al., 1999). A clear understanding of relationships among basal angiosperms has obvious major implications for interpreting the morphology of the early angio- sperms and subsequent patterns of floral evolution. Early hypotheses proposed that the first angio- sperms had large, Magnolia-like, flowers (Arber Parkin, 1907; Bessey, 1897, 1915; Cronquist, 1981, 1988; Takhtajan, 1969, 1997). However, Stebbins (1974) stressed that the concept of the earliest flower as large, strobiloid, and Magnolia- like was not consistent with the amount of special- ization that occurs within the Magnoliaceae. Steb- bins (1974) proposed that the earliest flowers were moderate in size. Using a combination of informa- tion from both extant and fossil Magnoliidae, En- dress (1987c) suggested that the earliest angio- but that the transition to unisexuality was relatively easy, the perianth was sperm was bisexual, undifferentiated and could be easily lost, and the number of floral parts was labile. Early phylogenetic studies focused attention on paleoherbs (Nymphaeaceae, Piperaceae, and Chlor- anthaceae) as possible first-branching extant angio- sperms (Donoghue & Doyle, 1989; Doyle et al., 1994; Hamby & Zimmer, 1992), suggesting that early flowers were small, with a trimerous perianth, and with few stamens and carpels. Recent topolo- gies (Doyle & Endress, 2000; Graham & Olmstead, 2000; Mathews & Donoghue, 1999, 2000; Parkin- son et al., 1999; Qiu et al., 1999, 2000; D. Soltis et al., 2000; P. Soltis et al., 1999a; Zanis et al., 2002) that place Amborella as sister to other angio- sperms suggest instead that the earliest flowers were small to moderate in size, with an undiffer- entiated perianth, stamens lacking a well-differen- tiated filament, and a gynoecium composed of one or more unilocular ovaries. The diverse array of early angiosperm fossils is consistent with this hy- pothesis (Crane, 1985; Crane et al., 1995; Friis et al., 1994, 1997, 2000). Patterns of evolution of specific floral characters in basal angiosperms have also been examined us- ing a phylogenetic framework (e.g., Hufford, 1996; Albert et al., 1998). Using the topologies of Dono- ghue and Doyle (1989) and Chase et al. (1993), Hufford (1996) found that the laminar stamen may have evolved independently several times in the Magnoliidae. Using the rbcL topology for angio- sperms (Chase et al., 1993), Albert et al. (1998 reconstructed perianth architecture to elucidate the evolution of the bipartite whorled perianth found in the eudicots. Albert et al. (1998) found that the single whorled perianth optimized as the ancestral character state for the angiosperms, based on a placement of Ceratophyllaceae as sister to all an- giosperms in the rbcL tree. To explore relationships among basal angio- sperms further, we added a sixth gene, 265 rDN to the data set of Qiu et al. (1999). A compartmen- talization approach (Donoghue, 1994; Mishler. 4; Rice et al., 1997) for these six genes was employed; this method facilitated the use of maxi- mum likelihood and more thorough parsimony methods of phylogenetic inference. Using our six- gene phylogenetic tree for basal angiosperms as a framework, we subsequently investigated the evo- 132 Annals of the Missouri Botanical Garden lution of perianth (1) phyllotaxis, (2) merosity (mer- ism), and (3) differentiation. MATERIALS AND METHODS 268 rDNA SEQUENCING The utility of entire 265 rDNA sequences for re- constructing angiosperm phylogeny has recently been demonstrated (Fan & Xiang, 2001; Fishbein 2001; Kuzoff et al., 1998). Although partial 268 rRNA and rDNA sequences have occasionally et al., been used to infer phylogeny (Buchheim & Chap- 199]; 1992; Ro et al., 1997), the great length of the gene (over 3200 bp) had, until recently, precluded the use of its entire man, Hamby & Zimmer, sequence for phylogeny reconstruction. The gene comprises both conserved core regions and more rapidly evolving expansion segments, thus enhanc- ing the utility of the gene for phylogenetic inference across several phylogenetic levels, from the angio- sperms as a whole (Kuzoff et al., clades such as Saxifragales (Fishbein et al., 2001), Cornales (Fan & Xiang, 2001), and Ranunculales (Kim et al., We generated and analyzed entire 26S rDNA se- quences for 44 angiosperms representing Magnoliales, submitted). Laurales, Chloranthales, Piperaceae, Aristolochi- aceae, Nymphaeaceae, Canellales, Austrobaileyales, monocots, and eudicots. The following gymnosperms served as outgroups: Ginkgo, Gnetum, Ephedra, and Larix. All species for which 268 rDNA sequences were included in this analysis are listed in Table 1 э. voucher information and GenBank numbers are also provided. We attempted to use the same species, and in many cases the same DNAs, used by Qiu et al. (1993) and D. Soltis et al. (1997, 2000). Familial and ordinal circumscriptions mostly follow those given in АРС II (200 шш аа and sequencing of 26S rDNA gen- erally followed the methods of Kuzoff et al. (1998). For PCR amplification, we used either ITS3 or N- пс2651 as the forward primer and 3331rev for the reverse primer. Їп some cases we amplified 26S rDNA in two portions. The 5' half of the gene was amplified using either ITS3 or N-nc2681 as the for- ward PCR primer and 1839rev as the reverse prim- The 3' half of 268 rDNA was amplified using N-nc2657 as the forward primer and 3331rev as the reverse primer. For the sequencing of 26S rDNA, the following primers were used: N-nc20S1, N-nc2683, N-nc2685, №пс2657, N-ne26S9, N- nc26811, №-пс26513, N-nc26814, 268rev, 641rev, 950rev, 1229rev, 1499rev, 1839rev, 2134rev, 2782геү, 3058rev, 333lrev. (Kuzoff et a 1998) and 1998) to smaller The vast majority of 265 rDNA sequence is eas- ily aligned by eye; however, several small portions of some of the expansion segments were more dif- ficult to align visually. Clustal X (Thompson et al., 1997), with gap opening set to 10 and gap exten- sion set to 0.2, was therefore used to obtain an initial 268 rDNA alignment, which was further re- fined by eye. PHYLOGENETIC ANALYSES USING COMPARTMENTALIZATION Mishler (1994) proposed compartmentalization as a method to reduce a large data set to a smaller, more manageable size, to decrease the effect of "spurious homoplasy,” and to maximize the amount of information used in phylogenetic analyses by al- lowing different data sets to be used for the local and global analyses. This approach also facilitates the application of maximum likelihood to the com- partmentalized data, an approach that could not be employed with the large, more complete data set. In brief, an initial global analysis is performed on the large data set. Well-supported clades (compart- ments) are identified from the global analysis using bootstrap, jackknife, or decay values (Bremer, 1988; Farris et al., 1996; 1985). Smaller, focused analyses are then performed to es- Felsenstein, tablish relationships within each of the compart- ments. Upon completion of the smaller analyses, relationships within compartments сап be con- strained and relationships among compartments can be inferred, or a hypothetical ancestor for each compartment can be reconstructed using dere or maximum likelihood methods (Yang et al., and used to infer relationships among s ments (Donoghue, 1994; Mishler, 1994; Rice et al., Using the five-gene data set of Qiu et al. (1999), we conducted a global parsimony analysis with 100 replicates of random taxon addition to search for multiple islands of most parsimonious trees (Mad- dison, 1991) and TBR branch swapping. Internal support was estimated using the bootstrap (Felsen- stein, 1985) with LOO replicates with 100 random MULPARS, and TBR branch swapping. Each of the following compart- taxon addition replicates, ments appeared well supported in our global par- simony analysis, receiving greater than 90% boot- 1999; Parkinson 1999a): Nymphaeaceae, strap support (see also Qiu et al., 1999; P. Soltis et al., Austrobaileyales, Magnoliales, Laurales, Piperales, et al., Canellales, Chloranthaceae, monocots, and eudi- cots. For each compartment we then reconstructed a hypothetical ancestral sequence for each of the Volume 90, Number 2 2003 Zanis et a 133 Perianth Evolution in Basal Angiosperms Table 1. Voucher information and GenBank Accession numbers for the 265 rDNA sequences used in this study. GenBank Accession Family Species Voucher number Acoraceae Acorus gramineus Aiton Kuzoff (1998) AF036490 Amborellaceae Amborella trichopoda Baill. Plunkett, G AY095449 Annonaceae Asimina triloba (L.) Dunal. Qiu 15 AY095451 Araceae Spathiphyllum wallisii Hort. Chase 201 (NCU) 095473 Aristolochiaceae Aristolochia macrophylla Lam. Qiu 91019 AY095450 Lactoridaceae Lactoris fernandeziana Phil. Stuessy et al. 11784 (OS) AY 095463 Asphodelaceae Bulbine succulenta Compton UCI Arb. 7174 AY095471 Austrobaileyaceae Austrobaileya scandens C. T. White Qiu 90030 0954: Berberidopsidaceae Berberidopsis corallina Hook. f. Chase 555 (K) AF389242 aceaee Buxus se ay "rens L. Chase 203 (NCU) AF389244 Cabombaceae Cabomba Qiu 97027 AY095453 Calycanthaceae Calyc part uid ntalis Hook. & Arn. ult. WSU AY095454 Canellaceae Ceratophyllaceae Chloranthaceae Chloranthaceae Eupteleaceae Fumariaceae Ginkgoaceae Gnetaceae Ephedraceae Gomortegaceae Hamamelidaceae Hernandiaceae Himantandraceae Magnoliaceae Magnoliaceae Menispermaceae Monimiaceae Nelumbonaceae Nymphaeaceae Pinaceae Piperaceae Platanaceae Poaceae Ranunculaceae Sabiaceae Saururaceae Taccaceae Tofieldiaceae Trimeniaceae Trochodendraceae Winteraceae Winteraceae Canella winterana (L.) Gaertn. Ceratophyllum demersum L. Chloranthus multistachys Pei Hedyosmum bonplandianum L. _ polyandra Sieb. & Zucc. entra eximia Torrey с. iloba L. Gnetum gnen m Ephedra odd Gomortega Кеш ( "ira 1. M. Johnston Hamamelis virginian Hernandia nymphae if lia Galbulimima belgraveana a Muell.) Sprague Magnolia denudata Des Liriodendron chinense Hemsl. ) Sarg. Tinospora caffra Miers Peumus boldus Molina NUNG lutea (Willd.) Pers. Platanus occidentalis L. Oryza sat Ranunc ulus Ен Milne-Redhead & Turrill Sabia swinhoei Hemsl. Saururus cernuus L Tacca chantieri André Pleea tenuifolia Michx Trimenia moorei (Oliv.) Philipson Trochodendron 44 ys Siebold . Forster ч. — Drimys winteri Tasmannia spits d K. ‚экн! 920010 Endress 97-102 Qiu 9001 (NCU) Reznicek 9756 (MICH) Cult. WSU Kuzoff et al. (1998) Kuzoff et al. (1998) eda Hoot 910 Univ. Zurich Bot. Gard. P. H. Weston 929 S. ge 1010 (NPRI) Qiu d 2131 (NGB) Strybing Arb. Hoot 9212 (UWM) Qiu 91029 Cult. WSU Qiu 91048 Qiu P90005 (NCU) мел, е! pe (1985) Chase K) Wagner hn (HAST) Suh 128 (US) Chase 175 (NCU) Chase 152 (NCU) ANBG 701680 Qiu 90026 (NCU) Nickrent 3013 (SIU) Qiu 90032 AY095455 AY 095456 AF: 389249 AF389262 AY095475 AF036488 AF036489 AY095460 AF036495 AY095402 AY095459 AY( is AF jedes AF389259 —— 095476 pos 5467 AF274602 M11585 AF389269 AF389272 AY095468 AY095474 AY095472 AY095470 AF274070 AF036491 AY095469 five genes analyzed by Qiu et al. (1999) and our 26S rDNA data set using a maximum likelihood approach as implemented in Phylogenetic Analysis using Maximum Likelihood (PAML) (Yang, 1997); we employed a general time-reversible model with rate heterogeneity as our model of molecular evo- lution. In addition to the ancestral sequence for each of the compartments noted above, our final compartmentalized data set also included Cerato- phyllum and Amborella, critical individual taxa that were not members of any compartment. Maximum likelihood represents an alternative to 134 Annals of the Missouri Botanical Garden the parsimony method of reconstructing ancestral sequences (Yang et al., 1995), employing informa- tion from a model of substitution and branch length estimates. Comparing both parsimony and maxi- mum likelihood methods for inferring ancestral se- quences, Yang et al. (1995) revealed that when se- quences were from closely related species, both maximum likelihood and parsimony methods were accurate. Thus, when dealing with invariant sites, or less variable sites, both parsimony and maximum likelihood methods often produced identical re- sults, and both had high levels of accuracy in re- constructing ancestral sequences. In contrast, both methods may be unreliable at reconstructing an- cestral sequences for highly variable sites (Yang et al., 5). All of the hypothetical ancestral sequences for the six genes analyzed (five from Qiu et al., 1999, plus the 265 rDNA data set presented here) were ultimately combined for subsequent parsimony and maximum likelihood analyses; we will hereafter re- fer to this as the compartmentalized data set. Both parsimony and maximum likelihood estimation methods were performed using PAUP* 4.0 (Swof- ford, 1998). Parsimony analyses of the compart- mentalized data set were conducted using the branch-and-bound search strategy with the initial upper bound computed via stepwise addition with furthest sequence addition. Maximum likelihood parameter values were es- timated from the single most parsimonious tree ob- tained from the parsimony analysis described. We used the general time-reversible model of molecu- lar evolution, accounting for invariant sites and rate 1996). model of molecular evolution was done using MO- DELTEST 3.0 (Posada & Crandall, 1998). Fifty-six models were compared using the Akaike informa- tion criterion (AIC). method does not require models to be nested and heterogeneity (Swofford et al., The AIC model comparsion selects models for good fit but penalizes models for unnec essary 1998). The parsimonious tree was 39743.36269, the proportion (Posada & Crandall, -İn likelihood score for the single most parameters of invariant sites was estimated to be 0.516797, and the gamma shape parameter was estimated to be 0.744558. These estimates along with estimates for the general time-reversible model and base fre- quencies were used in maximum likelihood analy- sis using heuristic searches similar to the parsi- mony analyses. Internal support for both maximum likelihood and parsimony analyses was estimated using bootstrap analysis with 100 replicates with 100 random taxon additions per replicate, with TBR branch swapping. The choice of CHARACTER ANALYSES Perianth character states were mapped assuming equally weighted parsimony using MacClade 3.0 (Maddison & Maddison, 1992). We employed the “all most parsimonious states” trace option. A syn- thetic phylogenetic tree was used to expand the tree in regions that were not sampled thoroughly in the analysis of basal angiosperms. That portion of the topology dealing with relationships among the basal angiosperms was generated here (see below); eu- dicot relationships are those from the jackknife consensus tree of P. Soltis et al. (1999a) and D. Soltis et al. (2000). Phylogenetic relationships de- picted within. Nymphaeaceae follow Les et al. (1999). Relationships depicted within the monocots are a portion of one of the shortest trees obtained from a combined morphological and rbcL data set analyzed by Chase et al. (1995). CHARACTER DEFINITIONS Because definitions for the perianth vary (En- dress, 1994a; Greyson, 1994; Weberling. 1989), we refer to the perianth as the sterile structures above the bract(s) if present, and immediately below the spore-producing organs in a flower. The organs of a perianth (i.e., sepals and petals) share similarities with bracts, a scale-like leaf subtending the organs of a flower, as well as staminodes, structures that represent sterilized stamens. For example, sepals have many features in common with bracts: both are green and lack clear anatomical differentiation into palisade and spongy parenchyma. Petals and staminodes share several characteristics as wel (Esau, 1965). Both are laminar in shape and usu- ally colorful (due to chromoplasts or pigments) (Esau, 1965). Descriptions of the perianth charac- ters and states followed previous workers' interpre- tations for each of the groups studied and are pre- sented in the Appendix. Perianth phyllotaxis is the arrangement of peri- anth organs (i.e., sepals and petals) on an axis (En- dress, 1994a). The states we coded are: absent (i.e., there is no perianth), a single whorl, two whorls, multiple whorls, and spiral. We also explored al- ternative character-state coding by reducing the different forms of whorled phyllotaxis (i.e., single whorl, two whorls, multiple whorls) into a single character state, whorled. Phyllotaxis is difficult to divide into states because it is labile in many groups (Doust, 2000, 2001; Endress, 1987b). Fur- thermore, the distinction between spiral and whorled is not always clear. That is, spiral and whorled may not be fundamentally different forms of phyllotaxis. Spiral and whorled perianth phyllo- Volume 90, Number 2 2003 Zanis et al. 135 Perianth Evolution in Basal Angiosperms taxis likely represent a continuous (quantitative) character rather than a discrete (qualitative) char- acter (Stevens, 1991). Perianths with whorled phyl- lotaxis appear to have organs arranged in the same plane. These organs appear equally spaced and also appear to have been initiated simultaneously (Endress, 1987b). However, ontogenetic investiga- tions have indicated that in some cases floral organs typically identified as being whorled actually result m the presence of long plastochrons [the time interval between the initiation of two consecutive organ primordia; (Endress, 1994a; Tucker, 1960)| inserted between a number of small plastochrons [e.g.. Huber (1980) for Solanaceae; Erbar & Leins (1985) for Apiaceae]. Thus, both spiral and whorled phyllotaxis may have the organs developing in a spiral sequence (Endress, 1987b, 2001). Perianth merosity refers to the number of peri- anth parts present within each whorl. We recog- nized five character states: three organs per whorl, two or four organs per whorl, five organs per whorl, or an indeterminate number of organs per whorl. Indeterminate refers to a range with no fixed num- ber of parts; typically, these are perianths with nu- merous spirally arranged parts Differentiated perianths are those having an out- er whorl (or series) that is clearly differentiated from the inner whorl(s) (series). A differentiated perianth is commonly referred to as being com- posed of sepals and petals. In contrast, an undif- ferentiated perianth is one that lacks clear distinc- tion between the outer and inner whorls; these have been traditionally recognized as tepals (Cronquist, 1988; Takhtajan, 1997). Albert et al. (1998) sug- gested that there must be at least two whorls pre- sent for unambiguous interpretation of sepals and petals, that is, an outer whorl that is clearly distin- guishable from the inner whorl. Often, when only a single-whorled perianth is present, it is referred to as a calyx as a matter of convention (Cronquist, 1981, 1988). For example, those Aristolochiaceae characterized by a single-whorled perianth are re- ferred to as having sepals (Cronquist, 1988; Takh- 1997; Tucker & Douglas, 1996); develop- mental data for Asarum (Leins & Erbar, 1985) a Aristolochia (González & Stevenson, 2000) support tajan, this interpretation Perianth differentiation can often be difficult to score in those taxa having a perianth of a single whorl. We used the following character states in our analysis of perianth differentiation: undifferentiated (no clear distinction into outer and inner whorls), differentiated (clear distinction between inner and outer whorls), single whorl, and absent. CHARACTER-STATE CAVEATS Several studies have considered the treatment of polymorphic characters (Mabee & Humphries, 1993; Maddison, 1993; Nixon & Davis, 1991; Wiens, 1999). There are three options: (1) code the polymorphic state as “missing,” (2) code polymor- phic taxa with a putative ancestral state, by recon- structing the ancestral state from a phylogeny of the taxa in question, and (3) code the polymorphic taxa as polymorphic. We coded several taxa as poly- morphic. For example, Canellaceae are coded as trimerous, tetramerous, or pentamerous (Wilson, 66). We tested the robustness of our results by using alternative codings (Donoghue & Ackerly, 1996; Weller et al 9 In addition to uncertainty in scoring taxa, due either to difficulty in interpreting structures or to polymorphism, there is uncertainty in the tree to- E and the impact that alternative tree topolo- gies may have on ancestral character-state recon- E (Donoghue & Ackerly, 1996). We examined the sensitivity of our character-state re- constructions by examining reconstructions on al- ternative trees. For example, we examined char- acter-state reconstructions on trees that placed Chloranthaceae sister to the clade containing the monocots and Ceratophyllum, and on trees that had Amborella and Nymphaeaceae as a clade sister to the rest of the angiosperms. We omitted outgroups from our analysis of peri- anth evolution. Previous morphological analyses that included both extant and fossil seed plants in- dicated that the angiosperms were part of an an- thophyte clade containing the Gnetales, Bennetti- tales, and Pentoxylon. This anthophyte clade was sister to the seed fern Caytonia (Donoghue & Doyle, 2000; eyes 1996, 1998). Recent analyses e.g., Bowe et al., 2000; Chaw et al., 2000; P. Soltis et al., 1999b) indicate that gymnosperms are mono- phyletic, with the Gnetales nested within the co- nifers or as the sister clade to the conifers. Fur- thermore, it is difficult to assess the homology of the angiosperm perianth with perianth-like struc- tures in the gymnosperms (e.g., Crane, 1988; Huf- ford, 1996). Nevertheless, we explored the effect of assuming different plesiomorphic states for the an- giosperms on our inferences of perianth evolution. RESULTS ANALYSES PHYLOGENETIC The compartmentalized data set had 14 hypo- thetical taxonomic units and three additional taxa with aligned sequences of 12,215 bp of sequence 136 Annals of the Missouri Botanical Garden data per taxon. Of these sites, 9658 were constant, 1239 were parsimony-uninformative, and 1318 were parsimony-informative. Parsimony analysis of the compartmentalized data resulted in a single — 0.741 most Tasa tree of 4218 steps (CI RI = 0 . The topology obtained from the and maximum ee analysis of the compartmen- talized data set had а -ln likelihood score of 39,743.36269. Both parsimony and maximum like- lihood analyses of the compartmentalized data re- sulted in identical trees that are also entirely con- sistent with the shortest tree of Qiu et al. (1999), as well as the tree we obtained in our first global analysis. The same early-diverging angiosperms are present in the same order: Amborella, Nymphae- aceae, and an ана niu e and: baileyaceae/Tri ceae. Bootstrap support for these relationships is 100% In recent studies of basal angiosperm phylogeny e.g., Qiu et al., 1999, 2000; Zanis et al., 2002), bootstrap or jackknife support 7096 has been observed for relationships among the major remain- = ing clades of basal angiosperms (i.e., Piperales, Ca- nellales, Magnoliales, Laurales, and Chloranthaceae). monocots, However, some relationships, such as the sister group of the monocots or the position of the Chloranthaceae, remain uncertain. In the compartmentalized analysis, bootstrap sup- port for relationships among the compartments (clades) is much higher (> 90%) for most nodes than the internal support found in both our global de and in the recent epis parsimony alyses (Qiu et al., 1999, 2000; Zanis et al., 2002), Canellales and Piperales form a strongly supported (100%) sister group that is, in turn, sister to a well-supported (100%) Laurales/Magnoliales clade. This magnoliid clade is also strongly sup- ported (100%). Ceratophyllaceae are strongly sup- ported (100%) as sister to the monocots. There is also strong support (86%) for the placement of the monocot/Ceratophyllaceae c clade as sister to a large Pi- The only relationship that remains uncertain is the placement of Chloranthaceae within this clade. Al- though Chloranthaceae appear as the sister group clade comprising Chl . Canellales, — perales, Magnoliales, Laurales, and eudicots. to the large clade of eudicots and magnoliids (Pi- perales, Canellales, Laurales, and Magnoliales), this placement receives bootstrap support < 50%. The complete 26S rDNA sequences have made an important overall contribution to the increase in internal support. For example, in a bootstrap anal- ysis of the compartmentalized data set using par- simony with 265 rDNA sequences removed, sup- port for the Ceratophyllaceae + monocot clade is only 56% (Fig. 2). Phylogenetic analyses of the 26S rDNA data alone (tree not shown) reveal a topology highly similar to, although less resolved than, that obtained for basal angiosperms based on three (D. 2000; P. Soltis et al., 1999a), five (Qiu 1999, 2000), or more genes (Zanis et al., Soltis et al., et al., 2002) CHARACTER ANALYSES All perianth characters were mapped onto the synthetic Table 2 lists minimum and maximum number of changes be- tree described above. the tween states for each of the characters examined. Much of the tree is equivocal when examining peri- anth phyllotaxis as a five-state character (Fig. 3). By combining all the whorled character states (i.e., single whorl, two whorls, and multiple whorls) into a single state, less of the reconstruction is equivocal (Fig. 4). Each node of the Amborellaceae/Nym- phaeaceae/Austrobaileyales grade is equivocal with whorled or spiral perianth being equally parsimo- nious states for the base of the angiosperms. How- ever, the ancestral state for the large clade contain- the Ceratophyllaceae, clade, Chloranthaceae, - [>] magnoliid monocots, and eudicots is a whorled perianth (Fig. 4). The mapping of number of perianth parts onto our topology indicates that it is equally parsimonious to have an ancestral peri- anth with parts in threes or indeterminate (Fig. 5). The ancestral condition for the large clade contain- ing the magnoliid clade, Chloranthaceae, Cerato- phyllaceae, monocots, and eudicots is a trimerous perianth (Fig. 5) Mapping of perianth differentiation results in much of the tree being equivocal with multiple most parsimonious scenarios for the evolution of perianth differentiation (Fig. 6). Clearly a differ- entiated perianth evolved multiple times in the an- giosperms. Our analyses indicate that there may have been as few as two to as many as six changes to a differentiated perianth from an undifferentiated perianth (Table 2). Futhermore, some differentiated perianths may have evolved from spirally arranged perianths, whereas other differentiated perianths, such as in Saruma (Aristolochiaceae), may have evolved from ancestors that had a perianth com- posed of a single whorl. We also conducted analyses to assess the sensi- tivity of our reconstructions to alternative topolo- gies. Our analysis of perianth phyllotaxis and num- ber of perianth parts is robust (little effect on ancestral reconstructions) to alternative topologies (data not shown). However, the analysis of perianth Volume 90, Number 2 Zanis et al. 2003 137 Perianth Evolution in Basal Angiosperms agnolia 53 L бе xlendron 7 „пса 58 Ө 64 Sipe 100 86, А therosperm a 8 100 Daphnandra 2 Joryphora Gomortega 58 За elk ^ ва pelea 71 94 Poda Шит 71 09 ора dyos 100 С C eraophvihan demersum submersum 100 Рета 1 -— 99 Pinus 0 0 Gnetum Welwitschia Metasequo ME 100 А S Magnoliales Laurales Winterales Piperales eudicots Chloranthaceae Ceratophyllaceae monocots Austrobaileyales Nymphaeales Amborella outgroups Figure l. Single most parsimonious tree from global analysis. Numbers above the branches indicate bootstrap lues differentiation is sensitive to alternative topologies. Laurales is reconstructed as having a differentiated For example, if Chloranthaceae are sister to the — perianth (data not shown). In addition to examining clade containing the monocots and Ceratophyllum, the sensitivity of our character analyses to alter- then the ancestral state for the clade containing the native topologies, we examined the effect of coding eudicots, Canellales, Piperales, Magnoliales, and terminal taxa as monomorphic versus polymorphic. 138 Annals of the Missouri Botanical Garden 100/99 100/100 Austrobaileyales Magnoliales Laurales Wihterales Piperales iin ; Cudicots w/o Ranunculales Ranunculales Chloranthaceae 100/56 monocots Ceratophyllaceae Nymphaeales Amborella Gnetales 98/99 Ginkgo ——— 0.01 substitutions/site Figure 2. bootstrap values. First value : support obtained from five genes (without 268 rDNA). Br Pinus/Larix Metasequioa/Podocarpus Single most parsimonious tree from compartmentalization analyses. Numbers above branches indicate above branch indicates gne support e six genes; second value indicates bootstrap anch lengths calculated using the general time-reversible model of molecular evolution accounting for invariant sites and rate he buic The results indicate that our analyses are not af- fected by changes from polymorphic to a fixed char- acter state. Lastly, assuming different plesiomorph- ic states for the angiosperms does affect our inference of ancestral character-state. reconstruc- tion. Thus, if we assume that the trimerous perianth is the plesiomorphic state for the angiosperms, then the trimerous perianth maps as the ancestral state for the deepest nodes of the tree. However, assum- ing that the plesiomorphic state was a perianth with an indeterminate number of parts does not differ from what is shown in Figure 4. If we assume that lacking a perianth is the plesiomorphic state for the angiosperms, we then find that there have been multiple origins of both a differentiated and undif- ferentiated perianth. Assuming other plesiomorphic character states for perianth differentiation had lit- tle impact on our inferences of perianth evolution. DISCUSSION COMPARTMENTALIZATION The compartmentalization approach allowed us to reduce the number of taxa from 105 in the Qiu et al. (1999) data set to 14 hypothetical (ancestral) taxonomic units and three operational taxonomic unils so that we could employ both a more thorough parsimony analysis and the maximum likelihood method. The topology that resulted from the com- partmentalized analysis is entirely consistent with respect to the phylogenetic placement of basal an- giosperms with the trees obtained in both our initial global analysis and the five-gene analysis of Qiu et al. (1999) (Fig. 1). Identical topologies were ob- tained using the entire data set of Qiu et al. (1999) or the hypothetical ancestral sequences using both maximum likelihood and maximum parsimony Volume 90, Number 2 Zanis et al. 2003 Perianth Evolution in Basal Angiosperms MAGNOLIALES _ LAURALES CAN PIPERALES EUDICOTS CHL MONOCOTS AUS NYMPHAEACEAE v о g 9 > 8 - РА КЕ 92.989 85 ت‎ 8 е 8828922 283 о E > 8 5 2958528 Ы 222937 2 S 8 $23 38 8 3.2 8 S35889ESEREBSÓIS.f] © 8 8232222258 $8 * E: = ЕЕ Ed SSpS Sages Rw 885s ES = Ы ч ТЕНЕ ШЧО ЕН ad EE THESSSEHEHISTEREREHHEHHHEEHEHHEHEHEETLHTHERHHEERHEEETHEHIE RS SERES SESRESS KY ا‎ CUE ay) we Wy 4 bd S N » р: A Ф xX single whorl ` ШШ multiple whorls ШШ absent uncertain equivocal [eers] E 3. Synthetic cladogram showing parsimony reconstruction of perianth phyllotaxis using MacClade 3.03. Five character states are employed with the whorled phyllotaxis state divided into three separate character states: two whorls, single whorl, and multiple whorls. Potomog/Zoster = Potomogetonaceae/Zosteraceae; CAN = Canellales; CHL = Chlor- anthaceae; AUS = Austrobaileyales. methods on the final compartmentalized data set provement in resolution and support, but the use of (Fig. 2). Nevertheless, further work is needed to additional data. understand the limits and biases associated with ur six-gene topology reveals the same order of compartmentalization analyses. early-branching angiosperms observed in other re- Phylogenetic analyses of the 265 rDNA se- сеп! studies (Mathews & Donoghue, 1999; Parkin- quence data alone yielded topologies with the same son et al., 1999; Qiu et al., 1999, 2000; D. Soltis early-branching angiosperms, as well as the same et al., 2000; P. Soltis et al., 1999a; Zanis et al., major clades of basal angiosperms (i.e. Piperales, 2002): Amborellaceae, followed by Nymphaeaceae, Chloranthales, Magnoliales, Laurales, Canellales, and the Austrobaileyales clade (Austrobaileyaceae/ monocots) observed in analyses of three-gene (D. Тп i Пет /Schisand ) Soltis et al., 2000; P. Soltis et al., 1999a), five-gene We found high levels of bootstrap support (> (Qiu et al., 1999), and larger (Zanis et al., 2002) 90%) in the compartmentalized analysis for all data sets. The addition of 268 rDNA sequences to nodes, except the node supporting Chloranthaceae the five-gene data set played an important role, in- as sister group to the clade of Magnoliales, Laura- creasing bootstrap support for relationships. Forex- les, Canellales, Piperales, Aristolochiales, and eu- ample, the addition of 26S rDNA sequences pro- — dicots. Our analyses show strong support for the vided high bootstrap support for the placement of sister relationship of Canellales and Piperales the monocots among the basal angiosperm lineages (100%); Laurales are sister to Magnoliales (100%); and the placement of Ceratophyllum as the sister Laurales/Magnoliales and Canellales/Piperales also group to the monocots (Fig. 2). However, compart- rm a well-supported magnoliid clade (100%); mentalization of the three-gene data set (D. Soltis Ceratophyllaceae are sister to the monocots et al., 2000; P. Soltis et al., 1999a) did not lead to 0). greater resolution ог support among basal angio- sperms (P. Soltis et al., 2000). Thus, it is not the The increase in bootstrap support between the analytical approach alone that accounts for the im- initial global analysis (as well as the broad analysis of Qiu et al., 1999) and the compartmentalized Annals of the 140 Missouri Botanical Garden NYMPHAEACEAE AUS MONOCOTS Djjo40qui pa 3 26 3 ler ошәши Lo + оздор уа E U llotaxis unordered o 83 зу ES 2-3 b: az | q bil |) U (Í | U ps 9d 1 1 j () 1 1 1 Phy C spiral BB absent FE uncertain ES equivocal 121507 /80шоюд Y NEAN ^N oeaoeaoopoui&Omé bh, SF ES à SAR Е, с N DE TY > x E 4 CHL EUDICOTS CAN PIPERALES LAURALES MAGNOLIALES шпдудопмәг) I шпшоХрәң e. А. К. 3) D 5 Synthetic « 4. igure + в 5 E ч n I { Clade 3.0 single whorl, and multiple whorl charac whorled. Potomog/Zoster 4 1 onstruction of perianth phyllotaxis using Мас Austrobaileyales. . with two whorls Ü I |! ladogram showing parsimony re Three character states are employed (spiral, whorled, absent states (from Fi CAN sleraceae eae/Zo = Potomogetonac - ) S = E hloranthaceae; AU "d into a single character state C 3) combine HL = Ig. . anellales; € " ( оәгодш уо МҮМРНАЕАСЕАЕ AUS ршәшиЈе оќәродол5п pa f dimerous or tetramerous ШП ШЕШ absent 7 ў Ё g 2 a Merosity unordered C indeterminate trimerous pentamerous uncertain ES equivocal 4 o EUDICOTS $1012UTa CAN PIPERALES se»orjeundsojouya әрәзеЗәшошогус MAGNOLIALES LAURALES a3eao?uouuyo ag a Ф e c =| b сс 7 =< Ф e x с ed “Ye Sy ro سا‎ | ~ al о < Ж Š il ep Bua n € = — ar . Q a `0 =? n © os а = -= Ф 5 а = = z - Ы 33 © = с ar -= O 9 & | 5 et لے‎ oT -L ME wer D^ =n р Ф Ed up rs = Ф A e = E of = 9 ous ш М > |! n [= © 2, ZI g~ «e Б>] Synthetic cladogram showing p = tomog/Zoster = Potomogetonaceae/Zosteraceae igure ` Volume 90, Number 2 3 Zanis et al. 141 Perianth Evolution in Basal Angiosperms MAGNOLIALES LAURALES — CAN PIPERALES EUDICOTS CHL MONOCOTS AUS NYMPHAEACEAE E] © 2 33 2 3 9 5 2 o S LE s&s v z РЕ 55 Б 9 882228 ugg v 5 sf 8 оо 5 5 боб = 8358353227 32 S Б ЕСЕРГЕ 4 8 EPs eg {з ЕЕН гк SHEL АЧЫ =. 3 в SSBESSSTSESs ong og 5 g $22229582389 3 Ё 5 SES Nossa 8 3 seag S FE E EEE EEE ESSE SS ERS OE SESSIE SIEI gE SES SES SES SBE SESE TS ELETE Ss. Я с Ж = THEHSHHTHIHIHHRSHHHBHHIHBHHHIHTHEHHIBHHISHDPBE p = ч = t я — 5 ў c р b z E" «шат оа солна «ада EG GG FREEEEEEEPEEFEERFEEEPEEEEEERECCETEEEEEHETEE 4 WV ” as 2 | : (P / X 27 : p V Ne. Figure 6. Potomog/Zoster — Potomogetonaceae/Zosteraceae; CAN — ales analysis is most likely due to both the minimization of homoplasy as a result of the construction of the hypothetical ancestral sequences (Mishler, 1 and, as noted, the addition of 265 rDNA sequences. However, some areas of the phylogeny remain poor- ly resolved and supported despite the use of data for six genes and approximately 12.000 base pairs of sequence data per taxon. For example, the node supporting the placement of the Chloranthaceae is weakly supported in both the Qiu et al. (1999) anal- ysis and our compartmentalized analyses of the six- gene data set. The low bootstrap support for this node is likely the result of short internal branch lengths. For example, the branch leading to the Chloranthaceae and all remaining angiosperms is only 17 steps with ACCTRAN optimization. The uncertainty surrounding the placement of the Chloranthaceae supports Mathews and Donoghue's (1999) suggestion that after the appearance of the Amborellaceae, Nymphaeaceae, and Austrobailey- ales lineages, there was a rapid radiation of the remaining basal angiosperm lineages. The hypoth- esis of a rapid radiation is plausible and consistent with long-standing hypotheses on the origin and di- vergence of the angiosperms especially given that fossils of Chloranthaceae, Laurales, Magnoliales, Differentiation unordered undifferentiated different outer vs. inner whorl E single whorl ШШ absent E22 uncertain EZ equivocal A synthetic cladogram showing parsimony reconstruction of perianth differentiation using MacClade 3.03. — Chloranthaceae; AUS — Austrobailey- Canellales; CHL and eudicots are all common by the mid Cretaceous (Crane et al., 1995; Friis et al., 1997). CHARACTER ANALYSES The origin of the perianth is uncertain, and var- ious hypotheses have been proposed. Weberling (1989) outlined four views regarding the origin of the perianth: (1) the whole perianth is derived from foliage leaves (following Glück, 1919; Prantl, 1887; Velenovsky, 1910); (2) the perianth is derived from sporophyll structures (Celakovsky, 1900; Nemejc, 1956); (3) the calyx is derived from bracts and the corolla is derived from modified stamens (Drude, 1887; Goethe, 1790; Naegeli, 1884); and (4) the calyx is derived from bracts and the corolla has a different origin, neither calyx nor staminal (Weber- ling, 1989). Features of the perianth that have at- tracted attention are perianth phyllotaxis (Endress, 1987c), the number of parts (Kubitzki, 1997), dif- ferentiation of the perianth (Albert et al., 1998), and, recently, patterns of symmetry in asterids (Donoghue et al., 1998; Ree & Donoghue, 1 We have examined here the evolution of phyllotax- is, number of parts (merosity), and differentiation, characters that are not necessarily independent. 142 Annals of the Missouri Botanical Garden Table 2. Parsimony based character analyses: The minimum and maximum number of changes between states for each of the perianth characters examined using the chart changes option in MacClade 3.03 (Maddison & Maddison, 1992 Perianth phyllotaxis: Five states 0: 'Iwo Single Multiple Spiral whorls whorl whorls Absent From: Spiral — 0—3 0-2 0—2 0—1 Two whorls 1-4 04 0-3 25 Single who 0-2 14 — 0—2 0—3 Multiple whorls 0-1 0—3 0-2 — 0-2 Absent 0-1 0-3 0-3 0-3 —— Perianth phyllotaxis: Two states To: Spiral Whorled Absent From: Spiral — 0-2 0 Whorled 3-5 — 6 Absent 0 0 — Number of perianth parts To: Indeter- minate Threes Twos Fives Fours Absent From: Indeterminate — 04 0 0-1 0 0 Threes 1-5 — 0-1 0 1-2 5 Twos 0 0 1 0 l Fives 0 0 0 — 0 0 Fours 1 О 1 О — 0 Absent 0 0 0 0 0 v Perianth differentiation To: Undiffer- Single entiated Differentiated whorl Absent From: Undifferentiated — 2-6 0-1 24 Differentiated 2-6 — 0—4 1-4 T 0-1 0—4. — 0-2 Absent 0-2 0-2 0-2 — For example, Kubitzki (1987) discussed the origin of the trimerous perianth from a spiral (indetermi- nate) ancestor. Thus, merosity and phyllotaxis are sometimes intertwined. Moreover, perianth differ- entiation is dependent on having more than a sin- gle-whorled perianth (Albert et al., 1998). Our analysis indicates that when perianth phyl- lotaxis is considered a five-state character (spiral, two whorls, single whorl, multiple whorls, absent), ancestral reconstruction is ambiguous for many of the internal branches (Fig. 3). However, the anal- ysis of perianth phyllotaxis as a three-state char- acter (absent, spiral, whorled) indicates that while the ancestral state for the angiosperms is ambigu- ous, the ancestral state for the large clade contain- ing the magnoliid clade, Chloranthaceae, Cerato- phyllaceae, monocots, and eudicots is a perianth based on a whorled ground plan (Fig. 4). However, we have treated perianth phyllotaxis as a discrete character rather than a continuous character (Ste- vens, 1991), and interpretation of our results should reflect this treatment. For example, the transition from a spiral to a whorled arrangement brings floral parts closer together through the gradual shortening of the floral axis (Takhtajan, 1980). Perianth merosity has been of particular interest with regard to the origin of the trimerous flower (Kubitzki, 1987). Our analysis of perianth merosity indicates that the ancestral condition for the angio- sperms is equivocal. However, the ancestral char- Volume 90, Number 2 2003 Zanis et al. Perianth Evolution in Basal Angiosperms 143 acter state for the large clade containing the mag- noliid clade, Chloranthaceae, Ceratophyllaceae, monocots, and eudicots is a trimerous perianth (Fig. 5). There may have been as many as four origins of the trimerous perianth from a perianth with an indeterminate number of parts, based on our to- pology (Fig. 5). Also, the indeterminate perianth may have originated as few as two or as many as five times (Fig. 5). Kubitzki (1987) suggested that the amount of morphological change required in the transition from a spiral (indeterminate) perianth to a trimerous perianth could be small; moreover, he suggested that the trimerous condition is a mor- phological constraint and that the return from a trimerous perianth to a spiral (indeterminate) peri- anth or pentamerous perianth is not possible. Sig- nificantly, however, our analysis of perianth mer- osity indicates that there have been from one to five transitions from trimery to an indeterminate peri- anth (Fig. 5 and Table 2). Some of the changes from a trimerous perianth to an indeterminate perianth occur in Laurales. For example, the perianth of Ca- lycanthaceae, which has an indeterminate number of parts, may have evolved from an ancestor with a trimerous perianth. However, Laurales are unique among basal angiosperms due to the presence of a hypanthium, which may affect the development and arrangement of the perianth. Perianth differentiation has long been studied (Eames, 1931; Hiepko, 1965; Kosuge, 1994). cent investigations of the evolution of perianth dif- ferentiation have come from both apr die and developmental genetic perspectives (Albert et al., 1998; Kramer & Irish, 1999, 2000). dendi un perianths have been proposed to have evolved mul- tiple times from different structures in different groups (Hiepko, 1965; Kosuge, 1994; Takhtajan, 991). Our analysis of perianth differentiation in- dicates that there may have been as few as two to as many as six changes to a differentiated perianth from an undifferentiated perianth (Fig. 6) (Table 2). Thus, our results agree with earlier interpretations that a poca perianth (Hiepko, 1965; Ko- suge, 1994; 1991) evolved multiple times. Some differentiated perianths may be those in which the calyx is derived from bracts and the tajan, corolla derived from modified stamens, a corolla ‘omposed of “andropetals” (Kozo-Poljanski, 1922; Takhtajan, 1980). Other differentiated perianths may have a corolla that is derived from modifie bracts, referred to as “bracteopetals” (Kozo-Poljan- ski, 1922; Takhtajan, 1980). The ontogeny of petals and sepals (Kosuge, 1994) and the number of vas- cular traces (Carlquist, 1969; Melville. 1969; Puri, 1951; Smith, 1928) have been used as criteria for deciding if petals are staminal in origin or sepal/ bract in origin. Thus, the differentiated perianth found in Nymphaeaceae, Ranunculales, and Cary- ophyllales is considered to be staminal in origin, whereas the petals of Magnoliales and Austrobail- eyales are considered to be bract in origin (Takh- tajan, 1980). These are just two of several scenarios for organ differentiation. CONCLUSION Compartmentalization of a six-gene data set yielded trees congruent with many recent phylo- genetic analyses (Graham & Olmstead, wi a thews & Donoghue, 1999, 2000; Qiu et al., 2000; D. Soltis et al., 2000: P. Soltis et al., me Zanis et al., 2002). We also observed high levels of bootstrap support for most of the nodes in the com- partmentalized tree; the values we obtained were higher than those in any previous study. However, these values should be interpreted cautiously as the potential biases of compartmentalization analyses are unknown. The ancestral states for perianth phyllotaxis and merosity are unclear. However, the ancestral state for the large clade containing the magnoliid clade, Chloranthaceae, Ceratophyllaceae, monocots, and eudicots is a whorled perianth, with reversals to spiral phyllotaxis and an indeterminate number of parts in some lineages. These results support the hypothesis of floral lability (“open organization”) for basal angiosperms; intricate synorganization oc- curred with the origin of eudicots (e.g., Albert et al., 1998; Endress, 1987c, 2001). Our analysis of perianth evolution across basal angiosperms indi- cates that there are multiple most parsimonious re- constructions of perianth phyllotaxis and perianth ae reatistion. However, the trimerous perianth oc- curred early in angiosperm evolution. In fact, the ancestral state for the clade containing the mag- noliid clade, Chloranthaceae, Ceratophyllaceae, monocots, and eudicots is trimery. Much of our mapping of perianth differentiation is equivocal for the deepest nodes of the tree, although a differen- tiated perianth has clearly originated multiple times. Some differentiated perianths may evolved from spirally arranged perianths, whereas other differentiated perianths, such as that of Sar- uma (Aristolochiaceae), may have evolved from an- cestors that had a perianth composed of a single whor have Literature Cited Aboy, H ogy of Ceratophyllum demersum. M.Sc. University, Ithaca E. 1936. A Study of the Anatomy and Morphol- T hesis, Cornell 144 Annals of the Missouri Botanical Garden Albert, V. A., M. . Gustafsson & L. DiLaurenzio. 1998. Ontogenetic ыо s, molecular pi Eu tal genetics, and the к petal. Pp. 349—374 in D. E. Soltis, P. S. Soltis Joyle (editors), Molec- ular vw s of Pla B Il. Kluwer, Boston. APG II. . An update of the Angiosperm Phylogeny Group caesa cation of the orders and families of flow- ering plants. Bot. J. Linn. Soc. (in press). Arber, E. A. N. & J. Parkin. 1907. On фе origin of an- giosperms. Bot. x me 1. Soc. 38: 2€ Armstrong, J. E. & S. C. Tucker. 1986. Floral developmen in. Myristica ко aceae). Amer. J. Bot. 1131— 1143 =й T. К. Wilson. 1978. Floral Ж ты of ~ fieldia (Myristic aceae). Amer. J. Bot. 6 —14‹ Bailey, I. W. & В. С. L. Swamy. 1948. fte uM poda Вай, А new morphologic al type of vesselles cotyledon. J. n d Arbo a а Barkman, T. J., enery, Li eal, J. Lyons-Weiler & C. W. de EE Жо! ae pendent and com- bined analyses of sequences from all three genomic compartments converge on e root of flowering plant Do oen Proc. Natl. Acad. Sci., U.S.A. 97: 13166— 3171. ss di- ie C. E. 1897. аист ied taxonomy of the an- giosperms. Bot. Gaz. 24: 145-17 ————. 1915 The Ыы. a of flowering plants. Ann. Missouri Bot. Gard. 2: 109-164. Bowe, L. M., 6. Coat & C. (| s 2000. Phy- logeny of seed plants based on al e genomic com- partments: Extant gymnosperms are enda ca tic is nd Gnetales’ closest re doy tives are c Acad. Sci., U.S.A. 97: 4092409 Bremer, K. 1988. The limits of amino acid sequence data in M phylogenetic: reconstruction. Evolution 42: -803. Brizic у, G. K. 1959. Variability in the floral parts of Go- mortega (Gomortegaceae). Willdenowia 2: 200—207. Buchheim & R. L. Chapman. 1991. Phylogeny of the calonia green йар шке: A study of $5 and 26S rDNA sequence data. Bi decane 25: 85-10 Carlquist, S. 1969. Toward a table e of floral anatomy. ЖОШ 19: 332-362. Celakovsky, L. J. 1900. Uber den phylogenetischen En- twicklungsgang der B. und über der Ursprung der амел ыт I. Theil. Sitzungsber. 2 i Ges. onifers. Proc. Natl. ss. Prag, Math. -Naturwiss. Cl. 1 i i W., E. Soltis, R. - Olmstead, Morgan D. H. Les, B. D. Mishler, M. uv A Price, . G. Hills, a -L. Qiu, K. A. ir H. Rettig, F Cont. J. D. dede R. Manhart, "К, J. Sytsma, Н. J. Michaels, W. J. Kre Karol, W. D. Clark, M. E. Golenberg, G. H. Learn, Jr., S. W. Graham, S. ( Barrett, S. айана & V. A. Albert. 1993. Phylo- genetics of seed plants: An analysis of nucleotide se- quences apes the plastid gene rbcL. Ann. Missouri Bot. Gard. 80: 528—580 ——-— 5 Stevenson, P. Wilkin & P. J. Rudall. 1995. Monoe sen systematics: A combined analysis. Pp. 35-730 in P. J. Rudall, D. F. Cutler & C. J. Humphries aeris Pus otyledons: ми and Evolution. d deck Gardens Chaw, S.-M., C. L. Parkinson. Y. Cheng, T. M. Vincent & phylogenetic perspe .د‎ 1996, of the Gnetales. Int. J. Pl. Sci. — 10 J. D. Palmer. 2000. Seed plant phylogeny inferred from all three plant genomes: Monophyly of extant gymno- sperms | к origin of Gnetales from conifers. Proc. Natl. Acad. Sci., U.S.A. 97: 4086—4091. Crane, P. R 1985. Phylogenetic analysis of seed plants and the origin of angiosperms. Ann. Missouri Bot. Gard. 72: 716-793. 988. clades and relationships in the "higher" PIDE. Pp. 218-272 in C. B. Beck (ed- itor), Origin and Evolution of Gymnosperms. Columbia Univ . Pre ess, New York. Friis & K. R. Pedersen. 1995. The origin and carly diversification of angiosperms. Nature 374: 27-33. Cronquist, A. 1981. An Integrated System of Classification of Flowering Plants. Columbia Univ. Press, New York. ——. 1988. The Evolution and Classification of ii ering ‘Plants, w ork Botanical Garden, New Dalec: n, R. M. T., P. F. Clifford & P. Е Yeo. Families of dis Minos 'otyledons: Structure, Evolution, and ۰ Springer, Berlin. Dengler, . 1972. Ontogeny of the vegetative and e apex А н ы occidentalis. Canad. J. Bot. 5 9-1356. Banche: M. J. 1994. Progress and prospects in recon- MEET plant phylogeny. Ann. Missouri Bot. Gard. 81: 405—418 é D. Ackerly. 1996. Phylogenetic unc boh ties gen aa. analyses. Philos. Trans., Ser. В: 1241-1249. E J. ^. Doyle. 1989. Phylogenetic analysis of an- giosperms and the nu of Hamamelidae. Pp. 17—45 in P. R. Crane Blackmore (editors), Evo- d erae and F Fossil History of the Hama- *. Vol. 1: Introduction and ‘Lower’ Hamamelidae. Clarendon bl. 0 — ——-& ————. 2000. Demise of the anthophyte hy- pothesis? Curr. Biol. ‘na R106—-R 109. —— R. Н. Ree & D. A. Baum. 1998. Phylogeny ape the мн of flower symmetry in the Asteridae Trends PI. Sci. 3: Vp ed Douglas, A. W. & S. C. Tuc io т. 1996a. Comparative floral ontogenies among Sears Туа: inc ene, Bellendena (Proteaceae). = er. a Bot. 83: 8-15 — — Ё ————. 1996b. Inflorescence DM and flo- ral OON 2 in Grevilleoideae (Proteaceae) with ‚шр. on the nature of the flower parts. Int. J. Pl. 97: 341-372. Doust, A. N. 2000. Comparative жү d in Win- leraceae. Апп. : 36€ 2001. The developmental ied ol florida variation in Drimys winteri (Winteraceae). Int. J. PI. Sci. 162: 697-7 Doyle, J. A. 1994. The origin of the pues rm flower: A clive. Pl. Syst. Evol. 152: 7—30. Seed plant phylogeny and + relationship 157 (Suppl.): 53—539. Molecules, morphology, fossils, and the re- Molec. Phy- Missouri ТТ ра of angiosperms and Gnetales. logenet. Evol. 9: 448—462. & P. — é K. s ч 2000. Morphologic ‘al phyloge- netic analysis of basal angiosperms: by goes and ms with Em ular data. Int. J. РІ. . 161 (Suppl.): 5121—51 xd A. Le е 1996. Phylogenetic analysis and character evolution in Annonac e on Mus. Natl. Hist. Nat., B, Adansonia 18: 27€ Volume 90, Number 2 2003 Zanis et al. 145 Perianth Evolution in Basal Angiosperms ‚ M. J. Donoghue & E. A. Zimmer. 1994. Integra- tion “ morphological and ribosomal RNA data on the origin of the angiosperms. Ann. Missouri Bot. Gard. 81: 9—450. Drinnan, A. N., P. R. Crane т S. B. Hoot. 1994. Patterns ` of floral evolution in the per с s non- magnoliid dicotyledons кыйса, РІ. vol. (Suppl.): 93-122. ‚ О. 1887. Die systematische und geographisc һе Anordnung der Phanerogamen. Pp. 175—496 in Schenk (editor), Handbuch der Botanik III, 2. B PM Eames, A. J. 1931. The vascular anatomy of the flower with paer a the theory of carpel Alec a Amer. J. Bot. 47: 147—188. Endress, Р. K. ves osi und Me ie der mantandraceae. Ber. tsch Bot. Ж 90: 83—103 80. Floral structure and deret A of Hor- tonia (Monica Pl. Syst. Evol. 133: 199-221. 1983. The early floral development 2 Austro- baila, ва Jahrb. Syst. 103: 481—497 6. Reproductive structures and phylogenetic € of extant primitive angiosperms. Pl. Syst. Evol. : 1-28. Es The Chloranthaceae: Reproductive struc- tures and phylogenetic position. Bot. Jahrb. Syst. 109: 3-220. —— ——. 1987b. Floral e and evolution. Bot. Jahrb. E 108: 417—43€ The early E of the angiosperm iwa A Ecol. Evol. 2 304. 1994a. Diversity ^ prm Biology of Tropical Flowers. Cambridge Univ. Press, Cambridge. 94b. Evolutionary aspects of the floral struc- tures in а РІ. Syst. Evol. 8 (Suppl.): 175— 183 . 2001. The flowers in extant basal ым апа inferer ances on ancestral flowers. Int. J. Pl. Sci. 1111- ^. 3 Hufford. 1989. The diversity of stamen structures and dehiscence patterns among Magnoliidae. Bot. J. Linn. Soc. 100: 45-85. & A. Igersheim. 1997. Gynoecium diversity and systematics of the Laurales. Bot. J. Linn. Soc. 125: 93- 168. & . 2000. Gynoecium structure and evo- lution in basal angiosperms. Int. J. Pl. Sci. 161 (Suppl.): S211 -5223. ^ B. Sampson. 1983. Floral structure and re- Банга of the Trimeniaceae (Laurales). J. Arnold Arbor. Hu 447—473 A. lg heim; К. B. Sampson & С. E. Schatz. 2000. Floral structure of Takhtajania ibd i its Parce position in Winteraceae. Ann. Missouri Bot. Gard. 8 347- TI Erbar, C. & P. Leins. 1981. Zur epu in Magnolien- iron Beitr. Biol. grum 56: 225-2; & .l ur Sequenz von qM SUN B einigen Майдай, Bot. Jahrb. Syst. 103: 433- M eun ————. 1985. Studien zur Organsequenz in Apiaceen- Sane n. Bot. Jahrb. Syst. 105: 379—400 Duos a & P. Leins. 1999. Development and ms of ee organs in Ranunculaceae. Flo- ra 194: 317-332 Esau, K. 1965. Plant Anatomy. John Wiley & Sons, New York. Fan, C. . Xiang. 2001. use ns relationships within ree (Cornaceae) based on 268 rDNA se- quences r. J. Bot. 88: 1131-1138, Farris, J. S., V А. pen M. Kallersjé, D. Lipscomb & = С. Klu 6. Parsimony jac н ssa iig eithbor joining Cladistics 12: 99- Veiis: J. 1985. Confidence тен on phylogenies: An approach using the bootstrap. Evolution 39: 783- 791. Fishbein, M. C., C. Hibsch-Jetter, D. E. Soltis & L. Huf: ford. 2001. Phylogeny of bx y Cp (angiosperms, eu- dicots): — of a rapid, ancient radiation. Syst. 47. 963. The morphology and еи of Ссн ^ Arnold Arbor. 42: 299-2 Friis, Ё P. R. Crane & K. R. Pedersen 1997. Fossil history ul танца angiosperms. 121-156 in K. Iwatsuki & P. H. Raven (editors), cu. and Diver- sification of Land Plants. Springer- -Verlag, New York. edersen & P. R. Crane. 1994. Angiosperm floral нна from the Pu ч retaceous of Portugal. PI. Syst. Evol. 8 (Suppl): 3 T ханаа a cmm structure and pese of hasal angiosperms from the early ر‎ (Barremian or Aptian) of western Portugal. I 182. nt. J. Pl. Sci. 161 (Suppl.): 5169-5 Glück, H. 1919. Blatt- und blütenmorphologische Stu- dien. Je zotha. г evenit 2000. Perianth devel- opment ig псом s of Aristolochia. Flora 195: 370— Grahaun, S. W. & R. G. Olmstead. 2000. Utility of 17 chloroplast genes for inferring the phylogeny of the bas- a vii gp Amer. J. Bot. 87: 1712-1730. Greyson, R. 199 E редот of Flowers. Oxford Univ. Press, m Hamby, R. K. & E. А. ae 1992. Ribosomal RNA as a phylogenetic tool in on os Pp. 50-91 in P. S. Soltis Soltis & J. J. Doyle (editors), FEE ular m. of Plants. Chapman and Hall, New , E. L. Schneider & S. Carlquist. 2000. Floral pw nw of Nelumbo nucifera Loco) Int. (Suppl.): 5183-819 965. Ие ak ИРИ und en- Perianth bei den Acide Bot. Jahrb. Syst. 84: ‚ J. W. Kadereit, F. и Blattner, К. B. Jork, A. E. арас & Р. К. Crane. 1997. Data congruence and phylogeny of the a s.l. based on four data sets: atpB and rbcL sequences, trnK restriction sites, and morphological characters. Syst. Bot. 22: 575- Huber, Н. 1993. Aristolochiaceae. Pp. 129-137 in K. Ku- V. Bittrich (editors), The Fami- s, Vol. 2. Flowering Plants: Dicotyledons: Magnoliid, Hamamelid and Car- yophyllid dics Springer, Berlin. Huber, K. A. Morphologische und entwicklungs- шае Untersuchungen an Bliiten und Bliiten- standen von Solanaceen und von Nolana paradoxa Lindl. (Nolanaceae): Dissertationes Botanicae, v. 55. Cramer, Vadus. Hufford, L. 1996. The morphology and evolution of male 146 Annals of the Missouri Botanical Garden reproduc E. structures of Gnetales. Int. J. Pl. Sci. 157 (Suppl.): 595-8112. — —— & P. R. Crane. 1989, A preliminary p analysis of the “lower” Hamamelidae. Pp. 175-192 i . R. Crane & S. Blackmore (editors), А оп, Sys- tematies, and Fossil History of the Hamamelidae, Vol. . Clarendon Press, Oxford, U.K. en A. & P. K. Endress. 1997. Gynoecium diversity and systematics of the Magnoliales and Winteroids. Bot. J. Linn. Soc. 124: 213-271. Karol, K. G., Y. Suh, G. E. Schatz & E. A. Zimmer. 2000. Molecular evidence for the phylogenetic position of Takhtajania in the Winteraceae: Inference from nuclear ribosomal and c ca gene spacer sequences. Ann. Moni Bot. Gard. 8 4—432. . 1993. Menispermacene. Pp. 402—418 in K. Kubitzki, J. G. Rohwer & V. Bittrich Spa The Families and Genera of Vasu Plants, Vol. ow- Yoliid, ат апа t Berli in. etal oe in Ranunculaceae. PI. 2 S Ol. Le Kozo- Poljanski, B. M. 6 An Introduction to Phyloge- netic нз da the Higher Plants. Voronezh. [In Russian.] Kramer, E mec hanisms c sdb petal development. Nature 144—148. . F. Irish. 1999, Evolution of genetic 399: | —— —— —. 2000. тч of the petal and sta- men developmental programs: Evidence from compar- ative studies of lower eudicots d ш angiosperms. Int. J. Pl. Sci. 161 (Suppl.): S29-S Kubitzki, K. 1987. Origin and is ance of trimerous -28. wer & V. Bittrich. 1993. The Families and Genera of en = Plants. qra Berlin. Kuzoff, R. K., J. A. Sweere, D. E. Soltis, P. S. Soltis & E. x с 1998. The нин potential of a 2 sequences in plants. Molec. Biol. Evol. Leins, P. ác Erbar. 1985. Ein Beitrag zur Blütenen- twic ыш, де Aristolochiaceen, einer Vermittlergruppe zu den аео Bot. Jahrb. Syst. 107: 343—368. Les, H. ie s and affinities of the Cerato- е см Тахоп 37: 345. ‚ D. К. Garvin i 1. F. Wimpee. 1991. Molecular of ancient aquatic angiosperms. atl. Acad. Sci. U.S.A. 88: 10119-10 123. 2. L. Schneider, D. J. Padgett, P. S. Soltis Soltis & M. J. Zanis. 1999, Phylogeny, кета ай m floral d of water lilies (Nymphaeaceae; Nym- phaeales): A synthesis of non-molecular, rbcL, matK. and 185 ee data. Syst. Bot. 24: 28—46. Loconte, H. & D. W. Stevenson. Magnoliidae. Сыйы 's 7: 267- Lyons-Weiler, J., ( Hoelzer Tausch. 1996. Relative Арак Singh Neem (RASA). I. The po hs measure of phylogenetic signal. Molec. Biol. Evol. 13: 74 9-757. Mabee, P. M. г. Humphries. 1993. Coding d data: Examples from allozymes and ontogeny. Syst. Bio : 166-181. evolutionary history Ni 1990. Cladistics of the di P. Maddison, D. R. . The discovery and importance of multiple sands E most- parsimonious trees. Syst. Zool. Maddison: W. Р. 1993. Missing data versus missing char- acters in к төбе з analysis. Syst. Biol. 42: 576—581. & D. R. Maddison. 1992. MacClade 3.04: Anal- ysis of Phylogeny and Character Evolution. Sinauer, . Crane & P S. Herendeen. 1999, Phy- logenetic ай; diversity, and синин ‘ation of eudi- Missouri Bot. Gard. 86 72. Magallón- Puebla, S., P. S. оем & P. is е 1997. Quadriplatanus ош! ius gen. et sp. : Sta- minate and pistillate p alanaceous etn from ne Late Cretaceous (Coniac ш of Georgia, U.S.A. Int. J. РІ. m 158: 373-394. Manchester, S. R. 1986. Vegetative and reproduc tive mor- phology of an extinct plane tree (Platanaceae) from the of western North America. Bot. Gaz. 147: 200- cots. Ann. Eoc 'ene 5. Mathews, S. & M. J. Donoghue. 1999. The root of angio- sperm phylogeny duplicate phytochrome genes. Scienc Dn. ен пшр phylogeny in- us from Julie ate phytochromes A and C. Int. J. Pl. . 161 (Su 541-555. тв. ш С. A 1980. Phylogeny of the gl n vith an evaluation of classifications. Syst. Bot. 5: 149— Melville, R. 1969. и in floral structure and evolu- tion. Kew re E 33-180. Mishler, ae analysis of molecular and sni mu Kind Amer. J. Phys. Anthropol. 94: 143- 156. Moseley, M. К. 1961. Morphological studies of the Nym- phaeaceae. П. The flower of Nymphaea. Bot. Gaz. 122: 33- pe . Morphological studies of Nymphaeaceae. 1. Developmen of the flower of Nuphar. Phytomor- phology 2]: —252. E P. 5. Williamson & H. Kosakai. 1984. MEE 'al studies of the Nymphaeaceae. XIII. Con- tributions to the ve dod and floral structure of Ca- bomba. pes ^d Bot. 71: 902-924. Naegeli, C 4. Mec ed h- physiologisc ‘he Theorie der еа чан, Munic g & Leipzig. Nandi, O. I., M. W. i P. K. Endress. 1998. A combined cladistic insi. of angiosperms using rbcL Ann. Missouri Bot. Gard. and non-molecular data sets. 85: 137-212. Nemejc, F. 1956. On the problem of the origin and phy- logenetic Wie eae of the Angiosperms. Acta. Mu Nat. dei 1 n 144 Nixon, K. C. & J.I qu ‚ 1991. Polymorphic taxa, miss- ing values and cladistic analysis. Cladistics 7: 233— Padgett, D. J., D. H. Les & G. E. Crow. 1999. Phyloge- netic relationships in Nuphar (Nymphaeaceae): Evi- dence from morphology, oera a m nuclear ribosomal DNA. Amer. J. Bot. 86: же Parkinson, С. pm & J. ^el 1999, Multigene idis identify the Pon yis ast lineages of n pides plants. C s pe 9: 1485-1488. Pigg, K. . Stockey. . Platanaceous pd from b Eu ‘ene of Alberta jme Rev. Palaeo- bot. Palynol. 70: 125-146. Posada, D. & K. A. Crandall. 1998. Modeltest: Testing the model of DNA substitution. Bioinformaties 14: 817— 818. = Prantl, K. 1887. Beitrüge zur е pen Systematik der Ranunculaceen. Bot. Jal 5-273. irb. Syst. 9: 2 Volume 90, Number 2 3 Zanis et al. 147 Perianth Evolution in Basal Angiosperms Puri, V. 1951. The role of floral anatomy in the solution р morphological problems. Bot. Rev. 17: 471—553. Y.-L.. M. W. Chase, D. H. Les & C. L. Parks. 1993. "Molecular phylogenetics of the Magnoliidae: Cladistic analyses of nucleotide — of the plastid gene rbcL. Ann. Missouri Bot. Ga 587—600. „J. bo F. Bernasconi-Quadroni, D. E. Soltis, P. S. Soltis, M. Zapis, E. A. Zimmer, Z. en: , V. Savolai- nen & M. W. Chase. of ‘the earliest angiosperms from mines a ay a and nuclear genomes. = e 402: 404—407. n Bernasconi-Quadroni, D. E. Soltis, P. 5. Soltis, M. " Zanis. E. A. Zimmer, Z. Chen, V. Savolai- nen & M. W. Chase. 2000. Phylogeny of basal angio- sperms: Analysis of five genes from three genomes. Int. J. Pl. Sci. 161 (Suppl.): 53-827. Ree, R. H. & M. J. Donoghue. 199 tep matrices and = interpretation of homoplasy. E Biol. 47: 582- 588. йө, К. A., M. J. Donoghue & R. С. Olmstead. 1997. Analyzing large data sets: rbcL 500 revisited. Syst. Biol. 46: 554—563. Ro, K., C. S. Keener & B. A. McPheron. 1997. Molecular phylogenetic study of the Ranunculaceae: Utility of the nuclear 26S ribosomal DNA in inferring сиз соон Molec. Phylogenet. Evol. 8: Robertson, Ң. E Tucker. 1979. Scale toe of Illicium Jordani, with emphasis on stamen and carpel development. Amer. J. Bot. 66: 605—61 Savolainen, V. W. Chase, C. М, jue s. D. E. Soltis, C. e M. F. Fay, A. De Bru livan & Y.-L. . 2000. ао of rali plants based as a combined analysis of экы» atpB and rbcL gene sequences. Syst. Biol. 49: Smith, G. H. 1928. Vascular anatomy of Ana flowers II. Ranunculaceae (continued), Menispermaceae, Caly- canthaceae, рее Вог. Сах. 85: 152-177 Soltis, D. E., Soltis, M. W. Chase, М. E. Mort, D. C. Albach, M. ps V. m W. H. Hahn, S. B. L^ M. F.F Fi M. Axtell, S. M. Swensen, L. M. Prince, W. J. Kress, K. C. Nixon & J. S. Farris. 2000. Angio- sperm phylogeny inferred from 18S rDNA, rbcL, and atpB sequences. Bo = D M. Chas, L. J. Сок, W. J. Kre s & K. J. Sytsm 1997. Angiosperm phylogeny inferred from 18S ribo- somal DNA sequences. Ann. Missouri Bot. Gard. 84: -49. ‚ A. Senters, S. Kim, J. D. Thompson, P. S. Soltis, M. J. Zanis, L. Ronse de Craene, P. K. Endress & J. S Farris. 2003. Gunnerales are sister to other core eudi- cots, and exhibit ee oca of early-diverging eu- dicots. Amer. J. B 3. Soltis, P. S., D. E. Soltis & M. W. Chase. 1999a. Angio- sperm phylogeny inferred from po I as a tool for — g^ Nature 402: 40 P. G. Wolf, D. L. i -M. Chaw &R Wu 1999b. The phylogeny of bud plants өтө from 188 rDNA sequences: Pushing the аш of rDNA signal? Molec. Biol. Evol. 16: 1771-17 . M. J. Хапіѕ & S. Kim. 2000. twi Tis Loo of angiosperms: Relationships and implications for floral evolution. Int. J. Pl. Sci. 161 (Suppl.): S97- 107. Stebbins, G. L. 1974. Flowering Plants: Evolution Above arvard Univ. Press, Cambridge Character states, inorphalogicsl я vari- ation, and phylogenetic analysis: A review. Syst. Bot. 16: 553—583. Swamy, B. G. L. 1949. Further газ ret te to the mor- phology of Degeneriaceae. JA nold Arbor. 30: 10—38. Swofford, D. L. 1998. PA «40: ‘Phylogenetic Analysis 5 chusetts. Phylogenetic inference. Pp. 407—425 in D. M. Hillis, B. K. Mable (editors), Molecular System- atics, 2nd ed. pe Sunderland, Massachusetts. lowering Plants: Origin and Dis- . Outline of the rac of flowering plants Мого) Bot. Rev. 46: 225—359. . Evolutionary Trends in Flowering Plants. "EN Od Univ. Press, New Yor 1997. Diversity and Classification of Flowering Plants. Columbia Univ. Press, sll Ba Flexible strategies for multiple Ке os ment aided by quality analysis tools. Nucl. Acids Res 4: 4876-4882. Thorne, R. F. 1992. An км зви of the flow- ering P 2 13: 365-38 Tucker, S. € a of i i apex of Michelia fuscata. ree J. Bot. 47: 266-27 & A ouglas 1996. T al structure, devel- opment, and renun m of paleoherbs. Pp. 141-175 in D. W. Taylor & L. J. Hickey (editors), Flowering Plant Origin, Evolution and Phylogeny. Chapman and Hall, New York. Van Beusekom, C. F. 1971. Revision of Meliosma (Sabi- aceae), section Lorenzanea excepted, living and fossil, geography and و ا‎ Blumea 19: 355-529 Velenovsky, J. 1910. Vergleichende Morphologie der Pflanzen. ا‎ ios der Blüte der Phanerogamen 4: 733-12 Vink, W. Ten The Winteraceae of the Old World. 1 Pseudowintera and Drimys—Morphology and taxonomy. Blumea 18: 225-354. The Winteraceae of the Old World. II. Zyg- ogynum—Morphology and taxonomy. Blumea 23: 219- 25 von Balthazar M. & P. K. Endress. 2002. Flowers of Bux- aceae. Int. J. Pl. Sci 163: 847-876. Walker, J. W. & A Walker. 1984. Ultrastructure of Lower Cretaceous angiosperm pollen and the origin and early с of flowering plants. Ann. Missouri Bot. Gard. 7 2]. Weberling, r pes Morphology of Flowers and Inflores- ) The ката of self-incompatibility in angiopserms: A phylogenetic approach. /n P. Hoch & A. Stephenson (editors), Experimental and Molecular Approaches to Plant Biosystematics. Monogr. Syst. Bot. Missouri Bot. Gard. 53: 355-382. Wiens, J. J. 1999. Polymorphism in systematics and com- parative ut Ann. Rev. Ecol. Syst. 30: 327-367. Wilson, T. K. 1966. The comparative morphology of the Vide nsn IV. Floral EPUM and conclusions. г. J. Bot. 53: 336- Am Yang. Й 1997. РАМІ: А ne package for Phyloge- 148 Annals of the Missouri Botanical Garden netic Analysis by Maximum Likelihood. 555-556. ‚ 5. Kumar & M. Nei. 1995. infigne of ancestral nucleotide and amino acid se quences. s 's 141: po ). Zanis, M. J., D. E. Soltis, P. S. Soltis, S. Mathews & M. J. Donoghue. 20 The root of the angiosperms revis- ited. Proc. Natl. Acad. Sci., U.S.A. 99: 6848-6853. CABIOS 15: APPENDIX 1. DESCRIPTIONS AND REFERENCES FOR PERIANTH CHARACTERS USED IN THIS STUDY Outgroup. With the exception of Gnetum, none of the extant gymnosperms has a m nth. The extinct Ben- nettitales have a perianth. However, Crane (1985) sug- gested that the perianth found in the Bennettitales is not yle, 1996, l . Thus, we have scored the зл in the outgroup as absent. Amborellaceae. Bailey and Swamy (1948) described the male and female flowers in detail, noting that the perianth is not differentiated into calyx and corolla, but instead composed of tepals. The perianth is arranged in a cyclic ог spiral a arrangement, and the number of tepals ranges from six to eight in the female flower. This work was fur- ther corroborated by recent work done by Endress and lgershei im (2000 which shows a spirally arranged undif- el perianth; the number E tepals ranges us 9 › 11 in male flowers and is in female flowe Arist E АП genera of Pes uere a a trimerous perianth that consists of a single whorl, with the exception of Saruma, which has two whorls that are dif- ferentiated (Huber, 1993; Leins & Erbar, 1985). In some species of Asarum, petals apparently begin to de ‘velop, but the only traces are small, threadlike structures (Leins & Erbar, 1985). Gonzalez and Stevenson (2000) demonstrat- ed that the uniseriate perianth in Aristolochia is derived from the outer whorl of a biseriate perianth. Austrobaileyales. Austrobaileyaceae, ' liciaceae (APG II, scored 7 or vosed of iisandraceae, and Il- 2003). These families have all been having an undifferentiated, spirally arranged perianth with an indeterminate number of parts (Endress, 1983; Endress & Sampson, 1983; Robertson & Tucker 1979). Buxaceae. Buxaceae have unisexual flowers. The sta- minate flowers have a perianth consisting of two or three whorls of dimerous tepals (Drinnan et al., 1994). The pis- tillate flowers typically have a spirally Ен ee with 6 to 20 tepals (Drinnan et al., 1994). However, Drin- nan et al. (1994) noted that the tepals in ilie ‘pistillate flowers are similar to the f the staminate flowers and are basically bracts w i i Austrobaileyales are comp lrimeniaceae, Schi died as Шен а eae as being unc йегет] and dimerous, with the RS r of whorls being polymorphic (two or three whorls). Canellaceae. Canellaceae have a differentiated, whorled perianth; the number of whorls varies from two to multiple. The perianth is composed of three fleshy sepals: the num- ber of petals ranges from 4 to 12, with each whorl being either trimerous, tetramerous, or pentamerous (Wilson A new method of 1966). We coded the perianth as differentiated and poly- morphic for the number of whorls and the number of parts. Ceratophyllaceae. Ceratophyllum has a that subtend the reprodi ауе been eds interpreted as а perianth or. bra (Aboy, 1936; Endress, 1994b; Les, 1988). Some reports suggest the presence of additional flowers between these scales, indicating that the scales may best be interpreted as bracts (Aboy, 1936). In our analysis we coded Cerato- phyllaceae as lacking a perianth, following Aboy (1930) and Les et al. (1993). Chloranthaceae. Within the Chloranthaceae only the fe- male flowers of Hedyosmum have a i a perianth (Endress, 1987a as lacking a perianth and Бакен, as having a single trimerous, whorled perianth. Lactoridaceae. Lactoridaceae have a /single-whorled, trimerous ge nth. Laura Laurales consist of seven families: meaning eae, duni ycan aceae, rl of scales = Atheros- ithac eae, Auraceae, bers of this group have а hypanthium. Atherospermata- eae have an indeterminate number of ends imis perianth parts that are spiral or whorled. Calycanthaceae have an undifferentiated perianth that is pe S arranged; the number of p. is indeterminate (Dengler, 1972 mortegaceae have a spiral, undifferentiated perianth with an indeterminate кеп of parts (Brizicky, 1959). Her- nandiaceae have a perianth with one or two whorls ilm 59 to буе undifferentiated parts (Kubitzki et al., 1993). Laur- aceae have an undifferentiated perianth of two whorls, each whorl consisting of three parts. Monimiaceae have a complex perianth, with some members having a well-de- veloped perianth and others having an enlarged receptacle and diminutive perianth. The perianth in all deus 'eae is undifferentiated and varies in number "aris; some taxa have a spiral ропада (e.g.. ~ Endress ‚ 1980; however, see Doyle & Endress, 2000) and others ave a whorled perianth. Siparunaceae have a cach of a single whorl and a variable number of parts (Doyle & Endress, ^ )- — — = Magnoliales. Magnoliales are composed of six families: Annonaceae, Degeneriaceae, Eupomatiaceae, Himantan- draceae, Magnoliaceae, апа Myristicaceae. Annonaceae have a trimerous, whorled, nuc re е perianth (follow ing Doyle & нен, 2000; е Thomas, 1996). Degeneriaceae have a im | trimerous, spirally arranged perianth Биши, 1949). Eupomatiaceae have a calyptra; however, it is uncertain if "m structure is sepal or bract in origin (Endress, 1977); thus, we coded it as "T". Himantandraceae have two sepal-like structures that surround the bud; however, it is uncertain if We struc- tures are perianth or bracts (Endress, 1977). Thus, we coded Himantandraceae as *?". S Tei ыы have a perianth with numerous par B (n »parently indeterminate; e.g.. Magnolia), but in y Деш а ол and certain species of Magnolia, the per Eo is in three trimerous hin (Er- xw & Leins, 1981, 1983; Tucker, 1960). Myristicaceae lave a бед of a ad trimerous, whorl (ere & Tucker, 1986; Armstrong & Wilson, 1978). Monocots. Monocots have me мз that lack a perianth and other taxa that have either a differentiated or an un- differentiated perianth. In nearly all taxa that have anth, the perianth is whorled and trimerous; only a few taxa have а pentamerous perianth (Chase et al., 1995; — а а peri- Volume 90, Number 2 Zanis et al. 149 2003 Perianth Evolution in Basal Angiosperms Dahlgren et al., 1985). We have used the basal lineages in our analysis. The perianth is complex within this group of the monocots as plac «иет for our analyses. Nelumbonaceae. Nelumbonaceae have been considered to have a spirally Heini: perianth with two (to five) s pals -— 10—30 petals (Hayes et al., 2000). aeaceae. Nymphaeaceae are composed of tw clades, sometimes treated as the distinct d Cabom- baceae and Nymphaeaceae s. str. (APG II, 2002; Les et al., . Les et al. (1999) considered the two families to constitute М poro se papi have either a e gerens (Endress, 2001; Les et al., 1999; Moseley , 1984) or undifferentiated (Doyle & Endress, 2000) print the phyllotaxis is considered to be whorled. Within Nymphaeaceae, periant th phyllotaxis has been re- т Albert et al., 1998; sider e & Endre 2000; е ены 2001: 1972) Nuphar id Barclaya each has a Фон ы рабы nth, whereas in Nymphaea and Victoria the differentiation is gradual. The number of A ‘tals and ех varies throughout the family in genera uphar i is itt red to have a perianth E ive, six, or nine sepals Paiga et al., 1999). T be base d on а trimerous iai plan the number of A pet wi Nymphaea and from 50 to с in Victo are scored as having a trimerous, iren perianth, and are scored as polymorphic эм perianth differentiation. In aeaceae, each taxon was scored as polymorphic for phyllotaxis and for number of perianth parts (see text for discu ssion eae. Piperaceae lack a perianth. ^ onem eae. The architecture т the perianth of Pla- tanus is uncertain (Drinnan et al., 1994). С b (1981) defined the perianth as ops | of three or four poorly hé oh sepals that alternate with small petals. Hufford and Crane (1989) coded Platanus as having an undiffer- entiated preis Doyle and Endress (2000) coded Pla- tanaceae as having a one- or two-whorled, undifferentiated perianth, with the number of parts two, four, or five. Stud- ies of fossilized Platanaceae indicate that the ener: was diverse, with some members having a well-developed un- differentiated perianth and others having staminate flowers "gg & S Stockey, 1991). We ыа Platanaceae as e an undifferentiated peri- anth, polymorphic for number of whorls (one or two whorls), and polymorphic for the number of parts (trim- erous, DU. dimerous/tetramerous). Proteaceae. The perianth in Proteaceae consists of four petaloid send (Douglas & Tu ipe 1996a, 1996b), some- times referred to as sepals, and two to four scales, some- times referred to as petals (Takhtajan, 1997). We coded n h the number of whorls (a uncertain ed as кзн dimerous о T 1996b). These different coding: have no impact r conclusions (see also D. Soltis et al., 2003). The family i is coded as tetramerous in T gure Ranunculales. We used seven families d Ranunculales 1 terms of merosity, phyllotaxis, and differentiation (Drin- nan et al., 1994; Kosuge, 1994). In Papaveraceae, the perianth contains one dimerous P of sepals and two dimerous petal whorls (Hoot et al., 1997). Trimerous flow- ers are n d in some т of E raceae, and some enera contain species in which both ums and trim- erous periamths are ass (Hoot et al., Marra some laxa in Papaveraceae lack petals ul : al., 1997). Fupteleac eae lack a per pth, ou gh the lower Pisa two small prophylls (En- varies from dimerous to trimerous (Foster, 1963). Lardi- zabalaceae have flowers with a two- Wage undifferenti- ated, trimerous е peran (Drinnan et al., enisper- whorled, Ихсар, trimerous perianth a two-whorled dimerous (Meacha la tremely variable perianths (Коше, 1994). Petals appear to have evolved multiple times within this family, and pet- als/nectaries are usually е as derived from sta- mens (Erbar et al., 1999; Kosuge, 1994). Sepals range from five to eight and when an. are present they range from five to seven. We coded many m of the order as polymorphic for all three characte tosids, asterids, Santalales, and Cass phyllales. For the ix dani coding of are coded as having a puc that is trimerous, averni erous, or tetramerous Sabiaceae. асе eae have а two-whorled perianth that is composed of three to five sep five petals (Drinnan et al., 1994). Van = жан qun that the flowers of Sabiaceae are pentamerous and t trimery has arisen through a series of reductions Saxifragales. Saxifragales are a po early- diverging group of eudicots (Fishbein et al., 2 The perianth morphologies in this group vary, and we vs dud the group pu for each of the perianth characters analyzed here. Perianth merosity is coded as absent, indeterminate, tetramerous, and pentamerous. Perianth differentiation is coded as undifferentiated. Phyllotaxis is coded as absent or in two whorls Saururaceae. The family comprise five uum and only seven species. The flowers are perfect and without a peri- anth. Trochodendraceae. Trochodendraceae comprise absent, differentiated, and may be interpreted as a perianth, stamens on lateral flowers. We coded Té orls of undifferentiated, dimerous tepals and авт as lacking a perianth. nteraceae. The phylogeny of Winteraceae indicates that Takhtajania, Tasmannia, and Drimys are Каи а sisters to the rest of the family (Karol et al., ). Pat- terns of variation in perianth structure and жй | in Winteraceae are complex, with some members having а £ Annals of the Missouri Botanical Garden differentiated, whorled, dimerous, perianth and others a perianth with spiral phyllotaxis (Doust, 2000, 2001; Vink, 1970, 1977). ки has а perianth that begins as dimerous, then changes to оа and eventually Бе- comes pentamerous (Е ndress et al., 2000). Tasmannia has a perianth that has four sepals and EL petals in a whorled arrangement. Drimys has a perianth that varies in terms of phyllotaxis as well as number of sepals and petals. In fact, phyllotaxis in Drimys may vary on the same plant Doust, 2001). We coded the family as having a differen- tiated perianth. Because all of the early-branching mem- bers ч the family have a whorled phyllotaxis, we coded Winteraceae as whorled for phyllotaxis and polymorphic for Ше number of perianth parts. -~ MOLECULAR SYSTEMATICS, EVOLUTION, AND POPULATION BIOLOGY IN THE MUSTARD FAMILY (BRASSICACEAE)! Marcus Koch,? Ihsan A. Al-Shehbaz,? Klaus Mummenhoff* and ABSTRACT The present review summarizes results from the past decade on the —€— s, Lae ај genetics, and evolutionary biology of the mustard family, Brassicac eae (Cruci erae). The research of vari author The review is useful in view си the N increasing work on а да,а of A. helana is only very s discussed and presented research should move beyond assessing genetic relationships or бык: апа qur also pare aa ane developmen and evolution, the molecular basis of various homoplastic characters, the nature of the genome, and many other ne challenges that are emerging from detailed molecular studies of A. thaliana. e Arabidopsis, Brassicaceae, Cruciferae, evolution, literature review, molecular systematics, polyploidy, population biology, speciation. The Brassicaceae are an important family for three primary reasons. First, the family includes several crop plants grown worldwide, some of which have been cultivated since prehistoric times. Vari- ous species are grown for oil, mustard condiments, forage and fodder for animals, or as vegetables (Crisp, 1976; Simmonds, 1986). The most т tant members belong to the genus Brassica L., cluding varieties of B. oleracea L. (broccoli, dol sels sprouts, cabbage, cauliflower, kale, kohlrabi, savoy), B. juncea (L.) Czern. (Indian mustard), B. nigra (L.) W. D. J. Koch (black mustard), B. napus L. var. napobrassica (L.) Rchb. (rutabaga). B. napus var. napus (rape), and B. rapa L. (summer turnip rape, Chinese mustard, Chinese cabbage). Other lo- cally important crops are Lepidium sativum L. (cress), L. meyenii Walp. (maca), Armoracia rusti- cana P. Gaertn., B. Mey. & Scherb. (horseradish), Raphanus sativus L. (radish), Sinapis alba L. an le Cochlearia officinalis L. and Cardamine amara L (bittercress), Eruca vesicaria (L.) Cavan. var. sativa (Mill.) Thell. (rucola or erugula), and Eutrema wa- sabi (Siebold) Maxim. (wasabi). Second, many spe- cies of the genera Aethionema R. Br., Alyssum L., Arabis L., Aubrieta Adans., Draba L., Erysimum L., Hesperis L., Iberis L., Lobularia Desv., Lunaria L., and Matthiola R. Br. are cultivated as ornamentals (Al-Shehbaz, 1984). Third, Arabidopsis thaliana eynh. (thale cress) is considered to be the perimental work in various fields of biology, in- cluding plant genetics, physiology, development, pathology, genetic engineering, and related fields. In recent years, A. thaliana has become a mode system for plant molecular biology, which has cul- minated in the recent publication of its complete genomic sequence (The Arabidopsis Genome Ini- tiative, 2000), also available via the World Wide Web —((http//www.nature.com/genomics/papers/ a-thaliana.html)). Arabidopsis thaliana has a small genome, is fast growing, easy to cultivate with min- imal space and care demands, and self-pollinating ! We are grateful to C. Donovan Baley, Neil A. Harriman, Thomas G. Lammers, Karol Marhold, and two anonymous We thank Victoria Hollowell n et editorial advice and Mark ? [nstitute of Botany, University for Agricultural Sciences Vienna, Gxegor-Mendel- учан 33. A-1180 Vienna, Austria. koch (edv 1.boku.ac.at. que Aa en Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. ihsan.al- shehbaz@mobot.org. otanik, Universität Osnabrück, Barbarastr. 11, ^ эреле ot EE Uni-Osnabrueck.DE D-49069 Osnabrück, Germany. Mummen- ANN. Missouni Bor. GARD. 90: 151-171. 2003. 152 Annals of the Missouri Botanical Garden with a high fecundity (Meyerowitz, 1989; Patrusky, 1991). Moreover, genetic modifications by transfor- mation are routinely accomplished to generate mu- tants or to over-express particular genes. Stock cen- ters. distribute seeds of different accessions and DNA libraries, YACs (yeast artificial chromosomes), and BACs mutants, mapping populations, (bacterial artificial chromosomes) for molecular analysis in genetics and development. A central question for this research program is how to apply knowledge gained from the model system (i.e., lab- oratory lines/strains of A. thaliana) to wild plant species. А wealth of information on A. thaliana is being assembled electronically (, Gene tics 156: 833-838. 2. Das System der Cruciferen. Oesterr. Jones, B. M. G. & J. R. Akeroyd. 1993. Arabis. In: T. G. \ ‚А. RAN p O. Chater, J. R. Edmondson, V. pardus D. H. каа S. M. Wal- 5 =m = = 2 æ = = © "rj m = a E = d c N 352-356. Саным TM ode Cambridge, New York Judd, W. S., R. W. Sanders & M. J. 1994. Donoghue. аа. aceae). Bot. d family pairs: dra ped phylogenetic analysis. Harvard Pap. Bot. 51. Kamm, A., I. Galasso, T. Se as, S. Heslop-Harrison. 1995. Analysis of a repetitive DNA family from Arabi- dopsis arenosa and к, — Arabidopsis species. Pl. Molec. Biol. 27: Kawabe, A., H. fan n, R. к à N. T. Miyashita. 1997. Nucleotide polymorphism in the acidic chitinase locus (Chi) го of the wild plant ак thal- iana. joe E vol. 14: 1303-1315 um, H.-G C. TH P. S. Vroom & 1998. MS ular evidence for an African origin of il waiian endemic Hesperomannia (Asteraceae). Pro Natl. Acad. Sci. U.S.A. 95: 15 145 n D. J., J. Kroymann, P. Brown, A. Figuth, D. sen, J. Doa nzon & T. Mitchell- Olds, 2001. Ge- netic one! of natural variation in Arabidopsis glucos- inolates. Pl. Physiol. 126: 811—825. Koc th, M. prealpir ne . Jansen. due Genetic differentiation and speciation in Cochlearia (Brassicaceae): Allohexaploid Cochlearia bavarica (Brassicaceae) compared to its dip- loid ance pe Austria. PI. & 1. е к иса in Germany and .E vol. gree т 2000. Molec ular systematics of the C OMM Yinshania (Brassicaceae): Evidence from plastid and nuclear ITS DNA sequence data. Ann. Mis- souri Bot. Gard. 87: 246-272. М . 2X 2: 35-4 z - Ч —— ecular data indicate com- plex intra- and interc кееш differentiation of Amer- ican чү (Brassicaceae). Ann. Missouri Bot. Gard. 89: 88-109. & m Hurka. 1999. [sozyme analysis in the poly- ploid complex Microthlaspi perfoliatum (L.) F. K. Meyer: Morphology, biogeography and evolutionary hisce Flo- ra 194: 33— ave umme nhoff. 2001. Thlaspi s. str. (Brassi- cac ene) versus Thlaspi s.l.: ipi un 'al and anatom- ical characters. in po light ay ITS data. Pl. Syst. pis 9-225. p )NA sequence —— pa 7 Isoelektrische Fok- A der reinheiten der Rubisco in Thlaspi Brassicaceae): Weitere pp auf eine Formengat- tung. Feddes Repert. 104: 371—381. . H. Hurka & K. M ELE 19906. Chloroplast DNA restriction site variation and RAPD-analyses in Cochlearia (Brassicaceae): Biosystematics and specia- tion. Nordic J. Bot. 16: 585-603. . M. Hurka. 1998a. Isozymes speciation and evolution in the pn Cochlearia L. Acta 111: 411—425. . Mummenhoff & H. em 1998b. Systematics and ponia history of heavy metal tolerant Thlaspi caerulescens in Western Europe: Evidence from genetic studies based on isozyme analysis. Biochem. Syst. € Ecol. 26: 823-838. & — — & ——— . Molecular biogeog- ache and evolution of the Mica ж сш s.l. polyploid complex (Brassicac oplast DNA and nuclear pae s restriction site variation. Сапай, J. euh 1€ Bishop М Т. а. 0145. 1999a. xe не systematics of Arabidopsis and Arabis. PI. Biol. 531. . Mummenhoff & H. iie 1999b. Molecular TS netics of Cochlearia (Brassicaceae) and allied genera based on nuclear ind ITS DNA sequence Volume 90, Number 2 2003 Koch e al. 165 га d of Brassicaceae Research analysis contradict traditional concepts of their evolu- попа “| Жоки 1. Syst. Evol. 216: 207—230. —— ———, В. Haubold & T. Mitchell- uem 2000. Compar- ative pen ы analysis of chalcone synthase and al- cohol dehydrogenase loci in permets Arabis, and re- lated genera (Brassicaceae). Molec. Biol. Evol. 17: 1483-1498. А ———.2001а of the Brassicaceae: malK and 534—5 es . Molecular systematics Evidence from coding plastidic nuclear chs sequences. Amer. J. Bot ———, Weisshaar, J. Kroymann, B. Haubold & T. Mitc "a Pn 2001b. ne di phylogenetic anal- ysis of the CHS and apetala3 promoter reveals new clues for pere function determination. Molec. Biol. 891. Evol. 18 uthmann & K. G. Bernhardt. 2003a. Pop- arian EE s of amara L. (Brassicaceae) in dynamic habitats: Genetic музы of seed bank and p populations. Basic Appl. Ecol. (in press). ‚ €. Dobeš & T. Mitchell- Olds, 2003b. ерлш liybrid ed in natural populations: Concerted ev ution of the internal transcribed spacer of nuclear ri- bosomal DNA (ITS) in — Алене ап Arabis aivan- Brassicaceae). Molec. › T. H. Lan, K. erson. 1994. € lomparative mapping of Arabidopsis thal- iana and Brassica oleracea chromosomes reveals is- ңа of conserved organization. Genetics 138: 499— ^ x ч es | 7T х ч — Ыы, м. 2002. Vergleichende Biogeographie europäisch- er Gebirgspflanzen: Molekular Untersuchungen an der r montana 1 e und morphometrische nontan-subalpinen Anthyllis „ (Fabaceae) und der hochalpinen Pritzelago 1 (L.) О. Kuntze (Brassicaceae). Doctorate Disser- Mainz renis Cuvillier Verlag Góttingen. ew H. & M. Aguade. 2000. Nucleotide variation al the chalcone isomerase locus in Arabidopsis thaliana. Genetics 155: ا‎ Kusaba, M., K. . Hendershot, J. Vreba B. Nasrallah & ‘Nasrallah. 2001. нигин in the pend Агаар sis: Characterization of the 5 loc H the outcrossing A. P rata i ч быа. ге ie sic PL Cell a 627-6 Lagercrantz, parative 1 mapping between Ar- a nigra indicates that h extensive ge- nome replication accompanied by chromosome fusions and des rearrangements. 150: rassica. genomes Genetics D. Lydiate. 1996. ут теч genome map- ping in Brassica. Genetics 144: 1903-1910. ‚ J. Putterill, G. Coupland & p. Lydiate. 1996. pes mapping in Arabidopsis id Brassica, fine scale ome colinearity and congruence of genes con- Dollar жайда time. Pl. J. 9: 43-20. Lakshmikumaran, M. & M. S. Negi. 1994. toja tural anal- ysis of two length чо. of be | pales nic spacer from Eruca sativa a 24: € )21. Lan, T. H. & А.Н. Paterson: о е тта mapping of quantitative trait loci sculpting the cu oleracea. Genetics ae 1927-1954. . Comparative mapping of QTLs determine the piant s size e Brassica oleracea. Theoret. Appl. Genet. 103: 3 ‚ T. A. Del Minis. с Р. Reischmann, J. [ps B.E Кона, J. McFerson, S. Kresovich & А. Н. rd of Brassica erson. 2000. An EST-enriched comparative map of Brassica oleracea and Arabidopsis thaliana. Genome Res. 10: 776—788. Lannér, C. 1998. Relationships of wild Brassica species with chromosome number 2n — 18, based on compar- ison of the DNA sequence of the chloroplast intergenic region apes trnL (UAA) and trnF (GAA). Canad. J. Bot. 76: 2 : ее, J.-Y., 3 "Mu mmenhoff & J. L. Bowman. 2002. Allo- ploidization and evolution of species with reduced floral structures in uà dium L. Natl. cad. Sci. U.S.A. (in press). Lenoir, A., B. саса S. Warwick, - Does 1997. Evolution of SINE Cruciferae plant species. Molec. th 941. (Brassicaceae). Proc. "di Picard & J.-M. retroposons in Evol. 14: 934— Les, D. Н. 1994. Molecular systematics and taxonomy of lake cress (Neobeckia aquatica; Brassicaceae), an im periled d mustard. Aquatic Bot. 49: 149-165. Lihová, J., K. Marhold & B. Neuffer. 2000. Taxonomy of Cardamine amara (Cruciferae) in the Iberian Peninsula. Taxon 49: 747-763. Liljegren, S. J., G. S. Ditta, Y. Eshed, B. Savidge, J. L. Bowman & M. F. Yanofsky. 2000. Shatterproof MAD- »ox genes control seed dispersal in Arabidopsis. Nature 404: 766—770. inde, M., S. Diel & B. Neuffer. 2001. E Capsella bursa-pastoris (L.) Medik. Pee ac His an- alysed by a cosegregation of phenotypic character (QTL) and molecular markers. Ann. Bot. (Oxford) 8 91-99, Lowman, A. C. & M. D. Purugganan. 1999. Duplication of the Brassica oleracea APETALAI floral homeotic gen and the evolution of domesticated cauliflower. J. Hered. x y . Vekemans terns il y s & A. L. Jacquemart. 1999. Pat- »ymie variation within Calluna vulgaris populations at seed bank and adult stages. Heredity 82: 32—440. Maluszynska, J. & J. S. Heslop-Harrison. 1993. Physical mapping of rDNA loci in Brassica species. Genome 36: 774-7 1 Marhold, K., M. Huthmann & Н. Hurka. 2002a. Evolu- tionary history of the polyploid complex of puri amara (Brassicaceae): Isozyme evidence. Pl. Syst. Evol. 233: 15—28 ‚ J. Lihová, M. ‚ В. Grupe & B. Neuffer. 2002b. jw Я in Cardamine LT ceae) in the € : Evidence from iege og and molecular die Bot. J. Linn. S« 5-294. Martin, J. P. ). Sanchez- Yelamo) hte re- lationships among species of the genus Diplotaxis (Bras- sicaceae) using inter-simple i re Theor. Appl. че net. 101: 1234-124 Mayer, M. S. & P. S. Soltis. 1994. The evolution of ser- pentine endemics: "S chloroplast DNA phylogeny ү: ч Streptanthus glandulosus complex (Cruciferae). Bot. 19: 557-574. & peat. markers. & | Soltis. 1994. The evolution of the Streptanthus glandulosus complex (Cruciferae): G netic divergence and gene flow Amer. J. Bot. 81: 1288-1299. McCue, K. A. & T. P. Holtsford. 1998. Seed bank influ- ences on genetic diversity in the rare annual Clarkia m (Onagraceae). Amer. J. Bot. 85: 30-36. i ‚ J. Bishop, J. Z. Lin, J. Н. Richards, A. Sala ore hell- Olds. 2001. Local adaptation across a e- w in serpentine endemics. 166 Annals of the Missouri Botanical Garden climatic gradient despite small Es ‘live Mp end size in the rare о _госКсгевв. . Roy Biol. :i., Ser. B 268: 17 721. Meekins, J. F., Н. Е. Ballari & B. C. McCarthy. 2001. Genetic variation and molecular biogeography of North American invasive n bee ies (/ oe etl: ata, v 'aceae). Int. J. Pl. . 162: 161-16 Meyer, F. K. 197 WM он der “Thlaspi’ Але Ки- ropas, “Aik pm e Feddes Repert. 84: 449-47 . v 9. кина һе Europas, Afrikas und V Geschichte, Mor- phologie und Chorologie. Feddes Repert. 90: 129-154. |. M. 1989. Arabidopsis, a useful weed. Cell Revision der * Vorderasiens. 1. "Thlaspi"^-Arten Mitchell, A. p. & P. B. Heenan. 2000. Systematic rela- tionships of New Zealand endemic Brassicaceae in- ferred е nrDNA ITS sequence Pes Svst. Bot. 25: 98-10 —— — & ——. 2002. Genetic variation within the Pa- pus (Brassicac ban) Soa based on oo епі AFLP data. J. Коу. Soc. New Zealand 32: 427—443. b DNA variation in the F upstream є t locus of the wild plant Arabidopsis thaliana and Araba gemmifera. Molec. 18: Biol. Evol. 64-171 nan & R. Terauchi. 1996. Intra- and inter- specific variation of the alcohol dehydrogenase loc region in wild api Arabis gemmifera and , e E thaliana. 136. е Biol. Evol. 13: 433— ‚ H. Innan & R. Terauchi. 1998. Intra- and омор fic . DNA variation and codon bias of the me diis ا‎ Adh) locus in з Arabis and Ara- 429. bidopsis spec ies. . Biol. Evol. 15: 1420- Moment 089, ee in der Gattung Lepidiur (Bi rassicaceae): Tsoelektrische Fokussi- erungsmuster der | nte тонер von Ribulose-1,5-Bis- phosphat ( (RUBISCO) als sys- und ы кын ай Marker. Ph.D. Thesis, e оГ Osnabrück, Osnabrück. . 1995. Should Cardaria draba (L.) Desv. be clas- Шей wae the genus Lepidium L. (Brassicaceae)? Ev- idence from bani E pride composition of RUB- ISCO. Feddes Repert. 106: 25-28. — ——— & H. Hurka. 1990. Evolution of the рх Capsella bursa-pastoris (Brassicaceae): Pur есіле fo- cusing analysis of Rubisco. Pl. Syst. Evol. 172: 205- 213. —— & 99]. Isoelectric focusing analysis of RUBISCO in loudium (Brassicaceae), sections P Lepiocardamon, and Cardamon. Biochem. Syst. & Ecc 19: 47—52 —— — & — ———. |994. Subunit polypeptide composi- tion of Rubisco en the origin of i Lie Arabi- dopsis suecica (Brassicaceae), Biochem. Syst. Ke 312. 22: 807—< Ф & —— — Allopolyploid origin of Arabi- dopsis suecica me d аш onmorphogenic mutants. J. Evol. Biol. ‹ Vx mel & J. Schmitt. 1999. Evolution of phenotypic ee A comparative approach in the е neighborhood of Arabidopsis thaliana. J. Evol. Biol. 12: 779-791. pA U. 1995. Molekulare Untersuchungen zu realgeschichte von Arabis alpina L. (Brassicaceae) in Europa, Vorderasien und Afrika. Ph.D. Thesis, Univer- sity of Osnabrück, Osnabrii Pollard, H., M. Cruzan & M. Pigliucci. 2001. Comparative studies of reaction norms in Arabidopsis. 1. Evolution of response to NE Evol. Ecol. Res. 3 1: S. Prakash, А. Mukhopadhyay & D. Pen- DNA patterns: EE x чиш classifications K. Prantl (adios), Die natürlichen уны e Vol. Ш,2 чы von Wilhelm Engelm eipzi Price, R. A. & I Al-Shehbaz. 2003. етт (Brassicaceae), а reek ыа case of fruit modification in و‎ Ann Missouri Bot. Gard. (in press). . D. Palmer & I. A. Al-Shehbaz. 1994. System- atic a of Arabidopsis: A molec “Шаг and mor- phological perspective. Pp. 7-19 in Meyerowitz & C. Somerville (editors), Arabidopsis. Cold Spring Har- bor — dA Ps New York. Purugganan, M. D. & J. I. Suddith, 1998. Molecular pop- ulation eds of the Arabidopsis CAULIFLOWER reg- ulatory gene: Nonneutral evolution and naturally occur- ring variation in floral homeotic function. Proc. Natl. Sci. U.S.A. 95: 8130-8134. 9. Molecular population genetics of floral bande loci: Departures from the equilibrium- neutral model at the APETALA3 and PISTILLATA genes of ene a thaliana. Genetics 151: 839 L. Boyles & J. I. Suddith. 2000. Variation and E at the CAULIFLOWER floral homeotic gene accompanying the evolution of domesticated Brassica oleracea. md 155: 855-862. Quiros, C. F., F. Grellet, J. Sadowski, T. Suzuki, G. Li & T. Wroblewski. 2001. Arabidopsis and Brassica compar- lated au pese 157: 1321-1330 D. ladie Ann. Век . Ecol. Syst. эш 467—501. ask, L., E. а ndréasson, B. . S. Eriksson, B. Pon- po dene & J. Meijer. 2000. Myrokinase: Gene family volution and herbivore defense in Brassicaceae. РІ. Molec. Biol. 42: 93-113. Reboud, X. & C. Zeyl. 1994. Organelle inheritance in plants. Heredity 72: 132-140. Richardson, J. E., F. M. Weitz, M. F. Fay, Q. B. C. С H. P. Linder, E. Reeves & M. W. De 2001. E Bi recent origin of species richness in the Cape flora of South Africa. ipn 412: 181-183. Rieseberg, L. H. & C. R. Linder. 1999. Hybrid classifi- cation: Insights froni des p mee ce of ex- ават аа Ecology 80: : Sinervo, C. R. Li м T Feb! M. NA 1995. Hybrid speciation accompanied by geno- mic reorganization in wild sunflowers. Nature 375: 313— 316. —— J. Whitton & К. Gardner. 1999. Hybrid zones and the genetic architecture of a barrier to gene between two sunflower species. Genetics 152: 113-727. Rodman, J. E., R. A. Price, K. Karol, E. Conti, K. J. 168 Annals of the Missouri Botanical Garden Sytsma & J. D. Palmer. 1993. Nucleotide sequences of ne lee gene indic а еи of mustard oil plants. . Mi ya ri i Bot. "e^ 386—699. кз ae d ane 1994. Nuc quon sequences s of rbcL confirm the cap- paralean affinity of the Australian endemic Gyroste- Austral. Syst. Bot. 7: 57—69. SS ت‎ K. . Sytsma. 1996. Mole- c ев, morphology, and Dahlgren’ s expanded order Cap- . 21: 289-307. monaceae, parales. £ m 1: Rollins, R. The Cruciferae of Continental North America. Е Univ. Press, Stanford. Rossberg, K. Theres, m carkan, R. Herrero, T. Sc диши. K. Schuhmacher, G. Schmitz & R. Schmidt. 2001. C omparative sequence кы sis eiae: exlensive microcolinearity in the lateral suppre (саш; Arabidopsis, and Capsella genomes. PL. Cell 13: sor regions of the Row a 1995. The breeding systems of six species of Ar- abis. Amer. J. Bot. 82: 86° -877 . 20 p atterns of association between crucifers < and thei ‘ir flower mimic pathogens: Host ju common = co-evolution or co-speciation. Evolution : 4l- mps are more & [à н. e 1989. Evidence for apomixis in Ara e J. Hered. € RO Ryder, C. D Ja 1. л. 2( ОК SARE в genome organisatic the Brassica oleracea genome a with collinear regions of the / ган thaliana genome. Genome 44: 808—817 Saal, B., J. p lieske, J. Hu, C. F. Quiros & D. Struss. 2001. Microsatellite markers b genome analysis in Brassica. А and C genomes and genetic mapping in Brassica oleracea. Boon Appl. Genet. 102: 695—699, Sadows & i S. 1998. Organization. of an Malone файла gene cluster on chromosome 4 in- cluding the RPS2 gene in the Brassica nigra genome. xo Genet. 96: 408—474. aubier, M. Delseny часы < G. J. King. : Two regions of . Assignment of rapesee a. microsatellites to the A Theor. & C. F. Quiros. 1996. Ge snetic ‘on physical mapping in Brassica diploid spe- cies of a gene cluster de y in Arabidopsis thaliana. ven, Genet. 251: 298—306. M. J. 1997. A non e tric approach lo es- sence of rate con- 232 ~ olec Sanderson, timating dive 'rgence Sa in th stancy. me lec. Biol. Evol. . 1998. Estimating rate к= time in У ‘шаг phy- loge nies: Ве s the molec ilar clock? Pp. 242-264 in Р. S. Soltis, E. E. Soltis & J. J. Doyle (e dos Molec- ular Сынык и i Plants П. Кине Boston. . 2002. Estimating absolute rates of molecular e TE lution and divergence ge ^k d likelihood ap- proach. igi Biol. Evol. 01-109. Se [ia A.-C. Elven & ^ к hmann. 2002. lecular- ec al approach solves taxonomic нң troversy in arctic Draba (Brassicaceae). Canad. J. 30: 59-71. Se = тир, M. Н. 1998. The effect of enzyme heterozygosity n growth in a strict elf-in- A Arabis i hin (Brassicaceae). 128: 21—31. R. 2000. Synteny: Recent advances and future Pl. Biol. : 3: 9 1-1 )2. ‘a 2001. C Wow struc ا‎ genomics in the alr 'aceae family. Pl. Ph siol. Bioc he m. 39: 1—11 A mo- outcrossing species, the s Hereditas Schmidt, prospects. Curr. Opin. 1 A= = z = © = = 5 an & Stotz, H., J. I Schulz, О. E. 1936. Cruciferae. Pp. 227—658 in A. Engler & i Prantl (editors), Die natürlichen Pflanze = n, vol. P. А е von Wilhelm care Leip Sharbe 1 Mitchell-Olds. . Recurrent (mM ploid origins ge КООР variation in the Arabis holboellii ыи) pie rassicaceae). Here dity 87: 59-08. —., В. Haubok > Mitchell- Olds. 2 2000. Genetic isolation by eese in ae idopsis thaliana: Biogeog- i ws post- Ces чаі colonization of Europe. Molec Ecol. 99-2118. Rd D. I. A. P. Parkin, R. Mayerhofer, D. J. Lydiate & . G. Good. 2000. Arabidopsis thaliana: A source of cani date eru resistance genes for Brassica napus. Genome : 13: —400. Simillion, C., E ae ele, . Van Montagu, M. Zabeau & Y. de Peer. 20 i hidden Aa gs a- n past of, jid thaliana. Proc. Natl. Acad. Sci. S.A. 99: 27-13632. МАНА N. en ses 1986. Evolution of Crop Plants. Harlow. &W Longman, V. 1995a. variation within and among different cultivars and landraces of and B. Wei wea L. тыйа, 91: 95b. ne isozymes in Band campe stris T and genetic divergence among m 'rent = ies of Brassiceae. Theor. Appl. Genet. 91: Simonsen, eneen. Genetic Brassica campestris L. based on iso- zymes. The Or. 6-35: 53-36 T D. Е & P. S. Soltis. 1993. Molecular ган апа EE eed nature of polyploidy. C.R.C. Crit. Rev. РІ. S 2: 243-273 é . 1999, Polyploidy: i^ 'urrent formation a sia genome evolution, Trends Ecol. Evol. 14: 348— Soltis, P. S. & D. E. Soltis. 2001. GE pi ular systematics: a and using the tree of life. Taxon 50: 663- 671. Song, K.. P. Lu, К. Tang & T. Osborn. 1995. Rapid ge- nome change in synthetic polyploids of Brassica and its i alions for Pe ue evolution. Proc. Natl. Acad. Sci. U.S.A. 92: -1123. Stehlik, Ll. А. Tribsc D & P. Schónswetter. 2001. grin gemeinsames Meeting zur Phylogeographie von a tischen und alpinen Pflanzen in Zürich, 1—3 Juni 2001. Bauhinia 15: 69—90. Sishop, С. W. Bergmann, M. Koch, P. Albe she A. G. Darvill & J. M. Labavitch. 2000. Evidence for divergent selection of amino acid sites in the evo lution of fungal ное к their En in- hibitors. Physiol. Molec. Pl. Path : 117-13 Strauss, S. Y . Siemens, M. B. n her & T. vind hell- Olds. 1990, йн 'al costs of plant resistance to her- bivores in the currency of pollination. Evolution 53: 1105-1113. Sweeney, Р. W. & R. A. Price. 2000. Polyphyly of the genus Баша (Вгаѕѕісасеае): wap nce m trnk in- tron ae p sequence E . Syst. Bot. 25: 468—478. Syvanen, 02. Rates of hea RNA КОЛ аге unique i о in Eukaryotes. J. Molec. Evol. 55: 85-9]. Szymanski, D. B., | Lloyd & M. D. ا‎ 2000. Progress in ede genetic analysis of tric nca initiation and үйөр анн} їп, Arabidopsis Trends I Sci. 5: 214-219 Thompson, J. D. & . Lumaret. 1992. The evolutionary dynamics of polyploid plants: Origins, он nt ds Ecol. Evol. 7: 302— M. E А and ии e. Trenc › Dehal, F. Jarrin, J. Hu, Tole do, J.. | Volume 90, Number 2 2003 Koch et a 169 Review of Brassicaceae Research Shehbaz & C. К. Quiros. 1998. Genetic dcc il of pn ns enii (Brassicaceae) -i y molecular ar Aie (Oxford) 82: 523—53 lisz, J. Fischer & T. P. Heind 1993. struc- Tonsor, S. p x bur je Dr study of population genetic s e: Seed bank to adults in Plantago lanceolata. Evo- lution 47: 833—843. Mi y a , €. König, R. Samuel, W. Pinsker & T. „ 2002. SS genetic variation in Bis- cule ila pics New focus d Irene Manton's theory. Р! уо]. 2 eee 18 Tsukaya, H., J. Yoke , Н. а da, Н. Kuroiwa, T. Ku- roiwa & K. Iwats Mig 1997. Morphological, physiologi- cal and molecular genetic characterization of ف‎ sis тайга, with reference to A. thaliana. J. Pl. H 110: 15— Urbanska, . H. Hurka, E. Landolt, B. Neuffer & K. оодоо: 1997. Hybridization and evolution in Cardamine (Brassicaceae) at Urner Boden, Central Switzerland: Biosystematic and molecular evidence. РІ. Syst e 204: 233—256. Uyenoya na, M. К. 2000. Evolutionary dynamics of self- inc e alleles in Brassica. Genetics 156: 351— 359. Van der Zwan, C., S. A. Brodie & J. J. Campanella. 2000 n, K. Karkkainen, E. -Baena- 1997. Ееее of micro- satellies in Arabis petraea газі Arabis lyrata, outcrossing relatives of Arabidopsis thaliana. Molec. Biol. Evol. 14: 220-229 Baldwin & L. Constance. 1998. Nuclear uth American species of San- rg (Apiaceae). Proc. Natl. Acad. Sci. U.S.A. 95: 235- T T. J., D. G. Brown & S. D. Tanksley. 2000. The origins of Pav duplications in Arabidopsis. Science 290: 2114 ds Walther, G. n E Post, P. Convey, A. Menz mesan, T. J. B e, J.-M. Fromentin, О. Hoegh- Guldberg & F. Bairlein. 2002. e responses to Nature 416: 3 . Al-Shehbaz “1998 Generic чи - el, C. Par- 99]. Molecular systematics of Brassica and allied genera (Subtribe Brassicinae, Bras- siceae)—C ge p and cytodeme congru- ence. Theor. ages t. 82: 81-92. ——— & ——— B de ds relationships in sub- tribe Все oia. tribe Brassiceae). Canad. J. Bot. 71: 906-918. Evaluation of the subtribes Mor- e Savignyinae, Vellinae and Zillinae (Brassi- ceae, tribe Brassiceae) using chloroplast DNA restric- n site variation. Can: E J. Bot. 72: 1692-1701. & 1997 Phylogenetic HR нат of po DNA restriction site variation in subtribes Raphaninae and Cakilinae ии еае, bê Brassi- ceae). MUN J. er E 960-9 Molec us phylogenies from Teo to m ation in ая and allies (Tribe Bras- siceae, Brassicaceae). Opera Bot. 132: 159-168. & I. E арыы 1992. Molecular sys- ON of Brassica and allied genera (Subtribe Bras- sicinae, Brassiceae)—Chloroplast DNA variation in the genus S ien Theor. Appl. Genet. 83: 839—850. 1. A. ehbaz, R. A. Price & C. Sauder. 2002. na of Sin mbrium (Brassicaceae) based on ITS sequences of nuclear ribosomal DNA. Canad. J. Bot. ors, E. R. & B. A. Schaal. 1996. Biased gene conver- sion is not occurring а rDNA repeats in the Bras- sica triangle. Genome 39: 150-154. Widmer, A. & M. Baltisberger. 1999a. Molecular evidence for allopolyploid speciation and a single origin of the narrow endemic Draba ladina (Brassicaceae). Amer. J. Bot. 86: 1282-1289 . Extensive кы сы chlo- roplast DNA (cpDNA) bn in the e Draba ai- zoides L. Реасе и relationships and population structure. . Ecol. 8: 1405-1415. Yang, Y.-W., W. Н. J. = es T. H. Won oe Genetic polymorphism of seven populations of Gapeella bursa- pastoris on RAPD markers. Bot. Bull. Acad. Sin. 39: 7-21. i, P.-Y. Tai, D.-P. Ma & W.-H. Li. TE‏ س ИА Г: studibe of Brassica, Rorippa,‏ abidopsis and allied genera based on the internal tran-‏ scribed spacer ue of 185—255 rDNA. Molec. Phy-‏ logenet. Evol.‏ & W- H. th 1999b. Rates of кйш де йена in angiosperm mitochondrial А sequences in angiosperm жк DNA se- quences and dates of divergence between Brassica and other angiosperm lineages. J. Molec. Evol. 48: 597— 604. Zunk, K., K. Mummenhoff & H. Hurka. 1993. pesi jm уер и site variation in the puse m eae, tribe L idieae. Pl. Molec 2 Newsl. 3: 4 , och & Н. E Dus Phyloge- netic Mlutisnalipe d Thlaspi s.l. (subtribe Thlaspidi- nae, Lepidieae) and allied genera based on chloroplast DNA restriction-site variation. Theor. Appl. Genet. 92: 375—381. & —— —. 1999. Phylogenetic relation- Med in tribe Lepidieae ы ы based on chlo- roplast DNA restriction site variation. Canad. J. Bot. 77: 1504-1512 APPENDIX 1. MOLECULAR MARKERS IN THE STUDY OF SYSTEMATICS AND EVOLUTION OF BRASSICACEAE: A LITERATURE SURVEY The following list summarizes studies in which various methods have been used to study the evolution and sys- tematics of Brassicaceae. The compiled literature is sorted according to the marker systems used, listed alphabeti- cally by taxa and then by authors. The ni is restricted to studies at the populational level or abov I. Isoelectric focusing of ribulose-1, 5- bisphosphate carboxylase (IEF-Rubisco): Arabidopsis including y ipeo dia (C. A. Mey.) Hayek (Mummenhoff & Hurka, 1994). Capsella bursa-pastoris aegros & Hurka, 1990). Diplotaxis (Mummenhoff et al., 1993). Erysimum (Mummenhoff & rx na 1994). Lepidium (Mummenhoff, 1989, 1995; oo & Hurka, 1991; Mummenhoff et al., Thlaspi (Koch et al., 1993; ашыгы & Zunk, 1991). II. Isozyme analysis: 170 Annals of the Missouri Botanical Garden Arabidopsis including Cardaminopsis (Mummenhoff & urka, 1995). A. on subsp. petraea, as Arabis petraea L. (Schierup. 298). Ps (Roy & Rieseberg, 1989). A. fecunda (McKay et t al. 2001). A. serrata Franch. & Sav. (Ovama et al., Biscutella Was "Ber el al., 2003). Boechera A. & . as Arabis (Roy, 1995; Roy & Rieseberg, 1989 y Brassica (C һе vre et al., 1995a, 1995b B. insularis Moris (Petit et a E ). Brassicaceae (Chevre et al., 1995). те 'eae (Anderson & is k, 1999: Simonsen & . 1995a, 1995b). Capsella (Hurka, 1990; Hurka & pur Lo Hurka Neuffer, 1997; Neuffer & Hurka, 39). C. bursa-pastoris (Neuffer, 1996; f Kui 29). 1993). 1995; Simonsen & Непееп. = 1999; Neuffer & Hurka, 1999 Кеше r et al., rubella (Neuffer & Hoffrogge. 199 О). : ardamine (Urbanska et al., 1997). C. amara (Koch et al., 2003a; Marhold x 2 2002a). C. pratensis L. agg. (Franzke & Hurka, 2000). Ке Cochlearia (Koch et al., 1998a; Koch, 2 к) d (Brochmann, 1992; Brochmann et үз 1991, 92а, 1992b, 1992c; Scheen et al., Foss rella fendleri (A. Gray) S. Watson (C бї, 1996). Microthlaspi (Koch Б Hurka, 1999). Nasturtium (Bleeker et al., 1999). Raphanus sativus (Huh & a 2001). Rorippa (Bleeker & Hurka, 2¢ Tod: ple Nutt. (Mayer el a, 94). Thlaspi caerulescens J. & € ud the th et al., Warea carteri Small (Evans et al. 0). П. riction fragment uu по ЕЕЕ of (a) chloroplast (cp) DNA: Arabidopsis, ine e Cardaminopsis (Mummenhoff & Hurka, 1995; Price et al., 1994). Arabideae et D 1994). Brassica (Song et al., 1995; Warwick & Black, s 1993, 1994, ur 1997b; Warwick et al., 1992), Brassicinae (Warwick & Black, 1991, 1993, Le 1997a, 1997p). cioe Agr (Pradhan et al., 1994, 1997a). E inae (Wars ick & Black. pein (Urbanska et - Diplotaxis (Warwick et al., "d Draba (Broc чөерә et al., 2c). Lepidieae (Zunk et al., 1993; (po et al., Lepidium (Mummenholl et al., 1995). Microthlaspi (Koch et al., 199 ^ : Moricandiineae (Warwick & Black, 1994). Raphaninae (Warwick & Black, 1997a). Savignyinae (Warwick & Black, 1994). Sisymbrieae (Price et al., 1994). Streptanthus (Mayer & Soltis, 1994). Thelypodieae (Zunk et al., 1996). Thlaspidineae (Zunk et al., 1996). Thlaspi (Mummenhoff & Koch, 1994; Mummenhoff et al., 1997a, 1997b; Zunk et al., ) Vellinae (Warwick & Black, 19‹ Zillinae (Warwick & Black, 199. 4). (b) nuclear ribosomal DNA Brassica (Delseny et al., 199( m Malin t. & Heslop- Harrison, 1993; Waters & Schaal, 1996). 1998b). 1992; Warwick & Black, as 997). 1999). Eruca Mill. (Lakshmikumaran & Negi, 1994). Microthlaspi (Koc i et al., 1998c). NA (c) total nuclear I . Random amplified polymorphie DNA (RAPD): Capa bursa-pastoris (Neuffer, 1996; Neuffer et al., 1999; Yang et al., 1998). Cardamine (Neuffer Җ Jahncke, 1997; Urbanska et al., 1997). C. amara (Lihová is al.. 2000). C. pratensis agg. (Franz ác 5 Hurka, 2000). Cochlearia (Koc n et "s 26). = Draba (Scheen et al., m Lepidium me yenit md a RA al., 1998). Amplified fragment length polymorphism Arabidopsis thaliana (Sharbel et al., 2000). Ca rdamine amara orir ai et al., 2002b). Cheesemania О. E. Schulz, Pachycladon Hook. f. (Mitche Т "к Haas 2002). repetitive DNA and microsatellit Alliaria petiolata (M. Bieb.) Cavara & кам (Meekins et al., 2001). Mr eer including oo (Kamm et al., 1 der Zwan et al., 2000). A. Е АЕ petraea (van Treuren et al., 19€ ж (Harrison & Heslop-Harrison, 1995; Bux]: et al., 01). жоо (Martin & Sanchez-Yelamo, 2000). VII. (А Sequencing o d c ылы, реци genes: aturase Brassic 'üceae (Кос th et al., 2001а). idhF: Cardamine, including Dentaria (Les, 1994; Sweeney & Price, 2000). rbcL: Arabidopsis (Price et al., 1994; E et al., 1997). Capparales (Rodman et al., 1993, 1996). Armoracia, Nasturtium, Neobeckia, гаса Les 0999. (b) non- оса plastid trnL intron and spacer Arabis (Roy, 2001). A. holboellii Shu & Mitchell-Olds, 2001). Boechera (Roy )1). Brassica (Lannér, 1998). бачата, including Dentaria (Franzke p ү 1998; Sweeney & Price, 2000; Bleeker et al., 2002b). C. BN agg. (Franzke & Hurka, 2( T Caulanthus S. Watson, including Guillenia Greene (Pep- per & Norwood, 2001 Coe hlearia (Koch et al., 199 Jb). Draba (Koch & Al-Shehbaz, 2002). —. Halimolobos Tausch (Bailey et al., 2002). Lepidium (Mummenhoff et al., 20014). Mancoa Wedd. (Bailey et al., ue Pennellia Nieuwl. (Bai ey et al., Rorippa (Bleeker & Hurka, D. Гн eker et al., 2002a). Sphaeroc Vp Schauer (Bailey & Doyle, 1999; Bai- ley et al., 2002). Streptanthus T epper & Norwood, 2001). dien (Koch & Al-Shehbaz, 2000). uclear coding alcohol E (ADH): Vido (Miyashita et al., 98). 1. thaliana (Hanfstingl et "e үз) eae et al., daba Miyashita et al., 1996, 1998). 1996). E Volume 90, Number 2 2003 Koch e 171 al. е in of Brassicaceae "жан Brassicaceae (Koch et al., 2000). Leavenworthia Torr. eee et al., (d) nuclear coding S-allele Vi veo lyrata (Charlesworth et al., 2000). A. th a (Charlesworth eta 1998). e (Charlesworth et al., 000; Uyencyama, 2000). (e) nuclear coding cha fee ee (CHS): Brassicaceae (Koch et al., 2000, 2001a (i) nuclear codi Arabidopsis d (Kawabe et al., 1997). Arabis (Bishop m nuclear cadies chalco ا‎ sg including a шне (Kuittinen & Agu- ade, 2000). (К) nuclear coding floral homeotic genes (APETA- LA, PISTILLATA, CAULIFLOWE Arabidopsis бүрке echoes & Parurardic 1999; Pu- ‚ 1998, 1999). Brassica onum | ызы 1999; Purugganan et . 2000 @) “өз non-coding -— transcribed spacer of ribosomal DN Arabidopsis (O’ Kane et a "1997; ч et al., Arabis (Koch et al., 19993; Roy, 2001 Boechera (Roy, 2001; Koch et ы. 2003b). Brassica (Yang et al., 1999a, 1999b; Warwick et al., 02). 1999a). Brassicaceae (Heenan et al., 2002; Mitchell & Heenan. 2000). Loc Dentaria (Franzke & Mummen- f 9: Franzke et al., 1998; Bleeker et al., С. еен agg. (Franzke & Hurka, 2000). E including Guillenia (Pepper & Norwood, 2001; Warwick et al., 2002). Cochlearia (Koch et al., 1999b). Crambe (Francisco-Ortega et al., 1999). A rrr Webb & Berthel. а. "e et al., 2000). ba (Beil stein & Windham, 2002; Koch & Al-Sheh- pe 2002; Widmer & Baltisberger, а 1999р). Dryopetalum А. > (Warwick et al., 2002). Erucastrum C. ce et al., 2002). Halimolobos (Bailey et al., 002). Mostacillastrum О. E. Sc rers (Warwick et al., 2002). Neotorularia Hedge & J. Léonard (Warwick et al., 2002). Pachyphragma (DC.) Rehb. (Mummenhoff et al., 2001b). Pennellia (Bailey et al., 2002). Anderson ex Hook. f. (Warwick et al., 2002). Pritzelago (Kropf, 2002). Rom vanschulzta O. E. Schulz (Warwick et al., 2002). Sibara Greene (Warwick et al., Sisymbrium (Warwick et a E Sphaerocardamum (Bailey & Doyle, 1999; Bailey et al., 2002). е Nutt. (Warwick et al., 2002). eptanthella Rydb. (Warwick et al., 2002). Hid ug (Pepper & Norwood, 2001; Warwick et al., 2002 The lypodiopsis Rydb. (Warwick et al., 2002). m Endl. (Warwick et al., 2002). Thlaspi (Koch & Mummenhoff, 2001; Mummenhoff et Vellinae (Crespo et al., 20 A Warea Nutt. ү ЖАРИ et al., 2002). Yinshania (Koch & AI- ii 2000). (m) chalcone synthase puer region: Brassicaceae (Koch et al., 2001b). (n) aleohol dehydrogenase promoter re Arabidopsis halleri (L.) O'Kane & Al- Shehbaz a oe етт mifera (Matsum.) O'Kane & Al-Shehbaz (Miyashi- la, A. Beine (Miyashita, 2001). (о) apetala3 promoter region кан-и асеае тҮ: et al., 2001b). (p) pistillat Halimolobos “Bailey et al., 2002). 2). a naan (Bailey & Doyle, 1999; Bailey et al.. 902). Daan (Lee et al., 2002). coding mitochondrial nad4: Arabidopsis (Yang et al., 1999a). Brassica Mn et al., 1999a, 1999b). mparative mapping approaches, segrega- ая (Acarkan et al., 2000; Vision et al., 2000). ca or Brassica-Arabidopsis (Axelsson et al., 2000, 2001: Ma et al., 1996; Cavell et al., 1998; Con- ner et al., Lagercrantz, 1998; jim cur n оз 1996; Lan Pi dc 2000, 2001; ves et al., Rossberg et al.. ; Ryder et a 01; oa, 2000; Schmidt et = 2001: Sillito ы uL 2000). Capsella bursa-pastoris (Linde et al., 2001). Capsella rubella (Acarkan et al., 2000; Rossberg et al., 20 Review (Se :hmidt, 2000; Schmidt et al., 2001). hort interspersed = (SINE ): Brassiceae (Lenoir et al., 1997). HOW BIG IS THE GLOBAL WEED PATCH?! Clinton N. Jenkins? and Stuart L. Pimm?? ABSTRACT Invasive spec ies are a major global pest to both biodiversity and agriculture and thus are a high priority for conservation science. Governments rec ognize this : and are devoting increasing resources toward solving the p Even so, there is inadequate Mn ku on en: invasives occur and thus where society can best use these resourc Disturbed areas tend to be very rable to i areas, the g nvasives, espec cially the weedy species. We map the world's dunt obal weed pate h, using maps of original and current landcover. At least 29.4 million km? (ca. 23%) of the world’s ice- “free land area is disturbed and thus favorable for invasive species. This weed patch map corresponds well to known locations of some of the setting geographic priorities Key words: i world’s worst weeds, le nding support to our approach, Our results should help in for actions against invasive species. disturbance, global change, invasive species, landcover, weeds. and been moved— Organisms have moved around the planet for millennia, but never in the numbers and with such speed as today. Most are benign, but a dangerous few cause major environ- mental problems. Invasive species may thrive in their new environment and dramatically change the dynamics and composition of the ecosystem. Be- cause of our lack of vigilance, we now suffer large economic losses in agriculture (Pimentel et al., 2000), suffer disrupted water supply and riverine transport (e.g., zebra mussel (Dreissena polymorpha) and Asian clam (Corbicula fluminea), Pimentel et al., 2000), and witness the degradation or even re- placement of entire ecosystems (e.g., purple loose- strife (Lythrum salicaria L.), Thompson et al., 198 2000). € results in а concomitant growth in the Pimentel et al., »rowth of the global trade network problem, and this is unlikely to change in the near future. As the problem is inherently global, we must develop a global strategy to solve it. Much invasive species research focuses on iden- tifying characteristics of species that enable them to invade. The idea is to identify who will invade and then restrict their movement around the world. That is certainly a vital approach to solving the problem, but it need not be the only one. This pa- per takes an alternative approach. We attempt to identify where they will invade rather than who the invaders will be. WHERE Do SPECIES INVADE? The simplest answer to this question is “every- where"! Consider the flora of Britain: It includes Polygonum amplexicaule D. Don from the Hima- layas, Dianthus caryophyllus L. from southern Eu- rope. Papaver somniferum L. from western Asia, Coronopus didymus (L.) Sm. from South America, ubus spectabilis Pursh from North America, and caena novae-zelandiae T. Kirk from New Zealand (Blamey & Grey-Wilson, 1989). show that when a flora is well known it is likely to These examples include species from around the world. Britain's species-poor native flora now accommodates a glob- al flora. Species can be introduced from anywhere That said, that some species are more likely to be introduced to anywhere, it would seem. it is clear than others, and some areas are more likely to re- ceive invasive species than others. Our argument is that disturbed areas are the prime habitats for invasive species. Recognizing the absence of an ideal database, we selected a set of species from the World’s Worst Invasive Alien (2001) to see if they occur predominantly in disturbed areas. The Species list compiled by Lowe et al. top three plants, and the number of countries in which they occur, were Lantana camara L. [51]. Chromolaena odorata R. M. King & H. Rob. [38], and Leucaena leucocephala (Lam.) de Wit [37]. The counts of countries are a little misleading in ' The National Science Foundation adn this work through a Graduate Research Fellowship for Clinton N. Jenkins. Rik Leemans and Bas Eickhou and Betsy von Holle for their E. n this work. 37996, : ent address: 13 Na East Veo Mic dicas 48824- 1222, 3 27708, U.S.A. StuartPimm@aol.com. ANN. Missouni Bor. GARD. 90: 2D e of and i Biology, University of Tennessee, 569 Dabney Hall, Knc U.S.A. P Resources, Department of Fisheries and Wildlife, M ‚ jenkins@msu.edu. Nicholas School of the bonum. and Earth Sciences, Duke University, Box 90328, Durham, North Carolina ssisted with the BIOME potential vegetation map. We thank Todd Campbell oxville, Tennessee Michigan State University, 172-178. 2003. Volume 90, Number 2 2003 Jenkins & Pimm Global Weed Patch that they are dominated by small, usually Pacific island nations. Nonetheless, all three species have invaded large areas. Chromolaena odorata, native to pore 'al America and the мыз (Holm et al., 1977: 212-216; Holm et al., 9: 85). invades pastures and croplands in ie g tropical Africa and Asia, the Southeast Asian island nations, and Australia. Lantana camara, native to Central and 1977: 299-302; Holm South iiie (Holm et al., 979: 207), is a widespread invader of pas- et al., ui cud in Southeast Asia, Australia (where it is a Weed of National Significance), parts of southern Africa and Madagascar, the Mediterranean, and ex- treme southern parts of the United States. Leucaena leucocephala is native to tropical America (Holm et al., 1979: 214; Lowe 2001), but in its non- native range can form dense, almost monospecific et al., stands that render large areas unusable. Currently it occurs in China, most of the Pacific Rim islands, and Australia. All three species occur in heavily disturbed ar- eas, as do many other widely distributed invasive species, such as Mimosa pigra L., various species of Opuntia, and Ricinus communis L. Another way to view this problem is to survey disturbed areas. Throughout the tropics, it is our experience that disturbed habitats will have at least one and often several of these listed species. Often they will be common; sometimes they will be the dominant spe- cies. Searching for “worst weeds" on the Wor Wide Web produces a list of repeat offenders that cause economic harm to croplands and pastures. Most are exotics, but not all. In short, these and other examples suggest that invasive species occur predominantly in disturbed By "disturbed, system changes, such as conversion to croplands, ecosystems. " we mean major eco- grazing lands, urban areas, or anthropogenic ones, such as grasslands where there was once forest. We will not further belabor this connection be- tween invasive species and human-modified habi- tats, for it is well-established and not obviously controversial. Rather, we accept the connection and move to the problem of estimating how large a frac- tion of the Earth's land surface humans have mod- ified. Since invasive species in these disturbed habitats are often deemed to be weeds, we can re- phrase our question to: How big is the global weed patch? We proceed by estimating the size of the weed patch at first global, then regional levels. Globally, the area of disturbed habitats is large. Regionally, we find that our global estimates are too small, for there is much disturbance that they miss. It is not surprising that invasive species are such an eco- logical problem. A FiRST ASSESSMENT OF THE SIZE OF THE WEED PATCH Global assessments of landcover change provide a series of somewhat overlapping estimates of dis- turbed habitats. A rough estimate of the size of the weed patch comes from combining the areas of (1) croplands, with (2) degraded grazing lands, and (3) the areas of cleared forests that are not in produc- live use. (1) Of the ice-free land surface of about 129 mil- lion km?, croplands cover about 15 million Кт? of the planet. All but 4 million km? were converted from naturally wooded or forested ecosystems (Pimm, 2001 (2) The Mes drylands cover roughly 61 million m2—an estimate that includes deserts, grasslands, shrublands, and savannas. The area varies from au- thor to author depending on what one means by “dry” (Pimm, 2001; Vitousek et al., 1986; Olson et al., 1983). Most of the world’s drylands suffer from desert- ification—a large portfolio of mostly human-caused problems that depress plant productivity. Some 23 million km? of the drylands have damaged vegeta- tion (Dregne, 1986). This often means the spread or increase of unpalatable plant species following overgrazing by cattle, goats, and other livestock Young & Longland, 1996; Archer, 1994; Bahre & Shelton, 1993; Van Auken, 2000). Those will some- times be native species. For instance, mesquite (na- tive Prosopis spp.) has dramatically increased in density over much of the southwestern United States (Bahre & Shelton, 1993; Archer, 1994; Van Auken, 2000). More often, the weeds will be exotic invasives such as Opuntia in Australia and Africa. (3) Another 40 million km? of the land surface has forests or woodlands of one kind or another, another 8 million km? are tundras, and the remain- der includes wetlands and urban areas (Olson et al., 1983 The conversion of forests to other habitats is more complicated, since most of the world’s crop- lands were once forests—and so have already been counted above as croplands. Most of the converted forests are in temperate regions. About 2 million m? of these forests have also been converted to grazing lands (Pimm, 2001). Modern human actions have shrunk the world’s tropical humid forests from an original area of from 14 to 18 million kn? to about 7 million km? at present (Myers et al., 2000; Pimm, 2001). Again, 174 Annals of the Missouri Botanical Garden the exact numbers depend somewhat on what one means by “humid.” Yet only about 2 million km? of croplands are in what was formerly humid for- ests. Some 5-9 million km? of humid forests have been converted to nominally grazing land, though much of it has very low stocking rates (Pimm, 2001). Summing these three pieces suggests that 15 million km? of present-day croplands, 23 million km? of drylands, 2 million km? of temperate forest converted to grazing lands, and from 5 to 9 million km? of additional forested land not producing crops have sustained sufficient changes to their vegeta- tion to make them target areas for invasive species. The combined total is just under half the ice-free land surface. The potential weed patch is huge. This approach is inevitably approximate and must miss many details. In particular, it does not map where these disturbed lands are. To both refine these estimates and provide a check on their ac- curacy, we will now explore detailed estimates of landcover. We do so first at a global scale, then at regional scales. Our analyses relate primarily to the once-forested half of the planet since the remote sensing imagery on which we rely does not so read- ily detect the damage to drylands. A SPATIALLY EXPLICIT GLOBAL ASSESSMENT OF CONVERTED FORESTS For the global analysis, we use a Geographic In- formation System (ERDAS Imagine v 8.5) to com- bine a map of presumed original vegetation with an estimate of current landcover. The result is a global map where each pixel has information about its original vegetation and if it has changed, or not changed, into a different type of landcover. he original vegetation map is from the Inte- grated Model to Assess the Global Environment (IMAGE) project (Leemans & van den Born, 1994; Alcamo et al., 1998; IMAGE team, 2001). In their Terrestrial Vegetation Model, the IMAGE team uses a modified BIOME model to estimate potential nat- ural vegetation using climate and soil characteris- ties. For a detailed description of this model, see Prentice et al. (1992) and Leemans and van den Born (1994). The resolution of this map is one-half degree of latitude and longitude. Color maps are available from Alcamo et al. (1998) and IMAGE team (2001). The current landcover map is from the Global Land Cover Characterization (GLCC) (Loveland et al., 2000). A digital version is available at (http:// /glec.html). This project used (13230 TT lola edcdaac.usgs.gov/g a one-year sequence of AVHRR satellite imagery to identify landcover in ca. 1992. As the primary concern of our study is disturbance (i.e., areas vul- nerable to invasion), we focus on the disturbed classes of the GLCC (croplands, mosaics of crop- lands and natural vegetation, and urban areas). The resolution of this map is approximately one km? at the equator. Ve also identify areas that have changed land- cover, but not necessarily into croplands or urban areas. For example, the conversion of forests into grasslands for grazing will not appear as disturbed, but it obviously is. (Grasslands are a natural type of vegetation, but not where the original vegetation was a humid tropical forest.) We do this to assess potential error causes and means of improving on our main analysis. The current data sets are not adequate for a com- plete and detailed analysis of landcover changes of all ecosystem types. The BIOME and GLCC maps use different classification schemes for vegetation that make matching corresponding classes between them somewhat arbitrary. For example. the GLCC map has an "open shrubland" class that corre- sponds to some grassland in the BIOME map. How- ever, open shrubland also includes areas that are obviously not grassland, such as central Australia, which the BIOME map classifies as hot desert. It is uncertain that these changes represent land deg- radation: more likely, they represent differences in classification schemes. RESULTS AND DISCUSSION Not surprisingly, the global analysis confirms that humanity has disturbed a large fraction of the world (Fig. 1). Of the ca. 129 million km? of ice- free land, ca. 27 million km? appear to have been converted to croplands, mosaics of croplands and natural areas, and urban areas. Table 1 shows the total disturbed area originating from each vegeta- tion type (rightmost column). Of the total disturbed area, 80% comes from just six vegetation types (bold numbers in Table 1) that originally covered 47% of the land. Disturbance concentrates in tem- perate climates (temperate forests, warm mixed for- est. grassland/steppe) and the drier subtropical and tropical vegetation (scrubland, savanna). Again, not surprisingly, most of this disturbance coincides with the world’s human population and croplands, most- ly in the Northern Hemisphere. Grasslands, scrublands, and savannas have lost from a fourth to a third of their original area (Table 1), but these numbers may be misleading. Distur- bance from livestock grazing on such ecosystems is difficult to detect by satellite. Most of the very cold Volume 90, Number 2 Jenkins & Pimm 175 2003 Global Weed Patch 1 Map of global disturbed areas. Black areas include croplands, mosaics of croplands and natural vegetation, urban areas, and original tropical woodlands and forests that are now grassland, savanna, or w woody savanna. Original vegetation is from the BIOME map and current landcover is from the GLCC map. This map uses a geographic projection and is not equal area. areas (tundra and boreal forest) and the very dry However, the GLCC data do not include separate (deserts) escape major disturbance. categories for dry habitats (such as grasslands) that Of the combined disturbed areas, 14.2 million are purported to originally have been forest (Love- km? are from once-forested areas (Table 1, sum of land et al., 2000). What happens if we assume that disturbed boreal forest, cool conifer, temp. mixed these are also converted landscapes? We approxi- mate the area of forest to dryland conversion by identifying tropical woodlands and forests that are now grasslands, savannas, or woody savannas. The area is about 2.4 million km?. Adding in this piece and decid. forest, warm mixed forest, and tropical woodland and forest). This seems to be a low esti- mate, for we should compare it with the 11 million km? of forests that are now croplands plus a further minimum estimate of 5 million km? of cleared trop- suggests that 16.6 million km? of forest have been ical forests not converted to croplands (above and converted. This is close to the lower estimate of 16 Pimm, 2001). million km? based on combining independent es- e 1. Original area of each vegetation type, the area disturbed as croplands, mosaics of croplands and natural auta, and as urban areas. Drylands are areas that were originally tropical woodlands or tropic sal forests (BIOME are now grasslands, savannas, or woody savannas (GLCC map). All areas are expressed as 1000s of square kilometers. Totals may be different from the sum of the parts bene of кке ый Area (1000 sq. km.) Total Vegetation type Original Croplands Mosaic Urban Drylands disturbed* Tundra 6096 7 13 0 20 Wooded tundra 2586 5 6 0 11 Boreal forest 16,144. 399 119 6 523 Cool conifer 809 686 10 1074 emp. mixed forest 6471 2151 1706 67 3924 Temp. decid. forest 4677 1842 837 62 2741 arm mixed forest 6189 1503 118: 3l 2719 srassland/steppe 17,800 1975 2078 32 4085 'se C 7 232 6 512 Serubland 2013 1356 19 3387 Savanna 15,926 2115 1962 11 4748 Tropical woodland 7485 830 842 10 1629 1682 Tropical forest 8893 718 843 fi 767 1569 Total 128,824 15,178 11,556 260 2397 26,995 * Does not include drylands. 176 Annals of the Missouri Botanical Garden Table Area disturbed An deforestation and through edge effects in the Amazon example. The area within 300 meters of an ue includes 4986 km? of “nat- ural" edge forest (see text). All areas are expressed as square kilometers. Area Total area | 300 m Area of Total area Year forest deforested from edge disturbed Original 112,919 0 4986 0 1992 100,974 11,945 17,801 24,760 timates of forest converted to croplands (11 million km?) and tropical forests converted to grazing lands (5 million km?); see above and Pimm (2001). This broad agreement is encouraging, but it also points to the difficulty in translating. landcover maps, with their inevitably arbitrary classifications of vegetation cover, into ecologically sensitive mea- surements of human impacts. That is a conclusion on which we now expand by considering two areas of forest conversion in more detail. А FiNE-SCALE ANALYsIS OF TROPICAL FORESTS We selected two tropical forest areas in Brazil with different disturbance histories to try to identify what the global analysis is missing. The Amazon is relatively intact but has had high rates of recent deforestation (i.e., within the last 30 years, Skole & Tucker, 1993). Northern Mato Gros- so state, in the southeastern Amazon Basin, is our example of such recent anthropogenic disturbance. In contrast, most of the Atlantic Forest was defor- ested more than 30 years ago (Fundação SOS Mata Atlántica, 1998). ' the surrounding area is our example of such his- toric disturbance. ‘he state of Rio de Janeiro and 'or the Amazon example, we calculate defores- tation using forest cover maps from the Tropical Rain Forest Information Center (TRFIC, 2002). We use maps for 1992 to match the year of the GLCC map. We also simulate the undisturbed condition by replacing the deforested class in the 1992 map with forest. Table 2 shows the deforestation statistics for the Amazon example. Comparison of TRFIC maps shows that as of 1992, 100,974 km* of the original 112,919 km? of forest remained, yielding 11.945 km? of deforested (disturbed) area The global analysis significantly underestimates disturbance in this region. According to the GLCC, the Amazon example shows just 3943 km? of dis- turbed area. That is only about a third of the re- gional estimate in the previous paragraph. Another 3582 km? is grassland, savanna, or woody savanna. Including these as disturbed brings the total to 7525 km’, which is still less than two-thirds of the TRFIC estimate. Even after correcting the GLCC map for unnatural drylands, it still misses a third of the disturbed area. For the Atlantic Forest example, we map forest ETM+ Using standard supervised classification cover in 1999 using Landsat 7 + satellite im- agery. techniques, we classify seven Landsat images into forest and non-forest classes. We do not distinguish between natural and plantation forests, but mono- specific plantations (e.g., eucalyptus) are not a large proportion of remaining forest in this area (pers. obs.). The World Wildlife Fund ecoregion map pro- vides an estimate of original forest cover (Olson & 1998). In the Atlantic Forest example, the analyses of Landsat imagery show that 91,993 km? of forest, of an original 127,850 km”, has been lost to defores- tation. The area of disturbed lands in the GLCC map is just 51,851 km?, 56% of the Landsat-de- rived estimate. However, adding in forests convert- ed to drylands increases the area to 111,000 km?, an overestimate of disturbance. In both examples, the best fit of estimates from GLCC data to detailed regional estimates comes Dinerstein, only when we recognize the conversion of forests to obviously disturbed habitats (such as croplands) and less obvious categories (such as grasslands and savannas) that could be natural ecosystems, but are not. Two factors contribute to these deficiencies of the GLCC map. One, the AVHRR imagery used making the GLCC has limited ability to discrimi- nate vegetation types. The AVHRR sensors spec- tral bands are too wide and poorly positioned for mapping vegetation, yielding inevitable errors. This appears to contribute to the overestimation of de- forestation in the Atlantic Forest example. The re- maining forest is simply misclassified. The second factor is that the resolution of the regional analyses (0.0009 km?) is much finer than the GLCC data (1 km?) in the global analysis. This finer resolution enables better detection of small areas of defores- tation. This likely contributes to the underestimate in the Amazon example, where there are many small patches of deforestation. We do not know if these error rates are general for the global analysis. These are only two examples from relatively small areas. What they do indicate is a need for better global mapping of landcover. At the time of writing, the GLCC was the best glob- al data set available, but good prospects exist for refining our results. Efforts are under way using Volume 90, Number 2 Jenkins & Pimm 177 2003 Global Weed Patch MODIS (http:// | lu/land find the GLCC estimates are both too small. In short, html) iad oe (http: сва geocover.com) sat- ellite imagery to map the world at 250-meter and 30-meter resolutions, respectively. Vegetation map- ping is a primary consideration in these sensor's designs, so the resulting maps should better dis- criminate vegetation types and their level of dis- turbance. The improved spatial resolution should also better detect small areas of disturbance. THE MISSING EDGES We have mapped disturbed areas as best possi- ble using current data sets, but another part of the weed patch is still missing from our estimates. Laurance (1997) found forest edges to be vulnera- ble to invasives because of “edge effects” disturb- ing the forest community. Although we cannot as- sess how much edge forest contributes to the global weed patch, the one km? resolution is too coarse; we can estimate it for the Amazon and Atlantic Forest examples. The regional forest maps have some errors that we must first correct. They have small gaps of non- forest, some of which may be natural, but many of which are small classification errors. These are in- significant for the earlier analyses, but can cause large areas of forest to appear to be near an edge, even if that edge is a single 30-m pixel. To account for this, we replace patches of non-forest smaller than 2 hectares with forest. The use of 2 hectares is conservative. It is larger than the presumed clas- sification errors, but may also include some truly deforested areas and natural gaps. The result is a probable underestimate of edge forest. mazon example, 17.801 km? of forest is within 300 meters of an edge (Table 2), the distance that Laurance et al. (1998, 2000) detected com- munity changes in Amazonian forest fragments. n the However, in the original state this region already had 4986 km? of natural edge forest due to rivers and savannas. Adding the edge forest, minus nat- ural edge, to the earlier estimate of disturbed area yields 24,760 km? of disturbed area M 2), ap- Е twice the original estimat tlantic Forest example, 29,373 km? of the remaining forest is within 300 meters of an edge. Adding this to the deforested area yields 121,366 km? of disturbed km? of undisturbed forest. The map of original for- area, leaving just 6484 est (WWF ecoregion) is too general to reliably cal- culate a “natural” amount of edge as done for the Amazon. hen we incorporate these regional estimates of edges into the calculations of disturbed area, then much disturbance—and much habitat for invasive species—occurs on a scale too small to detect with global landcover maps. CONCLUSIONS Invasive species are a growing problem for the world, both ecologically and economically. In re- sponse to the problem, governments are devoting increasing amounts of resources toward the preven- tion, control, and eradication of invasives in many parts of the world. To enable efficient use of these resources, the scientific community needs to iden- tify where invasives are likely to be a problem or become a problem in the future. Disturbed ecosystems are often favorable for in- vasives, a conclusion confirmed by comparing our map of disturbed areas with the distributions of some of the worst invaders. These disturbed areas, the global weed patch, occupy at least 29.4 million km? (23%) of the ice-free land surface. Other than the overgrazed drylands, which our analyses are unable to detect, the numbers broadly agree with independent estimates of disturbed area. The advantage of our approach is that it shows where those disturbed areas are and thus where the invaders are likely to be. Combining this with in- formation on who the invaders are likely to be should help in efficiently allocating resources to solve the invasive species problem. Literature Cited Alcamo, J., R. Leemans & E. Kreileman. 1998. Global Change э enarios of the 21st Century: Results from the IMAGE 2.1 Model. Pergamon & Elsevier Science, Ox- ord, U. Archer, S. western grasslands and savannas: 1994. Woody plant encroachment in south- ates, - and А. L proximate causes. 39-45 i av . Lay- cock & R. D. Pieper (editors), colo gs Implications of Livestock Herbivory in the West. The y for 3 . Shelton. 1993. Historic vegetation change, ddnde increases, and es in southeast- ern Arizona. J. свое 20: 489 Blamey, Grey-Wilson. 1989. tthe Illustrated Flora of Britain and Northern Europe. Hodder & Stoughton London. Dregne, 4 E. 1986. Desertification of arid lands. P. 12 in F. El-Baz & M. Н. A. Hassan (editors), Physics of De- ав. Kluwer Academic Publishers, Dordrecht, The Netherlands. Fundação SOS Mata Atlantica/INPE/Instituto Socioam- biental. 1998. Atlas da Evolução dos Remanescentes Florestais da Mata Atlántica e Ecossistemas Associados no Período de 1990-1995. Sao sige Brasil. Holm, L. G., D. L. Plucknett, J. V. Pancho & J. berger. 1977. The World’s Worst Weeds: кый and Biology. Univ. Press of Hawaii, Honolulu. 178 Annals of the Missouri Botanical Garden — ———, J. V. Pancho, J. P. Herberger & D. L. Plucknett. 1979. A Geographical Atlas of World Weeds. John Wi- ley & Sons, New York. IMAGE team. 2001. The IMAGE 2.2 Implementation of the SRES Scenarios: A а онна d of emis- sions, climate change and impacts in the 'entury. RIVM CD-ROM public ation 481508018, a In- stitute for Public Health and the Environment, Bilthov- en, The Netherlands. Laurance, W. F 1997. Hyper-disturbed parks: Edge ef- fects and the ecology of isolated reserves In tropical Australia. Pp. 71-83 in W. F Laurance & R. O. Bierregaard, Jr. (editors), "Tropical Pes Bran Ecology, Management, and Conservation of ено ad C ommunities. Univ. Chic 07 Press, Chic — ———, L. V. Ferreira, J. ankin-de s & 5. С. TEM e. 1998. Rain iol rus ntation and the dy- namics of Amazonian tree communities. Ecology 79: 2032-2040. = 'oncelos & T. E. Lovejoy. 2000. Forest loss and ines ntation in the Amazon: Implications for wildlife conservation. Oryx : —45. | den Born. 1994. Determining the tential global Wissibution of natural vegetation, ee and agricultural productivity. Water Air Soil Pol- lut. 133-161. ae T. R., B. C. Leemans, К. & G. J peed, J. F. Brown, D. O. Ohlen, Z. Zhu, L. Yang & J. W. Merchant. 2000. Developmen S a dahal land cover characteristics database and IG DISCover from 1 ex AVHRR data. Int. J. aa ET ns- ing 21: 1303-1: Browne a S. Boudjelas. 2001. 100 of the World’s Worst Invasive Alien Species. Invasive Species Specialist Group, IUCN/SSC GISP. Myers, N., R. A. Mittermeier, C. A. B. G. Mittermeier, G. TRFIC. 2002. da Fonseca & J. Ken гор о oe for ——— жш. Nature 403: 8 Olson, D. M. & E. Dinerstein. "1908. n. pun 200: A representation approach to conserving the Earth's most biologically valuable ecoregions. ынем Biol. 12: 5 2-515 Olson, J. S E aive таныш publication no. 5862 . Tennes Pimentel, D. L. Lac h. R Zuniga & D. Morrison. 2000. Environmental and economic costs of PASS nous species in hus г States. Bioscience 50: 5 Pimm, S. L. The World According x Pimm: A Sci- entist Audits n Earth. McGri aw-Hill, New York. ч . Cramer, S. P. Harrison, ү аы ans, R. . Mons rud & A. M. ion 1992. A global biome model based on plant physiology. and ла sol En n climate. J. Biogeogr. 19: 117-13: Skole, „. Tucker. 1993. Tropical deforestation and ten ane ntation in the Amazon: e ‘lite data from 1978 to 1988. Science 260: 1905- E Thompson. D. С.. R. L. Stuckey & E. Yr NN 1987. pread, impact, bet control of x loosestrife (Lyth- rum salicaria) in North Americ ү wet ands. Fish and Wildlife Кеѕеаге 'h Report no. 2. U.S. Fish and Wildlife Service, Washington, D.C. Tropical Rain Forest Information Center. RF at: (http://www.bsrsi.msu.edu/trfic/) (accesse «d n 28 October 2002). Van Auken, O. W. 2000. Shrub invasions of North Amer- ican grasslands. Annual Rev. Ecol. Syst. 31: 197-215. Vitousek, P. M., P. R. Ehrlich, A. H. Ehrlich & P. A. vae n. 1986. Human шо of = products of photosynthesis. Biosc : 368-37 Young, J. A. & W. oun. 1996. bids of alien plants оп Grant Basin rangelands. Weed Technol. 10: l. 384-39 ‘a ier & L. of Pede . Oak Ridge г Allison. 1983. Carbon 1 i Ecosystems. ORN National "ean REVISION OF ARDISIA Jon M. Ricketson? and SUBGENUS AURICULARDISIA — 120" J. Pipoly HP (MYRSINACEAE)! ABSTRACT A taxonomic revision of the Neotropical Ardisia = Auric ponia (Lundell) Ricketson & Pipoly is presented, and a key to the Neotropical subgenera of Ardisia is provided. The group comprises taxa formerly placed by Lundell in his segregate genera Auriculardisia, Amatlania, = Veris Ardisia ia subg. Auriculardisia is defined by its unique calyx lobes, which are asymmetrical, usually notched j elow PUn and always auriculate basally. Six sections are recognized within Ardisia subg. yes nra comprising 15 laxa: sect. Auriculardisia (Lundell) Ricketson & Pipoly (4 species), sect. Pleurobotryae Ricketson & Pipoly (1 species), sect. Fagerlindia Ricketson & Pipoly (7 species). sect. Wedelia Ricketson & Pipoly (8 species). sect. рел з (Lunde П) Ricketson & Pipoly (З species, including 6 subspecies), and sect. Palmanae Ricketson & Pipoly (47 species, including 2 subspecie). Thirteen new species and two new subspecies are described and illustrated. The new combinations Ardisia pellucida Oerst. subsp. pectinata (Donn. Sm.) Ricketson & Pipoly, Ardisia liebmannii Oerst. lx pu (Lundell) Ricketson & Pipoly, and Ardisia croatii Lundell subsp. correae (Lundell) Ricketson & Pipoly are proposed. Ardisia carchiana and Ardisia zakii are transferred to аша as Geissanthus carchianus (Lundell) Ric ketson & Pipoly and Geissanthus zakii (Pipoly) Rick- etson & Pipo Key words: Ardisia section Amatlania, Ardisia, section and subgenus Auriculardisia, section Fagerlindia, Myrsi- naceae, „жез section Palmanae, section Pleurobotryae, section Wedelia RESUMEN Se presenta una revisión taxonómica del subgénero neotropical Ardisia subg. Auriculardisia (Lundell) Ricketson & Pipoly y una clave para separar a los subgéneros neotropicales de Ardisia. El grupo comprende taxones anteriormente © б por Lundell en sus géneros Auriculardisia, Amatlania y Valerioanthus. Ardisia subg. Auriculardisia se define por el caracter único de sus lóbulos calicinos asimétricos, normalmente incisos justo debajo del um y siempre auriculados en la base. Se reconocen seis secciones dentro del Ardisia subg. oo а, q 75 taxones, listados a continuación: Auriculardisia (Lundell) Ricketson & Pipoly (4 especies), se ле Ricketson & Pipoly (1 especie), sect. Fagerlindia Ricketson & Pipoly (7 especies), sect. Wedelia Ricoh & Pipoly (8 ampa sect. Amatlania (Lundell) Ricketson & Pipoly (3 especies, incluyendo 6 subespecies ) y sect. саф Ricketso Pipoly (47 especies, incluyendo 2 subespecies). Se np us e ilustran trece especies y dos subes Van үл como nue para la ciencia. Se proponen tres combinaciones nuevas: Ardisia pellucida Oerst. subsp. pectinata (Donn. Sm.) Ric Hon: & Pipoly, Ardisia liebmannii Oerst. m poe бл (Lundell) Ricketson & Pipoly y Ardisia croatii е; subsp. correae (Lundell) Ricketson & Pipoly. Se transfieren Ardisia carchiana y Ardisia zakii al género Geissanthus, como Geissanthus carchianus (Lundell) d & Pipoly y Geissanthus zakii (Pipoly) Ricketson & Pipoly. The pantropical Ardisia Sw. is the largest genus circumscription has been problematic owing to a in the family Myrsinaceae, containing perhaps as — lack of comprehensive treatment since that of Mez many as 500 species (Chen & Pipoly, 1996). Its (1902) in Engler's Das Pflanzenreich a century ago. We thank the ича Botanical Garden and the Flora Mesoamericana Project, as well as the Fairchild Tropical cles. for their support to bring Ricketson and Pipoly together to work on the project. We gratefully acknowledge the loans from many herbaria an ains the study possible, along with our collaborators who supplied us with much critical material. For assistance with the scanning e pee micrographs we thank George Taylor of the Electron Microscope Laboratory at the Florida International University. Jack Fisher of Fairchild Tropical Garden pen assisted wit the plates and reviewed the morphology section, for which we are very grateful. We also very m appreciate the review of the taxonomic concepts section by Walter Judd of the University ‘of Florida. Scott Zona, Tim Utteridge, and Glen and Janet Hahn, Dean and Gail Larson, and Allan and Hope Schultz for their generous support of our work. i ally, the assistance of Jon Ricketson’s volunteer, Mary Bard, greatly facilitated the annotation and movement of specimens. ? Missouri Botanical Garden, P.O. Box 299, St. Louis Missouri 63166-0299, U.S.A. jon.ricketson@mobot.org. rchild Tropical Garden, 11935 Old Cutler Road, Coral Gables (Miami), Florida 33156-4242, U.S ‘pipoly@faire hildgarden.org. ANN. Missouri Bor. GARD. 90: 179-317. 2003. 180 Annals of the Missouri Botanical Garden The genus has traditionally been separated from all others in the Myrsinaceae by what were interpreted as free filaments, and pluriseriate ovules (Mez 1902). Pipoly and Ricketson (1998a), found that the stamens in all Ardisia are actually ۰ however, connate basally by their filaments to form a hyaline, inconspicuous tube, but that the tube is free from the corolla. While only a few groups within Ardisia s.l. have been segregated as separate genera in the Paleotropics (Sadiria Mez (1902), Afrardisia Mez (1902), Tetrardisia Mez (1902), Hymenandra A. DC. (1834), Parardisia Nayar & Giri (1986)), there has been an enormous increase in separation of species groups from Ardisia to new genera in the Neotropics, starting with Aublet’s (1775) descrip- tion of /cacorea (against which Ardisia is conserved) followed by Alphonse de Candolle (1841), Ducke (1930), and finally Lundell (1963, 1964, 198 1a, 1981Ь, 1981c, 1981d, 1982). Lundell's contribu- tion was clearly the most extensive, consisting of Ardisia subg. Auriculardisia sect. Amatlania (Lundell) Ric- ketson & Pipoly), Auriculardisia Lundell (— Ardisia subg. Auriculardisia (Lundell) Ricketson & Pipoly). Hymenandra А. DC., p.p.). Ctenardisia Ducke), Gentlea Graphardisia Lundell (= Ardisia subg. Pipoly & Ricketson, 1998a), Ardisia subg. Ardisia), Oerste- dianthus Lundell, Synardisia (Mez) Lundell, Vale- rioanthus Lundell (= Ardisia subg. Auriculardisia sect. Auriculardisia (Lundell) Ricketson & Pipoly), and Zunilia Lundell (= Ardisia ng Graphardisia Mez, Pipoly & Ricketson, 1999 group is comprised of over 800 names, it will be some time before each species has been reviewed the segregation of Amatlania Lundell (= ee” Chontalesia Lundell (= м Yunckeria Lundell (= Lundell, Graphardisia Mez, — Ibarrea Lundell (= ecause the and the entire group revised. In the meantime, we suggest using the key to the Mesoamerican genera that we recently published (Pipoly & Ricketson, 1999a) to identify gpa to generic levels (Ar- disia vs. Ctenardisia, Hymenandra, Gentlea, Synardisia). Lundell (198 : 342) separated his ge- nus Auriculardisia from Ardisia s.l. using the fol- lowing key: and . Plants glabrous or pubescent, not scaly; sepals symmetrical, ovate, longer than wide, thick: flowers strictly racemose or d did Pd Ardisia ‚ Plants с conspic uously : sc caly, furfuraceous or rare sly lepidote; sepals asymmetrical, depressed-orbicu- lar, elliptic or broadly ovate, and mostly wider than ong; шешш flowers in heads, subcorym- bose, rely umbellate or spicate, the rachis sometimes accrescent with the fruits lion 'oming ra- cemose Auric ү Our concept of {һе Auriculardisia group of spe- cles as a part of the larger genus Ardisia includes another group, Amatlania, segregated from Ardisia by Lundell (1982). Lundell stated that Amatlania resembles another segregate genus, Oerstedianthus Lundell, because both have filaments with gland- tipped hairs [a character Pipoly and Ricketson 1998a, 1999b) observed in Ardisia subg. Graphar- disia|. but differs notably in the nature of the in- dument. We also discovered that Lundell's (1982) genus Valerioanthus belonged to the Auriculardisia group. based on Am, asymmetric calyx lobes with subapical notches. Lundell (1982) defined Va- lerioanthus by its coarse red hirsute indument, with its auriculate, trichomes up to 2 mm long. During the course of this study, we noted with interest that the species Lundell (1982) included in the group, inc sc Valerioanthus nevermannii (Standl.) Lundell (= disia nevermannii Standl. in sect. aio Ж у ursinus (Lundell) Lundell (= Ardisia ursina Lundell 1 sect. Auriculardisia), and V. hirsutissima (Lun- dell) Lundell (which we have placed in synonymy -* under Ardisia ursina) have strikingly distinct ves- titure that in no way serves as a uniting character state. As defined in this treatment, Ardisia subg. Auriculardisia represents the largest group of Ar- disia species in the Neotropics and is comparable in size and taxonomic complexity only to subgenus Acrardisia, which ranges from Southeast Asia and the Pacific to Mesoamerica and the Caribbean. While preparing a treatment of Ardisia subg. Au- riculardisia for inclusion in the Myrsinaceae for lora Mesoamericana, we saw the need to present a full account of the group with detail not possible and as a precursor to our eventual Flora Neotropica treat- in the format of Flora Mesoamericana, ment. Herein we formally propose a new taxonomy for Ardisia subg. Auriculardisia to include six sec- tions, of which four, sect. Fagerlindia, sect. Pleu- robotrya, sect. Wedelia, and sect. Palmanae, are new; we reduce the genus Amatlania to a section of Ardisia subg. Auriculardisia; and finally, we cir- cumscribe the type section Auriculardisia. Delimi- tation of sections, while admittedly artificial, is de- signed to facilitate identification and to illustrate the salient character states of each group of species within the subgenus. Unfortunately, it was not pos- sible to conduct a phylogenetic analysis, so the ex- tent to which each section is monophyletic is cur- rently unknown. Ardisia subg. Auriculardisia sect. Amatlania con- tains three species and eight taxa. Amatlania was described as a separate genus by Lundell (1982: 38), based largely on the “reddish glandular pu- Volume 90, Number 2 2003 Ricketson & Pipoly 181 Revision of Ardisia subg. Auriculardisia bescence" of the branchlets and inflorescences. We have determined (see Morphology, Vestiture) that the vestiture found among this suite of species is actually composed of papillae or papillae mixed with uniseriate, multicellular glandular villous tri- chomes (Fig. 3A). Other features that define the section include the various toothing to serration of the leaf margins and the dense yellow glandular- granules within the corolla tube and on the fila- ments. Recognition of this group of species as a separate genus would leave y usse undefined. all species of Amatlania, the stamens are E basally by their filaments to form an inconspicuous tube free from the corolla tube, clearly placing them within Ardisia. In addition, all species in Amatlania have asymmetric calyx lobes with a sub- apical notch and an (albeit more imperfectly formed) auricle on one side, which also indicates that this group of species belongs to Ardisia subg. Auriculardisia. We therefore reduce Amatlania from generic to sectional status within Ardisia subg. Au- riculardisia. Ardisia subg. Auriculardisia sect. Auriculardisia, containing four species, is defined by the combi- nation of its terminal inflorescence with branches terminating in congested, glomerate corymbs, each of which is subtended by a persistent, enlarged in- florescence branch bract, and finally, the individual flowers subtended by a persistent, enlarged flora bract as long as the flowers. This section includes the rare and very interesting species Ardisia dodgei, which has unique red to pink floral bracts and is visited by hummingbirds. The section includes two species placed in the genus Valerioanthus by Lun- dell (1982) based on his misinterpretation of their vestiture (now considered synonyms of A. ursina), along with A. dodgei, A. glomerata, and A. nervo- sissim Las subg. Auriculardisia sect. Fagerlindia contains seven species and is defined by exhibition of Fagerlind’s Architectural Model (see Morpholo- gy), combined with terminal, pendent inflorescenc- es. Because the vegetative and reproductive shoot leaves are dimorphic, many herbarium specimens are incomplete, containing the flowering shoot only, or with a vegetative shoot flowering precociously. This has led to taxonomic over-description and a fundamental lack of understanding of the devel- opmental dynamics of these plants. Members of section Fagerlindia are normally inconspicuous el- ements of the understory. Ardisia subg. Auriculardisia sect. Palmanae is clearly the largest section in the subgenus, with 47 species. The section is admittedly artificial and is defined by the subsessile to sessile inflorescences. Whether this feature has arisen several times with- in subgenus Auriculardisia, or whether section Pal- manae represents a group without any other distin- guishing character state, is unknown at this time and will require more rigorous study. Ardisia subg. Auriculardisia sect. Pleurobotryae is monotypic, containing only Ardisia pleurobotrya Donn. Sm., and is defined by the strictly lateral inflorescences, long naked peduncles to 6.2 cm long, dense, overlapping chocolate brown lepidote scales on most plant parts, and pendent flowers on long, usually sigmoid pedicels. The lateral inflores- cences and chocolate brown lepidote scales are unique within subgenus Auriculardisia, and within the genus as a whole are otherwise frequent only in subgenus Akosmos Mez of Ardisia subg. жин уун элен sect. Wedelia is de- fined by a subshrubby to small arborescent habit, and terminal, columnar to sub-columnar and often pendent inflorescences, on peduncles at least % the length of the inflorescence and subtended by large, foliaceous bracts. Like species of section Fagerlin- dia, these plants are inconspicuous members of the forest understory. The most salient feature of the section is the columnar to sub-columnar inflores- cences, similar to the branches found in the basally branched, pyramidal inflorescence of Ardisia pseu- doracemiflora (sect. Palmanae). MORPHOLOGY HABIT AND ARCHITECTURE Species of Ardisia subg. Auriculardisia are ter- restrial, small trees and shrubs, rarely subshrubs (A. pellucida). The vast majority of taxa (sect. Amat- lania, sect. Palmanae, and Wedelia) display Scarrone’s Architectural Model (Hallé et al., 1978; Bell, 1991), which is defined by an orthotropic monopodial, rhythmically growing principal axis (stem or trunk) differentiated from the branches, which are produced in pseu- doverticillate tiers. The branches are composed of orthotropic shoots that produce terminal inflores- Auriculardisia, sect. cences; thus branching is sympodial by substitu- tion. In a mature individual, there is often a dense canopy crown whose ultimate units may appear pla- giotropic by gradual bending as they develop, a phenomenon documented by Fisher and Stevenson 1981). Species of Ardisia subg. Auriculardisia sect. Pleurobotryae exhibit Rauh's Architectural Model, is characterized by a monopodial, rhythmi- ~ whic cally growing, readily distinguishable trunk that de- velops pseudoverticillate tiers of branches morpho- genetically identical to itself. All branches are 182 Annals of the Missouri Botanical Garden orthotropic and monopodial, with lateral inflores- cences that do not affect development of branches. The most unusual model we found was that of Ardisia subg. Auriculardisia sect. Fagerlindia, de- scribed herein, which exhibits deme Archi- tectural Model (Hallé et al., 78). Fagerlind's Model is defined by its eee rhythmically growing trunk, with tiers of modular, plagiotropic branches that branch sympodially. The modular branches form sylleptically and sympodially, often by apposition, and with spiral phyllotaxy. First, a trunk (“vegetative shoot") develops rhythmically "vegela- producing successive pseudoverticels of tive" leaves until the apex loses dominance and latent lateral (“axillary”) buds are released, pro- ducing successive, monopodial pseudoverticels of shoots, each of which is sylleptic, with a long hypo- podium and exhibiting rapid extension growth. The sylleptic branch shoots produce first а prophvll, then pseudoverticels of “reproductive shoot" leaves (similar to the leaves of the "vegetative shoot" but notably smaller), and then may either lose domi- nance and once again produce “vegetative shoot” leaves, branch sympodially by apposition growth (repeating the module) without flowering, or pro- duce a terminal or pseudoterminal (acting as a ter- minal) inflorescence, followed by sympodial branching by substitution. The pseudoterminal in- florescences are lateral and form just below the branchlet apex. While the terminal bud is visible it appears that the shoot does not grow significantly While not taxonomically significant, it is interesting that all 5 until the inflorescence is in late fruit. members of the recently discovered Mesoamerican component of the boreotropical genus Hymenandra also exhibit Fagerlind's Model and fused anthers (Pipoly & Ricketson, 1999a). However, members of Ardisia subg. Auriculardisia sect. Fagerlindia may be readily distinguished from Hymenandra by their free anthers. TRUNK AND BRANCHLETS In this treatment, the principal axis formed by a "afit attains a diameter greater than or equal to 2.5 cm diameter at breast height (DBH); if it does not, it negatively geotropic shoot is termed a “trunk is termed a "stem." Branchlets are defined as the upper 10 cm of any branch. Branchlets may be stout or slender, straight or flexuous, terete, subte- rete, somewhat angulate with longitudinal ridges, or b bearing well-developed interpetiolar ridges. The branchlets always bruise blue-purple when cut, due to the few to numerous lyso-schizogenous resin ca- nals present in the secondary phloem. The resin canals of the pith in the primary growth region near the apex gradually coalesce into a poorly defined resinous area during secondary xylem formation and branchlet thickening. The bark is smooth, or is rarely exfoliating (Ardisia vesca Lun- ell). mostly ^ — VESTITURE The vestiture of the stems or branchlets, the pet- ioles, leaves, and peduncle is normally uniform within a given species. Much confusion is evident in the literature regarding the term “furfuraceous- lepidote" because the structure. of the covering scales of the leaves had not been thoroughly inves- tigated. While a study of the covering scales for the entire family is clearly warranted, we here define the. principal types of scales and other vestiture found in Ardisia subg. Auriculardisia. The most striking type of trichome found in the subgenus is the cupuliform scale, which is sessile to subsessile, with small lobes to arms, composed of one cell each, variously connate to the next to form a deep cup or a small cup with radiating arms (Fig. IA-1D). These trichomes are quite large. ranging in size from 200 to 300 jum in height. The cupuliform scales may be mixed with much smaller scales of the same type or translucent. subentire scales, here referred to as minutely furfuraceous- lepidote scales, composed of 8 to 12 cap cells and one stalk cell, approximately 30-50 jum in diam- eter, shown in the middle left area of Figure 1C and throughout in Figure 1D. Furfuraceous-lepidote scales are those that do not form a discernible cup and which may be entire to subentire but never with radiating arms, and range in size from 200 to 300 jum in diameter (Fig. 2B). The apparent thickness of the cap cells, owing to their respective lignification, varies greatly, from extremely thin, almost translucent (Fig. 2B) to ap- pearing opaque (Fig. 2A, 2D). In many taxa of the subgenus, varying quantitites of hydropotes may be present, mostly on the abaxial leaf surface (Fig. 2A). They are often confused with lepidote scales, but their obvious subsidiary cells and unusual cap morphology easily distinguish them. A good example is seen on the left side of Figure 2A, found in Ardisia panamensis Lundell. Hydropotes (“water drinkers”) were described by Mayr (1915), then Grüss (1927a, 1927b), and fi- nally, Gessner and Volz (1951), all of whom re- ported these structures from studies of submerged aquatic plants. They were first. reported for the Myrsinaceae by Pipoly (1987) in Cybianthus Mart. subg. Grammadenia (Benth.) Pipoly, but are now Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 183 Figure 1. reclined habit. —C. Ck D. EL scales e т those of А. brenesii, but more stipitate e-up is cupuliform scales Zuchowski 9128 (MO); D fun B. Boyle & С Godt 932 (FTG).) Scale bars: А, С: known for most subgenera of Cybianthus (Pipoly, 98), Hymenandra (A. DC.) A. DC. ex Spach (Pi- poly : Ricketson, 1999a), and Wallenia Sw. (Pan- fet, 2002). T аа from lepidote scales by their 5 to subsidiary cells, a central foot cell, a basal cell, a stalk cell, and up to 12 cap cells. In early ontogeny, a mucilaginous substance accumulates around the basal cell. Later the cap breaks off, cilaginous ring around the basal cell, whereupon they function as hydathodes in water regulation. Based on elegant ultrastructural and chemical stud- ies by Lüttge (1964) and Lüttge and Krapf (1972), we now know that these structures actively trans- port minerals into the leaf through the mucilaginous substance secreted around the stalk. The "e These structures may easily be distin- ~ eaving a mu- inous substance stains easily with Sudan IV poly, 1987). Given the nutrient-poor, moist to ves vial tropical life most Myrsinaceae, we expect to find that these structures are ubiquitous within the family. Hydropotes are zones inhabited by > furfurac uos o sc ode (f. A 3 from W. Haber & W. Zuchowski 9358 (MO); е scales in Ardisia subg. ee —A, B. Cupuliform scales of A. brenesii; note semi- ssiramea, showing cup opened and flattened at maturity. — . furfuracea. Note that cupuliform scales resemble C ae W. Haber & W. 200 um; B, D: : easily distinguished from translucent lepidote scales by their subsidiary cells. Ardisia nevermannii Standl. is unique within Ar- disia subg. Auriculardisia sect. Fagerlindia be- cause of its ferrugineous hirtellous vestiture, com- posed of stiff trichomes less than 2 mm long. These densely packed hairs are apparently unicellular, with subulate apices and bulbous bases (Fig. 2C). Lundell (1982) did not adequately distinguish the hirtellous tomentum of Ardisia nevermannii from the rufous multicellular, ONE trichomes such as those of Ardisia ursina, which consist of a uniseriate row of glandular cells that terminate in a stellate cluster of arms (Fig. 3C, 3D). If they are dense, the vestiture is referred to as stipitate-stel- late tomentose, or if they are less so, it is termed stipitate-stellate tomentellous. The arms frequently break off, leaving an apparent glandular-villous tri- chome, indistinguishable from those found on the branchlets and inflorescences of section Amatlania, such as Ardisia pellucida (Fig. 3A). While we can- 184 Annals of the Missouri Botanical Garden Figure A. оа Note walled e: transluce № tively quii size a | пеКпеѕ from B. Hammel et 2 19091 (ЕТС); ( D: 50 jum; В: 250 pm; С: 300 рт not dismiss the possibility that the glandular-villous trichomes of section Amatlania (Fig. 3A) are de- rived from the glandular apie late trichomes found in section Palmanae (Fig. 3C), we have not found any remnant arms on the villous trichomes of any member of section Amatlania. At times, the stipitate-stellate trichomes form a layer of vestiture above the minute, subsessile translucent scales in- distinguishable from those sometimes co-occurring with cupuliform-lepidote scales. The inflorescence vestiture among members of Ardisia subg. Auriculardisia sect. Amatlania is par- ticularly notable for the well-developed glandular- papillae (Fig. 3A), which in Ardisia pellucida subsp. pellucida are also accompanied by glandular villous trichomes. These trichomes preserve poorly and detach easily during the pressing and drying process, so many specimens lose most of the in- dument with time and handling. The papillae are very similar to those found in Cybianthus subgenera Grammadenia (Pipoly, 1987) and Microconomorpha Leaf vestiture in Ardisia subg. Auriculardisia. —A. Minute furfuraceous-lepidote sc (f), hydropote that scale is еж к To —B. Minute furfuraceous-lepidote scales, A. « tellous trichomes, s greater than the other species. (A ‚ from P. Acevedo 6875 (FTG); D from б. pam 2626 (FTG).) Scale bars: ^i һ), pm — riculata. Note t n а F pius 'e Mea 'pidote scales, . n J. Pipoly 7065 (FTG); I А. nevermannil. and are The 1983), are epidermal touch origin, when (Pipoly, somewhat sticky to the alive. chemical substance they contain is unknown. LEAVES The leaves of all species of Ardisia subg. Auri- culardisia are alternate or pseudoverticillate, ex- stipulate, and simple. Given that all branching is sylleptic, rapid extension growth of each shoot is followed by production of one or two prophylls, fol- lowed by fully formed leaves. In those species ex- hibiting Fagerlind's Model, the easily distinguish- able “vegetative shoot leaves" are much larger than the "reproductive shoot leaves" and have a smaller length-to-width ratio. While the leaves of the prin- cipal axis and the branches have spiral phyllotaxis, the leaves of the branches are reoriented in a plane (dorsiventral) through a twisting of the internodes rather than by distichous phyllotaxis. The blades chartaceous to coriaceous, or are mostly rarely Volume 90, Number 2 Ricketson & Pipoly 185 2003 Revision of Ardisia subg. Auriculardisia Figure 3. Vestiture in Ardisia subg. Auriculardisia. —A. Inflorescence rachis with glandular- pie | ep) and а villous (gv) trichomes, А. ziii icida subsp. pellucida. This is re Sine sentative of section Amatlania. nflo- rescence rachis with minute, overlapping furfuraceous-lepidote scales, A. wedelii. —C. Ab: Mw lea Герн near mid- “р ia ursina, showing the glandular, stipitate-stellate trichomes. —D. Inflorescenc ; hs ct margin with glandular, stip tellate tric bens: A. ursina. (А from C. Martinez "e (FTG); B athe: | P Eos 7716 (MO); С, D from б. McPherson 14042 (FTG).) Scale bars: A-C: 200 рт; D: 50 p membranaceous, elliptic, oblanceolate or oblong, cences on long peduncles terminating in corymbs. apically acute to long-acuminate, basally cuneate, The panicles are normally pyramidal and much to obtuse or rounded, or rarely auriculate. broader than long. However, an exception is found The blades may also be decurrent on the petiole in section Wedelia, where the panicles are columnar or not, conspicuously or inconspicuously (visible (much longer than broad) to sub-columnar (longer but not raised) or prominently (raised) punctate or {һап broad). The rachis may be straight, flexuous, punctate-lineate, variously furfuraceous- or cupu- or rarely geniculate. The inflorescence bract is nor- liform-lepidote. hirtellous-tomentose, villous glan- mally early caducous; the inflorescence branch dular tomentose, or stipitate-stellate tomentose. The bracts may be foliaceous, large, and often envelop margins may be entire, crenate, serrate, pectinate, the terminal flower cluster, at least in early ontog- dentate, flat, or revolute. The petioles may be ob- eny solete or subobsolete, canaliculate or marginate. The К; С ап be 4-, 5-, ог 6-merous, with asymmetric calyx lobes that Eo a subapical notch INBLORESCENGES AND FLOWERS and a prominent auricle on one side, a feature that Within Ardisia subg. Auriculardisia, the most defines the subgenus. The corollas are usually common inflorescence is a panicle, with flowers ar- white to pink. The corolla lobes may be ovate, el- ranged in corymbs or glomerules at the end of the — liptic, or oblong, symmetric to asymmetric, appar- branches, rarely appearing subracemose. The inflo- ently epunctate, pellucid, black or orange punctate, rescence is a terminal panicle except in section and/or lineate. The stamens are fused basally by Pleurobotryae, which has obviously lateral inflores- their filaments to form a hyaline, inconspicuous 186 Annals of the Missouri Botanical Garden tube, and the filaments may be flat or terete, character states. We should emphasize that our straight or gradually broader toward the base. The anthers may be ovate, lanceolate, linear-lanceolate, oblong, apically apiculate, subulate, caudate, mucronate, acute, or acuminate, basally cordate to subcordate or sagittate, dehiscent by subapical pores opening into wide longitudinal slits, with the - connective inconspicuously to conspicuously or prominently punctate dorsally. The pistil is most often obturbinate but may be obnapiform or ellip- soid; the ovary is globose, subglobose, ellipsoid, ovoid, or oblong. The placenta may be globose, sub- globose, or ellipsoid, with apparently uniseriate (few ovules in a high anthotactic spiral), biseriate or pluriseriate, few to many ovules. FRUIT The fruit is a one-seeded drupe that may be glo- bose to subglobose, inconspicuously (pellucid) punctate, or prominently black punctate and/or lin- ale. TAXONOMIC CONCEPTS, AND TERMINOLOGY NOTES ON KEYS, NOMENCLATURE, The keys are artificial and designed to expedite identification of herbarium specimens. An attempt has been made to emphasize vegetative characters to increase their usefulness with sterile material. Quantitative and qualitative data presented in keys and descriptions for floral parts and bracts were taken from organs rehydrated from herbarium spec- Measurements from imens by boiling in water. these range from 10% to 15% greater than those measurements taken directly from dried material. Data regarding stem diameters, inflorescence ra- chises, pedicels, and leaf and fruit shape and size were taken from dried herbarium specimens. A List of Species and Subspecies and an Index to Exsic- catae are provided in Appendices 1 and 2, respec- tively. The numbering system of Appendix 1 has been used for the species and subspecies in the text to help the reader locate taxa within the article. These numbers are also used in the sectional keys. Any correlations with phylogenetic relationships are coincidental. Our concept of a subgenus is here defined as: a group of species that comprises a major basal phy- logenetic subdivision within a monophyletic genus, whose formal recognition enables better under- standing of the phylogeny of the entire genus. A corollary to that is our concept of the section, de- fined as: a group of species within a formally rec- ognized subgenus that appears to share a common ancestry, based on the fact that they share unique work has thus far produced only preliminary clad- ograms, and that all homologies have not been worked out. However, we believe it is more appro- priate to formulate a preliminary phylogenetic hy- pothesis rather than have no working framework until the final analysis is done. By definition we advocate the use of sectional rank only within for- mal subgenera. Our concept of subspecies follows that of Pipoly 1987: 46), who defined a subspecies as: "groups of populations within a single lineage of ancestor- ~ descendant populations that show variation by unique combinations of plesiomorphies, or homo- plasic apomorphies, correlated with biogeography and/or ecology. This rank is primarily used to con- vey information regarding variation in the life his- tories of these populations and character state dif- ferences hypothesized to be the result of this variation. The subspecific rank in no way attempts to predict speciation events.” Morphological terms in this treatment follow Lindley (1848) and Pipoly (1987, 1992) for the in- florescence, rachis pedicels, and floral parts. De- scription of leaf morphology follows Hickey (1984), trichome description follows Theobald et al. (1984), and basic cell and tissue terminology follow Met- calfe (1984). Lundell (1981b, 1986) published 68 new names and combinations for binomials in Ardisia and Myrsine L. In both papers, he indicated that he was opposed to placement of the taxa involved in those genera and that the combinations and new names were provided in anticipation of future circumscrip- tions of the genera. We agree with Morales (1997) that in doing во, Lundell inadvertently invalidated his new names and combinations, according to the International Code of Botanical Nomenclature, Ar- ticle 34.1 (Greuter et al., 2000). For Ardisia spe- cies, we (Pipoly & Ricketson, 1998b) corrected this situation by making 27 combinations and 9 new names, of which 14 belong to members of Ardisia subg. Auriculardisia. In this monograph, we cite Lundell’s invalid combinations solely for the pur- poses of clarity, given the preponderance of no- menclatural synonyms complicated by the fact that the same epithets have been used for Ardisia as well as for many species in the segregate genera. While we are cognizant of the fact that these dell names have no nomenclatural status, we felt it Uun- was important to summarize all relevant nomencla- tural matters in one place to avoid confusion. TAXONOMIC TREATMENT Ardisia Sw., Prodr. 3: 48. 1788. TYPE [conserved]: Ardisia tinifolia Sw. Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 187 Trees ог shrubs, rarely suffrutescent herbs. | Leaves exstipulate, alternate or pseudoverticillate, normally punctate or punctate- lineate, petiolate or sessile. Flowers bisexual, 4- to 5(to 6)-merous; perianth with imbricate or quincun- cial aestivation (either dextrorse or sinistrorse), the or simple, lobes shortly connate; stamens included, the fila- ments connate into an inconspicuous tube, always free from the corolla tube, the anthers free, often 3 times longer than wide, dehiscent by apical or sub- apical pores, subapical pores opening into longi- tudinal slits or simple longitudinal slits, rarely transversely septate; pistil obturbinate, the ovary ovoid or oblong, the ovules few to numerous on a basal placenta, at times appearing uniseriate, some- times biseriate, most commonly pluriseriate, the style elongate, normally exserted at anthesis. Fruits drupaceous, l-seeded, punctate or punctate-lin- eate, sometimes longitudinally costate, with a some- what fleshy exocarp and crusty or slightly bony en- docarp; seeds covered by membranous remnants of placenta. Distribution. About 400 to 500 species, pantro- pically distributed but with the highest diversities in central Malesia and southern Mesoamerica. For Mesoamerica, we expect that of the nearly 800 names for the group, we will recognize approxi- mately 120 species. For groups within Ardisia native to or naturalized in Mesoamerica, we present the following key to the subgenera: KEY TO THE SUBGENERA OF ARDISIA IN THE NEOTROPICS la. Leaf blade margins crenate with a translucent vasc lule in the sinus of each « tion; appearing uniseriate subg. Pico qaae Leaf blades entire, crenate, serrate or pectin but never with nodules in the sinus of each cre- nation or indentation; ovules numerous, plurise- riate. 2a. Calyx and corolla lobes dextrorsely imbri- cate; apically free ا‎ of the filaments less than % as long as t ularized nodule 'Tena- ovules few, — — - e anthers. 3a. Inflorescence a ri ee termi- nal panicle; anthers widely ovate, de- hiscent by wide longitudinal slits sub- confluent apically at anthesis s Дабаа а es ence a а lateral umbel; anthers lanceolate, dehiscent by narrow longitudinal, non-confluent slits subg. Tinus uw — . Calyx and corolla lobes sinistrorsely imbri- cate or quincuncial; apically free portions of the ии longer than % as long as the anth 4a. rien r white, rarely light pink, not conc Mi. dehiscent by 6: 5 yellow gitudinal slits or subapic al pores open- 5a. Inflorescence branches spicate; branchlets mostly glabrous; subapical notch along the margin, symmetrical, Tur than wide, not > basally -........ subg. Ardisia . Inflorescence branches corymbose branchlets mostly with a subapical notch along the margin, asymmetrical, auriculate subg. ee . Anthers concolorous (drying gray). « hiscent by terminal to a e dedi g -— pore wider than the the- ca are it. 6a. Branchlets and n E inflorescence perianth densely - iss d blac e and/or punc- ally clasping the developin subg. Graphardisia зиф :hlets and infloresc ence rac 118 furfur: 6b. arely conspicuousl never prominently punctate-lineate, the inflorescence and floral bracts "crescent, never шеге, the ийги. эла fruit subg. /cacorea Treatments of Ardisia subg. Graphardisia were published by Pipoly and Ricketson (1998a, 1999b) and of Ardisia subg. Acrardisia by Pipoly and Ric- (2000). Treatment of the other subgenera Ardisia, subg. Ісасогеа, subg. Tinus, and subg. Crispardisia) are now under way (Ricketson & Pipoly, in prep.). ketson (subg. TAXONOMIC TREATMENT OF ARDISIA SUBG. AURICULARDISIA Ardisia Sw. subg. Auriculardisia (Lundell) Ric- ketson & Pipoly, comb. et stat. nov. Basionym: Auriculardisia Lundell, Phytologia 49: 341. 1981. TYPE: Ardisia glomerata Lundell. : 50. 1982. TYPE: Ar- = Valerioanthus never- коре Lundell, Wrightia 7 disia nevermannii Standl., mannii i (Standl. ) Lundell]. Stems or branchlets and leaves furfuraceous-lepidote (the scales tawny, Subshrubs or small trees. or rarely chocolate brown in color, appressed or flat, 188 Annals of the Missouri Botanical Garden thin, without veins visible), cupuliform-lepidote (chocolate or rufous, like a bowl with lobes, or arms, stiff yet translucent, with prominent venation visible), mixed lepidote, hirtellous-tomentose (the hairs simple, apparently one-celled, long and some- what stiff), rufous stellate-tomentose, the trichomes on multicellular, uniseriate stalks (the star section frequently breaking off and the hairs then seeming villous), minutely glandular-papillate (normal ru- fous), or villous-tomentose (indistinguishable from stipitate-stellate with star arms broken off). /n/lo- rescences terminal or rarely (A. pleurobotrya) lateral panicles that are pyramidal, or rarely (sect. Wede- lia) columnar to sub-columnar, the branches bear- ing flowers in corymbs, glomerules or rarely pseu- doracemes; pedicels stout to almost obsolete. Flowers with the perianth sinistrorsely imbricate or quincuncial; calyx lobes essentially free, subapi- cally notched, basally auriculate on one side; co- rolla lobes often slightly asymmetric, conspicuously or prominently punctate and/or punctate-lineate; the hyaline and inconspicuous staminal tube elo- bate, the apical free portions of the filaments mostly longer than % anther length, at times subequaling to slightly longer than the anthers, the anthers nar- rowly ovoid or lanceoloid, apically apiculate, cau- date-apiculate, subulate-apiculate or rarely mucro- nate-apiculate, basally subcordate, cordate, o = sagittate, dehiscent by subapical pores opening into longitudinal slits, the pores not wider than the slits: pistil obturbinate, the ovary globose to subglobose to oblongoid, the style slender, at times punctate and/or punctate-lineate, the placenta globose or ovoid, the ovules few to numerous, in 2 or more series, when few, at times appearing uniseriate. Fruits subglobose to globose, conspicuously or prominently punctate and punctate-lineate, smooth or costate. Distribution. Auriculardisia is Hidalgo, Pueblo, Ve- racruz, and Oaxaca, Mexico, south throughout Me- Ardisia subg. known from San Luis Potosí. soamerica to Colombia, Ecuador, and in parts of Peru and Venezuela. It grows from sea level to 3300 m in elevation. Ecology. Ardisia subg. Auriculardisia occurs in a wide variety of moist or humid, wet and pluvial habitats in primary, secondary, remnant, and dis- turbed areas, oak-pine woodlands, premontane, montane, cloud, and elfin forests. The subgenus contains 70 species, including 75 taxa, and is defined by asymmetric calyx lobes that are usually subapically notched and always basally auriculate. KEY TO THE SECTIONS OF ARDISIA SUBG. AURICULARDISIA la. Individual flowers subtended by a persistent fo- liaceous floral bract, equal to the size of the flow- ers; inflorescence with each flower subtended by a = rsistent, branch DEIGI сыы Le TERRE AA Auriculardisia Individual flowers often ied by persistent r caducous, non-foliaceous floral bracts muc h smaller than the flowers; inflorescence with flo ers not subtended by a foliaceous inflorescence branch bract. 2a. Inflorescences lateral (axillary) "P sect foliaceous inflorescence act. e ^ © Ф > . Pleurobotryae 2b. {ийген 'ences stric Чу panra or pseudo- terminal. 3a. Leaf blades dimorphic, those of the re- productive shoots usually smaller and often a different shape from mr of the vegetative ones t. Fagerlindia 3b. Leaf blades monomorphic, dks shoots not specialized into vegetative and re- productive ш, and the leaves all of the same general shape and size. 4a. Peduncles at least % the length of the inflorescence sect. Wedelia 4b. Peduncles less than % the length of the inflorescence. and inflorescence ra- chis minutely rufous glandular papillate or mixed with scat- tered uniseriate multicellular ле villous iis 5a. Pedicels Amatlania 5b. I Podios ls and л, е ra- € se chis furfuraceous-lepidote and/ or cupuliform, and/or rufous stipitate-stellate villous, the trichomes with uniseriate, mul- ticellular stalks below tel- ate group of arms, the аце arms often breaking off, leav- ing an apparently villor us tri- chome, the indument ни persistent . Palmanae TAXONOMIC TREATMENT OF ARDISIA SUBG. AURICULARDISIA SECT. AMATLANIA Ardisia subg. Auriculardisia sect. Amatlania (Lundell) Ricketson & Pipoly, comb. et stat. nov. Basionym: Amatlania Lundell, Wrightia 154 982. TYPE: Ardisia liebmannii Oerst. [= Amatlania liebmannii (Oerst.) Lundell]. Shrubs or small trees. Branchlets slender, te- rete, densely and minutely rufous glandular papil- late or mixed with scattered uniseriate multicellular glandular villous hairs, often glabrescent with age. Leaves monomorphic; blades membranous to cori- aceous, elliptic to oblong or obovate to oblanceo- late, mostly inconspicuously or conspicuously but often prominently punctate and punctate-lineate; Inflorescences terminal, petioles slender, terete. Volume 90, Number 2 2003 Ricketson & Pipoly 189 Revision of Ardisia subg. Auriculardisia erect, bipinnately to quadri-pinnately paniculate, pyramidal, shorter or longer than the leaves, usu- ally loosely congested corymbs; inflorescence bracts foliaceous, usually caducous; inflorescence branch bracts and floral bracts often caducous, margins variously toothed, usually with irregularly finely serrate “pectinate” teeth, or regularly dentate or serrate, or rarely entire; pedicels slender, terete. Flowers 5-merous, white, light pink, light purple. red, or blue-violet; calyx lobes essentially free, membranous to chartaceous, ovate to lanceolate, basally auriculate; corolla membranous, the lobes ovate to lanceolate, inconspicuously to conspicu- ously and usually prominently punctate and punc- tate-lineate, densely yellow glandular-granulose within the corolla tube; stamens connate: the fila- ments apically free, connate basally into an elobate tube, free from the corolla tube, epunctate, densely yellow glandular-granulose, the anthers ovoid to lanceoloid, apically apiculate, basally sagittate to cordate, dehiscent by subapical pores, opening into wide, longitudinal slits, the connective punctate; pistil glabrous, the ovary oblong, the style slender, erect, epunctate to inconspicuously punctate, the ovules pluriseriate. Fruits globose, inconspicuously or conspicuously, often prominently punctate and punctate-lineate, costate. Distribution. Ardisia subg. Auriculardisia sect. Amatlania is found from San Luis Potosf, Hidalgo, Pueblo, Veracruz, and Oaxaca, throughout the Me- soamerican region. In South America it is found from Venezuela, Colombia, Ecuador, and Peru. It grows from 10 to 1700 m in elevation Ecology. Members of Ardisia subg. Auricular- disia sect. Amatlania grow in a variety of humid habitats, including tall moist, humid, wet and plu- vial forests, and lowland, premontane, and mon- tane elevations. There are a few species that occur in gallery forest in seasonally dry areas as well. They generally have broad ecological tolerance for disturbance as long as water is not scarce. Ardisia subg. Auriculardisia sect. Amatlania is defined by (1) its vestiture of sparse to dense mi- nute papillae or those papillae mixed with dense to sparse uniseriate, multicellular villous hairs; (2) leaf margins variously toothed, usually with irreg- ularly finely serrate “pectinate” teeth, or regularly dentate or serrate, or rarely entire; (3) dense yellow glandular-granules within the corolla tube and on the filaments, similar to those found in Ardisia subg. Graphardisia. KEv TO THE TAXA OF ARDISIA SUBG. AURICULARDISIA SECT. AMATLANIA la. Leaf blade sogn pr finely serrate, the teeth 7 to 12 per cm, ctinate" ...-----.---------- e Ardisia pellucida 2a. Vestiture of the branchlets, inflorescence ra- у chis, branches, and pedicels with a mixture — —— Ls M | lular glandular villous hairs; calyx sparse to densely and prominently punctate and punctate- lineate. За. Calyx lobes ovate, 1.7-1.9 X 0.9-1.4 mm; corolla lobes 2.2-2.7 mm sa an- thers ovoid, 1.2-1.4 mm long -------- A . Ardisia реа куа ps тее ida 3b. br lobes ovate to lanceolate, 3. .5-1.7 mm; co Eu lobes nee mm long; anthers lanceoloid, 3—3.1 mm . Ardisia pellucida e юла . Ve stiture of the branchlets, inflore 4a. Leaves oblanceolate; calyx lobes snm 2.1-2.2 X .5 mm; ыш я lobes 5— 5.2 x 2.8-2.9 mm; anthers lanceoloid, 3-3.1 X 0.8-1 mm; pistil 4.7—4.8 mm ee A. Ardisia pellucida бер: pec Ab. Leaves elliptic; calyx lobes lanceolate, 5-2.6 X 0.9-1 mm; corolla lobes 3.6— 3.8 X 2.1-2.6 mm; anthers ovoid, 0.9— | X 0.6-0.7 mm; pistil 3.8-3.9 mm un style 2.7-2.9 mm long -------------- 6. Ardisia pellucida subsp. thomascroatü lb. Leaf TEM margins entire to regularly dentate to {л E e б . Corolla lobes 5.6-5.8 X 2.6-2.9 mm; an- thers lanceoloid, apically subulate-apicu- late, 2.5-2.6 mm n Ки 4.6—4.7 mm lon Р 7. Ardisia schippii Corolla Tobes: 2. 7- 3. 2x 1.62 mm; anthers linear-lanceoloid, apically apiculate, 1.9— 2.2 mm long; styles 3.1-3.6 mm long --------- I - Ardisia liebmannü 6a. Vestiture Ж s Мыз and inflores- hr اا‎ ilo. tri- 6 2. Ardisia liebmannii subsp. liebmannii 6b. Vestiture of branchlets and inflores- cence branches with a mixture of dense- ly and minutely rufous glandular papil- lose and densely erect uniseriate аьей glandular villous tri- chomes; vestiture of leaves usually gla- 190 Annals of the Missouri Botanical Garden brous above or sparsely and minutely lular hairs; corolla lobes. 3. x one 4 25-0:5-mm long . Ardisia liebmannii subsp. pow nsis Ardisia liebmannii Oerst., Vidensk. Meddel. Dansk Naturhist. Føren Kjøbenhavn 1861: 129. 1862. Icacorea liebmannii (Oerst Standl., Contr. U.S. Natl. Herb. 23: 1110. 1924. Amatlania liebmannii (Oerst.) Lundell. Wrightia 7: 40. TYPE: Mexico. Vera- cruz: prope Amatlan [de los Reyes], July 1842 (fl), Æ Liebmann 7A (holotype, C!). — Shrubs 0.3-6.1 m tall. Branchlets 1-4 mm diam., glabrescent, or of scattered to densely and minutely rufous glandular papillose, or with a mixture of densely and minutely rufous glandular papillose and scattered to densely erect uniseriate multicel- lular glandular villous trichomes. Leaves with blades chartaceous, oblong to elliptic, 4.8-21.5 х 2.5-9.1 cm, apically acuminate to long acuminate, with an acumen 0.6—4.2 cm long, basally acute to cuneate, decurrent on the petiole, inconspicuously punctate and punctate-lineate, usually glabrous above or sparsely and minutely papillose especially along the midrib, glabrous below or sparsely to scattered minutely papillose, the indument of the midrib often with a mixture of densely and minutely papillose and densely simple multicellular hairs, the midrib impressed above, conspicuously raised below, the secondary veins 10 to 19 pairs, promi- nulous above and below, the margins regularly den- tate to serrate, flat; petioles slender, marginale, 0.6— 2.7 mm long, the leaves. Inflorescences erect, bi- to tripinnately paniculate, 2.5-14 2-13 ст, pyramidal, shorter to longer than the leaf blades, the rachis with vestiture like the branchlets, the branches terminating in 4- to 10-flowered corymbs: peduncles 0.4—3.4 ст long. the lower branches often subtended by leaves: in- vestiture as in inflorescence branch bracts membranous, oblong, 1.2-7.5 X 0.2-3 mm. apically acute, basally sessile, midvein obscure to florescence bracts unknown: impressed above, prominulous below, prominently punctate and punctate-lineate, papillate and/or vil- lous as in the leaves, the margins regularly entire, minutely erose, hyaline, sparsely glandular cilio- late; floral the inflorescence branch bracts, but ovate or deltate, 0.6-2 X 0.2- 0.8 mm, usually glabrous above or sparsely and bracts similar to minutely papillose, glabrous below or with scat- tered minute papillae; petioles slender, terete, 2.8— 6 mm long, inconspicuously punctate and punctate- lineate, papillate and/or villous like the branchlets. Flowers 5-merous, pink to light purple or red; calyx 0.9-1.3 mm, apically acute to acuminate, nearly epunctate to sparsely and conspicuously or inconspicuously to lobes chartaceous, ovate, 1.4—1.9 х prominently punctate and punctate-lineate, gla- brous adaxially, sparsely and minutely papillate abaxially, the margins erose, hyaline, sparsely glan- dular 3.6-4 mm long. the tube 0.4—0.6 mm long, the lobes ovate to lanceolate, 2.7-3.2 X 1.6-2 mm, apically acute, few inconspicuous and prominently punctate and punctate-lineate, glabrous throughout, but densely yellow glandular-granulose adaxially, apically ciliolate; corolla membranous. above staminal tube as well as between the junction of the corolla tube and the lobes, the margins en- tire, hyaline; stamens 3.4—4 mm long, the filaments 1.6-2 mm long, the tube 0.3-0.6 mm long, the api- cally free portions 1-1.7 nm long, epunctate, 1 densely yellow ше the anthers 1.9-2.2 X acuminate apiculate, Tm cordate, the connec- lanceoloid, 1 ).6-0.8 mm, apically long tive epunctate to few conspicuous and prominently punctate; pistil 4—4.5 mm long. the ovary oblong. 0.8-1 mm long, the style 3.1-3.6 mm long, epunc- tate, the ovules 8, biseriate. Fruits globose, 3.6—5 mm diam., usually apically inconspicuously punc- tate and punctate-lineate, prominently costate. Distribution. Ardisia liebmannii is found from Hidalgo. Pueblo, Oaxaca, Veracruz, and Chiapas Mexico, growing from 550 to 1700 m in elevation. Ecology. Ardisia liebmannii occurs in primary and secondary lower montane forests, montane rain forests, and evergreen wet forests. Within Ardisia subg. Auriculardisia sect. Amat- lania, Ardisia liebmannii appears to be most closely related to A. schippii based on its entire to regularly dentate or serrate (not pectinate) leaf margins. However, A. liebmannii may be distinguished from A. schippii by its smaller corolla lobes to 3.2 х 2 mm, shorter linear-lanceoloid anthers to 2.2 mm long with apiculate apices, and the shorter styles to 3.6 mm long. During the 1930s, J. F. Macbride of the Field Museum of Natural History photographed Neotrop- ical types and "authentic specimens" of European herbaria. Unfortunately, the widely circulated pho- tograph of Ardisia liebmannii, F neg. 22951, is from F. Liebmann 7 (C). not of the type specimen, £ Liebmann 7A, and should only be considered a pa- ralype. Volume 90, Number 2 2003 Ricketson & Pipoly 191 Revision of Ardisia subg. Auriculardisia 1. Ardisia liebmannii subsp. jalapensis (Lun- dell) Ricketson & Pipoly, comb. et stat. nov. Basionym: Ardisia jalapensis Lundell, Wrightia 6: 104. 1980. Amatlania jalapensis (Lundell Lundell, Wrightia 7: 40. 1982. TYPE: Mexico. Veracruz: Mpio. de Xalapa [Jalapa], km 7 San Andresito, 1320 m, 4 Aug. 1976 (fr). Zola 610 (holotype, LL!; MEXU!, XAL!). Figure 4 м сагг. isotypes, ЦА Shrubs 1-6 m tall. Branchlets 1-3 mm diam., with a mixture of densely and minutely rufous glan- dular papillose and densely erect uniseriate mul- ticellular glandular villous trichomes. Leaves with blades 5.4—15.8 X 2.5-6 cm, with an acumen 0.6— 2.3 cm long, usually glabrous above or sparsely and minutely papillose especially along the midrib. gla- brous below or sparsely and minutely papillose. the indument of the midrib with a mixture of densely and minutely papillose and densely simple multi- cellular hairs; petioles 0.8—1.7 mm long, vestiture as in the leaves. Inflorescences 3—10.5 X 3—12 cm, vestiture as in the branchlets, the branches termi- nating in 4- to 8-flowered corymbs; peduncles 0.8— 1.7 em long; inflorescence branch bracts 2-6 X 0.3-1.4 mm, densely glandular papillose and/or vil- lous; floral bracts 0.8-1.3 X 0.2-0.4 mm; petioles 4—4.3 mm long. Flowers pink to violet; calyx lobes 1.4—1.5 X 0.9-1 mm; corolla 3.6-3.8 mm юш the tube 0.5-0.6 mm long, the lobes 3.1—3.2 X 1.6- 1.9 mm; stamens 3.4—3.5 mm long, the filaments 1.6-1.7 mm long, the staminal tube 0.5-0.6 mm long, the apically free portions 1-1.2 mm long, the anthers aac. 1.9-2 X 0.6-0.7 mm; pistil 4.4—4 m long, the ovary 0.9-1 mm long, the style 3. 534 6 mm long. Fruits 4.2—4.8 mm diam. Distribution. sis is found in Hidalgo, Pueblo, and Veracruz. Mex- ico, growing from 800 to 1700 m in elevation. Ardisia lieb- Ardisia liebmannii subsp. jalapen- Ecology and conservation status. mannii subsp. jalapensis occurs in montane oa and pine forests. It apparently has some ecological tolerance because it has been collected in second- ary forests, and data suggest that it is currently un- der threat. Within Ardisia subg. Auriculardisia sect. Amat- lania, Ardisia liebmannii subsp. jalapensis is most closely related to A. liebmannii subsp. liebmannii by its leaf blade margins regularly dentate to ser- rate, corolla lobes to 3.2 X 2 mm, linear-lanceoloid anthers to 2.2 mm long with apiculate apices, and styles to only 3.6 mm long. Ardisia liebmannü subsp. jalapensis may be easily distinguished from subspecies liebmannii because of the vestiture of the leaves usually glabrous above or sparsely and minutely papillose especially along the midrib, gla- brous below or sparsely and minutely papillose, the indument of the midrib with a mixture of densely and minutely papillose and of densely simple mul- ticellular hairs along the midrib and the vestiture on its branchlets and inflorescence branches with a mixture of densely and minutely rufous glandular papillose and densely erect uniseriate multicellular glandular villous trichomes, the longer corolla lobes to 3.2 X 1.9 mm, the shorter anthers to 2 mm long. and the longer styles to 3.6 mm long. pe cimens examined. MEXICO. Hidalgo: Mpio. Te- Doria, : © S ® С 5 £e =] 9s а. T =R м 5 le B 5 Р دو‎ д poaxtla, 21 June 1977 (fl), M. Martínez et al. 90 (MEXU, MO); Mpio. de Pahuatlan, Xopanapa 8 km al SW de Pa- huatlan, 22 May 1986 (fl), P. Tenorio L. & C. Romero de T. 11375 (MEXU, MO); Mpio. Ahuacatlán, path to Zapo- Пап, Agua Dulce, 4 km SE of Ahuacatlán, 2 July 1987 (fl), С. Toriz A. et al. 584 (FTG, MEXU). Veracruz: Mpio. Huatusco, trail to Горение 4 ky from the Huatusco— Coscomatepec 1 Aug. 1979 (fl), 5. Avendaño К. & J. Calzada 403 (Е LL); Mpio. hale Cerro Cercano, Río Seco, along Huatusco-Coscomatepec Hwy., 7 Nov. 1979 (fr), S. Avendaño R. 556 (F, LL); Mpio. Juchique de Ferrer, Cerro de Villa Rica, 6 May 1981 (fl), G. Castillo C. et al. 1764 (F, L Hernández M. & I. Calzada 1536 (F, ME chique Ferrer, La Cima, Las Hayas, 21 June 1972 (fl), R. Hernández M. 1565 (F, MEXU); Mpio. Tlapacoyan, ca. 6 km by air S of Tlapacoyan on road to Altotonga, 11 July 1982 (fr), M. Nee & G. Diggs 24883 (F). 2. Ardisia liebmannii subsp. liebmannii. Figure 5. Зуп. nov. /сасогеа crenipetala (Mez) Standl., U.S. Natl. Herb. 23: 1110. 1924. ius eripi (Mez) Lundell, Wrightia 7: 41. 1982. TYPE: Mexico. Veracruz: Orizaba, 1856 (fl), M. Botteri 146 reed designated by C. Lundell (1982), G!: isolectotypes, GH!, LL!, US)). Ardisia rekoi Lundell, in i Bot. Mus. Leafl. 9: 185. 1941. Syn. nov. TYPE: titlin, barranca Nin-du Au San Antonio Eloxo- chitlán, 18°21'N, 096°45'W, 1100 m, 24 July 1938 (fl), i, ES cnm & B. Reko 273 (holotype, MICH!, LL -12!; isotype, GH!). Amatlania pida Lundell, Phytologia 56: 19. 1984. Syn. Ardisia elliptifolia Lundell, k up 61: 63. 1986. non Ardisia elliptica Thunb., Nov. Gen. РІ. € 119. 1798. TYPE: Mexico. calde Mpio. de Cý- maltepec, km 149 carr. Tuxtepec, Puerto Eligio, 800 m, 17 June 1966 (fr), G. Martínez- Calderón 884 (holotype, XAL not seen; isotype, MO)). Ardisia crenipetala Mez, in Engl., Pflanzenr. IV. 236 (Heft 9): 91. E Shrubs 0.3-6.1 m tall. Branchlets 1.5-4 mm diam., glabrescent, or of scattered to densely and 192 Annals of the Missouri Botanical Garden Volume 90, Number 2 2003 Ricketson & Pipol 193 Revision of Ardisia subg. Auriculardisia minutely rufous glandular papillose, or with a mix- ture of densely and minutely rufous glandular pa- pillose and scattered erect uniseriate multicellular glandular villous trichomes. Leaves with blades 4.8-21.5 х 2.5-9.1 cm, with an acumen 1—4.2 ст long, usually glabrous above, scattered minutely papillose below; petioles 0.6-2.7 mm long. vesti- x 2-13 the branches terminating in 4- to 10-flowered corymbs: ture as in the leaves. Inflorescences 2.5-14 cm, with the vestiture like the branchlets, peduncles 0.4—3.4 cm long: inflorescence branch 1.2-7.5 X 0.2-3 mm, with scattered minute papillae; floral bracts 0.6-2 X bracts membranous, 0.3-0.8 mm; petioles 2.8-6 mm long, papillose or villous as in the branchlets. Flowers lavender or pink to light purple or red; calyx lobes 1.5-1.6 1-1.3 mm; corolla 3.9-4 mm long, the tube 0.4— 0.5 mm long, the lobes 2.7-2.8 X 1.9-2 mm; sta- mens 3.9—4 mm long, the filaments 1.9-2 mm long, the staminal tube 0.3-0.4 mm long, the apically free portions 1.5-1.7 mm long, the anthers 2.1-2.2 0.7-0.8 mm: pistil 4-4.1 mm long, the ovary 0.8—0.9 mm long, the style 3.1—3.2 mm long. Fruits 3.6-5 mm diam. Distribution. Ardisia liebmannii subsp. lieb- mannii is restricted to Chiapas, Oaxaca, and Ve- racruz, Mexico, growing from 550 to 1680 m in elevation. Ardisia lieb- mannii subsp. liebmannii occurs in primary an Ecology and conservation status. secondary lower montane forests, montane rain for- ests, and evergreen wet forests. Because it is rela- tively uncommon, it should be considered threat- Within Ardisia subg. Auriculardisia sect. Amat- lania, Ardisia liebmannii subsp. liebmannii is most closely related to A. liebmannii subsp. jalapensis (see under that мт ies for a discussion). Ardisia liebmannii subsp. liebmannii may be easily distin- guished from subspecies jalapensis because the vestiture of the leaves is usually glabrous above, with scattered minute papillae below, and the ves- titure on its branchlets and inflorescence branches is glabrescent, or of scattered to densely and mi- nutely rufous glandular papillose, or with a mixture of densely and minutely rufous glandular papillose and scattered erect uniseriate multicellular glan- dular villous trichomes, the shorter corolla lobes to € 2 mm, the longer anthers to 2.2 mm long, N and the shorter styles to 3.2 mm long. Populations corresponding to the type of Ardisia crenipetala are unique only for the more scattered and slightly longer multicellular glandular villous hairs in the inflorescence; otherwise it matches Ar- disia liebmannii subsp. liebmannii in all other re- spects. In Mez’s (1902) original description of Ar- disia crenipetala, he listed three syntypes, Botteri 1, and Conzatti 169, from G-DC, P, and ЄН. Suwdles (1924) did not list any specimens in his treatment for Mexico. Lundell (1982), however, in- directly lectotypified the Botteri 146 collection at G as the lectotype in his treatment of Amatlania. Ardisia rekoi was originally described as being related to Ardisia nigrescens Oerst. However, study of the type shows that it has auriculate calyx lobes, which clearly place it in Ardisia subg. Auricular- disia. The type of Amatlania elliptica was collected in fruit and represents populations unique for their slightly wider leaves with shallow dentate margins. Specimens examined. MEXICO. Chiapas: Mpio. of Tenejapa, slope at the sumidero at the market place of Yochib, paraje of Kotol Te', 16 July 1965 a. D. Breedlove 11081 (LL); boca Highlands, 6 P E, by road, of Bochil on hwy. 195, 10 Aug . 1965 (fr Кая a © ® ^ t a a ~ ~ ~ Mirad or, 15 К Valle Nacional, 16 Oct. (fr), J. Meave del Castillo et al. 1514 (MO); Mpio. Angel Albino Corzo, bg near Rancho Viejo of мы Finca Pru- sia, 23 Jan. 1968 (fr), A. Shilom T. 3567 (F, LL, NY). Veracruz: ы del Borrego, Oct. 1856 (fl), M. Botteri 481 (G, K Cantón de Cordoba, Colonia Melchor Ocampo, 19 June 1896 (fl), C. Conzatti 169 (GH); Orizaba, 1855 (fl), M. Cuming es (G); Mpio. ep El Haya camino Naolinco, Misantla, 27 Б, 1971 (fl), J. робни 242 (MEXU); Orizaba, е 1905 (fl), A Purpus 1242 (F, GH, MO, NY); El Paso Berero between Atzacan and Dos | Figure 4 (left). surface. —C. Detail of basal area of abax aharia surface. шыл Stamen, vein margin Zola B. 610 (LL); C, Avendaño R. 556 (F).) Figure 5 nn leaf surface. —C. Detail of basal area of abs abaxial surface. oe Stamen, dus ral margin. —H. amen, adaxial surface. —I. Fruit. (А tam M. E-H from M. Marinës et Д 90 (MEXU); D from R. Hernández M. 1565 (MEXU); I from 5. Stamen, Ardisia ipid subsp. jalapensis. —A. Flowering branch. —B. Detail of margin of abaxial leaf ial leaf surface. 1 —D. Detail of inflorescence. — ower. —F. Stamen, ‚ B drawn from holotypé, Ardisia liebmannii subsp. liebmannii. —A. Flowering branch. —B. т of margin of abaxial xial leaf surface. —D. Detail of p ence. F.S Flower. štamen, adaxial surface. —l. -C ah fon the holotype, 7 (F)) F. Liebmann 7A (C); D from К пена T ( 7); E-H from S. Sohmer 9517 (F); 1 Ais A. Shilom T. 356 194 Annals of the Missouri Botanical Garden Rfos about 10 km M of Orizaba, 15 oF 1974 (fl), Sohmer 9517 (F, MEXU); Tenango al N Río Blanco, 11 July 1983 (fl), R. pani C. & H. А pA (MEXU), 2 (MO). Ardisia pellucida Oerst., Vidensk. Meddel. Dansk Naturhist. Fgren Kjøbenhavn 1861: 130. 1862. /cacorea pellucida (Oerst.) Standl., Contr. U.S. Natl. Herb. 23: 1110. 1924. Amat- lania pellucida (Oerst.) Lundell, Wrightia 7: 40. 1982. TYPE: Mexico. Veracruz: prope Pi- tal, (fl), Æ Liebmann 29C (holotype, С!, F neg. 22055! Shrubs or subshrubs 0.3—5(—10) m tall, 2.5—6 cm diam. Branchlets 2-9 mm diam., scattered to densely and minutely rufous glandular papillose, or with a mixture of densely minutely rufous glandular papillose and scattered erect uniseriate multicel- lular glandular villous hairs. Leaves with blades membranous to chartaceous, elliptic, obovate to ob- lanceolate, 6.5—64.2 X 3-20.6 cm, apically acute to acuminate, with an acumen 0.4—4.8 cm long, ba- sally obtuse to cuneate or oblique or auriculate, usually decurrent on the petiole, inconspicuously or conspicuously and usually prominently punctate and punctate-lineate, glabrous above, sparsely and minutely papillose, denser along the midveins and secondary veins, the midrib impressed above, con- spicuously raised below, the secondary veins 12 to 39 pairs, prominulous above and below, the mar- gins irregularly finely serrate, the teeth 7 to 12 per cm, “pectinate,” flat; petioles slender, marginate, 0.9-3.6 cm long, glabrous above, sparsely to dense- ly and minutely papillose. Inflorescences erect, bi- to quadri-pinnately paniculate, 2.8—28.2 2.6— 22.7 cm, pyramidal, shorter or longer than the leaves, vestiture of the rachis similar to the branch- lets, the branches terminating in 5- to 19-flowered corymbs; peduncles obsolete to 2.1 cm long, the lower branches often subtended by leaves; inflores- cence bracts obsolete; inflorescence branch bracts P Pani. linear or oblong to elliptic, 2.8—12.9 0.2-2.9 mm, apically acute to attenuate, sparsely to densely and conspicuously or prominently punc- late and punctate-lineate, the margins entire except minutely erose apically, hyaline, sparsely glandular ciliolate or irregularly finally serrate “pectinate.” the midvein obscure to impressed above and raised below, the secondary veins obscure, floral bracts 0.6-3.3 х 0.2-0.8 mm, api- the veins obsolete, membranous, linear, cally acute, basally sessile, sparsely to densely inconspicuously or conspicu- ously or prominently punctate and punctate-lineate, the margins entire except minutely erose apically, hyaline, sparsely glandular ciliolate; pedicels slen- der, 3.1-12.5 mm long, inconspicuously punctate and punctate-lineate, sparsely to densely minutely papillose. Flowers 5- or 6-merous, lav- ender to red-violet or blue-violet to deep purple; calyx lobes essentially free, membranous to char- laceous, ovate to lanceolate, 1.7-3.2 0.9-1.7 mm, apically acute to long acuminate, notched be- low the apex, rarely lacking a notch, basally auric- ulate, sparsely to densely and inconspicuous or terete, conspicuously punctate or punctate-lineate, and of- ten prominently punctate and punctate-lineate, gla- brous adaxially, sparsely minutely papillose, the margins erose, hyaline, sparsely glandular ciliolate; corolla membranous, 3.5-6.1 mm long, the tube 0.7-1.4 mm long, the lobes ovate to lanceolate, s X 1.8-2.9 mm, apically acute, epunctate | bearing a few inconspicuous or conspicuous ا‎ and punctate-lineations, glabrous ada- xially, or densely yellow glandular-granulose api- cally above the staminal tube as well as between the junction of the corolla tube and lobe, glabrous outside or sparsely minutely papillose medially, the margins entire, hyaline; stamens 2.3-4.8 mm long, the filaments 1.3-2.5 mm long, the staminal tube 0.2-0.8 mm long, the apically free portions 1-1.8 mm long, epunctate to slightly punctate, glabrous to densely yellow glandular-granulose, the anthers ovoid to lanceoloid, 0.9-3.1 X 0.6-1 mm, apically apiculate, basally cordate to sagittate, the connec- tive epunctate to inconspicuously punctate; pistil 2.5-4.8 mm long, the ovary oblong, 0.9-1.5 mm long, epunctate or conspicuously punctate and punctate-lineate, the style 1.5-3.3 mm long, epunc- tate or inconspicuously punctate, the ovules 9 to 37. Fruits globose, 3.9-5.8 mm diam., inconspic- uously punctate and punctate-lineate, costate. Distribution. Ardisia pellucida is found from San Luis Potosf, Pueblo, Veracruz, Oaxaca, Tabas- co, and Chiapas, Mexico, and throughout Mesoa- merica. In South America it is found from Vene- zuela, Colombia, Ecuador, and Peru, growing from 35 to 1650 m in elevation. Ecology. Ardisia pellucida occurs in primary or secondary humid, moist, wet, or pluvial tropical for- ests from the lowlands to premontane levels, rarely in cloud forests where it occurs on low, isolated mountain pea rdisia в is a highly variable species viti Ardisia subg. Auriculardisia sect. Amatlania. From the other species in section Amatlania, it may be recognized by its leaf blade margin irregularly finely serrate with teeth 7 to 12 per cm, “pecti- nate," whereas the leaf blade margins in Ardisia schippii and A. liebmannii are regularly dentate to Volume 90, Number 2 2003 Ricketson & Pipol 195 y Revision of Ardisia subg. Auriculardisia serrate with teeth 3 to 5 per cm or, very rarely, entire. 3. Ardisia pellucida subsp. lancetillensis Ric- ketson & Pipoly, subsp. nov. TYPE: Honduras. Atlántida: hills above Lancetilla, 1500 ft. [457 m]. 15 July 1934 (fl), T. Yuncker 4588 (holo- type, MO!; isotypes, A!, F!, NY!). Figure 6. Subspecies haec a subsp. pellucida, payee calycinis 3.1-3.2 (nec 1.7-1.9) mm longis 1.5-1.7 c 0.9-1.4 mm latis, lobulis corollinis 4.3—4.5 (non 2. 2-2. " mm lon- gis, denique antheris 3.0-3.1 (non 1.2-1.4) mm longis praeclare distat Shrubs 1—3 m tall. Branchlets 5-9 mm diam., sparsely to densely minutely rufous glandular pa- pillose, or with a mixture of rufous glandular pa- pillae and uniseriate multicellular glandular villous trichomes. Leaves with blades oblanceolate, 13.7— 64.2 X 3.9-20.6 cm, with an acumen 0.7—3.6 cm long, the secondary veins 21 to 29 pairs; petioles 0.9-3.6 cm long. Inflorescences bi- to tripinnately paniculate, 8.5—28.2 X 5.6-22.7 cm, shorter or longer than the leaves, the peduncle, vestiture of the rachis, and branches similar to the branchlets, the branches terminating in 11- to 19-flowered cor- ymbs; peduncles obsolete to 4.8 cm long; inflores- cence branch bracts linear, 3.8—12.9 х 0.8-2.9 mm, densely and prominently punctate and punc- tate-lineate, the veins obscure, the margins minute- ly erose, hyaline, sparsely glandular ciliolate; floral bracts 1.5-3.3 X 0.2-0.6 mm, densely and prom- inently punctate and punctate-lineate; pedicels 3.1-12.5 mm long, inconspicuously punctate and punctate-lineate, densely minutely papillose. Flow- i blue-violet, violet-purple to deep purple; calyx lobes ovate to lanceolate, 3.1- 3.2 X 1.5-1.7 mm, apically acute to long acumi- nate, without notch below the apex, sparsely to ers 5-merous, lilac, prominently punctate and punctate-lineate; corolla .4—5.5 mm long, the tube 0.9-1 mm long, the lobes 4.3—4.5 X 2.2-2.3 mm, sparsely but con- spicuously punctate and punctate-lineate, densely yellow glandular-granulose adaxially, apically above the staminal tube, as well as between the corolla tube and lobe junction, sparsely and mi- nutely papillose medially outside; stamens 3.8—4 mm long, the filaments 1.3-1.6 mm long, the sta- minal tube 0.2-0.5 mm long, the apically free por- 1.1-1.2 mm, inconspicuously punctate, densely yellow glandular-granulose, the anthers lanceoloid, 3-3.1 X 0.9-1 mm; pistil 3.1-3.2 mm long. the ovary 0.9-1 mm long, conspicuously punctate and punctate-lineate, the style 2.1-2.2 tions mm long, inconspicuously punctate, the ovules 35 to 37. Fruits 5.5-5.8 mm diam. Distribution. Ardisia pellucida subsp. lanceti- llensis is endemic to the area around the Lancetilla Valley, near Tela and the Jardín Botánico Lancetilla Biological Reserve in Atlántida, Honduras, growing between 10 and 600 m in elevation. Ecology and conservation status. Ardisia pellu- cida subsp. lancetillensis occurs in disturbed pri- mary and secondary rain forests, which receive up to 4500 mm annual precipitation. An exhaustive search was conducted to look for this subspecies, without success, so we may assume the subspecies is threatened. Etymology. Lancetilla Valley region, near Tela, Atlántida, on the north coast of Honduras. We are very grateful to Ing. Ciro Navarro, Director of the Jardín Botánico de Lancetilla, who brought this subspecies to our attention and helped Pipoly look for it in the field. Within Ardisia subg. Auriculardisia sect. Amat- The subspecies is named for the lania, Ardisia pellucida subsp. lancetillensis shares with subspecies pellucida a similar vestiture with a mixture of densely minutely papillose and scat- tered, erect, long, simple multicellular hairs on the upper branchlets and throughout the inflorescence branches, but subspecies lancetillensis may be dis- tinguished by its larger ovate to lanceolate calyx lobes to 3.2 mm long, larger corolla lobes to 4.5 mm long, and larger lanceoloid anthers to 3.1 mm long. Paratypes. HONDURAS. Atlántida: Reserva del Jar- dín Botánico de Lancetilla, Sep. 1979 (fl), M. Antonio V. 15 (MO); Lancetilla Valley, Lancetilla, 22 June-27 July 1929 (fl), A. Chickering 200 (F); Lancetilla Valley, ca. 10 mi. SE of Tela, in forest preserve along Río Lancetilla, on trail to water reservoir, 3 Aug. 1977 (fl), T. Croat 42632 (MO); along road for municipal water supply of Tela, Lan- cetilla Botanical perdent, on road ca. 2 mi. WSW of Tela and S of main hwy., 9 Feb. 1987 (fr), T. Croat & D. Han- non 64637 (MO); basin in front of old dam, 21 Aug. 1999 (fl), R. Cruz 225 (HJBL); permanent Plot area, Lancetilla Reserve, 12 Nov. 1978 (fr), D. Hazlett 2984 (HJBL); Lan- cetilla, 10 Aug. 1978 (fl), D. Hazlett 3058 (HJBL); upper dam, Lancetilla Valley, 29 July-10 Aug. 1951 (fl), К. How- ard et al. 500 (A): Tela, Lancetilla Valley, above Experi- above stream but (MO); trail between arboretu 18 Jan. 1994 (fr), C. Nelson 17 325 (HJBL); vicinity of San Alejo, at base of hills S of San Alejo near Río San Alejo [ca. 15 km SW of Tela], 22-27 Apr. 1947 (ster.), P. Stand- ley 7729 (F); Lancetilla Valley, пе 20 Mar. 1928 (fr), US), 53322 (F, US); Lancetilla > 30 July 1962 (fl), G. Webster et wy "2681 (F, LL); Ta River 2 mi. above Puerto Sierra, 18 Jan. 1903 (fr), P. Wilson 79 (NY, US) 4. Ardisia pellucida subsp. pectinata (Donn. Sm.) Ricketson & Pipoly, comb. et stat. nov. 196 Annals of the Missouri Botanical Garden Volume 90, Number 2 2003 Ricketson & Pipoly 197 Revision of Ardisia subg. Auriculardisia ger pes pectinata Donn. Sm., Bot. Gaz. 12: . 1887. Icacorea -r (Donn. Sm.) cH Саш U.S. Natl. Herb. 23: 1110. 1924. Ardisia pellucida Oerst. var. pectinata (Donn. Sm.) Lundell, Wrightia 3: 99. 1964 Amatlania pellucida (Oerst.) Lundell var. pec- tinata (Donn. Sm.) Lundell, Wrightia 7: 40. 1982. Amatlania pectinata (Donn. Sm.) Lun- dell, Phytologia 55: 235. 1984. TYPE: Gua- temala. Alta Verapaz: Pansamalá, 3800 pp [1158 m]. June 1886 (fl), H. von Türckheim 942 (holotype, US!, US neg. 2381*; isotype, GH!). Fi Shrubs 3—5 m tall. Branchlets 4—7.5 mm diam., scattered to densely rufous glandular papillose. Leaves with blades oblanceolate, 6.5—41 X 3-12.3 cm, with an acumen 0.7—4.2 cm long, the second- ary veins 12 to 31 pairs; petioles 0.9—2.1 cm long. Inflorescences tri- to quadri-pinnately paniculate, 6.2—20.1 X 6.1—22.5 cm, shorter or longer than the leaves, vestiture of the rachis, branchlets, abaxial bract surfaces, and pedicels similar to the branch- lets, the branches terminating in 5- to 16-flowered corymbs; peduncles obsolete to 2.1 cm long; inflo- rescence branch bracts linear, 3.2—4.9 х 0.2-0.8 mm, sparsely and inconspicuously punctate and punctate-lineate, the veins obsolete, the margins minutely erose, hyaline, sparsely glandular cilio- late; floral bracts 1.5-2.6 X 0.2-0.6 mm, sparsely and inconspicuously punctate and punctate-lineate; pedicels 6.2-7.8 mm long, inconspicuously punc- tate and punctate-lineate. Flowers 5- or 6-merous, lavender to red-violet; calyx lobes ovate, 2.1—2. 1.4-1.5 mm, apically acute, sparsely inconspicu- ously punctate and punctate-lineate; corolla 5.9— 6.1 mm long, the tube 0.7—1.1 mm long, the lobes 5.2 x 2.8—2.9 mm, mostly epunctate, rarely or dd: with glandular-granulose papillae around the top of the corolla tube inside, sparsely and mi- nutely papillose medially outside; stamens 4.3—4.8 mm long, the filaments 1.8-2.2 mm long, the sta- minal tube 0.5—0.7 mm long, rarely or sparsely with glandular-granulose papillae on the staminal tube and filaments, the apically free portions 1.1-1.7 mm long, epunctate, the anthers lanceoloid, 3-3.1 X 0.8-1 mm, basally subcordate; pistil 4.7—4.8 mm long, the ovary 1.4—1.5 mm long, epunctate, the styles 3.2-3.3 mm long, epunctate, the ovules 21 to 22. Fruits 5.5—5.8 mm diam. Distribution. Ardisia pellucida subsp. pectinata is found from Mpio. Rayón in Chiapas, Mexico, to Baja Verapaz and Izabal, Guatemala, growing at 50—)1158-1851 m elevation. Ecology and conservation status. — Ardisia pellu- cida subsp. pectinata occurs in primary premontane and montane wet forests. Because of its restricted distribution, it should be considered threatened. Etymology. The specific epithet was derived from the Latin, meaning divisions, like a comb” referring to the teeth on the leaf blade margins. Within Ardisia subg. Auriculardisia sect. Amatla- nia, Ardisia pellucida subsp. pectinata shares a similar vestiture with subspecies thomascroatii, sparsely to densely, minutely, rufous glandular papillose through- out. However, Ardisia pellucida subsp. pectinata may be distinguished by its oblanceolate leaves, smaller but wider ovate calyx lobes to 2.2 X 1.5 mm, larger corolla lobes to 5.2 X 2.9 mm that are sparsely and minutely papillose medially outside and sparsely glandular-granulose inside around the corolla tube and filaments, the larger lanceoloid anthers to 3.1 X 1 mm, the larger pistil to 4.8 mm long, and the longer style to 3.3 mm long. *with narrow closely-set Specimens examined. MEXICO, Chiapas: Mpio. Ray- ón, near Puerto del Viento, 9 mi. NW of Pueblo Nuevo So lis tahuacán along road to Tapiula, 20 Aug. 1965 (fl), D. Bre elo ve 11998 (F, LL, NY); Selva Negra 10 km above Rayón mezcalapa along road to ru 2 Jan. 1981 (fr), D. Breedlove & B. Keller 49317 (LL, ), NY). e MALA. Baja Verapaz: along dirt olie h mi. NE of P rulhá, 17 July 1977 (fl), T. Croat pe 325 (MO). Izabal: Rio Juyamá, SE of Cheyenne, about 15 mi. SW of Bananera, 8 Apr. 1940 (fl, fr), J. Lana 39114 (F [inflorescence galled]. 5. Ardisia pellucida subsp. E Figure 8. Ardisia myriodonta Standl., J. Wash. Acad. 17: 1927. Syn. nov. Алайа irr ida (€ die ) ug <— Figure 6 (left). Jetail of inflorescence. — late a margin. —H. Fruit. (A Silanes enc ce) & Ardisia pellucida a lancetillensis. Flow =E. Sta men, abaxial surface. C drawn from isotype, T. Yuncker 4588 (GH); A —A. Flowering branch. —B. Detail of abaxial leaf Dres —К. Stamen, adaxial surface (leaf) & B from ne ре. . — С. Stamen, Т Yuncker 4588 (Е): D-G from G. Webster et al. 12681 (LL); Н from P. Standley 53322 (V).) Ae 7 (right). Detail of inflorescence. — en, abaxial surface Tec margin. —H. Fru saa ron holotype, H. von Tiirckheim Ardisia pellucida gal pectinata. —A. Flowering branch. —B. Detail of perde Lond surface. — D. Flower. —E. Stam Am Stamen, adaxial surface. —G. Stamen, US); C from isotype, H. von Türckheim Y).) t. (A, 2 ( 942 (GH); D-G from D. ане 11998 (Е); Н from D. Breedlove & В. EE. 49317 (N Annals of the 198 Missouri Botanical Garden ((OW) OZZSEĘ 017) iL *ad&jo[ou шолу umeap су—ү) "әәрипв [erxepe uwes “Q— ‘шеш [213948] *uaure]S :J— ‘voRpins [erxeqe ‘чәшес 7— чәмо]д '(]J— '92u22sa10pur Jo req 77)— `әәвипз peo [erxeqe jo eLA *g— "youg1q Зицәмор "ү "1po42spuiow] "dsqns. Dpionjod. oisipay (apu) бәллү CCTD 6601 1m 12 iJ zonbzpA ‘W woy H (чула) 09r jourumg а wo 9-0 (AMA) 028 421964 ¥ шош 7) :() 26g чиршдәту у ‘adXjojoy шолу имелр gp ^y) ama "Н ‘шлеш [R19je| “йәш; *£)— -»oegns [erxepe. “иәшец Э сәәвипв [erxeqe ‘uwes 7— moy '(]— '22uooso10gur jo jeq ')— uns jeo[ [erxeqe jo peq ^g— “Yours Zuuawo[4 "у “ppionyad *dsqns vpronyəd visipiy (ә) 9 aniy E 3 Volume 90, Number 2 2003 Ricketson & Pipoly 199 Revision of Ardisia subg. Auriculardisia var. ms (Standl.) Lundell, Wrightia 7: 40. 1982. TYPE: Panama. Panamá Zone, Barro се Island in Gattin Lake, ca. 120 m or less, 18—24 Nov. 1925 (fr), P. Standley “10848 (holotype, US!, US neg. 2376!). Subshrubs 0.3-3(-10) m tall, 2.5-6 cm diam. Branchlets 2—6.5 mm diam., sparsely or densely ru- fous glandular papillose, or with a mixture of rufous glandular papillae and uniseriate multicellular glandular villous hairs. Leaves with blades 8.7—42.9 х 3.7-14.6 cm, with an acumen 0.4—4.8 cm long, the secondary veins 14 to 29 pairs; petioles 1.1— 3.1 cm long. /nflorescences bipinnately paniculate, 2.8-15.2 x 2.6-10.2 cm, shorter than the leaves, vestiture of the rachis, branches, abaxial bract sur- faces, and pedicels similar to the branchlets, the branches terminating in 9- to 16-flowered corymbs; peduncles obsolete to 2 cm long; inflorescence branch bracts linear or oblong to elliptic, 2.8-11.2 X 0.6-2.6 mm, midvein obscure to impressed above and raised below, densely and prominently punctate and punctate-lineate, the margins minute- ly erose, hyaline, sparsely glandular ciliolate or ir- regularly finally serrate “pectinate”; floral bracts 1— ) mm, densely and prominently punctate and punctate-lineate; pedicels 3.2—4.: mm long. inconspicuously punctate and punctate- lineate. Flowers lavender, red-violet to deep purple; calyx lobes ovate, 1.7-1.9 X 0.9-1.4 mm, apically acute, slightly notched below the apex, sng p punctate and punctate-lineate; corolla 3.5-3.8 m long, the tube 0.7—1.3 mm long, the vus 2.2-2.7 x 1.8-2.2 mm, with few but conspicuous puncta- tions and punctate-lineations, densely yellow glan- dular-granulose adaxially. apically above the sta- minal tube as well as between corolla tube and lobe junction, sparsely and minutely papillate medially outside; stamens 2.3-2.5 mm long, the filaments 1.3-1.5 mm long, the staminal tube 0.3-0.4 mm long. the apically free portions 1-1.1 mm, densely yellow glandular-granulose, the anthers ovoid, 1.2— 1.4 X 0.6-0.8 mm, pistil 2.5-2.8 mm long, the ovary 0.9-1.2 mm long, conspicuously punctate and punctate-lineate, the style 1.5-1.7 mm long. epunctate or inconspicuously punctate, the ovules 9 to 22. Fruits 3.9-5.7 mm diam. Distribution. Ardisia pellucida subsp. pellucida is found from San Luis Potosí, Pueblo, Veracruz, Oaxaca, Tabasco, and Chiapas. Mexico, through Honduras, Rica, and Panama. In South America it is found Belize, Guatemala, Nicaragua, Costa from Venezuela, Colombia, Ecuador, and Peru. It grows from 35 to 1650 m in elevation. Ardisia pellu- cida subsp. pellucida occurs in moist, humid, wet. Ecology and conservation status. or rain forest, especially at the margins of natural gaps, and at the forest edge along watercourses. It can tolerate moderate to deep shade, but not com- paction of soil. At this time we see no immediate threat to this subspecies. “Tapacajete” (L. Williams Common Names. Within Ardisia subg. Auriculardisia sect. Amat- lania, Ardisia pellucida subsp. pellucida appears to be more closely related to subspecies lancetillensis than to the other subspecies because of the mixed vestiture on the vegetative parts consisting of mi- nute papillae and villous trichomes. However, sub- species pellucida may be distinguished from sub- species lancetillensis by its smaller ovate calyx lobes to 1.9 mm long, smaller corolla lobes to 2.7 mm long, and smaller ovoid anthers to 1.4 mm long. The type of Ardisia myriodonta is in fruit; how- ever, because of its similar vestiture and small ovate calyx lobes it matches A. pellucida subsp. pellucida in all respects. Specimens examined. MEXICO. Chiapas: Mpio. Oco- zocoautla de Espinosa, 32 km N of Ocozocoautla along o Mal Paso, 19 Oct. 1965 (fr), D. Breedlove & P. Raven 13562 (LL); Mpio. Palenque, 6-12 km 5 of Palen- que on road to Ocosingo, 27 July 1972 (fl), D. Breedlove 26511 (LL). Oaxaca: Mpio. im María Chimalapa, ca. 5 km SW of Santa María along ro á 17 July 1984 (fl), H. Hernández С. 211 (MO); Dtto. the chitán, Los Angeles 20 km NW of intersection with Matías Romero—Acayucan ba s on path before Martín Dehes 23 Oct. 1987 (fr), C. de od 1074 (FTG, MEXU, MO). Puebla: Mesa de San Diego, Mar. 1951 (fr), H. Bravo = 268 (MEXU); Mpio. кы, Negras, El Salto, 9 km N of La Ceiba, 25 Feb. 1987 (fr), A. Campos et al. 64 i MEXU). San Luis Potosí: Mpio. Xilitla, Poblado Xum- chiaio, 12 Aug. 1976 (fr), J. Calzada 2572 (MEXU); 10 mi. NW of Tamazunchale, 3 July 194 0 (f), C. Hitchcock & L. Stanford 6930 (V). of dress along road between Teapa and Tacotalpa, 3.1 . E of Teapa along stream and cliffs ca. 4 mi. S of Wy., 19 du 1987 (fr), T. Croat & D. Hannon 65343 (LL, MO). Ver z: Mpio. Catemaco, Arroyo Basuras, Sontecoma- pan, “10 Hu 1971 (A); Misantla, July 1912 (fl, fr), R. Her- iia M. 1227 (F), C. Purpus 5959 (BM, F, GH [2], LL. ; А Htidalgdlidan, La Escuadra, 17°18'N, ae 8'W, 16 Sep. 1974 (fr), M. Vazquez T. et al. 1055 LL, MEXU). BELIZE. Toledo: headwaters Río Grande, 21 Apr. 1933 (8). W. geen 8-559 (F). GUATEMALA. a Verapaz: near Tactic, 5 Apr. 1939 (ster.), P. Standley 70506 (F, LL). Escuintla: ecc of Río Burrión, NE of Escuintla, 16 Mar. 1941 (fl, fr), Р ү зен 89600 (Е). ~ 'an, Sie (fr), J. Steyermark 49320 (F). and Laguna Izabal, Montafia del Mico, 4 Apr. : E hia 38745 (F, LL). Petén: Chinc Er T km n Luís on Sebol Road, 9 Oct. 1966 (fr), E. Con- treras p (LL [3]; La Cumbre, 4 km E on Río Purula Road, 21 Sep. 1975 (fr), C. Lundell & E. Do 19914 (LL). Santa Barbara: Sololá, Aug. 1891 (fl), W. Shannon 200 Annals of the Missouri Botanical Garden 1256 |J. зн Ms 170| (US). HONDURAS. Atlán- tida: vicinity eiba, lower slopes of Mt. Cangrejal, June-Aug. 1938 A T. ms et al. 8832 (F, G, MO, NY, TEX, US). NICARAGUA. Boaco: Cerro Mombachi- to, SE of Boaco, 30 5s p. i (fr), P. Moreno 3193 (MO); upper SW slope of Cerro Mombac s S of road between Boaco and Camoapa, 3 Oct. 1979 W. Stevens et al. . Jinotega: Сайо VER Río Boc ау, 9 Маг. 1980 (fr) W. Stevens 16651 (MO). Riva ~ — a о = = 5 3 Е z | = t = = 5 o zm 2 دع‎ — z = 2 e = a = — — = 5 — = E © e Е y: a = - = 5 E z = = 2 9 Зарой, 12 Sep. 1982 (fr), J. Sandino 3575 (MO, NY). A PEA along the banks of Río Prínzapolka, ca. 2 km S of Waní, 16 Mar. 1979 (fr), J. Pipoly 4713 (MO, NY); Mpio. de Waspám, Reserva Bosawás, between Cerro la Francia and the summit of Hill Tara (Asang, nuh- ni), 20 Jan. 1996 (fr), R Rueda et al. 3974 (MO). COSTA RICA. Alajuela: about 5 km S of Canalete ! з the Río Zapate and along the new road to Upala, 12 Nov. 197: Llanura E San (fr), W. Burger & R. Baker 10008 (F); Carlos, 21 Feb. 1966 (fr), A. Molina R. et al. 17638 (F). Ca ago: А fo -— ventazón behind main building di CA- TIE . 30 July 1985 (fl), M. Grayum & B. Ham- Guanacaste: N of Río Las Flores, ca. | km E of Río Tenorio, Hacienda Montezuma, 24 Jan. 1985 „ МО); vicinity of . 1926 (fr), P. Sande & J. Valerio 44443 (US). Heredia: finca La Selv va, зан OTS Field Sta- tion on the Río Puerto "t Mi E of its junction with the Río Sarapiquí, along W River R a 1 Pen 1978 (fl), M. Grayum 1353 (DUKE), 7 in 1980 (fl), B. Hammel 8480 (DUKE). rn near Bratsi ca. 30 mi. SE of Limón along Río Sixaola, . 1977 (fr), Т Croat 43275 (LL, Cantón de a ‘a, Reserva Indígena Talamanca, vie > Talamanca, Amubri, Soki, 17 June 1994 (fr), G. Ga- llardo 256 (FTG, MO). Puntarenas: along Quebrada Bo- nita, Carara Reserva, 25 July 1985 (fr), M. Grayum et al. 5718 (MO); along N Fork (known loc pd as ue brada Mona"), of Quebrada Bonita, Carara Res 11 June 1986 (fl), M. Grayum et al. 7602 (LL, м i2). ) PANAMA. Colón: Ac hiote, 2 a bag 73 (fl), R. Dressler 4403 (MO). Darién: 38.6 mi. E of Bayano Dam Bridge, near Río Топ, 17 “May 1980 (fr), T Antonio 4640 (LL, MO); Cana, near Río Setigandi, 18 Apr. 1980 (fl), A Gentry et al. 28551 (MO); vicinity of Paya, Río р trail between Paya & Payita, 10 June 1959 (fl), W. Stern et al. 395 (MO). Panamá: Canal Zone, Barro уй Island, with Armour, 1 Apr. 1970 (fl), 9257 ae Armour 6, 24 May 1969 (fl, fr), (DUKE); s Station, 6 = 1862 (fl, f Ds . Hayes s.n. (BM). VENEZUELA. 7 CAEN de Pavia, along Rio Yasa, near "Guasáma," md ove Kasmera, Biological station of the Universidad del Zulia. he SW of Machiques, 26-27 Aug. 1967 (fr) J. Steyermark & J. Fernández 99746 (NY). COLOMBIA. Ma 2 ا‎ Alto Río Buritaca, Alto de Mira, on the trail to Quebrada Julepia, : 3 July 1989 (fl), 5. Madrifián & C. Barbosa 192 (С O). Ri- saralda: Mpio. Pereira, Hacienda ы 4 km from Cerritos Fin Cerritos-Pereira Hwy., 27 Nov. 1989 (fr), Р. Silverstone-Sopkin et al. 5709 (МО). ECUADOR. Маро Cantón Archidona, Carretera Hollin—Loreto, entre Avila у Loreto, Huiruno, comunidad Quichua, 24 Nov. 1989 (fr). C. Cerón 7699 (MO); is "VO Ms 'afuerte, a 1-2 km al N de la población en line . aguas arriba del Río Napo, caserio Quichua, 26 Feb. i T (fr), J. Tals & F. Coello 4254 (QCA), 4274 (ОСА). PERU. Madre de Dios: Prov. Manu, Río Alto Madre de МЈ c. = = уе ~ Dios, forest near chacra of Sr. Carpio, halfway between Shintuya & Manu, 10-11 Aug. 1974 (fr), R. Foster et al. 3253 (MO). Ue yali: near Peru-Brazil border, d Sapallal, B tary of Quebrada Shesha, base of € аз Cachoeiras, 19 June 1987 (fr), A. Gentry & C. Diaz 56457 (MO). 6. Ardisia pellucida subsp. thomascroatii Ric- ketson & Pipoly, subsp. nov. Costa Rica. San José: about 1 mi. beyond divide be- tween San Isidro del General and coastal town of Domincal, 900 m, 22 May 19706 (fl), T. Croat 35270 (holotype, MO!). Figure 9. Propter ramulos, rhachides aor оо рейс 'elos dense minuteque papillosos, atque calycinis incons- picue punctatis et lineato-punc als, a pei pectinata si- milans, sed ab ea Yes calycinis 2.5-2.6 (поп 2.1-2.2) mm po 0.9-1.0 (nec 1.4-1. 5) 1 mm latis, lobulis coro- linis 3.6-3.8 (nec 50-5. 2) mm longis, anteris ovoideis ad apices acutis (non lanceolatis ad apices caudatis) per- facile recognoscitur. Shrubs | m tall. Branchlets 4.5-5 mm diam., scattered to densely minutely rufous glandular pa- i x 5- cm long, the sec- pillose. Leaves with blades elliptic, 13.5-28.5 11 cm, with an acumen 1.5-1.7 ondary veins 20 to 23 pairs; petioles 1.2-1.7 long. Inflorescences tripinnately paniculate, 12—12.5 4—14.5 cm, shorter than the leaves, vestiture of em the rachis and branches similar to the branchlets, the branches terminating in 5- to 7-flowered cor- 7 cm long: inflorescence branch bracts linear, 3.5—7.3 X 0.4—0.6 mm, conspicuously punctate and punctate-lineate, below ymbs; peduncles 1.6-1.7 in- sparsely and minutely papillose, the veins obsolete, the margins entire, minutely erose, hyaline, sparse- х 0.2- 0.4 mm, inconspicuously punctate and punctate- lineate; pedicels 5-7.3 mm long, inconspicuously sparsely and mi- ly glandular ciliolate; floral bracts 0.6—3.2 punctate and punctate-lineate, ;»-merous, lavender; calyx X 0.9-1 mm, apically nutely papillose. Flowers 5 lobes lanceolate, 2.5-2.6 long acuminate, without a notch below the apex, inconspicuously punctate and punctate-lineate; co- rolla 4.8—5 mm long, the tube 1—1.4 mm long, the lobes 3.6-3.8 2.4—2.6 mm, mostly epunctate, densely yellow glandular-granulose adaxially, api- cally above the staminal tube, as well as between the corolla tube and the lobe junction, glabrous out- side; stamens 3.2-3.3 mm long, the filaments 2.4— 2.5 mm long, the staminal tube 0.7-0.8 mm long, the apically free portions 1.7-1.8 mm long, epunc- tate, densely yellow glandular-granulose, the an- thers ovoid, 0.9-1 х 0.6-0.7 pistil 3.8-3.9 mm long, the ovary 0.9-1 mm long, conspicuously punctate and punctate-lineate, the style 2.7-2.9 mm long, epunctate, the ovules 26 to 29. Fruits mm; ра unknown. Distribution. Ardisia pellucida subsp. thomas- Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia croatii is known only from the type collection and is endemic to San José, Costa Rica, growing at 900 m in elevation Ecology and conservation status. Ardisia pellu- cida subsp. thomascroatii occurs in tropical moist forest. Unfortunately, no further ecological data are available from the label. Given the accessibility of the site, and lack of collections, this subspecies should be considered threatened. Etymology. It is an honor to dedicate this taxon to Thomas B. Croat of the Missouri Botanical Gar- den, scholar, gentleman, and preeminent authority on the systematics and ecology of Neotropical Ara- ceae and the genus Anthurium in particular. Within Ardisia subg. Auriculardisia sect. Amat- lania, Ardisia pellucida subsp. thomascroati shares a similar vestiture with subspecies pectinata, sparsely to densely, minutely, rufous glandular pa- pillose throughout. However, subspecies thomas- croatii may be distinguished by its elliptic leaves, longer but thinner lanceolate calyx lobes to 2.6 X ] mm, smaller corolla lobes to 3.8 X 2.6 mm that are glabrous outside and densely glandular-granu- lose inside around the corolla tube and filaments, the smaller ovoid anthers to | X 0.7 mm, the small- er pistil to 3.9 mm long, and the shorter style to 2.9 mm long. 1. Ardisia schippii Standl., Field Mus. Nat. Hist., Bot. Ser. 12: 412. 1936. Amatlania schippii (Standl.) Lundell, Wrightia 7: 40. 1982. TYPE: Belize. Toledo: Temash River, 200 ft. [61 m]. 8 Aug. 1935 (fl, fr), W. Schipp 1365 (holotype, F!, F neg. 68249!, LL neg. 1971-99!; isotypes, A!, BM!, G!, GH!, K!, MICH!, MO!, NY!). Fig- ure 10. Ardisia izabalana Lundell, Wrightia 5: 88. 1975. Syn. nov. тайата izabalana (Lundell) мк Phytologia 413. 1983. TYPE: Guatemala. Izabal: El Estor, 6 km S, 30 Jan. 1975 (fr), C. I 18898 (holotype, LL!, F ne F neg. 55617!, LL!). peris & E. Contreras g. 55616; isotypes, LL!, 5-26 cm scaltered to Shrubs or small trees 4—10.7 m tall, Branchlets 3—4.5 mm diam., densely minutely rufous glandular papillose, often diam. glabrescent with age. Leaves with blades coria- ceous, elliptic, 7.8-25.9 X 2.6-9.1 cm, apically acute, with an acumen to 1 cm long, basally acute to obtuse, often oblique, decurrent on petiole, in- conspicuously punctate and punctate-lineate, gla- brous above, sparsely and minutely papillate along the midrib, often on the blade, usually glabrous, the midrib impressed above, conspicuously raised below, the secondary veins 11 to 22 pairs, promi- nulous above and below, the margins entire to reg- ularly dentate or serrate, the teeth 3 to 5 per cm when present, flat; petioles slender, marginate, 1.2— 2.8 cm long, glabrous above, papillose below. /n- florescences erect, tripinnately paniculate, 7.5—24 X 9—23 cm, pyramidal, mostly longer than the leaf blades, peduncle, branches, and pedicels minutely papillose, the branches terminating in 5- to 10- flowered corymbs; peduncles 1.1-3.3 cm long, the lower branches often subtended by leaves; inflores- cence bracts unknown; inflorescence branch bracts membranous, oblong, 2.1-2.4 X 0.9-1.1 mm, api- cally acute, basally sessile, scattered, inconspicu- ously punctate and punctate-lineate, the veins ob- scure, the margins minutely erose, hyaline, sparsely glandular ciliolate; floral bracts membranous, ovate to deltate, 1. scattered inconspicuously punctate and punctate- X 0.9-1.1 mm, apically acute, lineate, the veins obscure, the margins minutely erose, hyaline, sparsely glandular ciliolate: pedicels slender, terete, 3.8—6.3 mm long, inconspicuously punctate and punctate-lineate. Flowers 5-merous, lilac to al calyx lobes chartaceous, ovate, 1.7— 8 х 1.2-1.3 mm, apically acute to acuminate, few seiten punctate and punctate-lineate, gla- brous adaxially, sparsely and minutely papillose, the margins erose, hyaline, sparsely glandular cil- iolate; corolla membranous, 6.6—6.8 mm long, the tube 1-1.2 mm long, the lobes ovate to lanceolate, 5.6-5.8 X 2.6-2.9 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, but densely yellow glandular-granulose adaxially, apically above staminal tube, as well as between the corolla tube and the lobe junction, the margins entire, hyaline; stamens 4.4—4.6 mm long, the fil- 2.5-2.6 mm long, the staminal tube 0.8— 0.9 mm long, the apically free portions 1.6-1.8 mm, aments epunctate, densely yellow glandular-granulose, the anthers lanceoloid, 2.5-2.6 X 0.7-0.8 mm, apically long subulate-apiculate, basally cordate, the con- nective inconspicuously punctate; pistil 5.9-6 mm long, the ovary oblong, 1.3-1.4 mm long, promi- nently punctate and punctate-lineate, the styles 4.6—4.7 mm long, epunctate, the ovules 28 to 32. mm diam., punctate and punctate-lineate, conspicuously cos- Fruits globose, 3. prominently tate. Distribution. Ardisia rpm is found in Cayo and Toledo, Petén, Guatemala, growing dom 61 to erapaz, Izabal, and 1900 m in Belize, and A elevation. Ecology and conservation status. Ardisia schip- pii occurs in primary rain forests. While it is cer- tainly not common, at this time there are no data to suggest that the species is threatened. 202 Annals of the Missouri Botanical Garden Volume 90, Number 2 2003 Ricketson & Pipol 203 y Revision of Ardisia subg. Auriculardisia Within Ardisia subg. Auriculardisia sect. Amat- lania, Ardisia schippii appears to be most closely related to A. liebmannii (see under that species for similarities). However, A. schippii may be distin- guished from A. liebmannii by its larger corolla lobes to 5.8 X 2.9 mm, longer lanceoloid anthers to 2.6 mm long with subulate-apiculate apices, and the longer styles to 4.7 mm long. The populations corresponding to the type of Ar- disia izabalana are unique only for their narrow leaf blades with more entire margins. The vestiture is sparser than in A. schippii, but otherwise they match perfectly. The type of Ardisia izabalana is in late fruit. Specimens examined. BELIZE. Cayo: vicinity of Cu vas S of Millionario, 29-30 May 1973 (fl), T Croat 23587 (F. FTG, LL, MO); Vale E И dell 6204 (F, LL, NY). (0), С. banal 6277 (F. GH, LL. NY, Ewa GUATEMALA a Verapaz: An apris betw. Pozo #4 and lagune s Lagartos, 29 July (fr), A Lundell & E. мич 19576 (LL [2]. MO). bal: Puerto Mendez, Cadenas, on new bids | Road, jon 2.5 km from village, 17 Aug. 1969 (fr), E. Contreras 8964 (LL [2], TEX); Cadenas, 6 km SW, 9 July 1970 (fl), E. Contreras 10154 (DUKE, LL [3]. NY), (fr). 10164 (F, LL 3]. NY); Puerto Mendez, on Río Dulce km, 6 Sep. 1970 (fr), E. Contreras 10217 (DUKE, LL [3]. NY); ELE fr), E. Contreras 11491 (LL en Santo Toribio and fl, fr), E Contreras 2675 (DUKE, LL б TEX); La Cumbre, W of km 142/143, 500 m the n 10 Sep. 1975 (fr), C. Lundell & E. Con- treras 19838 (LL [2]. MO). TAXONOMIC. TREATMENT OF ARDISIA SUBG. AURICULARDISIA SECT. AURICULARDISIA Ardisia subg. Auriculardisia sect. Auriculardi- sia. Subshrubs or small trees to 20 m tall, 10 cm diam. Branchlets stout, terete, densely furfura- ceous-lepidote, cupuliform-lepidote (the scales with varying number of lobes or arms, at times of two sizes), mixed furfuraceous- and cupuliform-lepi- dote, or stipitate-stellate tomentose, the terminal star often breaking off, leaving an apparently vil- lous trichome, the indument normally persistent. Leaves monomorphic; blades membranaceous to co- riaceous, elliptic, obovate to oblanceolate, promi- nently, inconspicuously (pellucid) or conspicuously punctate and punctate-lineate, the margins entire, flat; petioles at times obsolete, or slender or stout and much longer, then canaliculate or marginate. Inflorescences terminal, erect, pinnately to tripin- nately paniculate, shorter than the leaves, the flowers on the secondary or tertiary branches glomerate, corymbose or of congested cor- ymbs, rarely pseudoracemose; inflorescence bracts, inflorescence branch bracts, and floral bracts mem- branaceous to coriaceous, usually pu floral bracts white, light pink to red, 3.7-9.7 X 1-8 mm; pedicels stout, terete, often ud and ac- pyramidal, crescent or incrassate with maturity. Flowers 5- or 6-merous, green to white or pink to purple in hues; calyx lobes membranaceous to coriaceous, the mar- gins entire, hyaline; corolla membranaceous to co- riaceous, the lobes variously connate, ovate to lanceolate, the margins entire, hyaline; stamens with obvious filaments, connate basally into an elo- bate inconspicuously hyaline tube, the anthers free, narrowly ovoid to linear-lanceoloid, lanceoloid, to ovoid, apically apiculate, basally cordate to sub- cordate, dehiscent by subapical pores, opening into wide, longitudinal slits, the connective inconspic- uously to conspicuously punctate dorsally; pistil glabrous, the ovary oblong, the style slender, erect, inconspicuously or conspicuously, or prominently punctate, the ovules few to numerous, in 2 to nu- merous verticels. Fruits globose to depressed-glo- bose, often conspicuously to prominently punctate and punctate-lineate. Distribution. Four species, from Puntarenas and San José in Costa Rica, southward to Coclé, Panamá, and San Blas in Panama, to Antioquia and Vaupés in Colombia. They grow in habitats from 5 to 1400 m in elevation. Ecology and conservation status. Members of section Auriculardisia occur in primary premonta- ne, cloud and elfin forests and pluvial forests. АП members are considered threatened. € Figure 10 (left). rescence. —D. Flower. Stamen, Ardisia schippii. —A. Flowering branch. —B. ака of abaxial leaf surface. —C. gen of inflo- abaxial surface. —F. Stamen, ral margin. —G. Stamen, adax ace. — Н. Fruit. (A-C drawn from буйур, W. Schipp 1365 (F); D-G from C. ет 6277 (СН); Н бот E. ae 10217 (LL).) Figure 11 (right). rescence, inflorescence bran F. Stamen, adaxial surface. — Ardisia. dodgei. ch bract, and floral bracts - —A. ا‎ branch. —B. Detail of abaxial leaf surface. —C. Detail of "eid —D. pu amen, lateral margin. — 9881 (MICH); C-G from F бызаа 530 (F); H from R. PCS 354 (M jr floral bract. —E. Stamen, abaxial surfac t. (A, : drawn from isotype, C. Dodge & V. bs 0).) 204 Annals of the Missouri Botanical Garden Ardisia subg. Auriculardisia sect. Auriculardisia is defined by the combination of its terminal inflo- rescence with branches terminating in congested, glomerate corymbs, each of which is subtended by a persistent, enlarged inflorescence branch bract, and finally, the individual flowers subtended by a persistent, enlarged floral bract, as long as the flow- ers. KEY TO THE bene OF ARDISIA SUBG. AURICULARDISIA SECT. AURICULARDISI/ la. Plants with a mixture of minute, sessile trans- lucent scales, the scales often early caducous, and tomentose or villous to hirsute rufous s stipi- tate-stella te tric -homes, the stellate branches on then appearing glandular-villous; calyx lobes narrowly lanceolate, 7.2-8.2 mm long, apically attenuate 11. Ardisia ursina Plants wih a mixture of primarily furfuraceou lepidote scales and cupuliform scales кет stipitate-stellate trichomes, when present. the stipitate-stellate trichomes lowe than 0.4 mm and barely discernable from the scales; calyx lobes ovate, oblate or orbicular suborbicular, 1.8-5.7 mm long, apically cael to truncate, often slightly emarginate 2a. Floral bracts red or pink; calyx The cori- aceous, 5.4—5.7 X 4.8-5.2 mm; corolla lobes 8. 3-8.5 x 3.2- a 5 mm; anthers linear- lanceoloid, 4.5-5.7 X 0.9-1.1 i a basally subcordate Ardisia dodi 2b. Floral bracts white to light din not red; calyx lobes membranaceous to chartaceous, P x -— — - inc onspic uous, 1.8-2. .9—4.7 mm; corolla lobes 4.5— 5.7 X 1-2.7 7 mm Jong: anthers ovoid to lanc iod 2.64.1 1.2-1.5 mm, basally cordate to = Ja. Petio ae ulate, 34.5 ст long; leaf blade s flat above, the sec bed veins caning above and below; perianth blue-gray or lavender; calyx lobes chartaceous, oblate, 2.3-2.7 4.3—4.7 mm; corol 4 mm cd the lobes 5.3—5.7 mm long; anthers 3.£ mm long, the filaments d а! base . Ardisia glomerata 3b Petioles pipi to ne a. auriculate sile to 0.3 cm long; leaf blades bullate Ku. the dary veins deeply impressed above, prominent be- low; perianth light green to pink; calyx lobes pa ev) bicular, 1.8-2.2 X 1.9-2.1 mm; corolla 6.8-7 mm long, the lobes 4.5—4.8 mm long; anthers 2.6-2.8 mm jong, the fil- aments not basally Үн us ed А ). Ardisia nervosissima seconc 8. Ardisia dodgei Standl., Publ. Field Mus. Nat. Hist., Bot. Ser. 18: 888. 1938. Auriculardisia dodgei (Standl.) ШЕЙ, Phytologia 49: 343. 1981. TYPE: Costa Rica. San José: low hills above Río Paquita, 5-50 m, 15 Aug. 1936 (fl), C. Dodge & V. Goerger 9881 (holotype, F!, F neg. 68147!, LL neg. 1971-45!; isotypes, CR!, GH!, MICH!, LL neg. 1971-15!, MO!, LL neg. 1971-46!, NY!, PH!, UC!, US!, US neg. 23711). Figure 11. Shrubs or small trees to 20 m tall, 10 em diam. Branchlets 3—7 a mixture of furfuraceous-lepidote and cupuliform mm diam., densely tomentose with scales and with scattered stellate and stipitate-stel- late trichomes. Leaves with blades chartaceous to coriaceous, elliptic, 12.7-25.2 X 5.2-11.3 cm, api- cally acute to acuminate, with an acumen 0.8-2.4 em long, basally acute to obtuse, decurrent on the petiole, mostly inconspicuously (pellucid) punctate and punctate-lineate, glabrate above, furfuraceous- lepidote below. occasionally with cupuliform scales along the midrib, the midrib impressed above, prominently raised below, the secondary veins 22 to 64 pairs, prominulous above and below; petiole stout, marginate, 4—12 mm long, mostly glabrate above, below with a mixture of dense furfuraceous- lepidote and cupuliform scales and with sparse stellate and stipitate-stellate trichomes. /nflores- cences bipinnately or tripinnately paniculate, 2-7.5 2.5-7 cm, the rachises densely tomentose with a mixture of furfuraceous-lepidote and cupuliform scales with scattered stellate and stipitate-stellate trichomes, the branches terminating in 5- to 10- flowered glomerate corymbs; peduncle 0.4—1.4 em long: inflorescence bract caducous, membranous, widely ovate to oblong, 1.1-2.6 X 0.8-1.6 em, api- cally rounded to truncate, prominently punctate and punctate-lineate, glabrate above, furfuraceous- lepidote below, the midrib inconspicuous, the mar- gins entire, flat; inflorescence branch bracts similar to the inflorescence bracts, but persistent, ovate, 0.8-2.1 X inflorescence bracts, but red to light pink, usually 8-9.7 X flower; pedicels 3.6—4.2 mm long, inconspicuously 0.6-1.6 em; floral bracts similar to the persistent, 5.2-8.5 mm, enveloping the punctate-lineate, with scattered. furfuraceous-lepi- dote scales. Flowers 5-merous, white or light pink; calyx lobes coriaceous, very widely ovate to sub- orbicular, 5.4—5.7 X 4.8—5.2 mm, apically rounded or truncate, conspicuously and somewhat promi- nently punctate and punctate-lineate, glabrous ada- xially, scattered furfuraceous-lepidote abaxially; co- rolla chartaceous, 10.3-10.5 mm long, the tube 2— 2.2 mm long, the lobes ovate, 8.3-8.5 x 3.2-3.5 mm, apically obtuse to rounded, conspicuously punctate and punctate-lineate, glabrous adaxially, sparsely furfuraceous-lepidote abaxially; stamens Volume 90, Number 2 2003 Ricketson & Pipoly 205 Revision of Ardisia subg. Auriculardisia 9.6-9.8 mm long, the filaments 5.2—5.4 mm long. the staminal tube 2.5-2.6 mm long, the apically free portions 2.7-2.8 mm long, pellucid punctate, 0.9-1.1 mm, basally subcordate, the connective conspicu- the anthers linear-lanceoloid, 4.5—5.7 X ously punctate; pistil 7-8.4 mm long, the ovary 0.9-1 mm long, the style 6.1—7.4 mm long, con- spicuously punctate and punctate-lineate, the ovules 36 to 39. Fruits depressed-globose, 6.8—7.2 mm diam., conspicuously punctate. Distribution. Ardisia dodgei was considered en- demic to the area of the Osa Peninsula in San José and Puntarenas, Costa Rica, until Pipoly (19912) reported. disjunct populations from Antioquia and Vaupés, Colombia, growing from 5 to 700 m in el- evation Ecology and conservation status. Ardisia dodget occurs in primary wet or pluvial forests. Because of its restricted distribution, it should be considered threatened. Etymology. Ardisia dodgei was named in honor of the late Carroll William Dodge (1895—1988), li- chenologist, mycologist, and staff member at the Missouri Botanical Garden from 1931 to 1963. Within Ardisia subg. Auriculardisia sect. Auri- culardisia, Ardisia dodgei is distinguished by its red or pink persistent floral bracts to 9.7 mm long and to 8.5 mm wide, the large corolla to 10.5 mm long with lobes to 8.5 mm long, and linear-lanceo- loid anthers to 5.7 mm long. Spec cimens examined. COSTA RICA. eserva Forestal Golfo D Ресен Aguilar 469 (CR. M Chal, La Parcela, 24 Sep. 1996 (ster.). de MO); 25 km W de Chacarita by road, be nd Chacarita, 29 June 1991 (fl), B. Band & M. Nepokroeff 18251 (CR, FTG, INB, MO); Cortés, Los Mo- gos, headwaters of Quebrada Taboga, 24 Nov. 1991 (fl), A Herrera 4977 (CR, FTG, INB, MO); Península de Osa, Rancho Quemado, SE Sector, 16 Sep. 1992 (fr), J. Marín & D. Marín 520 (CR, FTG, INB d Cantón de Golfito, Parque Nac ie rw ovado, Península de Osa, Corcova- , E. Alfaro 249 (CR, INB, MO); Es- July 1951 (fl), P. Allen 6258 (CAS, F [2]. € Puerto е Osa, 40 km W « Gómez P. 19496 (LL [2]. MO): Canton de Osa, Sierpe, Los Mogos, por nage of Río Agr affluent of Quebrada Taboga, 15 Dec. 1990 (fl), G. Herrera 4791 (CR. FTG, INB, MO); Parque Nac ‘ional С orcovado, Upper Ollas Trail, 21 June 1988 (fl), C. Kernan & P. Phillips 629 (CR, MO); Cantón de Osa, Hanc sho е o, 5 July 1991 (fl), F Quesada 530 (CR, F, FTG, INB, MO); Cantón de Golfito, Refugio Nacional de Fauna ais estre Соо, 28 Feb. 1994 (fr), G. Rivera et al. 2237 (CR, К); betw. Golfo Dulce and Río Térraba, Dec. 1947 (fl), A. Skutch 5310 (F, MICH [2], US); Cantón de Osa, Rincón de Osa, Entrance to Chocua- » 27 Aug. 1992 (fr), N. Zamora et al. 1861 (CR, INB, MO). COLOMBIA. Antioquia: highway to the sea, near Villa Arteaga, 6 Dec. 1948 (fl), Е López & M. Sánchez M. 28 (COL, US) ا‎ Río i» cni and Cerro de Isi- bukuri, district alon untain summit, 30 Nov. 1951 (fl), H. García- ê 137664 (COI c . Ardisia glomerata Lundell, Amer. Midl. Nat- uralist 29: 486. 1943. Auriculardisia glome- rata (Lundell) Lundell, Phytologia 49: 344. 1981. TYPE: Panama. Coclé: hills N of El Va- lle de Antón, trail to La Mesa, ca. 1000 m, 2 Sep. 1941 (fl), P. Allen 2741 (holotype, MO!, LL neg. 1971-51!; isotypes, A!, US!). 12. ‘igure Small trees to 8.5 m tall. Branchlets 6.5-11 mm diam., densely tomentose with a mixture of furfu- raceous-lepidote and cupulate scales and stipitate- stellate tomentellous on multicellular stalks to mm long, appearing glandular-villous as in A. ur- sina with the stellate portion broken off. Leaves with blades membranous to chartaceous, elliptic, 31.5— 39.5 X with an acumen 0.7-2.3 cm long, basally acute or acuminate, decurrent on the petiole, mostly incon- 9.6—16.2 cm, apically acute to acuminate, spicuously punctate, but with a few punctate-lin- eations above and below, glabrate above, below with a mixture of dense furfuraceous-lepidote and scattered cupulate scales, except densely so along midrib, the midrib impressed above, prominent be- low, the secondary veins 25 to 75 pairs, prominu- lous above and below; petioles stout, canaliculate, 3—4.5 cm long, glabrous above, below with a mix- ture of dense furfuraceous-lepidote and scattered cupulate scales. /nflorescences bipinnately or tripin- х 6-16 cm, the rachis densely tomentose as in the branchlets, the branch- nately paniculate, 12-24 es terminating in 5- to 9-flowered glomerate cor- mbs; peduncle 1.7-3.8 cm long; inflorescence bract unknown; inflorescence branch bracts often persistent, membranous, oblong, 0.9-1.2 X 0.4—0.7 cm, apically acute, inconspicuously (pellucid) punctate and punctate-lineate, glabrate above, fur- furaceous-lepidote and stellate-tomentose below, the midrib inconspicuous, the secondary veins in- conspicuous, the margins entire, flat, glandular-cil- iolate; floral bracts similar to the inflorescence branch bracts, white to light pink, but ovate to ob- long, 4.5-7.2 X 3.84.5 mm; pedicels 3.4-6.7 mm long, inconspicuously punctate-lineate, furfura- ceous-lepidote and stellate-tomentose. Flowers 5- or 6-merous, blue-gray or lavender; calyx lobes chartaceous, oblate, 2.3-2.7 X 4.3—4.7 mm, api- cally truncate and slightly emarginate, conspicu- ously punctate and punctate-lineate, glabrous ada- Annals of the 206 Missouri Botanical Garden Volume 90, Number 2 2003 Ricketson & Pipoly 207 Revision of Ardisia subg. Auriculardisia xially, furfuraceous-lepidote abaxially; corolla coriaceous, 8-8.4 mm long, the tube 2.7-2.8 mm long, the lobes ovate, 5.3-5.7 X 2.4-2.7 mm, api- cally obtuse to rounded, cucullate, conspicuously punctate and. punctate-lineate, glabrous adaxially. sparsely furfuraceous-lepidote abaxially; stamens 6-6.9 mm long, the filaments 3-3.2 mm long, ba- sally widened above the junction with the tube, the staminal tube 1.4—1.6 mm long, the apically free portions » 4—1.6 mm long, the anthers lanceoloid, x 1.3-1.5 mm, basally cordate, the con- nective conspicuously punctate dorsally; pistil 5.4— 6.2 mm long, the ovary 0.9-1.1 mm long, the style 4.5-5.1 mm long, conspicuously punctate апа punctate-lineate, the ovules 19 to 21. Fruits de- pressed-globose, 6.2-8.2 mm diam., densely and conspicuously punctate. Distribution. Ardisia glomerata is endemic to Cerro Pilón and adjacent areas near El Valle in Coclé, Panama, from 600 to 1000 m in elevation. Ecology and conservation status. Ardisia glo- merata is found in cloud forests, a life zone that is rapidly disappearing in Panama. Because of its re- stricted distribution, it should be considered threat- ened. Etymology. The specific epithet was derived from the Latin glomer, meaning to form or collect closely together into a sphere, and refers to the con- gested flowers of the inflorescence. Within Ardisia subg. Auriculardisia sect. Auri- culardisia, Ardisia glomerata is most closely related to A. nervosissima because of their small oblate or orbicular to suborbicular calyx lobes, which are apically rounded to truncate, and chartaceous to coriaceous corollas. However, A. glomerata can easily be separated from A. nervosissima because of its canaliculate petioles to 4.5 cm long, flat leaf blades, chartaceous, oblate calyx lobes to 2.7 mm long and to 4.7 mm wide, and coriaceous corolla with lobes to 5.7 mm long and to 2.7 mm wide and sparsely furfuraceous-lepidote abaxially, and anther to 4.1 mm long ‘imens examined. PANAMA. Coclé: betw. Cerro nin gee El Valle, 15 Aug. 1967 (8), J. Duke & Ј. Dwyer 13964 (LL, MO); foothills of Cerro Pilón near E] Valle, 5 Oct. 1967 (в), Ј. Duke & M. Correa А. 14692 (LL); Cerro Pilón, El Valle, 4 Jan. 1968 (fr), J. Duke & B. Lallathin 14989 (FTG, LL, MO); El Valle de Antón at the foot af Cerro Pilón, 15 Aug. 1967 (fl), J. Dwyer & M. горі А. 7938 (ЕТС. LL, МО); La Mesa, Cerro Pilón area, 19 1968 (fl), J. Dae 8321 (MO); woods adjacent to c diss farm, La Mesa above El Valle, 3 Jan. 1974 (fr), J. Dwyer 11866 (LI МО). T 10. Ardisia nervosissima Lundell, Wrightia 4: 62. 1968. Auriculardisia nervosissima (Lun- dell) Lundell, Phytologia 49: 345. 1981. TYPE: Panama. Coclé: El Valle, 800—1000 m, 28 June 1967 (fl), J. Duke 13150 (holotype, LL!, F neg. 55645!; isotypes, GH!, LL!). Figure 3. Shrubs or small trees to 10 m tall. Branchlets 3.5— 8.5 mm diam., densely furfuraceous-lepidote, the chartaceous, x 3.2- 9.6 cm, apically acuminate, with an acumen 0.3— scales persistent. Leaves with blades elliptic, obovate or oblanceolate, 9.2-26.5 X 1.7 em long, basally acute to nearly auriculate, de- current on the petiole, prominently punctate and punctate-lineate, sparsely furfuraceous-lepidote above, more densely so below along midrib and to- ward base, strongly bullate above, the midrib im- pressed above, prominent below, the secondary veins 22 to 35 pairs, deeply impressed above, prominent below; petioles stout, marginate to nearly auriculate, subsessile to 3 mm long, densely fur- furaceous-lepidote. Inflorescences pinnately to bi- pinnately paniculate, 6.2-12.5 X 4-8.5 cm, to- mentose with rufous translucent cupuliform scales, at times with arms, the scales 0.1—0.6 mm long, and with flat furfuraceous-lepidote scales below, the branches terminating in 3- to 7(or 9)-flowered glom- erate corymbs; peduncle 0.5-1.4 cm long; inflores- c— Figure 12 (left). Ardisia glomerata. —A. sessile, flat scale (right). — showing floral bract. —F. Sta Fruit. (A-C 11866 (MO).) men, abaxial surface. —G. Figure 13 (right). vestiture types showing sessile cupulate with age and drying, id ped flat (right) scales bract and floral bracts. —E. Flow Ardisia e dits . —A. Flo Flowering branc types, sessile stellate (left). "D -stellate on a multicellular stalk zr the ray ail of inflorescence, showing inflorescence branch peus and floral bracts. —E. Flow Stamen, C & E-H drawn from holotype, P. Allen 2741 (MO); D from isotype, P. Allen 2741 (US); I hm pi Dwyer owering bran eft) or multic 'ellular stipitate (center) scales, the scales terminally rotate, ш эйр er and floral bract. —F. бша tail of abaxial leaf surface. —C. Trichome s terminally rotate (center), "wl a h. —B. adaxial surface. —H. Stamen, lateral margin. en ch. —B. Detail of abaxial leaf surface. —C. Detail of ail of inflorescence, showing inflorescence branch abaxial surface. —G. Stamen, adaxial surface. H. Stamen, lateral margin. —1. Fruit. (А-1 drawn from 5. Pow » al. 6058 (MO).) 208 Annals of the Missouri Botanical Garden cence bracts unknown; inflorescence branch bracts usually persistent, membranous to chartaceous, lanceolate to elliptic, 0.5—4.1 X 0.3-1.4 ст, acrop- etally smaller, apically acute to obtuse, densely and prominently punctate and punctate-lineate, tomen- tose as in the rachis, glabrescent, the midrib con- spicuously impressed above and prominent and conspicuous below, the secondary veins inconspic- uous, the margins entire, sparsely glandular-cilio- i the branch bracts, but white to light pink, widely ovate, 4.8-5.7 X 3.44.3 mm; pedicels 0.5-5.1 mm long, tomentose like the inflorescence rachis. Flowers 5- late; floral bracts similar to inflorescence merous, light green to pink; calyx lobes membra- 9-2.1] mm, apically broadly rounded to truncate, promi- nous, orbicular to suborbicular, 1.8-2.2 x nently punctate and punctate-lineate, glabrous ada- xially, sparsely furfuraceous-lepidote abaxially; co- rolla coriaceous, 6.8-7 mm long, the tube 2.2-2.: X 2.1-2.3 mm, apically flat, rounded to obtuse, prominently punc- mm long, the lobes ovate, 4.5—4.8 tate and punctate-lineate, glabrous throughout; sta- mens 5.8-6 mm long, the filaments 3.7-3.9 mm long, not widened above the tube junction, the sta- minal tube 1.3-1.5 mm long, the apically free por- tions 2.3-2.5 mm long, 0.4—0.6 mm КЕ epunc- tate, the anthers ovoid, 2.6-2.8 X —1.4 mm, basally sagittate, the connective inc din. punctate dorsally; pistil 5.3-5.8 mm long, the ovary 0.9-1.1 mm long, the style 4.4—4.7 mm long, prom- inently punctate, the ovules 19 to 21. Fruits glo- bose, 6.4—7.2 mm diam., densely and prominently punctate and punctate-lineate, glabrous. Distribution. Ardisia nervosissima is endemic to Panama, in Coclé (Cerro Caracoral, Cerro Pilon, La Mesa, and El Valle) and Panamá (Cerro Campana and Cerro Trinidad), growing from 800 to 1100 m in elevation. Ecology and conservation status. Ardisia ner- vosissima is found in elfin and cloud forests. Be- cause of its restricted distribution, it should be con- sidered threatened. Etymology. The specific epithet refers to the strongly nerved leaves. Within Ardisia subg. Auriculardisia sect. Auri- culardisia, Ardisia nervosissima is most closely re- lated to A. glomerata (see under that species for similarities). However, A. nervosissima can easily be separated from A. glomerata because of its subses- sile, marginate to auriculate petioles to 3 mm long, bullate leaf blades, membranous and orbicular to suborbicular calyx lobes to 2.2 mm long and to 2.1 mm wide, and coriaceous corolla with lobes to 4.8 mm long and to 2.3 mm wide and glabrous aba- xially, and anther to 2.8 mm long. PA pct ЕУ PANAMA. Coclé: Cerro Pilón, El Valle, 1. 1968 (fr), J. Dike & үн Lallathin 14968 (LL, MO), a (LL, MO); Cerro Caracoral, 19 Jan. 1968 (fr), J. Duke & J. Dwyer 15132 О); Cerro E 14 E d ). J. Dwyer & B. Гапа 6677 (ЕТС, МО); El Valle, between Cerro Carac vue and Cerro Gaital, 18 (жЕ 1982 (fl), S. Knapp et al. 6058 (LL, MO); Cerro Pilon, Sep. 1968 (fl, fr), B. L Lallathín 3. 2 (FTG [2], MO); divide SW of La Mesa at end of logging road, : Dec. 1982 (fl, p B. Stein & C. Hamilton 972 (LL, MO, NY); | 2 km W of Cerro Pilón. 22 2 July Ju (fl), G. Sullivan 519 (FTG, LL [3], MO). Panamá: Parque Nacio- Altos de C na, near summit of Cerro Campana, Montenegro 10638 ; Cerro Campana, forest о p | Su-Lin, 8 io 1966 (fr), J. Pip 6650 (MO); C ‚ NE side of Cerro шч 4 Feb. 1971 (fr), R. ыд 2097 (DUKE, erro Deben 23 July 1983 (fl), C. Hamilton et al. ДРЕ (LL, MO, NY); Cerro eem 31 Dec. 1985 (fl, fr), G. McPherson 7904 (LL, MO). —. — S5E = = 11. Ardisia ursina Lundell, Wrightia 6: 92. 1979. ndell) — Valerioanthus ursinus ( Lundell, Wrightia 7: 50. 1982. Auriculardisia ursina (Lundell) Lundell. Phytologia 57: 450. 1985. TYPE: Panama. Panamá: El Llano—Cartf Road, 10 km from Inter-American Hwy., 5 5 Oct. 1974 (fl), S. Mori & J. Kallunki 2314 (ho- lotype, MO!, LL neg. 1979-5!). Figure 14. Ardisia fibris Lundell, Phytologia 61: 64. 1986. Syn. nov. Valer и hirsutissima (Lundell) Lun- dell, Phytologia 63: 1987. TYPE: Panama. Co- clé: Continental Divide above El Copé, 08°38'N, Mir d W, 650—750 m, 27 Nov. 1985 (fr), G. de Ne- A. Henderson, H. Herrera, G. McPherson & L. Brako 6408 (holotype, LL!; isotypes, BM!, CAS!, CR!, MEXU!, MO!, NY!, РМА!). Figure 14 (left). F rescence. —D. Flower. —E. Stamen, abaxi: «жане from holotype, 5. Mori & J. Kallunki 2314 6408 (MO E of Ardisia hirsutissima Lundell).) Figure 15 (right). Ardisia apoda. —А. rescence ^ mature fruit. ( Ardisia ursina. —A. Flowering branch. В. De . —F. Ste H. Fruit. —I. M of кен -stellate trichome with long arms, uniseriate on multicellular stalks. (A, О); C-G from G. McPherson 14042 (M( Flowering branch. —B. A-C drawn from holotype, J. Steyermark 41923 — tail of abaxial leaf surface. —C. Detail of inflo- amen, lateral margin. —G. Stamen, adaxial surface. — B & I drawn )); H from G. de Nevers et al. Detail of abaxial leaf surface. —C. Detail of inflo- (F).) Volume 90, Number 2 Ricketson & Pipoly 209 2003 Revision of Ardisia subg. Auriculardisia 210 Annals of the Missouri Botanical Garden with a mixture of scattered minute, sessile translucent Trees to 6 m tall. Branchlets 3—7 mm diam., scales, the scales often early caducous, and densely tomentose or villous to hirsute rufous stipitate-stel- late trichomes throughout, the stellate branches on uniseriate, multicellular stalks to 1.5 mm long, the apical stellate arms often caducous, the hairs then blades membranous, elliptic to narrowly obovate, 6.3—23.7 appearing glandular-villous. Leaves with X 2.2-6.9 cm, apically acuminate, with an acumen 3-16 mm long, basally acute, decurrent on the pet- iole, prominently punctate and punctate-lineate. above and below with a mixture of sparse to scat- tered, minute, sessile translucent scales, the scales often early caducous, and with scattered villous or hirsute rufous stipitate-stellate trichomes, denser along midvein, otherwise as in the branchlets, the midrib impressed above, prominently raised below, the secondary veins 21 to 35 pairs, prominulous above and below; petioles slender, canaliculate, 2— 6 mm long, with a mixture of scattered minute, ses- sile translucent scales, the scales often early ca- ducous, and scattered to densely tomentose or vil- lous to hirsute rufous stipitate-stellate trichomes above and er Inflorescences bipinnately panic- ulate, 4.2-8.4 X 2.3-4.7 cm, scattered minute, with a mixture of sessile translucent scales, the scales often early caducous, and densely tomentose or villous to hirsute rufous stipitate-stellate tri- chomes, otherwise like the branchlets, the branches terminating in 3- to 9-flowered glomerate corymbs: peduncles 0.2-1.5 cm long; inflorescence bracts early caducous, unknown; inflorescence branch bracts persistent, membranous, ovate to elliptic, 64.2 X cm, apically and basally acute, conspicuously and prominently punctate and punc- tate-lineate, with a mixture of scattered minute, sessile translucent scales, the scales often early ca- ducous, and scattered villous or hirsute rufous stip- itate-stellate trichomes as in the inflorescence ra- chis, prominently raised below, the secondary veins 9 to the midrib impressed above, pairs, prominent above and below, the margins ане flo- ral bracts similar to the inflorescence branch bracts but white to light pink, 3.7-6.9 x 1-1.7 sile, the veins inconspicuous; pedicels 1.5-3.1 mm mm, ses- long, inconspicuously punctate and punctate-lin- eate, vestiture as in the inflorescence rachis. Flow- ers 5-merous, white with a pink tinge; calyx lobes x 2-2.3 mm, apically attenuate, prominently punctate and membranous, narrowly lanceolate, 7.2-8.2 punctate-lineate, glabrous adaxially, with a mixture of sparse minute, sessile translucent scales, the scales often early caducous, and scattered villous or hirsute stipitate-stellate trichomes abaxially; co- rolla membranous, 9-9.4 mm long, the tube 2.1— 2.7 mm long, the lobes ovate, 6.7-6.9 х 3.64.2 mm, apically ) acum prominently punctate and punctate-lineate, glabrous throughout; stamens 6.6-6.8 mm long, the filaments 3.9-4.1 mm long, the staminal tube 1.3-1.7 mm long, the apical free portion 2.4-2.6 X 1-1.4 mm, the an- thers ovoid, 3.8—4 X 1.5-1.7 mm, basally subcor- date, the connective inconspic uously punctate dor- .4—7.7 mm long, the ovary 1.7-1.9 mm long, the style 5.5—5.6 mm long, inconspicuously acute t inate, ~ sally; pistil 7 punctate, the ovules 24 to 29. Fruits globose, 5— 5.8 mm diam., prominently punctate and punctate- lineate. Distribution. Ardisia ursina is endemic to Pan- ama, Coclé, and San Blas, Panama, growing from 300 to 1400 m in elevation. Ecology and conservation status. Ardisia ursina occurs in primary, premontane rain forests. Because it is relatively uncommon, it should be considered threatened. Etymology. the bear,” The specific epithet, meaning “of refers to the rufous stipitate-stellate in- dument of the branchlets, leaves, and inflorescence that resembles the fur coat of a bear. Ardisia ursina is unique within Ardisia subg. Au- riculardisia sect. Auriculardisia because its branch- lets, leaves, and inflorescence branches have an in- dument with a mixture of minute, sessile translucent scales, the scales often early caducous, and tomentose or villous to hirsute rufous stipitate- stellate trichomes, the stellate branches on unise- riate, multicellular stalks to 1.5 mm long, the apical stellate arms often caducous, the hairs then ap- pearing glandular-villous. Although the branchlets and inflorescence branches of Ardisia dodgei, glomerata, and A. nervosissima also can have stip- itate-stellate trichomes, they are considerably smaller, to only 0.4 mm long, and inconspicuously mixed with both furfuraceous-lepidote and cupuli- form scales. Ardisia ursina is also the only member of the section with narrowly lanceolate calyx lobes to 8.2 mm long with an attenuate apex. The type of Ardisia hirsutissima was collected in fruit, and except for a slightly darker and thicker vestiture, it is identical to A. ursina. Specimens examined. PANAMA. Coclé: above El I troso sawmill at Continental Divide, 24 hrs 1980 n. К. Sytsma 1821 (LL, MO). Panamá: along El Llano-Cartí road, along creek E of ien 4 Жү 1989 if), С. McPherson 4042 (F, FTG, MEXU, PMA). San Blas: Cerro Obu, E dus 1986 (fl), G. P Neren et al. 8046 (LL, MO); gandi, Sendero Wedar, uly 1986 (fl), J. Mc- Donagh et al. 202 (MO); Cerro EL trail from Шы Sidro, 20 Dec. 1980 (fr), K. Sytsma et al. 2770 (LL, MO). Volume 90, Number 2 2003 Ricketson & Pipoly 211 Revision of Ardisia subg. Auriculardisia TAXONOMIC TREATMENT OF ARDISIA SUBG. AURICULARDISIA SECT. FAGERLINDIA Ardisia subg. Auriculardisia sect. Fagerlindia Ricketson & Pipoly, sect. nov. TYPE here des- ignated: Ardisia brenesii Standl. uoad sepala asymmetrica ad bases auriculata atque ramulos furfuraceo- lepidotos ad Ardisiam subg. Auricular- disiam pertinet. Ab aliis sectionibus subgeneris foliis di- morphicis, atque planta exemplar architecturalem Fager- lind exhibente praeclare distat. Small subshrubs or trees exhibiting Fagerlind’s Architectural Model (Hallé et al., 1978), to 7 3 cm diam. Trunk, vegetative shoots, and reproduc- tive shoots slender, terete, hirtellous-tomentose, or furfuraceous-lepidote and/or with сири огт scales. Leaves dimorphic; vegetative shoot leaves with blades membranous, often glabrous above, usually similar to the shoots below; petioles stout, obsolete to petiolate; reproductive shoot leaves with lades similar to the vegetative shoot leaf blades, but usually smaller. Inflorescences terminal or pseu- doterminal, pendent, pinnately to tripinnately pa- niculate, pyramidal, usually longer than the leaves, the branches terminating with flowers in loosely m tall, congested corymbs; inflorescence bracts usually persistent, foliaceous; inflorescence branch bracts and floral bracts caducous, the floral bracts much smaller than the flowers; pedicels slender, terete. Flowers 5-merous, white, light pink, light purple or red; calyx lobes essentially free, membranous to chartaceous, ovate to suborbicular, basally auricu- late; corolla membranous, the lobes ovate to lan- ceolate, conspicuously and often prominently punc- tate and punctate-lineate; stamens with the filaments apically free, connate basally into an elo- bate tube, free from the corolla tube, epunctate, the anthers ovoid or narrowly ovoid to lanceoloid or linear-lanceoloid, apically apiculate, cuspidate, su- bulate, mucronate or caudate, basally sagittate or cordate, dehiscent by subapical pores, opening into wide, longitudinal slits, the connective punctate; pistil glabrous, the ovary oblong, the style slender, erect, inconspicuously or conspicuously, rarely prominently punctate, the ovules pluriseriate. Fruits globose, inconspicuously or conspicuously, often prominently punctate and punctate-lineate, often costate. Distribution. Seven species, including one pop- ulation from Veracruz, Mexico, and one population in Izabal, Guatemala, then southward from Nica- ragua (Río San Juan) throughout Costa Rica to Bo- cas del Toro, Panama, with a disjunct population in the Chocó of Colombia. They grow between sea level and 1500 m in elevation. Ecology. Species of Ardisia subg. Auriculardi- sia sect. Fagerlindia occur in primary, secondary, and remnant, premontane wet and pluvial forests and in evergreen Liquidambar—Quercus forests. Etymology. The name "fagerlindia" is derived from the acd: tural model exhibited by the mem- bers of this section. Ardisia subg. Auriculardisia sect. Fagerlindia is defined by the species' exhibiting Fagerlind's Ar- chitectural Model (Hallé et al., 1978), and its ter- minal, pendent inflorescences. One of the outstand- ing features exhibited by Fagerlind's Model is that the flowering shoots bear leaves markedly different in shape and size from those of the vegetative shoots. KEY TO THE ТАХА OF ARDISIA SUBG. AURICULARDISIA SECT. F'AGERLINDIA la. Trunks and shoots, leaf blades, and inflorescence irtellous-tomentose, the trichomes ap- 0.8-1. p mm lon 6. Ardisia nevermannii у; Теш wid shoots, leaf tni d inflorescence rachises. furfuraceous- -lepidote and/or with cu- mm long. rface a mixture of urfuraceous-lepidote and/or orm scales; corolla lobes 4.5— 3.1— aia 2 long P . Ardisia brenesii . Indument of lower leaf sil : nu ur- furaceous-lepidote sc i гаа obes 1.3— 4.2 mm long; anthers 1.2-3.1 n Ja. Vegetative shoot leaf blades 7 5-21 5 X .9-7.6 cm 4a. Leaf blades broadly «ниш, асше rachises parently unicellular, — erect c сири for 4.8 mm long; anthers N = mucronate-apiculate --------------------- 13. Ardisia bastonalensis Leaf blades — elliptic to ob- lanceolate or spathulate, gradually A = florescence bracts persistent; corol- la lobes black punctate and punctate-lineate; anthers not mu- Eran bae apiculate. af blades crenate to dentate: 1.7 ly punctate, reds IE ees UN 7. Ardisia preset 5b. Leaf blades rita peduncles 212 Annals of the Missouri Botanical Garden 5.5-15.5 em long; calyx lobes 1.3-1.5 X 0.8-1.0 mm; corolla lobes 3.2-3.4 X 1.9-2.0 mm; anthers 1.8-2.0 mm long, on apically free filaments 0.7—0.€ mm long: fruits 7.0-7.5 mm diam., Gi dinis pellu- cid-punctate, = "slate disia gordonii Mi vds hot Тенг ы 25.2—42.6 9.4—11.1 ст. dins hlets 7.5-8.5 mm diam.; veg- ative к leaf blades 39. 842 6 3b. a 2.4—12.8 em long; calyx lobes IL. 2 mm long sees - 12. Ardisia apoda йкы Меш 4.5 5—5.4 mm diam.; veg- tative shoot leaf blades 25.2-34.6 cm long; apically abruptly acumi- petioles of reproductive shoot 6b. - A D 3. Ardisia А rensis 12. Ardisia apoda Standl. & Steyerm., Publ. Field Mus. Nat. Hist., Bot. Ser. 23: 219. 1947. Icacorea apoda (Standl. & Steyerm.) Lundell. Phytologia 49: 347. 1981. Auriculardisia apo- da (Standl. & Steyerm.) Lundell, Wrightia 7 266. 1984. TYPE: Guatemala. Izabal: Cerro San Gil, 300—900 m, 25 Dec. 1941 (fr), J. Ste- yermark 41923 (holotype, Е!, Е neg. 68130!, LL neg. 1971-18!). Figure 15. Small trees to 6.1 m tall. Trunk and vegetative shoots 7.5-8.5 mm diam., densely furfuraceous-lep- idote; reproductive shoots similar, but 3—4.2 mm diam. Leaves dimorphic; vegetative shoot leaves with the blades membranous, Pro to slightly ob- lanceolate, 39.8—42.6 x 9 cm, apically long attenuate to an indistinct acumen, basally auricu- late, inconspicuously punctate and punctate-lin- eate, glabrous above, sparsely furfuraceous-lepidote below except denser along the midrib, the midrib impressed above, prominently raised below, the secondary veins 28 to 36 pairs, slightly depressed above, prominulous below, the margins entire to slightly crenulate, flat or revolute; petioles stout marginate, obsolete to 0.3 cm long, furfuraceous- lepidote; reproductive shoot leaves with the blades similar to the vegetative ones but 9.3-12.9 X 3.4— 3.6 cm, the secondary veins 9 to 21 pairs; petioles similar to the vegetative ones. /nflorescences pin- nately or bipinnately paniculate, 12.4—12.8 ст long. longer than the leaves, the rachis, branches, and pedicels densely furfuraceous-lepidote, the branches terminating in 7- to 12-flowered corymbs; peduncle 3.1—6.5 cm long; inflorescence bracts persistent, oblong, 1.4—1.6 X 4.8-5.4 mm, apically acute to rounded, midrib prominulous above and below, prominently punctate and punctate-lineate, the margins entire, flat; inflorescence branch bracts similar to the inflorescence bracts, but 1.7-2.2 X 0.4—0.9 mm: floral bracts unknown, early caducous; pedicel 7.4-9.1 mm long, prominently punctate and punctate-lineate. Flower color unknown; calyx lobes chartaceous, suborbicular to oblate, 1—1.: 1.1-1.3 mm, apically acute, prominently punctate and punctate-lineate, sparsely furfuraceous-lepi- dote, the margins minutely erose, hyaline, sparsely glandular-ciliolate; corolla, stamens, and pistil un- known. Fruits 7.4-8.5 mm diam., prominently punctate and punctate-lineate, slightly costate. Distribution. Ardisia apoda is only known from Cerro San Cil, Izabal, Guatemala, growing from 300 to 900 m in elevation. Ecology and conservation status. is found on damp, forested slopes and ravines. Be- cause it is only known from the type collection, it should be considered threatened. Etymology. The specific epithet refers to the sessile leaves. Because this species is only known from the type, the relationships of Ardisia apoda are uncer- tain. However, within Ardisia subg. Auriculardisia sect. Fagerlindia, Ardisia apoda may be most easily confused with Ardisia tortuguerensis by virtue of its large vegetative shoot leaves to 42.6 X 11.1 cm, gradually tapering to the base, the relatively thick stems to 8.5 mm in diameter and pedicels to 9.1 Ardisia apoda — А. Flowering al surfac Ardisia bastonalensis. Flower Stamen, Figure 16 (left). inflorescence. —D. abaxi: е. و branc h. —B. Detail of abaxial leaf surface. —C. Detail of | Stamen, lateral margin. —G. Stamen, adaxial surface. U).) (A-G drawn from т R. Cedillo T. & G. Higareda 2890 (ME X Figure 17 (right). Ardisia brenesii. —A. vesliture ee of a mixture of e sessile flat or d sc m of inflorescence. » nen, abaxial surface. Fruit. (A, B Pas from R. Riviere 3 358 (MO); C-H Атоо 11183 (F).) Flowering branch. —B. viri of abaxial leaf surface. —C. Detail of amen, lateral margin. H.S from А m et al. 8993 (MO); I Коп W. Burger & T. Volume 90, Number 2 Ricketson & Pipoly 213 2003 Revision of Ardisia subg. Auriculardisia sk gy T2 ES. K 214 Annals of the Missouri Botanical Garden mm long. However, Ardisia apoda is easily sepa- rated from А. tortuguerensis by its much thicker branchlets to 8.5 mm in diameter, much longer veg- etative shoot leaves to 42.6 cm long with attenuate- acuminate apices, longer inflorescences to 12.8 cm long and larger calyx lobes to 1.2 X 1.3 mm. 13. Ardisia bastonalensis Ricketson & Pipoly, sp. nov. T Veracru pio. Catemaco, Rancho La Chingada, 1 10 km al SE de Tebanca, camino a Bastonal, 22 Nov. 1984 (fl), К. Cedillo TI. & С. Higareda 2890 (holo- type, MEXU!; isotype, US!). Figure 16. exico. Propter folia dimorpha etiam inflorescentiam termina- lem ad Ardisiam sectionem Fagerlindiam pertinet. Spe- cies haec ab aliis speciebus sectionis laminis foliaribus ellipticis ad bases acutis, petiolis conspicuis 12-27 mm longis, pedunculis longioribus, perianthiis pellucido- punctato-lineatis denique antheris mucronatis statim se- parabilis Subshrubs 0.9 m tall. Trunk and vegetative shoots 3—7 mm diam., sparsely and minutely furfuraceous- lepidote; reproductive shoots similar, but 1-2 mm diam. Leaves dimorphic; vegetative shoot leaves with the blades membranous, elliptic, 8.8—15.1 х 3.8-7.6 cm, apically acuminate, with an acumen 5-9 mm long, basally acute, slightly decurrent on the petiole, conspicuously punctate and punctate- lineate, glabrous above, sparsely furfuraceous-lep- idote below, glabrescent, the midrib impressed above, prominently raised below, the secondary veins 40 to 45 pairs, slightly raised above and be- low, the margins entire, flat; petiole slender, mar- ginate, 12-27 mm long, glabrous above, sparsely furfuraceous-lepidote below, glabrescent; reproduc- tive shoot leaves with the blades similar to the veg- 2.2-5.1 em, ondary veins 23 to 29 pairs; petioles similar to the etative ones, but 3.2-9.4 х the sec- vegetative ones, but 4—12 mm long. /nflorescences pinnate, 1.6—4 X 0.8—5 cm, shorter than the leaves, the rachis and pedicels with densely furfuraceous- lepidote scales, the branches terminating in 11- to 17-flowered corymbs; peduncles 4—9 mm long: in- florescence bracts and branch bracts unknown (ear- ly caducous); floral bracts membranous, oblong, 2— 2.6 X 0.4—0.6 mm, apically acute, inconspicuously punctate and punctate-lineate, the margins minute- ly erose, apically hyaline, sparsely glandular cili- olate; pedicels 5.2-11.2 mm long, inconspicuously punctate and punctate-lineate, with scattered to dense minute furfuraceous-lepidote scales. Flowers i lobes membranous to chartaceous, 1.4-1.6 X 1.1-1.3 mm, apically acute to rounded, prominently punc- white; calyx suborbicular to widely ovate, tate and punctate-lineate, pellucid, glabrous ada- Ardisia limonensis Lundell, xially, glabrous or sparsely furfuraceous-lepidote, the margins minutely erose, hyaline, sparsely glan- dular ciliolate; corolla 5.4—5.6 mm long, the tube 1.2-1.4 mm long, the lobes narrowly ovate to ovate, 4—4.2 X punctate and punctate-lineate, pellucid, glabrous 2.2-2.4 mm, apically acute, prominently throughout, the margins entire, hyaline; stamens .3—4.4 mm long, the filaments 1.4—1.6 mm long, the staminal tube 1.2-1.3 mm long, the apically free portions 0.2-0.3 mm long, epunctate, the an- thers lanceoloid, 2.9-3.1 X 1.2-1.6 mm, apically mucronate-apiculate, basally sagittate, the connec- 5.6-5.9 mm long, the ovary 1.6-1.9 mm long, the styles 3.8—4 tive inconspicuously punctate; pistil 5 mm long, epunctate, the ovules 10 to 12. Fruits unknown. Distribution. Ardisia bastonalensis is endemic to the area around Bastonal, in the Mpio. de Ca- temaco, Veracruz, Mexico, growing at around 500 m in elevation. Ecology and conservation status. Ardisia bas- tonalensis occurs in tall evergreen Liquidambar- Quercus forests on clay soils. Because of its re- stricted distribution, it should be considered threatened. Etymology. The specific epithet refers to the type locality, around Bastonal, Veracruz, Mexico. ithin Ardisia subg. Auriculardisia sect. Fager- lindia, Ardisia bastonalensis is most closely related to A. tilaranensis and A. gordonii by their smaller vegetative shoot leaves less than 21.5 em long and 7.6 em wide. However, the elliptic leaves with acute bases, conspicuous petioles to 2.7 ст long in the vegetative shoot leaves and long peduncles to 9 mm long, the pellucid (not black) punctate-lin- eate perianth, and the markedly mucronate-apicu- late anthers clearly set it apart from these other two species. Paratype. MEXICO. Veracruz: Mpio. de Catemaco, Cumbres de Bastonal, 19 Nov. 1974 (fl), R. Cedillo T. 424 (MEXU, US). 14. Ardisia brenesii Standl., Publ. Field Mus. Ser. 18: 855. 1938. Auricular- disia brenesti (Standl.) Lundell, Phytologia 49: 342. 1981. TYPE: Costa Rica. Alajuela: Ca- taratas (Los Angeles) de San Ramón, vicinity of San Ramón, 17 Apr. 1935 (fl), A. Brenes 20537 (holotype, F!, F neg. 68135!, LL neg. 1971-23!; isotypes, NY! [2]). at. Hist., Bot. Figure 17. Wrightia 6: 79. 1979. Syn. nov. Auriculardisia egere (Lundell) Lundell, Jb deis a 49: 344. ] TYPE: Costa Rica. Limón: r km W of GA. S border Hacienda La Volume 90, Number 2 2003 Ricketson & Pipoly 215 Revision of Ardisia subg. Auriculardisia Suerte, E of sentry pue 10°30'N, wi 226. 40 m, 15 Mar. 1978 (fl, fr), C. Davidson, A. iner, Middleton & B. Аи» 7009 (holotype, m F ne 55677!; isotypes, F!, F neg. 68446!, LL!). Shrubs or small trees 2-7 m tall, 1-3 cm diam. Trunk and vegetative shoots 4—11.5 mm diam., in- dument with a mixture of dense ferrugineous-lepi- dote and scattered erect cupuliform scales, 0.1—0.9 mm long, the scale margins lobed or with 2 to 8 arms; reproductive shoots similar, but 1.2-3.5 mm diam. Leaves dimorphic; vegetative shoot leaves with the blades membranous, elliptic to oblanceo- late, 22.5-53.7 X 5.4-22.6 cm, apically acumi- nate, with an acumen 1.3-3.2 cm long, basally auriculate, prominently punctate and punctate-li- neate, glabrous above, indument with a mixture of dense ferrugineous-lepidote and scattered erect cu- puliform scales, 0.1—0.9 mm tall, the scale margins lobed or with 2 to 8 arms, the midrib impressed above, prominently raised below, the secondary veins 23 to 36 pairs, slightly depressed above, pro- minulous below, the margins entire or regularly cre- nate, flat or revolute toward the base; petioles stout, marginate, 0.3-0.8 cm long, indument similar to the vegetative shoots; reproductive shoot leaves with the blades similar to the vegetative shoots but oblong, 4.9-23.9 X 0.7-8.8 mm, the secondary veins 3 to 30 pairs; petioles similar to the vegeta- tive ones but 0.1—0.5 em long. Inflorescences bipin- nately or tripinnately paniculate, 9.2—35.5 cm long, longer than the leaves, indument as in shoots, the branches terminating in 7- to 19-flowered corymbs; peduncle 3.1-19.4 cm long; inflorescence bracts early caducous, oblong, 1.5—8.6 X 0.5-2.4 cm, api- cally acute to rounded, prominently punctate and punctate-lineate, glabrous above, indument below similar to the vegetative shoot leaves, the midrib prominulous above and below, the margins entire, flat or revolute toward the base; inflorescence branch bracts similar to the inflorescence bracts, but 1.1-3.3 x 0.5-14.9 mm; floral bracts similar to the inflorescence bracts, but 0.8-1.2 X 0.2-0.5 mm; pedicels 5.4—10.2 mm long, pellucid punctate and punctate-lineate, indument similar to the veg- etative shoots. Flowers pink, purple, or red; calyx lobes membranous to chartaceous, ovate to subor- bicular, 1.2-1.4 X 0.9-1.2 mm, apically acute, prominently punctate and punctate-lineate, with sparsely ferrugineous-lepidote or cupuliform scales, the margins minutely erose apically, hyaline, sparsely glandular-ciliolate; corolla 5—5.3 mm long, the tube 0.5—0.7 mm long, the lobes widely ovate, 4.5-4.8 X 2.2-2.4 mm, apically acute, prominently punctate and punctate-lineate, sparsely furfura- ceous-lepidote abaxially, the margins entire; sta- mens 3.8—4.1 mm long, the filaments 1-1.2 mm long, the staminal tube 0.3-0.7 mm long, the apical free portions 0.3—0.7 mm long, the anthers narrowly ovoid to lanceoloid, 3.1-3.4 X 1.1-1.2 mm, api- cally apiculate, basally sagittate, the connective punctate; pistil 2.9-4 mm long, the ovary 1-1.2 mm long, the style 1.7-2.8 mm long, inconspicuously punctate, the ovules 12 to 14. Fruits 6.4—7. diam., prominently punctate and punctate-lineate, costate. Distribution. Ardisia brenesii occurs from Río San Juan, Nicaragua, through Alajuela, and Here- dia to Limón, Costa Rica, with a disjunct popula- tion in San Blas, Panama, growing at О to 1500 m in elevation. This is the first report of the species from Panama. Ecology and conservation status. Ardisia bre- nesii occurs in primary, secondary, and remnant premontane wet forests. Fieldwork by Pipoly in Al- ajuela, Costa Rica, revealed that the species occurs at a density of less than a dozen individuals per hectare, normally on slopes beside small water- courses. While it is certainly not common, at this time there are no data to suggest that the species is threatened. Etymology. This species was named in honor of Alberto M. Brenes of the Museo Nacional de Costa Rica. Ardisia brenesii is unique within Ardisia subg. Auriculardisia sect. Fagerlindia because of the mixture of ferrugineous-lepidote and cupuliform scales present throughout the plant, although it clearly belongs to the section because of the di- morphism between the leaves of the vegetative and reproductive shoots. Ardisia brenesii may be most easily confused with A. tortuguerensis because of its large, strikingly dimorphic leaf blades to 53.7 cm long, those of the vegetative shoots elliptic to oblanceolate with obvious acumen to 3.2 cm long. and gradually tapering toward the base. However, Ardisia brenesii, with its mixed vestiture, apiculate- subulate anthers, much larger calyx lobes to 1.4 mm long, and corolla lobes furfuraceous-lepidote ko poi is easily recognized. e type of Ardisia limonensis is identical to pop- ulations of A. brenesii, except it is more robust. It is important to note that the MO holotype of A. limonensis is made of two vegetative shoot leaves and an inflorescence, and both the F and LL iso- types are made up of reproductive shoot leaves with an attached inflorescence. Specimens examined. NICARAGUA. Río San Juan near Río San Juan at “El Relos," ca. midpoint between El Castillo and Delta de San Juan, 23 Mar. 1961 (fl), С. 216 Annals of the Missouri Botanical Garden Bunting & L. Licht 791 (F |2], LL, NY); Buena Vista 1 km W of mouth of Río San Juan, : З Sep. 1983 (fl, fr), А Martínez & R. Riviere 2072 (MEXU, US); trail between e | Castillito or Caño de Oro, toward a erro el Gigante, 15 Sep. 1982 (fr), К. ۰ : 2159 (MEXU); Río San Juan; El 2 (fr), R. Riviere 356 ‚ К. Rueda et al. 1798 (HULE, MO); Reserva Ha SR slopes faci s € of Río San Carlos, on Js Rio San Juan, 11 Feb. 1996 (fl), R. Rueda et al. 405 Spa a ч Mpio бап m. del Norte, Reserva bns Maí › El Gigan te, 5 km del Río San Juan, 21 Sep. 1998 ү "R Rueda et al. 8877 (HULE, MO). COSTA RICA. Alajuela: on Caribbean slope between San Lorenzo and Los An E es de San Ramón, above the Río San Lo- renzo, 20 Sep. 1978 (fr), W. Burger & T. Antonio 11183 (F); Cantón de San Ramón, Hein Forestal San Ramón, Cordillera de Tilarán, trail to the Station, 6 Mar. 1995 (fl), G. Carballo 514 (CR, FTG, INB, MO); Reserva Biológica Monte jon valley of Río Peñas Blancas, 24 Apr. 1987 (fl), W. Haber & E. Cruz 6997 (CR, e MO, US); Re 'serva Forestal de San Ramón, ca. 10 km W of pd along Hío San Lorencito, 30 May-1 June "1986 (fl, fr), B. Ham- mel et al. 15261 (LL [2], MO [2]; Cantón de San Ramón, Reserva Forestal San Ramón, Volcán Muerto jc El Sahino Trail, 26 Apr. 1993 (fl), E. López et al. 42 (CR, F. FTG INB, MO); Reserva Biológica Monteverde, Estación e 2 Oct. 1990 (fl), М. Obando et al. 201 (CR, INB, MO); Los Angeles de San Ramón, 20 July 1984 (fr), J. Pipoly 7113 (CR, NY, TEX, US). Heredia: Cantón de eir aad Cuenca del Sarapiquí, Sardinal Hills, 15 km N of Puerto Viejo, 11 Jan. 1997 (fr), B. Hammel 20661 E MO). Limón: hills 3.5 airline km S of Isla Buena Vista in the Río Colorado, 16 airline km SW of Barra de Colorado, 15-16 Sep. 1986 (fl, fr), G. Davidse & G. He rera 31185 (MO); Refugio Barra del Colorado, between jn Chirripocito and Río Sardina, “sardinal” on Chirripó Atlántico quadrangle, 11 Nov. 1988 (fl), M. ты эн et al. 8993 (CR, M( n Cantón de Pococi, R.N.ES. Barra del Colorado, Llanura de Tortuguero, Sector ues 4 Jan. 199] (fr), E. Rojas 206 (CR, INB, MO); El Cedral, xii km N of Cariari, on the farm of Mario Chavarría P., 13 Sep. 1994 (fl), К. Thomsen 1024 (C, MO). PANAMA. San Blas: trail from Río Esadí to Cerro Banega, 21 Dec. 1985 (fl, fr), G. de Nevers & H. Herrera 6646 (MO [2] = = 7 = 15. Ardisia gordonii Ricketson & Pipoly, sp. nov. TYPE: Panama. Bocas del Toro: above Chiriquí Grande, 08?55'04"N, 082?10'04"W, 300 m, 26 Dec. 1986 (fl), G. McPherson & J. Aranda 10163 (holotype, MO!; isotype, LL!). Figure 18. Ob laminam secus margines integerrimam lobos caly- cinos 1.3-1.5 mm ги 0.8-1.0 latosque lobos c йн 3.2-3.4 mm longos 1.9-2.0 mm latos denique antheras 1.8-2.0 mm longas 1.0-1.1 mm latas, A. bastonalensi val- de arcte affinis, sed ab ea habitu fruticoso (non ded coso) lobulis calycinis 0.8-1.0 (non 1.1—1.3) mm latis, antheris 1.8-2.0 (non 2.9-3.1) mm longis 10-1. 1 (nec .2-1.6) mm latis, denique pistillo 4.2—4.4 (non 5.6-5.9) longo facile distinguitur. Small trees to 4 m tall. Trunk and vegetative shoots 3—4.5 mm diam., densely furfuraceous-lepi- dote, the scales flat, pallid, reproductise shoots sim- ilar to the vegetative shoots, but 1.5-2.5 mm diam. Leaves dimorphic; vegetative shoot leaves with the blades membranous, narrowly elliptic to oblanceo- late ог spathulate, 11.5-21.5 X 3.2-5.6 cm, cally acuminate, with an acumen 5-14 mm long, api- gradually tapering to an auriculate, amplexicaul base, prominently punctate and punctate-lineate above and below, glabrous above, densely furfura- ceous-lepidote, more scattered on the midrib, the midrib impressed above, prominently raised below, the secondary veins 37 to 45 pairs, prominulous above and below, the margins entire, flat; petioles obsolete; reproductive shoot leaves with the blades similar to the vegetative shoot leaf blades, but 6.4— 9.8 X 1.5-3.4 cm. Inflorescences bipinnately panic- ulate, 10-22 X 5.5-8 cm, longer than the leaves, densely furfuraceous-lepidote, the branches termi- nating in 5- to 12-flowered corymbs; peduncles 5.9-15.5 em long; inflorescence bracts similar to the vegetative shoot leaf blades, but oblanceolate 34.3 X 0.4-0.8 em; а bracts persistent, membranous, oblong, 4.5— 5.6 X 1.4-2.1 mm, apically acute, the veins absent, or oblong, 1.€ inflorescence саве punctate апа punctate-lineate, gla- brous above, sparsely furfuraceous-lepidote below, the margin minutely erose, hyaline; floral bracts similar to the inflorescence branch bracts, but 0.7— 2.1 X 0.4-0.8 mm: pedicels 0.6-1.1 cm long, prominently punctate and punctate-lineate, sparse- y furfuraceous-lepidote. Flowers pink; calyx lobes X 0.8-1 mm, apically acute, prominently punctate and punctate-lineate, membranous, ovate, 1.3—1.5 glabrous adaxially, sparsely furfuraceous-lepidote Figure 18 (left). re Sei "lower Stamen, abaxial surface. O).) Figure 19 (right). G. Stamen, adaxial surface. Ardisia gordonii. —A. Vid. branch. —B. pues of abaxial leaf Suid —C. —E. —EF. Stamen, lat —H. Fruit. (A, B & H drawn from holotype, G. McPherson & J. Aranda 10163 (MO); ¢ (M Ardisia nevermannii. —A. Flowering branch. —B. Detail of abaxial leaf ce се. —C indument, stiff, mostly erect modified е —D. Detail of е —E. Flower. — en, lateral margin. — Detail of adaxial surface. McPherson 8753 ral margin. —G. Stamen, G from G. . Detail of ex surface. —l. Fruit. (A, B drawn from M. cman el v 7925 (MO): C-H from C. Barbosa 6591 (МО); I n L. Poveda A. et al. 4174 (F)) Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 217 218 Annals of th онт Botanical Garden abaxially, the margins minutely erose, hyaline, 1926 (fl), P. Standley & J. Valerio 48603 (ho- sparsely glandular ciliolate; corolla 4.4—4.6 mm lotype, US!, LL neg. 1971-73!, US neg. 2377!). ve the tube 1-1.4 mm long, the lobes ovate, 3.2— X 1.9-2 mm, apically acute, prominently black punctate, glabrous throughout, the margins entire, hyaline; stamens 2.5-2.8 mm long, the filaments 1.7-1.8 mm long, the staminal tube 0.9-1 mm long, the apically free portions 0.7—0.9 mm long, epunc- tate, the anthers ovoid, 1.8-2 X 1-1.1 mm, apically apiculate, basally deeply cordate, the connective inconspicuously punctate; pistil 4.2—4.4 mm long, the ovary 1-1.1 mm long, the styles 3.1-3.4 mm long, inconspicuously punctate, the ovules 8 to 11. Fruits 7—7.5 mm diam., inconspicuously pellucid- punctate. Distribution. Ardisia gordonii is endemic to the area above Chiriquí Grande around the area of Cer- ro Pila de Arroz, in Bocas del Toro, Panama, grow- ing at 300—500 m in elevation. Ecology and conservation status. Ardisia gor- donii occurs on slopes of premontane wet forest. Because of its restricted distribution it should be considered threatened. Etymology. It is an honor to dedicate this spe- cies to Gordon McPherson, a curator at the Mis- souri Botanical Garden. Gordon is an indefatigable collector, prodigious floristician, a noted specialist in African and Central American floras, and above all, a scholar and gentleman. Within Ardisia subg. Auriculardisia sect. Fager- lindia, Ardisia gordonii appears to be most closely related to A. tilaranensis by the narrowly elliptic to oblanceolate or spatulate leaf blades gradually ta- pering to an auriculate base, the very short or ob- solete petioles, persistent inflorescence bract, black punctate corolla lobes, and apiculate anthers. How- ever, Ardisia gordonii can easily be separated from ói tilaranensis by the entire leaf blades, calyx lobes o 1.5 X 1 mm, much larger, black punctate corolla lobes to 3.4 X 2 mm, larger anthers to 2 X 1.1 mm on longer apically free filaments to 0.9 mm long, and larger non-costate fruit to 7.5 mm in diameter. Paratype. PANAMA. Bocas del Toro: along road to Chiriquí Grande, 10 road mi. from C ше Divide and 2 mi. along pipeline access road E of hwv.. on Cerro Pila 8 de Arroz, 10 Mar. 1986 (fl, fr), G. Mc она 8753 (LL, MO). 16. "m nevermannii Standl., J. Wash. Acad. Sci. 17: 524. 1927. Valerioanthus nevermannii (Standl.) T Wrightia 7: 50. 1982. Auri- culardisia nevermannii (Standl.) Lundell, Phytologia 57: 450. 1985. TYPE: Costa Rica. Limón: Finca Montecristo, on the Río Reven- tazón, below El Cairo, ca. 25 m, 18-19 Feb. Figure 19 Shrubs or subshrubs 3 m tall. Trunk and vegeta- tive shoots 6-9 mm diam., hirtellous-tomentose, the trichomes apparently unicellular, the hairs mm long; reproductive shoots similar, but 2.5—4.5 mm diam. Leaves dimorphic; vegetative shoot leaves with the blades membranous, narrowly oblong to oblanceolate, 34.2—42.5 X 10.2-13.9 ст, apically acuminate, with an acumen 1.3-2.9 ст long, ba- sally obtuse or slightly auriculate, prominently punctate and punctate-lineate, hirtellous-tomentose above and below, the hairs 0.8-1.8 mm long, much denser along the midribs, the midrib impressed above, prominently raised below, the secondary veins 26 to 42 pairs, slightly impressed above, pro- minulous below, the margins entire, flat; petioles stout, canaliculate, 0.7-1.8 cm long, hirtellous-to- mentose above and below, the hairs 1.4—1.8 mm long; reproductive shoot leaves with the blades sim- ilar to the vegetative ones but 5.5-27.5 X 2.7-8.9 mm, the secondary veins 13 to 31 pairs; petioles similar to the vegetative ones but 0.2-1.2 cm long. Inflorescences pinnately or bipinnately paniculate, 11—22.5 cm long, longer than the leaves, hirtellous- tomentose, the hairs 1.4—1.8 mm long, the branches terminating in 3- to 8-flowered corymbs; peduncle 4.1-12.7 cm long; inflorescence bract unknown; in- florescence branch bracts early caducous, oblong. 0.5-2.6 X 1.1-8.2 mm, apically acute to rounded, prominently punctate and punctate-lineate, hirte- lous-tomentose, the hairs 0.8-1.8 mm long, the midrib prominulous above and below, the second- ary veins inconspicuous, the margins entire; floral bracts similar to the inflorescence branch bracts but 0.4—1.6 X 0.2-0.4 mm, basally sessile; pedicel 6.2-7.6 mm long, conspicuously punctate, hirte- llous-tomentose, the hairs 1.2-1.8 mm long. Flow- ers white or light pink; calyx lobes membranous to chartaceous, ovate to suborbicular, 1-1.2 X 1 1.4 mm, apically acute, prominently punctate and punctate-lineate, sparsely hirtellous abaxially, the hairs 0.3-0.8 mm long, glabrous adaxially, the mar- gins irregular, minutely erose, hyaline, sparsely glandular-ciliolate; corolla 4.3—4.6 mm long, the tube 0.3—0.7 mm long, the lobes widely ovate, 3.9— 4.3 X 2.1-2.3 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the margins entire; stamens 3.1—3.3 mm long, the filaments 0.9-1 mm long. the staminal tube 0.3— 0.7 mm long, the apically free portions 0.3—0.6 mm long, the anthers narrowly ovoid to lanceoloid, 2.4— 2.7 X 0.9-1.1 mm, apically caudate, basally sag- Volume 90, Number 2 200 Ricketson & Pipoly 219 Revision of Ardisia subg. Auriculardisia ittate, the connective punctate dorsally; pistil 3.1— .3 mm long, the ovary 0.5-0.6 mm long, the style 2.5-2.7 mm long, inconspicuously punctate, the ovules 7 to 9. Fruits 7.2-1.9 i nently punctate and punctate-lineate, costate. mm diam., promi- Distribution. Ardisia nevermannii occurs in San José and Limón, Costa Rica, and the Chocó of Co- lombia. It is not known from Panama, but should be expected in the Darién, growing from sea level to 850 m in elevation. Ecology and conservation status. Ardisia never- mannii occurs in wet and pluvial forests. While it is certainly not common, there are no data to sug- gest the species is threatened at this time. Etymology. Standley (1927: 524) stated that "the species is named for Mr. Ferdinand Never- mann, a keen student of Costa Rican Coleoptera Ardisia nevermannii is unique among all the members of Ardisia subg. Auriculardisia sect. Fa- gerlindia because of its hirtellous-tomentose vesti- ture throughout the plant and the caudate apices of the anthers. 2 examined. COSTA RICA. Limón: Ham- bur, own river from Reventazón, 2 May 1930 (fl), G. к: 685 (М); Zona Protectora Barbilla, W side of plateau — санара: of N fork of Río Dantas Froni headwa rada varreal, Río Barbilla drain- dillera de Talamanca, i Apr. s & A. Chacón 2683 (CR, FTG, INB, MO); Río Reventazón drain- age basin, 23 Oct. Tes (fr), P. Shank & A. Molina К. 4412 (F); Finca Montecristo, on (he Río Reventazón, be- $ 9 Feb. 1926 (fl), P. Standley & J. Valerio ; Ha mbur Ld on the Ri : 6 (fl, P. Standley & J. Valerio 48754 (US), 48824 (U = San José: Carrillo Station, Brau- Carrillo, 19 Apr. 1984 (fr), L. Gómez et al. 21159 (LL, MO); Parque Nacional Braulio Carrillo, Estación Carrillo, 27 Nov. 1986 (fr), L. Poveda A. et al. 4174 (F); Laguna on hills along Río Corinto, Parque Nacional Braulio Ca- rrillo, 16 Aug. fr), P. Sánchez & N. Zamora 561 (MO). COI OMBIA. Ki Mpio. de Nuquí, rd miento Termales, uebrada Piedra, э, 85 1994 (fr), P. Aoda- Кик. et al. 6875 (FTG, US); iil: on Morro de Mico to the scenic е е. ‘jurubidá” pun erg ard and тэдн оп “Со va" trail north- 15 May 1990 (fl, fr), C. Barbosa 6591 (CHOCO, Dd 12, MO [2]: NW of Alto Curiche, 20 May 1967 (fr), e & J. Idrobo 11248 (LL); Mpio. Bahía de Solano, а an Nacional Natural Ensenada de Utría, trail between Punta Diego and Caída Cocalito, 18 Apr. 1990 (fr), J Espina et al. 3639 (CHOCO, MO); Mpio. Bahía de Solano, Jos en Nacional Natural Ensenada de Utria, trail between t e and Cocalito beach, 23 Apr. 1990 (fr), J. Espina et a 3827 (CHOCO, MO). 17. Ardisia tilaranensis Standl., J. Wash. Acad. 17: 524. 1927. Auriculardisia tilaranensis (Standl.) Lundell, Phytologia 49: 345. 1981. TYPE: Costa Rica. Guanacaste: Quebrada Se- rena, SE of Tilarán, ca. 700 m, 27 Jan. 1926 (fr), P. Standley & J. Valerio 46169 (holotype. US!, US neg. 2389!, LL neg. 71-115!). Figure 20. — Subshrubs to 3 m tall. Trunk and vegetative shoots 2.5-5 mm lepidote, reproductive shoots similar to the vegeta- tive shoots, but 1.5-3.5 mm diam. Leaves dimor- phic; vegetative shoot leaves with the blades membranous, narrowly elliptic, 7.5—17.2 X 1.94.2 cm, apically attenuate to acuminate, with an acu- men 0.7-1.4 cm long, gradually tapering to an au- riculate base, prominently punctate and punctate- sparsely diam., densely appressed furfuraceous- lineate. essentially glabrous above, appressed furfuraceous-lepidote below, the midrib impressed above, prominently raised below, the secondary veins 21 to 34 pairs, slightly depressed above. prominulous below, the margin crenate to dentate, flat; petioles stout, marginate, subobsolete to 2.5 mm long, glabrous above, densely furfura- ceous-lepidote below; reproductive shoot leaves with the blades similar to the vegetative ones, but 3.5-14.7 X 1.1-4.3 em. Inflorescences bipinnately to tripinnately paniculate, 5.4-9.2 X 2.5-5.6 ст, longer than the leaves, the rachis and pedicels densely furfuraceous-lepidote, the branches termi- nating in 7- to 13-flowered corymbs; peduncles 0.4—1.7 em long; inflorescence bracts persistent, membranous, ovate to lanceolate, 0.4—3.7 X 0.3— 0.9 cm, apically acuminate, otherwise as in the re- productive shoot leaf blades; inflorescence branch bracts caducous, membranous, ovate, 1.5-3.2 X 1.4-2.6 mm, apically acute, basally auriculate, prominently punctate and punctate-lineate, essen- tially glabrous above, sparsely fu dote below, the midrib жакыр above, slightly prominent below: pedicels 4.5—6.2 mm long, prom- inently punctate and punctate-lineate, densely fur- furaceous-lepidote. Flowers pink; calyx lobes mem- branous to chartaceous, ovate, 1-1.1 X 0.6-0.7 mm, apically acute, prominently punctate, sparsely furfuraceous-lepidote, the margins minutely erose, hyaline, sparsely glandular-ciliolate; corolla 2.8— 2.9 mm long, the tube 1—1.1 mm long, the lobes ovate, 1. 9 x 1.2-1.3 mm, apically acute, prominently black punctate and punctate-lineate, glabrous throughout, the margins entire; stamens 2.1-2.2 mm long, the filaments 1—1.1 mm long, the staminal tube 0.4—0.6 mm long, the apically free portions 0.6-0.7 mm long, epunctate, the anthers ovoid, 1.2-1.3 X 0.4—0.5 mm, apically apiculate, basally cordate, the connective inconspicuously 220 Annals of the Missouri Botanical Garden Е л Y PES. 3 « eja sc TET) Volume 90, Number 2 2003 Ricketson & Pipoly 221 Revision of Ardisia subg. Auriculardisia punctate; pistil 2.9-3 mm long, the ovary 0.6—0.7 mm long, the style 1.8-2.1 mm long, prominently and conspicuously punctate, the ovules 10 to 1: Fruits 6-6.5 mm diam., prominently punctate, slightly costate. Ardisia tilaranensis is endemic to located in Guanacaste Distribution. the Cordillera de Tilarán, and Alajuela, Costa Rica, growing from 600 to 1400 m in elevation. Ecology and conservation status. ranensis occurs in premontane humid forests. We believe that the protected status of the areas where this species has been recorded is enough to protect it for the time being. Etymology. The specific epithet referred to the type locality, along the Cordillera de Tilarán in Ardisia tila- Cantón Guanacaste, Costa Rica. By virtue of its subsessile leaf blades gradually tapering to an auriculate base and apiculate an- thers, Ardisia tilaranensis is most closely related to A. gordonii within Ardisia subg. Auriculardisia sect. Fagerlindia. However, Ardisia tilaranensis can be easily separated from A. gordonii by its crenate to dentate leaf blade margins, the smaller calyx to 1.1 mm long and corolla lobes to 1.9 mm long, shorter, apically free portions of the fila- ments to 0.7 mm long, and the smaller fruits to 6.5 mm in diameter. Specimens examined. COSTA RICA. Alajuela: Rese T- va Biológica Monteverde, Bosque Eterno de los Niños, Quebrada Agua Gata, 25 Jan. 1990 (fr), E. Bello C. 1810 (CR, FTG, INB, MO); forest trail from Macadamia Village o summit of Cerro Chato, 3 Oct. 1991 (fr), V. Funk et al. 10984 (US); Reserva Monteverde, Poco Sol, 13 km S For- tuna, чә ege 1989 (fl), W. Haber & W. Zuchowski 9358 (CR, INB, MO); Cantón Upala, SN El Retiro, slopes of эң Montezuma, 23 July 1993 (fl), G. Herrera 6541 (CR, F); S slope of Cerro Chato, 25 Feb. 1989 (fr), G. Russell et al. 976 (US). Guanacaste: hills on wa guna de Arenal, 18 July 1962 (fl), C. Brown 17416A (F, LSU); Quebrada Grande, Tilarán, Esperanza-Las Nubes Tra il, on Continental Divide, 24 Feb. 1987 (fl, fr), W. Ha- Bello C. 6710 (MO); El Silencio, near Tilarán, E jo (fr), P. Standley & J. Valerio 44729 (US), 44763 (US); Los Ayotes, near Tilarán, 21 Jan. P. Standley & J. Valerio 45422 (US). 18. Ardisia tortuguerensis Ricketson & Pipoly, 1926 (fr), sp. nov. TYPE: Costa Rica. Limón: Parque Na- cional Tortuguero, Lomas de Sierpe, 4 km NE of gate at Parque Nacional, along the Río Sier- ре, 10°24’N, 083?33' W, 100 m, 15 Aug. 1988 fl, fr), R. Robles, С. Herrera, L. Flores & M. Rojas 2052 (holotype, MO!; isotype, CR!). Fig- 1 Ob laminam foliarem ad basim gradatim contractam, d A. apoda ramulos crassos atque M wr longos cum a apod. es abrupte acuminatis iubes 9.2-11.5 (non 2.4—12.8) cm longis, denique lobulis bim 1.3—1.5 (non 1—1.2) mm longis praeclare distat Small shrubs, height unknown. Trunk and vege- tative shoots 4.5-5.4 mm diam., dense furfura- ceous-lepidote to short cupuliform scales, the scales sessile or slightly stalked, 0.1—0.2 mm tall, lobed or with 2 to 8 arms, very similar to those of A. brenesii but of one size instead of two; reproduc- tive shoots similar, but 1.5-2.3 mm diam. Leaves dimorphic; vegetative shoot leaves with the blades membranous, oblanceolate, 25.2—34.6 X 9.6-11.1 cm, apically abruptly acuminate, with an acumen 0.3-0.5 em long, basally gradually tapering to ап auriculate base, prominently punctate and punc- tate-lineate, furfuraceous-lepidote, the sparse above, denser below, much denser basally below and along the midrib, the midrib impressed scales above, prominently raised below, the secondary veins 28 to 37 pairs, slightly depressed above, pro- minulous below, the margins entire, flat or revolute; petioles stout, marginate, subobsolete to 0.3 cm long. densely furfuraceous-lepidote; reproductive shoot leaves with the blades similar to the vegeta- tive ones but 10.3—19.4 X 4—7.4 cm, the secondary veins 25 to 32 pairs; petioles similar to the vege- tative ones but 0.2-0.7 em long. Inflorescences bi- pinnately or tripinnately paniculate, 9.2-11.5 X 6.1—8.5 cm, longer than the leaves, indument as in the vegetative shoots, the branches terminating in 5- to 9-flowered corymbs; peduncles 0.5-1.5 cm long: inflorescence bracts and branch bracts early caducous, unknown; floral bracts chartaceous, ob- 1.4-1.7 X 0.5-0.6 mm, apically acute, prom- long. € Figure 20 (left). inflorescence. —D. Flower. —E. Stamen, abaxial surfac Ardisia tilaranensis. —A. << branch. —B. Detail of abaxial leaf surface. —C. Deta il of ғ. —F. Stamen, lateral margin. —G. Stamen, adaxial surface. —H. Fruit. (A-H drawn from W. Haber & E. Bello C. 6710 (MO).) Figure 21 (right). و‎ ce. — "lo =p t. (A-H ower. — 6E. Ardisia tortuguerensis. —А. vi inim branch. —B. Detail of i pa leaf surface. E. Stamen, abaxial surface. H don from holotype, R. Robles et E 2052 (MO).) ail of —EF. Stamen, adaxial surface. —G. Stamen, uis margin. 222 Annals of the Missouri Botanical Garden inently punctate and punctate-lineate, glabrous above, densely furfuraceous-lepidote below, midrib and secondary veins inconspicuous, the margins entire, hyaline; pedicels 4.3-8.1 mm long, conspic- uously punctate, indument as in the vegetative shoots. Flowers light purple; calyx lobes membra- nous to chartaceous, ovate to suborbicular, 1.3-1.5 X 1.1-1.3 mm, nently punctate and punctate-lineate, sparsely fur- apically acute or rounded, promi- furaceous-lepidote, the margins entire, minutely erose, hyaline, sparsely glandular-ciliolate; corolla 3.1-3.4 mm long, the tube 0.5-0.6 mm long, the lobes ovate, 2.6-2.8 X 1.6-1.9 mm, apically acute, prominently punctate and punctate-lineate, gla- brous adaxially, furfuraceous-lepidote abaxially, the margin entire; stamens 2.9-3 mm long, the fila- ments 1.1—1.3 mm long, the staminal tube 0.6—0.8 mm long, the apically free portions 0.5-0.7 mm long, epunctate, the anthers narrowly ovoid to lan- ceoloid, 1.9-2.1 х 0.7-0.8 mm, apically apiculate- cuspidate, basally sagittate, the connective punc- tate; pistil 3.2—4 mm long, glabrous, the ovary 1— 1.2 mm the style 2-2. long, y long, inconspicuously punctate, the ovules 8 to 10. /m- mm mature fruits 3—4.2 mm diam., prominently punc- tate and punctate-lineate, appearing non-costate. Distribution. Ardisia known only from the type in Limón, Costa Rica, growing tortuguerensis is at 100 m in elevation. Ecology and conservation status. Ardisia tortu- guerensis occurs on fairly steep, densely forested slopes with well-drained soils. Because it is only known from the type. it should be considered threatened. Etymology. The specific epithet refers to the area in which it is found, Parque Nacional Tortu- guero, along the east coast of Costa Rica in Limón. Within Ardisia subg. Auriculardisia sect. Fager- lindia, Ardisia tortuguerensis most closely resem- bles A. apoda, because of the large vegetative shoot leaves 25.2—42.6 x 9 rensis is easily separated from A. apoda by the nar- -A-11.1 cm. Ardisia tortugue- rower branchlets to only 5.4 mm in diameter, small- er vegetative shoot leaf blades to 34.6 cm long with abruptly acuminate apices, shorter petioles of re- productive shoot leaf blades to 0.7 cm long, shorter inflorescences to 11.5 cm long, and smaller calyx lobes to 1.5 mm long. TAXONOMIC TREATMENT OF ARDISIA SUBG, AURICULARDISIA SECT. PALMANAE Palmanae Ricketson & Pipoly, sect. nov. TYPE here des- Ardisia subg. Auriculardisia sect. ignated: Ardisia palmana Donn. Sm. Quoad lobulos calycinos най è bens er dos os sub apicibus subincises ad bases auriculatos ad Ardisiam subg. Auriculardisiam eu Ab aliis ета ии sub- generis inflorescentiis terminalibus pedunc ies obsoletis (versus breves) insidentibus perfacile cognoscitur. Shrubs or small trees. Branchlets straight or flex- uous, slender to stout, terete, subterete, with fine longitudinal ridges, or angulate, with densely fur- furaceous-lepidote and/or cupuliform scales, rarely stipitate-stellate tomentellous (A. liesneri), the ves- titure mostly persistent, but at times glabrescent. Leaves monomorphic, with blades membranous to coriaceous, elliptic to oblong or obovate to oblan- ceolate, at times inconspicuously punctate and/or punctate-lineate, the margins flat or revolute. /nflo- terminal, erect, pinnately to tripinnately paniculate, pyrami- rescences mostly pendent, sometimes dal, or obpyramidal, mostly shorter than the leaves, the branches loosely to tightly congested into cor- ymbs; peduncle short to obsolete, inflorescence bracts usually early caducous, inflorescence branch bracts and floral bracts often small and early ca- ducous; pedicels slender to stout, terete, obsolete to short. Flowers 5- or 6-merous; calyx lobes es- sentially free, membranous to coriaceous, asym- metric, widely ovate, oblate to ovate, subapically notched, basally auriculate: corolla membranous to coriaceous, the lobes narrowly ovate, lanceolate, or oblong, inconspicuously to conspicuously or prom- inently punctate and/or punctate-lineate; stamens connate, the filaments connate basally into an elo- bate tube, free from the corolla tube, epunctate, glabrous, the anthers ovoid to lanceoloid, basally lobate or cordate, dehiscent by longitudinal slits, the connective punctate; pistil glabrous, the ovary ovoid, the style slender, erect, epunctate or punc- tate and/or punctate-lineate, the ovules pluriseriate. Fruits globose to depressed globose, inconspicu- ously or conspicuously to prominently punctate. Distribution. Forty-seven species from Belize and Guatemala, southward through Mesoamerica to the Chocó Floristic Province of Panama, Colombia, and western Ecuador, from sea level to 2200 m in elevation. Ecology. The majority of the species of Ardisia subg. Auriculardisia sect. Palmanae are very lo- calized, occurring in montane and cloud forest hab- itats, occasionally in premontane pluvial forest and rarely in tall, lowland wet forest. Ardisia subg. Auriculardisia sect. Palmanae is defined by the terminal, sessile to subsessile inflo- rescences. In the absence of phylogenetic studies, we cannot be sure if this character arose more than once, so we cannot be sure about the monophyly of the section. Volume 90, Number 2 Ricketson & Pipoly 223 2003 Revision of Ardisia subg. Auriculardisia KEY TO THE ТАХА OF ARDISIA SUBG. AURICULARDISIA SECT. PALMANAE la. pa leaf surface with a dense экз of pm and furfuraceous-lepidote scales Inflorescences pendent; pedicels 6— m long; corolla tube sparsely furfuraceous- -lepidote abaxially; - anthem 44.1 mm long; styles 9.1 #7 ree mm long 1. Ardisia lundelliana Inflorescences erect; pedicels 0—4 mm long; c кы tube glabrous abaxially; enter : 6-2. 6 mm N c styles 2.1—5.2 mm lon 3a. e hlets Чыны; leaf blades elliptic to oblanceolate, calyx lobes 0.9-1.4 mm wide; corolla lobes 1. 4a. a without longitudinal ridges, with large petiole scars; inflorescences ba liene bipinnately paniculate; pedicels obsolete to 1.2 mm long; calyx lobes -— oadly rounded, 0.9-1.2 mm wide; corolla lobes 1.9-2.1 X 0.8-1 mm; anthers 16 "f ые 31. “Ardisia conglomerata Branchlets angulate, with longitudinal ridges, the petiole scars small, inconspicuous; inflo- rescences pyramidal, tripinnately paniculate; pedicels 2—4 mm long; calyx lobes acute, 1.3- mm wide; corolla lobes 3.5-3.7 X 1.9-2.2 mm; anthers 2.1-2.4 X 0.6-0.8 mm 34. Ardisia crassiramea olla 4b. Branchlets flexuous; leaf blades oblong to narrowly elliptic; calyx lobes 1.4—1.9 mm wide; coro 3b. lobes 3.7—4.3 mm long. 5a. Calyx lobes 1.7-1.9 mm wide; corolla lobes oblong, apically acuminate, 1-1.2 mm wide; stamens 3.2-3.6 mm ^ anthers 2.4-2.6 X 1-1.2 mm; styles 2.1-2.3 mm long -------------- Wr, CUP ERRORES сез eee 52. Ardisia mcphersonii 5b. Calyx lobes 1.4—1.7 mm wide; corolla lobes ovate, apically acute, 2.3-2.5 mm wide; stamens 5.8—6 mm long; anthers 2-2. 1.1 mm; styles 4.5-5.2 mm long 43. Ardisia i ipia lb. Abaxial leaf йй e with only one type of scale, either of densely cupuliform or furfuraceous-lepidote scale or oc т with glandular stipitate-stellate hairs. x lobes 1.7 mm long or longer. EN Petioles slender, 1-3 mm diam. alyx lobes as wide as long or wider than long, deeply notched below the apex. . Pedicels 4.5 mm or longer. 10a. Corolla lobes 6.1—6.4 mm long; anthers 3.7—4.3 mm long. ranchlets with furfuraceous-lepidote sc EM leaf ee coriaceous: ins lobes oblate, 2 m; corolla m wide; anther 7-3.8 X 1. 2- 1.4 mm; styles 5.1-5.3 mm pes rdis ia erassipedicellaia 11b. y iis with cupuliform scales and stalked «ате blades chartaceous; calyx lobes orbicular, - A] 9» amie lobes 2.5-2.7 mm ad anthers 4.2-4.3 X 1-1.1 mm; styles 50. jena mes m lon not prominent, the avid ve- 5.9 r 10b. Corolla lobes 3.1—5.2 mm long; anthers 1. iex mm . Abaxial leaf punctations inconspicuot nation flat; calyx lobes 1.7-2 mm gd corolla lobes 3. 1-4 Pd unguiensis n pro- 12b. Abaxial leaf punctations prominently raised, the ЕР rey minulous; calyx lobes 2-2.9 mm long; corolla lobes 4.5-5.2 X : 3a. Furfuraceous-lepidote scales on the branchlets, leaves E^ inflores- cence ferrugineous; secondary leaf venation prominulous above and below; calyx lobes ovate to orbicular, apically acute to obtuse, 2.3— mm wide; anthers 2.6-2.8 X 3m 13b. Furfuraceous-lepidote scales on the branc ШУ leaves and in cence rufous; secondary leaf venation impressed above, тооно raised below; calyx lobes oblate, apically broadly rounded to truncate, —3 mm wide; anthers 3-3.4 X T EHI o saca E O rdisia didit oir a a flor 2-2.5 mm "M corolla 1.5-1.7 Ardisia croatit | 14а. Petioles 4—13 mm long; calyx lobes lobes 3.4—3.6 mm wide; anthers 3—3.1 X - 36. piya croatii ruben. croatii 14b. Petioles 15-24 mm long; calyx lobes 2.6-2. ps corolla lobes yc: Mos 3.3-3.4 Ad du "ES 35. porns croatii perm correae 9b. Pedicels 4.5 mm long or shorter. 15a. Styles less than 5 mm lo 16a. Branchlets Серан: P blades coriaceous; calyx lobes ovate; anther 1 mm wide; styles 3-3.5 mm long 22—525. 59. е "ulii Branchlets terete; leaf blades е» or membranous; calyx lobes su orbicular to orbicular; anthers 1.1-2. ide; styles 4—4.7 mm i 17a. Leaf blades chartaceous; calyx coriaceous; d chartaceous, the 16b. 224 Annals of the Missouri Botanical Garden 6b. 7b. tube glabrous outside; anthers 1.9-2.2 mm long; styles 44.2 1 long; plants of Ecuador |... 25. Fired awarum 17b. Leaf blades membranous; calyx chartac eous; corolla diis the tube sparsely furfurac Ede outside; anthers 2.2-2.7 mm long; styles 4.4—4.7 mm long; plants of Costa Rica - ی یک ج‎ нн 39. iis Шодада ec 15b. Styles 5 mm long or fidei l8a. Ci alyx lobes 3.5-3.7 mm wide; саа lobes 5.3-5.6 mm long; anthers 3.6— 3.8 mm long; styles 6.9-7.1 mm long |... 4T. тар. edid 18b. Gales lobes 1.8—3 mm us bris dis 4.5 5—5.2 mm long; anthers 2 3.2 mm long: styles 5—5.7 mm lor 19а. Leaf blades 5.2-12.6 zy. L3 2 cm; E ES 0—1.2 mm long; corolla lobes 1.8-2 mm wide; anthers 2.2-2.4 X 0.9-1 mm — —— ÁO EROR PR 26. Ardisia enh anaes 19b. Leaf blades 12.6-34.8 X 4-9, 7 em; pedicel | Fa .5 mm long; co- rolla lobes 2.3-3 mm wide; anthers 2.6-3.2 х 1.1-1.5 mm. 20a. Trees 3-25 m tall; blades inconspicuously punctate and punc- tate-lineate on the upper surface; flowers white or light yellow; calyx " in coriaceous, 2.2—3 mm wide; anthers 2.6-2.9 х 1.3- 1.5 т . Ardisia fimbrillifera 20b. Trees 30 40 m tall; blades dom ed punctate и punctate lineate on the upper surface; flow di to light pink; cas lobes chartaceous, 1.8-2 mm Mid anth s 3. 1-3. гк 1.1-1.2 Ф № : н. Adai pseudoracemiflora 8b. Calyx lobes longer than wide, rarely deeply notched below ud apex. 21а. Calyx lobes 3-5 mm long 22a. Calyx lobes 4.8-5 х < 3.94.1 mm: styles 5.9-6.4 mm long 37. Ardisia da- rienensis 22b. Calyx lobes 3-3.7 X 2.6-3.4 mm; styles 5—5.6 mm long 23a l eaf jee 16.4—32.8 X 4.4—9.8 cm: inflorescence 8. 2-32.4 X 4.8— 1 ; flowers mostly 5- merous, rarely 6-merous; calyx ge 3- 3.4 ы 2 6-2. 2 mm; corolla lobes 5.8-6 X 2.8-3.2 mm; anthers 3.: 3.5 X 1.3-1.5 mm; sty T 5-5.2 mm long |... 27. Ardisia cnm 23b. Leaf blades 4.8-11.5 X 1. ТА, 2 cm; uS 4.8-7.2 X 345 em; flowers all 5-merous; calyx lobes 3.4—3.7 X 3.2-3.4 mm; corolla lobes 4 E | X 2.2-2.5 mm; anthers 3-3.1 X 1.2-1.3 mm; styles au DR O PME 44. Ardisia ыла E 21b. Calyx кш 1. 9: mm long. 24a. Leaf blades basally auriculate; petioles subobsolete to 5 mm long mE - 24. Ardisia auriculata 24b. Leaf blades basally obtuse, acute or cuneate; нен 4-19 mm long. 25a. Calyx lobes 2.7-2.9 X 2.4-2.7 mm: fruits 4—5 mm diam. MEO 22. Ardisia angucianensis 25b. Calyx lobes l. 1-2. 4 X 1.2- 2 mm; fruits 5-9 mm diam. sees " Е Es = 54. Ardisia башна Petioles thick, more than 3 mm i dia 26a. Calyx lobes 2.6-2.8 mm aes onis lobes 5.4—5.5 mm long: anthers TS E 1-3.3 mm long; styles 5 5.4-5.6 mm long |... 3. Ardisia megistophylla 26b. Calvi lobes 2.1-2.6 mm long; corolla lobes 4—4.5 mm n long; anthers 2. 3- 3 mm long; styles 2-5.1 mm long. 27a. Leaf blades longer ц 51 em long; branchlets 15-20 mm Фіат. ; pu les 4-6.3 cm long: side lobes 1.5-1.7 mm wide Md er Ardisia aguirreana 27b. Leaf blades shorter than 51 em long; branchlets 5-10.5 mm diam.; un les nearly sessile to 3.5 em long: corolla Жир 1.9-2.6 mm wide. 28a. Calyx lobes ovate, 2.5-2.6 X 1.8-1.9 mm; anthers 2.3-2.6 X 1.3-1.5 mm; styles 2-2.2 mm long; fruits 8-9.8 mm diam. 0 29. pop DE 28b. Calyx lobes ered to oblate, 2.1-2.3 X 2.1-3.1 mm; anthers 2.8-3 X 0.9- 1.1 mm; styles 4.6-5.1 mm long; fruits 4.5—6 diam. 29a. Leaf тн coriaceous, rufous furfurac s above and below; Pe t- ioles 4—5 mm diam.; calyx lobes 2.8-3.1 mm wide: йк" 2 4.4—4.5 2.2-2.3 mm; styles 5—5.1 mm long Айша crassipes 29b. Leaf blades membranous, св above Sud below; n 3—4 mm diam.; calyx lobes 2.1-2.4 mm wide: corolla lobes 44.2 .0-2.7 mm; styles 4.04.7 mm long 0 n 28. pis cartagoana Calyx lobes less than 1.7 mm long. 30a. Pedicels 3.5 mm long or longer. Volume 90, Number 2 Ricketson & Pipoly 225 2003 Revision of Ardisia subg. Auriculardisia 3la. Branchlets 8-10 mm in diam.; leaf sii 45.8—46.6 X 15.7-21.5 cm; ys ences 32 x 25-28 ст; calyx lobes 3.3-3.6 mm wide -........... rdisia ean 31b. Branchlets 1-6 mm diam.; leaf blades 2- 16.6 X 0.6—5.7 cm; inflorescences pi 24 x 2-19 cm; calyx lobes 0.7-3.2 mm wide. 32a. Corolla Tm 1.2-1.8 mm wide; anthers 1—1.9 mm long. 33a. Calyx lobes 1-1.4 mm long. 34a. Branchlets with ar aig a ridges forming up to 5 angles, 3-5 mm diam.; calyx lobes 1.2-1.4 mm long; corolla lobes 2.8-3 X 1.6-1.7 mm; anthers 0.7-0.9 mm wide; gle 3-3.1 mm long; fruits 3-5 mm diam. ----------------- n—————— E OC Ero ME UAM 61. Аа tarariae 34b. Branchlets terete or angled, but without interpetiolar ridges, 1—3(—3.5) mm diam.: calyx lobes 0.9-1.2 mm mg corolla lobes 2.3-2.6 X 1.2-1.5 mm; anthers 0.5-0.7 mm wide; styles 1.3-2.9 mm long; fruits 4.3-8 mm diam. 35a. Branchlets angled, sparsely and minutely ferrugineous furfuraceous- lepidote; calyx lobes 1-1.3 mm wide; corolla lobes 2.4-2.6 X 1.2- 1.4 mm; viai 1.8-1.9 mm long; styles 1.3-1.5 mm long; fruits 7— @ mim diam: не шин eine 62. Ardisia ipsa aulis 35b. Branc pep terete, densely M and furfuraceous-lepidote; lyx lobes 0.7-0.8 mm wide; corolla lobes 2.3-2.4 X 1.4-1.5 mm; ane rs 1-1.1 mm long: ча 's 2.6-2.9 тт add fruits 2 34.1 Айа tenuis dia PERPE ы ы ы са 33b. Calyx lobes 1. LAT 7 mm long. 36a. € es s 3-6 mm diam.; petioles 1.1-2.4 mm long; calyx lobes 1.6— 30. Ardisia coloradouna 36b. Brane т 1-3 mm diam.; pe stioles 4-7 mm m long; calyx lobes 0.8—1 n wide. 37a. Branchlets 2-3 mm diam.; pedicels 3.6—4.8 mm esl corolla lobes : 3.] X 1.7-1.8 mm: anthers 1.6-1.7 X 0.6-0.7 mm; styles 3.1— m long ызы с ы oe dL. din eucuneata 37b. аги hlets 1-2 mm diam.; pe edicels 7.2-16.3 mm long: enel lob .7-1.8 X 1.2-1.3 mm: anthers 1.8-1.9 X 0.8-0.9 mm; styles 1.6 ].9- mm long e TR E 23. Ardisia atropurpurea 32b. Corolla won .9-3.3 mm wide: anthers 2.2-4.2 mm long. 38a. Branchlets 1-3.5 mm jum medic els 12— meus mm lone Em 56. Ardisia panamensis 38b. Branc thlets (2—)4—5 mm dim 2 pedicels 4—9 mm long. Leaf blades coriaceous; pedicels 4—5 mm long; calyx lobes coriaceous, oblate, 2 Ll 2 mm "ide; apically rounded; corolla lobes chartaceous, 5— 3.3 mm: mpi s 3.2-4.2 X 1.2-1.4 mm; styles + к mm long: hg 9. 98 Vari Mam. eae en aa емее . Ardisia dwyeri 39b. Leaf Madeo membranous; - pedicels 6-9 mm long; calyx beds me ar a us o chartaceous, 1.1-1.2 mm wide, apically acute; corolla lobes membra- nous, 3.3-3.4 X 1.9-2 mm; anthers 2.3-2.4 X 0.8-0.9 mm; styles 3.4— 3.5 mm long; fruits 4-5 mm diam. s 66. Ardisia vesca 30b. с els 2.5 mm long or shorter. . Calyx lobes as long as or much longer than wide. Ala. Leaf blades more than 18 cm Ara ‘blades 24.5-51 х 9.2-15.4 em; calyx lobes 1.4—1.6 X 1.3-1.5 mm; corolla lobes 3.8—4 X 1.8-2 mm; anthers 2.3-2.4 X 1-1.1 mm; styles 3. ja 9 mm lon O ES EEE LEAN S NS SUB R 9. Ardisia knappii 42b. Leat blades 18. 5-20 X 6.7-7.9 cm; calyx е 1.2-1.3 X 0. ES mm; le qam pie —].8 X 1-1.2 mm: anthers 1.1—1. ).6—0.7 mm; styles 1.2 m lor m — À NONE. A Ardisia E Alb. en blades lese than 15 em long. s 3-5 mm diam.; calyx lobes 0.9-1 mm wide; corolla lobes 3—3.2 X 2 mm: anthers 1.7-1.8 mm long: styles 3.1-3.2 mm lo ng кай nearer E ые ез bmp) Aree a. 43b. Branc pe 5-8 mm diam.; calyx lobes 1.4—1.6 mm wide; Salle Lobos 3 X 1.6-1.9 mm; anthers 2.2-2.4 mm Е styles 1.9-2 mm long ... 64. joris tysonit 40b. Calyx lobes wider ay long. 44a. Leaf blades coriaceous 45a. Branchlets Бо оа, checking and exfoliating: leaf blades 23-23.5 X c em; petioles stout, marginale, 1-1.3 mm long; calyx lobes membranous, orbic to oblate, 0.8-1 х 0.6-1.1 mm oe 60. Ardisia pne NS 45b. Branchlets smooth; leaf blades 11. 3-14 L9 x 2.9.4. 4 cm; apes slender, can- aliculate, 3-6 mm longs calyx lobes chartaceous, ovate, 1.3-1.6 X 1.3-1.8 mm . 48. risa hugonensis 44b. Leaf blades ое уон to M NUR ous. 226 Annals of the Missouri Botanical Garden 46a. Calyx lobes 0.9—1.1 X 1.1—1.3 mm; styles 1-3 mm long Calya lobes 1.1-1.5 X 1.4-2 mm; 3 46b. 1.4-1.5 47a. Calyx lobes 2.9 mm long; styles 5 . Calyx lobes 1.1-1.4 2.5 mm long; styles حي ]~ zi‏ = 19. Ardisia aguirreana Pipoly, Caldasia 17: 419. 1995. TYPE: Colombia. Chocó: area of Baudó, on left bank of river Baudó, about 1.5 km up- stream of estuary, practically opposite the E- most houses of Puerto Pizarro, about 20 m in- land from the shoreline of the bay in front of Estero del medio, ca. 5 m above shoreline, [0— 100 m], 11 Feb. 1967 (fl, fr), H. Fuchs & L. diu 21851 (holotype, COL; isotypes, F!. US! [2]) For illustration, see Pipoly (1995: 419, fig. 1). Small shrubs ca. 3 m tall. Branchlets stout, terete, 15-20 mm diam., raceous-lepidote. Leaves with blades chartaceous, oblanceolate, 51.5—67 X 12-17.5 em, apically acu- minate, with an acumen 0.5-1.2 cm long, basally densely appressed rufous furfu- cuneate, decurrent on the petiole, inconspicuously punctate and punctate-lineate, glabrous above, densely furfuraceous-lepidote below, the midrib im- pressed above, prominently raised below, the sec- ondary veins 75 to 85 pairs, nitid above, promi- nently raised below, the margins entire, revolute: petiole stout, 3—5.5 em long, 5-8 mm diam., gla- brous above, furfuraceous-lepidote below. /nflores- cences erect, tripinnately paniculate, 12-2 12 cm, pyramidal, shorter than the leaves, rachis, branches, abaxial bract surfaces, and ped- the icels furfuraceous-lepidote, the branches loosely congested into 6- to 11-flowered corymbs; pedun- cles 4—6.3 ст long; inflorescence bracts unknown: inflorescence branch bracts caducous, membra- nous, ovate to lanceolate, 1.5—4.2 X 0.9-1.5 mm, apically acute, conspicuously punctate and punc- tate-lineate, glabrous above, furfuraceous-lepidote below, the margins entire, sparse glandular cilio- late; floral bracts similar branch bracts, but 2-2.5 X stout, subobsolete to 2.5 mm long, conspicuously to the inflorescence 0.2-0.3 mm; pedicels punctate and punctate-lineate, furfuraceous-lepi- dote. Flowers 5-merous, light violet gray: calyx lobes chartaceous, ovate, 2.1-2.6 X 1.2-1.4 mm, apically acute, conspicuously punctate and punc- tate-lineate, glabrous adaxially, furfuraceous-lepi- dote abaxially, the margins entire, erose, hyaline, sparse glandular ciliolate; corolla chartaceous, 6— 6.2 mm long, the tube 1.9-2.2 mm long, the lobes narrowly lanceolate, 4—4.1 X 1.5-1.7 acule, conspicuously punctate and punctate-lin- mm, apically 38. Ardisia dukei styles 3 3.5-5.4 mm long. 5 mm Png s lobes 4.1—4.4 mm we ne 2.8- .3-5.4 mm long |... 21. sia d mm long; ios lobes 2.7-3.7 mm ong pi o 3.5-3.6 mm long . Ardisia xum eate, glabrous throughout, the margin entire, hya- line; stamen 4.9—5 mm long; the filament 2.9-3 mm long. the staminal tube 1.1—1.4 mm long, the api- cally free portions 1.6-1.8 mm long, the anthers lanceoloid, 2.4—2.6 X 0.8-1 mm, apically broadly apiculate, basally lobate, the connective conspicu- ously punctate; pistil 4.8—5 mm long, glabrous, the ovary ovoid, 1—1.1 mm long, styles 3.7—4 mm long, conspicuously punctate, the ovules 24 to 30. Fruits (immature) globose, 5—7 mm diam., conspicuously punctate and punctate-lineate, inconspicuously costate. Distribution. Ardisia aguirreana is found in Chocó in the area of Baudó and the Parque Nacio- nal Natural de Utria, Colombia, growing from 5 to 100 m in elevation. Ecology and conservation status. Ardisia agui- rreana occurs in secondary vegetation on relatively wet, slightly swampy ground. Because of its re- stricted distribution, it should be considered threat- ened. Etymology. of Jaime Aguirre, ex-director of the Instituto de Museo de Historia Natural, This species was named in honor Ciencias Naturales, Universidad Nacional de Colombia. Common Name. Arrayán (Fuchs & Zanella 8. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia aguirreana is closely related to A. megistophylla, A. cogolloi, A. crassipes, and A. car- tagoana because of its long calyx lobes with thick petioles. Ardisia aguirreana is separated from А, megistophylla by its shorter calyx lobes to 2.6 mm long, corolla lobes to 4.1 mm long. anthers to 2.6 mm long and styles to 4 mm long. Ardisia agui- rreana is easily distinguished from the other related taxa by its longer leaf blades to 67 ст long, larger branchlets to 20 mm in diameter, longer peduncles to 6.3 cm long, and narrower corolla lobes to 1.7 mm wide. COLOMBIA. Chocó: Mpio. de Utría, along Quebra- Espina et al. 3751 (€ HOC О, МО); Parque Nacional de Utría, SE of the En- senada de Utría, entering through Charco de las Ballenas, 2 June 1990 (fl), К. García C. & E. Agualimpia 360 (CHO- CO, FTG, MO). Specimens examined. Quibdó, Parque Nacic mal Natural de L 0 — 2 — еј = 20. Ardisia albisepala (Lundell) Pipoly & Ric- Volume 90, Number 2 2003 Ricketson & Pipoly 227 Revision of Ardisia subg. Auriculardisia ketson, Sida 18: 511. 1998. Auriculardisia al- bisepala Lundell, Wrightia 7: 266. 1984. Ar- disia buena (Lundell) Lundell, Phytologia TYPE: Panama. Ve- raguas: trail on ridge to summit of Cerro Tute, Cordillera Tute, 1 km past Escuela Agrícola Altos de Piedras, W of Santa Fé, 08°30'N, 081?06'W, 950-1250 m, 15 Dec. 1981 (fl). 5 Knapp & K. Sytsma 2548 (holotype, LL: iso- types, MO!, NY!). Figure 22. 1986, nom. inval. Trees 6-8 m tall. Branchlets slender, terete, 3—5 mm diam., densely cupuliform lepidote. Leaves with x 2.2-3.9 em, apically long acuminate, with an acumen 6-14 blades membranous, elliptic, 7.2-10.7 mm long, basally obtuse, decurrent on the petiole, prominently punctate above and below, glabrous above, furfuraceous-lepidote below except mixed with cupuliform lepidote scales along midrib. the midrib impressed above, prominently raised below. the secondary veins 45 to 51 pairs, prominulous above and below, the margins entire, flat to slightly inrolled; petioles slender, canaliculate, 8-13 mm long, glabrous above, mixed lepidote below. /nflo- rescences erect, bi- to tripinnately paniculate, 5.2— 13.7 X 4.5-11.7 leaves, cupuliform lepidote, the branches loosely em, pyramidal, longer than the congested into 3- to 7-flowered corymbs; peduncles obsolete to 1.4—1.7 inflorescence branch bracts unknown: cm long; inflorescence bracts unknown; floral bracts caducous, membranous, lanceolate. 1.2-1.4 X 0.4—0.5 mm, apically acute, prominently punctate, glabrous adaxially, furfuraceous-lepidote. the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; pedicels stout, obsolete to 1.5 mm long, prominently punctate. furfura- ceous-lepidote. Flowers 5-merous, light pink: calyx lobes membranous, ovate, 1.2-1.3 X 0.9-1 mm, prominently punctate and punctate-lineate, gla- brous adaxially, scattered furfuraceous-lepidote abaxially, the margins irregular, minutely erose, hy- aline, sparsely glandular ciliolate; corolla membra- nous, 4.4—4.5 mm long, the tube 1.2-1.5 mm long. the lobes lanceolate, 3—3.2 X 2-2.2 mm. apically acute, prominently punctate and punctate-lineate, glabrous throughout, the margins entire, hyaline: stamens 3.1-3.2 mm long, the filaments 1.9-2.1 mm long, the staminal tube 0.6-0.7 mm long, the apically free portions 1.3-1.4 mm long, the anthers narrowly ovoid to lanceoloid, 1.7—1.8 х 0.8—0.9 mm, apically apiculate, basally lobate, the connec- tive conspicuously punctate; pistil 4—4.2 mm long, glabrous, ovary oblongoid, 1-1.1 mm long, the style 3.1-3.2 mm long, inconspicuously punctate, the ovules 18 to 20. Fruits unknown. Distribution. Ardisia albisepala is endemic to Cerro Tute in е Panama, growing from 950 to 1250 m in elevation Ecology and conservation status. Ardisia albi- sepala occurs in lower montane rain forests and cloud forests. Because of its restricted distribution, it should be considered threatened. Etymology. The specific epithet was derived from the Latin “albus” meaning a dull white, and “sepala” referring to the light-colored calyx lobes. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia albisepala is most closely related to A. tysonii because of its short calyx lobes that are as long as or much longer than wide, short pedicels, leaf blades less than 15 em long and less than 6.5 cm wide, and long petioles. However, A. albisepala can easily be separated from A. tysonii by its thin- ner branchlets to 5 mm in diameter, narrower calyx lobes to 1 mm wide, longer and wider corolla lobes to 3.2 X 2.2 mm, shorter anthers to 1.8 mm long, and longer styles to 3.2 mm long. PANAMA. Veraguas: “Cerro ridge x from former Escuela Agrícola, Santa Fé, A ^ р. 1983 (fl), C. Hamilton & R. Dressler 3109 (LL. MO es imens examined. 21. Ardisia anchicayana Ricketson & Pipoly. sp. nov. TYPE: Colombia. Valle del Cauca: Alto Yunda, Río Anchicayá, 1000 m, July 1972 (fl), S. Hilty Jy-28 (holotype, US!: iso- type, ARIZ!). Figure 23. Ob lobulos calycinos parvos et longiores quam latiores atque foliorum laminas T eas pedicellos ties va sed ab ea lobulis caly- n longis, каша e 14 im longis, anthe longioribus usque ad 2.9 mm denique stylis k feriens usque ad 5.4 mm longis perfacile statimque distinguitur. Trees to 15 m tall. Branchlets stout, terete, 5—6 mm diam., bark exfoliating, furfuraceous-lepidote. Leaves with blades membranous to chartaceous, el- liptic to oblong, 20-36 6.9-9.2 cm, apically acute, with an acumen unknown, basally acute, de- current on the petiole, inconspicuously punctate and punctate-lineate, glabrous above, furfuraceous- lepidote, the midrib impressed above, prominently raised below, the secondary veins 75 to 85 pairs, prominulous above and below, the margins entire, inrolled; petioles stout, canaliculate, 3—7 mm long. glabrous above, furfuraceous-lepidote below. /nflo- rescences likely pendent, tripinnately paniculate, 23.5-27.5 X 10-15 em, pyramidal, shorter than densely furfuraceous-lepidote. the the leaves. branches loosely congested into 5- to 7-flowered corymbs; peduncles obsolete, the lower branch sub- Annals of the 228 Missouri Botanical Garden CSA) 82-47 pg `$ ‘dojoq шолу umeıp g *y) 1940[4 *— “Yoursg #шләмо]] *y— ‘рирбротуәир pisipiy (nsu) ес әл ((OW) #РС& рш ^N у (риу `ç 79d&jost шолу umeap g ^y) типи] *g— “Yoursq зицәмо "ү “pypdasiqyy visipay рәр ze әлү Volume 90, Number 2 2003 Ricketson & Pipol 229 y Revision of Ardisia subg. Auriculardisia tended by leaves; inflorescence bracts unknown; in- florescence branch bracts caducous, membranous, ovate to oblong, 1.5-2 X 0.4-0.6 mm, apically acute, prominently punctate and punctate-lineate, glabrous above, furfuraceous-lepidote below, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflo- rescence branch bracts, 1.2-1.7 0.4—0.6 mm; pedicels stout, 1—1.5 mm long, inconspicu- ously punctate and punctate-lineate, furfuraceous- 2-merous, X 1.8-2 mm, apically but lepidote. Flowers white; calyx lobes membranous, ovate, 1.4—1.5 acute to rounded, prominently punctate and punc- tate-lineate, glabrous adaxially, sparsely furfura- ceous-lepidote, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 5.4—5.6 mm long, the tube 1.2-1.5 mm long, the lobes narrowly ovate, 4.1—4.4 X 2.4— 2.5 mm, punctate-lineate, glabrous throughout, the margins mm long, the fil apically acute, prominently punctate and entire, hyaline; stamens 5. aments 2.6-2.8 mm long, ilie staminal tube 0.5— 0.7 mm long, the apically free portions 1.9-2.3 mm long, the anthers narrowly ovoid, 2.8-2.9 X 1.1- 1.2 mm, apically apiculate, basally lobate, the con- nective conspicuously punctate; pistil 6.4—6.7 mm long, glabrous, the ovary oblongoid, 1.1-1.3 mm long, the style 5.3-5.4 mm long, epunctate, the ovules 18 to 24. Fruits unknown. Distribution. only from the type collection from Valle del Cauca, Colombia, growing at 1000 m in elevation. Ardisia anchi- Ardisia anchicayana is known Ecology and conservation status. cayana occurs in premontane pluvial forest, on the western slopes of the western Andean cordillera. This area of the Chocó Floristic Province is known for annual precipitation of over 8000 mm, and is also known for very large individuals of commer- cially valuable timber species. Therefore, all the forests of the area are under serious threat. Etymology. The specific epithet comes from the type collection эш the Río Anchicayá in Valle de Cauca, Colomb Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia anchicayana is most similar to A. palmana because of the calyx lobes wider than long. membranous to chartaceous leaf blades, and short pedicels. However, Ardisia anchicayana may be separated from A. palmana by the longer calyx lobes to 1.5 mm long, longer corolla lobes to 4.4 mm long, longer anthers to 2.9 mm long, and longer styles to 5.4 mm long. 22. Ardisia angucianensis Ricketson & Pipoly, sp. nov. TYPE: Costa Rica. Puntarenas: Can- ton de Osa, Fila Costefia, Fila Cruces, head- waters of Río Piedras Blancas, Cerro Angucia- na, western slopes, 08°48' 567%, 083^10'37"W, 1400-1600 m, 10 Dec. 1993 (fr), B. Hammel 19280 (holotype, MO!; isotype, INB not seen). Figure 24 Propter Tes. calycinos longiores quam latiores ovatos usque ad 2 m longos atque folia ad apices acuminata ad bases ac uta ЁЗ nigropunctatae arcte similis, sed ab еа lobulis calycinis 2.7-2.9 (non 1.7-2.4) mm longis, 2.4— 7 (nec 1.2-2.0) mm latis, оя 4—5 (поп 5—9) mm diametro statim separabilis. SM Trees to 5 m tall. Branchlets slender, terete, 4—5 mm diam., densely and minutely appressed rufous furfuraceous-lepidote, often glabrate with age. Leaves with blades membranous to chartaceous, el- ірис to narrowly elliptic, 28.8-33.3 X 9.9-11.6 cm, apically acuminate, with an acumen 10-12 mm long, basally acute, decurrent on the petiole, prom- inently raised below, prominently punctate and punctate-lineate above and below, glabrous above, densely and minutely appressed rufous furfura- ceous-lepidote, the midrib impressed above, prom- inently raised below, the secondary veins 42 to pairs, prominulous above, the margins entire, flat; petioles slender, canaliculate, 11—13 mm long, 2— 3 mm diam., glabrous above, sparsely furfuraceous- lepidote below. Inflorescences erect, bipinnately pa- niculate, 29. 17.5-18.2 cm, pyramidal, longer than the leaves, the rachis, branchlets, and pedicels densely and minutely appressed rufous furfuraceous-lepidote, the branches loosely con- gested into 5- to 7-flowered corymbs; peduncles 2.7-2.9 cm long; inflorescence bracts and branch bracts unknown; floral bracts unknown; pedicels stout, 2-2.8 mm long, inconspicuously punctate and punctate-lineate. Flowers 5-merous, calyx lobes chartaceous to coriaceous, suborbicular to or- bicular, 2.7-2.9 X 2. mm, apically acute to rounded, prominently punctate and punctate-linea- te, glabrous adaxially, furfuraceous-lepidote aba- xially, the margins irregular, minutely erose, hya- line, sparsely glandular ciliolate; corolla, stamens, and pistil unknown. Fruits (immature), globose, 4— 5 mm diam., prominently punctate and punctate- lineate, glabrous. Distribution. Ardisia angucianensis is known only from the type collection in Puntarenas, Costa Rica, growing from 1400 to 1600 m in elevation. cology and conservation status. Ardisia an- gucianensis occurs in premontane wet forest on limestone. Because it is only known from the type, it should be considered threatened. 230 Annals of the Missouri Botanical Garden NN S б AN LE fi ys, » Wer PER Lg SHS о ау 4 pu AN 2272 ye UY SS ae кесу \ P NS Ma y =, Ardisia angucianensis. —A. Flowering branch. —B. Fruit. (A. B drawn from holotype. B. Hammel 19280 (MO).) Figure 24 (left). C. Fruit. (4, B drawn from isotype, J. Folsom et al. 4998 (MO); С from J. Folsom et al. 6617 Ardisia atropurpurea. —A. Flowering branch. —B. Flower. Figure 25 (right). (MO).) Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia Etymology. The specific epithet comes from the location where the type was collected on Cerro An- guciana in the Cantón de Osa, Puntarenas, Costa Rica. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia angucianensis is most similar to A. nigropunctata due to its calyx lobes wider than long, to 2.9 mm, acuminate leaf apices, and acute leaf bases. However, A. angucianensis can be sep- arated by the longer and wider calyx lobes to 2.9 х 2.7 mm and smaller fruits to 4.5 mm in diameter. 23. Ardisia atropurpurea Lundell, Phytologia 48: 134. 1. Auriculardisia atropurpurea (Lundell) Lundell, Phytologia 49: 342. 1981 TYPE: Panama. Panamá: from Tortf to the Pi- lota del Toro, the mountain overlooking Tortí Arriba, 400—700 m, 27 Aug. 1977 (fl). J. Fol- som, G. Alonzo de Monte & relatives 4998 (ho- lotype, LL!; isotypes, FTG!, MO!). Figure 25. Shrubs 1—4 m tall. Branchlets slender, terete, 1— 2 mm diam., densely furfuraceous-lepidote. Leaves with blades membranous, elliptic, 2.8-7.1 х 0.8- 2.6 cm, apically acuminate, with an acumen 6-9 mm long. basally acute, decurrent on the petiole, inconspicuously punctate and punctate-lineate, gla- brous above, densely furfuraceous-lepidote below, the midrib impressed above, prominently raised be- low, the secondary veins 12 to 20 pairs, obscure to prominulous above and below, the margins entire, flat; petioles slender, canaliculate, 4-6 mm long, glabrous above, densely furfuraceous-lepidote be- low. Inflorescences erect, pinnately to bipinnately paniculate, 5— 2-7 cm, pyramidal, longer than the leaves, densely furfuraceous-lepidote, the branches loosely congested into 5- to 7-flowered corymbs; peduncle 1—4 mm long, the lower branch- es subtended by leaves; inflorescence bracts un- known; inflorescence branch bracts caducous, membranous, ovate to oblong, 1.3-2.3 X 0.5-0.8 mm, apically acute, prominently punctate and punctate-lineate, glabrous adaxially, mixed cupu- liform and furfuraceous-lepidote, the margins irreg- ular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflorescence branch bracts, but 1-1.3 X 0.3-0.6 mm; pedicels slender, 7.2-16.3 mm long, prominently punctate and punctate-lineate, mixed furfuraceous- and cu- puliform lepidote. Flowers 5-merous, deep purple; calyx lobes chartaceous, ovate, 1.5-1.6 X 0.9- mm, apically acute, prominently punctate and punctate-lineate, glabrous adaxially, furfuraceous- lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla س membranous, 2.5-2.6 mm long, the tube 0.7-0.8 mm long, the lobes lanceolate, 1.7-1.8 X 1.2-1.3 mm, apically acute, prominently punctate and punctate-lineate, glabrous adaxially, furfuraceous- lepidote abaxially, the margins entire, hyaline; sta- mens 2.3-2.4 mm long, the filaments 0.9-1 mm long. the staminal tube 0.2-0.3 mm long, the api- cally free portions 0.6-0.8 mm long, the anthers ovoid to narrowly ovoid, 1.8-1.9 X 0.8-0.9 mm, apically apiculate, basally cordate, the connective conspicuously punctate; pistil 2.6-2.7 mm long. glabrous, the ovary ovoid, 0.9-1 mm long, the style 8 mm long, prominently punctate and punc- tate-lineate, the ovules 12 to 14. Fruits globose, 4— — 7 mm diam., prominently punctate. Distribution. Ardisia atropurpurea is apparently endemic to the area above Tortí Arriba in Panamá, Panama, growing from 400 to 700 m in elevation. Ecology and conservation status. Ardisia atro- purpurea occurs in lowland wet forests. Because it is so poorly known, no conservation status evalua- tion can be made. Etymology. The specific epithet comes from the Latin “atro” meaning dark and “purpurea” meaning purple, and refers to the dark purple flower color. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia atropurpurea is most closely related to A. eucuneata because of the small calyx lobes, long pedicels, narrow corolla lobes, and short an- thers. However, Ardisia atropurpurea is separated from A. eucuneata by its thinner branchlets to 2 mm in diameter, longer pedicels to 16.3 mm long, shorter and narrower corolla lobes to 1.8 1.3 mm, longer and wider anthers to 1.9 X 0.9 mm, and shorter styles to 1.8 mm long. Specimen examined. PANAMA. Panamá: area sur- rounding Rancho Chorro, mountains above Torti Arriba, Canazas mountain chain, 3 Dec. 1977 (fl, fr), J. Folsom et al. 6617 (LL, MO). 24. Ardisia auriculata Donn. Sm., Bot. Gaz. 24: 395. 1897. Auriculardisia auriculata (Donn. Sm.) Lundell, Phytologia 54: 285. 1983. TYPE: Costa Rica. Limón: forests of Suerte [Guppies], Llanura de Santa Clara, 900 ft. [274 m]. Feb. 1896 (fr), J. Donnell Smith 6640 (holotype, US!, US neg. 2367!, LL neg. 1971- 19!). Figure 26. Shrubs or small trees 1—8 m tall, 2-7 cm diam. Branchlets slender, terete, the bark longitudinally ridged, 3-7.5 mm diam., densely and minutely fer- rugineous furfuraceous-lepidote, the scales early caducous. Leaves with blades membranous to char- taceous, elliptic or oblong to oblanceolate to ob- Missouri Botanical Garden Annals of the 232 ((OW) EIS "I? 12 рәтү 7) шоу q ЧОҢ) ESPI ]ojan) 7) X aqy ‘Gq *ad&jo[ou шолу UMPIP э-ү) "ma4 '([— чәмор 77)— "Рп лзәмору *q— "uoueJq SULIQMO] | "ұу unapam pisipay (аи) PZ әлү COD IA ECOOL `P 12 Yunq 4 шолу 4 SN) 0799 umus jjouuoq "f *ad&jo[oq шолу umeıp 7) у V) зит 7— MOJA 'g— “Yoursq Sunmo "V— -p;ppounp mapay (цә) OZ 21n31 у ww ج‎ V 2 m Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 233 ovate, 12.3—46.7 X 3.4—17.7 ст, apically acumi- nate, with an acumen 2-13 mm long. basally auriculate, prominently punctate and punctate-lin- eate, glabrous above, scattered minutely ferrugin- eous furfuraceous-lepidote below, the midrib im- pressed above, prominently raised below, the secondary veins 32 to 48 pairs, prominulous above, slightly to prominently raised below, the margins entire, flat; petioles stout, marginate, subobsolete to 5 mm long. 2-3 mm diam., glabrous above, furfu- raceous-lepidote below. Inflorescences erect, bipin- nate to tripinnately paniculate, 15—40 X 10-30 cm, pyramidal, longer than the leaves, the rachis dense- ly and minutely ferrugineous furfuraceous-lepidote, the branches loosely congested into 3- to 9-flowered corymbs, sparsely furfuraceous-lepidote; peduncle nearly obsolete to 8.6 cm long, densely furfura- ceous-lepidote, the lower branches subtended by leaves; inflorescence bracts unknown; inflorescence branch bracts caducous, membranous, oblong, 10— 18 х 2-4.5 mm, apically acute, inconspicuously punctate and punctate-lineate, glabrous above, fur- furaceous-lepidote below, the margins irregular, mi- nutely erose, hyaline, sparsely glandular ciliolate: floral bracts similar to the inflorescence branch bracts, but 1.1-1.8 X 0.5-0.8 mm; pedicels slen- der, 610.2 mm long, inconspicuously punctate and punctate-lineate, very sparsely furfuraceous-lepi- dote, glabrescent. Flowers 5-merous, greenish, pink, purple; calyx lobes membranous, 2.3-2.7 X acuminate, prominently punctate and punctate-lin- white, ovate, 1.4-1.6 mm, apically acute to eate, glabrous throughout, the margins irregular. minutely erose, hyaline sparsely glandular ciliolate: corolla membranous, 5.6-6.2 mm long, the tube 0.4—1.2 mm long, the lobes ovate, 4.8—5. 3 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the margins entire, hyaline; stamens 4.7—5.1 mm long, the fil- aments 2.2-2.5 mm long, the staminal tube 1—1.4 mm long, the apically free portions 0.9-1.4 mm long, the anthers ovoid, 2.5— 1.3-1.5 mm, apically apiculate-mucronate, pee cordate, the connective prominently punctate; pistil 6.2-6.4 mm long, glabrous, the ovary oblong, 1.3-1.4 mm long, the style 4.6—4.7 mm long, epunctate, the ovules 19 to 21. Fruits globose, 5.5-7 mm diam., promi- nently punctate and punctate-lineate. Distribution. Ardisia auriculata is found on the Atlantic Slope of Nicaragua in Jinotega, Río San Juan, and Zelaya, throughout Costa Rica (except Cartago and Puntarenas), and Panama in Bocas del Toro, Colón, Coclé, Panamá, San Blas, and Vera- guas, growing from sea level to 1200 m in elevation. Ecology and conservation status. Ardisia auri- culata occurs as an understory tree in lowland and premontane wet and pluvial forests. It apparently has some tolerance for disturbance and has a broad habitat range. Therefore, we do not believe it is threatened at this tim Etymology. The specific epithet was derived = rom the Greek, “auricle,” or ear, and refers to the lobed leaf bases. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia auriculata is closely related to both A. angucianensis and A. nigropunctata because of its large calyx lobes to 2.9 mm long, which are longer than wide, and slender petioles. Ardisia au- riculata can easily be distinguished from both A. angucianensis and A. nigropunctata by its auricu- — ate leaf bases with pedicels nearly obsolete to 5 mm long. Specimens examined. NICARAGUA. Jinotega: Bo- caycito, 28 Dec. d (fl), J. Atwood et al. 6903 (MO). Río San Juan: near Cafio Chontalefio, 20 km NE of El Castillo, Río Indio а 7-9 Mar. 1978 (fr), D. Neill 3381 (HNMN, MO); Mpio. Castillo, Reserva Indio-Maíz, Cerro el Diablo, 6 Jan. 1997 (ster.), R. Rueda et al. 5457 (HULE, MO). Zelaya: Atlanta, Сайо el Tigrillo, La Pica- da, 8 Nov. ). A. Laguna vai (MO); Río Puita “La Richard” 200 m SE, 13 Nov (fl), P. Moreno & J. Sandino 12977 (HN NMN, MO). Caño Montecristo, mouth of Caño El Consuelo, 7 Feb. 1982 (fr), P. Moreno 15054 (MO); E of Nueva Atlanta, 20 Feb. 1994 (fr), R. Rueda 3291 (HULE, МО); кро D Bonanza, Re- serva Bosawás, Cerro Cola Blanca, 2 June 1997 (ster. Rueda & I. Coronado 6495 (HULE, MO). COSTA RIC A. Alajuela: San Ramón, 22 Feb. 1931 (fr), A. Brenes gon (F); 3 km NNE of Bijagua, along new road to Upala, 7-t Nov. 1975 (fl), W. Burger & R. Baker 9806 (F, NY); e ribbean slope between San Lorenzo and Los Angeles de San Ramón, above the Río San Lorenzo, 20 Sep. 1978 (fl). W. Burger & T. Antonio 11192 (DUKE, F); Cantón de Gua- tuso, Cordillera Чор 5 km N of Lago е, near Lago Coter, 14 Oct. 4 (fl), A. rc et al. 320 (CR, F, nal cie 'ano, aeu trail S volcano, т of road along шыш and N of Río oa Caliente 1 Oct. T D. V Fi = T — Monteverde, Тум. Сетто Negro, lef bu Río Pei ñas Blancas, 29 Mar. 1987 (fr), W. Haber & E. Bello C. 6848 (MO); Cantón de Upala, Dos Ríos, 5 km S of Brasilia, right bank Río Pizote, 28 Oct. 1987 (fr), G. Herrera 965 CR, MO); La Fortuna, San Carlos, 2.5 km E Cerro Chato, bs rata Río Fortuna, 7 Nov. 1989 (fr), Q. Jiménez & Elizondo 740 (CR, ed ve MO); Cantón de Agua 4 cas, Hacienda La M , Río San Rafael, 8 Feb. (fr), L. Williams et a 29080 (F, MO, NY, US). Guana- : Parque Nacional Guanacaste, Ratna ión Pitilla, La- ‚ C. Moraga 284 (INB, MO). Heredia: Parque Nacional Braulio Carrillo, Estación El Ceibo, 29 Oct. 1989 (fl), " PME & B. Hammel 26 (СК. FTG, MO); property of I . Holdridge, ca. stream on the Rio Puerto xod 5-6 Jan. Burger & E oe U. 4234 (CR, DUKE, F, G, GH). Li- món: near Río Toro Amarillo, Guapiles, 10 Feb. 1965 (fr), R. cuts 66352 (MO, LL neg.). San José: Parque — ~ Y 234 Annals of the Missouri Botanical Garden Nacional Braulio Carrillo, Mere. Trail, Quebrac Gonzalez and tributaries, 5 Oct. fl), J. ое et al. 1855 (CR, FTG, INB, MO). BAM Bocas de Toro: s road to Chiriquf Grande, 26 Oct. 1985 (fl), G. McPherson 7382 (MO). Coclé: Coclecito road, eleva- tional tabem ‘tion from 1 mi. beyond the pos A the Mn top, 12 Jan. 1986 (fr), G. de Nevers et al. 6738 (LL ;olón: walking upstream from bridge over Río , 19 Jan. 1980 (fr G. de Nevers et al. 4076 (MC i Jkup- Neba n. to the falls, 30 Oct. 1991 (fl), H. Hm et al. 7 (FTG, INB, MO). oe vic inity of Escuela nem hal Alto Piedra near Sant: our walk along road beyond school, 1 Dec. 1979 (A). T. panies 3001 (MO). ~ 25. Ardisia awarum Ricketson & Pipoly, sp. nov. TYPE: Ecuador. Esmeraldas: San Lorenzo de Cantón, Reserva Indígena Awá, Parroquia Ri- caurte, comunidad Balsarefio, Río Palabr, 01°09'N, 078°31'W, 100 m, 15-29 Apr. 199] (f, D. Rubio & C. Quelal 1453 3 (holotype, MO; isotypes, FTG!, ОСМЕ not seen). Figure 21 Ob lobos calycinos 'ulares vel orbiculares, at- me laminam non coriaceam, A. ee arcte similis ed ab ea laminis — aua s (non membrana- ceis), calycis coriaceis (non memb ae necnon ant- heris 1.9-2.2 (non 8-1: 3) mm jones praeclare distat. Trees 8-25 m tall, 10—15 cm diam. Branchlets slender, terete, 4.5-6.5 mm diam., densely and mi- nutely appressed rufous furfuraceous-lepidote. Leaves with blades chartaceous, oblong elliptic to oblanceolate, 16.2—26.7 »8 cm, apically acuminate, with an acumen 8-14 mm long, basally acute, decurrent on the petiole, the midrib im- pressed above, prominently raised below, the sec- ondary veins 32 to 38 pairs, inconspicuously punc- tate and punctate-lineate above and below, glabrous above, densely and minutely appressed rufous fur- furaceous-lepidote, inconspicuously raised above and below, the margins entire, inrolled; petioles slender, marginate, 8-14 mm long, 1.5-2.5 mm diam., glabrous above, furfuraceous-lepidote below. Inflorescences erect, tripinnately paniculate, 8—21 10-18 сш, pyramidal, usually longer than the leaves, peduncle, the rachis, branchlets. abaxial bract surfaces and pedicels densely cupuliform and furfuraceous-lepidote, the branches loosely con- gested into 3- to 7-flowered corymbs: peduncle nearly obsolete to 1.7 cm long; inflorescence bracts unknown; inflorescence branch bracts unknown; 1-1.4 prominently punctate and floral bracts membranous, ovate. mm, apically acute, punctate-lineate, glabrous adaxially, furfuraceous- lepidote abaxially, the veins unknown, the margins irregular, minutely erose, hyaline, sparsely glan- dular ciliolate; pedicels stout, 1.5-3 mm long, in- conspicuously punctate and punctate-lineate, sparsely furfuraceous-lepidote. Flowers 5-merous, white; calyx lobes coriaceous, suborbicular, 2-2.5 X 2-2.5 mm, apically acute to rounded, promi- nently punctate and punctate-lineate, glabrous ada- xially, sparsely furfuraceous-lepidote abaxially, ob- solete flat scales with the margins entire or with small teeth, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla char- taceous, 5—6.2 mm long, the tube 1.4-2 mm long, the lobes lanceolate, 3.9-4.5 X 1.7-2.1 mm, api- cally acute, prominently punctate and punctate-lin- eate, glabrous throughout, the margins entire, hy- aline; stamens 3.7—4.2 mm long, the filaments 1.5— 2 mm long, the staminal tube 0.5-0.7 mm long, the apically free portions 1-1.3 mm long, the anthers ovoid, 1.9-2.2 X 1.1-1.5 mm, apically apiculate, basally deeply cordate, the connective conspicu- ously black punctate; pistil 5.2-5.6 mm long, gla- brous, the ovary oblong, 1.2—1.4 mm long, the style 4—4.2 mm long, prominently punctate, the ovules 46 to 48. Fruits red, then burgundy, then black at maturity, globose, 8-9.2 mm diam., prominently black punctate. Distribution. Ardisia awarum is endemic to the Reserva Indfgena Awa in San Lorenzo, Ecuador. We would anticipate that its range extends to the south- em m of Nariño, Colombia, on the western slopes of the Western Cordillera, growing between (80-3200 and 600(—1000) m in elevation. cology and conservation status. Ardisia awa- rum occurs in primary, pluvial premontane forests, apparently along the forest margins. Because the majority of the collections are from protected areas, it does not appear that the species is threatened. Etymology. The specific epithet is in honor of the Awa, a group of indigenous people found along the western slopes of the western Andean Cordil- lera, from Esmeraldas, Ecuador, north to central- western Narifio, Colombia. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia awarum is very similar to A. dun- lapiana because of its leaf blades coriaceous, calyx coriaceous, corolla chartaceous, the tube glabrous outside, shorter anthers to 2.2 mm long, and shorter styles to 4.2 mm long. Paratypes. ECUADOR. Carehi: Tulcán Cantón, Pa- поша Chical, sector Gualpi medio, Reserva Indígena Awá, trail to San Marcos N « 27 Mas 1992 (fl). G. Tipaz et al. 1028 (FTG, MO). Es- f the community center, 23— Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 235 meraldas: San Lorenzo Cantón, Reserva Étnica Ама, Centro ig pegs =з July 1992 (fr), C. Aulestia et al. 70 (FTG, MO, QCNE, US); Reserva nion Awa, Parro- quia Mataje, Centro VAM 21 Sep. 1992 (fr). C. Aulestia et al. 377 (FTG, MO, QCNE); Reserva fien ЛАН. Ра rroquia Mataje. Centro Mataje, 21 Sep. 1992 (fr), C. Au- lestia et al. 513 (FTG, MEXU, MO, NY, QCNE): Parroquia Alto Tambo, sector 7 Cristal, 13 Apr. 1992 (fr). G. Tipas et al. 777 (FTG, MO); Reserva Indígena Awá, Ric 19-24 Oct. 1992 (fr). QCNE aurte, G. Tipaz et al. 2044 (FTG, MO, 26. Ardisia blepharodes Lundell, Wrightia 4: 55. 1968. Auriculardisia blepharodes (Lundell) Lundell, Phytologia 49: 342. 1981. TYPE: Costa Rica. Cartago: El Muñeco on the Río Navarro, 1400-1500 m, 6-7 Mar. 1926 (fr). P. Standley & R. Torres R. 51266 (holotype, US!, LL neg. 1971-22!; isotype, GH!). Figure 28. Small trees 3—5 m tall. Branchlets slender, terete, 2-5.5 mm diam., densely and minutely appressed rufous furfuraceous-lepidote. Leaves with blades membranous to chartaceous, narrowly oblong, nar- rowly elliptic to oblanceolate, 5.2-12.6 X cm, apically abruptly acuminate, with an acumen 6-12 mm long, basally acute, decurrent on the pet- iole, prominently punctate and punctate-lineate above and below, glabrous above, densely and mi- nutely appressed rufous furfuraceous-lepidote, the midrib impressed above, prominently raised below. the secondary veins 17 to 29 pairs, prominulous above and below, the margins entire, flat; petioles slender, marginate, 6-11 mm long, 0.5-1.5 mm diam., glabrous above, furfuraceous-lepidote below. Inflorescences erect, bi- to tripinnately paniculate. 9.2—106.7 the leaves. the rachis, branchlets, abaxial bract sur- .1-12.5 cm, pyramidal, longer than faces and pedicels furfuraceous-lepidote, the branches loosely congested into 3- to 6-flowered corymbs; peduncle 0.8-2.6 cm long, the lower branches subtended by leaves; inflorescence bracts unknown; inflorescence branch bracts caducous., membranous, ovate, 0.8-2.2 X 0.4-1.2 mm, api- cally acute, prominently punctate and punctate-li- neate, glabrous above, furfuraceous-lepidote below. the margins irregular, hyaline. sparsely glandular ciliolate; floral bracts similar to but 0.5-0.9 X 0.4-1.1 mm: pedicels stout, obsolete to 1.2 mm minutely егоѕе, the inflorescence branch bracts, long, inconspicuously punctate and punctate-lin- eate, furfuraceous-lepidote. Flowers 5-merous, pink; calyx lobes chartaceous to coriaceous, orbic- ular to ovate, 2.2-2.8 X 2.4-2.8 mm, apically obtuse to rounded, prominently punctate and punc- tate-lineate, glabrous adaxially, furfuraceous-lepi- dote abaxially, the margins irregular, minutely erose, hyaline, T glandular ciliolate: corolla chartaceous, 6.9— 2.2-2.4 mm long, the lobes ur d ovate to lanceolate, 4.5—4.9 X punctate and. punctate-lineate, glabrous adaxially. m long, the tube 2 х 1.8-2 mm, apically acute, prominently furfuraceous-lepidote abaxially, the margins entire, hyaline; stamens 6.5.6.9 mm long, the filaments .5—4.7 mm long, the staminal tube 0.8-0.9 mm long, the apically free portions 3.6-3.9 mm long, the anthers narrowly ovoid, 2.2-2.4 X 0.9-1 mm, apically apiculate, basally deeply cordate, the con- —6.6 mm long, glabrous, the ovary oblong, 1.1-1.3 mm long, the style 5 nective conspicuously punctate; pisti —5.3 mm long, prominently punctate, the ovules 15 to 21. Fruits globose, 5-6 mm diam., inconspicuously punctate. Distribution. Ardisia blepharodes is endemic to the Cordillera Talamanca, in Limón and Cartago. Costa Rica, growing from 1200 to 1500 m in ele- vation. Ecology and conservation status. Ardisia blep- harodes occurs in moist premontane forests. Be- cause it is known from so few collections, no further information is available about its conservation sta- ут Хову. The specific epithet was derived from the Greek “Blepha-” meaning relating to eye- lashes or eyelids, referring to margins fringed with ” ee, * hairs or ciliated, and “odes” or meaning like or resembling, specifically referring to the glandular ciliolate calyx lobes. thin Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia blepharodes is easily distinguished from A. hagenii by its smaller leaves to 12. long, narrower calyx lobes to 2.8 mm wide, shorter corolla lobes to 4.9 mm long, shorter anthers to 2.4 mm long, and shorter styles to 5.3 mm long. Ardisia blepharodes is separated from both A. fimbrillifera and A. pseudoracemiflora by its smaller and narrow- er leaf blades to 12.6 X 3.2 cm, shorter pedicels obsolete to 1.2 mm long, narrower corolla lobes to 2 mm wide and shorter and narrower anthers to 2.4 X ] mm. pecimens examined. COSTA RICA, Cartago: knoll W “of Quebrada Casa Blanca, Tapantí, 26 Dec. 1984 (fl), M. Grayum et al. 4648 (FTG, LL, MO), 22 June 1985 (fl). M. T: туит & К. с 5434 (СВ, FTG, LL, MEXU, MO). Limón: mountain between Cerro Chimú and Cerro Ма 30 Apr. 1985 (fr), L. Gómez et al. 23564 (FTG O, NY). — 27. Ardisia capitellata Lundell, Wrightia 6: 67. 1979. Auriculardisia capiteliata pieni. Lun- dell, Phytologia 49: 981. TYPE: Costa Rica. Puntarenas: Pen coffee fincas along Annals of the 236 Missouri Botanical Garden ((OW) 9 sogon 7M шолу 7) :(jjpopum] орокту pisipipmouny Jo 9dsjost OW) V£€9Iz uəavy 4 шош g ЧАМ) 82997 10017) р cadAjost шолу umeap y) in1] у— ләмо]] “G— “Yoursq Зипәмор] "V— °оорјәпарә терү (аи) бр әзі C(OW) Fers ләш 7M X шток) W шоу 7) (OW) 8bOr T? 1 um&na:) "jy шо g (SQ) 99216 "y seo] у у *appuvis q ‘edsyojoy шолу umeap ү) игид 71)— JMOL 7g— "qougiq Зипәмор “y— ‘saposvydayq pisipay әр BZ әп; [а] m, | Volume 90, Number 2 2003 Ricketson & Pipoly 237 Revision of Ardisia subg. Auriculardisia Río Coto Brus, near Cotán, 23 km N La Unión, on Panama border, 9 Aug. 1974 (fl), T Croat 26678 (holotype, LL!, F neg. 55613; isotypes, MO!, NY!). Figure 29. pec latisepala Lundell, Wrightia 7: 269. 1984. . Ardisia latisepala (Lundell) Lundell, Phyto- lo ogia 61: 65. 1986, nom. inval. Ardisia а, О ds & Ricketson, Sida 1 18: 1998. РЕ: a Rica. Puntarenas: on and pie Wil- son's finca a, м. km S of San Vito de Java, ca. 4000 ft. [1219 m], 19 Aug. 1967 (fl), P. Raven 21653A (ho- lotype, F!, F neg. 683245; isotype, MO!). Shrubs or small trees to 9.1 m tall, and 10 ст diam. Branchlets slender, terete, 3—7.5 mm diam., densely and minutely appressed rufous furfura- ceous-lepidote. Leaves with blades chartaceous to coriaceous, elliptic to narrowly elliptic, 16.4—32. х 4.4—9.8 cm, apically acuminate, with an acumen 0.5-1.7 ст long, basally acute, decurrent on the petiole, inconspicuously punctate and punctate-lin- eate above and below, mostly glabrous above, densely and minutely appressed rufous furfura- ceous-lepidote below, denser along the midrib and secondary veins, the midrib impressed above, prominently raised below, the secondary veins 39 to 57 pairs, prominulous above and below, the mar- gins entire, flat to slightly revolute; petioles slender, 6-18 mm long, 1-3 mm diam., glabrous above, fur- furaceous-lepidote below. Inflorescences erect, bi- to 4.8-31.5 cm, pyramidal, longer than or as long as the leaves, the tripinnately paniculate, 8.2-32.4 X peduncle, rachis, branches, and pedicels densely and minutely appressed rufous furfuraceous-lepi- dote. the branches loosely congested into 4- to 9- flowered corymbs: peduncles 0.6-3.6 cm long. the lower branches subtended by leaves; inflorescence bracts and branch bracts unknown; floral bracts usually persistent, membranous, ovate, 2.5—4.5 X 0.8-3.2 mm, apically rounded to acute, inconspic- uously punctate and punctate-lineate, glabrous above and below, the midrib unknown, the second- ary veins unknown, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate: pedicels stout, 0.9-2.5(-4) mm long, inconspicuously punc- tate and punctate-lineate. Flowers 5-merous, rarely 6-merous, white to pink or light purple: calyx lobes chartaceous to coriaceous, orbicular to ovate, 3—3.4 x 2.6-3.2 mm, apically rounded, conspicuously red punctate and punctate-lineate, sparsely furfu- raceous-lepidote abaxially, glabrous adaxially, the margins entire, sparsely glandular ciliolate; corolla chartaceous, 7.2-7.6 mm long, the tube 1.4-1.6 mm long, the lobes narrowly ovate to lanceolate. 5.8-6 X 2.8-3.2 mm, apically acute to subulate, inconspicuously red punctate and punctate-lineate, glabrous throughout, the margins entire, opaque; stamens 3.9-4.2 mm long, the filaments 2.6-2.8 mm long, the staminal tube 0.7-0.8 mm long, the apically free portions 1.8-2.1 mm long, the anthers lanceoloid, 3.2-3.5 X 1.3-1.5 mm, apically broad- ly apiculate, basally deeply cordate, the connective conspicuously punctate; pistil 6—6.2 mm long, gla- brous, the ovary ovoid 1—1.1 mm long, styles 5—5.2 mm long, epunctate, the ovules 22 to 24. Fruits depressed globose, 7-9 mm diam., inconspicuously punctate and punctate-lineate, inconspicuously costate. Distribution. Ardisia capitellata is endemic to the southwestern portion of the Cordillera Talaman- a, Puntarenas, Costa Rica, growing from 1200 to 1800 m in elevation. Ecology and conservation status. Ardisia capi- tellata occurs in primary and secondary lower mon- tane. wet. forests, occasionally in disturbed areas. Because it is known from so few specimens, its cur- rent conservation status is unknown. Etymology. The specific epithet refers to the compact, capitulate appearing corymbs of the inflo- rescence branches. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia capitellata is very similar to A. ge- neralensis, but is distinguished from it by the short- er and narrower leaf blades to 32.8 X 9.8 cm, the smaller inflorescence 32.4 X 31.5 cm, flowers 5- or 6-merous, smaller and narrower calyx lobes to Ww 1 x 3.2 mm, shorter and narrower corolla lobes to 6 X 3.2 mm, and shorter and narrower anthers to 3.5 Populations corresponding to the type of Auri- х 1.5 mm. culardisia latisepala Lundell match the type of Ar- disia capitellata Lundell in all respects except for slightly larger flowers. Specimens examined. COSTA RICA. Puntarenas: Cantón de Coto Brus, Parque Indígena La Amistad, Cor- dillera Talamanca, Estación Pittier, Río Gemelo, 30 Jan. r), L. Angulo et al. 22 (ЕТС, INB, МО); foothills ‚ Cordillera Talamanca, in the area of Sitio Cotón (Cotonsito), along the road to Coto Brus, 3—4 5 (fl), 6. Davidse 24578 (LL, МО); Cordillera eck. area around Rio Canasta, 9.5 airline km NW of Ag Caliente, between Cerro Frantzius and Cerro Pittier, 6 Sep. 1984 (fl), G. Davidse et al. 28363 (LL, MO); Zona Protectora Las Tablas, Estación Biológica Las Alturas, Pa- ralelo Trail, + 331 (CR, FTG, digena La Аай: Cordillera del a, tier, Gemelo Trail, 14 May 1995 (fl), B. Gamboa R. 247 (INB. MO); Cantón de Coto Brus, Cuenca Térraba-Sierpe. Estación Biológica Las Alturas, trail to Cerro Есһапаї, 28 Oct. 1997 (fl), B. Gamboa R. 1902 (INB, MO): along trail tween Las Cruces Botanical Garden and Río Java, ca. 35 km SE of San Vito de Coto Brus, 12 Sep. 1985 (fl). - 238 Annals of the Missouri Botanical Garden M. Grayum et al. 5973 (FTG, LL, MO); Las Alturas, 26 Aug. 1974 (fl), P. Maas & B. Alpin 1492 (F); Cantón de Coto Brus, Parque Indígena La Amistad, Cordillera Tala- manca, Estación Pittier, Pittier Trail, 1 Aug. 1995 (fl), E. Navarro 170 (CR, FTG, INB, MO); Wilson's Farm, 6 km 5 of San Vito de Java, 16 Aug. 1967 (fl), P. Raven 21827 (F. MO, NY); Cantón de Coto Brus, Parque Indígena La Amistad, C erc aeuo 'a, trail to Altar aged ded Ca- nasta, 28 Jan. 1995 (fr Villalobos 6 (FTG, INB, MO): Cantón de Cat Brus, = Indígena La jud. ta d Cor- dillera Talamanca, Estación Pittier, 12 June 1995 (fl), R. Villalobos 205 (FTG, INB, MO). — — = 28. Ardisia cartagoana Lundell, Wrightia 6: 68. 1979. Auriculardisia cartagoana (Lundell) Lundell, Phytologia 49: 343. 1981. TYPE: Costa Rica. Cartago: along road between Juan Vinas & Turrialba, 7 km W Turrialba, along ditch, 1 July 1976 (fl), T. Croat 36841 (holo- type, MO!, F neg. 55674: isotype, LL!). Figure 30 Shrubs or small trees 1—8 m tall. Branchlets stout, terete, 5-10 mm diam., densely rufous furfura- ceous-lepidote. Leaves with blades membranous. 15.4—32.9 X 5.1-8.7 cm, apically acuminate, with an acumen 6-10 mm long, basally oblanceolate. acute, decurrent on the petiole, inconspicuously and conspicuously punctate and punctate-lineate, glabrous, the midrib impressed above, prominently raised below, the secondary veins 40 to 51 pairs, nitid above, prominulous below, the margins entire, flat to slightly inrolled; petioles stout, marginate, nearly obsolete to 1.5 em long, 3-4 mm diam., gla- brous. Inflorescences erect, bi- to tripinnately pa- 13.1-28.4 X 9.7-18.8 cm, usually longer than the leaves, mixed minutely cu- niculate, pyramidal, puliform and furfuraceous-lepidote, the branches congested to loosely congested into 5- to 9-flowered racemes with flowers clustered in a pseudocorymb at apex; peduncles 1.5-3.5 ст long, the lower branches subtended by leaves; inflorescence and branch bracts unknown; floral bracts caducous. membranous, 1.7-2 X 1.4—1.8 mm, apically acute, conspicuously punctate and. punctate-lineate, gla- brous adaxially, furfuraceous-lepidote abaxially, the margins irregular, minutely erose, hyaline sparsely glandular ciliolate; pedicels stout, obsolete to 2 mm long, inconspicuously punctate and punctate-lin- eate, furfuraceous-lepidote. Flowers 5-merous. calyx lobes chartaceous, X 2.1-2.4 mm, apically rounded, inconspicuously punctate and. punctate- white to pink violet; orbicular to oblate, 2—2.3 lineate to nearly epunctate, glabrous adaxially, fur- furaceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular cilio- ale; corolla membranous, 5.2—5.5 mm long, the tube 1.1-1.3 mm long, the lobes ovate, 4-4.2 X 1.9-2.7 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the mar- gins entire, hyaline; stamens 4.8-5 mm long, the filaments 2.3-2.4 mm long, the staminal tube 0.6- 0.7 mm long, the apically free portions 1.6-1.8 mm long, the anthers lanceoloid, 2.9—: I-1.1 mm, apically apiculate, basally deeply cordate, the con- nective conspicuously punctate; pistil 6.4—-6.5 mm long, glabrous, the ovary oblong, 1.8-1.9 mm long, the style 4.64.7 16 to 21. Fruits (immature) globose, 4.5-5.2 mm mm long, epunctate, the ovules long. prominently punctate and punctate-lineate. Distribution. Ardisia cartagoana is endemic to Costa Rica and has been found only in the Cordil- lera Talamanca near Cerro Turrialba (Cartago) and the Fila de Matamá at the headwaters of the Río Boyei (Limón). growing at 1110 to 1300 m in ele- vation. Ecology and conservation status. Ardisia car- tagoana occurs in premontane wet forest. It has been found in forest remnants near sugar cane fields and thus is very tolerant of disturbance and not considered threatened. Etymology. The specific epithet referred to the Province of Cartago, Costa Rica, where the type was collected. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia cartagoana is closely related to A. megistophylla, A. aguirreana, A. cogolloi, and А. crassipes because of its long calyx lobes with thick petioles. However, А. cartagoana is separated from A. megistophylla by its shorter calyx lobes to 2.3 mm long, shorter corolla lobes to 4.2 mm long. shorter anthers to 3 mm long, and shorter styles to 4.7 mm long. Ardisia cartagoana is easily distin- guished from A. aguirreana by its shorter leaf blades to 32.9 cm long, smaller branchlets to 10 mm in diameter, shorter peduncles to 3.5 em long, and wider corolla lobes to 2.7 mm wide. Ardisia cartagoana is separated from А. cogolloi by its smaller and wider orbicular, not ovate calyx lobes 2.3 X 2.4 mm, longer and narrower anthers 3 X 1.1 mm, longer styles to 4.7 mm long, and smaller fruits to 5.2 mm in diameter. Ardisia cartagoana differs from A. crassipes by its membranous not co- riaceous leaf blades, which are glabrous above and below, thinner petioles to 4 mm in diameter, nar- rower calyx lobes to 2.4 mm wide, shorter corolla lobes to 4.2 mm long, and shorter styles 5.2 mm long. Specimens examined. COSTA RICA. Cartago: Cantón de Turrialba, Valle del Reventazón, Grano de Oro, Moravia de Chirripé, 10 Oct. 1992 (fl, fr), P Campos 57 (CR, FTG 239 Ricketson & Pipoly Volume 90, Number 2 2003 Revision of Ardisia subg. Auriculardisia 20881 70 12 Kodig ‘{ чолу o (WAVE) 86r 70 12 0110907) "y "ed&opoq шолу имер g S v) CON) N1} 7)— чәмо 'g— Wqgoueiq SuLI2M0[] ‘ү "10110309 тері ausu) те әлпйї у] C(OIN) 4€ sodum;) q шолу D ‘g (OW) [#898 1904) 1, ‘adAjojoy шо UMPIP V) сип 7)— чәмор] £— “yours duuoMo[| "ү ‘оиро9рирә visipay— (әр) OF әлпдї 4] 240 Annals of the Missouri Botanical Garden INB, MO), 29 June 1993 (fl), P. corps 76 (FTG, INB, MO). Limón: Cantón и о Cordillera Talamanca, № flank of Fila de Mata t headwaters of Río Boyei, 17 Aug. 1995 (fl), M. run 11046 (INB, MO) 29. Ardisia cogolloi Pipoly, Caldasia 16: 277. 1991. TYPE: Colombia. Antioquia: Mpio. de Urrao, Parque Nacional Natural *Las Orquí- deas," Sector Venados, right bank of upper Río Venados, 06°34'N, 076?19' W, 1150-1300 m. 26 July 1988 (fl), A. Cogollo, J. Ramírez & О. Alvarez 3498 (holotype, JAUM!; isotypes, COL!, FMB!, HUA!, MO!). Figure 31. Trees 3-15 m tall, 9.5-26 cm diam. Branchlets stout, terete, 6.5-10.5 mm diam., densely rufous furfuraceous-lepidote. Leaves with blades charta- ceous, elliptic, 14.4—50.2 X 5.9-17.6 cm, apically obtuse with an abrupt acute acumen, acumen 0.6— 0.9 cm long, basally cuneate, decurrent on the pet- iole, prominently and inconspicuously punctate and punctate-lineate above and below, glabrous above, densely rufous furfuraceous-lepidote below, the midrib impressed above, prominently raised below, the secondary veins 38 to 45 pairs, nitid above. raised below, the margins entire, subrevolute peti- oles stout, canaliculate, 1.5-3.5 em long, 4—7 mm diam., glabrous above, furfuraceous-lepidote below. Inflorescences erect, bipinnately paniculate, 15.4— 6 X 4.5-13.6 cm, pyramidal, shorter than the leaves, the rachis, branchlets, abaxial bract surfac- es, and pedicels densely furfuraceous-lepidote, the branches loosely congested into 2- to 7-flowered corymbs; peduncle 0.2-1 cm long; inflorescence bracts and branch bracts unknown; floral bracts ca- ducous, membranous, ovate, 1-1.5 X 0.5-0.8 cm. apically acute, prominently punctate and punctate- lineate, glabrous above, furfuraceous-lepidote be- low, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; pedicels stout, obsolete to 0.7 mm long, prominently punctate and punc- tate-lineate, furfuraceous-lepidote. Flowers 5-me- rous, white; calyx lobes chartaceous, ovate, 2.5-2.6 3-1.9 mm, apically obtuse, prominently punc- tate and. punctate-lineate, glabrous adaxially, fur- furaceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular cilio- late; corolla membranous, 6.2-6.5 mm long, the tube 2-2.3 mm long, the lobes narrowly oblong, 4.2-4.5 X 2.4-2.6 mm, apically acute, prominently punctate and. punctate-lineate, glabrous adaxially, furfuraceous-lepidote abaxially, the margins entire, hyaline; stamens 5.2-5.9 mm long, the filaments 3.6-3.8 mm long, the staminal tube 1-1.1 mm long, the apically free portions 2.5-2.9 mm long, the an- thers ovoid, 2.3-2.6 X 1.3-1.5 mm, apically apic- ulate, basally cordate, the connective conspicuous- ly punctate; pistil 3.7-3.8 mm long, glabrous, the ovary oblong, 1.6-1.7 mm long, the style 2-2.2 mm long, epunctate, the ovules 19 to 27. Fruits globose, 8—9.8 mm diam., prominently punctate. Distribution. Ardisia cogolloi is endemic to the Cordillera Occidental of Colombia, growing from 800 to 1750 m in elevation. Ecology and conservation status. Ardisia cogo- lloi occurs in premontane pluvial forests, where it is ecologically significant, frequently attaining a di- ameter at breast height over 15 cm. Because the majority of the collections are from protected areas, it does not appear that the species is threatened. Etymology. This species was named in honor of Biól. Alvaro Cogollo Pacheco, Director of Re- search at the Fundación Jardín Botánico, “Joaquin Antonio Uribe," in Medellín, Colombia. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia cogolloi is closely related to A. me- gistophylla, A. aguirreana, A. crassipes, and A. car- tagoana because of its long calyx lobes with thick petioles. However, Ardisia cogolloi is separated from A. megistophylla by its shorter calyx lobes to 2.6 mm long, shorter corolla lobes to 4.5 mm long, shorter anthers to 2.6 mm long, and shorter styles to 2.2 mm long. Ardisia cogolloi is easily distin- guished from A. leaf blades to 50.2 cm long, smaller branchlets to 10.5 mm in diameter, shorter peduncles to 1 cm long, and wider corolla lobes to 2.6 mm wide. Ardisia cogolloi is separated from both A. crassipes and A. aguurreana by its shorter cartagoana by its larger and. narrower ovate, not orbicular calyx lobes to 2.6 X 1.9 mm, shorter and wider anthers to 2.6 X 1.5 mm, shorter styles to 2.2 mm long, and larger fruits to 9.8 mm in di- ameter. Specimens ае COLOMBIA. Antioquia: Mpio. de Urrao, Parque Nacional Natural *Las Orquídeas," Ve- reda Calles, ere bank of Rfo Calles, on the range NW of the Cabaña er 13 Чом, 199 3 s ial ашн E Las Orquídeas,” ч bom eq on the range NW Я т.), J. Pipoly et al. 16729 (JAUM, MO), 16782 (JAUM, MO). 16601 (JAUM, MO), 16958 а МО); Mpio. de Jrrao, Parque Nacional و‎ Las Orqufdeas,” V Calles, Alto de гоа . 1 km from С abana C Dec. 1993 (ster.), J. Pipoly et al. 17648 МО); Mpio. de Urrao, Parque Nac үш Natural *Las Or- quídeas," Vereda Calles, right bank of Río Calles, 9 Dec. 1993 (fl). J. Pipoly et al. 18000 (FTG, JAUM, MO); Mpio. de paleo Vereda Venados, Parque Nacional Natural Las Orquídeas, sitio La Miquera, 3 Feb. 1995 (fr), J. Pi- 2 et al. 18307 (JAUM, MO). ^ © O & = =. = 5 = ae E = = p " ^^ [ool A = — 30. Ardisia coloradoana Lundell, Wrightia 6: Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 69. 1979. Auriculardisia coloradoana (Lun- dell) Lundell, Phytologia 49: 343. 1981. YPE: Panama. Chiriquí: Cerro Colorado, 34— 35.6 km above Rfo San Félix, 13-14.6 above turnoff to Escopeta, 1390—1410 m, 15 July 1976 (fl), T. Croat 37235 (holotype, MO!, F neg. 55684!). Figure 32. Treelets or trees 2—6 m tall. Branchlets slender, terete, 3—6 mm diam., densely cupuliform and fur- furaceous-lepidote. Leaves with blades chartaceous l.l- 3.6 cm, apically long acuminate, with an acumen to coriaceous, elliptic to oblong, 3.4—10.8 х 6-12 mm long, basally acute or cuneate, decurrent on the petiole, inconspicuously punctate and punc- tate-lineate, nearly glabrous above, densely ap- pressed rufous furfuraceous-lepidote, the midrib impressed above, prominently raised below, the secondary veins 23 to 29 pairs, prominulous above and below, the margins entire, flat, at times drying inrolled; petioles slender, canaliculate, 1.1-2.4 ст long, glabrous above, densely appressed furfura- ceous-lepidote below. Inflorescences erect, bi- to tri- pinnately paniculate, 5-24 X 4.5-14 cm, pyrami- dal, longer than the leaves, peduncle, the rachis, branches and pedicels densely cupuliform and fur- furaceous-lepidote, the branches loosely congested into 3- to 7-flowered corymbs; peduncles nearly ob- solete to 3.8 cm long, the lower branches subtended by leaves; inflorescence bracts unknown; inflores- cence branch bracts caducous, membranous, ob- long, 2.5-2.9 X 1-1.3 mm, apically acute, promi- nently punctate and punctate-lineate, glabrous above, appressed furfuraceous-lepidote below, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflo- but 2.8-4.1 X 0.6-0.9 mm; pedicels slender, 3.5—5.1 mm long, inconspic- rescence branch bracts, uously punctate and punctate-lineate, mixed cu- puliform and furfuraceous-lepidote. Flowers 5-me- rous, white or cream; calyx lobes chartaceous, 6 X 1.6-1.9 mm, apically conspicuously punctate and orbicular to oblate, 1.4—1. acute to rounded, punctate-lineate, glabrous adaxially, furfuraceous- lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate: corolla membranous to chartaceous, 3.94.1 mm long, the tube 1-1.4 mm long, the lobes lanceolate, 2.7-2.9 X 1.6-1.8 mm, punctate and punctate-lineate, glabrous throughout, apically acute, inconspicuously the margins entire, hyaline; stamens 2.9-3 mm long, the filaments 1.4—1.6 mm long, the staminal tube 0.6-0.7 mm long, the apically free portions 0.7—1 mm long, the anthers ovoid to narrowly ovoid, 1.5-1.7 X 0.7-0.9 mm, apically apiculate, basally deeply cordate, the connective conspicuously punc- tate; pistil 3.8—4 mm long, glabrous, the ovary ob- long, 1.1-1.5 mm long, the style 2.5-2.7 mm long, nearly epunctate, the ovules 19 to 23. Fruits glo- bose, 6.5-7.2 and punctate-lineate. mm diam., inconspicuously punctate Distribution. Ardisia coloradoana is endemic to Cerro Colorado and Cerro Hornito along the Con- tinental Divide, at the junction of Bocas del Toro and Chiriquí, Panama, growing at 1290 to 1950 m in elevation. Ecology and conservation status. Ardisia colo- radoana occurs in cloud forests, where it is locally common. Label data indicate that it is common among forest remnants and thus tolerates moderate amounts of disturbance. Therefore, we do not be- lieve it is under threat at this time Etymology. The name refers to id type loca- tion on Cerro Colorado. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia coloradoana is most closely related to А. atropurpurea because of the small calyx lobes, between 1.4 and 1.7 mm long, long pedicels, nar- row corolla lobes, and short anthers. However, A. coloradoana is easily distinguished from А. eucu- neata and A. atropurpurea by its thicker branchlets to 6 mm in diameter, shorter petioles to 2.4 cm long, and wider calyx lobes to 1.9 mm wide. Ardisia coloradoana 1s most easily confused with A. pleurobotrya (sect. Pleurobotryae) based on the leaf size and shape. However, it is easily separated by its terminal inflorescence and short pedicels to 4.8 mm long, not lateral inflorescences with long pedicels to 10.5 mm long, as in A. pleurobotrya. Specimens examined. p: dica Bocas del Tor mi. beyond Campamento Chamí, + 12 mi. from San Félix, 20 June 1986 (fl), W. f Arch 16283 (LL, MO); js мш of Cerro Colorado, « on "ers se pars Di- om Chamí ‚ 12 Apr © (fl), G. Mc- Pherson so (LL, MO). Border of зби del Toro and hiriqui: Cerro Colorado, road along top, 13 Aug. 1977 г), J. "RU et al. 4710 (MO). Chiriquí: Fortuna Dam region, along trail to Cerro Hornito, Pate de Macho, on S ridge of watershed, 17 Jan. 1989 (fr), G. McPherson 13566 (FTG, MO, ЖЄ; e HJ — 31. Ardisia conglomerata Lundell, Wrightia 6: 71. 1979. Auriculardisia conglomerata (Lun- Sie Lundell, Phytologia 49: 343. 1981. YPE: Panama. Veraguas: NW of Santa Fé, 2 М from Escuela Agrícola Alto - Piedra, on ridge top below summit Cerro 1975 (fr), 5. Mori & J. Kallunki ps (holo- type, МО!, F neg. 55675). Figure 33. Trees 10 m tall. Branchlets stout, terete, with Annals of the 242 Missouri Botanical Garden ((OW) 626 nqunj]py ^f 3p Voy "< *2d5jo[ou шолу umeip Ң-ұ) inig ‘H— ‘әәвупѕ [erxeqe uwe 74)— "шлеш үеләдеү ‘URIS :4— ens [erxepe “UDUIRYS pee зэмођ} permed jo perq ‘Gg “PIUDISAIIOYUL je erq у— `әәерпѕ Jeo] jerxeqe jo perq i fee "Wuouegaq BSULIOMOT J v— "D]D42u10]8u02 оер ausu) єє ans] ((OW) 28291 «yq M “Woy umeap g ^y) 2290[4 ^g— yuraq guus«o[4 "V— ‘рирорргороэ mspay QJ) се on314 SSS Ex ay " N i PLL S 1 IE с = D qu хэ c£ Volume 90, Number 2 2003 Ricketson & Pipoly 243 Revision of Ardisia subg. Auriculardisia large. swollen leaf scars, 6-9 mm diam., with a mixture of densely cupuliform and furfuraceous- lepidote scales. Leaves with blades coriaceous, el- liptic, 9.6-16.5 X 3.6-6.8 cm, apically acuminate, with an acumen 0.6—0.9 cm long, basally obtuse, decurrent on the petiole, inconspicuously punctate and punctate-lineate, glabrous above at least with age. below with a mixture of densely cupuliform and furfuraceous-lepidote scales, especially so along the midrib, the midrib impressed above, prominently raised below, the secondary veins 22 to 28 pairs, prominulous above and below, the mar- gins entire, revolute; petiole stout, marginate, 0.8— 1.2 cm long, glabrous above at least with age, below with a mixture of densely cupuliform and furfura- ceous-lepidote scales. Inflorescences erect, bipin- X 44.8 ст, obpyramidal, rachis, nately paniculate, 3—3.9 shorter than the leaves, the peduncle. branchlets, and pedicels with a mixture of densely cupuliform and furfuraceous-lepidote scales. the branches congested into 2- to 8-flowered corymbs: peduncles to 6 mm long; inflorescence bracts per- sistent, chartaceous, oblong, 1.2-1.4 X 0.6-0.8 cm, apically acute, scattered inconspicuously punctate and punctate-lineate, glabrous above, below with a mixture of scattered cupuliform and furfuraceous- lepidote scales, the midrib inconspicuous, the sec- ondary veins obscure, the margins entire, revolute; inflorescence branch bracts similar to the inflores- cence bracts, but 4.5—6.7 X 2.5-5 floral bracts persistent, membranous, ovate, 0.6-1.2 X .7-0.9 mm, apically acute, scattered inconspicu- ously punctate and punctate-lineate, glabrous о mm: above, sparsely furfuraceous-lepidote below, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; pedicels stout, obsolete to 1.2 mm long, inconspicuously punctate, with a mixture of scattered cupuliform and furfuraceous-lepidote scales. Flowers [measurements from a single, par- tial, old flower] 5-merous, color unknown: calyx lobes membranous to chartaceous, orbicular to very widely ovate, 0.9-1.2 х 0.9-1.2 mm, apically ob- tuse to rounded, prominently punctate and punc- tate-lineate, glabrous throughout, at least with аре, the margins irregular, minutely erose, hyaline. sparsely glandular ciliolate; corolla membranous, 2 6 mm long, the tube 0.5-0.7 mm long. the lobes ovate to narrowly ovate, 1.9-2.1 X 0.8-1 mm, apically acute, prominently punctate and punctate- lineate, glabrous throughout, the margins entire, hyaline; stamens ca. 2.8 mm long, the filaments to 1.6 mm long, the staminal tube ca. 0.4 mm long. the apically free portions ca. 1.2 mm long, the an- thers lanceoloid, ca. 1.6 X 0.6 mm, apically apic- ulate, basally subcordate, the connective incon- spicuously pistil 4—4.4 mm long, glabrous, the ovary oblong to globose, 1-1.2 mm diam., the style 3-3.2 mm long, conspicuously punctate, the ovules unknown. Fruits globose, 4— punctate; 5.5 mm diam., prominently punctate. Distribution. Ardisia conglomerata is known only from the type collection on Cerro Tute in Ve- raguas, Panama, growing at 700 to 800 m in ele- vation. Ecology and conservation status. Ardisia con- glomerata occurs in cloud forests near the summit of Cerro Tute. Because of its restricted distribution, it should be considered threatened. Etymology. The specific epithet meaning clus- tered, often spherically so, referred to the small compact corymbs of the inflorescence Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia conglomerata is one of a number of species that have branchlets covered with a mixture of dense cupuliform and furfuraceous-lepidote scales. Ardisia conglomerata can be confused with A. lundelliana owing to the coriaceous, elliptic leaves with obtuse to rounded bases and the rela- tively stout, terete branchlets. However, Ardisia conglomerata is easily separated from A. lundellia- na because of its erect inflorescence with much shorter pedicels obsolete to 1.2 mm long and much shorter anthers to А Likewise, within Ardisia subg. Auriculardisia sect. Palmanae, another species that may be con- fused with Ardisia conglomerata is А. crassiramea, which also has branchlets covered with a mixture of dense cupuliform and furfuraceous-lepidote scales and straight branchlets. However, A. conglo- merata differs by the branchlets terete, with large petiole scars, inflorescences bipinnately paniculate and obpyramidal, the pedicels obsolete to 1.2 mm long, the orbicular to widely ovate wider calyx lobes to 1.2 mm wide, the smaller corolla lobes to 2.1 X 1 mm, and the smaller anthers to 1.6 X 0.6 mm. Ardisia conglomerata is known only from the ho- lotype, which is in young fruit. A fragment of the holotype is located at LL and contains pieces of a single flower consisting of a corolla tube with two attached lobes and two filaments with a single loose anther. The floral measurements in the above de- scription come from this material. It appears that this corolla had failed to completely fall from the expanding ovary. 32. "pruna erassipedicellata Lundell, Wrightia 6: 13, t. 140. 1979. Auriculardisia crassipedi- т D undell) Lundell, Phytologia 49: 343. PE: Panama. Veraguas: NW of Santa 244 Annals of the Missouri Botanical Garden Fé, 4.2 km from Escuela Agrícola Alto de Pie- dra, 25 Feb. 1975 (fr), S. Mori & J. Kallunki 4830 (holotype, LL!, F neg. 55643 isotype, MO!). Figure 34. mod roseiflora Lundell, Wrightia 7: 271. 1984. nov. p rdisia roseiflora Pit., Fl. Indo-Chine f 1930. Ardisia dressleri Lundell, Phytologia Cox is inval. Ardisia epi Pipoly & Ric клн Sida 18: 511. 1998. TYPE: Panama. Co- clé: trail from et Divide near iden above El Copé, to Río Blanco del Norte, 08?40'N, 080*36' W, 350—700 m, 20 Feb. 1982 (fl), S. үе J. Mallet & R. Dressler 3646 (holotype, MO! [ur cate |). Treelets or trees to (4—)12 m tall. Branchlets slen- der, terete, 5-7.5 mm diam., densely furfuraceous- lepidote. p with blades coriaceous, elliptic, 4.2—32. .7-9.8 ст, apically acuminate, with an acumen pee mm long, basally acute, decurrent on the petiole, inconspicuous, inconspicuously punctate and punctate-lineate, essentially glabrous above, furfuraceous-lepidote below, the midrib im- pressed above, prominently raised below, the sec- ondary veins 30 to 55 pairs, the margins entire, inrolled; petioles slender, marginate, 12-19 mm long, 2.5-3 mm diam., glabrous above, densely fur- furaceous-lepidote below. /nflorescences erect, bi- to tripinnately paniculate, 6.8-32.4 X 6.2—26.6 cm, pyramidal, usually longer than the leaves, the ra- chis, branches, and pedicels densely furfuraceous- lepidote, the branches loosely congested into 3- to 7-flowered corymbs; peduncles 0.4—2.8 cm long, the lower branches often subtended by leaves; in- florescence bract unknown; inflorescence branch bracts caducous, chartaceous, oblong, 7-8.2 x 2— 2.8 mm, apically acute, prominently punctate and punctate-lineate, essentially glabrous above, furfu- raceous-lepidote below, the margins entire, minute- ly erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflorescence branch bracts, but 0.6-1.3 0.6—1.4 mm; pedicels stout, 7—12 mm long, inconspicuously punctate and punctate- lineate, | pink; calyx lobes coriaceous, oblate, 2.4— furfuraceous-lepidote. Flowers 5-merous, 4.5 mm, apically rounded, inconspicuously punc- tate and punctate-lineate, glabrous adaxially, fur- furaceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular cilio- ate; corolla coriaceous, 9.6-9.8 mm long, the tube 3.7-3.9 mm long, the lobes lanceolate, 6.1—6.3 x 2.0—3.4 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the mar- gins entire, hyaline; stamens 7—7.1 mm long, the filaments 4—4.2 mm long, the staminal tube 2-2.1 mm long, the apically free portions 2-2.1 mm long, the anthers narrowly ovoid to lanceoloid, 3.7-3.8 1.2-1.4 mm, apically apiculate, basally lobate, the connective conspicuously punctate; pistil 6.9— 7.3 mm long, glabrous, the ovary oblong, 1.8-2 mm long, the style 5.1—5.3 mm long, promine ntly punc- to 45. prominently tate and punctate-lineate, the ovules Fruits globose, 6.8-7.5 mm diam., punctate and punctate-lineate. Distribution. Ardisia crassipedicellata is en- demic to the Cordillera Central near and along the Continental Divide in Bocas Del Toro, Coclé, and eraguas, Panama, growing at 350 to 900 m in el- evation Ecology and conservation status. Ardisia cras- sipedicellata is found in premontane forests. Be- cause it is relatively uncommon, it should be con- sidered threatened. Etymology. The spec ific epithet was derived from the Latin "crassi" or thick and “pedicellata” meaning pedicelled, referring to the thick pedicels. ithin Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia crassipedicellata has flowers that are most similar to those of A. liesneri because of their large calyx lobes that are as wide as or wider than long, long and slender pedicels, relatively long co- rolla lobes for the section, and lanceoloid anthers. However, A. crassipedicellata is easily separated from A. liesneri by its densely furfuraceous-lepidote scales, coriaceous leaf blades, larger oblate calyx lobes to 3 X 4.5 mm long, shorter and wider anthers to 3.8 X 1.4 mm, longer corolla lobes to 6.3 mm, and shorter styles to 5.3 mm long. The type of Auriculardisia roseiflora Lundell is in young bud, and without mature pedicels. While the floral parts are smaller, they match those of A. crassipedicellata in all qualitative aspects. Specimens examined. PANAMA. Bocas del Toro: along road above Chiriquí Grande, 25 Oct. 1985 (fr), G. McPherson 7. id (LL, MO). Coclé: Riveras sawmill, El Potroso, 7 km N of El Copé, Alto Calvario, 15 Jan. 1977 fl). J. Folsom io (MO); El Copé, El Potroso, first bottom of ridge after 3 small dips but really not so, Atlantic slope of Alto Calvario, 26 Feb. 1977 (fr), J. Folsom & R. Lantz 1893 (MO, F neg. 55623). ~ T 33. Ardisia crassipes Lundell, Wrightia 4: 57. 1968. Auriculardisia crassipes (Lundell) Lun- dell, Phytologia 49: 343. 1981. he ma. Bocas del Toro: Robalo Trail, N slopes of Cerro Horqueta, 6000—7000 ч [1829-2134 m]. 5-7 Aug. 1947 . P. Allen 4991 (holo- type. МО!, LL neg. 1971-29; isotypes, G! [2]. GH!, LL!). Figure 35. Pana- кы я Lundell, Wrightia 4: 181. 1971. Syn. Auriculardisia horquetensis (Lundell) Lundell 245 Revision of Ardisia subg. Auriculardisia Ricketson & Pipoly Volume 90, Number 2 2003 (пәри viuapnsoy nisipiy jo ad&yo[ou) *(TT) o£gz чәцоя ^y y чору ‘$ wo 7) (YO) 7166F vy qd ‘edAjost шолу имелр q X g ^v) лип (1— moj 7)— :9»ue»sarogu[ *q— :xede qougaq peo ^y— ‘sedisspso терү (їч) се oan3i4 ((OW) 6c£1 шо$]о у] "f шолу 4 (OW) O£gr 1y f $ How S *9d&jost шолу имелр 7) ў y) гип] 7)— “png 1940[4 *q— ‘әиел 8ишәмоү *y— :7p]j221padisspao visipay— QY) PE әлді у >‏ رکو VAS‏ : ESS‏ Specimens examined. PANAMA. Darién: Cerro Pirré, 11 Apr. 1967 (fr), N. Bristan 593 ne on Cerro Pirré, 14 Dec. 1962 (fr), ГА Duke 6561 (LL, MO); Cerro Pirré, 9—10 Aug. 1967 (fl), J. Duke & T. Elias 13750 (GH, LL, MO, B ridge top area N of Cerro Pirré, between Cerro Pirré and Rancho Plástico, 14 Nov. 1977 (fr), J. Folsom et al 6341 (LL, MO); Cana-Altos de Nique, trail on ridge between Río Setegantí and Río Alto Tuira, trail to Paletón, SE of goldmine camp and airstrip, 19 Apr. 1992 (fr), R. Foster 14148 (MO, PMA); SW ridge leading to Alturas de Nique on the Colombian border, 28 Dec. 1980 (fr), R. — Annals of the Missouri Botanical Garden Hartman 12337 (LL |2], MO); Cuasi-Cana Trail жуз ы Сего campamento and La Escalera to “páramo,” E of Tre 30 Apr. 1968 (fr), J. Kirkbride & J. Duke 1279 (Е “М Y): on ridge of Cerro Pirré, 14 Sep. 1989 (fl). G. | МЕРА 14069 (ЕТС, МО); Cana and vicinity, 17 Apr.-8 June 1908 (fr), R. Williams 833 (NY). San Blas: Cerro Obu, 25 June 1986 (fl), G. de Nevers et al. 8077 (LL, MO) —. 38. Ardisia dukei Lundell, Wrightia 4: 45. 1968. Ісасогеа үө (Lundell) Lundell, Phytologia 49: 348 . Auriculardisia dukei (Lundell) Lundell, vba 7: 267. 1984. TYPE: Pan- ama. Darién: peak between Río Bales & Río Aretí at their confluence, ca. 300 ft. [91 m], 13 Sep. 1966 (fl), J. Duke 8741 (holotype. MO!; isotypes, LL!, US!). Figure 40. 10—10.5 cm diam. with Trees with height unknown, Branchlets slender, terete, 4-6 mm diam., densely furfuraceous-lepidote and сири огт scales. Leaves with blades membranous, elliptic to widely obovate, 20.9-31.9 X 8.1—11.4 cm, apically acuminate, with an acumen 7—14 mm long, basally cuneate, decurrent on the petiole, inconspicuously punctate above and below, furfuraceous-lepidote above, with a mixture of furfuraceous-lepidote and the midrib impressed cupuliform scales below, the secondary above, prominently raised below, veins 60 to 68 pairs, prominulous above and below, the margins entire, flat or inrolled; petioles slender, marginate, 9-21 mm long, furfuraceous-lepidote above and below. /nflorescences erect, bi- to tripin- nately paniculate, 27.8-28.9 X 12-21 cm, pyra- midal, longer than the leaves, furfuraceous-lepi- dote, the branches loosely congested into 5- to 8-flowered corymbs; peduncles obsolete; inflores- cence bracts unknown; inflorescence branch bracts unknown; floral bracts usually persistent, membra- nous, ovate, 0.9-1. ).7-0.9 mm, apically acute, prominently punctate and punctate-lineate, gla- brous above, furfuraceous-lepidote below, the mar- gins irregular, hyaline, sparsely glandular ciliolate; pedicels slender, 0.9-1.5 mm long, inconspicuously punctate and punctate-lin- minutely erose, eate, indument as in the branchlets. Flowers 5-me- calyx lobes membranous to charta- ovate, 0.9-1.1 X 1.1-1.3 mm, acute, prominently punctate and punctate-lineate, glabrous adaxially, furfuraceous-lepidote abaxially, hyaline, apically the margins irregular, minutely erose, sparsely glandular ciliolate; corolla and stamens unknown: pistil in young fruit 3.2-5.0 mm diam., glabrous, the ovary in young fruit globose, 1.2—2 mm long, the styles in young fruit 1—3 mm long. prominently punctate, the ovules (according to Lun- dell, 1968) 22 to 24. Fruits unknown. =] Distribution. Ardisia dukei is endemic to : small peak in Darién, Panama, and was collected at ca. 91 m in elevation. Ecology and conservation status. is known from one of the most remote but highly exploited areas of the Darién, so its rarity is prob- ably a reality and not a collecting artifact. There- fore, we expect that the species is at least threat- This species ened if not in danger of extinction. Etymology. This species was named in honor of James A. Duke, ethnobotanist and colleague, who has dedicated his life to the understanding of economic and medicinal tropical plants around the world. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia dukei is most similar to A. anchi- cayana and А. palmana because of its short calyx lobes that are wider than long. membranous to chartaceous leaf blades, and short pedicels. How- ever, Ardisia dukei may be separated from both А. anchicayana and A. palmana by the shorter calyx lobes to 1.1 X 1.3 mm, and shorter styles to З mm long. 39. Ardisia dunlapiana P. H. Allen, Rain Forests Golfo Dulce 409. 1956. Auriculardisia dunla- piana (P. H. Allen) Lundell, Phytologia 49: 343. 1981. TYPE: Costa Rica. Cantón de Osa, vicinity of Esquinas Experi- ment Station, sea level, 16 Apr. 1949 (fl), P. Allen 5274 (holotype, US!, LL neg. 1971-48!; . 68148!, MO!, NY!, US). Puntarenas: isotypes, F!, F neg Figure 41. 8—40 cm diam. densely Shrubs or trees 2-20 m tall, Branchlets slender, terete, 3-5 mm diam., and minutely appressed furfuraceous-lepidote. Leaves with blades membranous, elliptic, 7.5—18.6 X 3.1—5.6 ст, apically acuminate, with an acumen 6-21 mm long, basally acute, decurrent on the pet- iole, prominently punctate and punctate-lineate, glabrous above, furfuraceous-lepidote below, the midrib impressed above, prominently raised below, the secondary veins 23 to 31 pairs, prominulous above and below, the margins entire, flat; petiole slender, marginate, 8—13 mm long, 1-2 mm diam., glabrous above, rufous furfuraceous-lepidote below. Inflorescences erect, bipinnately paniculate, 7-31 X 3-15 em, pyramidal, usually longer than the leaves, the rachis, branchlets, abaxial bract surfaces, and pedicels densely and minutely appressed rufous furfuraceous-lepidote, the branches loosely con- gested into 5- to 12-flowered corymbs: peduncles nearly obsolete to 3.2 em long, the lower branches subtended by leaves; inflorescence bracts unknown: 255 Ricketson & Pipoly Volume 90, Number 2 2003 Revision of Ardisia subg. Auriculardisia ((OW) 960] чәғшоц ^N шо 7) (ON) #226 Yay а ^"9d&osi шолу имрар g ^y) "maj 70— moy ^(— "quer Suuowo[J "V— -pupidpjunp терү (їч) [p 94n314 ((OW) 1529 amq Г *әЧХїоүочц шолу uwap g ^y) maj oinjeunup *g— ^qouexq 2uue0[4 ү “leynp pisipiy — "(yep OF әлпйї 4 ) ED boc 256 Annals of the Missouri Botanical Garden inflorescence branch bracts similar blades, but elliptic to oblanceolate, 1.8—5.4 х 0.5- 2.4 cm, the secondary veins 13 to 25 pairs; inflo- rescence branch bract petioles similar to the leaf blade petioles, but nearly obsolete to 12 mm long: floral bracts caducous, membranous, ovate, 1—1.4 1-1.3 mm, apically acute, prominently punctate and punctate-lineate, glabrous adaxially, the mar- gins irregular, minutely erose, hyaline, sparsely glandular ciliolate; pedicels slender, 1.5-3.5 mm long, inconspicuously punctate and punctate-lin- eate. Flowers 5-merous, white or light pink; calyx 2.2-3 mm, apically rounded, prominently pellucid to orange lobes chartaceous, orbicular, 2.2-3 x punctate and punctate-lineate, scattered. furfura- ceous-lepidote abaxially, the margins irregular, mi- nutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 5.4—7.1 mm long, the tube 1.5-2.2 mm long, sparsely furfuraceous-lepidote abaxially, eee adaxially, the lobes narrowly ovate, 4.2— inently orange punctate and punctate-lineate, com- X 2-2.5 mm, apically acute, prom- pletely glabrous, the margins entire, hyaline; sta- mens 4.2-5 mm long, the filaments 2.5-3.4 mm long, the staminal tube 1—1.2 mm long. the apically free portions 1.5-2.2 mm long, the anthers ovoid to oblongoid, 2.2-2.7 X 1.1-2.1 mm, apically apicu- late, basally deeply cordate, the connective con- spicuously punctate; pistil 5.2-5.8 mm long, gla- brous, the ovary ovoid, 1.1-1.2 mm long, the style 4.44.7 ovules 31 to 36. Fruits globose, 8-11 mm diam.. mm long, inconspicuously punctate, the prominently punctate and punctate-lineate. Distribution. Ardisia dunlapiana is endemic to the Osa Península, Puntarenas, Costa Rica, growing from sea level to 900 m in elevation. Ecology and conservation status. Ardisia dun- lapiana occurs along water courses in lowland wet forests. Because it thrives in medium light exposure conditions, it is resilient enough to stand some in- tervention, and is locally common in some areas. Given that it occurs in several reserve areas, we believe this species is not threatened. Etymology. This species was named in bones of Albert Atkinson Dunlap, a mycologist. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia dunlapiana is most close ‘ly related to А. awarum but differs by its leaf blades mem- branous, calyx chartaceous, corolla membranous with its tube sparsely furfuraceous-lepidote outside, longer anthers to 2.7 mm long, and longer styles to 4.7 mm long. Specimens examined. COSTA RICA. Puntarenas: Forestal Golfo Dulce, Rancho Quemado, 17 Oct. 1991 (fr). to the leaf R. Aguilar 565 (CR, INB, MO); Cantón de Golfito, Coastal A Range, Fila Cruces, headwaters of Río Piedras Blan- as, 12 May 1994 (fl), R Aguilar & F. Quesada 3265 (C F ГС, INB, MO); Reserva Forestal Golfo Dulce, Osa Pe ínsula, Rancho Quemado, ca. 15 km W of oar e Río Riyito, 2 June 1988 (fl), B. Hammel et al. 16974 (CR, F, FTG, LL, MEXU, MO, NY, US); Cantón de Golfito: Reserva Рожни) Golfo Dulce, Osa Península, Playa Cam- panario or San Josecito, Sierpe, 20 June 1991 (fl), P. Har- mon 242 (INB, MO): Parque Nacional Corcovado, Pavo Forest, 14 July 1988 (fl). C. Kernan 695 (CR, FTG, MO, US): Parque Nacional Corcovado Sirena, Los Bs Кары. 8 July 1989 (fl), C. Kernan 1216 (CR, FTG, MO); Cantón de ( Golfito, Rancho Quemado, forestry station, Eloy Cubero (fr Marín 107 (CR MO); near . W of Rincón de Osa, 8 Aug. 1967 (fl), P. Raven 21643 (G, MO); Parque Nacional Corcovado, S. Esquinas Golfito, Golfo Dulce ca. Río Es s as, Estación Esquir Ee с. June 1993 (fl), M. Segura & F. Quesada 93 (FTG, 3, MO); Cantón de Golfito Parque TAM ional Cor- a, EIN de Coto Colorado, Estación Esquinas, Sec- ción Esquinas, 25 Aug. 1993 (fl), M. Segura & F. Quesada МО); Cantón de Golfito Parque Nacional Corcovado, Valle de C Yn C a near Estación Esqui- nas, 8 Oct. 1993 (fr), M. Segura & F. irc 189 (INB, MO): Osa ее, тн па, 3.5 km W of Rincón, at n N of the house of Don Quecho, i ‘of BOSCOSA sta- on, 21 June 1993 (fl), K. Thomsen 399 (C, FTG), 6 Oct. 1994 (f ). А. Thomsen 1056 (C, MO). —. — — 40. Ardisia dwyeri Lundell, Wrightia 4: 145 1970. Auriculardisia dwyeri (Lundell) Lundell, Phytologia 49: 344. 1981. TYPE: Panamá: Cerro Jefe, roadside thicket, 2900 ft. [884 m], 20 Aug. 1967 (fl), J. Dwyer & S. Hay- den 8082 (holotype, LL!, F neg. 55651!, LL neg. 71-1675 isotypes, GH!, MO!, US!). Figure 42 Panama. Ardisia conglobata Lundell, Wrightia 6: 70. 1979. Syn. nov. Auriculc ME deer (Lundell) Lundell, EA 49: : 1981. TYPE: Panama. Panamá: Llano-Cartf Шы 9.8 km from Inter-American Hia : 100—1200 ft. [335-366 m], 28 Dec. 1974 (fl). S. Mori, J. Kallunki & B. Hansen 4172 (holotype. LL!, b neg. 55621!; isotype, MO!). — — T— 1.5-8 m tall, Branchlets slender, terete, 4—5 mm diam.. Shrubs or small trees 6-10 cm diam. densely furfuraceous-lepidote. Leaves with blades coriaceous, elliptic, 4—15.4 X 1.1-5.7 em, apically acute to acuminate, with an acumen 3-6 mm long. basally cuneate, decurrent on the petiole, incon- spicuously punctate and punctate-lineate, glabrous above, densely to sparsely furfuraceous-lepidote below, the midrib impressed above, prominently raised below, the secondary veins 24 to 32 pairs, indistinct to. prominulous above and below, the margins entire, flat; petioles slender, marginate, 5— 10 mm long. glabrous above, densely to sparsely furfuraceous-lepidote below. Inflorescences erect, 4-9 x 2.5-5 pinnately to bipinnately paniculate, 257 Ricketson & Pipoly Volume 90, Number 2 2003 Revision of Ardisia subg. Auriculardisia ((OW) 10 ((OIN) 86E иоәт әр 7) = әлә әр °<) *ad&jost шоу имвлр э-ү) "maj onjeurui 77)— Moy *g— "qouerq Зицәмоүд 'V— `рурәипәпә pisipiy (їч) ep әл] g un ung “9 woaz 7) (TT) 2808 чәрХәң `$ o Anq f ‘edAjojoy шолу umeip g é V ) эшл] 7)— uewop4 'g— "qoueaq #ппәмо[] "ү ‘amp pisipiy— "(yep cr әү Annals of the Missouri Botanical Garden cm, pyramidal, usually shorter than or as long as the leaves, densely cupuliform and furfuraceous- lepidote, the branches loosely congested into 5- to 11-flowered corymbs; peduncle 1-1.8 ст long, the lower branches subtended by leaves; inflorescence bracts unknown; inflorescence branch bracts ca- ducous, membranous, often foliaceous. oblong, 1.2— 9.7 X 4.9-10.4 mm, apically acute, inconspicu- ously punctate and punctate-lineate, glabrous above, densely furfuraceous-lepidote below, the midvein inconspicuous or impressed above, prom- inently raised below, the secondary veins indistinct, the margins entire, flat; floral bracts caducous, membranous, ovate, 2—5.7 X 2.4-3.1 mm, apically acute, prominently punctate and punctate-lineate, glabrous above, furfuraceous-lepidote below, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; pedicels stout, 4—5 mm long. inconspicuously punctate and punctate-lineate, fur- furaceous-lepidote. Flowers 5-merous, intense white, light pink to purple; calyx lobes coriaceous, 1.1-1.5 prominently punctate and punctate-lineate, gla- oblate, 2.8—3.2 mm, apically rounded, brous adaxially, furfuraceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely 6.3-1.5 mm long, the tube 1.3-1.5 mm long, the lobes ovate, 5-6 X nently punctate and punctate-lineate, glabrous ada- glandular ciliolate; corolla chartaceous, 2.3-3.3 mm, apically acute, promi- xially, furfuraceous-lepidote on the tube and me- dially abaxially, the margins entire, hyaline; stamens 5-6 mm long, the filaments 2.5-3 mm long, the staminal tube 0.8-1.5 mm long, the api- cally free portions 1.5-1.8 mm long, the anthers lanceoloid, 3.2-4.2 X 1.2-1.4 mm, apically apic- ulate-cuspidate, basally cordate, the connective conspicuously punctate; pistil 5.8-6 mm long, gla- brous, the ovary globose to oblong, 0.9-1 mm long, prominently punctate, the style 4.9-5 mm long, conspicuously punctate, the ovules 24 to 27. Fruits globose, 8-9.8 mm diam., prominently punctate. Distribution. Ardisia dwyeri is most common on Cerro Jefe and along the El-Llano—Cartf Road, in Panamá, but also has a disjunct population along the Santa Rita Ridge in Colón, Panama, growing from 335 to 1007 m in elevation. Ecology and conservation status. Ardisia dwyeri occurs in roadside thickets and premontane and cloud forests. Because of its restricted distribution it should be considered threatened. Etymology. This species was named in honor of John Dwyer, curator at the Missouri Botanical Garden and author of the Rubiaceae in the Flora of Panama. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia dwyeri is most closely related to A. vesca because of its short calyx lobes, long pedicels, thin branchlets, wide corolla lobes, and long an- thers. However, Ardisia dwyeri can be separated from A. vesca by its coriaceous leaf blades, shorter pedicels to 5 mm long, coriaceous much wider ca- lyx lobes to 3.2 mm wide with rounded apices, chartaceous, longer and wider corolla lobes to 6 X 3.3 mm, longer and wider anthers to 4.2 X 1.4 mm, longer styles to 5 mm long, and larger fruits to 9.8 mm in diameter. The type of Ardisia conglobata is unique only for its small axillary inflorescence of congested flower heads, but because it is in bud, this is not signifi- cant. In all other respects it matches the type of A. dwyeri. or лаз PANAMA. Colón: Santa Rita Ridge, 8 f Transisthmus Hwy., 28 July 1968 (fl), J. Dw n et M 8999 (F ‘TG [2]. MO, NY). Panamá: Cerro Jefe, slopes beyond radio tower, 3 Nov. 1985 (fl, fr), G. Mc Pn 7425 (LL. MO); El Llano-C arii Road, from InterAmerican Hwy, 4 Oct. 1974 (fl, fr), J. Kallunki 2245 (LL, MO); Dtto. Panamá, Cerro Jefe, 19 b 1984 (fl), J. Pipoly & R. Bethancourt 7043 (MO, NY, A), 10 July 1976 (fr), G. Sullivan 201 (MO). 41. Ardisia eucuneata (Lundell) Pipoly & Ric- ketson, Sida 18: 512. 1998. Auriculardisia eu- cuneata Lundell, Phytologia 57: 449. 1985. Ardisia eucuneata (Lundell) Lundell, Phytolo- gia 6l: 7, nom. inval. TYPE: Panama. San Blas. Nusagandi, trail from camp NW to a quebrada, 09?19'N, 078°15’W, 300 m, 31 July 1984 (fl), G. de Nevers & C. de León 3598 (holotype, LL!; isotype, MO!). Figure 43. Shrubs 2—2.5 2-3 mm diam., densely rufous furfuraceous-lepi- m tall. Branchlets slender, terete, dote, glabrescent. Leaves with blades membranous, oblong to narrowly oblong to narrowly oblanceolate, 4.1-12.4 X 0.9-3.1 cm, apically long acuminate to caudate, with an acumen 10—18 mm long, basally prominently punctate and punctate-lineate above and below, cuneate, decurrent on the petiole, mostly glabrous above, densely rufous furfura- ceous-lepidote, the midrib impressed above, prom- to 21 pairs, prominulous above and below, the margins inently raised below, the secondary veins 15 entire, flat; petioles slender, canaliculate, 4—7 mm long, glabrous above, furfuraceous-lepidote below. Inflorescences erect, bipinnately paniculate, 3-9 х 2—6 cm, pyramidal, shorter than the leaves, the pe- duncle, branches, and pedicels densely furfura- ceous-lepidote, the branches loosely congested into 3- to 7-flowered corymbs; peduncle obsolete to 8 Volume 90, Number 2 Ricketson & Pipoly 259 Revision of Ardisia subg. Auriculardisia mm long, the lower branches subtended by leaves: inflorescence bracts unknown; inflorescence branch bracts caducous, 1.5-2.4 X 0.4—0.6 mm, apically acute, glabrous adaxially, fur- furaceous-lepidote, the margins irregular, minutely membranous, oblong, erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflorescence branch bracts, but 0.4—0.7 X 0.2-0.3 mm; pedicels slender, 3.6— 4.8 mm long. prominently punctate and punctate- lineate, mixed cupuliform and furfuraceous-lepi- dote. Flowers 5-merous, light pink or purple: calyx lobes membranous, 1.4— 1.6 х 0.8-1 mm, apically acute, prominently punc- tate and punctate-lineate, glabrous adaxially, fur- ovate to narrowly ovate, furaceous-lepidote, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 3.5-3.6 mm long, the tube 0.5—1 mm long, the lobes narrowly ovate, 2.4—3.1 X 1.7-1.8 mm, apically acute, prominently punctate and punctate-lineate, glabrous adaxially, furfuraceous- lepidote abaxially, the margins entire, hyaline: sta- mens 2.8-2.9 mm long, the filaments 1.3-1.6 mm long. the staminal tube 0.4—0.6 mm long, the api- cally free portions 0.9-1.2 mm long. the anthers narrowly ovoid, 1.6-1.7 X 0.6—0.7 apiculate-cuspidate, basally cordate, the connective mm, apically conspicuously punctate; pistil 4.2-4.3 mm long. glabrous, the ovary ovoid, 1—1.1 mm long. the style 3.1-3.3 Fruits unknown. mm long. epunctate, the ovules 7 to 9. Distribution. Ardisia eucuneata is endemic to the Nusagandi area in San Blas, Panama, growing at 300 to 400 m in elevation. Ecology and conservation status. Ardisia eucu- neata occurs in river flood plains. Because of its restricted distribution, it should be considered threatened. Etymology. The specific epithet was derived from the Greek “eu” meaning well, good, thorough- ly, completely, or truly and “-cuneate” referring to the cuneate leaf bases. ‘ithin Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia eucuneata is most easily confused with A. atropurpurea (see under that species for similarities). However, Ardisia eucuneata is easily distinguished by its branchlets to 3 mm in diame- ter, pedicels to 4.8 mm long, corolla lobes to 3.1 х 1.8 mm, anthers to 1.7 X 0.7 mm, and the styles to 3.3 mm long. Specimens examined. PANAMA. San Blas dort di, Wedar Trail, 19 July 1986 (fl), J. MeDénaah et al. (LL, MO 42. Ardisia fimbrillifera Lundell. Wrightia 4: 180. 1971. Auriculardisia rp e dell) Lunde ytologia 49: : TYPE: Costa Ries. Heredia: ас. 700 ft. [213 m]. 2 June 1971 (fl, fr), G. Proctor 32238 (holotype, LL!, F neg. 55663!; isotypes, F!, LL!). Figure 44. Lun- 1981 Trees 3—25 m tall, 4.2-30 ст diam. Branchlets slender, terete, 4.5-7.5 mm diam., densely furfu- raceous-lepidote. Leaves with blades coriaceous, oblong elliptic to oblanceolate, 17.2—34.8 X 4-9.3 cm, apically acuminate, with an acumen 6-21 mm long, basally acute, decurrent on the petiole, in- conspicuously punctate and punctate-lineate above, prominently punctate and punctate-lineate below, sparsely furfuraceous-lepidote above, densely fur- furaceous-lepidote below, the midrib impressed above. prominently raised below, the secondary veins 33 to 45 pairs, inconspicuously raised above and below, the margins entire, inrolled; petioles slender, marginate, 6-15 mm long, 2-3 mm diam., sparsely furfuraceous-lepidote above, densely fur- furaceous-lepidote below. Inflorescences erect, tri- pinnately paniculate, 11-37 X 7-27 cm, pyrami- dal, usually longer than the leaves, the rachis, branchlets, abaxial bract surfaces, and pedicels furfuraceous-lepidote; the branches congested to loosely congested into 5- to 11-flowered corymbs; peduncles 1.6-3.9 cm long, the lower branches subtended by leaves; inflorescence bracts unknown; inflorescence branch bracts unknown; floral bracts ovate, 0.9-1.4 X 0.9-1.5 prominently punctate and caducous, membranous, mm, apically acute, punctate-lineate, glabrous above, furfuraceous-lep- idote below, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; pedicels stout, 2-3.5 inconspicuously punctate and punctate-lineate, densely furfuraceous-lepidote. mm long, Flowers 5-merous, white or light yellow; calyx lobes coriaceous, suborbicular, 2-2.8 X 2.2-3 mm, api- cally rounded, prominently punctate and punctate- lineate, glabrous adaxially, sparsely furfuraceous- lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla chartaceous, 6.2-6.9 mm long, the tube 1.4-1.9 mm long, the lobes elliptic to lanceolate, 4.5—5.2 х 2.3-3 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the mar- gins entire, hyaline; stamens 5—5.4 mm long, the filaments 2.6-2.8 mm long, the staminal tube 0.8— 1.2 mm long, the apically free portions 1.4—2 mm long, the anthers ovoid, 2.6-2.9 X 1.3-1.5 mm. apically apiculate, basally deeply cordate, the con- nective conspicuously punctate: pistil 6.4—7.3 mm long, glabrous, prominently punctate and punctate- 260 Annals of the Missouri Botanical Garden Су М РҮ, d: M \ js 5 UNS yY qi So Wart GN { ть. ^ Б УИ Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia lineate, the ovary oblong, 1.4-1.6 mm long, the mm long, prominently punctate and punctate-lineate, the ovules 46 to 51. Fruits glo- 7-9.5 mm diam., prominently punctate and punctate-lineate. style 5—5.7 bose, Distribution. Ardisia fimbrillifera is distributed from Río San Juan in Nicaragua through eastern Costa Rica (Alajuela, Heredia, Limón), and Pana- ma (Panamá, Veraguas, San Blas) to Chocó, Colom- bia, growing from 10 to 800(-1200) m in elevation. Ecology and conservation status. Ardisia fim- brillifera is locally infrequent in lowland tropical wet forests or, rarely, in premontane forests. It has been collected in locations that are fairly remote and many of them are protected, so we see no im- mediate threat to this species. Etymology. The specific epithet was derived from the Latin * or "fimbrill" and * referring to the calyx lobes bearing a fringe of glan- ‘fimbri” “fera” dular cilia. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia fimbrillifera is easily distinguished from A. hagenii by its narrower calyx lobes to 3 mm wide, shorter corolla lobes to 5.2 mm long, shorter anthers to 2.9 mm long, and shorter styles to 5.7 mm long. NICARAGUA. Río San Juan: Buena Vista a of the Río San Juan delta, 13 Sep. 1983 (fr), E. Martens S. & К. Riviere 2100 (MEXU, МО); near Сайо е 20 km NE of El Castillo, 18—21 Apr. 1978 (fl), D. Neill & P. Vincelli 3602 (HNMN, LL, , NY); Hío San jd 2 km E of Buena Vista, 14 Sep. 1982 (fr), R. Riviere 340 (HNMN, MO, NY, US); Mpio. del Castillo, Reserva Indio- Maíz, along Сайо el P: avón, 3 Specimens алыма, uerto, 11 May 1993 (f), F Araya et al. 321 (CR, FTG, INB, MO). Heredia: Parque Nacio- nal eg Carrillo, Estación ко Magsasay, 9 July 1990 Alcázar et al. 137 (CR, FTG, INB, MO); Finca sp Río Puerto mo E of junction with the Río Sarapiquí, Holdridge ee 3000 m line, 18 Feb. 1981 (fl), J. Folsom 9025 (CR, DUKE); Horquetas, 1-12 km SW of Vie gc on f: to Finca Plastico, near Finca ‚ 21 Apr. 1988 (fl), B. Hammel & Limón: Cantón de Talamanca, Plastico and Rar R. Robles 16714 (MO). Bratsi, Alto Lari, between Surayo and Dapari 50 m N of Río Dapari mouth, Pare, at Río Lari, 25 Feb. 1992 (fr). R. Aguilar & H. Schmidt 940 (INB, MO); Refugio Nacio- nal de Fauna Silvestre Barra del Colorado, Llanura de Tortuguero, Puerto Lindo, 24 July 1995 (fr), F Araya 792 и F, FTG, INB, MO); Hacienda Tapezco-Hda. La Suer- 29 air km W of Tortuguero, 10 Mar. 1978 (fl), C. Da- е et al. 6835 (F, MO); Cantón de Pococi, Parque Na- : a de Tortuguero, Estación Agua — Fría, Sendero Agua Fría hasta entrada Sendero Aguacate 1 Dec. 1990 (fr), J. Solano 274 (CR, FTG, INB, MO) NAMA. Panamá: Cerro Jefe, near n Indio, 17 Feb. 1968 (fl), J. Duke 15229 (MO); a Indio, slopes of Cerro Jefe, 18 Oct. 1971 (fr), A. Gentry "2150 (MO); on El al. 665 (MO, PMA). Veragu Santa Fé, ca. N of Escuela Agrícola Alto ^an Piedra, 17 Oct. 1974 (fr). S. Mori & J. Kallunki 2591 (MO); NW of Santa Fé, 1.8 km N of Escuela Agrícola Alto de Piedra, 23 Feb. 1975 (0), S. Mori & J. Kallunki 4775 (LL, МО); 7 Oar W of Santa Fé on new road past agricultural school, 12 Apr. 197 4 (fl), M. Nee 11191 (LL, MO, US). See e có: Mpio. de Riosucio, Zona de Urabá, Cerro del Cuchillo, 15 Nov. 1987 (fr), D. Cárdenas 835 (JAUM, MO); Río San Juan Basin, near Docordó, 29 Mar. 1979 (fr), E. Forero et al. 4354 (COL, MC 43. Ardisia furfuracea Standl., J. Wash. Acad. Sci. 17: 525. 1927. Auriculardisia furfuracea (Standl.) Lundell, Phytologia 49: 344 1 TYPE: Costa Rica. Heredia: Cerro de Las La- jas. N of San Isidro, 2000-2300 m, 7 Mar. 1926 (fl), P. Standley & J. Valerio 51556 (ho- lotype, US!, LL neg. 1971-34!, US neg. 2372). Figure Ardisia duriuscula Lundell, Wrightia 7: 24. 1981. nov. Auriculardisia рта (Lundell) Lundell, Phytologia 49: 344. . TYPE: Costa Rica. He- redia and San José I: de Zurqui, along the Río Para Blanca (Pac "s drainage), 10°03'N, 084701" W, 1600-1800 m, 6-7 Feb. 1977 (fl), W. Р. С. Vis- conti & J. Gentry 10288 (holotype, Е!, F neg. 68149!; isotype Кк ә trichomata Lundell, Phytologia 63: 1987. Syn. nov. Ardisia pire: (Lundell) Li dell, Phytologia 63: 463. YPE: Costa Rica. San José: roadside РА p Alto La Palma to Bajo La Hondura, ca. 10 km NE of San Vincente de Moravia, 1260—1550 m, 24 Feb. 1978 (fl), К. Wilbur 24906 eriam DUKE!) Syn. Small trees 2—6 m tall, to 3.7 cm diam. Branch- ¢ Figure 44 (left). Ardisia fimbrillifera. —A. Flowering branch. —B. Flower. —C. Fruit. (A drawn from holotype, б. Poo: 32238 (LL); B from M. Grayum 9761 (MO); C from R. Riviere 340 (MO).) Figure 45 (right). Ardisia furfuracea. —A. Flowering branch. —B. Flower. —C. Fruit. (А & B drawn from holotype, P. Standley & J. Valerio 51556 (US); C from G. Mora et al. 291 (MO).) 262 Annals of the Missouri Botanical Garden lets flexuous, stout, terete, 5-9 mm diam., with a mixture of densely cupuliform and furfuraceous- lepidote scales. Leaves with blades coriaceous, ob- long to elliptic, 16.2-47.8 X 3.8-13.2 cm, apically acuminate, with an acumen 0.3-1.6 cm long, ba- sally acute to cuneate, decurrent on the petiole, inconspicuously punctate and punctate-lineate, gla- brous above, below with a mixture of densely cu- puliform and furfuraceous-lepidote scales, the mid- rib impressed above, prominently raised below, the secondary veins 47 to 64 pairs, prominulous above and below, the margins entire, flat; petiole stout, marginate, 0.7-2.2 ст long, glabrous above, below with a mixture of densely cupuliform and furfura- ceous-lepidote scales. Inflorescences erect, bi- or tripinnately paniculate, 17.4—46.7 X 15.2-29.6 cm, pyramidal, as long as or longer than the leaves, with a mixture of densely cupuliform and furfura- ceous-lepidote scales, the branches terminally con- gested into 5- to 8-flowered corymbs; peduncle nearly sessile to 2.7 em long, the lower branches subtended by leaves; inflorescence bracts unknown: inflorescence branch bracts early caducous, mem- d to chartaceous, ovate to oblong, 1.3-1.8 0.4—0.5 cm, apically acute to rounded, or nearly so, with scattered prominent punctations and punctate- lineations, glabrous above, below with a mixture of densely сири огт and furfuraceous-lepidote scales, the midrib inconspicuous, the secondary veins obscure, the margins entire, flat; floral bracts early 0.8-1.3 0.8-1.3 mm, apically acute, prominently punctate caducous, membranous, ovate, and punctate-lineate, glabrous above, scattered fur- furaceous-lepidote below, the margins irregular, mi- nutely erose, hyaline, sparsely glandular-ciliolate: pedicels stout, 0.8-2.1 mm long, inconspicuously punctate-lineate, densely furfuraceous-lepidote. Flowers 5- or 6-merous, white; calyx lobes charta- ceous, orbicular to oblate, 1.3-1.6 X 1.4-1.7 apically rounded, often prominently and inconspic- uously mm, punctate and punctate-lineate, glabrous adaxially, densely furfuraceous-lepidote abaxially, the margins irregular, minutely hyaline, sparsely glandular-ciliolate; corolla membranous. 5-5.3 mm long, the tube 0.7-1.4 mm long. the lobes ovate, 3.9—4.3 X 2.3-2.5 mm. apically acute, sparsely prominently punctate and punctate-lin- eate, glabrous throughout, the margin entire, hya- line; stamens 5.8—6 mm long, the filaments 3.9—4.1 mm long, the staminal tube 0.6-0.7 mm long, the apically free portions 3.2—3.5 mm long, the anthers ovoid, 2-2.2 X 0.8-1 mm, apically apiculate, ba- sally cordate, the connective inconspicuously punc- erose, tate; pistil 5.6-6.5 mm long, glabrous. the ovary BB : oblong, 1.1-1.3 mm long, the style 4.5-5.2 mm long, inconspicuously punctate and. punctate-lin- eate, the ovules 39 to 50. Fruits depressed globose, 4.2-5.6 X 5.5-7.2 mm, conspicuously punctate. Distribution. Ardisia furfuracea is endemic t the mountains of central Costa Rica in Cartago, He- redia, and San José, growing at 1260 to 2000 m in elevation. - Ecology and conservation status. Ardisia furfu- racea occurs in premontane to montane wet forest and often in remnant fragments of forests found in pastures. Although it is found over a wide area, it is known from a relatively small number of collec- tions and should be considered threatened at this time. Etymology. The specific epithet means scurfy, covered with bran-like scales referring to the ves- titure throughout the plant. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia furfuracea is most closely related to A. mephersonii by virtue of the flexuous branchlets and oblong to narrowly elliptic leaf blades. How- ever, Ardisia furfuracea can be distinguished from А. mcphersonii by its narrower calyx lobes to 1.7 mm wide and wider corolla lobes to 2.5 mm wide. longer stamens to 6 mm long, smaller anthers to 2.2 X | mm, and longer styles to 5.2 mm long. It appears that Lundell (1981c, 1987) did not note the similarities between Ardisia furfuracea and populations represented by the types of Ardisia du- riuscula and Auriculardisia trichomata. The types of Ardisia duriuscula and Auriculardisia trichomata are identical to that of Ardisia furfuracea in all re- spects. Specimens ср COSTA RICA. Cartago: Тарап- tí Reserve, 7 Dec. жаз L. Gómez ни L [2]. МО); Canton de Pz ARE on > Nacional Tapanti, Valle del leventazón, Sector Dos Amigos, Чече һо ge Trail, 20 July 1994 (fr), G. Mora et al 3 1 (INB, ). Heredia: Parque Nations] Braulio Carillo Transect Tit to Pon of trail, 20 Tus 1992 (ster), B. Boyle & C. Godt 932 FTG, INB y Parque National Braulio meo Ттап- sect Trail, s їр ^ of trail, ca. 50 minute walk gd Ws 2 Sep. 1992 (fr). B. Boyle & N. Snow 1052 (Е МО); along B wy Rafael, Atlantic slope of En 'án CMM va, 12 Арг 1986 (fl). М. Grayum 7060 (CR, ЕТС MO, US). San ce along C R 220 between Alto de ja donus and Bajo La Hondura, 5-8 km N of ys come 17 Oct. 1974 (fl). J. Utley & K. Utley 1422 , F). 44. Ardisia generalensis Ricketson & Pipoly. sp. nov. TYPE: Costa Rica. San José: basin of El General, 675—900 m, Mar. 1942 (fl), A. Skutch 5025 (holotype, US!: isotype, MICH!). 46. Figure ropter laminam ellipticam vel obovatam, atque inflo- rescentiam bipinnatipaniculatam, flores dense corymbosos Volume 90, Number 2 2003 Ricketson & Pipol 263 y Revision of Ardisia subg. Auriculardisia gerentem A. darienensi arcte similis, sed ab ea petiolis canaliculatis (non marginatis) 4—8 (nec 8-28) mm longis 0.5-1 (nec 2-3) mm diametro, laminis foliaribus membra- naceis iam chartaceis) —(nec 55 ad 70-) jugis necnon lobulis calycinis ad apices acutis (nec b ac perfacile recognoscitur. nervis secundariis 21 ad 25 Trees to 6 m tall. Branchlets slender, terete, 3—4 mm diam., densely and minutely appressed rufous furfuraceous-lepidote. Leaves with blades membra- nous, elliptic to narrowly obovate, 4.8-11.5 X 1.7— 2 cm, apically acute, with an acumen 4—7 mm long, basally acute to cuneate, decurrent on the petiole, inconspicuously punctate and punctate-li- neate above and below, glabrous above, densely and minutely appressed rufous furfuraceous-lepi- dote below, the midrib impressed above, promi- nently raised below, the secondary veins 21 to 25 pairs, prominulous above and below, the margins entire, flat; petioles slender, canaliculate, 4—8 mm long, 0.5-1 mm diam., glabrous above. /nflores- cences erect, bipinnately paniculate, 4.8—7.2 X 3— 4.5 em, columnar to narrowly pyramidal, longer than the leaves, the rachis, branchlets, and pedicels densely furfuraceous-lepidote, the branches rarely subtended by leaves, the branches terminating in - to 7-flowered corymbs; inflorescence bracts and floral bracts caducous, ovate, 1.2-1.8 X 0.6-1.1 mm, cally acute, inconspicuously punctate and punc- EM bracts unknown; membranous, api- tate-lineate, glabrous throughout, the margins irreg- ular, minutely erose, hyaline; pedicels slender, 3.6- 5.1 cm long, nearly epunctate. Flowers 5-merous, white or pink; calyx lobes coriaceous, ovate, 3.4— 3.7 X 3.2-3.4 mm, apically acute, mostly epunc- tate, glabrous adaxially, furfuraceous-lepidote aba- xially, the margins chartaceous to coriaceous, 7.9-8.1 mm long, the entire, hyaline; corolla tube 3.3-3.6 mm long, the lobes narrowly ovate to lanceolate, 4.4—4.7 X 2.2-2.5 mm, apically acute. mostly epunctate, glabrous throughout, the margins entire, hyaline; stamens 6 .2 mm long; the fil- ament 3.3-3.4 mm long, the staminal tube 2.1-2.3 mm long, the apically free portions 1—1.3 mm long. the anthers narrowly ovoid to lanceoloid, 3—3.1 X 1.2-1.3 mm, apically apiculate, basally lobate, the connective inconspicuously punctate; pistil 8—8.3 mm long, glabrous, the ovary oblong, 2.6-2.7 mm long, the style 5.4—5.6 mm long, epunctate, the ovules 25 to 29, Fruits unknown. Distribution. Ardisia generalensis is endemic to the basin of El General in San José, Costa Rica. growing from 675 to 900 m in elevation. Ecology and conservation status. Ardisia gene- ralensis occurs in wet forests. The area contains a number of Darién and Chocó floristic elements. such as Cybianthus pastensis (Mez) Agostini and Cybianthus montanus (Lundell) Agostini. Because it is known only from two old collections, it should be considered threatened. Etymology. The specific epithet refers to the lo- cality of El General where the type was collected. thin. Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia generalensis is most similar to A. darienensis (see under that species for similarities). However, it is easily separated by the shorter and thinner, canaliculate petioles to 8 X 1 mm, mem- branous leaves with less numerous secondary veins to 25 pairs, and the apically acute calyx lobes. Paratype. COSTA RICA. San José: basin of El Gen- eral, Aug. 1945 (fr), A. Skutch 5237 (F, MO, NY, US). 45. Ardisia gigantea Ricketson & Pipoly, sp. nov. TYPE: Panama. Veraguas: NW of Santa Fé, 8.8 km from Escuela Agrfcola Alto de Pie- dra, Pacific slope, 21 Dec. 1974 (fr). S. Mori, J. Kallunki, T. Cochrane, B. Cochrane, B. Han- sen, R. Kowal & M. Nee 4003 (holotype, MO!; isotype, LL!). Figure 47. Propter folia magna ad apices ao ша ا‎ minute adpresso-furfuraceo-lepidota atque laceos secus margines erosos cum vii aguirreana primo intuitu и Wem sed ab ea epi, foliaribus 45.8—46.6 X 15.7-21.5 (non 51.5-67.0 X 7.5) cm manifeste habe inconspicue) punctatis et ineat puncta necnon lobulis we oblatis (non ovatis) 1.1-1.€ 3.3-3.6 (nec 2.1-2.6 2-1.4) cm ad apices r (nec acutis) perfacile Ri oed Trees to 8 m tall, to 30 cm diam. Branchlets stout, terete, 8—10 furfuraceous-lepidote. Leaves with blades membra- nous, elliptic, 45.8—46.6 х 15 acute, with an acumen 1.8-2.1 cm long, basally mm diam., densely appressed rufous .7-21.5 ст, apically acute, decurrent on the petiole, prominently punc- tate and punctate-lineate, glabrous above, furfura- ceous-lepidote, the midrib impressed above, prom- inently raised below, the secondary veins 48 to : pairs. prominently raised above and below, the mar- gins entire, inrolled; petioles slender, marginate, 1.2-1.7 5 mm diam., glabrous above, furfuraceous-lepidote below. Inflorescences erect, cm long, 4—5 bi- to tripinnately paniculate, to 32 X 25-28 cm, pyramidal, shorter than the leaves, the rachis, branches, and pedicels furfuraceous-lepidote, the branches loosely congested into 3- to 7-flowered corymbs; peduncle obsolete; inflorescence and branch bracts unknown; floral bracts unknown: pedicels slender, 9-13 mm long, inconspicuously punctate and punctate-lineate, furfuraceous-lepi- dote. Flowers 5-merous, color unknown; calyx lobes 1.1-1.6 х 3.3-3.6 mm, chartaceous, oblate, api- 264 Annals of the Missouri Botanical Garden Volume 90, Number 2 2003 Ricketson & Pipoly 265 Revision of Ardisia subg. Auriculardisia cally rounded, prominently punctate and punctate- lineate, glabrous adaxially, sparsely to densely fur- furaceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular cilio- late; corolla, stamens, and pistil unknown. Fruits globose, 9.5-13.5 cm diam., prominently punctate. Distribution. Ardisia gigantea is known only from the type collection, from near the Escuela Agrícola Alto de Piedra in Veraguas, Panama, growing at 500 to 1000 m in elevation. Ecology and conservation status. The species occurs in the premontane wet forest zone, but its population biology and consequent conservation status are unknown. Etymology. The nod is based on the excep- tionally large leaf blades Within Ardisia subg. ү тн sect. Pal- manae, Ardisia gigantea is most similar to А. agui- rreana, but is distinguished by its larger leaves to 46.6 cm long, and distinctive oblate calyx lobes. Ardisia gigantea is known only from a fruiting spec- imen: the holotype is of a single leaf with an at- tached inflorescence. The LL isotype consists of a single leaf. 46. Ardisia glandulosomarginata Oerst., Vi- densk. Meddel. Dansk Naturhist. Føren Kjøbenhavn 1861: 128. 1862. Auriculardisia киен (Oerst.) Lundell, Phyto- 28 3. TYPE: Costa Rica. Car- tago: in monte a 8000-9000 ft. |2438— 2743 mj], Jan. 1847 (fl, A. Oersted 25 (holotype, C!, Е neg. 22948!; isotypes, С!, F!, LL!, M!). Figure 48. logia 51: "n leptopoda Lundell, Phytologia 57 198 nov. agen leptopoda (Lundell) L wade n Phytologia 61: 1986, nom. inval. Ardisia lepto- үз (Lundell) us & Ricketson, Sida 18: 513. I anama. Chiriquí: К of Guade ойе е the Río Chiriquí Viejo, ca. 2 mi. NE of Cerro Punta, ridge of vin a i 7000 ft. [2134 m], 13 Jan. 1971 (fl, fr), &. bur, J. Teeri & R. Foster 13111 ты, F!, F neg. 68323!; isotypes, KE!, LL!, MO!, US). Shrubs or small trees 1.5-12 m tall, 7.6-18 cm diam. Branchlets slender, terete, 3—6.5 mm diam., densely appressed ferrugineous furfuraceous-lepi- dote. Leaves with blades membranous to nearly co- riaceous, elliptic, 4.2724.2 X 2.6-8.3 cm, apically acuminate, with an acumen 0.8-2.1 cm long, ba- sally obtuse or acute, decurrent on the petiole, prominently punctate and punctate-lineate, with a few scattered scales above, densely and minutely appressed ferrugineous furfuraceous-lepidote be- ow, the midrib impressed above, prominently raised below, the secondary veins 37 to 42 pairs, prominulous above and below, the quaternary veins prominulous below, the margins nearly entire to deeply crenulate, flat; petioles slender, canalicu- 11-19 mm glabrous above, densely ferrugineous furfuraceous- late, long, 1-2 mm diam., essentially lepidote below. /nflorescences erect, bi- to tripin- nately paniculate, 4-20 X 4—22.5 cm, pyramidal, usually longer than the leaves, the rachis, branch- lets. abaxial surfaces of all bracts, and pedicels densely ferrugineous furfuraceous-lepidote, the branches loosely congested into 5- to 11-flowered corymbs; peduncle nearly obsolete to 2.1 cm long, the lower branches often subtended by leaves; in- florescence bracts absent; inflorescence branch bracts very early caducous, membranous, ovate, 1.9-3.2 X 0.5-1.5 mm, apically acuminate, prom- inently punctate and punctate-lineate, glabrous adaxially, ferrugineous furfuraceous-lepidote aba- xially, the margins irregular, minutely erose hya- line, sparsely glandular ciliolate; floral bracts sim- ilar to the inflorescence branch bracts, but 1.2-1.8 X 0.4—0.6 mm; pedicels slender, 7.1—8.7 mm long, prominently punctate and. punctate-lineate, ferru- gineous furfuraceous-lepidote. Flowers 5-merous, light green, white or cream to light pink; calyx lobes membranous to chartaceous, widely ovate to orbicular, 2.3-2.5 X 2.3-2.5 mm, apically acute to obtuse, prominently punctate and punctate-lineate, glabrous adaxially, ferrugineous furfuraceous-lepi- dote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 6.2-6.4 mm long, the tube 1.2-1.5 х 3.2-3.8 mm, apically acute, prominently punctate and punctate- mm long, the lobes ovate, 4.5—5 lineate, glabrous throughout, the margins entire, hyaline; stamens 4.5—4.7 mm long, the filaments 2.2-2.4 mm long, the staminal tube 0.7—1 mm long, the apically free portions 1.4—1.5 mm long, the an- thers narrowly ovoid to lanceoloid, 2.6-2.8 X 1.2- € Figure 46 (left). 5025 (MICH).) Figure 47 (right). Ardisia gigantea, —A. Ardisia generalensis. —A. Flowering branch. —B. Flower. (A, . Flowering branch. B drawn from holotype, A. Skutch it. (А & С drawn from holotype, S. Mori et al. 4003 (MO); B from isotype S. Pun el a put (LL).) Annals of the 266 Missouri Botanical Garden ((OW) 9802 ‘P 12 ddvuy "s шолу 7) (OW) ГЕР То Ja рш "N шолу g (ON) 800Z “aR uoa ‘H X uaF uoa 7) ‘ad Ajojoy шолу umeap ү) n1] у— ләмо]] "g— ‘әиел Зицәмор ү сттиәЎрү visipay ausu) ep әлү оролда ^y x M voquoyy `g шол g (9) CZ P2181200 COW) T8101 24.1 M X espravg “9 шолу 7) (OW) 201 'y "ed&jo[oq шолу umeap y) "map у— JMOL] "g— Yuraq Zurowop4 "ү -pipuizipuiosompup]2 msipay — "(op gp әлү = ^ T Ф, Volume 90, Number 2 2003 Ricketson & Pipol 267 y Revision of Ardisia subg. Auriculardisia 1.3 mm, apically apiculate-cuspidate, basally deep- ly cordate, the connective conspicuously punctate: pistil 6.2-6.4 mm long, prominently punctate, gla- brous, the ovary oblong, 1.4-1.5 mm long, the style 4.9—5 mm long. prominently punctate and punc- tate-lineate, the ovules 29 to 33. Fruits globose, 6.8-9.2 mm diam., prominently punctate. Distribution. Ardisia glandulosomarginata is found in all provinces of Costa Rica except Gua- nacaste, and is common from the Costa Rican bor- der to central Panama, growing from 1075 to 3050 m in elevation. Ecology and conservation status. dulosomarginata normally occurs in very wet cloud to elfin forests with occasional populations at the upper limit of premontane and montane forests. It has been collected in areas with remnant forests, and apparently has some tolerance for disturbance. Therefore, we do not believe it is threatened at this Ardisia glan- time. Etymology. The specific epithet refers to the prominently raised punctations along the abaxial leaf margin. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia glandulosomarginata can be sepa- rated from A. croatii by the ovate calyx lobes to 2.5 mm, with acute to obtuse apices, the early cadu- cous, ferrugineous furfuraceous-lepidote scales, and the secondary veins prominulous above and below. The type of Auriculardisia leptopoda is in bud and notable only for its petioles that are slightly longer than the median for the species. COSTA RICA. Alajuela: Coli- Specimens examined. iau мы, nas de San Pedro de San Ramón, 25 Oct. Brenes 4495 (F, NY); Volcán Pods, S slope of cr 6 Apr. 1930 (fl), G. Guidoni А (G, W); Cantón s Mus Ruíz, gei 17 Jan @, r Smith Н-145 (F). € e SW slope of Хоса ап 3385А (Е); Rancho Flores, 22 Feb. 1890 (fl). A. Tonduz 2143 (BR); Finca La Selva, the OTS Field Station on the x br hin се E of its junction Ma, the Río Sa- ra f o Tres Rios at 2500 m 7 Jan. 1995 (i. n ы. 63199 (DUKE, MO). 1853 "Cordillera Ta- lamanca, headwaters of the unnamed W branch of the Río Teribe, between the Río Sini and the Continental Divide at Cerro Bekom, along river, 21-27 Mar. 1984 (fl), G. Da- vidse et al. 25719 (LL, MO); Refugio Barra del Colorado Forests and pastures between Río Chirripocito and Río 10?38'N, 083?45' W, 19 Apr. 1990 (fl), M. Gra- Zeledón, proe Indíg m n he coil of Río Bruns a y Río Blanco, Finca San Carlos, 5 5 Apr. 1995 (fl). R. Aguilar & О. Gar- rote 3834 (CR, FTG, INB, MO); Cantón de Coto Brus, Zona Protectora Las Tablas, Cuenca Térraba-Sierpe, 8 km NE of Progreso, T Feb. 1997 (fl), B. Gamboa & A. Picado 1052 (INB, : é Parque xum. Chirripó, Cuenca Térral Bonito, 5 May 1997 (fl), R. Aguilar 5063 (INB. MO); Co- pey, June 1898 án A Tonduz 11824 (US). PANAMA. Bo- s del Toro: border of Bocas del нди, trail veh Continental Divide NE of Boq s slope on Bocas del Toro side of trail, 23 (fr). B. Pod 7363 (MO). Chiriquí: E slope of Volcán de € uu (Ваги), WNW of Boquete, 19 Nov. 1975 (fl. fr), 6. Davidse & W. D'Arcy 10184 (LL, MO); vicinity of Bo din Finca Collins, 15 Mar. 1963 (fl), W. Stern et al. 2053 (MICH, MO); ca. 3.7 km E of bridge NE of Cerro Punta on road through Bajo Grande, 9 Nov. 1980 (fl, fr), W. Stevens 18242 (MO). Panamá: Cerro Jefe, 10—13 mi. bevond Goofy Lake, 12 Feb. 1966 (fl, fr), J. Duke 8028 MO). 47. Ardisia hagenii Lundell, Wrightia 4: 59. 1968. Auriculardisia hagenii (Lundell) Lun- dell. Phytologia 49: 344. 1981. TYPE: Pana- ma. Chiriquí: Boquete region, Horqueta, [6500 ft.| 1981 m, 17 Apr. 1940 (fl). C. von eid & W. ron Hagen 2008 (holotype, MO!. LL neg 11-166; isotype, NY!). Figure 49. Auric [sape chiriquiana Lundell, lis; ee 7: 267. 1984. т. nov. Ardisia ell) Lundell, еня 61: 62. quiana (L еы роу & grae Sida 18: : 1998. TYPE: Pana í: trail up Cerro E ate Macho, s N, 082725 WW. 1500-1900 m, 7 Jan. 1983 (fl), B. Stein, B. Schmalzel & D. Roubik 1223 (holotype, LL!; isoty уре, r O!). Auric | sein toroana Lun . Wrightia 7: 273. 1984 nov. deos toroana (L undell) Lundell, Phyto- ed 61: € inval. Ardisia toroana (Lun- de 2), Pipoly & Ric uL Sida 18: 514. : Panama. Bocas del Toro: 15 km up the Chan- Ms river to І. ee dam site no. 1, near camp- site on jg to r NE of campsite, 800—900 ft [244-274 m]. 12 Dec. 1979 (fl), Т Antonio 3079 ee LLN: isotype, MO!). Shrubs or small trees to 11 m tall, to 12.5 em diam. Branchlets stout, terete, 3.5—5.5 mm diam., densely and minutely appressed rufous furfura- ceous-lepidote. Leaves with blades chartaceous to coriaceous, elliptic to narrowly elliptic. 13.5-24.6 х 3.6-7.7 cm, apically acuminate, with an acumen 0.6-1.6 cm long, basally acute, decurrent on the petiole, inconspicuously punctate and punctate-lin- eate above and below, mostly glabrous above, densely and minutely appressed rufous furfura- ceous-lepidote below, more so along the midrib and secondary veins, the midrib impressed above, prominently raised below, the secondary veins 42 to 49 pairs, slightly impressed above, prominently raised below. margins entire, flat; petioles slender, 7-28 mm long, 1-3 mm diam., glabrous above, fur- 268 Annals of the Missouri Botanical Garden furaceous-lepidote below. /nflorescences erect, bi- to tripinnately paniculate, 9.1-20.5 х 4.2-20.5 cm, pyramidal, longer than the leaves, the rachis, branchlets, abaxial bract surfaces, and pedicels densely furfuraceous-lepidote, the branches loosely congested into 4- to 9-flowered corymbs; peduncles 0.8—3.2 mm long, the lower branches subtended by leaves; inflorescence bracts unknown; inflorescence branch bracts usually persistent, chartaceous, ovate, 2.5—4.5 X 1.5-2.2 mm, apically acute, in- conspicuously punctate and punctate-lineate, fur- furaceous-lepidote abaxially, glabrous adaxially, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflo- rescence branch bracts but membranous to char- taceous, 1.4-2.3 X 1.6-2.8 mm; pedicels stout, 1— 2.2 mm long, inconspicuously punctate and punc- tate-lineate, furfuraceous-lepidote. Flowers 5-me- rous, pink to red-violet; calyx lobes coriaceous, or- bicular, 2.6-2.8 X 3.5-3.7 mm, apically rounded, conspicuously black punctate and punctate-lineate, furfuraceous-lepidote abaxially, glabrous adaxially, the margin entire, erose, hyaline, sparsely glandu- lar ciliolate; corolla chartaceous, 8.6-8.9 mm long, the tube 1.2-1.3 mm long, the lobes ovate, 5.3-5.6 X 2.6-2.8 mm, apically acute, conspicuously black punctate and punctate-lineate, glabrous throughout, the margins entire, slightly hyaline; stamens 7.2— 7.3 mm long, the filaments 4.3—4.4 mm long, the staminal tube 1.2-1.3 mm long, the apically free portions 3.0—3.2 mm long, the anthers lanceoloid, 3.6-3.8 X 1.1-1.4 mm, apically apiculate, basally lobate, the connective conspicuously punctate; pis- tils 8.6-8.9 mm long, glabrous, the ovary oblong, 1.8-2 mm long, the style 6.9-7.1 mm long, con- spicuously punctate, the ovules 50 to 55. Fruits globose, 5.2-6.7 mm diam., inconspicuously punc- tate and punctate-lineate, inconspicuously costate. Distribution, Ardisia hagenii is endemic te western Panama, in Bocas del Toro and Chiriquí, growing from 1000 to 2100 m in elevation. “cology and conservation status. Ardisia hage- nit occurs in montane wet, cloud, and elfin forests. While it is certainly not common, at this time there are no data to suggest the species is threatened. Etymology. This species was named in honor of Christine and Wolfgang von Hagen, collectors of the type specimen. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia hagenii may be most easily con- fused with A. pseudoracemiflora because of its el- Прие to narrowly elliptic leaves and inflorescence much longer than wide. However, Ardisia hagenii is easily separated from A. pseudoracemiflora by its - wider, coriaceous, orbicular calyx lobes to 3.7 mm wide, longer corolla lobes to 5.6 mm long, longer anthers to 3.8 mm long, longer styles to 7.1 mm long. fewer secondary veins of the leaf blades, and thicker branchlets. The von Hagen collections are atypical of the species. The inflorescence branches appear fasci- aled, causing some branches to be swollen and re- duced, giving the inflorescence a “columnar” ap- pearance as noted by Lundell (1968). Unfortunately, Lundell failed to understand the atypical nature of his type and never annotated any other specimens as Ardisia hagenii. Instead, Lun- dell described Auriculardisia chiriquiana Lundell, which has normal branches and an inflorescence in young bud. The type of Auriculardisia toroana is in flower and unique only because of the hyaline mar- gin of its calyx lobes. However, the types corre- sponding to both Auriculardisia chiriquiana and A. toroana match Ardisia hagenii in all other respects. Specimens examined. PANAMA. Bocas del Tore ud Changuinola, 1 km from Corriente ane ai 2 ıa, 18 Jan. 1980 (fl), M. Correa et al. 3180 MA); SE and NE of € eris 1 km from IRHE, 19 P 1980 (fl), M. Correa d al. 3323 (MO, РМА); along C bag fey, and near the Corriente о School, 25 Feb. О (fr), M. Correa 3985 (MO, РМА); Fortuna Dam reg Y 1, gabs Mie service road, 7 Dec. 1985 (fr), G. Me Pherson 7827 (F, FTG, LL, MO, PMA). Chiriquí: "c road between Fortuna Lake dud Chiriquf Grande, 4.5—5 km N of dam over Fortuna Lake, 8 Mar. 1985 (fl), T. Croat ii M. oe 60032 (LL, MO); 3.5 mi. NE of Boquete, та of road along Río Palo Alto, 19 Nov. 1978 (fl), / Hound 5736 (МО); S slopes of Cerro Pate Macho along fo Palo Alto, 11 Nov. 1981 (fr), S. Knapp et al. 2086 (LL, MO, NY); near Fortuna Dam, along Quebrada de Arena, S i pd. e Divide, 5 Dec. 1985 (fl), G. Mc- Pherson 7782 (F . MO): along trail to Cerro Pate Macho, 6 Feb. e а G McPherson & M. Merello 8296 (FTG, LL, MO); Fortuna Dam lon, above N edge of lake, 27 Apr. 1986 (fl) G. McPherson 9083 (F, FTG, MEXU, MO, NY, РМА); SE slopes and summit of Cerro Pate о, trails from Río Palo Alto, 4 km NE of Bo- quete, 26 May 1981 (fl), K. Sytsma et al. 4831 (LL, MO); Fortuna Dam region, along Quebrada Arena, S of Conti- nental Divide, 15 Jan. 1989 (fl), G. McPherson 13541A (FTG, MEXU, MO, PMA) 48. Ardisia hugonensis (Lundell) Pipoly & Ric- ketson, Sida 18: 5 998. Auriculardisia hu- gonensis Lundell, Wrightia 7: 268. 1984. Ar- disia hugonensis (Lundell) Lundell, Phytologia 6, nom. inval. TYPE: Colombia. Chocó: Mpio. de Quibdó, Corregimiento de Guayabal, Río Hugón, ca. 80 m, 12 Sep. 1976 т), E. Forero & R. Jaramillo 2812 (holotype, NY* isotype, MO!). Figure 50. —. m Trees 10 m tall. Branchlets 4.5-6 mm diam.. smooth, densely appressed rufous furfuraceous-lep- 269 Ricketson & Pipoly Volume 90, Number 2 2003 Revision of Ardisia subg. Auriculardisia (LOI) 6162 {илз TY моң s шолу 7) (OW) СЕРР WMH ^W oy ddouy `$ woy g (OW) £781 ddpuy `$ *od&josr шолу umeıp ү) сип] 77)— Moy “q— "qoueqq Зицәмор4 "y— "nddpux терү (8ш) [€ omar CAN) 2182 ojmupapf ^w 29 042404. 77] doy шолу umep g *V) in11 'g— "qyougeiq Зиләмо "Y— 'sisuauogm visipay— "(yop OS 2314 К ылы Ame i PRE | \ E 5 p Н M OS Annals of the Missouri Botanical Garden idote. Leaves with blades coriaceous, elliptic, 11.3— 14.9 X 2.9-4.4 cm, apically acuminate, with an acumen 4—9 mm long, basally acute, decurrent on the petiole, inconspicuously punctate and punctate- lineate, glabrous above, furfuraceous-lepidote be- low, the midrib impressed above, prominently raised below, the secondary veins 25 to 33 pairs, prominulous above and below, the margins entire, inrolled; petioles slender, canaliculate, 3-6 mm long, glabrous above, furfuraceous-lepidote below. Inflorescences erect, tripinnately paniculate, 11.8— 13.1 X 9.5-10.5 cm, pyramidal, mostly longer than the leaves, densely cupuliform and furfuraceous- lepidote, the branches loosely congested into 3- to 7-flowered corymbs; peduncles obsolete to 5 mm long, the lower branches subtended by leaves; in- florescence bracts unknown: inflorescence branch bracts unknown; floral bracts unknown: pedicels stout, terete, 1-1.8 mm long, inconspicuously punctate and punctate-lineate, furfuraceous-lepi- dote. Flowers 5-merous, color unknown: calyx lobes chartaceous, ovate, 1.3-1.6 X 1.3-1.8 mm, apical- ly acute to rounded, prominently punctate and punctate-lineate, glabrous adaxially, glabrous to sparsely furfuraceous-lepidote abaxially, the mar- gins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla, stamens, and pistil un- known. Fruits globose, 6.8—10.4 mm diam.. prom- inently punctate and punctate-lineate, glabrous. Distribution. Ardisia hugonensis is known only from the type and is endemic along the Río Hugón in Chocó, Colombia, growing at about 80 m in el- evalion. Ecology and conservation status. Ardisia hugo- nensis occurs in the wettest of Neotropical forests, a true pluvial forest that may receive up to 11 m of rain per year. With increasing expansion of hu- man populations in the area, the species should be considered threatened. Etymology. The specific epithet refers to the Rio Hugón from where the type was collected. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia hugonensis is most closely related to А. smurfitana because of its short calyx lobes that are as long as or much longer than wide, short pedicels, and very long and wide coriaceous leaf blades. However, A. hugonensis differs from A. smurfitana by its smooth branchlets, smaller leaf blades to 14.9 x em, slender, canaliculate, shorter petioles to 6 mm long, and chartaceous, ovate, larger calyx lobes to 1.6 X 1.8 mm. 49. Ardisia knappii (Lundell) Pipoly & Ricket- son, Sida 18: 513. 1998. Auriculardisia knap- pu Lundell, Phytologia 55: 235. 1984. Ardisia knappii (Lundell) Lundell, Phytologia 61: 65. 1986, nom. inval. TYPE: Panama. San Blas: 23-29 km from Pan-American Highway on El Llano—Cartf Road, 09*22'N, 078?69' W, 300- 400 m, 28 Oct. 1981 (fl). S. Knapp 1843 (ho- lotype. LL!: isotype, MO!). Figure 51. Shrubs or treelets 2.5—5 m tall. Branchlets stout, terete, 8—14 mm diam., densely and minutely ap- pressed rufous furfuraceous-lepidote, glabrescent. Leaves with blades membranous and chartaceous, elliptic or oblong, 24.5-51 Xx 9.2-15.4 em, apically acuminate, with an acumen 5-18 mm long, basally cuneate, decurrent on the petiole, prominently punctate and punctate-lineate, sparsely furfura- ceous-lepidote above, densely furfuraceous-lepi- dote below, the midrib impressed above, promi- nently raised below, the secondary veins 60 to 70 pairs, prominulous above, prominently raised be- low, the margins entire, inrolled; petioles stout, marginale, 3-8 ст long, sparsely furfuraceous-lep- idote above, densely furfuraceous-lepidote below. Inflorescences erect, tripinnately paniculate, 929.5 X 5.5-17 ст, pyramidal, shorter than the leaves, densely mixed cupuliform and furfuraceous-lepi- dote, the branches congested, nearly glomerulate, into 15- to 25-flowered corymbs; peduncles obso- lete to 1 cm long, the lower branches subtended by leaves; inflorescence bracts unknown; inflorescence branch bracts caducous, membranous, ovate to ob- long, 1.4-2.9 X 1.4-1.9 mm, apically acute, the veins unknown, prominently punctate and punc- tate-lineate, glabrous above, furfuraceous-lepidote below, the margins irregular, minutely erose, hya- line, sparsely glandular ciliolate; floral bracts sim- ilar to the inflorescence branch bracts, but ovate, 1.2-1.5 X 0.7-1 mm; pedicels slender, 1—1.7 mm. prominently punctate and punctate-lineate, furfu- raceous-lepidote. Flowers 5-merous, purple to ma- genta; calyx lobes membranous, ovate, 1.4—1.6 X 1.3-1.5 mm, apically acute, prominently punctate and punctate-lineate, glabrous adaxially, furfura- ceous-lepidote, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 5.1—5.2 mm long, the tube 1.2-1.3 mm long, the lobes ovate, 3.8—4 X 1.8-2 mm. api- cally acute, prominently punctate and punctate-lin- eate, glabrous adaxially, mixed lepidote abaxially, the margins entire, hyaline; stamens 3.9—4.1 mm long, the filaments 1.8-1.9 mm long, the staminal tube 0.8-0.9 mm long, the apically free portions 0.9-1.1 mm long, the anthers narrowly ovoid, 2.3— A X 1-1.1 mm, apically apiculate, basally sub- cordate, the connective conspicuously punctate: N Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 271 pistil 4.7—4.8 mm long, glabrous, the ovary oblon- goid, 0.9-1 mm long, the style 3.7—3.9 mm long. prominently punctate, the ovules 12 to 15. Fruits globose, 6—7.3 mm diam., prominently punctate. Distribution. Ardisia knappii is found on Cerro Canta Gallo in the Indio-Maíz, in Río San Juan, Nicaragua, and along the El Llano-Cartí Road in Panamá, Veraguas, and San Blas in Panama. lt is not currently known from Costa Rica. It grows from 0 to 450 m in elevation. Ecology and conservation status. Ardisia knap- pii occurs in tropical wet forests along ridges and steep slopes. Not enough is known of its population biology to. accurately determine its conservation status. Etymology. This species was named in honor of Sandra Knapp (BM), preeminent authority on the systematics of Neotropical Solanum. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia knappii is most closely related to A. pulverulenta because of its short calyx lobes that are as long as or longer than wide, short pedicels, and large leaf blades. However, Ardisia knappii dif- fers from A. pulverulenta by its thicker branchlets to 14 mm in diameter, longer and wider leaf blades to 51 X 15.4 wider inflorescence to 17 cm wide, longer and wider calyx lobes to 1.6 X 1.5 mm, longer and wider corolla lobes to 4 X 2 mm, wider anthers to 1.1 mm wide, and longer styles to 3.9 mm long. cm, Specimens examined. dee AGUA. Río San Juan: Mpio. de San Juan del Norte, Reserve Indio-Maiz, Río Indio, Cerro Canta Gallo, 14 Sep. 1998 (fl), R. Rueda et al. 8551 (HULE), 15 Sep. 1998 (fl), R. Rueda et al. 8611 (HULE, MO), 16 Sep. 1998 8, К. сив et al. 8656 (HULE, MO). PANAMA. Bocas del Toro: Escudo Veraguas, SE side of island, ا‎ Nor 1990 (В), 7 э M. Pe- terson 8522 (MO, US). Panamá: 8.2 mi. (on new road, 8.6 mi. on old road), from Pan-American Highway on the El Llano-Cartí Road, 24 Mar. 1982 (fl), S. Knapp & M. Huft pas E L, MO); 10 mi. re is Pan-American ha way El Llano-Cartí Road, 21 Apr. 1982 (fl). : King. et bet 4738 (LL, MO); El Llano-Cartí Road, m km from junction with Inter-American Highway. 31 Oct. 1974 (fl, fr), S. Mori & J. Kallunki 2919 (МО); on El Llano—Cartf road, near Nusagandi, along trail to waterfall, de i DT vide, з xi 1986 (fl), J. Mv et al. 394 Nusagandi, ca. 20 km on El Llano-Cartí Road, trails n В 95 1 Мау 1992 (fl), n ле et al. 908 (MO, PM AL 50. Ardisia liesneri Lundell, Wrightia 6: 106. 0. Auriculardisia liesneri nns 1) Lun- dell, Phytologia 49: 344. 1981. TYPE: Costa ica. Puntarenas: Osa ok Corcovado National Park, slopes above Llorona, 08°36'N — Ке, © " 083?42' W, 0-200 m, 13 July 1977 (fr), R. Lies- ner 3266 (holotype, MO!; isotype, CR!). Figure 52. Trees 2.5-8 m tall, 4-10 cm diam. Branchlets slender, terete, 2—7 mm diam., densely cupuliform lepidote and stipitate stellate-tomentellous. Leaves with blades chartaceous, elliptic, 6.6-17.4 X 2.6— 5.8 cm, apically acuminate, with an acumen 7—17 mm long, basally acute, decurrent on the petiole, inconspicuously punctate and punctate-lineate, gla- brous above, cupuliform lepidote below, with ad- ditional stipitate-stellate trichomes along the mid- rib, the midrib impressed above, prominently raised below, the secondary veins 21 to 41 pairs, prominulous above and below, the margins entire, flat; petioles slender, canaliculate, 8-13 mm long. 1—3 mm diam., glabrous above, vestiture below as in branchlets. Inflorescences erect, bipinnately pa- niculate, 4—8 X 3-7 em, pyramidal, usually shorter than the leaves, cupuliform lepidote and stipitate- stellate tomentose, the branches loosely congested into 5- to 12-flowered corymbs; peduncles 0.5-1.2 cm long; inflorescence bracts unknown; inflores- cence branch bracts membranous, oblong, 6-18 X 3.2-4.2 mm, apically acute, prominently punctate and punctate-lineate, glabrous above, vestiture as in branchlets below, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflorescence branch bracts, but spathulate, enclosing the bud (similar to Geis- santhus), 2.5—4.2 X 1.5-2.5 mm; pedicels slender, 4.5-6.2 mm long, inconspicuously punctate and punctate-lineate, sparsely furfuraceous-lepidote and stipitate-stellate. Flowers 5-merous, pale or- ange to light pink or pink-tan; calyx lobes charta- ceous, orbicular, 1.7-1.9 X 1.7-1.9 mm, apically acute, prominently punctate and punctate-lineate, glabrous adaxially, sparsely furfuraceous-lepidote abaxially, the margins irregular, minutely erose, hy- aline, sparsely glandular ciliolate; corolla membra- nous, 8.5-8.7 mm long, the tube 2.1-2.3 mm long, the lobes lanceolate, 6.2-6.4 X 2.5-2.7 mm, api- cally acute, conspicuously punctate and punctate- lineate, glabrous or rarely with a few furfuraceous- lepidote scales, the margins entire, hyaline: stamens 6.6-6.8 mm long, the filaments 2.8-3 mm long, the staminal tube 1.2-1.4 mm long, the api- cally free portions 1.4-1.8 mm long, the anthers linear-lanceoloid, 4.2—4.3 X 1-1.1 mm, apically apiculate, basally sagittate, the connective conspic- uously punctate; pistil 7.1-7.3 mm long, glabrous, the ovary oblong, 1.2-1.4 mm long, the style 5.7— 5.9 mm long. mostly epunctate, the ovules 33 to Annals of the 272 Missouri Botanical Garden ((OW) 96€ "y 021207) "jy шолу Э ЧОЙ) 2965 jouumg ^g шолу 4 (SN) rez zepupuze 4 "y *ad&jopou шолу имвар y) ити 7) MO ^7g— “Yours ёицәмор "y— “pypcydojsiow msipiy (щш) єс oanzi4 S CON) TET шару f шолу 7) OW) 66891 `P 19 wun ‘g шоу g (ON) дог 1ausar] M "ad&jopou шолу имер y) N14 77)— ләмор *g— "goueaq Зицәмор "ү causan pisipiy (үрә) cc әли Volume 90, Number 2 2003 Ricketson & Pipoly 273 Revision of Ardisia subg. Auriculardisia 38. Fruits globose, 6.5—9.3 mm diam., inconspic- uously punctate and punctate-lineate. Distribution. Ardisia liesneri is endemic to the Osa Peninsula of Puntarenas, Costa Rica, growing from sea level to 400 m in elevation. Ecology and conservation status. neri occurs as an understory element in primary lowland wet forest. Its extremely limited distribu- tion makes it vulnerable to threat, even though most of its known range occurs in protected lands. Etymology. This species was named in honor of Ronald L. Liesner, senior curatorial assistant at the Missouri Botanical Garden Within Ardisia subg. Даана sect. Pal- manae, Ardisia liesneri has flowers that are most similar to those of A. crassipedicellata (see under that species for similarities). However, A. liesneri is easily separated from A. crassipedicellata by the mixture of cupuliform lepidote and glandular-stel- late tomentum, chartaceous leaf blades, the smaller orbicular calyx lobes 1.9 X 1.9 mm, smaller corolla lobes 6.4 mm long, longer and narrower anthers to 4.3 X 1.1 mm long, and longer styles to 5.9 mm Ardisia lies- long. pee арар. COSTA RICA. Puntarenas: Reserva эё '€ confluence o (fl), К. yp 220 (CR, FTG, INB, MO); Cantón de Osa, Reserva Forestal Golfo Dulce, Раббана de Osa, Los Mo- gos, ipe Chal, 30 July 1993 (fr), R. Aguilar & M. Segura 2064 ‚ FTG, INB, MO); ped depen Park, along тш 2 to San on 2] 7 (fr), € Hartshorn 1882 (F, FTG, MO); dais Forestal Сай Dulce, Osa afl trocha de La Tarde road, e km SW o La Palma, S of Rincón de rr along ridge E the Río Rincón valley, 28 Apr. 1988 (fl), B. Hammel & R. Robles 16737 (CR, INB, MO); Reserva Forestal Golfo Dulce, Osa Peninsula, Rane ho агайы са. 15 km W of NW end of valley, near Fila үз 1988 (fl), B. к et al. 16899 (CR, FTG, LL, MEXU, MO, NY); Cantón de Osa, Reserva ri Golfo Dulce, Rancho Quemado Valley, ca. 15 km W of Rincón, S side of valley along фа ее ы, Quebradóna and Hío Ri- yito, 11 Sep. 1990 (fr), B. Hammel et al (INB, MO); Cantón de Osa, Reserva Forestal Golfo T Ran- e cho Quemado, ca. 15 km of Rincón, in range before ancho Quemado Va They: 1 ae 1992 (fr), B. Hammel & i pos Bet B, MO); Parque Nacional Cor- os Forest, 26 May 1989 (fl), C. Kernan & p Phillips 1126 (CR, FTG, MO); Cantón de Osa, pod Forestal Golfo Dulce, Ap nd de Osa, Rancho Quemado ue 21 Sep. 1993 (fr), A. Marín & H. Gutiérrez 66 (INB, MO); Cantón de Osa, Cerro Rancho Quemado, Rincón, 20 dp 1991 (fr), J. Marín 134 (CR, INB, MO); Aguabuena, 3.5 km W of Rincón, permanent plot 1 km of BOSCOSA station, 9 Sep. 1992 (fr), K. Thom- sen 95 (FTG). Ardisia lundelliana Pipoly, Ann. Missouri Bot. Gard. 78: 524. 1991. ТҮРК: 51. Panama. Chiriquí: vicinity Fortuna Dam, forested slopes along ridge at S boundary of watershed, 08°45'N, 082°15'W, 1250 m, 28 Apr. 1986 (fl, fr), G. McPherson 9107 (holotype, MO! [uni- cate |). For illustration, see Pipoly (1991b: 524, fig. 1). Tree to 4 m tall. Branchlets stout, with fine lon- gitudinal ridges, 5—7 mm diam., with a dense mix- ture of cupuliform and furfuraceous-lepidote scales. Leaves with blades coriaceous, elliptic, 7-18 X 4— 7.7 cm, apically acute to short-acuminate, with an acumen 3—6 mm long, basally obtuse to rounded, lecurrent on the petiole, inconspicuously punctate and punctate-lineate, glabrous above, with a dense mixture of cupuliform and furfuraceous-lepidote scales below, the midrib impressed above, promi- nently raised below, the secondary veins 39 to 4 pairs, prominulous above and below, the margins entire, revolute; petioles slender, marginate, 1.1-2 cm long, glabrous or sparsely furfuraceous-lepidote above, with a dense mixture of cupuliform and fur- furaceous-lepidote scales below. Inflorescences pen- dent, pinnate to bipinnate paniculate, 8—15 X cm, globose, shorter than the leaves, vestiture of the rachis, abaxial bract surfaces, branchlets and pedicels similar to the branchlets, the branches loosely congested into 4- to 9-flowered corymbs; peduncles 2.3-2.7 mm long; inflorescence bracts unknown; inflorescence branch bracts membra- nous, lanceolate, 1.5-3.5 X 0.6-1.6 mm, apically acute, the veins inconspicuous; floral bracts similar to the inflorescence branch bracts except acute, 1.8-2 х 0.8-1 mm; pedicels slender, recurved, 6— 14 mm long, inconspicuously punctate and punc- tate-lineate. Flowers 5-merous, n pink; calyx lobes coriaceous, widely ovate, 5.2-6 X 3.6—4 mm, apically acute, prominently punctate d punctate- lineate, glabrous adaxially, with a dense mixture of cupuliform and furfuraceous-lepidote scales aba- xially, margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous to chartaceous, 9.9-10.1 mm long, the tube 2.7-2.9 mm long, the lobes narrowly ovate, 7—7.8 X 3.4— 3.6 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout except the tube sparsely furfuraceous-lepidote basally aba- xially, the margin entire, hyaline; stamens 6.9-7.1 mm long, the filaments 3.5-3.6 mm long, the sta- minal tube 1—1.2 mm long, the apically free por- tions 2.4—2.5 mm long, the anthers lanceoloid, 4— 4.1 X 1.2-1.3 mm, basally lobate, the connective conspicuously punctate; pis- til 14.4—15.1 mm long, glabrous, the ovary ovoid, 5.3-5.7 mm long, the style 9.1-9.4 mm long, prom- apically apiculate, 274 Annals of the Missouri Botanical Garden inently punctate and punctate-lineate, the ovules 19 to 24. Fruits (immature) globose, 5—7 mm diam., inconspicuously punctate. Distribution. Ardisia lundelliana is known only from the holotype collection, growing around the Fortuna Dam in Chiriquí, Panama, at around 1250 m in elevation. Ecology and conservation status. Ardisia lun- delliana is a ridge-top species in montane forests. Because of its restricted distribution, it should be considered threatened. Etymology. of the late Cyrus Longworth Lundell, specialist in Myrsinaceae for over 60 years. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia lundelliana is one of a number of This species was named in honor species that have branchlets covered with a mixture of dense cupuliform and furfuraceous-lepidote scales. In sterile condition, Ardisia lundelliana may be most easily confused with A. conglomerata (see under that species for similarities). However, Ar- disia lundelliana is easily separated from A. con- glomerata because of its pendent inflorescence with much longer, recurved pedicels to 14 mm long and much larger anthers to 4.1 mm. 52. Ardisia mcphersonii 1994. TYPE: Antioquia: Mpio. Frontino, area called Murrí, in W-Central part of Antioquia, ca. 15 km from Nutibarra, 06°40'N, 076°20'W, 1875 m, З Nov. 1988 ( buds). G. McPherson, J. Zarucchi, F. Roldán & O. Escobar 12954 (holotype, HUA!; iso- types, MO!, US!). For illustration, see Pipoly (1994: 38, fig. 1). Pipoly, Novon 4: 38. Colombia. — — Trees 4—6 m tall, to 7 cm diam. Branchlets flex- uous, slender to stout, subterete, 5—7 mm diam., with a mixture of densely cupuliform and furfura- ceous-lepidote scales. Leaves with blades charta- ceous, oblong to narrowly elliptic, 18-33 х 5.8- 8.5 acumen 1.2-1.5 cm long, basally acute, decurrent em, apically abruptly acuminate, with an on the petiole, prominently punctate above and be- low, essentially glabrous above, with a mixture of densely cupuliform and furfuraceous-lepidote scales, more densely so along the midrib, the mid- rib impressed above, prominently raised below, the secondary veins 30 to 38 pairs, prominulous above and below, the margins entire, revolute; petioles stout, canaliculate, 1.2-1.5 cm long, essentially glabrous above, below with a mixture of densely cupuliform and furfuraceous-lepidote scales. Inflo- rescences erect, tri- to quatripinnately paniculate, 21-30 x 20—30 cm, pyramidal, nearly as long as the leaves, the peduncle, rachis, secondary branch- es, and pedicels with a mixture of densely cupuli- form and furfuraceous-lepidote scales, the branches loosely congested into 5- to 8-flowered corymbs; peduncle 2.6-3.1 ст long, the lower branches sub- tended by leaves; inflorescence bracts and branch bracts unknown; floral bracts early caducous, char- taceous, lanceolate, 1—1.4 х 0.3-0.4 mm, apically attenuate, prominently punctate and punctate-lin- =, eate, glabrous above, below with a mixture o densely сири огт and furfuraceous-lepidote scales, the margins entire, densely lepidote; pedi- cels slender, 1.2-1.8 mm long, inconspicuously punctate and punctate-lineate, with a mixture of densely cupuliform and furfuraceous-lepidote scales. Flowers 5-merous, cream; calyx lobes char- taceous, suborbicular to oblate, 1.4—1.8 X 1.7-1.9 mm, apically rounded to obtuse, densely and prom- inently punctate and punctate-lineate, glabrous adaxially, furfuraceous-lepidote abaxially, the mar- gins irregular, minutely erose, hyaline, sparsely glandular-ciliolate; corolla coriaceous, 4.2—4.4 mm long, the tube 0.3-0.5 3.7-4.1 X 1-1.2 mm, apically acuminate, promi- nently punctate and punctate-lineate, glabrous mm long, the lobes oblong, throughout, the margin entire, erose, hyaline: sta- mens 3.2-3.6 mm long, the filaments 1-1.2 mm long, the staminal tube 0.3-0.5 mm long, the api- cally free portions 0.5-0.9 mm long, the anthers narrowly ovoid to lanceoloid, 2.4-2.6 X 1-1.2 mm, apically apiculate, basally sagittate, the connective epunctate; pistil 3.5-3.9 mm long, glabrous, the ovary ovoid, costate, 5-angled, 1.4—1.6 mm long. 1-1.2 mm diam., the style 2.1-2.3 mm long, epunc- tate, the ovules 24 to 35. Fruits unknown. Distribution. Ardisia mcphersonii is known only from the Murrí area of Mpio. de Frontino in Anti- oquia, Colombia, growing from 1700 to 1990 m in elevation. Ardisia mcphersonii occurs in montane and cloud forests. Ecology and conservation status. Because of its restricted distribution, it should be considered threatened. Etymology. This species was dedicated to Gor- don McPherson of the Missouri Botanical Garden, colleague and specialist in Madagascan Euphorbia- ^ eae. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia mcphersonii is most closely related to A. furfuracea (see under that species for similar- ities). However, Ardisia mcphersonii can be distin- guished from А. furfuracea by its wider calyx lobes 1.9 mm wide and narrower corolla lobes to 1.2 mm wide, shorter stamens to 3.6 mm long. larger Volume 90, Number 2 2003 Ricketson & Pipoly 275 Revision of Ardisia subg. Auriculardisia anthers to 2.6 X 1.2 mm, and shorter styles to 2.3 mm long. Specimens examined. . Alt to de Cuevas, COLOMBIA. Antioquia: Mpio 10 km W i 1988 (fl bud), J. Zarucchi et al. 7232 (MO 53. Ardisia megistophylla Lundell, Wrightia 4: 147. 1970. Auriculardisia megistophylla (Lun- dell) Lundell, Phytologia 49: 344. 1981. TYPE: Colombia. Chocó: Costa del Pacifico, эрен de Utria, 5 June 1950 (fr), А. Fer- nández 251 (holotype, US!, LL neg. 1971-69!). Figure 53. Ardisia atrata Lundell, Wrightia 6: 60. 1979 10V. "ulardisia е (Lundell) b Prunes 49: 342. 1981. TYPE: Panama. Colón: Río Gaunche in forest along Río dern 3-7 km above bridge on forested slope above river, 300—700 ft. [91—21: ~~ ^a С lI ШУ, 4902 (holotype, MO!, К neg. 55670!, LL 1979-33; isotype, LL!). Ardisia macrostachya Lundell, Wrightia 6: 81. 1979. Syn. Auric dp macrostachya (L dcus Lundell. Phyoloría 49: 344. 1981. TYPE: Panama. Panamá: Gorgas Men it Labs yellow fever research camp. Trees 2-8 m tall. Branchlets stout, terete, 9-15 mm diam., densely rufous furfuraceous-lepidote. Leaves with blades chartaceous, elliptic to oblong or narrowly oblong, 23.2—61.4 X 7.1—21.3 cm, api- cally acute, with an acumen 6-17 mm long, basally acute, decurrent on the petiole, prominently punc- tate and punctate-lineate above and below, nearly glabrous above, furfuraceous-lepidote below, the midrib impressed above, prominently raised below, the secondary veins 70 to 85 pairs, prominently raised above and below, the margins entire, inro- lled; petiole stout, marginate, 1.4—5.5 cm long, 4— 12 mm diam., nearly glabrous above, furfuraceous- below. Inflorescences erect, bi- to X 9-19 cm, the branches con- lepidote tripinnately paniculate, 12—37 midal, shorter than the leaves, gested into 9- to 13-flowered corymbs; peduncles obsolete to 2 ст long, the lower branches subtend- pyra- ed by leaves; inflorescence bracts unknown: inflo- rescence branch bracts caducous, membranous, ovate to oblong, 3.2-8.3 X 1.2—4.8 mm, apically acute, prominently punctate and punctate-lineate, glabrous above, furfuraceous-lepidote below of dense, obsolete flat scales with the margins entire or with small teeth, the midrib impressed above, prominulous below, the secondary veins inconspic- uously raised, the margins entire, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflorescence branch brac ‘ts, but 0.6— 1.1 X 0.5-0.8 mm; pedicels stout, 0.5—3 mm long, prominently punctate and punctate-lineate, furfu- raceous-lepidote. Flowers 5- or 6-merous, white, light pink, pink-blue, or red; calyx lobes charta- ceous, ovate, 2.6-2.8 X 1.8-2 mm, apically acute, prominently punctate and punctate-lineate, gla- brous adaxially, furfuraceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 7-7.2 mm long, the tube 1.5-1.8 mm long, the lobes ovate to narrowly ovate, 5.4—5.5 X 2.2-2.3 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the margins entire, hyaline: stamens 5.5-5.7 mm long, the filaments 2.7-2.8 mm long, the staminal tube 1—1.1 mm long, the apically free portions 1.6—1.8 mm long, the anthers lanceoloid, 3.1-3.3 X 0.9-1 mm, apically apicu- late, basally cordate, the connective conspicuously punctate; pistil 6.4—6.8 mm long, glabrous, the ova- ry ovoid to oblong, 1-1.2 mm long, the style 5.4— 5.6 mm long, epunctate, the ovules 11 to 13. Fruits globose, 6-9 mm diam., prominently punctate. Distribution. Ardisia megistophylla ranges from central Panama to Chocó, Colombia, growing from 9] to 1000 m in elevation. Ecology and conservation status. tophylla occurs in lowland pluvial riverine forests. Ardisia megis- Apparently, it is a fairly rare species and should be considered threatened. The specific epithet was derived very big or very large. tymo огу. from the Greek “megisto,’ and “phyla,” leaves. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia megistophylla is most closely relat- ed to A. aguirreana, A. cogolloi, A. crassipes, and A. cartagoana because of its long calyx lobes and thick petioles. Ardisia megistophylla can easily be separated from all these by its longer calyx lobes to 2.8 mm long, longer corolla lobes to 5.5 mm long, longer anthers to 3.3 mm long. and longer styles to 5.6 mm long. The type of Ardisia atrata is unique only for its slightly longer pedicels and slightly larger calyx lobes. The type of A. macrostachya is unique only for its slightly subsessile petioles and slightly larger calyx lobes. PANAMA. Coclé: along river Specimens examined. from La Junta near Lin 5898 (MO). Colón: E ses Rita Ridge. mere Pen 11 276 Annals of the Missouri Botanical Garden Jan. 1968 (fr), M. Correa A. 598 (MO); vicinity of Sar Trees 1.5-10 m tall, 2.6-25 ст diam. Branchlets Miguel de la Borda, 21 Apr. 1970 (fl, fr), T Croat 9559 MO); swampy area a резо h, 21 Apr. 1970 (fr), Croat 9868A (MO); swampy area between beach and for est, 25 Apr. 1970 (tl). T е 10071А (MO); Santa Rita Ridge, 1 Mar. 1971 (fr), T. Croat 13839 (MO); head waters ia s ааа near fork with Río Nombre de Diosito, where ido trail road crosses the ridge, 21 due "1978 3 (fl), imel 3952 (MO); Santa Hin ys d Hd., 20-22 km E Transisthmus Hwy., 25 Sep. ‚ К. Sytsma 1326 (LL, MO). San Blas: Caro M 16 Oct. 4 (fr), G. de Nevers et al. 4046 (MO). raguas: pin of Río Concepción, 4 Dec. 1967 (fr). W. pex et al. 2808 (LL, MO, MOCZ). COLOMBIA. Chocó: N ridge of Alto de Bue >y, eo Dos Bocas del Río Mutatá, tributary of Río El Valle, ESE of El ins 8 Aug. 1976 (fl), A. Gentry & M. onus pA (LL, M — 54. Ardisia nigropunctata Oerst., Vidensk. eddel. Dansk Naturhist. Føren Kjøbenhavn 861: 127. 1862. Auriculardisia nigropunctata (Oerst.) Lundell, Phytologia 54: 285. 1983. TYPE: Costa Rica. Cartago: Monte lrasá [Ira- 201], 8000-9000 ft. [2438-2743 m]. Jan. 1847 (fr), А. Oersted 28D (lectotype, designated here, C!, F neg. 22953!). Figure 54. Ardisia Virgo iin Mez, in Engl., Pflanzenr. IV. 236 (Heft 9); € 02. nov. Auriculardisia chontalensis e ыо Phstologia 54: Panama. Bocas del Toro: и. de Chiriquí and its neighborhood, Nov.—Dec. 5 (fl). J. Hart 136 (lec- = дш '@ here, ^ D neg. 71-1475; isolec- type, US aniio mammosa L undell, Wrightia 4: 60. 1968. Syn. nov. ula ee mammosa (Lundell) Lundell, Phyto- lagia 54: 285. 1983. TYPE: Nicaragua. Granada: summit of ML Mombacho, near Granada, 1160 m, 24 Dec уй» (fl), V Grant 869 thalotmne: Al, LL neg. 19 71-65 Auric ulardisia UR Lundell, Phytologia 56: 4 . Syn. nov. Аа" quadrata (Lundell) ua n CRAS 61: ta т une ide nie T : Costa Rica. Puntarenas: foothills of the Cor- diloni Talamanca, around Tres Colinas, mo m 20 Mar. 1984 (fl, fr), G. Davidse, G. Herrera Ch. & R. Warner 25645 (hislotype, LL!; isotypes, INB nol 1, ud Lundell, Phytologia 63: 74. 987. Syn. nov. Ardisia moraviana 1 undell) Lun- "id Phytologia 63: 463. 1987. T : Costa Rica. San José: roadside leading from Alto La Palma to Bajo La Hondura, ca. 10 km NE of San Vicente de Moravia, 1260-1550 m, 24 Feb. 1978 (fr), R. Wilbur 24919 (holotype, DUKE!). moraviana slender, terete, 2-7 mm diam., densely and mi- nutely appressed ferrugineous furfuraceous-lepi- dote. Leaves with blades membranous to coriaceous, oblong or elliptic to oblanceolate, 6-45 X 2-16 em, apically acute to acuminate or rounded, with an acumen 5—25 mm long, basally obtuse to acute or cuneate, decurrent on the petiole, prominently punctate and punctate-lineate, mostly glabrous above, densely and minutely appressed rufous fur- furaceous-lepidote, the midrib impressed above, prominently raised below, the secondary veins 30 to 50 pairs, bullate above, prominently raised be- low, the margins entire to dentate or serrate, flat; petioles slender, marginate, 4—18 mm long, 1-3 mm diam., glabrous above, furfuraceous-lepidote below. Inflorescences erect, bi- to tripinnately pa- niculate, 7—52 X 5—28 cm, pyramidal, usually lon- ger than the leaves, the rachis straight to genicu- ate, the peduncle, rachis, branches, and pedicels furfuraceous-lepidote, the branches loosely con- gested into 3- to 9-flowered corymbs; peduncle nearly obsolete to 4.2 em long, the lower branches subtended by leaves; inflorescence bracts unknown: inflorescence branch. bracts caducous, membra- nous, lanceolate to oblong, 1.3—4.3 X 0.5-1.2 cm, apically acute, prominently punctate and punctate- lineate, glabrous above, furfuraceous-lepidote be- low, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to but 0.5-2.1 mm; pedicels slender, furfuraceous-lepi- the inflorescence branch bracts, 0.2-0.5 dote. Flowers 5- or 6-merous, white to pink or light purple: calyx lobes membranous to chartaceous, ovate, 1.7-2.4 X 1.2-2 mm, apically acute, prom- inently punctate and punctate-lineate, glabrous adaxially, furfuraceous-lepidote, the margins erose, hyaline, sparsely glandular ciliolate; corolla mem- branous, 4.9—5.8 mm long, the tube 0.8—1.1 mm long, the lobes narrowly ovate to lanceolate, 3.9—5 m .5-2.5 mm, apically acute, prominently punc- tate and punctate-lineate, glabrous adaxially, fur- furaceous- iid abaxially, the margins entire, hyaline; stamens 3.3-4.5 mm long, the filaments 1.9-3.6 mm es the staminal tube 0.5—0.8 mm long, the apically free portions 1.4-2.8 mm long, the anthers narrowly ovoid, 1.9-2.8 X 0.8-1.4 mm, Figure 54 (left). Ardisia nigropunctata. —A. Flo Figure 55 (right). Ardisia palmana. —A. Tonduz 12632 (7406) (К); C from D. Acevedo 78 (MO).) owering branch. lectotype, A. Oersted 28D (C); B from R. Wilbur 63199 (MO).) Flowering branch. —B. Flower. —( ЕЕ —В. Flower. —C. Fruit. (A & C drawn from ;. Fruit. (A, B drawn from isotype, A. Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 277 278 Annals of the Missouri Botanical Garden apically apiculate-mucronate, basally deeply cor- date, the connective conspicuously punctate; pistil 8.5-9.3 mm long. the ovary oblong, 1-1.1 mm long, the style 7.5-8.3 glabrous, prominently punctate, mm long, prominently punctate and punctate-lin- eate, the ovules 12 to 20. Fruits globose, 5-9 mm diam., prominently punctate. Distribution. Ardisia nigropunctata is widely distributed from Belize to Panama, growing at 5 to 3200 m in elevation. Ecology and conservation status. Ardisia nigro- punctata occurs in moist forests, from premontane to cloud forests, and is always found along the for- est margins. Its tolerance to disturbance and broad habitat range combine to give the species great re- siliency, and it is not considered threatened. Etymology. black punctations throughout the plant The specific epithet refers to the Common Names. “High Ridge eee Berry” ' (P. Gentle 6406); "Blossom berry igni (Р. “Manchador” (S. Record 25); "Uva Montafiera” (R. Rueda et al. 2674 Pipoly (unpublished data) has observed that this "Blossom berry? — species 15 the favorite for adorning altars in reli- gious services throughout northeastern Nicaragua. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia nigropunctata is most similar to А. angucianensis (see under that species for similari- ties). However, A. nigropunctata can be separated by its shorter and narrower calyx lobes to 2.4 X 2 mm and larger fruits to 9 mm in diameter. In the description by Oersted (1862: 127) two collections were listed, one flowering in May 1847 and the other in fruit collected from *monte Barba," in January 1847 from “monte Irasu,” both present at C. Oersted 28C, and the lrasu collection is labeled as A. Oersted 28D. We hereby designate the A. Oersted 28D collection at C as the lectotype because it is The collection from Barba is labeled as / by far the more complete collection. | his original description, Mez (1902) listed fice collections for Ardisia chontalensis, R. Tate 228 from Nicaragua (in young fruit), B. Seemann 59 from Chontales, Nicaragua (in bud), and J. Hart 136 from Chiriquí, Panama (in flower), all from K. All three sheets are fragmented with no attached leaves. We here select the original K specimen of J. Hart 136 as the lectotype because it is in flower. Ardisia nigropunctata, as with many species of Ardisia, exhibits continuous quantitative variation over a broad range of leaf and floral part sizes, and this has caused much over-description. Material corresponding to the type of А. chontalensis is unique only for its less geniculate inflorescence, but matches A. nigropunctata in all other respects. The type of A. mammosa is unique only for its buds with strongly contorted and constricted corolla, which Lundell (1968) characterized as *mammose" or nipple-shaped. The type of Auriculardisia qua- drata is unique for its slightly larger and more gla- brous leaf blades with conspicuously punctate and punctate-lineations; however, these fall into the variation of А. nigropunctata as circumscribed by us. The type of Auriculardisia moraviana is in fruit and unique only for its large, glabrous leaf blades, which are similar to those of A. quadrata; however, the fruits match those of Ardisia nigropunctata ex- actly. Specimens examined. BELIZE. Cayo: Chiquibul, Raspac 'ulo Camp, 28 Mar. 1996 (fr), A. eis 1347 (BM, MO). Stann Creek: Stann Creek Valley, 17 mi., 7 Feb. 1940 (fl), P. Gentle 3205 (A, K, LL [2], ш Н, NY, US). Stann Creek Valley, Mountain ds Ridge, 30 Mar. 1940 (fr), P. Gentle 3294 (A ‚ MO, NY). Toledo: Maya наг ge abire Reserve, river- C Camp and AC Camp he эе 'op- `, upper Bla den Branch, 8 May 1996 (fr), ( M. Meadows 35751 (BRH, FTG, M SEL) be Mo. Mountains, Bladen Nature Reserve, West Snake Creek, 28 May 1997 (fl), D. Holland & B. Kid id (BRH, MO, SEL). G UATEMALA. Alta Verapaz: betwee ri. and Yakapur, 10 Mar. 1942 (fr), J. сўзини 8 (Е). Izabal: Los iine to Entre Ríos, 1 Mar. 1926 fr), j wu 25 (MAD 15); along Río Bonita, 21 Dec. 1941 (fl), J. Steyermark 41701 (F). Petén: on Sebol Road, 18 bo [o € Luís, 20 Nov. 1966 (fl), E. х 6619 (LL [2]. HONDURAS. Cortes: Cienaga tract near А Azul, ii aen 31 Dec. 1952 ФИ 8 Williams & үл Williams 18793 (US). Gracias a Dios: Camp Tiro, 2 mi. NW of Bulebar on third N branch of Que brada Tiro, trib- utary of Río Platano, between Río Platano and JB 30 Mar. 1981 (fl). J. б 1161 Olancho: Quebrada Catacamas cerca de la Oresa del ed dp . 19 je = Ф DR rà Montafia Pefia Blanca. Сик 'amas, 28 А Tu A. Molina R. 8338 (F, NY). nta Barl ‚Се tract, area E of Lake Yojoa, 9 к 1952 (fl). "P ‘Allen 6445 (F). Yoro: Cerro between Río Guán Guán and Río Texí- Leán Valley, W end of the Cordiller S i Nombre > Ж 6 Nov. 1988 (fl), J. Mac СА et al. 3231 (LL, A - ARAGUA. Chontales: Cerro Dd ca. i" | Сиара, З Jan. 1984 (f A A. Grijalva el al. 33804 (MO). Granada: Volcán Mombacho, hacienda Las Delicias, ca. 10 km al SE, ciudad Granada, 21 Mar. 1984 (fr), A. Gri- jalva et al. 367. 5 (HNMN, MO). e Mpio. de Wi- wilí, a e to Bosawás, Macizos del Cerro Kilambé, 9 Apr. 1998 (fr), К. Rueda & 1. Coronado 8114 (HULE. M atagalpa: Macizos de Peñas Blancas, SE side drainage of Quebrada El Quebradón. slopes N & W of da. San Martín, on border with Dept. Jinotega, 18-20 Jan. 1982 (fr), W. Stevens et al. 20996 (DUKE, HNMN, NY). Nueva Segovia: along Quebrada Tastaslí branch of Río Solonlí, 3 km S of Jalapa, 6 Apr. 1977 (fr), D. Neill 1659 (MO). Río San Juan: valley of Río Indio, 6 km upstream from the junction with Caña La Pimienta, 24 Feb. 1977 (fr). D. Neill 1503 (HNMN, MO): Mpio. San Volume 90, Number 2 2003 Ricketson & Pipoly 279 Revision of Ardisia subg. Auriculardisia Juan del Norte, delta 1 km E and 2 km N, 8 July 1995 (fr), R. Rueda et al. 2674 (FTG, HULE, MO). Rivas: Isla de Ometepe, slopes of Volcán Maderas S of Hacienda Magdalena, 28 Nov. 1982 (fl), P. Moreno 18873 (MO); Isla پت‎ a Volcán pepe. N slope, 21 Jan. 1983 (fr), ? Moreno 19789 ), NY). Zelaya: NE Nicaragua, re- ва of Braggman's put near Beca 451, 16 Dec. 1927 (fl), Е Englesing 79 (F, К); Monkey Point, Caño El Pato, 4 5 km from canyon, 25 Oct. 1981 (fr), P. Moreno 12377 (MO, NY); near Bil Tingnia, 6 km NW of Bonanza, 13 May 1978 (fl), D. Neill 3974 (HNMN, MO); Bonanza, on grounds of Neptune Mining Co., 26 Feb. 1979 (fr), J. P poly 3522 (HNMN, MO, NY); Estación Experimental " Recreo, 7 Feb. 1985 (fr), D. Ríos 305 (MO, NY). Without locality: s.d. (fl, fr), E. Friedrichsthal 6 (K), 1867-1868 (fl), К. Tate 228 (К). есе» RICA. Alajuela: E slopes of Volcán Miravalles. W of Bijagua, near the Río Zapote, 11— 12 Feb. 1982 (fl), W. Burger et al. 11676 (CR, F, LL, NY); Reserve Forestal, San apri A dad 1983 (fr), A. Car- valjal 355 (LL MO, NY). € r bridge over Río рө de Orosi at Tapanti, 2 "рее. 1978 (fl, fr). T. Antonio 72 (F, LL); Turrialba, pop between Hacienda Moravia 1 Calaveras, 3 Aug. 1995 (fl), G. Herrera 8273 (CR, F). nA еса enitn de Tilarán, San Gerardo Er Río айо айо Fincas Quesada and Arce, 5 Dec. 1991 (fl), т Bello C. & E. Cruz 4298 (FTG, INB, МО). aM е Cantón x Sarapiquí, Parque Nacional Braulio Carrillo, Puesto El Ceibo, on ridge crest 250 m E of Transect trail, 5 Mar. 1994 (ster.), B. Boyle et al. 2920 (FTG, INB, MO): Monte Barba [Barva], May 1847 (fl), A. Oersted 28C (C). Limón: Braulio Carrillo National Park, trail S of Quebra- da Gonzales, 11 Aug. 1992 (ster.), M. Bóhlke 2 (F); Limon River, Nov. 1904 (fl), H. Pittier s.n. (NY, US). Pun- tarenas: Cantón de Osa, Fila Costefia, Fila hen: head- waters of Río Piedras Blancas, Cerro Anguciana, W slope. 9 Dec. 1993 (fr), B. Hammel et al. 19258 (CR, FTG, INB, MO). San José: Cantón de Pérez Zeledón, Cuenca Térra- ba pups; Estación Santa Elena, 13 Sep. 1997 (ster.), E Alfaro & M. рш 1387 (INB, MO); Cantón de Pérez Zeledón, C же s ordillera Télumáne a, Camino a Cerro Chirripó, 31 July 1996 (ster.). B. Gamboa & A. Rojas 538 (INB, M Ee. Bocas del Toro: 1.5 mi. W of Almirante, 15 Oct. 1965 (fl), К. Blum kd (MO); near headwaters of Río Culebra ca. 5 km F of Cerro Pate Macho, 11 Feb. 1979 (fr), B. Hammel ero MO vicinity of Chiriquí Lagoon, Water Valley, 4 Dec. 1940 (fl), H. von Wedel 1823 (F). Chiriquí: Fortuna Dam area, N fork of ada de Arena, song river, 6 Feb. 1984 (fr), H. О). Panamá: Canal Zone, Zone, 5 Dec. 1973 | (fl, fr), А. Gentry & M. Nee 8691 (LL. MO). Veraguas: road from Santa Fé, past Ag School to base of Cerro Tuti, along first stream, flows from Cerro Tuti, 4 Feb. 1977 (ster.), J. Folsom 1593 (MO); Islotes de Cativo, 1841 (fr), К. Friedrichsthal 613 (W [2], F neg. 1990 55. Ardisia palmana Donn. Sm., Bot. Gaz. 27: 434. 1899. Auriculardisia palmana (Donn. Sm.) Lundell, Phytologia 49: 345. 1981. TYPE: Costa Rica. San José: in sylvis prope La Palma, 1460 m, Sep. 1898 (fl), A. Tonduz 12632 (7460) (holotype, US! [2]; isotypes. M!, CR not seen, F!, F neg. 68244!, G! [3]. Шш rufa Lundell, Wrightia 4: 182. GH!, K!, LL!, LL neg. 71-125!, M!, NY! [2]). Figure 55. 1971. Syn. Auriculardisia rufa (Lundell) Lundell, Phytologia 49: 345. 1981. : Panama. Chiriquí: mountain slopes SE of Cerro Punta, 6500-7000 ft. [1981— 2134 m]. 22 May 1971 (fr), G. Proctor 32020 d куре, LL!, F neg. 55644!; isotypes, F!, К 81, K!, LL', MICH! ize LOL Lundell, Wrightia 6: 64. 1979. Syn. nov. Auriculardisia т. {Ы undell) Lundell, Phytologia 49: 342. 1981. anama. — . Kallunki 5733 (holotype, LL, k M — Ardisia g ж Lundell, Wrightia 6: у eger gentryi (Lundell) 1 жа lox pem em 14. 1981. TYPE: Colombia. Chocó: N ridge of Alto de СТ "SW of El Valle, 500-1150 m, 8 Aug. 1976 ы А. Семгу & М. Кайеп Ene ле ЛЛ, F neg. 55656!; isotypes, LL!, Ardisia TT Lundell, Wrightia 6: n Э. Syn. nov. Auriculardisia ometepensis (Lundell) Lodi Phytologia 49: 345. 1981. TYPE: Nicaragua. Rivas: Isla Ometepe, Lago de Nicaragua, Volcán Maderas, N slope on wind beaten ridge near summit, 1200 m, 24 Feb. 1978 (fr), D. Neill & P. Vincelli 3298 (ho- lotype, МО!). тж eurubiginosa Lundell, Phytologia 56: 413 . Syn. nov. Ardisia eurubiginosa (Lundell) exi т Phytologia 61: 86, nom. inval. Ardisi d (L ameet: F. Morales, Phytologia 83: 11 m Cerro Echandi on the border, 09?*03—04'N, 082?50-51'W, Mar. 1984 (fr), G. Davidse, L. бп 25486 ternational eT m 1&9 Gómez, G. Herrera, C. Chacón & I. Chac (holotype, n», isotypes, MO!, NY!). Auriculardisia microcalyx Lundell, Wrightia 7: 270. 1984. Syn. nov. Nor Ardisia microcalyx Lundell, Wrightia 4: 46. 1968. rine azaharensis Lundell, Phytologia 61: 62. 1986, nom. inval. Ardisia azaharensis Pipoly & Ricketson, Sida 18: 511. 1998. TYPE: Costa Rica. Alajuela: 15 km NW of San Ramón by air, Cerr Azahar, headwaters of Río San Pedro, by mund. 9 km NW of San Ramón ip Piedades Norte, then 3 more km NW to pa Paz, M v to cluster of houses, then eft again on jeep r 45 km to top of ridge. 10°09’ 30N, 084°34-35' W, 1400-1500 m, 14 May 1983 (fr), R. Liesner, E. Judziewicz, J. ling че B. Pérez G. & A. Carvajal 15575 (holotype, LL; is lypes, F!, MO!). Small to large trees 2—22 m tall, 10-35 cm diam. Branchlets slender, terete, 3-10 mm diam., densely furfuraceous-lepidote. Leaves with blades membra- nous or chartaceous, elliptic or oblong, 8.3—32.7 X 3.1-8.8 ст, apically acuminate, with an acumen 2—19 mm long, basally acute, decurrent on the pet- iole, the midrib impressed above, prominently raised below, the secondary veins 52 to 65 pairs, prominulous above and below, prominently punc- tate and punctate-lineate, densely furfuraceous-lep- 280 Annals of the Missouri Botanical Garden idote, mostly glabrescent with age above, the mar- gins entire, flat to inrolled; petioles slender, canaliculate, 5-12 mm long, sparsely to densely furfuraceous-lepidote. Ив erect, tripin- 14-38 cm, dal, longer than the fees mixed cupuliform and nately paniculate, 14— pyrami- furfuraceous-lepidote, the branches loosely con- gested into 3- to 12-flowered corymbs; peduncles obsolete to 18 mm long, the lower branches sub- tended by leaves; inflorescence bracts unknown: in- florescence branch bracts unknown; floral bracts 1.6-1.9 х 1-1.2 apically acute, glabrous adaxially, furfura- caducous, membranous, ovate, mm, ceous-lepidote abaxially, the margin minutely erose, hyaline, sparsely glandular ciliolate; pedicels stout, 1-2.5 mm long, inconspicuously punctate and punctate-lineate, cupuliform lepidote. Flowers 5-merous, white to light pink; calyx lobes charta- 1.1-1.4 X 1.4—1.9 mm, apically acute to rounded, prominent- ceous to coriaceous, orbicular to oblate, ly punctate and punctate-lineate, glabrous ada- xially, sparsely furfuraceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 4—4.8 mm long, the tube 1.1—1.4 mm long, the lobes narrowly ovate to ovate, 2.7—3.7 2.2-2.5 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the margins entire, hyaline: stamens 4.1—4.4 mm long, the filaments 2.3-2.5 mm long, the staminal tube 0.6-0.8 mm long, the apically free portions 1.6-1.9 mm long, the anthers 1.9-2.5 X 0.8-1.2 basally subcordate, the pistil 4.9-5 1.3-1.4 mm long, the style 3.5-3.6 mm long, epunctate to narrowly ovoid to lanceoloid, mm, apically apiculate, connective conspicuously punctate; mm long, glabrous, the ovary oblongoid, inconspicuously punctate, the ovules 30 to 33. Fruits globose, 5-9 mm diam., prominently punc- tate and punctate-lineate. Distribution. Ardisia palmana is distributed from Rivas, Nicaragua, throughout Costa Rica to Bocas del Toro and Chiriquí, Panama, growing from 600 to 2850 m in elevation. Ecology and conservation status. Ardisia pal- mana occurs in primary and secondary forests and forest edges from lower premontane, montane, ev- ergreen cloud forests, cloud forests, and elfin for- ests. This species should not be considered threat- ened. Etymology. The specific epithet was derived from the name of the type locality, La Palma, San José, Costa Rica. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia palmana is most similar to A. an- = chicayana (see under that species for similarities). However, Ardisia palmana may be separated from A. anchicayana by the shorter calyx lobes to 1.4 mm long, shorter corolla lobes to 3.7 mm long, shorter anthers to 2.5 mm long, and shorter styles to 3.6 mm long. The original handwritten collection labels state, 1459 m, 25 IX 1898, Ad. However, two different handwritten Instit. "Foréts de La Palma, Tonduz 12632." labels сап be found, one from the Herb. Physico-Geogr. Nat. Costaricensis, the other from the Herbier Boissier. In John Donnell Smith's orig- inal description he Latinized the type collection stating, "In sylvis propre La Palma, alt. 1460 m, Sept. 1898, Tonduz n. 12632 herb. nat. Cost." John Donnell Smith made mass distributions of this col- lection with labels titled “Ex Plantis Guatemalen- sibus Necnon Salvadorensibus, Hondurensibus, Ni- caraguensibus, Costaricensibus, Quas Edidit John 99 r Donnell Smith.” The remaining label for this spe- “7460. Ardisia Palmana, Donn. Sm. in Bot. Gaz. xxvii. 434. La Palma, Prov. San José, Cos- ta Rica, alt. 1460 m. M.[month] Sept. 1898. Leg. [collector] Tonduz, (n. 12,632 herb. nat. Cost.)." Thus the number 7460 has often been associaled which is the Donnell Smith Herbarium number rather than Tonduz’s collection cles states, with this collection, number. It is important to note that the holotype at US consists of two sheets comprising a single gath- ering, one with the inflorescence and the other with the branchlet apex. Therefore, no lectotypification is necessary. Ardisia palmana, as with many species of Ardi- sia, exhibits continuous quantitative variation over a broad range of vegetative and floral parts, and this has caused much over-description. Material corresponding to the type of Ardisia rufa is unique only for its slightly more rufous indument. The type of A. boquetensis is unique only for its thinner leaves and apparently few flowers; however, it is in fruit and appears to have lost most of its flowers. The type of A. gentryi is unique for its smaller flow- ers, which are primarily in bud. The type of A. ometepensis is in young fruit and only unique for slightly larger and more numerous ovules and slightly thicker pedicels. The type of Auriculardisia eurubiginosa is unique for its slightly narrower leaf blades and thicker calyx lobes. The type of A. mi- crocalyx is in fruit and unique for its slightly small- er calyx and thicker, shorter pedicels. However, all these specimens fall into the variation of A. pal- mana as circumscribed by us. Specimens examined. NICARAGUA. Rivas: near summit and upper slopes of Volcán Maderas above Bal- güe, Isla Ometepe, 14 Sep. 1983 (fl), M. Nee & W. Robleto Volume 90, Number 2 2003 Ricketson & Pipoly 281 Revision of Ardisia subg. Auriculardisia T. 28089 (MO). COSTA RICA. Alajuela: vicinity of Bajos del Toro, road along Río Gorrión, ca. 10 (by air) ENE of Zarcero, 5 Dec. 1982 (fr), W. Alverson 2004 (MO, WIS); Cantón de San Ramón, Cordillera Tilarán, Monteverde, San Gerardo Biological Station, 31 Mar. 1995 (fl), D. Pen- neys 288 (CR, FTG, INB, MEXU, MO); Reserva Forestal San Ramón, ca. Colonia Palmarefia, 20 July 1984 (fl), J. Pipoly 7123 (MO, NY, TEX); Palmira, Región of Zarcero, 12 Oct. 1937 (fl), А. Smith 2d (F, MICH, MO). Cartago: Volcán Turrialba, 17 Sep. 1965 (fl), L. Bernai 10602 (G); vicinity of Quebrada Casa и Тарап, 29 Sep. 1981 (fl), M. Grayum 3941 (LL. MO); Panamerican Highway 5 of Cartago, between Tejar and Empalme, La Trinidad, EAS ee km Hed San José, 11 Sep. 1964 (f), K. Lems 56 (F, NY, US): vicinity of La Cangreja н a km S i El Tejar, Cordillera Talamanca, 1 Feb. 1963 (fr), L. Williams e et bu 24118 (F, GH). Guanacaste: PR de Liberia, Parque Nacional Guanacaste, Estación Cacao, 17 Dec. 1990 (f r), C. Chávez 486 (FTG, INB, MO); Zona Prot. Tenorio, Tilarán, Pe › Tilarán, Tierras Morenas, Río 993 (fl), G. анн 226 (INB, e Sarapiquí, Parque Nacional Braulio Carrilo, er ‘ión Magsasay, 23 jane 1990 (fr), D. Acevedo 78 (СК. K [2]. MO); Cantón de Barv ‚ Montana = ~ Е - = ~n C2 e = ao © = = y. US): í Gerro ed 9 o San Isidro, Cerro ba ar 3 Apr. 1976 (fl), J. Utley & K. Utley 4 (DUKE, F) Los Cartagos, 24 Feb. 1937 (fr). M. Valerio 1579 n. Cantón de San Rafael, Cerro С qoam. 16 Dec. 1993 (fr), G. Vargas et aa 1628 (CR, K). Limón: Cantón de Talamanca, Parque Nac ‘ional Cordillera Talamanca, Cordillera ee Queb. K uisa, Ujarrás Trail to San José Cabécar, 25 Mar. 1993 (fr), A. Peu 836 (F TG, INB, MO). Puntarenas: Monteverde, upper Sar 1 Luís ley on a slope, 20 Oct. 1985 (f), E E. Bello C: (CR, F, FTG, LL, , NY, US). Pontarenas & Alaje uela on or е p Cantine ntal Divide нн > km ESE of Mon- everde, 17-20 Mar. 1973 (fr), W. Bur, n Gentry 8593 (F): mo rde, ш community, 24 Feb. 1985 (fr). W Haber 1347 (LL, MO). San José: La Pa ima area, 5 of the escarpment and inns Divide, NE of San aac: 16-17 Nov. 1969 (fl, fr), W. Burger & R. Liesner 6202 (DUKE, F, MO, NY); Zona Prot. Cerro de Escazú, N slope of Cerro Rabo de Mico., Río Poás Basin, 9 Oct. 1991 (fl), J. Morales 152 (CR, INB, MO); коз E of San José 28 Oct. 1960 (fl), C. Palmer s.n. (NY); foréts de Bid 3» Flores, 22 Feb. 1890 (fr), / " Tonduz 2131 (BR); Tablazo, 23 Jan. 1935 (fr), M. Valerio 1194 (F i PANAMA. Bo« -= ;ordillera T: adw aters of the Culubre, 6 airüne km NW of the pea rro Ec lani on the Costa Rican-Panamanian inte emational Pin T, 3 Mar. 1984 (fl), G. Davidse et al. 25252 (LL, MO). m N of m Ríos on q slopes 2250 m, 22 Nov. 1979 (fr), T. Antonio 2707 (LL, MO); Las Cumbres, hogback ridge N of Que brada к; near town of Cerro Punta, 22 July 1971 (fl, f T). T. Croat & D. Porter 16093 (LL, MO, US). 56. Ardisia panamensis Lundell, Wrightia 3: 19 966. Ardisia pallidiflora Standl., J. Wash. Acad. Sci. 17: 523. 1927, nom. superfl., non Ardisia pc E ips J. Straits Branch loy. Asiat. * 2. Auriculardisia panamensis (Lundell) Len Phytologia 49: | TYPE: Panama. Chiriquí: between Alto de las Palmas and top of Cerro de la Hor- queta, 2100-2268 m, 18 Mar. 1911 (fl), H. Pittier 3255 (holotype, US!, LL neg. 1971-79!, US neg. 2380!). Figure 56. Shrubs 1—2 m tall, to 3 cm diam. Branchlets slen- der, terete, 1-3.5 mm diam., sparsely to densely and minutely rufous furfuraceous-lepidote. Leaves with blades membranous, elliptic, 2-8.4 X 0.8-3.2 cm, apically acuminate, with an acumen 0.3-1.7 mm long, basally acute, decurrent on the petiole, prominently punctate and punctate-lineate above and below, glabrous above, sparsely and minutely furfuraceous-lepidote below, the midrib impressed above, prominently raised below, the secondary veins 28 to 36 pairs, prominulous above and below, the margins entire, flat; petioles slender, canalicu- ate. 5-16 mm long, glabrous above, furfuraceous- lepidote below. Inflorescences erect, pinnately, rare- ly bipinnately, paniculate, 3—6.5 -8 cm, pyramidal, usually as long as the leaves, sometimes shorter or longer than the leaves, densely to sparse- ly furfuraceous-lepidote, the branches loosely con- gested into 6- to 14-flowered corymbs; peduncles 0.9-1.8 ст long, the lower branches subtended by leaves: inflorescence bracts unknown; inflorescence branch bracts caducous, membranous, oblong, 3.1— 4.2 X 0.8-1.4 mm, apically acute, prominently punctate and punctate-lineate, glabrous above, fur- furaceous-lepidote below, the margins irregular, mi- nutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflorescence branc bracts. but 1.4-2.6 X 0.5-0.9 mm; pedicels slen- der, 12-15 mm long, epunctate to inconspicuously punctate and punctate-lineate, usually glabrous, rarely with a few scattered furfuraceous-lepidote 5- or 6-merous, white, pale pink or scales. Flowers light purple; calyx lobes membranous, orbicular to 0.9-1.2 rounded, with few prominently punctate and punc- ovate, 1-1.2 mm, apically acute to tate-lineate, glabrous throughout, the margins irreg- ular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, —4.5 mm long, the tube 1.1-1.5 mm long, the lobes ovate, 3-3.2 x 23-24 mm, apically acute, with few promi- nently punctate and punctate-lineate, glabrous throughout except sparsely furfuraceous-lepidote on tube abaxially, the margins entire, hyaline; sta- mens 3.1-3.2 mm long. the filaments 1-1. m long, the staminal tube 0.5—0.6 mm long, the api- cally free portions 0.5-0.6 mm long, the anthers Annals of the Missouri Botanical Garden 5 (MO): C from T. B. Flower. —C. Fruit. (A drawn from holotype, H. Pittier 3255 (US); B from J. Pipoly 7 A. Flowering branch. Ardisia panamensis. Figure 56 (left). Cochrane et al. 6277 (MO).) (A-C drawn from holotype, L. Ё. Mora 2292 (JAUM).) B. Detail of bud. —C. Flower. Ardisia pseudoracemiflora. —A. Flowering branch. Figure 57 (right). | Volume 90, Number 2 2003 Ricketson & Pipoly 283 Revision of Ardisia subg. Auriculardisia lanceoloid, 2.2-2.3 X 0.6-0.8 mm, apically subu- late-apiculate, basally deeply cordate, the connec- tive conspicuously punctate; pistil 4—4.1 mm long, glabrous, the ovary oblong, 1.2-1.3 mm long, api- cally prominently punctate, the style 3.2-3.4 mm long, epunctate, the ovules 16 to 18. Fruits globose, 6-10 mm diam., prominently punctate. Distribution. Ardisia panamensis is endemic to Cerro Horqueta in o Panama, growing at 1750 to 2268 m in elevation Ecology and conservation status. Fieldwork by Pipoly has revealed that Ardisia panamensis occurs along the forest edge at the junction of the montane and cloud forests. It is locally common, but should be considered threatened because it borders the life zone preferred for coffee plantations. The specific epithet was derived Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia panamensis belongs to a group of species including A. dwyeri and A. vesca because of its short calyx lobes, long pedicels, thin branch- lets, wide corolla lobes, and long anthers. However, Ardisia panamensis is easily separated from the oth- er species by its extremely long pedicels to 15 mm long, and on most specimens the highly geniculate branches of the inflorescence and persistent sec- ondary branch bracts are highly distinctive. Specimens examined. PANAMA. Chiriquí: C ordille ra О); Cerro Monata, 8 km NW of Boquete, at top, 20 Oct. 1980 (fr), P. Maas & R. | абы 4952 (NY, U); Dtto. Boquete, Cerro Horque la, 28 June pes (fl). J. Pipoly 7065 (DUKE, FTG, i rail to Cerro Horqueta, 15 May 197 Я d fr. G. Proctor 31907 (LL [2]: 5 Ke of Cerro "e A of Boquete, 21 Jan. 1971 (A, fr), R. Wilbur et al. (DUKE) 57. Ardisia pseudoracemiflora Pipoly, Caldasia 16(78): 279. 1991. TYPE: Colombia. Nariño: Mpio. de Barbacoas, on the road from Barba- coas to Junín, 1050 m, 7 Aug. 1962 (fl), L. Mora 2292 (holotype, COL!; isotypes, PSO!, US!). Figure 57. Trees 30—40 m tall. Branchlets stout, terete, 6—7 mm diam., densely and minutely rufous furfura- ceous-lepidote. Leaves with blades chartaceous, el- liptic, 12.6-29.6 X 5.1-9.7 cm, apically acute, with an acumen 2-7 mm long, basally obtuse, de- current on the petiole, prominently punctate and punctate-lineate, densely furfuraceous-lepidote, of- ten glabrescent above, the midrib impressed above, prominently raised below, the secondary veins 55 to 64 pairs, prominulous above and below, the mar- gins entire, inrolled; petioles slender, canaliculate, 5-11 mm long, 2-3 mm diam., furfuraceous-lepi- dote above and below at first, then glabrescent above. Inflorescences a pseudoracemose pinnate panicle, 17-22.5 X 3-8 cm, usually shorter than the leaves, the rachis, branches, abaxial floral bract surfaces, and pedicels furfuraceous-lepidote, the branches loosely congested into 3- to 5-flowered corymbs, the lower branches often subtended by leaves; inflorescence bracts unknown; inflorescence branch bracts caducous, membranous, oblong, 2.9— 1 X 1-1.5 mm, apically acute, prominently punc- tate and punctate-lineate, glabrous above, furfura- ceous-lepidote below, the margins irregular, mi- nutely erose, hyaline, sparsely glandular ciliolate: floral bracts similar to the inflorescence branch bracts, but 1.6-2.5 X 0.5-0.9 mm: pedicels stout, 1.8-3.3 mm long, inconspicuously punctate and punctate-lineate, furfuraceous-lepidote. Flowers 5- merous, white to light pink; calyx lobes charta- ceous, ovate, 2.6-2.8 X 1.8-2 mm, apically round- ed, prominently punctate and punctate-lineate, glabrous adaxially, furfuraceous-lepidote abaxially, the margin entire, minutely erose, hyaline, sparsely glandular ciliolate; corolla chartaceous, 6.4—6.6 mm long, the tube 1.2-1.3 mm long, the lobes ovate, 2.4—2.5 mm, prominently punctate and punctate-lineate, gla- brous throughout, the margin entire, hyaline; sta- mens 6—6.2 mm long, the filaments 2.9-3 mm long, the staminal tube 1—1.1 mm long, the apically free portions 1.8-2 mm long, the anthers lanceoloid, 3 1.1-1.2 mm, apically cuspidate-apicu- late, basally cordate, the connective conspicuously punctate; pistil 7.6-7.8 mm long, glabrous, the ova- mm long, the style 5.1-5.5 mm apically acute, ry globose, 2.3-2.5 long, slender, erect, inconspicuously punctate, the ovules 28 to 35. Fruits unknown. Distribution. Ardisia pseudoracemiflora is known only from western slopes of the western An- dean cordillera, near Barbacoas, in Nariño, Colom- bia, growing at 1050 m in elevation. Ecology and conservation status. Ardisia pseu- doracemiflora occurs in one of the floristically rich- est areas of the Cordillera Occidental of Colombia. It occurs in montane pluvial forest, which receives well over 8000 mm of rain per year. The area is also known to house many endemies, including Clusia garciabarrigae, C. niambiensis, C. tetragona, Ardisia niambiensis, four more undescribed Myrsi- naceae, and many new species from other families, especially Araceae and Gesneriaceae. The area is undergoing rapid development, and while there are Annals of the Missouri Botanical Garden several nature reserves, many forests have been threatened by new roads, communication towers, and rural electrification projects. For these reasons, the species should be considered threatened. Etymology. Тһе specific epithet refers to. the overall shape of the inflorescence, resembling a ra- ceme. Within Ardisia subg. Pal- manae, Ardisia pseudoracemiflora may be most eas- Auriculardisia sect. ily confused with A. hagenii because of its elliptic leaves and pseudoracemose inflorescence. Howev- er, А. pseudoracemiflora is easily separated from A. бериш by its narrower, chartaceous calyx lobes to 2 mm wide, shorter corolla lobes to 6.6 mm long, shorter anthers to 3.2 mm long, shorter styles to 5.5 mm long, larger number of secondary veins of the leaf blades, and thinner branchlets. 58. Ardisia pulverulenta Mez, in Engl., Pflan- zenr. IV. 236 (Heft 9): 88. 1902. Auriculardisia pulverulenta (Mez) Lundell, Phytologia 54: 285. 1983. TYPE: Panama. Veraguas: Cap Corrientes, Feb. 1848 (fl), B. Seemann 1093 (lectotype, designated by Lundell (1968), K!, LL neg. 1971-88!; isolectotype, BM!). Figure 58. Shrubs or trees. terete, exfo- liating, 3-3.5 Leaves with blades membranous, elliptic, 18.5-20 Branchlets slender, mm diam., furfuraceous-lepidote. 0.7—7.9 cm, apically acute, with an acumen 5— 7 mm long, basally obtuse, decurrent on the petiole, inconspicuously punctate and punctate-lineate, gla- brous above, furfuraceous-lepidote below, the mid- rib impressed above, prominently raised below, the secondary veins 36 to 42 pairs, prominulous above revo- and below, the margins minutely crenulate, lute; petioles slender, marginale, 3-5 mm long, gla- brous above, densely furfuraceous-lepidote below. Inflorescences erect, pinnate or bipinnately panic- 11-13.5 than the leaves, furfuraceous-lepidote, the branches ulate, 2.5—4.2 cm, pyramidal, shorter loosely congested into 5- to 7-flowered corymbs; peduncles obsolete; inflorescence bracts unknown: membra- inflorescence branch bracts caducous, nous, ovate, 1.5-2.2 X 1.2-1.3 mm, apically acute, س prominently punctate and punctate-lineate, gla- brous adaxially, furfuraceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflo- rescence branch bracts, but 1—1.3 X 0.5-0.7 mm: pedicels slender, 1.5-2.5 mm long. prominently punctate and punctate-lineate, furfuraceous-lepi- dote. Flowers 5-merous, appearing light pink 1.2-1.3 X red; calyx lobes membranous, ovate, 0.7-0.8 mm, apically acute, prominently punctate and punctate-lineate, furfuraceous-lepidote, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 2.4-2.6 mm long, the tube 0.6-0.9 mm long, the lobes nar- rowly ovate, 1.7-1.8 X 1-1.2 mm, apically acute, prominently punctate and punctate-lineate, gla- brous throughout, the margins entire, hyaline; sta- mens 1.7-1.8 mm long, the filaments 0.7—0.8 mm long, the staminal tube 0.3—0.4 mm long, the api- cally free portions 0.3-0.5 mm long, the anthers 1.1-1.2 X 0.6-0.7 basally lobate, the connective conspicuously punc- ovoid, mm, apically apiculate, _ ate; pistil 1.5-1.6 mm long, glabrous, the ovary globose, 0.2-0.3 mm long, the style 1.2-1.4 mm long. prominently punctate, the ovules 12 to 16. Fruits unknown. Distribution, Ardisia pulverulenta is known only from the type collection, with a vague description of location, and no elevation. Ecology and conservation. status. Because of the vague description the ecology of Ardisia pulve- rulenta is unknown. However, because it is known only from the type, it should be considered threat- ened. Etymology. Тһе specific epithet comes. from Mez’s description of the “flores pulverulento-lepi- doti" meaning the lepidote scales of the flowers covered with a fine bloom or powdery matter. Ardisia pulverulenta is most closely related to А. knappii (see under that species for similarities). However, A. pulverulenta differs from A. knappii by its thinner branchlets to 3.5 mm in diameter, short- 7.9 em, shorter er and narrower leaf blades to 2X nar- rower inflorescence to 4.2 cm wide, and narrower calyx lobes to 1.3 X 0.8 mm, shorter and narrower corolla lobes to 1.8 X 1.2 mm, narrower anthers to 0.7 mm wide, and shorter styles to 1.4 mm long. 59. Ardisia ruedae Ricketson & Pipoly, sp. nov. TYPE: Nicaragua. Río San Juan: Mpio. de San Juan del Norte, Reserva Indio-Maíz, down riv- er LO km from Cerro Canta Gallo, La Chiripa hunting trail, 11°07'N, 083754' W, 100 m, 18 Sep. 1998 (fl), R. Rueda, I. Coronado, W. Ve- lásquez & Y. Rubi 8765 (holotype, MO!; iso- type, HULE not seen). Figure 59. Propter laminam foliarem ellipticam, pedicellos usque ad 4.5 mm longos atque stylos usque ad 3.5 mm longos A. дилат. similis, sed ab еа ramulis semiteretibus (non teretibus), laminis foliaribus coriaceis ue membra- naceis) nerviis secundariis 38 ad 53 (nec 23 ad 31)-jugis, lobulis calycinis ovatis (non orbicularibus) denique ant- heris 0.8-1.0 (non 1.1-2.1) mm latis statim cognoscitur. i Ricketson & Pipoly 285 2 $ r Volume 90, Number 2003 Revision of Ardisia subg. Auriculardisia ppany "M Wor q (ОҢ ) ((OW) 22001 1” 12 928 `P 12 Dpany "4 *ed&jo[oi шоу имрір J-y) зиш "(J— moja ')— “png moy *q— ‘Ҷоиза Suuawo[4 "V— ‘appans pisipay 'Qu3u) бс әл CONS) £60] чиошәәс̧ ‘gq wor umeap g *y) чәмо]] ^g— ‘YouRsq #@шләмор] ‘W— "ориәрпләата pisipay (әр) 8S ANSIA 286 Annals of the Missouri Botanical Garden Trees 5—20 m tall. rete, 3.5—5.5 mm diam., densely and minutely ap- Branchlets slender, semi-te- pressed rufous furfuraceous-lepidote. Leaves with blades coriaceous, elliptic, 5.5-15.6 X 2.7-7.1 ст, apically acuminate, with an acumen 5-13 mm long, basally acute, decurrent on the petiole, prominently punctate and punctate-lineate above and below, glabrous above, densely and minutely appressed furfuraceous-lepidote below, the midrib impressed above, prominently raised below, the secondary veins 38 to 53 pairs, inconspicuously raised above and below, the margins entire, flat; petioles slender, marginate, 6-14 mm long, 2-3 mm diam., glabrous above, densely and minutely appressed rufous fur- furaceous-lepidote below. Inflorescences erect, bi- pinnately paniculate, 4.8—10.5 X 4—7.5 em, pyra- midal, usually shorter to slightly longer than the leaves, the rachis, branchlets, abaxial bract surfac- es, and pedicels densely and minutely rufous fur- furaceous-lepidote, the branches congested into 5- to 9-flowered corymbs; peduncles 1.1-2.8 em long, the lower branches subtended by leaves; inflores- cence bracts unknown; inflorescence branch bracts floral bracts caducous, 0.8-1.2 X 0.8-1.2 apically acute, prominently punctate and punctate-lineate, gla- brous above, densely and minutely appressed ru- unknown; membranous, ovate, mm, fous furfuraceous-lepidote below, the margins irreg- ular, minutely erose, hyaline, sparsely glandular ciliolate; pedicels stout, 1.8—4.5 mm long. incon- spicuously punctate and punctate-lineate, densely and minutely appressed rufous furfuraceous-lepi- dote. Flowers 5-merous, yellow to purple; calyx lobes coriaceous, ovate, 2.4-2.6 X 2.5-3 mm, api- cally rounded, prominently punctate and punctate- lineate, glabrous adaxially, densely furfuraceous- lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla coriaceous, 6.4—6.7 mm long, the tube 2-2.1 mm 4.44.6 X 1.8-2 mm. apically acute, prominently punctate and punctate- lineate, glabrous throughout, long, the lobes lanceolate, the margins entire; stamens 5-5.5 mm long, the filaments 1.8-2.2 mm long, the staminal tube 0.8-1 mm long, the apically free portions 2.8-3 mm long, the anther ovate, 2.6- 2.9 X 0.8-1 mm, apically apiculate, basally deeply cordate, the connective conspicuously punctate; pistil 4.5-5.1 mm long, glabrous, the ovary oblong, 1-1.6 mm long, the style 3-3.5 mm long, promi- nently punc tate and punctate-lineate, the ovules 38 to 44. Fruits globose, 7.58.2 mm diam., incon- spicuously punctate. Distribution. Ardisia ruedae is endemic to the Atlantic Slope of Nicaragua, in Río San Juan and Zelaya, growing at 50 to 412 m in elevation. Ecology and conservation status. Ardisia ruedae occurs in tall, pluvial forests on lateritic soils, and label data indicate that it is locally common. Rain- fall in this area of Central America is comparable only to the wettest area of the Chocó Floristic Prov- ince of Colombia. Because of the remoteness of the populations, and the fact that the land it occurs on is protected, it is probable that this species does not face immediate threat. Etymology. It is our pleasure to name this spe- cies in honor of Ricardo Rueda, dean of arts and sciences at the Universidad Nacional Autónoma de Nicaragua-León, who is a specialist in Verbena- ceae. Within Pal- manae, Ardisia ruedae is similar to А. dunlapiana Ardisia subg. Auriculardisia sect. because of its elliptic leaves, pedicels up to 4.5 mm long, and styles up to 3.5 mm long, but may be separated from it by the semiterete branchlets, co- riaceous leaf blades with more secondary veins, ovate calyx lobes, and shorter anthers to 2.9 mm ong. Paratypes. Meus ici Río San Juan: Mpio. de San ie del Norte, Reserva Indio-Maíz, Cerro El Gigante, ‚ 1996 же ), R. Rueda et al. 4458 (HU LE, МО), . Rueda et al. 9005 (HULE, MO). Ze- la aya: Mpio. P Nuva Guinea, эрши. Indio-Maíz, Río Pijibaye entre el cano Bijagua y El Cerro Chiripa, 13 Jan. 1999 (fr), R. Rueda et al. id mt L E. MO), (fl), К. Rueda et al. 10027 (HULE, . 15 Jan. 1999 (fr), R. Rueda et al. 10131 (HULE. ow 60. Ardisia smurfitana hicketson & Pipoly, sp. nov. E: Colombia. Valle del Cauca: Bajo Calima, Concesión Pulpapel-Buenaventura, carretera Nacional km 28, ca. 100 m, 03°55'N, O77°W, 26 July 1989 (fl, fr), M. Monsalve B. 3111 (holotype, CUVC!; isotypes, FTG!, MO!). Figure 60. ropter lobulos calycinos 0.9-1.3 X 0.7-1.1 mm, atque laminam foliarem elliptic am vel oblongam A. pulverulentae similis sec ae ea laminis foliaribus coriaceis (non mem- branaceis). inflorescentiis 25—32 (non 2.5-4.2) em oa lis о lobulis pes yeinis 0.8-1.0 X 0.6-1.1 (non 1.2-1.3 .4—0.8) mm praeclare distat. Unknown habit or height. Branchlets stout, te- rete, horizontally checking and exfoliating. 8—11 mm diam., densely furfuraceous-lepidote. Leaves with blades coriaceous, elliptic to oblong, 23-23.5 X 5-5.6 em, apically acute, with an acumen 5—7 mm long, basally acute to cuneate, decurrent on the petiole, inconspicuously punctate and punctate-lin- eate, glabrous above, furfuraceous-lepidote below, 287 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia Volume 90, Number 2 2003 (OW) 28882 `P 12 asprang “9 *ad&josr шолу имвлр g ^y) N14 77— moy 'g— "queiq uto] ү 7901040] Disipsy ((OW) IIIE ^g ?apisuogy ‘W ‘edAjojoy шолу umeıp g ^v) зит ^g— qoueiq #шләмор] "ү ‘puDjfinus терү Уу uw Pi 4 7 ee T COO WW) (аца) [9 әл (yop 09 әли 288 Annals of the Missouri Botanical Garden the midrib impressed above, prominently raised be- low, the secondary veins inconspicuous, 68 to 79 pairs, the margins entire, revolute; petiole stout, marginate, 1—1.3 mm long, glabrous above, furfu- raceous-lepidote below. /nflorescences erect, bipin- nate to tripinnately paniculate, 24—25 X 28—32 cm, pyramidal, usually longer than the leaves, furfura- ceous-lepidote, the branches loosely congested into 5- to 11-flowered corymbs; peduncles 8-10 mm long, densely furfuraceous-lepidote; inflorescence and branch bracts unknown; floral bracts unknown: pedicels slender, 1-1.3 mm long, inconspicuously punctate and punctate-lineate, densely to sparsely calyx 0.8-1 х 0.6-1.1 mm, apically acute to rounded, prominent- furfuraceous-lepidote. Flowers 5-merous; lobes membranous, orbicular to oblate, ly punctate and punctate-lineate, sparsely furfura- ceous-lepidote, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla mm long, the tube 0.9-1.2 mm long, the lobes narrowly lanceolate, 2.6-2.9 х membranous, 3.9—4. 1.4—1.6 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the mar- gins entire, hyaline; stamens 2.8-3.7 mm long, the filaments 1.9-2.1 mm long, the staminal tube 0.2— 0.4 mm long, the apically free portions 1.5-1.7 mm 1.6-1.7 X 1.6-1.7 apically apiculate, basally lobate, the connective long, the anthers ovoid, mm, conspicuously punctate; pistil unknown. Fruits glo- bose, 1.4-1.6 mm diam., conspicuously punctate and punctate-lineate. Distribution. Ardisia smurfitana is known only from the type collection in Bajo Calima in Valle del Cauca, Colombia, growing at about. 100 m in ele- vation. Ecology and conservation status. Ardisia smur- Мапа occurs in lowland pluvial forest. Because it is apparently endemic in the most valuable timber stand on the continent, it should be considered threatened. Etymology. The specific epithet does not refer to blue, elf-like creatures, but rather, is named for the Smurfit Carton Colombia Company. whose gen- erous subsidies provided a group of investigators with opportunities to better study and understand the flora of Bajo Calima. Pal- manae, Ardisia smurfitana is most closely related Within Ardisia subg. Auriculardisia sect. to А. hugonensis (see under that species for simi- larities). However, A. smurfitana differs from A. hu- its branchlets horizontally checking larger leaf blades to 23.5 X 5.6 em; shorter, stouter, marginale, petioles to 1.3 mm gonensis by and exfoliating: long; and membranous, orbicular to oblate, smaller calyx lobes to 1 X 1 mm. 61. Ardisia tarariae Lundell, Phytologia 61: 67. 1986. Auriculardisia tarariae (Lundell) Lun- dell. Phytologia 63: 75. TYPE: Costa Rica. Limón: Cordillera Talamanca, Atlantic slope. Cerro Tararia, locally known as Tres Pi- cos, 09°09'N, 082°58'W, 2400-2600 m, 10 Sep. 1984 (fl), G. Davidse, G. Herrera Ch. & M. Grayum 28882 (holotype. MO!). Figure 61. LL; isotype, Treelets to 4 m tall. Branchlets slender, the in- terpetiolar ridges forming up to 5 angles, 3-5 mm diam., densely furfuraceous-lepidote. Leaves with blades membranous to chartaceous, elliptic to ob- lanceolate; 2.2-16.6 X 1 nate, with an acumen 4—14 mm long, basally cu- —5.2 ст, apically acumi- neate, decurrent on the petiole to the stem, prominently punctate and punctate-lineate above and below, glabrous above, densely furfuraceous- lepidote below, the midrib impressed above, prom- inently raised below, the secondary veins 23 to 2 pairs, prominulous above and below, the margins entire to shallowly crenate, flat; petioles slender, marginale, 6-19 mm long, glabrous above, furfu- raceous-lepidote below. Inflorescences erect, bipin- nately paniculate, 9-22 4—19 ст, pyramidal, longer than the leaves, the rachis and branches mi- nutely rufous cupuliform lepidote, the branches loosely congested into 5- to 11-flowered corymbs: peduncle obsolete to 0.5 em long, the lower branch- es subtended by leaves; inflorescence bracts un- known; inflorescence branch bracts caducous, membranous, oblong, 2.3-3.2 X 0.4—0.8 mm, api- cally acute, prominently punctate and punctate-lin- eale, glabrous above, furfuraceous-lepidote below, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflorescence branch bracts, but 0.9-2 X 0.3— 0.4 mm; pedicels slender, 4.5—6 mm long, recurved upward, inconspicuously punctate and punctate- lineate, sparsely furfuraceous-lepidote. Flowers 5- merous, light pink; calyx lobes membranous, sub- 1.2-1.4 prominently punctate and punctate-lineate, gla- — orbicular, —1.2 mm, apically acute, brous throughout, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla 3.6-3.8 mm long, the tube 0.6-0.8 mm long, the lobes ovate to narrowly ovate, 2.8-3 membranous, X 1.6-1.7 mm, apically acute, prominently punc- tate and. punctate-lineate, glabrous throughout, the margins entire, hyaline; stamens 3.8—4 mm long, the filaments 2.5-2.6 mm long, the staminal tube Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 289 0.5-0.6 mm long, the apically free portions 2-2.1 mm long, the anthers ovoid, 1.5-1.7 X 0.7-0.9 apically subulate-apiculate, um deeply cordate, mm, the connective conspicuously punctate; pistil 3.7-3.9 mm long, glabrous, the ovary globose, 0.7-0.8 mm diam., the style 3-3.1 mm long, epunc- tate, the ovules 18 to 19. Fruits globose, 3-5 mm diam., prominently punctate. Distribution. Ardisia tarariae is known from the Cordillera Talamanca near the Costa Rica-Panama border in Limón and Puntarenas, Costa Rica, grow- ing at 2100 to 2750 m in elevation. Ecology and conservation status. Ardisia tara- riae occurs at the interface of upper cloud and elfin forest. In this geographic area, the forests are dom- inated by oaks, but also have sizeable Podocarpus populations as well. Because it is known from so few specimens, its current conservation status 18 The specific epithet was derived from the type locality on Cerro Tararia. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia tarariae is most closely related to A. tenuicaulis and A. tenuis because of its very small calyx lobes, less than 1.4 mm long, long ped- Ar- disia tarariae can easily be distinguished from both related taxa by its larger calyx lobes to 1.4 X mm, longer and wider corolla lobes to 3 X 1.7 mm, icels, narrow corolla lobes, and short anthers. wider anthers to 0.9 mm wide, longer styles to 3.1 mm long, and smaller fruits to 5 mm in diameter. Specimens examined. COSTA RICA. Limón: Cordil- lera Talamanca, Atlantic slope, Valle de Sile ncio, area a N sta PTE border, 4 vidse et al. 28656 (LL, MO), 28703 (MO); Cordillen ra Ta- lamanca, Atlantic slope, unnamed cordillera between the Río Terbi and the Río Sinî. 13 Sep. 1984 (fr), С. Davidse et al. 29065 (LL, MO). ^ ра nas: Cantón de Coto Brus, Parque Incipe: na La Amistad, Сог Ше Talamanca, Es- tación Altamira, path from the Casa de Coca to the Valle del Silencio, 17 Apr. 1995 (fr), L. Angulo 199 (INB. MO). ~ 62. Ardisia tenuicaulis Lundell, Wrightia 6: 110. 1980. /cacorea tenuicaulis (Lundell) Lundell, Phytologia 49: 352. 1981. TYPE: Panama. Chiriquí: Fortuna Dam site, 1400—1600 m, 15 Sep. 1977 (fr), J. Folsom, R. Dressler & K. Dressler 5561 (holotype, MO!). Figure 62. Shrubs to 1 m tall. Branchlets slender, angled, 1.5-3 mm diam., sparsely and minutely ferrugine- ous furfuraceous-lepidote. Leaves with blades mem- branous, elliptic, 3.7-8.8 X 1.6-3.9 cm, apically acuminate to caudate, with an acumen 5-12 mm long, basally acute to obtuse, decurrent on the pet- iole, prominently punctate and punctate-lineate, glabrous above, sparsely and minutely furfura- ceous-lepidote, the midrib impressed above, prom- inently raised below, the secondary veins 20 to 25 pairs, prominulous above and below, the tertiary veins above and below, the margins entire, flat; pet- ioles slender, canaliculate, 5—7 mm long, glabrous above, sparsely furfuraceous-lepidote below, the margins entire or with small teeth. /nflorescences terminal pendent, bipinnately paniculate, 4—8 X 4— 7 em, pyramidal, longer than the leaves, sparsely furfuraceous-lepidote, the branches loosely con- gested into 5- to 7-flowered corymbs; peduncles 4— 6 mm long, the lower branches subtended by leaves; inflorescence bracts unknown; inflorescence branch bracts caducous, membranous, lorate, 1.9— 3.7 X 0.5-0.9 mm, apically acute, prominently punctate and punctate-lineate, glabrous adaxially, usually glabrous abaxially or with a few scattered scales, the margins irregular, minutely erose, hya- line, rarely glandular ciliolate; floral bracts similar bracts, but 0.8-1.1 X 0.2—0.3 mm; pedicels slender, recurved, 4—5.2 mm to the inflorescence branch long, epunctate to inconspicuously punctate and punctate-lineate, nearly glabrous or sparsely fur- furaceous-lepidote. Flowers 5-merous, light pink; [measurements from buds] calyx lobes membra- orbicular to oblate, 1-1.2 х 1-1.3 apically acute, prominently punctate and punctate- nous, mm, lineate, glabrous throughout or sparsely furfura- ceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular cilio- late; corolla membranous, 2.7—3 mm long, the tube 0.3—0.4 mm long, the lobes ovate to narrowly ovate, 2.4-2.6 X 1.2-1.4 punctate punctate-lineate, glabrous throughout, mm, apically acute, prominently the margins entire, hyaline; stamens 2-2.2 mm long. the filaments 0.2-0.4 mm long, the staminal tube 0.1-0.2 mm long, the apically free portions 0.1-0.3 mm long, the anthers ovoid to narrowly 1.8-1.9 X basally оре, the connective conspicuously punc- ovoid, 0.5-0.7 mm, apically apiculate, late; pistil 2.2-2.4 mm long, glabrous, the ovary ovoid, 0.8-0.9 mm long, the style 1.3—1.5 mm long, epunctate, the ovules 5 to 7. Fruits globose, 7—8 mm diam., inconspicuously punctate. Distribution. Ardisia tenuicaulis is known only from the Fortuna Dam area of Chiriquí, Panama, growing from 1200 to 1600 m in elevation. Ecology and conservation status. Ardisia tenui- caulis occurs in premontane wet forests. It is known only from one locality, and given the intensity with which the area has been surveyed, we should as- sume the species is rare and thus threatened. Annals of the 290 Missouri Botanical Garden | | ((OW) 2629 I? 12 u0s]o4 р шолу 7) (QIN) 6F02 12021) "V YP A42?) "y шо g (SN) ZOLET SPA ; X AmMq T 79dsjost шолу umeip y) mig 77)— сләмор "q— ‘yoursq Зицәмор ү snua терү — "(auSu) EQ oanzi4 | ((ON) r9ce n 12 шо$]о 4 T *edÁjo[ou шолу 9 (OW) FIIOI ррирліү T БУ UOSLIYU IW 9 шолу UMPIp Ч ‘v) "una e “png IƏMO| 4 'H— “YIURIG SULIOMO] J D === "s2mpoimmuo] Disipiy "(yop c9 ainsi J Volume 90, Number 2 2003 Ricketson & Pipoly 291 Revision of Ardisia subg. Auriculardisia Etymology. The grece epithet was derived from the Latin “tenui” psum slender or thin and *-caulis" meaning sten Within. Ardisia sulle, Auriculardisia sect. Pal- manae, Ardisia tenuicaulis has been confused with A. panamensis, and is vegetatively very similar. However, the extremelv small calyx lobes align it more closely with A. tenuis, from which it is sepa- rated by the longer calyx to 1.2 mm long. longer corolla lobes to 2.6 mm long, longer anthers to 1.9 mm long, shorter style to 1.5 mm long, and much larger fruits to 8 mm in diameter. Ardisia tenuicau- lis is most closely related to A. tarariae and A. ten- uis because of its very small calyx lobes to 1.2 X 1.3 mm, longer pedicels 5.2 mm long, narrower co- rolla lobes to 1.4 mm wide, and shorter anthers to 1.9 mm long. i a imen examined. PANAMA. Chi Fortuna Dam, in valley S of lake, 25 Dec. MP haan & J. Aranda 10114 (MO) riquí: vicinity of 1986 (fl). G. 63. Ardisia tenuis Lundell, Wrightia 4: 149. 70. Icacorea tenuis (Lundell) Lundell, Phytologia 49: 352. 1. Auriculardisia ten- uis (Lundell) Lundell, Wrightia 7: 273. 1984. TYPE: Panama. Darién: Cerro Pirré, 2500— 4500 ft. [762-1372 m], 9-10 Aug. 1967 (fl). J. Duke & T. Elias 13762 (holotype, LL, LL neg. 1979-4!; isotypes, GH!, MO!, US!). Figure 63. Мм Ardisia p Lundell, Phytologia 48: 134. 1981. Syn. Auriculardisia pirreana (Lundell) Lundell, Гый да, 49: 345. . TYPE: Panama. Darién: Cerro Pirré, ridge top near Rancho Plastico, 1200 m, 10-20 July 1977 (fl). J. Folsom, R. Hartman & R. Dressler 4251 (holotype, LL!; isotypes, MEXU! MO!). Shrubs 3 m tall. Branchlets slender, terete, 1—3.5 - = = © mm diam., densely cupuliform and furfuraceous- lepidote. Leaves with blades membranous, elliptic to narrowly ovate, 2.1—11 X 0.6-3.7 cm, apically caudate-acuminate, with an acumen 6—19 mm long, basally obtuse to rounded, decurrent on the petiole, prominently punctate pna punctate-lineate, gla- brous above, sparsely lepidote below across the blade, more densely mixed cupuliform and furfuraceous-lepidote along the midrib, the midrib impressed above, prominently raised below, the secondary veins 25 to 34 pairs, prominulous above and below, the margins entire to minutely crenulate, flat; petioles slender, marginate, 4—9 mm long, glabrous above, densely cupuliform and fur- furaceous-lepidote below. Inflorescences erect. bi- pinnate to tripinnately paniculate, 5-16 X 5-11 cm, pyramidal, longer than the leaves. indument of dense scales, obsolete or short stalked, flat or cu- pulate with the margins entire or deeply toothed, the branches loosely congested into 5- to 9-flowered corymbs; peduncle nearly obsolete to 0.9 cm long. the lower branches subtended by leaves; inflores- cence bracts unknown; inflorescence branch bracts caducous, membranous, oblong or ovate, 1.3—6.2 X 0.3-2.5 mm, apically acute, or short petiolate, the midrib impressed above, prominently raised below, prominently punctate and punctate-lineate, gla- brous above, furfuraceous-lepidote as in leaves be- low, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflorescence branch bracts, but 0.4—0.7 X 0.1—0.2 mm; pedicels slender, 4.5—6.2 mm long, inconspicuously punctate and punctate-lineate, densely cupuliform and furfuraceous-lepidote. Flowers 5-merous, cream or white; calyx lobes chartaceous, ovate, 0.9-1.2 X 0.7-0.8 mm, apical- ly acute, prominently punctate and punctate-lin- eate, glabrous adaxially, mixed lepidote abaxially, hyaline, sparsely glandular ciliolate; corolla membranous, .5-0.6 mm long, the lobes ovate to aine son ovate, 2.3-2.4 X 1.4-1.5 the margins irregular, minutely erose, 2.9-3 mm long, the tu mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the margins entire, hyaline; stamens 2.1-2.3 mm long, the fil- aments 1.1—1.3 mm long, the staminal tube 0.3— 0.4 mm long, the apically free portions 0.7—1 mm long, the anthers ovoid, 1-1.1 X 0.6-0.7 mm, api- cally cuspidate-apiculate, basally subcordate, the connective conspicuously punctate; pistil 3.1—3.6 mm long, glabrous, the ovary ovoid, 0.5-0.7 mm long, the style 2.6-2.9 mm long, prominently punc- tate and punctate-lineate, the ovules 4 to 6. Fruits globose. 4.3—4.7 mm diam., prominently punctate. Distribution. Ardisia tenuis is known from Cerro Pirré in Darién, Panama, and on Alturas de Nique in Chocó, Colombia, on the Panama-Colombia bor- der. The collection bearing the label Lawrance 114 is probably mislabeled, because the flora of Cerro Chapón in Boyacá is totally unrelated to that of the Chocó, and we know of no non-weedy species of angiosperms that are shared between the two. It has een collected at 750 to 1500 m in elevation. Ecology and conservation status. Ardisia tenuis occurs in pluvial montane, cloud, and elfin forest. Although locally abundant, because of its restricted distribution it should be considered threatened. Etymology. The specific epithet was derived from the Latin meaning thin, fine or slender, and refers to the very thin branchlets and inflorescence rachis. 292 Annals of the Miu. Botanical Garden Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia tenuis is most closely related to A. tarariae and A. tenuicaulis because of its very small calyx lobes, less than 1.4 mm long with long ped- icels, narrow corolla lobes, and short anthers. Ar- disa tenuis can easily be distinguished from A. ta- rariae by its smaller calyx lobes to 1.2 X 0.8 mm, shorter and narrower corolla lobes to 2.4 X 1.5 mm, narrower anthers to 0.7 mm wide, shorter styles to 2.9 mm long, and larger fruits to 4.7 mm in di- ameter. Ardisia tenuis can be separated from А. te- nuicaulis by its narrower calyx lobes to 0.8 mm wide, shorter but wider corolla lobes to 2.4 X 1.5 mm, shorter anthers to 1.1 mm long, longer styles to 2.9 mm long, and smaller fruits to 4.7 mm in diameter. The type of Ardisia pirreana is unique only for its slightly larger inflorescence and leaf blades: however, these fall well within the range of A. ten- uis. Specimens examined. PANAMA. Darién: Cerro Pirré, 07 r), N. Bristan 468 (MO); Serranía de of Serranía de Pirré above Cana Gold > Cana and Río Escucha Ruido, 27 July . T. Croat “ее MO); Parque Nacional del Da- X S branch of Río E ig "Pus ‘uro, 21 Oct. 1987 (fl), H. NATI el a 39. 39 (MO) ridgetop area N of Cerro Pirré, between ( › Pirré top and Rancho Plastico Nov. r). Н Folsom et seni 629 7 (MO); нени D "a го 972 (fl), A. Gentry & A. Clewell 7049 (F, МО); Parque Nac ional ER D: arién, Panama—C м bor- mine at he ныны of N brane : of Río Pucuro slopes of Cerro او‎ ına, ca. 6 km ` Cerro 0 G. de Never rs et i: 8523 il L МО); Сыра ig in a of Cana-Cuasi. trail, Mar. 940 & R. Terry 1568 (F). CE [Doubtful, probably represents а label mix-up—Boyacá: Chapón region, 100 mi. NW of Bogotá, 25 May 1932 (fr), A. Lawrance ПА (GH, МО). Chocó: SW idge leading to Alturas de Nique on border with Panama, 29 Da . 1980 (fl, fr), R. желез 12376 (LL, MO); Alturas de Riu and ridge to SW, 31 Dec. 1980 (fl, fr), R. Hartman 12473 (LL, MO). T — - m m ea» » = ~ — т” т T = 64. Ardisia tysonii Lundell, Wrightia 4: 165. 1971. Auriculardisia tysonii (Lundell) Lundell, d 49: 345. 1981. TYPE: Pa : Cerro Jefe, in Clusia forest, 2700— о d [823-914 т], 27 Jan. 1966 (fr), Е. Tyson, J. Dwyer & К. Blum 3279 (holotype, MO!, LL neg. 71-117*; 99646!, MO!). Figure 64. Panama. isotypes, LL!, F neg. ашы sais l undell, Wrightia 6: 86. 1979. Syn. nov. Phyto- € oclé: La Mesa above El Valle de Jes ca. of Cerro Pilón, on slopes of steep knife-like ds 900—930 m, 22 Mi rue July 1976 (fr), Т. Croat 37431 (holotype, MO!, LL neg. 1979-3!; isotype, LL!). — Shrubs 3-7 m tall, 4-12 cm Branchlets stout, terete, 5—8 mm diam., densely fur- or trees diam. furaceous-lepidote. Leaves with blades coriaceous, elliptic, 3.7-12.2 with an acumen 3-8 mm long, basally acute, de- 2.1-6.2 mm, apically acute, current on the petiole, inconspicuously punctate and punctate-lineate above and below, glabrous above, densely furfuraceous-lepidote below, the midrib impressed above, prominently raised below, the secondary veins 27 to 38 pairs, prominulous above and below, the margins entire, flat; petioles slender, marginate, 8-17 mm long, glabrous above, furfuraceous-lepidote below. /nflorescences erect, 4—12 cm, dal, usually longer than the leaves, sparsely furfu- bipinnately paniculate, 6-14 х pyrami- raceous-lepidote, the branches loosely congested into 5- to 7-flowered corymbs; peduncles nearly ob- solete to 2 ст long, the lower branches subtended by leaves; inflorescence bracts unknown: inflores- cence branch bracts caducous, membranous, ovate to oblong, 0.8—4.5 X 1.2-1.8 mm, apically acute, prominently punctate and punctate-lineate, gla- brous above, furfuraceous-lepidote below, the mar- gins irregular, minutely erose, hyaline, sparsely glandular ciliolate; floral bracts similar to the inflo- X 0.5-0.8 mm; pedicels stout, 0.7-1.8 mm long, inconspicuously rescence branch bracts, but 1—1.4 punctate and punctate-lineate, furfuraceous-lepi- dote to glabrescent. Flowers 5-merous, white to light pink; orbicular lo ovate, calyx lobes membranous to chartaceous, 1.2-1.6 X 1.4-1.6 mm, apically acute to rounded, prominently punctate and punc- tate-lineate, densely furfuraceous-lepidote, glabres- cent, the margins irregular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 4.1—4.2 mm long, the tube 1.2-1.3 mm long, the 2.9-3 X 1.6-1.9 mm, cally acute, prominently punctate and punctate-lin- eate, lobes narrowly ovate, api- glabrous throughout, the margins entire, hyaline: stamens 3.2-3.3 mm long, the filaments 1.3-1.4 mm long, the staminal tube 0.3-0.4 mm long, the aed free portions 1—1.1 mm long, the 7-0.9 apically apiculate, basally lobate, the connective conspicu- anthers ovoid, 2.2-2.4 X 0. mm, ously punctate; pistil 3-3.2 mm long, glabrous, the ovary oblongoid, 1.2-1.3 mm long, the style 1.9-2 mm long. prominently punctate, the ovules 23 to 25. Fruits globose, 5.5-7 mm diam., prominently punctate. Distribution. Ardisia tysonii is known from the area around Cerro Jefe and Cerro Azul in Panamá. the area around Cerro Tute in Veraguas, and the 293 Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia ((OW) 2826 22N “W чолу D (TT) ZEOLT 1 12 £nuag сү *ed&jopoq шолу имвлр g ү) "игид 7)— moja ^g— "qoueiq #шцәмо]д "ү ‘sisuainaun vistpay (YBU) GQ amat ] шоу 7) (OW) Fer jouumg ‘g шол) g (ON) 6225 vosk; ^3 ‘edMojoy woy uweip y) чәмо]д 7)— Зи ^g— “youesq Зипцәморд “y— C "ros Disipay (әр) #9 әіп814 ( OW) PEE vpuray ‘f 9 ошѕәрүрд Annals of the Missouri Botanical Garden area around La Mesa near Cerro Pilón in Coclé, Panama, growing from 823 to 1219 m in elevation. Ecology and conservation status. Ardisia tysonii occurs in very wet cloud and elfin forests and is relatively common. For this reason, it is not endan- gered at the moment. Etymology. This species was named in honor of Edwin L. Tyson, the collector of the type. Within Ardisia subg. Pal- manae, Ardisia tysonii is most closely related to A. Auriculardisia sect. albisepala (see under that species for similarities). However, A. tysonii can easily be separated from A. albisepala by its thicker branchlets to 8 mm in di- ameter, wider calyx lobes to 1.6 mm in diameter, shorter and narrower corolla lobes to 3 X 1.9 mm, longer anthers to 2.4 mm long, and shorter styles to 2 mm long. Lundell (1979) failed to compare the type of Ar- disia pilonensis, which is in fruit, with that of A. tysonii. Examination of both revealed that they are identical. Specimens examined. PANAMA. Panamá: summit of Cerro Jefe, on road below tower, 20 Jan. 1984 (fl), H Churchill 4286 (LL. MO, PMA): Cerro Jefe, along trail in d of road, right hand fork, W past radio tower, fr), B. Hammel 7 (MO): Dtto. Pana- 8 June 1984 (ster.), Pipoly 7026 (MO, NY); along der from PUR Azul to Cerro Jefe, 19 Jan. 1969 (fr), E. Tyson 5309 (MO): Cerro Jefe about 10 km beyond Cerro Azul in the mountains above Tocumen ede 8 Jan. 1975 (fr), R. Wilbur & J. Luteyn 19465 (D ; Cerro Jefe, at tower, 24 Jan. 1987 (fr). I. Valdespino p | Aranda 334 (MO, NY, РМА). Ve- raguas: NW of Santa Fé, 2 km from Escuela Agrícola Alto de Piedra, ridge top below Pac of Cerro Tute, S 1975 (fr), 5. Mori & J. Kallunki 5258 (LL, Cerro Tute, ca. 10 km NW of ex A on ridge top. | May 1975 (fr), 5. Mori 6278 (LL, = 65. Ardisia unguiensis Lundell, Wrightia 6: 112. 980. Auriculardisia unguiensis (Lundell) Lun- dell, Phytologia 49: 345. 1981. TYPE: Colom- bia. Chocó: Serranía del Darién, W of Unguía near Panama border, 550—1000 25 Julv 1976 (fl), A. Gentry, Н. León & L. Forero 17032 (holotype, LL': isotypes, FTG!, MO!). Figure 65 Auriculardisia nebulosa Lundell, Wrightia 7: 270. 1984 Syn. nov. Ardisia nebulosa (Lundell) Lundell, Phyto- logia 61: 65. 1986, nom. inval. Анана nebulosa p ا‎ Pipoly & Ricketson, Sida 18: 513. 1998. Y1 anama. Panamá: Cerro Jefe, 50-900 m, 1980 (fr . K. Sytsma 1980 E LL! ypes, BM!, MO !). Auriculardisia parv iflora Lundell, Wrightia 7: 271. 1984. Syn. nov. Non Ardisia oo Talbot, Syst. List Trees Bambay (ed. 2 902. Ardisia bristanii Lundell, Phytologia ay A ‘ ee nom. inval. Ardisia bristanii Pipoly & Ric weed Sida 18: 511. 1998. TYPE: Panama. Darién: Cerro Pirré, 4 Aug. 1967 i N. Bristan 1236 (holotype, Ust: isotypes, LL!, MO!, РМА!). Small shrubs or trees 4—25 m tall, to 20 cm diam. Branchlets slender, terete, 3-5 mm diam., densely and minutely rufous furfuraceous-lepidote, often glabrescent. Leaves with blades coriaceous, elliptic, 7.8-17.1 X an acumen 4—9 mm long, basally acute, decurrent 2.9-6.6 cm, apically acuminate, with on the petiole, inconspicuously punctate and punc- tate-lineate, glabrescent above, densely furfura- ceous-lepidote below, the midrib impressed above, prominently raised below, the secondary veins 24 to 29 pairs, inconspicuous above and below, the margins entire, flat; petiole slender, canaliculate, 7—19 mm long, 1-2 mm diam., glabrous above, fur- furaceous-lepidote below. Inflorescences erect, bi- to 7-11.9 X 8.1-12.6 cm, pyramidal, longer or shorter than the leaves, the tripinnately paniculate, 6. rachis, branchlets, abaxial bract surfaces, and ped- icels densely furfuraceous-lepidote, the branches loosely congested into 3- to 7-flowered corymbs: peduncles 0.5-2.3 cm long, the lower branches subtended by leaves; inflorescence bracts absent: inflorescence branch bracts unknown; floral bracts ovate, 0.9-1.1 X 0.5-0.7 prominently punctate and caducous, membranous, mm, apically acute, punctate-lineate, glabrous above, the margins irreg- ular, minutely erose, hyaline, sparsely glandular ciliolate; pedicels slender, 5-7.8 mm long, incon- spicuously punctate and punctate-lineate, indu- ment as in the branchlets. Flowers 5-merous, lav- ender, calyx lobes chartaceous to coriaceous, 1.7-2 2.5-3.1 rounded to truncate, oblate, mm, apically broadly prominently punctate and punctate-lineate, glabrous adaxially, sparsely fur- furaceous-lepidote abaxially, the margins irregular, minutely erose, hyaline, sparsely glandular cilio- late; corolla coriaceous, 4.8-5.1 mm long, the tube 1-1.8 mm long, the lobes ovate to lanceolate, 3.1— 4.] X 1.6-2.2 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the margins entire; stamens 2.6—3.4 mm long, the filaments 1.5-1.8 mm long, the staminal tube 0.7— 0.9 mm long, the apically free portions 0.6-1.1 mm long, the anthers ovoid to lanceoloid, 1.8-3 X 1— 1.4 mm, apically apiculate, basally lobate, the con- nective conspicuously punctate; pistil 4.7—5.4 mm long, glabrous, the ovary ovoid to oblong, 1.2-1.4 mm long, the style 3.5—4 mm long, prominently punctate and punctate-lineate, the ovules 45 to 48. 6.5-8 mm diam., punctate and punctate-lineate. Fruits globose, inconspicuously Distribution. Ardisia unguiensis is distributed Volume 90, Number 2 2003 Ricketson & Pipol 295 y Revision of Ardisia subg. Auriculardisia from Cerro Azul and Cerro Jefe in Panamá to Cerro Pirré in Darién, Panama, growing at 550 to 1372 m in elevation Ecology and conservation status. Ardisia un- guiensis occurs in premontane wet forests, cloud forests, and elfin forests. It has been found in areas where only remnant forest occurs and appears to be fairly resilient, doing well in full-light situations. Because of its ability to live in full sunlight, we not believe it is threatened at this time. Etymology. Named for the town of Unguía, Chocó, Colombia, where the type was collected. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia unguiensis may be separated from both A. glandulosomarginata and A. croatii by its inconspicuous abaxial leaf punctations, flat quater- nary venation, smaller calyx lobes to 2 X 3.1 mm, dense long glandular-ciliolate along the entire mar- gin, and smaller corolla lobes to 4.1 X 2.2 mm The type of Ardisia unguiensis is in very young fruit, while the type of Auriculardisia nebulosa is in mature fruit, and the type of A. parviflora is in bud. The types of both A. nebulosa and A. parviflora are notable only for the more densely black punctations of the floral parts, with no other significant feature to separate them. Specimens examined. PANAMA. Darién: Cerro Pirré, 9—10 Aug. 1967 (fl bud), J. Duke & T. Elias E13662 (LL, MO, MOCZ); forest trail N from Ensenada del Guayabo, 18 km SE Jaqué, 13 Jan. 1983 (fr), М. Garwood et al. 246 (BM, MO); Parque Nacional del Darién, W slope of Cerro Mali, on ridge between N & S branches of Río Pucuro, ca. 18 km E of Pucuro, 22 Oct. 1987 (fr), B. Aa e et al. 16432 (CAS, F, FTG, LL, MEXU, MO, PMA); S Real, region called Alturas del Nique, near Cana mine, Pi old Camino Real toward Colombia, 22 Aug. 1987 (fr), G. McPherson 11533 (FTG, LL, MO). Panamá: 6.5 km by iad N of Lago Cerro Azul, 13 Jan. 1974 (fr), M. Nee 9287 (LL, MO, NY); Cerro Jefe, 4.7 mi. above Goofy Lake, 27 hee: 1980 (fr), K. Sytsma et al. 2828 (LL, MO). 66. Ardisia vesca Lundell, Wrightia 6: 93. 1979. Auriculardisia vesca (Lundell) Lundell, Phyto- logia 345. 1981. TYPE: Panama. Coclé: near Continental prade along lumbering road 900 m, 19 Jan. 1978 (fl, fr), (holotype, MO!, К neg. 55676!). Figure 66. ы ae minima Lundell, Wrightia 6: 1979. Icacorea n (Lundell) Lundell, Mtl 49; 350. 1981. : Panama. Coc lé: La Mesa, above El Valle de aa ca. 2 km W of Cerro 2 on Moss of steep knife-like ridge, 900—930 m, 22 July 1976 (fl bud), T. Croat 37461 (holotype, МО!, К neg. po iso- type, LL n Shrubs or trees to 4 m tall. Branchlets slender, terete, longitudinally ridged, exfoliating, 2-4.5 mm diam., densely appressed rufous furfuraceous-lepi- dote, glabrescent. Leaves with blades membranous, elliptic, 3.4—9.6 X with an acumen 5-12 mm long, basally obtuse to 0.8—3.1 cm, apically acuminate, rounded, decurrent on the petiole, prominently punctate and punctate-lineate above and below, ter- tiary venation prominulous above and below, gla- brous above, sparsely furfuraceous-lepidote below, the midrib impressed above, prominently raised be- low, the secondary veins 27 to 36 pairs, prominu- lous above and below, the margins entire, flat; pet- ioles slender, canaliculate, 5-7 mm long, glabrous above, sparsely furfuraceous-lepidote below. /nflo- rescences erect, pinnate to bipinnately paniculate, cm, pyramidal, shorter than the leaves. appressed furfuraceous-lepidote, the branches loosely congested into 5- to 7-flowered corymbs; peduncles nearly obsolete to 0.8 cm long, the lower branches subtended by leaves; inflores- cence bracts unknown; inflorescence branch bracts caducous, membranous, oblong, 2.7-3.8 X 0.8-1.3 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout, the margins irregular, minutely erose, hyaline, sparsely glan- dular ciliolate; floral bracts similar to the inflores- cence branch bracts, but 1-1.6 X 0.5-0.7 mm; pedicels slender, 6-9 mm long, inconspicuously punctate and punctate-lineate, furfuraceous-lepi- dote or often glabrous. Flowers 5- or 6-merous, deep pink to pale purple; calyx lobes membranous to chartaceous, ovate, 1.4-1.5 X 1.1-1.2 mm, api- cally acute, prominently punctate and punctate-lin- eate, sparsely furfuraceous-lepidote, the margins ir- regular, minutely erose, hyaline, sparsely glandular ciliolate; corolla membranous, 4.5—4.7 mm long, the tube 1.2-1.3 mm long, the lobes ovate, 3.3-3.4 x 1.9-2 mm, apically acute, prominently punctate and punctate-lineate, glabrous throughout except sparsely furfuraceous-lepidote abaxially, the mar- gins entire, hyaline; stamens 4—4.2 mm long, the filaments 1.9-2 mm long, the staminal tube 0.8— 0.9 mm long, the apically free portions 1-1.2 mm long, the anthers lanceoloid, 2.3-2.4 X 0.8-0.9 mm, apically apiculate, basally cordate, the con- nective conspicuously punctate; pistil 4.4—4.5 mm long, glabrous, the ovary ovoid, 1-1.1 mm long. epunctate, the style 3.4—3.5 mm long, epunctate, the ovules 9 to 11. Fruits globose, 4-5 mm diam., prominently punctate and punctate-lineate. Distribution. Ardisia vesca is known from only a few locations in Coclé, Panama, growing at 850 to 930 m in elevation. Ecology and conservation status. Ardisia vesca occurs in cloud forests along the Continental Di- 296 Annals of the Missouri Botanical Garden Volume 90, Number 2 2003 Ricketson & Pipoly 297 Revision of Ardisia subg. Auriculardisia vide. It is apparently rare and should thus be con- sidered threatened. Куто The specific epithet refers to the bladder-like shape of the flower buds. Within Ardisia subg. Auriculardisia sect. Pal- manae, Ardisia vesca is most closely related to A. dwyeri (see under that species for similarities). However, A. vesca differs from A. dwyeri by its membranous leaf blades, longer pedicels to 9 mm long, membranous or chartaceous much narrower calyx lobes to 1.2 mm wide with acute apices, shorter and narrower membranous corolla lobes to 3.4 X 2 mm, shorter and narrower anthers to 2.4 X 0.9 mm, shorter styles to 3.5 mm long, and smaller fruits to 5 mm in diameter. The type of Ardisia minima is in very young bud and is notable only for its slightly larger leaves, and although both the types of A. vesca and A. minima were described at the same time, A. vesca 1s se- lected over A. minima because it has a few flowers in anthesis. Specimen examined. PANAMA. Coclé: vicinity of d a Mesa, beyond El Valle, on N slope of Cerro Сана, July 1987 (fl), С. MePherson 11230 (LL, MO). TAXONOMIC TREATMENT OF ARDISIA SUBG. AURICULARDISIA SECT. PLEUROBOTRYAE Ardisia subg. Auriculardisia sect. Pleurobotr- yae Ricketson & Pipoly, sect. nov. TYPE here designated: Ardisia pleurobotrya Donn. Sm. uoad sepala asymmetrica ad bases. auric :ulata atque teralibus, eundi usque ad 6.2 cm lc squamis lepidotis dense manifeste superpositis brunneis- que induta atque floribus pedicellis sigmoideis insidenti- bus perfacile cognoscitur. Within Ardisia subg. Auriculardisia, section Pleurobotryae is monotypic and is defined by its strictly lateral inflorescences, long naked pedun- cles, to 6.2 cm long, dense furfuraceous-lepidote scales on most plant parts, and pendent flowers on long, usually sigmoid pedicels. 67. Ardisia pleurobotrya Donn. Sm., Bot. Gaz. 25: 148. 1898. Auriculardisia pleurobotrya (Donn. Sm.) Lundell, Phytologia 49: 345. 1981. TYPE: Costa Rica. Alajuela: Potrero del Alto, Volcán Pods, 2450 m, 20 July 1888 (fl), Н. Pittier 389 (lectotype, designated here, US!, US neg. 2382!, LL neg. 1971-85!; isolectoty- pes, В!, BR!, CR!, С! [2], W!). Figure 67. Auriculardisia pleurobotrya var. parva Lundell, Phytologia 55: 236. 1984. Syn. nov. TYPE: Panama. Bocas de Toro: Upper ni To Ayer m, Aug. 1983 (fl, fr), L. Géme n, 1. Chacón & С. Herrera 21919 debris. 111; ж-а МО!, РМА!). Shrubs or trees to 20 m tall, 35 ст diam. ВгапсА- lets stout, terete, 4—6 mm diam., densely furfura- ceous-lepidote, the scales sessile, flat. Leaves monomorphic; blades coriaceous, elliptic to oblan- ceolate, 2.4—11 .9-3.7 ст, apically acumi- nate, with an acumen 0.4—1.2 mm long, basally acute to acuminate, decurrent on the petiole, in- conspicuously pellucid punctate and punctate- lineate, glabrous above, densely furfuraceous-lepi- dote below, the midrib above, prominently raised below, the secondary veins 6 to impressed 15 pairs, obscure above and below, the margins en- tire, revolute; petioles slender, canaliculate, 0.4— 1.5 mm long, glabrous above, furfuraceous-lepidote below. Inflorescences lateral, erect, pinnately or bi- 2.4—8.2 cm, py- ramidal, longer than the leaves, the rachis incon- pinnately paniculate, 3.4—13.6 X spicuously and —punctate-lineate, furfuraceous-lepidote like the branchlets, the branches terminating in 3- to 8-flowered corymbs; punctate peduncle cm long; inflorescence bracts early caducous, unknown; inflorescence branch bracts caducous, membranous, oblong, 2.4—7.1 X 0.5-2.1 mm, apically acute to rounded, prominent- ly punctate and punctate-lineate, glabrous above, furfuraceous-lepidote below like the branchlets, the margins erose, hyaline; floral bracts similar to in- florescence branch bracts, but 2.4—3.9 X 0.5-1.5 mm; pedicels slender, terete, 6.2-10.5 mm long, usually sigmoid at anthesis, inconspicuously punc- tate and. punctate-lineate, densely appressed fur- furaceous-lepidote. Flowers 5-merous, white to light pink or yellow; calyx lobes chartaceous, suborbic- 8 х ular to ovate, 1.4— .3-1.8. mm, apically € Figure 66 Hammel 952 (MO).) Figure 67 (right). d pem E —D. Flower. —E. —H. Ardisia pleurobotrya. 1 Stamen, abaxial s p А, Ardisia vesca. —A. Flowering branch. —A. oe branch. —B. Detail » abaxial leaf surface. —C. rface. —F. Stamen, adaxial s B drawn from J. Pipoly 7124 (MO): C 1-С from A. Jiménez M. 3949 ( (F); H from R. 5a 679 (MO). —B. Flower. —C. Fruit. (A-C drawn from holotype, B. Detail of ace. —G. Stamen, lateral margin. Annals of the Missouri Botanical Garden acute, densely and conspicuously punctate and punctate-lineate, densely furfuraceous-lepidote me- dially, the margins irregular, erose-fimbriate, the fimbriae composed of simple or stellate trichomes; corolla membranous, 4.1—5.5 mm long. in tube 1.3-1.8 mm long, the lobes ovate, 3.9—4.9 х 2.8- 3.1 mm, apically acute, prominently ne and punctate-lineate, glabrous, the margins entire; sta- mens 3.8-4.1 mm long, the filaments 1.4—1.6 mm long, the staminal tube 0.7—0.9 mm long. incon- spicuously punctate and. punctate-lineate, the api- cally free portions 0.5-0.9 mm long, the anthers narrowly ovoid, 2.4-2.6 X 0.7-1.1 mm, apically subulate, basally sagittate, dehiscent by subapical pores, the connective punctate dorsally; pistil 3.3— 4.5 mm long, the ovary obturbinate, 0.9-1.1 mm long, the style 2.3-3.5 mm long, conspicuously punctate and punctate-lineate, the ovules 14 to 16. diam., prominently Fruit globose, 7.7—9.1 mm punctate and punctate-lineate, glabrous. Distribution, Ardisia pleurobotrya is found throughout the high mountains of Costa Rica and in Bocas del Toro and Chiriquí, Panama, from 1100 to 3300 m in elevation. Ecology and conservation status. Primary and disturbed or remnant, oak, montane, cloud, and elf- in forests. It appears to be quite common in pro- tected areas on mountaintops, and Pipoly has ob- served it thriving in public, open areas on Volcán Poás, so we believe it is not threatened at this time. Etymology. The specific epithet was derived from the Greek, "pleuro," meaning lateral, in a sideways position, and “botrya,” meaning bunch, which refers to the elongate lateral inflorescence branches with clustered corymbs. ` (Standley 35090). Ardisia pleurobotrya has asymmetric calyx lobes Common Name. “Tucuico’ with auriculate bases and furfuraceous-lepidote in- dument, which clearly place it in subgenus Auri- culardisia. However, the strictly lateral inflores- cences with long naked peduncles with pendent flowers on long pedicels clearly mark this species. No other taxa within subgenus Auriculardisia have strictly lateral inflorescences. Although Donnell Smith’s types are generally as- sumed to be at US, his original description of Ar- disia pleurobotrya listed four different collections without designating a type, necessitating the need to select a lectotype. We hereby designate the H. Pittier 389 collection at US as the lectotype. This collection has generally been regarded as the type and has been photographed by both LL (neg. 1971- 85) and US (neg. 2382). Lundell annotated this sheet as a type but failed to officially designate it as a lectotype. Lundell’s concept of Auriculardisia pleurobotrya var. parva came solely from his holotype, now at LL, which has eight inflorescences in very young bud and one inflorescence slightly older, with smaller leaves. However, the isotype at MO has two branches, one similar to the holotype, but with larg- er leaves more typical of Ardisia pleurobotrya. The in fruit with mature calyx and Lundell other branch is matches species in all respects. (1984b: 236) stated that variety parva “a be a diminutive form of Auriculardisia pleurobotrya (Donn. Sm.) Lundell, and all parts of the variety appears to are smaller than the species," which was true given the nature of the material he had to work with. vila RICA. Alajuela: Volcán A. Endres s.n. W near Volcán T), A din M. 3949 (F); Parque Nacional Volcán Poás, 19 July 4 (fl). J. Pipoly 7124 (DUKE, F, G, K, MO dus А Ub d of Volcán Poás from rim of rats ` lo a distance 2 mi. МА mu) toward Varablanca, 28 Jan. 1971 (fl, fr), R Wilbu Teeri 13677 (DUKE, LL). Cartago: terminus of road to Volcán Turrialba, vicinity of Finca Quemados, 2 May 1971 (fl, fr), Е Almeda 5 70 (DUKE); Volcán Irazû, pastum kde Specimens examined. CR route 40 from Cartago to the crater, 21 June 1983 (fl, fr), K. Barringer et al. 3242 (CAS, F, NY, TEX): ка below crater of Turrialba Volcano. 75 qun 1965 (fl, fr), R. Lent 679 (F, LL, MO); vicinity of La Picada and je а mer on the E slope of Volcán Turrialba, 21 Feb. 1978 fr), R. Wilbur 24671 (DUKE). Cartago-San Hb i 22 km SE of El Empalme, along the Interamerican . 1969 (fl), W. Burger & R. Liesner 6481 (F, INB, Nacional Braulio Carrillo. bus iiie 20 June 1990 (fl, fr). B. Ари 96 (CR, FTG, INB ear Porrosati on the S slope of Volcán Barva, 22 jube '1968 (fl. fr), W. Burger & R. Stolze 6082 (F, MO, NY); Potero del Alto, SW slope of Volcán de Pods, Jan. 1889 (fl), H. doe s.n. BR); Laguna de Barva, 6 Feb. 1890 (fl), А. Tonduz 1 (BR [2]. US); between Sacramento Er оч " а а! PORE of Volcán Barva, 12 Apr. » (fl), J. Utley & Utley 2035 (F, NY). Limón: с de Tiana 'a, Atlantic о НЯ d iue ra between Río v d x fr), G. Davidse et al. 29¢ CI e Nac ы. Dur SE slopes, Cerro Biricuac ua, Río Da- ás and San José Cabécar, 5 Apr. 1993 fl). G. Herrera & W. Gamboa 6248 (FTG, INB. MO) Pun- tarenas: Cantón de Coto Brus, Zona Protectora Las Ta- у —. i ~ w = = С & 5 3 = © Р =” pps ® 3 р, 2E E iz] 3 E E iz zi = - С ч, 13 Aug. 1997 (fl), Е. Alfaro 1344 (CR, INB, MO); Cor- dillera de Ur m e sopes 3 Cerro Echandi, 23 Aug. 1983 (fl), vide et al. 23883 (LL, MO, NY); Cantón de Buenos cae Ujarraz, iie of Cerro Betsd, ш the Continental Divide, Pacific slope, 6 Oct. 1989 (fl). . Herrera 3626 (CR, F, FTG, MO, USJ). San José: area of Cerro de la Muerte, Cordillera de Talamanca, 1 Feb. 1963 (fl, fr). L. Williams et al. 24150 (F, G). PANAMA. Bocas del Toro: Cordillera de Talamanca, 2—5 airline = Volume 90, Number 2 0 Ricketson & Pipol 299 y Revision of Ardisia subg. Auriculardisia km NW of the peak of Cerro Echandi on the Costa Rican— Panamanian international border, 1 & 9 Mar. 1984 (fl), G. Davidse et al. 25124 (CR, LL, MO). Chiriquí: ridge N of Potrero veg 15 Mar. 1979 (fl, fr), W. D'Arcy & B. Ham- : ЕТС, MO, PMA); Volcán Bar, along road F of Juin add 17 Nov. 1978 (fl, fr), B. Hammel 5655 (MO): Cerro Pando, valley of the upper Río Chiriquí, 13 Mar. 1938 (fr), P. White 14 (MO). TAXONOMIC TREATMENT OF ARDISIA SUBG. AURICULARDISIA SECT. WEDELIA Ardisia subg. Auriculardisia sect. Wedelia Ric- ketson & Pipoly, sect. nov. TYPE here desig- nated: Ardisia wedelii Lundell. Quoad lobulos calycinos fere liberos asymmetricos sub apicibus incises ad bases auriculatos ad Ardisiam subg. Auriculardisiam pertinet. Sectio haec ab aliis sectionibus subgeneris foliis monomorphis atque inflorescentiis longe- pedunculatis columnaribus vel subcolumnaribus perfacile cognoscitur. Few-branched subshrubs, shrubs, or small trees, to 4 m tall. Branchlets stout, terete, vestiture of usu- ally dense furfuraceous-lepidote scales or a mixture of furfuraceous-lepidote scales and cupuliform scales or a mixture of dense cupuliform scales and dense, stellate trichomes or erect, stipitate-stellate trichomes on long multicellular stalks, 0.7—0.9 mm long with long multiple uniseriate arms. Leaves monomorphic; the blades membranous to charta- ceous, inconspicuously to conspicuously and often prominently punctate and punctate-lineate: petioles slender to stout, marginate, sessile to 1.6 cm long. Inflorescences terminal, erect, pinnately to tripin- nately paniculate, columnar to sub-columnar, usu- ally shorter than the leaves, usually loosely con- inflorescence bracts usually gested corymbs: inflorescence branch bracts persistent, foliaceous: and floral bracts caducous; pedicels stout, terete. Flowers 5-merous, white, light pink, light purple. red, or light orange; calyx lobes essentially free, membranous to coriaceous, lanceolate or ovate to suborbicular, basally auriculate; corolla membra- nous to chartaceous, the lobes ovate to lanceolate, inconspicuously to conspicuously and usually prominently punctate and punctate-lineate; sta- mens connate, the filaments apically free, connate basally into an elobate tube, free from the corolla tube, epunctate, the anthers ovoid or narrowly ovoid to lanceoloid, dehiscent by subapical pores. open- ing into longitudinal slits, the connective punctate; pistil glabrous, the ovary oblong, the style slender, erect, inconspicuously or conspicuously, rarely prominently punctate, the ovules pluriseriate. Fruits globose, inconspicuously or conspicuously, often prominently punctate and punctate-lineate, smooth to costate. Distribution. Members of Ardisia subg. Auri- culardisia sect. Wedelia occur from Jinotega and Zeyala in Nicaragua to Darién, Panama, where they grow between sea level and 1400 m in elevation. Ecology and conservation status. Species in Ar- disia subg. Auriculardisia sect. Wedelia occur in primary, secondary, disturbed, or remnant forests, swamp forests, evergreen forests, premontane or montane wet forests, and cloud forests. Members of this section should be considered threatened or en- dangered. Ardisia subg. Auriculardisia sect. Wedelia is de- fined by its subshrubby to small arborescent habit, terminal, columnar to sub-columnar inflorescences on peduncles at least % the length of the inflores- cence, and subtended by large, foliaceous bracts. KEY TO THE le OF ARDISIA SUBG. AURICULARDISIA SECT. WEDELL la. Indument of branchlets, petioles, and inflores- cence branches mixed, of two types: open cu- puliform lepidote, the other dense, erect, stipi- tate-stell hairs on long multicellular trichomes, the stalks 0.7-0.9 mm long with mul- 71. Ardisia heterotricha nflores- i n Ф ec ment of branchlets, “petioles, and i cence branches = furfuraceous-lepidote 2.2-2.5 mm bout corolla lobes 2.3-2.5 mm wide; anthers ovoid, 2.6-2.8 X Л-1.3 mm; pistils 5.3-5.7 mm long; styles 4.4—4.6 mm ies ي‎ шад ————Ó— 69. Anlisiafolsomi 3b. Indument of the iip wein d in rescences of densely furfuraceous- ind. dote scales, leaves oblanceolate, 42.2 45.4 ст long; calyx lobes 1.8-2 mm long; corolla lobes 1.9-2 mm wide; an- thers oblong, 2.5-2.6 X 1-1.1 mm; d mA 3.8—4 mm long; "es 2.8-3 mm long a Ara hammelii І Aunt 1.2-2.4 X 0.6-1 mm; s filaments separate, at least above the tube throat. 4a. Leaves apically obtuse, with an acumen > cm long; calyx lobes 0.8-1 T x 0.7-1 mm; corolla d 2-2.4 mm NC URN E Ardisia kennedyae 4b. Leaves apically ac p to acuminate, with an acumen 0.6-2.9 cm long; ca eme lobes 1.6-2.5 X 1-2.4 mm; corolla lobes 3—4.4 mm long. 5a. Алл» 2. 1-24 a 0.8-1 mm. es 4 mm wide; Н mm long; stamens 3.64 mm gina ond. long 68. Ardisia pM 300 Annals of the Missouri Botanical Garden 6b. Calyx lobes 1.1-1.3 mm wide; corolla lobes lanceolate, 3.8- 4.2 mm long; stamens 2.9-3.1 mm long; anthers oblong, 2.3— 2.4 mm P dieu talamancensis 5b. Anthers 1.2-2. du X 0.50.8 mm Та. Corolla lobes 0.9-1.2 mm wide; stamens 4.2-4.9 mm long; filaments 2.7-2.9 1 long, the apically free Jd of the filaments 2.2-2.4 mm long, the tube 0.5-1.6 mm long; pistil 4.8-5.2 mm long; tyle 4—4.6 mm long St wedelii Corolla lobes 2-2. 2 mm wide; ns Tb. = ~ 3 3 = 3 = A oS Js n e Kug y А A co mm long 73. Ardisia n mame yensis 68. Ardisia conoidea Lundell, Wrightia 4: 56. 1968. Auriculardisia conoidea (Lundell) Lun- dell, Phytologia 54: 285. 1983. TYPE: Costa Rica. Cartago: Tapanti, 1300 m, 15 July 1937 (fl), M. Valerio 1624 (holotype, F!, F 68140!, LL neg. 1971-25!). Figure 68. neg. Auric м ER EY Lundell, Wrightia 7: 272. 1984. on Ardisia pde e ia Mez in Engl., Pflan- ^ nr. IV. 236 Hefi 9): 12 902. Ardisia zarceroana aundell, а 61: 68. 986, nom. inval. Ardisia ipd rales, Phytologia 83: 111. 1997. TYPE: Costa Rie "a. pts C ое Central near San Juan de “т about 15 km N d = arcero, ca. 13 ar 7 Feb. 1965 (fl, fr), L. Williams, A. Molina R., T. eere & D. Gibson 28998 е, F!, F neg. 68321!, LL! [fragment]). Subshrubs to 2 m tall. Branchlets 5— mm diam., with. densely furfuraceous-lepidote scales. Leaves with blades chartaceous, 24.4—39.6 2.1-12.9 em, apically acuminate, with an acu- oblanceolate, men 1.4—2.1 cm long, gradually tapering to an auriculate base, prominently punctate above and below, nearly glabrous above, densely furfuraceous- lepidote scales below, even more dense along the midrib, the midrib impressed above, prominently raised below, the secondary veins 39 to 53 pairs, prominulous above, prominent below, the margins entire, revolute; 0.5 long, densely furfuraceous-lepidote below. Inflores- cences bipinnately petioles subobsolete to cm or tripinnately paniculate, co- 15.2-33.6 X 5.1-13.4 rachis and densely furfuraceous-lepidote, the branches terminating in 6- to 11-flowered corymbs; peduncles (1.5-2.7)4.7-11.8 em long; inflores- cence bracts caducous, oblate, 2.5-3.2 X 5.46.9 lumnar, branches cm, mm, apically acute, prominently punctate and punctate-lineate, densely furfuraceous-lepidote, midvein and secondary veins obscure, the margins flat; the inflorescence bracts but lanceolate, 4.2-5.6 X 0.9-1.4 mm; floral bracts similar to the inflores- cence branch bracts, but 1.2-2.1 X 0.6-1.2 mm, the margins minutely erose, hyaline; pedicels 3.9— entire, inflorescence branch bracts similar to 5.6 mm long, enlarged apically in fruit, conspicu- ously punctate and punctate-lineate, densely fur- furaceous-lepidote. Flowers white or pale red to dark purple: calyx lobes chartaceous, suborbicular to ovate, 1.8-2.2 X 2.2-2.4 mm, apically acute or rounded, scattered conspicuously punctate, sparse- ly furfuraceous-lepidote abaxially, the margins mi- nutely erose, hyaline, sparsely glandular-ciliolate; corolla membranous, 3.6-3.9 mm long, the tube 1.4-1.6 prominently punctate and 0.4—0.6 mm long, the lobes ovate, 3-3.5 X mm, apically acute, punctate-lineate, glabrous throughout, the margins entire; slamens 3.6—4 mm long, the filaments 1.9— 2.] mm long, the staminal tube 0.5—0.7 the apically free portions 1.4-1.6 mm pe epunc- 3 mm long, tate, the anthers ovoid to lanceoloid, 0.8-1 mm wide, apically ice ee basal- ly deeply cordate, the connective punctate: pistil glabrous, the ovary 0.6-0.9 mm long, the style 4.4—4.5 mm long, prominently punc- 5.1-5.4 mm long, tate and punctate-lineate, the ovules 9 to 12. Fruits globose, 6.1—8.2 mm diam., prominently punctate and punctate-lineate, slightly costate. Distribution. Ardisia conoidea is endemic to Al- ajuela and Cartago, Costa Rica, growing from 700 to 1350 m in elevation. Ecology and conservation status. Ardisia conoi- dea occurs in disturbed and remnant montane, ev- ergreen, and cloud forests. Because of its restricted distribution, it should be considered threatened. Etymology. The specific epithet referred to the pedicels, which are enlarged apically, and cone- shaped in fruit. Ardisia conoidea may be distinguished from the other species of Ardisia subg. Auriculardisia sect. Wedelia by its larger anthers to 2.3 X 1 mm, larger calyx lobes to 2.2 X 2.4 mm, smaller corolla lobes to 3.5 X 1.6 mm, and larger stamens to 4 mm long with broad filaments. The only type material of Ardisia conoidea is from the holotype at F, which is in very poor con- dition. The specimen is in young bud and the in- florescence has not yet expanded, a condition that would at first glance exclude this species from sec- tion. Wedelia. shriveled. The leaf blades are also smaller and However, the type material correspond- Volume 90, Number 2 2003 Ricketson & Pipoly 301 Revision of Ardisia subg. Auriculardisia ing to Auriculardisia sessilifolia is in much better shape with a mature inflorescence and leaves and clearly shows that it belongs to this species. Specimens examined. COSTA RICA. Alajuela: along road from San Ramón northward through Balsa, ca. 13.8 km N of bridge over Quebrada Volio and ca. 4.6 km N of bridge over (apparently) Río Balsa, at small stream (Río San Luis?) 29 Aug. 1979 (fl), W. Stevens 13784 (MO) Гү road ru San Ramón northward through Balsa, ca. 7 km N of bridge over Quebrada Volio and ca. 7.5 km NS bridge over (арау) e Balsa, at bridge over я Río Cataratas, 29 Aug. 1979 (fl). W. Stevens 13863 O). Cartago: Dust E of пау Casa Blanca, wn 26 Dec. 1984 (fr), M. Grayum et al. 4649 (LL. MO). 69. Ardisia folsomii Lundell, Wrightia 6: 76. 1979. Auric ulardisia. folsomii (Lundell) | dell, Phytologia 49: : 1. TYE ma. Coclé: Atlantic Slane NW El Copé along aun- E: Pana- Río San Juan near fork with Río Tife, 5—6 hour walk on trail to El Copé sawmill, ca. 1200 ft. [366 m]. 9 June 1978 (fl), B. Hammel 3316 (holotype, LL!, LL neg. 1979-2!; isotype, MO). Branchlets 6-9 indument a mixture of densely furfuraceous- Subshrubs 0.3—3 m tall. diam., mm lepidote scales and chocolate brown cupuliform scales with irregular arms (not as pronounced as in A. brenesii). Leaves with blades chartaceous, ob- ovate to widely obovate, 24.5—42.4 X 12.5-23 cm, apically abruptly acuminate, with an acumen — 0.5-0.9(-1.2) cm long, basally auriculate, promi- nently punctate above and below, scattered furfu- raceous-lepidote below, but denser along the midrib yelow, the midrib impressed above, prominently raised below, the secondary veins 21 to 49 pairs, prominulous above, prominent below, the margins entire, revolute; petioles sessile to 0.5 cm long. densely furfuraceous-lepidote below. /nflorescences bipinnately or tripinnately paniculate, columnar, 8.6-26.4 X 3.2-15.4 ст, the rachis and branches as in branchlets, the branches terminating in 8- to 14-flowered corymbs; peduncles 5.8-12.7 cm long; inflorescence bracts caducous, ovate, 2.9-3.8 X 10 4 mm, apically acute, basally sessile or nearly so, prominently punctate and punctate-lin- eate, vestiture as in the branchlets, midvein and secondary veins obscure, the margins entire, flat; inflorescence branch bracts caducous, ovate, 0.18— 2.8 х 0.12-1.2 cm, progressively smaller, apically acute, sessile, prominently punctate and punctate- lineate, densely furfuraceous-lepidote, midvein and secondary veins obscure, the margins entire, flat; floral bracts p to the inflorescence branch bracts, but 1 x 0 nutely erose, End pedicels 5.2-6.7 ).9-1.4 mm, margins mi- mm long, inconspicuously punctate-lineate, vestiture as in the branchlets. Flowers pale red to dark purple; calyx lobes chartaceous, suborbicular to ovate, 2.2— 2.5 x 2-2.2 mm, apically acute or rounded, prom- inently punctate and punctate-lineate, sparsely fur- furaceous-lepidote abaxially, the margins minutely erose, hyaline, sparsely glandular-ciliolate; corolla mm long, the tube 1.2-1.5 2.3-2.5 prominently punctate and membranous, 5 mm long, the lobes lanceolate, 4—4.7 X apically acute, punctate-lineate, glabrous adaxially, sparsely fur- mm, furaceous-lepidote abaxially, the margins entire; stamens 4—4.3 mm long, the filaments 2.1-2.3 mm long, the staminal tube 1.7-2 mm long, the apical free portion 0.3-0.5 mm long, epunctate, the an- thers ovoid, 2.6-2.8 X 1.1-1.3 mm, apically cus- pidate-apiculate, basally sagittate, the connective punctate; pistil 5.3—5.7 mm mm long, the style 4.4—4.6 mm long, conspicuously long, the ovary 0.9-1.1 punctate and punctate-lineate, the ovules 9 to 1: Fruits globose, 5.8-6.8 mm diam., prominently and conspicuously punctate, slightly costate. Distribution. Ardisia folsomii is endemic to the area around the El Potroso sawmill at Alto Calvario near the Continental Divide, above El Copé, Coclé, Panama, growing at 700 to 1300 m in elevation. Ardisia folso- mii occurs in cloud forests. Because of its location Ecology and conservation status. near the lumber camps and sawmills in the area, it should be considered threatened. This spec of James P. Folsom, indefatigable plant collector, Etymology. cies was named in honor taxonomic specialist in Dichaea (Orchidaceae), and director of the Huntington Botanical Gardens. Within Ardisia subg. Auriculardisia sect. Wede- lia, Ardisia folsomii is most closely related to A. hammelii by its small anthers to only 2.8 X 1.3 mm, the filaments fused well above the tube throat. However, Ardisia folsomii can easily be separated by its indument of the branchlets and inflorescenc- es densely furfuraceous-lepidote, the shorter leaves obovate to widely obovate to 42.4 cm long, the lon- ger calyx lobes to 2.5 mm long, the longer corolla lobes to 2.5 mm wide, the longer anthers ovoid to 2.8 X 1.3 mm, the longer pistils to 5.5 mm long, and the longer styles to 4.6 mm long. есітепѕ examined. тА. Coclé: El Copé on Pac tific side 2 hour walk from sawmill, 16 Oct. 1979 (fr), T. Antonio 2111 (MO); Alto Calvario, ж est above sawmill, on Continental Divide, 5.2 mi. above El Copé, 1979 (fr), T. Croat 49215 (MO); Кыйы Divide, abant 10 km from La Pintada, on road from La Pintada to Co- clecito, 16 Aug. pe qn, R. Dressler s.n. (FLAS, LL); near Cascajal, ca. 10 km from La Pintada, 31 Aug. 1980 fr), R. Dealer n. (FLAS [2]): usto camp at Alto йө, 302 Annals of the Missouri Botanical Garden Volume 90, Number 2 2003 Ricketson & Pipoly 303 Revision of Ardisia subg. Auriculardisia Calvario, 7 km N of El Copé, 14 Jan. 1977 (fr), J. Folsom 1308 (MO); Alto Calvario, 20 Apr. 1977 (fl), J. Folsom & A. Jaslon 2686 (MO); Continental Divide area N of El Copé, an active lumber camp at a site known as Alto Calvario, just above El Potroso, 28 May 1980 (fl), J. Fol- som & J. Mauseth 7853 (TEX); along road from La Pin- deda to El Copé by way of Piedras Gordas, sawmill above El Copé, 20 Apr. 1978 (fl), B. Hammel 2629 (MO, F neg. 55624); El Copé sawmill, on peak right of road just before, S of, sawmill, 28 July 1978 (fl), B. Hammel 4148 (MO above El Copé, 27 Nov. 1985 (fr), G. McPherson 7691 (MO); ауе Сайо Sucio and waterfall at base of Cerro Tife, ca. 4 hr. hike, Caño Sucio is 1 hr. W of the Río Blanco, which is a 5 hr. hike N of the Continental Divide and El Po:roso sawmill, 13 Dec. 1980 (fr). K. Sytsma et al 2515 (MO); above El iae pin at ‚ N of El Copé, 13 May 1981 (fl), K. Sytsma & L. Andersson 4527 (MO). — т у 70. Ardisia hammelii Lundell, Wrightia 6: 78. 1979. Auriculardisia hammelii (Lundell) Lun- dell, Phytologia 49: 981. TYPE: Pana- ma. Colón: S approac h а Cerro Bruja from Río Escandaloso, 2600 ft. [792 m], 18 May 1978 (fl), B. Hammel 5141 (holotype, MO!, F neg. 55667; isotype, PMA!). Figure 70. Shrubs to З m tall. Branchlets ca. 9 mm diam., densely furfuraceous-lepidote, the scales overlap- ping. Leaves with blades chartaceous, oblanceolate, 42.2—45.4 X 13.9-15.4 cm, apically acute to acu- minate, with an acumen 1—1.4 cm long, basally ta- pering to a somewhat auriculate base, inconspicu- ously punctate above and below, nearly glabrous above, densely furfuraceous-lepidote below, espe- cially denser along the midrib, the midrib im- pressed above, prominently raised below, the sec- ondary veins raised above and below, the margins entire, flat; potio subobsolete to 0.5 cm long, l densely furf s-lepidote below. Inflorescences bipinnately paniculate, sub-columnar, ca. 3 х 10.8 em, the rachis, branches, abaxial bract sur- faces, and pedicels densely furfuraceous-lepidote, the branches terminating in 9- to 11-flowered cor- ymbs; peduncles ca. 12.3 cm long; inflorescence bracts caducous, ovate, 2.9-5.8 X 1.4—1.6 cm, api- cally acute, basally sessile, midvein and secondary veins obscure, inconspicuously punctate and punc- tate-lineate, densely furfuraceous-lepidote, the margins entire, flat; inflorescence branch bracts similar to the inflorescence bracts, but 0.3-1.8 X 0.1-0.9 ст; floral bracts similar to the inflores- cence branch bracts, but 0.9-1.6 X 0.3-0.6 mm, the margins minutely erose, hyaline; pedicels 4.2— mm long, inconspicuously punctate-lineate, densely furfuraceous-lepidote. Flowers pale purple; calyx lobes chartaceous, suborbicular to ovate, 1.8- 2 X 1.9-2.4 mm, apically acute or rounded, con- spicuously punctate, densely furfuraceous-lepidote, the margins minutely erose, hyaline, sparsely glan- dular-ciliolate; corolla chartaceous, 5.96.1] mm long, the tube 1.6-1.8 mm long, the lobes lanceo- late, 4.2—4.4 X 1.9-2 mm, apically acute, incon- spicuously punctate and punctate-lineate, glabrous throughout, the margins entire; stamens 3.9—4.1 mm long, the filaments 1.8-2.1 mm long, the sta- minal tube 0.8—0.9 mm long, the apically free por- tions 1-1.1 mm, the anthers oblong, 2.5-2.6 X l- l.l mm, apically apiculate, basally sagittate, the connective punctate; pistil 3.8—4 mm long, the ova- гу 0.9-1 mm long, the style 2.8-3 mm long, prom- inently punctate and punctate-lineate, the ovules 14 to 17. Fruits unknown. Distribution. Ardisia hammelii is known only from the type collection from Cerro Bruja, Colón, Panama, growing at 792 m in elevation. Ecology and conservation status. Ardisia ham- melii occurs in cloud forest. Because it is only known from the type collection, it should be con- sidered threatened. Etymology. This species was named in honor of Barry Hammel, Curator at the Missouri Botanical Garden, taxonomic authority in the Cyclanthaceae and Clusiaceae, prodigious field botanist, and co- editor of the Manual de Plantas de Costa Rica proj- ect. Within Ardisia subg. Auriculardisia sect. Wede- lia, Ardisia hammelii is most closely related to A. folsomii (see under that species for the similarities). However, Ardisia hammelii can easily be separated by its indument of the branchlets and inflorescenc- € Figure 68 (left). Ardisia conoidea. inflorescence. —D. Flower. . Stamen, abaxial surfac iM Grayum et al. 4649 Figure 69 (сим). . Flo inflorescence. ower. —E. Stamen, abaxial surf —H. Antonio 2111 ( —A. Flowering wey —B. Detail of e leaf онон —C. . Fruit. (A, B drawn fro (L 1 ый sessilifolia Lundell): н {тот М. Сгауит 4649 (МО).) Ardisia folsomii. —A. M branch. —B. St ruit. (A-C draw from E B. Hammel 3316: D-G drawn from B. Raine! 4148 (MO). MO). tail of . Stamen, adaxial su —G. Stamen, latera in margin. ); €-G from L. Williams et аг 28998 (F dn of Auricu- etail of ек leaf surface. —( Detail of e. —G. Stamen, Tes 1 margin. H drawn fom T. —К Stamen, фес al su 304 Annals of the Missouri Botanical Garden A doy Оу Акаш Оло Жула Se eg Seth Volume 90, Number 2 2003 Ricketson & Pipoly 305 Revision of Ardisia subg. Auriculardisia es of densely furfuraceous-lepidote scales, leaves oblanceolate, longer, to 45.4 cm long, the shorter calyx lobes to only 2 mm long, the thinner corolla lobes to only 2 mm wide, the smaller anthers ob- long to only 2.6 X 1.1 mm, the smaller pistils to only 4 mm long, and the shorter styles to 3 mm long. 71. Ardisia heterotricha (Lundell) Pipoly & Ricketson, Sida 18: 512. 1998. Auriculardisia heterotricha Lundell, Wrightia 7: 268. 1984. Ardisia heterotricha (Lundell) Lundell, Phyto- logia 61: 64. 1986, nom. inval. TYPE: Pana- ma. Panama: on road near slopes of Cerro Jefe, 2400 ft. [732 m], 20 Jan. 1980 (fr). T. Antonio, H. Moore & F. Putz 3417 (holotype, MO!). Fig- ure 71. Shrubs or small trees to 1.2 m tall. Branchlets 4— 5.5 mm diam., vestiture a mixture of two types: sessile cupuliform scales with varying numbers of arms, the other uniseriate, multicellular stipitate- stellate trichomes, the stalks 0.7-0.9 mm long with multiple uniseriate arms, the stellate portion at times breaking off and the remaining hair appear- ing villous. Leaves with blades membranous to chartaceous, oblong to narrowly oblong, 33.3-35 X 12-12.4 ст, apically acuminate, with an acumen 2.5-2.8 cm long, basally acute, decurrent on the petiole, prominently punctate above and below, gla- brous above, vestiture below mixed, scattered on the blade and denser on the midrib and secondary veins, like that of the branchlets, the midrib im- pressed above, prominently raised below, the sec- ondary veins 24 to 40 pairs, prominulous above, prominent below, the margins entire, revolute: pet- ioles 0.8-1.6 cm long. vestiture like that of the branchlets. /nflorescences bipinnately or tripinnate- ly paniculate, 8—8. .5 em, sub-columnar, the rachis and branches with vestiture like that of branchlets, the branches terminating in 6- to 9- flowered corymbs; peduncles 2.3-2.7 cm long: in- florescence bracts unknown; inflorescence branch bracts early caducous, oblong, 2.8—3.4 X 0.7—0.9 mm, apically acute or rounded, basally sessile or nearly so, prominently punctate and punctate-li- neate, glabrous above, vestiture below like that of branchlets, midvein and secondary veins obscure, the margins entire; floral bracts similar to the inflo- rescence branch bracts, but 0.8-1.4 X 0.7-1 mm; pedicels 3.1—5.7 mm long, 1.3-2.1 mm diam., con- spicuously punctate, nearly glabrous, but with scat- tered furfuraceous-lepidote scales. Flower color un- lobes chartaceous, ovate to 2.2-2.5 X 2.5-2.8 mm, apically acute, prominently punctate and punctate-lineate, known; calyx suborbicular, vestiture abaxially like that of the branchlets, the margins minutely erose, hyaline, sparsely glandu- ili corolla unknown; stamens unknown; pistil unknown. Fruits globose, 7.2-8.6 mm diam., prominently and conspicuously punctate, non-cos- lar-ciliolate; tate. Distribution. from the type collection from Cerro Jefe, Panamá, Ardisia heterotricha is known only Panama, growing at 732 m in elevation. The summit of Cerro Jefe is a cloud forest with an unusual open Ecology and conservation status. canopy, dominated by Colpothrinax aphonopetala R. Evans (Arecaceae) and containing numerous en- demic species. Many taxa otherwise known from much higher altitudes are present on Cerro Jefe. There were no details regarding the ecology of this species on the label, and given that a long history of collection at this locality has generated only the type collection, the species should be considered m e. Etymology. The specific epithet was derived from the Greek, “hetero,” meaning different, and 8 ` “tricho,” meaning hairs, referring to the two types of trichomes present on most plant parts. Ardisia heterotricha is unique within Ardisia subg. Auriculardisia sect. Wedelia because of its vestiture of a mixture of two types: sessile cupuli- form scales with varying numbers of arms, the other uniseriate, multicellular stipitate-stellate trichomes, the stalks 0.7—0.9 mm long with multiple uniseriate arms, the stellate portion at times breaking off and the remaining hair appearing villous. 72. Ardisia kennedyae Ricketson & Pipoly, sp. nov. TYPE: Panama. Bocas del Toro: Chan- < Figure 70 (left). inflorescence. —D. Flowe abaxial surface. (A-G drawn fron һо В. b 3141 (MO).) Figure 71 (right). i . —D. De mature fruits tail of vestitu e d pd ш Flowering branc h. —B. De S Ardisia heterotricha. —A. Flowering branch. —B. Detail of abaxial leaf surface. —C. re consisting of a mixture of two types, one of dens with multiple arms, and the other of stipitate-stellate hairs on long multicellular stalks with multiple tail of abaxial leaf surface. —C. Detail of tamen, adaxial surface. —С. Band. lateral margin. Jetail of ‚ве beri flat uniseriate arms. (A—D drawn from holotype, T. Antonio et al. 3417 (MO).) 306 Annals of the Missouri Botanical Garden guinola, swamp forest near Luzon, 20 June 1973 (fl, fr), H. Kennedy 3258 (holotype, MO!). Figure 72. Quoad folia obovata ad bases acutas ad apices acumi- edicellos graciles А. ie Pai arcte af- finis, sed ab ea lobulis calycinis 0.8-1.3 1 2.2-5 mm longis 0.7-1.0 (nec 2. 2.0-2.2) mm latis a apices acutis (nec obtusis vel rotundatis), lobulis ue 2.0—2.4 (non nalas atque p d 4.0—4.7) mm longis (nec 2.3-2.5) mm latis, de- nique antheris 1.2-1.7 (non 2.6-2.7) mm longis statim distinguitur. Subshrubs to | m tall. Branchlets 5—6 mm diam., with densely furfuraceous-lepidote scales. Leaves with blades membranous, obovate to oblanceolate, 17.3-25.2 X 7.8-11.2 ст, apically obtuse, with an acumen 0.2—0.6 cm long, basally acute, decurrent on the petiole, punctations obscure, glabrous above or sparsely furfuraceous-lepidote, densely furfura- ceous-lepidote below, denser along midrib, the midrib impressed above, prominently raised below, the secondary veins 23 to 31 pairs, prominulous above, prominently raised below, the margins en- revolute; petioles 0.3-1 cm long, glabrous е, densely furfuraceous-lepidote below. Inflo- rescences pinnately paniculate, columnar, 10.5— 2-2.5 cm, peduncle, and branches densely furfuraceous-lepidote, the branches termi- nating in 8- to 12-flowered corymbs; peduncle 7— 7.4 cm long; inflorescence bracts persistent, ob- long, 1.4-2.3 0.3-0.7 apically acute, prominently punctate and punctate-lineate, densely furfuraceous-lepidote, midvein and secondary veins obscure, the margins fat; inflorescence branch bracts persistent, oblong, 1.7-2.3 х 0.6— 0.8 mm, progressively smaller, apically acute, ses- sile, prominently punctate and punctate-lineate, glabrous above, densely furfuraceous-lepidote be- low, midvein and secondary veins obscure, the mar- gins entire; floral bracts similar to the inflorescence branch bracts, but 1—1.7 X 0.2-0.8 mm; pedicels 4.5-5.4 mm long, inconspicuously punctate and punctate-lineate, densely furfuraceous-lepidote. Flowers pale pink to purple; calyx lobes membra- nous to slightly chartaceous, ovate to suborbicular, 0.8-1.3 X 0.7-1 mm, apically acute, prominently punctate and punctate-lineate, scattered furfura- the margins minutely tire, rachis, cm, entire, ceous-lepidote abaxially, erose, hyaline, sparsely glandular-ciliolate; corolla membranous, 2.8-3.2 mm long, the tube 0.6—0.7 mm long, the lobes ovate, 2-2.4 X 1-1.4 mm, api- cally acute, prominently black punctate and punc- tate-lineate, glabrous adaxially, glabrous to sparsely furfuraceous-lepidote abaxially, the margins entire: stamens 2.2-2.7 mm long, the filaments 1.1-1.3 mm long, the staminal tube 0.6-0.7 mm long, the apically free portions 0.4—0.7 mm long, epunctate, the anthers ovoid, 1.2-1.7 X 0.7-0.8 mm, apically obtuse with a minute apiculum, basally cordate; pistil 1.7-1.9 mm long, the ovary 0.4—0.5 mm long, the style 1.2-1.4 mm long, epunctate, the ovules 7 to 11. Fruits (immature) globose, 2.5-3 mm diam., prominently punctate, slightly costate. Distribution. Ardisia kennedyae is known only from the type collection in Bocas del Toro, Panama, at or near sea level. Ecology and conservation status. nedyae occurs in swamp forests and secondary veg- еіапоп on terra firme. Because it is only known from the type, it should be considered threatened. Ardisia ken- Etymology. It is a pleasure to dedicate this species to Helen Kennedy, prodigious plant collec- tor and preeminent authority on the systematics of Neotropical Marantaceae. Ardisia kennedyae is unique within Ardisia subg. Auriculardisia sect. Wedelia because its leaves have obtuse apices to only 0.6 cm long and the smallest flowers have calyx lobes to only 1.3 X 1 mm and corolla lobes to only 2.4 mm long. 43. Ardisia mameyensis Ricketson & Pipoly, sp. TYPE: Panama. Darién: Mamey, beside ver, 8 Mar. 1982 (fl. fr). C. Whitefoord & A. Edd 430 (holotype, MO!; isotypes, BM! [2]) Figure 73. nov. pali. Propter inflorescentiam bongipeaiins 'ulatam. bracteam subtentam Wedeliam | pertinet. Species haec ab aliis ar ebus sectionis staminum fila- mentis minutis 1.1— n longis atque antheris ovoideis apiculatis ad bases a Mess perfacile separabilis. foliaceam ad sectione Shrubs to 2.4 m tall. Branchlets 3.5—5 mm diam., densely furfuraceous-lepidote, the scale margins overlapping among them. Leaves with blades mem- branous and chartaceous, elliptic, 11.5-23.6 X 5.2-7.1 em, apically long acuminate, with an acu- men 0.6-2.4 ст long, basally obtuse, decurrent on the petiole, inconspicuously punctate and punctate- lineate, glabrous above, densely furfuraceous-lepi- dote below, the midrib impressed above, promi- nently raised below, the secondary veins 25 to 31 pairs, prominulous above and below, the margins inrolled; petioles subobsolete to 0.6 cm long, gla- brous above, furfuraceous-lepidote below. /nflores- cences pinnately to bipinnately paniculate, colum- nar, 7.5-12.9 X 1.8-3.8 cm, rachis and branches densely furfuraceous-lepidote. the branches termi- nating in 5- to 9-flowered corymbs; peduncles 3.8— 8.1 em long: inflorescence bracts persistent, 1 mem- branous, ovate to lanceolate, 1.6-5.1 X 1 em, apically acute, basally auriculate, prominently Volume 90, Number 2 2003 Ricketson & Pipol 307 y Revision of Ardisia subg. Auriculardisia punctate and punctate-lineate, glabrous above, densely furfuraceous-lepidote below, the midrib im- pressed above, prominently raised below, the sec- ondary veins 25 to 3] pairs, prominulous above, obscure below, the margins entire, inrolled: inflo- rescence branch bracts caducous, membranous, ovate, 2—4.6 X 0.2-0.6 mm, apically acuminate, the veins obscure, prominently punctate and punc- tate-lineate, glabrous above, densely furfuraceous- ya- ine; floral bracts similar to the inflorescence branch bracts, but 0.8-1.8 X 0.7-1 mm: pedicels 2.5-5 mm long, inconspicuously punctate and punctate-lineate, densely furfurac eous-lepidote. Flowers pink: calyx lobes membranous, ovate, 1.7— lepidote below, the margins minutely егоѕе, 2 х 1—1.2 mm, apically acute, inconspicuously punctate and punctate-lineate, glabrous adaxially, sparsely furfuraceous-lepidote abaxially. the mar- gins entire, minutely erose, hyaline, sparsely glan- dular ciliolate: 44.2 long, the tube 0.6-0.8 mm long, the lobes narrowly є lanceolate, 3.2—3.4 X 2-22 corolla membranous, mm mm, apically acute, prominently punctate and punctate-lineate, gla- brous throughout, the margins entire, hyaline; sta- mens 2.4—2.5 mm long, the filaments 1.1-1.2 mm long, the staminal tube 0.3—0.5 mm long, the api- cally free portions 0.7-0.8 mm, epunctate, the an- thers ovoid, 1.6-1.8 X 0.7-0.8 mm, apically apic- the conspicuously punctate; pistil 3.7-4.1 mm long, the ulate, basally subcordate, connective ovary 1-1.4 mm long, the style 2.5-2.8 mm long, epunctate to inconspicuously punctate, the ovules 10 to 17. Fruits (immature) globose, 4—4.7 diam., mm prominently punctate, style base persistent. Distribution. Ardisia mameyensis is known only from the type. collected near Mamey in Darién, Panama, below 100 m in elevation. Ecology and conservation status. Ardisia ma- meyensis occurs along riverbanks in wet forests. Be- cause it is known only from the type it should be considered threatened. Etymology. The specific epithet refers to the lo- cality where it was discovered. Ardisia mameyensis is distinguished within Ar- disia subg. Auriculardisia sect. Wedelia by its an- thers to 1.8 X 0.8 mm, larger corolla lobes to 2.2 mm wide, shorter stamens to 2.5 mm long, the fil- aments to 1.2 mm long. the apically free portion of the filaments to 0.8 mm long, the tube to 0.5 mm long, the shorter pistil to 4.1 mm long, and shorter style to 2.8 mm long. 74. Ardisia talamancensis Ricketson & Pipoly, TYPE: Costa Rica. Limón: Cantón de sp. nov. Talamanca, Sukut, de = juntas de Río PER y Río Sukut, 1.5 km aguas arriba sobre margen derecha, О 09°24’ 30"N, 082°58' = 350 m, 7 July 1989 (fl), С. Herrera 3172 (ho- lotype. МО!; isotypes, CR not seen, FTG!, INB not seen). Figure 74. Propter inflorescentiam longipedunculatam bracteam foliaceam subtentam ad | genes Wedeliam | pertinet. Species haec inter alias noideae similis, sed ab ea antheris oblongoideis din anguste ovoideis) 2.3-2.4 (nec 2.1-2.3) mm longis ad apices rotundato-(nec subulato-) apiculatis statim separabilis. Subshrubs 0.5-0.7 m tall. Branchlets 7-8 mm diam., 4 the scales nearly overlapping. Leaves with blades membranous to chartaceous, oblanceolate, 28.8—30.7 X 7 densely furfuraceous-lepidote, cm, apically acute to acuminate, with an acumen 1.1-2.9 ст long, gradually tapering to an acute base, inconspicuously punctate, glabrate above, densely furfuraceous-lepidote below, denser along the midrib, the midrib impressed above, promi- nently raised below, the secondary veins 43 to 56 pairs, prominulous above, prominently raised be- low, the margins entire, flat; petioles subobsolete to 0.5 em long, densely furfuraceous-lepidote below. Inflorescences bipinnately paniculate, columnar, 16.9-20.5 X 4.8-5.6 cm, densely furfuraceous-lepidote, the branches termi- rachis and branches nating in 7- to 11-flowered corymbs; peduncle 7.5— 11.2 cm long: inflorescence bract persistent, linear oblong, 5.3-7.8 X spicuously punctate, glabrous above, densely fur- 0.6-1 cm, apically acute, incon- furaceous-lepidote below, denser along the midrib, midvein and secondary veins obscure, the margins entire, flat; inflorescence branch bracts early ca- ducous, unknown; floral bracts persistent, lanceo- late, 1.4—2.8 X 0.9-1.2 mm, apically acute, sessile, prominently punctate and punctate-lineate, gla- brous above, densely furfuraceous-lepidote below, midvein obscure, the margins entire, minutely erose, hyaline; pedicels 2.5—6.2 mm long, incon- spicuously punctate, densely furfuracous-lepidote. Flowers rose; calyx obes chartaceous, ovate, 1.6— 1.9 х 1.1-1.3 mm, apically acute, scattered prom- inently punctate and punctate-lineate, sparsely fur- furaceous-lepidote, the margins minutely егоѕе, hyaline, sparsely glandular-ciliolate; corolla mem- branous, 5.2—5.4 mm long, the tube 1—1.2 mm long, the lobes lanceolate, 3.8-4.2 X 1.3-1.5 mm. cally acute, prominently punctate and punctate-lin- api- eate, ке throughout, margins entire; stamens mm long, the filaments 1.7-1.8 mm long, the de tube 0.9-1 mm long, the apical free portions 0.8—1 mm long, the anthers ovoid to lan- ceoloid, 2.3-2.4 х 0.8-1 mm, apically rounded- 308 Annals of the Missouri Botanical Garden apiculate, basally cordate, the connective punctate; pistil 5.2-5.3 mm long, the ovary 0.7—0.8 mm long, the style 4.4—4.5 mm long, inconspicuously punc- tate, the ovules 9 to 11. Fruits unknown. Distribution. Ardisia talamancensis is known only from the type collection in Limón, Costa Rica, growing at 350 m in elevation. Ecology and conservation status. Ardisia tala- mancensis occurs in premontane rain forests. Be- cause it is known only from the type collection, it should be considered threatened. Etymology. The specific epithet refers to the region in which it was found, the Cantón de Tala- manca of Limón, Costa Rica. Ardisia talamancensis may be distinguished from the other species of Ardisia subg. Auriculardisia sect. Wedelia by its larger anthers to 2.4 X 1 mm. smaller calyx lobes to 1.9 X 1.3 mm, larger corolla lobes to 4.2 X 1.5 mm, and smaller stamens to 3.1 mm long. Ardisia wedelii Lundell, Amer. Midl. Natu- гапы! 29: 486. 1943. t wedelti (1 uundell) Lundell, Phytologia 54: 98: TYPE: Panama. Bocas del Toro: 1 чы 1940 (fl), H. von Wedel 299 (holotype, MO!, LL neg. 71-121!). Figure 75. 75. Wrightia 6: 63. 1979. Syn. nov. uricu жо boltenii (L чын l) Lundell, Phytologia 19: 342. 1981. TYPE: Panama. Veraguas: valley of Río Dos Bocas lane road bet tween Escuela Agríc ola Alto Piedra and Calovebora, 15.6 km NW of Santa é, along Santa Fé trail, steep forested hill E of river, 450—550 m, ЗІ Aug. 1974 (fr). T Croat 27652 (ho- lotype, MO!, F neg. 556711). Ardisia boltenti Lundell, D Shrubs or small trees 0.2—4 m tall. Branchlets 5.1-9.5 mm diam., densely furfuraceous-lepidote, the scales appressed and touching. Leaves with blades chartaceous, oblong or elliptic to narrowly elliptic or oblanceolate, 18.7—41.2 X 6.8-15.2 em, apically acuminate, with an acumen 1.6-2.9 ст prominently punctate long, basally auriculate, above and below, sparsely furfuraceous-lepidote below. the midrib im- above, more densely s pressed above, prominently raised below, the sec- ondary veins 29 to 48 pairs, prominulous above prominent below, the margins entire, revolute; pet- ioles 0.3-0.6 cm long, densely furfuraceous-lepi- dote. /nflorescences pinnately or bipinnately panic- ulate, 8.5-25.2 X 1.2-7.1 and branchlets densely furfuraceous-lepidote, the branches terminating in 12- to 27-flowered cor- ymbs; peduncles 3.5-9.2 cm long; inflorescence bract 0.8-11.2 х 0.3-3.3 em; inflorescence branch bracts early caducous, oblong, 3.5-13.6 X 2.5-5.3 columnar, cm, rachis mm, apically acute or rounded, prominently punc- and punctate-lineate, sparsely furfuraceous- E T lepidote above, densely so below, midvein and sec- ondary veins obscure, the margin entire; floral bracts similar to the inflorescence branch bracts, but 1.3-2.2 X 0.6-1.2 mm; pedicels 5.1-8.7 mm long, inconspicuously punctate, densely furfura- ceous-lepidote. Flowers orange-white, pink to red to purple; calyx lobes coriaceous, thick medially, 1.8-2.5 х 1.3-5 apically acute, prominently punctate and punctate-lineate, sparse- ovate, mm, y furfuraceous-lepidote, the margins minutely erose, hyaline, sparsely glandular-ciliolate; corolla membranous, 4.3-5.1 mm long, the tube 0.7—0.9 mm long, the lobes lanceolate, 3.2—4.4 X 0.9-1.2 mm, apically acute, prominently punctate and unctate-lineate, glabrous adaxially, sparsely fur- furaceous-lepidote abaxially, the margins entire; stamens 4.2—4.9 mm long, the filaments 2.7-2.9 mm long, the staminal tube 0.5-1.6 mm long, the apically free portions 2.2-2.4 mm long, epunctate, 1.5-2.1 х 0.5-0.8 basally deeply the anthers ovoid to lanceoloid, mm, apically subulate-apiculate, cordate, the connective punctate; pistil 4.8—5.2 mm long, the ovary 0.6-1 mm long, the style 4-4.6 mm long. epunctate, the ovules 10 to 12. Fruits globose, 4.8—5.7 mm diam., punctate, slightly costate. prominently and conspicuously Distribution. soamerica from Nicaragua to Panama, growing from sea level to 1400 m in elevation. Along the Caribbean coast of Me- Ecology and conservation status. Ardisia wede- Figure 72 (left). bae kennedyae. inflorescence. —D. Detail of calyx, showing asymmetric, surface. —G. Stamen, adaxial surface. 3258 (MO).) Figure 73 (right). inflorescence. —D. Flower. —E. . Fruit. (A Eddy 430 (MO )) iiia mameyensis. Flo amen, abaxial su —A. Flowe ring branch. auriculate ovate c E lobes. H. Stamen, lateral margin. —l. — —JB. Detail of abaxial leaf surface. —C. Detail of E. Flower. —F. Stamen, abaxial t. (A-I drawn from holotype, H. Kennedy —A. Flowering branch. ms Detail of abaxial oe surface. —C. Detail of irface. —F. Star . B drawn i lb C. Whitefoord & A. Eddy 430 (BM: C-H from ae C. Whitefoord & A. ‚ lateral margin. . Stamen, adaxial surface. ; Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia 309 Annals of the 310 Missouri Botanical Garden Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia lii inhabits primary and remnant evergreen forests and secondary growth in the tropical-premontane wet forest transition zone, or rarely, in cloud forests. At this time, it does not appear to be threatened. Etymology. This species was named in honor of H. von Wedel, who collected a large number of specimens from Panama. Ardisia wedelii is distinguished within Ardisia subg. Auriculardisia sect. Wedelia by its anthers to 2.] X 0.8 mm, with smaller corolla lobes to 1.2 mm wide, larger stamens to 4.9 mm long, the longer filaments to 2.9 mm long, the apically free portion of the filaments to 2.4 mm long, the tube to 1.6 mm long, the larger pistil to 5.2 mm long, and longer style to 4.6 mm long. The type of Ardisia wedelii is in poor condition, with its inflorescence very young and the peduncle although similar specimens clearly show the long peduncle. When Lundell (1979) described Ardisia boltenii he failed to com- pare it with A. wedelii. However, upon close ex- not fully expanded, amination of all additional material, the type of Ar- disia boltenii clearly matches А. wedelii in all aspects, and the species is synonymized herein. NICARAGUA. Jinotega: Ki- lambé, “Cerro San Pedro,” 25 Mar. 1981 (fr), P. Moreno 7511 (MO). Río San Juan: Mpio. El Castillo, Sábalos, comarca Las Maravillas, 15 Jan. 1995 (fl), R. Rueda et al. 2870 (HULE, MO); NW of Nueva Atlanta village. 17 Feb. 1994 (fr), R. Rueda et al. 3178 (HULE, MO); Mpio. Cas- tillo, Reserva Indio-Maíz, Cerro el Diablo, 9 Jan. 1997 (fl. ). R. Rueda et al. 5610 (HULE, MO). Zeyala: 2 km NW of Rosita, 4 Jan. 1974 (fl), J. Atwood et al. 6975 (MO); El Esc obillo, 2 km from Colonia Serrano, trail to Yolania, 29 (fl), J. Sandino 3341 Specimens examined. malai 29-31 Oct. 1977 (fr), W. Stevens 4799 (MO); Costa Riquita, ca. 1.8 km SW of Colonia Naciones Uni- das, above (S of) road between Colonia Nueva León and Colonia Naciones Unidas, 6—7 Nov. 1977 (fr), W. Stevens 5061 (MO). COSTA RICA. Alajuela: ca. 13 km NNE of Bijagua along new road to Upala, 7-8 Nov. 1975 (fr), W. Burger & К. Baker 9810 (DUKE, F [2]; Cantón de Upala, Dos Ríos, 1 km before Río Pizote along the road from Dos MO, US); N slope of Volcán Arenal, in lava flows above Río Guillermina, 24 Feb. 1989 (fr), G. Russell et al. 942 (US); ca. 9 km N of R July 1976 (fl), J. Utley & K. Utley 5349 (DUKE). nacaste: between Guachipelín and Volcán La Vieja, 26 May 1932 (ster.), A. Brenes 15551 (F); edge of road ЗО km Е of PanAm Hwy., on road to Upala, 25 Aug. 1980 (fr). J. Kress et al. 9593 (DUKE). Heredia: Finca La Selva, the OTS Field Station on the Puerto Viejo just E of its junction E: the Río Sarapiquí, far side of research trail small gap at edge of trail, 5 July 1981 (fl), B. Hammel 10950 (DUKE). Limón: Cantón de Limón, Cor- ге de Talamanca, along divide between Río Xikiari Río Boyei above Cabécar village of Almirante (not on ipe maps), 13 Aug. 1995 (fr), M. Grayum 10928 (FTG, INB, ji pg: bap gi Estación Agua Fría, 8 km SE of pa ong Los Raudales Trail, 27 Oct. 1987 (fr), К. Robles 1160 (CR, MO). San José: W part of Mon- C (CR, MO); El е San Isidro Jan. 1891 (fr), H. Pise 3504 (BR). Р eren peur road from For- tuna Dam to Chiriqui (randa, from Continental Divide, 2 Aug. 1984 (fl), H. md 5929 (MO); Fish Creek lowlands, aha of Chiriquí Lagoon, 7 May 1941 (fl. Н. von Wedel 2393 (MO). Chiriquí: Fortuna Dam area, unnamed creek to E of road flowing into Río Hornito near Quebrada Moro, 16 June 1984 (fl), H. Churchill 5497 (LL, MO). Veraguas: along Río Dos Bocas, ca. 1: beyond Santa Fé, 25 July 1974 (fl), T. Croat 25791 (MO); Cerro Tute, W of Santa Fé, beyond Alto de Piedra, 17 Oct. 1985 (fr), G. McPherson 7158 (MO); 8.8 km from Escue ‘la Agricola Alto de Piedra, 16 Nov. 1974 (fr), Mori & J. Kallunki 3211 (MO, F neg. 55683); 14 a NW of Santa Fé, on road to Calovebora, Panama Hwy. : Aug. 1975 (fl), 5. Mori & A. Bolton 7650 (MO, b ied 55682). EXCLUDED NAMES ш сесси Lundell, Wrightia 3: 25. 1962. Amat- esi ge a andell) Lundell, Wrightia 7: 40. Я нараз: Pinabeto, near Mo- tozintla, 2585 m, .9 agi 1945 (fr), E. Matuda 5462 (holotype, LL!, F neg. 55688; isotypes, LL! [2]). This species is a synonym of Gentlea micranthera (Donn. Sm.) Lundell. In fruit, members of Ardisia subg. Auriculardisia sect. Amatlania can be diffi- cult to distinguish from species of Gentlea because of the dense to scattered papillae on the inflores- cence branches. However, the calyx lobes in Gent- lea are symmetric and lack auriculate bases. Also the inflorescences in Gentlea are usually smaller ج Figure 74 (left). inflorescence. —D. F . Stamen, abaxial surfac (A-G drawn from holoty pl pe. 6. Herrera 3172 (MO).) Figure 75 (right). inflorescence. —D. —AH. Fruit. (A, B fien fron Baker 9810 (F); H from G. Hews 1093 (MO).) н a wedelii. Flower. —E. Stamen, abaxial surface. sia j talamancensis. —A. prias branch. —B. Detail of abaxial leaf surfac low —F. Stamen, adax 1 T. Croat 27652 (MO holotype of Ardisia boltenii Lundell); C-G ба 7 Burger & R. e. —C. Detail of —F. Stamen, adaxial surface. —G. Stamen, lateral margin. —A. Flowering branch. —B. Detail of abaxial leaf е чанар —C. Detail of ial surface. —G. Stamen, lateral margin. Annals of the Missouri Botanical Garden and shorter than the leaves. Careful examination of the type material showed that its inflorescence is small and shorter than the leaves, the calyx lobes are symmetric and without auriculate bases, and the leaf margins are slightly crenate. These are characters typical of Gentlea micranthera and leave no doubt about the true identity of this taxon. Geissanthus carchianus (Lundell) Ricketson & Pipoly, comb. nov. Basionym: par carchia- na Lundell, Wrightia 7: 23. disia carchiana (Lunde Il) Жеп Phytologia 49: 342. 1981. 1 Blancas, 20 km below Maldonado on the Río San Juan, 900—1000 m, 27 May 1978 (fr), M. Madison, T. Plowman, H. Kennedy & L. Besse 4625 (holotype, F!; isotype, SEL!). . Auricular- ^E: Ecuador. Carchi: Peñas This taxon is actually a species of Geissanthus, and the new combination is made here. Lundell (1981b) transferred this taxon to his segregate ge- nus Auriculardisia based on its supposed affinity to Ardisia megistophylla Lundell (large leaves?), a member of Ardisia subg. Auriculardisia sect. Pal- manae, and noted that "the calyx is atypical." The type has a few fruits, the calyx clearly shows char- acteristics consistent with Geissanthus, and the ca- lyx tube splits below the basal calyx lobe and forms a callus with uneven sinuses between the calyx lobes. The calyx lobes clearly are not auriculate basally, thus eliminating the possibility that this taxon belongs to Ardisia subg. Auriculardisia. In- stead, Geissanthus carchianus has affinities to G. longistamineus (A. C. Sm.) Pipoly. Geissanthus zakii (Pipoly) Ricketson & Pipoly, comb. nov. Basionym: Ardisia zakii ро, Sida 17: 449. 1996. TYPE: Ecuador. cha: Carretera Quito-San Juan d palme, Km 59, 16 km NW of road, 1700—2000 m, 23 Sep. 1986 (fl, fr), V. Zak 1298 (holotype, МО!; isotypes, FTG!, ОСМЕ not seen). ichin- This taxon also turns out to be a species of Geis- santhus, and the new combination is made here. The type is in fruit but none is attached to the rachis. The calyx clearly has characteristics con- sistent with Geissanthus in that the calyx tube splits below the basal calyx lobe and forms a callus with uneven sinuses between the calyx lobes, and the calyx lobes clearly are not auriculate basally. This definitively eliminates the possibility that this taxon belongs to Ardisia subg. Auriculardisia. Literature Cited Aublet, 1775. Histoire des Plantes de la Guiane Fran- çoise, j s Pierre-Francoise Didot., Paris. Deutse y i allé, F., мана o 97-114. ds n Wrightia 4(2): 53-7 ——. 19 logia + к 13 1(2): 38-50. ol: 624 Bell, 1991. An Illustrated ae to one Plant Mort Oxford Univ. Press, New Candolle, A. de. 1834. A review of es E order Муг» ipo Trans. Lir . London 17: 95-1: ; Deine mémoire sur les ыз ж Ann. Sei. Nat. Bot., ser. 2, 16: 65-97. Chen, C. & J. J. Pipoly. 1996. Myrsinaceae. Pp. 1—38 in Wu е ng-yi & P. Н. Raven (editors), Flora of China, s э. Science Press, Bel ing, and Missouri Botanical n Press, St. Lo bus ‘ke, А. 1930. и pM ou peu connues б région amazonienne. Archiv. Jard. Bot. Rio 5: 188. Fisher, J. & D. W. Stevenson. 1981. Occurrence M reac- tion wood in branches of dicotyledons and its role in tree architecture. Bot. Gaz. 142: 82-95. Gessner, F. & G. Volz. 1951. Die Kuticula der Hydropoten von Nymphaea. Planta 39: 171-174. згешег, W., J. McNeill, F. R. Barrie, a У I cC H. M. B urdet, V sod editore :). 2000. Tenan Code of о al Nomenclature (Saint Louis Code). Regnum Veg. 13 Griiss, J. 1927a. Die Luftblatter dg Nymphaeaceen. B Deutsc : Bot. Ges. 45: 45445 927b. Die Haustoren ie Nymphaeaceen. Ber. Ges. 45: 459—466 A. Oldeman & P. B. йан 1978. Trop- Forests. An Architectural Analysis. Springer- E ча New York. Hickey, L. 1984. A revised classification of the architec- ture of i donous leaves. Pp. 25—39 in С. R. Met- calfe & L. Ch: alk (editors), on of the n Vol. 1. Systematic Anatomy of Leaf and Stem, with a Brief History of the Subject. Clarendon Press, Oxford. Lindley, J. 1848. Ilustrated Dictionary of Botanical Terms. Excerpt from illustrated dictionary of botanical terms by John Lindley. Pp. 346—383. [Reprint, with an introduction by Alice Eastwood, Stanford University, School A Farth Sciences, 1964. Lundell, 963. Studies a the American Myrsina- ceae Г ‘Wrightia 3(5): 7 4. Studies of ha md 'rican Myrsinaceae mel we A New genera and spec ies of American 3-73. 79. Мейд Мел Мугзїпасеае—П. Wrightia 6(4): 60-100. E m of American Plants—XX. Phyto- m. po ‘al Myrsinaceae—1V. Phytologia 48: Du gp Neotropical Myrsinaceae— VI. Phytologia 49: 341-35 54. 19814. 7(1): 23-25. 1982. Neotropical Myrsinaceae V. Wrightia Neotropical Myrsinaceae VII. Wrightia . 1984a. Neotropical Myrsinaceae—XIV. Wrightia 7(A): 266-275 D. . 1984b. Neotropical Myrsinaceae—XI. Phytologia 55: 235-242. dite Neotropical Myrsinace: . Phytologia (d к of American plants—XXII. Phyto- logia 63: 73-7 üttge, U. 1964. Майны ‘he Untersuchun- Volume 90, Number 2 2003 Ricketson & Pipoly Revision of Ardisia subg. Auriculardisia gen über die мота der Hydropoten von Nymphaea. Protoplas sma 59: 157-162. ›. Krapf. 1972. Die Ultrastruktur der Nymphaea-Hydropoten in Zusammenhang mit ihrer Funktion als salztransportierende nates Cytobiologie See Mayr, E. 1915. Hydropoten an — Sumpf-pflan- zen. Beih. po Centralbl. 32: 278— Metcalfe, Some SEE ia EI and tis- sues. Pp. 5462 in C. R. Metcalfe & L. Chalk pede Anatomy of the ре Pad Vol. 1. Systematic Ana omy of Leaf and Stem, vith a Brief History of the Sub. ject. Clarendon Press, Oxford. Mez, C. 1902. Мутвіпас фе . In: А. Engler (editor), Pflanzenreic i IV. 23€ 5 (Heft 9) 9): 1-437. „р з, J. F. 1997 098 Una nueva especie y seis nue- mbinaciones en las Myrsinaceae de Costa Rica y Ex. Руна B3: 109-112. Nayar, M. I 1986. Parardisia—A new genus in the ж n ое э with two species. Bull. Bot. Surv. pe 28 Oersted, А. Das 47-25 802. Myrfineae Centroamericanae et Me- D xicanae. d nsk. мде |. Dansk Naturhist. Føren ко > 17-142. Panfet, а Familia Myrsinaceae en Cuba. Unpublished pee Thesis, Universidad de la Habana, zuba. Pipoly HI, J. J. 1983. Con Бам toward а monograph of Cybianthus (Myrsinac . Notes е subgenera Stapfia e Mieroconomerpha Wrightia 7 : 235-244. A natic revision O genus‏ پچ Б subgen 1us Grammadenia (Myrsinaceae). York Bot. Gard. 43: 1—76. género Ardisia Swartz (Myrsinaceae >) en Colombia. Caldasia 16(78): 211-284. ————. 1991Ь. Ardisia lundelliana, a new species of T isinacest from Panama. Ann. Missouri Bot. Gard. 78: 524—520. 1992. The genus Cybianthus subgenus Cono- morpha (Myrsinaceae) in Guayana. Ann. Missouri Bot. Gard. 79: 908-957. ‚ of Ardisia (Myrsi )f the Cordillera бие кй di of Pipes aid Eouados 4: 38-44. . 1995. Dos nuevas « specie s del genero Ardisia Sw. (Myrsinaceae) de la pariso floristica Chocoana de T ae ty 17(82-8 24. ——À The genus Сынай (Myrsinaceae) in as ы Peru. Sida 18: 1—160. & J. M. Ric to 1998a. A revision of the genus Ardisia subgenus Crapl hrdisia (Myrsinaceae). Sida 18: 433472. . 1998b. New names and combinations in Neotropical Myrsinae ae. Sida 18: 503-517 — & 9a. Discovery of the Indio-Male- sian genus Hsc i (Myrsinaceae) in the Neotrop- ics, and its boreotropical implications. Sida 18: 701— 746. ———— & ————. 1999b. Additions to the genus . Ardina subgenus C A (Myrsinaceae). Sida 18: 114: 1160. 1 . 2000. Discovery of Ardisia subgenus Acrardisia Minns eae) in re Another bo- reotropical "veg Sida 19: 275-183. Stafleu, F. A. & R. S. Cowan. 1985. Taxonomic Literature, 2nd Ed. Vol. V: Sal-Ste. Regnum Veg. Standley, P. C. 1924. Myrsinaceae. /n: Aen and Shrubs of Mexico dace т rophulariaceae). Contr. US. os nm 23(4): 849-1312. . New plants ч. Central America—X. J. 3 end Sci. 17: 520-528. Theobald. W. L., J. Krahulik & R. C. ш 1984. Tri- chome деси йн and classificatio )-53 in С. R. Metcalfe & L. Chalk (editors), M of the Di- cotyledons, Vol. 1. Systematic Anatomy of Leaf and Stem, with a Brief History of the Subject. Clarendon Press, Oxford. APPENDIX 1 LIST OF SPECIES AND SUBSPECIES cee subg. Auriculardisia sect. Amat . liebmannii Gait subsp. e q me Ric- gine son & Pi 2. A. liebmannii subsp. liebmannii З. А. pellucida Oerst. subsp. lancetillensis Ricketson & ly J — Pipoly 4. A. pellucida subsp. pectinata (Donn. Sm.) Ricketson г Pipoly А A. pellucida subsp. pellucida A. pellucida subsp. thomascroatii Ricketson & Pipoly A. schippii Stanc il Ardisia subg. Auriculardisia sect. Auriculardisia r ecc > 0 E Я US а. Җа E: $ ot ® 11. Ardisia ursina Lundell Ardisia subg. Au uric Wee e sect. uc 12. Ardisia apoda Standl. & S 13. Ardisia inlet еа & Pipoly 14. Ardisia n an 15. Ardisia gordo qv & Pipoly 16. Ardisia nevermannii Standl. 17. Ardisia tilaranensis Stan 18. Ardisia tortuguerensis Ric sketson & Pipoly Ardisia subg. Auriculardisia sect. Palmanae 19. Ardisia aguirreana Pipoly 20. piera albisepala (L adeli) Pipoly & Ricketson 2 —_ аю Ricketson & Pipoly 22. Ardisia angucianensis Ricketson & Pipoly 23. Ardisia кл Lundell 24. Ardisia auriculata Donn. Sm. 25. Ardisia awarum Ricketson & Pipoly ell 29. Ardisia cogolloi Pipoly E Ardisia coloradoana Lundell Ardisia conglomerata Lundell S Ardisia crassipedicellata Lundell 33. Ardisia crassipes Lundel 34. Ardisia crassiramea Standl 35. Ardisia croatii Lundell subsp. correae (Lundell) Ric- ketson & Pi ipoly 36. Ardisia croatii subsp. croatii 31. Ardisia darienensis Lundell 38. Ardisia dukei Lundell 39. Ardisia dunlapiana Р. Н. Allen 10. Ardisia dwyeri Lunde 41. Ardisia eucuneata (Lundell) I 42. Ardisia fimbrillifera Lundel 43. Ardisia furfuracea Standl. "poly & Ricketson Annals of the Missouri Botanical Garden 44. Ardisia generalensis Ricketson & Pipoly 45. Ardisia gigantea Ricketson & Pipoly 46. Ardisia glandulosomarginata Oerst. 47. Ardisia hagenii Lundel 48. Ardisia hugonensis (Lundell) Pipoly & Ricketson 49. Ardisia knappii (Lundell) Pipoly & Ricketson 50. Ardisia liesneri Lundel 5l. Ardisia lunde lliana Pipoly . Ardisia pseudoracemiflora Pipoly Ardisia RUM enta Mez 59. ны edae Ricketson & Pipo 60. Ardisia анне Ricketson & Pipoly . Ardisia tarariae Lunde 65. Ardisia unguiensis Lundell 66. Ardisia vesca Lundell Ardisia subg. Auriculardisia sect. Pleurobotryae 67. Ardisia pleurobotrya Donn. Sm. agree subg. Auriculardisia sect. Wedelia Ardisia conoidea Lunde Pipoly & Ricketson 72 Ardisia kennedyae Ricketson & Pipoly Ardisia mameyensis Ricketson & Pipoly э Ardisia talamancensis Ricketson & Pipoly 75. Ardisia wedelii Lundell APPENDIX 2 INDEX TO EXSICCATAE The figures in es ntheses refer to the numbers from the list of species and subspecies. C nig ‘tion numbers in boldface type ndis type specime Acevedo, D. 78 (55); anco P. 6875 (16); Aguilar, R. 220 (50), 354 (8), 469 (8), 565 (39). 4699 (8). 5063 (46); Aguilar, R. & O. Garrote 3834 (46); Aguilar, R. & B. Hammel 26 (24); Aguilar, R. & F. Quesada 3265 (39); Aguilar, R. & H. Sc _ 940 (42); Aguilar, R. & M. yip 2004 us Alcázar, E., et al. 137 (42); Alfaro, ‚ 1344 (67); Alfaro, E. & M. Segura — 45 (54); meine F. 370 (67); (55); n o, L., 199 \ ‚ 872 (54), 1849 (36), 1999 (36), 2111 (69), ), 2707 (55), 3001 (24), 3079 (47), 3354 (24), 3542 (24), 3985 (0), 4025 (36), 4640 (5), Er 4); Antonio, T., et al. 34 71) Amonio V., M. B. 96 ; Araya, T 667 (42), 7 3i (4: 2); Atwood, J. : Avendafio R., S. 556 (1); Avendaño R., S. & J. Calzada 403 (1). Barbosa, C. 6591 (16); Barringer, K., et al. 3242 (67); Barringer, K. ;. Christenson 3385 A (46); ge C., E. 1810 (17), 3083 (11), 3097 (55); Bello C., E. & E. Cruz 4298 (54); Bernardi, L. 10602 (55), 10615 (63): Blum, K. 1378 (54); Blum, K. & E. Tyson 2316 (5); Bóhlke, M. 106 (2); -3 B. 744 (61). (54); Botteri, M. s.n. 146 (2), 481 ( & A. Boyle 940 C. 8: 30 (67. 842 (46), 1378 (42); Boyle, 1 л Pyrs — hg = © * — . 1107 (55), 2808 (42). : Я .. Н. 268 (5); Breedlove, D. 6298 (2), 11081 (2), (4). 26511 (5), 27954 (2), 52485 (2); Breedlove, 4 Almeda 48412 (2), 57278 (2), 57299 (5), 57993 (5); Breedlove, D. & B. Keller 49317 (4); Breedlove, D. & P. Raven 13562 (5); Brenes, A. 3815 [121] (55), 3828 [134] (55), 4033 [451] (84), 4495 (46), 5652 [244] (34), 6346 . 6358 (46), 13459 (24), 15551 (75), 20340 (46), 20536 (24), 20537 (14); Bristan, М. 468 (63), 593 (37), 1236 (65), 1397 (5); Brown, C. 163 (67), 17416A (17); m G. & L. n ht 791 (14); Burger, W. 4349 (67); Burger, W. & T. Antonio 11183 (14), 11192 ( (24); Burger, W. & R. Baker 9806 (24), 9810 (75 №. & J. о G28 a 5). 5608 59) 8704 (34), T Li iesner т 6202 (55), 6481 . 4234 (24), 4271 (24); & R. Stolze 6082 (67); Burger, W., 10288 (43), 11423 (46), 11676 (54), 12058 (46). Calzada, J. : (46), 3 355 (54); | ме A., el k ;., et al. 1764 (1); Cedillo T., R. 424 (13); Cedillo T., R. & G. Higareda 2890 Cerón, C. 7699 (5); Chacón, Hi (24); Chacón, I. & С. i >ra 1616 (24); Chávez, 55); Chickering, А. 200 (3); Churchill, C. 6 ү ud hill, H. 4286 (64), pu (75), 5 929 (75): € vs & A. Churchill 6131 (54); Churchill, i 54); Clark, J. 2141 (35); Clark, J. & D. 35); С faye J.. et al. 236 (35), 1677 (35), 2375 73 Coch- rane, T., et al. 6277 (56); Cogollo, A., et al. 3498 (29 7312 (29); Contreras, E. 2675 (7), 6348 (5 ), 6619 (54), 6920 (5), 6921 (5), 8846 (5), 8964 (7), 10154 (7), 10164 (7), 10217 (7), 11491 (7); Meu C. 169 (2); Correa A., M. 598 (53), song es wi rrea & R. Dressler 462 (40), 828 (40); Correa A., ү зе gro. 10638 (10); a et al. 2 (75), 2690 (35), 3180 (47), З (47); Cowell, J. 230 (5); Croat, T. 9257 (5), 9420 (5), 9859 (53), 9868A (53), 10071A (53), 10400 (5), 10590 5). 13682 (46), 13730 (46), 13839 (53). (5), 14529A (5), 23587 (7), 25595 (36), 25791 (75). 25907 (36), 26189 (24), 26678 (27), 27322 (5), 27327 (36), 27652 (75), 34825 (46), 34974 (46), 35270 (6). 36734 (54), 36841 (28). 37235 (30), 37431 (64), 37461 (66), 37800 (63), 40165 (5). 41325 (4), 42632 (3), 43275 (5), 43382 (46), 44258 (54), 46936 (24), 46965 (24), 48725 (75), 49215 (69), 55713 (35), 66549 (75), 66938 (36), 68190 Qu. 19375 (24); Croat, T. & J dur il. (24); Croat, T. & M. Grayum = 60032 (47); Croat, . Hannon 64637 (3), 65343 (5) Croat, T. & D. B ions (55 5), 16106 (46); Croat, T. & G. Zhu 76457 (75), 77198 (5); Cuadros, H., et al. 3939 c 63); Cufodonti, G. i51 (o) 455 (67). : 539 (46), 685 (16), 712 (67); ае M. s.n. (2). D'Arcy, W. 3906 (5). "10671 (46), 10799 (46), 11047 (46), 12166 (5: 5), 16283 (80); D'Arcy, W. & B. Hammel 12513 (67); у, W >. McPherson 16195 (63); УАгсу, W., et al. 12568 (46), 12572 (46), 12890 46), 12893 (46), 13152 (46), 15880 (54); Davidse, G. 2457 27), 36365 (54), iin To Davidse, G. & W. Duo 10184. (46); Des . & G. Herrera Ch. 29125 (46), = ~ Volume 90, Number 2 2003 Ricketson & Pipoly 315 Revision of Ardisia subg. Auriculardisia 2939] (46), 31185 (14); Davidse, G. & D. Holland 36675 (54); Davidse, С. & M. Meadows 35751 (54); Davidse, G., et al. 23828 (46), 23883 (67), 25124 (67), 25252 (55), 25486 (55), 25645 (54), 25708 (46), 25719 (46). 25764 (46), 28363 (27), 28656 (61), 28703 (61), 28882 (61), 29021 (67), 29065 (61); Davidson, C., et al. 6835 (42), 7009 (14); Davidson, M. 987 (46); Dodge, C., & (8); Dodson, C., et ye 8677 (35); Don- ; Dorantes, J. 2 s.n. (69), 4403 (5), 5918 (69); Dryer, 7 577 (3 (55), 1078 (34); Duke, J. 5335 (63), 5404 (5), 6561 (37), 13150 (10), 14154 (5), i M. Correa A. 14692 (9); (9), 15132 (10); төшү : i T. ‘Elias E13662 (65), (37), 13762 (63); T Duke, J. & J. ges a 11045 (5); Lallathin 14968 (10), 14989 (9), 15004 m 1433 (5), po „© ү le Correa А. 8 (9); Dwyer, J. s. 3. Lallathin 8677 (10); Dwyer et al. 8196 (64), 8202 (40), 8231 (40). 8247 (40). 8999 (40). Endres, A. s.n. (67), s.n. (67); Englesing, F. 79 (54) T oy al. 3639 (16), 3751 (19), 3827 (16); Espi- 56 (34); Espinoza, R. & R. Aguilar 1173 (24); Etiónne, С ѕ.п. (67). Fernández, n 10 (46), 251 (53), 267 (67), 269 (46), 836 (55); Fernández, A. & E. García 1511 (67), 1492 (67): Fernández, A., et al. 331 (27), 1078 (67); Folsom, J. 1308 (69), 1329 (32), 1593 (54), 4219 (63), 4251 (63), 5561 (62), 5898 (53), 8720 (54), 8943 (54), 9025 (42). 9115 (24), 9219 (42), 9403 (42), 9681 (42), 9782 (9) icd . & L. Collins е 1718 (5); Folsom, Ј. & 1371 (64); Folsom, J. & A. Jaslon 2686 (69); Fes J. & R. Lantz 1893 (32); sd J. & J. Mauseth 7853 (69); , J., et al. 1845 (64), 4710 (30), 4998 (23), 5476 (54). 6297 (63), 6341 (37), 6597 (5), 6617 (23), 6833 (5). 7842 (5); Forero, E. & R. Jaramillo 2812 (48); Forero, ., et al. 4 (42); Foster, R. 870 (5), 1458 (5). 2097 (10), 14148 oy Foster, R. & J. Terborgh 5280 (5); Fos- Й 3 (5); Franlie, p^ 316A (54); Friedrichst- hal, E. 6 bar pm (54); Fuchs, H. & L. Zanella 21851 (19); Fuentes, Z. 701 (34); P V., et al. 10923 (24), E. (17). Gallardo, G. 256 j foe R., B. 247 (27), 1902 (27); Gamboa R., B. & A. Picado 606 (54), 1052 (46); Gamboa R., B. & A. Riek 538 (54); García, D. 322 (54); García C., F. & E. Agualimpia 360 (19); Garefa-Barriga, H. 13766A (8); ра N. 2592 e Garwoo & А. Dominguez 1585 (5); Garwood, N., et al. 246 (6 5); Gay, H. 1515 (67); Gentle, P. 3205 (54), 3294 (54), 3982 (54). 5825 (5), 6406 (54), 7187 (54), 9665 (54); Gentry, A 2146 (40), 2150 (42), 3003 (24), 3649 (36), 5943 (46), 5979 (46), 6000 (55), 6006 (46), 8766 (24), 17032 (65), 43583 (5); Gentry, A. & A. Clewell den (63); Gentry, A. & C. Diaz 58457 (5); Gentry, Fallen 17348 84 17418 (53); Gentry, A. & 5. Mori psy y; Gentry, ; M. Nee 8691 (54); Gentry, A. & E. Tyson 1659 (5); os фа et al. 3446 (40), 3524 tio 27248 (5), 28551 (5). 76104. (52); Godfrey, В. 66352 (24), 66363 (54); Gómez P, L. 19169 (54), 19282 (43), 19496 (8), 19727 (46). 20078 (46); Gómez, L., et al. 20982 (24), 21159 (16), 21511 (46), 21919 (67), 22293 (55), 23564 (26), 23646 (67); Gómez-Laurito, J. 9259 (54), 9686 (46), 9706 (24), 9788 (55), 10569 (14), 11087 (55). = (46), 11410 (14), 11842 (24); Gómez-Pompa, A. 1133 (2); González, J. 664 (46), 749 (67); Gorden, B. 25D (75); Grant, J. & A T ү J. Rundell 2170 (46); Grant, V. 820 (54), 869 (54); Gra- уп m, M. 1353 (5), 1810 (5), 2337 (24), 3941 (55), 6774 42), 7060 (43), 7213 (46), 7290 (67), 7516 (67), 9761 (42), 10928 (75), 10984 (54), 11046 (28); Grayum, M. & B. Hammel 5754 (5); Grayum, M. & С. Herrera 4832 (5). 7905 (24); Grayum, M. & B. тти 37 87 (54); Grayum, M. & P. Sleeper 3847 (34); Grayum, M. & R. Warner 5434 26); Grayum, M., et al. 4648 (26), 4649 (68), 5718 (5). 5973 (27), 7602 (5), 7925 (16), 8993 (14), 5866 (75); Greenman, J. & M. Greenman 5601 (67); Grijalva, A. & D. Ríos 3456 (54); Grijalva, A., et al. 3380A (54), 3675 (54). Haber, W. 712 (55), 716 (55), 1347 (55); Haber, W. & E. Bello C. 1671 (54), 2428 (34), 6542 (55). 6552 Se 6710 (17), 6848 (24), 7655 je Haber, W. & К. 5651 (24), 5704 (24), 6997 (14); Haber, W. & W. 2 9128 (34), 9358 (17), 0033 (34): Haber, W., al. 1835 (5); Hagen, C. von & W. von Hagen 2005 (46), es (47); Hahn, W. 144 (5); nain. C. & W. D'Arcy 1339 (5); Hamilton, C. & R. Dressler 3017 a 3109 (20): как C. & Н. Stockwell 1462 (5 n Hamilton, et al. 4105 (10); vent 2029 (6 ^ p (46), 3141 (70), 331 2 (36), 4148 (69), 4599 (36), к, (64), (24). 5655 (67), 5736 (47), 5839 (36), 6119 (54). 7363 (46), 7 А 8741 (5), 8994 (5), 9314 (54), 9358 (5), 9766 (5), 9775 (24). 9944 (54), 10520 (42), 10659 (54), 10958 (5), 10959 (75), 11230 (66), 11247 (24), 1 (5). 19280 (22), 20661 (14); Hammel, B. & R. Aguilar 18572 (50) Hammel, M. Ne же 18251 (8); | К. Robles 16714 (42), 16737 (50); Ham- as — л N Nw м NY (5 3), 4937 (24), 6477 (46 А (14), 16237 (5), 16432 d 16899 (50). 16974 a^ р 38 (50), 19091 (24), 19258 (54); че А shi > Wiiteloote 659 (75); ы. Р. 242 (39) Hart, ч ded Hartman, R. 12337 (37), 12376 (63). 12473 (63); dos С. 973 (24), 975 (54), 1048 (42), 1106 (42), 1391 (54), 1399 (55), 1568 (24), 1780 pen 1882 (50); Hatheway, W. 1260 (46); pum T. (54); Hayes, S. s.n. (5); Hernández G., Н. 211 (5); 2 nández M., R. 1227 (5), 1565 к, 3338 а Hernández M. R. & I. Calzada 1536 (1); Herrera Ch., G. 146 (54), 965 (24), 1093 (75), 3172 (74), 3626 (67), 4165 (54), id (54), ey: (8), 4977 (8). 6341 ( d 8273 (54), 8708 24); Herrera Ch., G. & W. Gamboa 6248 (67); Herrera, H., et al. 1027 Ap. 1387 (24); Hill, i et al. 17806 (67); Hilty, S. Jy-28 (21); Hitchcock, C. & L. Stanford 6930 (5); Hoffmann s.n. (46); I чү (67); Holland, D. & B. Kid 105 (54); Holst, 3 (54); yc R.. et al. ei ups Hunnewell, F. v^ (46): Huft, M. & Т. Jacobs 1991 Ibarra M., e et al. 3720 (2). Jacobs, B. 2394 (5), 5 (5), 2928 (42); om J. & F. Coello 4254 (5), 4274 (5); Jiménez M., , 962 (67), 1866 (54), 2564 (46), 3626 (42), dem (67), 3988 (67); Jiménez, Q. & L. Elizondo 740 (24). Kant, A. 31 (5); Kennedy, H. 1155 (24), 3258 (72): Kennedy, Н. & R. Foster 2191 (24); Kennedy, Н. & G hips 3093 (5); Kernan, C. 695 (39), 1216 (39); Pw & P. Phillips 629 (8), 1126 (50); Kirkbride, J. M 683 (35), 728 Ay 781 (75), 809 (54), 1265 a 1279 (37); Knapp, S. 416 (24), 1013 (24), 1843 (49): Knapp, S. & M. Huft 4423 (49), 4578 (64); Knapp. S. & K. Sytsma 2493 (36), 2548 (20); Knapp. S., et al. 2086 (47), 3646 (32), 4065 (54), 4738 (49), 6058 (10); Koch, T3 316 Annals of the Missouri Botanical Garden S. 8772 (1); Koptur, S. 297 (54); Kress, J., et al. 9593 (75); Kuntze, Lt (46). Laguna, A. Ay Dess B. 8-2 (10); Lao, E. 385 (55); Laan . 114 (63); Lawton, R. 1281 (34); Lems, K. 5156 (55); R. 079 (67), 1354 (54), 1922 (42), SS 4096 (54); León, J. 3254 (67); Lewis, W., et al. 986 (54), 2808 (53); Liebmann, К. 7 (2), ТА (2) 29€ (5), 15342 (5); e пег, За 960 (24), 14118 (5). 3266 (50); FEE R. . Judziewicz 14664 (24); Liesner, R., et al. 75 s т inden, M. s.n. (5): ae E. 6020 (46); Lis ie B. 107 (24); Long, L. 79 (75 López, E., et al. 14); López, F. & _ Sig shez M. 28 Lundell, C. E e ), 6277 (7), 16466 (5); Lundell, Contreras 18898 (7), Pn im 19838 (7), ies 4 (5), 20259 (5); Lutevn, J. 689 (67). 757 (5). P. & B. Alpin 1492 (27); dd P. & R. deed 4952 (56); Maas, P., et al. 1585 (24); Madison, M. 176 (5); MacDougal, J. 779 (46); MacDougal, J., et al. 2 lu c (41), 394 (49), 3170 (3), 3231 (54 ); Madriñán, S. & . Barbosa 192 (5); Magaña, M. & S. Zamudio n (5): 1, А. & Н. Gutiérrez 66 (50); Marín, J. 107 (39), 134 (50); Marín, J. & D. Marín 520 (8); Marín, J. & F. Quesada 263 (54); Martínez, C. 1074 (5); Martínez S., E. 2159 (14), 2199 (42); Martínez S., 5 & R. Riviere 2072 (14), : (42); Martfnez-Calderon, G. s.n. (5), 884 (2); ieri s ~ = = к= К. 381 (55); Martínez . M., et al. 90 (1); McDonagh, J., al. 202 (11); McDowe 11. 1 ый д, 754 (24), 780 (54) McPherson, G. 675: (35). 6753 (35). 6861 (64), 715t (75), 7161 (24), 7236 (54), 7342 (32), 7382 (24), 7390 (75), 7425 (40), 7639 (36). 7691 (69). 7744 (55), 7782 47), 7784 (54), 7785 (54), 7827 (47). 7904 (10), 8007 24), 8069 (55), 8089 (54), 8388 (54), 8511 (24), 8592 (54), 8688 (75), 8706 (35), 8753 (15), 8802 (30), 9083 47), 9107 (51), 9265 (46), 9402 (55), 9680 (75), 9747 40), 9821 (46), e (04). 10033 (64), 10406 (75). 11230 (66), 11533 (65), 101 (5), 13541A (47), 13566 (30), 4042 (11), 14009 md 14079 (63); McPherson, G. & J. Aranda 10114 (62), 10163 deg Mcl oo G. & M. Merello 8296 (47); McPherson, С. R EE dsan 15921 (55), 15948 ino: rs (2). pe (49); Mc Pherson, G., et al. 1 (52); Meave del е ө J., et al. 1514 (2); Mé "s T. A. 6215 (2), 6315 (2), 6513 (2); Miranda, F. 7217 (2); Molina R., A. 8338 (54); Molina T R., A., et al. x d и 17038 (5), 17704 (24), 17822 (46); Мопго, А. 7 (54); Monsalve B., M. 3111 (60): Montenegro, E. TA (33); Mora, G., et E x 372 (54), 386 (54), 427 (54), 493 (54); Mora, L. 2292 (57): Moraga, C. 284 (24), 332 (54); Morales, F. & € a 1783 (67); Morales, J. p (55), 4235 (46); Morales J Abarca 1819 (46), 5217 (46); Morales, J.. 1855 (24), 3400 (4€ j. 5480 (34); Moreno, : “BA (54): Mo- 3193 (5). 6458 (54), 7511 Кы, 10585 (5), 12377 (54), (5 . 14984 (54), 15054. (24), 15169 (54) wi (54). (54). 16128 (54), 17270 (5), (5). 19789 (54); Moreno, P. & J. Henrich 847 T Me reno, P. & J. Sandino 7623 (75), 12752 (54), 12925 (54), 12977 (24), 13101 (54) 6278 (64), 7094 (5). 8001 eu A. 7330 (33), 7650 (75); Mori, 5. & J. Kallunki i (40). 2314 (11), 2388 ed 2575 (24), 2591 (42), 2634 (24). 2721 (64), 2919 (49), 3082 o 3211 (75), 3417 (53), : (42), 48: dps 5258 (64). 5275 (31), 5733 (55); Mori, S. & J. Withe Pio 1948 (24); Mori, S., et al. 3731 (64), 4003 (: e 4172 (40), 6077 (64); Mowbray, T. & R. Wilbur 1259 (54); Murphy, H. 1041 (46) reno, 10583В (5). Navarro V., О (27); Navarro V., E. & A. Picado d (46); Nee, Т, к (65). M nri (55). els 42); Nee, M. & G. Diggs 2488: >, M. & A. Gentry 8723 (24); Nee, M. & hod n 18520 (2 y Nee, M. & W. Robleto T. 28089 (5: E Neill. D. 779 (54), 1331 (54), 1503 (MO), 1504 (42), 1659 (MO), 1876 (5), 3381 (24), 3974 (54); Neill, D. & P. Vincelli 3298 (55), 3602 (42); Nevers, G. de & H. Herrera 6646 (14); Nevers, G. de & C. de León 3598 (41); Nevers, G. de, et al. 4046 (53), 4076 (24), 6408 (11), 6738 (24). 8046 (11), 8061 (36), 8077 (37), 8523 (32) , 8528 (63); Nevling, L. < F. Chiang 1621 (2); Nufiez, P., et al. 14257 (5); Nufiez, T., et al. 522 180 (24), 201 (14); Oersted, A. 25 4): Hs К. 2383 (5); Opler, P. 1594 (24); Ortiz, bandi: N., et al. (46), 28C (51), 28D (5 627 (54), 889 (54), 891 (5), 1 (54), V., et al. 1 3509 (35); Palmer, C. s.n. (55), 85 ma , R., et al. 665 (42), 908 (49); Penneys, D. 55); Peralta R. 609 (42); Sou P. 8522 (49); Picado, A. & B. Gamboa R. 342 (46), 359 (46); Pipoly, J. : 4713 (5), 7026 (64), 7064 (33), 7065 (56), : 1096 89, 7113 (1. d 2 55), 7124 (67); Pipoly, J. & R. Bethancourt 7043 (40); Pipcly. J., et al 16729 (29), us (29), 16601 o 16958 (29), 17048 (29), 18000 (29), 18307 (29); Pitman, N. & J. Clark 942 (35); Pittier, H. 389 (67), 680 (55), 2044 (46), 2395 (54), 3255 (56), 3504 (75), 8513 (5), 14112 (46), s.n. ж, 518 (67) 1035 (24), Poveda A., L. 302 (55), 34); (16); Proctor, G. 31907 Eel ) 5). Poveda A., L., et al. 4174 31925 (33), 32008 (46), 32238 (42), ee Purpus, C. Quesada, F. 530 (8 Ramirez, V. i В. Неше 159 (24); Ramírez, V. & F. Morales 145 (46); Raven, P. 21643 (39), 21653A (27), 7 (27): Record, 5. 6 (54), 25 (54); Reveal, J. & J. 4922 (63); Ríos, D. 305 (54); Ríos, P., et al. 143 (54); Rivera. G. 66 (67), 231 (55), 1645 (46); Rivera, G., et al. 2237 (8); Riviere, R. 337 (14), 340 (42), 358 (14); Robles, R. 1160 (75), 2131 (7: 2 дер, R. m A. Chac ‘on 2683 (16); ie "s, f. 54), 501 (55 Relig С. R. 29‹ ) (46) ыы! G. 226 920 (5). 1145 (2); Rojas, E. 173 abe " G. 3493 (67); Rubio, D. & C. Quelal 1453 (25): Rueda, R. 3291 (24), 8564 (42), 8670 (42), 9731 (42); Rueda, R. & I. Coronado 6495 (24), 7675 (23), 8114 (54); Rueda, R., et al. 1409 (54), 1798 (14), 2674 (54). 2870 (75), 3178 (75). av d^ 3438 (54), 3936 (54), 3974 (5). 4054. (14), ere . 4458 (59), 4528 (59), 5133 (42), 5301 (42), 5457 (24), 5610 (75), 6184 (42). 6268 (54), 8198 (42), Sei (49), 8611 (49), 8656 (49), 8765 (59), 8877 (14), 9005 (59), 9300 (54). 10022 (59), 10027 (59), 10131 (59), 10202 (54); Russell, G., et al. 942 (75), 976 (17). 32020 (55), 32130 (24 1242 (2), 5959 ( TA (3 л Sánchez, P. & 3575 Sánchez, J. & A. Cascante-M. 351 (54); : N. my 561 (16); ire г Ad 3341 (75), : (5), 5 54): Sandino, J.. 2923 (54); San Román, M. s. i Schatz, G. & "n uai 600 (24); Saunders, T (54). 1162 (54), 1203 (54); Schipp, W. 124 (54). 1248 (54), 5-559 (5), 1365 (7); Schmalzel, R. 1161 (24); Schmalzel, R. & D. Roubic k 1334 (33); Schnell, C. 1119 (55); Schubert, B. & J. Cordoba 929 (67); Schultes, R. & B. Reko 273 (2); Seemann, B. 59 (54), 1093 (58); Se- gura, M. & К. Quesada 93 (39), 148 (39), 189 (39); Shank, › & А. Molina R. 4319 (54), 4412 (16), 4740 (54), 4777 (54). 4872 (54), 4891 (75). 4983 (54), 4995 (54); Shan- بے Volume 90, Number 2 2003 Ricketson & Pipoly 317 Revision of Ardisia subg. Auriculardisia non, W. 1256 (5); Shilom T., A. 3567 (2), 3678 (2); Sil- verstone-Sopkin, - et al. 5709 (5); Sinaca C., S. & С. Ibarra M. 1094 Skutch, A. 3296 (55), 3538 6). 5025 (44), 5237 (44), 5310 (8); Smith, A. 34 (46), (46), H-90 (46), H-145 (46), 501 (55), 2015 (46), bu (46), 2040 (55), 2403 (46), 2244 (46), 10039 (46), 10074 (46); Smith, D. 208 (42), 290 (54), 478 (42), 607 (54), 618 (24), 2507 (24); Smith, H. 2738 3 (5), 2782 (5); Smith, J. 6640 (24), 6641 (46); eov S. 9517 (2); Solano, J. 274 (42); Solís R., К. 152 (46); Solomon, J. 2465 (6 i E i rry, J. 5 (54), 1004 (5); Standley, P. 7729 (3), 33144 is 5) 34157 (46), 34172 (46), 34442 (46), 35672 (54), 38593 (46), 40841 (5), 40848 (5), 41393 (54), 42001 (46), 42272 (46), 42765 Я 5 53157 (3), 53322 (3), 55460 (54), 70506 (5), 71096 (5). 89600 (5), 90697 (5); Standley, P. & R. Torres R. 47429 0) 20: 51266 (26); Standley, Р. (46), 44443 (5), 44729 (17 . 45422 (17), 45509 (5), 45853 n. (16), 48603 (16), 48754 (16), 48824 (16), 50843 (55 51556 (43), 51891 (54), 52114 (46), 52218 (46); Starry, D. 57 (5); Stein, B. 1306 (46); Stein, B. & C. i 972 (10); Stein, B., et al. 1223 ( 1 : Chambers 96 (55); Pen. W., et al. 395 (5), 569 (5), 2053 ; Stevens, W. 4799 (75), 4987 (54), 5061 (75), 6509 (55), 6841 (5), Кем сы 13784 (68), 13863 (68), 14005 (46). 18242 (46), 18744 (54); Stevens, W., et al. 14591 (5), 16651 (5), 20990 (54), 20997 (54), 21317 (54); Ste- yermark, J. 38745 (5), 39114 (4), 41701 (54), 41788 (54), 41795 (54), 41923 (12), 42022 (54), 44888 (54), 45311 (5). 49320 (5); Steyermark, J. & J. Fernández MA s (5); Steyermark, J., et al. 122965 (5); Stone, D. 3143 (67); Stone, D. & A. Welden 3440 (34); Stork, H. or (67), 3308 (67), 3361 (67); Sullivan, G. 201 (40), 519 (10); Sytsma, К. 1326 (53), 1642 (24). 1774 (36), 1821 (11), T (65), 1994 (40), 2022 (40), 2344 (40); Sytsma, K. . Andersson 4527 (69); Sytsma, K. & T. Antonio 3011 Dd 3015 (36); Sytsma, K. & W. Stevens 2085 (46), 2120 ck Sytsma, K., et al. 2515 (69), 2770 (11), 2828 (65), | (47). ч R. 228 (54); Taylor, C. 3875 (46), 3910 (46); Tay- lor, C. & C. Skotak 4348 e (24), 4749 (46); Tay- lor, J. 17845 (5 Tellez V., O., et al. 4894 (54), 5011 E Tenorio L., P. & C. Mota de k 11375 (1); Tenorio ^, et al. 12641 (: Em eph. M. 7 (46): Terry, M. & “Thomsen. T 95 (o0, 2044 (25 “1949 (67), Pag ИЕ 399 (39), 1024 (14 Todzia, C. 1080A E Tonduz, A. 1728 vis 1956 (46), 2131 (55), 2143 (46), 4314 (46), 9322 (54), 10778 (67), 11745 (54), 11824 (46), 12565 (54), 12632 [7460] (55); Toriz A, G. 584 (1); Torres C., H. Hernández 3261 (2), 3262 2 Tuc ке С. 2617 б Тап O., R. 2153 (5); Türckheim, Н. von on, E. 813 (46), 5309 (64); Tyson, Эх » Loftin 5903 (46); Тузоп, E., et al. 3279 (64), 4 (i Utley, J. & K. Utley 1422 vg 2035 (67), 2249 (67), 2393 (34), 3250 (24), 4022 24, 4082 (24), 4463 (55), 5347 (5), 5349 (75), K. 584 ; Valdespino, I. & J. a Т (64); Valerio R., J. 1317 (46); Valerio, M. 130 (5), 208 (67), 1041 (46), 1194 (55). 1579 (55), 1624 e Vargas, G., et al. 1470 (46), 1628 ay к» G. Я 5 24); \. M.. et al. 967 (e 1055 (5); Vega, S D Vilas R. 6 (27), 205 (27); Vogel, S. 73 (55 LEM j^ Webster G. 22135 (67); Webster, G., et al. 12278 (67) OW 2 1344 (54), ins 64. 2393 09) Wendt, T. et al. 5736 tefoord, C. & A. Eddy 430 (73); W 14510 (67), 14619 (67), 19934. (67), 19948 (67), : 723 (67), 24260 (46), 24498 (46), 24504 (46), 67 1 (67), 24906 (43), 9 (54), 28168 (54). 28177 (54), 28298 (54), 28307 pons 29460 (67), 30269 : 37186 (54). 38351 (54). 3887: 004. ‚ 63183 56. 63199 pe 63252 (54), 63266 (54), 63376 (54). 63638 (54), 63719 (54), 64991 (54); Wilbur, R. & F. Al- meda 17029 (67); Wilbur, R. & B. Jacobs 34432 (5); Wil- bur, R. & J. Luteyn 18391 (46), 19465 (64); Wilbur, R. & G. Moore 70233 (54); Wilbur, R. & T. Mowbray 63533 (24); Wilbur, R. & D. Stone 8718 (6 7); Wilbur, R. & J. Teeri 13294. (46), 13677 (67); Wilbur, R. & K. Williams 66813 Мең Wilbur, R., et al. 10988 (46), 13111 o 56), 23485 En 23508 (67); Williams, L. S i R. Williams 18793 (54); Wiliams | & A. Molina R. Eo! (54); Williams, L., et al. ‚ 24118 (55), 24150 (67), 28503 (46), 28618 Ба 28998 (68), 29080 (24); Willia 3 (37); Wilson, P. 79 (3); Woodson, R., et al. 807. (46). Yuncker, T., et al. 4588 (3), 8832 (5 Zambrano C., O. 1052 (5); Zamora, N., el al. 1540 (67), 1861 (8); Zamora V., N. & P. Sánchez E. — Zarucchi, J., et al. 7232 (52); Zola B., M. 610 Volume 90, Number 2, pp. 129-318 of the ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on June 13, 2003. www.mobot.org/mbgpress dam P 4 & John J. Pipoly Ш Annals . of the Missouri Dotanical Garden 2003 G Annals of the Missouri Botanical Garden Volume 90, Number 3 Summer 2003 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. All manuscripts are peer-reviewed by quali- fied, independent reviewers. Authors should write the Managing Editor for informa- tion concerning arrangements for publishing in the ANNALS. Instructions to Authors are printed in the back of the last issue of each volume and are also available online at www.mobot.org/mbgpress. Editorial Committee Victoria C. Hollowell Editor, | Missouri Botanical Garden Amy Scheuler McPherson Managing Editor, Missouri Botanical Garden Diana Gunter ` Associate Editor, Missouri Botanical Garden Kevin Brown - MBG Press Assistant, Missouri Botanical Garden Barbara Mack Administrative Assistant 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 Charlotte Taylor issouri Botanical Garden Henk van der Werff Missouri Botanical Garden For subscription information contact ANNALS OF THE MISSOURI GARDEN, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2003 is $145 per volume U.S., $155 Canada & Mexico, $180 all other countries. Four issues per vol- ume. The journal Novon is included in the sub- scription price of the ANNALS. annals@mobot.org (editorial queries) http://www.mobot.org/mbgpress 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. Periodicals postage paid at St. Louis, MO and additional mailing offices. POSTMASTER: Send address changes to ANNALS OF THE MISSOURI BOTANICAL GARDEN, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. The Annals are abstracted and/or indexed in AGRICOLA (through 1994), APT Online, BIOSISG, ingenta, ISIG databases, JSTOR, Research Alert®, and Sci Search®. © Missouri Botanical Garden 2003 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. ть; 5 ^t Desrmr c 3 T # ANCI/NICA 720 AQ 100^ rn. fi ar GO IFFT ww АЫ Volume 90 Annals Number 3 of the 2003 Missouri Botanical Garden SYSTEMATICS OF EURASIAN Inés Álvarez Fernández? AND NORTH AFRICAN DORONICUM (ASTERACEAE: SENECIONEAE)! ABSTRACT The genus Doronicum (Asteraceae: Senecioneae) comprises perennial herbs distributed in Europe, North Africa, and Asia. A worldwide revision of the genus recognizing 26 species and 4 subspecies is presented. In the present taxonomic treatment no infrageneric groups are recognized. Seven names are newly lectotypified herein: Arnica doronicum Jacq., Doronicum caucasicum M. Bieb., Doronicum portae Chabert, Doronicum scorpioides Lam., Doronicum souliei Cavill., Doronicum thibetanum Cavill., and Doronicum turkestanicum Cavill. A new chromosome count is provided for D. “ко subsp. diazit. Key words: Asia, Asteraceae, Doronicum, Europe. North Africa, Senecioneae. The genus Doronicum L. (Asteraceae: Seneci- This genus belongs in the Senecioneae, one of oneae) includes rhizomatous herbs with yellow or the largest and most complex tribes in the Astera- green-tinted radiate capitula. All phyllaries are сеае with 123 genera and around 3200 соч similar, generally herbaceous and arranged in two (Cassini, 1819; Bentham & Hooker, 1873; Hoff- or three rows. Cypselae are cylindric to obovate- mann, 1892; Nordenstam, 1977; Bremer, 1994). Its cylindric with 10 longitudinal ribs and bear a pap- 26 species constitute a presumably natural group pus of white-tinted minutely scabrous capillary (Bremer, 1994; involucre without shorter supple- bristles. The pappus can be absent in ray flowers mentary bracts, phyllaries herbaceous arranged in of some heterocarpic species. two or three rows, and cypselae cylindric to ob- 'T am grateful to Gonzalo Nieto Feliner for suggesting and guiding this work and for his constructive — on the manuscript, the curators of AV, B, B-W, BC, ВСЕ, BM, BOLO, BP. BR, BRNM, C, CL, COI-WILLK, E, FI, С, BOIS, G-DC, GAZI, GE, ЄН, GRM, GZU, HVR, IRAN, JACA, A K, LAU, LD, LE, LEB, LINN, LY, E MACB, МАЕ, MO, NAP, NY, P-HA, P-LA, RO, S, SANT, SZB, UPS, W, WU, ZA, and the Sánchez Pedraja personal herbarium for the loan of specimens or for the sii ie provided, Jasmin Jakupovik and Jesás Sanz for information and comments on the chemistry of Doronicum, Alberto Herrero and Félix Mufioz-Garmendia for their help with ig n Jerez, and oo Pradillo for technical assistance. This work has been supported by grant DGES PB96- 0849 of the Spanish Dirección General de Ensefianza Superior e Investigación Científica ? Real Jardín E CSIC, Plaza de Murillo, 2, E -28014 Madrid, Spain. ines@ma- -rjb.csic.es ANN. MISSOURI Bor. GARD. 90: 319—389. 2003. 320 Annals of the Missouri Botanical Garden ovate-cylindric with 10 longitudinal ribs). Its geo- graphical area extends from Europe and North Af- rica to Asia, growing in mesic woods and open rocky places with moist soil, and near watercourses, from sea level to 5000 m of elevation. Although there are several regional studies of the genus (Turkey, Edmondson, 1973, 1975, 1978; Ar- menia, Avetisyan, 1980; Iberian peninsula, Cha- 1992), there is only one previous worldwide revision of Doronicum (Cavillier, 1907, 1911). Fifteen new species (Diels, 1922; Widder, 1925; Sergievskaja, 1949; Widder & Rechinger, 1950; Edmondson, 1973, 1978; Cha- cón, 1987; Pérez & Penas, 1990; Pérez et al., 1994; Chen, 1998) and six hybrids (Bornmiiller & Koch, 1930; Widder, 1934, 1948; Stace, 1991) were sub- sequently described. Two species included in Cav- Belgium, Duvigneaud, illier’s monograph have since been transferred to other genera (i.e., D. hookeri C. B. Clarke ex Hook. to Nannoglottis (Kitamura, 1980), and D. thibetan- um Cavill. to Aster (Álvarez Fernández & Nieto Fel- 2000)). Cavillier (1907, 1911) studied the morphology of the genus in great detail, especially the classification is of rather limited value since these iner, indumentum, but his proposed infrageneric groups are obscurely defined based mainly on non- exclusive characters. As a result, classifying newly described taxa in any infrageneric framework problematic. The need to evaluate the newly described spe- cies and to assess the infrageneric taxonomy pro- vides justification for this work. The objectives were to study as many morphological characters (quali- tative and quantitative) as possible so that (1) only entities that could be consistently diagnosed were recognized in the taxonomic treatment, and (2) spe- cies were classified as groups on the basis of shared synapomorphies (secondary homologies, De Pinna, MATERIAL AND METHODS More than 50 qualitative and quantitative mor- phological characters were studied in ca. 4300 dried spec imens from the following herbaria: В, BC, BCF, BM, BR, BRNM, COI-WILLK, E. FI, G, GAZI, GH, GZU, НУК, IRAN, JACA, JE. K, LAU, LE. LINN, LY, MA, MACB, MAF, MO, NY, RO, S, SANT, UPS, W, WU, ZA. and the Sánchez-Pedraja personal herbarium. From other institutions, only photographs and photocopies of specimens, or ad- information, were available: . B-W. G-BOIS, G-DC, GE, GRM, LD, LEB, PAL, and SZB. ditional BOLO, BP, C, CL, NAP, P-HA, P-LA, A list of species and subspecies and an index to exsiccatae are presented in Appendices | and 2, respectively. Observations were made directly or with the aid of binocular lenses. Microcharacters of indumen- tum and cypselae were studied by SEM. Quanti- tative characters were recorded using a Brown & Shape Plus digital caliper (model 599-57 1-3). Mea- surements were made on herbarium specimens, af- ter flattening and drying. Distribution maps for each taxon are based on the specimens studied. Note that geographical ar- eas and follow Hollis and Brummitt 1992), and major political divisions for countries countries —. were included when that information was available. TAXONOMIC HISTORY The name Doronicum is apparently derived from the Arabic word “darawnay,” used for at least two different plants (Dozy, 1877). The pre-Linnaean botanists (Dioscorides, 1554, 1557; Dodoens, 1574) and other Greek authors referred to species of Do- ronicum as Aconitum pardalianches, and the plant was probably introduced in Western culture by Av- 1587). The genus Doronicum was described by Lin- icenna (Dodoens, 1574; Dalenchamps, naeus (1753) to include four species, only two of which are currently accepted in the genus: D. par- dalianches and D. plantagineum. The remaining species correspond to Senecio and Aster, respec- tively. Further, one species of Arnica described by Linnaeus (1753), A. scorpioides L.. also belongs in Doronicum as recognized by Lamarck (1786). Sev- eral pre-Linnaean authors also confused species Aster, and with Doronicum (Dalenchamps, 1587; Clusius, 1601; i 1623; Tournefort, 1700). In particular, the overall morphological similarity between Arnica and Do- of Senecio, Arnica Bauhin, ronicum suggested their close affiliation until the 970s. However, Nordenstam’s (1977) micromor- phological study of style, anthers, and pollen de- finitively has excluded Arnica from the Seneci- oneae. The cypsela dimorphism that occurs in some species has been a relevant feature in the taxonom- ic history of the genus. Lamarck (1786) referred to the he Чегос агрїс and homocarpic species as “ar- niques” and Necker (1790) even proposed the different genus Aronicum "doronies," respectively. for the homocarpic species, and his classification had wide acceptance in the 19th century (de Can- 1838; Koch, 1843; Hausmann, 1851; Rei- 1854; Schur, 1866; 1867; Si- 1886). although only Aronicum dolle. chenbach, Ardoino, monkai, some authors recognized as a section of Doronicum Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) 1854; Willkomm & Lange, 1870; 1892; Beck, 1893). Aa classification persisted until Bentham and Hooker (Ambrosi, man, 1879; Hoffmann, (1873a) placed Aronicum as a synonym of Doroni- cum. Tausch (1828) also recognized these two groups but referred the heterocarpic species to Par- dalianches, not Aronicum Cassini (1817) established the genus Grammar- thron with two species, G. biligulatum and G. scor- pioides, now subsumed within Doronicum (Jacquin, 1773; Lamarck. 1786); de Candolle (1836) de- scribed the monotypic genus Fullartonia (F. ka- maonensis), now D. kamaonense (DC.) Alv. Fern. (Alvarez PNE RM 2( 1838, de Candolle proposed two sections within Doronicum: section Eudoronicum, including some species of Senecio, and the monotypic section Chromochaeta, with Do- ronicum linifolium (Wall.) DC., now also in Senecio (Maguire, 3). Webb in Webb and Berthelot (1846) treated 1833-1835: t: section of Plon genus Pericallis D. Don (in Sweet, 228) including five species as a icum (D. cruentum, D. echinatum, D. papyraceum, D. tussilaginis, and D. webbii), but it is now treated — at its original rank (Nordenstam, 1978 Cavillier (1907, 1911) divided his study of the genus Doronicum, the first devoted to the study of the homocarpic species (1907), and the second to the heterocarpic ones (1911). concluded (1911) that this character was not useful to delimit natural groups. Cavillier proposed a new classification (1911) that included 3 sections, 7 subsections, and 34 species within Doronicum: sec- However, he later tion Doronicastrum (subsect. Corsica, subsect. Aus- triaca, subsect. Cardiophylla, subsect. Macrophyl- la, subsect. Pardalianchia, subsect. Plantaginea, and subsect. Grandiflora), section Soulieastrum (D. stenoglossum Махіт.), and section Hookerastrum (D. hookeri С. B. Clarke ex Hook.). i sectional treatment is not satisfactory, since section owever, this Hookerastrum was described on the basis of a spe- cies from another tribe (Nannoglottis hookeri, As- tereae) and the phylogenetic position (Alvarez Fer- nández et al., 2001) of section Soulieastrum’s only taxon precludes recognition at the sectional level without artificially splitting the bulk of the genus. e subsections in Doronicum were defined (Cav- illier, 1911) mainly from the shape of basal leaves, the size of leaves, and the presence of cypselae dimorphism. Phylogenetic study of the genus (Al- varez Fernández et al., 2001) concluded that these are not synapomorphic characters, and therefore Cavillier's classification does not recognize natural groups. Even after Cavillier’s work the relevance of het- erocarpy was claimed red (Gorschkova, 1961). In the Flora of the U.S.S. posed by Willkomm А Lange (1870) was merged with that of Cavillier (1911) to distinguish two sec- tions in Doronicum (Gorschkova, 1961): section Aronicum (ser. Altaica, ser. Carpatica), and section „ the classification. pro- Pardalianches (ser. Austriaca, ser. Cardiophylla, ser. Macrophylla, ser. Pardalianches, and ser. Plan- taginea). More recently, Edmondson (1978) fol- lowed Cavilliers classification (1911) instead of Gorschkova's (1961), but described the monotypic subsection /saurica (D. cacaliifolium Boiss. Heldr. within section Doronicum, which corre- sponded to Cavilliers section Doronicastrum Xe) The phylogenetic analysis herein confirms what a preliminary morphological study suggested: the morphological characters used are too labile to pro- vide a sound classification at the infrageneric level 2001). The molecular data from nuclear ribosomal and chloroplast DNA (Álvarez Fernández et al., suggest some groups but without enough support to recommend formal taxonomic groupings. One ex- ception is a Mediterranean group of species (D. plantagineum group), which receives good support from both molecular and morphological data. To avoid adding to the already complex taxonomic his- tory of the genus, a formal infrageneric treatment is not proposed here, but is deferred against new evidence, As already mentioned in the introduction, after Cavillier’s revision and until the present work, a large number of taxonomic actions were taken with- in Doronicum. These include the description of new species (Sergievskaja, 1949; Edmondson, 1973, 1978; Chen, 1998, among others) and one subsec- tion (Edmondson, 1978), and a few lectotypifica- tions (Chacón, 1987; Pérez et al., 1997; Jarvis & Turland, 1998). Despite all of these actions the ge- nus was still lacking nomenclatural stability. Thus, during this study and immediately preceding this work, several nomenclatural and taxonomic clari- fications were done (Alvarez Fernández & Nieto Feliner, 1997, 1999, 2000; Álvarez Fernández, 2001). The lectotypification of 16 names of Doron- icum in current use (Álvarez Fernández & Nieto Feliner, 1999) gave the genus nomenclatural sta- bility leading up to this revision. In this work, ad- ditional lectotypification of seven names belonging in Doronicum, although not in current use, is also presented to consolidate and clarify as far as pos- sible the nomenclature of the genus. Despite the efforts made to locate type material for all the names in current use, 4 out of 30 names (i.e., carpaticum, D. clusii, D. corsicum, and D. orientale) 322 Annals of the Missouri Botanical Garden still required further investigations for lectotype designation. Because at present these names clear- ly represent different recognized taxonomic entities, they are cited herein as names in current use, al- though their formal identity is not conclusive until lectotypes are designated. GEOGRAPHICAL DISTRIBUTION Half of the 26 recognized species of Doronicum are distributed in Europe and North Africa. Seven of the remainder are from southwestern Asia (Iran, Iraq, Caucasus, and Turkey), and 6 species are dis- tributed in central Asia (Turkistan, Altay, Tibet, Yunnan, and the Himalayas). With the exception of Doronicum orientale, each species is restricted to one of the three well-delim- ited areas: Europe, southwestern Asia, central Asia. Doronicum orientale is distributed in Europe and southwestern Asia, abundantly in the eastern Med- iterranean (Greece, western Turkey, southern Italy, and Lebanon-Syria), and scattered in central Eu- rope, where its proximity to inhabited places sug- gests possibly having escaped from gardens. Delim- iting the natural of distribution is also difficult in the case of two other European species: D. plantagineum and D. pardalianches. Both were used as ornamental plants in previous centuries (Pena, 1571; Miller, 1787) and now are considered alien plants in the United Kingdom (Harron, 1986; Clement & Foster, 1994). The species discussed above (D. orientale, D. areas plantagineum, and D. pardalianches) occur in sim- ilar mesic habitats from sea level up to subalpine regions, but not in high mountain habitats (the up- per tree-line). Doronicum hungaricum, occupying similar habitats in Eastern Europe, can be consid- ered vicariant with D. plantagineum in this region. Doronicum austriacum is widely distributed in Eu- rope, most abundantly in the Austrian Alps, Mac- edonia, and Ukraine, always in subalpine regions, and in the Iberian peninsula it occurs only in a few localities in the eastern Pyrenees. The strictly alpine species of the genus in Eu- rope are represented by Doronicum grandiflorum, D. clusii, and D. glaciale. The first of these species is the most widely distributed of them. It is abun- dant in the Alps, the Pyrenees, and in the Canta- brian range (northern Spain). In addition, there are two specimens from Corsica dated 1878 and 1917, suggesting its extinction on this Mediterranean is- land, which has well known floristic affinities with the Alps (Briquet, 1901). Doronicum clusii is pre- sent in the Alps and Carpathians, while D. glaciale is restricted to the Alps (mainly the Austrian Alps). where it can coexist with D. clusii. The alpine and subalpine habitats of the central and northern half of the Iberian peninsula (except the Pyrenees) are occupied by D. carpetanum, under which four sub- species are recognized. Other alpine to subalpine species in central and eastern. Europe аге Doronicum columnae and D. carpaticum. The first is widely distributed from It- aly to Romania, and D. carpaticum is restricted to the Carpathians. The remaining two European species are endem- ics, D. cataractarum in the Austrian Alps and D. corsicum in Corsica, and both occur in subalpine habitats. The represented by seven species. Only one, D. oblon- gifolium (from the Caucasus), is morphologically quite different from the others. Three of them are widely distributed: D. macrophyllum (Caucasus and northern Turkey), D. dolichotrichum (Caucasus and and D. maximum (east- genus Doronicum in southwestern Asia is south of the Caspian Sea), ern Turkey, and south of the Caspian Sea). Of the three remaining species a limited number of spec- imens are known, and this results in a scattered distribution. All of the central Asian species (D. altaicum, D. briquetii, D. falconeri, D. gansuense, D. kama- onense, D. stenoglossum) overlap at least in one point of their distributions. MORPHOLOGY RHIZOMES All representatives of Doronicum are perennial rhizomatous herbs. The shape and structure of the rhizome are constant within each species, but are not exclusive to any one. These characters are use- ful, sometimes indispensable, to discriminate be- tween species. There are fleshy or woody (or some- what woody) rhizomes in Doronicum. This character Cavillier (1911: as was described by C vided histological diagrams, “non tubéreux,” To distinguish be- tween these, observations on fresh material are re- fleshy rhizomes respectively. quired, although when pressed, flatten easily while the woody ones retain their orig- inal more or less terete shape. When fresh, fleshy rhizomes are succulent and brittle, while woody ones are fibrous and tough. Fleshy rhizomes are easily recognized in some European species (e.g., D. plantagineum, D. pardalianches, D. hungari- cum), while woody rhizomes are well represented in Asian species (e.g.. D. macrophyllum, D. maxi- mum, D. stenoglossum). In a few cases, rhizomes are fleshy to somewhat woody and cannot be as- Volume 90, Number 3 003 Álvarez Fernández 323 Doronicum (Asteraceae) signed to either type (e.g., D. grandiflorum, D. ca- taractarum). ithin a species, rhizome internodes may have roughly constant length and width (e.g.. D. altai- G), or may vary in length and width, cum, Fig. 1 i Most species resulting in stolon-like structures. have the former condition, and only D. orientale (Fig. 1A), D. plantagineum, D. pardalianches, and sometimes D. hungaricum have clearly irregular in- ternodes. Sometimes rhizome scales remaining from the sheath of basal leaves from previous years (e.g., D. carpetanum, Fig. 1F). The persistent remains of basal fibers from old petioles occur, for example, in Doronicum oblon- gifolium, but they are frequently absent (e.g.. D. austriacum, Fig. Hyaline, shiny, and smooth trichomes are some- nodes have brown-tinted times present on the younger nodes of rhizomes and also in the axils of basal leaves. Sometimes these trichomes are long, abundant, entangled, and white to yellow, and they can cover a large part of the rhizome (e.g., D. orientale). Such rhizomes were re- ferred to as “ériopode” by Cavillier (1911: 199) in contrast to “gymnopode” rhizomes, which lack this indumentum (e.g., D. columnae). In many cases it is difficult to see trichomes on rhizomes, because they are short and scarce and can be covered with leaf remains (e.g., D. carpetanum, D. clusii, D. gla- ciale, and D. grandiflorum). uds are evident on some fleshy rhizomes (e.g.. D. риди, Fig. IB). These stem buds can be oer seen in plants two years or older, but these must be collected carefully. Sometimes the scales that cover young buds can also be observe wo species each have unique хоне. Doron- icum cacalüfolium has r rhizomes with spherical, swollen internodes, alternating with LT constrictions, sometimes covered by a fibrous net. In the second type, seen in most D. stenoglossum, the main stem is inserted on a convex a ino kino surface. Sometimes, pieces of : oody organ perpendicular to the stem were es и лг үз the whole structure has not been seen, it is presumed to be a kind of distinct woody rhizome, but further study of the subterra- nean organ is needed. Adventitious roots are present in Doronicum stenoglossum and sometimes in D. kamaonense a few centimeters above the subterranean woody or- gan, suggesting that the lowest vertical part of the stem was buried. STEMS The stems in Doronicum are always erect, fistu- lose, cylindric, and slightly ribbed, green when fresh, and pale yellow to brown when dry. Generally stems are straight, but zigzag stems occur in some species (e.g., D. austriacum). The stems are often simple and end in a single capitulum. When bear- ing several capitula, the stem is branched only in the upper part. Exceptions are seen in D. stenog- lossum and D. kamaonense, which sometimes have branches on the lower part of the stem. The main stem terminates in a capitulum, which matures first. Further capitula, if any, are on ter- minal lateral branches, which for the most part overtop the main head. Each species generally has a characteristic number of capitula, e.g., one in D. falconeri, up to 5 in D. pardalianches, and more than 5 (up to 20) in D. corsicum. LEAVES Leaves are simple and alternate. Leaf characters have been traditionally used in the taxonomy of the genus (Cavillier, 1911), but their usefulness is lim- ited to the specific level. The shape and size of leaves are variable even within a single specimen for some species. For this reason, basal leaves (those inserted on the rhizome nodes) and cauline leaves are necessary for descriptive purposes. Sim- ilarly, cauline leaves are distinguished by position as lower, middle, and upper, i.e., inserted in the basal, middle, and upper third of the stem, respec- tively. In some species, basal and lower cauline leaves are usually absent at flowering time Basal leaves are petiolate, the petiole ж short and wide (e.g., D. briquetii), or much longer than the leaf blade (e.g., D. columnae). In species with arge basal leaves, sheaths are conspicuous (e.g.. D. macrophyllum, D. maximum, and D. dolichotri- chum). Acropetally, along the stem, the petiole eradually shortens, often leading to fiddle-shaped leaves. The upper cauline leaves are reduced, ses- sile, and ovate to bract-like. This leaf transition is marked in D. austriacum, D. carpetanum, D. ma- crophyllum, and D. pardalianches. Leaves may be orbicular, ovate, elliptic, and ob- ovate, as well as fiddle-shaped or bract-like. The base of leaves can be cordate (Fig. 2A). subcordate, truncate, or attenuate (Fig. 3A, C, F, gins are generally entire to subentire, sometimes markedly dentate (e.g., D. cacaliifolium, D. colum- nae, D. corsicum, and D. grandiflorum). Number and arrangement of cauline leaves de- . Leaf mar- termine to a large extent the architecture of the plant. In some species the number of cauline leaves is low (2 to 4, e.g., D. orientale) and they are con- fined to the basal third of the stem. In other leafy species (D. austriacum, D. corsicum, or D. altai- 324 Annals of the Missouri Botanical Garden D i n Figure 1. Rhizomes in Doronicum. —A. garicum (drawn from Grundl s.n., G B). —D. Doronicum cataractarum (drawn fro 52, — Е. Doronicum carpetanum sub altaicum ( ›‚ as D. longifolium). — پچ SSS)‏ EN 3‏ w. :1]‏ جت Е === Doronicum orientale (drawn from Willing 3515, В). —B. Doronicum hun- oli 1. Doronicum austriacum (drawn from Strid et al. 18585, m Hépflinger s.n., BM). —E. Doronicum columnae (drawn from Sladen 9/ sp. carpetanum (drawn from Luceño & Vargas 208, MA). —G. Doronicum drawn from Krasnoborov et al. 959, K) Volume 90, Number 3 Álvarez Fernández Doronicum (Asteraceae) Doronicum pardalianches (drawn from Rivas-Goday s.n., MA). —A. Basal leaf. —B. HE 29493, K). —C. Upper cauline — .—E. Ше! ins leaf. B —G. asal Figure 2. А,В. of basal leaf. C, D. Doronicum dolichotrichum (drawn from Davis & Hedge Indumentum of upper cauline leaf. E, F. Doronicum briquetii (drawn from Rock 22380, —F. Indumentum of upper cauline leaf. G, H. Doronicum altaicum (drawn from Joubert et al. 959, leaf. —H. Margin of basal lea 326 Annals of the Missouri Botanical Garden 2 1ст \ Figure 3. А, B. Doronicum канын (drawn from bip d 83/56, ae —A. Basal leaf. —B. Indumentum of basal leaf. C-E. Doronicum glaciale (drawn from Steininger B). —C. Basal leaf. —D, E. Indumentum of basal leaf. F-H. Doronicum clusii (drawn from Castroviejo et al. 11 615. MA). a Basal leaf. —G, H. Indumentum of basal leaf. Т, J. Doronicum oblongifolium (drawn from Albury et al. 3176, K). —1. Basal leaf. —J. Margin of basal leaf. ты Volume 90, Number 3 2003 Álvarez Fernández 327 Doronicum (Asteraceae) cum) leaves are arranged along the stem. The larg- est leaves are usually seen at the middle or basal parts of stems. Leaf venation is a good taxonomic character, eas- ily observed in dry specimens and preferably from basal leaves. For its description and categorization the terms proposed by the Leaf Architecture Work- ing Group (1999) are used. Most species have an actinodromous venation for first vein category (e.g.. D. grandiflorum, D. carpetanum, D. reticulatum) in which all secondary and tertiary veins are more or less equally evident. Pinnate venation for the first vein category occurs in central Asian species. In this type, the tertiary veins are not well marked, and both the secondary veins and the main vein are equally prominent and thick (e.g., D. altaicum, D. gansuense, D. kamaonense, D. stenoglossum). The acrodromous type of venation for the first vein category is restricted to a European group of spe- cies (D. hungaricum, D. orientale, D. plantagi- neum, and D. columnae). Intermediate cases be- tween the latter and the actinodromous type occur in D. columnae, D. carpaticum, and D. pardalianch- es, and between pinnate and actinodromous vena- tion in D. clusii and D. glaciale. HABIT Four main habit classes can be distinguished: (1) An "orientale" type: solitary capitulum with a scapose stem, sometimes bearing bract-like leaves: a few cauline leaves (2 to 4) inserted in the basal third of the stem. It is displayed by some European species (e.g., D. orientale, D. columnae, D. plan- tagineum). (2) An “altaicum” type: generally a single capitu- lum; a mostly leafy stem and a variable number of uniform leaves (4+) along the stem length, or at least in its lower half. This is present in some cen- tral Asian as well as European species (е... D. altaicum, D. falconeri, D. grandiflorum). 3) A “macrophyllum” type: several capitula; stem branched in the upper third; large cauline leaves (3 to 5) mainly in the lower half of the stem, and bract-like leaves on the upper stem. This is re- stricted to the southwestern Asian species (4) A “corsicum” type: several нат wn a айе number of + uniform leaves (5+) + evenly insert- ed along the stem. This is characteristic of D. cor- sicum and D. austriacum. In some species the habit does not correspond to these patterns (e.g., D. kamaonense and D. sten- oglossum, which are sometimes branched from the base), and sometimes intermediate patterns occur e.g., D. pardalianches, D. cacaliifolium). 8 p CAPITULA All Doronicum species have radiate, hemispheric to widely campanulate, homochromous capitula (Fig. 5A, E) with yellow or green-yellow corollas. Capitulum diameter ranges from 8 to 15 mm (e.g.. D. cacalüfolium and D. kamaonense) and 7 to 8 cm (e.g.. D. falconeri and D. cataractarum). The recep- tacle is convex to hemispheric, glabrous or pubes- cent. In fruit, the base of the capitulum is some- times widely turbinate. owers are female with strap-like or narrow- ly elliptic to slightly obovate rays, generally ending in three or two teeth, sometimes entire (e.g., D. al- taicum). Disk flowers are hermaphrodite, actino- morphic, and narrowly funnel-shaped. Phyllaries are herbaceous to slightly papery at the base in some species (e.g., D. austriacum) and arranged in 2 or 3 rows, the outer being wider than the inner. In most species the phyllaries are clearly shorter than the ray flowers, but they can be almost equal or even longer than them (e.g., D. stenoglos- sum, D. pardalianches). Phyllaries are ovate-trian- gular, ovate-elliptic, or ovate-lanceolate to linear. The phyllary apex is usually acute, except in D. e where it bears a sessile gland (Fig. 4A— ^hyllary margins are entire, except in D. haus- pn "tii where they are slightly fimbriate. A group of species (D. orientale, D. plantagineum, D. hun- garicum, D. carpaticum, and D. columnae) have phyllaries with ciliate margins, bearing thin, stiff, acute, and equidistant trichomes (0.2-1.5 mm) (Fig. 5E-G). FRUITS Some species of the genus have dimorphic cyp- selae (heterocarpy), evident primarily by the ab- sence of a pappus in ray flowers. Cypselae without pappuses are also generally larger than those with a pappus in the same specimen (Fig. 6A, B). Het- erocarpy occurs in other genera of Asteraceae such as Senecio, Crepis, Erigeron, Leontodon, and Het- erotheca, as well as in other families, Caryophyl- aceae, Apiaceae, Poaceae (Zohary, 1950). These morphologic differences serve different functions. eterotheca latifolia, in which the type of di- morphism is similar to Doronicum, Venable and Levin (1985) suggested that the pappose i rw are dispersed away by wind, while epappose selae fall near the mother plant. This double dis- persal strategy implies the potential to colonize dif- ferent habitats (Venable & Levin, 1985; Tanowitz et al., 1987; Imbert et al., 1996). Although plants with incompletely developed pappuses on ray flo- p< rets are found in some species (e.g., D. carpaticum, 328 Annals of the Missouri Botanical Garden Figure 4. A-D. Doronicum gansuense (drawn from Rock 12192, E, as D. thibetanum). —A. Capitulum. —B. Phyllary. —C. Apex of phyllary. —D. Ray flower. E-H. Doronicum db iin е from Rock 12941, СН). — Capitulum. —F. Phyllary. С. indumentum of A base of phyllary y flow 329 Álvarez Fernández Volume 90, Number 3 2003 Doronicum (Asteraceae) illing 14441, B). —E. Capitulum. K). —A. Capitulum. —B. Phyllary. —C. In- H. Doronicum orientale (drawn from Wi , (drawn from Davis 14381 Figure 5. A-D. Doronicum cacaliifolium dumentum of phyllary. —D. Ray flower. E —H. Ray flower. —F. Phyllary. —G. Indumentum of phyllary. Annals of the Missouri Botanical Garden м 1 LI NN n „А т! Ом УУЛА ООА y E a BA * xy cy SELON Sos оез" 3 WA ue LIO Mes NN 8 v D ۶ = i-a D A Г ў | | A ~ = " pd ze Figure 6. A, B. Doronicum i e дй um (from Álvarez et n ме МА). —А. Cypsela from a гау flower. —B. Cypsela from a disk flower. —C la of Doronicum carpetanum subsp. diazii irn Álvarez et al. 924, MA). —D. Base of caducous pappus in mde cs stenoglossum (from еи Ww К Gilbert 1205, E). Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) D. carpetanum), dimorphism is generally either present or not in Doronicum. Such intermediate sit- uations were considered hybrids by Cavillier (1911). In the present taxonomic treatment, since hybrids were not confirmed, intermediates are in- cluded with the closest species typselae are cylindric to abends: have 10 lon- gitudinal ribs, and are black, brown. brown-red, o olive-green. When an indumentum is present, this occurs mainly on the ribs, except for D. stenoglos- sum, where trichomes are spread across the cypsela surface. In most species, the pappus consists of 2 or 3 rows of minutely scabrous, white or yellow- tinged capillary bristles. A single row of bristles is present in some southwestern Asian species (e.g. D. macrophyllum, D. maximum) as well as in D. stenoglossum. This latter species is functionally heterocarpic since the pappue in ray flowers is ca- ducous. A thick base, or the pappus falls (Fig. 6D). Surfaces of cypselae are warty to slightly warty “crown,” remains when (Fig. 7A). reticulate-grooved (Fig. 7B). grooved to slightly grooved (Fig. 7С). or smooth (Fig. 7D). INDUMENTUM This was the most relevant taxonomic character for Cavillier (1907, 1911) at the specific. level. Other authors also considered the type and ar- рет of indumentum very important (Pérez et al., 1994). While the indumentum pattern iden- fies certain species (e.g., D. clusii, D. glaciale, D. haussknechtii, D. hungaricum, D. dolichotri- chum), any quantitative variation within the same indumentum type is not useful in distinguishing species. The latter quantitative criterion led to the erroneous characterization of the species D. aus- 1997). When indumentum is scarce, it may be most dense on the upper third triacum (Pérez et al., of the stem, on veins, margins and abaxial surfac- es of leaves, as well as in the basal part of outer phyllaries (abaxial surfaces). In Doronicum, the adaxial surfaces of the phyllaries are always gla- brous. The following types of trichomes have been recognized in Doronicum: Eglandular trichomes (1) Multiseriate trichomes with blunt apices that are formed at least by two rows of rectangular cells, these trichomes are usually 0.3-0.5 mm long, but sometimes up to 1.5 mm (in D. plantagineum) or even up to 4.5 mm (in D. hungaricum). They occur on stems, leaves, and phyllaries, more frequently in European species but are almost absent in cen- tral Asian species. (2) Uniseriate trichomes with blunt apices that are formed by a single row of rectangular or square cells, these trichomes are 0.1—0.4 mm long. This is the most common pubescence, occurring on stems, leaves, and phyllaries in almost all species. In D. cacaliifolium (Fig. 5A—C phyllary surface and differ from others by their en- ) they cover the abaxial larged basal part and curved apex (Fig. 8B). This type of trichome also occurs in D. clusii, in which the cells are clearly rectangular, but the trichome length is up to 5 mm, and they may form an en- tangled covering on the leaf margins. ~ 3) Multiseriate trichomes with acute apices. These consist of at least two rows of fusiform cells ending in one or two cells with an acute apex. Four sub- types occur: (3a) generally of more than two rows of cells, charac- Stiff trichomes (0.5-2.5 mm long), consisting teristic of leaves and phyllaries in D. glaciale and D. clusii (Fig. 3C—H). (ЗЬ) Somewhat stiff trichomes (0.2-1.5 mm long). sometimes crooked (cilia). They are only present on the margins of phyllaries of D. orientale (Fig. 5E— G). D. plantagineum, D. hungaricum, D. carpati- cum, and D. columnae. (3c) Trichomes (ca. 0.3 mm long), formed by two or three cells. They occur on the cypselae of almost со z all Doronicum species (Fig. (3d) Trichomes (0.5-5 mm long) that end in two acute cells, and are present only on leaves and pet- ioles of D. pardalianches. Glandular trichomes The following types of glandular trichomes are recognized in this work: (1) Short-stalked glandular trichomes. These con- sist of 4 to 8 cells, 0.05-0.3 mm long, and are present on the stems, leaves, phyllaries, cypselae, and flowers of all the species. Two subtypes are distinguished: (la) diameter as the cells of the trichome stalk (Fig. 8C). Trichomes with apices (of 2 cells) of the same (1b) Trichomes with capitate apices (of 3 or 4 cells) and wider than the sta (2) Long-stalked glandular trichomes. These con- sist of more than 6 cells, 0.3—5 mm long, and are present on the stems, leaves, and phyllaries in many Doronicum species. Two subtypes are distin- guished: —. 2a) Trichomes with apices of 3 or 4 cells of the same diameter as stalk cells. — 2b) Trichomes with apices (of more than 4 cells) exceeding stalk cells (Fig. 8D). This latter type is present in a few species (D. macrophyllum, D. hun- 332 Annals of the Missouri Botanical Garden IA — A S ; < т 7 f a 7 j Mf, —— 3 Р | زک‎ =, x E / '/ 4 * , 2 э ж Saige Р 4 d Ms "n pr pL EE: es "p — ی‎ i d ps » = LC G E P Р А j E = E i - v »» i r ن‎ tvm x А E т / м. рз 7 Figure 7. Cypselae surfaces in Doronicum. —A. Doronicum pardalianches (from Almaraz Doronicum reticulatum (from Baytop & Baytop 20972, E, as ithynicum). —C el al. 1015, MA). —B. . Doronicum cacaliifolium (from Davis . —D. Doronicum dolichotrichum (from Davis & Polunin 24383, E). Volume 90, Number 3 Álvarez Fernández 333 Doronicum (Asteraceae) Fig —A. Uniseriate eglandular trichomes with acute apices on а cypsela of Doronicum maximum (from Davis РЧ P 24113, E). eg Uniseriate eglandular trichomes with blunt apices on a phyllary of Doronicum . —C. Uniseriate glandular trichomes on the margins of a pic of Doronicum cataractarum (from Fest 571, B). р, Apex of a multiseriate glandular trichome from the base of the capitulum in Doronicum kamaonense (from Polunin et al. 401, G, as D. roylei). cacaliifolium (from Davis 14551, 334 Annals of the Missouri Botanical Garden garicum, and D. kamaonense). Only in this last spe- cies is the apex markedly obconical and can be observed near the capitula without a magnifying glass. CHEMISTRY Several chemical analyses on Doronicum macro- phyllum (Bohlmann & Grenz, 1979), D. parda- lianches (Bohlmann & Abraham, 1979), D. hun- garicum (Bohlmann et al. 1980) and D. grandiflorum (Reynaud & Raynaud, 1984, 1986; Reynaud et al., 1985) have resulted in the isolation of 52 different compounds. Two of them are pyrrol- izidine alkaloids, a few are aromatic compounds: 10 benzofurane-derivatives, 4 phenols, and 2 flavo- noids. Twenty-one monoterpenes, mainly thymol-de- were also isolated. rivatives, The remaining com- pounds are 5 diterpenes with kaurane structure, 2 triterpenes (one of them a lupane-derivative and oth- er oleanane-derivative), and 6 sesquiterpenes, most of these germacradiene-derivative. Since only four species were analyzed, the tax- onomic usefulness of these chemical characters at the specific level cannot be assessed. However, the presence of alkaloids of the pyrrolizidine group (Bohlmann & Grenz, 1979) is consistent with the inclusion of the genus within the tribe Senecioneae, which is characterized by these alkaloids (Robins, 1977). CHROMOSOME NUMBERS Chromosome numbers in Doronicum are rather constant. Most reports included herein were ob- tained from the literature (Lindqvist, 1950; Skal- inska, 1950; Baksay, 1956; Contandriopoulos, ; Favarger & Huynh, 1964; Polatschek, 1966: Favarger & Kiipfer, 1968; Lovka et al., 1972; Kuz- 1973; г Kjellqvist, 1974; Garbari et al., 1980; 1980; Belaeva & Siplivinsky, 1981; Van Loon & Oudemans, 1982; Kuzmanov & Georgieva, 1983; Strid & Franzén, 1983; Davlianidze, 1985; Strid & Anderson, 1985: Chacón, 1987; Lippert & Heubl, 1988; Tasenkev- itch et al., 1989; Vir Jee & Kachroo, 1989; Baltis- 1991; Ruiz de Clavijo, 1993). Only when no original sources were available (i.e., D. manov & Ancev, Lóve Van Loon, berger, catar- actarum, D. macrophyllum, among other counts), data from indexes of plant chromosome numbers (Fedorov, 1969; Moore, 1982; Goldblatt, 1985, 1988; Goldblatt & Johnson, 1994, 1996, 1998) were cited, and they are indicated by an asterisk after the number. However, chromosome counts were made here for D. carpetanum subsp. diazii (2n 60) and D. carpetanum subsp. kuepferi (2n = 60). Material was cultivated from fresh rhizomes or seeds. Apices of secondary roots as well as imma- ture disk flowers were used for counting. Both were fixed in 3:1 ethanol: acetic acid for 48 hours and then kept in 70% ethanol at =20°С. Chromosomes were stained with acetic orcein. 30, char- group in Senecioneae, The basic chromosome number x acteristic of the “cacalioid” is also considered to be the basic number in Do- ronicum (Bremer, 1994), although previous authors suggested x = 10 (Fernandes & Queirós, 1971; Má- jovsky & Murín, 1987). known for Chromosome numbers, 19 taxa, are consistent with both hy- potheses, although the fact that most of them are RUDI E of 30 (2n number x — 60) may support the basic 30. Only D. carpetanum subsp. car- petanum, D. carpetanum subsp. pubescens, D. plan- tagineum, and several populations of D. clusii and D. pardalianches had higher counts, at 2л = 120 Other species show both ploidy levels (e.g., D. al- 30*, 60; D. macrophyllum, n = 30*, 60; and D. oblongifolium, 2n — 60, 40), suggesting that polyploidy is common in the genus. Doronicum oblongifolium (2n — — taicum 2n = 40) is the only count that is inconsistent with x — 30, and it should be recount- ed. PHYLOGENY The first phylogenetic hypothesis for Doronicum sect. Doronicastrum (i.e., D. altaicum, D. austriacum, D. balansae, D. briquetii, D. cacaliifolium, D. carpa- ticum, D. carpetanum, D. clusii, D. cordatum, D. cor- sicum, D. dolichotrichum, D. falconeri, D. glaciale, D. grandiflorum, D. haussknechtii, D. longifolium, D. macrolepis, D. macrophyllum, D. maximum, D. oblon- gifolium, D. orientale, D. pardalianches, D. plantagi- neum, D. portae, D. reticulatum, D. thibetanum, D. thirkei, D. turkestanicum, and D. sented by Cavillier (1911) Cavillier (1911 tions (sect. Doronicastrum, sect. Soulieastrum, and viscosum) was pre- mentioned that his three sec- чч” sect. Hookerastrum) were not closely related, and thus the genus so circumscribed is polyphyletic. Cavillier considered the Doronicum subsections included in section. Doronicastrum to be natural groups, and his classification was developed ac- cordingly. He thought that subsection Plantaginea (D. cong D. longifolium (= D. hungari- cum herein), D. oblongifolium, and D. falconeri) was monophyletic and the most ancient group in the genus. Doronicum subsect. Pardalianchia (D. pardalianches, D. roylei (= D. kamaonense herein), D. reticulatum, and D. atlanticum (= D. plantagi- neum herein)) was a grouping derived from subsec- Volume 90, Number 3 2003 Álvarez Fernández 335 Doronicum (Asteraceae) tion Plantaginea. Doronicum subsects. Cardiophyl- la (D. carpetanum, D. orientale, D. carpaticum, and D. cordatum (= D. columnae herein)) and Macro- phylla (D. macrophyllum, D. dolichotrichum, D haussknechtii, D. maximum, and D. cacalitifolium) were derived from subsection Pardalianchia, and were also natural groups. In contrast, Cavillier placed the probable origin of subsection Grandiflo- ra (D. altaicum, D. briquetii, D. grandiflorum, D. glaciale, and D. clusii) among several members of subsection Plantaginea. The monotypic subsec- tions Corsica and Austriaca would be derived from different members of subsection Pardalianchia. Cavillier (1911) noted a trenc from ancient plants bearing one capitulum and bas- Morphologically, al leaves with a truncate or attenuate base to evolved plants with several capitula and basal leaves with cordate bases. Biogeographically this is not easy to reconcile since these groups include members from different continents. The phylogenetic analysis of Alvarez Fernandez et al. (2001) was based on morphological evidence as well as molecular data (nuclear ribosomal ITS and chloroplast trnL-F sequences). When morphological characters were mapped onto the most parsimonious topologies, only three were free from homoplasy (scapiform habit, acrod- romous basal leaf venation, and ciliate involucral bracts). All of these characters are synapomorphic for the D. plantagineum clade, and the strict con- sensus can be seen in Figure There are only two clades with bootstrap support above 9096, both basal. The Doronicum clade had 100% bootstrap support in all analyses, using Lig- ularia and Tussilago as outgroups. Although the use of only two outgroups does not provide stringent conditions for testing the monophyly of the ingroup, other evidence also indicates the monophyly of Do- ronicum. In particular, sequences aligned well within Doronicum but were difficult to align with any other genera suggesting that Doronicum, as cir- cumscribed in this work, is monophyletic. The oth- er clade with strong bootstrap support (9796) in- cludes all the species of Doronicum except D. corsicum, which is sister to it (Fig. 9). This note- worthy result has strong biogeographic implica- The next taxon that is a derivative of this Corsican endemic is D. pardalianches, followed by tions. the D. plantagineum group, which has 85% of boot- strap support. All these species are European, mostly Mediterranean, suggesting that early diver- sification took place on the European continent, specifically within the Mediterranean Basin. TAXONOMY Doronicum L., Sp. Pl. 885. 1753. Pardalianches Tausch, Flora 11: 182. 1828. Doronicum sect. Doronicastrum Cavill., Annuaire Conserv. Bot. Genéve 13-14: 337. 1911. Doroni- cum subsect. Pardalianches [Pardalianchia| Cavill., Annuaire Conserv. Jard. Bot. G lianches (Tausch) Gorschk., Bobrov, Fl. URSS 26: 773. 1961. Doronicum ser. Pardalianches (Tausch) Gorse in Schischk. & Bobrov, Fl. URSS 26: 778. 1961. TYPE: Herb. Clifford, 411.1 [sine collector] (lectotype, designated by Llamas et al. in Jar- vis & Turland (1998: 360), BM!). Aronicum Arun Elem. Bot. 1: ee 1790, nom. inval. (ICBN . V, rw et al., ). саптап Саа Bull. S d Paris: 32. 1817. TY “Doronicum ea scorpii brachiata,” Herb. dede X: 16 [sine collector] үү ыл сех by Álvarez in Jarvis & Turland (1998: 353), UPS!, photograph). [7 Doronicum e L.] Fullartonia DC., Prodr. 5: 281. 1836. TYPE: “Comp. angl. des Indes 1830” [sine collector], ex herb. de Can- dolle (lectotype, designated by Álvarez Fernández (2001: 294), G i act daa [= Doronicum ka- maonense b ) Doronicum subse ral "Cavill. Jard. Bot. E 13-14: 997. cated; Annuaire Conserv. PE: not lo- Doronicum subse Jard. Bot. E 13-14: 338. Austriaca ie ) Gorschk., in Seh sok. & Bobrov, Fl. URSS 26: 774. 1961. TYPE: [sine collector] ex herb. dolest (LINN n.° dept 4!) (lectotype, des- ignated by Pérez et al. (1997: Doronicum po Cardiophylla ui 5‹ . Geneve 13—14: 338. 1911. Doroni- "Саон (Cavill. Gorschk., in Schischk. & Bobrov, Fl. URSS 26: 775. 1961. TYPE: not designated; although this subsection was umm ү а a type cana its name alid (ICBN, Art. 37.1, Greuter et al., ае subsect. Мас violis Cavill., Ae a. serv. s rd. Bot. Geneve 13-14: 338. 1911. Doroni- Annuaire Con- TYPE: North Caucasus. Beschtau, [F А. Е Mar- schall von Bieberstein s.n.], ex herb. Marschall von Bieberstein (lectotype, designated by Álvarez Fer- nández & Nieto Feliner (1999: 804), LE!). Doronicum subsect. Plantaginea Cavill., Annuaire Con- serv. Jard. Bot. Genève 13-14: 338. 1911. Doroni- c сЕ ser. Plantaginea (Cavill.) Gorschk., d Schischk. & Bobrov, Fl. URSS 26: 779. 1961. TYPE: Herb. Clifford, 411.2 [sine collector] ^ni авд Чы ы by Llamas et al., in Jarvis & Turland (1998: 360), BM! Doronicum subsect. Grandiflora Cavill., Annuaire Con- v. Jard. Bot. Geneve 13-14: 338. 1911. TYPE: “Arnica altaic. pall., tige simple unifl. haute de 4 ou 336 Annals of the Missouri Botanical Garden X D. altaicum X D. briquetii X central and central-eastern Asia X D. gansuense 9 southwestern Asia 51 X D. falconeri E Europe and northern Africa X D. kamaonense X D. stenoglossum B D. carpetanum subsp carpetanum Р Ш D. clusii E D. glaciale 5 I— m D. арап subsp diazi ———- W D. grandiflorum ———- W D. carpetanum subsp kuepferi B D. Selen subsp pubescen ө D. dolichotrichum 9 D. haussknechtii 9 D. cacaliüfolium 9 D. reticulatum € D. macrophyllum subsp. macrophyllum 62 53 — ® D. maximum | — 9 D. macrophyllum subsp. sparsipilosum € D. oblongifolium E D. austriacum 73 E D. hungaricum 100 E D. plantagineum Ш D. columnae € Ш D. orientale B D. pardalianches Ш D. corsicum outgroups Figure 9. Strict consensus from 228 most parsimonious trees oe from as combined ле of three data sets in а. (morphology, nrITS, cpDNA trnL-F sequence data, Alvarez Fernández et al., 2001). и values above 50% are shown above the branches. Outgroups: Tussilago pes " and edari кы (L.) Cass. Geographical distribution of taxa: central and central-eastern Asia (crosses), southwestern Asia (circles), and Europe and northern Africa (squares). Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) 7 pouces, fl. tres grande, fleurit au com. * de mai [sine collector], ex herb. Lamarck (lectotype. desig- nated Ьу Álvarez Fernández & Nieto Feliner (1999: 803), P-LA!, photograph). T um sect. а Cavill., Ann uaire Conserv. ard. Bot. Genève : 338. 1911. TYPE: China. Tibet: Kiala, Ton rie > A. Soulié n (lectotype, designated here, G!; isotype, K!), [= Doronicum sten- zlossum Maxim. Doronicum ser. Altaica ws ge in Schischk. & Bobrov, Fl. URSS 26: 768. TYPE: [Asia] "E Sumis montium altaicorum* a collector], ex herb. Pallas (lectotype, de и by Álvarez Fernández & Nieto Feliner (1999: 801), BM! Doronicum ser. Carpatica Gorschk. in ie hischk. & Bob- rov, Fl. URSS 26: 771. 196 YPE: not located; protologue citation: Fes eus in der alpinen am Szurul ii um subsect. /saurica J. В. Edm., Notes Roy. Bot. Edinburgh 37(1): 72. Я z: Turkey. Sane . Mt. Ghei-Dagh, [T. Heldreich| 1043 (lecto- type, designated by Alvarez Fernández & Nieto Fel- !) iner (1999: 802), G-BOIS!; isotype, B 10—150(+) ст tall. Rhizomes fleshy or woody, glabrous to pubescent, sometimes Perennial herbs, with buds. Stems terete, fistulous, slightly ribbed, erect, branched or not, scape-like to leafy, some- times with persistent leaf remains forming dark scales or fibers at the base, yellow to brown-tinged variable, eglandular с © when dry. Indumentum r glandular, abundant near the capitulum to very scarce at the base of the plant, sometimes absent. Leaves alternate, simple, entire to dentate, pubes- cent or glabrous, with actinodromous or pinnate- sometimes somewhat actinodromous venation, acrodromous. Basal leaves sometimes reduced to cataphylls or absent at flowering time, petiolate, with orbicular, ovate, elliptic or obovate blades, generally with a blunt apex; base of blade truncate, attenuate or cordate; petiole generally as long as the blade or longer, sometimes shorter. Lower cau- line leaves similar to basal leaves, sometimes ses- sile. Middle cauline leaves sessile, ovate, elliptic, obovate, or fiddle-shaped, semi-amplexicaul, with blunt or acute apex. Upper cauline leaves similar, sometimes bract-like. Capitula 1 to 20(+), heter- ogamous, arranged in cymose synflorescences, ter- minal when solitary, radiate, 0.8-8 ст diam. in- cluding rays. Involucre much shorter than rays or rarely exceeding them. Phyllaries arranged in 2 to 3 rows, similar, herbaceous to somewhat papery at base and margins, ovate-elliptic to obovate-elliptic, or linear, generally with acute to tapering apex; margins entire, sometimes ciliate or minutely fim- briate; pubescent on the abaxial surface, rarely gla- brous, indumentum absent on the adaxial surface. Receptacles convex, glabrous to pubescent. Flowers with yellow or green-yellow shaded corollas, uni- form in color. Style branches short and blunt, ad- axial papillate. Anthers without appendages. Ray flowers female, arranged in 1 row. Rays oblong-el- liptic to linear; apex with 2 or 3 teeth or acute: sometimes pubescent at the base. Disk flowers, tu- bular, bisexual. Cypselae homomorphic (all cypse- lae with pappus) or dimorphic (ray flower cypselae without pappus), cylindrical, elliptic-obovate, with 10 ribs, light green, or black. Surface smooth, grooved or warty, rown to brown, brown-red, olive glabrous, or pubescent. Pappus arranged in | to 3 rows, with minutely scabrous capillary bristles, white to yellow tints. Base chromosome number x Distribution. Doronicum is found mostly be- tween 25? and 55° longitude in Asia, Europe, and North. Africa (Morocco and northeastern Algeria). growing in forests, open rocky places with moist soil, meadows, and near watercourses, from sea lev- el up to 5000 m in elevation. Kkv TO SPECIES OF DORONICUM m- — “~~ £e < ‚ flowers without pappus Dens plant) ies E RE RR EN Ray flowers with pappus, alt though | some- times Doo developed (homocarpic plant) Base of blade of basal leaves truncate or at- tenuate; plants usually bearing a adi ca- BUE aah 2. Base of blade of basal leaves c ordate or sub- cordate; plants bearing one to several capit- ulli анаа 3a(2). Basal leaves with acrodromous venatior 3 Basal leaves with ae DE or actinodromous vena 4(3). Basal leaves oblong- “elliptic: indumentum of the adaxial surface of basa s generally consisting of eglandular bien es @-5 mm long), sometimes sparse ..... 17. D. hungaric um 2(1). => е [d mich (up to 5(3). sometimes persisting fibers a stem; margins of leaf blades slightly swollen, sometimes with white-tinted d neos tri- . D. oblongifolium gins of leaf blades flat, dab. ҮЗЕН D. falconeri e black, zie эме sur- ен cordate and margins entire to sub- entire 1. __ . D. pardalianches 338 Annals of the Missouri Botanical Garden 11(10). 12(10). 14(13). 14". 15(14). All cypselae brown, red-brown, or olive- e addis: subcor- = or truncate and margins entire or den- ate ез Ош phyllaries ciliate (cilia not glandular) Outer phyllaries not ciliate |... Rhizome woody to somewhat woody, и ). D. columnae Rhizome fleshy, with pubesc ent nodes _ 9 Basal leaves with ovate blades, re ordate truncate at base 24. А ит Basal leaves with reniform to sidus ovate blades, cordate to subcordate at base — 22. D. orientale Rhizome fleshy with trichomes on nodes, sometimes scarce and absent on the oldest nodes | " Rhizome woody to somewhat woody, glabrous short Upper part of stem and phyllaries glabrous or with short-stalked glandular tric hones (up to 0.3 mm), sometimes also with scattered eglandular trichomes (up to 0.4 mm); plants bearing more than 2 capitula; Ms gend. gla- brous . cataractarum Upper part of stem with TN stalked glan dular trichomes (up to 5 mm), sometimes also with eglandular trichomes and short- stalked glandular trichomes; plants bearing 1 to 6 capitula: receptacle glabrous or pu- escent ›. 12. carpetanum es with pinnate-actinodromous venation; em and base of phyllaries жик и glandular tri- gland markedly obconical; some bus s with adventitious roots and some- times branched near the bas 18. одеса with actinodromous venation: аберне nse Uppe r without P ERE A roots and branche di the upper part of stem Stem leafy (more than 6 cauline eaves); mid- sent at flowering time: receptac ‘le gene ‘rally pubescent 2. D. austriacum Stem not leafy (less than 6 A leaves): middle and upper cauline leaves generally shorter than the adjacent internodes: basal leaves and lower cauline тт leaves sometimes prn at flowering time; receptacle glabrous glabrate 14 | ин; with a single c apitulum (exe 'eptionally о 3); basal leaves dentate; pe us of basa ien es 0.5-2 mm wide Plants with two to several « leaves dentate to entire; ). D. columnae capi basal petiole basal more than 2 mm wide, sometimes with‏ ا sheath more than 3 cm long |...‏ " asal leaves with a reniform blade and den- long ә. 18(17). 18’. 19(16). 19’, 20(1). — 20'. 21(20). 21’. 22(21). 22'. 23(22). 23'. 24(23). tate margins; phyllaries covered with seri- ceous uniseriate eglandular trichomes (0.2— 0.4 mm); rhizome E covered with scarious remains or fibe ). cac aliifolium Basal leaves ovate to anne ovate with cor- date to subcordate base or reniform, margins dentate to entire; , pubes- cent (eglandular or Шанда), but "s seri- — x phyllaries glabrous ceous; rhizome not moniliform Upper part of stem with long-stalked glan- dular trichomes or pubescen Upper part of stem glabrous to glabrate or with very short white eglandula restricted to the base of capitula Phyllaries with subulate apex and dark-col- ored longitudinal veins 25. D. reticulatum Phyllaries with acute but r not subulate apex. sins not dark-colored r trichomes f 18 ane part of stem and upper cauline leaves with white multiseriate glandular and/or aul trichomes (0.5—3 mm), sometimes attered, sometimes also glandular D. dolic homie hum Upper ке of stem репе тай wandulerd ps without white multiseriate ко - che ). D. mac nam Base of capitula with short white eglandu trichomes (ca. 0.2 mm); margins of serae sometimes slightly fimbriate or E iar — 16. D. haussknechtii Base of сара кенне margins of phyl- aries entir 20 1отеѕ ә, < D. maximum Corollas with pale yellow to green tints; linear (0.5 rays; pappus consist capillary bristles, caducous (at least in ray ы lower stems ан with adven- lous roots ›. D. stenoglossum C qs yellow: ravs elliptic to шр llip- tic (1.2-4.5 mm wide); »hyllaries erect to patent or patent, generally shorter than rays; pappus with more than one row of trichomes, not caducous (sometimes poorly developed in ray flowers); lower stems without adventitious roots All cauline leaves sessile: plants bearin ig several capitula . 10. At least lower cauline leaves petiolate; ven »earing ). corsicum one to several capitula .. Rhizomes with tric homes on nodes, generally cove ‘red by scarious remains very short of basal leaves, these sometimes scarce 23 Rhizomes glabrous |... 27 Stems generally more than 50 cm; basal leaves with cordate to subcordate base; plants bearing several capitula; base of ca- pitula glabrous or with short-stalked glands, generally scattered _ T. Stems generally less ). cataractarum dium 50 cm; basa leaves with subc ordate, truncate or attenuate eglandular or ma stalked glandular Pars 24 andular, sometimes also with scattered ee trichomes —. 25 Leaf margins mainly Volume 90, Number 3 2003 Álvarez Fernández 339 Doronicum (Asteraceae) 24". Leaf margins with e trichomes, ese sometim N е thes es scare@ -a Pappus of ray flowers well developed, similar to pappus of disk flowers: cypselae a c isk flowers pubescent, sometimes with glan suis 15. D. eruditorum 25'. Pappus of ray flowers poorly developed: cy selae of disk flowers mainly glandular . D. carpetanum oh imu (ac ule, в i 25(24). mn) sometimes ч with glands, scattered (Fig. 3C—E) 20'. Leaf margins pubesc cent (tangled, hyaline short-stalked 8. D. clusii eglandular trich and glands (Fig. 3F-H) omes 27(22). Basal leaves with cordate base ------- sid tiende. 5. D. « carpets um 20. Mie leaves with truncate, attenuate, or sub- e ase а 28 28(27). Phyllaries E blunt. apex that bears a ses- sile gland |... 13. D. gansuense 28'. Phyllacies with acute apex, sibus a sessile gla ME ИНИНИ 29 29(28). Base of capitula with Tong- stalked glandular trichomes (1—5 mm); plants мше опе са- ришат __. 2. Base of ca apitula. gli abrous or p (e 1- stalked glandular trichomes (up to 1.5 mm). scattered; plants bearing 1 to 4 capitula . D. altaicum . briquetii 1. Doronicum altaicum Pall., Acta Acad. Sci. Imp. Petrop. 2: 271, tab. 16. 1779. Aronicum altaicum (Pall.) DC., Prodr. 6: 320. 1838. Ar- nica altaica (Pall.) Turez., Bull. Soc. Imp. Na- 1838. TYPE: [Asia] [sine collec- turalistes Moscou 1: 95. “E Sumis montium altaicorum” tor], ex herb. Pallas (lectotype, designated by Álvarez Fernández & Nieto Feliner (1999: 801), BM). Plant up to 80 cm tall. Rhizomes somewhat woody, glabrous, generally with scaly leaf remains. dumentum of short-stalked glandular trichomes and eglandular trichomes (up to 1 mm), more abundant near the capitula, sometimes glabrate to glabrous. Leaves entire to slightly dentate. Basal leaves gen- erally absent at flowering time; blade 5-8(10) X 2.5-3 cm, ovate, elliptic or obovate, with attenuate base, and blunt or subacute apex, with actinodrom- ous to pinnate-actinodromous venation; petiole (2)7-10(27) ст long, 3—4(7) mm wide. Lower and middle cauline leaves (2.5)3-8(11.2) X (1)3—5(6 cm, similar to basal leaves or sessile, elliptic to — obovate, sometimes widely ovate to suborbiculate, semi-amplexicaul, with blunt apex. Upper cauline leaves 2.5—7(8.5) X (0.5)0.2—4(5 middle cauline leaves or ovate-lanceolate and with .5) em, similar to subacute apex. Indumentum similar to the adjacent part of stem, sometimes glabrous. Capitula 1 to 4; (2.5)4—5 em diam. including rays: involucre shorter than rays, (3)3.5—4.5 ст diam. Phyllaries herba- ceous, (0.7)1—1.2(1.7) em long, 1.2-2 ovate-lanceolate to subulate. Indumentum similar mm міс €. to the upper part of stem, sometimes glabrate. Re- nr les glabrous. Flowers with yellow corollas. Ray flower corollas 1.2-2.1 ст long, (1.5)2-3.5 wide, obovate-elliptic, apex with 2 or 3 teeth, some- o mm times toothless, acute or blunt. Disk flower corollas 4—5.5 X 2-2.5 mm. Cypselae dark brown, with smooth or slightly grooved surface, homomorphic, ca. 2.8 X 1 mm, generally glabrous. sometimes with scattered eglandular trichomes or glands. Pap- 5 mm, white to brown yellow-tinted. 60 (Belaeva & Splivinsky, 1981, as D. bargusiense Serg.; *Gold- pus up to Chromosome number, 2n = 30*, blatt & Johnson, 1998, see comments below). Illustrations. tab. 16); Figures 7, 2G 10 Pallas (1779: B Central Asia (Turkistan and Altai region to lake Baikal). Woods, meadows, and near watercourses, altitudi 1300-3400 m (Fig. 11). In central Asia there are some species morpho- Distribution. logically similar to Doronicum altaicum (i.e., D. bri- quetii, D. falconeri, and D. gansuense). All of them have the same habit (solitary capitulum and mostly leafy stems with uniform leaves). but only one of these, D. falconeri, overlaps its area of distribution with D. altaicum. The character used to distinguish between these two species is the presence of a pap- pus in the ray flowers of D. altaicum versus its absence in D. falconeri. The remaining similar spe- cies are differentiated based on the indumentum (long trichomes on the base of capitulum in D. bri- quetii vs. short trichomes in D. altaicum (Fig. 10)). and on the apex shape of phyllaries. which is blunt (due to a sessile gland) in D. gansuense (Fig. 4B. C) and acute (lacking the sessile gland) in D. al- taicum The citation of the chromosome number 2n = 30 for Doronicum altaicum was found in Goldblatt and Johnson's (1998) index, but the original source for this data was not seen. Selected err ee examined. KAZAKHSTAN. Lepsoi i Tentekom, 11 Sep. 1931, Enden s.n. (LE): Belogore, Tuk- schinskoe, 15 fra 1949, Fedorov et al. s.n. (LE); Alatau, Richter 2772 (GH). RUSSIA. Altay: montes del Altai, Ayu-Kel, río Baijs, Castroviejo & Valdés Bermejo 14032 (MA): Tigeretzkyi, 31 July 1891, Krylov s.n. (5); Tomskaya, 340 Annals of the Missouri Botanical Garden Figure 10. A, B. Doronicum altaicum (drawn from Krasnoborov et al. 959, К). —A. Capitulum. —B. Indumentum of the base of capitulum. C, D. Doronicum briquetii (drawn from Rock 22380, E). —C. Capitulum. —D. Indumentum of the base of capitulum. Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) 60 80 100 120 60H + + . + T Wm & ее е See e }, $ e e XE " К A “з. „А. y T / P d l Teen. \ P i 60 80 100 120 Figure 11. Bijskij, uri ‘i July 1913, ee s.n. (LE); Elek- , Katuni, 12 July 1947, е а di culina s.n. (LE) Жаша ма, Y o 192: = E = 2 = 5 к= | Л. =: LE). Chita: the ea Smirnov 314 405) (LE); Yangi, Belogore, Tukshinskoe, Pedot rov et ر‎ ) July is uraginskij, Mt. Moskva, 22 Send rin s.n. (L 1962, Krasnoboroe & Ershova s.n. (LE); Sayan, Alan, Taskalik Dzhoya i Klaya, Krasnoborov 8179 (LE); INN Ara- дало. 8 June 1892, Krylov (LE); Enisejskaya gub., Kanskij, Kuznetsov 927 (LE); Sidi, Sisima i Mani Eniseya, Volkov 185 (LE). Tuva: Sayan, Sayanskij, Aldi- Ishkina, pec 8172 (LE); ү Tajhinskij. зај an, ибо! Kara-Chol’, 959 (K); Pij-K Khe ski Sayan, Lomonosova & Shau E ei e toe g- Khol, Ulug-Taiga, 3 $ Vidrina s.n. LE). “West Siberia: Tomskaya, Fe Sek 15 June 1909, Iljin s.n. (LE). 2. Doronicum austriacum Jacq., Fl. Austriac. 2: 18, tab. 130. 1774. Arnica austriaca (Jacq.) Hoppe, in Sturm, Deutschl. Fl. 10: 16. 1814. TYPE: [sine collector] ex herb. Linnaeus (LINN n.? 1002.4!) (lectotype, designated by Pérez et al. (1997: 3)). Distribution map for: Doronicum altaicum (®); Doronicum briquetii (А); Doronicum kamaonense (© — Plant up to 150(+) em tall. Rhizomes woody to somewhat woody, glabrous, and generally without scaly or fibrous leaf remains. Stems generally branched in the upper part, leafy, internodes gen- erally shorter than the adjacent leaves. Indumen- tum of glandular trichomes, uniseriate and multis- 2 mm). sometimes only eglandular trichomes, sometimes glabrate, more abundant near the capitula. Leaves eriate eglandular trichomes (ир to entire to slightly dentate. Basal leaves absent at flowering time, petiolate, ovate to orbicular, with cordate to subcordate base and blunt apex, with actinodromous venation. Lower and middle cauline leaves 6.5-19 X 4.5-12.5 cm, similar to basal leaves or sessile, fiddle-shaped, semi-amplexicaul. Upper cauline leaves 2.5-13 X 0.7—5 ст, lanceolate, generally with acute apex. Indumentum similar to the adjacent part of the stem. Capitula (1)2 to 16; 3-7 em diam. including rays; involucre generally shorter than rays, 1.5—3.5 cm diam.; pe- 16 ст long, 0.5-2 mm diam. Phylla- ries herbaceous, sometimes slightly papery at the ovate- duncles 1.5- base or at the margins, ovate-subulate, generally with acute apex; the outer 0.5-1.8 cm long, 1.2-4 mm wide; the inner 0.6-1.4 cm long, 0.7-3 mm wide. Indumentum of glandular and eglandular tri- chomes, sometimes glabrous. Receptacles pubes- cent, rarely glabrate. Flowers with yellow corollas. Ray flower corollas (1.2)1.5—3.5 cm long, (1)2—4 342 Annals of the Missouri Botanical Garden mm wide, oblong-elliptic to obovate-elliptic; apex generally with 3 teeth. Disk flower corollas 4.5—5.5 X 1-1.5 mm. Cypselae brown-tinted to olive-green with grooved-reticulate m dimorphic. Cypse- lae of ray flowers 2-3.5 X 0.7-1.3 mm, glabrous or glabrate, without pappus. Cypselae of disk flow- ers 1.5-3 X 0,7—1 mm, pubescent, with white pap- 60 (Skal- 6; Kuzmanov & Ancev, 1983) jus 3—6 mm. Chromosome number 2n = inska, 1950; Baksay, 195 1973; Strid & Franzén, —— Illustrations. Jacquin (1774: tab. 130); Hegi (1928: 713, fig. 421); Sávulescu (1964: pl. 189, fig. 2); Bolòs & Vigo (1995: 839); Figures ІС, 12A-D. Distribution. Europe (Carpathians, Balkans, Alps, Apennines, and eastern Pyrenees). Cultivated and naturalized at least in Great Britain. Growing in forest, meadows, near watercourses, and in moist rocky places, altitude 300-2200 m (Fig. 13). Doronicum austriacum is a variable species with regard to phyllary shape, number of capitula and, in particular, type and abundance of indumentum. Based on the protologue, where there is no mention of the presence of glandular trichomes, Pérez et al. (1997) characterized it as a non-glandulose species and accordingly, they chose a lectotype with no glandular trichomes. Although their lectotypifica- tion is technically correct, the distinguishing char- acters given for this species are erroneous since many populations from Greece are glandular. Ac- cording to Pérez et al. (1997), these glandular pop- ulations might be included in Doronicum carpetan- um, a different species as recognized here. Although both D. austriacum and D. carpetanum are similar, they basically differ in type of rhizome (fleshy and scarcely pubescent in D. carpetanum vs. woody to somewhat woody and glabrous in D. aus- triacum), but not in type of indumentum, which can be glandular in both species. These two species do not overlap their areas of distribution. In addition, there are two other species morpho- logically similar to Doronicum austriacum (i.e., D. cataractarum and D. pardalianches) that may over- lap their distributions with D. austriacum. The characters to distinguish between those species and D. austriacum are the type of rhizome (fleshy rhi- zomes in D. pardalianches vs. woody to somewhat woody in D. austriacum), the fruit color at maturity (black cypselae in D. pardalianches vs. brown-tint- ed to olive-green in D. austriacum), the indumen- tum on the base of capitulum (scarce and short (up to 0.4 mm long) to glabrate in D. cataractarum vs. pubescent (glandular or not) in D. austriacum (Fig. A, К, F)), and the heterocarpy in D. austriacum versus homocarpy (sometimes pappus poorly de- veloped) in D. cataractarum. Selected specimens examined. ALBANIA. Korçë: Os- trovicé, Moskopolé, Alston & pen 2119 (BM, K). AN- DORRA. Pla de Sorteny, 7 Aug. 1948, Losa & Montserrat s.n. (BCF). AUSTRIA nr Plócken pass, 4 Aug. 1972, oid s.n. (B ie а h: Lackenhof, 19 July 1933, Cufodontis s.n. erósterreich: Ob- erar eon usee, Vachsteingebiet, Basc а H531 (В). Salz- burg: Boden der Alpen Lofers, Spitzel 972 (B, NY). Steiermark: valle Se 'haftal prope urbem Graz, a 1907, Fritsch s.n. ‚ E). Tirol: Zillertaler Alps, Gerlostal- Zillertal, Kramer 1366 (NY). ре ARIA. Blagoevgrad: Mt. Pirin, lago een = July 1993, Carrasco, Burgaz & Martín- ges '0 s.n. (MACB). En Rila ae in valle ah Se Mire & Bergmann = А 0). irae Pamprovo, Smolyan, у & к Wood Brey (BM). Veliko Türnovo: Béla Cherkva, July 1909, Stribrny s.n. (К). CZECHOSLOVAKIA. Tatra Mag- na, V Studená, 14 July MO). FRANCE. Cantal: UOmbriére, commune de Raulhac, Puyfol 4925 (BM, NY). Loire: Pilat, route de Pélussin au Crét de l'Oeillon, Barbezat 1834 (K). Pyrénées-Orien- tales: Cerdagne, massif du Carlite à Matanégre, Sennen 3968 (BC, MA). GERMAN Hirschkamm, ; . Wermsdorf, 12 Аш 1878, Oborny s.n. (К). ر‎ : Pisoderion, Alston & Sandwith 441 (BM, К); K Polunin 8268 (E); Kozani, Mt. Pieria, vu Kataphygion, Rechinger 17881 (В, ( HUNGARY. Szepes, Zips, Hohern-Tatra, Креон. bachtal, Ў Aug. 1911, Nydrddy s.n. (MA). ITALY. Tos- Etruria, Firenze, As 'olungo, 25 Aug. 1917, Fiori s.n. (В, BM, К). POLAND. Olkusz: Klucze, Biala, 31 May 1971, Frey & Sztyler s.n. (LE, NY). ROMANIA. Gorj: Oltenia, valle Gilort, Rinca, 28 June 1972, D. & M. Cîrtu s.n. (B, BM, MA, NY). UKRAINE. Ivano-Frankovskaja: m Chernij Cheremosh, Popadinets, бейтап et al. 1650 (LE). Zakarpatskaja: Veritskij, 19 June 1950, Igoshina s.n. (LE). YUGOSLAVIA, Crna Gora: between Ramna and Banasnic ‘ay i 328 (BM, E). ska: Gerovo, F & S. G. Gardner 2586 (ВМ); Z aka gora circa a May 1874, Vukotinović s.n. (ZA). Makedonija: Gebiet der Mala Rupa, Biesalski 469 (B); Goleinica-planina, va, Bornmüller 4269 (B). Slovenija: Pekel-Schluchbi Ohonica südlich Borovnica, südwestlich von Lubljana, 28 May 1966, Lippert s.n. (LE) 4 075 ~ — od 25, Suza s.n. Annuaire Con- Bot. Genéve 10: 197. 1907. TYPE: ndia. Kumaun, near the Kalam glacier, J. Ё Duthie 3066 (lectotype, designated by Álvarez Fernández & Nieto Feliner (1999: 802), G!). to 80 tall. woody, glabrous, generally with scaly leaf remains. Stems not branched, 3. Doronicum briquetii Cavill., serv. Jard. Plant up cm Rhizomes somewhat leafy completely up entire stem, internodes generally shorter than adjacent leaves. Indumentum of short- and long-stalked glandular trichomes (1—5 mm), more abundant near the rarely only eglandular trichomes. Leaves entire to slightly dentate. Basal leaves gen- capitula, erally absent at flowering time: blade (1.2)2—4 х ( — ).9)1—2(3) ст wide, ovate, elliptic or obovate, with Volume 90, Number 3 Álvarez Fernández 343 2003 Doronicum (Asteraceae) Figure 12. A-D. Doronicum austriacum (drawn from Strid et al. 18585, B). —A. Capitulum. —B. Phyllary. С. Indumentum of a phyllary. —D. Ray flower. E, F. Doronicum cataractarum (drawn from Hópflinger s.n., BM). —E. Capitulum. —F. Base of the capitulum. Annals of the Missouri Botanical Garden Figure 13. attenuate base, and blunt or subacute apex, with aclinodromous to pinnate-actinodromous venation; petiole (0.9)1.5—3(5) ст long, (1)2—3(5) mm wide. 2.3) x .5(6.2) cm, sessile, ovate-elliptic to ob- Lower x middle cauline leaves (3)4—9( (0.5)1.5— жен semi-amplexicaul. Upper cauline leaves 2—4(5.5) X (0.3)1—2(3) em, similar to middle cauline leaves or ovate. Indumentum similar to the adjacent part of stem, sometimes with glandular margins. Capitula solitary, (4)5—7(8) em diam. in- cluding rays; involucre shorter than rays, rarely equaling them, (3)3.5—4.5(5) cm diam. Phyllaries herbaceous; the outer (1)1.5-2(2.5) em long, 1.5- 1-2 ст long, 1-2(3.5 mm wide, ovate-lanceolate to subulate. Indumen- —. м 2.5(3) mm wide; the inner tum similar to the upper part of stem, more abun- dant at the base. Receptacles glabrous, rarely with scattered short-stalked glandular trichomes. Flow- ers with yellow corollas. Ray flower corollas (1.8)2— 3 em long, 2-3(3.5) apex generally with 2 or 3 teeth. Disk flower co- rollas 3.5-5 X homomorphic, 1.5-3 X 1 mm, glabrous, sometimes mm wide, obovate-elliptic. 1.5-2 mm. Cypselae dark brown. with scattered eglandular trichomes or glands. Pap- pus up to 6 mm, white to yellow-tinted. Chromo- some number unknown. Illustrations. Figures 2E, F, 10C, D. Distribution. Central and southern China (prov- inces of Sichuan, Tibet-Qinghai, and Yunnan), and the Himalayas. Open moist rocky places and woods, altitude 3000-5000 m (Fig. 11). As already discussed (see comments for D. al- Distribution map for Doronicum austriacum. taicum), there is a group of central Asian species morphologically very similar to each other. Doron- icum briquetii is included in this group, but its dis- tribution may only overlap that of D. falconeri. The character used to distinguish between them is the heterocarpy in D. falconeri versus the homocarpy in D. briquetii. There are other species outside this morphological group. D. kamaonense and D. sten- oglossum, that also overlap part of their area of dis- tribution with D. briquetii. But D. stenoglossum is quite different from D. briquetii in noticeable char- acters (i.e., numer of capitula, color, shape and size of ray flowers, and shape and size of phyllaries) as well as D. kamaonense (i.e.. number and size of capitula and type of indumentum). (See also com- ments under D. gansuense. Selected specimens examined. CHINA. Sichuan: Mt. Konka, Risonquemba, К nka ling, Rock 16834 (E, K, GH, ‚ №); Dun 3270 (S, UPS); Sikang. angi ing, T: ; Tapaoshan, Smith 11473 (S, UPS). Tibet- Qinghai: bnc dus n divide behind Tzekon, Forrest 666 (E); Bei-lua Shan, p e ene eas divide, Forrest 13164 (E); Oika-gur-pu, Mekong-Salween divide, Sarong, Forrest 14526 (BM, E, K, W); Dü Chu valley, E > Pashö, Kham, Hanbury-Tracy 22 (BM); D l. a, King- don Ward 5866 (E); Sobbé La, ۰ کک‎ 5 (BM); Tha Chu valley, Kingdon Ward 19592 (BM, S UPS); 2 a as valley, Lusha Chu, ot et al. 4752 (BM, S); Kongbo, Tsangpo valley, Pero La, Ludlow et al. ps (BM, E. e e ‘Tsangp зо valley, shong La, Ludlow et al. 5258 (BM, E, UPS): Paus La, near Paka, Ludlow et al. 5870 (BM. E, UPS); above atl Dzong, Pome, Ludlow et P 13148 (BM, E, UPS Pasum Chu, Kongbo. Ludlow et al. 13955 BM. E der Budi Tsepo La, Kongbo. Ludlow et Bi 15261 M UPS): Kongbo, Nyoto Sama, Ludlow et al. 15604 (BM, Do- Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) 345 Figure 14. ©). E); Tsarung, Wuli-la, Rock 22380 (GH, NY); Balti, Thále La to Bagmaharál, Schlagintweit 5962 (GH). Yunnan: Chien Chuan & Hsi valleys, Forrest 7662 (E): d plateau mountains, Forrest 13009 (E); Mekong—Salwee divide, Forrest 14413 (E, K); Mekong-Yangtze divide, A. wa, Forrest 25695 (E, G, К); Wei-Hsi area, Forrest 30434 (BM, E); inter fluvios Lu-djiang Salween et Djiou-djiang. in jugi Tschiangschel, Handel-Mazzetti 1765 (W); A-tun- tse, 3660 m, Kingdon Ward (E); Lichiang, McLaren 167D (BM); Mt. Habashan, Ndaku, Likiang range. drainage basin, Rock 9681 (E, GH, К); Hung-po, Tung-chu-ling, Rock 22891 (BM, E, GH, К, MO, NY): Dokerla, A-tun-tze, Wang 64905 (ЄН); MckoneSal- ween divide, Sila, Yii 22268 (E, GH). INDIA. Uttar Pra desh: Tilvie-Garheval, Chimpul opposite Bandarpunc h Duthie 849 (G, K). Yangtze 4. Doronicum cacaliifolium Boiss. & Heldr., Diagn. Pl. Orient. ser. 1, 11: 31. 1849. TYPE: Turkey. Taurus, Mt. Ghei-Dagh, [7: Heldreich] 1043 (lectotype, designated by Álvarez Fer- nández & Nieto Feliner (1999: 802), G-BOIS!; isotype, BM!). Plant up to 50(+) cm tall. Rhizomes woody to somewhat woody, glabrous, moniliform, sometimes with fibrous leaf remains. Stems branched in the upper part, leaves mainly distributed in the lower middle, upper internodes generally longer than the adjacent leaves. Indumentum of uniseriate eglan- dular trichomes (0.2-0.4 mm), sometimes with short-stalked glandular trichomes and a few mul- tiseriate eglandular trichomes, more abundant near the capitula and sometimes glabrous at the base. Leaves dentate. Basal leaves generally present at flowering time; blade 5.5-8 X 5-9 cm, orbicular to suborbicular, with cordate base and blunt apex, mountains of Distribution map for: Doronicum cacaliifolium (®): Doronicum dolichotrichum (А); Doronicum reticulatum with actinodromous venation; petiole 8-19.5 cm long, 2-2.3 mm wide. Lower cauline leaves with blade ca. 6 X 8 cm; petiole ca. 6.5 cm long, 1.5 mm wide, similar to basal leaves. Middle and upper 1.2-5.5 cm, sessile, fiddle- shaped, semi-amplexicaul, the upper leaves ovate cauline leaves 3—4 X to bract-like. Indumentum similar to the adjacent part of the stem or glabrate. Capitula 2 to 13; ca. 3.5 em diam. including rays; involucre shorter than rays, 1-2 ст diam. Phyllaries herbaceous, 6—7 X 2.5 mm, ovate-elliptic to ovate-subulate, generally with acute apex. Indumentum of uniseriate eglan- dular trichomes (0.2—0.4 mm), sericeous, abundant. Receptacles glabrous or glabrate. Flowers with yel- low corollas. Ray flower corollas ca. 1.7 cm long, 2 mm wide, oblong-elliptic, apex generally with 3 teeth. Disk flower corollas ca. 5 mm long. Cypselae brown-tinted, with slightly reticulate surface, di- morphic. Cypselae from ray flowers 2-3 X 0.9-1.2 mm, glabrous or glabrate, without pappus. Cypselae from disk flowers 2-2.7 X mm, pubescent (eglandular trichomes), with pappus white-tinted, 3—4 mm. Chromosome number unknown. Figures 5A-D, 7C, 8B. Illustrations. Distribution. Southern Turkey (Antalya and Konya provinces). Growing in shady rocky places, elevation 1800-2300 m (Fig. 14) Most of the Doronicum species from Turkey are similar morphologically (rhizomes woody to some- what woody and glabrous, several heterocarpic ca- pitula, and a few but very large leaves). These spe- cies have a few diagnostic characters; sometimes only one of these is consistent, making species Annals of the Missouri Botanical Garden identification difficult. Doronicum cacaliifolium is one of the best delimited species within this mor- phological group, and is also the one that has the most restricted area of distribution. The distinctive characters are its exclusive type of rhizome (mo- niliform) and the shape, size, and type of indumen- tum of phyllaries (Figs. 5A-C, 8B). the closest species within its morphological group is D. reticulatum from western Turkey (Figs. 14, 26), but this one has different shape, size, type of indumentum, and color of phyllaries (Fig. 26Е-С). Outside this morphological group, only D. orientale Geographically, may overlap its area of distribution with D. caca- liifolium (Figs. 14, 27), characters to distinguish between them. In addition to the difference in their phyllary characters (Fig. 5A-C, E-G), rhizomes from both species are quite but there are noticeable different (fleshy and pubescent in D. orientale vs. woody to somewhat woody, glabrous, and monili- form in D. cacaliifolium). Selected Pre examined. TURKEY. Antalya: Ak dag, Davis 14381 (E, G, K, МО, W); Ak dag. Davis 14551 (K). Konya: Hadim-Alanya, оног 5581 (GAZ 5. Doronicum carpaticum (Griseb. & А. Schenk) Nyman, Syll. Fl. Eur. Suppl.: 1. 1865. Aronicum scorpioides var. carpaticum Griseb. A. Schenk, Arch. 342. 1852. dium um carpaticum (Griseb. & А. Schenk) Schur, Verh. Mitth. Siebenbürg. Ver- eins Naturwiss. Hermannstadt 10: 137. 1859, Doronicum grandiflorum subsp. carpaticum Schenk) Rouy, Rev. Bot. Syst. ( Bot. 1: 53. 1903. TYPE: not located; protologue citation: "Sie- benbürgen: in der alpinen Region der südlich- en Karpaten, z. B. am Szurul (Fuss), Schur bis 7000" austeigend." Naturgesch. 18: as “carpathicum.” (Griseb. „€ PT. nach Plant up to 50 em tall. Rhizomes woody to some- what woody, glabrous, sometimes with leaf remains forming dark scales on nodes. Stems not branched, generally scape-like. Indumentum of uniseriate, multiseriate, and glandular trichomes, scattered, glabrous in the lower part. Leaves dentate to slightly dentate. Basal leaves generally present at flowering time; blade 2-4 ovate with cordate base and with blunt or subacute 2—4 cm, orbicular to broadly apex, with an actinodromous venation that some- times tends to be acrodromous; petiole 4-11 em long, 0.5-1 mm wide. Lower and middle cauline leaves 3—6 X 1.4—3 mm, similar to basal leaves or Upper sessile, fiddle-shaped, semi-amplexicaul. 3 cauline leaves 1.5— 2.5 cm, ovate-elliptic to ovate-lanceolate, sometimes bract-like. Indu- mentum of white-tinted scattered multiseriate eglandular trichomes (up to 1.5 mm), uniseriate eglandular trichomes mainly on the edge of the blade, and scarce short-stalked glandular tri- chomes. Сарша solitary, 3.5—5 ст diam. includ- ing rays; involucre shorter than rays, 2.5-3 cm diam. Ранее herbaceous, 0.9-1.2 cm long, 1.5— 4 mm wide, ovate-subulate; margins sometimes cil- iate, with acute, stiff and equidistant multiseriate eglandular trichomes (up to 0.6 mm long). Indu- mentum mainly glandular, and sometimes with un- iseriate eglandular trichomes. Receptacles glabrous o almost glabrous. Flowers with yellow corollas. Ray flower corollas 1.4-2.2 ст long, 2-2.8 mm wide, oblong-elliptic, apex generally with 3 teeth. Disk flower corollas up to 4 mm long. Cypselae (not seen at maturity), brown-tinted, homomorphic, 1.5 X 0.8 chomes) ^ mm, scarcely pubescent (eglandular tri- glabrate. Cypselae from ray flowers sometimes with a poorly developed pappus. Pappus up to 3.5 mm, white-tinted. Chromosome number 2n = 60 (Tasenkevitch et al., 1989). Illustrations. Săvulescu (1964: pl. 98, fig. 3). Distribution. Europe (Carpathians). Meadows, shady rocky places, and near watercourses, altitude 1200-2200 m (Fig. 15). The type material of Aronicum scorpioides DC. var. carpaticum Griseb. & A. Schenk could not be found, and although the protologue matches the di- agnostic features of this taxa, its identity here is tentative and the formal synonymy needs to wait until clarification. and D. closely related species, which differ only by their Doronicum carpaticum columnae are heteromorphic versus homomorphic fruits, respec- tively. Simonkai (1886) described а new species, including specimens without well-developed pap- pus: Aronicum barcense. Later, this name was con- sidered by Cavillier (1911) to be a hybrid species (D. carpaticum X D. columnae). | cannot confirm the hybrid origin for this taxon and therefore in- clude it as a synonym of D. carpaticum due to the presence of pappus, although poorly developed. Selected. specimens jp Kalofer, Wagner 77 (G). HUN Beszteroze-Na- szód: monte Clisia ad Rodnam, p Aug. 1902, Degen s.n. (LE). ROMANIA. Maramures: paur Hagy Pietrosz, 5 July 1907, Filarszky & Jávorka s.n. duc Mt. Ceahlau, ‚11 July 1922, е u s.n. (BM, С, К, О, 5). Fogaras: Bilea Cascada, 17 Aug. 1960, Liebenow s.n. (B); Transsilvaniae alpes Arpasenses, Vurtop, 9 Aug. 1883, Simkovics s.n. (B); Vertop et Vertopul, Simonkai 1816 (B, BM, С, K, S); circa lacum Bila, Arpás, 23 July 1914, Tuzson s.n. (BT ККА Frankovskaja: Verkhovinskij r-n, BU LGARIA. Plovdiv: ARY r. Chernij Cheremosh, Volume 90, Number 3 2003 Álvarez Fernández 347 Doronicum (Asteraceae) Figure 15. nae (9). ur. Popadinets, Geltman et al. 1636 (LE); Verkhovinskij r- n, 30-33 km k Yuyuz ot Verkhovini g. Chivchin, бейтап et al. 1880 (LE); Karpati khr. Chernogora, 7 July 1964, Y dpi s.n. (LE). Stanislavskaja: Mt. Goberla, Rakhov- skij 13 July 1958, Fodor s.n. (LE). Zakarpatskaja: Ha khovekij r-n, 27 July 1976, Borodina et al. s.n. (LE) YUGOSLAVIA. Srbija: Mt. Gnila greda supra vallem Dobrido dispersum, prope Trebinje, Aug. 1891, Vandas s.n. (K). 6. Doronicum carpetanum Boiss. & Reut. ex Willk. & Lange, Prodr. Fl. Hispan. 2: 108. 1870. Doronicum plantagineum subsp. carpe- . & Reut. ex bor. Rouy, Rev. 34. 1903. TYPE: Spain. jar i sierra de Guadarrama, Рейа Lara, July 1858 [Р E. Boissier ѕ.п. | (lectotype, designated by Chacón (1987: 267). COI- WILLK!). Plant up to 120 cm tall. Rhizomes fleshy, with shining white-tinted short trichomes on nodes, sometimes with buds, and generally with leaf re- mains. Stems branched in the upper part or simple. Indumentum of multiseriate eglandular trichomes (up to 1.5 mm) and short-stalked glandular tri- chomes at the middle part of stem, and also long- stalked glandular trichomes (up to 5 mm) at the upper part of the stem, sometimes glabrous at the base, sometimes mainly glandular, more abundant near the capitula. Leaves entire to slightly dentate. Distribution map for: Doronicum carpaticum (©); Doronicum cataractarum (А); Doronicum colum- Basal leaves generally absent at flowering time; blade 2.5-9 subcordate to truncate base and blunt or subacute X 2—7 cm, ovate to orbicular, with petiole (1.5)3.5-13.5 cm long, (0.5)1-3 mm wide. Lower and middle cauline leaves 3—11(15) X 1.5-7.5(10) cm, similar to basal leaves or sessile, fiddle-shaped, apex. wit actinodromous | venation; sometimes ovate, semi-amplexicaul. Upper cauline leaves 1-7.5(9) X 0.24 cm, sometimes bract-like. Indumentum similar to the adjacent part of the stem. Capitula 1 to 6, 2.5—6(7) cm diam. including rays; involucre shorter than 7)1-16(21) cm long, 0.5-2.5 mm wide. Phyllaries herbaceous, ovate-lanceolate, rays, 1.54 ст diam.; peduncles (0. ovate-subulate to narrowly elliptic; the outer 0.8— 2.2 ст long, 1—3(3.5) mm wide; the inner 0.7-2.1 cm long, 0.5-2.5 mm wide. Indumentum of long- stalked glandular trichomes and sometimes also with eglandular trichomes. Receptacles pubescent or glabrous. Flowers with yellow corollas. Ray flower corollas (1.2)1.5-3 em long, 1.7-5 mm wide, ob- ovate-elliptic, apex generally with 3 teeth. Disk flower corollas 4—7(8) 1-2.5 brown-tinted to olive-green, with grooved-reticulate mm. Cypselae to somewhat warty surface, dimorphic or һототог- phic. Cypselae from ray flowers (1.5)3 .5 mm, glabrous or glabrate, with or without pap- pus. Cypselae from disk flowers (1.5)2—4 х 0.2— — Annals of the Missouri Botanical Garden 1.3 mm, with eglandular or glandular trichomes, with pappus. Pappus (2.5)4—5.5 mm, white. Chro- mosome number 2n — 60, 120 (Fernandes & Quei- 1971, as D. pardalianches). м Illustrations. Figures 1 F, 6C, 16. Distribution. North of the Iberian peninsula and mountains in central Spain and eastern Por- Open moist rocky places, cliffs, from sea level to tugal. screes, woods, and near watercourses, 2500 m elevation (Figs. 17, 18). This taxon can be confused with Doronicum aus- triacum due to their similarities (e.g., habit, leaves, capitula, habitat; see comments under this species), but the presence of this latter species in the Iberian peninsula is only based on a few gatherings more than 50 years old from Andorra and Cerdagne in the Pyrenees. This species must be searched for in these areas in the Pyrenees. Doronicum carpetanum is variable with regard to quantity of indumentum, size, ploidy level, and presence of pappus in the ray flowers. Variation in these characters follows geographical patterns, and in most cases these populations can be distin- guished morphologically. Hybridization events both contemporary and in the origin of one of the sub- species here recognized (subsp. Фай) cannot be discarded in this group. However, further investi- gation is needed, and at present, taxonomic rec- ognition at the subspecific level is preferred to han- dle the intraspecific variability (see also comments for D. grandiflorum). In the present taxonomic treatment the following subspecies are recognized: KEY TO SUBSPECIES OF DORONICUM CARPETANUM 1. Receptacle glabrous or glabrate; plants gener- ally bearing one « Apes ulum l' Rec aptae i pubescent; plants bearing one to several capitula ....... ер: C ub of disk flowers with mo stly glan dular trichomes эЬ. D. carpetanum subsp. diazii 2'. Сурѕејае of disk flowers with mostly eglandular trichomes |... 6c. D. ca nn subsp. kuepferi 3(1). Lower and ı middle e leaves with eglan- dular trichomes, sometimes al« ud glandular trichomes оа. D. rane subsp. pubescens ower and middle cauline ves glabrous or with mainly оч tric . D. шш. suben: carpetanum 2(1). T o — 6a. Doronicum carpetanum subsp. carpetan- um Plants glabrous or glabrate, generally g in the upper part, sometimes also with scattered eglandular trichomes. Receptacle pubescent. Cyp- selae dimorphic, the inner with eglandular tri- andular chomes. Chromosome number 2n = 120 (Chacón, 1987) Illustrations. Figures 1F, 16A-D. Distribution. Massifs in the center of the Ibe- rian peninsula in Spain, plus some scattered pop- ulations in the north. Open moist rocky places and near watercourses, altitude 900-2300 m (Fig. 17). Selected specimens examined. SPAIN. Castilla-La Mane pm »uadalajara, Cantalojas, s Negra, 20 June 1985, р" Cardiel, s.n. (MACB). Castilla y León: Ávila, Solana de Ávila, laguna del Duque, arroyo Malillo, Álvarez & Yagüe 931 (MA); Salamanca, Candelario, sierra de Béjar, 28 June 1979, Amich et al. s.n. (MA); Burg Pineda de la Sierra, pico Mancillas, 14 July 1984, Benedi et al. s.n. (MA, MAF); Ávila, sierra de Majarreina au des- sus de iun ‘as prés Plasencia, Bourgeau 2508 (С ч WILLK, NY): e El Calvitero, 16 July 197€ Carrasco et Pe CB); ep Puebla de E sierra Calva de Porio. a July 1973, Casaseca s.n. (MA); T b o Trigaza, Costo & Fernández Quirós 5868 ( M co arco de Avila, sierra del Barco, н и. el al. 7133 a A); Segovia, Cerezo de Arriba, pico del Lobo, или et al. 10709 (MA); Burgos, valle de Valdelaguna, sierra de Neila, Mt. Haedillo, Gil Zúñiga & Alejandre 225-88 (MA); Ávila, Hoyos del Espino, Las Chorreras, Luceño & Vargas is 208 (MA); Soria, sierra E bollera, rio Racioncillo, July iola MACB); Soria, sierra iai, 6 July 1979, Mendiola s.n. (MACB); Soria, El Bercolar, sierra Cebollera, 17 Jul 1980, Pe sn. CB); Soria, Laguna Negra de Ur- bión, 15 Ju ‚ Navarro s.n. 7 7: ДЕ pm s.n. (MAF); Zamora, San Martín de Castañeda. El Cabezo, 22 June 1987, Roa s.n. (MA); Ávila, sierra de Gredos oriental, puerto de Mijares, July 1984, Sánchez- Mata s.n. (MAF); Avila, Navalguijo, garganta de los Ca- balleros, 3 June 1990, Sardinero si n. (MAF); Soria, Santa Inés, Majadarrubia, Segura Zubizarreta 12525 (MA); Ávi- la, San Martín del Pimpollar, Segura Zubizarreta 22654 (MA). Comunidad de Madrid: e a de Guadarrama, laguna de Penala ara, Almaraz et al. 802 (MA). La Rioja: Ezcaray, cerro de San Lorenzo, M et al. 805 (MA); mi de Pi iqueras, Sandwith 5684 (K); Zaldierna, sierra e la Demanda, pico Torocuervo, s.n. (MA). País К Alava, La | , 1 July 1985, Uribe-Echebarría s.n. (MA). Principado de Asturias: lagos de Saliencia, Luceño & Vargas 2569’ (MA) 6b. Doronicum carpetanum subsp. diazii (Pérez Morales & Penas) Álv. Fern., Novon 11: 294. 2001. Doronicum diazii Pérez Morales & Penas, Lagascalia 15: 155, fig. 2. 1990. TYPE: Spain. León, Abelgas, Puerto Bermejo, July 1974, C. Romero s.n. (holotype, LEB 4290 not seen). Plant up to 70 em tall, glabrous or glabrate in the lower part, glandular at the middle and upper part, sometimes also with scarce eglandular tri- Basal leaves chomes. Stems generally simple. Volume 90, Number 3 Álvarez Fernández 349 2003 Doronicum (Asteraceae) Figure 16. A-D. Doronicum carpetanum subsp. carpetanum (drawn from Luceño & Vargas 208, MA). —A. Capit- ulum. —B. Phyllary. —C. Indumentum of a phyllary. —D. Ray flower. E-H. Doronicum carpetanum subsp. kuepferi (drawn from Navarro & Valle s.n., MA). —E. Capitulum. —F. Phyllary. —G. Indumentum of a phyllary. —H. Ray flower. 350 Annals of the Missouri Botanical Garden -0 b 0 | € P sH + + | фат / QU. e ia / M e а SM 4, Реа Д ax s в 94 / Чг ° oet зү, i > ev ЕА ета ج‎ " тай » \ D did d sd ) Si е , С / | «au ә rs o 1 Jj 5 | / ہر‎ = d / к= / “м y T \ ( SE ips E Ns c ©? Ü p» o Р 1 y № ) « E bees gine / E 7 ENS EU — S P add жаа. ee or d е PA pt. / po «ИЙ m - 35 + / 9 — EIE „ t Pi \ ie h 0 5 0 3 Figure 17. Distribution map for Doronicum carpetanum subsp. carpetanum. 10 5 0 5 T +7745 o ASOD ® 40 -5 0 5 e 18. Distribution map for: Doronicum carpetanum subsp. diazii (А); Doronicum carpetanum subsp. Kuepferi (©). Fi (ө); Du carpetanum subsp. pubescens ~ Volume 90, Number 3 2003 Álvarez Fernández 351 Doronicum (Asteraceae) sometimes present at flowering time; blade 4—6 X 3-4 cm; petiole 6.5-10 cm long, 1-2 mm wide. Lower and middle cauline leaves 3.5-8.5 X 2-6 cm. Upper cauline leaves 1.2-5.5 X 0.5-2 cm. Ca- pitula | to 2(4), 3-4 cm diam. including rays; in- volucre 2-2.5 ст diam. Phyllaries 0.9-1.4 cm long, .5-3 mm wide. Md glabrous to glabrate. ong, 3—3.5 mm wide. Cypselae panes dimorphic, ray flowers sometimes with pappus poorly developed. The in- Ray flower corollas ner cypselae mainly glandular. Chromosome num- ber 2n = 60 (new count reported here: Spain. León, Riolago, Alvarez et al. 924 (MA 611192)). Illustrations. Pérez & Penas (1990: 156, fig. 2); Figure 6C Distribution. Northern Iberian peninsula (Can- tabrian range) and central-eastern ranges (Picos de Urbión) in Spain. Open moist rocky places and screes, altitude 1700-2100 m (Fig. 18). Although type material was not available, plants collected at the type locality and also several spec- imens identified by Pérez (one of the authors) as Doronicum diazii were studied. Selected specimens examined. SPAIN. Castilla y León: León, San Emiliano, Riolago, pico Penouta, Alvarez et al. 924 (MA); Soria, Sierra de Urbión, Ceballos & Vi- cioso 1136 (MA); Soria, Covaleda, macizo de кы la- rga, Gil Zúñiga & Wehr 320/93 (MA); Soria, sierra de Urbión, Laguna Negra, Harrold & McBeath 240 (E); Soria, del pico de к E al pico Tres Cruces, 26 July 1982, Navarro s.n. (M ни рейа Репоша, DES 27 July 1988, Puente yu Morales s.n. (M F); Soria, sierra de Urbión, нея 5340 (К). E de Asturias: Somiedo, һгайа de Murias Longas, 25 Aug. 1985, Aedo s.n. (MA). 6c. Doronicum carpetanum subsp. kuepferi (R. Chacón) Álv. Fern., Novon 11: 294. 2001. Doronicum kuepferi R. Chacón, Anales Jard. Bot. Madrid 43: 269. 1987. TYPE: Spain. Cá- ceres, sierra de Majarreina, cerca del pico del . Rivas Goday s.n. isotypes, MA 348763!, MAF 119444). Plants up to 70 cm tall, glabrous or glabrate in the lower part, glandular at the middle and upper part, sometimes also with scarce eglandular tri- chomes. Stems generally simple. Blade of basal leaves 2.5-5 X 2—6 cm; petiole 4.5-14 cm long, 1-2 mm wide. Lower and middle cauline leaves 3— 6 X 1.54 cm. Upper cauline leaves 1-3 X 0.2— 1.5 ст. Capitula 1 to 2(4), 2.5-5.5 ст diam. in- cluding rays; involucre 2-3.5 cm diam. Phyllaries 0.9-1.7 cm long, 0.5-2.5 mm wide. Receptacles gla- brous to glabrate. Ray flower corollas 1.4—1.7 cm long, 1.7—4 mm wide. Cypselae dimorphic, ray flow- ers sometimes with pappus poorly developed. The inner cypselae mainly with eglandular trichomes. Chromosome number 2n — 60 (Chacón, 1987; re- counted and confirmed here: Spain. Ávila, Portilla de Talamanca, Álvarez & Yagüe 933 (MA 611198)). Figure 16Е-Н. Illustrations. Distribution. Central-western of the Iberian peninsula (Sierra de Gredos). Open moist rocky places, screes, an near watercourses, altitude 1800-2500 m (Fig. 18). Selected specimens examined. SPAIN. Castilla y León: Ávila, Solana de Avila, Portilla de Talamanca, Al- varez & Yagiie 933 (MA); Avila, El Calvitero, 16 July e Gredos, sierra de Gredos, Dresser 846 (E); Salamanca, sierra de Béjar, El Trampal, Nieto Feliner et al. 2736 (MA); Avila, pue La Serrota, 5 July 1997, Palacio et = s.n. (MA); Salamanca, deo ч 27 July 1900, Pau s.n. (LY); Salamanca, sierra de Béjar, Hoya- moro, 22 xn 1983, Rico s.n. (MA CB): Ávila, Siue de Gredos, El Morezón, 26 July 1958, Rivas Goday s.n. (МАЕ); Ávila, sierra x Béjar, La Ceja, 26 July 1989, Rivas Martínez et al. s.n. (MAF); Avila, puerto de Villa- toro-Villanueva del Campillo, 20 May 1982, Sánchez- Mata et al. s.n. (MAF); Salamanca, sierra de Béjar, La Hoya, circo de la Рейа Negra, uly 1990, Sardinero s.n. (MAF); Ávila, sierra de Tormantos, Puerto Castilla, circo de El Barco, 23 Aug. 1990, Sardinero s.n. Salamanca, Candelario, Calvitero, Valdés Be E et. di 5812 (MA). Extremadura: Cáceres, sierra Maj na, cerca del Pico del Telégrafo, 7 Aug. 1946, Riv К Codie s.n. (MA, MAF). = > = 6d. Doronicum carpetanum subsp. pubescens (C. Pérez Morales, A. Penas, F. Llamas & C. Munibe 50: 11. 1998. Doronicum pubescens C. Pérez Mo- rales, A. Penas, F. Llamas & C. Acedo, La- zaroa 14: 7. 1994. TYPE: Spain. León: puerto del Pontón, 12 June 1992, A. Penas & M. E. García s.n. (holotype, LEB 47120 not seen). Acedo) Aizpuru, in Aizpuru et al., Plants mainly with eglandular trichomes at least in the middle part, also glandular in the upper part. Blade of basal leaves 4—6.5 X 3.5-5.5 cm; petiole 7—9 cm long. Capitula 4.5—5(7) cm diam. including rays. Receptacles pubescent. Cypselae dimorphic, the inner with eglandular trichomes. Chromosome number 2n = 120 (Chacón, 1987, as D. carpetan- um). Distribution. Northern Iberian peninsula in Spain and central Portugal (Serra da Estrela). Open moist rocky places, cliffs, woods, and near water- courses, altitude 50-2200 n Although type material of Doronicum pubescens 352 Annals of the Missouri Botanical Garden was not seen, plants collected at the type locality as well as several specimens identified by Pérez (one of the original authors) as Doronicum pubes- cens, were studied. Selected specimens examined. PORTUGAL. Beira Alta: Manteigas, Se erra da Estrela, Mondeguinho, ышк et al. 1296 2 July 1982, Ada. s.n. (MA); July 1983, Aedo s. n. (МА); , pt Fue os ре, Harrold & McBeath 158 (E); Vega de Pas, puerto de Estacas de Trueba, Pardo de Santayana a Moralis 1690 (M E Sede. 1 slo, 27 ig Patino et al. : , Rivas бубон el al. s.n. LM. RE 23: Castilla y Aedo et al. 3631b (MA); León, Enc 'inedo, laguna de | (MA); León, Oseja de Sajambre, puerto del Pontón, Ál- MA); León, Boca de Huérgano, Cor- Pue bla de Lillo, 18 July 1974 Cueto Ancino, 18 July 1951, Borja s.n. (MAF); León, circo Cebollero, puerto de San Isidro, 16 July 1974, Casaseca & Б Díaz s.n. : León, puerto de las Señales, 27 July 1979, Casaseca el ve A); León, entre le col de Р кеша, adas et Posada d Valdeón, Char- . Palacios del Sil, pico Catoute, 15 ; León, puerto de Mart tinez et al. s.n. Zamora, Portilla del Padornelo, 24 July 1972, Val- Nocedo, pode 16 June 1981, Rivas (MAF); T dés Bermejo s.n. ;alicia: Lugo, Cervantes, monte Camporredondo, Degrade, pico Tres Obispos, Alvarez et al. 926 (MA Corufia, Caaveiro, 25 Apr. 1981, Amich et al. s.n. (MA); La Coruña, Puente Carreira, 29 Mav 1953, Bellot s.n. (MA, MACB, ; Orense, sierra do Inver- nadeiro “n əza de Val do | C usb. 10 July 1973, Cas- troviejo s.n. (MA); Orense, Viana del Bollo, montaña de Ramilo, Merine 18 (MA); La Coruña, Malpica, As Porte- — ^ — las, > 1994, Sofiora s.n. (SANT). Asturias: supra Pajares, 14 July ); Cangas de Narcea, vega de Renfos, Muniellos, Silv а Рап- МАЕ). do et al. 1394 (MA, MACB, 7. Doronicum cataractarum Widder, Feddes Repert. 22: 115, Taf. 25-27. 1925. TYPE: Kärnten, Koralpe, Im obersten WeiBwassergra- ben, nahe der Waldgrenze bei etwa 1630 m, zwischen den Felsblócken des Baches, 20 Aug. 1923, К J. Widder s.n. (holotype, GZU!; isotypes, GZU!). Plant up to 100(+) em tall. Rhizomes woody to somewhat woody, scarcely pubescent to pubescent, and generally with leaf remains forming dark scales on nodes. Stem branched in the upper part, leafy, internodes generally shorter than the adjacent leaves. Indumentum of uniseriate eglandular tri- chomes (up to 0.4 mm), also with short-stalked or subsessile glandular trichomes near the capitula, sometimes glabrate. Leaves slightly dentate to den- tate. Basal leaves sometimes present at flowering; blade 8-20 X 8.5-19.5 ст, ovate to orbicular, with cordate to subcordate base and generally blunt apex, with actinodromous venation; petiole 21-26 ст long, 2.5—4.5 mm wide. Lower and middle cau- leaves 7-19 leaves or sessile, fiddle-shaped, semi-amplexicaul. Upper 0.8-2 cm, sometimes bract-like. line 6-15 cm, similar to basal cauline leaves 2.5—4.: ovate to. ovate-lanceolate, Indu- mentum similar to the adjacent part of the stem, sometimes with uniseriate and multiseriate eglan- dular trichomes (up to 2 mm). Capitula 2 to 14, 4— 8 em diam. including rays; involucre shorter than rays, 2.5-5 mm diam.; 0.7-1.5 mm wide, sometimes with turbinate base peduncles 3-16 cm long. during fruit (up to 12 mm width). Phyllaries her- baceous, 1.2-1.8 em long, 1.5-3.5 mm wide, ovate- lanceolate to elliptic, generally with acute apex. In- dumentum of short-stalked glandular trichomes and uniseriate eglandular trichomes, sometimes glab- rate. Receptacles glabrous. Flowers with yellow co- rollas. Ray flower corollas 2.5-3.5 ст long. 1.7-3 mm wide, oblong-elliptic to obovate-elliptic, apex generally with 3 teeth. Disk flower corollas 4—5 X 2—3 mm. Cypselae brown-tinted, with grooved-retic- ulate surface, dimorphic. Cypselae from ray flowers ca. 3.5 X 1 mm, glabrous or glabrate, without pap- pus or sometimes with a poorly developed pappus. Cypselae from disk flowers 2.5— ).8 mm; pappus 4—5.5 mm, white. Chromosome amnibus 2n — 00 (data obtained from several indexes of plant chromosome numbers: Fedorov, 1969; Goldblatt, 1985; Goldblatt & Johnson, 1994, 1996; original sources not seen). Illustrations. Widder (1925: Taf. 25-27); Hegi (1928: 716, fig. 424); Figures 1D, 8C, 12E, 12F. Distribution. Europe (Austrian Alps). In gullies and rocky places near watercourses, altitude 1600— 1900 m (Fig. 15 Morphologically, the closest species to Doroni- cum cataractarum is D. austriacum (see comments above), and there are only slight and few differenc- es between them. Doronicum cataractarum is an endemic from the Austrian Alps, which is included within the area of distribution of D. austriacum. The characters used to distinguish them are the scarcely pubescent lo pubescent rhizomes of D. cataractar- um versus glabrous rhizomes of D. austriacum; base of capitula glabrous to glabrate with short-stalked or subsessile gladular trichomes in D. cataractarum versus base of capitula glabrate to pubescent or with long-stalked glandular trichomes in D. aus- triacum. In addition, although these two species have dimorphic cypselae (ray flowers without pap- pus and disk flowers with pappus), this is not a very Volume 90, Number 3 2003 Álvarez Fernández 353 Doronicum (Asteraceae) stable character in D. cataractarum and sometimes the ray flowers in this species have a poorly devel- oped pappus, which is never present in D. austria- cum. Selected specimens examined. AUSTRIA. Karnten: Koralpe, am Bache im Himmelreich, 22 Aug. 1934, Drob- ny s.n. (B); Koralpe bei Deutschlandberg, Fest 571 (B): Koralpe, Bachufer unterhalb der Grillitschhütte, 30 July 1950, че sn. Es G); Koralpe, Grosses Kaar, Sep. 1953, Patzak s.r Waldgrenze, 20 Aug. 1923, Widder s.n. (GZU); Weidwasaencnhe n, Grillit- schhiitte, Sturzbach, 19 Aug. 1928, Widder s.n. (MAF). Steiermark: Ronde des Seebaches, Seekar der Koralpe, 24 Aug. 1936, Widder s.n. (G); Seebach der Koralpe, See- kar, 28 Aug. 1939, Widder s.n. (B) —. T 8. Doronicum clusii (All. Tausch, Flora 11: 178. 1828. Arnica clusii All., Auct. Syn. Stirp. Taurin.: [18]. 1773. Aronicum clusii (All.) W. D. J. Koch, Syn. Fl. Germ. Helv.: 382. 1837. TYPE: not located; protologue citation: “in Al- bula Rhaetica & aliis Rhaetorum alpibus." Plant up to 40 cm tall. Rhizomes fleshy to some- what woody, with shining white-tinted short tri- chomes on nodes, generally with leaf remains. Stems generally not branched, with leaves mainly at the base or in the middle basal part of stem. Indumentum of eglandular and glandular tri- chomes, more abundant near the capitula. Leaves entire to dentate. Basal leaves generally present at flowering time; blade 7(8.5)-(1.5)2 X 1-2.5(3.5) cm, elliptic to ovate-elliptic, truncate or attenuate base, blunt to acute apex, with actinodromous is pinnate-actinodromous venation; petiole (0.8)2— ст long, 1-3(4) mm wide. Lower and middle cau- line leaves 2.5-10 X 0.7—2.5(3.5) cm, similar to basal leaves or sessile, ovate-elliptic to narrowly elliptic, semi-amplexicaul. Upper cauline leaves 1.5-5.5 х 0.5-3 cm, similar to middle cauline leaves, or ovate-lanceolate. Indumentum of stiff, acute, and shiny multiseriate eglandular trichomes (up to 2.5 mm), and thin, tangled uniseriate eglan- dular trichomes (up to 2 mm), mainly on leaf mar- gins, also short-stalked glandular trichomes on leaf blade. Capitula l(to 4), rays; involucre shorter than rays, 2.5-5 cm diam. Phyllaries herbaceous; the outer 1.2-2 cm long, 1.5-3.3 mm wide; the inner 1-2 cm long. 1-2.3 mm wide, ovate-lanceolate to widely subulate. In- 7.5 ст diam. including dumentum similar to the upper part of stem. Re- ceptacles glabrous. Flowers with yellow corollas. 8-2.5 cm long, 2.5—4.5 mm wide, obovate-elliptic, apex generally with 3 teeth. Disk flower corollas 4—5 X 1.5-2 mm long. Cyp- selae brown, with grooved-reticulate surface, ho- momorphic, 1.5-2.5 X 0 Ray flower corollas 1. 2—1 mm, with eglandular trichomes or glabrate. Pappus up to 5.5 mm, white. 6 20 (Skalinska, 950, as Aronicum clusii; Tasenkevitch et al., А *Goldblatt & Johnson, 1996, see comments below). Illustrations. Reichenbach (1854: tab. 63, fig. 2); Hegi (1928: fig. 432); Sávulescu (1964: pl. 99, fig. е Кеѕтегцӣ & Moravetz (1956: fig. 1); Figure 3F Chromosome number 2n = 60*, Distribution. Europe (Alps and Carpathians). Open moist i places and screes, altitude 1500— 3000 m (Fig. 1 type um of Arnica clusii could not be found. and although the protologue matches the di- agnostic features of this taxon, its identity here is tentative and the formal synonymies need to wait until clarification. There are three European species, Doronicum clusii, D. glaciale, and D. grandiflorum, that are morphologically similar, and whose areas of distri- ution overlap in some places in the Alps occu- pying the same habitats. All of them have homo- morphic fruits (all cypselae with pappus), ovate to elliptic basal leaves with truncate or attenuate ba- ses, and rhizomes fleshy to somewhat woody with short trichomes on nodes, generally with a single capitulum or sometimes a few (2 to 4). All of them grow in open moist rocky places in the mountains, preferably the upper tree line to 3000 m in eleva- tion. Although they only differ in the type of in- dumentum, it is a very constant character (more noticeable on the leaf margins). In D. clusii and D. glaciale, stalked glandular trichomes are absent or scarce, while they are common in D. grandiflorum Fig. 3B). In contrast, long (up to 2.5 mm) non- glandular trichomes are present in both D. clusii and D. glaciale (Fig. 3E, H). Differences between D. clusii and D. glaciale are slight, and it is difficult to delimit them. Typical individuals from D. clusii and from D. glaciale qe the indumentum as it H and E, respectively), but —. is represented (Fig. 3 some individuals have a very few scattered thin and tangled uniseriate trichomes, which are abundant in typical D. clusii and absent in typical D. gla- ciale. Because of this, the present taxonomic treat- ment includes those exceptional individuals within D. clusii, although the existence of hybrids between these two species is not rejected. The citation of the chromosome number 2n — 60 for D. clusii was found in Goldblatt and Johnson's (1996) index, but the original source for this data was not seen. Selected specimens examined. pig cie —— 1 : Grafenalpe, Krakaudorf, July ‚ Fest s.n. (B); тена Tauern, Наше Styrie, » m 1868, “Ober. 354 Annals of the Missouri Botanical Garden ? Ў => Е T ER pu H zi UNE + Мыз + 7 50 d Е AL _ Dem er А ы p ay” ® е P Me Ы n "e d e Porge Ў f e vet s к. ce эы EN Ld о бнт b ©, حر‎ C R Dx c А sud a SOS. v. Е [ c ce ои m y | pt ~, ЕЯ A. \ У е 0 "Жы S н, ees Ч саш eU "e с i j TMA, 5 CN a. UN imul. АК eed Ce 40 2 Je кы (i) us C SS ` age C Hao | N : YO E 7 ° E Lp ул 9 ھک‎ v соата 8 E! f Se ге کیا‎ N ` DS e p SS p А 2. 5 ре Ioue Sm Y хе) s Nes үм, ae е С G be ъ . ! A TM Figure 19. leitner s.n. (B, LE). Tirol: Paznaun, Fladner Massio, 2 A : g. . (B); 2 Au. 1 e s.n. Ae Otathaler boudin Ober- Aug. T "d 1933. ter s.n. T B) . Mt. Muttenjoc th, 16 Aug. 1890, eee s.n. 5, Born- CZECHOSLOVAKIA. Vysoké Tan in val- le montana Mlynica a, 6 Aug. 1933, Dostdl s.n. (BM, MA, a pod Gerlachovsky Stit, July 1895, Fitko s.n. + Maj agas Tátra, lacum Ké An Zöld- tó, Kümmerle & 9 790 (B, О); Brezno, Mt. Dumbier, 3 Aug. 18‹ Kupcok s.n. (E). FR CE. Alpes-Maritimes: Mt. Bissa, col de Tende, aad. au 139 (COI-WILLK, G, K). ITALY. Lombardia: Bormio, passo dello Stelvio, Álvarez et e 1355 (MA); a Bormio, Mt. SE nre 31 July 19 i el E m s Bergamasques, 31 July 1910, гриме d 1. (G). Tren- tino-Alto Adige: Trento, е col del Cu `, Alvarez et al. 1353 (MA). Valle d'Aosta: psy zum Colle Pinter, am Bergbach, in Felsen, 1982, Royl & Schiers s.n. (B). POLAND. Zakopane, Beskid, i Karpaten, Hempel 2786 (B); Rysy, lacum Czarny Staw, Tatry Wysokie, Tatri Alti, 3 Sep. 1938, Madalski s.n. (B). SWITZERLAND. Graubünden: Alp d'Ischolàs, Engadine, Binz 405 (MO); Pontresina, e h, paso del Bernina, Lejner, C viejo et al. 11615 (MA). Tessin: San Bernardino, 28 July 1920, Partei s.n. "s Valais: près Zermatt, 14 Aug erpécle-Barcolla, Bonnier 1888, M sn. (MA); F m 3 ( ; We Е Quellflur, Damboldt 679/70 (B); ж) Grashang bei Spielbode nalm, Saas-Fee, Dam- жу 714/70 (B); E. F 's alpes de Taesch, 5 June 1908, Palibin s.n. (LE )SLAVIA. Visoki Verch, Liptau, July 1894, Hinas s.n. M Distribution map for: Doronicum clusii (©); Doronicum corsicum (C1); Doronicum glaciale (9). 9. Doronicum columnae Ten., Fl. Napol.l, Prodr.: 49. 1811. TYPE: Italy. “Majella” [M. Tenore s.n.| (lectotype, designated by Álvarez Fernández & Nieto Feliner (1999: 802), АР!). Plant up to 70 ст tall. Rhizomes woody to some- what woody, glabrous, generally with leaf remains forming dark fibers or scales on nodes. Stems not branched, generally scape-like. Indumentum of un- iseriate, multiseriate eglandular trichomes, short- stalked and long-stalked glandular trichomes, scarce at the base, more abundant near the capit- ula, Leaves dentate to slightly dentate. Basal leaves generally present at flowering time; blade 1.5-7 X 2-6.5 ст, orbicular to broadly ovate with cordate to subcordate base, with blunt or subacute apex, with actinodromous venation that sometimes tends to be acrodromous; petiole thin and stiff, 4—15(24) cm long, 0.5(-2) mm wide. Lower and middle cau- line leaves 1.9—8(10.4) X 1.2—6(7) em, similar to basal leaves or sessile, ашсын, а semi-am- plexicaul. Upper cauline leaves 1.5 0.8—3 cm, ovate-elliptic to ovate-lanceolate, some- times bract-like. Indumentum of uniseriate eglan- dular trichomes (up to 0.5 mm), conspicuously on the blade edge. Capitula 1(2 to 3), 2.8-7 cm diam. including rays; involucre shorter than rays, rarely equaling them, 1.8—4 cm diam. Phyllaries herba- Volume 90, Number 3 2003 Álvarez Fernández 355 Doronicum (Asteraceae) ceous, ovate-subulate, generally with acute apex; the outer 0.7—1.8 cm i inner 0.7-1.6 cm long, 0.5-1.8 mm wide; margins long, 0.7-2.8 mm wide, the sometimes ciliate, with acute, stiff and equidistant multiseriate eglandular trichomes (up to 0.6 mm). Indumentum mainly glandular, but also with unis- eriate eglandular trichomes. Receptacles glabrous or pubescent. Flowers with yellow corollas. Ray flower corollas 1-3 cm long, 1.5-3.8 mm wide, oblong- elliptie, apex generally with 3 teeth. Disk flower corollas 3.5—4.5 X 1.8-2 mm. Cypselae brown, with grooved-reticulate surface, dimorphic. Cypselae from ray flowers 2-2.5 X 0.7-1 mm, glabrous to glabrate, pas pappus. Cypselae from disk flow- ers 1-2.3 0.4 chomes; рарриѕ 3—4 mm, white. Chromosome num- ber 2n — 60 (Garbari et al., 1980; Van Loon, 1980; Strid & Franzén, 1983; Lippert & Heubl, 1988; Baltisberger, 1991). Illustrations. Reichenbach (1854: tab. 64, fig. 1); Hegi (1928: 715, fig. 423); Sávulescu (1964: pl. 98, fig. 2); Figure 1E. mm, with eglandular tri- Distribution. Europe (Balkans extending to central Greece, Carpathians, Alps, and Appenines). Meadows, shady rocky places and gullies, from sea level up to 2700 m in elevation (Fig. 15). Doronicum columnae is a polymorphic species morphologically similar to D. carpaticum and D. orientale. These species share the habit (scape-like stem with a few caulinar leaves bearing a single capitulum), the shape of basal leaves (orbicular to ovate with cordate to subcordate base), and the cil- iate margins of phyllaries (Fig. 5E, F), although this latter character is not constant in D. columnae and D. carpaticum. The most distinctive character be- tween D. orientale and both D. columnae and D. carpaticum is the type of rhizome, which is fleshy with pubescent nodes in D. orientale versus woody to somewhat woody and glabrous in D. columnae and D. carpaticum. 'There is only one character to distinguish D. columnae and D. carpaticum: the di- morphic cypselae (ray flowers without pappus) in D. columnae versus the homomorphic cypselae (all flowers with pappus) in D. carpaticum. Some spec- imens that have poorly developed pappus in the ray flowers are included in D. carpaticum, although the hybrid nature of them is not rejected. In addition, there are some exceptional specimens of D. colum- nae that have a few capitula instead of a single one and that can be confused with another sympatric species, D. pardalianches. It is easy to distinguish between them by comparing their rhizomes, which are woody and glabrous in D. columnae while fleshy with pubescent nodes in D. pardalianches. Besides, cypselae in D. pardalianches turn black at maturity, which p a duri character in the genus. e for Doronicum columnae was diffu (Alvarez Fernandez & Nieto Feliner, 1999). Since there is no collection date, doubt re- mains concerning this issue. Based on historical records, Tenore visited the type locality himself de- scribing several new species in his Prodromus in 1811. Taking into account this fact, and without other suitable type material, this was the best choice as lectotype. Selected specimens examined. ALBANIA. Gjirokas- tër: Mali Gjer, Alston & Sandwith 1528 (BM, K). Korçë: Ostrovice, Moskopolé, Когсё, Alston & Sandwith 2065 hkodér: Nikéi, Klementi, Baldacci 457 (BM). Tirané: Dajti, Pennington 41 (K). AUSTRIA. Tirol: Schlern, auf Felsgeroll in der Klamm, 14 July 1903, Beh- rendsen s.n. (С). BULGARIA. Grad Sofiya: а Р ES e June 1974, Markova, Cerneva & Gergino (BM 5, MA). Plovdiv: Kalofer, Wagner 77 (BM). So. fiya: E Musala, 11 Aug. 1976, Beck s.n. (B). GER- M. NY. Bayern: "= TEN АРА н) Hirschbichl Bind-Alm Mittereis ) & Podlech 25818 (NY). GREECE. Ípiros: Mt. S Atchley 945 (K); Papignon, Mt. Gamila, Lancaster 120 (BM). Makedhonía: Chaliki, Mt. poesia Sintenis 733 (B, E, G, K). Nísoi Ayaíou: Mt. Korax Aetoliae "P dh 23 July 1879, Heldreich s.n. (К, LE). Stereá Ellas— voia: Trapeza, Katafigon, Oeta, Balls & Gourlay B3231 (BM, E, K). Thes salía: Mt Olympus, Archibald 326 (E). oed Mehr p in valle Kazár, Schneider 1 (B, BM, К, MO). Veszprém: Bihania, in valle Izvor, T se, 22 May 1906, Gulyás s.n. (G). ITALY. Abruzzi: La M nella: Вин 127 (В). о Nellino, I н 1913, Pellanda s.n. (С). Em Forli, Foresta di Campigna, C QR E iiu) 8743 (С, MA, MAF). Lombardia: Como, Mt. Barbisino, Val Massino, "m July 1850, prod s.n. (LE). Poe Gargano, Mt. S. Angelo, 29 May 18 Porta & Rigo s.n. (LE). Toscana: Mt. Senario Prope ES je 20 Apr 1856, Caruel s.n. (LE). Trentino-Alto Tre Gardena, Sella Gruppe, yos et 2 1354 (MA). "ROMANIA. Alba: Piatra Strutu prope pag, Avram lancu, 24 May 1973, Gergely & Toader s.n. (B, ‚ K, MA). : Mt. Bucegi, m. Predeal, S of = i Ploiesti, оран 15 (BM). Hunedoara: Mt. Retezat lacum Zanoaga, 9 Aug. 1933, Borza & Nydrddy s.n. (BM, G, К, MO). Prahova: E Baiului, Muntele Cumpatu, 17 June 1983, Zamfir s.n. (B). YUGOSLAVIA. Bosna i Her- cegovina: Mt. Frebovic, Sarajevo, Beck & Fiala 232 (С, K. 3 Makedonija: Usküb, ad Divin Treska, Bornmüller 4263 NY). Srbija: Belgrad, Ripanj, 8 May 1887, Bornmüller s.n. ( D 10. Doronicum corsicum (Loisel.) Poir., in Lam., Encycl. Suppl. 2: 517. 1811 sica Loisel., Fl. Gall. Aronicum corsicum (Loisel.) DC., Prodr. 6: 319. 1838. TYPE: not located; protologue ci- tation: “in Corsica, ad rupes aquis fluentibus irriguas (D. Richard. Herb.).” . Arnica cor- 356 Annals of the Missouri Botanical Garden Plant up to 100(+) ст tall. Rhizomes woody to somewhat woody, glabrous. Stems branched in the upper part, leafy, internodes generally shorter than the adjacent leaves. Indumentum glandular and also with uniseriate and multiseriate eglandular tri- chomes, abundant near the capitulum, sometimes glabrous in the lower part. Leaves dentate to slightly dentate. Basal leaves absent at flowering time, sim- ilar to cauline leaves. Cauline leaves oblong-ellip- tic, sessile, slightly auriculate, semi-amplexicaul, acute apex, pinnate-actinodromous venation. Mid- dle cauline leaves 7-16 X 2-5.5 cm. Upper cau- line 3.5-10 X 1-2.5 cm. Indumentum scarce, with uniseriate and multiseriate eglandular leaves trichomes, and short-stalked glands. Capitula sev- eral, 5(+), ca. 5 cm diam. including rays; involucre much shorter than rays, 2-2.5 cm diam.; peduncles 3.5-7 cm long, 1 mm wide. Phyllaries herbaceous, ovate to ovate-lanceolate, generally with acute apex, sometimes slightly papery at the base or at the margins; the outer 0.6-0.8 ст long, 1.7-2.5 mm wide; the inner 0.7-1 cm long, 0.7-1.6 mm wide. Indumentum mainly of uniseriate eglandular trichomes, sometimes also with multiseriate eglan- dular trichomes and glandular trichomes. Recepta- cles glabrous or pubescent. Flowers with yellow co- rollas. Ray flower 0.4 cm, oblong-elliptic to obovate-elliptic, apex generally with 3 teeth. Disk flower corollas up to 0.7 cm long. Cypselae brown and with a smooth to grooved sur- face, homomorphic, ca. 3 X 1 mm, glabrous or gla- corollas ca. 2.5 brate; pappus ca. 5.5 mm white. Chromosome — 60 (Contandriopoulos, 1957 number 2n Illustrations. | Loiseleur-Deslongchamps (1807: tab. 20) Distribution. Corsica. In forests and gullies and on rocky slopes that are sometimes inundated, al- titude 700—1750 m (Fig. 19). The type material cited in the protologue of Ar- nica corsica refers to one sheet from the D. Richard herbarium. This collection should be in P, but no material was found there. Unfortunately, the pro- tologue includes only a crude and incomplete il- lustration that is not appropriate as a lectotype. Al- though the protologue and the illustration included both match the diagnostic features of this taxon, its formal identity needs further investigation in the search of an appropriate lectotype or a neotype. Currently, this is the only species of Doronicum 'rowing in Corsica (see comments under D. gran- diflorum). Although D. corsicum is morphologically similar to D. austriacum, they differ basically in the homomorphic cypselae in D. corsicum versus di- morphic in D. austriacum. In addition, D. corsicum has uniform, elliptic, sessile caulinar leaves with dentate margins, while D. austriacum presents dif- ferent types of caulinar leaves in the same speci- men (acropetally, petiolate to fiddle-shaped and ovate), with entire to slightly dentate margins. Selected ое examined. FRANCE. Corse: forét d'Artone bei Evisa, 20 July 1932, Aellen s.n. (MA); Cal- acuccia, Golo, Sidossi. July 1912, Cousturier s.n. (NY); Mt. d'Oro, 12 July 1916, Forsyth-Major s.n. (K); valleé de сш prés Vriario, 18 July 1906, Ce s.n. ; Calvi, Mt. Sollier, 1822, Jacquemont s.n. (NY); Tav- Corte, Kralik 538 (E, K); Fiumorbo, Pazzi du Mt. К); l'Inc udine, Lambinon M d A); Lit de la Restonica, près de Corte, Mabille (BM, K); "n Niolo, Requien 250 (BM, К); жын к july: 1878, Rei май s.n. (COI-WILLK, E, К, NY); foret d’Attone, 1885, Reverchon s.n. (B, E, NY) ignano, Renoso, Kralik 638a (E, 4 NS x 11. Doronicum dolichotrichum Cavill., Annu- aire Conserv. Jard. Bot. Genéve 13-14: 252. 1911. TYPE: Transcaucasus. Gourie, descente du mont Khino au défilé Goghieti, 18 July 893, М. M. Alboff s.n. (lectotype, designated by ww Fernández & Nieto Feliner (1999: 802), G). -— aiio 2d Widder & Rech. f., Oesterr. Bot. Z. 97: ) Syn. nov. TYPE: Transcaucasus. Azerbaidjan, Sari Chaman, Mirdamadi K2381 ( lotype, Doronicum hakkiaricum J, R. Edm., = Roy. Bot. vds YPE: Edinburgh 32(2): as 1973. Syn. nov. T ey. Hakkâri, K P. Н. Фан 15 : О. Р л. 124383 (holotype, E isotypes Doronicum bracteatum J. R. Edm., e E. Bot. Gard. burgh 32(2): 257. Ls Syn. nov. TYPE: Iraq. Edin n us | ( . Guest & E. R. Lud. Arl Gird Dagh, near Rust, E. low- Hewitt 2928 а K!). Plant up to 100(+7?) ст tall. Rhizomes woody, generally without leaf remains. Stems branched in the leaves distributed along the stem, upper internodes generally longer than the adjacent leaves. Indumentum of triangular, white-tinted, multiseriate eglandular trichomes (1— 3 mm), sometimes scattered, long-stalked glandular glabrous, upper part, trichomes (0.5-3 mm), sometimes abundant near the capitula, and occasionally uniseriate eglandular trichomes and short-stalked glandular trichomes, sometimes glabrous at the base. Leaves entire to dentate. Basal leaves sometimes present at flower- ing time; blade 6-15 X 8.5-18.5 em, orbicular or ovate, with cordate base and blunt or acute apex, with actinodromous venation; petiole 4.7-23 cm long, 3-3.5(6) mm wide, with sheathing base, sheath ca. 5 cm long. Lower and middle cauline leaves with blade 10-26 X 5-21.5 cm, similar to leaves or sessile, fiddle-shaped, semi-am- petiole 12-20 ст long, 3.5-5.5 mm basal plexicaul: Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) wide. Upper cauline leaves 4—9 X 1.6-7.5 cm, ses- sile, ovate to obovate, or bract-like. Indumentum similar to the adjacent part of the stem, sometimes also with uniseriate eglandular trichomes on mar- gins and on veins in the abaxial surface of leaves. Capitula 2 to 8, 4—5.5 cm diam. including rays; involucre shorter than rays, 2.5-3.5 cm diam.; pe- duncles 2-9 ст long, 1.5-2.3 mm wide, sometimes turbinate at the base of capitula (4—7 mm wide) in fruit. Phyllaries herbaceous, ovate-lanceolate to ob- ovate-lanceolate with acute apex; the outer 1.1—1.7 em long, 2-5.5 mm wide; the inner 1.2-1.5 ст long, 1.5-2.3 mm wide. Indumentum similar to the upper part of stem. Receptacles glabrous. Flowers with yellow corollas. Ray flower corollas 1.5-2.5 cm long, 2-3.8 mm wide, obovate-elliptic to ob- long-elliptic, apex generally with 3 teeth. Disk flow- er corollas 4—5 mm long. Cypselae brown, with smooth or slightly reticulate surface, dimorphic. Cypselae from ray flowers 2.84.3 X 1-1.5 mm, glabrous or glabrate, without pappus. Cypselae 1-1.5 mm, with eglandular trichomes; pappus ca. 4.5 mm, from disk flowers 2-2.5 X sometimes white. (Only immature cypselae seen.) Chromosome number unknown. Figures 2C, D, 7D, 20A-D. Illustrations. Distribution. Northeastern Turkey, Caucasus, and south of the Caspian Sea. Growing in woods, open moist rocky places, meadows, gullies, and near watercourses, altitude 800-3300 m (Fig. 14). The type of indumentum is critical for distin- guishing among species of Doronicum in south- western Asia. The variability of this character both in type and abundance in D. dolichotrichum (Fig. 2C, D) sometimes makes it difficult to identify ma- terial conclusively. Plants with intermediate char- acters between D. maximum (Fig. 23B, C) and D. macrophyllum (Fig. 26B, C) have Ben found. The diagnostic characters used to separate Do- ronicum bracteatum (Edmondson, 1973), D. hak- kiaricum (Edmondson, 1973), and D. hyrcanum (Widder & Rechinger, 1950) overlap substantially with those of D. dolichotrichum. The patterns of indumentum variability also match those of D. dol- ichotrichum. Accordingly, these names are consid- ered synonyms. CAUCASUS. North ur dagh, — 183 (LE). aucasus: Carthalinia, Tibia uri, A. H. & V F Brotherus 500b (BM, G); Kac ‚ Mt. Choczal- dagh. 12 July 1898, Fomin s.n. (LE); rg nia, Migri Gjunej, inter m. Gieljedzhik et m. Ketshmas, 20 Aug. 1932, Karjagin & Grossheim s.n. (К); Tiflisskaja, Gorijskij. Gora Tschra- Tschard, 7 July 1916, Krylov & Schteinberg s.n. (LE); Ba- tumskaja, Artvinskij, Arsiyanskij, 16 June 1910, Nesterov Selected specime ns examined. s.n. (LE); Nachitschevan, Mt. Ketshal-dagh, 4 Aug. 1934, Prilipko & Isaev s.n. (К); Nor-Bajazet, Gri-zür, 16 Aug. 1928, Schelkovnikov & Kara-Murza s.n. (LE); Mt. Alagéz, Kaznafar, 31 July 1931, Tamemshian & Maleer s.n. (E); Chokhatauri, BA hanî Meskhetski, Mt. vut MR 22 July 1979, Vašák & Esvandzhia s.n. (G, N. Azer- baidjan, Maku to Khoy, аы. Assadi е pee s ian 30306 (E); Zanjan, Vansar mountains, 20 June 1983, Moussavi et al. s.n. (IRAN); Ostan 2, Dimelo, Schmid 5989 (G, W); Gorgan, Ketul, Sharif 224 (W); Azerbaidjan, Ahar, Hassano, 8 Au Termeh s.n. ; Azer- baidjan, Kalibar, Nabidjan, Kouhha-ye Doghroun, 26 June 1978, Termeh et al. s.n. (IRAN). IRAQ. Helgord aus Каш & Serhang 24553 (К); Qandil n mountain, Qala Diza, rdc in H (BM). TURKEY. Artvin: Yalniczam Silsi- lesi, Savsat, Albury et al. 3159 (K 5 po enig Kordevan dag, а ач at Kütül vayla, Davis & Hedge 30342 (BM, E, K, W); Yalnizcam-Gebirge bei Karaköy, Savsat, Raus 4864 (B). Erzurum: Bingóze Köyü yaylas, Yildiz Dagi, Tatli 5171 (E). Hakkari: Kara dag, Davis & Polunin 24383 (BM, E, m EM Yagmurlu dag between Sarikamis and Karuar s & Hedge 30819 (E, К); Arpogay, Kaya Diplei, dard 2340 (GAZI). 12. Doronicum y sagen а B. Clarke, Fl. Brit. Ind. 3: 333. . TYPE: Karakorum, C. B. Clarke 30258 А Ib. designated by ÁI- varez Fernández & Nieto Feliner (1999: 803), Doronicum ipio um Cavill., Annuaire — em ve 13-14: "e 1911. Syn. nov. Russia. Siberia: nt. sent. milit.," M. J ME jon 385 (lectotype, parei here, G!). “Mo Plant up to 90 cm tall. Rhizomes woody to some- what woody, glabrous, generally with leaf remains forming dark scales. Stems not branched, leaves generally arranged all along the stem, internodes generally shorter than adjacent leaves. Indumen- tum of glandular trichomes (up to 2 mm), more abundant near the capitula, sometimes only very scarce eglandular trichomes or glabrate. Leaves en- tire. Basal leaves sometimes absent at flowering time; blade 2-7(9) х (0.2)1—2.5(3.2) cm, elliptic to obovate, with attenuate base and generally blunt apex, with actinodromous to pinnate-actinodromous venation; petiole 1—4.5 cm long, 2-5 mm wide. Lower and middle cauline leaves (3)5-12.5 X (0.3)1.5—4(5.5) em, similar to basal leaves or ses- sile, obovate, ovate, elliptic, or almost fiddle- shaped, semi-amplexicaul, with blunt apex. Upper cauline leaves 2.5-6(7.5) X (0.4)0.5—1.5(3.6) cm, similar to middle cauline leaves or ovate-lanceo- D late. Indumentum similar to the adjacent part of stem, sometimes also with uniseriate eglandular tri- chomes and glands on margins. Capitula. solitary. )5—7.5 em diam. including rays; involucre shorter cm diam. — than rays, rarely equaling them, 3—5 Phyllaries herbaceous, ovate-lanceolate to subulate: the outer 1.2-2(3) em long, (1)1.5-2.5(3.5) mm 358 Annals of the Missouri Botanical Garden Figure 20. A-D. Doronicum dolichotrichum (drawn from Davis & Hedge 29493, K). —A. Capitulum. —B. Phyllary. —C. Indumentum of a phyllary. —D. Ray flower. E-I. Doronicum kamaonense (drawn from Polunin 56/170 E, as D. roylei). —E. Capitulum. —F. Indumentum of the base of capitulum. —G. Phyllary. —H. Indumentum of a phyllary. —1. Ray flower. Volume 90, Number 3 2003 Álvarez Fernández 359 Doronicum (Asteraceae) Figure 21. wide; the inner 1-2 cm long, 0.7-2 mm wide. In- dumentum similar to the upper part of stem, very scarce at the apex. Receptacles glabrous. Flowers with yellow corollas. Ray flower corollas 2-3 cm long, 1-2.5(3) mm wide, obovate-elliptic, apex with 2 or 3 teeth, sometimes without teeth, acute. Disk flower corollas 4.3—5.3 X 1.3 mm. Cypselae brown, with smooth surface, dimorphic, generally glabrous, sometimes with scattered eglandular or glandular trichomes. Cypselae from ray flowers without pap- pus. Pappus up to 5 mm, white. (Mature cypselae not seen.) Chromosome number unknown. Distribution. Central-western China (provinces of Tibet-Qinghai and Xinjiang), Mongolia, Turkis- tan, Pamir, and Himalayas. Woods, open rocky places, gullies, and near watercourses, altitude 1800—5000 m (Fig. 21). Cavillier (1911) recognized two sympatric spe- cies, Doronicum falconeri and D. turkestanicum. The character claimed to distinguish them (shape and based on it the differentiation into two groups of of leaves) is vague and quite polymorphic, species is not easy to make. Some specimens de- termined by Cavillier as D. falconeri and D. tur- kestanicum were included in a multivariate. mor- phometric study (Álvarez Fernández & Nieto Feliner, 2001) resulting in no discrimination at all. Because of the lack of consistency in the delimi- tation between these two species, D. turkestanicum, Distribution map for: Doronicum gansuense €); Doronicum falconeri (А); Doronicum stenoglossum (C). which was described later, is here treated as syn- onym of D. falconeri. Despite the recognition of Cavillier’s species Do- ronicum turkestanicum in a local floristic study (Gorschkova, 1961), this name was still lacking a type designation. Thus, in this work the best pre- served specimen chosen among Cavillier's citations as D. turkestanicum is designated as its lectotype — see synonym above). The morphological similarities of Doronicum fal- coneri with other central Asian species is discussed above (see comments for D. altaicum and D. bri- quetit). Selected specimens examined. Tibet-Qing hai: Ata Kang La, Nagong, "ote Vini [ 0876 (ВМ, i Xinjiang: Thianschan, v sch, ГА 1, Brocherel 39 (G); Sairam, 18 July 1878, Райхон s.n. "LE , 5); Tien Shan, Urumqi river, Liston 818-1 (MO); Mts. Bogdo- ola et Urumtschi, Merzbac 5 has LE). INDIA. Jttar “Pradesh: (LE). Punjab: Rotang, Kulu I (K); Rupin pass, Dhaola Dhar, ‘Simla Hill, Sherriff 7405 (BM, E). JAMMU-KASHMIR. Kagan valley between Ba- lakot and Babusar pass, Abel 94 (BM); Astor, Alampi Lá, = m = E, K); Burzil, Koelz 9429 (GH, NY); Srinagar, Vis r, Lancaster 160 (BM); Haramukh, Lud- low & Men 7850 (BM, E, UPS); Karakoram, Gharesa сае Nagar, die 6238 (B, BM, E); Karakoram, His- par glacier, Turmun-Makerum, Russell 1235 (BM); Nafran, Lidder Stewart 12638 (NY). KAZAKHSTAN. Talgarskoe, ~ Annals of the Missouri Botanical Garden 14 June 1909, aes s.n. (LE); Alma-Atinskij, Alma- Atinekoe, L EH y . Dubiansky & Basilevskaja s.n. ee Sd drin 25 Aug. 1930, Matveeva Ket ini n pass, 19 June 1378. Regel s.n. (LE); ‚ Katon-Karagaj. 10 Aug. 1930, Smirnow s.n. KIRG е Fergana, Ak-basoga, 31 June 1901, Alexcenko s.n. (LE); Sir-Darinsk, Tian-shan, Ala- med, 6 July 1910, Golbek s.n. (LE); Boamskoe, Issik-Kul- skaya, Terskej-Alatau, Turgen-Aksu, Kujliu, Inilchek, 23 July 1965, Grudzinskaja s.n. (LE); Zailyijskij Alatau, Se- mirechenskaya, Vernenskij, Lipsky 1179 (LE) P Alatau, Tekes, 26 May 1950, Medvedeva et al. s.n. (LE Semirechenskaja, Pishpekskij. Vi p HM aM E Ala-ar- chi, Sovetkina 428 (LE). MONGOLIA. Changai, Kondra- tieva 68 (LE); Khara-C Pra Khairkhan. Duru, Pobedi- mova 339 (LE); Я schan, Przevalski 101 (LE): Kobdosekij, Bulugun, "ien 13101 (LE). RUSSIA. Al- tay: Ojrotiya, Koshagachskij, Chujskij, Che ш ve 17 Aug. 1937, Shatakelberg & mons s.n. a E). 1 ISTAN. Zaalajskij, Gordaba, 2 July ‚©. А. Fedchenko s.n. (LE). UZBEKISTAN. жее. )з С Frag. Schart jugi Alaici, 12 July 1900, Tranzschel s.n. (LE). = 13. Doronicum gansuense Y. L. Chen, Acta анон, Sin. 36: 73. 1998. TYPE: China. : Tebbu Мап, J. А Rock 12102 (holo- ре РЕ по! ѕееп). Doronicum cavillieri Álv. Fern. & Nieto Fel., Ann. Bot. Fenn. 37: 250. € Syn. nov. TYPE: China. Gansu: T'ao river ba Minshan range, Kuang ke, J. К Rock 12389 БИШР, NY isotypes, BM!, GH!, LE!). Plant up to 30 cm tall. Rhizomes somewhat woody, glabrous, generally with leaf remains. Stems not branched, with leaves all along the stem, inter- nodes generally shorter than adjacent leaves. In- dumentum of glandular trichomes (up to 2 mm), more abundant near the capitula, sometimes gla- brous at the base. Leaves entire. Basal leaves some- times absent at flowering time: blade 1.4-3.5 X 1.5—3 em, orbicular, suborbicular, or elliptic, with truncate or attenuate base and blunt apex, with ac- tinodromous to pinnate-actinodromous venation; petiole 3-8.5 cauline leaves 3-5 X 1-3 cm, sessile, em long, 0.8-1.5 mm wide. Lower and middle ovate-elliptic to widely pun . semi-amplexicaul. Upper cauline leaves 1.5 ст, similar to middle cauline leaves. a en of uniseriate eglandular trichomes (up to 0.5 mm), short-stalked glandular trichomes, and sometimes also long- stalked glandular trichomes, mainly on leaf mar- gins, scarce, sometimes glabrous. Capitula solitary, 3-5.5 than rays, 2-3 cm diam. 1.2-1.4 cm long, 1.5-2.5 late to widely subulate, with blunt apex (bearing a sessile gland). Indumentum similar to the upper cm diam. including rays; involucre shorter Phyllaries herbaceous, mm wide, ovate-lanceo- part of stem, more abundant at the base. Receptacles glabrous. Flowers with yellow corollas. Ray flower corollas 2-2.5 cm long, 2-3 mm wide, obovate-el- liptic, apex generally with 2 or 3 teeth. Disk flower corollas ca. 5 X 2 mm. Cypselae brown, homomor- phic, ca. 3 X 1 mm, with eglandular trichomes or glabrate, sometimes glandular. Pappus up to 4 mm, white. (Mature cypselae not seen.) Chromosome number unknown. Illustrations. Chen (1998: 36, fig. 1); Álvarez Fernández & Nieto Feliner (2000: 251, fig. 1); Fig- ure 4 Distribution. Central China (provinces of Gan- su and Sichuan). Woods, rocky places, and grassy slopes, altitude 3000-3700 m (Fig. 21) ter the name Doronicum cavillieri (Álvarez Fernandez & Nieto Feliner, 2000) was published, the authors realized that a previously published name of which they were not aware corresponded to the same species. Although the type material of Doronicum gansuense was not available for the pre- sent study, plants from the type locality, as well as from other localities cited in protologue were ex- amined. Morphological relationships of D. gan- suense with other central Asian species is discussed above (see comments for D. altaicum and D. bri- дие). Selected e ns examined. CHINA. Gansu: T'ao river, Merku valley, Rock 12192 (E, K, NY, S, W); T'ao river, Minshan range, ML Kuang ke, Rock 12389 (BM, GH, LE, NY ); Tebbu, о Lo 13020 (E, GH, K, LE, NY, S, W); Lianhuashan, e Xian, Wang 91161 (MO). Si у E 8 iuf to Songpan, 1989, Chamberlain et al. s.r 14. Doronicum glaciale (Wulfen) Nyman, Syll. Л]. Eur.: 1. 1855. Arnica ages Wulfen, in Jacq., Collectanea 1: . 1786. Aronicum glaciale (Wulfen) Rchb., "t Germ. Excurs. 1: 234. 1831-1832. Doronicum hirsutum subsp. glaciale (Wulfen) Rouy, Rev. Bot. Syst. Géogr. Bot. 1: 55. 1903. TYPE: Austria. Malnizer Tauern [sine collector], ex herb. Wulfen (lec- totype, designated by Alvarez Fernández & Nieto Feliner (1999: 803), W!). Arnica doronicum Jacq., Fl. Aperi. id tab. 92. 1773. баши гоп ОЛЕ" Са 1 Cuvier, Dict. D. DD um m doronicu (Jacq.) . Germ. a curs. 1: 233. 1831-1832. TYPE: “Arnica doronicum" [sine collector], ex ы rb. Lin- naeus (lectotype, designated here, LINN n? 1001.41). Auc 6 5.3 ores ےک‎ 8 “к — Plant up to 30 ст tall. Rhizomes fleshy to some- what woody, with shining white-tinted short tri- generally with leaf remains. chomes on nodes, Stems not branched, with leaves mainly at the base Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) or on the middle lower part of stem. Indumentum with eglandular trichomes more abundant near the ca- pitula. Basal leaves generally present at flowering time; blade 1.54.5 X 1-2 cm, elliptic to ovate-elliptic, with truncate or attenuate base, blunt to subacute apex, of short-stalked glandular trichomes, also Leaves entire to slightly dentate. with actinodromous to pinnate-actinodromous ve- nation; petiole 2—6 cm long, 1-2.5 mm wide. Lower and middle cauline leaves 3—7.5 X 1—4 cm, similar to basal leaves or sessile, ovate-elliptic to narrowly elliptic, semi-amplexicaul. Upper cauline leaves 1.5-3.5 X 0.3-2 or ovate-lanceolate. cm, similar to middle cauline leaves, Indumentum of stiff, acute, and shiny eglandular trichomes (up to 2.5 mm), mainly on leaf margins, and also short-stalked glandular trichomes. Capitula solitary, 4—5.5 cm diam. including rays; involucre shorter than rays, .5-3.5 cm diam. Phyllaries herbaceous, ovate-lan- ceolate to widely subulate; the outer 1.2-1.5 ст long, 1.5-3.3 mm wide; the inner 1-1.4 cm long, 1-1.3 mm wide. Indumentum similar to the upper part of stem. Receptacles glabrous to glabrate. Flow- ers with yellow corollas. Ray flower corollas 2-2.2 cm long, 2.5-3 mm wide, obovate-elliptic, generally with 3 teeth. Disk flower corollas up to 4 mm long. Cypselae brown, with grooved-reticulate apex surface, homomorphic, up to 2 mm long. with eglandular trichomes or glabrate. Pappus up to 4 mm, white. Chromosome number 2л = 60 (Polat- schek, 1966, as D. calcareum; Lovka et al., 1972). Illustrations. Jacquin (1773: tab. 92, 1789: tab. 586); Sturm (1814: tab 19.2); Reichenbach (1854: tab. 62, tab. 63 figs. 1, 3); Hegi (1928: fig. 431): Figure 3C—E. Distribution. Europe (central-eastern Alps). Open moist rocky places and screes, altitude 1000— 2800 m (Fig. 19). In 1773 Jacquin described Arnica doronicum. This species corresponds to е and al- though Jacquin named it first, the ICBN (Greuter et al., 2000) does not allow the use E the same epithet for both genus and species; thus the correct epithet is “glaciale” given by Wulfen in Jacquin (1786). Part of Jacquin's collection is kept at LINN. When this herbarium was studied, one sheet of Jac- quin’s material labeled as “Arnica doronicum" was found. To clarify as far as possible the identity of Jacquin's name, its lectotype is designated here (see D. glaciale synonyms). Doronicum glaciale is a species very similar to D. clusii and to a lesser extent to D. grandiflorum. Similarities and differences among these species are discussed above (see comments for D. clusit). Selected specimens examined. AUSTRIA. Karnten: Gartnerkofel zur Kühnveger Alp., 15 July 1928, Bothe s.n. Lesachthal, 25 Ju ily 1907, tte, Bärental, Karawanken, Hodgkin 168 (K); Wolayer S See, Qailtal, Rodelen. 220 (K); Gloknerhauses am Steige zur Pasterze, = а ly 1899, Schulz s.n. (B). Salzburg: Radstädter Tauern, Plu Men- bad Sastein, July 1 971 (B, K); Gamskaar Kogel, Wyatt 28 (K); Sulzka nothaler Alps, Wyatt 89 (K). T Grafenalpe pro- pe Krakaudorf, July 1902, Fest s.n. (K); Mt. Hoc re hwab, Aug. 1887, Steininger s.n. (G, LY). Tirol: pm in valle Isolae, 10 Aug. 18 Kleiner Ispeltal, Johannishütte, July 1929, inis s.n. (B); Gröden, Dolomiten, Hochjoch, Geislen gruppe, 27 July 1907, Bornmiiller s.n. (B); montem Weisspitz pr. Herzing, Aug. 1888, Huter s.n. (К); Pfumpenseetauern, Bavaria, 25 July 1955, Launert s.n. (BM); Lienz, auf der Zoche, Hersc 'h- baummer Alpe, 31 ме HE Mer s.n. (B); со June 1878, Treffer s.n. (B); Tristen in Weissenbach, July 1890, diu s.n. (6): Bendelstein, клр Wy att 56 (K). GERMANY. Bayern: Steinige z pe in bs ara Prettan, Aug. 18 92. Treffer s.n. Trentino-Alto Adige: Trento, Campitello " Fassa, col Rodella, yet et al. 1350 (MA). SWITZERLAND. Fur- ka pass, Rhónegletscher, 24 July 1886, Bornmiiller s.n. (B). 15. Doronicum grandiflorum Lam., Encycl. 2: 313. 1786. TYPE: “Arnica altaic. pall., tige simple unifl. haute de 4 ou 7 pouces, fl. tres grande, fleurit au com. + de mai" [sine col- lector], ex herb. Lamarck (lectotype, designat- ed by Álvarez Fernández & Nieto Feliner (1999: 803), P-LA!, photograph). Doronicum sc orpioides Lam., m 2: 313. 1786. TYPE: Arnica scorpioide: " [sine collec is ex herb. Lund (lectotype, pie ee here, photo- srap Doronicum porte Chabert, Bull. = Bot. France 53: 547. 906. Syn. nov. TYPE: Austria. Valbona, € 1893, P. Porta s.n. (lectotype, E here, F = — Plant up to ca. 70 cm tall. Rhizomes fleshy to somewhat woody, with shining white-tinted short trichomes оп nodes, generally with leaf remains. Stems generally simple. Indumentum of short- and long-stalked glandular trichomes, also with multi- seriate eglandular trichomes more abundant near the capitula. Leaves slightly dentate to dentate. Bas- al leaves sometimes present at flowering time; blade 3-6(7) X (1)2—5 ст, ovate-elliptic, ovate or sub- orbicular, with subcordate, truncate or attenuate base, blunt apex, actinodromous venation; petiole 10(20)-(1.5)4 em long, 1—4(6) mm wide. Lower and middle cauline leaves (1)4—9(13.2) X (1)1.5—6(7) cm, similar to basal leaves or sessile, ovate-elliptic, sometimes almost fiddle-shaped, semi-amplexicaul. pper cauline leaves 1.2—5.5 X 0.5-2 cm, similar to middle cauline leaves, ovate to ovate-lanceolate. Annals of the Missouri Botanical Garden Figure 22. Indumentum mainly of glandular trichomes, some- times also with eglandular trichomes. Capitula (to 4), 3.5-5.5 cm diam. including rays; involucre shorter ibus rays, 2—4 cm diam. Phyllaries herba- ceous, ovate-lanceolate to widely subulate; the out- ег 1-2(2.5) cm long, 1.5—4 mm wide; the inner 1— 2 cm long, 1-2.8 mm wide. Indumentum of glan- dular trichomes, sometimes also with eglandular trichomes. Receptacles к to glabrate. Flowers with yellow corollas. Ray flower corollas 1.3-2.6 cm long, 24.5 generally with 3 teeth. Disk flower corollas 3.5—5 1-2.5 mm. Cypselae brown, with grooved-retic- ulate surface, homomorphic, 2.5—4.5 0.5-1.5 mm, with eglandular trichomes or glabrate, some- mm аА obovate-elliptic, apex times with glandular trichomes. Pappus up to 5.5 mm, white. Chromosome number 2n — 60 (Favar- ger & Huynh, 1964; Favarger & Küpfer, 1968). Illustrations. Jacquin (1776: tab. 349); Sturm (1814: tab 19.2); Reichenbach (1854: tab. 62); Sá- vulescu (1964: pl. 188 fig. 1); Bolòs & Vigo (1995: 837); Figure 3A, B. Distribution. moist rocky places, screes, and near watercourses, altitude 900—3000 m (Fig. 22). Variable in size and shape of basal leaves and Central-western Europe. Open type of indumentum. Based on the latter, two spe- cies were recognized in the previous revision of the genus (Cavillier, 1907), Doronicum viscosum and D. Distribution map for Doronicum grandiflorum. portae, which are separated virtually solely by glan- dular versus mainly eglandular indumentum, re- spectively. Although type material of D. viscosum was not seen, several populations from the type lo- cality were studied. The type material of D. portae was found at КЇ, and its lectotype is designated above. Abundant intermediates preclude recogni- Поп of these two species, which are here placed as synonyms of D. grandiflorum. To clarify as far as possible the confusion around the epithet "scorpioides" the lectotype for Doroni- cum scorpioides is designated above. In the proto- logue, Lamarck mentioned “A Arnica scorpioides Г... refering to Jacquin’s plate Arnica scorpioides L. (Jacquin, 1776: tab. 349). Although the type of Lin- naeus’s name corresponds to D. pardalianches L., Jacquin's plate represents a specimen of D. gran- diflorum, as well as the only sheet kept at P-LA with the handwritten *A Arnica scorpioides L." (see comments below about D. pardalianches). Although similar in their morphology, the differ- ences in the type of indumentum among Doronicum grandiflorum, D. clusii, and D. glaciale are quite clear. In D. grandiflorum the trichomes are never entangled and always have a blunt apex (generally ending in two cells), and the stalked glands are similar to these trichomes (Fig. 3B) but with a glan- dular apex (generally two or more cells containing a brown substance). In D. glaciale the trichomes are stiff with an acute apex, and in D. clusii, which Volume 90, Number 3 2003 Álvarez Fernández 363 Doronicum (Asteraceae) also has this latter type of indumentum, trichomes are very thin, long, and entangled (see comments under D. clusii). Some populations of Doronicum grandiflorum from the Cantabrian range in northern Spain have broadly ovate to suborbicular, scarcely dentate to subentire basal leaves, and are difficult to distin- guish from D. carpetanum subsp. diazii. These two taxa are similar but the nature of their relationship is not clear. A likely hypothesis, which should be investigated, is that D. carpetanum subsp. diazii is of hybrid origin, its putative progenitors being D. grandiflorum and D. carpetanum subsp. carpetan- um. Although widely distributed in the European mountains, the most recent collection of Doronicum grandiflorum from Corsica was in 1917, suggesting that it is now extinct there. Selected specimens examined. ANDORRA. Mt. Canil- lo, 27 June 1847, Bourgeau s.n. (С, LY). AUSTRIA. Kürnten: Dobratsch, 20 July 1928, Widder s.n. (MAF). Oberósterreich: Windischgarsten, Oberleitner 73 (B). Steiermark: Totes Gebirge, Feuertal, Rechinger 2377 (BM, E, K). Tirol: Paznauntal, Ischgl, Townsend 93/574 (К). Vorarlberg: Schruns, 27 July 1895, Bornmiiller s.n. (B). FRANCE. Alpes-Maritimes: col de Tende, Bourgeau 140 (COI-WILLK). Ariége: Artigues, étang du Laurenti, Charpin & Dittrich 17387 (С). Basses-Alpes: Le Lautaret Dauphiny, July 1906, Brown s.n. (E). Corse: Mt. Rotondo, Forsyth-Major 292-23 (К). Isère: Mt. Obiou, 7 1864, Borel s.n. (К). Haute-Savoie: Mt. Vergy, Timothée 4264 (BR). Hautes-Alpes: Vars, Saint-André d'Embrun, Sieber 85 (С, К). Hautes-Pyrénées: Vignemale, 24 July 1850, Hennecart s.n. aida Loire: Pilatus, 10 Sep. 1884, Hamilton s.n. (E). GERMANY. Bayern: Bavaria, Mts. Krotten- к. 15 July E Bornmiiller s.n. (B). ITALY. Fri iuli-Venezia Giulia: Mts. Baldi, 8 July 1870, Rigo s.n. ap iguria: Carro, lago della Seala, 3 Aug. 1912, Beger n. (B). 3). Lomba rdia: Sondrio, Bormio, Valpisello. Aug. 1920, Longa s.n. (BM). Piemonte: Cüneo, Colla dell Piz- zo, Alpes d'Ormea, Charpin & Salaman 17480 (С). Tren- tino-Alto Adige: Avisio, inter Peniam et jugum Fedaja. 18 July 1906, Handel-Mazzetti s.n. (G). Valle Grand St. Bernard pass, Brummitt 5494a (K). Veneto: Belluno, Forcella Beccher, 11 July 1970. Deva & Mortin Aug. Posets, Almaraz & Cano 288 (MA); anc, ibón de Ip, Vogt 3960 (B). Санаа: Vega de Lié- bana, laguna de Pefia Prieta, Álvarez et al. 943 (MA); Pi- cos de Europa, inter lacus Los Pozos et La Canalona, Rechinger 1701 (W). Castilla y León: León, San Emili- ano, Peña Ubifia, 8 July 1990, Aedo s.n. (MA). Cataluña: Lérida, val de Arán, torrent de Barrongueta, Vogt & Prem 7327 (B). SWITZERLAND. Graubünden: Piz Padella, nu 27532 (B). Valais: val d'Entremont, Bec Rond, 4 . 1927, Cuatrecasas s.n. (MAF). Vaud: Ormond-Des- i д pic Chaussy, col des Mosses, Kramer 8672 (MA). YUGOSLAVIA. Slovenija: Julijske Alpe, Vrh Krnic, lac- um Bohinjsko, Wraber 9748/4 (B, BM). 16. Doronicum haussknechtii Cavill., Annuaire Conserv. Jard. Bot. Genéve 13-14: 255. 1911. TYPE: Turkey. Beryt dagh, H. K. Haussknecht 1029 (lectotype, designated by Álvarez Fer- nández & Nieto Feliner (1999: 804)). ек be. i J. R. Edm., Notes Roy. Bot. Gard. Ed- inburgh 32(2): 256. 1973. Syn. nov. TYPE: Turkey. Giresun, Karagöl, C. Tobey 1484 (holotype, E!). Plant up to 100(+) cm tall. Rhizomes woody, gla- brous, with or without leaf remains. Stems branched in the upper part, leaves distributed along the stem, upper internodes generally longer than the adjacent leaves. Indumentum of multiseriate and uniseriate white eglandular trichomes (ca. 0.2 mm), abundant near the capitula, sometimes glabrate. Leaves entire to dentate. Basal leaves sometimes present at flow- ering time; blade ca. 11 X 12-12.5 cm, orbicular or ovate, with cordate base and generally blunt apex, with actinodromous venation; petiole 12-20 cm long, with sheathing base, sheath ca. 6-8 cm long. Lower cauline leaves with blade ca. 23 X 18 em; petiole (0.7-)37.5 ст long, similar to basal leaves. Middle cauline leaves ca. 16 X 9 cm, ses- sile, fiddle-shaped, semi-amplexicaul. Upper cau- z line leaves ca. 5 X 2.5 cm, similar to middle cau- leaves or ovate to obovate, sometimes bract-like. Indumentum similar to the adjacent part of the stem and with short-stalked glandular tri- chomes, generally scarce, more abundant on mar- gins of leaves. Capitula up to 17(+), ca. 4 cm diam. including rays; involucre shorter than rays, ca. 2 line cm diam.; sometimes turbinate at the base of ca- pitula in fruit. Phyllaries herbaceous, ca. 1.3 cm long. 2 mm wide, ovate-lanceolate to obovate-lan- ceolate and acute. Margins scarcely and slightly fimbriate. Indumentum similar to the upper part of stem, sometimes with scarce multiseriate eglandu- lar trichomes. Receptacles glabrous or glabrate. Flowers with yellow corollas. Ray flower corollas 1.3-1.6 cm long, 2—4 mm wide, obovate-elliptic to oblong-elliptic, apex generally with 3 teeth. Disk flower corollas 1.5 mm. Cypselae brown, with smooth to slightly grooved surface, dimorphic. Cypselae from ray flowers 1.5-2 mm, gla- brous or glabrate, without pappus. Cypselae from disk flowers 3.5—4 X eglandular trichomes; pappus 3.5—4.5 mm, white. Chromosome number unknown. mm, sometimes with Figure 23F-J. Illustrations. Distribution. Northern and central Turkey (provinces of Giresun, Kayseri, and Maras). Mead- ows and near watercourses, altitude 2100—2600 m (Fig. 24). 364 Annals of the Missouri Botanical Garden ure 23. A-E. Doronicum maximum (drawn from Davis et al. 20588, E). —A. Capitulum. —B. Ph llary. —C. Indumentum of a phyllary. —D-E. Ray flower. F-J. Doronicum haussknechtii (drawn from Davis et al. 20010, E). — F. Capitulum. —G. Indumentum of the base of the capitulum. —H. Phyllary. —1. Indumentum of a phyllary. —J. Ray Volume 90, Number 3 Álvarez Fernández 365 2003 Doronicum (Asteraceae) 30 40 50 50 "t | — © 40 4 Figure 24. Distribution map for: Doronicum haussknechtii (04): Doronicum macrophyllum subsp. macrophyllum (9): Doronicum macrophyllum subsp. sparsipilosum (+); Doronicum maximum (А); Doronicum oblongifolium (C). There are several Turkish species included in the same morphological group (see comments for D. cacaliifolium and D. dolichotrichum above) that are distinguished from each other only based on the type of indumentum. Doronicum haussknechti is included in this group, and morphologically the most similar species is D. maximum, which also overlaps part of its area of distribution with D. haussknechtii (Fig. 24). In both D. haussknechti and D. maximum, the indumentum on the phylla- ries is very scarce or even absent (Fig. 23B, С, Н, D, while the rest of the Turkish species have pu- bescent or glandular phyllaries. The characters to distinguish between these two species are the white pubescence at the top of the peduncle (base of the capitulum) in D. m E (Fig. 23F, G), which is glabrous in D. m (Fig. 23A), and the scarcely fimbriate margins of phyllaries in D. haussknechtii (Fig. 23H, 1), which are entire, some- times with glands in D. maximum (Fig. 23B. C). The diagnostic characters used to separate Do- ronicum tobeyi (Edmondson, 1973) overlap sub- stantially with those of D. haussknechtii, and its patterns of indumentum match those of D. knechtii. Accordingly, this name is considered a synonym auss- Selected specimens examined. TURKEY. Kayseri: Is- ikdagi, Karlidere, Duman. & Aytaç 5413 (GAZI). Maraş: Goksun, Binboga dag, Isik dag, Davis et al. 20010 (BM, E, K). 17. Doronicum hungaricum Rchb. fil., Icon. Fl. Germ. Helv. 16: 34, tab. 65, fig. 1. 1854. TYPE: icon in Reichenbach (1854: tab. 65, fig. 966 I 1-8) (lectotype, designated by Ál- varez Fernández & Nieto Feliner (1999: 804)). Plant up to 80 cm tall. Rhizomes fleshy, glabrate to scarcely pubescent, with inconspicuous shining white-tinted trichomes on nodes, thick and short, sometimes stoloniform, with buds. Stems generally unbranched, scape-like. Indumentum mainly glan- dular, with short-stalked and long-stalked glandular trichomes, sometimes also uniseriate and multiseri- ate eglandular trichomes, more abundant near the capitula. Leaves entire, rarely subdentate. Basal ie es generally present at flowering time; blade 4— x 1-3 em, oblong-elliptic with truncate or atten- uate base, blunt apex, with acrodromous venation; petiole 4—8 em long, 1-2 mm wide. Lower cauline leaves 3-11 X (0.5)1-2.6 cm, similar to basal leaves or sessile, elliptic to fiddle-shaped, some- times semi-amplexicaul. Upper cauline leaves 2—5 X 0.4—1.3 cm, ovate-lanceolate, sometimes bract- like. Indumentum with uniseriate eglandular tri- chomes and short-stalked glands, scarce. Generally also with long multiseriate eglandular trichomes (up to 5 mm), mainly on the adaxial surface of middle vein. Capitula 1(2 to 3), 3-6 mm diam. including rays; involucre a little shorter than rays or equaling them, 2.5-4.5 em diam. Phyllaries herbaceous, ovate-subulate, generally with acute apex; the outer 1-1.5 ст long, 1—1.5 mm wide; the inner 1.1-1.8 cm long, 0.7-1 mm wide. Margins sometimes cili- ate, with acute, stiff and equidistant multiseriate eglandular trichomes (up to 1 mm). Indumentum mainly glandular. Receptacles glabrous or scarcely pubescent. Flowers with yellow corollas. Ray flower corollas 1.4—2.5 elliptic, apex generally with 3 teeth. Disk flower cm long, 1-2(3) mm wide, oblong- corollas up to 4 mm long. Cypselae brown, with 366 Annals of the Missouri Botanical Garden Figure 25. rugose-reticulate surface, dimorphic. Cypselae from ray flowers 2-2.3 X 0.6-1 mm, generally glabrous, Ipso. pappus. Cypselae from disk flowers 1.7— X 0.7-1 mm, with eglandular trichomes; pappus up to 3.5 mm, white. Chromosome number 2n = 60 (Baksay, 1956). о Reichenbach (1854: tab. 65, fig. 1-8); Săvulescu (1964: pl. 98, fig. 1); Figure Distribution. Eastern Europe (Balkans, Carpa- thians, and Ukraine). Forests and meadows, alti- tude 160-1900 m. Cultivated and sometimes nat- uralized (Fig. 25). The name Doronicum plantagineum var. hungar- icum Sadler (1840) was published before the ac- cepted name for ps species, Doronicum hungari- cum Rchb. are inc luded in re protologue of Reichenbach’s 4). Plants collected by Sadler specific name. However, in the protologue Rei- chenbach did not mention the earlier name, and so his name is not based on Sadler's. The name Doronicum longifolium Rchb. (1831— 1832) is clearly a synonym of Doronicum clusii (All.) Tausch. when Grisebach (1846) However, Distribution map for: Doronicum hungaricum (Wl); Doronicum plantagineum (e). combined it as Doronicum plantagineum var. lon- gifolium (Rchb.) Griseb., his description and geo- graphical distribution were those of D. hungaricum, not D. clusii. Later, the same author (Grisebach Schenk, 1852) explicitly treated Reichenbach's name as a synonym of D. plantagineum var. hun- garicum Sadler. Thus, the names D. longifolium auct., non Rchb., and D. plantagineum var. longi- folium (Rchb.) sensu Griseb. are synonyms of D. hungaricum Rchb. Doronicum enorm um could be confused with D. clusii, D. glaciale, and the Caucasian D. oblon- gifolium because of the elliptic entire basal leaves and similar habit in some specimens of those spe- cies, but in the case of D. hungaricum the rhizome is fleshy with pubescent nodes, the basal leaf ve- nation has an acrodromous pattern, and the phyl- lary margins are ciliate to somewhat ciliate. All these characters together lead to the inclusion of this species in the morphologic and phylogenetic "plantagineum" group (see Phylogeny above and ig. 9). Within this group the most closely related species is D. plantagineum, which differs mostly in the shape of basal leaves (ovate in D. plantagineum vs. elliptic in D. hungaricum) and in the type o Volume 90, Number 3 2003 Álvarez Fernández 367 Doronicum (Asteraceae) indumentum. Doronicum hungaricum is considered the vicariant species of D. plantagineum in eastern Europe, although it has a more restricted area than D. plantagineum has in western Europe and north- ern Africa (Fig. 25). Selected specimens examined. BULGARIA. Pa- zardzhik: Belovo, May 1894, Stribrny s.n. (E. К); Ses- trimo, May 1907, ‚ред s.n. (E). Plovdiv: Krichim M ^ se: Mt. Rhodope ad y 1900. Stribrny s.n. (W); Rhodope ad Stanimaka, May "1900, Střibrný s.n. (G). Varna: Varna, Gilliat-Smith 554 (К); Kamcyr, Sehacidis 300 (B, BM, MO). HUNGARY. Baranya: Mecsek prope Pécs, 25 May 1922, Boros s.n. 4 i Mész egy, 10 May 1870, Vrabélyi s.n. (B). Pest: Kamera- erdo, 18 May 1885, Degen s.n. (W); Mt. Hárshegy prope Budapest, May 1886, Degen s.n. (B); Mt. Kamen prope Pomáz, 16 May 1904, Degen s.n. (LE); vallis Farkas- vilgy Prope Budam, Filarszky & Schilberszky s.n. (B, BM, E, G, K, MO, NY); Leopoldifeld bei Ofen. 1873, Freyn s.n. (W); Жейн, Севек bei Budakeszi, 24 May 1933, Korb s.n. (W); M y, Nagykovácsi supra Budapest, T Apr. 1912, Каныңа А Szurák & Ti- mkó s.n : E, G, K, MO, NY, W); Kammerwald, prope Budam, Richter 520 (B, BM, G, W). Tolna: Mt. sókás, Simontornya, 27 May 1875, Tauscher s.n. (G, NIA. Alba: Blaj, Apr. 1923, Pop s.n. (E, RAINE. Sirashenski, 28 May 1955, LE); Chernaya, 22 Apr. 1961, Fodor s.n. (LE); Zlotij, "Bender, 23 Apr. 1909, Paczoski s.n. (LE); Stramenskogo, 6 May 1948, Shirokova s.n. M "d i GOSLAVIA. Srbija: Belgrad, Tapeider, 1888, Bor müller s.n. (B); Gabrovac prope Nisch, Pora 2200 (BM, G, K, W). ъа 18. Doronicum kamaonense (DC.) Álv. Fern., Novon 11: 294. 2001. Fullartonia kamaonen- sis DC., Prod. 5: 281. 1836. TYPE: angl. des Indes 1830" [sine collector]. ex herb. de Candolle (lectotype, designated by Álvarez Fernández (2001: 294), G-DC!, photograph). "Comp. d eee roylei DC., Prodr. 6: 321. 1838. TYPE: Cach- e, J. Е Royle 232 IN designated by Al- varez Fernández & Nieto Feliner (1999: 805), G- DC!, photograph). Plant up to 130 cm tall. Rhizomes woody to somewhat woody, glabrous. Stems branched in the upper part or sometimes from the base, internodes generally longer than the adjacent leaves. Indu- mentum in the lower part of stem made up of mul- tiseriate, retrorse and white-tinted eglandular tri- chomes (up to 4 mm), sometimes absent, upper part of stem generally glandular, with long-stalked glan- dular trichomes (up to 4 mm), sometimes also with uniseriate or multiseriate eglandular trichomes, rarely without glands; apex of glandular trichomes capitate, with more than 6 cells, peduncle capillary. Leaves entire to slightly dentate. Basal leaves gen- erally absent at flowering time; blade 3.5—6.5 X 3.5-7 cm, ovate to ovate-elliptic, with attenuate, truncate, or subcordate base and generally blunt apex, with pinnate-actinodromous venation; petiole 5-16 em long, 1.5-2 mm wide. Lower cauline leaves 3-11 X (0.8)2.5-9 cm; petiole 3-13 cm long, 1.5—4 mm wide, similar to basal leaves. Mid- dle cauline leaves 7-15.5 X 3-10 cm, sessile, fid- dle-shaped, semi-amplexicaul. Upper cauline leaves (0.8)1.5—7.5(11.6) х (0.1)0.8—4.5(6.2) cm, ovate-lanceolate, sometimes bract-like. Indumen- tum similar to the adjacent part of the stem. Ca- pitula 2 to 18, 1.5—4 cm diam. including rays; in- volucre shorter than rays, 0.8-3 cm diam.; .5) em long, 0.5-1.5(2) mm wide. Phyllaries babies. ovate-subulate, gen- peduncles 1—10(18 erally with acute apex; the outer 0.6-1.2 cm long, 1-3 mm wide; the inner 0.6-1.2 cm long, 0.5-1.5 mm wide. Indumentum similar to the upper part of stem. Receptacles glabrous. Flowers with yellow co- 1.5 cm long, 1.2-2 mm wide, oblong-elliptic, apex generally with 3 teeth. Disk flower corollas 2.5—4 X 1-2.5 mm. Cyp- selae brown to brown-red with grooved-reticulate surface, dimorphic. Cypselae from ray flowers 2— 3.6 X 1-1.5 mm, glabrous or glabrate, without pap- pus. Cypselae from disk flowers 2-3 X 1-1.5 mm, with eglandular trichomes; pappus (1.7)2—4 mm, white to yellow. Chromosome number 2n = 60 (Vir Jee & Kachroo, 1989, as D. roylei). rollas. Ray flower corollas 0.8— Figures 8D, 20E-I. Illustrations. Distribution. Central-southern Asia (Jammu- Kashmir to Nepal, Bhutan, and Tibet). Forests and m (Fig. 11) There is only one species in central-southern meadows, elevation 1900— Asia, Doronicum stenoglossum, which could be- come confused with D. kamaonense because of their similarities in habit. The differences between them are remarkable, since both have unique char- acters within the genus. Doronicum stenoglossum has corollas pale yellow to green shaded, linear ray flower corollas, and linear phyllaries (Fig. 4E, F). In D. kamaonense the type of indumentum at the base of the capitulum (glandular trichomes with a capillar peduncle and capitate apex bearing 6 or more cells; Figs. 8D, 20E, F) is a character to dis- tinguish it from other species. Although the area of distribution of D. kamaonense overlaps in part with D. briquetii and D. falconeri (see comments for these species and Figs. 11, 21), there are no no- ticeable morphological similarities between those and D. kamaonense. The name Doronicum roylei DC. was in use until the recent realization that the name Fullartonia ka- 368 Annals of the Missouri Botanical Garden maonensis DC. represents the same species and that it has priority (Álvarez Fernández, 2001) Selected specimens examined. _BHUTAN-SIKKIM. Sik- kim, Gharu napo, Cooper 867 (E). CHI Shingbe, Me La, Ludlow et al. 20406 (E). Tibet-Qinghai: Chumbi, Cooper 230 (E). thi tona, Tehri, Koelz 22052 (NY); Garhvál, Gaurikünd via Tríjugi Nardin and Maser Tal to Bílung, 1855, Schlagin- tweit s.n. (ЄН). JAMMU-KASHMIR. Grorai, Clarke 29287D (BM, LE); Sonamurg, Clare 30842 (BM, K dar eni Drummond 14024 (E, K); Satrundi, Chamba, 13 Aug. 1897, Lace s.n. (E); Srinagar, Gulmarg, Khillan- marg tear p v 206 (BM); Sinthan pass, Ludlow & Sherriff 9292 (BM, E); Karauli forest, near Rampur, Jhelun valley, nd & Sherriff 7719 (BM); Kishenganga valley, Osmaston 28 (К); Khelanmarg, Polunin 56/170 (B, BM, E); Sind valley, Stainton 7894 (Е); Shanda-Kel, Kish- enganga valley, road to Nanga Parbat, R. R. & I. D. ап 17786 (NY); Harwart. Timins 174 (BM, E); Hazara, Mokspuri, Murree hills, Webster & Sack 5715 (G, GH, K, | W). NEPAL. Lamrak, Dhwoj 196 (BM); Ghurchi Lagua, Polunin et al. Hi (BM, E, G, UPS); Barbaria Lekh, Po- lunin et al. 89 (E, UPS); Balangra pass, Polunin et al. 2622 (BM, E, UPS); Ratamata, Chakure Lekh, Polunin et al. 401 (BM, E, G, UPS); Mailung e Stainton 7400 BM); Chalike Pahar, Stainton et al. . BM, E, UPS); ‚ Stainton et бө 6051 (E, UPS); ad, ~ Tabata et al. 1072 (GH); Rara, Mugu, Ta- PA et e 2900 (BM, GH); Kali Lagna, Jumla, Tabata et al. 19. ae H); between Chautra and Maure lekh, Jum- la ioe Tabata et a 3309 (GH); Merghang, Wigram 5 (E, K). I — ч 8 jom 1 —. 19. Doronicum macrophyllum Fisch., Cat. Jard. Gorenki ed. 2: 40. 1812. TYPE: North Caucasus. Beschtau [E A. F Marschall von Bieberstein s.n.|, ex herb. Marschall von Bie- berstein (lectotype, designated by Álvarez Fer- nández & Nieto Feliner (1999: 804), LE!). Plant up to 120(+) cm tall. Rhizomes woody to somewhat fleshy, glabrous, with or without leaf re- mains. Stems branched in the upper part, leaves distributed along the stem or mainly on the basal part of stem, upper internodes generally longer than the adjacent leaves. Indumentum of uniseriate (ca. 0.2 mm), multiseriate (up to 1 mm) eglandular tri- chomes, and glandular trichomes (0.5-2 mm) gen- erally abundant near the capitula, sometimes gla- brate at the base. Leaves entire to dentate. Basal eaves sometimes present at flowering time; blade (8)19-26(30) X (7))7-23 ст, orbicular or ovate, with cordate base and blunt or acute apex, with actinodromous venation; petiole 7-18 cm long, with sheathing base, sheath 3—8(10.7) cm long. Lower cauline leaves with blade 6-24 X 5-19.5 ст; pet- iole 9-33 em long, 1—5 mm wide, similar to basal х 2-16.5 semi-amplexicaul. Upper X 0.9-14(15) cm, similar to leaves. Middle cauline leaves 5-21.5 эст, sessile, fiddle-shaped, cauline leaves 3-17 middle cauline leaves or ovate to obovate, some- times bract-like. Indumentum similar to the adja- cent part of the stem, more abundant on margins and on veins on the abaxial surface of leaves. Ca- риша 2 to 13, 3—5.5(7) em diam. including rays; involucre shorter than rays, 1.5—3.5(4.5) em diam.; peduncles (1.5)3-10.5(16) cm Ted 1-1.6 wide, sometimes turbinate at the base of capitula mm (1.3 cm wide) in fruit. Phyllaries herbaceous or sometimes slightly papery at the base or at the mar- gins, 0.6—1.5(2) cm long, 0.8—3(4) mm wide, ovate- lanceolate to obovate-lanceolate or subulate. In- dumentum similar to the upper part of stem. Receptacles glabrous or glabrate. Flowers yellow. Ray flower corollas 1.5(1.9)-3(3.5) em long, (1.7)2.3—3.3(5) mm wide, obovate-elliptic to ob- long-elliptic, apex generally with 3 teeth. Disk fl 4—6 1.5-2.5 mm. Cypselae brown, with smooth to grooved surface, dimorphic. Cypselae from ray flowers 3—4.5 X 1—1.3 mm, gla- brous or glabrate, without pappus. Cypselae from disk flowers (2)2.3—4.5 X 0.5-1(1.5) mm, some- times with eglandular trichomes; pappus (1.5)3—5 mm, white. Chromosome number 2n = 30, 60 (data obtained from two indexes of plant chromosome numbers: Fedorov, 1969; Goldblatt, 1988; original sources not seen). ow- er corollas (3. Illustrations. Figures 24, 26A-D. Turkey and Caucasus. Growing in woods, open moist rocky places, mead- i 00—370 The characters distinguishing this species from Distribution. Northern ows, and near watercourses, altitude 15 ) m. others in southwestern Asia are mainly based on the type of indumentum (see comments on D. dol- ichotrichum). Doronicum macrophyllum is а poly- morphic species, and within it, two subspecies can be distinguished: KEY TO THE SUBSPECIES OF DORONICUM MACROPHYLLUM l. Plants generally with more than 3 capitula and more than 3 cauline leaves (including bract-like leaves — ——————— КЕКЕ 19 . D. macrophyllum subsp. macrophyllum l'. Plants bearing | to 3 capitula and 2 or 3 cauline leaves (including bract-like leaves) 19 ». D. macrophyllum subsp. unc 19a. Doronicum macrophyllum subsp. crophyllum ma- Doronicum mac rolepis 154 & Sint., in Freyn, Bull. Herb. oissier 3: 351. . Syn. nov. TYP E: Turkey. Gü- müsc s in valle Bojükde re tractu Karagülldagh supra Artabir, P. posa 7173 (lectotype, eine ad by Edmondson 1975: 142), G!; BM!, ~ isotypes, B!, Volume 90, Number 3 Álvarez Fernández 369 2003 Doronicum (Asteraceae) Figure 26. A-D. Doronicum macrophyllum subsp. Me ics o re. from Hohenacker s.n., K). —A. Capitulum. — B. Phyllary. —C. Indumentum of a phyllary. —D. Ray flower. E-H. Doronicum ber qunm (Зена from Bornmiiller 9620, B). —E. Capitulum. —F. Phyllary. —G. Indumentum of à phyllary. a Ray flow 370 Annals of the Missouri Botanical Garden pi E not seen, К!, LD! photograph, P not seen, Annuaire Conserv. Jard. Bot. . TYPE: Turkey. La- Balansa s.n. (lec- eda LM Cavill., enève 14: 260. 19 саа près de Djimil, . B. totype, designated by Tragen (1975: 140), G!; isotype, W!). . Syn. nov Rhizome woody. Stem with more than 3 cauline leaves. Plants generally bearing more than 3 capit- ula, which are sometimes turbinate at the base (1.3 cm wide) in fruit. Chromosome number unknown. Illustrations. Avetisyan & Oganesyan (1995: tab. 174); Figure 26A-D. Distribution. Northern Turkey and Caucasus. Growing in woods, open moist rocky places, mead- ows, and near watercourses, altitude 1500-3700 m (Fig. 24). All the specimens from the only collection of Do- ronicum macrolepis differ from D. macrophyllum subsp. macrophyllum in size of capitula. Although the population is somewhat anomalous, the name is treated as а synonym of D. macrophyllum subsp. macrophyllum. The diagnostic characters used to separate D. balansae (Cavillier, 1911) overlap sub- stantially with those of D. macrophyllum subsp. macrophyllum. Besides, the patterns of indumen- tum variability match those of D. macrophyllum subsp. macrophyllum. Accordingly, this name is considered a synonym. Selected ко е roi CAUCASUS. North Caucasus: Dagestan, Dargi, Maara, Akuscha, Alexcenko 12861 (LE); Shar roj, Ыы Serchikhi, Averianov et (LE); Sc i «bur. dagh, ry stania, Becker 132 (LE); Digoriya, Tators, Digor-Tors, Aug. | E. & N. Busch s.n. (LE); Stavropolskij, С 'haevo- Cherkesskaya, Pastvishnogo, Geltman et hs 1179 ; Checheno- In- ‚ Ge feltman et га. 2358 idi ansc im , Didi = hva, A. Н. & V HU : Ario su, E. & N. Busch 47 a Teberda, 18 June 1968, Ehi vald & es s.n. (В); € ne Mts. Tzkura-Tzkharo, 9 July 1923, Juzepezuk s LE); Azer- bajdzhan, Kuba, Leze, Schach-dagh, 8 a 1935, Karja- gin s.n. (NY); Azerbayan, Baku, Gandzha, Rashnar-dagh, 30 June 1929, iul e s.n. (LE); Aze rbayan, Karabach, Lyzagorsk, 30 June 9. Kolakovsky s.n. (LE); Armenia, Daratschitschach, adis 400 (LE); Tiflis, Mt. Saa hire, 6 July 1919, Schischkin s.n. (BM): Svanetia, sepes inter flumina Hippum et Ingur, Sommier & Levi 09 (G Gruzniskaya, ierit чн) ka, Tsve let y Cher- epanov 1032 (LE); Goris, . Karabakhskoie, Brun, 27 July 1975, Vašák s.n. (B ч "Té flis, Wittmann 294 (LE). IRAN. M Qareh Dagh, Aliabad, Lamond & Ter- meh 4876 (E, IRAN, K, W). TURKEY. Artvin: Yu- sufeli, iT Каска Daglari, Aytag 2933 (GAZI). Kars: Ardahan, seis & Buchner 82-94-38 (W). Rize: Ikizdere, Ballikóy er, Cevresi, Sulak, Giiner & Vural 5974 (GAZI); Camlihemsin. Yukari Атаки, Cayirlik, Cokyillik, Güner & Vural 6115 (GAZI); Ikizdere, Gólyay- la-Cihantepe, Giiner & Vural 6643 (GAZI). Trabzon: So- ganli Daglari, Bayburt, Edmondson 851 (E); Zigana Pas- W shóhe, Sorger & Buchner 82-89-3 (W). 19b. Doronicum macrophyllum subsp. spar- sipilosum (J. R. Edm.) Alv. Fern., Novon 11: 295, 2001. Doronicum bithynicum subsp. spar- sipilosum J. udm., Notes Roy. Bot. Gard. Edinburgh 32: 258. 1973. TYPE: Turkey. “Il- gaz Daglari, 35 km S of Kastamonu, roadside on N side of pass top,” 27 July 1971, J. R Edmondson 463 (holotype, E!; isotypes, GI. А ISTF not seen, K!, W!). Plant up to 90 cm tall. Rhizomes woody to some- what fleshy. Stems with 2 or 3 cauline leaves. Blade of basal leaves 6.5-13.5 X 7-11 ст. Blade of lower cauline leaves 7.5-9.6 X 9.3-12 cm. Middle and upper cauline leaves 4.5-11 X 3-6 cm. Capitula 2 or 3, ca. 6.5 em diam. including rays; involucre ca. 3.5 em diam.; peduncles up to 16 cm long. Receptacles glabrous. Ray flower corollas 1.9—3.5 cm long. 1.7—4 mm wide. Disk flower corollas 3.5— 4.5 mm long. Pappus 1.5-3.5 mm (on immature ovaries). Chromosome number unknown. Distribution. Northern Turkey (provinces of Bolu and Kastamonu). Growing in woods and open moist rocky places, altitude 1700-2200 m (Fig. 24) Selected са low ect TURKEY. Bolu: Ala dag, Kartal Kaya, s & Coode 37370 (E, K). Kasta- monu: llgaz Dag. е et al. 38312 (Е, К); Ilgaz, Ka- rakol, Cankiri, Nydegger 19037 (С). This subspecies combines characters from two species of Doronicum. Rhizome, phyllary shape, and indumentum are similar to D. macrophyllum subsp. macrophyllum, while size and leaf arrange- ment are similar to D. orientale. It was described as a subspecies of D. bithynicum J. R. Edm., which is considered a synonym of D. reticulatum Boiss., but the diagnostic character for this species (ovate phyllaries with a dark-colored major venation and a long tapered-acute apex) does not occur in sub- species sparsipilosum. The subspecies has the phyl- lary type and other characters of D. macrophyllum subsp. macrophyllum, from which it differs mainly in having fewer capitula and in leaf arrangement, and there are intermediate specimens. Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) 20. Doronicum maximum Boiss. & A. Huet, in Boiss., Diagn. Pl. Orient. ser. 2, 3: 31. 1856. TYPE: Turkey. Tech-Dagh, A. Huet du Pavil- lon s.n. (lectotype, designated by Álvarez Fer- nández & Nieto Feliner (1999: 805), G-BOIS!, photograph; isotypes, BM!, С!, Plant up to 100(+) cm tall. Rhizomes woody, gla- brous, with or without leaf remains. Stems branched in the upper part, leaves distributed along the stem, upper internodes generally longer than the adjacent leaves. Indumentum of multiseriate and uniseriate eglandular trichomes (ca. 0.2 mm) and glandular trichomes (up to 1.7 mm), scattered, sometimes gla- rous. Leaves entire to dentate. Basal leaves some- times present at flowering time; blade 11-20 х 10— 21.5 ст, orbicular or ovate, with cordate base and generally blunt apex, with actinodromous venation; petiole 8-36 cm long, with sheathing base, sheath ca. 3 em long. Lower cauline leaves with blade 4.5— 17 X 5.5-24 ст; petiole 6.5—40 cm long. 2-3.5 mm wide, similar to basal leaves. Middle cauline leaves 3-13 х 2-14 semi-amplexicaul. Upper cauline leaves 1.2—5.5 1-3.5 to obovate, sometimes bract-like. Indumentum sim- cm, sessile, fiddle- shape d, э X cm, similar to middle cauline leaves or ovate ilar to the adjacent part of the stem, generally scarce, more abundant on margins of leaves. Ca- риша З to 18(+), : involucre shorter than rays, 1.5-2 ст diam., some- times turbinate at the base of capitula (ca. 5 mm cm diam. including rays; wide) in fruit. Phyllaries herbaceous, ovate-lance- olate to obovate-lanceolate and acute, glabrous or glabrate; the outer 0.7-1.2 cm long, 2-2.3 mm wide; the inner 0.9-1.2 cm long, 1.5-2.3 mm wide. Receptacles glabrous. Flowers with yellow corollas. ay flower corollas 1.2-2 cm long, 1.5-3 mm wide, obovate-elliptic to oblong-elliptic, apex generally with 3 teeth. Di Cypselae brown, with smooth to slightly grooved isk flower corollas x 1.5 mm. surface, dimorphic. Cypselae from ray flowers 2. 3.3 X 1-1.3 mm, glabrous or glabrate, without pap- pus. Cypselae from disk flowers 1.5-2.3 X 0.5 mm, sometimes with eglandular trichomes: pappus 3-5 mm, white. Chromosome number unknown. Illustrations. Figures 8A, 23A—E. Distribution. Eastern Turkey, south of Caucasus and south of the Caspian Sea. Moist rocky places and near watercourses, altitude 1700-3300 m (Fig. 24) Doronicum maximum is morphologically close to those southwestern Asian species with a “тасго- phyllum” type of habit, and the type of indumentum is the only character to distinguish among them. Within this “macrophyllum” group it is more sim- ilar to D. haussknechtii than to any other (see com- ments on D. haussknechtii), but the almost absolute absence of indumentum at the base of the capitula in D. maximum makes it different (Fig. 23A—F). Selected specimens examined. IRAN. Mt. Elvend, Pa- bot 1717 (G); Azerbaijan, Chalil Kuh, Selvana, Renz 48989 (E, G, W); Kurdistania, Mt. Takht- йан, June 1898, Strauss s.n. (B); Azerbaidjan, Rezaich, Silvaneh, 24 June 1970, Termeh s.n. (W). IRAQ. Arl Gird Dagh, Algurd Dagh, Rawandiz, Guest & Ludlow-Hewitt 2928 (К). TUR- KEY. Bitlis: Tatvan, Sorger 64-41-5 (W). Erzurum: Mt. Tech-Dagh, July 1853, Huet du Pavillon s.n. (BM, С, К); Bachrunse, Karliova-Cat, Karliova, Nydegger 17340 (С). Giresun: Balabandaglari, Kilinc Tepe, Tamdere, Davis et al. 20588 (BM, E, K). Hakkari: Cilo Tepe, Cilo yayla, Davis & Polunin 24113 (BM, E). Maras: Gaglayancerit, Engizek Dagi, Duman 4068 (GAZI). Mus: vallis ce Sauk, ha gee Gumgum, Warto, Kotschy 363 (B. G S, UPS, W) 21. Doronicum oblongifolium DC., Prodr. 6: 321. 1838. TYPE: “Doronicum plantagineum, En. pl. cauc. n? 674," 1832, C. A. Meyer (lec- totype. designated by Álvarez Fernández & Nieto Feliner (1999: 805), G-DC!, graph). photo- Plant up to 50 cm tall. Rhizomes woody to some- what woody, glabrous, generally with leaf remains forming fibers or dark scales. Stems not branched, leaves mainly in the lower middle of the stem. In- dumentum of white eglandular trichomes (up to 2.5 mm), more abundant near the capitula, also with scarce glandular trichomes, sometimes glabrous at the base. Leaves entire to slightly dentate. Basal leaves generally present at flowering time; blade (1.8)2-6 x (0.9)1.5—3 cm, elliptic, with attenuate base, and generally blunt apex, with actinodromous to pinnate-actinodromous venation; petiole 3—10 cm long, 1-3 mm wide. Lower and middle cauline leaves 3.5-8(9.5) X 1.4-2.5 cm, similar to basal leaves or sessile, elliptic to ovate-elliptic, some- times widely ovate to suborbiculate, semi-amplex- icaul, with blunt apex. Upper cauline leaves (1.6)3—6 х (0.2)1-2 em, similar to middle cauline leaves or ovate-lanceolate, sometimes bract-like. Indumentum similar to the adjacent part of stem, sometimes with white, tangled, uniseriate eglan- dular trichomes (up to 1 mm), more abundant on leaf margins. Capitula solitary, 4.5—7.5 cm diam. including rays; involucre shorter than rays, 2-5 ст diam. 1-1.5(2) em long, 2.5-5 mm wide, ovate-lanceolate to elliptic. Indu- art of stem, some- Phyllaries herbaceous, mentum similar to the upper times abundant. Receptacles glabrous. Flowers with yellow corollas. Ray flower corollas 2-3.5 cm long, 372 Annals of the Missouri Botanical Garden 2.5-5.5 mm wide, obovate-elliptic, apex with 2 or 3 teeth. Disk flower corollas 4—5 mm long. Cypselae brown, with striate-reticulate to warty surface, di- morphic, ca. 4 X 1 mm, generally glabrous, some- times with eglandular trichomes or glabrate. Cyp- selae from ray flowers without pappus. Pappus up to 4.5 mm, white. Chromosome number 2n = 60* (Davlianidze, 1985; *Fedorov, 1969). Illustrations. Avetisyan & Oganesyan (1995: tab. 175); Figure 3I, J. Distribution, Caucasus. Open moist rocky plac- es, s "on watercourses, altitude 1400-3900 m (Fig. sum онай is distinctive among the species from Caucasus. While the rest of the species in this area (exc ent D. orientale) have the “macrophyllum” type of habit, D. oblongifolium bears only one capitulum and has elliptic basal leaves making it similar in habit to other European or central Asian species (e.g., D. clusii, D. hun- garicum, D. falconeri, among others). Besides, the type of rhizome (woody, glabrous, and with fibrous leaf remains) is quite different from that of D. or- ientale (fleshy and with pubescent nodes). In ad- dition, D. oblongifolium has a dumentum on margins of basal leaves (Fig. 31, J). The citation of the chromosome number 2n — 60 special type of in- for Doronicum oblongifolium was found in Fedo- rov's index (1969), but the original source for this data was not seen. Selected specimens examined. CAUCASUS. North 1, Kaitag, Tabassaran, Urgah, Dshufu 13586 (LE); Checheno-In- ‚ Ave rianov a al. 242 & N. A. Busch s.n. (LE); Kaepe . June DM Кынай s.n. E. Digoria, rode tia, es Ruprecht 156 (LE); Tindal, Mts. жоне Aatschabala, 10 e 1861, Ruprecht s.n. . Transcaucasus: Azerbajdz n an, Mt. ord s 15 Aus. 1929, Achverdov & Vcg s.n. (LE); Carthali- nia, Zhra Zhraras, A. H. & : Brotherus 501 (BM, ( S); Armenia, Alügez, 20 po 1932 & (LE); Azerbajdzhan, Gandzha, 1928, Vig s.n. (LE); Aragac, lacum Kari, اا‎ lian 12787 ; Aze srbayan, Nac shitshevan, Zang, inter weis d et Kjavin-Kaja, 1 July 1928, Garr s.n. 2 Sarial, May 1838, Hohenacker s.n. (B, E, si isi Daralogez, ih 25 2, d i jagin & - Saflev s.n. (LE, S); Azerbayan, Nagornogo Kara- bacha, Gadrutskij, Znarat, 27 May 1948, Ki mich & of (LE); Azerbajdzhan, Karabach, inter Lysagorse e 1929, ——— s.n. (LE); Tiflis, en . 7 Aug. 1928, Kozlov НА s.n. (LE); Geor- „: Bakuriani, 9) Мау 19 Kozlovsky s.n. a, Chevsuretiya, о Choki Aug. 1982, Menitskij s.n. (LE); Azerbayan, Nachitsc hu Mt. Agdaban, 17 July 1934, Prilipko & Isaev s.n. (LE) Armenia, Mt. Alajos, Radle 142 (LE); Murov-dag, Giam- nko & Pomor tr] I = Koshkardagh, T — 4 T = ish, Elisabetholsk, 15 July 1909, энеми г s.n. (LE); ts. Areguni, ad lacum Sevan, Kras ‚ Shorsha, 8 Oct. 1974, Vasák s.n. (B, G); ak, ME Aragac, 13 \ ( s.n. (С, №); Armenia, Razdan, Mt. Ali- beg, Cakhkdzor, 4 July 1982, Vasdk s.n. (W). TURKEY. Artvin: Yalnizcam Silsilesi, Savsat, Albury et al. 3176 (K); Çoruh, Ardanuç, Kordevan dag, Yalnizcam Daglari, Davis & Hedge 30365 (E, K Erzurum: Dumluda, Sorger & Buchner 82-123-9 (W). 22. Doronicum orientale Hoffm., Com. Soc. Phys. Med. Moscou 1: 8. 1808. TYPE: not lo- cated; protologue citation: "Habitat passim cir- ca Zehet in Iberia." Doronicum caucasicum M. Bieb., Fl. Taur.-Caucas. 2: 321. 808. TYPE: *ex Caucaso iberico. Adam [Adam.? s.n.], ex herb. Marschall von Bichorsinin (lectotype, aad here, LE!). Plant up to 140 em tall. Rhizomes fleshy, pubes- cent to very pubescent, with shining white-tinted trichomes on nodes, stoloniform, sometimes with buds. Stems unbranched, scape-like. Indumentum of uniseriate and multiseriate eglandular trichomes and short-stalked glandular trichomes. Leaves en- tire to slightly dentate. Basal leaves sometimes pre- sent at flowering time; blade (2)4—7(8.5) x 2(3)- 6(7.5) cm, reniform to widely ovate with cordate base and generally blunt apex, with acrodromous venation; petiole (1.8)4—10(20) em long. (0.5)1— 2(3.5) mm wide. Lower cauline leaves (3.2)5—7.5(9) X (1.8)3—5(7.7) em, similar to basal leaves or ses- sile, fiddle-shaped, semi-amplexicaul. Upper cau- 4-7(10.5) х (0.4)3-6(7.6) cm, ovate-elliptic to ovate-lanceolate, sometimes bract- like. Indumentum mainly of short-stalked glandular line leaves (1. also with uniseriate and multiseriate 3(4)-(6 em diam. including rays; involucre shorter than rays or equaling them, (2.3)3— Phyllaries with acute apex; the outer (1)1.5-2(2.5) em long, (1)1.5-2 (3) mm wide: the inner 1—1.5(2) em long, 0.5-1.5(2) mm wide. Margins ciliate, with acute, stiff and equidistant multiseriate eglandular tri- chomes (up to 1.5 mm). Indumentum mainly glan- trichomes, eglandular trichomes. Capitula 1(2 or 3), 5(5.5) mm diam. herbaceous, ovate-subulate, generally dular, but also with eglandular trichomes. Recep- tacles generally pubescent, sometimes glabrous. Flowers with yellow corollas. Ray flower corollas (1.4)2-2.5(3) em long, 2-2.5(3.8) mm wide, oblong- elliptic, apex generally with 3 teeth. Disk flower corollas 3.5—4(5) X 1.2-1.5 mm. Cypselae olive- green or brown, with warty or rugose-reticulate sur- face, dimorphic. Cypselae from ray flowers (1)1.5— 2(2.3) х (0.2)0.5-0.8(1. pappus. Cypselae from disk flowers (1)1.3—1.5(1.8) —0.7(1) mm, with eglandular trichomes; pap- 7) mm, glabrous, without Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) EN ай - i : | 327 10 20 Figure 27. Distribution map for Doronicum orientale. pus (2.5)3-5 mm, white. Chromosome number 2n = 60 (Lindqvist, 1950, as D. cordatum; Baksay, 1956; Strid & Anderson, 1985). Illustrations. Sadler in Nendtvich (1836: tab. 2): Sávulescu (1964: pl. 99, fig. 2); Figures 1A, 5E-H. Distribution. Eastern Mediterranean region, from Syria to Sicily, and Caucasus (absent in north- ern Africa?). Cultivated and naturalized in central Europe. Forests, meadows, rocky places, and shady gullies, from sea level up to 2000 m in elevation (Fig. 27) Both Doronicum orientale Hoffm. and D. caucas- icum M. Bieb. were described іп 1808 with the month of publication unknown. One year later the former name was choosen by Willdenow (1809) as the valid name, and, hence, it is the correct name (ICBN, Art. 11.5, Greuter et al., 2000). The type material of D. orientale has so far not been located, and although its description in the protologue matches the taxonomic identity recognized as D. orientale by Willdenow (1809), its formal identity needs to wait until a lectotype is designated. Since both thors in floristic studies, the lectotype for D. cau- The only sheet found in Marschall von Bieberstein's herbar- names were alternately used by different au- casicum M. Bieb. is designated above. ium that matches the protologue was selected as the lectotype. The possible occurrence of Doronicum orientale in North Africa requires further work (see com- ments for D. pardalianches and D. plantagineum). Selected specimens examined. LEBANON-SYRIA. hoe agri Byzantini, J. & К. B ош 11965 (B); Far- b 5332 (B, E, K). CA casus: Kuban, Busch & Klopotov "t (L E, NY): bb Gelendzhikskii, Pschadi, Pschala, (S oia et al. 2360 (LE). Transcaucasus: kn ad rivulum Kura, 22 Mar. 1861, Ruprecht s.n. (LE). TURKEY. Adana: Bah- се, Dildil dag, Haruniye, Davis & M 26107 D (BM, К). Afyon: Darkiri, Hisaralan, Aytaç 1167 (GAZI). ponen Direkli, Üçoluk, Peker 1178 (GAZI). Antalya: Termessos, C. & M. North 42 (E). Aydin: Karacasu, Baba Dag, Seki, Davis 41627 (E, K). Balikesir: Kaz Dag. Sar- ied Dudley 34813 (E, (K). Bursa: Kebabdere, Karmis, Coode 3368 (E). Жым "Paphlagonis, ' Tesis a drei azdagh, Sintenis 3995 (B, BM, BRNM 1 1 Çiçi , Coode & Jones 2733 (E. К). Mugla: Inceali E. Walter 342 (B). Trabzon: Trabzon, Stainton 8113 (E, K). Yozgat: Akdagmadeni, Aktas, Cur- tis 144 (E). е Niederósterreich: Theresienthal bei Gratzen, 19 May 1899, dein s.n. (B, 5. BULGARIA. Plovdiv: Stanima ka, 10 June 1889, Střibrný s.n. (G). GREECE. ros: Arta, Заман, Willing 33327 (B). адно. Naussa, Mt. Vermion, К. Н. & F. Rechinger 8758 (BM). Nísoi Aiyaíou: Samos, Mt. Кай, Ара, Da- s 1654 К (E, К); Nenedes, cuia Samos, K. H. & F. pem 3794 (BM, K, G). Pelopónnisos: Lirkio, Ke- falovrison, García 953 (MA). Stereá Ellás-Évvoia: r4 tafigon, Oeta, Balls & Gourlay B3264 (BM, E, K); M Hymetti, Orphanides 196 (COI-WILLK, E). Thesealia: Pelion, Kissos, Beauverd 272 (G). HUNGARY. Baranya: Pécs, Mecsek, 25 Apr. 1946, Papp s.n. (5). Tolna: silva Gurovica prope Szekszárd, 13 May 1914, Hollós s.n. (5). LY. Basilicata: Potenza, in silva Pallareta, 20 Apr. 1924, Gavioli s.n. pies Calabria: Pizzo, 20 Apr. 1938, Lenander s.n. (S). Sicilia: Parco delle Madonie. Piano Sempria, Nieto 3888 (MA). YUGOSLAVIA. Makedonija: Doiran, Marianska planina, Hudowa, Bornmüller 4265 (B, NY). Srbija: Rakovika, prope Belgradum, Petrovič 2. 340 G, K, MA). — = 23. Doronicum pardalianches L., Sp. Pl.: 885. 1753. TYPE: Herb. Clifford, 411.1 [sine col- lector] (lectotype, designated by Llamas et al. in Jarvis & Turland (1998: 360), BM!). Annals of the Missouri Botanical Garden Sp. PI.: 884. ы Aster scorpioides (L.) Scop., Fl. Carniol. ed. 2, 2: 169. 1771. Gram- acea scorpioides (L.) Cass., in Cuvier, Dict. Sci. Nat. 19: 294. 1821. ied um scorpioides (L.) Rchb., Fl. Germ. Excurs. 1: . 1831-1832. T “Do- ronicum radice scorpii Wem hiata," Herb. Burser X: 16 [sine collector] (lectotype, designated by Álvarez in Jarvis & Turland (1998: 353), UPS!, photograph). Arnica sc ۰ L, Plant up to 150(+) ст tall. Rhizomes fleshy, pu- bescent to scarcely pubescent or glabrate with shin- ing white-tinted trichomes on nodes, stoloniform, sometimes with buds. Stems scarcely branched in the upper part, with few leaves mainly distributed along the basal % of the stem, internodes generally longer than the adjacent leaves. Indumentum of thin and acute multiseriate eglandular trichomes (up to 5 mm) in the lower part, uniseriate eglan- dular trichomes and glandular trichomes in the middle and upper part, abundant near the capitula. Leaves entire to slightly dentate. Basal leaves some- times present at flowering time; blade 3.6-16.5 х 3.3-14 em, ovate with cordate base and blunt apex, with acrodromous to actinodromous venation; peti- ole 4.5-27 em long, 1—4.5 mm wide. Lower cauline leaves 3.3-22 x 2.3-11 cm; petiole (3.4)6-10(27) em long, 1-1.5 mm wide, similar to basal leaves. Middle cauline leaves (2.7)5—9(15.3) X (1.6)3— 6(10) cm, sessile, fiddle-shaped, semi-amplexicaul. Upper cauline leaves (1)2—6(10) X (0.2)1—2(5.5) ст, ovate-elliptic to ovate-lanceolate, sometimes bract-like. Indumentum similar to the adjacent part of the stem. Capitula (1)2—7, 2-5.1 ст diam. in- cluding rays; involucre almost equaling rays, some- times exceeding them, 1—3.3 em diam.; peduncles (0.5)5—7(20) em long, (0.5)0.8-1(2) mm wide. Phyl- laries herbaceous, (1)1.2—1.4(1.7) em long, (0.7)1— 1.5(2.7) mm wide, ovate-subulate, generally with acute apex. Margins sometimes ciliate, with acute, stiff and equidistant multiseriate eglandular tri- chomes. Indumentum of glandular and eglandular trichomes. Receptacles pubescent or glabrate. Flow- ers with yellow corollas. Ray flower corollas 1.1— 2.5 ст long, 2-3.5 mm wide, oblong-elliptic, apex generally with 3 teeth. Disk flower corollas 4—6 X 1-2.5 mm. Cypselae black and with warty surface in maturity, dimorphic. Cypselae from ray flowers 1.7-3.5 X 0.7-1.3 mm, glabrous, without pappus. Cypselae from disk flowers 1.2-1.8 X 0.7-1 mm, with eglandular trichomes; pappus (2.5)3—4 mm, white. Chromosome number 2n 60, 120* (Lindqvist, 1950; *Moore, 1982, see comments be- ow) Illustrations. Jacquin (1776: t. 350); Reichen- bach (1854: t. 64, fig. 2); Sávulescu (1964: pl. 97, fig. 2: pl. 189, fig. 1); Bolds & Vigo (1995: 839); Figures 2A, B, 7A Distribution. Northeastern Iberian peninsula and central Europe. Cultivated and naturalized at least in Great Britain and Northern Europe, so that the limits of the natural distribution are uncertain. Forests, meadows, hedges, and near watercourses, from sea level up to 1800 m in elevation (Fig. 28). As indicated above, in the protologue of Arnica scorpioides L., several pre-Linnaean synonyms are included. This name has been treated as a synonym of Doronicum grandiflorum Lam. by all the authors that combined it, probably because Jacquin (1776: 26, t. 349) illustrated it with a plant of D. grandi- florum Lam. However, the lectotype designated by Álvarez in Jarvis and Turland (1998) represents Doronicum pardalianches L., since all of the origi- nal elements of A. scorpioides belong to D. parda- lianches L. Formally, all the combinations based on Arnica scorpioides L. are homotypic synonyms of Arnica scorpioides L. and thus synonyms of D. par- dalianches L., even though the descriptions and ref- erences in protologues correspond mainly to Doron- icum grandiflorum Lam. (see also comments above on D. under D. grandiflorum Lam.). Desfontaines (1798) cited Doronicum parda- lianches in North Africa: “in cacumis Atlantis pro- pe Belide,” but all the specimens from North Africa represent D. plantagineum. A few populations have broadly ovate basal leaves with subcordate bases that are similar to D. orientale, so their identity is uncertain. Doronicum pardalianches has similar basal leaves, but is quite different from both D. orientale and D. plantagineum in habit, number of capitula, number of cauline leaves, and color of cypselae. Although no sheet from Desfontaines's lo- cality was seen, the presence of D. pardalianches in North Africa is unlikely. Desfontaines's descrip- tion matches D. plantagineum, or even D. orientale whose presence in North Africa is questionable). Hybridization between D. plantagineum and D. or- ientale in this area is a possibility. (See comments for D. orientale and D. plantagineum.) Determining the native distribution of Doronicum pardalianches is difficult. Records are scattered in central Europe, but absent in the Iberian peninsula, except for northeastern Spain, where it is notably abundant, exactly in the gap presented by D. plan- tagineum (Figs. 25, 28). This suggests these species do not overlap in their presumably natural areas of distribution, and that the native area of D. parda- lianches reaches southwestern Europe in northeast- scorpioides Lam. ern Spain. Volume 90, Number 3 2003 Álvarez Fernández 375 Doronicum (Asteraceae) 60 50 40 10 Figure 28. Distribution map for Doronicum pardalianches. The citation of the chromosome number 2n — 120 for Doronicum pardalianches was found in Moore's (1982) this data was not seen. index, but the original source for Selected. specimens Vp det ANDORRA. Sant Mi- quel d'Engolasters, 'aldes, Almaraz et al. 1015 (MA). AUSTRIA. Niederdsterreich: Sachsen, Schlop- berg Hartenstein, May 1904. ) BEL- GIUM. Hainaut: Marchienne-au- ү ета 76 B 325 (MA). FRANCE. Basses-Alpe co près de Digne. 2 June 1808, pir s.n. (K). Cote d'Or: bois de Saulon, 19 June 1873, Bonnet sn. ius Doubs: bois de Chátay à Uzelle, Paillot 2279" (B. G, NY). Haute- Savoie: St. Pierre-de- Rumilly a St. desde түк 3459 (С). Наше s-Alpes: Séuse, 23 July 1885, Girod s. (G). Hautes-Pyrénées: bois de Sia pres Luz. uly 1872. Borderè s.n. (K) Isere: Grenoble, 2 June 1850, Chabert Mine 120 (MA). hg rand 43, Handris s.n. (С). S 31 ae oo Grandmaison s.n. O. Seine-et- Oise: "Paris, Duby 263 (MA). GERMANY. Bay- ern: Steiniger Wald, Baumberg, 20 June 1908, Harz s.n. erlin: i Berolinum, Borussia. 10 July 1887, рання s.n. (С). Halle: Winningen, 1 June 1857, Wirtgen s.n. (B, G). Hessen: T. Rheinland-Pfalz: ien Birtgen, Lechler 48 (B, GREAT BRITAIN. England: Ledbury. gens E (К); Woodchester Park, West Glouc ester, Lousle 4 (K); Bucks, between Marlow and Medneuham, (ж 37370 (К). Scotland: Inverness, Edinburgh, Syme 653 (K). IT- ALY. Lombardia: Mt. Bronzone sur Tavernola, 16 May 1910, Wilczek s.n. (G). Piemonte: Turin, Val Salice, 5 May 1870, Joad s.n. (Е). NORWAY. Bergen, Fredholm 1189 (NY). SPAIN. Aragón: Huesca, sierra de Guara, Nocito, barranco Fuente Espátula, Álvarez et al. 801 (MA); Patiles, 2 Aug. 1988, Aseginolaza & Gómez royo de la Vena, collado de Basses, Almaraz et al. 1008 (MA); Lérida, Alta Ribagorza, bosque de Besiberri, 10 Aug. 1987, Arán & Tohá s.n. (MA); V d Опе Е Саѕаѕ 605 (МА); Сегопа, San Fe > Pallar- ols à la ) . (K). Vaud: Graugette prope Pudet. 28 May 1870, Faurat s.n. (BM). . Doronicum plantagineum L., Sp. Pl.: 885. 153. TYPE: Herb. Clifford. С 2 ы col- lector] (lectotype, designated by Llamas et al. in Jarvis & Turland (1998: 360), BM!). Plant up to 150 cm tall. Rhizomes fleshy, pubes- cent to very pubescent, with shining white-tinted trichomes on nodes, stoloniform, sometimes with buds. Stems generally unbranched, scape-like. In- dumentum mainly glandular, with short-stalked and long-stalked glandular trichomes (up to 0.7 mm), sometimes also uniseriate and multiseriate eglan- dular trichomes, more abundant near the capitula. Leaves entire to slightly dentate. Basal leaves some- 376 Annals of the Missouri Botanical Garden times present at flowering time; blade (2.5)4—8(12) X (1.5)2.5-6(9.5) cm, ovate with truncate, attenu- ate or subcordate base, blunt or somewhat acute petiole (2)4— 8(19) cm long, (0.5)1.5-3(7) mm wide. Lower cau- line leaves (1.5)3—7(19) X (1)2—5(8.5) em, similar to basal leaves or sessile, fiddle-shaped to ovate- apex, with acrodromous venation; elliptic, semi-amplexicaul. Upper cauline leaves (1)2-4(9.5) х (0.1)0.7—2(5.5) ст, ovate-elliptic to ovate-lanceolate, sometimes bract-like. tum mainly glandular, with short-stalked and long- stalked glandular trichomes, also with uniseriate (0.2 mm), and multiseriate (up to 2 mm) eglandular trichomes. Capitula 1(2 or 3), 3—4(6.5) em diam. including rays; Indumen- involucre almost equaling rays, sometimes exceeding them, 3—4(5.5) em diam. Phyllaries ovate-subulate, generally with acute apex; the outer (1)1.5-2.5(3) cm long, (1)1.3-2(3) mm wide; the inner (1)1.5-2(2.5) em long, 0.5—0.7(2) mm wide. acute, stiff and equidistant multiseriate eglandular trichomes (up to 1.5 mm). herbaceous, Margins ciliate, with Indumentum mainly glandular. Receptacles glabrous or scarcely pubes- cent. Flowers with yellow corollas. Ray flower co- rollas (1.1)1.5-2.5(3) em long, (1.5)-2(3) mm wide, oblong-elliptic, apex ETN with 3 teeth. Disk ) X (1)1.3-1.5(2) mm. Cypselae olive-green or HA with warty surface, dimorphic. : (0.7 flower corollas 4(4.3)—4.5(: Cypselae from ray flowers 2—2.8(4) X —1.3 mm, generally glabrous, without pappus. Cypselae from disk flowers (1.5)2—2.7(3) X (0.7)1— 1.3 mm, with eglandular trichomes; pappus (3)3.5— 4.5(5) mm, white. Chromosome number 2n = 120 (Lindqvist, 1950; Fernandes & Queirós, 1971; Live & Kjellqvist, 1974; Ruiz de Clavijo, 1993). Illustrations. he h (1854: tab. 65, fig. 2); Hegi (1928: 711, fig. 420); Valdés et al. (1987: 77); Bolds & Vigo (1995: 838): Figure 6A, B. Distribution. Southwestern Europe (Portugal and Spain) and northern Africa (Morocco and Al- geria). Cultivated and naturalized in Great Britain and central Europe. Limits of its native range un- certain. Forests, meadows, hedges, and on shady moist rocky places, altitude 400-2200 m (Fig. 25). Doronicum plantagineum is variable for some characters (e.g.. size and robustness of the plants, size and shape of basal leaves, number of leaves and capitula, type and abundance of indumentum). Cultivated, naturalized plants, and a few natural populations tend to have basal leaves broadly ovate-elliptic to elliptic, with slightly dentate mar- gins, attenuate bases and somewhat acute apices. Some authors (Rouy, 1893, 1903a, 1903b; Legrand, 1894; Coutinho, 1939; Fournier, 1939) have given taxonomic recognition to these trends. The tinctive, generally being more robust and pubescent than the European, and having broadly ovate basal North African populations are the most dis- leaves with subcordate bases. The shape of the bas- al leaves in these plants does not allow a clear distinction between Doronicum plantagineum and D. orientale. These North African populations have been treated as subspecies or varieties of D. pa dalianches or D. plantagineum (Chabert, 1892; Barratte, 1893), or as a separate species, D. atlan- ticum (Chabert, 1891; Rouy, 1893), as in Cavillier's monograph (1907, 1911). A multivariate morpho- metric analysis (РСА and DA) of Doronicum (А1- varez Fernández & Nieto Feliner, 2001) reveals no morphometric support for the segregation of these populations as a species from the European popu- lations of D. plantagineum. On the other hand, a phylogenetic analysis based on morphological, nu- clear ribosomal (ITS), and chloroplast (trnL-F) data (Alvarez Fernández et al., 2001) showed differenc- es in ITS sequences between these populations that somewhat support separate species status despite the poor morphological differentiation. Introgres- sion from D. orientale into populations of D. plan- tagineum is not ruled out as the cause of sequence differences. Until further work is done, these pop- ulations are provisionally included in D. plantagi- neum m The bulk of the records of Doronicum plantagi- neum are from the Iberian peninsula; a gap occurs in northeastern Spain and southern. France, and most of the French records are from near Paris. This is the only representative of the genus in North Africa (northern Algeria and Atlas) and thus seems to be native to the Iberian peninsula and North Africa, its current area of distribution being ex- panded by human action. Selected ок examined. ALA Had, pic des Cédres, Alston Simpson E Tala Guile Boghni, Davis кч ‚ E); Ka- b к Magris, Reverchon 391 (BM, С). MOOR Ifrane, Azrou, rd 10476 (BM); Ain Leuh, Jahandiez 5386 (B, MA). FR е Louze, Brienne-le Chateau, Retz 8 5 (G. . MACB, MAF). айна Илк: тал гн а» aubois 2872 (BR, MA). Seine-et-Oise: Port-Villez, 11 May 1873, Delacour s.n. (K); forêt de Bondy, Paris, 15 May 1846, Kralik s (K); Yvelines, Véthenil, bois du Coudray, Lawalrée 15708 BH); Verrier go Buisson, Essone, forêt de Verrières, Retz 67398 (BR, AF). Somme: dd de Lize près п каг РА Brutelette s.n. (G). Var: ia Ma iret, у 1884, Leresche s.n. (В). GREAT BRITA с y аю Saling, Essex, Fox 780 (B, W); tham, 14 June 1887, Woodward s.n. (K). Scotland: Blair, Culross, 18 Apr. 1872, Pamoa s.n. (K); Amiston ERIA. Теше! el y 36 (BM); yo Volume 90, Number 3 2003 Alvarez Fernández Doronicum (Asteraceae) Wood, Edinburgh, Syme 654 (G). PORTUGAL. Algarve: entre Monchique e Alferce, Malato-Beliz et al. 3120 (MA). Alto Alentejo: Castelo de Vide, Amieira, Malato-Beliz 196 (MA). Beira Alta: serra da Estrela inter ыруг Че Manteigas et Poco do Inferno, — et al. 4844 (MA). Beira Baixa: Fundao, 25 Mar. 926, Camis & Mend оса ‚ Rainha ntes: С. & J. Vasconcellos 7877 (HVR). SPAIN. An- da luin: pn sierra Morena, Despefiaperros, collado de los Jardines, Cuatrecasas 3526 (MAF); Cádiz, Grazalema, sierra del Pinar, López & Morales 3008 GF (MA). Aragón: Zaragoza, entre ah et Tobed, sierra de Algairén, D et al. 63591 (G, MAF). Cantabria: Vega de Lié- bana, entre ide y Bárago, Álvarez et al. 966 (MA). Castilla-La Mancha: reir grin Retiendaá, hos del río Jarama, Álvarez et al. 954 MA (MA); о de Asturias: sn Belefio, 10 Apr. 1998, Medina s.n. ión de Murcia: Moratalla, sierra de La Muela, Álvarez et ys 1103 (MA). 25. Doronicum reticulatum Boiss., Diagn. Pl. Orient. ser. 1, 4: 12. 1844. TYPE: Turkey. Tmolus Bogdagh, Lydia, [Р E. Boissier] 3969 (lectotype, designated by Álvarez Fernández & Nieto Feliner (1999: 805), G-BOIS!). —— bithynicum J. R. Edm., Notes Roy. Bot. Gard. inburgh 32(2): 257. 1973. Syn. nov. TYPE: Tur- io Bursa, Olympi Bithyni, P. M. К. Aucher-Eloy 3269 (holotype, G!; isotypes, BM!, К!, К! Plant up to 80(+) cm tall. Rhizomes woody to somewhat woody, glabrous, sometimes with leaf re- mains forming scales on nodes. Stems branched in the upper part, leaves mainly distributed in the lower middle portion, upper internodes generally longer than the adjacent leaves. Indumentum of un- iseriate eglandular trichomes, rarely with a few multiseriate eglandular trichomes, sometimes with glandular trichomes, more abundant near the ca- pitula and sometimes glabrous at the base. Leaves entire or subentire. Basal leaves sometimes present at flowering time; blade 7.5-16.5 X 7—15 cm, or- bicular or ovate, with cordate base and blunt apex, with actinodromous venation; petiole 10—17.5 cm long, with sheathing base, sheath (1)3—5 cm long. Lower cauline leaves with blade 8-9.5 X 7—8 cm; petiole 10-13 em long, 1.5-2.5 mm wide, similar to basal leaves. Middle cauline leaves 6.3—10 X 2.6-7 cm, sessile, fiddle-shaped to obovate, semi- amplexicaul. Upper cauline leaves x l- 1.6 cm, sessile, ovate to obovate, or bract-like. In- dumentum similar to the adjacent part of the stem, sometimes abundant on veins on the abaxial sur- ace of leaves. Capitula 3 to 5, cm diam. including rays; involucre shorter than rays or equaling them, 3.5—4 cm diam.; peduncles up to 11 em. Phyllaries herbaceous, sometimes slightly papery, ovate-lanceolate with very tapering acute apex, generally with 8 to 12 longitudinal veins dark-colored; the outer 1.5-2 cm long, 2.2-5.5 mm wide; the inner 1.4—1.6 cm long, 1.2-3 mm wide. Indumentum of glandular trichomes. Receptacles glabrous. Flowers with yellow corollas. Ray flower corollas 1.7—3 cm long, 4—6 mm wide, obovate-el- liptic to oblong-elliptic, apex generally with 3 teeth. isk flower corollas -3 mm. Cypselae tinh with slightly grooved-reticulate surface, di- morphic. Cypselae from ray flowers 3.5 X 0.8 mm, glabrous or glabrate, without pappus. Cypselae from disk flowers 3 X 1 mm, sometimes with eglan- dular trichomes; pappus ca. 4 mm, white. (Com- pletely mature cypselae not seen.) Chromosome number unknown. Figures 7B, 26E-H. Illustrations. Distribution. Western ac (Bolu, Bursa, and Konya provinces). Grow in woods and open moist rocky places, idi 1800-2200 m (Fig. 1) — ` Doronicum reticulatum is morphologically simi- lar to those species with a “macrophyllum” habit in southwestern Asia, but it is quite distinctive be- cause of its unique type of phyllaries (Fig. 20E, F), which are ovate-lanceolate ending in a long taper- ing apex, and with 8 to 12 longitudinal veins dark- colored. Doronicum reticulatum grows only in west- ern Turkey where there is no overlap with any other species of the “macrophyllum” group, although it is geographically close to D. cacaliifolium. The only species that overlaps its area is D. orientale, which is morphologically quite different (i.e., habit, type of phyllaries, type of rhizome; Figs. 5E, F, 26K, F). See also comments on D. cacalüfolium and D. ma- crophyllum subsp. sparsipilosum. The name Doronicum bithynicum J. R. Edm. was given by Edmondson (1973) for a pro parte of the illegitimate name d thirkei Schultz Bip. ex Boiss. (Boissier, 1 79). Boissier's name includes his D. reticulatum МЫ (1844), which was collected in Tmolus Bogdagh, and also plants from Mt. Olym- pus in Bithynia. Edmondson (1973) considered these to be two different species, and he gave the name D. bithynicum for those plants from Mt. Olympus, reserving the name D. reticulatum for those from Tmolus Bogdagh. In this study, however. no diagnostic characters to separate D. bithynicum (Edmondson, 1973) and D. reticulatum were found. 378 Annals of the Ais Botanical Garden Accordingly, these names are here considered syn- onyms. Selected. specimens volets TURKEY. Bolu: Ala dag Karlalkoy, Alpay 2642 (E); Kóroglu, Buchner 83- 70- W); Ala dag. Капа! Kaya, Davis & Coode 37372 ^ Bursa: Uludag, Aytaç & Ekici 6229 (GAZI); U йар, А. : Baytop 20972 (E); Uludag, Bithyniae, vie 18013 (G); Mt. Olymp., July 1873, Pichler s.n. (К); vedere above Bursa, Polunin 15054 (E); Uludag, 1968, Sorger 68-53a-6 (W); Uludag bei Bursa, Н. & E. Walter 641 onya: Phrygia, Akscheher, Mt. Sultandagh, 9620 (B). Bornmüller 26. Doronicum stenoglossum Maxim., Bull. Acad. Imp. Sci. Saint-Petersbourg 27: 1881. TYPE: “China occidentalis, Regio Tan- gut (prov. Kansu)," 1880, . Przevalski s.n. (lectotype, designated by Álvarez Fernández & Nieto Feliner (1999: 805), K!). Doronic um souliei Cavill., T canis ipee Jard. eve 10: 235. 1907. Syn. . TYPE: China. us bet, Kiala, Tongolo, n A. Soulié 335 (lectotype, des- ignated here, G!; isotype, K!). Plant up to 120 ст tall. Rhizomes woody, gla- brous, without leaf remains, sometimes with adven- titious roots at the base of stem. Stems simple or branched, sometimes branched from the base, with leaves all along the stem. Indumentum of long- stalked glandular trichomes (0.5-1 mm), and some- times also eglandular blunt trichomes (ca. 1 mm), more abundant near the capitula, sometimes gla- brous to glabrate at the base. Leaves entire to very slightly dentate. Basal leaves absent at flowering time; blade ca. 6 X 3 cm, oblong elliptic, with at- tenuate base and blunt to subacute apex, with ac- tinodromous to petiole ca. 13 ст long, 2 mm wide. Lower and mid- pinnate-actinodromous venation; dle cauline leaves 3.5-15 X 1.1—6 cm, similar to basal leaves or sessile, almost fiddle-shaped to ovate-elliptic, semi-amplexicaul. Upper cauline leaves 2-6 X 0.5-2.5 cm, similar to middle cauline leaves, or ovate-lanceolate, sometimes bract-like. Indumentum scarce, mainly glandular, sometimes also eglandular blunt trichomes, more abundant on margins of upper leaves. Capitula 2 to 11, em diam. including rays and phyllaries; involucre longer than rays or equaling them. Phyllaries her- baceous, very narrowly triangular-subulate, almost linear, erect, acute; the outer 1.3-2.1 em long, 1— 2 mm wide; the inner 1.3-1.8 cm long, 0.7-1.5 mm wide. Indumentum only present in the lower half, similar to the upper part of stem, absolutely. gla- brous in the upper half. Receptacles glabrous. Flow- ers with pale yellow to green corollas. Ray flower corollas 1—1.3 em long, 0.5-1.8 mm wide, linear, apex generally with 3 teeth. Disk flower corollas 3.5—4.3 X 0.5-0.8 mm, narrowly obconical. Cyp- selae brown to brown-red, with warty surface, ho- momorphic, 2-3 X 0.7-1.8 mm, glabrous to gla- rate, with scattered short eglandular trichomes (ca. 0.1 mm). Pappus 4—6 mm, consisting of one row of white to white-yellow capillary bristles; pappus from ray flowers caducous as a whole crown. Chro- mosome number unknown. Figures 4E-H, 6D. Illustrations. Distribution. South-central China (provinces of Gansu, Sichuan, Tibet-Qinghai, and Yunnan). Open moist rocky places, woods, meadows, and near wa- tercourses, altitude 3000-5000 m (Fig. 21) Doronicum stenoglossum is the most distinctive species within the genus. In fact, historically it was treated as a different section (Soulieastrum), under D. souliei Cavill. by Cavillier (1911). Together, the large amount of autapomorphies makes its appear- ance quite different, specially regarding the capit- ula. The shape, size, and color of flowers and shape and size of phyllaries are unique within the genus E, F). In a phylogenetic analysis based on lee ular data (Álvarez Fernández et al., 2001) it is deeply nested within a group of central Asian species that share no morphological synapomor- phies at all, indicating that its distinctive charac- ters are autapomorphies and that the subgeneric treatment in this case is not appropriate. (See also comments under D. briquetii and D. kamaonense.) Cavillier (1911: 360) included Doronicum sten- oglossum in his section Soulieastrum, but because ~ J he could not see the type material, he concluded that the treatment was tentative. He mentioned that based on its protologue, the most similar species is D. souliei Cavill. When type material of both names was studied, the same identity was determined in both cases. Thus, the name D. souliei is treated as a synonym of D. stenoglossum, which has priority. To clarify the identity of D. souliei, a lectotype was designated above based on Cavilliers citation in the protologue. Two sheets that match his citation were found at G and K, respectively, and the one best preserved was selected as lectotype. Selected specimens examined. CHINA. Gansu: Hai Tchoang ze, Licent 4730 (ВМ, К); Тао river, Merku valle "y, Rock 12941 (GH); Tebbu, Drakana, Wapaku, Rock 14599 (Е, K, NY). Sichuan: Plpppane hsien, Fang ‚К, NY): Mts. Hu-li, Forrest 16825 (E, К); Mt. Poe 25 July 1885, Potanin s.n. (LE); Guma-Kika, 6 Aug. 1885, Potanin s.n. (LE); Mt. Mitzuga, Muli Gomba, Rock 16566 (E); Sikang, Kangting, Tachienlu, Chungo Valley, Hsin- lientzü, Peg 11378 (BM, GH, MO, UPS, W); Dongrergo, Smith 3587 (E, UPS); Hsioeh-shan, 19 July 1922, Smith 3880 с Tachienlu, Hadjaha, Stevens 392 (W); Sung- Volume 90, Number 3 2003 Álvarez Fernández 379 Doronicum (Asteraceae) pan, 1914, Weigold s.n. (W). Tibet-Qinghai: Dari Darlag Xian, Sainaniuda, Jimai Güymai Xiang, Huang He, Bar- tholomew & Gilbert 1205 (E, MO); Reting, Ludlow & Sher- riff 8931 E). Yunnan: Lai-cha-tse-ka, Hsia-Chung- tien, Feng 1893 (GH); Lichiang, Forrest 2663 (BM, E); Lp plateau, Forrest 10586 (BM, E, K, W); Muli, )schungdien, Tschako, Handel-Mazzetti 1343 (М); be- tween rpm Tungshan, Tuinaoko and Tsilikiang, Rock 9747 (E, GH, NY); Muli, Wachin, Jin-chang, Yü 14593 (BM, E, GH) TAxA ExcLUDED FROM DORONICUM Doronicum thibetanum Cavill., Annuaire Conserv. Jard. Bot. Genève 10: 225. 1907. TYPE: “Thi- bet" 1882, J. Murr s.n. (lectotype, designated here, G-BOIS!). This taxon belongs in the genus Aster sect. Al- pigeni subsect. Homochaeta (cf. Álvarez Fernández & Nieto Feliner, 2000). To clarify the identity of this name that was in current use until a recent study (Álvarez Fernández & Nieto Feliner, 2000), its lectotype was designated above. Literature Cited Álvarez Fernández, I. 2001. Four new combinations in Eurasian Doronicum L. (Asteraceae, Senecioneae). No- von 11: 294— —— & С. Nieto Feliner. 1997. On the lec totypification of Doronicum carpetanum (Compositae). Taxon 46: 763 & 21 Lectotypification of 16 species names in Dosis um (Asteraceae, Senecioneae). Taxon 48: 801—306 1 . 2000. A new species of Doronicum (Asteraceae, Sena ioneae) from central China. Ann. Bot. Fenn. 37: 249-25 & 1. A multivariate р th to as- sess the — utility of morphometric character: in yen EUST eraceae, быс жире Folia Put 36: 423—4 E uertes Aguilar, J. L. Panero & С. > Fel- iner. 2001. A phylogenetic ids of hd. um m (As. e, Senecioneae) based on morphological, nuclear ribosom al (ITS), and wow (trnL-F) evidence. Mo- c. Phylogenet. Evol. 20: 41—64 Ambrosi, F. 1854. Flora Ну Tirolo meridionale, Vol. Padov ong H. J. B. 1867. us ded du département des Alp exon Men Add V. E. 0. Rod a L. v Armenii (Do- ronicum L. бог іп Armenia). Biol. Zhurn. Viam 33: 532—534. ———— & Oganesyan (editors). 1995. Flora Armenii, Vol. 9. Koeltz Scientific Books, Havlickuv Bro E L. 1956. Cytotaxonomical studies on the Flora of ry. Ann. Hist.-Nat. Mus. Natl. Hung. 7: 321-334. mae M. 1991. Cytologic ey investigations of some Greek plants. Fl. Medit. 1: 164 Barratte, M. 1893. p scorpioides Willd. et Bum austriacum L. existent-ils en Algérie? Bull. Soc. Bot. France 40: 115-119. Bauhin, C. 1623. Pinax Theatri botanici. Basilea Beck, С. 1893. Flora von Nieder-Osterreich, Vol. 2. Wien. Belaeva, V. A. & V. posu 1981. IOPB chromosome number reports. Taxon : 60. & J. D. Hooker. 1873a. Genera Plantarum, —— —. 1873b. n papers to Ben- к? Hookers Genera Plantarum, Vol. 13. Mene Bohlm man .R. quias 1979. Ein neuer Ses- алое "und andere Inhaltsstoffe aus s Doi cum ру dii Phytochemistry 18: 668—671 & M 1979. Neue ln бойни aus Doronicum Ы жын ай Phytochemistry 18: 179— ‚ А. Dhar & M. Ahmed. 1980. Thymol derivatives from Doronicum hungaricum. Phytochemistry 19: 851. Boissier, P. E. 1844. Diagnoses Plantarum Orientalium no- varum, ser. 1, 4. Leipzig. . 1875. Flora Orientalis, 3. Gen eu 0. de & J. Vigo. 1995. Flora dels | ud Catalans, . 3. Barcino, Barcelona. TAM pik J. & F. Koch. 1930. Doronic um e x ое (hybr. nova). Кереп. Spec. . Regni Veg. 27: Bremer, K. 1994. Asteraceae: Cladistics & Classification. Timber Press, Portland, Oregon uir J. 1901. Recherches sur ls Flore des montagnes e la Corse et ses origines. Annuaire Conserv. Jard. Bot sare 5: 12-119. Candolle, A. P. de. 1836. Je systematis naturalis regni i aon Vol. 5. Par 1838. Prodromus Мо РЯ naturalis regni ve- жи ES 6. Paris. Cassini, Н. 1817. Apergu des genres ou sous-genres nou- veaux formes par M. Henri Cassini dans la famille des Synanthérées. Bull. Soc. Philom. Paris 3: 32 1819. Suite du sixieme mémoire sur la famine des sy ynanthérées, J. Phys. Chim. Hist. Nat. Arts 88 E 189-20: Cavillier, 1907. Étude sur les Doronicum à fruits hom- omorphes. Annuaire Conserv. Jard. Bot. Genéve 10: 177—2: oi . Nouvelles études sur le genre мате ит. Annuaire Cans . Jard. Bot. Geneve 13-14: 195—368. Chabert, M. 891. Tréisihme note sur la боге ABA Bull. Soc. pn France 38: 385—386. 392. Quatriéme note sur la Йоге d'Algérie. Bull. c. Bot. France 39: 334—336. C b. ón, R. 1987. Contribución al estudio taxonómico del género ping i 1; (Compositae) е - E : enínsula Ib- érica. А ge . Bot. Madrid 43: 2 Chen, Y. * o new species of E Linn. (Compositae) from China. Acta Phytotax. Sin. 36: 73- t am t, E. J. & M. C. De. 1994. Alien Plants of the pe Isles. BSBI, Lon a C. 1601. “ёз аш Esos historia. Antwer- Contandriopoulos, J. 1957. Contribution à l'étude caryol- ogique des TNT" de la Corse. Ann. Fac. * Marseille 26: 5 Coutinho, A. X. T Йога de Portugal, ed. 2. Lisboa. -—— 5 J. 1587. Historia generalis plantarum, Vol. . Lyo yo жы М. T. 1985. Chromosome numbers in the rep- resentatives of the flora from Georgia. Bot. Zhurn. 70: 698—700. Annals of the Missouri Botanical Garden De Pinna, M. C. C. 1991. Concepts and e of апо in the clac ез paradigm, Cladistics 7: 367—394. Moi R. L. 3. Flora Atlantica, Vol. 2. Paris. Diels, a о um limprichtii nov. spec. Керегі. Speo. XE eg. 12: 507. Шол Р. 1554. De celle ‘a materia libri sex. Lyon. 1557. bes de las yerbas y plantas (translated into Spanish by J. Jarava). Anvers p R. 1574. Purgantium ... h › А. 1877. Vol. 1. Leyde. дел 'aud, J. 1992. Le genre Doronicum L. en E legie dans les régions к Nat. Mosana 45: 81—92. Edmonton, ye R. 1973. Notes on Doronicum z Gard. و‎ 32: 255-258. 975. Doronae um L. 37-145 in р P avis (editor), Flora of Turkey and ihe East Aegean Islands, Vol. 5. c Bum Press, Edinburgh. m enus pig: L. Roy. m бы Edinburgh 37: Favarger, C. & L. Huynh. 1964. IOPB ка num- bers report II. Taxon 13: 208. — —— & P. Küpfer. 1968. Contribution à la cytotaxon- omie de la flore = des Pyrénées. Collect. Bot. (Bar- celona) 6: 325-35: Fedorov, A. (кйш. 1969. Chromosome Numbers of Flow- ering Plants. [Reprinted by О. Koeltz, 1974.] Koenig- stein. Fernandes, A. & M. Queirós. 1971. Contribution à la con- naissance cytotaxonomique = Spermatophyta du Por- tugal. Bol. Soc. i ot. 45: 5- Fournier, P. 1939 son-les-Grancey Garbari, F., P. Ме & С. istoriae. Antwerpen. Supplément aux dictionnaires arabes, in SW т 5 Ф — = ~ е, uv — МЈ in Iran. Notes Les Quatre ы de la France. Poin- Monti. 1980. Ped hp somk a Flora italiana. Inform. Bot. Ital. 12: 110. Goldblatt, P. (editor). 1985. Index to plant vise numbers 1982-1983. Monogr. Syst. Bot. Missouri Bot. Gard. 13: 1-224. — == (CD ~ ———— (editor). 1988. Index to plant chromosome num- bers 1984- 1985. Monogr. Syst. Bot. Missouri Bot. Gard. : 1-204. 1994. Index to Lo ;. Johnson ере pu . Monogr. Syst oe numbers rs ek Missouri Bot. Gard. 51: 1-267 —— & (e ius 1990, Index to acm chro- mosome numbers Wigs 993. Monogr. Syst. . Mis- ri Bot. Gard. 58: & (e "lius. onn Index to pm chro- mosome че uJ ВА . Monogr. Syst. . Mis- souri Bot. Gard. 6 o p hav 1961. a 'um L. Pp. 765-782 in Flora of .5.5.К. [Edited by B. K. Shishkin & E. G. Bobrov: English translation, 1995.] Dehra Dun, India. Greuter, W. J. McNeill, F. R. Barrie, H. ==, = ux г = ›. 5рїс eam florae rumelicae et bithynicae, Vol. 2. ‘rans hw — & A. Schenk. . Iter [ш um. Arch. Na- turgesch. 18 : 342. Harron, J. 1986. Flora of Lough Neagh. University of Ul- ster, Coleraine. я F. 1851. Flora von Tirol, Vol. 1. Innsbruck. Heg E G. 1928. уы Flora von ive -Europa, Vol. aa von J. F. Lehmanns, Münche Hoffmann, 1892. Doronicum L. Engler ^ P A. E. P ix nri e ы Vol. 4. Le Hollis, S. & R. K. Brummitt. 1002, World Geographical Sc Deme for Rec тн Plant Distributions. International Working Group on Taxonomic Databases for Plant Sci- ences, Hunt Institute for Botanical Documentation, Car- negie Mellon University, Pittsburgh. Imbert, E., J. Escarré & J. Lepart. 1996. Achene dimor- phism and E population АВ in MN sancta (Asteracea Sci. 157: 309-315 eque y i 73. Flora [bru 'eae, Pur . Viennae. —— LE Austriaceae, Vol. "i сац i Flora Austriaceae, Vol. 4. Viennae. 1786. Collectanea ad дик 'am, chemiam, et historiam denen spectantia, Vol. 1. Viennae. Jarvis, C. E. & N. Turland. 1908, ‘Typifie ation of Linnaean us ific oi varietal names in the Compositae (Astera- sae). Taxon 47: 347-370. itus s. 1980. Voie Seca hookeri (C. B. a ex Hook. 1 EA arr Acta totax. Geobot. 31: . 1843. * nopsis florae germanicae et hel- г ssa 294. in Pp. С. А. Die nattirlic hen ا =) = M. number reports, Taxon 22:. & S. Georgieva. dee (T RR number reports. Taxon 32: 66 Po 780. Encyclopédie méthodique. Bo- tanique, Vol. 2. Paris. Leaf Architecture Working Group. 1999. Manual of Leaf Architecture—Morphological description and categori- zation of dicotyledonous and net-veined monocotyledon- ous angiosperms. Smithsonian Institution, Washington, Э.С egrand, A. 1894. Flore analitique du Berry, ed. 2. Bourges I dE K. 1950. Some results of a рей al investi- galion of Eo gris Hereditas 36: 94—102. Linnaeus Species Plantarum. Stockholm Lippert, Heubl. 1988. Chromosomenzahlen von Haven а aus Bayer = апртепге anden Gebieten. Ber. Bayer. Bot. Ges. 59: 13- LoiseleurDeslone 'hamps, J. L. A. 1807. Flora Gallica. Lu- 1. Kjellqvist. 1974. Cytotaxonomy of Spanish plants IN. Dicotyledons: Caesalpiniac eae—Asteraceae. Ганчави е Е -211 Lovka, nik, A. e & D. XE 1972. IOPB is ш. г reports. Taxon 21: 337—339 Maguire, В. 194: : 3. A monograph of the genus e 'a. Brit- to onia 4: 386—510 ský, J. & A. Murín. 1987. Karyotaxonomicky Preh!’ ad Flory Berane Veda Vydavatel ‘stvo Sloven- islava. n P. 1787. Dictionnaire des jardiniers, nouvelle édi- tion, Vol. 3. London. Moore, D. M. 1982. Flora Europaea Check-List and Chro- mosome Index. eai Univ. Press, Cambridge ecker, N. J. de 1790. Elementa botanica. Neu wied. Кыз h, С. M. rend Enumerationem plantarum in ter- ritorio йт ооа udae Nordenstam, B. . Senecioneae and Liabeae system- alic review. n. ge 30 in V. H. Heywood, J. B. Har- borne & B. L. Turner (editors), The Biology and Chem- istry of je С ;ompositae. Academic Press, London. . 1978. Taxonomic studies in the быы, Opera Bot. 44: 1—84. tribe Senecioneae Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) Nyman, C. F. 1879. Conspectus florae europaeae, Vol. 2. Orebro. Pallas, S. 1779. и altaicum. Acta Acad. Sci. Imp. Petrop. 2: 271-27 Pena, P. 1571. one P S nova. Londor Pérez, С. & А 1990. з algunos p ibéricos. AN аба 15: 151-160 F. Llamas & C. jtm 1994. Doronicum pubescens sp. nov. yer 14: 5-12 Llamas, C. o & A. Pen nas. 1997. Typifi- сайоп e definition “ po nicum austriacum Те q. (Asteraceae). a Newslett. 30: 1— Polatsc lek. A. 5. Cytotaxonomische Beiträge zur flora der us өш эө Oesterr. Bot. Z. 113: 141-142 Reichenbach, Н. G. L. 1854. elveticae, ~ 16. Leipzig. Reic ere L. 183 Icones florae germanicae et 31-32. Flora germanica excursoria, Vo zig. Reine ri, T & D. Moravetz. 1956. Contributii la studiul Florei Republicii Populare Romine. Comun. Acad. Re- publ. Populare Romíne 6: 661. Reynaud, J. & J. Raynaud. 1984. Sur la présence d'Onopordine chez Doronicum grandiflorum Lam. (Composées). Pharmazie 39: 126. & 1986. Les не de Doronicum grandiflorum. Bioc "hem. Syst. & Ecol. 91-193. Becchi & J. Raynaud. 1985 j -Hydroxyace- tophenone derivatives fram "idc grandiflorum. J. od. (I Epis a) 48: 3 Robins, D. J. 197 p chemical review. 831—850 in V. н. evwood, J. B. Harborne & B. сее юа) The rus and Chemistry of the Com- . Academic Press, London. Rouy., М. С. 1893. Note sur le Doronicum scorpioides Willd. ue Soc. Bot. France 40: 186—189 . Le genre Doronicum dans la flore euro- péene et rien la flore atlantique. Rev. Bot. Syst. Geogr. Bot. 22. о, ОЗЬ. Le genre Doronicum dans la flore euro- péene et dum la flore atlantique. Rev. Bot. Syst. Ge Bot. 3-40. Ruiz de Clavi ijo. E. 1993. Números cromosómicos para la flora espafiola 664—690. Lagascalia 17: 170. — 1840. Flora comitatus pestiensis, ed. 2. Buda- 7" desi T. 1964. Flora Republicii Populare Romine, Vol. 9. Academiae Reipublicae Popularis Romanicae. Schur, F. 1866. ulia onm plantarum Transsilvaniae. Wie Sergjevskaja L. P. 1949. Doronicum bargusinense Serg. Sist. Zametki Mater. Gerb. Krylova Tomsk. Gosud. Univ. мн —2: 14—15. Simonkai, L. 1886. Enumeratio florae transsilvanicae. Bu- dape и M. 1950. Studies in chromosome numbers of Polish angiosperms. Acta. Soc. Bot. Poloniae 20: 45- 68. Stace, C. A. 1991. New names and combinations for three hybrids. Watson 18: 321. Strid, A. & I. A. Anderson. 1985. Chromosome raging of Greek mountains plants. Bot. Jahrb. Syst. 107: 206. & ranzén. 1983. Chromosome ilis dia in po plants from Greece. Willdenowia 13: 32€ $ J. 1814. Deutschlands Flora, Vol. 38. i Sweet. R. 1833-1835. The British Flower Garden . 2. 3. London. Tanowitz, B. D., P. F. Salopek & B. E. Mahall. 1987. HE ferential germination of ray and disc achenes in Hem zonia increscens (Asteraceae). Amer. J. Bot. 74: 303- 312. Tasenkevitch, L. І. Vysotskaja & N. K. Vorobetz. 1989. ا‎ к yir ies in rare and endemic spe- cies of vascular plants from the Ukranian Carpathians Bot. Tes 74: 1669-1670 1828. Ueber Arnica und Doronicum. Flora s Tournefor J. P. 1700. Institutiones rei herbariae, Vol. 1. Valdés, B., S. Talavera & E. Fernández Galiano cy 1987. а ا‎ de Andalucía Occidental, Vol. 3 Ketres elona. Van Loon, s c 1980. IOPB chromosome number reports. Taxon i 720. . J. M. H. fe WE some нА герогіѕ. Тахо ы 34: Venable, D. L. & = chromo- evin. 1985. eae - s v нн in Heterotheca oka J. Ecol. тн jos; U. D. & P. Kachroo. 1989. Cytogeography of some endemic taxa of {еч Himalaya. Proc. Indian Natl. Sci. Acad., B 3: Webb. P. B.& 5 Adis in 1846. Histoire Naturelle des les € i s. . Phytogr set Canaria, Vol. 2. Paris. Widder, F. J. Eine e Pflanze der Ostalpen— us оа Мас e cataractarum- und ihre Verwandten. Repert. Spec. Nov. Regni Veg. 22: 1 pei 34. Die Bastarde der pma um Arten. Mitt. Naturwiss. Me Steiermark 71: 132-146 1948. Der neue Bastard "iuis um columnae X iniu = Doronicum prennii Widder. Phyton (Horr “Г B 1-104. . H. Rechinger. € Compositae novae per- sicae, p et kurdicae (D. hyrcanum). Oesterr. Bot. Z. < 36 eK C L. 1809. cuan plantarum. Berlin Willkomm, M. & J. 1870. Prodroms florae TT Vol. 2 ا‎ ttgart. Zohary, M. 1950. Evolutionaty trends in the fruiting head of Compositae. Evolution 4: 103—109, APPENDIX 1 LIST OF SPECIES AND SUBSPECIES Doronicum ен Pall. д (lectotype) by Álvarez Para & Nieto Feliner, 19 2: Doronicum 1 oe Lib [LT by Pérez, Llamas, Acedo & Per 91] 3. quisi ns Cavill. [LT by Álvarez Fernán- um p Nieto Feliner, 1999] . Doronicum cacaliifolium Boiss. & Heldr. [LT by Ál- varez Fernández & Nieto Feliner, Doronicum carpaticum (Griseb. & A. Schenk) Ny- “ onicum carpetanum Boiss. & Reut. ех Willk. [LT i Chacón 1987] pisi: carpetanum Boiss. & Reut. ex Willk. abs. carpetanum "Шем Ө мел. vend Boiss. & Reut. ex T C. z Morales & A. Penas) Álv. F 382 Annals of the Missouri Botanical Garden subsp. pubescens (C. Pérez Morales, A. Penas, F. Llamas & C. Acedo) Aizpuru 7. Doronicum cataractarum Widder 8. Doronicum clusii (All.) Tausch 9. Doronicum columnae Ten. [LT by Álvarez Fernández & Nieto Feliner, 1999] 10. Doronicum corsicum (Loisel.) Po 11. Doronicum dolichotric i Cavill. [LT by Alvarez Fernández & Nieto Fe liner, 1999] 0 Clarke [LT by Alvarez ibaa K 399] = Doronicum turkes- е Cavill. T T here] . Doronicum gansuense Y. L. Chen Doronicum vind iale Мы n) Nyman [LT by Álva- rez Берне ре & Nieto Fe p 1999] = Arnica doronicum Jacq. [LT here} 15. о ит лче, Lam. [LT by Álvarez Fer- nández & Nieto Feliner, 1999] — Doronicum ne Chabe rt [LT here] = Doronicum scorpioides Lam. |LT here] 16. agi кенне ез Cavill. [LT by Álvarez беге 'z & Nieto Feliner, 1999] Doronicum hungaricum Rchb. fil. [LT by Álvarez Bau & Nieto Feliner, 1999| 18. Doronicum ا‎ ii (DC.) Alv. Fern. [LT by Al- varez Fernández, 2001 | 1¢ Joronicum ЖОО Fisch. [LT Ьу Álvarez xus & Nieto Feliner, 1999 9a. Doronicum ба s Fisch. subsp. macro- phylla um I9b. Doronicum macrophyllum Visch. subsp. sparsipi- losum (J. R. Edm.) Álv. Fern. 20. Doronicum maximum Boiss. x Huet [LT by ÁI- ios z Fernández & Nieto Feliner, 19‹ Doronicum oblongifolium DC. b T by Álvarez Fer- sinl & Nieto Feliner, 1999 Doronicum orientale es = шш um caucasicum M. 23. Doronicum pardalianches 1 Ж Acedo & Penas in Jarvis & Turland, 24. е ит он L. [LI Ts Llamas, Pérez, Acedo & Per n Jarvis & Turland, 1998] 25. Boo um re flam Boiss. [LT by Álvarez Fer- nández & Nieto Feliner, 1999 26. Doronicum sensu Maxim. [LT by Alvarez Fernandez & Nieto Feliner, 1999] = Doronicum souliei Cavill, [LT here] іер. [LT here А Llamas, Pérez, APPENDIX 2 INDEX TO EXSICCATAE Specimens are listed alphabetically by collector, fol- lowed by collection number or date, only number is unavailable. The number in parentheses cor- responds to the number in the List of Species and Sub- species above. J. Abel, 94 (12); A. Achverdov & A. Doluchanow, 15 Aug. 1929 (21), 17 Aug. 1929 (21); Adamovic, May 1896 (17), June 190? (22), July 1900 (9), July 1905 (9), July 1917 (5); C. Aedo, 25 July 1982 (6d), 2 July 1983 (6d), 25 Aug. 1985 (6b). 16 May 1987 (24), 2 Aug. 1987 (15). 8 July 1990 (15); C. Aedo et "m 2888 (23), 36031b (ӨЧ); P. abis ч? July 1932 (10); R. Ajdarova, 21 May 1983 (12); J. R. Aketagd, 53 (9); I. з ua 29 July 1893 (21); Al. se 2 May 1963 ): Albury, Cheese & Watson, 705 (22), 1751 v 2), 3159 (11), 3176 (21); J. A. Alejandre, 333-86 (24); J. A. Alejandre et al., 1460 (6d); R. C. Al- when collector ~ Т = exander, 1842 (2) (14); Alexcenko, 12861 (19a), 8 July 1897 (19a), 31 a 1901 (12); Alexcenko & Woronow, 13586 (21); Ch. S. Ali & Metz, 157 (12); T. Almaraz, 26 e 1997 (15): i ee & I. Alvarez, 977 (24), 979 ; T. Almaraz, I. Alva M. A. Garefa, 1008 (23), ted (23), 1015 (23); T. Айшага: I. Álvarez, M. A. García & j Duno, 984 (24); T. tye I. Álvarez, M. A. García . Medina, 802 (6a); T. Almaraz, I. Álvarez, iet & E. Monasterio- ii 804. (6a), 805 ке T. Almaraz & A. Cano, 275 (15), 288 (15); Alpay, 2641 (22 2642 (25); A. H. G. Alston & N. Y. Sandwith, 45 (9), 441 2). 691 (9), 2119 (2), 2065 Ө, 1528 (9); A. H. G . D. Simpson, 37736 (24); I. Álvarez, (6 PN I. Álvarez, M. Car ‘fa, Jansen, L. Se queira, 1296 (6d); I. A. García & E. Mon- asterio-Huelín, 924 (6b), 926 (64) I. Álvarez & N. Yagüe, 929 (ба), 931 (6a), 932 (6a), 933 (6c), 935 (6c), 936 (6d). 937 (6d), 941 (6d), 942 (6d), 944 (15); 800 (23). = ay 3), дыл 05, 954 (24), 966 (24). (6c); I. Álvarez, A. Herrero & N. Yagüe, 1350 (1: 2 (8), 1354 (9. 1355 (8); E Amich, 22 Apr: 1978 (24), 1 May 197 8 (24 "i Amich. & F. Herrero, 17 May 1 o 24); F. Amich, E. Rico & J. Sane hez, 26 July 1977 (15), 28 June e a); F. Amich, E. Rico, J. Sánchez & X. Giráldez, 25 Apr. 1981 (6d); Anania; 28 May 1955 (17); H. Andres, 695 (23); R. M és, 18 July 1974 (6d); R. Ansin, 392 (22); A. Aparicio, P. Murillo & S. Silvestre, 28 June 1984 (24); J. M. Aparicio & P. M. Uribe-Echev- arría, 27 July 1993 (15); H. Appleton, p (12), 902 (12); V. J. Aran & M. J. Tohá, 10 Aug. 1987 (23), 13 May 1995 (24); M. Arbella, 16 June 1981 (15); J. C. Archibald, 326 (9), 585 (2); Ar-I eybuey, 532 (9); E. Armstrong, 19 May 1949 (23); A. Arvat, 7 May 1934 (17); Arvet-Touvet, Cha- boisseau & Fauré 7 (8); С. Aseginolaza & D. Gómez, 21 July 1988 a». : Aug: 1988 (23); Assadi & Mozaffar- ian, 30306 (11); S Atchley, 945 (9), 2250 (9); L. Ater- TS am, T ido, Aug. 1918 (15 E Ain 'her-Eloy, 3268 (22), 3269 (25), 3848 (22), 4751 (22), 2 Apr. 1865 (22); Audibert, 1828 (24); Ausserdorfer, July 1863—65 (9), 6 July 1865 (9), 13 July 1865 (9), 10 Aug. 1873 (14); Averianov et al., 2421 (21), 2461 (21), 2837 (19a), 3200 (19a); Z. Aytaç, 1167 (22), 2933 (19a); Z. Aytaç & M. Ekici, 6229 (25); G. V Aznavour, 27 Mar. 1892 (22), 10 May 1896 (22), 7 Apr. 1889 үне 12 Apr. 1904 (22). . Bübler & I. Quasdorf, 625 (9), 758 (9), 965 (9); Backir, 183 (11), 220 (19a); C. Baenitz, 7 Aug. 1895 (2); Balansa, 728 (25), 1866 (19а); A. Baldacci, 14 (9), 25 (9), 371 (9), 457 (9); J. Ball, Sep. 1842 (15), . (2), 6 Aug. 1858 (2), 25 Aug. 1860 (9). (15), 18 June 1862 (23), 3 July 1862 (15), 15 Aug. 1 (15), 29 Aug. 1869 (9), 3 Aug. 1875 (9), о 1877 (9), Sep. 1883 (15), 24 Aug. 1884 (15), Aug. 5 1887 (2); P. W. Ball, 28 May 1958 (23); К. (22), 1049 22): E. K. Balls & W. B. Gourlay, 3231B (9), ae (22); L. Baniova & L. Kuprienova, 14 June 1960 ; V. Йаа Баш; 18 July 1973 (2): W. Barbey, 919 (22), 23 n 1873 (22); R. Barbezat, 1834 (2); T. H. Barbour, Aug. 1922 (18); C. Barclay, 1521 (22); I. Bane зга, 29 June Er (24); J. Barth, 20 May 1873 (17), 17 June 1885 (2), June 188 . 15 Aug. 1888 (9); B. Febre ti & A Gilbert, ovs E. Baschant, Aug. 1922 (2), June 1925 23), 14 May 1934 (23), July 1935 (8), 24 June 1937 (2), | July 1937 (8), 1943 (2), 1 July 1947 (2), Aug. 1948 5), Aug. 1950 (15); Bayern, 1860 (19a); A. Baytop, 4108 22); A. & T. iue v 25); T. Baytop, 11296 (22): . Beauverd, 2 1. Beck, 6 Aug. 1968 (8), 13 Aug. 068 (2 ), 6 an = = TES 11 Aug. 1976 (9); G. Beck D — 2 ы C) RSS — Volume 90, Number 3 2003 Álvarez Fernández 383 Doronicum (Asteraceae) von Mannagetta, June 1885 (9); G. Beck von Mannagetta & F. Fiala, 232 (9); A. Becker, 132 (19a), 220 (19a); H. Beger, 24 July 1909 (2), Aug. 1911 (15), 19 July 1912 s 3), 23 July 1912 (15), 3 Aug. 1912 (15), 1 May 1913 23), 7 July 1920 (15), 25 Aug. 1920 (8), July 1929 (14), 10 Aug. 1939 (2); O. Behr, July 1934 (2); W. Behrendsen, 4 July 1902 (15), 14 July 1903 (9); F. Bellingham, 20 July 1961 (9); F. Bellot, 30 Mar. 1940 (24), 29 May 1953 (бо), 16 May 1965 (24); D. Belmonte, 18 May 1982 (24), May 1982 (24); ا‎ 19 May 1911 (24), Aug. (ба); Вепеаї, Blanché, Molero & Vallés, 14 July 1984 (ба); J. L. Benito & D. Goñi, 14 rA hp (15); Berg, 12 Oct. 1924 (14); E. Berger, 30 May 1 23); Bernard, 20 July 1901 (15), Aug. 1902 (15); жеш 18013 (25): J. Ber- nátsky, 20 May 1898 (9); W. Bernouilli, 14 Aug. 1888 (8); S. Besalet, 163 (22); S. H. Bickham, 876 (23); Bierbach, May 1903 (9), June 1903 e» July 1903 (9); E. Biesalski, 469 (2); C. Billot, m 6 (2); Binder, Hage mann, Hempel & Raus, 297 (9): “Binz, 405 (8) (15); E. Blanco, 1348 (24); G. b 10476 (24 ; M. О. els ani, 2 Aug. 1912 (8), 28 Aug. 1912 (15); S Bogolubow, 6 June 1909 (12), 14 June 1909 (12); Bst 1842 (22); Boissier & Reuter, July 1858 a ا‎ 19 June 1873 (23). 26 May 1873 (23), 6 May 1 24); Bonnier, 163 (8); М. L. Bor, 11948 (12): e July 1866 (23), July 1872 (23), July 1873 (23); A. Boreau, 27 Apr. 1846 (24): J. M. Borel, June 1850 (23), 13 July 1862 (15), 7 Aug. 1864 (15), 5 Aug. 1866 (15), 2 June 1868 (23); J. Borja i S. Rivas Martínez, 1962—63 (15); Borja, 18 July 1951 (6d); J. Bornmüller, 76 (17) (22). 127 (9), 176 et 688 (22), 1165 (2), 1166 (2), 2624 (22), 2371 (9). : 4264 (22), 4265 (22), 4266 (22), 4267 (22). : (2) - ). 4621 (22), 46022 à 4623 (22). 9607 (22). 9620 (25 ‚ 9666 (22), 14245 (22 1886 n 8 May 1887 (9), 22 July 1887 (2), | 22), 4 May 1888 (9), 24 July 1888 (8), 1 June 1804 (23). 15 July 1894 (15). 9 Aug. ES 5), 1 July 1895 (8), 10 Aug. 1895 (15), 13 Aug. 1895 (15), 7 June 1887 (9), 9 July 1895 (8), 1 June 1896 (2), 3 Sep. 1896 14), 27 June 1906 (14), 18 July 1907 (9), 27 July 1907 (14), 30 July 1907 (14), 10 June 1908 (2), July 1911 (23), 28 June 1912 (9), 7 July 1912 (9), 9 July 1912 (9). 19 July 1912 (8), 24 July 1912 (8), 25 July 1912 (8), 7 Aug. 1912 (2). July 1916 (2), 28 Aug. 1922 (15). 8 ы 1925 (8), 21 July 1925 (15), 28 July 1925 (15), 4 June 1931 (9), 26 July 1932 (8), 2 Aug. 1932 (8); J. & F. i goo 11965 (22); Bornmüller & Schuller, 1918 (9); A. B. Borodina et al., 27 July 1976 (5); A. Boros, 25 May ps (17) A. Borza & E. I. Nyárády, 9 Aug. 1933 (9); Boilie, hy: 1907 (23), 15 July 1928 (14); Bouchard, 3 Aug. 1 (15). Sep. 1909 (15); H. Bourdot, July 1889 (23): E pg au, 19 (15), 139 (8), 140 (15), 2508 (ба), 2671 (6a), 21 dune 1849 en 20 Apr. 1860 (22), July 1862 (15): 1842 (15); E. Brandis, Apr. 1898 (9); J. Briquet, : (23), iege (23); C. E. Britton, 27 May 1905 (23): J. lioc h- 9 (12), 69 (12); A. H. & V. F. Brotherus, 500a (19а). 1) 500c (11), 501 (21), 501a (21); V. F. Brotherus, P 23); R. Brumm, Aug. 1900 (14), July 1903 (8), Aug. 1904. (8) (15), July 1906 (15), Paire 5494a (15), 6554 (22); ). Chater, 135 (24); Brunner, July 23); Brutelette, May 1862 (24); Buades, Mazimpa- ka, E & Ron, 20 May 1976 (24); J. Bubela, 24 July d 3 (2); J. Васе, nu 1931 (17); P. Buchner, 83-70- 3 (25); Bührer, 2 Aug. 1942 (15); Burgos & Cardie ue (24), 20 ы 1985 (6a); К. кы 9 June 1889 (22); J. Burtt-Davy, 121/27 (2); E. & М. A. Busch, 23 Му en и. —. erel, : 500b fe — 2 (19a), 47 (19a), 8 July 1913 (21), 10 July 1913 (21), 14 June 1925 (19a), 4 July 1925 (21), 13 July 1925 (19a), 28 June 1927 (21), 1 July 1927 (21). 3 July 1927 (19a), 12 Aug. 1927 (19a), 9 July 1928 (19a), 21 July 1928 (19a), 28 July 1928 (19a), 12 Aug. 1928 (19a), 19 July 1929 (19a), 1 Aug. 1929 (11), 5 Aug. 1929 (19a), 8 Aug. 1929 (19a), 18 Aug. 1929 (19a), 12 July 1930 (19a), 19 July 1931 (19a), 25 July 1931 (19a), 7 Aug. 1931 (19a), 31 Aug. 1931 (21), 17 July 1932 (21), 20 July 1932 (21), 4 Aug. 1932 (21), 18 July 1933 (19a), 20 July 1933 (19a), 1 ты 1933 (21), 23 Aug. 1933 (19а), 6 Sep. 1935 (19а); . Busch, 2 July 1903 (19a), 14 July 1903 (21), 18 ur 1904 пиг М. A. Busch & B. Klopotow, 668 (22) 16 Apr. 190 A. esae 20 May 1949 (24); Callier, Hirte & Scholz, 21 Aug. 1893 (2); C. Calvo, 25 July 1984 (15); F. Cámara, 17 July 1935 (ба); Cardiel & Burgos, 20 June 1985 (24); Carrasco, Burgaz & Martín-Blanco, 27 July 3 (2); Сата, Casaseca, Fernández Díez & Velayos, 16 Pu 1979 (6a); Carrasco & Velayos, 13 Apr. 1987 (24); risso & Mendoga, 25 Mar. 1926 (24); Caruel, 20 Apr. 1856 (9); B. Casaseca, 17 May 1968 (24), 30 Mar. 1973 (24), 19 July 1973 (6d); B. Casaseca, & F. J. Fernández Díez, 16 July 1974 (6d); B. Casaseca, Fernández Díez & were 8 Mar. 1977 (24), 4 Aug. 1977 (6c); Casaseca et . 27 July 1979 (6d); J. L. Castillo & R. Cordero, 26 : P/212 (15), 846 (6c); J. Drobny, 22 Aug. 193 4 . Drummond, 654 (24), 14024 (18), 14120 (18). 14137 "e 22598 (12). 18 Apr. 1872 (24); W. Dubiansky ; N. Basilevskaja, 16 July 1927 (12); Duby, 263 (23); Dadia, 34753 (22), 34813 (22); H. Duman, 4068 (20), 5581 (4); H. Duman & Z. Aytaç, 5413 (16); F. so & С. Mortin, 11 July 1970 (15); J. F. Duthie, 847 (1¢ 4126 (18), 4127 (12), 3066 (3), 11037 (18). pr 4 (12). 13402 (18), 13540 (12), 6 Sep. 1892 (18), 16 Aug. 1899 (18); J. Duvigneaud, 76B 325 (23), 23 May 1979 (22); A Duzenli, 588 (22); A. Dzejver. 29 June 1901 J. R. Edmondson, 292 (2), 328 (2), 463 (25), 851 (19a). 30 June 1971 (22); Edmondson & McClintock, 2260 (22); Efremov, 10 Aug. 1951 (2); Egger, Phi 1914 (22); Ego- rova, Tsvelev & Cherepapov, 2061 (19a); R. Ehwald & N. endt, 18 June 1968 (19a); Eichenfeld, 2 Aug. 1882 (9). 3 July 1894 (15); M. Ekici, 1323 (22); T. Ekim, 27 (22); Hno. Elías, 3414 (24), 17 May 1907 (24); G. Eliçin, 3412 (22); E. Ellman & e аА 383 (24); Elsmann, Мау 1827 (9), 1825 (9); € . Enden, 11 Sep. 1931 (1); En- dress, Aug. 1830 (2 YEN Sep. 1831 (15), 1831 (2); V. e 20 July 1903 (14), 18 July 1907 (14), 25 July 1907 (14), 1 Aug. 1912 (15), 4 Aug. 1930 (15), 30 June Ted (2); Enko, 23 Apr. 1903 (22); Erichsen, Aug. 1891 (15); Н. Ern, 7423 (22); R. Estébanez López, July 1915 (6d). W. P. Fang, 4129 (26); J. E. Farreny, 13 July 1975 (23); R. Farrer, 144 (13); B. A. Fede 21 29 June 1908 (12), 5 1): | рә О 9 July 1908 (12), 2 Aug. 1912 (2), 15 July 1937 (19a & B. A. Fedchenko, 2 July vn (12). 13 Aug. 190 (12); A. Fedorov et al., 219 (1), 6 July 1948 (1), 25 July 1948 (1), 27 July 1949 (1), 7 Aug. 1949 (1), 15 Aug. 1949 (1); К. M. Feng, 1893 (26), 2171 (3); A. idem" К. Sousa & J. Matos, 21 Apr. 1952 (24); C. Fernández, 2 May 1982 (24); C. Fernández & E. Gutiérrez is 860310 (24); M. C. Fernández Arroyo, 10 Apr. 1982 (24); J. Fernández Casas, 605 (23), 1012 (15), 26 July 1969 (15); J. Fernández Ca asas & García Guardia 792 (6a); J. Fernández Casas et al., 7130 (24); B. Fernández de Be- toño & Alejandre, 1524-85 (24), 348-88 (6a); F. J. Fer- nández Díez, 30 June 1974 (6c), 30 July 1974 (6c); Fer- nández Díez, Amich, Rico & Sánchez, 30 May 1980 (6a); F. Fernández González & R. Gavilán, 14 July 1988 (24); J. Fernández Horcajo, 6 May 1984 (24); B. Fest, 571 (7), July 1905 (14); Fetissow, 8 July 1878 (12), 18 July 1878 12), 29 June 1880 (12), 8 July 1881 (1), July 1881 (12), 24 July 1882 (12), 1 Aug. 1882 (12); Feuilleaubois, 2872 24); О. Fiedler, 15 July 1935 (15); Figueiras, 1 Aug. 1978 (6a); F. Filarszky, 4 July 1900 (8); F. Filarszky & S. Jávorka, 5 July 1907 (5); A. Fiori, 25 Aug. 1917 (2) е Fischer, 1822 (22); Fitz & Patzak, Sep. 1954 (2); S Pak 10 July 1958 (5), 13 July 1958 (5), 21 July v ‚ 10 нүн 1959 (2), 17 June 1959 (5), 12 July 1960 2 (17); A. Fomin, 12 July 1898 (11), 16 22); E. Formánek, Aug. 1891 (2); G. Forrest, 3), 7662 (3), 2663 (26), 10586 (26), 13009 (3). 13164 (3). 14413 (3), 14526 (3), 16825 (26), 25695 (3), 28568 (26), 30434 (3); Forsyth-Major, 919 (22), 12 July 1916 (10), 21 July 1916 (10), 18 July 1916 (10), 18 Aug. yn 5 Aug. 1917 (10); H. E. Fox, 780 (24), May 24); А. Franchet, May 1870 (24); E. F. ris hi, 2l Lo 2); R. Franzén & A. Andersson, 680 (9); J. Fraser, 9 May 1904 (2: Yi A. Fredholm, 1189 (23); W. Freiberg, 2 Aug. 1942 (8); L. Frey & Z. Sztyler, ЗІ May 1971 (2); | Freyn, 1873 (17): K. Fritsch, July 1907 (2 ): Fritz, 307 (9); . Fuente, 6 June 1980 (24); Fuentes, 6 Aug. 1977 (6a); 15 May 1975 (24); P. — К کر‎ — — E. Fuertes Lasala, Furse, 3995 c P p Эа). O. Gabriela, 6 July 1990 (9); E. Gabrielian, 12787 eN); R. Gagnidze & I. Mikeladze, a eo 1965 (19a); € nidze & Tschelidze, 4 May 1976 (22); A. Galán & E. Monasterio-Huelín, 9 June ns (24) J. S. Gamble, 30247 (15), June 1914 (15); Gander, 17 July 1806 (8); M. Gandoger, 1880 (22), Aug. 1899 (15), June 190? (6d), Aug. 1900 (15); X. E García Martínez, 28 July a^ (6a), 29 July 1984 (6a); X. R. EU Martínez et al., 21 Apr. 1985 (24); I. García Mijangos, 21 Apr. 1989 ai. M. A. García, 953 (22), 960 (22), 5 je "es (22); ¢ June 1867 (14), 28 July 1870 (9); M. F. & S. G. Gardner, 2202 (6d), 2586 (2), 2674 (9); G. G initi: June 1879 (23), June 1880 (2), 20 July 1881 (15), 21 July 1886 (15); О. Gavioli, 20 Apr. 1924 (22), 14 May 1924 (22), 3 May 1931 (22), May 1931 (9) 12 May 1933 (22), June 1938 (22); 1. Gavrilov, 1179 (12), 1 July 1928 (21); Gebler, 1); D. Geltman et al., 500 (2). 980 (2). 1174 Joh 1179 Gea, 1636 (5), 1650 (2), 1880 (5). 4 P. A. Genty, 10 Apr. 1901 (15); A. George, en 2 Aug. К 3); С. А. Сегага, 20 or ; loader, 21 May 1973 (О 5 (15); P. E. Gibbs, 69483 (6d); M i Alejandre, 225-88 (6a), 407-88 (ба), rt 88 (24). sas; (6b), 14 July 1990 (15); B. Gilliat- Smith, 554 (17), 2711 (9); J. Giménez, 16 Apr. 1991 (24); € L. Giraudias, 24 July 1893 (8); Girod, 1 Apr. 1884 (9), 2 July 1885 (23); Giuseppi, 50 (9); Godet, 8 Aug. 1891 (15); A. Golbek, 6 July 1910 (12); V. P. Goloskokov, 19 Aug. 1948 (12), 14 July 1952 (12); D. N. Golovnin, 158 (1); D. Volume 90, Number 3 2003 Álvarez Fernández Doronicum (Asteraceae) Gómez, 12 July 1990 (15); D. Gómez, A. Martínez & C. е 63591 (24); D. Gómez et al., 1 Мау 197 ак 86 ómez Hernández, 1 May 1984 (24); Gómez Мап- eque, T & Vargas, 2326 (6a); Hno. Gonzalo, = hie 1927 (23); W. G. Gorodetzky, 9 May 1916 (12): B. Gorskij, 22 June 1908 (12); J. Goudot, Aug. 1890 ig C. N. Goulimy, 228 (9), 3 May 1957 (9); F. Graf, July 1867 (14); Grande, June 1905 (9); O. Grebenchikoff, 5 July 1936 (2), 16 July 1937 (2), 16 July 1938 (9); W Greuter, 15544 (9), 15719 (2), 16225 (2); W. Greuter & erxmüller, 16941 (22); W. Greuter et al., 14749 (9 G o С. Grey-Wilson, July 1968 (9); I. Grintes- мї 3 cu, 11 July 2 (5); E. Gros, 11 June 1919 (24), 19 June 1929 (24). : Mn 1931 (24); R. Gross, 14 May 1914 (23), 4 (23), 26 a 1918 (22), 17 May 1918 (9); O. n 29 July 1923 (21); H. pe ae July 1875 (9). “July 1884 (9); J. Groves, 5 July 15), 13 July 1913 (9); V. I. Grubov & L. I. Ivanina, 6 Aug. 1945 (19a), July 1965 (12); I. A. —— des 1965 (12), 23 July 1965 (12); I. Grundl, M 13 (17); M. Mee р July 1905 (9); I. A. Gubanov, 19 s ay 1958 (12); S Gudoshnikov & V. Dirin, 20 July 1967 (1); сае p* (17), 286 (17); pres & Ludlow-Hew- itt, 2028 (20); J. Guicciardi, Aug. 1855 (9); K. M. Gui- chard, 3/59/TUR (22); E. Guinea, 16 Apr. 1960 (24); F. Guiol, 216 (22), 624/738 (9), 30 July 1931 (9), Aug. 1932 9); A. Gulyás, 21 May 1906 (9), 22 May e (9); A. Güner, 6121 (19a); A. Güner & M. Vural, 6115 (19a), 6643 (19a); A. Güner, M. Vural 9272 (22); H. Günther, 25 Aug. 1939 (8); H. : 18 1906 (10). . Hackel, 26 May 1876 (24); J. Hafellner & S. Titze. oe (0 ); I. Hagemann, H. Scholz & W. Schwarz, 85 (22), 187 (9), 332 (9), 553 (9); H. Hal, 13 Apr. 1869 (9); Hal- ácsy, July 1878 (2); Hamilton, 10 Sep. 1884 (15), 21 July 1859 (15). fue 1892 (8); Hanbury-Tracy, 22 (3); Handel- Mazzetti, 763 (19a), 1343 (26), 1765 (3), 18 July 1906 (15). 17 July 1927 (9); A. Hansen & Н. Nielsen, 1440 (22); A. Hardy, July 1881 (15); P. Hariot, 23 May 1873 (24); R. M. Harley, 17721 (2); R. M. Harley & D. Peev, 11907 (2); G. W. Harris, 259 (22), 433 (22); T. D. Harri- son, 18 May 1978 (24); P. Harrold & R. J. D. McBeath, 158 (60) 240 (6b), 252 (6b), 346 (15); Harz, 20 m 23); E. Hausser, 21 Apr. 1885 (9); Haussknecht, : Aug. 1865 (16); A. & F. v. Hayek, July 1907 (8): od: 16 July 1913 (5); Hayes, 1839 (9 ); M. Heard, Aug. 1926 (15); Heiland, May 1867 (22), July 1878 (23), 23 Apr. 1879 (9); T. Heldreich, 122 (22). 1249 (22), 25 Apr. 1828 t Apr. 1843 (22), May 1844 vi June 1844 22 Y Н zm : (22); Heltman, 28 June 1958 (2); W. Hempel, 2786 (8), 16 June 1964 (2), 17 June 1964 (9); J. Hennecart, 24 July 1850 I E. Hennipman et al., 475 (22), 741 (22), 1272 (22); К. М. Hepper, 4739 (23); К. Heras & J. А. Alejandre, 1192- 85 (ба), 1982—85 (ба); J. Herranz et al., 1 June 1990 (24); A. Herrero, 2 Aug. 1996 (15), 23 Aug. 1996 Mee „ Herrero, 2 May 1987 T Pilas 44 M 45 (9); P. Hiepko, 1 9); A. W. Hill, 1896 (15); T. N. Ho, A, i tholomew & M. Gilbert, m o E. Hodgkin, 168 (14), 220 (14); H. Hofmann, 29 July 1901 (2); R. F. Hohen- acker, June 1834 e May 1838 (21), 1838 (19a) 2, June 1842 (19a); Hollós, 13 May 1914 (22); Holm-Nielsen, n 1966 (23); D. Hüner & S. ei 1673 (22), 1718 (9); J. D. Hooker, 1862 (15); F. Hópflin- ger, 30 July 1950 (7). 1 Aug. 1961 (8); J. F. Horcajo, 6 May 1984 (24); J. Houska, 11 May 1939 (9); A. Huber, 17 Aug. 1930 (15); Huet du Pavillon, July 1853 (20), July 1854 (2); E. & A. Huet du Pavillon, 362 (9), 12 June 1855 (22), 19 Mar. 1856 (22); J. Hulják, 26 Apr. 1913 (17), 26 May 1913 (17); L. H. Hurst, 71 (9); J. Hutchin- son, 29 May 1958 (23); Hutchinson, Matthews & Riley, 60 (15), 156 (15); R. Huter, 1126 (9), 1814 (9), 17 July 1871 (2), 17 July 1878 (2), 20 July 1878 (9), July 1880 (8). Aug. 1888 (14), 18 July 1907 (9); Huter, Porta & Su d (22); F. Ch. Hy, Apr. 1903 (24). М. Igoshina, 19 June 1950 (2), 28 June 1950 (5); i. 15 June 1909 (1); ise, 1 Aug. 1867 (8); Inayat, 19662a (18), е (18 уап! Е (19a), 7 July 1964 (5); noe 12 July 1913 (12); 1. Izco & R. Mart., 25 July 1967 (6c); Izco, Ladero & Demetrio, 23 May 1968 (24); A. Izuzquiza et al., 146 hes 447 (24). Jabornegg, Aug. 1875 (15), 1886 (8 B. Jackson, 565 ges 11 Apr. 1927 (24); aont г 1821 (2), 1822 (10) (15); F. Jacquemoud, 3459 (23), 3888 (14); E. мамі. i (24), 436 (24), 10 May 1907 (24); J. "ri ly 6 (15), June 1890 (2), 19 May 1899 (22); D Hs July 1914 (12); Janka, 18 May 1884 (17), Apr. 1885 (17); Jaquet, 12 Aug. 1904 (15); U. Jath, 11 S ind (22); S. Jávorka, PF 1913 (9), 27 Apr. TI : ert, 7 June 1 24), 20 May 1894 (23), 2 M: 1894 (2: 3), 27 May 1897 (24), 6 May 1900 (24): ыч my, 1868 (2); A. J. Jhorp, 13 July 1886 (15); G. Jlié, 1887 1868 (15), 5 May 1870 (23), 22 July pes S. P. Thornton- Wood, 9876 (2); yse Jackson 6638 (9); S. Juzep- e (19a). . Karelin & Kirilov, 463 (12), 1621 (12) 20 Aug. 1932 (19a), 8 July 1935 (19a); I. Karjagin & A. "pps 20 Aug. 1932 (11) (19a); I. j ev, E dvi 1931 (21); I. pad & J. Я ons (1 Kasumova, 30 June 1929 TUM 15 July 1930 m. ау & Yaltirik, 3368 (22); C. Keck & T. 1879 (14), 17 July 188 Бак 40/8 dh > a S E x3 gs Куг; со ~J] л : ingdon War ‚ 4711 (26). 5866 (1) (3). 10876 (12), 12125 (3), ot (3), 1913 (3); Z. Kiogkova, 127 (19a); Kirpichnikov, 27 May 1948 (21); Klemenc, 1894 (12); B. Klopotov, 28 June 1909 (1); J. Knoph & R. Vogt, 2407 (23); O. mpi ca 12 July 1913 (12), 31 July es 2); O. Knorring & Z. Minkwitz, 20 Aug. 1911 (12); . 123 (14) (15); ы La 20964. (18), 22052 (18), 9429 (2, 9716 (12); St. Kogeoucharov, 50 (9); Е. Köhler & C. , 13 July re (19a); A. Kolakovsky, 10 June 1927 TON 17 June 1929 (21), 30 June 1929 (19a); Ko- vie 1663 (19a), June 1844 (21); P. M Kolovski, 1913 (1); V. 1. Komarov, en 1902 (1); : gy s i (12); A. Konnov et al., 742 (12); E. E 8 July 1908 (2). 24 May 1933 (17), т ре 1936 (23); Е. Korotkova & Z. Klimovskaya, 112 (12); S. Korshinsky, 1504 (12); Kosan- in, 14 June 1924 (9); T. Kotschy, 147 (22), 363 (20); W. Kotte, 18 ws 1932 (22); тна, 100 (17), July 1843 (17); С. Kozij, 16 July 1938 y (2); V. Kozlovsky, 7 Aug. 1928 (21. 24 May 1936 (21); Kralik, 538 (10), 538a (10), 638a (10), 16 May 1844 (24). 18 May Merry 15 May 1846 (24); K. U. Kramer, 1366 2), 8672 (15); I. Krasnoborov, 8172 (1), 8179 (1); I. Kras- may & Chanmicun, 240 (1); I. Krasnoborov & Ersho- 22 Aug. 1962 (1); 1. Krasnoborov, Hrubov & Jakov- ye 959 (1); I. Krasnoborov & Merzliakova, 8181 (1); I. © =ð, 1e {л pa = ZB == 386 Annals of the Missouri Botanical Garden 4 Aug. 1962 (1); Ё Krasnoborov & E. Se therbisc hkii, A (1): Р. N. Kr ios & E. ; ДИ чс жөр 4 Tine 1916 7 July 1916 Ж); M. Kuhn, 21 Aug. 1864 (2); А. Капта & Pavlova, July 1947 (1); J. B. о 535 n. B. Kümmerle, J. Szurák & G. Timk6, 28 Apr. 1912 B. Kümmerle & G. Timkó, ef: 19) Kuptok, 3 ne 1898 (8 ); Kuschakewicz, 14 July 8 (12); B. E manov, 76267 (9), 80713 (9), e n prs 80 (9); N ; Kuznetsov, 927 (1), 2180 (1), 4203 (1), 20 July 1912 Em 'aita, 182/20 (2: үй J. H. Lace, 13 Aug. 1897 (18), Aug. 1899 (18); M. Ladero, 17 Apr. 1965 (24). 1966 (24), 6 Apr. Dus (24), 29 Apr. 1968 (24); M ero, b ‘ano, del Águila & M. Sánchez, 21 Apr. (24); M. Ladero я Rivas Martínez, 10 July 1974 (15); aínz vi Sánchez Pedraja, 7 May 1991 (24); J. Lambinon, 86/C 20/264 (10); qeu 4876 (19a); Lamotte, 14 July 1867 (2); C. R. Lancaster, 120 (9), 160 (12), 188 (12), 206 (18). 221 (18), 23 July 1979 (19a); T. E. Lankester & T. A. S. Pearson, 1357 (12); Lansac & Pu Feliner 1469 (6a); E. Launert, 25 July 1955 (14); A. Lawalrée. 15798 (24), 26010 (23); С. 24); M. Laza, 19 Apr. 1935 (24); C. . 48 (23); p un 468 (1); Legrand, 18 Apr. 1893 (24); Lehmann, May 1904 (23); F. Lemperg, 307 (9): H. DE 20 Apr. Ure (22); С. roe 313 (9); С. Leredde, y 1948 (23); L. Leresche, 28 May 1884 (17); A. I. ak 2 July 1928 (19a), m 1930 (22), 16 June 1930 (19a), 15 July 1930 (21); A. I. Leskov & A. P. Rusaliev, 24 July 1929 (22); ie ig 1011 (24); К. Levier, 9 Aug. 1874 (9); Lewin, 20 June 1892 (1); Lewis, 3 (23); E ‚ 4730 (26); Liebenow, 17 Aug. 1960 (5); W. Lippert, 452 (14), 453 (15), 22 July 1963 (15), 28 May 1966 (2); W. Lippert & D. Podlech, 25818 (9); Lip- pert & Zollitsch, 22 Apr. 1964 (22); S. J. Lipsc ocn 2D ( 2); V. I. Lipsky, 821 (12), 1179 (12), 1 (12), 7 (12), 3587 (12), 6 July C P 2). Б, 1890 (19a), 12 May 1895 (22), 19 June a; A. Liston, 818-1 (12); D. Litvinov, 16 June , 8 July 1914 (11); P. Litzler, 75/837 (2); P. S Lloyd ^ е Megan, 81 (12); F. Lobbichler, 559 (12); А. Е, Lomax, 14 July 1892 (6d), 13 June 1893 (6c); M. Lomo- ). Shaulo, 732 (1); M. Longa, 31 July 1911 (8). 1919 (8), Aug. 1920 (15); G. López, 2042 (24); G. López & R. Morales, 2310 (24), 3008 (24); G. López. G. Moreno & E. Valdés, 23 July 1975 (15); López, Mirones, Peral & Sánchez Pedraja, 3 July 1994 (6d); M. J. López Pacheco, 23 May 1979 (24); Lorenz, 27532 (15), 27535 X M Losa, June 1929 (ба); Losa & Montserrat, 7 Aug. 1948 (2). "E 1950 (6d); Losa x е Goday, May 1959 (24): . Lousley, 954 (23), € 24), 14 May 19064 (23), 25 Tune ne 1968 (15), 4 Aug. 1 e ); H. G. Lübeck, May 1878 (9). June 1880 (23), Ss 1881 (23); M. Luceño, F. Muñoz . Lucefio & P. Vargas 208 (6a), 2569" (ба), 31788 (€ (6 за), 7 als 1986 (ба); (18), 682 (12), 26 June 1939 (18); F. Ludlow 1505 (12), 2369 (3), 7719 (18), 7850 (1 2), : 9292 (18), 9360 (12), y — T— -— ES - > ~ — — - > A 0 جا‎ эз z Jy ^^ oo! - : i EL (3), 52: 58 (3). 5 0 (3). 10); J. Madalski, Ae pe (5). 3 488 (19); D. Mai, 7 July 1986 Us n 19: 38 (8); | Vy n, Maire & M. Petitmengin, 876, 1906 (9); om 196 233 (24), 979 (6d); Malato-Beliz et al., 3120 (24), ‚ 3982 (24), 4309 (24); Н. Malicky, dor (22), \ 22); L. Malyshev et al., 9 Apr. 1957 (1); К. Maly, 5 Aug. 1911 (2), 6 May 1950 (22); Manissadjian, 1 May 1894 (22), 25 May 1906 (22); V. Manakjan, 5 July 1962 (21); Mansanet & Ladero, 3 May 1968 (24); quet, 83/56 in E. Margais, 21 May 1885 (23); cowicz, 100 (22); A. Margittal, 952 s July 1917 (5 5). July 1933 (5); Y. Е Marin, 21 July 19 1); M. Markova, Z. Cerneva & Р. rog i 16 June T (9); Мае, 12 May 1878 (23); U с Mantelli, 10 dd 1893 i (24), May 1980 (24); B. Mathew & D. paa 49 (8): B. Mathew & A. J. Tomlinson, 4386 (22); t 422 (8); E. P. Matveeva, 25 Aug. 1930 (12); Matveeva & Tkatch- enko, 12 July 1947 (1); A. & M. ic 5094 (1); M. Mayor et al., 27 July 1981 (15); Н. D. McLaren, 144 (26), 167D (3); L. Medina, 10 Apr. 1998 (24); Medvjedev, 230 (22); L. 1. Medvedeva et al., 26 May 1950 (12); A. Mee- bold, 987 (18), May 1928 (23); R. Meinertzhagen, 13 May 193 33 (22), 7 June 1933 (9); R. Melville, 23 Apr. 1957 M. A. Mendiola, 5 July 1979 (6a), 6 July 1979 (6a), 17 July 1980 (6a); A. Mendoça, & J. Vase oncellos, 6202 (24); А. du et al., 5198 (24); Y. L. M June 1974 (21). i i W. Lippert, 25119 p G. Merzbacher, 880 (12), 5 (1: 2), July 1903 (12); C. A. Meyer, 674 (21); D. E. и 239 (2); Miller, July 1902 (2); madi, K-2381 (11); J. Molero, May 1974 (24), : 1976 (24); A. Monasterio, 10 May 1943 (24), 22 Apr. 1945 (24); E. Monteil, 2 July 1916 (15); G. Montserrat, 1 Aug. 1987 (23); P. Montserrat, 16 June 1958 (23), 7 Aug. 1958 (15), 4 Aug. 1967 (15); P. & J. M. Montserrat, 11 Aug. 1967 (15); P. Montserrat, J. M. Montse rat & il- y 1978 (23); P. Montserrat & L. 1980 (23); P. Montserrat et al., 20 Aug. 1991 (2: 3), 1993 (15), 9 Aug. 1993 (15); H. Mora, 1836 (24); Moreno Moral, Patallo & Sánchez Ped- raja, 882/96 (15); Moreno Moral & Sánchez Pedraja, 496/ 96 (6d), 5 July 1991 (6d); P. Morthier, 23 Aug. 1883 (15); Moussavi, Habibi & Tehrani, 20 June 1983 (11); М. Е ER 2] July 1878 (2), 23 July 1882 (2); Murr, 1880 15); m et al. 106 (9); m 11120 (2) (23). geli, 3 Aug. 1837 ; Naumann, Nava md J. Valle, 28 dna 1982 (6x y; € July 1982 (6b). M May 1984 (24), 15 july: 1985 (6a); I Neé, June 1786 (24); V. Nekrasova, 2 July Nekrasova & L. Aleksandrov, 43 (22 : (23); Nendtvich, May 1866 (22); P. V. 'Nes "sterov, 25 June 1907 (1), 16 June 1910 (11); A. Neumann, 2 May 1959 (23); F. Niedereder, 9 July 1904 (2), 16 June 1905 (2); G. 2); G. Nieto oe 3888 (22): G. Nieto ‚ 1543 (ба Nieto Feliner et al., 2736 (Өс), 30 May jg (23); Yoda 647 (11); Norris, “к 1945 (22); C. & M. North. 42 2 (22); Nüsser, 82 (12); Nyárády, 28 Aug. 1911 (2), (2); M. Муде gger, 17340 (20), 19037 (25), 40085 (22), 40323 (22), 40766 (22), 16 Apr. 1976 (22); К. Nyman, Apr. 1844 22). Oberleitner, 73 (15), 8 July 1864 (2), 30 July 1865 (14), 19 July 1867 (8), 21 July 1868 (8), E June 1869 (2), 2 Aug. 1872 (8), 28 July 1874 (14); L. Oberneder, 5817 (15), 5935 (2); R. & L. iih 6497 (8); A. Oborny, 12 Aug. 1878 (2); H. Ocakverdi, 2340 (11); V X = I ә», Volume 90, Number 3 Álvarez Fernández 387 2003 Doronicum (Asteraceae) ; B. M. Ogievskie, 25 June 1913 (1); T. С. Orphani- (11); Raus & Royl, 5057a (9); A. Rawi & I. Serhang, des. 196 (22), 340 (9), July 1854 (9), July 1854 (9), : May 1857 (22); B. B. Osmaston, 28 (18); B. Ovchinnikov & M. Usov, 301 (12); Owerin, 26 June 1861 (19a); Ch Ozanon, 3 Aug. 1858 (15). H. Pabot, 1717 (20); J. Paczoski, 6 May 1901 (17), s Apr. 1909 (17); J. Paillot, 2279', 24 May 1859 (23); F jarón, 683 (24); Palacio, Carrillo & Ferrero, 5 July 1907 (6c); P. Palézieux, 1 Aug. 1898 (9); I. V. Palibin, 5 June 1908 (8), 7 Aug. 1908 (15); A. Pallarés, 15 Apr. 1990 (24), July 1994 (15), May 1996 (24); W. Panknin, 24 May 1936 (23); J. npe dg: 1868 (17), 8 Apr. 1906 (17): J. Papp. Apr. 1946 (22); D. Parascan, 7 June 1960 (5): D. Parascan, К. hein & D. Radu, 30 June 1960 (9); M. Parda ri Santayana & R. Morales, 1690 (64); Parla- tore, 18 Apr. 1856 (9), 1863 (22), 1866 (2); Parseval- yi nage 31 May 18?2 (23); N. L. Pastushov, 20 Apr. 1925 Ne Patino, Uribe-Echevarría & Valencia, 27 May apa Patrin, 1780 (1); Patzak, Sep. 1953 (2) (7); : Pau, 2386 (24), 27 July 1900 (бс); A. Pavai, 1875 (17), 187? (9); М. V. Pavlov, 601 (12); V. Payot, 1855 (15): S. Peker, 1178 (22); G. dran е Мау 1913 (9); А. М. Е. García, 31 Mar. 3 (24); А. i M. Herrero, 3 May en (24); E. Penkovskaya ; 1. Krasnoborov, 1 July 1964 (1); R. V. Pennington, 22 (9), 25 (9), 30 (9), : „ Pérez Chiscano, | June л 1991 (6a); C. Pérez Morales, 27 July 198 Morales et al., 11 July 1992 (6d); E. Buc 1849 (23); C. Persson, 5 Aug. 1934 (12); H. Pesmen, 727 (22): H. Pesmen & A. Güner, 2213 (22); S. Petrovič, 2200 (17). 2340 (22), Apr. 1882 (22), Apr. хааа ALTA Apr. 1887 (22). June 1887 (2); F. Petzi, 193 (2); V. V. etw 108 (12); T. Pichler, 81 (22), 21 diu 1864 (15), (9), June 1872 (9), July 1873 (25), June 1874 (22). July 1874 (25), Apr. 1876 (22), 1878 (14), May 1890 (9), July 1892 (14); C. Pinard, 1843 (22); Pirker, Royl & Fleischer. 315 (2); C. J. Pitard, July 1906 (15); E. Pobedimova, 45 (12), 339 (12); D. Podlech, 37607 (9); Poisson, May 1879 24); A. Polatschek, July 1969 (2), 9 May 1981 (23); А. ij. 18 July 1947 (1); О. Polunin, 56/170 (18), 5332 (22), 6238 (12), 8268 0) 14033 (22), 15054 2 Э, 15940 (22), 18 В Арг. 1956 (22); О. Polunin, W. R. i . 89 (18), 401 (18), 2622 (18), ud (18) 1923 — am, = TE m 3 КА E 2069 (1; M. G. P 325 (11); Popovic, 8 July 1954 (8); A. Toile: 23 July 1928 (19a); A. Poretsky & G. Schultz, 9 Sep. 1927 (19a): P. Porta, July 1911 (15), Aug. 1858 a i June 1867 (15), July 1889 (9), Aug. 1893 (15); a & G. Rigo, 160 (24), 257 (24), 319 (9), 29 ve ee 9: Post, 12 Aug. 1893 (22); G. N. i wv. -€—— -25 Jui 1885 (26), 6 Aug. 1885 . 24 June 1893 Rozhetits & Shishkin, E July 1940 (2); 26); Preissmann, 31 May 1886 (2); С. m & R. Vogt 5026 (2); F. Prenn, 13 Aug. 1945 (15); Fic 1830 (1) 22); Prescott-Decie, Lt. (18); M. P. Price, 1910 (1); W. R. Priee, 492 (22); L. Prilipko, 19 June 1932 (21); L. Prilipko & J. Isaev, i July 1934 (21), 4 Aug. 1934 (11): E. Pritzel, July gh ); N. E Przevalski, 101 (12), 333 d 7 July 1877 (12); M. I. Ptaschizky, 16 July 1908 12); " Puente & C. Pérez mulia 27 July 1988 (6b); J. Puyfol, 4925 Каар, 25 May 1895 (9); E. I. Rac sca & V.I. Grubov, 4 Sep. 1949 (2); G. к 40 (19а), 142 (21), 149 (21), 400 (19a); B. Rainha, 1047 (24); J. Ramsbottom, 1918 (22); V. Rastetter, 16 May 1993 (23); Raus, 4864 eras Ww т am, 24553 (11); K. H. Rechinger, 1701 (15), 3224 (15), 17881 (2), 20984 (9), 22674 а 23159 (22), 38768 (9), 54337 (22); K. H. & F. Rechinger, 851 (22), 661 (22), 3128 (2), 3794 e 3927 (22), 8758 (22), 10735 (2), 25 July 1928 (8); K. H. Rechinger sheffer, 711 (9), 13 Ж Кес бан 2377 (15); A. Regel, 686 (12), 19 July 1877 (12), 26 Aug. 1877 (12), 11 Sep. 1877 (12), 19 June 1878 0: 2), 20 June 1878 (12), 22 June 1878 (12), 24 July 1878 12, 4 June 1879 (12), 14 June 1879 (12), 15 June 1879 , 16 June 1879 (12); C. Regel, 1 June 1964 (22), 10 Ds 1964 (22); : | Aug. 1933 (8); J. Renz, 48989 (20); Renz et al., \ 1974 (11); Requien, 250 (10); К. Ressman, 1878 (14); B. Retz, 46062 (24), 67398 (24), 89985 (24), 17 July 1856 12); E. Reverchon, 16 Aug. 1872 (15), 18 July 1878 (10), 25 July 1878 (15), 28 Aug. 1878 (10), 21 July 1885 (10), 22 June 1886 (23), June 1898 (24); V. V. Reverdatto, 26 July 1942 (1); M. Reymond, 27 July 1931 (12); H. H. Rich, 1165 (18); A. Richter, 520 (17), 2772 (1). May 1872 (17), 13 May 1900 (17), 25 June 1900 (2), 12 May 1902 (9), 15 May 1907 (17), 31 July 1908 (9), 11 July 1909 (9); K. Richter, 17 July 1887 (2); E. Rico, 27 May 1983 983 es y E. Rico, X. Giráldez & T. Rom- 70 (9), 430 9. ~ 8 July 1870 (15 Rivas Goday, 15 May 1924 (24), 18 May 1941 ( (24), 30 Apr. 1944 (23), 22 Apr. 1945 (24), 7 Aug. 1946 (6c), 18 Apr. 1957 (24), 26 July 1958 (6c), 20 May 1959 (24), 1 ар ~ (24); S. Rivas Goday & F. Bellot, 30 Mar. 1940 2 . Rivas Goday & M. Ladero, 3 Apr. 1969 (24); 5. Tied Padus. M. Ladero & Valdés, 12 July "d (6a); S Rivas ( Goda ay & Monasterio, 10 Aug. 1947 (6a); S. Rüvas Goday, S. Rivas Martínez & M. Ladero, 28 June 1970 (ба), 30 [n 1973 (6a); 5. Rivas Goday & E. 11 July 1974 (6d); S. Rivas Martínez, 25 Aug. 25 July 1958 (6c), 13 July 1965 (15), 1 June IT (24), 24 Aug. 1978 (6a); Rivas Martínez, M. Costa & J. Izco, 11 July 1973 (ба); S. Rivas Martínez & J. Izco, 25 Aug. 1967 (6c); S. Rivas Martínez, M. Ladero & M. Mayor, 1 June 1966 (24); S. Rivas Martínez p E. Valdés وش‎ 7312 (6d); S. Rivas Martinez et al., 13 July 1965 (15), 13 July 1977 (6d), 16 June 1981 (64), : pr) 1982 (6a) (24), 26 July 1989 (6c); : Rivera, 25 Apr. 1978 (24); A. Roa, 22 June 1987 (ба); V. 1. Roborowski, 205 fa ; J. F. Rock, 9681 (3), 9747 (26), A (3), 12192 (13), 12389 (13), 12941 (26), 13020 (13), 14599 (26), 16566 (26), 16834 (3), 22380 (3), 22891 (3), 23067 (1); L. Rodin, 1231 (12); J. Rodrfguez-Oubina & I. Cruces, 20 June 1997 (6d); Rog- ers, 234 (23), 717 (22); J. Rohlena, Aug. 1912 (9), June 1922 (9), July 1933 (9); I. Roldugin, 4871 (12); i on 30 May 1886 (9), June 1892 (9); C. Romero, July 19 (15); К. Ronniger, 29 July 1883 (2), 29 July 1892 di : july 1918 (8), 20 May 1924 (9), 21 July — ), 24 July 1930 (2); R. J. Roshevitz, 11 June 1908 (12), 27 June 1908 (12), 7 July 1908 (12), 9 July 1908 (12), 1 Aug. 1908 (12), 3 July 1909 (12); H. Ross, July 1888 (22), May 1901 (22), June 1901 (22); S. Rossi & A. Malladra, Aug. 1890 (8) (15); W. Róssler, 8 (2); E. Rostan, 1850 (8), July 1860 (15); J. ب و‎ i July 1978 (9), 17 Apr. g. 1933 (6d); W. tuer g = = < — 2 Со: ра] w ‚тм, = -l > — N — =. gs 5 кі, — Sep. 1986 (8); Royl, Hempel & Richter, 19 Sep. 1986 (8); Royl & Ketelhut, 4 Sep. 1987 (15); Royl & Schiers, 1982 388 Annals of the Missouri Botanical Garden (8); A. Rozeira, 19 May 1946 (24); A. Rubio, 18 May 1991 (24); N. I. Rudstov, 13 July 1934 (12); Ruprecht, 154 (21), 22), 10 June 1861 ‚ 1861 . 1861 1291 (12), 1424 (12). 18 po (2); Saint-Lager, 3 July 1896 (22). 7 y ied 1919 (23); G. Samuelsson, 21 lI 1933 Sh 26 Apr. E 1 G. Samuelsson & nder, 22 A 1931 (22); ínchez-Mata, 9 June teg (2. 1), July 1984 (6a), 9 Sie ires Oak D. Sánchez-Mata, S. Laorge & D. Belmonte, 20 May 1982 (6c); Sánchez Pedraja, 21 Apr. 4) un shez Pedraja & Tapia Bon, 226/96 (24), 598b/96 (6d); №. Y. Sandwith, 3737B (23), 5340 (6b), 5684 (6a), 18 May 1935 (24), 30 June 1935 (23); V. Sa- poshnikov, 16 June 1902 (12), 30 June 1902 (12), 6 July 1912 (12), 22 June 1913 (12), 13 July 1913 (12), 14 Aug. 1923 (1); V. TT & Shishkin, 15 July 1912 (12); : Saposhnikov & T. Tripolitova, 7 July 1915 (12); F. Sap- a & E. E. Galiano, 9 Aug. 1952 (15); S. Sardinero, 3 Tite 1990 (6a), 12 July 1990 (6a), 14 July 1990 (6c), 3 Aug. 1990 (6c), 23 Aug. 1990 (6c), 21 June 1991 (6a), 12 July 1991 (6c), 19 July 1991 (6a), 25 July 1990 (6c). 29 July 199] 108) E. M. Saunders, June 1915 (18); D. Sauter, 971 (14); D. Sauter & A. Traur isteinet, May 1836 (14); Schafferer, 16 Aug. 1890 (8); K. Scheer, 5 May 1918 (22); A. B. Schelkovnikov, 29 June 1909 (21 ), 15 July 1909 (21), 2 July 1911 (19a); A. B. Schelkovnikov & E Kara-Murza, 13 July 1927 (19a), 15 5 Tus T T КЕ > { > = s $ © dE e S = = en О = J 8 l, July 1880 (8), 28 duly 8); М. Sc Кашы 1137 (22); B. Schischkin, 26 May 1912 (12), 7 June 1912 (12), 2 Apr. 1917 (22), 6 July 1919 (19a), 4 July 1920 (11); B. Shishkin & G. Sum- nevicz, 24 June 1931 (1); ne 5902 (3). 1855 (18); F. Sc D. 438 (12), 5989 (11); . Schneider, 11 (9). 300 . K. Schneider, & Lait 881 (2); J. Schne ah 22 Mo 1929 (17); Schneller, 23 June 1859 (17); H. Scholz & P. Hiepko, 988 (24); M. Schreiber, 10 Aug. 1911 (8); H. Schrenk, 8 Aug. 1925 (15); F. Schuh- 2), 2] ; R. Se kala 6 M 23), 19 Jub 1894. 2). 27 Tuly 1899 (14), 31 July al | Aug. 1899 (9), 6 July 1902 (23); R. & O. Schulz, 18 May 1895 (9), ied 1896 (9); A. Sc jn her, 1 June 1933 (23), 4 June 968 (23); Schur, July 1839 (2), 27 July 1850 (5); O. Se hwarz, 642 (22); Schwarzer, 28 Apr. 1848 (9); Schwerdt- К A. Segura Zubizarreta, 12525 (ба), 527 (6a), 22654 (6a), 16 May 1970 (24), 10 June D^ (24), 10 Apr. 1961 (24), 8 June 1967 (24), 23 Apr. 1970 V Seidl, 90 (21); Seidlitz, 33 (19a); R. Seligman, 53 (9), 103 (9); Sendtner, 301 (9), 302 (2); F. Sennen, 1973 (15), p^ (2). 4433 (15), 3 Aug. 191? (15). May 1910 (23), 4 Aug. 1916 (15), 10 Aug. 1916 (15), 27 July 1931 (15); К. Sennen & Elías, 1906 (24); M. Serim. 17 S (22); J. A. Sesé, 7 July 1992 (23); J. A. Sesé et al., 18 July 1987 (15); Sharif, 224 (11); B. A. шы е ni 1. G. Knorring, 17 Aug. 1937 (12); G. Sherriff, 7405 (12); Shirokova, 6 May 1948 (17); H. Sholz, 19 May 1974 (23); T. B. Shrestha, п (18). 5180 (18); о 3809 (11); Cup dis 15 (9); К. G. Sieber, 85 (15); W. Siehe, 118 (22). 1895 (22); F. i Silva Pando et al., 1394 (6d); L. Simkovies, 20 Apr. 1873 (22), 9 Aug. 1883 (5); Simonkai, 1816 (5); P. Sintenis, 80 (22), 196 (22), 244 (22), 414 (2), 2), 733 (9), 882 (22), 3844 (22), 3902 (22), 3995 ‚ 9929 (22), 7173 (19a); P. Sintenis & J. Bornmüller. 17 May 1891 (22); W. Sladen, 9/4/451 (2), 9/4/452 (9); J. м ai^ TT Smarda, Aug. 1934 (8); P. Smirnov, 10 Aug. 1930 (1); V. I. Smirnow, 314 (1), 2968 (1), s s 1910 (1), 4 Aug. 1936 (1); A. M. Smith, tre it ; H. Smith, 3270 (3), 3477 (3), 3587 (26), 3880 oA, (3); Soc. Rochelaise, 1881 (24); N. D. d 76, (12), re L (2 Soleirol, 2303 (10); ;F M Sofiora, 19 dune 1994 (б T , 82-119- 187 (12), 319 (12), 428 (12); M. Sovetkina & S. Chausova, 1963 (12); V. Spitzel, 1150 (15), 972 (2), May 1836 (2); Sredinski, 1873 (19a); J. . Stainton, 8113 22), 7381 (9), 7400 (18), 7894. (18), 7961 (9), 8011 (2); ). A. Stainton & Henderson, 6197 (19a); J. D. А "tai ton, Sykes & Williams, 3145 (18), 6051 (18); $ (18); H. ON ine July 1881 (2), July 1883 (2), 5 1887 (14), 6 Aug. 1887 (2); W. Steinitz, 24 May 1879 (17), 10 May 1881 (17), 20 Apr. 1882 (17), 10 May 1889 7); W. Steiniz & J. Briquet, 2 May 1880 (17); H. Sterk- ing et al., 28 Apr. 1885 (9); Sterneck, July 1904 (15), July 1908 (9); C. Steurer, 17 Aug. 1886 (15); H. Stevens, 392 (26); J. L. Stewart, 1733 (18); R. R. Stewart, 350a (18), 9367 (18), 6854 (18), 10370 (18), 12638 (12), 20301 (12), 20755 (12), 21539a (18), 21839 (12), 23132 (18); R. R. & I. D. Stewart, ee M rs (18); A. Stork, d (22), 17 May 1987 (22 Stranskij, 25 July 1911 (2); T. Strauss, m 1898 b i ‘Sttibrng, 10 June 1889 (22), 17 Apr. 1893 (9), 13 May 1893 (9), May 1894 (17), 19 May 1895 nes 9 May 1896 (17), Aug. 1896 (2), 11 May 1898 (22), 14 May 1899 (22). 5 Aug. 1899 (2), 10 Aug, 1899 (2), 12 Aug. 1899 (2), 26 May 1900 (17). May 1901 (17), Aug. 1903 (2), July 1909 (2), May 1910 (17 7) (22); Strid, Christiansen & Laulund, 26639 (22); Strid, Christiansen, Moller, 26147 (22); Strid & Papani- colau, 16449 (2); Strid et al., 18182 (9), 18205 (9), 18862 2), 19540 (2); I. Strupinsij, 25 June 1913 (1); Stubben- dorf, 1848 (1); E. Stuckenberg, 29 June 1912 (19a); Stud. biol. Rheno-Trai., 327 (15), 68-1979 (2); Sukatev, Ras- ч Ls Brizshev, 1606 (1); V. S. Summerhayes, 2035 (23), : V. S. Summerhayes & P. F. Hunt, 3767 23); "i Susanna, 776 (24); J. Suza, 14 July 1925 (2); J. T. Syme, 653 (23), 654 (24); "scu 6 May 1885 (17), 1890 (17); A. Sztehlo, Aug. 1876 H. Tabata, Rajbhandari & Tsuc m 1072 (18), 3309 18); H. Tabata et al., 12900 (18), 19355 (18); K. Taman- jan, 4 June 1988 (21); J. Tamemshian & W. Maleer, 31 July 1931 (12); Targioni, 12 July 1857 (2); A. Tatli, 5171 11); ees Apr. 1865 (17); J. A. Tauscher, 27 May 1875 (17) . Tebbutt, 15 ee di (23); H. G. Tedd, 1648 2 mere 123 A. Tengwall, 278 (22); Terme 0 : Aug. 8 (11). p^ Fans 1970 (20), 20 July 1971 (19a); Te on Moussavi & Habibi, 26 June 1978 (11); Terai July 1871 (9); C. G. H. Thedenius, Aug. 1903 (23); W. Thesiger, 1154 (11); Thomas, 1869 (23); H. P. & H. E Thompson, 199 (9); Rev. & H. P. Thompson, 77 (9), 6 (9), 523 (9); Thompson, 1ш 192: ); М. K. Timins, 174 (18); Timo- 15); V. S. Titov, 15 June 1910 (1 ), 22 June 1910 (1) 1 July 1914 (12); V. I. Tkatchenko, 182 (12); С. Tobey, | (22), 570 (22), Pis (22), 1484 (16), 2563-3 (22); ' . 931 (22); . Tolmacev, 10 July 1964 (2); M. Toma, 7 June 1967 (1 n. . Topa, 3482b (9); S. Topali, 23 Aug. 1937 (9), 8 July 1938 (9); A. Topitz, 8 July 1885 (2). 20 June 1887 (22); T (22), July 1889 (22); Torges, 21 Ma: July 1887 (15), 6 Aug. 1887 (15), 20 Apr. 1902 > (23 ), — jui. ~ — — rs T — Volume 90, Number 3 2003 Álvarez Fernández 389 Doronicum (Asteraceae) June 1902 (2 July 1903 (15), 13 Au y 1912 (15); C. C. Townsend, 84/55 (15), 90/247 (2), 93/574 (15), 94/32 (2), 4 June 1952 (23), 25 May 1953 3), 21 July 1902 (9), 30 July 1902 (15), 22 g. 1905 (15), 30 July 1907 (15), 15 G. Treffer, June 1878 (14), 13 July 1878 (8), vá 3 July 1888 (14), 24 July 1890 (14), 13 July 1896 . 30 Aye aeos 4); A. W. qs 2 1932 (23), ; W. y Trevelyan, 4 May 1846 (24); L. C. 14); W. P. Tre rr 14 pon 1912 (1), ; N. Tsvelev & S. Cherepanov, 1032 ; A. I. Tiari, 17 June 1912 (1); ); Tuhrman, 46 (1); бе иш, 1828 (1), 1836 (1 ); С. Turesson, 848U (1); Turland, 314 (22); J. Tuzson, 23 July 1914 (5); Y. Tzin- zerling, 2015 (1). J. Ullepitsch, 88 (8), July 1894 (8); barría, 228 (24), 1 1985 (6a), 30 June 1987 (2 Urrutia, 14 July 199 Valdés, Demetrio & E 14 May 1969 (24); E. Val- dés Bermejo et al., 4233 (6d); E. Valdés Bermejo et al., 5812 (6c); V. Valdés Bermejo, 24 July 1972 (6d); O. Val- entine, 28 July [20 (8); Vallin, 20 May 1911 (23); E. Valon, 3 (15): 1 (5), Sep. 1891 (15): | g. 1975 (15); P. Vargas, 2151 24); V. Vaxák, 26 May 1973 (22), 8 Oct. 1974 (21), 9 Oct. 1974 (21), 13 July 1975 (21), 27 July 1975 (19a), 4 July 1982 (21); V. Vašák & G. Esvandzhia, 22 July 1979 (11), 24 July 1979 (11); ^s Vašák & A. Zlatník, 28 May 1974 (12); G. B., F. G. & J. Vasconcellos, 7877 (24); Y. Y. Vasiliev, 15 May 1930 (22); S. Vautier, 23 May 1954 (23); Veitch, July 1912 (14); A. Velasco, 6 Apr. 1977 (24), May 1980 (24); M. Velayos 7574 (64d); B. Verlot, 27 Aug. 1879 (15); Veselsky, 2 (17); J. Vetter, 30 June 1916 (2), 11 June 1931 (2), e. 1932 (2); E. Vetterhall, 18 June 1898 (9); C. Vicioso, 22 Apr. 1908 (24). May 1912 (24), Aug. 1912 (6a), June 1916 (24), 5 May 1918 (24), 18 June 1918 a 5 June 1934 (24); C. Vicioso & 2 (6a); S e Vidrina, 3 July 1971 (1); . Villar 5 TA 974 (2); ~, P. M. Uribe-Eche- 3); P. Я MN 17 Aug. 1 үа (15); 1897 ; J. Vivant, 14 May 1970 (24); (24), К; 4), 3502. (6c), 3960 (15): erprieler, 13937 (22), 13942 (22), 14039 (9 ): C. Prem, 7069 (15), 7315 (15), 7327 (15); V Volkóv, 185 (1); E. V. Volkova & L. I. Ivanina, 5 July 1964 (2). 10 July 1964 (2); E. V. Volkova, Ivanina, Egorova & Matsen- ko, 7 July 1964 (5); E. A. Volkova et al., 7279 (12); V. J. Vorobiev & B. N. Klopotov, 1906 (21); Vrabélyi, 11 May 1867 (17). 10 May 1870 (17); Vukotinovié, 17 May 1874 (2). 20 May 1880 (2), 29 May 1880 (2), 23 June 1880 (2), hps 1897 (2); ч = a 1849 >. Wagenitz, 730 (15); J. Wagner, 76 (2), 77 (9), k 1897 (1); Walger, 29 May 1941 (22); E. Wall, 7 433 (22), 24/535 (22); ns 1918 (22); H. & E. Walter, 342 (22), 641 (25), 27 July 1967 (2), 28 July 1967 (2); H. Walter, E. Walter & a "Bilger, 4561 (22), 4693 (22); Walz, 25 June 1900 (9), 30 June 1900 (5); C. W. Wang, 64905 (3); G. H. Wang, 91161 (13); W. ыша, 15776 (15); С. L. Webster & J. Sack, 5715 (18); W. Weck, 22 Aug. 1891 (15); Weddell, July 1840 (23); Н. Weigold, Dn Weiler, 4 Aug. 1956 (12); Welwitsch, Apr. 1840 Werdermann & D. Meyer, 29 (8), 133 (9), 181 n T Re Mg 1893 (22); F. J. Widder, 20 Aug. 1923 5 (7), 1 Aug. 1926 (7), 20 July 1928 (15). 16 y 1947 (15); е а. ae (9), 2155 (2), 12 May 1843 (9); J. м күт 28 July 1864 (14); C. Wigram, 45 Tn Wilander, July 1871 (23): E (23), 16 May 1910 (23), 27 July 1911 (9), : (15), 22 July 1928 (15); F. E. Wilde, 4026 (22); Willing, 2896 (22), 3281 (22). 3515 (22), 3820 2 4832 (9). 6884 (2), 7054 (2), 7071 (2), 8305 (22). 9102 (22), 9126 (9), 9217 m 9566 ө, 9614 ( (9), 9885 Я (22). 14695 (9), 15139 (22), 15152 (9), 15649 (22), 17141 (9). 17596 (9), 22806 (22), 23151 (22), 26433 (22), 26604 (22), 29069 (22). 29732 (22), 30047 (22), 33124 (22), 33327 (22). 34599 (22); Wi TEE 1146 (24); E. H. Wilson, 3870 (26); Wimmer, 1857 (2); Winter, 107/3276 (8). 10 Aug. 1892 (9); Wirtgen, 1 June 1857 (23), 31 May 1895 (23); T. Wisniewski, 392 (21); Witte, July 1918 (15); E. Witting, оа 883 (14); Wittmann, 294 (19а); J. Wolff, Мау 1886 ; W. Wollny, 26 July 1902 (2); J. J. Wood, 188 (22): P ta 14 June 1887 (24), 28 July 1905 (8); Y. Wo- ronow, 18 June 1903 (1 y L DE 739 (1); W. C. Wors- dell, June 1895 (23), June 1904 (23); J. n 66 (15), July 1875 (15), July 1882 Tw July 1885 (15), Aug. 1885 (8). July 1887 (14) (15), Aug. 1887 (8); T. Wraber, 9748/4 (15). 13 July 1959 (15), 22 Aug. 1959 (2); Wright, 25 July 1898 (15); W. Wróblówna et al., 6 July 1952 (8); J. hi Wyatt, 28 (14), 45 (2), 56 (14), 59 (15), 66 (2), 69 (2), 8 (14), 114 (15), 124 (15), 291 2) Xatard, 182 (15), 1821 us N. I. Yakubova & B. Y ~ D > 2 un = = E 2 ©; — 9 e м = c — = e ae da -- NC July 1855 (6a), : i 12137 (3), "di (26), 12871 (26), 14593 (20). o, 22268 (3); A. A. Yunatov, 10071 (12), 13101 (12). IL sd dn 18 July 1974 (19a), 12 Aug. 1977 2: M. Zamfir, 17 June 1983 (9); O. pni 10 May 1925 (22); O. m APA et al., 9 Aug. 1928 (19а); Zeh- zad, 1305 (11); Zeller. 15 May 1904 (23); H. Zerny, 19 m VA a 2] july 1930 (2); W. Zovits, 1829 (21), 7 0 (22). A SYNOPTIC REVIEW OF Peter Goldblatt’ THE AFRICAN GENUS HESPERANTHA (IRIDACEAE: CROCOIDEA E)! ABSTRACT Although revised in the past 20 years for the two major centers of its range, the southern African winter-rainfall zone and the eastern southern African Drakensberg, the ов African genus Hesperantha remains inadequately understood. цит — exploration in southern Africa has resulted in the discovery of several new species and of populations of species known only from the type or very few collections. Species are listed here in a revised taxonomic order toge e pei keys bai the genus in the southern African winter-rainfall zone aes in eastern southern Africa and tropical Africa that have a summer-rainfall climate. Included here are 11 new species, a shift in the application of the name Н. ا‎ to olen called H. vernalis, recognition of H. leucantha for plants previously nés H. candida, and a series of novel observations relating to species delimitation, biology, geography, and taxonomy. Important range extensions are also noted for poorly known species, among them H. ciliolata, Н. flava, Н. quadrongula, and H. tere oe Including the novelties described in this account, 79 species of d are now recogn ized, 4 in tropical Africa, 37 in summer- e southern Africa, mostly of the Drakensberg, and 4 nter-rainfall bes Africa. орати longicollis and Н. coccinea are sha red between tropical and eastern icu: rn | Afric 'a, and H. radiata and H. bachmannii between the winter- and summer-rainfall zones of southern Africa. The new species from the southern African winter-rainfall zone are: H. decipiens, from Namaqualand, i allied to H. radiata; H. glabrescens, from the Rog- geveld Escarpment, closely related to H. pilosa; H. malvina, also related to H. pilosa, from cliffs on the Anysberg in "s Little Karoo; H. rupic ola, a litho ^hyte from western Bushmanland, possibly most closely related to H. acuta; and 1. sufflava, a member of section ii alba from Malmesbury in Western Cape sat 'e. New species from the se African summer- „rainfall zone are: H. altimontana, a к blooming, white-flowered species of the high Drakensberg of Lesotho and KwaZulu- Natal: H. brevistyla, a dwarf plant from Free State and adjacent KwaZulu-Natal; H. debilis, dis and allied to the widespread H. bachmannii, from the Albany aad of Eastern Cape Province; H. ee om Lesotho, which has small, purple flowers; H. saxicola, of rocky outcrops in Mpuma langa, South Africa, hich has жеш һие flowers j with $ short anthers; and H. stenosiphon, a long-tubed, Fd flowered species with blackish ann from Eastern Cape, South Africa. Key words: Africa, Моде Mond. Crocoideae, Hesperantha, Iridaceae, systematics. Despite the publication in the past 20 years of have been recorded in the past 10 years and new revisions and floristic accounts of Hesperantha Ker collections have been made of plants that are here Gawl. covering most of its range across southern recognized as the new species H. decipiens, H. gla- and tropical Africa (Goldblatt, 1984; Hilliard & — brescens, H. malvina, H. rupicola, and H. sufflava. Burtt, 1986; Goldblatt, 1986, 1987, 1993; Gold- Important range extensions or additional popula- blatt & Manning, 1996) and extensive fieldwork by tions in the winter-rainfall zone include those for many botanists, this sub-Saharan African and H. ciliolata, H. flava, H. quadrangula, and H. ter- largely southern African genus of Iridaceae sub- — etifolia, while the rare H. minima, a species first family Crocoideae G. T. Burnett (1835) (syn. Ixioi- collected in 1830, was rediscovered in Namaqua- deae Klatt, 1866, as subordo Ixieae) continues to land in 1991. yield novelties, significant range extensions, and Collecting in the summer-rainfall region of east- new collections that contribute to our understand- ern southern. Africa over the past 15 years has ing of known species. In the southern African win- yielded three new species, Hesperantha exiliflora ter-rainfall zone, several major range extensions from subalpine elevations in Lesotho, H. brevistyla ! I thank Margo Branch, John Manning, and Yevonn Wilson-Ramsay for the illustrations, Clare Archer for her assistance with historical questions, Mervyn Lótter, Cameron McMaster, Ingrid Nanni, Lendon Porter, and Pieter Winter for help i in the field and with the documentation of new taxa, John Manning for critical comments, and Roy Gereau for revising my Latin descriptions. Collecting permits were provided by the Nature Conservation authorities of Mpumalanga, Northern pus and Western Cape Provinces, South Africa. * B. A. Krukoff Curator of African Botany, Missouri Botanical Garden, P.O. Box 299, St. Louis. Missouri 63166- 0299, U.S. N peter.goldblatt@mobot.org. ANN. Missouni Bor. GARD. 90: 390—443. 2003. Volume 90, Number 3 2003 Goldblatt Hesperantha Review from the northern high Drakensberg of Free State and KwaZulu-Natal in South Africa, and H. steno- siphon from Eastern Cape Province, South Africa. Additional records suggest that white-flowered Н. hygrophila, as defined by Hilliard and Burtt, in- cludes three species. Hesperantha hygrophila has distinctive leaves with raised and winged margins and midrib, features that define the species (Hil- liard & Burtt, 1986). A second species has nearly plane, somewhat leathery leaves and a robust habit with an erect, many-flowered spike, while the third, confined to cliffs in the Long Tom Pass area of Mpumalanga, has trailing stems and leaves, and a few-flowered spike. The earliest name for the for- mer is Gladiolus inconspicuus, here transferred to Hesperantha, and the second is a new species, H. saxicola, Late winter- and spring-flowering plants included in H. baurii (which flowers in mid to late summer) by Hilliard and Burtt and corresponding to H. modesta Baker are recognized as a distinct species. In addition, spring-flowering plants with large, white flowers from the northern Drakensberg allied to H. schelpeana are recognized as a new pecies, H. altimontana. First discussed by Hilliard KE Burtt (19 having a white flower with a longer perianth tube, H. altimontana has plane, falcate leaves whereas 86) as differing from H. schelpeana in H. schelpeana has slender, terete leaves. The enig- matic high-altitude Drakensberg endemic, H. pub- inervia, originally known from fragmentary materi- al, has been rediscovered, and the range of H. grandiflora extended northward to The Sentinel in Free State, South Africa. Study of living plants in the southern Drakensberg has shown that a second species was included in H. grandiflora by Hilliard and Burtt (1986). These plants match the type of H. galpinii (Foster, 1948), currently a synonym of H. grandiflora, and indicate the need for an ex- panded definition of H. woodii, of which Н. galpinii must be a synonym. Lastly, reexamination of the type collection of H. candida makes this species an earlier name for H. vernalis Hilliard & Burtt. The species called H. candida by Hilliard and Burt as- sumes its earlier name, H. leucantha. Additional species described for the southern African winter-rainfall zone (Goldblatt, 1987) and here render my 1984 account of Hesperantha out of date and the keys valueless. Likewise, for eastern southern Africa the only post-Flora Capensis (Bak- er, 1896) treatment of the genus by Hilliard and Burtt (1986) dealt only with the species of the waZulu-Natal—Lesotho region and now requires expansion even there. Keys for the entire genus are provided here, one for the southern African winter- rainfall zone, and the other for tropical and eastern southern Africa combined. This account includes expanded descriptions and geographic information for H. pubinervia, H. woodii, and the incompletely understood species of Limpopo and Mpumalanga Provinces, H. schlechteri and H. brevicaulis. This last species is the only long-tubed Hesperantha with pink flowers from the northern provinces of South rica. The difficulty in distinguishing herbarium spec- imens of several closely allied species, including. for example, Hesperantha grandiflora from Н. woo- dii, H. hygrophila from H. inconspicua, and H. glar- eosa from H. schlechteri, emphasizes the importance of fieldwork and knowledge of living plants, partic- ularly for a genus like Hesperantha in which the basic floral morphology is highly conserved and useful taxonomic characters include the timing of anthesis, perianth and anther color, and the orien- tation of floral parts. It is likely, too, that critical characters for some species lie in the capsules and seeds, or in the corms. These are, however, seldom collected: fruiting material because it is absent at flowering time, and corms because they are often difficult to dig up (and because collectors are often reluctant to destroy plants). For Hesperantha, col- lectors should try to record time of opening and closing of the flower, presence or absence of scent and scent characteristics), flower color, presence —. of nectar, and any other feature not evident when the plant is pressed. MORPHOLOGY AND DIAGNOSTIC CHARACTERS OF HESPERANTHA Species of Hesperantha are small to medium- sized, deciduous geophytes, and with the exception of the rhizomatous H. coccinea, cormous rootstock (Fig. 1). Although I formerly subdivided — = e genus into four sections (Goldlbatt, 1982, 1984), I now recognize only three sections, section Concentrica (62 species), section Hesperantha (8 species), and section Radiata (9 species), based largely on corm characters. The corm body is asym- metric with a lateral ridge produced from the base from which the roots emerge. The woody corm tu- nics usually reflect the internal asymmetry in sec- tion Concentrica, but in sections Hesperantha and Radiata the corms are more or less symmetric and bell-shaped (but the flat base is often oblique) (Fig. IC-E). Section Radiata is additionally distin- guished by a bract character, the outer bracts unit- ed in the lower half around the spike axis, and flowers with a curved perianth tube. Species of sec- tion Hesperantha also appear to comprise a close- knit assemblage based on their distinctive corm, 392 Annals of the Missouri Botanical Garden Figure 1. Morphology of Hesperantha. —A. Н. acuta (Goldblatt 6373, MO), with small corm with concentric tunics. —B. H. humilis (van Zyl s.n., MO), acaulescent habit with imbricate corm tunics. —C. H. spicata (Goldblatt 5774, MO, NBG), with large, flat-based corm. D-F. Corm detail. —D. H. radiata (Goldblatt 5179A, MO), corm with tunics. —E. H. falcata (Goldblatt s.n., no voucher), intact corm and with the woody tunics removed to show the lateral projection from which roots are produced. —EF. H. fibrosa (Goldblatt 6101, MO), corm with tunics. —G. Flower of H. pilosa (Goldblatt 5810, MO) with tube opened vertically to show filament insertion and separated gynoecium with inferior ovary, slender style as long as the perianth tube and long style branches. Drawn by Margo Branch from live plants. Seale bar 1 em; corms D and E much enlarged. Volume 90, Number 3 2003 Goldblatt 393 Hesperantha Review and I continue to recognize this section, restricted to the southern African winter-rainfall zone. Comparison with the related genera Romu- lea Maratti and Geissorhiza Ker Gawl. (which also have asymmetric corms with woody tunics) suggests that the asymmetric corm with concentric tunics (Fig. 1A, F) that fragment into vertical segments, characteristic for Hesperantha sect. Concentrica, is ancestral. In modification of this type of corm, the tunics split mainly from the base and as a result rei older tunics partly overlap the newer ones (Fig. 1B). This corm defined a second section /mbricata (Goldblatt, 1982). Sections Hesperantha and Radia- ta both have bell-shaped corms with an oblique to horizontal base (Fig. 1C—E). In section Radiata the tunics mostly have concave, somewhat scalloped segments. The classification has not proved entirely workable, particularly for the summer-rainfall zone, where corm tunics seldom accumulate over several seasons. Most species there have corm tunics of the concentric type that taper above into prominent, fairly stiff points, but sometimes the accumulated tunics take on the appearance of the imbricate type. Most of the species of the summer-rainfall zone are so morphologically similar to one another in other ways that sectional separation on the basis of minor corm differences does not seem warranted. The dis- tinction between sections Concentrica and Imbri- cata thus no longer seems useful, and they have been united under the first name. No infrasectional groups are recognized in this large section of 62 species. Leaves of Hesperantha species are generally plane (Fig. 1А—С) and reflect few taxonomically significant specializations. The leaves of H. spicata often have undulate or crisped margins (Fig. 1C). The midribs are usually slightly thickened. and the margins are frequently also slightly raised. Hesper- antha juncifolia and Н. teretifolia have centric leaves, round in transverse section, and in the latter the surface is vertically ribbed with the rib edges microscopically ciliate. A few species have pilose (H. pilosa, H. pseudopilosa, H. pubinervia, H. gla- rescens) or minutely ciliate leaves (H. ciliolata, H. teretifolia). Leaf number is often constant in a spe- cies and is a useful defining character. Leaf number ranges from several and indeterminate in number in a species to consistently four, or three, often with the lower three or two basal and the remaining one cauline and largely sheathing. А minute scale-like leaf, borne on the stem shortly below the spike, characterizes several species allied to H. pilosa. Flowers are borne on aerial or largely subterra- nean flowering stems that are usually unbranched, and as in most Crocoideae, are arranged in spikes (Figs. 1A-C, 2, 4, 5). Flowering phenology is con- stant in a species and, except for minor shifts due to seasonal variation in temperature, timing of crit- ical rainfall, or elevation, flowering occurs at the same time each year. The bracts are green and sim- ilar in texture to the leaves, or tend to become dry above. The inner bracts have two main veins, a bifid apex, and often have membranous margins. The perianth always has a well-developed perianth tube, typically ranging in length from about half as long as the tepals to elongate and up to three times as long. In Hesperantha quadrangula, however, the perianth tube is ca. 3 mm long, and about one-third to one-quarter as long as the tepals. The perianth tube is straight in most species of sections Concen- trica and Hesperantha except in H. bachmannii, H. bulbifera, and Н. grandiflora where the tube is curved outward near the apex as it is in section Radiata. The tepals typically spread at right angles to the tube, the flower thus being rotate to hypocrateriform (Fig. ТАС). The tube is narrow with a short, ex- panded upper portion at the base of which the fil- aments are inserted (Fig. 1G). The filaments are filiform and erect, and bear linear, longitudinally dehiscent anthers that are twisted at the top of the filaments and face inward. In several species the anthers are articulated on the filaments and lie hor- izontally. Characteristic of the genus is the style, which divides shortly below the top of the perianth tube into three long, diverging to laxly spreading branches (Fig. 1A, B, С), stigmatic in the upper half or for almost their entire length. T of Western Cape Province, South Africa, Hesper- antha cedarmontana, Н. elsiae, and Н. saldanhae, are unusual in having the style branches and, ex- cept in H. saldanhae, the stamens included in the ree species perianth tube. The eastern southern African H. grandiflora has a zygomorphic flower with the tube curved at the apex, the tepals oriented vertically, and the stamens and style branches unilateral and declinate. In other species with a curved tube, the stamens lie in a drooping, more or less pendent cluster. Flower color is fairly conservative, and many species of the southern African winter-rainfall zone have white or cream flowers while most eastern southern African species have pink flowers. A few predominantly white-flowered species have popu- lations with a yellow perianth (Hesperantha acuta, H. falcata), and some populations of H. pilosa have white, blue, or magenta flowers. Flower color is of- ten associated with times of opening and closing of the perianth. In general, colored flowers open dur- ing the day (usually only in the morning or after- 394 Annals of the Missouri Botanical Garden noon) and white flowers open in the afternoon or evening and close during the night. White peri- anths that open in the evening appear derived (Reeves et al., 2001a, 2001b). If this is correct, the eastern. southern. African species, most of which have pink to mauve flowers open during the day. constitute the ancestral type. White, crepuscular to nocturnal flowers would then be derived and the few winter-rainfall zone species with pink or yellow flowers then represent a reversal to an ancestral condition. Scent is a common feature of white-flowered spe- cies of the winter-rainfall zone of southern Africa and is otherwise rare, although some populations of the mauve-flowered Hesperantha ciliolata of the winter-rainfall zone have scented flowers, as do the ite- or cream-flowered H. longicollis and H. ra- diata (and perhaps H. ballii) of summer-rainfall southern Africa. Scent is variable and to the human nose either pleasant and sweet to sweet-spicy, о acrid-musty, or bitter. Scent can be inconsistent within species, and may sometimes be absent in some populations of otherwise scented species, while in widespread species like H. falcata scent is variable, ranging from narcissus to frangipani, somewhat musty, or d absent. Scent is often weakly developed at anthesis and the intensity reaches a peak еи ап hour after flowers open in some species; it falls in intensity again before the flower begins to close. Scent is an unreliable char- acter because it varies so much among populations, sometimes even changing under different condi- tions and times of sampling. Capsules, and in particular seeds, vary across the genus, although they are seldom recorded. Cap- sules are usually globose to oblong, but may be cylindric in section Radiata, Seeds are primitively globose (sometimes weakly faceted by pressure in the capsule) and have a flattened chalazal end (Goldblatt & Wagner, 1984). Notable variants in- clude Hesperantha spicata, which has seeds with a loose, white spongy coat, and H. е which is distinguished from H. pilosa partly by it large seeds with a spongy coat (Goldblatt, 1987 \ Seeds of H. coccinea, described in more detail un- der that species in the systematic account, have a loose coat that contains a much smaller globose seed body, and I assume they are both aerodynamic and dispersed by water, as they are unusually buoy- ant. In eastern southern Africa several long-tubed species, including H. grandiflora and Н. huttonii, have winged seeds (Hilliard & Burtt, 1986), which provides support for the monophyly of the long- tubed species with this character. Seeds of H. sco- pulosa, which also has long-tubed flowers, are nar- rowly ovoid-oblong, have a very long persistent, twisted funicle, and have no wing at all. Seeds of Hesperantha species, especially of east- ern southern Africa, are not well known, and every effort should be made by collectors to obtain ripe seeds as well as flowering material. Because cap- sules mature several weeks after flowering in most species, seed collection is seldom possible unless a later visit to the site can be made. Chromosome number is conservative. Nearly all 30 species counted are diploid with an ancestral base number of x = 13 (Goldblatt, 1984, 1987: Goldblatt & Takei, 1997). The tropical African Hes- perantha petitiana is polyploid, with tetraploid or hexaploid populations (Goldblatt, pir and it may have a derived base number of x The only count for Н. baurii (Goldblatt & Takei, pee is also 2n 24. Examination of vouchers for earlier counts ts ( ;oldblatt, 1971) shows that the reports of 2n = 206 for H. baurii and H. longituba are both 13 is uncommon in the Iridaceae but is shared with the for H. brevicaulis. The basic number of x — largely Cape genus Geissorhiza. The base of x = 12 is probably secondary in the genus, and an example of dysploid reduction (Goldblatt, 1985, 1990) and 2001a, 2001b) con- firm the close relationship of these two genera. Both morphology molecular data (Reeves et al., Reeves's analysis of generic relationships using se- quences from four plastid DNA regions of the chlo- roplast genome yields a bootstrap value (BS) of 99% for the clade including the one species of Geissorhiza and two of Hesperantha. One of the two species of Hesperantha in that study was H. cocci- nea, only species of the rhizome-bearing Schizos- tylis Back. & Harv., ушу in Hesperantha by Goldblatt and Manning which was reduced to synon- — 1996). A second molecular study using the plastid DNA exon matK confirms the close relationship be- tween Hesperantha and Geissorhiza (BS 100%) (Goldblatt et al., 2003). Neither the matK, nor the Reeves et al. study have provided any well sup- =) ported indication of the relationship within Crocoi- deae of the Hesperantha—Geissorhiza clade, which is unresolved. The basic floral morphology is so constant in Hesperantha that species identification often rests on vegetative characters, especially the nature of the corm tunics in the southern African winter-rain- fall zone, or on leaf number and color and the rel- ative lengths of the floral parts, especially the peri- anth tube and the stamens. Flowering time is also constant within a species so that this can safely be used as an aid to identification. Volume 90, Number 3 2003 Goldblatt 395 Hesperantha Review THE SHORT-TUBED SPECIES OF EASTERN SOUTHERN AFRICA The short-tubed species of Hesperantha from eastern southern Africa, most of which have pink flowers, are difficult taxonomically and need fur- ther study. Hilliard and Burtt’s (1986) account of the genus for KwaZulu-Natal and adjacent areas constituted a major advance in the understanding of these species. Using this treatment, I have iden- tified three species that appear to be new and de- scribe them below. Hilliard and Burtt did not deal in detail with all the species that occur north of the KwaZulu-Natal area, that is, in Swaziland and what are now Mpumalanga and Limpopo Provinces of South Africa. This left H. brevicaulis (Baker) G. J. Lewis. H. rupestris N. E. Br. ex R. C. Foster, H. schlechteri (Baker) R. C. Foster, H. similis N. E. Br. ex R. C. Foster, and Gladiolus inconspicuus Schlechter not, or incompletely, accounted for and evidently endemic there. All except H. brevicaulis appear to be closely allied to the widespread and common Н. baurii. The fairly robust H. rupestris is distinguished by its white flowers with red on the reverse of the outer tepals, tall stature, and four leaves (Hilliard & Burtt, 1986). Hesperantha schlechteri, based on one ample collection from Limpopo Province, and H. similis appear to rep- resent the same species, which often has branched stems and five leaves. The latter, based on Wilms 1443 from Devil's Mpumalanga), is readily matched by several col- Knuckles (Long Tom Pass in lections from this area between Sabie and Lyden- burg that usually have five leaves, the lower four basal and with firm, narrow blades 1.5—3 mm wide, a flexuose stem usually looped above the sheath of the uppermost leaf, and large pink flow- ers, the outer tepals coppery on the outside. Hes- perantha baurii and Н. glareosa have four (or three) leaves. only two basal. In H. baurii they аге often longer and wider than the leaves of H. schlechteri, while H. glareosa has even narrower leaves. Leaf number and potential for branching are remarkably consistent in most species of Hes- perantha and may be relied upon as useful taxo- nomic characters, and thus H. schlechteri can usu- ally be distinguished by its unusual leaf number and frequent branching. The only other species of Hesperantha from eastern. southern Africa with short-tubed, pink flowers that occasionally have five leaves are H. brevistyla and H. leucantha (H. candida sensu Hil- liard & Burtt), both somewhat different plants with pale pink flowers (I have seen no white-flowered plants in the field or herbarium, although Hilliard and Burtt described the latter as sometimes having white flowers). Hesperantha brevistyla has small flowers, the tepals ca. 7 mm long, and short style branches reaching only to the lower third of the short white anthers, while H. leucantha normally has a relatively long perianth tube, mostly 12-15 mm long. Tepals in this species are 10-17 mm long, and living plants that I have examined have whitish anthers and pollen, the anthers are 5—7 mm long, and the style branches appear excep- tionally long, sometimes exceeding the anthers by 2-3 mm. In contrast, the more common Hesperantha bau- rii has bright yellow anthers and pollen, deep pink to almost magenta tepals, and style branches just barely exceeding the anther apices. Hesperantha leucantha, as understood by Hilliard and Burtt, seems to me too loosely delimited and I have re- defined it, referring plants with particularly small flowers with a tube 3.5-7 mm long and tepals 6- 8 mm long from interior Lesotho to the new H. exiliflora. Plants with an erect stem and large, white flowers often fading pink, from Mpumalanga and interior lowland and coastal KwaZulu-Natal are referred to H. inconspicua, which may distin- guished from H. hygrophila by their more or less plane leaves. Hesperantha hygrophila in contrast has leaves with the midrib and margins raised and the edges winged, thus arching over the laminar surface, and a prominent pair of secondary veins. These features are difficult to see in dry speci- mens, especially in H. inconspicua in which the leathery leaf blade dries to leave the midrib prom- inent and the non-vascular part of the leaf partly collapsed. Populations of another white-flowered species from the Long Tom Pass area of Mpuma- langa that grow on damp cliffs, have trailing leaves and stems. linear leaf blades, and short anthers represent another new species, H. saxicola, a col- lection of which was included in H. leucantha by Hilliard and Burtt. Among the remaining short-tubed specimens 1 have examined, spring-flowering plants with pink flowers from the sandstone belt of coastal KwaZulu- Natal and adjacent Transkei correspond closely with Hesperantha modesta, described by J. G. Bak- er in 1892, and tentatively included in Н. baurü by Hilliard and Burtt. Apart from the difference in flowering time, plants can readily be distinguished from summer-flowering H. baurii by having spikes of only two or three (rarely more) flowers and usu- ally three or sometimes four leaves, the lower two basal and with long blades and the remaining one or two largely to entirely sheathing. 396 Annals of the Missouri Botanical Garden BIOGEOGRAPHIC NOTE Including the novelties described in this ac- count, and Hesperantha coccinea, which has been transferred to the genus from Schizostylis (Goldblatt & Manning, 1996), Hesperantha now includes species, 4 in tropical Africa, 37 in summer-rainfall southern Africa, mostly of the Drakensberg, and 42 in winter-rainfall southern Africa. Most species are fairly narrow endemics, but H. petitiana extends from eastern Zimbabwe to Ethiopia, H. longicollis from the Vaal River, in Gauteng Province, South Africa, to Malawi, and Н. radiata (including Н. ty- soni, which was recognized as a separate species by Hilliard & Burtt, 1986) from Namaqualand in the west across the Western and Eastern Cape as far east as Swaziland. The southwestern and south- ern Cape (from the Bokkeveld Mountains to Port Elizabeth) remains the most species-rich area with 30 species, 18 endemic; the Drakensberg of East- ern Cape, KwaZulu-Natal, and Lesotho has 22 spe- cies, 15 endemic; the western (winter-rainfall) Karoo has 20 species, 8 endemic; and Namaqua- land-Bushmanland has 10 species, 6 endemic. Di- versity decreases northward: southern Africa north of the Vaal River axis has 12 species, 5 endemic, while tropical Africa has 4 species, 2 of which, H. ballii and H. petitiana, are endemic. FLORAL BIOLOGY For a genus of only modest floral variation, Hes- perantha species show considerable diversity in their pollination systems (Manning & Goldblatt, 1996; Goldblatt & Manning, 2000; Goldblatt et al., in press). In general, the short-tubed pink-, lilac-, or blue-flowered species are pollinated by a range of apid bees, mostly Apis mellifera and species of Anthophora (in the southern African winter-rainfall zone) or Amegilla (in the summer-rainfall zone), sometimes in combination with hopliine beetles, e.g., Н. baurii, Н. pauciflora. White-flowered diur- nal flowers are pollinated by the same suite of bees as well as Halictidae, but the white flowers that are crepuscular are pollinated by small night-flying moths, mostly species of Noctuidae and Drepano- gynidae that settle on open flowers, and if the flow- ers open before sunset, also by apid bees. Several long-tubed pink- or purple-flowered species are known to be pollinated by long-proboscid flies. These include the nemestrinids Prosoeca gangl- baueri (Н. grandiflora, H. scopulosa, and Stenobasipteron wiedmannii (H. brevicaulis) in eastern southern Africa, and P. peringueyi and P. sp. (H. latifolia, H. oligantha) in the winter-rainfall zone (Goldblatt & Manning, 2000; Goldblatt et al., ы, woodii in press). Other species with similar flowers must be assumed to have the same reproductive biology. Most likely, some cream- to yellow-flowered, long- tubed species of the winter-rainfall zone (e.g., H. muirii, Н. pallescens) will prove to be pollinated by horseflies (Tabanidae) of the genus Philoliche. The common red-flowered form of the long-tubed H. coccinea is pollinated by a guild of large butterflies of the families Papilionidae (Papilio spp.) and Sa- tyridae (Aeropetes tulbaghia). The less common pink-flowered form is presumably also pollinated by the long-proboscid fly, Prosoeca ganglbaurii. Lastly, the yellow-flowered H. vaginata is pollinat- ed by hopliine scarab beetles, and yellow-flowered populations of H. falcata and H. pauciflora are pol- linated by these beetles in combination with apid Ф Hesperantha Ker Gawl., Ann. Bot. (Kónig & s 1: 225. 1804. TYPE: Hesperantha falcata (L. f.) Ker Gawl. Most recent revisionary accounts: Goldblatt, J. S. African Bot. 50: 123. 1984; ape ums Notes Roy. Bot. Gard. Edinburgh 43: 436. 1986. Schizostylis Backh. & Harv., Curtis's Bot. Mag. 90: pl. 5422. 1864. TYPE: Schizostylis coccinea Backh. & Harv. In the account of the species that follows the keys, treatment of species is deliberately inconsis- tent. As this is a synoptic review, when no new information about species has come to hand I have provided only its name, primary nomenclature, and a reference to the most recent revisionary account, Goldblatt (1984, 1987) for species of the southern frican winter rainfall zone, Hilliard and Burtt (1986) for species of KwaZulu-Natal and adjacent areas. When new biological, geographic, or taxo- nomic information is available, this is presented in narrative. form. Only significant range extensions are listed in full at the end of the species entry; new collections made since the appropriate revision was published are only listed if they expand the range of the species. Full descriptions are provided for new species and for those that were incomplete- ly known in the past or are resurrected from syn- onymy. Exsiccatae are cited below following the quarter- degree square system in use in southern Africa as outlined by Edwards and Leistner (1971). All of these specimens have been examined unless stated to the contrary. It serves no useful purpose to provide a single key to the entire genus; thus two keys are provided, one for the winter-rainfall zone and one for tropical Volume 90, Number 3 Goldblatt Hesperantha Review and eastern southern Africa. The species are listed in the account that follows in taxonomic order with- in sections in a sequence that reflects as far as possible my current understanding of their relation- ships. Appendix 1 provides a list of the species in alphabetical order, together with their correspond- ing number in this article. KEY l. SOUTHWESTERN AND S تو KEYS TO THE SPECIES OF HESPERANTHA Note: For leaf number include the entirely sheath- ing leaves on the upper part of the stem and ex- amine more than one plant since the character is, to a limited extent, variable. Measure leaf width near the middle of one of the lower leaves. OUTHERN AFRICAN WINTER-RAINFALL ZONE (NAMAQUALAND, THE WESTERN KAROO, AND THE SPECIES OF THE 5 TH) SOUTHERN CAPE AS FAR EAST AS PORT ELIZABE? Flowers with the perianth tube curved at the apex and thus facing to the side or nodding; outer floral bracts with the margins united around the axis in the lower part (sometimes only near the base); corm an da with a flat or oblique base, the tunics Sara forming scalloped, concave segments (sect. Radiata Flowers deep р stamens and s doin thes included in the perianth tube ..................... 78. H. elsiae Flowers white to cream or pale Mua amens and style branches exserted. Bract maris united pes basally: perianth t tube 6-8 mm long а. 71. H. brevifolia 3b. = ct margins united for at least 2 mm. Flowers large with tepals 16—25 mm long and tube 15-25 mm long; perianth pale = with darker veins -2-000000 9. H. muirü a pee гра with tepals 10-17 mm long and tube 5-12 mm long; perianth bis ог 'am. Ба. € terete 2. H. juncifolia ves + plane, often somewhat thicker in the middle and occ asionally Meis cross- تا ت SE‏ Q2 Бы à Та. Bracts usually united around stem for 6-10 mm, i.e., about half their length; spike straight with the bracts parallel to the axis; plants mostly 20—60 cm high 79, with (1—)5 to 15 flowers per spike _. H. radiata 7b. Bracts usually united around stem for about one third their length ‘spike flexuose with the bracts diverging from the axis; plants short, mostly 5-10 ст high, with 1 to 5 flowers per spike. Ва. Corm base jen prominent horizontal spines; bracts united р the axis for up to 3 n Н. marlothii Corm base words spines; bracts united around the axis for s mr 74. H . decipiens 8b. Flowers with the perianth tube straight, rarely curve ds at йе: арех иг рез or half nodding; outer floral bracts with the margins free to the base, corm tunics asymmetric or symmetric and with a round or flat base, = пеуег idus the tunic layers forming scalloped sections (sects. Concentrica hd Hesperantha). s acaulescent or stem extending barely "Ph the in but sheathed by the leaf a 10a. Tae small, tepals 8-10(-12) mm long |... . H. hantamensis 10b. Flowers larger, tepals 13-25 mm long. Perianth Ma elongate, 30-45 mm long; tepals with dark blotches toward the ari — 98. Н. luticola 11а. 11Ь. digi tube less than 30 mm long (mostly 18—28 mm); коні and d within, lowers yellow, closed during the day and opening at sunset 2. H. flava PN Flowers m to red-purple, open Ыш. e day, Sie at night. m with a rounded base; leaves falcate, obtuse or subobtuse, leathery and with ge thickened margins ss 24. H. "te 13b. Corm with a flat base; leaves sword-shaped, acute, fairly soft- — with t je d not or barely thickened |... s 1. H. latifolia 9b. Plants with an aerial ste 4a. Leaves pilose to sc hid ciliate, sometimes visible only microscopically. 15a. Leaves hollow or * solid and with narrow longitudinal grooves; d scabrid ciliate along 16a. Leaves + round in cross section; анн = to cream, nocturnal, opening in the late afternoon and пее during the night LLorem a 13. H. teretifolia 16b. ненә oval in cross section; flowers pink, mauve, or purple, open in the day an sing in the early afternoon 2. H. ciliolata 15b. Leaf blade flat or he ады and midrib somewhat raised; pilose along the sheath and blade margins and v 17а. Leaves sword- irr to oblong; scale-like leaf below the spike mostly 12-20 m 10. H. оа long, often pilose |... ss 17b. Leaves bn to narrowly sword-shaped; scale-like leaf below the spike 3-5(-10) mm long, usually glabrous Annals of the Missouri Botanical Garden 14b. l8a. ae dwarf, to 5 cm high with spikes l-flowered; flowers magenta; leaves y hairy (rarely evidently glabrous) 22005000000. glabrescens 18b. Plants Bou at least 10 cm high, with spikes rarely with less than 2 fewer rs; flowers white, blue, um or mauve; leaf blades and sheath of the third leaf Dd ш densely ha 19a. Lea vin ШШ, conspicuously hairy; leaf tips usually acute od of open sandy slopes and flats) 0 H. pilosa 19b. Leaf blades : sparsely hairy (often visible only under the mic аен 1 tips obtuse (plants of rocky cliffs) с. MEET Ee . H. malvina Leaves and stems smooth. a. Perianth tube curved outward near the Rn flowers half подато. 32. H. bachmannii 20b. Perianth tube straight throughout, flowers facing upward. ¿orms triangular to bell-shaped in cuin. with a horizontal or oblique flat base. :orm base usually with prominent radiating spines or the margins toothed; flow- ers mostly pink to reddish purple, rarely yellow 23a. Perianth tube 6-11 B long |... ERROR ЕЛҮ 66. Н. pauciflora Jb. Perianth tube 15-25 mm long . 67. H. latifolia 22b. Corm base with small teeth or scarcely serrated, but without prominent spines; flowers white, cream, or yellow 24a. S = TI N tyle dividing at or halaw the middle of the perianth tube; style branches and sometimes the anthers partly or completely included in the perianth tube. 25a. Flowers 20-25 mm diam. with tepals 10-12 mm long; stamens and style branches fully included in the perianth tube Manca 65. H. cedarmontana 25b. Гомо са. 18 mm diam. with tepals 9—10 mm long; anthers exs but style branches reaching only to the mouth of the tube кемерлер ныр sy ete - 70. Н. saldanhae 24b. Style dividing just below the оці of the perianth tube; anthers and style branches fully exserted and filaments usually at least partly exserted za the d oe tube. 26a. Flowers small, and secund on a straight spike; perianth tube slightly curved, 4—6 mm long; tepals 4—7 mm eo leaves either plane and sometimes with cris ped margins, or terete to ovoid in cross ripe and hollow; seeds with ind angles айша тш) апа pg co with whitish spongy cells TERN ЕЛЕЕ: 9. H. spicata 26b. Flowers medium to large, not obviously secund, on a ud or flex . (5-)7-14 mm long; tepa traight margins; seeds globose or the sides slightly Hatiened ~ panur s coat dark br TOW 27a ^^ ее) pale ЖЫ. mE | Н. sufflaa 27b. Plants with 2 or more e basal leaves and usually 1 шы ог cauline and largely sheathing; perianth tube 5-9 mm long; te- pals (9—)12—18 mm long, usually longer than the tube; flowers white, cream, or yellow 28a. Bracts green, ond to truncate and often with a reddish margin; leaves usually at least four; flowers usually remote from the leaves, borne on the upper gj of the stem .... жена ELLE 63. pure typical form 28b. Bracts oio or bec oming membranous and above, and then + acute; es often only 3; flow ecd ۰ borne dass to the сое and from about the middle he stem |... H. falcata: pentheri “ trifolia UR 21b. Corms а often * asymmetric, 29: b. with one š aide somewhat flattened. a. Flowers shades of pink, mauve, or = A urple. p eed tube elongate, 20-35 mm long, exceeding the tepa 3 eaves linear, 2-3 mm wide; tube 25-30 mm ns and. шла са. 8-10 mm long aaa . oligantha 31b. ds sword-shaped, 7-10 mm wide; tube ca. 20 mm i long and fil- ents ca. 2 mm long |... sss s vh hs. ). Н. purpurea 30b. Perianth tube 7-12 mm long, usually shorter than to about as long as t tepals. 32a. Basal leaves sword-shaped to linear, 40-100 X 2—4 mm; leaf margins and often the midrib much thickened; stem lacking a sc ae 2 leaf in the upper part of the stem 0 7. H. fibrosa Volume 90, Number 3 2003 Gold Hesperantha Review blatt 399 32b. Basal leaves ovate, 12-20 X 4—8 mm, margins and midrib € thickened; stem with a short, scale-like leaf in the ud Рап of t stem 29b. Flowers lie. cream, or yellow 33a. 33b. “аймыз pw 3-12(-17) mm long, shorter than to rarely moe as m as pals. the te 9. H. um Viene з mostly 16-20 mm long, as long as to slightly rud than pallescens 34a. Plants consistently with two basal нүз = a third entirely sheath- linear 36a. mm long; leaves linear-filiform, less than 1 16 36b. Tepals ca. 10 mm long; leaves linear, 1.4—3.5 mm wi d osing the lower half of the nts tiny, mostly 4—5 cm high; 7 5-10 mm long; leaves em. H. minima ide 4. H. rupicola 35b. moderate in size, usually at least 10-20 cm li tepals m long; leaves linear to lanceolate and 1-7 m 37a. Sheathing leaf inflated and quadrangular in cross Elus 34b. Plants mostly with four leaves, sometimes the uppermost + scale- like. perianth tube 2—3 mm long; anthers 3—4 mm lon 14. H. quadrangula 37b. ж leaf not inflated or noticeably quadrangular іп rz 4.5-6 mm long tion; perianth tube 5-10 mm ane 15 nthers H. flexuosa 38a. Flowers bright yellow. 38b. 39a. Tepals 10-17 mm long; anthers ca. 7 mm long n 3. H. acuta long; anthers 9-15 mm long. = in size, the tepals ca. 20 mm ong, u ormly ye 40b. ellow . H. Боа о 12-18 cm tall, with a well ши тне aerial stem; flowers large with tepals 25—35 cm long, usu- A but not invariably marked with dark chocolate Flowers white, rarely cream. H. vaginata гез elliptic in cross section and hollow, without : thickened midrib . rivulicola 41b. Leaves + parallel- sided ї in cross section, not hollow, and with midrib slightly thickened. 42a. Corm tunics + imbricate or concentric but corms relatively large, mostly 10—14 43a. Perianth 43b. Perianth tube 13-17 mm long mm diam. tube 6-9 mm long 2. H. namaquana 42b. Corm tunics concentric and corms fairly small mostly 3-8 mm diam 4a. Stem with a short, sheathing leaf in the upper half 44b. Stem without a short, sheathing leaf in the up- acuta per half. 45a. Leaves (3—)4 to 5, the lower 3 basal, lin- 45b. rocks) d (3-)4, the lower 2 basal, o obtuse, normally prostrate; perianth (plants of mountain habitats growing 1 Annals of the Missouri Botanical Garden KEY 2. SPECIES OF TROPICAL AND EASTERN SOUTHERN AFRICA (THE SUMMER-RAINFALL PART OF THE SUBCONTINENT WITH A PROLONGED DRY SEASON IN THE WINTER AND SPRING MONTHS— EXTENDING FROM THE EASTERN CAPE PROVINCE, SOUTH AFRICA, EAST OF PORT ELIZABETH TO ETHIOPIA AND CAMEROON) la. — — = Flowers white to cream, ше outer tepals usually brown on the outside; perianth tube curved just below the I 33. H. bulbifera = E e = - = > ® = „ч d =. = = РА -—- = c ~ t = = © ec — 2b. Bract ная united around ded spike axis for up to half their length. anth tube 18-30 mm long; plants flowering in spring (plants of Free State, northern provinces of South Africa, Zimbabwe, Zambia, and Malawi) |... 77. H. longicollis 3b. dig tube 8—15 mm long; plants flowering in the early to mid summer, November to January. 4a. Spike with | or 2 flowers; bract margins united for ca. 3 mm (plants of eastern Zimbabwe) E 76. H. ballii 4b. Spike mostly with 5 to 10 flowers; bract margins united around the axis in the lower half (plants of South pue Lesotho, n Swaziland 5. H. radiata o creamy yellow, sometimes the outer tepals i on the За = Flowers variously pink to mauve, red, or white Е ааа tube mostly straight, w viet curve ius in two species but flowers not nodding. s produced in the spring (dry season) when foliage leaves are absent or partly emergent; leaves one or two, fully пата. dien pese. and produced on separate shoots is asionally old, partly to e а foliage leaves still attached to the flowering stem at flowering time). 6a. Flow a жен close to gro it ren stie above ground by the Seb tube 9-14 mm long; en short, 6-7 mm . H. crocopsis ong 6b. Flowers borne abo ove ground on a short flowering stem, the perianth tube short or loni: tepals Р. Perianth tube 6-10 mm long; tepals white or pale pink inside, the outer tepals slightly to strongly marked purple to brownish on the outside; leaves two or more, + е -------- — 29. H. ahamia 7b. Perianth tube 20-27 mm dione tepals uniformly whitish cream; leaf solitary, falcate, plane 30. H. Mi шшш ob. Flowers produc ed in the spring, summer, or autumn (dry or rainy season) but always bearing fully р E foliage leaves on the flowering ste ianth tube (14—)18—60 mm long, periar ball tube sometimes curved near the apex; seeds (where known) pye with a prominent wing or threadlike appendage at one or both ends. nth tube strongly curved near the apex and tepals vertically oriented; stamens and ie ee hes unilateral and declinate; anthers 8-13 mm long, dark brown асаа 97. H. Иста 9b. Perianth tube straight or икн curved above, the tepals а еца or weakly scending; stamens ascending to erect, and style branches spre g but not unilateral; ie. rs 3-9 mm long, yellow to whitish, or dark brown to black in Ga species. 10a. Leaves and often the stem weak and trailing; plants of cliffs and rock outcrops lla. Plants with cormlets in the lower leaf axils (sometimes lost when handled); leaves 4 to 6 (plants of the Eastern Cape) ss 58. H. huttonii 11b. not bearing cormlets in the lower leaf axils; leaves 4 or sometimes 3. . Flowers mauve pink; perianth tube mostly 18-30 mm long; anthers 6— mm lor күзе of oon dan and Li pud Province) _ 60. H. brevicaulis 12b. ionem gosse pinkish purple; perianth tube 18—42 mm long; anthers 3-6 mm long (plants of KwaZulu-Natal and nie : 13a. Perianth tube 18-25 mm long; anthers 3—4 mm long -.... 61. H. curvula 13b. Perianth tube mostly 30-42 mm long; anthers iA 5—6 mm long 62. Н. scopulosa 10b. Stem erect and leaves usually firm and upright; plants of grassland, open slopes, s and marshes. |a. Filmen 6-12 mm long. Plants of streams, growing in water; rootstock a rhizome or vestigial, but illary cormlets produced in the aerial nodes; flowers pink or red, gees . coccinea spre: 15b. Plants of grassland, roc ocky slopes, or grassy marshes; rootstock a corm with woody Side and stems lacking axillary cormlets; flowers shades of pink to mauve-pink. e 54. H. woodii 14b. Filaments 3-6 mm long 6a. e scabrid-pubescent; stamens erect and style idis hes evidently re maining suberect and noticeably shorter than the anthe . H. uini 16b. жн es glabrous; stamens and style branches asc жеи to АРГЫШ, the style branches as long as or longer than the stamens 17a. Plants e in the spring, SURE to Oc (бег; d white аа, 8. H. неви Volume 90, Number 3 2003 Goldblatt 401 Hesperantha Review 8b. Periant sometimes ridged on the angles but never wing 17b. Plants flowering in the summer and autumn, December to April; flow- nk. ers pin 18a. Perianth tube mostly 45—60 mm long; anthers blackish ........... 55. H. stenosiphon 18b. Perianth tube mostly 14-21 mm long; anthers yellow. 19a. Spike with 5 to 11 flowers; leaves 3-5.5 mm cn peri- anth tube 14—20 mm long 3. H. pulchra 19b. ye with 1 to 3 M leaves 1.5-2 mm on pes ube 23 mm leng ан 59. H. hutchingsiae h tube 3-15(-22) mm long, Vid imis ide (where known) globose or slightly angled, ed. 20a. Mine of southern Africa, dowering i in the spring, August-October, before the main rainy 21 2 ason; flowers pale pink or white la. Perianth tube mostly 14-22 mm long; flowers white, opening in the later afternoon 28. H. longituba and closing after is . Perianth tube mostly 4—12 mm long; flowers white or pink to mauve-pink, the outer tepals often flushed or feathered pink to mauve outside, open during the day, closing = = in the later afternoon. 22a. ier usually 3 (rarely 4), the lower 2 basal and with long, linear blades 2—3 wide, the upper l(or 2) largely or entirely sheathing: spike usually 2—3- flow ered (plants of coastal KwaZulu-Natal and Transkei) -................ 49. H. modesta 22b. Leaves usually 4, 2 basal, l subbasal and partly sheathing and a short, sheathing (plants of interior southern Africa, from southern Mpumalanga to Grahamstown in the south). 23a. Leaves oblong-lanceolate, 2-6 mm wide; spike mostly 2—6-flowered; peri- 26. H. candida anth tube (7—)9—12 mm long 23b. Leaves linear, 1-2 mm wide; spike 1—2-flowered; perianth tube 4-6 mm 27. H. debilis ong 20b. ers of southern or tropical Africa, flowering in the wet season, summer and autumn (in frica mainly December to April, rarely in November); flowers mostly pink to mauve or white to cream. 24a. ME growing in rock outcrops, often on cliffs, stem weak and drooping; leaves + inea 25a. diss pink, mauve, or white; perianth tube 6—8 mm long; anthers 6—7.5 mm 1) 51. H. gracilis long (plants of KwaZulu-Natal) |... 25b. epa white, sometimes flushed pale pink on fading; perianth tube йкы тт g; anthers ca. 4 тт long poe of Mpumalanga) -..................... 6. H. saxicola 24b. "n ы hanging from cliffs, st ct. 26a. Leaf blades with thic kened. margins, secondary veins, and midrib, the marginal and и thic наи flattened rather than rounded and arc ^ over the blade 7. Н. hygrophila . Leaf Bade with or жеен thic ened s margins and midrib, but sec иш veins not or hardly thickened and midrib and marginal thickenings rounded, not ob- viously flattened in outline nor arching over the blade surface ite to cream or palest yellow entirely or pink or red only on the e c 2 owers white reverse of the outer tepals. 28a. big i tepals speckled or uniformly bright red on the reverse. . Ste 12 cm high (taller i gi cultivated plants); tepals long (plants E pes . H. alborosea m usually 7— I mm long; anthers 3—4 mm tal 29b. Sem m 30—45 cm high: tepals 12-15 mm jen anthers m long el of Mpumalanga) -------------------- 48. H. rupestris кр тун faintly pink оп the 28b. Tee uite white to cream, or reverse of the outer pa especially on fading. 30a, Flowers tiny, the tepals ca mm long, perianth tube ca. 4—5 mm long, and anthers ca. 2-3 mm long; leaves soft-texture 50. H. umbricola 30b. Flowers medium to large, the tepals 12-20(-23) mm long, peri- anth tube (5—)6—8.5 mm long, and anthers 5—8 mm long, leaves firm, with thickened margins and midrib. 31b. Tepals uniformly cream to pale yellow, remaining cream on fading, 14—20(-23) mm long; filaments 3—4 mm long; 402 Annals of the Missouri Botanical Garden 27b. anthers mostly golden-brown, dark when dry, 5-7(-8) mm ORE el a 44. H. lactea 31b. Tepas sea white : cream or faintly flushed pink e rse of the outer tepals especially on fading, 12- 16 mm long; ү кен 2.5-3 mm long; anthers cream, not darker on fading, mostly 6.5-8 mm lo ong eee 5. H. inconspicua Flowers pink to mauve niy: 32a. Perianth tube 3-6 mm long. 33a. Leaves 5.5-7 mm wide, falcate; spike with 1 or 2 flowers; tepals 10-12 mm long o 43. Н. ingeliensis 33b. Leaves mostly 0.7-2 mm wide, lin 34a. Stem slender and evidently soft. "with l to 3 flowers widely spaced; tepals 6—8 mm lor Bg нае 40. H. к йо 34b. Stem slender but firm ind wiry, „ mostly with 3 to 8 flowers tepals 10—14 mm long 35a. Leaves usually 5 ا‎ plants depaupe most- ly 2-4 mm wide; tepals 13-16 mm ir rs 4— 4.5 mm long LH es hteri 35b. Leaves us NT 4, mostly 1-2 mm e ep 10— 12 mm long; anthers 4.5-5 mm long — 37. H. glareosa 32b. Perianth tube ari 6-13(-15) mm long. 36a. Leaves 4 or 5, the lower 3 basal, often about as long as the stem or slightly longer; anthers 2.3—5.5 mm long, pollen white: perianth tube 7-12(-15) mm long. Та. Tepals 7-8 mm long; unm tube 7-9 mm long; anthers base and with darker green | longitudinal veins ss - 42. H. brevistyla 37b. Тара 12-16 mm long; perianth tube 5-14 mm long; an- thers 4—6 mm long; style branches 10-12 mm long, e и the anthers in the closed flower; reverse o ilie ou “е, not pale green with dark veins, + uniformly colored ou 38a Perianth ۹ 12-14(-15) mm long; anthers 4.5-5.5 mm long: perianth pale pink; stem and ws weak and trailing J. H. leucantha 38b. P ih tube to 10 mm long; anthers © mm long; pud middle to deep pink: stem к= leaves + erect H. schlechteri 36b. Leaves normally 4, sometimes 3, with the lower 2 Seri usually about half to two-thirds as long as the stem; anthers mostly 3-7 mm long, pollen yellow or white: perianth tube (4.5—)6-12 mm long. 39a. Flowers mauve-pink or almost white; anthers 3—4.8 mm long: capsules usually 8-14 mm long, often exceeding the bracts; seeds usually angular, 1.0-1.2 mm diam. (plants of tropical Africa: Zimbabwe to ne and Cameroon) 36. H. petitiana 39b. Flowers ens bright pink; anthers 5-9 mm n long, pollen yellow; capsules usually 6-9 mm long and enclosed by the bracts; eed usually uniformly globose, 0.9-1.0 mm diam. (plants of eastern southern Africa). 40a. Outer tepals iiid oink to reddish on the outside, ca. 16 X 9 mm; leaves often 3, the lower 2 basal, leathery, the midrib and margins not raised cele мү fleshy surface os alive); spike mostly : o 4 flowers 35. Н. baurii subsp formosa 40b. Outer tepals n not or barely darker than the ou ca. 12-15 X 5-6.5 mm; leaves 4, the lower 3 ba sal or subbasal; firm to leathery, the adah i margins often slightly thickened n aised i е e the sur- face; spike mostly with 5 t E flowers --------------- 5. H. bauri subsp. baurii а. EE N т. = >= Ф E Volume 90, Number 3 2003 Golablatt Hesperantha Review 403 I. Hesperantha sect. Concentrica Goldblatt, nn. Missouri Bot. Gard. : 949. TYPE: Hesperantha pilosa (L. f.) Ker Gawl. He кас” аў еа Goldblatt, Ann. Missouri Bot Gard. 2. Syn. nov. TYPE: Керем кы jd Plants with small to large asymmetrical corms, + rounded at the base, usually with a flattened lateral ridge; older corm tunics splitting from the base into vertical segments. Spike with floral outer bract margins free to the base. Flowers variously colored, usually with a straight perianth tube (curved near the apex in H. bachmannii, H. bulbi- fera, and H. grandiflora). Species 1—62. Occurring across the entire range of the species from Western Cape Province, South Africa, to Ethiopia. 1. Hesperantha erecta (Baker) Benth. ex Baker, Handbk. Irideae 150. 1892. Geissorhiza erecta Baker, J. Bot. 14: 238. 1876. TYPE: South Af- rica. Western Cape: Olifants River, July-Aug. 1830, J. F. Drége 8468 (lectotype, designated by Foster (1948: 11), K!; isotypes, K!, 11. P!). Last revisionary account: Goldblatt, J. 5. African Bot. 50: 45. 1984 2. Hesperantha namaquana Goldblatt, J. S. Af- rican Bot. 50: 47. 1984. TYPE: South Africa. Northern Cape: Bitterfontein, Kareebergen, 24 July 1896, R. angen 8304/5 (holotype, K!: isotypes, B!, B PRE!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 47. 1984 Restricted to southern Namaqualand in Western Cape Province, South Africa, Hesperantha nama- quana was re-collected in 1999, along the banks of a seasonal stream northeast of Bitterfontein. This seems to be the habitat for the species, and until now has not been reported. This specialized habi- tat, in a largely arid landscape, may explain why it is so seldom seen. Additional specimens. SOUTH AFRICA. Western Cape: 30.18 (Kamiesberg) NE of WE E road to Kliprand, along seasonal stream (CD), 7 Gold- blatt & Manning 11681 (MO, NBG). 3. Hesperantha acuta (Licht. ex Roem. & Schult.) Ker Gawl., Gen. Irid. 91. 1827. Ixia acuta Licht. ex Roem. & Schult., Syst. Veg. 1: 383. 1817. TYPE: South Africa. Northern Cape: foot of the Roggeveld Mts., Aug. 1805. M. Н. C. Lichtenstein s.n. (holotype, B!). Figure 1А. Last revisionary account: Goldblatt, J. S. African Bot. 50: 48. 1984. As circumscribed by Goldblatt (1984), Hesper- antha acuta included both white- and yellow-flow- ered populations, the latter referred to H. tugwelliae by R. C. Foster. No information has become avail- able to suggest this treatment was incorrect. Both H. acuta and H. falcata have populations of plants with either white or yellow flowers. In H. falcata the unscented yellow-flowered plants open during the day and close at night, the opposite phenology to the sweetly fragrant white-flowered plants. This is not the case in yellow-flowered populations of H. acuta from the Swartberg Mountains and Prince AI- bert (Goldblatt & Porter 11859, 9 Sep. 2001, MO, NBG: Goldblatt & Porter 12191, 10 Sep. 2002, MO, NBG). Flowers of these plants opened at sun- set when they produced a sweet scent, and closed again at sunrise, the same pattern as in white-flow- ered plants. Corms of H. acuta seem unusually var- ied and in some populations conform closely to the section Concentrica type with a rounded base and oblique lateral ridge, while others have an oblique but flat base, approaching the corms of section Hes- perantha, which are bell-shaped with a broad flat base. Too few collections of the species have corms for me to detect a pattern associated with either geography or some other factor. This calls for future investigation. 4. He -sperantha rupicola Goldblatt, sp. nov. T South Africa. Northern Cape: Bush- ен Farm Naab, E of Springbok, S-facing quartzite rocks, 14 Aug. 2000, P. Desmet 3009 (holotype, NBG!). Plantae 3-5 cm altae eramosae, cormo globoso, foliis З omnibus basalibus, laminis plus minusve linear 3.5 mm latis marginibus costaque vel 2-3-)flora, floribus albis telis exterioribus extus lev- iter malvinotinctis, m longo. x 3 i — tepalis ca. antheris ca. 5 mm longis, styli ramis ca. 3.5 mm longis. Plants 3-5 em high, unbranched. Corm globose with an obliquely flattened side, 9-11 mm diam., tunics dark brown, the layers + imbricate. Leaves 3, all basal, spreading or drooping, as long to twice as long as the stem and up to 9 cm long, the blades + linear to falcate, 1.4—3.5 mm wide, soft-textured, the midrib and margins barely thickened. Stem erect, unbranched. Spike 1-, occasionally 2- or 3- flowered; bracts 9-10 mm long, soft-textured, pale green or becoming membranous above, the inner slightly shorter than the outer, membranous with 2 green keels. Flowers white, the outer tepals slightly flushed mauve on the outside; perianth tube funnel- 404 Annals of the Missouri Botanical Garden shaped, ca. 10 mm long; tepals spreading, ovate, ca. 10 X 3-4 mm, obtuse. Filaments ascending, ca. 4 mm long, inserted at the mouth of the tube; an- thers ca. 5 mm long, shortly tailed, pale yellow, pollen whitish. Ovary ovoid, ca. 3 mm long; style dividing near the mouth of the tube, the branches spreading over the tepals, reaching nearly to the anther tips in the closed flower. Capsules and seeds unknown. Flowering. August. Distribution. South Africa, Northern Cape, i shade on south-facing slopes on cliffs and among boulders, in quartzite or granite rocks, interior Na- maqualand and western Bushmanland. rcge e ola was evidently first collect- ed in 1977, . H. Oliver, H. Tólken, and F. Venter, on ен Mountain near Soutkloof, in Bushmanland in central Northern Cape Province. It was subsequently brought to my attention in 1999 by Phi | versity of Cape Town's Institute for Plant Conser- vation Expedition to Namaqualand. Only a few plants have ever been found in flower, and corms can only be extracted with difficulty from the rock crevices in which they usually grow. Philip Desmet nilip Desmet and members of the Uni- reports that the species occurs on several more hills in Bushmanland. In 2001 a small population was discovered southeast of Kliprand, extending the to the south. Rather than wait until more adequate material can range of the species some 150 km be obtained, I have decided to describe the species here, hoping that by publishing the incomplete de- scription others may be encouraged to seek addi- tional plants, including capsules and seeds ‘he material available suggests that Hesperantha rupicola may be distinguished by the short stature, - to 3-flowered spike, moderate-sized white flow- ers, and soft-textured, spreading to drooping leaves. The corms are relatively large, have an oblique flat side, and more or less concentric tunics. Similar corms are found in H. acuta, and this suggests a relationship with this interior southern Cape and Karoo species. The similarity in the size of the flow- ers to those of the Namaqualand species, H. flex- uosa, may be convergent for the latter species has small corms with a rounded base and spikes of sev- eral flowers. Paratypes. SOUTH AFRICA. Northern Cape: 29.18 (Aggenys) Aggenys Mountain, Soutkloof WSW of Aggenys farm, shady e н (BB), 24 Aug. 1977, Oliver, Tolken & Venter AU Жш; enys, aoe. 5- facing cliffs on Vill, irs Э, Desmet et al. in IPC Expe- dition 364 3018 Genel yon) 13 bs toward Loeriesfontein bon Kliprand—Vanrhynsdorp road, low 2 Eg Morphology of H. oligantha. Scale bar 1 cm. Figure 2. Single flower much enlarged. Drawn by Margo Branch from live plants (Thomas & van Jaarsveld 8967, NBG). granite hill (DB), 7 Aug. 2001, Goldblatt & Manning 11690 (MO, NBG) 5. Hesperantha oligantha (Diels) Goldblatt, J. S African Bot. 50: 87. 1984. Lapeirousia obtu tha Diels, Bot. Jahrb. Syst. 44: 117. 1910 TYPE: South Africa. Northern Cape: Hanan Mts., Oct. 1900, F L. E. Diels 725 (holotype, B!; isotype, MO!). Figure 2. Last revisionary account: Goldblatt, J. S. African Bot. 50: 87. 1984. Known only from the type collection when the genus was revised for the southern African winter- rainfall zone (Goldblatt, 1984), Hesperantha oligan- tha was rediscovered in 1986 on the summit pla- teau of the Hantamsberg above Calvinia in the western Karoo (Thomas & van Jaarsveld 8967 NBG), and was then re-collected in 1994. (Goldblatt & Manning 10043, MO) in the course of research on the pollination biology of long-tubed, purple- flowered lridaceae pollinated by long-proboscid flies (Goldblatt et al., 1995; Manning & Goldblatt, 1996). These new collections confirm the narrow range and unusual morphology of H. oligantha (Fig. 2) and show that it is a species of seasonal streams, seeps, and shallow pools. The corms of the new collections cast doubt on the relationships of H. oligantha for they appear to have concentric rather Volume 90, Number 3 2003 Goldblatt 405 Hesperantha Review than imbricate tunics (Goldblatt, 1984), although this is by no means certain owing to the limited material available. Hesperantha oligantha is, how- ever, now placed among species with similar con- centric corm tunics in the taxonomic sequence. 6. Hesperantha montigena Goldblatt, J. 5. Af- rican Bot. 50: 51. 1984. TYPE: South Africa. Western Cape: Hex River Mts., 11 Oct. 1980, E. E. Esterhuysen 35528 (holotype, BOL!; iso- types, K!, MO! Last revisionary account: Goldblatt, J. S. African Bot. 50: 51. 1984 7. Hesperantha و‎ Goldblatt, J. S. Afri- can Bot. 50: 60. 1984. TYPE: South Africa. Northern Cape: Calvinia district, stream beds below the Hantamsberg, 16 Sep. 1980, Р. Goldblatt 5807 (holotype, MO!; isotypes, B!, BOL!, E!, K!, NBG!, PRE!, S!, US!, WAG!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 60. 1984 When described by me in 1984, Hesperantha ri- vulicola was known from two records, one from the slopes of the Hantamsberg at Calvinia and the other from *van Wyk's farm" near Nieuwoudtville, an un- certain locality, there being several farms in the area with owners of that name. Hesperantha rivuli- cola has since been found in streams on the farms Oorlogskloof and Matjiesfontein (Goldblatt & Man- ning 9466), both located south of Nieuwoudtville, the latter owned by the van Wyk family. The flowers are pollinated by large anthophorine bees when they open after 16:00H and after dark by a range of moths that settle on the flowers (Goldblatt et al., in press). 8. Hesperantha malvina Goldblatt, sp. nov. TYPE: South Africa. Western Cape: Anysberg, farm Tapfontein, W of Matjiesgoedkop, ledges on cliff face, 3 Sep. 1987, J. W. Lloyd 1120 (holotype, NBG!). pu 10-27 cm e eramosae, cormo ovoideo ca. 6 iam., tunici osis concentricia, a * sparse villoris vel нечен inferio summo omnino ея таг- alibus is ca. 3 mm sir antheris 5.5—6.0 mm longis, styli ramis ca. 7 mm lon Plants 10-27 cm high, stem unbranched. Corm ovoid, ca. 6 mm diam., tunics woody, concentric. Leaves 4, the lower 2 basal, linear-oblong, the third largely sheathing and with a short free apex, the uppermost entirely sheathing, partly membranous, 8-22 mm long, inserted a short distance below the spike, blades ca. 2 mm wide, the margins and mid- rib slightly thickened and sparsely long-hairy in the lower half. Spike 1- to 3-flowered; bracts ca. 10—14 mm long, green or flushed purplish, the margins and apex membranous. Flowers mauve, open during the day; perianth tube cylindric below, expanded near the apex, 8-9 mm long; tepals subequal, el- liptic, 13-14 X 4—5 mm, spreading at right angles to the tube. Filaments ca. 3 mm long; anthers 5.5-6 mm long, yellow, pollen yellow. Ovary ovoid, ca. 3 mm long; style branches ca. 7 mm long, reach- ing the upper third of the anthers in bud, spreading when mature. Capsules and seeds unknown. Flowering. Late September to early October. Distribution. South Africa, Western Cape, Little Karoo, Anysberg, on damp, south-facing sandstone cliffs and rocks. Assigned to Hesperantha pilosa when first named in the herbarium, owing to the sparsely hairy leaves, H. malvina actually bears little further re- semblance to that species. The flowers are in gen- eral larger, with the broadly elliptic tepals up to 14 mm long and 4—5 mm wide, while the leaves are linear-oblong and somewhat obtuse at the tips. In H. pilosa the tepals are seldom as long as 14 mm, are rarely more than 3 mm wide, and the leaves are usually linear or linear-lanceolate and acute. The habitat, rocky sandstone cliffs, is quite differ- ent from that of Н. pilosa, which grows on granitic or sandy flats or lower slopes. Hesperantha malvina is evidently restricted to the Anysberg in the west- ern Little Karoo of Western Cape Province. To date it is known only from the holotype, and must be assumed to be rare and very local in distribution. 9. Hesperantha pilosa (L. f.) Ker Gawl., Ann. Bot. (Kónig & Sims) 1: 225. 1804. Ixia pilosa L. f., Suppl. Pl. 92. 1782. TYPE: South Africa. Western Cape: hills around Cape Town, Sep., C. P. Thunberg s.n. (lectotype, designated by Goldblatt (1984: 54), Herb. Thunberg 979, PS!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 54. 1984. Since the treatment of Hesperantha pilosa in my 1984 account of the genus, subspecies latifolia has been raised to species rank as H. pseudopilosa (Goldblatt, 1987). Nevertheless, this relatively com- mon Western Cape and western Karoo species re- mains a fairly variable series of populations in both vegetative and floral morphology. Plants have small 406 Annals of the Missouri Botanical Garden corms with concentric tunics, three basal leaves, the uppermost largely sheathing and enveloping the stem for over half its length. The stem may be smooth or sparsely villous and always bears a bract- like scale a short distance below the spike. The leaf blades are linear or narrowly sword-shaped with slightly thickened margins and midrib, and the margins, midrib, and a pair of secondary veins bear long, usually spreading hairs. The spike is straight to slightly flexuose, and mature, healthy plants bear several flowers, either in a loose spiral or in secund arrangement. The flowers are unremarkable in the genus and are usually white, occasionally blue, mauve, or pur- ple, with the outer tepals flushed dull red to brown on the outside, but white-flowered plants have the anthers held more or less horizontally and parallel to the spreading tepals and short style branches, seldom reaching to the middle of the anthers in the closed flower (when the stamens and style branches are aligned). Most populations in the Bokkeveld Mountains and in the Roggeveld have blue to mauve flowers, but otherwise seem identical except that the anthers are ascending to suberect and the style branches are sometimes longer, often reaching to the anther apices in bud. White flowers typically open shortly before sunset and close again during the night, whereas blue or mauve flowers open in the morning and close in the early afternoon. It would be preferable to recognize the populations of plants with colored flowers as a separate species (for which the name H. puberula is available). How- ever, there is doubt as to whether the populations with colored flowers represent a single clade or multiple origins of a colored perianth. Moreover, there is little apart from perianth color to reliably distinguish these plants. A remarkable collection of what appears to be Hesperantha pilosa from the Hex River Mountains (Esterhuysen 36196A, ber, consists of plants with only two foliage leaves. BOL), flowering in Decem- The lower leaf is linear-lorate and 4 mm wide, with an obtuse apex, and the second leaf sheaths the lower two-thirds of the stem. The margins, midrib, and secondary veins all bear the long hairs that are diagnostic for H. pilosa, but these seem unusual in being slightly reflexed. The plants are taller than usual for H. pilosa, and the spikes bear only one or two flowers. The style branches appear longer than those of any other white-flowered H. pilosa and in the closed flower reach the top of the anthers. The habitat, cool shaded slopes below a waterfall, is unusual for H. pilosa, which normally grows in light bush on exposed slopes and flowers in Sep- tember and October. It seems unlikely that the hab- itat is responsible for the unusual morphology and late flowering time of the Esterhuysen collection. These plants more likely represent a distinct spe- cies. Additional material is needed before an in- formed decision can be made about these plants. 10. Hesperantha pseudopilosa Goldblatt, S. Af- rican J. Bot. 53: 461. Hesperantha ii dani eplaced name: latifolia Goldblatt, rican . 50: 1984. Non Hesper- unn latifolia pe de Vos. 1974. TYPE: South Africa. Northern Cape: Roggeveld, slopes of Sneeukrans, 22 Sep. 1981, P. Gold- blatt 6339 (holotype, MO!; GI). isotypes, K!, When first raised from subspecies rank as sub- species latifolia of Hesperantha pilosa in 1987, H. pseudopilosa was known from isolated sites on the okkeveld Plateau (west of Nieuwoudtville), Sneeu- krans on the Roggeveld Escarpment, and the north- ern foothills of the Kleinswartberg (Goldblatt, 1987). Additional populations (see specimens listed below) have since been found at several more sites on the Bokkeveld Plateau and along the Roggeveld Escarpment, as well as north of Matjiesfontein in the southern Karoo, which substantially fills in what seemed to be an erratic distribution pattern. The species now appears to have a nearly continuous range along the interior edge of the southern Afri- can winter-rainfall zone from the Bokkeveld Es- carpment in the north to the northern slopes of the lein Swartberg The mostly after 1 white Rm open relatively late in the day, 18:00H ent. They are visited by a range of settling moths, mostly of the family Noctuidae (Goldblatt et . and then produce a strong sweet sc al., in press). Additional specimens. SOUTH AFRICA. Northern Cape: 31.19 (Calvinia) Nieuwoudtville trekpath (AC), 27 Aug. 1999, Goldblatt 11108 (MO); Wild- flower Reserve, 25 July 1983, Perr y & br (NBG). 3120 (Williston) farm Knec кай W of Mk pos (CC), Oliver 8917 (NBG). 32.20 (Sutherland) S of Sutherland, farm Verlatekloof (DA), 26 Aug. 1986, Cloete & Haselau 47 (NBG); 2 km N of the top of Komsberg Pass (DB), 31 Aug. 1993, Goldblatt & Manning 9677 (MO, NBG). Western Cape: 33.20 елин. hills between Verlatekloof апа Matjesfontein (DC), 3 Aug. 1998, Gold- blatt & d ded 10951 (MO). 33.20 (Montagu) hill ca. 2 km W of Tweedside Station, 1200 m (AB), 12 Aug. 1988, Vlok 1988 (M( » 11. Hesperantha glabrescens Goldblatt, sp. nov. TYPE: South Africa. Northern Cape: farm Hottentotskloof, ca. 15 km SW of Sutherland on Bo- Visrivier road, 2 Oct. 1999, P. Goldblatt & I. Nünni 11190 (holotype, NBG!; isotypes, K!, MO!). Volume 90, Number 3 Goldblatt 407 Hesperantha Review Plantae 3—5 cm altae eramosae, cormo ov /oideo 4—6 m 3. infe ribus biflora), flore magenteo, tubo perianthii ca. tepalis ca. 8 X 2 mm, filamentis ca. 2 mm lot ca. 3 mm longis, styli ramis ca. 3.5 mm уне autels, Plants 3-5 ст high, stem unbranched, bearing a bract-like scale a short distance below the spike. Corm ovoid, 4—6 mm diam., tunics woody, concen- tric. Leaves 3, the lower 2 falcate, the upper sub- erect, largely sheathing, the blades 1—1.3 mm wide, the margins and midrib slightly thickened, the mar- gins and veins sparsely hairy. Spike l(or 2)-flow- ered; bracts ca. 6 mm long, purplish, becoming dry and brown above. Flowers magenta, open in the morning and closing in early afternoon; perianth tube cylindric, expanded near the apex, ca. 3.5 mm long; tepals subequal, elliptic, 8 X 2 mm, spreading at right angles to the tube. Filaments ca. 2 mm long; anthers 3 mm long, yellow, pollen yellow. Ovary ovoid, ca. 2 mm long; style branches ca. 3.5 mm long, purple, spreading. Capsules and seeds un- known. Flowering. Late September and early October. Distribution. South Africa, Northern Cape, Roggeveld Escarpment southwest of Sutherland, on moist clay flats along watercourses. Closely related to, and initially appearing to be a depauperate form of, the widespread Western Cape Hesperantha pilosa, H. glabrescens was first collected in 1998 by the Cape Town botanist Nick Helme. It can be dini hod from its ally in sev- eral features. The leaves are sparsely hairy, unlike the fairly densely hairy or even downy leaves of H. pilosa, and the spikes are mostly l-, or rarely 2- flowered, whereas H. pilosa typically has several, sometimes up to 10 flowers per spike. More unusual is the small size of the plants, only 3—5 cm high, shorter than any И. pilosa. The flowers are magenta in color, and quite small, with a perianth tube ca. 3.5 mm long and tepals about 8 mm long. Flowers of H. pilosa are often larger, usually white, although sometimes blue or magenta, have a perianth tube 6-10 mm long, and tepals (8—)1 2-15 mm long. The decision to recognize H. glabrescens was made in part because of the consistent morphology and be- cause plants of what I consider true H. pilosa grew nearby (Goldblatt & Nanni 11191), always in the shade of shrubs, and always taller with the spikes up to 12 ст high, and with 2 to 4 flowers. These plants had plane, densely pilose leaf blades, a stark contrast to those of H. glabrescens. The flowers of the two species were virtually identical in size and pigmentation at the type locality. Paratypes. SOUTH AFRICA. Northern Cape: 32.20 Sutherland) 15 km SW of Sutherland on Bo-Visrivier road (BC), 25 Sep. 1998, N. Helme 1554A (NBG). — es 12. Hesperantha ciliolata Goldblatt, J. S. Afri- can Bot. 50: 984. South Africa. Northern Сере: Жош Escarpment, farm Geelhoek, 21 Sep. 1953, J. P. H. Acocks 17176 holotype, PRE!). —. Last revisionary account: Goldblatt, J. S. African Bot. 50: 59. 1984 Described in 1984 from a single collection from the farm Geelhoek in the center of the Roggeveld Escarpment, Hesperantha ciliolata is distinguished mainly by its straight, ribbed leaves, oval to terete in transverse section, with fine scabrid cilia on the rib edges. The type population had violet flowers according to the collection information. А second population was found in about 1995, some 50 km to the north, on the farm is d west of Middelpos, and these plants, which flowered in cultivation in 1998, have mauve-pink flowers that are open in the morning and close in the early afternoon. The flow- ers of the Botuin collection have an unusual acrid- sweel odor, characteristic. of the orchid genera Corycium Sw. and Pterygodium Sw. The scent is so unusual and distinctive that I wonder whether it may attract the same bees that pollinate those or- chids, species of Rediviva (Melittidae), females of which visit the orchids to collect floral oils for nest provisioning (Steiner, 1989). A third collection from south of Sutherland (Goldblatt & Nénni 11189) consists of plants with pale blue-mauve to light pur- ple flowers, also opening in the morning and closing at ca. 14:00H. The flowers of this population are evidently unscented. Plants from the Voetpadsberg east of Touwsrivier (Goldblatt & Ndnni 11200, in fruit; Goldblatt & Porter 11877, in flower) have the same overall ap- pearance as Hesperantha ciliolata except that the leaves mostly have only two grooves, or sometimes a pair of secondary ones strongly raised, but the edges of the raised parts are ciliate-scabrid, as are the edges of the ribs of the sheathing, upper leaf. The flowers closely resemble those of Goldblatt & Nanni 11189 in their bluish color but have a light, musky acrid scent. These plants must be regarded as a minor variant of this otherwise Roggeveld Es- carpment species and represent a new record for the Cape Floral Region. The seeds of the Voet- padsberg plants are wedge-shaped, thus strongly angular with flat faces, the edges of which are slightly winged. The seeds of Roggeveld popula- tions of H. ciliolata are unknown. 408 Annals of the Missouri Botanical Garden € specimens. SOUTH AFRICA. Northern Cape: 32.20 (Sutherland) S of Sutherland (AC), 2 € 1999, bere & Nünni 11189 (MO, NBG). Western Cape: 33.20 (Montagu) foot of the Voetpadsberg, 21.5 km E of Touwsrivier, foot of sandstone slope (AC), 3 Oct. wee (fr), Goldblatt & Nünni 11200 (K, MO, NBG, PRE), Sep. 2001, Goldblatt & Porter 11877 (К, MO, NBG, P T: АС). 13. Hesperantha teretifolia Goldblatt, S. Afri- can J. Bot. 53: 460. 1987. TYPE: South Af- rica. Northern Cape: Roggeveld Escarpment between Middelpos and Calvinia, 13 Oct. 1983, P. ин 7090 (holotype, NBG!; iso- types, K!, MO!, PRE!, S!, STE!, WAG!). Described in 1987 from plants from the Roggev- eld Escarpment near the farm Botuin west of Mid- delpos, Hesperantha teretifolia is unusual in having centric, more or less terete, minutely grooved, hol- low leaves. A feature not noted in the protologue is that the leaf surface between the grooves is densely covered by minute papillae (or extremely short, cil- ia-like hairs) visible only under the microscope at 10X magnification, Specimens collected 10 years later, in September 1997, close to the type locality, growing in rocky sites among low bushes, clearly show this unusual epidermal feature, which is rare in the genus. A third population was discovered in 1998 at Uitkyk Farm, some 60 km south of the type locality, by N. A. Helme. Like the northern popu- lation, these plants have white flowers that open late in the afternoon, mostly after 16:30H, and are then sweetly scented. The leaves differ slightly in having the papillate hairs mostly confined to edges of the grooves, and the ridged areas between the grooves in the upper part of the leaf sometimes have a transparent center. Hesperantha teretifolia may be less rare than available records indicate, and exploration of the more rugged and remote parts of the Roggeveld Escarpment will likely yield additional populations. SOUTH AFRICA. Northern В and Middelpos, Additional specimens. Cape: 31.19 (Calvinia) between bun Botuin, 16 Sep. 1997, Goldble tt & Manning 10747 (MO, NBG); 32.20 (Sutherland) SW ofS Шш on the road to Calvinia, Farm Uitkyk (AA), 26 Sep. 1998, Helme 1556 (NBG 14. Hesperantha quadrangula Goldblatt, J. S. African Bot. 50: 62. 1984. TYPE: South Af- rica. Northern Cape: Hantamsberg, slopes be- low the summit cliffs, 16 Sep. 1980, P. Gold- blatt 5795 (holotype, MO!; isotype, NBG). Last revisionary account: Goldblatt, J. S. African Bot. 50: 62. 1984 Known from only two collections from the Han- tamsberg at Calvinia in Northern Cape Province when desc ies s quadrangula was found on the veld Escarpment west of Mid- dead in e (Goldblat & Nünni 11157), a range extension of some 60 km. The stony habitat of the Roggeveld population is similar to that on the Han- tamsberg, although the ground is only slightly slop- ing, unlike the steep flanks of the oe where the earlier collections were made. The white flowers open at 16:00—16:30H and then и ке а sweet, rose-like fragrance. Surprisingly for a white- flowered species, the tepals close again at nightfall, and the flowers are completely closed by 19:30H. The flowers are visited by a range of bees including Anthophora diversipes and Apis mellifera and by the occasional hopliine scarab beetle (Goldblatt et al., in press). The species is remarkable for the partic- ularly short perianth tube, ca. 3 mm long, and may reliably be identified by this feature alone or in combination with the brown cataphyll, two basal foliage leaves, and sheathing upper leaf, which is quadrangular in cross section. ж specimens. SOUTH AFRICA. Northern Cape: 31.20 (Williston) road to M 55 km SE of Calvinia, farm Knechtsbank (CC), 19 Sep. 1999, Goldblatt & Nünni 11157 (MO, NBG, PRE). 15. e flexuosa Klatt, Abh. Nat. Ges. Halle 15: 394. 1882. TYPE: South Africa. Northern poe Namaqualand, Kamiesberg, Elboogfontein, Aug.-Sep. 1830, J. Е Drége 2639 (holotype, B [“Herb. Lubeck”]!; isotypes, G!, P!) Last revisionary account: Goldblatt, J. S. African Bot. 50: 63. 1984. 16. oo minima او‎ R. C. Foster, Contr. Gray Herb. 135: 77. 1941. Geissorhiza minima Baker, Y Bot. (London) 5: 239. 1876. TYPE: South Africa. Northern Cape: Nama- qualand, Kamiesberg “Modderfonteinsberg,” Oct. 1830, J. Е Drége 2632 (lectotype, des- ignated by Foster (1941: 77), К!; isotypes, B!, G!, K!, L', MO!, P!, S. Last revisionary account: Goldblatt, J. S. African Bot. 50: 65. 1984. The diminutive Hesperantha minima was de- scribed by J. G. Baker for a dwarf species from Namaqualand collected in about 1830 by the no- table collector J. F. Drége, and then referred to the related genus Geissorhiza. R. C. Foster correctly transferred the species to Hesperantha in 1941. When the genus was revised in 1984, the plant was still known from this single collection, one so well Volume 90, Number 3 2003 Goldblatt 409 Hesperantha Review represented in the world's herbaria that there was no doubt that it represented a good species. The locality was also quite clear. Drege's collection is from “Modderfonteinsberg,” i.e., the mountain on the farm Modderfontein, where he also collected the plants later to be described as Xenoscapa uli- ginosa Goldblatt & J. C. Manning (Goldblatt & Manning, 1995). That species was rediscovered in 1990 by John Rourke and E. C. Nelson while ex- ploring Sneeuberg, the second highest mountain in Namaqualand, reaching 1591 m, and located partly on the farm Modderfontein. When I visited the site in 1991 to collect more material of the Xenoscapa, plants resembling H. minima were seen in leaf, growing in moss on thin soil on wet granite domes. Subsequently, flowering plants were located and their identity confirmed as H. minima (Goldblatt 9246, MO). Both H. minima and X. uliginosa are known from only this one locality, presumably the very site discovered by Drege over 170 years ago. 17. Hesperantha fibrosa Baker, Handbk. Iri- deae 149. 1892. TYPE: South Africa. Western Cape: Kleinrivier Mts., Aug., C. L. Zeyher 3960 (lectotype, designated by Goldblatt (1984: 66), K!; isotypes, B!, G!, K!, S!, SAM!, W!, 7!) А Last revisionary account: Goldblatt, J. S. African Bot. 50: 66. 1984. Collections of Hesperantha fibrosa from “Eland- skloof” (Esterhuysen 3160, BOL) suggested that the species occurs in the Cold Bokkeveld as well as to the south in the Caledon and Bredasdorp Districts of Western Cape Province where it was once com- mon (Goldblatt, 1984), ‘More likely the Elandskloof of the Esterhuysen plants is the less well known place of that name in the Bredasdorp Mountains near Napier, which lies within the expected range of the species. Unless the species is re-collected at Elandskloof in the Cold Bokkeveld, it must be as- sumed that H. fibrosa does not occur there and that it has a coherent distribution range in the southern Cape. 18. Hesperantha cucullata Klatt, Abh. Naturf. Ges. Halle 15: 393. 1882. TYPE: South Afri- ca. Northern Cape: Hantam Mts., Aug., H Meyer 9 (lectotype, designated by Goldblatt (1984: 72), B!; possible isotypes, S!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 72. 1984 19. Hesperantha truncatula Goldblatt, S. Afri- can J. Bot. 53: 461. 1987. TYPE: South Af- rica. Western Cape: northern foothills of the Kleinswartberg, farm Vleiland, 10 Sep. 1983, J. H. J. Vlok 662 (holotype, NBG!; isotypes, MO!, PRE!). 20. Hesperantha purpurea Goldblatt, J. S. Af- rican Bot. 50: 85. 1984. TYPE: South Africa. Northern Cape: farm Perdekraal, NW of Cal- vinia, 12 Sep. 1981, P. Goldblatt 6246 (holo- type, MO!; isotypes, K!, NBG!, PRE!, S!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 85. 1984 21. Hesperantha vaginata (Sweet) Goldblatt, J. S. African Bot. 36: 298. 1970. Geissorhiza va- ginata Sweet, British Fl. Gard. 2: pl. 138. 1836. TYPE: South Africa. Without precise lo- cality, British Fl. Gard. 2: pl. 138. 1836 Last revisionary account: Goldblatt, J. S. African Bot. 50: 72. 1984 Hesperantha vaginata is the only species of the genus known to be pollinated exclusively by hopli- ine scarab beetles, Scarabaeidae: Hopliini (Gold- blatt et al., in press). The unscented, yellow flowers, in most populations conspicuously patterned with chocolate-brown, are visited by numerous Clania glenyonensis individuals on warm days after the te- pals unfold after 14:30H. The beetles use the flow- ers as sites for assembly, mate selection, and cop- ulation, but they also consume pollen, which appears not to affect the fitness of the species, which reproduces well (Goldblatt et al., in press). 22. Hesperantha karooica Goldblatt, J. S. Af- rican Bot. 50: 79. 1984. TYPE: South E Northern Cape: Calvinia, foot of the Hantams- berg, 25 Aug. 1968, F J. Stayner s.n. (holo- type, NBG 87606!; isotype, STE!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 79. 1984 Although so closely resembling Hesperantha va- ginata that there remains doubt about the status of H. karooica, no plants resembling either species have been re-collected at Calvinia, the type locality, or elsewhere that match the small flowers and di- minutive size of the type plants. In the absence of intermediates the species must continue to be rec- ognized. Annals of the Missouri Botanical Garden 23. Hesperantha hantamensis Schltr. ex R. C. oster, Contr. Gray Herb. 166: 15. 1948. TYPE: South Africa. Northern Cape: Calvinia, dolerite hills, Aug. 1921, К. Marloth 10262 (holotype, B!; isotypes, PRE!, STE! Goldblatt, J. S. African Last revisionary account: Bot. 50: 84. 1984 24. Hesperantha humilis Baker, J. Bot. 14: 239. 1876. TYPE: South Africa. Northern Cape: Roggeveld near Jakkalsfontein, 7 Aug. 1811, W. J. Burchell 1320 (holotype, K!). Figure 1B. Last revisionary account: Goldblatt, J. S. African Bot. 50: 84. 1984 Hesperantha humilis is readily recognized by the more or less acaulescent habit and 1- to 3-flowered spike of deep pink flowers with an elongate peri- anth tube. As in other species of the genus with a similarly colored perianth, the flowers open during the day and close in the mid afternoon. The peri- anth tube, 17—24 mm long, seems to indicate that the flowers are pollinated by a long-tongued insect. However, the tube is narrow and the walls tightly envelop the style, leaving no room for an insect's tongue. The little nectar produced is forced into the top of the tube and is accessible to insects with c mouth parts no more than 5 mm long. 25. си flava С. J. Lewis, S. African . 23: 255. 1933. TYPE: South Africa. e m Whitehill, 31 July 1937, R. H. Compton 4276 (lectotype, designated by Gold- blatt (1984: 82), BOL!; isotype, K!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 82. 1984 hen described in 1933, the acaulescent, yel- low-flowered Hesperantha flava was known from a single population near Whitehill in the southwest- ern Karoo of Western Cape Province. Even in 1984 when the genus was revised for the winter-rainfall zone (Goldblatt, 1984), it was known to me from only two more populations, one near Steinkopf in northern Namaqualand in Northern Cape Province, and the other from Matjesfontein, near Whitehill, a disjunction of some 500 km. Such disjunctions are not common in the southern African flora and were thus suspect. Two more populations are now known from the area between these two stations, one from near Kliprand in southwestern Namaqualand (Oli- ver 9846A, NBG), and the other between Middelpos and Calvinia (Manning s.n., NBG) in the western Karoo. The range now forms a coherent pattern, and H. flava may be assumed to extend more or less continuously in suitable habitats along the interior edge of the winter-rainfall zone from Steinkopf in northern Namaqualand to Matjesfontein and White- hill. Its flowering time, early winter, often in May or June, largely explains why it is so poorly known, for little collecting is done at this time of the year. 26. не candida Baker, Handbk. Iri- deae l: 892. TYPE: South Africa. Free State, 1861. ү Н. Cooper 746 (holotype, К!; NH!, PRE!). isotypes, E!, а vernalis pow ta Burtt, Notes Roy. Bot. Gard. y ria Н 43: 1986. Syn. nov. TYPE: South Africa. azul Natal Estcourt, S side Kambe rg, 9 Sep. 1973, Е B. Wright 1530 (holotype, ; not seen; isotypes, MO!, NUN. Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 423. 1986, as Hes- perantha vernalis. Although I believed Hesperantha candida to be conspecific with H. longituba (Goldblatt, 1984), Hilliard and Bu (1986) argued convincingly that these were two separate species, H. candida called by them Н. vernalis. The two differ mainly in floral features, especially the length of the perianth tube, timing of flower opening, and sometimes perianth color. One of just a few eastern southern African species of the genus that flower in the spring, H. candida may be recognized by the moderate-sized to fairly large flowers, the tepals mostly 15-20 mm long. usually pink to mauve-pink but occasionally white, with a perianth tube (5-)7-12(-14) mm long, and anthers 5-7.5 mm long. Plants bear leaves when in flower, although the leaf tips may be dry, or in exceptional cases are lacking, having been grazed or burnt off before the flowering stem has emerged. Vegetatively similar, even apparently identical, H. longituba consistently has white flow- ers with a longer perianth tube, usually at least 14 mm and up to 22 mm long. I admit to sometimes finding it difficult to distinguish herbarium speci- mens of H. candida from H. longituba, for the range in the length of the perianth tube in the two species overlaps and the perianth of H. candida is also sometimes white, although then usually the outer tepals are flushed or feathered with pink, whereas flowers of H. longituba are, as far as | know, uni- formly white. Leaves of Hesperantha candida are typically sword-shaped or nearly linear with a raised midrib and margins and an otherwise plane surface. A few collections from the Eastern Cape (e.g., Bester 850 and 866, PRU) consist of plants with oblong to nearly ovate leaves that appear to lack thickened margins. Other collections, among these Bester 830 Volume 90, Number 3 2003 Goldblatt Hesperantha Review (PRU), are intermediate, having the lowermost leaf sword-shaped and the other basal leaves oblong, leaving no doubt that the broader-leaved plants be- long to the same species, H. candida. Whatever the shape of the leaves, the leaf surface between the margins and midrib is always plane and smooth, with secondary veins not thickened and seldom vis- i Hesperantha candida was called H. vernalis by Hilliard and Burtt (1986) because they associated the type (Cooper 746) with what I consider a second species, H. leucantha. This is a fairly slender, pale pink- to mauve-flowered species of the Drakens- berg and nearby that flowers from December to February. Like H. candida the flowers open during the day but they seem generally smaller, with tepals 12-15 X 3-5 mm long and anthers 4.5—5.5 long, whereas those of H. candida have tepals most- ly 15-20 X 6-9 mm, and anthers 5-7.5 mm long. Leaves of the two species also differ, for those of mm H. candida usually have a prominently thickened midrib and margins, are usually fairly broad (2—5 mm wide in the type, but on most other collections 3-6 mm wide), and the leaf tips are often dry and the leaves of H. leucantha are usually quite slender, mostly 1.5-2.5 mm wide, have only slightly thick- ened midribs and margins, and are rarely dry at the tips. Flowering time, which would be useful addi- tional information for the interpretation of H. can- dida, is unknown. Cooper made the type collection n 1861, and according to records (Gunn & Codd, 1981) he crossed the Lesotho frontier at Ficksburg. reaching Harrismith on 25 September where he re- mained until 10 October. He may well have col- lected the type of H. candida at this time. Unfor- tunately, the details of his subsequent travels are sometimes broken at anthesis. In contrast, incomplete. Records indicate that he was in Pie- termaritzburg in July of the following year, having reached there via Ladysmith. Whether he was still in the Free State that summer is evidently unknown so that the month of his collection of H. candida cannot be determined with any confidence. After re-examining the lectotype (which is duplicated at E, G, PRE, and Z), I remain convinced that candida is the spring-flowering species allied to H. longituba. Hesperantha candida has a fairly wide distribu- tion, plants extending not only from the Witteberg in the Eastern Cape to Harrismith in the Free State, but farther north into southern Mpumalanga, if | have correctly placed Rogers 21282 (BOL, K, PRE), from Carolina. Plants of this gathering have flowers with a tube 11—14 mm long, tepals ca. mm long, and anthers ca. 6 mm long. Collections provisionally assigned here also extend the range southward as far as the Hogsback in the Amatola Mountains (Giffen s.n., GRA, PRE), Grahamstown (Britten 2800, GRA; Jacot Guillarmod 8575, GRA; Schénland 775, PRE; Snijman 464, NBG), and Al- exandria (Archibald 5977, PRE). The plants from Grahamstown, at the southwestern extremity of the range, have the smallest flowers, the tepals ca. 12 X 6 mm, found in the species. Some specimens also have more flowers per spike than has been recorded elsewhere. They have 4 to 6 flowers per spike versus the more usual 1 to 3 in populations further north. These plants are puzzling, but it is more likely that they represent a southern and low- er elevational form of H. candida than an unde- scribed species. Sometimes identified as H. falcata, they do not have the bell-shaped corm of this south- ern and western Cape species. 27. Hesperantha debilis Goldblatt, sp. nov. TYPE: South Africa. Eastern Cape: Grahams- town Nature Reserve, W of Dassie Krantz, Oct. 1951, R. H. Martin s.n. (holotype, RUH 9554!). Plantae 10-15 cm altae plerumque pauciramosae, cor- mo ovoideo 4—5 mm diam., tunicis lignosis concentricis, foliis 4, inferioribus dudbus basalibus, linearibus 1-2 mm latis summo ey marginibus costaque leviter incras- а l- vel 2-flora, floribus albis, tubo perianthii 4—6 mm pata майт» 12-14 X ca. 5 mm, filamentis 2.5— 3.0 mm longis, antheris ca. 5 mm longis, styli ramis ca. 10 mm longis antheras excedentibus. satis, spic Plants 10-15 ст high, stem usually branched. Corm ovoid, 4—5 mm diam., tunics woody, concen- Leaves 4, the lower 2 basal, linear, the third partly sheathing, free in the upper half, the upper- most largely sheathing, 15—20 mm long, inserted in the upper third of the stem, blades the margins and midrib slightly thickened. Stem tric. 1-2 mm wide, erect, slender, usually branched at the i ndo node and sometimes at the second node. Spike 1 or 2-flowered; bracts 8—11 mm long. green or flushed purplish, the margins and apex membra- nous. Flowers white, the outer tepals flushed red on the outside; perianth tube cylindric below, expand- ed near the apex, 4—6 mm long; tepals subequal, elliptic, 12-14 X ca. 5 mm, spreading. Filaments 2.5-3 mm long; anthers ca. 5 mm long, ?yellow, pollen ?yellow. Ovary ovoid, ca. 3 mm long; style branches ca. 10 mm long, overtopping the anthers by up to 2 mm in the closed flower, spreading in the open flower. Capsules globose, ca. 3 mm long; seeds unknown. Flowering. October. 412 Annals of the Missouri Botanical Garden Distribution. South Africa, Eastern Cape, near type. The relationships of H. longituba presumably Grahamstown and the Suurberg, on sandy slopes. Plants first collected in the 19th century in the Grahamstown district of Eastern Cape Province and still poorly known, match no known species of Hes- perantha although they are not particularly distinc- tive. Described here as H. debilis the species may be recognized partly by the small corms 4—5 mm in diameter, lax, linear leaves 1-2 mm wide, and a slender, branched stem bearing one or two flowers on each branch. It may be most closely allied to the eastern southern African Н. candida, a more robust plant, the stems of which are rarely branched. The bracts of H. debilis are soft in texture and have broad membranous margins. Paratypes. SOUTH AFRICA. Eastern Cape: 33.26 (Grahamstown) New Years River, Albany (AC-AD), with- out date, Barber 255 (K); near Grahamstown (BC), without date, Bolton s.n. (K 28. Hesperantha longituba (Klatt) Baker, Gard. Chron. 7: 652. 1877. Geissorhiza longituba latt, Linnaea 35: 383. 1867-1868. TYPE South Africa. Eastern Cape: Somerset East, date unknown, J. H. Bowker s.n. (lectotype. че у by Goldblatt (1984: 69), К!; iso- types, K!, S Mpeg Pied var. bicolor Baker, Fl. Cap. 6: 63. rantha 2. е R. C. Foster, p дал erb. 166: 948. TYPE: South Afri- ca. Somerset Fast, Bosc ы P. MacOwan s.n. (lec- totype, designated by Foster (1948: 5), K!; isotypes, BOL!, G!, GRA!, M!, WU!, Z). Last revisionary account: Goldblatt, J. S. African Bot. 50: 69. 1984 Largely a species of the southern Karoo and ad- jacent dry mountains of the Eastern Cape, Hesper- antha longituba flowers in the spring when the veld is usually dry. The white flowers open in the late afternoon, according to Hilliard and Burtt (1986), and close after dark. The species may be recog- nized by the relatively large flowers with tepals 15— 22 mm long and a perianth tube usually 16-22 mm long, and by the broad, leathery leaves with prom- inent margins and midrib. Unlike most species of eastern southern Africa, the corms are large and have overlapping tunic layers notched into seg- ments below, like those of several western Karoo species, e.g., H. cucullata, H. vaginata, included in section /mbricata (now sect. Concentrica) by Gold- blatt (1984). Most other eastern southern African species have corms with concentric tunics that frag- ment along vertical fracture lines into discrete seg- ments and appear to be a fundamentally different lie with this predominantly winter-rainfall group of species. Among the spring-flowering species of the east- ern southern African summer-flowering zone, Hes- perantha candida is the only one that may be con- fused with Н. longituba. That species (called Н. vernalis by Hilliard & Burtt, 1986) differs mainly in the length of the perianth tube, mostly 14—22 mm long (and associated bracts), and its flowering phenology. The latter feature cannot be determined from most herbarium specimens, but Hilliard and Burtt have determined that flowers of the shorter- tubed species (tube mostly 7-12 mm long) are di- urnal (confirmed by notes on some herbarium col- lections) and those of the longer-tubed species are crepuscular. This provides additional evidence for regarding them as separate species. Plants from the Aliwal North and Stutterheim areas are intermedi- ate in tube length and make the distinction between he two sometimes seem arbitrary. Collections from further west, at Cradock (Marloth 2152, PRE), Somerset East (van der Walt 186, PRE), and Graaf Reinett (Linger 2106, PRE), always have a longer perianth tube, sometimes up to or even exceeding 2 mm. As indicated in the key, H. longituba may be distinguished by having a perianth tube mostly 16-22 mm long and white flowers that open in the later afternoon and remain open at night. In con- trast, H. candida has a perianth tube mostly 7—1 mm long and white or pale pink flowers that are open during the day and closed at night. Hesper- antha candida was included in H. longituba by Goldblatt (1984), but Hilliard and Burtt (1986) ar- gued convincingly that they are separate, albeit al- lied, species. 29. Hesperantha schelpeana Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 3 1979, TYPE: Lesotho. Black Mts. Sani and Mokhotlong, 5 Nov. 1973, О. liard & B. L. Burtt 7075 (holotype, E not seen; isotype, NU!). Figure 3 between Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 434. 1986 Collections made since Hesperantha schelpeana was described in 1979 make it necessary to amplify the circumscription of the species. Plants may have one or two flowers (Bester 3098, Strever 433, PRU), or even three (Bester 2961, PRU) or more (Hilliard & Burtt 13117, NU), and although described as leafless when in flower, plants usually bear two sheathing leaves (the cataphylls of Hilliard & Burtt’s description) as well as a basal cataphyll that Volume 90, Number 3 Goldblatt 413 2003 Hesperantha Review | B), full size. Drawn by John Manning from pressed plants and accompanying photographs (H. schelpeana: € ue s.n, NBG; H. altimontana: Cubitt s.n., NBG (flowering Slim?) and Goldblatt 11239, NBG (leaf)). Scale bar 1 сп 3 Figure 3. Morphology of H. аи (A) and Н. altimontana 414 Annals of the Missouri Botanical Garden is colorless, and occasionally the sheathing leaves overlap and may even have a short, unifacial tip. Collections from Naude's Nek (Strever 433, PRU), Lapa Munnik Pass (Bester 3098, PRU), and Saal- boom Nek (Hilliard & Burtt 13117, NU) show that plants occasionally bear one long basal leaf and — sometimes a second shorter one at flowering. These longer leaves are often dry or dying back at flow- ering time, but are evidently occasionally still at least partly green. The leaves are about 2 mm wide, and it appears from the way they dry that they are As they dry the soft internal tissue shrinks, leaving each surface with tw oval in section when alive. ч - or more narrow but irregular, longitudinal grooves, making it appear as if the living leaf had thickened veins and thinner intercostal areas. Ev- idently new leaves are normally produced on a sep- arate shoot after flowering, but sometimes the fully developed leaves of the past season persist until next flowering. The occasional presence of leaves at flowering time makes it possible to confuse Hesperantha schelpeana with H. candida, which flowers at the same time of year and normally has leaves present on the flowering stem (unless destroyed by fire or grazing). Properly developed plants of H. candida have four leaves, two basal, a third subbasal and partly sheathing, and a short, sheathing leaf in the upper part of the stem. Hesperantha schelpeana, however, never has more than three leaves, the low- ermost slender and exceeding the stem (when still present at flowering), the second much shorter, of- ten largely sheathing, and a third, short, sheathing leaf usually present on the upper part of the stem. The leaf blades also differ, and H. candida can be distinguished by the fairly broad leaves with clearly thickened and raised margins and midrib and the surface plane and smooth elsewhere. Several spec- imens assigned in herbaria to Н. candida (or its schel- peana in which the foliage leaves have persisted synonym Н. vernalis) seem to me to be H. later in the year than usual, and then the two spe- cies can only be separated by examination of leaf number, size, and shape of the blade. ıe flowers of Hesperantha schelpeana were de- scribed by Hilliard and Burtt as whitish or pale pink with the outer tepals feathered red-purple out- side. Plants from The Sentinel on the Free State— KwaZulu-Natal border in the Drakensberg Mts., at the northern limit of the species (Cubitt s.n., NBG). have pink flowers, shading darker at the tips of the tepals, with the throat dark brown, providing a stark contrast to the deep yellow anthers and pollen. The reverse of the outer tepals is so strongly feathered with brown as to appear nearly uniformly pigment- ed. Some plants from the southern end of its range (e.g.. Bester 2961, 3018, PRU) have white flowers with the outer tepals barely tinged with mauve near the tips. 30. Hesperantha altimontana Goldblatt, sp. TYPE: South Africa. Free State: Drak- The Sentinel, ca. 2500 m, 10 Cubitt s.n. (holotype, NBG!). nov. ensberg Mts., Oct. 1979, G. Figure 3B. Plantae 10-12 cm altae eramosae, cormo ovoideo, 8— 12 mm diam., foliis caulis florentis 2 omnino vaginantibus 2.5-3.5 cm longis, foliis ps anthesis ү tis solitariis faleatis 6-8 cm X mm, spica l-flora, flore cremeo- albo, tubo perianthii 20-27 mm longo, oe 20-23 х 9-10 mm, filamentis ca. 3 mm longis, antheris ca. 10 mm longis, styli ramis ca. 10 mm longis. Plants 10-12 ст high, erect, unbranched, flow- ering without the leaves. Corm ovoid, 8-12 mm diam., with relatively soft, imbricate tunics. Leaves of the flowering stem) 2, entirely sheathing, 2.5— 3.5 em long; foliage leaves produced later in the ж jii season, solitary, falcate, to 6-8 ст long, 3—4 mm wide, leathery, the midrib hyaline but not raised when alive. Spike 1-flowered; bracts ca. 20 mm long, green, flushed with purple, the inner slightly shorter than the outer. Flowers uniformly creamy white, shading to pale yellow both inside and out- side the tube; perianth tube 20-27 mm long, cylin- drical, expanded near the mouth; tepals spreading, elliptic-ovate, 20-23 X 9-10 mm, subacute. Fila- ments erect, ca. 3 mm long, inserted in the mouth of the tube; anthers diverging, ca. 10 mm long, yel- low, pollen yellow. Ovary oblong, ca. 5 mm long; style dividing in the mouth of the tube, the branch- es reaching to the upper third of the anthers in bud, ca. LO mm long, axly spreading above. Capsules and seeds unknown. Flowering. October. Lesotho and South Africa in Free State and KwaZulu-Natal Provinces, ground in the high Drakensberg. Distribution. stony, open Hesperantha altimontana is presumably allied to the two other southern African Drakensberg species that flower early in the season and without their foliage leaves, H. crocopsis and H. schelpeana. Of the two, it is most like H. schelpeana in its fairly large flowers, but it differs in having a perianth tube 20-27 mm long and uniformly creamy white peri- anth, whereas H. schelpeana, with a perianth tube 5-10 mm long. has whitish or pale pink flowers with the outer tepals flushed with darker pink to purple on the outside (Fig. 3A). Hesperantha cro- copsis has much smaller flowers than H. altimon- Volume 90, Number 3 2003 | 415 Goldblatt Hesperantha Review tana, the tepals only 6-7 mm long, but a perianth tube 9-14 mm long, and like H. schelpeana the outer tepals are red to purple on the outside. Ilus- trations of H. schelpeana (Hilliard & Burtt, 1979) and Н. altimontana (Trauseld, 1969: 41, number 670, as Hesperantha species) show the difference between the flowers of the two species. This plant was known to Hilliard and Burtt who discussed the few available specimens under Hes- perantha schelpeana, concluding that they probably represented a new species. Hesperantha altimon- tana is thus the third species of the genus, all of the high Drakensberg, in which the foliage leaves are produced after flowering on new shoots on the corm near the base of the flowering stem. Foliage leaves are lacking on flowering specimens, but plants from ridges west of Thaba Ntlenyana in Le- sotho (Goldblatt 11239, NBG) that appear to be H. altimontana have a distinctive, falcate leaf with a plane blade like no other known member of the genus. This collection is assumed to be vegetative H. altimontana. The illustration of H. altimontana (Fig. ЗВ) is based on the collection and photo- graphs made by the photographer Gerald Cubitt and vegetative specimens from Thaba Ntlenvana. Paratypes. LESOTHO. Black Mountain Pass on slopes 3200 m (CA), 5 Feb. 2000 (fr), SOUTH AFRI psa KwaZulu-Natal: 29.29 ова rberg) Giants Castle Game "m The a. (BC), 25 Oct. 1966, oo 670 (NU, PRE), date unknown, aan ld 1104 (PRE not seen). 29.29 (Underberg) top of W of Thaba Ntlenyana, wre 11239 (NBG). 31. Hesperantha crocopsis Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 37: 302. TYPE: Lesotho. Mokhotlong District, above Mashai Pass, 7 Nov. 1977, O. M. Hil- liard & B. L. Burtt 10489 (holotype, E not seen; isotype, NU!). Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 435. 1986. 32. Hesperantha bachmannii Baker, Bull. Herb. Boissier ser. 2, 1: 863. 1901. TYPE: South Africa. Western Cape: near Hopefield, date unknown, F E. Bachmann 1177 (lecto- type, designated by Foster (1948: 4), Gl; types, B!, 7! iso- Last revisionary account: Goldblatt, J. S. African Bot. 50: 90. 1984 33. Hesperantha bulbifera Baker, J. Bot. 14: 183. 1876. TYPE: South Africa. Eastern Cape: Somerset East, Boschberg, Nov. 1876, P. MacOwan 2215 (lectotype, designated by Goldblatt (1984: 95), K!; isotypes, BOL!, G!, K!, PRE!, WU!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 95. 1984 Hesperantha bulbifera remains a puzzling spe- cies, for although readily recognized by the white flower with a curved perianth tube and soft-textured leaves, it has an unusually scattered distribution across the southern African summer-rainfall zone. Populations extend from Somerset East in Eastern Cape Province in the south to Limpopo Province and Sabie in Mpumalanga in the east, always in wet habitats such as wet cliffs and within the spray of waterfalls. It is poorly distinguished from the re- lated H. bachmannii by larger size, a characteristic cormlet in the lower leaf axils, and sometimes a longer perianth tube, but for the present it seems advisable to continue to recognize the species. Al- though mostly flowering in the summer months, the type collection, from Somerset East, was in bloom when collected in September. The collection from the Soutpansberg, Venter 6205 (PRE, July, 1981) and another from Iron Crown in the Wolkberg (Ven- ter s.n., PRE, 15 Oct. 1985), extends the range of the species substantially to the north of its next nearest sites at Thabazimbi (Venter 1936, PRE) and Sabie (Cunliffe sub Moss 4311, K, Z) (Goldblatt, 1984). 34. Hesperantha pallescens Goldblatt, J. S. Af- 50: 88. 1984. TYPE: South Africa. Western Cape: below Piekenierskloof Pass, 3 Sep. 1980, P. Goldblatt 5645 (holotype, MO!; isotypes, B!, BOL!, C!, K!, M!, NBG!, P!, PRE!, S!, US!, WAG!). rican Bot. Last revisionary account: Goldblatt, J. S. African Bot. 50: 88. 1984. My field studies at the only known site for the species indicate that this narrow endemic is seri- ously threatened. Plants grow at the edges of cul- tivated fields at the foot of Piekenierskloof Pass, north of Piketberg. The long-tubed, pale yellow flowers open in the morning and close at sunset and are unscented, Perianth tube length, 18-22 mm, and self-compatibility (Goldblatt, 1984) suggest the species may have a specialized pollinator, probably a long-proboscid fly (Goldblatt et al., in press). Annals of the Missouri Botanical Garden 35. Hesperantha baurii Baker, J. Bot. 14: 182. 1876. TYPE: South Africa. Eastern Cape: Transkei, Baziya Mountain, Mar., L. R. Baur 628 (holotype, K!; isotype, BOL!). Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 426. 1986. 35a. Hesperantha baurii subsp. baurii See Hilliard & Burtt, Notes Roy. Bot. Gard. Ed- inburgh 43: 426. 1986. 35b. Hesperantha baurii subsp. formosa Hil- liard & Burtt, Notes Roy. Bot. Gard. Edin- burgh 43: 430. 1986. TYPE: Lesotho. Sani top, W of border post, 9500 ft., 16 Jan. 1976, О. M. Hilliard & B. L. Burtt 8829 (holotype, NU; isotypes, E not seen, MO). By far the most common species of Hesperantha in eastern southern Africa, H. baurii extends from Satana's Nek near Engcobo in the Eastern Cape in the south through KwaZulu-Natal, Lesotho, the eastern Free State to Mpumalanga Province. Fre- quent in grassland from 500 to 2000 m, it favors relatively well-watered grassy slopes and may oc- casionally even be found in seasonal seeps or rocky pavement, where it flowers from mid to late sum- mer, December to March. Flowers are bright pink with large deep yellow anthers and pollen and when open from mid morning to about 16:00H they can make a striking display. In combination with these features, H. baurii can be recognized by the rela- tively short perianth tube, mostly 7-10 mm long. tepals 12-18 mm long, and a remarkably consistent vegetative morphology. with four leaves, the lower two largest and basal, the third leaf sheathing in the lower half, and the uppermost leaf usually short, inserted in the middle of the stem and entirely sheathing. Plants are erect, and only the most ro- The spike ranges from weakly to markedly flexuose and bust individuals may have a single branch. bears up to 12 flowers, occasionally more, and spikes with as many as 18 flowers have been re- ported (Hilliard & Burtt, 1986). My circumscription of H. baurii largely agrees with that of Hilliard and Burtt except that I regard the late winter- to spring- flowering plants from coastal Transkei and. Kwa- Zulu-Natal included by them in the species as the separate Н. modesta, which with its presumed syn- onym H. subexserta are here removed from H. bau- Apart from the spring flowering time, H. mo- desta may be recognized by the fewer-flowered spikes, mostly bearing 2 or 3 flowers, and the pres- ence of only 3, or occasionally 4 leaves, of unusu- ally soft texture. Subspecies formosa of Hesperantha baurii, de- scribed by Hilliard and Burtt (1986) for plants with few-flowered spikes (usually with 2—4(—6) flowers), normally only 2 basal and 1 sheathing, cauline leaf, and somewhat larger flowers, with tepals 15-21 mm long versus 12-16(-18) mm in subspecies baurii, from the higher parts of the KwaZulu-Natal Drak- ensberg and adjacent Lesotho, remains somewhat puzzling. It is usually easy to distinguish this sub- species from more typical H. baurii, but plants from slightly lower altitudes appear to be intermediate between the two taxa. I have not seen enough plants in the field to make апу change to Hilliard and Burtt's taxonomy. Their record of the co-occurrence at Sani Pass of subspecies formosa in fruit when subspecies baurii was in flower suggests that this plant deserves at least subspecies rank (or should even be recognized as a separate species). In the key to the genus (p. 402), I have included both subspecies, but care should be taken in identifying intermediate plants, which may not be accommo- dated in the key. A collection from damp sites in the Ngeli Forest (Balkwill & Cadman 2886, NU), flowering in late January, must be mentioned here. Plants bear 4 leaves, the lower 3 basal and the uppermost in- serted in the middle part of the stem. All the leaves have long blades, reaching almost to the base of the spike and 1 mm wide or less, narrower than any recorded in Hesperantha baurii. Despite the narrow blade, the midrib appears in the dry state to be strongly thickened and perhaps winged, with | Тһе slender stem bears only 4 flowers, themselves rel- the wings reaching over the blade surface. atively small: the tube 5-6 mm long, the tepals 9- 10 X ca. 3 mm, and the anthers ca. 4 mm long. These dimensions are all smaller than any encoun- tered in plants that can be confidently placed in H. baurii. From the scanty material available it is im- possible to say if this plant represents a new spe- cies or a depauperate form of H. baurii growing in shady conditions. 36. Hesperantha petitiana (A. Rich.) Baker, J. Linn. Soc. 16: 96. 1877. Ixia petitiana А. Rich., Tent. Fl. Abyssinia 2: 309. 1850. TYPE: Ethiopia. Tigre, near Mai Gouagoua, L. R. Quartin-Dillon & A. Petit s.n. (lectotype, designated by Goldblatt (1986: 138), P!). Last revisionary pice Goldblatt, Ann. Mis- souri Bot. Gard. 38. 1986 Unspecialized i in ed aspect, pale pink- or white-flowered Hesperantha petitiana seems hardly to differ from the common southern African H. bau- Volume 90, Number 3 2003 Goldblatt Hesperantha Review rit (Goldblatt, 1986). Although Hilliard and Burtt (1986) suggested that H. petitiana could readily be distinguished from H. baurii by the smaller anthers, 4—4.8 mm long versus 5-9 mm in H. baurii, ma- terial I have examined of H. petitiana has anthers 3—5 mm long. This tropical African species may also be distinguished in fruit from similar, short- tubed eastern southern African species by the larg- er capsules, (7—)10—15 mm long and often slightly exceeding the bracts. In H. baurii the capsules are mostly 6-9 mm long, thus shorter than the bracts that subtend them (Goldblatt, 1993). The two spe- cies also appear to have different seeds, those of H. petitiana often being angular and slightly larger. 1.0-1.2 mm diam., compared with those of H. bau- rii and its immediate allies, which typically have more or less uniformly rounded and small seeds, 0.8-1.0 mm in diameter. Spikes of H. petitiana of- ten bear only 2 to 4 flowers on nearly straight spikes, but sometimes up to 10 flowers, whereas flexuose spikes of 8 to 12 flowers or more are more common in Н, baurit. 37. Hesperantha glareosa Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 424. 1986. TYPE: South Africa. KwaZulu-Natal: Underberg, headwaters of Mlahlangubo River, 23 Jan. 1982, О. M. Hilliard & B. L. Burtt 15367 (holotype, NU!; isotype, E not seen). Last revisionary account: Hilliard & Burtt, Notes. Roy. Bot. Gard. Edinburgh 43: 424. 1986. Although Hesperantha glareosa is largely a spe- cies of the Drakensberg Mountains of KwaZulu-Na- < tal, specimens cited by Hilliard and Burtt (1986 from Mpumalanga (then the Transvaal) are similar to some depauperate specimens of H. glareosa and may be this species. It is impossible to say if these slender, one- or two-flowered plants (e.g.. Galpin 14329, BOL, PRE; Jacobsen 4780, PRE) are de- pauperate, four-leaved H. schlechteri, or truly H. glareosa as Hilliard and Burtt (1986) believed. The flowers are moderate in size, the tepals 10—14.5 mm long, but the perianth tube length given by Hilliard and Burtt for the species, (3—)5-15 mm, must be a typographic error, for all plants І have seen, including most of those examined by Hilliard and Burtt, have a perianth tube 5—7 mm long. Their key is correct in placing H. glareosa under those species with anthers 3—5.5 mm long, but specimens I have examined have anthers mostly 4.5—5.5 mm long. Depauperate and/or poorly pressed, and ap- parently small-flowered plants from Lesotho cited under H. glareosa by Hilliard and Burtt are prob- ably H. exiliflora of which they had no adequate material in 1986. Dieterlen 1095 (PRE, SAM) and Werger 1662 (PRE), which they cited, are probably that species. Differences between H. glareosa and H. exiliflora are discussed under H. exiliflora, spe- cies number 40 below. 38. Hesperantha schlechteri (Baker) R. C. Fos- ter, Contr. Gray Herb. 114: 64. 1936. Geissor- hiza schlechteri Baker, Bull. Herb. Boissier, ser. 2, 1: 863. 1901. TYPE: South Africa. Lim- popo Province: Woodbush, 27 Mar. 1894, R. Schlechter 4701 (syntypes, BOL!, G!, K!, P^). и similis №. E. Br. ех К. C. Foster, Contr. Gray Herb. 166: 24. 1948. Syn. nov. TYPE: South Africa. Mpuma iba Sabie District, Devil's Knuckles, Apr. 1887, F Wilms 1446 (holotype, K!; isotypes, B!, Z!). Because of the confusion over the identity of Hesperantha schlechteri and H. similis, both based only on their respective types (Foster, 1941), I in- clude an extended description of the species as cir- cumscribed here. Neither species was treated in the revision of the genus for part of eastern southern Africa (Hilliard & Burtt, 1986), and thus no mod- ern account of the species exists. Likewise, | cite all specimens that I have seen in support of my circumscription of H. schlechteri. Plants 18—30 cm high, stem usually unbranched, flexuose and often looped above the sheaths of the upper leaves. Corm ovoid, 8-12 mm diam., tunics woody, concentric. Leaves mostly 5, occasionally 4 usually the lower 4 (or 3) basal or subbasal and largest, the uppermost one or two inserted in the middle of the stem and sometimes entirely sheath- ing or with short free unifacial tip, linear, 2—4 mm wide, the margins and midrib not, or barely thick- ened when alive. Spike mostly 4- to 8-flowered, flexuose; bracts mostly 9-12 mm long, green, the inner about two thirds as long as the outer. Flowers pale to deep pink, the outer tepals with a coppery brown flush; perianth tube slender below, widening near the top, 5—7(-10) mm long; tepals subequal, lanceolate-ovate, 13—16 X 5 mm, spreading + at right angles to the tube when fully open in the morning. Filaments 2-3 mm long; anthers 4—4.5 (-6) mm long. Ovary ovoid, ca. 2.5 mm long: style branches 10—12 mm long, laxly spreading, exceed- ing the anthers by ca. 2 mm in the closed flower. Capsules oblong-obovoid, 7-9 X ca. З mm; seeds subglobose or the sides flattened by pressure, ca. 0.6 mm long. January to mid March. South Africa, Limpopo Province Flowering. Distribution. and Mpumalanga, along the Drakensberg escarp- 418 Annals of the Missouri Botanical Garden ment from Woodbush in the north to Long Tom Pass in the south, on rocks and thin soil on sandstone pavement. Although first collected in 1887 near Lydenburg in Mpumalanga Province, South Africa, by the Ger- man apothecary Friedrich. Wilms, е schlechteri was described in 1901 by Ј. С. from a later gathering made by Rudolf Sc ie "i in 1894 in what is now Limpopo Province of South Africa. Baker referred the species to the related genus, Geissorhiza, and it was only in 1936 that С. schlechteri was transferred to Hesperantha by the The Wilms collec- tion formed the basis for a second species, H. sim- ilis, described by Foster in 1948. Then known from relatively few collections, the two species remained American botanist R. C. Foster. poorly understood, and some specimens from Mpu- malanga that closely match the type of H. similis fairly closely were referred to “H. sp. (= Codd 9481, PRE)” by Hilliard and Burtt (1986: 429), who also identified a few specimens in herbaria as H. glareosa, a species otherwise largely restricted to Lesotho and the KwaZulu-Natal Drakensberg. 1 similis in H. easily distinguished from all the similar short- tubed, here include H. schlechteri, which is pink-flowered species of Hesperantha in eastern southern Africa, including H. glareosa, by normally having 5 narrow leaves with blades 2-4 mm wide and a flexuose stem, often with a pro- nounced loop above the sheaths of the uppermost leaves. The pink flowers have a relatively short perianth tube, 5—7(—10) mm long, and anthers are mostly 4—4.5 flowers with a perianth tube of similar length, an- mm long. Hesperantha glareosa has thers 4.5-5.5 mm long, and consistently only 4 leaves, the blades mostly 1-2 mm wide. Hesperantha schlechteri appears to have a fairly narrow distribution, occurring in the Woodbush and Wolkberg of Limpopo Province and the higher parts of the Mpumalanga Escarpment, which extends a short distance to the north and south of Long Tom Pass. Plants cited below from Mariepskop have four or five leaves, the blades only about 1 mm wide, and may not be correctly placed in H. schlechteri. Codd 9481 (PRE), above, from near MacMac Falls, is puzzling, for the The collection mentioned plants have five leaves, the blades of the lower ones 4—6 mm wide, unusually broad for any species of Hesperantha in Mpumalanga. The flowers are quite small, with tepals ca. 7 mm long and a tube ca. 5 mm long, and the anthers 4 mm long. It may be an unusual form of H. schlechteri or an undescribed Mac Mac Mpumalanga several times and failed to locate any species. I have visited the Falls area in Hesperantha there. А few more collections, includ- ing Goldblatt & Manning 9814 (MO, NBG) from Dullstroom, also represent this form. SOUTH AFRICA. Limpopo: ы kberg, rocky grassland on ae : to Serala Pea . 23 Feb. RE Goldblatt & I 11954 (MO, Ae E Mpur nalanga: 24.30 (Pilgrims Rost) summit of Black Hill, Pilgrims Rest (DC), 1 Mar. 193 Galpin 14329 (BOL, K); Mariepskop (DB), 10 Apr. 1958, van der Schiff 4382 (K, PRE), 25.30 (Lydenburg) Dullst- room, farm Verlorenvalei (AC), 27 Mar. 1985, Krynauw 265 (LYD, PRE); Mauc :hsberg (BA), 22 Dec. 1932, Smuts & Gillett 2286 (PRE); summit of Mt. Anderson, Mar. 1933, M 22486 (BOL); summit plateau, Mt. Anderson, Mar. 3, Galpin 21485 (BOL), Galpin 13779 (PRE); pas of jene Tom Pass, Feb. 1972, Goldblatt 611 (BOL), 6 Feb 1994 (fr), Goldblatt & Manning 9835 (MO, NBG | 6 Feb. 1994, Goldblatt & Manning 9836 (MO, е 2 Маг. 1996, Goldblatt & Manning eo MO, NBG); pe гов Pass, Whisky Spruit, 11 Feb. 1986, о 1032 (LYD Mokobu Nature Reserve, 20 i 1953, Marais 43 (PRE). Additional specimens. 23.30 (Tzaneen) < —. —. 39. Hesperantha leucantha Baker, Handbk. Ir- ideae 150. 1892. TYPE: South Africa. Kwa- Zulu-Natal, Oliviershoek Pass, 14 Jan. 1886, J. M. Wood 3437 (holotype, K!). Роот ser. е macra Baker, Bull. Herb. ‚4, асһасһе, ус, . 1904. TYPE: Lesotho. Mt. ae H. Jacottet 1937 (holotype, Ch), As outlined above under Hesperantha longituba, l regard H. leucantha as the correct name for the species called H. candida by Hilliard and Burtt (1986). The type of H. candida, described by J. С. Baker in 1892, is the spring-flowering species that Hilliard and Burtt called H. vernalis. Hesperantha leucantha can be recognized by the fairly straight stem, few-flowered spike, and pale pink to pale li- lac flowers with a relatively long perianth tube com- pared to the narrow tepals. In the plants I have seen from the Witzieshoek area of Free State, not far from the type locality of H. leucantha, the anthers and pollen are always white and the horizontally spreading style branches are unusually long in re- mm long. Although Hil- liard апа Burtt described the flowers as white or lation to the tepals, ca. 15 pale pink to lilac-mauve, white-flowered plants are unusual, and probably not part of the normal pat- tern of variation. Among the ample material at К, . NBG, and PRE Slide that appear to match the distinctive appear- there are no white-flowered ance of H. leucantha. The specific epithet leucan- tha (white flower) is therefore misleading, but I sus- pect that Baker assumed that the flowers were white. The species is distinctive in having very pale pink flowers, white anthers and pollen, and partic- ularly long, spreading style branches, which in the closed flower usually exceed the anthers by 3—4 Volume 90, Number 3 2003 Golablatt 419 Hesperantha Review mm. The weak, often inclined to drooping stem, and drooping leaves are also a notable feature of H. leucantha. Hesperantha leucantha is sometimes confused with H. glareosa, a species with deep pink flowers with a much shorter perianth tube, mostly 4—6 mm long. while the tepals are 10—14.5 mm long. Hes- perantha glareosa also has four leaves, the lower three basal and with firm, narrow blades, mostly 1-2 mm wide. In Н. leucantha the perianth tube is usually 12-15 mm long, and exserted from the bracts unless these are unusually long, and the te- pals are 9-12(-15) mm long across its main range in interior Lesotho, eastern Free State, and northern KwaZulu-Natal. The erect stem and flexuose spike of H. glareosa also help prevent confusion between these two species, for H. leucantha typically has a more or less inclined to drooping stem and straight spike, and fairly soft-textured leaves, weakly trail- ing over rocks. Plants from north of the Vaal River, in Gauteng and Northwest Province, included (as Hesperantha candida) in H. leucantha by Hilliard and Burtt, are somewhat atypical in their greater height and slen- der habit, but for the present it seems best to in- clude them here. Sheets at K and PRE of a collec- tion from Milner Park, Johannesburg (Moss 18285 are annotated H. mossii by N. E. imens referred to H. leucantha from Mt. Anderson (Galpin 13781, BOL, PRE), in what is now Mpu- malanga, are here assigned to the new species H. 2 rown. The spec- saxicola, as they have relatively large white flowers, exceptional for H. leucantha in both size (tepals 15 X 7 mm and small anthers only ca. 4 mm long) and the white perianth, and they occur in what would otherwise be outside its recorded range. Geissorhiza Machache, thought to come from the Transvaal by J. G. Baker, who described the species in 1904, matches closely collections of Hesperantha leucantha from northern Lesotho and southern Free State. The two plants of macra from Mt. and its type collection have spikes of one or two pink flowers, soft-textured leaves about 2 mm wide, and the perianth tube ca. 13 mm long. Until now the type has not been matched with any Hesperantha species, although it was known to belong to the genus by Foster (1948). Mt. Machache is actually in eastern Lesotho some 32 km from Maseru where Jacottet collected (Gunn & Codd, 1981), thus well within the expected range of H. leucantha. 40. Hesperantha exiliflora Goldblatt, sp. nov. YPE: Lesotho. 3 km from New Oxbow Inn on road to Moteng Pass, subalpine grassland, 2550 m, 3 Feb. 1987, D. J. B. Killick 4477 (holotype, PRE!; isotype, MO!). Plantae inr ec cm altae eramosae, cormo conico prope basem iam., tunicis lignosis, foliis usitate 3 inferioribus falcatis summo E maxima parte vagi- nens. laminis inferioribus 1.2-2.3 mm latis marginibus costaque leviter incrassatis, Bonet ad 3 3-flora, floribus roseis in ore dest ик perianthii 3.5—5.0 mm longo, tepalis бту ilamentis са. 2.5 mm longis, kan ahi 2: 233 mm me styli ramis 3.0— 3.5 mm longis. Plants mainly 12-25(-30) cm high. Corm сопіс, ca. 8 mm diam. near the base, with woody tunics soon breaking into elliptic segments tapering above into short points. Leaves normally 4, the lower 2 basal and longest, reaching to between the upper third of the stem and the top of the spike, the third leaf inserted shortly above the ground and largely sheathing, the uppermost leaf entirely sheathing and bract-like, 12-32 mm long, inserted in the middle to upper third of the stem, the blades of the lower leaves + linear, 1.2-2.3 mm wide, firm and erect, the midrib and margins slightly raised, hya- line at least when dry. Stem erect, unbranched. Spike mostly 1- to 3-flowered, lax and + straight; bracts 6—8(—9.5) mm long, green, the upper margin transparent, the inner bract slightly shorter than the outer, * membranous with 2 green keels, 2-lobed at the tip. Flowers pink, pale yellow in the mouth of the tube; perianth tube funnel-shaped, 3.5-5 mm long; tepals spreading, elliptic, 6-8 X ca. 2.5 mm, acute. Filaments ascending, ca. 2.5 mm long; an- thers ascending, 2.5-3.3 mm yellow, pollen yellow. Ovary oblong, 3—4 mm long; mouth of the tube, the branches 3—3.5 mm long, diverging above, reaching to between the middle of the anthers to just beyond the apices in the closed flower. Capsules and seeds long, shortly tailed, style dividing below the unknown. Flowering. January and February Distribution. Lesotho, Drakensberg plateau in subalpine grassland, often in damp sites. Although collections of Hesperantha exiliflora were at first associated with the Drakensberg Moun- tain species H. glareosa in both herbaria and the literature, | can see no particular reason to consider them closely related, beyond their common occur- rence. Plants of the few known populations of H. exiliflora have remarkably small flowers, the peri- anth tube 3.5—5 mm long and the tepals 6-8 mm long. The only other species of pink-flowered Hes- perantha from eastern southern Africa with com- parably small flowers, H. ingeliensis is rather dif- ferent for it has tepals 10—12 mm long and broader, falcate leaves, 5.5-7 mm wide. This is quite unlike the straight, slender leaves of H. exiliflora, which are only up to 2.3 mm wide and appear to have a 420 Annals of the Missouri Botanical Garden fairly soft texture. The straight stem and seemingly consistently few-flowered spikes of H. exiliflora are unlike the rather wiry, flexuose spikes of H. glar- eosa, which has larger flowers with a tube 5-6 mm long and tepals 10—12 mm long. Hesperantha exiliflora seems most closely allied to a second new species, H. brevistyla, of the north- ern high Drakensberg, which has pale pink flowers 7 X 3.5 mm, .9 mm long, and short ca. 6 mm long, that reach to about the lower third of the anthers at most. with a tube 7-9 mm long, tepals ca. short white anthers 2.5-3 style branches, The flowers of H. exiliflora dry a dark purple color, while those of H. brevistyla dry pale pink. In contrast, the an- thers of H. the style branches are 3-3.5 mm long, reaching to between exiliflora аге yellow and the middle of the anthers to just beyond their tips in the closed flower. LESOTHO, 29.28 (Marakabei) 1 km E of above the dip tank 5 km E of Ha 3 Feb. 2000, ы r-Smith 195 (BOL); Mountain Ка, 32.5 km E of Thaba Putsoa, 1500 m (AC), 20 Mar. 1983, Pala 5060 (PRE). 29.29 (Un- derberg) top of Sani Pass, marshy turf (BA), 20 Feb. 1985, Manning 550 (NU). Paratypes. Tiping wy 41. Hesperantha alborosea Hilliard & Burtt, Bot. Gard. Edinburgh 43: 421. South Africa. KwaZulu-Natal: а Cobham, Upper Polela Cave area, 13 Feb. 1979, О. M. Hilliard & B. L. Burtt 15367 CRIME К not seen; isotype, NU!). Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 421. 1986 42. Hesperantha brevistyla Goldblatt, sp. nov. TYPE: South Africa: Free State, Sentinel trail, rocky pavement below sheer basalt cliffs on trail to the chain ladders, 5 Mar. 2002, P Goldblatt & J. Porter 11989 (holotype, NBG!; isotypes, MO!, PRE!). Figure 4. Plantae 4—10(-16) ст altae eramosae, cormo nd prope basem ca. 3 mm diam., tunicis lignosis, foliis 4 ve 5, З vel 4 inferioribus falcatis summo pro maxima ies vaginante, laminis inferioribus 2—4—5) mm latis margi- nibus costaque leviter incrassatis, spica 1—4(—5)-flora, flo- ribus pallide roseis in ore tubi pr. tubo perianthii tepalis 7-8 3.5 mm, filamentis 1.8-2 mm fannie, antheris 2. 5-41 mm longis, styli ramis ca. 6 mm longis. 7—9 mm longo, са. 3 mm Plants 4—10(— diam. near the base, with woody tunics soon break- 16) ст high. Corm conic, ing into elliptic segments tapering above into short 3 or longest, about as long as the spike, uppermost leaf points. Leaves 4 or 5, the lower і basal and inserted between ground level and the middle of the stem, and partly or rarely entirely sheathing, the blades of the basal leaves falcate to suberect, firm textured, 2—4(—5) mm wide, the midrib and margins slightly raised, hyaline when dry. Stem unbranched. 1—4(—5)-flowered, + straight; bracts 9-13 mm long, green, the inner bract slightly shorter than the outer, 2-lobed at the tip. Flowers pale pink, cream in the mouth of the tube, outer tepals pale greenish with darker green veins; perianth tube funnel-shaped, 7-9 mm long, erect, Spike expanded in the upper 2 mm; tepals spreading hor- i 7-8 X ca. 3.5 mm, acute. Fila- ments ascending, 1.8-2 mm long; anthers ascend- ing. 2.5—4 mm long. white, pollen white. Ovary oblong, ca. 3 mm long, enlarging rapidly after pol- lination; style dividing at the base of the upper part of the tube, the branches suberect, reaching to be- izontally, elliptic, tween the base and middle of the anthers, ca. 6 mm long. oa oblong, 8—12 mm long; seeds angu- ar, + prismatic, reddish brown, ca. X 1 mm, the edges forming membranous ridges. Flowering. | February to mid March. South Africa, Drakensberg of Free State and KwaZulu-Natal, and probably also in Lesotho, on basalt slopes and Distribution. in the northern rocky pavement in shallow damp soil. The earliest collection of Hesperantha brevistyla appears to have been made by Hilliard and Burtt in 1984 on the trail from Royal Natal National Park to The Sentinel at Basuto Gate. The collection was referred by them to H. leucantha (as H. candida) without comment (Hilliard & Burtt, 1986). Plants ‘ollected nearby оп the trail to The Sentinel in 1999 (Goldblatt & Manning 11053, MO, NBG) rep- resent the same species, which I reluctantly iden- tified as H. leucantha following Hilliard and Burtt. The following year, 2000, I found growing nearby, however, but at lower elevation on Cave Sandstone slopes, quite typical H. leucantha (e.g.. Goldblatt & Nünni 11232, MO, NBG; Goldblatt & Porter 11991, MO, NBG), which has larger flowers with longer anthers and particularly long, spreading style branches. This population made it clear that the smaller plants growing on basalt are not this species, but represent a separate taxon that is read- ily characterized by small, pale pink flowers with the outer tepals pale greenish pink on the outside and veined dark green, and unusually small white anthers, 2.2—4 mm long. Perhaps the most distinc- tive feature of the species, which I am calling H. brevistyla, are the remarkably short, suberect style branches that reach to the base or lower third of the anthers. In most other species of the genus, the Volume 90, Number 3 Goldblatt 421 2003 Hesperantha Review e 4. Morphology and floral details of Hesperantha brevistyla. Scale bar 1 em; single flower and seed much Ж hon Drawn by John Manning from photographs and pressed plants (Goldblatt & Porter 11989, MO, NBG, PRE). Annals of the Missouri Botanical Garden laxly spreading style branches reach at least to the anther apices in the closed flower and in H. leu- cantha usually exceed the anthers by 3—4 mm. Comparison of Hesperantha brevistyla with Н. leucantha now seems inappropriate for there is no reason to regard the two species as particularly closely allied, despite the shared pale pink color of the flowers and the whitish anthers. Hesperantha brevistyla may, instead, be most closely allied to a second small-flowered Drakensberg species, H. ex- iliflora, which has pink flowers with a tube 3.5—5 mm long, tepals 6—8 X ca. 2.5 mm, yellow anthers 2.5-3.3 mm long, and style branches 3-3.5 mm long, reaching to between the middle of the anthers to just beyond their tips in the closed flower. This contrasts with H. brevistyla, which has a tube 7—9 mm long, tepals ca. 7 X 3.5 mm, white anthers 2.5—4 mm long, and style branches ca. 6 mm long, reaching to the lower third of the anthers at most. 'aratypes.. SOUTH AFRICA. Free State: 28.28 (Bethlehem) Drake nsberg prepa trail to The Sentinel, 5 Feb. 1999, Goldblatt & Manning 11053 The Pudding, in shallow wet vigi 6 Mar. 2002, Gold- blatt & Porter 11989 (MO, NBC ега 28.28 (Ве thlehem) Royal Natal Nat ional Park, r Jasuto Gate, . 7200 ft. (DB), 18 Feb. 1984, Hilliard & pm 17089 n MO, NU); Drakensberg plateau near Mont-aux-Sour- ces, damp flats along trail from the chain ladders to Tugela Falls, 5 Mar. 2002 (in fruit), Goldblatt & Porter 11977 (MO, NBG, PRE). 43. Hesperantha ingeliensis Hilliard & Burtt, Gard. Edinburgh 43: 424. South Africa. KwaZulu-Natal: Alfred District, Ngeli Mountain, 4 Jan. 1969, O. M. Hilliard & B. L. Burtt 5838 (holotype, NU!; isotype, E not seen). Last revisionary account: Hilliard & Burtt, Notes 1986. Roy. Bot. Gard. Edinburgh 43: 421. 44. Hesperantha lactea Baker, Handbk. Irideae 151. 1892. TYPE: South Africa. KwaZulu-Na- tal: Verulam, Nov., J. M. Wood 1118 (lectotype, designated by Hilliard & Burtt (1986: 431), K!) Last revisionary account: Hilliard & Burtt, Notes Hoy. Bot. Gard. Edinburgh 43: 431 Readily identified by its creamy, sometimes pal- est yellow flowers with a tube 6-8.5 spreading tepals 14—20(-23) mm long, and promi- nent anthers up to 8 mm long, Hesperantha lactea mm long, is well known from coastal and near KwaZulu-Natal, but is actually recorded from as far south as the Transkei and from Hlobane in the Vry- inlerior heid District in northern KwaZulu-Natal (Hilliard & Burtt, 1986). Like other white- or cream-flowered Hesperantha species of eastern southern Africa the flowers are open during the day (opening 11:45- 12:30H and closing 4:40—5:15H at Inchanga, near Durban). This contrasts sharply with most white- flowered species of the genus from the southern Af- rican winter-rainfall zone, the flowers of which open in the mid to late afternoon and close long after dark, and sometimes only after midnight (Goldblatt, 1984; Goldblatt et al., in press). 45. Hesperantha inconspicua (Baker) Gold- blatt, Basionym: Gladiolus incon- spicuus Baker, Bull. Herb. Boissier ser. 2, < comb. nov. 1005. 1904. TYPE: South Africa, as “Trans- vaal, Donkerhoek, 4 Jan. 1894," but both lo- cality and. collection number, and perhaps date, are incorrect, R. Schlechter 4188 (holo- type, С!). Plants 25—45 ст high. Leaves mostly 4, the lower 3 basal or subbasal, the uppermost inserted in the middle of the stem, partly to entirely sheathing, blade reaching to about the base of the spike, p егу, narrowly sword-shaped-linear, mostly 3 mm wide, the midrib and margins slightly thic k ened, the midrib rounded. Stem + erect, often branched from the axil of the third leaf. Spike most- 2-flowered; bracts green or purplish at the 10-15 than the outer. tips, mm long, the inner slightly smaller Flowers white, sometimes faintly flushed pink on the reverse of the outer tepals, of- ten only on fading, unscented; perianth tube slen- der, expanded near the tip, 7-8.5 mm long; tepals subequal, spreading at right angles to the tube, 12- 16(-18) thers 6.5—8 mm long, cream, pollen yellow. Ovary X 6-8 mm. Filaments 2.5-3 mm long; an- ovoid, 2-3 mm long, style branches reaching to apex of the anthers in the closed flower. Capsules subglobose, 4—7 mm long; seeds angular-prismatic 1.5-1.8 X 1.2 mm to = ovoid, Flowering. December to mid March. Distribution. South Africa, from the Blyde Riv- er hills and Lydenburg in Mpumalanga to the Um- tamvuna Gorge in southern KwaZulu-Natal. Although known since the 1890s when it was first collected in the Transvaal by Rudolf Schlech- ter, Hesperantha inconspicua has largely been over- looked. The species was described by J. G. Baker in 1904 who referred it to Gladiolus. The locality and collectors number on the type are confusing, for Schlechter was not at the purported type local- ity. Donkerhoek, east of Pretoria, at that date but Volume 90, Number 3 2003 Goldblatt 423 Hesperantha Review at Hammanskraal to the north. A collection bearing the same number as the type of Gladiolus incon- spicuus is Dicoma gerrardii Harv. ex F. C. Wilson. Since H. inconspicua is not known from the vicinity of Donkerhoek or Hammanskraal and the collection number is evidently incorrect, Schlechter may be assumed to have collected the type in December 1893, when he was in the Dullstroom—Lydenburg area of Mpumalanga Province, or less likely in March 1894, when he traveled in the mountains of Limpopo Province (where the species has not been recorded). Subsequently the collection must have been mislabeled. A later collection made by Ernest Galpin in the 1930s in the Little Berg in KwaZulu-Natal was im- mediately recognized as being a species of Hesper- antha but it was not associated with Schlechter’s plants. The Galpin collection from the Little Berg and a few more specimens of the species were in- cluded in H. hygrophila by Hilliard and Burtt (1986). Hesperantha hygrophila is largely a species of the KwaZulu-N grow on wet rocks or in marshes. Its leaves are atal Drakensberg, where plants distinctive among the eastern southern African spe- cies, being pale green, without the common gray bloom of most Hesperantha species, and the blades have thickened margins, prominent secondary veins and a raised midrib, which is flattened rather than rounded in outline and the edges of the thick- ened part form wings that extend outward over the laminar surface. The vegetative aspect and flowers of Hesperantha inconspicua are unexceptional among the eastern southern African species of the genus except that the perianth is white, sometimes faintly flushed pink on the reverse of the outer tepals, especially on fading. That the type has not until now been associated with any species of Hesperantha is not surprising. Only the holotype is known, at the Ge neva Herbarium, and consists of a plant mostly in bud. The single mature flower was boiled up by G. J. Lewis who at once realized it belonged in Hes- perantha. She annotated the sheet as H. baurii but did not publish her conclusion. Hilliard and Burtt thought the plant might be H. rupestris, which has somewhat smaller white flowers, with the outer te- pals usually red on the outside. Among the few other white or cream-flowered species of eastern southern Africa, H. inconspicua may also be con- fused with the coastal KwaZulu-Natal species, H. and dull yellow to brownish anthers and pollen. The leaves of the two species are virtually identical, both hav- ing thickened margins, a midrib, and the other veins obscure and not visible lactea, which has creamy-yellow flowers slightly raised, rounded when alive. Plants are typically fairly tall, mostly 40—60 cm high, KwaZulu-Natal are often smaller, sometimes no ut collections from southern more than 15 ст high. They may be a separate — axon. One collection from Ісмака River Gorge near Port Shepstone (van Wyk 7197, PRU) has par- ticularly small flowers and soft-textured leaves, and these plants are reminiscent of the Swaziland en- demic H. umbricola, which has the tepals ca. 5 mm long, a perianth tube ca. 4 mm long, and anthers ca. 2 mm. The resemblance is probably due to con- vergence. The related Hesperantha saxicola, which also has white flowers, often fading pink, stands out both in its shorter stature and unusual habitat, rocky out- crops and cliffs. These plants have drooping leaves and stems that are 10—18 cm long and have flex- uose spikes of up to five large flowers with tepals mostly 15-16 X 7 mm and anthers ca. 4 mm long. Hesperantha inconspicua often has smaller flowers with tepals 12-15 mm long, but longer anthers, 6.5-8 mm long, and more flowers per spike. Additional specimens. SOUTH заран Mpumalan- ga: 24.30 (P sni ees 2 km from Graskop on road to Blyde River Canyon (DD), 14 Mar. ul — & Burtt 14334. (NU). 25. 30 Ки» y grassland S of Mrge (AC), 8 Goldblatt 10867 (MO, NGB, PRE); vlei on farm Mathes, 20 Feb. 1985, Carser s.n. (1 i D). 26.30 (Carolina) Ermelo, % mi. W of Vossman's Beacon А 20 Feb. 1951, Codd 6384 (РКЕ). KwaZulu- Natal: 27.30 (Vryheid) near Vryheid (DD), Jan. 1936, Pole о 3897 (РКЕ). 29.29 (Underberg) е athkin P sx top of the Little Berg, under rocks (AB ar. 19 Galpin 11884 (PRE); Highmoor Forest he in ile oe iuam & Vahrmeijer 3582 (PRE); Coleford, S of Na- serve above Endewana River, 25 12 1976, Hilliard & p» 9565 (NU). 29. " ОШО ) Karkloof, Mbona P rocky se ‚ Ке 000, Goldblatt & Nünni 5 (MO, NBC), 30 Dec. on Nanni 153 (NBG); 8 mi. AN of Pietermaritzburg, marshy ground (CB), Nov. 1939, Thomas 9 (NBG). 30.30 (Port Shepstone) Um- tamvuna Nature Reserve, NE face of Iron Crown, е rocks (CC), 23 Dec. 1983, Abbott 1583 (PRU); River 4 Gars ‚ веер in shallow soil, 24 Jan. 1986, van 1 Wyk 7179 (PRU). 31.30 (Port Edward) Umtamvuna Waterfall (AA), 25 Oct. 1962, Strey 4468 (PRE). 46. Hesperantha saxicola Goldblatt, sp. nov. TYPE: South Africa. Mpumalanga: rocks at the top of Mt. Anderson, Mar. 1933, E. E. Galpin 13781 (holotype, PRE!; isotypes, BOL!, К!). Plantae 10-25 cm altae ex scopulis trahentes, foliis usi- 5-2.2 mm latis, caul ae OD eramoso, spica 2- ad 4-fl floribus albis, tubo perianthii (7—)11 mm longis, tepalis subaequalibus patentibus (11-)15-16 X 5.5-7.0 mm, fi- lamentis 2.5-3.0 mm longis, antheris ca. 4 mm longis, ramis styli antheras excedentibus Plants 10-25 cm high, trailing from cliffs, simple 424 Annals of the Missouri Botanical Garden or branched, corm with the tunics extended above as a neck of fibers. Leaves mostly 4, the lower 3 basal or subbasal, the uppermost inserted in the middle of the stem and partly to entirely sheathing, shorter than the basal, blades of the basal leaves reaching to about the base of the spike, soft-tex- tured, + linear, mostly 1.5-2.2 mm wide, the mid- rib and margins hardly thickened. Stem weak and arching toward the ground, unbranched. Spike 2- to 4-flowered; bracts green, often drying at the tips, 10-15 about as long as the outer. Flowers white, probably mm long, the inner two-thirds as long to unscented; perianth tube funnel-shaped, expanded in the upper half, (7—)1 1 mm long; tepals subequal, spreading at right angles to the tube, (11—)15— X 5.5-7 mm. Filaments 2.5-3 mm long; anthers ca. 4 mm long, cream, pollen yellow. Ovary ovoid, an 1.5-2 mm long, style branches exceeding the an- thers by ca. 1.8 mm in the closed flower. Capsules and seeds unknown. Flowering. | March-April. Distribution. South Africa, in Mpumalanga, and possibly Limpopo Provinces, on wet cliffs and rock seeps. Known to Hilliard and Burtt (1986) from a single collection (Galpin 13781) made in 1933, on Mt. Anderson in Mpumalanga, the species was referred by them to Hesperantha leucantha (which they called H. candida) in their study of the genus in KwaZulu-Natal and surrounding areas. That spe- cies, centered in the KwaZulu-Natal and Lesotho highlands, is generally less robust and has smaller, pale pink flowers, and except for the Mpumalanga collection they cited, does not occur along the Mpu- malanga escarpment. Hesperantha saxicola is con- fined to cliffs and damp rocks and grows in rock crevices, often associated with moss. Plants have trailing stems 10—25 cm long, long linear leaves up to 2.2 mm wide, and spikes of 2 to 5 fairly large flowers. The tepals are 11—16 mm long and either uniformly white or faintly flushed with pink mauve on the outside, while the anthers stand out in their relatively small size, ca. 4 mm long. Two more recent collections from nearby, close to the top of Long Tom Pass, Arynauw 335 (LYD) and Linder 3203 (PRE), closely match the Galpin plants. A collection from Mt. Sheba, a short dis- tance to the north of Long Tom Pass, appears to belong here, but the flowers have tepals slightly flushed with purple on the outside, leaves less than 1 mm wide, and flowers with a perianth tube ca. 7 mm long and tepals ca. 11 mm long, somewhat smaller than in the populations to the south. Similar small-flowered plants with a white perianth from God's Window (Létter 324, NBG) may also be this species. A recent collection made in mid February on Formosa Mt. east of Lydenburg (Burrows 7296) is included here with reservation. The plants appear to have white flowers (possibly fading pale mauve) but the collection notes describe the flowers as pink-mauve. Two of the three plants of the collec- tion have branched stems and the spikes have up — to five flowers, whereas other specimens of Hesper- antha saxicola that | have seen have unbranched stems and spikes with at most four flowers. The flowers do, however, have the distinctive short sta- mens with anthers only about 4 mm long. A feature not noted in other collections of H. saxicola is the corm, which has the tunics extended upward as a neck of fibers. Specimens examined. SOUTH AFRICA. Mpumalan- ga: 24.30 (Pilgrims Rest) Mt. Sheba Nature Reserve (DC), Apr. 1972, Goodman s.n. (J). 25.30 (Lydenburg) rocks at the top of Mt. Anderson (BA), Mar. 1933, Galpin 13781 (BOL, K, PRE): es Tom Pass, farm De Kuilen, on low cliffs, 13 Mar. . Krynauw 335 (LYD); between the summit of Lon 5 Pass and Mt. Anderson, shady cliffs, 20 Mar. 1982, Linder 3203 р Formosa Mt., SE of look-out tower, cool SE-facing slopes in rock cracks at 2100 m, 18 Feb. 2001, Fus 7296 (BKH, NBG). 47. Hesperantha hygrophila Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 40: 278. 1982; 43: 433. 1986. TYPE: South Africa. KwaZulu-Natal: Alfred District, Ngeli Moun- 1969, O. M. Hilliard & B. L. Burtt NH!, tain, 2 Jan. 5762 (holotype, E not seen; isotypes, NU!) Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 433. 1986. As described by Hilliard and Burtt (1986), Hes- perantha hygrophila has unremarkable white flow- ers fading pink with age and distinctive leaves, pale green in color, with the midrib, margins, and a sec- ondary pair of veins quite clearly thickened. The midrib is flattened and the edges of the flattened ridge arch slightly over the laminar surface. Plants with this leaf type are restricted to the KwaZulu- Natal-Lesotho Drakensberg and the Ngeli range to the south. White-flowered plants from the Kwa- Zulu-Natal Midlands and Little Berg (e.g., Galpin 11884, PRE, referred to H. hygrophila) have leaves quite typical of most other eastern southern African Hesperantha in their slightly raised midrib, rounded in outline, and slightly thickened margins plants represent another species, the сасна пате for which is Gladiolus inconspicuus, now transferred to Hesperantha, the identity of which has long been Volume 90, Number 3 Golablatt 425 2003 Hesperantha Review uncertain. С. J. Lewis examined the type in prep- branched. Corm conic, 7-9 mm diam. near the aration for her revision of Gladiolus in South Africa and referred the specimen to H. baurii. Apart from the leaf differences, H. inconspicua is generally more robust than H. hygrophila, has many flowers per spike (up to 16 in plants from Goldblatt & Nän- ni 11245, › branched in the axil of the third lea . NBG), and sometimes has the stem This narrower definition of Hesperantha hygro- phila leaves the species fairly uniform as regards vegetative morphology. The flowers are usually white, but specimens collected at Highmoor (Gold- blatt & Nünni 11248) have pale pink flowers, oth- erwise identical to typical H. hygrophila both veg- etatively and florally. The flowers of this population open between 8:30 and 9:00H and close again by Hesperantha hygrophila may also be confused with the cream(or pale yellow)-flowered H. lactea, which grows in drier, well-drained grassland at al- titudes of up to 800 m, rather than on wet rocks and scree in mountains mostly above 2000 m (Hil- liard & Burtt, 1986). deep yellow to light brown and tend to dry a brown- In H. lactea the anthers are ish color, a feature that conveniently separates it from H. hygrophila in which the pale cream anthers remain pale when dry. The differences in leaf ve- nation also separate the two, but this feature is not always easy to see in dried specimens. 48. Hesperantha rupestris N. E. Br. ex R. C. Foster, Contr. Gray Herb. 166: 23. 1948. TYPE: South Africa. Mpumalanga: Waterval Boven, among rocks, 29 Mar. 1929, C. E. Moss 17314 (syntypes, K!, PRE!). A fairly robust species, plants sometimes stand- ing 45—50 em high, Hesperantha rupestris is distin- guished from the closely allied H. baurii in having white flowers, the outer tepals flushed dark pink to red on the outside, as well as in the height of the stem. It is restricted to rocky habitats in the central Mpumalanga highlands. As in most other eastern southern African Hesperantha species in which the timing of flower opening is known, the flowers of H. rupestris are diurnal, being open in the morning, according to information on some herbarium sheets. 49. Hesperantha modesta Baker, Handbk. Iri- deae 150. 1892. TYPE: South Africa. Kwa- Zulu-Natal: Umlaas Location [or Bevaan River on the type at K, evidently in error], 17 July 1885, J. M. Wood 3201 (holotype, K!; isotype, NH not seen). Plants mainly 15-25 cm high, erect, un- base, with woody tunics soon breaking into elliptic segments tapering above into short points. Leaves 3. occasionally 4, the lower 2 basal and longest. reaching to about the middle of the stem, the upper one (or two) leaves smaller, 4.5-6 mm long, insert- ed in the lower third to middle of the stem, sheath- ing for most of their length, with a short free tip. the blades + linear, 2-3 mm wide, firm and erect, the midrib and margins slightly raised. Spike lax, mostly 2—3-flowered; bracts 12-15 mm long, green, the inner bracts about two-thirds as long as the outer, translucent with 2 green keels, shortly forked at the tip. Flowers bright mauve-pink, pale yellow in the mouth of the tube; perianth tube 6-9 mm long. cylindrical, expanded near the mouth; tepals X 3.5-5 mm, acute. Fil- mm, ide on the tepals spreading. elliptic, 10-1 aments erect, ca. « above the mouth of the tube, decurrent; anthers di- verging, ca. 5 mm long, shortly tailed, yellow, pol- len yellow. Ovary oblong, ca. 2.5 mm long; style branches reaching to the anther apices in bud, ca. 10 mm long, laxly spreading in the open flower. Capsules and seeds unknown. Flowering. August to October. Distribution. South Africa, KwaZulu-Natal and coastal Transkei in the sandstone belt from Durban to Port St. Johns, possibly also in Zululand near Eshowe, in marshy grassland, vlei edges, anc stream banks. A full description of Hesperantha modesta is pre- sented here because it was included in H. baurii by Hilliard and Burtt (1986) and no complete de- scription is available. It is difficult to assess the immediate relationships of this spring-flowering species for its unremarkable morphology suggests no particular affinity, except a general one to the H. baurii complex of eastern southern African grassland species. It can, however, readily be rec- ognized by the presence of just three or sometimes four leaves, the lower two basal and with long blades, and the upper one or two inserted on the stem and short and largely sheathing. The leaf lades have a fairly soft texture and slightly thick- ened margins and midrib. Plants flower in late win- ter and spring. Other species in the H. baurii alli- ance flower in the summer and typically have four, or sometimes five leaves, usually three of them bas- al or subbasal. An exception, H. baurii subsp. for- mosa, has either three or four leaves, but this is a high Drakensberg plant that flowers in January and February and has a large, deep pink perianth. The flowers of H. modesta are relatively large, in the middle of the range found in H. baurii, but judging Annals of the Missouri Botanical Garden from the available specimens it appears unusual in having only two to four flowers per spike. So few flowers per spike, even when associated with a rel- atively robust plant body, the unusual leaf number, and late winter to spring flowering make it clear that these plants should be recognized as a separate species, when the criteria for distinguishing species in Hesperantha are followed. There is confusion about the type locality of Hes- perantha modesta for the isotype at the KwaZulu- Natal Herbarium has the locality, Umlaas location, whereas the sheet at K with the same number and The lat- ter, in the interior in the Vryheid District, seems date is purportedly from the Bevaan river. unlikely for any spring-flowering Hesperantha bear- ing foliage leaves at flowering time. The discrep- ancy was noted by Hilliard and Burtt, who were not able to resolve this conflict after examining Wood's collecting registers. Hesperantha modesta was pro- visionally included in А. baurii by Hilliard and Burtt, although they discussed it and a few collec- tions of like plants separately. 1 concur with Hilliard and Burtt's observation that Hesperantha subexserta (Baker, 1896), based on a Medley Wood collection from Botha's Hill in the sandstone hills between Pietermaritzburg and Durban, flowering in October, may also belong here. These plants have two basal leaves, and one or two cauline, sheathing leaves of the soft texture typical of H. modesta, but the spike of one of the plants of the type collection has up to eight flowers and a short lateral branch. Other specimens of H. modesta have two or three, or at most four flowers per spike. Plants resembling H. subexserta should be sought again at the type locality, no great dis- = ance from Umlaas Location where the type collec- tion of H. modesta was most likely made. A collection from Eshowe, Zululand (Lawn 1179. NH), well to the north of the recorded range of Hesperantha | modesta, is provisionally included here. The rather poor condition and crowded mounting of the specimens makes them difficult to identify for certain, but the flowering time, October, and marshy habitat, suggest this species. Additional specimens. SOUTH AFRICA. KwaZulu- Natal: 28.31 (Nkandhla) Reservoir Marsh, Eshowe (CD), 3 „Оер 1949, Lawn 1179 (NH). 31.30 (Port Edward) Mtam- Reserve, marsh at Etheldale (AA), 11 Oct. 1986, Goldblatt 7897 (E, MO). Eastern Cape: 31.29 (Port St. Johns) Ntsubane Forest ecu near Fraser falls, seepage eie in pay sand (AC), 22 Aug. 1976, Venter & Vorster 39 PRE); near Magwa, jur malian of Magwa Fe а а (ВС), 23 Aug. 1984, Balkwill, Man- ning & Ge ottliffe Norris 1914 (NU); 3 mi. inland from Port Grosvenor, in a bog (BD), 23 Aug. 1969, Strey 8894 (PRE). 50. Hesperantha umbricola Goldblatt, S. Afri- can J. Bot. 53: 459. 1987. TYPE: wad. Near Mbabane, among rocks, 21 Feb. 1982, P. Goldblatt 6610 (holotype, MO!; isotypes, E!, K!, NBG!, NU!, PRE!, S!). The tiny white flowers of Hesperantha umbricola, only ca. 11 mm long, with a perianth tube ca. 4—5 mm long, tepals 5-6 mm long, and anthers 2-3 mm long, make the species unmistakable in the genus. Like other white-flowered species of Hesperantha in eastern southern Africa, flowering is diurnal. There appear to be no records of the species other than the type collection. I suspect that Н. umbricola is allied to the white-flowered H. inconspicua, but its much smaller flowers that do not fade pink, and narrow, soft-textured leaves seem to confirm that it is indeed a separate species. 51. Hesperantha gracilis Baker, Handbk. Iri- deae 149. 1892. TYPE: South Africa. Kwa- Zulu-Natal: base of perpendicular rocks at Is- angwaan, Apr., J. M. Wood 923 (holotype, K!). Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 422. 1986. 52. Hesperantha pubinervia Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 419. 1986. TYPE: South Africa: KwaZulu-Natal, Royal Natal National Park, Mont-aux-Sources, Feb. 1927, (holotype, BOL!). Prescott-Decie s.n. Last revisionary account: Hilliard & Burtt, Notes Hoy. Bot. Gard. Edinburgh 43: 419. 1986 Plants 12-30 em high, erect, unbranched. Corm conic, 7-9 mm diam. near the base, with woody tunics soon breaking into elliptic segments tapering above into short points. Leaves 4, the lower 2 basal, the upper smaller and partly to entirely sheathing, the blade + linear, 2.5—3.5 mm wide, + erect, with scabrid hairs on the margins, midrib, and second- ary veins, the midrib prominently thickened, the margins less so. Spike 1—3(—4)-flowered; bracts 20— 24 mm long. green. Flowers dull salmon pink, pale yellow-green in the mouth of the tube edged in darker salmon; perianth tube 20-25 mm long, un- usually slender, the hollow interior filled by the style and without nec vd tepals spreading at right angles to the tube, 14—18 X 3.5 mm, subacute. Filaments erect, ca. 5 mm long; anthers erect and + contiguous, ca. 4.5 mm long, pollen yellow. Ova- ry narrowly ovoid, ca. 4 mm long; style branches remaining suberect in the open flower, ca. 6 mm long, reaching to about the upper third of the an- Volume 90, Number 3 2003 Golablatt 427 Hesperantha Review thers and emerging between them. Capsules oblong, (10-12-14 mm long; seeds angular-prismatic, ca. 1.2 X 1.0, the edges forming membranous ridges. Flowering. February and March. Distribution. South Africa and probably Leso- tho, in the high northern Drakensberg on the slopes below The Sentinel and on the Mont-aux-Sources plateau, in rocky grassland. Growing along a well-used path to The Sentinel and Mont-aux-Sources in eastern Free State and adjacent KwaZulu-Natal, Hesperantha pubinervia nevertheless seems to have been overlooked there, which makes it appear that this area of the Drak- ensberg is poorly collected. My observations in gen- eral confirm the original description made from just one gathering. However, one feature, the length of the style branches, is not consistent with the pro- tologue, for plants on the slopes of The Sentinel have ascending style branches shorter than the sta- mens and about 6 mm long, less than the 11 mm length recorded by Hilliard and Burtt. The branches reach to the upper third of the anthers style and in one flower, perhaps not fully developed. only to the anther bases. The tepals are an unusual salmon pink, unique in the genus. The color con- trasts markedly with the deep pink to magenta flow- ers of sympatric and co-blooming Н. baurii and Н. scopulosa, and with the pale pink flowers of H. brer- ist yla. The capsules, not previously described, are ob- long and about as long as the bracts, mostly 12-14 mm long. The numerous seeds are angular (pris- matic) and about 1 mm long. The specimens on which the amplified description are based are cited below. Additional specimens. SOUTH usb Free State: of The Sentinel ail n ladders to "gel Falls, 5 ae 2002 (f, Goldblatt & Pole 11977 (MO, NBG, P 53. Hesperantha pulchra Baker, Handbk. lri- deae 150. 1892. TYPE: South Africa. Eastern Cape: Transkei, Baziya Mountain, Apr., L. К. Baur 159 (holotype, K!; isotype, B!). Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 413. 1986. Extending from the Amatola Mountains of the Eastern Cape north through the Transkei to Naude’s ek, the late-flowering Hesperantha pulchra has until now seemed fairly well understood. It is read- ily recognized by the bright pink flowers, relatively short filaments, 3—6 mm long, and perianth tube of according to Hilliard & Burtt, KwaZulu-Natal, north of the Tugela River, which Hilliard and Burtt mentioned as like H. pulchra but with a tube ca. 15 mm long (Gerstner 6713, PRE) and ca. 12 mm (Fakude 3, NH), fit uncomfortably in this otherwise Eastern Cape species. We need to know more about the Zululand plants, which could perhaps be accommodated in H. baurii equally well except for the late flowering, in April and May. I hesitate to treat these populations as representing an undescribed taxon given our current knowledge about H. pulchra. An interesting collection from Mt. Thomas, near Stutterheim in the Amatola Mountains (McMaster s.n., NBG), collected in later flower on 11 February, with ripe capsules at the base of the spike. resem- bles Hesperantha pulchra except for the early flow- ering. Plants have narrowly elliptic capsules 25 mm long that contain large seeds, up to 4 mm long with the wings at either end ca. 1 mm long and the seed body ca. 2 mm long. Capsules of H. pulchra are described by Hilliard and Burtt as 10—17 mm long with seeds ca. 1.25 mm in diameter, with weakly The Mt. Thomas or strongly developed wings. plants may represent a novelty. Echoing Hilliard and Burtt's remark regarding Hesperantha pulchra, we need to know more about both the Zululand and Mt. Thomas plants. Addi- tional collections, especially in fruit, will be helpful in understanding the range of variation in this spe- cies. 54. Hesperantha woodii Baker, Handbk. Irideae 150. 1892. TYPE: South Africa. KwaZulu-Na- tal: Richmond district, Peak of Byrne, Apr. or May 1883, J. M. Wood 1868 (isotypes, K!, NH!). Figure 5A. e R. C. Foster, Contr. Gray Herb. 1‏ و Hilliard & Burtt, N ES 43: 426 TYPE: sg Africa Lesotho, valley above Buffalo River falls, Mar. 1904, E. E. Galpin 6856 (lec- totype, م‎ шей һеге, BOLI; ie B!, GRA not seen, K!, NH not seen, PRE!, SAM!). Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 413. 1986 Plants 20-45 cm high, erect, unbranched. Corm conic, 8-10 mm diam. near the base, with woody tunics soon breaking into triangular segments ta- pering above into short points. Leaves usually 4, the lower 2 basal, longest, reaching to between the middle of the stem and the apex of the spike, the upper smaller and partly to entirely sheathing, the 428 Annals of the Missouri Botanical Garden Figure 5. —A. T! Flowers of du с (Goldblatt & Ро barely curved perianth tube and lar blades + linear, 2-3 mm wide, firm and erect, the midrib prominently raised, the margins slightly thickened. Spike mostly 2—4-flowered; bracts 22— 30 mm long, green, becoming dry and brownish above. Flowers bright mauve-pink, pale yellow in the mouth of the tube; perianth tube (13—)22-38 mm long, cylindrical, slightly curving toward the apex, expanded near the mouth, with nectar in the base; tepals weakly ascending, elliptic, 23-27 Xx 6— 7.5 mm, acute. Filaments ascending, 6—8 mm long; anthers diverging, 9-10 mm long, tailed in the low- er 2 mm, yellow, pollen yellow. Ovary ellipsoid, 6—8 mm long; style branches 18-20 mm long. longer than the stamens, laxly spreading. Capsules oblong, 16-18 mm long; seeds unknown. Flowering. Distribution. rica February and March. Lesotho and South Af- 1 KwaZulu-Natal and the Eastern Cape, on Southern | se ы. and rocks often near streams and in montane habitats. Living plants of a long-tubed Hesperantha (Fig. 5A) in the southern Drakensberg near Naude's Nek rs of 2 وی‎ жыт erar, E ч ш 11054, MO, NBG, from Naude's Nek). —B. 1 The Sentinel). Compare suberect flowers with ge, оын ond ы. y T w ооойй with the secund flower with strongly curved pe rianth tube and darkly colored anthers of H. grandiflora. in Eastern Cape Province proved difficult to iden- tify using Hilliard and Burtt’s (1986) key, which led me first to H. grandiflora, a species with a curved, elongate perianth tube (26-)33-55 mm long. and zygomorphic flowers with unilateral stamens and style branches. The Naude’s Nek plants, however, had actinomorphic flowers with symmetrically dis- posed stamens and style branches, but a weakly curved perianth tube. They seemed to closely match H. galpinii, a species regarded as a synonym of H. grandiflora by Hilliard and Burtt, and careful examination of the type material shows that H. gal- pinii probably does not have a zygomorphic flower and nor do a few similar specimens cited under H. grandiflora. The type locality of H. galpinii cannot easily be revisited to examine plants there, and the question of stamen and style branch orientation cannot be readily resolved. Subsequent comparison of several collections matching Hesperantha galpinii led me to H. woodii. It became clear that specimens such as McClean 763 (PRE), which have flowers with a perianth tube 26—32 mm long, assigned by Hilliard and Burtt to Volume 90, Number 3 2003 oldblatt 429 Hesperantha Review H. woodii, closely match Н. galpinii and that this referred by Hilliard and Burtt. Although the di- mensions of the perianth tube in H. woodii are 14— 28 mm according to Hilliard and Burtt’s circum- scription of the species, the majority of specimens they cited have a tube exceeding 20 mm and I have seen none with a tube less than 18 mm long. Peri- anth tube length in the type collection is ca. 20 mm. The addition of H. galpinii and some longer- tubed plants from Naude's Nek to Н. woodii thus leaves the definition of that species little changed, although the upper extreme of the perianth tube becomes 38 mm. Hesperantha woodii can be dis- tinguished from H. grandiflora by tepal. stamen, and style branch orientation, rather than by peri- anth tube length, which overlaps too much to be 5). Hes- perantha grandiflora has a perianth tube (26—)33— useful in comparing the two species (Fig. 55 mm long, strongly curved at the apex, vertically oriented tepals, and unilateral, declinate stamens and style branches. The geographic range of Hesperantha grandiflora extends from The Sentinel in the north through the Drakensberg to Mt. Currie (the type locality) in the south, and to Naude's Nek and Barkly Pass in the west. Thus both H. grandiflora and H. woodii occur in the Naude’s Nek area and in the mountains above Kokstad. The transfer of specimens matching H. galpinii to H. woodii extends the geographical range of the species very little, for H. woodii has already been recorded at Mhlahlane, near Umtata in the Eastern Cape, although not from southern Lesotho. Examination of collections at the herbaria at Kew and Zurich has revealed an interesting historical record for Hesperantha woodii, Drége 4540, which must have been gathered in 1832 when J. F. Drege traveled overland from Grahamstown to Port Natal (Gunn & Codd, 1981). Thus, Drege may be cred- ited with the discovery of the species, although Wood's and he is commemorated in the name of the spe- specimens, gathered in 1883, are the type cies. Additional specimens. LESOTHO. 28.30 (Matatiele) summit slopes of Maquaba peak, near Quacha’s Nek (BA),13 Mar. 1936, Tengi 14247 (K). SOUTH AFRICA. KwaZulu-Natal: 29.29 (Underbe Cobham Forest Sta tion, Sipongweni Caves (CC), 13 Apr. 1972, Hilliard 5509 K, NU); river banks, Underberg, Ма; 1938, McClean 753 (K, PRE); Polela, Glengarif, Marwaqa, 26 Mar. 1977, Ren- nie 815 (NU); Sunset Farm, Polela District, 17 Feb. 1979, Rennie 1006 (NU). иа ips 30.28 (Matatiele) 1-2 km W of Naude’s Nek, F 1999, Goldblatt & Man- ning 11054 (K, MO, NBG, PRE). 55. Hesperantha stenosiphon Goldblatt, sp. nov. TYPE: South Africa. Eastern Cape: Stutterheim district, Moonstone farm, steep grassy slope, 1120 m, 20 Mar. 2001, C. McMaster s.n. (ho- lotype, NBG!; isotypes, MO!, PRE!). Figure 6. Plantae 25—50 cm altae eramosae, cormo globoso ca. 12 mm diam., foliis 4 inferioribus duabus basalibus, en- m 2—3 mm latis, spica (2-)4—3(-12)-flora, brac- s 18—24 mm longis, floribus roseis albescentibus prope orem tubi, perianthii tubo 45—60 mm longo recto, tepalis 8-21 X 7-8 mm, filamentis ca. 3 mm Su antheris 9-10 mm longis, ramis styli ca. 8.5 mm lon Plants 25—50 cm high, erect, unbranched. Corm globose, ca. 12 mm diam., with woody tunics soon breaking into segments tapering above into short points. Leaves usually 4, the lower 2 basal, longest, reaching to between the middle of the stem and the apex of the spike, the upper smaller and partly to entirely sheathing, the blades + 2-3 mm wide, firm and erect, the midrib and margins slight- ly thickened. Spike (2-4—9(-12)-flowered; bracts 18-24 mm long, green, becoming dry and brownish linear, near the tips. Flowers bright pink, whitish in the mouth of the tube; perianth tube 45—60 mm long, cylindrical, slightly curved or straight, barely ex- panded near the mouth; tepals ascending, elliptic, 18-21 x 7-8 mm, 47 ca. 3 mm long; anthers diverging, 9-10 mm long, subacute. Filaments suberect, dark brown to blackish, pollen yellow. Ovary ovoid, ca. 3.5 mm long; style branches ca. 8.5 mm long. in the closed flower reaching to just below the an- ther apices. Capsules and seeds unknown. Flowering. March and April. Eastern Cape, Cathcart district, among rocks partly shaded by bush and small trees Distribution. or in rocky grassland on hill tops. the striking, long-tubed Hesperantha stenosiphon was discovered Evidently unknown until 2000, by the naturalist, Cameron McMaster, in the Cat cart area of Eastern Cape Province, South Africa. 'a. The plant is probably most closely allied to the long-tubed H. grandiflora and H. woodii and shares with them the linear leaves with slightly to mod- erately thickened margins and midribs as well as the pink perianth with an elongate tube, usually more than 20 mm long. Hesperantha stenosiphon is readily recognized by the symmetrically arranged stamens with unusually short filaments, only about 3 mm long, long blackish anthers, and spike of (2 to)4 to 9 flowers. Hesperantha grandiflora has flow- ers that face to the side with vertically oriented tepals and unilateral, downcurving stamens and style branches, filaments 8-14 mm long, and brown anthers and pollen, while H. woodii has similarly 430 Annals of the Missouri Botanical Garden SN / | My FON | Figure 6. Morphology and floral details of Hesperantha stenosiphon. Scale bar 1 ст. Drawn by John Manning from photographs and pressed plants (Goldblatt & Porter 12005 (K, NBG, MO, PRE) Volume 90, Number 3 2003 Goldblatt Hesperantha Review large, but nearly upright flowers, symmetrically dis- posed stamens with filaments 6-8 mm long, and yellow anthers and pollen. The two latter species have flowers with a perianth tube (18-)22-55 mm long, whereas H. stenosiphon has a tube 45—60 mm long. The elongate perianth tube suggests that the flowers are pollinated by the long-proboscid fly, Prosoeca ganglbaurii, which also pollinates H. grandiflora and H. woodii (Goldblatt & Manning, 2000; Goldblatt et al., locality where plants were common and in full in press). At the Bombazi bloom, however, I found no long-proboscid flies, but flowers were instead visited by honey bees, which collected pollen after failing to reach the nectar contained in the lower part of the perianth tube, well beyond the reach of their tongues. Common at the two sites where it is recorded, Hesperantha stenosiphon is nevertheless rare, and appears to be restricted to the hills east of the main Amatola range that extend toward the valley of the Great Kei river. Plants are mostly confined to dol- erite outcrops on the upper slopes and summits of these hills where they seem to prefer partly shaded situations close to shrubs and small trees. Paratypes. SOUTH AFRICA. Eastern Cape: 33.26 (Stutterheim) Cathcart district, ca. 50 km NE of Stutter- heim, Bombazi farm, Mar. 2001, McMaster s.n. (NBG), Mar. pes нат & Porter 12005 (К, NBG, MO, PRE); Bolo dis Moonstone farm, on steep grassy slopes, 1076 m, 27 Feb, 2002, McMaster s.n. (MO). De 56. Hesperantha coccinea (Backh. & Harv.) Goldblatt & J. C. Manning, Novon 6: 263. 96. Schizostylis coccinea Backh. & Harv Curtis’s Bot. Mag. . 5422. 1864. TYPE: South Africa. asiem о without precise locality, Curtis's Bot. Mag. 90: pl. 5422. 1864. With a flower structure exactly like that of any Hesperantha species, and virtually identical, except in the red color, to that of long-tubed species like H. woodii, H. coccinea seems well placed in the genus to which it was transferred in 1996 (Gold- blatt & Manning, 1996). Hesperantha coccinea was long regarded as the sole species of a separate ge- nus, Schizostylis, distinguished in subfamily Cro- coideae by its rhizomatous rootstock. A plant of stream banks and marshes, H. coccinea is believed to have lost its corm because this xeromorphic fea- ture is not adaptive in such mesic conditions. Cu- riously, the aerial leaf axils of some populations produce a small corm of the asymmetrical shape typical of Hesperantha. There are two color morphs of H. coccinea, the typical, and more widespread red form, which extends from the Amatola Moun- tains in Eastern Cape Province, South Africa, to Zimbabwe, and a pink form local in the northern Drakensberg and Witwatersrand in Gauteng Prov- ince. The name Schizostylis pauciflora has been ap- plied to plants from Witpoortjie in the western Wit- watersrand that are depauperate and have spikes of few flowers. Unusual for Hesperantha, the com- mon red-flowered form of H. coccinea is pollinated by a guild of large butterflies of the families Papi- lionidae (Papilio spp.) and Satyridae (Aeropetes tul- baghia) (Goldblatt et al., in press). Seeds of Hesperantha coccinea, not previously described, are unusual in the genus. Approximately 2 X 1.2 mm, the more or less prismatic (segmental) seeds have a loose, translucent, light brown seed coat with a smooth outline, but with slightly devel- oped ridges on the angles of the segments. Within the translucent coat a small, more or less spherical seed body, ca. 1 mm in diameter, is evident. The surface cells of the seed coat are aligned in straight files and have a domed (colliculate) outer wall and appear empty. This seed is somewhat aerodynamic and will float for some time before becoming wa- terlogged. The seed is evidently adapted for dis- persal by water, a not unexpected adaptation in this semi-aquatic plant. Other species of Hesperantha have a seed coat that closely envelops the seed body, but other features conform to the pattern de- scribed for the genus (Goldblatt & Wagner, 1984). 57. Hesperantha grandiflora G. J. Lewis, J. 5. African Bot. 7: 30. 1941. Acidanthera tysonü Baker, Handbk. Irideae 187. 1892, non Н. ty- sonii Baker (1892), — H. radiata (Jacq.) Ker Gawl. TYPE: South Africa. KwaZulu-Natal: waterfall near Mt. Currie, Apr. 1883, W. Tyson 1151 (= Herb. Norm. Austro-Afr. 895) (holo- type, K!; isotypes, BOL!, NBG!, PRE!). Figure 5B Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 410. 1986 One of the most distinctive of the species of pink-(or red-)flowered Hesperantha of eastern southern Africa that have flowers with an elongate perianth tube at least 15 mm long, the high Drak- ensberg H. grandiflora is easily recognized by its flowers with a perianth tube (26-)33—55 mm long, curved outward near the apex and tepals held more or less vertically (Fig. 5B). It also has unilateral stamens and style branches that are declinate, thus arching downward above the lower (abaxial or an- terior) tepal. A feature of H. grandiflora not re- corded in the literature is that it has reddish brown anthers and pollen, whereas most other species have yellow or cream to whitish pollen. The darkly 432 Annals of the Missouri Botanical Garden colored pollen is still evident in the type collection after more than 100 years, and I have confirmed this feature in living plants seen at Sani Pass, Gi- ants Castle Pass, on The Sentinel trail on the Free State-KwaZulu-Natal border, and at Barkly Pass in 5B). Plants from the latter two localities constitute range extensions to Eastern Cape Province (Fig. the north and west of the range of H. grandiflora and represent the first record of the species from the Free State. Seeds of this population conform to 1986): they are 1.5-2 mm long and have a small wing at the description given by Hilliard and Burtt ( opposite ends of the globose seed body ca. 0.8 mm long (the seed is described as 1—1.5 mm diam. by Hilliard & Burtt). Examination of living plants collected in the southern Drakensberg near Naude's Nek Pass has shown that Hesperantha grandiflora as circum- scribed by Hilliard and Burtt included plants with more or less upright flowers, a weakly curved tube and symmetrically disposed, ascending stamens, quite different from H. grandiflora. As explained above, these plants are better referred to H. woodii as is H. galpinii, which was treated as a synonym of H. grandiflora by Hilliard and Burtt. A collection at the Kew Herbarium made by J F. Drege in 1832 is almost certainly this species, and represents the earliest record of Hesperantha grandiflora. The specimen is identified as “Gladi- olus spilanthus," a later synonym for the Western Cape G. gracilis, but has no specific locality infor- mation and it does not appear to be listed in the report of Drége’s travels (Meyer, 1843). SS Si imens. SOUTH AFRICA: Free State- -Natal: 28.28 D hen grassy e on road to The Se нон trail alor —KwaZulu-Natal = VOT- der (DB), 4 Mar . 2 002, bene & Porter ee 31.27 (Lady Frere) Barkly Ap moist gully on Cave S: пее slope near top of pass (BB), 6 Mar. 2002, Goldblatt & [n 11995B (MO, NBG). 58. Hesperantha huttonii (Baker) Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 40: 278. 1982. Acidanthera huttonii Baker, J. Bot. 14: 339. 1876. TYPE: South Africa. Eastern Cape: Stockenstrom Division, Katberg, date unknown, H. Hutton s.n. (holotype, K!). Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 415. 1986 A species of shady rock outcrops and forest mar- gins, Hesperantha huttonii has long been thought to be restricted to the Amatola Mountains between Stutterheim and Adelaide in Eastern Cape Prov- ince, South Africa. Apart from the long-tubed, pale pink flowers with large yellow anthers, 7-10 mm long, it can be recognized by the fairly broad, flac- cid, drooping leaves, suberect to drooping stem, and the presence of a cormlet in the axils of the lower leaves (Hilliard & Burtt, 1986). Plants from coastal Transkei (Flanagan 2514, PRE; Cloete 1661, NH), some 250 km to the east, best match the species. However, the Flanagan collection ap- pears to lack axillary cormlets and has a perianth tube ca. 23 mm long, while the Cloete collection has flowers with a tube 15-17 mm long. Hesper- antha huttonii typically has a tube 30-39 mm long with a lower limit 21 mm. Provisionally, the Trans- kei plants must be included in H. Auttonii until more information becomes available. The seeds of this species are distinctive. Borne in capsules 12-20 mm long, they are approximately 2.5-3.5 mm long and ca. 1.25 mm wide, and have a membranous wing at either end, the wings each about half as long as the seed body, which is about 1.2-1.5 X 1 mm (Goldblatt & Porter 12011, MO, NBG, from the Kologha Forest, near Stutterheim). Slightly smaller seed dimensions, ca. 1.5 X mm, provided by Hilliard and Burtt (1986), prob- ably indicate variation in seed size across popu tions of the species. The seeds recall those of Glad- iolus, which have a broad circumferential wing (Goldblatt & Manning, 1998). The long-tubed flowers of Hesperantha huttonii are pollinated by the long-proboscid nemestrinid fly, Stenobasipteron wiedmannii, which has been re- corded visiting the species in the Kologha Forest (Goldblatt et al., 19-23 mm long, thus well suited to acquire pollen in press). This fly has a proboscis loads on its thorax as it probes the perianth tubes of H. huttonii, which are 23-35 mm long at this site and contain nectar at the base. Indirect evi- dence for the pollination of H. huttonii by the same y has been recorded by Potgieter and Edwards (pers. comm.) who found pollen of H. huttonii on a fly caught visiting Plectranthus ciliatus E. Mey. ex Benth. as well as Plectranthus pollen on Hesper- antha anthers, presumably carried there by Steno- basipteron, an important pollinator of Plectranthus species. Additional specimens. SOUTH AFRICA. Eastern Cape: 31.29 (Port St. n eng rau Lupatana, sand- HUN c d face above the river, ca. 15 m, in mats of moss ) Dec. 1991, Cloete 1661 (NH); Port St. Johns, Mun 2514 (PRE). EE 59. Hesperantha hutchingsiae Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 414 1986. TYPE: South Africa. Eastern Cape: Transkei, Mhlahlane Forest Station, Mjika, 21 Mar. 1985, A. Hutchings & Plumstead 1621 (holotype, E not seen). Volume 90, Number 3 3 Goldblatt Hesperantha Review 433 When first described by Hilliard and Burtt (1986). Hesperantha hutchingsiae was known only from the type collection from a marsh at Mhlahlane near Umtata. A second collection has come to hand from near Naude’s Nek some distance to the east. This collection, Strever 755, has three flowers per spike, and the flowers have a perianth tube 23 mm long. In the type collection the spike has only two flowers and the perianth tube is 21 mm long. The new record slightly expands the range of variation in the species. Notes on the Strever collection in- dicate that the anthers are dark purple, an unusual character in the genus, but recorded in H. gran- diflora and H. stenosiphon, both of which have large flowers with a substantially longer perianth tube. (26—)33—55 mm long in Ё. grandiflora and 45—60 mm long in H. stenosiphon. TE specimens. SOUTH AFRICA. Eastern Cape: 30.28 (Matatiele) valley of Lehana’s Pass, Mauden Nek. moist area on S slope below dolerite rocks (CA), 1 =” Mar. 1988, Strever 755 (NU). 60. hpc gris brevicaulis (Baker) G. J. Lew- . S. African Bot. 6: 30. 1941. Acidanthera beide Baker, Fl. Capensis 6: 132. 1896. ҮРЕ: South Africa. Barberton, Devil's Bridge, Makwonga Range, Mar. 1891, E. Е. Galpin 1252 (holotype, K!). Plants 10-25 ст high, the stem drooping, un- 7-9 mm diam. near the + S cm branched. Corm conic, base, tunics unknown. Leaves 4, occasionally 5 linear, 2—4 mm wide, the lower with fairly soft-tex- tured blades trailing distally, the midrib slightly raised, the uppermost leaf largely to entirely sheathing. Spike mostly 2—3(—4)-flowered; bracts green, soft-textured, the outer 20-28 mm long. Flowers dull mauve-pink, pale yellow in the mouth of the tube; perianth tube (18—)22—35 mm long. cy- lindrical, slightly curving toward the apex and ex- panded near the mouth, with nectar in the base; tepals spreading + at right angles to the tube, )-25 X 7-8 mm, subobtuse. Filaments ascending, 5-8 mm long; anthers diverging, 6-10 mm long, pollen жй Ovary 4—5 mm long; style branches axly spreading, 12-15(-20) mm long, alternating with the stamens and exceeding them in the closed flower. Capsules narrowly ovoid, 15-20 mm long; seeds unknown. Flowering. March to May, sometimes in Feb- Distribution. South Africa, Mpumalanga and Limpopo Provinces, along the eastern escarpment on steep rocks and cliffs, the corms growing in damp moss and in rock crevices. Although the type collection of Hesperantha brev- icaulis is from the mountains near Barberton in Mpumalanga Province, and has somewhat longer stamens than more recent collections from the Sa- bie-Graskop part of the Escarpment some 90 km to the north, there seems no difficulty in regarding all these collections as a single species, the only long-tubed, pink-flowered member of the genus from the northern provinces of South Africa. Hil- liard and Burtt (1986) concluded that it differed from all of the long-tubed species treated in their account of the KwaZulu-Natal, Lesotho, and East- ern Cape species of the genus, but Retief and Her- man (1997) did not include H. brevicaulis in their flora of the northern provinces of South Africa. Hes- perantha brevicaulis flowers relatively late in the season and is seldom seen before the last week of March. Like other species of the genus with similar, long-tubed flowers, it appears to be adapted for pol- lination by long-proboscid flies. The nemestrinid fly Stenobasipteron wiedmannii has been recorded vis- iting the species at God's Window near Graskop (Goldblatt & Manning, 2000). Plants from the Wolkberg in Limpopo Province (Davidson 3153, J; Goldblatt & Porter 11953, МО, G) seem at first to represent this species, but they flower earlier in the season, beginning to bloom as early as the middle of February. They also have flowers with a somewhat shorter perianth tube, 13-16 mm, compared with (18—)22-28 mm for the Sabie-Graskop populations and ca. 35 mm in the type. from Barberton. Provisionally I include the Wolkberg populations in Hesperantha brevicaulis. Plants from Serala have the following features that seem to differ significantly from populations to the south: bracts (11—)15—18 mm long, the inner about two thirds as long; flowers with a periant tube 13— 16 mm long; tepals 20-22 х 7-9 mn ; filaments (4—)5—6 mm; anthers 5—6.5 mm long; ovary ca. mm long; style branches 15-18 mm long. The shorter perianth tube (and associated shorter bracts) suggests a less specialized pollination sys- tem, perhaps including bees and nemestrinid flies with somewhat shorter probosces than Stenobasip- teron wiedmannii, the proboscis of which is up to 25 mm long. Equally puzzling is a recent collection from Mt. Prospect, near Lydenburg (Burrows 7309, BKH), collected in early flower in mid February. This plant has long-tubed mauve flowers with a tube 30- 32 mm long, tepals ca. 18 mm long, short filaments ca. 3 mm long, and anthers ca. 6 mm long. The four leaves have well developed blades and there is no sign of a largely sheathing upper leaf. The blades have a firm texture, an apparently dropping Annals of the Missouri Botanical Garden The plant thus fits poorly in Hesperantha brevicaulis be- flowering stem, and a spike of five flowers. cause of the firm leaves, absence of a sheathing upper leaf, spike with more than four flowers, short Additional material is needed before a decision can be made filaments, and relatively short anthers. on its status. Additional e SOUTH AFRICA. онова 23.30 (Tzaneen) Wolkberg Mountain, below cliffs (CC), Apr., ere 3154 (J); Wolkberg, damp cliffs below Se r- ala Peak, 23 Feb. 2002, Goldblatt & Porter 11953 (К, MO, NBC, PRE). Mpumalanga: 24.30 (Pilgrim’s Rest) acMac Falls, in grass on sheer cliffs (DD), 14 Mar. 1959, Germishuizen — (PRE); God's Window and The Pin- nacle, 20 Apr. 1969, . 1 s. S (PRE); rock oute rops near God's Window, ‚ Goldblatt 72 (J) Apr. 1967, Goldhlat A o Kluge 2505 D Mons Ven- ter dies ч PU, де .30 (Lydenburg) Mt. Prospect, slopes eam, 2010 m, 18 Feb. 2001, Bares 7309 Wn. Ви Iskloof Nature Reserve, crev- ices in cliffs (BC), 2 May 1988, Burrows 4070 (BKH, J). ~ 61. Hesperantha curvula Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 416. 1986. TYPE: South Af KwaZulu-Natal: Underberg, Bushman’s Nek, Thamathu Pass, 5 ‘eb. 1976, O. M. Hilliard & B. L. Burtt 8981 (holotype, NU!; isotypes, E not seen, MO!). Africa. Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 416. 1986. 62. Hesperantha scopulosa Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 417. 1986. TYPE: South Africa. KwaZulu-Natal: Underberg, Bamboo Mountain, 8 Mar. 1977, O. M. Hilliard & B. L. Burtt 10074 (holotype, NU; isotype, E not seen). Last revisionary account: Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 417 Fairly common on wet basalt cliffs and rocks along the approach to The Sentinel in Free State Province, and known from several localities on ba- salt and sandstone in KwaZulu-Natal, Hesperantha scopulosa belongs to a guild of plant species with long-tubed pink flowers that are pollinated by the long-proboscid fly, Prosoeca ganglbaurii (Goldblatt 1999, 2000). Interestingly, although flowers with a long perianth tube usually secrete & Manning, ample amounts of nectar, which is retained in the tube, H. scopulosa seems to be deceptive. The tube is very narrow, and I have not been able to detect any nectar in the tube in two populations I exam- ined. Successful pollination of these flowers must depend on their similarity to those of nectar-pro- ducing species growing nearby, among them Glad- iolus microcarpus G. J. Lewis, Hesperantha gran- diflora, and Zaluzianskya microsiphon (Kuntze) K. Schum. (Scrophulariaceae). Seeds of Hesperantha scopulosa, collected on The Pudding below The Sentinel (Goldblatt & Manning 9856, NBG), are narrowly ovoid-oblong, ca. 0.4 mm, and have a long, irregularly twisted per- sistent funicle several times longer than the seed body. The chalazal end is flattened and has a small membranous flap of tissue, perhaps a vestigial wing. These seeds conform to the description of Hilliard and Burtt (1986) who called them oblong- elliptic in shape, 1.25 X 0.5 mm, with a narrow wing on one end, and with a long pale funicle. The only other comparable seeds in the genus may be those of H. gracilis, which Hilliard and Burtt de- scribed as narrowly obovoid and with a long slender funicle. П. Hesperantha sect. Hesperantha Plants with relatively large symmetrical, bell- shaped corms with a flat, horizontal or oblique base; older corm tunics remaining entire, not split- ting from the base. Spike with floral outer bract margins free to the base. Flowers variously colored, with a straight perianth tube Species 63-70. Restric ted to the southern Afri- can winter-rainfall zone. 63. Hesperantha falcata (L. f.) Ker Gawl., Ann. Bot. (Kónig & Sims) I: 225. 1804. /xia falcata L. f., Suppl. pl. 92. 1782. TYPE: South Africa. Western Cape: hills around Cape Town, with- out date, C. P. Thunberg s.n. (holotype, Herb. Thunberg 9052B, UPS!). киш: iet € Kew Bull. 1906: 26. 1906. ГҮР uth Africa. Cape: Olifants River sil Clanwilliam, Sep. 1894, A. фы 686 (lec- totype, designated by Foster (1948: 21), K!). Hesperantha Mas R. C. uet pd s Herb. 160: э TYPE: South Africa. Western Cape: Buls- oe is pn "es R. Schlechter 8378 (holotype, B!; Western ы BOL!, . K!, MO!, P!, PH!, PRE!, US!, /1). Last revisionary account: Goldblatt, Ј. S. African Bot. 50: 97. 19 Although both H. pentheri and H. trifolia were recognized by R. C. Foster (1948) in his preliminary account of Hesperantha, they were included in H. falcata in the 1984 revision of the genus for the winter-rainfall zone (Goldblatt, 1 ey un- doubtedly represent distinct н races and an argument can be made for their recognition as separate species or subspecies; they are treated as separate entities in my key to the winter-rainfall spe- Volume 90, Number 3 2003 Goldblatt Hesperantha Review cies. Plants matching H. trifolia almost always have only three leaves, and whether dwarfed by poor growing conditions or robust, they have relatively large white flowers, often only two or three per spike, always evenly spaced along the stem, and a pink to light reddish pigmentation on the reverse of the outer tepals, rather different from H. falcata, in which the flowers are somewhat crowded at the tip of the spike. The bracts of H. pentheri and H. trifolia are soft- textured or even membranous, thus unlike the firm green bracts of typical H. falcata, which have prom- inent veins, a reddish margin, and obtuse apex. Plants matching H. pentheri have deep cream to light yellow flowers, the outer tepals often flushed dull red to purple on the outside. Both flower relatively early, mostly in August (typical H. falcata flowers in Sep- tember or October), and are confined to the north- west Cape. in the Olifants River valley and sur- rounding mountains where typical H. falcata does not occur. The most northerly record of plants matching H. pentheri is from the Kobee Valley, north- east of Vanrhynsdorp (Goldblatt & Porter 11794, MO, NBG), not far from the southernmost popula- tions of H. pauciflora, which also have pale yellow flowers. The two can readily be distinguished by their different corms, those of H. pauciflora having spines radiating from the base. 64. Hesperantha sufflava Goldblatt, sp. nov. TYPE: South Africa. Western Cape: Malmes- bury, sandy gravel slopes in renosterveld, 14 Aug. 1999, P. Goldblatt & 1 Nénni 11087 (ho- lotype, NBG!; isotypes, K!, MO!, PRE!, WAG!). Figure 7. Plantae 8—15 cm altae prope basem saepe ramosae, cor- mo campanulato basi plano tunicis lignosis Ven arn leviter denticulatis 10—18 mm diam., foliis 3 omnibus bas- alibus vel summo subbasali, lanceolatis vel ШТ 2-7 mm latis, spica (2-vel)3-ad 7-flora flexuosa, floribus pal- lide flavis tepalis externis abaxialiter pallide hor tubo perianthii 12714(-16) mm longo recto, tepalis 7-9 (-10) X 4—5 mm, filamentis ca. 2 mm longis. antheris ca. ongis, ramis styli ca. 12 mm longis. in medio tubi perianthii divisis, apic es ie ae non attingentibus. Plants mostly 8—15 from near the base in the axil of the uppermost leaf. 10-18 tunics woody, the margins denticulate. cm high, often branching Corm bell-shaped with a flat base. mm diam., Leaves 3, all basal or the uppermost subbasal, lan- ceolate to falcate, the upper leaf partly sheathing, with a short free unifacial tip, 2-7 mm wide, slight- ly fleshy, the midrib not visibly thickened. Spike slightly flexuose, (2-)3—7-flowered; bracts 10-14 mm long, green, often red along the upper margins, diverging from the stem, the outer about as long as or slightly shorter than the stem. Flowers pale yel- low, the outer tepals flushed light brown on the out- side; perianth tube slender, straight, 12-14(-16) mm long; tepals subequal, ovate, slightly spooned, mm, spreading + at right angles to the tube when fully open after 15:00H. Filaments ca. 2 mm long, inserted at the base of the tepals; anthers ca. 4 mm long, erect, yellow. Ovary oblong, ca. 2.5 mm long; style branches ca. mm dividing in the middle of the tube, exserted for ca. ong, 7 mm and then weakly diverging, reaching to about the upper third of the anthers in bud. Capsules ob- — ong, -9 X ca. 4 mm; seeds + globose or weakly angled by pressure, ca. 1.3 mm long. Flowering. Late July and August. Distribution. South Africa, Western Cape, local in the Malmesbury District in renosterveld on sandy gravel. A member of section Hesperantha, H. sufflava at first appears to be intermediate between the com- mon, usually white-flowered H. falcata and the much rarer H. spicata. At the type locality H. spi- cata grows together with H. sufflava and is obvi- ously quite different in its secund spike of small, pure white flowers, distinctive in the slightly curved perianth tube 4—6 mm long and tiny tepals 4—7 mm long. Moreover, flowers of H. spicata open at about 18:30H when they release a strong, sweet, narcis- sus-like scent with a strong clove component. Flow- ers of H. sufflava are pale yellow and open at about 15:00H, and they have a quite different, slightly acrid, pyrethrum-like odor. Particularly unusual are the style branches, which divide in the middle of the perianth tube, and only the upper 7 mm are exserted. In nearly all Hesperantha species the style divides at the throat of the perianth tube and not within the narrow part of the tube. Superficially Hesperantha sufflava is not much like H. falcata for the flowers seem much smaller. The tepals, 7-10 mm long, are smaller than in most populations of H. falcata, but the perianth tube is substantially longer, usually 12-16 mm long, and is always longer than the tepals (Fig. 5). In H. fal- cata the tepals are usually 12-18 mm long, excep- tionally only 9-11 mm in southern Cape coastal populations, and the perianth tube is 4—9 mm long, thus usually shorter than the tepals. Apart from the flowers, H. sufflava is distinctive in always having only three leaves, all basal or the uppermost one subbasal, and the stem often has a branch produced from the axil of the upper leaf. Most populations of both H. spicata and H. falcata have three (or more) basal leaves and one, largely or entirely sheathing cauline leaf. 436 Annals of the Missouri Botanical Garden Paratypes. SOUTH. AFRICA. Western Саре; 33.18 the base of the tepals and in the throat (or they are (Cape Town) epis sandy gravel slopes in renos- terveld 1.7 km from town center on road to Tulbagh (DB). 12 Aug. 2000, Goldblatt & Vänni 11383 (K, MO PRE, WAG). ~ FUIT 65. Hesperantha cedarmontana Goldblatt, J. S. African Bot. 50: 106. 1984. TYPE: South Africa. Western (os Pakhuis Mts. W of L poldt's Grave, 27 Sep. 1981, P. е 5403 (holotype, MO!; isotypes, K!, NBG!, PRE). . African Last revisionary account: Goldblatt, J. S Bot. 50: 106. 1984. 66. Wap pauciflora (Baker) G. J. Lew . Fl. Pl. Africa 18: pl. 682. 1938. Thitonia pair Baker, Handbk. Irideae 193. 1892. YPE: South. Africa. Northern Cape: Nama- 1893, H. Bolus 9622 (lectotype, designated by Goldblatt (1984: 108), BOL!). een near Naries, Sep. Last revisionary account: Goldblatt, J. S. African Bot. 50: 108. 1984. The typically pink- to purple-flowered Hesper- antha pauciflora extends from northern Namaqua- land to the Bokkeveld Escarpment at the northern edge of the Cape Floristic Region of South Africa and is locally common at the southern end of its geographic range. Plants with pale vellow flowers from the extreme south of its range at Papkuilsfon- tein, south of Nieuwoudtville (e.g., Goldblatt 11102, O), are included here, expanding the range of variation in Н. pauciflora. They appear to differ in no other significant way from the pink-flowered populations and have identical bell-shaped corms with prominent radiating spines, spikes of only 2 to 4 flowers, and submembranous bracts, dry near the tips. The flowers also have the same daily open- ing and closing pattern. On warm days the tepals unfold after 13:00H and close again after 17:00H (Goldblatt et al., by a variety of bees and by hopliine beetles, both in press). The flowers are visited of which appear to be legitimate pollinators of the species. There are also significant differences between the Namaqualand and Bokkeveld Escarpment pop- ulations. Plants from Namaqualand have dark pink to Jui flowers with a white throat (Goldblatt, 984: fig. 4), stamens with relatively long filaments, 4—8 mm piel anthers 6.5-9 mm long, and the style branches reach the anther apices or exceed them by up to 2 mm. On the Bokkeveld Escarpment, about 100 km south of the nearest Namaqualand populations, the flowers are either uniformly pale to deep pink, or often have darker pigmentation at uniformly pale yellow). The stamens in these pop- ulations have filaments only 2-3 mm long while the anthers аге 7-9 mm long, and the style branches barely reach the anther tips or up to 2 mm below them. These differences appear to reflect popula- tion divergence due to isolation, and perhaps sig- nify a pollinator shift, which remains to be studied. 67. Hesperantha latifolia (Klatt) M. P. de Vos, African Bot. 40: 252. 1974. Syringodea latifolia Klatt, Abh. Naturf. Ges. Halle 15: 403. 1882. TYPE: South Africa. Northern Cape: Kamiesberg, Ellenboogsfontein, Sep. 1330, J. F Drége 2633 (lectotype, en by de Vos (1974: 252), В!; isotypes, P!, Last e account: Goldblatt, J. S. African Bot. 50: 111. Restricted to oe elevations in the Kamies- berg Mountains of central Namaqualand, this win- ter-rainfall zone species is relatively common in shallow soils overlying granite pavement. Plants are typically short, usually less than 5 cm high, but robust plants growing in rock crevices or through ow bushes may reach 15 cm. The dark red-purple flowers with a perianth tube 15-25 mm long are now known to be pollinated by the long-proboscid fly, Prosoeca peringueyi, which also pollinated sev- eral other long-tubed species with similarly colored flowers in Namaqualand, among them Babiana dre- gel Bak., Lapeirousia silenoides (Jacq.) Ker Gawl. (Iridaceae), and Pelargonium incrassatum (Andr.) Sims (Geraniaceae) (Goldblatt et al., 1995; Gold- blatt & Manning, 2000). 68. Hesperantha luticola Goldblatt, J. S. African Bot. 50: 113. 1984. TYPE: South Africa. North- em Cape: between Midddelpos and Calvinia, Farm Knechtsbank, 21 Aug. 1974, M. Е Thomp- son 2529 (holotype, STE!; isotype, PRE!). ast revisionary account: Goldblatt, J. S. African Bot. 50: 113. 1984. An aspect of the acaulescent Hesperantha luti- cola not known when the species was described is the fact that the perianth tube is virtually closed. The tube, 30—45 mm long. might be expected to be hollow and to contain nectar as it does in many other long-tubed species of the genus. The walls of the tube are, however, thick, and they closely en- velop the style leaving no internal cavity. The small amount of nectar produced by the flowers is forced into the upper, slightly wider part of the tube where it is accessible to insects with relatively short pro- bosces. The perianth tube thus appears to serve as Volume 90, Number 3 2003 Goldblatt Hesperantha Review Е. 3 ^ i» sate бү”, 5] РЫР) $2755 y mm Ls = ге Figure 7. Morphology and floral and capsule and seed details of H. sufflava. Drawn by John Manning from live plants (Goldblatt & Nünni 11087, MO, NBG). Scale bar 1 em; single flowers and seed much enlarged. a pseudopedicel, raising the tepals, stamens, and style branches above the basal cluster of leaves. The ovary remains close to or below ground level, and is thus protected from damage during its mat- uration. The ripe capsules are borne a short dis- tance above the ground. 69. Hesperantha spicata (Burm. f.) N. E. Br., Kew Bull. 1929: 136. 1929. Ixia spicata Burm. f., Prod. Pl. Cap. 1. 1768. TYPE: South Africa. Without precise locality or date, probably cul- tivated in Holland, N. L. Burman s.n. (holo- type, herb. Burman, G!). Figure 1C. 438 Annals of the Missouri Botanical Garden Last revisionary account: Goldblatt, J. S. African Bot. 50: 114. 1984. In my 1984 account of Hesperantha in the south- ern African winter-rainfall zone I treated Hesper- antha spicata as comprising three subspecies: spicata, with falcate basal leaves with crisped margins, subsp. graminifolia (Sweet) Gold- blatt, with narrow, erect leaves with plane margins, and subsp. fistulosa (Baker) Goldblatt, with terete leaves. While the two former subspecies seem sat- subsp. isfactorily defined, a new record of subspecies fis- tulosa from the farm Joostenbergkloof, west of Paarl, expands our understanding of this plant, pre- viously reported only from wet flats in the Porter- ville district, well to the north of Paarl. Plants at the Joostenbergkloof site grew in a wet seep and were in full flower (though closed during the day time) in mid September while nearby in well drained, stony sand, plants of subspecies spicata were in fruit. Despite the similarity of the flowers of subspecies spicata and subspecies fistulosa, the two subspecies seem well separated from one an- other not only in leaf morphology and habitat, but in flowering time. Alternative treatment of the latter as a separate species for subspecies fistulosa seems equally acceptable. 70. Hesperantha saldanhae Goldblatt, J. S. Af- rican Bot. 50: 119. 1984. TYPE: South Africa. Western Cape: granite rocks at Vredenburg, 8 Aug. 1962, G. J. Lewis 5977 ые NBG!). . African Last revisionary account: Goldblatt, J. S Bot. 50: 119. 19 Hesperantha saldanhae remains known only from the type collection made by C. J. Lewis in August 1964. Plants were collected on une rocks at Vre- denburg. Repeated visits to the presumed type lo- cality, a prominent cluster of exposed granite rocks at the edge of the town, have failed to reveal any sign of the species. Only moderate disturbance at the site makes it seem unlikely that the species is extinct due to human activity. Nevertheless, H. sal- danhae must be considered seriously endangered, and possibly extinct. П. Hesperantha sect. Radiata Goldblatt, Bot. 09: 377 Hesperantha radiata (Jacq.) Ker Gawl. Ann. TYPE: Missouri Gard. Plants with relatively large symmetrical, bell- shaped corms with a flat, horizontal or oblique base; tunics often scalloped below and fringed at the lower margins, older corm tunics not splitting from the base into segments but often the tunics forming scalloped, concave segments. Spike with floral outer bract margins united around the spike axis for up to half their length. Flowers variously colored, most species with a strongly curved peri- anth tube, nearly straight in Н. juncifolia and Н. elsiae. Species 71—79. Mainly in the southern African winter-rainfall zone, but H. radiata extends from Namaqualand, South Africa, in the west to Swazi- land in the east, H. longicollis is eastern southern African, and H. ballii is endemic to Zimbabwe in tropical Africa. 71. Hesperantha brevifolia Goldblatt, J. S. Af- rican Bot. 50: 121. 1984. TYPE: South Africa. Western Cape: Piketberg, Zebrakop, 16 Dec. 1971, Е. FE. Esterhuysen 35320 (holotype, MO!; isotypes, B!, BOL!, BR!, C!, E!, K!, M!, NBG!, P!, PRE!, S!, US!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 121. 1984 Known since 1800 when plants were collected by British botanist John Roxburgh near Tulbagh in Western Cape Province (Roxburgh s.n., G), this spe- cies of section Radiata is rare and has not often been recorded. It was only described in 1984, by which time populations were known from Piketberg, the Cold Bokkeveld, and the mountains north of Bainskloof (Goldblatt, 1984) as well as Tulbagh. It ias only become evident recently that Hesperantha brevifolia extends to the north as far as the slopes of the Nardouwsberg between Clanwilliam and Klawer, some 80 km from the next closest station (Goldblatt & Manning 10720B, MO, NBG; Maguire 1032, NBG). The flexuose stem, bract margins unit- ed around the axis only at the base, and short leaf blades make it relatively easy to recognize H. brevi- folia within section Radiata. 72. Hesperantha juncifolia Goldblatt, J. S. Af- rican Bot. 50: 135. 1984. TYPE: South Africa. Western Cape: Bredasdorp, Ratelrivier, lime- stone flats, 29 Sep. 1970, P. Goldblatt 403 (ho- lotype, BOL!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 135. 1984 Known only from the type collection when first described, Hesperantha juncifolia has since been re-collected twice (Esterhuysen 36425, BOL, MO, Esterhuysen 36371, BOL, MO) at Brandfontein along the west coast of Cape Agulhas, a short dis tance from the type locality (Goldblatt, 1987). The new collections confirm that the species is a local endemic of wet depressions on coastal limestone flats and distinguished from the related H. radiata Volume 90, Number 3 Golablatt Hesperantha Review by the terete leaf blades, outer bract sheathing the stem only in the lower third, and the straight peri- anth tube 5-6 mm long. Flat-leaved H. radiata has flowers with a curved perianth tube mostly 7—15 mm long, and the outer bract sheathing the stem for about half its length. Additional biological notes were recorded by Goldblatt (1987), including the fact that the white flowers open early in the morning and close at about 16:00H, unusual for white-flow- ered species of section Radiata. 73. Hesperantha marlothii R. C. Foster, Contr. Gray Herb. 166: 20. 1948. TYPE: South Af- rica. Northern Cape: Sutherland District, near Waterkloof, Sep. 1921, Marloth 10412 (holo- type, B!; isotypes, PRE!, STE!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 130. 74. Hesperantha decipiens Goldblatt, sp. nov. TYPE: South Africa. Northern Cape: Kamies- berg, northern slopes of Rooiberg, 19 Sep. 1991, P. Goldblatt 9258 (holotype, NBG!; iso- types, K!, MO!, PRE!). Figure 8 Plantae 10-15 ст altae piane eramosae, cormo campanulato basi obliqua plana 8-12 mm diam., boys lignosis concentricis, foliis 4 тыи vel falcatis 1.2— nedio incrassatis, spica plerumque e3- ad 6-flora flexuosa, floribu s albis vel cremeis odoratis tepalis externis abaxialiter ex rufis еы tubo perianthii 10— 13 mm longo ad apicem curvato, tepalis 12-14 X 2.5- 3.5 mm, filamentis ca. 4 mm longis, antheris ca. 5 mm longis, ramis styli antheras parum excedentibus. Plants 10—15 em high, stem usually unbranched. Corm ovoid with an oblique flat base, 8-12 mm diam., tunics woody, concentric, somewhat scal- loped into concave segments with fringed edges. Leaves 4, the lower 3 basal and largest, the upper- most inserted in the middle of the stems and some- times entirely sheathing or with a short free uni- facial tip, linear, 1.2-2 mm wide, slightly fleshy, slightly thickened in the midline. Spike mostly 3— 6-flowered, flexuose; bracts 10—15 mm long, green, becoming dry above, the outer with margins united around the stem for 3-5 mm, the inner about as long as the outer. Flowers white to cream, the outer tepals pale pink to reddish or brown on the outside, strongly scented of stocks when open after dark; perianth tube slender, recurving above, 10-13 mm long; tepals subequal, lanceolate, 12-14 X 2.5-3.5 mm, spreading + at right angles to the tube when fully open in the later afternoon. Filaments ca. 4 mm long; anthers 5—7 mm long. Ovary oblong, ca. 3 mm long; style branches ca. 12 mm long, weakly spreading, slightly shorter or slightly longer than [— AA Figure 8. Morphology of H. decipiens, full size. Drawn ba imens (Gold- by Yevonn Wilson-Ramsay from pressec blatt 9258, K, MO, NBG). Scale bar 1 the anthers. Capsules oblong-obovoid, 6-8 X ca. + winged, ca. | 3.5 mm; seeds angular, the edges mm long. Flowering. August and September. Distribution. South Africa, Northern Cape, ex- tending from near Springbok in the north through the Kamiesberg to the northern Knersvlakte, on thin, sandy gravel on granite pavement. Evidently first collected їп 1897 by Rudolf Schlechter and only occasionally since then, Hes- perantha decipiens is still relatively poorly record- ed. Confused with either the widespread И. radiata 440 Annals of the Missouri Botanical Garden or its relative from the Roggeveld and Bokkeveld Mountains, H. marlothii, H. decipiens has the outer bract margins united around the spike axis and flat- based corms that define section Radiata. Superfi- cially it seems to merely represent a northern series of populations of H. marlothii. That species is, how- ever, defined by a few-flowered, flexuose spike, out- er bract margins united only near the base, and corms with prominent lateral spines. Hesperantha decipiens has the flexuose, few(or several)-flowered spike of H. marlothii, but the outer bract margins are united for up to 5 mm, about a third of their length, and more significantly, the corms lack lat- Instead the corms closely resemble radiata in their scalloped, slightly fringed lobes. Close examination of the flowers shows that the style branches of H. decipiens usu- ally exceed the anthers by 1-2 mm and the anthers are 5—7 mm long. Robust specimens of H. marlothii have anthers 6—8 mm long and the style branches reach only to about the middle of the anthers. Less robust specimens have anthers 4.5—6 mm long, but the style branches are still slightly short of the an- ther apices. Hesperantha decipiens can be distin- guished from H. radiata because that species has the outer bract margins united for at least half their length and a straight spike typically bearing more than 8 flowers. The capsules of H. decipiens are slightly shorter than the bracts and 6—8 mm long. whereas those of H. radiata are usually slightly lon- ger than the bracts and 8—10 mm long. Mature cap- sules are seldom collected so that it is uncertain that this distinction holds for all populations of both species. SOUTH AFRICA. Northern Cape: 29.17 ass, damp sand on Paratypes. (Springbok) near Wilde Lora ie granite (DC), 9 Sep. . Goldblatt 5751 (MO). 30.17 (Hondeklipbaai) hills : ia "Rietklaof (BD), 11 Sep. 1897, Schlechter 11202 (B, GRA, K, 4). 3018 sd Ka- miesberg 2 km S of Leliefontein, 17 Sep. 2 pe Goldblatt & Porter 12226 (MO, NBG); farm Welkom, lower eastern slopes of Rooiberg (AC), 9 Sep. 1980, m 5768 (M, O, PRE); Смар Kloof, northern approach to Rooib- seo. 19 Sep. 1991, Goldblatt 9258 (MO, NBG, PRE): N of FM tower m Leliefontein, 20 Sep. 1991, Goldblatt & Mi 10008 (MO, NBG). Western Cape: 3118 (Van- rhynsdorp) Knersvlakte, near gypsum mine N of the Sish- en rail-line (BC), 21 Aug. 1983, Bean 1272 (BOL) = 75. Hesperantha radiata (Jacq.) Ker Gawl., Ann. Bot. (König & Sims) 1: 225. 1804. Ixia radiata Jacq., Icones РІ. Rar. 2: pl. 280. 1782. TYPE: South Africa. Without precise locality. Jacq.. Icones Pl. Rar. 2: pl. 280. 1782 Hespe d tysonii Baker, Handbk. Irideae 151. 1892. TY South Africa. KwaZulu-Natal: streambanks ear Kokstad, Oct. 1883, W. Tyson 1585 eh Un Kt isotypes, B not seen, GRA not seen, NH!, S Hesperantha iere var. y caricina Ker Gawl., Curtis's Bot. Mag. 21: pl. 790. 1804. TYPE: South us a. due pe ise locality, Curtis's Bot. Mag. pl. Heron uil Salisb., Trans. Hort. Soc. 1: 321. 181 a new name for H. radiata var. y caricina .. Curtis's Bot. Mag. 21: pl. 790. 1804. Hes- perantha caricina (Ker Gawl.) Klatt, Abh. Naturf. ves. Halle 15: 395 (Erganz. 61). 1882, nom. Шер. superfl. pro H. tenuifolia (based on the same type). "y Gaw Last revisionary accounts: Goldblatt, J. S. Afri- can Bot. 50: 123. 1984. Hilliard & Burtt, Notes Roy. Bot. Gard. Edinburgh 43: 436. 1986 (as H. Lysonii). The circumscription of Hesperantha radiata has been unsettled since Hilliard and Burtt (1986) maintained that the eastern southern African plants included in Н. radiata by Goldblatt (1984) repre- sent a separate species, H. tysonii. Clearly popu- lations in eastern southern Africa are more uniform than in the winter-rainfall zone, favor moist habi- tats, and flower mostly in November or December (although the type collection of H. tysonii was made in October). These authors do not, however, list any morphological features in which the eastern plants differ from the western apart from having less woody corm tunics (in the few specimens available that have corms). If these are separate species I fail to see how to distinguish them or to determine what the populations from the area between KwaZulu- Natal-Lesotho and the southwestern Cape (the eastern. Cape and Karoo mountains) should be called since they cannot be distinguished from many collections from either the summer- or winter- rainfall zone. As a matter of practicality, if nothing more, H. tysonii must be regarded as conspecific with H. radiata. The two can only be keyed out if distribution and flowering time are invoked. In an effort to distinguish eastern populations from those in the winter-rainfall zone I have reex- amined a selection of specimens of Hesperantha ra- diata from its entire range. Plants from the winter- rainfall zone differ slightly in having anthers (4—) 5.5-7 mm long and a perianth tube (6-)8—12 mm long, and thus mostly slightly smaller flowers than eastern populations in which anthers are 7-8 mm long and the perianth tube is 9—14 mm long (Hil- liard & Burtt provided dimensions of 5.5-9 mm long for anthers and 7-13 mm for the perianth tube based on plants they examined from eastern south- ern Africa). These overlapping dimensions reflect a surprising consistency in flower size from Nama- qualand and the western Karoo in the winter-rain- Volume 90, Number 3 2003 Goldblatt Hesperantha Review fall zone west to the northeasternmost populations in Swaziland and Dullstroom in Mpumalanga in the summer-rainfall part of the subcontinent. The Dullstroom population is a particularly distinctive form (Drews 188, NBG) in which the flowers are uniformly cream and the tepals large, ca. 17 X 5 mm, the anthers 8.5 mm long, but the perianth tube, ca. 10 mm long, is well within the expected range. All that remains is a possibly weak distinc- tion in the corm tunics. Paucity of material with corms makes comparison difficult, but corms of a collection from Lesotho (Dieterlen s.n., SAM 2408) differ not at all from some specimens from the Western Cape (e.g., Lewis 5988, NBG). I suggest that Н. radiata must be considered one of a rela- tively small number of species that occur across southern Africa bridging the opposed climate re- gimes of the subcontinent. More puzzling to me are collections of small- flowered plants of Hesperantha radiata from the Cape Peninsula and hills around Stellenbosch and Somerset West in the extreme southwest of the southern African winter-rainfall zone. These plants have many-flowered, unusually crowded spikes in which the bracts slightly overlap one another and the uppermost leaf is always entirely sheathing and reaches almost to the base of the spike. The flowers have tepals 7-9.5 mm long, a tube 6-7 mm long. and anthers 4—5 mm long. The stems also have a weakly developed neck of fine fibers around the base. Including these plants in H. radiata makes winter-rainfall H. radiata appear even more vari- able than do the dimensions mentioned above. Par- ticularly notable is the fact that the small-flowered plants bloom later than the typical ones, and they may be sympatric, as for example. Oliver 4332, NBG (21 August), and 4756, NBG (17 October), both collected in 1973 in the hills at Langverwacht near Stellenbosch. The August-flowering plants are typical H. radiata and the October-flowering ones the small-flowered form with crowded spikes. Spec- imens of the latter, collected by C. F. Ecklon and C. L. Zeyher in the early 19th century, tated Н. setacea Eckl. (e.g., Ecklon & 2 Trid. 233—89.9) (Ecklon, 1827), while some sheets at the Kew Herbarium are annotated H. tenuifolia. This is R. A. Salisbury’s (1812) name for H. radiata var. y caricina of Curtis’s Bot. Mag. pl. 790 (Ker- Gawler, 1804). The epithet alludes to the charac- teristic narrow leaves, also, however, found in some populations of larger-flowered plants that corre- spond to the type of H. radiata. 76. Hesperantha ballii Wild, Kirkia 4: 136. 1963. ah PE: Zimbabwe. Chimanimani Mts., Point 71, July 1961, Ball 948 (holotype, RENI. Last revisionary account: Goldblatt, Ann. Mis- souri Bot. Gard. 73: 135. 1986. 77. Hesperantha longicollis Baker, Bull. Herb. 904. TYPE: South Africa. Gauteng (as Transvaal), spots 11 Sep. 1898, P. Conrath 600 (syntype, K!). Boissier ser. 2, 4: 1004. Last revisionary account: Obermeyer, Fl. Pl. Af- rica 46: pl. 1810. 1980 78. Hesperantha elsiae Goldblatt, J. S. African Bot. 50: 136. 1984. TYPE: South Africa. West- ern Cape: ar T River Kloof, above Disa Pool, 11 v. 1979, P. Goldblatt "es (holotype, MO!; р: аю K!, NBG!, PRE!, US!, WAG!). Last revisionary account: Goldblatt, J. S. African Bot. 50: 136. 1984. 79. Hesperantha muirii (L. Bolus) G. J. Lewis, J. S. African Bot. 7: 32. 1941. Acidanthera muirii L. Bolus, Ann. Bolus Herb. 1: 5; 1915. TYPE: South Africa. Western Cape: Riversdale District, Farm cra Oct. 1931, J. Muir 1087 (holotype, BOL! Last revisionary account: Goldblatt, J. S. African Bot. 50: 133. 1984 Literature Cited Mi G. J. 1892. Handbook of the Irideae. Reeve, Ash- cent. . 1896. o 'eae. In pi is Thiselton-Dyer, Flora Capens is 6: 7-17 ent Garden, London 35. ‘One a Botany, inc luding a Gen: eral History of the Vegeta ingdom, in which Plants are Arranged According to ths System of Natural Affin- ities. " Rens peut London. Ecklon, Topographisches Verzeichniss der т kei ode der С. F. Ecklon. Reise Verein, Ess- inget aa D. & O. A. Leistner. 1971. A degree square reference system for citing biological records in south- ern Africa. Mitt. Bot. Staatssaml. München 10: 501- 509 Foster, R. C. 1941. Studies in the Iridaceae. П. A revision of Geissorhiza Ker-Gawl. 8 19 worthy species of Hesperantha. Contr. Gray Herb. 166: Goldblan, P. 1971. Cytological and morphological studies uthern African Iridaceae. J. S. African Bot. 37: 517-460. . Corm morphology in Hesperantha (Irida- ceae, я and a proposed infrageneric taxonomy. Ann. eins ri Bot. Gard. 69: 370-371 : revision of Hespe muda (Iridaceae) in . J. S. African the winter ini area of southern Africa Bot. 50: 15 442 Annals of the Missouri Botanical Garden 985. Systematics of the southern African genus Geissorhiza (Iridaceae —Ixioideae). Ann. Missouri Bot. Gard. 72 E 17441. ————. 1986. Notes on the systematics of жерш а: in tropical Africa. Ann. Missouri Bot. Gar 134-139 w species and notes on southern African He. оа їс: ıe). S. African J. Bot. 53: 459-463. —————. 1990. Phylogeny and e un of Iridaceae. Ann. Missouri Bot. Gard. 77: 607—627. س‎ . Iridaceae. In ( /. Pope (editor), Flora Zamhe 'siaca а 12( E —100. Flora Zambesiaca Managing Committee, Londor & J. C. Manning. 1995. Phylogeny of the African genera Anomatheca and Freesia с ceae-Ixioideae), and a new genus e apa. Syst. Bot. 20: 161—178 ——— Ф —— —. 19 edicion of Schizostylis (Iri- daceae: Ixioideae) in tram Novon 6: 262-204. — & ———. [0 Gladiolus in Southern. Africa: Systematics, вое. p Evolution. Fernwood Press, Cape Tow С — ———. 1999, The long-proboscid fly pollina- ion eras > п Gladiolus (Iridaceae). Ann. Missouri Bot. Card. 86 — & ——— ). Long- -probosc "id fly es in southern Milos. Ann. Missouri Bot. Gard. 87: 146— 170. & Takei. 1997. Chromosome cytology of Iri- daceae, base n numbers, patterns of variation and modes of karyotype change. Ann. Missouri Bot. Gard. 84: 285— 304. ——— . Wagner. 1984. A survey of seed surface uec 00 in А (Iridaceae). Ann. Missouri Bot. Сагс ou 31—190. —— E & P. Bernhardt. 1995. Pollination biology e irousia iiis nus Lapeirousia (Iridaceae) in southern Africz divergence and adaptation for long- о fly latin Ann. Missouri Bot. Gard. 82: 517—5 ——, J. PPM J. Davies, V. Savolainen & S. Rezai. а Cyanixia, a new genus for the Socotran endemic, Babiana socotrana oe subfamily Cro- coideae). e gh J. Bot. 60: 3 —— ub "Ben ihardt & J. .. Manning. In press. Floral bs vof He читти ТУТИ eae: Crocoideae): How minor shifts i in floral presentation change the pol- lination а rei Мін Bot. Gard. Gunn, M. & L. E. Codd. 1981. Botanical Exploration of Southern A а. А. А. Balkema, Cape Town. Hilliard, O. M. & B. L. Burtt. 1979. Notes on some plants from southern Africa chiefly p Natal: УШ. Notes Hoy. Bot. Gard. сн à 37: 204—325 — & —— ». Hespe s (Iridaceae) in Natal E ie i o Bot. Gard. Edinburgh 43: 407— к. RN J. 18 carex-leaved Hesperantha. Curtis's 190. Klan, à MP теш Revisio Iridearum (Conclusio). Linnaea 34: 04. Hesperantha radiata var. n Bot. Mag. 21: pl. о » (a & P. Goldblatt. 1996, The Prosoeca per- ingue: yt (Diptera: Nemestrinidae) pollination syndrome in southern Africa: Long-tongued Ps and their tubular flowers. Ann. Missouri Bot. Gard. 83: 67-86 Meyer, E. 1843. Zwei pflanzenge deii һе Doc umente. Flora 26, Beihefte 1-230. Reeves, G., P. Goldblatt, M. W. Chase, T. De Chies. A. V. Cox, B. Lejeune, P. J. Rudall & M. F. Fay. 2001a. Mo- lecular systematics of Iridaceae: A combined analysis of four plastid DNA sequence matrices. Pp. 29-42 in M. Colasante & P. J. Rudall нө Irises and Iridaceae: Biodiversity and Systematics. Proceedings of the International Iridaceae erus (nce. Ann. Bot (Roma), nuov. ser., 52 (2). — M. ww P. Goldblatt, и J. Rudall, M. F. Fay, A. V. Сох, . LeJeune & T. Souza-Chies. 2001b. Molecular systematics of Iridaceae: Evide nce from four plastid DNA regions. Amer. J. Bot. 88: 2074—2087. Retief, E. & P. P. J. Herman. 1997. Plants of the Northern Provinces of South Africa: Keys and Diagnostic Char- acters. Strelitzia 6. Salisbury, К. A. 1812. A the cultivation of rare plants. Trans. Hort. Soc. 1: Steiner, K. E. 1989. m pollination of Disperis (Orchi- daceae) by oil-collecting bees in southern Africa. Lin- dleyana 4: 164—183. Trauseld, W. R. 1969. Wild Flowers of the Natal Drak- ensberg. Purnell, Cape Town. APPENDIX 1 LIST OF SPECIES WITH NUMBERING AS PRESENTED WITHIN THE ARTICLE Geissorhiza macra see H. leucantha Pe iude acuta cente еле seen 3 H. alborosea о _ KS ME . 41 H. Kissed mu MON кыа 30 Н. bachmannii — . 32 H. ballii |... _ 76 Н. Ваши 35 Н. brevicaulis |... 60 H. brevifolia Lo aure ere __ 71 Н. brevistyla 42 Н. in ОДИ "ERR 33 Н. candida . 26 H. candida var. bicolor see > H. longituba H. cedarmontana — | AERE = . бо H. ciliolata |... 12 IT: COCGIBER eL see ee eke 56 Н. erocopsis 00 31 H. cucullata |... 0 o 18 H. curvula 0 Ll 61 Н. debilis ese ee sa aks 27 Н. decipiens |. Т: Н. elsiae |. 78 Н. erecta Sees eks 1 H. exiliflora |... 0 10 Н. falcata 2 __ 63 H. fibrosa 00000050000 17 H. flava |... i z m 25 H. flexu e OIE 15 H. Жай, вее Н. woodii H. glabrescens | SS " - 211 H. glareosa ee a ee _ 37 H. gracilis 0 __ RS крл . 9l H. кан | "S uir Lec iie ce 04 H. hes ا‎ ee к л 23 H. humilis TN m ки . 24 H. hutchingsiae -" t E РО .. 99 Н. huttonii 58 H hygrophila — 47 H. inconspicua 2 _ 15 ingeliensis ааыа аады 43 Volume 90, Number 3 Goldblatt 443 2003 Hesperantha Review И ikarooica а ee 22: -H quadrangula "————— 14 HI. Таса tee ge ae ae [C o c з= PR 75 ADT CT se en RUM: 67 Н. radiata var. caricina, see H. radiata H. leucantha М0 30: ИШИМИ aa i cdi onion 7 H. longicollis i aoa — M" TM ИО ыы a e a mE ындан данны 48 H Rond. сыы деке a ada 28. Н. rupicola. a e сне инн 4 НСО e rtp a e 68. Н. saldaRhûê e e mea 70 NM malina есте ананна а е неле inion 8 H.saxicola a ыс нынын i a nn 46 Н. marlothii ase ا‎ 73 H. şehelpeanā E 29 Н. mma ile aaa 16 .Hsehleahieri ura a 38 H. modesta oovan 49 Н. scopulosa saa ga 62 РЕАЛЕ ЕЛИНИН a ee 6 H. similis, see H. schlechteri HARF e ERES dO: FE SpICOld e ныды и e 69 Н, namaguana i a e 2 H.stenosiphon sa a 55 FAUT 01 e أ‎ 5. JH. SUPPL a e e a ee 64 Bel RAE asas e 34 H. га see Н. radiata H. pauciflora. sasea gaan 66- H teretfolid a a 13 H. pentheri, see H. falcata H. trifolia, see H. falcata Heli... eco eaae aaa 36 M tpuncatula. eu orient aA 19 H. pilosa neonin a Eaa нан 9 Н. tysonii, see H. radiata H. Wi NGI agen 10 H. umbricola a ee 50 Н. púbinervia „+з acer pment 2: с мылын» a o 21 H. pulchra e tage saan 53 H. vernalis, see Н. candida H NR ass a m 20 JE woodii n Өф BIOGEOGRAPHY AND Н. Zhu,” H. Wang, B. Li, and FLORISTIC AFFINITIES OF ROVE THE LIMESTONE FLORA IN SOUTHERN YUNNAN, CHINA' ABSTRACT The forests on limestone in southern. Yunnan, in tropical southwest China, were inventoried, and their floristic composition and biogeographical affinities are discussed. These limestone forests were characterized phanerophytes making up ca. 78% of the total species and those with mesophyllous leaves comprising 75%. Ecological species groups based on their habitat preferences were discerned from field observations: the species exclusive to the limestone habitats make up 10% and the preferents make up ca. 12% of the total limestone flora. From these limestone fore ts, 1394 vascular plant species belonging to 640 genera and 153 families were recorded. Based on their о, 15, 12 e elements at the generic level and nine at the specific level were recognized. About 90% of the seed ee t genera (over 90% of the species) were tropical; furthermore, 35% of the seed plant genera (65% of the species) ave tropical 8180 абое, In a арра аа por apina floras itom southern China nd ab dn E ай, ls ih This Titae il ra is thus tropical and p rt of i tropical Asian flo та at its ا‎ калы Key words: biogeography. China, limestone ба southern Yunnan. Amestone in tropical China occurs mainly in GENERAL. GEOGRAPHY Yunnan and Guangxi Provinces of southern and | s | "ООН Xishuangbanna, the southern part of Yunnan, central China. Because of the great diversity of . : : i ia т which borders Burma and Laos, is a mountainous edaphie conditions and topography, vegetation . : x ` area at the northern margin of tropical Southeast Asia (Fig. 1). Basically, the study area has a moun- tainous topography with the mountains running types on limestone are extremely diverse and rich in endemic taxa. Limestone vegetation in southern China has been destroyed as much as other vege- : ? north-south and decreasing in elevation southward. tation types even though these limestone areas are Altitude varies from 480 m in the lowest valley in the south to 2400 m at the top of the highest moun- tain in the north. The limestone strata occur mainly more difficult to access and to farm. Limestone veg- etation is also more vulnerable because it recovers much more slowly on usually thin soils. Our re- in southeastern Xishuangbanna and range in alti- search was conducted mainly in the area of Xish- tude from 600 to 1600 m uangbanna, in the southern part of Yunnan, where about 19% (3600 km?) of the total area is limestone (Liu et al., 1990). Most of this limestone area is The region of Xishuangbanna has a typical trop- ical monsoon climate with an annual mean tem- perature of 22°C, annual temperature accumulation still forested and is receiving increasing attention (the sum of daily temperature means where they are for its biodiversity and its urgent need of conser- > 10°C) of 8000°C, and annual precipitation vary- vation. Primary floristic works in southern Yunnan ing from 1200 to 1556 mm, of which more than have been written (Zhu et al., 1996, 1997, 1998a, 80% falls during the rainy season between May and ) 1998b; Wang et al., 1997). This paper represents a the end of October (Xu et al.. 1987 synthesis of its floristics, physiognomy, and biogeo- The rock substrate is hard limestone of Permian graphical affinities. origin with a rugged topography. The soil is mainly ! This project “ funded by The National Natural Science Foundation of China (40271048), the Chinese Acad emy of Sciences (The Fund for Top One Hundred Young Scientists and KSCX2-1-06B), ee the Yunnan Natural Science Foundation, The senior е thanks Хи Zaifu for his great help with his research a u Zheng-yi and Zhang Hong- da, his academic advisors. He particularly thanks E. Tanne r and P, Grubb for their ж in nan ing d ~ and preparing this paper during his visiting scholar’s year at the University of Ca cambridge. Finally, he thanks T. C. Whitmore, who has greatly supported ee helped him in his research, and two anonymous reviewers for their constructive comments on this article. ishuangbanna Tropical Botanical Garden, The Chinese Acade my of Science, Mengla, Yunnan 666303 P. R. China. E Me address for corresponde nee: zhuhua@public.km.yn.en; zhuh@xtbg.ac cn ‘Prince of Songkla University, Hatyai 90112, Thailand. ANN. MISSOURI Bor. GARD. 90: 444—465. 2003. 445 Southern Yunnan Limestone Flora Zhu et al. Volume 90, Number 3 2003 "e[nsuruoq- AP[PN ‘Budeg, ()[ — ^2uojsour e[nsuruoq LLW ou ‘67 "pue[reu N "oep3uorq?) *g— `шрщәгд N "Suonudon?) :*; — eup *surejuno[y Sue43uo(| *9— "eun ‘udeny ^c— euo AS *Suibur[ne) p— сшщ) AS 'surejunoj ueqsurbe(p “¢— eut) MS 0288007] Ce “Pole 2182524 In() WE (9 914], 3235) seloy [8401891 po1eduioo pue ‘euyy `иеиипд uIəy nos ‘guueqsuenysIy и! 14 цәтеәвә1 эп jo SUOTJBVO'T i 21n31 J °$ 1 o0ZI о511 о011 о501 о001 956 06 098 o08 ` ] Е (mJ А LES 1 ® Ы S APA j tI [ete —— Р ludo lsemejns sjejeuio|4 OOO! OSL 005 osz Q d eu оәшоя о ч g =”. к= D lew Panis id 4 (> А . ` b È | A: n e . "x e UE с . Jeol 2 афо еўиет ys E r, 2, ci 2. - 3 = URS H . gees ешо unos 0 E : By ad S S| зев Ese. = а - puereur - E uid fud ES ipul 3 уййн ri ‚7 Y a aL |, چ‎ “= A L < B í JeuiueAA Д m S iu rd. "ect d e6 leg i ail у y 2 p^ lu а Г е Е 8 -ynog euluo ed Tue 4c è ч Р | a л IH 35 qi jedan r3 rn n u De 521 021 ET o01] RT 001 56 M o$8 208 446 Annals of the Missouri Botanical Garden brown, coarse in texture, and composed of loamy limestone with a pH of ca. 6.75 and ca. 3.56% organic matter (Liu et al., 1990). METHODS A complete floristic inventory was made based on the identification of more than 5000 plant spec- imens the habitat ii southern Yunnan during 1985—1995 and deposited mainly at HITBC and SYS. The flora of the vege- collected from limestone ~ tation on the limestone consisted of 153 families of vascular plants, including 640 genera and 1394. species. An initial floristic analysis was made based on the inventory (Zhu et al., 1996). vegetation types occur on the limestone—tropical seasonal rain forest, tropical seasonal moist forest, Three main and tropical montane dwarf forest—which were se- lected for establishing plots. For the tropical sea- sonal rain forest, seven separate plots ranging in size from 2000 to 2500 m? were established. For the tropical seasonal moist forest seven separate plots ranging in size from 500 to 2000 m2 were laid out. For the tropical montane dwarf forest, only two plots of 10 by 10 m were made due to its restriction limestone summits. These different plot sizes were used because of the differential coverage of forest type and site restrictions. The structure and species composition of the vegetation on the lime- stone were analyzed based on plot data already published (Zhu et al., 1998a). In the present paper, plant inventory lists of the two main forest types (excluding montane dwarf forest) were compiled from sample plots separately for the physiognomic life form and leaf size) analysis. The criteria for —. life form and leaf size classes suggested by Raun- kiaer (1934) and the importance value index (IVI) suggested by Curtis and McIntosh (1951) were used in the physiognomic or ecological analysis. Ecolog- ical species groups were discriminated from field observation and correspond to groups used in Shi- mizu (1964) and Chin (1977). Species-level bio- geographical affinities were assessed for the total flora of the limestone vegetation. The floristic sim- ilarities between the limestone flora of southern Yunnan and the floras on limestone and non-lime- stone habitats from southwest China, northern Vi- etnam, northern Thailand, and the Malay Peninsula were also discussed. CLASSIFICATION OF LIMESTONE VEGETATION Based on plant physiognomy, forest profile, flo- ristic composition, and habitat, the primary lime- stone vegetation can be classified into three vege- tation types, i.e. tropical seasonal rain forest, tropical seasonal moist m. and tropic ‘al montane dwarf forest (Zhu et al., 1 formations, including nine кен меге гес- ognized: iin these, six (1) Ravine seasonal rain forest (including the Po- metia tomentosa—Alphonsea monogyna community and Pometia tomentosa—Celtis philippensis var. wightii community); (2) Lower hill seasonal rain forest (including only the Celtis philippensis var. wightii-Lasiococca com- beri var. pseudoverticillata community ); (3) Evergreen moist forest (including the Osmanthus polyneurus—Dracaena cochinchinensis community and Lasiococca comberi var. pseudoverticillata—Cleis- tanthus sumatranus community); (4) Semi-evergreen moist forest (including the Bombax insignis-Colona floribunda community and Bombax insignis-Garcinia bracteata community); 5) Evergreen dwarf forest (including only the Pho- tinia angusta—Pistacia weinmannifolia community); (6) Semi-evergreen dwarf forest (including only the Ficus nertifolia—Dracaena cochinchinensis commu- nity). = Detailed descriptions and ecological analyses of the communities have been reported earlier (Zhu et al., 1998a). Here the classification of the limestone vegetation is concisely enumerated so that the bio- geographical components of the limestone vegeta- tion can be better understood. TROPICAL SEASONAL RAIN FOREST Tropical seasonal rain forest on limestone, just as the regional tropical seasonal rain forest off lime- stone, shares characteristics with the equatorial lowland rain forest. These forests are mainly ever- green, but there are some deciduous trees in the emergent layer. This is equivalent to the tropical semi-evergreen rain forest of Southeast Asia (Whit- more, 1984), or the tropical of India-Burma (Champion, semi-evergreen forest 1936), as well as the evergreen seasonal of tropical America (Beard, 1944, 1955). In southern Yunnan, limestone forests occur in wet valleys and on lower slopes of hills or mountains below 1000 m altitude. This same forest type also occurs in northern Thai- land (Smitinand, 1966) and North Vietnam (Thin, The tropical seasonal rain forest represents Southeast forest these 1997), although different names were used. Asian tropical rain forest at its latitudinal and al- titudinal limits. The ecological structure of the trop- ical seasonal rain forest on limestone is almost ex- actly the the rain forest. off limestone in the Xishuangbanna region (Zhu. 1992, same as seasonal Volume 90, Number 3 2003 Zhu et al. 447 Southern Yunnan Limestone Flora Life forms of the limestone forest in southern Yunnan. Table 1. Thero- Geoph phytes Phanerophytes Liana All Cham 156 Hph Megaph Mesoph Microph Nanoph Epiph Woody Herb Para Life form* 34 Number of species Limestone seasonal rain forest 2.8% 19.3% 1.2% 3.6% 33.7% 13.7% 68% 48% 62.7% 124% 1.2% 0.4% Percentage of total species (14800 m? of 7 plots, total 249 species) 124 24. 5 2 62 36 18 Number of species Limestone seasonal moist forest 0.9% 0.9% 29% 17% 9.9% 14% 58.8% 13% 3.3% 8.5% 12.8% 2.3% Percentage of total species (9650 m? of 7 plots, total 211 species) * Life form (Raunkiaer, 1934); Megaph Mesophanerophyte (perennials 8 to 30 m high); Microph = Microphanerophyte Megaphanerophyte (perennials over 30 m high); Mesoph Epiphytes; Therophytes Parasitic; Epiph = ;eophyte (perennials, dying back above ground); Para = G Chamaephytes (perennials less than 0.25 m high above ground); Geoph (annuals). 1997). Most species in the seasonal rain forest on limestone are also found in the adjacent non-lime- stone seasonal rain forest, but the latter is more diverse with additional species, which are not found on the limestone. TROPICAL SEASONAL MOIST FOREST Tropical seasonal moist forest occurs on the mid- dle and upper limestone slopes ranging from 650 to 1300 m altitude. This vegetation type abuts the seasonal rain forest and was called monsoon forest by some Chinese authors (Liu, 1987; Wu, 1980). The term seasonal moist forest is preferred here because the forest is not equivalent to Schimper's monsoon forest (Schimper, 1903), in spite of the fact that it is affected by seasonal dryness and con- tains a variable percentage of deciduous trees. The seasonal dryness in the region is compensated to some extent by dense fog accompanied by low tem- peratures in the same months (November to April) (Whitmore, 1984). Some deciduous trees, such as Gmelina arborea Roxb., Anthocephalus chinensis (Lam.) Rich. ex Walp., and Homalium laoticum Gagn. var. glabretum C. Y. Wu, shed leaves toward the end of the dry season, while others, such as Cratoxylon cochinchinensis (Lour.) Bl., Ficus reli- giosa L., and Elaeocarpus varunua Buch.-Ham. ex Mast., shed their old leaves as new ones develop. This suggests that deciduousness in the region 15 more frequently associated with locally dry habitats than the seasonal dryness of climate. Therefore, us- ing the term monsoon forest for the evergreen or semi-evergreen forest on limestone is confusing be- cause Schimper’s monsoon forest is more or less completely leafless during the dry season. MONTANE DWARF FOREST Montane dwarf forest occurs only on the tops of hills and summits of mountains at altitudes above 900 m. There is only one dwarf tree layer with a canopy height of 7-15 m. Epiphytic orchids, such as Eria hainanensis Rolfe and Bulbophyllum ni- grescens Rolfe, and non-vascular epiphytes (bryo- phytes and lichens) are abundant. In some sites small woody climbers, such as Derris caudatilimba How (Papilionaceae) and Pristimera arborea (Roxb.) A. C. Smith (Hippocrateaceae), are also fre- quent. PLANT PHYSIOGNOMY OR ATTRIBUTES From plot data, life form spectra (Raunkiaer, 1934) of the two main forest types (seasonal rain forest and seasonal moist forest) are compiled in 448 Annals of the Missouri Botanical Garden Table 2. Physiognomic characteristics of the limestone forest in southern Yunnan. Leaf form Leaf texture Leaf size Forest type > C P L Na Mi Me Ma Limestone EM Percentage of species 72.3 27.7 47.9 52.1 0 13.8 76.6 9.6 rain for Percentage of Importance Value Index (IVI? 76.3 23.7 52.8 47.2 0 38 9] 53 Bonne NE Percentage of species 68 32 51.5 48.5 1 215 742 3.1 moist forest? Percentage of Importance Value Index (IVI) 74.8 25.2 41.7 58.3 0.4 23.3 66.5 9.7 ! From 14800 m? of 7 plots, total of 94 tree species > 5 cm DBH. ? From 9650 m? of 7 plots, total of 97 tree species > 5 ст DBH. 3 IVI = Relative dominant density + Relative рин. у + Relative dominant breast area (Curtis & McIntosh, 1951). S: Simple leaves; C: Compound leaves; P: Papery leaves; L: Leathery leaves; Ma: Macrophyll (large to 164,025 mm’); Me: Mesophyll (to 18,222 mm’); Mi: Mic Тр (to 2025 mm?); Na: Nanophyll (to 225 mm?) (Raunkiaer, 1934). Table 1. Leaf size spectra, leaf form, and leaf tex- weighting the species by IVI, the percentage of ture are shown in Table 2. Both forest types were — leathery leaves decreases in seasonal rain forest but dominated by phanerophytes. Including lianas, increases in seasonal moist forest due to the pres- these perennial phanerophytes make up 73.9- ence of some species with these leathery leaves 83.2% of the total species, while annual chamae- such as Cleistanthus sumatranus (Miq.) Muell.-Arg. phytes account for only 12.4—1396. However, the (Euphorbiaceae) and Dracaena cochinchinensis. seasonal moist forest shows lower percentages of woody lianas as well as megaphanerophytes and ECOLOGICAL SPECIES GROUP mesophanerophytes, but higher percentages of epi- Based on the study of the limestone floras of Ja- phytes as well as microphanerophytes and nano- pan and Taiwan, Shimizu (1964) divided limestone phanerophytes than the seasonal rain forest. plants into five ecological groups: Both forest types have species with mesophyllous leaves making up ca. 7596 of the total tree species, but the forests show clear differences if the species are weighted by importance value index (IVI). This increases the percentage of mesophyllous perenni- als and decreases the percentage of micro- and ma- (1) plants exclusive to limestone habitat; (2) plants selective for and found mainly in lime- stone (3) plants preferring and dominant on limestone; 4) taxa indifferent, with no special association with limestone; crophyllous trees in seasonal rain forest, while the : | (5) plants found only occasionally on limestone or opposite trend is seen in seasonal moist forest. Sea- sonal moist forest occupies much more rugged hab- itats with thinner and drier soils, and has more mi- To Shimizu, these first three groups were char- crophyllous species. In weighting by IVI, the — acteristic species for the limestone habitats and in increase in percentage of macrophyllous trees in particular his exclusive and selective taxa were cal- seasonal moist forest is mainly due to the dominant — cicoles. Chin (1977) accepted this classification evergreen species Dracaena cochinchinensis (Lour.) and similarly categorized plants on limestone in the S. C. Chen (Agavaceae), with its long leathery lan- Malay Peninsula into four groups, combining selec- ceolate leaves, and the dominant deciduous tree tive and preferent plants. Similar ecological species species Colona floribunda (Wall. ex Voigt) Craib groups have been later recognized by Chinese bot- (Tiliaceae) also with large leaves to 30 cm long. In anists (Liang et al., 1985; Liu et al., 1994). strangers to limestone. Table 3. The ecological species groups of the limestone flora of southern Yunnan. Ecological species groups (see Shimizu, 1964; Chin, 1977) Number of species % Plants found only on limestone: endemic to southern Yunnan 24 1.7 not endemic to southern Yunnan 117 8.4 Plants dominant on limestone 170 12.2 Plants no restriction on limestone 858 61.6 Plants found occasionally on limestone 225 16.1 Total 1394 100 Volume 90, Number 3 2003 Zhu et al. 449 Southern Yunnan Limestone Flora Table 4. Predominant families found in limestone forests of southern Yunnan. No. of о. 0 . of o. of genera species %* genera species % Orchidaceae 35 86 26.9 Verbenaceae 6 20 43.5 Rubiaceae 34 58 64.1 abiatae 13 20 30.1 Euphorbiaceae 21 58 60.4 Gesneriaceae 13 19 63.3 Papilionaceae 22 55 43.4 Sterculiaceae 7 18 46.8 Moraceae 7 48 73.8 Dioscoriaceae 1 18 64.3 Vitaceae 7i 38 79.5 Menispermaceae 10 17 60.7 Acanthaceae 26 36 65.4 Liliaceae 10 16 64.0 Rutaceae 11 35 71.4 Araceae 10 15 44.1 Asc am eae 16 35 53.0 Compositae 8 15 14.6 Urticaceae 12 35 53.0 Myrsinaceae 4 14 38.2 Lauraceae 10 35 45.0 Commelinaceae 7 14 60.9 Apocynaceae 19 33 58.9 Zingiberaceae 6 14 42.4 i 12 30 83.0 Myrtaceae 1 13 45.8 Annonaceae 12 30 50.9 Tiliaceae 3 12 57.1 Cucurbitaceae 9 24 52.3 Mimosaceae 6 12 66.6 Rhamnaceae 19 21 10.4 Anacardiaceae 1 11 64.7 Piperaceae 3 20 54.8 Convolvulaceae 4 11 37. Ulmaceae 5 1] 100 iz the no. of species on limestone Yo = x 100 the total no. of species in southern Yunnan Following Shimizu and Chin’s classifications, we divided the limestone flora of southern Yunnan into these four ecological species groups (Table 3). In our study, 141 vascular plant species are restricted to limestone habitats and thus are exclusively found ere. These include the following common species Celtis philippensis var. wightii, Amoora calcicola, Murraya tetramera, Pistacia weinmannifolia, as well as species in Agapetes, Sageretia, Tupistra, and Pristimera. Of these, 24 species are endemic to southern Yunnan. Taxa exclusive to limestone make up about 10% of the total limestone flora, which agrees with the results from Longgan limestone (ex- clusive taxa, 13%) (Liang et al., 1985) and Longhua limestone (exclusive taxa, 10%) (Liu et al., 1994) from Guangxi Province in China. Both the exclu- sive and preferent taxa make up 22.3% of the total sum. They could be termed as characteristic spe- cies for limestone habitats (see Appendix 1). This is similar to the results from Longgan in Guangxi (with these characteristic species making up 20% of the total sum) (Liang et al., 1985) and from the Malay Peninsula (27.5%) (Chin, 1977). THE FLORA AND Irs BIOGEOGRAPHY In the limestone forests of southern Yunnan, Chi- na, 153 families of vascular plants including 640 genera and 1394 species and varieties, were re- corded, of which seed plants compose 129 families, 558 genera, and 1269 species (see Appendix 1). More than 80% of the species also occur in the non-limestone habitats of the Xishuangbanna re- gion. The limestone flora makes up about one quarter of the total species of the regional flora. (The flora of the Xishuangbanna region was primarily docu- mented with 3336 native species of 1218 genera and 207 families of seed plants; see Li, 1996.) Some families show relative preference for lime- stone habitats (with more than 60% of the total number of species in the region on limestone), for example, Acanthaceae, Euphorbiaceae, Gesneri- aceae, Meliaceae, Menispermaceae, Moraceae, iaceae, Rutaceae, Vitaceae, and Ulmaceae (Table 4). Other families, such as Hip- pocrateaceae, Icacinaceae, and Vacciniaceae, show Rh amnaceae, Rubiace an even stronger preference for limestone (found almost exclusively in limestone habitats in southern Yunnan), although they are not among the predom- inant families in species richness. The distribution types of Chinese seed plants at the generic level were documented by Wu (1991). Based on Wu’s document, 544 of the 558 genera of seed plants from the limestone forest of southern Yunnan can be divided into 12 distribution types or geographic elements (14 genera, which are cos- mopolitan in distribution, are not included in the geographic statistics). One thousand two hundred forty-four of the 1269 species of seed plants from the limestone forest can be recognized in nine dis- 450 Annals of the Missouri Botanical Garden Geographic affinities of the limestone forests of southern Yunnan. Table 5. Percentage Geographic element at the generic level Percentage of species Geographic elements at the specific level of genera (see Wu, 1991) l. Pantropic 21.1% Pantropic 2. Tropical Asia-Tropical America disjunct 3. Old World Tropics Tropical Asia-Tropical America disjunct " 2: 13.8% 3. Old World Tropics 4. Tropical Asia to Tropical Australia 5. Tropical Asia to Tropical Africa 4. Tropical Asia to Tropical Australia [Tropical Asia to Tropical Africa 5. 35.3% . Northern Temperate 8. Temperate Eastern Asia and Northern America disjunct 9. Old World Temperate 7 6b. Mainland Southeastern Asia to Malaysia 6c. Southern Asia to Mainland Southeastern Asia Mainland Southeastern Asia to Southern China 6d. 10. Temperate Mediterranean, Western Asia to Central Asia 0.5% 7. Eastern Asia 8. Southern China 9. Eastern Asia 0.4% Endemic to China 12. 18.6% 100% Endemic to Yunnan Total of 1244 species 100% Total of 544 genera tribution types based on their geographic distribu- tion (25 species of seed plants are not included due to insufficient distribution references) (Table 5). At the generic level, the geographic elements of trop- ical distribution (1—6, Table 5) compose 90% of the total genera; the geographic elements of Sabe distribution (7-10, Table 5) make up only 6.7%. the specific level, the species that are of typic al tropical distribution (1—6, Table 5) account for 70.2% of the total species. Among these, the geo- graphic types that are considered to be from trop- ical Asia make up 64.5% of the total species from limestone forests in Xishuangbanna. If the species from the tropical areas adjacent to Xishuangbanna from southern China and Yunnan are included, these tropical species compose more than 90%. This indicates that the limestone flora at Xishuang- banna is principally tropical in nature and repre- sents the tropical Asian flora at its northern tropical margin. In a floristic comparison with nine similar floras, both limestone and non-limestone, from southwest China, northern Vietnam (Thin, 1997), northern Thailand (Smitinand, 1966), and the Malay Pen- insula (Chin, 1977, 1979; Burkill & Henderson, 1925) (Table 6), the limestone flora of southern Yunnan displays explicit taxonomic affinities to the tropical floras and shows a closer affinity to the floras from the Malay Peninsula than to other floras from subtropical China (the floras of Huapin and Dongyang, see Li et al., 1986; Xu, 1984), even though these Malaysian floras lie farther away geo- graphically from southern Yunnan. Our limestone flora in southern Yunnan shares the most genera with the limestone flora of northern Vietnam (Cuc- phuong, see Thin, 1997) among those floras com- pared in this study. The similarity at the generic level between our limestone flora and the limestone flora of northern Thailand (Doi Chiengdao) (Smitin- and, 1966) is less than would be expected from its geographic proximity. This lack of correspondence could be because the plant list for Doi Chiengdao used here for comparison is an incomplete one con- sisting of only 512 species, less than half reported for most other sites in Table 6. The limestone flora of Xishuangbanna did not show a higher floristic similarity to other regional limestone floras than to non-limestone floras in our comparison. It appears that limestone floras develop from local or regional floras, supported also by the fact that only about 10% of the total species of limestone floras (the exclusive group) are restricted to limestone habitats in our study. The floristic relationships between our limestone flora in southern Yunnan and neighboring floras of Volume 90, Number 3 2003 Zhu et al. Southern Yunnan Limestone Flora Table 6. Comparison of floristic similarities between the limestone habitats of UE southern Yunnan, and the limestone and non-limestone habitats from southwestern China and southeastern Asi Size of flora Shared taxa Similarity Location Habitat (Seed plants) by both floras coefficients 2. Longgan, SW Chir limestone 149 families 118 91.2 22°14-33'N, 106°46' E 669 genera 371 66.5 1363 species 3. Daqin е ио SW China non-limestone 182 fam 126 97.4 22°14'N 871 gen. 389 69.8 1813 spp. 4. Gulinqing, SW China limestone 3 fam 116 89.9 22°36'N, 104°E 496 gen 261 52.6 1095 spp. 5. Huapin, SW China non-limestone 151 fam. 83 72.8 25°31-39'N, 109°50'E 475 gen. 150 Joss 1051 spp. 6. Don сЕ е SW China limestone 116 fam. 86 73.7 25°14'N, 107°56'E 367 gen 153 41.6 736 spp 7. Cucphuong, N Vietna limestone 167 fam 120 93 20°14—24' N, 1052444 'E 60 gen. 428 76.7 1661 spp. 8. Chiendae, N Thailand limestone 101 fam. 93 92.1 19?2'N, 98°54 342 gen. 181 52.9 512 spp. 9. Malay peninsula limestone limestone 117 fam. 93 81.6 1-6°N, 100—104^E 535 gen 244 51.6 1112 spp. 10. Taiping, Malay peninsula non-limestone 115 fam. 94 82.5 4°N, IOI^E 682 gen 243 51.6 Notes: The direct аа of species composition between the different floras is not very significant before the floras at is specific level i is not m References: location 2 (Chen, 1985) 3 (Daqinshan Forest station of Guanxi Forestry Bureau, pi et al., 25). tropical Asia and southern China were discussed by Zhu (1997). This limestone flora shares all fam- ilies and 8896 of its genera with the flora of Indo- china (Lecomte, 1907-1951; Aubréville et al., 1960—1996), 96% of its families and 68% of its genera with the flora of the Malay Peninsula (Rid- ley, 1967; Keng, 1978), 73% of its woody plant genera with Burma (Kurz, 1877), and more than 97% of its families and more than 80% of its genera with other tropical floras of south China (including Hainan Island; see Wu, 1994). The limestone flora of Xishuangbanna demonstrates strong affinity to other tropical Asian floras. Literature Cited Aubréville, A., Tardieu-Blot & J. E. Vidal (editors). 1960—1996. Баң du Cambodge. du Laos et du Viet- nam. No. 1-28. Museum лык d'Histoire Naturelle, the taxa revised; therefore, the comparison of floristic similarities between the different 1980); 4 (Li, 1) ә 1986); 6 (Xu, 1984); 7 (Thin, 1997); 8 (Smitinand, 1966); 9 (Chin, 1977, 1979); 10 (Burkill & ТШЕН Beard, J. 5. 1944. Climax vegetation in tropical America. Ecology е 127—158. ————. 1955. The classification of tropical American vegetation pes. Ecology 36: 359—412. Burkill, H. . Henderson. 1925. The flowering P ants s i. in the Malay Peninsula. Gard. Bull. Straits Settlem. 3: 300—459. C hampion H. G. 1936. A preliminary survey of the forest pes “a India and Burma. Indian Forest Rec., n. EL C a Е P. 1985. Limestone Flora in Longgang, Guangxi, China. U published M.S. Thesis, Zhongshan University. [In Chinese. | Chin, S. C. 1971, 1979. The limestone hill flora of Malaya, I, П. Gard. Bull. ирле ЗО: 16 T 32: 64-203. Curtis, J. T. & R. P. McIntosh. 1951. upland forest continuum in ie prairie-forest ipi tun region of Wis consin. om = ak оа Fore on o xi Forestry. Bure 30. Plant lis of СО Guangxi. [U тане dide in Chinese. | eng, Н. 197 gren and Families of Malayan Seed Plants. Singapore Univ. Press, Singapore. - 452 Annals of the Missouri Botanical Garden Kurz, W. S. 1877. Forest Flora of British Burma 1 & 2. Wang, H., H. Zhu & B. С. Li. 1997. Vegetation on lime- Office of the Superintendent of Government Printing, Calcutta. [Reprint 1974, International Book Distribu- lors Гаване. . (editor). 1907-1951. Flora generale de L'Indochine. Tome 1—7. Masson et Cie Editeurs, Paris. Li, B. 1987. Flora " Gulinqing Nature Reserve in south- eastern э, Unpublished M.S. Thesis, Zhongshan University. [In Chinese. Li, S. K., S. F. Yuan, L. F. Liu & Z. Z. Chen. 1986. The flora of Huapin. n Reports on the Huapin Forest Area in Guangxi. Shandong Science Press, Jinan. [In Chi- nese. Li, Y. H. (editor). 1996. List of Plants in rdum uae Yunnan National Press, oe D, Chinese.] Liang, C. F., J. Y. Liang & L. F 985. A report on the exploration of the flora of Logan Guihaia 5(3): 191—209. [1 ` Chinese with т abstract. | Liu, L. H. 198 . Rain forest. /n: "Wa (editor) Veg- etation of та Science Ча ijing. [In Chinese.] Liu, L., - Hu, Y. C. Yang, W. W. Liu & R. X. Guo гету 1990. Reports on Land and Ee 'onomy of Xish- uangbanna. Yunnan People's Press, Kunming. [In Chi- nese . Ye, G. Zhang & H. Chen. 1994. Floristic wind ysis of the Los natural reserve. Acta Bot. Aus sinica 9: 1— n Chinese with English abstract và Raunkiaer, C E n The Life Forms of Plants and Statis- tical Plant Geography Oxford Univ. Press, Oxford. Ridley, Н. 7. The Flora of Malay Peninsula, I-V. L. Reeve, london Lis gat Se himper, A. F. W. . Plant-Geography upon a Phys- iological Hee Oger! Univ. Press, Oxford. Shimizu T. 1964. Studies on the limestone flora of Japan and Taiwan. Pas II. J. Fac. Textile Sci. Technol. Shinsu Dose A 12: 1-88 Smitinand, T. 1966. The vegetation of Dao Chiengdao, a limesto Chie nema, north Thailand. Nat. ist. *. 21(1-2): 93-128. Thin, N. N. 1997 The vege tation. Ed Cucphuong national park, Vietnam. Sida 17: 719 пе massive in stone in Xishuangbanna, dl China. Guihaia 101-117. [In C pw with English abstract. Whitmore, T. C. 4. Tropical Rain Forests of the Far East, 2nd ed. С l ме ө Press, Oxford. Wu, C. Y. (editor). 1980. Vegetation of China, pp. 363- 397. = Pig Press, Beijing. [In Chinese.] . The areal-types of € و‎ penen of seed plants. re Г Bot. Yunnan Suppl. IV: 1-1: Wu, T. L. (editor). 1994. A € На of Flowering Plants of Islands and Reefs of Hainan and E Prov- Science Press, Beijing thir Xu, Y. C., H. Q. Jiang & F. Quan (editor. 1987. Reports on the Nature Reserve of Xishuangbanna. Yunnan Sci. estone Flora of Dongyang a "China. Unpublished M.S. Thes fione Шор. [In Chinese.] Zhu, H. Tropical rain fore ast vegetation in Xish- tanum. C hin, Geogr. Sci. 2: 64-73. . 19 7. Ecologic ^al and biogeographic al studies on the tropic al rain forest of south Yunnan, SW China with a special reference to its relationship with rain forests of pial iie : Biogeogr. 24: 647—602. .H. 3. G. Li & Z. F. Xu. 1996. A phyto- КЕБУ d research on forest flora of анин hills in Xishuangbanna. Guihaia 16: 317—330. [In Chinese with English abstract.] 1997. Floristic re- lationships be wee n the limestone flora of Xishuangban- Asia and south a Bot. Yunn. 19: 357-365. [In Chinese with English ae | H. Wang & B. G. Li. 1998a. The Structure, Spe- cies Composition and Diversity of the Limestone Veg- etation in pa shuangbanna, SW China. Gard. Bull. Sin- gapore 50: 5-33. ^ . H. Wang, B. G. Li & Z. F. Xu. 1998b. Research on the tropical seasonal rainforest of Xishuangbanna, south Yunnan. Guihaia 18: 371—384. [In Chinese with English abstract.] Zhu et al. 453 Volume 90, Number 3 2003 Southern Yunnan Limestone Flora (NAM *ed&iopoy) gozo Puy iL М noq) "M а $ ue) 7] 7A stusofipios гел тә] 1121280 (онин) &©14 (OMLIH) 221№ эр-опд) ою :(8.11H) 122E ту-иод 17 (NAY *ed&o[ou) Fg20I-6S 179 12) Suas], 7) "X umdapoowosnu sadiq ЯҮЯЭУЧЯаІа (NAM *ed&io[ou) 26201-68 1-2uaus ted :C)g.LIH) #082 (ОЯН) F18 (OLIH) 2122 myu-up 17 ТНУ А `2 270919709 роошү ЯҮЯЭУІТЯИ ansH ‘Н ‘Н Sisuaupuund wnusadsosai (ОЯН) SS9ET »p-on» ov] ‘(NANY ‘edAlopoy) 7 [7€ “(Он ИН) O€BEE muy-upy 17 : ənsH ‘Н `H sısuəunjĝuəu umuLiadso4214d (O&.LIH ‘dhos: ‘NAY *ed&jo[ou) pgop 1ту-ирд 17 (SAS) BZEYD ЧУЯОУЛПҮЙ (ОЯН *9d4iost ‘NAY ‘edAjojou) 82007 1My-ung 17 SuenH “Н а Jy Зиед 7D "A Sisuappsuaw D38103 N (NAM *ed&io[ou) оре (IALIH) TEZ 1u-uvy 17 Is], q ^H smsuaunjguoui тел pnsnopu vapur] JAV3OVHOVTI (Онин dos) 2001-66 1-Suays tad пж ие оошо ЯУЯЭҮІНЯМЅЯЭ (NAM "9d4yopou) OFOOT-6¢ 1Ї-®иәц 124 Sueng sisuanau& sno4an() AVAIVIVA suey’) |], “A 11] uowajsouosuy Suey’) |], “A 1mqupÁ& u01017) = NAM `әЧАзоүоц) 97 ep 1mj-upg v] (28.LIH ‘dost ‘NAY ‘edMopoy) zgego my-ungx 17 (SAS) 060140 aVaADVIGdOHd 13 (DS.LIH *ad&jost ‘NOM *edÁjo[ou) CIcp 1т{-ио 1] ТМ CH 2121270270 so1Kdsow] ЯУЯЭУМЯЯЯ (NAM "9d&io[og) 7pg9[ 1-#иәц 124 rq `H Sisuappsuaw pa400s01(] AVAIVANOISOIG (NAM *edAopou) ezr 1туү-ио 17 (OLIH) Zcore Fuayo-unyy Fuvyz NS д 'Н $P Burg а № wngjajdojopd лел пету unigup]y AVADVISNVIV (млм ‘edAjost ‘HDA *edÁjo[ou) eege guauz-Supn*) әд ӘД `7 `©) sisuan]duaui DUIMA AVAIVAVOV (SAS ‘adAqojoy) 2611 PNH тул nu^? ‘Н x Bueyy 7L H 21270212019117 u smjoponjdpg AVAIVHLNVOV ueuuny UIIYINOS 0] әгшәриә pue syeyiqey 3u0]s23uit| 0] 9AISn[oxo Exe] ju?[d рээ (дәм Mi АЕА! PDAS SIN ‘i ыы #шрзоәәр әле suıadsor3ue jo вәшви Aqrurej au "([ui'iseA/uo1e35/] € M/2100qour310qoui//:diu) SODIdOULL pue (8101г: ин. INd] ut pue рәвтлә1 4риәдәг souo 10у jdooxo soureu ogroedseagur pue ojroads ло} ao1nos әлцрүлоцупр әц sr әротитс̧ SsuD mdog 2021]qnd1ay 010] "ueuunyg шәцпоѕ ut иоце}әәл 2uojsautq әці JO sdnois satsads [e2130[00H `q xipuaddy ) әйдәү ÁueSo[Auq unrsdsorguy ay) 01 Annals of the 454 Botanical Garden issouri M (ОНИН) 9226 “(OMLIH) 0055 MY-UDK YT (SAS) EZE4D (SAS) 6247 (OS.LIH) 8006 “(OMLIH) 2006 ?p-on*) on (ОНИН) 10201 vomnpedxy (ЭНИН) 99191 75 12 vp-ony ор, O&S.LIH) SZZP My-uv4 т] ~ (SAS) 02842 (SAS) $1249 (SAS) Z08492 (SAS) r84) (SAS) 60649 (SAS) SPE?) (ОНИН) 91? 2uog Buv P DNH NYZ (SAS) £8642 (SAS) F184) (SAS) 9BEYD (SAS) S1842 (DALIH) 0197 (OALIH) £F01 mg-up т] (SAS) 9649 (ОНИН) OS9ET vp-on:) ору ЯЛІН) F£001 1ЇЎиәц 124 (ОНИН) 2621 2uojg Funy (OS.LIH) ZIZI ?p-onz) ору ЯЛІН) SHOOT :/-2u2us 124 (DALIH) £I97E uonpodxs (DALIH) £c6r£ 'OSLIH) OZIZE vompedxj (O9 LH) 82121 #иоц-#иәцу Fung (ОЯН) 091£ (IALIH) 661g 2Zuog Funy (ОНИН) 116 O8LIH) 061€ my-un тү (SAS) 981147) (SAS) ESEYD (ОНИН) Z62£ vp-on:-onj (SAS) ZIZTYD *(S AS) CLITYD (SAS) E749 (SAS) 612149 (SAS) £61149 “AY ритиипр 112147] 3V3409VNID9VHOH TEA uisu xpquiog ЯҮЯОУЭУЧИАОЯ Wouos рта pijpsnuiaM asnely ^W y [аи snmund smyupnjnuo-) [Pu] sisuaupuunt snjjpydoydiowy AVAOVUV BUPRIS] asUaUDUIDY DiugiJsouiso-) r] Ld 9 dueisy, asuauaidzuid ‘1P۸ әиујр uo1puapouy uo(] `©) (хоу) ojpununop musoupzy AVADVNADOdYV “Ə 2770]272d отоо] 18ү хә s[|ot(q VƏDA naoudoanpy шоці, 5 "J OOH punga snappy 'douse«) x jourq psoumnbs Dosuoud]y ичп $1]]ош nasuoud]y 'dousg«) Ņ July риртиод pnosuoud]y AVADVNONNV йш “LIL X nA Д 9 (918179) umapnf uo4puapooixo[ SUW "LOL POM A 9 ттлт “IBA sisu2uosp; трио "ugoug1jg xo "sstog niofruupuiuom DIJD]SId 3VADVIGHVOVNV хоу 240j1u49] DuapDona(] "douse«) хә әлә] DUDIPOQUIDI DUIDIDIG] AVADVAVOV "JIEM 22412909 niZioqunu | чиш ^W ^W хә әзер cg 7 г) styonisorsd sauuppiqoas эшэ (saaN xa "ew ) ур] SNYJUDIDIA} d] '"siopuy ‘ү snuoZpaja] smpupoppppon(] nu? ‘н (ourodg) D40yfiya2u1012 "PA D40jfiXD] snidpi3oapuy JVAOVHINVOV ивиипд uIəynos 0} эшәриә you mq syeyiqey эио}ѕәші] 0] 3Aisn[oxo EXE], UP [qd рээ (ОЯН “9441081 ‘NAY *ed&jo[ou) EEG] рр-от) ор Oe] "(T 4) sisuanjguoui Dunu21&qq AVAOVAHL 'penunuo) °[ xtpueddy Zhu et al. 455 Volume 90, Number 3 2003 Southern Yunnan Limestone Flora (SAS) €SEYD (SAS) 6£4D (SAS) Z9EYD (SAS) C184) (SAS) €9SYD (SAS) 62149 (DALIH) 6262 'O8.LIH) S60£ my-ui 17 (ОЯН) 679E mu-up т] (OLIH) 672£ My-udK 1] (O8.LIH) 6£2& (OALIH) 8126 17-и 1] (OLIH) $866 1Ї-®иәц 124 (DALIH) Zz1£ 2uog Fung (SAS) 16242 (OALIH) 9202 14-и 17 (SAS) £2749 (SAS) Тё6ЧЭ (SAS) €F8'1) (ОНАН) 2727 muy-uvy 1] (SAS) 61647 (SAS) GEE1YD (OS.LIH) 22901 142-19 "YZ (SAS) 69FYD (ОЧ АН) €€9c Fu0H zung (SAS) 0ZITUO (ОЯН) рөө vp-on*) оюу (OGL IH) 082€ my-upy r] (ОНИН) €CLE008A (ОЯН) 0962 UPnb-opyg zuo], (OMLIH) [PIE m4-um 7] (ОЯН) 6251 Зчон Fung (SAS) PEYI (Оян) 621 Зон Buy (Og. LIH) 9822 MY-UPA 17 (SAS) E9OTYD (SAS) STOTYD (OQLIH) &29£1 PP-0N9 Ov], (OELIH) SSF 174-0 17 (SAS) [2847 (ОНИН) 769£I (OALIH) 68261 PP-END on] douse«) snupiuoq uouia]souo n] Joxoog 1244079 smpuny Ayd aun, g 827024510 U01047) "Siy-q[onjy (PIN) smupagpums snyruns121) AVEAOVIGNOHd 14 ‘sway nuuvu s3jodp3y əş 7g `2 119401 sojodpz y suea "^A Dorupuumnq sajodp3y 3V39VORTA [Dmg 9 РІД D4251n2D1u9] 2400501(] [mang o шед vasaaynd D2402s01(] [ping 9 ureaq 222u1/2 D2402501(] [peng 3 шед vssadsp ә102801(] [Pang 9 ULI Dupryop4p n2400s01] 3VA4)9V1HODSOIG чиш "M "M 21091903 aununpav;) AVYAAD OHIO səjdelg м `9) qno?) у ‘dausey) 14papJap "лел sisuauis ртшрихриј, so|delg ^A °9 (SWH) szsuauzs DIWDUAPUT "qxoy Dsoua2n1 DUD10 ‘хоу »7Djnoiund Dun40d AVAIVTIATOANOD повцә1п Y query 718402 D1u0u424 qus `Э (T) popu njosd&o04214 AVLISOdWOD ем DIDUIWNID snssiogouy JTVAOVLLAWUgSNOD quiuis 77 ^y psopmos n42ounjstuq uius °) 'y (21914) DUDIPOQUID) D42UISH4q yg °) ^y (nH) sisuaupuun& pjJotu3uasacq чиш 77) y риртрләш 0]]2142u28207] AVAOVULSV TAD Agr] 112111p0q єр) AVAOVGTHVddvV) jdeis SUDIDA n3202004D0Q ЯУЯЭУХПЯ penunuo? "| xipuadd V Annals of the 456 Missouri Botanical Garden (OLIH) c£8 Онин) 7£9 2uog Fury ‘(OALIH) 996g 1пү-ию 17 (ОНИН) ££€rg uompodx3 (ОНИН) 0I£r£ uounpodx (O&8.LIH) 20 'COg8.LIH) 6201 тЧ-ищ 17 (O&.LIH) ZS8F (ОЯН) 886Z mqg-up ту (Онин) "ws ni-Zuouz Fung *C)H1IH) 662€ mu-uot 17 81IH) 6021166 13-ovg 17 (O8.LIH) 8626] np-on?) OY] (DALIH) 6S9FT un-Zuif m) (OS.LIH) 8101 1-2ueug tag (O8.LIH) 28111 Dp-ons) on (O&8.LIH) £18£ (ALIH) £0rf muy-u» т] (ОНИН) 69001-6¢ -BUŞ 124 *(O8.LIH) €z9cg чопрәйху (ОНИН) 06261 vp-on») on] (O8.41H) 2666-6 -®иәце tag ASKS) 90849 (SAS) P2947) (ОНИН) #801 1-Suous 124 (ОНИН) 20681 Dp-on9 on] t8. LIH) 2686 18иәцс̧ 124 (SKS) 2F6u) (O&.LIH) 222€ (OMLIH) 29002 14-чу т] (SAS) 61912 (ОНИН) 222% “(OMLIH) 2082 my-ue 17 (SAS) 64) (SAS) FOEYD (O&8.LIH) 0017 їпү-ир т] (SAS) POSYD (SAS) 86£4D (ОЯН) 0027 1m-up, т] (SAS) 81149 (SAS) 260142 (ОНИН) TIPI "4-и 17 < [ng $72704d штшләйѕол2ә 4 9nsH 9 nA Д `D asuadunjduno шпш-әб$олә1,] ЯУЯЭҰАТУИ umg рори отріотиәу ЯУЯЭУМІП ITM CH ?1ofumnop n&an20)]d a7) TVAOVHIVI 'qxoy umqpaonjoaut Dui2401duikg ЯУЯЭМІИУТ TT MH УМ А `7 огорхоу лел x&qpoono Disopqoy TEA штлоріалоа ошшәјѕоцішох) Buen 7D) "A NA Д 7D) smprupipjupx snajo?) ЯҮІМІЯУТП "una[s (еМ) огрирләј DapuDydwoy 3IVAOVNIOVOI ТІСНА X9 NM 7A 7D) 0jD2]2D4Q рти) JTVHVUXALLLOS 91817) Muay Daoqoumua() uo( ‘q 727201195 snjouois«] "UIS CA y anyangan sndapoout&pi(] mung "T^g C^vT ^H) $70/11022Q sndap2onydpny JTVAIOVBIANSH1S Зирци "X (quei) sisuaumis “rea sisuoip] sipppKaor) ЯУЯЭУІЧУИЛЯ "dause«) sisuaurquo; psoydog MOH ни ST "gpuog ( л) DUliSstD40p0 от ЯУЯЭҮЯҮЯ 'penuguo) ‘ү xtpueddy Zhu et al. 457 Volume 90, Number 3 2003 Southern Yunnan Limestone Flora (OS.LIH) 0526 174-ирд 17 (SAS) OSSYD (28.LIH) 62F01 (OALIH) 81701 142-194 "Y7 (SAS) 66€4D (ОЯН) €766£ PP-0n9 ою (ОЯН) FIZE ттү-ир 17 (SAS) OSBYD (SAS) 86249 (OLIH) ZOEE muy-unjg 17 (SAS) 05640 (O&8.LIH) 65681 FuvA-uvg uat) (ОНИН) Z0IFF vp-on:) on] (OLIH) OSIF 114-04 1] (SAS) 8STYD (Оян) £899 142-124 "47 (SAS) ZZ49 (SAS) 62949 (оян) 1607 'OS8LIB) [OIG #чоң чоц (IALIH) 0FF9I PP-END on] (OQLIH) OZH 11-0 т] (D&8.LIH) 0Z£cr vp-onz) on] (Оян) 0961 #иоң Suny :(()8.LIH) 6251 vp-on2-ov] ‘(OALIH) c£00c 14-чу 17 (Онин) Z29ST »p-onz) on] (9911Н) 86887 vp-on:) от (SAS) 10612 (оян) ZZEE muy-upg v] (SAS) Z6TYD (SAS) [SEY (OMLIH) SIZI O&LIH) 0141 £uog дио (SAS) 6ZSYI (SAS) 6££42 noq) "w d $ чәч? "T A Dppj20u2141 чел рзорподроо OsD]NVUAA "zv[N-pueH D40jfixv] n1124220$ Udy’) "| 'A хә nA д 79 (8чәцо у ng) sisuagupignig тел ррләор D1u200H ‘bry »orupapf piupnoy ЯУЯЭУМИУНЯ гәцәләү 11324402 vn[opuns4T 3VAOV'IMATHd шәу (‘Эа `9) оиорофшоә ел DAYIDISO}da] nruo4adaq ЯУЯЭУЧННаІа "| Suan 004107) AVWIVd pur] s1su2u142 D10p1104f “YPS osuagunjgupmy umpadoqmg “YS (2»ueg) niso4quin um]j&udoqmg 3VA9VIIHOHO "ze|y-pueg (чиш “дү 7A) unupi “IBA әвиәитѕ штит AVAOVATO ziny Dyofapissaquny? sisdo]p1s]DA] AVAIVOVIO "еМ 02024ә57шәѕ aUISIA ур AVAOVNISHAN 01$ ‘Н әрә ртируйә1 от `5 `Н sisuagunjgunjo viuvidojg "douea sisuauanyamns DIÁ) ЧҮЯОУЙИНЯА4 МАЙ suey Muay oruodun]y NA ‘A `” (әнә) оуозәарлјәј Dioowy DA CA 7D suomonsa, 08ү AVAOVITAW 'penunuo) "т xipueddy Botanical Garden issouri Annals of the M 458 (O&.LIH) 760FF PP-0n9 ov, (Эрн) £2001 1иәцс 124 (SAS) 68140 (O&.LIH) 222 (OLIH) OOZE My-uD{ т] (ОНИН) ZO9ET Noy-umf FuvyZ (SAS) S264) (SAS) 1642 L PEL (SAS) 16649 (SAS) 01242 (SAS) gceu» (ОНАН) 7€2£ mu-u»i 17 (SAS) 2947) (SAS) FISU) (SAS) 61242) (SAS) 22649 (SAS) 21649 (OG.LIH) 96£ 27-и т] (ОНИН) 6798 142-194 NYZ (Og. LIH) ОТР 14-и TT (SAS) 62842 (SAS) 22812 IZEYD (SAS) ZEYD 2 (Онин) [IEF (OALIH) 16E MY-UD4 17 (SAS) 68014) (SAS (O&.LIH) Z02£1 noy-uvif Bunyz (DALIH) £6£ 14-D 17 (ЭНИН) 502? muy-uvg т] (Онин) 2965 (OALIH) 595 myg-uig 17 (SAS) 9642 (ОНИН) 6927 'COW8LIH) €19g mu-up 17 (OS.LIH) £Z9ct uompodx (DALIH) 2221 »p-on) op, (DALIH) 60££ (OALIH) 912E my-um Т] (OLIH) /28£ 'OgLIH) OFZP myg-ub т] (ОНИН) LIF O8.LIH) £221 my-uv 17 (SAS) 02242 (SAS) досу? (ОНИН) [S9ET Dp-on4 ony (SAS) 91849 (SAS) 19847 SULY auojsaur] О] jua19j21d exe[ диву P226 "UA 2/22U42D] D21/2$242utu041 TVIOVHAHIONIZ "Qpue[q adsodna niuzilspaja[ ‘Чәче (uosweT) umigqnp puuzispajo ‘douses) xo әшәгд umupipoquipo Duigspala[ AVAOVLIA 194219] SISUNUDIS 11124404) JTVAOVNHISWVTA Чо} ^q D) ?1Du242 st42044 utune[[m^) зу *douge«) 12722pp2 рәр "gueg n4qpos pniuoin]Jad ‘douse punssiuo4n poj10di] BURN `L ^M saproruza]ps DuiaJs0]0]7] ejesep ummofroopq4ou Dui2150]0]] IVADVOLLH 1 ouipedaos 74 ("uoue[q) srsuaddipryd авл nin $1172 JVAOVWTI qIP1) njsnqoa piuidan AVAOVWIAHdV.LS ‘TIEN Sou рәт AVAOVEATHVNIS pug С1рчә$ FP ure?) szsuauiddipd vizoquapur] AVAOVBIVTI1HdOH)DS Suenp DIUDI} DADLIN I, Р]РАРН pyofiysaryonea DADA JA AVADVLIY "Nouisag ю1]оЙлфрә® njjoind ag ^W Pylundjod ngoang u2) °) M хә моң snaogi3uo] “IBA snoipispipd uogodououicg] JUUA snorpur smympopuum(] TVAOVIV MH 'penunuo? "р xipuoddy Zhu et al. 459 Volume 90, Number 3 2003 Southern Yunnan Limestone Flora (O&.LIH) 9F6E 1nu-ub 17 | (SAS) 92649 (ОНИН) F9£0I 142-94 nuz (ОНИН) 92691 PP-0N9 ор (SAS) S9ZYD (SAS) 621142 (SAS) 90242 (O&.LIH) 222 2f 17 (SAS) 091 [42 (Оян) €Z16 vp-on:) ov] (SKS) 691142 (O&.LIH) 72201 1-2uaug 124 (SAS) ОРЄЧЭ (SAS) ££cu2 (SAS) 9F9Y4D (SAS) 28€42 (SAS) FFSV) (Оч ИН) 1c9c£ чотрәйху ‘(SAS) 8EFYD (OLIH) S61 FF Dp-on+) OD] (OGLIH) #26 "uy-unq TT (SAS) 09642 (OS.LIH) 29981 Noy-unip дируу (Og.LIH) PLESE vp-on:) ор, (SAS) [6947 (DALIH) 92601 Fuoy-Busyz7 Вир (SAS) 91247 (SAS) 6947) (SAS) 98447) (0g d IH) £cSF£ иотпрәах Я (SAS) 62/47 (SAS) £249 SAS) #9249 (SAS) CFCUO (SAS) 869492 — (OLIH) 66729 17-и 17 (SAS) OZFYD (SAS) €Z94D (ОЯН) 6£1£ 14-и 17 (OALIH) 81£F2 (DALIH) ZI£rZ 1102 итоиуир) (SAS) TIZIYI (SAS) 66142 (SAS) Z81149 (O&.LIH) 6282 1my-u»y 17 (SAS) ZFZYD (SAS) 9847 (OLIH) [IOI 1myj-uvi 17 (SAS) 1642 (SAS) ОТЕЧО (SAS) 08147 (SAS) GZIU2 (SAS) ZETYO (O&.LIH) SZZg »p-on2 on] (SAS) I6IIV2 (OLIH) 868% 1uy-ung 1] (O&g.LIH) 122g my-ung 17 (SAS) 2642 (SAS) 8084) (SAS) 9ZEYD "] 29192 xpquog AVADVOVEWNOU ‘yosully pnjDjso4d nruo3ag "osuu[ sippnéap piuosog ‘[suaH 72u18D2np nıuotag AVAIVINODAG Hoyos sisuastuoysuoy vsoydopiydoyy поцәс (‘qxoy) vasməap Dsoydoprydvyy у чооң 0ә7и0219 »i$020]07) Pug wnupaiaid ошәиот]8ү AVAIVAV тгәцдә(] (әшпү) snsogpma sndapooxo[ поор (хоу) punssionua; Diuaps1D JN TL $ SYANS "qp OA (FT JL d 9 PUISI) sipmpxps sayruvunusof 'utuejso^) ю$о]]14 nXog r] ` d $ 80815, psoau2u nXog A91 20 DAOH чя YH (J "p »souio vAoyy 131M хә TEMA umofim] puiguu-) "pnuqos (29H) 2478914 ошәишлх) шәр ([YBA) гоштш D1p14281([ Suvits] (“лә”) njoambso nipuyosw r[ | d Queis) esuaunuund nunpig "pd CUIU d» uen) sisuozupsjup] nujojsonuy JIV30VNADOdV out 7724011 D101do1n]y шоці, 59 "у "ооң 2£n2urpui 1201d241]y Buey ^L M nuny? sn AVAIVNONNV Zueny `7) 7L Jy uersg 4 7 (еее) 7021ds12u0] оХуәрјѕоишәс̧ ny д 7D (uag ^y) оррипдодоа vsodsopidnyy puy umnjjayomd uma пери” (aum[g) n40jfixp] snjdpagoupuy JTVAOVHINVOV 'penuguo) ст xtpueddy Annals of the 460 Missouri Botanical Garden (ОНИН) c62£ ''OSLIH) 8291 »p-on? ору (OALIH) OSEPE uompodxg (SAS) 2224) (ОНИН) 202Z 2uog Зиру (SAS) F847) (ОЯН) #222 Fuoy д (SAS) 6257 (SAS) 92647 P (SAS) P8247) (OLIH) 26722 Fuvn-1y¢g onyz (ОЧ ИН) 999£] рр-от‹) on] (OS.LIH) OOST 14-и т] (SKS) 2642 (SAS) S747) (ЭНИН) £crc my-uvg тү ASKS) TETYD (ЭНИН) #691 14-и 17 (SAS) BEYD (SAS) 9274) (SAS) 60£42 (ОНИН) 66£ 1mj-uvq тт (SAS) 22647 (O&8.LIH) OSZP ту-и т] . (ОНИН) с/с Mmy-ung т (OMLIH) 9SE un&-Zupmyjs noz (SAS) 20942 (ОНИН) 00£01 1-9иәцс 124 :()g,LIH) 98971 un&-2uif n) (SAS) HIZINI (SAS) 91€42 (ОНИН) PZT PP-OND ою (SAS) 9z£u2 (SAS) 10214) (O&8.LIH) 292 1102 unouyuy (GLI) 2666 W-Fuays 124 (SAS) 72064) (ОЯ ИН) /92r£ uompodxg (ОНИН) OZPOT 142-94 "YZ *COSLIH) SETH їпү-ир 17 (O&.LIH) 2112 тЧ-ио 17 (SAS) S6142 (SAS) ESTYD (SAS) 98142 (SAS) £947) (OMLIH) 2822 Fupn-1yg оруу (SAS) 21947 (O&.LIH) 92Z€ '(OS.LIH) 09gg my-up т] DA ‘д 72 (deis) vyAydosopw °3 sisuouis vaoqvsvg mng “Т “gq (29ueg) Dınauo131p »20qpaDq mng 7T ^g (Yyoursy) suaosafns n20qpnind 79 Ч PIOIpIURP onay) JTVIOVBIPINSH5 OW Y "[[OZ шоута Dunoon) 4 AVAOVILYIOOVL uun(q (ичп) sadi uospuaporpsofiny m ‘иеше sssuaupuun& Dna jy qeg nxajnua2 piunmng AVAOVEVA quis ‘f ^f (sung) suvog sisdoranqung пу CSCH (9N) »ipjjaouoaopnosd ‘1P۸ 14200102 0320201817] "Jod "N y xeq штур]оцә@йаәлф u01pص12]7)‎ `ЧәиЗрс) umsoa]2D4q uoip13]7) uueujjoH q (sway) un74do1onu iea шпируиош nusapiuuy “Ye sisuajaypis nuapguoouapy AVAOVIGUOHd 14 “SIUM. 9 1әрцә}р sisuaupuun sostdsorg AVADVNAGA эа CT) 9»?iuodpf vi2unoq AVLISOdWOD SuoH “A '(] suapua?s napup4oq AVAIVNITANWOD "4sseq wngojanbuinb wumaajduag IVA3OVOBIATLdOIQHUV) Japyey Pups) шпитдд AVAOVITOATHd VD unu? 4 P1744doın suupddp?) ziny отојиріфшәш таар” A1V459Vanivddvo "suuof ‘W “| 222ups] 012143 JV39VYNIOVHOS 'penuguo) `q xipuaddy Zhu et al. 461 Volume 90, Number 3 2003 Southern Yunnan Limestone Flora (O&8.LIH) 2222 чон Suny (Og. LIH) 967 OSLIH) УРЕ 10? umouyu] ЭН (ОЯН) 299g 1my-u»y 17 (SAS) ZFYD (SAS) 02142 (SAS) 0£142 (DALIH) £cc£ 1nu-uvg v] (OS.LIH) 11-88 Zuog Sung ў оту туў (OALIH) 2085 114-00 17 (SAS) TEYD (NOM) SETE my-uv 1] (SAS) 18649 (SAS) 99540 (SAS) 08912 (SAS) OSZ 1102 итоиҹиг] (Оян) 86922 un&-Zuif 117) (SAS) 80842 (OLIH) S8E 1muy-upg "T (ОЯН) peg un&-Zupnyg nog (OLIH) [SOI 170-00 17 (SAS) SFY?) (онин) ZOLE #иоң Fung Sg LIH) 1269 Fuoy-uayz Funk XDALIH) [6Р6 uoupodxg (SAS) 712492 (ОНАН) PS2 17-и IT (SAS) PEEHI (SAS) 1949 (SAS) T9149 (SAS) 61142 (SAS) 69247) (ОНИН) £€0£ mu-u»i 17 (SAS) #98492 (OLIH) 219% чотрәйхӯ (SAS) 82749 (SAS) 22142 9226 иотрәйху (O&8.LIH) 0F6c (OALIH) 262 “(OMLIH) 1262 Fu0H Fung — (ОЯН) 862£ Pp-ony on] (O&.LIH) $0081 Fury-1f ung *()g.LIH) ZSZ mu-upy 17 (SAS) 1647 (DALIH) 06991 np-on*) on] (Онин) [6181 Zuvi-if. ung (SAS) $0142 19Q9]9[4 224724402 visip4y JTVIOVNISHANW 19u107) (Sury) Ta гел D4a21qnd sn21 4] Suey) TR КОШУУ piuDApn?) AVADVUONW "zve[y-pueg ("qQouez4) suaosaqnd стел D01719 Duoo[ vr] "^H umjpoids um] xoskqq ITH Ņ DA A `Э ummgponuo] umj&xosA(T BULY DUNJA “IRA DiuDpnqn] vispaqnu) TVAOVI TAI 'qxoy рѕојра ю1үпәлә1 AVAOVATVIA uau? ^N `$ (doq) sisuourquoj Tea sisuajpuguaq agpjdig AVAOVIBVIC TVA "учоон 272m олут, yluny xo əjíoy wnypjound umjpuozjog AV3OVEIUI SƏƏN (SAAN) хә Түү) D70]022un] aqaoyd ‚чен Ņ ‘Yyluag (SPAN хә "[[EM) Djnguoja рәх Bueny н d $ P4 "A d ?7ofnuejpp vosir] dAVADOVH TV челт) ош auisapouaudg log 1 Gua uospuapo.ay’) AVAOVINV'T DA A 7D (qurq) sapodyoois 121104511 ‘wg D1]0fiisoddo 023004q092]07) AVLVIGV'! шү хә ләләр] “ч оллртипр so1&pody AVAIVNIOVOI saaN (qXOY) 8221115 smunjpoo4puo(] AVANIAVUD 'penuguo) "т xipueddy Annals of the 462 Missouri Botanical Garden ( (SAS) 22212 (SAS) 97947 (SAS) 82647 (SAS) SZPY) (ОНИН) 0101 71102 итоиҹиг) (SAS) 28847 (OLIH) STEE my-ung т] (ОНИН) 9SOpp рр-от) Ov] :()8.LIH) Z9crg uoupodxg ЭНИН) 6981 1j-uiy 17 (SAS) ETEYD (SAS) 2664) (SAS) O€EYD (SAS) 92142 (ОНИН) £££ un&-Zupm[s noz (SAS) OFITYD (SAS) 22247 (ОНИН) 6921 1пү-ию т] (O@LIH) 8S7 OS.LIH) £Zrc MY-UD 17 (SAS) 96947 (ОНИН) 86r] їтү-ирд 17 (SAS) COTYD (SAS) 2124) (ОЯН) F00 Zuog SR wong NYZ (SKS) SEY?) (ОНИН) F182 (OALIH) 9rc my-ue т] (SAS) 62274) (SAS) 812197 (SAS) 16247 'COgLIH) 681 174-и т] (SAS) S94) GAS) 880142 | (OMLIH) 11822 un&-Zuif my (OQ.LIH) €99 1102 umourtu;) (SAS) QIZIY (SAS) FZHYD (SAS) SBSYD (OMLIH) 8£2c muy-uvq 17 (SKS) FOBYD (ОНИН) 62029 nq иә) (OS.LIH) POOR INY-UDK 17 (OLIH) 62ГЕ muy-uo 17 (ОНАН) S97 Zurix-njg оруу (ЭНИН) STITI unb-i2q дир (O81LIH) cog 1пу-июд т] (ОЯН) 8£7£ 1mj-u»q 17 (SAS) £1249 (ALIH) EZZ mu-ur, 17 РЕ SAS) Z6FYD (SAS) TOSH 3ueng 77) 77) apum um Axoijjupg axel] штјәру штуХхоцјиру 3ueng 7) 77) (uospA ^H 7 $ pyy) кшш ‘IRA UIT DULID unj&xouupg xoef CT) оторплира охоту a[3UMG (unu?) X шәр) оурСудолэтш охоту "Sua1dg (Т) пио олоту "IPUI CPIM) 2170/2019 suso2£q) JTVAOVLOH oJ pstawunqns J psondsənə PIOS "Youradg sisuaguos2uoi onang IMO] DSOUIID njjoand шәр D40udojpida» юлох] AVADVIJ HH әллә] sisuadoyp snydiz7 поцо “MY а $ U2U) 7] "A sua»saqnd лвл Dd4)2012] 080111124 AVADVNAVHY эа 7 л чәй Ud 9 7021104» чәй, ың pupau. (oy pnuosadod AVAOVUAd!d 3oueH DIPUDIYIO D4040) AVWTVd "Agr] njoambsa wnicydorg "] »joqumapo Douay AVAOVATTVXO "] papur тцәш8әү ЯУЯОУНЭОМУЧОЧО əunjg (три) psomzZunp nipidoaj "uus (puy) vypoyd mpaiəy әдип (рир) snnbijqo snqyn[204]sp*) ЯУЯЭҮаІНЭНО Uenon "IN ‘f $P "IN у хә пем Cqxoyp umzojfianpo штӮ 2с AVAOVLYAW ‘penuyuoy `p xipuaddy Zhu et al. 463 Volume 90, Number 3 2003 Southern Yunnan Limestone Flora (OA.LIH) £09 172-009 17 (O@LIH) 8998 102-190 пуу (OMLIH) FEZ 12-009 17 (O&.LIH) PSSZ muy-uvy т] (ОНИН) OF PPE uompodx (OMLIH) 2798 102-190 "17 :()8.1IH) 799£1 Pp-ony ov] OS.LIH) 999r 17ү-ирд 17 (ОНИН) 67091 Pp-ony оюу (Суң 11Н) 990£ Zuog Funy (SAS) OFZIYD — — (SAS) ӨОРЧО (SAS) PS9YD (ОНАН) ZPOL »p-onz) on] (SAS) 89949 (SAS) 91949 (SAS) £€912) (ОНАН) 6816 PP-END оз ASKS) FSO) (SAS) EOF) (SAS) 71849 ASAS) ZOSYD (ОНАН) €76/1 mu-upi 11 (SAS) 89749 (SAS) FF8492 (ОНИН) 6€£6 1F@uayç 1244 ASKS) FIZYD (SAS) B9FYD — (ЭЧАН) £6222 un&-zuif m) 06091 vp-on:) on] (Суң 11Н) OLZE ?[-uog IT (SAS) 25949 AS AS) 18847 (ОНИН) 7971 ®иоң Fung (s, (O&.LIH) 6FI Ячон Виз (SAS) 6669 (GAS) 88479 (SAS (ОНИН) [P917 »p-onz) on] (SAS) ТТЕЧЭ (OLIH) 91r61 un&-wif 157) ASKS) FOOTYD — (O@.LIH (ОЯН) 9691 174-и0д т] (OMLIH) 78001 1-9иәцс̧ 124 (SAS) ZE64D (SAS) 00242 eumgg um0fijmo wndAydosjuy AIVIOVAHAOLENV yereyy ^f unupisappui штјитрү 'ppeg Coop) 112402238 pa ‘хел штуррпрә шпитрү "т ummpnpo wnjuvipy "| suauaa-snjndpo umjupipy 3 V3) V.ILNVIGV syqey ouojsauri[ 0} juo19jo1d єтАцЧаорАлә, "unuog ^w (шед X Bury) sz]qD12ads рәцәѕәләшшо f Joyeg хә "үем umaojfimuoa] ‘лел штора wN1Y I pay] jdeis wmaanp-ouis umi5kpog 3VADVHMHIDNIZ 319 ў SPIA 0170/112 81014 "joue[|d рирәрѕирүрд SMA unu? 9 шәрі aasauignd muziuspaja 'douSe«) штшләйѕоиош IEA ѕиәэѕәфтлә DuiZi)spajag "dause«) 1(napjop muzispajo ‘Lay umaojfiympo оштра r] "] D JEP) s?]p1u2140-041$n 5. DNA "| pypnbav dae] AVAOVLIA шәр (wung) naum2ups тю1т]отпо,] 13u44 “HD лори рәр ттеу --риең 24ofipmuoui рәр 'ppow (Weg yong) 247s2dn4 ui21s017] 'ppoA Сәәпу җ p[oqetgS) ғулрә risnagauqo(q 'Pp9A Dun122u102 ni2un[oog ITVADOVOLLHU( pang | ("douSe«)) 11724011 Du0]07) qre1) (TEM) Ppunquoyf риојог) AVAOVTILL (ОНИН) 17<8 171j-uvg т] чя "M P4opfipnu sojoumaja AVAOVTAWVALAL (Оян) гов PnH "Y7 nu^ `0 `$ x Busy’) 7D) "A szsuaunuun niosidp] IVIOVWMIAHdVILS 'penunuo?) ү xipueddy Annals of the 464 Missouri Botanical Garden (OS.LIH) $22 тпЎ-орд r] (Онин) 86266 vp-on*) ony (ОНИН) 01091 Dp-on ор (ОНИН) órZer 172-009 ту (O8.LIH) 0£2€1 Pp-eny ony (DALIH) $8086 117-009 т] (ОЯН) &6cF£ uomnpadxy (ОНИН) 2228€ np-on) on] (O&.LIH) SEZE y-upy 11 (DALIH) 962 1n3-ong т] (ОНАН) 19121 ?p-0n9 ор (ОНАН) /Z€£f рр-опх) on] (ОЯН) £29 m3-ong т] (ОНАН) Z9F01 142-194 nuz (OALIH) ózZcr unf-2ui) r] (ОНИН) £€r£r Pp-eny оюу (ЭЯ11Н) 6£Zc mMy-uv4 r (ОНИН) 6£Z£ 1пү-июд т (O&.LIH) 088 1m3-opg т] (ЭНИН) 86£6 71102 unouu;) t(8LIH) ZEZ 119-009 т] OLIH) 018 mu-uv т] м. Sutu7) (epoyey) 100: q вло] AVADVIGOdATOd чәши], (Т) ртоупәітр sıdəjosydəy AVADVaGIda TOUHdAN xurT (7]) sounjawozp9 DUIWIDIFOLA AVAIVGLLINOWAH јә (12xeg) uopoyəp wnyosdjog IY) 77) (uoq `q) 07ә0]уәоә suo1doag AVADVGTHaA LdOAYG “ppeg (MS) opia ошик "us f (Ney xo azuny) 12umj40f орик] suyo ^H 1009 трик AVAOVIYVNAYG Sul") хә UY) 77) sisuauiquo] nifjouad] AVADVGILIATOd эшлә “OI “(ppeg) »7Dpidsno piynspauuny ашу (чоон) njmuutid-opifijpuuid pmipoiupjy duly?) (159?) DID] тротшорү ЯУЯЭУІЧАНІУ BUY’) (S114) `H) sapro&udoajup suadonoay емее (H9N хә ozuny) wuniosopəyə umiuo]dspuowukg A319 у "ооң хә "рем $uDizDa umiuajdsy "jsuasoy pjooixos umiuajdsy ооң umpzuojoad шттиәуаѕү иц) ^H ungoaliapi umiuajdsy [Seid `7) umsioxo umuajdsy Sul") asuaun2o4]snp umruajdsy AVADVINATdSV ‘panuquo) сү xipuaddy Zhu et al. 465 Volume 90, Number 3 2003 Southern Yunnan Limestone Flora (OMLIH) 9c£1 174-и т] (OALIH) 98086 118-000 т] (ОЯН) S8181 ?wvijaf ung (Онин) 2226 1l-Zuaig 124 (OLIH) 2892р unt-Zuif) 17 (ОНИН) OFIOL Mpunysuay (OLIH) 27201 "puvi12uopy (ОЧ ИН) 8Z [Sp m53-ong 17 (ОЯ ИН) 597096 m3-ong т] (ОНИН) 16 115-009 17 narıpıP], C T noinispapd snsosoqoA") suyo (ədoH) отара snJ08s0]2Á ") ЯУЯЭҮ Ld XTAH.L SueA “Нә Y UI) (ррәя) s7suaonuns D40123], ‘pedon ([s24q- 9) suauumoop 140122] Зирм ^H 7D) y Sun?) (ezuny) pxoaop sisdoqua7) AVIAOVHBIVLOAXL лә], (J umng) »10frmnuo] Disosopiay’) nA Ny SY suyo оѕоиитјорпәѕі si421dogma]y dV4AO9VaOnrdLdONIsS "urxe[y (A2314) зу ^XooH) njpnuiajnd IEE вис (MŞ) suoagoau1 njgouidvjag AVAOVTTANISV TAS suy ^H sapi042jdoiunon 84910] AVAOVGINaA Ld 'penunuo) '[ xtpueddy THE POPULATION SS > 2 i anguy Jaffré,* Peter Bernhardt,’ STRUCTURE AND FLORAL Vincenza Pontieri," Peter H. Weston,’ BIOLOGY OF AMBORELLA Dave Malloch, Hiroshi Azuma,? TRICHOPODA Sean W. Graham,'? Marc A. McPherson,'' (AMBORELLACEAE)! Hardeep S. Rai,? Rowan F. Sage, and Jean-Louis Dupre" ABSTRACT The shrubs and small trees of Amborella trichopoda are functionally unisexual and the populations are dioecious, male biased, and occur primarily in clumps. Floral size oe Teponed for this species was confirmed by differ- ences in floral biomass. At the level of the inflorescence, ther significantly greater numbers of male versus female flowers/inflorescence. No differences were observed сас iile and female plants in height, stem number, and diameter at the ground level. Male flowers bear 6 to 21 stamens and female flowers 3 to 6 spirally arranged carpels and staminodes that mimic the fertile androecia in male flowers. Flowering within a population was synchronous, pe flowers of Amborella trichopoda are both insect- and wind-pollinated. A wide variety of insects ranging in size from ca. 1 mm to 7 em in length pollinate the flowers, indicating a generalist pollination system. Beetles involved in pollination dwell in the forest litter but also spend hours on the leaves, flowers, and branches dd on pollen. Pollen i is the reward fori insects as there is an absence of detectable floral volatiles and nectars, and anthers lack secretions or food bodies. A free-flowing stigma secretion was occasionally present, but it was not consumed by pollinators. Structural studies indicate that the stigma is of the dry-type, and the pollinators probably visit female flowers because of the mimetic role of the staminodes. The combination of wind and insect ү оа exhibited in A. trichopoda is rare i basal angiosperms. Gall midges, parasitoid wasps, and thrips utilize floral lissue as a breeding site, impeding repro- duction. Two species of gall-inducing midges (Cecidomyiidae) insert egg(s) into the gynoecia of developing flower buds, conve itis one or more ovaries into ll. Роа wasps (Chalcidae) lay eggs i n the galls that develop into larvae that prey upon the midge maggots. The Cecidomyiidae expanded with the ape rms, but the earliest fossils of gall- inducing gall Е occur in id е a uy 'eptive mechanisms involving numerous floral traits in small bisexual and unisexual flowers are comnx the ANITA group and other basal angiospern ey words: Amborella tric pes eris үе ANITA group, beetles, dieer y, t midges, insect/wind pollina- tion, New Caledonia. 5 The woody evergreen plant Amborella trichopoda nies constructed with a variety of genes place it as Baill. (Figs. la, 2a, b, 3b), the only species of Am- the sister group of the rest of the flowering plants borellaceae, is endemic to New Caledonia in the (Mathews & Donoghue, 1999; Qiu et al., 1999; Sol- South Pacific (Jérémie, 1982). This species has at- tis et al., 1999 ). It is a member of the paraphyletic tracted worldwide attention, as molecular phyloge- ANITA group, a grade comprising the first three ' This research was funded by a National Geographic Grant (#6974-01) to PB, LBT, and TLS; a financial award to LBT from Tulane University; a Natural Sciences and Engineering Research Council of Canada grant to TLS. The author: are in демей to P. H. T lor his r support; IRD for use of laboratory facilities; G. Lowry II for инт on acquiring chemicals; in di чо h s harig for assistanc > with chemical purchases; J. asker for translation of French; Ree 'eves pn gall à nsect rearing; J. Fras 'hnical assistance; ; Susanne Moghlia for help in loc do plants and information conc erning Ms ipe. bee thank the following sc ientists or identific ation of insects: Michael W. Gates, Eric Grissell, Steve Lingafelter, Stuart McKamey (USDA-ARS, Syst. Ent. Lab.), Warren Steiner (Smithsonian- Eatomolosy), Ray a (USDA-ARS, em "Ent. Lab, retired). Michael W. Gates contributed to editing the manuscript and provided key ~ - ie] 4 e 5 > ч 3 © > = rzi De D ~ c = чш a X = © "T © 2 T 3 = erences and information on insects. ? Cell and Molecular Biology Department, Tulane үш (Uptown), New Orleans, Louisiana 70118, U.S.A ape ا‎ .edu. +6813 Department of Botany, University of Toronto, Toronto, Ontario, M5S3B2, Canada. "alu de Botanique et d' Ecologie Végétale биек IRD, В.Р. A5, 98848, Noumea, New Caledonia. ° Department of Biology, Saint Louis University, St. . Missouri 63103-2010, U.S.A. ° Royal Botanic Gardens, Mrs. Macquaries Road, е 30 )0, Australia. E Departme nt of Botany, Graduate School of Science, Kyoto Un Ub. Kyoto 606-8026, Japar 1011,12 р of Biological Sciences, CW 405, Biological Sciences Centre, Alberta, T6G2E9, Canada. A о. du Laboratoire d'analyses, IRD, B.P. A5, 98800, Noumea, New Caledonia. ANN. Missouni Bor. GARD. 90: 466—490. 2003. University of Alberta, Edmonton, Volume 90, Number 3 2003 Thien et al. 467 Amborella branches of angiosperm phylogeny [Amborellaceae, Nymphaeaceae and Austrobaileyales (Austrobail- eyaceae, llliciaceae, Schisandraceae, and Trimeni- aceae); APG II, 2003]. There is some controversy over the precise branching order at the base of the flowering plant tree (Parkinson et al., 1999; Bark- man et al., 2000; Graham & Olmstead, 2000; Gra- ham et al., 2000; Qiu et al., 2000, 2001; Zanis et al., 2002), but Amborella is clearly the sole living representative of a lineage that emerged at (or at least very close to) the base of the crown (extant) flowering plants. Recent discoveries of many small fossils of bi- sexual and unisexual flowers in the mesofossil flo- ras of the Early Cretaceous (Barremian—Aptian) can be attributed to relictual extant lineages (lllici- aceae, etc.), further supporting the ANITA grade as 2000). The discovery of complete flowering branches of the ex- a set of ancient lineages (Friis et al., tinct, basal family (Archaefructaceae, with two spe- cies) from the Lower Cretaceous (124.6 mya; Yixian Formation) in western China (Sun et al., 2002) also supports the basal phylogenetic position of ANITA taxa. In a cladogram constructed with molecular characters and morphological characters relevant only to fossils, Archaefructaceae are the sister group to all flowering plants and Amborellaceae are basal to all extant angiosperms (Sun et al., 2002). Archaefructus was monoecious with male and fe- male “flowers” borne on different parts of the same em. Amborella trichopoda is dioecious with flowers of both sexes (Figs. 1, 3b) arranged in botryoids or poorly ramified panicles (Bailey & Swamy, 1948; Jérémie, 1982; Endress & Igersheim, 2000). Flow- ers of both sexes are small, dull white or cream in color (Bailey & Swamy, 1948; Endress & Iger- sheim, 2000). Male flowers, which are larger than the female, develop a terminal abortive organ that has been described as a potential reduced gynoe- cium, and female flowers develop abortive stamens (staminodes, Bailey & Swamy, 1948; Endress & Ig- ersheim, ). Male and female flowers reportedly develop secretory tissue on the anthers and stigmas, respectively (Endress & Igersheim, 2000). The fruit of A. trichopoda is an ovoid red drupelet containing a similarly ovoid seed; gall-like fruits have also been reported (Endress & Igersheim, 2 In spite of the abundant information on the re- productive development and anatomy of Amborella trichopoda, virtually nothing is known about its pol- lination biology and breeding system(s) under nat- ural conditions. It has been suggested that A. tri- chopoda may be insect-pollinated with secretions on anthers and stigmas serving as insect rewards (Endress & Igersheim, 2000), and moths, poten- tially attracted by a nocturnal floral scent, have been observed visiting greenhouse-grown plants in the evening (Collett, 1999). Moreover, Sampson (1992) noted that staminodes of female flowers may in some instances develop abnormal pollen, thereby providing a deceit reward for pollinators. Endress 2001), however, noted that A. trichopoda may be wind-pollinated. Fossil evidence indicates that in- м sects have been important as selective forces іп the evolution of seed plants (Baker & Hurd, 1968; Re- gal, 1977; Crepet & Friis, 1987; Grimaldi, 1999), and Labandeira (1998) noted that pollen consump- tion by an extinct lineage of insects occurred before the end of the Carboniferous based on the contents of insect coprolites. There is insect pollination in some extant gymnosperms (Norstog, 1987). How- ever, whether there are floral scents and anther and stigmatic secretions in A. trichopoda requires in- vestigation. In general, basal angiosperms show in- sect pollination systems emphasizing beetles (Co- leoptera) (Diptera), (Thysanoptera) and bees (Hymenoptera) are more likely to play secondary roles, and wind pollination 1990, and flies while thrips is noted as uncommon to rare (Endress, 2001; Thien et al., 2000 It has been documented that male flowers are larger than female flowers of Amborella trichopoda, but it is not apparent if dimorphisms exist in other reproductive or vegetative traits. Dimorphisms in vegetative and reproductive features are common in dioecious species and are often correlated with spe- cific modes of pollination (Feil & Renner, 1991; Feil, 1992; Renner & Feil, 1993). The purpose of the present study was to address some of the unexplored issues regarding the repro- ductive biology of Amborella trichopoda. The spe- cific questions addressed by studying the species in its naturally occurring environment were: (1) Is A. trichopoda insect- and/or wind-pollinated? (2) Do flowers produce floral volatiles and floral re- wards such as secretions on anthers and stigmas and, if so, what is their chemical composition? (3) What is the population structure (distribution pat- tern) of male and female plants? Three locations in New Caledonia, Méé, Plateau de Dogny, and Col d’Amieu, were utilized to focus on these issues in 2001 and 2002. MATERIALS AND METHODS ECOLOGICAL, EDAPHIC, AND FLORISTIC CHARACTERS OF AMBORELLA TRICHOPODA New Caledonia, located in the Southwest Pacific approximately 1200 km east of subtropical Queens- Annals MU. PUR Garden Figure l. Features of fruit, gall, and oviposited floral buds in Amborella trichopoda. —a. Red fruits are distinguished from larger green galls. Floral buds marked with arrowheads are slightly black indicating the presence of fungus. Bar n. = 5 mm. —b. Arrowhead denotes cecid larva within a recently apne female flower. Bar Volume 90, Number 3 2003 Thien et al. Amborella 469 land, Australia, is rich in gymnosperms and relic- tual representatives of the Cretaceous—early Tertia- ry Gondwana floras of Australasia (Lowry, 1998; Morat et al., 1984; Jaffré et al., 2001) including basal angiosperms. Amborella trichopoda occurs in rain forests with acid soils along a chain of high mountains from Plateau de Dogny to the high valley of the Tipindjé River (Jérémie, 1982; see soil map. Pintaud et al., 2001). In the rain forests, A. tricho- poda occurs within the 1500 mm (and above) rain- fall isohyet determined by Pintaud et al. (2001). The restricted distribution of Amborella tricho- poda on the island can be correlated with some wet conifer and palm sites (Jaffré, 1994; Pintaud et al., 2001). Nearly all palm populations occur within the 1500 mm isohyet, with maximum species diversity in four areas receiving > 3000 mm of rainfall per year (Pintaud et al., 2001). The four areas are des- ignated Pleistocene rain forest refugia by Pintaud et al. (2001). In the dry periods of the Pleistocene these refuge areas may have remained within the 1500 mm isohyet (Pintaud et al. 2001; Jaffré, 1994). Some populations of A. trichopoda currently occur in the Massif des Levrés Pleistocene refuge (number 2) of Pintaud et al. (2001) Herbarium specimens (IRD, Noumea, New Ca- ledonia) indicate two flowering periods, one from early March to the end of May and the other in August (Jérémie, 1982); however, in August 2001 no plants flowered at Mee and Plateau de Dogny. The primary flowering time of Amborella trichopoda is during the warm season (mid November to mid April), in which rainfall is high with frequent trop- ical depressions and cyclones (Lowry, 1 Jaffré (unpublished data) suggests ДЕ ам tri- chopoda occurs on acid soils rich in Al, in habitats with high precipitation, but is not found on ultra- mafic soils (peridotites and serpentines). The soils at Plateau de Dogny have a pH of 4.2 and contain 10.6% Al, and at Col d'Amieu soils contain 14.0396 Al with a pH of 4.14 (Jaffré, unpublished data). The accumulated Al content of some plants at Plateau de Dogny are as follows: Rapanea Aubl. sp.—30 ppm; Phelline Labill. sp.—40 ppm; A. tri- chopoda—75 ppm; Ixora montana Schltr.—7 ppm; Psychotria balansae (Baill.) Guillaumin—163 ppm; Hedycarya baudouini Baill.—375 ppm; Psy- chotria L. sp.—6875 ppm; Zygogynum crassifolium (Baill.) Vink—9250 ppm; Knightia stobilina (La- bill. R. Br.—10,625 ppm; and Melastoma denti- culata Labill.—11,375 ppm. Mabberley (1993) re- corded A. trichopoda as accumulating Al; however, the tissue values provided by Jaffré suggest Al ac- cumulation is negligible. Site description. Natural populations of Ambor- ella trichopoda used in the present study are lo- cated at three sites: (1) Col d'Amieu 500-800 m near Sarraméa, Province Sud; (2) Plateau de Dogny 500—800 m. near Sarraméa, Province Sud; and (3) Mée, located on the eastern side of the island (570 m). The Col d'Amieu site was the most accessible and hence was the primary study area visited in August 2001 and March 2002. Plateau de Dogny was the secondary study area visited in March— April 2001 and August 2001. The Мёё site was visited in August of 2001 and was used to deter- mine the production of galls and observe insect/ plant interactions. During all site visits conducted in August, flowering was examined to evaluate whether it corresponded to the information from herbarium sheets (IRD, Noumea, New Caledonia). The forest floor of the Col d'Amieu site (Fig. 2). a mature rain forest on the eastern slope of the mountain range, is dominated by ferns, especially Marattia attenuata Labill. and Orthiopteris firma (Kuhn) Brownlie (1 m in height). The arborescent fern Cyathea vieillardii Mett. and the palm Burre- tiokentia vieillardii (Brongn. & Gris) Pichi-Serm. are common in the forest. An unidentified species of Freycinetia Gaudich. is also abundant. Other un- derstory plants include Amborella trichopoda, Psy- chotria baillonii Schltr., Hedycarya cupulata Baill., Phelline brachyphylla Baill., sifolium Guillaumin, Tapeinosperma acutangulum Zieridium pseudobtu- Mez, and Casearia silvana Schltr. Trees in the un- derstory include Cryptocarya macrocarpa Guillau- min, Calophyllum caledonicum Vieill., Garcinia L. sp., Hernandia moerenhoutiana Guillem., Syzygium Gaertn. sp., Harpullia austrocaledonica Baill., Schefflera Foster & Foster f. sp., Niemeyera bal- ansae (Baill.) Aubrév., and Ficus L. spp. The forest differs from the nearby “forét Persan" on the Col d'Amieu massif (Jaffré & Veillon, 1995) in its steeper slopes (with a slope aspect of 50-65% vs. gr poe at slightly higher elevation (540 m vs. 500 m) with greater precipitation (> 2000 mm vs. ca. 1800 mm). DIMORPHISM IN SELECT VEGETATIVE AND FLORAL FEATURES In dioecious breeding systems it is necessary to determine the various factors that affect reproduc- tive success, i.e., number of male versus female plants, number of flowers produced by each sex, pollen to ovule ratios, etc. (for techniques see: Kearns & Inouyne, 1993; Real, 1983; Dafni, 1992). An initial assessment of dimorphism in secondary sex characters in Amborella trichopoda was con- 470 Annals of the Missouri Botanical Garden Figure 2. trunk has been broken. Arrow indicates leaf he rbivory. ducted by establishing six 10 X 30 m plots laid eastern. slope of Col d'Amieu (Fig. 4). Within each plot, the spatial dis- out at random along the A tribution of male and female plants of A. trichopoda Plant ground level, and stem diameter were determined number = was mapped. height, of stems a for all individuals within each plot (voucher spec- at RYD and MO; Table The gender of each plant was determined by inspecting imens deposited all accessible open or opening flowers in the lower canopy and inspecting plants for the presence of fruit. As well, branches (N = 4 to 6 per plant; N = 75) with flowers and floral buds were randomly selected and removed from the higher canopy (ca. 2.5 m), 3 females and 3 males from each of 3 plots. They were immediately placed in water and taken to the these unattached branches continued to develop aboratory for observations. It was noted that flowers for up to two weeks. All male flowers and floral buds from these branches were inspected in the laboratory with a dissecting microscope to de- termine if a functional carpel was present. Anthers and staminodes were squashed with Alexander's Triple stain (Alexander, 1969) to assess the number of aborted and non-aborted pollen grains per an- ther. The sampled branches were also used to de- termine the number of flowers/inflorescence, and the dry weight of male and female flowers at an- thesis. A transect from 625 to 730 m at Plateau de Dog- ny was laid out to determine the gender of individ- Growth form and habitat of Amborella trichopoda at Col d'Amieu. Note һу ede `ъ up i Arrowhead shows point where detritus on forest floor. —b. Forest habitat ual plants. Five branches each from 5 males and 3 females were removed and floral buds were exam- ined for the development of functional carpels and pollen, and to quantify flowers/inflorescence and in- florescence/cm. Plots were not established at Pla- teau de Dogny due to the steep slopes (55—70°), unpredictability of weather, and difficulty of access. To determine floral longevity, 3 male and 3 fe- male plants from the Col d'Amieu population were marked with colored plastic wire, and the phenol- ogy of flowers was examined for each of 8 consec- utive days. Flowers were observed with a 10X hand lens and the following traits recorded: tepal move- ments, presence or absence of stamen and stigmatic secretions, period(s) of secretion, stamen dehis- cence, duration of pollen retention in anthers, and onset of necrosis. Floral longevity was also deter- mined on flowers opening on materials brought into the d'Amieu. aboratory from Plateau de Dogny and Col All statistical analysis of morphological traits was conducted using one-way ANOVA or the Mann-Whitney Rank Sum test (SigmaStat 2.03). Mann-Whitney Rank Sum tests comparing data col- lected from Col d'Amieu and Plateau de Dogny showed no significant differences; therefore all data from these two sites were pooled. WIND DISPERSAL OF POLLEN Experiments to assess the possibility of wind pol- lination were conducted at Col d'Amieu in March 471 Thien et al. Volume 90, Number 3 Amborella 2003 etroleum jelly- = 2002. Pollen traps consisting of р | Б Sess. = laced within the canopy a : © == сеу covered slides were plac f dale E Pga bb f roximately 5 to m from - E ЕЯ [нинин distances of арр бы аии. E 55 немо 1а and female plants at 1 to 5 m above is ed | © hopes isnsYo ; ays, placed in 8. SESS = = = gg The traps were removed after four days, DE 1 © 7 E эа 2 ie a microscope slide box that was then seale > $*5525 i n Jd t and transported to the University of Toronto. E: 5 = — ape, e : : : ial inter- © ЕТЕ о i M | : ] with differential i > : N 5 © slides > then observec á os ~ N Slides wore antify the number of Amborella E ference optics to quantity (Fi 5) = у= : irc 1 1g. 5). ———— : ains present g t = = 2 us trichopoda grains ү 3 9 alas о £165 tE adt $ PET 4 с E = INSECT POLLINATION S.S н Pes Гета ! z. = |eez-eec ‚2. : ans of Amborella tri- 5 E E 3 SASA ej e Insect activity on floral organs of i uM e E HESINAN = -honod s documented for 2 to 3 hours E ^ eb & | єз Са " Сї сч сһоро а ма : | h 2002. Male and fe- 5 م ل ب ف ا٤ پوك‎ 4 d ays at Col d'Amieu in Marc . 3 S ° 8| نم نم تم‎ сс days at Cc served at the site on a clear © = i. ale plants were also observe m E male [ t and continuing for 3 А — nse ® 7 night beginning at I BE Lucian Lith = © ais tw eo hours. It was noted that Ti har insects (not dap- ES rr 9 3 ther inse T E z^|3e "n cm the cecid gall ecosystem Tt h lls and in the E & tl SS + + 4 d flowers) lay eggs in the ga es = t tured on По " x “okm tte uii le and female buds of A. trichopoda. This ee ы Wa | teins) Cl нч ин та!е а шге‹ е |с 6 Б insects (captur ^ E SE qaa 1 ۹ 9 however, emphasizes pollinating Ё A nn A $ 2 3 с 2 MO | on flowers) and insect > o 1 age EF 5 E: le of insects visiting plan = sample ; k Ё £ = | 4 examined for pollen loads as described by B = Е ®_@ |жж с 9 |= 90). Insect specimens were sent to = ® Ор on desi hardt (1990). -— Бс and ide- 3 занн a ae К, Smithsonian Institution for iden = a S ae x z 53 = чост s 8 position. 8 „чш = ЕЕ |с s o 0 кр N => NN Ss MISC 5 = 2 Bee LLINATION s E «шур |, ij BAT. NATURAL RATES OF OPEN PO | Е i libio 1 la tricho- 5 n © Natural rates of pollination in tear ie f uu S aa € 5 1 e пит E ЕБ | Є кей лү б: da were determined by recording t ; < Lz^laee-c-o|s po found in carpels in each flow t gu SSu a tle length of pollen tubes foun nus Иң 3 bb” | tl H - ; le flowers with fresh, E T ьо ННооо | = А ximately 200 female x as iwi dh iA rd n nw en CPP icked from 11 plants along = 23/5555 $ = | а receptive stigmas were p 1 04у я Е" i 2 from 625 to 730 m at Plateau " Ес " "з a transect fro Б 95% ethanolgldrial E "t s and fixed overnight in a } tored in : Z acetic acid mixture. The material was s ; ми ты, -~ рам. ра. A - © 1-17 8 E ENSEM M- l. The carpels were excised and soft e SEIS a Шеш а € 7096 ethano $e E bibite andira- 5 Е = tl WWW d 10% solution of s Ф & ++ “а ened in à dred В = U EEE S 5s E ter at room temperature for 24 hours, then prep E р" == хос о Ж microscopy fol- 8 2 o D|uu-x = { softening and e pifluo rescence 3 بے بے |2 ك‎ 1 : T hardt (1990), except that E Е = | аата = lowing Goldblatt i Bernhar zd. € J = noeciu E 8 E all carpels within the same aid material re- Е а 8 d under a single cover slip (no &. Б Е сое 5 squashe но к lv siens of galling Е л ai & rau. maining). Flowers showing early sig à | نک ت ق تب‎ NU Ẹ S FINSEN AN AN t were discarded. o Ф = P- m'a = = č = L б< Маг = E HISTOCHEMICAL AND ‚р кай I Locis oon BE TOT EM ~ THER: ^ oz NESS CHARACTERIZATION OF AN = =- = — 4 ? еѕепі оп ‚5 E © To determine whether secretions were pr I. pus € sti s, or other floral tissue, flowers г Ф anthers, stigmas, . 7 Га T = O C ut > ant ers, А il 2001) and Col d Amieu ER gE|-- * Plateau de Dogny (Apri v Annals of the Missouri Botanical Garden Figure 3 —a. Scanning electron micrograph illustrating lipid pollen coat on pollen grain. Tissue prepared with c ‘ryofixation. Bar = 1.3 um. —b. The beetle Neoadelium Јаше (Azszab) on female inflorescence. Large вор arrowhead denotes positive uen nium red staining of stigma indicating pectin, Small black arrowhead marks in situ positive aniline blue black staining of stigma indicating protein. Features of the pollination biology of Amborella trichopoda. Volume 90, Number 3 Thien et al. 473 2003 Amborella (March 2002) were examined in the natural popu- glass cartridges were subsequently sealed in alu- lations or on cut branches in the laboratory. То as- minum foil and stored in a freezer (—20°C). sess the presence of various chemical constituents drops of the following stains Mies applied M flowers CHARACTERIZATION OF FUNGI IN FLORAL GALLS from both populations: 196 alcian blue in 396 acetic acid for acidic polysaccharides, pectins, and mu- cilages (Jensen, 1962; Benes, 1968); 0.05% ruthe- nium red for pectic substances (Gurr, 1965); 1% aniline blue-black in 7% acetic acid for proteins (Fisher, 1968); and Sudan black B in 70% ethanol for lipids (O’Brien & McCully, 1981). Nonspecific esterase activity was detected on flowers immedi- Male and female plants were inspected in the field for the presence or absence of fruit-like galls. Inspection indicated that female plants only pro- duced fruit-like galls. However, holes indicating the emergence of insects were regularly observed in 2 mm male floral buds. Similar observations were made on female floral buds in addition to the fruit- like galls. The fruit-like floral galls of Amborella trichopoda were collected from plants growing in the Col d’Amieu (2001 and 2002) and Plateau de Dogny (2001) and placed in closed sterile vials ately removed from branches, described by Dejong et al. (1967). Controls were run by omitting the substrate. Male and female flowers were also sampled at anthesis and from 1 to 3 days post-anthesis and were prepared for cryofixation and freeze-substitu- where insects were allowed to emerge. Branches tion at IRD (Institut de recherche pour le dével- oppement). Freeze-substituted flowers were then prepared for light, scanning, and transmission elec- galls/node as well as fruits/node. Insects from the tron microscopy after warming to room temperature floral buds and fruit-like galls were stored in 70% as described by Koehl (2002). To determine the ETOH and sent to the Smithsonian Institution for presence of pectins and arabinogalactan proteins identification and deposition. (AGPs) in the extracellular matrix (ECM), immu- Dissection of both parasitized floral buds and flo- nolocalization at the level of the TEM was con- ral galls indicated the presence of fungal growth ducted using monoclonal antibodies JIM5 and therein. To identify the fungi, intact flowers and JIM13 specific to highly esterified homogalacturon- galls were placed in sealed sterile vials and trans- ans and AGPs, respectively, as described by Koehl ported to the University of Toronto. Within 5 days, (2002). the galls were split into halves using a flame-ster- used to determine flower number/node (see above) were also used to quantify the number of fruit-like ilized razor blade and then pulled apart. This al- FLORAL AND LEAF ODORS lowed the larval chamber to be exposed without ‘ introducing contamination material. All galls ex- Volatile compounds were sampled by head space Б B methods and analyzed using GC-MS as described by Azuma et al. (1997, 2001). A mini-pump (model MP-2N, Shibata Scientific Instrument, Tokyo) was used to extract air from the enclosed flowers, buds, and leaves of male and female plants. Glass car- tridges (7 mm X 5 ст) containing adsorbent (30 mg Tenax GR, Mesh 60/80, GL Sciences, Tokyo) tissue. The culture medium contained Modified were inserted into the flow line (silicon tubing) to Leonian's Agar (Malloch, 1981) supplemented with capture the chemicals. A total of 12 samples (in- chlortetracycline and incubated at 21°C for several amined contained insect larvae or pupae and/or ev- idence of insect activity (with insect exit holes). Using a dissecting microscope, small pieces (4—5 per gall) of the chamber wall or frass of 10 galls were aseptically transferred to petri dishes to isolate (grow) fungi that were living in the plant cluding controls) were taken for various lengths of days. Modified Leonian’s Agar was chosen because time during day and night periods to span a com- it is an all-purpose medium that generally allows plete day. The pumps were driven at approximately good growth for most fungi. The resulting colonies 150 ml/min for 2 to 6 hours (plus control). The were then subcultured on Modified Leonian Agar. ae —c. Tissue on ovipositor (arrow) of cecid captured on petroleum jelly-covered slide placed on female plant. Bar = 25 um. —d. Pollen tetrad possibly of A. trichopoda (arrow) : adjacent to mouth part of the same cecid qid in Figure r = 13 um. —e. Pollen tubes (arrows) growing on stigma of A. она (open pollinated). Ваг = 25 рт. — f. Pollen tube (arrows) entering micropyle and embryo sac of ovule of А. trichopoda (open dn Dos = 25 pm. Figure abbreviations: L, lipid pollen coat; S, stigma; OV, ovule. 474 Annals of the Missouri Botanical Garden ELEVATION - 565 m с 9 d c B g Е, P. р F #597 d e 226 9 do ? d g? g 9 ELEVATION AM -540m [ef x ELEVATION - 500 m пе 4. Map of 10 X 30 m plots at Col d'Amieu showing distribution of male and female plants of Amborella igi trichopoda. RESULTS DIMORPHISM IN SELECT VEGETATIVE REPRODUCTIVE FEATURES AND No significant differences were observed at Col d'Amieu between male and female individuals with respect to height, number of stems at ground level, and stem diameter (Table 1). In general, Amborella trichopoda is an evergreen shrub (Fig. 2) with a slightly scandent habit (Bailey & Swamy, 1948), or it forms a small tree 7-8 m tall (Posluszny & Tom- linson, 2003). The main shoot (trunk) of A. tricho- poda tends to branch at or near the soil surface (Fig. 2). Adventitious roots occur on the trunks of some plants. Trees of A. trichopoda with tall and large diameter trunks are rare, although four indi- viduals were recorded with trunk heights ranging from 1 to 5 meters (3.04, 5.0, 1.3, and 1.0 m); th trunks. of these. individuals tended to be broken and/or hollow and rotten (Figs. 1, 2). Signs of her- bivory and necrosis are common on leaves as is the growth of lichens. All plants examined at Plateau de Dogny and Col d'Amieu were strictly male or female. No female flowers, fruits, or fruit-like galls were identified on any male plants and no functional carpels were identified on male flowers, thereby confirming pre- vious reports that the plants are functionally uni- sexual and the populations are dioecious (Bailey & Swamy, 1948; Endress, 200 ale flowers pro- duced 9552 + 996 pollen grains/flower lowers/4 iod with a pollen abortion rate of 7.9% + 4.996 (М = 12 flowers/4 plants), whereas stam- inodes on hus produced no pollen grains, abort- ed or otherwise. The plots from the Col d'Amieu population of Amborella trichopoda were all male biased (Table 2 1). The male and female plants at lateau de Dogny were not mapped, although the male to female ratio was 1. . N = 118 plants in a transect from 625 to 730 m. Although female Volume 90, Number 3 Thien et al. 475 2003 Amborella $= T T T T T T T T T T 150 F - E | "d I2 120r 1 со = р ° e e e | ©) x * Zz “г, S ult : | 60 + ° - : die e ° T e e e ө LLI Ш зор ° $ ы BS E © Г 6 e e ° e ° $ е 7 а. о e e ө ө ® ө ө $ s еи t+ L П L П П 1 П П П П 12 3 4 5 6 7 8 9 10 11 12 TRAP NUMBER Figure 5. Number of pollen grains captured on petroleum jelly-covered slides placed within the canopy of female as (traps 1—3) and male (traps 4—6) p la nts of Amborella trichopoda, respectively, and within З т raps 7-9) and 10 m uae 10—12) of the co-occurring ا‎ and females, from three separate locations for four days and male plants did occur singly, more commonly they were clustered, but preliminary AFLP ses of their DNA indicate the plants do not form clones (Graham et al., in prep.). Flowering and non- analy- flowering branches from adjacent male and female plants were often intertwined. THE MORPHOLOGY AND PHENOLOGY OF REPRODUCTIVE ORGANS The male and female flowers of Amborella tri- chopoda are small, flat structures with organs ar- ranged on a floral platform that rips as the flower expands (Endress & Igersheim, 2000; Posluszny & Tomlinson, 2003). The tepals of flowers of both sex- es are cream-colored (the ovary is green in female flowers), and flower size ranges from 3 to 5 mm across (male flowers are slightly larger; Bailey & Swamy, 1948; Endress & Igersheim, 2000). The flowers are produced on short (4 cm) paniculate inflorescences in the axils of leaves (Money et al., 1950; Endress & Igersheim, 2000) and are also cauliflorous on stout branches and trunks. The male flowers bear 6 to 21 short stamens with strap-like filaments and triangular anthers (Baillon, 1869; Bailey & Swamy, 1948; Endress & Igersheim, 2000). Each female flower usually contains O to 4 staminodes and 3 to 9 spirally arranged carpels. The number of floral organs in both female and male flowers is highly variable, a pattern common to the ANITA grade and other families of basal an- giosperms (e.g., Magnoliidae sensu Endress, 1990). Flowering of male and female plants was syn- chronous at Plateau de Dogny and Col d'Amieu. Significant differences were observed between the number of female and male flowers per inflores- cence (number of female flowers/inflorescence — 4.2 + 1.3, N, 106 inflorescences; number of male flowers/inflorescence = 4.6 + 1.6, М, 98 inflores- 0.05). Significant differences were ob- served in dry weight of female versus male flowers 011 g + .006 g/, N, 10 flow- 0.0226 g + .006 g/, М, 10 flowers/9 plants; P — .05). 127 buds and flowers (at 14 nodes) over 8 days on 3 male plants (7 branches) cences; P — at anthesis (female — ers/9 plants; male — Observations of at Col d'Amieu indicate the buds open throughout the day and night. This begins as the tepals form a small, circular aperture during expansion at the perianth apex. Approximately 18 to 45 hours are required for the process to reach completion with the dehiscence of the outermost stamens. Pollen is dispersed for 36 to 96 hours. Inspection of fresh, unfixed anthers indicated that no exudates were present on the anthers. The average life span of a "typical" male flower from bud opening to complete anther dehiscence of pollen ranges from 4 to 5 days. As the perianth expands the rim of the floral platform gradually curls downward, spreading the stamens. Once the anther sacs are empty of pollen the center of the flower becomes concave, the floral platform curls further and the greenish white sta- mens turn brown, and the entire flower abscises, usually within one day Observations of 29 nds and 15 flowers at 4 nodes 476 Annals of the Missouri Botanical Garden over 9 days on one female plant (2 branches) at Col d'Amieu, showed the flower bud opening process was similar to that of male plants. The time from bud opening to open flowers with receptive stigmas (visual inspection), varied from 18.5 to 24 hours, with stigmas appearing receptive for up to 24—30 hours. When the perianth first opens the staminodes and stigmas are held upright but gradually reflex as the flower progressively ages. The staminodes lie perpendicular to the carpels when the flower is fully open. As the flower ages the green-white stigmas and tepals gradually turn. brown or faint red to yellow and abscise after 1 to 2.8 days. HISTOCHEMICAL AND STRUCTURAL ANALYSIS OF THE ANTHERS AND STIGMAS Inspection of fresh, unfixed anthers combined with histochemical and structural analysis indicat- ed that no exudates were present on the anthers. As well, no exudates were observed to be present on the anther when cryofixation was used prior to preparation for SEM. The irregularly shaped cuticle was thick and blistered with an underlying pectin- rich layer (Fig. 6a, b). The cuticle was uniform throughout. except for a reduced number of blis- tered regions at the connective tip. Stigmas varied as to whether an apparent free- flowing secretion was present or not. An apparent free-flowing stigmatic secretion was observed only on 3 of more than 1200 female flowers under nat- ural and laboratory conditions from plants at Pla- teau de Dogny and only 30 of approximately 1000 flowers at Col d’Amieu. The apparent free-flowing secretions were only present at Col d’Amieu during the second week of observations after a day of rain. The following observations apply whether the free- flowing stigmatic secretion was apparent or not. Stigmas were histochemically positive for pectins and proteins (Fig. 3b). A thin cuticle with positive esterase activity covered the multiseriate stigmatic papillae, which was subtended by a heterogeneous extra cellular matrix (ECM) containing low-esteri- fied pectins, and moderate amounts of AGP (Fig. 6d, e). This ECM extended between all files of cells comprising the stigma and into the region of an- giospermy (Fig. 6f). FRUITS AND GALLS The small (5 X 7 mm) drupaceous fruit of Am- borella trichopoda (Bailey & Swamy, 1948; Endress & Igersheim, 2000) has slightly compressed sides, a fleshy mesocarp with red juice (Fig. la), and a stony endocarp (Bailey & Swamy, 1948; Endress & Igersheim, 2000). Both fruits and galls (Fig. la) were observed at all three locations during August and March. Observations of mature fruits and galls at Col d'Amieu and Plateau de Dogny suggest both remain attached to the plant through overlapping flowering seasons. The mean number of fruit/node was not significantly different from that of galls/ node (fruit/node = 1.7 + N, 259 nodes; galls/ 1.7 + 1.3, М, 183 nodes; P = 0.231) There was an expected significant difference in dry node = weight between fruits and galls (fruit dry weight/5 fruit = 0.21 g + 0.02; gall dry weight = 0.17 g + 0.03; t = 3.508, df = 18, P = 0.003). Only 5 out of 550 galls dissected appeared to contain a seed. The number of fruits and galls developing per flow- er ranged from none to 5. Both fruits and galls de- veloped in the same flower(s). DEVELOPMENT AND ASSOCIATED ENTOMOLOGY OF CARPEL GALLS Insects oviposited either within the whorls of ini- tiating organs of both male and female flowers or within the ovary locule of Amborella trichopoda ~. Fig. 1b). Insects developing following ovipositing within the whorls of organs were easily identified as the floral buds were swollen and frequently dark in appearance due to the presence of a dark fungus Fig. la). Ovipositing within the ovary locule only resulted in the production of galls. Insects isolated from the floral buds where ovipositing did not in- duce gall development included Rileya, Megastig- mus, and Thysanoptera. Pollen placed on the stigma of an oviposited ovary failed to germinate. he breeding system of Amborella trichopoda is impacted by two species of gall-inducing gall midg- es (Cecidomyiidae; Roskam, 1985; Hawkins & Gagné, 1989; Gagné, 1989). The most common gall midge (maggot) in the galls of A. trichopoda is a new species in the genus Asphondylia, probably en- demic to New Caledonia. It is known that many species of Asphondylia have a needle-like chitin- ized ovipositor to place eggs into plant tissue (like buds). Pupal development of Asphondylia takes place in the galls of A. trichopoda. In other seed plants, emergence of adults is usually correlated with specific plant tissue formation of the host. Typ- ically males emerge first and after mating females search for the host plant (Harris, 1994). The in- sect's adult life tends to be brief (usually hours or days, but some live longer) and eggs hatch within days or weeks, followed by development of the lar- vae (Harris, 1994). The development time of the larvae feeding in the galls of A. trichopoda is un- known. As noted, the primary flowering season of A. trichopoda extends from March to May, allowing Volume 90, Number 3 2003 Thien et al. Amborella adequate tissue for repeated infection of the female flower buds. The insects insert an egg(s) into the carpels of developing flowers (in 2 mm buds) on female plants (Fig. 1а). Subsequently, with emer- gence and feeding of the larva (Fig. 1b), the carpel develops into a hollow gall (ca. 8 X 9 mm), nour- ished by vascular traces from the plant. In the ma- ture gall, fungi line the cavity of the gall and are perhaps eaten by the developing maggot (Roskam, 1992; for galls in general see Dreger-Jauffret & Shorthouse, 1992, and Mani, 1964). In a sample of 10 galls, each larval chamber contained fungi, but the surrounding gall tissues did not. The fungal material clearly differed from gall to gall in color (white, black, or brown) and texture. Seven of the 10 galls yielded only 1 fungal species, while 3 contained 2 species of fungi. Spor- ulating black or brown colonies have been identi- fied as Phoma and Cornucopiella. All fungi were mitosporic (Fungi Imperfecti). Four of the 17 iso- lates (from the 10 galls) produced white colonies: 2 have tentatively been placed in the genera Ac- remonium and Gliocladium, but 2 of the isolates remain sterile and unidentified. Isolates of two floral buds and one pupa (cadav- er) from inside one of the galls also yielded fungi. The buds yielded isolates of Fusarium spp.. and the cadaver was infected with vigorously sporulat- ing species that produced apothecium-like coni- diomata of Cornucopiella. Floral buds with Fusar- ium did not develop mature anthers or carpels. Non-sporulating fungi were common in the iso- lates, as has been found in other insect-associated fungi, e.g., fungi cultivated with attine ants (Mu- chovej et al., 1991), and in ambrosia fungi (Cassar & Blackwell, 1996), in which the fungus exists in an asexual yeast-like form. Several fungi have previously been reported from galls produced by Asphondylia species (insects). Many of the fungi appear to be common plant en- dophytes, but their relationship to the insects is not clear (Wilson, 1995). Macrophomopsis coronillae (Desm.) Petrak has been repeatedly isolated from galls produced by Asphondylia species in north temperate regions since the early 20th century (Kehr, 1999) and has been thought to be used by the larvae as food, since Neger (1909, 1910) first identified it and labeled it as *Ambrosiapilz." None of the isolates from galls of A. trichopoda could be identified as species of Macrophomopsis. The non- sporulating strains, however, require further study before they can be identified. GALL INSECT ASSOCIATES Each gall is essentially a small ecosystem with its own species composition of parasitoids (Zwolfer, 1979; Hawkins & Goeden, 1984; Hawkins & Gag- né, 1989). The galls produced by Asphondylia sp. nov. on Amborella trichopoda also have a guild of associated insect taxa, mostly parasitic wasps (Hy- menoptera). Although the exact relationships have not been elucidated in this particular system, it seems most likely that these wasps are primary or secondary parasitoids in the gall system, consuming the larvae of gall-forming flies. The most common chalcidoid is a new species of Rileya (Chalcidoidea: Eurytomidae), which resembles three other species, R. insularis and R. cecidomyiae from the New World. and an undescribed Australian species. Rt- leya species to date are associated exclusively with galls made by the Cecidomyiidae, and several Ri- leya species have been documented from non-eco- nomically important cecidomyiids in the United States (Plakidas, 1982; Plakidas & Weis, 1994; Waring & Price, 1989; Hawkins & Goeden, 1984 However, Rileya cecidomyiae Ashmead (cited as К. tegularis Gahan by Hawkins & Goeden, 1984) can switch from entomophagy to phytophagy during its — larval existence. This mixed feeding habit is ap- parently an important component in parasitoid suc- cess (Hawkins & Goeden, 1984). One unidentified braconid wasp (Ichneumonoidea: Braconidae) also emerged from galled A. trichopoda fruit. It is likely parasitic and may attack the gall formers or its par- asitoids. The genus Systasis (Chalcidoidea: Pteromalidae) was collected on and near flowering plants of Am- borella trichopoda. This genus is cosmopolitan, with over 30 species worldwide and 21 from Australasia (Boucek, 1988). The various species are seed feed- ers or parasites of Cecidomyiidae, and some host associations may be mixed phytophagous and en- tomophagous (Boucek, 1988). Two larvae of Megastigmus were found in galls of Amborella trichopoda. The genus Megastigmus Dalman (Chalcidoidea: Torymidae) comprises about 125 described species, most of which are believed to be seed-eaters. A few species, however, parasit- ize gall-formers, their parasitoids, and in some cas- es they also consume plant tissue. The most com- mon host insects are Cynipidae, but gall-forming Eurytomidae and Pteromalidae also serve as rare hosts, as well as two species of parasitic Torymidae. One species is known to parasitize Cecidomyiidae in flowers of Eucalyptus spp., and two attack gall- forming Eriococcidae (Grissell, 1999). Many hosts remain to be discovered. FOLIAGE AND FLORAL FORAGERS AS PROSPECTIVE POLLINATORS A wide variety of insects visit floral and foliage organs of Amborella trichopoda. Most notably, the 478 Annals of the Missouri Botanical Garden Figure 6. Micrographs of anther and stigma of Amborella ligi sque All tissues prepared by cryofixation, —; t Epidermal cells of terminal region of the anthers. Note absence of secretions. Arrow shows blistered appearance o cuticle. Bar = 3 рт. —b. Transmission elec tron | micrograph of anther epidermal cell showing me alization of pectin (arrowhead) be low thick cuticle. Bar = 0.2 pm. —c. Scanning electron mic rograph of stigmatic cells showing ruptured thin cuticle (arrowhead) and localized uplifts of pectin (arrow). Bar = 6 um. —d. Transmission electron Volume 90, Number 3 2003 Thien et al. 479 Amborella majority of these insects also occur in the litter of the forest. Insects collected and observed on both flowers and vegetative parts of Amborella trichopoda at Col d'Amieu are listed by identification number in Table 2. Two species of Curculionidae (Cryptor- hynchinae) are common pollinators of A. trichopoda at Col d'Amieu (Table black cryptorhynchine ca. 4—5 mm in length was collected and observed at all study sites. In the — — 2). One species, a sma flowering season this insect was observed on male and female flowers and leaves of A. trichopoda. However, in August 2001 (A. trichopoda was not in flower) it was again recorded on the leaves of A. trichopoda but not observed on other plant species in the populations. The larger cryptorhynchine (2-3 cm in length) spends hours walking on the branches, leaves, and flowers of A. trichopoda, and pollen of A. trichopoda was removed and identified from the ventral surface of one individual (Table 2). When the rostrum (beak or snout) of the insect is folded, this weevil becomes difficult to detect due to its cryptic coloration, as its body then resembles a seed or a fungus. Neoadelium fauveli (Zaszab) (Coleoptera: Tene- brionidae: Adelinae) belongs to what is considered to be the most primitive tribe of the family (Fig. 3b; Doyen et al., 1989; Hawkeswood, 1987). The genus Neoadelium is endemic to New Caledonia and contains five known species (Kaszab, 1982; Matthews, 1998). The insects were observed scoop- ing pollen out of the anthers, and their mouth parts were covered with pollen. Microscopic examination of the hind gut contents of two male specimens and the foregut of another (collections 3, 17) revealed abundant wood bits and other plant tissue, but no pollen. Pollen feeding has been observed in some unrelated tenebrionids (Steiner, 1988) and by P. Bernhardt and P. Weston (in prep.) on Eidothea hardeniana P. H. Weston & Kooyman, a Proteaceae (Weston & Kooyman, 2002). Members of Adelinae are not known to feed on pollen but “оп leaf litter and other accumulations of dead plant matter— some in rotten wood" (Matthews, 1998). Otherwise, members of tenebrionid groups are regarded as un- common flower and pollen foragers, and have been unrecorded previously as prospective pollinators (Bernhardt, 2000). However, Neoadelium fauveli is flightless, as its hind wings are very reduced and presumably would not be as efficient a pollinator as winged beetles. Long-horn beetles (Coleoptera: Cerambycidae: Lamiinae) were observed at night on female plants of Amborella trichopoda walking on leaves and flow- ers. This species is flightless and rare. The larvae of Cerambycidae are generally xylophagous (Hawkeswood, 1987), while adults often feed on flowers or fruits (Bernhardt, 2000). This particular species may be associated with fungus or lichens as an adult (Lingafelter, pers. comm.). Lamiinae are well developed in Australia, and several endemic species occur in New Caledonia (Gressitt, 1982). Other taxa collected on Amborella trichopoda are listed in Table 2 (Nos. 1, 2, 5, 7, 8, 11, and 14), and general comments are offered below on what is known about their habits. The aderid and the cor- ylophid (Nos. 1 and 2) are usually found on vege- tation and are most likely detritivores. The mirid insect (No. 5) is probably phytophagous, although predatory species do occur. The aphrophorid insect (No. 8) and the cicadellid (No. 7) are phytophagous. the latter being endemic to New Caledonia. The unidentified coccinellid insect (No. 20) is likely predaceous both as a larva and an adult on various sessile insects (scales, aphids, etc.). The braconid insect (No. 11), collected from a leaf, was likely searching for a host plant to parasitize. The un- identified Microlepidoptera insect (No. 14) may be an important pollinator, as Collett (1999) reported nocturnal moths visiting and pollinating flowers of A. trichopoda in greenhouses at the Santa Cruz Ar- boretum, California. Some Microlepidoptera retain pollen-scraping mouth parts, e.g., Sabatinca sp. that pollinate Zygogynum in New Caledonia (Thien et al., 1 Insects attracted to the small flowers of Ambor- ella trichopoda range in size from approximately 1.0 mm to over 7 cm in length. The larger insects walk on the branches and crawl through the inflores- cences feeding on pollen, and in the process their ventral surface is coated with pollen. The flat plat- form morphology of male and female flowers and the adhesion of the pollen grains aid in pollen transfer as the ventral surface of the insect scrapes over the flower. Pollen is also deposited on an in- < micrograph of stigmatic cells show = m. —e. among stigmatic epidermal cells. of multicellular stigmatic trichome. pur Transmission electron micrograph of stigmatic epidermal cells ee tans (JIM13—arrowhead) in the extracellular matrix. 0 Bar = 25 um. Figure abbreviations: C, cuticle; E, extracellular E S, single cell wing thin cuticle (arrowheads) and immunolocalization of pectins (JIM5—arrow). Bar showing immunoloc ти of arabi- Ваг = 0.4 wm. —f. Light micrograph s wing exudate Annals of the 480 Missouri Botanical Garden (pnq) | (0c) I €I (вә#рїш j[e3) әврихшортәә”) ?191du (0) Т П 9epruooP1g (pnq :0) Т OI SISDISAG *aepr[euro12]q :eapropro[eu?) (0) 1 (pnq ‘0) Т ZI ‘6 ‘sou “ds papy oepruoj&in :eapropro[eu? ?19jdouauiÁH I (0) I VI ?191dopido[oaot]g (0D I S (s8nq jeop) aepury вәшәң (DI L pind таюу :aepr[[opeor?) (ZDI g aepuioudoaqd y ?191douioH sak 61 шә @—/ (вәпәәч peusoYy-sU0]) әвшише” :aepro&qure127) (0) I 0c (səpəəq pitqAper) 3epr[[2uro207) © 8 (0) 1 (0) T 8I 9 (шш с—р *[[eurs) pəsur nous I c EDI (sure1a &ueui) [ Ol ^p шә ¢—Z "josur mous әйте] (o&urqoru&uiojd &i7)) әрршоцүпәлпг) S I c (0 8b) c (€) I LI 'SI ‘€ noanpf umijappoow `(вәпәәч Зигхлер) әвицәрұ :aepruoriqoua] (2) 1 c (səpəəq Jeo] әҹц-1ив) әрриәрұ ¢ (LI) I I (вәпәәч ѕп8ипу әјпшш) oeprudo[&ro7) v191doo[o7) ld I ld I ld I ld 1 N чохе, әүешә J 9[?]N ә[ешә4 рәләѕу po1nide? are suourroods тәцәпод ‘1amo pue Jes uo pa41osqo ospe pue paunjdeo = sa :sjoosur paunjdeo uo surg иә[оа jo 1aquinu = ( ) 16бӘАРӘ[ = “(SA Aep ІТ 1240 uone AI9sqo Je s1inou cc A[91eurxoadde) вшорэ[еэ MƏN "notury. P 199 ye оройо) DjJo40quiy Je sjue[d 3[Eui3] pue эеш uo p234129sqo pue paindes S39asu] "uonnijnsu] иегиоѕцищшс əy} Je рәџѕодәр ‘] :sıemoy = [4 Yequinu иопәә[[оә = N € PL Volume 90, Number 3 2003 Thien et al. Amborella Table 3. Number of pollen tubes in stigmas (carpels) of Amborella trichopoda. Data recorded from 208 flowers collected in 1 day from 11 plants along a transect at Plateau de Dogny. Plant Altitude Number of Number of Carpels Flowers Flowers number meters flowers carpels with tubes with no tubes with tubes 1 625 14 58 26 3 11 2 640 26 119 38 6 20 3 640 7 3l 12 1 © 4 660 6 25 7 3 3 5 660 17 80 11 13 4 6 690 27 133 227 13 14 7 700 17 74 1 14 3 8 710 15 70 12 7 8 9 725 26 121 18 14 12 10 730 27 134 19 15 12 11 730 20 98 9 20 6 99 Totals 208 943 178 109 99 18.8% 52.4% 47.5% Unpollinated Pollinated sect's body by other flowers in the same inflores- cence as it crawls over individual flowers to feed on pollen. The male flowers of Amborella trichopoda produce large quantities of pollen, as indicated by its high pollen to ovule ratio (9552 to 6), which suggests both that wind pollination takes place and that ample pollen is available as food and for de- position and transport on the bodies of insects. In addition, the floral phenology of the paniculate in- florescences maintains several functional flowers per inflorescence throughout the flowering season, potentially increasing the rate of pollination. e parasitic Hymenoptera occur in large num- bers on the flowers of Amborella trichopoda (Table 2), although the precise number of observations was not recorded because it was not possible to identify the various species in the low light conditions in the forest. The role of parasitic Hymenoptera as potential pollinators of A. trichopoda is unknown. We are aware of only a few documented examples of pollination of a plant by a parasitoid (Ferguson & Donham, 1999; Nilsson, 1979; Dixon et al., 1990; Feil, 1992; Feil & Renner, 1991). The Cecidomyiidae (No. 13) collected from male and female flowers of Amborella trichopoda appear to be the same species reared from a gall. An insect captured from the male flower carried pollen on its legs, but more of these small insects must be cap- tured and examined for pollen to determine what role they play in pollination. WIND POLLINATION Pollen of Amborella trichopoda was dispersed onto the petroleum jelly-covered microscope slides placed within the crown of male and female plants up to 30 m from the nearest male (Fig. 5). It should be noted that all pollen traps were continuously hung for four days, whereas the average life span of a female flower is one day. Significantly, although the number of grains captured per slide was quite i grains were dispersed onto slides in clumps of 7 to 12 suggesting the presence of a lipid pollen coat, an interpretation subsequently con- firmed by SEM of cryofixed anthers (Fig. 3a). Eight of the 58 slides also captured at least one insect species each of Cecidomyiidae and Thysanoptera. One cecid had both plant material associated wit its ovipositor and microspores associated with its variable, mouthparts (Fig. 3c, 3d). ANALYSES OF OPEN-POLLINATED FLOWERS Pollen germination in pollinated flowers resulted in numerous pollen tubes reaching the base of the stigma (Fig. 3e), but only one to three tubes were observed to enter the ovary locule and one to enter the ovule (Fig. 3f) Flowers of Amborella trichopoda collected along a transect at Plateau de Dogny (a one-day sampling period) were examined for the presence and num- ber of pollen grains on stigmatic surfaces and pol- len tubes penetrating gynoecium tissues (Table 3). In a sample of 208 flowers collected from 11 plants along a transect from 625 to 730 m elevation, 52.496 contained gynoecia in which all the carpels lacked adherent grains and pollen tubes (not pol- linated; Table 3). The data also show that the 208 flowers bore 943 carpels, but only 178 (18.896) car- pels had pollen grains adhering to their stigmas 482 Annals of the MEUS Botanical Garden and/or penetrating pollen tubes (Table 3). The max- imum number of germinating-penetrating pollen tubes in a carpel was 13 taken from a flower col- lected at 625 lyzed contained one to five pollen tubes. Only five m. All other pollinated carpels ana- flowers (2.496), collected on five separate trees, contained gynoecia in which all of the carpels on the receptacle contained at least one germinating grain. These fully pollinated gynoecia were taken from plants 1, 2, 3, 5, and 11 (Table 3) representing a total of 19 pollinated carpels. Although this is only a single sample over a long flowering season, we noted that the stigmas of female flowers re- mained receptive for 1 to 2.8 days, and this may be indicative of the actual rate of natural (open) pollinations. In another sample in which 127 buds and flowers were observed over 8 days (April 2001) at Col d'Amieu, none of the flowers set fruit. FLORAL ODORS Labels on herbarium sheets (IRD, Noumea, New Caledonia) indicate the flowers of Amborella tricho- poda produce a sharp odor. Observations of flow- ering plants in the greenhouses of A. trichopoda indicate that the plants produce an odor late in the evening (Collett, 1999, and pers. comm.). One au- thor (LBT) branches of A. trichopoda. While carrying the spec- detected an odor from flowers on cut imens from the field to the laboratory, a faint odor was detected that smelled like licorice, scented hay, and also at times like feces. Other efforts to detect any odor from the flowers at all hours of the day failed. No specific signals for odor were detected from the samples. More sensitive techniques are currently being devised to attempt capture of odors from the buds and flowers. It has been well dem- that in parasitoid search for host plants (Godfray, 1994), and it there- fore seems likely that they are produced by Am- 2. onstrated odors are involve borella trichopoda. DISCUSSION INSECT VERSUS WIND POLLINATION Field observations, insect collections, and pol- len-load analyses of floral foragers as well as wind traps strongly suggest Amborella trichopoda is pol- linated by representatives of families placed in the insect order Coleoptera and other taxa of insects, as well as by air currents, reflecting а generalist mode of pollen deposition (sensu Faegri & van der Pijl, 1979). Amborella trichopoda shares a suite of characters common to many tropical, woody di- oecious species that occupy the understory with a generalist mode of pollination. In these instances, the dioecious species produce numerous small flowers with few floral attractants ranging in color from white to yellow to pale green with pollination conferred by unspecialized pollinators that typify 1975; 1995). dioecious such understory habitats (Bawa & Opler, 1991; Sakai et al., Insects effecting pollination in Mayer & Charlesworth, many tropical species are generally small. In contrast, both small and large insects appear to disperse pol- len in A. trichopoda. The pollination of Amborella trichopoda by Cer- ambycidae, and the large-bodied Curculionidae, and Tenebrionidae (of which the latter two spend long periods of time walking on the intertwined flowering/vegetative branches of male and female plants) is a distinct mode of beetle pollination (see Bernhardt, 2000). were bisexual, these large insects would undoubt- f the flowers of A. trichopoda edly effect geitonogamy in the absence of any self- incompatibility mechanism. Although pollen feed- ing has tenebrionids (Steiner, some unrelated 1988; Bernhardt, Kooyman & Weston, in prep.), the role of the Adeliini, to which been observed in N. fauveli belongs, as a primary pollinator repre- sents an entirely novel relationship between plants and insects. While both wind and insect pollination occur in Amborella trichopoda the question remains as to which mode of pollination provides the primary and most dependable mode of pollen transport to re- ceptive stigmas. Our current data suggest that wind pollination is a “failsafe mechanism" for insect pol- lination and vice versa. The relative importance of insects versus air currents probably fluctuates with seasonal (climatic) pressures as in some Salix spp. 1990). below the canopy layer of a wet, (Vroege & Stelleman, hile flowering far broad-leafed, ev- ergreen, tropical forest is not associated typically with wind-pollinated angiosperms (Faegri & van der Pijl, 1979) neither is a pollinator spectrum dominated currently by flightless beetles. Whether or not the winged insects associated with the ecol- ogy of the galls on A. trichopoda act as pollinators remains to be determined. This feature will be valu- able to address since cecids and parasitoid wasps that have been demonstrated to use floral tissue as a breeding site in other angiosperms also effect pol- Orchidaceae; 1990; Ferguson & Don- lination (Monimiaceae; Feil, 1992: Nilsson, 1979; Dixon et al., 1999). Insect and wind pollination in Amborella tricho- мІ, Һат, poda most probably share similar modes of pollen dispersal for one important reason. The range of Volume 90, Number 3 2003 Thien et al. 483 Amborella pollen vectors, the dispersal of pollen in sticky- adherent clusters, and the population structure of A. trichopoda at Col d'Amieu indicate that clumped (leptokurtic) pollination has a greater chance of success than either linear (sensu Richards, 1986) or long-distance pollination. This is particularly im- portant as male bushes surround and overlap soli- tary females. In this respect, the combination of insect and short-distance wind pollination in A. tri- chopoda shows some similarities with insect and “gravitational” pollination (sensu Meeuse et al., 1990) in Ephedra spp. Given the phylogenetic position of Amborella tri- chopoda, our results on the pollination biology of the species are consistent with the hypothesis that the first flowering plants had a generalist pollina- tion mechanism (Bernhardt & Thien, 1987). Fossil evidence indicates that the earliest pollen vectors lacked both elongated nectar-collecting probosces and specialized behaviors to shake, vibrate, or ma- nipulate anthers to remove pee other edible 1997, rewards (see Grimaldi, 1 : Labandeira, i t has been pos- 1998; Bernhardt & Thien, 1987). 1 ited that these early pollinators included beetles with chewing mouth parts, “short-tongued” but mandibulate moths, sphecid wasps, and early flies now placed within suborder Brachycera (Grimaldi, 1999; Kato & Inouye, 1994; Labandeira, 1998) Do any members of the generalist system rep- resent a long-term association with Amborella tri- chopoda? Wind and weevils (Curculionidae) appear to be the most likely candidates. Although wind pollination is rare among basal angiosperms (Thien et al., 2000), it also occurs within the ANITA grade, i.e., in Brasenia schreberi J. F. Gmel. (Nymphae- aceae; Osborn & Schneider, 1988) and in Trimenia moorei (Oliv.) W. R. Philipson (Trimeniaceae; Bern- hardt et al., 2003). Wind pollination is also an op- tion for Saururus cernuus L., a eumagnoliid species within the Saururaceae (Piperales; Thien et al., 1999; Pontieri & Sage, 1999), which, along with Trimenia moorei, is also pollinated 2 a wide va- riety of insects (Bernhardt et al., ciation between curculionids and planted is an The asso- ancient one (Farrell, 1998), and weevils remain im- portant generalist and specialist pollinators of a wide variety of cycads (Norstog, 1987) as well as basal angiosperms and eumagnoliid species (Bern- hardt, ). For example, members of the genus Elloschodes are the only known pollinators of the relictual Australian eumagnoliid genus Eupomatia (Eupomatiaceae; Bernhardt & Thien, 1987) The mesofossil floras of the Early Cretaceous (Barremian-Aptian) contain many small flowers (Friis et al., 2000). Indeed, the stamens of some of these fossil flowers are similar to those of Amborella trichopoda and members of other basal angiosperm families, including members of the ANITA group (Friis et al., 2000). Although the insects associated with A. trichopoda show more morphological simi- larities corresponding to taxa of the mid-Tertiary taxa, the insect-flower interactions seen in modern A. trichopoda may reveal novel interactions. Nota- bly, the Cucurlionidae and the Tenebrionidae are actually part of a detritus fauna (e.g.. insects crawl- ing from litter to flowers) as reported for fly polli- nators of another ANITA species, Illicium floridan- um Ellis (Illiciaceae: Thien et al., . This feature is significant, as it may give insight into the evolution of floral presentation in the early flowers of the late Jurassic-early Cretaceous (see Mitter & Farrell, 1991, for discussion of insect-plant phylo- genetic relationships). Based on these observations and analyses of pol- len grains attached to insects, it is tempting to sug- gest that non-volant insects (such as wingless bee- tles), maturing within the forest litter, played a unique role in the pollination of the earliest angio- sperms growing under a gymnosperm forest canopy. Along these lines, Feild et al. (2001), using Ат- borella trichopoda as a guide to extrapolate physi- ological characteristics of early angiosperms, noted cautiously that success in shady, wet forest under- story habitats may account for the initial ecological success gained by angiosperms in Cretaceous land- scapes. However, it is also possible that the rela- tionship between non-flying beetles and dioecious A. trichopoda may be merely the consequence of a relictual angiosperm restricted to an island refugi- um in the southern Pacific basin. Carlquist (1974) noted that both wingless insects derived from winged ancestors and plants with unisexual flowers derived from ancestors with bisexual flowers are re- current evolutionary “trends” on tropical Pacific is- lands FLORAL REWARDS VERSUS POLLINATION-BY-DECEIT Insect pollination in angiosperms as well as other seed plants is closely tied to the evolution of at- tractants and rewards produced by the plant for the insect (Thien et al., 2000). Within the basal angio- sperms, attractants include flower color and odors (pollen vs. floral epidermis) and rewards often com- bining physical warmth, breeding sites (copulation vs. brood chambers), pollen, nectar, food bodies, and other secreted metabolites (Thien et al., 2000). Heat and the production of volatile substances are regarded as ancient mechanisms for enlisting insect pollinators that may have evolved before floral pig- 484 Annals of the Missouri Botanical Garden mentation patterns. The first floral odors are posited to have been associated with a lipid-rich pollen coat (van der Pijl, 1960; Porsch, 1950, 1954; Thien et al., 2000). In addition, “protonectar” produced by wet stigmas is believed to be a relictual reward that evolved before the first interfloral (nuptial) nec- tar gland (Endress & Igersheim, 2000 In the absence of detectable leaf or floral vola- tiles, starchy food bodies, free-flowing anther se- cretions, or an edible liquid associated consistently with the stigmatic surfaces, the primary attractant and reward for insects visiting male flowers of Am- borella trichopoda is pollen. Pollen was the only floral component documented to be actively col- lected, in particular by the wingless tenebrionid, N. fauveli. Notably, a lipid-rich pollen coat was de- tected on the pollen grains of A. trichopoda. The pollen coat of A. trichopoda aids in the dispersal of pollen from anthers in clumps (even during wind dispersal). Although the pollen coat has been noted to function in increased pollen removal from an- thers by pollen vectors (Endress, 1994), this feature represents a novel observation for a member of the ANITA group. It is tempting to speculate that the lipid coat may function as a pollinator reward. As well, the pollen of A. trichopoda may have an odor as observed in other species (Dobson & Bergström, 2000), which must be detected with alternative techniques, but the precise role of the pollen coat in the pollination biology of A. trichopoda remains to be determined. In contrast to the pollen coat operating as a po- tential attractant and reward, a free-flowing stig- matic secretion was rarely observed in female flow- ers. Structural and histochemical observations of the stigma of Amborella trichopoda provide evi- dence that it is of the dry-type. The receptive sur- face in dry stigmas consists of a thin protein layer associated with non-specific esterase activity that is secreted onto the surface of the cuticle. The cuticle of dry stigmas remains intact and is only partially disrupted at anthesis. Although wet stigmas may possess an overlying pellicle layer, the receptive surface is composed of free-flowing exudates asso- ciated with a ruptured cuticle. In A. trichopoda, free-flowing liquid is not always present at the stig- matic surface and when present, the structural na- ture of the thin cuticle remains unchanged. The presence of a random stigmatic secretion in A. tri- chopoda parallels observations from a recent field study on the pollination biology of Sarcandra gla- bra (Thunb.) Nakai (Chloranthaceae). Droplets are also secreted occasionally by carpels of S. glabra (Tosaki et al., 2001). However, when these droplets are present they are consumed by insects associ- ated with pollination of 5. glabra. In contrast, no insect was ever found to consume the infrequent, stigmatic secretions of A. trichopoda, suggesting that this fluid is not a significant source of proto- nectar for prospective pollinators. In fact, previous descriptions of wet stigmas in some members of the ANITA group require rein- terpretation. Wet-type stigmas described in //licitum floridanum (Illiciaceae) and Trimenia moorei (Tri- meniaceae) have been shown recently to be dry- types (Koehl, 2002; Bernhardt et al., 2003). The presence of a dry-type stigma in these basal groups calls into question views that a wet stigma was the plesiomorphic condition functioning in pollinator attraction or pollen capture, recognition, retention, and germination (Endress & Igersheim, 2000; Pon- tieri & Sage, 1999). Based on the flowers of extant taxa, it is more likely that a protonectar based on stigmatic secretions evolved independently and along early diverging lineages. Why would insects collecting pollen of Amborella trichopoda ever visit a female plant, as these female flowers contain none of the edible rewards associ- ated with basal angiosperms such as protonectar, pollen. or starch food bodies normally consumed by beetles (Bernhardt, 2000)? We suspect the polli- nation mechanism of A. trichopoda incorporates a form of cryptic dioecy (sensu Mayer & Charles- worth, 1991) in which the comparatively large and succulent staminodes in female flowers mimic the fertile androecia in male flowers (Mayer & Charles- worth, 1991). In cryptic dioecy (sensu Mayer & Charlesworth, 1991), one or both of the functionally unisexual morphs appear to have perfect, hermaph- roditic flowers, making the dioecious condition dif- ficult for both insects (and botanists) to detect. This mode of floral presentation has not been described in many taxa but has evolved independently in some unrelated families including the Myrtaceae, and Trochodendronaceae (Mayer & Charlesworth, 1991). Within the ANITA group (e.g., Nymphaeaceae and other basal angiosperms), floral deception mechanisms in both bisexual and small unisexual flowers attract insects to increase pollination (see reviews in Bernhardt & Thien, 1987; Bernhardt, 2000). Amborella trichopoda utilizes staminodes with no functional pollen and a dry stigma that pro- duces little or no secretions (dry-type stigma). Pol- len is the sole source of food for the visiting insects. Mayer and Charlesworth (1991) noted that in ap- nectar is absent or scarce, Araliaceae, parent androdioecy, which fits the pollinator-attraction hypothesis. In another member of the ANITA group, Trimen- ia moorei, the stigma of the small bisexual flower Volume 90, Number 3 2003 Thien et al. Amborella is also of the dry-type and produces no nectar yet appears to be wet (producing nectar) due to the high pectin content of the cell wall (Bernhardt et al., 2003). In addition, the flowers produce floral odors and one chemical, 2-phenylethanol, is known to attract insects and elicit a reflex extension of the proboscis (Bernhardt et al., 2003). Indeed, flies vis- iting the flowers of T. moorei were observed and filmed extending their probosces to touch the shiny stigma even though no nectar was ever present (Bernhardt et al., 2003). The presence of deceptive mechanisms involving numerous floral traits in small unisexual and bi- sexual flowers of extant members of the ANITA group and other basal angiosperms suggests floral deception played an important role in the evolution of the early angiosperms. These mechanisms (Dufay & Anstett, 2003) provided adaptations whereby small flowers could attract insects to increase rates of pollination at a relatively low cost in resources. POLLINATION AND LIMITS TO SEED SET While it is obvious that many potential saada are lost to gall maggots in this species, our ob of fruits set at Col d'Amieu and analyses ake open- pollinated carpels at Plateau de Dogny suggest strongly that a sheer lack of pollination may also limit seed set in Amborella trichopoda. At certain times in the flowering season and in certain popu- lations of A. trichopoda pollen fails to reach the receptive stigmas on the majority of carpels regard- less of whether the major pollinators are air cur- rents, moths, or wingless, leaf-gleaning insects. This limit to fecundity occurs throughout the an- giosperms when species have discontinuous popu- lations and relatively long flowering seasons but are incapable of mec Жала! modes of self-pollination (see Lipow et al., As Amborella йер is a mass-flowering plant, a low conversion rate of tiny flowers into large fruits may be expected provided it is under- stood that the comparative rate of fruit set between mass-flowering species is extremely variable. Under certain circumstances, for example, almost half of the female flowers of A. trichopoda are likely to be pollinated by single or clumped pollen grains, while only a third of the stigmas of the bisexual florets of Acacia retinodes J. M. Black receive po- 1984). The gynoecium of Acacia is a single carpel containing several ovules, lyads (Bernhardt et al., while the gynoecium of A. trichopoda is limited to a variable number of free carpels and each carpel contains a solitary ovule (Bernhardt, 1989). Polli- nation in А. trichopoda must occur several times if all carpels are to mature into fruits; the sample in this study indicates that this event occurred in less than 3% of the standing crop of open and receptive flowers. CECIDOMYIIDAE Is the relationship with the Cecidomyiidae and Amborella trichopoda indicative of a long-term as- sociation between angiosperms? Gall midges were derived from the Fungivoridea, a Mesozoic repre- sentative (Rohdendorf, 1974). The oldest fossil members are found in the Cretaceous; none of these, however, were phytophagous (Gagné, 1977). Fossil. Cecidomyiidae found in the Upper Oligo- cene-Lower Miocene are representative of present- day taxa (Rohdendorf, 1974; Gagné, 1977). The larvae of all species of extant Cecidomyiidae are detritus eaters (decaying plants, fungi, and live my- celia) except for tribes Cecidomyii, Asphondyliini, Lasiopterini, and Oligotrophini, which can form galls (Roskam, 1992). Asphondyliini and a few La- siopterini form galls in which a layer of fungal hy- phae may serve as food for larvae (Roskam, 1992). It is hypothesized that the radiation of the gall midges coincided with the rise of the angiosperms (Roskam, 1985, 1992; Harris, 1994), which offered new adaptive feeding sites for the insects in the cambium of damaged stems and flowers and in vas- cular systems with large sap flows. Roskam (1992) suggested the diversification of the Cecidomyiidae might also be due to the development of flowers (nutrient-rich sites). Their transition from myceto- phagy to phytophagy completed by the end of the Cretaceous (Mamaev, 1975) may have been facili- tated by fungal infections in damaged flowers (many phytophagous gall midges feed on flowers), and once begun, other plant organs were exploited as 1992; Mamaev, 1975). (1992) further suggested that galls evolved as a re- food (Roskam, Roskam action of the host plant to the irritation of the larvae Roskam, 1992). Most gall- inducing е midges display a narrow host range (Roskam, 1 In extant ТРЕ .l., по galls were recorded (excreted juices, etc.: on plants infected by gall midges, gall wasps, or gall-making sawflies (Roskam, 1992). However, Feil (1992) and Feil and Renner (1991) found that gall midges pollinated members of the Siparuna- ceae s. str, in which eggs are laid in the flowers but gall induction is not involved. In contrast, 25.5% of species in the Rosidae are infected by gall midges, 11.5% by gall wasps, and 19.2% by gall-making sawflies. In the Hamamelidae 8.3% of the species were infected by gall-inducing midges, Annals of the Missouri Botanical Garden 69.196 by gall wasps, and 13.796 by gall-making sawflies (Roskam, 1992). The numbers of fern and gymnosperm species infected with galls were 1.6% by gall midges, 096 by gall wasps, and 4.196 by gall-forming sawflies (Roskam, 1992). The gymno- sperms are thought to have been colonized by gall- inducing Cecidomyiidae later than the angiosperms (Roskam, 1992 Amborella trichopoda, the sister species of all ex- tant angiosperms, is the first member of the ANITA group to be infected by a gall-inducing Cecido- mylidae. The Cecidomyiidae diversified with the angiosperms, but the earliest fossils of gall-induc- ing gall midges occur in the Miocene (Roskam, 1992; Zwülfer, 1978). Phytophagy and gall induc- tion are derived trophic types in the Cecidomyiidae (Roskam, 1992). Thus the gall-inducing gall midg- es are unlikely to have played a role in the early — evolution of the angiosperms (instead they were probably a late invasion similar to the colonization of the gymnosperms). Present-day interactions be- tween A. trichopoda and the gall-inducing gall midges (not to be confused with other interactions of phytophagous Cecidomyiidae in basal angio- sperms; see Feil, 1992; Feil & Renner, 1991) are consistent with aspects of the sequential model of co-evolution or interaction, i.e., no correlation be- tween the evolutionary age of the plant and the in- sect species (Jermy, 1984) Furthermore, Hickey and Doyle (1977) de- scribed a fossil gall on Sassafras potomacensis Ber- ry (Upper Cretaceous, 115 mya, Maryland) that re- sembles a cynipid spangle gall wasp on extant oaks (Larew, 1992), but extant Sassafras (Lauraceae) do not bear such galls (Larew, 1992). Apparently the identification of the leaf bearing the gall is uncer- tain, which limits analysis of the gall with regard to a specific lineage (Larew, 1992). Recent fossil finds of gall wasps indicate that their history ex- tends back to the Upper Cretaceous, and the host plants of cynipids then were species of Papavera- ceae (Ronquist & Liljeblad, 2001). AMBORELLA AND THE EVOLUTIONARY ECOLOGY OF DIOECY Using multivariate analysis, Renner and Ricklefs (1995) correlated the occurrence of dioecy (at the family level) with several attributes, including mon- oecy, climbing growth, biotic dispersal of diaspores, abiotic pollination (wind, water), a shrub habit, and presence in tropical habitats. Only 5% of the extant angiosperm taxa are dioecious, but this arrange- ment of sexual functions is widely distributed in basal as well as advanced taxa (Renner & Ricklefs, 1995). In the ANITA group, dioecy is restricted to the Amborellaceae and some species in the Schis- andraceae (Thien et al., 2000). However, a diverse group of unisexual and bisexual flowers occur in the mesofossil floras (Early Cretaceous) of Portugal (Friis et al., 2000) indicating that dioecy may have evolved early in the flowering plants. Despite putatively severe effects of dioecy on fe- male fitness and strict evolutionary requirements, how long-term evolutionary stability (Heilbuth, 2000; Heilbuth et al., 2001). More important, the fossil record indicates that some are obviously ancient etc., some lineages of dioecious seed plants —. e.g.. Chloranthaceae, Schisandraceae, cycado- phytes, Gnetophyta, etc.) and often show a distri- bution that is highly disjunctive yet very broad and often intercontinental surviving in an equally broad range of tropical and subtropical ecosystems. As mentioned above, dioecious plants are relatively di- verse and dense on oceanic islands including New Caledonia (Carlquist, 1974; Baker, 1959; Godley, 1979; Sakai et al., 1995; Bawa, 1980, 1981). It is hypothesized that on oceanic islands seed dispersal by birds and bats as well as other ecological factors may have played a large role in the evolutionary success of dioecy (Carlquist, 1974; Cox, 1982, and pers. comm.). As noted above, the male and female plants in the study population Col d'Amieu of Amborella tri- chopoda are clumped with female plants tending to be surrounded by male individuals. The reduction in potential seed production on female plants as a result of the gall-inducing Cecidomyiidae, the loss of pollen in males to endophytic parasitoids, and the apparent low dispersal ability, probably com- bine to affect the spatial distribution and fitness of the plants. ny factor that alters gain curves toward line- arity (even if the evolutionary stable strategy for sex allocation is bisexual) can cause dioecy (Charles- worth, 1984). Did floral parasitism (brood parasit- ism) induce dioecy in Amborella trichopoda by al- tering male and/or female gain curves? Weevils do not pollinate flowers of Sagittaria latifolia Willd., but their clear preference for laying eggs on inflo- rescences with male flowers favors higher seed set in all female inflorescences (Muenchow & Dele- salle, 1992). Female wasps actively pollinate the female flowers of figs, but the larvae they leave as eggs ultimately consume some of the ovules (Dufay & Anstett, 2003). Large bats pollinate Freycinetia but they eat the stamens and in the process damage ovules resulting in division of the sexes (dioecy; Cox, 1982; Feil, 1992). It is presumed that Ambor- ella trichopoda once possessed bisexual flowers, but Volume 90, Number 3 2003 Thien et al. 487 Amborella the mechanism(s) inducing dioecy and pathway are unknown. SUMMARY The male and female plants of Amborella tricho- poda are clumped but not clonal in a population. Male and female flowers are small flat structures; male flowers are functional for 4 to 5 days and fe- male flowers 1 to 2.8 days. A lipid coat is present on the pollen and the stigma is of the dry-type. 2. Amborella trichopoda exhibits a wind and insect pollination breeding system; staminodes of female flowers mimic fertile male stamens. The insect vis- itors and pollinators vary greatly in size, and many live in the forest litter and forage on the flowers of A. trichopoda for pollen. 3. Two species of Cecidomyiidae parasitize the car- pels of Amborella trichopoda; this turns the carpels into galls and reduces seed set. Several species of parasitoid insects in turn prey upon the cecid insect larvae in the galls. While dioecy is common in the ANITA group. the combined wind and insect pollination breeding system of Amborella trichopoda utilizing staminodes that mimic fertile anthers in female flowers is rare in dioecious angiosperms in general. 5. Deceptive floral mechanisms are common in the ANITA group and other basal angiosperms and may have played a role in the early evolution of flow- ering plants. Literature Cited Alexander, M. P. 1969. Differential deem of "е апа non-aborted pollen. Stain Technol. 44: 11 APG II Une cii Phylogeny Group). 2003. ген ndis of the Angiosperm Phylogeny Group classification for the orders and families ^ot т iut: АРС II. Bot. m 436 Linn. S Azuma, H., iie Au d ка ма, x е J. С. و ج‎ did au yes L. B. Thien & S. Kawano. 7. Chemi ical divergence in floral scents of Magnolia ra allied genera (Magnoliaceae). Pl. Syst. Biol. 12: 69— 83. —— ——, M. Toyota & Y. Asakawa. 2001. Intraspecific var- iation of floral scent chemistry in ndun kobus E J. Pl. Res. 114: 411 Bailey, I. W. & B. G. L. Swamy. 1948. pom tricho- poda Baill., à new morphological type а vesselless di- cotyledon. J. Arnold Arbor. 29: 245-25 Baillon, H. 1869. Histoire des Plantes, Val 1. Hachette. Pari aris. Baker, H. 1955. Self-incompatibility and establishment af- ter — dispersal. Evolution 9: 347—349 . 19 a а, tive methods as factors in spe- c alii in n. plants. Cold Springs Harbor Symp. Quant. Biol. 24: 177-191. D. Hurd, Jr. 1968. Intrafloral ecology. Annual t Ecol. rns 13: 385-414 фы. T. J., G. Chenery, J. R. McNeal, J. Lyons- Weiler, W. J. тир. ыле A. D. Wolfe & С. Pamphilia. 2000. bisected: and combined a ка Bawa, K. S. 1980. Evolution of dioecy in flowering plants. € Ed Ecol. Syst. 11: 15- . Modes of pollination, vohli and Memes structure in a tropical lo land pra Abst. XIII “1 Bot. Congr. Sydne A. Opler. 1975. Dioecism in tropical forest trees. Evolution 29: 167-179. Benes, K. 1968. On the p of plant cell walls with alcian blue. Biol. Pl. 334—347. Bernhardt, : 1989. The floral eal of Australian Aca- cia. Pp. 263-282 in C. H. on & J. L. Zarucchi dium. Advances in “Legume Biology. Monogr. Syst. Bot. Missouri Bot. Gard. 29. . 1990. Pollination ‘ecology of Oxalis violacea ‘sng alidaceae) а ng a controlled grass fire. Evol. 171: . 2t Ds Gortva evolution and ROSA radia- lion of beetle-pollinated angiosperms. PI. . Evol. 222: 293-320 Thien. 1987. Self-isolation and insect pol- lination in the primitive angiosperms: New evaluations of older hypotheses. Pl. d st. Evol. 156: 159-176. . Kenrick & R. B. Knox. 1984. Pollination bi- ology and the breeding E of Acacia retinodes (Le- guminosae; Mimosoideae). Ann. Missouri Bot. Gard. 71: 17-29. e, Р. ps pa H. Azuma, M. Lam, L. B оен & T Bruhl. The ply dn of Trimenia moorei (Trimeniaceae) Y Flora 1 volatiles, insect/wind pol- len vectors, and stigmatic үр in a basal io ero Ann. Bot. 92: 1-1 Boucek, Z. 1988. Australian Chale 'idoidea (Hymenoptera). A B жөн. ч Revision of Genera of Fourteen Fam- ilies, with a Reclassification of Species. CAB Interna- tional, Wallingford, Oxon, U.K., Cambrian News, Aber- ystwyth, Wales. Carlquist, S. 1974. Island Biology. Columbia Univ. Press, New or Cassar, & M. Blackwell. 1996. Convergent origins of wir We Mycologia 88: 596—6€ Charlesworth, D. 1984. grege and the evolution of dioecy. Biol. J. Linn. Soc. 23: Бн att, R. 1999. Arboretum ч} ба makes headline ws. Bull. Univ. Calif., Santa Cruz Arbor. Assoc. 23: ~ 1 2. Cox, P. A. 1982. Vertebrate pollination and the mainte- nance of dioecism in Freycinetia. Amer. Naturalist 120: 65-80. Crepet, W. L. & E. M. Friis. 1987. The evolution Wes pollination n sms in angiosperms. Pp. 181—201 in E. M. Friis, W. G. Chaloner & P. R. Crane Reed The Фир ч Angiosperms and their Biological Con- bridge Univ. Press, New York. 992. Pollination Ecology: A Practical Ap- mtd h. Oxford Univ. Press, New York. Dejong, D. W., E. F a & A. C. Olson. 1967. Oxi- doreductive айй hydrolytic enzyme patterns in plant suspension culture cells: Local and time relationships. Exp. =, Res. 47: 139—156. Dixon, , J. S. Pate & J. Kuo. . The Wes ee jul haan жш) ые lla pe neri. Pp. 37—62 in J. Arditti (editor), Orchid Biology. 488 Annals of the Missouri Botanical Garden Reviews and Perspectives, V. Timber Press, Portland, Oregon. one H. E. е "pus The Ln and evolution of pollen dur rs. Pl. . Evol. 2 X Doyen, J. T., G. Matthews & J. К. one: Eod Clas- sification and annotated checklist of the Australian gen- era of Tenebrionidae (Coleoptera). Invertebrate Taxon. 3: 229-260. Dreg ger-Jauffret, F & J. D. Shorthouse. 1992. Diversity of gall-inducing ae and their galls. Pp. 8-33 in J. D. Shorthouse & O. Rohfritsch (editors), Biology of Insect- Induced Galls. Oxford Univ. Press, Dufay, M. & M-C. Anstett. 2003. Conflicts ы. еп plants and pollinators that reproduce within inflorescences 00: 3-14 ). Evolution of reproductive structures and functions in асы чы aus rms (Magnoliidae). em. New York Bot. Gard. : —34. ————. 199 i \ me prom Biology of Topic ү“ ш rs. € 'ambridge Univ. P Ca س‎ T - = ^" y 2. 2 апа зон es on ancestral flowers. Int. J. Pl. Sci. 16: 1111-1140. < A. Igersheim. 2000. The reproductive struc- tures of the basal angiosperm Amborella trichopoda (Am aee Int. J. Pl. Sci. 161: 5237-8248. pis B. 1998. “Inordinate fondness” explained: Why e there so many beetles? Science 281: 555-559 бы. . & L. van ijl. 1979. The Principles of л Ecology, 3rd ed. Pársimion Press, New York l'e E 992. Reproductive « ecology of dioecious Sipa- una DM лу in Ecuador—A case of gall midge pollination: Bot. J. Linn. Soc. 110: 171-203 —— —— & 5, S. Renner. 1991. The pollination of Siparuna (Моча е че) by gall-midges (Cecidomyiidae): Anoth- er likely ancient association. Amer. J. Bot. (Abstr.) 78: 37- 186. Feild, T. S., T. Brodribb, T. Jaffré & N. M. Holbrook. 2001. с of leaf anatomy, photosynt thetic light use, and xylem hydraulies to light in Amborella trichopoda (Ambore llaceae). Int. J. Pl. Sci. 162: 999-1008 Ferguson, C. . Donham. 1999, Pollinator of негой Lady's slipper ubi а уи о (Or- chidaceae) Р Oregon. N. Amer. Orchid J. 5: 180—184. Fisher, D. B. 1968. CEP staining of ribboned epon sec- tions for lich microscopy. Histoc hei mie 16: 92—96. Friis, E. M., К. К. Pedersen & P. К. Crane. 2000. Repro- ductive structure and organization of kan angiosperms from the Early Cretaceous (Barremian or Aptian) of western n Int. J. Pl. Sci. 161: 8169-8182. Gagné, R. G. 1977. Cecidomyiidae (Diptera) from Cana- dian eros ine Entomol. Soc. Wash. 79: 57—62. . The ph oW Midges of North Amer- Ica. ак i Ay Ithaca, New York. Godfray, H. C. 994. Paced ee pave and Evo- lutionary Ecology Princteon Univ. Press, New Jersey. Godley, E. J. 1979. Flower biology in New Zealand. New Zealand J. Bot. 17: 441—460. Goldblatt, P. & P. Bernhardt. 1990. Pollination biology of Nivenia (Iridaceae) and the evan e B self- о га Во 93— Graham, S. W. & R. ‚Шы 2000 Utility of 17 c hloroplast genes for ү мага the PE Ape uh the bas- al angiosperms. Amer. J. Bot. 87: 1712-17 ————, P. A. Reeves, A. C. E. Burns & 2 с TE 2000. Microstructural changes in noncoding chloroplast DNA: Interpretation, evolution, and utility of indels and 1eterostylous — =- inversions in basal angiosperm phylogenetic inference. Int. J. Pl. Sci. 161: 583-896. Gressitt, J. L. 1982. Ecology and be ae of New Guinea Coleoptera (Beetles). Pp. 7 134 in J. sitt (editor), Biogeography and Jm of New Guinea, Vol. 2. Dr. W. Junk, The Hague. Grimaldi, D. 1999, The co-radiations of pollinating insects and angiosperms in the Cretaceous. Ann. Missouri Bot. Gard. 86: 373—406. xri . E. 1999. An annotated catalog of world Me- gastigminae (Hymenoptera: Torymidae). pea Gurr E 1965 „ Gres- Chak Entomol. Inst. 31: 1-9 . The Rational Use of Dyes in Biology. Hill, 'idoidea: J2. lon. ‚ К. M. 1994. Gall midges (Cecidomyiidae): Clas- кз айй and biology. Pp. 203-211 in M. А. J. Williams а р : ll Organisms, Interaction, Popula- tions. The Sys ae 5 pes Axford. TS UI" T n n. of E Angus & Robertson, ga Is К, South Wales. Hawkins, B. A. Gagné. 1989. Determinants of asse bisce size in or D: и: zecidomyiidae (Dip- tera). "e ge 8l: — —— 4 К. Go cm on Organization of a para- sitoid | community associated with a complex of annie on ае x spp. in southern California. Ecol. Entomol. -292. — EC. J. € ;. 2000. Lower species richness in dioecious clades. iu l. Naturalist 156: 221-24 . P. Otto. 2001. The consequences of ives y an seed dispersal: Modeling the seed-shadow 38 3. „ Пуеѕ & S handicap. Evolution 55: Hickey, L. J. & J. A. Doyle. 1977. Early Cretaceous ~ evidence for e EAR e a Bot. Rev. 43: 3-1( i T. 1994. Distribution T ecology of e cnr. of “кз Рр. 171— in N. J. En г К. $. Hill (editors), Conifers з the ees D аы Melbourne Univ. Press, Melbo & J.-M. ара 1995. а) lar floristic 1 forest on schist in New Cale- donia: A comparison ‘with an iens rain us "st. Bull. Mus. Natl. Hist. Nat., Sér. 4, sect. Adansonia 17: 201-226 characteristics of a — ———, P. Morat, J.-M. Veillon, F. Rigault & G. Dagostini. 2001. ( Comin and characterization of the native flora of New Caledonia. пае Sc е el Techniques ns IRD, Centre de Noum 13. Jensen, W. A. 1962. iae al Histoc hs tnis Principles and Practice. Freeman, London. Jérémie, J. 1982. Ашен 'eae. Pp. 157-160 in A. Au- breville, J.-F. Jérémie (editors), Flore de la Nouvelle Calédonie et Dependances. Museum National a Histoire Naturelle, Laboratoire de Phanérogamie, Par- Le roy & J em. T. 1984. е zA insect/host plant relation- s . Amer. Naturalis : 609—630. Kaszab, Z 82. Die Tene los ən Neukaledoniens und der ош азы (Coleoptera). Folia Entom. Hungar- ica. 43: 1-294. Kato, M. & T. posa 1994. Origin of insect pollination. Nature 368: Kearns, C. hw Inouye, 1993. Techniques for Pol- oe Biologists, Univ. Press Colorado, Niwot, ado. Kehr. V. 1999. Die “Ambrosiagalle” als tripartites Inter- aktionssystem: Macrophomopsis coronillae (Desm.) Pe- 2 Colo- Volume 90, Number 3 Thien et al. 489 2003 Amborella trak in Gallen von Asphondylia i er ES an Lotus , F. W. 1909. Arbevsiapilse. П. re и 2 Biblioth. Mycol. cornic p L. Koehl, 002. Functional is live se of Il- licium есе (Illiciaceae). MSc., University of To- ronto. Labandeira, C. 1997. Insect mouthparts: Ascertaining the paleobiology of insect feeding strategies. Annua g . 1998. How old is the flower and the fly? Science 280: d Larew, Н. С. 1992. Fossil galls. Pp. 50—59 in J. D. Short- house & ‘0. Rohfritsch (editors), Biology of Insect-In- duced Galls. Oxford Univ. Press, New Yor Lipow, S. R., P. Bernhardt & N. Vance. 2002. Comparative rates of pollination and fruit-set in widely separated Ep. И? : a rare orchid еи саб тив) Int. J. Pl. Sci. 163: 775-782 Lowry II, P. E 1998, Diversity, ud and extinction in the p of New Caledonia: A review. Pp. 181—206 in C & P. P. Lowry II (editors), Rare, Threat- . and p hier Floras of Asia and the Pacific Rim. Institute of Botany, Academia Sinica Monograph Series No. 16, Taipei. Mabberley, D. J. 1993. The Plant-Book. Cambridge Univ. Press, Cambridge. Malloch, D. 1981. Moulds: Their г NUM and Identification. Univ. Toronto Mamaev, B. M. 1975. Evolution of Сай Forming 8 )اع‎ Gall Midges. The British г Division, Wetherby. Row original: Le Mani, M. S. 1964. Ecology of Plant Calls. Dr. W. Junk. The Hag Mathews, s. & M. J. Donoghue. 1999. The root of angio- sperm phylogeny inferred from duplicate phytochrome genes. Science 286: 947-950. Matthews, E. G. 1998. Classification, phylogeny and bio- geography of the genera of Adeliini (Coleoptera: Tene- brionidae). Invertebrate Taxon. 12: 685—824. Mayer, S. S. & D. Charlesworth. 1991. pid = in flowering alanis, kae Ecol. Evol. 6: 320— , A. e Meuer, O. W. P. Ка 5. М. 1990. Ento пы. in the dioecious gymna- sperm Ephedra aphylla Forsk. (= E. alte with some notes on Ephedra campylopoda c Ш. Further anthecological studies ем исар impor- lance of entomophily. Israel J. Bot. : 123. Mitter, C. & B. Farrell. 1991. Mac revlon aspects of insect-plant relationships 8 in E. Bernays (editor), Insect-Plant ta ol Ш. CRC Press, ога. 22 . W. Bailey & В. С. L. Swamy. 1950. The idw p relationships of the Monimiaceae. J. Arnold Arbor. 31 2-404. Morat, Ph., J.-M. Veillon & Н. S. Mackee. 1984. Floristic relationships of New Caledonian r forest ame gams. Pp. 71-128 in F. J. Radovsky. n H. Sohmer penr Biogeography A the Tropical Pa- Assoc. Sys Lawrence, Kansas, and Bernice P. Bishop Vocum. Honolulu. Hawaii zw J. J.. C. Della Lucia & R. M. C. Mucho- . 1991. TS sp. nov. zs a nest of эе Fu. ycol. . 95: 1308-1311. .А. Delesalle. 1992. enm of weevil monoecious and female inflores- Amer. Midl. Naturalist c к= herbivory оп pip cences of Sagittaria latifolia. 127: 355-361 pone. Ber. Deutsch. Bot. Ges . 1910. Ambrosio эц Ш. Waiters Бааша an Ambrosiagallen. Ber. Deutsch. Bot. Ge 455- 480. Nilsson, L. A. 1979. The pollination ecology of Herminium mier A eee t. Not. 132: 537-549. lorstog, L. 987. Cycads and ihe origin of insect pol- lination. Amen Sci. 75: 270-279. O’Brien, T. P. & M. E. McCully. 1981. The Study of Plant Structure: Principles and Selected Methods. Termarcar- phi, Victoria. Osborn, J. M. & E. L. Schneider. 1988. Morphological studies of the Nymphaeaceae sensu lato. ХУІ. The floral biology of panels’ schreberi. Ann. Missouri Bot. Gard. 75: 778-794, Parkinson, C. s & J. D. Palmer. 1999. Multigene analysis identify pi three earliest m of extant flowering plants. Curr. Biol. 9: 1485 Pintaud, J.-C., T. Jaffré & H. Puig. 2001. Clg = New Caledonian palms and possible evidence of P tocene rain forest refugia. C. R. Acad. Sci. Paris, Sci. 324: n Eo Plakidas . Notes on gall inhabitants of Asphon- dylia олоон Osten Sacken (Diptera: Ceci- don mn in western Pennsylvania. J. New York En- жел "an 2-4. A. Weis 1994. Depth associations and utili- n the parasitoid guild of РОИ Envi zation pred rns pudbeckiaeconspieua oe Cecidomyiidae). ronm. Entonx 1. 23: 115-121. Pontieri, V. & T. L. Ар. e. 1999. Evidence for stigmatic self-ine ompatibiin pollination induced ovule enlarge- ment and transmitting tissue exudates in the paleoherb, SM cernuus L. (Saururaceae). Ann. Bot. 84: 507— а a O. 1950. Geschichtliche Lebenswertung der Kas- — Oesterr. Bot. Z. 97: 359-372. А Geschlee thtsgebundener Bliitenduft. Oes- 101: 359-372. P. B. Tomlinson, 2003. ig ts of inflo- ulative basal lerr. E ^ Posluszny, U. & scence and floral development in the PV rm е trichopoda Арген, anad. J. Bot. —39. Tue .-L., J. Lee, a ade oni- Mieux у D. E. Soltis, P. S. Soltis. M. Zanis, er, Z. Chen, V. Sal- volainen & M. Chase. 1999, The earliest angio- sperms: Evidence frat mitt ror ees and nu- clear genomes. Nature 402: 404—407 Phylogeny of basal‏ .2000 س angiosperms: iru sis of io genes from three genomes t. J. Pl. к^ 53—527. ‚ J. Lee, . Whitlock, F. Bernasconi-Quadroni & O. e 2001. Was the ANITA rooting of the angiosperm phylogeny affected ET ao attrac- tion? Molec. Biol. Evol. 18: 174 T" L. 1983. Pollination Biology. Heim Press, Flor- Regal. P. J. 1977. Ecology and evolution of flowering plant E Science 196: 622—629. Renner, S. S. & J. P. Feil. 1993. Pollinators x tropical dioecious angiosperms. Amer. J. Bot. 80: 1100—1107. — ——— & К. E. Ricklefs. 1995. ay and its pe in the wur plants. Amer. J. Bot. 82: 596-606. Richards, . 1986. Plant Breeding [S Allen and Unwin, ae 490 Annals of the Missouri Botanical Garden Rohdendorf, B. 1974. The Historical Developme nt of Dip- era. Translated from the Russian by - Moore and 1. Thiele. Univ. Alberta Press, Edmor Ronquist, F. & 1. Liljeblad. 2001. Evolution of the gall wasp-host ro association. Evolution 55: 2503-2522. Hoskam, J. € Evolutionary patterns in call midge- 1081 slant assoc ie SS Cec domi n i Би, 128: 193-213 ——— Evolution 7 the gall-inducing guild. | n J. D. Shorthouse & O. Rohfritsch (e du Biology of Insect-Induced Galls. Oxford Univ. Press, New York Sakai, A. K., W. L. Wagner, D. M. Ferguson & D. R. Herbst. 1995. Origins of dioecy in the Hawaiian flora. Ecology 76: 2517-2529. Sampson, F. B. 1992. Pollen morphology of the s ellaceae and Hortoniaceae (Hortonioideae: Monim aceae). Grana 32: 154—162 Soltis, P. S., D. E. Soltis & M. sperm apii ny inferred genes 1999, Angio- from КЫ chloroplast for comparative biology. Nature 402: 402-40: Steiner, us E Jr. 1988. Larval ن‎ and biology of East- ern U. S. Epitragini (Coleoptera: Tenebrionidae), with the P 'overy of a new form of pidum ‘gin-trap.” [Ab- stract.| Proc. XII Intl. Congr. Entomol. Vancouver, Can- ada: 51. Sun, С. Q. Ji, „ Dilcher, S. Zheng, K. C. Nix Wan. 2002. i lacini Часеае, a new basal angiosperm family. Scie p 296: 899-904. Thien, L. B., Wik & L. Yatsu. 1983. The repro- ductive holy of a m тш floridanum Ellis. Amer. J. Bot. 719-72 ———, P. e qued G. W. G ibbs, O. Pellmyr, € TE strüm, I. Groth & G. McPherson. 1985. The pollination of Zygogynum (Winteraceae) by a moth, Sabatinca (Mi- cropterigidae): An ancient association? Science 227 540-543 e 7-5 —— Ellgaard, M. Devall, S. Ellgaard & P. Ramp. 1999. ыш structure and reproductive biology of Saururus cernuus L. (Saururaceae). Pl. Sp. Biol. € 55. — a & S. Kawano. 2000. New perspectives on the О biology of basal angiosperms. Int. J. PI. < 61: 5225-3235. жы Ү Ка S. Renner & Н. Takahashi. 2001. Pollina- tion E nin glas (С dicia 'eae) in. natural populations in Тара 1. J. РІ. Res. 114: 423—427. van der Pijl, L. 1960. Ecologie al: pn of flower evolu- tion. I. Phyletic evolution. Evolution 14: 403-416. Vroege, P. W. & P. Stelleman. 1990. Insect and wind pol- lination in Salix pens L. and Salix caprea L. Israel J. Bot. 39: 125— Waring, G. & P. Price. 1989. Parasitoid pressure and the radiation of a gall-forming group (Cecidomyiidae: As- phondylia spp.) on creosote bush (Larrea tridentata). 93-299, . Kooyman. 2002. Systematics of Eidothea (P roteace ваг), w species, E. oo north-eastern New South Wales. Telopea 9: 821—832 Wilson, D. 1995. Fungal a s which invade insect galls: Insect pathogens, benign ie id or fungal inquilines? Pars 103: 255-20 Zanis, M. J., D. E. Soltis, P. S. Soltis, з Mathews & M. J. Donoghue 2s The root of the angiosperms revis- ited. Proc. Natl. Acad. Sci. U.S.A. 99: Fun Zwülfer, H. 1978. Med und Ergebnis evolution von phytophagen und xe касың sek und höheren д anzen. Sonderb. Naturwiss. Verei se der Cen ‘gies and counterstrategies in insect ка systems competing for space and E in اق‎ flower heads and plant galls. Fortschr. Zool. 25: 353 Volume 90, Number 3, pp. 319—490 of the ANNALS OF THE En RI BOTANICAL GARDEN was published on October 10, 20( issouri Botanical Garden ТОО 0313 0 www.mobot.org/mbgpress CONTENTS kivanni of Eurasian and North ‘African Doronicum (Asteraceae: Senecioneae) ............ Inés Álvarez ЖИНИ A Synoptic Review of the African Genus Hesperantha (Iridaceae: Crocoideae) -....... Peter Goldblatt Biogeography and Floristic Affinities of the Limestone Flora in Southern Yunnan, China H. Zhu, H. Wang, B. Li & P. Sirirugsa The Population Structure and Floral Biology of Amborella trichopoda (Amborellaceae) ..... RE Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, e. Vineenza Pontieri, Peter H. Weston, Dave Malloch, Hiroshi Azuma, ; Sean W. Graham, Marc A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre 466 Cover illustration as fede NE Plate TO li Кө Rossica, edited ui PS, Pallas, K.F. Friedrich, and J.J. Weitbrecht and published by Petropoli, Leningrad, 1784-1788. A book — ` of hand-colored copper пиа of plants indigenous to European and Asian areas of the _ Russian Empire, and one of a CE collection. of "ex books i in age же of the Missouri Ч Botanical Garden. Lot Annals ofthe | Missouri Dotanical Garden umber Volume 90, Number 4 Annals of the Fall 2003 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. All manuscripts are peer-reviewed by quali- fied, independent reviewers. Authors should write the Managing Editor for informa- tion concerning arrangements for publishing in the ANNALS. Instructions to Authors are printed in the back of the last issue of each volume and are also available online at www.mbgpress.org. Editorial Committee Victoria C. Hollowell Editor, Missouri Botanical Garden Amy Scheuler McPherson Managing Editor, Missouri Botanical Garden Diana Gunter Associate Editor, Missouri Botanical Garden Kevin Brown MBG Press Assistant, Missouri Botanical Garden Barbara Mack Administrative Assistant 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 Charlotte Taylor Missouri Botanical Garden Henk van der Werff Missouri Botanical Garden For subscription information contact ANNALS OF THE MISSOURI GARDEN, % Allen Marketing & Management, Р.О. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2004 is $150 per volume U.S., $160 Canada & Mexico, $185 all other countries. Four issues per vol- ume. The journal Novon is included in the sub- _ scription price of the ANNALS. annals@mobot.org (editorial queries) http://www.mbgpress.org THE ANNALS OF THE MISSOURI BOTANICAL GARDEN (ISSN 0026-6493) is published quar- terly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. Periodicals postage paid at St. Louis, MO and additional mailing offices. POSTMASTER: Send address changes to ANNALS OF THE MISSOURI BOTANICAL GARDEN, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. The Annals are abstracted and/or indexed in AGRICOLA ae 179 APT Online, BIOSIS®, ingenta, ISIG databases, JSTOR, Research Alert®, and Sci Searc © Missouri Botanical Garden 2003 The mission of the Missouri Botanical Garden is to discover and share к about a and their environment, in order to өре and enrich life. Thi + Sd 19924 nanence of Paper). : ла au patie ad Sy: ee Volume 90 Number 4 2003 Annals of the Missouri Botanical Garden WZ A REVISION OF THE IMBRICATE GROUP OF STYRAX SERIES CYRTA (STYRACACEAE) IN ASIA! Yelin Huang, Peter W. Fritsch, and Suhua Shi? ABSTRACT veral taxonomic treatments of Styrax — the Asian species of Styrax series Супа defined species S. hookeri and S. eastern Asia, гне асеае, shiraianus, S. supaii, ey words: Styrax, To help re with imbricate eee aestivalion, Our revision оотрги ses combined distribution from [Japan south to Sumatra and west 1 and S. serrulatus are clarified. Styrax agrestis var. curvi т and lectotypes are ides ted for S. duclouxii, S. floribundus, S. p aarian S. hookeri, S. hookeri var. y S. hypoglaucus, S. а 5. heus "hti, S. macranthus, Styracaceae) exist in regional floras of Asia, but the Asian spec ies of the genus have not been comprehensively revised since 1907. ctify this, we conducted a taxonomic revision of nnanensis, S. obassia, S. perkinsiae, S. serrulatus var. latifolius 5. . Keys, descriptions, and я maps are provided for all specie Styrax series Cyr Styrax L. comprises about 130 species of trees and shrubs distributed in eastern and southeastern a, the New World, and the Mediterranean region (Fri itech, 1999). The range of this genus is typical of many plant groups distributed among the refugia of Tertiary mixed-mesophytic forests in the North- ern Hemisphere, except that it also includes a large Neotropical component that extends south to north- ern Argentina and Uruguay (Fritsch, 1999, 2001). Styrax is by far the largest and most widespread of the 11 genera in the Styracaceae sensu Fritsch et (2001) and Fritsch (in press a). Characters unique to Styrax in relation to the family include a stamen tube attached high (vs. low) on the corolla, al. the presence (vs. absence) of placental obturators, bitegmic (vs. unitegmic) ovules, and an indu- ! We thank the curators of the herbaria listed in the Materials and Methods section, who kindly made specimens available. Juan Ochoa for help with gae maps, and Meg S PC deus 2 aries tura cript. This w 302: 300: 30 and 30825104 to S. Shi. providing E funds for Y. Hua ? State {Нә for Biocontrol, ladbO3zsu Ai Eran We thank the School g); Issssh@zsu.edu.en (S. Shi). California Academy 3 artment of e pfritsc ا‎ alacademy.or, of Sciences, e are especially grateful to Bruce Bartholome w for help with the collection database and ArcView software, stalcup for illustrations. We also thank ssistance, and Walter Judd, Jun W en, anchi Gandhi and and an anonymous re akeside Foundation of the California Academy of Sciences for ang to spend nine i i in the U.S.A for this study. of Life Sciences, Sun Yatsen po Guangzhou 510275, China. San Francisco, California 94118-4599, U.S.A. ANN. Missouni Bor. GARD. 90: 491—553. 2003. 492 Annals of the Missouri Botanical Garden rate (vs. thin) seed coat; these have been identified as putative synapomorphies for the genus (Fritsch, 1999; Fritsch et al., 2001). The combination of the following characters also serves to delimit Styrax from other genera of Styracaceae: absence of bud scales, presence of pseudoterminal fertile shoots (except in 5. macrocarpus W. C. Cheng; presumably a reversal), a short hypanthium, unarticulated ped- icels, glossy stamen filament trichomes that are cir- cular in cross section, a 3-carpellate ovary, the presence of mesocarp, and a seed-to-carpel ratio S (Fritsch et al., 2001; Fritsch, in press a). Like other Styracaceae, Styrax has a vestiture of stellate trichomes (in some cases modified to peltate scales or rarely simple trichomes), generally twice the number of stamens as petals, and introrsely dehis- cent anthers with a large, linear connective (Fritsch et al., 2001; Fritsch, in press a). TAXONOMIC HISTORY AND PRESENT OBJECTIVES In the most recent worldwide monograph of the genus (Perkins, 1907), Styrax was divided into sec- tion Styrax, with 16- to 24-ovulate gynoecia (most of the genus) and section Foveolaria (Ruiz & Pav.) Perkins, with 3- to 5-ovulate gynoecia (2 Neotrop- ical species). Section Styrax was in turn divided into series Styrax (= series Imbricatae Perkins, in- valid name) and series Valvatae (Gürke) Perkins, each defined, as the names suggest, on the basis of corolla aestivation. Despite the use of aestivation type for infrageneric delimitation, Perkins (1907) acknowledged that some species of Styrax placed in series Valvatae are occasionally slightly imbri- te (= "subvalvate"), whereby the edges of the corolla lobes are contiguous but oblique in cross section, with a mixed condition of valvate and sub- valvate aestivation sometimes occurring within one and the same flower. On this basis, Steenis (1932), in a revision of the Malesian species of Styrax, did not recognize either series of Perkins and placed several imbricate-flowered species of series Styrax (S. subpaniculatus Jungh. & de Vriese, S. porteri- anus G. Don, and S. subdenticulatus Miq.) under 5 serrulatus Roxb., a species with otherwise valvate to subvalvate aestivation Fritsch (1999) conducted à morphological phy- logenetic analysis of Styrax and revised the infra- generic classification of the genus based on the re- sults. In addition to corolla aestivation type, several other characters diagnosed the deep divergences of the Styrax topology, whereas clades diagnosed by a reduced number of ovules per gynoecium were highly nested. In the recireumscribed sectional and series classification, the clade corresponding to the deciduous section Styrax (about 33 species) was supported by the presence of young shoots with scattered stalked stellate trichomes distinct from the rest of the vestiture patterns (vs. without stalked trichomes unless accompanied by a dense tomen- tum consisting of trichomes of the same general type) and membranaceous (vs. subcoriaceous) co- rolla lobes, whereas the clade corresponding to sec- tion Valvatae Gürke (about 97 species) was sup- ported by valvate (vs. imbricate or subvalvate) corolla aestivation, the evergreen (vs. deciduous) condition, sides of the corolla straight (vs. convex) in bud, and concave (vs. planar) stamen filaments. The delimitation of these two species groups cor- responds roughly to a geographic distribution in warm-temperate versus humid-tropical regions, re- spectively. Within section Styrax, the clade corre- sponding to series Styrax (3 species, western North America, Mediterranean region) was supported by strictly pseudoterminal (vs. pseudoterminal and lat- eral) inflorescences, whereas that corresponding to series Cyrta (Lour.) P. W. Fritsch (about 30 species, eastern. and Asia, southern North America) was supported by glandular-serrate (vs. southeastern entire) leaf margins. The character states of corolla aestivation delim- ited in the morphological analysis reflected the dis- tinction made previously (Perkins, 1907; Steenis, 1932) between a truly valvate type of corolla aes- tivation and the subvalvate type. The Fritsch (1999) demonstrated that valvate tion as defined by Perkins has evolved at least twice in Styrax, once in the most recent common results of aestiva- ancestor of the evergreen species and once (as the subvalvate condition) in the deciduous species. Therefore, according to Fritsch's revision, valvate aestivation is possessed by all members of section Valvatae, imbricate aestivation is possessed by all members of series Styrax and some members of se- ries Cyrta, and the remaining members of series Cyrta possess the subvalvate type. The morpholog- ical analysis of Fritsch (1999) supported the idea of Hwang (1999) that imbricate aestivation is the primitive state in Styrax. A molecular phylogenetic based on DNA sequences from the internal tran- scribed spacer (ITS) region of nuclear ribosomal DNA, both separately and in combination with mor- phology, provided strong support for the series clas- sification of Fritsch (1999, 2001). The ITS phylog- eny the classification, although a combined analysis recov- ered a topology consistent. with it. A family-wide analysis of Styrax sectional ambiguous as to phylogenetic analysis based on DNA sequences of the chloroplast genes rbcL and trnL-F in combi- Volume 90, Number 4 2003 Huang et al. 493 Revision of Styrax Series Cyrta nation with ITS sequences and morphology (Fritsch et al., 2001) provided some support for section Sty- rax, but support for section Valvatae was ambigu- ous. The placement of Huodendron Rehder as sister to Styrax with strong support rendered the original state for corolla aestivation in the genus ambiguous because both genera are polymorphic for this char- acter. The monophyly of the subvalvate members of series Cyrta as predicted from morphology was not supported by the ITS results (Fritsch, 2001) be- cause several major clades contained both subval- vate and imbricate species. For example, S. japon- icus Siebo ^ucc., a species with imbricate aestivation, grouped strongly with S. formosanus Matsum., a species wit subvalvate aestivation. Fritsch (2001) concluded that reticulate evolution may at least partly explain the discordance between morphological and molecular data in series Cyrta but cautioned that the absence of chromosome counts for most species hinders further progress on this issue. During the course of our study it became clear that, despite the discordance between morphologi- cal and molecular data and the conclusions of Per- kins (1907) and Steenis (1932), corolla aestivation is species-specific without exception in series Cyrta and apparently is one of the few discontinuous and potentially phylogenetically informative characters in the series. This indicated to us that the treatment of S. serrulatus by Steenis (1932), in which a mix- ture of distinctly imbricate and subvalvate types of aestivation was postulated, would require careful re-evaluation. Furthermore, our study of regional floristic treatments of the genus for Asia (Steenis, 1932, 1949; Hwang, 1987; Svengsuksa & Vidal, 1992; Yamazaki, 1993; Hwang & Grimes, 1996; Y. Lee, 1996; Long, 1999) suggested that the spe- cies of series Cyrta are often poorly understood tax- onomically across political boundaries, likely through the lack of comprehensive examination of Asian types and other collections. Here we provide a taxonomic revision of the members of Styrax series Супа in Asia with im- bricate corolla aestivation (17 species). Restricting our revision to the species with imbricate aestiva- tion provides a practical limit to the scope of the study and is not meant to suggest that the group is necessarily monophyletic. A revision of the remain- ing species of series Cyrta (i.e., those with subval- vate corolla aestivation) is anticipated as part of a comprehensive revision of Styrax. The four North American species of the series (S. a Lam., S. glabrescens Benth., S. grandifolius Ait., and S. jaliscanus S. Wats., all with imbricate aes- tivation) are treated in more taxonomic detail else- mericanus where (Gonsoulin, 1974; Fritsch, 1997; Fritsch, in prep.), but are included in the key to species and various discussion sections to provide complete coverage and facilitate identification of cultivated material. GEOGRAPHIC DISTRIBUTION AND ENDEMISM The 30 or so species of Styrax series Cyrta occur in temperate lowland to tropical montane forests of eastern and southeastern Asia and North America with 80—300 cm mean annual precipitation and without a prolonged dry season (Fritsch, This intercontinentally disjunct distribution is com- mon in many plant and animal groups, and is best explained by periods of relatively warm climate in the Tertiary that allowed widespread Northern Hemisphere distributions and transcontinental mi- gration of lineages across the Bering and North At- lantic land bridges. Cooling and drying trends over the course of the Tertiary eventually restricted these lineages to ura refugia" today (Wolfe, 1975; Тлеу 1985a, b, 2000; Tiffney & Manchester. 2001). Most of the species in the series (ca. 26) occur in eastern and southeastern Asia, consistent with the general pattern of higher species richness in eastern Asia versus eastern North America in plant genera disjunct between these two regions (Wen, 1999). The remaining 4 species occur in the eastern United States (S. americanus, 5. grandifol- ius), southwestern Mexico (Jalisco and Nayarit; 5. jaliscanus), and eastern and southern Mexico to Costa Rica (S. glabrescens). In Asia, the imbricate members of Styrax series Супа exhibit a combined distribution from Hok- kaido, Japan, south to Sumatra, Indonesia, and west to Mechi, eastern Nepal. The 17 Asian species here recognized are all endemic to the region of interest. Most species possess a range that overlaps that of at least one other species in the group, except the southernmost species S. curvirostratus (B. Sveng- suksa) Y. L. Huang & P. W. Fritsch, 5. porterianus, and S. subpaniculatus. Styrax japonicus and obassia Siebold & Zucc. both exhibit a disjunct dis- tribution among China, Korea, and Japan. Styrax japonicus extends along the Ryukyu Islands south to the northernmost islands of the Philippines, by- passing Taiwan. The most common and widespread species in more or less relative order are 5. Japon- icus, S. obassia, S. odoratissimus Champ. ex Benth., S. hookeri C. B. Clarke, and S. tonkinensis (Pierre) Craib ex Hartwich. Species that can be considered narrow endemics are S. buchananii W. W. Sm., 5. chrysocarpus Н. L. Li, 5. curvirostratus, S. macro- carpus, S. porterianus, S. shiraianus Makino, S. sub- Annals of the Missouri Botanical Garden paniculatus, S. supaii Chun & F. Chun, and S. wil- sonii Rehder, i.e., 53% of the Asian species of the group (Table 1). Two other series of Styrax have representatives in Asia. Styrax officinalis L. of series Styrax occurs in the eastern Mediterranean region and extends into southwest Asia in Cyprus, Israel, Jordan, Leb- anon, Syria, and Turkey. All ten or so species of series Benzoin P. W. Fritsch are endemic to eastern and southeastern Asia. The species of series Ben- zoin are easily distinguished from those of series Cyrta by the following characters: plants evergreen (vs. usually deciduous), bases of young shoots with- out stalked ferrugineous or fulvous stellate tri- chomes unless these accompanied by a dense to- mentum consisting of trichomes of the same general color and type (vs. ferrugineous or fulvous stalked trichomes present, distinct from the rest of the ves- шиге), sides of the corolla straight or nearly so (vs. convex) in bud, corolla lobes subcoriaceous (vs. enl кыша or chartaceous), seeds depressed- globose (vs. generally ellipsoid, resting on the hi- lum on a flat surface instead of the sides between the hilum and apex), and seed coat crackled (i.e., coarsely reticulate-sutured; vs. generally smooth or with other types of patterns; see Fritsch, 1999). Al- though the geographic ranges of series Сула and series Benzoin overlap nearly completely, the spe- cies of series Cyrta tend to occur in warm-temper- ate regions, whereas those of series Benzoin tend to occur in subtropical to tropical regions. MORPHOLOGICAL AND TAXONOMIC CHARACTERS Here we discuss the principal diagnostic char- acters used in the systematics of the imbricate group of Styrax series Cyrta in Asia. HABIT All species herein are deciduous shrubs or trees except perhaps Styrax curvirostratus and S. subpan- iculatus, which may be at least semi-evergreen. The tree species are typically less than 20 m tall but occasionally attain a height of greater than ЗО m. ‘Two species are known only as shrubs (Styrax lim- prichtii Lingelsh. & Borza and S. wilsonii). Styrax rugosus Kurz is typically a shrub but can occur as macrocarpus, S. a small tree to 6 m, whereas S. obassia, and S. supaii are typically small trees or rarely shrubs. Styrax grandifolius from the south- eastern United States often forms colonies through root-suckering, but this habit is not known in any Asian species of Styrax. LEAVES Leaves are generally alternate but display two general patterns of phyllotaxis, one with more or less uc spaced alternate leaves (Styrax buch- апапи, S. chrysocarpus, S. curvirostratus, S. odor- atissimus, S. porterianus, S. subpaniculatus, and S. tonkinensis), the other with the two most basal leaves opposite or subopposite (sometimes one or both of these are replaced by scales). The latter condition, occurring on each new shoot, is nearly constant in S. hemsleyanus Diels, S. macrocarpus, S. obassia, and S. ѕирай, but less so in the remain- ing species, especially S. limprichtii and S. rugosus. The petioles of larger leaves are dilated at the base and completely cover the bud in two northern spe- cies (S. obassia and S. shiraianus); this feature is unique to these two species within Styrax. Petiole length differs greatly within and among species and is of diagnostic value in some instances (e.g., S. macrocarpus). The margins of the laminae are near- y always serrate or dentate, with each tooth tipped by a gland. Occasionally (e.g., 5. japonicus, S. por- terianus, S. subpaniculatus), some leaves are entire except for the tooth-like gland. The size and shape of the leaves are variable The leaves of Styrax hookeri, all relatively within many species. and S. common and widespread species, 5. Japonicus, odoratissimus, are especially variable. parallel and perpendicular to the secondary veins The tertiary veins are more or less sub- in most species, but in S. chrysocarpus, S. curviros- tratus, S. japonicus, S. limprichtii, and S. supaii they are more or less reticulate. Typically the leaves of sterile shoots are larger than those of fertile shoots. VESTITURE Although trichome types and the density of pu- bescence on various parts of the plants are useful characters for identifying some Styrax species, high infraspecific variation is common in the genus (Fritsch, 1996, 1997, in prep.). Many species in our revision exhibit such variation (S. buchananii, S. hemsleyanus, S. hookeri, S. japonicus, S. limprichtii, 5. odoratissimus, and S. subpaniculatus). The lower laminar surface in these species can be essentially glabrous or sparsely to densely pubescent, the pu- bescence (if present) consisting of short or long stellate trichomes, or a mixture of both. In diag- nostic terms, these differences are a matter of de- gree rather than kind and exhibit no correlated gaps with other characters, elevation, or geography see discussions under each species in the Taxo- —. nomic Treatment section). The pubescence on re- productive parts of Styrax in our treatment can be Volume 90, Number 4 Huang et al. 495 2003 Revision of Styrax Series Cyrta Table 1. Species distribution, richness, and endemism, by country. *, endemic. No. species/ Country No. endemics Species Bhutan 1/0 S. hookeri China 12/7 S. chrysocarpus, *S. hemsleyanus, S. hookeri, S. japonicus, *S. lim- prichtii, *S. macrocarpus, S. obassia, *S. odoratissimus, S. rugosus, *S. supail, S. tonkinensis, *S. wilsonii India 1 [or 2]/0 S. hookeri, ? S. japonicus Indonesia 1/1 *S. subpaniculatus Japan 3/1 S. japonicus, S. obassia, *S. shiraianus Laos 2/0 5. japonicus, S. tonkinensis Malaysia 1/( S. porterianus Myanmar 5/1 "s buchananii, S. hookeri, S. japonicus, S. porterianus, S. rugosus Nepal 1/0 S. hookeri North Korea 2/0 S. joponicus S. obassia Philippines 1/0 S. japonicus South Korea 2/0 S. japonicus, S. obassia Thailand 2/0 S. porterianus, S. rugosus Vietnam 3/1 *S. curvirostratus, S. japonicus, S. tonkinensis used to identify species such as S. chrysocarpus sometimes two or more arising from the same node (with tric home color), 5. uas (with trichome — (e.g., S. buc hananii, S. hemsleyanus, S. obassia, 5: length), and S. supaii (with trichome type). None- — odoratissimus, S. subpaniculatus, and S. tonkinen- theless, in some species pubescence presence and sis). Lateral T on es are 1- to 2-flowered or amount on reproductive parts varies continuously racemose; they are shorter than the pseudoterminal or sporadically, e.g., on the inner surface of the inflorescence and occur in the leaf axils immedi- corolla lobes and style (S. hookeri), the pedicel and ately below it. We agree with Perkins (1902, 1907) calyx (S. japonicus), or on the surface of seeds (S. thal inflorescence length and flower number per in- odoratissimus and S. tonkinensis). In S. japonicus, — florescence are relatively constant within most spe- the amount and density of pubescence is strongly cies of our revision, and have used these as fun- associated with geography, whereby the most pu- damental key characters. Only S. odoratissimus, S. bescent plants occur in the southernmost portion of tonkinensis, and two North American species (5. the range and least pubescent and glabrous plants glabrescens and S. grandifolius) exhibit significant in the northernmost portion. varialion in this respect (hence each must fall out twice in the key). INFLORESCENCES А А 2 ы FLOWERS All inflorescences in members of series Cyrta are produced on the shoots of the current growing sea- Flowers are bisexual and actinomorphic with a son except those of Styrax macrocarpus, which con- short hypanthium (see Dickison, 1993) adnate to sist of single flowers produced on shoots of the pre- the basal third or less of the ovary wall. Flower vious growing season. The inflorescences of S. length ranges from 0.7 to 3.2 cm. Some species macrocarpus are unique within the genus in this — (e.g., Styrax curvirostratus, S. hemsleyanus, S. hook- respect, although several other genera of Styraca- eri, S. japonicus, S. macrocarpus, S. obassia, and S. ceae show the same pattern (Fritsch et al., 2001). “оле have generally larger flowers than oth- This feature has presumably been derived indepen- ers, especially those whose flowers are consistently dently in the ancestor of these genera and in S. less than 1. 5 em long (e.g., S. odoratissimus, S. por- macrocarpus because Huodendron, the sister group — terianus, S. subpaniculatus, and S. wilsonii). of Styrax, possesses the common state in Styrax. In The Lui pedicels (15—50 mm) of Styrax japonicus species other than S. macrocarpus, inflorescences distinguish this species from all others in our revi- are both pseudoterminal and lateral; occasionally sion. Except for most specimens of S. japonicus, the only pseudoterminal inflorescences are produced abaxial surface of the gamosepalous calyx in Styrax on some shoots of some species, but lateral inflo- is always covered with stellate trichomes. Abaxial rescences can always be found. Pseudoterminal in- calyx pubescence can be used in species identifi- florescences are always racemose or paniculate, cation, e.g., the presence (vs. absence) of various Annals of the Missouri Botanical Garden amounts of scattered dark yellow, orange, or brown stiff stellate trichomes in addition to the base to- mentum (S. hemsleyanus, 5 S. hookeri, S. limprichtii, S. obassia, S. rugosus, S. ка aod S. wilsonii), and a more sparsely pubescent to glabrous region than the rest of the calyx within 1 mm of the margin (S. buchananii, S. curvirostratus, S. hookeri, S. ja- ponicus, S. macrocarpus, S. odoratissimus, S. porter- ianus, and S. shiraianus) versus a calyx that is even- ly pubescent throughout. The calyx margin can be truncate, undulate, irregularly lobed, or distinctly dentate. If the margin is dentate, then the teeth are usually contiguous or separated by a shallow con- cave portion. Styrax supaii is distinguished from all other species by the long, simple or 2-armed tri- chomes covering the abaxial surface of the calyx, and long calyx teeth (4—5 mm long). The gamopetalous corolla is completely white or rarely flushed with pink, and it is nearly always sparsely to densely stellate-pubescent on both sides. Some specimens of Styrax hookeri are gla- brous adaxially. The corolla tube is almost always shorter than the lobes, usually ranging from 2 to 5 mm long. Only S. shiraianus possesses a corolla tube (10—12 mm long) longer than the lobes. In our species, the 5 (to 7) lobes range from 5 to 26 mm long and from 2.5 to 11 mm wide. The stamens are adnate to the corolla tube prox- imally, free distally, and are twice the number of corolla lobes. The corolla lobes and stamen fila- ments become both free and distinct at approxi- mately the same point along the floral axis in all species. Filaments range from 1.5 to 10 mm long and are usually equal or slightly unequal within a flower, but sometimes are distinctly alternately un- equal in length, especially in Styrax supaii. The filaments are flexuous at mid-length in some spe- cies (S. buchananii, S. curvirostratus, S. odoratis- simus, S. subpaniculatus, and sometimes S. porter- ianus). The filaments are of equal width throughout у. hemsleyanus, S in S. Miele. S S. obassia, S. rugosus, S. shiraianus, and S. мена and dis- tally attenuate in the rest. Filament pubescence varies from absent (e.g., S. obassia) to proximally pubescent (e.g., S. hemsleyanus) or densely pubes- cent throughout (e.g., S. buchananii, S. curvirostra- tus, S. subpaniculatus). The amount of filament pubescence is variable in S. hookeri and S. tonkinensis. Anthers are wider than the distal portion of the filament except in S. cur- subpaniculatus, and $. odoratissimus, S. * obassia, S. virostratus, S. tonkinensis, where they are more or less the same width as the adjacent filament apex. The connec- tives are glabrous to stellate-pubescent. The length of the anthers, ranging from 2 to 7 (to 10) mm, is useful for species identification. The ovary is always apically pubescent; it ap- pears to hold little taxonomic value in the group under revision. The style is filiform and varies from glabrous (e.g., Styrax hemsleyanus) to densely hir- sute (e.g, S. buchananii, S. curvirostratus). The amount of style pubescence varies substantially within some species, thus limiting its usefulness in species identification. The number of ovules per carpel in the group under study is often difficult to ascertain due to the small size of the placental re- gion. The few samples that we have examined in- dicate that the number is variable, but is rarely less than 5 or more than 8 per carpel. FRUIT The fruit is usually globose, ovoid, or ellipsoid. Sty- rax curvirostratus has a cylindrical fruit, and that of S. macrocarpus is occasionally pyriform. The apex may be rounded (e.g.. S. subpaniculatus), apiculate through persistence of the style base (e.g., S. japoni- cus), or rostrate (S. curvirostratus and S. odoratissi- mus). Fruit size ranges from 0.5 to 3 ст long and from 0.4 to 2.5 ст wide. The outer surface of the pericarp is white-gray to gray-yellow stellate-tomen- tose or -pubescent, except in S. chrysocarpus, in which it is golden yellow stellate-tomentose. The inner sur- face of the pericarp is typically glabrous or sparsely pubescent; only in S. chrysocarpus and S. macrocarpus is the pericarp densely pubescent inside. In both of these species the fruit is apparently indehiscent, al- though the limited material available for study leaves the constancy of this character in doubt. The fruit is unquestionably indehiscent in S. porterianus and S. subpaniculatus; the fruit is dehiscent by two or three valves in the other species. Styrax porterianus and S. subpaniculatus are similar in fruit dehiscence to the North American species S. glabrescens and S. gran- difolius, which nearly always possess an indehiscent fruit. Styrax porterianus is the only species in our treatment with a fleshy pericarp (ca. 2 mm thick). The pericarp of all other species is dry. The thickness of the pericarp is variable within the dry-fruited species, but 5. macrocarpus always has a pericarp greater than 1 mm thick, distinguishing this species from all other dry-fruited species in our revision. SEEDS The seeds of the imbricate group of series Cyrta are globose, ovoid, or ellipsoid, from beige to brown and smooth to finely reticulate-fissured or irregu- larly rugose (i.e., with a wrinkled appearance). Seed coat pubescence occurs in Styrax curvirostratus, S. Volume 90, Number 4 2003 Huang et a 497 Revision Styrax Series Cyrta macrocarpus, S. odoratissimus, and S. tonkinensis. his pubescence is absent, however, in some in- dividuals of each of these species. The seed coat is usually tuberculate in S. tonkinensis. These tuber- cles are sometimes arranged in a stellate pattern, in which case they often resemble stalked stellate trichomes. CHROMOSOME NUMBERS Only three species in our revision have been counted: Styrax hookeri (n = 8, Arora, 1961; Mehra & Bawa, 1969; Mehra, 1976), S. japonicus (n > 20, Manshard, 1936; n = 8, Yamazaki, 1993), and 5. obassia (n = 8, Manshard, 1936; Yamazaki, 1993). Chromosome numbers of two North American spe- cies of series Суна have also been reported (S. americanus, n ; S. grandifolius, п = l6: Gon- soulin, 1974). fion; these numbers and reports for species in the other series, the base number of Sty- rax is inferred to be x = 8 (Fritsch, 2001). Poly- ploidy is thus far known with relative certainty only in series Cyrta (S. grandifolius; the old number for S. japonicus of n > 20 must be oe" in light of the more recent number of n = ECOLOGY AND ECONOMIC IMPORTANCE According to herbarium specimen labels, species of the imbricate group of series Cyrta are found most often from 500 to 2700 m elevation in Asia. Several species (Styrax japonicus, S. obassia, S. odoratissimus, S. porterianus, S. subpaniculatus, 5. supaii, and S. tonkinensis) occur additionally or ex- clusively at elevations less than 500 m; only 5. hookeri extends to 3000 m or higher. Some has ies (S. hemsleyanus, S. hookeri, S. japonicus, 5. odor- atissimus, and S. tonkinensis) have a wide elevation range (2000 m or more in extent). Many of these species are found in a variety of habitats, such as open woodlands, pastures, mountain slopes, road- sides, high-elevation forests, and successional ar- eas. Many species show a distinct preference for mesic microhabitats, such as canyons, draws, and other riparian situations. Styrax species are most frequently pollinated by bumble bees and honey bees (Gonsoulin, 1974; Sugden, 1986; Kato & Hiura, nators reported for Styrax species are papilionoid wasps, and other groups of bees (e.g., carpenter bees, hal- ictids, anthophorids; Copeland, 1938; Gonsoulin, 1974; Sugden, 1986; Saraiva et al., 1988; Tamura & Hiura, 1998). Both nectar and pollen serve as ther polli- butterflies, syrphid flies, sphingid moths, floral rewards for pollinators, although there are no specialized structures recognizable as nectaries. The stellate trichomes present on the exterior sur- ce of the corolla in most species of Styracaceae have been suggested as an adaptation for support- ing large pollinators, which use them as “toe holds” to gather nectar and pollen (Sugden, 1986). The flowers yrax are sweetly fragrant (Perkins, 1907; Copeland, 1938; Fritsch, pers. obs.). Nearly all species have exclusively hermaphro- ditic flowers. Partial self-incompatibility has been suggested for Styrax obassia (Tamura & Hiura, 1998), the only member of series Cyrta examined for breeding system. Obligate xenogamy is docu- mented for several species of Styrax from other se- 1988). Morpho- logical gynodioecy is reported for ten gere in series Valvatae (Wallnéfer, 1997; Fritsch, 1999, press b), but experiments to confirm functional gyn- ries (Sugden, 1986; Saraiva et al., odioecy in these species have not been conducted. Little data exist on the dispersal mechanisms of Styrax. Fruits of S. obassia are dispersed by ground mice and food-hoarding birds (Kato & Нига, 1999). After the fruit wall has become detached, the seeds of the riparian species S. faberi Perkins, a valvate-flowered member of series Cyrta, remain attached to the receptacle by the hilum. The seeds, which would otherwise sink, can thus be trans- ported in water by the floating infructescence (P. Fritsch, pers. obs.). The seeds of S. japonicus, an imbricate-flowered species of series Cyrta that ex- libits the same type of seed attachment, may also be dispersed in this way. The seeds of S. ameri- canus reportedly have been found attached to the feet of waterfowl (Ridley, 1930), but this is probably not a primary means of dispersal of Styrax species because the surface of the seeds is generally smooth and curved, and therefore not obviously adapted for attachment. e benzofuran egonol and its glycosides occur in the seed oil of several species of Styrax. The fruit of Styrax contains significant amounts of je- gosaponin, a potent defense chemical. Various spe- cies of Styrax also contain styracitol, 8-phenyl eth- yl alcohol, and coniferin (Hegnauer, 1962; Gibbs, 1974) In many species of Styrax, a balsamic resin (ben- zoin, gum benjamin) exudes from the bark and wood tissues following injury to the cambium. This resin consists chiefly of coniferyl cinnamate, cin- namyl cinnamate (styracin), and coniferyl benzoate associated with cinnamic and benzoic acids; minor components are fragrant benzaldehyde, vanillin, and styrene (Hegnauer, 1962; Langenheim, 2003) It is used medicinally as an antiseptic and expec- torant, and in the flavor and fragrance industries (Pratt & Youngken, 1951; Duke, 1985; Langen- 498 Annals of the Missouri Botanical Garden heim, 2003). The best known source of benzoin is S. benzoin Dryand., a species of series Benzoin. Within series Сула, three species have been re- ported as sources of benzoin (Burkill, 1966): S. ser- rulatus, S. subpaniculatus, and S. tonkinensis, the latter two of which are included in our revision. The benzoin from S. tonkinensis is called zoin" because of its source in “the western parts of (Burkill, Indochina and eastern parts of Siam” 1966: 2146). We have not seen any specimens of S. tonkinensis from Thailand to confirm its occur- rence there. The oil extracted from the seeds of some species in our revision can be used to make soap or lubri- cating oil (e.g., Styrax hemsleyanus, S. japonicus, S. obassia, and S. odoratissimus; Tai Pan, 1981; Hwang, 1987), or medicinally as an antiseptic to treat scabies (S. tonkinensis; Hwang, 1987). The young leaf of S. japonicus is used as tea in certain regions of China (K. M. Feng 11082, Yunnan), and the fruit of this species can be used as a source of sugar extract to brew wines (P. C. Tam 63659, Hu- nan). The flowers, leaves, fruits, and galls of some species are used as Chinese herbal medicines (e.g.. S. hemsleyanus and S. japonicus; Tai & Pan, 1981). Several Styrax species of series Cyrta native to Asia are cultivated for ornament (Raulston, 1992). Sty- rax japonicus and S. obassia are most commonly cultivated, but also occasionally planted are S. hemsleyanus, S. limprichtii, S. odoratissimus, S. shi- raianus, S. tonkinensis, and S. wilsonii. Many cul- tivars of 5. japonicus have been developed (Rauls- ton, 1992). Most species of Styrax series Cyrta serve as the primary host for aphids of the family Hormaphidi- dae (tribe Cerataphidini). These aphids produce conspicuous galls of various shapes on the vege- tative and reproductive shoots of Styrax. Most in- dividual cerataphidine aphid species use a single species of Styrax as primary host, although it is common for several species of aphid to parasitize 1997). Often the shapes of the galls produced by aphid species are the same species (Stern et al., characteristic of particular species of Styrax, e.g., spiral galls on S. paralleloneurus Perkins and cor- alline galls on S. subpaniculatus. The aphids pro- duce a sterile soldier caste that defends the rest of the colony from predators. The morphology of these aphids and their galls, aphid behavior, and soldier production have been studied extensively ; Docters van Leeuwen, 1922; Aoki, 1982; Kurosu ‚ 1990, 1991, 1997; Aoki & Kurosu, 1993; Aoki et al., 1998; Kurosu et al., 1998), and the evolution of soldier production has been investi- pu > (у e.g. gated in ecological and phylogenetic contexts "Siam ben- (Stern, 1994, 1998; Stern & Foster, 1996). Evi- dence for co-evolutionary patterns of host-switching analyses of both the 1995) and Styrax (Fritsch, 1999), The four North American species of series Cyrta comes from phylogenetic aphids (Stern, and S. shiraianus from Japan are apparently not D by these aphids (P. Fritsch, pers. obs.; S. Aoki, pers. comm.); neither are any species of series Styrax or series Valvatae (P. Fritsch, pers. obs.). Thus, this interaction is apparently restricted to eastern and southeastern Asia and associated is- lands. MATERIALS AND METHODS Nearly 5000 herbarium specimens from 22 her- baria (A, AAU, BM, BO, BR, С, CAS, DS, E, GH, IBK, IBSC, K, KUN, KYO, L, MO, P. PE, TAI, TI, and UC) were examined for this study. All descrip- tions were derived from examination of herbarium specimens. Flowering and fruiting times, elevation ranges, habitats, distributions, common names, and uses were derived from label information. Descrip- - ions of leaves generally refer to those of the fertile branches; leaves of sterile branches are consistent- ly larger and often possess more variation in tri- chome quantity and quality than those of fertile branches and thus are less useful for species iden- tification. Leaf and petiole measurements were tak- en from the larger examples on each herbarium sheet. At the proximal ends of the twigs many de- ciduous species of Styrax have small leaves of roughly equivalent size among species, and the in- corporation of these into descriptions would make species identification more difficult. Flowers are described at the stage of anthesis except where not- ed. Calyx dimensions are presented as height (from the end of the pedicel to the distal margin) times width at the apex, and thus include the short hy- panthium. Fruit length was measured from the base of the fruiting calyx to the tip of the fruit (the calyx is persistent). Fruiting measurements were taken rom mature fruits where possible. Often immature fruits are the only types available for examination which on a herbarium sheet, in case the larger fruits on the sheet were measured. Most observa- tions were made by eye or with the aid of a dis- secling microscope (maximum magnification = 04x Because our study is based primarily on herbar- ium specimens, we employ the morphological spe- cles concept, as discussed in Stuessy (1990), for species recognition. We base our species on the exislence of correlated gaps in states among mor- phological characters, and treat clinal patterns as Volume 90, Number 4 Huang et al. Revision of Styrax Series Cyrta infraspecific variation that requires no formal tax- onomic recognition. We explain our decisions on circumscription under each species, often in the context of the relevant taxonomic work of previous authors. We assume that the morphological differ- ences among the species we recognize have a ge- netic basis, as can be inferred from examination of several species in a common garden setting (e.g., Styrax japonicus, S. obassia, and S. odoratissimus at the University a California Botanical Garden, Berkeley, California, U.S.A.), and regard the spe- cies we have recognized as hypotheses to be tested as new morphological data become available. Ap- pendix 3 provides an alphabetic listing of species in the Taxonomic Treatment, including synonyms and excluded names. The dots in the distribution maps are based on the specimens cited in this revision (see Taxonomic Treatment and Appendices 1 and 2). For collections in which geographic coordinates were not indicated on specimen labels (most collections), we estimated coordinates based on descriptive label information about the location of the collection. Our estimate was aided with a variety of published maps, atlases, and gazetteers, particularly the (United States) Na- tional Imagery and Mapping Agency (NIMA) da- tabase of foreign geographic feature names, with access provided by GEOnet Names Serve at = sometimes the tubercles ar- my rmations oncave margin; filaments of S. tonkinensis lamina visible through the р е if present inita nearly tomentum in 5. sees ag "mn calyx f filam runcate, undulate, or irregularly ea y the lobed, the teeth not contiguous if present; nts Nie i distally, flexuous at middle (occasionally straight in 5. pons ve smooth, glabrous, appressed-stellate-pu- escent, or lepidote (seeds unknown in S. bucha nies Coroll: times as ЭН аѕ wide; pseudoterminal inflorescences usually =e > 5 Ф (л = - | Ni ng M niculate; fruit apex rostrate, iin merely apiculate: seeds usually *10 ressed-stellate-pubescent or lepidote ‚ S. odoratissimus paniculate; fr buchananii); seeds glabro 9 Conn mm long; flowers mm long |... imes as long as wide; pseudoterminal inflorescences usually ng as uit apex — or subacute, rarely also apiculate (fruit unknown in 5. es (at bou pm and style densely stellate- min anthers 5—7 of 3-1.6 ст long; calyx stellate-hirsute, arm trichomes av- ‚5. buchananii eraging ca. 9b. Connectives and style glabrous: anthers 3—4 mm long; flowers 0. 9-1 E cm long: 2m 14. 5 calyx tomentose, arms of trichomes < 0.2 mm long -.............. 2b. Pseudoterminal inflorescences 1 10a. Petiole of larger leaves dilated at base, covering the bud: inflorescences distally congested; p m long; corolla tube 10-12 mm long ....... s S. коз ш long. = 7-flowered (3- to 11-flowered in S. shiraianus). el shiraianus 7 cm wide; fruit indehiscent (rarely . with corolla 10-28 mm long; North America. dehiscent by 3 v 13a. Tree to Е т, 11-23 X 6-10 mm: fruit 10-17 es no 13b. Tree to 6 m, often suc NE. extensively from roots; leave iu races corol 1 lob es 8— 6 X 3-7 mm: fruit 8—12 t suckering Dus roots; leaves membranaceous; corolla lobes —]9 mm; Mexico and M per renum СЦЕ И: glabrescens Benth. la X 6—8 mm: азнав oM States ------- —— S. randi Ait. 12b. Distalmost leaves usually T cit wide nom > 7 cm wide in S. jaliscanus, odoratissimus, 5 subpanic ulatus, en tonkinensis); (mit Какен or if indehis- ent, then corolla ГЕ: mm long; Аѕ 14a. Calyx truncate, deren MS lobed or toothed, if toothed then the teeth not contiguou 1 mm of inge calyx aba sially Boos or if stellate trichomes present, within margin more sparsely mbescen than the rest of the calyx or subglabrous to glabrous, somewhat scarious, brown when Longer pedicels on eac = twig 15—50 mm long usually ‘equal to or longer 15 15 a. b. than subtended flower |... sss .J Longer pedic a on each twig 2—10(-13) mm long, usually shorter than aponicus subtended flow All flower TS a او‎ from shoots of the previous growing sea- son; petioles < 1(-2.5) mm long: pericarp dry, (1-)1.5—3 mm thick; inner surface of pericarp PES y appressed- Leh > —— 55 [ба 16b. 5 іе ast some > flowers paired or in racemes arising from ы of the urrent growing season; petioles > 2.5 mm long; peri thic k or s inner surface of pericarp glabrous or н airs pubes- 1 — =] 1 m a. : Stems of young f« fertile shoots gener rally > . 5. macrocarpus СРИ of young fertile shoots generally < 0.6 mm wide at the narrowest points proximally; pedicels slender, 0.2-0.6 mm wide proximally; calyx toothed, the teeth linear-subulate at least at apex but often wider proximally, 0.5-1.2 mm long; corolla lobes 1-5 mm wide, apex acute; North America (eastern United 3 . americanus Lam. mm wide proximally (often narrower distally); pedicels stouter, (0.4—)0.6-1 mm wide Annals of the Missouri Botanical Garden 14b. € proximally; calyx truncate, undulate, irregularly lobed, toothed, if toothed the teeth deltoid to linear-deltoid: morals lobes 3-13 mm wide, a x obtuse or acute-acuminate; Asia. 18a. Flowers (1.3—)1.5-2.5 ст long: vus lobes (1112-18 mm iis ide (3. Aes X 7(-1 n nm; m e s eredi Шш. brin often abaxially pecially dani sc cattered among the base to- en rs 3-5 mm done. wider than distal portion of filament; fruit subglobose or ovoid, (1.0-)1.5-2 ст long: apex acute, аша ee 93 S. hookeri 10b. Tectiary ani quaternary veins of fanina conspicu- ously raised adaxially (as well as abaxially), the ter- tiaries irregularly reticulate; calyx abaxially without scattered stiff stellate trichomes; filaments 4—5 mm long, of equal width throughout; anthers 5-6 mm long, as wide as or narrower than distal pirer of filament; fruit cylindrical to obliquely ovoid, 2-2. cm a apex usually rostrate, rostrum up to 2 ет ong. S. curvirostratus 18b. hi тз < с 5cm n long; corolla lobes 9-11 mm dene: calyx 34-5 3—4 mm; filaments 1.54 mm long: pericarp not longitudinally striate. 2( nnectives (at least pronis) s style densely кн сайа pericarp dry, 0.5-1 mm thic k, smooth or slightly rugose; s seeds usually appressed- stellate-pubescent or lepidote, rarely glabrous; ma- ture leaves light green to yellow-gre en when dry, chartaceous or thick-chartaceous 5. iom EE 20b. С onnectives and style glabrous: pe ric arp (e 2 mm thick, deeply rugose when dried; seeds d brous; mature leaves green to d green when dry, membranaceous or thin-chartaceous к 11. S. porterianus low E: 5 ; Calyx distinctly dentate, the t tee d usually poa or separated i» a rub concave portion; calyx abaxially within 1 mm of d margin evenly pubescent, ч color and texture + similar to the rest of the calyx. 21b. Trees to 30 m tall; petiole 8-12(-15) mm e pericarp not oe striate, apex rostrate; seec ds densely tuberculate, sometimes екы ercles i s loni inensis Shrubs to 2.5 m tall (сойлей s a tree to 6 m in S. ne eu ole mm long: pericarp longitudinally striate, ane x rounded or apiculate; ded. smooth or finely retict Ja оше glabre 22a. Lamina 1—2.5(-4) X 0.7-2(-2.5) cm; "fruit 0.4—0.6 cm ic^ S. wilsonii 22b. Lamina : x 2-8 cm; fruit = 0. Ti 'm wide. 23a. Sec ie veins of lamina 7—10 on each side of midvein: inflo- rescence rachis gray-green tomentose; calyx gray-green lanate throughout; North America (western Мехіс‹ 9, تن = а. s. Jaliscanus i$, Wats. 23b. Secondary veins of TM 4—7 on each sire af сои мп; Н rescence rachis yellow or orange tomentose; calyx yellow, yel- low-brown, or orange tomentose, often also mile various amounts of larger scattered dark yellow, orange, or brown stiff stellate trichomes, especially proximally; Asia. 24a. Quaternary as well as the tertiary veins of lamina abaxi- ally prominent and raised in young leaves; rachis with stalked trichomes; fruit 0.8-0.9 ст wide —__ 12. S. rugosus 24b. Only the tertiary veins of lamina abaxially prominent and raised in young leaves; rachis without E tricho omes; fruit 1.0-1.5 em wide 0 . S. limprichtü D Volume 90, Number 4 2003 Huang et al. 503 Revision of Styrax Series Cyrta 1. Styrax buchananii W. W. Sm., Notes Roy. Bot. Gard. Edinburgh 12: 234. 1920 [as S. *Buch- ananW |. ' 3 Kachin State: Myitkyina in Mara Nantan forest, Kaukkwe Valley, 606 m, Mar. 1912, E. M. Buchanan 51 (holotype, E!; isotype, E!). Figure 1. yanmar. Styrax serrulatus var. latifolius Perkins, in Engl., Pflan- zenr. IV. 241 (Heft 30): 37. 1907. TYPE: Myanmar. Mandalay my W. Griffith 3670 (lectotype, des- i K [loan accession no. H2000/01016- 29]!; isotypes, GH!, K [loan accession no. H2000/ 01016-301). Small trees. Young twigs densely yellow-brown stellate-pubescent; older twigs becoming gray, subglabrous. Petiole 3—4 mm long. Two most prox- imal leaves on each shoot alternate. Lamina of fer- tile shoots 6-11 X 4—6 cm, those of sterile shoots to 16 X ll cm, chartaceous, ovate-oblong; apex slightly acuminate to obtuse, base rounded or broadly cuneate, rarely truncate; adaxially sparsely yellow-gray pubescent with 2- or 3-armed or stel- late trichomes; abaxially sparsely to densely yel- low-gray the through the pubescence, the pubescence especially prevalent on veins; margin remotely irregularly ser- stellate-hirsute, surface visible rulate apically; secondary veins 5 or 6 on each side of midvein, adaxially faintly prominent, abaxially prominent; tertiary veins parallel and perpendicu- lar to the secondaries, plane or slightly sunken adaxially, abaxially prominent. Fertile shoots 19— 30 cm long, 3- or 4-leaved. Inflorescences arising from shoots of the current growing season; lateral inflorescences racemose, 2—5 cm long, 3- to 5-flow- ered, often with 1 or 2(to 4) flowers occurring in the same leaf axil; pseudoterminal inflorescences usually paniculate, sometimes racemose, 9-13 ст long, 10- to 22-flowered, lateral branches 2 to 5, sometimes with 2 or 3 short lateral racemes from the base of inflorescence, rachis and branches densely yellow stellate-pubescent. Pedicel (1—)3 mm long, densely yellow stellate-hirsute; bracteoles л ca. 3 mm long, linear, usually positioned at the base of pedicels. Flowers 1.3-1.6 cm long. Calyx 4—5.5 X 4—5 mm, cupuliform; adaxially white ap- pressed-pubescent with 2- or 3-armed or stellate trichomes; abaxially densely yellow stellate-hirsute, arms of trichomes averaging ca. 1 mm long, within | mm from the margin more sparsely pubescent, somewhat scarious, brown when dry; margin trun- cate, undulate, or irregularly lobed, the teeth mi- nute and not contiguous if present. Corolla 0.8-1.1 cm long, white, tube ca. 3 mm long, glabrous, lobes 5, 9-13 X 34.5 mm, 2.4-3.0 X as long as wide, lanceolate or stellate-tomentose on both sides. Stamens 10; filaments 3—4 mm long, ovate-lanceolate, slightly flexuous at middle, distally attenuate, densely white to yellow stellate-hirsute throughout, arms of the trichomes predominantly pointing up- ward; anthers 5—7 mm long, wider than distal por- tion of filament; connectives (at least proximally) densely stellate-hirsute. Style densely white stel- late-hirsute nearly throughout, thinning distally; stigma 0.4—0.7 mm wide, punctiform. Fruit un- known. Illustrations. None previously published. Phenology. Flowering: February-April. Fruit- Distribution. Myanmar (Kachin State, Manda- lay Division, and Sagaing Division); Figure 2. abitat. In valley forests; 600-1500 m Styrax buchananii is geographically isolated in Myanmar from all other Styrax species except | hookeri and S. japonicus, from which it can ай be distinguished by its longer inflorescences and more numerous flowers. It has been only rarely col- lected throughout its range and is only known from flowering collections. This species was first. described by Perkins (1907) as a suggested that it likely represented a new species, but the available material at the time of description (Griffith 3670 from the Ruby Mines District (Smith, 1920) of Mandalay Division, Myanmar) was inad- equate for a proper assessment of species status. Similarly, Smith (1920), in describing S. buchan- was uncertain variety of Styrax serrulatus. Perkins anii based solely on Buchanan 51, whether S. serrulatus var. latifolius Perkins should be listed as a synonym. With the benefit of addi- tional material available to us, we affirm that Grif- fith 3670 and Buchanan 51 represent one and the same species, based on the combination of imbri- cate corolla aestivation, pubescent style, many- flowered inflorescences, and other key characters present in these collections. Furthermore, the gen- eral locality of Griffith 3670 lies in the vicinity of all other collections of S. buchananii. Smith (1920) described S. buchananii with valvate corolla aes- tivation, but our observations confirm that all spec- imens cited in the protologue of S. buchananii have distinctly imbricate aestivation. Several shared characteristics, e.g., the two most proximal leaves on each shoot of the current year subopposite to opposite and stellate pubescence covering nearly the whole length of the filaments and styles, suggest that Styrax buchananii is a close relative of S. odoratissimus. Styrax buchananii can be distinguished from S. odoratissimus by the short- er petioles (3—4 vs. 5-12 mm), ulate (vs. typically racemose) inflorescences, corol- the typically panic- 504 Annals of the Missouri Botanical Garden Figure 1. Styrax buchananii. —A. Flowering branch. —B. Leaf surface, adaxial view. —C. Leaf surface, abaxial view. —D. Stellate trichome from the abaxial side of the leaf. —E. Flower. —F. Calyx + gynoecium, median long- section, —G. Part of corolla + androecium, opened. —H. Stamen, lateral view. Based on Kingdon-Ward 20550. Volume 90, Number 4 Huang et al. 505 2003 Revision of Styrax Series Cyrta hio 130 * Styrax buchananii A " Styrax chrysocarpus PN Ps a Styrax macrocarpus 2 * Styrax wilsonii p ^ 800 L ped Kilometers — . y madens a jee SA o ^ Ps d а = = Є m è £m P / C Р | k 2d ~. ü “ x i: yom [| | <, 2 м uf 5 J 3 2 b „>т ( r N км чүл Miia, Ж EE ыб. е кый A ur E iit à m : VN P wA h j Ом. (or г. Y t І е = Ü а е d ACE ui. a! € “Sa " ee ^ е " Figure 2. Geographic distribution of Styrax buchananii, S. chrysocarpus, S. macrocarpus, and S. wilsonii. la lobes 2.4—3 (vs. 1.7-2.2) times as long as wide, and a distribution (northern Myanmar) that is out- side the known range of S. odoratissimus (China). Another probable close relative of Styrax buch- ananii is S. chrysocarpus, a species whose range in unnan Province is located between those of S. buchananii and S. odoratissimus. Styrax chrysocar- pus has a leaf texture and average petiole length (5-8 mm) similar to the other two species, and in all three the two most proximal leaves are alternate. The differences between S. buchananii and S. chry- socarpus are not entirely clear in the absence of data from flowers (S. chrysocarpus) and fruits (S. buchananii). The protologue of Styrax serrulatus var. latifolius cites both B and K specimens of Griffith 3670. The B specimen has presumably been destroyed, and thus we have chosen one of the two K specimens that we have seen as a lectotype. Neither sheet har- bors Perkins's annotation, but that with loan acces- sion number H2000/01016—29 has better flowering material. Thus, we have chosen this sheet as the lectotype. Additional specimens examined. MYANMAR. Kachin State: Myitkyina at Lamaing, E. M. Buchanan 21 (E); Japing Valley, G. Forrest 21083 (E); Myitkyina Dist., Sum- prabum Subdivision, Hlingnan, Y. Hla & C. Koko 3746 ci . road to иара: Ј. Н. Гасе 5737 . J. Н. Lace 5774 (Е, К); Sumpra Bum, F. F. K. Ward 20550 (A, BM). Sagaing Division: Patkoi Range, border betw. Burma [Myanmar] & India, R. 5. Hole 17 (K). 2. Styrax chrysocarpus H. L. Li, J. Arnold Ar- bor. 25: 312. 1944. TYPE: China. Yunnan: Pingbian Miaozu Zizhixian, 1400 m, 9 July 1934, H. T. Tsai 62505 (holotype, A!; isotypes, IBSC!, KUN!, РЕ!). Trees 7-20 m tall. Young twigs yellow-brown stellate-tomentose; older twigs dark brown, subgl- abrous. Petiole 5-8 mm long. Two most proximal leaves on each shoot alternate. Lamina 10-20 X cm, chartaceous, oblong-ovate to oblong; apex acute to slightly acuminate; base rounded or roadly cuneate; adaxially sparsely yellow-gray stellate-pubescent, arms of the trichomes up to 0.2— 0.3 mm long, the pubescence especially prevalent on veins; abaxially densely yellow-gray stellate-hir- sute, arms of the trichomes up to 0.5-0.6 mm long, 506 Annals of the Missouri Botanical Garden the surface remaining visible through the pubes- cence; margin subentire or remotely irregular ser- rulate apically; secondary veins 5 to 10 on each side of midvein, adaxially plane or slightly sunken, abaxially prominent; tertiary veins reticulate, abax- ially prominent. Flowers unknown. Infructescences arising from shoots of the current growing season, apparently racemose, 1- to 5-fruited, yellow stel- late-tomentose. Fruiting pedicel 4—5 mm long. Fruiting calyx 5—6 X 10—15 mm, cupuliform, red- brown, the margin not appressed to the fruit, gla- brous adaxially, densely stellate-pubescent abaxi- ally; margin irregularly 5- or 6-crenately lobed, lobes ca. 4 X 10 mm. Fruit 1.6-1.8 X 1.0-1.2 cm, ovoid, apex shortly pointed, apparently indehiscent; pericarp dry, 0.3—0.5 mm thick, outside golden yel- low stellate-tomentose, inside densely pale yellow appressed-pubescent. Seeds dull dark-brown, ovoid, smooth, glabrous. Illustrations. C. Y. Wu, Fl. Yunnan, 3: 430, pl. 123 (1—3). 1983. Phenology. Flowering: unknown. Fruiting: July. Distribution. China (Yunnan); Figure 2. In ravine forests; 1400-1500 m. Huang-guo-an-xi-xiang Habitat. Vernacular (Hwang & Qi, 1983). Styrax chrysocarpus is known with certainty only names. 1985), Mao-guo-an-xi-xiang (Wu, from Pingbian Miaozu Zizhixian, southeastern Yun- nan Province. This species is easily distinguished from other members of Styrax by its golden yellow fruit and densely pale yellow pubescent inner sur- face of the pericarp. A sterile specimen with aphid galls collected between 1550 and 1650 m elevation in Yongshan Xian, extreme northeastern Yunnan Province (H. T. Tsai 51156), might be this species. Its leaves, however, are glabrous, unlike the dense- ly hirsute upper and lower surfaces of those in the type. More fertile material from the vicinity of Tsai's localities is highly desirable to better understand the taxonomy of this species. Although only fruits are available for compari- son, careful analysis of vegetative and fruit mor- phology suggests that Styrax chrysocarpus is most likely allied to other deciduous species with im- bricate aestivation. Sterile specimens of 5. chryso- carpus are similar to some specimens of S. buch- ananii and S. odoratissimus in the relatively large leaves, the lower laminar surface somewhat rough to the touch, and the strictly alternate leaves. Sty- rax chrysocarpus consistently differs from S. odor- alissimus, however, in its shorter infructescences and larger yellow stellate-hirsute fruit, and differs from most specimens of S. odoratissimus in its gla- brous seeds (differences between S. chrysocarpus and S. buchananii are addressed in the discussion under 5. buchananii). Furthermore, none of the dis- tributional ranges of these three species overlap: 5. buchananii is restricted to Myanmar, S. odoratissi- mus to southeastern China, and S. chrysocarpus to eastern Yunnan Province, China. Amed. CHINA. — T. Tsai 62522 (A, KUN, PE); Yongshan Xian, Н. T Additional n imens 6 (A. IBSC, )). Tsai 511: E A, BC еш, 3. Styrax curvirostratus (B. Svengsuksa) Y. L. Huang & P. W. Fritsch, Styrax agrestis var. curvirostratus B. Svengsuk- stat. nov. Basionym: sa, Flore du Cambodge du Laos et du Viétnam 26: 176. 1992. TYPE: Vietnam. Lam Dong: Massif du Lang Bian, between Dankia and Dangle, 1000-1200 m, 25 Oct. 1930, Е. Po- апе 18626 (holotype, P not seen; isotype, P). Figure 3. Trees to 15(-20) m tall. Young twigs dark gray brown, sparsely gray-white. stellate-pubescent; older twigs dark brown or nigrescent, glabrescent. Two most proximal leaves amina 6-11 X 34.5 em, thick-chartaceous, elliptic to oblong; apex short- Petiole 7-10 mm long. on each shoot alternate. acuminate to acuminate; base rounded to broadly cuneate; glabrous, rarely abaxially sparsely short- stellate-pubescent on the veins and vein axils, both surfaces glossy, bright green when dry; margin en- tire or slightly undulate, rarely irregularly dentic- ulate; secondary veins 5 or 6 on each side of mid- tertiary reticulate and conspicuously raised on both sides. vein; and quaternary veins irregularly Fertile shoots 8-15 em long, 3- to 5-leaved. Inflo- rescences arising from shoots of the current growing season; lateral inflorescences 1(2)-flowered; pseu- doterminal inflorescences l- or 2-flowered or ra- cemose, 1—3 cm long, (1- to)3- to 5-flowered, rachis yellow stellate-tomentose. Pedicel 8-9 mm long, yellow stellate-tomentose; bracteoles 1-2 mm long, linear, positioned at various places along the ped- icel but mostly near the base. Flowers 1.6-1.8 em long. Calyx 6-7 X 6.5-7 mm, cupuliform; adaxially densely white appressed-stellate-pubescent, proxi- mally becoming sparsely pubescent with white 2- or 3-armed trichomes; abaxially densely yellow stellate-pubescent, within 1 mm from the margin more sparsely pubescent or glabrous, somewhat scarious, brown when dry; margin truncate, undu- late, or irregularly lobed, the teeth minute, not con- tiguous if present. Corolla 0.9-1.2 em long, white, tube ca. 4 mm long, slightly pubescent, glabrous Volume 90, Number 4 Huang et al. 507 2003 Revision of Styrax Series Cyrta Figure 3. ADEM curvirostratus. —A. dudit branch. —B. Leaf surface, adaxial view. —C. Leaf surface, n view. —D. Flower. —E. Stamen, lateral view. —F, G. Fruit. —H. Seed. A-E based on don et al. VH 4 based on Poilane "18626; G, H based on Chenier 38674. proximally, lobes 5, 12-13 X 5-6 mm, obovate to equal width throughout, densely white stellate-vil- obovate-elliptic, apex acute, densely pale pony lous throughout, arms pointing upward; anthers 5— stellate-hirsute on both sides. Stamens 10; fila- 6 mm long, as wide as or narrower than distal por- ments 4—5 mm long, strongly flexuous at middle, of tion of filament; connectives glabrous. Style densely 508 Annals of the Missouri Botanical Garden 115 ш Styrax curvirostratus e Styrax porterianus А Styrax subpaniculatus 0 400 n Kilometers Figure 4. white stellate-pubescent throughout, conspicuously 3-angular and 3-furrowed, stigma ca. 0.2 mm wide, capitate. Fruit 2.0-2.5 X 1.1-1.5 cm, cylindrical to oblique-ovoid, apex usually rostrate, rostrum up to 2 ст long, dehiscent; pericarp dry, 0.3-0.4 mm thick, outside irregularly longitudinally striate, gray stellate-tomentose, inside minutely downy-pubes- cent. Seeds brown, ellipsoid, smooth to finely retic- ulate-fissured, glabrous or occasionally appressed- stellate-pubescent. Illustrations. B. Svengsuksa & J. E. Vidal, Flore du Cambodge du Laos et du Viétnam 26: 173, pl. 31 (10—11). 1992 (as S. agrestis var. curvirostra- tus). Geographic distribution of Styrax curvirostratus, S. porterianus, and S. subpaniculatus. Phenology. Flowering: April, May. Fruiting: January, September, October. Distribution. Vietnam (Binh Thuan, Dac | Khanh Hoa, and Lam Dong); Figure 4. Habitat. In primary, closed, evergreen broad- leaved mountain forests; 1000—1700 m. Styrax curvirostratus is the only imbricate spe- ас, cies of Styrax documented in southern Vietnam; it is thus easily distinguishable from the several sym- patric members of the genus with valvate aestiva- tion. This species is distinguished from most other imbricate species by its long-rostrate, cylindrical to oblique-ovoid fruit 2-2.5 X 1.1-1.5 em. The other species in this group with at least some rostrate- fruited individuals are S. hookeri, S. odoratissimus, Volume 90, Number 4 Huang et al. 509 9 Revision of Styrax Series Cyrta and S. tonkinensis. These species possess smaller (less than 2 cm long) fruit with a shorter rostrum (typically less than 2 mm long) than S. curvirostra- tus. Features of S. curvirostratus shared with S. odoratissimus and S. buchananii are the (1) densely white stellate-villous filaments and style, (2) trun- cate, undulate, or irregularly lobed calyx with non- contiguous teeth if present, and (3) sparsely pubes- cent or subglabrous calyx within 1 mm of the margin, without larger stiff stellate trichomes. In addition, S. curvirostratus and S. buchananii have longer anthers (5—6 mm long) than the other im- bricate species of series Cyrta. Styrax curvirostratus occasionally possesses appressed-stellate-pubes- cent seeds, as in most individuals of S. odoratissi- mus. Styrax curvirostratus can be distinguished from both S. buchananii and S. odoratissimus by its larger calyx 7 mm), longer flowers (1.6-1.8 cm long), and longer, wider, straight (vs. flexuous) filaments of equal width throughout (vs. narrowing distally). Moreover, 5. curvirostratus 1s easily separable from S. buchananii by its shorter inflorescences (1—3 cm vs. 9-13 cm long) with few- er (1 to 5 vs. 10 to 22) flowers. Styrax curvirostratus can be recognized when sterile by the reticulate and distinctly raised quaternary veins on both sur- faces of the lamina. This species was first collected in Lam Dong in 1930 (Poilane 18626), but was left undescribed un- til Svengsuksa and Vidal (1992) assigned this spec- imen and several others to Styrax agrestis (Lour.) G. Don, a species with valvate corolla aestivation, as a new variety. The variety was based on fruiting specimens only, as no flowering material was avail- able. Styrax curvirostratus typically shares with S agrestis a rostrate fruit, which separates these two species from most others in Southeast Asia, and the ranges of the two taxa overlap, with that of 5. agres- tis the larger. It was thus not unreasonable for Svengsuksa and Vidal to place S. curvirostratus as a variety of S. agrestis, Eis insuishable i in fruit from the typical variety by its shorter petioles and ped- icels, usually glabrous seeds, and more conspicu- ous rostrum. Recently, however, a flowering speci- men (Averyanov et al. VH4544) was collected at a locality within the range of S. curvirostratus that matches the vegetative morphology of this taxon in every respect, yet has distinctly imbricate, rather than valvate, corolla aestivation and fewer flowers per inflorescence than S. agrestis (1 to 5 vs. 5 to 10). These features clearly distinguish S. curviros- tratus from S. agrestis. Furthermore, the conspicu- ously reticulate quaternaries on both surfaces, long anthers, and other features listed above distinguish this taxon from all other species of Styrax, thus warranting its recognition at the species level. Additional specimens examined. VIETNAM. Dac Lae: N de Ninh-Hoa, Massif de la Mère et l'Enfant, E. Poilane 6578 (P). Khanh Hoa: Phu Khanh, Massif du . B. Chevalier 38674 (P). Lam Dong: Lac Duong, Mun. Da Chay, 35 km NE from Dalat City, L аре et 4 VHA4544 (AAU, CAS); Massif du Haut Dankia & Dangle, F. ойле 23457 (Р), Ноп Ва, А. Donai, 23569 P. 4. boi hemsleyanus Diels, Bot. Jahrb. Syst. 530. 1 [as 5. “Hemsleyana”|. TYPE: China. Sichuan: Wushan Xian, 1885—1888, A. Henry 5676 (lectotype, designated here, A!; isotypes, BM!, GH!, IBSC[2]!). Styrax diu cur. var. griseus Rehder, in Sarg., Pl. Wil- : 291. 1912. TYPE: China. Hubei: Changyang Tu ujiazu Zizhixian, 1212-1818 m [1300-2000 m, gc June 1907, F. un Wilson 2574a (holo- А!; isotypes, BM!, К!, К Styras huanus Rehder, J. Arnold ie 11: 167. 1930 [as S. "Huanus"] TYPE: China. Sichuan: Nanc nE Shi, 2273-2576 m [1200-2700 m, protologue]. 3 June 1928, W. P. Fang 1376 (holotype, А!; isotypes, BM!, DS!, Е!, IBSC!, K!, PE[4]!). Trees to 12 m tall. Young twigs densely gray- brown stellate-pubescent; older twigs dark brown, glabrescent. Petioles 10—24 mm long, neither di- lated nor covering the bud. Two most proximal leaves on each shoot subopposite to opposite. Lam- ina 7—15 X 4—9 mm, chartaceous, elliptic or ovate- elliptic, rarely broadly elliptic, gray-green to dark green when dry; apex acute to short-acuminate; base oblique and subrounded to broadly cuneate, often shortly decurrent into petiole; adaxially sparsely gray-white pubescent with 2- or 3-armed or stellate trichomes; abaxially glabrous or sparsely to densely gray-white stellate-pubescent or -tomen- tose; margin subentire or serrate apically; second- ary veins 7 to 10 on each side of midvein, tertiary veins subparallel, abaxially prominent. Fertile shoots (12—)15-20 cm long, 2- to 4-leaved. Inflo- rescences arising from shoots of the current growing season, racemose; lateral racemes 4- to 9(to 13)- flowered, often with solitary flowers occurring in the same leaf axil; pseudoterminal racemes 1(10 3), 8— 15 em long, 8- to 15(to 20)-flowered, rachis yellow- brown stellate-tomentose. Pedice mm long, yellow-brown stellate-tomentose; mm long, subulate or linear, positioned at the base or middle part of pedicel, sometimes those toward the base of the inflorescence leaf-like. Flowers 1.5— 2.5 cm long. Calyx 4-8 X 3-6 mm, narrow-cupu- liform; adaxially densely appressed-pubescent, proximally becoming sparsely pubescent with white bracteoles 2—3 Annals of the Missouri Botanical Garden Tm | [ı10 [130 e Styrax hemsleyanus m д * Styrax rugosus А Styrax supaii 0 400 800 = — i —€— ARE Kilometers | - | | : = ae TUM J iv 45 Е. 45 X \ ^ & ( В ME: ref, \ y ۹ 1 f 7 کی‎ | E A m © © = 2 b ` ) . e. ы 18 | ‹ ce f 3 ғ > M of i ; . "NA $ \ e 0 T ° a? of à. 5 s, “С 4 30) JAN e? cx А c. v © ‘ate . v ” Ga: f ve \ \ ~ L | a | * € 4 LE J 8 . E * "m JEU , & à A Jo es e e + Сл í М \ g * $ N j му Figure 5. Geographic distribution of Styrax hemsleyanus, S. rugosus, and S. supaii. 2- or 3-armed trichomes; abaxially yellow-brown Illustrations. Prain, Bot. Mag. 136: t. 8339. stellate-tomentose throughout, often also with vari- ous amounts of larger dark brown stiff stellate tri- chomes especially proximally; margin with 5 un- evenly distributed teeth 2-3 mm long, unequal, subulate or deltoid, contiguous, densely pubescent on both sides. Corolla 1.1—1.7 em long, white, tube 4—5 mm long, glabrous, lobes 5 or 6, 12-15 x 4.5- 5 mm, elliptic to elliptic-obovate, apex acute, adax- ially subglabrous except distally, abaxially pale yel- low stellate-tomentose. Stamens 10 to 12; filaments 6-7 mm long, straight, relatively broad, of equal width throughout, ventrally + pubescent proximal- ly, glabrous distally; anthers 3.5—4.5 mm long, wid- er than distal portion of filament; connective subgl- abrous. Style glabrous; stigma 0.4—0.5 mm wide, capitate. Fruit 0.8—1.3(—1.6) X 1-1.5 em, globose to ovoid, apex apiculate, dehiscent; pericarp dry, 0.1-0.4 mm thick, outside slightly longitudinally rugose, yellow-brown to gray-yellow stellate-tomen- tose, inside sparsely appressed-stellate-pubescent or glabrous. Seeds brown, ovoid, nearly smooth, sometimes irregularly rugose or finely reticulate-fis- sured, glabrous. 1910; W. P. Fang, Ic. Pl. Omei. 1(1): t. 47. 1942: Anonymous, lc. Cormophyt. Sin. 3: 337, fig. 4628. 1974; F. T. Tai & T. C. Pan in W. P. Fang, Fl. Sichuan. 1: 418, fig. 161. 1981 (as S. huanus): ibid.: 426, pl. 165. 1981; C. Y. Wu, Fl. Yunnan. 3: 430. pl. 123 (4—7). 1983; S. M. Hwang & C. J. Qi in W. C. Cheng, Sylva Sin. 2: 1602, fig. 797. 1985 (as S. huanus); ibid.: 1619, fig. 812. 1985; S. M. Hwang, Fl. Reipubl. Popularis Sin. 60(2): 85, pl. 28 (6—7). 1987 (as S. huanus); ibid.: 96, pl. 32 (8— 14). 1987; W. Q. Yin in Y. C. Хи, Ic. Arbor. Yun- nan. 2: 896, pl. 472 (1-6). 1990; S. Y. Wang in B. Z. Ding, Fl. Henan 3: 230, fig. 1775 (5-8). 1997; Z. Y. |C. Y.] Wu & P. H. Raven, Fl. China Ill. 15: 197, fig. 197 (6—7). 2000 (as S. huanus); ibid.: 201, fig. 201 (8-14). 2000. Phenology. Flowering: March, May, June. Fruiting: February, May-September. Distribution. China (Gansu, Guizhou, Henan, Hubei, Hunan, Shaanxi, Shanxi, Sichuan, and Yun- nan); Figure 5 Habitat. In relatively mesic, semi-open mixed Volume 90, Number 4 2003 Huang et al. 511 Revision of Styrax Series Cyrta forests on mountain slopes and in ravines; 700— 2700 m. Vernacular names. Sichuan; Fang, 1942), Hui-mao-lao-gua-ling (Chi- na, Yunnan; Wu, 1983), Jin-shan-an-xi-xiang (Chi- na, Sichuan; Tai & Pan, 1981), Lao-gua-ling (Chi- na, Henan; Anonymous, 1974), Ma-lin-guang (China, Shaanxi; J. О. Xiang 6053), Mai-pao (Chi- na, Sichuan; J. H. Xiong et al. 91179), Mo-pao (China, Henan; Anonymous, 1974), Nan-chuan-an- xi-xiang (China, Sichuan; Tai & Pan, 1981). Styrax hemsleyanus is a relatively common spe- He-si-li-ye-mo-li (China, cies, occurring mainly in the mountains at middle elevations surrounding the Sichuan basin. It can be distinguished from sympatric imbricate-flowered species by the combination of the subopposite to opposite two most proximal leaves on each shoot, long, multi-flowered pseudoterminal racemes, and prominent calyx teeth. Styrax hemsleyanus is sim- ilar in these respects only to S. obassia, a species ranging farther to the east, but these two species are easily distinguished even when sterile by the petiole base of the larger leaves, which covers the Fur- thermore, the rachis of the raceme is pubescent in bud in S. obassia but not in S. hemsleyanus. S. hemsleyanus and glabrous or nearly so in 5. obas- 1. Rehder (1930) described Styrax huanus from Nanchuan Shi in southeastern Sichuan, considering the white-stellate tomentum on the lower laminar surface and the longer and glabrous stamen fila- ments as distinguishing it from S. hemsleyanus. Rehder (1912) also differentiated S. hemsleyanus var. griseus from the typical variety by the presence and quantity of pubescence on the lower laminar surface. Hwang (1987) treated this variety as a syn- onym of S. hemsleyanus, but agreed with Rehder on the status of S. huanus, citing the leaf pubes- cence difference and the type of trichomes as jus- tification for recognizing two species. Contrarily, we have found no basis for recogniz- ing any taxon other than a single species among these entities. Filament length and pubescence quantity exhibit a complete range of variation among individuals of Styrax hemsleyanus and S. huanus. Pubescent-leaved individuals of S. hem- sleyanus have a combined distribution largely over- lapping that of glabrous-leaved individuals, occur- ring from Yangcheng Xian, Shanxi Province (e.g.. T. W. Liu & Z. F. Zeng 226, 245, 1285, and 1393), at the extreme northern edge of the species’ range, south to Zhenxiong Xian, Yunnan Province (Exp. NE Yunnan 1161), whereas glabrous-leaved indi- viduals occur throughout the range of the group. The pubescent-leaved individuals also exhibit no obvious elevation or habit distinctions, and seem to occur sporadically, often near collection localities of glabrous-leaved individuals. Furthermore, sev- eral collections show an intermediate amount of pu- bescence between the types of S. hemsleyanus and S. huanus, even among collections from the vicinity of the type locality of S. huanus. In addition to the density of pubescence on the abaxial leaf surface, Hwang (1980) considered Sty- rax huanus distinguishable from S. hemsleyanus based on leaf trichome types. We consider this dif- ference to reflect merely the length of the stellate trichomes. The arms of the trichomes on the abaxial surface of the leaves are small (averaging ca. 0.15 mm long) in S. huanus versus some specimens of S. hemsleyanus (averaging ca. 0.5 mm long), but careful inspection of all collections of this group available to us indicates that arm length varies con- tinuously. In the protologue of Styrax ет three collections Wc are cited by Diels: A. Henr 5676, A. Henry 6895, and B. von emerge 2078. Because Diels's herbarium was B, we assume that the material on which the description of S. hem- sleyanus was based has been destroyed. We there- fore have designated the А specimen of A. Henry 5676 as the lectotype because duplicates are ap- parently more widely distributed than those of the other two syntypes (in particular, A. Henry 5676 is represented by two duplicates in a Chinese her- barium (IBSC), unlike either of the other syntypes). and only the A. Henry sesses collection locality data. There is no evidence of Diels's handwriting on the type material that we 676 specimen from А pos- have examined. Selected specimens examined. CHINA. Gansu: Kang Xian, Yang-ba-xiang, Z. Y. Zhang 16612 (PE). Guizhou: Jiangkou Xian, Niu-wei-he, Exp. Fan-jin-shan & Feng- huang-shan 402110 (IBSC, PE). rw Lushi Xian, Lao-chun-shan, K. M. Liou 4421 PE) Song Xian, Sang-shi, Long-di-man, Henan iue Dept. 1074 (PE): Xixia Xian, T L. Ph ichuan Xian, Pu-cha- biao- ben 20304 1 (PE); Shennongjia Linqu, т `n- ibi -jia ‚ Sino-Amer. Bot. Exp. (1980) 1 IC); Wufeng Tujiazu Zizhixian, H. 7 P" 5861 (IBSC, PE) Hunan: Sangzhi Xian, Ba-mao-xi-xiang, - ping-shan, B. G. Li 750286 (PE). Shaanxi: Ankang ‚ Tao- he-gong- p^ P. Y. Li 7778 (KUN); Fuping Xian, . К. T. Fu 4849 (РЕ); Long Xian, Shen-si, С "ki collector unknown 2346 (A); BG ; Ningshan Xian, Jiang-kou-xiang, J. Q. Xing 6053 (IBK): Pingli Xian, Da- dang-fu- V P. Y. Li 1380 (KUN); Shangzhou Shi, Long- j 6 ee et al. 34 (IBSC); Weinan Shi, Qing- gang- км i-gou, Z. B. Wang 15652 (IBSC, KUN PE); Zhashui Xian, Qing-ling-shan., collector unknown 66 Annals of the Missouri Botanical Garden (PE); Zhen’an Xian, X. X. Hou et al. 601 (IBSC); “енш Xian, Zhong-hong-xiang, P. Y. Li 2209 (KUN). Shan xw Xian, Sang-lin, Shu-pi-gou, Gan-qi-tong. 7. W c Z. B. Zeng 1285 (CAS). Sichuan: Chengkou Xian, Ku T. L. Dai 105634 (KUN, PE); Ebian Yizu Zizhixian, Wa-shan, E. H. Wilson 2578 (A, BM, E); Emeis- han Shi, E-mei-shan, W. P. Fang 14826 (A, KUN); Du- jiangyan Shi, W. P. Fang 2225 (А, E, IBSC, К); dinyang Xian, Sichuan Economic Pl. Exp. 2483 (PE); Xian, near Ta-chien-lu, 9 2 re 406 Ek Mp ме hos Shi; W P. Fang ‚К, m. Pingshan Xian, Sichuan Eco- nomic Pl. Exp. 1206 (PE); Pingwu Xian, H. L. Tsiang 19 (IBSC); Tianquan Xian, А C. Tai & C. M. Teng 4215 (KUN); Wushan Xian, A. Henry 5676A (IBSC); Wuxi Xian, Hong-chi-ba, G. H. Yang 59375 (IBSC, KUN, PE). Yun- nan: Zhenxiong Xian, Hua-shan, Exp. NE Yunnan 1161 (KUN). B. Clarke, in Hook. f.. Fl. 5. Styrax hookeri C. i 9 S. “Hookeri”]. Brit. India 3: 7 YPE: India. Sikkim: 5 7j m, J. D. Hooker s.n. (lectotype, designated here, K! [loan accession no. H2000/01016, fl branch]; isotypes, BM!, BR!, C!, K!, L[2]). Styrax macranthus Perkins, Bot. Jahrb. Syst. 31: 487. 1902. TYPE: China. Yunnan: Liichun Xian, region of Feng C gs Ling, 2121 m [2000 m, protologue], S of the Red River, A. Henry 10644 (lectotype, desig- nated here, i isotypes, A!, BM!, E[2]!, IBSC[2]!. 0!, PE!). Styrax сүи деш ыл. : Engl., а nr. IV. 241 (Heft 30): 7 Й : India. Assam: Mt. Sillet (Per- E 4400B еи ре, В нү o . 2 Styrax roseus Dunn, Bull. Misc. Inform. 8 1911: 273. TYPE: China. Sichuan: Ebian Yizu Zizhixian, . Wu [from protologue]. Wa-shan d 1912), OR m [2600 m, protologue], July 1903, E. H. Wil- оп.“ 1065 (holotype. K!; isotypes, Tar es IBSC!). r Rehder, in Sarg . Wilson. 1: 292. 2 [as 5 Рао |. P China. Sichuan: Chien Yiz m, qt 1 2576 (lec up e s A al is ae BMI, E! yee jd Notes Roy. Bot. Tengyu ; 13, G. Korres 9869 (holotype, Е!; isotypes, AL Shrubs or trees to 10 m tall. Young twigs gray- brown stellate-puberulent; older twigs purplish Petiole (2.5—24—6(-10) mm long. Two most proximal leaves on each shoot sub- 3—4(—6) cm but sometimes distalmost lamina smaller, char- brown, glabrescent. opposite to opposite. Lamina 6—8(-12) X taceous to thick-chartaceous, oblong, lance-ovate, or narrowly elliptic, often dark green when dry: apex acuminate to caudate, rarely acute, slightly oblique; base often slightly oblique, rounded to broadly cuneate, rarely shallowly cordate or nar- rowly cuneate; adaxially sparsely gray-white (rarely yellow-brown) pubescent with simple or 2- or 3- armed to stellate trichomes, glabrescent; abaxially glabrous or sparsely to gray-white stellate-pubes- cent to -tomentose, pubescence especially preva- lent on the veins and especially longer on the axils of the midvein and secondary veins; margin glan- dular-serrulate and slightly revolute; secondary veins 5 to 7 on each side of midvein, tertiary veins subparallel and perpendicular to the secondary nerves, together with the quaternaries adaxially plane and abaxially prominent. Fertile shoots 4—12 cm long, 3- to 5-leaved. Inflorescences arising from shoots of the current growing season; lateral inflo- rescences l- to 3-flowered; pseudoterminal inflo- rescences l- or 2-flowered or racemose, 2—4 cm long, (1)2- or 3(to 6)-flowered, rachis yellow stel- late-tomentose. Pedicel (2-)5-8(-13) mm long, yel- low-brown stellate-tomentose; bracteoles 3—4 mm long, subulate or linear, positioned at various plac- es along the pedicel but mostly near the middle, more rarely near the base, sometimes those toward the base of the inflorescence leaf-like. Flowers (1.3—)1.5-2.5 ст long. Calyx (3.5-)5-7(-9) х 4- 6(-11) mm, cupuliform; adaxially covered with 2- or 3-armed to stellate appressed trichomes, becom- ing glabrous proximally; abaxially yellow stellate- tomentose, often also with various amounts of larger scattered gray, tawny, orange, or brown stiff stellate trichomes especially proximally, within 1 mm from the margin more sparsely pubescent, somewhat scarious, brown when dry; margin truncate, undu- late, irregularly 2- or 3-lobed, or toothed, the teeth if present minute to 1 mm, deltoid to linear-deltoid, not Corolla (0.8—)1.2-1.9 cm long, white or pink, tube 3—4 mm long, glabrous, lobes 4(5), (11-)12-18 X (4—)5-10 mm, obovate to ob- ovate-elliptic, contiguous. adaxially appressed-stellate-pubes- cent or nearly glabrous, abaxially densely pale yel- low stellate-pubescent. Stamens 8 to 10; filaments 5-7 mm long, straight, distally attenuate, densely pubescent proximally, glabrous or sparsely stellate- pubescent distally, pubescence especially prevalent along the margin; anthers 3-5 mm long, wider than distal portion of filament; connectives glabrous. Style usually + out, occasionally subglabrous; stigma 0.2—0.5 wide, capitate. Fruit (1.0—)1.5-2 X (0.7— subglobose or ovoid, apex acute, occasionally short- white stellate-pubescent through- mm J1-1.5 em, rostrate, dehiscent; pericarp dry, 0.1—0.3(-0.6) mm thick, rarely up to 0.9 mm thick, outside at least Volume 90, Number 4 2003 Huang et a Revision 2 Styrax Series Cyrta 110 * Styrax hookeri 0 800 1 J Kilometers Бош, Н _ кыз. ML E \ B = v \ { - Mag p oy tw А E ig j 2 2 ئ‎ d AA ^ аф \ D ee d ON ы и Я — = л! ыш! ЖЕ ee u er АЙЫ. e « \ of А À eet ` Е VN 66 oe À 130 Figure 6. Geographic distribution of Styrax hookeri. faintly longitudinally striate and + rugose when dry, gray-yellow stellate-tomentose, inside glabrous. Seeds beige or brown, subglobose or ovoid, smooth, glabrous. Illustrations. | Anonymous, Ic. Cormophyt. Sin. 3: 337, fig. 4627 (as S. roseus). 1974; F. T. Tai & T. C. Pan in W. P. Fang, Fl. Sichuan. 1: 428, fig. 166. 1981 (as S. roseus); C. Y. Wu, Fl. Yunnan. 3: 424, pl. 120 (7-10). 1983 (as 5. perkinsiae); ibid.: 433, pl. 124. 1983 (as S. roseus and 5. macranthus); S. M. Hwang & C. J. Qi in W. C. Cheng, Sylva Sin. 2: 1607, fig. 802. 1985 (as S. perkinsiae); ibid.: 1621, fig. 814. 1985 (as S. macranthus); ibid.: 1622, fig. 815. 1985 (as S. roseus); T. L. Ming in C. Y. Wu, Fl. Xizang. 3: 869, fig. 335 (1—3). 1986; S. M. Hwang, Fl. Reipubl. Popularis Sin. 60(2): 90 pl. 30 (6—9). 1987 (as S. perkinsiae); ibid.: 101, pl. 34 (6—10). 1987 (as S. uv ibid.: 103, pl. 35 (1—6). 1987 (as S. roseus); W. Q. Yin in Y. C. Xu, Ic. Arbor. Yunnan. 2: 894, pl. 471 (7-12). 1990 (as S. roseus); D. G. Long in Grierson & D. С. Long. Fl. Bhutan 2(2): 577, fig. 58(e-g). 1999 (as S. gran- diflorus); Z. Y. |С. Y.] Wu & P. H. Raven, Fl. China Ill. 15: 199, fig. 199 (6—9). 2000 (as S. perkinsiae): ibid.: 203, fig. 203 (7-12). 2000 (as 5. macranthus); ibid.: 204, pl. 204 (1—7). 2000 (as S. roseus). Phenology. Flowering: Fruiting: April-November, January. Distribution. Bhutan (Lhun ТЫ, Tongsa, and Wangdi Phodrang), China (Guangxi, India March-September. Tashigang. Guizhou, Sichuan, Xizang, and Yunnan), (Arunachal Pradesh, Assam, Meghalaya, Nagaland, Sikkim, and West Bengal), Myanmar (Kachin State), and Nepal (Mechi); Figure 6. Habitat. wooded habitats and forest edges on mountain slopes; 730-3352 m Vernacular names. Da-rui-ye-mo-li (China, Sichuan; Exp. E-shan 155), Fen-hua-an-xi-xiang In a variety of open or semi-open (Hwang, 1980), Fen-hua-ye-mo-li (SW China; Anonymous, 1 ‚ Feng-chun-an-xi-xiang (Hwang, 1980), Lü-chun-an-xi-xiang (Hwang, 1987), Mai-mu (China, Sichuan; Z. T. Guan 8448), Mao-zhu-ye-mo-li (China, Yunnan; Wu, 1983), Qing-ye-dong-gua-shu (China, Guangxi; 5. О. Chen 14376), Rui-li-an-xi-xiang (Hwang, 1980), Rui-li- ye-mo-li (China, Yunnan; Anonymous, 1974), Shui- liang-zi (China, Sichuan; Sichuan Economic Pl. Annals of the Missouri Botanical Garden Exp. 169), Trali Shing (Bhutan; А Ludlow et al. 18802), Wa-shan-an-xi-xiang (Tai & Pan, 1981), Xi-shu-mai-mu (China, Sichuan; Z. T. Guan 8197), Yun-nan-ye-mol-li (China, Yunnan; Wu, 1983). Styrax hookeri is a common and widespread spe- cies, occurring at relatively high elevations from eastern Nepal along the Himalayas through Assam, India, and extending into southwestern China. It is apparently most common in Yunnan Province. Our those of Perkins Styrax hookeri differs from (1907) and Hwang (1987). We spes with Perkins (1907) that this species is not treatment of in extreme variant of 5. serrulatus, as suggested by Clarke (1882). Perkins (1907) treated S. hookeri narrowly by simultaneously recognizing 5. caudatus Perkins (Assam, India) and 5. macranthus Perkins Later, Perkins (1910) distinguished S. hookeri var. yun- (southern and eastern. Yunnan Province). nanensis Perkins from the typical variety by its smaller and narrower leaves. This collection is geo- graphically isolated (northeastern Yunnan Prov- ince) from variety hookeri sensu Perkins (Himala- yas). Three new species of Styrax from the provinces of Yunnan and Sichuan (S. е Rehder, S. roseus Dunn, and S. shweliensis W. Sm.) were subsequently described by various au- thors. Their types, along with that of S. macranthus, are centrally located between the apparently dis- junct localities of S. hookeri sensu Perkins (i.e. its known range as of 1910). These species were de- limited primarily by poorly defended features of the Hwang (1987) considered S. macranthus, S. perkinsiae, and S. ro- leaves, inflorescence, and calyx. seus to be separate species, and treated S. shwe- liensis as a synonym of S. perkinsiae. Because Hwang's treatment was in a regional flora of China, Hwang apparently did not examine collections of S. hookeri sensu Perkins from outside China. She may also not have had access to the type of S. hookeri var. yunnanensis, which she cited as a syn- onym of S. grandiflorus Griff. (= S. japonicus). Sty- rax hookeri var. yunnanensis has a shorter pedicel and a calyx with scattered orange stiff trichomes, among other features, that clearly distinguish it from S. japonicus and establish its placement within our concept of S. hookeri. ccess to many more collections than were avail- able to either Perkins or Hwang has allowed us to reassess these high-elevation taxa. We interpret the highly overlapping range of morphological variation exhibited by this group as warranting only a single widely distributed species. Although trichome type and the amount of pubescence on various parts of the plants are diagnostic characters in the delimi- tation of some Styrax species, these features are highly variable in S. hookeri. The vestiture on the inner surface of the corolla lobes and the lower laminar surface consists of either long or short stel- late trichomes, or else is lacking; that on the lower laminar surface can be sparse to dense. The styles are usually densely stellate-pubescent nearly throughout, at least proximally, but in some speci- mens from northeastern Yunnan Province and the Khasi Hills in Meghalaya, India, they are glabrous. Several other characters are also variable across the range of S. hookeri (e.g., leaf shape and size, flower and fruit size, petiole and pedicel length). We de- tect no gaps in character state variation, either as- sociated with gaps in other characters or with geo- graphic or ecophysiographic variables, for use ir =, recognizing any of the synonyms of 5. hookeri. Although recognizing species segregates of Sty- rax hookeri is not warranted, there has clearly been some regional isolation among populations of this species resulting in geographically correlated (al- though not discontinuous) morphological trends. For example, individuals with the most densely pu- bescent abaxial leaf surfaces occur in western and central Yunnan Province, with consistently. gla- brous or sparsely pubescent populations to the west in the Himalayas. Some collections from the edge of the species’ range exhibit slightly atypical fea- tures. The collections from the Khasi Hills (e.g., C. B. Clarke 43631A) have leaf margins with more nu- merous and prominent serrations than are typical in the species. This variation, however, also appears in other areas scattered throughout the species’ range. Perhaps the most distinctive morphological variants within 5. hookeri come from the provinces of Guangxi and Guizhou (e.g., X. H. Song 272 and 907, C. Wang 41180, S. О. Chen 14376, and Exp. Guizhou 6836). lance-elliptic, subcoriaceous leaves and/or relative- These specimens have narrowly ly small fruits ca. 7 mm wide. We have opted against the formal recognition of these populations of S. hookeri at an infraspecific level because many specimens collected from areas scattered through- out the range of the species exhibit intermediacy in these characters. Differences between Styrax hookeri and all sym- patric imbricate-flowered species of Styrax are ad- dressed in the discussions under S. buchananii, S. hemsleyanus, 5. limprichtii, S. odoratissimus, and S. rugosus. Flowering individuals of 5. hookeri with a pedicel length approaching that of S. japonicus can usually be distinguished by the presence of scat- tered orange or brown stiff long-stellate trichomes on the calyx. Fruiting individuals are more easily distinguished, because the pericarp of 5. hookeri is usually thinner and at least faintly longitudinally Volume 90, Number 4 2003 Huang et al. 515 Revision of Styrax Series Супа striate (vs. irregularly rugose), and the seeds are smooth (vs. usually finely reticulate-fissured or ir- regularly rugose). Sterile specimens of S. hookeri can be distinguished from those of S. japonicus by a tendency toward elliptical leaves with acuminate to caudate apices and subparallel tertiary veins that are conspicuously raised only abaxially, versus a tendency toward subrhombic leaves with acute to slightly acuminate apices and narrowly reticulate tertiary veins that are conspicuously raised on both surfaces. These characters, however, exhibit some degree of overlap The closest putative relatives of Styrax hookeri (i.e., 5. limprichtii, S. rugosus, and S. wilsonii) oc- cur at relatively high elevations scattered through- out southwestern China and northern Myanmar. These species share with 5. hookeri a typically subglobose or ovoid fruit ا‎ or apiculate at the apex and with usually at least a faintly longi- tudinally striate pericarp, seed surfaces thal are smooth or irregularly rugose, and a calyx with usu- ally various scattered orange or brown stiff stellate trichomes larger than those of the base tomentum. Styrax hookeri is easily distinguished from them. however, by the characters in couplet 14 of the key. The presence of abaxially glabrous or sparsely pu- bescent leaves and a densely pubescent style can often be used to distinguish S. hookeri from these species as well but are not as reliable. According to the protologue, the type locality of Styrax roseus is Mt. Wu ("Wushan" in Pinyin) in Sichuan Province. Hwang (1987) interpreted this as Wushan Xian in eastern Sichuan, but the label on the type indicates that the locality is in western Sichuan. Rehder (1912) confirmed that the western Sichuan locality is correct by citing the type local- t. Wa (a variant of Mt. Wu) in western Sich- uan, also the type locality of S. perkinsiae. ity as The type material of Styrax hookeri at K consists of two sheets of J. D. Hooker s.n. from Sikkim. both of which possess flowering and fruiting branches. Individuals of S. hookeri flower and fruit at different times of the year within the same geographic re- gion, indicating that these branches were collected on different dates. Thus, we interpret the material as consisting of four syntypes. This conclusion is supported by the writing “2 Styrax Sikkim” fol- owed by Hookers initials in his handwriting on each of the sheets, implying that there are two Sty- rax specimens on each sheet. There appears to be no basis for a decision regarding selection of the most appropriate specimen as the lectotype other than the condition of the material and the fact that one of the sheets possesses what is likely to be a field label in Hookers handwriting. Thus, we have lectotypified on the largest branch with the most reproductive material on this sheet. In further sup- port of our selection, this branch is also the largest and most floriferous of those on either sheet. The holotype of Styrax macranthus at B is pre- sumably destroyed. It is possible that Perkins only saw the specimen at B; none of the other sheets of A. Henry 10644 that we have examined possess Perkins's annotation label, and no herbarium other han B is mentioned in either Perkins (1902) or Perkins (1907) to confirm Perkins's examination of Dad additional material. On this basis, we have chosen the K specimen of A. Henry 10644 as the lectotype, because Kew was the location of Henry's head- quarters. The holotype of Styrax hookeri var. yunnanensis at B is presumably destroyed. We have designated the specimen at P as the lectotype because it is the only duplicate specimen that we have seen, and it possesses Perkins's annotation. The protologue of Styrax perkinsiae cites E. H. Wilson 2576 as the type. There are two sheets of this number at A, but each has a different date. The word “holotype” is written on one of the sheets, but this is apparently not in Rehders handwriting and it is not clear who wrote it. As such, these sheets must be considered syntypes. We have cho- sen the specimen that was collected in July 1908 as the lectotype because the material has more flowers for examination than the 17 September 1908 collection. Also, because the word “holotype” is written on this sheet, designating this sheet as the lectotype will avoid the risk of undue confusion. elected specime dy examined. BHUTAN. Lhun Tshi: Пе hung, Khoma Chu, F Ludlow et al. 18802 (A, BM). egre Yonpu ун near Tashigong Dzong, F Ludlow et al. 12593 (BM, E). Tongsa: 1 km S of Tongsa, Grierson & " С. Long 1107 (Е, di Рона ie Chu Valley, А Ludlow & С. Sherriff 3133 (BM, E). HINA. Guangxi: Nandan Xian, C. Wang а (A, l AS, IBSC); Rongshui Miaozu Zizhixian, чый -fang-xiang, Jiu-wan-da-shan, Chen 14376 (IBK, IBSC, KUN, PE). Guizhou: Anlong Xian, Long-shan-xiang, Exp. Guizhou 4737 (KUN); Bijie Shi, Sheng-ji-xiang, P. H. Yu 240 (KUN): Dafang Xian, Bai-na-qu, Jiu- long-shan, Exp. Bi-jie 847 (PE); Libo Xian, Dong-ting, X. H. Song 272 (K, MO); Panxian Ba-da-shan, Exp. An- n. 890 (KUN): Minglong: Xian, Exp. S Guizhou 205 (KUN); Xingyi S Guizhou 6836 (IBSC, PE). Sichuan: Baoxing Ps g-shan, Tuan-niu-ping, Nan-shui-bei-diao-dui 1871 ( ^E) Ebian Yizu Zizhixian, T. T. Үй 653 (A, IBSC, PE); uf Shi, E-mei-shan, G. H. Yang 55400 (IBSC, KUN, РЕ); Ganluo Xian, Hai- tang, Sichuan pae Pl. Exp. 4086 (КОМ, PE); Han- yuan Xian, X. Zhao 511 (РЕ); Leibo Xian, Ma-hu- ia^ Рец xiang, Tang-jia- is Sichuan Economic Pl. Exp. 315 (KUN, PE); Mabian Yizu Zizhixian, Da-zhu- wr Shan- Xian, 5. К. 840104. mu-gang, T. H. Tu 5494 (PE); Mao KUN); Meigu Xian, Shu-dang-xiang, Sic A mint 516 Annals of the Missouri Botanical Garden Pl. Exp. 13556 (PE); Mianning Xian, from Guanling Xian to Muli Xian, 5. К. Wu 2204 (KUN); M x A Zizhix- ian, from Guanling Xian to Muli Xian, 5. K. Wu 2: KUN, PE); Pingshan Xian, Wu- Ai han, 0. S. Zha ‚ You-jia- ping > Guan 8059 (PE); C. Hsieh Hor (IBSC, PE); о Xian, е dm Н. L. Tsiang 35129 (IBSC, E Xuy- ong Xian, Yi-shui-qu, е m mic Pl. | (KUN); Yanyuan Xian, Ni-ba-sha LS. Zhao 309 РЕ): Jao-an, Da- die a S Economic РІ. PE). Xizang lu Motuo Xian, Han-mi, é i & S. Z. Cheng 5062 (PE). — brad Shi, San- И -qiao, China-USSR team |, РЕ); Bine ‘huan Xian, Ji- zhu-shan, S. Y. Bao м . Ww 00 ке — ~ . Ching 22673 (KUN, PE[2]): з shan, Sino- British Exp. RR ай 850 (А, KUN); Eshan Yizu Zizhixian, Huang-cao-ling, Exp. E- shan 88155 (KUN): jue i Shun-ning, Wu-mu- exa " Үй 16624 (A, E, PE); Fugong Xian, Fen- qua p. Qinghai & Me E KUN); Fumin ouk omn nping, Н. F. Handel-Mazzetti 6119 (A, E); Fuyuan Xian, Shi-ba-lian-shan, Xiao-nao-chang, Exp. P us he 2356 (KUN); Gengma Daizu Wazu Zizhixian, Xi-shan, China-USSR team 5570 (IBSC, PE); Gongshan da iit Мили Zizhixian, from Gong-shan to "n long, Da-ba-di, Gao- li- -gong-shan, P Y. Mao wa ig ‚ РЕ); poe Zizhixian, Feng-kua-shan, M. "à i 3493 (IBSC, KUN); Lanping Baizu Pumizu Zizhixian, Bing-zhong, Luo- he, X. Е Deng 791361 (KUN); Lijiang Naxizu Zizhixian Lichiang Range, D. McLaren L100A (BM); Longline а Salwin- Кк divide, T. T. Yü 20294 (A, E, PE); Lüchun Xian, Feng-chun-ling, 5 of Red River, А. m 10644 (A, BM, E[2], IBSC[2]. K, MO, PE); Lushui Xian, from Ya-kou to Pian-ma, S. K. Wu 8478 (KUN); Ruili Shi, Luckoag-Salween divide, G. Forre e cn Shuangbai Xian, Shuang-bai-si-c а W. C. Yin 490 (IBSC, KUN[2], PE); MCA Lahuzu Wazu Bulangzu eye Zizhixian, Tai-ping-xiang, J. S. Xing 832 „РЕ Suijiang Xian, Luo-han-ping, B. S. Sun ; Tengchong Xian, Lang-ya-shan, D. Y. Xia ; We ishan Yizu Huizu Zizhixian, Wu-liang- shan, Me Malis. Y Tsiang 12204 (IBSC); Weixi Lisuzu Zizhixian, Wei-deng-xiang, Exp. Qinghai & Xizang 6603 (KUN); Wenshan Xian, Lao-jun-shan, K. M. Feng 22401 (IBSC, KUN); Yangbi Yizu Zizhixian, Shi-zhong. xiang, Shang- chang, Sino- British Exp. e 269 (A, E. К, KUN); Yanjin Xian, Cheng-feng-shan, NE Yunnan (1970s) 1163 (KUN); Yao'an Xian, Tai- -ping-xiang, Y. Chen ¢ B. Bai 562 (KUN); Yiliang Xian, Cao-tian-ma, Exp. NE Yunnan (1970s) vin (KUN, PE); Yongping Xian, betw. Sha-yang & Chu-tong, G. Forrest 21112 (A, BM, Е, K, с ы |, ОС) . Yongshan Xian, H. T. Tsai 50936 (A, IBSC[2]. N, PE); Yuanjiang | ҮЗ Yizu Daizu Zizhixian, Hou- io a Lin 770497 (KUN); Yuxi Shi, Gao-lu-shan, 5. K. Wu 57 (KUN); иан Shi, Tang- Mar pa, К. eres 4951 (P; des Xian, Snow Range, T. T. Yü 17074 (A, E, KUN, PE); Zhenxiong Ae Mo- b: X. W. Li 173 (IBSC). RE Arunachal Pradesh: Pachakshiri Dist., Lalung, А Ludlow x al. 3713 (BM. E). Assam: Dr. King’s collector s.n. (BM, L). wir Khasi Hills, J. D. Hooker & J. J. P s.n. (BM, « ‚ К, 1). Nagaland: Naga Hills, Kohima, W. N. ae 2. 25269 (L, UC). Sikkim: J. D. Hooker s.n. (BM, BR, С, 1]2]). ayer Bengal: Takdah, Darjeeling, H. Hara & к Togashi 2141 (B ; KYO). MYANMAR. Kachin State: N Triangle (С sp Tama Bum), К F. К. Ward 20990 (A, BM, E). NEPAL. Mechi: Salpa Dara, J. D. A. Stainton 8332 (BM) = 6. Styrax japonicus Siebold & Zucc., Fl. Jap. 1 53. 1837-1838 [as S. "japonicum"]. Супа ja- ponica (Siebold & Zucc.) Miers, Ann. Mag. at. Hist., ser. 3, 3: 279. 1859. TYPE: Japan. Kumamoto Pref., Simabara, /. Keiske s.n. (lectotype, designated here, L [accession по. 908240-682] not seen; digital image of lectotype!). Kyushu: Styrax р Griff., S. Not. РІ. Asiat. 4: 287. 1854 [as randiflora"]. ' у ы, W. Gr si K!; iso lype, GH Styrax japonicus var. "acri Gilg, Bot. Jahrb. Syst. : (Beibl. 75): 58. TYPE: China. лө Qingdao Shi, Lao- Eras Aug. 1907, O. Nebel s.n. (ho- lotype, B destroyed). 3671 (Perkins, 1907) (holotype, Styrax cavaleriei b Lév., Repert. Spec. ye Regni Veg. 4: 331 7 [as S. * e YPE: China. Guizhou: iod Xian, 7 May 1903, J. Pis s 997 (holotype, E!; isotype, A!). Styrax ка Lév., ite epert. Spec. 1 Regni Veg. 4: Le S. "Bodinieri |]. TYPE: жшк Guin Cua Shi, vicinity of ae , Colleg ‚ Apr. 1898, E Bodinier 2221 (holotype, E!; "icd to W bo ‚ А! SUME СЕР Perkins, 1910 [as S. nan: * Nain Yizu Repert. gens “Duclouxit” |. 1 Zizhixian, near т. bos. Regni Veg. PE: ex Yun- у Tsin, 20 E 904, Е Ducloux 2716 pedea "ds 'signated , P!). PN СА Н. Lé Кереп. Spec. Nov. Regni Ve 1912. TYPE: China. Cnr Dushan Xian, ne 1902, E. Bodinier s.n. (holotype, E*; iso- A!, Е!). Styrax нот Hayata, Icon. Pl. Formos. 5: 121. 1915. ix japonicus var. kotoensis (Hayata) Masam. & ке, е Кер. Taihoku Bot. Gard. 3: 65. 1933. TYPE: China. Taiwan: Taitung Xian, Kotosho [1 anyu Hie July 1912, Y. Tashiro, T. Kawakami & S5 aki 44 [collection number not indicated in protologue | (holotype, ТІ; isotype, IBSC!). Styrax jippei-kawamurai Yanagita, J. Soc. Forest. 15: 693. 1933 [as S. "Jippei- саен " Styrax japonic us var. Jippei-kawamurai (Yana H. Hara, Enum. Sperm. Jap. 1: D dia 8 [as s Puis var. Jip- pei-Kawamurat"]. rax japonicus f. jippei-kawa- murai (Yanagita) T Yam: azaki, Fl. Japan dis 1993 [as S. “japonicus f. jippei- pori "| TYPE ue Honshu: Shizuoka Pref., O Shima Island, Jan. ), J. Kawamura s.n. (type isla rial missing). Styrax йы е var. yS Masam., ans. Nat. Hist. Soc. Taiwan 25: 2 935. TYPE: Japan Ryu- 4 Islands: d P с. Oct. 1923, Ipse 1. (holotype, TAI not seen). Soros t philippinensis Merr. & Quisumb., Philipp. J. Sci. б Y PE: Philippines. a C a- miguin үе t. Malabsing, 9 Mar. 1930, H Edaño 79248 Шш, NY not seen; еа Ly). Styrax japonicus var. zigzag Koidz., Acta Phytotax. Geo- Volume 90, Number 4 2003 Huang e 517 eta Revision a Styrax Series Cyrta bot. 6: 212. 1937. TYPE: Japan. Honshu: Iwate Aena Rikuchiu, Higashiiwaigun, Ohtsuhomura, C. Toba s.n. (holotype, KYO not seen). Styrax japonicus f. parviflorus Y. Kimura, J. Jap. Bot. 16: 59. 19 ые. Iwaya-mura, 30 May 1937. | ” Yosioka 23 (holotype, TI!). Styrax japonicus var. angustifolius Koidz., Acta Phytotax. Ge 55. 1941 [as = Japonicum var. dra Бар — : Wakayama Pref., el 940. С. Koidzumi s. s.n. "bw YO! omentosus Hatus., J. Jap. Bot. 29: s S. "Japonicum var. tomentosum’ °]. Sty- rax ponies sí tomentosus See ima) T. Yamazaki, E: Japan. Ryukyu Is- lands: Kagoshima Pref., e hlad Group, Nak- anoshima Island, Apr. 1936 and 18 Aug. 1933 [1934 from protologue], T. Naito s.n. (holotype, FU not seen; photo of holotrpé, TI!). Styrax Japonicus f. rubicalyx Satomi, J. Geobot. 6: 110. 1957. TYPE: Japan. Honshu: Ishikawa Pref., Kaga, ен -pass, Asakawa-mura, Kahoku-gun, 20 July . N. Satomi s.n. (holotype, KANA not seen). Sb j japonica s var. longipedunculatus Z. Y. Zhang, Fl. Tsinlingensis 1(4): 395. 1983 [as S. “japonica var. longipedunculata"]. TYPE: China. Gansu: Wen Xian, Bi-kou-zhen, Bi-shan-gou, Quai-miao, 750 m, 3l dd 1967, C. L. Tang 1739 nie. HW not een bos Japoniei “ су Z. Y. Zhang, Fl. Tsinlin- > 1(4): . 1983 [as 5. “Japonica var. nervil- T t T d 9028 (holotype, HW. not seen; isotype, m Beg papers : 3 censa T. Yamazaki, Fl. Japan 3a: 04. 1 YPE: Japan. Honshu: Tokyo Pref. To- kyo, cultivated. Sep. 1991, T Yamazaki s.n. (ho en). lotype, TI not s Shrubs or trees to 8(-10) m tall. Young twigs brown, sparsely gray-yellow or pale yellow stellate- pubescent; older twigs gray or nigrescent, glabres- cent. Petiole (2—)4—7(-10) mm long. Two most prox- imal leaves on each shoot (when both present) subopposite to opposite. Lamina 3-11 х 2—5(-7) cm, chartaceous to thick-chartaceous, oblong-ellip- tic, ovate-elliptic, ovate to ovate-lanceolate, or sub- rhombic; apex acute to slightly acuminate; base cu- neate to broadly cuneate or subrounded, often decurrent into petiole; adaxially sparsely stellate- pubescent when young, especially prevalent on veins, glabrescent; abaxially glabrous except along the vein and the axils of the secondary veins; mar- gin entire to apically remotely serrate; secondary veins 5 to 8 on each side of the midvein, tertiary veins reticulate, conspicuously raised on both sur- faces. Fertile shoots 2-9 cm long, 1- to 4-leaved. Inflorescences arising from shoots of the current growing season; lateral inflorescences 1- or 2-flow- ered; pseudoterminal inflorescences 2-flowered or racemose, 1—4 cm long, 2- to 5-flowered; rachis glabrous or pubescent. Pedicels (10—)15-50 mm long, the longer pedicels on each twig > 15 mm long. usually equal to or longer than subtended flower, slender, glabrous or stellate-pubescent; bracteoles 3—5 mm long, linear or subulate, usually positioned at the base of pedicels, sometimes those toward the base of the inflorescence leaf-like. Flow- ers (1.2-)1.5-2.5(-3) em long. Calyx 4-7 X 3—5 mm, cupuliform to funnelform; adaxially glabrous; abaxially glabrous or sparsely to densely white or gray-yellow stellate-pubescent, if stellate trichomes present, within 1 mm from the margin more sparse- ly pubescent or glabrous, somewhat scarious, brown when dry; margin with 5 irregularly spaced trian- gular-ovate teeth 0.5-1 mm long or sometimes less, not contiguous. Corolla (0.8—)1.0-1.6(-2.3) cm long, white, occasionally pink, tube 3-5 mm long, glabrous, lobes 5 or 6, 11-20 x (3-)5-7(-9) mm, ovate, oblong-ovate, obovate, or ovate-lanceolate, apex obtuse, densely appressed-stellate-pubescent on both sides, sometimes sparsely pubescent adax- ially. Stamens 10 to 12; filaments 5-6 mm long, straight, slightly broadened proximally and white- villous, distally attenuate and glabrous; anthers 4— 5(-10) mm long, wider than distal portion of fila- ment; connective glabrous. Style proximally white stellate-pubescent, distally glabrous; stigma 0.2— 0.4 mm wide, punctiform. Fruit 0.8-1.5 х 0.8-1 cm, ovoid or ellipsoid, apex apiculate, usually de- hiscent by 3 valves from the base; pericarp dry, .4—1.0 mm thick, dry, outside coarsely and irreg- ularly rugose when dry, gray or gray-yellow stellate- tomentose, inside glabrous. Seeds brown, ellipsoid, smooth or finely reticulate-fissured to irregularly rugose, glabrous. Selected illustrations. Siebold & Zucc., Fl. Jap. 1: t. 23. 1835; Griff., Ic. Pl. Asiat. 4: t. "gi 1854 (as S. grandiflorus); Regel, Gartenfl. 17: t. 583. 1868; Hook. f., Bot. Mag. 98: t 0. Th (as S. serrulatus Roxb.); Gard. Chron. ser. 2, 24: fig. 166. 885; Gartenflora 36: fig. 89. 1887; Dippel, Handb. Laubholzkunde 1: fig. 207. 1889; Gard. Chron. ser. 3, 65: 279, fig. 140. 1919; Addisonia 7: t. 231. 1922; Merr. & Quisumb., Philipp. J. Sci. 56: 316, pl. 1. 1935 (as л E W. P. Fang, Ic. Pl. Omei. 1(1): t ; Anonymous, lc. Cormo- phyt. Sin. 3: d P eR 1974; F. T. Tai & T. C. Pan in W. P. Fang, Fl. Sichuan. 1: 424, fig. 164 10). 1983 (7-10 as S. grandiflorus); L. Yang in Y. K. Li, Fl. Guizhou. 2: 541, fig. 231. 1984 (including S. japonicus var. calycothrix); S. M. Hwang & C. Qi in W. C. Cheng, Sylva Sin. 2: 1614, fig. 808. 518 Annals of the Missouri Botanical Garden [110 130 quem | tyrax japonicus | |o 400 800 | | Kilometers | | ^s ( [b ( , "e \ , ( P > n f 8 \ Р ; il . 2 of e | e ° Na / \ he Bt ОШ 2 g ' | р ve EN `з 9 : ч ^ a 5 N A . " \ ^. et SA e?e u C е $ NC э_ y ° е 5 » E. \ e e MULA C ме i f e * у æ e Е o9 МЫ . е. . 66 ов $ < ° ee le uU. “ee " e ee е v -R Ы t - o e “be * yg "ы И о oe 956 e. . e € Ca ^r > ee v: oe ee v * ee? ә í • Ф P of Су ga we Figure 7. Geographic distribution of Styrax japonicus. 1616, fig. 809. 1985 (as S. grandiflo- rus); S. M. Hwang, Fl. Reipubl. Popularis Sin. 60(2): 93, pl. 31. 1987 (1-11; 8—11 as 5. grandi- florus); S. M. Hwang in F. H. Chen, Fl. Guangdong 1985; ibid.: 1: 387, fig. 419. 1987; ibid.: 387. fig. 420. 1987 (as S. grandiflorus); J. Q. Liu in L. G. Lin, Fl. Fu- jian. 4: 351, fig. 284. 1989; X. M. Liu in X. H. Qian, Fl. Anhui 4: 65, fig. 1769. 1991; 5. n Wang in B. Z. Ding, Fl. Henan 3: 230, fig. 1775 (1—4). 1997; Z. Y. |C. Y.] Wu & P. H. Raven, FI. China Ш. 15: 200, fig. 200 (1—11). 2000 (8-11 as S. gran- diflorus). Phenology. Flowering: January-October, De- cember. Fruiting: February- November. China. (Anhui, Gansu, Guizhou, Distribution. Fujian, Guangdong, Guangxi, Henan, Hubei, Hunan, Jiangsu, Jiangxi, Shan- dong, Shanxi, Sichuan, Taiwan, Yunnan, and Zhe- jiang), Japan и Kyushu, Ryukyu Islands, and Shikoku), Laos (Houaphan and Xiangkhoang), Myanmar (Kachin State), North Korea (Pyongyang City), Philippines (Babuyan Islands and Batan Is- lands), South Korea (Cheju, Inchon City, Kangwon, Kyonggi, North Cholla, North Kyongsang, Seoul Hainan, Shaanxi, City, South Cholla, South Chungchong, and South Kyongsang), and Vietnam (Cao Bang): Figure 7. Habitat. In a variety of open wooded habitats, in woodlands and forest edges, successional habi- ats, rarely in dense shade, mostly in mesic micro- — habitats, such as canyons, draws, ravines, and other 3-2700 m Vernacular names. Benigaku-egonoki (Japan; Satomi, 1957), Chun-shu (China, Sichuan; Sichuan Pl. Exp. 12), Da-hua-an-xi-xiang (Hwang, 1980), Da-hua-ye-mo-li (Anonymous, 1974), Diao-gong-zai (China, Guangxi; Z. А Hunag 15), Egonoki (Japan; Shigetaka Suzuki AA1108), Er-wan-tao (China, Guangdong; X. Q. Liu 24221). Ganboku-egonoki (Japan; Koidzumi, 1937). Gou- bi-zi-shu (China, Hunan; /„ H. Liu 1877), Gou- ла -zi (China, Zhejiang: Zhejiang Bot. Res. Team ei-Cha-hua (China, Sichuan; Tai & Pan, 1981), Hime-egonoki (Japan; Anonymous, 1940), Hosoba-yegonoki (Japan: Koidzumi, 1941), Hou- feng-teng (Hwang & Qi, 1985), Hui-lu-tui (China, Henan; collector unknown 334), Jun-qian-zi (China, Shaanxi; P. C. Kuo 2180), Kótó-egonoki (Japan: Masamune & Suzuki, 1933), Lai-xiang-mei (China, riparian situations: Economic Volume 90, Number 4 2003 Huang et a 519 Revision of Styrax Series Cyrta Hunan; P. C. Tam 63659), Lan-yu-an-xi-xiang (Li, ION Lan-yu-ye-mo-li (Li, 1978), Li-jia (China, ; P. C. Tam 62899), Ling-dang-hua (China, Guizhou: R. B. Jiang 521), Mao-e-ye-mo-li (Anon- ymous, 1974), Mo-li-bao (China, Guizhou; P. C. Tsoong 1032), Mu-jie-zi (China, Hubei; Hwang & i, 1985), Mu-xin-zi (China, Sichuan; Z. R. Zhang 25145), Ni-chi-yang (China, Jiangxi; Jiangxi Nor- 1243), Padong mao (Laos; А. F G. Kerr 20941), Ru-xiang-shu (China, Yunnan; Y. Y. Hu 580582), Sa-ye-shu (China, Hunan; P C. Tam 61328), Sei-ton-kwa цен Siebold, 1835-1841), Shui-dong-gua (China, ; О. Н. Lu 2441). 1835—1841), Ttaejuk пати (South Korea; В. К. Yinger et al. 2525), Yang- huai-zi (China, Yunnan; L. S. Xie & M. Cai 440), Yang-jiao-shu (China, Sichuan; H. F Zhou 11150). Yao-bai-he (China, Shaanxi; J. Q. Xing 9018), Ye- hua-bei (Hwang, 1987), Ye-mo-li (China, Guizhou: К. M. Lan 351), Ye-ping-guo (China, Hubei: Liu 142), Ye-wu-wei-zi-shu (China, Guangxi; Z. К Hunag 19), Ye-xun-zi (China, Henan; P C. Kuo 3945), Zhuang-shu (China, Sichuan; Tai & Pan, mal Univ. Guangxi Tsisjano-ki (Japan; Siebold, Probably the most common species of Styrax in Asia, S. japonicus occurs from Japan to Myanmar south to Vietnam and Laos, with a few outliers in the far northern Philippines and the islands of Lan- yu and Hainan, China. Styrax japonicus is distinguished from all other species of Styrax by pedicels that are usually great- er than 1.5 cm long (vs. = 1.3 em long) and equal to or longer (vs. shorter) than the subtended flower. Populations on Hainan Island and in several local- ities in Yunnan Province, China, have the shortest pedicels (as short as 10 mm, although pedicels on specimens from these areas can be found that are at least 15 mm). Also, some specimens of S. hookeri in Yunnan Province have unusually long pedicels that approach the кч of the shortest pedicels of S. japonic us. is species 1s highly variable across its range. The pubescence on the calyx consists of a sparse to dense layer of stellate trichomes, or is absent. Like other widespread species of Styrax on other continents (e.g., S. americanus, S. glabrescens, S. sieberi Perkins), S. japonicus exhibits variation in the size and shape of the flowers, fruits, and es- pecially leaves. Consequently, many varieties and forms of this species have been recognized. O study, however, does not reveal consistent а nations of characters for use in delimiting infraspe- cific taxa in 5. japonicus. For example, the type of S. japonicus f. parviflorus Y. Kimura has extremely small flowers and leaves, but collections with one or both of these features have also been collected in Yunnan Province (e.g., H. T. Tsai 55793). Such a pattern suggests that this extreme represents a sporadic variant rather than a taxonomically signif- icant geographic entity. Other populations in the Ryukyu Islands and Honshu Island, Japan, and Lanyu Island, China, have flowers and leaves that are larger than is typical for the species. These pop- ulations have been recognized in some works as varieties kotoensis (Hayata) Masam. & Suzuki, jip- pei-kawamurai (Yanagita) H. Hara, and iriomotensis Masam. Although the presence of a relatively high percentage of individuals with these features in in- sular East Asia might prompt the question of why such a pattern exists, the individuals themselves merely possess extremes of completely continuous characters that can be found in other parts of Asia. We therefore do not formally recognize such plants. i and Hwang (1987) maintained Styrax grandiflorus as a species distinct from 5. Perkins ( japonicus by its densely pubescent pedicel and ca- lyx and apparently its general range, extending far- ther south than S. japonicus. The leaves of S. gran- diflorus also tend to be elliptic (vs. subrhombic) i but these features are apparently only weakly correlat- with an acute (vs. slightly acuminate) apex, ed with pedicel and calyx pubescence from south to north. Several specimens from Japan, Korea, and Shandong Province, China, possess a densely pu- bescent calyx, whereas specimens with glabrous pedicels occur sporadically throughout southern China (e.g.. Yunnan Province and Hainan Island). Furthermore, many collections exhibit an interme- diate amount of pubescence. The pubescent phase also exhibits no obvious elevation or habit distinc- tions and seems to occur sporadically, often near collection localities of the glabrous phase. A similar pattern of variation in pubescence density unac- companied by geographic or ecological separation occurs in S. hemsleyanus and S. hookeri, and those species have not been subdivided in this revision. For the same reason, we have subsumed S. gran- diflorus under S. japonicus. The distribution of Styrax japonicus exhibits some notable patterns. No specimens from main- land China have been collected south of the Nan- ling Mountains, which extend along the northern border of Guangxi and Guangdong Provinces, but the species has been collected south of the main- land on Hainan Island. Long-distance dispersal is not likely as an explanation for this distribution because the fruit of S. japonicus appears not to pos- sess high vagility (Fritsch, 1999). A more likely explanation is that intervening populations have gone extinct due to habitat changes (vicariance). 520 Annals of the Missouri Botanical Garden Hainan was connected with mainland China until the early Quaternary (Chang, 1962), and insofar as Hainan Island is considered part of the Guangdong floristic region (Chang, 1962), the appearance of the Qiongzhou Strait separating Hainan Island from the mainland seems not to have had a major influ- ence on the flora of Hainan. It is also possible that populations have become extirpated through human disturbance The disjunct distribution of Styrax japonicus be- tween Lanyu Island (Taiwan) and the Philippines is paralleled in about 110 other flowering plant spe- cies, suggesting that land connections between the two islands are likely to have existed previously (Chang, 1994). The flora of Lanyu Island appears to have greater similarity to the flora of the Phil- ippines than to Taiwan in that 46 genera not ap- pearing on Taiwan are shared by Lanyu and the Philippines (Chang, 1994). The Taiwan Strait may have first appeared in the late Mesozoic, after which Taiwan contacted Mainland China several times (Chang, 1994). Although the floras of Taiwan and the mainland share many species, Taiwan does possess some distinctive floristic characteristics. The absence of S. japonicus from Taiwan suggests that the evolution of Styrax has proceeded in iso- lation on this island. Styrax japonicus is very sim- ilar to the Taiwanese endemic species 5. formosan- us, differing mainly by imbricate (vs. valvate) corolla aestivation and a slightly longer pedicel. Phylogenetic analysis of DNA sequences of the ITS region (Fritsch, 2001) strongly suggests that 5. ja- ponicus and S. formosanus are sister species. Thus, it appears that S. formosanus on Taiwan has spe- ciated from 5. japonicus ancestral stock and that the imbricate-flowered species of Styrax do not con- stitute a clade (see Taxonomic History and Present Objectives). The locality of W. Griffith 3671 (the type of Sty- rax grandiflorus) is in the Naga Hills, either in the Sagaing Division of Myanmar or Nagaland, India. We could not determine the geographic coordinates of the specific localities mentioned in the proto- logue of this species (“Nempea” and "Namtuzceh") with sufficient precision to map them. The collec- tion appears to represent the westernmost locality of S. japonicus known. No specimens were cited in the protologue of Styrax japonicus. New species in volume 1 of Flora Japonica were described by J. G. Zuccarini based on data supplied by von Siebold. The only material that we have seen from the von Siebold herbarium consists of on-line images of two L collections from a database of the von Siebold collections main- National Herbarium Nederland tained by the (). We chose I. accession number 908240-682, as the lectotype because it has better flowering material than Z. Keiske 64 (accession number 908240-688). Furthermore, the /. Keiske s.n. collection bears in- Keiske s.n., sect galls of the same general type as those that appear on the illustration accompanying the pro- tologue, whereas /. Keiske 64 does not possess galls. The type material of Styrax Jippei-kawamurai Yanagita (J. Kawamura s.n.) is missing. Yanagita worked at the National Forestry Agency in Tokyo, the herbarium of which is now part of the herbar- Tokyo. None of Yanagita's specimens can be found in this herbar- ium of the Tama Forest Museum, ium or are known elsewhere (H. Ohba, TI, pers. comm. ). ected specimens examined. CHINA. Anhui: Huangshan Shi, Huang-shan-qu, Huang-shan, M. P. Po & Yao 79022 (A); Jinzhai Xian, Bai- -ma-zhai, G cai-gou, K. Yao 8965 (A, CAS, i peng- -dian, Z. W. Xue 830187 (IBSC); Xie & L. Zheng 97133 (CAS). Fujian: Xin- chun-xiang, Exp. Wi Gu-shan, Bai-yun- "dong, L. Xian, L. G. Lin 1406 (PE); B Xian, Xin-qiao-xiang, G. L. Cai 464 (IBSC, KUN). Gansu: Hui Xian, Fan-ba, Z. B. Wang 19392 (KUN); Kang Xian, Yang-ba-xiang, from Nao-hui-ba to Yang-ba, Z. Y. Zhang 16760 (PE); Wen Xian, Xiao-wan-li, Bi- feng- gou, Bi-kou, X. Wang 98 (MO). Guangdong: Heping Xian, G. C Zhang 35 (IBSC); Liannan Yaozu Zizhixian, Jin-keng- xiang, Р. C. Tam 59492 (IBSC, KUN, PE); Lianshan алм Yaozu Zizhixian, P. C. Tam 58283 (KUN); Ru- 'uan Yaozu Zizhixian, Xi-shan-xiang, Ba-bao-shan, Wang 44043 (IBSC, KUN, MO, PE); Shantou Shi, Wu- king-fu, 1906, J. M. Dalziel s.n. (E); Wengyuan Xian, Long-xian, X. Q. Liu 24221 (IBSC). Guangxi: Guanyang Xian, Dou-yan-lin, Z. Z. Chen 52458 (IBK, IBSC, KUN); Xian, Niu-wei, Ba-wang-shan, Exp. Hong- shui- he 89-1109 (KUN); Lingui Xian, Huang-sha-xiang, Chen 50983 (IBK, IBSC, KUN[2]); Lingyun Xian, Loe- Steward & H. C. Cheo 415 (A, BM); xian, San-men-xiang, D. A. Huang 60211 (IBK, IBSC); T Miaozu Zizhixian, Luo- dong- m Jiu-wan-da-shan, 5. Q. Chen 14442 (IBK, IBSC, ‚ РЕ); Xing’an Xian, Liang-jin-kuang-xiang, Mao- Z. Z. Chen 51257 (IBK, IBSC, KUN); Zi- yuan m Shuen-yuen, T. S. Tsoong 81668 (A). Guizh- ou: Anlong Xian, Long-shan-xiang, Exp. uerus 4481 CUN, PE). Bijie Shi, Bao-he-xiang, P. H. Yu 331 (KUN PE) Dushan Xian, Shui-li-guang-li-qu, Exp. 1i bo 1115 ) (KUN); Duyun Shi, Yun-fou-shan, Tuyun, Y. E d s (IBSC[3], PE); Е наана ling-shan, Л. Y (PE); Hae Xian, 521 (IBSC); рен i Xian. W ing ); Jiangkou Xian, Tai-ping "Hiver above confluence 1 Hei-wan sgh SE , side of р; ioe Bot. Exp. 274 (A, BR, S. РЕ); К jiang-xiang, Lei-gong- or and ыр S deus 2102 (KUN, PE); Leishan Xian, Z. P. Jian probe (KUN); Libo Xian, Jie-na, n H. Song 558 (K, MO); n Xian, J. Cavalerie 997 (A[2], E); Nayong Xian, Ju-ren-qu, Exp. iE 358 CUN, PE); Panxian Tequ, P. C. Tsoong 1740 (PE); Ping- e: S С hoh-tsuen, А Longsheng Gezu Zizhi ~ — 2 Volume 90, Number 4 2003 Huang et al. Revision of Styrax Series Cyrta tang Xian, Exp. S Guizhou 2745 (PE); Pu'an Xian, Qing- shan- -xiang, Exp. An-shun 1353 (KUN, PE); Qingzhen Shi, un-gui-shan, Zhu-sha-dong, Exp. Sichuan & Guizhou 1860 (PE); Rongjiang Xian, Yue-liang-shan, Exp. S Guizh- ou 2902 (PE); eus Xian, Ma-xi, tid -cun, Exp. Wu- ling-shan 2598 (KUN); Shiqian Xian, Fu-yan, Mai-zi- "ys Exp. Wu-ling-shan 1989 (KUN); Shuisheas Xim, P. С. Tsoong 1786 (PE); Songtao Miaozu Zizhixian, Gao-diao- xiang, Huang-tang-ping, Exp. Wu-ling-shan 616 (KUN); y Steward et „К, PE[2]); Tongzi Xon Tien- chu-tze, Tungtze, y bow 5004 (PE); Wengan Xian, Yong- he- xiang, Exp. Li-bo 2240 (KUN); Xingyi Shi, Ba-ling-xiang, Exp. “Guizhou 7361 (IBK, PE); Xishui Xian, Guan-du-qu, Exp. Bi-jie 1491 (PE); Yinjiang Tujiazu Miaozu Zizhixian, Su-jia-po, Xiao-jia-he, e P. Jian 31437 (PE); Zhenning Buyizu Miaozu Zizhixian, Tschenning-Huang-tsauba- Yun- nan, H. F. Handel- Ма 10310 (А, C, E); Zunyi Shi, Liang Feng Yah, A. N. Steward et al. 137 (A, BM, E, Т. PE). inan: Baisha Lizu Zizhixian, iot -men, Exp. - Hainan ip (IBSC[2]); Chengmai Xian, Bai-shi- SE Gu- dong-cun, C. I. Lei 376 (A, IBSC[2], d UC Ж Gone: zhong L М ee Zizhixian, Hong-mao-shan Т Tsang & H. Fung 491 (BM, IBSC, PE). Henan: Baofone Xian, Pu-cha-biao- I 18727 (PE); Lushi Xian, from Da-quai- di to Qi-he, J. Q. Fu 2210 (KUN); Neixiang Xian, Bao- tian-man Nature Reserve, Da-hong-si River, D. E. Bouf- ford et 26287 (AAU, Е); Shange hend Xian, Pu-cha- [тше ben 10363 (PE); Song Xian, Hong-luo-he, К. J. Guan et al. 1905 (PE[2]); Tongbai Xian, Fu-niu-shan, Henan Forestry m 59 (PE); Weihui Shi, gu chi, cha- ae ben 34393 (PE); Xin Xian, Wu-ma, Pu-cha- ben 8269 (PE); Xingyang Shi, Ji-gong-shan, es USSR team pos (PE); Xixia жип, се . Guan et а 1405 (РЕ[2]); i Pu-cha- © беп 6239 PE). Hubei: Badong. Xian, A. Henry 1430 (K); Baokang Xian, E. H. Wilson 2134 (К); Changyang Tujiazu cart ian, Huo-jia-ping, T. P. Wang 11375 (PE); Enshi Shi, wan-chang, L. Y. Dai & C. H. Qian 616 (PE); d Xian, H. J. Li 5516 (KUN); Jianshi Xian, Hua- s ping, D W. B. Lin 70 (PE); Lichuan Shi, Shui-shan- ba, Y ang-he- . W. C. Cheng & C. T. Hwa 559 (A, PE, UC); Shen- nongjia Linqu, Shen- -nong-jia Forest Dist., NE of Guan- men-shan along the S side of the Shi-cao River, Sino-Amer. Bot. Exp. (1980) 763 (А, E, KUN, UC); Songzi Xian, Mo-pan-zhou, Père C. peu 17704 (A); Wufeng Tujiazu Zizhixian, H. J. Li 6802 (KUN, PE); Xianfeng Xian, Qing-shui-kuang-qu, W. Pi Lin 575 (PE); Xingshan ar M Hing-shan, H. J. Li 1064 (PE); Yich- Shi, T'O, A. Henry 3926 (K); m Xian, Wu- end K R. Liu 142 (PE); Zhuxi Xian, K. M. Liou 8776 (PE); Zigui Xian, H. J. Li 318 (PE). Hunan: Baojing ‚ X. L. Yu 91440 (КОМ); Changsha Shi, collector un- nown 27495 (PE); Chengbu Miaozu Zizhixian, Jin- кап sha . Z. Lin 11145 (IBSC); Cili Xian, Suo-xi-luo N ture y ein Exp. W Hunan 1087 (PE Lan-zhu-ping, P. C. 2 61. 328 (IBK, Xian, Zhang-jia-ba, Z Xian, X. D. Yun 104 (IBSC); Fenghuang Xian, Yong-shui, Exp. Hunan 614 (PE); Hengshan Xian, Guang-ji-shi, P. C. Tam 63944 (IBK, IBSC); Jianghua Yaozu Zizhixian, Hæluo-kou-xiang, B. G. Li 5149 (PE); Longshan Xian, u-ya-xiang, L. H. Liu 1877 (KUN); Ningyuan Xian, Jiu- wan-shan, P. C. Tam 61690 (IBK); Sangzhi Xian, Ba-mao- xi-xiang, Tian-ping-shan, B. G. Li 750013 (PE); Wee Shi, P. C. Tam 64023 (IBK); Wugang Shi, Yun-shan, P. C Tsoong 1241 (PE); Xinhuang Dongzu Zizhixian, Lien Exp. Hunan 281 (PE); Xinning Xian, Shun-huang- shan, Q. Z. Lin wy (IBSC); Yizhang Xian, Mang-shan, quan- xiang, P. H. Liang 63552 (IBK, MO); Yongshun ian, Xiao- Bey om X Yu 91655 (KUN); 7 Um Zizhixian, Nan-mu- ping, ا‎ unknown 490 (KUN). Jiangsu: Ganyu Xian, Liu-li ar Hai- chow f Hers H636 (A); Lianyungang Shi. Yun-tai-shan K. Yao 8497 (MO); Yixing Shi, R. C. Ching 4825 (K). pea 7 Xian, hi ee жашы d S. Yue 3551 (IBSC, Dayu Xian, Zuo-bo- e et al. 611 DR IBsC. KUN); ques Shi, Deng shan, J. Xiong 2349 (PE); Jiujiang е Lu-shan, А Lan col. Lianhua Xian, Wu-gong-shan, ion g. Exp. Jiangxi 377 (PE); Dada. Shi. Da-fen-qu, 7 n Yue pris (PE); Nankang Xian, Fu-shi-xiang, M. Q. Nie et al. 9797 (KUN); е epe Da-long-xiang, 5. S. Lai 5182 yia KUN); Taihe Xian, S. 5. Lai 558 (I Tonggu Xian, Long-men, S. S. Lai prs (PE); Wuning Xian, r -ping- xiang, S. S. Lai 2695 dope PE); Wuyuan Xian ‘hing 3273 (A, E, K, UC); Xunwu Xian, Jian-xi- xiang, Bi-jia-shan, J. S. Yue ке (IBSC, KUN, PE); Yit- Huang- gang- xia =) т rod uping "n Yue-ba- der ера gou (KUN); ter di Xian, Caan, S. B. He 614 (KUN); Shangzhou Shi, Si-ji-he, P. Y. Li 8461 (KUN); Shanyang Xian, Xiao-he-kou-xiang, Hei-gou-da-dui, Z. Y. Zhang 15926 (PE); Shiquan Xian, Liang-he-xiang, J. Q. Xing 8048 (IBK); Xixiang Xian, Xia-guan-kou, Lao- cheng, J. О. Xing 1843 (PE); Yang Xian, Hua-yang, К. Т. Fu 5240 (IBK, PE); Zhenping Xian, P. Y. Li 2692 (KUN, PE); Ziyang Xian, Feng- du uo- dia an, P. C. Kuo 2180 (PE). ioi ‚ Shang-guo-dui, T. Ww Liu à u Xian, Tien-pa-ho, W. P. Fang 10307 (А, DS, Е, IBSC, PE[2]); Chongqing Shi, Bei- pei-qu, Jin-yun-shan, Exp. Sichuan & Guizhou 192 (PE); Da Xian, Sui- ting -fu, W. P. Fang 10249 (BM, Rd PE); Dujiangyan < та м. Nan-yue & Lu-zi-tang, D. Е. Bouf- ford & E (aw stia 24853 (A. AAU, CAS, L, MO); Ebian Yizu Zizhixian, Sha-ping, Z. S. Zheng 230 (KUN): Emeishan Shi, E-mei-shan, 2 il Yang 55688 (IBSC, PE); Fengjie Xian, Zhu-yuan-xiang, Z. R. Zhang 25586 (IBSC, KUN, PE); Hanyuan Xian, Sichuan Economic Pl. Exp. 1013 (KUN); Hechuan Shi, X. L. Sun 5597 (PE); Jianyang Shi, Hong-jia-yan-he, Sichuan Economic Pl. Exp. 2226 KUN); Leibo Xian, Xi-ning-xiang, Q. S. Zhao 428 (PE); Li Xian, Suo-luo-gou, R. Li 46764 (IBSC); Mabian Yizu Zizhixian, Е T. Wang 22866 eh KUN, PE[3]); Nan- chuan Shi, Jin- fo-s an, С. F Li 1 (IBSC, KUN, PE); => Xian, Yong-xing-qu, D. Y. Peng 45496 (IBSC); Shi, Mou-tao-chi, Ma-hwang-au, C. T. Hwa 16 (PE): Wan- yuan Shi, K. L. Chu 2179 (PE); Wushan Xian, Dang-yang- xiang, G. H. Yang 59092 (IBSC, KUN, PE); Wuxi Xia . Taiwan: Taitung Xian, Lanyu Island, W slope of Hung T'ou-shan, W. L. Wagner 6721 (CAS, MO). Yunnan: Dali Shi, T N. Bos 16474 (IBSC[2], KUN); Eshan Yizu Zizhixian, Exp. E-shan 88441 (KUN); Funing Xian, Jar- 522 Annals of the Missouri Botanical Garden gei, C. W. Wang 89592 (IBSC, KUN, PE); Fuyuan Xian, Huang-ni- he, Exp. Hong-stiui- he 2943 (KUN); Ge ngma Daizu Mice Zizhixian, C. W. We ang 72938 (^, IBSC, KUN PE[2]; Guangnan Xian, Mao-yi-xiang, Q. A. Wa 9740 (KUN); Jiangcheng Hanizu Yizu Zizhixian, Y. H. Li 5408 (KUN); Jingdong Yizu Zizhixian, Me onus "o meng- lung, C. W. Wang 78470 (^, IBSC, KUN, PE); Jinghong Shi, Meng-soong, Dah-meng-lung, C. = dg 78470 (A, IBSC, KUN, РЕ|2]); p Miaozu Yaozu Daizu Zizhix- ian, i-e Fu lin- -qu, В. Y. Qiu 57007 (KUN); Kunm- Shi, i an, К. M. juna 10406 (KUN, PE); Longch- s S. Yang 8311 (KUN); Longling Xian, H. T. Tsai fpes (A, B, IBSC, KUN, PE); Lüchun Xian, Fen-shui-ling, Lei-bo Valley, D. D. Tao 238 (IBSC KUN[2]); Lufeng Xian, W of Lufeng City, Sino-Amer. Bot. Exp. (1984) 1307 (A, CAS, KUN); Luxi Xian, Lo-shiueh- shan, H. D. McLaren U219 (С, E); Malipo Xian, Chung- dzai, К. M. Feng 12740 (A, KUN, PE[2]; Mengzi Xian, Yang-cao-tang-xiang, Y. Y. Hu 580574 (KUN); Nanjian Yizu Zizhixian, А Ducloux 2716 (Р); Pingbian a Zi- . Liang-zi-xiang, Yao-shan-qu, P Mao 4154 1 i shi, тшй Exp. Hong-shui-he 2065 (KUN); Shuangjiang Lahuzu Wazu Bu- langzu Daizu Zizhixian, from Shuang-jiang to Tai-ping- xiang, J. S. Xing ЖО Жар Suijiang Xian, Mo- dao-xi, B. S. Sun 141 (IBSC, KUN); Tengchong Xian, Dong-shan-xiang, on ai- v. H. Li 11. 357 (€ AS); Wen- shan Xian, Lao-jun-shan, M. Feng 11082 (A, KUN, PE); Xinping Yizu Daizu saan Mao-er-shan, Exp. Yu- xi 2992 (KUN); Xundian Huizu Yizu Zizhixian, Hay tien, F. Duc i qu (P); Yiliang e from Cao-tian-ma to Niu-jie, NE Yunnan 905 (KUN); Yingjiang Xian, C. D. Tao 13063 (KUN); اا‎ B ue Daizu Zi- і He-ping-shui-ku, 739 (KUN); Yuanyang Kian, S. C. Ho 85196 (IBSC ' Zhanyi Xian, Xiao-ma-la, Y. H. Li 148 (KUN, s | Zhenxiong Xian, Shi- jia-wan, P. H. Yu 1096 (IBSC, М, PE). Zhejiang: Jin- yun Xian, Yan-ling-keng, Wen-yang, S. Y Chang 1740 MO); Kaihua Gai. Gu-tian-miao, г X. Wang 2123 (РЕ); Longquan Shi, Feng-vang-shan, 5. Y. Chang 3319 (MO): Rui’an Shi, Shi-yang, 5. Y. Zhang ТК (MO, PE); Sui- chang Xian, Shui- chang: Zhejiang Bot. Res. Team 25807 (MO, РЕ); Taishun Xian, Wu-ling-van, 5. Y. 7 (KUN, PE); Wuyi Xian, Xi-lian-x Fang, Z. J8311260 (IBSC ). JAPAN. Honshu: ; ; Aomori Pref., gun осе 5 Yosioka 23 (Т1): ikawa-gun, Ishikawa-cho, H. Iketani 1117 (МО); Cin f., Mizunami-shi, Matsuno-ko, S. Tsugaru et al. 23572 (KYO): Pref., Usui-gun, Malsuida-machi, J. Murata & T. T. Chen 7672 (TI); Hiroshima Pref., Saeki-gun, Yuki- ida Horus -gawa, 1979, (AAU); B Pref., Mt. Rokko, Н. Mürai 3092 (A); Ish- , Enuma-gun, di H. Muroi 2243 (A); Iwate Iwayama, H. Muroi 5010 ( gawa Pref., . Yokohama, 16251 (BR); Kyoto Pref., Funai-gun, Wachi-cho, Mt. Cho- ro-ga-dake, S. Tsugaru et al. 18431 (МО); Г sushi-gun, M. Hiroe 16424 (UC); piss Pref., ma-cho. side of Mt. Otakamori, E. Wood & D. E. Boufford 3967 (A, CAS); Nagano Pref., ut hi-kumagun, Jkuwa-mura, betw. Noziri & Mt. Aterayama, along the River Ayera-gawa, G. Murata & H. Nishimura 122 (AA U, KYO); Nagasaki Pref., Tsushima Island, Shimoagata-gun. Mitsushima-cho, Sumo, K. Mimoro 1840 (MO); Nara Pref., Usui- Nara-shi, Ninnikusen-cho, H. [ketani та 56 (MO); Nigata ref., Morimachi, Minami-Kambara, Y. Ikegami 17502 (A); Okayama Pref., Maniwa-gun, M. Hos 16409 (UC); Osaka Pref Fukuoka 5852 . Kawac hinagano City М. (AAU, С, E, K, L, UC); Saitama Pref., г nr -gun, Niiza- machi, Heirinzi, 1966, H. Ohashi s.n. Shiga Pref., Taka: es AR Makino-m nac a Minarnimakino, H. Ohashi et al. 8 3 (А); S 1-gun, Shi- S ~ RS >t @ —. 810 (P); Tokyo Pref., p oj. IT 9714 (A, AAU, C, К, ). PE): Mina Pref., Mt. Tonami in town of Lun zi s 360 (MO); Wa- kayama Pref., Kii, G. Koidzumi s.n. (KYO[2]): Ee Pref., Mer um igh Inosawa, H. Ohashi et al. 10779 (A): — ie . Mukaidoi, Tukuyama, H. Migo s.n. (A). : Kagoshima Pref., Phsumi, Yaku-shima Island, rw: ү Miyoshi 10778 (К); 3 eps Pref., Futae Pass, 1983, Y. Endo s.n. (M Toi, K. Kondo n Koyasan, Okinawa Island, Kunigami. Nago-dake, т. el a 6157 (E, GH, K IC). Shikoku: eee Pref., Mitoyo- gun, Toyono- mura, n Takahashi 1197 (A, KUN, PE) uaphan: Muang awin, Clueng Kwang. A. FC Kerr 20941 (BM. K, L). Xiangkhoang: betw. Muong Ham & Ta Thom, J. Vidal 880B (P). МҮАММ/ i State: N Triangle (Hkunlum), £ F K. Ward 20632 (A, BM, E). NORTH KOREA. Pyongyang City: Chonbuk, Chonju, Wansan Chilbong, up B. Y. Sun s.n. dA HIL IPPINES. Babuyan Islands: Camiguin Island, n absing, G. E. Edaño 79248 (L). Batan "yet Daan Island, Mt. Matarem, M. Ramos 80424 (BO, L). SOUTH KOREA. Cheju: ! Mein. sju-gun, Shinye-ri, D. А. ud ford et al. 25729 (C E). Kangwon: Koge n, E. H. Wil- d 9328 (A). b Keiki, Kosyo Ё. H. b 87. 54 A, K). North Cholla: rium m Moak san, D. E. Bouf- ford et al. 25808 (CAS, s North Kyongsang: Dae жалаш, 1987, Y. S ae 1. (A, MO). Seoul City: e Moran 4327 (BM, BR, r GH, MO, UC). Sout : Mt. Moodung, 1984, Kim s.n. (A). South C FEER hong Sosan Gun, жы» sland, В. R. Yinger et al. 2525 (А). South Kyongsang: тта Keisyonando Chosen Nippon, А. Uno 23243 (A). VIET- NAM. Сао Ва ng: Mt. Pia Oac, 1997, U. URN s.n. (CAS). 7. Styrax limprichtii Lingelsh. & Borza, Repert. Nov. Regni Veg. 13: 386. June 1914 [as S. "Limprichtii^|. TYPE: China. Yunnan: Chuxiong Shi, Tschu-hsiung-fu, 2000 m, 24 Aug. 1913, К. G. Limpricht 920 (lectotype, designated here, WRSL not seen; photo of lec- Spec. totype, А!, PE!; isotype, A!). Styrax langkongensis W. W. Notes Roy. Bot. Gard. Edinburgh 8: 208. Se ir 1914. TYPE: China. unnan: xian unknown, hills at the S end of the Lang- En Valley 2121-2727 m, May 1910, е orrest 558 дошу, E! isotypes, BM[2]!. TBS, Кр "PE UC). Shrubs to 2.5 m tall. Young twigs gray-yellow or yellow-brown stellate-tomentose; older twigs dark Volume 90, Number 4 2003 Huang et al. 523 Revision of Styrax Series Cyrta purple, glabrescent. Petiole 1—3 mm long. Two most proximal leaves on each shoot alternate or more often subopposite to opposite. Lamina 3.5—7(—9.5) 5 cm, chartaceous, elliptic to obovate; apex obtuse to slightly acuminate; base rounded to broadly cuneate; adaxially densely stellate-pubes- cent when young, becoming sparsely pubescent; abaxially white stellate-tomentose, rarely subgla- brous, often with additional scattered orange or dark brown stellate pubescence especially preva- lent on veins and the two most proximal leaves on each shoot; margin serrate or nearly entire (but still glandular), often irregularly dentate apically; sec- ondary veins 5 or 6 on each side of midvein, ter- tiary veins reticulate, plane or slightly sunken, ab- axial surface of the secondary and tertiary veins obscured by the tomentum, only the tertiaries abax- ially prominent and raised in young leaves. Fertile shoots 1-7 ст long, 3- to 5-leaved. Inflorescences arising from shoots of the current growing season; lateral inflorescences 1(2)-flowered; pseudoterminal inflorescences 2-flowered or racemose, 1—2 long, 2- or 3(4)-flowered, rachis yellow or orange stellate-tomentose, stalked trichomes absent. icel 3-4 mm long, densely pubescent; bracteoles 3-5 mm long, subulate, positioned at various places along the pedicel but mostly near the middle, more rarely near the base, sometimes those toward the base of the inflorescence leaf-like. Flowers 1.5-2.0 cm long. Calyx 5-6 X 5—6 mm, cupuliform; adax- ially sparsely appressed-pubescent with white 2- or 3-armed or stellate trichomes, becoming glabrous proximally; abaxially yellow-brown or orange stel- late-tomentose throughout, often also with various amounts of larger scattered orange or brown stiff stellate trichomes, especially proximally; margin distinctly dentate, the teeth (0.6—)1—1.5(-2) mm long, subulate to deltoid, unequal, usually contig- uous or separated by a shallow concave portion. Corolla 1.0—1.4 cm long, white, tube ca. 4 mm long, glabrous, lobes 5, 9-11 4—6 mm, elliptic to ovate-elliptic, short-stellate-pubescent on both sides. Stamens 10; filaments 5—6 mm long, straight, proximally broadened, densely white stellate-pu- bescent, trichomes up to 0.5 mm long, distally su- bulate-attenuate and glabrous; anthers 4—5 mm long, wider than distal portion of filament; connec- tives glabrous. Style proximally stellate-pubescent and distally glabrous, sometimes sparsely stellate- pubescent or glabrous throughout; stigma 0.2—0.4 mm wide, capitate. Fruit 1.4—1.6 X 1—1.5 ст, glo- bose, apex rounded or apiculate, dehiscent by 3 valves from apex; pericarp dry, 0.3-0.6 mm thick, outside regularly longitudinally striate throughout, rugose, gray stellate-tomentose, inside glabrous or cm minutely downy-pubescent. Seeds brown, ovoid, finely reticulate-fissured, glabrous. Cormophyt. Sin. Illustrations. Anonymous, lc. S. M. Hwang, Fl. Reipubl. Popularis Sin. 60(2): 90 pl. 30 (1—5). 1987; W. Q. Yin in Y. C. Xu, Ic. Arbor. Yunnan. 2: 892, pl. 470 (7-10). 1990; Z. Y. [C. Y.] Wu & P. H. Raven, Fl. China Ill. 15: 199, fig. 199 (1—5). 2000. Phenology. | Flowering: February-October. Fruiting: April, June- November. Distribution. China (Sichuan and Yunnan); Fig- ure 8. Habitat. In relatively sunny, dry stony pastures, more often in forests on open rocky slopes; 1400— 271 Vernacular na Chu-xiong-an-xi-xian mes. -n 1980), Chu-xiong-ye-mo-li (Anonymous, 974). Styrax limprichtii, a much-branched shrub that rarely exceeds 2.5 m, occurs only in northeastern Yunnan Province and adjacent southwestern Sich- uan Province. The ranges of Styrax limprichtii and S. rugosus are contiguous. These two species are morphologi- cally similar in many respects, as demonstrated by their adjacency in the key, and are probably sister taxa. They nonetheless exhibit enough differences throughout their ranges to justify the recognition of these two entities as species (see couplet 24 of the key). Furthermore, S. rugosus usually occurs at low- r elevations (7 m) than S. limprichtii 1400-2750 m). Besides these differences, S. lim- prichtii differs from S. rugosus by a tendency toward shorter calyx teeth, shorter bracteoles, shorter shoots, and less rugose leaves. These characters, however, exhibit some degree of overlap. Also sim- ilar to S. limprichtii is S. wilsonii (see comparisons under that species). Styrax limprichtii, S. rugosus, S. wilsonii, and the western Mexican endemic S. jaliscanus are all sim- ilar morphologically, sharing a shrubby habit, a ca- lyx that is densely pubescent throughout abaxially and distinctly dentate with the teeth contiguous or nearly so, a petiole < 5 mm long, and an evenly The striate pericarp ~ longitudinal-striate pericarp. appears to be restricted to these species within Sty- rax. Phylogenetic analysis of DNA sequences of the ITS region, however, places S. limprichtii at the base of series Cyrta, and S. jaliscanus groups with the rest of the North American species of series Cyrta (Fritsch, 2001). Nonetheless, support for these positions is weak, and this group is in need Annals of the Missouri Botanical Garden [110 = Styrax limprichtii | * Styrax obassia Kilometers Figure 8. of more detailed study to resolve the apparent dis- crepancy between molecular and morphological data. Only two imbricate-flowered species of Styrax are sympatric with S. /imprichtii. Styrax hookeri is normally easily distinguished from S. limprichtii by its tree (vs. shrub) habit. Styrax limprichtii can be further distinguished from small individuals of S. hookeri by its distinctly dentate (vs. usually trun- cate or undulate) calyx, which is + pubescent throughout (vs. more sparsely pubescent within mm of the margin). Styrax limprichtii also tends toward leaves with acute to blunt (vs. acuminate to caudate) apices, coarsely serrate margins, espe- cially distally (vs. more finely serrate), and a rugose — vs. smooth) surface abaxially. Moreover, S. limpri- chtii also usually has densely pubescent laminae abaxially, and anthesis often occurs before the full expansion of the leaves, whereas the leaves of S. hookeri can be densely pubescent to glabrous abax- ially and anthesis occurs at the same time as or after full leaf expansion. The only other species of imbricate-flowered Styrax sympatric with S. limpri- chtii is S. japonicus, occurring near Dali Shi and Yunnan Province. Kunming Shi, This species is Geographic distribution of Styrax limprichtii and S. obassia. readily distinguished from S. limprichtü by its lon- ger pedicels and nonstriate (vs. longitudinally stri- te) pericarp. The collections of K. G. Limpricht numbers 896, 920, and 973 are all cited in the protologue of 5. limprichtii without a clear indication of type, and thus are all syntypes of this name. We have chosen the К. С. Limpricht 920 specimen at WRSL as the lectotype because it is apparently the most widely distributed of the three collections. Selected spec Hag examined. CHINA. Sichuan: Muli Zangzu Zizhixia Forrest 22394 (A, E, K); Luo-bo- xiang, Nan-shui- hs ls a 5709 (KU N, PE); Mu-li Val- ley, mtns. betw. Mu-li & Ku-lu, J. F Rock 24150 (A, E, JC); from Ke-tze to Ku-ba-dian, 7! T. Үй 7216 (A, KUN, p. from Tuo-li-gou to Ke-tze, T. T. Yü ded (A, KUN PE); Panzhihua Shi, Da-bao- a г Qinghai & Xizang 11362 (KUN); Yanyuan Xiar F. Handel- Изин. 2068 (A, E) Nan- к bei-diao- dui 5564 (PE) Wei-luo-he, Mao-niu-shan, Nan-shui-bei-diao-dui 5967 (KUN, PE) Yunnan: Binc us Xian, Hi-zu-sha : Sin-tien, Pin-tchouan, А Ducloux 4627 (P); Ji-zhu-shan, H. Li 503 х №); Gan-dian, Y. Q. Lin 11 (KUN); from Xia-vang to xi, T. №. Liou 21495 (IBSC, PE); Niu-jing, T. N. Liou 21688 (IBSC, PE); Simeon Yen, 5. Ten 351 (E, UC); Ji-shan, H. C. Wang 1988 (KUN, IBSC, PE); Chux- iong Shi, Guang-ba-he Reservoir, S. C. Huang 20 (KUN); — —. — - Volume 90, Number 4 2003 Huang et al. 9 525 Revision of Styrax Series Супа d E Sino- Amer. Bot. к. (1984) 1256 (A, CAS, y W slopes of the Sung-kuiei Range. G. Forrest 23057 (A, E, K); Xiao- ix shan, Erhai Sino-British Exp. Cang-shan 1 (A, KUN), Y. re. 11337 (IBSC, KUN); Wu-tai-feng, X ee м 1740 (IBSC[2]. KUN[2]); Yang-tze divide, Е i na li lake, J. K. Ward 3831 (E); Dayao Xian, Shi-yang-qu, Coll. Team for Oil Pl. (1965) 650302 (KUN); Kang-jia- shan, P. Di 60022 (KUN); Eryuan a San-ying-qu, Jiao- shi-he, Exp. NW toa 6389 (KUN, PE); Heqing Xian, Bai-yan, R. C. g 24523 (KUN, PE): from Sarchatze to C hiang- Ing near ‘Sune ^as K. M. Feng 801 (A, KUN); Dsolin-ho, Н. F. Handel-Mazzetti 6224 (A, Е); Ti chiang Range, H. D. rao 5114 (BM); Jsu-yung, H. D. McLaren 1 14F(AA) (C, E); Pai-ching, H. D. McLaren F199 AAU, E); Sung-kuei, H. D. McLaren e233 (E); He-chuan- xiang, W. C. Wang 390 (KUN); Kunming Shi, Yi-ping- lang, T. N. Liou 16614 (IBSC[2], PE); Lanping cen Pum- izu Zizhixian, in the Lang-kong Valley, G. Forrest 9954. (BM, E, K, PE, UC); Lijiang Naxizu Zizhixian, T ze-li on Yang-tze River, R. C. Ching 20264 (A, KUN, PE); Tai- ngo- koo, R. C. Ching 21670 (KUN); Tze-li on Yang-tze, . Ching 22139 (A); Shu-di-du-kou, Exp. Qinghai & p 638 (KUN[2]); near Jin-sha-jiang, Exp. SW China (Guizhou, Sichuan & Yunnan) 200 (PE); Tuscus -xiang, Bai- shui-he, K. M. Feng 21567 (KUN, PE); Li-chiang Range. H. D. McLaren 46No.2 (BM); gi рее betw. Ta-li-fu & Li-kiang, J. F Rock A (A, E, UC); betw. Li-kiang & Ta-li-fu, J. Е Rock 6397 ( A, UC); E Li- jet shan, Sung- kwe-ho-chin Range, J. E Rock 6268 (A, UC); betw. Li- kiang, Tung-shan, Tui-nao-ko, & Tsi-li- En , J. F Rock 8520 (А, UC); Yulung-shan, C. K. Schneider 3965 (IBSC); E cong -qu, H. Sun 771038 (KUN); C. Y. Zhao 21670 ‚ PE); Shi-er-lan-gan-ban-shan, Y. X. Zhao 22139 Eu 1 Ninglang Yizu Zizhixian, from Ku-ba-dian to Tuo- li-gou, T. T. Yü 7309 (A, KUN, PE); Yangbi Yizu Zizhix- i i i i F. Rock T T bres zhou, Sichuan & aipe s.n. (PE); Yao-chou, T McLaren 205F (C); Yongsheng Xian, Song- ab Eo Qinghai & Xiang 692 ip: Jong- ues ect 16929 (E); Xin-liang-gong-she, C. X. Lü 2168 (KUN); Yunlong Xian, Jin-yue-liang, Xiang-liao- en (Coll. Team for Perfume Pl.) 156 (KUN); Zhongdian Xian, mtns. NE of Yang-tze Bend, G. Forrest 10696 (A, BM, PE, UC); Chung-tien plateau, G. Forrest 12653 (AAU, BM, E, PE). 8. ку заяаг W. C. Cheng, Contr. Biol. . Assoc. Advancem. Sci., Sect. Bot. 10: 242. 38 [as S. "macrocarpa"|. TYPE: China. Hunan: Yizhang Xian, Mang-shan, 800 21 Aug. 1937, W. C. Cheng 7000 [from protologue| (holotype, PE!; isotype, PE! [no collection number indicated on either sheet |). Styrax shejiangensis S. M. Hwang & L. L. Yu, Acta Bot. o Sin. 1: 75. 1983. [eds China. Zhejiang: Jiande Xian, 27 June 1958, Y. Y. Ho 29344 бош type, IBSC!; isotype, IBSC!). Shrubs to 2 m tall or trees to 9 m tall. Young twigs densely gray-brown stellate-pubescent, older Мм twigs becoming gray, glabrescent. Petiole < 1(-2.5 mm long. Two most proximal leaves on each shoot subopposite to opposite. Lamina 2.5-17 X 2-7.5 cm, chartaceous, elliptic to obovate-elliptic; apex acute; base cuneate, broadly cuneate or rounded; sparsely stellate-pubescent on veins when young, otherwise glabrous; margin subentire or apically slightly serrate, secondary veins 6 to 10 on each side of midvein; tertiary veins subparallel, adaxially plane or slightly sunken, abaxially raised. Pedicel 7—12 mm long, white stellate-tomentose; bracteoles 3—5 mm long, ovate-lanceolate, positioned at the base or middle part of pedicel. Flowers 2.3-3.2 ст long. solitary, arising only laterally from shoots of the previous growing season, opening before the leaves. Calyx 5—7 X 7—9 mm, cupuliform; adaxially glabrous; abaxially gray stellate-tomentose, within | mm from the margin more sparsely pubescent or glabrous, somewhat scarious, brown when dry; mar- gin with 4 to 6 broadly deltoid teeth, subglabrous on both sides. Corolla 1.6-2.6 cm long, white, tube 3—4 mm long, glabrous, lobes 5-7, 1.6-2.3 X 0.8- 1.1 cm, elliptic or narrowly elliptic, apex obtuse to acute, sparsely white stellate-pubescent on both sides. Stamens 10 to 12; filaments 8—10 mm long, straight, proximally broadened and ventrally dense- ly white stellate-villous, distally attenuate and gla- brous; anthers 5—6 mm long, wider than distal por- tion. of Style proximally sparsely white stellate-pubescent prox- imally, distally glabrous; stigma 0.2-0.4 mm wide, punctiform. Fruit 1.8-3 X 1.0-2.5 ст, ovoid to pyriform, apex rounded or apiculate, p in- m thick, out- side smooth, gray or pale вме cd. filament; connectives glabrous. dehiscent; pericarp dry, inside densely appressed-pubescent with long sim- ple. 2-armed, or stellate white trichomes. Seeds brown or dark brown, ellipsoid to ovoid-ellipsoid, glabrous, sometimes sparsely irregularly rugose, white stellate-villous. Illustrations. W. C. Cheng, Contr. Biol. Lab. Chin. Assoc. Advancem. Sci., Sect. Bot. 10: 243, Ad 25. 1938; S. M. Hwang, Acta Bot. Austro Sin. 1: 76, fig. 1. 1983 (as S. zhejiangensis); S. M. Hwang & C. J. Qi in W. C. Cheng, Sylva Sin. 2: 1618, fig. 811. 1985; S. M. Hwang, Fl. Reipubl. Popularis Sin. 60(2): 98, pl. 33 (1-2). 1987; S. M Hwang, Fl. Reipubl. Popularis Sin. 60(2): 98, pl. 33 (3—5). 1987 (as S. zhejiangensis); Z. Y. |С. Y.] Wu & P. H. Raven, Fl. China Ill. 15: 202, fig. 202 (1—5). 2000 (3—5 as S. zhejiangensis). Phenology. Flowering: April, May. Fruiting: June, July, September, October. Distribution. China (Guangdong, Hunan, and Zhejiang); Figure 2. 526 Annals of th Missouri Botanical Garden Habitat. In shady mountain forests, and along er, ovoid fruit. We regard the tertiary veins of S. the margins of mixed forests in ravines; 570—950 m. Vernacular names. Da-guo-an-xi-xiang (Hwang, 1980), Zhe-jiang-an-xi-xiang (Hwang, 1983). Styrax macrocarpus, occurring in mountainous regions between 570 and 950 m elevation, has been collected only rarely. This species is easily distin- guished from sympatric species of Styrax by its much shorter petioles (less than 1(-2.5) mm long); solitary, lateral, and larger (2.3—3.2 cm long) flow- ers that open before the leaves on shoots of the previous year; and a larger (1.0-2.5 em wide) fruit, with a pericarp (1—)1.5—3 mm thick that is densely appressed-pubescent inside with long simple or 2- armed white trichomes. Hwang (1983) described Styrax zhejiangensis from a single fruiting specimen collected in Jiande Xian, Zhejiang Province. Although she considered a possible relationship of S. zhejiangensis with S. macrocarpus based on their common solitary, lat- erally produced flowers, Hwang (1983) felt that the combination of a smaller, pyriform (vs. ovoid) fruit, broadly elliptic to ovate-oblong (vs. elliptic to ob- ovate-elliptic) leaves, and sparsely stellate-pubes- cent (vs. glabrous) seeds was sufficient justification for the recognition of a new species. There is, how- ever, ample evidence for a close affinity between the two taxa, so much so that we regard the two as conspecific. In addition to their solitary, laterally produced flowers, S. macrocarpus and S. zhejian- gensis share many other features such as the two most proximal leaves on each shoot subopposite to opposite; the petiole very short or absent; the leaves with a similar number of secondary veins and an entire or indistinctly toothed margin; the calyx adaxially glabrous, abaxially gray stellate-tomen- tose, within | mm from the margin more sparsely pubescent or glabrous, somewhat scarious, brown when dry, margins with 4 to 6 broadly deltoid teeth, subglabrous on both sides; the fruiting calyx patel- liform, not appressed to the fruit; the fruit ovoid to pyriform with the apex rounded or apiculate, a smooth gray to brown stellate-tomentose surface, and an inner fruit wall densely appressed-pubes- cent with long simple or 2-armed white trichomes; and seeds ellipsoid to ovoid-ellipsoid, with an ir- regularly rugose testa. Furthermore, characters that reportedly distin- guish Styrax zhejiangensis from S. macrocarpus are unreliable or do not otherwise serve to delimit the two taxa. According to the protologues of the spe- cies, S. zhejiangensis tends to have wider leaves with reticulate tertiary veins and a smaller, pyriform fruit, whereas 5. macrocarpus tends to have narrow- er leaves with subparallel tertiary veins and a larg- zhejiangensis as subparallel rather than reticulate and the fruit shape of S. macrocarpus (i.e., speci- mens outside of Zhejiang Province) as encompass- ing both ovoid and pyriform variants, as was also Hwang (1987). Hwang (1987) de- scribed S. macrocarpus as a tree (6-9 m tall) with observed by glabrous seeds, and S. zhejiangensis as a shrub (less than 2 m tall) with pubescent seeds. Our exami- nation of more collections than were available to Hwang (1983, 1987), however, has revealed several specimens of S. macrocarpus that exhibit a shrub habit with intermediate height (ca. 4 m; e.g., X. Ç 28884 and G. L. Shi 14815). stellate trichomes distributed on the surface of the Liu The number of seeds of S. zhejiangensis varies from several to doz- ens, even on those from the same plant. High in- fraspecific seed pubescence variation is present in two other species in this revision (S. odoratissimus and 5. tonkinensis), so this character by itself can- not be used to justify the recognition of S. zhejian- gensis. Most collections of Styrax macrocarpus are from Mang-shan, Yizhang Xian, southern Hunan Prov- ince, where the type collection was made. All other collections were made far from the type locality ex- cept one specimen from adjacent Ruyuan Xian, northern Guangdong Province (Z. L. Chen 30610). This species exhibits a discontinuous distribution between Fengkai Xian, western Guangdong Prov- ince (G. L. Shi 14815 and Exp. Guangdong 5185), » 29344), and the region of the type locality. This disconti- Jiande Xian, Zhejiang Province (Y. Y. nuity may be an artifact of human-induced extir- pation between the known localities, rather than the species’ original distribution, because the original vegetation of the whole region encompassing the range of 5. macrocarpus has been heavily modified by human disturbance. Additional specimens examined. CHINA. Guang- ong: Fengkai Xian, Qi-xing-xiang, Exp. Guangdong 5185 (IBSC y Hei-shi- -ding, G. L. Shi › (IBSC); Ru- л Yaozu Zizhixian, Tian-men-zhang, Z. L. Chen 30610 (IBSC). Hunan: Yizhang Xian, Mang: in nd known (PE); Mang-shan, Shui-kou-miao, 5. Q. Chen 2889 (AAU, BR, IBK, IBSC C, KUN, РЕ); dap ee E Q. Chen 5408 (IBS 3 (IBSC A Mang-shan, Yang- gong done, P. H. Liang 85107 (IBK, IBSC); Mang- О 28884 (IBK, IBSC, PE), зе nan-lin- shan, X. Q. shi-xi-dui 137 (BS 9. Styrax obassia Siebold & Zucc., Fl. Jap. 1: 93. 1839. TYPE: Japan. I. Keiske 287 (lectotype, [accession no. 908241— 452] not seen; digital image of lectotype!). designated here, Shrubs or trees to 15 m tall. Young twigs brown Volume 90, Number 4 2003 Huang et a 527 Revision of Styrax Series Cyrta stellate-pubescent; older twigs dark purple, gla- brescent. Petiole of larger leaves 10—15(-20) mm long, dilated at base and covering the bud. Two most proximal leaves on each shoot subopposite to opposite, smaller than distal leaves, with petioles not dilated at base or covering the bud. Lamina 5— 17 X 4-15 obovate, or suborbicular; apex cm, chartaceous, broadly elliptic, broadly abruptly caudate-acuminate; base subrounded to broadly cu- neate; adaxially glabrous except for sparse gray pu- bescence on major veins, abaxially gray-white stel- late-tomentose; subentire or remotely apiculate-dentate; secondary veins 5 to 8(to 10) on each side of midvein, tertiary veins + parallel and margin perpendicular to the secondary nerves, abaxially prominent. Fertile shoots 14—26 cm long, 1- to 3- eaved. Inflorescences arising from shoots of the current. growing season, usually pseudoterminal, occasionally also lateral; lateral inflorescences usu- ally consisting of a single flower: pseudoterminal racemes 10—18 cm long, 10- to 18(to 23)-flowered, rachis glabrous or nearly so. Pedicel 4—6(—10) mm long, white stellate-tomentose, sometimes with larg- er scattered brownish stellate trichomes; bracteoles 3—5 mm long, linear, positioned at the base or mid- dle part of pedicel, sometimes those toward the base of the inflorescence leaf-like. Flowers 1.2— cm long. Calyx 5—6 X 4—5 mm, сатрапшаіе; adax- ially glabrous; abaxially white stellate-tomentose N throughout, often also with various amounts of larg- er yellow or brownish stiff stellate trichomes espe- cially proximally; margin with 5 or 6 lanceolate to narrowly deltoid, irregularly distributed teeth, con- tiguous. Corolla 1.0-1.5 pink, tube 4—5 mm long, glabrous, lobes 5(6). 13- 16 X 4-6 mm, elliptic, apex acute, white stellate- tomentose on both sides. Stamens 10(12): filaments 6—8 mm long, straight, of equal width throughout, long. сш long, white or rarely subglabrous or glabrous; anthers 4—5 mm equal to filament in width or narrower; connective glabrous. Style proximally stellate-pubescent, oth- erwise glabrous; stigma 0.1—0.3 mm wide, puncti- Fruit 1.4-2.0 X 0.7-1.2 cm, ovoid to subo- void, apex rounded or apiculate, dehiscent by 2 valves; pericarp dry, 0.2-0.5 mm thick, outside coarsely and irregularly rugose, white or yellow- form. brown stellate-tomentose, inside glabrous. Seeds dark brown, ellipsoid, smooth, glabrous. Illustrations. Siebold & Zucc., Fl. Jap. 1: t. 46. 1835; Gard. Chron. ser. 3, 16: 513. 1888, 2 507. 1897; Hook. f., Bot. Mag. 115: t. 7039. 1889; Dip- pel, Handb. Laubholzkunde 1: fig. 205. 1889; Na- kai, Sylv. Korea 13: t. 13. 1923; Nakai, Trees Shrubs Japan ed. 2: fig. 157. 1927; Anonymous, Ic. Cormophyt. Sin. 3: 338, fig. 4629. 1974; S. M. Hwang & C. J. Qi in W. C. Cheng, Sylva Sin. 2: 1601, fig. 796. 1985; S. M. Hwang, Fl. Reipubl. Popularis Sin. 60(2): 85, pl. 28 (1—5). 1987; X. M. Liu in X. H. Qian, Fl. Anhui 4: 64, fig. 1766. 1991; 5. Y. Wang in B. Z. de Fl. Henan 3: 229, fig. 17 74. 1997; Z. Y. [C. Y.] Wu & P. H. Raven, Fl. China Ш. 15: 197, fig. 197 (1—5). 2000. Phenology. Flowering: May-July. Fruiting: June—Novembe Distribution. China (Anhui, Hubei, Hunan, Jiangsu, Jiangxi, Liaoning, Shandong, and Zhe- jiang), Japan (Hokkaido and Honshu), North Korea, and South Korea (Cheju, Kangwon, Kyonggi, North Chungchong, North Kyongsang, Seoul City, and South Kyongsang); Figure 8 abitat. In mountain slopes, and in deciduous forests in ra- mesic, open mixed forests on — vines; 9—1400 m. Vernacular names. | Lao-dan-pi (China, Shan- dong; Shandong Wild Pl. Exp. [1959] 89), Lao-kai- pi (Anonymous, 1974), Oho-ba zisja (Japan; Sie- bold, 1835-1841), Sei ton kwa (Japan; Siebold, 1835-1841), Shan-zhen-zi (China, Shandong; Hwang & Qi, 1985), Shu-ling-hua (China, Anhui; Exp. Anhui 219), Yu-ling-hua (Hwang, 1987; Anon- ymous, Sin. Yun-jin-du-juan (China, Zhejiang; Y. Y. Ho 23309). Styrax ae occurs at the extreme northern edge of the range of Styrax in Asia, extending from the island of Hokkaido (northern Japan) through North and South Korea to southeastern China. Sty- rax obassia is the only species of Styrax that occurs as far north as Hokkaido and Liaoning Province, northeastern China. It appears to be a relatively common component of wet temperate forests in Ja- pan. The dilated petiole base covering the bud sep- arates S. obassia from most other species of Styrax except S. shiraianus, which is distinguished from S. obassia byi its shorter shoots (4—8 cm) and inflo- rescences (2-3 cm), much shorter pedicels (less than | mm), and generally narrower leaves (to 9.5 cm) with more robust teeth. These species range farther north than any other species of Styrax in Asia except S. japonicus, suggesting that the dilated petiole base is an adaptation to temperate condi- tions in lieu of bud scales. Besides the dilated pet- iole base, S. obassia differs from other sympatric species by its larger flowers with glabrous filaments and styles (cf. S. odoratissimus) and shorter pedi- cels and longer inflorescences (cf. S. japonicus). See also 5. hemsleyanus for additional comments. This wide-ranging species is not extremely var- iable morphologically, and no unusual specimens were encountered in this revision. This could ex- 528 Annals of the Missouri Botanical Garden plain the highly constant treatment regarding this species since its description in 1839 No specimens were cited in the protologue of Styrax obassia. New species in volume 1 of Flora Japonica were described by J. С. Zuccarini based on data supplied by von Siebold. The only material that we have seen from the von Siebold herbarium consists of on-line images of two L collections from a database of the von Siebold collections main- Nederland One « accession number 950161- tained by the National Herbarium (). © = these collections (L 812) bears the stamp “Herbarium Ch. D'Alleizette" and a typeset (not handwritten) label that contains a reference to aximovicz, who collected in Japan between about 1860 and 1866, long after von Siebold was there (from 1823 to 1830). The label also bears an indication that this specimen was des- ignated for exchange. D'Alleizette was a plant col- lector residing in Bordeaux, France, and apparently never collected in Japan. Based on these data, we conclude that this specimen was not part of von Siebold and Zuccarini's original material. The other еи (I. Keiske 287, L 908241-452) bears labels that are consistent in ко: and format with most others in von Sie- accession. number bold's herbarium. Keiske was one of von Siebold's collaborators while von Siebold was in Japan. Al- though only a single leaf constitutes this specimen, it is clearly recognizable as that of S. obassia on the basis of, among other features, the overall ob- ovate-orbicular shape and a coarsely dentate mar- gin with the teeth most prominent apically. On this basis, we have lectotypified the name S. obassia on the L specimen of I. Keiske 28 The specific epithet “obassia” is derived from the common name for the species in Japanese. Because the epithet is a noun in apposition, it should not be modified to 2000 Article 23.5), as was done by Hwang and Grimes (1996). "obassis" (see Greuter et al., Selected specimens examined. CHINA. Anhui: Jinzhai iei Bai-ma-zhai, Xi-da-wa, K. Yao 8928 (A, CAS, K. M Yuexi Xian, E China lt Station 7007 (IBSC). Hu- : Luotian Xian, _ Hei 1251 (PE). Hunan: Hengshan Xian, Heng-shan, S. Q. Chen d 31 vee Jiangsu: vi lector unknown 2337 by Ching 3253 (A, E, fu-shan, Tian-yu- ine ү, Kk E 5891 (PEI]. prid ing: Dandong Shi, An-dong, Zhen-jiang-shan, G. Sato 5250 (PE) Fengcheng Shi, Feng-huang-shan, Z. W (PE). Shandong: Pingyi Xian, Zhou 6389 (PE); Qingdao Shi, Lao-shan, А N. Meyer 275 (A, DS, UC); Yantai Shi, Kun- кеш Shandong Wild Pl. Exp. 89 (PE). Zhejiang: Anji Xian, Tian-mu-shan, M Y. Ho 22113 (IBSC, PE); Lin'an Shi, Da- is qe Y YH 23309 (IBSC, MO, PE); Tiantai Xian, б. R. Chen D I: wv 2 А LE =ч Ч т = 5 ~ x P T = = = x (Bi = Б = A PE); Zhuji Shi, Y. Y. Ho 24026 (IBSC, MO). JA- pu Hokkaido: Haku-unloke, collector unknown (K), in 7 (A); Sapporo, Yezo, 1903, S. Arimoto s.n. (GH, MO); ARA 1861 Maximovicz s.n. (BM); Shirileshi, Okushiri, 1890, K. Miyabe & E. Tokubuchi s.n. (GH); Fur- ano City, Yamabe, K. Sohma & M. Takahashi 535 (A, MO). Honshu: Aomori Pref., Shimokita- -gun, Kawauchi- cho, K. Deguchi 5737 (MO); Fukui Pref., Nanjo-gun, Ima- jo-cho, N FEY of E ike, С Murata & H. Nishimura 5663 (A, AAU, KYO, L, PE, TI); 2 Pref., Takane-mura, as -gawa, H. Kana BM, BR, Е, К, UG Gumma Pref. Tano-gun, mura, Narahara, Es Akegasawa & Shionosawa, J. Mur- ata dig (A); ; ibo Pref., Tabu-gun, Oya-cho, Ikada, Tentaki, Murata 1030 (A, AAU, C, E, L, MO, UC); Ishikawa a f.. Shiramine- sone Akatani, Akatani-rindo, S. Tsugaru et al. 22237 O); Iwate Pref., Morioka, Mt. erm H. Muroi 5028 (A); Kyoto Pref., Kitakuwata- aig -cho, Ashiu, from Sugou to Kadzura-goya, M. jn et al. 1293 (TI); Miyagi Pref., Mono-gun, Kitakami- machi, eek S side of Okinakura- -yama, D. Е. Boufford E. W. Wood 25412 (A, MO); Nagano Pref., Suwa-gun, Fujimi-cho, Hanan nd zawa, T. Xn e 22540 (AAU); Nara Pre . Kurokami-yama, H. Mur 6962 (A); Niigata Pref., ra ‘uwonuma, Mikuni, Y. Ike- gami 2628 (A); Saitama Pref. ., Chichibu-gun, Kamiizumi- mura, a J. Murata et S deb AAU[2]. PE); ); Shiga Pref., Ika-gun, Kinomoto-cho, Harikawa, С. Murata & S Kitamura 2362 (AAU, E, "n L, UC Meu Pref., Ashi-gun, Tonbara-cho, Mt. Oyorogi, K. Mim o & 5. Thu- garu 3195 (A, MO); Tochigi Pref., Nikko. 1864. S. Tschon- oski s.n. (A, BM, C, B Tokyo Pref., инаая -gun, M Mitake-Nanayo Fall, | S. Kobayashi 099 (С Pref., Nishi-murayama-gun, Nishika River, S. Tsugaru & T Takahashi 6607 (MO). NORTH KOREA. Locality unknown: : unknown (IBSC); Wolgoic eid 1914, К. € (PE). SOUTH d A. Cheju: Halla-san, T. Ts e (A, C, E). g Kons E. H. Wilson 10422 (BM, K). b ou near Duigen, E. H. Wilson 8467 (А) Nort h Chungchong: Hwa mona -san, 33 mi. SE of Tae- jon, egi In Cho 8276 (E). North Kyongsang: Port Chusan, C. Wilford 934 (A, К). Seoul City: Tobong-san, R. Moran 5209 (BM, BR, E, GH, L, MO, UC). Soutl yongsang: S'onch'ong Dist., au of Chiri-san, Chiri- san Natl. Park, А Kirkham & Boyce KFBX86 (К). —. 10. Styrax odoratissimus Champ. ex Benth., Hooker’s J. Bot. Kew. Gard. Misc. 4: 304. 1852 as S. "odoratissimum" |. TYPE: China. Hong Kong: ravines of Mt. Victoria, J. G. Champion 138 (holotype, K!; isotypes, E!, K[3]!) Styrax о Perkins, Bot. Jahrb. Syst. 31: 1902. TYPE: China. Province unknown: чш. [from S 1907 |, Hillebrand s.n. (holotype, В 'stroyed). Styrax m а . Kew 1906: 161. & Е. Н. Wilson, Bull. Misc. 1906 [as S. “Veitchiorum” |. TY PE: China. Bus. Fang Xian, 2100-2400 m [protologue], June 1907, E. H. eei 2015 (holo- type, K not seen; isotypes, А[2]!, IBSC!). Trees to 10 m tall. Young twigs sparsely short- yellow-brown stellate-pubescent; older twigs pur- plish or dark brown, glabrescent. Petiole 5-12 mm Volume 90, Number 4 2003 Huang et al. 529 Revision of Styrax Series Cyrta long. Two most proximal leaves on each shoot al- ternate. Lamina 4—15 X 2-8 cm, chartaceous to thick-chartaceous, ovate, ovate-elliptic, or elliptic, dull light green to yellow-green at maturity when dry; apex acute to short-acuminate; base broadly cuneate to rounded; adaxially usually glabrous ex- cept midvein, abaxially usually glabrous except midvein and axils of secondary veins, sometimes yellow-brown stellate-tomentose or -hirsute but sur- face remaining visible through the pubescence: margin entire or remotely serrulate apically, sec- ondary veins 6 to 9 on each side of midvein; tertiary veins subparallel, densely, adaxially plane ог slightly sunken, abaxially prominent. Fertile shoots 7—15 ст long, 3- to 5-leaved. Inflorescences arising from shoots of the current growing season; lateral inflorescences l- to 2-flowered or racemose, 3—5 cm long, (3- to)5- to 7-flowered; pseudoterminal in- florescences usually racemose or rarely paniculate, 3-8 cm long, 5- to 7(to 11)-flowered, rarely 1-flow- ered, rachis and branches yellow stellate-tomen- tose. Pedicel 4—9 mm long, yellow stellate-tomen- tose; bracteoles 2—4 mm long, subulate, positioned at various places along the pedicel but mostly near the base, more rarely near the middle. Flowers 1— 1.5 em long. Calyx 3—4(—5) X 3—4 mm, cupuliform: adaxially glabrous; abaxially yellow stellate-tomen- tose, within 1 mm from the margin more sparsely pubescent or glabrous, somewhat scarious, brown when dry; margin truncate, undulate, or irregularly lobed, the teeth minute, not contiguous if present. Corolla 0.6-1.0 ст long, white, tube 3—4 mm long. glabrous, lobes 5(6), 9-11 X 4—6 mm, 1.7-2.2 X as long as wide, elliptic to obovate-elliptic. Stamens 10(12); filaments 1.5-3 mm long, slightly flexuous at middle, proximally broadened, distally attenuate, densely white stellate-pubescent throughout; an- thers 4—5 mm long. wider than distal portion. of filament; connectives (at least proximally) densely appressed-stellate-pubescent. Style densely white stellate-pubescent nearly throughout, distally thin- ning; stigma 0.2-0.5 mm wide. punctiform. Fruit 0.8-1.0 X 0.6—0.8 cm, usually subglobose. sionally ovoid, apex rostrate, rarely merely apicu- осса- late, dehiscent; pericarp dry, (0.3—)0.5—1.0 mm thick, outside smooth or slightly rugose, gray-yellow stellate-tomentose, inside sparsely appressed-stel- late-pubescent. Seeds brown, ovoid, slightly rugose, usually appressed-stellate-pubescent or lepidote, rarely glabrous. Illustrations. Miers, Contr. Bot. I: t. 29. 1851- 1861; Hu, Bull. Fan Mem. Inst. Biol. 3: pl. 16. 1932; Anonymous, Ic. Cormophyt. Sin. 3: 336, fig. 4626. 1974; S. M. Hwang & C. J. Qi in W. С. Cheng, Sylva Sin. 2: 1620, fig. 813. 1985; S. M. Hwang, Fl. Reipubl. Popularis Sin. 60(2): 101, pl. 34 (1—5). 1987; J. Q. Liu in L. G. Lin, Fl. Fujian. 4: 352, fig. 285. 1989; X. M. Liu in X. H. Qian, Fl. Anhui 4: 67, fig. 1771. 1991; S. Y. Wang in B. Z. Ding, Fl. Henan 3: 231, fig. 1776 (1-2). 1997; Z. Y. [C. Y.] Wu & P. H. Raven, Fl. China Ill. 15: 203, fig. 203 (1—6). 2000. Phenology. Fruiting: March-November. China (Anhui, dong, Guangxi, Guizhou, Hong Kong, Hubei, Hu- Flowering: March—July, September. Distribution. Fujian, Guang- nan, Jiangsu, Jiangxi, Shanxi, Sichuan, Yunnan, and Zhejiang); Figure 9. Habitat. leaved forests on mountain slopes, along streams in 30-2100 m. Vernacular names. In relatively mesic, semi-open, broad- ravines; Bai-mu (China, Guangxi; G. X. Li 54), Er-huan-dong-gua (China, Hunan; P. C Tam 61731), Gou-len-cai (China, Guangdong; K. P To et al. 12645), Tsoong 681), Hong-la-jiu-shu (China, Guangdong: W. T. Tsang 20435), Huang-ye-shu (China, Anhui; Exp. Anhui 2376), Ji-gu-duan (China, Jiangxi: J. Xiong 1860), Mao-ye-mo-li (China, Zhejiang; Y. Y. Ho 24610), Mao-ye-shui-dong-gua (China, Gu- angxi; S. Q. Chen 14692), Niu-zi-shu (China, Ji- angxi; C. M. Hu 5271), Ru-bai-ye-mo-li (Hwang & Qi, 1985), Shan-long-yan (China, Guangdong; W. T. Tsang 21712), Shuang-chi-shan-mo-li (China, Hu- пап: 4. Н. Shen 1235), Ye-jin-gu (China, Zhejiang: Zhejiang Bot. Res. Team 25888), Ye-ling-li (China, Zhejiang: Y. Y. Ho 26445), Yu-xiang-ye-mo-li (Chi- na, Shanxi; T. W. Liu & Z. B. Zeng 1372), Xiang- ye-ye-mo-li (China, Zhejiang; Zhejiang Bot. Res. 28350). Yu-xiang-ye-mo-li (Anonymous, Fen-fang-an-xi-xiang (Hwang. 1980). ;uang-ye-mo-li-bao (China, Guizhou; P. C. Team 1974). The Chinese endemic species Styrax odoratissi- mus is one of the most common and widespread species treated in this revision. This species is most abundant in eastern and southeastern China, grad- ually decreasing in abundance northward and west- wart pm odoratissimus exhibits much morphologi- cal variation across its range. The lower laminar surface is mostly glabrous, but is sometimes stel- late-tomentose or -hirsute, although the surface al- ways remains visible through any pubescence pre- sent. In addition, leaf size, inflorescence length, and flower number vary significantly throughout the range of the species. It nonetheless can easily be distinguished from other sympatric species with im- bricate corolla aestivation (i.e., S. hemsleyanus, 5. hookeri, S. japonicus, 5. macrocarpus, S. ѕирай, and 530 Annals of the Missouri Botanical Garden Trio 130 == i Q e Styrax odoratissimus о 800 N L —— —— " м | Kilometers з = — ——s Е ^d eg 45 t 1 J C 4 V x 5 ғ ке. 3 y NES et \ E NG Ф C рє У. У E" " 1 ` f ~- 7 v У ~ \ Й ; ‘ ) E E 4 ( —4. ^ N \ tA N EE ГА ша - ^ A ра c o L 4 B { i ў 4 30 4 ке зо ; $a ‚7 Айй. с E ۰ (— ‘ р «| ee ("Lal S $ Є, MR 4 E Л » ч f / aver 7 | e J x < I d | 5 fe C ч | A | | \ | ? Dw m MN ў Б! "mi we e AX e х f Lo) sae hs к.с) Г Lys ر‎ VAN c. "o, ә vl x pf? NR, ИЕ. HM | J Figure 9. S. tonkinensis) by the combination of its distinc- tively flexuous filaments and densely pubescent stamens and styles. Furthermore, the appressed- stellate-pubescent or lepidote seeds often served to distinguish it from other species of Styrax, although some glabrous individuals exist. Sterile herbarium specimens can usually be distinguished from other sympatric species by leaves that are yellow-green in the dried state and often slightly scabrous adax- ially. Moreover, the petiole and proximal portion of the midvein are both usually red-tinged in the liv- ing state. It is unknown whether this feature is re- stricted to this species within the genus, but it seems at least to distinguish 5. odoratissimus from sympatric species. Perkins (1907) considered differences in inflo- rescence length and the leaf margin sufficient to separate Styrax veitchiorum Hemsl. & E. H. Wil- son, a species described from flowering material collected in Fang Xian, Hubei Province, China, from S. odoratissimus. Subsequently, Rehder (1912) identified fruiting material from the same county (Wilson 308) as S. veitchiorum. We consider the characters used by Perkins (1907) to separate these two taxa to vary continuously. The fruit of S. veit- Geographic distribution of Styrax odoratissimus. chiorum differs from that of most other fruiting col- lections of 5. odoratissimus in its glabrous (vs. pu- bescent or lepidote) seeds. Glabrous seeds, however, appear sporadically in other individuals of S. odoratissimus. Because no other characters are apparent for use in delimiting the two entities, our treatment follows that of Hwang (1987) in placing 5. veitchiorum as a synonym of S. odoratissimus. Perkins (1902) described Styrax prunifolius Per- kins based on a specimen from China (Hillebrand s.n.) but later (1907) treated this species as a syn- onym of S. odoratissimus. We assume that the ho- lotype of S. prunifolius was at B because that is where Perkins conducted her work on Styrax. and that this has been destroyed. Having seen no du- plicate material of Hillebrand s.n., we here place S. prunifolius under S. odoratissimus following the precedent set by Perkins (1907), and in considering its similarity to S. odoratissimus as inferred from the original description. Some specimens of Styrax odoratissimus collect- ed at the northwestern edge of the species’ range, especially on and near Mt. Emei (E-mei-shan i Pinyin) in Sichuan Province, exhibit atypical fea- tures (i.e. a more cylindrical fruit, larger leaves, Volume 90, Number 4 2003 Huang et al. Revision of Styrax Series Cyrta and glabrous seeds; e.g., China-USSR team 1853; Ching & Shun 80; H.-C. Chow 7547 and 8016; Rev. E. Faber 195; W. P. Fang 2462, 7560, 12624, and 16790; X. Y. Huo 5654; C. B. Peng 6070; S. L. Sun 540; T. H. Tu 347 and 407; G. H. Yang 55029). Several of these have been identified as 5. hem- sleyanus, but are easily distinguished from that spe- cies by the alternate (vs. subopposite to opposite) arrangement of the two most proximal leaves on each shoot. Furthermore, the fruits available for study from these specimens are rostrate, as in 5. odoratissimus, and the flower buds (no open flowers are available for study) do not deviate from the range of variation within S. odoratissimus as defined here (e.g., the filaments and styles are densely pu- bescent throughout, the connectives are slightly pu- bescent, and the filaments are flexuous). The atyp- ical fruits and seeds observed in these specimens are distributed sporadically across the range of 5. odoratissimus, i.e., in the provinces of Anhui (C. 5. Fan & Y. Y. Li 221), Guangdong (H. G. Liu 490 and P C. Tam 58332), Guangxi (L. X. C 500159), Hunan (P. H. Liang 83722), and Yunnan (B. X. Sun et al. 254) in addition to Sichuan and Hubei mentioned above, and thus have no apparent e З taxonomic significance. Thus, we place the Emei specimens under this species, with the caveat that specimens collected at anthesis would be highly desirable for corroboration. The closest relatives of Styrax odoratissimus ap- pear to be several species from the southern part of the range of Styrax in Asia. These include 5. buchananii, porterianus, and 5. du атр The relation- well understood, S. chrysocarpus, S. curvirostratus, 5. ships among these species are but each shares with 5. оё ний а аю the alternate arrangement of the two most proximal leaves on each shoot; a lower laminar surface that is visible through any pubescence that may be present (vs. a tomentum that obscures the surface); a calyx that is truncate, undulate, or irregularly lobed, the teeth not contiguous if present, and the outer surface within 1 mm of the margin more sparsely pubescent than the rest of the calyx or subglabrous to gla- brous, somewhat scarious, and brown when dry, but without scattered orange or brown stiff long stellate pubescence. Differences between S. odoratissimus and these species are addressed in respective dis- cussion sections of each species. Selected ier ове CHINA. Anhui: Dong- zhi Xian, Xiang-ling, C. M. Tan 971113 (РЕ); Huangshan Shi, Huang-shan p n ang-shan, / Wang 3780 (IBSC, PE); Jingde Xian, Ou-yuen, P. oils 25676 P Jixi Xian, Exp. Anhui 1059 (PE); Qimen Xian, Li-xi liu-feng, Y. F Xiao & W. Z. Xie 152 (IBSC): бл aa Xian, Jiu-hua-shan, 5. C. Sun 1204 (A); She Xian, Huang- shan, M. Chen 1061 (IBSC, PE); Shucheng Xian, Wan-fu- shan, M. P. Dee 11153 (PE); Xiuning Xian, Wu-cheng, Exp. ‘Anhui 2344 (PE). Fujian: Changting Xian, Gui-long- shan, C. M. Hu 3737 (IBSC, KUN, PE); Chong’an Xian, Xin-chun-xiang, Exp. Wu-yi-shan 11 (IBSC, РЕ); Fuding Xian, Tong-mu-xiang, Tong-shan, P. X. Qiu 1487 (РЕ|2]); Fuzhou Shi, hillside near U niversity Foo-chow, T. S. Ging k y Guangze Xian, Chu-fu-xiang, Xia-yang-da- fed Y. T. Zhang 79025 (IBSC); Gutian Xian, Y. G. Yan 6215 (KUN); Hua'an Xian, Xin-kou, P. C. Tsoong 648 (IBSC[2], ы Liancheng Xian, Zhang- бор 1932, Y. Ling s.n. (PE); Minhou Xian, H. H. Chung 2742 (A, UC); Nanping Shi, Yan- бнт ев G. S. He 4256 (MO); Ninghua Хіап Hui-hua, Shui- kou-xiang, K. M. Wu 60225 (IBSC); Shan- Xi-qing, Pl. Res. Exp. i Shouning Xian, R. C. Ching 2241 (A[2], IBSC, UC); Shun- chang Xian, Tian-ping, Hou-shan, M. S Y. Li 4584 (PE); ewe Xian, Long-an, Chen-keng, Wu-niu- wan-shan, M. 5. Z. Y. Li 252 (IBSC); Xiamen Shi, Ban-tou а © L. Cai 38 (IBSC); Yong'an Xian, D. S. Wang 453 (PE); Yongchun Xian, Fang-guang, ve pid iubar : 273 (PE); е Xian, Н. Н. Chung 2615 (А, IBSC , UC). Gua : Da bu Xian, Tong-gu- н W. T. ко. 21712 (A[3], BM. IBSC, K, PE[2], UC); Feng- shun Xian, Da- tian-xiang, Bei-xi, X. G. Li 200955 (IBSC, РЕ); Heping Xian, Li-ming-shan, б. C. Zhang 256 (IBSC); Huaiji Xian, Hei-chong, Yuan-shan-lin-chang, Z. Y. Li 1681 (MO); Huiyang Shi, Luo-fu-shan, Hua-sou-tai, N. K 2 41677 (IBK, IBSC, КОМ, РЕ); Jiaoling Xian, Si-hu- ang. L. Tang 4630 (IBSC, PE); Lechang Shi, Heo-tse- ling. Da-lang, Y. Tsiang 1386 (A, IBSC, UC); Liannan Yao- zu Zizhixian, Jin-keng-xiang, P. C. Tam 59535 (PE); Lianshan ws Yaozu Zizhixian, -xiang, C. Tam 58332 (IBK, IBSC, PE) Lian- zhou Shi, oci xiang, Exp. Nan- "i 272 (IBSC); Long- men Xian, San-jiao-shan, Cong-hua, W. T. Tsang 20435 (PE); Meizhou Shi, NER ың X Ы Li 202464 (IBK, IBSC, PE); Nanhai Shi, Shih-pi-keng, Hao-shan, S. 5. Sin 9444 (A) Pingyuan Xian, Cha-gan-xiang, Huang-zhu- ping, L. Tang 4380 (IBSC, PE); Qujiang Xian, Long-tou- shan, S. P. Ko 50337 (IBK, IBSC, MO, PE); Renhua Xian, Jen-hwa Dist., Shi-bi-xia-cun, Wan-chi-shan, W. T. Tsang 26345 (A, E, IBSC); Ruyuan Yaozu Zizhixian, Qing-xi- dong. S. P. Ko 52889 (A, IBK, IBSC); Shantou Shi, Wu- king-fu, 60 mi. W of Swa-tow, J. M. Gilchrist 79 (IBSC); Shaoguan Shi, Exp. и 1244 (IBSC); Shixing Xian, Chang-ke E Che-ba- 7. L. Zhang 56031 (MO); uhua Xian, Chang-bu- eres Гян -shan, X. б. Li 201687 (IBK, IBSC, PE); Xinfeng Xian, Ah-p'o-kai-shan, Cha-ping-cun, Y. W. Taam 721 (A, CAS, К, KYO); Yang- e Xian, Wu-yuan-xiang, L. Tang 1069 BSC, KUN); Yingde Shi, Sha-kou-xiang, Hua-shui-shan, C. 163471 (IBSC); Zengcheng Shi, 2d 20. 301 Ny. TE Bose эш, Wu- nos shan, «E P P2 Guilin Shi, Di jan bd yuan-cun, Hin -shan E" : Tsang 28311 (A, IBSC); Hezhou Shi, bia 'xiang, H. Chen et al. 500072 (IBK, IBSC); Huanjiang Maonanzu с Wu- hua -shan, Jiu-ren, H Qin 895180 (К); Jingxi Xian, ng, Long- yang- -shan, S. P. Ko 55648 (A, I xiang, Da-ling, D. Н. Qin et al. 65266 (PE Niu-wei, Ba-wang-shan, Exp. Hong-shui-he 1085 (KUN); 532 Annals of the Missouri Botanical Garden Lingchuan Xian, Qi-fen-shan, Z. Z. Chen 53822 (KUN): Lingui Xian, Huang-sha-xiang, Z. Z. Chen 51016 (IBK, IBSC, KUN, PE); Longsheng Gezu Zizhixian, po n -xiang, Guang-fu Coll. Team 707 (IBK ‚ IBSC, KUN y Quan- zhou Xian, Shan-chuan-xiang, Bao-ding- ut C. H. Tsoong 83331 (IBK, IBSC, PE); Xing’an Xian, Wu-tong- shan, 7. M. Tsui 250 (A, IBSC, К, PE); Yongfu Xian, He- shun-xiang, G. X. Li 54 (IBK, IBSC); Ziyuan Xian, Chuen 1 5 Tsoong 82058 (A, IBK). Guizhou: , Shi-pan-xiang, Shi-hui-dui, Exp. j (PE); Jiangkou Xian, Fan-jing-shan, Exp. Hunan & Guizh- ои (1983) 2626 (KUN); Libo Xian, Wei-zi, 185 (MO); Shuicheng Xian, P C. Ti Wuc ie Xian, Lian-tai-shan, P. € PE[2]; Yinjiang ver Qing- du-he, Zhang et al. 4025t Hill, Hong Kong Is lud. B. Bartholomew 1916 (C. bei: Badong Xian, Ge-zi-he, Z Y uud 618 (P E Fang ; Үс hang Shi, 1888, ‚ Suo-xi-yu Nature . Y Xi et al. 443 (PE): Dao Sn Niu- dass 61731 (IBK, IBSC); Hengyang Shi, Li-mu-you, » C Tam 62348 (IBK, IBSC); Jianghua Yaozu Zizhixian, He-luo-kou- -xiang, B. G. Li 5149 (IBSC); Lin- gling Xian, ag sa Huang-jiang-yuan, 5. Q. Chen 674 (IBK, IBSC); л Xian, ius] hua, X. б. Li 203380 (hsc PE Xian, Bao-mao-xi, 7! К. Cao Tongdao Dongzu izhixian, 1 C. Tien 1028 (IBSC); Xinhuang Dongzu Zizhixian, Tian-lei forest farm, Zhong-nan-lin-shi-xi-dui 163 (IBSC); Xinning Xian, Jin-shi-zhen, Dong-tou-cun, L. B. Luo 93 (BM, BR, CAS, IBSC, PE); Yizhang Xian, Же Si Jin-quan-xiang, P. Н. Liang 83707 (IBK, IBSC, MO); Zixing Shi, Ping-jiang- xiang, Luo-jia-qiao, P. H. a 86298 (IBSC, MO) oe M Shi, pud yk 5. H. Mao et al. 44 E). Jia е Xian, Du-jiang- “Hu 2758 Ls TBS KUN, PE); Boyang Xian, a Pid Q. H. Li & C. А 1146 (PE); Chon- gren Xian, Kou-ling, Tsoong-je n, Y. Tsiang 101: 10 (IBSC, UC); Chongyi Xian, Mi-xi, Ji-gong-zui, e et al. 8625 (IBK, IBSC, KUN); Dayu Xian, isis long M. Q. Nie et al. 6700 (IBSC); Dingnan Xian, Da Xiong 1860 (PE); Dongxiang Xian, Q. H. Li ne get 1470 (PE); m hang Xian, Ping- ы -xiang, i M Hu 5271 (IBSC, PE); Huichang Xian, 3342 (IBK[2], IBSC, KU! E Xi-hu-xiang, Q ¢ C ees 23 4 ( (PE): TS Shi j^ p et al. 5008 ae Jiujiang Shi, Lu-shan, Sai-yin . Nie 7265 (КОХ); Leping Shi, Li-jun-shan, Da-he pris 5 Н. Li & C. Chen 1335 (PE); Lichuan Xian, Yan-chuan-qu, Wu-yi-shan, M. X. Nie & S. S. Lai 2881 (IBSC, KUN[2]); Longam Hg we zhi-shan, near Lin- wu- -dong- -cun, S. K. . BM); Nanfeng Xian, San-xi- xian g X. X. fune Poss “BSC, PE); Nankang Xian X Yang 6: ud (IBSC); Quannan Xian, Zhu- shan-xiang, Yao-shan, J. Xiong 723 (PE); Ruijin Shi, Qing-xi-xiang, Lian- E C. M. Hu 4252 (IBSC, KUN[2 PE); Shangrao Shi, Wu-yi-shan, M. X. Nie & S. S. Lai 4331 (IBSC, KUN); Shangyou Xian, Guang-gu-shan, M. Q. Nie F y^ 6342 (IBK, KUN); Shicheng Xian, Jing-kou- xiang, . Hu 4585 (KUN, PE); Suichuan Xian, Qi- ling-xiang, . Lai et al. 235 (PE); Wuning Xian, Yi- shan-gong-she, S. S. Lai 2464 (KUN, PE); Xiushui Xian, iod -sha- Tb Xiang-jia- Ene S. S. Lai 3458 (KUN): Yifeng Xian, Xi-keng, S. A. Lai et al. 433 (PE); Yihuang Xian, Bai-zhu- “xiang, X. X. Yang 16820 (IBSC); Zixi Xian, Ma-tou-shan-xiang, Wu-yi-shan, M. X. = Guan-shan, Nie & S. S. Lai 3530 (IBSC, KUN). Shanxi: Yangcheng Xian, Gan-qi-tong, Shu-pi-gou, T W Liu & Z. Е : Emeishan ii e ‚ К, PE); Fengj Chu 1266 (IBSC). Yuna: Yanjin Xian, Cheng- fei: eism. Exp. NE Yunnan 1163 (KUN). Zhejiang: Anji Xian, Long-wang-shan, C. Wang 18532018 (IBSC); Chun’an Xian, Lin- qr xiang, Xia-keng, Zhejiang Bot. Res. Team 27581 (MO ‚ РЕ); Hangzhou Shi, Bei-gao-feng, Ning-ying-shi, 5 Ching 512 (MO); Jiande Xian, from Jian-de to Shuang-xi-kou, ( €T Gu-tian- -miao X. Wang 2099 (PE); Lin'an Shi, Shun-xi, б. i J811 2140 (PE) Lishui Shi, Da-gang-tou, Xiao-jing, 3 Y Zhang 6054 үү PE); Longquan Shi, Feng-vang-shan, H. Y. Zou 123 (A); Ри ngyang и Tes an-ke, S of Ping-yung, R. C. Ching 2080 (A, IBS ); Qingtian Xian, Tsimp- ien, Y. L. Keng PE); НЕ Xian, Long- gong, S. Y. Zhang 3450 (PE); Suichang Xian, d jia-ping, К. : Ching 1622 (A, UC); Taishun Xian, Jin-fen, Liao-yan, . Y. Chang 8514 (MO); Tiantai Xian, Tian-tai-shan, R. С Ching 1434 (A, IBSC); Wencheng Xian, Da-jun, Jing- ning, S. Y. Chang 5178 (MO); Wuyi Xian, Xi- ve ae R. J. Jin et al. J8311012 (IBSC ); Xianju Xian, S. Y. Char 7772 (MO); Yunhe Xian, Chen-chiong, 40 mi. S of Sia- chu, К. C. Ching 1809 (A, E, IBSC, UC); Zhuji Shi, Wu- qian, X. B. Li et al. J8212029 (PE). 11. Styrax porterianus G. Don, Gen. Hist. 4: 5 1838 [as 5. var. rugosus Steenis, Bull. 249. 19 "Porterianum" |. Styrax serrulatus ard. Bot. Buiten- zorg, sér. 3, 12: 32. TYPE: Malaysia. Pulau Pinang: Pinang, Wall. Cat. No. 4401 (G. Porter s.n.) (holotype, BM!; isotypes, К[3]!). Styrax floribundus Griff., Not. PI. Asiat. 4: 287. 1854 w S. “floribunda”]. TYPE: Myanmar. Tenasserim: be tween panies and Mergue, Apr. 1835, W. Griffith s.n. (lectotype, designated here, K [loan accession no. H2000/01016- 380]: prop no. H2000/01016-39]!, F Styrax АНЕ Н. К. Fletcher, Bull. Misc. Inform. Kew 1937: 509. 1938. TYPE: ac Pattani: Betong, 20 iR © Aug. a A. F. G. Kerr 7494 (holotype, К!; isotypes, BM!, E . [loan accession Trees to 20 m tall. Young twigs dull red or gray tomentose; older twigs gray, glabrescent. Petiole 3— 7 mm long. Two most proximal leaves on each shoot alternate. Lamina 5-11 X 3—5 ст, membranaceous or thin-chartaceous, ovate- to elliptic-oblong, green to dark green at maturity when dry; apex slightly acuminate; base usually oblique-rounded, rarely oblique-cuneate, short-attenuate; adaxially gla- brous except along the major veins; abaxially gla- brous or sparsely to densely white stellate-pubes- cent, pubescence especially prevalent on veins and the most proximal two leaves on each shoot, surface remaining visible through the pubescence; margin entire or usually remotely serrulate; secondary nerves 5 or 6 on each side of midvein; tertiary veins + parallel and perpendicular to the secondaries, Volume 90, Number 4 2003 Huang et al. 533 Revision of Styrax Series Cyrta faintly prominent on both sides. Fertile shoots 5— 12 cm long, 2- to 5-leaved. Inflorescences arising from shoots of the current growing season; lateral inflorescences l- to 2-flowered or racemose, 2-3 cm long, l(to 5)-flowered; pseudoterminal inflores- cences racemose, 2—4 cm long, 3- to 5(7)-flowered, rachis red-gray stellate-tomentose. Pedicel 3—10 mm long, densely stellate-pubescent; bracteoles 0.5-2 mm long, linear, positioned at various places along the pedicel but mostly near the base, more rarely near the middle, sometimes those toward the base of the inflorescence leaf-like. Flowers 0.7—1.3 cm long. Calyx 3—4 X 3—4 mm, campanulate; adax- ially glabrous or sparsely short-appressed-stellate- pubescent; abaxially gray stellate-tomentose, within | mm from the margin more sparsely pubescent, somewhat scarious, brown when dry; margin trun- cate, undulate, or slightly 5-lobed, the teeth minute and not contiguous. Corolla 0.5—0.9 cm long, white, tube 2-3 mm long, glabrous proximally, lobes 5. 10-11 х 3-4 mm, linear-lanceolate, apex acute- acuminate, adaxially sparsely stellate-pubescent, abaxially densely so. Stamens 10; filaments 3—4 mm long, straight or flexuous at middle, distally attenuate, moderately to densely white stellate-pu- bescent on both sides, thinning to glabrous distally: anthers 3—4.5 mm long, wider than distal portion of filament; connective glabrous. Style glabrous: stigma 0.4—0.5 mm wide, subcapitate. Fruit. 0.9— 1.5 х 0.8-1.2 cm, subglobose to globose, apex rounded or short-apiculate, indehiscent; pericarp fleshy, ca. 2 mm thick, outside deeply rugose when dried, gray stellate-tomentose, inside glabrous. Seed brown, ellipsoid to ovoid-ellipsoid, nearly smooth, glabrous. Illustrations. Miers, Contr. Bot. I: t. 29. 1851— 1861; Steenis, Bull. Jard. Bot. Buitenzorg, sér. 3, 12: 222, fig. 3 (5). 1932 (as S. serrulatus var. ru- у . Tree Flora of Ma- laya 3: 265, fig. 1. 1978 (as S. serrulatus var. ru- gosus). Phenology. Flowering: March—May, July. Fruit- ing: March—May, July, August, November Distribution. Malaysia (Kedah, Pahang, Perlis, and Pulau Pinang), Myanmar (Tenasserim), and Thailand (Chumphon, Krabi, Nakhon Si Thammar- at, Pattani, Phangnga, Phuket, Satun, Songkhla, Su- rat Thani, and Trang); Figure 4 Habitat. In mesic, mixed primary forests; 50— От Vernacular names. Fa La Mi Bai Leg те S. Phusomsaeng 241), Kam Yan (Thailand: A. Kerr 7494), Lang Ka Re (Thailand; A. А G. is 15300), 8505) Styrax porterianus is the only species of Styrax Pang Ka Re (Thailand; A. К G. Kerr кы with imbricate corolla aestivation known from the Malay Peninsula, where it is endemic. Styrax por- terianus appears to be most closely related to S. odoratissimus and S. subpaniculatus. All three spe- cies share relatively small flowers, a truncate, un- dulate, or irregularly lobed calyx margin, and an abaxial laminar surface usually visible through any pubescence present. Furthermore, they all occur at relatively low elevations. Styrax porterianus is eas- ily distinguished from S. subpaniculatus by its shorter raceme and thinner, dry and rigid pericarp, and from Styrax odoratissimus by the characters in couplet 20 of the key. Despite the placement of Styrax porterianus into series /mbricatae by Perkins (1907), Steenis (1932) considered this species to be a variety of S. serru- latus (var. rugosus Steenis). Perkins (1907) placed Styrax serrulatus in series Valvatae based on its val- vate corolla aestivation. Steenis's concept of S. ser- rulatus, however, contained both imbricate and val- vate types of aestivation based on Perkins's (1907) assertion that this and several other species in se- ries Valvatae include individuals that exhibit a mix- ture of these types, even within the same flower. In contrast, we consider aestivation type to be a reli- able taxonomic character with which to distinguish S. porterianus from S. serrulatus. In several species of series Cyrta (those listed by Perkins under series Valvatae and several other more recently described species), a subvalvate condition occurs whereby the edges of the corolla lobes are contiguous but oblique in cross (see Steenis, 1932: 10c). This is malis ve different, however, from section the strictly imbricate corolla aestivation observed in all specimens of S. porterianus (see Steenis, 1932: fig. 10d). Styrax porterianus differs from 5. serrulatus s. str. in other aspects of both the foliage and fruit, as Steenis recognized. In S. porterianus, the margin of the lamina is entire or at most remotely denticulate, whereas that of S. serrulatus is distinctly toothed. The fruit of S. porterianus has a fleshy pericarp at maturity (Putz & Ng, 1978; unique among species of section Styrax) that is rugose and ca. 2 mm thick in the dried state. In contrast, the pericarp of S. serrulatus is dry and rigid at maturity, nearly smooth, and less than 1 mm thick. Furthermore, the ranges of these two entities are geographically distinct, with S. serrulatus in the Himalayan region and not extending as far south as the Malay Pen- insula (P. Fritsch, unpublished data). The sum of these differences warrants the recognition of S. por- 534 Annals of the Missouri Botanical Garden terianus at the species level. Fletcher (1938) ap- pears to have understood the significance of these differences as well in describing S. betongensis Н. R. Fletcher from Thailand, apparently unaware of the earlier name. e have seen two sheets of W. Griffith’s collec- tion of Styrax floribundus from К and one from К. None of these display any indication of holotype status. Because Griffith's herbarium was transferred to K, we have chosen a lectotype from among the two К specimens. The two sheets offer little evi- dence for a decision on proper lectotypification, and we could not locate a literature source with a sample of Griffith's handwriting. On one of the K sheets, however, the locality is spelled as in the protologue (*Mergue"), whereas on the other it is spelled differently (“Мегди”), suggesting that the locality information on the latter was transcribed incorrectly some time after the original collection was made. On this basis, we have designated the sheet with the protologue spelling of the locality as the lectotype. Additional specimens examined. MALAYSIA. аа Jeniang, Kedah, bin Kiah, Sidek 5345 (C, 1). Pahan Tembeling, Ulu Sg., NW Tanjong aig M. E Bin Haji ое Nur & M. Noor MS2027 (C, 1). Perlis: aki bukit, M. S. Kiah bin Hadji Perm ^ (BM. К, E. Pu- Mu Pinang, узы tor unknown (BM), 1890, col- lector unknown (E), C. Curtis 1538 (BM, L), V derson 18 (L); Pinang Is and, 1824 .n. : Peuara Bakir, 1896 Spic МЕИ (BM); Polo Boe- long, collec dd unknown 1189 (K). ' jis IAIL P Chum- o»hon: Kao Po Ta luang Kaew, Rar ў d 339 (L). ). Krabi: ái n, Ms Khon, В. К ros 101 (C, К, L). pare Si Thammarat: Tung Song, N. ES Bunnag 92 (BM y ); Ban Kram, Nakawn Sritamarat, A. F. G. Kerr 15300 n [2]. К); Ban Kram, Palatung, А. К. G. Kerr 15302 (BM). Phangnga: Khao Phra Mi, Flor ra of сун Project 4th Exp. (1972) 30878 (19 u о, К. Geesink T. Santisuk P3 (AAU, C, Е 1). Phu ket: Saul: Tung nui, . Kerr gs 9 "iv. E, K): Kaokatawam, A. К ла m 505 (BM, E, К); Lanta, А F. G. Kerr 18988 HS , E, K). Satun: Е Natl. Park AAU). Maxwell XE from Talo Wao to Talo Oo Dang, G. Congdon 507 So a: Dist. Haad Yai, Ko Hong Hill. J. F 85346 (AAU, BM, 535 у Dist., Khao pia Natl. атага, JF ‚ Gahrome Galls, N Maxwell ^ 5669 (L). Surat ie Жз .F т Ker 12519 (BM, E rang: ‚ Khao C hong, S. S. Phusomsaeng 241 (AAU, C, E, TR 12. Styrax rugosus Kurz, J. Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 40(2): 61. 1871 [as S. “ru- gosum" |. TYPE: Myanmar. Pegu: hills between Sittang & Salween, 1212 m, Brandis s.n. (as Brandis 936, 1907) (holotype, CAL not seen). Perkins, Shrubs or trees to 6 m tall. Young twigs yellow- brown stellate-tomentose; older twigs purplish and glabrescent. Petiole 2-3 mm long. Two most prox- imal leaves on each shoot alternate, or more often subopposite to opposite. Lamina 3-7 X 2-3 em on fertile branches. those on sterile branches usually larger, to 9 X 4.5 em, chartaceous, ovate-oblong, ovate, or elliptic: apex acute or more often acumi- nate: base rounded to broadly cuneate, often slight- ly oblique; adaxially rugose and densely covered with simple and 2- or 3-armed to stellate trichomes when young, becoming sparsely pubescent or rarely glabrous: abaxially gray-yellow stellate-tomentose; margin serrate or apically dentate, secondary veins 4 to 7 on each side of midvein, tertiary veins par- allel. quaternaries as well as the tertiaries abaxially prominent and raised in young leaves. Fertile shoots (4—)6—-10(-12) em long, 3- to 5-flowered. In- lorescences arising from shoots of the current к, growing season; lateral inflorescences usually 1- or 2-flowered: pseudoterminal inflorescences race- mose, 2—4(—6) ст long, 3- to 6-flowered, rachis yel- low stellate-tomentose, also intermixed with stalked trichomes; bracteoles 4—12 mm long, linear, posi- tioned at various places along the pedicel or at the base of the calyx, sometimes those toward the base of the inflorescence leaf-like, margin conspicuously serrate. Pedicel 3—4 mm long, stellate-tomentose. Flowers 1.4-1.6 cm long. Calyx 4.5-5 X 3.5-5 mm, сири огт; adaxially sparsely appressed-pu- bescent with short white 2- or 3-armed or stellate trichomes; abaxially yellow stellate-tomentose throughout, often also with various amounts of larg- er scattered dark yellow or orange stiff stellate tri- chomes, especially proximally; margin distinctly dentate, the teeth usually contiguous or separated by a shallow concave portion; teeth 2—3 mm long, anceolate to subulate, apex acuminate, densely stellate-pubescent on both sides. Corolla 1.0—1.2 cm long. white, tube 4—5 mm long, glabrous prox- imally, lobes 5, 5-10 X 4-5 mm, elliptic to ob- ovate, adaxially subglabrous, abaxially densely pale yellow stellate-pubescent. Stamens 10; filaments 7— 8 mm long, straight, of equal width throughout, densely white stellate-villous proximally, trichomes up to 0.5 mm long, becoming glabrous distally; an- thers ca. 5 mm long, wider than distal portion of filament: connectives glabrous. Style glabrous o sparsely white stellate-villous; stigma 0.2—0.4 mm wide, Fruit 0.7-0.9 X 0.5-0.6 cm, ovoid, apex rounded or apiculate, dehiscent; peri- carp dry, 0.2-0.3 mm thick, outside irregularly lon- gitudinally striate throughout, yellow-brown stel- ale- ~ punctiform. tomentose, inside glabrous or sparsely downy-pubescent. Seeds brown, ovoid, smooth, gla- brous. Volume 90, Number 4 2003 Huang et al. 535 Revision of Styrax Series Cyrta - Illustrations. C. Y. Wu, Fl. Yunnan. 3: 426, pl. 121 (1—5). 1983; S. M. Hwang & C. J. Qi in W. С. Cheng, Sylva Sin. 2: 1605, fig. 800. 1985; S. M. Hwang, Fl. Reipubl. deer Sin. 60(2): 85, pl. 28 (7-12). 1987; W. Q. Yin in Y. C. Xu, lc. Spas Yunnan. 2: 894, pl. 471 (1-9. 1990; Z. Y. |C. Y Wu & P. H. Raven, Fl. China Ill. 15: 198, fig. 198 (8—14). 2000. Phenology. Flowering: March—July. Fruiting: April, July, August, October, November. Distribution. China (Yunnan), Myanmar (Man- dalay Division and Shan State), and Thailand (Chiang Mai, Loei, and Mae Hong Son); Figure 5. Habitat. In relatively sunny, mixed forests on mountain slopes; 700—1650(—2300) m. Vernacular names. — Zhou-ye-an-xi-xiang (Hwang, 1980), Zhou-ye-ye-mo-li (Anonymous, 1974). Styrax rugosus occurs primarily in open forests — at middle elevations in northwestern Thailand, cen- tral and. southern Myanmar, and southern. Yunnan Province, China. Numerous specimens are avail- able from throughout most of the geographic range of this species, especially at the extreme northern (Jingdong Yizu Zizhixian, Yunnan Province) and southern (Chiang Mai Province, Thailand) edges. In addition to its close morphological similarity to 5. limprichtii (see discussion under that species), 5. rugosus is also sympatric with three other imbri- cate-flowered Styrax species in southern. Yunnan Province (S. hookeri, S. japonicus, and S. tonkinen- sis), from which it is easily separated by the prom- inently long calyx teeth and rugose leaves. The lon- ger pedicels and glabrous abaxial leaf surfaces of S. japonicus, the larger fruit of S. hookeri, and the longer petioles and tuberculate seeds of S. tonki- nensis also can be used to distinguish these species from S. rugosus. Additional specimens p inue CHINA. Yunnan: Jingdong Yizu i-sheng-miao, China-U. SSK SC KUN); near Jiu-tsun, Meng-ku- 1); Huang-cao-ling, ing 101327 (KUN); Z. H. Yang 101681 (KUN[2]); Me ehai Xian, K. L. Le 235 (KUN); Fo-hai, C. W. Wang 74113 (A, Tt KUN, PE), 77088 (A. PE) Nan-chiao, C. W. Wang 75068 (A, KUN, PE), 75198 (A, IBSC, KUN, Mo-jiang & Pu-er, s.n. (ТЇ); Maymyo, Buchanan 25 (E); T mN C. E. Partinin 680 (K). Shan State: Mt. Mo-la-hein, А G. Dickason 8750 (A, E, L); Laungyi, A. Khalil DII89 (A); Paugmi State, near Leja, W. A. Robertson 152 (K). Localit known: C. B. Collett 800 (K); Thaymyo, F. G. Dickason 6008 (A). T HAI- LAND. an Bo Luang, C. F. van Beusekom ; E, K, L); Doi Intanon, “Danish Exp. (1958/1959) 3295 (C K); Bo Luang, ae 8. Thailand Project Second Exp. (1968) 1913 (AAU, C, 1); Doi Angka, Dci Pa Maun, Н. B. G. Garrett 376 ү K, L); Chiang Dao Dist., Doi Sahm Meun Range, Doi Chiam, A. Griffith 2 (CAS, L); Me Jun, A. F. С. Kerr 6201 (BM, К), 6201A (BM, E, К); from Sop Aep to Pha Mawn (Ban Yang), G. Murata et he T15602 Bn peni Sanam, C. Phengklai et al. 4150 (C, K, L). Loei: J collector unknown DI189 (A); Phukrading, 7 Smiinand 328 (A). Mae Hong Son: Jawm Tong, Mae S tidge, Mae Soi Subdist., Awp Luang Natl. Park, near Ban bus G em (Mong Village), J. F Maxwell 91535 (AAU, P, CAS), 93944 (CAS, L). Locality unknown: Bo L aang, t Geesink et al. 5776 (AAU, C, E), T5776 (L), A. F G. Kerr 4201A (К), 8855 (BM, К); Hoi, Pu Jang, A. F G. Kerr 8855A (E "EI 13. Styrax shiraianus Makino, Bot. Mag. (Tokyo) 12: 50. 1898 [as S. “Shiraiana”]. Strigilia shi- raiana (Makino) Nakai, Trees Shrubs Japan 1: 256. 1922. TYPE: Japan. Honshu: Shizuoka Pref., Sugura, Araizawa in Abe-gori, Herb. Sc. Coll. Imp. T Tokyo s.n. (lectotype, desig- nated here, TI!). Styrax shiraianus. var. discolor Nakai, J. Jap. Bot. 14: 631. 1938. ' (fr), Т! Nakazima s.n. (type material, TI missing). Trees to 8 m tall. Young twigs purple-gray, yellow or brown stellate-tomentose; older twigs gray, gla- brescent. Petiole of larger leaves 8-15 mm long, dilated at base and covering the bud. Two most proximal leaves on each shoot subopposite to op- posite, smaller, with petioles not dilated at base or X 7-9.5 cm, char- broadly obovate or rhomboid-orbicular; covering the bud. Lamina = aceous, apex rounded or short-caudate; base cuneate or cu- neate-rounded; adaxially deep green, with scattered simple or 2- or 3-armed to stellate trichomes, es- pecially prevalent proximally, glabrescent; abaxi- ally pale green to pale white, sparsely stellate-pu- bescent, glabrescent except in the axils of the midrib and secondary veins; margin proximally glandular-serrulate, distally irregularly grossly den- tate; secondary veins 4 to 6 on each side of mid- vein; tertiary veins parallel, abaxially prominent. Fertile shoots 4-8 ст long, 2- to 4-leaved. Inflo- rescences arising from shoots of the current growing season; lateral racemes usually 1-flowered; pseu- doterminal inflorescences racemose, 2—3 cm long, 3- to 11-flowered, distally congested, rachis yellow stellate-tomentose. Pedicel < 1 mm long, densely white and brown stellate-villous; bracteoles ca. 6 mm long, linear or setaceous, positioned at the base of pedicel, often those toward the base of the inflo- rescence leaf-like. Flowers 1.5-2 cm long. Calyx campanulate, 4—6.: mm; adaxially densely appressed-pubescent with 2- or 3-armed or stellate 536 Annals of the Missouri Botanical Garden [i40 [144 2 V Q 100 — 200 Kilometers 38 34| | A г” c x А e | wd X Y 4 А | \% Se - SM /7 Z Е put % Figure 10. trichomes; abaxially stellate-tomentose, often also with various amounts of larger yellow or brownish stiff stellate trichomes, especially proximally, with- in 1 mm from the margin more sparsely pubescent or glabrous, somewhat scarious, brown when dry; margin 5- to 8-toothed, teeth 1.5-2 mm long, del- toid, contiguous, apex acute. Corolla 1.0-1.5 cm long, white, tube 10—12 mm long, proximally gla- brous, distally pubescent, lobes 5, 6-8 х 2.5-4 mm, ovate, apex acute, stellate-tomentose on both sides. Stamens 10; filaments 3—4 mm long, straight, of equal width throughout, sparsely stellate-pubes- cent; anthers 2-3 mm long, wider than distal por- ++ tion of filament; connective glabrous; style proxi- mally stellate-pilose, distally glabrous; stigma 0.4— 0.6 mm wide, punctiform. Fruit 0.8—1.0 х 0.6-0.8 Geographic distribution of Styrax shiraianus. cm, ellipsoid to subglobose, apex rounded or apic- ulate, dehiscent by 2 or 3 valves from apex; peri- carp dry, 0.3-0.7 i stellate-tomentose, Seeds brown, ellipsoid, smooth, glabrous. mm thick, outside smooth, white inside sparsely pubescent. Illustrations. Perkins in Engl., Pflanzenr. IV. 241 (Heft 30): 71, fig. 9. 1907; Nakai, Trees Shrubs Japan 1: 256, fig. 141. 1922 (as Strigilia shiraiana); Perkins, Ubers. Gatt. Styrac.: fig. 9. 1928 Phenology. Flowering: May, June. Fruiting: July-November. Distribution. Japan (Honshu, Kyushu, and Shi- koku); Habitat. In open deciduous forests; 600- 1500 m. Volume 90, Number 4 2003 Huang et al. Revision of Styrax Series Cyrta Vernacular names. Ko-hakuunboku (Japan: 1901, T. Makino s.n.), Uraziro-kohakuunboku (Ja- Nakai, 1938). Styrax shiraianus is endemic to Japan (but see pan; below), occurring on the islands of Honshu, Shi- koku, and Kyushu. It appears to be a rare species, relatively little material being available for study. This species is easily distinguished from other spe- cies of Styrax series Cyrta by the racemes with dis- tally congested flowers, long (10—12 mm) corolla tube, and very short pedicel (less than 1 mm). When only sterile specimens are available, the only other taxon with which S. shiraianus might possibly be confused is S. obassia. Both species have peti- oles that are dilated at the base and cover the bud, unlike all other species of Styrax, in which the bud is exposed. Sterile material of 5. shiraianus can be distinguished from that of S. obassia by its smaller leaves that are abaxially glabrous or nearly so (vs. densely gray-white stellate-pubescent) and irregu- larly grossly deltoid-dentate (vs. subentire or re- motely apiculate-dentate) leaf margins. Apparently based on an erroneous observation of valvate corolla aestivation in Styrax shiraianus, Na- kai (1922) transferred this species to Strigilia Cav.. a genus described by Cavanilles (1789) and taken up by Miers (1859) to accommodate many South American species of Styrax. Later Nakai (1938) transferred it back to Styrax. Styrax shiraianus has been reported from South Korea by Nakai (1938) on the basis of two collec- tions from "Tiisan" (Chiisan) Mountain (5. Okamoto “Zennan” purs Cholla) Province and Tei-daigen s.n. from s.n. from > (South. Kyongsang) Province). Subsequently, the species was ‚ listed in three references on the Korean flora (T. B. Lee, 1989; W. T. Lee, 1996; Y. N. Lee, 1996), but lo- cality or source information was not specified in any of these works. Nakai worked at TI until 1943, and from this we assume that the Korean material of 5. shiraianus is stored at TI. We have not seen any material of S. shiraianus, however, from Korea among our loans from TI or other herbaria. Fur- thermore, we have not observed any photographs of living plants of S. shiraianus from Korea. The Flora of Korea (Y. N. L graphs of nearly all Korean species, including 5. ee, 1996) contains color photo- japonicus and S. obassia, but a photograph of 5. shiraianus is notably lacking. No specimens of 5ty- rax at SNU in Seoul, South Korea, have been iden- tified as S. shiraianus (C.-W. We cannot be certain, therefore, Park, pers. comm.). that the. Korean specimens cited by Nakai are not merely misiden- tified individuals of, e.g., S. obassia. our collections were cited in the protologue of Styrax shiraianus: Aug. 8, 1884, T. Makino s.n.; Aug. 1885, T. Makino s.n., K. Watanabe s.n.; and Herb. Sc. Coll. Imp. Univ. Tokyo s.n. The Makino herbarium (MAK) houses none of these specimens (M. Wakabayashi, pers. comm.), and TI has only the last of these (one sheet). Therefore, we have chosen to lectotypify on the only sheet of the syn- types known to exist among these herbaria. The TI herbarium does not have type material of S. shi- raianus var. discolor (H. Ohba, pers. comm.). Additional specimens examined. JAPAN. Honshu: Gifu Pref., Pas p Nakatugawa-shi, near Okunodair: foot of Mt. . Hidehiko 14 (KYO); Mino, K. a 2771, Net pos 7198, 9048 (А); Hyogo Pref., suhiko, H. Muroi 38 (A); Nagano Pref., Shinano, jo 1905, G. Jack s.n. (A, GH); Shinano, Ihida-shi, Mt. е Ohdaira, 1961, Е Miyoshi s.n. (А); Shinano, Nishichikuma-gun, Ohtaki-mura, M. Mizushima 2379 (A); Kodzuke. Agatsuma-gun, Sawada-mura, Shima hot well, M. Илана 2958 (А); ри нА е Ри At "i ice G. Murata & H. Nishimura 906 (AA C, E, K, JC); Shinano, F. H. Wilson 7012 (A. Er ӨН, К); ELA Pref., Okayama-ken, Ushiroyama ai- dagun, 1951, K. Uno s.n. (A); Shiga c. ‚ Shiga- -gun, Sgi- ga-cho, Yakumogahara in Hirasan Mts., G. Murata 55807 (A. KYO); Tochigi Pref., Nikko, 1901, collector unknown (А), 1914, collector unknown (E), 1915, collector pe n Я Makino s.n. (А, ТІ), 105785, 121299 (CAS), 105786 (A), те №. Mochizuki s.n. (А), 1920, Н. Takeda s.n. (BM), Е. Н. Wilson 7710 (A); Nikko-shi, Mt. сым -yama, Y. Tateishi 10287 (А). Kyushu: Kumamoto Mt. Ichibusa, Higo, 1908, collector unknown (E), tw E. E. Harmsen s.n. (L), 1910, N. Mochizuki s.n. (E). 1917, Tashiro s.n. (А); Sobosan, Pere U. J. Faurie 3272 ); Kagoshima Pref., Mt. Kirishima, 1938, T. Naito s.n. A). Shikoku: Tokushima Pref., Mt. Tsurugi. M. Hiroe 1341 (C, UC); Ehimi Pref., Kamiukena-gun, Omogokei, 1940, С. Murata s.n. (A); Kochi Pref., Iyo sikoku, Yogo Ikkaku, /. Yogo 9510 (A). =" Naki- = 14. Styrax subpaniculatus Jungh. & de Vriese, in de Vriese, Pl. Nov. Ind. Bat. 9. 1845. Styrax serrulatus var. mollissimus Steenis, Bull. Jard. Bot. Buitenzorg, sér. 3, 12: 250. 1932. TYPE: Indonesia. Sumatra: province unknown, Tob- ing Dist, Battalands, 900 m, 1860—1862 (Steenis, 1932), Е W. Junghuhn s.n. (holotype, accession no. 90631-105]!; isotype, L [ac- cession no. 908239-1494]!). m Styrax ителүен Miq., Fl. Ned. Ind., Eerste Bijv. 474. 1860 [as 5. “subdenticulatum”|. TYPE: Indo- nesia. peat province unknown, western Sumatra, Battang Baroes 1856 (Steenis, 1932), ype, U not seen; nu i image of holotype!; isotype, BO not s Styrax айаш ше, Bull. Jard. Bot. Buenos sér. 3, 12: 1932. TYPE: Indonesia. Sumatra: Su- matera Barat, E coast, Manindjau, Kp. Silajang, 500 m, 7 July 1922, Forest Research ч жн b.b. 396. (holotype, BO!; isotype, L!). 538 Annals of the Missouri Botanical Garden Trees to 33 m tall. Young twigs yellow-brown stellate-tomentose, terete; older twigs dark brown, glabrescent. Petiole 3-9 mm long. Two most prox- imal leaves on each shoot alternate. Lamina of fer- tile shoots 4—8.5 X 2—5 cm, those of sterile shoots 14.5 X 7.5 сш, ceous, ovate, ovate-oblong, elliptic, or lanceolate; membranaceous to thick-charta- apex acuminate to caudate; base subrounded broadly cuneate, slightly attenuate, sometimes oblique; adaxially subglabrous except on the midrib and the primary abaxially nearly glabrous to stellate-pubescent or -tomentose, the surface usually remaining visible through the pubescence; margin entire or indistinctly toothed, nerves, glabrescent; occasionally revolute; secondary veins 6 to 8 on each side of midvein; tertiary veins + parallel and perpendicular to the secondaries. Fertile shoots (12—)15-21 cm long, (1- to)3- to 5-leaved. Inflores- cences arising from shoots of the current growing season; lateral inflorescences l- or 2-flowered or racemose, 3—8 cm long, (1- 10)5- to 13-flowered; pseudoterminal inflorescences racemose or panic- ulate, 7-17 cm long, 9- to 20(to 23)-flowered, eral branches 2 to 7, sometimes with 2 to 3 addi- at- tional racemes from base of inflorescence, rachis and branches yellow-brown tomentose. Pedicel 4— 6.5 mm long, stellate-tomentose; bracteoles 1—3 mm long, subulate or linear, mostly positioned at the base of the pedicel. Flowers 0.9-1.2 ст long. Calyx 3—4 X 3—4 mm, campanulate; adaxially gla- brous or finely short-appressed-stellate-pubescent; abaxially yellow tomentose, arms of trichomes < 0.2 mm long, densely gray-white stellate-pubescent throughout; margin truncate, undulate, or irregular- ly lobed, the teeth minute, not contiguous if pre- sent. Corolla 0.5—0.8 cm long, white, tube 2.5-3 mm long, glabrous proximally; lobes 5, 7-9 x 2.5— ) mm, 2.3-2.8X as long as wide, oblong-elliptic, apex obtuse or acute, tomentose on both sides. Sta- mens 10; filaments 2.5-3 mm long, slightly flexu- ous at middle or occasionally straight, distally at- tenuate, densely white stellate-pubescent; anthers 3—4 mm long, equal to filament in width or narrow- er; connectives glabrous. Style glabrous; stigma 0.3—0.5 mm wide, punctiform. Fruit 0.7—1.0 X 0.8 ст, obovoid or globose, apex rounded or sub- 0.6— acute, rarely also apiculate, indehiscent; pericarp dry, 0.2-0.5 mm thick, outside smooth, gray to- Seed brown, mentose, inside downy-pubescent. ovoid, nearly smooth to irregularly rugose, glabrous. Jungh. & de Vriese, in de Vriese, 2. 1845; Steenis, Bull. 12: 222, fig. 3 (3). Illustrations. РІ. Nov. Ind. Bat.: pl. З, 1-1 Jard. Bot. Buitenzorg, sér. 3, 1932 (as S. S. oliganthes). oliganthes); ibid.: 242, fig. 9. 1932 (as Phenology. Flowering: February-April, Octo- ber. Fruiting: May-August, October. Distribution. Indonesia (Sumatra); Figure 4. Habitat. in montane rain forests; 100—1600 m. Vernacular names. Кајое lomlang kajoe (R. 5, Boeea 9285), kajoe komajan (J. E. Teysmann 965HB), or kajoe keminjan (Perkins, 1907 ex F. A. W. Miquel) Styrax subpaniculatus is the only species of Sty- In mesic, mixed primary forests, and rax with imbricate corolla aestivation known from the island of Sumatra, Indonesia, where it is en- demic. Steenis (1932) considered this species to be a variety of S. serrulatus (var. mollissimus Steenis), a species placed by Perkins (1907) in series Val- vatae on the basis of its valvate corolla aestivation. Using the same reasoning as that outlined in the discussion of S. porterianus, we consider S. subpan- The consistently imbricate corolla aestivation in S. sub- iculatus a species distinct from S. serrulatus. paniculatus sharply delimits this species from S. serrulatus, which in our view possesses a subvalvate type of corolla aestivation. Styrax serrulatus is geo- graphically distinct from 5. subpaniculatus, occur- ring in the Himalayas and vicinity but not extend- ing as far south as the Malay Peninsula or Sumatra. Styrax subpaniculatus can also be distinguished from S. serrulatus by its usually pubescent (vs. gla- brous or nearly so) abaxial leaf surfaces and the truncate or undulate (vs. distinctly toothed) calyx margin. Styrax porterianus has many features in common with S. subpaniculatus, but has shorter (2—4 vs. 7— 17 cm long), strictly racemose (vs. often paniculate) inflorescences, and a fruit with a fleshy (vs. dry and rigid) pericarp that is deeply rugose (vs. smooth) in the dry state. In addition, the ranges of S. subpan- iculatus and S. porterianus are completely non- overlapping, the latter being restricted to the Malay Peninsula. Styrax subpaniculatus is also similar to S. buchananii and S. odoratissimus but distinguish- able from both by its glabrous connectives and styles. In addition, the larger flowers (1.3—1.6 vs. 0.9-1.2 cm long) and longer anthers (6-7 vs. 3—4 mm) are useful characters to distinguish S. buch- ananii from S petals (4—6 vs. stellate-pubescent or lepidote (vs. glabrous) seeds . subpaniculatus, whereas the wider 2.5-3 mm) and usually appressed- readily distinguish S. odoratissimus from S. subpan- iculatus. Steenis (1932) described Styrax oliganthes based on a single fruiting collection from western Suma- tra. Although hesitant to describe this species as Volume 90, Number 4 2003 Huang et al. Revision of Styrax Series Cyrta new from only fruiting material, Steenis felt that the combination of densely pubescent abaxial leaf sur- faces and apparent lack of any brown leaf pubes- cence (ie. only white trichomes) provided suffi- cient justification. for the recognition of а new species. Steenis postulated Styrax benzoides Craib and S. tonkinensis as close relatives of S. oligan- thes, with 5. benzoides distinguishable by its inde- hiscent fruit and S. tonkinensis by its tuberculate seeds. We agree that neither species could possibly ith 5. oliganthes: t oliganthes differs from S. tonki- nensis in its truncate or undulate (vs. distinctly den- д be conspecific w besides smooth seeds, S. tate) calyx margin and rounded (vs. rostrate) fruit apex. Styrax benzoides has Һе depressed-globose seeds of series Benzoin (see Fritsch, 1999); those of Styrax oliganthes are ellipsoid, clearly establish- ing its inclusion in series Cyrta. Steenis did not consider a possible relationship of Styrax oliganthes with S. subpaniculatus. None- theless, there is ample evidence of affinity between these two entities. Both can reach a height of 30 m or more, which is uncommonly tall for species of Styrax; the leaves are of the same general dimen- sions, with equivalent numbers of secondary veins on each side of the midvein and an entire or in- distinctly toothed margin; the fruiting calyx margins are truncate or undulate; the fruit is indehiscent, + subglobose to slightly obovoid, smooth, and of similar general dimensions and color; the seeds are glabrous; finally, the locality of S. oliganthes is well-embedded within the general range of S. sub- panic ‘ulatus, both being restricted to Sumatra. "rj E rthermore, characters that reportedly distin- guish Styrax oliganthes from S. subpaniculatus are not reliable or otherwise do not serve to delimit the two taxa. The densely pubescent abaxial leaf sur- faces in S. oliganthes differ from all collections of S. subpaniculatus known, but the degree of pubes- cence in S. subpaniculatus varies continuously from nearly none to nearly covering the entire surface. Variation in the amount of infraspecific abaxial leaf pubescence is common in species of Styrax, in- cluding several in this revision (e.g., S. hemsley- anus, 5. hookeri). Steenis stated that there are only white trichomes on the abaxial leaf surface of 5 oliganthes, but upon inspection at 64X magnifica- tion we observe scattered yellow, orange, and even brown stellate trichomes. The inflorescences of S. eame are reportedly few-flowered, unlike the many-flowered condition of the pseudoterminal in- donne of S. subpaniculatus. Only infructesc- ences, however, are known in S. oliganthes. Typi- cally, more flowers than fruits are borne on each reproductive structure in Styrax, and thus it is often difficult to infer the number of original flowers, or the structure and length of an inflorescence, from fruiting material. Furthermore, as in 5. subpanicu- latus, several pseudoterminal infructescences on the holotype of S. oliganthes are branched. We examined several other features not men- tioned by Steenis (1932) in considering the sepa- ration of the two species. The arms of the trichomes on the leaves abaxially average ca. 0.1 mm long in S. oliganthes versus those on most specimens of 5. subpaniculatus (averaging ca. 0.4 mm long), but close inspection of all collections of S. subpanicu- latus available to us indicates that trichome length is a continuously variable character. The leaves are ^4. thick-chartaceous in S. oliganthes whereas in most specimens of S. subpaniculatus they are membra- naceous, but one specimen in bud (Boeea 8857) has leaves that are nearly as thick as S. oliganthes and several more have leaves that are notably thicker than usual. The seeds of S. oliganthes are irregu- I arly rugose whereas those in S. subpaniculatus are smooth, but many species of Styrax exhibit infra- specific variation for this character (e.g., 5. Japon- icus). Ultimately, we can detect no distinctive char- acters upon which to base the separation of 5. oliganthes from S. subpaniculatus. The only reference made to collections of Styrax subdenticulatus in the al gorda is indicated. with '(T.)." an abbreviation for J. E. Teysmann. Accord- ing to Steenis (1932), i ans made three collec- tions of Styrax from the type locality cited in the protologue. Two of these are identified by Steenis as 5. paralleloneurus (J. E. Teysmann 963 and 966), and the third is specified by Steenis as the type of S. subdenticulatus (as “Teysmann 965HB |B, UJ” B in this case is BO, Herbarium of the Botanic Gardens, Buitenzorg, Java). Miquel’s herbarium was ~ U, but no indication of type status or any other annotation of Miquel exists on the U specimen of this collection. Although we have not seen the col- lections of S. paralleloneurus made by Teysmann from the type locality of 5. subdentic rs the two species are easily M adt with vegetative characters. For example, the leaf surfaces of S. sub- denticulatus are visible through the pubescence, whereas those of 5. paralleloneurus are not. Thus, a mistake in identification of these specimens by Steenis is extremely unlikely. On this basis, we feel confident that the Teysmann 965HB specimen at U is the holotype of S. subdenticulatus (and thus there is no need to lectotypify in this case). Additional e imens examined. INDONESIA. SU- MATRA. Aceh: Saurauja, Blangkedjeren, A. H. G. Alston 14716 (BM, L); Gajolanden, Goempang to Koengke, € G. J. Van Steenis 9802 zi ‚ L); Gunung Leuser Natl. 540 Annals of the Missouri Botanical Garden Park, from Kutacane to Be rar n, Kulam, near Agu- san, pass betw. Alas & Palok, Whitmore TCW33 (L); Gunung Leuser Nature Reserve, ше Машаа, 6 m SW from the mouth of Lau Ketambe, J. Wilde & B. E. E. de Wilde 15756 (BO, L); н Le user Nature Reserve, upper amas River Valle ey, ca. 15 Kutacane . 0. ). de Wilde & B. E. E. de Wilde 18342 (К, 1). Bengkulu: G. Kaba, near ree Angat, hot springs, H. O. Forbes 2866 (BM. GH, L). Sumatera Barat: Ayer mancior, O. Beccari 699 (BM, L); Pinang-Pinang plot, Ulu Gadut, M. Hotta 26604 (BO); Pesisir Selatan, 12 km W of Muarasako, Y. Laumonier YL5961 (K, L); Pajakumbuh, Mt. 6 Р Maradjo 87 (1), W. Meijer 3175 (BO); гүз umbuh, slope of Mt. = W. Meijer 3186 (BM, L). ipid ^) gla Res. Palembang, Pasemah Lands, near Miu . O. Forbes 2335 (BM, GH, I oeli, чүш. Sitoemba, preven 3 h 5225 (L); Таран, Angkola & Sipirok near Kampoeng Бабан Kola, t Re мып Institution b.b. 5249 (L); г Petjeren, "Forest Research Insti- tution b.b. 6854 (L); an, а Н. Н. Bartlett ., L); Kaban Djahe, A. H. Batten-Pool 5 (L); Adian Rindang, Asahan, vic n of Hoeta Tomoean Dolok, R. 5. Boeea 8857 (А, K, L, UC); Asahan (NE of Tomoean Dolok & E ‚ К. 5. Boeea 9285 (A, K, L, UC); S Tong- Karoland, J. dip а 3418 (L); NW Si- А. Lórzng 5641 (L); L); Bandarbaru, near . A. rri 14129 (L); E Mt. Sibajak, 7 « L); Gunung Leuser Natural Park, Bes qu River area, W. - te Fire il Forest Reserve, upper . J. O. de Wilde & B. E. E. de Wilde 21156 (L). Locality anki nown: ee Ajoeb 728 (L); Sumatra, si О. Forbes 2835 (BM), 7866 (L); Karo m nea e La bs & Kanpo eng Sigarang, Forest Reseai arc h area à b 8 (L); нр Langsdin, 1913, J. С. : qiu. atra, E Batavae, 1857, W. de Vise n (11): Е еа of Sumatra, H. S. Yates T (BM, L, 1467 (^, BM, IBSC, UC). ~ 15. Styrax supaii Chun & F. Chun, Sunyatsenia 3: 34. 1935 [as S. TYPE: China. Guangdong: Ruyuan Yaozu Zizhixian, pp sien Dun [Qi-xian-gou], 9 May 1934, B Kwok 80419 IE designated us IBSC!; isotypes, A!, IBSC “Supait’ |. Shrubs to 2 m tall or trees to 6 m tall. Young twigs brown or dark brown, densely stellate-pubes- cent; older twigs dark purple, glabrescent. Petiole 2—5 mm long. Two most proximal leaves on each shoot subopposite or opposite. Lamina 4—8 X 2—5 mm, chartaceous to thick-chartaceous, rarely mem- branaceous, ovate to obovate; base rounded to broadly cuneate; adaxially with a few simple or 2- or 3-armed to stellate trichomes when young, gla- brescent; abaxially sparsely stellate-pubesc ent, gla- brescent; margin coarsely serrate, deeply 3- to 5- dentate or lobed apically, lobes serrate-triangular or lanceolate, often remotely apiculate-serrate along the whole margin, up to 0.5 mm long; secondary veins 3 to 5 on each side of midvein; tertiary veins reticulate, adaxially plane, abaxially raised. Fertile shoots 2-5 cm long, 2- to 5-leaved. Inflorescences arising from shoots of the current growing season; lateral inflorescences 1- or 2-flowered; pseudoter- minal inflorescences 2—3 cm long, 2- or 3-flowered, rachis and pedicel sparsely short-stellate-pubes- cent, with additional long simple or 2-armed tri- chomes. Pedicel 10—15 mm long; bracteoles 3—4 mm long, linear or subulate, positioned at the base or middle part of pedicel. Flowers 1.5-1.8 cm long. ;alyx 5—6 (excluding teeth) X 4—5 mm, obconical; adaxially sparsely appressed-pubescent with long simple trichomes; abaxially with numerous simple or 2-armed trichomes ca. 1-1.5 mm long, stellate tomentum only sparsely distributed near the base, otherwise absent; margin distinctly dentate, the teeth 4—5 mm long, narrowly lanceolate or deltoid, Corolla 0.9-1.3 cm long, white, tube ca. З mm, glabrous, lobes 5, 14-15 X unequal, contiguous. 4—5 mm, lance-elliptic, adaxially sparsely pubes- cent with white 2- or 3-armed to stellate trichomes along the costae or distally, otherwise glabrous, abaxially densely stellate-pubescent. Stamens 10, conspicuously alternately unequal in length; fila- ments 4—5 mm long, straight, proximally broadened and white stellate-villous, distally attenuate and glabrous; anthers 4—6 mm long; connectives gla- brous. Style glabrous; stigma 0.2—0.4 mm wide, 1.0-1.5 х 0.7-0.9(-1.3) cm, ovoid or ellipsoid, apex apiculate to short-rostrate, punctiform. Fruit dehiscent; pericarp dry, 0.3-0.6 mm thick, outside longitudinally striate and rugose, rarely smooth, densely white stellate-villous, inside glabrous. Seeds brown, ovoid, smooth, glabrous. my Chun & F. Chun, Vae nia 3: ; Ни & Chun, lc. Pl. S pl. 248. 1937; : M. Hwang & C. J. Qi in W c. Cheng, Sylva Sin. 2: 1617, fig. 810. 1985; S. M. Hwang in F. H. Chen, Fl. Guangdong 1: 388, fig. 421. 1987; 5. M. Hwang, Fl. Reipubl. - Sin. 60(2): 96, pl. 32 (1-7). 1987; Z. Y. [C. Y.] Wu & P. H. Raven, Fl. China Ill. 15: 201, fig. 201 (1—7). 2000. Phenology. Flowering: May, June. Fruiting: June—November. Distribution. China (Guangdong and Hunan); Figure 5. Habitat. In mixed woods or near roadsides, and usually in relatively dry, disturbed habitats; 310— 900 m Vernacular name. Lie-ye-an-xi-xiang (Hwang, c Styrax supaii is known only from the mountain- ous regions of Yizhang Xian, Hunan Province, and Zizhixian, Guangdong Province, Ruyuan Yaozu Volume 90, Number 4 2003 Huang et al. Revision of Styrax Series Cyrta China. Based on the few specimens available for study, we believe this taxon must be a rare com- ponent of the vegetation. This distinctive species is easily identified by its long calyx teeth (4—5 mm) and a calyx covered with long simple trichomes (av- eraging 1.5 mm long), stellate calyx trichomes be- ing absent nearly throughout. It also can easily be distinguished from sympatric species by its coarse- ly serrate to deeply 3- to 5-dentate or lobed leaves apically, and stamens alternately differing in length by 1-2 mm Additional specimens examined. CHINA. Guang- dong: Ruyuan Yaozu Zizhixian, Daikiu [Da-qiao- ue Hou- zhi-e]. 1933 [5 June 1934; protologue], S. P. Ko 52797 (IBK, IBSC[2], PE). Hunan: Yizhang Xian, Mang. -shan, жт! -dong, 5. Q. Chen 3552 (IBK, IBSC); Mang-shan, B. € 86 (IBSC); Dong-shan-keng, Mang-shan, P H. Bis pe (IBK, IBSC); Rong-j n ng, Q. Lin 167 (IBSC); Mang-shan, Zhong-nan-lin-shi-xi-dui 94 (IBSC). 16. Styrax tonkinensis (Pierre) Craib ex Hart- wich, Apotheker-Zeitung 28: 698. 1913. An- thostyrax tonkinensis Pierre, Fl. Forest. Coch- inch. 4: t. 260. 1892 [as A. *Tonkinense" |. TYPE: Vietnam. Province unknown: Tu Phap, 12 May 1887, B. Balansa 4332 (lectotype, designated by Svengsuksa & Vidal (1992), P not seen; isotype, Р!). pros M E Bot. Jahrb. Syst. 31: 486. VT Yunnan: Simao Shi, eastern E 1600 n m, wi Henry 12006 (lec p des- ated here, K!; isotypes, A!, E!, IBSC! !, PES). Styrax subniveus Merr. & Chun, хаза ]: » 1930. TYPE: China. Guangdon echang Shi, May 1929, c L То 20732 ipn IBSC!; isotype, PE!). Trees to 30 m tall. Young twigs gray-brown stel- late-tomentose, older twigs dark brown, cent. Petiole glabres- 12(-15) mm long. Two most proxi- mal leaves on each shoot alternate. Lamina 5-18 X 4—10 cm, chartaceous to thick chartaceous, el- liptic to ovate; apex short-acuminate; base rounded to cuneate; adaxially glabrous except the major veins when young, glabrescent; abaxially gray or white stellate-tomentose, arms of trichomes very short, uniform, surface completely concealed by the tomentum; margin entire or apically 2- to 3-cre- nately toothed on young leaves; secondary veins 5 6 on each side of midvein; tertiary veins sub- parallel, adaxially plane or slightly sunken, abaxi- ally prominent. Fertile shoots (7—)10—25 cm long, 3- or 4-leaved. Inflorescences arising from shoots of the current growing season; lateral inflorescences l- or 2-flowered or racemose, 3—5 cm long, l- to 7-flowered; pseudoterminal inflorescences race- mose or paniculate, (5—)7—20 cm long, (6- to)8- to 18(to 23)-flowered, lateral branches 2 to 5, some- times with 2 or 3 lateral racemes from base of in- florescence, rachis and branches yellow-brown stel- Pedicel 5—10 long, yellow-brown stellate-tomentose; bracteoles 3-5 mm long, subulate or linear, positioned at the mid- dle of pedicel or base of calyx. Flowers 1.2-1.5(- 1.7) ст long. Calyx 3-4 X 2.5-3 mm, cupuliform; adaxially appressed-pubescent with white 2- or 3- armed or stellate trichomes; abaxially densely gray- white stellate-pubescent throughout; margin dis- tinctly dentate, glandular-dotted, the teeth 0.3- 0.7(-1.2) mm narrow-deltoid, usually contiguous or rarely separated by a shallow concave portion, unevenly distributed. Corolla 0.8—1.1(—1.3) cm long, white, tube 3—4 mm long, glabrous prox- imally, lobes 5, 10-15 X 3-4 mm, lance-ovate or oblong-elliptic, late-tomentose. mm long, white stellate-tomentose on both sides. Stamens 10; filaments ca. 4 mm long, straight, of equal width throughout, moderately to densely white stellate-villous throughout, some- times thinning apically; anthers ca. 5 mm long, as wide as filament; connective glabrous or short-stel- late-pubescent. Style glabrous; stigma 0.2-0.5 mm punctiform. Fruit 0.8-1.2 X 0.7-1.1 cm, subglobose, apex rostrate, irregularly dehiscent by З valves from apex; pericarp dry, 0.8-1.1 mm thick, outside nearly smooth, gray stellate-tomentose, in- Seeds brown or dark brown, ovoid, densely tuberculate, wide, side sparsely downy-stellate-pubescent. sometimes the tubercles arranged in stellate for- mations. Illustrations. Pierre, Fl. Forest. Cochinch. 4: t. 200. 1892; Anonymous, Ic. Cormophyt. Sin. 3: 338, fig. 4630. 1974 (as S. hypoglaucus); C. Y. Wu, Fl. Yunnan. 3: 424, pl. 120 (1—6). 1983; L. Yang in Y. K. Li, Fl. Guizhou. 2: 548, fig. 234 (5-7). 1984; S M. Hwang & C. J. Qi in W. C. Cheng, Sylva Sin. 2: 1603, fig. 798. 1985; S. M. Hwang, Fl. Reipubl. Popularis Sin. 60(2): 85, pl. 28 (8—13). 1987; J. О. Liu in L. G. Lin, Fl. Fujian. 4: 357, fig. 290. 1989; W. Q. Yin in Y. C. Xu, Ic. Arbor. Yunnan. 2: 892, pl. 470 (1—6). 1990; B. Svengsuksa & J. E. Vidal, ёк wi Cambodge du Laos et du Viétnam 26: 169, pl. 30, 4—7. 1992; Z. Y. [C. Ү.] Wu «Р. Н. Raven, Fl. s Ш. 15: 197, fig. 197 (8-13). 2000 Phenology. Flowering: April-July, September, November, December. Fruiting: January, February, April-December. Distribution. China (Fujian, Guangdong, Gu- angxi, Guizhou, Hunan, Jiangxi, Yunnan, and Zhe- jiang), Laos (Houa Phan, Luang Prapang, Phong- sali, and Xieng Khouang), and Vietnam (Bac Can, Cao Bang, Ha Tay, Lai Chau, Lao Cai, Ninh Binh, 542 Annals of the Missouri Botanical Garden 110 [130 А = = 2 | | e Styrax tonkinensis | | | | 0 400 800 | А ` Kilometers = кз ш. m" 45 L С * ш { d l / e › \ j 5 | E CUR К, y 4 | ge^ ) L3 | / У ' L , 5 > ў C = j \ 5 x А g Г РТИ pw et сё ү к M ( : | Е t p o? i " 30 " AA m | 1 À e. { LJ Duo v Ju Pee < S | , : о с QS ) : Spe ж l ° 25 {з Te t v^. n 1 TR OG 7 \ ONV 9| VA Q, л? 0 SR MT e ч pay Figure 11. Geographic distribution of Styrax tonkinensis. Phu Tho, Son La, Thanh Hoa, Tuyen Quang, and — Niu-you-shu (China, Yunnan; P Y. Mao 2941), Yen Bai); Figure 11. Habitat. and along edges of mixed forests in relatively dis- turbed sites; 30-2400 m Ba-fan-long (China, 94). Bai-bei-an-xi-xiang (China, Guangdong; Exp. Guangdong 587), Bai-bei-mu, jd hua-mu (China, Guangxi; S. Q. A63811), Bai-hua-lang (China, Guangxi: Y. X. Lin 16575), Bai-hua-lang-shu (China, Guangdong; W. T. Tsang 25866), Bai-hua-shu (China, Yunnan; W. X. Liu 566), Bai-hua-shu-guo (China, Yunnan: Exp. Wen-shan 259), Bai-hua-zhan (China, Guangdong; K. P. To et al. 12025), Bai-mai-an-xi-xiang (Hwang. 1987), Bai-ye-an-xi-xiang (China, Guangdong; Exp. Paese 70), Bai-ye-ye-mo-li (China, Hunan: Q. n 514), Da-qing-shan-an-xi-xiang (China, Gu- Hwang, 1987). Dian-gui-mo-li-hua (China, Guizhou; P. C. Tsoong 1094), (Hwang & Qi, 1985), Dou-zha-shu (China, Yunnan: P. Y. Mao 2468), Ge-jian-ge (China, Guangxi; Y. K. Li P1122), Jie-yong (China, Yunnan; P. Y. Mao 2779), Jing-guo (China, Y. К. Li P901), Mei-lu-zai (China, Guangxi; Exp. Guangxi 3468). In open forests оп mountain slopes. Vernacular Yun- W. names пап; Zhong angxi; Dian-gui-ye-mo-li Guangxi; Qing-shan-an-xi-xiang (China, Guangxi; H. N. Qin 245), Shi-chi-yang (China, Guangxi; F H. Xie et al. 3727), Tai-guo-an-xi-xiang (Hwang, 1987), Xiao- jie-yong (China, Yunnan; K. M. Feng 5149), Yue- nan-an-xi-xiang (Anonymous, 1974). Styrax tonkinensis is a relatively common com- ponent of primary and secondary forests and dis- turbed sites across southern China and the northern regions of Laos and Vietnam. The gray to white stellate-tomentose abaxial surface of the lamina (of- ten nearly glaucous in appearance) appears to be a constant character in S. tonkinensis, and serves to distinguish it from most other sympatric species of the imbricate members of series Cyrta. Three spe- cies occurring within the range of S. tonkinensis have at least some individuals with a stellate-to- mentose abaxial laminar surface (S. hookeri, S. lim- prichtii, and S. rugosus). Styrax limprichtii and S. rugosus differ from S. tonkinensis by their smooth and glabrous (vs. tuberculate) seeds, leaves with shorter petioles and more prominent teeth, a calyx with scattered orange or brown stiff stellate pubes- cence, and a fruit with a rounded or apiculate (vs. rostrate) apex and a longitudinally striate (vs. Volume 90, Number 4 2003 Huang et al. 543 Revision of Styrax Series Cyrta smooth or irregularly rugose) pericarp. Styrax hook- eri differs by its truncate, undulate, or irregularly lobed calyx with the teeth not contiguous if present, and the outer surface of the calyx within 1 mm of the margin more sparsely pubescent than the rest of the calyx and somewhat scarious. Styrax tonkinensis was described first as Anthos- tyrax tonkinensis Pierre from specimens collected in Vietnam (B. Balansa 4332 and 4358). (1902) may well have overlooked this taxon, be- Perkins cause S. macrothyrsus Perkins was published from one of the type collections of Anthostyrax tonkinen- sis (B. Balansa 4332), and no reference was made to Anthostyrax Pierre in Perkins's 1907 monograph. Realizing Perkins’s error, Hartwich (1913) made the transfer to Styrax. Perkins (1902) described 5. hypoglaucus Perkins from а specimen collected from Simao Shi, Yunnan Province, China (Henry 12006). Styrax hypoglaucus supposedly differs from S. tonkinensis by its 6- to 10-flowered racemose or sparsely branched inflorescences 5—6 cm long (vs. multi-flowered paniculate inflorescences 17-18 cm long; Perkins, 1902, 1907). Chun (1930) described S. subniveus Merr. & Chun based Merrill and on a specimen collected from Lechang Shi, Guang- 20732). They considered this species to be allied to 5. hypoglaucus and S. dong Province (C. L. Tso 2 tonkinensis. According to their protologue. S. sub- niveus has racemose or narrowly paniculate inflo- rescences 3—8 cm long with few to many flowers. We agree with Hwang (1980) that both of these species are synonyms of Styrax tonkinensis. The constancy of such features as the entire or weakly toothed leaves that are densely pubescent abaxially, relatively small flowers, dentate calyx, glabrous style, rostrate fruit, and especially the tuberculate seeds (which occur nowhere else in the genus) all serve to delimit this species. As in S. odoratissimus, inflorescence length and flower number per inflo- rescence exhibit notable variation in S. tonkinensis. This is reflected in the key to species, in which 5 tonkinensis falls out twice because of the variation in these characters. Suvatti (1978) cited Styrax tonkinensis from east- ern Thailand, but we have not seen any specimens of this taxon from that country. Styrax tonkinensis was introduced to the island of Java after World ar II for reforestation purposes (Backer & van den Brink, 1965). The holotype of Styrax hypoglaucus at B is pre- sumably destroyed. Perkins may have only seen the specimen at B because none of the other sheets of A. Henry 12006 that we have examined possess Perkins's annotation label. No herbaria are cited in either Perkins (1902) or Perkins (1907) to establish whether Perkins examined additional material. We have chosen the K specimen of A. Henry 10644 as the lectotype because Kew was the location of Hen- ry's headquarters. Selected | specimens gramina: CHINA. Fujian: Hua'an Xian, Xin-kou Tsoong 635 (IBSC, PE); Ji- anning Xian, Wu-yi-shan, from Hong-du to Pi-keng, H. Y. Zou 20266 (MO); Taining Xian, Xin-qiao- xiang, G. D. Ye 2137 (IBSC). Guangdong: Fengkai Xi Huang-gang-shan, C. Gaozhou Shi, Fen-zhi-ling, KU » PE) ee Shi, . G. Jip 36 Tang 25866 (А, CAS, E. BSC): ا‎ Shi. Le. ni shan, N. K. Chun 41251 (IBK); Lechang si Yang-guo- tian, Zhong- shan, S. P. Ko 54545 C, KUN, PE); Ge-c Карай; s Tang 2413 И РЕ); ем Xian, е auehan, near 19, us shut: H. С. Exp. СЕ 587 (IBSC); xiang, C. 41987 (IBK, IBSC KUN): Кашан Shi, San-tang-xiang, Long- Em C Wang 41442 (IBK. IBSC[2]. MO); Yangshan Xian, Wu-yuan-xiang, L. Tar 1082 (IBSC, KUN); Yingde Shi, Sha- kou- -xiang, Hua-shui- shan, P. H. Liang 84294 (IBK, IBSC); Zhaoqing Shi, Ding- hu-shan, б. L. Shi 13948 (IBSC). Guangxi: Bama Yaozu Zizhixian, Ling-lu-xiang, Y. K. Li P1122 (IBK, IBSC, PE); pm Xian, Song-shan- “xiang, Y. X. Lin 16575 (IBSC, PE); Bose Shi, Ba-ko-shan Ching 7398 (^. IBSC, PE, UC ) Cangwu Xian, Tong- luo-shan. (IBK, IBSC); Daxin Xian Xian, Huang-lian-shan, C. C. Chang 13769 (IBK, IBSC); Hezhou Shi, т "m ^. Chen et al. 500132 (IBK, IBSC); Jing n, Z. J. ü 1458 (IBK); Jinxiu Yao- zu Zizhixian, Ye: T an, на, C. Wang 39419 CAS, IBSC, L); Lingchuan Xian, Gong-ping-qu, А H. Xie et al. 3183 (IBK); [nec Xian, Yan-dun-xiang, C. F Liang 33787 (IBK); Lingyun Xian, Yu- Pes i ر‎ La , X. Q. Liu 28504 (IBK, E SC, KUN s ona Zi : 5. К. e au ey (A); dpi Gezu Zi Da. die -xiang, м -fu Coll. Team 306 (IBK, IBSC, MO. PE); Longzhou Xian, Da-qing- shan, C. C. Chang 11921 iet KUN); Nanning Shi, R. C. Ching 7957 (A, IBSC, PE); Ningming Shang-si- xiang, C. C. Chang 13025 (IBSC, KUN); Pingguo Xian, Na-lu-xiang, Y. K. Li P901 (IBK, IBSC, a Pingxiang Shi, йлн Institute of Botany 2 (IBK); Pubei Xian, Long-men-xiang, W. C. Chen 61 (IBSC); Wen Shi, San- wan-da-shan, 5. Q. ке 4141 (IBSC); Rongshui Miaozu Zizhixian, Ping-shi-xiang, Jiu-wan-da-shan, 5. Q. Chen 16565 (IBK, IBSC, KUN. PE); Rong Xian, Ta- tseh-t suen, A. N. Steward & H. C. Cheo 1085 (A, BM); Shanglin Xian. a 25360 (IBSC); А n-da-shan, W. ; Tianlin Xian, Mao-bi-liang, 7. T. Li 600853 BSc. KUN, PE); Xingan Xian, Liang-jin-kuang- xiang, Mao-er-shan, Z. Z. Chen 51517 (IBSC, KUN); Yong Xian, Sheng- ч bo g, J. Е Qin 700397 (IBK); Zi- ٤ Z. Z. Chen 51906 (IBK). Guizh- ‚Р, С. goo 1094 ЫЗ. Нипап: P. C. Tam 63707 (IBK); Dongkou Xian, Xue-feng- shan-qu, Ba-qu, Shui-wei, C. T Lee 2472 (IBSC, PE[2]): Hengshan Xian, Heng-shan, C. J. Qi 58 (IBSC); Jianghua Yaozu Zizhixian, An-ning, Hunan Forest Institute 6214 Ta-ming-shan, S. S. Annals of the Missouri Botanical Garden at uu) P C. T. Lee 2279 (IBSC); MS) Xian, Q. Z 1 (IBSC); Yizhang Xian, Mang-s shan Z. Lin 5 ‹ y et Zixing Shi, Ping-jiang- xiang, I Liang 80286 К. 9C, MO). Jiangxi: Dayu Xian, Zuo- bos xiang, M. Q. Nie et al. 9637 (IBK, IBSC, KUN); Shangyou m from Sheng-shui to Xi-long, Exp. Jiangxi 718 (PE). : Cangyuan Wazu Zichixian. Ban-hong- “xiang, r H. id 725 (IBSC, KUN); Funing Xian, Jar-gei, Wang 89220 (IBSC, KUN, PE); ias Daizu har Zi- zhixian, Xi-shan, China-USSR team hd SC. РЕ); Hekou Yaozu Zizhixian, Wu-tai-shan, W. X. n 566 (IBSC, KUN, PE); Jianshui Xian, H. T Tai 53147 (IBSC, KUN. PE i U >” PA — oo , Konta & Н. Takahashi CH3721 (КОМ); Jinghong Shi, . W. Wang 73643 (A, KUN). Jinping Miaozu Yaozu Daizu Zishixian HN ng-ping- xiang, China-USSR team 1510 Ñ >> mee = 2 a Sag C. W. Wang 74118 (A, IBSC, KUN, PE[2]); Mengla Xian, K. Wu et al. 289 (K CUN ); Pingbian pror Ziz Liang-zi- xiang, ` San-cha-he, P. Y. Mao 4083 (IBSC, KUN. M, Pu'er Hanizu Yizu Zizhixian, Marne A. Henry 13693 (A, E, K); Shizong Xian, S. C. Ho 85251 (IBSC); мны jak Lahuzu Wazu Bulangzu Daizu Zi- zhixian, ‚ Bang- tuo, J. S. Xing 1082 BSC, KUN, PE); Si- —. . К. M. Feng 22614 (IBSC, CUN); Xichou Xian, Lian- hua- tang, Jin-ping-shan, 5 Z. Wang 889 (KUN); Yanshan Xian, Pie-shih-eih, C. á Wang 84747 (KU N, PE); Yuanjiang Hanizu Ns Daiz Р Xi-gui-he, G. - Tao 38695 (KUN m usd , Fen-shui-ling, S. Ho 85159 (IBSC). Zhejiang: Longa Shi, Feng- Seen H. Y. Zou 454 (A). LA an: Tasseng de Samneua Muong de E (P, UC). Luang Prapang: NE de NR Nu g Prabang, E. Poilane 20726 (P). Phongs Poilane 26003. Xieng Khouang: km 220, ж Vinh & и E. Poilane 16779 (P). VIETNAM. Вас Can: : Kiet, E. Poilane 1831 (A, P). Cao Bang: pan t U. Kurosu s.n. (CAS ). Ha Tay: Da Chong, P + Pételot 5755 (A, P). Lai Chau: betw. Tsinh Ho & Mn Nua N of Lai Chau, Я Pon 25690 (P). Lao Cai: Cha- pa, P. А. Pételot 32: AS, Р, UC). Ninh Binh: Kho, Trung Giap, К jn 469 (P). Phu Tho: Phu Ho. P. A. Pételot 1033 (P, nes Son La: Pha Din, 1995, Kurosu & S. Aoki s.n. (CAS). Thanh Hoa: from Hoa Binh to Chobo, - Poilane 13018 (A, P). Tuyen Quang: Nui La, Ha Tuyen, А Fleury 37970 (P). Yen Bai: Bao Ha, E. Polani 23204 (P). Y Borel? 17. Styrax wilsonii Rehder, in Sarg., Pl. Wilson. 1: 293. 1912 [as S. “Шот” |. TYPE: China. Sichuan: Baoxing Xian, Mu-pin, 1300-1700 908, E. H. Wilson 884 (lectotype. designated here, A [accession no. 18452]!; iso- types, A[3]!, BM!, E!, K[2]!). m, June | Shrubs to 2 m tall. Young twigs densely ferru- gineous stellate-pubescent. Older twigs dark brown, glabrescent. Petiole < 2 mm long. Two most prox- imal leaves on each shoot opposite to subopposite. Lamina 1-2.5(-4) X 0.7-2(-2.5) ст, chartaceous, obovate, rhomboid, or rarely elliptic-ovate; apex acute to short-acuminate; base cuneate; adaxially sparsely stellate-pubescent along the major veins, otherwise glabrous; abaxially finely gray-white stel- late-tomentose, also with scattered yellow-brown or dark brown short stellate trichomes on major veins and the two most proximal leaves on each shoot; margin coarsely serrate or apically 2- to 4-dentate; secondary veins 4 to 6 on each side of midvein, adaxially slightly sunken, abaxially prominent; ter- tiary veins inconspicuous, plane or slightly sunken adaxially, faintly prominent abaxially. Fertile shoots 1-2.5 em long, 2- to 4-leaved. Inflorescences aris- ing from shoots of the current growing season; lat- eral inflorescences usually 1-flowered; pseudoter- minal racemes 1-2 cm long, 3- to 5-flowered, rachis yellow stellate-tomentose. Pedicel 2-3 mm long, yellow or brown stellate-tomentose; bracteoles 0.5-1 mm long, subulate or linear, usually posi- tioned at the middle of pedicel, sometimes those toward the base of the inflorescence leaf-like. Flow- ers 0.9—1.1(—1.3) cm long. Calyx 2-3 X 3-3.5 cupuliform; adaxially sparsely white appressed-pu- э mm, bescent with 2- to 3-armed or stellate trichomes; abaxially gray-white stellate-tomentose throughout, often also with larger scattered orange or brown stiff stellate trichomes, especially proximally; margin teeth narrow-deltoid, un- distinctly dentate, the evenly distributed, usually contiguous or rarely separated by a shallow concave portion. Corolla 0.6—0.8(-1.0) ст Pu rd tube ca. 3 mm long, glabrous, lobes 5 3.5—4 mm, narrowly oblong, adaxially ade us ent except at the apex, abaxially pale yellow stellate-tomentose. Sta- mens 10(12); filaments 4.5-5 mm long, straight, distally slightly attenuate, ventrally white stellate- pubescent, becoming glabrous distally; anthers ca. 3 mm long, wider than distal portion of filament; connective glabrous. Style glabrous; stigma ca. 0.2 X 0.4—0.5 cm, subglobose, apex rounded or apiculate, dehiscent; mm wide, punctiform. Fruit 0.5—0.6 pericarp dry, 0.2-0.3 mm thick, outside longitudi- nally striate, gray tomentose, inside glabrous. Seeds brown, ovoid to globose, smooth, glabrous. Illustrations. Prain, Bot. Mag. 148: t. 8444. 1912; F. T. Tai & T. C. Pan in W. P. Fang, Fl. Sichuan. 1: 420, fig. 162. 1981; А i Hwang, Fl. Reipubl. Popularis Sin. 60(2): 8 . 29 (1-6). 1987; Z. Y. [C. Y.] Wu & P. H. de Fl. China Ill. 15: 198, fig. 198 (1—7). 2000. Phenology. Flowering: May, June, September. Fruiting: April, September. Distribution. China (Sichuan); Figure 2. Volume 90, Number 4 2003 Huang et al. 545 Revision of Styrax Series Cyrta Habitat. mountain slopes; 700-1500 m Ai-mo-li (Hwang, 1987), Xiao-ye-an-xi-xiang (Hwang, 1980), In relatively sunny, open forests on Vernacular names. Xiao-ye-ye- mo-li (Anonymous, 1974). Styrax wilsonii is known only from middle (1000—1700 m) elevations of Baoxing Xian, Sich- uan Province, China. It is similar to the more wide- spread S. limprichtii in its shrub habit, scattered to dense orange or brown stiff stellate trichomes on the calyx, globose fruit with longitudinally striate pericarp, and flowering time usually before the full expansion of the leaves, such that we initially con- sidered whether the two species might be best treated as varieties of a single species. Styrax wil- ѕопи can be readily separated from 5. limprichtii, however, by its smaller leaves, flowers, and fruit. In addition, the abaxial laminar surface of 5. wilsonii possesses a tomentum of uniform height, whereas that of S. limprichtii possesses a layer of longer trichomes in addition to the white base tomentum, or is glabrous or nearly so. The apparent disjunc- tion between these two species is likely to be real rather than an artifact of inadequate collecting be- cause numerous collections of other species of Sty- rax have been made in the intervening areas of Sichuan Province. The morphological differences together with the discontinuous distribution provide sufficient evidence for treating S. /imprichtii and 5. wilsonii as separate species. We have seen four sheets from A of Styrax wil- Н. Wilson 884. Two of these indicate a collection date of June 1908, one a col- sonii labeled as Е. lection date of September 1908, and one a collec- tion date of October 1910. The protologue indicates that E. H. Wilson 884 is the type, but does not thus, must be regarded as syntypes. We have chosen the indicate a date of collection; these sheets June 1908 sheet with accession number 18452 as the lectotype because it possesses the best flower- ing material for examination. Furthermore, on the other June 1908 sheet (accession number 18453) is written "isotype" (with handwriting unknown but probably not Rehder’s). Thus, alternatively desig- nating 18453 as the lectotype would cause undue confusion. Additional yon — CHINA. Sichuan: Baoxing Xian, C. Pei 8120 (PE), T. P. Soong 9476 (IBSC, PE), 39476 (KUN), T. ^ Yü 1903 (IBSC, PE); Liang-he- kou, X. B. Zhang & Y. X. Ren 4507 (PE); Wu-long. X. B. Zhang & Y. X. Ren 4534 (PE); Ming-ling-xiang, Zhuang- zi-he-ba, X. B. Zhang & Y. X. Ren 4640 (PE); Yan-bi- cun, X. B. Zhang & Y. X. Ren 4957, 4982 (PE). ExcLUDED NAME Styrax bashanensis S. Z. Qu & К. Y. Wang, Bull. Bot. Res., Harbin 9(1): 27. 1989. TYPE: Chi- na. Shaanxi: Zhenping Xian, 1190 m, 28 May 1976, K. Y. Wang 548 (holotype, NWFC lost). We have located no authentic material referable to this name; the type is missing at NWFC. The description is consistent with the characters exhib- ited by some specimens of S. hookeri distributed near the periphery of this species’ range H. Song 272 and 907, C. Wang 41180; мга lance-elliptic, subcoriaceous leaves and/or relative- ly small fruits ca. 7 mm wide) where Wang's col- lection is located. There is sufficient uncertainty in the nature of these characters, however, to preclude the placement of this name in synonymy. Literature Cited Anonymous. ase Miscellaneous. J. Jap. Bot. 16: 56—00. 1974. Iconogra "p 1а pol Sinicorum, Vo $ K'o 0 ste ch’u pan she, Beiji Aoki, r 1982. реш omi -like sec uu instar larvae of Ps eudoregma shitosanensis (Homoptera, my var a) found on its pr ин ost. Kontyû, Tokyo 50 5: — —— & U. Kun 1993. The gall, soldie nomic position "n the aphid Tuberaphis taiwana moptera). Jap. s eom 61: 361—3 ‚ T. Fukat фе Н. Ishikawa. 1998. Cer- ataphi jamuritsu, a Eins al aphid producing sol- diers in large, hard galls. К йош. Sci. 1: 327—333. г C. M. 1961. New rs ada taxo- Ho- — Arora : chromosome report. П. Bull. Bot. Surv. India 3i 37 Backer, C. A. . C. B. van den Brink, Jr. 1965. Flora of Java, Vol. 2. N. V. P. Noordhoff, Groningen, Nether- ands. Burkill, I. H. 1966. A Dictionary of n Eco 'onomic Prod- ucts d the Malay Peninsula, Vol. of Agri- culture and Co-operatives, Kuala a ur. Cavanilles, ER Septima Dissertatio Botanica. F. A. Didot, Pari Chang, H. T. 1962. The characteristic of poros flora. Acta Sci. Nat. Univ. Sunyatseni 1: 1 ————. 1994. The characteristic of di aste dd flora of Taiwan. Pp. 129-145 in Н. T. Chang (editor), do Zhongshan Univ. Press, Mw ou. Clarke, C. B. 1882. Styraceae. 2-590 in J. D. oder (editor. Flora of British 2. Vol. 3. Lovell ndon. Copel: a. T F. 1938. The Styrax of northern California and the relationships of the Styracaceae. Amer. J. Bot. 25: 771-780. Dickison, W. C. 1993. Floral anatomy of the Styracaceae, including observations on intra-ovarian trichomes. Biol. J. Linn. Soc. 112: 223-255. Docters van Leeuwen, W. M. 1922. Ueber einige von Aphiden an Styrax-arten ее Gallen. Bull. Jard. Bot. Buitenzorg 4: 147—162 Duke, J. A. 1985. CRC Handbook of Medicinal Herbs. CRC Apa ер Raton, Florida. Fang, W. P. . Icones Plantarum Omeiensium, Vol. (1). E ur ias University, Chengdu. Annals of the Missouri Botanical Garden ips her, E R. 1938. Contributions to the flora of Siam. Р р Kew 1937: 505-51( s jm P W 1996. Isozyme ins of intercontinental disjuncts within Sb rax (Styracaceae): Implications for the e Tethyan hypothesis. Amer. J. Bot. 83: 342-355 1 1997. A revision of Styrax (Styracaceae) for west- 1 Texas, — Ann. ee ie en . 1999. P V ny of Styrax based on morpholog- ical c фы. ‘ters, with implic ations is Es us infrageneric classification t. Bot. 24: 356—378. ————. 2001. Phylogeny and oes P the flow- ering plant genus Styrax racac based on chlo- roplast DNA restriction sites and DNA sequences of the ipod icon 'ribed spacer region. Molec. 108. Evol. 87-4 sens nd Mesoamerica. Missouri Bot. and Phylogenet. ——. jd press a. Styracaceae. /n: K. Kubitzki (editor), The Families nie Genera of Flowering Plants. Springer- Verlag, Berlin. —— — —. In press b. New species and taxonomic changes in Styrax (Бред асеае ) cn M а а. Novon. ————,‚ C. M. Morton, Chen . Meldrum. 2001. Pls ny and TT of he Sty racaceae. Int PI. Sci. 162 (6 Suppl.): 595-81 Gibbs, R. D. 1974. emat of the Flowering Plants, 4 vols. dapes Ys Univ. P Gonsoulin, G. J. 1974. A revision of Styrax (Styracaceae) in Ned America, { entral America, and the ( Sida 5: 191—258. Greuter, W McNeill, F. R. Barrie, H. M. Burdet, Demoulin, T. S. Pileusitàs. D. H. Nicolson, P. C. Sin J. E. Skog, P. Trehane, N. J. Turland & D. Kiih (editors). 2000. International Code of Botanie al Nomenc са = Louis Code). Regnum Veg. 138. Hartwich, € . 1913. Ueber die Siam-Benzoe. Apoth- d адаы 28: 698-6099. 9602 . Che ^motaxonomie Caribbean. der Pflanzen. Vol. © — The Genera of Flowering Plants. Clarendon P jp LN rd. 198( Hwang, 5. M. A preliminary study on the family Styracaceae from б. Acta Phytotax. Sin. 18: 154— 67. . 1983. New taxa of Styrax from China. Acta Bot. Ruso Sin. 1: 7 K 15-1. ———, . Styracaceae. Pp. 77-150 in Flora Reipub- licae кы Sinicae, Vol. 60. Science Press, Beijing. ————. 1999, Systematic Rn and geographic. distri- bution of Styracaceae. Pp. 319-331 in A u (edi- or), The ( казые i. of nd Families and c Science Press, Beijing. —— & Саш . 1996. Styracaceae. Pp. 253-211 in Z. Y. Wu & P.H. Rave n (editors), Flora al China, Vol. 15. Science Press, Beijing, and Missouri Botanical Gar- I dp ys —————. 1907 Kurosu, U. & S. Aoki. 1990. Formation of a "Cat's m by the eun "бетщ nekoashi ае sd ›. J. Entomol. 166. & p "In cipient galls of the soldier- ue ing арй СЕТА bambusae (Нотор- tera). A ap. J. a 59: 663—609. = J7. Cerataphis vandermeermohri (Home im га), : n | aphid with е falling off their huge gall ›. J. e e 65: 278- === еее & S. . 1998. с E 'rnation of two tropical gall- forming “хи ‚ Astegopteryx styra- cophila and A. pallida (Homoptera). Entomol. Sci. 1: 21—26. Langenheim, J. Н. 2003. Plant Resins. Chemistry, Evo- lution, Ecology, and Ethnobotany. la nd, a „ В. 1989. Illustrated Flora of Korea. Hyangmunsa, Timber Press, Port- 996. Colored Standard Illustrations of Korean AK ademi Sojok, Seoul. 1996. Flora of Korea. Kyohak, Seoul. "от заснеме ds 106— Г in H. L. Li, T. s. T . Huang, T. Koyar Jevol (editors), Flora 7 Tainan Vol. 4. E poc h Publishing. ui Long, D. 9. Styracaceae. Pp. 578-57 . J. C. Lei ^ 4 G. Long, Flora of SS Roy- al Botanic Garden Edinburgh and Royal Government of U.K. Bhutan, к shield, Manshard, E. 1936. Embryologische Untersuc nes nan Styrax on m et Zucc. Planta 25: —383. Masamune, G. Suzuki. 1933. A of as lec ton in the ‘and of Kizan. Ann. Rep. Taihoku Bot Gard. 3: 49-76. Mehra. P. N. 1976. Cytology of Himalayan Hardwoods. Sree juae Press, с ша. & list К Bawa. 1969. Chromosomal evolution in Кыт г haw Evolution 23: 466-481. errill, D. Chun. 1930. Contributions to our olde P ^d S dune flora. Sunyatsenia 1: ers, J. 1859. On the natural order Styraceae, as distin- E from the Symplocaceae. Ann. Mag. Nat. Hist., ser. : 394—404. Nakai, 1 T. 1922. Trees and Shrubs Indigenous in Japan Proper, Vol. . Seibido Shoten, Nihombashi, Tokyo. 1938. Nota A plantas asiae orientalis (V). J. 14: 629-€ 1902. eae zur Kenntnis der Styracaceae. 78—488. Jap. Bot. Perkins, J. 30 2i racaceae. Pp A. Engler (e itor), Ба Pflanze nreic a "m E s "x 30. Enge жые Leipzig. . 1910. Neue abd aceae aus Ostasien I, Керем. . Nov. n Veg. 8: 82-8: den, St. Louis. B R. ¢ Jr. 1951. Pharmacognosy. J. «с 1985. Styracaceae. Pp. 1596-1650 in B. Lippincott, Philade phia. W. €. Cheng MNA Sylva Sinica, Vol. 2. Chinese For- Putz, F. E F. S. P. Ng. 1978. Styracaceae. Pp. 262-264 estry Press, Beijiı in F. S. Р. Ng (editor), Tree Flora of Malaya, Vol. 3. 18. Kato, E. & T. Hiura. 1999. Fruit set in Styrax obassia Styrac aceae): The effect of light availability, display size, and local floral density. Amer. J. Bot. 86: 495- 501. > ee Koidzumi, G. 1937. Contributiones ad a m florae asiae orie malis. Acta Phyto : 210-244. . 1941. Contributiones ad оо m florae asiae MEE ke ‘ta Phytotax. Geobot. 10: 54-63 ах. Geobot. Rannon, J. C Press, C ЕБ ————. 19 Longman, London. 1992. И А comprehensive review of ntal enum Amer. Nurseryman, No- E 992: 23—36 a f. uie l, Rehder, A. 1912. Styraceae. Pp. 289-296 in C. S. Sargent (editor), Plantae ie Vol. 1. Cambridge Univ. Massachusetts. . New species, varieties and combinations Volume 90, Number 4 3 Huang et al. 547 Revision of Styrax Series Cyrta from the herbarium and the collections of the Arnold Arboretum. J. Arnold Arbor. 11: 153-1 Ridley, H. N. 1930. The Dispersal of Plants throughout the э Lovell Reeve, Ashford, Kent. Saraiva, L., O. Cesar & R. Monteiro. 1988. Biologia da ы is e sistema de reprodução de Styrax campo- rum Pohl e S. ferrugineus Nees et Mart. (Styracaceae). Revista рее; Bot. 11: 71-80. . А new form of Styrax japonicum. J. Geo- Siebold, о k von. 1835-1841. Flora Japonica. Lugduni batavorum, pm Smith, W. W. 0. Notes on certain Asiatic Styracaceae. Notes Roy. E а Edinburgh 12: 231-236 Steenis, C. G. G. J. van. 1932. The Styracaceae of Neth- erlands India. Bull. Jard. Bot. Buitenzorg sér. 3, 12: : — J49. Styracaceae. Рр. 50-56 in С. G. G. J. v Steenis io, Flora Malesiana, series 1(4). Noordhoff. Kolff N Batavia Stern, D. L. Ея А пиено analysis of soldier evo- sb in the aphid family Ba ете Ргос. Ңоу. inu ser. B, Biol. Sci. 256: 203-209 5. Phsloeanetie а аа that chide, rather than n determine gall morphology. Proc. Roy. Soc. London, ser. B, Biol. Sci. 260: 85—89 . 1998. Phylogeny of the tribe Cerataphidini (Ho- bus and the evolution of the horned soldier aphids. — p P 155-165 . Foster. 1996. The evolution of soldiers in ahi. КҮ Rev. 71: 27-79 . Aoki & U. Kurosu. 1997. Ретона aphid axonomic affinities and life cycles with molecular data: A case study of the tribe Cerataphidini (Hormaphidi- Aphidoidea: Hemiptera). Syst. Entomol. 22: 81— dae: c D Stuessy, T. F. 1990. Plant Taxonomy. The Systematic Eval- uation of Comparative Data. Colombia Univ. Press, New ork. Sugden, E. A. 1986. Anthecology and pollinator efficacy of Styrax (pns subsp. redivivum (Styracaceae). Amer. J. Bot. 73: 919—930. Suvatti, C. 19 78. "Flora of Thailand, dittayasathan, Bangko Syengsu sa, B. K. K. J. E. Vidal. 1992. Styracaceae. . 145-195 in P. Morat (editor), Flore du Cambodge di Laos et du Viét étnam, Vol. 26. Muséum National d'Histoire Naturelle, Par Tamura, S Hiura. 1998. Proximate factors affecting fruit set and seed mass of Styrax obassia in a masting O7 Vol. 1. Ratchaban- 1981. Pp. 409-433 in W. P. Pans Later Flora Sic bns anica, Vol. 1. Sichuan People Press, Chengdu. Tiffney, B. H. 1985a. The Dp North Atlantic Land Bridge: Its importance in ‘Tertiary and modern. phyto- geography of the Northern Hemisphere. J. Arnold Ar- bor. 66: 243-273. . 1985b. Perspectives on the origin of the floristic Eis between eastern Asia iid eastern North America. J. Arnold Arbor. 66: 73-94. 0. Ge 'ographic and climatic influences on the Te eous and Tertiary history of pinus 'rican floristic similarity. Acta Univ. Carol. Geol. 44: 5-16. & S. R. Manchester. 2001. n influence of phys- ical environment on phytogeographic continuity and асеае. phylogeographic ени in the Northern Hemi- sphere. Int. J. Pl. Sci. 162 (6 Suppl.): 53-517. Wallnófer, B. 1997. A revision of Styrax L. section Pam- Walln. (Styracaceae). Ann. 720. Wen, J. 1999. Evolution of eastern Asian and eastern North American disjunct distribution in flowering plants. ED Rev. Ecol. Syst. 30: 421—455. Wolfe, J. A. 1975. Some aspects of plant geography of the Northern A ea during the Late Cretaceous and Tertiary. Ann. Missouri Bot. Gard. 62: 264-279. Wu, C. Y. 1983. Styracaceae. Pp. 406-437 in Flora of Yunnan, Vol. 3. Science Press, Beijing. Yamazaki, T. 1993. Styracaceae. Pp. 104-106 in K. Iwat- suki, T. Yamazaki, D. E. Boufford & H. Ohba (editors), Flora of Japan, Vol. 3a. Kodansha, Tokyo APPENDIX 1. List of species. l. Styrax buchananii W. W. Sm. 2. Styrax chrysocarpus H. L. Li 3. Styrax curvirostratus (B. Siena) Y. L. P. W. Fr 4. Styrax nd Diels 5. Styrax hookeri C. B. Clarke 6. Styrax japonicus Siebold & Zucc. 7. Styrax limprichtii Lingelsh. & n Huang & 11. Styrax porterianus С. 12. Styrax rugosus Kurz rax shiraianus Makino 14. Styrax subpaniculatus Jungh. & de Vriese 15. Styrax supaii Chun & F. Chun l6. Styrax tonkinensis po Craib ex Hartwich 17. Styrax wilsonii Rehde APPENDIX 2. Index to exsiccatae. All specimens examined by the REDE are listed al- phabetically by collector, followed by collection numbers correspond to those more than two persons participated in s collection, аі the first collector listed on the label is cited. 236 Team 641 (10); 1251 (10); vf (10); 1944 (5). Ajoeb 728 (1 4). T. Akagi in 1985 (6). A. кеш in jas (6). C. d'Alleizette i in 1908 (6): s.n. (4); s . А. бх 14716 (1 Amano 6962 (6 "€ jer е1 a 192 (6). H. Ando in 1965 (6); in 1967 (6 ). Anony- mous Ja Ps (IBSC); 3d (9) (IBSC); 1—32 (5) (KUN); 53 (16) (PE); 66 (4) (PE); 7824-70 (10) (PE); 86 (10) (IBK); 87 (6) (PE); 96 (6) (BM): 101 (6) (BM); 124 (16) (PE); 144 (16) (PE); 184 (5) (KUN): 186 (10) (PE); A); 195 (5) (KUN); 201 (6) (IBSC); 240 (6) (PE); 250 (9) (PE); 252 (6) (E); 265 (9) (PE); 273 (10) (PE); 279 (6) (C); 284 (9) (PE); 294 (10) (PE); L297 (6) (PE); 334 (6) (PE); 345a (6) (PE); 400 (10) (BM); 490 (6) (KUN); 522 (6) (KUN); 74-522 (6) (IBSC); 528 (6) (РЕ); 550 (10) (PE); H.III589 ; 784 (5) (KUN); D941 (6) (PE); 1160 ©) (PE): 1189 (11) (K); 1326 (16) (IBK); 1344 (5) (E): | 2 (5) (Е); 1681 (16) (IBK); H1795 (4) (A); 83-2052 (10) (PE); : (6) (BM, E); 2161 (6) (E); 2162 (6) (Е); 2163 (6) (E); 2337 (9) (BR); 2346 ^ (A); 2705 (6) (KUN); 2742 (6) (PE); 3544 (6) (PE); 3584 (6) (PE); 3746 (1) (K); 4499 (16) 1N - =, о © ~ = м rm — 548 Annals of the Missouri Botanical Garden (IBK); 5093 (6) (IBSC); 6061 (9) (A); 6473 (16) (IBSC); 7047 (16) (IBSC); 10153 (6 Д 5 11840 (9) (А, МО); 12835 (9) (MO); 27495 (6) (PE); 31010 (9) (IBSC); 40225 (16) (P); 69965 (6) (IBK); PAD (10) (KUN); 90244 (16 (IBSC); 18515037 (10) (PE); 8521239 (10) (PE). S. Ari- moto in 1903 (9). L. Averyanov et al. VH4544 (: B. Balansa 4332 (16); 4339 (16); 4358 (16): 4365 (16): g oa s . Y. Bao 1 (7); 4 (5); 169 (6); 175 (6); 217 (6); 391 (5). S. P. Barchet in 1906 wu B. ИР аи T a H. H. Bartlett 8077 (14). A. H. Batten-Pool 5 (14). R. K. Beattie & Y. Kurihara 10753 (6); 10814 (6). О. Вессап 699 (14). Beijing Youth Team (Guizhou, 1986) 54 C. К. van Beusekom М = Phengkhlai 1078 (12); \ 96 (9). J. Bisset 4605 (9). von Rosthorn pe (6). E. M. Bodinier in 1902 (6); s.n. (6); 1099 (10); 2221 (6). R. S. Boeea 8857 S eran 4). P. H. F. Bon 338 (10). A. id 8 (16); . M. Borel 1 (16); 2 (16); 3 (16); 7 (16); 8 (16); 13 s ee 17 (16). D. E. Boufford & B. B. 24085 (6); 24853 (6). D. E. Boufford & E. W. Wood 25412 (9). D. E. Boufford et al. 25729 (6); 25808 (6); 26287 (6). F. S. A. Bourne in PA (6). H. S. Bowes 3199 (5). P. W. as & P. W. r 131 (€ 2 wW. P. Brooks 511 (9); 52511 . M. Buc hana 21 (1); 25 (12); 51 (1). Me T Cai 38 (10); 464 (6). K. H. Cai 850 (16); 1089 6). T. R. Cao 90621 (10). Z. Y. Cao 191 (6 3 29 (6); 90 (6). J. Cavalerie 997 (6); 1062 (€ ME 4526 (6); 8190 (€ ›). С. Н. Cave in 1912 (5); in 1913 (5); in 1917 (5); in 1919 (5 5); in 1922 (5). J. G. Champion 138 (10). D. پا‎ 5481 (6). T. R. Chand 1846 (5); 7065 (5); 7586 . С. Chang 602 (16); 11921 (16); 13025 (16); 13709 ^de 13901 UT 13910 i C. E. C hang 29. 50 (6); 5 1 — O x ` = £5 - — ое © ш тю ү Со PEN w Z — S. Y. С ; 1740 (6); 3319 (6); 5178 (10); 5822 Th 6258 (10); 7 m: (10); 8: j^ 4 (10). Y. L. Chang 2512 B. Y. Chen 2949 (10). C. Chen 1520 (10). G. R. A 2368 (10); 2442 (9). H. С Chen et al. 500072 (10); 500132 (16); 500159 (10). L. X. Chen 2500132 (16); 500159 (16). M. Chen 1061 (10); 1 176 (10). S. Chen 519 (9). S. Q. Chen 674 (10); 2889 (8); 3431 (9); 3552 (15); 3573 (10); 4141 (16); 4875 (16); 5408 (8); 5649 (10); 10192 (16); 12850 (16); 13215 (16); 14263 (6); 14376 (5); 14420 (16); la (6); 14692 (10); 14709 (6); 15240 (6); 15255 (16); 15281 (16); "pe (6); 15367 na 15908 (6); 16408 (6); Te (16). S. Y. Chen 5649 0). T. C. Chen 410 (6): 514 00) 885 (16): 580 (6 1028 (10). W. C. Chen 61 (16). Y. Chen & B. Bai 562 (5). Z. L. Chen 30585 (16); 30601 (10); 30603 (16); 30605 (10); 30610 (8); 30613 (10); 30614 (10). Z. Z. Chen 50892 (10) 50893 (10); 50983 (6); 50995 (10): 51016 (10); 51055 (6); 51074 (6); 51257 (6); 51517 (16); P51517 (16): 51906 (16); 52034 (6); : iiis (6); 52659 (10); 53822 (10). W. Cheng 103 (10). W Cheng in 1937 (8); s.n. (8); 2926 (6); 3732 (10) m (10); 6198 (5); 6332 (5); 6540 (5): ; 10441 4 10638 (4); 12008 (4); 11022 (4). W. C. Chen БИ . Hwa 559 (6); 662 (6); 975 (6). X. Cheng et al. ) Y. Q. € deu ME (10); 170129 (10); 0202 по, КУУТ Н. C. Cheo & W. F. Wil- son 229 (10). K. H. Cheo C302 un 38674 (3); reed (16). C. C. Chi 5256 (10). C.-Y. Chiao 1634 (5); 2715 (6); 2800 (6); 2848 (6). C. P. Chien 623 (16). Chin & Shun 80 (4); 137 (4). Ching: USSR team 18 (12); 46 (16); 217 (12); 346 (6): 431 (6): 832 (6); 1510 ¿ 1713 (10); 1853 (10); a ni od (16); pui (5); 3070 (16); 3726 (16); 5570 (5); 5610 (16 6268 (5); 8531 (4); 9556 ? (16): 9688 (16); 185312 (10), A. J. B. Chevalier pn Vietnam Exp. s.n. (16). Ching & Shun 80 (10). R. Ching 137 (4); 1415 5 (10) 1434 (10); 1622 (10); 1809 am 2080 (10); 2241 (10); 2911 (10); 3253 (9); 3273 (6); 4699 (10); 4825 (6); 5966 (6); 7096 (16); 7160 (6); 7398 (5); 24523 (7); 24887 (7). R. C. Ching & C. L. (10); 485 (10). L. H. Chiu 50078 (16). C. : (4). H.-C. Chow 832 (6); 7547 (10); 8016 (10). 1266 (10); 2179 (6); 2963 (5). T. S. Chu 60946 (10). D. C. Chun 414 (10); 885 (16); 886 (6). N. K. Chun 41130 (16); 41251 (16); 41677 (10); 41913 (10). mo 3699 (6); 3707 (6); 3708 (6); 4081 (6); 4159 (6); : (6): 4990 (10); 5079 (10); 6051 (10); 7369 (16); 9744 ( io. 9808 (10); 10623 (10). Chung In Cho 8276 (9). H. H. Chung 1345 (10); 1867 (10); 2615 (10); 2667 (10); 2742 (10); 2867 (10); 3438 (10); 7003 (10); 8455 (10). Z. S. Е ‘hung 81981 (10). С оа Lijiang-Dali Exp. 1446 (5). С. агке 728B (5); 26889B (5); 27995B (5); 349441) (5); 13631A (5). Coll. Team for Oil Pl. 650302 (7); ча (7). C. B. Collett 800 (12). С. Congdon 507 (11). J. d. 1171 (6). P. Courtois 25676 (10); 28596 no. 36304 (6). J. M. Cowan s.n. (5). Cultuurtuin van Techn- ische Gewassen in 1936 (6). C. Curtis 1538 (11). L. Y. Dai & 771 (6); 1484 (6); s (6) T. L. Dai 1296 (4); 1511 (4); 100551 (4); 103315 (6); 105634 (4). J. M i (6). Danish s (1958/1959) 3295 (12). K. Deg guchi 4819 (6); 5737 (9). K. Deguchi & S. Tsugaru 3819 (6). J. Delavay in 1883 (7 ); 1017 (7); 2536 (7 y 2782 (7); : (7); 4354 (7); 4394 (7). R. P. Delavay s.n. |. 2018 (6); 2505 (6). M. B. Deng 4134 (9); 4223 (9); : (9); 4768 (10); 4803 (10); 11153 (10). M. P. Deng & К. Yao 79022 (6). X. F. Deng 4 (5 Ji Ds 361 (5). P. Di 60022 i F. G. Dic о 6008 (12); 8750 (12). M. Dickins s 1877 (6). X. Y. Dong & Y. * Xiong 93565 (10). P. Dors sett & W. : Morse 719 . Dransfield 3418 (1 "n F. Ducloux 689 (6); 2291 (€ E 2716 (€ »); 2717 (6); 4626 (5); 4627 (7); 4951 (5). S. T. Dunn 2897C (10). ina. Work Station 6855 (6); 7007 (9). Y. 77 (16). G. E. Edafio 79248 (6). H. J. Elwes & K. Watanabe . € 2% ‚ Ёз хр. Anhui 59 (9); 219 (9) 359 (9); 423 (9); Үү jos 2344 (10); 2376 (10). Exp. An-shun 70 (6); 660 (6); 890 (5); 1353 (6). Exp. ee 358 (6); 847 (5); 491 (6). Exp. Da-yao-shan 11 хз An Qe 14243 (10). Exp. d 88155 jin-shan & F uang-shan 31159 6 400532 (6); 400566 (б); 400838 (6); 400911 (6); 401959 6); 402061 (6); 402110 (4); 402476 (10). Exp. Gao-li- gong-shan (1997) 9518 (5). Exp. george 70 (16); 144. (16); 249 (6); 445 (16); 587 (16); 1244 (10); 1265 (6); 5185 (8). Exp. Guangxi 455 (10); a r^ 3468 (16); 3627 (16). Exp. Guizhou 2924 (10); 3034 (10); 4042 (6); 4481 6); 4737 (5); 4809 (6); 6836 (5); 7361 (6). 711 (6). Exp. Henan 714 (6); 868 (6); 945 (6); 1254 (6); 1405 (6); 1511 (4); 1905 (6); 2188 (6). Exp. Hokkaido EHOK105 (9). Exp. Hong-shui-he 89-999 (10); 1085 (10); 89-1109 (6); 2065 (6); 2336 (6); 2356 (5); 2943 (6). Exp. Hubei 14022 (6). Exp. Hunan 281 (6); 614 (6). Exp. Hu- nan & Guizhou 2626 (10); 3279 (6); 3802 (6). Exp. Ji- angxi. 377 (6); 718 (16); 1411 (6); 1652 (6). Exp. Jin-fo- shan 477 (6); 1205 (6). Exp. Li-bo 1115 (6); 1188 (6); 2240 (6); 2248 (6). Exp. Long-sheng 55 (16); 151 (6). Exp. ja chun 43 (16); 803 (6); 866 (16); 902 (16); 1194 (16). xp. N Guizhou 373 n ДО (6); 1360 (6): 1588 (6): 2046 e Exp. Nan-ling 55 (10); 272 (10); 560 (10). Exp. NE 7A Z - + — ~ p Volume 90, Number 4 2003 Huang et al. 549 Revision of Styrax Series Cyrta m c ); 309 (5); 568 (5); 905 (6); 1161 (4); 1163 . NW Yunnan 4010 (7); 6389 (7). Exp. Qinghai Xizang 20 (7); 517 (7); 638 (7); 692 (7); 74-4029 (5); 6603 (5); 7245 (5); 9653 (5); 11362 (7). Exp. Qing-ling (No. 3 Team) 968 (4). Exp. 5 China 1996 (16); 2658 (10). Exp. S Guizhou 205 (5); 958 (6); 1104 (6); 1274 (6); 1590 (6); 1753 (6); 2008 (6); 2102 (6); 2182 (6); 2745 (6); 2902 (6); 3615 (6). Exp. Sang-zi 737 (4). Exp. SE Guizhou 50113 (6); 50609 (6); 50741 (6); 50915 (6); 50919 (6); 51245 (6). Exp. Sichuan & Guizhou 123 (6); 192 (6); 415 (6); 1774 (6); 1860 (6). Exp. SW China (Guizhou, Sichuan, Yunnan) in 1965 (7); 200 (7). Exp. W Hunan 81 (10); 495 (6); 1087 (6). Exp. Wen-shan 65-138 (16); 60243 (106); 60259 (16); 68275 (16). Exp. Wu-ling-shan 40 (6); 224 (6); 616 (6); 697 ( (6); 772 (6); 912 (6); 1989 (6); 2345 (6); 2598 (6). Exp. Wu-yi-shan 11 (10); 160 (10); 80-261 (6); 80-472 (6); 912 (6); 932 (6); 1624 (6); 1812 (6); 2409 (10); 6839 (10); 400668 (6); 400829 (6); 400912 (6); 401195 (10); 401260 (6). Exp. Yu-xi 2351 (5); 2373 (5): 2992 (6); 3055 (6); 3067 (6); 89480 (6). Exp. Zhan-jiang 2909 (16); 3648 (16). Exp. Zi-yun-shan 272 (6); 412 (6 910 (6); 932 (6). Rev. E. Faber 195 (10). C. S. Fan & Y. Y. Li 221 (10). W. D. Fan 79 (16); 179 (16). M. Y. Fang 23912 (6); 24815 (6). W. P. Fang 942 (4); 1056 (6); 1133 (4); 1376 (4): 1401 (4); 2225 (4); 2462 (10); 2636 (5); 2787 (5); 2873 (5): 6558 (5); 7560 (10); 10249 (6); 10307 у, 12624 (10): S : : ; 14328 (6); 5); ; 16304 (6): 16478 (6); 16790 (10); 18828 (6); 18851 (A. W. С ang et al. 30738 (4); 31017 (4); 2: (о); 34596 (5 y 34792 (5); 35129 (5). W. Z. Fang 27 (10). 1 Deng 975102 (10). Rev. Pere P. G. ; 14: (6); 772 (6); 1073 (4). Pere U. J. Faurie | in 1905 (9); 238 (9); 303 (6); 328 (9); 425 (6); 670 (9); 725 (6); 726 (6): 721 (6); 728 (9); 1875 (9); 1876 (6); 2511 (6 3272 (13); 4281 (9); 5928 (9); 13031 (6); 13215 (6). (7); 5149 (16); 7455 (5); 7938 (5); 8236 (5); 8789 (5) 10406 (6); 11082 (6); 11453 (16); 12267 (16); 12740 (6): 13452 (16); 21567 (7); 22004 (6); 22401 (5); 22614 (16): 50170 (6). Fengel 13 (10). F. Fleury 469 (16): 30203 (16): d Exp. (summe r, "X T — "nr 249 (1. р: (14). С. Forrest ій ra (6). 3585 (2); Researc ch ind b.b. 3965 ( 6854 (14); 8618 E : 4899 7899 ( 7); 11945 (6); 12410 (6); 12653 (7); 14221 (6) (6); 15710 (6); 16049 (6); 16929 (7); 17521 (6); (6); 18249 (5); 18455 (6); 18504 (5); 18927 (5); e (5); 18957 (5); 20300 (5); 20855 (5); 21083 (1); 21112 (5); 21803 (5); 22394 (7); 22927 (5); 23051 (1); 23237 (7); 24039 (6); 24039F (6); 24445 (6); 24681F (5); 25191 (6); 25649 (5); 26380 (6); 27397 (6); 27962 (5); 29552 (6); 29792 ©; 30908 (5). F. R B, Fer 37280 (6); 37361 : . H. E. Fox in 1912 (6). J. Q. Fu 582 (6): 1894 (4); Ж (4); 2210 (6). K. T. ja 1906 (6); 4849 (4): 5240 (6). L. К. Fu 655 (10). S. & T. Fujii 1792 (6); 0 1850 (6). №. Fukuoka 5852 (6); 7461 (6); 11562 (6). №. Fukuoka & M. Ito 173 (6). №. Fukuoka & №. Kurosaki 1574 (6 ). . S. Gamble 261 (5); 3139A (5); 3141A (5); 6884A (5); 6890A (5); 7009 (5); 9552 ( oh 27995A en M. X. Gao T al. 2 . X. F. Gao 17 E (16 Gao 1694 (6). DEAR sa 823045 (10). H. B. G. ut 376 (12). J. Q. Ge 21235 (16). R. Geesink p T. Santisuk 5275 (11). R. Geesink et al. 5776 (12); T5776 (12). J. м. Gilchrist 79 (10). eu 152 (16). J. L. Gressitt 2414 (6). . C. Grier- son & D. G. Long 1107 (5). A. Griffith 2 эч D. Griffith 2268 (5). W. Griffith in 1844 (11); s.n. (11); 309 (5); 2267 (5); 3670 (1); 3671 (6); 3673 (5); 3679 (5). K. J. Guan 75256 (10). Z. T. Guan 329 (6); 411 (4); 7252 (4); 7668 (5); 7779 (5); 8059 (5); 8197 (5); 8448 (5); 8931 (5); 8988 (5). Guangdong Wood Exp. (1970s) 775 (6). Guang-fu Coll. Team 260 (6); 306 (16); 326 (16); 376 (16); 541 (16); 707 (10); 752 (6); 796 (16); 968 (6). Guangxi Forestry чк” 192 (10). er Institute of Botany 1 (16); 2 ои (16); 5 (16); 8 (16); 9 (16); 10 (16); 11 (16); 1 8 (16); 19 (16). Guangzhou Geographic Institute E Z. Сиб 342 (4); 3945 (6). J. Q. Guo 21235 (16). . Han 174 (16). H. F. OA in 1851 (10); 890 do E . F. Handel-Mazzetti 743 (6); 2068 (7); 6119 (5); 6224 (7); 9060 e 9456 (5); 10310 (6); (6). H. А гаќа 790 wh 2141 (5). H. Hara et al. in 1974 (9); 6143 (5): 6144 6) E i 3 5. Hasegawa s.n. (6); 2586 cde (6). S. Hatusima & X G. S. He 767 (10); 1476 (16); "v (16); 4256 (10): 4464 (16); 5168 (16); 5664 (16); 6049 (10); 6342 (10). J. He 2246 eee 2278 (10). S. B. He 614 (6). Y. Q. He & C. L. 2 (6). G. Hei 1251 (9). Hon. Forestry wo 59 ы 027 (6); 1074 (4). М. К. Henderson 18 (11). А. Hen n. (6); 1430 (6); 1918 (6); 2116 (Ө); 2815 (6): ae 3876 ( (6); 3926 (6); 4120 (6); 5495 (6); 5639 (6); 5639A (6); 5676 (4); 5676A (4); 5769 (6); 5779 (6); 5977 (4); 5980 (6); 6120 (6); 6895 (4); es (6); 7 (6); 7427 (4); 8882 (4); 10055 (6); 10644 (5); 12006 (16); 120064 (16); 12006 (16); 13673 (16); 13693 (16). Herb. Sc. Coll. Imp. Univ. Tokyo s.n. (13 ы (4). y^ Наки c 3). — 16); 4 пог i A 15014 ( PA 16234 (б); 16310 (6); 16409 (6); (6). S. C. Ho 85124 (16); 85159 (16); 85196 (6); 85200 (6); 85251 (16); 86872 (4). Y. Y. Ho 1059 (10); 5654 (10); 20567 (10); 21866 (10); 21996 (9); 22063 (10); 22113 9); 22494 (10); 22981 (10); 23036 (10); (10); 23281 (10); 23309 (9); 23747 (9); 23790 (10); aon (9); 24026 (9); 24197 (10); 24449 (9); 24610 (10): 25488 (9); 26445 (10); 28442 (10); 28554 (10); 28794 (10); 28938 (9); 29107 (9); 29245 (10); je (8). R. S. Hole 17 (1). T. Hong & B. J. Geng 217 (10). J. D. Hooker s.n. (5). J. D. Hooker & J. J. Hooker s.n. (5); 1300 (5). A. Hosie 35 (6). K. Hosoi 2165 (9); 2476 (9). M. Hotta 26604 (14). X. X. Hou et al. 601 (4). C. C. Hsieh 39893 (5); 40316 (5); (5); 40463 (5); 40488 (5); 40864 (5); 41104 (5); 41134 (5); 41303 n 41703 (5); 42119 y^ 42238 (5); ў . Y. K. Hsiung 5891 (9). С. M. Hu 2138 (10); 2758 um 2881 (10); 2928 e 3262 (10); 3342 (10); 3737 (10); 4252 (10); 4331 (10); 4383 (10); 4585 (10); 5271 (10). H. H. Hu 801 (6). S. Y. p in IE (6); 8359 (10). W. K. Hu 10234 (5); 34792 (5) 22469 (1 6). Y. Y. Hu 580574 (6); 580582 (6): A (6). C. Huang 163313 (10); 163471 (10); 164273 (16). D. A. га 60082 (16); 60138 (16); 60206 (6); 60211 (6); 60347 (6). M. X. Huang 112049 (16); 112239 (10); (10); 112448 (10); 112743 (8). О. B. Huang — Hunan Chinese Herb Medicine Institute 7300081 (10). Hunan Forest Institute 77-403 (6); 77-407 (10); 77-409 (16); 6214 (16). Hunan Normal Univ. 386 (6); 441 (10). 1. Hurusawa 10647 (6). C. T. Hwa 16 (6); 427 (6). J. L. — "n 550 Annals of the Missouri Botanical Garden Hwang et al. 190816 (10). K. Ichikawa 200547 (6). Y. Ikegami 2628 (9); 17502 (6). H. "i es 1117 (6); Im 10585 (9). I Im & T. Karahara 9714 (6); ¢ DS s K. Inoue 1834 о ты (6). №. К. Ip 39 (10); 69 (10). M. Ito 675 (6). M. Ito et al. 1293 (9). T. Iwasaki in 199] (6). K. Iwatsuki et n 93 (6); 141 (6); 466 (6); 666 (6). К Brod & N. iem in 1965 (6); 620 (6) G. Jack in pu . J. Jerasaki in 1906 (6). Z. H Ji et al. 467 (6) PS an 31365 (6); 31437 (6); 31650 (6); 50113 (6); n (6); 50741 (6); 51245 (6); 61356 (10); 400450 (6). Jiang 8321266 (10). R. B. Jiang 521 (6). Jiangxi Normal Univ. 1160 (6); 1186 (6); 1242 (6); 1243 6). R. J. Jin et al. J8311012 (10). F. W. Junghuhn in 1840 (14). J. Jutila et al. 392 (6). H. Kanai & H. Ohashi in 1973 6859 9) M. katoet al 120 (6). 44 (6). 1. Keiske s.n. (6); s.n. Keng : 311 чо, 800 (10; 2371 s.n. (9); 521 (6); 522 (9). S. M. Hwang 4684 (s 1733 (6); 2256 e: „өк = (9); 731182 (9). I. Kato T. Kawakami & S. Sasaki А (12): ) (11) 14659 (11): : TT a. Р 1): 20941 (6). DI189 (12). M. S. Kiah bin Hadji 35302 (11). Y in 1984 (6); in 1987 0) А, Kimura et al. in 195 56 (9). Dr. . Kirino 360 (6). F. Kirkham & \ X86 (9). ©. P Ko 50337 (10); 52797 (15); 52889 (10): 53046 (6); 54212 (6); 54282 (10); 54507 (16); 54545 (16); 55648 (10); 55960 (6). S. Kobayashi in Sees (6); A (9); 16251 (6); 16480 (6). W. М. Koelz 2339 (5); 23810 (5); 25269 (5); 30684 (5). К. Kondo in a (6); 2228 (6); 8196 (9). F. Konta & H.' lakahashi CH3721 (16). H. Kuenzler 2176a (6); 2197a (6). Kunming Work E 163 (6): 7160 (16): 50170 (6); 50939 (6). P. € Kuo 342 (4); 343 (4); 1636 (4); 2180 (6); : (6). S. Kurata & T. Nakaike 1605 (6). T. К 3825 (6). Л. Kurosawa et al. 615 (9). U. Kurosu in 1997 (6). U. Kurosu & S. Aoki in 1995 (16). S. P. Kwok 80404. (10); 80419 (15). J. H. Lace in 1902 (5); 5737 (1); 5752 (6); 5774 (1). S. Lai 235 (10); 236 (10); : 2 (6), 433 (10); 558 (6 » 789 (16); 900 (6); 2374 (6); 2464 (10); 2695 (6); 2881 (10); 3302 (10); 3458 (10); 3495 (10); 3530 (10); 4331 (10); 4337 (6); а (6): 5008 (10); Pu (6). Rev. J. La- mont 437 (10). . Lan 85351 a H. Lau 441 (10); 708 (10); 9576 с РЯ НР S. К. Lau 4432 (10); 28504 (16). Y. S. Lau 133 M > Laumonier YL5961 (14). Y. W. Law 508 (6). K. L. Le 235 : T. Lee 2279 (16); 2472 (16). T. C. Te zt 3 - 3107 (4); 3284 (6). C. I. Lei 375 (6); 376 (6); 701 (6). J. H. Léveillé 3319 (6). B. G. Li dg (15); 150 (16); 5149 ws gu (6): > 75017 ls 750286 (4). Li Cheng 3440 (5); 2 (5. G. B. Li А. ү J8112140 (10). G. ч т i 61097 (4); 61237 (6); 61872 (4); 61931 (6); 61933 (6); 62322 (4): 63675 тар; 63893 (4): .H.L E 1962 al. 1974 0x 2034 (6). H. J. Li — веј E Li зт 192 (6). " K. Li NAT 3558 (5). M. S. Li & Z. Y. T 25 (6); 4584 (10); 5247 (10); 5674 (10). P. Y. Li 1380 (4) 2209 (4); 2343 (4); 2692 A 7778 (4); 7813 (4); 8461 (6); (6). Q. H. Li Chen 613 (10); 834 (10); 1146 (10); 1335 (10); Т: a D 1806 (10). R. Li eha . B. Li et al. J8212029 (10); J8212112 (10). X. (6 Li 200955 (10); 201687 (10); 202464 (10); 203380 o эже, X. W. Li 127 (Ө); 173 (5). Y. Li 10623 (10). Y. H. Li 148 (6); 1987 (16); 3224 (16); 5408 (6); 5739 (6); 11499 (16); 11725 (16). Y. K. Li in 1976 (16); 515 (6); P901 (16) P1122 (16); 9322 (10). Z. J. Li 1458 (16); 3223 (16). Z i Li 74 (16). Z. T. Li 70137 i 600853 До; 604103 6). Z. X. Li 36 (10): 37 (10). Z. Y. Li 272 shoes MN in (9) D Team 11126 7). W. Y. al. (16). 30252 P 30291 (10). E т EN (10); 33787 (16). F. € » Mang 197 (6). L. К. Liang 21996 (16); 22001 (16): : 35655 (16). P. H. Liang 83552 (6); 83707 (10); 83722 (10); 84294. (16); 84351 (10); 84394A (16); 84418 (16); 84522 (16); 84568 (10); 85107 ( » 85544 (15); 86286 (16) 86298 (10). H. 0). H. W. Limpric ‘ht ‹ 920 (7). L. K. Lin 4 10201 по: 70 (6); ‹ Y 11039 (6); 11051 (6): i А (4); 257 (6); 266 (6); 575 (6). Y. Q. Lin 11 (7). 3 16575 (16). Z. W. Lin 603 (0). ‹ 0. . Ling in 1932 (10); 3487 (10); m 5310 (fo Ying Ling 218.01 )). К. j 1938 (6); 5328 (4); 8643 (6); 8776 ^ 1); 14551 (6); T w c = = em їл — E Zz — л ~ - - Б N. - 16609 bi М. en D. Li iu 2339 (10): 10850 {бу 10856 (0) ie 1017 164 (16). H. $5 Liu sre . Wang 82253 (16). J. t vu ›). L. H. Liu ee (6); € (10 10579 (10). L. F. Liu 576 (6); 10469 (10): 10489 с (6); 5646 (6); 5852 (6). „ Liu 80381 (10): 890381 (10). T. W. Liu & Z. F. Zeng d (6); 220 (4); 235 (10); 245 (4); T 1372 (10); 1393 . X. Liu 566 (16); 694 (16). X. L. Liu 4791 (9). X. Q. E 24221 (6); с (16); oni (8); 29012 (10). Y. Li ju | 128 (6): 563 (6). Y. S. Liu 1225 (5); 1269 (: d 2212 (5). G. R. Long 830125 (6). Long-gang Compl. Exp. 10593 2 11682 (6). J. Lórzing 5641 (14); 14096 (1: 0: 14129 (14); 15167 . X. Lü Oe (10); и (6); 2150 dn. 2781 ( E 1545 (16); em 7, W. Lii Ludlow & G. Sherriff 31 у" 5). F. Ludlow et al. 12593 (5); 18802 (5); 20510 (5); 20562 ( B. Luo 93 (10). X. R. Luo 2066 (16). 3713 (5); 5); 20995 (5). L. Y. B. Luo 2987 (6); 3009 (6). Z. C. Luo 201 (6); 843 (10 W. W. Ma 2589 (5). Y. F. Mai 60651 (10). E. E. Maire s.n. (5); 57 (6); 1321 (6); 1365 (6); 1413 (6); е (6). T. Makino i in PA Bs 3); 105693 (6); 105695 (6); Е (6); ; 105786 (13); 121299 (13). Н. ; 427 n. 2394 (16); ; 2779 no 2941 (16); 3982 (16); 4083 (16); H E = 1326 (1); p (16); 5570 (5); 6173 (16); 7160 4 (10); 5 . P. Maradjo 87 (14). : 489 (6). S. Masajiro 4585 (6). К. 09); in 1862 O); in 1862 (6); in (9). J. F. Maxwell. 85346 ü M 85535 ü 1); 85669 ü 1): 9] ee (12); 93944 (12). H. Mayr in 1886 (9). H. D. McLaren 46 No.2 ; 1100А (5); AD102 (5); U105A (6); 114F (AA) (7): E (7); J: F114 (7); S114 (7); F122 (5); F142 (5); aalol (7); S161 (7); F199 (7); 205F (7); F205 (7); U219 (6); E (6); 233C (7); e233 (7). 1. C. V : Mohr in 1931 (14). W. Meijer 3175 (14) ` 7614 (10). F. P. Metcalf & T. Meyer 275 (9); 277 (6); 440 ur Bristol 148 (6). H. Migo i in 1927 (9); in 1932 (6); in 1937 10); in 1954 (6). R. G. Mills in 1914 (9). Milne 187 (6). K. Mimoro 1840 (6). K. Mimoro & S. Thugaru 3195 (9). der Meer —. Volume 90, Number 4 2003 Huang et al. Revision of Styrax Series Cyrta K. Miyabe & E. Tokubuchi in 1890 (9). F. Miyoshi in 1961 (1 3); 1363 (6); 2167 (6); 2756 (6); 3076 (6); 5139 (6); 10778 (6); 19506 (9). M. Mizushima 300 (9); 686 (6); 2035 А 2379 (13); 2958 (13); "^ (9); 11365 (6). U. Mizushit M. Mizushima 814 Mochizuki in 1904. Я іп 1910 (13); in 1910 (6). 5 G. J. Mohnike s.n. (6). Mokim in 1898 (6). Y. Momiyama 234609 (6); bono (6). R. Morar 1 4327 (€ 2); 5039 (6); 5177 (6); 5209 (9); 5352 (9. f Миша ш 1940 аз); 1030 € )); 55807 (13). G. Mura ag S. Kitar з. Murata & Н. Nishimura = (6); 906 ü 3x "5663 e ;. Murata et al. ; prs (12). J. Murata 1769 a J: 4854 (6). J. T. Chen 7672 (6). J. Murata & H. T. Im 16301 Murata et al. 84 Ly. hog E s 741 (9); 759 (6); 827 O) 927 (os 979 (б): „А ); 3028 (6); 3092 ( 6); 4763 (6); 4880 (0): 4940 (6): 4978 0; 2006.3 9); 5147 (6); 5714 5182 ( ); 6625 (6); 6913 T. Naito in P (6); in 1933 hys in 1938 (13); in 1971 'akano in 19 . K. Nakayama & F. S. Nanba in ce j^ 5). Nan-shui-bei-diao- dui 1871 (5); 2 ); 5709 (7); 5967 (7); 7592 (5); 9051 (5); 9239 (5). T. Nemoto 350 (6); 408 (6). M. X. Nie 726 (10). M. X. Nie & S. S. Lai 2881 (10); 2928 (10); 3530 (10); 4331 (10). M. X. Nie et al. 1906 (6); 1960 (6); 6700 (10); 7265 (10); 8342 E: 8625 (10); 9125 (6); 9637 (16): 9644. re 9781 (10); 9797 (6); 9817 (10). H. Nishimura s.n. (6). C. Niyomdham 339 a E: Nordic Arboretum Exp. 197 6 to South Korea 104a . Nozawa in 1885 (9). 5. О ; H. Sakai T5 ^n Н. Ohashi in 1966 (6); 11983 (9). H. Ohashi et al. 7011 (6); 8653 (6); 21881 (9). H. Ohba 677050 (6). \ 8903030 (6). } . J. Ohwi 9127 (9); 9182 (6); 9343 (9). R. Oldham in ino (6); s.n. (6); 201 СЯ 536 (6); 666 (6). Н. D. Orleans s.n. (6). 7. Parkinson 680 (12). C. Pei 8120 (17); 10327 (4). Б Peng 6070 (10). D. Y. Peng 45496 (6); 45550 (6): ss (5); 46494 (5). Н. Peng 445 (12): 517 (€ j: 1851 (12). P. A. Pételot in 1928 (16); 1033 (16); | 5155 dc Phengklai et al. 4150 (12). J. Phillips in 5. Phusomsaeng 241 (11). J. B. L. Pierre 3288 (5). 4 LT Korea (1984) 2093 (6); 2227 (6); 2525 (6). РІ. эн атин in the Republic of Korea (a tumn, 1989) 298 (9). P Korea (spring, 1989) 32 (9): 119 (9). Pl. Res. Exp. in Fujian 52459 (10); 53444 (10); 62506 (10); 76060 (10). E. Poilane 1831 (16); 1880 (16); A1880 (16); 2020 (16): 2021 (16); 6578 (3); 12572 (16); 12620 (16); 13018 (16); 16779 (16); 16906 (16); 18626 (3); 18835 (16); 18885 (16); 20726 (16); 23457 (3); 23569 (3); 25294 (16); 25690 (16); 26003 (16). A. E. Pratt 406 (4). Pu-cha-biao-ben 479 (6); 745 (6); 3174 (6); 6239 (6); 8269 (6); 10363 (О); 18727 (6); 20304 (4); 34084 (6); 34393 (6); 34510 (4); 34604 (4); 34645 (4); 34666 (4); 35018 (6 C. J. Qi S8 (16). Y. Y. Qian 37 (16); 3719 (16). D. H. Qin et al. 65266 (10). H. N. їл ош, o = 5 ә RC „© Dp © D — Qin 245 (16); 895180 (10). 71265 (6). B. Y. Qiu 50170 (6); k 52476 (5); 54496 (6); 55269 (6); 56119 (16); 57004 (16); A (6); 59599 (6). P. X. Qiu 1487 (10). Z. D. Qu 1136 16). Z. X. Qu 1017 (6); 1074 (6); 1390 (4). N. Rabil Bunnag 92 (11). M. Ramos 80424 (6). A. Reh- P 1362 (7); 2676 (6). H. Y. Ren 11818 (10). C. Ritchie in 1866 (5). W. A. Robertson 152 (12). J. F. Rock 3198 Pl. ека in the Republic of — 7); 5058 (7); 6397 (7); 8268 (7); 8520 (7); 10243 (5): 10546 (7); 17075 (5); 22044 (5); 24150 (7); 24572 (7). K. Saito & H. Okazaki in 1966 (9). H. Sakurai in 1906 9). Ж Sakurai in 1905 (6); in 1910 (9 ). B. Sangkhachand 1014 (11). C. 5. Sargent in 1892 (6); in 1903 (6). S. Sasaki in 1920 (6); 4 L. NSM47 (6). G. Sato 4529 (9); 5250 (9 ). Y. Sato in 1968 (6). Y. Sato et al. in 1971 (6). P. A. Savatier [-Fouillade| s.n. (9): 810 (6); 2035 (9). T. Sawada in 1927 (6). C. K. Schneider 1145 (7); 1402 (7); 3542 (7); 3965 (7). Service Vaca ees s 30204 id F. H. Sha 560 (6); 571 (6). M. Shah Bin Haji Moham r & M. Noor MS2027 (11). Shi andong ї niv. 196 (6); ы (6). Shandong Wild Pl. Exp. 89 (9); 738 (9). S. J. Shen 273 (6). Z. H. > 3 End 0 L. Shi 13948 (16); 14142 (16); 14170 (16); 5 (8); 14841 (16). Y. S. Shiao 49126 (6). i 22540 (9). K. Shiota 2767 (6); 2768 (6); 2769 (9); 2770 (9); 2771 (13); 5358 (6); 5765 (6); 5846 (13); 6525 (13); 7198 (13); 9048 (13); 9049 (9). Y. M. Shui 1974 (6); 2274 (6); 2420 (6); 2 oe (6); 2487 (6); 3056 (6). Sichuan Eco- nomic Pl. Exp. 12 (6); 169 (5); nets 315 (5); 351 (5): 159 (6): 571 (5) ay (5); E (5); 950 (5); 988 (5); Ts (6): 1206 (4): 1270 (6): (5); 1312 (5 5); 1437 (4); 1737 (5); 2204 (5); 2226 (6); 2483 (4); 3703 (5); 3813 (5); 4086 (5) 13556 (5). Sichuan Univ. 11146 (6); 108826 108905 (6); 1107: vu ›). B. К. Sidek S345 (11). Pere C. Silvestri 17704 (6). C. J. Simons s.n. (5). S. S. Sin 9444 (10); 21326 (10); 22254 (16); 22305 (16); 25360 (10): 50091 (6). Sino-Amer. Bot. Exp. (1980) 348 (4); 763 (6): 1133 (4): 1390 (4); 1484 (6). Sino-Amer. Bot. Exp. (1984) 366 (5): 1256 (7); 1307 (6); 1505 (6). Sino-Amer. Guizhou Bot. Exp. 274 (6); 958 (6); 1089 (6). Sino-American Yun- tai Botanical Exp. Team (SAYTBET) 45067 (6); 45207 (6). Sino-British Exp. Cang-shan 1 (7); 269 (5); 275 (5); 850 (5); 998 (5); 1212 (5). aic Exp. 289 (16); (16); 1703 о vu s.n. (5). H. Smith 1661 (6); 2114 (5); 10090 (5). Mrs. R. K. Smith in 1937 (6); in 1938 (6). W. W. Smith in 1908 ( (5). T. Smitinand 328 (12). K. Sohma in 1976 (9). K. Sohma & M. Takahashi 535 (9). K. Soma el al. iw (6). X. H. su 185 (10); å ; A. Sontag in 1894 (6); in 1895 (6). T. P. Soong ; 39476 (17). Specimens n ак 33 (4). „J. TU 420 ( Nee n420 (16); 4. й . Stainton 5308 "s d (5); 8332 (5). IE — - = ~ . €. Cheo 415 (6); 1085 (16). A. Pt 352 (6). S. Sugaya & C. Kimura en (9); 10422 (9). S. Sugaya et al. ‚ B. S. Sun 141 (6); 359 (5); 618 (5); 676 (5). B. Y. [s in 1988 (6). H. Sun SH85 (5); 518 (16); 771038 (7). S. C. 5 5. C. Sun & 271 (4). S. L. 2255 (4); 24. 12 (5). x. L. Sun 5597 (6). Shigetaka Suzuki. AA14-71 (9); 118 (9); UC10-186 (6); UC647 (6); UC742 (6); AA1200 (6). Y. Җ. ' Ж? 558 (10); 571 (10); 721 (10); 1563 (10); 2110 (10). Taeko 125. (6); MSM125 (6). M. Tagawa 2275 (9). F. t Tai & C. M. Teng 4215 (4). C. Takahashi et ИТ in 1974 (9). M. Takahashi 1197 (6); 1896 (6); 1899 6). M. Takahashi & Y. Yuki 417 و‎ 5). T. Takahashi in 1972 (6); s.n. (9); 353 (6); 1969 (6). H. Takeda in 1904 (9); in 1908 (6); in 1920 (13). A Tébeliara 459 (6). P. C. Tam 57344 (16); 57592 (16); 58283 (6); 58332 (10); 59492 (б); 59535 (10); 61328 (6); 61690 (6); 61731 (10); 62023 (6); 62348 (10); 62817 (10); 62899 (6); 63507 (10); 63659A (6); 63707 (16); 63925 (16); 63944 (6); 64023 AAI 108 — ~ - 552 Annals of the Missouri Botanical Garden 6). S. Tamaki in 1909 (6). M. Tamura et al. 26603 6). . M. Tan 97573A (10); 971113 (10). H.-C. )). L. Tang 1069 (10); 1082 (16); 1178 (10); 2413 (16); 4380 (10); 4630 (10); 5105 (10). P. L. Tang 60996 (10). 5. G. Tang 7524 (10); 7614 (10). T. Tang 137 (4); 23268 (4). T. F. Tang 108 00); 134 me D. D. Tao 164 (16); 238 . Tao & P. 1. Chiou 59599 16). T. Taquet 1108 (9); ; 3033 ( 5); 3034 i (б); 3036 (9 ): 3039 (9). Tashiro . Z. Tashiro in 1917 (6). rd ishi i (6); 10287 (1 М 1: 3890 (9). Y. Tateishi et al. £ 940 (9). S. Te т 8І (7); 182 (7); 210 (7); 351 (7). Н. К. | eng 122 "Wr S. W. T 584 (6); 90211 (6). W. Teng 90472B (6). J. Teysmann 965HB (14). K. P. To et al. 12025 (16); 12267 s 12645 (10). T. H. To 70 (4). M 6). M. Togashi in 1968 (6); in 1978 (12). А 16); 50936 (5); 51035 (5); 51090 (5); 51156 (2); 52068 5); 52110 (5); 52144 (6); 52736 (6); 5: (16); 55793 2); 58355 (5); 58439 (5); 60806 (16); 61212 mis 61 A 16); 62505 (2); 62522 (2); 62766 (2); 73643 (16); 7 (16); 588079 (6). W. T. Tsang in 1928 (6); 621 (6) 1, dc 17370 (6); 20301 (10); 20435 (10); 21712 (10); 22053 (16); 22067 (16); 22655 (16); 24105 (16); 25866 (16): жо (10); 27113 (10); 28311 (10); 29752 (10). W. T. Tsang & Н. Fung 491 (6); LU491-18025 (6); LU18025 (6 ^ Tsang et al. 98 (6); LU98-17629 (6); 491 (6). S Tse Kulî i in 1864 (9); in 1866 (9). Z. H. Tsi 91351 (6). C. J. Tsiang 5343 (6). H. L. Tsiang s.n. (6); 19 (4); 10162 (6); 34596 (5 5); 35129 (5). H. L. Tsiang & ES 34596 (5); 34792 ©) S. P. Tsiang 16646 (16); 16648 (16). Y. Tsiang 107 (10); 351 (10); 1386 (10); 2262 (16); 5004 (6); ; 10140 (10); 11337 (7); 5); 5); 12204 (5); 12348 (12). C. L. Tso 407 (10); 20152 (10); 20645 n 20732 (16); 20797 D 20856 (16); 21164 (10); 21756 (10); zw (10). С. Tsoong 81668 (6); 83331 (10); 83529 . K. тк D48 (10); D225 (10); 265 (10); 403 us 480 (10); 620 (10). P.-C. Tsoong 437 (10); 635 (16); 648 (10); 681 (10); 916 (6); 976 (6); 1032 (6); 1094 (16); 1241 (6); 1275 (6); 1277 (6); 1332 (6); 1740 (6); 1786 (6); 3675 (10); 4300 (10) T. S ЖО 81668 (Ө); 81981 (10); 82058 (10); 83529 (6). S. L. M 20205 (10). S Pens 14456 (6); ea (6). S. Tsugaru & M. Sawada 1855: 3 (9). S. Tsugaru ` Takahashi 6607 (9) 13519 (9); 14822 (6); P (9). Я То et al. 18431 (6): i. (9); 23572 (6). ' Tsui 250 (10). T. H. Tu 70 (4); 103 (5): 203 (4): EH 289 (4); 347 (10); 407 (10); “те ); 5597 (6). K. Ueda 496 (6); 512 (9). K. Uno in ne (6); in 1951 (6); in 1951 (13); 18516 (6); 22537 (6); 2 (6). Jules Vidal 880B (6); 1504 (6); 1575 aa 1 (6). W H. de Vriese in 1857 (14). W. L. Wagner 6721 (6). E. Н. m 8401 (6). F Walker & S. Tawada 6590 (6). E. H. Walker et al. EE 51 (6). N. Wallich 4401 (11); ОУ N. К. Walter 33698 (5). 5. B. Wan 27426 (4); 27436 (4); 27440 (4); 27463 (4). C. Wang 39419 (16); 41180 (5); 41251 (16); 41442 (16); 41987 (16); 44043 (6); 89617 (6); 164273 (16). C. W. Wang 72938 (6); 73643 (16); 73669 (16); 74113 (12) 74118 (16); 74200 (16); 75068 (12); 75198 (12); 77088 (12); 78470 (6); 80687 (16); 82253 (16); 84747 (16); 85972 (16); 87296 (6); 87425 (6); 87552 (6); 87603 (6); 87834 (10); 88441 (6); 88773 (16); 89220 (16); 89414 (16); 89492 (6); 89592 (6); 89594 (6); 89617 (6); 90046 (6). C. Z. Wang 841 w^ D. S. Wang 453 (10); 697 (10). F. C. Vë 10467 (4). F. T. Wang 22848 (5); 22866 (6); 23029 (4); 23268 (4). H. C. Wang 1718 (7); 1740 (7); 1988 (7). H. Y. Wang 981 (4). J. X. Wang 1426 (10); 1741 ( ( ( ( (6 ( ork (10); 2055 (10); 0); 2123 (6); 2124 (10). К. С. /a 16). L. Wang & S. L. Tsou 60842 (6). M. J. Wang 3487 (10); 3780 (10). S. Z. Wang 889 (16); 1043 (6). S. X. Wang 462 (4). T. H. Wang 12044 (6). T.-P. Wang 11165 6); 11375 (6); 11422 (6); 11480 (4). W. C. Wang 390 (7); 18532018 (10). X. Wang 98 (6). X. Z. Wang 7371 (10). Y. C. Wang 91 (6); 779 (6). Z. Wang 1611 (9). Z. B. Wang 11165 (6); 11375 (6); 11451 (6); 15652 (4); 19392 (6). Z. T. Wang et al. 870095 (6); Sd (6). Z. Y. Wang 618 (10). O. Warburg 6635 (6). F. F. K. Ward 6630 (5); 18818 (5); 20550 (1); 20632 (6); AA 21553 (5). J. К. Ward 3831 (7). K sonra in 1891 (6); in 1899 (6). Whitmore TCW3348 (14). W. J. J. O. de Wilde & 4 = E. de Wilde 15730 (14); 18342 (14); 211 m 4). C. ford 816 (6); 934 (9). Wilson 1100 (4). E. 1901 (4); in © 3); s.n. (4); ta G ; ; 1734 (10): 3015 2099 (1 ~ 2756 (6); 4065 (5); 6007 (6); 6988 (6); 7003 (9); 7012 3); 7180 (6); 7395 (9); 7462 (6); 7581 (9); 7710 (13); 3086 (6); 8467 (9); 8530 (9); 8753 (9); 8754 (6); 9316 9); 9328 (6); 9454 (6); 9516 (9); 10422 (9); 10618 (6). E. W. Wood & D. К. Boufford 3776 (6); 3967 (6). С Wright in 1853 si 2); p (6): Ба J. € C. R. Wu et al. 18413128 (10). C. Y. Wu 8623 (6). K. M. Wu 60225 (10). ). A. Wu 9022 (6 2); 9040 ae Be (5); 9465 (6); 9669 (6); 9740 (6). 5. K. Wu 57 (5); 2203 (5); 2204 (5); 61- 3837 (5); 6670 (5); 6988 (5); or (5); 8478 (5); 84028 y 613718 (6); 840104 (5). Y : ). Xi-da-an-kang Coll. Team ] à 0 (10); 443 (10). D. Y. Xia BG58 a C. Xiang-liao-dui (Coll. Team for Pe ште РІ.) 156 T 85132 (6); 85269 (5). Y. F. Xiao & W. Z. Xie 152 i et al. 3183 (16); sc ( 6); 3808 (6); е (6). A Ж Cai 440 (6); 847 (5). Z. W. Xie Kn d | 33 J. Q. Xing m (б); 6053 (4); т, »); 8652 (6); NE (6); 8963 (6); 9018 (6); 9028 (6). J. 5. Xing 763 (6); 832 (5); 1082 (16); 50455 (6); 50939 (6). J. H. Xiong et al. 30738 (4); 31017 (4); 90467 (6); 90659 (4); 90786 (6); 91077 (4); 91179 (4); 91185 (6); 91281 (6); 91602. (4); 91652 (4); 92089 (6); 92095 (6); 93258 (6); 93282 (6). J. Xiong 723 (10); 1052 (10); 1860 (10); 2349 (6); 2734 (10). 3434 (16); 3559 (5); oo i 5); 4921 D 5062 (5); 5231 (5); 5378 (5). S 140 (6); 644 (4). X. H. Xu et al. 16 (16). Z. W. Xue 47 10 487 (10); E (6). i 9). Y. G. Yan 6215 (10). B. M. Yang 2166 (4). G. H. Yang 5821 (6); 55005 (4); 55022 (4); 5); 55688 (6); eui Pi : 1 (5); 56741 (5); 5 ; 59092 (6); 59162 (4) 5 65343 (6); 65407 (4). J. S. CURE 8311 (6. J. X. ' Yang 3004 (6). K. H. Yang in 1984 (9). L. Yang 757 (10). S. X. Y. 201 (6). X. X. Yang 16820 ао; 650367 (10); 650492 (10). Y. B. Yang 33 (4). Y Yang 3497 (5). Z. B. Yang 1059 (6). Z. H. Yang 8: 5829 101176 (6): 101327 (12); 101681 (12). C. W. Yao 2708 (4); 3203 (6); 3748 (5); 3778 | 4563 (5); 4739 (5); 4855 (4). Н. W. Yao 4563 (5): 17: 39 (5 5). К. Үао 8497 (6): 8928 (9); 8965 (6). em, = 4 c => T == aa “Ж = — T le ^^“ (9); in 1906 (6). R. Yatabe 11209 (6). (14); 1467 (14). G. FA ы 2137 (16). Ye-ti- V -san-dui 347 (4). Z. C. Ye 413 . €. Yin 490 (5); 5 (6). J. S. Ying et al. 188 "a Vb. (6): 560 (6). H. E Yip 364 (16); 453 (16). 1. Yogo 9510 (13). К. Yonekura 395 (6); 3259 (6). O. Yongsok 6529 (9). S. Yoshioka 23 (6). P. H. Yu 240 (5); 302 (6); 331 (6); 334 (6); 336 (6); 1096 (6). S. L. Volume 90, Number 4 Huang et al. 553 2003 Revision of Styrax Series Cyrta & J. F. Qin 700454 (6); 700556 (6). T. T. Yi 853 (5); 14383 (6); 15926 (6); 16612 (4); 16760 (6). C. Y. Zhao EA (5); 936 (5); 1903 (17); 2011 (5); 2034 (5); 2629 (5); 21670 (7). Q. S. Zhao 309 (5); 428 (6); 454 (6); 504 (5): 3055 (5); 3066 (5); 3263 (5); 3629 (5); 3692 (5); 3987 27 (5); 1547 (5); 1617 (6). К. Е. Zhao 14 (6); 60 (10); (5); 7216 (7); 7309 (7); 14198 (7); dw: ); 16624 (5); 6 (6). Y. X. Zhao s.n. (5); 224 (5); 511 (5); 22139 (7). 17074. (5); 20294. (5); 22079 (5). X. L. Yu 91440 (6); E X. Zhao 224 (5). Zhejiang Bot. Res. Team 25639 (10); 91655 (6). S. F. Yuan & L. F. Liu E (10); 5646 (6); 25752 (10); 25807 (6); 25856 (10); 25888 (10); 26071 5805 (10); 5850 (6). S. F. Yuan 6492 (16). J. : v 1810 (6); 26385 (10); 26490 (10); 26754 (10); 27031 (10); (6 2d 355] (6); 4669 (6); 5021 (6). X. D. Yun 104 (6). 27581 (10); 28350 (10); 29382 (9). Zhejiang Forestry Col- Zhang et al. 180 (10); 300 (10). C. L. au 56031 lege J8023047 (10). J. H. Zheng du ). Z. S. Zheng 230 (10). G C. Zhang 35 (6); 53 (10); 256 (10). G. S. Zhang (6). Zhong-nan-lin-shi-xi-dui 64 (16); 94 ru 137 (8); 105 (16). J. X. Zhang & B. Н. Chen 215 (16). Q. T. Zhang 163 (10); 173 (10); 30913 (6); aes (б). S. Q. Zhong 8532 (16). S. Y. Zhang 502 (10); 633 (10); 690 (10); 2441 А60960 (6); A63811 (16). С. S. Zhou 103 (6). H. F. Zhou (10); 2658 (10); 2705 (10); E e (6): 3286 (6); — 10903 (6); 11150 (6); 26228 (10); 26359 (6); 26529 (6): 3319 (6); ا‎ 5076 (6); 5178 (10); 5251 (10): 5471 26661 (6); 26685 (6): 26686 (6); 108265 (6); 109331 (6). (6); 5477 (6); 5481 (6); 5667 (6); 5738 (6); 5740 (6); 5822 T. Y. Zhou 73 (6); 458 (9); 738 (9); 1187 (9) 2297 (6); 010); 6011 (10); е 6613 (6): 6806 (10); 6993 (10). 6389 (9); 13323 (6); 13351 (6). С. X. Zhu 105 (6). Н. Q . Zhang & Y. en 4507 (17); 4534 (17); 4640 Zhu 228 (10); 267 (9); 801 (10) T. P. Zhu & Z. F, Liu nes 4957 (17); 4982 (17). Y. T. Zhang 79025 (10). Z. R. 373 (6); 1360 (6); 2046 (6). Z. C. Zhu et al. 34 (4). Zhang 25051 (6); 25145 (6); 25586 (6); 25715 (6). Z. S. Zhu 1910 (16); 1917 (16); 1925 (16); 1928 (16); 1929 Zhang et al. 1602 (6); 1906 (6); 1960 (6); 400268 (6); (16). R. Zimmermann 345 (9); 422 (6). H. Zollinger 535 400456 (6); 400532 (6); 400731 (6); 401756 (6); 402476 (6). H. E Zou 123 (10); 454 (16); 800 (16); 20266 (16). (10); 402501 (10). Z. W. Zhang J8311260 (6). Z. Y. Zhang H. Y. Zou & F. Y. Yuan 847014 (10). APPENDIX З. Index to scientific names. Numbers in parentheses correspond to taxon numbers in the text. Synonyms and excluded names are italicizec Adnaria Raf. ........ 500 var. Jippei-kawakamii (Yanagita) H. Hara (б) 516 Anthostyrax К. = 500 var. kotoensis (Hayata) Masam. & Suzuki (6) 516 tonkinensis Pierre (16) 500 var. longipedunculatus Z. Y. Zhang (6) —........ 517 Супа Lour... 500 var. nervillosus Z. Y. Zhang (Ө) |... 517 Japonica (Siebold & Zucc.) Miers (6) ——..... 516 var. tomentosus Hatus. (6) 517 Darlingtonia Torr. 499 var. zigzag Koidz. (6) .. 516 Epigenia Vell. 499 Jippei-kawamurat Yanagita СО сыш шас 516 Foveolaria Ruiz & Pav. 499 kotoensis Hayata (6) 516 Styrax L. 499 langkongensis W. W. Sm. (7) 522 agrestis (Lour.) G. Don eae | 500 limprichtii Lingelsh. & Borza (7) -2 522 var. curvirostratus e Svengsuksa (3) _. 506 macranthus Perkins (5) 512 bashanensis S. Z. Qu & К. Y. Wang 2... 545 macrocarpus W. C. Cheng (8) 225 be "ongensis H. R. Flete her (11) |... .. 9532 mac idcm Perkins (16) 243 bodinieri H. Lév. (6) . . 56 obassia Siebold & Zucc. (9) 226 buchananii W. us Sm. (D 1 503 odoratissimus Champ. ex Benth. (10) 528 caudatus Perkins (5) _. 512 oliganthes Steenis (14) -............... —— 331 кеншен HL Le (6) 516 perkinsiae Rehder (SE 512 chrysocarpus Н. L. Li (2) . 505 philippinensis Merr. & Quisumb. (6) .................... 516 curvirostratus (B. Svengsuksa) Y. L. Huang & роги is Don (1 oe › W. Fritsch (8) ууу... m | 506 prunifolius Perkins (10) 328 duclouxii Perkins (6) 516 сш» Dunn (5 ) жашыса ысы 512 floribundus Griff. (11) . 532 serralatas Ro (12) + E a hA 310 var. latifolius Perkins (1) 503 - ollissimus Steenis (14) 937 var. griseus Rehder (4) 509 иго hak (HD 532 hookeri C. B. Clarke (5) 51 h d M = shiraianus Makino (1 535 var. yunnanensis Perkins (5) 512 var, discolor Nakai (13) 535 huanus : 509 shweliensis W. W. Sm. (5) 512 hypoglaucus Perkins (16) 541 subdenticulatus Miq. (14) 537 japonicus Siebold & Zucc. (6) 1 516 subniveus Метт. & Chun (16) |... : 541 f. jippet-kawamurai (Yanagita) T. Yamazaki (6) _ ethene ulatus Jungh. & de Vriese (14) ......... 537 е лень O16 виран Chun & F. Chun (15) 0 540 f. parviflorus Y. Kimura (О uoces sic мое RS кай inensis (Pierre) Craib ex Hartwich (16) ..... 541 f. pendulus T. Yamazaki (б) secans 517 touchanensis H. Lév. (6) 516 f. rubicalyx Satomi (6) 517 veitchiorum Hemsl. & Е. H. Wilson (10) |... . 9528 f. tomentosus (Hatusima) T. Yamazaki (6) ....... 517 wilsonii Rehder (17) . 544 var. angustifolius Koidz. (0) |... 517 zhejiangensis S. M. Hwang & L. L. Yu (8) 525 s Eas у S. M. H wang & L. L. Yu (8) : var. calycothrix Gilg (6) 516 Strigilia Ca 499 var. iriomotensis Masam. (0) — 516 shiraiana (Makino) Nakai (1: 3) 535 PHYLOGENY OF W. M. M. Eddie? T. Shulkina.;? CAMPANULACEAE S. STR. $ Gaskin" R. C. Haberle? and INFERRED FROM ITS SEQUENCES OF NUCLEAR RIBOSOMAL DNA’ ABSTRACT inety-three taxa comprising thirty-two genera (plus four outgroups from gon 'eae) of ile Campanulaceae based on ITS sequences of nuclear ribosomal DNA consensus tree е b strap values was constructed, in addition to a phylog were used to estimate a phylogeny rom 2629 most parsimonious trees, a strict im showing branch lengths. The topologies of scussed in relation to the pollen and as a venio within the family, in addition to chromosome number and аас ‘al distribution. The results show that there is a major dichotomy between the colpate/colporate pollen alliance (platycodonoid taxa) and the porate polle n alliance Жети көш od id ‹ ampanuloid taxa). Both these major alliances are monophyletic. Within the porate alliance there are two major clades, the wahlenbergioids and the campanuloids. The a clade is further subdivided into two major clades representing the Rapune "ulus and the Campanula s. str. groups of taxa, plus three smaller clades that are considered as “transitional” taxa. It is argued that the family originated in a eben те West Gondwanaland and that tectonic processes are res sponsible for the major dichotomy in the family. The colpate/colporate eph 'odonoids Кинини ө remained relatively relictual in Asia, whereas the porate taxa s of the Northern and Southern Hemispheres. The campanuloid line sage spread over the Northern Hemisphere from a gom evolutionary center in the Mediterranean region and is represented in North eae пса only by the Rapunc ulus group. The wahlenbergioi lineage i widely dispers ed across the southern continents and oceanic islands but has a spread over much r secondary provides insights for future investigations and a phylogenetic framework that can be tested with other data sets. Its limitations for phylogeny reconstruction are briefly discussed. More extensive taxon s i ang additional data sets are required to refine these results and for a new classification of the Campanulaceae to be ed. торо Key words: (ae eae, evolution, Gondwanaland, ITS, nuclear-ribosomal DNA, phylogeny. Classification systems of the bellflower family vided the family into two subtribes, the Campanu- eae s. str.) have traditionally followed leae and the Wahlenbergeae, based on the mode of the рони ments of Boissier (1875, 1888) and capsule dehiscence (Table 1). Schónland divided Schénland (1889-1894) and, together with the re- the family into three subtribes, separating Platy- finements of Charadze (1949, 1970, 1976), Fedorov codon A. DC., Musschia Dum., and Microcodon А. 1957), and others, can ultimately be traced back DC. in his subtribe Platycodinae on the basis of to the arrangement of De Candolle (1830) who di- calyx lobe position in relation to the locules of the (€ ! W.M.M.E. thanks Tina Ayers and Randy Scott for their pne d in Flagstaff un for facilities at Northern Arizona University in 1998. Others who deserve spec " me ntion inc тен Oliver, Andrew Hudson, lan Hedge, Martin Ingrouille, Susana Neves, Mark Chase, Mike Fay, Peter Lewis, Marcia p Jim and din Arc d iem and Per Hartvig. For technical support we thank Kavita Vyas ar nd С ‘hristine Green. The Regius Keepers of the Royal Botanic Garden, Edinburgh, and the Directors of the Royal Botanic Gardens, Kew, are thanked for use of 1 acilities. Assistance by the staff of the Goulandris Natural History Museum (Athens), the Royal Botanic Garden, ол -« the Darwin Library University of Edinburgh) i is gratefully acknowledged. For funding, W.M.M.E. acknowle dges Iniversity of London (The Central Research Fund; The Keddy Fletcher-Warr Studentship of Birkbeck C olle ge), at University of iue ов (The Molecular seii Fund of the Institute of Cell and Molecular Biology: The James Rennie Beques t), and Royal Botanic Garden, Edinburgh (The Edinburgh жо Garden (Sibbald) Trust). R.C.H. acknowledges assistance with collections from and discussions with Nancy Morin (The Arboretum at Flagstaff) and Barbara Ertter (Jepson Herbarium, University of California, Berkeley). We also Keri Tom Givnish and Tom Lammers for p suggestions on an earlier version of the manuscript. This research was supported by NSF grant DE B 9982092 to R.K.J. ? Section ۷ ил Biology, Institute of Cellular and Molecular Biology, and Plant Resources Cent r. University of Texas at Austin, Austin, Texas 78712, U.S.A. jansen@mail.utexas.edu. * Missouri "bee Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. ‘Present address: United States Department of ане id ultural Ming оли e, Northern Plains Agri- cultural Research Laboratory, 1500 North Central Avenue, Sidney, Montana 59270, ° Present address: Office of Lifelong Learning, University of Edinburgh. 11 Bucc ye h Plac е, Edinburgh, Scotland, U.K. weddie l @staffmail.ed.ac.uk —. ANN. Missouni Bor. GARD. 90: 554—575. 2003. Volume 90, Number 4 2003 Eddie et al. 555 Phylogeny of Campanulaceae Table 1. Classification of Campanulaceae (A. P. de Candolle, 1830). Subtribe I (Wahlenbergeae) Subtribe II (Campanuleae) Capsule with apical (valvate) dehiscence 15 59 Blume (baccate capsule) Сапагіпа L. (baccate capsule) b A. DC Codonopsis Wall. asione L Lightfootia UHer. Prismatocarpus UHer. oella Wahlenbergia W. Roth Capsule with lateral (porate) dehiscence Adenophora Fisch. Campanula L. Merciera A. DC. (indehiscent) Petromarula Vent. ex Hedw. f. Phyteuma 1 Specularia A. DC. Symphyandra A. DC. Trachelium L. ovary (Table 2). Such natural classifications were essentially based on morphology of the calyx (e.g.. the presence or absence of appendages between the lobes) or of the mode of capsule dehiscence (e.g.. whether it is apical and valvate or lateral and por- ate). Many authors (e.g., Hutchinson, 1969; Caro- lin, 1978; Cronquist, 1988; Takhtajan, 1969) con- sidered Cyananthus A. DC. to be the most primitive genus within the family based on its superior ovary. These various classifications were generally use- ful in floristic works, especially during the 20th century when much of the research on the Cam- panulaceae was of a regional, floristic nature. Fre- quently, various authors have used their own mod- ified system with many nomenclatural changes, and great confusion has resulted. Considerable conflict still exists as to the number of genera recognized. Generic distinctions in the family are often subtle, being based on a suite of characters best observed in living plants. In addition, species of the Cam- panulaceae appear to be prone to considerable phe- notypic plasticity (Eddie, 1997; Eddie & Ingrouille, 1999) as well as ontogenetic variation, and this has led to a burgeoning of the literature with superflu- ous species names. w generic monographs that have been completed, although excellent, often lacked a global perspective, and have contributed little to the establishment of a new, more generally accepted classification of the family. Reconstruc- tion of the phylogeny of the Campanulaceae has been hindered by a lack of consensus as to what constitutes a genus and the failure to apply impor- tant character combinations (e.g., cytological and palynological characters), which could potentially Table 2. Classification of the Campanulaceae (Schénland, 1889-1894). Tribe Campanuleae Subtribe Campanulinae Subtribe Wahlenberginae Subtribe топ Adenophora Fisch. sect. Rapunculus Boiss. Canarina L. Heterocodon Nuttall Michauxia L Her. Ostrowskia Regel Peracarpa J.D. Hooker & T. Thoms. Phyteum sect. Cylindrocarpa Rgl. dranthum G Tracheli um C ampanumoea Blume Hedraeanthus Grisebach Het ле a A. DC Jas E ha (J. D. Hooker) Lem. Lightfootia U Her. { . DC. Prismatocarpus U Her. ا‎ Hochst. Roella Siphoc d reichelia Wahlenbergia W. Roth Microdon А. sect. in A. DC. sect. Caelotheca A. DC. Musschia Dum. Platycodon А. DC. 556 Annals of the Missouri Botanical Garden highlight major discontinuities at the generic, trib- al, and subtribal levels. Many species have been placed, for convenience, in Campanula L., Asyneu- ma Grisebach & Schenk, and Wahlenbergia Schrad. ex W. Roth, and this has further complicated our understanding of phylogenetic relationships. In- deed, some of the intrageneric taxa in these large genera are probably more deserving of generic sta- tus than some of the currently recognized segregate genera. The so-called satellite genera of Campan- ula do not appear to be any closer to each other than they do to Campanula, and there is no evi- dence to suggest that Campanula, despite its nu- merical superiority, is ancestral to them. It is thus often easier to define what Campanula is not rather than what its actual boundaries are. Thus, to some extent, the genus Campanula is conceptually use- genus may 66 ” less and its continued use as a “core be misleading. The same is probably true for Asy- neuma and Wahlenbergia. nowledge of inter- and intrageneric relation- ships within the family has steadily increased dur- ing the latter half of the 20th century. Cytological studies, beginning with the seminal investigations of Gadella (1962, 1963, 1964, 1966, 1967). Con- tandriopoulos (1964, 1966, 1970, 1971, 1972 976, 1980a, b, 1984), Contandriopoulos et al. Damboldt (1965a, b, 1966, 1968, 1969, 1970, 1975, 1976, 1978a, b), Phitos (1963a, b, 1964a, b, 1965), and Podlech and Dam- boldt (1964) have vastly increased our knowledge of intrageneric relationships, particularly of the ge- nus Campanula. number in the Campanulaceae is n — 17. appears to have evolved independently several times in relatively unrelated genera (e.g.. in Cam- panula, Nesocodon Thulin, Ostrowskia Regel, and Canarina L.). Forty-two percent of the pub- lished chromosome counts of the family Campan- ulaceae s.l. have this number (Lammers, 1992). The base number in the family has been suggested to be x = 8 (Bócher, 1964; S Ару 1984), but Raven (1975) suggested that x = 7 i: the ancestral number. An ancestral boe аби of x — 7 is supported by counts for e (Ku- mar & Chauhan, 1975; Hong & Ma, 1991) It was Avetisian (1948, 1967, 1973, who first drew attention to the different pollen morphol- ogies within the family and gave a schematic pre- sentation of pollen evolution based on aperture The most common chromosome and this m types. She further pointed out that pollen with col- pate and colporate apertures are typical of those taxa found in the tropics, whereas those with porate apertures are typical of taxa from temperate re- gions. Dunbar (1973a, b, c, 1975a, b, 1981, 1984) and Dunbar and Wallentinus (1976) extended Av- etisian’s work by providing excellent surveys of pol- len from numerous genera of the Campanulaceae, and this has been augmented by Morin (1987), Nowicke et al. (1992), and Morris and Lammers (1997). Several of these studies suggest that some of the genera are artificially grouped together in De Candolle’s and in Schónland's arrangements. be- cause of the limited criteria used as the basis for their classification systems Seed morphology has been examined for a num- ber of taxa, principally those of North America (Shetler & Morin, 1982, 1986) and Eurasia (Be- 1984a, b, 1985, 1986; Oganesian, 1985). Life-form in the Campanulaceae has been studied intensively by Shulkina (1974, 1975a, b. 1977, 1978, 1979, 1980a, b. c. 1986a, b. 1988) and Shulkina and Zykov (1980), but these data have not been incorporated into a cladistic analysis. Sero- lyaev, logical studies have been done on the tribe Phy- teumatae (Gudkova & Borshchenko, 1991), while Gorovoi et al. (1971) conducted a limited chemo- taxonomic survey of Russian Far-Eastern taxa. Ko- lakovsky (1980, 1982, 19864, 1986b. 1987), Ko- akovsky and Serdyukova (1980). Lakoba (1986) did some pioneering carpological investi- gations of the family, but so far this work has not been corroborated and it remains to be seen wheth- er their segregate genera will be accepted. Few molecular phylogenetic studies of the Cam- panulaceae have been undertaken. Cosner (1993) and Cosner et al. (2004) used chloroplast DNA cpDNA) structural rearrangements to establish a phylogeny of the family based on 18 genera, while Cosner et al. (1994) determined rbcL sequences for and ‚ж, several genera as part of a study of interfamilial relationships of the Campanulales. Eddie (1997, and unpublished data), using cladistic and phenetic methodologies, investigated the morphology of most of the genera of the Campanulaceae, in addition to molecular variation of 23 to 29 taxa using internal transcribed spacers (ITS) and matK/trnK-intron se- quence data from nuclear ribosomal (nrDNA) and cpDNA, respectively. For molecular variation with- in and between genera, ITS sequences have been used by Ge et al. (1997) for Adenophora Fisch. and by Kim et al. (1999) for Hanabusaya Nakai. Ha- berle (1998) examined relationships among the families Campanulaceae, Cyphiaceae, Nemaclada- ceae, Cyphocarpaceae, and Lobeliaceae using ITS sequence data. This study is an attempt to reconstruct the phy- logeny of the Campanulaceae s. str. using nrDN ITS sequences and to compare the results with cer- tain characters that have traditionally been used in Volume 90, Number 4 2003 Eddie et al. Phylogeny of Campanulaceae 557 the classification of the Campanulaceae (i.e., cap- sule morphology and presence/absence of calyx ap- pendages in addition to chromosome numbers, pol- len, and geographical distribution). It is the first time that molecular methods have been applied to a broad sample of taxa (93 species in 32 genera) within the family. This study is also the first part of more extensive investigations of the Campanu- laceae using a variety of molecular markers, in- cluding the sequences of chloroplast genes matK and rbcL, as well as chloroplast genome rearrange- ments and morphological data. Ultimately these studies should lead to a new comprehensive clas- sification of the Campanulaceae. MATERIALS AND METHODS TAXA SAMPLED AND SOURCES OF PLANT MATERIAL ITS sequences for 93 taxa of the Campanulaceae were used, including a number of which were pre- viously published and available from Genbank (Fu et al., 1999; Ge et al., 1997; Kim et al., 1999; Schultheis, 2001; K. Dotti, unpublished data) (see Appendix 1). Many of the samples represent taxa that are commonly accepted as genera or sections within genera because of their obvious morpholog- ical discontinuities and that provide a broad sample of the family. The nomenclature used in this study is based on the classification system used by Fe- dorov (1957), but the names given to the major groups or clades are merely for convenience and not based on any particular classification system. Added to the data set were four outgroups from the Lobeliaceae ко bacigalupii Weiler, Lobelia aberdarica R. E. ries, L. tenera unth, and L. nS LX bringing the total number of taxa in the data set to 97. There is overwhelming evidence from both morphological (Lammers, 1992; Gustafsson & Bremer, 1995) and molecular (Cosner et al., 1994; Gustafsson et al., 1996; Jansen & Kim, 1996; Albach et al., 2001) studies that the Lobe- liaceae are an appropriate outgroup for the Cam- panulaceae sensu stricto. DNA samples were ob- tained from living plants cultivated at The Institute of Cell and Molecular Biology (ICMB), University of Edinburgh, Scotland, U.K., The Royal Botanic Garden Edinburgh (RBGE), Scotland, U.K., The University of Texas at Austin (UT), U.S.A., and the Missouri Botanical Garden (MO), St. Louis, U.S.A. For sources of material and location of vouchers, see Appendix 1. DNA ISOLATION, AMPLIFICATION, AND SEQUENCING Genomic DNA was extracted following the CTAB protocol of Doyle and Doyle (1987) or with minor modifications such as the addition of PVP-40 and/ or BSA. Double-stranded DNA from the ITS and the intervening 5.85 subunit of the 185—265 nr DNA was amplified using standard PCR procedures (Kim & Jansen, 1994). The basic primer sequences were those of White et al. (1990) or the modifica- tions by Yokota et al. (1989). Purification of the PCR products was by means of Qiagen QIAquick spin columns (Qiagen Corp.), and sequences were obtained from ABI Prism 377 Automatic DNA Se- quencers (Perkin Elmer, Applied Biosystems Di- vision). For each taxon, forward and reverse se- quences were obtained, and the results were saved as electropherograms and edited using the pro- Codes Согр.), > QUENCE NAVIGATOR, ver. 1.0.1 (Perkin Elmer, Applied Biosystems Division). SEQUENCE ALIGNMENT The boundaries for the ITS region were deter- mined by comparisons with published ITS sequenc- es of Nicotiana rustica L. (Solanaceae, Venkates- warlu & Nazar, 1991), Krigia Schreb. (Asteraceae, Kim & Jansen, 1994), Madiinae Benth. (Astera- ceae, Baldwin, 1992), and Gentiana L. (Gentiana- 1996). Alignment of ITS proved to be problematic, particularly at the 3’ end of the ITS2 region close to the 26S subunit. Due to a high level of ambiguity, this region was deleted at 205 bases downstream from the start of the ITS2 region. ceae, Yuan et al., The highly conserved 5.8 subunit was not available for all taxa and therefore was not included in phy- logenetic analyses. The multiple alignment used in this study was created by CLUSTALX (ver. 1.64b; Thompson et al., Slow/Accurate dynamic programming option. Di- vergent sequences (> 40%) were delayed in the 1997) in several stages using the alignment procedure. Insertions from individual taxa, which created gaps and had no apparent ho- mology with the rest of the taxa, were removed, and another round of alignment was initiated. A range of gap penalties from 5.00 to 20.00 and gap exten- sion penalties from 3.00 to 8.00 were initially tried with various combinations until a consistent align- ment was established using a gap penalty of 8.00 and a gap extension penalty of 5.00. Minor final adjustments to the alignment were done manually. The чилеп і is porci ak http: Неи. мері. utexas | htm>. All new ай һауе been submitted to Genbank. 558 Annals of the Missouri Botanical Garden Table 3 of nr DNA in the Campanulaceae. Base composition and nucleotide divergence in the aligned partial sequences of ITS] and ITS2 regions Sequence parameter ITS] + partial ITS2 Aligned le e Constant sites riable sites Informative sites Unambig. transversions Ts/Tv rati Avg. base frequencies* 10 A = uninformative) 1.254 20.8 — 30.6 = 292 T= 19.3 * Missing data and gaps excluded. PHYLOGENETIC ANALYSES A search for the most parsimonious tree was ini- tiated using the PARSIMONY option of PAUP 4.068 (Swofford, 2001) with ACCTRAN, MUL- TREES, TBR, and COLLAPSE ZERO LENGTH BRANCHES options. All characters were given equal weight and were unordered. Gaps were treat- ed as missing data. The HEURISTIC search algo- rithm was chosen, with 1000 random addition rep- licates and with a limit of 2000 trees saved per replicate. The amount of support for monophyletic groups was evaluated by 1000 bootstrap replicates and a 50% cut-off value for the bootstrap consensus tree (Felsenstein, 1985). Consistency Indices (CI) (Kluge & Farris, 1969) were also computed. The Retention Index (RI) and the gl statistic (Hillis & Huelsenbeck, 1992) were also computed, the latter after computing the tree-length distribution. of 100,000 random parsimony trees by means of the RANDOM TREES command. RESULTS AND DISCUSSION The total aligned length of the 1751 and partial ITS2 (including gaps) was 497 bp. There were 81 constant characters, 71 variable characters that were parsimoniously uninformative, and 345 par- simoniously informative characters (Table 3). Par- simony analyses е 2629 trees with. 2130 steps, a CI of 0.3703 (excluding uninformative characters), and ч of 0.7583 (Figs. 1, 2). The gl statistic for 100,000 trees randomly sampled was —0.327694 indicating that the ITS data set is sig- nificantly skewed from random and contains con- siderable phylogenetic information (Hillis & Huel- senbeck, 1992). For other statistics of the aligned sequences see Table 3. Multiple ITS types were not detected, and in one case there were two separate samples of the same species (Adenophora divaricata Franch. & Sav.) that did not come out together. The branch lengths are very short for the Adenophora clade overall, which indicates that most of the spe- cies have very similar ITS sequences. The differ- ences between the two samples of A. divaricata sug- gest either misidentification of the original sample or population differences in the ITS sequences. The taxonomic categories used in classifications are unequivocal and the amount of molecular di- vergence (and hence phylogenetic signal) within and between taxa at each level in the taxonomic hierarchy varies. For a family such as the Campan- ulaceae, which has numerous monophyletic genera and sections, the use of ITS sequences is justified by the phylogenetic signal obtained, but there may be substantial trade-off due to problems with align- ments. The difficulties associated with sampling across a wide spectrum of taxa in the Campanula- ceae should lessen as we are able to refine our molecular analyses at different levels in the taxo- nomic hierarchy, in conjunction with other sources of data. Due to high ambiguity at the generic leve in the Campanulaceae, ITS sequence data may be approaching the limits of usefulness for phyloge- netic reconstruction, whereas at the species level, there may not be enough signal, and many species may be spuriously placed with each other. For ex- tensive discussion of the utility and limitations of the ITS region in the reconstruction of angiosperm phylogeny, see Baldwin et al. (1995), (2003), and Goertzen et al. (2003). Coleman MAJOR CLADES IN THE ITS TREE The topology of the strict consensus tree (Fig. 1) shows that there are two major clades of the Cam- panulaceae. This major dichotomy in the family is supported by pollen data. For convenience, these two major clades are referred to as alliances and are named on the basis of their pollen types. The taxa in the smaller of these two alliances comprise Volume 90, Number 4 2003 ddie et al. 559 Phylogeny of Campanulaceae Edraianthus pumilio qe Edraianthus graminifolius L_— Campanula latifolia 98 аа tchihatchewii cam mpanula ba roata ape|2 3s 's aeadejnuedwe> 100 100 exe) jPeuonisueJ |. SQIOInNVdWV2 32NVITIV 31VHOd O £ 3 o ك‎ z = w 8 3 2 S əpep snjnoundey Ade ru velia oe petiolata Adenophora poranio А 82 Adenophora serena Githopsis diftus sa Беа بسا‎ = [ — Roella cilia ы Heterocodon rariflorum rater ocapsa congest OOS dicentrfolia Codonopsis lanceolata 87 e pilosula Е а + EE и kawakami 32NVITIV 3LVUOd102/31Vd 103 5 71 ampanu ia тил Fan Й Campanula punctata 88 Campanula lanata Campanula alliariifolia 74 100 Campanula grossheimii Campanula hohenackeri Campanula kolenatiana Campanula siegezmundii Campanula glomerata Symphyandra pendula Symphyandra armena 74 Campanula sosnowskii Campanula C" Campanula ` Campanula raddea s Campanula ossetica : Symphyandra hofmanni 54 [—— Campanula tridentata ——— Campanula sarmatica Campanula mirabilis 1 96 Campanula mollis |. Me C Campanula edulis Feeria angustifolia L—— — Azorina vidalii Diosphaera rumeliana a (Roucela) erinus helium caeruleum Hooks lactiflora | 87 p тее M | peregrin. urea . Jasione sessiliflora : ana aes Jasione пата == зоо а Wahlenber ia hederacea a SGIONONODALV ld sqioIDu3gNa31HVM Figure 1. 4 outgroups of the Lobe numbers above the "i Rd 's are bootstrap acters), RI — ceae based on pars E | Со | жа um oo едед ФФ E . بوچ‎ BO ao Sa ?5 (dnoi6yno) rcentages 0.7583.| Nodes without Блин. ыган 's que less than 509€ imi Ромео, bacigalupii Lobelia tupa 3V32VI13801 Strict conse m of 2629 most рио trees with 2130 steps for 93 taxa of the Campanulaceae and simony ms of the “р ITS] and ITS2 sequence data. The ).3 ` 1000 replicates. [CI = 103 (excluding е pend 560 Annals of the Missouri Botanical Garden Edraianthus pumilio (2 = Edraianthus venio NE 32) Campanula latifolia (2n = ichauxia tchihatchewii ampanula barbata "m = 34) Campanula tridentata (2л = 34) Campanula sarmatica (2n = 34) Ca ке mirabilis E = 102?) nula mollis (2n = 24) ге edulis (2n = 56) а angi с оппа Diosphaei теала ana (2л = ampanula а ова) ңы ы = ж т = Y Cam Trachelium caeruleu Jasione crispa (2л ri | Jasione montana = 12) Jasione maritima (2л = 12) Jasione laevis (2n - 60) 2nz 12) : Legousia falcata (2n = 26) Cam astrum americanum (2n = 58) Physoplexus comosus Phyteuma Spicata (2n= ae Adenophora divaricata 1 (2n = =). Adenophora банны (2n= 34) Adenophora strict Roella ciliata ance marae, саан ае Codonopsls s dicont ا‎ (2п= Cyananthus ovatus (2n= 14) 2л = Бен pilosula (2n 16) с деу а (2п=16) sa (2-16) рки иеа k obivit 2 nz16) narina Mee бл = 34) Lobelia aberd: Campanula Campanula pyramidalis ‘tn 34) Campanula lusitanica (2n = Codonopsis lanceolata Codonopsis I (2n= 16) Platycodon der ra (2n = 18) Campanumoea javanica — japonica marula pinnata Campanula Ру, (2л = 32 Campanula persici Cam o rotundifolia (27 panula herminii (ane = 2 Стра. a par (2n= (2n 18) Cai eem ia tenera ==] а Downingia 5 Lobelia tupa — 10changes e 2. Phylogram of one of the 2629 equally parsimonious trees for 93 taxa of the Campanulaceae and 4 ba anis of the Lobeliaceae based on parsimony analysis of e combined ITS] and partial ITS2 sequence data. A e bar representing 10 changes is shown on bottom left corn Volume 90, Number 4 2003 Eddie et al. 561 Phylogeny of Campanulaceae genera such as Codonopsis, Platycodon, Canarina, etc., which are all distinguished by their possession of either colpate or colporate pollen (Avetisian, 1948, 1967, 1973, 1986; Dunbar, 1973a, b. c 1975a, b, 1981, 1984) and are also referred to as the platycodonoid group in this paper. The colpate/ colporate alliance is strongly supported with a 100% bootstrap value and is the only clade with taxa that have baccate fruits (Canarina, Campan- umoea Blume, and Cyclocodon W. Griff.), although the majority have dry capsules. Geographically, the colpate/colporate alliance is mostly distributed in the tropics or subtropics, from Southeast Asia and the western Himalayas to Ussuriland, Korea, and Japan, and from Indonesia and the Philippines to New Guinea. The genus Canarina is unique within this alliance in being essentially African, but it is disjunct, with one species in Macaronesia and two species in the mountains of East Africa. The taxa in the larger alliance comprise the remainder of the and they are distinguished by " Campanulaceae, their porate pollen. The porate alliance has only weak support with a 5596 bootstrap value. It is far larger numerically than the colpate/colporate alli- ance and is divided into two major groups, the wah- lenbergioids and the campanuloids. This huge al- liance is distributed mostly in the temperate regions of the world, although a few wahlenbergioid and campanuloid taxa extend to the tropics. Al taxa within these two groups have capsules that are predominantly dry and dehiscent. In the discussion that follows, we describe the major groups in the two alliances and how they compare with data from morphology, chromosome number, and geography. THE COLPATE/COLPORATE ALLIANCE (THE PLATYCODONOID GROUP) There is strong support (10096) for the mono- phyly of the colpate/colporate alliance, although the major clades within this alliance are only partially resolved. Canarina, Cyananthus, and Codonopsis all. subg. Obconicapsula D. Y. Hong form a po- lytomy with the remainder of taxa, including Co- Platy- Blume. donopsis, Leptocodon (J. D. Hooker) Lem.. codon, and Campanumoea javanica Codonopsis subg. Obconicapsula is somewhat iso- lated morphologically and, to a lesser extent, geo- graphically (central Himalayas) from the rest of Co- donopsis. It has an ovary that bulges upward above the level of the calyx lobes and an incomplete nec- tar dome. These features, together with the overall appearance of the flower, recall Platycodon. Cy- ananthus comprises highly adapted perennial and annual species of very high elevations in the moun- tains of southern Asia. Because of its superior ovary it has and low chromosome number of 2n = 14, traditionally been considered the most ancestral ge- nus of the Campanulaceae (Hutchinson, 1969; Cronquist, 1988; Takhtajan, 1969). However, it also has specialized ecological characters such as deep taproots and prostrate lateral branching, both of which are characteristic of alpine plants. The iso- lated position of Canarina is supported by both ge- — ж m — ariensis (L. remainder of the platycodonoids have 2n — 16 or 18. Bootstrap support for the clade comprising Lep- tocodon, the remainder of Codonopsis, plus Cam- panumoea javanica and Platycodon is moderate 74%). Support for the minor clade containing the bulk of Codonopsis plus Platycodon and C. javanica is strong (88%), but the clade with only the latter two genera is moderately supported (70%). The taxa of Codonopsis are morphologically less diver- и gent from each other, whereas C. javanica and Pla- tycodon are considerably divergent. Hong and Pan (1998), on the basis of pollen morphology, seed coat. and gross morphology. restored the genus Cy- clocodon, which was formerly included in Campan- umoea s.l. as C. celebica Blume and C. lancifolia Roxb.) Merr. They considered Cyclocodon to be more closely related to Platycodon than to Cam- panumoea s. str. (i.e., C. javanica Blume and C. inflata (Hook. f.) C. В, Clarke). Campanumoea and Cyclocodon have baccate fruits but would appear to be rather distant from Canarina. т; THE PORATE ALLIANCE (THE WAHLENBERGIOID AND CAMPANULOID GROUPS) The monophyly of the porate alliance is weakly supported (55%) and it comprises two very unequal clades, the wahlenbergioids and the campanuloids. This is undoubtedly an artifact of the undersam- pling of wahlenbergioid taxa. The wahlenbergioid group. The sister relationship of the two wahlenbergioid taxa, Craterocapsa Hil- liard & B. L. Burtt and Roella L., has strong boot- strap support (94%). These two genera have tradi- tionally been considered closely related (Hilliard & Burtt. 1973). Both are from southern Africa, al- though Craterocapsa ranges north to the mountains of eastern Zimbabwe. Since only three traditionally accepted wahlenbergioid genera were available for molecular analysis, the discussion of the results for this group is relatively straightforward, but caution should be observed for such a small sample. Wah- lenbergia hederacea L. falls within the campanuloid group and is therefore distant from the other two Annals of the Missouri Botanical Garden wahlenbergioid genera. This is surprising because this species has traditionally been considered as chromosome typically wahlenbergioid. It has a number of 2n = 36, which is not particularly un- usual, but it is isolated in western Europe. and has a vegetative morphology that is rather atypical for the wahlenbergioids. Although all modern Euro- pean workers have never questioned the wahlen- bergioid nature of W. hederacea, this species was recognized as a separate genus by some early work- rs (Schultesia Roth, Valvinterlobus Dulac, Aikinia Salisb. ex A. DC.) and it was assigned to Roucela by Dumortier (1827). The majority of species of Wahlenbergia are distributed in the Southern Hemi- sphere. Some species (e.g., W. trichogyna Stearn) have 2n = 36, but the majority have 2n = 18 (see also Petterson et al., 1995; Crawford et al., 1994; 2000). In the study of Cosner et 4), the Australian species, W. gloriosa Loth- Anderson et al., al. (200 ian (not sampled), was found to be in the same clade as Roella ciliata L. The campanuloid group (Campanula s. str., "tran- This huge group forms an unresolved polytomy consist- sitional" taxa, and Rapunculus clades). ing of two major clades and three smaller ones. This basic division is partially in agreement with mode of capsule dehiscence (there are exceptions such as Edraianthus in the Campanula s. str. clade and Adenophora and subsection Heterophylla in the Ra- punculus clade) and presence or absence of calyx appendages, two characters that have traditionally been used in intrageneric classifications of Cam- panula (Boissier, 1875, 1888; Fedorov, 1957). One large, well-supported clade (81%) comprises those taxa centered around Campanula s. str. (i.e., mostly those taxa belonging to the sect. Medium DC.), but also genera such as Trachelium, Diosphaera, Azo- rina, etc. The second large clade has moderate sup- port (69%) and comprises those taxa centered around Campanula sect. Rapunculus (Fourr.) Boiss. (the Rapunculus clade). Two smaller clades have strong support (100%) and consist of several tran- sitional genera such as Jasione L., Musschia, and Gadellia Shulkina, while the third small clade com- prises Wahlenbergia hederacea alone. THE CAMPANUIA S. STR. CLADE The Campanula s. str. clade includes a small number of mostly monotypic genera that are con- siderably more divergent than the majority of taxa in this clade. Some have upright flowers (e.g.. Trachelium caeruleum L., Diosphaera rumeliana (Hampe) Bornm., Feeria angustifolia (Schousb.) Buser, Campanula [subg. Roucela (Dumort.) J. Damboldt| erinus L., Campanula mollis L., and Campanula edulis Forssk.), but Azorina vidalii — Wats.) Feer, with its nodding flowers, is a conspic- uous exception. With Trachelium removed, boot- strap support for this clade is 93%. = 28) belongs to a rather distinct group of annual campanuloids of the Med- Campanula (subg. Roucela) erinus (2n iterranean, which superficially resemble C. mollis and C. edulis, but its capsules are discoid and the calyx appendages are absent. The corolla approach- es the hypercrateriform shape of Trachelium corol- las to some extent. The flowers of Diosphaera Buser are quite similar to those of Trachelium and it has ). but there are conspicuous differences between the two gen- the same chromosome number (2n — 34 era, both vegetatively and in the form of the inflo- rescence. The two genera are often united, but they are disjunct geographically in the Mediterranean. Calyx appendages are absent in both genera. Azorina Feer is quite isolated morphologically (vegetatively and in branching pattern), but its vague resemblance to Campanula bravensis Bolle Smith of the Cape Verde Is- lands, together with its chromosome number of 2n and C. jacobaea С. = об, may link it rather tenuously to Campanula subsect. Oreocodon Fed. (but see also Thulin, 1976: 354). Support for the clade that comprises Azorina, Feeria, Campanula mollis, and C. edulis is weak (58%), but when Azorina is removed support for the taxa is 100%. traditionally been associated with Trachelium, but remaining Feeria angustifolia has morphologically it is quite distinct. In some re- spects, particularly the globular, more compact shape of the inflorescence, and the valvate apical dehiscence, it approaches Jasione L., but the chro- mosome number for Feeria angustifolia is 2n = 34 (vs. 2n = sequences with those of both Campanula mollis and 12 for Jasione). The similarity of its ITS Cs . edulis does not accord with its morphology. Cam- panula mollis and C. edulis are probably closely related to each other, and this relationship is strongly supported in the ITS tree (96%). These two species belong to a group of annual and perennial = 24, 28, 54, 56), which range rom Macaronesia, North Africa, and the Iberian campanuloids (2n = Peninsula south to the equator in Tanzania. They have basal dehiscence and appendages between the 1929; Quézel, 1953 1976). This group probably links up with Campan- calyx lobes (Maire, Thulin, ula subsect. Oreocodon of the western Himalayas and south-central Asia, which is characterized by species such as C. incanescens Boiss., C. cashmer- colorata Wall. The remaining taxa in the Campanula s. str. tana Royle, and С. эй, clade are weakly supported (58%) as a monophy- Volume 90, Number 4 2003 Eddie et al. 563 Phylogeny of Campanulaceae [епс group. They are mostly Eurasian and North African, although at least one species in this alli- ance occurs as far east as the Aleutian Islands (C. chamissonis Fed. subsect. Scapiflorae (Boiss.) Fed., not sampled), and another south to the equator in northern Tanzania (C. keniensis Thulin, also not sampled). The isolated species C. mirabilis Albov (subsect. Spinulosae (Fom.) Fed.) is the sister taxon to all the others. The small clade formed by Ed- raianthus pumilio (Schultes) A. DC., E. gramini- folius (L.) A. DC., and C. latifolia L. is weakly sup- ported 2 50%). The two species of Edraianthus DC.) DC. are confined to the mountains of so suiheastórn Europe, and are rather dissimilar mor- phologically. Edraianthus pumilio has solitary flow- ers on multiple inflorescence stems, whereas E. graminifolius has a glomerulate inflorescence. Mor- phologically, E. pumilio may be closer to Campan- ula (Petkovia Stefanoff) orphanidea Boiss. (not sam- capsule pled), which has a similar mode of dehiscence (Hartvig, 1995) and similar corollas (C 2 orphanidea has 2 raianthus was for- merly considered to be wahlenbergioid because of the apical rupture of its capsule, but its overall morphology is very similar to Campanula and its chromosome number (2n = 32) is more typical of campanuloid taxa. Campanula latifolia is rather isolated in the Campanula s. str. clade. It belongs to a distinct group of tall mesophytic species from Eurasia that lack appendages and have nodding flowers on long spicate inflorescences (e.g.. C. trachelium L., C. bononiensis, C. rapunculoides 1... etc.). In general morphology this group (subsect. Eucodon (А. DC.) Several other minor groups within the Campan- Fed.) resembles Adenophora. ula s. str. clade have moderate to strong support. Michauxia tchihatchewii Fisch. & C. A C. barbata L. have a bootstrap value of 98%. relationship is surprising since the morphology of Meyer and This these two species is very divergent. The monophyly of the two, yellow-flowered species from the Euro- pean Alps, C. thyrsoides L. and C. petraea L., is moderately supported (71%). Collectively, four taxa form a strongly supported clade (85%). The long branches (Fig. 2) show clearly that these these four taxa are all very divergent from each other. In some cases, relationships in the Campanula s. str. clade are in accord with classification of Fedorov (1957), whereas in other instances there is conflict. For examp i azica Kharadze, C. sosnowskii Kharadze, and C. “bellidifolia Adam have a support value of 74%, which agrees with their placement in section Scapiflorae (Boiss.) Fed. In contrast, C. hohenackeri Fisch. & C. A. Mey. (subsect. Trilo- culares Boiss.) and C. grossheimii Kharadze (sub- sect. Eucodon) have bootstrap support of 100%, but their relationship conflicts with Fedorov’s arrange- ment. TAXA THE “TRANSITIONAL” The clade comprising Musschia, Gadellia, and e two species of Campanula sect. Pterophyllum не, (C. peregrina L. and C. primulifolia L.) is strongly supported (100%). Musschia aurea Du- mort. is an endemic of Madeira together with its congener, M. wollastoni Lowe, whereas C. peregrina and C. primulifolia are disjunct between the east- ern Mediterranean region and the western Iberian Peninsula, respectively. Gadellia lactiflora (M. Bieb.) Shulkina is endemic to the Caucasus region. Morphologically, Musschia is different from the oth- er three taxa except for a vague similarity of form, robustness, and disposition of the stigmatic lobes. Its capsule is 5-loculed, prismatic, and opens with numerous transverse slits. Its chromosome number — 32. Gadellia was erected by Shulkina (1979) for Campanula lactiflora M. Bieb. based on its distinct growth form and chromosome number (2n — 36). It has open, upright flowers and dehisces somewhat medially/apically. Campanula primulifol- ia was placed in the genus Echinocodon (= Echin- ocodonia Kolak.) by Kolakovsky (1986b). Campan- ula peregrina was acknowledged to be very close to C. primulifolia by Damboldt (1978b) and was placed in the section Pterophyllum. Bootstrap sup- port for a close relationship between these two spe- cies is 87%. Despite their strong resemblances, the chromosome number for C. primulifolia is 2n = 36, while C. peregrina is recorded as 2n = 26 (Gadella, 1964). However, Marchal (1920) recorded the for- mer also as 2л = 26, so these findings require clar- ification. The genus Jasione L. is strongly supported as a monophyletic group (100%). Within the genus, J. crispa (Pourr.) Samp. is sister to all the others sam- pled, but the clade formed by them is weakly sup- ported (64%) and relationships among species within the group are unresolved. The relationship of Jasione to other taxa of Campanulaceae is un- resolved in the ITS tree. Jasione has most frequent- ly been associated with the wahlenbergioid alli- ance, although it does bear some resemblance to Feeria Buser with which it shares a similar mode of capsule dehiscence, but it has а chromosome number of 2n = 12 (vs. 2n = 34 for Feeria). THE RAPUNCULUS CLADE The Rapunculus clade has moderate support (6996) and has a number of smaller clades that are 564 Annals of the Missouri Botanical Garden all relatively divergent from each other morpholog- ically. In terms of branch length, the taxa within the Rapunculus clade are much more divergent overall than the taxa within the Campanula s. str. clade (Fig. 2). Githopsis Nuttall and Heterocodon Nuttall are rather divergent in morphology from each other, particularly that of the capsule (see McVaugh, 1945), but are probably closely related and have strong bootstrap support (100%). They are sister to the remaining members of the Rapunculus clade. Most of these taxa are either Mediterranean or North American in distribution. The majority of taxa within this clade have open, upright flowers that are rather stellate in form, and the capsule opens apically or medially by a pore, but there are conspicuous exceptions (see below). None of the taxa in the Rapunculus clade has calyx appendages. The irregular rupture of the capsule apex in Gith- opsis may represent a derived condition, but this is not to imply that its ancestral state was lateral (e.g., it may be derived from an apical valvate condition similar to that present in the wahlenbergioid alli- ance). In Adenophora, Hanabusaya, and Campan- ula rotundifolia L. (the sole representative of the harebell group sampled, Campanula subsect. Het- erophylla Fed.), the nodding and the capsule opens basally. The inclu- sion of these taxa within the Rapunculus clade is surprising. Morphologically these taxa seem to be owers are campanulate and more closely allied to groups within the Campanula s. str. clade (e.g., C. latifolia and its allies in sect. Eucodon). When Githopsis and Heterocodon are removed, the remaining taxa of the Rapunculus clade have 100% bootstrap support. Within this clade the Tex- an endemic annual Campanula reverchoni А. Gray is sister to all the remaining taxa, although support for this group is weak (< 50%). Within this clade there are several small groups with moderate to strong support. The clade comprising Adenophora and Hanabusaya is strongly supported (99%), al- though species relationships are largely unresolved. This confirms the close relationship between Han- abusaya and Adenophora suggested previously by Eddie (1997) and by Kim et al. (1999), and it ten- tatively suggests that Hanabusaya is closest to the two species А. stenanthina (Ledeb.) Kitagawa and A. paniculata Nannf. (sect. Thyrsanthe (Borb.) Fed.). Support for the clade uniting these three taxa is weak (« 50%). The remaining species of Aden- ophora form an unresolved polytomy, although there is weak support for a group consisting of А. hima- layana Feer (sect. Pachydiscus Fed.) and A. lobo- phylla D. Y. Hong (sect. Microdiscus Fed.). The sister group to the Adenophora/Hanabusa ya clade is only weakly supported, but it contains sev- eral well-supported smaller groups. These taxa are divergent morphologically and have a wide range of chromosome numbers. The group containing the serpentine endemic from the Balkans, C. hawkin- siana Hausskn. & Heldreich (2n = 22), and Ibe- пап endemics C. lusitanica Loefl. (2n = 18), C. herminii Hoffmanns & Link. (2n = 32), and C. ar- vatica Lag. (2n = 28), is strongly ws (98%), while the clade with C. stevenii M. Bieb. (2n = 32) (2n = 16, 18) Tas a support The two morphologically divergent and C. persicifolia L. value of 99%. species, C. arvatica and C. rotundifolia (2n = 34), are sister species with 77% bootstrap support. Campanula carpatica Jacq. (subsect. Rotula Fed.) does Е appear to be as close to C. pyramidalis L. P Y 34), but it does resemble C. herminii from i [Бап Peninsula. Campanula pyramidalis is part of the “isophylloid” group of species (e.g., С. isophylla Moretti, C. garganica Tenore, C. versicolor Andrews [not sampled], etc.), which is centered in Italy and the western Balkan Peninsula and i somewhat intermediate between the Phyteuma L./ Asyneuma alliance and those species that could be considered as typically rapunculoid (e.g.. Campan- 1965a). many species in this group hybridize ula carpatica, etc.) (see also Damboldt, However, freely. and numerous hybrids involving C. carpatica are known in cultivation (Lewis & Lynch, 1989). Thus, the ITS data suggest that this grouping is a natural one. Broader sampling would perhaps have helped clarify the positions of the “isophylloid” and Heterophylla groups. The Phyteuma clade includes morphologically similar species and has strong (p da support (9196). Petromarula Vent. ex Hedw. all the other taxa, followed by Asyneuma japonicum (Miq.) Briq. (Endl.) Schur and Phyteuma has a bootstrap sup- f. is sister to The clade comprising Physoplexis port of 81%, but relationships within this group are unresolved. The long branches in this clade (Fig. 2) suggest these taxa are relatively divergent. The sister group of Phyteuma and closely related genera includes Eurasian genera such as Legousia Dur. and several diverse North American taxa, such as Triodanis Raf., Campanula divaricata Michx., and Campanulastrum americanum (L.) Small. This clade is weakly supported with a bootstrap value of 50%. sometimes considered to be congeneric with Legou- sia (McVaugh, 1945, 1948), these taxa are all rather divergent morphologically. In Asyneuma, Phyteu- ma, Petromarula, the species such as Campanula pyramidalis, and the less than Apart from Triodanis, which Physoplexis, "isophylloid" American taxa such as Campanulastrum and Triod- Volume 90, Number 4 2003 Eddie et al. 565 Phylogeny of Campanulaceae anis, the capsule opens apically or medially by a more irregular pore. Morphologically, C. divaricata resembles Adenophora somewhat, and the capsule opens basally. In other respects, such as the open stellate shape and upward orientation of the flower, the majority of the other taxa in this clade are typ- ically rapunculoid (e.g., C. rapunculus L., C. patula L., etc.). CONCLUSIONS Overall, there is a remarkable congruence be- tween the ITS tree and traditional ideas on species relationships within the Campanulaceae (Eddie, 1 The insights of early workers such as De Candolle and Boissier have proved to be remark- ably clear, and their classification systems have, on the whole, been logically consistent with our find- ings on phylogeny. This study also supports the se- rological studies of Gudkova and Borshchenko (1991) and the cpDNA аи of Соѕпег (1993) апа Cosner et al. (200 The ITS trees indicate that i colpate/colporate alliance (the platycodonoids) is sister to the re- mainder of the Campanulaceae (Eddie, 1997, 1999; Shulkina & Gaskin, 1999). This is in agreement with phylogenies of the Campanulaceae based on cpDNA structural rearrangements (Cosner et al., 2004). In comparison with the porate taxa, the col- pate/colporate taxa show considerably more molec- ular divergence, although the wahlenbergioid taxa were under-sampled. As a group, the colpate/col- porate alliance has not radiated into drier, more temperate regions and its area of greatest diversity remains the region between the eastern Himalayas and southwest China. It is hypothesized that Os- trowskia (not sampled) represents a minor element of this alliance, which has evolved in the dry, tem- perate, and highly seasonal environments of Central Asia and thus displays features that parallel certain porate taxa, particularly the mode of capsule de- hiscence. Canarina is clearly part of this alliance and was misplaced in the classifications of De Can- dolle (1830) and Schónland (1889—1894), although its chromosome (2n — 34) is anomalous within the platycodonoid group. These results also suggest that pate/colporate taxa (see Hong & Pan, 1998). Within vaccate fruits evolved several times in the col- this alliance there are combinations of certain mor- phological features that also occur in the porate taxa, e.g., valvate apical dehiscence, a nectary pro- tected by expanded basal filaments (nectar dome). and colored pollen, and these may afford some clues about possible links between the two major alliances of the family. The wahlenbergioids probably branched off early in the evolution of the porate alliance and consti- tute the only major group in the Southern Hemi- sphere. They have radiated most in southern Africa, although distinctive taxa occur on islands of the Atlantic, Indian, and Pacific Oceans. Several spe- cies of Wahlenbergia have ovaries that are almost superior, while Nesocodon from Mauritius has flow- ers that recall some species of Codonopsis in the colpate/colporate alliance. In contrast, the campan- uloids are dominant over much of the Northern Hemisphere. The relative isolation of monotypic or small, distinctive genera within the two main cam- panuloid clades (e.g., the Rapunculus and Campan- ula s. str. clades) suggests that the group as a whole evolved in the Mediterranean Basin and spread rapidly over the Northern Hemisphere. The Rapun- culus clade is considerably heterogeneous both cy- tologically and morphologically, although all taxa within this clade are exappendiculate. Many of the species were included in section Rapunculus (Fourr.) Boiss. (Boissier, 1875). It is the most geo- graphically widespread clade, most diverse in the Mediterranean Basin, and the only one that has spread into North America (apart from Campanula chamissonis in the Aleutian Islands). The numeri- cally small but diverse campanulaceous taxa of North America probably contain many relicts from pre-glacial times and represent several relatively independent groups derived from the main rapun- culoid radiation in Eurasia (Shetler, 1979). An ear- ly radiation of the Rapunculus group in the North- ern Hemisphere would explain the distinctiveness of subgroups (e.g., Phyteuma, Petromarula, and re- lated genera) that are associated with the European Alpine orogenic events and fluctuating Mediterra- nean sea levels during the Tertiary period (Eddie, 1984; Favarger, 1972; Greuter, 1979). It would also explain the presence of endemic genera such as Githopsis in California and the other rather hetero- geneous taxa in North America, e.g., Heterocodon and diverse Campanula annuals in California (see Morin, 1980), China, and southern Asia (e.g., Hom- ocodon D. Y. Hong and Peracarpa J. D. Hooker & T. Thoms.). The ancestral group(s) that eventually gave rise and the harebell group riae о may be re- » Adenophora, Hanabusaya, lated to some of the American taxa such as C. divaricata and C. E Small (not sampled), and may also have been ancestral to the predomi- nantly appendiculate Campanula s. str. group, of which the mesophytic, exappendiculate species such as C. latifolia, C. trachelium, etc. (sect. Eu- codon), may be the least morphologically modified descendants. 566 Annals of the Missouri Botanical Garden Species of the Campanula s. str. clade are mostly appendiculate, have basal dehiscence, and are cy- tologically more homogeneous, particularly those species in Campanula and Symphyandra. Many of them were included in Campanula sect. Medium (DC.) Boiss. (Boissier, 1875). Much of the radiation of this group is associated with the mountain-build- ing processes of Eurasia, from the Atlas Mountains in the west to the western Himalayas. Subcenters of high diversity for the Campanula s. str. clade include the Balkan Peninsula, Anatolia, and the Caucasus Mountains. Campanula, as it is currently constituted, is clearly polyphyletic. The more di- vergent taxa in this clade are found mainly in the Mediterranean basin and are placed in small or monotypic genera (e.g., Azorina, Diosphaera, Ed- raianthus, Feeria, and Michauxia). Since De Can- dolle's monograph of 1830, Edraianthus has been associated with the wahlenbergioid group, but it is clearly campanuloid, although its exact relation- ships within the Campanula s. str. clade remain unclear (see also Hilliard & Burtt, 1973). Symphyandra А. DC. is now generally consid- ered to be artificial (Greuter et al., 1984; Ogane- sian, 1995), and this analysis supports that conclu- sion. However, the the recognized by Fedorov (1957) are all quite distinct, and we suggest that the species formerly included four sections of genus in this genus should be re-examined and not nec- essarily included in Campanula without substantial evidence. The generic status of Symphyandra odon- tosepala (Boiss.) E. Esfandiari (not sampled) and the Iranian endemic genus Zeugandra Р. Н. Davis (not sampled) also need to be reassessed. Sym- phyandra hofmanni Pant. seems to be rather distant from the bulk of species in Campanula, whereas S. pendula (M. Bieb.) DC. and S. armena (Stev.) A. DC. are much closer. Several genera may best be regarded as transi- tional between the wahlenbergioid group and the campanuloid group. Musschia is probably better placed with the campanuloids, but it is somewhat intermediate morphologically between the two ma- jor porate groups and shows some resemblance to wahlenbergioids such as Heterochaenia A. DC. from Réunion. It does not appear to be close to Platy- codon or Microcodon A. DC. as in the arrangement of Schénland (1889-1894). On the basis of ITS se- quence similarity to Gadellia, we suggest that the distinct morphological evolution of Musschia on Madeira was relatively rapid. Jasione also appears to be basal within the complex of Northern Hemi- sphere genera but its exact relationships remain unclear. On the whole it appears to have more af- finities with campanuloid taxa. In the cpDNA tree of Cosner et al. (2004), Jasione forms an unresolved polytomy with Symphyandra, Edraianthus, Cam- panula, and Trachelium. Chromosome numbers (Fig. 2) are lowest overall in the colpate/colporate alliance, although the low- est recorded diploid number is for Jasione (2n = 12). Within the Rapunculus clade, with the excep- tion of the clade comprising Adenophora and Han- abusaya, the chromosome numbers are diverse and are consistently lower than numbers recorded for the Campanula s. str. clade, which аге predomi- nantly 2n = 34. If we accept the premise that there has been a general increase in chromosome number during the evolution of the Campanulaceae, then the platycodonoids are ancestral to all other groups and the wahlenbergioids and rapunculoids are an- cestral to the campanuloids s. str. This accords well with our knowledge of pollen morphology and evo- lution in the family, as well as the morphology of the capsule in the different groups. However, the diploid number 2л = 34 occurs in several unre- lated lineages (Campanula, Nesocodon, Canarina, and Ostrowskia) and probably evolved indepen- dently in each of these genera. This analysis suggests that the ancestral home of the Campanulaceae may be in the region of eastern Asia (of current geography) (see also Hong, 1995; Cosner et al., 2004), but such an interpretation can- not be easily reconciled with the distribution of many genera within the family or with closely re- lated families such as the Lobeliaceae, Cyphiaceae s. str, or Nemacladaceae (Eddie, 1984, 1997, 1999). Carolin (1978), citing the distribution of Cy- ananthus in India, concluded that the Campanu- family that evolved primarily in western Gondwanaland. Bre- mer and Gustafsson (1997), using nucleotide sub- stitutions in rbcL, suggested an East Gondwanaland aceae are essentially ап African origin at the end of the Cretaceous for the astera- ceous alliance of families, and that the group sub- sequently diversified and expanded to West God- wanaland before the breakup of the supercontinent. On the basis of atpB-rbcL spacer sequence data, E. B. Knox (pers. comm.) has stated, *. .. The inter- pretation is that Сурма and the Lobeliaceae orig- inated in southern Africa because the eight ‘basal’ lineages are entirely or predominantly African, and many of these are restricted to southern Africa."; 7... The Lobeliaceae, Cyphiaceae, and Campanu- m 'eae go back at most 40—50 MYA, and I do not think that the id s patterns can be attri- buted to Gondwanaland." If the family had arisen in Asia one would have expected platycodonoids to be represented in Eurasia and in North America. The presence of the colporate genus Canarina in Volume 90, Number 4 2003 Eddie et al. Phylogeny of Campanulaceae 567 Africa and Macaronesia suggests that the family may have been more widespread in Africa and around the Indian Ocean than now, but this addi- tional hypothesis does not conflict with ап Asian, rican, or a Gondwanaland origin for the family. The major dichotomy in the family between the col- pate/colporate and the porate taxa suggests that ma- jor tectonic processes in the early to mid Tertiary period are implicated in its evolutionary history. A fragmenting West Gondwanaland origin, with the Asian platycodonoid taxa as relictual in land mas- ses that now form the region of the eastern Hima- layas and western China, seems a more likely sce- nario, and this would accord well with the hypothesis (Eddie, 1997) that the more basal mem- bers of the wahlenbergioid group are essentially southern or oceanic in their distribution (e.g., /Ve- socodon, Heterochaenia, Berenice L. R. Tulasne, and the shrubby species of Wahlenbergia from New Zea- land, St. Helena, and the Juan Fernandez Islands). The endemic genera of the Cape Region of South Africa probably represent a very early radiation of the wahlenbergioid group in the fynbos vegetation as the climate there cooled and became more arid during the mid to late Tertiary (Eddie & Cupido, 2001). The ITS phylogeny does not necessarily reflect a species phylogeny (Doyle, 1992), but it does pro- vide a basis for inferring possible relationships within and between taxa at several taxonomic levels and provides insights for future investigations. It also provides a phylogenetic framework that can be tested with other data sets. We must await more extensive taxon sampling and data from other genes (both nuclear and chloroplast), as well as intrage- neric analyses and chloroplast genome rearrange- ment studies in order to refine these results. At the same time it must be emphasized that refined data sets of floral morphology and developmental studies are also desirable before a new classification of the Campanulaceae can be proposed. Literature Cited Albach, D. C., P. S. Soltis, D. E. Soltis & R. G. Olmstead. 2001. Phuleseiietio analysis of Asterids based on se- quences of four genes. Ann. Missouri Bot. Gard. 163-212. Anderson, С. J., С. Bernardello, P. Lopez, T. F. Stue Сабы 2000. Reproductive biology of Wahlen- bergia (Campanulaceae) endemic с таш Crusoe Island (Chile). Pl. Syst. Evol. 233: 109-123 P E. 1948. Palynologica caucasica Ш. Pollen of e Caucasian representatives of the family Ls ы. Trudy Bot. Inst. Akad. Nauk Armjansk. SSR. 5 199-206. [In Russian.] Morphology of the pollen of the family Campanulaceae and closely related families ene cleaceae, Lobeliaceae, Cyphiaceae) in connection with questions of their systematics and apg ont и Bot. Inst. Acad. Sci. Armenia 16: 5-41. [In Russ . 1973. Palynology of the Order Campanula s.l. 90-93 in Т, Kuprianova (editor), Pollen and SN re Morphology of the Recent Plants. Proc. 3rd Int. Pa iip 1971. [In Russian. | 6. Palynomorphology of = families Campan- d. it ur e, and P hragmataceae. Bot. Zhurn. (Moscow & pny Т1: 1003—1010. [In Russian. | Baldwin, B. G. 1992. Phylogenetic utility of the internal a spacers of nuclear ribosomal DNA in nts: An TE from the Compositae. Molec. Phy- juan Evol. 1: „М. J. E J. M. Porter, M. F. lk. ре а ki, с. S. Campbell & M. J. Donoghue. 1995. The ITS region of nuclear ribosomal DNA: A valuable source of evidence on angiosperm phylogeny. Ann. Missouri Bot. Gard. 82: 247-277. Belyaev, A. A. 1984a. Seed anatomy in some ^s ned tives of the Campanulaceae. Bot. Zhurn. (Mosco Leningrad) 69: a [In Russian. ] =, 19 Surface ultrastructure and some morpho- lo EM Вебе of seeds in the representatives of the family Campanulaceae. Bot. Zhurn. (Moscow & mn 69: 890—898. [In Russian.] 985. New data on the anatomical structure g cover and ultrastructure of the seed surface o two pee s the pou Bri و‎ ragma (Cam- panulaceae). Bot. rn. (Moscow & Leningrad) 70: 2 an. 986. Features of the anatomy and ultrastructure of the surface of а seed coat in some species of critical ih era of the family Campanulaceae. Bot. Zhurn. (Mos- му & Leningrad) m 1371-1375. n Russian.] Bae her, T. 1964. ( the ок persicifolia group. Tidskr. 58 um 18. ко e 384— ns and aberrations Svensk Bot. 875. Campanulaceae. /n: Flora orientalis 3 a Georg, Geneva and Basel. А er 8. px b In: Flora orientalis, sup- plementum. H. Geo eneva and Basel. Bremer, K. & M. H. c. ey ae 1997. East Gondwana ancestry of the Wee alliance of families. Proc. Nat. Acad. Sci. U.S.A. 94: 9 Candolle, A. P. de. 16. Monographie des Campanulées. Paris. Carolin, R. C. 1978. The systematic relationships of Bru- пота. Brunonia 1: 9—29. C haradze, A. 1949. A treatment of the systematics of the оо эи Sist. Geogr. Rast. 28: 89-102 ——— 6. The genus Campanula L. s.l. in A Cau- casus "al tus). Zametki Sist. Genus Rast. 32: 45— 56 Calemin; A. W. 2003. ITS2 is a double-edged tool yj eukaryote evolutionary comparisons. Trends Genet. 315. C bntandHoponies, J. 1964. Contribution à l'étude caryol- Campanulées de Grèce. Bull. Soc. Bot. France 111: 222-235. 966. Contribution à l'étude nae des Campaniae de Grece. II. Bull. Soc. Bot. France 113: 453-493 ogique des Annals of the Missouri Botanical Garden 1970. Contribution à l'étude о des Campanulacées s du Proche-Orient. Bull. Bot. France 117: 55- 1971. Contato a Ps و‎ des E ées du Proche-Orie s Asyneuma de Turquie. Bull. Soc. Bot France | 17 i 300-220 (1970 | publ. 1971). —— ——. 1972. Contribution à l'étude cytotaxonomique des € Өк poen ées du Proche-Orient: 3. Bull. Soc. Bot. France 119: 75-94. 1976. эз à l'étude cytotaxonomique des Сараш ées du Proche-Orient: 4. Bull. Soc. Bot. France 123: 33—46. 1980a. Contribution à l'étude cytotaxonomique des Campanulac ées d'Iran. Rev. Biol. Ecol. Medit. 7: —36. 1980b. Contribution à l'étude cytotaxonomique du к C сатрапа en Afrique du nord et centrale. Bol. S rot. 53: —906. n ол. and evolution of the genus Campanula i in the Mediterranean Region. Pp. 141—158 in Plant Biosystematics. Academic Press, Canada. — ——, P. Quézel & A. Pamukcuoglu. 1972. и lacées nouvelles du pourtour méditerranéen oriental. Ann. Univ. курса Sei. 46: 53-61 (1971, publ. 1972) — ——— & J. Zaffran. 1974. A propos des cam- panules du e aizoon en Grèce meridionale et en B Bu "i Soc. Bot. France 120: 331—340 (1973, Pull 197 = qm rger & N. Galland. 1984. Contribution à m oye Moi des ‚шшш eae du Mar- *. Bull. Sci. (Rabat) 8: 10 4. бм М. E. 293. Phylogenetic ind Molecular Evolu- tionary Studies of Chloroplast DNA Variation in the Campanulaceae. Ph.D. Dissertation, Ohio State Univer- sity. س ‚ R. K. Jansen & T. G. Lammers. 1994. Phyloge- netic relationships i in a тер based оп rbcL sequences. Pl. Syst. 90: 79— === à Кыа & R. К. ш 2004. Chloro- plast DNA rearrangements in Campanulaceae: Phylo- genetic wes of highly rearranged genomes. Molec. Phylogene E Crawford, D. J., T. TE Stuessy, T. G. Lammers, O. M. Silva & P. Pacheko, PR Allozyme variation and evolution- ary relationships among three species of Wahlenbergia ^ a rime 2d 1 the Juan Fernandez Islands. Bot. . 151: 119 into A. e The Evolution and Classification of ‘lowering Plants, 2nd ed. The New York Botanical Gar- len, the bed Damboldt, J. 1965a. Zytotaxonomische Revision der iso- phyllen Visi 9. in Europa. Bot. Jahrb. Syst. 84: 302—358. 55b. Campanula tommasiniana Koch und C. volducinans . & S. zur Zytotaxonomie zweier medi- terra т Reliktsippen. Oesterr. Bot. Z. 112: —400. ——— ›. Zur Systematik und кш уоп Сат- pu arvatica Lagasc. Ber. Deutsch. Bot. € 112: 299—308. . 1968. Vorarbeiten zu einer Revision od Gattung Asyneuma (Campanulaceae). I. Willdenowia 5: 35-5: ——— 'orarbeiten zu einer Re vision dar Сон ig ине (Campanulaceae). Zur systematischen Stellung von Asyneuma scoparium (Boiss. & Hausskn.) Bornm. Willdenowia 5: 241—243. yes. 79: — — — — c Aegean * oie Revision der Gattung Asyneuma. Boissiera 17: I-12 75. е and classification in the ge- nus о. L. Inte use Botanical Congress: XII. Leningrad 1975. i51 Materials for a ias af Turkey 32: Ca panulaceae. Notes Roy. Bot. Gard. Edinburgh 35: 39- 52. )78a. Asyneuma Griseb. & Schenk. Pp. 65-81 in P. з Davis (editor), Flora of Turkey and the East Aegean Islands, Vol. 6. Univ. Edinburgh Press, Edin- burgh. 978b. Campanula L. Pp. 2—64 in P. H. Dav i, ы of Turkey and the East n TEM Vol. 6. Univ. pin a Press, ШТ Doyle, | J. 19€ ene trees and spe ees: Molecular systematics as one-character asa P st. Bot. 17: 144—163. a Doyle. 1987. A rapid isolation procedure for small ne of n шы tissue. Phytochem. Bull. Bot. Soc. Amer. 19 Dumortier, B. C. J. 1827. Flo ra Bs Dunbar, A. 197 Ba. A short report on the fine structure of some С pow ecd ollen. Grana a : 25-28. — ———. 1973b. Pollen ontogeny in e species of Cam- кер она A study by пн idm Bot. Not. 126: 277-315 ‚ 1973c. A review of the ultrastruc "ure and ontog- eny of some ME xollen. Grana 13: 85-9: ——. 1975 1 pollen of Campanulaceae and re slated asg 8 with spec ial reference to the surface ultrastruc- ES о eae subfam. Campanuloideae. Bot. Not. : 73-101. eo b. On pollen of Campanulaceae and related а with special reference to the surface ultrastruc- ture: П. Campanulaceae subfam. Cyphioideae and sub- fam. Lobelioideae, Goodeniaceae, Sphenocleaceae. Bot. Not. 128: 102-118. . 1981. The p ation of soluble material on 1 the surfac ‘e and in the cavities of the pollen wall of Cam- ш ы and ша eae. Micron 12: 47- › = oan morphology in Campanulaceae IV. Nordic i acd -19 = стане 1976. On pollen of Сат- vene eae E 2 numerical taxonomic investigation. Bot. Мо 1. vip —72. Eddie үе A Systematic Study of the Ge nus Musschia жге with versity and Evolution in the Campanulaceae: Campan- uloideae. M.Sc. Thesi i ing. oC . 1997. A Global reassessment of the Generic Re- jatünahins in the Bellflower Family (Campanulaceae). Ph.D. Thesis, University of Edinburgh. . 1999, Molecular and morphologic ‘al approaches to an understanding of phylogeny in the е ceae. International Botanical Congress: XVI. St. Lou 1999, Ab: we 16.1.1. & C. N. Cupic йо, 2001. Some observations on the reproductive аа т of hs wahlenbergioid genera of the family Campanulaceae s. str. from the fynbos veg- etation of Sou п Africa. tan 20001, Albuquerque, New Mexico. Abstract Ingrouille. 1999. Рао in the vel oculed" species of the genus Campanula, sect. билиши (Campanulaceae). Nordic J. Bot. 19: 153-169 Volume 90, Number 4 2003 Eddie et a Phylogeny of Campanulaceae 569 Favarger, C. 1972. Endemism in the montane floras of Europe. 4 in D. H. Valentine (editor), Tax- onomy, Preis and Evolution. Academic Press, London. p A. A. 1957 (1972). Campanulaceae. Pp. , 459-475 in B. K. Shiskin & E. G. Bobrov, а m Vol. е vue and Leningrad. [Flora of the ie S.S.R., Vol. 24: Dipsacaceae to Campanulaceae. Is- ael Progr. for 3 ‘ientific Translations, Jerusalem, 1970. | быаны, 1. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783— 791. Fu, R. Z., J. Wang, Y. B. Zhang, Z. T. Wang & P. C. Shaw 999, Differentiation of medicinal Codonopsis species from adulterants by polymerase chain reaction-restric- tion fragment length polymorphism. Planta Med. 65: 18-650. Gadella, Т. W. J. 1962. Some Ma ciue observations in the pu EU serie Proc. Kon. Ned. Akad. We- tensch., C. 65: 269-27 .] E Some clog observations in the genus Ned. A H Puis П. Proc. Kon. Akad. Wetensch., 270-283. . 1964; 1962- 1965. Cytotaxonomic studies in the genus Campanula. Wentia 11: 1-104; 9—13. 1966. Some notes on the n of ge nera in the Campanulaceae. I, П. . Kon. Ned. Akad. Wetensch., C. 69: 502—521. 1967. The taxonomic significance of two artifi- cub produced hybrids in the genus шы. Acta Bot. Neerl. 15: pear Ge, S., B. A. Schaal & I . 1997. A —€— of the status of Морат pins based on ITS dp HR s, with reference to the utility of ITS low n Adenophora. г ча Phytotax. Sin. 35: 2 5-395, iu L. R., J. J. Cannone, R. R. tell & R. . 2003. ITS secondary structure putet Lon comparative analysis: Implications for sequence align- pd and PUE of the Asteraceae. Molec. Phylo- . Evol. 29: 216-234. Geri, P. G a Ponomarchuk & L. І. Strigina. 1971. Dye duds study on Russian ira rn Cam- panulac D m chemistry 10: 2419-24 Greuter, W. The origins and жаша " island flo- газ as ee by the Aegean archipelago. Pp. 87— 106 in D. Bramwell (editor), Plants and Islands. Aca- demic Press, : ondon. urdet & G. Long (Editors). m Med- Chec list, T 16 em Jard. Bot. Genève. 120- Campan- u Gudkova, I. Y. & G. Вела, 1991. The serolog- ical study of the Campanulaceae. The ше ге- urn. (Moscow & banno 76: 809-817. [In et in Gustafsson, M. H. G. & K. Bremer. 1995. Morphology and m interrelationships of the Asteraceae, Ca- lyceraceae, Campanulaceae, Gooden iA and related families бе Amer. J. Bot. 82: 250-265. ‚ A. Backlund & B. Bremer. 1996. Phylogeny of the б оза sensu lato based on rbcL sequences with particular reference to the Goodeniaceae. РІ. Syst. Evol. 199: put Haberle, R. С 8. Phylogenetic Systematics of Pseu- iua ge the American Cyphioids (Campanulaceae, sensu lato). M.S. Thesis, Northern Ar- izona University, Fla, taff. Hartvig, P. 1991. ie A Pp. 369-387 in A. Strid & Kit Tan (editors), The Mountain Flora of Greece, Vol. 2. Edinburgh Univ. m Edinburgh. Hilliard, O. & B. L. Burtt. 1973. Notes on some plants of southern Africa, roh im Natal. П. Notes Roy. Bot. Gard. e —338. Hillis, = Conse k. 1992. Signal, noise, and viai in molecular phylogenetic analyses. J. Hered. 8 95. is D. сз 1995. The geography of the Campanulaceae: Jn the distribution centres. Acta Phytotax. Sin. 33 е ——4@[ 1991. Systematics of the genus Cy- ananthus 1 ех Royle Acta Phytoax. Sin. 29: 25-51. — & K.-Y. Pan. 1998. The restoration of the genus Cyclocodon олаг. каны and its evidence from pol- len and seed-coat. Acta Phytotax. Sin. 36: 106-110, illus. Hutchinson, J. 1969. The Families of Flowering Plants, 2 s. pio d ani Press, London Jansen, R. K.-J. Kim. 1996. Implications of chlo- roplast ona gee for the classification and phylogeny of the Asteraceae. Pp. 317-339 in D. J. N. Hind & H. Beentje (editors), Compositae: Systematics. Puc 'eedings of the International Compositae Conference, Kew, Vol. 1. Royal Botanic Gardens, Kew. Kim, K.-J. & R. K. Jansen. 1994. Comparisons of per genetic hypotheses among different data sets dandelions (Krigia, Asterac po Additional insu from internal hir ribed spac cfr ira r ribosomal DNA. Pl. Syst. vol "190: Kim, Y.-D., nu d . Lee, xd Jansen. 1999. Molecular hm e for the reel position of Hanabusaya asiatica Nakai (C oe ceae), an endemic species in Korea. J. Pl. Biol. 42: 168-173. Kluge, A. G. & J. S. Farris. 1969. Quantitative pm and the evolution < anurans. Syst. Zool. 18: . Novye dannye k SE ko- : Nard dannye po karpologii HODIE е hikovykh (New evidence on the carpology of relict Campanulaceae) Soobshch. Akad. Nauk Gru- zinsk. S.S.R. 105: 116-120 68. Carpology of the Сатрапшасеае and ра of taxonomy. Bot. Zhurn. (Moscow & Lenin- : 1155-1166. [In Russian.] I86b. Echinocodon, a new Mediterranean genus "EM с eae family. Soobshch. Akad. Nauk ine ae 121: 385-388. — grad) 7 7. Sistema semeistva Campanulaceae starogo sveta. tee of the Campanulaceae family from the Old World). Bot. Zhurn. (Moscow & Leningrad) 72 1572-1579 & Serdyukova. 1980. Novye dannye k s tematike kavkazskikh kolokol'chikov (Ad systematicam specierum Caucasicarum generis Е L. notitia Ade novae). orem Es Geogr. Rast. 36: Kumar, V. & K. . Chauhan. 1975 5. cen ft of Cy- n inis Wall Cat. Proc. Indian Sci. Congr. Assoc. 62, Group I ы. R. V. 1986. г жн of the axicorn—An or- gan of the Campanulaceae. Soobshch. Akad. Nauk Gru- zinsk. S.S.R. 123: 141-144 Lammers, T. С. 1992. Circumscription and phylogeny of 570 Annals of the Missouri Botanical Garden и Ann. Missouri Bot. Сага. 79: 388- Thin P. & M. Lynch. 1989. Campanulas x. Croom Helm, Bromley, Kent. aire, Ң. 1929. r peg d Nord 1 20: 17 Marchal, ja Hm Recherches sur les variations numé- ue des iced iun dans la série végétale. Mém. cad. Roy. S е ique, СІ. Sci., Ser. 2: 3-108. Mc зне К. 1945 . The genus Triodanis Rafinesque, and its o relalionships to Specularia and C ER Wrightia : E oon E Per à l'étude de la flore de ». Bull. Soc. Hist. Nat. Afrique N. 48. Generic status of Triodanis and Specularia. prex 50: 38—49. Morin, N. R. 1980. Systematics of the annual California campanulas (Campanulaceae). Madrofio 27: 149-163. — —— —. 1987. Pollen surface morphology in North Amer- ican Campanula, a and Heterocodon (Campan- ulace mer. . 74: 746-747. Morris, о" а 1997. Circumscription of Coc т 515 and rd allied genera ni асани and Leeptocodon (Campanulaceae: Campanuloideae Palynological а Bot. Bull. Acad. Sin. 38: 271-284. Nowicke, J. Shetler & N. Morin. 1992. Exine structure d е че Сатрап же ио еае) t. G nd 5-80 ral fea atures al seeds i S species. Ann. Missouri Bo боле dn. M. К. 1985. Struct C аат and Sena ч араара). (Moscow & Leningrad) 7 947—955 Synopsis of Caucasian E eae. 3 )8. ~ Zhurn. 995 Candollea 50: in Petterson, J. A., E. G. Williams & M. land flo Zealand J Bot. : Phitos, D. 1963a. OMA rupestris- аби Phyton (Horn) 10: 124— 1963b. Eine neue aus si Sw Mitt. Bot. München 5: 121-1 . Beiträge zur Kenntnis p челек then Campana ve п. Ber. Deutsch. Bot. 7: 49-54. 964b. raped: Campanula- ee de Agais. 08-230. Oeste g Ls Z. MI бы on Z. Podlech, D. & J. sei 1964. Zylotaxonomisc the Bei- trage zur Kenntniss der Campanulaceen in Europa. Ber. Deutsch. Bot. Ges. 76: 360—369 (1963, publ. 1964) Quézel, P. 1953. Les Campanulacées d'Afrique du Nord. Feddes ou rt. 56: 1—65. 954. Contribution à = d de L аи аи gu Soc. Sci. Nat. Maroc 34: 297 . 1975. The bases n tolog ier гео Во. Сага. Sc hünland, S, -18 . W. Engelmann, Leipzig. Shulkina, T. V. 1974. Morphology of seedlings in some of the genus Campanula L. in the flora. Bot. Zhurn. (Moscow & Leningrad) 59: representatives S.S.R 439—447. [In Russian.] . Evolution of Campanula L. life forms. Inte rational Boisnical Congress: XII. Leningrad 1975. Abst 235. rem 5b. Peculiarities of seasonal development in [In iris | C — eae. РІ. S I аша : 1: —. 19 I. Dawson. 1995. Contributions to a aa atlas of the New Zea- га—34. У n (Campanulaceae). New 490. ge zur Kenntnis der Campanula 127. в r Û Camp Art der Gattung vi anula the stem in the € Campanulaceae s. ow hari ‘lokularen Campanula-Arten. 2: 449-498 angiosperm D 62: 724—7( 94. Campanulaceae. /n: A. crn rantl ue, Die natiirlichen Pflanzenfamilien some M ies of the genus To rie L. Bot. Zhurn. (Moscow & Leningrad) 60: 109-117. [In Russian.] Types of life forms and their importance for the taxonomy of Campanula L. Bot. Zhurn. (Moscow < Leningrad) 62: 1102-1114 1. [In Russian. | 978. Life forms in Campanulaceae Juss., their ipi al distribution and connection with taxono- . Bot. Zhurn. (Moscow & Leningrad) 63: 153—169. lin Russian. | 1979. De positione systematica Campanula lac- tiflora Bieb. Novosti Sist. Vyssh. Rast. 16: 175-179. [In Russian. . 1980a. Ge 'ographic 'al distribution of наан life unk of the section Campanula of the USSR flora Byull. Moskovsk. Obshe h. Ispyt. Prir. Otd. B 85: 73. 980b. Predstaviteli semeistva C оа еае у E d mira. Probl. Bot. (Novosibirsk), 14: 81— 86 (1979). ine c Ref. Zhurn., Biol. 2: V me . 1980c. The significance of life form characters special reference to the family Evol. 136: 233-246. . 1986a. Zhizne 'nnye formy kolokol’chikovykh, ikh — UM i svyaz's sistematikoi (Leningrad). Botan. In-t AN SSSR 1986. 197 pp. [illus. (Ruko- pis Че >p. у. VINIT ў о = .86., no.5535-V)]. Abstract in Ref. s о . On us ш апсе of Big Pad уя сһагас Е rs in shit systematics. Bot. Zhurn. (Mos 569—579. [In Russian.] . Architectural models in the ( up п беа and possible ways 1. (Moscow & Leningrad) 7 £ for re «АА каше ‘eae s. str. family, transformation Bot. —16. [In = » . 1999. Gross morphology and molec- nulaceae. International ular data in puntos of Campa XVI. St. Louis 1999, Abstract Botanical Congre 16.1.3 E. Zykov. 1980. The anatomical structure of str. in relation to the evolution of life forms. Bot. Zhurn. (Moscow & Lenin- гай) 65: domi [In Russian.] Schultheis, L. 2001. Systematics of Downingia (C ei nd ulaceae) Wir on molecular sequence data: Implica- tions for floral and chromosome evolution. Syst. Bot. 26: 603-621. Shetler, S. G. 1979. Variation and Evolution of the Nearc- tic pies x i ase subsect. Heterophylla), Vols. | and Il. Thesis, University of Michigan, Ann Arbor. é . Morin. 1982. Seed morphology in North American НА eae. International Botanical Con- gress: 'dney 1981. Abstracts, p. 136. Abstracts Australian Ас ademy of Scienc ез ый ——— 19 Seed ds of North Missouri Bot. Gard. : ж go ^^ C ы & American Campanulaceae. Ann. 73: 653-688. Swofford, D. L. 2001. PAUP: Phylogenetic Analysis Using Parsimony (and other Methods), Vers. 4.068. Sinauer, 69. оо Plants: Origin and Dis- persal. E uo Institution Press, Washington, D.C. m J. D., T. J. Gibson, F. Plewnia & n 1997. The Clustal facer HM strategies for multiple sequence align- ment aided by quality analysis tools. Nucl. Acids Res. 882. 24: 1876—18 Volume 90, Number 4 2003 Eddie et al. Phylogeny of Campanulaceae Thulin, M. 1976. е keniensis Thulin sp. nov. and notes on P species. Bot. Not. 128: 350-350. Venkateswarlu, K. & 99]. A conserved core structure in ts rRNA intergenic region from tobacco, Nicotiana rustica. Pl. Molec. Biol. 17: 189— 194 White T. Bruns, S. Lee & J. Taylor. 1990. Ampli- fication and direct sequencing of fungal ribosomal RNA genes for phylogenetics. Pp. 315—322 in M. Innis, D. Gelfand, J. Sninsky & T. White (editors), PCR Proto- Yuan, "YM. P cols: A Guide to Methods and Applications. Academic Press, San Diego. Yokota, Y., T. Kawata, Y. lida, A. Kato & S. Tanifuji. 1989. N E r sequences of the 5.85 rRNA gene and жш, transcribed рае ег Regions in carrot and "e bean ribosomal DNA. J. Molec. Ev z 29: 294— qu & J. J. Doyle. 996. евз РЯ phylogeny of the pue T bci inferred. from nucleotide sequences of the internal transcribed ut - ers Uo of nuclear B oma! DNA. Amer. J. Bot 641-6 Missouri Botanical Garden Annals of the 572 PEPLEEAV "IcOccC AV GEPLEEAV "OcOcct AV CE£FICEAV '6I0CCE AV LEPLEEAV “BLOCCEAV ОЄЎ1ЄЄАУ “LLOGCEAV 6CPIEEAV “OLOZZEAV BCPLEEAV “SLOCCEAV LOVIEEAV "TIOGCEAV O9cV Itt AV “ELOZZEAV CEVIEEAV "CIOCCEAV PEVLEEAV “LLOGCEAV ESPLEEAV “OLOZZEAV GCVLEEAV “600ECEAV [ct LEEAV "800cc€ AV OZPIEEAV “LOOZZEAV СРЕВІЯУ “LEPESTAV OLO60AV “COLO60AV 120604V `с120604У OLO60AV “B0L060AV OIPLEEAV '900cct AV 020604V “POLO60AV O0L060AV *0020604YV I20604V “PILO60AV 120604V '8120604V 010604V '9020604V 120604V "91206041V 120604V “OLLO60AV BIPIEEAV “SOOZZEAV (OW *RIBIOO*) “snseone’)) F6lc ONW W (XAL 'e2nuog) ZZZ әләм `S (H94) 20076 APPA “ZOOK6 DON (OW "PIS1024) "snseone-)) F980c 011242] W (H954) 91086 ?!Pp3 (XAL AMI) 910°S6 opp (XAL “PIqery Ipneg) 29/9 әпәиәогу `Ç (XAL LUBNA VSN) OST 2742qDH 7) M (4 ‘stout VSN) 9egg saourumT (191, *р18109‹) “snsbone’)) "us U1YSDD (XAL "ATP 002 ISE pioqujoay f (4H94) £0076 DON (OIN L13109) *snseone")) “u's DUIYINYY 4 ссс91 MOU (CHO) FISFOF8FSF APPA (6661) 1F 19 wry (2661) T? 19 22) (2661) “TR 19 99 (1661) 1F 19 22) (uedef) 2600661 Я9ЯН €00€286l HOU (LN) soz Urysey (LA) LZZ SPAeN (LA) с00ъ6 ?!pp3 (LN *OW) t80c шҷѕве) (LN) 91086 ?!pp3 (LN) 910€6 2!PPA (LN) 002" ISG Z ә1рря (LA) €00b6 ?!pp4 GL ty €or :UIXSB4) (LA) ws 191807) (LA) ҰТЗРОРВРСР этррч GLO) 28096 этрря (LN) 99096 әїРРЯ Хәр ү) ә “Yost y 11232Du2t[ou отирішт? pur] у ‘ѕиешуонң 111419 npmupdurm-) Q2t31p[2H y "uxssnepn рит1ѕғитутру рјпирашр? azpRleyy 17_ш12у015 рупирішр» "| 0702әш0]9 Djnupduip?) “1 $nuiu2 ррпирішт”) ‘yss10 4] szmpa pgmupdur?) XY D2D214Dà1p ngmundurm) ‘boef рәттә pnmunpdur) sumpy ют]оЙйр1]әф nmupdumr) "| 2209109 nQmupdur?) ‘Be родо omundun) ozpe1eu?) розши pqnundui-) PIA 2770/11401]. omundun) 192] CSM) пүорта Duozy ‘bug (DIN) unomuiodpbf punautsy 1491927 рирәітрт pn4oudouapy “bi, vns n401douapy вмезеџу ("әрә") ритуғириәјѕ Dsioydouapy Бу (9907 xy “qaig) naojijoui24 Dsoydouapy учив орорпотиой psoydouapy ejeAep Sisuauosisow тіоціоиәрү BUOH “A ‘q 0744040] vsoydouapy 1994 nup&npppum, Dsioydouapy “ABS XN “YOUR njponiparp n40udouapy "PS X ‘Yue J] DIDIUDAIp ргоціоиәрү әвәде[пивіше”) (ZSLI 'ISLD J9quinu ПО15$5Ә22Р yueques) (@[qPJIPAE пәм UISLIO jo &nunoo) saouanbos pausiqnd 10} әәиәләЈәг 10 *Ioquinu uotssa2oe uapaes [peorugjoq “иопешлоуш ләцәпод (Kıoqısoda1) pue Joqul nu uoIss32)2P V Nd A]uoujne pue әшеи uoxe поч ў 'suomneor[qnd рәџә ey) ur punoj aq ueo saouanbas pousr[qnd ло} [ешәјеш jo sao1nog -ur&uo1oe umueqrog pue чәдшпи иопәә[[оә *10]29|[09 Aq рәмооу st suauroads ләцәпол цим [euoje]y "suapge«) [porugjog 10} suotrAa1qqe әці ^o[[oj srequinu uorssa2oy (JHIA) Ysinquipy jo Хивләлтип] рив (X41) unsny ‘sexag, jo Хивләлти[] *19]u27) вәәлповә 1ue[q (J) unsny ‘sexag, Jo AISI3A1U (DN) aspuquie’) *uonosqo?) ngmupdum:) euoney (Ig L) isi (OW) uapıes [eoruejog Linossip (Яо) Mey 'suspre?) oruejog [e4oy (Чоңу) ysinquipy suapıes) oruejog [роу әле suorjeraa1qqe [еиоцпуѕиј *seaoer[oqoT pue aeaoe[nuedure?) ay) urpim saroods pue еләиәЗ Aq 4q[eorjeqeud[e pasig ‘ұма 1u Jo uorzar с] ey] 10} рәәиәпһәѕ exej, `I xipuadd y 573 Eddie et al. Volume 90, Number 4 2003 Phylogeny of Campanulaceae 198p€ LJV (6661) "Ie 19 n4 Ацо UaysduD] sisdouopo:) O98PE LIV (6661) ‘TE 19 n4 juuew оупѕопа sisdouopor) LIC800HV (6661) 18 19 n4 ddiu7) рѕора sisdouopor) LESIETAV (6661) Te 19 n4 чие nsoa4au sisdouopo-) OS8PElLAV (6661) ЛЕ 19 n4 чие ру$әрош sisdouopo-) LOPIEEAV “BPOZZEAV 0€602861 OAH (LA) €c0€6 ә!рря учоон 29 `чїшәң (оопу X дәт) »mjoa2un] sisdouopo;) OOPIECAV “LPOZZEAV (4 ‘ивмте) 6679 8әшштт (LA) ‘u's 190807) emen 1una, sisdouopor) OSPLIEEAV “OPOZZEAV (jedan) ZSgoZ661 ADU (LA) 20866 ә!їрря "WS M MM Prpofiamuaotp sisdouopo;) 8СР1ЄЄ АУ 'STOCCEAV (вривүвү A1eue) шед) сє0//61 99H (LA) 8096 ә!їрря aye, (7]) ғєиәырирә DutDuvy) c98T€ LIV (6661) Te 19 n4 eum[g pornapf naoumundur) LOPLEEAV “PROZZEAV (XAL) 05096 ?!PP3 “DON (LN) 06096 әїрря пеш (7) umunou2um umajspjnupdum;) OSPIEEAV ‘ЄРОССЄДУ (OW ‘2181095 *snseone?)) 7 pp wysog f (LAON) LTH чуве "qergos vmruəpı рүпиріш) SSPIEEAV “ChOTZEAV (H94) ws 21pp3 “ON (LO) u's e:pp3 7] saprosiky) omundun) PSPIEEAV “LPOZTEAV (Ig,L ‘881095 ‘snseoneg) gc] unjspz) "f (LA) ZOE шв) ‘qrg "JN 2иәләјѕ рупир4шог) ESPIEEAV “OPOZZEAV (IL ‘2181095 ‘snseoneg) ghp u14sD9 f (LN) ple ugse9 azpeaeu?) 144smousos pundum) CSTI£€AV “6LOTZEAV (IAL ‘snspone’)) “u's puiymuys 1, (L n) Z9p use рә npunuzidais тираш?» ISPIEEAV “BEOSTEAV (194, ‘8131095 *snseone)) "ws ритуртус̧ iI (LA) gey чүс) megy DIDUDS pjupdui) OSTLEEAV 6tVICC AV BPVIEEAV LVVICEAV 9VV I€£ AV SVVICCAV PPVLEEAV CVPLESAV "LE0CCE AV '9t0CcE€ AV '"StO0CCE AV "VEOGCEAV "££OCCEAV `бЄОССЕ AV “LEOSZEAV 'OEOZZEAV CVVICEAV ` 6c0cc€ AV (4 VSN) p128 si2unur] (XAL sexa], *"VS0) £0000 PPA (191, “RIBIOI) ‘snseone’)) "urs риту |] (CHO) 68096 APPA “DION (H94) 26096 PPA ‘76096 ЭЭМ (XAL ‘[ean110q) 666 SAIN) є (29u£14) €ZZO986L ADU (HHO) 22096 APPA 'CL€6961 ADA (XAL Хә) 20066 PP (LA) ‘u's 190807 (LA) t0000 ?t!ppal (LA) 4€ UIYSES (LO) 68096 ?!ppa ил) 66096 PPA “| pyofipunjos отирішр») ABI‘) "V TuOt242d24 Djnupdutr) "Ajnei[ pupapppa ppgupdurm-) “| sipppiuma(d ppupdurmr) "weg pypjound pjmupdurr) "| pyofynuiud opmupbdunm:) "| papijad pjnupduir) “| pyofinssad njmupdurm?) “1 puidasad pjnupdupy [PPIEEAV ‘BZOSZEAV (Ig ‘snseone)) gc punmmus iL (LN) 89p uses) "qarg D211ass0 njnupdum;) OPPIEEAV “LZOZZEAV (XAL ‘ureds) ggz әләм `S (LA) O€% $9^9N "1 =ош vjmuvdurm;) 6EPIEEAV “OZOZZEAV CV0cL661 ADAH (LA) 96096 әтрря ^oq[y =/19олш Dynuvdupy 9€PICCAV “SZOZZEAV (XAL “TP3nu0g) 9ZZ әләм "S (LA) 9@@ әләм "рәо”] »oruppsn] vjupdur;) LEPIELEAV "PCOCCEAV J9qonoA ou *'srourure'] (Ln) ‘u's 191807) S оор] ојпиріш»? OEPLEEAV ‘EZOZZEAV (AHO) [S096 APPA ‘76096 ЭЭМ (LA) 1€096 ?!ppa ‘ALLY DIDUD] DjnuvdwuD?) CEPIEEAV ‘ZZOZZEAV (Ig ‘snseone)) gy vunmus iL (LN) 99p шв әрі гү `7) риоџоиәоұ DjnuDduD?) (ZSLI ‘ISLI (әүҷереле иәцм шо Jo (А:оџѕодәл) рив Joquinu UOISSIIIR yueques) Аципоә) saouanbos pousi qnd 10] әәиәләЈәл 10 *1oquinu uorssaooe uapaes [eoruejoq ‘иоцешлоуит I99noA Joquinu uorssa2oe YNA Aquoqne pue 2ureu uoxej 'panuguo) [| xipuadd y Annals of the 574 Missouri Botanical Garden COT [t€ AV cot L€€ AV lor I€EC AV OOt [CE AV 68V LEE AV 88V L£C AV L8VICCAV 98T [££ AV S8VTI€£ AV TSVICCAV CSV ICCAV CST I€££ AV [IST TEE AV O08T IEE AV OLVIEEAV BLVIEEAV LLVIEEAV OLVIEEAV CLVLEEAV VZVICEAV ELVLEEAV GLVICEAV [2E ICE AV OZ TICEAV 69T [££ AV 989v 1€£ AV LOVIEEAV 99v | ££ AV SOT [££ AV VOY ICE AV COT [£C AV COVIEEAV '080cct AV '620cc€£ AV "820cct AV '0L0cct AV '690cct AV "£90cct AV "C90cct AV "I90cc€ AV "090cct AV “OSOGTE AV `ӨСОССЕ AV “LEOGCE AV '9C0cc€ AV '£C0cc€ AV “OSOZZEAV “OPOETE AV (XAL "puepoos) 470086 APPA (XAL 's*xoL VSN) ZEL ?1490»H 7D) M (HOM L80086 APPA (141, ^"5nseone?) ‘u's ритуртус̧ сү €080S261 AOA (НЯ) u = sza&y у (4H94) 92096 APPI (uredS) 67078261 ADA (4) £666 %әшшт (XAL 922212) 99096 appa (XAL ‘JHIA 0£0€6 APPA "ws OAH (edəN) 188126861 498 (XAL '‘JH9A) ££0€6 ?1Pp3 (XAL ‘JH9A) 21026 ?!pp3 (9Y VdS) £786 әЎрәң » SAPS (408 "ureds) 96°96 әЎрәң X SAPS (409 "ureds) 67796 әЎрәң 29 SAPS (H94) €£0€6 Appa (XAL AHI) £00€6 Ppa (XAL POJE VSA) or] 4490H 7) M (e210N yMoS) O8EZ 861 ЯОЯН JIYINOA ou "Uu LIO [A TIZ€696l ЯЭЯЧ (XAL 9220194) 6ZFZI 1” 12 Anf y 5 6L IOT66I 4OgMH 1£609961 Я9ЯН (HHO3I) €F0€6 ?1pp3 (SO) 621 42uso7) (HH93 'ouiose p grr `q 2218028 X Іл (LN) MPO0086 PPA (LN) ZEL ?p9qeH CL) 180086 ?!pp3 (La)ecc uses (LN) V£690€7 ?!pp4 (LA) 8SZ092 ?tpp4 (LN) “u's 190607) (LN) 92096 ?!pp3 (LA) 06096 э!рря (LN) ‘u's 191807) (LA) 900€6 21PPA (LN) 99096 21pPPA (LA) O€0S6 21PPA (LN) "ws әтрру (LN) 1066 э!рря (LO) #066 ?!pp4 (LN) 21026 ә!їРРЯ (LO) 6b әїрря (LN) €€0€6 этрря (LN) €90€6 э!рря (LN) 6FL ?poqeH (LA) 810€6 PPA (L0) “u's 12u807) (LN) 600€6 ?!Pp4 (LN) 410086 ?!Pp4 (LN) 6TTOt6 ?1PPA (LN) 6c0€6 ?'pp4 (LA) €*0€6 ?!pp4 (LA) "is 190607) (LN) ВРО әтрря “| рәәрләрәу miz1aquapnnag, "пә (HNN) pdap2ojda] siuppou "ү шпәјтләрә umiauopa эа ¥ (9214) оуприәа ріриоХуашАс̧ queg luupufoy Dipupdyduitg Эа 'V (4316) Duawsd рірирАцішл$ “ү 0mipo орәоу Эа су Cboe[) snaoyipupag UOPOIAID] "| pipoids puna Aud “1 240]m21040 vumaÂyd NY (7]) vsowos sixajdostyd "gv ("D pypuuid DINADWOL ad зошт] pənd D1yassn Vr) » “Yost y 1102490447 DMIXNDYN W шә] $7772042 uopooojd»] yosiy (71) su2uoa-umgnoods pisnodv] yos (чәр) отор Disnodo] "nay x cssiog | Daojfipissos 2uoispf "| Dupjuow auoisp[ Ou113 N хә шоу "IN И! (&qn(]) оштри auoisp[ jung WIR] sz12D] 2uoispf "dures (umnog) pdsiu) 2uoispf ‘HNN unaojfupa uopo20491a]] EJEN 022010 DXDsnqpung Aeı<) "y pnsnffip sisdoyny eurx[nuos ("әт N) P1011120] Diyjappy Jasng (*qsnouos) pyofiisngup nəə] Эа ^V (seipiqos) ound snupimaps3 Эа ү СЛ) smjofrimupad. snyiiÍapg шиоя (dwe) pupijauma p42ridsoiT "pguog хә "Прем sn7DQ0] snyjuDUDAD "^g Y PENH D2522u02 nsdp2042]047) (ZS. "ISLD I9quinu UOISSIIIR yuequs5 (9|qejre^e иәцм ursuo Jo &gunoo) soouanbos pousijqnd 10J 32U219]21 10 *I9quinu uorssa2oe uapıe3 [eoruejoq ‘иопешлојуи ләцәпод (4лоџѕодәл) pue їәдшпи ПО158Ә2ОР ү Nd Ayuoyyne pue эшви UOXEJ, ‘panuquo) “|, xtpueddy 575 Eddie et al. Volume 90, Number 4 2003 Phylogeny of Campanulaceae SCOYSOAV OEVEOLAV SEVEOLAV 00697 LAV (8661) moq 1002) sro i[nuos (1002) st2ui[nqos (1007) st2ui[nqos —. ujunw рләиә nijoqo] “| pdn} pyeqo] sau “YD р ә Sey You pounbpioqo 0172401] Ioj zidnjpoziopnq pizuiumo(] IBIIBT[IGO"] (ZSLIIS.LD I9quinu UOISSIIIE yuequəs) (9]qe[reAe uam uso уо Аципоә) saouanbas pausi[qnd 10J әәпәлә}әл 10 *I9quinu UOISS900R чәртез [eoruejoq “иоцешлоуиг ләцәпод (&io]sodo4) pur Хулощпе pue әшри uoxep I9qui nu UOISSIDJIE VNA 'penunuo-) “| xipusddy MORPHOLOGICAL STUDIES TOWARD AN IMPROVED CLASSIFICATION OF CAMPANULACEAE S. STR.! Tatyana V. Shulkina,? John F. Gaskin,?? and W. M. M. Eddie? ABSTRACT Growth and seedling morphology of 144 spec ies representing 30 genera of Campanulaceae s. str. were studied. Two types of seedlings were found: Group A, with vegetative characters. Thus, plant rhythmic seasonal growth ас npn é leaf arrangement, continuous growth (at leas Taxa in Group A are distributed И іп Asi s from Group have an opposite lea long i dep period, and sympodial branching. Plants hun (ыны B have a an elongated epicotyl and elongated ی‎ x Group B, with a shortened (not visible) epic otyl and usually shortened internodes. These Thu rrelated with other cdi in ontogenesis), a spiral iode types appear to zd f arrangement (at lea in the non-flowering period), and sympodial and monopodial brane hing, a, whereas representatives in Group B ide. occur almost worldw The groups do not coincide with current taxonomic авес ‘ations but correspond remarkably well with the distribution ‘of other characters such as pollen- grain арен and correlate with groups init on molecular analysi refore, these two groups may reflect two ineages. and seedling morphology a e of taxonomic significance in a diss ulaceae and can be used for treatments in кшк кеп with other characters. чайдагы changes, which are supported by molecular data, are proposed. Key words: Campanulaceae, growth and seedling morphology and development, taxonomy. Campanulaceae herein are treated in a narrow circumscription (without Lobeliaceae) as a mono- phyletic group with a distinct geographical distri- bution and with well-defined morphological char- acters. Campanulaceae s. str., despite their size and importance in temperate floras, remain unrevised. This family, with about 50 genera and 800 species distributed worldwide, is the largest and most prim- itive and basal one within the order Campanulales (Lammers, 1992; Takhtajan, 1997). Although rep- resentatives of the family occur on all continents except Antarctica, the vast majority of genera and species are found in temperate regions of the Olc World. Raven and Axelrod (1974) considered the family to have a Laurasian—African origin. The cen- ters of distribution and diversity include the Med- пв East Asia, and South Africa (Shulkina, Kolakovsky, 1995; Hong. 1995; Eddie, De Candolle's (1830) comprehensive monograph on the Campanulaceae provided a solid basis for all subsequent works. He divided the family into two tribes and later added a third tribe to accom- 1839). Schónland (1 also divided the tribe Campanuloideae modate Merciera (De Candolle, (Campanulaceae s. str. here) into three groups, based on mode of dehiscence and ovary position, but these three groups differed in composition from those of De Candolle. These two classifications be- came the basis for all future treatments (Table 1). Although the current systems differ greatly from the old ones in number of genera, as many taxa have been added during the last century, it is easy to understand what classification each particular au- thor is following. Schénland’s treatment has been used often and remains a currently useful refer- ence. Fedorov (1957), on the contrary, followed in gen- eral De Candolle’s position and published a de- tailed classification for Campanulaceae growing in the former Soviet Union (FSU). Fedorov proposed 8 tribes (6 new) based on capsule dehiscence, co- rolla shape, and presence and shape of appendages between the calyx lobes. Kolakovsky (1995) pro- posed a new system with 4 subfamilies and 22 internal fruit structure. He tribes based on ' We are indebted to Ihsan Al-Shehbaz, Peter Stevens, and Peter Hoch for reading the text and providing useful remarks. Many thanks to Victoria Hollowell for careful editing. T. Shulkina thanks the people who helped her obtain | Campanulac ji seeds from р н of the world: R. Kamelin, A. Kolakovsky, N. Morin, H. B. R ? Missouri fun al Garden, Po. m 299, St. * Present address: USDA, ARS-NPARL, ^ Section of Integrative Biology, University of Tex - M. Brenan, A. Cronquist, A. Dolukhanov, R. Gagnidze, {ycroft, and P. Wendelbo ouis, Missouri 63166-( P.O. Box 463, Sidney, isa 59270, as at Austin, ЕТЕ ae hulkina@ iobot.org. Texas 78712, U.S.A. ae address: Office of Lifelong k. Learning, University of Edinburgh, 11 Buccleuch Place, Edinburgh, Scotland, U.K. weddiel @staffmail.ed.ac.u ANN. Missouni Bor. GARD. 90: 976—591. 2003. Volume 90, Number 4 2003 Shulkin UT Studies of Campanulaceae 577 Table 1. Treatment of genera in Campanulaceae. De Candolle (1830, 1839) Schénland (1897) Fam. Campanulaceae Subfam. Campanuloideae Tribe Campanuleae Subtribe Campanulinae denophora Adenophora Campanula Campanula Musschia Heterocodon Michauxia ichauxia Petromarula Ostrowskia Phyteuma Peracar Symphyandra Phyteuma Trachelium Legousia Symphyandra Trachelium Tribe Wahlenbergieae Subtribe Wahlenberginae noea Campanur Campanumoea Canarinc P eue da Cephalostigma Codon Codonopsis Cy Mord Edraianthus Edraianthus asione Heterochaenia Lightfootia Jasione Microcod Leptocodon Platycodon Merciera Prismatocarpus Prismatocarpus Roella Rhigiophyllum Wahlenbergia Roella Wahlenbergia Tribe Merciereae Subtribe Platycodinae Merciera Microcodon Musschia Platycodon described 9 new genera within Campanula, which have not yet been included in the Vascular Plants of Russia and Adjacent Countries (Czerepanoy, 1995) due to their contradictory descriptions. Takh- tajan (1997) divided the family into 4 subfamilies and 16 tribes, taking into consideration not only the fruit structure but also pollen-grain structure. ovary position, as well as the presence or absence of appendages between the calyx lobes. Subfamily Cyananthoideae includes the genera Cyananthus, Codonopsis, Campanumoea, Leptocodon, and Pla- tycodon; subfamilies Ostrowskioideae and Canari- noideae are monotypic. The last subfamily, Cam- panuloideae, consists of 12 tribes and includes all y Kolakovsky were not included in the system. Hong (1995) ten- remaining genera. Genera described tatively divided the genera into 6 unnamed groups based primarily on various morphological charac- ters. Eddie (1997) divided the family into two major tribes, with the differences between them consid- ered to not warrant subfamilial status. Eddie's Pla- tycodoneae subdivided into the following subtribes: Echinocodinae, Codonopsinae, Platycodinae, Campanumoeinae, Ostrowskiinae, Cyananthinae, and Canarininae. His Campanuleae comprised the following: Wahlenberginae, Jasioneinae, Musschi- inae, Azorininae, and Campanulinae. There is considerable disagreement among all prior classifications of Campanulaceae. Further- more, there is no common opinion about generic limits or higher relationships among the major sub- divisions of the family. Taxonomic problems in this family can be explained by the fact that nearly all of these earlier classifications had a geographical rather than biological basis. Thus, floristic treat- ments differ considerably in the generic delimita- tion of the Campanulaceae for the former U.S.S.R. (Fedorov, 1957), Europe (Fedorov & Kovanda, 1976; Tutin, 1976), Turkey (Damboldt, 1976), and China (Hong, 1983). Genera crossing diverse geographical regions need multidisciplinary study, including research on the development of vegetative organs, morphology and anatomy of fruits and seeds, pollen grain struc- ture, as well as molecular and serological data. As stated by Takhtajan (1997: 6), phyletic relationships and construct phyletic line- “We cannot establish ages using only floral characters. It is all the more impossible to reconstruct phyletic lineages on the basis of the characters of the vegetative organs only.” The greater the number of characters from different correlation groups taken into consider- ation, the closer we can approximate the evolution- ary phylogeny of the family. Vegetative characters in higher plants are ac- corded only a limited place in classification, de- spite the angiosperms being first divided into two great subclasses according to the number of coty- ledons as early as the 13th century by Albertus Magnus. Publications in which the importance of vegetative characters is supported are nol numer- ous (e.g.. Stebbins, 1974; Tomlinson, 1984). How- ever, life forms and growth patterns, ultimately in- fluencing the structure of the mature plant, are often ignored or little emphasized because of the common opinion that all these characters are adap- tive. However, life forms include many distinctive vegelative characters that can be taxonomically valuable if they are stable. The goals of this study are (1) to study vegetative organs and development of life forms in represen- tative species across Campanulaceae; (2) to select characters that are common to species groups that may have taxonomic value; (3) to compare the groups suggested by these characters with formal 578 Annals of the Missouri Botanical Garden classifications for congruence with other morpho- logical and molecular data We examined seedling morphology, growth pat- terns, leaf arrangement, seasonal development and behavior, and branching patterns before and after first flowering in studied plants. Special attention was paid to genera whose placement varies in cur- rent systems: Azorina, Campanulastrum, Canarina, Edraianthus, Musschia, Ostrowskia, and Platyco- don. Also included were representatives of recently segregated genera: Annaea (= Campanula), Gad- ellia (= Campanula), Hemisphaera (= Campanula, subsect. Scapiflorae), Neocodon (= Campanula, sect. Rapunculus), and Theodorovia (= Campanu- la) MATERIAL AND METHODS Plants of 144 species in 30 genera were exam- ined (Table 2). The studied genera represent taxa from 2 tribes of De Candolle (1830), 3 subtribes of Schónland (1889), 8 subtribes of Fedorov (1957), 17 tribes of Kolakovsky (1995), and 14 tribes of Takhtajan (1997), and they provide a representative sample of the Campanulaceae. Almost half of the studied taxa were formed by the species of Cam- panula (65) and other genera (13) of the flora of the FSU. All new genera described by Kolakovsky were split from Campanula as well. As the most complete classification for this group was made by Fedorov (1957), the list of studied species was mainly arranged according to the system published in the Flora of the U.S.S.R. All plants were grown at the Komarov Botanical Institute (St. Petersburg, Russia) and a few (Azorina vidalii, Campanulas- trum americanum, Campanula kemulariae, C. tata, Canarina canariensis) also at the Missou- Missouri, U.S. The taxonomic н of all plants was confirmed i Botanical Garden (St. Louis, when flowering. Vouchers are partly deposited in the general herbarium at the Komarov Botanical Institute (LE) (e.g., Canarina canariensis, Shulkina, 1978; with no numbers as is typical in Russian herbaria), and partly in the herbarium at the De- partment of Living Plants collections at the Ko- marov Botanical Institute. "ants were grown outdoors or in greenhouses, depending on the plant's requirements. Seeds were collected in nature throughout the former Soviet Union and the midwestern United States by the se- nior author. They were also obtained from other col- lectors undertaking field trips on the islands of Ma- caronesia, or in the Middle East or South Africa, as well as from different botanical gardens. Seeds were sown in the greenhouses at the Komarov Bo- tanical Institute in early spring (March) during 1973-1990. Observations were made every other day during germination and early stages of seedling growth and once a week for maturing plants. Sam- ple size per collection (species investigated) was 20 to 50 plants whenever possible, but in some cases fewer seedlings were available. The period of ger- mination and cotyledon size and shape were noted, and the first leaves were examined. Seedlings were illustrated when they had first leaves and fully de- veloped cotyledons, approximately one, rarely two, months after first appearance. At this time seed- lings were transplanted into larger pots. Most plants were planted in summer in an ex- perimental plot in the open air, but some (Azorina, Canarina, Diosphaera, Musschia, Roella) were kept in greenhouses. Some portions of outside plants were brought back into greenhouses in late autumn to study the presence and length of their dormant period. Leaf arrangement and branching patterns were examined throughout the year as were the presence of green leaves or renewal buds during the winter months. The timing and position of new growth were recorded in early spring. Plants were dug out, and the development of their underground organs was checked in the first year and while flow- ering. Life forms of some species (above- and un- derground organs) were also studied in nature by the senior author in the Caucasus, southern Siberia, Central Asia, the Russian Far East, the Carpathi- ans, the Mediterranean, and the midwestern United States. RESULTS AND DISCUSSION GROSS MORPHOLOGY The Campanulaceae include plants with varied life forms. As shown in previous studies (Shulkina, 1978) most species, genera, are perennial herbs, and these are found including members of about 30 throughout the range of the family. Annuals, mainly in the Mediterranean region and the New World, rarely in East Africa and Australia, also very rarely in East Asia, are present in 11 genera. Some Af- rican annuals are relatively long-lived plants (e.g., 10-12 months), whereas the Mediterranean annuals are usually Wahlenbergia undulata lives short-lived (e.g., Brachycodonia fastigiata, 1-2.5 months). Thirteen. genera consist completely or partly of arborescent and semi-arborescent plants. These dwarf trees and shrubs occur in the Azores, Madeira, the Mascarenes, Reunion, and South Af- Musschia, Bere- Three genera include her- rica (e.g., Azorina, Heterochaenia, nice, Prismatocarpus). Volume 90, Number 4 Shulkina et al. 579 2003 Morphological Studies of Campanulaceae Table 2. List and location of taxa studied, number of species used/general number of species in each genus. АП vouchers are at the Komarov Botanical Institute, St. Petersburg, Russia (LE). Type species are in bold. Adenophora 8/40, Eurasia A. stenanthina (Ledeb.) Kitag. Perennial Altay A. kurilensis Nakai Perennial Korea* ^ lilüfolia (L.) A. DC. Perennial E Europe nikoensis Franch. & Sav. Perennial Korea* pereskiifolia (Fisch. ex Roem. & Schult.) G. Don Perennial Siberia A. tetraphylla (Thunb.) Fisch. Perennial Sakhalin A. trachelioides Maxim. Perennial Far East A. triphylla (Thunb.) A. DC. Perennial China* Asyneuma 4/50, Disjunct, Europe & E Asia A. japonicum (Miq.) nd Perennial Far East A. otites (Boiss.) Born Biennial France* A. pulchellum (Fisch. à Mey.) Bornm. Biennial E Caucasus A. salignum (Waldst. & Kit. ex Besser) Fed. Perennial Е Caucasus Azorina 1/1, Azores A. vidalii (Wats.) Feer Dwarf tree Portugal * Brachycodonia 1/1, Mediterranean, E Caucasus, C Asia B. fastigiata (Dufour ex A. DC.) Fed. Annual C Asia Campanula L. 70/300, Northern Hemisphere seclion Campanula subsection Quinqueloculares Boiss. ;. medium L. Biennial France* C. crispa Lam. Biennial Caucasus subsection Spinulosae (Fomin) Fed. C. mirabilis Albov Perennial W Caucasus subsection Triloculares Boiss. C. sibirica L. Perennial Siberia C. caucasica Bieb. Perennial E Caucasus C. hohenackeri Fisch. & C. A. Mey. Perennial Caucasus C. komarovii Maleev Perennial ' Caucasus C. longistyla Fomin Perennial W Caucasus subsection Phasidianthe Fed. C. imeretina Rupr. Perennial W Caucasus subsection Tulipella Fed. C. punctata Lam. Perennial Far East subsection Dasystigma Fed. C. alpi с Perennial Carpathians subsection Annuae (Boiss.) Fed. — Roucella Dumort. C. erinus L. Annual France* C. propinqua Fisch. & C. A. Mey. Annual Armenia subsection Campanula atifolia L. Perennial Caucasus C. bononiensis L. Perennial N Caucasus C. ies Hi C. Koch Perennial W Caucasus C. m АДЫ, & Kuth. Perennial W Caucasus С с. жө Boiss Perennial Е Caucasus C. rapunculoides 1. rennial Е Europe C. trachelium L. Perennial E Europe subsection /nvolucratae (Fomin) Fed. C. glomerata L Perennial E Europe 580 Annals of the Missouri Botanical Garden Table 2. Continued. C. cephalotes Nakai C. oblongifolia (C. Koc ms Charadze C. trautvetteri Grossh. & subsection Cordifolia (Fomin) Fed. C. alliarüfolia Willd. C. dolomitica E. Busch C. makaschvilii E. Busch subsection n Latilimbus Fed. i na Bieb. 94 E Kola c irinae Kuth. C. sarmatica Ker-Gawl. C. sommieri Charadze subsection Trigonophyllon Fed. C. dzychrica Kolak C. autraniana Albov Fomin) Fed. ~ subsection Symphyandriformes C. kolenatiana C. ^. Mey. ex Rupr. ;. bayerniana Кирг. ;. choziatowskyi Fomin _ kemulariae Fomin C. ossetica Bieb. C. raddeana Trautv. subsection Oreocodon Fed. C. incanescens Boiss. C. kachetica Kantsch. C. kantschavelii Zagareli subsection Scapiflorae (Boiss.) Fed. = Hemisphaera Kolak C. anomala Fomin C. aucheri A. DC bellidifolia Adai ;. biebersteiniana Roem. & Schult. ;. chamissonis Fed. c ciliata Steven C. saxifraga Bieb. C. tridentata Schreb. subsection Rupestris (Boiss.) Fed. C. karakuschensis Grossh. = Theodorovia Kolak. C. lehmanniana Bunge = Hyssaria Kolak. subsection Hypopolion Fed. C. hypopolia Trautv. subsect. Heterophylla (Nym.) Fed. C. rotundifolia 1.. C. polymorpha Witasek section Rapunculus (Fourr.) Boiss. subsection бесик н Fed. — Neocodon Kolak. ik = = Neoc odon Kolak. C. alberti Trautv. = Neocodon Kolak. C. altaica Ledeb. = Neocodon Kolak. C. beauverdiana Fomin = Neocodon Kolak. C. hemchinica €. Koch = Neocodon Kolak. Cor rapun C. i anum pedes & Scher Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Pere nnial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial China* S Caucasus S Caucasus Caucasus Caucasus W Caucasus Caucasus W Caucasus W Caucasus C Caucasus N Caucasus W Caucasus W Caucasus 'Tbilisi* S Caucasus S Caucasus W Caucasus 'Tbilisi* Caucasus > Asia Caucasus Caucasus N Caucasus Caucasus Caucasus Caucasus Far East E Caucasus N Caucasus Caucasus S Caucasus C As a Caucasus E Europe Carpathians N Caucasus Carpathians C Asia S Siberia S Caucasus W Caucasus Volume 90, Number 4 03 Shulkina et al. Morphological Studies of Campanulaceae 581 Table 2. Continued. C. hieracioides Kolak. = Annaea Kolak. C. lambertiana А. DC. = Neocodon Kolak. lak C. patula L. = Neocodon Kolak. C. persicifolia L. = Neocodon Kolak. C. po ontica Albov = Neocodon Kolak C. stevenii Bieb. = Neocodon Kolak. C. turzcaninovii Fed. subsection Rotula Fed. C. carpatica Jacq. subsection Melanocalyx Fed. C. uniflora L. subsection Odontocalyx Fed. C. lasiocarpa Cham. Campanulastrum 1/1, North America C. americanum (L.) Small Canarina 2/3, Canary Islands, disjunct E Africa C. canariensis (L.) Vatke C. eminii Aschers. Codonopsis 6/30, E and C Asia C. clematidea (Schenk) C. B. Clarke C. pilosula (Franch.) Nann C. ussuriensis (Rupr. & SD Hemsl. C. vincifolia Kom. Cyananthus 4/23, E. Asia C. lobatus Wall. ex Benth. C. inflatus Hook.f. & Thomson C. integer Wall. ex Benth. C. microphyllus Edgew. Diosphaera 1/3, Middle East D. hysterantha Rech.f. & Schiman-Czeika Edraianthus 4/24, E. Mediterranean E. rni ur (L.) A. DC. E. horvatii Lakusic E. pumilio (Portenschlag) A. DC. E. sutjeskae Lakusic E. tenuifolius (Waldst. & Kit.) A. DC. Gadellia 1/1, Caucasus С. lactiflora (Boiss.) Schulkina Githopsis 3/4, W North America G. calycina Benth. G. diffusa A. Gray G. pulchella Vatke G. specularioides Nutt. Jasione 3/20, Europe, N Africa J. heldreichii Boiss. & Orph. J. laevis Lam J. montana L. Legousia 3/20, Europe, N Africa, Americas L. falcata (Ten.) Fritsch Perennial W Caucasus Perennial N Caucasus Perennial E Europe Perennial E Europe Perennial W Caucasus Perennial S Caucasus Perennial S Siberia Perennial Carpathians Perennial N Siberia Perennial Far East Biennial MO, U.S.A. Perennial Spain* Perennial France* Perennial C Asia Perennial China* Perennial England* Perennial Far East теп Far East Perennial Japan Perennial Great Britain* Perennial Great Britain* Perennial Austria* rennial Great Britain* Perennial Spain* Perennial Italy* 'rennial Yugoslavia* Perennial Yugoslavia* Perennial France* Perennial France* Perennial Caucasus Annual W North America Annual W North America Annual W North America Annual W North America Biennial France* Ann., bien. France* Biennial Switzerland* Annual Spain* 582 Annals Жөн ч а Garden Table 2. Continued. L. hybrida (L.) Delarbe Annual Greece* L. pentagonia (L.) Druce Annual France* Leptocodon 1/2, Е Asia L. gracilis (Hook.f.) Lem. Perennial Great. Britain* Michauxia 1/7, Е Mediterranean M. laevigata Vent. Perennial Caucasus Musschia 2/2, Madeira Islands M. aurea (L.) Dum. Shrublet Great Britain* M. wollastonii Lowe Dwarf tree Great Britain* Ostrowskia 1/1, C Asia, Afghanistan O. magnifica Regel Perennial Central Asia Peracarpa 1/1, E. Asia P. circaeoides (F. Schmidt) Feer Perennial Russian Far East Physoplexis 1/1, Europe (Alps) P. comosa (L.) Schur Perennial Switzerland * Phyteuma 7/40, Europe P. betonicifolium Vill. Perennial France* P. ое Sternb. & Hoppe Perennial France* P. orbici "idi Perennial E Europe P. spice E Perennial E Europe* P. vagneri ГА. Kern. Perennial E Europe* Platycodon 1/1, E Asia P. grandiflorus (Jacq.) A. DC. Perennial Russian Far East Popoviocodonia 1/1, Russian Far East P. uyemurae (Kudo) Fed. Perennial Russian Far East Roella 1/25, South Africa R. ciliata L. Perennial South Africa Sergia 1/2, C Asia S. sewerzowii (Regel) Fed. Perennial C Asia Symphyandra 4/12, Е Mediterranean S. armena (Steven) A. DC. Perennial Caucasus S. cretica А. DC. Perennial Greece* S. hofmannii Pant. Biennial France* S. pendula (Bieb.) A. DC. Perennial Switzerland* Trachelium 2/7, Mediterranean L. Shrublet Italy* T. rumelianum Hampe Shrublet Italy* Wahlenbergia 6/150, Southern Hemisphere, Europe, SE Asia W. albomarginata Hook. f. Ann., per. Great Britain* W. gracilis (Forst.) A. DC. Ann., per. Great Britain* W. hederacea (L.) Reichenb. Ann., per. Great. Britain* W. procumbens A. DC. Ann., per Great Britain W. undulata (L. f.) A. DC. Ann., per. South Africa Zeugandra 1/2, Middle East (Iran) Z. iranica P. H. Davis Perennial Iran Number of species follows Mabberley (1997), except C oo (Shrestha, 1992) and Edraianthus (Lakuxic, 1974). Taxonomie division within M follows Fedorov (1957). * Species of cultivated origin Volume 90, Number 4 2003 Shulkina et al. 583 Morphological Studies of Campanulaceae baceous vines | (Campanumoea, | Canarina, Codonopsis). The different life forms in the Campanulaceae have been accommodated within several commonly used gross morphological systems (Du Rietz, 1931; Raunkiaer, 1934; Serebrjakov, 1962). Comparison between these gross morphological groups and tax- The same life form may be present in different tribes, onomic classifications shows no agreement. and individual tribes may include more than one life form. A single genus can include life forms with different life spans (e.g., Campanula includes pe- rennials, biennials, and annuals). Closely related species sometimes have different types of adapta- tion (e.g., C. hohenackeri has a well-developed pri- mary root system, whereas mature plants of C. cau- casica have rhizomes, and both species belong to the same subsection Triloculares). Therefore, the life form groups arranged according to existing mor- phological systems do not correlate with Campan- ulaceae taxonomic classifications, and perhaps the current taxonomic systems do not reflect. natural groups within the family. SEEDLING MORPHOLOGY The initial and early stages of plant growth are significant to the survival of seedlings in various kinds of environments (Stebbins, 1971, 1974). Seedling morphology (along with other characters) has been useful for the delimitation of taxa above the generic level s some families such as Cras- sulaceae (Ohba, 19 and Sapotaceae (Bokdam, 1977), at the generic lev- el within the tribe Cynometreae of the Fabaceae (Léonard, 1957), and species level in the genus Calophyllum (Stevens, 1980). АП species studied in Campanulaceae have epi- geal (aboveground) germination yledons usually oval in shape with an apical notch. Cotyledons may be as large as 6.3 X 8.3 mm (Can- 1.0 mm 78), Gesneriaceae (Burtt, 1977). — Fig. 1), with cot- arina canariensis) and as small as 1.5 X (Gadellia lactiflora). The primary leaves emerge in two or three weeks, and cotyledons persist during the first two months of development. The position of the primary leaves varies, and as a result there are two different types of seedlings within the fam- ily (see Table The first seedling group, “Group A,” gated epicotyl (1—11 cm long) and elongated first internodes. The length of the epicotyl and inter- nodes may vary even in one species under different has an elon- conditions. Young plants of some species can pro- duce shorter internodes vorable habitats. Thus, Ostrowskia magnifica grown when occurring in unfa- in the open air in St. Petersburg might have inter- nodes 2—3 cm long, whereas in greenhouses and in its native habitat in Central Asia the internodes are 10 em or more. Although the length of the epicotyl and internodes may vary, they are always present. A second group, *Group B," has no visible epi- cotyl, and the first internodes are practically ab- sent. Leaves appear immediately above cotyledons and form a rosette. Seedling morphology in each genus is relatively uniform. It is true not only with oligotypic genera such as Canarina, Musschia, Sergia, and Trachel- ium, but also with rich genera such as Campanula and Phyteuma. Species of Campanula studied here belong mostly to the flora of the FSU and include representatives of both sections and the 24 subsec- tions of Fedorov's (1957) classification. Campanula is morphologically heterogeneous, and seedlings differ markedly in size, first leaf shapes, and de- velopment patterns. However, they all form rosettes at the beginning of growth. All examined species do not have an epicotyl and the first internodes are very short. On the basis of seedlings, Annaea, Hem- isphaera, Neocodon, Theodorovia—genera de- scribed by Kolakovsky—do not stand apart notice- ably from the other Campanula species. The only Campanula that has ап elongated. seedling is C. lactiflora, which is now segregated in Gadellia. Other consistent characters of vegetative and re- productive organs are common to species of each seedling group. In Group A, the plant is sympodial and its leaves are opposite (Campanumoea, Can- arina, Cyananthus, Os- trowskia, Platycodon), at least in ontogenesis. In Codonopsis, Leptocodon, such mature plants the leaf arrangement may re- main opposite or become whorled (Canarina, Os- trowskia) or spiral (some species of Cyananthus). These species are perennials; only Legousia and some Cyananthus species are annuals (Shrestha, 1992). All have sympodial growth patterns, and shoots die every year even if they do not terminate in a flower, and the next years shoots come from axillary buds. Following the first year's growth, the plants are dormant during the unfavorable season. be it cold or dry. Thus, Codonopsis, Leptocodon, and Platycodon, which occur in eastern Asian regions without snow cover, have a deep dormant period in winter. Canarina (Canary Islands and East Africa) and Ostrowskia (Central Asia) are both geophytes growing in a climate with a long dry period (spring and summer for Canarina and summer for Ostrows- kia), during which they are dormant. Even in the greenhouses with constant warmth and humidity these plants have a deep dormant period. In seedling Group B, all plants have a rosette of 584 Annals of the Missouri Botanical Garden T 5 cm A. - 5 cm Ld Figure l. Seedling morphology. Examples of plants with shortened epicotyl (a to 1). First row: —a. Campanula latifolia. —b. Symphyandra armena. —c. Brachycodonia fastigiata. —d. Adenophora liliifolia. —e. Popoviocodonia uyemurae. Second row (scale a as above): —f. Mic "hauxia laevigata, —g. Phyteuma spicatum. —h. Asyneuma е8 шт. —4. Sergia sewerzowii. —j. Edraianthus атш ои Third row: —k. Сатратиамгит americanum. —]. warpa circaeoides. Examples of plants with an elongated epicotyl (m to t). —m. Gadellia lactiflora. —n. Azorina haha —о. Мше hia aurea. Fourth row: в oum grandiflorus. —Qq. С‹ »donopsis pilosula. —r. Cyananthus microphyllus. —s. Canarina canariensis. —t. Ostrowskia magnifica, with pen enh in the first year and primary stem the next year. All plants taken from collections at LE. Volume 90, Number 4 2003 Shulkina et al. 585 rd Studies of Campanulaceae Table 3. Seedling grouping in Campanulaceae. roup А Gro (elongated epicotyl) (reduced epicotyl, rosette-formers) Azorina Adenophora Campanumoea Asyneuma Canarina rachycodonia Codonopsis Campanula Cyananthus Campanulastrum Gadellia Cryptocodon Legousia Cylindrocarpa Leptocodon Phi iie aera Musschia Githop. Ostrowskia Bonds Platycodon Jasione Michauxia Peraca ja Physopl Popovioc enis Roella Sergia Symphyandra Trachelium Zeugandra leaves or at least the first internodes are shortened in early ontogenesis. Plants have spirally arranged leaves. The group includes annuals, biennials, pe- rennials, and semi-arborescent forms with various types of seasonal development and branching. An- nuals can have 2 to 4 leaves in a rosette, and the epicotyl axis terminates in a flower. All subsequent reproductive branches usually arise from the mer- istems in the upper leaf axils just beneath the ter- minal flower, e.g., Campanula erinus, Githopsis ca- lycina. In biennials an epicotyl axis produces a rosette of leaves during the first year (up to 100 leaves in Campanula medium) and elongated internodes the next year that terminate in a flower or in thyrsoid inflorescences, panula barbata, C crispa, C. medium, Michauxia laevigata. In many perennials, e.g.. Campanula latifolia, C. ‚ Asyneuma pulchellum, Cam- Adenophora lilüfolia, Asyneuma japonicum, glomerata, C. alliarüfolia, the main stem comes into flower after the production of short nodes over 2—4 years of growth. Further stem growth occurs from axillary buds after a dormant period. In some species mature plants do not have aboveground ro- settes of leaves, but two or three pairs of scale leaves, below ground, e.g.. all examined species of Adenophora. In the Mediterranean region, some plants retain green leaves not only in a basal ro- sette, but also at the mid-fertile nodes, and sub- sequent branches derive from axillary buds. Their perennial stems are lignified, forming arborescent semi-shrubs, e.g., Trachelium caeruleum. There are also many species of Group B with an indeterminate apical meristem. Many of these Cam- panulaceae develop a basal rosette that can over- winter under the snow. Reproductive branches are axillary, often leafless. In this case the branching pattern is clearly monopodial, e.g., Campanula an- omala, Edraianthus graminifolius. The main rosette can persist or be replaced by axillary ones that also grow monopodially and bear second-order repro- ductive leafless stems. In some cases plants are monopodial but reproductive stems are leafy, e.g., species from subsection Trigonophyllon, such as Campanula autraniana. The other extreme is Cam- panula karakushensis, where the main rosette pro- duces cataphylls only, and it is the axillary stems that are leafy and bear an inflorescence. In Cam- panula polymorpha, C. rotundifolia, and C. uniflora the apical meristem does not participate in forma- tion of the plant body. The epicotyl axis produces a rosette of two or three leaves, after which the apical meristem diminishes. Axillary elongated stems are produced by basitonic (sympodial) branching terminating in inflorescences. Successive branches are produced from lower leaf axils on these second-order reproductive stems, and the branching pattern. becomes basically sympodial. All examined species of each subsection of Cam- panula have a similar branching pattern, and this character is of taxonomic value within this genus. Perennials of Adenophora, Astrocodon, Asyneu- ma, many species of Campanula, Cryptocodon, some Phyteuma, and Popoviocodonia enter dorman- cy after their initial anthesis. Also, perennial Cy- lindrocarpa, Diosphaera, Edraianthus, Jasione, Physoplexis, Sergia, Symphyandra, Trachelium, and though life form and seasonal rhythm vary widely within Group B, all plants of this group, including biennials and annuals, start as rosette plants. An interesting correlation was found between seedling types and pollen grains. Dunbar (1973) and especially Avetisjan (1986, 1988) studied pol- len within the family, the latter describing four groupings divided into nine types. These roughly sort into two assemblages corresponding to or co- incident with seedling Groups A and B (see Fig. 2). The first pollen assemblage includes meridio- nal-zonocolpate, equatorial-colporate, and colporo- idate pollen grains and includes Cyananthus, Co- donopsis, Leptocodon, Ostrowskia, Platycodon, and Canarina. 'The second pollen assemblage has por- Bra- ate grains and includes Asyneuma, Azorina, Annals of the Missouri Botanical Garden Figure 2. Pollen grains: —a. Musschia aurea, porate, Lowe 161283 (MO). —b. rr a ЖУРТУ "d, LM Central Asia, Turkestan Range. Shulkina s.n. (LE). —c. M canariensis, colporate, Crosby 11425 (MO). —d. Gadellia lactiflora, porate, Caucasus, Teberda, Shulkina s.n. (LE Volume 90, Number 4 2003 Shulkina et al. 587 Morphological Studies of Campanulaceae Figure 3. Two ope of p b arivides (A), Platycodon grandiflorus (B), Cyan inthus integer (С). lants within the Campanulaceae family. i Plants with elongated seedlings: Canarina . Plants with rosette seedlings: Campanula mirabilis (D), den pon tetraphylla (E), Campanula tridentata (F), yel vidalii (С). chycodonia, Campanula, Campanulastrum, Ed- raianthus, Gadellia, Jasione, Legousia, Michauxia, Musschia, Peracarpa, Popoviocodonia, Roella, Ser- gia, Symphyandra, and Trachelium, covering those species of Campanula that were segregated into An- naea, Hemisphaera, Neocodon, and Theodorovia. Almost all species of seedling Group A have col- pate, colporate, and colporoidate pollen grains, whereas species of seedling Group B have only por- ate pollen grains (Fig. 3). There are some excep- species of Azorina, Gadellia, Musschia, and Legousia develop an elongated epicotyl (seedling tions: 588 Annals of the Missouri Botanical Garden Group A), though their pollen grains are porate, as in seedling Group B. Three of these genera, Azo- rina, Gadellia, and Musschia, have spiral leaves in the earliest seedling stages. Azorina and Musschia both occur in a warm unseasonable climate and have continuing monopodial growth throughout the year, exceptional within Group A. The only char- acter that associates Azorina and Musschia with group A is their elongated stem in the first. year, which is pronounced to as long as 50-70 ст. Spe- cies of these two genera have arborescent life forms unusual within Campanulaceae. Azorina vidalii, which grows in the Azores Islands, is a dwarf tree or shrub to 1.5 m high (Feer, 1890; Vasilevskaya & Shulkina, 1976). During the first year it develops an elongated epicotyl and stem with elongated in- ternodes. Subsequently, the internodes become shortened, but the main stem remains vegetative, and the axillary branches, all with elongated nodes, produce inflorescences in 2 to 3 years and die after fruiting. Two species of Musschia occur on the Ma- deira Archipelago. Musschia wollastonii is a mono- carpic, unbranched dwarf tree to 1.5 m tall when flowering, with a rosette of large leaves (to 70 cm long) elevated above ground. The stems produce elongated internodes during the first year and short- ened ones in following years. It comes to flower in 2 to 5 years and the flowering period lasts 4 to 6 months. The stems are crowned by long inflores- cences (70-90 cm), and plants die after fruiting. Musschia aurea is a dwarf shrub 0.4—0.7 m high. The main stem has elongated internodes in the first year and rather shortened ones in the following years. All axillary branches are equivalent in length to the main one. The plant grows 2 to 5 years before flowering and inflorescences are terminal on the re- productive branches, which are monocarpic and die after fruiting. Molecular data from ITS sequences also UP the position of Azorina (Shulkina & Gaskin, 1999) and Musschia (Eddie, 1984; Eddie et al. Pei this issue) within Group B. Gadellia lactiflora (seedling Group A herein) was segregated from 1979). gated seedlings, an unusual growth pattern with a Endemic to the Caucasus, Campanula (Shulkina, Gadellia has elon- dormant period and sympodial growth after the first — year; an unusual chromosome number (2n = 36 and morphology (Gadella, in flowers such as narrow filaments; pollen grains 1964); some peculiarity with two pores (Shulkina, 1979); and an unusual seplicidal fruit, which is dehiscent by pores and regularly cracks along the septa up to the axis col- umn (Kolakovsky, 1986). Molecular (Eddie et al., 2003), serological (Gudkova & Borschenko, 1986), and seed morphology (Belyayev, 1984, 1985) data also support its segregation from Campanula. At the same time, it has many characters in common with Campanula, including its spiral leaf arrange- ment, which proves that its elongated stem is of secondary origin ne last exception in seedling Group A is the genus Legousia, the taxonomic position of which has been controversial (MeVaugh, 1948; Fedorov, 1957) within Campanulaceae. The prismatic cap- sules and almost rotate corollas distinguish it from all other related taxa, sensu Phyteumateae (Fedo- rov, 1957). Shetler and Morin (1986), who investi- gated the seed structure of the North American Campanulaceae, also concluded that the taxonomic position of Legousia is unclear and more study is needed. Serological studies revealed differences separating Legousia from other genera within Phy- teumateae (Gudkova & Borshchenko, 1991), and its elongated seedling is also a character that suggests reconsideration of its taxonomic position. Molecular studies (Eddie et al., 2003) show Legousia is nearer to Campanulastrum than Phyteuma. This division. within Campanulaceae based on seedling type almost completely coincides with De Candolle’s (1830, 1839) system. De Candolle's work included only half the genera now known, but the comparison is potentially useful. De Candolle rec- ognized two major groups: Wahlenbergieae and Campanuleae (a third tribe, Merciereae, includes a single South. African genus, Merciera, with 3 spe- cies, which was unfortunately unavailable for this study). The tribe Wahlenbergieae includes genera with "capsula apice dehiscens," whereas the tribe Campanuleae has plants with "capsula lateralitier dehiscens" (De Candolle, 1830). Almost all plants from his tribe Wahlenbergieae have "elongated" seedlings (Group A), whereas plants from Campan- uleae have a “rosette” type of seedling (Group B). A few exceptions need further discussion. De Candolle's division was based on external fruit structure. He placed Edraianthus and Jasione in the tribe Wahlenbergieae because both have api- cally dehiscent capsules. Kolakovsky (1982, 1995), who studied internal fruit. structure, showed that fruits of many genera їп Campanulaceae have a special organ (special tissue) that helps to open a capsule. The list of genera with an axicorn (as it was named by Kolakovsky) includes Adenophora, Asyneuma, Campanula, Michauxia, Phyteuma, Po- poviocodonia, Sergia, and also Edraianthus and some other genera of Group B. This axicorn opens a pore on the lateral wall of the fruit in Campanula and other mentioned genera, while in Edraianthus it irregularly tears apart the membranous top of the capsule. Thus, capsules in Edraianthus and Cam- Volume 90, Number 4 2003 Shulki 589 мок Studies of Campanulaceae panula open in different places but by the same mechanism. An explanation probably lies in a type of inflorescence of Edraianthus. All species of this genus have capitate inflorescences surrounded by bracts (Lakuzié, the fruit facilitates seed dispersal more readily than a basal or lateral opening. Therefore, the capsule of Edraianthus differs in its dehiscence mechanism from those of other genera with apical valves. In- deed, Edraianthus is related to our seedling group with basal rosettes, and this о is sup- ported by molecular data (Eddie et al., On the other hand, the groups in EE sys- tem are very heterogeneous in growth and seedling characters. Thus, according to morphological divi- sion Canarina should be excluded from the Cam- panula alliance. Edraianthus is closely related to Campanula and not to genera with apical capsule dehiscence and should be excluded from the sub- tribe Wahlenberginae. Musschia should be exclud- ed from subtribe Platycodinae. The taxonomic po- sition of Legousia should be reconsidered. There is greater similarity between our morpho- (1997) system. His first three subfamilies (Cyanthoideae, Ostrowskioi- logical groups and Takhtajan’s deae, Canarinoideae) include genera with an elon- gated epicotyl, our Group A. All studied species within our Group B, with a shortened epicotyl, be- long to his subfamily Campanuloideae. Anomalous taxa (Azorina, Musschia) in which the elongated epicotyl may be of secondary origin are also in this subfamily, but isolated in separate tribes. Data from molecular biology, such as chloroplast DNA structural changes, can contribute to Cam- panulaceae classification and have already been used in phylogenetic reconstruction of the Lobeli- 1993). Recent molecular anal- yses of the Campanulaceae based on rbcL sequenc- es (Cosner et al., 1994) and nuclear ribosomal DNA ITS sequence data of 93 taxa (Eddie et al., 2001, 003) support two major lineages within the family (Shulkina & Gaskin, aceae (Knox et al., CONCLUSIONS Seedling morphology appears to be a useful char- acter for the classification of Campanulaceae, with two major groups evident. The first one, seedling Group A (Campanumoea, Canarina, Leptocodon, Ostrowskia, Platycodon, and Cyanan- Codonopsis, thus) share elongated seedlings, opposite leaves (at least in the early stage), sy peciam branching. and dormancy after the first year. Flowers are mostly in cymose inflorescences (Platycodon, Canarina, Os- trowskia), rarely solitary in the high mountain spe- 1973), and the apical opening of cies (Cyananthus). The ovary is superior, half-in- ferior, or inferior. The pollen grains range from 6- to 10-colpate (Cyananthus), colporate (Canarina, Platycodon), or colporoidate (Campanumoea), seen as primitive types within the Campanulaceae (Av- etisjan, 1988). Capsule dehiscence is mostly apical by valves (Codonopsis, Leptocodon, Platycodon, Cy- ananthus), or lateral by cracks (Ostrowskia); there is no axicorn. Almost all taxa are diploid, with 2n = 14,1 Arano & Saito, 1979). Vessels with scalariform perforation plates are found in Cyanan- thus, Platycodon, and Canarina (Shulkina & Zykov, 1980). Genera of this group occur mostly in East Asia, and only Canarina has a disjunct distribution (Macaronesia and eastern Africa). Most genera are monotypic or oligotypic (Canarina, Leptocodon, Os- trowskia, Platycodon, Campanumoea), and many are considered paleorelicts with unclear relation- ships (Hedberg, 1961; Popov, 1963): their taxonom- ic positions vary in different systems. тоир іп Campanulaceae includes genera with rosette seedlings, spiral leaf arrangement, and different branching patterns (sympodial, monopo- dial). Immature plants of this group show no fixed dormancy; mature plants have various seasonal growth patterns and different life forms. The flowers are usually in cymose inflorescences modified into umbel-like, spike-like, and solitary forms. The ova- ry is inferior. All studied species have porate (in- cluding zonoporate and pantoporate) pollen grains. Fruit dehiscence varies, but the capsules never have apical valves; an axicorn is sometimes pre- sent. The chromosome numbers vary greatly, with numerous polyploid lines, but x — 17 in many Gadella, 1964). The representatives of this group —. are widely distributed. Taxonomically the group in- cludes the tribes Campanuleae (6 genera) and Phy- teumateae (6 genera), among them large genera such as Campanula (300 spp.) Asyneuma (50 spp.). е (40 spp.), and Phyteuma (40 spp.). merous smaller genera with restricted ranges are Githopsis (western North America), Ed- raianthus (Apennines and Balkan Peninsula), and Michauxia (Turkey, the southern Transcaucasus, Iran). There are also monotypic and oligotypic gen- Azorina (Azores), Cryptocodon (Pamiro—Alay Mountains), Cylindrocarpa (Karatau, Tien Shan), Musschia (Madeira Islands), Physoplexis (southern Alps), Popoviocodonia (Russian Far East), Sergia (Tien Shan), and Zeugandra (northern lran), etc. These habitats and environmental conditions obvi- ously vary greatly, and the plants of the group have numerous life forms. АП data lead us to conclude that the basal rosette and a shortened type of seed- ling represent morphological apomorphies. There is Annals of the Missouri Botanical Garden strong evidence that the "elongated" type of seed- ling is plesiomorphic and characterizes more prim- itive Campanulaceae forms. Therefore, two evolutionary directions, two line- ages, can be traced within the family which cor- respond to the above two groups and probably re- flect differences in the environments occupied by ancestral types. The recently recognized genus Gadellia (Shulkina, 1979) and the critical Campan- ulastrum (Small, 1903) are supported by morpho- logical and molecular data, and both fall outside of Campanula s. str. The sister taxa to Gadellia in the ITS study (Eddie et al., 2003) is Musschia aurea, and this supports Gadellia as a genus distinct from Campanula, Campanulastrum americanum (Cam- panula americana) of the “rosette Group B” is not close to the Campanula alliance. Studies of pollen grain (Avetisjan, 1988), chromosome number and morphology (Gadella, 1964), seed-coat morphology (Shetler & Morin, 1986), and molecular data (Eddie et al., 2003) support segregation of Campanulas- The genus Campanula is highly heteroge- i Further trum. neous and should be studied carefully. morphological and molecular investigations are needed to increase our understanding of monophy- letic groups within this family. In Campanulaceae similarities due to convergent and parallel evolu- tion occur both in reproductive and vegetative structures. All characters should be used in con- junction with others. Literature Cited Arano, . Saito. 1979. The karyotypes and chr mosome » evolution i in family С Күт eae (Japan) of Asterales. Kromosomo П (15-16): 433—447. Avetisjan, E. M. 1986 E of the families Campanulaceae, Sphe nocleaceae and Pen eaea ceae. Bot. Zhurn. (Moscow P Leningrad) 7 1003— 1009. үм ros —— 3. P eos of the superorder м си zia institute of Botany, Erevan. [In Russian.] Be es у, 984. Seed anatomy in some re odis tives of the ( Ия family. Bot. m . (Moscow & Le pop 69: 585—594. [In Russia Comparative Anatomy of Seeds within the‏ س Campanulac eae Family. Thesis, rh Komarov Botan-‏ ical dn en mingra id. [In Russian.]‏ Bokdam, J. 1977 sepes e of some African тунбу еае н their taxonomic signific Meded. Rilke Landings hool 20: 1-84. Burtt, B. L. sification above the genus, as кр by Gesneriaceae, with aepo ‘ls from dba groups. Pl. Syst. Evol. Suppl. 1: 97-10 C andolle, A. 6 18: a Monographie des е аайы aris. 839. Campanulac eae. oe VII: M 1497. Cosner, ansen & ». Lammers. 1994. Phylogenetic re uen uh in the € Campanulales based rbcL sequen Syst. Evol. 190: Czerepanov, S. K. ^ el л апсе. 2. =~] 1995, Vascular Plants x PAM and Adjacent Countries. Cambridge Univ. Press, Cam- bridge. Damboldt, J. 1976. Materials for a flora of Turkey Campanulaceae. Notes Roy. Bot. Gard. Edinburgh 35 39-52 Pollen ontogeny in some species of A study by electron microscopy. Bot. Dunbar, A. 1973. Campanulaceae. Not. 126: 277-315. Du Rietz. 1931. Life Forms of Terrestrial. Flowering PES Univ. Press, "e vs ala. Eddie, W. M. M. I9 A Systematic Study of the Genus Mussc du бо with Reference to Character Di- versity and Evolution in the E NUM A Sc. Thesis, University of Rea . A Global Reassessment of enr ;eneric Re- m in the Bellflower Family (Campanulaceae Ph.l Pu University of Pabst о 1 — ——, T. V. Shulkina, J. F. Gaskin, R. Haberle € Tons n. e )1. Reconstruction of d Kwon of tii Campanulaceae s. str. using ITS sequences of nuclear dapi DNA. B 2001. Albuquerque, New Mex- "E Abstract 445. ш “ae. = —_, —, ———, —— $ 3. Phy- logeny of Ganpanulece eae s. str. “inferred fro m TIS se- quences of nucle аг са DNA. Ann. Missouri Bot. Gard. 90: 554—57 Fedorov, A. A. 1957 (editor), Flora SSSR. Moscow- v nnd — —«& M. anda. Tutin et al. (edito rs), Flora Europaea 4: ridge p Press, Cambridge. бы: Н. 1890. Beitrage zur Systematik 97 Aye و‎ der ergs een. Engler Bot. Jahrb. Gadella, T. W. G. 1964. Cytotaxonomic Aro in dn ge- nus Campa. We itia. 11: 1-104. Gudkov ¥ G. P. Borshchenko. 1986. караш s ^i sides within the Campanulaceae fami Abstract: 51-53 in Chemosystematics and Evolution of Vascular Plants. All-Union Conference, the MBG, Mos- cow. [In Russian. | #2 . Shishkin Nauk, Campanulac je In: B. 24: 450. Academia In Russian. | 1976. - Campanula L. In: T. G. 74—92. Cam- 1 1991. The serological study of the Campanulaceae. The iilo аны relations in the tribe Phyteumateae. Bot. Zhurn. (Moscow & Leningrad) 76: 809-817 2 sian. | Hedberg, 0. d of the genus Canarina L. ep are d. Sven. Bot. Tidskr. 55: 17—62. Hong, ges -Y. 1983. ( Campania = Fl. Reipubl. Popularis Sir (е, cience Press, Beijing. . 199 the geography of the ( Ce pales: On Acta Phytotax. Sin. 33: 521- the Шешип centers. 530. Koi E. . R. Downie & J. D. Palmer. 1993, Chlo- roplast genome rearrangements and the evolution of gi- ant lobelias from herbaceous ancestors. Molec. Biol. Evol. 10: 414—430. Kolakovsky, A. A. 1982. The biological *mechanism" en- suring dissemination in pv thus. Soobshch. Akad. Nauk Gruzinsk. S.S.R. 105: 361—363 . 1986. piger of C lace: of taxonomy. Zhurn. (Mose ow & Leningrad) 7 1155-1168. [In Russian, | 1995. The Campanulaceae — ~ and о ms Family. Agent, Mos- y = cow. [In Russian. | Lakušić, R. 1973. Das Natürliche System der Sari onen und der Arten der Gattung Edraianthus DC. God. iol. Inst. u Sarajevu. 26. Posebno Izdanje. Volume 90, Number 4 Shulkina et al. 591 Morphological Studies of Campanulaceae . 1974. Edraianthus DC. God. Biol. Inst. u Sara- u. 26. Розета Izdanje. «one: 992. Circumscription and phylogeny of the dS Ann. Missouri Bot. Gard. 79: 388— 413. Léonard, J. 1957. Genera des Cynometreae et Amherstiea africanes (1 potum Caesalpinoideae). Essai de blastogénie appliqué à d systématique. Mem. Acad. Roy. Belg. Coll. 30: 1-305 г 1997. The Plant-Book, 2nd ed. Cam- b i кее R. 1948 "ularia. Rhodora 50: Ohba. H. 3. Generic and infrageneric classific vation of the uc 2 Sedoideae (Crassulaceae). J. Fac. Sci., . Generic Sta of Triodanis and Spe- 3849. Univ. Tokyo, е 12: 139-198. Popov, M. С 196 PES Princ iplos and Cone epts of Phy- togenetics . Moscow. [In Russ Raunkiaer, C. 1934. Plant Life Forms. Oxford Univ. Press, Oxford. Raven, P. H. & D. J. Axelrod. 1974. Angiosperm bioge- ography ar di past continental movements. Ann. Missouri Bot. Gard. 61: 529—673. Schénland, S. 1889. Campanulaceae. In: A. Engler & К. Prantl (editors), Die natürlichen Piausetfamilié i IV, : 1-70. W. Engelmann, Leipzig pos 1. G. 1962. Ecological Plant Morphology. Shkola, Moscow. [In Russian. Shetler, S. G. & N. R. Morin. 1986. Seed morphology in Northern Ame rica Conpaselas eae. Ann. Missouri Bot. 53 The و ا‎ Revision of the Genus Cyananthus Wall. ex Benth. Thesis, Komarov Bolanical Institute, St. Petersburg. [In Rus 55 Shulkina, T. V. 1978. Life forms in C e MN eae, the geographical distribution and connection with taxono- my. Bot. Zhurn. (Moscow & Leningrad) 63: 153—169. [In Russian.] . 1979. De positione systematica Campanula la tiflora Bieb. Novosti Sist. Vyssh. Rast. 16: 175-179. [in Russian. | & J. F. Gaskin. 1999. Gross morphology and mo- data in taxonomy of gt ЖЕ pacer? Interna- Ab- : XVI. St. Louis 1999, lecular tional. Botanical Congress stract 16.1. : ES Zykov. 1980. The anatomic: өк structure of the stem in = Campanulaceae s. str. ation to the evolution of life forms. Bot. = urn. (Mosco ow & Lenin- grad) 65: 627-638. |In Russ Small, J. K. 1903. Flora of he сайа тп United States. Published by ys New York. Stebbins, G. L. 19 1. Adaptive radiation of reproductive characteristics in dur اد‎ seeds and seedlings. Ann. Rev. Ecol. Syst. 2: 237—260. ‚ 1974. Flowering Plants. Evolution above the Spe- cies Level коч Univ. I ridge. Stevens, P. 80. A revision of the Old W World аре ies ч alors (Guttiferae). J. Arnold Arbor. 61: 117— ress, Cam тари. А. L. 1997. Diversity and Classification. of Flowering Plants. Columbia Univ. Press, New Y Tonbneon, P. B. 1984. Vegetative кык PE е nigmas in be to plant systematics. Pp. 49—66 i . H. Heywo . Moore (ec liio: Current Con- cepts in Pant Lies a Academic Press, ondon. Tutin, T. G. 6. Campanulaceae. /n: T. G. Tutin et al. (editors), ds Europaea 4: 74—102. r Univ. Press, Cambridge. Vasilevskaya, V. K. & T. V. Shulkina. 1976. Morphological and analomical structure of the arborescent plant Azo- rina vidalii. Trudy Moskovsk. Obšč. Isp. Prir. 42: 131— ). [In Russian. | PHYLOGENY OF SAURURACEAE BASED ON MORPHOLOGY AND FIVE REGIONS FROM THREE PLANT GENOMES' Shao-Wu Meng,?* Andrew W. Douglas," De-Zhu Li, Zhi-Duan Chen,’ ? and Jun-Bo Yang? Han-Xing Liang,’ ABSTRACT P id lic re ieu of the six extant species of four genera of the Saururaceae are resolved based on sec ITS from the nuclear genome; rbcL and trnL-F fre chondrial genome. Zippelia ini agen an from a genus of Piper: data: 188 in separate and combined analyses of s data sets, the combined DNA lence data sequence data. Forty-nine PALO ual cal char in this family, again using A a as outgroup. Whether the topologies of S sets, morphological data sets, morphologic val d sets, all are кк congruent. In all juence m the chloroplast genome; and matR from the mito- eraceae, is used as an outgroup. Results are presented acters reconstruct the phylogeny ururaceae are based on individual genomic the combined DNA sequence anc analyses, the monophyly of Saururus and Gymnotheca, re- spectively, is strongly че ун d ше sister relationship between Gymnotheca and Saururus is well supported. In s analysis of nuclear DNA data Anemopsis is the ister to abrir and i sni and with Saururus sister re Gymnotheca; however, in the analyses of the o the A sets, Anemopsis is the sister group of Houttuynia, and p sister group to all other Saururaceae, with Houttuynia n ther dat vemopsts—Houttuynia clade lies sister to the Sine acide a clade. The rui that the Anemopsis— و‎ ese с dee comprises the sister group of Saururus-Gymnotheca clade is novel and differs from previous phylogenetic opin ey words Saururus. Anemopsis, genomes, Gymnotheca, 1 morphology, multigene data, phylogeny, Saururaceae, Saururaceae are a core member of the paleoherbs (Tucker & Douglas, 1996) and are an ancient fam- ily with six species in the four relictual genera Sau- rurus, Gymnotheca, Anemopsis, and Houttuynia (Liang, 1995). These are all perennial herbs with simple flowers that bear bracts without perianths. Saururaceae have an East Asian—North American disjunction, with Anemopsis californica Hook. & Arn. and Saururus cernuus L. in North America, Houttuynia cordata Thunb., Gymnotheca chinensis Decne., Gymnotheca involucrata Pei, and Saururus chinensis (Lour.) Baill. in East Asia. Due to their basal systematic position and interesting geograph- ical pattern of distribution, Saururaceae have been of much phylogenetic interest, although they are a small family including six species. Current view- points on the phylogeny of Saururaceae diverge, based on gross morphology, cytology, and floral morphogenesis. Wu and Wang (1957) included Sa ururus, Circaeoc arpus, Anemopsis, Gymnotheca, and Houttuynia in Saururaceae and thought that Circaeocarpus, Anemopsis, Gymnotheca, and Hout- tuynia derived directly from Saururus one after an- other. Later, they (Wu & Wang, 1958) realized that the recently published Circaeocarpus (Wu & Wang, 1957) was in fact a member of Piperaceae, and Circaeocarpus saururoides C. Y. Wu and Zippelia begoniaefolia Blume ex Schult. & L. H. Schult. were synonymous. Considering their biogeography, Wu (1984) later thought Anemopsis and Houttuynia to be products of a vicariance event, and S. chi- nensis and cernuus were products of another * Author for correspondence (present address): Decision Systems Lab, Institute = Infocomm Research, 21 Heng Mui Keng Terrace, Singapore 119613. ' We expres e KIB), Cha-Cha Huang, Ying-Xue Sun, ;. Hollowell, Anthony R. Brach wmeng@i2r.a-sta . an anonymous reviev А а cial Grants of the N E b a He onto, bonn 650204, People's Republic of China. oratory of 100093. People's Republic of China. ' Department of Biology, University of Mississippi, University, Mississippi 386 GARD. 90: 592—602. 2003. ANN. Missouni Bor. Ya-Ping Hong, and Bao- Hua E wer, and Shirley “ng Song Ge, and Hang Sun for ж ‘Ipful discussions; is to redu.sg or T viis ss our hearty thanks to Zhen- Hua Guo, Yong. Yan Chen, Feng dr Mir ing Gao, Xin Tian, Bo Tian all IBC AS) for ihr assistance; to Victoria ucker for helpful чы to Xiao-Quan Wei Niu rok kindly nios sin the itg S primer sequences y tional ed Science Foundation of € tà (NSFC ) (39725001 1068, wc 30830020, en NSFC3 30030. nd Biogeography, не Institute of Botany, Chinese Academy of Sciences, Systematic and Evolutionary Botany, Institute of Botany, Chinese Academy of Sciences, Bei jing 77, U.S.A. Volume 90, Number 4 Meng et al. 593 2003 Phylogeny of Saururaceae Table 1. List of species and vouchers** for Saururaceae and its outgroup (Piperaceae). Species Origin Collection no. 188 ITS Ingroup Anemopsis californica Hook. & Arn. U.S.A. Louisiana. Liang 97017 AF197576* AF21592 Gymnotheca chinensis Decne он Yunnan: Xuanwei. Meng 99003 AY032643 AF203062 Gymnotheca involucrata Pei China. Sichuan: Emei. Liang 97015 AY032644 — AF20362 Houttuynia cordata Thunb. China. Guizhou: Xingyi. Meng 99301 AY032645 АЕ20362 Saururus cernuus L U.S.A. Louisiana. Liang 97016 AY032646 АЕ22330 Saururus chinensis (Lour.) Baill. China. Yunnan: Mengla. Meng 99525 AY032647 — AF21592 Outgroup Zippelia begoniifolia Blume ex Schult. & L. H. Schult. China. Yunnan: Menglun. Meng 99715 AY032648 AF203063 vouchers are deposited at KUN. * Obtained from GenBank. vicariance event. Based on basic chromosome num- bers of Saururus, Anemopsis, and Houttuynia, Oka- da (1986) proposed that Anemopsis and Houttuynia were derived from Saururus. Lei et al. (1991) sup- ported Okada’s opinion and thought that Gymnoth- eca was the more derived taxon based on chromo- some number. Based on a cladistic analysis of morphological and ontogenetic characters, Tucker et al. (1993) inferred that Saururus was the first to diverge from the ancestral Saururaceae, followed by Gymnotheca, with Houttuynia and Anemopsis as sis- ter taxa. Combining data from gross morphology. embryology, palynology, cytology, and flower development (Liang & Tucker, 1990; Liang, 1991, 1992, 1994, 1995), Liang (1995) proposed that the ancestors of Saururaceae were divided into anatomy, two lineages: Gymnotheca—Anemopsis and Sauru- rus—Houttuynia. In spite of the fact that certain Saururaceae, such as Saururus, Anemopsis, and Houttuynia, have been represented in recent studies of higher-level phy- logenetic relationships within the angiosperms (e.g., Chase et al., 1993; Soltis et al., 1997; Mathews & Donoghue, 1999; Qiu et al., 1999; Soltis et al., 2000), further investigation into the molecular sys- tematics of Saururaceae has been needed to deter- mine phylogenetic relationships within this family of basal angiosperms. Our phylogenetic assessment of Saururaceae is based on five genic regions from all three plant genomes: the 18S ribosomal R gene and ITS spacer (including 5.85) from the nu- clear genome; rbcL and trnL-F, its intron and gene spacer, from the chloroplast genome; and matR from the mitochondrial genome. Forty-nine mor- phological characters were also selected for phy- logenetic analysis (Appendix 1). These morpholog- ical characters were comprised of subsets from gross morphology, anatomy, embryology, palynology, cytology, and flower development. Generally, 185, rbcL, and matR genes have been used to recon- struct higher-level phylogeny, such as relationships among orders, families, or distantly re genera (e.g.. Chase et al., 1993; Soltis et al., iu et al., 1999), while ITS and trnL-F have bars been used for genera, species, and lower-level 1995; Bayer & Starr, We selected these five gene regions because questions (Baldwin et al., 1998). taxa in this family likely diverged at diverse points in time. MATERIALS AND METHODS PLANT MATERIALS All six species of the ingroup, Anemopsis cali- fornica, Gymnotheca chinensis, G. involucrata, Houttuynia cordata, Saururus cernuus, S. chinensis, and one designated outgroup, Zippelia begoniaefol- ia of Piperaceae, were collected from natural pop- ulations. Vouchers are deposited in the herbarium of Kunming Institute of Botany (KUN), Chinese Academy of Sciences, Kunming (see Table 1). The GenBank accession numbers of all relevant se- quences are included. DNA EXTRACTION, PCR, AND SEQUENCING Genomic DNA was extracted from silica-gel- dried or fresh leaves using a modified CTAB pro- cedure (Doyle & Doyle, 1987). PCR amplifications were conducted at a thermocycler (Perkin- Elmer 9600). It consisted of initial denaturization at 94° (4 min.), followed by 35 cycles of 94°C denaturi- zation (1 min.), 55°C annealing (1 min.), and 72°C extension (90 sec.), with a final extension for 7 min. at 72°C. The 18S primers used for amplification CTAGAGCTAATA- — and sequencing were 5' 594 Annals of the Missouri Botanical Garden CGTGCAAC 3' (I21F) and GTGGACTTC 3 (1692R). The primers of ITS, rbcL, trnL-F, and matR followed White et al. (1990), Feng et al. (1998), Taberlet et al. (1991), and Meng et al. (2002), respectively. PCR products were separated with 1.5 2 % agarose TAE gel and were purified using Wizard PCR Preps DNA Puri- fication System. Sequencing reactions were per- formed using PRISM Dye Terminator Cycle Se- quencing Ready Reaction Kit (Applied Biosystems, Inc.). The products of sequencing reaction were electrophoresed on an ABI PRISM 310 Genetic Analyzer (Applied Biosystems, Inc.), which per- forms capillary electrophoresis and can ensure ac- curate base stretch above 1200 bp in one sequenc- ing (guidebook about sequencing from Applied Biosystems, Inc.). Each studied DNA segment is sequenced twice from two ends in opposite direc- tions. SEQUENCE ALIGNMENT ANALYSES AND PHYLOGENETIC Contiguous DNA sequences were compiled using ка Inc.). were aligned using Clustal X (Thompson et al., 1997) and Mega 2b3 (Sudhir et al., 2000). Maxi- mum parsimony analyses were performed using PAUP 4.0 blO (Swofford, 2001). We branch-and-bound search option with furthest ad- Seq™ (Applied Biosystems, All sequences used the dition sequence. Gaps were treated as missing data. A bootstrap analysis was performed with 1000 rep- licates (Felsenstein, 1985). COMBINED DNA DATA ANALYSIS We first analyzed individual genomic data sets after combining 18S and ITS to represent nrDNA, combining rbcL and trnL-F to represent cpDNA, and finally using matR DNA to represent mito- chondrial genomic data. All DNA sequence data were then combined into one matrix to analyze. These combined data sets were analyzed using the same settings as the individual genomic data sets. After the phylogenetic tree was reconstructed from the combined DNA data sets, the matrix of all DNA ERI was re-analyzed with characters. re- weighted according to rescaled consistency indices (RC) (Farris, 1989). Separate matrixes of each genomic data set as well as the matrix of combined DNA data sets are available upon request from the corresponding au- thor. GATAAGGTTCA- MORPHOLOGICAL DATA ANALYSIS Forty-nine morphological characters were select- ed to reconstruct the phylogeny in this family (Ap- pendix 1). These characters were derived from her- barium specimens and literature (e.g.. Liang & et al., 1990, 1995; Liang, 1991, 1992, 1995; Tucker, 1975, 1980, 1981, 1982a, b, Tucker et al., 1993; Lei et al., 1991; Carl- 1995; Meng & Liang, 1997). Six char- acters were vegetative; 3 were from stem anatomy; Tucker 1994, 1985; quist et al., 15 were from floral morphology; 10 were from floral anatomy; 5 were from pollen; 8 were from embry- ology: and 2 were from cytology. Within this mor- phological character matrix (Appendix 2), 37 char- acters were treated as binary and 12 as multi-state. The analysis of Saururaceae was conducted using PAUP 4.0 blO (Swofford, 2001). АП characters were first equally treated as weighted and unor- dered. Other settings are the same as those for the molecular data. After a morphological phylogenetic tree was reconstructed using the above setting. the morphological matrix was also analyzed with char- acters reweighted according to rescaled consistency indices. ANALYSIS OF THE MORPHOLOGICAL COMBINED DNA AND DATA After replacing A, G, C, T from DNA sequence with 0, 1, 2, 3, respectively, we combined all DN data sets with the morphological one into а common matrix and re-analyzed. All characters were unor- Other PAUP 4.0 were the same as those for the combined dered and equally weighted. settings in molecular data. RECONSTRUCTION OF CHARACTER EVOLUTION Using the program WinClada v. 0.9.99m 7.5 beta (Nixon, 1999), we analyzed the morphological ma- trix again and recovered a phylogenetic tree topol- ogy similar to that from the PAUP 4.0 analysis. The distribution of each morphological character was then analyzed to investigate the evolution of each morphological character in Saururaceae. The com- bination of this heuristic search, 1000 replications, with one starting tree per replication, using a mul- tiple TBR and TBR search strategy, with zero ran- dom seed, and a slow optimization, was used in the maximum parsimony analysis. RESULTS NUCLEAR GENOME DATA ANALYSIS The alignment of 7 sequences resulted in a ma- trix of 2279 aligned positions, of which 195 were Volume 90, Number 4 2003 Meng et al. 595 Phylogeny of Saururaceae Gymnotheca chinensis Gymnotheca involucrata Saururus cernuus Saururus chinensis лур гуй гу Linn th 17. "4 Anomnners ‚л р J 7: L.L . C7: II [9] J The single most ж tree for Saururaceae based on nuclear genome data. Tree length = 434, el. ü- - 7 0.8963. RI = 0.7384, and RC = variable and uninformative but 137 were parsimo- ny-informative. Our percentage of phylogenetic-in- formative sites was 696. The uncorrected sequence divergence ranged from 0.09% to 6.517% within Saururaceae and from 9.233% to 10.11% between the outgroup and the ingroups. Pairwise distance comparisons of all data sets, including the individ- ual genomic DNA data sets, the combined DNA data sets, and the morphological data sets, are available from the corresponding author. A single most parsimonious tree of 434 steps was obtained from nrDNA (18S and ITS), with Cl = 0.8963, RI = 0.7384, and RC = 0.6618 (Fig. 1). The monophyly of Saururus (100% BS: bootstrap percentage) and Gymnotheca (100% BS) was strongly supported. Anemopsis i was the sister group of other Saururacege. Houttuynia cor- 0.6618. Bootstrap values are found above branches data was sister to the Saururus-Gymnotheca clade (85% BS). Saururus was sister to Gymnotheca (99% — We CHLOROPLAST GENOME DATA ANALYSIS Sequence alignment yielded 2400 bp. 166 of which were variably uninformative and 76 of which were parsimony-informative. The percentage of phylogenetic-informative sites was 3.167%. uncorrected sequence divergence ranged from 0.043% to 2.543% among Saururaceae sampled and from 7.846% to 8.38% between the outgroup in Piperaceae and the ingroup taxa. The A single most parsimonious tree of 270 steps was yielded for cpDNA data sets (СІ = 0.9444, КІ = 0.8315, and RC = 0.7853; Fig. 2). The monophyly 596 Annals of the Missouri Botanical Garden 100 Gymotheca chinensis 76 100 Gymmotheca involucrata Sawurus cernuus 9] Saururus chinensis [— — — —— Anemoypsis californica — — — — Houttuynia cordata Figure 2. The single most Dok tree for Saururaceae based on chloroplast genomic data. Tree length = 0.7853. Bootstrap values are found above branches. 270, CI = 0.9444, RI = 0.8315. and RC of Gymnotheca (100% BS) and Saururus (100% BS) was strongly supported. Moreover, A. californica was the sister group of H. cordata (91% BS), and the Anemopsis-Houttuynia clade was sister to the Saururus—Gymnotheca clade. The sister relation- ship of Saururus and Gymnotheca was supported (76% BS). MITOCHONDRIAL GENOME DATA ANALYSIS Sequence alignment yielded 1777 bp, 59 of which were at variable sites and 19 at parsimony- informative sites. The percentage of phylogeneti- cally informative sites was 1.07%. The uncorrected sequence divergence ranged from О to 1.753% эрер Saururaceae sampled and from 1.439% to 2.064 A single most parsimonious tree of 63 ste ps was % between the outgroup and the ingroup taxa. yielded for mitochondrial genomic data (matR), and y topo ogy of the tree matches dayi 2, with = 0.9524, RI = 0.8696, and RC = 0.8282. T, of Gymnotheca (99% BS) Saururus (99% BS) was resolved with strong internal support. CI The Anemopsis californica was the sister group of H. cor- data (73% BS). was sister to the Saururus-Gymnotheca clade. Sau- The Anemopsis—Houttuynia clade rurus was sister to Gymnotheca (74% BS). COMBINED MOLECULAR DATA ANALYSIS After all molecular data were combined, there were 6456 bp in the matrix: 633 of them were var- iable, 232 were parsimony-informative. The per- centage of phylogenetic-informative sites was 3.59%. ranged from 0.111% to 3.365% among included The uncorrected sequence divergence Volume 90, Number 4 2003 Meng et al. 597 Phylogeny of Saururaceae Saururaceae and from 6.712% to 7.073% between the outgroup and the ingroups. A single most parsimonious tree of 775 steps was obtained for the combined molecular data sets, with the topology of the tree as in Figure 2, wit - 0.9084. RI = 0.75, and RC = 0.6813. The mono- phyly of Saururus (100% BS) and Gymnotheca (100% BS) was strongly supported. Anemopsis cal- ifornica was the sister group of H. cordata (52% BS), and the Anemopsis—Houttuynia clade sister to the Saururus-Gymnotheca clade. Saururus was then sister to Gymnotheca (100% BS). A stable topology was generated after the matrix of combined DNA data sets was re-analyzed once with characters reweighted according to RC (base weight — 2). The topology was still identical to the previous one (Fig. 2). However, the following pa- rameters and каша values were much higher: tree length = 1276, CI = 0.9953, RI = 0.9839, and RC = 0.9792. Again, the monophyly of Sau- rurus (100% BS) and Gymnotheca (100% BS), and sister-group relationships between Saururus and Gymnotheca (100% BS), and between Anemopsis and Houttuynia (100% BS), were strongly support- ed. MORPHOLOGICAL DATA ANALYSIS Of the 49 characters considered, 19 were vari- able-uninformative and 29 were parsimony-infor- mative. The percentage of phylogenetic-informative characters was 59.18%. The uncorrected character divergence ranged oii 2.041% to 61.702% within the Saururaceae and from 53.061% to 63.83% be- tween the outgroup and the ingroups. A single most parsimonious tree of 71 steps was produced for the morphological matrix, and the to- pology of the tree was again congruent with Figure 2, with CI = 0.8451, RI = 0.7442, and RC = 0.6289. The monophyly of Saururus (100% BS) and Gymnotheca (83% BS) was strongly supported, with A. californica sister to H. cordata (72% BS), and Anemopsis—Houttuynia sister to Saururus—Gym- notheca. The sister relationship between Saururus and Gymnotheca was weakly supported (57% BS). After the morphological matrix was re-analyzed, with characters reweighted according to RC (base weight = 2), a stable topology was obtained. This also resembled previous topologies. However, the following parameters were much higher: tree length = 101, CI = 0.9703, RI = 0.94, and RC = 0.9121. Again, the monophyly of Saururus (100% BS) and Gymnotheca (95 ; relationships between Saururus and Gymnotheca % BS). and sister-group (93% BS), and between Anemopsis and Houttuynia (99% BS), were strongly supported. ANALYSIS OF THE COMBINED DATA SETS OF DNA AND MORPHOLOCY After the molecular and the morphological data were combined, there were 6505 bp in the matrix, 681 of which were variable and 261 of which were parsimony-informative. The percentage of phylo- genetic-informative sites was 4.0196. The uncor- rected sequence divergence ranged from 0.157% to 3.799% among Saururaceae sampled and from 37% to 7.479% between the outgroup and the ingroups. single most parsimonious tree of 846 steps was obtained for the combined molecular and morpho- logical data sets, and the topology of the tree cor- responded to Figure 2: CI = 0.9031, RI = 0.7492, and RC = 0.6766. The monophyly of Saururus 100% BS) and Gymnotheca (100% BS) was strong- ly supported. кой: en ч was the sister group of H. cordata (8296 BS), with Anemopsis and sister to аан and Gymnotheca. —. Houttuynia Saururus was the sister group of Gymnotheca (100% BS). ANALYSIS OF MORPHOLOGICAL CHARACTERS A phylogenetic tree was obtained when we ana- lyzed the morphological matrix using WinClada v. 0.9.99m 7.5 beta, with its topology corresponding to Figure 2. After analyzing the distribution of each character and its state on the phylogenetic tree, characters 3, 12, 14, 16, 19, 28, 30, 33, 35, 36, and 39 were realized as homoplasious, with the oth- er characters homologous in Saururaceae (Fig. 3). A "homoplasious character" means that its diverse states are due to convergent, parallel, or reverse evolution and not due to inheritance from a com- mon ancestor. Such a character still contributes to constructing the phylogenetic tree in a cladistic analysis (see Fig. 3), but it is prone to mislead if overweighted in building a phylogeny. DISCUSSION THE PHYLOGENY OF SAURURACEAE In all analyses, the monophyly of Saururus and d is resolved with high bootstrap sup- port. The combined analysis of molecular data and шы :al data strongly supports the monophy- у of Saururus and Gymnotheca, and the sister- group relationships between Anemopsis and Hout- : and tuynia, between Gymnotheca and Saururus, between the Anemopsis—Houttuynia clade and the 598 Annals of the Missouri Botanical Garden 5 7 9 1723242535 4041 43 44 46 47 21112121111112 145 1461930333638 111220293234 111221 8 133138 AAAA LAA A4 0110 0000010000 Figure 3. 136 1461930333639 1022121101 18 21 23 25 26 28 29 32 47 49 Zippelia begoniaefolla Anemopsis californica 101 1111 14 15 19 28 35 37 39 42 45 48 Houttuynia cordata 10212111111 = Gymnotheca chinensis 12 —O— Gymnotheca involucrata 1 232 td Saururus chinensis 110 —— Saururus Ceruus Distribution of morphological character states. The numbers above branches indicate characters; the numbers below branches refer to corre sponding character states. Solid black circles represent homologous characters with empty circles representing homoplasious characte Gymnotheca—Saururus clade. Similarly, strong sup- port is seen separately from analyses of chloroplast genomic data, mitochondrial genomic data, mor- phological data, and combined DNA data. Depar- ture occurs in our analysis of the nuclear genome data sets (185, ITS): Anemopsis is the sister group of all other plants of Saururaceae, with Houttuynia then sister to Saururus and Gymnotheca, rurus sister to Gymnotheca (Fig. 1). are surprising and differ from all the other phylo- genetic opinions on Saururaceae based on morpho- logical data (Wu & Wang, 1957, 1958; Okada, 1986; Lei et al., 1991; Tucker et al., 1993; Liang. 1995). Our results disagree with the phylogenetic tree of Saururaceae of Wu and Wang (1957, 1958), but partly confirm their relationships as seen by Wu (1984), who proposed that Anemopsis and Houttuy- and Sau- These results nia may be vicariant genera, and S. chinensis and 5. cernuus may be vicariant species. Vicariant gen- era and species may be interpreted as “sister groups" in a phylogenetic sense because whether they are two vicariant genera or two vicariant spe- cies, they are from an immediate common ancestor. The sister-group relationships between Anemopsis and Houttuynia, and between S. chinensis and S. cernuus are well supported in our study. Our results are not congruent with Okada (1986) and Lei et al. (1991), who considered Saururus as the basal genus of Saururaceae, and Anemopsis and Houttuynia to be derived from ап ancient Saururus. Lei et al. (1991) further suggested that Gymnotheca was most distantly derived from any ancestral Saururus, as was supported by Tseng (1982) in the Flora Rei- publicae Popularis Sinicae and by Xia and Brach (1999) in the Flora of China. In terms of the close Volume 90, Number 4 2003 Meng et al. 599 Phylogeny of Saururaceae relationship of Anemopsis and Houttuynia, our re- sults partly agree with Tucker et al. (1993), who generated a tree similar to our combined molecular tree (c ompare Fig. 2 herein and fig. 5 in Tucker et al., 3), but with low bootstrap values; they treat- ed Saururus as the first derived genus in Saurura- ceae and believed that Saururus bore many ple- siomorphies. Tucker et al. (1993) used Mit Cabomba, Chloranthus, Lactoris, outgroups of Saururaceae and Piperaceae. Accord- ing to the present understanding of angiosperm phylogeny (APG, ‚ Chloranthus, Cabomba, and Magnolia, lying too dissi from Saururaceae, may not be the best choices for outgroups of Sau- ruraceae, although Lactoris and Saruma may serve and Saruma as as outgroups of Saururaceae (Parkinson et al., 99: Graham & Olmstead, 2000; González & Ru- dall, 2001). However, Piperaceae are preferable to Lactoris and Saruma to function as the outgroup of Saururaceae (APG, 1998; Mathews & Donoghue, 1999; Qiu et al., 1999; Soltis et al., 2000). Also at issue is the interpretation of character 20 in Tucker et al. (1993: 621), whether a pair of stamens orig- inated from separate primordia or from a common primordium. The stamens of Houttuynia have been confirmed to originate from separate primordia (Tucker, 1981; Liang, acter was variably coded in Tucker et al. (1993). When we correct for this and re-analyze, using Pi- peraceae as outgroup, the topology resembles Fig- owever, this char- ure 2, and the bootstrap supports rise. SN d weighted the rhizomatous character She “separate to support the monophyly of Saururaceae. (1995: е eated "stoloniferous" initiation of bract-flower" as synapomorphies that supported the sister relationship of Gymnotheca "common primordium and * and Anemopsis, and treated initiation. of bract-flower" as a synapomorphy for Saururus and Houttuynia. However, the ontogeny of the bract-flower in Saururaceae was homoplasious. SELECTION OF THE OUTGROUP IN THE PHYLOGENETIC RECONSTRUCTION OF SAURURACEAE Hennig (1966) pointed out that a sister group is the preferred outgroup, and one of the main tasks of phylogenetic analysis is to look for these. What then is the sister group of Saururaceae? Hutchinson (1959) and Cronquist (1981) both treated Pipera- ceae, Saururaceae, and Chloranthaceae in Pipera- les. Melchior (1964) circumscribed Saururaceae, Piperaceae, Chloranthaceae, and Lactoridaceae in Piperales. For Dahlgren (1983), Thorne (1983), and Takhtajan (1987), Piperales were restricted to Sau- and although Takhtajan ruraceae Piperaceae, (1997) further distinguished Peperomiaceae from Piperaceae. Chase et al. (1993) supported the sister relationship between Piperaceae and Saururaceae in rbcL analysis, as did Hoot et al. (1999), includ- ing atpB, rbcL, and 18S. Additional support was provided by Mathews and Donoghue (1999) using duplicate phytochrome genes (PHYA and PHYC), from Qiu et al. (1999) for rbcL, atpB, 185, matR, and atpl, spanning three genomes, as well as Soltis et al. (2000) from atpB, rbcL, and 185. In conclu- sion, the sister relationship between Piperaceae and Saururaceae has been well established (Tucker et al., 1993; Wu et al., 1998; APG, 1998; Mathews & Donoghue, 1999; Qiu et al., 1999; Soltis et al., 2000). Piperaceae are the sister group of Saurura- ceae, confirmed not only by morphology but also molecular systematics, and therefore the better out- group for study. IS SAURURUS THE SISTER GROUP OF THE REST OF SAURURACEAE? Previous authors (Wu & Wang, 1957, 1958; Oka- da, 1986; Lei et al., 1991; Tucker et al., 1993; Liang, 1995) postulated that the ancestral Saurur- aceae were similar to extant Saururus in having free carpels, free stamens, and superior ovaries. Sau- rurus was considered to have the following ances- tral features: a stamen number of 6 (character 15, Appendix 1); stamens free (character 16) and hy- pogynous (character 14); carpels superior (charac- ter 19) and free (character 21); stamens and carpels free and not adnate (character 18); and placenta marginal (character 23). These morphological char- acters and others in Appendix 1 were scored fol- lowing traditional opinion about morphological character evolution. Nonetheless, our analysis here- in differs from previous phylogenetic trees based on morphology (Wu & Wang, 1957, dou! еі et al., 1991; Tucker et al. 1995), in that the Anemopsis-Houttuynia ане lies sister to the Gymnotheca—Saururus clade. We think ^^ WN w ^» the difference is due to the following three reasons at least. First, our morphological characters exceed others, in that 49 from diverse sources were con- sidered. Okada (1986) and Lei et al. (1991) mainly used chromosome numbers to reconstruct the phy- logeny of Saururaceae. Tucker et al. (1993) used 35 characters restricted to morphology and ontog- eny. Second, the method of analysis is different. Wu and Wang (1957, 1958), Okada (1986), Lei et al. 1991), and Liang (1995) did not use cladistic anal- ysis, whereas Tucker et al. (1993) and our study did. homoplasious in Saururaceae, e.g.. — Third, certain morphological characters are characters 14 Annals of the Missouri Botanical Garden (stamen position), 16 (stamen fusion), and 19 (ovary position) (Fig. 3). However, hypogynous stamens (our character 14), free stamens (our character 16), and superior ovary (our character 19) were over- weighted by previous researchers (Wu & Wang, 1957, ses, Saururus is not the first derived genus within 1958; Liang, 1995). In our included analy- Saururaceae (Figs. 1, 2). Our analysis is supported by the following documents. Tutupalli and Chaubal (1975) studied the constituents of essential oils of A. californica, H. and S. pointed out that each species in Saururaceae had cordata, cernuus. They its own characteristic essential oil type, and that Saururus was not the least specialized genus ac- cording to its chemosystematics. After comparing wood and stem anatomy of A. californica, H. cor- data, and S. cernuus, Carlquist et al. (1995) thought that Anemopsis possessed likely ancestral character states, such as relatively abundant secondary growth and tracheids. Buddell and Thieret (1997) put Anemopsis before Saururus when they described Saururaceae in the Flora of North America. Literature Cited APG (Angiosperm Phylogeny Group). 1998. An ordinal classification for the families of шк plants. Ann. Missouri Bot. Gard. 85: 531—55 Baldwin, B. G ‚ Sanderson, | м. Ропе | Wo- jciechowski, (s Campbell & M. J. Donoghue 1995 ›. The ITS region of nuclear Бш DNA: A valuable source of evidence on к rm phy logeny: Ann. Mis- souri Bot. Gard. 82: 247-25 Bayer, R. J. & J. R. Starr. 1998, Tribal phylogeny of the Asteraceae based on two non-coding chloroplast se- quences, the trnL intron and о intergenic spac- er. Апп. Missouri p Gard. t э: 242-250. Budell. ©. F, I & Тн. E. Meyer, Lizard's- al Fa amily. Pp. 37—38 in Flora of North i еі Editorial Committee, Flora of North America, Vol. ео n e and Ha- mamelidae Oxford Univ. Pre ‚ New Yor ; NS 1995. Wood and stem anatomy of Saururaceae with reference to ecology, phylogeny, ds origin of the monocotyledons. 13 a 100) 3-150. . Saururaceae Chase, M. W., D. Soltis, R. G. Olmstead, D. аа р. Н. ep в. i M. R. Duvall, R. c H. G. Hills, Y. 23 K. A. Kron, J. H. ee E. Conti, J. D. Palmer, 1 D Manhart, K. J. Sytsma, H. J. i ). Clark, M. illiams, р, A. Tadek, C. E e L. E. Eguiarte, E. hrs G. earn, Jr., 1 Barrett, S. Dayanandan & V. N genetics of seed plan quences кз ч пачи gene rbcL. Gard. 80 Cronquist, А ү aL re Integrated System of Classification of Flowering Plants. Columbia Univ. Press, New York. З. C. H. . 1993. Phylo- 1 analysis of icis. tide se- Ann. Missouri Bot. 777 5 re x G. 1983. Pp. 777-183 in R. K. Brummitt (ed- itor), 1992. Vasc es 9 Families and Genera. Roval bn Gardens Doyle, J. J. & ‚8, Doyle. 1987. А rapid DNA isolation procedure for "eps quantities of fresh leaf material. , Phytoc hem. ig : 11-15 ‚19 i rete ntion лише and the rescaled consistency sit x. Cladistics -419 Felsenstein, J. 1985. € и 'e limi ts on | phyloge nies: An approach using the bootstrap. Evolution 39: 783- 791 Feng, Y. E hn ng, K. Y. Pan & D. Y. Hong. 1998. А revalu: ‘the systematic positions of the Cerci- diphy үн 'eae ж Daphniphy llaceae based on rbcL gene sequence analysis, with reference to the relationship in the lower Hamamelidae. Acta Phytotax. Sin. 36: 411— 422. González, F. & P. Rudall. 2001. The questionable affini- ties of Lactoris: Evidence from branching pattern, inflo- rescence КУКУ? and stipule development. Amer. J. Bot. 88: 214 d ч Graham, S. W. é R. © | Olmstead: 2000. Utility of 17 chloroplast genes for po mae the phyloge Ls A the bas- al ica Ane J. po: oc 17 12-17 7. Hoot, S. B., S. Magallón & P. R. Crane. 1999, Phylogeny of basal eudicots based on three molecular data sets atpB, rbcL, and 185 nuclear ribosomal DNA sequences. Ann. Missouri bt Gard. 86: es ied. J. ! The F Кыш of Flowering Plants, ed. С lare ae Press, Oxfor Lei. LG, S.-S Zhang & Z.-Y. Yu. and the evolution of Gymnotheca. Occid. S 11: 41—46. 1991. The karyotype Acta Bot. Boreal.- Liang, H. X. 1 991. Karyomorphology of Mad а “ phyloeeny of four S in Saururaceae. Act: Yunnan. : 3: 303-307 1992. Study on 1 the pollen morphology of Sau- K ча Bot. Yunnan. 14: 401—404 ruraceae. . 1994. On the systematic significance of floral or- ganogenesis in Saururaceae. Acta Phytotax. Sin. 425-4: ae . On the па eg ш! in Sau- ruraceae. ee "a с Yunna 5-267 —— 4 S. C. Tucker. 1990, Cien floral vase iatis in Saururaceae. Amer 607—623. 1 study of the J. Bot. 77 Floral ontogeny of Zippelia be-‏ . ڪڪ e and iis бе affinity: Saururaceae or Pi-‏ »eraceae? Amer. J. Bot. 82: 681—689.‏ Donoghue. 1999, The root of angio-‏ . › rm зин пу ‚Тө rred eus duplicate phytochrome‏ .2 es. 2 тепсе 286: 947—95‏ Mele ho . 1964. Pp. SEA R. K. Brummitt (ed-‏ itor), 1 Tue E ign Families and Genera. Royal‏ Botanik s‏ Meng. S.-W. & Н. a dus 1997. Comparative embry-‏ ology on “чен eae. Acta Bot. Yunnan. 19: Оң‏ Z.-D. Chen, D-Z. Li & H.-X. Liang. 200 2. Phy-‏ ‚ logeny of ae eae based on mitoc оны mat se-‏ quence da ta. J. Pl. Res. 115: 71-7‏ Nixon, K. ( ‚ 1999, кк lada (BET/ més ver, 0.9.9, Published‏ by the author Ithaca, New York.‏ бө, Н. 19t Кышы: and relationship in‏ e genera of jc eae and Piperaceae. Bot. Mag.‏ (Tokyo) 99: 289—29‹‏ — -— = е, Volume 90, Number 4 2003 eng et al. Phylogeny of Saururaceae Parkinson, Adams & J. D. Palmer. 1999, Multigene phar dun identi: the three earliest lineages of extant gen plants. Curr. Biol. 9: 1485-1488. Qiu, Y. L., J. Lee, F. Bernasconi-Quadroni, D. E. Soltis, P. S. Soltis, E^ Zanis, E. A. Zimmer, Z.-D. Chen, V. Savolainen & M. W. Chase. 1999. The a angio- sperms: Evidence from mitochondrial, plastid and nu- clear ge p Nature 402: 404—407. pr g E., P. S. Soltis, D. L. и L. W.J Кз s. B. Hoot, J. A. Sweere, R. E A. Kron, M. W. Chase, 5. M. Seri, E. S.-M. Chaw, L. J. Gillespie, W. J. К . Sytsma. 1997. Angiosperm phylogeny inferred rin 185 ribo- somal DNA sequences. Ann. Missouri Bot. Gard. 84: 1—49. Pn i. — ^x 5 . P. S. Soltis, М. W. Chase, M. E. Mort, D. C. Albach, M. n M Savolainen, W. H. Hahn, S. B er M. F. Кау, М.А tell, S M. Prince, W. J. Kress, г J. S. Farris. 2000. Angio- sperm аи пу inferred from 185 rDNA, rbeL, atpB е псе . Bot. inn. Soc. Sudhir, .M. Swense n, L. and 2.0. Pennsylvania State University, University Park, and Arizona State University, Tempe. „ 2001. PAUP: Phylogenetic Алар! „ Massachusetts Taberlet, P., L. € y ; J. Bouvet. 1991. Uni- versal primers for amplification of three non-coding re- gions of chloroplast DNA. Pl. Molec. Biol. 17: 1105- 1109 Ising pu A. 1987. Pp. 790-799 in R. K. Brummitt (ed- itor), 1992. Vasc ular Plant Families and Genera. Hoya b eli م‎ 19 Diversity and Classific ation of Flowering Plants. Combi Univ. Press, New York. sow J. D., T. J. Gibson, F. Инак F. ги & D. x es 1997. The CLUSTAL-X windows in- terface: кн xible strategies for multiple sequence ig me y aided by quality analysis tools. Nucl. 25: 4876—4882. Thorne, R. F. 1983. Pp. 770-776 in R. K. Brummitt (ed- itor), 1992. Vascular Plant Families and Genera. Royal Botanic Gardens, Kew. Tseng, Y.-C. 1982. Saururaceae. Pp. 4—11 in Y.-C. Tseng, T.-C. Chen, K.-F. Wu, P.-S. Chen & P.-Z. Zhu о, Flora Reipublicae Aa Sinicae. Tomus 20(1). S ence Press, Acids Beijing Tucker, S. C. 1975. Floral development in Saururus cern- uus l. Floral initiation and stamen development. Amer. J. Bot. 62: 289-301. 1980. сине and flower development in omia. Amer. J. Bot. 67: 686— the Piperüceat. I. Pepe 70: . 1981. Inflorescence and floral development in Houttuynia cordata (Saururaceae). Amer. J. Bot. 68: 1-1032 1982a. Inflorescence and flower development in the реа eae. П. Inflorescence development of Piper. Amer. 1 69: 743—752. 982b. Inflorescence and flower development in the B eae. Ш. е ‘ence development of Piper. камыт s 69: 1389 E 5. Initiation on n of inflorescence and pias er in ape sis californica (Saururaceae). Amer. "i ж 31. u .W dtm 1996. Floral structure, devel- ата апа i of paleoherbs: Saruma, Ca- mba, Lactoris, and s ed Piperales. Pp. 141-175 in D. W. Taylor & I - Hic ‘key (editors), Flowering Plant n Evolution & Phylogeny. Chapman & Hall, New ork. : H.-X. Liang. 1993. Utility of ontoge- nelic ‘ond conventional characters in determining phy- logenetic relationships of Saururaceae and Piperaceae (Piperales). Syst. Bot. 18: 614—641. Tutupalli, L. V. & M. G. Chaubal. 1975. Saururaceae Composition of essential oil from foliage of Houttuynia cordata and chemosystematics of Saururaceae. Lloy: 38: 9: pm White ‚ S. Lee & J. Taylor. 1990. Ampli- fication п аай вл ѕедиепс n Ч fungal ribosomal RN A genes for раен Рр. inis, D. H. Gelfar Lap Байан y T. . White (editors), PCR Protoc pi A (€ to Methods and Applications. Ас AE M ag San Diego. Wu, Z.-Y. . An Outline of Phytogeography (printed matter). 2. E iety of Botanists in Yunnan Province 1: 44-45. Wang. 1957. A preliminary study on tropical Ме subtropic sal flora in Yunnan I. Acta totax. Sin. 6 254. & Phy- 1958. Some corrections on the paper a ори study on tropical and subtropical flora in Yunnan 1. Acta Phytotax. Sin. 7: 193— = VAC. Tang, A.-M. Lu & Z.-D. 1998. On primary subdivisions of the Мормона Towards a \ classification of the angiosperms. Acta Phytotax. Sin. ch 358-402. ia N.-H. & A. S Brach. 1999, Saururaceae. Pp. 108- 109 in Zh.-Y. Wu & Raven (co-chairs of he ed- itorial ea Flora of China, Vol. 4. Science Press, Beijing, and Missouri Botanical Garden Press, St. Lou- is. APPENDIX | MORPHOLOGICAL CHARACTERS AND THEIR CHARACTER STATE CODES. Vegetative 1. Stem: erect (0), stolon (1), short stem with one node (2). . Te ead leaf of stem in reproductive period: green (0), white (1). о. on lamina: none (0), restricted to under- * aide (1), on both sides (2). 4. Leaf venation: pinnate (0), palmate (1). 5. Secondary venation: none (0), dichotomous (1), not dichotomous (2). 6. Areoles: incomplete (0), incomplete or imperfect (1). imperfect or perfect (2). Stem Anatomy . Number of stem vascular cylinders: 1 (0), 2 (2). Fiber in stem: disc 'ontinuous (0), continuous d 2 oration plate ty (0). simple (1). Floral Morphology со — — — pe in vessel members: scalariform 10. Floral symmetry: radial (0), dorsiventral or zygomor- phic (1 esl regular flower: none (0), present (1). Ti Color of inflorescence involucrum: green (0). green, showy (1). — not 602 Annals of the Missouri Botanical Garden 13. Flower-bract stalk: absent (0), present (1). 14. Stamen hypogynous (0), perigynous (1), epigynous (2). 5. Number of stamens: 6 (0 positio n 15 J: 16. Stamen fusion: free (0), connate 17. Anther dehiscence: stom alone entire length of anther (0), predominantly i in рш] position (1), in distal position (: 18. Adnation of stamens and carpels: free (0), partial fu- sion 19. Ovary position; superior (0), perigynous (1), inferior (2). 20. Number of carpels: 4 (0), : .1(2). 2]. Carpel adnation: free (0), full nde (1), single car- pel (2). 22. Style none (0), prese n 23. Placenta: marginal (0), pa pen (1), sal (2). 24. ta 's per carpel: greater than or in to 3 (0), less than 1 (1). presence: Floral Anatomy 25. Number of carpel vascular bundles: 2 (0), coadnate (1), 1 (2). 26. Vascular bundle fusion of stamens and carpels: free (0), partial fusion 27. Fusion of adaxial and abaxial carpel bundle: free (О), jartial fusion 28. raid asis of bract-flower: discrete bract and flower ini- tiation (0), common primordial initiation (1). 29. ene sis order of carpels: middle primordium first (0), vilateral primordium first (1), simultaneous appear- ance or single or common primordium (å 30. Genesis of stamens: discrete primordium (O common primordium ( ;enesis ordering of stamens: bilateral stamens first (0), middle stamens first (1). — — - — S^ Appendix 2. 32. Genesis pattern of median sagittal stamens: in pair (0), adaxial axis first (1), no adaxial or abaxial stamen (2). 33. Genesis pattern of bilateral stamen pair: discrete pri- mordium (0). common primordium 34. Median sagittal carpels: adaxial and abaxial carpels (0), adaxial only ü Pollen 35. Germinal aperture: anasulcate (0), anasulcate and an- atric D 'ate (1), inaperturate (2). 36. Small verruculae at the edge of foveolae of pollen tectum: absent (0), present (1), narrow bell of granule EI. 37. Microspore ла simultaneous (0), successive (1). 38. Type of minor tetrad: bilateral symmetry, T shape and * shape (0), Мше ral symmetry and + shape (1), T bilateral symmetr 39. Pollen abortion: Fn (0), present (1). Embryology 40. Layers of ovule integument: two (0), outer layer pre- sent but ошай (1), only inner layer present (2). 41. Micropyle: both inner and outer integuments (0), in- ner сл only ( 12. Nucellus: сгаѕѕіпис eê ' (0). tenuinucellate (1). 1 micropylar or chalazal — T t n present (1). 46. Perisperm: cellular type (0), nuclear tvpe (1). 47. Fruit type: capsule (1), berry (2). Cytology 48. Ploidy: diploid (0), polyploid (1 49. Base chromosome number: 11 (( y». not 11 (1). — The matrix of coded morphological characters. 1111111111222222222233333333334444444444. Taxon/Characters Zippelia begoniaefolia Anemopsis californica Houttuynia cordata 0021211111000000110010212110100100110201 101101201 20200707001 10101011111101110210211000100000000101 1 0011100100110210012111101111200201111210010010111 1 Saururus chinensis Saururus cernuus Gymnotheca chinensis Gymnotheca involucrata O1111 002111000100100000000100001 10010000 1 0000000000000 1001120001001201012011101110111110000010000000101 1001120001011201012011101110111110000010000000101 PHYLOGENETIC POSITION AND GENERIC LIMITS OF ARABIDOPSIS (BRASSICACEAE) BASED ON SEQUENCES OF NUCLEAR RIBOSOMAL DNA! Steve L. O’Kane, Jr.2 and Ihsan A. Al-Shehbaz? ABSTRACT The primary goals of s study were to assess the generic limits and monophyly of Arabidopsis and to investigate its relationships to related ta ITS-2) of nuclear boca DNA, including 5.88 rDN pene trees. An attempt was made to include all spe Hylandra as W айарга, Preudoarabidopsis and /anhea Key ш abidopsis. a Pa a in the family Brassicaceae. Sequences of the internal transcribed spacer region (ITS-1 and A, were used in maximum parsimony analyses to construct ecies currently or recently included in Arabidopsis, as well es suggested to be c ‘lose relatives. Our ipid show that. Arabidopsis, as trad alis recognized, 15 polyphyletic. (1) now includes species previously placed in species of iiie and oy excludes species now placed in Crucihimalaya, Beringia, Olimar- Cardaminopsis and Arabidopsis, Arabis, E Brassicaceae, Crucihimalaya, ITS phylogeny, Olimarabidopsis, Pseudoar- Arabidopsis thaliana (L.) Heynh. was first rec- ommended as a model plant for experimental ge- netics over a half century ago (Laibach, 1943). In recent years, many biologists worldwide have fo- cused their research on this plant. As indicated by Patrusky (1991), the widespread acceptance of A. thaliana as a model organism is attributed to the discovery that it has one of the smallest genomes of any flowering plant, a low chromosome number (n = 5), and that its genome contains few repetitive sequences and little intergenic spacer DNA. A s prising recent finding by Blanc et al. (2000), how- ever, showed that although A. thaliana has a re- markably small genome, much o DNA is present in more than one copy. In addition to these important attributes, A. thaliana has a short gen- eration time (four to six weeks), a small size (dozens can be grown in a small pot), and can easily be grown on synthetic media (Meyerowitz, 1989; Mey- erowitz & Pruitt, 1985). The species has been used extensively in developmental, evolutionary. and ge- the netic studies and has played a major role in un- derstanding the various biological processes in higher plants (see references in Somerville & Mey- erowitz, 2002). The intraspecific phylogeny of A. thaliana has been examined by Vander Zwan et al. 2000). Despite the acceptance of A. thaliana as a model organism and the sequencing and mapping of its nuclear genome (The Arabidopsis Genome Initiative, 2000; Cooke et al., 1996), little is known about the other species of Arabidopsis sensu lato, — and their closest relatives. A small number of molecular phylogenetic stud- ies have included a few members of Arabidopsis Hey sensu lato (Price et al., 1994; O’Kane et al.. 1996; Galloway et al., 1998; Koch et al., 1999, 20 2001; Yang et al., 1999). However, none of these studies attempted to examine all of the taxa either currently or previously included in the ge- nus, and they included only a small number of oth- er, sometimes distantly related, genera. The last comprehensive taxonomic account (Schulz, 1924). ! Research. and We were support National Geographic Society (NGS-5068-93), and 'hose lab n whose this work wa ed by grants from the National Science Foundation the University of Northern Iowa. Special thanks s initiated, and to the support of the Missouri Botanical Garden. We offer gratitude to the (DEB-9208433), the to Barbara Schaal, i many hosts, field companions, and herbarium curators who aided in this study, especially Abdulla Abbas, Nogman Aralbaev, Isa O. Baitulin, DM Berkuter ^i. Hang, Josef Holub, Kat Zbigniew Mirek, Klaus ште Noriaki Murakami, Railay and an anonymous re iko, Ram С haudhary, Gheorghe Diho e late Sigizmund Kharkevich, oru, Vladimir Dorofeyev, Yang Guang, Franta Krahulec, Hanna Kuciel, Karol Marhold, Nonna Pavlova, and Boris Syomkin. Reviews by Donovan y ewer improved this paper substantially. ? Department of Biology, жшше of Northern lowa, Cedar Falls, lowa 50614- Eo U.S.A. steve.okane@uni.edu. 3 Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S A. ihsan.al-shehbaz@mobot.org. ANN. MISSOURI Bor. GARD. 90: 603—612. 2003. 604 Annals of the Missouri Botanical Garden which recognized 11 species, is unsatisfactory. As many as 50 species have been placed in the genus, and, although many of these are now placed in oth- 1999), their phylo- genetic relationships remain unresolved. Monophy- er genera (Al-Shehbaz et al., ly of the genus has not yet been critically determined, and even basic biological information, such as chromosome numbers, generation time, and breeding system of the members of the genus, is acki Generic delimitation is perhaps one of the most difficult and frequently encountered problems in the systematics of the Brassicaceae (Al-Shehbaz, 1973; Rollins, 1993), and Arabidopsis clearly dem- aas been a lack of onstrates this problem. There agreement among taxonomists on the number of species that belong to Arabidopsis and on the char- acters that indicate its generic boundaries (e.¢., Ball, 1993; Lóve, 1961; Rollins, 1993). The generic limits of Arabidopsis have been highly unnatural, and there were no well-defined characters separat- ing it from several closely associated genera (but see our taxonomic revision based on the results of this current work, O’Kane & Al-Shehbaz, 1997; Al- Shehbaz & O'Kane, 2002a). Some individual Ara- bidopsis species have been transferred among sev- eral other genera. An example is A. thaliana, which on the basis of Schulz's (1924) synonymy was pre- viously placed in at least nine other genera, in- cluding Arabis L., Conringia Adans., Crucifera К. Н. L. Krause, Erysimum L., Hesperis L., Nasturtium R. Br., ophragma Čela Pilosella Kostel., Sisymbrium L., and Sten- rabidopsis has been closely associated with мМ three different genera, Cardaminopsis (С. А. Mey. Hayek, Arabis L., and Halimolobos Tausch. Schulz (1924, 1936) considered its nearest relative to be Halimolobos, and separated the latter аз being coarser herbs with the styles much narrower than the fruit, as opposed to Arabidopsis, which were seen as slender herbs with the styles slightly nar- rower than the fruit. These alleged differences are not mutually exclusive, and species recognized by him in one genus can easily be accommodated in the other. Lóve (1961) and Hylander (1957) indi- cated a relationship with Cardaminopsis based on natural interspecific hybridization. Hedge (1968) suggested a closer relationship between Arabidopsis and Arabis and indicated that the two differ only in the cotyledonary position. He further suggested that Arabidopsis wallichii (Hook. f. & Thoms.) Busch probably represents the link between the two gen- era. An (1987) and Jafri’s (1973) transfer of several species from Arabis to Arabidopsis was probably in- fluenced by Hedge’s view. Molecular-based results (O’Kane et al., 1996 1995; Mummenhoff & Hurka, 1994) (1961) and Hylander's (1957) hy- pothesis in showing A. thaliana to be most closely Kamm et al., agree with Lóve's related to species placed in Cardaminopsis. In an- ticipation of results published here and to make the names available for floristic works in progress, we previously published the needed nomenclatural in- novations for the genus Arabidopsis (O’Kane & AI- Shehbaz, genera to accommodate excluded species (Al-Sheh- 1999). In brief, Arabidopsis includes only 1997) and have established several new baz et al., => thaliana and species previously included, or suggested to be, in Cardaminopsis (Jones & Ake- royd, 1993a, 1993b). Species now excluded from Arabidopsis are placed in Thellungiella О. E. Schulz (Al-Shehbaz & O'Kane, 1995), Neotorularia Hedge & J. Léonard (Al-Shehbaz & O'Kane, 1997 lanhedgea Al-Shehbaz & O'Kane (Al-Shehbaz & j )9), Crucihimalaya Al-Shehbaz et al., Olimarabidopsis Al-Shehbaz et al., and Pseudoar- abidopsis Al-Shehbaz et al. (Al-Shehbaz et al., 1999), and Beringia Price et al. (Price et al., 2001). Our primary objectives are to determine the ge- — n O' Kane, neric limits of a morphologically coherent, mono- phyletic Arabidopsis and to reconstruct a robust in- terpretation of its phylogenetic neighborhood. A well-corroborated phylogeny of the group will allow better evolutionary interpretations to be made of the massive amounts of data now accumulating for A. thaliana. Workers will know which species to com- pare to A, thaliana when making interpretations of evolutionary processes. Furthermore, these initial steps will provide a better understanding of mor- phological character evolution in the Brassicaceae, a family of great economic importance fraught with taxonomic problems related to an under-developed understanding of character evolution and generic delimitation. MATERIALS AND METHODS TAXON SAMPLING We included representatives of all taxa (at least at the generic level) that are now or have been in- cluded in —— (e.g., Schulz, 1924; Hedge, 1965; Jafri, 1973; Al- Shehbaz, 1988; Ball, 1993). Taxa shown to o near Arabidopsis in other molec- ular studies have also been included (Price et al., 994; O'Kane et al., 1996; Galloway et al., 1998), as have a sampling of taxa from elsewhere in the Brassicaceae. Phylogenetic trees were initially root- ed by Cleome lutea Hook. of the Cleomaceae, a family basal to the Brassicaceae (Rodman et al., 1993; Judd et al., 1994; Hall et al., 2002). Included Volume 90, Number 4 2003 O'Kane & Al-Shehbaz Phylogenetic Position of Arabidopsis 605 taxa, as well as voucher information and some no- menclatural comments, are given in Table 1. Where possible, plant materials were collected in the field and dried in powdered silica gel. In some cases tissue was obtained from plants grown from seeds. Where fresh or dried material was not available, we used tissue from herbarium specimens; the se- quence for Arabis scabra All. was obtained from GenBan DNA EXTRACTION, SEQUENCING PCR AMPLIFICATION, AND Total DNA was extracted from dried tissue ground in a pinch of sterile sand by a modified CTAB procedure as previously described (O’ Kane 1996). scribed spacer region (including ITS-1, 5.95. and et al., Amplification of the internal tran- ITS-2) was done using the conditions given in O'Kane et al. (1996) except that some ITS regions were amplified as a single unit using primer ITS1— 185 (5' CGTAACAAGGTTTCCGTAGG 3’) and ITS-4 (White et al., 1990) rather than as two over- lapping pieces. PCR products were purified from 0.89€ agarose gels containing 1X TAE using Wizard PCR Preps (Promega). Sequences were obtained ei- ther manually using the fmol® DNA Sequencing System (Promega) or from the automated sequencer at the University of lowa using the same primers as were used to amplify the product. GenBank ac- cession numbers are given in Table 1. Sequences of the allotetraploid Arabidopsis suecica (Fries) Norrl. were obtained from cloned PCR products as previously reported (O’Kane et al., ) SEQUENCE ANALYSES ALIGNMENT AND PHYLOGENETIC Sequences were aligned with the computer program MALIGN 2.7 (Wheeler & „Саве; available at «ftp: //ftp amnh / / li g >) us- ing the bioa ende ally determined parameters: ; Meca 5 e: З. trailing 3, matrix 0 J3 2231 keepa2, score 2. The matrix parameters internal 7 233 2 3 0, аѕрг, spr, quick. weight transversions as 3, transitions as 2, intitial gaps E . gap extensions as 5, and initial and ending gaps as 3 (not a factor in our sequences). Most-parsimo- nious trees were found using PAUP* 4.0b4 (Swofford, 2000). In our analyses all characters were considered to be of equal weight and gaps were coded as missing data. Two hundred and fifty replicates of random ad- dition using Fitch parsimony were performed using Tree Bisection Reconnection (TBR), Mulpars, multi- state = polymorphism, gaps coded as missing, and Collapse branches if maximum length is zero. Boot- strap support was obtained from 500 replicates using a single round of simple taxon addition. Decay values 1988; Donoghue et al., 1992) were found using the Decay program 4.01 of Eriksson 1998). Clade Significance (Lee, 2000) was imple- mented in PAUP by the AutoCladeS program (T. Er- iksson: available at ). We found that this new measure of support indicates those clades that have (Bremer, —. the highest support based on the other two measures. The information content of the data was assessed by the gl statistic (Hillis & Huelsenbeck, 1992) based on 100,000 random trees and by the Permutation Tail Probability (PTP) (Faith & Cranston, 1992) based on 200 heuristic searches of randomized data (PAU parameters as above except simple addition was used rather than random addition). RESULTS In nearly all samples there was no evidence of heterogeneity among individual ITS copies. Rarely, two different bases were present at a given position as indicated by two bands on an autoradiograph or as two clear peaks on a chromatogram. In these cases the base position was coded using the appro- priate ambiguity code. We interpret this rare “het- erozygosity" as incomplete homogenization of the ITS copies and not as evidence of hybridization; an individual sequence would show much more vari- ation if hybridization were involved. Although ini- tial analyses used Cleome lutea as the outgroup, inclusion of this taxon added to the complexity of sequence gaps and to areas with ambiguous align- ments. Of the taxa included in this study, Berter- oella maximowiczii (Palib.) O. E. Schulz was found to be strongly supported as the basal-most taxon. Further analyses, then, used B. maximowiczii as the outgroup. The resulting multiple alignment of the internal transcribed spacer region (ITS) was 716 base-pairs in length. Sequences are deposited in GenBank (accession numbers in Table 1), and the full alignment is available from the first author. Two regions of the alignment that were extremely sen- sitive to alignment parameters and could not be improved by eye (114—139 and 461—507) were not used in the phylogenetic analyses. In all, 384 bases were invariant, 79 were parsimony uninformative, and 253 were parsimony informative. Parsimony searches yielded 24 distinct. most- parsimonious trees of length 951, consistency index jam CI) 0.48, consistency index excluding uninforma- tive characters (CIU) 0.43, retention index 0.73, and rescaled consistency index (RC) of 0.35. The gl statistic of the data was —0.5982, which Annals of the 606 Missouri Botanical Garden €SSLETAV eun) (OW) FS£6 7»quous-1V 99uq "V H Y әче, о) "zequoqs-y. (Mpy) Dour pépppumpon4)) 88SLE1TAV VSA wenn (OLSD £22£ ?uvy,o "xooH nem] әшоә) OLSLETAV VSN *unossry (OW) 1076 2UPY.O X 2294246-1 зрә (1) suorpd-psang vjjasdv;) PLOLELAV BIuUBWOY (OW) 96St n4oww] Y 2uDN () "Od хә “ZIpuy Dd4n230421u1 Du1]2auimm7) BLSLETAV ‘VSN `орелоол) (SO) rZI£ KJ9oN иоѕтеду "S (uospreqorg) 2/7290] DADIg €LSLETAV вищо (H9) 802 “%4 3p 0J z[uos ^q 70 (qed) nmoraounmpur vjjao42112g LSSLEIAV erssny (AD £S ujoqy pijofisanq ‘dsqns auey,() $ zequaus-[y ‘Pud ^v "y CO) vyofisanq mung 0€986X YUE fU234) шолу "HW D4QDOS Siqnap CLG2ELAV ise] Jey uerssny (OW) Z89€ oquem4og эр 2uvy () "] pmpuod догу Z9GLETAV ‘VSN ‘еиеиоур (OW) G29 гиру, 0) чом "Т9 ngpmu siqpay SLELAV ‘VSN ‘еиешор (OW) £29£ ?upy,() чоюу "S 11110۸] siqpay O9SLETAV uedef (OW) 689€ omy X әиру () “buy охорә?оу siqnay CLSLELAV VSN ‘еиешоу (OW) 9295 әиру. 0) поѕте № "S приошштар siqpiy 6SSLETAV erueuloy (OW) 8I9£ n4oqi X 2uny, 0) 7] pudyo siqpay Vecevn ərqndəy 42929 (OW) 9£9£ »opnp4y 3p 2uvy () (z a[dues) Sccevn uejsmeuapy (OW) FESZI Чэә[рода (Т әјашеѕ) ‘цихәң (7]) риту `y c9grcen pue[ur J] (Н) 0972 uauidum] “ION, (C14) Dans сү 98g1cen PIYBAOTS (VUE) (226D "ws раоўәлилоң ZEQUPYS-TV ў LYO (NYS) 729/22u сү 692961 seq ej ueissny (VTA) 08-8-2 %әрәЈәс̧ 3p олоо] 02796A 1S4 1F | ueIssny (VTA) 68-8-9 аә]әргә J » 40]p3408 6ESLELAV 1Se Те uerssny WTA) £107-66 NL 19 1[214944DWVN zequaus- IV » ouey. О (" 1) Dn3nD4]ad ‘dsqns zeqyueys- [V » JURY () ("D DIDIA] ү 292961 ise eq uerssny (ОЙ) 189g оўиәтзләд 3p 2uy O 897961 uedef (OW) 789 ?upy, () 99796A sey Tey ueIssny (OW) 629€ oquoqm*428] MW 9uDN 0 Zequaus-ry by 3UEYN.() Cod хә `цәв1]) DITIDYIUDY ‘dsqns zequ2us-|y » ЭПР () CT) DIDIA] "Y. OPSLELAV BIUBUIOY] (OIN) [I9E тоу] 29 2UP¥.0 zequaus-[V ў әчеу.0) (иәдпм) sz5u24120 *dsqns zequaus-[y y 3U€N () CT) HPY сү [PSLETAV puejod (OW) 221 leroy 1127101 “dsqns zequoqs-y X әче, (у (71) Hapy `y EÞSLETAV sey Jey uerssny (ОИ) €89E naomqoaq $ 2upy,() ZPSLELAV sey Jey uerssny OW) 929g Oyuamyiog 3 әиру, 0) PrSLELAV uede f AOW) £69£ ?upy,() zequauS-pV "y 9uex, o Cumsiejy) Diafnuwas ‘dsqns zequous-|y X әче) (7]) мәр сү ЧАЧКУ етле[ѕоЗпд (M) (2261) “u's мәшшпу zeqqueus-V P PUPY.C (HOU) отоло "y SUSLETAV aourl | (SONA( (0681) “U's 2807) zequaus-[V X 9uey,Q (Эа) sisuauuaqas `y ££cev pue[oq (OW) £998 ]o1ny x 2upy () L8IZSN erssny (qT) 62 IP 1 aajaazy 99grcen pu?[oq (OW) 0€9£ 3221 X 2uny,() 391[eEMe'] (7T) nsoua1n sisdopiqpay 19quinw Áp[poo] IJMO uonewozut IƏJYƏNOA uoxe] Jo wnueqiəy pue ләчшпи 10]29][05 "10]29[|0» SE пәлїЗ 51 uorjeuLriojur 19q5noA "pnis si ur p»pn[our səjdwes 10} sSJoquinu UOISS32)E gugequa«) pue UOT]EULIOJUI JIYONOA 'soouanbas WN [1 [еәпиәрг цим sop[duies ajorpur sjduosiodns jeoquapt ‘ısodap `1 FL 607 O'Kane & Al-Shehbaz Volume 90, Number 4 2003 Phylogenetic Position of Arabidopsis POSLELAV ppeue’) (OW) 4£64 шоцӯшрәт z[nuos ^4 О Сва) Vaulsnspos vjjoizunjjay | 6LSLETAV Хәзип, (ИЯ) 22981 sPpoq X suvg 9uPy,0 $ zequous-py. (бүчәлцәс) pad ojjoisunjjou | €9SLETAV ивент (ЯП) 2611 1” 1? aojang z[nuos “AO (AON су 70) Рптуйоуру vjjoidunjjon | 06S2ELAV ODIX9[N “иоә”ү o^9nN (OIN) £685 uo; 29 Шоу Ap1c) ^V. 125242 sisdipiju&g 68SLETAV odIXƏ W *e[rnueo-) (OW) 6872 “Gey op 8011104 SUT]JOY (Био) umgp1adouopui unuDp1D201204dÇ I8SLETAV VSN "euejuopy (OW) €29£ әиру, 0) AON "V 70 CPIM хә ueudois) ритә4урә ртуѕтојәшс 8SSLETAV UBSYEZE (OW) (2861) “U's 108140145 son ^V “YP WYO "zequous--y. ('qerg ^|) ру} уЧохо] sisdopiqpavopnosq Y8SLETAV VSN ‘орезо[оз) (OLSI) 6£2£ uoszopuy эр әиру, 0) zequaus-[V $ PUEYO (әчәэл‹)) nsoumad vuvsíyd €8SlELAV V'S N ‘виешоу (DISD F6ZE 2UPY.O ndapoowupip чел Аел) ^y. (oog) ndmooouitpip тиоѕца Z8SLELAV уп “Че (OLSI) 902 ?uvx.Q ‘qPAY oyofiman vuvséug S8SLETAV ‘VSN 'eesseuue], (OLSI) 2687 чошдәу zequaus-[y $ әиеҹ.0) (SUNJOY) szsuauois DiuosAD f 98SLETAV ‘VS ‘әәѕѕәииәј, (OLSI) #682 чә 3 иршдәу zequauS-|V $ 2uex,o (Suoy) pjidisuap ruos&ng LPSLETAV uvjsiueuajy (SA) 6SE8T 4ogunjoay aou "V “YP PUYO "zequaus-[v. (PA з "uosiog) vsouquin srsdopiqpapun]o 6bSLETAV ueystueysyjy (3) OS98M 091әриәд P PPIH aoud ‘V М $P 9uex,Q ‘zeqyeys-[y (ueudois) рутшта sisdopiqpapun) SPSLELAV UP1SZÁGIÁN (ч 1) "ws “7D Jo doqgn4sy әд "V "M suey.O "Zequaus-y ('suioQq T, » 3 чоон) DIINGDI sisdop1qD.iDUtjQ) 9/C€/€T4V рив[од (OW) 626 1022904 x vysmozg n1ppmorupd “dsqns ‘asaq (7]) рторпотира vysəy L9S1€14V "VAS `оотхәҢ MON (915) 22 42«amng 3133159) (UOS `S) nifofunaui] DIUALASIIO\] ILSLETAV uel] (M) РЕСОР 192unjooy pieuo2T ^f y» әврәң (JS9(]) 007710) n10gm40102A 9967 £L1V eut) (4d) "ws Zury-ovus preuoyT ‘f ў 93pog (AON ^V 7) suny ptp]na0]02N] SOSLELAV uejstueusgy (4) 62221 '[?2]pod oueN.() y zequaug-[y (espey) »jpdasouns niD]n40102] L6SLELAV оотхәр (OLS) 66£9 ПӘН мең $ “ооң 1427/02 nd4204] 89SLE1AV UBISTURYFY (М) [PSE mosvy әйт, $ zequeqs--y. (SWOYJ, X 7 00H) n40ffimunu. vaspayuny LOSLETAV ‘VSN “вреләм (OW) 6FSF I» 12 рә/злорү supo (жип) magav чел Аел) "y. vsnffip sngojour]og 6967 £ 14V оотхәң (OW) 196s puvo SUI[OY 290799 чел z[nuos ^3 ‘O (SWH) мәшура sogojounog O8S2€1AV ‘VSN *opeio[o?) (SD) 2612 әиоу.0) surpoy прируиәа ошәгтя LLSLETAV пец (OW) FISES 402unjooy] jdeis ('sstog xo "Zuejog) vpnu sisdoqpaqq C6SLELAV оогхә (OfS) [09 ПӘН “MBH ротор poing C6SLELAV “WS ‘Oorxeyy MƏN (DfS) Z9pp 42140d шо (шүәдиху) 1142217512 ndapootudaoun(q POSLELAV ueser (ЯТ) £Zr6 paepqmzuox эр yourapyny’) ^ossnun[ $ "zoA() (Aodog) smunssipiqns smjupispiouq SSSLETAV uejsmieuapy (У) 99 uq оца "y "M P әчеу,() "zequaus-py. (вшоцу, X JF YOH) nym рАрроштуотат) 19959617 uejsppeq (M) SS90£ 4o2unjooy aoud гү М 5 ery. "zequeus--y (‘ssoquie)) 091118 DÅD PUNIMI) ISSLETAV ues tyren, (qT) "ws unoumy aud "V "M $P 9?uex,o "zequeus-py. ('uosiog) nraoqruumoao np]pumpon4:) ZSSLETAV uejsmeu2jy (Я) 2796 919913 X oqjopuom aud "V М P 9uex,o "zequeus-py. (Хәр ^V 72) purssiour оАрроштутоти) 9SSLETAV eun) (OW) £££6 7qY2Y4S-1V aud “WY P әчеу,() "zequeus-py. (suroq] X 7j ^xoo) 20:020180 nApnppowun1n47) OSSLETAV (теш) 1d 42] (S) 259 4asivy eon "V H 29 әчеу,0) "zequaus-y (шило) мәұопәиу néppumpon4;) Joaquin \| Áji[eooq әолпос uone wozu J13q]2n0A uoxe, `рэпчцио? 19198], 608 Annals of the Missouri Botanical Garden indicates strong phylogenetic signal in the data (P < 0.01). The Permutation Tail Probability (РТР also indicated strong signal (Р = 0.005). Figure 1 shows the strict consensus tree of the 24 most par- М7 simonious trees. DISCUSSION RELATIONSHIPS AND CIRCUMSCRIPTION OF ARABIDOPSIS The relationships among the species included in this study are almost entirely consistent with results previously published for smaller taxon samples in а (e.g.. m Koch et а 99). Like the Brassicaceae focusing on Price et al., 1994; Galloway et al., al., 1999, 2000, 2001; Yang et a those studies, our research indicates that Arabidop- sis as traditionally circumscribed is a highly artifi- cial group. In fact, even the tribe Sisymbrieae, the traditional placement for Arabidopsis (Schulz, 1924, 1936; Al-Shehbaz, 1984, 1988), is itself. artificial. The confused circumscription of Arabidopsis, as based on morphological grounds, was noted in pre- 7: 602), for example, recognized that if Cardaminopsis and vious taxonomic treatments. Hylander (195 Arabidopsis are combined, as seemed likely, the limits of Arabidopsis “would thereby be consider- ably widened—or, perhaps more correctly, drawn in quite another way.” Jones (1964) also indicated that two species of Arabis, A. pedemontana Boiss. and A. cebennensis DC., might best be included in Cardaminopsis. Thus, at least as early as 1964, tax- onomic problems were anticipated in Arabidopsis. Cardaminopsis, and Arabis. esults from our study are sufficient to allow a revision of the taxonomy of the genus Arabidopsis. А strongly supported clade (bootstrap support 97%, decay index 6, clade support 0.014; see Fig. 1) containing А. thaliana, the type species of the ge- nus, defines the limits of a recireumscribed genus Arabidopsis. As stated above, we have anticipated the publication of these results by redefining the circumscription of Arabidopsis (O’Kane & Al-Sheh- baz, 1997), transferring species to previously rec- 1995, 1997), and lastly by erecting several new genera ognized genera (Al-Shehbaz & O'Kane, for species previously included in Arabidopsis (Al- Shehbaz et al., 1999; Al-Shehbaz & O’Kane, 1999, 2002a; Price et al., 2001). Names in bold type in Figure | represent new genera erected for species previously included in Arabidopsis by other authors, and the bold letter “A” indicates species previously included Arabidopsis. In every сазе we have made genera monophyletic (sensu. Hennig, 1966; holophyletic of Ashlock, 1971). Species previously included in Arabidopsis are now placed in Beringia, Cructhimalaya, Olimarabidopsis, Pseudoarabidop- sis, lanhedgea, Neotorularia, and Thellungiella. Ar- abidopsis sensu novo is distinguished from other genera in the Brassicaceae by having short petio- late but not auriculate or amplexicaul cauline leaves, the presence of simple trichomes, these of- ten mixed with few-forked ones but not stellate hairs, well-defined basal rosettes at least in young plants, white to lavender (rarely almost purple) but never yellow flowers, erect to slightly ascending поп-ѕассаіе or slightly saccate inner sepals, si- liques at least slightly torulose, much longer than they are wide and glabrous, compressed (rarely subterete to terete), seeds uniseriate in the silique, and cotyledons accumbent or rarely incumbent O'Kane & Al-Shehbaz, 1997). Habit ranges from annual to short- or long-lived perennials. Chromo- — some numbers vary from x = n = 5 in A. thaliana lo x = in the remaining species except for A. suecica, which is an allotetraploid (2л = 26) de- rived from A. thaliana (2n = 10) and А. arenosa 16) (Mummenhoff & Hurka, 1996, and references therein). (L.) Lawalrée (2n. = 1994; O'Kane et al., Keys to the species and subspecies are given in O'Kane and Al-Shehbaz (1997). As circumscribed here, the genera Cardaminopsis and Hylandra A, буе are united with Arabidopsis. Arabidopsis is a Dæ a monophyletic genus consisting of A. arenosa, A. « bennensis (DC.) O'Kane & Al-Shehbaz, A. croatica Schott) O’Kane & Al-Shehbaz, A. halleri (L.) O'Kane & Al-Shehbaz, A. lyrata (L.) O’Kane & Al- Shehbaz, A. neglecta (Schultes) O'Kane & Al-Sheh- baz, A. suecica, and A. thaliana. Although we have Boiss.) O Kane & Al-Shehbaz, it, too, clearly belongs in Arabidopsis — — no sequences of A. pedemontana based on its morphological relationship to A. ceben- nensis. All other species previously included in Arabi- dopsis are more distantly related, especially those Himalayan species now included in Crucihimalaya and the Middle Morphologically, Cructhimalaya differs Eastern and central Asian Olimar- abidopsis. from Arabidopsis in that it has at least some stellate trichomes, whereas Arabidopsis has forked tri- chomes. Olimarabidopsis differs from Arabidopsis in its yellow, rather than white or lavender petals, pu- bescent, rather than glabrous fruits, and auriculate rather than attenuate or petiolate cauline leaves. Al-Shehbaz et al. all species formerly placed in Arabidopsis. Based (1999) presented an analysis of on the results presented here, Arabidopsis is а ge- nus of circumboreal and circum-north-temperate species. Most species, however, are confined to Eu- rope. Workers conducting comparative research us- Volume 90, Number 4 O'Kane & Al-Shehbaz 609 Phylogenetic Position of Arabidopsis C. kneuckeri А C. ovczinnikoviiA C. mollissimaA C. himalaicaA C. stricta — Crucihimalaya OW bes A. C lasio Be eringia 1 bursifoli аА Sphaerocardamum E ois Halimolobos diffusa var. jaegeri Halimolobos palmeri var. acutiloba Nerisyrenia linearifolia Lyrocarpa coulteri “Polye olpate Synthlipsis greggii Dithyrea californica Clade Dimorphocarpa wislizenii Paysonia stonensis & P. densipila Physaria pest дез ср & P. didymocarpa ari (1) sa A 4. halleri subsp. halleri K 4. halleri subsp. gemmifer 4. halleri subsp. ovirensis 4. lyrata subspp. kamchatica & petraea K 0.014 _y— |4. arenosa, A. neglecta, A. suecica (clone 18) 97(6 V | 4. croatica A m4 سا‎ 0.014 98(6) . cebennensis ic Arabidopsis . thaliana (#2) A 0.001/100(13) { A. thaliana (#1) & A. suecica (clone 19) Д а) 56(1) Capsella bursa-pastoris —t Г — Arabis pendula * 52(2) | Neslia paniculata I Camelina microcarpa | Pseudoa rabidepsis toxophylla А melon skia calycina 100(3) 10004 Hus س‎ иа abis flagellosayk ? . irabis scabr Arabis sensu stricto |47 'abis арта? Drabopsis Neotoru pairs m abiti: A Neotorularia humilis A Braya & Dichasianthus subtilissimus Neotorularia torulosa Thellungiella halophila А Thellungiella salsuginea A (2) 0.001 79(2) |...) <0.001/100(13) 92(5) Berteroella maximowiczii Figure l. Strict consensus tree of 24 dad qos trees. Whole numbers indicate diner шо eese in parentheses are the Decay Index; dec yee values are Clade Significance. Dashed branches indicate branches with < 50% bootstrap support. р на bold “A” indicates species variously placed in Arabidopsis in ка vious taxonomic treatments. Large asterisks indicate species traditionally placed in Arabis. Genera in bold type are segregates from Arabidopsis named elsewhere as a result of these analyses (see Discussion). 610 Annals of the Missouri Botanical Garden ing А. thaliana as a model organism can now con- fidently the species of a better- circumscribed Arabidopsis, all of which are found use other E in the sister group to A. thaliana, as experimental organisms. This sister group relationship implies that all species of Arabidopsis are equally related to A. thaliana. Unfortunately, the sister group to the genus Arabidopsis cannot be given with confidence. A trichotomy appears below Arabidopsis (Fig. 1): (Arabidopsis clade)(“Capsella—Pseudoarabidopsis” clade)("*Crucihimalaya-Olimarabidopsis" clade). Galloway et al. (1998), using sequences of arginine decarboxylase and а much smaller taxon. sample (28 species from throughout the family), confidently placed Capsella in a sister group relationship to Arabidopsis. Unfortunately, their analysis did not include any members of the *Crucihimalaya—-Oli- marabidopsis” clade. Koch et al. (1999) found, us- ing ITS sequences (and with low bootstrap support). Capsella to be sister to Arabidopsis and Olimara- bidopsis, but their analysis did not include any members of Crucihimalaya. Koch et al. (2001) ob- tained similar results using plastidic matK and nu- clear Chs sequences. Additional work is needed to resolve this issue, but assuming that the “Capsella— Pseudoarabidopsis clade" Is sister to Arabidopsis appears to be a valid working hypothesis. TAXONOMIC IMPLICATIONS ELSEWHERE IN BRASSICACEAE THE Although our intent was not to study the genus Arabis in апу detail, our results mirror those of Koch et al. (1999, 2000, 2001) in showing Arabis, as traditionally recognized, to be polyphyletic even after A. lyrata L., A. pedemontana, and A. ceben- nensis are transferred to Arabidopsis (Fig. 1). In our analysis, the genus Boechera seems to be the proper home for x = A. lyallii A. Gray and A. though not all of the necessary generic transfers Löve & D. Löve 7 species like drummondii S. Watson, have been made. Arabis in its strictest interpreta- tion will consist only of those species in the clade with A. alpina L., the lectotype of Arabis. Arabis glabra (L.) Bernh. belongs to Turritis L. and A. pau- ciflora (Grimm) Garcke belongs to Fourraea Greu- ter & Burdet (Koch et al., 1999). But to which ge- nus does А. pendula L. or A. turrita L. belong (Koch et al., 1999)? Including species once thought to be related to Arabidopsis in our study has also raised other taxonomic questions. Neotorularia, Braya Sternb. & Hoppe, and Dichasianthus Ovez. & Jun- ussov form a well-supported clade (Fig. 1) with Neotorularia being paraphyletic. Thellungiella and Eutrema R. Br. also form a well-supported clade, with Thellungiella being paraphyletic. A surprising result of our study was the discovery of a clade of species all possessing pollen with more than the usual three colpi (see Polycolpate clade, Fig. 1). Palynological studies (Rollins, 1979; Rollins & Banerjee, 1979) showed that among some genera thought not to be closely related in the Bras- sicaceae, colpi range from four to ten. These gen- era, according to Schulz (1936), are as follows: Physaria (Nutt. ex Torr. & A. Gray) A. Gray (tribe Lepidieae, subtribe Physariinae), Dithyrea Harv. and its recent segregate Dimorphocarpa Rollins (Lepidieae, Iberidinae), Lyrocarpa Hoo arv. (Lepidieae, Lyrocarpinae), Nerisyrenia Greene (as Greggia A. Gray) and Synthlipsis A. Gray (Lepi- deae, Capsellinae), and Lesquerella S. Watson (Dra- beae). The results presented here also suggest that within this “polycolpate clade” taxonomic revisions are needed in Lesquerella and Physaria. We have recently united these two genera (Al-Shehbaz & O'Kane, 2002b), except that the auriculate-leaved species formerly placed in Lesquerella are recog- nized as a distinct genus, Paysonia O'Kane & Al- Shehbaz (O’Kane & Al-Shehbaz, 2002). Future work in the family will certainly vield fur- ther taxonomic alignments since there is rampant morphological convergence (Al-Shehbaz et al., 1999; Koch et al., 1999) and because previous tax- onomy in the family has relied heavily on fruit mor- phology (e.g.. Rollins, 1993; Al-Shehbaz, 1984) to the exclusion of floral and vegetative features (Al- Shehbaz et al., 1999). Molecular techniques in con- cert with a reevaluation of morphological characters are rapidly reshaping our understanding of the fam- ily (e.g., Bailey & Doyle, 1999; Bailey et al., 2002; Bowman et al., 2000; Koch et al., 1999, 2000, 2001; Mummenhoff & Koch, 1994; Mummenhoff et al., a, b, 2001а, b; Price & Palmer, 1996; 1996; Warwick & Black, 1993, 1997). Characterizing the membership of Arabidop- sis and sketching its relationships to related genera, Rodman et al.. we believe, contributes to this growing body of knowledge. Literature Cited Al-Shehbaz, 1. A. 1973. The biosystematics of s ge nus Thelypodium (Cruciferae). Contr. Gray Herb. 204: 3- 148. 1984. The tribes of Cruciferae (Brassicaceae) in the sni dicitar United States. J. Arnold Arbor. 65: 343-373. - 088. The genera of Arabideae (Cruciferae: Bras- sicac en in the iS тп United States. J. Arnold Arbor. 69: 85-10€ &S.L.O' Kane, Jr. 1995. Placement of Arabidop- Volume 90, Number 4 2003 O'Kane & Al-Shehbaz Phylogenetic Position of Arabidopsis 611 sis eiue in Thellungiella (Brassicaceae). Novon 5: 309— 1997. Arabidopsis gamosepala and A. tuemurnica belong to Neotorularia (Brassicaceae). No- von 7: 93—94. . 1999. че b on foedo: a neric name replacing the uq pai” da 56: 321-327. & 2a. Taxonomy and Arabidopsis due: eae). (22 August, 2002), € Somerville & E. M. Pii ipe (editors), The pono s sis Book. American Societ Plant Biologists, Rock- ville, Maryland. 401/10. 10олаЬ, 0001, Beim s.l. (Brassicaceae). Bot. J. Linn. 99. Soc. 125: & —. 1997b. Molecular phyloge- — of Thlaspi s.l. (Brassicaceae) based on chloro- plast DNA restriction site variation and sequences of the Ww ET rcu spacers of nuclear ribosomal DNA. ad. J. Bot. C [ Bie mann n J. БОКЕЙ, 2001а. last ‘DNA phylogeny and ee of the genus Lepidium дне aceae). Amer. J. Bot. 88: 2051-2063. «Н. тенин 20010 b. Pachyphrag- ceae) revisited: Molecular data indicate Chloro- ` та enis close B D: to Thlaspi s.str. Folia Geobot. 36: Lm O'Kane, S. L., Jr. ‚ А. Al-Shehbaz. 1997. A synopsis von 3-327. of; кашы 059 асеае). No 2002. Paysonia, a new genus segre- gated fiori Lesquerella (Brassicaceae). Novon 12: 379- 381 B. A. Schaal & I. A. Al-Shehbaz. 1996. The origins of me suecica (Brassicaceae) as indicat- 559-566. ed by nuclear sequences. Syst. Bot. 21: Patrusky, B. 196 Drs botanica (the fruit fly of plant pe Мы Price, D. cie 1996. Сй DNA based e nies and fruit shape in the tribe Arabideae (Brassicaceae). Amer. J. Bot. 83: 5187-8188. J. D. Palmer & I. A. Al-Shehbaz. 1994. System- atic relationships of Arabidopsis: A molecular and mor- phologic 'al perspective. Pp. 7-19 in E eyerowilz & C. R. Somerville (айо). Arabidopsis. Cold Spring Най d Press, New Yor Al-Shehbaz & S.L. O'Kane, Jr. 2001 ж (Боні асеае), a new genus of Arabidopsid affin- ities from Russia and North America. Novon 11: 332- 336. Rodman, J. E., R. A. K. Karol, E. Conti, K. J Sytsma & J. D. Ee 1993. Nuc оине sequences of E the rbcL gene indicate be id of mustard oil plants. Ann. Missouri Bot. Gard. 80: 686—699. ‚К. . A. Price & К. J. Sytsma. 1996. Molec ules, ee НЕ, and M e xpande sd order apparales. d Bot. 21: 307 RD R. C. 1979. Dithyrea and a related genus (Cru- ciferae). Publ Bussey Inst. Harvard Univ. 1979: 3-32. 993. The Cruciferae of Continental North Amer- ica. Stanford Univ. Press, Stanford. C. Banerjee. 1979. Pollens ies ruciferae. Publ. luce Inst. а а Univ. 1979: : Schulz, O. E. 1924. Crucife и ibrieae. i A. Engler (editor), Bun h IV. x 86): 1-388. Verlag von Mer d Engelmann, is ;zruciferae. In A Nani Se Pflanzenfam., мі 2, lag von Wilhelm Engelmann, Leipz Somerville, C. R. & Meyerow йк А dius 2002. The Arabidopsis Book . Africa Bobartia gladiata (L.f. ) Ker-Gawl. South Africa, W. Cape, Boucher 5263 (NBG) ietes grandiflora N. E. Br. Cultivated material, NBG Kirstenbosch, South Africa Ferraria crispa Burm. South Africa, W. Cape, Goldblatt & Manning 11665 (MO Ferraria foliosa G. J. Lewis South Africa, W. Cape, Goldblatt & Porter 11888 (MO Ferraria ferrariola (Jacq.) Willd. South Africa, N. Cape, Goldblatt & Porter 11765 (MO) Ferraria divaricata Sweet subsp. divaricata South Africa, N. Cape, Goldblatt & Manning 10176 (MO) Ferraria schaeferi Dinter South Africa, W. Cape, Goldblatt & Porter 11734 (MO) Moraea collina Thunb. South Africa, W. Cape, Goldblatt 2132 (MO, NBG) Moraea spathulata (L.f.) Klatt South Africa, E. Cape, Goldblatt 12232 (MO) Tribe Tigridieae Alophia drummondii (Graham) R. C. Foster Cultivated material, Missouri BG, U.S.A. Ennealophus euryandrus (Griseb.) P. Ravenna Cultivated material, RBG Kew, U.K. yelasine elongata (R. Grah.) P. Ravenna Cultivated material, ex hort. B. Mathew, Surrey, U.K. Tigridia meleagris (Lindl.) Nichols. Mexico, Kenton, Rudall & Howard 49-319 (K) Tribe Trimezieae Pseudotrimezia planifolia P. Ravenna Brazil, Harley et al. 25445 (K) Subfamily Crocoideae (Ixioideae) Tribe Pillansieae Pillansia templemannii L. Bolus South Africa, W. Cape, Goldblatt 7907A (MO, NBG) 616 Annals of the Missouri Botanical Garden Table 1. Continued. Taxon Collection data Tribe Watsonieae Lapeirousia neglecta Goldblatt & J. C. Manning Micranthus alopecuroides (L.) Rothm. Thereianthus racemosus (Klaff) G. J. Lewis Watsonia angusta Ker-Gawl Tribe Ixieae Babiana sinuata G. J. Lewis B. stricta (Aiton) Ker-Gawl. Crocus sieberi Gay Dierama pendulum (L. f.) Walp. ›. L. Meyer) кейш F. geben (Baker) Klaff Geissorhiza ri i Klatt Gladiolus graci 1 oo fale ata (L. f.) Ker-Gawl. ia polys Melasphae rula ramosa (Burm. f.) N. E. Br. tachya Romulea rosea (L.) E Sparaxis grandiflora ( a Delw һе) Ker-Gawl. Tritonia squalida (Aiton) Ker-Gawl. Tritoniopsis parviflora (Jacq.) С AJ 18 Tritoniopsis burchellii (N. E. Br.) Goldblatt Xenoscapa fistulosa (E. Mey. ex Klaff) Goldblatt South Africa, W. Cape, Goldblatt & Manning 9022 (MO) South Africa, Goldblatt & Manning 10431 (NBG) South Africa, W. Cape, Manning s.n. (NBG) South Africa, W. Cape, Snijman 971 (NBG) South Africa, W. Cape, Goldblatt 2545 (MO, NBG) os Africa, W. Cape, Goldblatt & Manning 10343 BG) а — RBG, Kew, U.K. Cultivated ma NBG Kirstenbosch, South Africa Cultivated mate iem NBG Kirstenbosch, South Africa Cultivated material, NBG Kirs tenbosc h, South Africa South Africa, W. Cape, Manning 2 (NBG) South Africa, W. Cape, Manning 2016 (NBG) South Africa, near George, Goldblatt & Manning s.n. South Africa, ex hort. Kirstenbosch South Africa, W. Cape, Cape Peninsula, Goldblatt & d s.n. South A ‚ W. Cape, Manning 2017 (NBG) быша оре МВС Kirstenbosch, South Africa South Africa, W. Cape, near Albertinia, Goldblatt & Manning s.n. South Africa, Goldblatt & Manning 2283 (NBG) South Africa, W. Cape, Goldblatt & Manning 9869 NBG) South Africa, W. Cape, Manning 2028 (NBG) tia, Nivenia, and Witsenia (Figs. 3A, F, G, 4D, E). which together form a clade (Manning & Goldblatt, 1991; Reeves et al., 2001). Septal nectaries are ab- sent from the monogeneric subfamily Isophysidoi- deae (/sophysis) (Fig. 2E) Perigonal nectaries and elaiophores (oil-produc- ing hairs) are mostly confined to subfamily Iridoi- deae, in which septal nectaries are always absent (except Diplarrhena; see above). However, there are at least two examples of perigonal nectaries or elaiophores in taxa outside Iridoideae: (1) one spe- cies of Nivenioideae (Aristea spiralis (L.f.) Ker- Gawl.) produces nectar that is secreted from small perigonal nectaries on the short perianth tube be- low the base of the free parts of the tepals (Fig. 4A); Daumann (1970) also reported perigonal nec- laries in А. africana (L.) Hoffmanns. (2) In Trito- niopsis parviflora (Jacq.) G. J. Lewis (subfamily Cro- coideae), floral oils are produced from a glandular epithelium at the base of the free parts of the tepals and in the mouth of the perianth tube (Manning & Goldblatt, 2002). Oil secretion is supplemented by production of sugary nectar from septal nectaries. Within subfamily Iridoideae, perigonal nectaries and elaiophores take several different forms. Dau- mann (1935) illustrated highly vascularized nectar- iferous regions present on different parts of the flower surface in different species of Iris. These in- cluded: (1) Nectaries present on the base of the perianth tube (comprising fused perianth and staminal tissue), especially (but not exclusively) in the interstaminal ensata Í. pseu- and also in two allied mono- typic genera, Belamcanda and Hermodactylis). regions (e.g.. in douglasiana Torr., 1. Thunb., /. foetidissima L., I. graminea L., dacorus L., I. sibirica L., (2) Nectaries present around the base of the style (e.g., in Z. dichotoma Pall.). (3) Nectaries present in a continuous region ex- tending from around the base of the perianth tube to around the base of the style (e.g., in Z. chamaeiris Bertol., I. germanica L., 1. kolpakowskiana Regel, I. pumila L., I. stolonifera Maxim., І. tectorum Max- im., /. tingitana Boiss. & Reut., /. warleyensis Fos- ter, Z. xiphioides Ehrh.) extending onto the bases of the inner tepals in /. sisyrinchium L. These аге not Volume 90, Number 4 2003 Rudall et al. Floral Nectaries in Iridaceae Table 2 Distribution of nectaries in Iridaceae (arranged according to the classification of Goldblatt, 1990, 1998a). Taxon Septal nectaries Other floral secretory structures Isophysidoideae (/sophysis) ^ absent (Fi abse ig. 5A) nt from Aristea (Fig. Nivenioideae 4B), Geosiris (fig. 4C), Pa- tersonia, present in Klat- па, Nivenia (Fig. ЗЕ, G) Wits (Fig. 4D, F Iridoideae absent (except in Diplar- rhena: Fig. 5B) Bobartia absent (Fig. 5A) ietes absent (Fig. 6B) Diplarrhena present (Fig. 5B) Iridoideae-Sisyrinchieae absent (Figs. 2C, D, 3E, 7A, 8A) Iridoideae—Irideae absent (Figs. 5A, 6B) lIridoideae-Tigridieae absent (Fig. 2F) Iridoideae-Trimezieae absent Crocoideae present in all taxa (Figs. 2 C) B, 5C, 6 А, absent absent, except in Aristea spiralis, in which small perigonal nectaries present (Fig. 4 often present absent absent absent nectar secreted from filament column in Olsynium (Figs. 7, 8); elaiophores present on filament col- umn in many species of Sisyrinchium subg. Sisy- rinchium (Fig. ЗС-Е) perigonal nectaries present in Moraea (Fig. 6D) and Ferraria, or on the walls of the perianth tube in Tris elaiphores present on tepal claws in some species (Figs. 3B, 9) elaiophores present on tepal claws in some species absent septal nectaries, because they do not extend into the unfused regions of the carpel margins within the ovary, but they could easily be mistakenly iden- tified as such in the absence of anatomical sections. Among other members of the Old World and North American tribe Irideae, in Moraea and Fer- raria the smooth glistening perigonal nectaries, when present, can readily be seen with the naked eye; they are located at the bases of the outer tepals (in most species of Moraea) or at or near the bases of all six tepals (in Ferraria and many Moraea spp.) (Vogel, 1954; Daumann, 1970; de Vos, 1979; Gold- blatt, 1986, 1987; Goldblatt & Manning, unpub- lished data). In some species of Ferraria the nec- taries may cover more than half the area of the tepal claws and measure up to 4 X 3 mm. Species of Dietes and Bobartia apparently lack nectaries en- tirely (Figs. 5А, 6B) and produce no nectar, as in- dicated by visual observation and the taste test (ap- plying likely nectar-bearing parts of the flower to the tongue) (Goldblatt, MEN we data). This ob- (1970) reports of per- igonal nectaries in Dietes bicolor (Lindl.) Sweet and D. iridoides (L.) Sweet ex Klatt (as both D. vegeta (L.) N. E. Br. and Moraea iridoides L.). In Moraea ochroleuca (Salisb.) Drapiez and M. lurida Ker- servation refutes Daumann's Gawl. nectar is produced along the length of all six tepal claws above the vascular strands but no nec- tariferous glands are evident (Goldblatt et al., 1998; Goldblatt & e 2002). Nectaries are not de- veloped pollinated br hopifine scarab beetles, e.g.. M. can- . C. Manning, M. tulbagh- in several species of Moraea that are tharophila Goldblatt ensis L. Bolus, and M. villosa Ker-Gawl. (Goldblatt et al., 1998; Goldblatt & Manning, 2002; Steiner, 1998 The New World Iridoideae are currently distrib- uted among tribes Sisyrinchieae, Trimezieae, an Tigrideae (Goldblatt, 1990, 1998a). These include numerous species that produce nectar from various different parts of the flower, although some appar- ently lack nectar entirely; for example, Cocucci and Vogel (2001) reported that Olsynium philippi (Klatt) Goldblatt (Sisyrinchieae) In many Iridoideae (and other Iridaceae) the sta- does not produce nectar. men filaments are free in the upper half and fused into a filament column at their base. In O. douglasii A. Dietr. the outer surface of the filament column is smooth, with large epidermal cells and dark- staining highly vascularized tissue beneath, indi- cating a nectary (Fig. 7A—D). The inner surface of the filament column is highly convoluted and thin- walled. The filament column is swollen just above the base; nectar presumably accumulates in a sub- epidermal space in this region. After exudation, nectar then accumulates in a semi-enclosed cham- 618 Annals of the Missouri Botanical Garden re 2. —A, of ovary (B); septal nectaries arrowed. — ovary (C) and style plus E tube — ~ ; nectaries absent ovary; nectaries absent. 0.5 mm for all images. ber around the narrow basal part of the column. This nectar chamber (also present in some Moraea species, e.g., M. miniata Andr.; Goldblatt & Bern- hardt, 1999) is enclosed by the short perianth tube (in Olsynium), the walls of which are almost in con- tact with the swollen portion of the filament column, or by the short tepal claws that abut the swollen lower portion of the filament column (in Moraea). Other species of Olsynium have a similar swollen (D . Gelasine кр (Iridoideae-' l'igridicae). TS of top of ovary, nectaries absent. Scale . Crocus sieberi (Croc ew. transverse sections (TS) of flower through base of ovary (A) and top ‚ D. Orthrosanthus е (Iridoideae-Sisyrinchieae), TS of flower through ophysis tasmanica (Isophysidoideae), TS of superior base to the filament column and a perianth tube and most likely also produce nectar in the manner described for О. douglasii. At least O. junceum (Е. Mey. ex Presl) Goldblatt is known to produce nectar (Forcone et al., ed from anticlinally elongated Saris cells on both surfaces of the filament tube (Fi Many other New World Iridoideae E trichome elaiophores, which are invariably unicellular and 1998), which is presumably secret- Volume 90, Number 4 Rudall et al. 619 2003 Floral Nectaries in Iridaceae ma Уа Са um 4 - ЕУ T e tru o» apr > > WILL e 3. —A. Klattia stokoei (Nivenioideae), TS of flower just proximal to top of semi-inferior ovary, showin septal nectaries (arrowed). —B. Tigridia meleagris (lridoideae-Tigridieae), TS of inner tepal with oil-producing tri- chomes (elaiophores). C-E. Ravine hium s pu Sisyrinchieae). —C. Longitudinal section (LS) of flower show- ing distal end of inferior ovary a © proximal e nd of perianth tube, staminal column, and central style, with elaiophores present on Aliana column. —D. TS of staminal column (with elaiophores) and central style. —E. TS of ovary, lackin septal nectaries. F, G. Nivenia rh (Nivenioideae). —F. TS of flower just proximal to top of semi- oper ovary, showing septal nectaries (arrowed). —G. TS of perianth/stamen tube and style. Scale = 0.5 mm for all images 620 Annals of the Missouri Botanical Garden Tom 4. wed. Aristea ecklonii (Nivenioideae), TS of ovary, (Nivenioidez kd TS of ovary, lacking se e nec laries. D, E, with one of three se E папа s arrowe — |00 um; in B-E m club-shaped, with a rounded tip. In Sisyrinchium subg. Sisyrinchium trichomes (elaiophores) present on the lower part of the filament column (Fig. 3C, D) secrete floral oils; this was first reported by Vo- gel (1971, 1974) and described in detail by Co- cucci and Vogel (2001). Similar trichome elaio- Witsenia maura (Nivenioideae). — 5 of ovary, with one of three septal nectaries arrowed. Scales: Aristea spiralis (Nivenioideae), LS of top of ovary with perigonal nectary at base of perianth tube ipo madagascariensis lacking septal nectaries. С. TS of top part of ov vary, in А phores also occur on the tepal claws in Tigridieae, including Alophia (syn. Eustylis), Cypella, Enneal- ophus (Fig. 9), Fosteria (a genus closely allied to, if not nested in Tigridia: Goldblatt, 1998a), Gelas- ine, Herbertia, many species of Tigridia (Fig. 3B), and Trimezieae, including Neomarica (Molseed, Rudall et al. Floral Nectaries in Iridaceae Volume 90, Number 4 2003 * a “= E - "ure 5. . Bobartia gladiata aper eri iui кыба ен eae), TS of s те of three septal nectaries arrowed. 20, p nectaries arrowed. Scales: in A, C, D TS of ovary, lacking septal nectaries. —B. Diplarrhena ptal nectary (arrowed). Babiana stricta е oideae), TS of ovary, with ritontopsis ҮТ" m (С! roc bei sae), TS of ovary, with one of three septal — 0.5 mm; in B — ) рт 621 622 Annals of the Missouri Botanical Garden IT e 09:01006". са { e Figure 6. —A. Watsonia angusta nn 'oideae), TS of ovary, with one of three septal nectaries arrowed. —B. Dietes grandiflora ( шы, sae), TS of ovary, lacking septal nectaries. adiolus gracilis (Croc db. TS of ovary, with « me of three septal nectaries arrow: a —D. Hae spathulata (Iridoideae-Irideae), TS of tepals, with perigonal nectary arrowed. Scale = 0.5 mm. Volume 90, Number 4 Rudall et al. 623 2003 Floral To 0 in lridaceae 2 « >з - * Е] + Ss € r, Figure T. е о (Iridoideae-Sisy rinchieae), serial TS of flower. —A. TS of ovary, lacking septal nec taries. —B. Just above top of inferior ovary, showing central style surrounded by staminal column pa rtially fused with perianth tube. ra Mor T s section sh lowing central style surrounded by staminal column in which the outer epidermal cells are axially elongated and the inner epidermis i is thin-walled and convoluted. —D. Enlarged view ps of staminal column showing extensive vascular tissue beneath outer side Seales: in A-C = 0.5 mm; in D 00 p.m. 624 Annals of the Missouri Botanical Garden ER os 8. . TS above top of inferior inner epidermal cells are ab elongated. Scale = 0.5 mm. 1968, 1970; Vogel, 1974; Lee, 1994). The elaio- phores develop from epidermal cells rather late in development, shortly prior to anthesis. Their distri- bution on the tepals varies, although in general the Olsynium vues (Iridoideae-Sisyrinchieae), TS of flower. lacking septal nectaries. —A. TS « vary, showing central style surrounded by staminal ae in which both outer and ovary, zone of glandular trichomes is restricted to the ad- axial surfaces of the smaller inner tepals, often con- centrated in pouches at the juncture of the tepal claw and limb. In Herbertia there are nectaries on Volume 90, Number 4 2003 Rudall et al. Floral Nectaries in Iridaceae Figure 9. Ennealophus euryandrus و‎ ача Жозе trichomes not gi uoo of olde i als. —C. T inner tepals. Seale = 0.5 mm the smaller inner tepals and the claws of the outer tepals. At least two genera of the tribe apparently lack nectaries: Sessilanthera and Cobana, which have porose anthers and on this basis are expected to be buzz pollinated. Absence of nectar is often associated with buzz pollination, and information e), TS of flowers through style. —A. r bud; short perigonal trichomes (elaiophores) present on inner S of bud те тази longer trichomes (elaiophores) present on inner surface of TS of young bud; available indicates that these two genera lack nec- taries and elaiophores (Molseed & Cruden, 1969; Ravenna, 1974). Eleutherine and Nemastylis also , 1970) but these species have conventional longitudinal anther de- hiscence. evidently lack nectaries (Molsee 626 Annals of the Missouri Botanical Garden Red-flowered and presumably hummingbird-pol- linated species of Tigridia, which were formerly placed in a separate genus, Rigidella (Goldblatt, 1998a), secrete copious amounts of sugary nectar from nectaries on the inner tepal claws (Cruden, 1971). Other Tigridieae may also secrete nectar from perigonal nectaries, although this is not ade- quately documented. Molseed (1970) mentioned fly and wasp visitors in some small-flowered species of Tigridia; as far as we know flies and wasps do not take floral oils as a reward. Lee (1994) described a novel mode of nectar pro- duction in Alophia drummondii Grah. in which the expanded, fiddle-shaped anther connectives secrete a sticky mucilage with dilute sugar content. During the later stages of anthesis this is concentrated by evaporation to form an alternative nectar source. The primary role of the mucilage is reportedly to promote pollen adherence on the dorsal thorax of visiting bees. DISCUSSION SEPTAL NECTARIES IN ASPARAGALES An understanding of sister-group relationships is critical for optimization of characters onto an ex- isting phylogeny. However, this is problematical for Iridaceae because relationships among the various lower asparagoid clades are still not satisfactorily resolved, partly because /xiolirion is a “wildcard” taxon on a very long branch in molecular analyses (Fay et al., 2000), tending to destabilize tree to- pology. Three taxa, /xiolirion (Ixioliriaceae), Dor- yanthes (Doryanthaceae), and Tecophilaeaceae, are commonly placed close to lridaceae in cladistic analyses (e.g., Chase et al., 1995; Rudall et al., 1997; Fay et al., 2000). Of these, the Australian bird-pollinated genus Doryanthes (ca. two spp.) is a good candidate for a sister group relationship with Iridaceae because it shares some synapomorphies, including presence of styloid crystals (and absence of raphides). Within Iridaceae, raphides are invari- ably absent and almost all taxa have styloids (Gold- blatt et al., 1984) with the exception of Sisyrin- chium and its close allies, which lack crystals altogether (a synapomorphy for this group: Rudall et al., 1986; Goldblatt et al., 1990). On the other hand, raphides are present (and styloids absent) in both /xiolirion and Tecophilaeaceae (Prychid & Ru- dall, 1999). Furthermore, both Doryanthes and Ir- idaceae have extrorse anther dehiscence, a char- acter that is otherwise unusual in Asparagales. Ixiolirion is a small Eurasian genus of ca. three species with epigynous flowers with septal nectaries in the top half of the ovary (Simpson & Rudall, 1998; Rudall, 2002). Fay et al. (2000) placed /xi- olirion with Doryanthes in an equally weighted mul- tigene analysis and with Tec o eae in a suc- cessively weighted analysis. Analysis of morphological data (Rudall, 2002b) supported a lose association between /xiolirion and Doryan- ^ thes, although /xiolirion shares a corm with Teco- 1998; Simpson & Rudall, 1998), while Doryanthes has a short vertical rhi- zome (Clifford, 1998). Septal nectaries are absent from some Tecophi- laeaceae (Simpson & Rudall, 1998), but present in both Doryanthes, in which copious nectar is se- creted (Newman, 1928; Patil & Pai, 1981; Kocyan & Endress, 2001) and /xiolirion (Rudall, 2002a). Loss of septal nectaries has apparently occurred de philaeaceae (Kubitzki, novo several times within Asparagales (Rudall, 2002a), including at least once in both Tecophilae- aceae and lridaceae. Therefore, despite their ab- sence from the putatively basal genus, /sophysis (see below, and introduction), the most likely sce- nario is that presence of septal nectaries is the ple- siomorphic condition for Iridaceae (Fig. 1). Total absence of septal nectaries (i.e., complete fusion of carpel margins) occurs sporadically within Asparagales (sensu АРС, 1998), in contrast to the related order Liliales, where absence of septal nec- taries and presence of perigonal nectaries both rep- resent highly consistent synapomorphies (Rudall et al., 2000). an alternative pollination mode; for example, sev- Total absence of nectar often indicates eral asparagoid families with vibratile (buzz) polli- nation and Solanum-type flowers lack nectaries en- tirely, including some Laxmanniaceae (e.g Arthropodium), some Hemerocallidaceae, some Te- cophilaeaceae, and apostasioid orchids (Vogel, 1981; Dressler, 1993; Bernhardt, 1995, cyan & Endress, 2001; Rudall, 2001, 2002a). Sep- nectaries are often absent from epigynous taxa (Rudall, 2002a) and always absent from “hyper- epigynous" taxa with gynostemia such as Orchi- daceae (Rudall & Bateman, 2002). NECTARIES IN IRIDACEAE Within Iridaceae, there is some topological con- flict at the deeper nodes resulting from analyses of different data sets, especially between molecules and morphology, although even the molecular to- pologies differed in some respects (Reeves et al., 2001). Analysis of four plastid DNA data sets (rbcL, rps4, trnL, trnL-F) for 57 genera of Iridaceae plus outgroups, found that in each of the individual trees, some genera or groups of genera were mis- placed re alive to morphological cladistic studies, Volume 90, Number 4 2003 Rudall et al. Floral Nectaries in Iridaceae but the combined analysis produced a tree in which all subfamilies were resolved as monophyletic, ex- cept that Nivenioideae formed a grade in which Crocoideae (Ixioideae) were embedded (Reeves e al., 2001). Among other genera of disputed affini- ties, Geosiris was placed within the nivenioid grade (Goldblatt et al., 1987), /sophysis was sister to the rest of the family, and Diplarrhena was placed in a well-supported position as sister to Iridoideae (Ru- dall & Goldblatt, 2001). Bobartia (formerly Sisyrin- chieae: Goldblatt & Rudall, 1992) has found sup- port as a member of Irideae (Iridoideae), by both Reeves et al.’s (2001) study and a molecular anal- ysis of the ITS2 gene (Donato et al., 2000). the combined (successively = (2001) is correct, Tasmanian genus /so- physis as the basal taxon ig 1), and also that the presence of septal nectaries is the plesiomorphic condition for Iridaceae (see above), then there are two possible evolutionary scenarios for this char- acter within the family: (1) suppression of septal nectaries in the ancestor of Iridaceae plus at least two de novo developments of septal nectaries, in Nivenioideae/Crocoideae and Diplarrhena (and probably a subsequent secondary loss in lridoi- deae), or (2) separate suppressions of septal nec- taries in /sophysis, the basal Nivenioideae (Aristea, Geosiris, and Patersonia), and Iridoideae, as illus- trated in Figure 1. The latter scenario would in- volve several separate losses of septal nectaries in Isophysis, Aristea, Geosiris, and Patersonia. How- ever, it is possible that Aristea, Geosiris, and Pater- sonia may ultimately form a clade, since Reeves e al. (2001) suggested that the existing topology. which conflicts with morphological data for these taxa, is questionable due to high levels of sequence divergence. In any case, this hypothesis, although less parsimonious, is perhaps ultimately more plau- sible, especially if (as seems likely) perigonal nec- taries evolved from septal nectaries (see below). Nivenioideae and Crocoideae. The subfamily Nivenioideae is a relatively heterogeneous group with respect to several characters, including nec- taries (Table 2). Reeves et al. (2001), in analyses of molecular data from four gene loci, found Niv- enioideae to be paraphyletic to Crocoideae (Ixioi- deae), although the three shrubby Cape genera Klattia (3 spp.), Nivenia (10 spp.), and Witsenia (1 sp.) together formed a well-supported clade. Klat- tia, Nivenia, and ipia all have septal nectaries (Figs. ЗА, F, С, 4D, E). In Nivenia, nectar ranges — from 1 to 2 wl in nue has a relatively high sugar concentration, and is sucrose-rich (Goldblatt, 1993). In contrast, nectar volumes in Klattia and Witsenia are substantial; as much as 77.9 wl of nec- ar per flower has been recorded in W. maura — Thunb. and 46.9 wl nectar in K. stokoei Guthrie (Goldblatt, 1993). In these genera nectar sugar con- centrations are low, 12-1546 sucrose equivalents, and nectars are hexose (glucose plus fructose)-dom- inant. Larger volumes of nectar, low sugar concen- trations, and dominance of hexose sugars in these two genera are associated with sunbird pollination compared with anthophorine bee and/or long-pro- boscid fly pollination in Nivenia (Goldblatt, 1993). Most species of the Afro-Madagascan Aristea (ca. 50 ѕрр.; Nivenioideae) do not produce nectar and lack septal nectaries; the presence of perigonal nectaries in A, spiralis is apparently an autapomor- phy for this species. Aristea spiralis is embedded within the derived subgenus Pseudaristea, which is characterized by derived large pollen grains with specialized 2-zonasulcate apertures (Goldblatt & LeThomas, 1997). Nectar production in A. spiralis appears to be correlated with its pollination by long-proboscid flies, evidently unique in the genus (Johnson, 1992; Goldblatt & Manning, 1997). Of the remaining genera of Nivenioideae, nectaries are absent from the Madagascan saprophyte Geosiris, and the salverform flowers of the Australasian Pa- tersonia, which have a long perianth tube. In this genus the tube evidently serves to raise the nearly sessile flowers out of the long inflorescence spathes that enclose the flower ithin subfamily Crocoideae (28 genera, cur- rently arranged in three tribes), nectar character- istics are highly variable in volume and concentra- typically — sucrose-rich to sucrose-dominant. This is associated with the di- verse pollination systems found in most genera of the нү; (Goldblatt et al., 1995, 1998, 2001, in press; Goldblatt & Manning, 2000; Bernhardt & Goldblatt, 2000). only significant departure from the pattern of high sucrose concentration is in some (but not = sunbird-pollinated species (Gold- blatt et al., 1999, 2001). Discharge of nectar from septal nectaries is highly variable (Smets et al., 2000). Suppression of nectar production in some species of Ixia subg. Ixia is associated with a shift in pollination strategy from anthophorine bees to hopliine beetles in section /xia (hopliines have no interest in. nectar and use flowers largely as sites for assembly, mate selection, and copulation) (Gold- blatt et al., 2000a). As far as is known, species of the other section Dichone of Ixia subg. Ixia are buzz pollinated by female anthophorine bees, and pollen is the sole floral reward (Goldblatt et al., 2000a). In subgenus /xia, suppression of nectar secretion is correlated with the narrowing of the diameter of tion, but are Annals of the Missouri Botanical Garden the perianth tube so that its walls tightly surround the style. functions as a stalk (pseudopedicel) for the out- The tube in these species effectively spread tepals. The narrowing of the tube so that it clasps the style occurs in species of several other genera, notably in Romulea and Sparaxis, but usu- ally traces of nectar are still secreted and presented at the mouth of the tube (Goldblatt et al., 2000a, b), and de Vos (1972) has confirmed that septal nectaries are present in Romulea. Often, no trace of nectar can be detected in flowers of these spe- cies, presumably because the nectar has already been removed by visiting bees. In two species of Hesperantha, however, nectar secretion appears to be completely suppressed and pollination is accom- plished by deceit, the flowers of these species re- sembling others that do secrete nectar (Goldblatt et al., in press). In Tritoniopsis parviflora, production of floral oils from glandular epithelium at the base of the free parts of the tepals (Manning & Goldblatt, 2002) is apparently unique among Crocoideae. This species is pollinated over part of its range by the oil-col- lecting melittid bee, Rediviva gigas. Species of Re- diviva are assumed to use floral oils as part of the food supply for larvae (Steiner & Whitehead, 1988, 1990) Iridoideae: Diplarrhena. The presence of septal nectaries in. Diplarrhena argues against its inclu- sion in the subfamily Iridoideae, which otherwise Rudall and Goldblatt (2001) erected a new tribe, Diplarrheneae, for this genus entirely lack them. but noted that its phylogenetic position remains un- certain. Molecular data did not fully resolve the positions of either Diplarrhena or Patersonia, since they occupied reversed positions in the rps4 topol- to the rbcL and trnL-F trees ogy with respect (Reeves et al., 2001), although in the combined (successively weighted) molecular analysis their positions were strongly supported (BS > 96%) with Diplarrhena as sister to Iridoideae, and Patersonia sister to the remaining Nivenioideae/Crocoideae. On the other hand, the presence of septal nec- taries in Nivenioideae/Crocoideae plus Diplarrhena tends to support the rps4 topology (Diplarrhena as sister to Nivenioideae/Crocoideae) rather than the combined topology (Fig. 1), although Diplarrhena lacks the repeated codon of the rps4 gene that is found in the Nivenioid—Crocoid clade (Souza-Chies 1997). similarities with some Nivenioideae; for example, et al., Diplarrhena shares morphological pollen of Diplarrhena strikingly resembles that of both pollen with a vermiform surface patterning (Rudall & Goldblatt, 2001). Flowers of Diplarrhena differ Patersonia, having distinctive inaperturate from those of most other Iridoideae, especially in the style branches, which are flat lobes, unlike the tubular style branches (sometimes with apical ap- pendages) and terminal or subterminal stigma lobes typical of Iridoideae. Conversely, the leaf anatomy of Diplarrhena is rather similar to that of Libertia (Iridoideae—Sisy- rinchieae) (Rudall, 1986, 1994), and the simple rhipidial inflorescences are consistent with those of other Iridoideae. Furthermore, although the sub- family Nivenioideae lacks clear synapomorphies and may well be paraphyletic (Reeves et al., 2001), Diplarrhena differs from them in some respects, no- tably inflorescence morphology, having flowers in simple rhipidia rather than the paired (or binate) rhipidia of all Nivenioideae (Goldblatt, 1990; Man- ning & Goldblatt, 1991). Diplarrhena also differs from Nivenioideae in that the root metaxylem ves- sels in all Nivenioid genera have only scalariform perforation plates, whereas in most other Iridaceae, including Diplarrhena, they are simple (Cheadle, 1963). Clearly, no single character сап indepen- dently test relationships, and a revised morpholog- ical analysis of Iridaceae in the light of new data may help to resolve this conundrum. Perigonal nectaries within Iridoideae. | Among other Iridoideae, there is apparently a preadapta- tion to production of perigonal secretory structures (character В in Fig. 1), since both perigonal nec- taries and elaiophores (oil-producing glands) are widespread in the subfamily but otherwise unusual within Asparagales (although some Orchidaceae have nectariferous tepaline spurs). The presence of sugary nectar (as opposed to oil) secreted from the filament column is an ашаро- morphy for Olsynium. The “diffuse” nectaries that occur in Moraea lurida and M. ochroleuca appear to be unique in the family, yet must have arisen independently from each other, since the two spe- cies belong in different subgenera (Goldblatt, 1986, 1998b). However, given their topological similarity and close phylogenetic relationship, it seems likely that the elaiophores found in some Sisyrinchium species may have evolved from perigonal nectaries. Daumann’s (1935) detailed work on nectaries in many different species of /ris (Iridoideae) may help to elucidate nectary evolution in Iridaceae. In some Iris species the nectariferous regions extend from around the base of the perianth/stamen tube and around the base of the style, so that the nectary exudes nectar at the same place as it would in an epigynous flower that possessed septal nectaries (e.g.. in members of subfamily MEDIEN. or in 2000). This raises the intriguing possibility that perigonal Dioscorea, Dioscoreaceae: Caddick et al., Volume 90, Number 4 2003 Rudall et a 629 l. Floral Nectaries in Iridaceae nectaries may have evolved from septal nectaries by heterochrony (temporal shifts), or possibly het- erotopy (Alberch et al, 1979; Bateman, 19 Baum & Donoghue, 2002; Rudall & Bateman, 2002), specifically reflecting later formation of nec- taries that are expressed in a more distal position on organ primordia. This hypothesis, which re- quires testing in a molecular-developmental con- text, is dependent on taxa with perigonal nectaries having evolved from ancestors with septal nectaries (as shown in Fig. 1) rather than from taxa that lacked septal nectaries entirely. It is noteworthy that perigonal nectaries, which are relatively un- are frequently associated common in monocots, with epigyny (e.g., in Iridaceae and Orchidaceae), although exceptions include some taxa of the order Liliales, in which perigonal nectaries are common and both epigyny and hypogyny occur (Rudall et al, 2000). Van Heel (1988) examined gynoecial and septal nectary development in some hypogy- nous Asparagales, but there have been very few such developmental studies on epigynous Aspara- gales, although Payer (1857) illustrated organogen- esis in Gladiolus communis (Iridaceae) and Mogen- sen (1969) in Agave (Agavaceae). Rudall (2002a) examined floral development in the taxonomically isolated lower asparagoid genus Xanthorrhoea (Xanthorrhoeaceae), in which the ovary is hypogy- nous at anthesis but epigynous at early develop- ment stages, and septal nectaries are extensive. These investigations all showed initial free devel- opment of organ primordia (two peripheral (ера! whorls, two stamen whorls, and finally three free carpels) and subsequent axial elongation of the re- ceptacular/placental region by elongation of their common bases. Furthermore, it seems unlikely that a de novo nectar source should have evolved in Iridaceae at the same time as suppression of septal nectaries, unless there is some developmental relationship between the two structures. Perigonal nectaries ap- pear to be simply an alternative source of nectar for the same insects, including bees of various fam- ilies, that also pollinate flowers of some Crocoideae, where nectar from septal nectaries is proffered. The absence or paucity of flowers adapted for special- ized pollinators in Iridoideae, including long-pro- boscid flies, sunbirds, moths, and large butterflies, so common in Crocoideae, seems related to the presence of a well-developed perianth tube and widespread floral zygomorphy in Crocoideae and not directly to nectary type. However, the source of nectar is in the ovary, and thus at the base of the perianth tube where nectar is retained and beyond the reach of short-tongued or short-proboscid in- sects. This factor is obviously significant in the de- velopment of the specialized pollination strategies so common within Crocoideae (Bernhardt & Gold- blatt, It is also гони that within Asparagales tri- chome elaiophores were exploited only in a subfam- ily in which perigonal nectaries were developed (i.e., Iridoideae). Oil collection is a widely exploited syndrome in the Neotropics and adjacent temperate zones to the north and south, and bees of several genera in different families utilize floral oils, hence the disproportionate development of oil secretion in the New an s. [n contrast to the situation in the New World, in which literally hundreds of Iridaceae secrete floral oils, in the Old World only one species of Iridaceae (Tritoniopsis parviflora) is known to do so, and only one genus of bees is adapted to exploit floral oil secretion. Literature Cited Alberch, P., S. J. Gould, G. F. Oster & D. B. Wake. Size and та іп ontogeny and phylogeny. ا‎ 5: 296-317 APG (Angiosperm Phylogeny Group). 1998. An ordinal classification for the families of flowering plants. Ann. Missouri Bot. Gard. 85: 531—553. Bateman, R. M. 1994. Evolutionary-developmental change in the growth architecture of fossil rhizomorphic ا‎ Scenarios constructed on cladistic founda- . Biol. Rev. 69: 527—597. йн. m & M. J. Donoghue. 2002. Transference of func- tion, heterotopy and the evolution of plant development. Pp. 52-69 in Q. C. B. Cronk, R Bateman Hawkins (editors), Developmental e and Plant Evolution. Taylor & Francis, Londor Be diis P. 1995. The floral ecology of ере саегшеа аг. аѕѕега Mew Gs Cunninghamia 20. 1996. Anther adaptation in =) ee Pp. 192— 220 in e G. D'A Keating (editors), The Anther: Form, Function, ES Phylogeny. Cambridge Univ. eg Cambridge P. Goldblatt. 2000. The Bs of pola мй ы in the iar 'eae of sout 301—308 in K. Wilso A. Mormon (editors), dium "heh cd and Evolution. CSIRO, Mel- bou Caddick, 1. , P. J. Rudall & P. Wilkin. 2000. Floral жеш ad and development in Dioscoreales. Feddes Hepert. 111: get cs 4 W., М. R. dii H. G. — Hills, J. ©. Сопгап, ^ Fay, L. R. == E e 9s Е: т = Ф ы ^ "n z c / ay, Caddick, K. M. Cal un & . 1995. Molecular ي‎ of kapes i 109-137 in P. J. Rudall, P. J. Cribb, Cutler & C. J. Humphries (editors), ои хм Systematics and Evolution. Royal Bo- tanic Gardens, S. Soltis, | P. S. Soltis, P. J. Rudall, M. F. Fay, Kores, T. J. Givnish Higher-level i stematics of the monocotyledons: An as- sessment of current knowledge and a new classification. D. Pp. 3-16 in k. L. Wilson & A. Morrison (editors), 630 Annals of the Missouri Botanical Garden Monocots: Systematics and Evolution. CSIRO, Mel- . 1993. The Woody Iridaceae. Timber Press, Port- meat land, Олок. Cheadle, . 1963. Vessels in Iridaceae. Phytomorphology 13: 2¢ Clifford, H. T. 1998. e ес ыйы Pp. 236-238 in К. Kubitzki (editor), The Far s and Ge enera of Vascular Plants, Vol. Flowering Plants Monocotsledons = Lilianae. ua Heidelber Cocucci, & S. Vogel. 200 of Sisyrinchium species (icu eae em and their pollinators in southe m es America. Flora 196: 26—46 Cruden, R 19 no systematics of Rigidella (lrida- ceae). Brittonia 2 7-225. Dahlgren, R. & H. T. Cito 1982. The Monocotyledons: A Comparative en Academic $i 'ss, London. ———, ——— & Yeo. 1985. The Families of the Miniocobledons. eee - Verlag. Berlin. Daumann, E. 1935. Die systematische Bedeutung des Blutennektariums der Gattung Iris. Beih. Bot. Centralbl. 53B: 525-625. 1970. Das Blütennektarium der Monocotyledo- riicksichtigung seiner syste- Feddes Oil-producing flowers nen unter besonderer Bei matisc ied und phslogenetise then Bedeutung. Кереп. 80: 463—: Je bs M. P 19 7. The gen nus Romulea in South Africa. . Afric ‘an Bot., Suppl. 1974a. Die Suid О genus Syringodea. J. 8. African 1 Bot 40: 201-25 . 1974b. шина, 7 nuwe = van die Irida- ceae. 7. S. African Bot. 40: 301-306 —— —. 1979. The Mou genus reium J. S. African Bot. 45: 295-375. . 1982. The i ‘an genus Tritonia Ker-Gawler 1. Afri vi Bot. 48: 105-1063. . The ^ ‘an genus Crocosmia Planchon. J. S. now А en 50: 463—502. . 1985. Re vision of the South African 4 nus Chas- manthe | eae). S. African J. Bot. 61 Donato, R., C. Leach & 1. G. Conran. 2000. xad of Dietes (Irid: aceae) inferred from ITS2 sequences. Pp. 407—413 in K. L. Wilson & D. A. Morrison (енеге), Е ттен JS: Monocots: Systematics and CSIRO, Mel- bourne. Dressler, R. L. 1993. Phylogeny and Classific ation of the Orchid dag Cambridge Univ. Press, Cambridge. Endress, P. K. 1995. Major Eee айа of n monocot flowers. Ms rid 79 in P. J. Rudall, P. J. Cribb, D. F Cutler . J. Humphries (editors "t ү нш тй ERE and Evolution, Royal Botanic Gardens, Kew. Fay, M. F., P. J. Rudall, S. Sullivan, K. L. Stobart, A. Y. de Bruijn, G. Reeves, К. Qamaruz- vane p -P. Hong, J. Joseph, W. J. Hahn, J. G. Conran & M. 2000. Phylogenetic studies of Asparagales pue on four plastid DNA loci. Pp. 360-371 i Wilson & D. А. Жел, eae S ln matics and Evolution. ww . Melbourne. Forcone, A., “Gale о & L. M. Bernardello. 1998. Floral nectar c ren mical composition of some species from Pa- Chase. tagonia. Biochem. Syst. Ecol. 25: 395—402. Goldblatt P. 1986. ps و‎ as of Southern Africa. Ann. оаа h | . 14. — iste жен : of the southern African genus m n 'eae-lridoideae). Ann. Missouri Bot Gard. de —509, das and classification of Iridaceae. : 607—621. Ann. eit Bot. Gard. 1998a. Iridaceae. Pp. 295-335 in K. Kubitzki (editas. ule: and Genera of Flowering Plants, Vol. S Flowering P М its—Monocotyledons—Lilianae. ns . T idelber ——— Jüb. Жел 'tion of Barnardiella, Galaxia, Gyn- ке и г Homeria, and Rc ссни in Mo- raea (Iridaceae: Irideae). Novon 8: 371— & P. Bernhardt. 1999, EN DM mechanics of "s species (Iridaceae) with a staminal column. nn. Missouri Bot. Gard. 86: 47-56. & . Thomas. 1997. Palynology, phylogenetic reconstruction and the classification of the Afro-Mada- gascan genus Aristea Aiton (Iridaceae). Ann. Missouri Bot. a 84: 263-284. & J. C. Manning. 1995. Phylogeny of the African genera i be a and Freesia (Iridaceae-lxioideae), and a new genus, ey dm к. Bot. 20: 161—178. ———— & ————. 1997. Ne p ew cies of Aristea sect. adn ue (lida eae) from South Africa and notes on the taxonomy and pollination biology of the section. No- von 7: 137-144. —— ( . 2000. The long- г id fly ay tion system in southern Africa. Ann. Missouri Bot Gard. 87: 146-170. — — — & ——. 2002. Notes and new species of Mo- raea (Iridaceae: pU a winter-rainfall south- ern d ‘a. Novon 12: : & P. J. Rudall. Ton Hep of the south- ern African ge ne Bobartia (Iridaceae-lIridoideae). S. African J. Bot. : 304—309. © h & P. J. Rudall. 1984. Occurrence of ets in Iridaceae and allied families and their phyloge netic significance. Ann. Missouri Bot. Gard. 71: 1013-1020. udall, V. I. Cheadle, L. J. Dorr & €. A. Affinities of the Madagascan endemic ceae or iC Meer ah Bull. Mus. sect. B, Adansonia 3: —, P. J. Williams: a genus Geosiris, Irida Nat. Hist. Nat., Paris, 4° sér., 9 s 239-248. = ]. Henrich. 1990. The genera of the AM ‘ит roe (Iridac [Re qr +): Phylogeny and relationships. Syst. Bot. 15: 497—510. ————, J. €. Manning & P. л 1995. Pollination biology of Lapeirousia subg. Lapeirousia (Iridaceae) in uthern Africa; Floral divergence and adaptation for long- un fly-pollination. Ann. Missouri Bot. Gard. 82: 517 — : Р. | ыа & J. € 1998. Pollination of petaloid geophytes by monkey beetles (Scarabaeidae: Rutelinae: Hopliini) in southern Africa. Ann. Missouri Bot. Gard. 85 215-230. . J Manning. anning & P. Bernhardt. 1999, Evidence of bird pollination i in Iridaceae of southern Africa. Ad- ansonia sér. 3, 21: 25-40 ——— & ——— a. Adaptive radiation of pollination mec ШАША ў in ше African genus /xia (lri- daceae: Crocoideae). Ann. Missan Bot. Gard. 89; 564— E77 í — ——— Ё — ——. 2000b. Adaptive radiation of palliation mec hanisms : in Sparaxis (lridaceae: Ixioi- deae). Adansonia sér. 3, 22: 5 ы — — — Ё ——. 2001. Radiation of pollina- tion еп ms in Gladiolus d eae: ога. іп a. Ann. Missouri Bot. Gard. 88: 713-73. southern ا‎ Nünni, P. Bernhardt & J. C. Manning. in press. » Volume 90, Number 4 2003 Rudall et al. Floral Nectaries in lridaceae Floral d of es (Iridaceae: Crocoideae): How minor shifts in floral presentation change the pol- Tete spem. i. Missouri Bot. Gard. 1 S. D. 1992. Plant-animal relationships. Pp. 175— in R. a. cd as Ecology of Fynbos. Ox- ford Univ. Press, Ca Kocyan, A. & P. K. m "2001. Floral structure and eee Apostasia and Neuwiedia (Apostasioi- deae) and t relationships to other Orchidaceae. Int. un -961. 8. Ixioliriaceae. Pp. = їп К. Ки- җе к. Тһе e and Genera of Vascular Plants, Vol. 3. Flowering Plan Aen i a Lilianae. шт Heide а rg. Lee, J. 1994. Fatty Oil Production by Alophia drummondii (Iridaceae) and Modified Oil Collection Behavior of Centris О a aa Master’s Thesis, University ustin. . & P. Goldblatt. 1991. Seed coat structure in the БЕУ Cape Iridaceae, Nivenia, Klattia, and Witsenia. Bot. J. Linn. aoe 107: 387-404 & The pollination of Tritoniopsis к iflora ( (hides eae) by the oil-collecting bee Rediviva gigas (H ymenoptera: Melittidae): The first record of oil- sec s on in Afric 171-176. Mogensen, H. L. 1969. Floral ontogeny and an interpre- lation of the inferior ovary in Agave parri. Canad. J. Bot. an Iridaceae. S. African J. Bot. Mosen, E 1968. Fosteria, a new genus of Mexican Iri- laceae. d 20: 232-234. 970. The genus Tigridia (lridaceae) уз ыч апа Pd America. Univ. Calif. Publ. Bot. 54: 1-128. & R. W. Cruden. 1969. Sessilanthera, a new genus of American Iridaceae. Brittonia 21: 191—193. Newman, 1. V. 1928. The life history of Doryanthes excel- sa. 1. Some ecological and vegetative features and spore production. Proc. Linn. m N.S.W. 53: 499—538. Patil, D. A. & R. M. Pai. 1981. The floral anatomy of "i m excelsa Du (Agavaceae). Indian J. Bot. 4 —9, Payer. J. B. 1857. Traité d’Organogénie Comparée de la Fleur. Victor Masson, Paris. Prychid, C. J. & P. J. Rudall. 1999. Calcium oxalate сгуз- tals in monocotyledons: Structure and systematics. Ann. Bot. 84: 725-739 Ravenna, Р. 1974. Соба a new e of Central Amer- ican wi Bot. Not. 127: 104— W. Chase, P. Colla PIR ‚М.Е ox, B. LeJeune za-C bu 2001. Molecular systematics of Iridaceae: pene 'e ide four i tid DNA regions. Amer. J. Bot. 88: 2074—2 Rudall, P. J. 1986. Taxonomic sienien of ә чеш in Australasian ud Nordic J. Bot. 6: 277— ]« tomy and systematics at Iridaceae. ms J. Linn. Soc. m 1-21. . 2001. Centrifixed anther attachment in monocot- gd. Kew Bull. 56: 965-973. 2002a. Homologies of inferior i вда septal nectaries in monocotyledons. Int. J. Pl. 261-2706. 2002 ». Unique floral structures e iterative evo- lutionary Do mes in Asparagales: pes from a mor- phological analysis. Bot. Rev. 68: 488—509. R. M. Bateman. 2002. Roles of IE zygomorphy and homeosis in floral evolution: The gy- nostemium and labellum of orchids and sd lilioid monocots. Biol. Rev. 77: 403-441. & P. Goldblatt. 2001. Floral anatomy and system- atic position of Diplarrhena (Iridaceae): A new tribe Di- plarrheneae. Ann. Bot. (R ‚ А. Y. tomic xil and ош ona] investigation of Sisyrinchium and allied genera. Bot. Gaz. (Crawfordsville) 147: 466— 477. . Furness, M. W. Chase & M. F. Fay. 1997. Mic DoE and pollen sulcus in Aspara- gales TT Canad. J. 75: 408-4. ong, C. A. Furness, J. G. Conran, G. Kite, K. L. d & M. W. Chase. 2000. Consider the lil- ies—Systematics of Liliales. Pp 347—357 in K. L. Wil- son & D. A. Morrison (editors), per dis and Vere CSIRO, Melbourn Schmid, . Functional ii of the mor- lu E es of septal nectaries. Acta Bot. Neerl. 25—128. pes Reproductive versus extra-reproductive . Septal nectary anatomy and phy- logeny of thy aermoracee. Sys st. Bot. 18: 593—613. ————. 1998. Reversal in ovary position from inferior to superior in the Ha IS doors. Int. J. Pl. Sci. 159 66— y a Rudall. 1998. Tecophilaeaceae. Pp. 429— 4 . Kubitzki (editor), The Families and Genera or A e Plants, Vol. 3. Flowering Plants—Monocot- Yledons— Liliana Springer, Heidelber Sm . P. Ronse Decraene, P. Caris & P. J. Ru- dall 2000. Floral nectaries in кок Distri- bution and evolution. Pp. 230—240 in K. L. Wilson & D. A. Morrison (abi. prse а and Evolution. CSIRO J, Melbour Souza-Chies, T. T., & B. Lejeune. 1997. P with parsimony and igo our DNE the plastid gene rps4. Pl. Syst. Evol. 2 Steiner, K. E. 1998. Beetle E of m moraeas in South ue РІ. Syst. Evol. 209: 47-6 — —— & V. X e 1988. The association be- tween oil-proc wm flowers and oil-collecting bees in оч inire of southern LB Monogr. Syst. Bot. uri Bot. G в. 25: 259— К adaptation to oil-se- creting — c and Diascia. Evolution 44: 1701—1707. Van Heel, W. A. 1988. On the development of some gy- noecia with septal nectaries. Blumea 33: 477—5 Vogel, S. 1954. Blütenbiologische Typen als Elements der M Lap pen ina Bot. Stud. 1: 1—338. 1. Olproduzierende Blumen, die durch A „ Carter, E. Besin A ete bestiubt werden. Naturwiss. 58: 5 ——.. 1974 р 4. Olblumen und Mrammelhde Bienen. 8 Akad. Wiss. Lit. Math. Naturwiss. Kl. 7: 285-574. 1981. Bestäubungskonzepte der еш sin und m Ausdruck im System. Ber. Deutsch. Bot. 94: 663—675. 632 Annals of the Missouri Botanical Garden STATISTICAL SUMMARY OF SOME OF THE ACTIVITIES IN THE MissoURI BOTANICAL GARDEN HERBARIUM, 2002 Vascular Bryophyte Total Acquisition of Specimens Staff Collections 36,701 3,355 40,056 Purc 3,510 0 3.510 Exchange 25.162 1.208 26,370 Gifts 8,405 995 9.400 Total acquisitions 13,718 5,558 19,336 Mountings Newly mounted at MO 90,652 19.980 110,632 Specimens mounted when acquired 30,441 0 30,441 Repairs Specimens repaired 22,982 n/a 22,982 Specimens stamped 1,622 n/a 1,622 Total repairs 24.604 0 24.004 Specimens Sent On exchange 34.185 4.576 38,761 As gifts 13,798 1,622 15,420 Total 47,983 6,198 54,181 Loans Sent Total transactions 320 19 339 Total specimens 24.400 1.393 25,793 To U.S. institutions Transactions 165 7 172 Specimens 11.971 176 12,147 'To foreign institutions Transactions 155 12 167 Specimens 12.429 1.217 13,646 То student investigators Transactions 37 0 Specimens 3,868 0 3,868 To professional investigators Transactions 283 19 : Specimens 20,532 1,393 21,925 Loans Received Transactions 252 9 261 Specimens 18,519 624 19,143 From U.S.A. From abroad Total Visitors 226 106 332 During 2002, 142,695 specimens were accessioned into the oo 110,632 mounted at MO, 30.441 mounted when ac quired, and 1622 old MO specimens stampe ed (and numbered). The total number of mounted, accessioned spec n in the herbarium on 1 January 2003 was 5,361.911 (4.9 " 54.890 vascular plants and 407,021 bryophytes). The 30,441 vascular plants listed as “mounted when ac quired" are all from the Reed Herbarium, purchased in 2001, and iin about 20 percent of that acquisition. Volume 90, Number 4 Crosby 633 2003 Statistical Summary The Garden's herbarium is closely associated with its database management system, TROPICOS (see (www.mobot. org/mobot/research)). The charts below summarize some of the statistics from TROPICOS both for the calendar year 2002 and as year-end totals. Note that the specimen records in TROPICOS are primarily based on MO specimens, meaning that about thirty-eight percent of the bryophytes (an increase of about one percent over 2001) and thirty-three ercent of the vascular plants (an increase of about one percent) in the herbarium are now computerized, with an overall total of about thirty-four percent (an increase of about two percent). TROPICOS records—Calendar Year 2002 Additions Bryophytes Vascular Plants Total Specimens 13,318 111,708 125,026 Names 3,202 17,261 20,463 Synonyms 4,840 14,307 19,147 Distributions 3,822 24,398 28,220 ypes 1,311 13,004. 14,315 Bibliography 1.782 2,318 4.100 TROPICOS records— Year-End 2002 Totals Bryophytes Vascular Plants Total Specimens 156,616 1,659,817 1,816,433 Names 103,633 812,662 916,295 Synonyms 68,255 406,549 474.804 Distributions 42.850 823,163 866,013 ypes 8,894. 291,621 300,515 Bibliogray ‚һу 25,056 66,972 92,028 Specimens in herbarium 407,021 4,954,890 5,361,911 Percent of specimens computerized 38 33 34 In TROPICOS, шегше -based Synonymy is always linked to a reference in Bibliography and directly with at least two records in Names, the synonym, often a basionym, and the correct name according to the reference. Additional qoum may be dd from these direct links, e.g., all other combinations of a КЕ treated as а e a synonym of 1 name are also synonyms of it. —Marshall R. Crosby ANNALS OF THE MISSOURI BOTANICAL GARDEN: CHECKLIST FOR AUTHORS The Annals publishes original articles in systematic botany and related fields. Papers whose purpose is the establishment of new nomenclatural entities in vascular plants and Rather, should be submitted to Novon for consideration. (Novon bryophytes аге not accepted. they instructions are available on the Web or from the man- aging editor.) CONDITIONS FOR PUBLICATION Authors are urged to contact the managing editor prior to submitting manuscripts. Because the electronic file will be used by the printer for typesetting, adhere to the items checklist. Authors' editors and are sent to dos printer in electronic form (not ges). Every the xa should be clearly labe led with the first author's last it is important to listed | in the format section of the ' files are formatted by the Pn 'amera-ready ‚ electronic file sent by name, the first four letters of the taxon involved, and an indication of the files contents (i.e., text, figure, table, etc. Маи 'ripts that have been reviewed before submission will be subject to the full normal review process initiated » review. It is the author's responsibility to supply missing bibliographie information; the editor does not perform library research except by prior arrange- ment with authors not having access to crucial reference materials. Communication with the editor at all points of the publication process is encouraged. PAGE CHARGES 'age costs аге $80 per di although. charges may sometimes be reduced or waived. Changes in proof made by authors will be billed е to the author at $3.50 per line changed. CORRESPONDENCE Amy McPherson, Managing Editor Missouri Botanical Garden Press Р.О. Box 29 St. Loi, U.S.A MO 63166-0299 Missouri Botanical Garden Press Monsanto Center 4500 Sha w Bd, Street addr res St. Louis, MO 63110-2291 Email: annals@mobot.org General ÛJ Text is in English or Spanish on numbered pages lL] Manuscript is submitted as Microsoft Word [pre (rab file either on diskette, CD, or Zip Dis k or by electronic transfer (email or ftp; contact the MBG Press Assistant for ftp instructions). LJ Electronic file of manuscript is labeled with first au- hor's last name, first four letters of the taxon involved, О U чы and an indication of what the file contains, e.g., Celis Cipu text.doc, Celis Cipu Figurel.tif, Celis em Та- blel.doc, One copy of Г double-spac еа printed manuscript, in- cluding abstract, legends erature Cited, and ool, is enclosed. The printed manuscript is the same as the electronic file Original illustrations, тет lats, or equivalents are included with the original submission, with the unde ое that же may have to be re- vised based o w. [JPEG ie are permissible for the review process, if necessary, but are not ac- > the 0 See the printer's electronic illustration requirements, below. | Authorities, abbreviated according to Brummit & Pow- I's Authors of Plant Names, are provided for species’ names the first time they are mentioned in the text; they are provided in a table that includes names of species. Letter is enclosed containing the author's full mailing address, telephone number, fax number, and e-mail address (required). tables, specimen lists, Lit- their FT ceptable for sip A ticas e 2. Format of Text Right-hand ee is not justified, and words are not hyphenated t One space is inserted after periods, colons, question marks, and all other punctuation marks. First sentences of paragraphs are а d with a tab. 1 tween them are used to ; measurements, 5—7 mm; altitudes, 1500— L] A single hyphen is used in са words, for ex- ample, 9-nerved, oval-triangula L] Common Latin words or Kou are not italicized (e.g et al., i.e., sensu, etc. O Only taxon names at T rank of genus and below are italicized. L] Case! accents, umlauts, and other diacritical marks have been include ÛJ All figures and tables are cited in the text and are numbered in the order in which they are to appear. 3. Style Recent issue of the Annals is used as a model. i Sac Manual of Style, latest edition, refere is used as a 4. First Page Footnotes are typed as double-spaced paragraphs on the first. page. The first footnote contains acknowledg- . The second footnote is the author's address. Ad- пя for additional authors are each separate foot- notes. No other footnotes occur, except in tables, where unavoidable. Abstract & Key Words A one- paragraph € precedes the text. Papers in Spanish have an English abstract in addition to a Spanish resumen. Volume 90, Number 4 2003 [] lide Е" O [] For species with [] -H О The abstract is concise (1 paragraph) and includes brief statements about the paper's intent, materials and methods, results, and significance of findings. A brief list of key words immediately follows the Ab- act. Taxonomic Treatment Species entries are organized as follows: Heading, ver- nacular name(s), Latin diagnosis necessary), de- sc ription, ac iens summary, specimens examined, discussion. The discussions are б and follow the same mih e.g., diagnostic characteristics, distinction from similar species variation, distribution and ecolo- 2 2 gy. nomenclature and typification, us One paragraph per basionym is leds as аа Тахоп author, literature citation, type citation, e.g.. Beilsch- miedia latifolia eh Sa. Nishida, Ann. Missouri at d. 86: 680. oo latifolia Nees, Sys 4. o TYPE: Peru. Locality not in- dicated, 1835 (fl), Matthew 1433 е E! types, BM!, E!, К!, LE not seen, OXF not seen). Synonyms based on different types are place ed in sep- 150- агаіе paragraphs, each beginning with the basionym, followed by other combinations (if appropriate), and citation of the type. A brief Latin diagnosis for each new taxon is provided rather than a complete Latin description. infraspecific taxa: Description and discussion are composite (incorporating all infraspe- cific taxa) and parallel with other species descriptions. Descriptions of infraspecific taxa are parallel with one another (in the same species). All synonyms are listed under. the appropriate infras pecu лбе tax rar surements are metric. ” Hyphe ns are us ж for parenthetical extremes “peduncle (8.2--)14.3-- 28.0(--31.9) cm long,” unless intermec alues are ate va with (2)4(6) lo RS Length width are x" in "^ indi: manner: lamina 36.4- -82.8 X 9.1--16.8 c When relevant, nomina nuda, misapplied names, and perfluous names are included in the discussion fol- qn the description, but are not part of the formal synonymy. i. on of Types Exclamation Nds are ROM for d ded pipa and types not seen are indicated as such US not see sn. 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 15 making € lec Ses ation, the phrase “designated here" is 8. Tables Tables are neat, double-spaced, and clearly presented. In most cases the printer will typeset these from hard copy. Captions are урей double-spaced as paragraphs at the tops of the tables Г] Each table starts on a separate sheet. О О [] [] LJ Specim [] CL] Herbariun 9. Abbreviations Periods are used after all abbreviations (which are minimized) except metric measures, compass direc- ions, and herbarium designation When dates are given as part of c vilis 'lion information, three-letter month abbreviations are used, except for months with four letters, which are spelled out in full. a abbreviated, and cities are spelled out. t. Louis, is acceptable. АНА reviated pe to B-P-H (Bo- tanico-Periodicum- EL m) and to B-P-H/S (Bo- tanico-Periodicum- ORC ду ама). Authors’ names are abbreviated according to Brummit & Powell's Authors of Plant Na Book titles are abbreviated according to Taxonomic Lit- erature, edition 2, but with initial letters capitalized. Book n are o Ыл out in the Literature Cited. If a m does not appear in B-P-H or TL-2, or if гө hin are not available, its or is fully p ates ar spelled out. Herbaria are icon eene to the most recent edition of Index Herbaric Abbreviated forms are not d for references in the text, except when citing the names of plants. If it is ssary to cite a particular page in the text, the form Smith (1998: 12) is used. 10. Specimens Examined If many specimens were examined, those cited in the ecimens collector, followed by collection number, followed by the number of the taxon in the text. Names (including initial(s)) of first and second collector are provided, “et al.” if three or more ens are cited in the text as follows: Additional specimens examined (o XICO. Oaxaca: Sierra San Pedro 12°37'N, 85?14' W, 950--1100 m, 3 Feb. 1987 gensen 665 (BM, G, K, US). [Dates and reproductive status are optional but are omitted from longer lists.] er in the same paragraph, e.g., UN- C d are run togethe COU A. Major political division: ssa GC TRY B. oe ie political division: ... Separate par- agrap ed for major нип. regions within major political divisions. 11. Specimen Vouchers and Genetic Sequences If the paper presents original data, associated herbar- ium vouchers are cited. [Vouchers for seed and/or other collections should be included where pertinent. De- p on the paper, reference to the original wild source may be required.] Vouchers are also cited from common names and uses taken from specimen labels. nd number, herbarium in w nd a clear annotation that the Te represents the voucher for the study in quest Nucleic acid or Kcd 'in sequences corresponding to equal or greater than 50 nucleotides are entered into m pj state the collector a the voucher is located, an 636 Annals of the Missouri Botanical Garden OO П с Û UUÛ Û [] U [] 0 ооо an appropriate data bank, e.g., GenBank/EMBL. The accession numbers are provided before publication. [Long sequences (exceeding two pages) will not be rou- 1 tinely published. Author accepts responsibility for establishing the ac- curacy of information provided. 12. Keys sys are clear and have been checked carefully for consistency an the descriptions. Leads of each cc 5 — e, et are para Dic hotomous m are indented. Infraspecific taxa are keyed separately, not in species "ys 13. Literature Cited The Literature ete contains full citations of all ref- erences cited in t All entries in i erence Cited are cited in the text. Spelling of author(s) name(s) and years of publication have been double-checked. All entries have been verified against or iginal sources, especially journal titles, accents, diacritical marks, and spelling in languages other than English, Periodicals are listed as follows r initial(s). Year. Full title of article. Journal abbreviated i -H/S. Volume: pages. No parenthetical part numbers after volume numbers are given unless each part is paginated separately. For more than one author, this style is followed: au- initial(s), second author's initial(s), name. s: author’s last name, thors last name, last name & third author's initial(s), last Books appear as follows: author's last name, initial(s). Year. Full Unabbreviated Title (edited by Editor), 3rd ed., Vol. 2. Publisher, City of Publicatior Citations of work “in prep.," unpublished theses and and similar references to inaccessible sources have been eliminated or kept to a minimum. They are not necessarily included in the Literature Cit- ed. dissertations, 14. Illustrations Author(s) name(s) and figure number(s) are clearly e or plate, on both originals e review electronic а each fi a- beled with the first author’ 5 last nam , the Brat four ‘igure ü ‚“... Fig- dicated on each oe ‚ copies. пе is letters of the taxon, and ‘ ~ file extension clearly indicates what e Scale E appear on illustrations, photographs, and maps. Magnifications/reductions are not indicated in cap- ions. High-quality copies of the figures are included with each copy of the manuscript. Illustrations are pre- -sized to fit either column width er than 11 X 16 in. 29 x 40 cm). [Maximum size for printed illustra- О О UU U OO O U ae. ELECTRONIC pl o A r ~ le) ài А‹ 32 tions is 57, X 8Y, in.] Oversized or delicate figures are submitted as photostats, high-quality photoc opies, or electronic files following the printer's requirements. Figures are numbered in Arabic numerals in the order of their citation in the text. Parts of figures are labeled with capital letters. Photographs are crisp black-and-white prints. If sub- mitted electronically, they are prepared according to the printers requirements. Figures are grouped into Wap plates when pos- sible; edges of photographs are abutted. No stripping is dod en plate or figure seg- ments. [Printer will insert vimine Edges of figures are square Maps include reference to шш and longitude and are bounded by a fine border. Scanning electron micrographs are free of conspicuous с Da ging. Axes on graphs are all labeled. Captions provide all explanatory text. separate from other text, one paragraph for each group of esr and following the style in current issues of the Anne E on maps are us and reduction has been taken into consideratior Captions are ARTWORK—PRINTER’S SPECIFICATIONS We accept line art, halftones and color figures on disk via electronic delivery. Please include a printout or e email of the file direc Чогу that inc 5 the file name(s), = size and kin You should rods figure sizing specifications with the journal prior to disk submission, and the figures on disk should be the final size des Graphic. files Е be eed 1 TIFF, EPS, PDF file format. Applications that are БРЕТ Adobe шан Corel Draw, Adobe Illustrator, and Adobe d but please, if possible, "Save tif, eps, or pdf file. All line art 1 should be scanned at 1200 d.p.i. All grays- cale and color images should be scanned al 350 d.p.i. Color images should be submitted in CMYK mode, not RGB or Indexed Color mode Include the screen and printer fonts used within the graphic file. Use only Adobe Type 1 postscript fonts. True Type fonts and System * dies fonts should not rma be usec 1. If the files have been compressed, please indicate what compression program has been used The disk submission should be accompanied by a hard copy of the figure that exactly matches what is on the disk in the event that the disk is unusable. `СЕРТАВІЕ ELECTRONIC ARTWORK MEDIA floppy disks; Syquest cartridges; EZ Flyer; ZIP disks; JAZ disks; Magneto Optical disks; CD ROM. Art- work may also be submitted via email attachment or FTP. Ri Please contact Caren idgway, e-mail address: kridgway@allenpress.com, for more Susan. Volume 90, Number 4, pp. 491—636 of the ANNALS OF THE И IRI BOTANICAL GARDEN was published on December 24, 200 ANNALS OF THE MISSOURI BOTANICAL GARDEN VOLUME 90 2003 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 2003 ISSN 0026-6493 VOLUME 90 2003 AL-SHEHBAZ, IHSAN A. (see Marcus Koch, Ihsan A. Al-Shehbaz & Klaus Mum- menhoff) 151 AL-SHEHBAZ, IHSAN A. (see Steve L. O’Kane, Jr. & Ihsan A. Al-Shehbaz) _ ALVAREZ FERNANDEZ, INES. Systematics of Eurasian and North African Do- ronicum (Asteraceae: Senecioneae) 319 AZUMA, HIROSHI (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hiroshi Azuma, Sean W. Graham, Mare A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre) 466 BAEZA, CARLOS M. (see Hanna Weiss, Tod F. Stuessy, Jiirke Grau & Carlos M. aeza 56 BARKER, NIGEL P., Н. PETER LINDER, CYNTHIA M. Morton & MARK LYLE. The Paraphyly of Cortaderia (Danthonioideae; Poaceae): Evidence from Morphology and Chloroplast and Nuclear DNA Sequence Data — 1 BERNHARDT, PETER (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hi- roshi Azuma, Sean W. Graham, Mare A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre) 466 CHEN, ZHI-DUAN (see Shao-Wu Meng, Andrew W. Douglas, De-Zhu Li, Zhi- Duan Chen, Han-Xing Liang & Jun-Bo Yang) 502 CHORNESKY, ELIZABETH А. & JOHN M. RANDALL. The Threat of Invasive Alien Species to Biological Diversity: Setting а Future Course 67 DAVILA ARANDA, PATRICIA (see Claudio Delgadillo M., José Luis Villaseñor Ríos & Patricia Dávila Aranda) 25 DELGADILLO M., CLAUDIO, JOSE LUIS VILLASEÑOR Ríos & PATRICIA DAVILA ARANDA. Endemism in the Mexican Flora: A Comparative Study in Three Plant Groups 25 DENSLOW, JULIE S. Weeds in Paradise: Thoughts on the Invasibility of Tropical Islands 119 DOUGLAS, ANDREW W. (see Shao-Wu Meng, Andrew W. Douglas, De-Zhu Li, Zhi-Duan Chen, Han-Xing Liang & Jun-Bo Yang) 592 DUPRE, JEAN-LOUIS (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hi- roshi Azuma, Sean W. Graham, Marc A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre) 466 EDDIE, W. M. M., T. SHULKINA, J. GASKIN, R. C. HABERLE & R. K. JANSEN. Phylogeny of Campanulaceae S. Str. Inferred from ITS Sequences of Nu- clear Ribosomal DNA _ 554 EDDIE, W. M. M. (see Tatyana V. Shulkina, John F. Gaskin & W. M. M. Eddie) FRITSCH, PETER W. (see Yelin Huang, Peter W. Fritsch & Suhua Shi) — 491 GASKIN, JOHN F. Molecular Systematics and the Control of Invasive Plants: A Case Study of Tamarix (Tamaricaceae) 109 GASKIN, JOHN F. (see W. M. M. Eddie, T. Shulkina, J. Gaskin, R. C. Haberle & R. K. Jansen) GASKIN, JOHN F. (see Tatyana V. Shulkina, John F. Gaskin & W. M. M. Eddie) GOLDBLATT, PETER. A Synoptic Review of the African Genus Hesperantha (Ir- idaceae: Crocoideae) GOLDBLATT, PETER (see Paula J. Rudall, John C. Manning & Peter Goldblatt) GRAHAM, SEAN W. (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hiroshi Azuma, Sean W. Graham, Mare A. McPherson, dinis 5. Rai, Rowan К. Sage & Jean-Louis Dupre) GRAU, JÜRKE (see Hanna Weiss, Tod F. Stuessy, Jürke Grau & Carlos M. Baeza) HABERLE, R. C. (see W. M. M. Eddie, T. Shulkina, J. Gaskin, R. C. Haberle & R: E /апзеп escas dapes unten itia re pM ины be ER E Em LUE HUANG, YELIN, PETER W. FRITSCH & SUHUA SHI. A Revision of the Imbricate Group of Styrax Series Cyrta (Styracaceae) in Asia JAFFRE, TANGUY (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hiroshi Azuma, Sean W. Graham, Marc A. McPherson, Hardeep S. Rai, Rowan К. Sage & Jean-Louis Dupre) JANSEN, R. K. (see W. M. M. Eddie, T. Shulkina, J. Gaskin, R. C. Haberle & R. K. Jansen) JENKINS, CLINTON N. & STUART L. Pimm. How Big Is the Global Weed Patch? Косн, Marcus, IHSAN A. Ar-SHEHBAZ & KLAUS MUMMENHOFF. Molecular Systematics, Evolution, and канап acd in the Mustard Family (Brassicaceae) LACH, LORI. Invasive Ants: Unwanted Partners in Ant-Plant Interactions? -.... Li, B. (see H. Zhu, H. Wang, B. Li & P. Sirirugsa) Li, DE-ZHU (see Shao-Wu Meng, Andrew W. Douglas, De-Zhu Li, Zhi-Duan Chen, Han-Xing Liang & Jun-Bo Yang) LIANG, HAN-XING (see Shao-Wu Meng, Andrew W. Douglas, De-Zhu Li, Zhi- Duan Chen, Han-Xing Liang & Jun-Bo Yang) LINDER, H. PETER (see Nigel P. Barker, H. Peter Linder, Cynthia M. Morton & Mark Lyle) LYLE, MARK (see Nigel P. Barker, H. Peter Linder, Cynthia M. Morton & Mark Lyle) Mack, RICHARD N. on Naturalizations and Invasions in the Eastern United States: 1634—1860 — < —< — ы. McPuknsow, МАКС A. (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, eter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hi- roshi Azuma, Sean W. Graham, Marc A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre) MALLOCH, DAVE (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hiroshi Azuma, Sean W. Graham, Marc A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre) MANNING, JOHN C. (see Paula J. Rudall, John C. Manning & Peter Goldblatt) MENG, SHAO-Wu, ANDREW W. DOUGLAS, DE-ZHU Li, ZHI-DUAN CHEN, HAN- XING LIANG & JUN-Bo YANG. Phylogeny of Saururaceae Based on Mor- phology and Five Regions from Three Plant Genomes MORTON, CYNTHIA M. (see Nigel P. Barker, Н. Peter Linder, Cynthia M. Morton & Mark Lyle) MUMMENHOFF, KLAUS. (See Marcus Koch, Ihsan A. Al-Shehbaz & Klaus Mum- menhoff) O'KANE, STEVE L., JR. & IHSAN A. AL-SHEHBAZ. Phylogenetic Position and Generic Limits of Arabidopsis (Brassicaceae) Based on Sequences of Nu- clear Ribosomal DNA PELLMYR, OLLE. Yuccas, Yucca Moths, and Coevolution: A Review — PiMM, STUART L. (see Clinton N. Jenkins & Stuart L. Pimm) PiroLY, JOHN J., Ш (see Jon M. Ricketson & John J. Pipoly HI) — PoNTIERI, VINCENZA (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, eter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hi- roshi Azuma, Sean W. Graham, Marc A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre) Ош, YIN LONG (see Michael J. Zanis, Pamela S. Soltis, Yin Long Qiu, Elizabeth Zimmer & Douglas E. Soltis) RAI, HARDEEP S. (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hiroshi Azuma, Sean W. Graham, Marc A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre) RANDALL, JOHN M. (see Elizabeth A. Chornesky & John M. Randall) — REICHARD, SARAH HAYDEN & PETER S. WHITE. Biological Invasions, the 48th Annual Systematics Symposium of the Missouri Botanical Garden. In- vasion Biology: An Emerging Field of Study RICKETSON, JON M. & JOHN J. Preoty Ш. Revision of Ardisia subgenus Au- riculardisia (Myrsinaceae) RUDALL, PAULA J., JOHN C. MANNING & PETER GOLDBLATT. Evolution of Floral Nectaries in Iridaceae 466 466 466 129 466 67 64 179 613 SAGE, ROWAN К. (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter Н. Weston, Dave Malloch, Hiroshi Azuma, Sean W. Graham, Mare A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre) SAGE, TAMMY L. (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hiroshi Azuma, Sean W. Graham, Marc A. McPherson, Hardeep S. Rai, Rowan К. Sage & Jean-Louis Dupre) SHI, SUHUA (see Yelin Huang, Peter W. Fritsch & Suhua Shi) SHULKINA, TATYANA V., JOHN F. GaskiN & W. M. M. EDDIE. Morphological Studies Toward an Improved Classification of Campanulaceae S. Str. -..... SHULKINA, T. (see W. M. M. Eddie, T. Shulkina, J. Gaskin, R. C. Haberle & R. K. Jansen) SIRIRUGSA, P. (see H. Zhu, H. Wang, B. Li & P. Sirirugsa) SOLTIS, DOUGLAS E. (see Michael J. Zanis, Pamela S. Soltis, Yin Long Qiu, Elizabeth Zimmer & Douglas E. Soltis) SoLTIS, PAMELA S. (see Michael J. Zanis, Pamela S. Soltis, Yin Long Qiu, Elizabeth Zimmer & Douglas E. Soltis) STUESSY, Top F. (see Hanna Weiss, Tod F. Stuessy, Jürke Grau & Carlos M. aeza THIEN, LEONARD B., TAMMY L. SAGE, TANGUY JAFFRÉ, PETER BERNHARDT, VINCENZA PONTIERI, PETER H. WESTON, DAVE MALLOCH, HIROSHI AZU- MA, SEAN W. GRAHAM, MARC A. MCPHERSON, HARDEEP S. RAI, ROWAN F. SAGE & JEAN-LOUIS DUPRE. The Population Structure and Floral Bi- ology of Amborella trichopoda (Amborellaceae) VILLASEÑOR Ríos, José Luis (see Claudio Delgadillo M., José Luis Villaseñor Vos & Patricia Dávila Aranda) WANG, Н. (see Н. Zhu, Н. Wang, B. Li & P. Sirirugsa) Weiss, HANNA, Top К. SruEssv, JORKE Grau & CARLOS M. BAEZA. Chro- mosome Reports from South American Hypochaeris (Asteraceae) — WESTON, PETER Н. (see Leonard B. Thien, Tammy L. Sage, Tanguy Jaffré, Peter Bernhardt, Vincenza Pontieri, Peter H. Weston, Dave Malloch, Hi- roshi Azuma, Sean W. Graham, Marc A. McPherson, Hardeep S. Rai, Rowan F. Sage & Jean-Louis Dupre) WHITE, PETER S. (see Sarah Hayden Reichard & Peter S. White) |... YANG, JuN-Bo (see Shao-Wu Meng, Andrew W. Douglas, De-Zhu Li, Zhi-Duan Chen, Han-Xing Liang & Jun-Bo Yang) ZANIS, MICHAEL J. Phylogenetic Analyses and Perianth Evolution in Basal Angiosperms ZHU, H., Н. WANG, B. Li & P. SIRIRUGSA. Biogeography and Floristic Affinities of the Limestone Flora in Southern Yunnan, China ZIMMER, ELIZABETH (see Michael J. Zanis, Pamela S. Soltis, Yin Long Qiu, Elizabeth Zimmer & Douglas E. Soltis)... 466 466 49] 129 444 129 LL ЇЇ Li LE www.mbgpress.org CONTENTS A Revision of the Imbricate Group of Styrax Series Супа (Styracaceae) in Asia —. Yelin Huang, Peter W. Fritsch & FERE Shi 491 Phylogeny of Campanulaceae S. Str. Inferred from ITS Sequences of Nuclear Ribosomal BNA W. M. M. Eddie, T. Shulkina, J. Gaskin, R. C. Haberle.& R. K. Jansen 554 Morphological Studies Toward an Improved Classification of Campanulaceae S. Str. ....... Tatyana V. Shulkina, John F. Gaskin & W. M. M. Eddie 576 Phylogeny of Saururaceae Based on Morphology and Five Regions from Three Plant Genomes ......... Shao-Wu Meng, Andrew W. Douglas, De-Zhu. Li, Zhi-Duan Chen, Han-Xing Liang & Jun-Bo Yang 592 Phylogenetic Position and Generic Limits of Arabidopsis (Brassicaceae) Based on Sequences of Nuclear Ribosomal DNA — Steve L. O’Kane, Jr. & Ihsan A. Al-Shehbaz 603 Evolution of Floral Nectaries in Iridaceae Paula J. Rudall, John C. Manning & Peter Goldblatt 613 Statistical Summary of Some of the Activities in the Missouri Botanical Garden Herbarium, 2002 Marshall R. Crosby 632 Checklist for Authors ; 634 Index for Volume 90 < Cover illustration. : Tam marix pentańdra. Plate LXXIX i in Fi ties Rossica, edited by PS. Pallas, - oe _ KF. Friedrich, and J.J. Weitbrecht and published by Petropoli, Cie 1784—1788. A book ` of hand-colored. copper engravings of plants indigenous to European and Asian areas of the . Russian Empire, ауа one of a re collection of such books in eis library of the Missouri "s Бинен Garden.